Final Report Summary - INHERITANCE (INtegrated HEart Research In TrANslational genetics of dilated Cardiomyopathies in Europe)
Executive Summary:
INHERITANCE is a small-scale focused research project that aimed at characterising the genetic basis of familial Dilated Cardiomyopathy (DCM) and translating basic knowledge of the aetiology into routine clinical practice throughout Europe. DCM is a life-threatening heart muscle disorder characterized by the presence of dilatation and systolic impairment of the left or both ventricles that is unexplained by abnormal loading conditions or coronary artery disease. The prevalence is 1 in 2,500 adults with an annual incidence of between 5 and 8 per 100,000. DCM is one of the leading causes of heart failure due to systolic dysfunction, being responsible for 10,000 deaths per year in Europe, and the commonest indication for cardiac transplantation in adolescents and adults.
Over the course of the project, INHERITANCE partners achieved genotype and phenotype data that has been used to identify important new information with immediate translation in the everyday clinical practice. The greatest contribution has been the characterisation of patients with laminopathies, a subtype of DCM caused by mutations in the gene encoding for the nuclear envelope protein, Lamin AC (LMNA) (6-8% of all DCM). The disease mechanism is haploinsufficiency with decreased expression of both gene and protein in affected hearts. INHERITANCE has shown that laminopathies are associated with a predictable phenotype that begins in early adulthood with conduction defects, and subsequent progression to heart block, ventricular arrhythmia and finally to heart failure. Male patients show a worse prognosis due to a higher prevalence of malignant ventricular arrhythmias and end-stage heart failure. Careful phenotyping has also allowed developing predictive models that identify patients who might benefit from early initiation of specific therapies such as anticoagulation and prophylactic ICD implantation. A DCM with right ventricular involvement caused by LMNA mutations identical to that seen in patients with ARVC has shed new light on the common biological pathways that cause heart muscle disease. Genes encoding for desmosome proteins have been shown to cause DCM. Together with new data on dilated cardio-dystrophinopathies (7% of all DCM in males) and less common genes (LDB3, desmosome genes, each accounting for less that 5% of all DCM) INHERITANCE also confirmed the role of clinical markers and of patho-functional assessment in the interpretation of genetic and clinical data.
The project generated the basis for a novel classification/nosology system for cardiomyopathies, the MOGE(S) system, that has been endorsed by the WHF and is supported by a free app at http://moges.biomeris.com/moges.html. As TNM for cancer, MOGE(S) is a flexible and expandable system that integrates phenotype and genotype information for cardiomyopathies.
Using Genome Wide Association Studies (GWAS) and animal models, we identified novel disease pathways, candidate genes and pathophysiologic mechanisms of myocardial damage; by linkage analysis, we discovered and characterised a novel autosomal recessive atrial DCM associated with mutations of NPPA. We further described novel autosomal recessive channelopathies presenting with both atrial and ventricular dilatation. The implementation of new high throughput gene sequencing (NGS) in large multicentre clinical series yielded to an extensive atlas of mutations and variants, that now require large-scale phenotyping tools and novel statistical strategies for the analysis of gene-phenotype correlations. This goal is now being achieved through systems that integrate large-scale phenotyping, bioinformatics and functional analyses. The NGS pipeline has proved to be reproducible and suitable for everyday practice. The large number of potentially pathogenic variants has revealed the complex genetic architecture of DCM that requires maintenance of research and collaboration between INHERITANCE partners as well as extension to other countries for EU-wide uniform translation.
Project Context and Objectives:
PROJECT CONTEXT
INHERITANCE is a translational multidisciplinary, multi-centre research project aimed at discovering novel genes and translate basic knowledge of the aetiology and pathophysiology of familial dilated cardiomyopathies (DCM) into routine clinical practice. The project involved 11 partners from 8 EU countries (IT, UK, SW, DE, FR, ES, DK, NDL). The consortium was generated taking into account the scientific and clinical expertise as well as the existing clinical series and facilities that were necessary to achieve the objectives of the project. The project stemmed from an uncovered clinical need that was the genetic characterisation of DCM and the establishment of diagnostic work-up for probands and relatives carriers of mutations of genes causing DCM.
INHERITANCE used existing state-of-the-art genotyping technologies in order to implement cost-effective protocols for rapid molecular diagnosis of specific forms of DCM. The project was structured in 5 research areas [clinical cardiogenetics, -omics (including genetic testing, transcriptomics, proteomics and metabolomics, animal models, structural and treatment), animal models, structural studies and bioinformatics] articulated in 11 WPs; two of the 11 WPs developed bioinformatics databases and systems for knowledge management and data analysis.
These genetic data have been used to integrate novel clinical diagnostic algorithms employing advanced non-invasive imaging modalities with information derived from original genomic, transcriptomic, metabolomic, proteomic and functional studies in blood and heart tissue samples obtained from patients with genetic forms of DCM. The findings have been used to develop translational programmes for preclinical diagnosis, new treatment strategies, family-tailored clinical work-up and cost-effective use of medical resources across the EU.
DEFINITIONS OF THE DISEASE
Familial dilated cardiomyopathy (FDCM) is clinically defined as idiopathic dilated cardiomyopathy (DCM) that is proven to occur in two or more related family members. DCM is defined as a myocardial disorder characterized by the presence of left ventricular dilatation and systolic impairment, in the absence of abnormal loading conditions (e.g. hypertension, valve disease) or coronary artery disease sufficient to cause global systolic dysfunction. Right ventricular dilatation and dysfunction may also be present. Therefore, current diagnosis of DCM is based on the clinical phenotype and does not distinguish the causes of the disease. Given that several genes may cause similar phenotypes all current strategies for treatments are phenotype-based.
EPIDEMIOLOGY OF FAMILIAL (GENETIC) DCM
The prevalence of DCM is approximately 1 in 2,500 adults, with an annual incidence of between 5 and 8 per 100,000. About 40,000 DCM cases are expected in the INHERITANCE European countries. In children, the incidence is much lower (0.5 to 0.8 per 100,000 per year), but DCM is the commonest cardiomyopathy in the paediatric population. DCM is responsible for 10,000 deaths per year in Europe and is the commonest indication for cardiac transplantation in adolescents and adults. In fact, DCM constitutes the most common cause of end-stage heart failure treated with heart transplantation worldwide. More than 50% of heart transplants have been performed to date in DCM patients. Up to 50% of individuals with DCM have at least a first-degree relative who is affected; a further 20% of family members have isolated left ventricular enlargement with preserved systolic function. Many of them subsequently develop overt DCM. In spite of this knowledge, the majority of patients were neither made aware of the familial risk nor offered genetic testing and less than 1% of patients in Europe benefitted from genetic testing. Genetic testing could add diagnostic value to the “early” instrumental markers in this 20% of relatives, thus contributing to a diagnostic anticipation and identification of individuals that will develop the full clinical phenotype. Early instrumental markers such as left ventricular dilatation without overt dysfunction or borderline left ventricular dysfunction with normal ventricular size would gain diagnostic value in mutation carriers.
Before INHERITANCE, family screening and genetic studies had identified more than 40 disease-causing genes. The commonest genes that cause DCM are those encoding proteins of the sarcolemma, cytoskeleton, sarcomere, nuclear envelope and energy generation. Most DCMs (up to 90%) are inherited as autosomal dominant traits, while a minority are X-linked recessive, autosomal recessive, and matrilineal.
Variation of age of onset, disease severity and prognosis amongst members of families carrying the same causative mutation suggest that modifier genes play a role in disease expression and response to pharmacological therapy. However, previously performed studies were small and showed discordant results on the effect of potential genetic modifiers. Only a large scale, genome-wide screening study in patients with known causative mutations, such as the one proposed in INHERITANCE, can examine this question.
PHENOTYPE CHARACTERISATION
The current clinical assessment of the different genetic disorders that cause DCM is relatively crude, with little emphasis given to the search for potential diagnostic cardiac and systemic markers. As a result, specific genetic diagnoses are often missed in everyday clinical practice and the implications for other family members overlooked. The symptoms and signs of DCM are highly variable and depend partly on the degree of left ventricular dysfunction and on the underlying aetiology. Presentation can be acute, often precipitated by concurrent illness or arrhythmia, or chronic, preceding the diagnosis by many months or years.
DCM is increasingly diagnosed as an incidental finding in asymptomatic individuals during routine examination or family screening. Associated cardiac traits (such as left ventricular non-compaction, conduction tissue disease, WPW, short PR) and non-cardiac traits such as hearing loss, retinitis pigmentosa, myopathy, palpebral ptosis, granulocytopenia, methylglutaconic aciduria, premature cataract, learning difficulties, can be present and are important in guiding physicians towards a possible aetiology. These traits are variably associated with the DCM phenotype and can contribute to characterising the phenotype in families with DCM caused by mutations of different genes. They constitute easy-to-recognise, low cost markers that should be identified in the clinical phase of the proband/family evaluation. The family screening, which is the strategy implemented in INHERITANCE, confirmed the hypothesis that clinical markers that are specifically recurring in DCM caused by different genes can be present in the early phases of the disease and disappear in late phases, thus further supporting the strategy of learning about the disease phenotype from more family members observed in different phases of the natural history of DCM.
The prognosis of DCM is highly variable and may also depend on aetiology. Early survival studies suggested mortality in symptomatic adults with idiopathic DCM approaching 25% at one year and 50% at 5 years. More recent reports have shown better outcomes, with five-year survival rates of approximately 20%, perhaps reflecting earlier disease recognition and treatment, and advances in medical therapy. However, recent studies have shown that the arrhythmogenic risk is extremely high in certain subtypes of DCM, such as cardio-laminopathies.
Role of biochemical markers in diagnosis and management
The role of biochemical markers in the diagnosis and monitoring of specific sub-types of familial DCM is still largely unexplored. Elevated levels of serum creatine kinase (sCPK) can be raised in dystrophin-related DCM in X-linked forms, or a LMNA gene defect in autosomal dominant DCM, but more disease-specific biomarkers are unknown. INHERITANCE aimed at searching for novel peripheral markers using proteomic and metabolomic analyses in paired myocardial samples and biological fluids (blood and urine) from genotyped patients and healthy carriers of causative mutations.
Influence of a genetic diagnosis on treatment
Nowadays, patients with DCM are treated in accordance with international guidelines for the management of heart failure with little consideration of the possible influence of the underlying aetiology on the response to treatment. Recent studies suggest that this might result in sub-optimal or inappropriate therapy in some patients. The most striking example is that patients with mutations in the nuclear envelope protein Lamin AC (LMNA) are at high risk of conduction disease relatively early in the course of their disease, but several studies have shown that implantation of a pacemaker alone does not prevent sudden death. Therefore, knowing that a DCM patient is carrier of a LMNA gene mutation might be of major importance when deciding on device therapy.
OBJECTIVES
The objectives planned in INHERITANCE were:
1. Clinical characterisation of DCM phenotypes associated with mutations of known disease genes. This objective specifically aimed at establishing priority clinical characterisation in clinically malignant cardiomyopathies and providing the prevalence of clinical markers that recur in DCM caused by mutations of same genes. The most common clinical markers associated with DCM are an ECG marker, the conduction disease (Atrio-Ventricular Block) and a biochemical marker (the increased Serum Creatine Phosphokinase).
2. Implementation of diagnostic gene testing in DCM population. Before INHERITANCE, less than 1% of patients and families in the EU received genetic advice and counselling. The Project aimed at increasing this very low percentage of patients throughout the EU, both within the partner centres’ and other countries, through the dissemination activities.
3. Development and validation of novel pre-genetic tests based on measurement of RNA transcripts in patients with pre-specified mutations from peripheral blood. This objective was particularly focused on DCM caused by mutations of the three major genes associated with DCM at the beginning of the project: LMNA, DYS and LDB3.
4. Identification of novel predisposing disease genes and modifiers in humans by GWAS and in animal models. This objective was planned because of the extreme phenotype variability in families with DCM caused by mutations of the same genes and even in the affected members of families with DCM caused by the same mutation.
5. Identification of disease-specific metabolites that can be measured in serum or urine. This objective was planned to explore possible metabolomics biomarkers in the different types of DCM.
6. Clarification of the pathophysiological mechanisms of myocardial damage and dysfunction in animal models and evaluation of old and novel treatments.
7. Structural characterisation and molecular simulation analyses on at least one mutated protein (LMNA). This objective specifically aimed at achieving structural information about the effects of recurrent mutations of LMNA on the protein.
8. Novel clinical approach with old drugs in patients with pre-clinical DCM. At the time of planning and revision of the new Document of Work, this objective was needed because the concept of early diagnosis could be strengthened by the identification of the mutation causing the disease in the probands’ relatives.
9. Generation of a web-assisted database and of a Wiki-based collaborative system. This objective was planned because of the large amount of data that were expected to be generated in the project and the need of an “intelligent” fast and easy tool for combining clinical markers with the genetic make-up of the different DCMs.
10. Generation of e-tools able to assist data collection and analysis for DCM, with potential translation to other types of cardiomyopathies.
11. Development of new diagnostic nosology for an easy description of key issues of cardiomyopathies caused by mutation of the different genes, considering the morpho-functional phenotype, the organs/tissues other than the heart eventually involved in the disease and the type of inheritance as well as the genetic defects.
While the achievement of each objective represented a major advance in the diagnosis and treatment of DCM patients INHERITANCE provided an integration of all the objectives into new international guidelines/classifications/nosology systems for describing the information obtained from clinical characterisation of the disease and genetic testing, thus facilitating the clinical work-up in probands and families and implementation in large family series.
Project Results:
WP1 - PHENOTYPES IN DCM AND GENO-PHENOTYPE CORRELATIONS
Idiopathic dilated cardiomyopathy (DCM) is a life-threatening heart muscle disorder characterised by the presence of dilatation and systolic impairment of the left or both ventricles that is unexplained by abnormal loading conditions or coronary artery disease. The commonest presentation is with symptoms and signs of heart failure, with a minority presenting with ventricular arrhythmia or sudden cardiac death. The diagnosis of DCM is still essentially descriptive and uses imaging techniques to detect the morphological and functional cardiac phenotypes. DCM affects approximately 1 in 2,500 adults with an annual incidence of between 5 and 8 per 100,000. It is one of the leading causes of heart failure due to systolic dysfunction, being responsible for 10,000 deaths per year in Europe, and the commonest indication for cardiac transplantation in adolescents and adults. Up to 50% of DCM is inherited as an autosomal dominant trait.
Before INHERITANCE, few patients were made aware of the familial risk, and less than 1% was offered genetic testing. More than 40 disease-causing genes were known, encoding proteins of the sarcolemma, nuclear envelope, cytoskeleton, sarcomere, desmosome and proteins important in energy generation. After INHERITANCE, the proportion of informed, counselled and genotyped families significantly increased in the partners’ centres; each partner contributed to national and international scientific dissemination, thus expanding the awareness that DCM is a familial disease in most cases and that clinical family screening is the first, low cost and most powerful tool to identify familial DCM. Family screening is now recognized as essential contributor to the family health, both prevention of major cardiac events and optimally tailored treatments(1,2). Two position papers of the WG on Myocardial and Pericardial Diseases of the European Society of Cardiology reflect the strategy of family screening and “red flags” as markers of disease and the collaborative research activities of the Partners of INHERITANCE(3,4).
Before INHERITANCE, genetic testing in everyday clinical practice was limited by the cost, time and complexity of conventional Sanger-based sequencing technologies. Advances in high throughput sequencing (HTS) technology are demonstrating the potential to solve this problem by analysing substantially larger genomic regions at a lower cost than conventional capillary Sanger sequencing. The INHERITANCE partners implemented next generation sequencing technology (NGS) both as centralized strategy and as local implementation for genetic testing, with the advantage that each centre is now progressing with its own NGS tool. However, NGS may also pose new challenges. The potential to identify a large number of rare variants means that large-scale phenotyping of relatives and families is required to understand genotype-phenotype relations. The emerging information from NGS analysis of more than 80 disease genes and family screening studies is the high number of patients that carry more than one mutation/variant in different genes. In spite of this apparent genetic complexity, the inheritance in DCM is autosomal dominant in most families, thus indicating one causal mutation and a variable number of “modifier” mutations. The emerging question is therefore how to establish the role (cause or modifier) of the identified mutations. The answer is emerging from cascade family screening, follow-up of relatives and functional investigations of the role of the mutant proteins in affected hearts, or in vitro or experimental systems.
Geno-phenotype correlations and nosology of cardiomyopathies.
Major achievements of the project include advances in geno-phenotype correlations and genetic-based risk stratification. Dilated cardiolaminopathies, which are associated with a very high risk of life-threatening arrhythmias (5,6), were one of the major targets of the research in the geno-phenotype correlation setting. LMNA codes a protein of the nuclear envelope that is essential for nuclear membrane integrity and function. Disruption of the protein due to gene mutations causes structural damage of the nuclear membrane. In the four years of the research, cardiac phenotypes associated with LMNA defects have been proven to be far more than DCM only. The association of DCM phenotype with conduction disease (CD) is the most common phenotype in the cardiology setting. Up to 80% of patients identified as carriers of LMNA mutation demonstrate CD (AVB). Conduction disease often precedes the onset of the cardiomyopathy. However, a minority of patients with cardiolaminopathy does not show CD; they can present with atrial fibrillation, ventricular arrhythmias, arrhythmogenic right ventricular cardiomyopathy (ARVC) or show prominent myopathy traits. Occasionally reported, atrial standstill, Left Ventricular Non-Compaction, sudden cardiac death without evidence of prior cardiomyopathy were not identified in INHERITANCE series (5).
During the project, a subgroup of patients with a usual phenotype of conduction disease and borderline or definite arrhythmogenic right ventricular cardiomyopathy was identified. In order to determine whether LMNA mutations are a common cause of this overlapping cardiac phenotype, a cohort of one hundred and eight patients from unrelated families with ARVC was genetically tested using conventional Sanger sequencing for five desmosomal genes (plakoglobin, desmoplakin, plakophillin-2, desmoglein-2, desmocollin-2) and LMNA. As expected, approximately 60% were positive for desmosomal gene mutations but (4%) without desmosomal gene mutations carried LMNA variants. These individuals had severe disease with conduction abnormalities on their resting 12 lead ECG and atrial fibrillation. This work shows that LMNA mutations can be found in severe forms of ARVC and that LMNA should be added to desmosomal genes when genetically testing patients with suspected ARVC, particularly when they also have ECG evidence for conduction disease or atrial arrhythmia. Therefore genes such as LMNA that typically cause DCM have been shown to cause ARVC too (7).
Dilated Cardio-desmosomalopathies. We explored the hypothesis that mutations in desmosomal genes account for a proportion of otherwise unexplained cases of DCM. 89 heart transplant recipients were selected for the study. All of them underwent genetic screening using conventional Sanger sequencing of five desmosomal genes (plakophillin-2, desmoplakin, desmocollin-2, desmoglein and plakoglobin), and the findings from genetic evaluation were correlated with the clinical phenotypes and histological characteristics in explanted hearts. All relatives of patients with pathogenic mutations or variants of unknown significance were offered clinical and genetic evaluation. The major finding was a prevalence of pathogenic mutations in 13% of patients; an additional percentage of patients (6%) had genetic variants of unknown significance. Unexpectedly, the histology of explanted hearts from patients with pathogenic mutations was identical to that of patients without mutations. Family screening showed evidence of co-segregation of mutations with DCM phenotype in five families. The implications of this study are that mutations in desmosomal genes are frequent in patients with advanced DCM undergoing cardiac transplantation and are a cause of familial disease (8). This result demonstrates that mutations in desmosome genes are not ARVC-specific but also cause DCM. The recently revised diagnostic criteria of ARVC (Marcus et al. Eur Heart J. 2010;31:806-14) introduced the DCM-like predominantly left and the biventricular variants (30%), thus reducing the difference between ARVC and DCM and leading to the novel definition of Arrhytmogenic Cardiomyopathy (ARC).
Dilated cardiodystrophinopathies demonstrated a phenotype characterised by severely dilated and dysfunctioning hearts with either increased sCPK or overt myopathy/dystrophy in more than 80% of cases in which DYS gene defects are identified (9). We identified DYS defects in 34 out of 436 patients (7.8%) (Onset age 34 ± 11 years, age range 17 to 54 years); 30 had proven X-linked inheritance. The 2 phenotypes included DCM with mild skeletal myopathy and/or increased serum creatine phosphokinase (n=28) or DCM only (n=6). The EMB showed defective dystrophin immunostain. The DYS defects consisted of 21 in-frame deletions and 11 out-of-frame deletions as well as 1 stop and 1 splice-site mutation. In the overall series of male patients with DCM in which DYS defects have been systematically searched for with MLPA and sequencing (n=542), we identified 42 (7.6%) probands carriers of large deletions in DMD gene (n=38) and point mutations (1 splice, 1 stop and 2 missense). The early data obtained in 2000 using simple multiplex PCR for DYS defects (Arbustini et.al. JACC 2000; 35:1760-8) gave a proportion of 6.5% mutations in consecutive series of males with DCM; by introducing sequencing of the gene in those cases that showed abnormal Dys immunostaining of the endomyocardial biopsies (EMB), the proportion increased to 7.8% (34/436)(9) and is stable (7.75%, 42/542) at the end of the project. This result confirms that most DYS mutations causing DCM are large deletions, but also that testing the gene only using MLPA would miss 10% of cardiodystrophinopathies.
INHERITANCE results also demonstrated a high proportion of DCMs caused by mutations of genes that recur in less than 5% of the cases. The morpho-functional classification describes the phenotypes (dilated, arrhythmogenic right, biventricular and predominantly left cardiomyopathy, hypertrophic and restrictive cardiomyopathies) but misses the description of the genetic basis of the different cardiomyopathies calling for a novel nosology system that can provide both morpho-functional and etiological information.
Prognosis
We demonstrated important differences between genetic subtypes of DCM. Dilated cardiolaminopathies are associated with a high risk of life-threatening ventricular arrhythmias even when the left ventricular dilatation and dysfunction are mild (5). In a median follow-up period of 43 months (interquartile range: 17 to 101 months), 48 of 269 LMNA mutation carriers (18%) experienced a first episode of Myocardial Ventricular Arrhythmia (MVA): 11 persons received successful cardiopulmonary resuscitation, 25 received appropriate ICD (Implantable Cardioverter Defibrillator) treatment, and 12 persons died suddenly. Independent risk factors for MVA were non-sustained ventricular tachycardia, left ventricular ejection fraction <45% at the first clinical contact, male gender, and non-missense mutations (ins-del/truncating or mutations affecting splicing). MVA occurred only in persons with at least 2 of these risk factors. There was a cumulative risk for MVA per additional risk factor (5) (Figure 1.1). In a further study we demonstrated that male patients show a higher prevalence of left ventricular impairment than females, but there were no gender differences in the frequency of traits such as atrioventricular block, atrial tachyarrhythmias and non-sustained ventricular tachycardia (6). The results of these studies are relevant for the formulation of novel guidelines for ICD implantation in patients with cardiolaminopathies; in fact, given that current guidelines recommend ICD implantation only in patients with severe left ventricular dysfunction. However, subgroups of patients with DCM caused by mutations of genes causing malignant cardiac phenotypes (i.e. Laminopathies) risk life-threatening ventricular arrhythmias even at Left Ventricle (LV) size and function of mild severity and in any case far from the criteria for primary ICD prevention in current guidelines. We recently proposed a novel algorithm for primary prevention with ICD in dilated cardiolaminopathies (Figure 1.2) (10). Instrumental markers, such as J-point elevation at ECG also have been considered as potentially informative for the risk of sudden death in cardiolaminopathies and in other types of DCM (11).
Conversely, dilated cardiodystrophinopathies are associated with a low risk of life threatening ventricular arrhythmias, even when the left ventricular dilatation and dysfunction were severe (9). As anticipated above, we identified DYS defects in 34 of 436 patients (7.8%). During a median follow-up of 60 months (interquartile range: 11.25 to 101.34 months) we observed 17 events, all related to heart failure (HF) (median event-free survival: 83.5 months). Eight patients (23%) underwent transplantation, and 9 (26%) died of HF while waiting for transplantation. Eight patients received an implantable cardioverter-defibrillator, although none had device intervention during a median follow-up of 14 months (interquartile range: 5 to 25 months). No patient died suddenly, suffered syncope, or developed life-threatening ventricular arrhythmias (Figure 1.3). Based on this study, we concluded that DYS-related DCM should be suspected in male patients with increased serum creatine phosphokinase (82%) and X-linked inheritance. The disease shows a high risk of end-stage HF but a lower risk of life-threatening arrhythmias.
The spectrum of disease genes is now higher than it was expected when INHERITANCE was planned. The sequencing of 40 disease genes by Sanger before INHERITANCE and of >80 disease genes by NGS in large series of probands after INHERITANCE demonstrates that the descriptive diagnosis of DCM does not provide sufficient information about each single DCM case/family. In addition, phenotype variability in families is influenced by the presence of more than one mutation in two or more genes. This observation implies that one of the two or three (or more) mutations plays a determinant role (causative mutation), while the second mutation like acts as modifiers of the phenotype.
Novel classification/nosology of cardiomyopathies.
The morpho-functional classification of cardiomyopathies is useful for current clinical management that is based on the phenotype but does not meet the recent evidence that more than 80 known disease genes may cause phenotypically similar but genetically different cardiomyopathies. The observation that genetic testing generates subgroups of DCM that can be classified on the basis of their cause rather than only on their morpho-functional phenotype led to the development of a novel classification and nosology system which describes the morpho-functional phenotype (M), the involvement of organ/tissues other than heart (O), the familial or genetic inheritance of the disease (G), the specific cause/etiology of the cardiomyopathy, genetic or non genetic, (E) and the optional stage/functional status of the cardiomyopathy (S) using both ACC-AHA stage A-D and NYHA class I-IV. The MOGES classification largely stemmed from INHERITANCE, and was endorsed by the World Heart Federation in 2013. Since genetic testing is now increasingly becoming a part of clinical work-up and, based on the genetic heterogeneity, numerous new names are being developed for the description of cardiomyopathies associated with mutations in different genes, a comprehensive nosology was needed that could inform about clinical phenotype, genotype as well as inheritance and involvement of organs other than the heart. In the quest for a genetic terminology, terms such as desmosomalopathy, cytoskeletalopathy, sarcomyopathy, channelopathy, cardiodystrophinopathy, or cardiolaminopathy, zaspopathy, myotilinopathy, dystrophinopathy, alpha-beta crystallinopathy, desminopathy, caveolinopathy, calpainopathy, sarcoglycanopathy, dysferlinopathy, merosinopathy, emerinopathy are now being used. Adding to the fact that such nosology would grow unmanageable, the genetic notation would neither define the phenotype nor the extent of systemic involvement. The MOGE(S) nosology does not aim at generating novel morphofunctional phenotypes. Analogously to the TNM system for cancer, MOGE(S) simply describes the clinical phenotype, the involvement of more organs/tissues, the familial or non-familial disease, the specific cause of the disease and, eventually, the clinical stage. MOGE(S) obligates clinicians to use a uniform language essential for future research, registries, surveys and international initiatives. The proposed nomenclature is supported by a web-assisted app (http://moges.biomeris.com) and reflects the diagnostic work-up of cardiomyopathy for evaluation of the manifest/asymptomatic disease, strategy for family screening and the result of genetic testing (Figure 1.4). MOGE(S) has been published contemporaneously in JACC and in Global Heart, the new official journal of the WHF (12,13), and is now matter of clinical and scientific interest; its implementation and exploitation has started (14,15). It is expected that the nomenclature would also help the development of collaborative registries and facilitate the description of complex genetics and grouping cardiomyopathies on their cause. Overall, results achieved in this WP constitute a major heritage of the project for future research developments.
WP2 GENETIC TESTING
The activities done in this WP aimed at implementing genetic testing using both Sanger-based technology and Next Generation Sequencing tools for the screening of known and novel disease and candidate genes, including analysis of DNA from probands that tested negative for all known genes. In parallel, the partners maintained the comparative evaluation of the efficacy of Sanger vs. NGS for mutation detection. We further considered that need to focusing on sequencing of rare disease genes that were not previously investigated and, finally, on discovering novel cardiomyopathies and related genetic causes.
Therefore, the genetic testing in INHERITANCE included:
• Sanger-Based sequencing of known disease genes
• NGS-based analysis of both mitochondrial DNA and nuclear DNA genes
• Linkage analysis
Consensus and quality
The INHERITANCE consortium first established consensus criteria for the definition of disease-causing genetic variants (pathologic), Genetic Variants of Uncertain Significance (GVUS), subpolymorphisms and polymorphisms to be adopted through research in all the consortium labs; the consensus criteria provided partners with a common tool for their clinical and genetic activity and reporting, as well as DB implementation. The relevance of this task is related to: 1) the evidence that some of the genes associated with familial DCM are hypervariable and show a high number of variations (i.e. PKP2) mostly private; 2) the unexpectedly high number of double compound mutations; 3) the provisional interpretation of novel missense genetic variants until their role in the pathogenesis of the disease is confirmed. Amongst the criteria for the definition of genetic variants as disease causing, the segregation of the mutations with the phenotype in families was maintained as fundamental and remains a major consensus criterion contributing to the definition of pathogenetic mutations.
The consortium also established an EU NetwOrk for quality Controls (ENOC) for Sanger-based genetic testing, starting from the gene that most commonly causes familial DCM, LMNA; the results in blindly circulated samples from 6 labs documented 100% concordant results for mutation detection. To date, the sensitivity and specificity of different methods for mutation detection was evaluated in ENOQ activities, with 100% concordance on pathologic mutations. We established a Consortium-restricted Web-assisted mutation DB INHMUTDB that includes all genetic variations identified in INHERITANCE.
Next Generation Sequencing (NGS) tools were used for investigating mitochondrial and nuclear genes.
Mitochondrial DNA genes.
Mutations in mitochondrial DNA (mtDNA) may cause maternally-inherited cardiomyopathy and heart failure. In homoplasmy, all mtDNA copies contain the mutation. In heteroplasmy, there is a mixture of normal and mutant copies of mtDNA. The clinical phenotype of an affected individual depends on the type of genetic defect and the ratios of mutant and normal mtDNA in affected tissues. We determined the sensitivity of next-generation sequencing compared to Sanger sequencing for mutation detection in patients with mitochondrial cardiomyopathy. We studied 18 patients with mitochondrial cardiomyopathy and two with suspected mitochondrial disease. We "shotgun" sequenced PCR-amplified mtDNA and multiplexed using a single run on Roche's 454 Genome Sequencer. By mapping to the reference sequence, we obtained 1,300x average coverage per case and identified high-confidence variants (Figure 2.1). By comparing these to >400 mtDNA substitution variants detected by Sanger, we found 98% concordance in variant detection. Simulation studies showed that >95% of the homoplasmic variants were detected at a minimum sequence coverage of 20x while heteroplasmic variants required >200x coverage. Several Sanger "misses" were detected by 454 sequencing. These included the novel heteroplasmic 7501T>C in tRNA serine 1 in a patient with sudden cardiac death. These results support a role of next-generation sequencing in the discovery of novel mtDNA variants with heteroplasmy below the level reliably detected with Sanger sequencing. This comparative strategy will assist in the identification of mtDNA mutations and key genetic determinants for cardiomyopathy and mitochondrial disease (16,17).
Nuclear genes.
We developed a Next Generation Sequencing (NGS) pipeline including nuclear genes pertinent with DCM. Our strategy was first to concentrate a large series of DNA of homogeneously phenotyped patients from the consortium(18-20) and then, in parallel, implementing local NGS tools for future autonomous translational activities (Fig. 2.2). In a first study, we established an array-based subgenomic enrichment followed by next-generation sequencing to detect mutations in patients with hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM). With this approach, we show that the genomic region of interest can be enriched by a mean factor of 2169 compared with the coverage of the whole genome, resulting in high sequence coverage of selected disease genes and allowing us to define the genetic pathogenesis of cardiomyopathies in a single sequencing run. We detected disease-causing mutations in 6 patients, out of which, 2 microdeletions, and 4 point mutations. Furthermore, we identified several novel nonsynonymous variants, which are predicted to be harmful, and hence, might be potential disease mutations or modifiers for DCM or HCM.
We further progressed with a NGS study in a large cohort of patients from the INHERITANCE consortium. The partners submitted 1041 DNA samples to the P3 centre as established. 719 samples have been sequenced to date as planned (Figure 2.2). 272 samples had been previously analysed by Sanger sequencing. Therefore a large amount of data for comparative evaluations of the Sanger-based vs. NGS-based sequencing has been made available. Comparative evaluation has been done considering both common SNPs and mutations. Genomic regions for NGS were selected on the basis of their relevance for DCM. Methods for NGS including enrichment of the target region was achieved by in-solution hybridization of fragmented genomic DNA against custom capture oligonucleotides. Enrichment was achieved by using the Agilent SureSelect System with multiplexing. The mean sequence coverage was approximately 8000-fold. Sequencing was performed on an Illumina HiSeq 2000 sequencing platform in a multiplex approach according to optimised protocols for target-enriched samples (Illumina; San Diego, CA). Demultiplexing of the raw sequencing data and generation of fastq files have been performed using CASAVA v.1.8.2. . Mapping of the sequencing reads against human genome hg19 was done with the burrows-whealer alignment tool (BWA) (Li, Bioinformatics, 2009). Variant calling and quality filtering of variants was performed with Genome-Analysis-Toolkit (GATK) as well as annotation of known variants present in dbSNP132 database (DePristo, Nature Genetics, 2011). Annotation of disease-causing variants, sequence coverage metrics and prediction of genomic function was done using an undisclosed prototype system including hardware and software. Sequencing libraries have been prepared along with the development of a standardized data analysis pipeline for the detection of known or novel variants to enable the correlation of genotype data with phenotype data of the investigated patients. Samples from 50-100 probands were selected for NGS analysis from each clinical centre.
A mean number of 4.9 megabases were generated per patient with a mean coverage over a target region of 2196 base pairs. An automated analysis, using galaxy-tools as a framework, was established to utilise different software algorithms for mapping, variant calling, filtering, annotation and interpretation of a functional effect on protein function of the found variants. In terms of prevalence, the highest number of variations was seen in the TTN gene that codes the giant sarcomeric protein Titin followed by LDB3 (CYPHER-ZASP) and PKP2 (Plakophillin 2). After filtering, the percentage of variations with “potential” pathogenic effects dramatically decreases. PKP2 and LDB3 variants remain highly represented. We should consider that both genes show high variability and LDB3 code alternatively spliced products. The alternative splice sites are seen by NGS as splice variants. A large number of variants identified by NGS is known and potentially relevant (HGMB Biobase). Each partner received results in order to compare existing Sanger-based data with NGS results and to complete the family studies for segregation evaluation. This is in fact a fundamental step because each proband was found to carry more than one variant demonstrating in silico prediction of a potential pathologic effect (20,21).
Rare or still candidate disease genes analysis
The results achieved in GWAS and zebrafish, as well as pathologic studies and mechanisms of diseases investigated in the project, provided novel candidate genes whose screening was implemented in the second period of the project. Novel genes, additional to those included in the NGS list, were proven as associated with DCM. We screened novel candidate or disease genes (i.e. ANT1 (22), ABLIM, ACADVL, ANO5, Corin, EMD. DOLK, FHL1, MLIP, MURC, NDUFAF2, NEBL, NKX2-5, POLG1, PYGM, TBX5) in consecutive series of DCM probands or in smaller series of probands with clinical or pathologic markers addressing to the test of these genes. Overall, we identified the causative mutation in 54 probands and families with DCM and variable phenotypic traits. This result further highlights the genetic heterogeneity of the DCM and suggests the need of a novel gene-based nosology for DCM. Functional in vitro studies and segregation analyses on families are ongoing with the aim of clarifying the effects of the mutations on the protein.
Discovery of a novel DCM: The Atrial Dilated Cardiomyopathy.
We identified a novel dilated cardiomyopathy only affecting atria in a geographic isolate of the north-East of Italy. We defined the disease as atrial DCM (ADCM). The disease was characterized by: 1) clinical onset in adulthood; 2) bi-atrial dilatation up to giant size; 3) early supraventricular arrhythmias with progressive loss of atrial electrical activity to atrial standstill; 4) thromboembolic complications; 5) stable, normal left ventricular function and NYHA functional class during the long-term course of the disease (Figure 2.3). By linkage analysis we mapped a locus at 1p36.22 containing the natriuretic precursor A (NPPA) gene. By sequencing NPPA we identified a homozygous missense mutation (p.Arg150Gln) in all living affected individuals of the 6 families. All patients showed low serum levels of Atrial Natriuretic Peptide (ANP). Heterozygous mutation carriers were healthy and demonstrated normal levels of ANP. We concluded that Autosomal recessive ADCM is a rare disease associated with homozygous mutation of the NPPA gene and characterized by extreme atrial dilatation with standstill evolution, thromboembolic risk, preserved left ventricular function and severely decreased levels of ANP (23).
The importance of this rare disease goes beyond the local benefits for the population and affected families and extends to primary atrial diseases, including the familial forms of lone AF (24). After the identification of autosomal recessive primary atrial cardiomyopathy ending in atrial standstill, we identified a further autosomal recessive atrial disease (Sick Sinus Syndrome, SSS) associated with novel compound mutations in SCN5A (25). The natural history of this SSS demonstrated an early LV dilatation: this opens novel avenues for research in overlapping ventricular and atrial cardiomyopathies.
WP 3 GENOME-WIDE ASSOCIATION STUDIES (GWAS)
The activities carried out in this WP aimed at identifying the genetic component of multifactorial/sporadic DCM and characterize the susceptibility to develop the disease or its variable phenotypic expression (disease progression) of either sporadic or familial DCM patients by performing Genome-Wide-Association studies (GWAS) in large populations of patients and matched controls. We therefore analyzed the population of patients and matched controls that have been already recruited (phenotypic data, blood samples and DNA extraction) through the network members (>2,000 sporadic DCM patients in Paris and Heidelberg, and similar number of controls). Homogeneous diagnostic criteria and geographic origin of patients will be considered, and controls have been matched for age, gender and geographic origin. Two separate initial analyses/ GWAS have been performed in parallel on different populations, one conducted in Paris, one in Heidelberg. Replication and combined analyses included both groups.
We conducted the first Genome Wide Association Study (GWAS) to identify new loci contributing to sporadic DCM. In the GWAS initiated in Paris, we identified 2 loci implicated in sporadic DCM (with respective P-values of 9.5x10-10 and 4.0x10-12 in the combined data set of 2365 DCM patients and 2410 controls, including “discovery” and “replication” populations). We also demonstrated that gene BAG3, involved in the second locus in sporadic DCM, is also involved in familial DCM as causal mutations in this gene were identified (Figure 3.1). The genes underlying the two loci, HSPB7 and BAG3, encode for non-structural proteins, which suggests new mechanisms underlying DCM-associated heart failure (26).
Subphenotype and variable expressivity analysis
We further studied these two loci and hypothesized that the loci could be involved in the severity of systolic HF (heart failure) and not only in the susceptibility to develop HF. We genotyped polymorphisms (SNPs) previously associated with DCM (rs2234962 for BAG3 locus, rs10927875 and rs945417 for HSPB7 locus) in a European population of 1160 patients with HF due to CAD (Coronary Artery Disease, ischemic-HF) as well as in a European cohort of 1141 patients with DCM. The severity of syst HF was assessed by left ventricle ejection fraction (LVEF), LV end diastolic diameter (LVEDD), age and NYHA class (dyspnea at inclusion). We observed that LVEDD was significantly associated with rs2234962 (BAG3 locus) both in DCM patients (591 patients with LVEDD available) and ischemic-HF patients (348 patients with LVEDD available) (p=0.0086 and 0.012 respectively). Other parameters reflecting severity of HF were not different according to this SNP or the others. Therefore severity of HF was not related to the two loci, except BAG3 locus that was associated with LV diameter in both populations. This gene could therefore appear as a prognostic factor in patients with HF.
Replication of the two loci in a population of Ischemic heart failure.
We further studied the two loci identified through the GWAS and hypothesized that the loci could also be involved in systolic heart failure due to coronary artery disease (ischemic-HF: systolic dysfunction LVEF<45%, significant coronary artery stenosis). We genotyped SNPs previously associated with DCM (rs2234962 for BAG3 locus, rs10927875 and rs945417 for HSPB7 locus) in a European population of 1160 patients with ischemic-HF and in 1612 controls. All patients and controls are of Caucasian origin, from France, Germany and UK. We observed that SNPs related to HSPB7 locus were significantly associated with ischemic-HF (MAF of rs10927875 and rs945417 were less frequent in patients than controls, adjusted p value 0.0017 and 0.02 respectively) whereas SNP related to BAG3 locus was not. Therefore, of the two loci previously associated with DCM, HSPB7 locus was also associated with ischemic-HF whereas BAG3 locus was not, suggesting differential involvement according to the underlying cause of HF.
Functional analyses of BAG3 gene mutations.
In order to unravel the molecular mechanisms associated with the expression of BAG3 mutations identified in familial DCM forms, we performed experiments in cells and animal models; rat neonate cardiomyocyte were infected with adenovirus expressing the 2 missense mutant of the BAG domain of BAG3. Transfected cells were rapidly degenerating with concomitant blockade of the autophagy pathway as measured through LC3-II marker down regulation. This was in accordance with the described role of BAG3 as a key regulator of autophagy as a mechanism controlling protein quality control and proteostasis. We subsequently confirmed that BAG3 mutant expressing cells were unable to rescue a luciferase sensor activity suggesting a default in protein refolding capacity. As HSP70 is central in proteostasis regulation and one of the known interactor of the BAG domain, we performed protein interaction assays between HSP70 and mutant BAG3 proteins. We observed a lack of interaction as compared to WT BAG3. These preliminary and unpublished results are indicative of BAG3 mutation mechanisms related to a lack of interaction with HSP70 leading to protein quality control dysfunction.
A new GWAS initiated in Paris.
We are finalizing the results of a new GWAS performed on the whole-exome nsSNPs based Illumina BeadChip. We genotyped more than 2000 cases and controls in the discovery step. Among the ~55.000 SNP with frequency over 0.5% in the exomeChip, at least 6 regions, including the two previously identified (BAG3 and HSPB7 loci), display significant association after Bonferroni correction. These preliminary results had to be replicated in independent population by individual TaqMan based genotyping in the 1000 cases and control population provided by Heidelberg partner. Specific statistical analysis of the rare polymorphisms is also to be performed (gene burden test) and would likely lead to the identification of more contributing loci. This ongoing work will strongly contribute to the discovery of new pathways leading to DCM and subsequently will feed further collaborative projects within the consortium.
A second large-scale case-control GWAS on sporadic DCM was initiated at Heidelberg in collaboration with national and international partners (27). We applied a three-staged case–control design. Stage 1 (screening phase) included 909 genome-wide genotyped individuals of European descent with DCM recruited between 2005–08 and 2120 controls from the PopGen and KORA population-based cohorts. In a first replication stage, SNPs on locus 6p21 were genotyped in 2597 DCM cases from Germany and Italy recruited between 2007 and 2011 as well as in 4867 controls from the population-based SHIP study (SHIP-0 and SHIP-TREND) and from Italy. In a second replication stage, the lead SNP was replicated in a cohort of 637 DCM cases and 723 healthy controls representing European Caucasians of French descent. Overall, we analysed more than 4100 DCM cases and 7600 controls. We identified and successfully replicated multiple single nucleotide polymorphism on chromosome 6p21 (Figure 3.2). In the combined analysis, the most significant association signal was obtained for rs9262636 (P = 4.90 × 10−9) located in HCG22, which could again be replicated in an independent cohort. Taking advantage of expression quantitative trait loci (eQTL) as molecular phenotypes, we identified rs9262636 as an eQTL for several closely located genes encoding class I and class II major histocompatibility complex heavy chain receptors (27).
A further genome-wide study aimed at investigating for the first time whole peripheral blood miRNAs as novel biomarker candidates for non-ischaemic heart failure with reduced ejection fraction (HF-REF). We assessed genome-wide miRNA expression profiles in 53 HF-REF patients and 39 controls. We could identify and validate several miRNAs that show altered expression levels in non-ischaemic HF-REF, discriminating cases from controls both as single markers or when combined in a multivariate signature. In addition, we demonstrated that the miRNAs of this signature significantly correlate with disease severity as indicated by left ventricular ejection fraction (Figure 3.3). Our data further denote that miRNAs are potential biomarkers for systolic heart failure. Since their detection levels in whole blood are also related to the degree of left ventricular dysfunction, they may serve as objective molecular tools to assess disease severity and prognosis (28).
WP4 TRANSCRIPTOMICS
The activities planned in this WP aimed at testing the hypothesis that quantitative gene expression (QGE) in RNA from affected myocardium and parallel blood samples taken from patients with DCM and known disease mutation can reflect or show correlations with the genetic defects. In addition, we tested the hypothesis that QGE and protein expression are correlated and that a lower gene expression in the affected myocardium correlates with lower or abnormal protein expression. We explored QGE of genes whose mutations have been proven as causally linked with DCM. We further investigated the expression of mutated proteins in the affected myocardium, both endomyocardial biopsies and hearts excised at transplantation.
Quantitative gene expression (QGE)
• Our first major target gene was LMNA, which is the most common and malignant disease gene in DCM (Figure 4.1 design of the study). We found that the QGELMNA is significantly lower in the RNA from myocardium and peripheral blood of mutation carriers than wild type controls. Using the comparative ΔΔCT method, we evaluated the QGE of LMNA (QGE(LMNA)) in peripheral blood and myocardial RNA from carriers of LMNA mutations, versus blood and myocardial samples from DCM(LMNAWT) patients and CTRL(LMNAWT) individuals. After generating reference values in normal controls, QGE(LMNA) was performed in 311 consecutive patients and relatives, blind to genotype, to assess the predictive value of QGE(LMNA) for the identification of mutation carriers. In parallel, Lamin A/C was investigated in myocardial samples from DCM(LMNAMut) versus DCM(LMNAWT) versus normal hearts (CTRL(LMNAWT)). LMNA was significantly underexpressed in mRNA from peripheral blood and myocardium of DCM(LMNAMut) patients versus DCM(LMNAWT) and CTRL(LMNAWT). In 311 individuals, blind to genotype, the QGE(LMNA) showed 100% sensitivity and 87% specificity as a predictor of LMNA mutations. The receiver-operating characteristic curve analysis yielded an area under the curve of 0.957 (p < 0.001). Loss of protein in cardiomyocytes' nuclei was documented in DCM(LMNAMut) patients. The reduced expression of LMNA gene in blood is a novel potential predictive biomarker for dilated cardiolaminopathies with parallel loss of protein expression in cardiomyocyte nuclei (29) (Figure 4.2).
• The second investigated gene was LDB3 (CYPHER ZASP). The QGELDB3 assay showed that this gene is not expressed in the RNA (at levels detectable by real time PCR) from peripheral blood samples. Vice versa the gene is expressed in the myocardium where we found that the mean 2^(-DCt) in LDB3 mutation carriers was 186,6374072 vs. 237,4355256 in wild type samples (p = ns). Therefore, we did not find significant differences in the QGELDB3 of mutated vs. wild type DCM patients. We conclude that mutated LDB3 gene is not underexpressed in the myocardial tissue of patients with dilated cardiozaspopathies.
• The third gene investigated in this task was DYS. We did not investigate QGEDYS in RNA from peripheral blood samples, while we extensively measured the QGEDYS in myocardial samples from patients with DYS-related DCM. We found a statistically significant underexpression of QGEDYS in the myocardium of patients that carry mutations of the Dystrophin gene vs. wild type controls. The mean 2^(-DCt) in DYS mutation carriers was 2,178374496 vs. 4,240265316 in the myocardium of DCM patients with wild type DYS. Therefore mutated DYS gene is underexpressed in the myocardial tissue of patients with dilated cardiodystrophinopathies and the level of underexpression is independent on the mutation type.
Expression of mutated proteins in affected hearts
Although genetic testing seems to be superior to biopsy of the myocardium in cardiomyopathy for the diagnosis(30), the interpretation of the role of the genetic mutations needs the myocardial investigation. Therefore we progressed with the investigation of the protein expression at the myocardial levels, in both mutated and wild type control samples (Figure 4.3).
• Patients with cardiolaminopathies demonstrated loss of expression of the Lamin AC protein in the membrane of the myocyte nuclei; vice versa, nuclei of non-myocyte myocardial cells preserved the expression of the protein. Haploinsufficiency is probably the mechanism by which mutations in the LMNA gene damage the nuclear integrity and function. We further progressed with the investigation of the Lamin AC expression in the myocytes of the atrio-ventricular node and found a far more severe loss of protein expression than that seen in ordinary myocardial cells. Western blot studies confirmed the immunohistochemical results, with decreased expression of the protein, irrespective of the mutation types (29).
• In patients with LDB3 mutations we observed variable results ranging from a homogeneous mild attenuation of to normal expression of the protein in affected hearts. This finding could not be considered as specific and in any case was not sufficient for suspecting a LBD3 mutation. Same results were obtained in hearts of patients carrying different types of mutations. The observation that myocardial QGELDB3 and protein expression do not significantly differ in mutated and wild type hearts adds to the clinical evidence of little segregation of mutations with the phenotypes (recorded in geno-phenotype correlation research in WP1 and WP2) and raises serious doubts about the unique causative role of LDB3 mutations in familial DCM.
• Vice versa, the immunostain of heart samples or endomyocardial biopsies with anti-Dystrophin antibodies demonstrated a decreased expression of the protein at the sarcolemmal membrane level. We observed that the extent and pattern of distribution of DYS immunostain may vary, but when control myocardium from patients/individuals with wild type dystrophin is compared, the difference of immunostain is strongly significant. The results are easily reproducible and DYS immunostain is now a standardised test for endomyocardial biopsies. These results have been recently confirmed in 7 additional affected hearts excised at transplantation. Therefore, myocardial samples from hemyzygous mutated males carrying DYS defects can be easily screened with immunohistochemistry, using commercially available antibodies (9).
• In patients with DCM associated with mutation of TNNI3, immunostaining with anti-troponin I3 antibodies showed mild attenuation of the intensity of the immunostain but this finding was not considered as sufficiently specific for suspecting a genetic defect in TNNI3.
• To date, the immunohistochemical studies of myocardium from DCM associated with “sarcomeric” defects did not provide diagnostic contribution.
The tissue studies done in the project provided the methods and facilities for expanding the research activities from clinically stable phases of the disease to end stage heart failure in patients with genetic DCM that were treated with LVAD as bridge to transplantation (31) and for exploring the tissue markers that could be target of molecular imaging in the near future (32). In fact, new research on molecular imaging is now stemming from the methods and results achieved in INHERITANCE.
Desmosomal protein expression in hearts of patients with mutations.
We showed that genes causing ARVC may also cause DCM and vice versa. The protein and gene expression pattern are therefore relevant to understand the mechanisms of disease related to the causative gene. Therefore we investigated how DSG2 and DPS mutations contribute to the pathogenesis of ARVC. Initial mutation analysis of DSG2 in 71 probands identified the first family reported with recessively inherited ARVC due to a missense mutation. In addition, three recognized DSG2 mutations were identified in 12 families. These results and further mutation analyses of four additional desmosomal genes indicated that ARVC caused by DSG2 mutations is often transmitted by recessive or digenic inheritance. Because desmosomal proteins are also expressed in skin tissue, keratinocytes served as a cell model to investigate DSG2 protein expression by Western blotting, 2D-PAGE, and liquid chromatography-mass spectrometry. The results showed that heterozygous mutation carriers expressed both mutated and wild-type DSG2 proteins. These findings were consistent with the results obtained by immunohistochemistry of endomyocardial biopsies and epidermal tissue of mutation carriers, which indicated a normal cellular distribution of DSG2. The results suggested a dominant-negative effect of the mutated DSG2 proteins because they were incorporated into the desmosomes (33).
Since desmoplakin is part of all desmosomes, which are abundantly expressed in both myocardial and epidermal tissue and serve as intercellular mechanical junctions, we investigated protein expression in myocardial and epidermal tissue of ARVC and CS patients carrying DSP mutations in order to elucidate potential molecular disease mechanisms. Genetic investigations identified three ARVC patients carrying different heterozygous DSP mutations in addition to a homozygous DSP mutation in a CS patient. The protein expression of DSP in mutation carriers was evaluated in biopsies from myocardial and epidermal tissue by immunohistochemistry. Keratinocyte cultures were established from skin biopsies of mutation carriers and characterized by reverse transcriptase polymerase chain reaction, western blotting, and protein mass spectrometry. The results showed that the mutation carriers had abnormal DSP expression in both myocardial and epidermal tissue. The investigations revealed that the disease mechanisms varied accordingly to the specific types of DSP mutation identified and included haploinsufficiency, dominant-negative effects, or a combination hereof. Furthermore, the results suggest that the keratinocytes cultured from patients are a valuable and easily accessible resource to elucidate the effects of desmosomal gene mutations in humans (34).
Epigenetic factors
Since DCMs show remarkable variability in their age of onset, phenotypic presentation, and clinical course, and the disease mechanisms that modify the occurrence and progression of the disease are both genetic and epigenetic factors that may interact with environmental stimuli, we examined genome-wide cardiac DNA methylation in patients with idiopathic DCM and controls. We detected methylation differences in pathways related to heart disease, but also in genes with still unknown function in DCM or heart failure, namely Lymphocyte antigen 75 (LY75), Tyrosine kinase-type cell surface receptor HER3 (ERBB3), Homeobox B13 (HOXB13) and Adenosine receptor A2A (ADORA2A). Mass-spectrometric analysis and bisulphite-sequencing enabled to confirm the observed DNA methylation changes in independent cohorts. Aberrant DNA methylation in DCM patients was associated with significant changes in LY75 and ADORA2A mRNA expression, but not in ERBB3 and HOXB13. In vivo studies of orthologous ly75 and adora2a in zebra-fish demonstrate a functional role of these genes in adaptive or maladaptive pathways in heart failure (35).
Overall, gene and protein expression in DCM hearts from patients that carry mutations in different genes can be considered easily feasible tools for investigating the mechanisms of disease. This is especially useful considering that the NGS tools are demonstrating that a large proportion of patients carry more than one mutation in different genes, and that future research must concentrate on the role that the different mutations play in the generation and evolution of the cardiomyopathy.
WP5 PROTEOMICS & METABOLOMICS
The activities of WP5 were planned to develop a systems biology framework for understanding the phenotypic and transcriptional changes identified in patients with DCM. Samples were provided by both human cohorts, as described in WP 4, and animal models of DCM, as described in WP 7. Samples were examined by metabolomics and, for tissue requiring further characterization, by proteomics, to validate proposed mechanisms of pathology in the other WPs. In addition, high throughput metabolomics was applied to human blood plasma samples to identify potential metabolic biomarkers of DCM. We measured metabolic changes in heart tissue and blood plasma from patients with DCM in order to identify potential biomarkers of disease or drug response as well as understand some of the pathological changes that accompany DCM. Given the priority assigned to cardiac laminopathies and the genetic heterogeneity of familial DCM as well as the complexity of the metabolomic tests, we decided to focus on laminopathies and on the comparative evaluation of metabolomics profiles in DCM compared with hypertrophic cardiomyopathies (36).
We first generated an omic database (ISA, Investigation/Study/Assay) for recoding, collecting and management of all data generated by this work package in INHERITANCE. We then examined the metabolic changes in a mouse model of DCM (Research area 3, LMNA knock out mouse). The purpose was to compare the metabolic changes found previously in human tissue with a mouse model, to identify metabolic commonalities that could be considered the key to the development of dilated cardiomyopathy, and to use these metabolic changes as markers of disease progression or drug efficacy. In order to provide a wide coverage of the metabolome we developed a number of high throughput assays for the quantification of metabolites in both heart tissue and blood plasma. This should be transferable to any cardiac disease model and we are currently publishing these assays. In summary we investigated lipidomics, carnitine metabolities, amino acids and nucleotides.
• Open profiling lipidomics: Liquid chromatography mass spectrometry (LC-MS) was used to profile total intact lipids extracted using the Folch procedure. The lipid extract includes a range of metabolites including triglycerides, diglycerides, phospholipids and cholesterol esters. The assay depends on, first, separating lipids by their polarity using liquid chromatography, and then detecting individual lipid species using high-resolution mass spectrometry. However, for any mass of an individual lipid specie there may be a large number of isomers which are isobaric (i.e. possess the same mass and hence will be indistinguishable by MS). To address this issue, tandem mass spectrometry has been performed to first fragment lipid species and then identify the parent compound using the mass spectra of the subsequent fragments. The total fatty acid complement of extracts of heart tissue was measured by gas chromatography (GC). Fatty acids are converted into fatty acid methyl esters as part of a transesterification derivatization and then analyzed by GC-MS. We have published these methods to allow others to adopt them for their metabolomics/lipidomics studies (37).
• Targeted analysis of carnitine metabolites. The carnitine shuttle allows the transport of fatty acids across the inner mitochondrial membrane, and given the central importance of fatty acid metabolism (beta-oxidation) in the heart, the carnitine derivatives are important measures of mitochondrial function and substrate selection. Because of this we developed a triple quadrupole LC-MS assay to profile 36 carnitine derivatives in heart tissue and blood plasma. In this assay either plasma or extracted tissue are analyzed (both aqueous and lipid fractions of the Folch extraction). Carnitines are measured on a Quattro Premiere LC-MS following butylation using 3M HCl in butanol. The original direct infusion method has now been replaced with a method using a C8 LC column to provide some chromatography to reduce ion suppression from polar lipids like phospholipids. A parent ion scan of fragments of 85 Daltons is performed to detect and quantify the carnitines. We have also applied this assay to other cardiac mouse models.
• Amino acid analysis. A quantitative assay was developed relying on polar reverse phase chromatography of the butylated amino acid derivatives. This assay is a targeted analysis of a range of amino acids and their derivatives including glutathione (both oxidized and reduced), nitrotyrosine, 3-hydroxy tyrosine, and dityrosine. This has allowed a range of amino acids to be detected in intact tissue in a quantitative manner. In addition we have extended this assay to include citric acid cycle intermediates.
• Nucleotides. An assay was developed for nucleotides to address two issues associated with DCM – i. changes in cardiac energetics and hence high energy phosphates such as ATP and ii. oxidized nucleotides as markers of ROS damage. The assay measures ATP, ADP, AMP, cAMP, GTP, GDP, GMP, cGMP, CTP, CDP, CMP, UMP, UDP, UTP NAD, acetyl-CoA, 8-oxo-guanine and 8-hydroxy-2-deoxyguanosine.
Patients with DCM vs. other cardiomyopathies
Application of our targeted analysis to the profiling of HCM and DCM human samples using valvular and ischemic cardiomyopathy disease tissue as controls were readily discriminated by principal component analysis of the metabolomics data. The model showed that the groups clearly separate by test centre and batch suggesting an effect associated with sample gathering. However, analysis of the individual clusters by a similar unsupervised method reveals a significant difference between control and disease groups. The analysis of the loadings plot shows that GTP, ATP, oxidised glutathione and ADP are increased in the disease group compared with the control group whereas glucose and fructose 6 phosphate, reduced glutathione, GMP and adenosyl methionine are increased in the control group. Given the test centre and batch clustering HCM samples were investigated further alone in order to glean biological information from the data and some discrimination could be determined according to genotype. Thus, metabolomics and lipidomics show promise in terms of discriminating individuals with different forms of DCM.
Analysis of plasma
We investigated 14 samples of unknown disease mutation, 1 of dystrophin gene mutation and 11 of the lamin A/C mutation from Beneficiary 1 and 20 samples of known gene mutation and 16 samples of unknown gene mutation from Beneficiary 9. These samples were examined using the assays developed for analysis of heart tissue in the hope of finding metabolic changes that extend to blood plasma and could be used to monitor non-invasively disease severity – an important issue in a set of diseases with variable penetrance not necessarily predicted by the gene mutation. High resolution 1H NMR spectroscopy, targeted amino acid analysis and the carnitine assay were performed to cover key metabolites in the metabolome. For each assay condition it was possible to discriminate blood plasma taken from beneficiaries 1 and 9, despite the samples being taken using an identical SOP. This observation is relevant for any future development planned in the setting of metabolomics studies, indicating that conditions other than those commonly respected and planned in a clinical SOP for genomics should be included and analysed.
Expanded model of the metabolome of DCM
Human DCM. In this task we report changes in metabolism in heart tissues of DCM/ HCM patients vs. healthy individuals. Then we compared this analysis with results on blood plasma. We confirmed the identity of the metabolites we analyzed using gas chromatography mass spectrometry (GC-MS) and nuclear magnetic resonance (NMR) spectroscopy. We found that GTP, ATP, oxidised glutathione and ADP are increased in the DCM disease group relative to the control group, whereas glucose and fructose 6 phosphate, reduced glutathione, GMP and adenosyl methionine are increased in the control group. We confirmed the identity of the metabolites we analyzed using GC-MS and NMR spectroscopy.
Mice DCM. (Figures 5.1 5.2) We used the assays described above to examine metabolism changes in heart tissue from LMNA null mice. These mice develop severe DCM in 5-6 weeks when homozygous and across 6-12 months when heterozygous. Analysis of the amino acid compliment of the heart tissue demonstrated that we could readily distinguish effects due to age, discriminating 2 week and 5 week-old animals. In addition, it was possible to discriminate homozygous animals from controls and heterozygous, indicating that DCM induced alterations in amino acid metabolism. In a similar manner we examined the carnitine profile of the heart tissue. However, only clustering according to age of animals was clearly detected by the multivariate statistics. Examining nucleotides and redox active compounds by LC-MS, again separation was detected associated with both age of animals and genotype, with a clear separation evident from 5 week old homozygous animals from 5 week heterozygous and control animals. This was in part driven by the concentration of glutathione present in the heart tissue.
We also analyzed heart tissue from 40 week old heterozygous LMNA mice. Tissue was analyzed in terms of amino acid content, acylcarnitines and polar metabolites (TCA cycle and glycolytic intermediates). While no separation could be detected for the acylcarnitine dataset, a weak model was built for amino acid metabolism and polar metabolites. In 40 weeks heterozygous laminopathic mouse hearts vs. controls (WP6) and after validation of the model using 100 random permutations of sample class membership we showed an increase of aspartate and glutamate in the heterozygous group with glutamine, alanine and phenylalanine decreased. Thus, we were able to determine both metabolic changes in homozygous and heterozygous mice associated with the LMNA mutation.
WP6 ANIMAL MODELS: ZEBRAFISH
The strategy of research for experimental models of DCM was addressed to both investigating phenotype in knock out, overexpression and transgenic zebrafish embryos and adults, to discover and characterize novel disease genes (35-38).
ZEBRAFISH studies were developed because a large proportion of familial and sporadic DCM cannot be explained by mutations in the disease genes that were known when INHERITANCE was planned (38). In fact, although classical human genetic approaches such as linkage analyses and candidate gene testing led to the identification of more than 40 DCM genes, known before INHERITANCE, some DCM families only counted one affected survivor but the family story and clinical report document a high number of DCM-related deaths. Therefore, conventional linkage studies are not feasible. Furthermore, for many mutations identified in human DCM the molecular mechanisms by which they lead to impairment of cardiac contractility and arrhythmias are not well understood. Large-scale forward and reverse genetic approaches in zebrafish were considered as powerful tools for the identification of disease genes for humans. Prior studies on zebrafish led to the identification of disease/candidate genes that cause DCM. Therefore, research in WP6, zebrafish, aimed at revealing the in vivo role of distinct DCM-associated regions (in cooperation with WP3), disease genes and distinct DCM gene mutations. With this approach, the specific aims were: 1) to characterize known DCM genes that are within the focus of INHERITANCE; 2) to evaluate the effects of selected mutations within frequent DCM genes, such as lamin A/C; 3) to characterize genes within genomic regions found to be associated with human DCM (WP3). We generated transgenic zebrafish lines for selected DCM-causing mutations. Knock-down, overexpression and transgenic zebrafish embryos and adults were functionally, structurally and molecularly characterized in detail (Figure 6.1). The results of these experiments revealed the functional role of newly identified genes associated with DCM and pathways crucial in the pathogenesis of DCM.(4 novel candidate genes: PINCH1 and PINCH2, SMYD1 and KCNQ2). The translational screening of these genes led to the identification of mutations in patients with DCM (38-41).
Functional, structural and molecular characterization of knockdown, overexpression and transgenic zebrafish embryos and adults.
We first established the strategies and methods for the characterization of genetically manipulated zebrafish as model organisms for human DCM. The goal to identify orthologs of human DCM genes, design antisense oligonucleotides and generate transient loss-of-function zebrafish was reached. Accordingly, these knockdown zebrafish represent valuable models to study the molecular pathways associated with DCM. The following zebrafish orthologs have been identified and are available for knockdown or overexpression studies within the other projects: d-Sarcoglycan, Myosin-light-chain 1, Myosin-light-chain 2, Vinculin, Lamin A, Muscle LIM protein, Myosin Binding Protein, Titin-cap, Desmin, a-Tropomyosin 4, Troponin-T2, Integrin-linked Kinase, Desmin, Nexilin, KCNQ2, Affixin, Ablim3, Dystrophin. The homology of these zebrafish proteins ranges between 40 to 95%. We designed specific antisense oligonulceotides to characterize loss-of-function effects for many of them. As an example, we characterized the novel DCM candidate genes PINCH1, PINCH2, Smyd1, KCNQ2. All knockdowns with the exception of KCNQ2 resulted in a DCM-like phenotype in zebrafish. This phenotype was characterized by decreased ventricular function (impaired Fractional Shortening) and reduced blood circulation and pericardial effusion (39,40).
To prove the causal role of a novel genetic variant for the establishment of a cardiomyopathy phenotype, overexpression studies of mutated genes in zebrafish were carried out. The following zebrafish orthologs of DCM (candidate) genes were full-length cloned into Gateway-destination vectors and selected novel human mutations were introduced in ILK, LMNA, PINCH1/2. For instance, we identified the novel, non-synonymous variant P70L in ILK gene by NGS. Next, we explored whether the P70L mutation affects protein confirmation. The mutation of the conserved residue is predicted to lead to changes in the secondary ILK protein structure. In the functional assays in zebrafish, we identified that ILK p.P70L is a hypomorph and hence cannot compensate for loss-of wt ILK function. Furthermore, as shown by co-immunoprecipitation, ILK P70L shows disturbed interaction with PINCH proteins, partners of ILK in the ILK-PINCH-Parvin mechanotransduction complex. Since alterations of the DNA can also be on an epigenetic level, we investigated the role of recently identified modifications of DNA methylation of several genes found in myocardium of patients with DCM and controls. For two genes, namely ly75 and adora2a we could dissect a functional role of the changed expression in the heart. These results are of great importance, since they underline a novel disease pathophysiology/mechanism of DCM, which might also be important for diagnosis and treatment.
As detailed above, the zebrafish allows to analyze the function of many different cardiac genes in a manageable time period. In fact, this leads to a rapid accumulation of functional data that need to be structured, organized and retrieved by the scientists. Hence, we decided to establish a project-internal zebrafish DCM database that allows fast access to data on DCM associated genes. Data encompass sequence homology, available Morpholinos, their corresponding knockdown phenotype and gene expression as well as subcellular localization. Data to be integrated were retrieved (from published and unpublished projects; references are included in the database; unpublished data are jointly generated with Prof. Dr. Rottbauer) and organized using standard bioinformatics tools and manual curation. Data were introduced in a database table format and include expression of genes, Morpholino sequences, target genes, quality controls measures, loss-of-function phenotypes and others. A first database for the integration of functional genomics data of DCM genes and candidates has been generated. We will further expand the system to assist ongoing and future research projects. We are currently investigating the functional effects of novel mutations identified by NGS mutation screening in the large patient cohort of the consortium.
Overexpression of selected DCM mutations in zebrafish and generation of transgenic lines.
Transgenic zebrafish lines are a valuable tool to dissect the impact of distinct gene mutations in the context of wild-type background or a loss-of-function zebrafish line. Accordingly, this task employs different strategies for the generation of transgenes, ranging from transient overexpression to stable transgenic lines. The first can be characterized in the embryonic state only, whereas the last can also be investigated in adulthood. Furthermore, transgene expression (wild type form of A) was introduced in gene A loss-of-function zebrafish to successfully demonstrate rescue of the phenotype. Hence, the transgenic approach is highly suitable to characterize the function of distinct genes and DCM gene mutations.
Further research in this WP included protein-protein interactions, in particular of contractile and regulatory proteins. Assembly, maintenance and renewal of sarcomeres require highly organized and balanced folding, transport, modification and degradation of sarcomeric proteins. However, the molecules that mediate these processes are largely unknown. We isolated the zebrafish mutant flatline (fla), which shows disturbed sarcomere assembly exclusively in heart and fast-twitch skeletal muscle. By positional cloning we identified a nonsense mutation within the SET- and MYND-domain-containing protein 1 gene (smyd1) to be responsible for the fla phenotype. We found SMYD1 expression to be restricted to the heart and fast-twitch skeletal muscle cells. Within these cell types, SMYD1 localizes to both the sarcomeric M-line, where it physically associates with myosin, and the nucleus, where it supposedly represses transcription through its SET and MYND domains. However, although we found transcript levels of thick filament chaperones, such as Hsp90a1 and UNC-45b, to be severely upregulated in fla, its histone methyltransferase activity – mainly responsible for the nuclear function of SMYD1 – is dispensable for sarcomerogenesis. Accordingly, sarcomere assembly in fla mutant embryos can be reconstituted by ectopically expressing histone methyltransferase-deficient SMYD1. By contrast, ectopic expression of myosin-binding-deficient SMYD1 does not rescue fla mutants, implicating an essential role for the SMYD1–myosin interaction in cardiac and fast-twitch skeletal muscle thick filament assembly (39).
Finally, taking advantage from the extensive research done in zebrafish, we isolated a mutation in zebrafish, bungee(bngjh177), which selectively perturbs valve formation in the embryonic heart by abrogating endocardial Notch signaling in cardiac cushions. We found by positional cloning that the bng phenotype is caused by a missense mutation (Y849N) in zebrafish protein kinase D2 (pkd2). The bng mutation selectively impairs PKD2 kinase activity and hence Histone deacetylase 5 phosphorylation, nuclear export, and inactivation. As a result, the expression of Histone deacetylase 5 target genes Krüppel-like factor 2a and 4a, transcription factors known to be pivotal for heart valve formation and to act upstream of Notch signaling, is severely downregulated in bungee (bng) mutant embryos. Accordingly, the expression of Notch target genes, such as Hey1, Hey2, and HeyL, is severely decreased in bng mutant embryos. Remarkably, downregulation of Histone deacetylase 5 activity in homozygous bng mutant embryos can rescue the mutant phenotype and reconstitutes notch1b expression in atrioventricular endocardial cells. We demonstrate for the first time that proper heart valve formation critically depends on Protein kinase D2-Histone deacetylase 5-Krüppel-like factor signalling (41) (Figure 6.2).
WP7 MICE MODEL
Lamins are proteins located next to the inner nuclear membrane of the cell as components of the nuclear lamina. Mutations in the LMNA gene are associated with several diseases including autosomal dominant Emery-Dreifuss muscular dystrophy and dilated cardiomyopathy. LMNA-mediated cardiomyopathy is often highly malignant; however the underlying mechanisms of disease are still unclear. Patients with a LMNA mutation have a serious prognosis due to high prevalence of heart failure and sudden cardiac death (3,4). Current treatment strategies to improve the prognosis are limited to implantable cardioverter-defibrillator and heart transplantation. However, no direct medical treatment has been described for these patients.
Mouse models of disease provide the opportunity to search for disease mechanisms and screen novel treatment strategies. Therefore, we will take advantage of mouse models of LMNA mediated disease, to enable pre-clinical screening of novel and existing strategies to treat these diseases. We developed a novel LMNA null mouse (LMNA(GT-/-) based on genetrap technology and analyzed its early post-natal development. We detect LMNA transcripts in the heart, the outflow tract, dorsal aorta, liver and somites during early embryonic development. Loss of A-type lamins results in severe growth retardation and developmental defects of the heart, including impaired myocyte hypertrophy, skeletal muscle hypotrophy, decreased amounts of subcutaneous adipose tissue and impaired ex vivo adipogenic differentiation. These defects cause death at 2 to 3 weeks post partum associated with muscle weakness and metabolic complications, but without the occurrence of dilated cardiomyopathy or an obvious progeroid phenotype (Figure 7.1). Our results indicate that defective early post-natal development critically contributes to the disease phenotypes in adult laminopathies (42).
Therefore, we planned to investigate both Lmna deficient and mouse models harbouring a previously characterized human LMNA DCM-causing mutation. Mice heterozygous for the Lmna mutant gene show a cardiac phenotype similar to that seen in human LMNA mutation–related dilated cardiomyopathy. Lmna heterozygous mice allow not only to investigate the pathophysiology of the myocardial damage and dysfunction but also to test the effects of existing drugs (e.g. beta-blockers, ACE inhibitors) to prevent or delay DCM and heart failure. Therefore, we investigated both existing and novel experimental models of cardiolaminopathy. Long-term survival was reduced in Lmna+/- untreated mice because of sudden death or heart failure necessitating euthanasia. Survival of medically treated groups was comparable to wild type untreated mice. In order to evaluate conduction disturbances that are inherent to the Lmna+/- phenotype and whether Metoprolol/Enalapril can prevent or delay these conduction disturbances, we performed ECGs at 50, 70 and 75 weeks of age. ECG data-analysis at longer intervals is ongoing. Furthermore, extensive molecular characterization (stress markers, fibrosis, apoptosis markers and oxidative stress) is ongoing to clarify the beneficial effect of Enalapril treatment in preventing Lmna+/- related DCM.
Existing model
After obtaining ethical approval to breed and use the Lmna A/C-null mice for transcriptomics, proteomics and metabolomics and testing existing drugs on these animals, we achieved the full Lmna knockout model and described the age-dependant phenotypes in homozygous and heterozygous mice. Heterozygous mice display a similar phenotype to laminopathies in humans. In addition, we collected hearts and blood plasma from control, heterozygous and homozygous Lmna-null mice for metabolomic studies. The mouse is a somatic Lamin A/C knockout, where exons 8 to part of 11 are replaced with the Pgkneo neomycin resistance cassette in reverse orientation to the Lmna gene. This full Lmna knockout model was generated in the lab of Colin Stewart (Sullivan et al 1999, Joural of Cell Biology). Homozygous LaminA/C-null mice display a delayed postnatal growth after two weeks. Muscular weakness and a dilated heart are apparent at 3-4 weeks, after which mice die at 8 weeks of age. This is accompanied by the known features of human loss of lamin A/C including failure to thrive, energy depletion and extremely shortened life span. Although the complete loss of lamin A/C does not reproduce human disease, as this is characterised by partial loss of lamin, or a dominant negative effect, heterozygous mice are a general reference model for human laminopathy. The heterozygous LaminA/C-null has only a cardiac phenotype encompassing conduction system defects at 10 weeks. After 40 weeks, the heart becomes dilated and less contractile. Overall, these mice display a similar phenotype as laminopathies in humans.
We upscaled the breeding of Lamin A/C-null mice to obtain enough mice for both “–omics” approach and for testing the effects of drugs. “-Omics” studies were done in hearts and blood plasma from control, heterozygous and homozygous Lmna-null mice at 2, 5, and 40 weeks of age (before and after the onset of dilated cardiomyopathy). Each experimental group consisted of 8 mice. Heart function and disease progression was monitored by echocardiography. Before sacrifice, we determined the body weight, heart weight and tibia length for each mouse. LMNA +/- mice develop a DCM phenotype similar to human DCM.
Novel Disease Models: H222P-Lmna knockin mouse model
We obtained ethical approval for the H222P-Lmna mice for –omics studies and to test the effect of existing drugs on progression of heart failure. We tried obtaining breeding couples of a knock-in mouse model that carries the H222P-Lmna missense mutation from the lab of Gisèle Bonne, which was previously described in human laminopathy (Arimura et al., Hum Mol Gen 2005). These mutant mice exhibit overtly normal embryonic development and sexual maturity. At adulthood, male homozygous mice display reduced locomotion activity with abnormal stiff walking posture and all of them die by 9 months of age. As for cardiac phenotype, they develop chamber dilation at 2 months and hypokinesia with conduction defects. Therefore, LmnaH222P/H222P mice represent an excellent model to study laminopathies as they develop a dystrophic condition of both skeletal and cardiac muscles similar to the human diseases. Current data show that ACE inhibitor Enalapril prevents development of DCM/HF in Lmna +/- mice, while Beta-blocker Metoprolol cannot prevent nor delay the progression to DCM in Lmna+/- mice. We are currently investigating the degree of fibrosis, apoptosis and oxidative stress in all experimental groups. In addition, ECG analysis is ongoing to determine whether Enalapril also prevents conduction disturbances. The results suggest that treatment of LMNA patients with Enalapril would be a very effective therapy in preventing the onset of DCM.
Survival of Lmna wildtype and Lmna +/- treated and untreated control mice.
To screen efficacy of existing treatment (e.g. β -blockers, ACE inhibitors) to prevent or delay DCM and heart failure, we designed an experiment in which we test Metoprolol (β-blocker) and Enalapril (ACE inhibitor) in Lmna mutant mice. Metoprolol is a widely prescribed drug to treat heart failure. It decreases heart rate, contractility and cardiac output, therefore decreasing blood pressure. Enalapril is an angiotensin converting enzyme (ACE) inhibitor, which lowers blood pressure by preventing constriction of blood vessels. This drug is also widely prescribed to treat heart failure. These drugs are often prescribed at a stage were heart failure is already apparent. In our study we want to examine the effects of these drugs before the onset of Lmna-related DCM specifically.
Experimental groups were as follows (n=20 mice per group):
• Wildtype: control group, no treatment
• Lmna+/-: control group, no treatment
• Lmna+/-: Metoprolol (30mg/kg/day)
• Lmna+/-: Enalapril (10mg/kg/day)
To determine whether early administration of β-blocking drugs prevents DCM, Metoprolol or Enalapril was administered in drinking water to male WT and Lmna+/- mice from 10 weeks of age. Heart function was monitored by echocardiography and ECG, after which mice were sacrificed at 75 weeks of age. Extensive morphology, histology and molecular characterization of hearts are currently ongoing. The experimental endpoint was in early 2013, when mice were sacrificed and hearts were harvested for morphological and molecular analysis. The Lmna+/- control and Metoprolol-treated mice developed dilated hearts. Interestingly, the Enalapril treated mice did not develop a dilated phenotype; however, the hearts were noticeably smaller than the Lmna+/- and Lmna WT control groups. When we compared the myocyte size of the Enalapril hearts with the Lmna+/- control and wildtype control, we observed smaller myocyte size in the Enalapril treated mice. Currently, heart function and progression of DCM has been monitored in all 4 experimental groups by means of echocardiography at 10, 40, 50, 65, 70 and 75 weeks of age. Left ventricular function gradually decreased in Lmna+/- mice as expected. Metoprolol treatment was unable to prevent or delay the decrease in left ventricular function of Lmna+/- mice. However, Lmna+/- mice treated with Enalapril were comparable to wildtype mice in LV function. This means Enalapril treatment is able to prevent the naturally occurring decrease in LV function of Lmna+/- mice. LV end-diastolic and end-systolic parameters were lower in the Enalapril treated mice, suggesting an enhanced LV contractility in this experimental group. Long-term survival was reduced in Lmna+/- untreated mice because of sudden death or heart failure necessitating euthanasia. Survival of medicine treated groups was comparable to wild type untreated mice. In order to evaluate conduction disturbances that are inherent to the Lmna+/- phenotype and whether Metoprolol/Enalapril can prevent or delay these conduction disturbances, we performed ECGs at 50, 70 and 75 weeks of age. However, ECG data-analysis is ongoing. Furthermore, extensive molecular characterization (stress markers, fibrosis, apoptosis markers and oxidative stress) is ongoing to clarify the beneficial effect of Enalapril treatment in preventing Lmna+/- related DCM.
Our current data show that ACE inhibitor Enalapril prevents development of DCM/HF in Lmna +/- mice, while Beta-blocker Metoprolol cannot prevent nor delay the progression to DCM in Lmna+/- mice. We are completing the study by investigating the degree of fibrosis, apoptosis and oxidative stress in all experimental groups. In addition, ECG analysis is ongoing to determine whether Enalapril also prevents conduction disturbances.
WP8 STRUCTURAL STUDIES
In order to identify and characterize the molecular effects of DCM-related mutations we undertook a structural biology approach, which provides atomic resolution models of selected mutant proteins, together with a biophysical characterization of their basic properties. X-ray crystallography allows to check the effects of mutations in terms of structure variability of the affected protein, providing a rational basis for the analysis of its overall stability. In the initial stage of the Project, our study was focused on the Lamin protein (LMNA), which hosts several of the currently known DCM mutations. LMNA is composed of a central α-helical “rod” domain, comprising four coiled-coil segments (Coil1A, Coil1B, Coil2A and Coil2B), non-α-helical “head” and “tail” domains. Part of our study was focused on Emerin (EMD), a protein which has been proven to interact directly with Lamin; moreover, mutations in the Emerin gene cause DCM, similarly to Lamin mutations. Emerin contains a LEM domain, a lamin-binding domain (EMDlbd) and a transmembrane domain.
Lamin We cloned, expressed and purified different constructs of the Lamin and Emerin genes and DCM-related mutants. Full length Lamin (LMNA, residues 1-660) was expressed and tentatively purified following conventional chromatographic methods used in protein biochemistry; however, the eluted protein was not pure and showed a marked tendency to aggregate and precipitate. Under these conditions, the crystallization of the protein was completely unsuccessful. The presence of disordered protein regions may affect the solubility/stability of the protein through aggregation. For this reason, a new construct (LMNAcentr) lacking the first 28 residues (preceding the Coil1A domain) and the last 115 residues (following the tail domain) was cloned during the second year. Also in this case we did not manage to obtain crystal hits. Size-exclusion chromatography experiments showed that LMNAcentr exists in a monomer/dimer equilibrium, while it is known that a unique aggregation status is fundamental to enhance crystallization. In vitro phosphorylation was proven to abolish linear assembly into polymers, but does not affect dimer formation. For this reason, a further p-likeLMNA construct was designed containing two mutations, S22E and S392E; both Ser residues are known to undergo phosphorylation. Indeed, mutations to glutamate mimic the phosphorylated status, due to their negative charge, and could help us isolate the homogeneous dimeric form of the protein for crystallization experiments. Expression and purification of the p-likeLMNA was successful, and the purified protein was used in wide crystallization screens; no crystals have been obtained so far. Various fragments of the protein, corresponding to different domains, were subcloned: Coil1A (residues 29-101), Coil2B (residues 303-387) and Tail (residues 432-544) domains. Moreover, mutants R335W, E347K, E342K, and E358K in the coil2B domain were expressed and purified.
Emerin: EMD is a serine rich protein, very difficult to express and purify since it contains a transmembrane domain, composed of a highly hydrophobic region, which leads the protein itself to aggregation. Moreover, EMD has been predicted to contain scattered disordered regions and low complexity regions, which compromise the protein stability. On a different approach, EMDlbd (Emerin lamin binding domain, residues 72-168) was expressed in the recombinant form and used for crystallization trials. The eluted protein was pure enough and stable in solution, but no crystals were obtained after thorough exploration of the crystallization conditions that led only to huge precipitation mass. Following these experiments, we believe that, in order to attain a defined conformational state, EMDlbd requires the presence of a binding partner. For this reason, we performed co-crystallization of EMDlbd with Tail domain, LMNA, LMNAcentr and p-likeLMNA, but no crystallization has occurred so far.
Experimental procedure
All the proteins were expressed in E. coli strain BL21(DE3) grown in Luria-Bertani medium. In the case of Coil1A, Coil2B and Coil2B mutants, the purification was performed under denaturing conditions. After induction, the cells were harvested and lysed by sonication. The supernatant was loaded on Nickel Sepharose Fast Flow resin (Ge-Healthcare). The fusion proteins were dialyzed overnight into the appropriate buffer for crystallization experiments.
Protein characterization
Recombinant Coil2B mutants and Coil1A were produced through a refolding procedure starting from denatured inclusion bodies; then, the proteins were characterized to verify their correct refolding after purification. CD analysis was performed in order to verify the correct folding of the purified protein. The CD spectra highlight a correctly folded protein composed of 87% α-helix secondary structure. This result is consistent with the Coil2B 3D-structure, which consists of a single α-helix. The remaining 13% of non α-helix is presumably referred to the His-Tag residues (6 residues) and the linker between HisTag and Coil2B added in the cloning procedure (26 residues). DLS measurements carried out at different Coil2B concentrations showed that the polydispersion (defined as the standard deviation of the hydrodynamic radius) of the protein samples was high at low concentration (1-5 mg/ml), while increasing concentrations decreased polydispersion. At 10 mg/ml Coil2B resulted monodisperse and thus suitable for crystallization experiments. HPLC-light scattering and analytical purification were performed to study the formation of the Coil1A/Coil2B dimer, however, no dimerization was detected so far.
Crystal screening
All the LMNA protein constructs were concentrated to 10 mg/ml using an Amicon Ultra centrifugal filter (Millipore) and used for vapour-diffusion sitting-drop trials assembled using an Oryx-8 crystallization robot (Douglas Instruments, East Garston, UK). All crystallization trials were performed at 20°C by mixing a 0.2 μl droplet of the protein solution with 0.3 μl of a reservoir solution. Crystals of Coil2Bwt, R335W, E347K, E342K and E358K mutants were obtained under several conditions. At the very beginning of the study, we observed that after two weeks at 20°C no crystals grew, but vitreous spheres were present under various conditions. To overcome the formation of undesired vitreous aggregation, the crystallization trials were first incubated at 20°C and successively at 37°C, at which temperature most of the spheres turned into crystals. In the case of the E347K mutant, numerous crystals, presenting different shapes and grown under different physico-chemical conditions in the initial screening, were obtained. In contrast, crystals of E342K and E358K grew under a single condition and, consequently, optimization of such growth condition was performed. The largest crystals were cryo-protected and flash-frozen in liquid nitrogen. X-ray diffraction tests were performed almost every month at the European Synchrotron Radiation Facility (Grenoble, France).
X-ray diffraction
We obtained diffraction data sets for Coil2Bwt, R335W and E347K. Crystals of Coil2Bwt and of the R335W mutant diffracted both to 3.2 Å, while crystals of E347K diffracted to 3.6 Å resolution (38). Despite the large number of crystallization and diffraction trials on E358K and E342K mutants, X-ray diffraction from all crystals was always extremely poor, possibly because these two mutants may represent a significant barrier for the correct formation of the coiled-coil dimer in the context of the quaternary assembly of the protein. Models of the E358K and E342K mutations were created, and a possible explanation on the effects linked to the mutations is proposed here below. The LMNA 1A-2B dimer was homology-modelled based on the corresponding domain of human vimentin A (PDB codes 1GK4, 1GK7). The E342K and E358K mutants were modelled on the assembled dimer (Figure 8.1). Residue 358 is exclusively acidic. The introduction of a positive charge in affected individuals was proposed to disrupt the overall acidic character (while altering the electrostatic charge) of this region of the helix. The model showed that the mutated residue lies at the interface between the two coiled helices, likely interfering with proper assembly of the dimer. In contrast, residue E342 lies on the surface of Coil 2B, on the opposite side with respect to Coil1A-2B dimer interface. Substitution with Lys, namely substitution with a positively charged residue, will not affect the formation of the dimer with Coil1A, as for the E358K mutation; however, it will possibly hinder the formation of tetrameric protofilaments, which assemble through the association of lamin coiled coil dimers in a head-to-tail fashion.
As observed for Coil2Bwt and R335W, crystals of the E347K mutant belong to the hexagonal space group P6522. The diffraction data were processed and scaled. The crystal structures were solved by molecular replacement using the known Coil2b structure as search model. A single protein molecule was located in the crystal asymmetric unit (Figure 8.2). The positioned model molecule was subjected to refinement. The amino acid sequence of the model was modified to match the mutant correct sequence and fitted to the electron density. The final model displays excellent overall stereochemical parameters, with 100% of the residues comprised within the allowed regions of the Ramachandran plot. Description of the structures is reported in “Significant results”.
Drug search
Since the structure of the full length LMNA and/or complex of the domains are not yet available, a search for a drug lead interacting with a single alpha-helix appears poorly meaningful. In fact, drug binding to a protein requires a binding cleft, or similar site, which would allow specific recognition and proper free energy of binding. Such sites are coded in the globular domains of proteins, or in protein/protein interface regions. The helical structures obtained during the project do not present such favourable potential druggable sites.
Gene product structures
The INHERITANCE project requires knowledge regarding the molecular bases of dysfunctions in genes linked to DCM (based on the work and discoveries of other Partners). In order to select crystallizable constructs of LMNA and Emerin, thorough information on all the domains was collected. In particular, it was discovered that in vitro phosphorylation of chicken lamin B by cdc2 Kinase abolishes linear assembly into polar head-to-tail polymers, but does not interfere with dimer formation. According to the PhosphoSite database, more than 30 phosphorylation sites in human A-type lamins have been reported. Phosphorylation of Thr19, Ser22 and Ser392 leads to depolymerization of lamin filaments during nuclear envelope breakdown in mitosis and meiosis. This information leads to the design of the p-likeLMNA. Furthermore, since it is very difficult to express and purify the EMD protein, due to its transmembrane domain that leads the protein itself to aggregation, basing on literature we designed the EMDlbd construct (Emerin lamin binding domain, residues 72-168), which was reported to bind the Tail domain of LMNA.
Significant results
We solved the crystal structure of Coil2Bwt, R335W and E347K mutant. Structures comprise amino acids 313 to 386 (sequence numbers refer to the full length LMNA). The Coil2Bwt structure consists of a single long α-helix. X-ray diffraction analysis revealed that the crystals present one polypeptide chain per asymmetric unit. However, the coil builds up a dimer through assembly with a symmetry equivalent molecule in the crystal lattice. The dimer is positioned on a crystallographic 2-fold axis, and exhibits a number of salt bridges, which stabilize the interactions between the two monomers. Superimposition of Coil2B and its mutants showed that the overall conformation of the molecules is strongly conserved, as expected considering the single residue mutations and crystal isomorphism. The side chains of Arg335 and Glu347 are shown to be exposed at the surface of the coiled coil dimer; nevertheless, the mutations occur at sites that are highly conserved in the known LMNA amino acid sequences. A mutation of Arg335 to Trp may directly interfere with the coiled-coil stability, either for steric hindrance or charge modification. A destabilization of such assembly, or the achievement of a looser coil, may be the cause of the pathogenic effect of the mutation R335W. The substitution of Glu347 with a positively charged residue, such as Lys, may interfere with the formation of putative intra- or inter-helical salt bridges, or other forms of polar interactions. Both mutations may alter the interactions of the coiled-coil dimer, and consequently interfere with the lamina assembly; alternatively, they may affect the binding of components within the nuclear lamina, including nuclear factors that have been proposed to interact/associate with LMNA. These results were included in the paper, which was published on Biochemical and Biophysical Research Communications (43).
WP9 - THERAPEUTICS AND IMPROVEMENT OF MEDICAL MANAGEMENT
In this WP we explored the possibility of:
1. Prevention or delay of the onset of the DCM phenotype in asymptomatic mutation carriers
2. Prevention or delay of DCM in animal models: the Myopalladin model
Prevention or delay of DCM phenotype in asymptomatic gene carriers (PSL, Paris)
We designed a Pilot Clinical Trial with angiotensin-converting enzyme (ACE) inhibitor in patients at a preclinical stage from families with Dilated Cardiomyopathy, entitled INHERITANCE “PRECARDIA” Study or PRE clinical mutation CARriers from families with DIlated cardiomyopathy and ACE inhibitors. The aim is to study the impact of ACE inhibitors (ACEi) on patients who carry a mutation but have not developed DCM yet through a double blind randomized multicentre trial. The hypothesis is that ACEi may delay or prevent the occurrence or delay the onset of DCM in these patients (pre-clinical stage), thus modifying the natural history of the disease (44).
The design of the study is as follows:
Inclusion criteria. Relatives with a DCM-causing mutation (from families with DCM) without obvious DCM but with minor left ventricular (LV) abnormalities: isolated slight LV enlargement or reduced systolic dysfunction (echocardiography).
Number of patients to enrol. N = 200.
Duration of the recruitment. Expected duration between 1 and 2 years.
Duration of the follow-up and treatment. Follow-up = 3 years.
Treatment. Perindopril 10 mg once a day versus placebo. Randomisation 1 / 1.
End points. Primary end-point: deterioration of LV end diastolic diameter/volume or Ejection fraction (echocardiographic or magnetic resonance imaging). Secondary end-points include the evolution of other echocardiographic parameters or hormonal biomarkers (including Mid-Regional pro-Adrenomedullin).
Main steps of the trial (summary) until the closure of the study:
- The study (clinical trial: NCT01583114, reference sponsor C10-44 Precardia, N° EudraCT 2010-023184-18) was formally approved by the sponsor (INSERM) in January 2011(44).
- The study was approved (August 2011) at the European level by the Clinical Trial Facilitation Group (CTFG) at the HMA (Heads of medicine Agencies - network of the Heads of the National Competent Authorities in the European Economic area) according to a unified VHP single process (VHP or Voluntary Harmonisation Procedure).
- The first participant was enrolled in December 2011.
- By November 2013, only 3 investigation centers were opened and only 12 participants were enrolled.
- Early December 2013: the enrolment was stopped (because of the end of the two–year period planned for enrollment), and an amendment was submitted by the PI to the sponsor.
- Decision by the sponsor (Inserm) to prematurely stop the trial during December 2013 with an official decision dated 28th January 2014. The decision was not related to scientific reasons, but only to logistic aspects of the trial. Progress did not meet the requirements for sustained feasibility of the study.
Enrollment of participants.
The enrollment was delayed due to the length of the administrative processes that delayed the opening of investigating centres. At the end of 2013, only three centres were opened and active: Paris (starting December 2011), Pavia (starting December 2012) and A Coruna (starting July 2013) (Annex 1).
Inclusions in 2011: 1 patient
07 001 RL 000 009 01/12/2011 France
Inclusions in 2012: 6 patients
07 002 PS 000 010 17/01/2012 France
07 003 GN 000 011 24/01/2012 France
07 004 B N 000 101 19/06/2012 France
07 005 J M 000 102 03/07/2012 France
01 001 BR 000 109 03/12/2012 Italy
01 002 GM 000 110 19/12/2012 Italy
Inclusions in 2013: 5 patients
01 003 MG 000 111 09/01/2013 Italy
01 004 MP 000 112 09/01/2013 Italy
01 005 VR 000 113 22/05/2013 Italy
01 006 PM 000 114 15/07/2013 Italy
07 006 PP 000 103 15/10/2013 France
TOTAL = Number of patients enrolled by 31/11/2013: 12 patients.
Decision of INSERM (sponsor) to stop the trial.
Decision. A letter written by Ms Sonia Gueguen (Pôle Recherche Clinique) in French, signed on the 11th December and addressed to the PI of the trial (Ph Charron), recommended to stop the trial. Then, there was a lot of exchanges and discussion between Inserm, the PI of PRECARDIA and the coordinator of INHERITANCE. In a second letter in English, Inserm (Clinical research department, Mrs Gueguen), confirmed the decision to stop the trial and indicated that the official closure of the trial was on the 28th January 2014. In the letters, the sponsor indicates that, unfortunately and despite the undeniable effort that had been made, there was no sufficient progress in the general advance of the project, especially because of the delay in centers opening / enrollment and also because of under-funding of the WP. Progress did not meet the requirements for sustained feasibility of the study. As a consequence Inserm decided to prematurely stop the trial. During further discussion, Inserm confirmed that the decision was NOT related to scientific reasons, but only to operational and feasibility aspects of the trial.
Prevention or delay of DCM in animal models: the Myopalladin model
We previously identified the Myopalladin gene as involved in monogenic DCM with the identification of several mutations causing DCM in independent DCM families (Duboscq-Bidot L et al. Cardiovasc Res, 2008). There are two main parts in the present project:
1) monitoring the cardiovascular phenotype of the knock-in mice to fully characterize the pathophysiological consequences of the expression of the P1112L mutation in heterozygous and homozygous mice.
2) once the disease and its time course are described, evaluating the phenotype reversion potential of siRNA, specifically targeting the mutant allele.
Mouse model production and phenotyping
Successful embryo delivery occurred in January 2011, resulting in the birth of the heterozygous mouse, F1 heterozygous mice and homozygous mice by November 2011. Echography was then performed in mice up to the age of 18 months. Results showed that from the age of three months, both heterozygous and homozygous mice showed clear signs of a dilated cardiomyopathy.
By month 9, we observed in homozygous mice that LV end systolic volume was increased as compared to wild-type (p < 0.001) and LV ejection fraction values was reduced (p < 0.0001). Of note, the DCM echography based phenotype was reliably observed at 18 month of age (Figure 9.1). Also interestingly, generally across the whole assessment period, heterozygote echocardiographic values are non-significantly different to those of homozygotes, which might be an indication of a dominant negative effect of the mutation, in line with what observed in human mutation carriers (Duboscq-Bidot, 2008).
Mice sacrificed for the six month characterization were weighed and then dissected. No gross morphological differences in the hearts from different genotypes were evident by eye. However, mouse hearts from both heterozygous and homozygous were heavier than wild-type hearts (p < 0.05 using unpaired two-tailed Student’s t-test for heterozygote hearts only) (Figure 9.2). Introducing the mutant allele which includes the P1112L mutation and the siRNA targeting sequence into a mouse strain could have generated unexpected alternative splice sites, resulting in potential abnormal myopalladin mRNA expression. Therefore we performed PCR on cDNA generated from RNA extracted from heterozygous and wild-type mice. Comparison of PCR products indicated that no aberrant splicing had occurred.
Results of shRNA expression in Myopalladin KI model
We validated the allele-specific target introduced into the recombinant vector utilizing an in vitro luciferase based assay. shRNA (specifically targeting the mutant mypn allele) were more stable in vivo and less susceptible to degradation than siRNA, inducing us to re-align our strategy to produce AAVshRNA constructs. We used the helper plasmid pDG9, resulting in an AAV2.9 vector, which was shown to demonstrate cardiotropism.
To measure whether specific shRNA knock-down of the mutant allele would have occurred, an allele-specific qPCR was designed and optimized (Figure 9.3) indicating that the allele-specific qPCR is capable of cleanly distinguishing between the two alleles. In addition, we used this qPCR to rule out that introducing the mutation and shRNA targeting sequence altered the tissue expression of myopalladin in the knock-in mice. We performed and analyzed the results of a large pilot AAV eGFP infection study in normal wild-type mice in order to define the optimal injection mode, dose and serotype best combination. We infected a total of 34 C57BL/6 mice at the age of six weeks with either AAV2.1 eGFP (serotype 1) or AAV2.9 eGFP (serotype 9) virus. Mice were sacrificed after four weeks of infection, and heart, leg muscles and liver samples were taken and processed for RNA extraction (for subsequent qPCR), protein extraction (for western blotting), and embedded in freezing medium for immunofluorescence.
Following the optimization of the AAV infection strategy using the AAV2.9-eGFP virus produced in-house, we generated the following AAV2.9 viruses: 1) AAV2.9-shALS9-dsRed; 2) AAV2.9-shLuc-dsRed; 3)AAV2.9-dsRed. The goal of the study is to counteract the development of the dilated cardiomyopathy caused by the mutant protein by the removal of the mutant allele via shRNA silencing. Prior to infecting the entire study cohort with our viruses, we tested the viruses in a second pilot study in order to determine their effectiveness. In order to ensure both adequate viral infection and shRNA production during the early stages of disease development, we injected small groups of young adult mice from the three genotypes at the age of 6 weeks (Figure 9.4) using the AAV2.9 eGFP.
Being the INHERITANCE funding period now terminated, we are continuing the project under an independent funding from ICAN foundation (ICAN, Paris, France). FP7 INHERITANCE funding will be acknowledged in any publication/communication arising from this work.
WP10 BIOINFORMATICS DATABASE (45)
The INHERITANCE project has required the recompilation and integration of all the generated data in a common database. To allow effective intra and inter group collaboration of both clinicians and basic researchers, the database had to be Web-based. The database includes multiple types of data, which have been recovered and integrated. No standard definitions existed for a large number of the variables that had to be recorded (as those provided by CARDS). Standardization of such variables is an immediate benefit of the Project. There were pre-existent data in several databases that had to be recovered and integrated in the common database.
The input of data had to be easy and user-friendly and the accuracy of the included data had to be warranted (validation tools and quality controls). Collaboration and data sharing are fundamental, but to ensure maximal collaboration the database management system had to warrant that each group has full control over the use of their own data. Ethical and legal issues about confidentiality, privacy and anonymous data had to be solved. Exploitation of the data and integration of tools for data query and analysis had to be considered in the design of the software for the database. Loss of the data had to be prevented at central and local levels (security copies of the database at different levels).
To this end, Health In Code and the INHERITANCE Consortium have developed a dedicated INHERITANCE database, and Health in Code has integrated the software with the Cardioregister database, an International/National Register of Familiar Cardiomyopathies. Cardioregister is a voluntary Database, web-based (internet access), designed to collect, exploit and download anonymous data of patients and families with cardiovascular diseases, offering the ability to produce customized reports with data. It is organized in families and it has quality controls established by the participating investigators. Cardioregister has been designed as an instrument useful not only for data collection but also to support decision-making and family screening. In the INHERITENCE project, Cardioregister has been integrated with the system i2b2, customized for the needs of the project by the University of Pavia (WP11). i2b2 is a data warehouse and data mining project that was designed to allow fast querying and interaction the data collected so far. I2b2 and Cardioregister are periodically synchronized. Together with the data collected in the INHERITANCE database and i2b2 (WP11), the Inheritance consortium had a large number of patients’ data already available for summary statistics and pooled data analysis. The joint exploitation of the newly collected data and the data available at the background of the project is a crucial factor for the achievements of the project goals.
Bioinformatics Database and e-CRF (45).
We generated and implemented a software and database supporting the interdisciplinary activities of the INHERITANCE project. After having achieved this aim in the first period of the project, in the second period we further developed the eCRF for the clinical trial PRECARDIA INHERITANCE (WP9), planned the release of the commercial product (eCRF) and performed the SWOT analysis for the evaluation of the strengths, opportunities and threats, as needed in a business venture. The software and the database were developed according to the consensus reached by partners during the first period of the research. The last version of the Inheritance Database is 1.1.3.17790 installed in September 2012. In addition, a PRECARDIA eCRF was released in the second part of the Project. A multidisciplinary team of computer engineers and cardiologists developed the software and the database following the design and the recommendations made by the INHERITANCE working group. The whole system was tested and verified by the team of Health in Code and submitted to Inheritance group that approved and validated it; the software was used during all the project duration. Verification by Health in Code ensures that the product was built according to the requirements and design specifications, and software validation carried out by INHERITANCE consortium ensures that the product actually meets their needs, and that the specifications were correct (see details below).
1- The Inheritance database (Fig. 10.1)
The introduction of clinical data is still open. The database already includes data from previous data sets and new patients and family from the different participant centres. The available version of the Database has been developed following the principles established in the project. It allows effective collaboration of the partners with the introduction of their data in a common base. The variables of the database have been defined according to either existing standards or the criteria of the consensus of the consortium. The system works with anonymous data and provides confidentiality and security for the stored data. Quality control systems have been implemented. The system also allows the exploitation of the data by the partners. Prevention tools to avoid loss of data have been implemented (security copies). The DB is structured in sections, extensively described in deliverables D10.1 D10.3 and first and second reports.
2- The Precardia eCRF (Fig. 10.2)
The PRECARDIA clinical trial is one of the most challenging and ambitious activity of the Inheritance Project (WP9). Within the limited funding of the INHERITANCE project, all the consortium members have to contribute to the implementation of all the requirements for the correct development of a clinical trial. Health in Code (Partner 6) as the leader of the Research Area 6 (Bioinformatics) contributes to WP9 with the development, maintenance and hosting of the specific software application dedicated to the electronic recording of the data generated by the clinical trial (eCRF).
The software has been designed and developed keeping recommendations and forms designed by the project consortium. A multidisciplinary team of computer engineers and cardiologists, following the design and the recommendations made by the INHERITANCE working group and guidelines to develop eCRF systems, have developed the software and the database. The whole system was tested and validated by the team of Health in Code and submitted to Inheritance group that approved it, and it was used during the project duration. The last modified version is 1.0.4689.26422 that includes all pending corrections and modifications, and was installed in November 2012.
Data represent a decisive component of the Inheritance project. For this reason, we pursued a strategy based on three directions within Inheritance: 1) First, we have developed and implemented database and management system software that supports the needs of the project and allows the start of the recovery and integration of clinical and genetic data. The data of the patients included by the partners of the Inheritance project are being inserted in this Database. The consortium has approved the whole system. 2) Second, we have developed a data warehouse and data mining tool (WP11) that is constantly synchronized with the INHERITANCE database to quickly analyse and summarize the available data. The consortium has approved the whole system. 3) Third, we have developed dedicated software for the electronic CRF that gives support to the PRECARDIA clinical trial. Data will be introduced in this specific database by the INHERITANCE partners and will be available for review and queries by the authorized researchers; granting the fulfilment of all the security, confidentiality and data protection legal and ethical requirements. The consortium has approved the whole system. The validation and verification of INHERITANCE database and PRECARDIA eCRF have been achieved.
Health in Code is now working to launch this new product (eCRF) on the market, and is designing a market study that collects information about:
• Environmental conditions
• Potentially target clients
• Principal competitors on the market
• Products that have been substituted.
• Client expectations about the software and its functionalities
• Zoning and sales volumes
• The advertising and promotional strategy best suited to reach target customers
Once these steps are completed, we will massively launch the product on the market; meanwhile it is being offered on a small scale to commercial customers and partners of Health in Code. A new line of business has opened for Heath in Code, due to the participation in the INHERITANCE project. This line is gathered to the software development and the new commercialization of an eCRF product developed within the PRECARDIA project. The eCRF is now a commercial product that can be adopted in cardiology clinical trials.
WP11 KNOWLEDGE MANAGEMENT AND DATA ANALYSIS
The main task of WP11 was addressed at providing the INHERITANCE project partners with decision support tools and knowledge discovery models. In particular, the database tools developed in WP10 could be empowered by a set of web-based functionalities including data analysis workflow management, intelligent querying of the phenotype data, text mining of the relevant literature, and case-based reasoning.
The data warehouse
We developed a data warehouse supporting fast phenotype data mining and exploration. An automated management system can retrieve both data on biologic existing knowledge and novel discoveries from the literature. The implementation of this system can support decision making clinically oriented genetic tests. The data warehouse is based on the open source software developed by the i2b2 project (https://www.i2b2.org). A customized version of the data warehouse has been populated through a set of automated queries that extract patients’ data from the INHERITANCE database (Cardioregister, see WP10). A web client (46) query interface allows ad-hoc queries to be created by research clinicians and returns aggregate numbers of patients that satisfy the queries, as well as anonymized specific patient’s record data. A set of tools for automatic biological knowledge retrieval tool and literature-based discovery was made available to the project partners.
The Literature Mining tool for the automated knowledge extraction from the scientific literature has been conceived to provide utilities for automated elaboration of textual content, so that it is possible to automatically extract and interpret the information available in literature (47). It is based on the GATE framework (http://gate.ac.uk) and implements an iterative version of the Swanson’s ABC model to find associations in the literature (48). It has been tested in the case of dilated cardiomyopathies and of hypertrophic cardiomyopathies.
We also implemented a system for supporting reasoning and decision-making, by intelligently analyzing the data on each single case, with the aim of helping diagnosis and prioritizing gene screening. Our system is able to provide, given a patient case, the most similar cases present in the project’s database; in this way it was possible to provide the clinician with the ability to reason on a subset of cases obtained through a human-like analogical reasoning. The peculiarity of the system is that it implements a novel method to take into account data similarity at both the data level (i.e. the values of the data) and the semantic level (i.e. the presence of similar, though not the same, symptoms).
Wiki-based system
Wiki-based systems are crucial for collaborative research and results dissemination. A first version of the wiki system has been made available to the project partners: http://www.labmedinfo.org:8123/mediawiki/index.php/Main_Page
The semantic wiki-based system has been purposely implemented for supporting the consortium members in sharing and disseminating data and knowledge. It consists of a collaborative system that is used to track project activities, share ideas and data, foster exchange of information between the investigators to support several activities of the INHERITANCE translational research project.
Scientific data that are considered within the wiki are: DCM related documents, genes and proteins. Although it is possible to add these types of data using the same manual procedure described for the organizational data, the INHERITANCE semantic wiki provides a tool to automatically acquire DCM related documents and to add to the system the genes and the proteins that are cited within these documents. The first step is to specify the file that contains (in plain text, PDF or MS Word format) the document to be acquired and analyzed; once the file has been uploaded to the wiki, the system starts the text analysis procedure in order to extract gene and protein names. When the text analysis is completed, the user is presented with a preview page (Figure 11.1) where the identified genes and proteins are highlighted. Now it is possible to specify the name of the DCM related document page and to add it to the wiki. Gene and protein pages, when created with the procedure described above, also exploit a literature retrieval tool that finds the most recent literature about the specific molecule and links it in the relative wiki page.
The interrogation tools directly available in the semantic wiki are:
• the built-in query tool of the Semantic MediaWiki,
• a summary page defined to synthetize all the project activities and participants information,
• the RelFinder browser, useful to look for relations between keywords inside the wiki and show the relations graph.
While the built-in query tool requires the user to learn a query language in order to perform an effective interrogation on the wiki pages, the latter two interrogation tools allow an easier interaction with the wiki user.
The RelFinder browser, directly accessible from the wiki’s home page, is a graphical interrogation tool that shows, in form of graphs, the information contained in the semantic wiki. Figure 11.2 for instance, shows the graph that links the DCM related documents with the genes extracted from them.
Application of the Unified Medical Language System (UMLS) based system for DCM
Finally we hypothesized that a Unified Medical Language System (UMLS) based system for Literature-Based Discovery in medicine can contribute to trance gene-phenotype association reported in the literature. Literature-Based Discovery (LBD) is a technique that can be used in translational research to connect the very sparse and huge information available in scientific publications in order to extract new knowledge. This paper presents an LBD system based on the open discovery paradigm exploiting NLP techniques and UMLS medical concepts mapping, to provide a set of tools useful to discover unknown relationships. The system has been evaluated on the problem of discovering new candidate genes potentially related to dilated cardiomyopathies (DCM), and can be used in any medical context to connect different type of concepts. The validation of the system involves reproducing the discovery of genes currently associated to DCM. Validation showed that the system is able to discover many gene-disease associations by using the literature available before their first publication in a scientific article (49,50).
REFERENCES
References 3,4 represent pertinent activities (scientific, dissemination) of partners of INHERITANCE in scientific societies (several authors are partners in INHERITANCE).
1. Serio A, Narula N, Kodama T, Favalli V, Arbustini E. Familial DCM: a clinically and genetically heterogeneous disease? Herz. 2012;37:822-829.
2. Meder B, Katus HA. Clinical and genetic aspects of hypertrophic and dilated cardiomyopathy. Internist. 2012;53:408-14, 417-8.
3. Rapezzi C, Arbustini E, Caforio AL, Charron P, Gimeno-Blanes J, Heliö T, Linhart A, Mogensen J, Pinto Y, Ristic A, Seggewiss H, Sinagra G, Tavazzi L, Elliott PM. Diagnostic work-up in cardiomyopathies: bridging the gap between clinical phenotypes and final diagnosis. A position statement from the ESC Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2013;34:1448-58.
4. Charron P, Arad M, Arbustini E, Basso C, Bilinska Z, Elliott P, Helio T, Keren A, McKenna WJ, Monserrat L, Pankuweit S, Perrot A, Rapezzi C, Ristic A, Seggewiss H, van Langen I, Tavazzi L; European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Genetic counselling and testing in cardiomyopathies: a position statement of the European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2010;31:2715-26.
5. van Rijsingen IA, Arbustini E, Elliott P, Mogensen J, Hermans-van Ast JF , Van der Kooi AJ , van Tintelen P, van den Berg M, Pilotto A, Pasotti M, Jenkins S , Rowland C, Aslam U , Wilde A, Perrot A, Pankuweit S, Zwinderman AH, Charron P, Pinto YM. Risk Factors for Malignant Ventricular Arrhythmias in Lamin A/C Mutation Carriers: a European Cohort Study. JACC 2012;59:493-500.
6. van Rijsingen IAW, Nannenberg EA, Arbustini E, Elliott PM, Mogensen J, Hermans-van Ast JF, van der Kooi AJ, van Tintelen JP, van den Berg MP, Grasso M, Serio A, Jenkins S, Rowland C, Richard P, Wilde AAM, Perrot A, Pankuweit S, Zwinderman AH, Charron P, Christiaans I, Pinto YM. Gender-specific differences in major cardiac events and mortality in lamin A/C mutation carriers. Eur J Heart Fail. 2013;15:376-384.
7. Quarta G, Syrris P, Ashworth M, Jenkins S, Zuborne Alapi K, Morgan J, Muir A, Pantazis A, McKenna WJ, Elliott PM. Mutations in the Lamin A/C gene mimic arrhythmogenic right ventricular cardiomyopathy. Eur Heart J. 2012;33:1128-36.
8. Garcia-Pavia P, Syrris P, Salas C, Evans A, Mirelis JG, Cobo-Marcos M, Vilches C, Bornstein B, Segovia J, Alonso-Pulpon L, Elliott PM. Desmosomal protein gene mutations in patients with idiopathic dilated cardiomyopathy undergoing cardiac transplantation: a clinicopathological study. Heart. 2011;97:1744-52.
9. Diegoli M, Grasso M, Favalli V, Serio A, Gambarin FI, Klersy C, Pasotti M, Agozzino E, Scelsi L, Ferlini A, Febo O, Piccolo G, Tavazzi L, Narula J, Arbustini E. Diagnostic work-up and risk stratification in x-linked dilated cardiomyopathies caused by dystrophin defects. J Am Coll Cardiol. 2011;58:925-34.
10. Disertori M, Quintarelli S, Mazzola S, Favalli V, Narula N, Arbustini E. The need to modify patient selection to improve the benefits of implantable cardioverter-defibrillator for primary prevention of sudden death in non-ischaemic dilated cardiomyopathy. Europace. 2013;15:1693-701.
11. Serio A, Narula N, Frontera A, Isabella Gambarin F, Arbustini E. Prevalence of J-point elevation in families with sudden arrhythmic death syndrome. J Am Coll Cardiol. 2012;59:1659-60.
12. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. The MOGE(S) classification for a phenotype-genotype nomenclature of cardiomyopathy: endorsed by the World Heart Federation. J Am Coll Cardiol. 2013;62:2046-72.
13. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. The MOGE(S) classification for a phenotype-genotype nomenclature of cardiomyopathy: endorsed by the World Heart Federation. G Heart. 2013; 8:355-382.
14. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. MOGE(S) nosology in low-to-middle-income countries. doi:10.1038/nrcardio.2013.219‑c1
15. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. Letter JACC; in press 2014
16. Zaragoza MV, Fass J, Diegoli M, Lin D, Arbustini E.Mitochondrial DNA variant discovery and evaluation in human Cardiomyopathies through next-generation sequencing. PLoS One. 2010;5(8):e12295.
17. Zaragoza MV, Brandon MC, Diegoli M, Arbustini E, Wallace DC. Mitochondrial cardiomyopathies: how to identify candidate pathogenic mutations by mitochondrial DNA sequencing, MITOMASTER and phylogeny. Eur J Hum Genet. 2011;19:200-7.
18. Haas J, Katus HA, Meder B. Next-generation sequencing entering the clinical arena. Mol Cell Probes. 2011;25:206-11.
19. Meder B, Haas J, Keller A, Heid C, Just S, Borries A, Boisguerin V, Scharfenberger-Schmeer M, Stähler P, Beier M, Weichenhan D, Strom TM, Pfeufer A, Korn B, Katus HA, Rottbauer W. Targeted next-generation sequencing for the molecular genetic diagnostics of cardiomyopathies. Circ Cardiovasc Genet. 2011;4:110-22.
20. Frese KS, Katus HA, Meder B. Next-Generation Sequencing: From Understanding Biology to Personalized Medicine. Biology 2013, 2, 378-398.
21. Jan Haas, Karen S. Frese, Barbara Peil, Wanda Kloos, Andreas Keller, Rouven Nietsch, Zhu Feng, Sabine Müller, Elham Kayvanpour, Britta Vogel, Farbod Sedaghat-Hamedani, Wei-Keat Lim, Xiaohong Zhao, Dmitriy Fradkin, Doreen Köhler, Simon Fischer, Jennifer Franke, Sabine Marquart,Ioana Barb, Ali Amr, Philipp Ehlermann, Derliz Mereles, Tanja Weis, Andreas Kremer, Vanessa King, Emil Wirsz, Richard Isnard, Michel Komajda, Alessandra Serio, Maurizia Grasso, Petros Syrris, Eleanor Wicks, Vincent Plagnol, Luis Lopes, Tenna Gadgaard, Hans Eiskjær, Mads Jørgensen, Diego Garcia-Giustiniani, Martin Ortiz-Genga, Maria G. Crespo-Leiro, Rondal H. Lekanne Dit Deprez, Imke Christiaans, Ingrid A. van Rijsingen, Arthur A. Wilde, Anders Waldenstrom, Martino Bolognesi, Riccardo Bellazzi, Stellan Mörner, Justo Lorenzo Bermejo, Lorenzo Monserrat, Eric Villard, Jens Mogensen, Yigal M. Pinto, Philippe Charron, Perry Elliott, Eloisa Arbustini, Hugo A. Katus, Benjamin Meder. Atlas of the Clinical Genetics of Human Dilated Cardiomyopathy Eur Heart J (under evaluation) 2014.
22. Arbustini E, Grasso M. Human "nuclear" mitochondrial cardiomyopathy a novel mouse model characterizes the disease. JACC Cardiovasc Imaging. 2011;4:11-5.
23. Disertori M, Quintarelli S, Grasso M, Pilotto A, Narula N, Favalli V, Canclini C, Diegoli M, Mazzola S, Marini M, Del Greco M, Bonmassari R, Masè M, Ravelli F, Specchia C, Arbustini E. Autosomal recessive atrial dilated cardiomyopathy with standstill evolution associated with mutation of Natriuretic Peptide Precursor A. Circulation Genetics 2012; 6:27-36,.
24. Disertori M, Mazzola S, Quintarelli S, Arbustini E. [New frontiers for an ancient disease: genetics of atrial fibrillation]. G Ital Cardiol. 2013 ;14:46-54.
25. Kodama T, Serio A, Disertori M, Bronzetti G, Diegoli M, Narula N, Grasso M, Mazzola S, Arbustini E. Autosomal recessive paediatric sick sinus syndrome associated with novel compound mutations in SCN5A. Int J Cardiol. 2013;167:3078-80.
26. Villard E, Perret C, Gary F, Proust C, Dilanian G, Hengstenberg C, Ruppert V, Arbustini E, Wichter T, Germain M, Dubourg O, Tavazzi L, Aumont MC, DeGroote P, Fauchier L, Trochu JN, Gibelin P, Aupetit JF, Stark K, Erdmann J, Hetzer R, Roberts AM, Barton PJ, Regitz-Zagrosek V; Cardiogenics Consortium, Aslam U, Duboscq-Bidot L, Meyborg M, Maisch B, Madeira H, Waldenström A, Galve E, Cleland JG, Dorent R, Roizes G, Zeller T, Blankenberg S, Goodall AH, Cook S, Tregouet DA, Tiret L, Isnard R, Komajda M, Charron P, Cambien F. A genome-wide association study identifies two loci associated with heart failure due to dilated cardiomyopathy. Eur Heart J. 2011;32:1065-76.
27. Meder B, Rühle F, Weis T, Homuth G, Keller A, Franke J, Peil B, Lorenzo Bermejo J, Frese K, Huge A, Witten A, Vogel B, Haas J, Völker U, Ernst F, Teumer A, Ehlermann P, Zugck C, Friedrichs F, Kroemer H, Dörr M, Hoffmann W, Maisch B, Pankuweit S, Ruppert V, Scheffold T, Kühl U, Schultheiss HP, Kreutz R, Ertl G, Angermann C, Charron P, Villard E, Gary F, Isnard R, Komajda M, Lutz M, Meitinger T, Sinner MF, Wichmann HE, Krawczak M, Ivandic B, Weichenhan D, Gelbrich G, El-Mokhtari NE, Schreiber S, Felix SB, Hasenfuß G, Pfeufer A, Hübner N, Kääb S, Arbustini E, Rottbauer W, Frey N, Stoll M, Katus HA. A genome-wide association study identifies 6p21 as novel risk locus for dilated cardiomyopathy.Eur Heart J. 2013 Jul 12. (PMID: 23853074]
28. Vogel B, Keller A, Frese KS, Leidinger P, Sedaghat-Hamedani F, Kayvanpour E, Kloos W, Backe C, Thanaraj A, Brefort T, Beier M, Hardt S, Meese E, Katus HA, Meder B. Multivariate miRNA signatures as biomarkers for non-ischaemic systolic heart failure. Eur Heart J. 2013;34:2812-22.
29. Narula N, Favalli V, Tarantino P, Grasso M, Pilotto A, Bellazzi R, Serio A, Gambarin FI, Charron P, Meder B, Pinto Y, Elliott PM, Mogensen J, Bolognesi M, Bollati M, Arbustini E. Quantitative expression of the mutated lamin a/c gene in patients with cardiolaminopathy. J Am Coll Cardiol. 2012;60:1916-20.
30. Thanaraj A, Meder B. Genetic testing is superior to biopsy of the myocardium in cardiomyopathy - yes. Dtsch Med Wochenschr. 2013;138:598
31. Narula N, Agozzino M, Gazzoli F, Concardi M, Pagani F, Favalli V, Kodama T, Mazzola A, D'Armini AM, Arbustini E. The pathologic basis of recovery. Heart Fail Clin. 2014;10(1S):S63-74.
32. de Haas HJ, Arbustini E, Fuster V, Kramer CM, Narula J. Molecular imaging of the cardiac extracellular matrix. Circ Res. 2014;114:903-15.
33. Rasmussen TB, Palmfeldt J, Nissen PH, Magnoni R, Dalager S, Jensen UB, Kim WY, Heickendorff L, Mølgaard H, Jensen HK, Baandrup UT, Bross P,Mogensen J. Mutated desmoglein-2 proteins are incorporated into desmosomes and exhibit dominant-negative effects in arrhythmogenic right ventricular cardiomyopathy. Hum Mutat, 2013;34:697-705
34. Rasmussen TB, Hansen J, Nissen PH, Palmfeldt J, Dalager S, Jensen UB, Kim WY, Heickendorff L, Mølgaard H, Jensen HK, Sørensen KE, Baandrup UT, Bross P, Mogensen J. Protein expression studies of desmoplakin mutations in cardiomyopathy patients reveal different molecular disease mechanisms. Clin Genetics. 2013;84:20-30
35. Haas J, Frese KS, Park YJ, Keller A, Vogel B, Lindroth AM, Weichenhan D, Franke J, Fischer S, Bauer A, Marquart S, Sedaghat-Hamedani F, Kayvanpour E,Köhler D, Wolf NM, Hassel S, Nietsch R, Wieland T, Ehlermann P, Schultz JH, Dösch A, Mereles D, Hardt S, Backs J, Hoheisel JD, Plass C, Katus HA, Meder B. Alterations in cardiac DNA methylation in human dilated cardiomyopathy. EMBO Mol Med. 2013;5:413-29.
36. Griffin JL, Atherton H, Shockcor J, Atzori L. Metabolomics as a tool for cardiac research. Nat Rev Cardiol. 2011;8:630-43.
37. Heather LC, Wang X, West JA, Griffin JL. A practical guide to metabolomic profiling as a discovery tool for human heart disease. J Mol Cell Cardiol. 2013;55:2-11.
38. Kloos W, Katus HA, Meder B. Genetic cardiomyopathies. Lessons learned from humans, mice, and zebrafish. Herz. 2012;37:612-7.
39. Just S, Meder B, Berger IM, Etard C, Trano N, Patzel E, Hassel D, Marquart S, Dahme T, Vogel B, Fishman MC, Katus HA, Strähle U, Rottbauer W. The myosin-interacting protein SMYD1 is essential for sarcomere organization. J Cell Sci. 2011;124:3127-36.
40. Meder B, Huttner IG, Sedaghat-Hamedani F, Just S, Dahme T, Frese KS, Vogel B, Köhler D, Kloos W, Rudloff J, Marquart S, Katus HA, Rottbauer W. PINCH proteins regulate cardiac contractility by modulating integrin-linked kinase-protein kinase B signaling. Mol Cell Biol. 2011;31:3424-35
41. Just S, Berger IM, Meder B, Backs J, Keller A, Marquart S, Frese K, Patzel E, Rauch GJ; Tübingen 2000 Screen Consortium, Katus HA, Rottbauer W. Protein kinase D2 controls cardiac valve formation in zebrafish by regulating histone deacetylase 5 activity. Circulation. 2011;124:324-34
42. Kubben N, Voncken JW, Konings G, van Weeghel M, van den Hoogenhof MM, Gijbels M, van Erk A, Schoonderwoerd K, van den Bosch B, Dahlmans V, Calis C, Houten SM, Misteli T, Pinto YM. Post-natal myogenic and adipogenic developmental: Defects and metabolic impairment upon loss of A-type lamins. Nucleus. 2011;2:195-207
43. Bollati M, Barbiroli A, Favalli V, Arbustini E, Charron P, Bolognesi M. Structures of the lamin A/C R335W and E347K mutants: Implications for dilated cardiolaminopathies. Biochem Biophys Res Commun. 2012;418:217-21
44. EUDRA CT 2011-001292-39 https://www.clinicaltrialsregister.eu/ctr-search/search?query=DCM
45. www.cardioregister.org/inheritance
46. i2b2inheritance.com
47. Nuzzo A, Mulas F, Gabetta M, Arbustini E, Zupan B, Larizza C, Bellazzi R. Text Mining approaches for automated literature knowledge extraction and representation. Stud Health Technol Inform. 2010;160:954-8.
48. Bellazzi R, Larizza C, Gabetta M, Milani G, Bucalo M, Mulas F, Nuzzo A, Favalli V, Arbustini E. Information technology solutions to support translational research on inherited cardiomyopathies. Stud Health Technol Inform. 2011;169:907-11.
49. Larizza C, Gabetta M, Milani G, Bucalo M, Mulas F, Nuzzo A, Favalli V, Arbustini E, Bellazzi R. Supporting Translational Research on Inherited Cardiomyopathies through Information Technology. Methods Inf Med. 2013;52:137-47.
50. Gabetta M, Larizza C, Bellazzi R. A Unified Medical Language System (UMLS) based system for Literature-Based Discovery in medicine. Stud Health Technol Inform. 2013;192:412-6.
Potential Impact:
INHERITANCE addressed the Topic HEALTH-2009-2.4.2-3: Translation of basic knowledge on inherited cardiomyopathies into clinical practice. The objective of INHERITANCE was to use advanced biotechnological tools and the pooled expertise of a consortium of expert clinicians and scientists to improve the health of European citizens with familial DCM by translating basic knowledge of disease pathophysiology into practical and effective treatments. The different components of the FP7 call indications “multidisciplinary approach bringing together genomics, proteomics, structural and functional studies with clinical investigation should lead to the development of new treatments” matched our clinical and scientific needs as the key elements of the project (Figure Potential Impact 1). Specifically:
• Clinical Phenotypes in probands and relatives
• Extended (cardiac and non-cardiac) clinical characterisation (“red flags”).
• OMICS Analyses
o Molecular diagnosis of disease causing mutations.
o Identification of disease-modifying and -predisposing genes and polymorphisms by Genome Wide-Association studies
o Quantification of mutated gene transcripts in paired samples of peripheral and heart tissues.
o Proteomics and metabolomics for the identification of novel specific tissue and circulating disease markers.
• Structural: Biophysical and X-ray structural studies of mutated proteins exploring the molecular bases of disease.
• Functional: Expression of clinically malignant disease-gene mutations in animal models.
• Therapeutics: Old drugs for novel use and new molecular targets for existing drugs, as well as disease-specific guidelines for treatments.
• The bioinformatics actions planned in the project are a central tool for integrating our results and data in novel disease-specific management algorithms.
The impact of INHERITANCE ranged from models of social-health innovation to the basic knowledge of the mechanisms of disease in DCM as well as biotechnological innovation for genetic testing that makes sustainable and reproducible the screening of multiple disease-genes and web-assisted dissemination and free of charge tools such as application for smartphones and tablets.
MODELS OF CARE
Customized family-centered model of care for heritable cardiovascular diseases: the family-centered model
In INHERITANCE we have established a patient/family centered model of care for familial diseases. The model replaces the current discipline-based model of care (Figure Potential Impact 2). In familial DCM more affected members show the disease at different stages. The family member that brings relatives to the attention is the proband who usually seeks medical attention because of symptoms. The simple genetic counselling provides first essential information about the family history. However, the key step for familial DCM is the clinical non-invasive family screening that may identify affected asymptomatic family members, or relatives in which the echocardiographic study demonstrates early morpho-functional changes not fulfilling the WHO criteria for the diagnosis of DCM. The addition of genetic testing and the identification of the mutation that causes the disease provide a further powerful tool for the diagnosis; identify carriers that are still clinically healthy; inform about the gene- and mutation-specific phenotype. Based on the evidence that more than 80 disease genes may cause a similar DCM phenotype and that each disease-gene may cause not only the DCM but also extra-cardiac traits such as hearing loss, myopathy, encephalo-myopathies, palpebral ptosis, gastrointestinal problems, renal disease, etc., the interdisciplinary evaluation is an obligate strategy that completes the clinical characterization of the disease, thus favouring a selective approach to genetic testing.
Customized Disease-Specific Innovation
In the family centered model of care, disease-specific diagnostic work-ups are designed and organized for offering selected appropriate evaluations of the DCM associated with defects of different genes. The multidisciplinary diagnostic work-up is planned by the referral center that organizes all specialist clinical evaluations, which may vary according to the phenotype presented by the proband; in this novel organization strategy, new professionals (diagnostic work-up assistants) have been educated to interact with the referral physicians (geneticists or cardiologists or both) that provide the plans for the evaluation of the probands and for cascade family screening. The clinical network with family doctors facilitates the family context and the disease setting. Information for probands and relatives is a key step: although written or web-supported information is useful and reaches a larger proportion of potential beneficiaries, the direct communication either in family counselling or in meetings organized for families provides a unique occasion for learning about emerging social-health needs that are related with the disease in the families.
Family care
Once clinical and genetic cascade family screening is completed, the picture of the family may demonstrate:
1. affected, symptomatic probands and relatives, fulfilling diagnostic criteria of cardiomyopathy
2. affected, asymptomatic relatives that were unaware of their disease before the family screening
3. family members that only show instrumental (echocardiographic or ECG-graphic) abnormalities suggestive for an early diagnosis
4. health carriers of the mutation that causes the disease in the family.
The above four groups of patients/individuals at risk need tailored monitoring programs that are planned according to pheno-genotype, family history, comorbidities and any other individual need that may require specific plans. The concept of the model is that of taking the responsibility of family health programs, thus limiting the discomfort and the costs for the family itself and supporting educational actions for the family members during the counselling that concludes each monitoring occasion.
Collective educational activities
Regular meetings for patients and families are a unique occasion for reciprocal exchange of information about the disease (how to live with a familial DCM), advancements in care strategies and translational research results. These meetings are usually organized with Charities and Associations of patients supporting the activities of the centres dedicated to familial cardiomyopathies.
Efforts to translate the basic science of genetic forms of DCM into clinical practice are now largely impacting on patients and families. Before INHERITANCE, the integration of clinical, genetic and other biological data into coherent disease specific paradigms was limited to sporadic meeting occasions, mostly dedicated to physicians, cardiologists in particular. After INHERITANCE, the extension of information to specialists other than cardiologists, as well as family doctors, highly contributed to increase the awareness of the DCM as familial disease and as a condition in which family screening, both clinical and genetic, is the crucial key strategy for family health programs.
Sustainable genetic testing: lower costs having expedite results
The project had the extraordinary merit of developed not only knowledge but also expertise on how to organize, implement and exploit genetic testing in each centre participating in the project. The strategy was that of centralizing the NGS of a large EU population and, in parallel, implementing local technologies and expertise and exploiting existing local resources, both human and technological. Before INHERITANCE, most centres were equipped for performing Sanger-based sequencing, a costly and time-consuming method. After INHERITANCE, all centres run their own NGS-based tests. At present, fast, low-cost sustainable tools for diagnosis and disease monitoring are available, and Sanger-based testing is limited to a confirmatory role for mutations identified by NGS.
SCIENTIFIC IMPACT
INHERITANCE generated a scientific legacy by creating a multidisciplinary transnational platform comprising disease-oriented clinical assessments and non-invasive imaging; genomic studies that identified causative mutations as well as predisposition and modifier genes; measurements of metabolites and cardioproteomes; information on structure of mutated proteins; functional assessments in animal models; bioinformatics tools that integrate the different scientific activities with clinical investigation in order to develop new treatment strategies. With 46 publications in peer reviewed journals (Impact factor calculated on ISI 2012=271.174) one PATENT (BIO-10917), the web site for Cardioregister (www.cardioregister.org/inheritance) the e-CRF for clinical trials, the Wiki-based system (http://www.labmedinfo.org:8123/mediawiki/index.php/Main_Page) and the new Unified Medical Language System (UMLS) based system for DCM (incorporated in the Wiki System) and the free of charge app for the novel classification of cardiomyopathies (http://moges.biomeris.com/moges.html) the project can be considered as example of good use of public funds for research throughout Europe.
The scientific results impacted on:
- The definition of the natural history of the different subtypes of DCM, based on family screening and re-screening studies.
- The elucidation of genotype-phenotype correlations with impact on diagnostic, prognostic and outcome prediction.
- The elucidation of pathophysiologic mechanisms of myocardial damage and dysfunction in disease-specific experimental models.
- The identification of novel disease genes: before INHERITANCE, the disease genes were about 40; after INHERITANCE, the list includes more than 80 genes, all included in the large-scale NGS study carried out in the project.
- The identification of genes that modify DCM evolution and severity.
- The development of transcriptomic tests on peripheral blood RNA that can guide genetic testing in the clinical setting.
- The characterisation of disease-specific cardioproteome and peripheral metabolome to identify markers for diagnosis and clinical monitoring.
- The crystallisation of key epitopes of mutated vs. wild-type proteins for future elucidation of the effects of mutations in the 3D structure of the mutated proteins and future disease-specific pharmacological research.
- Pre-clinical (animal models and in vitro testing on human monocytes) evaluation of novel treatment strategies, of existing drugs and novel molecular therapies, to generate the basis for future experimental validation and investment.
- The generation of a web-assisted database that also guides the clinical work–up of the different types of DCM.
- The generation of a bio-informatics Wiki-like tool initially for the consortium and thereafter for the scientific community.
IMPACT ON FAMILIES
The aim of INHERITANCE is to produce recommendations for the management of patients and families with DCM that can be applied in all clinical settings. Specific innovations will include:
- Promotion of routine family screening in the diagnostic work-up of patients with DCM.
- Development of phenotype guided genetic testing protocols.
- Development of treatment recommendations in pre-symptomatic mutation carriers with borderline abnormalities in left ventricular size and function.
- Generation of the evidence that current guidelines for the implantation of cardiac defibrillators in high-risk subgroups of DCM for the prevention of sudden death, such as cardiolaminopathies, must be revised.
- The generation, implementation and exploitation of simple bioinformatics tools that assist disease management.
- Establishment of the family-oriented model of care for DCM and other inherited cardiovascular diseases in European healthcare systems.
TECHNICAL IMPACT
INHERITANCE significantly contributed to the improvement of quality of health in Europe. High impact from the bio-informatics activities (developed by P6 and P10 in WP10 and 11 respectively) resulted in a user-friendly database and the knowledge management system. A commercially exploitable product from this work is a database that guides clinical and genetic diagnostic work-up (and the e-CRF, see scientific property). Free on the Web is the app that supports the novel classification/nosology system for cardiomyopathies (MOGE(S)). The app constitutes an innovative fast approach to the daily management of phenotype and genotype data.
SOCIO-ECONOMIC RELEVANCE
A population of about 40,000 European citizens is affected by DCM in the participating countries. The socio-economic consequences for patients and families can be profound, as many develop the disease at a young age. The costs to healthcare systems are correspondingly high as patients suffer many years of prolonged ill health and as a result consume valuable resources. Moreover, they are frequently too sick to contribute financially to their local communities and to society as a whole. Data generated by INHERITANCE first provided national health organisations with a novel model of care that can be used to develop family oriented services offering tailored treatments to patients with DCM and preventing the development of disease in relatives through the application of cascade clinical and genetic screening as well as early medical treatments.
HEALTH POLICY RELEVANCE
INHERITANCE contributed to validate a prototype for translational research impacting clinical care through the integration of conventional clinical tools with cutting edge genetic, structural, functional, metabolomic and proteomic technologies. The subject of the study, familial DCM, is not only an important healthcare problem in Europe as well as throughout the world countries, but is also a paradigm for other inherited diseases of similar or greater relevance to the health of European citizens. Fulfillment of the aims of INHERITANCE is providing policy makers with data that can be used to develop a healthcare model for inherited heart diseases to be applied to centres across the continent. The long-term result will be more efficient in targeting of resources and improved health for patients and families.
USE OF DISSEMINATION OF FOREGROUND
EU and National Scientific Societies
The collaboration implemented in INHERITANCE generated partnerships for future research and stimulated similar scientific programs in all European countries. The results of the project have been disseminated to the wider cardiologic community and European policy makers through national societies and the European Society of Cardiology. With more than 40,000 European cardiologists that are members of the ESC and participate in the ESC annual meeting, the scientific products of INHERITANCE have been disseminated. The project’s output was also disseminated at a national level via the annual meeting of the national societies of cardiology represented by the partners in INHERITANCE.
World Heart Federation
A significant achievement was the collaboration with the WHF and the endorsement of a novel classification/nosology description system for cardiomyopathies. While progressing with research in INHERITANCE we realised that genetic information was exponentially increasing and that cardiomyopathies were being described based on the name of the mutated gene/protein. A genetic terminology entered the scientific literature with diagnoses termed as desmosomalopathy, cytoskeletalopathy, sarcomyopathy, channelopathy, cardiodystrophinopathy, cardiolaminopathy, zaspopathy, myotilinopathy, dystrophinopathy, A-B crystallinopathy, desminopathy, caveolinopathy, calpainopathy, sarcoglycanopathy, dysferlinopathy, merosinopathy, emerinopathy etc. These terms do not describe the phenotype but rather specifically describe the cause (the disease gene). From a clinical point of view, cardiologists cannot diagnose a DCM as myotilinopathy or calpainopathy. In addition, the increasing evidence of overlapping phenotypes (DCM and ARVC, HCM and RCM, DCM evolving through DCM) and complex genotypes (> than one mutation) are contributing to elucidate the old concept of phenotype variability, incomplete penetrance. These concepts are now being explained with complex genetic make-up in members of the same families. To mantain the morpho-functional classification (HCM, RCM, DCM, ARVC) but also to inform about familial and non-familial cardiomyopathy and disease gene and mutation, and while waiting for the completion of the discovery of all disease genes, a novel nosology descriptor was generated. The acronym of this novel nosology is MOGE(S) that includes morpho-functional characteristics (M); organ involvement (O); genetic or familial inheritance patterns (G); and an explicit etiological annotation (E) with details of a genetic defect or underlying disease/cause along with functional status information (S) as per the American College of Cardiology/American Heart Association (ACC/AHA) (A to D) stage and New York Heart Association (NYHA) (I to IV) functional classes. The novel nosology is supported by a free web app at http://moges.biomeris.com/moges.html that can be easily used with smartphone and tablets. Proposed in December 2013, MOGES is already a widely circulated paper, object of comments, short reports on web sites (Figure Potential Impact 3). Modelled on the TNM system for cancer, MOGES is a flexible system that does not introduce novel cardiomyopathies but allows a diagnostic description including both morpho-functional phenotype (fundamental) and aetiology, both genetic and non-genetic.
Patients and families are also a key target in the INHERITANCE dissemination strategy. At present, P1 organises an annual meeting for patients and families dedicated to cardiomyopathies; P2 is vice-chairman of a major national cardiomyopathy charity and contributes to educational sessions for patients and carers across the UK. Similar activities will be implemented in all countries represented in the Consortium by other partners. P5 collaborates with a French charity that works for promoting knowledge on cardiomyopathies and supports patients in France.
Dissemination and exploitation of translational programs have been addressed to national health government institutions. In public health care systems, where citizens contribute to their health care, the upgrading of the national health programs takes years to introduce novel plans and dedicated investments. Inherited DCM has been around for a long time, a neglected disease in national health plans; for instance, in several European countries (including Italy), DCM is not even listed among rare diseases, even when genetically determined. Partners of INHERITANCE are promoting inherited DCM as a rare disease, to provide patients and families with the same rights of other groups of formally recognised rare diseases.
INHERITANCE DISSEMINATION MEETING (STRESA MAY 10,11, 2014 ITALY)
We have organised an international meeting specifically aimed at disseminating the results of INHERITANCE and promoting adhesion to INHERITANCE models of family-centred care in other countries where centres for inherited cardiomyopathies are just taking the first steps. In this meeting for the first time we will have a session dedicated to charities active in 3 countries (UK, FR, and IT). When looking for further similar charities in other countries we discovered that there are no similar initiatives in the other countries participating to the project. We will take advantage of this unique occasion for encouraging and helping these 3 charities to reciprocally interact, explore the needs in different EU countries, as well as the quality and the standards of care for cardiomyopathies. We are aware that fund restriction is now a major limitation for our research and that solidarity and collaboration with charities is a unique occasion for making health programs uniform though Europe. We will also ask charities to collaborate with researchers in writing a document highlighting the view points of patients and families in which 50% of the members are affected by ominous cardiac diseases in young age. A draft of the document is ready for the discussion in the meeting.
With this meeting, we also hope to gain more attention from national and regional health institutions with respect to cardiomyopathies, DCM in particular. The meeting in STRESA will be preceded by a masterclass on inherited cardiomyopathies, open to all professionals working in this setting, free of charge. The masterclass will be held in a meeting centre kindly provided by the Lombardia Region and will be endorsed by “Ordine dei Medici di Milano and Pavia”, MAGICA Onlus, Lombardia Region, Ministry of Health.
The meeting has been planned with the double aim of reporting and discussing the results of INHERITANCE but also opening a door on incoming scientific discoveries in molecular imaging and expansion of our collaborations with more centres and countries in Europe.
List of Websites:
The public website of INHERITANCE project is www.inheritanceproject.eu.
The Principal Investigator of the project is Eloisa Arbustini, MD, Center for Inherited Cardiovascular Diseases, IRCCS Foundation Policlinico San Matteo, Pavia ( Phone 0039 0382 501486, fax 0039 0382 501893, mail info.cardiomiopatie@smatteo.pv.it).
INHERITANCE is a small-scale focused research project that aimed at characterising the genetic basis of familial Dilated Cardiomyopathy (DCM) and translating basic knowledge of the aetiology into routine clinical practice throughout Europe. DCM is a life-threatening heart muscle disorder characterized by the presence of dilatation and systolic impairment of the left or both ventricles that is unexplained by abnormal loading conditions or coronary artery disease. The prevalence is 1 in 2,500 adults with an annual incidence of between 5 and 8 per 100,000. DCM is one of the leading causes of heart failure due to systolic dysfunction, being responsible for 10,000 deaths per year in Europe, and the commonest indication for cardiac transplantation in adolescents and adults.
Over the course of the project, INHERITANCE partners achieved genotype and phenotype data that has been used to identify important new information with immediate translation in the everyday clinical practice. The greatest contribution has been the characterisation of patients with laminopathies, a subtype of DCM caused by mutations in the gene encoding for the nuclear envelope protein, Lamin AC (LMNA) (6-8% of all DCM). The disease mechanism is haploinsufficiency with decreased expression of both gene and protein in affected hearts. INHERITANCE has shown that laminopathies are associated with a predictable phenotype that begins in early adulthood with conduction defects, and subsequent progression to heart block, ventricular arrhythmia and finally to heart failure. Male patients show a worse prognosis due to a higher prevalence of malignant ventricular arrhythmias and end-stage heart failure. Careful phenotyping has also allowed developing predictive models that identify patients who might benefit from early initiation of specific therapies such as anticoagulation and prophylactic ICD implantation. A DCM with right ventricular involvement caused by LMNA mutations identical to that seen in patients with ARVC has shed new light on the common biological pathways that cause heart muscle disease. Genes encoding for desmosome proteins have been shown to cause DCM. Together with new data on dilated cardio-dystrophinopathies (7% of all DCM in males) and less common genes (LDB3, desmosome genes, each accounting for less that 5% of all DCM) INHERITANCE also confirmed the role of clinical markers and of patho-functional assessment in the interpretation of genetic and clinical data.
The project generated the basis for a novel classification/nosology system for cardiomyopathies, the MOGE(S) system, that has been endorsed by the WHF and is supported by a free app at http://moges.biomeris.com/moges.html. As TNM for cancer, MOGE(S) is a flexible and expandable system that integrates phenotype and genotype information for cardiomyopathies.
Using Genome Wide Association Studies (GWAS) and animal models, we identified novel disease pathways, candidate genes and pathophysiologic mechanisms of myocardial damage; by linkage analysis, we discovered and characterised a novel autosomal recessive atrial DCM associated with mutations of NPPA. We further described novel autosomal recessive channelopathies presenting with both atrial and ventricular dilatation. The implementation of new high throughput gene sequencing (NGS) in large multicentre clinical series yielded to an extensive atlas of mutations and variants, that now require large-scale phenotyping tools and novel statistical strategies for the analysis of gene-phenotype correlations. This goal is now being achieved through systems that integrate large-scale phenotyping, bioinformatics and functional analyses. The NGS pipeline has proved to be reproducible and suitable for everyday practice. The large number of potentially pathogenic variants has revealed the complex genetic architecture of DCM that requires maintenance of research and collaboration between INHERITANCE partners as well as extension to other countries for EU-wide uniform translation.
Project Context and Objectives:
PROJECT CONTEXT
INHERITANCE is a translational multidisciplinary, multi-centre research project aimed at discovering novel genes and translate basic knowledge of the aetiology and pathophysiology of familial dilated cardiomyopathies (DCM) into routine clinical practice. The project involved 11 partners from 8 EU countries (IT, UK, SW, DE, FR, ES, DK, NDL). The consortium was generated taking into account the scientific and clinical expertise as well as the existing clinical series and facilities that were necessary to achieve the objectives of the project. The project stemmed from an uncovered clinical need that was the genetic characterisation of DCM and the establishment of diagnostic work-up for probands and relatives carriers of mutations of genes causing DCM.
INHERITANCE used existing state-of-the-art genotyping technologies in order to implement cost-effective protocols for rapid molecular diagnosis of specific forms of DCM. The project was structured in 5 research areas [clinical cardiogenetics, -omics (including genetic testing, transcriptomics, proteomics and metabolomics, animal models, structural and treatment), animal models, structural studies and bioinformatics] articulated in 11 WPs; two of the 11 WPs developed bioinformatics databases and systems for knowledge management and data analysis.
These genetic data have been used to integrate novel clinical diagnostic algorithms employing advanced non-invasive imaging modalities with information derived from original genomic, transcriptomic, metabolomic, proteomic and functional studies in blood and heart tissue samples obtained from patients with genetic forms of DCM. The findings have been used to develop translational programmes for preclinical diagnosis, new treatment strategies, family-tailored clinical work-up and cost-effective use of medical resources across the EU.
DEFINITIONS OF THE DISEASE
Familial dilated cardiomyopathy (FDCM) is clinically defined as idiopathic dilated cardiomyopathy (DCM) that is proven to occur in two or more related family members. DCM is defined as a myocardial disorder characterized by the presence of left ventricular dilatation and systolic impairment, in the absence of abnormal loading conditions (e.g. hypertension, valve disease) or coronary artery disease sufficient to cause global systolic dysfunction. Right ventricular dilatation and dysfunction may also be present. Therefore, current diagnosis of DCM is based on the clinical phenotype and does not distinguish the causes of the disease. Given that several genes may cause similar phenotypes all current strategies for treatments are phenotype-based.
EPIDEMIOLOGY OF FAMILIAL (GENETIC) DCM
The prevalence of DCM is approximately 1 in 2,500 adults, with an annual incidence of between 5 and 8 per 100,000. About 40,000 DCM cases are expected in the INHERITANCE European countries. In children, the incidence is much lower (0.5 to 0.8 per 100,000 per year), but DCM is the commonest cardiomyopathy in the paediatric population. DCM is responsible for 10,000 deaths per year in Europe and is the commonest indication for cardiac transplantation in adolescents and adults. In fact, DCM constitutes the most common cause of end-stage heart failure treated with heart transplantation worldwide. More than 50% of heart transplants have been performed to date in DCM patients. Up to 50% of individuals with DCM have at least a first-degree relative who is affected; a further 20% of family members have isolated left ventricular enlargement with preserved systolic function. Many of them subsequently develop overt DCM. In spite of this knowledge, the majority of patients were neither made aware of the familial risk nor offered genetic testing and less than 1% of patients in Europe benefitted from genetic testing. Genetic testing could add diagnostic value to the “early” instrumental markers in this 20% of relatives, thus contributing to a diagnostic anticipation and identification of individuals that will develop the full clinical phenotype. Early instrumental markers such as left ventricular dilatation without overt dysfunction or borderline left ventricular dysfunction with normal ventricular size would gain diagnostic value in mutation carriers.
Before INHERITANCE, family screening and genetic studies had identified more than 40 disease-causing genes. The commonest genes that cause DCM are those encoding proteins of the sarcolemma, cytoskeleton, sarcomere, nuclear envelope and energy generation. Most DCMs (up to 90%) are inherited as autosomal dominant traits, while a minority are X-linked recessive, autosomal recessive, and matrilineal.
Variation of age of onset, disease severity and prognosis amongst members of families carrying the same causative mutation suggest that modifier genes play a role in disease expression and response to pharmacological therapy. However, previously performed studies were small and showed discordant results on the effect of potential genetic modifiers. Only a large scale, genome-wide screening study in patients with known causative mutations, such as the one proposed in INHERITANCE, can examine this question.
PHENOTYPE CHARACTERISATION
The current clinical assessment of the different genetic disorders that cause DCM is relatively crude, with little emphasis given to the search for potential diagnostic cardiac and systemic markers. As a result, specific genetic diagnoses are often missed in everyday clinical practice and the implications for other family members overlooked. The symptoms and signs of DCM are highly variable and depend partly on the degree of left ventricular dysfunction and on the underlying aetiology. Presentation can be acute, often precipitated by concurrent illness or arrhythmia, or chronic, preceding the diagnosis by many months or years.
DCM is increasingly diagnosed as an incidental finding in asymptomatic individuals during routine examination or family screening. Associated cardiac traits (such as left ventricular non-compaction, conduction tissue disease, WPW, short PR) and non-cardiac traits such as hearing loss, retinitis pigmentosa, myopathy, palpebral ptosis, granulocytopenia, methylglutaconic aciduria, premature cataract, learning difficulties, can be present and are important in guiding physicians towards a possible aetiology. These traits are variably associated with the DCM phenotype and can contribute to characterising the phenotype in families with DCM caused by mutations of different genes. They constitute easy-to-recognise, low cost markers that should be identified in the clinical phase of the proband/family evaluation. The family screening, which is the strategy implemented in INHERITANCE, confirmed the hypothesis that clinical markers that are specifically recurring in DCM caused by different genes can be present in the early phases of the disease and disappear in late phases, thus further supporting the strategy of learning about the disease phenotype from more family members observed in different phases of the natural history of DCM.
The prognosis of DCM is highly variable and may also depend on aetiology. Early survival studies suggested mortality in symptomatic adults with idiopathic DCM approaching 25% at one year and 50% at 5 years. More recent reports have shown better outcomes, with five-year survival rates of approximately 20%, perhaps reflecting earlier disease recognition and treatment, and advances in medical therapy. However, recent studies have shown that the arrhythmogenic risk is extremely high in certain subtypes of DCM, such as cardio-laminopathies.
Role of biochemical markers in diagnosis and management
The role of biochemical markers in the diagnosis and monitoring of specific sub-types of familial DCM is still largely unexplored. Elevated levels of serum creatine kinase (sCPK) can be raised in dystrophin-related DCM in X-linked forms, or a LMNA gene defect in autosomal dominant DCM, but more disease-specific biomarkers are unknown. INHERITANCE aimed at searching for novel peripheral markers using proteomic and metabolomic analyses in paired myocardial samples and biological fluids (blood and urine) from genotyped patients and healthy carriers of causative mutations.
Influence of a genetic diagnosis on treatment
Nowadays, patients with DCM are treated in accordance with international guidelines for the management of heart failure with little consideration of the possible influence of the underlying aetiology on the response to treatment. Recent studies suggest that this might result in sub-optimal or inappropriate therapy in some patients. The most striking example is that patients with mutations in the nuclear envelope protein Lamin AC (LMNA) are at high risk of conduction disease relatively early in the course of their disease, but several studies have shown that implantation of a pacemaker alone does not prevent sudden death. Therefore, knowing that a DCM patient is carrier of a LMNA gene mutation might be of major importance when deciding on device therapy.
OBJECTIVES
The objectives planned in INHERITANCE were:
1. Clinical characterisation of DCM phenotypes associated with mutations of known disease genes. This objective specifically aimed at establishing priority clinical characterisation in clinically malignant cardiomyopathies and providing the prevalence of clinical markers that recur in DCM caused by mutations of same genes. The most common clinical markers associated with DCM are an ECG marker, the conduction disease (Atrio-Ventricular Block) and a biochemical marker (the increased Serum Creatine Phosphokinase).
2. Implementation of diagnostic gene testing in DCM population. Before INHERITANCE, less than 1% of patients and families in the EU received genetic advice and counselling. The Project aimed at increasing this very low percentage of patients throughout the EU, both within the partner centres’ and other countries, through the dissemination activities.
3. Development and validation of novel pre-genetic tests based on measurement of RNA transcripts in patients with pre-specified mutations from peripheral blood. This objective was particularly focused on DCM caused by mutations of the three major genes associated with DCM at the beginning of the project: LMNA, DYS and LDB3.
4. Identification of novel predisposing disease genes and modifiers in humans by GWAS and in animal models. This objective was planned because of the extreme phenotype variability in families with DCM caused by mutations of the same genes and even in the affected members of families with DCM caused by the same mutation.
5. Identification of disease-specific metabolites that can be measured in serum or urine. This objective was planned to explore possible metabolomics biomarkers in the different types of DCM.
6. Clarification of the pathophysiological mechanisms of myocardial damage and dysfunction in animal models and evaluation of old and novel treatments.
7. Structural characterisation and molecular simulation analyses on at least one mutated protein (LMNA). This objective specifically aimed at achieving structural information about the effects of recurrent mutations of LMNA on the protein.
8. Novel clinical approach with old drugs in patients with pre-clinical DCM. At the time of planning and revision of the new Document of Work, this objective was needed because the concept of early diagnosis could be strengthened by the identification of the mutation causing the disease in the probands’ relatives.
9. Generation of a web-assisted database and of a Wiki-based collaborative system. This objective was planned because of the large amount of data that were expected to be generated in the project and the need of an “intelligent” fast and easy tool for combining clinical markers with the genetic make-up of the different DCMs.
10. Generation of e-tools able to assist data collection and analysis for DCM, with potential translation to other types of cardiomyopathies.
11. Development of new diagnostic nosology for an easy description of key issues of cardiomyopathies caused by mutation of the different genes, considering the morpho-functional phenotype, the organs/tissues other than the heart eventually involved in the disease and the type of inheritance as well as the genetic defects.
While the achievement of each objective represented a major advance in the diagnosis and treatment of DCM patients INHERITANCE provided an integration of all the objectives into new international guidelines/classifications/nosology systems for describing the information obtained from clinical characterisation of the disease and genetic testing, thus facilitating the clinical work-up in probands and families and implementation in large family series.
Project Results:
WP1 - PHENOTYPES IN DCM AND GENO-PHENOTYPE CORRELATIONS
Idiopathic dilated cardiomyopathy (DCM) is a life-threatening heart muscle disorder characterised by the presence of dilatation and systolic impairment of the left or both ventricles that is unexplained by abnormal loading conditions or coronary artery disease. The commonest presentation is with symptoms and signs of heart failure, with a minority presenting with ventricular arrhythmia or sudden cardiac death. The diagnosis of DCM is still essentially descriptive and uses imaging techniques to detect the morphological and functional cardiac phenotypes. DCM affects approximately 1 in 2,500 adults with an annual incidence of between 5 and 8 per 100,000. It is one of the leading causes of heart failure due to systolic dysfunction, being responsible for 10,000 deaths per year in Europe, and the commonest indication for cardiac transplantation in adolescents and adults. Up to 50% of DCM is inherited as an autosomal dominant trait.
Before INHERITANCE, few patients were made aware of the familial risk, and less than 1% was offered genetic testing. More than 40 disease-causing genes were known, encoding proteins of the sarcolemma, nuclear envelope, cytoskeleton, sarcomere, desmosome and proteins important in energy generation. After INHERITANCE, the proportion of informed, counselled and genotyped families significantly increased in the partners’ centres; each partner contributed to national and international scientific dissemination, thus expanding the awareness that DCM is a familial disease in most cases and that clinical family screening is the first, low cost and most powerful tool to identify familial DCM. Family screening is now recognized as essential contributor to the family health, both prevention of major cardiac events and optimally tailored treatments(1,2). Two position papers of the WG on Myocardial and Pericardial Diseases of the European Society of Cardiology reflect the strategy of family screening and “red flags” as markers of disease and the collaborative research activities of the Partners of INHERITANCE(3,4).
Before INHERITANCE, genetic testing in everyday clinical practice was limited by the cost, time and complexity of conventional Sanger-based sequencing technologies. Advances in high throughput sequencing (HTS) technology are demonstrating the potential to solve this problem by analysing substantially larger genomic regions at a lower cost than conventional capillary Sanger sequencing. The INHERITANCE partners implemented next generation sequencing technology (NGS) both as centralized strategy and as local implementation for genetic testing, with the advantage that each centre is now progressing with its own NGS tool. However, NGS may also pose new challenges. The potential to identify a large number of rare variants means that large-scale phenotyping of relatives and families is required to understand genotype-phenotype relations. The emerging information from NGS analysis of more than 80 disease genes and family screening studies is the high number of patients that carry more than one mutation/variant in different genes. In spite of this apparent genetic complexity, the inheritance in DCM is autosomal dominant in most families, thus indicating one causal mutation and a variable number of “modifier” mutations. The emerging question is therefore how to establish the role (cause or modifier) of the identified mutations. The answer is emerging from cascade family screening, follow-up of relatives and functional investigations of the role of the mutant proteins in affected hearts, or in vitro or experimental systems.
Geno-phenotype correlations and nosology of cardiomyopathies.
Major achievements of the project include advances in geno-phenotype correlations and genetic-based risk stratification. Dilated cardiolaminopathies, which are associated with a very high risk of life-threatening arrhythmias (5,6), were one of the major targets of the research in the geno-phenotype correlation setting. LMNA codes a protein of the nuclear envelope that is essential for nuclear membrane integrity and function. Disruption of the protein due to gene mutations causes structural damage of the nuclear membrane. In the four years of the research, cardiac phenotypes associated with LMNA defects have been proven to be far more than DCM only. The association of DCM phenotype with conduction disease (CD) is the most common phenotype in the cardiology setting. Up to 80% of patients identified as carriers of LMNA mutation demonstrate CD (AVB). Conduction disease often precedes the onset of the cardiomyopathy. However, a minority of patients with cardiolaminopathy does not show CD; they can present with atrial fibrillation, ventricular arrhythmias, arrhythmogenic right ventricular cardiomyopathy (ARVC) or show prominent myopathy traits. Occasionally reported, atrial standstill, Left Ventricular Non-Compaction, sudden cardiac death without evidence of prior cardiomyopathy were not identified in INHERITANCE series (5).
During the project, a subgroup of patients with a usual phenotype of conduction disease and borderline or definite arrhythmogenic right ventricular cardiomyopathy was identified. In order to determine whether LMNA mutations are a common cause of this overlapping cardiac phenotype, a cohort of one hundred and eight patients from unrelated families with ARVC was genetically tested using conventional Sanger sequencing for five desmosomal genes (plakoglobin, desmoplakin, plakophillin-2, desmoglein-2, desmocollin-2) and LMNA. As expected, approximately 60% were positive for desmosomal gene mutations but (4%) without desmosomal gene mutations carried LMNA variants. These individuals had severe disease with conduction abnormalities on their resting 12 lead ECG and atrial fibrillation. This work shows that LMNA mutations can be found in severe forms of ARVC and that LMNA should be added to desmosomal genes when genetically testing patients with suspected ARVC, particularly when they also have ECG evidence for conduction disease or atrial arrhythmia. Therefore genes such as LMNA that typically cause DCM have been shown to cause ARVC too (7).
Dilated Cardio-desmosomalopathies. We explored the hypothesis that mutations in desmosomal genes account for a proportion of otherwise unexplained cases of DCM. 89 heart transplant recipients were selected for the study. All of them underwent genetic screening using conventional Sanger sequencing of five desmosomal genes (plakophillin-2, desmoplakin, desmocollin-2, desmoglein and plakoglobin), and the findings from genetic evaluation were correlated with the clinical phenotypes and histological characteristics in explanted hearts. All relatives of patients with pathogenic mutations or variants of unknown significance were offered clinical and genetic evaluation. The major finding was a prevalence of pathogenic mutations in 13% of patients; an additional percentage of patients (6%) had genetic variants of unknown significance. Unexpectedly, the histology of explanted hearts from patients with pathogenic mutations was identical to that of patients without mutations. Family screening showed evidence of co-segregation of mutations with DCM phenotype in five families. The implications of this study are that mutations in desmosomal genes are frequent in patients with advanced DCM undergoing cardiac transplantation and are a cause of familial disease (8). This result demonstrates that mutations in desmosome genes are not ARVC-specific but also cause DCM. The recently revised diagnostic criteria of ARVC (Marcus et al. Eur Heart J. 2010;31:806-14) introduced the DCM-like predominantly left and the biventricular variants (30%), thus reducing the difference between ARVC and DCM and leading to the novel definition of Arrhytmogenic Cardiomyopathy (ARC).
Dilated cardiodystrophinopathies demonstrated a phenotype characterised by severely dilated and dysfunctioning hearts with either increased sCPK or overt myopathy/dystrophy in more than 80% of cases in which DYS gene defects are identified (9). We identified DYS defects in 34 out of 436 patients (7.8%) (Onset age 34 ± 11 years, age range 17 to 54 years); 30 had proven X-linked inheritance. The 2 phenotypes included DCM with mild skeletal myopathy and/or increased serum creatine phosphokinase (n=28) or DCM only (n=6). The EMB showed defective dystrophin immunostain. The DYS defects consisted of 21 in-frame deletions and 11 out-of-frame deletions as well as 1 stop and 1 splice-site mutation. In the overall series of male patients with DCM in which DYS defects have been systematically searched for with MLPA and sequencing (n=542), we identified 42 (7.6%) probands carriers of large deletions in DMD gene (n=38) and point mutations (1 splice, 1 stop and 2 missense). The early data obtained in 2000 using simple multiplex PCR for DYS defects (Arbustini et.al. JACC 2000; 35:1760-8) gave a proportion of 6.5% mutations in consecutive series of males with DCM; by introducing sequencing of the gene in those cases that showed abnormal Dys immunostaining of the endomyocardial biopsies (EMB), the proportion increased to 7.8% (34/436)(9) and is stable (7.75%, 42/542) at the end of the project. This result confirms that most DYS mutations causing DCM are large deletions, but also that testing the gene only using MLPA would miss 10% of cardiodystrophinopathies.
INHERITANCE results also demonstrated a high proportion of DCMs caused by mutations of genes that recur in less than 5% of the cases. The morpho-functional classification describes the phenotypes (dilated, arrhythmogenic right, biventricular and predominantly left cardiomyopathy, hypertrophic and restrictive cardiomyopathies) but misses the description of the genetic basis of the different cardiomyopathies calling for a novel nosology system that can provide both morpho-functional and etiological information.
Prognosis
We demonstrated important differences between genetic subtypes of DCM. Dilated cardiolaminopathies are associated with a high risk of life-threatening ventricular arrhythmias even when the left ventricular dilatation and dysfunction are mild (5). In a median follow-up period of 43 months (interquartile range: 17 to 101 months), 48 of 269 LMNA mutation carriers (18%) experienced a first episode of Myocardial Ventricular Arrhythmia (MVA): 11 persons received successful cardiopulmonary resuscitation, 25 received appropriate ICD (Implantable Cardioverter Defibrillator) treatment, and 12 persons died suddenly. Independent risk factors for MVA were non-sustained ventricular tachycardia, left ventricular ejection fraction <45% at the first clinical contact, male gender, and non-missense mutations (ins-del/truncating or mutations affecting splicing). MVA occurred only in persons with at least 2 of these risk factors. There was a cumulative risk for MVA per additional risk factor (5) (Figure 1.1). In a further study we demonstrated that male patients show a higher prevalence of left ventricular impairment than females, but there were no gender differences in the frequency of traits such as atrioventricular block, atrial tachyarrhythmias and non-sustained ventricular tachycardia (6). The results of these studies are relevant for the formulation of novel guidelines for ICD implantation in patients with cardiolaminopathies; in fact, given that current guidelines recommend ICD implantation only in patients with severe left ventricular dysfunction. However, subgroups of patients with DCM caused by mutations of genes causing malignant cardiac phenotypes (i.e. Laminopathies) risk life-threatening ventricular arrhythmias even at Left Ventricle (LV) size and function of mild severity and in any case far from the criteria for primary ICD prevention in current guidelines. We recently proposed a novel algorithm for primary prevention with ICD in dilated cardiolaminopathies (Figure 1.2) (10). Instrumental markers, such as J-point elevation at ECG also have been considered as potentially informative for the risk of sudden death in cardiolaminopathies and in other types of DCM (11).
Conversely, dilated cardiodystrophinopathies are associated with a low risk of life threatening ventricular arrhythmias, even when the left ventricular dilatation and dysfunction were severe (9). As anticipated above, we identified DYS defects in 34 of 436 patients (7.8%). During a median follow-up of 60 months (interquartile range: 11.25 to 101.34 months) we observed 17 events, all related to heart failure (HF) (median event-free survival: 83.5 months). Eight patients (23%) underwent transplantation, and 9 (26%) died of HF while waiting for transplantation. Eight patients received an implantable cardioverter-defibrillator, although none had device intervention during a median follow-up of 14 months (interquartile range: 5 to 25 months). No patient died suddenly, suffered syncope, or developed life-threatening ventricular arrhythmias (Figure 1.3). Based on this study, we concluded that DYS-related DCM should be suspected in male patients with increased serum creatine phosphokinase (82%) and X-linked inheritance. The disease shows a high risk of end-stage HF but a lower risk of life-threatening arrhythmias.
The spectrum of disease genes is now higher than it was expected when INHERITANCE was planned. The sequencing of 40 disease genes by Sanger before INHERITANCE and of >80 disease genes by NGS in large series of probands after INHERITANCE demonstrates that the descriptive diagnosis of DCM does not provide sufficient information about each single DCM case/family. In addition, phenotype variability in families is influenced by the presence of more than one mutation in two or more genes. This observation implies that one of the two or three (or more) mutations plays a determinant role (causative mutation), while the second mutation like acts as modifiers of the phenotype.
Novel classification/nosology of cardiomyopathies.
The morpho-functional classification of cardiomyopathies is useful for current clinical management that is based on the phenotype but does not meet the recent evidence that more than 80 known disease genes may cause phenotypically similar but genetically different cardiomyopathies. The observation that genetic testing generates subgroups of DCM that can be classified on the basis of their cause rather than only on their morpho-functional phenotype led to the development of a novel classification and nosology system which describes the morpho-functional phenotype (M), the involvement of organ/tissues other than heart (O), the familial or genetic inheritance of the disease (G), the specific cause/etiology of the cardiomyopathy, genetic or non genetic, (E) and the optional stage/functional status of the cardiomyopathy (S) using both ACC-AHA stage A-D and NYHA class I-IV. The MOGES classification largely stemmed from INHERITANCE, and was endorsed by the World Heart Federation in 2013. Since genetic testing is now increasingly becoming a part of clinical work-up and, based on the genetic heterogeneity, numerous new names are being developed for the description of cardiomyopathies associated with mutations in different genes, a comprehensive nosology was needed that could inform about clinical phenotype, genotype as well as inheritance and involvement of organs other than the heart. In the quest for a genetic terminology, terms such as desmosomalopathy, cytoskeletalopathy, sarcomyopathy, channelopathy, cardiodystrophinopathy, or cardiolaminopathy, zaspopathy, myotilinopathy, dystrophinopathy, alpha-beta crystallinopathy, desminopathy, caveolinopathy, calpainopathy, sarcoglycanopathy, dysferlinopathy, merosinopathy, emerinopathy are now being used. Adding to the fact that such nosology would grow unmanageable, the genetic notation would neither define the phenotype nor the extent of systemic involvement. The MOGE(S) nosology does not aim at generating novel morphofunctional phenotypes. Analogously to the TNM system for cancer, MOGE(S) simply describes the clinical phenotype, the involvement of more organs/tissues, the familial or non-familial disease, the specific cause of the disease and, eventually, the clinical stage. MOGE(S) obligates clinicians to use a uniform language essential for future research, registries, surveys and international initiatives. The proposed nomenclature is supported by a web-assisted app (http://moges.biomeris.com) and reflects the diagnostic work-up of cardiomyopathy for evaluation of the manifest/asymptomatic disease, strategy for family screening and the result of genetic testing (Figure 1.4). MOGE(S) has been published contemporaneously in JACC and in Global Heart, the new official journal of the WHF (12,13), and is now matter of clinical and scientific interest; its implementation and exploitation has started (14,15). It is expected that the nomenclature would also help the development of collaborative registries and facilitate the description of complex genetics and grouping cardiomyopathies on their cause. Overall, results achieved in this WP constitute a major heritage of the project for future research developments.
WP2 GENETIC TESTING
The activities done in this WP aimed at implementing genetic testing using both Sanger-based technology and Next Generation Sequencing tools for the screening of known and novel disease and candidate genes, including analysis of DNA from probands that tested negative for all known genes. In parallel, the partners maintained the comparative evaluation of the efficacy of Sanger vs. NGS for mutation detection. We further considered that need to focusing on sequencing of rare disease genes that were not previously investigated and, finally, on discovering novel cardiomyopathies and related genetic causes.
Therefore, the genetic testing in INHERITANCE included:
• Sanger-Based sequencing of known disease genes
• NGS-based analysis of both mitochondrial DNA and nuclear DNA genes
• Linkage analysis
Consensus and quality
The INHERITANCE consortium first established consensus criteria for the definition of disease-causing genetic variants (pathologic), Genetic Variants of Uncertain Significance (GVUS), subpolymorphisms and polymorphisms to be adopted through research in all the consortium labs; the consensus criteria provided partners with a common tool for their clinical and genetic activity and reporting, as well as DB implementation. The relevance of this task is related to: 1) the evidence that some of the genes associated with familial DCM are hypervariable and show a high number of variations (i.e. PKP2) mostly private; 2) the unexpectedly high number of double compound mutations; 3) the provisional interpretation of novel missense genetic variants until their role in the pathogenesis of the disease is confirmed. Amongst the criteria for the definition of genetic variants as disease causing, the segregation of the mutations with the phenotype in families was maintained as fundamental and remains a major consensus criterion contributing to the definition of pathogenetic mutations.
The consortium also established an EU NetwOrk for quality Controls (ENOC) for Sanger-based genetic testing, starting from the gene that most commonly causes familial DCM, LMNA; the results in blindly circulated samples from 6 labs documented 100% concordant results for mutation detection. To date, the sensitivity and specificity of different methods for mutation detection was evaluated in ENOQ activities, with 100% concordance on pathologic mutations. We established a Consortium-restricted Web-assisted mutation DB INHMUTDB that includes all genetic variations identified in INHERITANCE.
Next Generation Sequencing (NGS) tools were used for investigating mitochondrial and nuclear genes.
Mitochondrial DNA genes.
Mutations in mitochondrial DNA (mtDNA) may cause maternally-inherited cardiomyopathy and heart failure. In homoplasmy, all mtDNA copies contain the mutation. In heteroplasmy, there is a mixture of normal and mutant copies of mtDNA. The clinical phenotype of an affected individual depends on the type of genetic defect and the ratios of mutant and normal mtDNA in affected tissues. We determined the sensitivity of next-generation sequencing compared to Sanger sequencing for mutation detection in patients with mitochondrial cardiomyopathy. We studied 18 patients with mitochondrial cardiomyopathy and two with suspected mitochondrial disease. We "shotgun" sequenced PCR-amplified mtDNA and multiplexed using a single run on Roche's 454 Genome Sequencer. By mapping to the reference sequence, we obtained 1,300x average coverage per case and identified high-confidence variants (Figure 2.1). By comparing these to >400 mtDNA substitution variants detected by Sanger, we found 98% concordance in variant detection. Simulation studies showed that >95% of the homoplasmic variants were detected at a minimum sequence coverage of 20x while heteroplasmic variants required >200x coverage. Several Sanger "misses" were detected by 454 sequencing. These included the novel heteroplasmic 7501T>C in tRNA serine 1 in a patient with sudden cardiac death. These results support a role of next-generation sequencing in the discovery of novel mtDNA variants with heteroplasmy below the level reliably detected with Sanger sequencing. This comparative strategy will assist in the identification of mtDNA mutations and key genetic determinants for cardiomyopathy and mitochondrial disease (16,17).
Nuclear genes.
We developed a Next Generation Sequencing (NGS) pipeline including nuclear genes pertinent with DCM. Our strategy was first to concentrate a large series of DNA of homogeneously phenotyped patients from the consortium(18-20) and then, in parallel, implementing local NGS tools for future autonomous translational activities (Fig. 2.2). In a first study, we established an array-based subgenomic enrichment followed by next-generation sequencing to detect mutations in patients with hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM). With this approach, we show that the genomic region of interest can be enriched by a mean factor of 2169 compared with the coverage of the whole genome, resulting in high sequence coverage of selected disease genes and allowing us to define the genetic pathogenesis of cardiomyopathies in a single sequencing run. We detected disease-causing mutations in 6 patients, out of which, 2 microdeletions, and 4 point mutations. Furthermore, we identified several novel nonsynonymous variants, which are predicted to be harmful, and hence, might be potential disease mutations or modifiers for DCM or HCM.
We further progressed with a NGS study in a large cohort of patients from the INHERITANCE consortium. The partners submitted 1041 DNA samples to the P3 centre as established. 719 samples have been sequenced to date as planned (Figure 2.2). 272 samples had been previously analysed by Sanger sequencing. Therefore a large amount of data for comparative evaluations of the Sanger-based vs. NGS-based sequencing has been made available. Comparative evaluation has been done considering both common SNPs and mutations. Genomic regions for NGS were selected on the basis of their relevance for DCM. Methods for NGS including enrichment of the target region was achieved by in-solution hybridization of fragmented genomic DNA against custom capture oligonucleotides. Enrichment was achieved by using the Agilent SureSelect System with multiplexing. The mean sequence coverage was approximately 8000-fold. Sequencing was performed on an Illumina HiSeq 2000 sequencing platform in a multiplex approach according to optimised protocols for target-enriched samples (Illumina; San Diego, CA). Demultiplexing of the raw sequencing data and generation of fastq files have been performed using CASAVA v.1.8.2. . Mapping of the sequencing reads against human genome hg19 was done with the burrows-whealer alignment tool (BWA) (Li, Bioinformatics, 2009). Variant calling and quality filtering of variants was performed with Genome-Analysis-Toolkit (GATK) as well as annotation of known variants present in dbSNP132 database (DePristo, Nature Genetics, 2011). Annotation of disease-causing variants, sequence coverage metrics and prediction of genomic function was done using an undisclosed prototype system including hardware and software. Sequencing libraries have been prepared along with the development of a standardized data analysis pipeline for the detection of known or novel variants to enable the correlation of genotype data with phenotype data of the investigated patients. Samples from 50-100 probands were selected for NGS analysis from each clinical centre.
A mean number of 4.9 megabases were generated per patient with a mean coverage over a target region of 2196 base pairs. An automated analysis, using galaxy-tools as a framework, was established to utilise different software algorithms for mapping, variant calling, filtering, annotation and interpretation of a functional effect on protein function of the found variants. In terms of prevalence, the highest number of variations was seen in the TTN gene that codes the giant sarcomeric protein Titin followed by LDB3 (CYPHER-ZASP) and PKP2 (Plakophillin 2). After filtering, the percentage of variations with “potential” pathogenic effects dramatically decreases. PKP2 and LDB3 variants remain highly represented. We should consider that both genes show high variability and LDB3 code alternatively spliced products. The alternative splice sites are seen by NGS as splice variants. A large number of variants identified by NGS is known and potentially relevant (HGMB Biobase). Each partner received results in order to compare existing Sanger-based data with NGS results and to complete the family studies for segregation evaluation. This is in fact a fundamental step because each proband was found to carry more than one variant demonstrating in silico prediction of a potential pathologic effect (20,21).
Rare or still candidate disease genes analysis
The results achieved in GWAS and zebrafish, as well as pathologic studies and mechanisms of diseases investigated in the project, provided novel candidate genes whose screening was implemented in the second period of the project. Novel genes, additional to those included in the NGS list, were proven as associated with DCM. We screened novel candidate or disease genes (i.e. ANT1 (22), ABLIM, ACADVL, ANO5, Corin, EMD. DOLK, FHL1, MLIP, MURC, NDUFAF2, NEBL, NKX2-5, POLG1, PYGM, TBX5) in consecutive series of DCM probands or in smaller series of probands with clinical or pathologic markers addressing to the test of these genes. Overall, we identified the causative mutation in 54 probands and families with DCM and variable phenotypic traits. This result further highlights the genetic heterogeneity of the DCM and suggests the need of a novel gene-based nosology for DCM. Functional in vitro studies and segregation analyses on families are ongoing with the aim of clarifying the effects of the mutations on the protein.
Discovery of a novel DCM: The Atrial Dilated Cardiomyopathy.
We identified a novel dilated cardiomyopathy only affecting atria in a geographic isolate of the north-East of Italy. We defined the disease as atrial DCM (ADCM). The disease was characterized by: 1) clinical onset in adulthood; 2) bi-atrial dilatation up to giant size; 3) early supraventricular arrhythmias with progressive loss of atrial electrical activity to atrial standstill; 4) thromboembolic complications; 5) stable, normal left ventricular function and NYHA functional class during the long-term course of the disease (Figure 2.3). By linkage analysis we mapped a locus at 1p36.22 containing the natriuretic precursor A (NPPA) gene. By sequencing NPPA we identified a homozygous missense mutation (p.Arg150Gln) in all living affected individuals of the 6 families. All patients showed low serum levels of Atrial Natriuretic Peptide (ANP). Heterozygous mutation carriers were healthy and demonstrated normal levels of ANP. We concluded that Autosomal recessive ADCM is a rare disease associated with homozygous mutation of the NPPA gene and characterized by extreme atrial dilatation with standstill evolution, thromboembolic risk, preserved left ventricular function and severely decreased levels of ANP (23).
The importance of this rare disease goes beyond the local benefits for the population and affected families and extends to primary atrial diseases, including the familial forms of lone AF (24). After the identification of autosomal recessive primary atrial cardiomyopathy ending in atrial standstill, we identified a further autosomal recessive atrial disease (Sick Sinus Syndrome, SSS) associated with novel compound mutations in SCN5A (25). The natural history of this SSS demonstrated an early LV dilatation: this opens novel avenues for research in overlapping ventricular and atrial cardiomyopathies.
WP 3 GENOME-WIDE ASSOCIATION STUDIES (GWAS)
The activities carried out in this WP aimed at identifying the genetic component of multifactorial/sporadic DCM and characterize the susceptibility to develop the disease or its variable phenotypic expression (disease progression) of either sporadic or familial DCM patients by performing Genome-Wide-Association studies (GWAS) in large populations of patients and matched controls. We therefore analyzed the population of patients and matched controls that have been already recruited (phenotypic data, blood samples and DNA extraction) through the network members (>2,000 sporadic DCM patients in Paris and Heidelberg, and similar number of controls). Homogeneous diagnostic criteria and geographic origin of patients will be considered, and controls have been matched for age, gender and geographic origin. Two separate initial analyses/ GWAS have been performed in parallel on different populations, one conducted in Paris, one in Heidelberg. Replication and combined analyses included both groups.
We conducted the first Genome Wide Association Study (GWAS) to identify new loci contributing to sporadic DCM. In the GWAS initiated in Paris, we identified 2 loci implicated in sporadic DCM (with respective P-values of 9.5x10-10 and 4.0x10-12 in the combined data set of 2365 DCM patients and 2410 controls, including “discovery” and “replication” populations). We also demonstrated that gene BAG3, involved in the second locus in sporadic DCM, is also involved in familial DCM as causal mutations in this gene were identified (Figure 3.1). The genes underlying the two loci, HSPB7 and BAG3, encode for non-structural proteins, which suggests new mechanisms underlying DCM-associated heart failure (26).
Subphenotype and variable expressivity analysis
We further studied these two loci and hypothesized that the loci could be involved in the severity of systolic HF (heart failure) and not only in the susceptibility to develop HF. We genotyped polymorphisms (SNPs) previously associated with DCM (rs2234962 for BAG3 locus, rs10927875 and rs945417 for HSPB7 locus) in a European population of 1160 patients with HF due to CAD (Coronary Artery Disease, ischemic-HF) as well as in a European cohort of 1141 patients with DCM. The severity of syst HF was assessed by left ventricle ejection fraction (LVEF), LV end diastolic diameter (LVEDD), age and NYHA class (dyspnea at inclusion). We observed that LVEDD was significantly associated with rs2234962 (BAG3 locus) both in DCM patients (591 patients with LVEDD available) and ischemic-HF patients (348 patients with LVEDD available) (p=0.0086 and 0.012 respectively). Other parameters reflecting severity of HF were not different according to this SNP or the others. Therefore severity of HF was not related to the two loci, except BAG3 locus that was associated with LV diameter in both populations. This gene could therefore appear as a prognostic factor in patients with HF.
Replication of the two loci in a population of Ischemic heart failure.
We further studied the two loci identified through the GWAS and hypothesized that the loci could also be involved in systolic heart failure due to coronary artery disease (ischemic-HF: systolic dysfunction LVEF<45%, significant coronary artery stenosis). We genotyped SNPs previously associated with DCM (rs2234962 for BAG3 locus, rs10927875 and rs945417 for HSPB7 locus) in a European population of 1160 patients with ischemic-HF and in 1612 controls. All patients and controls are of Caucasian origin, from France, Germany and UK. We observed that SNPs related to HSPB7 locus were significantly associated with ischemic-HF (MAF of rs10927875 and rs945417 were less frequent in patients than controls, adjusted p value 0.0017 and 0.02 respectively) whereas SNP related to BAG3 locus was not. Therefore, of the two loci previously associated with DCM, HSPB7 locus was also associated with ischemic-HF whereas BAG3 locus was not, suggesting differential involvement according to the underlying cause of HF.
Functional analyses of BAG3 gene mutations.
In order to unravel the molecular mechanisms associated with the expression of BAG3 mutations identified in familial DCM forms, we performed experiments in cells and animal models; rat neonate cardiomyocyte were infected with adenovirus expressing the 2 missense mutant of the BAG domain of BAG3. Transfected cells were rapidly degenerating with concomitant blockade of the autophagy pathway as measured through LC3-II marker down regulation. This was in accordance with the described role of BAG3 as a key regulator of autophagy as a mechanism controlling protein quality control and proteostasis. We subsequently confirmed that BAG3 mutant expressing cells were unable to rescue a luciferase sensor activity suggesting a default in protein refolding capacity. As HSP70 is central in proteostasis regulation and one of the known interactor of the BAG domain, we performed protein interaction assays between HSP70 and mutant BAG3 proteins. We observed a lack of interaction as compared to WT BAG3. These preliminary and unpublished results are indicative of BAG3 mutation mechanisms related to a lack of interaction with HSP70 leading to protein quality control dysfunction.
A new GWAS initiated in Paris.
We are finalizing the results of a new GWAS performed on the whole-exome nsSNPs based Illumina BeadChip. We genotyped more than 2000 cases and controls in the discovery step. Among the ~55.000 SNP with frequency over 0.5% in the exomeChip, at least 6 regions, including the two previously identified (BAG3 and HSPB7 loci), display significant association after Bonferroni correction. These preliminary results had to be replicated in independent population by individual TaqMan based genotyping in the 1000 cases and control population provided by Heidelberg partner. Specific statistical analysis of the rare polymorphisms is also to be performed (gene burden test) and would likely lead to the identification of more contributing loci. This ongoing work will strongly contribute to the discovery of new pathways leading to DCM and subsequently will feed further collaborative projects within the consortium.
A second large-scale case-control GWAS on sporadic DCM was initiated at Heidelberg in collaboration with national and international partners (27). We applied a three-staged case–control design. Stage 1 (screening phase) included 909 genome-wide genotyped individuals of European descent with DCM recruited between 2005–08 and 2120 controls from the PopGen and KORA population-based cohorts. In a first replication stage, SNPs on locus 6p21 were genotyped in 2597 DCM cases from Germany and Italy recruited between 2007 and 2011 as well as in 4867 controls from the population-based SHIP study (SHIP-0 and SHIP-TREND) and from Italy. In a second replication stage, the lead SNP was replicated in a cohort of 637 DCM cases and 723 healthy controls representing European Caucasians of French descent. Overall, we analysed more than 4100 DCM cases and 7600 controls. We identified and successfully replicated multiple single nucleotide polymorphism on chromosome 6p21 (Figure 3.2). In the combined analysis, the most significant association signal was obtained for rs9262636 (P = 4.90 × 10−9) located in HCG22, which could again be replicated in an independent cohort. Taking advantage of expression quantitative trait loci (eQTL) as molecular phenotypes, we identified rs9262636 as an eQTL for several closely located genes encoding class I and class II major histocompatibility complex heavy chain receptors (27).
A further genome-wide study aimed at investigating for the first time whole peripheral blood miRNAs as novel biomarker candidates for non-ischaemic heart failure with reduced ejection fraction (HF-REF). We assessed genome-wide miRNA expression profiles in 53 HF-REF patients and 39 controls. We could identify and validate several miRNAs that show altered expression levels in non-ischaemic HF-REF, discriminating cases from controls both as single markers or when combined in a multivariate signature. In addition, we demonstrated that the miRNAs of this signature significantly correlate with disease severity as indicated by left ventricular ejection fraction (Figure 3.3). Our data further denote that miRNAs are potential biomarkers for systolic heart failure. Since their detection levels in whole blood are also related to the degree of left ventricular dysfunction, they may serve as objective molecular tools to assess disease severity and prognosis (28).
WP4 TRANSCRIPTOMICS
The activities planned in this WP aimed at testing the hypothesis that quantitative gene expression (QGE) in RNA from affected myocardium and parallel blood samples taken from patients with DCM and known disease mutation can reflect or show correlations with the genetic defects. In addition, we tested the hypothesis that QGE and protein expression are correlated and that a lower gene expression in the affected myocardium correlates with lower or abnormal protein expression. We explored QGE of genes whose mutations have been proven as causally linked with DCM. We further investigated the expression of mutated proteins in the affected myocardium, both endomyocardial biopsies and hearts excised at transplantation.
Quantitative gene expression (QGE)
• Our first major target gene was LMNA, which is the most common and malignant disease gene in DCM (Figure 4.1 design of the study). We found that the QGELMNA is significantly lower in the RNA from myocardium and peripheral blood of mutation carriers than wild type controls. Using the comparative ΔΔCT method, we evaluated the QGE of LMNA (QGE(LMNA)) in peripheral blood and myocardial RNA from carriers of LMNA mutations, versus blood and myocardial samples from DCM(LMNAWT) patients and CTRL(LMNAWT) individuals. After generating reference values in normal controls, QGE(LMNA) was performed in 311 consecutive patients and relatives, blind to genotype, to assess the predictive value of QGE(LMNA) for the identification of mutation carriers. In parallel, Lamin A/C was investigated in myocardial samples from DCM(LMNAMut) versus DCM(LMNAWT) versus normal hearts (CTRL(LMNAWT)). LMNA was significantly underexpressed in mRNA from peripheral blood and myocardium of DCM(LMNAMut) patients versus DCM(LMNAWT) and CTRL(LMNAWT). In 311 individuals, blind to genotype, the QGE(LMNA) showed 100% sensitivity and 87% specificity as a predictor of LMNA mutations. The receiver-operating characteristic curve analysis yielded an area under the curve of 0.957 (p < 0.001). Loss of protein in cardiomyocytes' nuclei was documented in DCM(LMNAMut) patients. The reduced expression of LMNA gene in blood is a novel potential predictive biomarker for dilated cardiolaminopathies with parallel loss of protein expression in cardiomyocyte nuclei (29) (Figure 4.2).
• The second investigated gene was LDB3 (CYPHER ZASP). The QGELDB3 assay showed that this gene is not expressed in the RNA (at levels detectable by real time PCR) from peripheral blood samples. Vice versa the gene is expressed in the myocardium where we found that the mean 2^(-DCt) in LDB3 mutation carriers was 186,6374072 vs. 237,4355256 in wild type samples (p = ns). Therefore, we did not find significant differences in the QGELDB3 of mutated vs. wild type DCM patients. We conclude that mutated LDB3 gene is not underexpressed in the myocardial tissue of patients with dilated cardiozaspopathies.
• The third gene investigated in this task was DYS. We did not investigate QGEDYS in RNA from peripheral blood samples, while we extensively measured the QGEDYS in myocardial samples from patients with DYS-related DCM. We found a statistically significant underexpression of QGEDYS in the myocardium of patients that carry mutations of the Dystrophin gene vs. wild type controls. The mean 2^(-DCt) in DYS mutation carriers was 2,178374496 vs. 4,240265316 in the myocardium of DCM patients with wild type DYS. Therefore mutated DYS gene is underexpressed in the myocardial tissue of patients with dilated cardiodystrophinopathies and the level of underexpression is independent on the mutation type.
Expression of mutated proteins in affected hearts
Although genetic testing seems to be superior to biopsy of the myocardium in cardiomyopathy for the diagnosis(30), the interpretation of the role of the genetic mutations needs the myocardial investigation. Therefore we progressed with the investigation of the protein expression at the myocardial levels, in both mutated and wild type control samples (Figure 4.3).
• Patients with cardiolaminopathies demonstrated loss of expression of the Lamin AC protein in the membrane of the myocyte nuclei; vice versa, nuclei of non-myocyte myocardial cells preserved the expression of the protein. Haploinsufficiency is probably the mechanism by which mutations in the LMNA gene damage the nuclear integrity and function. We further progressed with the investigation of the Lamin AC expression in the myocytes of the atrio-ventricular node and found a far more severe loss of protein expression than that seen in ordinary myocardial cells. Western blot studies confirmed the immunohistochemical results, with decreased expression of the protein, irrespective of the mutation types (29).
• In patients with LDB3 mutations we observed variable results ranging from a homogeneous mild attenuation of to normal expression of the protein in affected hearts. This finding could not be considered as specific and in any case was not sufficient for suspecting a LBD3 mutation. Same results were obtained in hearts of patients carrying different types of mutations. The observation that myocardial QGELDB3 and protein expression do not significantly differ in mutated and wild type hearts adds to the clinical evidence of little segregation of mutations with the phenotypes (recorded in geno-phenotype correlation research in WP1 and WP2) and raises serious doubts about the unique causative role of LDB3 mutations in familial DCM.
• Vice versa, the immunostain of heart samples or endomyocardial biopsies with anti-Dystrophin antibodies demonstrated a decreased expression of the protein at the sarcolemmal membrane level. We observed that the extent and pattern of distribution of DYS immunostain may vary, but when control myocardium from patients/individuals with wild type dystrophin is compared, the difference of immunostain is strongly significant. The results are easily reproducible and DYS immunostain is now a standardised test for endomyocardial biopsies. These results have been recently confirmed in 7 additional affected hearts excised at transplantation. Therefore, myocardial samples from hemyzygous mutated males carrying DYS defects can be easily screened with immunohistochemistry, using commercially available antibodies (9).
• In patients with DCM associated with mutation of TNNI3, immunostaining with anti-troponin I3 antibodies showed mild attenuation of the intensity of the immunostain but this finding was not considered as sufficiently specific for suspecting a genetic defect in TNNI3.
• To date, the immunohistochemical studies of myocardium from DCM associated with “sarcomeric” defects did not provide diagnostic contribution.
The tissue studies done in the project provided the methods and facilities for expanding the research activities from clinically stable phases of the disease to end stage heart failure in patients with genetic DCM that were treated with LVAD as bridge to transplantation (31) and for exploring the tissue markers that could be target of molecular imaging in the near future (32). In fact, new research on molecular imaging is now stemming from the methods and results achieved in INHERITANCE.
Desmosomal protein expression in hearts of patients with mutations.
We showed that genes causing ARVC may also cause DCM and vice versa. The protein and gene expression pattern are therefore relevant to understand the mechanisms of disease related to the causative gene. Therefore we investigated how DSG2 and DPS mutations contribute to the pathogenesis of ARVC. Initial mutation analysis of DSG2 in 71 probands identified the first family reported with recessively inherited ARVC due to a missense mutation. In addition, three recognized DSG2 mutations were identified in 12 families. These results and further mutation analyses of four additional desmosomal genes indicated that ARVC caused by DSG2 mutations is often transmitted by recessive or digenic inheritance. Because desmosomal proteins are also expressed in skin tissue, keratinocytes served as a cell model to investigate DSG2 protein expression by Western blotting, 2D-PAGE, and liquid chromatography-mass spectrometry. The results showed that heterozygous mutation carriers expressed both mutated and wild-type DSG2 proteins. These findings were consistent with the results obtained by immunohistochemistry of endomyocardial biopsies and epidermal tissue of mutation carriers, which indicated a normal cellular distribution of DSG2. The results suggested a dominant-negative effect of the mutated DSG2 proteins because they were incorporated into the desmosomes (33).
Since desmoplakin is part of all desmosomes, which are abundantly expressed in both myocardial and epidermal tissue and serve as intercellular mechanical junctions, we investigated protein expression in myocardial and epidermal tissue of ARVC and CS patients carrying DSP mutations in order to elucidate potential molecular disease mechanisms. Genetic investigations identified three ARVC patients carrying different heterozygous DSP mutations in addition to a homozygous DSP mutation in a CS patient. The protein expression of DSP in mutation carriers was evaluated in biopsies from myocardial and epidermal tissue by immunohistochemistry. Keratinocyte cultures were established from skin biopsies of mutation carriers and characterized by reverse transcriptase polymerase chain reaction, western blotting, and protein mass spectrometry. The results showed that the mutation carriers had abnormal DSP expression in both myocardial and epidermal tissue. The investigations revealed that the disease mechanisms varied accordingly to the specific types of DSP mutation identified and included haploinsufficiency, dominant-negative effects, or a combination hereof. Furthermore, the results suggest that the keratinocytes cultured from patients are a valuable and easily accessible resource to elucidate the effects of desmosomal gene mutations in humans (34).
Epigenetic factors
Since DCMs show remarkable variability in their age of onset, phenotypic presentation, and clinical course, and the disease mechanisms that modify the occurrence and progression of the disease are both genetic and epigenetic factors that may interact with environmental stimuli, we examined genome-wide cardiac DNA methylation in patients with idiopathic DCM and controls. We detected methylation differences in pathways related to heart disease, but also in genes with still unknown function in DCM or heart failure, namely Lymphocyte antigen 75 (LY75), Tyrosine kinase-type cell surface receptor HER3 (ERBB3), Homeobox B13 (HOXB13) and Adenosine receptor A2A (ADORA2A). Mass-spectrometric analysis and bisulphite-sequencing enabled to confirm the observed DNA methylation changes in independent cohorts. Aberrant DNA methylation in DCM patients was associated with significant changes in LY75 and ADORA2A mRNA expression, but not in ERBB3 and HOXB13. In vivo studies of orthologous ly75 and adora2a in zebra-fish demonstrate a functional role of these genes in adaptive or maladaptive pathways in heart failure (35).
Overall, gene and protein expression in DCM hearts from patients that carry mutations in different genes can be considered easily feasible tools for investigating the mechanisms of disease. This is especially useful considering that the NGS tools are demonstrating that a large proportion of patients carry more than one mutation in different genes, and that future research must concentrate on the role that the different mutations play in the generation and evolution of the cardiomyopathy.
WP5 PROTEOMICS & METABOLOMICS
The activities of WP5 were planned to develop a systems biology framework for understanding the phenotypic and transcriptional changes identified in patients with DCM. Samples were provided by both human cohorts, as described in WP 4, and animal models of DCM, as described in WP 7. Samples were examined by metabolomics and, for tissue requiring further characterization, by proteomics, to validate proposed mechanisms of pathology in the other WPs. In addition, high throughput metabolomics was applied to human blood plasma samples to identify potential metabolic biomarkers of DCM. We measured metabolic changes in heart tissue and blood plasma from patients with DCM in order to identify potential biomarkers of disease or drug response as well as understand some of the pathological changes that accompany DCM. Given the priority assigned to cardiac laminopathies and the genetic heterogeneity of familial DCM as well as the complexity of the metabolomic tests, we decided to focus on laminopathies and on the comparative evaluation of metabolomics profiles in DCM compared with hypertrophic cardiomyopathies (36).
We first generated an omic database (ISA, Investigation/Study/Assay) for recoding, collecting and management of all data generated by this work package in INHERITANCE. We then examined the metabolic changes in a mouse model of DCM (Research area 3, LMNA knock out mouse). The purpose was to compare the metabolic changes found previously in human tissue with a mouse model, to identify metabolic commonalities that could be considered the key to the development of dilated cardiomyopathy, and to use these metabolic changes as markers of disease progression or drug efficacy. In order to provide a wide coverage of the metabolome we developed a number of high throughput assays for the quantification of metabolites in both heart tissue and blood plasma. This should be transferable to any cardiac disease model and we are currently publishing these assays. In summary we investigated lipidomics, carnitine metabolities, amino acids and nucleotides.
• Open profiling lipidomics: Liquid chromatography mass spectrometry (LC-MS) was used to profile total intact lipids extracted using the Folch procedure. The lipid extract includes a range of metabolites including triglycerides, diglycerides, phospholipids and cholesterol esters. The assay depends on, first, separating lipids by their polarity using liquid chromatography, and then detecting individual lipid species using high-resolution mass spectrometry. However, for any mass of an individual lipid specie there may be a large number of isomers which are isobaric (i.e. possess the same mass and hence will be indistinguishable by MS). To address this issue, tandem mass spectrometry has been performed to first fragment lipid species and then identify the parent compound using the mass spectra of the subsequent fragments. The total fatty acid complement of extracts of heart tissue was measured by gas chromatography (GC). Fatty acids are converted into fatty acid methyl esters as part of a transesterification derivatization and then analyzed by GC-MS. We have published these methods to allow others to adopt them for their metabolomics/lipidomics studies (37).
• Targeted analysis of carnitine metabolites. The carnitine shuttle allows the transport of fatty acids across the inner mitochondrial membrane, and given the central importance of fatty acid metabolism (beta-oxidation) in the heart, the carnitine derivatives are important measures of mitochondrial function and substrate selection. Because of this we developed a triple quadrupole LC-MS assay to profile 36 carnitine derivatives in heart tissue and blood plasma. In this assay either plasma or extracted tissue are analyzed (both aqueous and lipid fractions of the Folch extraction). Carnitines are measured on a Quattro Premiere LC-MS following butylation using 3M HCl in butanol. The original direct infusion method has now been replaced with a method using a C8 LC column to provide some chromatography to reduce ion suppression from polar lipids like phospholipids. A parent ion scan of fragments of 85 Daltons is performed to detect and quantify the carnitines. We have also applied this assay to other cardiac mouse models.
• Amino acid analysis. A quantitative assay was developed relying on polar reverse phase chromatography of the butylated amino acid derivatives. This assay is a targeted analysis of a range of amino acids and their derivatives including glutathione (both oxidized and reduced), nitrotyrosine, 3-hydroxy tyrosine, and dityrosine. This has allowed a range of amino acids to be detected in intact tissue in a quantitative manner. In addition we have extended this assay to include citric acid cycle intermediates.
• Nucleotides. An assay was developed for nucleotides to address two issues associated with DCM – i. changes in cardiac energetics and hence high energy phosphates such as ATP and ii. oxidized nucleotides as markers of ROS damage. The assay measures ATP, ADP, AMP, cAMP, GTP, GDP, GMP, cGMP, CTP, CDP, CMP, UMP, UDP, UTP NAD, acetyl-CoA, 8-oxo-guanine and 8-hydroxy-2-deoxyguanosine.
Patients with DCM vs. other cardiomyopathies
Application of our targeted analysis to the profiling of HCM and DCM human samples using valvular and ischemic cardiomyopathy disease tissue as controls were readily discriminated by principal component analysis of the metabolomics data. The model showed that the groups clearly separate by test centre and batch suggesting an effect associated with sample gathering. However, analysis of the individual clusters by a similar unsupervised method reveals a significant difference between control and disease groups. The analysis of the loadings plot shows that GTP, ATP, oxidised glutathione and ADP are increased in the disease group compared with the control group whereas glucose and fructose 6 phosphate, reduced glutathione, GMP and adenosyl methionine are increased in the control group. Given the test centre and batch clustering HCM samples were investigated further alone in order to glean biological information from the data and some discrimination could be determined according to genotype. Thus, metabolomics and lipidomics show promise in terms of discriminating individuals with different forms of DCM.
Analysis of plasma
We investigated 14 samples of unknown disease mutation, 1 of dystrophin gene mutation and 11 of the lamin A/C mutation from Beneficiary 1 and 20 samples of known gene mutation and 16 samples of unknown gene mutation from Beneficiary 9. These samples were examined using the assays developed for analysis of heart tissue in the hope of finding metabolic changes that extend to blood plasma and could be used to monitor non-invasively disease severity – an important issue in a set of diseases with variable penetrance not necessarily predicted by the gene mutation. High resolution 1H NMR spectroscopy, targeted amino acid analysis and the carnitine assay were performed to cover key metabolites in the metabolome. For each assay condition it was possible to discriminate blood plasma taken from beneficiaries 1 and 9, despite the samples being taken using an identical SOP. This observation is relevant for any future development planned in the setting of metabolomics studies, indicating that conditions other than those commonly respected and planned in a clinical SOP for genomics should be included and analysed.
Expanded model of the metabolome of DCM
Human DCM. In this task we report changes in metabolism in heart tissues of DCM/ HCM patients vs. healthy individuals. Then we compared this analysis with results on blood plasma. We confirmed the identity of the metabolites we analyzed using gas chromatography mass spectrometry (GC-MS) and nuclear magnetic resonance (NMR) spectroscopy. We found that GTP, ATP, oxidised glutathione and ADP are increased in the DCM disease group relative to the control group, whereas glucose and fructose 6 phosphate, reduced glutathione, GMP and adenosyl methionine are increased in the control group. We confirmed the identity of the metabolites we analyzed using GC-MS and NMR spectroscopy.
Mice DCM. (Figures 5.1 5.2) We used the assays described above to examine metabolism changes in heart tissue from LMNA null mice. These mice develop severe DCM in 5-6 weeks when homozygous and across 6-12 months when heterozygous. Analysis of the amino acid compliment of the heart tissue demonstrated that we could readily distinguish effects due to age, discriminating 2 week and 5 week-old animals. In addition, it was possible to discriminate homozygous animals from controls and heterozygous, indicating that DCM induced alterations in amino acid metabolism. In a similar manner we examined the carnitine profile of the heart tissue. However, only clustering according to age of animals was clearly detected by the multivariate statistics. Examining nucleotides and redox active compounds by LC-MS, again separation was detected associated with both age of animals and genotype, with a clear separation evident from 5 week old homozygous animals from 5 week heterozygous and control animals. This was in part driven by the concentration of glutathione present in the heart tissue.
We also analyzed heart tissue from 40 week old heterozygous LMNA mice. Tissue was analyzed in terms of amino acid content, acylcarnitines and polar metabolites (TCA cycle and glycolytic intermediates). While no separation could be detected for the acylcarnitine dataset, a weak model was built for amino acid metabolism and polar metabolites. In 40 weeks heterozygous laminopathic mouse hearts vs. controls (WP6) and after validation of the model using 100 random permutations of sample class membership we showed an increase of aspartate and glutamate in the heterozygous group with glutamine, alanine and phenylalanine decreased. Thus, we were able to determine both metabolic changes in homozygous and heterozygous mice associated with the LMNA mutation.
WP6 ANIMAL MODELS: ZEBRAFISH
The strategy of research for experimental models of DCM was addressed to both investigating phenotype in knock out, overexpression and transgenic zebrafish embryos and adults, to discover and characterize novel disease genes (35-38).
ZEBRAFISH studies were developed because a large proportion of familial and sporadic DCM cannot be explained by mutations in the disease genes that were known when INHERITANCE was planned (38). In fact, although classical human genetic approaches such as linkage analyses and candidate gene testing led to the identification of more than 40 DCM genes, known before INHERITANCE, some DCM families only counted one affected survivor but the family story and clinical report document a high number of DCM-related deaths. Therefore, conventional linkage studies are not feasible. Furthermore, for many mutations identified in human DCM the molecular mechanisms by which they lead to impairment of cardiac contractility and arrhythmias are not well understood. Large-scale forward and reverse genetic approaches in zebrafish were considered as powerful tools for the identification of disease genes for humans. Prior studies on zebrafish led to the identification of disease/candidate genes that cause DCM. Therefore, research in WP6, zebrafish, aimed at revealing the in vivo role of distinct DCM-associated regions (in cooperation with WP3), disease genes and distinct DCM gene mutations. With this approach, the specific aims were: 1) to characterize known DCM genes that are within the focus of INHERITANCE; 2) to evaluate the effects of selected mutations within frequent DCM genes, such as lamin A/C; 3) to characterize genes within genomic regions found to be associated with human DCM (WP3). We generated transgenic zebrafish lines for selected DCM-causing mutations. Knock-down, overexpression and transgenic zebrafish embryos and adults were functionally, structurally and molecularly characterized in detail (Figure 6.1). The results of these experiments revealed the functional role of newly identified genes associated with DCM and pathways crucial in the pathogenesis of DCM.(4 novel candidate genes: PINCH1 and PINCH2, SMYD1 and KCNQ2). The translational screening of these genes led to the identification of mutations in patients with DCM (38-41).
Functional, structural and molecular characterization of knockdown, overexpression and transgenic zebrafish embryos and adults.
We first established the strategies and methods for the characterization of genetically manipulated zebrafish as model organisms for human DCM. The goal to identify orthologs of human DCM genes, design antisense oligonucleotides and generate transient loss-of-function zebrafish was reached. Accordingly, these knockdown zebrafish represent valuable models to study the molecular pathways associated with DCM. The following zebrafish orthologs have been identified and are available for knockdown or overexpression studies within the other projects: d-Sarcoglycan, Myosin-light-chain 1, Myosin-light-chain 2, Vinculin, Lamin A, Muscle LIM protein, Myosin Binding Protein, Titin-cap, Desmin, a-Tropomyosin 4, Troponin-T2, Integrin-linked Kinase, Desmin, Nexilin, KCNQ2, Affixin, Ablim3, Dystrophin. The homology of these zebrafish proteins ranges between 40 to 95%. We designed specific antisense oligonulceotides to characterize loss-of-function effects for many of them. As an example, we characterized the novel DCM candidate genes PINCH1, PINCH2, Smyd1, KCNQ2. All knockdowns with the exception of KCNQ2 resulted in a DCM-like phenotype in zebrafish. This phenotype was characterized by decreased ventricular function (impaired Fractional Shortening) and reduced blood circulation and pericardial effusion (39,40).
To prove the causal role of a novel genetic variant for the establishment of a cardiomyopathy phenotype, overexpression studies of mutated genes in zebrafish were carried out. The following zebrafish orthologs of DCM (candidate) genes were full-length cloned into Gateway-destination vectors and selected novel human mutations were introduced in ILK, LMNA, PINCH1/2. For instance, we identified the novel, non-synonymous variant P70L in ILK gene by NGS. Next, we explored whether the P70L mutation affects protein confirmation. The mutation of the conserved residue is predicted to lead to changes in the secondary ILK protein structure. In the functional assays in zebrafish, we identified that ILK p.P70L is a hypomorph and hence cannot compensate for loss-of wt ILK function. Furthermore, as shown by co-immunoprecipitation, ILK P70L shows disturbed interaction with PINCH proteins, partners of ILK in the ILK-PINCH-Parvin mechanotransduction complex. Since alterations of the DNA can also be on an epigenetic level, we investigated the role of recently identified modifications of DNA methylation of several genes found in myocardium of patients with DCM and controls. For two genes, namely ly75 and adora2a we could dissect a functional role of the changed expression in the heart. These results are of great importance, since they underline a novel disease pathophysiology/mechanism of DCM, which might also be important for diagnosis and treatment.
As detailed above, the zebrafish allows to analyze the function of many different cardiac genes in a manageable time period. In fact, this leads to a rapid accumulation of functional data that need to be structured, organized and retrieved by the scientists. Hence, we decided to establish a project-internal zebrafish DCM database that allows fast access to data on DCM associated genes. Data encompass sequence homology, available Morpholinos, their corresponding knockdown phenotype and gene expression as well as subcellular localization. Data to be integrated were retrieved (from published and unpublished projects; references are included in the database; unpublished data are jointly generated with Prof. Dr. Rottbauer) and organized using standard bioinformatics tools and manual curation. Data were introduced in a database table format and include expression of genes, Morpholino sequences, target genes, quality controls measures, loss-of-function phenotypes and others. A first database for the integration of functional genomics data of DCM genes and candidates has been generated. We will further expand the system to assist ongoing and future research projects. We are currently investigating the functional effects of novel mutations identified by NGS mutation screening in the large patient cohort of the consortium.
Overexpression of selected DCM mutations in zebrafish and generation of transgenic lines.
Transgenic zebrafish lines are a valuable tool to dissect the impact of distinct gene mutations in the context of wild-type background or a loss-of-function zebrafish line. Accordingly, this task employs different strategies for the generation of transgenes, ranging from transient overexpression to stable transgenic lines. The first can be characterized in the embryonic state only, whereas the last can also be investigated in adulthood. Furthermore, transgene expression (wild type form of A) was introduced in gene A loss-of-function zebrafish to successfully demonstrate rescue of the phenotype. Hence, the transgenic approach is highly suitable to characterize the function of distinct genes and DCM gene mutations.
Further research in this WP included protein-protein interactions, in particular of contractile and regulatory proteins. Assembly, maintenance and renewal of sarcomeres require highly organized and balanced folding, transport, modification and degradation of sarcomeric proteins. However, the molecules that mediate these processes are largely unknown. We isolated the zebrafish mutant flatline (fla), which shows disturbed sarcomere assembly exclusively in heart and fast-twitch skeletal muscle. By positional cloning we identified a nonsense mutation within the SET- and MYND-domain-containing protein 1 gene (smyd1) to be responsible for the fla phenotype. We found SMYD1 expression to be restricted to the heart and fast-twitch skeletal muscle cells. Within these cell types, SMYD1 localizes to both the sarcomeric M-line, where it physically associates with myosin, and the nucleus, where it supposedly represses transcription through its SET and MYND domains. However, although we found transcript levels of thick filament chaperones, such as Hsp90a1 and UNC-45b, to be severely upregulated in fla, its histone methyltransferase activity – mainly responsible for the nuclear function of SMYD1 – is dispensable for sarcomerogenesis. Accordingly, sarcomere assembly in fla mutant embryos can be reconstituted by ectopically expressing histone methyltransferase-deficient SMYD1. By contrast, ectopic expression of myosin-binding-deficient SMYD1 does not rescue fla mutants, implicating an essential role for the SMYD1–myosin interaction in cardiac and fast-twitch skeletal muscle thick filament assembly (39).
Finally, taking advantage from the extensive research done in zebrafish, we isolated a mutation in zebrafish, bungee(bngjh177), which selectively perturbs valve formation in the embryonic heart by abrogating endocardial Notch signaling in cardiac cushions. We found by positional cloning that the bng phenotype is caused by a missense mutation (Y849N) in zebrafish protein kinase D2 (pkd2). The bng mutation selectively impairs PKD2 kinase activity and hence Histone deacetylase 5 phosphorylation, nuclear export, and inactivation. As a result, the expression of Histone deacetylase 5 target genes Krüppel-like factor 2a and 4a, transcription factors known to be pivotal for heart valve formation and to act upstream of Notch signaling, is severely downregulated in bungee (bng) mutant embryos. Accordingly, the expression of Notch target genes, such as Hey1, Hey2, and HeyL, is severely decreased in bng mutant embryos. Remarkably, downregulation of Histone deacetylase 5 activity in homozygous bng mutant embryos can rescue the mutant phenotype and reconstitutes notch1b expression in atrioventricular endocardial cells. We demonstrate for the first time that proper heart valve formation critically depends on Protein kinase D2-Histone deacetylase 5-Krüppel-like factor signalling (41) (Figure 6.2).
WP7 MICE MODEL
Lamins are proteins located next to the inner nuclear membrane of the cell as components of the nuclear lamina. Mutations in the LMNA gene are associated with several diseases including autosomal dominant Emery-Dreifuss muscular dystrophy and dilated cardiomyopathy. LMNA-mediated cardiomyopathy is often highly malignant; however the underlying mechanisms of disease are still unclear. Patients with a LMNA mutation have a serious prognosis due to high prevalence of heart failure and sudden cardiac death (3,4). Current treatment strategies to improve the prognosis are limited to implantable cardioverter-defibrillator and heart transplantation. However, no direct medical treatment has been described for these patients.
Mouse models of disease provide the opportunity to search for disease mechanisms and screen novel treatment strategies. Therefore, we will take advantage of mouse models of LMNA mediated disease, to enable pre-clinical screening of novel and existing strategies to treat these diseases. We developed a novel LMNA null mouse (LMNA(GT-/-) based on genetrap technology and analyzed its early post-natal development. We detect LMNA transcripts in the heart, the outflow tract, dorsal aorta, liver and somites during early embryonic development. Loss of A-type lamins results in severe growth retardation and developmental defects of the heart, including impaired myocyte hypertrophy, skeletal muscle hypotrophy, decreased amounts of subcutaneous adipose tissue and impaired ex vivo adipogenic differentiation. These defects cause death at 2 to 3 weeks post partum associated with muscle weakness and metabolic complications, but without the occurrence of dilated cardiomyopathy or an obvious progeroid phenotype (Figure 7.1). Our results indicate that defective early post-natal development critically contributes to the disease phenotypes in adult laminopathies (42).
Therefore, we planned to investigate both Lmna deficient and mouse models harbouring a previously characterized human LMNA DCM-causing mutation. Mice heterozygous for the Lmna mutant gene show a cardiac phenotype similar to that seen in human LMNA mutation–related dilated cardiomyopathy. Lmna heterozygous mice allow not only to investigate the pathophysiology of the myocardial damage and dysfunction but also to test the effects of existing drugs (e.g. beta-blockers, ACE inhibitors) to prevent or delay DCM and heart failure. Therefore, we investigated both existing and novel experimental models of cardiolaminopathy. Long-term survival was reduced in Lmna+/- untreated mice because of sudden death or heart failure necessitating euthanasia. Survival of medically treated groups was comparable to wild type untreated mice. In order to evaluate conduction disturbances that are inherent to the Lmna+/- phenotype and whether Metoprolol/Enalapril can prevent or delay these conduction disturbances, we performed ECGs at 50, 70 and 75 weeks of age. ECG data-analysis at longer intervals is ongoing. Furthermore, extensive molecular characterization (stress markers, fibrosis, apoptosis markers and oxidative stress) is ongoing to clarify the beneficial effect of Enalapril treatment in preventing Lmna+/- related DCM.
Existing model
After obtaining ethical approval to breed and use the Lmna A/C-null mice for transcriptomics, proteomics and metabolomics and testing existing drugs on these animals, we achieved the full Lmna knockout model and described the age-dependant phenotypes in homozygous and heterozygous mice. Heterozygous mice display a similar phenotype to laminopathies in humans. In addition, we collected hearts and blood plasma from control, heterozygous and homozygous Lmna-null mice for metabolomic studies. The mouse is a somatic Lamin A/C knockout, where exons 8 to part of 11 are replaced with the Pgkneo neomycin resistance cassette in reverse orientation to the Lmna gene. This full Lmna knockout model was generated in the lab of Colin Stewart (Sullivan et al 1999, Joural of Cell Biology). Homozygous LaminA/C-null mice display a delayed postnatal growth after two weeks. Muscular weakness and a dilated heart are apparent at 3-4 weeks, after which mice die at 8 weeks of age. This is accompanied by the known features of human loss of lamin A/C including failure to thrive, energy depletion and extremely shortened life span. Although the complete loss of lamin A/C does not reproduce human disease, as this is characterised by partial loss of lamin, or a dominant negative effect, heterozygous mice are a general reference model for human laminopathy. The heterozygous LaminA/C-null has only a cardiac phenotype encompassing conduction system defects at 10 weeks. After 40 weeks, the heart becomes dilated and less contractile. Overall, these mice display a similar phenotype as laminopathies in humans.
We upscaled the breeding of Lamin A/C-null mice to obtain enough mice for both “–omics” approach and for testing the effects of drugs. “-Omics” studies were done in hearts and blood plasma from control, heterozygous and homozygous Lmna-null mice at 2, 5, and 40 weeks of age (before and after the onset of dilated cardiomyopathy). Each experimental group consisted of 8 mice. Heart function and disease progression was monitored by echocardiography. Before sacrifice, we determined the body weight, heart weight and tibia length for each mouse. LMNA +/- mice develop a DCM phenotype similar to human DCM.
Novel Disease Models: H222P-Lmna knockin mouse model
We obtained ethical approval for the H222P-Lmna mice for –omics studies and to test the effect of existing drugs on progression of heart failure. We tried obtaining breeding couples of a knock-in mouse model that carries the H222P-Lmna missense mutation from the lab of Gisèle Bonne, which was previously described in human laminopathy (Arimura et al., Hum Mol Gen 2005). These mutant mice exhibit overtly normal embryonic development and sexual maturity. At adulthood, male homozygous mice display reduced locomotion activity with abnormal stiff walking posture and all of them die by 9 months of age. As for cardiac phenotype, they develop chamber dilation at 2 months and hypokinesia with conduction defects. Therefore, LmnaH222P/H222P mice represent an excellent model to study laminopathies as they develop a dystrophic condition of both skeletal and cardiac muscles similar to the human diseases. Current data show that ACE inhibitor Enalapril prevents development of DCM/HF in Lmna +/- mice, while Beta-blocker Metoprolol cannot prevent nor delay the progression to DCM in Lmna+/- mice. We are currently investigating the degree of fibrosis, apoptosis and oxidative stress in all experimental groups. In addition, ECG analysis is ongoing to determine whether Enalapril also prevents conduction disturbances. The results suggest that treatment of LMNA patients with Enalapril would be a very effective therapy in preventing the onset of DCM.
Survival of Lmna wildtype and Lmna +/- treated and untreated control mice.
To screen efficacy of existing treatment (e.g. β -blockers, ACE inhibitors) to prevent or delay DCM and heart failure, we designed an experiment in which we test Metoprolol (β-blocker) and Enalapril (ACE inhibitor) in Lmna mutant mice. Metoprolol is a widely prescribed drug to treat heart failure. It decreases heart rate, contractility and cardiac output, therefore decreasing blood pressure. Enalapril is an angiotensin converting enzyme (ACE) inhibitor, which lowers blood pressure by preventing constriction of blood vessels. This drug is also widely prescribed to treat heart failure. These drugs are often prescribed at a stage were heart failure is already apparent. In our study we want to examine the effects of these drugs before the onset of Lmna-related DCM specifically.
Experimental groups were as follows (n=20 mice per group):
• Wildtype: control group, no treatment
• Lmna+/-: control group, no treatment
• Lmna+/-: Metoprolol (30mg/kg/day)
• Lmna+/-: Enalapril (10mg/kg/day)
To determine whether early administration of β-blocking drugs prevents DCM, Metoprolol or Enalapril was administered in drinking water to male WT and Lmna+/- mice from 10 weeks of age. Heart function was monitored by echocardiography and ECG, after which mice were sacrificed at 75 weeks of age. Extensive morphology, histology and molecular characterization of hearts are currently ongoing. The experimental endpoint was in early 2013, when mice were sacrificed and hearts were harvested for morphological and molecular analysis. The Lmna+/- control and Metoprolol-treated mice developed dilated hearts. Interestingly, the Enalapril treated mice did not develop a dilated phenotype; however, the hearts were noticeably smaller than the Lmna+/- and Lmna WT control groups. When we compared the myocyte size of the Enalapril hearts with the Lmna+/- control and wildtype control, we observed smaller myocyte size in the Enalapril treated mice. Currently, heart function and progression of DCM has been monitored in all 4 experimental groups by means of echocardiography at 10, 40, 50, 65, 70 and 75 weeks of age. Left ventricular function gradually decreased in Lmna+/- mice as expected. Metoprolol treatment was unable to prevent or delay the decrease in left ventricular function of Lmna+/- mice. However, Lmna+/- mice treated with Enalapril were comparable to wildtype mice in LV function. This means Enalapril treatment is able to prevent the naturally occurring decrease in LV function of Lmna+/- mice. LV end-diastolic and end-systolic parameters were lower in the Enalapril treated mice, suggesting an enhanced LV contractility in this experimental group. Long-term survival was reduced in Lmna+/- untreated mice because of sudden death or heart failure necessitating euthanasia. Survival of medicine treated groups was comparable to wild type untreated mice. In order to evaluate conduction disturbances that are inherent to the Lmna+/- phenotype and whether Metoprolol/Enalapril can prevent or delay these conduction disturbances, we performed ECGs at 50, 70 and 75 weeks of age. However, ECG data-analysis is ongoing. Furthermore, extensive molecular characterization (stress markers, fibrosis, apoptosis markers and oxidative stress) is ongoing to clarify the beneficial effect of Enalapril treatment in preventing Lmna+/- related DCM.
Our current data show that ACE inhibitor Enalapril prevents development of DCM/HF in Lmna +/- mice, while Beta-blocker Metoprolol cannot prevent nor delay the progression to DCM in Lmna+/- mice. We are completing the study by investigating the degree of fibrosis, apoptosis and oxidative stress in all experimental groups. In addition, ECG analysis is ongoing to determine whether Enalapril also prevents conduction disturbances.
WP8 STRUCTURAL STUDIES
In order to identify and characterize the molecular effects of DCM-related mutations we undertook a structural biology approach, which provides atomic resolution models of selected mutant proteins, together with a biophysical characterization of their basic properties. X-ray crystallography allows to check the effects of mutations in terms of structure variability of the affected protein, providing a rational basis for the analysis of its overall stability. In the initial stage of the Project, our study was focused on the Lamin protein (LMNA), which hosts several of the currently known DCM mutations. LMNA is composed of a central α-helical “rod” domain, comprising four coiled-coil segments (Coil1A, Coil1B, Coil2A and Coil2B), non-α-helical “head” and “tail” domains. Part of our study was focused on Emerin (EMD), a protein which has been proven to interact directly with Lamin; moreover, mutations in the Emerin gene cause DCM, similarly to Lamin mutations. Emerin contains a LEM domain, a lamin-binding domain (EMDlbd) and a transmembrane domain.
Lamin We cloned, expressed and purified different constructs of the Lamin and Emerin genes and DCM-related mutants. Full length Lamin (LMNA, residues 1-660) was expressed and tentatively purified following conventional chromatographic methods used in protein biochemistry; however, the eluted protein was not pure and showed a marked tendency to aggregate and precipitate. Under these conditions, the crystallization of the protein was completely unsuccessful. The presence of disordered protein regions may affect the solubility/stability of the protein through aggregation. For this reason, a new construct (LMNAcentr) lacking the first 28 residues (preceding the Coil1A domain) and the last 115 residues (following the tail domain) was cloned during the second year. Also in this case we did not manage to obtain crystal hits. Size-exclusion chromatography experiments showed that LMNAcentr exists in a monomer/dimer equilibrium, while it is known that a unique aggregation status is fundamental to enhance crystallization. In vitro phosphorylation was proven to abolish linear assembly into polymers, but does not affect dimer formation. For this reason, a further p-likeLMNA construct was designed containing two mutations, S22E and S392E; both Ser residues are known to undergo phosphorylation. Indeed, mutations to glutamate mimic the phosphorylated status, due to their negative charge, and could help us isolate the homogeneous dimeric form of the protein for crystallization experiments. Expression and purification of the p-likeLMNA was successful, and the purified protein was used in wide crystallization screens; no crystals have been obtained so far. Various fragments of the protein, corresponding to different domains, were subcloned: Coil1A (residues 29-101), Coil2B (residues 303-387) and Tail (residues 432-544) domains. Moreover, mutants R335W, E347K, E342K, and E358K in the coil2B domain were expressed and purified.
Emerin: EMD is a serine rich protein, very difficult to express and purify since it contains a transmembrane domain, composed of a highly hydrophobic region, which leads the protein itself to aggregation. Moreover, EMD has been predicted to contain scattered disordered regions and low complexity regions, which compromise the protein stability. On a different approach, EMDlbd (Emerin lamin binding domain, residues 72-168) was expressed in the recombinant form and used for crystallization trials. The eluted protein was pure enough and stable in solution, but no crystals were obtained after thorough exploration of the crystallization conditions that led only to huge precipitation mass. Following these experiments, we believe that, in order to attain a defined conformational state, EMDlbd requires the presence of a binding partner. For this reason, we performed co-crystallization of EMDlbd with Tail domain, LMNA, LMNAcentr and p-likeLMNA, but no crystallization has occurred so far.
Experimental procedure
All the proteins were expressed in E. coli strain BL21(DE3) grown in Luria-Bertani medium. In the case of Coil1A, Coil2B and Coil2B mutants, the purification was performed under denaturing conditions. After induction, the cells were harvested and lysed by sonication. The supernatant was loaded on Nickel Sepharose Fast Flow resin (Ge-Healthcare). The fusion proteins were dialyzed overnight into the appropriate buffer for crystallization experiments.
Protein characterization
Recombinant Coil2B mutants and Coil1A were produced through a refolding procedure starting from denatured inclusion bodies; then, the proteins were characterized to verify their correct refolding after purification. CD analysis was performed in order to verify the correct folding of the purified protein. The CD spectra highlight a correctly folded protein composed of 87% α-helix secondary structure. This result is consistent with the Coil2B 3D-structure, which consists of a single α-helix. The remaining 13% of non α-helix is presumably referred to the His-Tag residues (6 residues) and the linker between HisTag and Coil2B added in the cloning procedure (26 residues). DLS measurements carried out at different Coil2B concentrations showed that the polydispersion (defined as the standard deviation of the hydrodynamic radius) of the protein samples was high at low concentration (1-5 mg/ml), while increasing concentrations decreased polydispersion. At 10 mg/ml Coil2B resulted monodisperse and thus suitable for crystallization experiments. HPLC-light scattering and analytical purification were performed to study the formation of the Coil1A/Coil2B dimer, however, no dimerization was detected so far.
Crystal screening
All the LMNA protein constructs were concentrated to 10 mg/ml using an Amicon Ultra centrifugal filter (Millipore) and used for vapour-diffusion sitting-drop trials assembled using an Oryx-8 crystallization robot (Douglas Instruments, East Garston, UK). All crystallization trials were performed at 20°C by mixing a 0.2 μl droplet of the protein solution with 0.3 μl of a reservoir solution. Crystals of Coil2Bwt, R335W, E347K, E342K and E358K mutants were obtained under several conditions. At the very beginning of the study, we observed that after two weeks at 20°C no crystals grew, but vitreous spheres were present under various conditions. To overcome the formation of undesired vitreous aggregation, the crystallization trials were first incubated at 20°C and successively at 37°C, at which temperature most of the spheres turned into crystals. In the case of the E347K mutant, numerous crystals, presenting different shapes and grown under different physico-chemical conditions in the initial screening, were obtained. In contrast, crystals of E342K and E358K grew under a single condition and, consequently, optimization of such growth condition was performed. The largest crystals were cryo-protected and flash-frozen in liquid nitrogen. X-ray diffraction tests were performed almost every month at the European Synchrotron Radiation Facility (Grenoble, France).
X-ray diffraction
We obtained diffraction data sets for Coil2Bwt, R335W and E347K. Crystals of Coil2Bwt and of the R335W mutant diffracted both to 3.2 Å, while crystals of E347K diffracted to 3.6 Å resolution (38). Despite the large number of crystallization and diffraction trials on E358K and E342K mutants, X-ray diffraction from all crystals was always extremely poor, possibly because these two mutants may represent a significant barrier for the correct formation of the coiled-coil dimer in the context of the quaternary assembly of the protein. Models of the E358K and E342K mutations were created, and a possible explanation on the effects linked to the mutations is proposed here below. The LMNA 1A-2B dimer was homology-modelled based on the corresponding domain of human vimentin A (PDB codes 1GK4, 1GK7). The E342K and E358K mutants were modelled on the assembled dimer (Figure 8.1). Residue 358 is exclusively acidic. The introduction of a positive charge in affected individuals was proposed to disrupt the overall acidic character (while altering the electrostatic charge) of this region of the helix. The model showed that the mutated residue lies at the interface between the two coiled helices, likely interfering with proper assembly of the dimer. In contrast, residue E342 lies on the surface of Coil 2B, on the opposite side with respect to Coil1A-2B dimer interface. Substitution with Lys, namely substitution with a positively charged residue, will not affect the formation of the dimer with Coil1A, as for the E358K mutation; however, it will possibly hinder the formation of tetrameric protofilaments, which assemble through the association of lamin coiled coil dimers in a head-to-tail fashion.
As observed for Coil2Bwt and R335W, crystals of the E347K mutant belong to the hexagonal space group P6522. The diffraction data were processed and scaled. The crystal structures were solved by molecular replacement using the known Coil2b structure as search model. A single protein molecule was located in the crystal asymmetric unit (Figure 8.2). The positioned model molecule was subjected to refinement. The amino acid sequence of the model was modified to match the mutant correct sequence and fitted to the electron density. The final model displays excellent overall stereochemical parameters, with 100% of the residues comprised within the allowed regions of the Ramachandran plot. Description of the structures is reported in “Significant results”.
Drug search
Since the structure of the full length LMNA and/or complex of the domains are not yet available, a search for a drug lead interacting with a single alpha-helix appears poorly meaningful. In fact, drug binding to a protein requires a binding cleft, or similar site, which would allow specific recognition and proper free energy of binding. Such sites are coded in the globular domains of proteins, or in protein/protein interface regions. The helical structures obtained during the project do not present such favourable potential druggable sites.
Gene product structures
The INHERITANCE project requires knowledge regarding the molecular bases of dysfunctions in genes linked to DCM (based on the work and discoveries of other Partners). In order to select crystallizable constructs of LMNA and Emerin, thorough information on all the domains was collected. In particular, it was discovered that in vitro phosphorylation of chicken lamin B by cdc2 Kinase abolishes linear assembly into polar head-to-tail polymers, but does not interfere with dimer formation. According to the PhosphoSite database, more than 30 phosphorylation sites in human A-type lamins have been reported. Phosphorylation of Thr19, Ser22 and Ser392 leads to depolymerization of lamin filaments during nuclear envelope breakdown in mitosis and meiosis. This information leads to the design of the p-likeLMNA. Furthermore, since it is very difficult to express and purify the EMD protein, due to its transmembrane domain that leads the protein itself to aggregation, basing on literature we designed the EMDlbd construct (Emerin lamin binding domain, residues 72-168), which was reported to bind the Tail domain of LMNA.
Significant results
We solved the crystal structure of Coil2Bwt, R335W and E347K mutant. Structures comprise amino acids 313 to 386 (sequence numbers refer to the full length LMNA). The Coil2Bwt structure consists of a single long α-helix. X-ray diffraction analysis revealed that the crystals present one polypeptide chain per asymmetric unit. However, the coil builds up a dimer through assembly with a symmetry equivalent molecule in the crystal lattice. The dimer is positioned on a crystallographic 2-fold axis, and exhibits a number of salt bridges, which stabilize the interactions between the two monomers. Superimposition of Coil2B and its mutants showed that the overall conformation of the molecules is strongly conserved, as expected considering the single residue mutations and crystal isomorphism. The side chains of Arg335 and Glu347 are shown to be exposed at the surface of the coiled coil dimer; nevertheless, the mutations occur at sites that are highly conserved in the known LMNA amino acid sequences. A mutation of Arg335 to Trp may directly interfere with the coiled-coil stability, either for steric hindrance or charge modification. A destabilization of such assembly, or the achievement of a looser coil, may be the cause of the pathogenic effect of the mutation R335W. The substitution of Glu347 with a positively charged residue, such as Lys, may interfere with the formation of putative intra- or inter-helical salt bridges, or other forms of polar interactions. Both mutations may alter the interactions of the coiled-coil dimer, and consequently interfere with the lamina assembly; alternatively, they may affect the binding of components within the nuclear lamina, including nuclear factors that have been proposed to interact/associate with LMNA. These results were included in the paper, which was published on Biochemical and Biophysical Research Communications (43).
WP9 - THERAPEUTICS AND IMPROVEMENT OF MEDICAL MANAGEMENT
In this WP we explored the possibility of:
1. Prevention or delay of the onset of the DCM phenotype in asymptomatic mutation carriers
2. Prevention or delay of DCM in animal models: the Myopalladin model
Prevention or delay of DCM phenotype in asymptomatic gene carriers (PSL, Paris)
We designed a Pilot Clinical Trial with angiotensin-converting enzyme (ACE) inhibitor in patients at a preclinical stage from families with Dilated Cardiomyopathy, entitled INHERITANCE “PRECARDIA” Study or PRE clinical mutation CARriers from families with DIlated cardiomyopathy and ACE inhibitors. The aim is to study the impact of ACE inhibitors (ACEi) on patients who carry a mutation but have not developed DCM yet through a double blind randomized multicentre trial. The hypothesis is that ACEi may delay or prevent the occurrence or delay the onset of DCM in these patients (pre-clinical stage), thus modifying the natural history of the disease (44).
The design of the study is as follows:
Inclusion criteria. Relatives with a DCM-causing mutation (from families with DCM) without obvious DCM but with minor left ventricular (LV) abnormalities: isolated slight LV enlargement or reduced systolic dysfunction (echocardiography).
Number of patients to enrol. N = 200.
Duration of the recruitment. Expected duration between 1 and 2 years.
Duration of the follow-up and treatment. Follow-up = 3 years.
Treatment. Perindopril 10 mg once a day versus placebo. Randomisation 1 / 1.
End points. Primary end-point: deterioration of LV end diastolic diameter/volume or Ejection fraction (echocardiographic or magnetic resonance imaging). Secondary end-points include the evolution of other echocardiographic parameters or hormonal biomarkers (including Mid-Regional pro-Adrenomedullin).
Main steps of the trial (summary) until the closure of the study:
- The study (clinical trial: NCT01583114, reference sponsor C10-44 Precardia, N° EudraCT 2010-023184-18) was formally approved by the sponsor (INSERM) in January 2011(44).
- The study was approved (August 2011) at the European level by the Clinical Trial Facilitation Group (CTFG) at the HMA (Heads of medicine Agencies - network of the Heads of the National Competent Authorities in the European Economic area) according to a unified VHP single process (VHP or Voluntary Harmonisation Procedure).
- The first participant was enrolled in December 2011.
- By November 2013, only 3 investigation centers were opened and only 12 participants were enrolled.
- Early December 2013: the enrolment was stopped (because of the end of the two–year period planned for enrollment), and an amendment was submitted by the PI to the sponsor.
- Decision by the sponsor (Inserm) to prematurely stop the trial during December 2013 with an official decision dated 28th January 2014. The decision was not related to scientific reasons, but only to logistic aspects of the trial. Progress did not meet the requirements for sustained feasibility of the study.
Enrollment of participants.
The enrollment was delayed due to the length of the administrative processes that delayed the opening of investigating centres. At the end of 2013, only three centres were opened and active: Paris (starting December 2011), Pavia (starting December 2012) and A Coruna (starting July 2013) (Annex 1).
Inclusions in 2011: 1 patient
07 001 RL 000 009 01/12/2011 France
Inclusions in 2012: 6 patients
07 002 PS 000 010 17/01/2012 France
07 003 GN 000 011 24/01/2012 France
07 004 B N 000 101 19/06/2012 France
07 005 J M 000 102 03/07/2012 France
01 001 BR 000 109 03/12/2012 Italy
01 002 GM 000 110 19/12/2012 Italy
Inclusions in 2013: 5 patients
01 003 MG 000 111 09/01/2013 Italy
01 004 MP 000 112 09/01/2013 Italy
01 005 VR 000 113 22/05/2013 Italy
01 006 PM 000 114 15/07/2013 Italy
07 006 PP 000 103 15/10/2013 France
TOTAL = Number of patients enrolled by 31/11/2013: 12 patients.
Decision of INSERM (sponsor) to stop the trial.
Decision. A letter written by Ms Sonia Gueguen (Pôle Recherche Clinique) in French, signed on the 11th December and addressed to the PI of the trial (Ph Charron), recommended to stop the trial. Then, there was a lot of exchanges and discussion between Inserm, the PI of PRECARDIA and the coordinator of INHERITANCE. In a second letter in English, Inserm (Clinical research department, Mrs Gueguen), confirmed the decision to stop the trial and indicated that the official closure of the trial was on the 28th January 2014. In the letters, the sponsor indicates that, unfortunately and despite the undeniable effort that had been made, there was no sufficient progress in the general advance of the project, especially because of the delay in centers opening / enrollment and also because of under-funding of the WP. Progress did not meet the requirements for sustained feasibility of the study. As a consequence Inserm decided to prematurely stop the trial. During further discussion, Inserm confirmed that the decision was NOT related to scientific reasons, but only to operational and feasibility aspects of the trial.
Prevention or delay of DCM in animal models: the Myopalladin model
We previously identified the Myopalladin gene as involved in monogenic DCM with the identification of several mutations causing DCM in independent DCM families (Duboscq-Bidot L et al. Cardiovasc Res, 2008). There are two main parts in the present project:
1) monitoring the cardiovascular phenotype of the knock-in mice to fully characterize the pathophysiological consequences of the expression of the P1112L mutation in heterozygous and homozygous mice.
2) once the disease and its time course are described, evaluating the phenotype reversion potential of siRNA, specifically targeting the mutant allele.
Mouse model production and phenotyping
Successful embryo delivery occurred in January 2011, resulting in the birth of the heterozygous mouse, F1 heterozygous mice and homozygous mice by November 2011. Echography was then performed in mice up to the age of 18 months. Results showed that from the age of three months, both heterozygous and homozygous mice showed clear signs of a dilated cardiomyopathy.
By month 9, we observed in homozygous mice that LV end systolic volume was increased as compared to wild-type (p < 0.001) and LV ejection fraction values was reduced (p < 0.0001). Of note, the DCM echography based phenotype was reliably observed at 18 month of age (Figure 9.1). Also interestingly, generally across the whole assessment period, heterozygote echocardiographic values are non-significantly different to those of homozygotes, which might be an indication of a dominant negative effect of the mutation, in line with what observed in human mutation carriers (Duboscq-Bidot, 2008).
Mice sacrificed for the six month characterization were weighed and then dissected. No gross morphological differences in the hearts from different genotypes were evident by eye. However, mouse hearts from both heterozygous and homozygous were heavier than wild-type hearts (p < 0.05 using unpaired two-tailed Student’s t-test for heterozygote hearts only) (Figure 9.2). Introducing the mutant allele which includes the P1112L mutation and the siRNA targeting sequence into a mouse strain could have generated unexpected alternative splice sites, resulting in potential abnormal myopalladin mRNA expression. Therefore we performed PCR on cDNA generated from RNA extracted from heterozygous and wild-type mice. Comparison of PCR products indicated that no aberrant splicing had occurred.
Results of shRNA expression in Myopalladin KI model
We validated the allele-specific target introduced into the recombinant vector utilizing an in vitro luciferase based assay. shRNA (specifically targeting the mutant mypn allele) were more stable in vivo and less susceptible to degradation than siRNA, inducing us to re-align our strategy to produce AAVshRNA constructs. We used the helper plasmid pDG9, resulting in an AAV2.9 vector, which was shown to demonstrate cardiotropism.
To measure whether specific shRNA knock-down of the mutant allele would have occurred, an allele-specific qPCR was designed and optimized (Figure 9.3) indicating that the allele-specific qPCR is capable of cleanly distinguishing between the two alleles. In addition, we used this qPCR to rule out that introducing the mutation and shRNA targeting sequence altered the tissue expression of myopalladin in the knock-in mice. We performed and analyzed the results of a large pilot AAV eGFP infection study in normal wild-type mice in order to define the optimal injection mode, dose and serotype best combination. We infected a total of 34 C57BL/6 mice at the age of six weeks with either AAV2.1 eGFP (serotype 1) or AAV2.9 eGFP (serotype 9) virus. Mice were sacrificed after four weeks of infection, and heart, leg muscles and liver samples were taken and processed for RNA extraction (for subsequent qPCR), protein extraction (for western blotting), and embedded in freezing medium for immunofluorescence.
Following the optimization of the AAV infection strategy using the AAV2.9-eGFP virus produced in-house, we generated the following AAV2.9 viruses: 1) AAV2.9-shALS9-dsRed; 2) AAV2.9-shLuc-dsRed; 3)AAV2.9-dsRed. The goal of the study is to counteract the development of the dilated cardiomyopathy caused by the mutant protein by the removal of the mutant allele via shRNA silencing. Prior to infecting the entire study cohort with our viruses, we tested the viruses in a second pilot study in order to determine their effectiveness. In order to ensure both adequate viral infection and shRNA production during the early stages of disease development, we injected small groups of young adult mice from the three genotypes at the age of 6 weeks (Figure 9.4) using the AAV2.9 eGFP.
Being the INHERITANCE funding period now terminated, we are continuing the project under an independent funding from ICAN foundation (ICAN, Paris, France). FP7 INHERITANCE funding will be acknowledged in any publication/communication arising from this work.
WP10 BIOINFORMATICS DATABASE (45)
The INHERITANCE project has required the recompilation and integration of all the generated data in a common database. To allow effective intra and inter group collaboration of both clinicians and basic researchers, the database had to be Web-based. The database includes multiple types of data, which have been recovered and integrated. No standard definitions existed for a large number of the variables that had to be recorded (as those provided by CARDS). Standardization of such variables is an immediate benefit of the Project. There were pre-existent data in several databases that had to be recovered and integrated in the common database.
The input of data had to be easy and user-friendly and the accuracy of the included data had to be warranted (validation tools and quality controls). Collaboration and data sharing are fundamental, but to ensure maximal collaboration the database management system had to warrant that each group has full control over the use of their own data. Ethical and legal issues about confidentiality, privacy and anonymous data had to be solved. Exploitation of the data and integration of tools for data query and analysis had to be considered in the design of the software for the database. Loss of the data had to be prevented at central and local levels (security copies of the database at different levels).
To this end, Health In Code and the INHERITANCE Consortium have developed a dedicated INHERITANCE database, and Health in Code has integrated the software with the Cardioregister database, an International/National Register of Familiar Cardiomyopathies. Cardioregister is a voluntary Database, web-based (internet access), designed to collect, exploit and download anonymous data of patients and families with cardiovascular diseases, offering the ability to produce customized reports with data. It is organized in families and it has quality controls established by the participating investigators. Cardioregister has been designed as an instrument useful not only for data collection but also to support decision-making and family screening. In the INHERITENCE project, Cardioregister has been integrated with the system i2b2, customized for the needs of the project by the University of Pavia (WP11). i2b2 is a data warehouse and data mining project that was designed to allow fast querying and interaction the data collected so far. I2b2 and Cardioregister are periodically synchronized. Together with the data collected in the INHERITANCE database and i2b2 (WP11), the Inheritance consortium had a large number of patients’ data already available for summary statistics and pooled data analysis. The joint exploitation of the newly collected data and the data available at the background of the project is a crucial factor for the achievements of the project goals.
Bioinformatics Database and e-CRF (45).
We generated and implemented a software and database supporting the interdisciplinary activities of the INHERITANCE project. After having achieved this aim in the first period of the project, in the second period we further developed the eCRF for the clinical trial PRECARDIA INHERITANCE (WP9), planned the release of the commercial product (eCRF) and performed the SWOT analysis for the evaluation of the strengths, opportunities and threats, as needed in a business venture. The software and the database were developed according to the consensus reached by partners during the first period of the research. The last version of the Inheritance Database is 1.1.3.17790 installed in September 2012. In addition, a PRECARDIA eCRF was released in the second part of the Project. A multidisciplinary team of computer engineers and cardiologists developed the software and the database following the design and the recommendations made by the INHERITANCE working group. The whole system was tested and verified by the team of Health in Code and submitted to Inheritance group that approved and validated it; the software was used during all the project duration. Verification by Health in Code ensures that the product was built according to the requirements and design specifications, and software validation carried out by INHERITANCE consortium ensures that the product actually meets their needs, and that the specifications were correct (see details below).
1- The Inheritance database (Fig. 10.1)
The introduction of clinical data is still open. The database already includes data from previous data sets and new patients and family from the different participant centres. The available version of the Database has been developed following the principles established in the project. It allows effective collaboration of the partners with the introduction of their data in a common base. The variables of the database have been defined according to either existing standards or the criteria of the consensus of the consortium. The system works with anonymous data and provides confidentiality and security for the stored data. Quality control systems have been implemented. The system also allows the exploitation of the data by the partners. Prevention tools to avoid loss of data have been implemented (security copies). The DB is structured in sections, extensively described in deliverables D10.1 D10.3 and first and second reports.
2- The Precardia eCRF (Fig. 10.2)
The PRECARDIA clinical trial is one of the most challenging and ambitious activity of the Inheritance Project (WP9). Within the limited funding of the INHERITANCE project, all the consortium members have to contribute to the implementation of all the requirements for the correct development of a clinical trial. Health in Code (Partner 6) as the leader of the Research Area 6 (Bioinformatics) contributes to WP9 with the development, maintenance and hosting of the specific software application dedicated to the electronic recording of the data generated by the clinical trial (eCRF).
The software has been designed and developed keeping recommendations and forms designed by the project consortium. A multidisciplinary team of computer engineers and cardiologists, following the design and the recommendations made by the INHERITANCE working group and guidelines to develop eCRF systems, have developed the software and the database. The whole system was tested and validated by the team of Health in Code and submitted to Inheritance group that approved it, and it was used during the project duration. The last modified version is 1.0.4689.26422 that includes all pending corrections and modifications, and was installed in November 2012.
Data represent a decisive component of the Inheritance project. For this reason, we pursued a strategy based on three directions within Inheritance: 1) First, we have developed and implemented database and management system software that supports the needs of the project and allows the start of the recovery and integration of clinical and genetic data. The data of the patients included by the partners of the Inheritance project are being inserted in this Database. The consortium has approved the whole system. 2) Second, we have developed a data warehouse and data mining tool (WP11) that is constantly synchronized with the INHERITANCE database to quickly analyse and summarize the available data. The consortium has approved the whole system. 3) Third, we have developed dedicated software for the electronic CRF that gives support to the PRECARDIA clinical trial. Data will be introduced in this specific database by the INHERITANCE partners and will be available for review and queries by the authorized researchers; granting the fulfilment of all the security, confidentiality and data protection legal and ethical requirements. The consortium has approved the whole system. The validation and verification of INHERITANCE database and PRECARDIA eCRF have been achieved.
Health in Code is now working to launch this new product (eCRF) on the market, and is designing a market study that collects information about:
• Environmental conditions
• Potentially target clients
• Principal competitors on the market
• Products that have been substituted.
• Client expectations about the software and its functionalities
• Zoning and sales volumes
• The advertising and promotional strategy best suited to reach target customers
Once these steps are completed, we will massively launch the product on the market; meanwhile it is being offered on a small scale to commercial customers and partners of Health in Code. A new line of business has opened for Heath in Code, due to the participation in the INHERITANCE project. This line is gathered to the software development and the new commercialization of an eCRF product developed within the PRECARDIA project. The eCRF is now a commercial product that can be adopted in cardiology clinical trials.
WP11 KNOWLEDGE MANAGEMENT AND DATA ANALYSIS
The main task of WP11 was addressed at providing the INHERITANCE project partners with decision support tools and knowledge discovery models. In particular, the database tools developed in WP10 could be empowered by a set of web-based functionalities including data analysis workflow management, intelligent querying of the phenotype data, text mining of the relevant literature, and case-based reasoning.
The data warehouse
We developed a data warehouse supporting fast phenotype data mining and exploration. An automated management system can retrieve both data on biologic existing knowledge and novel discoveries from the literature. The implementation of this system can support decision making clinically oriented genetic tests. The data warehouse is based on the open source software developed by the i2b2 project (https://www.i2b2.org). A customized version of the data warehouse has been populated through a set of automated queries that extract patients’ data from the INHERITANCE database (Cardioregister, see WP10). A web client (46) query interface allows ad-hoc queries to be created by research clinicians and returns aggregate numbers of patients that satisfy the queries, as well as anonymized specific patient’s record data. A set of tools for automatic biological knowledge retrieval tool and literature-based discovery was made available to the project partners.
The Literature Mining tool for the automated knowledge extraction from the scientific literature has been conceived to provide utilities for automated elaboration of textual content, so that it is possible to automatically extract and interpret the information available in literature (47). It is based on the GATE framework (http://gate.ac.uk) and implements an iterative version of the Swanson’s ABC model to find associations in the literature (48). It has been tested in the case of dilated cardiomyopathies and of hypertrophic cardiomyopathies.
We also implemented a system for supporting reasoning and decision-making, by intelligently analyzing the data on each single case, with the aim of helping diagnosis and prioritizing gene screening. Our system is able to provide, given a patient case, the most similar cases present in the project’s database; in this way it was possible to provide the clinician with the ability to reason on a subset of cases obtained through a human-like analogical reasoning. The peculiarity of the system is that it implements a novel method to take into account data similarity at both the data level (i.e. the values of the data) and the semantic level (i.e. the presence of similar, though not the same, symptoms).
Wiki-based system
Wiki-based systems are crucial for collaborative research and results dissemination. A first version of the wiki system has been made available to the project partners: http://www.labmedinfo.org:8123/mediawiki/index.php/Main_Page
The semantic wiki-based system has been purposely implemented for supporting the consortium members in sharing and disseminating data and knowledge. It consists of a collaborative system that is used to track project activities, share ideas and data, foster exchange of information between the investigators to support several activities of the INHERITANCE translational research project.
Scientific data that are considered within the wiki are: DCM related documents, genes and proteins. Although it is possible to add these types of data using the same manual procedure described for the organizational data, the INHERITANCE semantic wiki provides a tool to automatically acquire DCM related documents and to add to the system the genes and the proteins that are cited within these documents. The first step is to specify the file that contains (in plain text, PDF or MS Word format) the document to be acquired and analyzed; once the file has been uploaded to the wiki, the system starts the text analysis procedure in order to extract gene and protein names. When the text analysis is completed, the user is presented with a preview page (Figure 11.1) where the identified genes and proteins are highlighted. Now it is possible to specify the name of the DCM related document page and to add it to the wiki. Gene and protein pages, when created with the procedure described above, also exploit a literature retrieval tool that finds the most recent literature about the specific molecule and links it in the relative wiki page.
The interrogation tools directly available in the semantic wiki are:
• the built-in query tool of the Semantic MediaWiki,
• a summary page defined to synthetize all the project activities and participants information,
• the RelFinder browser, useful to look for relations between keywords inside the wiki and show the relations graph.
While the built-in query tool requires the user to learn a query language in order to perform an effective interrogation on the wiki pages, the latter two interrogation tools allow an easier interaction with the wiki user.
The RelFinder browser, directly accessible from the wiki’s home page, is a graphical interrogation tool that shows, in form of graphs, the information contained in the semantic wiki. Figure 11.2 for instance, shows the graph that links the DCM related documents with the genes extracted from them.
Application of the Unified Medical Language System (UMLS) based system for DCM
Finally we hypothesized that a Unified Medical Language System (UMLS) based system for Literature-Based Discovery in medicine can contribute to trance gene-phenotype association reported in the literature. Literature-Based Discovery (LBD) is a technique that can be used in translational research to connect the very sparse and huge information available in scientific publications in order to extract new knowledge. This paper presents an LBD system based on the open discovery paradigm exploiting NLP techniques and UMLS medical concepts mapping, to provide a set of tools useful to discover unknown relationships. The system has been evaluated on the problem of discovering new candidate genes potentially related to dilated cardiomyopathies (DCM), and can be used in any medical context to connect different type of concepts. The validation of the system involves reproducing the discovery of genes currently associated to DCM. Validation showed that the system is able to discover many gene-disease associations by using the literature available before their first publication in a scientific article (49,50).
REFERENCES
References 3,4 represent pertinent activities (scientific, dissemination) of partners of INHERITANCE in scientific societies (several authors are partners in INHERITANCE).
1. Serio A, Narula N, Kodama T, Favalli V, Arbustini E. Familial DCM: a clinically and genetically heterogeneous disease? Herz. 2012;37:822-829.
2. Meder B, Katus HA. Clinical and genetic aspects of hypertrophic and dilated cardiomyopathy. Internist. 2012;53:408-14, 417-8.
3. Rapezzi C, Arbustini E, Caforio AL, Charron P, Gimeno-Blanes J, Heliö T, Linhart A, Mogensen J, Pinto Y, Ristic A, Seggewiss H, Sinagra G, Tavazzi L, Elliott PM. Diagnostic work-up in cardiomyopathies: bridging the gap between clinical phenotypes and final diagnosis. A position statement from the ESC Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2013;34:1448-58.
4. Charron P, Arad M, Arbustini E, Basso C, Bilinska Z, Elliott P, Helio T, Keren A, McKenna WJ, Monserrat L, Pankuweit S, Perrot A, Rapezzi C, Ristic A, Seggewiss H, van Langen I, Tavazzi L; European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Genetic counselling and testing in cardiomyopathies: a position statement of the European Society of Cardiology Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2010;31:2715-26.
5. van Rijsingen IA, Arbustini E, Elliott P, Mogensen J, Hermans-van Ast JF , Van der Kooi AJ , van Tintelen P, van den Berg M, Pilotto A, Pasotti M, Jenkins S , Rowland C, Aslam U , Wilde A, Perrot A, Pankuweit S, Zwinderman AH, Charron P, Pinto YM. Risk Factors for Malignant Ventricular Arrhythmias in Lamin A/C Mutation Carriers: a European Cohort Study. JACC 2012;59:493-500.
6. van Rijsingen IAW, Nannenberg EA, Arbustini E, Elliott PM, Mogensen J, Hermans-van Ast JF, van der Kooi AJ, van Tintelen JP, van den Berg MP, Grasso M, Serio A, Jenkins S, Rowland C, Richard P, Wilde AAM, Perrot A, Pankuweit S, Zwinderman AH, Charron P, Christiaans I, Pinto YM. Gender-specific differences in major cardiac events and mortality in lamin A/C mutation carriers. Eur J Heart Fail. 2013;15:376-384.
7. Quarta G, Syrris P, Ashworth M, Jenkins S, Zuborne Alapi K, Morgan J, Muir A, Pantazis A, McKenna WJ, Elliott PM. Mutations in the Lamin A/C gene mimic arrhythmogenic right ventricular cardiomyopathy. Eur Heart J. 2012;33:1128-36.
8. Garcia-Pavia P, Syrris P, Salas C, Evans A, Mirelis JG, Cobo-Marcos M, Vilches C, Bornstein B, Segovia J, Alonso-Pulpon L, Elliott PM. Desmosomal protein gene mutations in patients with idiopathic dilated cardiomyopathy undergoing cardiac transplantation: a clinicopathological study. Heart. 2011;97:1744-52.
9. Diegoli M, Grasso M, Favalli V, Serio A, Gambarin FI, Klersy C, Pasotti M, Agozzino E, Scelsi L, Ferlini A, Febo O, Piccolo G, Tavazzi L, Narula J, Arbustini E. Diagnostic work-up and risk stratification in x-linked dilated cardiomyopathies caused by dystrophin defects. J Am Coll Cardiol. 2011;58:925-34.
10. Disertori M, Quintarelli S, Mazzola S, Favalli V, Narula N, Arbustini E. The need to modify patient selection to improve the benefits of implantable cardioverter-defibrillator for primary prevention of sudden death in non-ischaemic dilated cardiomyopathy. Europace. 2013;15:1693-701.
11. Serio A, Narula N, Frontera A, Isabella Gambarin F, Arbustini E. Prevalence of J-point elevation in families with sudden arrhythmic death syndrome. J Am Coll Cardiol. 2012;59:1659-60.
12. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. The MOGE(S) classification for a phenotype-genotype nomenclature of cardiomyopathy: endorsed by the World Heart Federation. J Am Coll Cardiol. 2013;62:2046-72.
13. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. The MOGE(S) classification for a phenotype-genotype nomenclature of cardiomyopathy: endorsed by the World Heart Federation. G Heart. 2013; 8:355-382.
14. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. MOGE(S) nosology in low-to-middle-income countries. doi:10.1038/nrcardio.2013.219‑c1
15. Arbustini E, Narula N, Dec GW, Reddy KS, Greenberg B, Kushwaha S, Marwick T, Pinney S, Bellazzi R, Favalli V, Kramer C, Roberts R, Zoghbi WA, Bonow R, Tavazzi L, Fuster V, Narula J. Letter JACC; in press 2014
16. Zaragoza MV, Fass J, Diegoli M, Lin D, Arbustini E.Mitochondrial DNA variant discovery and evaluation in human Cardiomyopathies through next-generation sequencing. PLoS One. 2010;5(8):e12295.
17. Zaragoza MV, Brandon MC, Diegoli M, Arbustini E, Wallace DC. Mitochondrial cardiomyopathies: how to identify candidate pathogenic mutations by mitochondrial DNA sequencing, MITOMASTER and phylogeny. Eur J Hum Genet. 2011;19:200-7.
18. Haas J, Katus HA, Meder B. Next-generation sequencing entering the clinical arena. Mol Cell Probes. 2011;25:206-11.
19. Meder B, Haas J, Keller A, Heid C, Just S, Borries A, Boisguerin V, Scharfenberger-Schmeer M, Stähler P, Beier M, Weichenhan D, Strom TM, Pfeufer A, Korn B, Katus HA, Rottbauer W. Targeted next-generation sequencing for the molecular genetic diagnostics of cardiomyopathies. Circ Cardiovasc Genet. 2011;4:110-22.
20. Frese KS, Katus HA, Meder B. Next-Generation Sequencing: From Understanding Biology to Personalized Medicine. Biology 2013, 2, 378-398.
21. Jan Haas, Karen S. Frese, Barbara Peil, Wanda Kloos, Andreas Keller, Rouven Nietsch, Zhu Feng, Sabine Müller, Elham Kayvanpour, Britta Vogel, Farbod Sedaghat-Hamedani, Wei-Keat Lim, Xiaohong Zhao, Dmitriy Fradkin, Doreen Köhler, Simon Fischer, Jennifer Franke, Sabine Marquart,Ioana Barb, Ali Amr, Philipp Ehlermann, Derliz Mereles, Tanja Weis, Andreas Kremer, Vanessa King, Emil Wirsz, Richard Isnard, Michel Komajda, Alessandra Serio, Maurizia Grasso, Petros Syrris, Eleanor Wicks, Vincent Plagnol, Luis Lopes, Tenna Gadgaard, Hans Eiskjær, Mads Jørgensen, Diego Garcia-Giustiniani, Martin Ortiz-Genga, Maria G. Crespo-Leiro, Rondal H. Lekanne Dit Deprez, Imke Christiaans, Ingrid A. van Rijsingen, Arthur A. Wilde, Anders Waldenstrom, Martino Bolognesi, Riccardo Bellazzi, Stellan Mörner, Justo Lorenzo Bermejo, Lorenzo Monserrat, Eric Villard, Jens Mogensen, Yigal M. Pinto, Philippe Charron, Perry Elliott, Eloisa Arbustini, Hugo A. Katus, Benjamin Meder. Atlas of the Clinical Genetics of Human Dilated Cardiomyopathy Eur Heart J (under evaluation) 2014.
22. Arbustini E, Grasso M. Human "nuclear" mitochondrial cardiomyopathy a novel mouse model characterizes the disease. JACC Cardiovasc Imaging. 2011;4:11-5.
23. Disertori M, Quintarelli S, Grasso M, Pilotto A, Narula N, Favalli V, Canclini C, Diegoli M, Mazzola S, Marini M, Del Greco M, Bonmassari R, Masè M, Ravelli F, Specchia C, Arbustini E. Autosomal recessive atrial dilated cardiomyopathy with standstill evolution associated with mutation of Natriuretic Peptide Precursor A. Circulation Genetics 2012; 6:27-36,.
24. Disertori M, Mazzola S, Quintarelli S, Arbustini E. [New frontiers for an ancient disease: genetics of atrial fibrillation]. G Ital Cardiol. 2013 ;14:46-54.
25. Kodama T, Serio A, Disertori M, Bronzetti G, Diegoli M, Narula N, Grasso M, Mazzola S, Arbustini E. Autosomal recessive paediatric sick sinus syndrome associated with novel compound mutations in SCN5A. Int J Cardiol. 2013;167:3078-80.
26. Villard E, Perret C, Gary F, Proust C, Dilanian G, Hengstenberg C, Ruppert V, Arbustini E, Wichter T, Germain M, Dubourg O, Tavazzi L, Aumont MC, DeGroote P, Fauchier L, Trochu JN, Gibelin P, Aupetit JF, Stark K, Erdmann J, Hetzer R, Roberts AM, Barton PJ, Regitz-Zagrosek V; Cardiogenics Consortium, Aslam U, Duboscq-Bidot L, Meyborg M, Maisch B, Madeira H, Waldenström A, Galve E, Cleland JG, Dorent R, Roizes G, Zeller T, Blankenberg S, Goodall AH, Cook S, Tregouet DA, Tiret L, Isnard R, Komajda M, Charron P, Cambien F. A genome-wide association study identifies two loci associated with heart failure due to dilated cardiomyopathy. Eur Heart J. 2011;32:1065-76.
27. Meder B, Rühle F, Weis T, Homuth G, Keller A, Franke J, Peil B, Lorenzo Bermejo J, Frese K, Huge A, Witten A, Vogel B, Haas J, Völker U, Ernst F, Teumer A, Ehlermann P, Zugck C, Friedrichs F, Kroemer H, Dörr M, Hoffmann W, Maisch B, Pankuweit S, Ruppert V, Scheffold T, Kühl U, Schultheiss HP, Kreutz R, Ertl G, Angermann C, Charron P, Villard E, Gary F, Isnard R, Komajda M, Lutz M, Meitinger T, Sinner MF, Wichmann HE, Krawczak M, Ivandic B, Weichenhan D, Gelbrich G, El-Mokhtari NE, Schreiber S, Felix SB, Hasenfuß G, Pfeufer A, Hübner N, Kääb S, Arbustini E, Rottbauer W, Frey N, Stoll M, Katus HA. A genome-wide association study identifies 6p21 as novel risk locus for dilated cardiomyopathy.Eur Heart J. 2013 Jul 12. (PMID: 23853074]
28. Vogel B, Keller A, Frese KS, Leidinger P, Sedaghat-Hamedani F, Kayvanpour E, Kloos W, Backe C, Thanaraj A, Brefort T, Beier M, Hardt S, Meese E, Katus HA, Meder B. Multivariate miRNA signatures as biomarkers for non-ischaemic systolic heart failure. Eur Heart J. 2013;34:2812-22.
29. Narula N, Favalli V, Tarantino P, Grasso M, Pilotto A, Bellazzi R, Serio A, Gambarin FI, Charron P, Meder B, Pinto Y, Elliott PM, Mogensen J, Bolognesi M, Bollati M, Arbustini E. Quantitative expression of the mutated lamin a/c gene in patients with cardiolaminopathy. J Am Coll Cardiol. 2012;60:1916-20.
30. Thanaraj A, Meder B. Genetic testing is superior to biopsy of the myocardium in cardiomyopathy - yes. Dtsch Med Wochenschr. 2013;138:598
31. Narula N, Agozzino M, Gazzoli F, Concardi M, Pagani F, Favalli V, Kodama T, Mazzola A, D'Armini AM, Arbustini E. The pathologic basis of recovery. Heart Fail Clin. 2014;10(1S):S63-74.
32. de Haas HJ, Arbustini E, Fuster V, Kramer CM, Narula J. Molecular imaging of the cardiac extracellular matrix. Circ Res. 2014;114:903-15.
33. Rasmussen TB, Palmfeldt J, Nissen PH, Magnoni R, Dalager S, Jensen UB, Kim WY, Heickendorff L, Mølgaard H, Jensen HK, Baandrup UT, Bross P,Mogensen J. Mutated desmoglein-2 proteins are incorporated into desmosomes and exhibit dominant-negative effects in arrhythmogenic right ventricular cardiomyopathy. Hum Mutat, 2013;34:697-705
34. Rasmussen TB, Hansen J, Nissen PH, Palmfeldt J, Dalager S, Jensen UB, Kim WY, Heickendorff L, Mølgaard H, Jensen HK, Sørensen KE, Baandrup UT, Bross P, Mogensen J. Protein expression studies of desmoplakin mutations in cardiomyopathy patients reveal different molecular disease mechanisms. Clin Genetics. 2013;84:20-30
35. Haas J, Frese KS, Park YJ, Keller A, Vogel B, Lindroth AM, Weichenhan D, Franke J, Fischer S, Bauer A, Marquart S, Sedaghat-Hamedani F, Kayvanpour E,Köhler D, Wolf NM, Hassel S, Nietsch R, Wieland T, Ehlermann P, Schultz JH, Dösch A, Mereles D, Hardt S, Backs J, Hoheisel JD, Plass C, Katus HA, Meder B. Alterations in cardiac DNA methylation in human dilated cardiomyopathy. EMBO Mol Med. 2013;5:413-29.
36. Griffin JL, Atherton H, Shockcor J, Atzori L. Metabolomics as a tool for cardiac research. Nat Rev Cardiol. 2011;8:630-43.
37. Heather LC, Wang X, West JA, Griffin JL. A practical guide to metabolomic profiling as a discovery tool for human heart disease. J Mol Cell Cardiol. 2013;55:2-11.
38. Kloos W, Katus HA, Meder B. Genetic cardiomyopathies. Lessons learned from humans, mice, and zebrafish. Herz. 2012;37:612-7.
39. Just S, Meder B, Berger IM, Etard C, Trano N, Patzel E, Hassel D, Marquart S, Dahme T, Vogel B, Fishman MC, Katus HA, Strähle U, Rottbauer W. The myosin-interacting protein SMYD1 is essential for sarcomere organization. J Cell Sci. 2011;124:3127-36.
40. Meder B, Huttner IG, Sedaghat-Hamedani F, Just S, Dahme T, Frese KS, Vogel B, Köhler D, Kloos W, Rudloff J, Marquart S, Katus HA, Rottbauer W. PINCH proteins regulate cardiac contractility by modulating integrin-linked kinase-protein kinase B signaling. Mol Cell Biol. 2011;31:3424-35
41. Just S, Berger IM, Meder B, Backs J, Keller A, Marquart S, Frese K, Patzel E, Rauch GJ; Tübingen 2000 Screen Consortium, Katus HA, Rottbauer W. Protein kinase D2 controls cardiac valve formation in zebrafish by regulating histone deacetylase 5 activity. Circulation. 2011;124:324-34
42. Kubben N, Voncken JW, Konings G, van Weeghel M, van den Hoogenhof MM, Gijbels M, van Erk A, Schoonderwoerd K, van den Bosch B, Dahlmans V, Calis C, Houten SM, Misteli T, Pinto YM. Post-natal myogenic and adipogenic developmental: Defects and metabolic impairment upon loss of A-type lamins. Nucleus. 2011;2:195-207
43. Bollati M, Barbiroli A, Favalli V, Arbustini E, Charron P, Bolognesi M. Structures of the lamin A/C R335W and E347K mutants: Implications for dilated cardiolaminopathies. Biochem Biophys Res Commun. 2012;418:217-21
44. EUDRA CT 2011-001292-39 https://www.clinicaltrialsregister.eu/ctr-search/search?query=DCM
45. www.cardioregister.org/inheritance
46. i2b2inheritance.com
47. Nuzzo A, Mulas F, Gabetta M, Arbustini E, Zupan B, Larizza C, Bellazzi R. Text Mining approaches for automated literature knowledge extraction and representation. Stud Health Technol Inform. 2010;160:954-8.
48. Bellazzi R, Larizza C, Gabetta M, Milani G, Bucalo M, Mulas F, Nuzzo A, Favalli V, Arbustini E. Information technology solutions to support translational research on inherited cardiomyopathies. Stud Health Technol Inform. 2011;169:907-11.
49. Larizza C, Gabetta M, Milani G, Bucalo M, Mulas F, Nuzzo A, Favalli V, Arbustini E, Bellazzi R. Supporting Translational Research on Inherited Cardiomyopathies through Information Technology. Methods Inf Med. 2013;52:137-47.
50. Gabetta M, Larizza C, Bellazzi R. A Unified Medical Language System (UMLS) based system for Literature-Based Discovery in medicine. Stud Health Technol Inform. 2013;192:412-6.
Potential Impact:
INHERITANCE addressed the Topic HEALTH-2009-2.4.2-3: Translation of basic knowledge on inherited cardiomyopathies into clinical practice. The objective of INHERITANCE was to use advanced biotechnological tools and the pooled expertise of a consortium of expert clinicians and scientists to improve the health of European citizens with familial DCM by translating basic knowledge of disease pathophysiology into practical and effective treatments. The different components of the FP7 call indications “multidisciplinary approach bringing together genomics, proteomics, structural and functional studies with clinical investigation should lead to the development of new treatments” matched our clinical and scientific needs as the key elements of the project (Figure Potential Impact 1). Specifically:
• Clinical Phenotypes in probands and relatives
• Extended (cardiac and non-cardiac) clinical characterisation (“red flags”).
• OMICS Analyses
o Molecular diagnosis of disease causing mutations.
o Identification of disease-modifying and -predisposing genes and polymorphisms by Genome Wide-Association studies
o Quantification of mutated gene transcripts in paired samples of peripheral and heart tissues.
o Proteomics and metabolomics for the identification of novel specific tissue and circulating disease markers.
• Structural: Biophysical and X-ray structural studies of mutated proteins exploring the molecular bases of disease.
• Functional: Expression of clinically malignant disease-gene mutations in animal models.
• Therapeutics: Old drugs for novel use and new molecular targets for existing drugs, as well as disease-specific guidelines for treatments.
• The bioinformatics actions planned in the project are a central tool for integrating our results and data in novel disease-specific management algorithms.
The impact of INHERITANCE ranged from models of social-health innovation to the basic knowledge of the mechanisms of disease in DCM as well as biotechnological innovation for genetic testing that makes sustainable and reproducible the screening of multiple disease-genes and web-assisted dissemination and free of charge tools such as application for smartphones and tablets.
MODELS OF CARE
Customized family-centered model of care for heritable cardiovascular diseases: the family-centered model
In INHERITANCE we have established a patient/family centered model of care for familial diseases. The model replaces the current discipline-based model of care (Figure Potential Impact 2). In familial DCM more affected members show the disease at different stages. The family member that brings relatives to the attention is the proband who usually seeks medical attention because of symptoms. The simple genetic counselling provides first essential information about the family history. However, the key step for familial DCM is the clinical non-invasive family screening that may identify affected asymptomatic family members, or relatives in which the echocardiographic study demonstrates early morpho-functional changes not fulfilling the WHO criteria for the diagnosis of DCM. The addition of genetic testing and the identification of the mutation that causes the disease provide a further powerful tool for the diagnosis; identify carriers that are still clinically healthy; inform about the gene- and mutation-specific phenotype. Based on the evidence that more than 80 disease genes may cause a similar DCM phenotype and that each disease-gene may cause not only the DCM but also extra-cardiac traits such as hearing loss, myopathy, encephalo-myopathies, palpebral ptosis, gastrointestinal problems, renal disease, etc., the interdisciplinary evaluation is an obligate strategy that completes the clinical characterization of the disease, thus favouring a selective approach to genetic testing.
Customized Disease-Specific Innovation
In the family centered model of care, disease-specific diagnostic work-ups are designed and organized for offering selected appropriate evaluations of the DCM associated with defects of different genes. The multidisciplinary diagnostic work-up is planned by the referral center that organizes all specialist clinical evaluations, which may vary according to the phenotype presented by the proband; in this novel organization strategy, new professionals (diagnostic work-up assistants) have been educated to interact with the referral physicians (geneticists or cardiologists or both) that provide the plans for the evaluation of the probands and for cascade family screening. The clinical network with family doctors facilitates the family context and the disease setting. Information for probands and relatives is a key step: although written or web-supported information is useful and reaches a larger proportion of potential beneficiaries, the direct communication either in family counselling or in meetings organized for families provides a unique occasion for learning about emerging social-health needs that are related with the disease in the families.
Family care
Once clinical and genetic cascade family screening is completed, the picture of the family may demonstrate:
1. affected, symptomatic probands and relatives, fulfilling diagnostic criteria of cardiomyopathy
2. affected, asymptomatic relatives that were unaware of their disease before the family screening
3. family members that only show instrumental (echocardiographic or ECG-graphic) abnormalities suggestive for an early diagnosis
4. health carriers of the mutation that causes the disease in the family.
The above four groups of patients/individuals at risk need tailored monitoring programs that are planned according to pheno-genotype, family history, comorbidities and any other individual need that may require specific plans. The concept of the model is that of taking the responsibility of family health programs, thus limiting the discomfort and the costs for the family itself and supporting educational actions for the family members during the counselling that concludes each monitoring occasion.
Collective educational activities
Regular meetings for patients and families are a unique occasion for reciprocal exchange of information about the disease (how to live with a familial DCM), advancements in care strategies and translational research results. These meetings are usually organized with Charities and Associations of patients supporting the activities of the centres dedicated to familial cardiomyopathies.
Efforts to translate the basic science of genetic forms of DCM into clinical practice are now largely impacting on patients and families. Before INHERITANCE, the integration of clinical, genetic and other biological data into coherent disease specific paradigms was limited to sporadic meeting occasions, mostly dedicated to physicians, cardiologists in particular. After INHERITANCE, the extension of information to specialists other than cardiologists, as well as family doctors, highly contributed to increase the awareness of the DCM as familial disease and as a condition in which family screening, both clinical and genetic, is the crucial key strategy for family health programs.
Sustainable genetic testing: lower costs having expedite results
The project had the extraordinary merit of developed not only knowledge but also expertise on how to organize, implement and exploit genetic testing in each centre participating in the project. The strategy was that of centralizing the NGS of a large EU population and, in parallel, implementing local technologies and expertise and exploiting existing local resources, both human and technological. Before INHERITANCE, most centres were equipped for performing Sanger-based sequencing, a costly and time-consuming method. After INHERITANCE, all centres run their own NGS-based tests. At present, fast, low-cost sustainable tools for diagnosis and disease monitoring are available, and Sanger-based testing is limited to a confirmatory role for mutations identified by NGS.
SCIENTIFIC IMPACT
INHERITANCE generated a scientific legacy by creating a multidisciplinary transnational platform comprising disease-oriented clinical assessments and non-invasive imaging; genomic studies that identified causative mutations as well as predisposition and modifier genes; measurements of metabolites and cardioproteomes; information on structure of mutated proteins; functional assessments in animal models; bioinformatics tools that integrate the different scientific activities with clinical investigation in order to develop new treatment strategies. With 46 publications in peer reviewed journals (Impact factor calculated on ISI 2012=271.174) one PATENT (BIO-10917), the web site for Cardioregister (www.cardioregister.org/inheritance) the e-CRF for clinical trials, the Wiki-based system (http://www.labmedinfo.org:8123/mediawiki/index.php/Main_Page) and the new Unified Medical Language System (UMLS) based system for DCM (incorporated in the Wiki System) and the free of charge app for the novel classification of cardiomyopathies (http://moges.biomeris.com/moges.html) the project can be considered as example of good use of public funds for research throughout Europe.
The scientific results impacted on:
- The definition of the natural history of the different subtypes of DCM, based on family screening and re-screening studies.
- The elucidation of genotype-phenotype correlations with impact on diagnostic, prognostic and outcome prediction.
- The elucidation of pathophysiologic mechanisms of myocardial damage and dysfunction in disease-specific experimental models.
- The identification of novel disease genes: before INHERITANCE, the disease genes were about 40; after INHERITANCE, the list includes more than 80 genes, all included in the large-scale NGS study carried out in the project.
- The identification of genes that modify DCM evolution and severity.
- The development of transcriptomic tests on peripheral blood RNA that can guide genetic testing in the clinical setting.
- The characterisation of disease-specific cardioproteome and peripheral metabolome to identify markers for diagnosis and clinical monitoring.
- The crystallisation of key epitopes of mutated vs. wild-type proteins for future elucidation of the effects of mutations in the 3D structure of the mutated proteins and future disease-specific pharmacological research.
- Pre-clinical (animal models and in vitro testing on human monocytes) evaluation of novel treatment strategies, of existing drugs and novel molecular therapies, to generate the basis for future experimental validation and investment.
- The generation of a web-assisted database that also guides the clinical work–up of the different types of DCM.
- The generation of a bio-informatics Wiki-like tool initially for the consortium and thereafter for the scientific community.
IMPACT ON FAMILIES
The aim of INHERITANCE is to produce recommendations for the management of patients and families with DCM that can be applied in all clinical settings. Specific innovations will include:
- Promotion of routine family screening in the diagnostic work-up of patients with DCM.
- Development of phenotype guided genetic testing protocols.
- Development of treatment recommendations in pre-symptomatic mutation carriers with borderline abnormalities in left ventricular size and function.
- Generation of the evidence that current guidelines for the implantation of cardiac defibrillators in high-risk subgroups of DCM for the prevention of sudden death, such as cardiolaminopathies, must be revised.
- The generation, implementation and exploitation of simple bioinformatics tools that assist disease management.
- Establishment of the family-oriented model of care for DCM and other inherited cardiovascular diseases in European healthcare systems.
TECHNICAL IMPACT
INHERITANCE significantly contributed to the improvement of quality of health in Europe. High impact from the bio-informatics activities (developed by P6 and P10 in WP10 and 11 respectively) resulted in a user-friendly database and the knowledge management system. A commercially exploitable product from this work is a database that guides clinical and genetic diagnostic work-up (and the e-CRF, see scientific property). Free on the Web is the app that supports the novel classification/nosology system for cardiomyopathies (MOGE(S)). The app constitutes an innovative fast approach to the daily management of phenotype and genotype data.
SOCIO-ECONOMIC RELEVANCE
A population of about 40,000 European citizens is affected by DCM in the participating countries. The socio-economic consequences for patients and families can be profound, as many develop the disease at a young age. The costs to healthcare systems are correspondingly high as patients suffer many years of prolonged ill health and as a result consume valuable resources. Moreover, they are frequently too sick to contribute financially to their local communities and to society as a whole. Data generated by INHERITANCE first provided national health organisations with a novel model of care that can be used to develop family oriented services offering tailored treatments to patients with DCM and preventing the development of disease in relatives through the application of cascade clinical and genetic screening as well as early medical treatments.
HEALTH POLICY RELEVANCE
INHERITANCE contributed to validate a prototype for translational research impacting clinical care through the integration of conventional clinical tools with cutting edge genetic, structural, functional, metabolomic and proteomic technologies. The subject of the study, familial DCM, is not only an important healthcare problem in Europe as well as throughout the world countries, but is also a paradigm for other inherited diseases of similar or greater relevance to the health of European citizens. Fulfillment of the aims of INHERITANCE is providing policy makers with data that can be used to develop a healthcare model for inherited heart diseases to be applied to centres across the continent. The long-term result will be more efficient in targeting of resources and improved health for patients and families.
USE OF DISSEMINATION OF FOREGROUND
EU and National Scientific Societies
The collaboration implemented in INHERITANCE generated partnerships for future research and stimulated similar scientific programs in all European countries. The results of the project have been disseminated to the wider cardiologic community and European policy makers through national societies and the European Society of Cardiology. With more than 40,000 European cardiologists that are members of the ESC and participate in the ESC annual meeting, the scientific products of INHERITANCE have been disseminated. The project’s output was also disseminated at a national level via the annual meeting of the national societies of cardiology represented by the partners in INHERITANCE.
World Heart Federation
A significant achievement was the collaboration with the WHF and the endorsement of a novel classification/nosology description system for cardiomyopathies. While progressing with research in INHERITANCE we realised that genetic information was exponentially increasing and that cardiomyopathies were being described based on the name of the mutated gene/protein. A genetic terminology entered the scientific literature with diagnoses termed as desmosomalopathy, cytoskeletalopathy, sarcomyopathy, channelopathy, cardiodystrophinopathy, cardiolaminopathy, zaspopathy, myotilinopathy, dystrophinopathy, A-B crystallinopathy, desminopathy, caveolinopathy, calpainopathy, sarcoglycanopathy, dysferlinopathy, merosinopathy, emerinopathy etc. These terms do not describe the phenotype but rather specifically describe the cause (the disease gene). From a clinical point of view, cardiologists cannot diagnose a DCM as myotilinopathy or calpainopathy. In addition, the increasing evidence of overlapping phenotypes (DCM and ARVC, HCM and RCM, DCM evolving through DCM) and complex genotypes (> than one mutation) are contributing to elucidate the old concept of phenotype variability, incomplete penetrance. These concepts are now being explained with complex genetic make-up in members of the same families. To mantain the morpho-functional classification (HCM, RCM, DCM, ARVC) but also to inform about familial and non-familial cardiomyopathy and disease gene and mutation, and while waiting for the completion of the discovery of all disease genes, a novel nosology descriptor was generated. The acronym of this novel nosology is MOGE(S) that includes morpho-functional characteristics (M); organ involvement (O); genetic or familial inheritance patterns (G); and an explicit etiological annotation (E) with details of a genetic defect or underlying disease/cause along with functional status information (S) as per the American College of Cardiology/American Heart Association (ACC/AHA) (A to D) stage and New York Heart Association (NYHA) (I to IV) functional classes. The novel nosology is supported by a free web app at http://moges.biomeris.com/moges.html that can be easily used with smartphone and tablets. Proposed in December 2013, MOGES is already a widely circulated paper, object of comments, short reports on web sites (Figure Potential Impact 3). Modelled on the TNM system for cancer, MOGES is a flexible system that does not introduce novel cardiomyopathies but allows a diagnostic description including both morpho-functional phenotype (fundamental) and aetiology, both genetic and non-genetic.
Patients and families are also a key target in the INHERITANCE dissemination strategy. At present, P1 organises an annual meeting for patients and families dedicated to cardiomyopathies; P2 is vice-chairman of a major national cardiomyopathy charity and contributes to educational sessions for patients and carers across the UK. Similar activities will be implemented in all countries represented in the Consortium by other partners. P5 collaborates with a French charity that works for promoting knowledge on cardiomyopathies and supports patients in France.
Dissemination and exploitation of translational programs have been addressed to national health government institutions. In public health care systems, where citizens contribute to their health care, the upgrading of the national health programs takes years to introduce novel plans and dedicated investments. Inherited DCM has been around for a long time, a neglected disease in national health plans; for instance, in several European countries (including Italy), DCM is not even listed among rare diseases, even when genetically determined. Partners of INHERITANCE are promoting inherited DCM as a rare disease, to provide patients and families with the same rights of other groups of formally recognised rare diseases.
INHERITANCE DISSEMINATION MEETING (STRESA MAY 10,11, 2014 ITALY)
We have organised an international meeting specifically aimed at disseminating the results of INHERITANCE and promoting adhesion to INHERITANCE models of family-centred care in other countries where centres for inherited cardiomyopathies are just taking the first steps. In this meeting for the first time we will have a session dedicated to charities active in 3 countries (UK, FR, and IT). When looking for further similar charities in other countries we discovered that there are no similar initiatives in the other countries participating to the project. We will take advantage of this unique occasion for encouraging and helping these 3 charities to reciprocally interact, explore the needs in different EU countries, as well as the quality and the standards of care for cardiomyopathies. We are aware that fund restriction is now a major limitation for our research and that solidarity and collaboration with charities is a unique occasion for making health programs uniform though Europe. We will also ask charities to collaborate with researchers in writing a document highlighting the view points of patients and families in which 50% of the members are affected by ominous cardiac diseases in young age. A draft of the document is ready for the discussion in the meeting.
With this meeting, we also hope to gain more attention from national and regional health institutions with respect to cardiomyopathies, DCM in particular. The meeting in STRESA will be preceded by a masterclass on inherited cardiomyopathies, open to all professionals working in this setting, free of charge. The masterclass will be held in a meeting centre kindly provided by the Lombardia Region and will be endorsed by “Ordine dei Medici di Milano and Pavia”, MAGICA Onlus, Lombardia Region, Ministry of Health.
The meeting has been planned with the double aim of reporting and discussing the results of INHERITANCE but also opening a door on incoming scientific discoveries in molecular imaging and expansion of our collaborations with more centres and countries in Europe.
List of Websites:
The public website of INHERITANCE project is www.inheritanceproject.eu.
The Principal Investigator of the project is Eloisa Arbustini, MD, Center for Inherited Cardiovascular Diseases, IRCCS Foundation Policlinico San Matteo, Pavia ( Phone 0039 0382 501486, fax 0039 0382 501893, mail info.cardiomiopatie@smatteo.pv.it).
