Skip to main content

Systems biology for the functional validation of genetic determinants of skeletal diseases

Final Report Summary - SYBIL (Systems biology for the functional validation of genetic determinants of skeletal diseases)

Executive Summary:
The SYBIL project (systems biology for the functional validation of genetic determinants of skeletal diseases) presents a paradigm of how a multidisciplinary group of scientists, clinicians, and technologists can work as an integrated consortium to achive critical mass for delivering new scientific knowledge in a large and disperate group of rare and common diseases.

In this context, and in the first instance, SYBIL employed a pipeline approach to identify and model genetic variants implicated in skeletal diseases using cell and animal models. Over 100 cell, 25 mouse and 15 zebrafish models were generated and validated. This represents a significant undertaking that generated a diverse portfolio of novel and relevant model systems to study skeletal disease mechanisms in detail. The increasing complexity of model systems developed by SYBIL allowed basic disease mechanisms to be studied and validated quickly in cell models prior to generating relevant animal (primarily mice) models to provide in vivo relevance through studying appropriate cells and tissues and ultimately the pre-clinical data required for translational exploitation. Mouse models are recognised as the vertebrate model system of choice for studying complex tissues such as bone and cartilage (eg. growth plate, articular cartilage) and the in-depth phenotyping of these mouse models have identified new disease mechanisms and reinforced known mechanisms, which would otherwise not have been possible.

The availability of a broad portfolio of validated cell and animal models of skeletal disease supported the extensive use of omics-based technologies to determine in detail the transcriptomic, proteomic and metabolomic profiles induced through the expression of known genetic variants. Transcriptomic profiles were generated for a range of different cell types (e.g. chondrocytes, osteoblasts, osteoclasts) that were supplemental with proteomics. Importantly, metabolomic profiles were generated for a number of cell and animal models of disease a represents novel approach for skeletal disease investigation. A significant achievement of SYBIL was to apply for the first time a global or ‘systems’ approach to the analysis of skeletal pathologies, in which the full extend of genetic defects would be revealed following an unbiased approach. The availability in SYBIL of large ‘omics’ datasets from a broad portfolio of validated models promoted the application of this systems biology approach.

A longer-term expectation of SYBIL was the translation of new knowledge acquired about disease mechanisms into novel therapeutic targets and related biomarkers. In this context SYBIL has explored a number of new therapeutic avenues for rare bone diseases ranging from small molecules to pepetides and enzymes. One highlight of SYBIL is the repurposing of carbamazepine (cbz) for the treatment of metaphyseal chondrodroplasia (MCDS), which encompassed cell and mouse pre-clinial studies that promoted orphan drug designation for use of cbz in MCDS. Cbz is now being tested in MCDS patients through a Horizon 2020 funded clinical trial (MCDS-Therapy: 754825).

In summary, SYBIL has successfully generated, validate and studied in depth a comprehensive collection of cellular and animals models of skeletal disease and have successful enhanced omics technologies through systems biology to generate new knowledge on disease mechanisms. Moreover, by targeting ER stress through drug repourposing SYBIL has transitioned from bench to bedside to deliver a potential new therapy for a rare bone disease. The knowledge generated in SYBIL will ultimately translate to more new therapies for this large unmet medical need in genetic bone diseases.
Project Context and Objectives:
WP1: The overall goal of SYBIL was an in-depth and advanced analysis of various genetic determinants of common and rare skeletal diseases (CSDs and RSDs) in order to better understand the underlying disease processes and to deliver new and validated disease biomarkers and therapeutic targets. To achieve this goal, SYBIL decided in the first instance to develop relevant cellular and animal models for in-depth genetic, molecular, biochemical and metabolic phenotyping. The logical first step of SYBIL was to select those genetic determinants for further study in the SYBIL pipeline. This cataloguing, prioritizing and selecting procedure was the main objective of WP1, which served as the starting point for SYBIL. The selection was based in the first instance on variants already identified by SYBIL partners. Many SYBIL partners have been studying for many years RSDs and CSDs and have established the pathogenicity of many variants for a large proportion of genes/mutations. In addition, SYBIL also considered a list of variants, including genes either recently identified by SYBIL partners, or variants reported in the recent literature with relevance to RSDs and CSDs. A specific and representative committee was established to prioritize the variants for functional validation. The committee was composed of 10 partners (UNEW, UNIVAQ, UNIMAN, UNIPV, INSERM, UA, PG, UKE, EVCYT, CNR). The prioritisation of genetics variants responsible for RSDs was based on preliminary experimental analyses. For CSDs caused by common or rare variants, functional characterisations were also considered. For each variant, we collected data regarding gene function and signalling pathways and predictions of the effect of the disease-causing variant.

WP2: WP2 was focused on the generation of various cell lines expressing genetic variants identified in WP1. To this end, SYBIL generated and characterised surrogate cell lines, primary cell lines derived from osteoblasts, osteoclasts and chondrocytes from mouse and human tissues. In parallel, SYBIL generated and characterized induced pluripotent stem cells (iPSC) from various human and murine tissues and differentiated them into various cell line progenitors ranging from mesenchymal to hematopoietic progenitors. In particular, SYBIL also set up and optimized novel protocols to obtain iPS clones from various human tissues including urine. Of note, SYBIL dedicated particular effort in comparing different procedures to reprogram cells by testing different approaches (integrating and non-integrating virus, plasmid and transdifferentiation approach). In parallel, SYBIL optimized protocols to differentiate human and murine iPSC and murine ES into various tissues both in mouse and in human. Generation of this platform was meant to constitute a prerequisite to study by OMICS approaches the molecular and cellular mechanisms in various preclinical models of bone diseases (WP5) and for the development of novel therapeutic strategies ranging from pharmacological approaches using new small molecules to lentiviral vectors or gene targeting approaches mediated by artificial nucleases (WP8). Finally, the creation of a data repository reporting in detail the cell lines and integrated with bioinformatics analysis was undertaken. Overall, this WP achieved the general objective to provide an important platform of more than 100 cellular lines representing various forms of skeletal disease, thus representing a suitable tool for the study of pathogenetic mechanisms as well as for testing novel therapeutic strategies.

WP3: A detailed knowledge of the molecular mechanisms that underlie rare and common skeletal diseases (RSDs and CSDs) is a prerequisite to develop diagnostic tools and therapeutic targets. The generation of relevant animal models for RSDs and CSDs in WP3 was a crucial integral interface of the SYBIL consortium which provided novel insights into the pathogenic mechanisms underlying skeletal defects in vivo. Knowledge for the selection of relevant models was obtained by WP1/2. The animal models were then indispensable for other work packages: WP4 where they were morphologically phenotyped at large scale, WP5 where transcriptome, epigenome, proteome, metabolome and secretome data were generated, WP6 where the data were analyzed by systems biology, WP7 where biomarkers were identified and validated that then finally in WP8 lead to the preliminary validation of therapeutic targets. Novel gene editing technologies (initially TALEN targeting, later CRISPR/Cas9 editing) were evaluated and applied, in addition to the well-established conventional methods. Furthermore, the reverse erythromycin system for the reversible switch on/off of gene expression was evaluated for its application in mouse models.

WP4: The objective of WP4 was to provide deep phenotyping of models of RSDs and CSDs. WP4 had the ambitious goal to integrate the research performed in most of the other WPs. This goal has been fully achieved, receiving information and materials from WP1-WP3 and providing relevant and well characterised models to WP5-WP8. Especially important was the integration with WP7 and WP8, through which several models phenotyped in WP4 were the subject of investigations aimed at identifying new biomarkers (WP7) and new therapies (WP8) for RSDs and CSDs. Both WPs were successful, thus strengthening the relevance and validity of the studies performed in WP4. Deep-phenotyping was obtained by a number of biophysical, biochemical and morphological approaches supported by SOPs and relevant ontologies. The work performed in WP4 used the ‘Omics Knowledge Factory’ developed in WP5, generating -omics profiles of skeletal biology in normal and pathological conditions.

WP5: WP5 had the overall aim to support other work packages with Omics data, ontologies for systematic interpretation and correlation of phenotype, histology, and molecular data, and a knowledgebase for data organisation. Complementary to the more hypothesis- and phenotype-driven analysis within WP4 the cellular and animal models for RSDs and CSDs provided by WP1 to WP3 were analysed by WP5 using systematic high-throughput approaches comprising transcriptome, proteome, metabolome, and epigenetic profiling. In parallel, software tools for NGS data analysis and experimental design were developed. These data were further analysed by bioinformatics and modelling within WP6. From there the results were fed into WP4 for improvement of mechanistic understanding, to WP7 for selection of biomarkers, and WP8 for novel treatment strategies.

WP6: The analysis of large datasets obtained from ‘omics’ technologies show that perturbations linked to CSDs and RSDs are rarely limited to a single gene, but generally extend far beyond the immediate neighbourhood of genes or pathways that are linked to known gene defects. The aim of WP6 was to apply for the first time a global or ‘systems’ approach to the analysis of skeletal pathologies, in which the full extend of these perturbations would be revealed following an unbiased approach. Our main analytical tool was the development of network-based models, which combined the known molecular interactions between genes and proteins with the perturbations observed by experimental ‘omics’ analysis, in different skeletal tissues and disease models. An additional objective of this work package was to develop new algorithmic and software tools to deepen and streamline these analyses by bioinformaticians, and also to make the results more easily accessible to biologists and clinical researchers. This has involved the expansion of modelling technology previously developed for cancer research towards skeletal diseases, the development of new systems biology methods for the identification of dysregulated pathways, subnetworks, regulatory factors and drug response overlaps in ‘omics’ datasets, and the creation of an integrated knowledgebase and analysis pipeline. All these developments were informed by large datasets generated from WP5 and legacy data and informed the identification of new biomarkers (WP7) and prediction of novel disease mechanisms and therapeutic targets (WP8).

WP7: The quantification of disease-specific biomarkers in biological fluids, such as serum or urine, is extremely useful for disease diagnosis and treatment monitoring. Although some biomarkers have been established to monitor skeletal remodeling and mineral homeostasis, the vast majority of skeletal disorders are currently diagnosed by other means, and their progression and potential treatment response is not controlled at the level of biomarker quantification. This does not only apply to several RSDs being studied in the SYBIL programme, but also to some CSDs, in particular osteoarthritis and diabetes-associated fractures. The aim of WP7 was therefore to identify peptide and non-peptide candidate biomarkers of RSDs and CSDs, to develop, if necessary, specific assays for their quantification, and to evaluate the diagnostic and prognostic value of a given candidate biomarker in both disease models and patients. In the context of the overall project WP7 was mostly fed by the generation of data within WP5 and WP6, where different Omics strategies were paired with Systems Biology approaches to identify candidate biomarkers for specific skeletal disorders, taking advantage of the models generated in WP2 and WP3 and phenotyped in WP4. The quantification of identified biomarkers by specific assays directly fed into WP4, as a part of the deep phenotyping approach, and into WP8, to monitor treatment effects at a biomarker level. While this workflow was most relevant for RSD-specific biomarkers, SYBIL also aimed at additional strategies (performed in WP5/6) to identify potential biomarkers for CSDs.

WP8: Skeletal disorders are difficult diseases to treat and, quite often, the pathological process begins before birth and can affect the whole skeleton. There are currently few specific therapeutic interventions in skeletal disorders; therefore, the generation of reliable and effective treatments requires novel and innovative research that can identify therapeutic targets to prevent, halt or modify skeletal disease progression. The main goal of the SYBIL consortium has been to gain knowledge of disease mechanisms and age-related changes and to deliver new and validated therapeutic targets. Specific disease mechanisms have been identified for several skeletal disorders through deep phenotyping activity of WP4 on cellular and animal models developed in WP2 and WP3 respectively, transcriptomics and proteomics data sets generated in WP5 and their integration in WP6 applying bioinformatics approaches. These activities have provided the necessary data resource and technological platforms for the development of innovative therapeutic approaches. Thus, in WP8 molecules targeting specific pathways and biological processes involved in skeletal disorders have been ranked on the basis of their effects, specificity and toxicity. The most effective have been further studied in animal models in order to optimize administration route, dosage, time and age of treatment. In addition, the experimental approaches aimed at assessing the efficacy of the treatment and the use of non-invasive biomarkers developed in WP7 have been validated during and at the end of the treatment. This work has provided POP data for the development of new therapeutic approaches and future personalized treatments of skeletal disorders.

WP9: The objectives of work package 9 were to facilitate dissemination of SYBIL work and knowledge to a variety of stakeholders, to facilitate collaborations, interactions and data sharing between the SYBIL researchers, to train and support young SYBIL researchers and to identify exploitable SYBIL technological developments and knowledge that could be continued/supported past the project. WP9 interacted with all the work packages on the project, gathered data from all partners and facilitated research exchanges and training that fed into several work packages of SYBIL. Publicly accessible website was built by UNEW and CERTUS and used as a primary means of interaction with SYBIL stakeholders. A password protected internal portal was also developed, in order to allow cataloguing of variants and models, sequencing requests, reporting and to facilitate collaboration and development of the system biology pipelines within SYBIL consortium. The disorders and variants prioritised in WP1 were published in lay language together with educational materials on the “Research” page of the publicly available SYBIL website in order to promote visibility of the SYBIL objectives to all stakeholders. The cellular and animal models generated and phenotyped in WP2-WP4 were presented at national and international meetings and SYBIL social media platforms were used to send out “soundbites” to the relevant stakeholders. Several of these models were established and/or phenotyped as part of WP9 exchange visits whereby a young researcher (PhD student or PostDoc) from one partner organisation would visit another SYBIL centre to learn new techniques and transfer knowledge within the SYBIL consortium. The interaction between the SYBIL young researchers was also facilitated by a social media-like discussion board website. Moreover, WP9 facilitated learning through webinars and project meetings that fed into the other work packages directly. SYBIL internal portal was used to catalogue all the variants and models and facilitate exchange of information and models between SYBIL researchers. The internal portal was used to develop the “Omics” knowledge factory (WP5) and to deposit SYBIL transcriptomics and proteomics data that allowed SYBIL scientists to build Systems Biology pipelines for streamlined and standardised data analysis (WP6). The pipelines were integrated into the internal portal. Ontologies relevant to skeletal biology were developed by CERTUS and gold standard images and standard operating procedures (SOPs) were published on the publicly available website for the benefit of the wider research community. Moreover, a WP9 exchange between CERTUS and UNIMAN led to development of publicly available Systems Biology tool, PhenomeExpress. SYBIL work from WP2-WP6 as well as the biomarkers identified in WP7 and potential treatment regimens studied under WP8 were published in highly cited journals and presented at national and international meetings. The biomarkers and drug treatments also led to 2 patents, 6 patent applications and a clinical trial, and several exploitable outcomes of SYBIL (drug treatments, clinical trials, models of disease, technique refinements and potential for spin-off companies) were identified at the end of the project. SYBIL also led to several national and international grant applications.

WP10: The objectives of this WP10 were: (i) To ensure that the work and tasks are completed on time, within budget and according to high quality standards; (ii) To ensure that reporting is performed on a periodic basis, in the most efficient and pragmatic way, according to Commission guidelines to provide all consortium members with all important and high-impact information that may influence the outcome of the project; (iii) To define procedures for handling the industrial property rights; (iv) To create an IPR database for results and knowledge generated from the research activities; (v) To promote women participation in skeletal diseases research projects and (vi) To develop consensus view on the ethical issues surrounding the use of model organisms in bone diseases research. The coordinating team was assisted by a managerial partner (FINOVATIS), ensuring professional management of this complex and large-scale European research project. WP10 was successful, meeting all objectives as planned.
Project Results:
WP1: The aim of the WP1 was to select genetic variants for functional validation in cellular and animal models. The prioritisation of genetic variants was an evolving process as new genes & mutations were being identified; however, a carefully selected cohort of variants was defined at the start of the project. These genetic variants were representative of the diverse range of gene products and genetic pathways that are fundamental for skeletal development and are associated with the relevant pathologies. A secure, web-based and user-friendly database (https://secure.sybil-fp7.eu/sp/variants) listing a total of more than 100 unique genetic variants in 50 different genes, was developed. In addition, two new databases were integrated into the existing system, namely those to record Research Activity and OMIC datasets.

It was decided to populate this database with both recently described novel variants and genetic determinants of known pathogenicity, which have already been identified by SYBIL partners, but have only been partly studied with limited functional validation. In addition, we also ensured that the database represented both RSDs and CSDs and that the spectrum of variants was a good representation of the molecular complexity and heterogeneity underlying skeletal disorders. To implement these databases systematically an initial version of a complete SYBIL data model was developed. This model inter-linked related elements of the system and connected back to the variant database. The SYBIL Variant database allowed each partner to create a simple record of their research work. This database was considered as a reliable resource for the proper selection of genetic variants reflecting the clinical and molecular heterogeneity observed in both RSDs and CSDs.

A specific and representative committee was established to prioritize the variants for functional validation. The committee was composed of 10 partners (UNEW, UNIVAQ, UNIMAN, UNIPV, INSERM, UA, PG, UKE, EVCYT, CNR). Criteria for variant selection included gene function, involvement in important signalling pathways, scientific innovation, involvement in RSDs or CSDs and potential disease-causing effect. Using these criteria specific variants were selected by the committee for further validation in mouse models, which included: COL10A1 p.Tyr632Ter (UNIMAN) ; WNT1 p.Gly177Cys (UKE) ; ACAN p.Asp2276Asn (UNEW) ; FGFR3 p.Asn540Lys (INSERM) ; LRP4 p.Arg1170Glu (UA) ; IMPAD p.Asp177Asn (UNIPV) ; CLCN7 p.Gly215Arg (UNIVAQ), MATN3 pTyr303Met (UK), miR-31 overexpression (EVCYT), FBN2 Phe1670Cys (UA), Slc10A7 Cond . KO (INSERM), Pop1 Cond KO (UKL-FR), Wnt1 Arg235Trp (UKE), Sost Val21Met (CHARITE), P4hb Tyr393Cys (UNIVAQ), Kif22 Pro143Lys (UNEW), Ano5 Thr491Phe (UKE), Trpv4 Arg594His (UKL-FR), Col2A1 Gly136Glu (UKE), Pls3 Cond. KO (UKE), Ifitm5 C-14T (UKK), NBAS 6237-3C>G (CNR), COL9A3 Exon 3 deletion (UNEW). Those mouse models were further studied in the other WPs.

WP2: The establishment of a portfolio of primary cell lines from animal models and patients has represented a crucial step for the subsequent molecular and cellular dissection of the pathophysiology of CSD and RSD. SYBIL has focused the efforts in generating and collecting cell lines from a number of disease models including the following:

Zebrafish models:
Zebrafish models carrying defects in COMP, P3h1, Crta,Pp1b and Tmem 38b

Human cell lines:
1. Surrogate (HeLa, HEK203T, HEK293-EBNA, CHO, ATDC5) human cell lines expressing genomic variants of molecules responsible for the pathogenesis of the following human bone disorders: Schmid Metaphyseal Chondrodysplasia (caused by mutations in COL10A1); Hyperostosis Cranialis Interna (caused by mutations in Zip14); Cenani-Lenz syndrome (mutations in LRP4); Sclerostosis (mutations in LRP4); Chondrodysplasia (mutations in MATN3); MED-PSACH disease (mutations in COMP); Type II Collagenopathy (mutations in Collagen II); Gerodermia Osteodysplasia (mutations in Gorab); Osteogenesis imperfecta type XIV (TMEM38B knock-out human Fetal Osteoblasts 1.19 cell line); Autosomal dominant Osteopetrosis (mutations in LRP5); Autosomal dominant Osteopetrosis (mutations in CLCN7); Acrofrontofacionasal Dysostosis type 1 (mutations in NBAS gene)

2. Human induced pluripotent stem cells (iPS) derived from various tissue sources as in vitro models of the following diseases: MED-PSACH disease (mutations in COMP); Chondrodysplasia (mutations in MATN3); Autosomal recessive osteopetrosis (mutations in TCIRG1); Autosomal dominant osteopetrosis (mutation in CLN7). In parallel iPS from various tissues of normal donors have bene generated as internal normal controls.


Murine cell lines:
1. Surrogate cell lines (RAW264.7 osteoclast-like line) have been generated to analyse the effect of novel mutations for the following diseases: Autosomal dominant osteopetrosis (mutations in CLN7).

2. Primary murine cell lines have been obtained from chondrocytes, fibroblasts, osteoblasts from the following diseases: Osteogenesis imperfecta (osteoblast, chondrocyte, fibroblast cell lines obtained from the Brtl mouse model carrying defect in col1a1); Osteogenesis imperfecta (osteoblast cell line obtained from the Amish mouse); Prolidase deficiency (osteoblast, chondrocyte, fibroblast cell lines obtained from the dal/dal mouse model); Diastrophic dysplasia (osteoblast, chondrocyte, fibroblast cell lines obtained from the Dtd mouse carrying mutation in Slc6a2gene); Gerodermia osteodysplasia (osteoblast cell lines obtained from Gorab knock out mouse model); Chondrodysplasia (fibroblasts obtained from the Impad1 mouse for functional studies of the gPAPP enzyme); Autosomal dominant osteopetrosis (osteoclasts obtained from the Clcn7G213R KI mouse).

3. Murine induced pluripotent stem cells (iPS) derived from various tissue sources as in vitro models of the following diseases: Autosomal recessive osteopetrosis derived from murine embryonic fibroblasts carrying defects in TCIRG1 gene; Autosomal recessive osteopetrosis derived from murine skin fibroblasts carrying defects in RANKL gene.

4. Twenty cellular lines carrying deletion, inversion in non-coding genomic regions encompassing Epha4, Sox9 and Foxg1 genes.

iPSC generation and differentiation protocols:
In the framework of WP2, SYBIL tested several reprogramming approaches (Sendai virus, lentiviral vectors carrying Oct4, Sox2, Klf4 and Myc reprogramming genes and non-integrating vectors such as mRNA nucleofection) to obtain iPSC from human and mouse tissues. Comparison of different techniques has allowed to identify Sendai virus as the most reliable and suitable method for the generation of iPSC. In parallel, SYBIL used various tissue sources, including fibroblasts obtained from skin, ligament, tendon, articular chondrocytes, peripheral blood. Particular effort was dedicated to the optimisation of a protocol for the generation of iPSC from cells obtained from the urine sediment. This novel source of cells allows speeding up and implementing the collection of specimens from patients affected by various RSDs and CSDs by using urine-derived cells. In particular, iPSC from urine sediment of a patient with autosomal dominant osteopetrosis type 2 due to defect in the ClCN7 gene were obtained and characterized to demonstrate pluripotency and karyotype.

Establishment and optimization of protocols to differentiate iPSC towards various tissues of interest have represented important goals. SYBIL developed and optimized differentiation protocols to obtain hematopoietic progenitors and /or myeloid precursors. These protocols were established for murine and human iPSC. In parallel, differentiation protocols were optimized to differentiate iPS towards mesenchymal progenitors with the final aim to obtain osteoblasts, chondrocytes, tenocytes as platform for functional studies and for the evaluation of novel therapies. These protocols have been established both in mouse models and in human iPS.

Data bank generation and coordination:
SYBIL reported the list of cellular models generated in the framework of WP2 to track bioresources. Data sets were integrated with bioinformatic analysis pipelines and represent a platform already available to the members of the consortium. This virtual biobank will be maintained and updated.

WP3: Within WP3 a broad portfolio of mouse and zebrafish models were generated. Although it was originally planned to evaluate and apply TALEN for the generation of genetically modified mice and zebrafish, in addition to conventional methods, CRISPR/Cas9 became the new gold standard for the introduction of smaller genetic modifications in mice and zebrafish and was therefore applied in WP3.

In mice, the following mouse models were generated for the following genes and associated diseases:
Acan Asp1983Asn Knock-In (KI); recessive skeletal dysplasia
Wnt1 Gly177Cys KI; severe early-onset osteoporosis
Clcn7 G213R KI; autosomal dominant osteopetrosis type 2
Col10A1 Tyr632Ter KI; Schmid metaphyseal chondrodysplasia
Fgfr3 Asn540Lys KI; hypochondroplasia
Lrp4 Arg1170Gln KI; lung development, digit development
Matn3 Thr303Met KI; osteoarthritis/chondrodysplasia MED & SEMD
Impad1 Asp177Asn cKI; chondrodysplasia and abnormal joint development
inducible miR-31 Rosa26 KI; N/A
Fbn2 F1670C KI; Marfan-like syndrome
IFITM5 S40L KI; dwarfism
Slc10A7 cKO; skeletal dysplasia with amelogenesis imperfecta
Pop1 cKO; connective tissue diseases
Sost V21M KI ; craniodiaphyseal dysplasia
Lrp5 V667M KI; lower bone mineral density and decreased strength
Kif22 P148L KI; spondyloepimetaphyseal dysplasia with joint laxity, leptodactylic (lepto-SEMDJL)
NBAS hu 6237-3 C>G KI; short stature syndrome / Pelger-Huët anomaly
Ano5 S500F KI; gnathodiaphysial dysplasia 1
Trpv4 R594H KI; skeletal dysplasia
Impad1 Asp177Asn KI; chondrodysplasia and abnormal joint development
P4hb Y393C KI; Cole-Carpenter syndrome

All models were completed and delivered or are in delivery. Most show relevant phenotypes and some were already published. Although some initial attempts using CRISPR/Cas9 technology failed (Fbn2, IFITM5) and were ultimately achieved using ES cell technology, we later used it successfully to generate the Ano5 and Trpv4 models. The wrap-up, however, is mixed: the effort to obtain clean and well-defined lines via CRISPR/Cas9 is still higher than the smooth and safe ES cell techniques. Unquestionably, this technology will have powerful applications and will likely replace existing techniques once sensitive procedural aspects are optimized.

CRISPR/Cas9 technology was successfully applied to generate unique animal models: P3h1 and Crtap knock out zebrafish as models for osteogenesis imperfecta (OI) type VII and VIII and Matn3a and COMP in frame deletion mutants as models for MED and PSACH were established. The OI and the chondrodysplasia models were deeply phenotyped in WP4 showing their relevance as disease models. The Mats3a and COMP lines are the first zebrafish models for MED and PSACH and will be of great value for testing therapeutic interventions such as those in performed in WP8. The other zebrafish lines, e.g. a Fgfr3 knockout as a model for osteochondrodysplasia are now ready for further analysis.

WP4: The aim of WP4 was to provide deep phenotyping of a number of cellular and animal models of RSDs and CSDs. These were pre-existing models that were generated before the start of the project and required detailed phenotyping to be used as models of RSDs and CSDs, as well as new models generated during the project as result of the variant selection performed in WP1, the generation and preliminary phenotyping of new cellular models obtained in WP2 and the generation and preliminary phenotyping of animal models performed in WP3. These models belonged to various species, including human and mouse for the cellular models, mouse and zebrafish for the animal models.

WP4 investigated over 40 different variants that were collected and described in deliverable D4.5. They included variants affecting the growth plate and the articular cartilage, as well as variants affecting the osseous skeleton. Chondrocytes, fibroblasts, osteoblasts, osteoclasts and mesenchymal stem cells were investigated, along with cell lines used as models in which the variants were overexpressed or silenced. Beside cartilage and bone, skin, lung, kidney, bone marrow, spleen and brain were also investigated in diseases that showed a systemic phenotype, in some case unrecognised until they were studied in SYBIL (e.g. in autosomal dominant osteopetrosis type 2).

WP4 identified many pathways affected in RSDs and CSDs. Most frequent pathways included primary cilium orientation, proliferation, ER stress and unfolded protein response, Golgi stress, vesicular trafficking, autophagy, apoptosis, enzyme and kinase activities, bone resorption and bone formation, gene expression and splicing. Pluripotency has been largely investigated in iPS cells. Overall the study has provided specific and reliable information on the cellular modifications induced in a large number of RSDs and CSDs. The majority of the variants investigated in WP4 altered the cellular phenotype and contributed to the pathogenesis of the diseases they were involved in. The studies were straightforward and were performed using methods well established in the participant laboratories. No drawbacks were encountered, therefore there were no deviations from the planned activities and no mitigation actions were necessary to complete the proposed work. The only delays were caused by new strategies for mouse generation that required correction and by the slow breeding of some mouse models. This delay did not cause problems for the achievements of all objectives of the WP, and these mouse models will be investigated by the beneficiaries in future work beyond SYBIL.

Finally, WP4 provided agreed protocols for the deep-phenotyping of RSD and CSD models in deliverables D4.1 to D4.3. The information accumulated on the biomechanical properties of cartilage and bone in animal models were collected in deliverable D4.7 while deliverable D4.9 offers a collection of gold-standard representative images and key scientific and phenotypic findings, available at http://www.sybil-fp7.eu/d4.9. We can conclude the WP4 was successful, truly representing a key and integrated WP4 for the all SYBIL project.

WP5: For such diverse bone and cartilage disorders like osteopetrosis, osteogenesis imperfecta, gerodermia osteodysplastica, Hajdu-Cheney syndrome, early-onset osteoporosis, hypophosphatemia, multiple epiphyseal dysplasia, early-onset osteoarthritis Omics profiles were generated. 250 samples from 15 in vitro and in vivo models were subjected to transcriptome profiling using an improved pipeline requiring lower RNA amounts and processing time. For five different RSDs and CSDs proteomics analyses were performed by 2D-DIGE, Label free technology, and MRM LC-MS/MS. Effects of genetic variants on cell metabolism were investigated in 15 in vivo and in vitro models. Basic data processing occurred within WP5, but higher level data integration and modelling was done within WP6.

Most important findings were novel connections between osteopetrosis and early-onset osteoarthritis and TGF-β and BMP signalling, and regulation of stemness by pathways confined to osteoclast differentiation. Inflammatory signatures were identified in Hajdu-Cheney syndrome and in osteocytes under stress. Circulating protein markers for ER-stress induced by unfolded ECM proteins were detected providing options for disease monitoring. miRNAs secreted by senescent cells were shown to regulate many mRNAs relevant for skeletal biology and could serve as biomarkers for CSDs and contribute to their pathogenesis.

Wetlab as well as bioinformatics tools for epigenetic profiling of chromatin modification and folding were developed. Using these technologies, a comprehensive enhancer map for genes relevant for limb development was generated. Based on these results analysis of disease states revealed enhancer adoption secondary to copy number variants as a common mechanism for congenital skeletal malformations like brachydactyly.

A SYBIL phenotyping and Omics ontology was set up to include all terms used within SYBIL, ensuring a common vocabulary and no nomenclature clash between researchers. This was prerequisite for successful data integration into the SYBIL central data repository, which was extended to cover the entire breadth of data, documentation using the SYBIL standard procedures and ontologies. Having an overarching model enables data analysis to be automated through defined pipelines. The data is maintained in a number of interrelated databases and data repositories, which will be kept functional beyond the official end of the SYBIL project.

The Human Phenotype Ontology (HPO) was extended within SYBIL to also cover common diseases. The hiPHIVE algorithm that sorts candidate disease genes according to phenotype relevance based on HPO terms. It also includes candidate genes based on the comparison to model organisms and on protein interaction networks of known disease gene products. Based on these achievements the software Genomiser was developed allowing prioritization of variants in the non-coding genome. A software for design of chromatin capture probes also aimed at a deeper investigation of the non-coding genome.

In summary, WP5 was an important motor for the whole consortium by providing high-throughput data and the necessary IT infrastructure and ontologies to build up a central knowledgebase that will have a stimulatory effect on research on CSDs and RSDs that reaches beyond the SYBIL community.

WP6: UNIMAN has developed processes for the integration of Omics datasets into network models, aimed at the analysis of disease samples and at the prediction of new disease related genes/proteins. This process was applied to several hundreds of existing and newly generated transcriptomics datasets to yield expression response profiles characteristic of skeletal disorders. The responses are primarily from bone and cartilage tissues and cover a range of experimental types including disease vs non-diseased tissues and genetic perturbations. The combination of these results leads to the identification of candidate genes and pathways with strong disease association. Upstream regulators and drugs counterbalancing the expression responses were identified for all these perturbations.

These results were made available on an interactive data portal (SkeletalVis – http://phenome.manchester.ac.uk/) freely available for the worldwide skeletal biology community. SkeletalVis is composed of exploration and comparison modules as well as a detailed help section. The exploration section allows visualisation of the detailed analysis; the experimental annotations with the type of perturbation and experimental platforms are also provided. The data portal allows searching of the differential expression table, which can be searched to find particular genes, copied or exported as text files for use with external tools. SkeletalVis provides detailed systems biology analysis with enriched pathways, drugs, transcription factors, which can be viewed in interactive tables to identify the key dysregulated biological processes. Active sub-networks can be viewed as interactive networks coloured by fold change. In addition, the shared response tool shows what are the most similar/dissimilar known perturbations to a new experiment, which is of interest to discover relations and generate hypotheses when confronted with a new type of perturbation.

ALACRIS revised and extended the mechanistic model of cellular pathways that is part of the predictive ModCellTM system to reflect molecular components and pathways related to bone-specific molecular processes and skeletal diseases. ModCellTM has been built on the premise that the onset and progression of a disease is associated with a malfunction in the complex biological networks occurring in specific cells or tissues. The model is essentially a computational representation of the complex molecular networks in cell types and tissues. Because ModCellTM has been developed for human cells, mapping of orthologous genes between mouse and human is required to make the mouse data applicable for ModCellTM. In order to overcome this, ALACRIS has implemented 16 mouse signalling pathways in the model. Next, ALACRIS adapted its drug database to the revised ModCellTM model. Currently, the drug database comprises over 500 drugs, each described by their targets and affinity metrics (e.g. Kd, IC50). Of these drugs, 62 were found to be potentially relevant for skeletal disease and predictive modelling of their therapeutic effects. The drugs identified also include candidates for repositioning.

ALACRIS has focused on analysis of the Unfolded Protein Response (UPR) signalling pathway, comprising 18 genes, including Hspa5 (BiP) and the three ER stress sensors, Ern1 (Ire1), Atf6 and Eif2ak3 (Perk). The UPR pathway model has been tested under different virtual conditions such as perturbation of ligands, receptors and regulators. After successful evaluation of the human and mouse UPR pathway tests, mRNA-seq data of human osteoclast, murine osteoblast and osteocyte samples (data provided by the CHARITE) were used for model initiation and simulation. Monte Carlo simulation predicted that the UPR signalling activation capacity in osteoblasts is much higher than in osteoclasts and osteocytes. In order to understand if the unfolded protein response is involved in the molecular pathology of gerodermia osteodysplastica driven by Gorab exon deletion and/or physical load, ALACRIS simulated the effect of unfolded protein on mouse osteocyte samples using sample-specific transcriptomics data (data provided by the CHARITE). ModCellTM predicted no significant changes in UPR signalling activation capacity between wild-type and mutant samples, as well as between samples with and without exposure to physical load.

ALACRIS also performed virtual drug screening using the following UPR signalling pathway inhibitors: 4u8C, kira6, ISRIB, GSK2606414 and GSK2656157. Monte Carlo simulation of drug effects for individualized models of osteoblast, osteoclast and osteocyte using sample-specific transcriptomics data, predicted that osteoblasts are more sensitive to Ern1 inhibitors 4u8C and kira6, and Eif2ak3 inhibitors ISRIB, GSK2606414 and GSK2656157, than osteoclasts and osteocytes. The strongest inhibition of UPR signalling was elicited by Ern1 inhibitors 4u8C and kira6, which was in line with the observed stronger activity of the Ern1 branch in the pathway.

Mucolipidosis type II (MLII) is a lysosomal storage disorder caused by inactivating mutations of the GNPTAB gene. Affected children usually do not survive the first decade of life, since the failure of lysosomal protein targeting caused by the GNPTAB mutations results in severe defects of many cell types. Osteoblasts gene expression datasets from a GNPTAB mouse model were provided by UKE to UNIMAN. A cell specific network for the MLII osteoblasts was constructed, then gene expression analysis was performed and active modules were detected to predict candidate gene targets for MLII.

Microarray data of rib cartilage samples coming from a mouse model of Desbuquois dysplasia supplied by UNIPV were analysed by UNIMAN to reveal the effect of a Cant1 knock out. Other data from wild-type and mutant zebrafish were analysed to reveal mechanisms of osteogenesis imperfecta.

CERTUS has integrated the systems biology analysis pipeline with their research note management system, enabling streamline analysis and exporting of existing and new datasets for further downstream analysis. The SYBIL knowledge model was developed to record research activity within the SYBIL project, maintaining knowledge of all resources made available by partners, the variants under investigation and the results of experimental work. All these resources are recorded in the SYBIL knowledgebase, which will be maintained for a significant period post project to allow partners to continue to access the data sets and pipeline, and to encourage further contributions from partners related to the work conducted within SYBIL.

WP7: In the context of WP7 several candidate biomarkers were identified, specific quantitative methods were developed (mostly for non-peptide biomarkers), and some of the quantifications clearly demonstrated the diagnostic and prognostic value of the specific biomarker approach. As expected, we also obtained few negative results, i.e. no significant difference between patients and healthy controls for some of the candidates. More specifically, serum concentrations of CHAD or THBS4 were not affected in individuals with osteoarthritis, and the same applied for WNT1 in individuals with osteoporosis. On the other hand, the SYBIL approach has led to the identification of promising biomarkers for RSDs and CSDs, which will be discussed separately in the following paragraphs.

With respect to protein biomarkers for RSDs, three disease-specific candidates were identified. First, serum levels of sclerostin were increased in individuals with sclerosteosis due to mutations of the LRP4 gene. Second, an increase of bioactive natriuretic peptides was identified in patients with a novel genetic disorder associated with tall stature, long digits and extra epiphyses. Third, serum IL-6 concentrations were found increased in a mouse model of Hajdu-Cheney syndrome, thereby confirming the initial results obtained by transcriptomic analyses. With respect to non-peptide biomarkers one relevant method was established to quantify the sulfation pattern of urinary chondroitin sulfate (CS). This method was extremely useful to diagnose two specific RSDs, i.e. diastrophic dysplasia (characterized by enrichment of non-sulfated CS) and mucopolysaccharidosis-IV (characterized by enrichment of CS sulfated at position 4). In both cases the simultaneous quantification of urinary 0S-CS, 4S-CS and 6S-CS was used to demonstrate a treatment effect in the context of WP8.

With respect to protein biomarkers for CSDs, three candidates were identified in the context of osteoporosis. Whereas serum CHAD concentrations were found significantly increased in women with postmenopausal osteoporosis, both, LCN2 and PENK1, where identified as biomarkers for disuse-induced osteoporosis. Although we were so far unsuccessful to define a long-sought biomarker enabling the diagnosis and evaluation of osteoarthritis, we made significant progress based on a Systems Biology approach in the context of this highly prevalent disorder. More specifically, by systematically analysing gene expression in knee cartilage from 60 patients (and 10 controls) it was found that there are two groups that can be separated by gene expression. Since some of the differentially expressed genes encode secreted proteins, it is now possible to determine their concentration in synovial fluid or the serum of the respective patients. With respect to non-peptide biomarkers, a commercially available system (osteomiRTM) was established to quantify serum concentrations of specific microRNAs. Here it was found that some of these were significantly associated with fracture incidence in patients with type II diabetes, which is extremely important, since these patients often have a bone mineral density in the non-osteoporotic range.

Taken together, WP7 was not only successful regarding the validation of several novel biomarkers for skeletal disorders. It also provided the basis for additional experiments beyond the SYBIL time frame. First, all SYBIL partners, and most likely other researchers, will continue to determine the identified biomarkers in additional patients of the respective disorders to confirm the present results, but also to expand the applicability of the developed assays. Second, many data generated in WP5 and WP6 are still utilized to identify potential biomarkers of further skeletal disorders, and their quantification will be continued by the SYBIL researchers. Third, the established method to quantify the urinary CS pattern is directly applicable to monitor treatment effects in patients with disorders of glycosaminoglycan synthesis or degradation (including diastrophic dysplasia and mucopolysaccharidosis-IV). Fourth, the commercially available osteomiRTM system will be continuously utilized by SYBIL partners and other researchers in order to confirm its applicability to predict bone quality and fracture risk. Taken together, since biomarker identification is generally considered as a difficult task, the outcome of WP7 has to be regarded as highly successful.

WP8: From previous results and from WP4, WP5 and WP6 activities, SYBIL Partners have identified different cellular functions and pathways involved in the pathogenesis of RSDs and CSDs that are crucial for correct bone and cartilage development, maintenance and homeostasis including autophagy and proteasomal degradation, ER stress, sulfation pathway, signalling pathways of proliferation and differentiation, protein synthesis, extracellular matrix degradation and bone resorption. Based on these data a vast portfolio of molecules, targeting specific cellular functions or pathways, have been considered including peptides, enzymes, molecular chaperones, tyrosine kinase inhibitors, amino acid derivatives and siRNA.

Drug repositioning has been considered as a good opportunity to speed up the treatment of patients; for this reason, different molecules already used to treat other diseases have been tested. Targeting endoplasmic reticulum stress with already FDA approved molecules (4PBA, carbamazepine) identified a novel therapeutic approach to ameliorate RSD outcome. Drug repositioning has been considered also for diastrophic dysplasia and gerodermia osteodysplastica using acetilcysteine. In addition, small molecules inhibiting specific tyrosine kinases and Wnt pathway have been studied. Finally, small peptides and proteins have been considered including RANK ligand, an enzyme replacement therapy (lysosomal arylsulfatase B) and a TGFβ antibody to treat Autosomal Recessive Osteopetrosis RANKL dependent, Mucopolysaccharidosis type VI and Gerodermia Osteodysplastica, respectively.

The studies have been performed in animal models of human skeletal disorders and/or in cells from the same models, from the patients or in Hela and HEK cells transfected with mutant proteins. The effects of small molecules on these in vivo or in vitro systems have been evaluated using a wide range of experimental approaches including histology, X-rays, micro-CT, skeletal staining, biochemical studies, expression at the mRNA and protein levels and bone resorption assays.

In the WP8 framework, reagents for transient or stable modification of the genome have also been considered in order to correct the gene defects in mouse and human iPSC derived from various skeletal rare diseases. Regarding transient modification, siRNA has been successfully used to treat a mouse model of autosomal dominant osteopetrosis linked to the Clcn7 gene and the molecule is now in preclinical optimization. Stable modification of the genome has been achieved using mouse and human induced pluripotent stem cells (iPSC). The feasibility of precise gene correction using homologous recombination in autosomal recessive osteopetrosis linked to the TCIRG1 and ClCN7 genes has been demonstrated in vitro. In parallel, the CRISPR-Cas9 technology has been tested to correct autosomal recessive osteopetrosis (ARO) caused by defects in the ClCN7 gene. Compared to the conventional gene therapy approaches (gene addition), these strategies offer the advantage of maintaining the control of gene expression by natural regulatory elements and avoiding the risk of dangerous side effects (insertional mutagenesis).

In conclusion, a collection of potential new therapeutic approaches has been prioritized in in vitro systems and in animal models in order to define: i) the administration route and time of administration, ii) the phenotype after the treatment and the methods used to evaluate the skeletal amelioration, iii) any side effect.

WP9: The non-classified information was disseminated to a wide audience through a publicly available website (https://www.sybil-fp7.eu). The website generated by CERTUS and UNEW included lay language information about the diseases and variants studied by SYBIL, SYBIL news and bi-annual newsletters, and standard operating procedures (SOPs) which were shared with the wider research community to set the gold standard in skeletal research. The website had on average 1,000 visitors a month throughout the project. A @SYBIL_news Twitter account (UNEW) was also used to disseminate SYBIL news and soundbites. Moreover, the SYBIL project resulted in more than 40 publications, over 85% of which were open access. SYBIL funding was cited in all SYBIL publications in order to emphasise document ownership. The results of SYBIL work were also presented at national and international conferences (ALL PARTNERS), through around 130 oral presentations and 90 posters. SYBIL researchers (ALL PARTNERS) were also involved in national and international public engagement events including the International Rare Disease Day and the European Researchers’ Night and organised national and international events reaching over 900 school children, 5500 members of the lay public, 1100 members of the patient organisations and over 550 clinicians and health professionals Europe-wide.

Mobility plan for young scientists was established in year 1 of the project. Overall, SYBIL facilitated 16 research visits between the SYBIL centres (UNEW/ UNIVAQ/ UNIMAN/ CERTUS/ EVCYT/ UNIPV/ UKK/ UKE/ UA/ POLYGENE/ ALACRIS/ CHARITE; 3 in year 2, 6 in year 3, 6 in year 4 and 1 in year 5), all of which resulted in knowledge transfer and were disseminated at national and international meetings. Several of these led to collaborative publications as well, and gave the junior researchers an experience of different research environment as well a highly valuable and transferrable skillset to apply in their future careers. Moreover, a ‘Systems Biology Hub’ was established on the password protected portal (CERTUS) in order to provide training in systems biology and to help junior research retain ‘ownership’ of their data and collective sense of belonging to the project. Junior ownership of data was further reinforced by actively promoting junior oral and poster presentations at national and international meetings (ALL PARTNERS), resulting in the total of 65 junior oral presentations (50% of all SYBIL oral presentations) and 38 junior posters in the 5 years of the project.

Training programmes for scientific and regulatory personnel were developed at year 1 and were updated throughout the project (ALL PARTNERS). Overall, SYBIL staff presented at and participated in 22 national and international workshops, some of which were open to the wider skeletal community and deposited on publically available websites. SYBIL also organised workshops for the junior researchers in the wider skeletal community, as satellite workshops at several international meetings (European Calcified Tissues Society ECTS meetings in 2016 and 2017 and Matrix Biology Europe MBE conference in 2014). Our junior researchers also participated in SYBIL annual meetings where they presented their data and forged new collaborations. Moreover, all the standard operating procedures (SOPs) from the SYBIL laboratories were deposited on SYBIL website to provide training and gold standards for the wider skeletal community.

We have implemented a communication infrastructure for the project in year 1 (CERTUS). The internal collaborative SYBIL portal was maintained by CERTUS throughout the project and served as a platform for exchange of information and large data sets for the “Systems Biology” applications. CERTUS have also developed several data analysis pipelines in collaboration with UNIMAN and made those available in the portal, further facilitating data comparisons and collaborations within the consortium. Moreover, a collaborative social media-like portal was set up for the duration of the project to enable junior communication and exchange of protocol and ideas.

SYBIL exploitation and commercialisation plans were discussed in year 3 and developed in following years. A plan to share the animal and cell models was drafted and implemented through the collaborative SYBIL portal, which will remain live past the project. Moreover, PhenomeExpress (http://phenome.manchester.ac.uk/) Systems Biology software was developed in collaboration between CERTUS and SYBIL and has been used for comparative analysis of SYBIL large datasets, thus initiating new collaborations. It is also freely available to the wider scientific community. Commercialisation hurdles have been overcome and SYBIL project led to two patents (UNEW and UNIVAQ) and six patent applications (UNIVAQ). SYBIL also resulted in one current (MCDS Therapy https://mcds-therapy.eu; UNIMAN/ UNEW) and one planned (UNIVAQ) clinical trial, R&D results for commercial exploitation (GATC), and general advancement of knowledge (ALL PARTNERS)

WP10: WP10 has been dedicated to the administrative, scientific and financial management of the consortium as well as coordination of communication and reporting.
Project’ strategic and operational management. During the full duration of the project, the management team at UNEW and FINOVATIS dedicated much of its time to setting up and deploying all actions, tools and systems necessary to ensure the good implementation of the project. This was done in order to ease partners’ work and avoid unnecessary administrative burden. In total, 6 SYBIL annual meetings were held (including kick-off and final meetings). An EC review meeting was also organised half-way through the project. These meetings included the PIs from all the partnering institutions as well as key personnel and collaborators of the project. Junior scientists and members of the SYBIL Scientific and Ethics Advisory Board (SEAB – as formed within the first month of the project) were involved in all the meetings. For each meeting, an agenda was prepared including time for scientific presentations and discussions in the form of chaired plenary sessions. During all meetings and as part of the project’s strategic management, Governing Board (GB) meetings were organised. The GB was in charge of overseeing the overall progress of the project towards its objectives, deliverables and milestones, while ensuring the proper administrative, legal and financial running of the project. One of the main managerial issue to be addressed was the use of the Reserve Funding that has been provisioned under UNEW to cover for training events & visiting exchange programme. Minutes were always prepared from each of these annual meetings. circulated to all GB members. Several ad hoc meetings and TCs were also organised. As part of the project’s operational management, the coordination team allocated much of its time in the preparation of the contractual reportings (financial and technical) by generating and providing templates and instructions to the consortium. Management tools have been prepared and circulated to partners (including reporting templates, guidelines and timeline, reminders on deliverables and milestones, etc.). Collection of information (including deliverable reports, dissemination activities and publications) were also centralised by the coordination team to be all uploaded onto the participant portal in due time. All deliverables of WP10 have been completed and returned on time.
IPR and knowledge management. SYBIL did undertake IPR/Knowledge management and monitoring in order to guide the consortium in its choice of research and achieve operational freedom. This was based on the fundamental ethical rules and principles recognised at the European level, ensuring that the SYBIL research is driven into the right directions. The access rights to pre-existing know-how and knowledge were defined in the general conditions of the EC contract and detailed in the Consortium Agreement. Confidentiality of information and knowledge was enforced as defined in the Consortium Agreement. Based on regular feedback from all PIs over the course of the project, information on partner-specific IP protection or exploitation matter was collected as reported in both (1) the final report and (2) onto the EC portal (sections on patents, exploitation as well as dissemination activities and publications).
Gender actions. In order to promote the participation of women in this project and related research projects, a set of gender indicators was produced in the form of a questionnaire to measure progress towards gender equality in skeletal diseases research. Partners were sent this questionnaire in order to monitor gender equality over the full duration of the project. Overall, women were well represented within SYBIL and gender balance well maintained over the full duration of the project.
Potential Impact:
WP1: During the course of the project, WP1 created a comprehensive portfolio of genetic variants for RSDs and CSDs. A carefully selected cohort of variants was defined at the start of the project. The selected variants were extensively studied, characterised and validated by a broad range of state-of-the-art techniques (WPs). These pre-clinical models are unique and very valuable resource. These new knowledges are an immediate and major impact on developing and delivering new therapies for patients across the full spectrum of skeletal diseases.

WP2: Genetic skeletal diseases have an important socio-economic cost since the majority of these rare diseases can not be prevented or treated by conventional therapies. To this end, the establishment of a comprehensive portfolio of cellular models represents a unique and valuable source to investigate molecular pathways involved in the pathogenesis of the disease thus allowing to unravel novel players that might be relevant for the development of alternative therapies. In parallel, the availability of such preclinical models represents the prerequisite to assess the efficacy of therapies including new drug candidates or asses the feasibility of gene therapy approaches. In vitro models here generated, have covered the diverse range of gene products associated with skeletal defects. Of note, the development of these cellular models has represented a platform on which novel cell therapy approaches including gene editing and gene therapy can be tested (WP8). Moreover, generation of iPS from various tissues including urine sediment represents a step forward to obtain iPS cells that once corrected can be used for autologous transplantation in a near future when novel techniques will allow to overcome the current clinical limitations. The social impact is also strengthened by the generation and maintenance of a data bank reporting all the cell lines newly generated in the framework of the project, that can be now released to the scientific community. Dissemination activities have represented a relevant part of the WP2. Data obtained have been presented at national and international meetings. Moreover, results and techniques newly developed in the framework of the project have been shown in the framework of educational workshops (PhD students) and in high schools or patient group events such as those organized by Telethon foundation.

WP3: According to the central role of WP3 within SYBIL the successful generation of animal models had a strong impact on WP4, WP5, WP6 and WP8 as it delivered very important tools for those work packages. The large number of animal models will, beyond SYBIL, have a long-lasting impact on the progress in the field of skeletal diseases. The scientific community as a whole will strongly profit, and the models will finally yield unravelling of disease mechanisms, detection of biomarkers and development of therapies which will definitively have socio-economic impact and give patients a better life. For example, MED and PSACH, caused by mutations in Matn3 and COMP are rare skeletal diseases where drug development is economically not attractive. The establishment of the Matn3a and COMP mutant zebrafish line will now allow the large-scale testing of small molecules at the larval stage in very short time and at high numbers which would not by possible in mice. This enables an unbiased approach by which really novel drugs could be developed. On the other hand, many of the mouse models accurately recapitulated the human morbidity and hence lend themselves for detailed investigation in a way that was not accessible yet. In sum, these demonstrate the unique usefulness of small rodents for the modelling of human disease and making rare diseases a possible object of research. Given the small numbers of some of the specific patient groups, this was hitherto impossible. Animal models generated by WP3 have been presented often by young investigators at national and international meetings, e.g. Gordon Conferences or Matrix Biology Europe in oral presentations and posters. For several publications, animal models were used that were generated by WP3. Taken together WP3 generated an extraordinarily large number of highly relevant animal models for SYBIL, but even more important for the whole skeletal research community and some of the models will have the potential to become “gold standard” disease models having strong impact diagnosis and drug development.

WP4: With a critical analysis of the results of WP4, we can say that this WP had a large impact on the whole project, allowing the consolidation of the SYBIL results in reliable and well established cellular and animal models. WP4 is also likely to have a strong future socio-economic impact contributing to lay the foundation for a better understanding of RSDs and CSDs that affect a large population of individuals in Europe and beyond. Just as concrete example, drug repositioning to treat RSDs characterised by ER stress has been hypothesised within SYBIL also with the contribution of WP4. This drug is now the object of a separate EU-funded project and quickly proceeded to clinical trial. WP4 has also contributed to identify systemic changes in diseases that were previously thought to affect exclusively the skeleton, and experimental therapies tested in SYBIL to cure the bone phenotype turned out to be effective also on the other affected organs. These findings may have future societal implications, reducing the burden on patients and families and the costs for the assistance of severely compromised patients who may benefit from these treatments. WP4 has contributed to the patenting of two treatments and has disseminated the results in public engagement activities and in a number of scientific meetings, including the Matrix Biology Europe congresses, the European Calcified Tissue Society congresses, the American Society for Bone and Mineral Research congresses, and other meetings attended especially by young investigators. Several SYBIL publications have been produced by the large contribution of WP4. Nicely, it has to be noted that most of them are the result of the integrated activities of several WPs, in which WP4 has often played a central and indispensable role. In conclusion, WP4 generated the expected results and was successfully completed. SYBIL partners inherited many information and models that will be further investigated beyond SYBIL, representing a strength for future collaboration and funding opportunities.

WP5: Systems Biology is not possible without systematic acquisition of high-throughput data, which was exactly the role of WP5. Therefore, this work package was crucial for the aims of the SYBIL consortium. The data and the computational environment provided by WP5 enabled discoveries in other work packages aiming at the development of novel biomarkers for disease monitoring and novel therapeutic approaches that will have socio-economic impact. In fact, based on the results generated within SYBIL a clinical study targeting ER stress has been initiated. The knowledgebase created within WP5 will have a long-lasting effect for the SYBIL community since it will be further maintained. In addition, WP5 provided data for WP6, which had as one output a public web interface (Skeletal Vis) for the evaluation of the role of candidate genes for skeletal disorders and for the analysis of differential gene expression data. An important contribution of WP5 was the extension of the Human Phenotype Ontology (HPO). The HPO allows for encoding phenotypes using computable vocabulary. Thus, this terminology plays a crucial role for the interpretation of big data and data mining and facilitates the generation of novel insights into human diseases, which will have immense socio-economic impact through the discovery of novel targets for diagnostics and therapies. Therefore, the HPO was adopted by Global Alliance for Genomics and Health as the recommended ontology for rare diseases. SYBIL partners are part of the Monarch Initiative (htp://monachinitiative.org) and have created a repository for deep phenotyping data on human diseases and model organisms that includes over 1000 skeletal terms. This public website includes a number of tools for data mining and retrieval that is used by researchers around the world. SYBIL fostered the development of HPO-based bioinformatics tools for the interpretation of NGS data (Exomiser, Genomiser), that are widely used within the community. In addition, the software GOPHER aiming at optimizing chromatin capture probe design is freely available. Results obtained by WP5 were disseminated in public engagement activities and in many scientific meetings, including the European Calcified Tissue Society congresses, the American Society for Bone and Mineral Research congresses, as well as the conference of the European and American Societies for Human Genetics. Hence, WP5 fully reached its envisaged aims and played a crucial role for the success of the whole consortium.

WP6: Within any large and lengthy research project, the ‘omics’ data made available for analysis must be carefully marshalled and curated. For SYBIL, this has entailed maintaining all relevant assets together with their full provenance, including details of the processes by which they were generated. A platform has been constructed, using best practice from software development, to support the systems biology analyses within SYBIL. The platform has become a significant and potentially lasting resource for researchers interested in rare and common bone disease. Moreover, CERTUS has developed a clear understanding of the issues surrounding data management in ‘omics’ centred projects and is active in their application in new projects similar in nature. Experience gained in SYBIL has informed both requirements gathering and systems design in the field of ‘omics’ data management. CERTUS expects to continue development in these areas and is working with a number of partners to improve data management in leading edge clinical genetic diagnostics. UNIMAN has disseminated outputs from SYBIL to a wide-range of scientists to share results and approaches at national and international conferences, including ISMB 2017, OARSI 2018, NetSci 2018 and MBE 2018 with both poster and oral presentations. Work was also disseminated in international, peer-reviewed journals. A key output of WP6 is the integration of ~250 cross-species, skeletal transcriptomic experiments and construction of the online intuitive data-portal (SkeletalVis – http://phenome.manchester.ac.uk/) to allow exploration and meta-analysis by the international research community. This consolidation of complex data into an accessible format is crucial to gaining meaningful information from these datasets in order to increase our understanding of skeletal biology and disease. The development of a systems biology approach for predicting drug response for bone disease by ALACRIS will enable new omics-driven approaches to be used for new targeted approaches, even for rare bone diseases. The expansion of such models and their pre-clinical validation will be important to improve current approaches to diagnosis and treatment of bone-diseases. In the long run, this should provide the basis for development of new drugs, drug repurposing and new medical and biotechnology research and commercial exploitation, for improving the well-being of patients.

WP7: Biomarkers are widely used in clinical practice to diagnose specific pathologies, such as hepatic disorders (by measuring activity of liver-specific enzymes) or prostate cancer (by quantifying serum levels of prostate-specific antigens), just to name two of many well-known examples. For the diagnosis of specific skeletal disorders biomarkers have only been established for few of these, for instance pyridoxal-5´-phosphate for hypophosphatasia or FGF-23 for hypophosphatemic rickets. In contrast, there are still no established biomarker approaches for the majority of rare skeletal disorders and for the two most prevalent skeletal disorders, i.e. osteoporosis (> 25 million patients in the EU) and osteoarthritis (> 40 million patients in the EU). Therefore, any knowledge, even in case of preliminary data, on candidate biomarkers that can be used to diagnose these disorders, has a high socio-economic impact and societal implication. In the context of WP7 we have established previously unknown biomarkers with diagnostic and prognostic value for specific rare skeletal disorders, such as sclerosteosis, diastrophic dysplasia or mucopolysaccharidosis-IV. Moreover, we identified novel biomarkers for disuse-osteoporosis and for fracture prediction in individuals with type II diabetes. Finally, although the respective large-scale quantification approaches are still in progress, WP7 has identified additional promising biomarker candidates, which might be extremely useful to monitor further skeletal disorders, in particular osteoarthritis, where a prognostic biomarker is urgently needed. Data from WP7 have been presented by various (mostly young) investigators in oral presentations at different highly regarded scientific meetings, including the SYBIL satellite meetings at ECTS 2014/2016. Moreover, there are so far 9 publications including biomarker data that were obtained in the context of WP7, and since additional manuscripts are currently in preparation, the scientific output is expected to be even higher. Key results were also announced in the SYBIL newsletter and in press releases from the specific universities. With respect to the methods required to quantify the respective protein biomarkers, it is important to state that necessary ELISA systems are commercially available for most of them, which allows other researches to reproduce and substantiate our findings. In terms of non-peptide biomarkers, i.e. sulfation pattern of urinary chondroitin sulfate and serum concentrations of specific microRNAs, we have described the detailed methodology in the respective publications. Most importantly, EVCYT developed the latter strategy into a commercially available system (osteomiRTM), which is now available for the scientific community as a quick 3-step system for 11 biomarkers of bone quality. Finally, although their applicability to diagnose specific skeletal disorders remains to be established in future experiments, it is important to state that PRIMM has developed specific antibodies in the context of WP7. These include five monoclonal antibodies against extracellular vesicles produced by human mesenchymal cell cultures. Since WP7 also generated data in mouse models demonstrating that extracellular vesicles serve as carriers for various cytokines and potential biomarkers for skeletal disorders, these antibodies may be of great interest for the scientific community. In that specific case several partners of the SYBIL consortium will first utilize these novel antibodies to confirm their potential relevance in biomarker quantification with the aim to publish these data in highly-ranked scientific journals. Once this has been achieved, it will be decided, how these interesting tools will be made available for the scientific community.

WP8: The final objective of SYBIL has been to provide the necessary data resource and technological platforms for the accelerated development of innovative therapeutic approaches, which are the basis for future personalised treatments of skeletal diseases. In 10 different skeletal disorders (including RSDs and CSDs) effective small molecules have been characterized and will be considered for further studies aimed at the development of therapeutic approaches. Huge efforts in collaboration with WP7 have been made also in the validation of non-invasive biomarkers in serum and urine important for monitoring disease progression and efficacy of treatments. These findings have future socio-economic impacts by reducing the burden on patients and their families and reducing the healthcare costs of patients that may benefits from these treatments. Results from WP8 were disseminated in public engagement activities and in many scientific meetings including Matrix Biology Europe, International Skeletal Dysplasia Society meeting, European Calcified Tissue Society, the American Society for Bone and Mineral research, European and American Societies for Human Genetics and several national meetings. Several publications have been already produced, but since additional manuscripts are in preparation, the scientific output is expected to be even higher. This research activity has allowed the generation of a database of new knowledge on therapeutic targets for RSD and CSD that summarises, for specific skeletal disorders, the phenotypic effect of small molecules/reagents targeting specific cell functions. This new therapeutic knowledge is a starting point for the development and testing of new molecules or reagents that will be available after the end of the SYBIL project.

WP9: The SYBIL project generated novel knowledge and models applicable to study and treatment of a variety of skeletal conditions. WP9 activities such as presenting at national and international meetings, SYBIL satellite meetings at prestigious conferences, high impact open access publications and SYBIL soundbites, allowed the dissemination of SYBIL knowledge and data to the wider research community, thus furthering the general knowledge in the field and setting gold standards for skeletal research. SYBIL dissemination channels under WP9 were used to share SYBIL knowledge and the standard operating procedures (SOPs) in order to ensure standardisation of data across the wider scientific community and building larger systems biology networks world-wide. SYBIL systems biology pipelines have been made available to the wider scientific community and all large SYBIL datasets have been uploaded to publicly accessible databases thus contributing to furthering skeletal research field. SYBIL dissemination activities, either stand-alone or as part of larger international initiatives such as the International Rare Disease Day and European Researcher’s Night, reached large audiences. WP9 supported the dissemination of information to a variety of stakeholders, empowering and engaging the members of the public and patient organisations and further strengthened the link between the patient organisations and the scientists in their respective centres and Europe-wide. Moreover, SYBIL public engagement and STEM activities served to inspire young minds and promote scientific research as a viable career option. Furthermore, young SYBIL scientists were able to participate in research exchanges and training that gave them transferrable skillsets and international networks that will be invaluable in their future careers. SYBIL work led to novel IP in furthering skeletal knowledge, Systems iology analysis pipelines, biomarker discovery and treatments. SYBIL generated several cell and animal models of common and rare skeletal conditions and associated disease mechanisms that were published through WP9 channels and together with SYBIL generated large datasets are now available to the wider scientific community. It also led to R&D improvements in the SYBIL SMEs’s processes and two SYBIL patents for treatment of rare skeletal conditions, with six more patent applications and potential company spin-off currently being processed. Moreover, a SYBIL-related clinical trial for treatment of a rare skeletal condition, MCDS Therapy, was H2020 funded based on SYBIL pre-clinical data and commenced in January 2018.
List of Websites:
http://www.sybil-fp7.eu