Final Report Summary - FIGHTINGDRUGFAILURE (Priorities and Standards in Pharmacogenomic Research: Opportunities for a Safer and More Efficient Pharmacotherapy)
Executive Summary:
Final publishable summary www.fightingdrugfailure.net
FIGHTINGDRUGFAILURE: PRIORITIES AND STANDARDS IN PHARMACOGENOMIC RESEARCH:
OPPORTUNITIES FOR A SAFER AND MORE EFFICIENT PHARMACOTHERAPY
Pharmacogenomics is the scientific discipline that relates inter-individual variation in drug response to genetic variation. While there have been major advances in basic genomic knowledge and genotyping technologies, translation into clinical practice has been slow. To help define which patients are going to respond to drugs and which patients are going to develop adverse drug reactions (ADR), this ITN took a concerted action across the EU for the training of 14 ESR and 3 ER to quicken the pace of translational research. This involved the generation and application of -omics data for combined analyses with clinical data of defined patient cohorts and to relate it to clinical outcomes of drug therapy as well as performing functional studies where possible. The sharing of patient materials, access to high-throughput technology platforms, pooled or meta-analyses, as well as transfer of knowledge and skills across laboratories of academic and industry partners including third partners hosting secondments clearly enhanced this translational research effort towards the promotion of uniform drug treatment standards and priorities throughout Europe and worldwide, and moreover improved the competitiveness of the EU in this emerging field of personalized medicine.
PREDICTION OF CANCER RISK AND DRUG TREATMENT OUTCOME (RBMF Stuttgart and N.N. Petrov St. Petersburg)
A focus has been on breast cancer which is the most frequent cancer in women with more than 1.34 million newly diagnosed cases each year and more than 400.000 annual deaths worldwide. Strategies of personalized risk assessment, molecular phenotypic classifications and the assessment of molecular predictors for drug treatment (tamoxifen, cytotoxic agents) outcome have been explored. Pharmacogenetic investigations showed that the polymorphic drug metabolizing enzyme CYP2D6 is a predictor of tamoxifen outcome in estrogen receptor (ER) positive breast cancer due to its role as the key enzyme in the formation of the active metabolite endoxifen. We showed that tamoxifen metabolism follows the same mechanism in pre- and postmenopausal patients across different ethnicities. Notably, premenopausal patients with impaired CYP2D6 metabolism showed a higher risk to experience breast cancer recurrence as compared to patients with full enzyme function, a finding that has been underscored by genetic and pharmacokinetic data thereby suggesting a potential for clinical use.
Because endocrine resistance also depends on the breast tumor molecular phenotype, we explored the role of tumor-associated microRNAs for the risk of recurrence in patients treated with adjuvant tamoxifen. Notably, increased expression of miR-126 and miR-10a has been identified as independent predictors for tumor relapse. Other than ER positive breast cancer that can be targeted by tamoxifen, triple negative breast cancer (TNBC) lack expression of hormone receptors (ER, PR and HER2), representing a heterogeneous group of aggressive cancers not amenable to current targeted treatments. We showed that microRNAs are suitable classifiers to differentiate between TNBC and ER positive breast cancer. These findings hold the potential to find gene regulation patterns potentially useful for the development of algorithms for breast cancer subtype classification and the discovery of novel drug target towards the control of TNBC. Studies on the prediction of response to neoadjuvant cytotoxic therapy of breast cancers characterized by germline mutations such as BRCA1 founder mutations showed that the frequency of pathological complete response was 28% in BRCA1 carriers as compared to 12% in mutation negative patients with similar clinical response rates. This translates into a prediction of greater than 70% response rate to the neoadjuvant treatment for BRCA1 mutation carriers which is in contrast to 25% in mutation-negative TNBC, again highlighting the need for novel drug targets and targeted therapies for TNBC.
At the level of breast cancer risk prediction, we showed that the HSD17B1_G allele protected women from developing breast cancer when they had used hormone replacement therapy (HRT) for more than 10 years. Instrumental to this finding was the availability of large population-based case-control collections from Germany, Sweden, and Australia (>5500 cases, >8000 controls). Importantly, we investigated more than 35 newly identified breast cancer susceptibility loci recently published by the BCAC and COGS consortium (Michailidou et al. Nature Genet 2013) for a potential role in tamoxifen outcome of ER positive breast cancer. We identified few genes and polymorphisms that hold the potential to be treatment outcome predictors and therefore should be followed up in the future within a treatment outcome context.
PREDICTION OF ADVERSE DRUG REACTIONS (University of Liverpool, Lab21 Cambridge and RBMF Stuttgart)
The work on carbamazepine (CBZ) have addressed one of the most commonly prescribed anticonvulsants and the observed hypersensitivity in susceptible patients was found to be associated with immune factors including the human leukocyte antigen alleles HLA-B*15:02 and HLA-A *31:01. In collaboration with the University of North Carolina at Chapel Hill, large inter-individual variability in response to CBZ and its toxic metabolites CBZ-10,11 epoxide and 9-acredinecarboxaldehyde has been demonstrated in unrelated individuals (N=331) of European and African ancestry. Several novel candidate genes for CBZ-induced toxicity were identified which can now be subjected to further investigations in patients with CBZ hypersensitivity and drug-exposed controls for their relevance in the susceptibility and control of CBZ toxicity. Investigations of the underlying immunological mechanism pointed to CD8 T-cells that can recognize CBZ in a HLA-A*31:01 dependent way, but the complete haplotype seems to be important for generating a polyclonal response as seen in vivo. Moreover, HLA-A*31:01 binding peptides have been identified that will now help to better understand the mechanisms of antigen presentation. Altogether, these studies have helped to further understand the mechanisms of CBZ hypersensitivity. Furthermore, we obtained initial data that pointed to a link between impaired CBZ treatment response in epilepsy and blood-brain barrier efflux ATP-binding cassette transporter C2 (ABCC2, MRP2). However, our in vitro data showed that CBZ is not a substrate for human ABCC2 and association analyses of the SANAD study (endpoint: time to first seizure and time to 12-month remission) showed no link between genetic polymorphisms -24C>T, c.1249G>A c.3972C>T and CBZ treatment outcome. Accordingly, ABCC2 appears to not play a role in CBZ treatment efficacy.
The work on the antiretroviral compound Nevirapine (NVP) used for the treatment of patients with HIV addressed its hypersensitivity reactions (e.g. fever, acute liver or skin toxicity). NVP is frequently used in developing countries and approximately 5% of the treated patients develop hypersensitivity reactions. We investigated clinical samples from Malawi in a prospective study and showed that HLA-C*04:01 predisposes to Stevens-Johnson syndrome which has been confirmed in a meta-analysis. Moreover, CD177 mRNA was elevated in cells taken from patients with acute hypersensitivity reactions and microarray analysis identified dysregulated pathways associated with hypersensitivity reaction. Our data suggest that mRNA species may be a promising avenue to pursue the identification of potential novel biomarkers of hypersensitivity.
Steroid-induced glaucoma is another frequent ADR that develops upon glucocorticoids treatment. We investigated underlying key molecules and regulatory pathways via meta-analysis of transcriptomic studies. By way of literature searches and review of publicly accessible microarray data repositories (GEO, ArrayExpress), we identified studies measuring changes in gene expression profiles in human trabecular meshwork cells upon dexamethasone treatment versus untreated controls. Raw data from three microarray studies on different Affymetrix chips were subjected to quality control, normalization and re-annotatation of gene expression sets by bioinformatics tools (R, Bioconductor) and merged to provide consistent input for further analysis for differential gene expression and regulatory networks. This set the stage for the identification of key molecular players responsible for the development of glaucoma which may serve as biomarkers in the future to allow identification of susceptible individuals prior to steroid treatment.
In cancer patients treated with 5-fluorouracil (5FU), a functional polymorphism in the multidrug resistance associated protein (MRP8)which is a member of the ATP binding cassette (ABC) superfamily of transporters, has been investigated for its role in the development of severe leucopenia. We demonstrated that rs17822471 is a risk factor for 5FU induced severe leucopenia. Notably, the ABCC11 protein expression in liver was considerably lower in carriers of the minor allele compared to homozygous wildtype allele carriers and the deleterious effect has been confirmed in recombinant expression experiments of the protein variant. Because MRP8 is expressed in normal leukocytes and marrow blasts, we propose that the T546M polymorphism leads to intracellular accumulation of the active 5FU metabolite to selectively increase the risk of leukopenia.
The promiscuous liver enzyme CES1, is involved in the metabolism of a number of important drugs including clopidogrel (antiaggregant), oseltamivir (antiviral), isoniazid (tuberculosis), irinotecan (anticancer), methylphenidate (psychostimulant) and cocaine. The role of genetic variability in CES-1 on clopidogrel pharmacokinetics and pharmacodynamics and on isoniazid pharmacokinetics was investigated. Also drug-drug interactions between clopidogrel and the anti-HIV non-nucleoside reverse transcriptase inhibitors efavirenz, nevirapine were assessed.
HEPATIC FACTORS OF INTER-INDIVIDUAL DRUG RESPONSE AND CHOLESTEROL GENES LINKED PATHOLOGIES (RBMF Stuttgart, University of Ljubljana, Lab 21)
The liver-specific organic anion transporters OAT7 (SLC22A9) was investigated for its expression in vitro localization as well as specific endogenous and xenobiotic substrate binding. We identified a novel OAT7 substrate and a non-synonymous SNP that affects OAT7 expression and uptake function. Using multivariate statistical testing, protein and mRNA expression levels were linked to non-genetic factors (e.g. age, sex, clinical information) and genetic variants. The cloning of a critical 5` upstream region of the SLC22A9 gene revealed transcription factor and nuclear receptor binding sites involved in the regulation of SLC22A9 expression in the liver.
The CYP51A1 enzyme, which is involved in the metabolism of cholesterol, was investigated in collaboration with the University of Iowa for the identification of common polymorphisms. These were shown to be associated with lower birth weight in babies and higher LDL cholesterol levels in mothers in the 2nd trimester of pregnancy. Moreover, we showed in a biochemically characterized Cyp51 +/- heterozygous mouse model, that the animals lacking one functional Cyp51 allele have an increased number of mitoses and apoptoses in the liver, even if their gross phenotype was normal.
A newly developed SteatoNet mathematical model demonstrated the association between cholesterol and non-alcoholic fatty liver disease (NAFLD). SteatoNet is based on ordinary differential equations and contains 160 metabolites, 60 proteins, 260 enzymes, 150 reactions, together with gene expression and protein regulation levels. Notably, the simplified model (22 metabolites, 51 enzymes, 26 proteins) still shows a conserved steady-state behavior highlighting the proper description of the essential regulatory networks. A major benefit is the multi-level structure (gene, transcript, enzyme, regulation, etc.) and the incorporation of liver interactions with other tissues through the blood. Validation analyses highlight the critical nature of cholesterol and ketone body transport and transcriptional regulators farnesoid X receptor (FXR), liver X receptor (LXR) and sterol regulatory element-binding protein 2 (SREBP2) as additional/novel NAFLD associated candidates.
In summary, this ITN demonstrated a broad coverage of relevant pharmacogenomics research topics addressing major diseases and ADRs as well as potential new drug targets. The results obtained fuel the promise towards improved pharmacotherapy. The aspect of better efficiency has been comprehensively addressed for tamoxifen and cytotoxic therapy in breast cancer with identified potential outcome predictors that include CYP2D6, BRCA1, miR-126 and miR-10a. A promising approach for the discovery of novel drug targets for those breast cancers for which targeted treatment is currently not available has been outlined. Aspects of safer pharmacotherapy have been addressed for frequently used drugs known to cause severe ADR such as 5FU, CBZ, NVP, clopidogrel and non-nucleoside reverse transcriptase inhibitors as well as glucocorticoids. Relevant identified drug targets for the avoidance of ADR include MRP8, CES, HLA-B*15:02, HLA*31:01 and HLA-C*04:01. Notably, specific hepatic factors and cholesterol associated pathologies have been elucidated to provide more potential drug targets and systems biology approaches revealed entire pathways to be scrutinized for drug treatment improvements. We clearly demonstrated that the exploration of new ideas and concepts as well as model systems and patient cohorts lead towards the advancement of hypothesis driven research that can provide the long-term prospects of improving patient benefits from drug treatment. Altogether this ITN collaborative effort has shown that genetic polymorphisms have a great potential as biomarkers to improve individualized treatment strategies.
Our scientific projects provided a structured high quality postgraduate training with unique opportunities in all aspects of pharmacogenomics at the leading edge of biomedical sciences and patient care. The newly acquired basic and specialized knowledge equipped 14 ESR and 3 ER for their future independent research to further excel and develop into versatile and highly competitive professionals in the public and private sectors. We capitalized on this by bringing together the complementary academic and economically oriented providers of research training from different countries including Germany, UK, Slovenia, Russia, Australia and the USA, thereby maximizing the overall quality of the pharmacogenomics research training in Europe. As a result we supply eligible professionals for future academic and private job markets who will be valuable members of the scientific community as well as industry and who will be able to take up leadership roles in this research field within European and worldwide networks to help to develop future generations of researchers capable of coping with and contributing to the medical and societal challenges of pharmacotherapy in the 21st century.
At the professional and societal level we provided greater visibility to the fact that the genome can be utilized for more accurate prediction of an individual’s drug response and selection of appropriate drug dosage. Because each Euro spent for initial treatment requires an additional Euro to cope with the management of adverse effects, millions of people in Europe and worldwide suffer from the enormous burden of morbidity and loss of human resources as well as economic misfortune and political challenge. The progress made in this ITN meets those challenges by providing further options in the application of pharmacogenomic principles by our newly trained young experts who will contribute to improve health care management of major diseases at the level of drug targets. Moreover, this can be taken forward for drug development by researchers and industry – an important issue given that pharmaceutical pipelines have been rather poor of late. This will have a profound impact on society as a whole and must be considered a major step forward in the health care management particularly in light of an ageing society and the emotional, clinical and economic burden associated with chronic illnesses. Our ITN delivers trained investigators to help maximize chances of success, promote uniform drug treatment standards throughout Europe and worldwide, and improve the competitiveness of the EU in this emerging and exciting field where opportunities for commercial exploitation exist now. This will also incite further public discussion on how the pretreatment genetic testing of patients as an a priori approach will contribute to improved treatment outcomes and reduced adverse effects, which in turn will require new responsibilities to be addressed by the European regulation.