Clinical implementation of pharmacogenetics

Introduction

It is well known that multiple pharmacogenetic biomarkers have been identified during the last decade. However, transferring these biomarkers into clinical practice has faced numerous problems [1–3], and few genetic variations have been determined during the course of treatment or during disease activity in clinical routine [4]. Four different points of view are considered: the role of pharmacogenomics for individualization of cancer treatment, the relationship of pharmacogenetics with pharmacodynamics and clinical evolution in combination transplant therapies, the needs and barriers of pharmacogenetics in clinical application, and the state of art of pharmacogenetic testing in Spain.

Most pharmacogenetic testing in oncology is carried out to predict drug efficacy, and there are multiple examples of successful clinical implementation of these biomarkers [5]. However, adverse reactions are also a limiting factor that can compromise the anti-cancer action of drugs used in the treatment of this disease, most of which are generally caused by germline mutations [6]. Therefore, risk-based personalized treatment seems to be a good strategy for cancer. Coordinated and multidisciplinary approaches could be the basis for cancer therapy individualization [7].

In recent years, there has been great improvement in transplanted patient care and in their clinical course. Better management of immunosuppressive therapy has been essential to maintain high viability of the allograft to prevent acute rejection. However, patients still cannot be stratified based on risk for rejection and individual response to immunosuppressive drugs. Polymorphisms in CYP3A5, ABCB1, and CYP3A4 genes are widely recognized and occasionally used for the optimization of transplant patient therapy [8]. However, the complexity of immunosuppressive therapies that usually combine several drugs and the pathophysiological particularities of the disease make this approach insufficient. New strategies combining pharmacogenetic, pharmacodynamic, and pharmacokinetic biomarkers to select the most effective immunosuppressive drug combination are needed. Gene-gene interactions are another point that must be considered. Using panels of donor-recipient genetic polymorphisms that also take into account gene-gene interactions is a new approach that can help predict the risk of rejection and adverse reactions [9].

Clearly, oncology is the field in which pharmacogenetics is most widely used in clinical practice. Although the main needs and barriers for pharmacogenetics application in oncology are similar to other fields, cancer has its singularities. For instance, detection of tumor-free circulating DNA, also called liquid biopsies, is revolutionizing the way tumor progression and evolution are monitored and personalized medicine are applied [10]. Detection techniques are essential to identify one mutation in a cancer cell among a high number of normal cells.

Few data are available for the study of the current implementation of pharmacogenetics in Spain. No studies have been performed, but the situation is similar into other European countries. Consortia such as Clinical Pharmacogenetics Implementation Consortium in the USA (https://www.pharmgkb.org/view/dosing-guidelines.do?source=CPIC) or the Dutch Pharmacogenetics Working Group in Europe [4] have created guidelines for clinical implementation of pharmacogenetics. These guidelines could help to implement pharmacogenetic tests in the health system. Regulatory agencies also play an important role defining mandatory, recommended, and informative biomarkers. National pharmacogenetic societies (in Spain, the Spanish Pharmacogenetics and Pharmacogenomics Society [Sociedad Española de Farmacogenética y Farmacogenómica, SEFF]) can play an active part promoting networks, guidelines for pharmacogenomics implementation, and training of health professionals.

We review the latest contributions in this field presented in the Clinical Implementation of Pharmacogenetics Symposium from the VII Conference of the SEFF held in April 20, 2015. Numerous and diverse pharmacogenetic studies are currently being implemented in clinical practice to improve patient management, and there are many successful examples of pharmacogenetic markers involved in clinical decision making. Some are briefly discussed below.

Pharmacogenomics for individualization of cancer treatment

Pharmacogenomics is finally becoming a part of oncology with implications in therapy selection, treatment avoidance, dosing, and risk prediction [11]. Despite the number of cancer treatments, few tools for individualization based on patient-specific molecular characteristics are implemented, although it is well known that multiple active anti-cancer regimes show wide intra- and inter-patient variability in both response and toxicity [12].

The logical sequence for the application of pharmacogenetics is discovery, validation, and, finally, application. Discoveries in cancer pharmacogenetics have been extensive, but more thorough understanding of the pharmacological basis of drug toxicity, efficacy, and resistance in cancer patients is needed. The presence of clinically predictive germline variants has also opened the door for objective predictors of toxicity in patients. This will allow for the development of robust risk/benefit models to facilitate the selection of the appropriate option for an individual patient between apparently equal treatments. Thus, the tumor but also the germline genome (transporters, drug metabolizing enzymes, etc.) definitively need to be studied [13, 14]. In this sense, the large knowledge gap regarding the uncovering of genes that regulate the activity (transport and metabolism) of many drugs we currently use is another barrier to clinical application [14, 15]. Genome Wide Associations Studies (GWAS) are still a crucial resource to improve our knowledge especially regarding drug toxicity, rendering important biomarker information despite a low sample number [16].

To avoid confusing “hope” with “hype”, we need to find the right biomarkers to ensure robust data set validation and apply them in patients as soon as possible [17, 18]. The Moffitt Cancer Center is an example of how to manage this new era of personalized cancer medicine and apply pharmacogenomic data by integrating the following components: a molecular diagnostics program, a clinical genomics action committee (a multidisciplinary team to discuss and evaluate each individual case from different perspectives, with different specialized oncologists, a geneticist, a biostatistician-bioinformatician, a molecular pathologist, a genetic counselor, a pharmacist, an anatomopathologist, etc.), a personalized medicine consultation service, and last but not least, a personalized cancer medicine clinical training program. In this center and according to a complete genetic screening, patients are given a “patient recommendations” report containing Food and Drug Administration (FDA)-approved therapies for the patient’s type of tumor and other types of tumors, taking into account the genomic alterations detected and the clinical trials being performed in the center where the patient was enrolled. This opens new treatment options not considered when using standard procedures. Decisions must be based on the genetic characteristics of the tumor using all the necessary techniques and knowledge.

The Clinical Risk Panel is an ongoing project at Moffitt. It is a “genomic risk panel” to ensure clinical pathway-driven care more than a gene-driven care, that is, using all information known with respect to drug pathways. This care is designed to adhere to cancer risk guidelines, to identify underlying predisposition to severe toxicity, and to mitigate the risk of untoward drug effects. It contains more than 250 gene assessments of actionable variants associated with cancer predisposition, cardiac/nerve/bone marrow failure syndromes genes, and pharmacogenomics. Its goals are to implement broad genetic testing processes to meet and exceed national cancer screening guidelines and to promote risk reduction during chemotherapy for adverse drug effects to minimize toxicities, enhance value, and improve the quality of patient care. We need to be able to distinguish and to move from clinical validity to clinical actionability and finally to clinical utility [19].

In conclusion, pharmacogenomics is a reality in oncology, but a multidisciplinary effort is still needed to integrate techniques and knowledge and consider the patient as a whole: the tumor and the body as a whole.

The data presented herein is based on the lecture that Howard McLeod gave at the VII SEFF Conference in Madrid in 2015.

Pharmacogenetics: pharmacodynamics and clinical course in combination therapies

There has been recent improvement in transplant patient care and in their clinical course. Optimal immunosuppressive therapy is essential to maintain high viability of the allograft to prevent acute rejection; however, we still cannot stratify patients based on risk for rejection and individual response to immunosuppressive drugs. These drugs are characterized by a narrow therapeutic range and high inter-individual variability in pharmacokinetics and pharmacodynamics, which may result in an increased risk of therapeutic failure if these agents are used at standard doses in all patients [20].

Drug concentrations in blood correlates with clinical outcomes, making therapeutic drug monitoring an important factor to individualize drug therapy and to obtain the best balance between efficacy and adverse events [21, 22]. Each patient must be considered individually, and thus, several factors such as age, body weight, enzymatic activities, kidney or liver function, concomitant therapies, and more recently, pharmacogenetics should be taken into account [23]. Thus, immunosuppressant dosage and therefore the inter-individual variability linked to graft rejection may depend on genetic diversity. There are many pharmacogenetic studies on the association of genetic polymorphisms and the response to immunosuppressive drugs and graft rejection, on the predisposition to genetic interactions between donor and receptor, and on single-nucleotide polymorphisms (SNPs) in drug metabolism [24].

Most of these pharmacogenetic studies have focused on SNP influence on the disposition and dose requirements of cyclosporine and tacrolimus [25, 26]. Individual differences in tacrolimus doses are partly due to polymorphisms of genes encoding proteins that play key roles in their absorption and distribution (P-glycoprotein) and in their metabolism (CYP3A4 and CYP3A5) [27, 28]. In this sense, blood concentrations of tacrolimus and dose requirements are strongly associated with CYP3A5 polymorphism [29]. Several studies have demonstrated that genotyping CYP3A5 may identify patients at risk of tacrolimus underexposure [30, 31] and that graft recipients carrying the CYP3A5*1 allele – either donor or recipient – require higher doses of tacrolimus than CYP3A5*3 homozygous patients [32]. Li et al. [33] developed a pharmacogenetics-based dosing model (including CYP3A5 genotype) for the prediction of the tacrolimus maintenance dose in renal transplant recipients, which may be useful to help clinicians prescribe the initial tacrolimus dose with greater safety and effectiveness. However, studies on the mechanisms associated with increased rates of acute rejection have reached contradictory conclusions [28, 34]. The ABCB1 genotype of either recipient or donor was associated with the distribution of tacrolimus, the incidence of acute rejection or other adverse events in kidney transplant recipients [27]. However, findings on the influence of ABCB1 polymorphisms on cyclosporine or tacrolimus disposition and efficacy are inconsistent [35, 36].The association of other polymorphisms with acute rejection have been also evaluated: rs1800795 in IL-6 [37], TNF-A-308G/A [38], TLR4 rs10759932 [39], GSTM1 and GSTP1 [40], and MRP2 C24T [41]. Unfortunately, mainly due to limitations in study design, results obtained are conflicting and cannot adequately assess the role of pharmacogenetics on the individual pharmacokinetic and pharmacodynamic profile of immunomodulators. Although many reports suggest promising genotypes for predicting acute rejection, immunosuppressant disposition, and adverse effects, these findings have yet to be implemented in routine clinical practice. Obstacles to the successful clinical use of predictive genetic polymorphisms include [42]

• many genetic influences that are still largely unknown;

• the relationship among gene, protein, and function, which is complex and often unstable;

• some studies have often ignored donor genotypes;

• studies have used a small sample and have not evaluated gene-gene interactions;

• non-pharmacogenetic factors: ethnicity, concomitant medication, allo-antibody production [43].

Therefore, it is interesting to consider new strategies that combine pharmacogenetic variables with pharmacodynamic and pharmacokinetic biomarkers to increase the choice of immunosuppressive treatments. Considering the complex physiopathology of graft rejection, donor-receptor gene interaction should also be considered to prevent this clinical event. Thus, when administering combination therapies, the genetic polymorphisms of each metabolizing enzyme or transporter protein with a potential effect on any of the drugs need to be considered – both alone and in combination. Therefore, new approaches predicting the risk of rejection must be based on panels of donor-recipient genetic polymorphisms that may reflect gene-gene interactions. Furthermore, genetic information needs to be evaluated together with proteomic and metabolomic data on specific populations.

The data presented herein is based on the lecture that Mercè Brunet gave at the VII SEFF Conference in Madrid in 2015.

Needs and barriers in pharmacogenetics implementation

Cancer is possibly one of the areas in which clinical implementation of pharmacogenetics has been faster [44]. A complete study of tumor biology and its environment can help select the best treatment, decrease its toxicity, and avoid the development of tumor resistance. Thanks to simple genetic tests, the selection of a specific treatment to treat a particular type of tumor has become a reality in oncology. There are several biomarkers that are essential for the selection of cancer treatment: KRAS for cetuximab treatment in colorectal cancer, HER status in breast cancer, ALK mutation, and crizotinibin lung cancer [5]. Another aspect to be considered is that patients play an increasingly active role in their treatments. In this era of increasing knowledge, patients demand information about different treatment options, and studies confirm that many would accept a delay in treatment to take a pharmacogenetic test if it could help them opt for a more personalized approach.

What does a biomarker need to do to be introduced in clinical practice? In short, a biomarker is only useful when it provides reliable, actionable, and predictable information to imply an alternative drug or drug-dosage regime based on health data in a cost-effective way, as several health technology approaches indicate [45–48].

However, certain needs and barriers are identified in cancer pharmacogenomics. The process of biomarker discovery and validation is often slow and problematic, and clinical implementation is always challenging. There is a lack of consistent evidence, and this results in clinical uncertainty produced by inadequate validation of biomarkers and inadequate evidence of clinical utility [3].

On the one hand, there are limitations in technology and study design from the very beginning of the drug development process. Ideally, the study of pharmacogenomic biomarkers should be carried out in the early steps of drug development from phase I and II trials. Nevertheless, even when this is prospectively undertaken, inappropriate preclinical trial design with inadequate and/or inaccurate animal models often limits the scientific value of the results. Both well-designed, forward-thinking trials and population-based studies are necessary to fully implement pharmacogenomics in clinical practice [49]. On the other hand, molecular data are difficult to translate into clinical practice because of disease complexity and inter- and intra-tumor heterogeneity. Large tumors are genetically diverse, and their molecular characteristics can change as the disease evolves and also in response to any given treatment leading to drug resistance and treatment failure [50]. For this reason, it may be necessary to repeat pharmacogenetic testing at different disease stages and probably after each treatment failure.

Promising new techniques and biomarkers are being studied. Circulating tumor cells and circulating free DNA are alternatives to invasive procedures, and their analysis may become a prognostic factor, acting as a “liquid biopsy” [51]. However, detection in the blood flow poses a technological challenge. Beaming digital polymerase chain reaction has demonstrated to be highly sensitive to find a few mutated molecules among a huge number of wild-type colonies. Evidence also indicates that microRNA can play a role as a diagnostic, predictive, and prognostic biomarker in cancer [52].

Nevertheless, the results of pharmacogenomic tests and their implications are not always appropriately understood by clinicians and other health-care professionals, which can lead to data misinterpretation and miscommunication. For cancer patients to benefit from these tests, the oncologist must first adopt the tests, be able to interpret results, and find them useful in managing patient care [53].

Finally, the implementation of pharmacogenomics also faces legal barriers. Regulatory agencies such as the US FDA and the European Medicines Agency (EMA) have officially recognized the importance of pharmacogenetics and pharmacogenomics. However, lack of regulatory and legislative harmonization between different agencies can hamper clinical trials and make reaching a consensus more difficult.

In summary, it is necessary for different research groups, health-care professionals, and regulatory institutions to work together to include pharmacogenomic testing in the design of future clinical trials and rigorously and consistently implement the knowledge pharmacogenomics provides.

The data presented herein is based on the lecture that Jesús García-Foncillas gave at the VII SEFF Conference in Madrid in 2015.

State of the art in pharmacogenetic testing in Spain

When the human genome was finally decoded a few years ago, great expectations arose. Scientists anticipated unraveling the origin of illness and developing new target-specific drugs; concepts such as personalized medicine and pharmacogenetics gained fame and relevance, and the number of published papers on the subject increased exponentially [54]. However, lately, it has come to a standstill. Research breakthroughs have not been as numerous and relevant as expected and the application of pharmacogenetics in clinical practice has been slow and problematic. The reasons for this are probably multifactorial. Spain is not isolated and most factors affecting clinical implementation of pharmacogenetics are common to other countries. Regulatory agencies, research consortia for clinical implementation of pharmacogenetics, and national or even international pharmacogenetic societies can play relevant roles in clinical implementation.

Regulatory agencies, including the US FDA and the EMA, now require that several drugs include pharmacogenomic information in their labeling [55, 56]. However, labeling does not always include specific actions based on the biomarker information, and there are important discrepancies between the specific recommendations issued by agencies. Nevertheless, national and international consortia are addressing some of the barriers to the implementation of pharmacogenetic testing in clinical practice. For instance, the Clinical Pharmacogenetics Implementation Consortium, the Pharmacogenomics Research Network, and the Pharmacogenomics Knowledge Base (PharmGKB) were created to provide peer-reviewed, updated, evidence-based guidelines for gene/drug pairs [57]. Currently, these guidelines can be freely accessed via the PharmGKB website, and they probably constitute one of the most widely used resources by researchers and health-care professionals. The Royal Dutch Association for the Advancement of Pharmacy also established the Pharmacogenetics Working Group to develop pharmacogenetics-based therapeutic (dose) recommendations. After systematic review of the literature, in 2011, they provided recommendations for 53 drugs associated with gene coding for 12 enzymes and proteins known to be involved in different drug pharmacokinetics and pharmacodynamics pathways [4]. In France, an online directory listing all homologated laboratories that perform pharmacogenetic tests nationwide was created to promote their clinical implementation and encourage collaboration between professionals [58]. The website also includes FDA and EMA genotyping information as well as practical recommendations for pharmacogenomics-based prescriptions issued at the European Science Foundation conference in 2010 [59]. Other European collaborative efforts include the Pharmacogenetics and Pharmacogenomics research network of the European Federation for Pharmaceutical Sciences and the establishment of the European Society of Pharmacogenomics and Personalized Therapy.

In Spain, SEFF was constituted in 2005 to offer a platform for researchers and institutions in the field and promote the sharing of knowledge and the development of integrated projects. As a SEFF initiative, in 2014, a directory of pharmacogenetics laboratories in the country was established to provide information about their services and contact information through the SEFF website [60]. According to the data received from the members of the society, HLA-B*5701, UGT1A1*28, and IL28B polymorphisms are the most extended tests in Spanish pharmacogenetic laboratories. However, laboratory portfolios fail to cover all the determinations recommended (or even required) by the regulatory agencies. There are undoubtedly other laboratories around the country that provide these services and do not belong to this network, mainly pathology laboratories. It is clear that all the determinations that must be done are in fact being carried out.

There are also wide differences in the incorporation of pharmacogenomics into clinical practice between regions and hospitals in the Spanish National Health System (NHS). The causes for this include the particular territorial organization in Spain where health competencies are transferred to autonomous regions, the lack of standard criteria to decide which tests must be carried out, and that pharmacogenetic testing is not perceived as necessary by authorities of the autonomous regions, and as a consequence, hospital managers make it difficult to incorporate them into daily practice. Each hospital has its own criteria (or no criteria at all) with respect to which tests should be performed. Even when the institution recognizes the need for a specific test, there is usually no consensus as to where the testing should be carried out, leading to disagreements between different in-house services and laboratories. In addition, there is no standard validated technology for genomic determination and health-care professionals often lack adequate pharmacogenomic training.

Both scientific societies and political institutions should become involved in and pursue initiatives to help incorporate pharmacogenetics into clinical practice. SEFF must take a leading role as the referential scientific institution. It should provide validated guidelines and recommendations and establish the minimally required quality criteria for the techniques. It should also take an active role in training professionals and promoting the development of national and international pharmacogenomics networks. From the political and official point of view, NHS institutions should establish which tests are necessary before starting a treatment, provide appropriate funding, and ensure that hospital managers address all these issues at the local level.

Only when the incorporation of pharmacogenetics becomes an official NHS priority and initiatives are adequately supported by scientific knowledge and powerful leading societies will it become an essential part in daily clinical practice.

The data presented herein is based on the lecture that Luis López-Fernández gave at the VII SEFF Conference in Madrid in 2015.

General conclusions

• The development of new techniques and the use of appropriately validated biomarkers will definitely improve cancer treatment in the coming years.

• Pharmacogenomic biomarkers should be combined with pharmacodynamic and pharmacokinetic variables to increase the choices and the efficacy of immunosuppressive treatments.

• Pharmacogenomic biomarkers need to be cost-effective and to change the clinical practice to be implemented.

• It is fundamental that health professionals, scientific societies, and regulatory agencies work together to ensure fast and adequate implementation of pharmacogenomics.