Role of pharmacogenetics in the use of CNS drugs: from drug pipeline to primary care?

Pharmacogenetics uses genetic analysis to explain interindividual variation in drug response, both in terms of safety and efficacy. Exact definitions vary widely (e.g., whether pharmacogenetic tests include those analyzing metabolic biomarkers rather than DNA) [1], but a broad definition of the field reveals a large number of applications for pharmacogenetic technology [2]. These include optimizing the drug development pipeline of pharmaceutical firms by screening out drug candidates likely to cause toxicity prior to clinical trials, and stratifying patient populations in trials. In the clinic, potential applications include allowing precise diagnosis of pathologies by physiological process rather than phenotype, as well as prediction of adverse drug reactions for licensed drugs prior to prescribing.

In recent years, there has been considerable academic and clinical interest in the application of pharmacogenetics to improve the way in which drugs are prescribed in neurology and psychiatry. This has been spurred by growing numbers of genotype–phenotype associations as the use of genotyping technology continues to expand. For example, the T102C allele of the serotonin (5-hyroxtryptamine [5-HT]) receptor 2A has been associated with susceptibility to tardive dyskinesia, while the A1 allele of the dopamine receptor gene DRD2 has been associated with a higher likelihood of positive response to treatment with the antipsychotics haloperidol and nemonapride [3]. Until advances in gene cloning started to reveal more about the genetics of neurotransmission, the brain had remained a blackbox, in which the modes of action of some pharmacological compounds were largely unknown. The hope now is that brain function and disease will be increasingly understood in detailed molecular terms, as functional genomics and neuroscience research support advances in pathophysiology [4,5]. Furthermore, with clearer understanding of interindividual variation in molecular structures of drug receptors, transporters or metabolizing enzymes, predictive tests will facilitate reduced occurrence of toxicity and nonresponse in the clinic. As a result, advocates of pharmacogenetics hope that treatment regimes can be optimized, rationalized and personalized, resulting in improving patient compliance and greater cost–effectiveness [4]. So large is the expected potential, that some analysts suggest CNS applications will lead the growth of pharmacogenetic testing as a whole, with three-quarters of genetic tests being in neuropsychiatrics by 2013 in a market worth tens of billions of dollars per annum [101].

So, is pharmacogenetic testing for CNS applications going to become a clinical reality in the near future? Certainly research interest in the field is booming. A literature review of genes with pharmacogenetic relevance carried out by Goldstein and colleagues suggested that the CNS is the most target-rich pharmacogenetic environment at present, with a dozen or so genes of interest validated by at least two studies [3]. Furthermore, a recent survey of public sector research institutes in the USA, Japan and the European Union (EU) found that a third of responding laboratories with a confirmed interest in pharmacogenetics were studying drug response relating to CNS targets, second only to the 55% of laboratories involved in studies of drug-metabolizing enzymes [6]. Pharmacogenetic testing is already widely used to support clinical development of new drugs by pharmaceutical firms [102,103]. Yet, detailed investigation of the biotechnology industry’s development pipeline yields only a small number of distinct products relevant to prescribing decisions using drugs acting on the CNS [6]. These include tests to predict patient response to vilazodone in treatment for depression, and patients’ risk of developing agranulocytosis when prescribed clozapine, both being developed by Clinical Data Inc. (MA, USA). In addition, Prediction Sciences’ (CA, USA) GeneR panel tests are being developed to aid the selection of the most appropriate medicine for treatment of depression, bipolar disorder and schizophrenia. However, all these tests remain at present in the developmental stages. Beyond drug targets within the CNS itself, by far the most interest in CNS-relevant pharmacogenetics has focused on known genetic variations of the cytochrome P450 (CYP) family of drug-metabolizing liver enzymes (implicated in the metabolism of many drugs, including widely used antidepressants, such as fluoxetine, sertraline and paroxetine). Interindividual genetic differences in patients relating to these enzymes are responsible for some patients being ultrarapid metabolizers of drugs, who are as a result less likely to obtain therapeutic benefits from standard dosages. Deficient metabolizers are at risk of toxicity if drugs build up in the body. As a consequence, a number of companies aim to develop genetic tests to characterize patients’ CYP genotypes. Perhaps the most well known of these, the DNA–mircoarray-based Amplichip^®, is already commercially available from Roche (Basel, Switzerland) in the USA and the EU. However, at an estimated $400 per test in consumables alone [ROCHE, PERS. COMM], its use in the clinic in the short term will doubtless depend on studies by health economists. Furthermore, even in the USA, proponents of CYP laboratory testing services are in the early stages of establishing themselves as a community and use is yet to commence in the UK’s National Health Service, despite it having a relatively well-developed network for genetic testing services [103,104]. Thus, while the science may be of interest, the technology and healthcare infrastructure to conduct these tests in support of clinical decision-making has yet to develop.

Lessons from the introduction of other pharmacogenetic tests, as well as genetic tests for inherited disease, suggest that a number of important barriers need to be overcome before the technology can become widely used in neurology and psychiatry [102–104]. These include, but are not limited to, those described in the following sections of this review.

Research

Causative associations remain to be proven in many cases. In general, replication of results in pharmacogenetic studies and research on the differential responses between different population groups require larger clinical studies, perhaps even building and comparing genetic databases from different countries [7]. The construction of such databases raises substantial patient privacy and consent concerns that may, in turn, increase regulatory hurdles for researchers [105]. Some have even suggested that those searching for significant genotype–phenotype associations will run into the same problem as researchers looking for common disease associations; that is, that phenotypes can only be explained when looking at combinations of genotypes [104]. This is already becoming evident, as a recent review of pharmacogenetic targets for drug transporters (e.g., the dopamine transporter gene SLC6A3 or serotonin transporter [SERT]) suggests that existing studies are not conclusive [8].

Development & regulation

Genetic tests for inherited disease, which are mainly used for clinical conformation of affected patients, have often not been subject to stringent regulation. However, with commercially produced kits or devices intended for potentially life-saving pharmacogenetic interventions, the need for proof of clinical validity in specific populations will be more compelling [9,105]. The resulting increases in trial costs and extended development times may reduce the commercial viability of some projects.

Clinical utility

Once a scientifically validated genotype–phenotype association is found, a number of factors may impinge on its clinical relevance. Age, illness and nutrition are variables that can all alter an individual’s response to the same drug at different times in their life [104]. Similarly, drug–drug interactions may produce variation as important as interindividual genetic differences and could make pharmacogenetic test results falsely reassuring. For example, patients treated with standard doses of fluoxetine and paroxetine reportedly exhibit the metabolic phenotype of CYP-deficient individuals [4]. Within the context of healthcare systems, a wide range of further factors will be important in determining utility [103,104]. These include the speed at which test results will be available, the frequency of individuals found with relevant variations and cost of not using a test (adverse drug reactions, nonresponse), the cost of the test, and whether the test is a compliment or a replacement to existing monitoring procedures [103]. Furthermore, at the level of clinical practice, a lack of evidence surrounding the benefits of testing, the relative value of pharmacogenetic information, and concerns regarding increasing time and workload burdens, may limit the actual usefulness of specific tests. As a consequence, establishing clinical use and creating a demand for pharmacogenetic tests is far from guaranteed, even when a good association has been established.

Education

The above factors suggest that education alone is not sufficient to ensure the uptake of pharmacogenetic tests where suitable. Nonetheless, many of the professionals in practice today have little or no grounding in genetics and, therefore, may be unsure when a test is warranted, or have difficulties in interpreting test results [102]. Creating the capacity to provide medical school training or continuing professional development courses in pharmacogenetics will be difficult in the short term, as expertise is nonexistent outside a few centers of research excellence [10].

Commercial incentives

At present, industry is playing the leading role in developing pharmacogenetic tests. Some 30 small firms are working on a range of tests, mainly for already licensed medicines. By contrast, large pharmaceutical companies are working on companion diagnostics for new drugs and have little interest in tests for established products [2,6]. Under these circumstances, there is a danger that CNS-related pharmacogenetic tests for widely used drugs will not reach the market owing to the lack of a significant commercial incentive and insufficient interest and investment from large companies.

Conclusions

Pharmacogenetics may hold promise for the future in neuropsychiatric fields, but it is more likely to provide an incremental advance, representing another tool in the medical armamentarium, rather than a revolution in treatment. Furthermore, the complexity of genotype–phenotype interactions and particular difficulties in neuropsychiatric treatment, such as compliance and multiple medicine usage, mean that determining which interventions should be adopted may be a long and difficult task. Combined with the health infrastructural barriers identified above, the evidence suggests that pharmacogenetics will not make major contributions to clinical practice in the short or medium term without significant coordinated efforts by policy makers, opinion leaders and, in some cases, industry, to ensure that useful tests are widely adopted. This is largely because the benefits of pharmacogenetic advances may only be realized at the health system or societal level. As a consequence, a comprehensive public policy framework will need to be adopted that will support industrial innovation, prevent market failure, protect the safety and rights of the public, support the clinical adoption of pharmacogenetics, and create the required health service infrastructure. The role of practioners at all levels (specialist clinicians, pharmacologists and those in primary care) in promoting best practice and establishing a strong evidence base will also be essential in order to ensure that pharmacogenetics can advance, albeit slowly, across a wide front. Despite the lack of immediate benefit, it is essential that professionals maintain an interest and awareness of the field, as cumulative advances may, in time, build into a significant domain of clinical knowledge.