Multiple Sclerosis: Advances in Clinical Trial Design, Treatment and Future Perspectives

Genetic polymorphisms that influence the activity of proteins regulating the pharmacodynamic and pharmacokinetic properties of drugs are key contributors to the variability in response to drugs between individuals. For example, cytochrome P CYP enzymes are the major enzymes involved in drug metabolism. Polymorphisms in these enzymes can result in either ultrafast metabolism of therapeutic drugs, thereby limiting their efficacy, or poor metabolism, thereby increasing the risk of toxicity [ 17 ]. Identification of genetic differences associated with variability in drug response would allow better-informed decisions regarding choice of treatment.

The aim of this article is to review basic definitions of genetic biomarkers and how they are being used in other disorders and then to review their current status in MS. Pharmacogenomics is the study of how genes affect drug response. Both inherited and acquired genetic variations may be involved. A genomic biomarker could, for example, be the degree of expression of a gene, the function of a gene, or the regulation of a gene [ 18 ]. Pharmacogenetics is a subset of pharmacogenomics.

It involves variations in DNA sequences as they relate to drug metabolism and response [ 18 ]. A genetic variation may range from a single nucleotide polymorphism SNP to loss of part of a chromosome. Key applications for pharmacogenetic biomarkers are the identification of responders and nonresponders to medications, avoidance of adverse events, and optimization of drug dose. Almost all therapeutic areas have at least one drug for which pharmacogenomic guidance exists, including psychiatry, rheumatology, gastroenterology, endocrinology, and dermatology; however, by far, the best represented therapeutic area is oncology Table 2.

The vast majority of drugs have guidance concerning variations in DNA sequence which relate to drug response, which are therefore classed as pharmacogenetic biomarkers. Examples of drugs with pharmacogenetic guidance are presented in Table 3. Pharmacogenetic information may pertain to many aspects of drug use, including definition of specific patient populations for which a drug is indicated or contraindicated, for which dose adjustments could be necessary, and also in which potentially serious adverse events could occur.

For the majority of oncology drugs, the inclusion of biomarkers in drug labels generally corresponds to a requirement or recommendation for genetic testing; however, in other therapeutic areas, there is no specific guidance on what actions should be taken based on biomarker information. In the area of infectious diseases, there are at least 17 drugs with pharmacogenetic information in their labeling [ 19 ].

For the majority of these, the guidance relates to the presence of a pharmacogenetic marker and the increased likelihood of adverse effects associated with treatment. Abacavir is a synthetic carbocyclic nucleoside analog with inhibitory activity against human immunodeficiency virus. The labeling of abacavir states that all patients should be screened for the presence of the allele before initiating or reinitiating abacavir therapy, unless the patient has a previously documented allele assessment [ 20 ].

Abacavir is contraindicated in patients who are positive for the allele because of the high risk of experiencing a hypersensitivity reaction [ 20 ]. Systematic analysis has indicated that testing for before initiating abacavir is cost-effective [ 21 ], and companion diagnostic tests are available for this allele, although no specific test is recommended in the drug label [ 22 ]. In oncology, there are at least 54 drugs with pharmacogenetic information in their labeling [ 19 ].

For over half, the pharmacogenetic information relates to a specific indication or usage [ 19 ].

It is not recommended for use in patients with wild-type BRAF melanoma, as safety and efficacy have not been demonstrated in this population. Accordingly, the requirement for the presence of this mutation is specified in the indications for vemurafenib within the drug label [ 23 ].

Despite the codevelopment and coapproval of the drug and diagnostic test, the label does allow for another FDA-approved test to be used if desired [ 23 ]. Only a handful of hematology drugs have pharmacogenetic information in their labeling, mostly relating to warnings and precautions [ 19 ]. Lenalidomide is one of the few drugs with pharmacogenetic information pertaining to its indication.

In an initial study involving patients with myelodysplastic syndromes who did not respond to treatment with recombinant erythropoietin, a greater proportion of patients with a 5q deletion no longer needed red cell transfusions after being treated with lenalidomide compared with patients with other karyotypes [ 24 ]. This observation was confirmed in further studies, leading to a defined indication for lenalidomide for the treatment of transfusion-dependent anemia due to International Prognostic Scoring System low- or intermediate-1 risk myelodysplastic syndromes associated with a 5q deletion abnormality with or without additional cytogenetic abnormalities [ 25 ].

No specific guidance is provided regarding testing for 5q deletion. Some drugs used to treat heritable genetic diseases, such as cystic fibrosis and certain inborn errors of metabolism, also have pharmacogenetic information in their labeling.

Specific guidelines are available to facilitate the interpretation of genotype tests to guide ivacaftor therapy [ 27 ]. In the field of schizophrenia, aripiprazole, an atypical antipsychotic, has pharmacogenetic information in its label relating to its use in patients with poor CYP2D6 metabolism. In endocrinology, glimepiride, a sulfonylurea indicated as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes mellitus, has guidance in its label relating to its use in individuals who are glucosephosphate dehydrogenase-deficient, owing to the risk of hemolytic anemia.

Finally, in the area of transplantation, mycophenolic acid is contraindicated in patients with hereditary deficiency of hypoxanthine-guanine phosphoribosyltransferase, as use of mycophenolate in such patients may cause exacerbation of disease symptoms. To date, the routinely used pharmacogenetic biomarkers reflect relatively simple, well-defined genetic changes.

Considerable efforts are now being made to establish pharmacogenetic biomarkers for polygenic diseases, such as cancer, chronic kidney disease, MS, cardiovascular disease, and neuropsychiatric illnesses. In these indications, numerous genes and their products are potentially involved in disease manifestation, drug metabolism, and drug mechanism of action, and the individual contribution of each gene may be small [ 28 — 30 ]. In , a systematic review of pharmacogenetic studies found that most had examined genetic variations in drug targets i.

However, this is no doubt a consequence of early studies using a candidate-gene approach to identify potential genes. The candidate-gene approach directly tests the effects of genetic variants of a potentially contributing gene in an association study. A higher frequency of a particular allele or genotype in a series of individuals with a specific disease or phenotype can be interpreted as meaning that the allelic variant or genotype is associated with that disease or the disease phenotype [ 31 ].

A key limitation of this approach is that it is dependent on knowledge of the biology of the disease being investigated in order to identify candidate genes for testing. In contrast, GWAS examine multiple genetic variants, typically SNPs, in different individuals to see if any variant is associated with a trait or response to a drug. The advantage of the GWAS approach is that it does not require specific knowledge of the disease in question; however, it is not without its own challenges.

For example, effect sizes for common variants are typically modest and large sample sizes are required to detect associations; single genomic regions can harbor variants with weak effects and also large effects; some associations implicate non-protein-coding regions; and correlations between genetic variants and phenotypes can be limited by the accuracy and validity of the phenotypic measurement [ 32 ]. For pharmacogenetic information to be routinely used in clinical practice, the genetic markers will require validation in large cohorts, particularly in cases where response is found to be influenced by numerous genetic variants.

Predictive diagnostic tests also require validation before use, and regulatory frameworks are evolving to ensure that diagnostic tests accurately identify the intended population for a corresponding treatment. Over the last decade, considerable effort has been made to identify pharmacogenetic markers in the field of MS. To date, efforts have been focused on the identification of markers that determine drug response, and there are no published data relating to pharmacogenetic markers to predict adverse drug reactions.

A review conducted in outlined a number of candidate genes for future study but noted the challenges in identification of pharmacogenetic markers for adverse drug reactions including the relative rarity of these reactions, the need for accurate characterization of the reaction, and accurate phenotyping of the patient [ 14 ]. Regarding pharmacogenetic markers for drug response, longitudinal data on drug response and disease worsening are necessary for their identification and potential validation.

At the time of writing, there were no data available on pharmacogenetic markers for natalizumab or fingolimod. Although there are no universally accepted measures of response for DMTs, most studies evaluate a clinically event-free status, for example, relapse-free status with no confirmed worsening on the Expanded Disability Status Score, as a measure of a positive response.

In addition, this approach unnecessarily exposes patients to potential adverse effects and can add to the substantial economic burden carried by healthcare providers. Not all of the significant associations have been validated, and further studies in large populations with independent validation are warranted.

These have been carried out in initial populations of between approximately and patients, with follow-up to determine response for up to 4 years. Although the findings of the earliest study were not independently validated, subsequent studies were validated in independent populations to further test their significance.

Account Options

Surprisingly, there has been little overlap across studies with regard to the genes that have been implicated as being significantly associated with treatment response. For example, the most recent study did not find an association for the genes for GPC5 and SLC9A9 [ 58 ], as was identified by two previous studies [ 54 , 57 ]. The reasons underlying the lack of consistent findings across the studies are unclear but may reflect the lack of an unequivocal definition of responder status, small sample sizes, and different populations.

It is also possible that the contribution of single alleles of candidate genes is very small and that combinations of alleles should be studied together to identify markers for therapeutic response. Multilocus analyses have been attempted and some have identified sets of alleles that show significant associations. Smoking and middle ear disease. Otolaryngol Head Neck Surg , , pp. Statistical and design considerations for multiple sclerosis clinical trials.

Surveillance of clustering near point sources. Statistics in Medicine , 15, pp. Time series models with univariate margins in the convolution-closed infinitely divisible class. Journal of Applied Probability , 33, pp. Zidovudine absorption and small intestinal function in HIV seropositive patients. Approximations to multivariate normal rectangle probabilities based on conditional expectations.

Journal of the American Statistical Association , 90, pp. An assessment of pNEM: A Bayesian analysis of bivariate survival data from a multicentre cancer clinical trial. Statistics in medicine , 14, pp. Bayesian multivariate spatial interpolation: A bootstrap based on the estimating equations of the linear model. Biometrika , 82, pp. Freezing influences the healing of rabbit medial collateral ligament autografts. Government, university, and community collaboration to develop a cardiac surgery database and report of morbidity and mortality while waiting for open heart surgery. Medinfo , 8 Pt 2, p.