Rheumatoid arthritis (RA) is the most common form of chronic inflammatory joint disease, affecting approximately 1% of the population [1]. No treatment cures RA. Therefore, reducing disease activity with early therapeutic intervention is key to minimizing the joint damage and functional decline [2, 3]. Initiation of therapy with disease-modifying antirheumatic drugs (DMARDs) within a few months after the diagnosis of RA is essential. A delay of 3 months in the introduction of medication has been shown to result in substantially more radiographic damage at 5 years [4].
Treatment of RA usually follows a stepwise approach (Figure 4.1) [5]. First, patients are treated with monotherapy DMARDs, with methotrexate (MTX), administered weekly in low doses, being the most commonly used drug [6–8]. If there is insufficient response and/or adverse drug events, the DMARD may be switched to another DMARD, a second DMARD may be added to the monotherapy, or therapy may be changed to a newer subgroup of DMARDs, so-called biological agents, either alone or in combination with other DMARDs [9–12]. Currently available biological agents inhibit the actions of tumour necrosis factor (TNF)-α and interleukin 1 (IL-1), two proinflammatory cytokines thought to play a pivotal role in the pathogenesis of RA [13]. More recently, drugs targeting other cytokines or altering the activation of T cells or eliminating pre-B cells have gained interest for treating patients with RA. Examples of these new agents are anti-CTLA4Ig (abatacept) and anti-CD20 (rituximab) [14–16].
Despite the fact that the understanding of the pathophysiology of RA has led to different treatment options and strategies for patients with RA, the response to treatment with DMARDs is often suboptimal. For example, only 46% of patients show a good clinical response with MTX monotherapy and 30% discontinue treatment because of toxicity, whereas for anti-TNF (adalimumab) therapy approximately 60% show good response and 22% discontinue treatment [11, 17, 18]. Remarkably, studies in the field of identifying determinants for effective and safe drug treatment for individual patients with RA are still scarce. To date, the choice of drug for therapeutic intervention is made empirically.
A key challenge is to improve drug therapy by targeting DMARDs in RA to those patients who are most likely to respond, thereby predicting the individual response with maximum efficacy and avoiding toxicity [13, 19–21]. The interindividual variability in toxicity and drug response has become the basis of current pharmacogenetic research. This chapter will discuss the principles of pharmacogenetics, address genetic variability in genes contributing to outcome in DMARD treatment and, finally, will consider future research perspectives and implications of pharmacogenetic testing in improving RA therapy.
PHARMACOGENETICS
Pharmacogenetics is the field that studies the influence of variations of DNA sequence on drug response [22]. Single nucleotide polymorphisms (SNPs) represent the most abundant source of genetic variation in humans. An SNP is a genetic variation characterized by a single nucleotide base change due to alteration, deletion or insertion of the base. Genetic variants are designated as polymorphic when the minor allele frequency of the SNP is at least 1% and the allele leads to small phenotypic changes.
Single nucleotide polymorphisms occur every few hundred bases in promoter regions, coding and non-coding sequences. Functional SNPs can alter promoter activity (regulatory SNPs), DNA, pre-mRNA conformation or mature RNA (alternative splicing), and they can influence the function or expression of the gene product – the protein [13]. In addition, copy number variants (CNVs) – defined as DNA segments that are 1 kb or larger in size present at variable copy number in comparison with reference genome – have also been shown to lead to phenotypic changes as a result of altering gene dosage, disrupting coding sequences
Establish diagnosis early Document baseline disease
activity and damage Estimate prognosis
Periodically assess disease
Change or add DMARDs
biological DMARDs No previous MTX treatment
MTX monotherapy
monotherapy
other monotherapy combination therapy combination therapy
combination therapy Suboptimal response
to MTX Adequate response with
decreased disease activity
Inadequate response (ongoing active disease after 3 months of
maximal therapy) Initiate therapy
Begin patient education Start DMARD therapy within 3 months
Consider NSAID Consider local or low-dose systemic
corticosteroids
Start physical therapy or occupational therapy
Figure 4.1 Stepwise approach of treating patients with rheumatoid arthritis. DMARD, diseasemodifying
or perturbing long-range gene regulation [23]. Moreover, variable numbers of tandem repeats (VNTR), especially in the promoter region of a gene, have shown to influence gene expression and are another source of genetic variation.
The current strategies used in pharmacogenetic studies include association studies of drug response with SNPs in ‘candidate genes’, which are genes selected on the basis of the pharmacokinetic and pharmacodynamic properties of the drug under study. Another strategy is the genome-wide analysis of SNPs, for example between responders and non- responders. This latter approach intends to associate independently evenly distributed SNPs across the human genome in (or in linkage with) candidate genes with response, without using any a priori knowledge about the pharmacology of a drug.
Although the candidate gene approach is appealing, it may fail for several reasons. This method does not take into account the potential role of other genes, including those genes whose function is not yet understood. It also does not account for gene duplications and other mechanisms that alter protein function, for example post-translational modifications. On the other hand, genome-wide analysis has the advantage of considering genes whose function in relation to the studied phenotype is not yet understood or recognized, but has limitations with regard to generating relatively high costs and requiring more sophisticated statistical approaches. For instance, testing thousands of SNPs for their association with clinical outcome increases the number of false-positive results, whereas a stringent adjustment for multiple testing with Bonferroni correction reduces study power and may lead to false-negative associations [24–26]. Even though the term ‘genome- wide’ indicates that allelic variants across the human genome are covered, currently available scans with 100 000–500 000 SNP arrays cover, at most, 5% of the estimated allelic variations in our genome. Therefore, associations in genome-wide studies do not only depend on association between genotype and clinical phenotype, power and sample size of the study to detect a significant difference, but also on allele frequencies at the marker and the candidate gene locus and the amount of linkage disequilibrium between genetic variants [27].
An interesting step between the candidate gene and the whole-genome pharmacogenetic analysis is the study of drug response in the downstream and interacting signalling pathways [20, 28]. Pathways of genes with allelic variants may be more important than individual genes, with the effects of polymorphisms in networks of genes acting together to create one phenotype. Thus, variability in drug response may reflect genetically determined changes in the biological environment in which drugs interact.
Nevertheless, it has to be realized that genetic factors are estimated to account for 15–30% of the interindividual variation in drug response. Only for certain drugs has an exceptionally high degree of interindividual variation in response been attributed to genetic variants [29]. Drug efficacy and safety are more likely to be a complex result of the influence of many genes interacting with environmental and behavioural factors [29, 30]. For that reason, the current examples of clinically applied pharmacogenetics represent genetic variations with relatively high penetrance in a monogenic trait; an allelic variant has a profound effect on the pharmacokinetic and pharmacodynamic factors of a drug, and that individual difference in one gene accounts for clearly recognizable phenotypes [28, 29]. An approach to increase the explained variation in drug response and to improve the usefulness of clinical pharmacogenetic testing is to develop models in which genetic and non-genetic factors important for the phenotype are combined.
Pharmacogenetic studies in complex traits such as RA are, as yet, developing. To date, the extent to which genetic factors influence drug response in RA is largely unknown. The unidentified nature of the disease makes the precise mechanisms of action of DMARDs uncertain and the candidate gene approach of limited use. In addition, disease activity fluctuations and disease progression may affect treatment response. However, recent reports do indicate that genetic variants affect DMARD response in patients with RA.