Individualization of Leflunomide Dosing in Rheumatoid Arthritis Patients

Ashley M Hopkins; Catherine E O'Doherty; David JR Foster; Richard N Upton; Susanna M Proudman; Michael D Wiese


Personalized Medicine. 2014;11(4):449-461. 

In This Article

Pharmacogenomic Variants Affecting Leflunomide Response Outcomes

Pharmacogenetic variants have also been shown to affect treatment responses. Pawlik et al.[54] examined a SNP within the first exon of DHODH (rs3213422, 19C>A), where the mutant A allele causes a Gln7Lys amino acid substitution in the N-terminus of the expressed DHODH protein. Within the Caucasian population approximately 25% carry the CC genotype of DHODH 19C>A, and Pawlik et al.[54] found that the C allele was more frequent amongst 147 patients who met the ACR20 response criteria after 6 months of treatment (p = 0.048). This association was not observed in a similar study conducted by O'Doherty et al.,[56] although carriage of DHODH haplotype 2 (a six marker haplotype; which interestingly includes the C allele of DHODH 19C>A) was associated with reduced response to therapy (p = 0.008; n= 56 at 3 months).[56] The different findings in the two studies could be explained by differences in the patient populations. In the O'Doherty et al.[56] study, more than 90% of participants initiated leflunomide in combination with other DMARDs less than one year after RA diagnosis, compared with the Pawlik et al.[54] study, where the disease duration was more than 9 years at the initiation of monotherapy with leflunomide.

Interestingly, Grabar et al.[58] found that carriage of the C allele of DHODH 19C>A was also associated with a reduced risk of toxicity in individuals who were primarily taking leflunomide as monotherapy. Grabar et al.[53] also found the CC genotype of a SNP in CYP1A2 (rs762551 (-163C>A)), which is present in approximately 5% of the Caucasian population, was associated with a 9.7 higher odds ratio of toxicity compared with those carrying an AA or AC genotype (p=0.002), where carriage of the A allele is commonly referred to as CYP1A2*1F .[53] SNP CYP1A2*1F is associated with increased enzymatic activity, particularly in tobacco smokers, and although CYP1A2 is partially responsible for the conversion of leflunomide to teriflunomide, there was no observable influence on teriflunomide concentration according to this genotype.[37,53] However, Wiese et al.[55] did not detect this association with toxicity for either DHODH 19C>A, or CYP1A2*1F. Further studies of the effect of CYP2C19 phenotype on leflunomide toxicity by genotyping CYP2C19*2 and CYP2C19*17 (rs4244285 and rs12248560, respectively), identified a linear trend for poor and intermediate metabolisers to cease leflunomide due to toxicity more frequently than extensive or ultra-rapid metabolisers (adjusted Hazard Ratio = 0.432 for each incremental change in phenotype, 95% CI 0.237 to 0.790, p= 0.006).[55] This followed Grabar et al.'s[53] study, which found no association between carriage of CYP2C19*2 and toxicity, despite previously reporting the loss of function allele to be associated with lower teriflunomide concentrations.[37,53] This suggests that the side effects of leflunomide may not be caused solely by teriflunomide, as poor and intermediate metabolisers should have lower concentrations and therefore a lower incidence of toxicity.[55] A potential explanation of this is that toxicity occurs from leflunomide directly, or alternatively a toxic metabolite is formed more abundantly in poor and intermediate metabolisers.[55] Currently, no other metabolites of leflunomide, other than teriflunomide, have been detected within the plasma, but studies in human hepatic microsomes have found that leflunomide can be converted to either teriflunomide or the metabolite methyl-hydroxyleflunomide (M2), whereas teriflunomide is converted to methyl-hydroxyteriflunomide (M1).[61,62] These findings have led to the hypothesis that the enzymes responsible for the formation of M2 are different from CYP2C19 and CYP1A2 which mediate conversion of leflunomide to teriflunomide, and that when CYP2C19 activity is reduced, increased formation of M2 is likely, which may contribute to the increased incidence of toxicity.[55] Similarly altered CYP1A2 activity may result in the formation of toxic leflunomide metabolites, although at this point it remains biologically unexplained how the increased enzymatic activity allele, CYP1A2*1F, results in the changed toxicity observed by Grabar et al..[53]

Finally, estrogen receptor (ESR) gene studies have found that SNPs in ESR1 (rs9340799 and rs2234693) have been associated with variable responses to leflunomide, with the AA genotype of rs9340799 and TT genotype of rs2234693 associated with a better response after 12 months.[59] These results are consistent with cell culture studies in which estrogen modulated the effect of leflunomide, studies which found that sex hormones are involved in regulating the immune response in RA and that in general, females respond poorer than men to treatment.[59,63–65]