Fcγ receptor type IIIA polymorphisms influence treatment outcomes in patients with inflammatory arthritis treated with tumor necrosis factor α–blocking agents




To determine whether polymorphisms in Fcγ receptor type IIIA (FcγRIIIA) correlate with clinical efficacy in patients with rheumatoid arthritis (RA) and patients with psoriatic arthritis (PsA) being treated with tumor necrosis factor α (TNFα) inhibitors.


The study group comprised 30 patients with RA and 5 patients with PsA. Patients were classified as being very good responders (n = 23) or nonresponders (n = 12) to therapy with TNF blocking agents. Whole blood was obtained from all patients, and DNA was extracted for FcγRIIIA analysis. The FcγRIIIA-158 polymorphism was determined using an allele-specific polymerase chain reaction assay.


The distribution of FcγRIIIA-158 genotypes was as follows: for F homozygous (F/F), 31.5%; for V homozygous (V/V), 11.5%; and for V/F heterozygous (V/F), 57%. Among very good responders, the distribution of alleles was as follows: for F/F, 48%; for V/V, 13%; and for V/F, 39%. Among nonresponders, the distribution of alleles was as follows: for F/F, 0%; for V/V, 8%; and for V/F, 92%. The low-affinity F/F homozygous genotype was found to be significantly associated with response to TNF inhibitor therapy (P < 0.01 by Fisher's exact test).


These results suggest that FcγRIIIA-158 polymorphisms may affect the outcome of treatment with TNF blocking agents. Better understanding of the factors affecting responses to these agents could result in improved outcomes of treatment.

The use of therapies targeting the key proinflammatory cytokine tumor necrosis factor (TNF) has remarkably improved the standard of care for the treatment of rheumatoid arthritis (RA) and other systemic inflammatory disorders. At present, 3 TNF inhibitors have been introduced into clinical practice: infliximab (a chimeric monoclonal antibody), adalimumab (a human monoclonal antibody), and etanercept (a fusion protein consisting of a dimer of the extracellular portion of p75 TNF receptor linked to the Fc portion of human IgG1). Data from research studies as well as growing clinical experience have shown that clinical responses to TNF inhibitor therapy are variable among patients who are treated with these agents. Approximately 25% of patients have marked improvement in all aspects of disease, 50% show a moderate response, and ∼25% experience only minor, transient clinical improvement or have no benefit from treatment (1). Despite substantial analysis and investigations among patients treated with TNF inhibitors, to date no disease-related features, patient characteristics, genetic associations, or other factors have been identified that reliably correlate with treatment outcome, either positively or negatively (2).

Immunoglobulin molecules mediate various effector functions via specific binding of their Fc portions to cell surface receptors. Engagement of these Fc receptor (FcR) molecules can affect cell-specific functions, including phagocytosis, degranulation, antibody-dependent cell-mediated cytotoxicity, cytokine release, and regulation of antibody production (3). FcR polymorphisms alter receptor function by enhancing or diminishing the affinity for immunoglobulins (4). Three classes of FcR that are capable of binding IgG antibodies are recognized: FcγRI (CD64), FcγRII (CD32), and FcγRIII (CD16) (3, 5). FcγRII and FcγRIII have multiple isoforms (FcγRIIA/C and B; FcγRIIIA and B). FcγRIIIA is expressed as a membrane-spanning receptor on macrophages, natural killer cells, and a subset of T lymphocytes. The most frequent polymorphism of FcγRIIIA is a point mutation affecting amino acids in codon 158 in the extracellular domain. This results in either a valine (V158) or a phenylalanine (F158) at this position.

Compared with the F/F allotype, the V/V allotype has a greater affinity for human IgG. In the general population, the approximate distribution of FcγRIIIA allotypes is as follows: for V/V, 1–20%; for V/F, 35–45%; and for F/F, 40–50% (6). Recently, it was suggested that allelic polymorphisms in FcγRIIIA may correlate with the efficacy of immunoglobulin-based therapeutics by their variable affinity for immunoglobulins. This, in turn, could influence antibody-dependent cell-mediated cytotoxicity, complement-dependent cytotoxicity, apoptosis, or other factors. In a study of non-Hodgkin's lymphoma, the therapeutic response to rituximab (anti-CD20 monoclonal antibody) was associated with the FcγRIIIA genotype. Patients homozygous for the FcγRIIIA-V158 genotype had a 100% response, while the responses in FcγRIIIA-F158 carriers were 67% and 51% for those who were heterozygous and F homozygous, respectively (6). Preliminary data also indicate that in patients with systemic lupus erythematosus, the FcγRIIIA genotype influences the extent of B cell depletion that occurs during anti-CD20 monoclonal antibody therapy (7). We hypothesized that there might be an association between polymorphisms in FcγRIIIA and the clinical response to TNFα inhibitors in patients with RA or psoriatic arthritis (PsA).


Eligible subjects were patients with RA or PsA who were 18 years of age or older. Each patient's treating rheumatologist was asked to categorize the patient as being a very good responder, a fair or good responder, or a nonresponder according to his or her response to 3 months of therapy with 1 of the 3 anti-TNF agents (infliximab, etanercept, adalimumab). Very good responders to anti-TNF treatment were patients who had ≥70% improvement in the physician's global assessment and the swollen and tender joint counts, as well as ≥50% improvement in 2 of the following measures: erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) level, patient's global assessment, and duration of early morning stiffness. Nonresponders to anti-TNF treatment were defined as patients who experienced no improvement in the physician's assessment and no improvement in the ESR, CRP level, joint count, patient's global assessment, or morning stiffness. Fair or good responders were patients who could not be classified into either of the other 2 categories.

This investigational analysis was conducted at the arthritis clinics of the University of California, San Diego. The institutional review board approved the protocol, and all patients gave written informed consent prior to participation in the study.

Determination of FcR polymorphisms.

Heparin-anticoagulated blood samples were collected, and chromosomal DNA was isolated using phenol–chloroform extraction. Samples were diluted in distilled water and stored at −20°C until used. FcγRIIIA genotyping was performed by allele-specific polymerase chain reaction (PCR) analysis, as described by Leppers-van de Straat (8). Forward and reverse primers were selected. Specificity of the reactions was achieved by gradually increasing the annealing temperature. PCR products were run on 3% agarose gels and visualized under ultraviolet light using a photoimager system (Figure 1).

Figure 1.

Fcγ receptor IIIA allotyping assay, showing detection of V and F polymorphisms by allele-specific polymerase chain reaction (PCR) (8).

Statistical analysis.

Differences in the frequencies of alleles between the 2 study populations (very good responders versus nonresponders) were analyzed by Fisher's exact test. The chi-square goodness-of-fit test was used to compare the distribution of genotypes among the study subjects with that among the general population.


Thirty-five patients (30 with RA and 5 with PsA) were included in the study. All patients had active disease despite being treated with disease-modifying antirheumatic drugs (DMARDs) prior to initiation of anti-TNF therapy. Twenty-four of the patients were white, 7 were Hispanic, 2 were African American, and 2 were Asian. Twenty-three of the patients were considered to be very good responders to TNF-blocking therapy, and 12 patients were classified as nonresponders. The mean ± SD disease duration was 9.5 ± 6.8 years among responders and 10.9 ± 6.9 years among nonresponders. Fourteen of the responders (61%) and 7 of the nonresponders (58%) were receiving methotrexate (Table 1); 9 of the responders (39%) and 5 of the nonresponders (42%) were receiving more than 1 DMARD, including methotrexate; and 11 responders (48%) and 5 nonresponders (42%) were receiving oral prednisone (all dosages were <10 mg/day) at the time when anti-TNF therapy was initiated. Among responders, 10 were treated with etanercept, 9 were treated with infliximab, and 4 were treated with adalimumab. Among nonresponders, 9 were treated with etanercept, 2 were treated with infliximab, and 1 was treated with adalimumab.

Table 1. DMARD use among responders and nonresponders to anti-TNF agents*
 Responders (n = 23)Nonresponders (n = 12)
  • *

    Values are the number (%). DMARD = disease-modifying antirheumatic drug; anti-TNF = anti–tumor necrosis factor.

Methotrexate14 (61)7 (58)
Prednisone11 (48)5 (42)
More than 1 DMARD9 (39)5 (42)

Overall, the distribution of FcγRIIIA-158 alleles in the study population approximated the observed distribution in the normal population, namely, 31.5% for low-affinity F homozygous (F/F), 11.5% for high-affinity V homozygous (V/V), and 57% for V/F heterozygous (P > 0.8). Among the very good responders, the distribution of alleles was 48% F/F, 13% V/V, and 39% V/F. Among nonresponders, the distribution of alleles was 0% F/F, 8% V/V, and 92% V/F (Figure 2). The low-affinity F/F allotype was found to be significantly associated with a response to anti-TNF therapy as compared with the V/V plus V/F allotypes (P < 0.01 by Fisher's exact test).

Figure 2.

Distribution of allotypes among study patients.


In this study of patients with RA or PsA, in which those who demonstrated a very good response to TNF-blocking agents were compared with patients who showed no clinical response to this therapy, we observed that polymorphisms in FcγRIIIA-158 correlated with therapeutic outcome. We purposely avoided including patients with intermediate responses in order to help minimize confounding factors. The low-affinity homozygous F allele had a statistically significant relationship with a very good clinical response.

All 3 TNF-blocking agents share an IgG1 Fc fragment, although the TNF-binding portions of the molecules differ. Infliximab and adalimumab are monoclonal antibody molecules, while etanercept includes functional TNF receptors. Therapeutic agents containing immunoglobulins or immunoglobulin-like structures offer pharmacokinetic benefits to the response to treatment, but the contribution of the Fc portion of the immunoglobulin among currently available TNF blockers remains to be defined. Anti-TNF agents neutralize the biologic activity of TNFα by binding to the soluble and transmembrane forms of TNFα and preventing soluble and transmembrane TNFα from binding to TNF; however, mechanisms other than simple neutralization of TNFα may be involved (9, 10). For instance, by binding to membrane-bound TNF, anti-TNF agents might induce apoptosis of the cells responsible for inflammation and reduce the number of activated cells at the site of inflammation (11, 12). It has been demonstrated that all 3 TNF antagonists can lead to apoptosis by inducing reverse signaling (9, 13). Moreover, because both infliximab and etanercept have been shown to cause down-regulation of interleukin-12–stimulated interferon-γ production, it is possible that binding of these drugs to membrane-bound TNF may effect reverse signaling (14).

In our study, all 9 patients with the low-affinity allotype (F/F) were good responders to TNF therapy. These results could be explained by FcγRIIIA-related changes in the elimination or distribution of the anti-TNF agents and, thus, their availability. Normal clearance of IgG is regulated largely by another type of FcR, the neonatal FcR, which is a distant member of the class I major histocompatibility complex protein family that is expressed by endothelial cells (15). Whether there is any association between FcγRIIIA and neonatal FcR remains to be determined. Immune complex clearance by macrophages and the reticuloendothelial system, however, is governed by the affinity of Fc receptors. If anti-TNF drugs are cleared from the circulation by being complexed to TNF, then a low-affinity receptor might allow for persistence of the therapeutic agent. The carriage rate of the F/V heterozygous allotype was higher in the nonresponder group (92%). F/V allotype affinity has been described as low or high in different studies (6, 7). It is possible that the F/V allotype behaves in a manner similar to that of the V/V high-affinity allotype and leads to faster elimination of anti-TNF agents.

The results of this study suggest that the FcRIIIA allotype may influence the treatment outcome in patients taking anti-TNF agents. However, this study has several caveats. First, the population studied was very small. Further examination of FcRIIIA polymorphisms in a larger population of arthritis patients treated with anti-TNF drugs is needed to confirm these preliminary results. Also, distinct results may be observed among patients with dissimilar genetic backgrounds, as has been observed for other genetic associations (16). Also, in this study, we focused on extreme clinical results; the associations observed may or may not hold if varying levels of response are included. Nevertheless, the substantial clinical response to TNF inhibitors and the high cost of these treatments justify efforts to identify patients who would or would not benefit from these agents. In addition, these findings might extend to other immunoglobulin-containing drugs used to treat other diseases. Better understanding of drug response could result in improved outcomes, reduced costs, and optimization of pharmacologic approaches.