Differential Efficacy of TNF Inhibitors With or Without the Immunoglobulin Fc Portion in Rheumatoid Arthritis - The ANSWER Cohort Study


 Background: Rheumatoid factor (RF) binds to the Fc portion of immunoglobulin and can bind to the Fc portion of anti-TNF monoclonal antibodies (TNFi), attenuating their clinical efficacy. We tried to determine whether the therapeutic efficacy is different between TNFi with or without the immunoglobulin Fc portion in rheumatoid arthritis (RA) patients depending upon the RF titres.Methods: RA patients treated with TNFi were enrolled from an observational multi-centre registry in Japan. TNFi treatments were divided into two groups based on the structural characteristics between TNFi with Fc (infliximab, adalimumab, golimumab, etanercept) and TNFi without Fc (certolizumab-pegol). Patients were classified into 4 groups according to RF titre quartiles. The sequential disease activity score in 28 joints using erythrocyte sedimentation rate (DAS28-ESR) were compared by the Mann-Whitney U test between TNFi with and without the immunoglobulin Fc portion in each RF titre group.Results: A total of 705 RA patients were included and classified into four groups according to quartiles of baseline RF titres (RF1; RF 0-15.0 IU/mL, RF2; 15.0-55.0, RF3; 55.0-166, RF4; 166-7555). In RF4, RA patients treated with TNFi without Fc had a significantly lower DAS28-ESR than those treated with TNFi with Fc (3.2 (2.3-4.2) vs. 2.7 (2.0-3.0)) (median, IQR) after 12 months of follow-up, while there were no significant differences in RF1, RF2, and RF3.Conclusion: TNFi without Fc (certolizumab-pegol) may be more efficacious than TNFi with Fc in RA patients with high RF titres.

License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Background Rheumatoid arthritis (RA) is a chronic autoimmune condition with synovitis. The treatment strategy for RA has totally changed since biological disease-modifying antirheumatic drugs (bDMARDs) emerged in clinical practice. Anti-TNF inhibitors (TNFi) were the rst bDMARDs introduced in RA clinical practice. To date, ve TNFi have been approved for RA treatment, including in iximab (IFX), etanercept (ETN), adalimumab (ADA), golimumab (GLM), and certolizumab-pegol (CZP). However, it is not fully known which TNFi have superior clinical e cacy over other TNFi in patients under speci c clinical conditions. The ve TNFi have different structural characteristics. IFX, ADA, and GLM are fully functional IgG1 monoclonal antibodies against TNF and contain the immunoglobulin Fc portion, with IFX being a chimeric antibody and ADA and GLM being fully humanized monoclonal antibodies. ETN is a soluble receptor against TNF fused with the Fc portion of human IgG1, and CZP is a monovalent Fab fragment against TNF fused with polyethylene glycol (PEG). Because CZP is a PEGylated Fab fragment, it lacks the Fc portion of immunoglobulin, while all other TNFi (IFX, ADA, GLM, and ETN) have an immunoglobulin Fc portion in their structures [1]. These structural differences in TNFi may lead to differential clinical responses against RA in patients with speci c clinical characteristics. However, few studies have analysed the differential clinical responses among different TNFi.
Rheumatoid factor (RF) is an IgM autoantibody against the Fc portion of immunoglobulin G (IgG). High RF titres are the major poor prognostic factors for RA according to the 2019 EULAR recommendations [2]. Because RF can bind to the immunoglobulin Fc portion, it might bind the Fc portion of certain TNFi and in uence their clinical e cacy. Therefore, it is possible that high titres of RF might attenuate the clinical effect of those TNFi containing the immunoglobulin Fc portion (TNFi-w/ Fc; IFX, ADA, GLM, and ETN) but not that of TNFi without Fc (TNFi-w/o Fc; CZP). However, no study has compared the clinical e cacy of TNFi with or without the immunoglobulin Fc portion according to RF titres.
In this study, we tried to investigate whether the therapeutic e cacy is different between TNFi with and without the Fc portion in RA patients depending on RF titres, particularly those with high RF titres. To this end, we analysed the data from the ANSWER cohort, a multi-centre registry of RA patients in Japan, which has sequential clinical data before and after bDMARD treatment [3,4].

Patients
The Kansai Consortium for Well-being of Rheumatic Disease Patients (ANSWER) cohort is an observational multi-centre registry of patients with RA in the Kansai district of Japan. Data from patients at seven institutes (Kyoto University, Osaka University, Osaka Medical College, Kansai Medical University, Kobe University, Nara Medial University, and Osaka Red Cross Hospital) were included. Patients with RA treated with one of ve TNFi (ADA, CZP, ETN, GLM, and IFX; including both intravenous and subcutaneous agents but excluding bio-similar agents, all of which were introduced between 2011 and 2018) and whose data on baseline RF titres and disease activity were available were included in our study. In this study, patients ful lled the 2010 RA ACR/EULAR criteria. In addition, age, sex, body mass index, disease activity (disease activity score in 28 joints using erythrocyte sedimentation rate (DAS28-ESR), health assessment questionnaire (HAQ)), disease duration, bDMARD naïve rates, concomitant use rates and doses of methotrexate (MTX) and prednisolone (PSL), RF, and anti-cyclic citrullinated peptide antibody (ACPA) were collected. RF (IgM RF) was measured by latex agglutination nephelometric assay. ACPA were measured by anti-CCP2 ELISA kit. Patients were classi ed into equal quartiles according to baseline RF titres (RF 1 -RF 4 ).
Outcome measures DAS28-ESR was assessed before and at 3, 6, and 12 months after TNFi initiation, according to baseline RF concentration quartiles. The differences in DAS28-ESR from baseline (dDAS28-ESR) to 3, 6, and 12 months after TNFi initiation were assessed in the same way. The 1-year overall retention rates were also assessed.

Ethics
The representative facility of this registry is Kyoto University, and this observational study (not clinical trial) was conducted in accordance with the Declaration of Helsinki and approved by each ethics committee of the seven institutes described above (Kyoto University approval number R0357).

Statistical analysis
An unpaired t-test or the Mann-Whitney U test was used when variables were continuous, and the chi-squared test or Fisher's exact test was used when variables were categorical. To deal with missing DAS28-ESR after TNFi initiation, data were analysed using the last observation carried forward method. All analyses were two-tailed, and P values of < .05 were considered statistically signi cant. All statistical analyses were performed with JMP Pro 14.0 for Mac (SAS Institute Inc., Cary, NC, USA).  Table 1. In RF 4 , ACPA titres (133 (36-300) vs. 346 (109-500)) (IU/mL) and HAQ (1.0 (0.5-1.6) vs. 1.5 (1.0-2.2)) were higher in patients with TNFi-w/o Fc than in those with TNFi-w/ Fc.

Discussion
To the best of our knowledge, this is the rst study to compare the e cacy of TNFi with and without the immunoglobulin Fc portion in RA patients according to baseline RF titres. The results revealed that TNFi-w/o Fc was more e cacious than TNFi-w/ Fc in RA patients with high RF titres (RF high -RA patients).
RF is a typical poor prognostic factor of RA. IgM RF ampli es the production of TNF from macrophages induced by the immune complex containing ACPA [5,6]. It has been demonstrated that high titres of RF or higher levels of serum TNF concentrations are associated with higher disease activity [7]. In a post hoc analysis of the RISING study, the baseline TNF concentration was increased in ACPA-positive and RF-positive patients, and those patients had attenuated clinical responses to IFX [6]. Thus, high titres of RF are an important factor that could lead to treatment resistance by TNFi. Establishing an appropriate treatment strategy for RF high -RA patients would be of considerable clinical importance.
However, it is not fully understood how RF affects the clinical e cacy of TNFi. Some reports showed that the presence or high titres of RF can be a predictive marker of unresponsiveness to TNFi, but a meta-analysis failed to con rm this nding [8,9]. RA patients with high baseline IgA RF titres were associated with poor clinical response to TNFi, including IFX, ADA, and ETN [10]. Furthermore, RF positivity was associated with a poor DAS28-ESR response to TNFi, including IFX, ADA, ETN and GLM [9,11]. However, other reports suggested that neither positivity nor titres of RF were associated with treatment resistance to TNFi [12]. These previous studies were all regarding TNFi-w/ Fc (IFX, ADA, GLM, and ETN), and there have been no data on TNFi-w/o Fc (CZP). Our results suggested that TNFi-w/o Fc, namely, CZP, might be e cacious in treating RF high -RA patients. We have some accountable mechanism for this nding as discussed below.
First, RF might bind to the Fc portion of TNFi and attenuate the complement-dependent cytotoxicity (CDC) activity of TNFi-w/ Fc. CDC activity is an important effector function of TNFi, particularly for fully functional monoclonal antibodies (IFX, ADA, and GLM), because it exerts cytotoxic effects against membrane-bound TNF-expressing cells. [4,13]. Additionally, the Fc portion is known to be required for CDC activity. Because RF binds to the Fc portion of immunoglobulin, RF interferes with the CDC activity of the anti-B cell monoclonal IgG1 antibody rituximab (RTX) in an Fc-dependent manner [14]. High RF titres were associated with reduced CDC activity by RTX in vitro [14]. Considering these reports, RF might bind to the Fc portion of TNFi-w/ Fc and attenuate their therapeutic effect.
Second, RF might affect the neonatal Fc receptor (FcRn)-mediated recycling of monoclonal antibodies, and the serum concentration of TNFi-w/ Fc might be reduced in RF high -RA patients. The in vivo kinetics of IgG are controlled by FcRn, and serum concentrations of monoclonal antibodies with Fc portions are controlled via FcRn-mediated recycling [15]. Circulating IgG in the blood enters monocytes or endothelial cells via uid phase pinocytosis and is encapsulated into endosomes. Inside endosomes, IgG binds to FcRn and is recycled via exocytosis, while IgG, which cannot be recognized by FcRn, is sorted and degraded by lysosomal functions [16]. Notably, in patients who have the IgG1 allotype G1 m17,1, IFX competes with their own IgG1 and cannot undergo the recycling process, resulting in a shortened half-life of IFX [17]. RF binds to monoclonal antibodies in the same region of the FcRn binding portion, the CH2-CH3 interface region, and thus RF-bound antibodies cannot undergo the recycling process su ciently [18,19]. Therefore, the half-lives of TNFi-w/ Fc might be shortened in patients with high RF titres, and higher doses of TNFi may be needed to control RA disease activity in RF high -RA patients [6].
CZP is a monovalent Fab' fragment with PEGylation and is the only TNFi that does not possess the immunoglobulin Fc portion [1]. This unique structure allows the therapeutic effects and pharmacokinetics of CZP to be unaffected by Fc-mediated CDC or FcRn-mediated recycling processes [20]. TNFi-w/ Fc may have similar clinical e cacy and pharmacokinetics even in patients with high RF titres. This might have led to the bene cial effect of CZP over other TNFi-w/ Fc in RF high -RA patients.
This study has some limitations, as it is a retrospective cohort study, and possible confounding factors could not be extracted. While baseline characteristics of TNFi-w/ Fc and TNFi-/wo Fc were different about RF titre, ACPA titre, and HAQ in RF 4 , it was not advantage but disadvantage for TNFi-w/o Fc. Therefore, it was still applicable that TNFi-wo Fc was bene cial for RF high RA patients. We also cannot exclude the possibility that our result is drawn from some unknown confounding factors. In addition, we could not perform adjusted statistical analysis with baseline clinical characteristics due to the small number of patients. Thus, future analysis is required. Mechanistic studies that analyse the binding between RF and the Fc portion of TNFi are also needed.

Conclusions
In conclusion, our analysis suggested that CZP can be a better treatment option in RF high -RA patients than other TNFi-w/ Fc. Our ndings may help promote precision medicine in RA, especially in RF high -RA patients.  Tables   Table 1. Baseline patient characteristics in each group (RF 1 -RF 4 ).

List Of Abbreviations
Baseline characteristics were compared between RA patients treated with TNF inhibitors with the Fc portion (w/Fc) and those without (w/o Fc).
Continuous data were analysed using an unpaired t-test or the Mann-Whitney U test, and categorical data were analysed using the chi-squared test or Fisher's exact test.