Impact of Single-Nucleotide Polymorphisms of CTLA-4, CD80 and CD86 on the Effectiveness of Abatacept in Patients with Rheumatoid Arthritis

Abatacept (ABA) is used as a first-line treatment in patients diagnosed with moderate and severe rheumatoid arthritis (RA). The interindividual response to ABA therapy is very variable in these patients. The objective of our study was therefore to investigate the role of polymorphisms of the CTLA-4, CD80 and CD86 genes, as well as that of clinical factors of the disease, in the response to ABA in patients with RA. A retrospective cohort study was carried out in 109 patients receiving treatment with ABA and diagnosed with RA. The genetic variables were analyzed using real-time PCR with TaqMan® probes. The patients were classified according to the European League Against Rheumatism (EULAR) criteria at 6 and 12 months from start of treatment. The independent variables associated with higher EULAR response were lower duration of previous biologic disease-modifying anti-rheumatic drugs and lower baseline values of the disease activity score 28 after 6 months of ABA treatment; and lower baseline patient’s visual analogue scale (PVAS) after 12 months. In addition, a significant association was found between duration of ABA treatment, non-administration of concomitant glucocorticoids and lower baseline values of the number of inflamed joints and erythrocyte sedimentation rate clinical variables, with remission of the disease after 6 months’ treatment with ABA. Finally, remission of the disease after 12 months’ treatment with ABA was associated with earlier age at start of ABA therapy and lower number of previous biologic therapies (BTs). The CTLA-4 rs5742909-T allele and the CTLA-4 rs231775-G allele were found to be associated with satisfactory EULAR response and low disease activity (LDA) after 12 months’ treatment with ABA (CTLA-4 rs5742909 T vs. CC; OR = 5.88; CI95% = 1.48–23.29 and OR = 4.75; CI95% = 1.35–17.94, respectively, and CTLA-4 rs231775 G vs. AA, OR = 3.48; CI95% = 1.20–10.09 and OR = 4.68; CI95% = 1.49–17.94, respectively). In conclusion, patients with RA treated with ABA showed better EULAR response and LDA rate when they had the CTLA-4 rs5742909-T or CTLA-4 rs231775-G polymorphisms; furthermore, this remission rate increased in patients that began ABA treatment earlier, those with a lower number of previous BTs and those with a lower PVAS value.


Introduction
Rheumatoid arthritis (RA) is a chronic autoimmune disease affecting the joints. It produces inflammation and structural damage, reducing patients' quality of life [1]. The worldwide prevalence of the disease varies between 0.3% and 1.2%, and it is the most common chronic inflammatory joint pathology in Spain, with a prevalence of between 0.3% and 1.6% [2][3][4][5]. The therapeutic strategy for RA has changed in recent years. Clinical remission or, failing that, low disease activity (LDA) has been established as the main objective [6]. Biologic disease-modifying antirheumatic drugs (bDMARDs), as well as targeted synthetic DMARDs (tsDMARDs), have played a key role in the prognosis of RA [7]. One of the therapeutic targets indicated in RA treatment is abatacept (ABA). This is a fusion protein formed by the extracellular domain of cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4), linked to the constant fragment (Fc) of human immunoglobulin 1 (IgG1). The mechanism of action of ABA is based on its binding with the CD80/CD86 complex, preventing the interaction of this complex with the CD28 transmembrane protein of T-lymphocytes by blocking the co-stimulation signal necessary for their activation [8].
Abatacept is used as a first-line treatment for moderate and severe RA in patients in whom the therapeutic objective has not been achieved after previous administration of DMARDs, according to the criteria of the European League Against Rheumatism (EULAR) recommendations [7]. There is great interindividual variability in treatment response with bDMARDs, with a therapeutic failure rate of approximately 30% [9]. Finding response biomarkers for these drugs would help to develop a therapy individually tailored to the patient, achieving the therapeutic objective more quickly and effectively. Interindividual genetic variability may also contribute to the response to ABA treatment. Single-nucleotide polymorphisms (SNPs) have been associated with response to bDMARDs and could be used as pharmacogenetic predictors [10][11][12][13][14].
Polymorphisms in the CD80 and CD86 genes have previously been studied for their association with autoimmune diseases [15][16][17][18]. The CD80/CD86 complex is a set of membrane glycoproteins consisting of two extracellular domains: a transmembrane domain and a cytoplasmic tail domain [19,20]. In the CD86 gene we find the rs1129055 (G > A) polymorphism, located in exon 8, which gives rise to an alanine (Ala) to threonine (Thr) amino acid substitution at codon 304 [21]. Similarly, in the CD80 gene we find the rs57271503 (G > A) polymorphism. Furthermore, SNPs in the CTLA-4 gene have been associated with peripheral tolerance. CTLA-4 is an immune system regulatory receptor, constituently expressed on the surface of activated T lymphocytes and in regulatory T lymphocytes, inhibiting their proliferation [22][23][24]. CTLA-4 shows a greater affinity for the CD80/CD86 complex than CD28 and activates an immune signal inhibiting T-lymphocyte proliferation. It is indispensable for terminating the immune response cascade, as well as for acquisition of peripheral tolerance for antigens, preventing autoimmunity [23,25].
All these receptors are related, in turn, to the mechanism of action of ABA [15,18,[26][27][28]. The presence of various SNPs in these genes may entail a conformational change in the proteins they encode, producing variability in the patient's treatment response. In this conceptual framework, our objective is to assess the role of SNPs of these three genes, CTLA-4 (rs3087243, rs231775 and rs5742909), CD80 (rs57271503) and CD86 (rs1129055), as possible response predictors (EULAR response, LDA and remission) in patients with RA treated with ABA.

Study Design
We conducted a retrospective cohort study.

Ethics Statements
This study was carried out in accordance with the Declaration of Helsinki, with the approval of the Ethics and Research Committee of the University Hospital Virgen de las Nieves. The subjects who participated in the study signed an informed consent for collection and genetic analysis of saliva samples and for their donation to the Andalusian Public Health System Biobank. The samples were identified by alphanumeric codes.

Study Population
The study included 109 Caucasian patients diagnosed with RA according to the American College of Rheumatology (ACR) classification criteria, recruited in the Rheumatology and Pharmacy Departments of the University Hospital Virgen de las Nieves in Granada between 2009 and 2019 [7]. Of the 109 patients recruited in the study, the response to ABA was evaluated in 105 after 6 months and in 92 after 12 months. The remaining patients did not meet the study's evaluation criteria. The participants were over the age of 18 years and being treated with ABA. The route of administration of the drug was intravenous (IV): 500 mg (<60 kg), 750 mg (60-100 kg) or 1000 mg (>100 kg), in weeks 0, 2 and 4, respectively, and subsequently every 4 weeks at the same dose, in 49 patients; and subcutaneous (SC): 125 mg/week, in 60 patients.

Sociodemographic and Clinical Variables
The sociodemographic variables included sex, smoking, age at diagnosis of RA, number of years with the disease, age at initiation and duration of ABA treatment, route of administration of the drug (IV or SC), concomitant glucocorticoids (GCs), concomitant conventional synthetic DMARDs (csDMARDs) (methotrexate (MTX), leflunomide (LFN)), number and duration of previous BTs and reason for suspension of ABA.

DNA Isolation
The saliva samples were collected with buccal swabs (Kit OCR-100). The DNA was extracted using the QlAamp DNA Mini Kit (Qiagen GmbH, Hilden, Germany), following the manufacturer's instructions for purifying DNA from saliva, and stored at −40 • C. The DNA concentration and purity were measured using a NanoDrop 2000 UV spectrophotometer with the absorbance ratio at 280/260 and 280/230.

Response Variables
The effectiveness of the treatment was evaluated according to the EULAR response, LDA and disease remission criteria, at 6 and 12 months from commencement of treatment with ABA.

Statistical Analysis
The descriptive analysis was performed using R 3.5.1 software. The quantitative variables were expressed as the mean (± standard deviation) for those that complied with normality and as the median and percentiles (25 and 75) for the variables that did not follow a normal distribution. Normality was confirmed by the Shapiro-Wilk test.
The bivariate analysis between the response and the sociodemographic, clinical and genetic variables was performed using Pearson's chi-square test or applying Fisher's exact test for the qualitative variables. For the quantitative variables, Student's t-test was applied to the variables that complied with normality, and the Mann-Whitney U test was applied for non-normal variables. The statistical power of genetic association was determined using G*Power 3.1.9.7. [35].
Multivariate analysis (logistic or linear regression) was used to calculate the adjusted odds ratio (OR) and a 95% confidence interval (CI 95% ) for the potential EULAR response, LDA and remission prognostic factors. The goodness of fit for each model was analyzed with the Hosmer-Lemeshow test and the omnibus test of coefficients, as well as calculating the Cox-Snell and Nagelkerke r 2 coefficients. All tests were two-sided, with a probability of 0.05 or less being considered statistically significant, and were performed using R 3.5.1 or PLINK toolset free-access software for whole genome association analysis [36][37][38].
The Hardy-Weinberg equilibrium and the haplotype frequencies were determined and Lewontin's D-prime (D') and the linkage disequilibrium coefficient (r 2 ) were calculated.
The linkage disequilibrium (LD) for each polymorphism was calculated with the PLINK genome association analysis program [39]. The analysis of the haplotype frequencies and their association with the responses analyzed in the study was performed using the snpStats program, a web-based tool for the analysis of association studies [40][41][42][43][44].

EULAR Response
In the bivariate analysis, a higher EULAR response was found in patients with fewer years of disease duration, who were administered the drug as monotherapy, without concomitant GCs and who had been receiving previous bDMARDs treatment for a shorter period. As for clinical variables, patients responded better to the therapy when their initial basal values for DAS28, NPJ, NIJ, PVAS, ESR and HAQ were lower (values are detailed in Table S16). The multivariate analysis showed that the independent variables associated with EULAR response at 6 months were lower duration of previous bDMARDs (OR = 0.97; CI95% = 0.95-0.99) and lower baseline values of the DAS28 (OR = 0.52; CI95% = 0.30-0.87) and PVAS (OR = 0.95; CI95% = 0.91-0.98) clinical variables. The multivariate analysis results are shown in Table 3.

EULAR Response
In the bivariate analysis, a higher EULAR response was found in patients with fewer years of disease duration, who were administered the drug as monotherapy, without concomitant GCs and who had been receiving previous bDMARDs treatment for a shorter period. As for clinical variables, patients responded better to the therapy when their initial basal values for DAS28, NPJ, NIJ, PVAS, ESR and HAQ were lower (values are detailed in Table S16). The multivariate analysis showed that the independent variables associated with EULAR response at 6 months were lower duration of previous bDMARDs (OR = 0.97; CI 95% = 0.95-0.99) and lower baseline values of the DAS28 (OR = 0.52; CI 95% = 0.30-0.87) and PVAS (OR = 0.95; CI 95% = 0.91-0.98) clinical variables. The multivariate analysis results are shown in Table 3.

Low Disease Activity (LDA)
In the bivariate analysis, LDA was found to be associated with a shorter period of treatment with previous bDMARDs and with administration of ABA by the SC route. Low disease activity was also associated with lower baseline levels of DAS28, PVAS and HAQ (values are detailed in Table S17). The CTLA-4 rs5742909-T allele showed a tendency toward association with LDA (T vs. CC; OR = 6.67; CI 95% = 0.84-52.74; Table S17). After performing the multivariate analysis, we found that a lower initial DAS28 value (OR = 0.69; CI 95% = 0.49-0.96; Table 3) indicated a tendency to a better therapeutic response.

Remission
In the bivariate analysis, disease remission was related to shorter disease duration, bionaive patients and also shorter duration of previous BTs, as well as a lower baseline value of the NPJ, NIJ, PVAS and HAQ clinical variables (values are detailed in Table S18). No association was found between the polymorphisms studied and remission of the disease at 12 months. The multivariate analysis showed a significant association with earlier age at the start of ABA therapy (OR = 0.96; CI 95% = 0.92-0.99; Table 3), lower number of previous BTs (OR = 0.56; CI 95% = 0.34-0.92; Table 3) and lower baseline PVAS (OR = 0.95; CI 95% = 0.93-0.98; Table 3).

Discussion
Abatacept is a biologic drug used as a first-line treatment in patients diagnosed with RA who have not reached a state of disease remission after administration of previous DMARDs. Interindividual variability in response to ABA is very wide in patients with RA. Our study included 105 patients diagnosed with RA and treated with ABA as first-/second-/third-or-more-line therapy for at least 6 months, and 92 maintained the treatment for at least 12 months. As regards the effectiveness of ABA, bionaive patients showed a higher EULAR response and greater disease remission after ABA treatment compared to patients who had already undergone previous BTs (78.57% vs. 46.74%; 64.29% vs. 30.43%; after 12 months of ABA). Our results are consistent with a study performed in 2716 Caucasian RA patients treated with ABA, in which the bionaive patients showed a higher EULAR response after 6 (OR-adjusted = 3.59, CI 95% = 2. 25-5.72) and 12 months (OR-adjusted = 4.29, CI 95% = 2.77-6.65) of ABA treatment [45]. Moreover, in the multivariate analysis of our results, the number and duration of previous BTs were significantly associated with remission of the disease and EULAR response after 12 and 6 months of treatment, respectively. Numerous observational studies and clinical trials have reached similar conclusions [46][47][48][49]. This hypothesis may be justified, given that the low effectiveness of BTs in many patients causes autoimmune and inflammatory processes to be triggered more rapidly, leading to a gradual and irreversible progression of the disease. Thus, BTs used in advanced phases of the disease lose their efficacy due to the extensive development of the pathology. Hence the importance of diagnosing the disease early and looking for biomarkers to help select the most effective therapy for each patient. Regarding biochemical markers, several studies have described the presence of positive ACPA as constituting an ABA treatment response factor by inhibiting T-lymphocyte activation, and consequently autoantibody production [50,51]. A study based on the "Orencia and Rheumatoid Arthritis" registry, which included 1003 Caucasian patients with RA, showed significant differences in ACPA values between patients with satisfactory and unsatisfactory EULAR responses (75.9% vs. 62.2%: p = 0.001) [51]. However, our study showed a tendency to a significant relationship between positive ACPA values and EULAR response after 6 months of treatment with ABA.
Polymorphisms in the genes associated with the drug's mechanism of action could explain the substantial variability of response in patients treated with ABA. In this study, we determined the role of SNPs on the CD80, CD86 and CTLA-4 genes in the treatment response. ABA is a fusion protein that acts by selectively modulating T-lymphocyte activation [52]. This modulation is performed physiologically by CTLA-4 [52]. Several SNPs located on this gene have been investigated for their association with autoimmune diseases [22,23,28,53]. The CTLA-4 rs5742909 polymorphism (C>T) is in the promoter region of exon 1 of the CTLA-4 gene, affecting the ATG transcription initiation codon [54]. This SNP could alter the transcription factor binding site [22,[55][56][57]. In addition, the CTLA-4 rs5742909-T allele has been linked to increased RNAm expression and higher levels of the CTLA-4 protein [53,58]. This theory supports the results obtained in our study, since the CTLA-4 rs5742909-T allele was associated with a higher EULAR response and lower LDA ratios after 12 months of ABA treatment, in the multivariate analysis. This SNP is presumably related to protection against autoimmunity and progression of the inflammatory response, which would explain the fact that patients who carry it show a greater response to ABA treatment [56,59]. A previous study analyzed the relationship of ABA with the CTLA-4 rs5742909 polymorphism in 39 Caucasian patients after 6 months of treatment, finding no significant association (p > 0.05) [18]. A study conducted in 200 Caucasian patients diagnosed with RA found higher levels of CTLA-4 in those carrying the CTLA-4 rs5742909-CT genotype (CT vs. CC; p < 0.001) [60]. Another of the CTLA-4 gene polymorphisms that have been studied is rs3087243 (G>A), located in exon 4 in the 3 UTR region [61]. Previous research has associated this SNP with a change in the levels of RNAm expression of the CTLA-4 molecule, also affecting its various isoforms [62,63]. It has also been observed that it could affect the peptide glycosylation level, resulting in a non-functional CTLA-4 molecule [23]. Specifically, the CTLA-4 rs3087243-G allele has been associated with a reduction in CTLA-4 production [62,64]. In our study, patients carrying the CTLA-4 rs3087243-G allele showed a higher EULAR response and greater LDA at 12 months. Furthermore, the CTLA-4 rs3087243-GG (GG vs. AA/AG) genotype was associated with greater remission of the disease after 6 months of treatment with ABA in the bivariate analysis; there was no significant association in the multivariate analysis. A study carried out in 200 patients with RA and 200 healthy Caucasian controls showed an association between the CTLA-4 rs3087243-GG and CTLA-4 rs3087243-GA genotypes and higher CTLA-4 levels in healthy controls (p < 0.001) [60]. Moreover, no previous study has shown significant results relating to the association between the CTLA-4 rs3087243 SNP and response to ABA [18]. Another of the SNPs studied was CTLA-4 rs231775 (A > G). In this case, a change from Thr to Ala occurs at position +49 of exon 1 of the gene [65]. This change of amino acid could cause incomplete glycosylation of the protein in the endoplasmic reticulum, which would affect production of the soluble form of the molecule. Previous research has shown a reduction in CTLA-4 production associated with the CTLA-4 rs231775-GG genotype [66][67][68]. In our study, we found that patients carrying the CTLA-4 rs231775-G allele showed a higher EULAR response as well as greater LDA after 12 months of treatment with ABA in the multivariate analysis. Remission of the disease after 6 months of treatment was also greater in patients with the CTLA-4 rs231775-G allele in the bivariate analysis, but in this case the association was not maintained in the multivariate analysis. In a study conducted in 715 Caucasian patients with multiple sclerosis and 527 healthy controls, it was observed that lesser CTLA-4 expression was associated with the least aggressive form of the disease, which could be explained by the fact that the CTLA-4 molecule promotes an increase in reservoirs of memory T lymphocytes, because it reduces induced T-lymphocyte apoptosis [69,70]. Consequently, lesser CTLA-4 production would result in a reduction in memory T lymphocytes, leading to less inflammation and a lower probability of inflammatory relapses [69]. Given the large inflammatory component involved in the development of RA, this hypothesis could explain the greater therapeutic response in patients who express lower levels of CTLA-4.
Abatacept binds selectively to the CD80/CD86 complex, like CTLA-4, preventing the latter from binding to the CD28 receptor and thereby suppressing T-cell activation [52]. Polymorphisms in these proteins could produce conformational changes in the receptor, affecting its interaction with ABA and causing the therapeutic response. The CD86 rs1129055 polymorphism (G > A) could give rise to a change in the protein's phosphorylation level, influencing the process of negative regulation of T-lymphocyte proliferation that occurs after its interaction with CTLA-4 [21,74]. In our study, we did not find a significant association with the response variables studied. Our results are consistent with the only previous study carried out in 39 Caucasian patients, which analyzed the association between the CD86 rs1129055 and therapeutic response to ABA, since in this case, too, no significant relationship with EULAR response was found after 6 months of ABA treatment [18]. Similarly, no association was found between the CD80 rs57271503 polymorphism and therapeutic response to ABA in our patients after 6 and 12 months of treatment. These results are in line with those obtained in the previous study comprising 39 Caucasian patients diagnosed with RA, which analyzed the relationship between the CD80 rs57271503 SNP and EULAR response [18]. No other studies have evaluated the role of these SNPs in the effectiveness of ABA in patients diagnosed with RA.
The limitations of our study include the limited sample size, which could be responsible for the loss of statistically significant association for the CD80 rs57271503 and CD86 rs1129055 SNPs. Thus, our sample size has a statistical power to detect genetic association of 60%. Nevertheless, all the patients were recruited from the same hospital cohort, following the same therapeutic protocols, by the same team of rheumatologists, which ensured the homogeneity and reliability of the clinical variables collected. All patients diagnosed during the study period were recruited, ensuring the representativeness of the sample. Despite the limited sample size, the effects observed in these patients were clear. Further studies in larger cohorts will be needed to confirm the prognostic value of the biomarkers, particularly the polymorphisms of the CTLA-4, CD80 and CD86 genes.
These results suggest that the CTLA-4 rs3087243, rs231775 and rs5742909 polymorphisms could act as predictors of response to ABA treatment in patients diagnosed with RA.

Conclusions
Patients with RA treated with ABA showed a better EULAR response and LDA rates when they had the CTLA-4 rs5742909-T or CTLA-4 rs231775-G polymorphisms. As regards the clinical variables, greater remission was observed after 12 months of treatment in patients who initiated ABA treatment earlier, with a lower number of previous BTs and a lower PVAS value at the start of treatment. Likewise, shorter duration of previous BTs is associated with a higher EULAR response after 6 months of treatment with ABA.
Supplementary Materials: The following are available online at http://www.mdpi.com/2075-4426/10/4/220/s1, Figure S1: Linkage disequilibrium, Table S1: Hardy-Weinberg equilibrium,  Table S14: Haplotype association with LDA at 12 months ABA, Table S15: Haplotype association with remission at 12 months ABA adjusted by age at ABA start, number of previous BTs, PVAS, Table S16: Predictors of EULAR response at 6 and 12 months of treatment with abatacept in rheumatoid arthritis patients (bivariate analysis), Table S17: Predictors of LDA at 6 and 12 months of treatment with abatacept in rheumatoid arthritis patients (bivariate analysis), Table S18: Predictors of remission at 6 and 12 months of treatment with abatacept in rheumatoid arthritis patients (bivariate analysis). Funding: This work was partly supported by a contract for Noelia Márquez Pete from the University of Granada and the Fundación de Investigación Biosanitaria de Andalucía Oriental (FIBAO). University Hospital Virgen de las Nieves Biobank was supported by grants co-funded by ERDF funds (EU) from the Instituto de Salud Carlos III (PT13/0010/0039).

Acknowledgments:
The authors would like to acknowledge all the patients which participated in the study, the professionals from the University Hospital Virgen de las Nieves that contributed to the management of the samples, especially the nursing team from the Rheumatology Department and the professionals of outpatient Pharmacy Service. The results of this research study are part of the doctoral thesis to be submitted by Noelia Márquez Pete at the University of Granada, carried out within the "Clinical Medicine and Public Health" doctoral program.

Conflicts of Interest:
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported, and that there are no competing financial interests in relation to the work described in this article.