Associations of HLA Polymorphisms with Anti-SARS-CoV-2 Spike and Neutralizing Antibody Titers in Japanese Rheumatoid Arthritis Patients Vaccinated with BNT162b2

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes Coronavirus Disease 2019. Anti-SARS-CoV-2 spike (S) and neutralizing antibodies (Abs) are measured to evaluate the efficacy of vaccines. Human leukocyte antigen (HLA) may be associated with vaccine efficacy. Here, we investigated the association of HLA polymorphisms with the production of anti-SARS-CoV-2 S or neutralizing Abs in vaccinated rheumatoid arthritis (RA) patients in Japan. Genotyping of DRB1 and DQB1 was conducted in 87 Japanese RA patients vaccinated with BNT162b2. Associations of allele or haplotype carrier frequencies with anti-SARS-CoV-2 S or neutralizing Abs were examined. DRB1*12:01 was significantly positively associated with the production of S Ab (p = 0.0225, odds ratio [OR] 6.08, 95% confidence interval [CI] 1.32–28.03). The DQB1*03:01 allele carrier frequency tended to be higher in high responders of S Ab. Allele carrier frequencies of DRB1*15:01 (p = 0.0102, OR 9.26, 95% CI 1.65–52.01) and DQB1*06:02 (p = 0.0373, OR 7.00, 95% CI 1.18–41.36) were higher in responders of neutralizing Ab. Haplotype and two-locus analyses of DRB1 and DQB1 suggested that DRB1 alleles were the primary drivers of these associations. Logistic regression analysis showed associations of these alleles independent of clinical characteristics. Independent associations were found between HLA alleles and anti-SARS-CoV-2 Ab production by vaccinated RA patients.


Introduction
Coronavirus Disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which emerged as a pandemic starting in December 2019 [1]. A quantitative reverse transcription polymerase chain reaction is conducted for a definitive diagnosis. Serum anti-SARS-CoV-2 nucleocapsid (N) antibody (Ab) titers are quantified in order to detect a previous history of COVID -19. The presence of anti-SARS-CoV-2 spike (S) Ab in sera of vaccinated individuals is indicative of vaccine responsiveness when the S protein sequence of SARS-CoV-2 is used in COVID-19 mRNA vaccines. Anti-SARS-CoV-2 neutralizing Abs are titered to estimate the protective efficacy of Abs against infection. Blocking enzyme-linked immunosorbent assays has been employed recently to detect neutralizing Abs [2][3][4]. It is thought that the long-term production of neutralizing Abs is important for protection against SARS-CoV-2 infection. Humoral responses against COVID-19 mRNA vaccines are quite variable between individuals, influenced inter alia by age, sex, the time between the last vaccination and blood collection, or adverse reactions to the vaccine [5][6][7]. Levels of anti-SARS-CoV-2 S and neutralizing Abs were reported to be lower in patients with rheumatic diseases including rheumatoid arthritis (RA) relative to healthy controls [8][9][10].
Human leukocyte antigen (HLA) polymorphisms were reported to be associated with the efficacy of vaccines against measles [11,12], hepatitis B [13][14][15], or influenza [16,17]. However, few studies on the association of HLA with levels of anti-SARS-CoV-2 S Abs have been reported [18,19]. Hence, we investigated the association of HLA polymorphisms with levels of anti-SARS-CoV-2 S-specific or neutralizing Abs in vaccinated RA patients in Japan.

Patients and Sera
Eighty-seven RA patients were recruited at the National Hospital Organization Tokyo National Hospital. All RA patients fulfilled the American College of Rheumatology Criteria for RA [20] or Rheumatoid Arthritis Classification Criteria [21]. None of the RA patients had any history of COVID-19 before serum collection. All these patients had been vaccinated twice against SARS-CoV-2 with mRNA vaccine BNT162b2 (Pfizer, New York, NY, USA), and sera were then collected prior to the third vaccination.
The study protocol was reviewed and approved by The Research Ethics Committee of the National Hospital Organization Tokyo National Hospital. Written informed consent was obtained from each RA patient. This study was conducted in accordance with the principles expressed in the Declaration of Helsinki.

Detection of Anti-SARS-CoV-2 N, S, and Neutralizing Abs
The IgG fraction of anti-SARS-CoV-2 N Abs was measured by Anti-SARS-CoV-2 N IgG chemiluminescent enzyme immunoassays (Fujirebio, Hachioji, Japan). At a cut-off value of 1.0 U/mL, none of the RA patients were positive for the anti-SARS-CoV-2 N Abs. Anti-SARS-CoV-2 S Ab titers were quantified by Elecsys Anti-SARS-CoV-2 S (Roche Diagnostics, Mannheim, Germany) detecting Abs including IgG against the receptor-binding domain on the S1 subunit of SARS-CoV-2. The cut-off value was 0.8 U/mL. RA patients with high levels of anti-SARS-CoV-2 S Abs were classified as high responders (S high, top 25th percentile of anti-SARS-CoV-2 S Ab distribution) or low responders (S low, bottom 75th percentile of anti-SARS-CoV-2 S Ab distribution). Anti-SARS-CoV-2 neutralizing Abs (neu) were quantified using SARS-CoV-2 Neutralization Antibody Detection Kits (Medical & Biological Laboratories Co., Ltd., Tokyo, Japan). The degree of neutralization was calculated as follows: (1-optical density value of sample/optical density value of blank) × 100 (%). RA patients with high levels of anti-SARS-CoV-2 neutralizing Ab were classified as high responders (Neu high, top 25th percentile of anti-SARS-CoV-2 neutralizing Ab distribution) or low responders (Neu low, bottom 75th percentile of anti-SARS-CoV-2 neutralizing Ab distribution). The results of anti-SARS-CoV-2 N, S, and neutralizing Ab titers for the RA patients were previously reported [22].

Genotyping
HLA genotyping of DRB1 and DQB1 loci was performed by the polymerase chain reaction with reverse sequence-specific oligonucleotide probes (WAKFlow HLA typing kits, Wakunaga, Akitakata, Japan), using the Bio-Plex system (Bio-Rad, Hercules, CA, USA).

Statistical Analysis
Data were analyzed by Fisher's exact test using 2 × 2 contingency tables or Mann-Whitney's U test. Deviation from Hardy-Weinberg equilibrium was detected by Genepop (http://genepop.curtin.edu.au/ (accessed on 15 November 2022) [23]. Associations of HLA allele, carrier frequencies or haplotype carrier frequencies were tested by Fisher's exact test using 2 × 2 contingency tables. To estimate the primary associations of alleles, two-locus analysis was employed with Fisher's exact test using 2 × 2 contingency tables. Multiple logistic regression analysis under the additive model was employed to examine whether HLA alleles are independently associated with the production of anti-SARS-CoV-2 S Abs or neutralizing Abs in RA patients. Simple linear regression analysis was also employed to estimate the association of a rheumatoid factor with anti-SARS-CoV-2 neutralizing Abs in RA patients.

Clinical Characteristics of RA Patients
Clinical manifestations of RA in the patient cohort are listed in Table 1. SARS-CoV-2 S Ab low responders were older than the high responders. The mean interval between the last vaccination and blood collection was longer in the low-responder group. Number or average values of each group are shown. Standard deviations or percentages are shown in parentheses. Differences were tested by Mann-Whitney's U test or Fisher's exact test using 2 × 2 contingency tables. * Fisher's exact test was employed. Ab: antibody, S: spike, RA: rheumatoid arthritis, S high: high responders for anti-SARS-CoV-2 S Abs, S low: low responders for anti-SARS-CoV-2 S Abs, Neu high: high responders for anti-SARS-CoV-2 neutralizing Abs, Neu low: low responders for anti-SARS-CoV-2 neutralizing Abs, DMARD: disease-modifying anti-rheumatic drug, csDMARD: conventional synthetic DMARD, bDMARD: biological DMARD, and tsDMARD: targeted synthetic DMARD.

Logistic Regression Analysis of HLA Alleles and Clinical Characteristics
To exclude the possibility of an influence of the clinical characteristics on the production of anti-SARS-CoV-2 Abs, conditional logistic regression analysis of HLA alleles and these characteristics was conducted ( Table 6). The results of the logistic regression analysis of DRB1*12:01 and clinical characteristics suggested an independent association of DRB1*12:01 with the production of S Ab. Age remained associated with anti-SARS-CoV-2 S Ab levels when conditioned on the other factors.
The results of the logistic regression analysis of DRB1*15:01 and clinical characteristics also suggested an independent association of DRB1*15:01 with the production of neutralizing Ab ( Table 7). The time between the last vaccination and blood collection remained associated with anti-SARS-CoV-2 neutralizing Ab levels when conditioned on the other factors. Thus, anti-SARS-CoV-2 S Ab titers were higher in RA patients with DRB1*12:01, while anti-SARS-CoV-2 neutralizing Ab levels were higher in those with DRB1*15:01.

Discussion
The present study revealed that DRB1*12:01 and DRB1*15:01 were associated with anti-SARS-CoV-2 S and neutralizing Ab levels, respectively, in Japanese RA patients vaccinated with BNT162b2. Although DRB1 and DQB1 are in strong linkage disequilibrium, haplotype and two-locus analyses of DRB1 and DQB1 suggested the primary roles of DRB1 alleles. These associations were independent of the clinical manifestations of the RA patients.
The association of HLA with anti-SARS-CoV-2 S Abs had been previously investigated in 56 healthcare workers [19]; no association was detected. However, in silico DRB1*15:01 was predicted to have a higher number of strong binding peptides derived from the entire sequence of the SARS-CoV-2 S protein [19]. HLA associations with anti-SARS-CoV-2 S Abs had also been analyzed in 100 Japanese hospital workers [18] and DQA1:03:03 was reported to be associated with the production of greater amounts of anti-SARS-CoV-2 S Abs. In that study, the association of DRB1*12:01 with higher levels of anti-SARS-CoV-2 S Abs was not reported. An additional consideration is that there is no linkage disequilibrium between DRB1*12:01 and DQA1:03:03 in Japanese populations [24]. These results are therefore not consistent with those of the present study, but RA patients clearly constitute an immunologically different population. This could be the reason for such different results, because anti-SARS-CoV-2 S Ab is sometimes not produced at all in certain patients with rheumatic diseases after vaccination [25]. Although haplotype and two-locus analyses suggested the primary role of DRB1 alleles, other genes in linkage disequilibrium with DRB1-DQB1 loci might also increase the production of anti-SARS-CoV-2 Abs. Adult-onset Still's disease was associated with DRB1*12:01 and DRB1*15:01 in Japanese populations [26,27]; several patients with adult-onset Still's disease after being vaccinated against SARS-CoV-2 were reported [28][29][30]. It was known that drug-induced adverse effects are common in adultonset Still's disease. These data suggested that these DRB1 alleles might confer higher drug responsiveness in conditions with inflammation.
The present study on the association of HLA with anti-SARS-CoV-2 S and neutralizing Ab in vaccinated RA patients does have some limitations. The sample size is modest and this is a single-center study performed in Japan. The distribution patterns of HLA alleles are different in other ethnic populations. Other loci in the HLA region would influence the production of anti-SARS-CoV-2 Abs. Larger multiethnic studies on the whole HLA region are warranted to confirm the findings of the present study. In the future, associations of HLA with the roles of memory B or memory T cells could be examined in vaccinated RA patients [10,31,32].

Conclusions
In conclusion, in the present study, independent associations of HLA alleles with anti-SARS-CoV-2 Ab levels were found in Japanese RA patients vaccinated with BNT162b2.