Comprehensive genetic and functional analysis of FcÎ³Rs in rituximab therapy for autoimmunity reveals a key role for FcÎ³RIIIa on NK cells

B cell depletion using rituximab is widely used to treat autoimmune diseases, but patient response varies. The efficacy of rituximab is limited by the efficiency of depletion. Strategies to improve response include altering rituximab dosing, switching anti-CD20-mAb, alternative B cell targets, or non-B cell targeted therapies. Implementing an appropriate strategy requires understanding of the mechanism(s) of resistance to depletion and, if this varies between individuals, a means to test for it. Rituximab kills B cells via a variety of Fc{gamma} receptor (Fc{gamma}R)-dependent mechanisms, including antibody-dependent cellular cytotoxicity (ADCC), as well as non-Fc{gamma}R mechanisms. We conducted a longitudinal cohort study in rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) using two national registries. Qualitative and quantitative FCGR functional variants were measured using multiplexed ligation-dependent probe amplification, supplemented by novel FCGR2C assays. We provide consistent evidence that FCGR3A, specifically increased number of copies of the FCGR3A-158V allele, was the major Fc{gamma}R gene associated with rituximab response, including clinical response in RA and SLE and depth of B cell depletion in the combined cohort. In SLE, we provide preliminary data suggesting increased FCGR2C ORF copies were also associated with improved clinical response. Furthermore, we demonstrated the impact of disease status and concomitant therapies on both natural killer cell Fc{gamma}RIIIa expression and rituximab-induced ADCC; demonstrating increased Fc{gamma}RIIIa expression and FCGR3A genotype were independently associated with clinical response and B cell depletion. Our findings highlight the importance of enhancing Fc{gamma}R-effector functions, may help stratify patients, and support ongoing development of next-generation CD20 depleting therapeutics.


Novel FCGR2C QSV assay and functional interpretation of FCGR2C genotyping 1
To supplement the MLPA panels, we developed a novel FCGR2C QSV assay to more accurately determine 2 FCGR2C copy number. To biologically validate this assay, we sought to determine whether expression of 3 FcγRIIc (CD32) on NK cells was associated with the number of copies of classical FCGR2C ORF alleles 4 (15). As a control comparison, we determined NK cell CD32 expression in individuals with a specific 5 FCGR locus rearrangement that has previously been shown to encode FcγRIIb expression on NK cells (20).

6
This CNR1 deletion ( Figure 1A) involves a deletion of both FCGR2C and FCGR3B with the remaining 7 FCGR2C gene bearing a STP allele.

8
Using early RA patients prior to treatment, we observed increased NK cell CD32 expression in those with 9 carriage of the classical FCGR2C ORF allele (p<0.0001, Figure 1B). Individuals carrying a deletion of 10 CNR1 demonstrated high mean levels of NK cell CD32 expression ( Figure 1B) with CD32 expression 11 present on the majority of NK cells (Figure 1C), which suggests FcRIIb was expressed on these NK cells. cORF were genotyped where matched NK cell CD32 (clone KB61) expression data were available. A 8 particular locus rearrangement including a deletion of one copy of FCGR2C and one copy of FCGR3B 9 (CNR1 del) and a STP allele on the remaining copy of FCGR2C, has been described to encode expression (p=0.04, not adjusted for multiple testing), with an inferior response for individuals homozygous for the H 1 allele (p=0.03), but no evidence to support an association between rs9427399 (Q27W) or SNP I123T 2 (rs1050501) in FCGR2B with any clinical response measures. The rs396991 (F158V) variant in FCGR3A 3 was first analysed at the genotypic level with some evidence that the V allele was associated with a better 4 response, particularly for SJC and 2C-DAS28CRP (Table 3); carriers of at least one V allele had 5 fewer 5 swollen joints than those with only F alleles (p=0.01) and 2C-DAS28CRP was on average 0.29 units lower 6 than those with only F alleles (p=0.02), after adjusting for baseline measures. Compared to patients with 7 two copies of FCGR3A (the majority), those with duplications were found to have a better clinical response 8 on the basis of 2C-DAS28CRP (p=0.03). There was also a suggestion that those with deletions had a worse 9 outcome, although numbers were very small (n=9). These facets were combined by examining the effect of 10 the number of copies of the V allele. Higher numbers of the FCGR3A-158V allele were associated with 11 better response in the form of SJC (p=0.01) and 2C-DAS28CRP (p=0.02). In a model including both the 12 number of copies of the V allele and the number of copies of the F allele, there was some evidence that 13 copy number (or number of copies of the F allele) also contributed to response when the 2-component DAS 14 was evaluated (p=0.04).

15
Integrated copy number and functional SNP analysis was also used to analyse quantitative and qualitative 16 FCGR2C and FCGR3B variation. There was no association between either classical STP/ORF 17 polymorphism (rs10917661) in FCGR2C or the NA1/NA2 haplotype in FCGR3B and clinical response at homozygosity for the H allele of FCGR2A was associated with an increased odds of BILAG MCR at 6 1 months (p=0.02), however, there was no effect of FCGR2A rs9427399 (Q27W) on clinical response, nor 2 I123T (rs1050501) in FCGR2B. Concordant with our findings in RA, the rs396991 (F158V) variant in 3 FCGR3A was associated with increased odds of BILAG MCR when analysed at the genotypic level, with 4 homozygosity for the V allele demonstrating a twofold improved response (p=0.03). Carriage of at least 5 one copy of the FCGR3A-158V allele was associated with a 1.8 fold improvement in odds of BILAG 6 response (p=0.04) and 1.9 fold improvement in odds of BILAG MCR (p=0.02). In contrast to RA, we did 7 not find any evidence of an association between FCGR3A copy number and clinical response, most likely 8 because the majority of patients carried two copies of FCGR3A (248/262). For each copy of FCGR3A V 9 allele there were increased odds of BILAG MCR at 6 months (p=0.01). At the genotypic level, the FCGR2C 10 classical STP/ORF genotype was associated with 2 fold increased odds of BILAG MCR at 6 months 11 (p=0.02), with carriage of the ORF allele being associated with a 2.2 fold improvement in odds of BILAG 12 MCR (p=0.02). Furthermore, duplications of FCGR2C had a 3.1 fold increased odds of BILAG MCR at 6 13 months (p=0.02), and a 2 fold improved response per copy of the FCGR2C ORF allele (p=0.02). For 14 FCGR3B, patients with duplications also had the highest odds of achieving BILAG MCR compared to 15 those with two copies (p=0.01). These associations with FCGR2C and FCGR3B copy number may not be 16 independent as 17/22 subjects with a FCGR3B duplication also had a FCGR2C duplication indicating 17 strong linkage disequilibrium between the genes. However, we found no evidence of association between 18 other FCGR3B allelic variation and SLE clinical response, which may indicate that the principle association

22
There were insufficient individuals from non-Caucasian backgrounds with 2 copies of FCGR3A, FCGR2C 23 and FCGR3B to accurately determine whether this group had different patterns of linkage disequilibrium 24 to the Caucasian population. A sensitivity analysis was therefore performed in Caucasians only, with 25 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Association of FCGR F158V genotype and copy number with complete B cell depletion in RA and SLE
Since peripheral B cells were analysed using HSFC in the Leeds cohorts only, data from both RA and SLE 6 were combined to increase statistical power (n=413). There was no significant difference in depth of 7 depletion between the disease groups (p=0.26), although there was a significant difference in age (p=0.002).

8
The baseline clinical variables associated with depletion in the combined group are shown in 9 Supplementary  CNV (FCGR2A and FCGR2B), no association with complete B cell depletion was observed. For the CNV 16 affected genes, higher number of copies of the FCGR3A V allele were associated with an increased odds 17 of B cell depletion (p=0.02), in both the adjusted and unadjusted analyses.

19
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 28, 2021. ;https://doi.org/10.1101/2021 doi: medRxiv preprint 1 Figure 2: Association of FCGR genotype and copy number with complete B cell depletion. Odds ratio 2 (OR), 95% confidence intervals (CI) and p-value for the effect of the indicated genotype or copy number 3 on complete B-cell depletion at 2 weeks post-rituximab, compared with reference genotype. The dots 4 represent the OR and the error bars denote the CI. All tests were performed using univariable logistic 5 regression, adjusted for age, concomitant disease-modifying anti-rheumatic drug, including 6 hydroxychloroquine, and baseline plasmablast count. year and DMARD-naïve; n=46) and established RA (diagnosed >2 years and receiving DMARDs; n=20).

3
FcγRIIIa expression differed between the three groups (p=0.01) and was reduced in early RA (p<0.001), 4 but not established RA (p=0.22), when compared to HC ( Figure 3A). This differential expression was not 5 secondary to altered ratios of CD56 bright to CD56 dim NK cells between RA patients and HC (Supplementary

3
The reduced degranulation in RA compared with HC, prompted us to explore the impact of DMARDs on 4 NK cell degranulation ex vivo ( Figure 3F). A significant reduction in the number of NK cells degranulating 5 was observed after 14 weeks of methotrexate in RA, compared to baseline, with the most marked reduction 6 occurring in individuals homozygous for the FCGR3A-158F allele (p=0.02).

7
Inclusion of a CD16 blocking antibody inhibited rituximab-induced NK cell degranulation in HC (p=0.003) 8 and RA (p=0.03, Figure 3G), and thus supported a major role for FcγRIIIa in NK cell-mediated ADCC ex 9 vivo. No significant reduction in NK cell degranulation was observed with CD32 blockade in HC or RA 10 (p=0.22, p=0.60; Figure 3G). None of these individuals had a deletion in CNR1 (FcRIIb expression) and 11 there was no clear association with FCGR2C genotype.

13
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Higher expression of FcRIIIa is associated with better clinical response and depletion in vivo
15 independent of NK cell number 16 We measured absolute NK cell numbers and NK cell FcγRIIIa expression on the day of the first infusion 17 of rituximab compared with subsequent depletion and response to explore these effector mechanisms in 18 vivo.

2
We measured NK cell FcγRIIIa expression in 18 RA patients and 17 SLE patients immediately prior to 3 rituximab and demonstrated a positive association between FcγRIIIa expression and EULAR response at 6 4 months (defined as improvement of at least 0.6 point to DAS28CRP≤5.1) in RA (p=0.03; Figure 4F), but 5 not BILAG MCR in SLE (p=0.55, Figure 4G). There was no difference in FcγRIIIa expression between 6 RA and SLE (p=0.14) and the groups were combined to explore the effect of FcRIIIa expression on B cell 7 depletion. Patients with complete depletion had higher NK cell FcγRIIIa expression at rituximab initiation 8 than those with incomplete depletion in RA and SLE (p=0.04: Figure 4H). 14 15 16 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 28, 2021.

1
We report the largest study, to date, of quantitative and qualitative functional variants at the FCGR genetic 2 locus in well-characterised RA and SLE cohorts, including subgroups where peripheral blood B cell (naive, 3 memory and plasmablast cell) levels were available before and after rituximab therapy. We provide 4 consistent evidence that FcγRIIIa is the major low affinity FcγR associated with both clinical and biological 5 (i.e. depth of B cell depletion) response to rituximab in both autoimmune diseases. More specifically, an 6 increase in the number of copies of the FCGR3A-158V allele, encoding the allotype with a higher affinity 7 for IgG1, was associated with improved responses. In addition, irrespective of FCGR3A genotype, we 8 observed that the higher the number of FCGR3A gene copies correlated with a better clinical response in 9 RA. To functionally support these genetic findings, we also demonstrated that FCGR3A genotype was 10 associated with NK cell-mediated degranulation in vitro, and increased NK cell FcγRIIIa expression was 11 associated with improved clinical response and depletion in vivo, thus providing further biological support 12 for our genetic studies.

13
In RA, we have presented data on the commonly used clinical endpoints of CRP, SJC and 3C-DAS28CRP.

14
We also included our more recently published 2C-DAS28CRP score (35) that includes revised weighting 15 of CRP and SJC to more closely reflect the more objective endpoint of ultrasound-proven synovitis, an 16 outcome measure that we have proposed as the RA disease activity measure of choice for genetic and 17 biomarker studies (36). In SLE we have used a BILAG-based endpoint rather than SLEDAI based. BILAG 18 is better for biomarker studies because, unlike the SLEDAI, it allows partial improvement, it allows severe 19 and moderate manifestations within each organ system to be weighted equally, and it does not include 20 serological componentsa serious confounder in studies of a B cell-targeted therapy.

21
Our studies have once again highlighted the challenges of analysing and interpreting the FCGR genetic 22 locus, even with commercially available assays. We present the first study that has undertaken a combined 23 quantitative and qualitative analysis of the FCGR locus to allow more comprehensive biological 24 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 28, 2021. ; interpretation. The association with FCGR3A was most apparent when analysed as the number of copies of 1 the 158V allele, consistent with previous studies in RA where traditional SNP analyses demonstrated 2 improved responses in those that carried the FCGR3A-158V allele (27, 37).

3
No consistent association signals were observed with other low affinity FCGR genes at this locus, 4 suggesting FcγRIIIa was the most important FcγR contributing to rituximab response. Increased NK cell-5 mediated degranulation was observed in individuals homozygous for the FCGR3A V allele, irrespective of 6 disease status, which combined with the association between FcγRIIIa expression on NK cells and in vivo 7 clinical response, independent of absolute NK cell numbers, supports ADCC being a major biological 8 mechanism. However, this does not preclude FcγRIIIa-mediated clearance of rituximab-opsonised B cells 9 by tissue macrophages and cells of the reticuloendothelial system by ADCP (38). Our data also reveal 10 potential disease or inflammation-specific factors that may impair ADCC, for example reduced NK cell 11 degranulation and reduced FcRIIIa expression in early RA patients compared with HC, which we showed 12 was correlated with serum IL-6 titre and autoantibody status, but not FCGR3A-158F/V genotype.

19
In SLE, higher number of copies of the FCGR2C ORF allele and FCGR2C gene duplications were also 20 associated with BILAG MCR at 6 months. At the time this study was carried out, the commercial MLPA 21 panels did not include probes for FCGR2C copy number per se and these panels were supplemented with 22 additional FCGR2C genotyping to aid interpretation. Our expression studies demonstrated that NK cell 23 CD32 expression correlated with carriage of the FCGR2C-ORF allele, but also revealed high expression 24 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 28, 2021. levels in individuals with simultaneous deletion of FCGR2C and FCGR3B, which leads to FcRIIb 1 expression. This rearrangement was observed in 52 RA and 62 SLE individuals in our genetics studies, 2 with no evidence it significantly impacted on depletion or response to therapy.

3
Our results have implications for the design of future clinical trials optimising rituximab therapy and 4 development of more effective B cell targeted strategies, which can be extrapolated to other B cell-mediated 5 autoimmune diseases. For example, patients with high copy number of FCGR3A-158V may be most 6 suitable for stratification to lower doses of rituximab (i.e. 500mg x 2) or even ultra-low dose regimen (i.e. 7 200mg x 1) (39) during repeat rituximab cycles. This may reduce long-term complications, such as late 8 onset neutropaenia (31) and infections secondary to hypogammaglobulinaemia (40). If successful this 9 would reduce drug costs and administration times. A previous clinical trial has shown that patients with 10 initial incomplete depletion can benefit from a higher dose of rituximab than currently licensed (1000mg x 11 3) (41). Entry into that study was determined after measuring initial depletion, but FCGR3A genotyping 12 may allow stratification prior to administering treatment. Secondly, confirmation that FcγRIIIa is the major 13 FcγR contributing to clinical response is of particular value to therapeutic antibody design and highlights 14 the need for the next generation of CD20 therapeutic antibodies to show equivalent ADCC potency in 15 individuals with both FcγRIIIa-158F/V allotypes, particularly when based on the native IgG1 background.

16
Indeed, there are many novel therapeutics in development with modified Fc regions that enhance ADCC.

17
Obinutuzumab was one of the first glycoengineered therapeutic mAbs to be FDA approved. Here control 18 of fucosylation during manufacture leads to increased ADCC, albeit at the expense of reduced CDC (42).

19
This type 2 CD20 mAb binds to a different CD20 epitope and is currently being investigated in a phase III (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

4
The challenge for pharmacogenetic studies is the assembly of large cohorts of sufficiently homogeneous 5 patients. Efforts were made in this study to minimize heterogeneity by using strict entry criteria and improve 6 consistency of time-points and definition of clinical response, however, some clinical and laboratory data 7 were missing, which is inevitable when large observation cohorts are utilised. Most notably, B cell data 8 were only available for the patient cohorts treated in Leeds. Hence, some analyses were performed by 9 combining data from RA and SLE. We have presented data on a number of outcome measures to support 10 future meta-analyses. For the RA cohort, 481 patients were genotyped. This sample size gives over 80%

Highly Sensitive Flow Cytometry
Peripheral blood B cell subsets were measured using HSFC at the accredited Leeds Haematological Genotyping 1 Genomic DNA was extracted from EDTA-anticoagulated whole blood using Qiamp mini spin columns 2 (Qiagen), Gene Catcher (ChargeSwitch, Thermo Fisher) and a manual phenol chloroform method. DNA 3 concentration was measured using UV spectrophotometry and samples diluted to 10ng/µl. 50ng total 4 template was used for each MLPA reaction. MLPA probe mix panels P110 and P111 (MRC-Holland,

5
Amsterdam, The Netherlands) were performed for every sample. Where a sample failed in the first reaction 6 due to insufficient template (Q fragment QC) the DNA template was concentrated using ethanol 7 precipitation and MLPA repeated. MLPA panels were supplemented with an in-house assay which yielded     improves on the nested model with overall copy number only, a likelihood ratio test was performed.

20
Since the RA and SLE cohorts were of mixed ethnicity, we assessed the potential for population 21 stratification by measuring pairwise LD between the relevant functional polymorphisms in the FCGR locus 22 for each sub group. We utilised Haploview to calculate r 2 LD between biallelic markers in individuals with 23 two copies of FCGR3A and FCGR3B (49).

24
Continuous variables were compared using Mann-Whitney test or Kruskal-Wallis H test, depending on data 25 distribution and number of independent groups for comparison. Spearman's test was used for all 26 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

5
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

17 18
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

9
All other authors declare no competing interest related to the work described in this manuscript.  18 All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 28, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021   1 Genes presented in chromosomal order on 1q23 centromere to telomere. 2 Coefficient, standard error (SE) and p-value for the effect of the indicated genotype or copy number on outcome at 6-months compared with baseline genotype or copy number. Positive coefficients for clinical response outcomes indicate a worse outcome. 3 FCGR3A, FCGR2C and FCGR3B are subject to copy number variation, analyses were performed according to biallelic genotype whereby the effect of heterozygosity and homozygosity for the rare allele were compared with homozygosity for the common allele. N: number where blood was taken immediately prior to first rituximab cycle.

Clinical outcomes: Rheumatoid arthritis
Clinical outcome measurements (SJC28, TJC28 and CRP, in mg/L) were taken at baseline (up to 6-weeks before the individual's rituximab treatment) and again at follow-up (6 months from the individual's first rituximab infusion, or 3 months if either the six month data were not available or the patient had received a further rituximab infusion prior to 6 months). Where CRP levels were recorded as below 5 mg/L, the lower limit of reliable detection, these were replaced by imputed values from a Uniform distribution from 0 to 5. Disease activity scores were calculated using the formula:

Treatment
All patients received a first cycle of therapy consisting of 100 mg of methylprednisolone and 1000 mg of rituximab given intravenously on days 1 and 14, with the exception of 3 RA patients who received 1g in total. All RA patients received rituximab MabThera® while 7/262 SLE patients received a biosimilar. Continuation of a stable dose or reduction of concomitant DMARDs and/or oral prednisolone was left to the clinicians' discretion with the aim to stop glucocorticoids if remission was achieved within 6 months.

Routine Laboratory assessments
All autoantibody and immunoglobulin assessments were determined using standard assays in the routine NHS diagnostics laboratory of each participating site. RA patients who ever had RF and ACPA titres of >40 iu/mL and >7 iu/mL, respectively, were defined as positive, to maintain consistency with previously published studies. For SLE, ANA was tested using indirect immunofluorescence and a panel of nuclear autoantibodies including anti-dsDNA and anti-ENA antibodies (Ro52, Ro60, La, Sm, and RNP). Complement (C3 and C4) and total serum immunoglobulin (IgM, IgA and IgG) titres were measured by nephelometry. Adult reference ranges are as follows: C3: 0.75-1.65 g/L, C4: 0.14-0.54 g/L), IgG (6-16 g/L), IgA (0.8-4 g/L) and IgM (0.5-2 g/L).

FCGR2C copy number assay
High sequence identity between FCGR2A, FCGR2B and FCGR2C prevents the MLPA probes in panels P110 and P111 from uniquely recognising the copy number variable FCGR2C without simultaneously hybridising to FCGR2B or FCGR2A. The resulting interpretation of FCGR2C gene copy number requires combinations of multiple probe intensities.