Genetic variations in low‐to‐medium‐affinity Fcγ receptors and autoimmune neutropenia in early childhood in a Danish cohort

Autoimmune neutropenia (AIN) in early childhood is caused by autoantibodies directed against antigens on the neutrophil membrane and is a frequent cause of neutropenia in children. Association of AIN with Fcγ receptor (FCGR) 3B variants is well described. In this study, we investigate genetic variations in the FCGR locus and copy number variation of FCGR3B. A total of 130 antibody‐positive AIN patients, 64 with specific anti‐HNA‐1a antibodies and 66 with broad‐reacting anti‐FcγRIIIb antibodies, were genotyped with a multiplex ligation probe assay and compared with healthy controls. Positive findings were confirmed with real‐time q‐PCR. We determined copy numbers of the FCGR2 and FCGR3 genes and the following SNPs: FCGR2A Q62W (rs201218628), FCGR2A H166R (rs1801274), FCGR2B I232T (rs1050501), FCGR3A V176F (rs396991), haplotypes for FCGR2B/C promoters (rs3219018/rs780467580), FCGR2C STOP/ORF and HNA‐1 genotypes in FCGR3B (rs447536, rs448740, rs52820103, rs428888 and rs2290834). Generally, associations were antibody specific, with all associations being representative of the anti‐HNA‐1a‐positive group, while the only association found in the anti‐FcγRIIIb group was with the HNA‐1 genotype. An increased risk of AIN was observed for patients with one copy of FCGR3B; the HNA genotypes HNA‐1a, HNA‐1aa or HNA‐1aac; the FCGR2A 166H and FCGR2B 232I variations; and no copies of FCGR2B 2B.4. A decreased risk was observed for HNA genotype HNA‐1bb; FCGR2A 166R; FCGR2B 232T; and one copy of FCGR2B promoter 2B.4. We conclude that in our Danish cohort, there was a strong association between variation in the FCGR locus and AIN. The findings of different genetic associations between autoantibody groups could indicate the presence of two different disease entities and disease heterogeneity.


INTRODUCTION
Primary autoimmune neutropenia (AIN) in early childhood is caused by antibodies directed against neutrophil-specific antigens, mostly located on immunoglobulin G (IgG) Fc receptor type 3b (FcγRIIIb) (Flesch & Reil, 2018). The condition is often benign, and most patients are in remission after 2-3 years (Bux et al., 1998). Information regarding the cause of this disease is limited due to the scarce data on the triggering aetiology.
IgG is the most abundant Ig class, constituting over 75% of circulating Ig (Bournazos et al., 2009). Fc-gamma receptors (FcγRs) are the cellular receptors for IgG, and binding of IgG complexes triggers various cellular immune effector functions. Six classic FcγRs are known in humans and are divided into one high-affinity receptor (FcγRI) and five low-to medium-affinity receptors (FcγRIIa, -b and -c and FcγRIIIa and -b). Most FcγRs are activating receptors capable of inducing a cellular response, with the exception of FcγRIIb, which is an inhibitory receptor (Nagelkerke, Schmidt, et al., 2019). The five genes (FCGR2A,FCGR2B,FCGR2C,FCGR3A and FCGR3B) encoding the low-to medium-affinity Fcγ receptors are located in a single cluster on chromosome 1q23.3, the FCGR2/3 locus. Genetic variations, causing functional changes, have been found in all five genes. These variations are associated with several conditions, such as autoimmune (Breunis et al., 2008;Lee et al., 2009;Lee et al., 2008;Li et al., 2010;Yuan et al., 2009), autoinflammatory (Asano et al., 2009;Khor et al., 2011;Onouchi et al., 2012) and infectious diseases (Adu et al., 2012;Chai et al., 2012), and the efficacy of immunotherapy (Ahlgrimm et al., 2011;Cartron et al., 2002;Hurvitz et al., 2012;Tamura et al., 2011;Treon et al., 2011). Genetic analysis of these variants is complex, deriving from a high degree of homology and linkage disequilibrium .
FcγRIIa (CD32a) induces many different cellular defence mechanisms and is found on monocytes, macrophages, dendritic cells, neutrophils and platelets. FcγRIIb (CD32b) is expressed on B cells, where it constitutes the only surface-expressed FcγR, and co-crosslinking of FcγRIIb with the B-cell receptor (BCR) inhibits activating signals induced by the BCR. FcγRIIb expression can also be detected on neutrophils but only in individuals with a 2B.4 promoter haplotype (Su et al., 2007). FcγRIIc (CD32c) is expressed on natural killer (NK) cells, neutrophils, monocytes and macrophages, but only in individuals who carry an open reading frame (ORF) of this receptor. However, in the majority of individuals, the gene is a pseudogene and is not expressed . The FcγRII class shares a characteristic structure that includes functional signalling motifs in their cytoplasmic domains and an immunoreceptor tyrosine-based activating motif for FcγRIIa and -c and an immunoreceptor tyrosine-based inhibition motif for FcγRIIb (Bournazos et al., 2009 (Bournazos et al., 2009).
The focus of this study is to investigate genetic variations in the lowto medium-affinity receptors and their association with AIN and the specificity of autoantibodies.

Study cohort
The study cohort included 130 patients diagnosed with autoantibodypositive AIN between 2004 and 2022 at the National Centre for Diagnostic AIN Testing, Department of Clinical Immunology, Aalborg University Hospital, Denmark. The inclusion criteria were the presence of neutropenia, an absolute neutrophil count below 1.5 × 10 9 cells/L in two repeated tests, age under 5 years at the time of diagnosis as suggested by Fioredda et al. (2022), and the presence of anti-neutrophil antibodies in the flow cytometric indirect granulocyte immunofluorescence test (Flow-GIFT) as previously described (Nielsen et al., 2021).
Patients with initial negative antibody screening were retested as suggested by Bux et al. (1998)
Interpretation of MLPA data does not only include a statistical estimation of the significance of the result, but also look at the probe ratio obtained and compare this with the arbitrary borders in Coffalyser.Net and with the thresholds given in the product description. This excludes situations in which a probe is seemingly statistically significantly different by coincidence. Similarly, the reaction quality and the variability of the reference probes and samples is evaluated with Coffalyser.Net. This protects against interpreting experimentally induced variability as true results. The SNP genotypes are concluded based on two or more mutation-specific probes from two individual tests confirming the result (Coffalyser.Net™ Reference Manual, MRC-Holland, The Netherlands).

Genotyping
Real-time polymerase chain reaction (PCR) using TaqMan  Copy numbers of FCGR3B found by MLPA in patients were also validated with this method, while controls found to have a copy number >2 with real-time q-PCR had their HNA-1 genotypes determined with MLPA. Data analysis was performed with CopyCaller® Software (Applied Biosystems).

Statistics
Statistical analysis was conducted using the statistical program Stata (Version 17.0; StataCorp, College Station, Texas). Gene frequencies were estimated by direct counting and they were compared using the Fisher exact test (Yao et al., 2019). Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Only SNPs with a frequency ≥1% are reported in this study. SNPs were tested for Hardy-Weinberg equilibrium with p values calculated based on the χ 2 distribution with 1 degree of freedom (Table S2). Genotype frequencies were compared with Cochran-Armitage trend test with p values calculated based on the χ 2 distribution with 1 degree of freedom. Bonferroni correction was used to adjust p values in case of multiple statistical testing.

Baseline characteristics
We included 130 patients diagnosed with AIN with a median age at diagnosis of 14.2 months (range, 3−54 months). The sex distribution was 49% females and 51% males. All AIN patients were positive for anti-FcγRIIIb antibodies, and of these had 49% anti-HNA-1a-specific antibodies. The patients were both investigated as a combined group and as two individual groups divided by their antibody specificity.
The two groups consisted of 64 patients which were anti-HNA-1apositive, and the remaining 66 patients which will be referred to as anti-FcγRIIIb-positive.
For FCGR2B I232T, we observed a significantly difference in genotype frequency (p = .015) and higher frequency of the 232I allele (OR = 1.83 (1.14-2.94)) among Danish AIN patients compared with the control group of European descent published by Nagelkerke et al.
Of the two antibody specificity groups, only the anti-HNA-1a-positive patients was significantly different from the European control group.
We observed a significantly lower frequency of the FCGR2B pro-  (Table S1).

HNA-1 alleles and FCGR3B copy numbers
Copy numbers from one to four of FCGR3B were observed in both Danish AIN patients and healthy Danish controls (see Table 2). There was a significant difference in the frequency of patients with only one copy in the anti-HNA-1a-positive group compared with the control group (OR = 3.15 (1.23-8.06)). We did not observe any associations between copy number and AIN for the combined group or for the anti-FcγRIIIb-positive patients.

DISCUSSION
In this study, we investigated the most relevant and functional SNPs cohort, and our group confirmed these findings and discovered that the presence of the HNA-1a genotype appeared to be responsible for all anti-HNA-1a-positive cases (Nielsen et al., 2021).
In the present study, we combined FCGR3B copy number and HNA-1 genotype data. With the use of MLPA, we were able to identify exact  According to Human Genome Structural Variation Consortium (HGSV) guidelines, we use amino acid numbering from the full protein (including signal peptides) and have done this throughout the manuscript. The position in the mature protein (without signal peptides) is shown between brackets for some of the SNPs that are commonly known by that position.  (3.17-11.17)) groups. Here, we observe that two copies of the HNA-1b genotype are an advantage against disease for both patient groups.
Interestingly, in the anti-HNA-1a-positive group, the HNA-1b genotype also appears to protect individuals who have both an HNA-1a and HNA-1b genotype, so HNA-1b seems to offer protection even in the presence of an HNA-1a genotype. In the group of individuals with more than three copies, the HNA-1aab and HNA-1abb genotypes were the most frequent types in the healthy blood donor control group, but interestingly, we did not observe any of these HNA genotypes in our patient cohort. Instead, we observed several cases with HNA-1aac (OR = 12.7 (1.4-114.4)), which seems to be a rare genotype in the Danish population, also based on a low frequency of HNA-1c (FCGR3B*03) in the population (Nielsen et al., 2012). A correlation between the HNA-1c genotype and increased copy number in Europeans has previously been published (Nagelkerke, Tacke, et al., 2019). We observed no individuals homozygous for HNA-1c, but we did observe both HNA-1ac and HNA-1bc genotypes in the patients with two copies, which was confirmed by both real-time PCR and MLPA. However, we cannot exclude the possibility that this can be a result of ethnicity and all HNA-1c genotypes found in our control group are indeed linked to a higher copy number of FCGR3B.
Neutrophils from homozygous FCGR2A 166H individuals have been shown to have increased phagocytosis and degranulation in response to bacteria and a higher binding affinity for IgG1 and IgG2 compared with FCGR2A 166R (Bruhns et al., 2009;Park et al., 1993;Trinchieri & Valiante, 1993). In resting neutrophils, FcγRIIa is in a low-affinity state, while binding is engaged by FcγRIIIb, but once neutrophils have been activated, FcγRIIa is converted to a high-affinity receptor that leads to FcγRIIa-dependent ligand binding and signalling (Nagarajan et al., 2000). The FCGR2A 166H variant has been found to be associated with autoimmune diseases such as Kawasaki disease (Duan et al., 2014) and childhood immune thrombocytopenia , the latter having many similarities to AIN.  (Su et al., 2007). We observed that lacking a 2B.4 promoter haplotype was associated with AIN with almost twice as high a risk compared with healthy controls, and that having one   (Blank et al., 2005;Su et al., 2004). A variant in FCGR2B, FCGR2B T232, has also been associated with SLE (Floto et al., 2005). Individuals with FCGR2B T232 are unable to inhibit activation receptors because the receptor is excluded from sphingolipid rafts, resulting in the unopposed proinflammatory signalling thought to promote SLE (Floto et al., 2005). The mechanism by which this single amino acid change in the transmembrane domain prevents access to lipid rafts is unclear. We observed comparable findings in our AIN cohort, where homozygosity for FCGR2B T232 was significantly associated with a higher risk of disease (OR = 2.12 (1.24-3.60)), especially for the anti-HNA-1a patients (OR = 3.08 (1.31-7.23)), but not for the anti-FcγRIIIb patients.
The FCGR2/3 locus contains a high degree of homology, and linkage disequilibrium needs to be taken into consideration for the identified associations. It is also important to highlight that CNV of FCGR3B includes deletion or duplication of a whole region, which also includes FCGR2C, and it is merely because of the HNA-1 association with AIN that we focus on CNV in FCGR3B. However, we cannot exclude the possibility that the association found for CNV in FCGR3B could also be related to other elements in the copy number region.
Homology between the three FcγRII receptor genes makes genotyping in this region difficult using standard techniques, which is why we chose to use the control group already published by Nagelkerke et al.
(2019) to compare with the findings in our study where a validated commercial real-time PCR assay was not available. This control group consisted of healthy individuals from Austria, Australia, the Netherlands and the United Kingdom who had all self-reported to be of European descent. However, it is a limitation to the findings of associations for FCGR2B I232T and FCGR2B promoter 2B.4 that we did not compare them to a Danish control group or tried to replicate our findings with another method, and the association to AIN must therefore be taken with precaution, until these results have undergone further replication.
In general, SNPs of Fcγ receptors are accompanied by the loss of inhibitory Fcγ expression and altered functions may result in unbalanced immunity and subsequently cause autoinflammation, which might be associated with susceptibility to AIN.
A novel finding in this study is the highly significant relationship between autoantibody specificity and genetic risk factors, a phenomenon previously described for other autoimmune diseases, such as RA and SLE, where several discrete subgroups can be defined based on autoantibody specificity (Reed et al., 2020). These subgroups also differ significantly in relation to genetic risk factors (Diaz-Gallo et al., 2022;Padyukov, 2022). We propose a similar heterogeneity in AIN. The presence of two distinct subgroups of patients with childhood AIN based on serological findings at the time of diagnosis and different genetic risk factors warrants further study.
Overall, we can conclude that there are several associations for AIN in the FCGR locus. However, with the exception of HNA-1a, our findings seem to only concern half of the patients diagnosed with AIN, those that have specific antibodies against HNA-1a. It is very interesting that the other half of the group seemed to be distinguished in regard to their genetic background. This indicates that there might be two different diseases underlying AIN diagnosis, and future investigation is relevant to clarify what this means for the clinical outcome.