Genomic dissection of the Fcγ receptor region in the context of monoclonal antibody therapy

Development of the anti-CD20 antibody, rituximab, heralded the start of monoclonal antibody (mAb) therapy as an effective means of treating cancer. Despite its undoubted impact, clinical responses remain variable and cures are rarely achieved. Evidence from pre-clinical models and human trials indicates that mAbs primarily act through engaging low-affinity Fc gamma receptor (FcγR)-expressing effector immune cells. The low-affinity FcγR genes, FCGR2A, FCGR2B, FCGR2C, FCGR3A and FCGR3B, are located within a highly homologous 200 kb region at 1q23, which is the result of an ancestral segmental duplication event (Figure 1). The locus also contains numerous single nucleotide polymorphisms (SNPs), many of which can affect receptor affinity and/or function and are associated with differential responses following mAb immunotherapy. Moreover, the region contains extensive copy number variation (CNV) that also can affect the expression and function of these receptors, but its impact on mAb immunotherapy remains unknown. To investigate the full impact of SNPs and CNV in the FcγR locus, we have optimised a number of sensitive and specific assays which are amenable to formalin fixed paraffin embedded (FFPE) material in order to apply them to clinical trial samples in a high-throughput manner. Initially we assessed the accuracy of established TaqMan and novel allele-specific (KASP) genotyping assays for FCGR2A-131H/R (rs1801274), FCGR3A-158F/V (rs396991) and FCGR2B-232I/T (rs1050501) SNPs by analysing 2085 DNA samples derived from peripheral blood lymphocytes (PBL) from a large, multi-centre cohort. Our data showed that although clear discrimination was possible at the FCGR2A-131H/R SNP, we needed additional selective Sanger sequencing to discriminate the FF/FV and IT/TT genotypes for the FCGR3A-158F/V and FCGR2B-232I/T SNPs, respectively. This difficulty in genotype discrimination in the cases of FCGR3A and FCGR2B is likely due to sequence homology with other genes and CNV in the gene regions which complicate assay design and interpretation of certain genotypes. Secondly, we applied a combined KASP genotyping and Sanger sequencing approach to matched PBL DNA and FFPE-extracted DNA from follicular lymphoma (FL) patients [n=14] and showed that while FFPE material was more likely to fail genotyping, successfully genotyped cases were concordant with the matched genomic DNA samples. FFPE samples which failed to amplify PCR products of at least 100 bp using the BIOMED-2 multiplex PCR protocol were more likely to fail genotyping assays. Finally, we assessed the ability of a multiplex ligation-dependent probe amplification (MLPA) assay to concurrently determine SNP genotype and CNV in the low-affinity FCGR locus in a cohort of 155 normal donors and DNA from seven matched PBL-/FFPE-derived FL cases. We employed a paralog ratio test (PRT) assay for FCGR3A and FCGR3B CNV confirmation. In our normal donors, MLPA and PRT results were concordant. 16% of normal donors harboured a deletion [n=15] or duplication [n=10] affecting the FCGR2C locus (A summary of regions of CNV is shown in Figure 1). CNV of FCGR3B was associated with variation at FCGR2C and no CNV was observed in FCGR2A and FCGR2B. CNV affecting FCGR3A was observed in 5% of donors with deletions and duplications in 4 and 5 donors, respectively. In the FFPE-derived DNA samples, we observed elevated variability in data quality that was most noticeable in probes targeting HSPA6, FCGR2C exon 4 and HSPA7. Poor quality data correlated with samples that failed to amplify at least the 100 bp PCR product using the BIOMED-2 multiplex PCR protocol. As such, the preclusion of HSPA6, FCGR2C exon 4 and HSPA7 probes and FFPE samples that failed to amplify any BIOMED-2 PCR product from the analysis permitted the production of high-quality MLPA data. Finally, we designed, and are currently optimising, a targeted re-sequencing platform (Haloplex, Agilent) to interrogate informative regions of the FcγR region, which includes those with unique sequence identify for CNV analysis, and those that include known SNPs. In conclusion, we have evaluated a suite of assays for the genomic analysis of the FcγR locus that are scalable for application in large clinical trials of antibody therapy. This work will ultimately provide a detailed architecture of the region and establish the importance of FcγR genetics in predicting response to antibody therapeutics.