Follicular lymphoma-protective HLA class II variants correlate with increased HLA-DQB1 protein expression

Targeted RNA-Based Oxford Nanopore Sequencing for Typing 12 Classical HLA Genes

Identification of human leukocyte antigen (HLA) alleles from next-generation sequencing (NGS) data is challenging because of the high polymorphism and mosaic nature of HLA genes. Owing to the complex nature of HLA genes and consequent challenges in allele assignment, Oxford Nanopore Technologies' (ONT) single-molecule sequencing technology has been of great interest due to its fitness for sequencing long reads. In addition to the read length, ONT's advantages are its portability and possibility for a rapid real-time sequencing, which enables a simultaneous data analysis. Here, we describe a targeted RNA-based method for HLA typing using ONT sequencing and SeqNext-HLA SeqPilot software (JSI Medical Systems GmbH). Twelve classical HLA genes were enriched from cDNA of 50 individuals, barcoded, pooled, and sequenced in 10 MinION R9.4 SpotON flow cell runs producing over 30,000 reads per sample. Using barcoded 2D reads, SeqPilot assigned HLA alleles to two-field typing resolution or higher with the average read depth of 1750x. Sequence analysis resulted in 99-100% accuracy at low-resolution level (one-field) and in 74-100% accuracy at high-resolution level (two-field) with the expected alleles. There are still some limitations with ONT RNA sequencing, such as noisy reads, homopolymer errors, and the lack of robust algorithms, which interfere with confident allele assignment. These issues need to be inspected carefully in the future to improve the allele call rates. Nevertheless, here we show that sequencing of multiplexed cDNA amplicon libraries on ONT MinION can produce accurate high-resolution typing results of 12 classical HLA loci. For HLA research, ONT RNA sequencing is a promising method due to its capability to sequence full-length HLA transcripts. In addition to HLA genotyping, the technique could also be applied for simultaneous expression analysis.

HLA RNA Sequencing With Unique Molecular Identifiers Reveals High Allele-Specific Variability in mRNA Expression

The HLA gene complex is the most important single genetic factor in susceptibility to most diseases with autoimmune or autoinflammatory origin and in transplantation matching. Most studies have focused on the vast allelic variation in these genes; only a few studies have explored differences in the expression levels of HLA alleles. In this study, we quantified mRNA expression levels of HLA class I and II genes from peripheral blood samples of 50 healthy individuals. The gene- and allele-specific mRNA expression was assessed using unique molecular identifiers, which enabled PCR bias removal and calculation of the number of original mRNA transcripts. We identified differences in mRNA expression between different HLA genes and alleles. Our results suggest that HLA alleles are differentially expressed and these differences in expression levels are quantifiable using RNA sequencing technology. Our method provides novel insights into HLA research, and it can be applied to quantify expression differences of HLA alleles in various tissues and to evaluate the role of this type of variation in transplantation matching and susceptibility to autoimmune diseases.

Differences in Human Leukocyte Antigen Expression Between Breast Implant-Associated Anaplastic Large Cell Lymphoma Patients and the General Population

BACKGROUND

Breast implant-associated anaplastic large cell lymphoma (BIA-ALCL) is an uncommon T-cell lymphoma associated with textured-surface breast implants. Human leukocyte antigen (HLA) polymorphisms have been described with other forms of lymphoma, but have not been described for BIA-ALCL.

OBJECTIVES

The aim of this study was to evaluate HLA polymorphisms in BIA-ALCL patients.

METHODS

We prospectively evaluated HLA alleles in patients with BIA-ALCL. HLA was analyzed by probe-based sequence-specific testing and sequence-based typing. The frequencies of HLA-A, HLA-B, HLA-C, HLA-DRB1, and HLA-DQB1 alleles were evaluated. Allele frequencies in the Caucasian European general population were obtained from the National Marrow Donor Program to serve as normative controls. We estimated the relative risk of BIA-ALCL with 95% confidence intervals from a t test.

RESULTS

Thirteen patients who had undergone BIA-ALCL and HLA testing were identified from 2017 to 2018. Patients carried 10, 11, and 9 HLA-A, HLA-B, and HLA-C alleles, respectively. There were 8 DRB1 alleles and 5 DQB1 alleles in the BIA-ALCL patients. The A*26 allele occurred significantly more frequently in the general population compared with BIA-ALCL patients (0.2992 vs 0.07692, P < 0.001).

CONCLUSIONS

Our results identify a difference between HLA A*26 in patients who develop BIA-ALCL and the general population, and may signify genetic susceptibility factors responsible for germline genetic variation in HLA in patients with BIA-ALCL. Further work is needed to elucidate if these alleles are predictive for BIA-ALCL in women with textured-surface breast implants.Level of Evidence: 4.

SNP variants associated with non-Hodgkin lymphoma (NHL) correlate with human leukocyte antigen (HLA) class II expression

Large consortia efforts and genome-wide association studies (GWASs) have linked a number of genetic variants within the 6p21 chromosomal region to non-Hodgkin lymphoma (NHL). Complementing these efforts, we genotyped previously reported SNPs in the human leukocyte antigen (HLA) class I (rs6457327) and class II (rs9271100, rs2647012 and rs10484561) regions in a total of 1,145 subjects (567 NHL cases and 578 healthy controls) from two major ethnic groups in Malaysia, the Malays and the Chinese. We identified a NHL-associated (P_{NHL_add} = 0.0008; OR_{NHL_add} = 0.54; 95% CI = 0.37–0.77) and B-cell associated (P_{Bcell_add} = 0.0007; OR_{Bcell_add} = 0.51; 95% CI = 0.35–0.76) SNP rs2647012 in the Malaysian Malays. In silico cis-eQTL analysis of rs2647012 suggests potential regulatory function of nearby HLA class II molecules. Minor allele rs2647012-T is linked to higher expression of HLA-DQB1, rendering a protective effect to NHL risk. Our findings suggest that the HLA class II region plays an important role in NHL etiology.

Functional intronic ERCC1 polymorphism from regulomeDB can predict survival in lung cancer after surgery

We searched for potential regulatory single nucleotide polymorphisms (SNPs) in excision repair cross-complementing group 1 (ERCC1) using RegulomeDB, a database integrating information from the Encyclopedia of DNA Elements (ENCODE) project, and investigated their association with survival after surgery in non-small cell lung cancer (NSCLC). Among 364 SNPs found within ERCC1 region using RegulomeDB, four top priority SNPs (rs2298881C>A, rs1049739A>G, rs10415949A>G and rs6509214G>T) were selected for this study. The four SNPs were investigated in 316 patients. A replication study was performed (n = 579). Of the four SNPs analyzed in the discovery set, rs2298881C>A and rs6509214G>T were significantly associated with survival outcomes. The association was consistently observed only for rs2298881C>A in the validation cohort. In combined analysis, rs2298881C>A was significantly associated with worse overall survival and disease-free survival (P = 0.0002 and 0.02, respectively). A decreased reporter gene expression for rs2298881 A allele was observed compared with C allele by luciferase assay (P = 0.02). ERCC1 rs2298881C>A, an intronic SNP, is the first genetic polymorphism with functional evidence of regulating its expression, and the SNP is associated with prognosis of NSCLC. Our result supports the role of RegulomeDB as a comprehensive source of prioritized candidate SNPs for genetic association studies.

Putative functional variants of XRCC1 identified by RegulomeDB were not associated with lung cancer risk in a Korean population

The Encyclopedia of DNA elements (ENCODE) project revealed that nearby or distantly located non-coding DNA regulates the expression of coding genes. RegulomeDB (http://regulome.stanford.edu) is a new database that can be used to predict whether a variant affects transcription factor binding and gene expression. We investigated the association between lung cancer risk and potentially functional polymorphisms of XRCC1 that were selected using RegulomeDB in a Korean population. A total of 185 polymorphisms of XRCC1 were evaluated using RegulomeDB. Strong evidence suggested that 10 polymorphisms, from among the 185, affected XRCC1 expression with scores of 1a-1f that were based on the RegulomeDB scoring system. The rs2854510 polymorphism was rare in Asians (minor allele frequency < 0.05). Eight polymorphisms were in strong linkage disequilibrium (LD). The rs2854509 polymorphism, which was one of the 8 polymorphisms in LD, and rs7248167, which was not in the LD block, were genotyped in 610 lung cancer patients and 607 age- and sex-matched controls. Additionally, four polymorphisms of XRCC1 (rs25487, rs25489, rs1799782, and rs3213245), which were investigated with regard to their association with lung cancer risk in previous studies, were also genotyped. Two polymorphisms (rs2854509 and rs7248167) that were predicted to affect XRCC1 expression based on their RegulomeDB scores were not associated with lung cancer risk (P = 0.31 and 0.93, respectively). When stratified according to age, gender, smoking status, and tumor histology, the two polymorphisms of XRCC1 were not associated with lung cancer risk. Among the four polymorphisms that were previously studied, only rs25489 of XRCC1 was significantly associated with lung cancer risk (dominant model, adjusted odds ratio = 0.61, 95% confidence interval = 0.46-0.83, P = 0.002). Although RegulomeDB is an attractive tool for predicting the regulatory potential of variants, the two polymorphisms that were selected using RegulomeDB were not associated with lung cancer risk.

Genome-wide association study identifies five susceptibility loci for follicular lymphoma outside the HLA region

Genome-wide association studies (GWASs) of follicular lymphoma (FL) have previously identified human leukocyte antigen (HLA) gene variants. To identify additional FL susceptibility loci, we conducted a large-scale two-stage GWAS in 4,523 case subjects and 13,344 control subjects of European ancestry. Five non-HLA loci were associated with FL risk: 11q23.3 (rs4938573, p = 5.79 × 10(-20)) near CXCR5; 11q24.3 (rs4937362, p = 6.76 × 10(-11)) near ETS1; 3q28 (rs6444305, p = 1.10 × 10(-10)) in LPP; 18q21.33 (rs17749561, p = 8.28 × 10(-10)) near BCL2; and 8q24.21 (rs13254990, p = 1.06 × 10(-8)) near PVT1. In an analysis of the HLA region, we identified four linked HLA-DRβ1 multiallelic amino acids at positions 11, 13, 28, and 30 that were associated with FL risk (pomnibus = 4.20 × 10(-67) to 2.67 × 10(-70)). Additional independent signals included rs17203612 in HLA class II (odds ratio [OR(per-allele)] = 1.44; p = 4.59 × 10(-16)) and rs3130437 in HLA class I (OR(per-allele) = 1.23; p = 8.23 × 10(-9)). Our findings further expand the number of loci associated with FL and provide evidence that multiple common variants outside the HLA region make a significant contribution to FL risk.