Genetic testing of Behçet's disease using next-generation sequencing to identify monogenic mimics and HLA-B*51

Abstract

Objective: Several monogenic autoinflammatory disorders and primary immunodeficiencies can present early in life with features that may be mistaken for Behçet's disease (BD). We aimed to develop a genetic analysis workflow to identify rare monogenic BD-like diseases and establish the contribution of HLA haplotype in a cohort of patients from the UK.

Methods: Patients with clinically suspected BD were recruited from four BD specialist care centres in the UK. All participants underwent whole-exome sequencing (WES), and genetic analysis thereafter by (i) examining genes known to cause monogenic immunodeficiency, autoinflammation or vasculitis by virtual panel application; (ii) scrutiny of variants prioritized by Exomiser using Human Phenotype Ontology (HPO); (iii) identification of copy number variants using ExomeDepth; and (iv) HLA-typing using OptiType.

Results: Thirty-one patients were recruited: median age 15 (4-52), and median disease onset age 5 (0-20). Nine/31 (29%) patients had monogenic disease mimicking BD: five cases of Haploinsufficiency of A20 with novel TNFAIP3 variants (p.T76I, p. M112Tfs*8, p. S548Dfs*128, p. C657Vfs*14, p. E661Nfs*36); one case of ISG15 deficiency with a novel nonsense variant (ISG15: p.Q16X) and 1p36.33 microdeletion; one case of common variable immune deficiency (TNFRSF13B: p.A181E); and two cases of TNF receptor-associated periodic syndrome (TNFRSF1A: p.R92Q). Of the remaining 22 patients, eight (36%) were HLA-B*51 positive.

Conclusion: We describe a novel genetic workflow for BD, which can efficiently detect known and potentially novel monogenic forms of BD, whilst additionally providing HLA-typing. Our results highlight the importance of genetic testing before BD diagnosis, as this has an impact on choice of therapy, prognosis and genetic counselling.

Keywords: Behçet’s disease; Exomiser; HLA typing; autoinflammatory disease; whole-exome sequencing.

Figure 1.

Bioinformatic genetic analysis summary. FASTQ files generated from whole-exome sequencing were aligned to human reference genome hg38 using BWA-MEM [23]. Variants were called from BAM files using TNscope [24, 25] for somatic variants; DNAseq for germline variants [24]; and ExomeDepth [26] for copy number variants (CNV). Somatic and germline VCF files were annotated using wANNOVAR [27]. CSV files were then filtered using virtual gene panels through an R script. The germline VCF file was also analysed using Exomiser [28]. The resulting variants were combined into a shortlist and reviewed by an expert multidisciplinary team (MDT) to identify clinically actionable variants. FASTQ files were also processed in OptiType [29] to determine HLA genotypes

Figure 2.

Clinical pictures of case 31. (A/B) Diffuse maculopapular rash with irregular borders and subtle telangiectasia affecting the forearm and shoulder, (C) bilateral exudative tonsillitis during a febrile episode, and (D) multiple discrete perianal lesions with abscess and fistula formation, and erythema and excoriation of the surrounding skin

Figure 3.

Comparison of the scores for the International Criteria for Behçet’s Disease (ICBD), the International Study Group (ISG) and the Paediatric Behçet’s Disease (PEDBD) criteria. Percentage of cases grouped by outcome of the genetic analysis that met the (i) International Criteria for Behçet’s Disease (ICBD) (score ≥4) [20], (ii) International Study Group (ISG) [21] criteria and (iii) Paediatric Behçet’s Disease (PEDBD) criteria (score ≥3) [22]. For ICBD and ISG criteria there were nine cases with monogenic disease; eight cases with HLA-B*51 associated Behçet’s Disease (BD): and 14 cases with likely polygenic BD. For PEDBD criteria, only paediatric cases were included, hence the analysis included 23 cases of which seven had monogenic disease, five cases had HLA-B*51 associated BD and 11 cases had likely polygenic BD