A WAS promoter variant underlying Wiskott-Aldrich syndrome in two kindreds

The global expansion of genomic medicine has improved access to genetic testing for patients with suspected inborn errors of immunity (IEI). Despite these advances, diagnostic rates remain relatively low, with current estimates suggesting that <50% of patients with suspected IEI receive a confirmed genetic diagnosis. This limitation is partly due to the focus of most clinical laboratories on variants within protein-coding regions or those impairing splicing (1). Expanding variant analysis beyond these protein-coding regions and variants leading to aberrant splicing is promising to improve the diagnosis of IEI. Such variants can be pathogenic through different mechanisms, such as disruption of the 5′ untranslated region (UTR) and 3′ UTR, or by disrupting cis-regulatory elements (CREs), including promoters or enhancers (2, 3). While the availability of tools to screen for variants disrupting 5′ UTR and 3′ UTR remains limited, many scores or strategies were developed recently for variants located in CREs, including the use of tailored cutoffs for combined annotation dependent depletion (CADD) and regulatory Mendelian mutation (REMM) scores (2). A specific tool to screen for such variants, promoterAI, was also recently published.

In this context, we investigated four male patients from two unrelated kindreds with features of IEI (Fig. 1 A). P1 and P2, from kindred A, are of French descent and were born to non-consanguineous healthy parents. P1 is 4 years old and has a history of eczema, celiac disease, and recurrent hemorrhagic immune thrombocytopenic purpura (ITP) (Fig. 1, B and C). P2 is 7 years old and suffered at the age of 2 years from an acute episode of ITP (Fig. 1, B and C). He also has a history of atopic dermatitis. P3 and P4 from kindred B were born to non-consanguineous Eritrean parents and currently reside in Switzerland. Both patients presented with eczema. P3, now 24 years old, was referred in early childhood for lymphadenopathy. At the age of five, he was diagnosed with EBV-driven Hodgkin lymphoma (mixed-cellularity subtype, EBV positive, stage IIAE) involving the cervical, supraclavicular, and mediastinal regions. Almost 2 years after completing chemotherapy, he experienced a relapse of Hodgkin’s disease with axillary, mediastinal, intra-abdominal, and pulmonary involvement (stage IVa). At 9 years of age, he developed autoimmune hemolytic anemia and ITP, initially treated with steroids. P3 has since experienced multiple ITP relapses and received treatment with steroids, intravenous immunoglobulins, and rituximab. He is currently stable under immunosuppressive therapy with mycophenolate-mofetil, prednisone, and sirolimus. The brother of P3, P4, is 32 years old and suffered from diffuse large cell lymphoma, classified as Murphy stage III (equivalent to Ann Arbor stage III), and was treated with chemotherapy. Until now, he has been in remission but has persistent inguinal lymphoproliferation, where biopsies were taken without any signs of malignancy. The sister and parents of P3 and P4 are healthy.

Given the suspicion of IEI, both kindreds were explored through high-throughput sequencing (HTS) approaches capturing exons, e.g., a panel for P1, and whole-exome sequencing in both P3 and P4. Screening of small nucleotide variants or copy number variants within the coding sequences and essential splice site variants of captured genes failed to identify any pathogenic variants. We thus reanalyzed for variants covered in noncoding regions that are partially covered even by exon-captured based HTS (1). We did not detect any noncoding variants predicted to impair splicing by tools such as spliceAI. We further screened on noncoding variants with high CADD score, adapting the algorithm by Villani et al. (2). We ended up identifying in both kindreds a single nucleotide variant located upstream of the WAS gene (hg38 X-48683790-C-T; NM_0000377.3; c.-64C>T). This variant is very rare because it is absent from gnomAD v4.1. The CADD v1.7 score was high (=20.5). In addition, this variant was predicted as deleterious by REMM v0.4 (=0.999) and promoterAI v1.0 (=−0.746). The variant was indeed located on a conserved nucleotide as highlighted by elevated phastCons (=1) and phyloP (=3.90) scores. This variant is predicted to be located in a CRE by ENCODE, and the c.-64C position lies within the WAS promoter region shown to be essential for hematopoietic-specific expression, as it contains critical transcription factor–binding sites (e.g., FLI1, GATA family motifs, and GC boxes that bind Sp1/Sp3) (Fig. 1 D) (4). By Sanger sequencing, this variant was confirmed to be present at a hemizygous state in P1, P2, P3, and P4, at a heterozygous state in the mothers of both kindreds and in the sister of P3 and P4, and was absent in the fathers of both kindreds (Fig. 1 E). At immunological level, P1, P2, P3, and P4 displayed T CD8 lymphopenia, with increased CD4/CD8 ratio (data not shown). In T-activated blasts of the patients, WASP expression was 22% and 10% of normal at the mRNA and protein levels, respectively (Fig. 1, G and H). Overall, according to criteria from the American College of Medical Genetics, these data are further supportive of categorizing the c.-64C>T variant as pathogenic and causal for the clinical phenotype of the four patients.

We thus report a pathogenic variant located within the promoter of WAS. We found this rare variant in two unrelated kindreds and evidenced decreased but not absent protein level amongst the T cells of the patients. This likely explained the clinical phenotype, characterized by intermittent periods of normal platelet counts and relatively mild but fluctuating disease manifestations such as eczema, autoimmunity, and lymphoproliferation. Although this presentation may appear mild and would historically have been considered “intermittent X-linked thrombocytopenia,” it would now be better classified as a class I WAS variant (5), and the occurrence of malignancies in two patients highlights the potential severity of such phenotypes. Clinically, this emphasizes that WASP deficiency should also be considered in patients who do not present with the classic triad (eczema, thrombopenia, and immunodeficiency) and who may even have episodes of normal thrombocyte counts. Recognition of such atypical presentations is important for timely genetic counseling and screening of female carriers, as well as for the treatment of the patients, who may benefit from hematopoietic stem cell transplantation or gene therapy. Regarding WAS locus, variants located within the core promoter element c.-28 could also be candidate pathogenic variants. The mechanism of pathogenicity for variants in CRE is usually decreased transcription of otherwise normal gene products, resulting in a quantitative rather than qualitative defect (3). Furthermore, some variants in CRE may also lead to additional alternative splicing. Interestingly, variants in CRE that can also result in increased transcript products have been reported in other genetic diseases such as developmental disorders (3), but not, to the best of our knowledge, for IEI. A limitation of the study is that we did not address the possible cell-specific impact of the variant (3). Overall, the variant we report here further exemplifies that the implementation of bioinformatics tools for noncoding regions is warranted to systematically screen genetic data for variants in such regions in genes for IEI.