Introduction

There are over 400 genes implicated in inborn errors of immunity (IEI) [1]. Identifying their variants is critical for patient care, genetic prevention, and translational research [2]. For this purpose, diagnostic exome sequencing has proven helpful in preventing primary immunodeficiency (PID) in tribal populations where founder mutations and autosomal recessive (AR) disorders are exceptionally common [3].

A national premarital screening program was implemented in the United Arab Emirates (UAE) in 2005 (Personal Status Act No 28, Article 27). This endeavor has eliminated sickle cell disease and beta-thalassemia major from the community. This study examines indigenous (novel) pathogenic variants of IEI, and suggests including them in the existing premarital screening program.

Methods

This retrospective data collection (Chart Review) study was approved through the ‘Tawam Human Research Ethics Committee’ (T-HREC); reference number: MF2058-2020-722 (AA/AJ/722; May 13, 2020), and informed consent to participate was waived. The study involved one center (Tawam Hospital, Al Ain City, Abu Dhabi, UAE). A review of 54 variants of IEI (detected by diagnostic exome sequencing between March 2016 and November 2019) was conducted between May 13, 2020 and July 10, 2020. Of these, 21 variants were novel and included in this report. Thirteen variants involved patients and eight were found on a screening test.

Variant information in public databases was consolidated using Ensembl Variant Effect Predictor (https://www.ensembl.org/vep). Gathered data included effect of variation, codon change, amino acid change, and variations from dbSNP, functional consequence, and pathogenicity score from algorithms, such as SIFT, PolyPhen, Condel, CADD, FATHMM, LRT, MetaLR, MetaSVM, MutationAssessor, MutationTaster, PROVEAN, REVEL, and VEST3 in dbNSFP (One-Stop Database of Functional Predictions and Annotations for Human Non-synonymous and Splice Site). The Mutation Significance Cutoff (MSC) Server was used to obtain gene-wise CADD MSC [4].

Multiple sequence alignment was used to evaluate conservation of amino acids at studied sites and compute the Jensen-Shannon divergence (JSD) values. Amino acid sequences of proteins from Homo sapiens (human), Pan troglodytes (chimpanzee), Mus musculus (house mouse), Rattus norvegicus (Norway rat), Canis lupus familiaris (dog), Equus caballus (horse), Bos taurus (bovine), Xenopus tropicalis (frog), Gallus gallus (chicken), and Danio rerio (zebrafish) were downloaded from NCBI RefSeq (S1 Table). Sequences were imported into Geneious 9.1.8 (https://www.geneious.com) using NCBI Accessions and multiple sequence alignment was performed using MUSCLE. Alignments were exported in FASTA format and used to obtain JSD values from “https://compbio.cs.princeton.edu/conservation/score.html” using the default options [5].

Homology model of the ATM (Ataxia telangiectasia mutated) protein with the variation (ATM:p.Val2823Gly) was generated using the SWISS-MODEL server (https://swissmodel.expasy.org/). The Protein Data Bank (PDB) structure 6K9L (4.27 Angstrom resolution cryo-EM structure of human dimeric ATM kinase) was used as the template. Protein structures were visualized using Visual Molecular Dynamics 1.9.3 [6].

Results

Table 1 lists the 21 novel variants of genes (19 single genes) related to IEI, sorted according to the International Union of Immunological Societies Phenotypical Classification [1]. Thirteen variants were found in symptomatic patients, while eight variants were found on a screening test in asymptomatic subjects. Fifteen variants are autosomal recessive (AR, 71%), three undetermined inheritance, one autosomal dominant (AD), and two X-linked. Nine variants are frameshift, three nonsense, four intronic (one pathogenic), and five missense (two pathogenic). Table 2 shows sequence alignment of the proteins at positions of the missense variants from ten species.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Novel variants of IEI.

https://doi.org/10.1371/journal.pone.0245888.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Sequence alignment of protein sequences at positions of the missense variants with the Jensen-Shannon divergence (JSD) values.

https://doi.org/10.1371/journal.pone.0245888.t002

One variant involves ataxia-telangiectasia (AT)

ATM:p.Val2823Gly has conflicting predictions of pathogenicity. Val2823 is conserved, with a permissible change to isoleucine (JSD: 0.7041, Table 2). Val2823 (Fig 1) is located in the kinase domain of the alpha helical C terminus. Gly2823 is predicted to disturb the stability of the helical conformation, thus, affecting the structure and function of the protein [9,10]. This assessment is consistent with the SIFT score of 0.00 (damaging) and the CADD score of 28.9 (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Structural model of wild-type and variant ATM protein.

The ATM protein structure is shown in white cartoon representation and the kinase domain is shown in yellow. Amino acids are shown in stick representation. The black boxed region is enlarged in the subsequent two images. (A) Structure of the conserved kinase domain of ATM, (B) wild type Val2823, and (C) variant Gly2823.

https://doi.org/10.1371/journal.pone.0245888.g001

Discussion

This study describes 21 novel variants of IEI. Their types include frameshift, nonsense, intronic, and missense. Our bioinformatics analysis shows the majority (15 of 21, or 71%) are pathogenic or likely pathogenic; of these, 13 are autosomal recessive. Thus, these mutations are amenable for genetic prevention via screening and counseling. Future studies, however, are needed to confirm their prevalence and natural history in individual patients.

The variants reported here are novel. Their prediction of pathogenicity is assessed by in silico computer tools described above. Thus, functional validation of these variants remains fundamental to attribute pathogenicity with certainty [13,14]. As previously shown for cancer predisposition, improved genomic analysis lowers the prevalence of variant of uncertain significance (VUS) [15]. The authors concluded that variant “interpretation by multiple experts in the context of personal and family histories maximizes actionable results and minimizes reports of VUS” [15]. Similar efforts are needed for novel variants of IEI.

Pathogenic variants of autosomal recessive disorders carry ‘fetal risk’ for homozygosity and compound heterozygosity, especially in tribal cultures with frequent consanguineous marriages. Fetal risk is, of course, higher in the presence of multiple pathogenic variants of a gene, especially within a tribe, such as the three variants of LRBA (Table 1). Another example is the six known pathogenic variants of ataxia telangiectasia (AT) in our community; one of these is novel and reported here. This entity is characterized by genomic instability due to defects in DNA repair [16]. Homozygous and double heterozygous affected individuals have immune derangements (e.g., thymic involution) and central nervous system degeneration. Heterozygous individuals may show mild clinical features, such as increased risk for cancer and hypersensitivity to radiation [17,18]. Thus, the AT variants carry an exceptional fetal risk for homozygosity and compound heterozygosity, especially in our tribal cultures that mainly practice consanguineous marriages. Such variants need to be built into a national screening program aiming at disease prevention [19].

Generation of homology models is often hampered by the unavailability of wild-type human or homologous protein structures. Hence, in this instance, the only variant that could be modelled was the ATM:p.Val2823Gly based on the 4.27 Å cryo-electron microscopy structure of ATM (PDB ID: 6K9L). Val283 is part of the helix designated as kα4c in the phosphatidylinositol 3-/4-kinase catalytic domain of the ATM serine-protein kinase [20]. This valine, or the physiochemically similar isoleucine, is conserved in all species considered here. While the precise role of the helix is not known, a change to a glycine, a notably achiral amino acid, at this position could potentially disrupt the stability of this helix.

Twenty-one IEI novel variants are categorized based on the IUIS Phenotypical Classification [1]. This allows simpler dissection of IEI molecular diagnosis in association with clinical and immunological profiles. It serves as a framework to determine IEI types and prevalence in the UAE. It will also enhance our immunological understanding of the variant to aid patient care with targeted therapy and prevention with family screening and counselling.

The novel IEI variants reported here include one variant of severe combined immunodeficiency (SCID); two variants of combined immunodeficiency (CID) with low CID and DNA repair defect; four variants of CID with syndromic features; one variant of antibody deficiency; two variants of defects in phagocytes; three variants of one gene related to immunodysregulation; two variants of defects in intrinsic and innate immunity; one variant of hemophagocytic lymphohistocytosis (HLH) and Epstein-Barr virus (EBV) susceptibility; three variants related to bone marrow failure; and two variants under ‘others’. PIGA and THSD1 variants under ‘Others’ are not reported in IUIS. However, the PIGA variant causes paroxysmal nocturnal hemoglobinuria (PNH) and is associated with cellular immunodeficiency. It is hypothesized that PNH impacts lymphocyte count through decrease in lymphopoiesis and complement-mediated damage [21]. A case series in UAE shows the THSD1 variant causes intracranial berry aneurysm 12, nonimmune hydrops fetalis, congenital cardiac defects, hemangiomas and hypogammaglobulinemia [22].

One of the major limitations of diagnostic sequencing and genetic testing is that the effect of a variation on function cannot be inferred from sequence alone. Considering the volume of next generation data produced, a large proportion of variants currently remain classified as variants of uncertain significance (VUS) in bioinformatics databases. While phenotype data could address this problem partly, systematic in vitro functional and mutational assays are necessary to pinpoint the precise role of a variant on the structure and function of a protein and its pathways [14]. Often, a single assay may not be sufficient for this purpose elevating the difficulty in establishing the pathogenicity of VUS. Nonetheless, in this study, based on computational predictors, we have attempted to predict the pathogenicity of novel variants identified in our population. However, as indicated, functional assays would be necessary to ascertain the precise pathogenicity of these variants.

IEI are heterogeneous disorders and are often underdiagnosed. Knowing the causative genetic variant offers a better prediction of its course and opens up opportunities for participating in clinical trials related to the defect. Genetic screening, on the other hand, offers early identification of people at risk; thereby, allowing more prompt interventions [19]. Thus, genetic testing is required for all individuals with immunodeficiency. Our results will aid in setting the framework of a national screening program aiming at genetic prevention of IEI.

Supporting information

Acknowledgments

The contributions of patients and Tawam Hospital staff are greatly appreciated.