Inherited Interleukin 2–Inducible T-Cell (ITK) Kinase Deficiency in Siblings With Epidermodysplasia Verruciformis and Hodgkin Lymphoma

Abstract

Biallelic mutations in the ITK gene cause a T-cell primary immunodeficiency with Epstein-Barr virus (EBV)–lymphoproliferative disorders. We describe a novel association of a homozygous ITK mutation with β–human papillomavirus (HPV)–positive epidermodysplasia verruciformis. Thus, loss of function in ITK can result in broad dysregulation of T-cell responses to oncogenic viruses, including β-HPV and EBV.

Cutaneous human papillomavirus (HPV) infection typically manifests with isolated papillomas. However, some patients in familial clustering develop extensive and protracted HPV infections, primarily the β-HPV types 5 and 8, with distinct cutaneous findings [1–3]; β-HPVs do not cause lesions in the general population. This clinical entity, epidermodysplasia verruciformis (EV), with autosomal recessive inheritance, is characterized by numerous cutaneous flat warts in childhood, which progress into verrucous tumors and evolve into squamous cell carcinomas later in life. The “typical” form of EV, not vulnerable to other infections, is caused by mutations in CIB1, TMC6, or TMC8, which impair keratinocyte-intrinsic immunity to β-HPV infection [3–5]. Mutations in other genes have been associated with “atypical” EV (Supplementary Table 1) [1–3]. Defects in those genes impair T-cell development or function, predisposing to β-HPV and other infections [1, 2]. We now describe an extended family with atypical EV in association with a mutation in an additional T-cell gene, ITK, encoding interleukin 2–inducible T-cell kinase (ITK).

METHODS

Case Report

The proband, a 44-year-old Iranian woman born to first-cousin parents, presented with multiple flat warts scattered on her whole body, prominently on the hands and legs, noted since age 17 (Figure 1A and Supplementary Figure 1). Histopathology of a skin lesion was consistent with EV (Figure 1B), and was positive for HPV types 5 and 8 (Figure 1C). Viral typing also demonstrated the presence of Epstein-Barr virus (EBV) in plasma with a load of 10 000 IU/mL. The proband’s younger sibling, a 40-year-old man (III-5 in Figure 1E), was diagnosed with Hodgkin lymphoma by computed tomographic scan showing a large peritoneal mass, and a needle biopsy confirmed the diagnosis (Supplementary Figure 1D and 1E). Clinical examination of skin, HPV typing, and histopathology of his skin lesions, similar to those in the proband, confirmed the diagnosis of EV. He also had high EBV load in plasma (Figure 1C). Both patients, therefore, suffered from atypical EV, associated with uncontrolled β-HPV and EBV infection. Short-term treatment with oral acitretin (10 mg/day) or local cryosurgery for skin lesions was largely ineffective. For Hodgkin disease, the patient received chemotherapy with ABVD (Adriamycin, Bleomycin, Vinblastine, Dacarbazine) protocol, which resulted in dramatic resolution of the intra-abdominal lymphadenopathies.

Figure 1.

Strategy for identification of the homozygous ITK mutation in a family with epidermodysplasia verruciformis (EV). A, Clinical presentation of EV manifesting with flat warts (arrows) on the hands of patient III-5. B, Histopathology of a flat wart in acral skin from patient III-3. Note epidermal thickening with hyperkeratosis, acanthosis, and hypergranulosis (hematoxylin and eosin staining, ×10 magnification); note koilocytes with large cytoplasmic halo (inset, ×40). C, Viral typing by real time quantitative polymerase chain reaction from a skin biopsy revealed the presence of human papillomavirus types 5 and 8, while Epstein-Barr virus was present in plasma (viral load, 10 000 IU/mL) of both patients III-5 and III-3 (data from III-3 is shown). D, Fluorescence-activated cell sorting of CD3⁺ T cells in patients III-3 and III-5 (upper and lower right panels, respectively) for T-cell receptor (TCR)–Vα24⁺ and TCR-Vβ11⁺. Reduced percentage of cells with these markers in patients with the ITK mutation (right panels: <0.02%), in comparison to 2 healthy controls (left panels: 0.36% and 0.42%), indicated lack of natural killer T cells. E, Pedigree of the family. Segregation analysis of the mutation is shown, with + sign representing the wild-type allele. F, Genome-wide homozygosity mapping was performed in patients III-3 and III-5, as well as in the healthy brother III-4 and the mother II-2. Regions of homozygosity (ROH) are represented by the blue vertical lines. The only ROH shared between the patients and not present in the unaffected individual was of 10.2 Mb on chromosome 5, containing the ITK locus. Locations of genes associated with EV are shown with black dotted lines. G, Whole exome sequencing and filtering of variants in patients III-3 and III-5, and the mother, II-2. Filtering criteria used to arrive at the pathogenic mutation in ITK are shown. H, Sanger sequencing confirmed that the c.362_364del mutation in ITK is homozygous in patient III-3 (lower panel); control DNA showed wild-type alleles only (upper panel). I, Structural analysis of the wild-type (I-1) and mutant (I-2) ITK protein motif in 3D structure. Interactions of Zn²⁺ ion with amino acids in the Cys₃His₁ motif are represented by orange dashed lines. Oxygen, hydrogen, nitrogen, sulphur, and carbon atoms are in red, white, blue, yellow, and cyan, respectively (stick-and-ball structures in I-1 and I-2). Backbone root-mean-square deviation (I-3), pairwise distance between Cys132–Zn²⁺, Cys133–Zn²⁺, Cys143–Zn²⁺, and His121–Zn²⁺ (I-4). Wilcoxon signed-rank test significance: *P < 2.2x10(-16) . Interaction energies of Zn²⁺ –Btk motif during the molecular dynamics trajectories in wild-type (blue line) and mutant (red line) structures (I-5). Abbreviations: EBV, Epstein-Barr virus; FITC, fluorescein isothiocyanate; HPV, human papillomavirus; ITK, interleukin 2–inducible T-cell kinase; MAF, minor allele frequency; MD, molecular dynamics; PE, phycoerythrin; RMSD, root-mean-square deviation; ROH, region of homozygosity; TCR, T-cell receptor.

Whole Exome Sequencing

DNA was isolated from peripheral blood of the patients and family members, and DNA from 2 patients (III-3 and III-5) and the patients’ mother (II-2) was submitted to whole exome sequencing (WES). Exons and flanking intronic regions of approximately 20 000 genes (45 Mb; 98% of consensus coding sequence exons) were captured by the Illumina TruSeq Exome Enrichment Kit and sequenced on a NextSeq 500 instrument (Illumina, San Diego, California). The data were analyzed as detailed in Figure 1G and the Supplementary Materials.

Homozygosity Mapping

Whole genome–wide single nucleotide polymorphism (SNP)–based homozygosity mapping was performed with DNA from 2 patients (III-3 and III-5) and their unaffected brother (III-4) and mother (II-2) by Infinium Global Screening Array-24 version 1.0 kit (~640 000 SNP markers; Illumina). The SNP genotyping data were analyzed by using PLINK (http://pngu.mgh.harvard.edu/) [6]. The regions of homozygosity (ROH) of ≥4 Mb were identified by an algorithm in PLINK.

Viral Typing

Skin and whole blood were collected from patients III-3 and III-5. Viral DNA was isolated with the High Pure viral nucleic acid extraction kit (Roche Diagnostic, Risch-Rotkreuz, Switzerland), followed by real time quantitative polymerase chain reaction by SYBR Green (Molecular Probes, Eugene, Oregon) for HPV types 5, 6, 8, 20, 21, and 38, as well as EBV, at 40 cycles of 30 seconds at 95°C, 60°C, and 72°C.

Fluorescence-activated Cell Sorting for Peripheral Blood Mononuclear Cell Immunophenotyping

Peripheral blood mononuclear cells (PBMCs) from whole blood from patients III-3 and III-5, as well as healthy controls, were isolated by density gradient centrifugation on Ficoll-Paque. CD3 MicroBeads were used to isolate T cells (Miltenyi Biotech, Bergisch-Gladbach, Germany). After gating on 1 × 105 CD3⁺ T-cell lymphocytes, the percentage of natural killer (NK) T cells (CD3⁺TCR-Vα24⁺ TCR-Vβ11⁺) was determined using the BD FACSCalibur (BD Biosciences, Singapore). Staining for fluorescence-activating cell sorting (FACS) was performed with fluorochrome-conjugated monoclonal antibodies (clones): CD3 (SK7; BD), TCR-Vα24 (C15), and TCR-Vβ11 (C21; Coulter-Immunotech, Miami, Florida).

T-Cell Proliferation Assay

Isolated PBMCs, 2 × 10⁵ per well from the patients and unrelated healthy controls, were cultivated in Roswell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum. Cultures were stimulated by 5 μg/mL phytohemagglutinin (PHA). After 72 hours of incubation, the treated and untreated cells were imaged under light microscope, and the cultures were examined by (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay during the last 24 hours of incubation.

Homology Modeling and Molecular Dynamics Simulations

Structural analysis of ITK in mutant and wild-type conformation was accomplished based on NMR structure of BTK motif (PDBID:2E6I). Mutant structure was modeled using the I-TASSER web server [7], and explicit molecular dynamics (MD) simulations were performed using the NAMD.2.12 package [8]. For details, see the Supplementary Materials.

RESULTS

Blood samples were collected from 11 family members, and DNA from patients III-3 and III-5 and their unaffected mother (II-2) was submitted for WES. WES data analysis, including multistep filtering of the variants, led to identification of homozygous variants in 3 candidate genes, ITK, BCAS1, and ZNF717 (Figure 1G); of these genes, only ITK is known to be related to immunity [9, 10], and mutations in ITK have previously been associated with EBV infection and lymphoproliferative syndrome 1 (Online Mendelian Inheritance in Man number 613011) [11].

Homozygosity mapping was performed in patients (III-3 and III-5) as well as healthy relatives (III-4 and II-2) (Figure 1E). Only a single 10.2-Mb ROH was shared by both patients, but not present in healthy family members (Figure 1F). This ROH on chromosome 5 harbors the candidate gene ITK, and not BCAS1 or ZNF717, providing additional evidence for causality of ITK mutations. The mutation in ITK (NM_005546) identified by WES and confirmed by Sanger sequencing, was a 3-bp deletion, c.362_364del (p.His121del) on exon 4 residing in a BTK-type zinc finger domain (aa119–148) containing the conserved histidine 121 (Figure 1H). This mutation was predicted to be disease causing in MutationTaster database, and it is not found either in ExAC, 1000 Genomes or gnom AD. Subsequent segregation analysis of 11 family members confirmed the homozygous ITK mutation as the associated variant (Figure 1E).

NK T cells are deficient in the peripheral blood of patients with ITK mutations [12, 13]. Immunophenotyping for NK T cells performed in our patients revealed a low number of NK T cells (Figure 1D). In contrast, FACS analysis revealed that the fraction of T cells expressing CD3 was normal (66.6 ± 1.08%) (Supplementary Figure 2). Finally, the T-cell proliferation assay showed that PHA-treated cells were capable of expanding and forming colonies, and there were no differences between the patients’ and control cells. These observations are consistent with previous reports (Chung et al [13] and Çağdaş et al [14]) (Supplementary Figure 3).

To examine the consequences of His121 deletion on structural dynamics of the BTK motif, MD simulation was performed after induction of the mutation using homology modeling approach. Zn²⁺ ion forms coordination complexes with 4 residues (ie, Cys132, Cys133, Cys143, and His121) in a tetrahedral geometry in the wild-type structure (Figure 1I-1). Lack of His121 resulted in elimination of 1 of the 4 Zn²⁺ interactions (Figure 1I-2). To assess structural stability of wild-type and mutant conformations, both proteins were monitored during the 15 ns MD trajectories (Supplementary Table 2). Deletion of His121 led to instability of the BTK motif as indicated by large fluctuations of backbone root-mean-square deviation in mutant conformation (Figure 1I-3). Since His121 is located at the conserved motif of zinc-finger domain (Cys₃His₁) assembly around a Zn²⁺ ion [15], we measured the distances between the center geometry of these residues and Zn²⁺ during the MD trajectories in both conformations. Figure 1I-4 indicates the pairwise distance between Cys132–Zn²⁺, Cys133–Zn²⁺, Cys143–Zn²⁺, and His121–Zn²⁺. Deletion of His121 in mutant ITK significantly increased the average distance between cysteine residues and Zn²⁺. The interaction energies between the Zn²⁺ ion and protein at the initial stage of MD simulations were lower than in the wild-type form (–144 vs –321 kcal/mol, respectively) (Figure 1I-5), demonstrating the essential role of His121 in stability of the structure. By continuing the simulation, Zn²⁺ was released from the mutant protein, while it remained bound to wild-type (–0.2 ± 2.7 kcal/mol vs –254 ± 7.0 kcal/mol, respectively) (Supplementary Video).

DISCUSSION

Collectively, our data suggest that the homozygous ITK mutation is responsible for the EV and EBV phenotypes in this family. The ITK gene product ITK is a nonreceptor tyrosine kinase involved in T-cell receptor signaling [9, 10]. In mice, ITK plays a role in NK T and T-cell (αβ and γδ) development, activation, and function [9, 16]. Mutations in ITK have previously been reported in association with EBV lymphoproliferative disorders (LPDs) [12, 14, 17, 18]. Our demonstration of previously unpublished ITK mutations in EV is novel, as susceptibility to β-HPV infection has not been reported in association with this gene. Additionally, the presence of EV in patient III-3, who did not have EBV-LPD, suggests that predisposition to cutaneous HPV infection is primarily related to loss of ITK, and not secondary to immune dysfunction from EBV-LPDs. Finally, association between β-HPV and EBV diseases has been reported in patients with other T-cell aberrations [19]. Consequently, it would be prudent to genotype all EV patients with documented β-HPV infection for mutations, potentially prognosticating the development of LPDs and assisting in planning for appropriate anticancer treatments.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank the family for cooperation. Carol Kelly assisted in manuscript preparation.

Financial support. This study was supported by the institutional funds of the Department of Dermatology and Cutaneous Biology of Thomas Jefferson University. The laboratory of Human Genetics of Infectious Diseases was supported by the National Institute of Health (R21 AI107508-02), The Rockefeller University, Institut National de la Santé et de la Recherche Médicale, Paris Descartes University, and the St. Giles Foundation.

Potential conflicts of interest. J.-L. C. has received personal fees from Genentech, ADMA Biologics, Nimbus, Vitae Pharmaceuticals, Kymera Therapeutics, Sanofi, Asahi Kasei, and Pfizer. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.