Disease Progression of WHIM Syndrome in an International Cohort of 66 Pediatric and Adult Patients

All observed CXCR4 variants were classified according to the 2015 standards and guidelines for the interpretation of sequence variants established by the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) [34]. Definitions of criteria for classifying pathogenic variants are published by Richards and colleagues [34]. Additional general guidance developed by the Clinical Genome Resource (ClinGen) for the PVS1[35] and PS3[36] criteria were also incorporated. Of note, the PVS1 criterion is applied only to loss of function variants; given that the disease mechanism of WHIM syndrome is gain of function, PVS1 was not applied to any CXCR4 variants. In lieu of PVS1, the PM4 criterion was applied to variants resulting in protein length changes. Although this criterion is typically applied to in-frame deletions/insertions, ClinGen recommendations (https://​clinicalgenome.​org/​docs/​variant-curation-standard-operating-procedure-version-2/​) suggest PM4 can also potentially be applied to variants that cause premature truncation, but not nonsense-mediated decay, such as those observed in the CXCR4 variants reported in association with WHIM syndrome that lead to production of C-terminally truncated receptors lacking residues critical for receptor internalization. Although the in silico predictor CADD [37, 38] suggests that the observed variants are likely to be damaging to protein function, the PP3 criterion was not applied to avoid double counting of evidence used to support the application of PM4. Additionally, the PP1 criterion used to support co-segregation of the variant within one or more families was applied if the genotype of the affected family member(s) was confirmed. The PS4 criterion (used at a moderate strength when a variant was observed in multiple unrelated patients with the same phenotype) was applied if the variant was observed in ≥ 3 unrelated individuals asserted to have WHIM syndrome. A specific phenotype sufficient for application of the PP4 criterion was defined as the presence of most features of WHIM, including myelokathexis.

The cDNA of human CXCR4 (GenBank accession no NM_003467) carrying the consensus Kozak sequence at the 5′ end (GCCGCCACCatg) and 36 nucleotides of the 3′UTR at the 3′ end was synthesized at Genewiz (Leipzig, Germany) and cloned in pcDNA3.1 (Hygro+) via NheI/EcoRV to give pcDNA3.1_CXCR4. CXCR4_WHIM syndrome variants were generated by site-directed mutagenesis, using the QuickChange II XL kit (Agilent Technologies), the primers (listed in table below) and pcDNA3.1_CXCR4 as template. S346Pfs*12 was synthesized at Genewiz and cloned in pcDNA3.1 (Hygro+) via NheI/EcoRV. Plasmids were generated using the Endofree Plasmid MaxiPrep Kit (Qiagen) and all plasmid preparations were sequenced (Microsynth, CMV forward, BGH reverse) (STable3).

K562 (CCL-243, ATCC) cells were cultured in IMDM (Gibco) containing 10% FCS (Sigma) and penicillin-streptomycin (Gibco). Starvation was performed in IMDM, 0.5% BSA (Sigma), penicillin-streptomycin overnight.

K562 cells (2.5 × 10⁶ cells in 250 μl/cuvette) were transiently transfected by electroporation with 25 μg pcDNA3.1_CXCR4 plasmids per cuvette in Ingenio Solution (Mirus), using a Gene pulser Xcell device (BioRad) and exponential decay protocol (250 V, 1000 μF) in 0.4 cm cuvettes followed by immediate transfer to 6-well plates containing 2.25 ml pre-equilibrated culture media. After 24 h, transfected cells (1 × 10⁵ cell/well) were seeded in 96-well plates and serum-starved overnight. Cells were then resuspended in warm incubation buffer (HBSS with Ca²⁺ and Mg²⁺ + 0.5% BSA + 20 mM Hepes pH 7.4) and stimulated with CXCL12 for 45 min or 4 h at 37 °C, 5% CO₂. After incubation, the cells were washed twice with cold incubation buffer and then stained with CXCR4 12G5-APC antibody (BD, 1:20 dilution in incubation buffer) for 20 min at 4 °C. After washing and resuspending in flow buffer (HBSS with Ca²⁺ and Mg²⁺+ 0.1% BSA + 20 mM Hepes pH 7.4), the samples were measured on Cytoflex (Beckman-Coulter) and analyzed in FCS Express software (De Novo Software). Cells were gated based on the FSC/SSC and isotype control and the mean fluorescence intensity (MFI) of CXCR4+ population was used in subsequent analysis.

Statistical analysis was performed in the GraphPad Prism software. p Values < 0.05 were considered statistically significant and set as follows: *p < 0.05; **p < 0.01; ***p < 0.001; ^nsp ≥ 0.05. Unpaired 2-tailed t test was used. The number of independent experiments (n) is stated in each figure legend.

The study cohort comprises 66 patients with WHIM syndrome, diagnosed according to the adapted ESID-PAGID guidelines: the majority met criteria for a definite diagnosis (64), and two cases met criteria for a probable WHIM diagnosis. Our cohort includes 57 unique unreported cases of WHIM syndrome, those reported previously were individual case reports (n = 9) referenced in Supplementary Table 1. Our cohort does not have major overlaps with any prior reports from academic centers with high expertise in Europe (Italy, France) or the National Institutes of Health in the USA. Table 1 summarizes the demographics of the subjects. Our cohort includes nearly twice as many children (n = 43, < 18 years old) as adults (n = 23, > 18 years old). The median age at diagnosis was 5.5 (range 2 weeks–51 years). There was a predominance of female patients (62% female, n = 41; 38% male, n = 25). The diagnosis of the majority of patients (79%, n = 52) was prompted by clinical signs (Sx) related to WHIM, rather than via positive family history (FH, 20%, n = 13) and/or newborn screening for SCID (NBS-SCID, 1.5%, n = 1) with low T cell receptor excision circles (TRECs). In 71% (n = 47) of patients, both genetic testing and diagnostic bone marrow biopsy were used to confirm WHIM diagnosis. In 26% (n = 17) of patients, diagnosis of WHIM syndrome was confirmed solely based on genetic testing, and the remaining 2 patients were diagnosed by bone marrow biopsy and clinical signs only. Only 22.7% (n = 15) of patients in our cohort presented with all four features of the WHIM acronym. An asterisk in parenthesis (*) indicates that a bone marrow biopsy was not performed but leukopenia was present. The most common phenotype combination was HIM with 24% (n = 16), followed by IM in 15.2% (n = 10), I(*) 7.6% (n = 5), WI(*) 7.6% (n = 5), HI(*) 6.1% (n = 4), WIM 4.5% (n = 3), HM 4.5% (n = 3), WHI(*) 4.5% (n = 3), M or (*) 1.5% (n = 1) each (Supplementary Table 1).

Of 66 patients, 64 cases were reported to have variants in the C-terminal domain of CXCR4. In one case, genetic information was lost, and in 2 patients, the diagnosis of WHIM syndrome was based on clinical presentation and bone marrow morphology, according to ESID-PAGID diagnostic criteria for WHIM syndrome. We identified 17 distinct variants in our cohort, resulting in 16 different amino acid changes (Fig. 1A). Four variants were nonsense-truncating variants and 12 were frameshift variants, all positioned (p.318 to p.346) in the intracellular region of the C terminus (p.303 to p.352), similar to prior reports [5] (Fig. 1B, C). Nonsense mutations were reported in 45 patients (68%) and frameshift mutations in 17 patients (26%). In 38 (64%) cases, mutations occurred de novo, while the other 21 cases were distributed among 15 pedigrees (SFigure1). Eleven of seventeen variants (64%) are previously unpublished in WHIM patients and represent 23% of our cohort (15 cases). As expected, p.R334* (47%, n = 31), and p.S338* (17%, n = 11) were the most common variants in our cohort. We identified two different genotypes (c.1013C>G and c.1013C>A) resulting in p.S338*, with the c.1013C>A (1/11, 9%) never reported in WHIM syndrome. Both variants create a premature stop codon, tCa>tGa, and tCa>tAa, respectively, thus truncating the C terminus. The 14 remaining amino acid changes were distributed among the remaining genotyped cases in only 1 to 4 cases, each (STable2). We observed a marked phenotypic heterogeneity in patients with the same genotype, even within the same family, with no obvious correlation between genotype and clinical manifestations (SFigure1). All 17 distinct variants in our study cohort, including the 11 novel variants, are predicted to be deleterious with a CADD PHRED value above 15. (range: 26.5 to 39) (STable2). All variants observed in the cohort were classified according to the ACMG/AMP 2015 standards and guidelines for the interpretation of sequence variants [34]. Of the 17 identified variants, 15 were classified as likely pathogenic or pathogenic (STable2). These classifications were primarily supported by the location of the variants in the critical C-terminus, absence in the gnomAD population database [39], availability of inheritance information gleaned from family studies (variant co-segregation with disease or variant de novo status), and presence of WHIM-specific phenotypic features observed in affected individuals.

The receptor desensitization defect is believed to be a hallmark of CXCR4 mutations found in WHIM syndrome patients and is the mechanism responsible for their gain-of-function phenotype [40]. To confirm that the newly identified CXCR4 variants (T318N fs*26, S319Lfs*2, R322Qfs*22, S324Ffs*21, S324Pfs*42, L326Q fs*17, K327R fs*17, S339L fs*27, V340L fs*27, S346* and S346Pfs*12) also displayed the same defects, we measured their internalization responses to the natural ligand CXCL12. CXCR4-WT, CXCR4-WHIM (R334* and E343K) and the novel CXCR4 variants were expressed in K562 cell line (lacking endogenous CXCR4 expression [41]), stimulated with increasing concentrations of ligand and stained for remaining CXCR4 expression on the cell surface (Fig. 1D). As a result of CXCL12 stimulation, CXCR4-WT receptor surface levels decreased in a dose-dependent manner. CXCR4-R334*, the most frequent WHIM syndrome variant, displayed a significant internalization defect at all CXCL12 concentrations used, while E343K, a missense variant found in WHIM syndrome patients [42], altered the internalization response only at lower ligand concentrations. All investigated new variants showed defective CXCR4 internalization. The S346* variant was the least affected, displaying significant impairment only at 1nM CXCL12. Interestingly, while a subset of this patient’s neutrophils demonstrates the typical myelokathexis changes with long internuclear strands, the majority of neutrophils had a normal morphology.

In addition to in silico and functional testing, we conducted a database search to find identical or similar (different mutation on cDNA level with identical amino acid sequence) somatic CXCR4 gain-of-function mutations in malignant diseases. Conditions included Waldenström’s macroglobulinemia, lymphoma, and IgM monoclonal gammopathy. All literature reported malignant variants could be identified as somatic CXCR4 gain-of-function mutations. Of the 11 novel variants, 6 variants possessed identical somatic nucleotide changes and 5 variants resulted in identical amino acid sequences but different nucleotide changes (STable2).

First, we assessed the initial type of WHIM-related illnesses that occurred within the first year of clinical presentation. Infections were the most common initial clinical manifestation of WHIM syndrome (88%), dominated by upper and lower respiratory tract infection (49%), followed by otitis media (23%), and skin infections (13%), while HPV-related manifestations and warts were uncommon (2%). Five percent of patients initially presented with congenital abnormalities, including malformation (3%, maxillofacial defects including orofacial cleft) and heart defects (2%, patent ductus arteriosus) (Fig. 2A). Seventy percent of symptomatic patients had their first initial WHIM clinical manifestation before the age of 1 year, and 96% had it before the age of 5 years. To account for any retrospective, observational bias, we calculated the age at initial clinical manifestation of those WHIM syndrome patients that were prospectively diagnosed based on positive family history and age < 1 year (n = 9). Over 50% of those prospectively diagnosed patients had their first WHIM-related manifestation before the age of 3 months and over 90% of prospectively diagnosed patients, had their first WHIM-related manifestation before the age before 1 year of age (Fig. 2B).

Clinical manifestations in our cohort (n = 66) varied across the life span; we therefore calculated the cumulative percentage of WHIM syndrome-related manifestations tested at a given age. STable1 displays the individual observational period, with a median of 12 years. The predominant clinical presentation in infancy was susceptibility to bacterial infections, more than 50% (n = 36) of them had documented bacterial infections during the first year of life. Overall, the most common infections were otitis media and pneumonia developing at a median age of 3 years. Neutropenia was diagnosed prior to 4 weeks of age in 26% of our study population (n = 66) and in 50% by the age of 6 months. Lymphopenia and hypogammaglobulinemia were diagnosed later in life, with more than half of the cohort being diagnosed at the age of 4 years (n = 34) and 10 years (n = 33), respectively. Three patients reported cutaneous warts during early childhood. The incidence of warts increased to more than 25% in adulthood, with a median reported onset of 14 years (n = 17) (Fig. 3A).

Patients that were prospectively diagnosed based on positive family history and < 1 year of age (n = 9), with a median observational period of 9.2 years, had a similar median onset of infections, with 50% developing their first infectious complication at 1 year of age. Neutropenia and lymphopenia were observed in 50% and 33% prior to 4 weeks of age, respectively. Neutropenia prevalence increased to 100% at age of 10 months, while lymphopenia was observed in 50% at the age of 10 months and increased to 80% at the age of 10 years. Hypogammaglobulinemia was rarely diagnosed in early life, with more than half of the cohort being diagnosed at the age 10 years. Warts were never reported in infancy, with one patient developing warts at age 2 years (11%) and 2 additional patients at age 6 and 9 years, respectively (22% and 33%). Patients presented a mean lowest value during follow-up of 852 ± 386 leukocytes/mm³ and 113 ± 95 neutrophils/mm³. In addition, we observed a mean lowest value during follow-up of ALC count of 526 ± 153 cells/mm³ and an AMC count of 51 ± 37 cells/mm³ (Supplementary Figure 2).

Furthermore, we performed a longitudinal analysis of patients’ laboratory values for those that neither received G-CSF or CXCR4 antagonists. Lab values were excluded if patients were actively treated for infection at this time point (n = 35, median observational period of 6.5 years). Patients with WHIM syndrome displayed moderate leukopenia prior to 4 weeks of age, with a median WBC count of 3920 cells/mm³ (n = 5), until age of 1 year with 3590 cells/mm³ (n = 8). However, with age, leukopenia progressively worsened with a median WBC count of 1350 cells/mm³ (n = 13) at age of 2 years, 1240 cells/mm³ (n = 13) at age of 5 years, and 830 cells/mm³ (n = 8) at age 10. Neutropenia varied through life; however, there was a trend towards a progressive loss of neutrophils with a median ANC count of 540 cells/mm³ (n = 6) around birth, to 1238 cells/mm³ (n = 8) at age 1 year, 547 cells/mm³ (n = 12) at age 2, and 197 cells/mm³ (n = 8) at age 5. The progressive loss was most noticeable with ALC: WHIM syndrome patients had overall normal lymphocyte counts at birth up to one year of age, with a median ALC of 2350 cells/mm³ (n = 4) and 1694 cells/mm³ (n = 6), respectively. However, we observed the development of a progressive lymphopenia in early childhood, with an ALC count of 939 cells/mm³ (n = 11) at age 2 and 787 cells/mm³ (n = 13) at age 5. Individual WBC, ANC, and ALC counts over time for every single patient can be seen in (Supplementary Figure 2). Analyzing individual patients longitudinally showed a progressive neutropenia in 20 patients (57%) and progressive lymphopenia in 25 patients (71%) (Fig. 3B).

Treatment data were available for 54 WHIM syndrome patients. Granulocyte colony-stimulating factor (G-CSF) was used in 55.7% (n = 30), IgG replacement therapy (IgGRT) in 53.7% (n = 29) and antibiotic prophylaxis in 38.9% (n = 23) to reduce susceptibility to infections (Fig. 4A). Therapeutic response was scored based on clinical judgment by treating physicians, for detailed definition see material and methods. G-CSF improved infectious complications and increased ANC counts in approximately 50% of patients, while 22% of patients showed improved ANC counts with no effect on susceptibility to infections and 26% of patients failed to increase ANC counts or ameliorate infection frequency in response to G-CSF treatment. Antibiotic prophylaxis and IgG replacement therapy were able to reduce infectious burden (> 50%) in 80% and 74% of patients, respectively, while partial response (< 50%) was observed in 10% and 22%, respectively. Only a minority of patients showed no response to antibiotic prophylaxis and IgGRT (10% and 4.35%), respectively (Fig. 4B). CXCR4 antagonist (plerixafor) was used in 5.6% (n = 3) of patients, which ameliorated the burden and frequency of infection and improved panleukopenia. Of patients, 5.6% (n = 3) underwent hematopoietic stem cell transplantation (HSCT) as definitive treatment for WHIM syndrome, with successful outcome in 2 patients; however, one patient died due to HSCT-associated complications. Details for the patients that underwent HSCT were described by Laberko et al. [43] Twenty-two percent of our cohort required no therapeutic intervention at the time of this study. Data on treatment response of different drug combinations were not considered in this analysis.

We observed a substantial diagnostic delay from first date of recorded neutropenia to the final molecular diagnosis of WHIM syndrome. FH and/or NBS played a crucial role in facilitating early recognition of WHIM syndrome: in 67% (n = 6/9) of subjects, diagnosis was prompted by FH or NBS within the first year of life with a median delay of 1.3 years; in comparison, in 14% (n = 7/49) of subjects, diagnosis was prompted by clinical signs within the first year of life with a median delay of 5 years (p = 0.0387) (Fig. 4C). Those patients with a diagnosis prompted by FH and/or NBS were more likely to receive early therapeutic interventions, e.g., G-CSF and IgGRT, with a median age of 2 and 2.6 years, respectively, compared to those with a diagnosis prompted by clinical signs with a median age of 6 and 9 years, respectively (IgGRT p = 0.04, G-CSF p = 0.09) (Fig. 4D).

Irreversible end-organ damage, including bronchiectasis and hearing loss, was significantly more common in those patients with a diagnosis prompted by clinical signs compared with those with a diagnosis prompted by FH/NBS. Of the former, 27% developed bronchiectasis (n = 9) and hearing loss (n = 3) secondary to infections, while only one patient in the FH/NBS group was diagnosed with bronchiectasis (8%, p = 0.03), and non-developed hearing loss. End-organ damage in WHIM appeared early in life with a median age of 9.6 years; however, we did not find a correlation of age and end-organ damage (p = 0.278) (Fig. 4E). Eighty-three percent of our cohort were admitted to the hospital due to WHIM-related complications, with a median first hospital admission at the age of 2 years. Patients with a family history of WHIM were less likely to be hospitalized compared to those whose diagnosis was prompted by clinical signs (69% vs 87%). We further observed that the age at which patients required inpatient treatment differed between those with a diagnosis prompted by FH/NBS and clinical signs, with a median age of 7 years and 2 years, respectively (Fig. 4F).

In our cohort of 66 patients with WHIM syndrome, we observed 12 patients (18%) with prominent autoimmune manifestations. In addition, we included and analyzed the cases of 2 patients with AI manifestations from Australia and the USA on the basis of a literature search and/or personal communication. Among patients with AI, we observed a predominance of female patients (71% females, 29% males). All WHIM syndrome patients with AI were non-Hispanic whites. AI manifestations were reported in patients that were significantly older than the overall cohort, with a median age of 22.1 years (range 6.2–47.2) compared with those without AI with a median age of 9.0 years (range 1.1–36.4). We could not identify any significant difference in lymphocyte counts, serum immunoglobulin levels or mutation between patients with and without AI. Taking into account the patients from this cohort and the literature search, the most frequent AI complications were autoimmune cytopenias (n = 8, 57.1%), followed by endocrinopathies (n = 4, 28.6%, type 1 diabetes [n = 3] and autoimmune thyroiditis [n = 1]), skin manifestations (n = 2, 14.3%, psoriasis [n = 1] and vitiligo [n = 1]) and two patients presented with arthritis and autoimmune hepatitis (each n = 1, 7.1%) (Fig. 5A). A total of 49.9% (7/14) of patients had more than one AI complication. Specifically, 35.7% of cytopenia cases presented with an additional autoimmune complication, while other autoimmune complications were more likely to occur in isolation (35.7%) (Fig. 5B). Immune thrombocytopenia purpura (ITP) was the most frequent autoimmune complication (n = 7, 53,8%), followed by Coombs-positive autoimmune hemolytic anemia (AIHA) (n = 5, 38%). Evans syndrome was observed in 4 patients (30.7%). No trilineage Evans syndrome was reported, with no autoimmune neutropenia observed in our cohort (Fig. 5C). Cytopenias had a nadir hemoglobin level of 6.4 g/dL attributed to AI (n = 5) and nadir platelet count of 37.000 cells/μL due to ITP (Fig. 5D).

Treatment outcomes were reviewed in detail in patients with ITP (n = 7) and AIHA (n = 5). Mild AIHA (n = 2, median hemoglobin nadir 6.9 g/dL) required no specific therapy. High-dose intravenous immunoglobulin (IVIG) and steroids were frequently used as first-line therapy. Disease control using first-line therapy was achieved in a subset of patients (ITP 27%, AIHA 20%) only. Most patients with ITP required second-line therapy, which most frequently consisted of thrombopoietin receptor agonist (n = 1), B cell depletion with rituximab (n = 2), and mTOR inhibitor, sirolimus (n = 1). Second-line treatment of ITP fully or partially improved the outcome in the majority of cases (93%). Splenectomy seemed to be effective for controlling AIHA (n = 1) but failed to improve ITP (n = 1) (Fig. 5E, F).