This study was approved by the Sidra Medicine ethics committee (IRB protocol 1601002512). Informed consent was obtained from all individual participants included in the study and/or their parents.

Peripheral blood mononuclear cells (PBMCs) were purified by Ficoll density gradient centrifugation. PBMCs were stimulated with 1 μg/mL of anti-CD3 and anti-CD28 antibodies. The cells were cultured 3 days in complete RPMI-1640 medium with 10% FBS at 37°C with 5% CO₂. Then the cells were cultured with 100 U/mL hIL-2 in fresh media, and fresh hIL-2-containing medium was supplemented every 2 days.

PBMCs of 1-2 million in 1 ml of RPMI 1640 medium with 20% FBS were cultured with cyclosporin A and B95-8 culture supernatant. Every following week, complete RPMI 1640 media with 20% FBS was added. After 3-5 weeks, once lymphoblast proliferation was established, the media was replaced with complete RPMI 1640 media with 10% FBS every two days for continued expansion of the cells or cryopreserved for subsequent culture and expansion.

Genomic DNA from blood of the patient was submitted to Invitae and sequenced using their Primary Immunodeficiency Panel, which included 207 genes at the time of analysis. The Invitae diagnostic testing results only found two variants as homozygous. The other VUS were heterozygous and not relevant to the disease phenotype. Familial segregation analysis was conducted by Sanger sequencing of these two homozygous variants. The following primers were used (IKBKB-F 5’-gggctctgtcttcctctgtt-3’, IKBKB-R 5’-gatctgagaggcaaagcagc-3’; AICDA-F 5’-ccagagccgcaataaaagtc-3’, AICDA-R 5’-gccttgtctctgagccattc-3’). IKBKB and AICDA PCRs were carried out using Taq PCR Master Mix Kit (Qiagen) on genomic DNA obtained from activated T cells and EBV-transformed lymphoblasts, respectively.

PBMCs were stained in 1X PBS 0.3% BSA with appropriate antibodies for analysis of isotype-switched memory B cells (APC-CY7 anti-CD19, PE-Cy7 anti-CD27, FITC anti-IgM, PE anti-IgD) and CD45RO-BV605-positive memory T cells (FITC anti-CD3, APC anti-CD4, APC-CY7 anti-CD8). Flow cytometry data were acquired on the NovoCyte (ACEA Biosciences) and analyzed with FlowJo version 10.

Cells were stained with surface marker antibodies (PE anti-CD25, FITC anti-CD3, and AF647 anti-CD127) for 45 minutes on ice. The cells were then washed, fixed/permeabilized using the eBioscience FoxP3 Transcription Factor Staining Buffer kit (ThermoFisher Scientific) and incubated overnight with V450 anti-CD4 and PE-Cy7 anti-FoxP3 (eBioscience) antibodies. Data were analyzed by FlowJo software version 10.

PBMCs were incubated with 1 µM of CFSE, washed, and then cultured with beads conjugated to a combination of anti-CD28, -CD3, and -CD2 antibodies (T cell activation/expansion kit, Miltenyi Biotec) for 3 to 5 days. On days 3 and 5, cells underwent staining using Viability dye (LIVE/DEAD™ Fixable Near-IR Dead Cell Stain Kit, ThermoFisher), V450 anti-CD4- and BV605 anti-CD8 (BD Biosciences) antibodies and analyzed on the flow cytometer.

Activated T cells were expanded in 100U/mL IL-2 and restimulated using 20 nM phorbol 12-myristate 13-acetate (PMA) and 1 mM ionomycin (Sigma-Aldrich) for 5 hours at 37°C. After the first hour of stimulation, Brefeldin-A (Invitrogen) was added to block cytokine secretion. After 5-hour stimulation, cells were washed and surface Fc receptors (Biolegend) were blocked for 15 min at 4°C. Cells were stained with APC-Cy7 anti-CD8, then fixed with BD Cytofix/Cytoperm (BD Bioscience) and stained with AF488 anti-CD4, PE anti-IFNγ, BV650 anti-TNFα, and APC anti-IL-2 (eBioscience).

For AID protein expression, 2x10⁶ EBV-transformed lymphoblasts were harvested, lysed in hot 1% SDS lysis buffer and sonicated. For IκBα and IKKβ protein analysis, 1x10⁶ activated T cells were lysed. Both lysates were subjected to SDS-PAGE using NuPAGE 4–12% Bis-Tris Gels (Invitrogen). Primary antibodies used included AID monoclonal antibody (#4949S, Cell Signaling Technology), recombinant Anti-IKK beta antibody (ab124957, Abcam), recombinant Anti-IκBα antibody (#4814, Cell Signaling Technology) and Anti-β-Actin Rabbit Polyclonal Antibody (HRP) (VWR).

For the analysis of NFκB pathway proteins, anti-CD3/CD28 activated T cells of the patient, their two siblings, and two healthy unrelated controls (HDs) were expanded in the presence of IL-2 for at least 8 days in Advanced RPMI medium (ThermoFisher #12633012) supplemented with fetal bovine serum L-glutamine, streptomycin, and penicillin. 5x10⁵ cells of each sample were stimulated with 20 ng/mL TNFα (Peprotech #300-01A) for the indicated times, after which the cells were washed twice with ice cold phosphate buffered saline. The samples were lysed in boiling lysis buffer containing 1% SDS 0.1 M Tris-HCl (pH 8.0) and then sonicated. The samples were resolved on 10% polyacrylamide gels at 140 V and transferred onto a PVDF membrane at 110 V for 1 hour on ice. The membrane was blocked with 5% milk in TBST and incubated with anti-phospho-p65 (Cell Signaling #3033) and anti-IκBα antibodies (Cell Signaling #4814) in 3% BSA TBST at 4°C overnight and then with goat anti-mouse (ThermoFisher #A-31553) and anti-rabbit (ThermoFisher #G-21234) peroxidase-conjugated secondary antibody (SeraCare #5220-0341) for 1 hour at room temperature. The membrane was treated with Pierce ECL Western Blotting Substrate (ThermoFisher #32106) and developed using ChemiDoc MP Imaging System (Bio-Rad).

Total RNA was extracted from EBV-transformed lymphoblasts generated from the patient, her siblings, and unrelated healthy donors (HDs), including age-matched control. cDNA synthesis was performed as previously described (2, 12) with a Cμ A (5′-GAGGCAGCTCAGCAATC-3′) primer. PCR was performed with the following primers: V3-23 leader exon (5′-GGCTGAGCTGGCTTTTTCTTGTGG-3′) and Cμ B (5′-TCACAGGAGACGAGGGGGAA-3′) (2) (95°C, 5 min; 35 cycles *(94°C, 30sec; 68°C, 30sec; 72°C 45 sec); 72°C, 20 min). PCR products were cloned using the TOPO TA cloning kit (Invitrogen) and V3-23 positive colonies were sequenced. Following manual quality verification of the sequencing chromatograms using UGENE software to confirm single peaks and minimal read length of 292 bp, IgH sequences were aligned to germline sequence (NG_001019.6) and analyzed by Igblast using the default settings to calculate sequence similarity (13). Mutational Frequency was calculated as = 100 - (the number of mutations were divided by the number of nucleotides in the V3-23 region).

The statistical analysis was conducted using GraphPad Prism 9 with Mann-Whitney test.

The present case is a Tunisian patient from consanguineous parents. She was seen aged 9 months in the emergency department with an episode of oral thrush and upper respiratory tract infection, but not admitted to hospital. She was first admitted to the pediatric ward, aged 2 years, with a history of fever, cough and shortness of breath. She was found to have bilateral pneumonia on chest X-ray. She deteriorated clinically and radiologically, and developed hypoxia with respiratory distress despite antibiotic treatment. She was admitted to the Pediatric Intensive Care Unit, treated with oxygen therapy, but did not need ventilation. She also had abdominal pain, constipation, and mild hepatomegaly was reported on ultrasound scan. She was found to have microcytic anemia, and noted to have failure to thrive - her weight was on the 3^rd centile. Her microbiology was positive for respiratory syncytial virus, cytomegalovirus (CMV), and serology was IgM+ for herpes simplex virus. Extensive clinical laboratory evaluation revealed marked hypogammaglobulinemia with elevated IgM levels, high CRP, and high vitamin B12. Cellular immunological investigation was performed, and the percentage of T, B and NK cells were within age-adjusted ranges, while memory CD45RO+ cells were elevated for age ( Table 1 ). She was within normal limits for lymphocyte proliferative responses to PHA, pokeweed and concanavalin A mitogens.

Based on these results, monthly intravenous immunoglobulin (IVIG) replacement therapy was commenced. Although the initial hepatic ultrasound showed mild hepatomegaly, this was repeated during follow-up and was normal. Liver function tests were performed and showed no evidence of transaminitis or elevated bilirubin. CMV PCR was positive at low levels initially (526 IU/mL), and then subsequently became negative within 6 weeks. The anemia was microcytic, and predominantly due to iron deficiency based on ferritin, transferrin saturation, and blood film appearances. The hemoglobin improved with oral iron supplementation. The patient’s overall condition improved with treatment, and she remains clinically stable under regular follow-up. She has not had any recurrent infections since commencement of IVIG, i.e. over 5 years of follow-up. She, her parents, and 2 sisters had an episode of COVID-19 infection in 2022. She receives trimethoprim/sulphamethoxazole prophylaxis twice daily, three times per week, and the family have been advised on the precautions against cryptosporidium infection.

The patient has dizygotic twin siblings, who are two years older and both currently in good health. One of them (Sibling-2) previously had poor weight gain and history of recurrent minor infections according to the mother. However, these infections did not require admission, and weight gain has improved with age.

Due to the presence of marked hypogammaglobulinemia with elevated IgM, a genetic panel for primary immunodeficiencies was ordered to investigate a possible underlying inborn error of immunity. Clinical genetic testing using the Invitae PID panel identified two homozygous missense variants of uncertain significance (VUS) in two candidate genes: IKBKB c.230G>A, p.R77Q and AICDA c.239G>C, p.W80S. Familial segregation analysis by Sanger sequencing of the variants confirmed that the proband was homozygous for the IKBKB p.R77Q variant whereas both parents were heterozygous carriers, one sibling was homozygous for the reference allele, and the second sibling whose history included repeated infections and low weight gain was also homozygous for the p.R77Q variant. IKBKB encodes for IKKβ, and this mutation, which is located in the kinase domain, changes the IKKβ amino acid residue 77 from arginine, a basic and polar amino acid, to glutamine, a neutral and polar amino acid (p.Arg77Gln). According to the literature and population databases, this variant was novel at the time and has since been reported as a VUS in ClinVar.

Sanger sequencing also confirmed that the proband was homozygous for the AICDA p.W80S missense mutation, which changed an aromatic residue to a polar, non-charged residue. Both parents and siblings were heterozygous carriers. The mutation is in exon 3 and affects an amino acid within AID’s catalytic domain known to impact both CSR and SHM.

Pedigree structures for both mutations are shown in Figure 1 ( Figures 1A, B ).

Both variants were classified as VUS, and several in silico variant effect prediction tools indicated that both variants might be deleterious. AID p.W80 and IKKβ p.R77 residues were highly conserved between different species ( Figure 1C ). For the AID p.W80S variant, CADD score was 29.9, PolyPhen-2 score was 1.00 (probably damaging), SIFT score was 0.005 (damaging), MutationTaster predicted disease causing, and PrimateAI score was 0.9045 (damaging). For the IKKβ p.R77Q variant, CADD score was 25.4, PolyPhen score was 0.951 (possibly damaging), SIFT score was 0.336 (tolerated), MutationTaster predicted disease causing, and PrimateAI score was 0.5646 (damaging).

When comparing the wild-type proteins of AID and IKKβ to their mutant forms, structure-based prediction of the variants’ effect on protein stability using DYNAMUT2 revealed a folding free energy change for both structures ( Figure 1D ). Interestingly, we observed that AID p.W80S exhibits higher impact on the stability than the previously reported AID p.W80R mutation (14) ( Figure 1D ).

A flow cytometric assessment of peripheral blood lymphocytes revealed normal percentage of total CD27⁺CD19⁺ memory B cells but a reduction of class-switched CD27⁺IgM^-IgD^- memory B cells compared to healthy donors (HDs) and siblings ( Figures 1E, F ).

To investigate potential links between the R77Q IKKβ variant and a biological defect, we assessed Treg and memory T cell levels and function. We found a similar level of CD4⁺CD25⁺FOXP3⁺ Tregs in the patient compared to HDs ( Figure 2A ). We also observed no difference in frequency of CD45RO⁺ memory T cells between the proband and the siblings ( Figure 2B ). Evaluation of T cell cytokine production revealed a normal secretion of IL-2, TNFα and IFNγ ( Figure 2C ).

We assessed the IKKβ and IκBα protein levels in PBMCs and activated T cell blasts to help rule out a potential hypermorphic impact of the p.R77Q variant, although it would be unlikely due to the recessive mode of inheritance. The immunoblot analysis of PBMCs revealed that our patient had a slightly decreased IKKβ expression and normal IκBα levels compared to controls ( Figure 3A ). IKKβ and IκBα expression, however, was comparable to that of HDs in activated T cell blasts ( Figure 3A ). If there was a gain-of-function, we would expect a decrease in IκBα levels due to increased phosphorylation and degradation.

We additionally evaluated IKKβ activity by assessing IκBα degradation and p65 phosphorylation after NFkB pathway activation with TNFa. Both phospho-p65 and IκBα levels in patient activated T cells were comparable to that of the HDs following stimulation with TNFα. ( Figure 3B , Supplementary Figure 1 ). Thus, the data does not support a possible hypermorphic nor hypomorphic function of the variant.

To further evaluate T cell function, we assessed the proliferative response of T cells to anti-CD2/CD3/CD28 beads after 3 or 5 days of culture. Our findings showed that patient and sibling-2 cells’ proliferative responses were comparable to HDs ( Figure 3C , Supplementary Figure 2 ).

We examined the levels of the AID protein in the EBV-transformed human B-cell lines of sibling-1, sibling-2, patient, and HDs to confirm the in silico predicted destabilizing effect of the W80S AID mutation. The findings showed that the patient cells displayed much lower AID expression than did controls and the heterozygous siblings’ cells ( Figure 4A ). Transcript levels of AICDA were not reduced in the patient compared to controls ( Supplementary Figure 3 ).

Therefore, we evaluated the functional impact of the W80S AID variant on SHM by comparing the mutational frequency within the variable region of the immunoglobulin heavy chain amplified from EBV-transformed B-cell lines from patient, siblings, and HDs. We found a near complete loss of SHM (0.10% ± 0.1915%) from the patient ( Figure 4B , Supplementary Table 1 ).

Variant classification was performed using the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) guidelines 2015 (15), and criteria were applied based on genetic, functional, and clinical data. These ACMG/AMP standards provide a systematic framework for interpreting sequence variants using defined criteria, which help determine the clinical significance of genetic alterations. For the IKKβ p.R77Q variant, careful review of the clinical history of the homozygous sibling found that she is currently healthy, which is inconsistent with the known fully penetrant phenotype of IKKβ deficiency, and meets the BS2 criterion. The normal NF-κB functional results were in favor of the BS3 criterion, resulting in classification to likely Benign/Benign. In contrast, classification of the AID p.W80S variant included moderate evidence from domain localization (PM1), functional impairment (PS3), and a strong phenotype match (PP4), leading to a classification of likely pathogenic/pathogenic ( Table 2 ).