A Novel Homozygous Stop Mutation in IL23R Causes Mendelian Susceptibility to Mycobacterial Disease

Purpose Mendelian susceptibility to mycobacterial disease (MSMD) is caused by inborn errors of IFN-γ immunity. The most frequent genetic defects are found in IL12 or a subunit of its receptor. IL23R deficiency in MSMD has only been reported once, in two pediatric patients from the same kindred with isolated disseminated Bacille Calmette-Guérin disease. We evaluated the impact of a homozygous stop mutation in IL23R (R381X), identified by whole exome sequencing, in an adult patient with disseminated non-tuberculous mycobacterial disease. Methods We performed functional validation of the R381X mutation by evaluating IL23R expression and IL-23 signaling (STAT3 phosphorylation, IFN-γ production) in primary cells (PBMCs, EBV-B cells) and cell lines (HeLa) with or without back-complementation of wild-type IL23R. Results We report on a 48-year-old male with disseminated non-tuberculous mycobacterial disease. We identified and characterized a homozygous loss-of-function stop mutation underlying IL23R deficiency, resulting in near absent expression of membrane bound IL23R. IL23R deficiency was characterized by impaired IL-23-mediated IFN-γ secretion in CD4+, CD8+ T, and mucosal-associated invariant T (MAIT) cells, and low frequencies of circulating Th17 (CD3+CD45RA−CCR4+CXCR3−RORγT+), Th1* (CD45RA−CCR4−CXCR3+RORγT+), and MAIT (CD3+CD8+Vα7.2+CD161+) cells. Although the patient did not have a history of recurrent fungal infections, impaired Th17 differentiation and blunted IL-23-mediated IL-17 secretion in PBMCs were observed. Conclusion We demonstrate that impaired IL-23 immunity caused by a homozygous R381X mutation in IL23R underlies MSMD, corroborating earlier findings with a homozygous p.C115Y IL23R mutation. Our report further supports a model of redundant contribution of IL-23- to IL-17-mediated anti-fungal immunity.1 Supplementary Information The online version contains supplementary material available at 10.1007/s10875-022-01320-7.


Introduction
Mendelian susceptibility to mycobacterial disease (MSMD) is caused by inborn errors of IFN-γ immunity [1][2][3]. Patients with MSMD are defined as individuals who are susceptible to infections by weakly virulent environmental mycobacteria and Bacillus Calmette-Guérin (BCG) disease post vaccination [4]. To date, 19 genes are implicated in MSMD (CYBB, IFNGR1,  IFNGR2, IFNG, IL12RB1, IL12B, IL23R, IL12RB2, ISG15,   IRF8, JAK1, NEMO, RORC, SPPL2A, STAT1, TBX21, TYK2, USP18, ZNFX1). The different genetic disorders are further defined by the nature of the causal alleles (null or hypomorphic), protein levels (normal, low, or absent), and the mode of inheritance (autosomal recessive: AR, autosomal dominant: AD, or X-linked recessive: XL) and whether they present as isolated MSMD or as a part of a broader spectrum (syndromic MSMD) involving infectious susceptibility to other pathogens and/or autoinflammation [1,5,6]. The most frequent genetic defects are found in IL12 or a subunit of its receptor (AR complete IL12p40 deficiency or AR IL12RB1 deficiency), followed by defects in the IFN-γ receptor (AD or AR complete/partial IFNGR1/2 deficiency) [1]. For other genes such as IL23R or IL12RB2, only a single kindred with functional validation of the corresponding genetic defect has been described [7]. This is not hypothesized to be due to increased rarity of loss of function (LOF) mutations in these genes, but due to the lower penetrance of these deficiencies [7]. In case of IL23R deficiency, two affected pediatric-onset cases from the same kindred with a phenotype of disseminated BCG disease were found to have a pathogenic homozygous mutation (p.C115Y) in the extracellular domain of IL23R. Although the mutation did not impair mRNA or protein expression, reduced membrane-bound IL23R and impaired IL-23 signaling was observed. Another IL23R mutation (p.R381Q) has been reported as a partial LOF variant and associates with pulmonary tuberculosis severity [8]. In this study, we describe a kindred with a homozygous mutation (c.1141C > T, R381X) in the intracellular domain of IL23R, presenting as adult-onset disseminated non-tuberculous mycobacterial disease. We used in silico prediction tools, immunophenotyping, and functional analysis of peripheral blood mononuclear cells (PBMCs) and cell lines to corroborate the genotype-phenotype relation in this patient.

gDNA Extraction and Whole Exome Sequencing
For gDNA extraction from whole blood, Purelink DNA Genomic DNA Mini kit (ThermoFisher Scientific) was used according to the manufacturers' instructions. Whole exome sequencing was performed using SureSelect Human All Exon V7 (Agilent) for exome capture (Macrogen®, the Netherlands).

PBMC Isolation
Whole blood was diluted 1:1 with RPMI 1640 and layered over lymphocyte separation medium (LSM) (MP Biomedicals, 0,850,494-CF). Tubes were centrifuged at 400 G for 25 min and PBMCs were harvested and stored in liquid nitrogen until FACS staining or used immediately for functional assays.
mRNA Isolation and Synthesis of cDNA mRNA was isolated from PBMCs using TRIzol reagent (Life Technologies). Five hundred microliters of TRIzol was added to PBMCs and frozen at − 80 °C. Later, samples were thawed and mRNA isolated according to manufacturer instructions. Complementary DNA (cDNA) was synthesized from RNA using the GoScript™ Reverse Transcription System (Promega) according to manufacturer instructions. qPCR Forward and reverse primers were purchased from Integrated DNA Technologies (IDT). A Taqman probe was designed on exon 2 (ccagacatgaatcaggtcactattcaatgg) with primers located on the span junction of exon1-2 (forward primer tcaaacaggttgaaagagggaaac) and exon 2-3 (reverse primer tcctccatgacaccagctga) and a Taqman probe on exon 8 (tgatcgtctttgctgttatgttgtcaattctttct) with primers on the span junction of exon 7-8 (forward primer acagggcaccttacttctgacaac) and on exon 8 (reverse primer agttcggaatgatctgttaaatatccc) HPRT1 or GAPDH was used as housekeeping gene for all performed qPCRs. The reaction was performed as followed: 0.6 µl of primers (0.30 µM), 0.4 µl of probe (0.25 µM), 6 ng of DNA (2 ng/µL) and 1 × Taqman Fast mastermix (ThermoFisher Scientific) were mixed and plated on a 96-well plate. The plate was run on the StepOneTM Real-Time PCR system (ThermoFisher Scientific) and analyzed with StepOne Software v2.3.

Lentiviral Transduction
EBV-B cells were plated in a 96-well plate at a seeding density of 1 × 10 5 cells in 200 μL of cRPMI and transduced using LV vectors as described above. After adding the lentiviral vectors, cells were spinoculated for 90 min at 35 °C and 800 G. Afterwards, they were resuspended and medium

Th Compartment
Frozen PBMCs were thawed and counted, and cell concentration was adjusted to 5 × 10 6 for each sample.

Statistical Analysis
Statistical analysis was only performed if three biological replicates were available in each group. To compare between healthy controls and patient, an unpaired Student t-test (if normally distributed by Shapiro-Wilk) or Mann-Whitney U test (no normal distribution) was used. To compare the effects of stimulation in the healthy control or patient group, a paired Student T test, one-way ANOVA, or a Friedman test with post hoc tests was used depending on normality of the data. Significance levels were defined as followed: *p < 0.05, **p < 0.01, ***p < 0.001. Bars represent the standard error of the mean (SEM).

Identification of a Homozygous R381X IL23R Mutation in a Kindred with MSMD
A 48-year-old patient from Turkish descent (Fig. 1a) was diagnosed at the age of 41 years with a disseminated multimycobacterial infection, with pulmonary (Mycobacterium avium complex, culture, and PCR proven), bone marrow (presence of acid-fast bacilli, not identified by culture nor PCR), and gastrointestinal (Mycobacterium tilburgii, PCR proven) manifestations (Fig. 1b). Despite tuberculostatic treatment during 10 months, he suffered from protracted febrile episodes, systemic inflammation, and general weakness (for detailed case description see supplementary data). Using whole exome sequencing, a homozygous stop mutation in IL23R (p.R381X) was found in the patient and in heterozygous state in his mother and all of his siblings except for one. He had one brother who died of lung cancer and two sisters who died at an early age (no genetic material available), although information on the cause of death could not be retrieved. The other siblings were healthy. The R381X mutation was reported once in heterozygosity in the healthy population database but was not found in homozygosity (gnomAD v2.1.1, ExAC, 1000 Genomes). In silico tools (9) predict the mutation to be deleterious (Fig. 1c, CADD 37) and it resides in the intracellular domain of the IL23R (Fig. 1d), before the JAK2 binding site which is necessary for further downstream signaling through STAT3. This residue is well conserved among mammalian species (Fig. 1e).

R381X Results in Near Absence of Membrane-Bound IL23R
IL23R undergoes extensive alternative splicing, with up to 24 isoforms produced, of which seven are not expressed due to nonsense-mediated decay and five of which are translated to produce a soluble form of IL23R (sIL23R) lacking  [10]. The remaining isoforms are expressed as membrane-bound receptors with varied extracellular domains. All transcripts encoding the putative isoforms contain exon 1, 2, and 3 and therefore we designed our primers to detect this region, measuring the total IL23R transcript level in PBMCs (Fig. 2a). As baseline expression of total IL23R is low, phorbol myristate acetate (PMA) and ionomycin were used to upregulate its transcription. Compared to healthy controls, the patient had tenfold lower IL23R mRNA levels both unstimulated and after PMA/ ionomycin stimulation. Similar results were achieved using a primer-probe set covering the 3' region, indicative of nonsense-mediated decay (FigS1a). Overexpression of a truncated IL23R (F380-HA) could not be detected (FigS1b) and flow cytometric detection of membrane-bound IL23R was severely decreased compared to healthy controls, although not completely absent in PBMCs or EBV-B cells of the patient (Fig. 2b, FigS1c-d). Thus, R381X results in near absent expression of membrane bound IL23R, primarily via nonsense-mediated decay.

R381X Is a Complete Loss of Function Variant When Expressed in HeLa Cells
As IL-23 signals through the IL23R via JAK2 to induce STAT-3 phosphorylation, we first tested whether this response was blunted by the R381X mutation in a HeLa cells. Since IL23R functions as a heterodimeric receptor, we co-expressed both IL23R (WT, R381Q, or R381X) and IL12RB1 to assess the STAT3 phosphorylation response in HeLa cells. The presence of WT and R381Q IL23R on the membrane after transfection was confirmed by confocal microscopy (Fig. 2c) and flow cytometry (Fig. 2d), where it colocalized with IL12RB1 (FigS1e). As expected, R381X IL23R was neither expressed on the membrane nor in the cytosol of HeLa cells (Fig. 2c-d). After stimulation of HeLa cells with IL-23, we observed a strong STAT3 phosphorylation in WT cotransfected HeLa cells, while in the case of a cotransfection with IL23R R381Q or R381X, STAT3 phosphorylation was only moderately increased or unchanged, respectively, compared to untransfected or IL12RB1 with empty vector cotransfected cells (Fig. 2e).
These findings indicate that the IL23R R381X is a complete LOF variant.

Patient EBV-B Cells and PBMCs Have Impaired STAT3 Phosphorylation in Response to IL-23, Which Can Be Restored by Transduction of WT IL23R
Next, we evaluated IL-23 signaling in patient PBMCs and EBV-B cells using flow cytometry ( Fig. 3a-b, FigS2) and Western blot (Fig. 3c-d, FigS3). In contrast to stimulation of patient PBMCs with IL-6 and IFN-α, which led to increased STAT3 phosphorylation, no increase was observed using IL-23, comparable to results seen in PBMCs of an IL12RB1-deficient patient (Fig. 3a-d). The IL-12 signaling pathway was still intact as a normal STAT4 phosphorylation response to IL-12 was observed in patient PBMCs (Fig. 3e,  FigS4). To evaluate if IL-23 signaling could be restored in the patient, we introduced WT IL23R by lentiviral transduction in patient EBV-B cells (Fig. 3f). The introduction of WT IL23R, but not empty vector, in patient EBV-B cells restored STAT-3 phosphorylation upon IL-23 stimulation. These results provide a causal link between the observed defect in STAT-3 phosphorylation downstream of IL-23 and the IL23R R381X variant.

IL23R Deficiency Is Characterized by Decreased Circulating Th17, MAIT, Th1* Cells, and Impaired Th17 Differentiation
To determine the impact of this signaling loss on the patient's immune system, we first compared the frequencies of leukocyte subsets in healthy controls and patient. The frequencies of B, NK, total T cells, and regulatory T cells were normal in the patient (TableS1). Relative to healthy control frequencies, CD4 + T cells were decreased and CD8 + T cells increased, but absolute numbers remained within normal range (Fig. 4a-c, TableS1). The frequency of mucosal-associated invariant T (MAIT) cells (defined as CD8 + Vα7.2 + CD161 + cells) was also severely decreased in the patient (Fig. 4d, FigS5). Further characterization of the CD4 + T cell compartment (Fig. 4e, FigS5) revealed a higher proportion of effector memory T cells (CD3 + CD4 + CD45RA − CCR7 − ) and T effector memory re-expressing CD45RA (TEMRA) cells (CD3 + CD4 + CD45RA + CCR7 − ), while naïve T cells (CD3 + CD4 + CD45RA + CCR7 + ) were decreased. In the CD4 + memory compartment (CD3 + CD45RA − ), the number of Th17 cells (CD3 + CD45RA − CCR4 + CXCR3 − RORγT + ) and Th1* (CD3 + CD45RA − CCR4 − CXCR3 + RORγT + ) were significantly decreased in the patient (Fig. 4f, FigS6-7). Given the low proportion of Th17 cells in the patient, we investigated the effect of IL-23 signaling on the generation of Th17 cells from naïve CD4 + T cells (CD45RA + ) upon stimulation with polarizing cytokines in vitro. IL-17 secretion, a surrogate marker for Th17 differentiation, was low at baseline in the patient and did not increase in a Th17 polarizing condition in contrast to the healthy controls (Fig. 4g).
These results are consistent with the known functions of IL-23 in lymphocyte biology, supporting a loss of IL-23 signaling in the patient.

IFN-γ and IL-17 Secretion by CD8 + , CD4. + T, and MAIT Cells in Response to IL-23 Is Blunted in the Patient
Based on previous observations, identifying IL-23 as a stimulator of IFN-γ especially in alpha-beta T cell subsets and MAIT cells [7], we tested whether the patient alpha-beta T cell subsets and MAIT cells had impaired IFN-γ secretion after IL-23 stimulation. Therefore, we incubated PBMCs in the presence of anti-CD3/anti-CD28 and added conditions with or without IL-23, IL-12, or IL-6. Addition of IL-12 resulted in a significant increase of IFN-γ production compared with the unstimulated condition in CD8 + , CD4 + , and MAIT cells from both healthy controls and the patient. However, we observed that addition of IL-23 was unable to induce IFN-γ production in the patient in contrast to the healthy controls ( Fig. 5a-d, Fig S8a-b). Similarly, this defect of IFN-γ secretion in response to IL-23 was also seen in patient NK and iNKT cells while IFN-γ responses to IL-12 remained normal (FigS9a-c). To characterize the consequences of this functional defect in the presence of a mycobacterial infection, we co-cultured healthy control and patient PBMCs with non-tuberculous mycobacteria to assess the mycobacteria specific IFN-γ. In the absence of available strains identified in our patient, the Mycobacterium abscessus strain (ATCC19977) was used. The addition of IL-23 in the patient did not result in an augmented IFN-γ response compared to Mycobacterium abscessus alone, in contrast to healthy controls where a mild increase in IFN-γ production was seen (Fig. 5e, FigS10). This defect was specific to IL-23, since addition of IL-12 to mycobacteria or PMA/ionomycin as single stimulant resulted in a comparable increase of IFN-γ production. Next, we measured IL-17 production in the supernatant of stimulated PBMCs with anti-CD3/anti-CD28 with or without addition of IL-23. IL-17 secretion was low in both conditions and did not increase upon IL-23 stimulation, whereas healthy controls had higher baseline levels but increased upon IL-23 costimulation (Fig. 5f).

Discussion
We report on a novel, homozygous IL23R (p.R381X) mutation as a cause of MSMD. IL23R is a heterodimeric receptor consisting of an IL12RB1 and IL23R subunit. After binding to IL-23, it signals downstream through TYK2 (attached to IL12RB1) and JAK2 (attached to IL23Rα) to induce phosphorylation of STAT3. The role of IL-23 signaling is best studied in the context of inflammatory disease. A welldescribed example is inflammatory bowel disease, where both mouse models, human observational studies and population genetic studies underscore the importance of IL-23 [11]. The connection of IL-23 signaling to mycobacterial disease is based on the observation that IL23a −/− mice have impaired long-term control of pulmonary Mycobacterium tuberculosis infection [12], the fact that exogenous IL-23 can complement IL-12 deficiency for restoring mycobacterial immunity in mice [13], the association of a partial LOF variant c.1142G > A (R381Q) with active pulmonary tuberculosis [8] and the report of a kindred with MSMD harboring a homozygous complete LOF mutation in IL23R [7]. In addition, a second patient with MSMD and a splice homozygous splice site mutation, c.367 + 1G > A, at the exon 3 of IL23R is reported but the effect of this variant on IL23R expression and signaling was not reported [14].
In contrast to the first reported kindred by Martínez-Barricarte et al. where both patients had a homozygous missense mutation (p.C115Y) in the extracellular domain resulting in impaired IL-23 signaling, our patient has an adult-onset phenotype and a stop mutation (p.R381X) located in the intracellular domain before the JAK2 binding site. The mutation resides in a well-conserved residue and was predicted to be LOF using in silico prediction tools. Functional testing revealed that our patient had severely decreased total IL23R mRNA levels and membrane-bound IL23R. The functional impact of this mutation was assessed by determining STAT3 phosphorylation following IL-23 stimulation in primary cells and in HeLa cells transfected with WT IL23R or the partial LOF variant R381Q [15,16]. The R381X mutation in our patient has a complete LOF effect in both overexpression and primary cells, which can be restored by lentiviral transduction of WT IL23R. The immunophenotype of the IL23R-deficient patient was characterized by a low number of MAIT cells, resembling the previously reported IL23R patient but also STAT3, IL12RB1 andTBX21 patients [6,17]. Furthermore, Th17 and Th1* cells were significantly reduced, differing from the previously reported patient where only a mild reduction was observed [7]. In addition, the relative proportion of CD4 + and CD8 + T cells in our patient differed from healthy controls in contrast to the previously reported IL23R deficient patient [7]. The effect of a homozygous stop mutation on Th17 cell development is consistent with previous data reporting reduced circulating Th17 cells in individuals carrying a heterozygous R381Q mutation [16]. Moreover, the observation of reduced Th17 cells is supported by studies demonstrating a central function for IL-23 in TH17 differentiation in mice [18]. In addition, we showed that the differentiation of naïve T cells to Th17 was impaired, confirming previous results [7]. Remarkably, our patient did not have any susceptibility to fungal infections. This supports the hypothesis that-at least in humans-the contribution of IL-23-to IL-17-mediated antifungal immunity might be redundant [7]. Finally, we assessed IFN-γ secretion by alpha-beta T cell subsets (CD4 + , CD8 + ) and MAIT cells upon IL-23 stimulation. IL-23 is a less potent IFN-γ inducer in T cells compared to IL-12 as described by Martínez-Barricarte et al., characterizing IL-12-and IL-23-specific effects in different immune subsets [7]. We observed a mild, but significant increase of IFN-γ production by alpha-beta T cell subsets and MAIT cells in healthy controls when stimulated with IL-23, while this was absent in the patient. IL-12 stimulation remained intact, confirming a selective IL-23 signaling defect. Furthermore, in the context of a non-tuberculous mycobacterial infection in vitro, the addition of IL-23 did not increase IFN-γ in the patient, in contrast to healthy controls. In addition, IL-17 secretion by stimulated PBMCs in the patient was low and did not increase upon IL-23 costimulation.
In conclusion, we identified a patient with MSMD and a novel stop mutation (p.R381X) located in the intracellular domain of IL23R. This report illustrates a mechanism of abolished membrane-bound IL23R expression. Functionally, this negatively affects the development of Th17 and MAIT cells and impacts IL-23-stimulated IL-17 and IFN-γ production by alpha-beta T cell subsets and MAIT cells. Our findings together with the report of a previously validated IL23R patient with MSMD and existing data on mycobacterial susceptibility of IL23a −/− mice link the genotype and phenotype in our patient. The absence of fungal infections in this adult patient and the previously reported pediatric onset patients are consistent with the hypothesis that the contribution of IL-23-to IL-17-mediated anti-fungal immunity might be redundant, despite the presence of decreased circulating Th17 cells and reduced IL-17 secretion in response to IL-23.
Author Contribution FS initiated the study, performed genetic analysis, designed and performed the experiments, analyzed the data, and drafted the initial manuscript.
FL assisted in designing and performing the experiments for qPCR, cloning, transfection and transduction, and revised the manuscript.
KDK assisted in performing experiments and revised the manuscript.
MW assisted in the lentiviral transduction and the conceptualization of the experiments and revised the manuscript.
MG and JN contributed to the conceptualization of the experiments and revised the manuscript.
EP assisted in confocal microscopy and revised the manuscript. PDM was responsible for the clinical care of the patient and revised the manuscript.
XB was involved in the diagnostic process and revised the manuscript.
RG assisted with the production of lentiviral vectors for transduction and revised the manuscript.
AL and SHB contributed to the conceptualization of the study and revised the manuscript.
RS initiated and supervised the study, contributed to the conceptualization of the study, and revised the manuscript.
Funding FS (11B5520N) and JN (11C3521N) are fellows of the Fonds Wetenschappelijk Onderzoek-Vlaanderen National Fund for Scientific Research (FWO). RS is FWO senior clinical investigator fellows (1805518 N, respectively) and received funding from KU Leuven C1 (C12/16/024). This work was supported by the VIB Grand Challenges Program.

Data Availability
The authors confirm that the data supporting the findings of this study are available within the article and/or its supplementary material.

Fig. 5
Impaired IFN-γ and IL-17 production in response to IL-23 stimulation. a FACS plots of IFN-γ secretion by CD4 + T cells in 1 HC and patient, pre-activated PBMCs with anti-CD3/CD28 with or without addition of IL-23 (100 ng/mL), IL-6 (10 ng/mL), or IL-12 (40 ng/mL) for 7 days. FMO IFN-γ was used for proper gating. Representative of 3 independent experiments b-d % of IFN-γ secreting CD8 + , CD4. + , and MAIT cells, pooled from 3 independent experiments. To assess differences between conditions within HCs or patient, a Friedman test (p < 0.0001) with post hoc Dunn's multiple comparison test was used. To compare HC and patient groups, a Mann-Whitney U test was used for each condition. e IFN-γ measured from supernatant of PBMCs in co-culture with Mycobacterium abscessus ATCC19977 for 3 days with or without IL-23 or IL-12. PMA ionomycin was used as a positive control condition. Data from 1 experiment. f IL-17 secreted by stimulated PBMCs with CD3/ CD28 with or without addition of IL-23 (100 ng/mL), data from 2 experiments. To assess response in HCs, a paired Student T-test was used ◂ Declarations Ethics Approval and Consent to Participate The study was conducted in accordance with the principles of the Helsinki Declaration and approved by the ethics committee of the University of Leuven (s58466). Written informed consent was obtained from patients and healthy control subjects.

Consent for Publication
Written informed consent for the publication of this research was obtained from the study participants.

Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. RS is member of the European Reference Network for Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (Project ID No 739543).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.