Three Patients With Disseminated Mycobacterial Infections Due to Severe Defects in Interferon Gamma Receptor Signaling: A Challenging Diagnosis

IFN–gamma receptor (IFNGR) signaling via STAT1 is crucial in the defense against intracellular pathogens. Defects in this pathway enhance the susceptibility to infection by otherwise weak pathogenic mycobacteria, resulting in a primary immunodeficiency called mendelian susceptibility to mycobacterial disease (MSMD). Here we describe three patients with MSMD caused by variants in the IFNGR1 or STAT1 genes. All three patients presented with disseminated non-tuberculous mycobacterial infections caused by M. avium , M. persicum or M. bovis BCG respectively. Whole-exome sequencing (WES) was used as the first line diagnostic approach, however in all patients additional analysis was crucial to make the definite diagnosis. In Patient 1, only one heterozygous autosomal recessive variant p.(Val63Gly) in the IFNGR1 gene was identified. Patient 2 was compound heterozygous for the pathogenic null p.(Val68Lysfs*6) variant and the hypomorphic p.(Ile37Thr) variant in IFNGR1. In Patient 3 a novel variant in the STAT1 gene c.1379A>T, p.(Asn460Ile) was identified. Additional genetic analysis identified a second novel complex Alu-insertion in the IFNGR1 gene in Patient 1. Functional analysis showed that Patients 1 and 2 had reduced expression of IFNGR1. All patients had reduced phosphorylation of STAT1 and absent induction of SOCS1 mRNA after IFN- γ stimulation. While STAT1 phosphorylation was normal after IFN – α stimulation in Patient 1 and 2, it was mildly reduced in Patient 3. We conclude that functional assays are crucial to assess the extent of IFNGR signaling defects when new combinations of bi-allelic or non-conclusive genetic variants are found, which is important in the determination of clinical treatment.


Introduction
Mendelian susceptibility to mycobacterial disease (MSMD) is a rare inherited disorder characterized by infection with weakly virulent mycobacteria including Calmette-Guérin bacillus (BCG) in otherwise healthy individuals (1,2). Mycobacterial disease usually begins in childhood and less frequently in adolescence and adulthood. Patients are not restricted to susceptibility to mycobacterial species only, as they are also vulnerable to salmonellosis, candidiasis and tuberculosis (2).
This pathway is pivotal to the innate immune defense against mycobacteria, bridging between myeloid cells and lymphoid cells (15). Upon mycobacterial infection, mononuclear phagocytes produce interleukin (IL) 12, which activates T cells and natural killer cells via the IL-12 receptor, an IL12RB1 and IL12RB2 heterodimer. Upon ligand binding the IL-12 receptor signals through TYK2 and JAK2, resulting in STAT4 phosphorylation, followed by STAT4 homo-dimerization and nuclear translocation and finally IFN-production. In response, IFN-binds to its receptor, IFNGR, a heterodimer of IFNGR1 and IFNGR2 chains. This ligand-receptor binding leads to downstream phosphorylation of JAK2, JAK1, and STAT1. Homodimer phosphorylated STAT1 (pSTAT1) will translocate into the nucleus and binds to the IFN-γ activation sequence (GAS) elements, enhancing transcription of IFN-γ-responsive genes. These series of events will eventually enable macrophage activation, differentiation, and further upregulation of proinflammatory cytokine production, including IL-2 and TNF-α (16). Signaling via IFN-is tightly regulated by the SOCS (Suppressors of Cytokine Signaling) 1 protein, which acts as a negativefeedback inhibitor (17). IFNGR1 and IL12RB1 deficiency are the most common causes of MSMD, accounting for almost 80% of all genetically diagnosed cases (14). The pathogenic variants in IFNGR1 can be either autosomal recessive (OMIM #209950) or dominant (OMIM #615978) (18), and can result in complete or partial functional defects (3,4,19,20). Lack of surface expression leads to a complete loss of function, while expression of mutated IFNGR1 can lead to either partial loss of function or complete loss of function. Complete loss of function in IFNGR signaling results in severe mycobacterial infections in early childhood and poor survival. However, variants with partial loss of function may result in a milder clinical phenotype that manifests later in childhood.
In this manuscript, we describe 3 patients with presentation of MSMD with novel (combination of) variants in genes involved in the IFN-γ ⁄STAT1 signaling pathway. These cases illustrate the clinical and diagnostic challenges that can be faced, and show that additional genetic and functional immunological assays help to explain the clinical phenotype and can be helpful in the diagnostics of suspected IFNGR signaling defects in patients.

Ethical approval
Following approval from the Erasmus MC Medical Ethics Committee (MEC-2013-026, MEC-2016-606, or MEC-2016-202), written informed written consent was obtained from all parents and healthy controls before blood donation in accordance with the Declaration of Helsinki.

Genetic analysis
Genomic DNA of all three patients was extracted from peripheral blood samples using standard procedures. A diagnostic WES-based PID gene panel analysis was performed, which included around 400 PID-associated genes based on the IUIS classifications 2017 or 2019 (21,22 Long range PCR was performed on DNA from patient 1 to amplify the genomic region surrounding exon 3 using the forward primer 5'-CCTGGTGAATTCTACTTTTCTTCAA-3' and the reverse primer 5'-AGTGTTTCTTAAGCATTGTGATAATTT-3' using Phusion high fidelity DNA polymerase (Thermo Fisher, Waltham, US). The longer PCR product in Patient 1, which was not observed in the healthy controls, was extracted from gel and purified using the gel extraction kit (Qiagen, Hilden, Germany). The purified band was sequenced and analysed using CLC main workbench 7 (Qiagen). The sequence of the insertion was uploaded in RepeatMasker (https://www.repeatmasker.org/cgi-bin/WEBRepeatMasker) (23).

IFN-gamma and IFN-alpha stimulation
Peripheral blood mononuclear cells (PBMCs) were isolated from fresh blood from 3 patients and healthy controls via density gradient centrifugation (Ficoll-Hypaque, GE Healthcare life sciences) as previously described (24). PBMCs were cultured in RPMI 1640 (Life Technologies) culture medium supplemented with 10% FCS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 g/ml streptomycin-sulfate in 96-well round bottom plates (Thermo Scientific) at a density of 5x10 6 cells/ml. Freshly isolated PBMCs were starved in 1ml serum free RPMI 1640 culture media for 1h at 37 °C in 5% CO2 incubator. After starvation, cells were centrifuged at 500xg for 5min. Cell  For CD64 membrane staining, freshly obtained PBMCs were starved in 1ml serum free RPMI 1640 culture media for 1h at 37 °C in 5% CO2 incubator. After starvation, cells were centrifuged at 500xg for 5min, and 1x10 6 cells were seeded in 96 well round bottom plate with/without IFN-γ (60IU/ml in MilliQ, R&D Systems, Abingdon, UK) in RPMI 1640 culture medium containing 10% FCS and antibiotics. Cells were cultured for 24h at 37 °C in 5% CO2 incubator.

SOCS1 expression
Freshly isolated PBMCs were starved in 1ml serum free RPMI 1640 culture media for 1h at 37 °C in 5% CO2 incubator. Around 2x10 5 to 3x10 5 cells were taken and remained unstimulated or stimulated for 0.5h, 1h, 2h, 4h and 6h with IFN-γ (60IU/ml in MilliQ, R&D Systems, Abingdon, UK) or IFN-α (10 4 IU/ml in PBS, PeproTech, London, UK) in a 37 °C water bath. After stimulation, cells were harvested, and RNA was extracted using GenEluteTM Mammalian Total RNA Miniprep Kit (#SLCC7097,Sigma-Aldrich) and reverse transcribed into cDNA using Reverse Transcriptase reaction with random primers (Thermo Fisher Scientific). Real-time quantitative PCR for SOCS1 and GAPDH was performed with SOCS1 primer/probe mix (Hs00864158_g1, Thermo Fisher Scientific) and GAPDH primer/probe mix (Hs99999905_m1, Thermo Fisher Scientific). All tests were run on a 7900HT Fast Real-Time PCR System(Thermo Fisher Scientific). GAPDH, the housekeeping gene expression, was used to normalize SOCS1 gene expression results.

Data visualization and statistics
Graph plots were made using Graphpad Prism V9.0.

Patient 1
A 6-year old girl was admitted to our pediatric ward with progressive cough, fever and tachypnea unresponsive to repetitive courses of common antibiotics. She is the only child of healthy, nonconsanguineous Caucasian parents. She had a remarkable medical history with multiple hospital admissions (>20) for bronchial hyperactivity, lower respiratory tract infections and a coarctectomy at the age of 2.

IFNGR1 expression is reduced in the patients with the IFNGR1 variants
Since variants in IFNGR1 can result in loss of expression of IFNGR1 we measured surface expression of IFNGR1 using the IFNGR1-specific monoclonal antibody (CD119, clone GIR-208).
The expression of IFNGR1 was severely reduced on monocytes of Patient 1 (Figure 2

Bi-allelic IFNGR1 variants leads to substantial reduction in STAT1 phosphorylation via IFN-γ signaling in monocytes
The pathogenic variant p.(Val63Gly) in Patient 1, was described previously as leading to autosomal recessive partial IFNGR1 deficiency (26). Patients reported with this pathogenic variant had low levels of IFNGR1 expression on the cell surface and low binding of IFN-to the receptor present on the cell surface. It was shown that stimulating monocytes with high doses of IFN-γ resulted in a better IFN-γ signaling, although still reduced compared to healthy controls (25). In line with this, we also observed that phosphorylation of STAT1 (pSTAT1) was absent in response to stimulation with low IFN-γ concentration compared with healthy controls (Figure 3). This reduction in pSTAT1 could not be attributed to reduced STAT1 expression, since the levels of STAT1 in Patient 1 were in the normal range of the controls (Figure 4). Furthermore, the phosphorylation of STAT1 upon IFN-α stimulation was normal, confirming that the defect in STAT1 phosphorylation is specific for IFN-γ signaling ( Figure 5).
The splice variant c.373+1G>T in Patient 2, has been described as pathogenic autosomal recessive variant and leads to complete IFNGR1 loss of function. The second variant p.(Ile37Thr) has also been described functionally as partial deficient (12,13). Upon IFN-stimulation, we found that the monocytes of Patient 2 had reduced expression of pSTAT1 compared with healthy controls  (Figure 3), but the phosphorylation of STAT1 was slightly decreased. Moreover, it seemed to be delayed upon IFNγ and IFN-α stimulation in monocytes compared to healthy controls (Figure 3 and Figure 5).

No upregulation of SOCS1or CD64 upon IFN-stimulation
Page 12 of 21 IFN-mediated STAT1 activation results in the transcription of IFN regulated genes, including SOCS1 and CD64. To further investigate the effect of the variants identified in the three patients, we determined if SOCS1 mRNA was upregulated upon stimulation with IFN-γ or IFN-α. While in healthy controls SOCS1 mRNA level clearly increased after stimulation, in Patient 1 and 2 no upregulation of SOCS1 was observed after stimulation with IFN- (Figure 6 A). In contrast, IFNα stimulation induced high levels of SOCS1 mRNA expression in Patient 1 and Patient 2 ( Figure   6B). These data confirm that there is a functional loss in IFNGR1 signaling in Patient 1 and 2, while signaling via the IFN-α receptor is still functional. In Patient 3, the expression of SOCS1 seems normal after 0.5h of stimulation with IFN-γ, but was reduced after 1 and 2 hours of stimulation. We also measured CD64 membrane expression in monocytes after IFN-γ stimulation in Patient 2. After 24h of IFN-γ stimulation, there was no increase in CD64 surface expression, while in the healthy control CD64 expression increased after IFN-stimulation (Figure 7).

Discussion
We describe three patients with presentation of MSMD with new (combination of) bi-allelic or de novo variants in genes involved in the IFN-γ ⁄STAT1 signaling pathway. All three patients presenting with disseminated non-tuberculous mycobacterial (NTM) infections caused by M. avium, M. persicum and M. bovis BCG, respectively. On presentation none of these patients were previously known with a primary immune deficiency. A general approach in diagnosing patients with disseminated NTM infections has been proposed in literature using flow cytometry, lymphocyte immunophenotyping and sequencing of genes known to be involved in NTM host immunity (16,26). Measurement of IFN-γ plasma levels during infectious episodes can also be helpful as patients with complete IFNGR deficiency will display highly elevated plasma levels of . However, since recent advances in technologies with regard to genomic analysis, a genotype-first approach for new patients with a suspicion of an inborn error of immunity has currently become a more typical approach (28). Since availability for rapid genetic testing is available in our clinical setting, we used whole exome sequencing with a PID gene panel as the first-line diagnostic approach.
In Patient 1, primary genetic analysis revealed only one heterozygous IFNGR1 variant. Because  (29). They have previously also been described in patients with X-linked severe combined immunodeficiency and with X-linked agammaglobulinemia (30). In Patient 2, heterozygosity for two different, known pathogenic variants in the IFNGR1 gene were found. Bi-allelic different hypomorphic variants in this gene have not been described extensively in literature and the functional effects in an individual case are unpredictable. In the third patient, a de novo novel variant in STAT1 was found.
AD STAT1 deficiency is a known cause of disseminated BCG infection and multifocal osteomyelitis, warranting molecular phenotyping in order to prove the pathological effect of this novel variant. These cases illustrate that interpretation of the significance of these novel (combination of) genetic variants in each of these cases can be a diagnostic challenge which we encounter more frequently using advanced genetics such as WES as first-line diagnostics. The genetic and functional analysis helps in determining the exact diagnosis, as the clinical symptomatology in Patient 2 and 3 could also fit with the diagnosis LCH. Furthermore, the analysis indicates also that we should continue the search for a second variant missed by WES-based analysis, when there is a strong clinical suspicion for the respective disorder.
In these cases, (functional) immunological phenotyping in combination with the outcomes of genetic testing is important in order to make a definitive diagnosis and predict the prognosis of an individual patient. This may lead to very specific therapeutic decisions. When the patient is considered to have the phenotype of complete autosomal recessive IFNGR1 deficiency for instance, allogeneic stem cell transplantation should be considered as a realistic therapeutic option, whereas a phenotype of partial IFNGR1 deficiency will be treated with watchful waiting and/or prophylactic antibiotics. Patients with novel combinations of variants such as in Patient 1 and 2 are diagnosed frequently, as novel variants in MSMD-related genes are still found (31)(32)(33). These cases should be carefully characterized using immunological phenotyping. Flow cytometric detection of the affected protein can be a rapid test in the clinical diagnosis of MSMD but is not always feasible or conclusive, depending on choice of antibodies (25,34). For instance, a previous report indicate that patients homozygous for the p.(Val63Gly) variant have reduced surface expression of IFNGR when using the GIR-208 IFNGR1 antibody, but nearly normal using a different monoclonal antibody (MMHGR-1) (25). In Patient 1, the expression of IFNGR1 was significantly reduced on monocytes of Patient 1 using the IFNGR1 GIR-208 antibody, which means that the variant on the other allele, can either not be detected by the IFNGR1 GIR-208 antibody, or more likely is not expressed since we also could not detect mRNA transcripts from this allele. Furthermore normal protein expression does not necessarily preclude normal signaling function (25,35,36).
In the patients presented in this study, who all had a clinical suspicion of MSMD and genetic variants in known MSMD genes, we chose to proceed with STAT1 phosphorylation and IFN-γ induced SOCS1 and CD64 expression assays as markers for IFN-γ signaling. The results of these assays helped in assessing the severity of the patients' phenotype. In Patient 1, the absence of STAT1 phosphorylation and SOCS1 mRNA expression combined with the elevated plasma IFN-γ level and life-threatening presentation of disease led to consideration of allogeneic hematopoietic stem cell transplantation as a treatment option, for which patient and her parents are counselled. Patient 2 is still under treatment for his disseminated NTM infection and recovering very well. In Patient 3 the STAT1 phosphorylation was slightly decreased and is therefore alone not conclusive to confirm the pathogenicity of the p.(Asn460Ile) variant in STAT1. Therefore we also tested the upregulation of SOCS1 mRNA upon stimulation with IFN-γ. These results were more conclusive and showed a lack of upregulation of SOCS1. The asparagine at position 460 in STAT1 is located in the DNA binding domain, and crystal structures suggest that Asn460 is the only amino acid that directly interacts via hydrogen bonds to DNA in the GAS sequence, which implies that this amino acid is crucial for STAT1 function (37,38). Together, these functional data and the fact that the variant is de novo strongly suggest that Patient 3 has an AD STAT1 deficiency.

Conclusion
Patients presenting with disseminated NTM infection in early or late childhood with bi-allelic or non-conclusive genetic variants in genes involved in NTM immunity need to be phenotyped by functional assays of the IFNG/STAT1 signaling pathway in monocytes. These assays are helpful in assessing the extent of IFNGR pathway signaling defects and in clinical decision making in how to treat the patient.

Acknowledgements
The research for this manuscript was performed within the framework of the Erasmus Postgraduate School Molecular Medicine.              Membrane CD64 expression in Patient 2 after IFN-γ stimulation Flow cytometric analysis of CD64 membrane expression on monocytes showed that Patient 2 does not upregulate CD64 after 24h stimulation with IFN-γ in contrast to the healthy control (HC). HC: n=1; Patient 2: n=1.