THP-1 cells were obtained from the American Type Culture Collection (ATCC #TIB-202). EBV-B cells from patients and healthy controls were derived from peripheral blood mononuclear cells (PBMCs) (27). Informed written consent was obtained in accordance with the Declaration of Helsinki and ethical approval from the Great Ormond Street Hospital for Children NHS Foundation Trust and the Institute of Child Health Research Ethics Committee (Reference Number: 06/Q0508/16). For cell culture details, see Supplemental Methods.

JAK1 knock down (KD) and scrambled control (Sc) cell lines were generated using lentiviral vectors expressing short hairpin RNA (shRNA) sequences as previously described (17). For further information, see Supplemental Methods.

THP-1 cells were stimulated or left unstimulated with 50 ng/ml IFN-γ (Invitrogen) for 24h. Total RNA from cells was extracted using RNAeasy kit (Qiagen) and converted to cDNA by reverse-transcription using Quantitect reverse transcription kit (Qiagen). Determination of mRNA level was performed by RT-PCR using specific primers ( Supplementary Table 2 ) and QuantiTect SYBR^® Green PCR Kit (Qiagen) according to manufacturer’s instructions. Fold changes were calculated using the DDCT2 (-Delta Delta C(T)) method and results normalized with respect to the values obtained for the endogenous ACTIN and GAPDH cDNA. See supplemental methods for assessment of MX1 and OAS1 following IFN-α stimulation.

THP-1 cells were stimulated with 10³ IU/ml IFN-α2b or 50 ng/ml IFN-γ. EBV- B cells were stimulated with 10⁵ IU/ml IFN-α2b or 500 ng/ml IFN-γ for 10 min. Cells were fixed using fix buffer I and permeabilized using Perm Buffer III (BD Biosciences) for 30 min at 4°C, and labelled with anti-pSTAT1 (clone 4a, BD Biosciences) or anti-STAT1 (clone 1/Stat1, BD Biosciences) antibodies for 60 min at room temperature. At least 10000 gated events were acquired on a BD LSRFortessa cytometer and data were analyzed using FlowJo software (Tree Star Inc., USA). Data on graphs is shown as relative increase (mean fluorescence intensity (MFI) of stimulated cells – MFI of unstimulated cells/MFI unstimulated cells).

For details of Mycobacterium bovis Calmette–Guérin (BCG) and Salmonella typhimurium, including culture, see Supplemental Methods. THP-1 cells were differentiated into macrophages using 10 ng/ml of phorbol myristate acetate (PMA) for 48 h and then were left unstimulated or stimulated with IFN-γ 50 ng/ml for 18 h before infection. Cells were infected using stocks (BCG) or bacteria in mid-log grow phase (Salmonella), using a multiplicity of infection (MOI) of 20:1 for BCG expressing-mCherry and 10:1 for BCG and Salmonella. Monolayers were incubated for 4 h with BCG and 30 min with Salmonella at 37°C in 0.5% CO₂. Infected cells were washed to remove extracellular bacteria. To kill extracellular bacteria after Salmonella infection, cells were incubated in complete medium with Gentamicin (100 µg/ml) for two hours. Subsequently, macrophages were incubated in fresh complete medium, in the presence or absence of IFN-γ (50 ng/ml) for different time points (detailed in the legends).

Cells were lysed at 3 days for BCG infection or 24 h for Salmonella infection with 0.05% SDS w/v in H₂O and serial dilutions plated out on Middlebrook 7H11 or LB agar plates followed by incubation at 37°C for 14 days for BCG infection or 12 h at 37°C after Salmonella infection. Bacterial survival, measured as CFU for each condition, was expressed as a percentage of the CFU counted in the untreated Sc control.

Macrophages differentiated from Sc and KD THP-1 cell lines using PMA, were left unstimulated or stimulated with IFN-γ (50ng/ml) before infection with BCG expressing-mCherry strains (MOI 20:1). After phagocytosis, cells were washed and incubated in complete medium in the presence or absence of IFN-γ (50 ng/ml) for the given time points. Cells were removed from the plate using Accutase^® solution (A6964, Sigma Aldrich), washed with PBS, fixed in 4% paraformaldehyde (PFA) for 10 min and analyzed by flow cytometry (BD LSRFortessa) using FlowJo.

200,000 THP-1 cells were differentiated on 35mm glass bottom dishes (Fluorodish) for microscopy experiments. For further details see Supplemental Methods.

BCG-lux were labelled with pHrodo™ (Invitrogen) at a concentration of 25mM according to the manufacturer’s instructions, except for omission of the 100% methanol step. Approximately 100,000 CFU were resuspended in 500 µl buffer at pH 7. Samples were then acquired on a BD Fortessa flow cytometer (BD LSRFortessa), and pHrodo fluorescence was measured in the PE-Texas Red channel, and analyzed using FlowJo.

Macrophages were left unstimulated or stimulated with IFN-γ (50ng/ml) before BCG infection (MOI 10:1). Percentage of apoptosis was determined using APC Annexin V apoptosis detection kit with PI (BioLegend 640932) according to manufacturer’s instructions by flow cytometry (BD LSRFortessa), and analyzed using FlowJo.

Primary dermal fibroblast and EBV-B cells viral assays were performed as previously described (24, 28–31). For further details see Supplemental Methods.

Statistical analysis was performed with Graph Prism Version 5.01. The following tests were used: One-way ANOVA; Wilcoxon-Signed Rank for pairwise comparisons; Mann-Whitney for unpaired comparisons. Results were expressed as mean ± SEM. P-values < 0.05 were considered significant.

As the patient with AR partial JAK1 deficiency presented predominantly with MSMD ( Table S1 ), we sought to establish a model to examine the role of JAK1 in myeloid cells during mycobacterial infection. As patient blood was not accessible, for this study we generated a THP-1 monocytic cell line with sub-total JAK1 knock-down (KD) to model partial JAK1 deficiency using lentiviral vectors expressing short hairpin RNA (shRNA) sequences. The JAK1 KD model does not fully recapitulate the impact of the patient’s variants which preserved expression of mutant JAK protein that had reduced signaling function (16) but allows assessment of reduced JAK1 function by reducing protein expression. Compared to control shRNA, JAK1 shRNA substantially reduced JAK1 messenger RNA expression for three out of four hairpins tested ( Figure S1A ). THP-1 cells transduced with JAK1 shRNA #3 were utilized for further studies. Compared with shRNA control cells, 25-30% JAK1 protein expression was detected in JAK1 shRNA #3 cell lines using western blotting, in keeping with partial knock down ( Figure S1B ). To test whether the level of JAK1 reduction was sufficient to impair JAK1 protein function, we studied JAK1-mediated phosphorylation of STAT1 in response to IFN-γ stimulation, using flow cytometry. We observed a significant decrease in STAT1 phosphorylation following IFN-γ stimulation in the THP-1 JAK1-KD cells compared to scrambled control shRNA cell lines (Sc) (p<0.05) ( Figures 1A, B ). Surprisingly, STAT1 phosphorylation in response to IFN-α stimulation was not significantly reduced ( Figures 1C, D ), suggesting that partial impairment of JAK1 function affects predominantly the IFN-γ response in THP-1 cells. Following stimulation with IFN-γ, upregulation of interferon regulatory factor 1 (IRF1) and Class II Transactivator (CIITA) mRNA was significantly lower in the KD than Sc lines, indicating impaired downstream gene transcription in JAK1 deficiency ( Figures 1E, F ). In contrast, upregulation of 2’-5’-oligoadenylate synthetase 1 (OAS1) and MX1 in response to IFN-α stimulation was preserved ( Figures 1G, H ). Our findings were unlikely to be due to alterations in the surface expression of IFN-γ or IFN-α receptors as surface levels of IFNGR1 and IFNAR2 were comparable between JAK1 KD and Sc cell lines at baseline and after stimulation with their ligand ( Figures S2A–D ). As seen in THP-1 cells, partial JAK1 deficiency had a clear impact on IFN-γ signaling in primary dermal fibroblasts and EBV B-cells from the patient with AR JAK-1 deficiency, resulting in significantly reduced STAT1 phosphorylation following IFN-γ stimulation ( Figures 1I, J ) and ( Figures S3A, B ), albeit to a lesser degree that seen in fibroblasts from a patient with complete IFN-γR deficiency (16) ( Figure S3B ).

To test the impact of reduced JAK1 function on IFN-γ-mediated host defense to intracellular pathogens we utilized BCG as a well-established model for mycobacterial infection (32). Macrophages differentiated from the THP-1-JAK1 KD and Sc cell lines were infected with BCG, with or without prior IFN-γ stimulation. JAK1 KD and Sc THP-1 cells were capable of internalizing BCG, as seen by confocal microscopy ( Figure 2A ). Using confocal analysis and a lysotracker dye which increases fluorescent intensity in low pH (33), both Sc and JAK1 KD cell lines were observed to traffic a proportion of internalized BCG into acidified compartments ( Figure S4A ).

To better quantitate BCG infection, THP-1 macrophages were co-cultured with mCherry-expressing BCG and analyzed by flow cytometry. Similar levels of bacteria were internalized by KD and Sc lines at 4 hours, and this was largely unaffected by IFN-γ stimulation ( Figures 2B, C ), indicating that loss of JAK1 does not significantly impact phagocytosis. As expected, at both 24 and 72 hours, IFN-γ stimulation significantly reduced mCherry fluorescence in Sc lines consistent with lower bacterial survival. In contrast, IFN-γ had no significant impact on mCherry levels in KD lines ( Figures 2B, C ). To confirm that JAK1 deficiency promotes intracellular BCG survival, KD and Sc lines were lysed on culture plates 3 days after infection and BCG titers were quantitated by counting colony forming units (CFU/ml). As seen in flow cytometry assays, BCG survival was higher in KD lines indicating an important role for JAK1 in controlling mycobacterial infection ( Figure 2D ). Similar findings were obtained with Salmonella typhimurium ( Figure 2E ), another intracellular pathogen known to require IFN-γ for control of the bacterial infection, with significantly higher bacterial survival seen in KD than Sc cells. Together these results demonstrate that JAK1-deficient myeloid cells permit enhanced intracellular mycobacterial and salmonella survival in vitro.

To further explore the mechanisms promoting enhanced bacterial intracellular survival in myeloid cells with reduced JAK1 function, phagosome acidification and apoptosis were tested as these are key IFN-γ-dependent steps in the control of mycobacterial infection (34–39). Following infection of THP-1-derived macrophages with pHrodo-labelled BCG, phagosomal acidification was measured by measuring fluorescence, which is released in the context of low pH. Even in the absence of IFN-γ stimulation, both Sc and KD THP-1 cells had relatively high levels of pHrodo fluorescence, which relates to the high baseline lysosomal content in PMA differentiated THP-1 macrophages (40) ( Figure 3A and Figure S4B ). Fluorescence intensity was increased after IFN-γ-stimulation in Sc cells lines consistent with additional IFN-γ-mediated induction of acidification ( Figures 3A, B ). In contrast there was no increase in acidification in the KD cell line following IFN-γ stimulation.

To test whether partial JAK1 deficiency is sufficient to impair IFN-γ-induced apoptosis, Annexin V/PI staining was measured by flow cytometry. Sc control and KD THP-1 cells had similar baseline levels of apoptosis which was not significantly increased 5 days after BCG infection alone ( Figures 3C, D ). In contrast, IFN-γ pre-treatment induced significant apoptosis at both 3 and 5 days after BCG infection in Sc control cells compared with untreated Sc control cells, an effect that was abrogated in KD cells ( Figures 3C, E and Figure S5 ). Together our data suggest that defective intracellular bacterial killing in myeloid cells with reduced JAK1 function is at least in part due to impaired IFN-γ-induced phagosome maturation and apoptosis.

To test viral susceptibility in JAK1 deficient cells, we utilized established viral infection models in both fibroblast and EBV-B cells (24, 41, 42). We have previously demonstrated reduced STAT1 phosphorylation following IFN-α stimulation in primary dermal fibroblasts from the patient with AR partial JAK1 deficiency (16). We also observed reduced STAT1 phosphorylation following IFN-α in EBV-B cells from the same patient with AR partial JAK1 deficiency, albeit to a lesser degree than that seen in IFNAR-deficiency ( Figures S6A, B ). For fibroblast infections, we used Parinfluenza virus 5 (PIV5) and highly attenuated recombinant strains of PIV5 (PIV5VΔC) that lack defined functional IFN antagonists (33, 34). This virus is weakly virulent forming only pinpoint plaques in cells that produce and respond to IFN but with ability to form large plaques if the IFN system is impaired. As previously shown (41), fibroblast monolayers from patients with complete STAT2 deficiency supported the formation of large plaques (infected cells) of PIV5 and PIV5VΔC, demonstrating uncontrolled viral infection resulting from failure of the IFN-α response ( Figure 4A ). Fibroblast monolayers from heathy control and the patient with partial JAK1 deficiency prevented large viral plaque formation indicating successful viral control ( Figure 4A ). This assay relies on production of endogenous IFN-α by infected fibroblasts and therefore reflects cellular response to physiological concentrations of IFN.

We also tested whether partial JAK1 deficiency altered the capacity of fibroblasts to respond to exogenous addition of IFN-α to control PIV5 infection. Using immunostaining to visualize intracellular virus, loss of viral fluorescence was seen during successful suppression of viral infection in healthy control fibroblasts treated with IFN-α ( Figure 4B ). Comparable viral suppression was mediated by patient fibroblasts after IFN-α stimulation suggesting partly preserved IFN-α responses in fibroblasts with reduced JAK1-function ( Figure 4B ). Using a separate model, fibroblasts from healthy controls and patient were infected with herpes simplex virus-1 (HSV-1), showing control of viral infection 24h after treatment with exogenous IFN-α ( Figure S7 ).

To further test the impact of partial JAK1 deficiency on host viral protection we used a separate model in which EBV-B cells are infected with VSV. As expected, healthy control EBV-B cells controlled viral infection when treated with exogenous IFN-α, evidenced by lower viral titers compared with untreated cells at 24h and 48h after infection ( Figure 4C ). In contrast, EBV-B cells from the patient with AR partial JAK1-deficiency exhibited no reduction in viral titers in the presence of IFN-α indicating a significantly reduced response to IFN-α. In this assay, VSV titers in infected patient EBV-B cells were comparable to STAT1-deficient EBV B-cells after 24h and 48h of infection ( Figure 4C ). Together these results suggest that partial JAK1 deficiency results in impaired viral protection with variable impact according to the cell type involved.