Human germline biallelic loss-of-function OSMR variants cause severe allergic disease

Inborn errors of immunity (IEIs) are a group of monogenic disorders that impair or dysregulate the function of the human immune system (1, 2). Primary atopic disorders (PADs) are a subset of the IEIs where severe allergic disease is the predominant clinical feature (3). The discovery of new PADs advances our understanding of the molecular mechanisms responsible for allergic inflammation and can transform the clinical care of affected individuals (3, 4, 5).

In this study, we describe a novel human PAD caused by germline biallelic loss-of-function (LOF) variants in the gene OSMR found in 10 patients from seven unrelated families spanning three continents. Oncostatin M (OSM) receptor beta (OSMRβ), encoded by OSMR (Online Mendelian Inheritance in Man [OMIM]: 601743), is a member of the interleukin (IL)-6 superfamily. OSMRβ functions as a shared receptor component in two cytokine signaling pathways, pairing with GP130 to form the OSM type II receptor and with IL-31Rα to form the IL-31 receptor, thereby mediating the biological effects of both OSM (OMIM: 165095) and IL-31 (OMIM: 609509). OSM is a monomeric glycoprotein produced by activated T cells, monocytes (6), macrophages (7), and neutrophils (8, 9), whereas IL-31 is mainly produced by CD4⁺ T helper cells (10). Human biallelic OSM deficiency causes a bone marrow failure syndrome (11, 12). The expression of the IL-31 receptor is specifically high in the dorsal root ganglion and is recognized for promoting itch in prurigo nodularis (13) and other pruritic skin conditions, while OSMRβ is widely expressed in a variety of cell types and tissues, including epithelial cells, fibroblasts, blood vessels, nerve cells, respiratory tissues, adipose tissues, and lymph nodes (14, 15). OSMRβ also plays a role in keratinocyte proliferation, differentiation, and inflammatory responsiveness (16). Consistent with its broad tissue expression, dysregulated OSMRβ signaling has been implicated in diverse inflammatory, fibrotic, and epithelial diseases (17), including inflammatory bowel disease (18, 19, 20), pulmonary fibrosis (21), inflammatory skin conditions such as IL-31-mediated pruritus (22), diffuse cutaneous systemic sclerosis (23), cutaneous T cell lymphoma (24), chronic autoimmune urticaria (25), and familial primary localized cutaneous amyloidosis (FPLCA) (26, 27, 28, 29, 30, 31). FPLCA is an extremely pruritic skin disorder characterized by amyloid deposits in the superficial dermis. Heterozygous variants in IL31RA and OSMR are associated with FPLCA, which is typically an autosomal dominant disorder with onset of symptoms in adolescence or adulthood (26, 27, 28, 29, 30, 31). Notably, pathogenic OSMR variants have previously been described exclusively in the context of autosomal dominant FPLCA, although a single patient with recessive OSMR deficiency presenting with systemic allergic disease was reported while this manuscript was under peer review (32).

Here we expand our understanding of the role of OSMRβ in human health and disease by describing a cohort of 10 patients with biallelic LOF variants in OSMR causing a new PAD.

We investigated 10 patients from seven kindreds with severe early-onset atopic disease that is most notable for treatment-resistant atopic dermatitis. Patients were identified by their clinicians as candidates for assessment of a monogenic disorder based on their severe phenotype (as detailed in the next section). The patients were of self-reported European (kindreds A, B, and C), South Asian (kindreds D and G), or Arab ancestry (kindreds E and F). All patients carried biallelic variants in OSMR (NM_003999.3). By sequencing their healthy parents and siblings (when available), we established an autosomal recessive pattern of inheritance.

Patient 1 (P1) (from kindred A) carried compound heterozygous missense OSMR variants (c.1046C>A, p.Ala349Asp and c.1307T>A, p.Val436Asp). The p.Ala349Asp variant was maternally inherited, while the p.Val436Asp variant was paternally inherited. Patients 2 and 3 (P2 and P3) (from kindreds B and C, respectively) were homozygous for the c.1307T>A, p.Val436Asp OSMR variant. P2 inherited the variants from their healthy heterozygous parents, while no pedigree information was available for P3. Patient 4 (P4) (from kindred D) was homozygous for the c.1979_1980delAC, p.Tyr660Serfs*16 OSMR variant. Patient 5 (P5) (from kindred E) was homozygous for the c.150dup, p.Gln51Thrfs*23 OSMR variant. Patients 6 and 7 (P6 and P7) (from kindred F) are siblings and were homozygous for the c.808C>T, p.Gln270* OSMR variant. Patients 8–10 (P8, P9, and P10) (from kindred G) are siblings and were homozygous for the c.1433del, p.Pro478Hisfs*18 OSMR variant. Patients 4–10 inherited their biallelic OSMR variants from their healthy heterozygous consanguineous parents (Fig. 1 A).

All identified variants localized to either the cytokine-binding domain or the extracellular fibronectin III domains of OSMRβ, regions required for receptor signaling. The fibronectin III domains, in particular, are essential for receptor dimerization and downstream signal transduction (33, 34) (Fig. 1 B). Both the p.Ala349Asp and p.Val436Asp variants result in the substitution of a nonpolar amino acid (alanine and valine, respectively) with a negatively charged aspartic acid at evolutionarily conserved positions. Both variants were predicted to be damaging by a variety of in silico tools (Fig. 1, B and C). The remaining OSMR variants, p.Gln51Thrfs*23, p.Gln270*, p.Pro478Hisfs*18, and p.Tyr660Serfs*16, are predicted LOF variants that introduce premature termination codons (Fig. 1 C).

The 10 patients presented with similar phenotypes consistent with an underlying PAD (Fig. 2, A–C). Specifically, they exhibited severe, widespread, treatment-resistant atopic dermatitis (Fig. 2 D), which began in their first month of life, combined with atopic blood biomarkers of peripheral blood eosinophilia (Fig. 2 B) and markedly elevated serum IgE (Fig. 2 C). Some patients also presented with failure-to-thrive (Fig. S1). P1 began treatment with dupilumab, a biologic inhibitor targeting IL-4 and IL-13 signaling, at 4.5 years of age, resulting in a significant improvement in skin disease (Fig. 2 E), and P6 and P10 have also recently initiated treatment with this therapy. Individual patient summaries are provided in Table S1. Clinical immunophenotyping of patients with OSMR deficiency was essentially unremarkable and did not reveal a consistent immunologic signature (Table S2).

To assess the functional impact of the p.Ala349Asp, p.Val436Asp, p.Gln51Thrfs*23, p.Gln270*, p.Pro478Hisfs*18, and p.Tyr660Serfs*16 OSMRβ variants, we modeled them in HEK293 cells. We selected HEK293 cells as our model system due to their low endogenous OSMRβ expression (35) (Fig. 3, A and B). As controls, in parallel, we tested wild-type (WT) OSMRβ and four OSMRβ missense variants reported in association with FPLCA (26, 27, 30, 31). WT OSMRβ and OSMRβ missense variants tagged with green-fluorescence protein (GFP) at the C terminus were overexpressed in HEK293 cells. Extracellular and total OSMRβ levels were assessed by antibody staining in the absence or presence of permeabilization, respectively, followed by flow cytometric analysis. Using this system, we demonstrated significantly decreased cell surface expression of the p.Ala349Asp and p.Val436Asp OSMRβ variant constructs compared to WT OSMRβ (P < 0.05) (Fig. 3, A and C). Cell surface expression was intermediate in cells transfected with FPLCA variants (p.Asn462Ser, p.Gly513Asp, p.Val631Leu, and p.Ile691Thr). Notably, total OSMRβ protein expression was indistinguishable between all groups, as shown by intracellular flow cytometry staining (Fig. 3, A and D).

We modified the experimental approach to assess the patient premature termination codon OSMR variants (Fig. 3, E–G). To detect potential truncated protein products, we introduced an N-terminal Myc tag into OSMR constructs carrying variants predicted to generate premature termination codons. Immunoblotting showed that the p.Gln270*, p.Pro478Hisfs*18, and p.Tyr660Serfs*16 variants produce truncated but stable OSMRβ proteins, whereas the p.Gln51Thrfs*23 variant did not yield detectable protein (Fig. 3 G). These truncated proteins lack the transmembrane domain of OSMRβ and would therefore be predicted to fail to localize to the cell surface (Fig. 1 B). However, because the p.Tyr660Serfs*16 variant occurs in exon 14, which comprises 174 nucleotides, we studied potential in-frame skipping of exon 14 and preservation of the transmembrane domain. Assessing membrane localization, flow cytometric analysis demonstrated that the p.Gln270*, p.Pro478Hisfs*18, and p.Tyr660Serfs*16 variants result in truncated OSMRβ proteins that localize predominantly to the intracellular compartment, characterized by high total OSMRβ expression and minimal extracellular OSMRβ expression, with no detectable signal from the C-terminal GFP tag (Fig. 3, B, E and F; and Fig. S2; shown for p.Tyr660Serfs*16). The truncated variants retain the N-terminal signal peptide required for endoplasmic reticulum (ER) targeting but fail to traffic to the cell surface (36), suggesting retention in the ER. Flow cytometry similarly failed to detect the p.Gln51Thrfs*23 variant, consistent with the immunoblotting results.

Together, these data support a LOF mechanism for the missense (p.Ala349Asp, p.Val436Asp) and truncating (p.Gln270*, p.Pro478Hisfs*18, and p.Tyr660Serfs*16) OSMRβ variants, whereas the p.Gln51Thrfs*23 variant results in a loss of protein expression. Despite distinct molecular lesions, all variants display markedly reduced cell surface expression of OSMRβ relative to WT and FPLCA variants, consistent with impaired receptor availability at the plasma membrane and consequent disruption of OSMRβ-dependent signaling.

To validate the findings from the in vitro HEK293 cell experiments, we quantified OSMRβ expression in primary fibroblasts from P1, P2, and P3 alongside three healthy controls (HCs); primary samples from P4 through P10 were not available. Cell surface expression was significantly lower in the primary patient fibroblasts compared to HCs (P = 0.02) (Fig. 4, A and B), whereas total OSMRβ expression was similar between patients and controls (Fig. 4, A and C). Notably, the cell surface OSMRβ expression in the healthy heterozygous mother of P1 (Family A-I-1) was comparable to the other two HCs, further reinforcing the autosomal recessive inheritance pattern of disease (Fig. 4 A).

Leveraging population-level genomic data to assess pathogenicity, we examined the p.Val436Asp variant, which is reported in gnomAD (v4.1.0) with 14 homozygous individuals, including nine from the UK Biobank (UKB). The relatively high minor allele frequency (MAF = 0.00339) and the presence of multiple homozygotes in an unselected population might ordinarily argue against a classical IEI. However, available UKB clinical data suggest enrichment of allergic disease or cutaneous features among individuals homozygous for p.Val436Asp, including elevated peripheral eosinophil counts or percentages (3/9) and documented allergic or dermatologic diagnoses in several individuals. These findings should be interpreted cautiously given the small sample size and likely incomplete clinical penetrance (Table S3). To place this observation in context, we systematically evaluated all homozygous coding OSMRβ variants reported in gnomAD with MAFs exceeding that of p.Val436Asp. In contrast to the patients’ OSMRβ variants, these higher-frequency population variants localized to the cell surface at levels comparable to WT OSMRβ and were predicted to be benign by AlphaMissense (Fig. 5, A–D), highlighting the discriminatory value of AlphaMissense predictions for normal cell surface expression of OSMRβ (37).

Given that cell surface OSMRβ expression was low in the overexpression system and on primary cells, we hypothesized that this would likely impair signaling through the OSM/OSMRβ axis (Fig. 6 A). We tested this hypothesis using primary fibroblasts from P1, P2, and P3. Activation of STAT5 after stimulation with OSM is exclusively mediated by the OSM type II receptor (OSMRβ/GP130) and was absent in patient cells. Similarly, patient cells displayed significantly decreased activation of STAT1 and STAT3 after OSM stimulation, with the residual activation likely reflecting signaling through the OSM type I receptor (LIFR/GP130) (Fig. 6, B–E). Notably, heterozygous control fibroblasts from a parent (Family A-I-1) were indistinguishable from the other HC fibroblasts, indicating that impairment of the OSM/OSMRβ axis requires biallelic LOF.

To establish the specificity of the signaling defect, we assessed signaling through other GP130-containing cytokine receptor complexes in primary patient fibroblasts. In contrast to OSM stimulation, both leukemia inhibitory factor (LIF)– and IL-27–mediated STAT activation were preserved. Stimulation with LIF, which signals through the LIFR/GP130 complex, induced comparable downstream responses in fibroblasts from patients and HCs (Fig. 6, F–H) (38). Similarly, stimulation with IL-27, which signals through the IL-27Rα/GP130 complex, induced comparable STAT1 phosphorylation in fibroblasts from patients and HCs (Fig. S3). Together, these data demonstrate that biallelic LOF variants in OSMR selectively impair OSMRβ-dependent signaling while preserving signaling through other GP130-associated receptor complexes.

To establish a causal relationship between the genotype and cellular phenotype (39), we used a lentiviral approach to re-express WT OSMRβ in primary dermal fibroblasts of P2. Lentiviral transduction rescued cell surface expression of OSMRβ (Fig. 7, A–C). Lentiviral transduction also rescued signaling through the OSM axis, as measured by pSTAT1 and pSTAT3 (Fig. 7, D–F and Fig. S4). Finally, transcriptomic analysis confirmed a profound defect in OSM-induced gene regulation in patient fibroblasts that was largely corrected by re-expression of WT OSMRβ (Fig. 7, G and H). In HC fibroblasts, OSM stimulation induced 523 significantly upregulated and 211 downregulated genes. In contrast, P2 fibroblasts transduced with empty vector (EV) showed no significant differential gene expression following OSM treatment. Re-expression of WT OSMRβ in P2 fibroblasts restored OSM responsiveness, resulting in 91 significantly upregulated genes and no downregulated genes. Notably, 85 of these 91 genes overlapped with OSM-induced genes in HC fibroblasts, demonstrating substantial rescue of the transcriptional program downstream of OSMRβ signaling (Fig. 7 H).

Differential pathway activation using gene set enrichment analysis (GSEA) showed significant enrichment in pathways for immune activation and cell proliferation in HCs after stimulation with OSM (Fig. 7 I). Consistent with established links between STAT1 activation and interferon signaling (40, 41), we found that OSM stimulation led to strong activation of interferon activation pathways (e.g., HC: Gamma_IFN normalized enrichment score 2.71, adj P value < 0.001), as well as STAT3, STAT5, and MYC pathway activation (Fig. 7 I and Table S4). These pathways were either reduced or absent in patient fibroblasts (Fig. 7 I and Table S4). Reinforcing causality, lentiviral rescue of P2 fibroblasts with WT-OSMR restored signaling in all immune-activated pathways downstream of OSMRβ. Collectively, these findings establish the role of OSMRβ in OSM-mediated immune activation of fibroblasts.

In this study, we define a new human PAD caused by germline biallelic LOF variants in OSMR in 10 patients with early-onset severe atopic disease. These variants all led to a lack of cell surface expression of OSMRβ protein and decreased OSM-mediated STAT1, STAT3, and STAT5 activation and phosphorylation. As this study includes only 10 affected individuals, defining the full clinical spectrum of this disorder and elucidating all the underlying mechanisms will require identification of additional patients and further mechanistic studies.

The identification of biallelic OSMR deficiency fills a previously unassigned disease node within the human IL-6 family cytokine signaling pathway (Fig. 8). At the most proximal level, biallelic complete LOF variants in IL6ST (GP130) disrupt signaling by multiple IL-6 family cytokines causing a severe multisystem disorder, termed extended Stüve–Wiedemann syndrome, encompassing lethal cardiopulmonary defects, skeletal abnormalities, autonomic dysfunction and additional variable clinical features such as congenital thrombocytopenia, eczematoid dermatitis, renal abnormalities, and defective acute-phase response (42). In contrast, loss of LIFR causes Stüve–Wiedemann syndrome, characterized by congenital skeletal dysplasia and autonomic instability (43). More restricted, predominantly immune phenotypes resembling STAT3-related hyper-IgE syndrome are observed in individuals with autosomal recessive hypomorphic or autosomal dominant-negative IL6ST variants (44, 45). Selective disruption of the IL-11 signaling axis further illustrates the functional specialization within this pathway, as defects in IL11RA or IL-11–selective variants in IL6ST produce predominantly craniofacial and skeletal phenotypes, including craniosynostosis and dental abnormalities (46, 47). More recently, biallelic deficiency of the OSM ligand itself has been shown to cause bone marrow failure, revealing a nonredundant role for OSM in human hematopoiesis (11, 12). In a parallel branch of this pathway, defects in IL31RA disrupt signaling through the IL-31 receptor complex and are associated with pruritic skin disease, including cutaneous amyloidosis in humans (31, 48), while experimental models further support a role for IL-31 signaling in pruritus, epithelial inflammation, and type 2 immune responses (10, 49, 50). Our findings establish OSMR LOF as a distinct receptor-level disorder that selectively impairs OSM-dependent signaling while preserving broader IL-6 family pathways, thereby defining a PAD that is mechanistically and phenotypically intermediate between ligand-level defects and disruptions of the shared co-receptor or down-stream JAK/STAT signaling. Together, these observations refine genotype–phenotype relationships across the IL-6 family and support a hierarchical model in which disease severity are shaped by both the position of the affected gene within the signaling cascade and the breadth of downstream cytokine responses that are lost.

Population genetics provide another layer of evidence supporting the involvement OSMR genetic variation in human immunity. Several independent genome-wide association studies have found that polymorphisms in OSMR associate with multiple chronic inflammatory conditions, including Crohn’s disease (51, 52, 53), ulcerative colitis (51, 52, 53), psoriasis (53), ankylosing spondylitis (53), and sclerosing cholangitis (53). Recently, microbial genome-wide interaction studies also found that interactions between single-nucleotide polymorphisms in the OSMR gene and early-life exposures, such as breastfeeding, are associated with gut microbiota alterations linked to asthma and atopy (54).

An important distinction we establish in this study is the difference between biallelic autosomal recessive OSMR deficiency and FPLCA. Whilst there is overlap between the conditions, there are key differences. The most notable are that FPLCA is typically an autosomal dominant disorder with disease limited to the skin and onset in adolescence or adulthood, while the patients we describe all had the onset of severe atopic dermatitis from infancy accompanied by other manifestations of their allergic diathesis. Notably, none of the OSMR variants identified in this study have previously been associated with FPLCA, despite multiple clinical reports implicating OSMR in this condition. In addition, heterozygous carriers of the LOF alleles identified here did not report skin disease, supporting a recessive disease mechanism distinct from the dominant inheritance of missense variants observed in FPLCA. Moreover, the amyloid deposits in the superficial dermis, which are a defining feature of FPLCA, were lacking in the skin biopsies of patients with biallelic OSMR deficiency (30). Functional data emphasize the differences between FPLCA and biallelic OSMR deficiency. We show that FPLCA variants lead to intermediate levels of cell surface expression of OSMRβ, which contrasts with the absence of cell surface expression we find with the autosomal recessive OSMR LOF variants. Our result is consistent with a previous study of OSMRβ expression by microscopy that found no differences in the localization of FPLCA-associated OSMRβ variants when compared to OSMRβ^WT (55). Ultimately, the distinct features and potential overlap between autosomal dominant FPLCA and autosomal recessive OSMR deficiency will be fully defined through the diagnosis of additional affected individuals.

Despite known species differences, the Osmr^−/− mouse accurately models the epithelial and signaling defects observed in human biallelic OSMR deficiency. In both settings, loss of OSMRβ results in impaired OSM-dependent STAT signaling and prominent abnormalities in epithelial homeostasis, with Osmr^−/− mice displaying increased epidermal thickness and our patients manifesting early-onset severe atopic dermatitis and barrier dysfunction (55). This concordance supports a conserved role for OSMRβ in postnatal skin biology across species. In contrast, systemic phenotypes reported in Osmr^−/− mice, including hematopoietic and skeletal abnormalities, were not observed in our cohort, likely reflecting species-specific ligand–receptor usage and context-dependent activation of OSM signaling pathways (38, 56, 57, 58, 59, 60). Together, these findings indicate that while murine models incompletely capture the full human phenotype, they accurately model the epithelial and signaling defects that define biallelic OSMR deficiency as a PAD.

Our data suggest that biallelic OSMR deficiency may exhibit variable expressivity and likely incomplete clinical penetrance. For example, in family G, although all three homozygous individuals were affected, the severity of their atopic dermatitis varied substantially, ranging from particularly severe to more manageable disease. Similarly, analysis of UKB data indicated that among nine individuals homozygous for OSMR c.1307T>A (p.Val436Asp), at least two had no documented abnormalities in eosinophil counts, allergic manifestations, or skin phenotypes, although phenotyping in UKB is incomplete. Independent clinical observations further support the pathogenic potential of this allele. Andersen et al. recently described a patient homozygous for OSMR c.1307T>A (p.Val436Asp) who presented with elevated IgE levels, atopic dermatitis, chronic pulmonary aspergillosis, and bone fractures (32).

To place these findings in the broader population genetic context, we systematically assessed cell surface expression of all homozygous coding OSMR variants reported in gnomAD with MAFs exceeding that of p.Val436Asp. In contrast to the patients’ OSMR variants, these higher-frequency population variants localized to the cell surface at levels comparable to WT OSMRβ and were predicted to be benign by AlphaMissense. These results indicate that homozygosity for OSMR missense variants does not uniformly result in impaired receptor surface expression and underscore the importance of integrating structural context, functional assays, and computational prediction tools such as AlphaMissense when interpreting OSMR variants.

Together, these observations suggest that biallelic OSMR LOF may represent an under-recognized genetic contributor to severe allergic disease. There is precedent for this pattern in other genetically mediated allergic disorders. For example, relatively common LOF variants in FLG impair epithelial barrier function and are strongly associated with atopic dermatitis and downstream allergic disease (61). More recently, JAK1 variants with mild-to-moderate GOF activity have been linked to relatively common presentations of inflammatory, allergic, and/or autoimmune disease (62). These examples highlight how variants with incomplete penetrance can nevertheless make substantial contributions to disease susceptibility when present in appropriate genetic contexts.

In conclusion, we identify biallelic LOF variants in OSMR in 10 patients from seven unrelated families, establishing OSMR deficiency as a new PAD and revealing an essential role for OSMRβ signaling in maintaining immune and barrier homeostasis in humans. With >50 genes now recognized to cause PADs (5), clinicians should consider inclusive next-generation sequencing approaches when evaluating patients with severe or atypical allergic disease, as timely molecular diagnosis can guide targeted therapy and improve clinical outcomes. More broadly, these findings highlight how the study of rare monogenic PADs can illuminate fundamental pathways governing allergic inflammation and immune homeostasis.

All study participants and/or their parents/guardians provided written informed consent for sequencing, data analysis, and publication of findings and images. Research study protocols were approved by The University of British Columbia Clinical Research Ethics Board (H15-0064), the NIH Institutional Review Board (NCT00852943, NCT01164241), or King Faisal Specialist Hospital and Research Centre (2241173).

Participants were included in our research study if they harbored biallelic variants in the OSMR gene and had clinical manifestations consistent with atopic disease. Participants ranged in age from infancy to 74 years of age. No participant attrition occurred during the study, and all enrolled participants were included in the analyses.

Whole-exome sequencing was performed on genomic DNA from patients as either trio-based or singleton analyses. For singleton cases, segregation of candidate OSMR variants was evaluated in available family members by Sanger sequencing. The identified OSMR variants were predicted to be damaging and segregated with disease and thus were selected for further analysis (63, 64, 65).

Data for this study were extracted under UKB application number 103789. Data processing was performed using the UKB Research Analysis Platform. Whole-exome sequencing data from 469,835 participants (data field 23158: Population-level exome OQFE variants, PLINK format – final exome release) were analyzed to identify homozygous carriers of the OSMR variant c.1307T>A, p.Val436Asp (NM_003999.3; rs34324145; SNV 5-38917567-T-A [GRCh38]). Homozygous genotypes were extracted using JupyterLab on the UKB DNANexus platform. Basic demographic information (sex, age at inclusion) and phenotype data (ICD-10–coded diagnoses, data field 41270; eosinophil count and percentage at first blood draw) were obtained using the cohort browser tool on the same platform.

Plasmids used for transfection studies contained full-length OSMR cDNA in a pCMV6-AC-GFP vector with a C-terminal GFP tag (Cat#: RG216943; OriGene Technologies). To generate OSMR LOF variants (c.1046C>A, c.1307T>A, c.1979_1980delAC, c.150dup, c.808C>T, c.1433del), homozygous population database variants (c.561T>G, c.1090T>C, c.1579G>A, c.1657G>A, c.1733G>A, c.2806C>T, c.2849C>A, c.2876C>G), and other literature-reported FPLCA variants (c.2072T>C, c.1891G>T, c.1538G>A, c.1385A>G), a Q5 site-directed mutagenesis kit (Cat#: E0554S; New England Biolabs) was used according to the manufacturer’s recommendations, with primer pairs noted in Table S5.

Plasmids used for immunoblotting transfection experiments were based on the same construct containing full-length OSMR cDNA cloned into a pCMV6-AC-GFP vector with a C-terminal GFP tag. A Myc tag was inserted between residues 27 and 28, immediately downstream of the signal peptide cleavage site (36), thereby generating an N-terminally Myc-tagged mature OSMRβ protein. The following primers were used to add the Myc tag as well as linker regions: forward primer 5′-GGC​GGC​GGC​GGC​AGC​GAG​CAG​AAA​CTC​ATC​TCA​GAA​GAG​GAT​CTG​GGC​GGC​GGC​GGC​AGC​GAA​CGT​TTA​CCA​TTG​ACT​CCT​GTA​TCA​C-3′ and reverse primer 5′-AGC​CAA​GAC​TTC​ACT​CTG​GTA​AGT​CC-3′. To generate OSMR frameshift and premature stop variants, c.150dup, c.808C>T, c.1433del, and c.1979_1980delAC, a Q5 site-directed mutagenesis kit (Cat#: E0554S; New England Biolabs) was used according to the manufacturer’s recommendations, with primer pairs noted in Table S5.

To generate lentivirus vectors, WT OSMRβ from the above plasmids was cloned into a Lenti vector with a C-terminal GFP tag (Cat#: PS100071; OriGene Technologies) using EcoRI-HF (Cat#: R3101; New England Biolabs) and NotI-HF (Cat#: R3189; New England Biolabs). The plasmids were packaged using third-generation packaging plasmids and transfected into HEK293T cells (RRID:CVCL_0063). This was also done for an empty Lenti vector, which was used as a control. Culture media was collected, centrifuged, filtered, concentrated, and stored at −80°C before use.

All OSMR variant plasmids were confirmed by Sanger sequencing and purified from 10-β competent Escherichia coli using a QIAprep Spin Miniprep Kit (Cat#: 27104; Qiagen).

Primary fibroblasts were isolated from a non-lesional skin punch biopsy (P1 and P2) or a surgical skin biopsy (P3). The skin was excised from the underlying connective tissue and placed in separate wells of a 6-well plate in DMEM (GE Healthcare) supplemented with 10% heat-inactivated FBS (Gibco, Life Technologies), 2 mM L-glutamine (HyClone, Thermo Fisher Scientific), 1 mM sodium pyruvate (Gibco, Life Technologies), and 1× Antibiotic-Antimycotic (Gibco, Thermo Fisher Scientific) for 3 wk at 37°C, or until a confluent monolayer of fibroblasts had formed. Primary fibroblasts were then lifted and frozen for future assays. Fibroblasts were Sanger sequenced to confirm the genotype of the cells before use in experiments.

Transient expression of OSMR variants in HEK293 (RRID:CVCL_0045) cells was accomplished using a Lipofectamine 3000 kit (Thermo Fisher Scientific) according to the manufacturer’s recommendation. Briefly, HEK293 cells were seeded at 8.0 × 10⁵ cells/well in a 6-well plate in 1.5 ml of DMEM supplemented with 10% FBS and incubated for 24 h at 37°C. Cells were transfected with 2.5 μg of plasmid DNA using P3000 and Lipofectamine 3000 reagents and harvested after 24 h. HEK293 cells were tested for mycoplasma contamination every 2 mo and were authenticated by short tandem repeat profiling by the manufacturer.

Stable expression of OSMRβ in primary dermal fibroblasts was accomplished using the lentivirus approach as previously described (66, 67). Briefly, primary dermal fibroblasts were infected with lentiviral particles in the presence of 5 μg/ml polybrene (Sigma-Aldrich), cultured, and expanded in DMEM supplemented with 10% FBS (Gibco, Life Technologies). Expanded cells were sorted on GFP expression using a BD FACS Aria (BD Biosciences) cell sorter.

OSMRβ expression and phospho-STAT detection were quantified using flow cytometry in transfected and primary cells. For cell surface OSMRβ expression, transfected HEK293 cells or primary fibroblasts were lifted and stained with OSMR PE (Cat#:12-1303-42; Thermo Fisher Scientific; RRID:AB_1633423) for 20 min. Expression was measured in GFP⁺ HEK293 cells or primary fibroblasts on an LSRII flow cytometer (BD Biosciences) and analyzed using FlowJo software (BD Biosciences; RRID:SCR_008520). For total OSMRβ and phospho-STAT detection, cells were fixed using BD Cytofix (Cat#: 554655; BD Biosciences) for 20 min at 4°C and permeabilized using Perm III for 30 min on ice (Cat#: 558050; BD Biosciences). The cells were then stained with OSMRβ PE, or with pSTAT1 BV421 (Cat#: 562985; BD Biosciences; RRID:AB_2737932), pSTAT3 PECF594 (Cat#: 562673; BD Biosciences; RRID:AB_2737714), and pSTAT5 PeCy7 (Cat#: 560117; BD Biosciences; RRID:AB_1645546). OSMRβ expression was measured in GFP⁺ HEK293 cells that were transfected with the p.His187Gln, p.Ala349Asp, p.Tyr364His, p.Val436Asp, p.Asn462Ser, p.Gly513Asp, p.Glu527Lys, p.Asp553Asn, p.Gly578Asp, p.Val631Leu, p.Ile691Thr, pPro936Ser, p.Ala950Glu, or p.Pro959Arg variants. OSMRβ expression was measured in PE⁺ HEK293 cells that were transfected with p.Gln51Thrfs*23, p.Gln270*, p.Pro478Hisfs*18, or p.Tyr660Serfs*16 variants (because these variants results in a truncated protein and thus the C-terminal GFP would not be translated). OSMRβ expression was measured in primary fibroblasts as well. Cells were analyzed on an LSRII flow cytometer (BD Biosciences), and flow cytometry data were analyzed using FlowJo software (BD Biosciences).

Immunoblotting was conducted as previously described (68, 69) to assess the phosphorylation status of STAT1, STAT3, and STAT5 in primary fibroblasts stimulated with OSM, LIF, or IL-27 and to quantify OSMRβ expression in HEK293 cells. Primary fibroblasts and HEK293 cells were cultured in DMEM. Cells were harvested in chilled radioimmunoprecipitation assay (RIPA) buffer (Cat# 89901; Thermo Fisher Scientific) supplemented with HALT protease and phosphatase inhibitor cocktail (Cat# 87786; Thermo Fisher Scientific) and then lysed for 15 min on ice. Cell lysates were separated by 12% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride (PVDF) membranes (Cat# IPFL00010; Immobilon-FL; MilliporeSigma). Membranes were blocked with 5% BSA in TRIS-buffered saline with Tween-20 for an hour and then incubated with primary antibodies in blocking buffer overnight at 4°C. The next day, membranes were washed and incubated with secondary antibodies for 1 h at room temperature and then imaged using a LI-COR Odyssey infrared scanner (LI-COR Biosciences).

The primary antibodies used were the following: pSTAT1 (Cat#: 9167; Cell Signaling Technologies; RRID:AB_561284), pSTAT3 (Cat#: 9138; Cell Signaling Technologies; RRID:AB_331262), pSTAT5 (Cat#: 9356; Cell Signaling Technologies; RRID:AB_331263), Myc-tag (Cat#: 2276; Cell Signaling Technologies; RRID:AB_331783), and β-actin (Cat#: 3700; Cell Signaling Technologies; RRID:AB_2242334). The secondary antibodies used were the following: goat anti-rabbit IgG DyLight 800 conjugated (Cat#: 611-145-002-0.5; Rockland Immunochemicals; RRID:AB_11183542) and goat anti-mouse IgG IRDye 680RD (Cat#: 926-68070; LI-COR; RRID:AB_10956588).

RNA was extracted, sequenced, and then preprocessed as previously described (68, 70, 71). Expression data were then normalized to reads between samples using the edgeR package in R (R Foundation; RRID:SCR_012802). Normalized counts were filtered to remove low counts using the filterByExpr function in edgeR (72). Differential expression between unstimulated and stimulated samples for healthy control fibroblasts, P2-EV fibroblasts, and P2-WT-OSMR was accomplished using Limma (73) (RRID:SCR_010943). Differentially expressed genes were defined as those with a fold change >1.25 and an Benjamini–Hochberg adjusted P value < 0.05.

Pathway analysis was done by first performing GSEA with 1,000 permutations using the Molecular Signatures Database Hallmark module. The signal-to-noise ratio was used for gene ranking and the obtained normalized enrichment scores, and P values were further adjusted using the Benjamini–Hochberg method. Pathways with an adjusted P value <0.05 were considered significant. Leading-edge genes from significant pathways between HC unstimulated fibroblasts and stimulated fibroblasts were identified. Expression levels of these genes were then determined in each of the three groups (healthy control fibroblasts, P2-EV, and P2-WT-OSMR). Sample level enrichment analysis scores were computed as previously described (74). Briefly, z-scores were computed for gene sets of interest for each sample. The mean expression levels of significant genes were compared to the expression of 1,000 random gene sets of the same size. The difference between observed and expected mean expression was then calculated and represented on heatmaps.

Fig S1 contains growth charts for P1 and P2, demonstrating failure to thrive. Fig S2 contains contour plots showing that the OSMR variants lead to a lack of OSMRβ cell surface expression in HEK293 cells. Fig S3 shows a representative immunoblot of patient and control fibroblasts stimulated with IL-27, indicating that signaling remains intact. Fig S4 contains additional data showing that WT OSMR transduction restores STAT signaling in primary fibroblasts. Table S1 summarizes the clinical characteristics of the 10 OSMR-deficient patients. Table S2 presents available clinical immunophenotyping data from five patients against published reference ranges (75). Table S3 summarizes phenotypical data of UKB listed homozygous carriers of the p.Val436Asp variant. Table S4 lists significantly enriched pathways identified in OSM-stimulated patient and healthy control fibroblasts. Table S5 provides the primer sequences used for site-directed mutagenesis of the OSMR gene.

Data are deposited at Gene Expression Omnibus (RRID:SCR_005012) with the following identifier: GSE303962.

We thank all the patients and their families for taking part in this study. We also thank Daniel MacArthur of the Broad Institute who through the creation of ExAC and gnomAD helped facilitate this collaboration.

This work was supported by grants from the Canadian Institutes of Health Research (PJT-178054 and OGB-205728; S.E. Turvey), Genome British Columbia (SIP007; S.E. Turvey), BC Children’s Hospital Foundation (S.E. Turvey), the Jeffery Modell Foundation (S.E. Turvey, J.J. Lyons), National Institute of Allergy and Infectious Diseases (4R00AI138586, J.J. Lyons), and National Institutes of Health (NIH) (DK062370, M. Boehnke). This research was supported in part by the Intramural Research Program of the NIH. The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered works of the United States government. However, the findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. S.E. Turvey holds a Tier 1 Canada Research Chair in Pediatric Precision Health, O’Sullivan Family Hospital Chair in Precision Health Research at BC Children’s Hospital, and the Aubrey J. Tingle Professor of Pediatric Immunology. S. Samra is supported by the University of British Columbia Four-Year Doctoral Fellowship. J. Körholz is supported by the Alliance4Rare (Eva-Luise und Horst Köhler Stiftung), the Else-Kröner Fresenius-Stiftung, and received a European Society for Immunodeficiencies (ESID) Research Fellowship (04/2025). O. Wegehaupt is fellow of the ESID Bridge Grant Fellowship 2025 and the Berta-Ottenstein-Program for Advanced Clinician Scientists, Faculty of Medicine, University of Freiburg, Germany.

Author contributions: Simran Samra: conceptualization, formal analysis, investigation, methodology, validation, visualization, and writing—original draft, review, and editing. Mehul Sharma: data curation, formal analysis, investigation, methodology, software, validation, visualization, and writing—original draft, review, and editing. Julia Körholz: data curation, formal analysis, investigation, methodology, validation, visualization, and writing—original draft, review, and editing. Yihui Liu: data curation, methodology, resources, and validation. Alyssa James: investigation and writing—review and editing. Christina Michalski: investigation, methodology, and writing—review and editing. Pariya Yousefi: investigation and methodology. Kate L. Del Bel: conceptualization, investigation, methodology, project administration, resources, and writing—review and editing. Henry Y. Lu: conceptualization, data curation, formal analysis, investigation, methodology, visualization, and writing—review and editing. Ashish A. Sharma: formal analysis and writing—review and editing. Maja Tarailo-Graovac: data curation and formal analysis. Joshua Dalmann: investigation. Lily Buder: investigation. Bhavi Modi: writing—review and editing. Ralf Wiedemuth: investigation. Liam Golding: writing—review and editing. Britt Drögemöller: validation and writing—review and editing. Géraldine Blanchard-Rohner: investigation and writing—review and editing. Christof Senger: formal analysis and writing—review and editing. Wingfield Rehmus: writing—review and editing. Julie S. Prendiville: resources and writing—review and editing. Massimo Mangino: data curation, formal analysis, and writing—review and editing. Colin J. Ross: conceptualization, methodology, resources, supervision, validation, and writing—review and editing. Clara D.M. van Karnebeek: conceptualization, funding acquisition, investigation, supervision, and writing—review and editing. Wyeth W. Wasserman: formal analysis, supervision, and writing—review and editing. Sergio D. Rosenzweig: investigation and writing—review and editing. Julie Niemela: data curation and methodology. Pascal M. Lavoie: resources, supervision, and writing—review and editing. P.M. Prathibha: Resources. Oliver Wegehaupt: investigation and resources. Catherine M. Biggs:funding acquisition, supervision, and writing—review and editing. Michael Boehnke: resources and writing—review and editing. Leena Kinnunen: data curation and investigation. Heikki A. Koistinen: investigation, resources, and writing—review and editing. Margaret L. McKinnon: conceptualization, investigation, and writing—review and editing. Jonas Maximillian Breuer: investigation and writing—review and editing. Jana Schonenkorb: investigation and writing—review and editing. Robert Brock: investigation and resources. Sarah Thull: investigation and writing—review and editing. Christian Netzer: investigation and writing—review and editing. Clara Velmans: investigation and writing—review and editing. Norah Altuwaijri: conceptualization and data curation. Issam R. Hamadah: conceptualization, data curation, formal analysis, funding acquisition, and writing—review and editing. Ruqaiah Altassan: resources. Ahmed Alfares: writing—review and editing. Sateesh Maddirevula: conceptualization, investigation, and resources. Siddaramappa Jagdish Patil: resources. Diana K. Bayer: conceptualization, data curation, investigation, visualization, and writing—review and editing. Jonathan J. Lyons: conceptualization, funding acquisition, methodology, supervision, validation, and writing—review and editing. Stuart E. Turvey: conceptualization, formal analysis, funding acquisition, investigation, project administration, resources, supervision, and writing—original draft, review, and editing.