Lupus-like autoimmunity and increased interferon response in patients with STAT3-deficient hyper-IgE syndrome

Capsule Summary:

We report on patients with STAT3 loss-of-function (LOF) mutations who developed lupus-like autoimmunity. Immune dysregulation seen in STAT3 LOF patients suggests a susceptibility to systemic autoimmunity with important implications in monitoring and management of these patients.

To the Editor:

Autosomal dominant hyper-IgE syndrome (AD-HIES) is a rare primary immunodeficiency caused by dominant negative loss-of-function (LOF) mutations in signal transducer and activator of transcription 3 (STAT3) [1]. STAT3 LOF patients have recurrent skin and lung infections, dermatitis, elevated serum IgE, and vascular, connective tissue, and skeletal abnormalities. Autoimmunity and auto-inflammation are frequent complications of other primary immunodeficiencies and have also been observed in patients with heterozygous gain-of-function mutations in STAT3 (ADMIO1) [2,3]. However, the relationship between STAT3 LOF and systemic autoimmunity remains unclear. Here, we describe a subset of STAT3 LOF patients that developed systemic lupus erythematosus (SLE) or SLE-like manifestations. SLE is a complex autoimmune disease characterized by aberrant type I interferon (IFN) responses with increased expression of IFN-stimulated genes (ISGs) in blood and affected tissues [4]. Neutrophil dysregulation is also implicated in SLE pathogenesis including extrusion of chromatin bound to granular proteins through neutrophil extracellular traps (NETs) [5]. Initially describedas an antimicrobial mechanism, NETs are now recognized as an important source of lupus autoantigens (e.g. double-stranded DNA, histones) and have proinflammatory effects in autoimmune and renal disorders [5].

We performed a retrospective medical chart review of the well characterized STAT3 LOF cohort at the NIH Clinical Center (ClinicalTrials.gov identifier NCT00006150 ) to identify patients with clinical and laboratory manifestations of SLE. Patients were classified as SLE if they met at least 4 of the 1997 revised American College of Rheumatology (ACR) classification criteria for SLE, whereas patients who only met 2–3 criteria were classified as SLE-like. We identified 10 out of 158 STAT3 LOF patients with SLE or SLE-like manifestations (STAT3 LOF + SLE, Table 1; ~6.8% prevalence in this cohort, compared to ~0.23% prevalence in the general population). Development of autoimmunity was not associated with HIES score, peak absolute eosinophil count, or peak IgE (Figure S1). Of these 10 patients, 5 were classified as SLE and another 5 had SLE-like features. For those with SLE, age of presentation was 8–22 years and 4 were female (including 2 patients who are fraternal twins and have been previously reported [6]). For those with SLE-like disease, age of presentation was 12–24 years and 3 were female. Clinical features in STAT3 LOF patients with SLE and SLE-like disease included nephritis, autoimmune cytopenias, discoid rash, and alopecia (Table 1). All patients had anti-nuclear IgG autoantibodies, and a majority also had anti-double stranded DNA IgG autoantibodies (90%) and decreased C3 and/or C4 (70%). Notably, 6 of the 10 patients developed nephritis (predominantly immune complex glomerulonephritis). One patient progressed to end stage kidney disease despite immunosuppressive therapy and underwent kidney transplantation. Patients were treated with a combination of immunosuppressants such as steroids, hydroxychloroquine, and/or mycophenolate, which were overall well-tolerated and led to disease remission without an increase in infectious complications.

Table 1.

Clinical characteristics of STAT3 LOF patients with lupus and lupus-like autoimmunity.

PATIENT	SEX	AGE AT SYMPTOM ONSET	AGE AT ANALYSIS	RACE	STAT3 MUTATION	CLINICAL	RENAL	HEMATOLOGIC	ANA	anti-dsDNA	C3/C4	SEROLOGIES	ACR 1997	SLICC 2012	ACR 2017	DIAGNOSIS	MEDICATIONS	PEAK IgE	PEAK AEC	HIES SCORE
1	F	8	22	W	K658E	Alopecia, arthritis	Class III and V nephritis (immune complex glomerulonephritis with subendothelial, intramembranous, subepithelial deposits, and tubuloreticular inclusions)	ITP	+	+	Low	Negative anti-ENA, RF, and anti-CCP	+	+	+	Lupus	IVIG, prednisone, hydroxychloroquine, methotrexate	3,343	340	73
2	F	15	39	W	R382W	Discoid lupus, alopecia	No	Leukopenia	+	+	Low	Negative ANCA, SSA, SSB, anti-Scl70, anti-Jo-1. Positive RF, anti-Sm, anti-RNP, ACA IgM, and ACA IgA	+	+	+	Lupus	Methotrexate, cyclophosphamide, hydroxychloroquine, IVIG, rituximab	3,880	325	58
3	F	21	28	AA	I653del	Arthritis	Class IV and V nephritis progressing to ESRD and renal transplant (immune complex glomerulonephritis with predominantly mesangial deposits, segmental endocapillary proliferation with small fibrocellular crescents, and moderate chronic tubulointerstitial injury)	No	+	+	Low	Negative RF and anti-CCP. Positive anti-ENA, SSA, and anti-SmRNP	+	+	+	Lupus	Mycophenolate, prednisone, tacrolimus, and hydroxychloroquine	5,121	330	54
4	F	21	28	AA	I653del	Arthritis, discoid lupus	Class III nephritis (immune complex glomerulonephritis with mild interstitial fibrosis and mesangial, subendothelial, and subepithelial deposits)	No	+	+	Low	Negative ENA, RF, and anti-CCP	+	+	+	Lupus	Mycophenolate, prednisone, hydroxychloroquine	7,969	1150	48
5	M	10	12	AA and W	R382W	Possible Jaccoud’s arthropathy	Microscopic hematuria	AIHA	+	+	Low	Negative, anti-ENA, RF, and anti-CCP	+	+	+	Lupus	IVIG, prednisone, hydroxychloroquine	2,936	730	67
6	M	12	15	PI and L	R417G	No	No	AIHA	+	−	Low	Negative anti-ENA	−	+	−	Lupus-like	IVIG, prednisone, rituximab	14,763	870	53
7	F	24	25	W	R382W	No	C3 glomerulopathy (subepithelial deposits and 56% global sclerosis)	No	+	+	Normal	Negative anti-ENA, RF, anti-CCP, anti-MPO, anti-PR3, and anti-centromere	−	−	+	Lupus-like	Mycophenolate, hydroxychloroquine	67,122	1690	63
8	M	22	34	W	R382W	Alopecia	No	No	+	+	Normal	Negative for anti-ENA, anti-TPO, and anti-thyroglobulin	−	−	−	Lupus-like	None	8,338	680	75
9	F	20	23	W	V463E	No	IgA nephropathy (biopsy findings not available)	No	+	+	Normal	Negative anti-ENA, anti-PR3, anti-MPO, and anti-centromere	−	−	+	Lupus-like	None	1,703	270	69
10	F	11	13	W	V637M	No	No	Leukopenia, ITP	+	+	Low	Positive anti-ENA and anti-Sm	−	+	+	Lupus-like	Hydroxychloroquine, IVIG	4,845	260	55

Patients (n=10) with STAT3 LOF mutations were identified to have clinical and serologic features of lupus. Demographic information, disease manifestations, clinical laboratory data, and historic treatment protocols are shown. Diagnosis of lupus (clear rows) and lupus-like (shaded rows) disease was made on the basis of fulfilling American College of Rheumatology (ACR) 1997 classification criteria. Updated classification criteria for systemic lupus, Systemic Lupus International Collaborating Clinics (SLICC) 2012 and ACR 2017, are also shown for comparison. Abbreviations are the following: N/A= not available, W= White, AA= African-American, L= Latino, PI= Pacific Islander, ESRD= end-stage renal disease, ITP= idiopathic thrombocytopenic purpura, AIHA= autoimmune hemolytic anemia, IVIG= intravenous immunoglobulin, ANA=anti-nuclear antibody, anti-dsDNA=anti-double stranded DNA antibody, AEC=absolute eosinophil count, ENA= extractable nuclear antigen, CCP= cyclic citrullinated peptide, PR3= proteinase 3, TPO= thyroid peroxidase, RF= rheumatoid factor, ANCA= anti-neutrophil cytoplasmic antibody, MPO= myeloperoxidase, ACA= anti-cardiolipin antibody, RNP= ribonuclear protein, anti-SSA = anti-Sjögren’s syndrome-related antigen A, anti-SSB = anti-Sjögren’s syndrome-related antigen B, IVIG= intravenous immunoglobulin. ¹Age at presentation; ² Delineation of lupus and non-lupus symptoms is unclear at this time.

We further characterized the autoimmunity associated with STAT3 LOF mutations in patient neutrophils and peripheral blood mononuclear cells (PBMCs). Unexpectedly, we found that all STAT3 LOF patients, regardless of whether they had SLE features or not, had increased expression of ISGs, including HERC5, IFI44L, MX1, RSAD2, IFI27 and CXCL10 (Figure 1A, S2). No significant differences were observed in ISGs between STAT3 LOF patients with or without SLE (Figure 1A, S2). IFN-α/β were undetectable in sera, suggesting the dysregulated type I IFN response may be related to the underlying STAT3 mutations (Figure S3). However, IFN cytokines are difficult to measure in serum and we cannot rule out that samples were below limit of detection of the assays used.

Figure 1. Increased interferon response, neutrophil extracellular trap (NET) formation, and anti-NET antibodies in patients with STAT3 LOF.

(A,B) RT-PCR for ISG expression and STAT/SOCS ratio in neutrophils and PBMCs from healthy controls (n=6), STAT3 LOF (n=7), and STAT3 LOF + SLE (n=4) patients. (C) Immunofluorescence for spontaneous NET formation at 3 hours in healthy controls (n=8), STAT3 LOF (n=7), and STAT3 LOF + SLE (n=5) patients; Myeloperoxidase (MPO, red), DNA (Hoechst, blue). (D) NETs were further quantified as percent of neutrophils forming extracellular traps. (E) Anti-NET IgG antibodies measured in serum by ELISA in STAT3 LOF (n=8), STAT3 LOF + SLE (n=5), and SLE (n=10) patients. Data were normalized to healthy controls (n=8). Statistics were calculated by non-parametric Mann-Whitney test. *p<0.05, **p<0.01, ***p<0.001, ns=not significant.

Notably, a previous report indicated that patients with STAT3 LOF mutations have corresponding upregulation of the STAT1 pathway due to impaired induction of suppressor of cytokine signaling 3 (SOCS3) [7]. As STAT1 is the major transcription factor involved in expression of ISGs, it is possible that the increased IFN response observed in STAT3 LOF is due to loss of negative regulation by the STAT3-SOCS3 axis. Supporting this concept, we found that STAT3 LOF patients had a significantly increased STAT1/SOCS3 ratio compared to controls (Figure 1B, S4). No differences were observed in STAT3/SOCS ratios (Figure 1B).

In addition to type I IFN responses, we further investigated neutrophil activity and dysregulation. Quantification of NETs by immunofluorescence demonstrated that neutrophils from STAT3 LOF patients formed more spontaneous NETs than neutrophils from matched healthy controls (Figure 1C–D). A similar trend was observed in neutrophils from STAT3 LOF patients with SLE features, but was not statistically significant, possibly due to their ongoing treatment with immunosuppressants. NETs are highly interferogenic and may trigger IFN production in STAT3 LOF patients, providing a second mechanism that could explain the observed increase in ISGs. Previous studies have also demonstrated that anti-NET antibodies can form immune complexes with NET proteins that deposit in tissues, leading to complement activation that promotes inflammation in various organs including the kidneys [8]. We found that both STAT3 LOF and STAT3 LOF + SLE patients developed higher levels of anti-NET IgG autoantibodies than healthy controls (Figure 1E, S5). As expected, SLE only patients also had increased levels of anti-NET IgG. There were no significant differences in anti-NET IgE (Figure S6).

In summary, we have identified a cohort of patients with STAT3 LOF and coinciding SLE-like autoimmunity with predominant renal involvement. STAT3 LOF patients have increased expression of ISGs, NET formation, and anti-NET autoantibodies, even in the absence of clinical autoimmunity. Previous work has shown that a subset of patients with heterozygous STAT3 GOF mutations develop early-onset autoimmunity [3]. Moreover, STAT3 upregulation has been implicated in SLE and there has been interest in targeting STAT3 to treat lupus nephritis in mouse models [9]. In contrast, our data indicate that STAT3 LOF is also associated with profound dysregulation of the immune system that may predispose towards autoimmune disease, and that STAT3 may have broad contributions to immune homeostasis. Therefore, we recommend screening for SLE-like disease in STAT3 LOF patients to minimize end-organ complications. Given the effects of STAT3 LOF mutations on type I IFN responses and neutrophil regulation, targeted therapies (like JAK inhibition) may be considered for patients with STAT3 LOF and SLE.