The STING HAQ haplotype and clinical non-penetrance in COPA syndrome

The field of autoinflammatory disorders has expanded considerably over the past 25 years. Alongside the discovery of novel genetic variants resulting in the inappropriate activation, or defective negative regulation, of inflammatory molecules, striking examples of clinical non-penetrance have been recorded, including among the inherited type I interferonopathies—autoinflammatory diseases characterized by constitutive activation of the type I IFN (IFN-I) pathway. For example, in a study of 74 individuals harboring pathogenic heterozygous gain-of-function mutations in IFIH1, 13.5% (seven of whom were aged over 50 years) were clinically asymptomatic (Rice et al., 2020). More recently, asymptomatic individuals homozygous for the most common mutation observed in symptomatic patients with RNASEH2B-related Aicardi–Goutières syndrome were reported, thereby highlighting the possibility of clinical non-penetrance also in the context of autosomal recessive disease (Crow, 2023).

COPA syndrome is a type I interferonopathy caused by autosomal dominant mutations in COPA (Watkin et al., 2015). The pathogenesis of COPA syndrome is directly linked to STING function, as COPA mutations result in a defect of the retrograde transport of STING (Deng et al., 2020; Lepelley et al., 2020; Mukai et al., 2021; Steiner et al., 2022). Where COPA function is disturbed, STING—an innate immune protein essential to the induction of an IFN-I response following the sensing of cytosolic DNA—remains activated in the Golgi, thereby mediating chronic IFN-I signalling. The frequency of clinical non-penetrance among individuals heterozygous for bona fide pathogenic mutations in COPA is remarkably high, having been previously estimated at 15–20% (Lepelley et al., 2020; Simchoni et al., 2023; David et al., 2026). Noting both that STING haplotypes can influence the potency of downstream IFN signalling (Yi et al., 2013) and the pivotal role of STING in the pathogenesis of COPA syndrome, Simchoni et al. recently evaluated the potential effect of STING haplotypes in the clinical penetrance/non-penetrance of COPA syndrome (Simchoni et al., 2025). In doing so, they found the common STING HAQ haplotype, combining three non-synonymous substitutions i.e., R71H, G230A and R293Q, to be present in all nine clinically asymptomatic individuals heterozygous for a COPA mutation (either R233H, A239P, E241K, V242D, or D243G) assessed in their cohort. Conversely, the STING HAQ haplotype was absent in all 26 individuals manifesting clinical disease that they tested. Experimentally, they further demonstrated that HAQ STING acts dominantly to dampen COPA-dependent STING activation. Given the potential importance of these data for clinical management, we decided to investigate the STING HAQ haplotype status of a separate cohort of individuals segregating pathogenic mutations in COPA.

We ascertained 23 individuals from 13 European families harboring an experimentally validated pathogenic mutation in COPA (Watkin et al., 2015; Bader-Meunier et al., 2021; Kechiche et al., 2024; David et al., 2026), for whom STING haplotype data from next-generation or Sanger sequencing were also available (Fig. 1, A and B). 15 individuals (65%) were assessed as clinically symptomatic, 14 of whom manifest classical features of COPA syndrome (lung: 87%, joint: 71%; kidney: 14%) detailed in Table 1 with a median age at onset of 4 years (range 0–14 years), and one who was diagnosed with isolated renal disease at age 50 years (see below) (Fig. 1 C). In contrast, seven individuals (29%) were clinically asymptomatic, with no features of COPA syndrome following clinical assessment (median age 42 years, range 30–50 years) (Fig. 1 C). Clinical non-penetrance in these seven individuals was confirmed by physical evaluation, and in five cases (not including F3.PIII.3 and F3.PIII.7) by investigations, including chest computed tomography (CT) scan, pulmonary function assessment, and testing for proteinuria. Consistent with these demographic and clinical data, there was a clear distinction in the level of IFN-I signalling recorded in the whole blood between the 15 symptomatic patients versus the seven asymptomatic individuals (Fig. 1 D). The clinical status of one individual (F3.IV.2), manifesting vitiligo in the absence of other disease features at age 18 years, was considered uncertain, and this HAQ-positive individual was not included in our further analysis.

Two asymptomatic individuals ascertained in our cohort, most recently assessed at the ages of 46 and 50 years, were positive for the HAQ haplotype. Notably, however, in contrast to the data of Simchoni et al. five clinically asymptomatic individuals, aged 30, 39, 39, 42, and 43 years at last evaluation, did not carry the HAQ haplotype. Further, the HAQ haplotype was present in a male (F3.II.4) diagnosed with lupus-like membranous glomerulonephritis during evaluation for intrafamilial kidney donation to his daughter. Although idiopathic membranous glomerulonephritis can occur in the general population, most typically in males over the age of 60 years, the absence of anti-PLA2R antibodies in both serum and kidney biopsy samples (Beck et al., 2009), along with full-house immunofluorescence, suggests the involvement of the COPA mutation in the development of the disease in this individual. Additionally, absence of lung involvement has been previously reported in COPA patients (F7.1 [Bader-Meunier et al., 2021] and F5.II.1 in this cohort), and the IFN score, assessed once in this patient, was elevated to a level comparable with that of other symptomatic patients with a COPA mutation (Fig. 1 D). During his most recent evaluation, at the age of 61 years, while clinically well, the patient remains dependent on hydroxychloroquine and anti-proteinuria treatment to maintain his renal function.

As new genes associated with inborn errors of immunity (IEIs) have been discovered, and the use of high-throughput sequencing techniques has become more widely available, the number of mutant genotypes associated with markedly variable clinical expression and, in some cases, frank clinical non-penetrance has increased (Gruber and Bogunovic, 2020; Kingdom and Wright, 2022). In the context of the IEIs, well-known examples include genes underlying primary immunodeficiencies (CTLA4) and autoinflammation (IFIH1, RNASEH2B, COPA, JAK1, NLRP3, and TNFRSF1A). Broadly speaking, this phenomenon is most likely explained by additional genetic factors, either protective or susceptibility alleles, as well as epigenetic modifiers and environmental triggers such as infections.

STING signalling is central to host immune defense but also to other biological processes such as cellular senescence, autophagy, and anti-tumor activity, reflecting the fact that STING evolved prior to the development of IFN-I signalling pathways. The STING HAQ haplotype shows a higher prevalence in East Asian populations compared with Europeans and sub-Saharan Africans, suggesting positive selection during migration out of Africa 50,000–70,000 years ago (Yi et al., 2013; Patel et al., 2017). From an immunological standpoint, evaluations of functional differences between STING alleles have been contradictory, with the relative activity of the 232R and HAQ haplotypes highly variable between reports (Koide et al., 2025). In vitro, Simchoni et al. showed that HAQ STING acts dominantly to dampen COPA-dependent STING activation and thus might be protective against the risk of developing COPA syndrome. Their data were remarkable in suggesting a binary distinction between clinical disease and clinical non-penetrance based on HAQ haplotype status, with all nine asymptomatic individuals heterozygous for a COPA mutation also being HAQ positive and an absence of the STING HAQ haplotype in all 26 individuals manifesting clinical disease. In contrast, in our cohort of European patients, we ascertained five clinically asymptomatic individuals harboring a pathogenic COPA mutation in the absence of the STING HAQ haplotype. This observation indicates that clinical penetrance in COPA syndrome involves factors beyond HAQ haplotype status (and which might differ between populations). The identification of an HAQ-positive male with isolated kidney disease starting later in life, which we consider most likely attributable to the recurrent R233H mutation in COPA, remains difficult to interpret. This finding raises three possibilities: (1) despite the presence of an HAQ allele, disease expression can occur; (2) the HAQ allele is protective against “complete” disease expression but can sometimes be associated with an “attenuated” phenotype; (3) the HAQ allele is completely protective against disease expression (and this patient’s phenotype is unrelated to his mutation status).

Considering other additive or protective genetic factors explaining penetrance in COPA syndrome, we found the frequency of the 232R allele to be the same in symptomatic and asymptomatic individuals (8/15 [53%] and 4/7 [57%], respectively). We also looked for variants in an in-house panel of ∼500 other IFN-related genes in whole-genome sequencing data available for 10 COPA mutations carriers, and no potential candidates emerged that were shared between asymptomatic or symptomatic individuals (data not shown). In addition, no obvious environmental or infectious factors were reported in our cohort (although, given the retrospective nature of our data set, additional evaluation of previous viral exposures is warranted). Of note, STING also serves as an innate immune receptor for bacterial cyclic dinucleotides (Ablasser and Chen, 2019). As such, STING status might modulate the effect of microbiota at (alveolar) epithelial barriers and play a role in the clinical variability seen in STING-mediated inflammation. Further, although not assessed systematically, in the one asymptomatic carrier of a COPA mutation tested, we found no neutralizing anti-IFN-I auto-antibodies which might explain an absence of clinical penetrance. Finally, while recent evidence has highlighted monoallelic expression as an explanation for clinical non-penetrance in at least 4% of IEI associated genes, COPA did not show monoallelic expression in a clonally expanded T cell model (Stewart et al., 2025). Other potential explanations for clinical non-penetrance such as somatic mosaicism and mutation reversion have not been reported in COPA syndrome.

Summarizing, our findings challenge the suggestion that STING haplotype status is the sole determinant of clinical penetrance in COPA as highlighted by the identification of five asymptomatic individuals not carrying the HAQ haplotype. While we cannot rule out the possibility of later onset disease in the clinically asymptomatic HAQ-negative individuals that we ascertained, disease penetrance has been previously estimated to be 89% by age 12 years, and in the group of affected individuals reported by Simchoni et al. in 2025, 83% of patients were symptomatic by 5 years of age, with 100% manifesting disease by age 18 years (Simchoni et al., 2023; Simchoni et al., 2025). However, given that the number of patients with COPA syndrome reported in the literature remains low, the full spectrum of clinical disease is likely yet undefined, a point possibly illustrated by the HAQ-positive male with apparently isolated renal disease that we describe here. Taken together, the recent findings of Simchoni et al. and the results that we present highlight the need to establish international consortia to study the phenomenon of clinical penetrance in COPA syndrome (and other rare autoinflammatory diseases), so as to gather sufficient numbers of patients to define the underlying mechanisms in further clinical and functional studies.

Patients with genetically confirmed COPA syndrome were included in this retrospective study. Patients were recruited from referent rare diseases centers in France and the United Kingdom. COPA variants were annotated according to HGVS nomenclature using the MANE select transcript NM_004371.4. STING haplotype was determined by next-generation or Sanger sequencing and annotated according to HGVS nomenclature using the MANE select transcript NM_198282.4. Data sharing was consistent with the requirements of the relevant local ethics committee for all patients.

Data collected included sex, clinical manifestations at presentation and during follow-up, and age at genetic diagnosis of COPA syndrome. Results of chest x-ray, chest CT, and pulmonary function tests were recorded. ILD was defined according to CT scan findings, and alveolar hemorrhage was determined according to CT scan and/or bronchoalveolar lavage and/or lung biopsy. The presence of honeycombing and/or traction bronchiectasis and/or inter and intralobular septal thickening radiologically defined pulmonary fibrosis. Biological parameters such as inflammatory markers, autoantibodies, and immunological status were collected, as well as treatment characteristics and the response to therapy.

Status of IFN biomarkers was determined by studying the expression of ISGs (by qPCR [Rice et al., 2013] or by NanoString [Kim et al., 2018]) in peripheral blood.

The study was approved by the Comité de Protection des Personnes (ID-RCB/EUDRACT: 2014-A01017-40; revalidated in 2022 and 2025). Written informed consent was obtained for all patients.

Analyses were performed using PRISM software (version 10, GraphPad Inc.). A P value <0.05 was considered significant.

Data underlying Fig. 1 are available in the published article. To maintain confidentiality, sequence data relating to individual patients are not available.

The authors thank all the patients and their families for their participation in this study.

C. David is supported by the Fondation Pour la Recherche Médicale (grant FDM202106013329) and Fondation Bettencourt-Schueller “Poste CCA-Bettencourt”. T. Wauquier is supported by the program “M2-MD” of Imagine Institute. The project was supported by the grant “Rolando Cimaz” from Francophone Society for Pediatric Rheumatology and Inflammatory Diseases (SOFREMIP) for whole-genome sequencing. Y.J. Crow acknowledges the European Research Council (786142-E-T1IFNs), a state subsidy managed by the National Research Agency (France) under the Investments for the Future program (ANR-10-IAHU-01), and a UK Medical Research Council Human Genetics Unit core grant (MC_UU_00035/11). P. Bastard was supported by the French Foundation for Medical Research (EA20170638020), the MD-PhD program of the Imagine Institute (with the support of the Fondation Bettencourt-Schueller), and a “Poste CCA-INSERM-Bettencourt” (with the support of the Fondation Bettencourt-Schueller).

Author contributions: Clémence David: conceptualization, data curation, formal analysis, investigation, methodology, resources, validation, visualization, and writing—original draft, review, and editing. Tifenn Wauquier: visualization and writing—original draft. Alix de Becdelièvre: investigation, resources, and writing—review and editing. Camille Louvrier: investigation, visualization, and writing—review and editing. Maud Tusseau: investigation. Cécile Masson: formal analysis. Luis Seabra: data curation, investigation, and methodology. Caroline Kannengiesser: visualization and writing—review and editing. Hayssam Al Arab: investigation and resources. Ibrahima Ba: resources and validation. Mary Brennan: resources. Alexandre Belot: data curation, formal analysis, and writing—review and editing. Nadia Nathan: investigation, resources, supervision, validation, visualization, and writing—review and editing. Hélène Maillard: resources and writing—review and editing. Héloïse Reumaux: data curation. Jérémie Sellam: conceptualization, resources, visualization, and writing—review and editing. Jacques Cadranel: formal analysis, validation, visualization, and writing—review and editing. Yves Hatchuel: investigation. Laurence Weiss: resources. Sébastien de Almeida: resources. Cinthia Rames: resources. Marie Wislez: resources and writing—review and editing. Clémentine Vigier: resources and writing—review and editing. Géraldine Labouret: resources. Claire Kastner: writing—review and editing. François Provot: validation. Julien Tarabeux: investigation. Elise Schaefer: investigation. Darragh Duffy: data curation and investigation. Vincent Bondet: investigation, methodology, resources, and writing—review and editing. Paul Bastard: resources and writing—review and editing. Anne Puel: investigation and writing—review and editing. Jean-Laurent Casanova: resources, supervision, and writing—review and editing. Gillian I. Rice: formal analysis, investigation, resources, and writing—review and editing. Brigitte Bader-Meunier: writing—review and editing. Yanick J. Crow: conceptualization, funding acquisition, methodology, supervision, writing—original draft, review, and editing. Alice Lepelley: investigation and writing—review and editing. Marie-Louise Frémond: conceptualization, funding acquisition, investigation, methodology, resources, supervision, visualization, and writing—original draft, review, and editing.