Genetics, Pathophysiology, and Treatment of Autoinflammatory and Autoimmune Diseases

National Human Genome Research Institute

Investigators

Linked publications & trials

Abstract

Section I: Disorders of the Inflammasome Among the topics our research group has studied, the disorders of the inflammasome stand out as probably the area in which we have the deepest understanding. The inflammasome is a macromolecular complex that is responsible for the activation of certain key inflammatory cytokines, in particular IL-1 and IL-18, and an inflammatory form of cell death denoted pyroptosis. The genes underlying several well-known autoinflammatory disorders, including familial Mediterranean fever (FMF), the three cryopyrinopathies – familial cold autoinflammatory disease (FCAS), Muckle-Wells syndrome (MWS), and neonatal-onset multisystem inflammatory disease (NOMID), and mevalonate kinase deficiency (MKD), are known to be at least in part inflammasome disorders. In the first section, I will describe 5 inflammasome-related projects studied during the current reporting period. They are: 1) The role of cofilin-1 in the regulation of the NLRP3 inflammasome; 2) The delineation of a new disorder, CHIVE-18, caused by the pyrin inflammasome; 3) Description of a new illness caused by an upstream gene in the mevalonate pathway, HMGCS1; 4) Extension of a new illness caused by mutations in PMVK, a downstream gene in the mevalonate pathway; and 5) Description of a new illness caused by mutations in NLRC3, a regulator of the NLRP3 inflammasome. Cofilin-1 is a Redox-Sensitive Guard of the NLRP3 Inflammasome NLRP3 has a pivotal role in nucleating one of the major inflammasomes. Mutations in the gene encoding NLRP3 cause a spectrum of autoinflammatory disease, the cryopyrin-associated periodic syndromes (CAPS), and the NLRP3 inflammasome is known to contribute to the pathology of other disorders ranging from cardiovascular disease to cancer to gout. Reactive oxygen species (ROS) play a key role in NLRP3 inflammasome activation induced by most extracellular activators. However, the molecular mechanism by which a change in cellular redox state leads to NLRP3 inflammasome activation, as well as the molecular pathogenesis of CAPS, has not been elucidated. Through an unbiased proteomic approach complemented with structural modeling, we identified cofilin-1, an actin-severing protein, as a negative regulator of the NLRP3 inflammasome. In resting cells, we found that cofilin-1 binds directly to the NBD and LRR domains of NLRP3, thereby blocking its oligomerization and association with NEK7—two essential steps for inflammasome activation. We further found that upon stimulation with NLRP3 inflammasome activators, cofilin-1 is oxidized by ROS and dissociated from NLRP3. CAPS-associated mutant NLRP3 exhibited reduced binding to cofilin-1 compared to wild-type, suggesting that the spontaneous inflammasome activation observed in myeloid cells from CAPS patients may result from diminished inhibition of NLRP3 activation by cofilin-1. We further discovered that four amino acid residues of cofilin-1, from 101 to 104 (Phe-Ile-Phe-Trp), were critical for NLRP3 interaction. Oxidation-independent peptides containing this NLRP3-binding motif suppressed inflammasome activation induced by endogenous CAPS-associated mutations and ex vivo NLRP3 activators such as ATP and nigericin, but not flagellin. Our bioinformatic structural analyses corroborate a model in which cofilin-1 plays a pivotal role in NLRP3 activation by ROS and support the potential use of cofilin-1-derived peptides in patients who are unresponsive to or intolerant of other forms of NLRP3 blockade. Cofilin-1 and NLRP3 modelling To investigate the structural basis of the NLRP3–cofilin-1 interaction, we generated full-length complex models using AlphaFold3. The predicted structure revealed a stable interface between cofilin-1 and both the NBD and LRR domains of NLRP3. To further resolve domain-specific interactions, we also modeled cofilin-1 with individual NLRP3 domains. Binding free energy calculations indicated the strongest interaction with the combined NBD–LRR region, consistent with pull-down assays using domain-deletion constructs. We next modeled eight pathogenic NLRP3 variants associated with CAPS in complex with cofilin-1. These substitutions consistently reduced the predicted binding affinity compared to wild-type NLRP3, in agreement with experimental evidence that such mutations impair NLRP3-cofilin-1 binding. We further evaluated the contribution of residues 101–104 in cofilin-1 to its binding to NLRP3. In silico alanine substitutions destabilized the complex, whereas replacement of Phe103 with tyrosine preserved binding affinity, in line with pull-down validation. To examine the role of redox regulation, we modeled oxidized cofilin-1 with a Cys39–Cys80 disulfide bond using Rosetta and ChimeraX. HADDOCK docking revealed weaker affinity of oxidized cofilin-1 for NLRP3, suggesting that ROS-induced oxidation destabilizes the interaction and facilitates inflammasome activation. Finally, alignment of the NLRP3–cofilin-1 complex with cryo-EM NLRP3–NEK7 and closed decameric cage structures demonstrated extensive steric clashes. These analyses show that cofilin-1 and NEK7 cannot bind to NLRP3 simultaneously and cofilin-1 bound NLRP3 cannot form the closed cage structure. These findings support a model in which cofilin-1 binds monomeric NLRP3 under resting conditions, blocking oligomerization and NEK7 engagement. Discovery of a Severe MEFV-Associated Autoinflammatory Disease, CHIVE-18, That Is Distinct from Familial Mediterranean Fever Herein we describe a new disorder we designate as CHIVE-18 (Compound Heterozygosity for I692del and V726A MEFV variants and Elevated IL-18) syndrome, presenting the clinical findings and inflammatory signatures of 45 CHIVE-18 patients and 8 others with two MEFV variants, including at least one copy of I692del. Clinical findings in CHIVE-18 patients often include transfusion-dependent anemia, arthritis, organomegaly, inflammatory bowel disease, and less frequently cytokine storm, AA amyloidosis, and neutrophilic dermatoses. Unexpectedly, we found that individuals who are homozygous for I692del present with milder phenotypes. We hypothesize that the coinheritance of the mutations affecting the I692 and V726 residues could be critical for the tertiary structure of B30.2 domain and interaction with an as-yet unidentified protein or molecule. Total serum IL-18 levels in CHIVE-18 patients were markedly elevated as compared to those of classical FMF patients and healthy controls. Serum total IL-18 and free IL-18 levels were comparable to the previously described levels of patients with pyogenic arthritis, pyoderma gangrenosum, and acne (PAPA) syndrome. Additionally, CHIVE-18 patients exhibited significantly increased levels of TNF, and interferons alpha and gamma. Patients’ myeloid cells showed enhanced STAT1 phosphorylation after stimulation with IFNgamma. Although IL-1beta was not detectable in serum samples, in situ cytokine staining of liver tissue from a previously reported CHIVE-18 patient with autoimmune hepatitis revealed high expression of IL-1beta, along with TNF and IL-18, localized to CD68+ tissue macrophages. Our preliminary data suggest that the mechanisms of pyrin activation in CHIVE-18 may differ from that of conventional FMF-associated variants known to trigger pyroptosis. CHIVE-18 patients usually require an IL-1 inhibitor and sometimes combination therapy with a TNF inhibitor. The importance of distinguishing this condition is emphasized by its association with episodes of cytokine storm that resulted in the deaths of one affected child and several presumably affected (but not previously diagnosed) siblings. Hyperinflammatory episodes should be treated aggressively to prevent morbidity and mortality. In summary, CHIVE-18 is a distinct and severe pyrin-associated disease. Discovery of a New Autoinflammatory Disorder Caused by an Upstream Gene in the Mevalonate Pathway, HMGCS1 The mevalonate kinase (MVK) pathway is a crucial metabolic pathway involved in the synthesis of sterols and isoprenoids. Non-sterol isoprenoids play significant biological roles in the post-translational modifications of various signaling molecules through a process called prenylation, in some circumstances influencing inflammasome activation. Defects in prenylation can disrupt protein trafficking and localization, leading to an increase in pro-inflammatory cytokine production. Enzyme deficiencies within the MVK pathway are known to cause systemic autoinflammatory diseases that exhibit a wide range of clinical manifestations. Moreover, anti-cholesterol drugs such as statins, which inhibit HMG-CoA reductase (the second enzyme in this pathway), may result in muscle inflammation. In this study, we identified biallelic loss-of-function variants in the HMGCS1 gene, which encodes the first enzyme in the MVK pathway, in four patients who presented with recurrent fever, arthritis, abdominal pain, and progressive myositis. Three patients from two unrelated families were homozygous for a rare variant, c.265C>T (p.Arg89Trp), predicted to be harmful by multiple algorithms. One patient was homozygous for a novel variant, c.572G>C (p.Arg191Pro), classified as a variant of uncertain significance (VUS). The enzyme activity of both mutant HMGCS1 proteins was found to be decreased, suggesting a deleterious effect on protein function. Preliminary results indicated that these missense variants do not affect protein expression or dimerization, which are critical for protein activity. Analysis of peripheral blood mononuclear cells (PBMCs) and CRISPR/Cas9-edited HMGCS1-deficient cells demonstrated defects in prenylation, confirming the involvement of the MVK pathway. Cells deficient in HMGCS1 exhibited elevated levels of IL-1β and IL-18 when stimulated with Pam3CSK4 and IFN-γ, indicating that HMGCS1 deficiency resulted in pyrin-dependent inflammasome activation. RNA sequencing of frozen muscle tissues indicated a strong type I and II interferon signature. UMAP analysis of single-cell RNA sequencing data from peripheral blood identified two major cell clusters: NK cells and monocytes. The most significantly upregulated pathways encompassed oxygen/CO2 transport genes, cell death genes, and NF-kappaB signaling. Intracellular cytokine staining revealed a substantial increase in IL-1β, IL-6, and IFN-γ in all monocyte subsets of two affected patients compared to unaffected relatives. Monocytes expressing inflammatory markers, such as CD64, Siglec-1, and CD11c, were increased with a marked expression in non-classical monocytes (NCM). Classical monocytes (CM), intermediate monocytes (IM), and NCM from affected siblings displayed reduced PSGL-1 levels, which correlated with the observed inflammatory responses, suggesting that activated monocytes infiltrated muscle tissues. A decrease in total NK cells was observed in two adult affected siblings, while T cells, especially Th1 and Th17 cells, were expanded. Additionally, T cells in the affected siblings showed elevated expression of IFN-γ, correlating with T cell infiltration in muscle tissue. Notably, activated T cells from the affected siblings exhibited increased levels of granzyme B and perforin but decreased CD107a, indicating that although these cells were activated, they demonstrated reduced degranulation activity. Treatment with a JAK inhibitor in the two adult patients and with anifrolumab in one young patient has shown promise in alleviating systemic and muscle inflammation. Further studies are underway to explore the spectrum of immune dysregulation associated with HMGCS1 deficiency. Extension of a New Autoinflammatory Disorder Caused by a Downstream Gene in the Mevalonate Pathway, PMVK Phosphomevalonate kinase (PMVK) catalyzes mevalonate phosphate to mevalonate pyrophosphate. PMVK deficiency is extremely rare, with only three reported patients presenting chronic refractory arthritis, recurrent inflammation, and elevated IL-1β. Geranyl-geranyl pyrophosphate supplementation reduces IL-1β production, implicating impaired protein prenylation in pathogenesis. Two sisters from consanguineous parents presented with severe autoinflammatory disease. The younger, now nine, developed diarrhea, fever, arthritis, and thrombocytopenia from eight months of age, progressing to chronic systemic inflammation despite multiple therapies, among which high-dose IL-1 blockade was the most effective one. She currently has chronic inflammation and glucocorticoid-related complications. The older sister, now 11, had recurrent fevers, abdominal pain, and arthritis from 2.5 years, initially responsive to colchicine and currently managed with etanercept. After PMVK deficiency was diagnosed, urinary mevalonic aciduria was demonstrated in one sister during active disease. Both patients were evaluated at the NIH Clinical Center, where treatment plans included increasing the anti–IL-1 dose for the younger sister. Genome sequencing performed at the NIH Intramural Sequencing Center identified a homozygous PMVK variant (p.Val106Met) in both sisters, while their unaffected parents were heterozygous carriers. The variant was confirmed by Sanger sequencing. It is extremely rare (one heterozygous carrier in gnomAD v4.1.0) and predicted to be pathogenic by multiple in silico tools (CADD 24.8; REVEL 0.43; α-missense 0.53). In silico thermodynamic predictions (DynaMut2, PremPS, INPS-3D) indicated that the V106M variant destabilizes the protein. Functional studies were initiated using PBMCs isolated from the patients and their parents. PBMCs were stimulated with LPS, and IL-1β secretion was measured by ELISA. Preliminary experiments showed no significant differences between patients and parents, though repeat studies are underway. Cryopreserved PBMCs were also sent for prenylome analysis at the Garvan Institute (Sydney, Australia). Unlike MKD, no clear prenylation defect was detected, although results were inconclusive and repeat assays are planned. Overall, our findings expand the very limited knowledge of PMVK deficiency and support the pathogenicity of the V106M variant in this family. Description of a New Illness Caused by Mutations in NLRC3, a Regulator of Innate Immune Signaling NLRC3 is a negative regulator of innate immunity, controlling NF-κB, STING/TBK1, and inflammasome pathways. No human disease had previously been linked to NLRC3. Our patient is a 24-year-old male born to consanguineous parents with recurrent febrile episodes from two months of age, including abdominal and chest pain, lymphadenopathy, myalgia, headache, oral aphthosis, and musculoskeletal pain. His disease was partially responsive to colchicine, with additional benefit from on-demand anakinra. Genome sequencing revealed a homozygous NLRC3 p.Leu90Pro (L90P) variant; his unaffected parents and brother were heterozygous carriers. This variant is absent in homozygous state in gnomAD and predicted to be pathogenic by multiple in silico tools. In silico stability predictions (DynaMut2, PremPS, INPS-3D) suggested that the L90P variant destabilizes NLRC3. Structural modeling of NLRC3 complexes with TRAF6, STING, and caspase-1 further predicted that L90P, but not a benign G82R variant, significantly reduces protein–protein binding affinity in these complexes, consistent with a loss-of-function effect. To investigate functional effects, we generated NLRC3 knockout THP-1 monocytic cells using CRISPR/Cas9. Upon LPS stimulation, KO cells produced markedly higher IL-1β, IL-6, and TNF-α than WT controls, consistent with hyperactive NF-κB signaling. Treatment with MCC950 partially reduced IL-1β in both WT and KO cells, whereas colchicine had no effect. Western blot confirmed increased pro–IL-1β in KO cells after stimulation. We are now reintroducing WT or L90P NLRC3 into KO cells to directly compare cytokine production and protein interactions with TRAF6, STING, and caspase-1. In parallel, we plan to study patient-derived PBMCs to validate these mechanistic insights. Overall, these findings establish NLRC3 as a novel human autoinflammatory disease gene and highlight its critical role as a negative regulator of innate immune signaling. Section II: Monogenic Autoinflammatory Disorders Involving Molecular Mechanisms other than the Inflammasome A second major area of investigation for our group has been the monogenic autoinflammatory disorders that do not fit easily into the category of inflammasomopathies. Over the last reporting period, these include two known illnesses, the deficiency of adenosine deaminase 2 (DADA2) and the retinal dystrophy with optic nerve edema, splenomegaly, anhidrosis, and headache (ROSAH) syndrome, and three newly identified illnesses in the UNC93B1, TYK2, and DHX36 genes. DADA2 The deficiency of adenosine deaminase 2 (DADA2) was initially identified and characterized by our group in 2014. It is caused by biallelic variants in ADA2, which were initially shown to cause endothelial damage and a shift towards pro-inflammatory M1 macrophages, but probably cause other as-yet uncharacterized molecular and cellular abnormalities. Affected individuals can present with a broad spectrum of phenotypes that can be grouped into three main clusters: inflammatory, immune dysregulatory, and hematologic. Usually patients have a primary phenotypic grouping but, upon thorough evaluation, have features of disease that are associated with the other phenotypic groups. The inflammatory features respond robustly to treatment with tumor necrosis factor inhibitors (TNFi) but the hematologic/immunologic features have a small to modest response which has resulted in these patients often being referred for hematopoietic stem cell transplant (hSCT). In DADA2 patients undergoing hSCT with a history of pure white cell aplasia (PWCA), there have been a number of graft failures resulting in the need for a second or third transplant. As our group has amassed over 80 patients with DADA2, we have had the opportunity to review greater than 20 bone marrows of patients primarily with pure red cell aplasia (PRCA) and/or PWCA. We have observed that in those patients, there is almost always a large infiltrate of polyclonal T-cells into their bone marrow. To investigate whether the T-cells may be contributing to the patients’ anemia and/or neutropenia, we added a T-cell directed therapy (cyclosporine) to their medication regimen. We have now treated 4 patients with PWCA (decreased ANC for 30 years, 30 years, 2 years, and 2 months) and 1 with PRCA (transfusion dependent for 20 years) using cyclosporine. In the cohort with PWCA, they all responded within 1-3 months with normalization of the ANC to greater than 1500 µL. In the patient with PRCA, his reticulocyte count went from 0.3% to 5% and his hemoglobin and hematocrit stabilized within 3 months of cyclosporine usage. He has not needed a transfusion in 3.5 months which has never happened since he developed PRCA at 8 months of age. There is a 5th patient with both PRCA/PWCA who has had cyclosporine initiated within the past 2 months. Results are pending. Repeat bone marrow biopsy on one of the PWCA patients with a 30-year history of ANC <500 µL revealed a marked decrease in the T cell infiltrates in the marrow and restoration of myelopoiesis. Two of the patients with PWCA were able to bridge to transplant successfully without any significant complications and are now cured. This therapeutic approach is promising for patients who a) do not have a sufficient donor available for hSCT; b) aren’t transplant candidates due to comorbidities from their DADA2; and/or c) do not want to undergo hSCT. Further research studies including single cell analysis of the lymphocytes is ongoing as we try to further understand the mechanisms underlying these hematologic manifestations of DADA2. ROSAH Syndrome ROSAH is a rare, autosomal dominant autoinflammatory disorder caused by gain-of-function mutations in ALPK1, an atypical kinase that is a sensor of 7-carbon sugars that are made by bacteria but not mammalian species and thereby can be sensed by the innate immune system. ROSAH was first recognized as a distinct genetic condition in 2019, and since that time our group has led efforts to define its clinical spectrum and underlying mechanisms. Over the past five years, in collaboration with Dr. Christina Kozycki, an Assistant Clinical Investigator in the National Institute of Allergy and Infectious Diseases, we have advanced the scientific understanding of ROSAH and systematically assessed the impact of immunomodulatory therapies on ocular manifestations, which are of particular importance given that untreated disease frequently progresses to blindness by mid-adulthood. In FY25, we continued to expand our clinical cohort and have now enrolled and deep-phenotyped 60 individuals with ROSAH. Each participant was enrolled on our longitudinal natural history protocol, which allows for capturing comprehensive ophthalmologic, immunologic, and systemic data to better characterize the clinical manifestations of the disease and elucidate the natural course over time. Insights gained from this growing cohort have not only deepened our understanding of disease heterogeneity but have also directly informed patient care, particularly with respect to identifying therapeutic strategies that may slow or prevent vision loss. We continue to evaluate the possible role of tocilizumab in ameliorating the advance of vision loss. By continuing to grow this unique cohort and strengthen our longitudinal dataset, we are laying the groundwork for future clinical trials and therapeutic development aimed at altering the trajectory of this blinding disease. Clinical observations in our ROSAH cohort have prompted us to examine the possibility of innate immune crosstalk beyond ALPK1’s canonical pathway. While ALPK1 is canonically associated with NF-kappaB activation through TIFA phosphorylation in response to microbial sugars, we observed that most individuals with ROSAH exhibit basal ganglia mineralization, a radiographic finding that has been associated with interferonopathies such as Aicardi-Goutières syndrome. This overlap raised the possibility that ALPK1 gain-of-function mutations may also influence interferon signaling pathways. In FY25, in collaboration with Dr. John Kehrl’s group in NIAID, we demonstrated that ALPK1 indeed amplifies interferon signaling through engagement of the STING pathway, a central sensor of cytosolic nucleic acids. In a study released as a preprint in July 2025, we showed that ALPK1 activation enhances both canonical and non-canonical STING outputs, including both canonical and noncanonical responses such as STING proton channel-dependent LC3B lipidation and NLRP3 inflammasome activation. Furthermore, ALPK1 signaling activates eIF2α, a key effector of the integrated stress response. Conversely, STING activation increases ALPK1 protein expression and triggers TIFA-threonine 9 phosphorylation. These findings support a model of bidirectional signaling between ALPK1 and STING, in which microbial and nucleic acid sensing pathways can amplify one another. Together, these findings highlight a previously unrecognized feedback loop between ALPK1 and STING that broadens our understanding of innate immune regulation and identifies new opportunities for therapeutic intervention in autoinflammatory disease. Identification of a New UNC93B1 Variant Associated with a Type I Interferon Signature Gain-of-function UNC93B1 variants have recently been associated with the development of early onset systemic lupus erythematosus (SLE) or chilblain lupus. We identified a four-year old Caucasian male with early onset lymphohistiocytic panniculitis, recurrent fevers, fatigue, myalgia, and a Type I interferon signature in the peripheral blood. Family history was remarkable for lupus panniculitis in his mother, maternal great-aunt and great-grandfather, and SLE with class III lupus nephritis in his deceased maternal grandfather. Trio whole genome sequencing detected a heterozygous single nucleotide variant in UNC93B1 (c.1574G>C, p.Arg525Pro) in the proband and his affected mother, and targeted sequencing confirmed the variant in the affected great-aunt. The UNC93B1 variant c.1574_1575delinsCT, p.Arg525Pro has been reported in four patients from a three-generation family who presented with chilblain lupus and a Type I IFN signature but no SLE. Functional assessment of the variant revealed an enhanced TLR8 signaling in the patients’ PBMC and monocytes and in THP-1 cells. Treatment with the anti-Type I IFN receptor monoclonal antibody, anifrolumab, was initiated in our patient and resulted in significant improvement of our patient’s skin rash and systemic inflammation. TYK2 Gain-of-Function Variants in Human Immunodysregulation TYK2 is a Janus kinase (JAK) essential for the promulgation of cytokine signals in innate and adaptive immune systems. We identified five patients with immunodysregulation harboring heterozygous, novel or ultrarare variants in TYK2, all predicted to be pathogenic in silico. Clinical phenotypes included linear scleroderma with aberrant immune composition in two patients carrying p.E41V and p.P441R, dermatomyositis in two patients with p.H667Q and p.E925K, and an autoinflammatory syndrome with mucosal ulceration, periodic fever, and rash in one patient carrying p.K642R. Functional studies demonstrated that these TYK2 variants drive enhanced signaling. Transient and stable transfections revealed increased basal TYK2 phosphorylation, heightened IL6 promoter activity, and heightened STAT phosphorylation upon cytokine stimulation. Ex vivo stimulation and phenotyping of available patient PBMCs showed hyper-responsiveness downstream of TYK2-mediated cytokines, via enhanced STAT phosphorylation and immune activation. Bulk RNA sequencing of whole blood further distinguished patient transcriptional signatures from healthy donors. Therapeutic testing revealed TYK2-driven hyperactivation with variant-specific drug responses. The TYK2-selective inhibitor deucravacitinib effectively suppressed STAT activation in cells carrying p.E41V, p.P441R, p.H667Q, and p.E925K. In contrast, the p.K642R variant, positioned within the inhibitor’s binding site, was completely resistant to TYK2-selective blockade with deucravacitinib (as would be expected) but remained responsive to broader JAK inhibition. Collectively, these results establish that the identified TYK2 variants represent the first germline hypermorphic variants in TYK2, directly linking enhanced TYK2 signaling to immunodysregulation. A Genotype-First Approach Delineates a DHX36-Associated Inflammatory Disease DHX36 encodes an ATP-dependent helicase that resolves G-quadruplex (G4) DNA and RNA to enable immune gene expression, while also serving as a viral nucleic acid sensor. Using a genotype-first approach, in collaboration with Dr. Laura Lewandowski of the National Institute of Arthritis and Musculoskeletal and Skin Disease we identified ten heterozygous, dominantly inherited or de novo variants in DHX36 in ten unrelated families with early-onset systemic inflammation characterized by overlapping Behçet’s-like and/or lupus-like features, often accompanied by recurrent infections. All variants were ultra-rare or novel, predicted to be pathogenic, and clustered within the protein’s helicase domain. Structural modeling predicted destabilization of DHX36 and impaired binding to G4 substrates and viral sensing partners such as RIG-I. In vitro assays confirmed reduced protein stability. Stable reconstitution of DHX36-deficient cells with mutant alleles drove increased stress granule formation, reflecting accumulation of cytoplasmic G4s, compared with wild-type rescue. In a complementary transient transfection system, poly(I:C) stimulation revealed defective NF-kappaB and ISRE reporter activation, consistent with impaired viral sensing, particularly in variants carried by patients with immunodeficiency. Together, these findings support a loss-of-function mechanism. Ex vivo studies of patient-derived cells corroborated these findings, showing increased stress granule accumulation and dysregulation of downstream interferon and stress signaling pathways, particularly in patient dendritic cells. Together, these results establish that heterozygous loss-of-function variants in DHX36 disrupt G-quadruplex resolution and viral sensing, linking impaired immune homeostasis to combined inflammation and immunodeficiency. Section III: Aberrant Disorders of SAA Amyloidosis One of the important downstream consequences of both monogenic and genetically complex inflammatory disorders is the development of serum amyloid A (SAA)-driven amyloidosis. In some cases, misfolded SAA leads to the accumulation of fibrils in the kidneys, gastrointestinal, thyroid, adrenal, or other organs, leading to morbidity and mortality. Prior to the development of antibiotics, chronic infectious disorders could also lead to amyloidosis. With advances in diagnosis, the advent of antibiotics, and the development of biologic therapies, SAA amyloidosis has become relatively uncommon, yet there remain patients with still unexplained AA amyloidosis. In this Section, we present discoveries we have made during the current reporting period that explain some of these previously unexplained forms of SAA amyloidosis. An SAA1 Promoter Variant Underlying AA Amyloidosis By performing whole genome sequencing (WGS) in a patient with idiopathic AA amyloidosis and kidney failure, we identified an ultra-rare homozygous promoter variant in the SAA1 gene, predicted to increase gene expression by two deep learning algorithms, Enformer and PromoterAI. The variant was subsequently identified in homozygosis in an affected brother, and in heterozygosis in the unaffected father. Interestingly, the proband had negative C-reactive protein (CRP), an acute phase reactant that typically is increased in the context of systemic inflammation and elevated SAA circulating levels, supporting the hypothesis of a variant specifically affecting SAA1 transcription. The variant disrupts the promoter motif for the putative transcription repressor ZNF766. As SAA1 is mainly synthesized in the liver and adipose tissue, we are currently generating induced pluripotent stem cells (IPSCs) from the patient’s peripheral blood mononuclear cells (PBMCs) to characterize the effects of this variant in a patient-derived hepatocyte cell line. In a previous report of AA amyloidosis due to a different heterozygous variant in the SAA1 promoter, 3 of 4 affected individuals treated with tocilizumab had a reduction in circulating SAA levels, indicating that modulating the IL-6 signaling pathway may represent a therapeutic option in this condition. The IPSCs-derived hepatocyte cell lines will be used to assess the efficacy of potential treatments. FRK in Systemic AA Amyloidosis We encountered a family with progressive systemic AA amyloidosis requiring maintenance dialysis or kidney transplantation. One of the patients in this family did not respond to anti-IL-6 receptor antibody therapy, with the CRP, one of the serum inflammation markers, remaining positive. Using WGS, we identified a novel germline heterozygous missense variant (p.Glu346Gly) in the FRK gene. Using Sanger sequencing in 6 family members, we confirmed the presence of the variant in 2 patients with systemic AA amyloidosis and in an asymptomatic 27-year-old relative. FRK is member of the BRK family of kinases, distantly related to Src family kinases, that phosphorylates STAT3 independently of IL-6 signaling, thereby increasing the expression of downstream genes such as CRP. We transfected this variant into two human hepatocyte cell lines (Hep3B and Huh7) and confirmed that, in the absence of IL-6 stimulation, STAT3 phosphorylation was enhanced, and the expression of the downstream amyloidosis-related genes SAA and CRP was increased. Gene expression was significantly suppressed by dasatinib, a pan-Src inhibitor approved by the FDA for chronic myeloid leukemia. This study suggests that a novel variant in FRK causes systemic AA amyloidosis and that FRK-targeted therapy, rather than conventional anti-inflammatory empirical treatment, may be effective for patients with this FRK variant. Section IV: Development of New Methodologies In collaboration with Dr. Steven Mount of the University of Maryland College Park, we have utilized bioinformatic methods to improve gene discovery by improving variant impact prediction. One such project is nearly ready for submission. Freyja: Incorporation of Recent Selection Signals Improves Variant Impact Prediction Accurate prediction of missense variant effects remains a major challenge in human genetics, particularly for immune-mediated diseases, where diagnostic yield remains disproportionately low compared with other rare conditions. In this project we found that conventional prediction tools frequently misclassify pathogenic variants in immune datasets, and that accurate interpretation requires systematically lower classification thresholds than for other disease genes. We hypothesized that this discrepancy reflects atypical patterns of recent selection. Using four variant datasets (ClinVar, Infevers, IUIS, OMIM) and eight genome-wide selection metrics (e.g., iHS, SDS, FST), we found that variants in genes under positive selection also require significantly lower thresholds, and that these genes are strongly enriched for immune pathways. To address this, we developed Freyja, an ensemble random forest model that integrates recent selection signals with structural features and evolutionary constraints. Freyja outperformed 45 existing tools across multiple benchmarks, achieving AUCs up to 0.96, and identifies overlooked pathogenic variants. We also provided variant curation guidelines that improve interpretation using widely used prediction tools, including both immune-specific recommendations and recommendations for genes under positive selection. Freyja and its selection-informed framework advance variant interpretation and disease gene discovery, offering a robust tool for clinical genomics and immune diagnostics. Section V: Somatic mutations in autoinflammatory disease Based on our discovery of VEXAS in a previous reporting period, we have been very interested in the discovery of other autoinflammatory diseases caused by somatic mutation. The Detection of Somatic Mutations Even at Low Variant Allele Fractions (VAFs) Although the catalog of genes implicated in the pathogenesis of monogenic systemic autoinflammatory diseases (SAIDs) continues to expand, a significant proportion of patients with suspected SAID still remains without a molecular diagnosis despite extensive genetic testing. A potential contributor to this diagnostic gap is mosaicism, a condition where only a portion of the cells in the organism harbor the genetic variant, resulting in a low variant allele fraction (VAF) that may not be detected by standard next generation sequencing and germline variant calling pipelines. In order to define the role of somatic mutations in the pathogenesis of SAIDs, we performed somatic variant calling using Mutect2 and LoFreq in standard depth exomes from patients with suspected SAID and their unaffected relatives, thus creating an in-house database of somatic variants in individuals with SAIDs. We assessed the presence of clonal hematopoiesis (CH) using both a classic approach based on known CH-driver variants and a barcode approach based on the number of passenger mutations. In individuals more likely to have a higher somatic mutational burden, including patients with late-onset disease, subjects with germline predisposition syndromes/defects in DNA repair genes or exposure to genotoxic agents, and patients with CH, to increase the sensitivity for low-VAF variants, we performed deep exome sequencing at 300X. In a subset of high-risk patients we additionally performed WGS at 100X using the Ultima Genomics duplex error-corrected sequencing ppmSeq methodology, which provides high specificity improving accuracy even at low VAFs. Standard depth exome somatic variant calling has been performed in ~1400 individuals, and analysis of deep exome sequencing and ppmSeq data is ongoing for 70 and 18 individuals, respectively. Section VI: The Role of NOTCH4 in Scleroderma For several years the Inflammatory Disease Section has collaborated with Dr. Pravitt Gourh, an Assistant Clinical Investigator in the National Institute of Arthritis and Musculoskeletal and Skin Diseases, leading a large US group studying scleroderma in the African American population. Dr. Gourh recently took a position in Novartis, the pharmaceutical company, and subsequently the IDS has taken a larger role in ensuring that the collaborative projects are completed. Gain of Function NOTCH4 Variants Disrupt Angiogenesis in Systemic Sclerosis We are currently making revisions on a manuscript submitted to the Annals of Rheumatic Diseases focusing on the genetic underpinnings of vasculopathy and fibrosis in systemic sclerosis (SSc) in African American (AA) patients. The study combines genetics, single-cell RNA sequencing, functional assays, and a mouse model to explore the role of NOTCH4 in SSc vasculopathy and the potential for NOTCH4-directed therapies. Gene-based testing identified NOTCH4 association at an exome-wide significance with SSc (P=1.6x10-7) and patients with severe vascular disease (P=3.5x10-7). The risk haplotype defined by the missense (c.2824C>T) and promoter (c.-117G>A) variants was enriched in AAs with SSc (11%) vs. controls, and the population attributable risk due to this haplotype in AAs with SSc was 2.6%, which was 52-fold higher than in European Americans. The SSc-associated NOTCH4 variants increased NOTCH4 expression and signaling, leading to decreased angiogenesis and increased endothelial-to-mesenchymal transition (EndoMT). Nailfold capillary abnormalities, decreased angiogenesis, and fibrosis of the vascular lumen are commonly seen in SSc. Genetic, chemical, antibody, or FDA-approved drug inhibition of NOTCH4 signaling rescued angiogenesis and returned EndoMT to baseline. From this study, we concluded that NOTCH4 variants are associated with SSc pathogenesis and vasculopathy, partly explaining the increased prevalence of SSc in AAs. The study highlights the need for further research and clinical trials in the inhibition of the NOTCH4 pathway as a strategy to treat the vascular and fibrotic manifestations of SSc.