Genome-wide association for sarcoidosis identifies novel risk loci and genetic heritability in African and European ancestries: a meta-analysis from the Finngen, Million Veteran Program, UK Biobank, and Biobank Japan datasets Authors: Andrea Ricci, Federica Andolfi, Daniele Sabbatini, Filippo Gozzi, Giada Di Betto, Paolo Ventura, Elena Buzzetti, Antonello Pietrangelo, Enrico Clini, Roberto Tonelli, Dario Andrisani, Brent Julius de Guzman Marinduque, Elisa Bergamini, Chiara Vecchi, Elena Pegoraro, Dario Gregori, Elena Corradini, Stefania Cerri Published: 2025 ABSTRACT AbstractIntroductionSarcoidosis is an inflammatory disease driven by immune-mediated mechanisms, characterized by the formation of epithelioid cell granulomas and a wide range of clinical manifestations. Its phenotype is the result of a complex interplay of genetic and environmental factors, the precise roles and interactions of which remain poorly defined.AimTo identify candidate genes and risk loci associated with sarcoidosis from large population datasets. To estimate the genetic heritability of the phenotype in selected ancestries.Population and methodsPublic summary statistics from the FinnGen release 12 (European ancestry), pan UK BioBank Project (UKBB - European and African ancestry), Million Veteran Program (MVP - European and African ancestry), and Japan BioBank (East Asian ancestry) were included for European, African and multi-ancestry meta-analysis through sample size-based analysis. Novel risk loci and single nucleotide polymorphisms (SNPs) significantly associated with the disease were critically reviewed on the basis of the available literature. For each risk locus, SNPs highly correlated with the lead SNP were selected based on Combined Annotation Dependent Depletion (CADD) scores. Genetic heritability (h2) scores were obtained through ancestry-specific linkage-disequilibrium score calculation.ResultOverall 9659 cases (7559 European, 1880 African, 220 East Asian) and 1,665,804 controls (1,361,726 European, 126,411 African, 177,667 East Asian) were analysed. Nineteen and two risk loci were identified in European and African ancestry, respectively; h2 scores were 0.25 (European) and 0.19 (African). Candidate non-MHC genes for further explorations through functional studies included IL23R, PUS10, ACOXL, PLCL1, FAM117B, BMPR2, PPARG, ESYT2, ANXA11, CCDC88B, ATXN2, CCL24, RP11−540O11.1, HOMER2, CD19, UBASH3A, RNF215, and others. Interferon gamma signaling, meiotic recombination/condensation of prophase chromosomes, and DNA methylation were the most enriched gene sets in European and multi-ancestry meta-analysis. Multi-ancestry meta-analysis was confronted with FinnGen+UKBB+MVP meta-analysis (released by FinnGen freeze 12) yielding consistent results (18 risk loci identified)ConclusionNineteen and two risk loci were significantly associated with sarcoidosis for European and African ancestries, respectively. Moderate genetic heritability was observed for both ancestries. A set of significantly associated non-MHC genes and SNPs was obtained to investigate functional validation. Although further studies are warranted, epigenetic alterations may contribute to the risk of developing sarcoidosisSupplementary informationThe online version contains supplementary material available at 10.1186/s13023-025-04097-1.Keywords: Sarcoidosis, GWAS, Meta-analysis, Rare diseases, UK Biobank, FinnGen, Million veteran program, Biobank Japan Introduction Sarcoidosis is an inflammatory disease driven by immune-mediated mechanisms, characterized by the formation of epithelioid cell granulomas and a wide range of clinical manifestations [1]. Although the disease occurs worldwide, its prevalence and clinical features have been shown to differ across ethnic groups [2–4]. This variability, along with familial clustering and an increased risk in relatives, underscores its multifactorial nature, reflecting the complex interplay of genetic and environmental factors, the precise roles and interactions of which remain poorly defined [5]. The identification of genetic variants through genome-wide association studies (GWAS), whole exome sequencing (WES), and whole-genome sequencing (WGS) have been conducted with the aim of uncovering the contribution of genetic components to the development of sarcoidosis [6]. These studies have not only confirmed associations with major histocompatibility complex (MHC) class II genes, particularly HLA-DRB1 and BTNL2 variants within the HLA class II region, but have also led to the identification of several additional genetic loci/genes that may confer susceptibility to sarcoidosis [7, 8]. The growing availability of data from large population-based GWAS datasets has boosted the potential to identifying new risk loci, as well as corroborating the significance of those already reported in the literature. Moreover, analysis from large multi-ethnic populations can provide additional evidence for the correlation between SNP expression and ethnic variability, given a particular phenotype. In this study, we conducted a meta-analysis of the publicly available summary statistics of four genome-wide association studies (GWAS) for sarcoidosis, including results from European (EUR), African (AFR), and East Asian (EAS) ancestries. Through bioinformatic analyses and critical comparison with the available literature, we sought to identify novel ancestry-specific risk loci, estimate genetic heritability of the disease, and highlight possible metabolic pathways of interest for functional validation. Methods Datasets and patient population The latest public updates of the following datasets were accessed: FinnGen freeze 12 [9]: https://www.finngen.fi/en/access_results, phenotype SARCOIDOSIS (released on 4 November 2024); the FinnGen study is a large-scale genomics initiative that has analyzed over 500,000 Finnish biobank samples and correlated genetic variation with health data to understand disease mechanisms and predispositions; the project is a collaboration between research organisations and biobanks within Finland and international industry partners https://www.finngen.fi/en; FinnGen freeze 12 [9]: https://www.finngen.fi/en/access_results, phenotype SARCOIDOSIS (released on 4 November 2024); the FinnGen study is a large-scale genomics initiative that has analyzed over 500,000 Finnish biobank samples and correlated genetic variation with health data to understand disease mechanisms and predispositions; the project is a collaboration between research organisations and biobanks within Finland and international industry partners https://www.finngen.fi/en; Pan-UK Biobank project [10] https://pan.ukbb.broadinstitute.org/, phenotype “D86 sarcoidosis” (last recorded update 16 March 2023); the UK Biobank is a large-scale biomedical database and research initiative that has recruited around 500,000 individuals aged 40 to 69 years and collected extensive genetic and phenotypic information https://www.ukbiobank.ac.uk/; Pan-UK Biobank project [10] https://pan.ukbb.broadinstitute.org/, phenotype “D86 sarcoidosis” (last recorded update 16 March 2023); the UK Biobank is a large-scale biomedical database and research initiative that has recruited around 500,000 individuals aged 40 to 69 years and collected extensive genetic and phenotypic information https://www.ukbiobank.ac.uk/; Million Veteran Program (MVP) [11] (last recorded update 22 August 2023): Phecodes Phe 697.EUR.GIA (European) and Phe 697.AFR.GIA (African American); the MVP is an ongoing prospective cohort study and biobank in the Department of Veterans Affairs Healthcare System designed to study genetic influences on health and disease among veterans https://www.ncbi.nlm.nih. gov/projects/gap/cgi-bin/study.cgi?study_id = phs001672.v12.p1; Million Veteran Program (MVP) [11] (last recorded update 22 August 2023): Phecodes Phe 697.EUR.GIA (European) and Phe 697.AFR.GIA (African American); the MVP is an ongoing prospective cohort study and biobank in the Department of Veterans Affairs Healthcare System designed to study genetic influences on health and disease among veterans https://www.ncbi.nlm.nih. gov/projects/gap/cgi-bin/study.cgi?study_id = phs001672.v12.p1; Biobank Japan (BBJ) https://pheweb.jp/, summary statistics available from [12] (published on 30 September 2021), phenotype sarcoidosis; BioBank Japan is a disease-oriented biobank that receives biological samples and clinical information from participating patients diagnosed with target diseases through cooperative medical institutions nationwide and stores such samples and information https://biobankjp.org/en/. Biobank Japan (BBJ) https://pheweb.jp/, summary statistics available from [12] (published on 30 September 2021), phenotype sarcoidosis; BioBank Japan is a disease-oriented biobank that receives biological samples and clinical information from participating patients diagnosed with target diseases through cooperative medical institutions nationwide and stores such samples and information https://biobankjp.org/en/. For each dataset, the following ancestry-specific samples size were reported: Finngen SARCOIDOSIS, r0.12, 5411 cases, 492,311 controls; Finngen SARCOIDOSIS, r0.12, 5411 cases, 492,311 controls; Million Veteran Program, EUR ancestry 1509 cases, 449,523 controls; Million Veteran Program, EUR ancestry 1509 cases, 449,523 controls; Million Veteran Program, AFR ancestry 1827 cases, 119,828 controls; Million Veteran Program, AFR ancestry 1827 cases, 119,828 controls; UK Biobank (pan-UKBB), EUR ancestry, 639 cases, 419,892 controls; UK Biobank (pan-UKBB), EUR ancestry, 639 cases, 419,892 controls; UK Biobank (pan-UKBB), AFR ancestry, 53 cases, 6583 controls; UK Biobank (pan-UKBB), AFR ancestry, 53 cases, 6583 controls; Japan Biobank, EAS ancestry, 220 cases, 177,667 controls. Japan Biobank, EAS ancestry, 220 cases, 177,667 controls. All datasets adopted the most recent GRCh38 assembly of human genome except BBJ, which followed the GRCh37 assembly. Therefore, chromosome coordinates in the BBJ dataset were substituted with coordinates from the UKBB reference dataset through rsIDs cross-referencing (R script available in Supplementary data). Prior to performing meta-analysis, all summary statistics were individually replicated. The BBJ, MVP and meta-analysis datasets lacked nearest gene specifications, so annotation through rsIDs cross-referencing from the UKBB reference dataset was performed (R script available in Supplementary data). Meta-analyses Dataset handling, SNP filtering, and plotting of Manhattan and qq plots was performed through R version 4.4.1 [13], using the following packages: parallel, vroom, data.table, dplyr, BiocManager, Biostrings, SeqArray, ggmanh, SeqArray, farver, labeling, readr, knitr, kableExtra. R markdown html reports for the main analyses are available in supplementary data. Sample size-based meta-analysis of summary statistics from EUR, AFR, and multi-ancestry were performed using METAL [14]. In all analyses, correction for genomic inflation λ was applied. A p- value < 5 × 10−8 was required to assess statistical significance. Prior to meta-analysis, the datasets were filtered according to a common set of SNPs. AFR-specific meta- analysis included 1880 cases, 126,411 controls; 21,388,456 common SNPs were analysed. EUR-specific meta-analysis included 7559 cases, 1,361,726 controls; 10,776,775 common SNPs were analysed. Multi- ancestry meta-analysis included 9659 cases, 1,665,804 controls; 5,586,486 common SNPs were analysed. A similar multi-ancestry meta-analysis (with no EAS ancestry) was issued with the FinnGen r0.12 freeze and can be accessed at https://mvp-ukbb.finngen.fi/pheno/D3_SARCOIDOSIS. Genetic heritability estimation Genetic heritability (h2) was estimated by means of linkage disequilibrium (LD) score regression through the LDSC tool [15, 16]. EUR and AFR LD scores were made available by the pan-UKBB project website [10]. Risk loci annotation Risk loci plotting and annotation was performed with Functional Mapping and Annotation of GWAS (FUMA GWAS) [17]; ANNOVAR [18] and Combined Annotation-Dependent Depletion (CADD) [19] were used for functional annotation. FUMA allowed also to perform tissue expression (TEA) and gene- set enrichment analysis (GSEA) through MAGMA [20] and the GENE2FUNC option, based on the Genotype-Tissue Expression (GTEx) project, version 8 [21, 22] for expression quantitative trait locus (eQTL) estimations. Among the available datasets, REACTOME [23] was used for GSEA. The MHC region was excluded from all analyses except GSEA. Plink 1.9 [24, 25] with GRCh38-based 1000 Genomes (1000 G) Project Phase 3 reference panel [26, 27], was used to clump variants according to linkage disequilibrium (r2 > 0.6). Plink and the 1000 G reference (EUR and AFR ancestries) were also used by FUMA to calculate r2 and minor allele frequency estimations for risk loci plotting and annotation. For each locus in EUR ancestry meta-analysis, SNPs used for mapping were manually selected based on their CADD score ( > 12.37). Literature search was conducted to critically assess the significance of the identified genes. Result Overall 9659 cases (7559 European, 1880 African, 220 East Asian) and 1,665,804 controls (1,361,726 European, 126,411 African, 177,667 East Asian) were analysed. Risk loci identification EUR ancestry Nineteen risk loci were identified in EUR ancestry meta-analysis (Fig. 1, regional plots are reported in Fig. 2): of these, at least six (see Supplementary Table 1) had not been previously associated with the phenotype (at p-value < 5 × 10−8 significance level) in any single GWAS. Manhattan plot for EUR ancestry meta-analysis Regional plots from EUR ancestry meta-analysis AFR ancestry Two risk loci were identified in AFR ancestry meta-analysis (Fig. 3), mainly coinciding with results from the MVP summary statistics (Fig. 4 and Supplementary Files). Manhattan plot for AFR ancestry meta-analysis Regional plots from AFR ancestry meta-analysis multi ancestry Eighteen risk loci were identified in multi-ancestry meta-analysis (Fig. 5): compared to EUR ancestry, additional loci were identified in chromosomes 1, 3, 10, and 12 (Fig. 6). Multi-ancestry meta-analysis was confronted with FinnGen+UKBB+MVP meta-analysis (released by FinnGen freeze 12) yielding consistent results. Manhattan plot for multi-ancestry meta-analysis Regional plots from multi-ancestry meta-analysis Genetic heritability Total observed scale h2 in AFR ancestry was 0.1879 (0.2089), with λ = 1.0016 (mean χ2: 1.0062, intercept: 0.9864). Total observed scale h2 in EUR ancestry was 0.2504 (0.0581), with λ = 1.0436 (mean χ2: 1.1116, intercept: 0.9469). Overall, mild-moderate and moderate genetic heritability was estimated for AFR and EUR ancestry, respectively; mild residual genomic inflation could be observed in EUR ancestry. Functional annotation of variants Tissue expression The most significantly represented tissues (p-value < 10−3) in the EUR ancestry meta-analysis were whole blood, spleen, terminal ileum, visceral omentum, mammary tissue, and lung (Fig. 7). TEA in multi ancestry meta-analysis highlighted a significant association with whole blood and spleen (Supplementary Figure S1). No tissues in AFR meta-analysis were associated with the phenotype at the required significance level; however, the most significantly represented tissues were adrenal gland, kidney cortex, and pancreas (Supplementary Figure S2). Tissue-enrichment analysis for EUR ancestry Assessment of significance genes and SNPs Critical assessment of putative functional roles for non-MHC genes significantly associated with the phenotype in EUR and multi ancestry is reported in Table 1. Overall, SNPs from the following genes have already been identified in previous studies on sarcoidosis (GWAS or other): CCDC88B, ANXA11, IL23R, FAM117B, CCL24, ATXN2/SH2B3. The following additional genes have been associated with sarcoidosis through functional or other experimental studies: PLCL1, PPARG, CD19. Indirect evidence of some involvement in the pathogenesis of sarcoidosis is available for PUS10 and ESYT2. The following additional genes have been involved in immune (dys)regulation: RNF215, RPS6KA4, LAT, UBASH3A, HOMER2, SUMO2P (indirect evidence). Finally, no clear evidence linking ACOXL, MAT1A, or MACROD1 to sarcoidosis or immune (dys)regulation was found from the available literature. Table 1Genes significantly associated with the phenotype were critically reviewed on the basis of known mechanisms of action, available evidence of a role in sarcoidosis or related phenotypes, and previous identification in GWAS studies (rsids which have been identified both in the literature and in the present study are in bold). Comparison with the very recent study by Yuan et al. [28] could not be performed as complete summary statistic could not be accessed yetGene/ProteinMechanismEvidenceGWASCCDC88B/GIPIE• role in suppressing apoptosis induced by endoplasmic reticulum stress in endothelial cells ( [29], not replicated in [30])• contributes to the maturation and activation of T cells [30]• expressed in lung cells [31]; has a role in decreasing apoptosis by binding GRP78 and IRE1, thus attenuating the IRE1/ASK1/JNK signaling pathway and expression of CHOP [29]• expressed in CD3+ cells (lymph nodes, thymus and spleen); its absence impairs T cell function and leadsto reduced cytokine production [30]Identified in [31]:• rs6633743 (high regulatory potential)• rs647152 (possible influence on protein function)• rs671976 (probable causative variant or in LD with causative SNPs)RNF215• negatively regulates type 1 IFN [32]• increased RNF215 expression blocksNF-κB p65 binding to the IFNB1 promoter [32], thus preventingIFNβ transcription [33]-ANXA11• regulates apoptosis in immune cells [34]• downregulated in immune cells exposed to activation stimuli [35]Identified in [35, 36]:• rs1049550 (C allele is associated with sarcoidosis risk) and [36];• rs2784773 (strongly associated with sarcoidosis) [35])IL23R• part of the IL23/IL17 signaling pathway leading to the production of various cytokines [37] and involved in granuloma formation [38]• encodes for a subunit of the IL23 receptor [6], which consists of IL12Rβ1 and IL23R and is found on CD4+ memory T cells [39]• differentiation in Th17/Th IL-17 cells and consequent production of cytokines such as IL17 and TNFα is attributed to the binding of IL23 to its receptor on CD4+ naive T cells [37]• IL23R stimulates the production of IFNγ byCD45RO memory T cells [39]• Th17/Th IL-17 cells are increased in peripheral blood and in the broncho-alveolar lavage (BAL) and are found within the granulomas of patients with sarcoidosis [40]• most sarcoidosis-related uveitis are genetically susceptible to IL23R [41];• no efficacy was demonstrated for ustekinumab in pulmonary sarcoidosis compared to placebo [42];• IL23 is upregulated in cutaneous sarcoidosis [43]• rs11465804 and rs11209206 are negatively associated with sarcoid-related uveitis [44];• the frequency of rs11209026 is decreased in patients with sarcoidosis and in patients with sarcoidosis and uveitis [6]• rs117633859 and rs6664119 are associated with reduced serum expression of IL23R [6];• rs12069782 in the IL23R promoter was associated with sarcoidosis in [31]• rs117633859 and rs117282985, upstream IL23R, are significantly associated with sarcoidosis in Japanese population [6]FAM117B• role in oxidative stress by activating KEAP1/NRF2signaling pathway [45, 46]• associated with sarcoidosis in [47]rs6748088 is significantly associated with susceptibility forsarcoidosis in EUR [31]ACOXL• role in oxidative stress (poor evidence) [48]• implicated in fatty acids metabolism [48]--PLCL1• involved in NLRP3 activation [49]• correlated to sarcoidosis in [50]• regulates fibroblasts through activation of NLRP3 inflammatory pathway in rheumatoid arthritis [49]; NLRP3 has been linked to Th17/Treg activation (with Th17 being a key player in the inflammation in sarcoidosis) [51]• the blockade of NLRP3 modifies Th17/Treg balance;• [50] directly correlates NLRP3 to sarcoidosis in granuloma formation-PPARG• involved IFNγ production and Th17 activity [52, 53]• in PPARγ knock-out mice, the instillation of MWCNT allowed the creation of a murine model of sarcoidosis [54, 55]• reduced in alveolar macrophages of patients with pulmonary sarcoidosis [52]• activation of PPARγ inhibits granuloma formation and reduces levels of inflammatory mediators CCL2and osteopontin by decreasing NF-κB [56]-CCL24• involved in Th2 inflammatory response [6] and in pulmonary fibrosis [57]• reduction in CCL24 expression may reduce Th2 inflammatory response favoring Th1 inflammatory response [6]• in systemic sclerosis, the blockade of CCL24 leads to a reduction in lung inflammation and fibrosis [87]• CCL24 is incresed in the BAL of sarcoidosis stage 3 patients compared to patients with Lo¨fgren syndrome[58] and in the acqeous humor of patients with uveits[59]Significantly associated with sarcoidosis in [6] the risk allele rs4728493 (intronic) is associated with a reduced expression of CCL24 and in particular to Lo¨fgren syndromeMAT1A• putative role in oxidative stress [60]--RPS6KA4/MSK2• regulation of inflammation [61]• reduced levels of MSK1 and MSK2 are correlatedwith an increase in TNF, IL-6 and IL-12 [61, 62]-MACROD1• involved in inflammation [63]--ATXN2,SH2B3--Identified in [31]:• rs3184504• rs653178 significantly associated with Lo¨fgren SyndromeLAT,RP11- 231C14.10,RP11-264B17.3• linker for activation of T cells [64]• T cell aberrant activation has been implicated in disease outcome [65]• the role of T cells in the pathogenesis of sarcoidosishas been extensively studied [66]-PUS10• plays a direct role in apoptotic pathways through a positive feedback mechanism involving caspase-3• defects in PUS10 transport or interactions lead toincreased cell survival [67]• BAL fluid lymphocytes from patients with sarcoidosis display a non-apoptotic morphology associated with endogenous caspase-3 activity [68]-ESYT2• lipid transport protein involved in plasma membrane lipid homeostasis [69] that clears diacylglycerol from the plasma membrane leading to down regulation of T-cells activation and reduction of IL-2 levels [70]• involved in fibroblasts migration (art 4)• absence of ESTY2 leads to high levels of diacylglycerol in the plasma membrane, thus increasing the activation of T cells• increases TCR signaling, stimulates CD4+ production of IL2 [70], a Th1-like cytokine whose expression in CD4+ and CD8+ T cells is increased in the BAL fluid of patients with sarcoidosis compared to controls [71]• mice with ESYT2 gene deficiency shows an alteredembryogenic fibroblasts migration [72]-UBASH3A/STS2• involved in reducing the activation and signaling of T cells and in the expression of cytokines such as IL-2 and IFNγ [73]• inhibits NF-κB signaling in type 1 diabetes [74]• STS-2 and STS-1 knock-out mice show an augmented response following T cell receptor (TCR) stimulation, enhanced levels IFNγ and IL2 compared to wild type mice [73]• knock-out for STS2 alone has a weaker role in TCR stimulation and IFNγ production [75]• ZAP70 is a transducer for TCR [76] and defect of UBASH3A/STS-2 and UBASH3B/STS-1 lead to enhanced ZAP70 activation through phosphorylation [73, 76] while defect in STS-2 alone is associated with weak ZAP70 phosphorylation [75]• overexpression of UBASH3A reduces IL2 andNF-κB levels [74]-HOMER2• acts as a negative regulator of T cell activation by modulating TCR and CD28 signaling• plays a role in regulating NF-κB, NFAT, and calcium signaling [77]• following stimulation with anti-CD3, T cells from Homer triple knock-out mice produced IL-2 in amounts 2 to 6 times higher compared to wild type.-SUMO2P• alterations in SUMOylation levels influence the differentiation of regulatory T cells (Tregs) and the transcription of IL-17 in Th17 lymphocytes [78]• although there is no direct evidence that alterations in the SUMO-2P1 pseudogene are associated with immune system dysregulation, SUMOylation itselfhas been implicated in immune regulation-CD19• B cells and Tfh2- and Tfh17-like cells – most effective cell type in supporting B-cell activity, particularly antibody production – may play a role in the occurrence and development of sarcoidosis and other autoimmune conditions [79]• significant lymphopenia involving CD4, CD8, and CD19 positive cells was common in sarcoidosis patients and correlated with disease severity [80]- Genes significantly associated with the phenotype were critically reviewed on the basis of known mechanisms of action, available evidence of a role in sarcoidosis or related phenotypes, and previous identification in GWAS studies (rsids which have been identified both in the literature and in the present study are in bold). Comparison with the very recent study by Yuan et al. [28] could not be performed as complete summary statistic could not be accessed yet • role in suppressing apoptosis induced by endoplasmic reticulum stress in endothelial cells ( [29], not replicated in [30]) • contributes to the maturation and activation of T cells [30] • expressed in lung cells [31]; has a role in decreasing apoptosis by binding GRP78 and IRE1, thus attenuating the IRE1/ASK1/JNK signaling pathway and expression of CHOP [29] • expressed in CD3+ cells (lymph nodes, thymus and spleen); its absence impairs T cell function and leads to reduced cytokine production [30] Identified in [31]: • rs6633743 (high regulatory potential) • rs647152 (possible influence on protein function) • rs671976 (probable causative variant or in LD with causative SNPs) • increased RNF215 expression blocksNF-κB p65 binding to the IFNB1 promoter [32], thus preventing IFNβ transcription [33] Identified in [35, 36]: • rs1049550 (C allele is associated with sarcoidosis risk) and [36]; • rs2784773 (strongly associated with sarcoidosis) [35]) • part of the IL23/IL17 signaling pathway leading to the production of various cytokines [37] and involved in granuloma formation [38] • encodes for a subunit of the IL23 receptor [6], which consists of IL12Rβ1 and IL23R and is found on CD4+ memory T cells [39] • differentiation in Th17/Th IL-17 cells and consequent production of cytokines such as IL17 and TNFα is attributed to the binding of IL23 to its receptor on CD4+ naive T cells [37] • IL23R stimulates the production of IFNγ by CD45RO memory T cells [39] • Th17/Th IL-17 cells are increased in peripheral blood and in the broncho-alveolar lavage (BAL) and are found within the granulomas of patients with sarcoidosis [40] • most sarcoidosis-related uveitis are genetically susceptible to IL23R [41]; • no efficacy was demonstrated for ustekinumab in pulmonary sarcoidosis compared to placebo [42]; • IL23 is upregulated in cutaneous sarcoidosis [43] • rs11465804 and rs11209206 are negatively associated with sarcoid-related uveitis [44]; • the frequency of rs11209026 is decreased in patients with sarcoidosis and in patients with sarcoidosis and uveitis [6] • rs117633859 and rs6664119 are associated with reduced serum expression of IL23R [6]; • rs12069782 in the IL23R promoter was associated with sarcoidosis in [31] • rs117633859 and rs117282985, upstream IL23R, are significantly associated with sarcoidosis in Japanese population [6] • role in oxidative stress by activating KEAP1/NRF2 signaling pathway [45, 46] rs6748088 is significantly associated with susceptibility for sarcoidosis in EUR [31] • role in oxidative stress (poor evidence) [48] • implicated in fatty acids metabolism [48] • correlated to sarcoidosis in [50] • regulates fibroblasts through activation of NLRP3 inflammatory pathway in rheumatoid arthritis [49]; NLRP3 has been linked to Th17/Treg activation (with Th17 being a key player in the inflammation in sarcoidosis) [51] • the blockade of NLRP3 modifies Th17/Treg balance; • [50] directly correlates NLRP3 to sarcoidosis in granuloma formation • in PPARγ knock-out mice, the instillation of MWCNT allowed the creation of a murine model of sarcoidosis [54, 55] • reduced in alveolar macrophages of patients with pulmonary sarcoidosis [52] • activation of PPARγ inhibits granuloma formation and reduces levels of inflammatory mediators CCL2 and osteopontin by decreasing NF-κB [56] • reduction in CCL24 expression may reduce Th2 inflammatory response favoring Th1 inflammatory response [6] • in systemic sclerosis, the blockade of CCL24 leads to a reduction in lung inflammation and fibrosis [87] • CCL24 is incresed in the BAL of sarcoidosis stage 3 patients compared to patients with Lo¨fgren syndrome [58] and in the acqeous humor of patients with uveits [59] • reduced levels of MSK1 and MSK2 are correlated with an increase in TNF, IL-6 and IL-12 [61, 62] Identified in [31]: • rs3184504 • rs653178 significantly associated with Lo¨fgren Syndrome LAT,RP11- 231C14.10, RP11-264B17.3 • T cell aberrant activation has been implicated in disease outcome [65] • the role of T cells in the pathogenesis of sarcoidosis has been extensively studied [66] • plays a direct role in apoptotic pathways through a positive feedback mechanism involving caspase-3 • defects in PUS10 transport or interactions lead to increased cell survival [67] • lipid transport protein involved in plasma membrane lipid homeostasis [69] that clears diacylglycerol from the plasma membrane leading to down regulation of T-cells activation and reduction of IL-2 levels [70] • involved in fibroblasts migration (art 4) • absence of ESTY2 leads to high levels of diacylglycerol in the plasma membrane, thus increasing the activation of T cells • increases TCR signaling, stimulates CD4+ production of IL2 [70], a Th1-like cytokine whose expression in CD4+ and CD8+ T cells is increased in the BAL fluid of patients with sarcoidosis compared to controls [71] • mice with ESYT2 gene deficiency shows an altered embryogenic fibroblasts migration [72] • involved in reducing the activation and signaling of T cells and in the expression of cytokines such as IL-2 and IFNγ [73] • inhibits NF-κB signaling in type 1 diabetes [74] • STS-2 and STS-1 knock-out mice show an augmented response following T cell receptor (TCR) stimulation, enhanced levels IFNγ and IL2 compared to wild type mice [73] • knock-out for STS2 alone has a weaker role in TCR stimulation and IFNγ production [75] • ZAP70 is a transducer for TCR [76] and defect of UBASH3A/STS-2 and UBASH3B/STS-1 lead to enhanced ZAP70 activation through phosphorylation [73, 76] while defect in STS-2 alone is associated with weak ZAP70 phosphorylation [75] • overexpression of UBASH3A reduces IL2 and NF-κB levels [74] • acts as a negative regulator of T cell activation by modulating TCR and CD28 signaling • plays a role in regulating NF-κB, NFAT, and calcium signaling [77] • although there is no direct evidence that alterations in the SUMO-2P1 pseudogene are associated with immune system dysregulation, SUMOylation itself has been implicated in immune regulation For each risk locus in EUR-ancestry meta-analysis, SNPs with CADD score > 12.37 were manually selected and reported in Table 2. In AFR ancestry meta-analysis, only two SNPs displayed CADD score. >12.37 (see Supplementary files). Table 2For each risk locus in EUR ancestry meta-analysis, mapped SNPs with CADD score >12.37 and r2>0.6 are reported under the lead variant (first row in bold)risk locusrsidbpr2GWAS P-valueAnnotationNearest GeneCADD1:67597977:C:Trs20248256759797712.96E-23intronicC1orf1413.432rs11209026677059580.9203591.268E-10exonicIL23R26.5rs11208997675609560.6428510.00226exonicC1orf14112.62rs2224501675325260.6191570.000000004463intergenicSLC35D113.75rs4655514675509210.6509730.001956intergenicC1orf14115.59rs7519768675555220.6291075.933E-10intergenicC1orf14112.99rs34388889675634300.774939NaNintronicC1orf14112.53rs12140736675866710.6312741.427E-10intronicC1orf14113.862:61224259:C:Trs101810426122425911.67E-08intronicPUS100.43rs6710043612182800.7878210.00009042intronicPUS1013.932:111610816:G:Trs1018333811161081611.16E-10intronicACOXL5.723rs155400511159895813.15E-08exonicACOXL17.28rs67325651116078320.898915.95E-09intronicACOXL15.15rs1340181111161610411.83E-08intronicACOXL21.5rs175251471116195220.9197224.95E-08intronicACOXL16.43rs93086901116388980.8281153.69E-08intronicACOXL12.69rs93086921116412630.8281153.57E-08intronicACOXL13.03rs288110271116539370.8345753.11E-08intronicACOXL12.39rs1131228611116576870.858706NaNintronicACOXL14.45rs15138261116607100.8345754.00E-08intronicACOXL13.452:198808578:A:Grs140109019880857811.98E-10intronicPLCL10.618rs7880181982655260.7396530.000002647exonicSF3B118.96rs7880231982833050.7396530.000002588exonicSF3B113.29rs85391983620180.7406390.000001763exonicHSPD118.44rs10642131989502400.7736670.0000001944exonicPLCL126.7rs7880071982336760.7109170.00000405intergenicNPM1P4616.84rs676578121983874010.6987020.00000002455intronicHSPE1–MOB4:MOB412.46rs38385841984167280.74175NaNUTR3MOB415.71rs113369561985407010.901653NaNUTR5RFTN213.6rs7006621986687510.9445760.00000001692upstreamPLCL112.6rs6968171987222250.9540870.000000008587intronicPLCL113.92rs5700361571987239040.954087NaNintronicPLCL112.66rs48508121987436550.9536460.000000008505intronicPLCL115.42rs9389291987808600.9813240.000000001489intronicPLCL112.86rs14642111987986300.9445170.00000001046intronicPLCL112.99rs556961341988215690.9813934.065E-10intronicPLCL112.56rs614356571988255060.6272210.00006861intronicPLCL114.38rs58375761988263630.94NaNintronicPLCL115.33rs133826971989205600.812660.0000003302intronicPLCL115.252:203579866:C:Trs19139091620357986611.16E-12intronicFAM117B1.08rs729250892033595660.9618325.25E-08intronicBMPR217.28rs1409745622034602020.713015NaNintergenicAC009960.115.01rs9339692034997200.646164NaNupstreamFAM117B13.85rs67480882035565260.9733921.39E-12intronicFAM117B13.853:12316339:A:Grs29721661231633911.36E-08intergenicPPARG0.3776:29607101:C:Trs31318562960710113.20E-20intergenicSUMO2P14481rs13195402264635750.8774331.17E-13exonicBTN2A123.7rs13195509264636600.7557872.664E-13exonicBTN2A122.5rs139332558266377240.889731NaNexonicZNF32219.82rs7756481271150690.8790680.00000139exonicHIST1H2AH16.39rs141138864272230640.704895NaNexonicPRSS1615.96rs16897515272780200.9663381.82E-08exonicPOM121L219.9rs200484277756740.8092017.57E-15exonicHIST1H2BL16.23rs200981278331740.7820087.32E-15exonicHIST1H2AL15.9rs200956278397460.9490393.42E-09exonicHIST1H3I16.41rs200973278584210.9849321.88E-09exonicHIST1H3J21.5rs34788973278792000.9472284.47E-16exonicOR2B223.2rs61742093278799820.9472289.87E-16exonicOR2B222.6rs1679709282283420.8442559.92E-11exonicNKAPL22.3rs33932084282688240.9407688.94E-14exonicPGBD114.75rs2230683288911760.925621.21E-17exonicTRIM2718.69rs404240295239570.8979481.45E-15exonicUBD:GABBR114.02rs6902389263221530.7292011.05E-11ncRNA exonicHIST1H3PS113.92rs6908156263228610.7292011.05E-11upstreamHIST1H3PS114.27rs6902392263221150.7292011.05E-11ncRNA exonicHIST1H3PS114.9rs9467740263832500.8721077.73E-12upstreamBTN2A212.85rs3832422263835230.872107NaNUTR5BTN2A213.39rs12207181264312850.8471827.81E-10ncRNA exonicBTN2A3P12.45rs34104395264782520.7557872.82E-13intergenicBTN2A113.34rs13198716265820350.8904331.49E-13intergenicABT112.54rs9393735265823270.7198294.74E-07intergenicABT112.51rs13203358265905780.6749912.01E-07intergenicABT113.55rs6456742266170750.7855482.24E-08intergenicRP11-457M11.612.79rs72845515266468710.7855482.02E-08intronicZNF32212.63rs12198053266676590.6642095.50E-08intergenicZNF32213.55rs12192446269956380.8964443.23e-7intergenicLINC0024013.3rs72838249269984890.9251269.65E-08intergenicVN1R12P13.41rs72838262270085170.9251261.58E-07intergenicVN1R12P13.7rs69343292715803312.40E-08intergenicRP11-209A2.113.14rs12198077271693070.8683640.000001443intergenicRP11-209A2.112.54rs72843629271752150.8683649.54E-07intergenicRP11-209A2.112.52rs67457459271983430.7112191.45E-10intergenicPRSS1613.45rs72843641272033350.8589028.25E-07intergenicPRSS1615.71rs728394772732700018.09E-15ncRNA exonicZNF204P14.36rs34573979274805260.9863282.38E-13intergenicXXbac-BPGBPG34I8.115.35rs56405707276402460.9863281.08E-13intergenicRP1-15D7.113.34rs9357045276889270.6902513.36E-07intergenicRP1-97D16.119.07rs9348774276889300.6902514.66E-07intergenicRP1-97D16.117.07rs13193480277025610.9729836.80E-14intergenicRP1-97D16.115.29rs4713119277128250.9551618.27E-09intergenicRP1-97D16.112.66rs9468225277457190.6096220.000007862intergenicRSL24D1P117.65rs200483277748240.8092015.99e-15upstreamHIST1H4PS113.69rs17751184277750280.9729834.02E-14ncRNA exonicHIST1H4PS114.92rs375576927278353221NaNUTR5HIST1H1B13.52rs200949278354350.7371872.81E-14upstreamHIST1H1B12.63rs45509595278409260.9729831.74E-15UTR3HIST1H4L15.55rs184666393278703241NaNintergenicRNU7-26P17.84rs71559054278967990.9472285.69E-16intergenicOR2W6P14.09rs67040724279055090.9472282.84E-16ncRNA exonicOR2W6P13.54rs148418547279143590.947228NaNintergenicOR2W6P13.54rs28360499279453960.9347965.44E-16ncRNA exonicOR2W4P13.56rs149947279724330.6923620.000009017intergenicIQCB2P13.59rs149900280145970.7089890.00001664ncRNA exonicOR2B7P14.67rs35572414280151110.750238NaNncRNAexonicOR2B7P12.49rs203888280215890.6832850.000008666ncRNA exonicOR2B8P15.75rs35155115280218530.7743580.000006234ncRNA exonicOR2B8P14.64rs188105280713930.7664480.000002204intergenicZSCAN12P115.23rs75874576280878320.9026522.66E-12intergenicZSCAN16–AS112.49rs9468300281268400.666462.69E-10UTR3ZKSCAN813.42rs1736904282192700.8373312.04E-08intronicZKSCAN412.41rs146407472282431751NaNncRNA intronicRP5-874C20.314.64rs13211507282573770.9550511.58E-16intronicPGBD113.79rs36005309282773460.955051NaNintergenicPGBD112.91rs56075693282903280.9697151.24E-16intergenicZSCAN3112.71rs10591593283155420.609775NaNintronicZSCAN3112.96rs357448192831833111.01E-16intronicZSCAN31:ZKSCAN313.57rs78371185283215021NaNintronicZSCAN31:ZKSCAN314.17rs71559081283720690.6601740.00004601intergenicZSCAN1212.97rs13209596283961900.6601740.00004634intergenicZSCAN2313.61rs7765989284002950.7033670.000005804intronicZSCAN2312.95rs11967609286561090.7807053.49E-07intergenicLINC0053313.29rs12193659286771440.8998014.65E-12intergenicRPSAP214.4rs1233583287155660.9268514.83E-17intergenicRPSAP214.21rs9257189287575550.9268512.93E-17intergenicNOL5BP14.92rs2091812879247713.66E-10intergenicXXbac-BPG308K3.512.38rs3135300288243970.9117161.64e-17intergenicXXbac-BPG308K3.613.51rs3135302288296370.9117161.81E-17ncRNA exonicXXbac-BPG308K3.6:RPL13P12.44rs3132392288386290.9117162.25e-17intergenicXXbac-BPG308K3.613.32rs3130746291531550.925625.32E-17intergenicOR2J4P12.86rs31290932917026114.64E-09intergenicOR2H4P13.4rs426595295124820.8654386.44E-08intergenicGPR53P14.95rs1115084442960351211.03E-10downstreamSUMO2P1196:33196397:A:Crs348592173319639711.71E-36intergenicHTATSF1P3.039rs9394145333997780.8818754.646E-13exonicSYNGAP117.25rs4713668336907960.9458191.426E-11exonicIP6K322.1rs1146990533317230611.38E-23upstreamSLC39A713.33rs2114483334385914.33E-22intergenicLYPLA2P112.78rs116323349334012910.6603551.39e-16intronicSYNGAP116.5rs9366821334440740.7769568.45E-20intergenicZBTB914.14rs62405954335248200.8749425.64E-29intergenicBAK112.68rs210122335746350.9801963.85e-12intergenicITPR312.96rs9368768335886100.7526230.0000002741UTR5ITPR315.94rs71565395335935770.9043111.07e-9intronicITPR313.36rs749847336650010.9324432.332E-11UTR3UQCC213.46rs9469566336676050.9494051.76e-11intronicUQCC214.21rs4713659336678390.9494051.948E-11intronicUQCC214.48rs6457739336738310.9494051.937E-11intronicUQCC220.2rs58420296336772140.9324431.103E-11intronicUQCC213.5rs2966336895200.9257071.972E-11UTR3IP6K314.81rs2967336895340.6620910.004241UTR3IP6K312.69rs4713674337081690.8443563.958E-12intronicIP6K316.92rs73747318337820480.7969130.0000002631intergenicMLN12.44rs17600945338022630.880960.000005024intergenicMLN20.9rs7746977338424810.677812.555E-10ncRNA intronicLINC0101614.81rs11758326344824980.8091880.00000002457intronicPACSIN118.957:75453254:G:Trs78116267545325411.13E-11upstreamCCL241.5527:158531641:A:Grs780376615853164112.25E-08intronicESYT21.718rs23054731585362670.9923990.00000005877exonicESYT222.2rs23054751585363450.9923990.0000000623exonicESYT222.7rs10617351585264550.9473050.000001459UTR3ESYT216.15rs8424501585674230.6873580.000001523intronicESYT214.26rs8424521585754400.6873580.000001293intronicESYT212.6110:81926702:A:Grs10495508192670219.08E-53exonicANXA1128.6rs12411782820129670.704490.000003082exonicAL359195.118.65rs25733608189173114.134E-10intronicPLAC914.58rs2819885818920020.657681.154E-11intronicPLAC912.69rs2819875819035170.7823274.26E-39intronicPLAC912.57rs3851053819185820.6359244.48E-28intronicANXA1114.57rs141872729819279350.979479NaNintronicANXA1115.34rs11202248819914750.9303163.77E-11intergenicRP11-40F6.113.82rs60492867820225600.88442NaNintergenicMAT1A14.14rs10788579821086100.9623072.57E-14intronicDYDC1:DYDC213.51rs35050604821966770.9561714.65E-14ncRNA exonicRP11-137H2.612.39rs10887881822046200.9872033.29E-10intergenicRP11-137H2.615.41rs7095954822092320.841765.37E-08intergenicTSPAN1412.88rs116857892822140650.841764.91E-08UTR5TSPAN1413.16rs12220642822189640.7079370.000001754intronicTSPAN1412.711:64107735:A:Grs6637436410773512.38E-23UTR5CCDC88B5934rs647152641091180.8071751.64E-18exonicCCDC88B13.15rs600377641249400.6829945.377E-11exonicCCDC88B14.93rs521950641277440.6829943.66E-11exonicRPS6KA414.41rs11542299641388050.8760144.45E-15exonicRPS6KA416.99rs60031276640129100.7477942.52E-15UTR5PPP1R14B18.33rs35989122640116140.992329NaNdownstreamFKBP214.87rs11231727640118540.909930.0000001063downstreamPPP1R14B12.39rs2510066640524470.808364.22E-18UTR5GPR13713.7rs542907641249800.6829941.097E-10UTR3CCDC88B16.0112:112007756:C:Trs65317811200775611.20E-23intronicATXN20.312rs79689601114266150.799158.10E-09intergenicRP1-46F2.215.51rs108499251114955180.8176970.0000001583intronicCUX216.4rs43784521115040330.8256540.00000006634intronicCUX213.24rs73991131117928480.624650.00001513intergenicCUX217.4915:41202608:C:Trs1819560654120260819.59E-09intergenicRP11-540O11.115.1115:83515292:A:Crs118563168351529211.32E-08UTR3HOMER22.07416:28951457:C:Trs116453022895145719.45E-12downstreamCD191.70719:10463118:C:Grs345364431046311811.95E-11exonicTYK225.5rs74956615104277210.8283255.24E-09UTR3CTD-2369P2.12:RAVER120.821:43855067:A:Crs18935924385506716.01E-09intronicUBASH3A11.1622:30776419:C:Trs7578703077641914.34E-09intronicRNF2150.304rs757660307931370.8003512.38E-08exonicRNF215:SEC14L219.14rs1860217306855600.9754589.45E-09UTR5GATSL3:RP1-130H16.1815.03rs2240422307009240.9803241.15E-08intronicTBC1D10A13.35 For each risk locus in EUR ancestry meta-analysis, mapped SNPs with CADD score >12.37 and r2>0.6 are reported under the lead variant (first row in bold) Gene-set enrichment analysis Interferon gamma signaling, meiotic recombination/condensation of prophase chromosomes, and DNA methylation were the most enriched gene sets in EUR (Fig. 8) and multi-ancestry meta-analysis (Supplementary Figure S3). PD1 and INFγ signaling, generation of second messenger molecules, and MHC class II antigen presentation were the most enriched gene sets in AFR ancestry (Supplementary Figure S4). Gene-set enrichment analysis for EUR ancestry Discussion In this study, we present a GWAS meta-analysis for sarcoidosis in European and African ancestries, aggregating publicly available summary statistics from the most recent releases of some of the largest genome biobanks. Moreover, our multi-ancestry meta-analysis is the first to include populations of African, European, and East Asian ancestries together. To date, only few GWAS have selectively focused on sarcoidosis so far [6, 31]: Fisher et al. identified 4 risk loci, whereas Meguro et al. identified 3 risk loci. Two risk loci from both studies overlap with our findings (nearest genes: ATXN2/SH2B3 and FAM117B in [31]; nearest genes: CCL4 and C1orf141-IL23R in [6]). Very recently, Yuan et al. have presented a meta-analysis in which they studied in depth the same phenotype [28]. Our work, which was conducted independently, aims to add new evidence to their study, as we employed different techniques and accessed different summary statistics. For instance, we included the results from the BioBank Japan project and used the latest Finngen release, whose summary statistics are the most informative concerning this phenotype. Compared to Yuan et al., our analysis appears to be more restrictive, as we found less risk loci [28]. Our results, however, are comparable to the multi-ancestry meta-analysis issued by FinnGen and including patients from the MVP and UKBB [81]. In this regard, it is of interest that at least three risk loci (lead SNPs: rs10181042, rs7803766, rs181956065; nearest genes: PUS10, ESYT2, RP11-540O11.1) have been identified in our study but not reported elsewhere (see Supplementary file 1). For functional annotation and subsequent analysis, European ancestry yielded the highest number of identifiable loci. In fact, the results from African ancestry meta-analysis mostly coincided with the single summary statistics from the MVP study, and only two known risk loci could be observed - although a potential trend was highlighted at least for a locus in chromosome 1, in the IL23R region. This is likely due to low sample numbers in controls, as well as - and especially - in cases of AFR ancestry. This highlights the urgent need to extend efforts toward larger case-control populations in AFR (and EAS) ancestries, compared to European populations. It is worthwile to note that multi-ancestry meta-analysis confirmed some statistically significant associations for trending loci in European ancestry, in particular in chromosomes 1 (lead SNP 1:67626416:C:G), 3 (lead SNP 3:12471871:C:T), 6 (lead SNP 6:34504042:G:T), 10 (lead SNP 10:64454497:C:T), and 12 (lead SNP 12:112007756:C:T). Meanwhile, some risk loci in chromosomes 7 (lead SNP 7:75453254:G:T), 15 (lead 15:41202608:C:T), 19 (lead SNP 19:10463118:C:G), 21 (lead SNP 21:43855067:A:C), and 22 (lead SNP 22:30776419:C:T) appeared private to EUR ancestry, as they lost statistical significance when including multiple ancestries (Figs. 6 and 4). It is particularly interesting that tissue-enrichment analysis in European ancestry (Fig. 7) highlights a significantly enriched expression in tissues (whole blood, spleen) involved in immunitary functions and activity of the reticuloendothelial system. Additionally, TSEA shows enrichment in the lung, the primary site affected by the disease, as well as in other tissues—including the terminal ileum, visceral omentum, and mammary tissue—whose involvement in sarcoidosis remains less clear. Crucially, such findings are lost when performing multi-ancestry analysis (Figure S1), perhaps due to LD distortions caused by the coexistence of ancestry-specific r2 values. In both European and African ancestries, mild-moderate genetic heritability of the trait has been evidenced, corroborating the assumption that sarcoidosis is a predominantly acquired autoimmune disease, to which some familiar risk appears to have an impact nonetheless. Gene-set enrichment analysis in EUR and multi-ancestry meta-analysis yielded some particularly interesting results, as the most represented metabolic pathways in both analyses were related to interferon gamma signaling and DNA methylation. Interferon-γ (IFNγ) signaling plays a pivotal role in driving the Th1-dominant immune response. Produced mainly by NK cells and Th1 lymphocytes, IFN-γ enhances macrophage activation and promotes the release of chemokines, facilitating the recruitment of Th1/17 cells, monocytes, Tregs, and B cells to inflamed lung tissue. It has been shown that this persistent immune activation fosters granuloma formation and sustains an exaggerated inflammatory response [82]. The hypothesis of methylation alterations as a risk factor for developing sarcoidosis is particularly intriguing, as this would adapt well to the aforementioned disease model in which both acquired and genetic risk factors coexist. Growing evidence highlights the role of epigenetic regulation in sarcoidosis [83]. Transcriptional changes have been observed in peripheral blood cells [84, 85] and transbronchial biopsies [86], suggesting an epigenetic influence on disease mechanisms. DNA methylation and histone modifications play a role in the differentiation of CD4+ and regulatory T (Treg) cells [87, 88], potentially shaping immune responses. Additionally, genes involved in chromatin remodeling, such as HDAC and other chromatin-modifying factors, have been linked to the disease [86]. In this perspective, our findings could be seen as further evidence supporting a role for epigenetics in the pathogenesis of sarcoidosis. Be that as it may, caution must be exercised when interpreting these results, as the histone protein genetic region is widely overposed to the MHC.Therefore, over-representation due to some degree of LD cannot be ruled out (although, only in GSEA the MHC region was included in the analysis as an attempt to account for LD distortions). Among the non-MHC genes that were represented in the risk loci (Table 1), at least 9 have been previously associated to sarcoidosis. In particular CCDC88B, ANXA11, IL23R, FAM117B, CCL24, and ATXN2/SH2B3 have already been identified in previous GWAS or other genetic studies. Among those, CCDC88B and ANXA11 are potential regulators of apoptosis [29, 34], whereas CCL24 was increased in broncho-alveolar lavage (BAL) of patients with stage 3 sarcoidosis compared to patients with Lo¨ fgren syndrome [58] and IL23R has been implicated in the formation of granulomas [38]. Concerning the latter, it is worth noting that previous attempts to use ustekinumab, an anti IL-12 and IL-23 monoclonal antibody, have been unsuccessful in the management of sarcoidosis [42]. Of great interest, PLCL1, PPARG, CD19 have been involved in sarcoidosis through functional studies [49–56]. These factors are probably the most promising for further investigations as possible targets for therapeutic strategies. Finally, at least 8 additional genes have been implicated in immune regulatory response, whereas no clear evidence linking ACOXL, MAT1A, or MACROD1 to sarcoidosis or immune (dys)regulation was found from the available literature. Further studies are required to define a possible involvement of these factors in the pathogenesis of the disease, which if confirmed would pave the way to previously unforeseen clues about the pathogenesis and therapeutic options for sarcoidosis. Among the limitations of this study, all those stemming from a meta-analysis of public summary statistics must be taken into account: in particular, as access to individual-level data was not possible, further delving into genomic analysis (e.g. through validation of polygenic risk scores or Mendelian randomisation) was not feasible. Furthermore, it is important to note that the available phenotype information for patients only stated whether they were affected by sarcoidosis, but did not specify other clinically relevant details such as the targeted organs or phenotype severity. Finally, although slightly different results can be obtained depending on the choice of the technical parameter to use for meta-analysis and functional annotation, the core results of the study have been robust even after repeating calculations with different specifics. Additionally, although further analysis (e.g. chromatin analysis) could be implemented through the available platforms, we put our best efforts into prioritising the most relevant information, with a focus on clinical interpretation of the results. Conclusion We present a GWAS meta-analysis for sarcoidosis in European and African ancestries, obtained from the most recent releases of publicly available summary statistics of the FinnGen, MVP, pan-UKBB, and BioBank Japan projects. Nineteen and two risk loci were significantly associated with sarcoidosis for European and African ancestries, respectively. Mild-moderate genetic heritability was observed for both ancestries. From comparison between European and multi-ancestry meta-analysis, novel risk loci private to EUR ancestry could be hypothesised; on the other hand, trending loci in EUR ancestry confirmed their significance when aggregating ancestries. Crucially, more efforts are needed to collect data from African and East Asian ancestries, as the current sample sizes in both control and - especially - cases are likely insufficient to obtain novel ancestry-specific information through meta-analysis. Critical revision and comparison with the available literature has highlighted several genes as putatively associated with the pathogenesis of sarcoidosis, thus paving the way to further investigation as possible disease-altering factors or therapeutic targets. Although further studies are warranted, epigenetic alterations may contribute to the risk of developing the phenotype. We hope that the interpretation of these new data will contribute to a more precise understanding of the etiology and pathophysiology of sarcoidosis at the molecular level, with the potential to aid in the development of more personalized approaches for diagnosis, prognosis, and treatment. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 Acknowledgements We want to acknowledge the participants and investigators of the FinnGen study, UK BioBank and pan-UKBB project, Million Veteran Program, and BioBank Japan. Abbreviations 1000 genomes (Project Phase 3 reference panel) African (ancestry) broncho-alveolar lavage Biobank Japan Combined Annotation-Dependent Depletion Genotype- Tissue Expression (progect) East Asian (ancestry) expression quantitative trait locus/i European (ancestry) Functional Mapping and Annotation (tool) gene-set enrichment analysis genome-wide association study linkage disequilibrium (score) major histocompatibility complex Million Veteran Program single nucleotide polymorphism tissue expression analysis T cell receptor UK Biobank whole exome sequencing whole-genome sequencing Author contributions Andrea Ricci: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing. Dario Andrisani: Supervision, Writing – review & editing. Antonello Pietrangelo: Methodology, Supervision, Writing – review & editing. Brent Julius De Guzman Marinduque: Data curation, Formal Analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Chiara Vecchi: Conceptualization, Methodology, Supervision, Writing – review & editing. Daniele Sabbatini: Conceptualization, Formal Analysis, Investigation, Methodology, Software, Supervision, Writing – review & editing. Dario Gregori: Methodology, Supervision, Writing – review & editing. Elena Buzzetti: Methodology, Supervision, Writing – review & editing. Elena Corradini: Conceptualization, Methodology, Resources, Supervision, Writing – review & editing. Elena Pegoraro: Methodology, Supervision, Writing – review & editing. Elisa Bergamini: Conceptualization, Methodology, Supervision, Writing – review & editing. Enrico M. Clini: Conceptualization, Project administration, Resources, Supervision, Writing – review & editing. Federica Andolfi: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. Filippo Gozzi: Conceptualization, Investigation, Methodology, Supervision, Writing – review & editing. Giada Di Betto: Conceptualization, Formal Analysis, Investigation, Supervision, Validation, Writing – review & editing. Paolo Ventura: Writing – review & editing. Roberto Tonelli: Conceptualization, Methodology, Resources, Supervision, Writing – review & editing. Stefania Cerri: Conceptualization, Investigation, Methodology, Project administration, Supervision, Writing – review & editing. Funding The authors declare that no financial support was received for the research of this article. The publication of this article was funded by the Italian funds for Departments of Excellence (project 020146_23_DE_DIPECC_23-27_DIDATT_ALTA_QUALIFICAZ). Data availability R markdown sheets are available in Supplementary Files. Further material can be made available upon reasonable request to the Authors. Declarations Ethics approval and consent to participate This study is a meta-analysis of publicly available genome-wide association study (GWAS) summary statistics. No new data were collected, and no direct interaction with human subjects occurred. All analyses were conducted on de-identified data. The original contributing cohorts obtained ethical approvals and informed consent from all participants. UK Biobank: approved by the North West Multicentre Research Ethics Committee (REC reference 11/NW/0382); UK Biobank: approved by the North West Multicentre Research Ethics Committee (REC reference 11/NW/0382); FinnGen: approved by the Coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa (approval number HUS/990/2017); FinnGen: approved by the Coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa (approval number HUS/990/2017); Million Veteran Program (MVP): approved by the VA Central Institutional Review Board; Million Veteran Program (MVP): approved by the VA Central Institutional Review Board; BioBank Japan (BBJ): approved by the research ethics committees at the Institute of Medical Science, the University of Tokyo, the RIKEN Yokohama Institute, and the cooperating hospitals. All contributing studies were conducted in accordance with the Declaration of Helsinki and relevant national and institutional ethical regulations. BioBank Japan (BBJ): approved by the research ethics committees at the Institute of Medical Science, the University of Tokyo, the RIKEN Yokohama Institute, and the cooperating hospitals. All contributing studies were conducted in accordance with the Declaration of Helsinki and relevant national and institutional ethical regulations. Consent to publish All the authors have approved the manuscript and consented to publish. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Footnotes Andrea Ricci and Federica Andolfi co-first author. Elena Corradini and Stefania Cerri co-last author. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Andrea Ricci, Email: andrea.ricci@unimore.it. Filippo Gozzi, Email: filippo.gozzi@unimore.it. References Associated Data Supplementary Materials Supplementary material 1 Data Availability Statement R markdown sheets are available in Supplementary Files. Further material can be made available upon reasonable request to the Authors.