Novel immune-related susceptibility loci associated with pediatric steroid-sensitive nephrotic syndrome identified by a transethnic genome-wide association study Authors: Xueying Yang, Zijian Tian, Xueting Chen, Jian Shao, Li Wang, Fenfen Ni, Rui Liang, Bo Zhao, Shipin Feng, Xiaojie Gao, Jingjing Cui, Xia Yang, Xuelan Chen, Xiaoqin Li, Baohui Yang, Junli Wan, Jia Jiao, Anshuo Wang, Qin Yang, Xuejun Yang, Daoqi Wu, Gaofu Zhang, Mo Wang, Kaixin Zhou, Han Chan, Haiping Yang, Qiu Li Published: 2026 ABSTRACT ABSTRACTBackgroundSteroid-sensitive nephrotic syndrome (SSNS) is a relatively rare glomerular disease with considerable clinical and genetic heterogeneity. Although efforts have been made to explore the genetic determinants of SSNS, the underlying genetic architecture has still been hampered by limited sample size and remains to be elucidated.MethodsWe conducted a genome-wide association study involving 510 sporadic pediatrics of SSNS and 1062 controls with Chinese ancestry, followed by a transethnic meta-analysis involving 2942 cases and 37 041 controls. We evaluated the colocalization probability between single-cell expression quantitative trait loci, plasma protein quantitative trait loci, and the novel loci using the colocalization method. We explored the potential causal effects using Mendelian randomization (MR) and assessed the clinical associations of polygenic risk scores (PRS) for SSNS.ResultsTwo novel genome-wide significant signals, rs4309200 from M6PR-KLRG1 (P = 4.70 × 10−9, OR = 1.23, 95%CI = 1.16–1.30) and rs9303279 from GSDMB-ORMDL3 (P = 7.25 × 10−9, OR = 0.82, 95%CI = 0.75–0.89), both highly related to immunological function, were identified for SSNS. Colocalization and MR analyses revealed that a genetic tendency toward higher M6PR expression in NK and CD4+ T cells with an effector memory or central memory phenotype (CD4 ET) was associated with an increased risk of SSNS (NK: PPH4 = 0.93; P = 2.04 × 10−4, OR = 1.60; CD4 ET: PPH4 = 0.91; P = 8.14 × 10−3, OR = 2.28), whereas higher ORMDL3 expression in B memory and CD4 ET cells was associated with a lower risk of SSNS (B memory: PPH4 = 0.93; P = 3.83 × 10−8, OR = 0.61; CD4 ET: PPH4 = 0/92; P = 1.82 × 10−9, OR = 0.49). The clinical correlation analysis confirmed that a higher PRS is associated with an earlier age of onset, and the rs56117924 from the NPHS1 gene may predominantly drive this association.ConclusionWe identified two novel immune-related loci associated with susceptibility to SSNS, nominating M6PR and ORMDL3 as potential contributors to SSNS risk in a cell-type-specific manner, and confirmed the association of PRS with earlier disease onset. These findings expand our knowledge of the genetic architecture of SSNS and highlight the importance of large-scale collaborations in uncovering genetic associations related to SSNS.Keywords: genome-wide association study, Mendelian randomization, polygenic risk scores, quantitative trait loci, steroid-sensitive nephrotic syndrome What was known: Pediatric SSNS is a relatively rare glomerular disease with a risk of end-stage kidney disease. Pediatric SSNS is a relatively rare glomerular disease with a risk of end-stage kidney disease. Although the latest GWAS meta-analysis of pediatric SSNS identified four novel significant associations across multiple populations, these identified loci explain only a small fraction of overall heritability. Although the latest GWAS meta-analysis of pediatric SSNS identified four novel significant associations across multiple populations, these identified loci explain only a small fraction of overall heritability. Genetic association studies in pediatric SSNS have been hampered by insufficient sample size, limited power, and a high degree of heterogeneity. Genetic association studies in pediatric SSNS have been hampered by insufficient sample size, limited power, and a high degree of heterogeneity. This study adds: Genome-wide significant association study identified rs4309200 from the M6PR-KLRG1 and rs9303279 from the GSDMB-ORMDL3 regions as two novel susceptibility loci for pediatric SSNS. Genome-wide significant association study identified rs4309200 from the M6PR-KLRG1 and rs9303279 from the GSDMB-ORMDL3 regions as two novel susceptibility loci for pediatric SSNS. Integrative multi-omics analysis provide evidence supporting a potential causal role of novel SSNS genetic risks in influencing M6PR and ORMDL3 gene expression in a cell-type-specific manner. Integrative multi-omics analysis provide evidence supporting a potential causal role of novel SSNS genetic risks in influencing M6PR and ORMDL3 gene expression in a cell-type-specific manner. Polygenic risk stratification demonstrated a dose-dependent association between allelic burden and earlier disease onset, with rs56117924 in the NPHS1 gene predominantly driving this association. Polygenic risk stratification demonstrated a dose-dependent association between allelic burden and earlier disease onset, with rs56117924 in the NPHS1 gene predominantly driving this association. Potential impact: Two additional novel associations in the immune regulatory gene regions of M6PR-KLRG1 and GSDMB-ORMDL3 further support the role of immune dysregulation in the pathogenesis of SSNS. Two additional novel associations in the immune regulatory gene regions of M6PR-KLRG1 and GSDMB-ORMDL3 further support the role of immune dysregulation in the pathogenesis of SSNS. Despite limited clinical correlations with PRS and individual genetic loci, the replicated association of higher PRS with younger age of onset in this independent Chinese cohort supported the possibility of uncovering more clinical findings with larger sample sizes and significant variants. Despite limited clinical correlations with PRS and individual genetic loci, the replicated association of higher PRS with younger age of onset in this independent Chinese cohort supported the possibility of uncovering more clinical findings with larger sample sizes and significant variants. INTRODUCTION Steroid-sensitive nephrotic syndrome (SSNS) is a relatively rare glomerular disease in children, characterized by significant clinical and genetic heterogeneity across populations. The incidence of SSNS ranges from 1.15 to 16.9 cases per 100 000 children, with notable variations observed based on ethnicity and geographical location [1]. Children of Asian descent have a higher incidence of SSNS compared with those of European descent, yet they tend to exhibit more favorable treatment responses and outcomes [2]. The observed disparities are likely driven by a combination of biological, genetic, and environmental factors, although the exact mechanisms remain largely unidentified. Various studies indicate that immunological disruption and circulating pathogenic factors are significant contributors to the biological pathogenesis of SSNS [3]. Specifically, systemic T-cell and B-cell dysregulation are proposed as potential immune mechanisms that may predispose individuals to the disease [4, 5]. A recent large-scale genome-wide association study (GWAS) meta-analysis involving 2440 individuals with SSNS identified novel significant associations across diverse populations, enhanced the comprehension of population heterogeneity from a genetic perspective, and highlighted the limitations of genetic association studies in SSNS due to inadequate sample sizes [6]. However, the genetic risk factors associated with SSNS in children of Chinese ancestry have not been reported. Despite GWAS excelling in identifying single-nucleotide polymorphism (SNP) associations of diseases, it falls short in pinpointing the exact functional effects of disease-associated variants [7]. The recent emergence of quantitative trait loci (QTL) studies, particularly protein QTL (pQTL) and single-cell expression QTL (sc-eQTL), has ushered in a new era in molecular profiling of complex and heterogeneous diseases by providing cell-type-specific biological insights into the associations between genetic variants and the genes they modulate [8–10]. In addition, Mendelian randomization (MR) [11], a novel genetic statistical approach for causal inference that employs genetic variants as instrumental variables to assess the influence of modifiable exposures, has been extensively used to elucidate the underlying mechanisms of diseases. In this study, a GWAS was performed in the Chinese Han population, followed by a meta-analysis to uncover additional genetic loci associated with SSNS. The cell-type-specific eQTL effects and potential causal effects of these novel loci were further assessed and estimated by integrating eQTL and pQTL from large-scale transcriptomic and proteomic investigations, and the MR approach. Last, an SSNS polygenic risk score (PRS) was generated using summary statistics from this transethnic meta-GWAS, and its clinical association was validated. MATERIALS AND METHODS Cohorts Patients with SSNS were recruited from six hospitals in China between July 2018 and December 2022. All patients enrolled in this study met the standard diagnosis of childhood SSNS, and those with secondary steroid-resistant nephrotic syndrome were excluded [12]. Detailed cohort inclusion and exclusion criteria followed our previously published protocols [13]. In addition, 1062 healthy control participants were recruited from the Children’s Hospital of Chongqing Medical University. Genotyping, quality control (QC), and imputation DNA was extracted from blood samples using the Blood Genome Column Medium Extraction Kit (Kangweishiji, China). All samples were genotyped with the CAS Array, which was designed for Chinese populations. The detailed procedure of genotyping has been described in our previous study [14]. After genotype calling, genetic data were processed using PLINK v.1.90 and v.2.00 according to the following QC steps [15, 16]. For individual-level, individuals with sex mismatch, abnormal heterozygosity, genotyping rate <95%, and identity by descent >0.1875 were excluded. For variant-level, variants with calling rates <98%, minor allele frequency (MAF <5%, and deviation from Hardy–Weinberg equilibrium (P < 10−6) in control participants were excluded. Principal component analysis (PCA) was performed to assess population stratification. Whole-genome imputation was performed with Beagle5.4 using the 1000 Genomes Project high coverage WGS as the reference panel. [17] The INFO value was calculated by bcftools v.1.21 with the “impute-info” plugin. Post-imputation QC was performed based on MAF < 5%, INFO < 0.3, and deviation from Hardy–Weinberg equilibrium. Supplementary data 1 displays the INFO and DR2 (empirical r2) of variants that reached suggestive genome-wide significance in the GWAS analysis, to demonstrate the imputation quality. The flowchart of the study design and the data processing is illustrated in Fig. 1. Flowchart of study design and the data processing for the current GWAS. CR, call rate; Het, heterozygosity; IBD, identity by descent; HWE, Hardy–Weinberg equilibrium. Genome-wide association analysis and conditional analysis A total of 1520 individuals and 6 952 003 SNPs were utilized for further association testing by fastGWA implemented in GCTA with the mixed linear model [18, 19]. The sex and first five principal components (PCs) were utilized as covariates in the association test. The genome-wide significance threshold was set to P < 5.0 × 10−8. Conditional association analysis was performed, including the lead SNP as the covariate. Manhattan and quantile–quantile (Q-Q) plots were plotted under R v.4.3.1. Linkage disequilibrium (LD) of this Chinese local cohort was computed using PLINK v.1.90 to evaluate the LD relationship of novel loci with previously reported loci. The regional plots were generated by LocusZoom with 1000 Genomes Project EAS as an LD reference. Furthermore, the LD heatmaps were generated by LDBlockShow software from original VCF files of genotype data within the region of chromosome 6: 32.580–32.675 Mb owing to the complex LD structure of the human leukocyte antigen (HLA) region between populations [20]. HLA fine-mapping The CookHLA v.1.0.1 and Beagle v.5.1 were used to conduct HLA fine-mapping with the HAN-MHC reference panel [21–23]. Supplementary data 2 displays the confidence score of the HLA-imputed genotype for each individual, which were calculated as the sum of the posterior probabilities of the two alleles for a heterozygous call. The HLA Analysis Toolkit was used to perform HLA association with a logistic regression model adjusted for sex and the first five PCs [24]. Also, 112 HLA classical alleles and 575 HLA amino-acid polymorphisms were eventually filtered by MAF > 0.01 to the association test from 172 imputed 4-digit HLA classical alleles and 6772 amino-acid polymorphisms. Statistical significance thresholds were adjusted with Bonferroni correction for multiple comparisons (Padjust< .05/112 for 4-digit HLA classical alleles and Padjust< .05/575 for HLA amino-acid polymorphism tests). Transethnic meta-analysis Transethnic, inverse-variance weighted, fixed-effects meta-analyses were performed in METAL [25] using summary statistics, respectively from a multi-population meta-analysis of 2440 cases and 36 023 controls from Admixed American, African, East Asian, European, Maghrebian, and South Asian populations, and a study of the Japanese population of 987 childhood SSNS patients and 3206 healthy control participants [6, 26]. Data preparation for METAL involved harmonizing SNP identifiers, alleles, and effect sizes across studies to ensure consistency. The genomic coordinates of GWAS summary statistics were transformed to the GRCh38/hg38 reference genome using the liftOver tool (http://genome.sph.umich.edu/wiki/LiftOver). The regional plots were constructed by LocusZoom with the 1000 Genomes Project EAS and the 1000 Genomes Project (all ancestries) as LD reference populations, respectively [27]. Colocalization analysis Colocalization analyses were performed between novel loci identified in this GWAS, and (i) cell-type specific eQTLs and (ii) plasma pQTLs within ±1 Mb of the novel loci, respectively, to investigate potential causal variant and assess whether gene expression and/or protein levels shared the same causal variant with SSNS susceptibility. Summary statistics of single-cell expression QTL (sc-eQTLs) for 14 immune cell types were obtained from the OneK1K cohort. This dataset was derived from 1 267 758 peripheral blood mononuclear cells collected from 982 healthy individuals of Northern European ancestry (https://onek1k.org/) [28]. Summary statistics of plasma pQTLs were obtained from two large-scale proteomic studies: the UKB-PPP (http://ukb-ppp.gwas.eu) and the deCODE study (https://decode.com/). The UKB-PPP study performed proteomic profiling on plasma samples from 54 219 UKB participants using the Olink Explore 3072 platform, capturing 2923 unique proteins. [9] The deCODE study conducted GWAS of plasma protein levels using 4907 aptamers in a cohort of 35 559 Icelanders [10]. The coloc.abf function from the coloc R package (v.5.2.3) was utilized with default prior probabilities (p1 = p2 = 1 × 10−4, p12 = 1 × 10−5). Significant colocalization was defined as a posterior probability for a shared causal variant (PPH4) ≥0.9. Colocalization signals between the GWAS and QTL data were visualized using the locuszoomr package [29]. Mendelian randomization (MR) MR analyses were performed using the TwosampleMR R package [30] for gene expression and/or plasma protein levels that showed significant colocalization with novel loci. For sc-eQTLs/pQTLs with only one SNP available, Wald ratios were calculated, whereas the inverse-variance weighted (IVW) method was employed for those containing multiple SNPs. When heterogeneity was present, the weighted median method was employed as an alternative to the IVW [31]. Heterogeneity and pleiotropy were assessed through Cochran’s Q test, MR-Egger regression, and the MR-PRESSO global test to evaluate result robustness. When these sensitivity analyses were unavailable, the Steiger directionality test was performed. Polygenic risk score (PRS) The PRS of SSNS individuals was calculated using the R package bigsnpr, which used the classical clumping and thresholding (C + T) approach with the effect sizes derived from the transethnic meta-GWAS summary statistics (a total of 2942 cases and 37 041 controls) from this study [32]. Clumping was done with 0.2 as the r2 threshold and a window size of 500 bp. Variants with P < 5.0 × 10−8 or P < 1.0 × 10−5 were respectively selected for calculation. After clumping and thresholding, 99 or 200 variants were included. The PRS was calculated by adding up the dosage of these variants, weighted by the beta coefficient from the GWAS summary statistics, for each individual. Statistical analysis Statistical analysis of clinical data was conducted under R v.4.1. The t-test was used to compare normally distributed variables (mean ± standard deviation). Non-parametric variables were shown as median with interquartile range (P25, P75) and compared using the Mann–Whitney U-test. Linear regression analyses, adjusting for five PCs and sex, were performed on genome-wide significant loci with the clinical parameter separately to explore genotype–phenotype correlations. Significance thresholds for multiple comparisons were adjusted by Bonferroni correction (Padjust< .05/6 for MR and Padjust< .05/10 for genotype–phenotype tests). RESULTS The Chinese population GWAS results A total of 510 sporadic pediatric patients with SSNS from six medical centers in China and 1068 healthy controls with no prior history of renal disorders were enrolled. The demographic characteristics, laboratory results, histological findings, and follow-up information are presented in Supplementary Table S1. The PCA, Q-Q plot, and inflation factor (λGC = 1.03) indicated no population stratification and genomic inflation (Supplementary Fig. S1, Fig. 2). Manhattan plot and quantile–quantile plot for the GWAS in Chinese pediatric SSNS. Manhattan plot, x-axis indicates chromosomal position, y-axis indicates log-transformed P values, red horizontal line shows genome-wide threshold of P = 5 × 10−8. Quantile–quantile plot (Q–Q plot), the x-axis indicates expected −log10 P values, y-axis indicates observed −log10 P values from GWAS. λGC denotes the calculated inflation factor. The results of the Chinese population GWAS are summarized in Fig. 2, revealing a significant association in the HLA-DRB1/DQA1 region [rs35372932 (effect allele C), P = 2.37 × 10−9, OR = 0.54, 95% confidence interval (95% CI) 0.34–0.74]. All significant SNPs were in high LD and clustered within the same LD block (Fig. 3). Conditional analysis on rs35372932 confirmed its role as the lead SNP (Supplementary Fig. S2). Besides, rs35372932 exhibited medium-low LD with rs9271602 (R² = 0.30) and rs9391784 (R² = 0.20) (identified by Downie et al. [33]) (Supplementary Table S2) and possibly forming an LD block with rs9271602 and rs1063355 (identified by Barry et al. [6]) based on subsequent conditional analysis (Supplementary Fig. S3). The next fine-mapping analysis of the HLA region revealed a protective effect in HLA-DQA1*06:01 (P = 1.08 × 10−4, OR = 0.44, 95% CI 0.29–0.67) (Supplementary Fig. S4, Supplementary Table S3). While no other loci or alleles reached genome-wide significance, the findings largely replicated previously reported signals and HLA alleles or amino-acid polymorphisms in this GWAS (Supplementary Table S4) [6, 26, 33–36]. Rs447783 in the NPHS1 region, as well as amino-acid residue haplotypes at HLA-DQA1 positions 47 and 52, were significant for independently replicating higher risk in East Asians (Supplementary Figs S4 and 5). Locus zoom plot for regions on chromosomes 6p21.32 identified in the Chinese cohort. (a) Locus zoom plot for regions of chr. 6:32.4–32.8 Mb. The x-axis indicates chromosomal position, left y-axis indicates log-transformed P value, right y-axis indicates recombination rate. Purple diamond indicates SNP with the smallest P value within each region. Locus zoom plot was created by LocusZoom (https://my.locuszoom.org) using 1000 Genomes Project EAS as LD reference. (b) Haplotype map and LDBlock of the HLA-DRB1/DQA1 region by LDBlockShow software from original VCF files of genotype data within the region of chr. 6: 32.580–32.675Mb. Two novel associations from the transethnic meta-analysis A fixed-effects meta-analysis of pediatric SSNS was undertaken utilizing the latest published report from a multi-population meta-analysis by Barry et al. [6]. Two new genome-wide significant loci were discovered: rs4309200 (A allele with OR 1.23, 95%CI 1.16–1.30, P = 4.70 × 10−9) at the 12p13.31 region of the M6PR-KLRG1 and rs9303279 (C allele with OR 0.82, 95% CI 0.75–0.89, P = 7.25 × 10−9) at the 17q21.1 region of the GSDMB-ORMDL3 (Fig. 4 and Table 1). In light of the previously observed population-driven heterogeneity, specifically the presence of risk loci specific to Europeans and East Asians in SSNS, we conducted a separate meta-analysis using the Chinese and Japanese cohorts, which represent East Asians, to also identify associations that may be masked by population heterogeneity [26]. Two additional genome-wide significant loci were discovered in comparison to the original Japanese population study: rs3769684 (T allele with OR 1.32, 95% CI 1.22–1·42, P = 1.82 × 10−8) in CD28 (identified by Barry et al. [6]) and rs4309200 (P = 2.26 × 10−8) in KLRG1 (Supplementary Fig. S6, Supplementary Table S5). Summary results of the transethnic meta-analysis. (a) Manhattan plot. All loci are labeled by the nearest gene with novel associations in red and the latest identified associations in blue. (b) Locus zoom plot for the lead SNP rs4309200 in the regions on chromosomes 12p13.31. (c) Locus zoom plot for the lead SNP rs9303279 in the regions on chromosomes 17q21.1. Table 1:Genome-wide significant SNPs from the transethnic meta-analysis.GenePos_hg38Top SNPEANEAOR95%CI P HLA-DQB16:32659937rs1063355TG0.500.43–0.589.75E-76NPHS119:35843280rs56117924AG1.811.71–1.921.36E-29AHI16:135425746rs7759971TC1.351.28–1.421.29E-16BTC4:74768255rs28862935AG1.461.36–1.561.68E-13CD282:203721233rs55730955AT0.750.67–0.832.74E-13CLEC16A16:10998462rs8062322AC0.750.67–0.831.39E-12TNFSF159:114796423rs6478108TC1.261.188–1.324.98E-11CALHM66:116448682rs2637681TG1.241.18–1.312.17E-10KLRG112:8969260rs4309200AC1.231.16–1.304.70E-09GSDMB17:39917715rs9303279CG0.820.75–0.897.25E-09aEA, Effect allele; NEA, Non-effect allele. Genome-wide significant SNPs from the transethnic meta-analysis. EA, Effect allele; NEA, Non-effect allele. M6PR and ORMDL3 gene expression showed causal association with SSNS The cell-type-specific eQTL effects of these two novel genetic risk loci and their potential causal target genes (eGene) were assessed by colocalization and MR approaches. Across the 14 immune cell types from OneK1K project, M6PR gene expression showed strong colocalization evidence in the NK cells (PPH4 = 0.93) and CD4+ T cells with an effector memory or central memory phenotype (CD4 ET) (PPH4 = 0.91), and ORMDL3 gene expression had strong colocalization associations with the B memory cells (PPH4 = 0.93) and CD4 ET cells (PPH4 = 0.92), within the colocalization region windows (Fig. 5a and b, Supplementary Data 3 and 4). These results are concordant with our observations of cell-type-specific eQTL effects in the scQTLbase and ImmuNexUT database (Supplementary Tables S6 and S7, Supplementary Fig. S7) [8, 37]. Further MR analyses of cell-type-specific M6PR and ORMDL3 gene expression in SSNS showed that higher M6PR expression in NK and CD4 ET cells were associated with an increased risk of SSNS (PWald_NK= 2.04 × 10−4, OR = 1.60, 95%CI = 1.25–2.05; PWald_CD4ET= 8.14 × 10−3, OR = 2.28, 95%CI = 1.24–4·20), whereas higher ORMDL3 expression in B memory and CD4 ET cells were associated with a lower risk of SSNS (PWald_B memory cells= 3.83 × 10−8, OR = 0.61, 95%CI = 0.51–0.73; PIVW_CD4ET= 1.82 × 10−9, OR = 0.49, 95%CI = 0.39–0.62) (Fig. 5c and d). The detail results of MR and following sensitivity analyses are shown in Supplementary Data 5 and 6. Overall, these findings suggest a potential causal link between novel SSNS genetic risks and the cell-type-specific expression of M6PR and ORMDL3. Association between genetically proxied M6PR and ORMDL3 expression, and SSNS. (a) LocusZoom plots illustrating evidence of genetic colocalization between M6PR expression in NK and CD4ET cells, and SSNS in the M6PR locus. (b) LocusZoom plots illustrating evidence of genetic colocalization between ORMDL3 expression in B memory and CD4 ET cells, and SSNS in the ORMDL3 locus. (c) Forest plots of MR results for the effects of genetically proxied M6PR and ORMDL3 expression, and SSNS. CD4 ET, CD4+ T cells with an effector memory or central memory phenotype; Bmem, B memory cells; Nsnps, number of SNPs used in MR. To investigate whether these novel genetic variants also play causal roles at proteomics level, we further conduct plasma pQTL analysis involving colocalization and MR. Among the candidate proteins from the identified novel loci, only M6PR protein was detectable within both UKB-PPP and deCODE project. However, the plasma pQTLs of M6PR protein showed no colocalization (PPH4 < 0.2) and an inverse disease association (PIVW_cis = 1.16 × 10−3, OR = 0.61, 95%CI = 0.45–0.82; PWM_cis&trans = 2.99 × 10−3, OR = 0.63, 95%CI = 0.46–0.85) with no detectable horizontal pleiotropy through either MR-Egger regression or MR-PRESSO global testing (Supplementary Data 7 and 8). Additional sensitivity analyses, including funnel plots and leave-one-out analyses, further verified its robustness (Supplementary Fig. S8). Correlations for susceptibility loci on the age of onset Finally, we generated a PRS to evaluate the previously reported correlations on the age of onset. As expected, this independent Chinese cohort confirmed the correlation of higher PRS with younger age of onset (Supplementary Table S8). Moreover, we found that this association was primarily predisposed by the risk locus on the NPHS1, as patients with risk allele A of rs56117924 on the NPHS1 locus had the most significantly lower age of onset than allele G (P = .007) when screening all the susceptibility loci (Supplementary Fig. S9). DISCUSSION To our knowledge, this is the first GWAS of pediatric SSNS in the Chinese Han population, not only confirming the prominent predisposing genetic factor mapping to the HLA-II region but also replicating SSNS susceptibility in the NPHS1 region. Besides, two novel susceptibility loci for pediatric SSNS were identified in the M6PR-KLRG1 and GSDMB-ORMDL3 regions, which are closely related to immunological function. Moreover, this study evaluated the molecular QTL effects of the novel loci and estimated their potential causative effects on SSNS. This study revealed certain disease signatures and investigated the underlying molecular mechanisms, providing further insights into the function of genetic factors and biological mechanisms of SSNS. The NPHS1 gene, encoding the critical podocyte slit diaphragm protein nephrin, has been established as a susceptibility locus for SSNS in East Asian populations, a finding initially reported in Japanese cohorts [26], and confirmed in subsequent ancestry-specific analyses [6]. Our study independently reinforces this genetic association within the Chinese population. Recent studies have confirmed the presence of circulating anti-nephrin autoantibodies in patients with podocytopathy, including childhood idiopathic nephrotic syndrome (INS), revealing a potential autoimmune podocytopathy phenotype [38–40]. This pattern appears analogous to the previously reported interplay between HLA-D/PLA2R1 genetics and anti-PLA2R autoimmunity in membranous nephropathy [41, 42]. These independent genetic and immunological observations lead us to hypothesize that genetic variations in the NPHS1 susceptibility region may contribute to disease susceptibility by influencing nephrin immunogenicity or disrupting immune tolerance: a notion that is biologically plausible. The correlation between the East Asian-specific NPHS1 risk loci and the relatively high incidence of SSNS in East Asian populations may provide a plausible explanation [43]. However, we emphasize that our present study provides no direct evidence to link these NPHS1 risk alleles with autoantibody production. Further integrated studies are required to determine whether and how NPHS1 risk loci affect the antigenicity of nephrin and whether these variants interact with specific HLA alleles to influence the development of autoimmune responses. The newly discovered lead SNP rs4309200 is located at the 12p13.31 locus within the M6PR-KLRG1 region, where the natural killer gene complex is situated, playing a critical role in regulating the activity of NK and T cells [44]. By integrating scRNA-seq eQTL and pQTL data with novel risk loci using colocalization and MR analyses, we identified potential causal genes at risk loci and the specific cell types they affect. Higher expression of M6PR in genetically proxied NK and CD4 ET cells was associated with an increased risk of SSNS and shared causal genetic variants. Although MR analyses on pQTLs showed that genetic predisposition to higher plasma M6PR levels was associated with lower SSNS risk, which is an inverse effect compared to cell expression, there was lack of colocalization evidence for plasma M6PR proteins. M6PR is a transmembrane glycoprotein crucial in delivering newly synthesized lysosomal enzymes bearing the M6P signal from the trans-Golgi-network to late endosomes [45–47]. Limited research has been conducted to investigate the function of M6PR, with some suggesting its involvement in the exocytosis of M6P-bearing ligands and the process of viral infection, such as rotaviruses and enterovirus 71 cell entry [48], the transport of varicella-zoster virus [49], and the efficient replication of Influenza A virus [50]. In fact, infections are one of the major predisposing factors for SSNS [43]. Although M6PR has not previously been associated with SSNS, our analyses point to it as a potential causal gene in NK cells and CD4ET cells for SSNS. This suggests that further experimental exploration is necessary to confirm the biological importance of dysregulated M6PR gene expression in SSNS, particularly in specific cell types. The second novel lead SNP rs9303279, located at the 17q21.1 locus, barely attained suggestive significance in the most recent multi-population meta-GWAS study due to insufficient sample size. This locus, housing GSDMB and ORMDL3, is associated with an array of autoimmune disorders and is the most consistently reproducible asthma-susceptibility locus identified to date [51, 52]. The shared pleiotropic genetic loci between asthma and SSNS possibly reflect common susceptibility mechanisms, as both are typically induced by respiratory viral infection. Moreover, the incidence for asthma was significantly higher compared with healthy control participants in a large study of 1340 children with newly diagnosed INS [53]. This further underscores the shared mechanisms of immune dysregulation leading to the high prevalence of coexisting diseases. Therefore, the comprehensive shared genetic architecture and causal associations between SSNS, asthma, and other autoimmune disorders require further exploration. Our subsequent multi-omics analyses point to ORMDL3 as the causal gene in B memory and CD4 ET cells. ORMDL3, a conserved endoplasmic-reticulum transmembrane protein, plays a critical role in immune regulation. Its involvement extends to key biological processes including lipid metabolism, inflammatory responses, and autophagy: pathways intrinsically linked to autoimmune disorders [54, 55]. Experimental evidence demonstrate d that ORMDL3 deletion disrupts immune homeostasis, leading to decreased splenic B-cell survival through increased apoptosis and impaired autophagy [56, 57]. Considering the pivotal role of ORMDL3 in immune homeostasis, understanding its function in SSNS could provide critical insights into the pathogenesis and potentially lead to the development of novel therapeutic strategies. Finally, this study confirmed that higher PRS was linked to an earlier age of onset in the independent Chinese cohort. Although potential overfitting exists when using GWAS-derived weights to calculate PRS within the same cohort, the intra-cohort clinical association, rather than inter-cohort disease risk association, still provides valuable insights into the clinical relevance of PRS. In addition, rs56117924 on the NPHS1 locus demonstrated the strongest association. As mentioned by Jia et al. [34], no eQTL effects of rs56117924 were observed in the current glomerular transcriptomic data, while regulating NPHS1 expression in an allele-specific expression manner. Based on these findings and prior knowledge, we hypothesize that the lower expression of nephrin may contribute to podocyte vulnerability, which could potentially explain the earlier onset age observed in carriers of the risk allele. However, limited clinical correlations with PRS and individual genetic loci have been observed till now due to limited sample size and significant variants. Therefore, it is necessary to perform large-scale GWAS by widespread collaborations to identify more associations and evaluate any potential personal or clinical utility. No doubt certain limitations exist in this study. First, GWAS meta-analyses were performed by accessing the GWAS summary statistics directly rather than individual-level data, which could lead to heterogeneity and bias in genetic marker estimation. Second, the mixed-ancestry population of our meta-GWAS summary data, coupled with the European origin of sc-eQTL and plasma pQTL data, may introduce potential biases due to LD structure and population matching. These biases limit the generalizability of our findings to other ethnic groups. However, based on the statistical power afforded by the best available data from European QTL cohorts and existing evidence suggesting that many eQTLs and pQTLs are conserved across populations [58–60], we believe that the colocalization results, while requiring cautious interpretation, still provide valuable biological insights into the potential regulatory mechanisms of the novel GWAS loci. Furthermore, considering the differences in growth and development, immune status, and environmental exposures between children and adults, there may be variations in plasma protein and peripheral blood mononuclear cell gene expression levels, as well as their genetic regulation. The confluence of differences in population, tissue origin, and age groups could introduce biases into our causal inferences. Therefore, future trials should thoroughly evaluate the causal associations identified in our study across diverse cohorts before clinical implementation. Finally, this study has not yet evaluated the contribution of rare variants, copy numbers, and insertion-deletions to SSNS. Further sequencing studies are required to assess the relative contributions to overall disease risk. CONCLUSIONS In summary, this GWAS on pediatric SSNS in a Chinese Han population identified two novel susceptibility loci, suggested that M6PR and ORMDL3 gene expression may drive SSNS risk in a cell-type-specific manner, and confirmed the clinical association of PRS in an independent cohort. Importantly, this study underscores the power of large-scale collaborations in uncovering genetic associations of SSNS. In the future, more extensive collaborations across diverse populations will be crucial for gaining a deeper understanding of the genetic basis of SSNS. ETHICS APPROVAL AND CONSENT TO PARTICIPATE Written informed consent was obtained from all participating children and/or their guardians. The study was reviewed and approved by the Ethics Committee of the China National Clinical Research Centre (Children’s Health and Disease, No. 2018-141). The study was performed according to the provisions of the Declaration of Helsinki. Supplementary Material