Skip to main content
PMC home page
eBioMedicine logoLink to eBioMedicine
. 2023 Jun 8;93:104647. doi: 10.1016/j.ebiom.2023.104647

Dissecting shared genetic architecture between obesity and multiple sclerosis

Ruijie Zeng a,k, Rui Jiang a,b,c,k, Wentao Huang a,b,k, Jiaxuan Wang a,k, Lijun Zhang a,c, Yuying Ma a,b, Yanjun Wu a,b, Meijun Meng a,d, Hekui Lan e, Qizhou Lian f,g,h, Felix W Leung i,j,∗, Weihong Sha a,b,c,d,∗∗, Hao Chen a,b,c,d,∗∗∗
PMCID: PMC10363440  PMID: 37300932

Summary

Background

Observational studies have associated obesity with an increased risk of multiple sclerosis (MS). However, the role of genetic factors in their comorbidity remains largely unknown. Our study aimed to investigate the shared genetic architecture underlying obesity and MS.

Methods

By leveraging data from genome-wide association studies, we investigated the genetic correlation of body mass index (BMI) and MS by linkage disequilibrium score regression and genetic covariance analyser. The casualty was identified by bidirectional Mendelian randomisation. Linkage disequilibrium score regression in specifically expressed genes and multimarker analysis of GenoMic annotation was utilised to explore single-nucleotide polymorphism (SNP) enrichment at the tissue and cell-type levels. Shared risk SNPs were derived using cross-trait meta-analyses and Heritability Estimation from Summary Statistics. We explored the potential functional genes using summary-data-based Mendelian randomization (SMR). The expression profiles of the risk gene in tissues were further examined.

Findings

We found a significantly positive genetic correlation between BMI and MS, and the causal association of BMI with MS was supported (β = 0.22, P = 8.03E-05). Cross-trait analysis yielded 39 shared risk SNPs, and the risk gene GGNBP2 was consistently identified in SMR. We observed tissue-specific level SNP heritability enrichment for BMI mainly in brain tissues for MS in immune-related tissues, and cell-type-specific level SNP heritability enrichment in 12 different immune cell types in brain, spleen, lung, and whole blood. The expressions of GGNBP2 were significantly altered in the tissues of patients with obesity or MS compared to those of control subjects.

Interpretation

Our study indicates the genetic correlation and shared risk genes between obesity and MS. These findings provide insights into the potential mechanisms behind their comorbidity and the future development of therapeutics.

Funding

This work was funded by the National Natural Science Foundation of China (82171698, 82170561, 81300279, and 81741067), the Program for High-level Foreign Expert Introduction of China (G2022030047L), the Natural Science Foundation for Distinguished Young Scholars of Guangdong Province (2021B1515020003), Natural Science Foundation of Guangdong Province (2022A1515012081), the Foreign Distinguished Teacher Program of Guangdong Science and Technology Department (KD0120220129), the Climbing Programme of Introduced Talents and High-level Hospital Construction Project of Guangdong Provincial People's Hospital (DFJH201803, KJ012019099, KJ012021143, and KY012021183), and in part by VA Clinical Merit and ASGE clinical research funds (FWL).

Keywords: Obesity, Multiple sclerosis, Body mass index, Shared genetic architecture

Research in context.

Evidence before this study

Several observational studies have revealed that children with a higher body mass index (BMI) are associated with a higher risk of developing multiple sclerosis (MS) in later life. However, the role of genetic factors in their comorbidity remains largely unknown.

Added value of this study

In this study, by leveraging data from genome-wide association studies, we found significant genetic correlation and identified shared risk SNPs between BMI and MS. We further identified a putative functional gene shared between obesity and MS. These findings could provide insights into the shared genetic architecture between obesity and MS.

Implications of all the available evidence

Our study highlights the importance of early detection and prevention of neurodegenerative disorders, including MS, amongst overweight and obese people. New therapeutics for the two diseases could be explored, and it is promising to develop modalities targeting both diseases based on their shared genetic architecture. Moreover, our findings might provide new understandings for the pathogenesis of the two diseases, and inspire future research to keep digging into their comorbidity.

Introduction

Obesity is a growing health and economic issue.1,2 It is often comorbid with a wide range of conditions, including neurodegenerative disorders.3,4 In recent years, growing evidence from observational studies has illustrated the associations between obesity and multiple sclerosis (MS). MS is a debilitating chronic demyelinating and neurodegenerative disease of the central nervous system, and leads to an impaired quality of life and disability.5

Several observational studies have revealed that children with a higher body mass index (BMI) have a higher risk of developing MS later in life.6, 7, 8 Amongst children with obesity, a 101% higher risk of developing MS compared to normal-weight children was revealed by a meta-analysis.9 In addition, previous Mendelian randomisation (MR) studies have indicated the potential causal effect of BMI on MS.10, 11, 12, 13 These epidemiological studies provide hints of shared genetic risk components between obesity and MS. However, more updated data are needed to evaluate their causality, and the shared genetic architecture between obesity and MS remains largely unknown.

Understanding the causality and biology underlying this association is important for deciphering the aetiology of MS and can potentially provide therapeutic insights. Several mechanisms have been proposed. Studies suggest that obesity aggravates central inflammation and disability in multiple sclerosis.14,15 Alterations in hormones and gut microbiota in patients with obesity might also account for the development of MS.16, 17, 18 Clearly, it is still very challenging to obtain a full picture of this complex link. Studying the shared genetic risk components could potentially provide a unique perspective in this direction.

Here, using large-scale genome-wide association study (GWAS) summary statistics (Fig. 1), we aimed to investigate the genetic correlation, causal association, and shared risk loci with potential functions between obesity and MS, to provide insights into their comorbidity.

Fig. 1.

Fig. 1

Open in a new tab

Overview of statistical analyses performed in the study. CPASSOC: Cross Phenotype Association; GNOVA: Genetic covariance analyser; GSMR: Generalised summary-data-based Mendelian randomisation; GWAS: Genome-wide Association Study; LD: linkage disequilibrium; MAGMA: Multi-marker Analysis of GenoMic Annotation; MTAG: Multi-Trait Analysis of GWAS; MR: Mendelian Randomisation; ρ-HESS: Heritability Estimation from Summary Statistics; scRNA-seq: single-cell RNA sequencing; SMR: Summary-databased Mendelian randomisation.

Methods

Datasets

GWAS summary statistics

Effect estimates for SNPs associated with BMI, which was calculated by weight and height information of included participants, were obtained from the GWAS meta-analysis by the Genetic Investigation of Anthropometric Traits (GIANT) consortium, which involved 2.4 million HapMap 2 SNPs from approximately 0.7 million participants.19 GWAS summary results for MS were derived from the meta-analysis by the International MS Genetics Consortium (IMSGC), which included 15 datasets involving 26,703 controls and 14,802 MS cases of European ancestry, and the diagnostic criteria had been reported in each included studies.20 The collection of samples, quality control, and imputation methods for each study have been previously described.19,20

Bulk-tissue RNA sequencing gene expression data

In the subsequent linkage disequilibrium score regression (LDSC)-specifically expressed genes (SEG) and summary-data-based Mendelian randomisation (SMR) analysis, we obtained bulk-tissue RNA-seq gene expression data from the Genotype-Tissue Expression (GTEx) project, which is a public data resource of gene expression in 53 nondiseased human primary tissues.21 We download the GTEx v6p dataset, which has been fixed.22 We then chose the lite version of the GTEx V8 expression quantitative trait locus (eQTL) summary data (P < 1 × 10−5). cis-eQTL summary statistics for whole blood were downloaded for the downstream analysis from eQTLGen, a meta-analysis of 14,115 individuals.23

Single-cell RNA sequencing gene expression data

We obtained four single-cell RNA sequencing (scRNA-seq) resources from human lung (N = 57,020 cells), spleen (N = 94,257 cells), and whole blood from 10X Genomics Chromium and mouse brain (N = 133,509,876 cells).24,25 The “EWCE” R package was utilised to process the scRNA-seq data and convert mouse genes to human gene symbols.26

Statistical analyses

Heritability and genetic correlation

LDSC is a useful method to estimate the genetic correlation for multiple traits or diseases.27 Based on the precomputed LD scores of the 1000 Genomes projects, which were calculated for SNPs in the HapMap 3 SNP set, we removed SNPs that did not match the reference panel (MAF ≤ 0.01 or INFO score ≤ 0.9) and reformatted new GWAS summary statistics.28 We estimated single-trait SNP heritability for BMI and MS using stratified linkage disequilibrium score regression (SLDSC) with the baseline-LD model. According to the recommendation, we set the population prevalence and observed sample prevalence as 0.0003 and 0.63, respectively, to convert observed scale heritability to the liability scale. Then we performed bivariate LDSC without constraining the intercept to estimate the rg value, representing genetic correlations between MS and BMI, and selected the suggestive (P < 0.05) genetic associations as the significant correlation.29,30 Sensitivity analyses were conducted based on LDSC with the single-trait heritability intercept constrained. Because there was no sample overlap in our two traits, we set all single-trait intercepts to 1 and all cross-trait intercepts to 0.

Genetic covariance analyser (GNOVA) was supplemented to estimate the SNP-based heritability and genetic correlation between BMI and MS.31 GNOVA estimates genetic covariance based on all genetic variants shared between two GWAS summary statistics. Genetic correlations were then calculated based on variant heritability and genetic covariance. Calculations were based on the 1000 Genomes Project’s European population-derived reference data using default parameters. In addition, sample overlap correction between two different sets of GWAS summary statistics was statistically calculated. Compared to LDSC, GNOVA provides higher estimation accuracy for genetic correlations and more powerful statistical inference.31

Cross-trait meta-analysis

To detect the shared risk SNPs in BMI and MS, we performed two cross-trait meta-analyses, including multi-trait analysis of GWAS (MTAG) and cross-phenotype association test (CPASSOC).32, 33, 34 MTAG is a generalised meta-analysis method that enhances statistical power to estimate the genotypic and phenotypic variance-covariance matrices to generate trait-specific estimates for each SNP. SNPs were restricted with a minor allele frequency (MAF) ≥ 0.01 and sample size N ≥ (2/3) × 90th percentile. MTAG adjusts for possible errors by using bivariate LD score regression when sample overlap is present. MTAG is suitable when all variants have the same effect sizes on traits and generate trait-specific association statistics. We calculated the upper bound for the false discovery rate (‘maxFDR’) to examine the assumptions on the equal variance-covariance. In addition, as a sensitivity analysis, CPASSOC integrates association evidence from multiple traits’ GWAS summary statistics when the variant is correlated to at least one trait. We utilised the SHet version to assume heterogeneous effects across traits. The SNP set was obtained by applying pairwise LD pruning with r2 = 0.2 using the software “PLINK”. We prioritised independent SNPs that were genome-wide significant (P < 5 × 10−8) in the cross-trait meta-analyses using both MTAG and CPASSOC and were in significant regions identified by ρ-HESS.

Identification of local genetic correlations

Heritability Estimation from Summary Statistics (ρ-HESS) is a method to estimate local SNP-heritability and genetic correlation.35 We estimated the local genetic correlations to examine whether BMI shared genetic correlation with MS at the local independent region in the genome using ρ-HESS. There were 1699 potential regions that were approximately LD-independent loci, with an average size of nearly 1.5 MB.36 Then we calculated the local SNP heritability for two traits and the genetic correlation between two traits using the 1000 Genomes Project as the reference provided on the ρ-HESS webpage. The pairwise GWAS (GWAS-PW) method was also used to examine shared genomic regions by BMI and MS.37 GWAS-PW can provide estimations of the posterior probability for the locus shared by both traits (PPA3), or the locus associated with both traits but by distinct causal variants (PPA4). The threshold of PPA3/4 > 0.8 was used in our models.

Mendelian randomisation

To explore putative causal relationships between BMI and MS, the R packages “TwoSampleMR” and “GSMR” were used for suggestive associations (P < 0.05). We undertook Mendelian randomisation analysis, mainly including five MR methods, MR-Egger, inverse variance weighting (IVW), weighted median, weighted mode, and generalised summary-data-based Mendelian randomisation (GSMR) with different assumptions about horizontal pleiotropy.38, 39, 40, 41 Briefly, when there is one single genetic variant, the Wald ratio is used to calculate the causative effect between the exposure and the outcome. Using the meta-analysis approach, IVW analysis can estimate the causal effects of two phenotypes. The MR-Egger method further added a weighted linear regression of the gene–outcome coefficients for non-measured horizontal pleiotropy, which allows for the presence of directional uncorrelated pleiotropy. The pleiotropy test and heterogeneity test were conducted by the MR-Egger intercept test and Cochran's Q statistic. We used single SNP effect analysis and MR-PRESSO analysis to detect pleiotropy and outliers.42 The variants were mainly selected based on three assumptions: (1) they are correlated with the exposure; (2) they are not dependant on confounding factors; and (3) they do not directly generate effects on the outcome. For these five methods, we selected SNPs with genome-wide significance (P ≤ 5 × 10−8) of the exposure trait as instrumental variables. F statistics for each instrument were estimated by F = β2/SE2, and an F statistic < 10 was regarded as insufficiently informative for further analysis.43 The study power (at α = 0.05) was calculated to illustrate a causal effect depending on the percentage of risk factor variance accounted for by instruments.44 The heterogeneity Q statistics were calculated, and the leave-one-out method was used as the sensitivity analysis.

LDSC-SEG analysis

We performed LDSC-SEG to investigate whether SNP heritability for BMI and MS is evident for trait-tissue relevance inference.45 The 1000 Genomes Phase 3 of European ancestry was utilised as a reference panel to calculate LD scores, and SNPs only in HapMap 3 with MAF > 0.05 were included as input. Based on the baseline model and all gene sets, we ranked genes from the GTEx project by computed t-statistics reflecting critical tissue types and their specific expression in 53 tissues. We obtained the top 10% of specifically expressed candidate genes with the highest t-statistic to estimate the significance of tissue type-specific SNP heritability enrichment. The coefficient P values were calculated based on the regression coefficient Z score, and the Benjamini-Hochberg FDR-corrected P value of < 5 × 10−3 was determined as significant for enrichment tissues across the two traits.

Cell type enrichment analyses using scRNA-seq datasets

We conducted Multimarker Analysis of GenoMic Annotation (MAGMA) cell typing to evaluate the gene-level genetic correlation between BMI and MS GWAS traits and cell type expression specificity.46 Cell types across the four tissues were considered significant in MAGMA with a P value < 0.05. The cell type specificity matrix for scRNA-seq used in MAGMA was calculated using Expression Weighted Cell Type Enrichment, “EWCE” and the “MAGMA_Celltyping” R package.47

Summary-data-based Mendelian randomization

We conducted SMR analysis to identify candidate risk genes with possible causal effects and SNPs significant in cross-trait meta-analyses of BMI and MS.48 We used GWAS and eQTL data to detect the association between trait-associated SNPs and gene expression. The heterogeneity in dependent instrument (HEIDI) test was applied to distinguish linkage in the causal association. Genomic expression data from GTEx V8 and cis-eQTL summary data from eQTLGen were used for the eQTL expression data of whole blood.23,49 As a default, we removed SNPs in strong LD with an r2 > 0.9 and with top associated eQTLs if MAF > 0.01. Expression probes with eQTL P ≤ 5 × 10−8 were selected as the top associated variants. All SNPs were extracted by genome-wide complex trait analysis (GCTA) software, ensuring their independence.50 SMR significant probes were selected using Bonferroni-corrected thresholds for SMR P values (0.05/number of probes) and HEIDI test P value thresholds >0.05 to indicate the lack of heterogeneity.

Validation of expressions of the risk gene

The expression profiles of adipose tissues, brain lesions, and peripheral blood mononuclear cells were evaluated for the risk gene by the datasets GSE9624, GSE38010, and GSE21942.51, 52, 53 The GSE9624 dataset involved omental adipose tissues from obese and normal-weight children.51 The GSE38010 dataset included histologically characterised MS brain lesions and control brain samples.52 The GSE21942 dataset contained peripheral blood mononuclear cells from MS patients and controls.53 The Mann–Whitney U test was used to evaluate the differences in gene expression between the diseased group and the control group. Analyses of the data were performed using R version 4.1.0 (R Project for Statistical Computing, Vienna, Austria), Python 2.7 (Python Software Foundation, Wilmington, US), or GraphPad Prism 8.0 (GraphPad Software, San Diego, US).

Ethics

Since the included studies had been approved by their independent review boards, additional ethical approval is not required in our study based on summary-level data.

Role of funders

Funders had no role in the in the study design, data collection, data formal analysis nor interpretation or writing of the report.

Results

Estimation of genetic correlations between obesity and MS

We used bivariate LDSC to estimate the genetic correlation (without constrained intercept) between BMI and MS (rg = 0.08, P = 3.45 × 10−8, LDSC, Supplementary Fig. S1 and Table S1). The liability-scale SNP heritability estimates were 21.2% and 4.6% for the BMI and MS traits, respectively. The intercept of genetic covariance was calculated at approximately 0.01, indicating mild sample overlapping between BMI and MS. After constraining the LDSC intercept on the assumption of no sample overlap, the genetic correlation was slightly weaker but remained significant (Supplementary Fig. S1 and Table S1). Analyses by GNOVA also demonstrated a positive and consistent genetic association between BMI and MS (Supplementary Table S1).

Identification of genomic risk regions for BMI and MS

Given the strong genetic relationships between BMI and MS, we conducted MTAG to improve our power to identify genetic SNPs shared between traits. A total of 39 genome-wide significant SNPs (P < 5 × 10−8) were revealed in both MTAG and CPASSOC (Supplementary Table S2), including 18 newly identified shared SNPs (rs11647753, rs11649612, rs12716972, rs12716974, rs2289292, rs3809624, rs4407979, rs4609871, rs2306589, rs2306593, rs8070260, rs8882, rs9906189, rs1044821, rs11667487, rs2302299, rs4808762, rs8112975). The maxFDR values for MTAG analyses of BMI and MS were 1.4 × 10−6 and 1.8 × 10−2 respectively. Furthermore, the MTAG results were highly consistent with those generated by CPASSOC, indicating that the MTAG results are reliable and that bias in MTAG assumptions is likely to be negligible. Eventually, 13 risk SNPs were found to be consistently significant when examined by ρ-HESS, and the shared loci were verified by GWAS-PW (PPA4 > 0.8, Supplementary Table S3).

The local genetic correlation was estimated, and 57 suggestively significant regions were identified (P < 0.05, ρ-HESS, Supplementary Table S4). There was close agreement in the average local genetic correlation in regions harbouring BMI-specific loci or MS-specific loci (Fig. 2a and b and Supplementary Fig. S2). We estimated the local single-trait SNP heritability for BMI (h2 = 22.4%) and MS (h2 = 23.5%) (Table 1). Compared with bivariate LDSC, genome-wide local genetic correlations calculated by ρ-HESS between MS and BMI (rg = 0.0428) were all largely consistent (Table 1).

Fig. 2.

Fig. 2

Open in a new tab

Local genetic correlations (rg) between BMI and MS. (a) Average local rg estimates for two traits in regions harbouring disease-specific risk variants, regions harbouring shared risk variants (“intersection”), and all other regions (“neither”). Local genetic correlations with estimates less than −1 or greater than 1 were forced to −1 or 1, respectively. Error bars represent the 95% confidence intervals (CIs), which were calculated using a jackknife method. (b) Density distribution of local rg estimates for two traits in disease-specific regions (red, green), intersection regions (blue) and other (purple) regions. BMI, body mass index; MS, multiple sclerosis.

Table 1.

Heritability and genetic correlation between body mass index and multiple sclerosis.

Method Body mass index Multiple sclerosis
Heritability h2 LDSC 0.2119 0.0457
Heritability h2 GNOVA 0.1885 0.1385
Heritability h2 ρ-HESS 0.2240 0.2350
Genetic correlation rg LDSC 0.0796
Genetic correlation rg GNOVA 0.0647
Genetic correlation rg ρ-HESS 0.0428
Open in a new tab

LD: linkage disequilibrium. GNOVA: Genetic covariance analyser; ρ-HESS: Heritability Estimation from Summary Statistics.

Evidence for causality between BMI and MS

We conducted bidirectional MR to explore the potential causal effect and whether the shared genetic background between BMI and MS was consistent with pleiotropy. The IVs were selected after evaluation based on the three assumptions (Supplementary Table S5). The F statistics for the instrumental variables ranged from 28.6 to 1360.3, indicating that the probability of weak instrument bias was low. We conducted various bi-directional MR methods to test the stability of relationships for a more stable result. We found evidence to support the causality of BMI on MS in five methods (IVW β = 0.22, SE = 0.06, P = 8.03E-05; GSMR β = 0.32, SE = 0.05, P = 6.15E-13; Fig. 3 and Supplementary Fig. S3 and Table S6) with no significant evidence of heterogeneity (IVW Q = 356.49, P = 0.99, Supplementary Table S6). The leave-one-out analysis indicated the effect was not driven by any single SNP (Supplementary Fig. S4a). The estimated power to detect a causal effect for BMI on MS was 100.0%. In the reverse analyses, all these methods except GSMR identified a nonsignificant causal effect of MS on BMI (IVW β = 4.47E-03, SE = 4.23E-03, P = 0.29; GSMR β = 0.04, SE = 0.01, P = 1.32E-3; Supplementary Table S6). The leave-one-out analysis was shown in Supplementary Fig. S4b.

Fig. 3.

Fig. 3

Open in a new tab

Summary of bi-directional MR analyses between BMI and MS. BMI, body mass index; GSMR: Generalised summary-data-based Mendelian randomisation; IVW: inverse variance weighting; MS, multiple sclerosis. Error bars represent the 95% confidence intervals for the associated MR estimates.

Tissue-level SNP heritability enrichment in BMI and MS

To identify in which tissue these shared SNPs take effect, we used the LDSC-SEG method to identify specific tissues in which genes with increased expression were enriched in SNPs, using publicly available GWAS data and genotype tissue expression data from GTEx. After adjusting for the baseline model, we identified FDR-significant (P < 5 × 10−3, LDSC-SEG) SNP heritability enrichment for BMI across 9 tissues (Fig. 4a and Supplementary Table S7), particularly for central nervous system (CNS)-related tissues, including the frontal cortex, anterior cingulate cortex, nucleus accumbens, putamen, caudate, hypothalamus, cerebellar hemisphere, cerebellum, and cortex. For MS, a total of 4 tissues were significantly enriched, particularly for blood and immune-related tissues, including the spleen and Epstein–Barr virus (EBV)-transformed lymphoblastoid cell lines (LCLs) (Fig. 4b and Supplementary Table S7).

Fig. 4.

Fig. 4

Open in a new tab

Tissue type-specific enrichment of SNP heritability for BMI and MS estimated using LDSC-SEG. (a) The heritability enrichment of tissues for BMI; (b) The heritability enrichment of tissues for MS. The x axis displays negative log10 P-values of coefficient Z-scores for each individual test. BMI, body mass index; MS, multiple sclerosis; SNP, single nucleotide polymorphism; LDSC-SEG: linkage disequilibrium score regression applied to specifically expressed genes.

Cell-level SNP heritability enrichment in BMI and MS

We utilised publicly available scRNA-seq datasets of four tissues enriched in LDSC-SEG, including brain, spleen, lung, and whole blood, to evaluate the gene-level genetic association with cell type expression specificity for BMI and MS. In the lung dataset, we found significant enrichment at P < 0.05 (MAGMA) for both BMI and MS in mature B cells, naive B cells, mast cells, natural killer (NK) cells, dividing NK cells, CD4+ T cells, CD8+ cytotoxic T lymphocytes, dividing T cells, regulatory T cells, activated dendritic (DC) cells, plasmacytoid DC cells, and monocytes. We observed a significant enrichment across NK_FCGR3Apos in spleen tissue. Cells in the brain and in peripheral blood mononuclear cell (PBMC) tissues were not coenriched in BMI and MS traits. The enriched cells in four tissues for BMI and MS are listed separately (Supplementary Figs. S5–S8).

Identification of shared functional genes for BMI and MS

We applied SMR to infer causality and identify the putative functional genes for BMI and MS, by jointly analysing GWAS summary data and whole blood eQTL summary data from eQTLGen and GTEx. We identified 12 shared risk genes, including FAM213B, GGNBP2, HLA-B, HLA-C, MAST3, MICB, MYO19, PRXL2B, TBX6, TNFRSF14, TRAF3, and ZNHIT3, of which only one gene, GGNBP2, was shared between BMI (PSMR = 3.69 × 10−19, PHEIDI = 0.11, topSNP: rs11263770) and MS (PSMR = 2.41 × 10−7, PHEIDI  = 0.98, topSNP: rs11650008) and passed the HEIDI-outlier test in cis_eQTL data (Supplementary Table S8). More importantly, GGNBP2 was also identified as one of the genetically shared variants in previous cross-trait meta-analysis phenotypes between BMI and MS.

Differential expressions of GGNBP2

The mRNA expressions of GGNBP2 in tissues of patients with obesity or MS, and control subjects were further evaluated. The expression of GGNBP2 was significantly increased in the adipose tissue of obese individuals compared to control subjects (Supplementary Fig. S9a, P < 0.05, unpaired t-test). In addition, GGNBP2 expression was also enhanced in peripheral blood nuclear cells of MS patients compared to that in healthy controls (Supplementary Fig. S9b, P < 0.0001, unpaired t-test). In MS brain lesions, the expression of GGNBP2 was generally raised compared to normal brain tissue, while studies involving more participants are warranted to increase the statistical power (Supplementary Fig. S9c, P = 0.09, unpaired t-test).

Discussion

In this study, we present evidence of causality and overlapping genetic architecture of BMI with MS. Our results provide new insights into their comorbidity, and might contribute to the prediction, diagnosis, and treatment of diseases.

We found a significant genetic correlation between BMI and MS, which supported the hypothesis that genetic factors play an important role in the comorbidity of obesity and MS. From the cross-trait meta-analysis, 39 SNPs were identified in both MTAG and CPASSOC analysis, in which 13 suggestively significant SNPs were located in the ρ-HESS-estimated significant genomic regions. We used two different statistical analysis methods, MTAG and CPASSOC, to reduce possible bias caused by potential sample overlap and further explored the risk SNPs obtained from MTAG with CPASSOC. Our results showed that the SNPs identified by MTAG were all consistently significant in the CPASSOC analysis, which improved the reliability of our findings. The causality of BMI on MS was consistent with previous studies,10,54 and the sensitivity analyses confirmed the robustness of the findings.

In addition to the IVW methods, different statistical methods for Mendelian randomisation analysis were used to evaluate the robustness of our results. The MR-Egger regression yields less biased estimates and lower power than the IVW method.55 The weighted-median method is able to provide effect estimates with lower bias than the IVW method, whereas a high type 1 error rate has been observed.56 The weighted-mode method generally has a decreased bias and a type 1 error rate, whereas its power to detect a causal effect is low.56 The GSMR method utilises the generalised least-squares approach to generate estimates, compared to the variance–covariance matrix used by the IVW method.57 Since three of the five methods for the causal estimates of BMI on MS were statistically significant, the effect was considered robust in this study.

Functional enrichment for gene expression in multiple tissues and cells was also investigated using the GTEx datasets. We identified 9 tissues, mainly the brain, with significant SNP heritability enrichment for the BMI trait. Growing evidence also suggests that susceptibility to obesity is distributed across multiple brain regions and is strongly associated with structural abnormalities.58, 59, 60 The enrichment results of the MS trait were mainly reflected in immune tissues, including the spleen and LCLs, which suggests the involvement of local immune responses in the development of MS. This is consistent with previous literature demonstrating a strong relationship between MS and immune dysregulation.61 EBV-infected B cells and plasma cells that accumulate in meningeal immune cell collections may contribute to the progressive development of MS.62 The lung could contribute to the activation and transformation of autoreactive T cells to a migratory mode, which allows them to enter the CNS and induce autoimmune conditions, including MS.63 Our findings provide evidence to support the involvement of immune responses in MS, and further experimental validation of the enriched tissues and cells is warranted.

Notably, we identified SNP heritability enrichments in different immune cells in the lung and spleen for both BMI and MS. MS is an immunologically heterogeneous disorder, and CD8+ T cells predominate in MS lesions.64 B-cell depletion may reduce the proinflammatory cytokines produced by B cells, CD4+ and CD8+ T cells, which can effectively reduce MS relapses.65,66 B cells interact with T helper cells to create a feedforward loop, and the highly pathogenic subsets enter the blood-CNS barriers, which can lead to MS pathological changes.67 NK and DC cells control T-cell activation in CNS autoimmunity, and reduce the risk of MS.68,69 Mast cells participate in the pathogenesis of MS by promoting angiogenesis.70 Our results may provide insights into the pathogenesis of obesity and MS, and the development of therapeutics targeting specific tissues and cell types.

In addition to cross-trait meta-analysis, we also used blood and tissue eQTL data to evaluate whether the BMI-MS association can be mediated by shared risk genes. Amongst the shared risk genes we identified, human leukocyte antigen (HLA) class genes have been identified to recruit macrophages into adipose tissues.71 In addition, the primary function of HLA class genes is antigen presentation to T cells, which is involved in CNS-directed autoimmunity and MS development.72 Striking interactions of BMI and HLA genotype with the risk of MS have been observed, which confirms our findings.73 Using SMR and HEIDI, we discovered that GGNBP2 might serve as another potential link between the two traits, and transcriptomic analyses using patient samples verified our findings. Prior studies have documented the biological relationship between GGNBP2 and BMI. GGNBP2 is a tumour suppressor gene involved in several types of cancer, such as glioma,74 breast cancer,75 and prostate cancer.76 GGNBP2 is also a shared gene for ALS and obesity-related traits.77 Further investigations into the mechanisms of GGNBP2 on obesity and MS pathogenesis are needed.

Numerous attempts have been made to evaluate the association between obesity and MS. Several MR studies have revealed the causal effect of BMI on MS.10, 11, 12, 13 However, the summary statistics used in previous studies were relatively outdated or smaller-scale, and previous evidence on their shared genetic basis is scarce. The prevalence of obesity is still rising globally, which is a great threat to public health.78 To date, no curative treatment is available for multiple sclerosis.79 Therefore, our study highlights the importance of early detection and prevention of neurodegenerative disorders, including MS, amongst overweight and obese people. New therapeutics for the two diseases could be explored, and it is promising to develop modalities targeting both diseases based on their shared genetic architecture. Moreover, our findings might provide a new understanding for the pathogenesis of the two diseases and inspire future research to continues to investigate their comorbidity. Our study could contribute to the improvement of disease management for patients with obesity or MS.

Caveats and limitations

Our study also has several limitations. Potential sample overlap might exist between the GWAS datasets from GIANT and IMSGC, both of which used samples from the Wellcome Trust Case Control Consortium (WTCCC) cohorts. However, this effect would be small since the samples from the WTCCC comprised ∼2.5% of the samples from the GIANT consortium. In addition, the results of this study were based on individuals of European ancestry, and therefore our findings have limited generalisability to other ancestral populations. The initially identified shared regions did not strictly pass the threshold after Bonferroni correction (0.05/1699), while it was exploratory and part of the regions could be validated by other methods, which were considered more reliable. Moreover, based on the information from our study, further in vitro and in vivo studies could be conducted to evaluate the function and mechanisms of GGNBP2 in obesity and MS.

Conclusion

In summary, we found a significant genetic correlation and identified shared risk SNPs between BMI and MS. We further identified a putative functional gene, GGNBP2, shared between obesity and MS. These findings could provide insights into the shared genetic architecture between obesity and MS, and contribute to a better understanding of their pathogenesis as well as the development of therapeutics.

Contributors

Conceptualisation and design: FWL, WHS, and HC; Funding acquisition: WHS and HC; Collection and assembly of data: JR, WTH, and JXW; Data analysis and interpretation: JR, WTH, JXW, RJZ, LJZ, YYM, YJW, MJM, HL, and QZL; Manuscript writing: RJZ, RJ, WTH and HC. RJ, WTH, and HC have directly accessed and verified the underlying data reported in the manuscript. All authors reviewed and approved the final manuscript.

Data sharing statement

Data source

GWAS summary statistics for MS are available by application from: https://imsgc.net/?page_id=31.

GWAS summary statistics for BMI are available by application from: http://portals.broadinstitute.org/collaboration/giant/index.php/GIANT_consortium_data_files.

The gene expression data are available from: https://www.ncbi.nlm.nih.gov/geo/.

The eQTL summary data for eQTLGen and GTEx are available from: https://www.eqtlgen.org/cis-eqtls.html; http://yanglab.westlake.edu.cn/software/smr/#eQTLsummarydata.

scRNA-seq data are available from: Whole blood: https://www.10xgenomics.com/resources/datasets; Lung and Spleen: https://doi.org/10.1186/s13059-019-1906-x; Brain: https://doi.org/10.1016/j.cell.2018.06.021; Small Intestinal Epithelium: https://singlecell.broadinstitute.org/single_cell/study/SCP44/small-intestinal-epithelium.

Code availability

The codes used in this study can be found at:

LDSC: https://github.com/bulik/ldsc.

PLINK: https://www.cog-genomics.org/plink/1.9.

MTAG: https://github.com/JonJala/mtag.

CPASSOC: http://hal.case.edu/∼xxz10/zhuweb/.

GSMR: http://cnsgenomics.com/software/gsmr/.

TwoSampleMR: https://mrcieu.github.io/TwoSampleMR/.

SMR: https://cnsgenomics.com/software/smr/#Overview.

LDSC-SEG: https://github.com/bulik/ldsc/wiki/Cell-type-specific-analyses.

MAGMA Celltyping: https://neurogenomics.github.io/MAGMA_Celltyping.

Declaration of interests

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.

Acknowledgments

This work was funded by the National Natural Science Foundation of China (82171698, 82170561, 81300279, and 81741067), the Program for High-level Foreign Expert Introduction of China (G2022030047L), the Natural Science Foundation for Distinguished Young Scholars of Guangdong Province (2021B1515020003), Natural Science Foundation of Guangdong Province (2022A1515012081), the Foreign Distinguished Teacher Program of Guangdong Science and Technology Department (KD0120220129), the Climbing Programme of Introduced Talents and High-level Hospital Construction Project of Guangdong Provincial People's Hospital (DFJH201803, KJ012019099, KJ012021143, and KY012021183), and in part by VA Clinical Merit and ASGE clinical research funds (FWL). We thank IMSGC and GIANT for providing access to the GWAS summary statistics. We acknowledge the included databases and authors that generously made the data publicly available. We thank American Journal Experts (AJE) for English language editing.

Footnotes

Appendix A

Supplementary data related to this article can be found at https://doi.org/10.1016/j.ebiom.2023.104647.

Contributor Information

Felix W. Leung, Email: Felix.Leung@va.gov.

Weihong Sha, Email: shaweihong@gdph.org.cn.

Hao Chen, Email: chenhao@gdph.org.cn.

Appendix A. Supplementary data

Supplementary Figs. S1–S9 and Tables S1–S8
mmc1.docx (1.5MB, docx)

References

  • 1.Chen X.-W., Ding G., Xu L., Li P. A glimpse at the metabolic research in China. Cell Metab. 2021;33:2122–2125. doi: 10.1016/j.cmet.2021.09.014. [DOI] [PubMed] [Google Scholar]
  • 2.Hu K., Staiano A.E. Trends in obesity prevalence among children and adolescents aged 2 to 19 years in the US from 2011 to 2020. JAMA Pediatr. 2022;176:1037–1039. doi: 10.1001/jamapediatrics.2022.2052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Ansari S., Haboubi H., Haboubi N. Adult obesity complications: challenges and clinical impact. Ther Adv Endocrinol Metab. 2020;11 doi: 10.1177/2042018820934955. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Mazon J.N., de Mello A.H., Ferreira G.K., Rezin G.T. The impact of obesity on neurodegenerative diseases. Life Sci. 2017;182:22–28. doi: 10.1016/j.lfs.2017.06.002. [DOI] [PubMed] [Google Scholar]
  • 5.Dobson R., Giovannoni G. Multiple sclerosis–a review. Eur J Neurol. 2019;26:27–40. doi: 10.1111/ene.13819. [DOI] [PubMed] [Google Scholar]
  • 6.Munger K.L., Bentzen J., Laursen B., et al. Childhood body mass index and multiple sclerosis risk: a long-term cohort study. Mult Scler. 2013;19:1323–1329. doi: 10.1177/1352458513483889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Høglund R.A.A., Meyer H.E., Stigum H., et al. Association of body mass index in adolescence and Young adulthood and long-term risk of multiple sclerosis: a population-based study. Neurology. 2021;97:e2253–e2261. doi: 10.1212/WNL.0000000000012957. [DOI] [PubMed] [Google Scholar]
  • 8.Hedström A.K., Olsson T., Alfredsson L. High body mass index before age 20 is associated with increased risk for multiple sclerosis in both men and women. Mult Scler. 2012;18:1334–1336. doi: 10.1177/1352458512436596. [DOI] [PubMed] [Google Scholar]
  • 9.Liu Z., Zhang T.T., Yu J., et al. Excess body weight during childhood and adolescence is associated with the risk of multiple sclerosis: a meta-analysis. Neuroepidemiology. 2016;47:103–108. doi: 10.1159/000450854. [DOI] [PubMed] [Google Scholar]
  • 10.Mokry L.E., Ross S., Timpson N.J., Sawcer S., Davey Smith G., Richards J.B. Obesity and multiple sclerosis: a mendelian randomization study. PLoS Med. 2016;13 doi: 10.1371/journal.pmed.1002053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Vandebergh M., Dubois B., Goris A. Effects of vitamin D and body mass index on disease risk and relapse hazard in multiple sclerosis: a mendelian randomization study. Neurol Neuroimmunol Neuroinflamm. 2022;9 doi: 10.1212/NXI.0000000000001165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Harroud A., Mitchell R.E., Richardson T.G., et al. Childhood obesity and multiple sclerosis: a Mendelian randomization study. Mult Scler. 2021;27:2150–2158. doi: 10.1177/13524585211001781. [DOI] [PubMed] [Google Scholar]
  • 13.Jacobs B.M., Noyce A.J., Giovannoni G., Dobson R. BMI and low vitamin D are causal factors for multiple sclerosis: a Mendelian Randomization study. Neurol Neuroimmunol Neuroinflamm. 2020;7 doi: 10.1212/NXI.0000000000000662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Stampanoni Bassi M., Iezzi E., Buttari F., et al. Obesity worsens central inflammation and disability in multiple sclerosis. Mult Scler. 2020;26:1237–1246. doi: 10.1177/1352458519853473. [DOI] [PubMed] [Google Scholar]
  • 15.Matthews P.M. Chronic inflammation in multiple sclerosis—seeing what was always there. Nat Rev Neurol. 2019;15:582–593. doi: 10.1038/s41582-019-0240-y. [DOI] [PubMed] [Google Scholar]
  • 16.Guerrero-García J.J., Carrera-Quintanar L., López-Roa R.I., Márquez-Aguirre A.L., Rojas-Mayorquín A.E., Ortuño-Sahagún D. Multiple sclerosis and obesity: possible roles of adipokines. Mediators Inflamm. 2016;2016 doi: 10.1155/2016/4036232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lee E.B., Mattson M.P. The neuropathology of obesity: insights from human disease. Acta Neuropathol. 2014;127:3–28. doi: 10.1007/s00401-013-1190-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Mirza A., Forbes J.D., Zhu F., et al. The multiple sclerosis gut microbiota: a systematic review. Mult. Scler. Relat Disord. 2020;37 doi: 10.1016/j.msard.2019.101427. [DOI] [PubMed] [Google Scholar]
  • 19.Yengo L., Sidorenko J., Kemper K.E., et al. Meta-analysis of genome-wide association studies for height and body mass index in ∼700000 individuals of European ancestry. Hum Mol Genet. 2018;27:3641–3649. doi: 10.1093/hmg/ddy271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Beecham A.H., Patsopoulos N.A., Xifara D.K., et al. Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat Genet. 2013;45:1353–1360. doi: 10.1038/ng.2770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.GTEx Consortium Genetic effects on gene expression across human tissues. Nature. 2017;550:204–213. doi: 10.1038/nature24277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Finucane H.K., Reshef Y.A., Anttila V., et al. Heritability enrichment of specifically expressed genes identifies disease-relevant tissues and cell types. Nat Genet. 2018;50:621–629. doi: 10.1038/s41588-018-0081-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Võsa U., Claringbould A., Westra H.-J., et al. Unraveling the polygenic architecture of complex traits using blood eQTL metaanalysis. BioRxiv. 2018:447367. doi: 10.1101/447367v1. [DOI] [Google Scholar]
  • 24.Madissoon E., Wilbrey-Clark A., Miragaia R., et al. scRNA-seq assessment of the human lung, spleen, and esophagus tissue stability after cold preservation. Genome Biol. 2020;21:1–16. doi: 10.1186/s13059-019-1906-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zeisel A., Hochgerner H., Lönnerberg P., et al. Molecular architecture of the mouse nervous system. Cell. 2018;174:999–1014.e22. doi: 10.1016/j.cell.2018.06.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Skene N.G., Grant S.G. Identification of vulnerable cell types in major brain disorders using single cell transcriptomes and expression weighted cell type enrichment. Front Neurosci. 2016;10:16. doi: 10.3389/fnins.2016.00016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Finucane H.K., Bulik-Sullivan B., Gusev A., et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat Genet. 2015;47:1228–1235. doi: 10.1038/ng.3404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Consortium G.P. A global reference for human genetic variation. Nature. 2015;526:68. doi: 10.1038/nature15393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Bulik-Sullivan B., Finucane H.K., Anttila V., et al. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015;47:1236–1241. doi: 10.1038/ng.3406. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Yao Y., Li C., Meng P., et al. An atlas of genetic correlations between gestational age and common psychiatric disorders. Autism Res. 2022;15(6):1008–1017. doi: 10.1002/aur.2719. [DOI] [PubMed] [Google Scholar]
  • 31.Lu Q., Li B., Ou D., et al. A powerful approach to estimating annotation-stratified genetic covariance via GWAS summary statistics. Am J Hum Genet. 2017;101:939–964. doi: 10.1016/j.ajhg.2017.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Turley P., Walters R.K., Maghzian O., et al. Multi-trait analysis of genome-wide association summary statistics using MTAG. Nat Genet. 2018;50:229–237. doi: 10.1038/s41588-017-0009-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Zhu X., Feng T., Tayo B.O., et al. Meta-analysis of correlated traits via summary statistics from GWASs with an application in hypertension. Am J Hum Genet. 2015;96:21–36. doi: 10.1016/j.ajhg.2014.11.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Li X., Zhu X. Cross-phenotype association analysis using summary statistics from GWAS. Methods Mol Biol. 2017;1666:455–467. doi: 10.1007/978-1-4939-7274-6_22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Shi H., Mancuso N., Spendlove S., Pasaniuc B. Local genetic correlation gives insights into the shared genetic architecture of complex traits. Am J Hum Genet. 2017;101:737–751. doi: 10.1016/j.ajhg.2017.09.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Berisa T., Pickrell J.K. Approximately independent linkage disequilibrium blocks in human populations. Bioinformatics. 2016;32:283. doi: 10.1093/bioinformatics/btv546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Pickrell J.K., Berisa T., Liu J.Z., Ségurel L., Tung J.Y., Hinds D.A. Detection and interpretation of shared genetic influences on 42 human traits. Nat Genet. 2016;48:709–717. doi: 10.1038/ng.3570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Burgess S., Thompson S.G. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol. 2017;32:377–389. doi: 10.1007/s10654-017-0255-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Burgess S., Butterworth A., Thompson S.G. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013;37:658–665. doi: 10.1002/gepi.21758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Bowden J., Davey Smith G., Haycock P.C., Burgess S. Consistent estimation in mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40:304–314. doi: 10.1002/gepi.21965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Hartwig F.P., Davey Smith G., Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol. 2017;46:1985–1998. doi: 10.1093/ije/dyx102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Verbanck M., Chen C.Y., Neale B., Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50:693–698. doi: 10.1038/s41588-018-0099-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Li B., Martin E.B. An approximation to the F distribution using the chi-square distribution. Comput Stat Data Anal. 2002;40:21–26. [Google Scholar]
  • 44.Brion M.J., Shakhbazov K., Visscher P.M. Calculating statistical power in Mendelian randomization studies. Int J Epidemiol. 2013;42:1497–1501. doi: 10.1093/ije/dyt179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Gazal S., Marquez-Luna C., Finucane H.K., Price A.L. Reconciling S-LDSC and LDAK functional enrichment estimates. Nat Genet. 2019;51:1202–1204. doi: 10.1038/s41588-019-0464-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.de Leeuw C.A., Mooij J.M., Heskes T., Posthuma D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol. 2015;11 doi: 10.1371/journal.pcbi.1004219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Skene N.G., Bryois J., Bakken T.E., et al. Genetic identification of brain cell types underlying schizophrenia. Nat Genet. 2018;50:825–833. doi: 10.1038/s41588-018-0129-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Zhu Z., Zhang F., Hu H., et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat Genet. 2016;48:481–487. doi: 10.1038/ng.3538. [DOI] [PubMed] [Google Scholar]
  • 49.GTEx Consortium The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369:1318–1330. doi: 10.1126/science.aaz1776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Yang J., Lee S.H., Goddard M.E., Visscher P.M. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88:76–82. doi: 10.1016/j.ajhg.2010.11.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Aguilera C.M., Gomez-Llorente C., Tofe I., Gil-Campos M., Cañete R., Gil Á. Genome-wide expression in visceral adipose tissue from obese prepubertal children. Int J Mol Sci. 2015;16:7723–7737. doi: 10.3390/ijms16047723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Han M.H., Lundgren D.H., Jaiswal S., et al. Janus-like opposing roles of CD47 in autoimmune brain inflammation in humans and mice. J Exp Med. 2012;209:1325–1334. doi: 10.1084/jem.20101974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Kemppinen A., Kaprio J., Palotie A., Saarela J. Systematic review of genome-wide expression studies in multiple sclerosis. BMJ Open. 2011;1 doi: 10.1136/bmjopen-2011-000053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Gianfrancesco M.A., Glymour M.M., Walter S., et al. Causal effect of genetic variants associated with body mass index on multiple sclerosis susceptibility. Am J Epidemiol. 2017;185:162–171. doi: 10.1093/aje/kww120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Bowden J., Davey Smith G., Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44:512–525. doi: 10.1093/ije/dyv080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Burgess S., Foley C.N., Allara E., Staley J.R., Howson J.M. A robust and efficient method for Mendelian randomization with hundreds of genetic variants. Nat Commun. 2020;11:1–11. doi: 10.1038/s41467-019-14156-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Zhu Z., Zheng Z., Zhang F., et al. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun. 2018;9:1–12. doi: 10.1038/s41467-017-02317-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Willeumier K.C., Taylor D.V., Amen D.G. Elevated BMI is associated with decreased blood flow in the prefrontal cortex using SPECT imaging in healthy adults. Obesity. 2011;19:1095–1097. doi: 10.1038/oby.2011.16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Opel N., Thalamuthu A., Milaneschi Y., et al. Brain structural abnormalities in obesity: relation to age, genetic risk, and common psychiatric disorders. Mol Psychiatr. 2021;26:4839–4852. doi: 10.1038/s41380-020-0774-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Green E., Jacobson A., Haase L., Murphy C. Reduced nucleus accumbens and caudate nucleus activation to a pleasant taste is associated with obesity in older adults. Brain Res. 2011;1386:109–117. doi: 10.1016/j.brainres.2011.02.071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Dendrou C.A., Fugger L., Friese M.A. Immunopathology of multiple sclerosis. Nat Rev Immunol. 2015;15:545–558. doi: 10.1038/nri3871. [DOI] [PubMed] [Google Scholar]
  • 62.Balfour H.H., Schmeling D.O., Grimm-Geris J.M. The promise of a prophylactic Epstein–Barr virus vaccine. Pediatr Res. 2020;87:345–352. doi: 10.1038/s41390-019-0591-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Odoardi F., Sie C., Streyl K., et al. T cells become licensed in the lung to enter the central nervous system. Nature. 2012;488:675–679. doi: 10.1038/nature11337. [DOI] [PubMed] [Google Scholar]
  • 64.Friese M.A., Fugger L. Pathogenic CD8+ T cells in multiple sclerosis. Ann Neurol. 2009;66:132–141. doi: 10.1002/ana.21744. [DOI] [PubMed] [Google Scholar]
  • 65.Comi G., Bar-Or A., Lassmann H., et al. Role of B cells in multiple sclerosis and related disorders. Ann Neurol. 2021;89:13–23. doi: 10.1002/ana.25927. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Jelcic I., Al Nimer F., Wang J., et al. Memory B cells activate brain-homing, autoreactive CD4+ T cells in multiple sclerosis. Cell. 2018;175:85–100.e23. doi: 10.1016/j.cell.2018.08.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67.Van Langelaar J., Rijvers L., Smolders J., Van Luijn M.M. B and T cells driving multiple sclerosis: identity, mechanisms and potential triggers. Front Immunol. 2020;11:760. doi: 10.3389/fimmu.2020.00760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Piacente F., Bottero M., Benzi A., et al. Neuroprotective potential of dendritic cells and sirtuins in multiple sclerosis. Int J Mol Sci. 2022;23:4352. doi: 10.3390/ijms23084352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69.Gross C.C., Schulte-Mecklenbeck A., Wiendl H., et al. Regulatory functions of natural killer cells in multiple sclerosis. Front Immunol. 2016;7:606. doi: 10.3389/fimmu.2016.00606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Ribatti D., Tamma R., Annese T. Mast cells and angiogenesis in multiple sclerosis. Inflamm Res. 2020;69:1103–1110. doi: 10.1007/s00011-020-01394-2. [DOI] [PubMed] [Google Scholar]
  • 71.Lumeng C.N., DeYoung S.M., Bodzin J.L., Saltiel A.R. Increased inflammatory properties of adipose tissue macrophages recruited during diet-induced obesity. Diabetes. 2007;56:16–23. doi: 10.2337/db06-1076. [DOI] [PubMed] [Google Scholar]
  • 72.Olsson T., Hillert J. The genetics of multiple sclerosis and its experimental models. Curr Opin Neurol. 2008;21:255–260. doi: 10.1097/WCO.0b013e3282fd10cc. [DOI] [PubMed] [Google Scholar]
  • 73.Hedström A.K., Bomfim I.L., Barcellos L., et al. Interaction between adolescent obesity and HLA risk genes in the etiology of multiple sclerosis. Neurology. 2014;82:865–872. doi: 10.1212/WNL.0000000000000203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Zhan A., Lei B., Wu H., et al. GGNBP2 suppresses the proliferation, invasion, and migration of human glioma cells. Oncol Res. 2017;25:831. doi: 10.3727/096504016X14816726393937. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Lan Z.-J., Hu Y., Zhang S., et al. GGNBP2 acts as a tumor suppressor by inhibiting estrogen receptor α activity in breast cancer cells. Breast Cancer Res Treat. 2016;158:263–276. doi: 10.1007/s10549-016-3880-2. [DOI] [PubMed] [Google Scholar]
  • 76.Yang Z., Wang Y., Ma L. Effects of gametogenetin-binding protein 2 on proliferation, invasion and migration of prostate cancer PC-3 cells. Andrologia. 2020;52 doi: 10.1111/and.13488. [DOI] [PubMed] [Google Scholar]
  • 77.Li C., Ou R., Wei Q., Shang H. Shared genetic links between amyotrophic lateral sclerosis and obesity-related traits: a genome-wide association study. Neurobiol Aging. 2021;102:211.e1–211.e9. doi: 10.1016/j.neurobiolaging.2021.01.023. [DOI] [PubMed] [Google Scholar]
  • 78.Agha M., Agha R. The rising prevalence of obesity: part A: impact on public health. Int J Surg Oncol (N Y) 2017;2:e17. doi: 10.1097/IJ9.0000000000000017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 79.Dumitrescu L., Constantinescu C.S., Tanasescu R. Siponimod for the treatment of secondary progressive multiple sclerosis. Expet Opin Pharmacother. 2019;20:143–150. doi: 10.1080/14656566.2018.1551363. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figs. S1–S9 and Tables S1–S8
mmc1.docx (1.5MB, docx)

Articles from eBioMedicine are provided here courtesy of Elsevier

RESOURCES