Causal effects of gut microbiota, metabolites, immune cells, liposomes, and inflammatory proteins on anorexia nervosa: A mediation joint multi-omics Mendelian randomization analysis
Zeyang Li,Tianyu Bi
DOI: https://doi.org/10.1016/j.jad.2024.09.115
2024-09-17
Abstract:Background: Anorexia nervosa (AN) is a significant psychological disorder influenced by environmental and genetic elements. Emerging research highlights the pivotal role of the gut microbiome in the development of diverse mental health conditions. This study aims to explore the causal effects and interactions of the gut microbiome, metabolites, immune cells, lipids, and inflammatory proteins on the risk of anorexia nervosa through mediation and multi-omics Mendelian Randomization (MR) analysis. Methods: This study used data from the FinnGen genome-wide association study (GWAS) of AN (N = 402,625), integrated with GWAS data on 473 of gut microbiota (N = 5959), 233 metabolites (N = 136,016), 731 immune cells (N = 3757), 179 lipids (N = 7174), and 91 inflammatory proteins (N = 14,824). This study used the univariate MR (UVMR), mediation MR analysis, and sensitivity analysis to assess the potential causal associations between these biomarkers and AN. Results: The inverse variance weighted (IVW) results suggest that 25 gut microbiota have causal effects on AN. Firmicutes E (OR: 0.294, 95 % CI: 0.107-0.806, P = 0.017), RUG147 (OR: 0.386, 95 % CI: 0.151-0.990, P = 0.048), CAG-977 (OR: 0.562, 95 % CI: 0.378-0.837, P = 0.005), Desulfobacterota A (OR: 0.651, 95 % CI: 0.466-0.909, P = 0.012), CAG-269 sp002372935 (OR: 0.673, 95 % CI: 0.483-0.937, P = 0.019), Klebsiella (OR: 0.684, 95 % CI: 0.566-0.827, P = 0.00009), Desulfovibrionia (OR: 0.706, 95 % CI: 0.538-0.926, P = 0.012), Klebsiella pneumoniae (OR: 0.737, 95 % CI: 0.600-0.906, P = 0.004), Desulfovibrionales (OR: 0.786, 95 % CI: 0.631-0.979, P = 0.031), CAG-776 (OR: 0.787, 95 % CI: 0.632-0.980, P = 0.032), Desulfovibrionaceae (OR: 0.788, 95 % CI: 0.635-0.978, P = 0.030). 13 gut microbiota were risk factors for AN, including Parachlamydiales (OR: 3.134，95%CI: 1.185-8.287, P = 0.021), Paenibacillus J (OR: 2.366，95%CI: 1.305-4.29, P = 0.005), Gillisia (OR: 1.947，95%CI: 1.135-3.339, P = 0.016), UBA1191 (OR: 1.856，95%CI: 1.221-2.822, P = 0.004), UBA7703 (OR: 1.843，95%CI: 1.032-3.289, P = 0.039), Faecalicatena sp002161355 (OR: 1.788，95%CI: 1.114-2.870, P = 0.016), Johnsonella ignava (OR: 1.742，95%CI: 1.031-2.944, P = 0.038), Staphylococcus aureus (OR: 1.614, 95%CI: 1.007-2.588, P = 0.047), Comamonas (OR: 1.522，95%CI: 1.004-2.307, P = 0.048), Ruminococcus D (OR: 1.24，95%CI: 1.050-1.464, P = 0.011), CAG-349 (OR: 1.198，95%CI: 1.048-1.370, P = 0.008), Ruminococcus D bicirculans (OR: 1.175，95%CI: 1.001-1.379, P = 0.048), CAG-177 (OR: 1.272，95%CI: 1.077-1.503, P = 0.005). Reverse MR analysis showed that causal effect of AN on 18 gut microbiota, but to a lesser extent. 12 metabolites have causal effects on AN. There are 7 protective factors, including glucose levels (OR: 0.700, 95%CI: 0.550-0.893, P = 0.004), isoleucine levels (OR: 0.769, 95%CI: 0.602-0.983, P = 0.036), phospholipids in large VLDL (OR: 0.856, 95%CI: 0.736-0.996, P = 0.044), total lipids in large VLDL (OR: 0.860, 95%CI: 0.740-0.999, P = 0.049), total lipids in small VLDL (OR: 0.863, 95%CI: 0.751-0.992, P = 0.038), free cholesterol in small VLDL (OR: 0.86, 95%CI: 0.752-0.996, P = 0.044), and free cholesterol in medium VLDL (OR: 0.866, 95%CI: 0.752-0.998, P = 0.047). There are 5 risk factors, including estimated degree of unsaturation (OR: 1.174, 95%CI: 1.009-1.367, P = 0.039), free cholesterol to total lipids ratio in small VLDL (OR: 1.199, 95%CI: 1.017-1.414, P = 0.031), phospholipids to total lipids ratio in small VLDL (OR: 1.216, 95%CI: 1.008-1.467, P = 0.041), total cholesterol levels in small HDL (OR: 1.241, 95%CI: 1.008-1.530, P = 0.042), and phospholipids to total lipids ratio in medium VLDL (OR: 1.280, 95%CI: 1.055-1.553, P = 0.012). Reverse MR analysis showed that AN had a causal effect on 15 metabolites. Mediation analysis reveals that the estimated degree of unsaturation mediates 0.69 % of the effect of Klebsiella pneumoniae on AN. Total lipids in small VLDL mediate 0.358 % of the effect of CAG-177 on AN, with a mediated proportion of 1.490 %. The mediation proportions for Estimated degree of unsaturation and Total lipids in small VLDL are relatively small. 36 immune cells have causal effects on AN. There are 7 protective factors, including Switched memory B cells %B cell (OR: 0.892，95%CI: 0.801-0.994, P = 0.038), CD127-CD8+ T cell absolute count (OR: 0.888，95%CI: 0.789-1.000, P = 0.049), IgD + CD24- B cell (OR: 0.917，95%CI: 0.862-0.975, P = 0.006), HVEM+ T cell (OR: 0.945，95%CI: 0.894-0.999, P = 0.045), CD40 + CD14 + CD16- monocyte (OR: 0.937，95%CI: 0.882-0.996, P = 0.038), CD64 + CD14 + CD16- monocyte (OR: 0.966，95%CI: 0.939-0.993, P = 0.016), CD8+ natural killer T cells (OR: 0.911，95%CI: 0.836-0.992, P = 0.032), HLA-DR+ T cells (OR: 0.921，95%CI: 0.866-0.980, P = 0.010), CD28-CD8+ T cells (OR: 0.886，95%CI: 0.792-0.991, P = 0.034). There are 26 risk factors. Reverse MR analysis showed that AN had a causal effect on 31 immune cells. AN increases the expression levels of five types of immune cells, including CD40 + CD14-CD16+ monocytes (OR: 1.087，95%CI: 1.004-1.177, P = 0.041), PDL-1+ CD14-CD16+ monocytes (OR: 1.082，95%CI: 1.002-1.168, P = 0.046), CD45+ CD33dim HLA-DR+ cells (OR: 1.145，95%CI: 1.019-1.287, P = 0.023), CD45+ basophils (OR: 1.164，95%CI: 1.036-1.307, P = 0.011), CD8+ natural killer T cells (OR: 1.102, 95%CI: 1.015-1.196, P = 0.020), and also decreases the expression levels of 26 immune cells. 6 liposomes showed exhibit protective effects against AN, including phosphatidylcholine (18:0_20:3) levels (OR: 0.852, 95%CI: 0.740-0.981, P = 0.026), phosphatidylcholine (O-18:2_18:1) levels (OR: 0.800, 95%CI: 0.672-0.952, P = 0.012), phosphatidylinositol (18:0_18:1) levels (OR: 0.873, 95%CI: 0.773-0.986, P = 0.029), phosphatidylinositol (18:1_18:2) levels (OR: 0.844, 95%CI: 0.734-0.971, P = 0.018), sphingomyelin (d38:1) levels (OR: 0.903，95%CI: 0.820-0.995, P = 0.039), and triacylglycerol (56:4) levels (OR: 0.786, 95%CI: 0.660-0.936, P = 0.007). There are 3 risk factors, including diacylglycerol (16:1_18:1) levels (OR: 1.208, 95%CI: 1.040-1.404, P = 0.014), phosphatidylcholine (18:1_18:3) levels (OR: 1.237, 95%CI: 1.003-1.526, P = 0.047), and phosphatidylinositol (16:0_20:4) levels (OR: 1.148, 95%CI: 1.003-1.314, P = 0.045). Reverse MR analysis showed that AN had a causal effect on 3 phosphatidylcholine (15:0_18:2) levels (OR: 1.075, 95%CI: 1.001-1.154, P = 0.048), phosphatidylcholine (O-16:2_18:0) levels (OR: 1.078, 95%CI: 1.002-1.159, P = 0.043), and triacylglycerol (51:1) levels (OR: 0.919, 95%CI: 0.850-0.994, P = 0.035). 6 inflammatory proteins have causal effects on AN, with protective factors including Glial cell line-derived neurotrophic factor levels (OR: 0.822，95%CI: 0.692-0.978, P = 0.027) and Interleukin-15 receptor subunit alpha levels (OR: 0.886, 95%CI: 0.789-0.995, P = 0.041) and risk factors including CC motif chemokine 4 levels (OR: 1.126, 95%CI: 1.011-1.254, P = 0.031), Interleukin-12 subunit beta levels (OR: 1.135, 95%CI: 1.033-1.248, P = 0.008), Monocyte chemoattractant protein-1 levels (OR: 1.152, 95%CI: 1.010-1.314, P = 0.035), and Sulfotransferase 1A1 levels (OR: 1.166, 95%CI: 1.006-1.351, P = 0.042). Reverse MR analysis showed that AN had a causal effect on Transforming growth factor-alpha (OR: 1.054，95%CI: 1.010-1.101, P = 0.016). Conclusions: This study used large-scale and novel GWAS data, for the first time reveals through mediation analysis and multi-omics MR analysis the roles of gut microbiota, metabolites, immune cells, lipids, and inflammatory proteins in the pathogenesis of AN. These findings provide new biomarkers and targets for further prevention and treatment strategies.