Abstract
The aim of this study was to assess the causal relationship between immune cells and anorexia nervosa (AN) by Mendelian randomization (MR). Data on immune cell phenotypes and AN were obtained from genome-wide association studies. Next, single nucleotide polymorphisms included in the MR analysis were screened according the basic assumptions. Furthermore, inverse variance weighted was used as the main method for MR analysis to evaluate the causal effect of immune cell phenotypes on AN. Finally, MR-Egger intercept, Cochran Q, and leave-one-out sensitivity analyses were used to assess horizontal pleiotropy, heterogeneity, and robustness, respectively. The MR analysis showed that NKT %lymphocyte (OR 1.201, 95% CI = 1.021-1.411, P = .027), NKT %T cell (OR 1.258, 95%CI 1.043-1.519, P = .017), double negative (DN) (CD4-CD8-) NKT %lymphocyte (OR 1.235, 95%CI 1.074-1.420, P = .003), DN (CD4-CD8-) NKT %T cell (OR 1.222, 95%CI 1.060-1.410, P = .006), CD8dim NKT absolute count (OR 1.225, 95%CI 1.045-1.436, P = .012), CD8dim NKT %lymphocyte (OR 1.214, 95%CI 1.043-1.414, P = .012), CD8dim NKT %T cell (OR 1.215, 95%CI 1.035-1.425, P = .017), CD16-CD56 on NKT (OR 1.193, 95%CI 1.014-1.402, P = .033), CD28 + CD45RA + CD8br %T cell (OR 1.020, 95%CI 1.002-1.037, P = .025) were associated with increased genetic susceptibility to AN. MR-Egger showed no horizontal pleiotropy (P ≥ .05). Cochran Q and sensitivity analysis showed that the results were not heterogeneous and were robust. This MR analysis revealed 9 immune cell phenotypes related to increased genetic susceptibility to AN, emphasizing the importance of NKT and CD8 in AN. This finding provides new insights for understanding the pathogenesis of AN and developing immune-targeted drugs.