Mendelian Randomization and GWAS Meta Analysis Revealed the Risk-Increasing Effect of Schizophrenia on Cancers

SIMPLE SUMMARY: The relationship between schizophrenia and tumors has sparked much interest and controversy. On the one hand, the health of patients with schizophrenia is affected by a variety of risk factors associated with cancer development, such as smoking, alcohol, and drug abuse. On the other hand, early investigations have found that patients with schizophrenia have a lower cancer incidence than the overall population. This phenomenon has prompted new theories on the processes underlying the protective effect. To explore the connection between schizophrenia and tumors, we used two-sample Mendelian randomization and GWAS meta-analysis methods on the GWAS summary data to assess potential genetic links between schizophrenia and 13 cancers. It was revealed that schizophrenia may lead to an increased risk of breast, ovarian, and thyroid cancers. Furthermore, the thyroid-stimulating hormone level is impacted by schizophrenia, which may further affect the estrogen and thyroid hormone levels. Meanwhile, based on our results, AS3MT, SFXN2, and PCCB may be potential biomarkers for preventing breast and thyroid cancers in patients diagnosed with schizophrenia. Therefore, we suggest that patients with schizophrenia should pay close attention to early risk warnings for breast, ovarian, and thyroid cancers. ABSTRACT: The causal relationship between cancer and Schizophrenia (SCZ) remains controversial. Some researchers have found that SCZ is a cancer-preventive factor in cohort studies or meta-analyses, whereas others have found the opposite. To understand more about this issue, we used two-sample Mendelian randomization (2SMR) on available GWAS summary results to evaluate potential genetic connections between SCZ and 13 cancers. We discovered that the genetic susceptibility to schizophrenia lead to an increasing risk of breast cancer (odds ratio [OR] per log-odds increase in schizophrenia risk: 1.049, 95% confidence interval [CI]:1.023–1.075; p = 0.00012; FDR = 0.0017), ovarian cancer (OR, 1.326; 95% CI, 1.267–1.387; p = 0.0007; FDR = 0.0045), and thyroid cancer (OR, 1.575; 95% CI, 1.048–2.365; p = 0.0285; FDR = 0.123). Secondly, we performed a meta-analysis based on the GWAS summary statistics of SCZ and the three significant cancers. Next, we associated genetic variants to genes using two gene mapping strategies: (a) positional mapping based on genomic proximity and (b) expression quantitative trait loci (eQTL) mapping based on gene expression linkage across multiple tissues. As a result, we identified 114 shared loci and 437 shared genes in three groups, respectively. Functional enrichment analysis shows that the most enriched biological pathways are related to epigenetic modification. In addition, we noticed that SCZ would affect the level of thyroid-stimulating hormone (OR, 1.095; 95% CI, 1.006–1.191; p = 0.0354; FDR = 0.177), which may further affect the level of estrogen and the risk of the above three cancers. In conclusion, our findings under the 2SMR assumption provide crucial insights into the risk-increasing effect of SCZ on three cancers’ risk. Furthermore, these results may provide insights into understanding the genetic predisposition and underlying biological pathways of comorbid SCZ and cancers.