Independent effect and joint effect of polygenic liabilities for schizophrenia and bipolar disorder on cognitive aging and education attainment

To elucidate the specific and shared genetic background of schizophrenia (SCZ) and bipolar disorder (BPD) to better understand their nosology, this study explored the independent and joint effects of polygenic liabilities for SCZ and BPD on cognitive aging and educational attainment among a collection of 80318 unrelated community participants from the Taiwan Biobank. Using the Psychiatric Genomics Consortium meta-analysis as a discovery sample, we calculated the polygenic risk score (PRS) for SCZ (PRSSCZ) and BPD (PRSBPD), shared PRS between SCZ and BPD (PRSSCZ+BPD), and SCZ-specific, differentiated from BPD, PRS (PRSSCZvsBPD). Based on the sign concordance of the susceptibility variants with SCZ and BPD, PRSSCZ was split into PRSSCZ_concordant and PRSSCZ_discordant and PRSBPD was split into PRSBPD_concordant and PRSBPD_discordant. Linear regression models were used to estimate the association with cognitive aging as measured by the Mini-Mental State Examination (MMSE) in individuals aged [&ge;] 60 years. Ordinal logistic regression models were used to estimate the association with educational attainment. PRSSCZ was negatively associated with MMSE (beta=-0.063, p<0.001), while PRSBPD was positively associated with MMSE (beta=0.04, p=0.01). A larger difference between PRSSCZ and PRSBPD was associated with lower MMSE scores (beta=-0.052, p<0.001). Both PRSSCZ_concordant and PRSSCZ_discordant were negatively associated with MMSE scores, without a synergistic effect. There was a complex interaction between PRSBPD_concordant and PRSBPD_discordant on the MMSE scores. PRSSCZ+BPD (beta=-0.09, p=0.01) and PRSSCZvsBPD (beta=-0.13, p<0.001) predicted a decrease in MMSE scores during the follow-up. PRSSCZ, PRSBPD, and PRSSCZ+BPD were positively associated with educational attainment, whereas PRSSCZvs BPD was negatively associated with educational attainment. This study provides evidence for the contrasting effect of polygenic liabilities for SCZ and BPD on cognitive aging and partially supports the hypothesis that the heterogeneity of SCZ and the positive association of polygenic liability for SCZ with education might be attributed to the shared part with BPD.