The aim of this study was to examine the genetic relationship between six psychiatric disorders (depression, bipolar disorder, schizophrenia, ASD, ADHD and anxiety) and COVID-19 susceptibility and severity using post-GWAS statistical methods. We found evidence of significant genetic correlations between ADHD, depression and COVID-19 susceptibility and severity, and between anxiety and severity after correcting for multiple testing. While prior studies have reported significant genetic correlations between depression and COVID-19 outcomes, (10) our study is the first to report significant positive correlations between ADHD, anxiety, and COVID-19. In contrast, we did not find significant genome-wide genetic association between bipolar disorder, schizophrenia and ASD and any COVID-19 outcome, which is in line with current literature. (32, 33)
To further investigate the shared genetic architecture between our psychiatric disorders of interest and COVID-19, we examined specific shared genomic regions. We found two regions that were shared between depression and COVID-19 susceptibility, while no region was shared with severity. Gene-based tests conducted on these regions revealed two genes in chromosome 17: MED24 and THRA. Alternatively, despite the lack of global genetic correlation between bipolar disorder and either COVID-19 susceptibility or severity, we found evidence of two shared genomic regions, one shared with susceptibility and the other with severity. While no significant genes were shared between bipolar disorder and COVID-19 severity, six genes were identified for COVID-19 susceptibility, among them MED24 and THRA – as was seen with depression. MED24, which codifies for a mediator complex, has been associated with white cells count. (34) However, the role of this gene in psychiatric disorders is still unclear. THRA is a thyroid hormone receptor (THR) gene. (35) THRs are involved in brain development and function, (36) and alterations in THRs, including THRA, have been reported in many psychiatric disorders. (37, 38) Interestingly, mutations in THRA have been associated to a reduced white blood cell count (34) and B cell deficiency in mice, (39) suggesting a role in the immune response. Additionally, THRA has been shown to play a crucial role in lung regeneration after infection. (40) Thus, alterations in THRA could increase the risk of mood disorders while also leading to a compromised immune system, what may underlie, at least in part, the reported increased susceptibility to COVID-19 infection reported in people with mood disorders in some epidemiological studies. (5)
This is further supported by our observed tendency towards a causal association between depression risk and increased COVID-19 susceptibility, but not severity, which is in line with current literature. (10, 12) Given that people with mood disorders have a higher risk of subclinical hypothyroidism, (41, 42) we could hypothesize that a reduced thyroid functioning, probably caused by THRA malfunction, could be increasing the susceptibility to COVID-19 in this population. However, a recent MR study suggested that COVID-19 susceptibility causes hypothyroidism and not the other way around. (43) Although the exact role of THRA in both mood disorders and COVID-19 susceptibility is unclear, evidence suggests that altered thyroid function may play a role in this association and warrants further investigation.
The lack of genetic correlation between schizophrenia, ASD and any COVID-19 suggested that genome-wide common genetic effects do not explain the association between these disorders and COVID-19. Nevertheless, we found evidence of 4 regions in chromosomes 4, 19 and 17 shared between schizophrenia and COVID-19. The region on chromosome 17 was also shared between ASD and COVID-19 susceptibility and severity. Gene-based analysis conducted on this region identified 9 genes that were significantly enriched in schizophrenia, ASD and COVID-19. These include CRHR1, the corticotropin releasing hormone receptor 1, which plays a role in the hypothalamic-pituitary-adrenal (HPA) axis activation and is crucial in the physiological response to stress. (44) Moreover, CRHR1 is important for the immune response as it exerts both indirect anti-inflammatory effects through the production of cortisol, which supress immune function, and direct proinflammatory effects on immune cells. (45)
Given its role in stress response, CRHR1 has been associated with several psychiatric disorders. For instance, increased methylation levels in CRHR1 have been linked to more negative effects on health care workers’ mental health during the COVID-19 pandemic. (46) Alterations in CRHR1 have also been linked to higher levels of proinflammatory cytokines in people with schizophrenia. (47) Results from animal studies showed that blockage of CRHR1 receptor in mice infected with streptococcus pneumonia increased neutrophil infiltration in lungs but did not confer resistance to the infection, (48) and mutations in CRHR1 have been associated with neutrophil and lymphocyte count. (49) Neutrophils are important in the fight against pulmonary infections, but dysregulations in neutrophil’s function are linked to uncontrolled inflammatory reactions that can result in lung damage and sepsis. (50) It is well known that pneumonia is the most common cause of COVID-19-related hospitalization. (51) Thus, it is possible that in people with psychiatric disorders, and specifically schizophrenia and ASD, alterations in CRHR1 are contributing to the increased severity and mortality reported in observational studies (2) through the dysregulation of neutrophil’s function. In order to validate these biological pathways, we sought for drug-gene interactions with the CRHR1 gene. We identified interactions with CRF1 antagonists such as Verucerfont, Pexacerfont or Emicerfont, which are under investigation as potential treatments for stress-induced alcoholism, (52, 53) and also with the antidepressant Fluoxetine. Interestingly, Fluoxetine has been reported to exert antiviral and anti-inflammatory activities against SARS-CoV-2 infection. (54, 55) Furthermore, CRHR1 also interacted to Budesonide and Triamcinolone, corticosteroids used to treat asthma, and Telavancin, an antibiotic used to treat pneumonia.
Surprisingly, despite global genetic correlations, we did not find any genomic region shared between ADHD or anxiety and any COVID-19. The polygenic architecture of these traits might explain the lack of shared genomic regions; while we could observe a strong association at a global level, caused by multiple genes with tiny individual effects, it might be challenging to identify specific genomic regions with substantial effects shared between our traits. Another reason might be the lack of statistical power to detect causal variants in the anxiety and ADHD GWAS, which is key to identify shared genomic regions with confidence. (56) Nevertheless, we observed a potential causal association between ADHD and increased susceptibility and severity of COVID-19, which confirm prior results obtained in smaller samples. (12) Demontis et al., found that almost all variants influencing ADHD also influenced smoking, 79% of which had concordant directions. (15) Given that smoking is a known risk factor for severe COVID-19, (57) it is plausible that the causal association between ADHD and increased severity of COVID-19 is driven by smoking habits rather than by shared genetic causes.
Finally, we should highlight that all significant genes shared between more than one psychiatric disorder and COVID-19 were located on chromosome 17, specifically in 17q12-q21, a region that has been linked to immune response. (58, 59) In addition, the potential causal genes identified in this study converge on the endocrine axis HPA and hypothalamic-pituitary-thyroid (HPT), two connected biological pathways that play an important role in stress and immune response. (60) Therefore, while different genes have been identified across different psychiatric disorders in relation to COVID-19 susceptibility and severity, our results suggest that immune responses may be behind the increased vulnerability of people with psychiatric disorders to COVID-19. These results, if confirmed, might open the way for new targets for suitable therapeutic approaches.
The main strength of our study is the use of datasets from the largest available GWAS of various psychiatric disorders and COVID-19 traits. Moreover, to the best of our knowledge, it is the first study to explore shared causal risk loci between six different psychiatric disorders and COVID-19 traits using consistent methodology. However, this study should be considered in the light of several limitations. First, all GWAS results used in this study were obtained from cohorts of European ancestry, so our results might not represent other ancestry groups. Second, the power of the original GWAS is key for most of the analysis performed in our study. Thus, a non-significant result does not necessary reflect a true lack of association. Third, MR needs a strict list of assumptions to be met, which is challenging when analysing polygenic traits. Even though we have endeavoured to exclude pleiotropic SNPs, we cannot certify that this assumption has not been violated. Fourth, our study does not account for complex gene-gene interactions that may play a significant role in the observed relationships between psychiatric disorders and COVID-19. Finally, we are missing the role of potential cofounding variables such as BMI or vaccination on COVID-19 susceptibility and severity, which could not be taken into account due to the study design.
In conclusion, our results support that the relationship between psychiatric disorders and COVID-19 risk is likely due to shared alterations in immune-related pathways such as thyroid and inflammatory dysfunction, and impaired stress response, and is not as a result of environmental factors alone. Exploring new targets and drug repositioning strategies for medications traditionally employed in the treatment of stress-related and thyroid disorders holds promising potential to enhance COVID-19 and similar viral infection outcomes among individuals with psychiatric disorders.