NEW YORK – Polygenic risk scores based on data from multiple ancestral populations improve predictions of inflammatory bowel disease for all populations, a new analysis has found.
Inflammatory bowel disease — which includes Crohn's disease and ulcerative colitis — affects about 3 million people in the US, according to the Centers for Disease Control and Prevention. While genome-wide association studies have identified more than 200 loci linked to IBD, many studies have focused on individuals of European ancestry, and have suggested that polygenic risk scores developed in one ancestral population may not perform as well in other populations.
Using data collected by Icahn School of Medicine at Mount Sinai's BioMe Biobank, researchers compared polygenic risk scores for IBD developed using European, African-American, and Ashkenazi Jewish case-control studies as well as a PRS developed using a combination of those studies. As they reported on Thursday in the journal Gastroenterology, they found that the risk scores from the combined data improved predictions across populations.
"The ability to accurately predict genetic disease risk in individuals across ancestries is a critical avenue that may positively affect patient outcomes, as early interventions and even preventive measures are being considered and developed," senior author Judy Cho from the Mount Sinai's Charles Bronfman Institute for Personalized Medicine said in a statement. "These findings support a need for greater genetic diversity, including more data on African American populations, to enhance disease risk predictions and reduce health disparities for all populations."
For each ancestry group, the researchers generated and applied a PRS using BioMe Biobank data. That dataset included 32,595 individuals thought to be representative of the New York City area population and, for this study, included 339 IBD cases and 19,202 controls. The prevalence of IBD in this cohort was in line with expected disease prevalence estimates.
The researchers then evaluated how well the PRS could predict disease status. Overall, predictions were best among the non-Jewish European ancestry population, while predictions were worse in the African-American cohort.
When they evaluated risk scores generated from a meta-analysis of the three ancestry cohorts, the predictive power of the score went up. The improvement was more significant among the non-Jewish European ancestry cohort and the Ashkenazi Jewish cohort than the African-American cohort. This decreased predictive power among African-Americans could stem from both a smaller reference dataset and increased genetic diversity within populations of African descent, the researchers noted. They added that the findings highlight the need for additional studies of African-American and non-European populations.
They additionally examined the contribution of rare variants typically associated with very-early-onset IBD to disease within the three ancestry cohorts. For individuals of European ancestry, they confirmed a role of an IL10RA variant in IBD and they noted association trends for two variants in LRBA with IBD among African-American patients.
LRBA variants lead not only to decreased LRBA protein expression, but also to cytotoxic T lymphocyte associated protein 4 (CTLA-4) protein levels. LRBA is a transmembrane protein with a role in vesicle trafficking, and LRBA alterations have been found to affect recycling CTLA-4 to the cell surface and influence autoimmune inflammation.
As the malaria drug chloroquine has been shown to increase CTLA-4 expression, the researchers suggested that future studies could help predict which IBD patients might benefit from that or similar drugs. "Since lowered LRBA and CTLA-4 expression can lead to IBD, it's encouraging that chloroquine is able to partially recover expression," first author Kyle Gettler, a postdoc at Mount Sinai, said in a statement.