Gene expression changes in blood RNA after swimming in a chlorinated pool
Lucas A. Salas , Laia Font-Ribera , Mariona Bustamante , Lauro Sumoy , Joan O. Grimalt , Sarah Bonnin , Maria Aguilar , Heidi Mattlin , Manuela Hummel , Anna Ferrer , Manolis Kogevinas , Cristina M. Villanueva
DOI:10.1016/j.jes.2017.05.011
Received November 28, 2016,Revised May 05, 2017, Accepted May 05, 2017, Available online May 16, 2017
Volume 29,2017,Pages 250-261
Exposure to disinfection by-products (DBP) such as trihalomethanes (THM) in swimming pools has been linked to adverse health effects in humans, but their biological mechanisms are unclear. We evaluated short-term changes in blood gene expression of adult recreational swimmers after swimming in a chlorinated pool. Volunteers swam 40 min in an indoor chlorinated pool. Blood samples were drawn and four THM (chloroform, bromodichloromethane, dibromochloromethane and bromoform) were measured in exhaled breath before and after swimming. Intensity of physical activity was measured as metabolic equivalents (METs). Gene expression in whole blood mRNA was evaluated using IlluminaHumanHT-12v3 Expression-BeadChip. Linear mixed models were used to evaluate the relationship between gene expression changes and THM exposure. Thirty-seven before-after pairs were analyzed. The median increase from baseline to after swimming were: 0.7 to 2.3 for MET, and 1.4 to 7.1 μg/m3 for exhaled total THM (sum of the four THM). Exhaled THM increased on average 0.94 μg/m3 per 1 MET. While 1643 probes were differentially expressed post-exposure. Of them, 189 were also associated with exhaled levels of individual/total THM or MET after False Discovery Rate. The observed associations with the exhaled THM were low to moderate (Log-fold change range: ?0.17 to 0.15). In conclusion, we identified short-term gene expression changes associated with swimming in a pool that were minor in magnitude and their biological meaning was unspecific. The high collinearity between exhaled THM levels and intensity of physical activity precluded mutually adjusted models with both covariates. These exploratory results should be validated in future studies.
