A new study raises a red flag against adding antibiotics to prevent contamination when growing cells in the lab, finding that it can induce unintentional genetic changes in the cells and distort test results.

The standard practice and resulting genetic changes may be especially concerning in pharmacogenomics experiments looking at how human cells respond to drugs, according to researchers at the University of California, San Francisco (UCSF).

While studying how genetic variations affect human response to drugs, Nadav Ahituv, professor of bioengineering in the School of Pharmacy, found that rifampin, a common antibiotic prescribed for tuberculosis and other infections, can cause significant changes in the expression of genes and regulatory elements in human liver cells.

Ahituv and graduate student Ann Hane Ryu then designed a simple experiment in which they compared a human liver cell line, HepG2, grown with and without the standard antibiotic cocktail, PenStrep, a combination of penicillin and streptomycin. They chose liver cells because the liver helps clear foreign materials from the body and would be the most responsive to drugs.

They let the cells grow for three weeks and analyzed them to identify changes in both coding and non-coding DNA.

As detailed in a paper published this month in Scientific Reports, they found altered expression in 209 genes, particularly in those related to drug and stress response. These included higher expression of genes known to be involved in apoptosis and the unfolded protein response.

Of particular concern is that several of the affected genes are transcription factors, which can in turn control the expression of many other genes, Ahituv was quoted as saying in a news release from UCSF this week. In addition, the researchers found changes in more than 9,500 gene regulatory elements, part of the 98 percent of human genome that is non-coding. These regulatory elements were located near genes involved in cell differentiation, nuclease activity and transfer RNA, or tRNA, modification.

Ahituv, senior author of the paper, hopes the study will serve as a warning for other researchers to take into account the effect of antibiotics, and perhaps even examine other variables in the laboratory that may unintentionally influence an experiment. And he would recommend that researchers studying drug response avoid the use of antibiotics, and instead take other precautions against contamination.