The behavioral effects of genetic mutations depend strongly on the mouse strain in which they are studied, and in some cases, the same genetic mutation can have opposite effects on behavior, such as increasing anxiety-driven behaviors in some mouse strains and decreasing them in others, researchers have found.

"There really is no such thing as saying this gene has this effect on this phenotype," Abraham Palmer told BioWorld Today. "When you express a mutation on a particular genetic background, the effect of the mutation is going to be filtered through all of the other unique aspects of that genotype."

Palmer is a professor of psychiatry at the University of California at San Diego and the senior author of the paper reporting the findings, which appeared in the Sept. 8, 2016, online issue of Neuron.

While there are several hundred mouse strains available commercially, the research reported in any individual paper typically is more often than not conducted with just one strain. And biomedical research overall skews heavily toward just a few strains.

In their experiments, the authors crossed mice of one of those workhorse strains, the C57BL/6J mouse, with 30 different mouse strains to compare the effects of two genes, CACNA1C and TCF7L2, that have been strongly flagged as disease risk genes by human genomewide association studies (GWAS).

"In humans, genetic variants [of CACNA1C and TCF7L2] are associated with several psychiatric disorders," though the TCF7L2 gene is primarily associated with type 2 diabetes risk, lead author Laura Sittig, who is a graduate student at UCSD, told BioWorld Today.

In earlier single-strain studies, mice with only one copy of TCF7L2 had shown changes in fear learning and anxiety levels, while having only one copy of CACNA1C appeared to have antidepressant effects in female mice and anti-anxiety effects in both sexes, among other effects.

The team then gave a battery of behavioral tests to the mice, and looked at the effects of having only one copy of either gene across all 30 strains. They also looked at sex differences in those effects.

Sex and gender has been receiving a lot of attention as a source of variability in recent years, with the NIH now requiring researchers to either include female animals and female-derived cell lines in their preclinical research, or to justify their exclusion.

Overall, the researchers found what they termed "extremely strong interactions with genetic background" in their experiments. In many cases, those interactions meant that lacking one copy of a gene had much stronger effects in some strains than others. But in four instances, they showed that lack of the same gene could have the opposite effect in different strains.

For example, some strains with only one copy of TCF7L2 were less prone to startle in response to a loud noise, while others were more so.

Other studies have shown that different strains have microbiome differences, and those differences can influence, for example, the susceptibility to infections.

While the team did not explicitly look at microbiome differences, Sittig noted that all mouse strains they used came from the same vendor, and were housed together, making significant differences between their microbiomes unlikely.

The team's findings also showed that strain effects eclipsed those of sex.

"We did some quantitative analysis of the effects of sex vs. the effects of genetic background," Palmer said, and the effects of genetic background were "much greater."

"The NIH has really been pushing investigators not to brush sex effects under the rug," he added. His team's research shows that "it would be equally important to spend the extra resources to examine the effects of several different genetic backgrounds."

Palmer did not want to give a specific number of strains that would be necessary to be confident in the results of a study, saying that "depends on so many factors."

"Certainly we are not suggesting that people need to do 30 strains," he said.

The authors consider their role to be neither that of a Pollyanna nor a Cassandra.

Palmer stressed that the findings "could be misread by a layperson, or a person with their own agenda, that doing research on mice is not productive. That's absolutely not our position."

In fact, Sittig added, the results are "also an opportunity" to learn about gene interactions, which, as the paper shows, are the final arbiters of a gene's effect.

Still, Palmer said, "practically, the main idea is that a single strain is probably not enough" to gain robust insights into the function of a given gene. "If you do a single strain," he said, "you just don't know where you are on that spectrum."