Animal-based research has led to nearly every major medical advancement in recent history. And mice, because they are very similar to humans in terms of genetics and biology, have played a vital role in studies on a range of diseases, from cancer to birth defects.
For example, mouse studies proved that alcohol can damage the developing brain, particularly in the third to fourth week when many pregnancies are not yet recognized. Findings like this led the U.S. Surgeon General to advise women who are pregnant, or considering becoming pregnant, to abstain from using alcohol and helped change the way doctors counsel patients on alcohol use. Mice are necessary for such research because some questions about the complex biomedical processes involved can only be answered by studying how they play out within a living organism.
And although it can yield tremendous breakthroughs, research of this nature can be quite challenging. Studying how genetics and the environment can influence development at different embryonic stages requires precisely timed pregnancies. But identifying whether a mouse is pregnant, especially during the earliest stages of gestation, is trickier than one might think, and standard methods are imprecise.
To address this issue, a research team from the School of Veterinary Medicine (SVM) and the Research Animal Resource Center (RARC) at the University of Wisconsin-Madison has developed a reliable, non-intrusive method for early pregnancy detection that also significantly reduces the number of mice required to conduct such research. They describe the method in a study published online on June 19 in the Journal of the American Association for Laboratory Animal Science.
“This study demonstrates that weight gain can be used reliably to distinguish pregnant from non-pregnant female mice earlier than any other available method,” says Rob Lipinski, assistant professor of comparative biosciences at the SVM and a co-author of the study. “We were motivated to do this study because our work focuses on very early development, when embryos are most sensitive to disruption, and when pregnancy is particularly difficult to detect accurately.”
How they arrived at this new method requires some explanation of how scientists conduct studies of biological development that involve mice. To examine the full spectrum of embryonic stages, it’s important to detect the exact onset of pregnancy. For the two breeds of mice involved in this study, the typical gestation period is about 20 days, and pregnancy can be detected on day zero by the presence of a copulation plug. But this method often yields a significant false positive rate, sometimes as high as 50 percent.
Pregnancy can also be confirmed through a visual inspection at 15 days or palpation at 10 days, but the latter can cause stress on the animals that confounds study results and often causes mice to be unfit for subsequent research. Regardless, many types of biological development studies, such as fetal alcohol research, require earlier pregnancy detection than 10 days to be conducted properly.
While completing postdoctoral work on craniofacial birth defects at the University of North Carolina at Chapel Hill, Lipinski became frustrated with the inefficiency of pregnancy detection and the animal use burden of his studies. He recalled bits of data buried deep in previous research studies that suggested pregnant mice gain more weight than non-pregnant mice. This led him to hypothesize that weight gain might be used to discriminate between the two.
To put the concept to the test, Lipinski collaborated with three staff members at RARC: Eric Sandgren, RARC director;
Jody Peter, manager of the RARC rodent facility; and Erin Plisch, a senior research specialist. The team divided the mice into two breeding populations. In one group, they weighed pregnant and non-pregnant mice from day 4.5 of gestation to day 11.5 and then conducted a linear regression analysis of the data, which verified that pregnant mice do indeed gain weight more quickly than non-pregnant mice.
In a second group, they verified pregnancy through the most common method, copulation plug detection, which yielded a 37 percent false positive rate. They also measured and recorded the weight gain of pregnant mice and used the data to establish an optimal weight gain discrimination threshold of 1.75 grams. When they applied this threshold as an optional identifier for pregnant mice, it produced a false positive rate of approximately 10 percent without excluding any pregnant mice.
“Basically, we reduced the number of animals required for our typical experiments by almost 25 percent,” says Lipinski. “This is not only better from an animal welfare standpoint, which is a high priority, but also for a laboratory’s bottom line. For my lab, the savings from associated costs of animals, per diems, reagents, and salary comes to almost $20,000 per year.”
“Our responsibility as researchers is to improve the efficiency of our work,” says Sandgren. “And our responsibility as members of an animal program is to improve the well-being of the animals under our care, including reducing the number that are involved. The results of this study support both of these objectives.”
The research team also included Galen Heyne, an associate research specialist in Lipinski’s lab, and undergraduate Cal Melberg, both of whom will begin the Doctor of Veterinary Medicine program in fall 2015. For Heyne, a future veterinarian who is passionate about animal welfare, the prospect of reducing reliance on animals in research was extremely appealing.
“There are a lot of labs around the world that do studies similar to ours,” says Heyne, referring to his work with Lipinski on cleft lip and palate, one of the most common birth defects in people. “So the potential to reduce the number of animals involved in this research is large if we can spread the word about this method.”
— Nick Hawkins