Plaque levels differ in popular Alzheimer’s mouse model depending on which parent’s variants are passed down

Transgenic mice typically help to standardize disease research, but not in the case of a common model of familial Alzheimer’s disease: The amount of amyloid beta plaque in 5XFAD mice—so named for the five familial Alzheimer’s disease variants they carry—varies depending on how they are bred, according to a study published 20 January in Neuron.

Less than 20 percent of more than 900 studies published between 2019 and 2024 that used the 5XFAD model reported breeding schemes, the work also shows, which is “such an experimental confound,” says study investigator Constanze Depp, a postdoctoral researcher in Beth Stevens’ lab at Harvard University.

The original model mice, first described nearly two decades ago, carry three disease-linked variants in the APP gene and two in the PSEN1 gene, all derived from people with early-onset, inherited Alzheimer’s disease. The animals develop plaques by 2 months of age.

But it turns out that when the mice inherit the transgenes from their father, they have twice as many plaques as when they carry maternally inherited transgenes, the new work shows. The results were independent of other variables such as age, sex and colony.

The findings may help to explain the replication crisis in Alzheimer’s disease research, says Kim Green, professor of neurobiology and behavior at the University of California, Irvine, who was not involved in this study. “Making sure that you generate your animal cohorts in a controlled and reproducible fashion is important, but it is incredibly difficult.”

The results also underscore the importance of considering confounders in animal model experiments, says Bettina Platt, chair in translational neuroscience at the University of Aberdeen, who was also not involved in the study. Female 5XFAD mice have higher levels of amyloid beta than males, according to a 2021 study she led. It has also been known, Platt says, “that in most transgenic animals there is quite a lot of genetic drift anyway, so even the same transgenic line bred in different laboratories may end up having huge differences in pathology.”

A

bout six years ago, Depp and her colleagues started to notice some unusual results while studying myelin integrity in 5XFAD mice. It was clear there was a lot of intrinsic variability in the model, says study investigator Andrew Octavian Sasmita, a postdoctoral researcher at University College Cork, which led the team to reevaluate a number of different characteristics of the model.

Sasmita, Depp and their colleagues used heterozygous 5XFAD mice that they bred by crossing one 5XFAD heterozygote parent with a wildtype C57BL/6 mouse. “We went back into the pedigree of how we bred the animals and figured out that one of the major differences was the breeding scheme that we had applied,” Depp says.

The offspring showed stark differences in amyloid plaque levels, the team discovered using quantitative light-sheet microscopy. And this pattern repeated itself in multiple team members’ mouse colonies and in 5XFAD mice obtained directly from the Jackson Laboratory and bred, which meant “we are really 100 percent sure that this is true,” Depp says.

As Sasmita and Depp told other researchers about their findings at meetings, some were surprised. Others, Depp says, were “maybe a bit suspicious and then just opted for one breeding scheme. But the data were never out there.” Indeed, Green says, his own lab did not report its breeding scheme in past publications because he never thought there was a need to, although his team always uses paternal transgene inheritance with 5XFAD mice.

T

he mechanism behind the biased plaque patterns could involve genomic imprinting, the process by which sperm and egg cells add epigenetic tags to certain genes to ensure that offspring express only one parent’s copy of the imprinted genes, Sasmita initially suspected. The “parental transgene difference effect was a hot thing back in the ’80s and the ’90s,” that people may have forgotten about, he says.

Neither the APP nor the PSEN1 gene is known to be imprinted, according to a database of imprinted genes. But the 5XFAD mouse model’s Thy1.2 promoter—a short strip of DNA that controls expression of the altered transgenes—could be, the team thought. And if so, it would influence expression levels of both transgenes.

In support of this hunch, mice with maternal inheritance of the transgenes showed lower APP protein levels than their paternally programmed counterparts, the researchers found. (They did not examine PSEN1 levels, Sasmita says.)

The same effect was absent in the APP^NLGF mouse model of Alzheimer’s, which does not use transgenes, Depp says. Still, other transgenic mouse models that use a Thy1-promoter could have similar imprinting-related issues that scientists need to be aware of, she adds. In line with that idea, a non-Alzheimer’s transgenic mouse model called ATeam that also uses a Thy1-promoter has a correlation between inheritance and protein expression similar to that in 5XFAD mice, the study shows.

It might be time to move away from transgenic models, partially due to these confounds, says Green, who is part of MODEL-AD, a consortium seeking out next-generation models of Alzheimer’s disease. Transgenics, he says “don’t recapitulate the phenotypes that we see in the human Alzheimer’s brains,” so there is a need for more physiologically relevant models. Although, he adds, the 5XFAD model still has its place in certain types of research, because plaques form after a few months—compared to more than a year in other mouse models.

For now, Depp says, she and her colleagues want others to be aware of this effect and try to control for it: “I think in the end, the idea would be always to breed paternally and keep this consistent among many, many, many generations.”