Effects of Actr2 gene Variants on Mouse Viability, Gene Expression, and Cytoskeletal Dynamics

Venous thromboembolism (VTE) is a prevalent human disease that exhibits significant heritability. Factor V Leiden (F5^L) is the most common known VTE risk factor, but modifier genes significantly influence VTE development. We recently identified a p.R258G missense mutation in the Actr2 gene (Actr2^+/G, ARP2 protein) as a genetic suppressor of lethal thrombosis in mice homozygous for F5^L (F5^L/L) and hemizygous for tissue factor pathway inhibitor (Tfpi^+/-). However, the antithrombotic mechanism of Actr2^+/G is unknown. We used genetic and genomic techniques to investigate the phenotypic effects of the Actr2^G variant on cells and mice.

We previously analyzed progeny from Actr2^+/G x Actr2^+/G (N7 backcross generations to C57BL6/J) to investigate the effects of Actr2^G on mouse survival. The progeny deviated from Mendelian frequencies, as only ~12% were Actr2^G/G mice (N=303; p<9x10^-7) and these mice have a significantly reduced lifespan of ~six months (N=131; p<0.0001). The Actr2^+/G mice, when compared to wildtype mice, were morphologically indistinguishable and were born at the anticipated frequency. In the context of F5^L, a F5^L/LActr2^+/G x F5^L/LActr2^+/G cross yielded only 3 F5^L/LActr2^G/G, which also significantly departed from expected frequencies (N=66; p<0.01). The F5^L/LActr2^+/G mice were born at the anticipated frequency. To determine if the mouse embryos were produced at normal frequencies, we examined pups at embryonic day 14.5 of this F5^L/LActr2^+/G x F5^L/LActr2^+/G cross. We observed normal frequencies at embryonic day 14.5, with 5 F5^L/LActr2^G/G embryos out of 18 total. Taken together, our results suggest that Actr2^G exhibits gene dosage lethality, with Actr2^G/G homozygosity detrimental for late stage embryonic development/early adulthood.

We next studied the mechanism(s) of Actr2^G thrombosis suppression by gene expression studies. RT-qPCR for multiple coagulation genes was performed in liver samples from Actr2^+/+ and Actr2^+/G mice. Serpine2 mRNA (Protease Nexin-1 (PN-1)) was significantly increased in Actr2^+/G mice (N=3; p<0.0001). We also observed an increase in plasma PN-1 by quantitative Western Blot analysis. To investigate the Actr2^G regulation of PN-1 expression, mouse N2a cells containing Actr2^+/G or V5-tagged PN-1 were generated using CRISPR/Cas9. Genomic analysis revealed homology directed repair (HDR)-mediated knock-in frequencies of 24.3% (9/37) for Actr2^+/G and 5.6% (1/18) for PN-1-V5. We then attempted to generate PN-1-V5 cells carrying mutant and/or Myc-tagged ARP2 (GMyc2). For the GMyc2 lines, the HDR efficiency for double knock-in was 15.6% (7/45). Analysis of all of our cell lines revealed however, that indels were also present within both the endogenous Actr2 and Serpine2 loci. Subsequent analysis of our GMyc2 cells revealed 5 Actr2 alleles. This suggests that cells avoid the loss of Actr2 function and highlights the difficulty with creating an Actr2 mutant cell model.

To further investigate the effects of the Actr2^G mutation, we isolated mouse embryonic fibroblasts (MEFs) from Actr2^+/+ and Actr2^G/G mice. The Actr2^G/G MEFs grew poorly in culture, and displayed a significant reduction in cell surface area 20 minutes after plating on fibronectin when compared to Actr2^+/+ (q=0.0003). Further analysis revealed F-actin aggregation in the root of cellular protrusions of Actr2^G/G MEFs. These results indicate cytoskeletal remodeling defects in Actr2^G/G cells. In summary, our studies indicate that mutations affecting ARP2 function can have profound effects on mouse and cell survival, as well as pronounced effects on specific physiological processes like blood coagulation and the transcriptional regulation of coagulation related genes.