Abstract
In humans, somatic activating mutations in PIK3CA are associated with skeletal overgrowth. In order to determine if activated PI3K signaling in bone cells causes overgrowth, we used Tg(BGLAP-Cre) and Tg(DMP1-Cre) mouse strains to somatically activate a disease-causing conditional Pik3ca allele (Pik3caH1047R) in osteoblasts and osteocytes. We observed Tg(BGLAP-Cre);Pik3caH1047R/+ offspring were born at the expected Mendelian frequency. However, these mice developed cutaneous lymphatic malformations and died before 7 weeks of age. In contrast, Tg(DMP1-Cre);Pik3caH1047R/+ offspring survived and had no cutaneous lymphatic malformations. Assuming that Cre-activity outside of the skeletal system accounted for the difference in phenotype between Tg(BGLAP-Cre);Pik3caH1047R/+ and Tg(DMP1-Cre);Pik3caH1047R/+ mice, we developed sensitive and specific droplet digital PCR (ddPCR) assays to search for and quantify rates of Tg(BGLAP-Cre)- and Tg(DMP1-Cre)-mediated recombination in non-skeletal tissues. We observed Tg(BGLAP-Cre)-mediated recombination in several tissues including skin, muscle, artery, and brain; two CNS locations, hippocampus and cerebellum, exhibited Cre-mediated recombination in >5% of cells. Tg(DMP1-Cre)-mediated recombination was also observed in muscle, artery, and brain. Although we cannot preclude that differences in phenotype between mice with Tg(BGLAP-Cre)- and Tg(DMP1-Cre)-mediated PIK3CA activation are due to Cre-recombination being induced at different stages of osteoblast differentiation, differences in recombination at non-skeletal sites are the more likely explanation. Since unanticipated sites of recombination can affect the interpretation of data from experiments involving conditional alleles, we recommend ddPCR as a good first step for assessing efficiency, leakiness, and off-targeting in experiments that employ Cre-mediated or Flp-mediated recombination.