Abcg2-Expressing cells fuse with existing cardiomyocytes

Abstract

Cardiac side population cells (cSPCs) were the first group of progenitor cells identified in the heart; however, their progenitor cell properties have only been established in cell culture and after transplantation. To determine whether cSPCs possess progenitor cell properties in vivo, we generated an Abcg2-driven, lineage-tracing mouse model. In this model, 75.8 ± 10.87% of cSPCs are labeled with GFP. Over a four-week chase period, there is a five-fold increase in cardiomyocyte-labeling from 0.17 ± 0.15% to 0.84 ± 0.24%. Surprisingly, 90.7 ± 2.66% of labeled cardiomyocytes arise from fusion events and not direct differentiation from cSPCs based on labeling in Abgc2MCM/+ R26mTmG/+ mice. Given the extensive labeling of bone marrow and endothelial cells in our model, we hypothesized that the increase in cardiomyocyte labeling over the four-week chase period arises from fusion of unlabeled cardiomyocytes with GFP-labeled bone marrow cells or endothelial cells. To test this hypothesis, we irradiated Myh7Cre/+ R26tdTomato/+ mice, transplanted them with hematopoietic stem (LSK) cells isolated from Abgc2MCM/+ R26GFP/+ and pulsed them with tamoxifen. At the end of a four-week chase period, a similar percentage of bone marrow side population cells, LSK cells and differentiated lineages were labeled with GFP. More importantly, only 0.02 ± 0.03% of cardiomyocytes were labeled with both GFP and tdTomato. No cardiomyocytes were labeled with only GFP. Next, we evaluated whether fusion with endothelial cells could account for the increase in cardiomyocyte labeling we observed. Using an endothelial-specific, lineage-tracing mouse model, Cdh5CreER/+ R26mTmG/+, we found no cardiomyocytes labeled at the end of a four-week chase period. Our results demonstrate that the primary mechanism of cardiomyocyte labeling in our Abcg2-driven, lineage-tracing mouse model arises from fusion of cardiomyocytes with lineage-traced cells, but not with bone marrow or endothelial cells. These results indicate that fusion in the heart may occur at much higher rates than previously assumed, and may play important roles in homeostatic maintenance of cardiac function.