Thesis presented November 09, 2022

Abstract:
BMP9 and BMP10 are two specific ligands of the receptor ALK1, a type I BMP receptor mainly expressed on vascular endothelial cells. Mutations in the BMP9/BMP10/ALK1/ENG signalling pathway are linked to vascular diseases including hereditary hemorrhagic telangiectasia (HHT) and pulmonary arterial hypertension (PAH). BMP9 is expressed by the liver, whereas the primary source of BMP10 is the heart. However, BMP10 is also found to some extent in the liver, where it could form a BMP9/10 heterodimer, that has been described as circulating form responsible for ALK1 activity on endothelial cells. The aim of my thesis project was to unveil BMP9 and BMP10 functions in cardiovascular homeostasis in both postnatal and adult mouse models deficient for BMP9 and/or BMP10 and to discriminate between cardiac and hepatic BMP10 roles. While BMP9 was constitutively deleted, BMP10 deletion had to be achieved using the ubiquitous Rosa26 CreER^T2 tamoxifen-inducible promoter, to bypass embryonic lethality.

Our results show that BMP10 iKO adults did not develop any obvious age-dependent cardiovascular defects, however, concomitant deletion of the two ligands (double-KO Bmp9^-/- R26CreER^T2 Bmp10^lox/lox) results in strong vascular defects, showing that both ligands participate to vascular homeostasis. In younger animals, we initially identified that tamoxifen-injected Rosa26-CreER^T2 BMP10 iKO mice showed strong hematopoiesis defects. Unfortunately, we demonstrated that the observed phenotype was an artefact induced by CreER^T2 toxicity. Indeed, upon tamoxifen administration to Rosa-26 CreER^T2 pups lacking a targeted floxed allele, we observed a similar phenotype to that of the Rosa Cre BMP10 iKO. In the bone marrow of these animals, highly proliferative haematopoietic cells were strongly depleted, and the sinusoidal vessels were completely disorganised. Our work shed light on the importance of using the right controls when working with inducible Cre dependent mice models. Conversely, in the context of newborn lymphangiogenesis, we confirmed a functional role of BMP10 in the lymphatic vasculature which is not due to the Rosa 26-CreER^T2.

In addition, we proved the importance of BMP9 and BMP10 in vascular development and angiogenesis using the postnatal retina angiogenesis model. In particular, combined loss of BMP9 and BMP10 resulted in a significant decrease of the vascular radial extension, veins enlargement, arterial loss of αSMA coverage and a strong increase in capillaries density. In order to discriminate between cardiac and liver BMP10 function, we generated BMP10 KO models using either the cardiac-specific α-myosin heavy chain (αMHC)-MerCreMer inducible promoter or lecithin-retinol acyltransferase Lrat-Cre constitutive promoter, expressed in hepatic stellate cells. We could confirm tissue-specific BMP10 invalidation in both mice models. Moreover, analysis of the plasma obtained from these animals allowed us to confirm the hepatic origin of the circulating active heterodimer composed by BMP9 and BMP10, yet their functional role remains to be determined.

Keywords:
BMP9 and BMP10, vascular functions, knock-out mouse models, angiogenesis