A Better Understanding of the Role of TGF-β Family Members in Tissue Fibrosis

Traditionally two different subfamilies of the transforming growth factor-β (TGF-β) superfamily have been associated with different effects on tissue fibrosis: the profibrotic TGF-β subfamily (TGF-β1, TGF-β2 and TGF-β3) and the antifibrotic bone morphogenetic proteins (BMPs) subfamily. TGF-β1 is the most widely studied profibrotic cytokine, as it regulates numerous processes involved in tissue fibrosis: synthesis of extracellular matrix proteins (ECM), apoptosis of the parenchymal cells, regulation of epithelial-tomesenchymal transition (EMT) and endothelial-to-mesenchymal transition [1,2]. On the other hand, BMPs (especially BMP7) have been considered as antifibrotic proteins [2-4]. TGF-β subfamily members activate ALK5 type I receptors –thus activating Smad2 and Smad3 proteinswhereas BMPs activate ALK1, ALK2, ALK3 and ALK6 receptors and thus activating Smad1, Smad5 and Smad8 proteins.

BMP9 and BMP10 were described as high affinity ligands for ALK1 receptors in endothelial cells [5,6]. These proteins show redundancy in several biological processes such as vascular development, and share numerous similarities in their functions, although BMP10 is necessary for embryonic development and BMP9 is not [7]. Our group has demonstrated that the ALK1 receptor has an antifibrotic role in an experimental model of obstructive nephropathy [8], as ALK1 (previously considered as an endothelial specific receptors) is expressed in fibroblasts and has an important regulatory role in ECM proteins synthesis [8]. In order to explore the antifibrotic role of the ALK1 receptor, we studied the effects of its canonical and high affinity ligand, the BMP9 protein, on renal fibrosis and ECM proteins regulation. We have recently observed that BMP9 stimulates cellular responses in fibroblasts through ALK1 and ALK5 receptors. Although previous studies demonstrated that BMP9 stimulates cellular responses in other non-endothelial cell types such as liver tumour cells, those studies did not show whether the effects are due to ALK1 activation [9].
The findings that we show in our recent studies address a profibrotic role of BMP9 in vitro, but further analysis is needed to unveil the potential interest of this molecule in fibrotic diseases. It will be necessary to evaluate the levels of BMP9 in fibrotic tissues as well as in plasma, either in patients or in in vivo experimental models of fibrosis.
It is well known that BMP9 is produced by hepatocytes in the liver and circulates in an active form at concentrations ranging between 2 and 12 ng/ml [16].
We demonstrate that BMP9 behaves as a profibrotic factor in vitro, promoting the synthesis of ECM proteins such as collagen I, fibronectin and connective tissue growth factor (CTGF) [10]. The identification of BMP9 as a profibrotic factor complicates the scheme of the BMP/TGF-β network. BMP9 does not induce the same effects than BMP7 -the most powerful antifibrotic cytokine, as their mechanisms of action are different. While BMP7 activates Smad1/5/8 through ALK3 we observed that BMP9 activates Smad1/5/8 through the ALK1 receptor and Smad2/3 proteins through the ALK5 receptor, being ALK1 essential in both processes. In this experimental scenario, both Smad1/5/8 and Smad2/3 are necessary for the induction of ECM proteins synthesis (Figure 1). In several cellular types, BMP9 can induce non-Smad signalling (MAPK, PI3K/AKT, NF-κB, Wnt etc.) [11]. In our experimental model with mouse embryonic fibroblasts, BMP9 induces Erk1/2 phosphorylation, which is necessary for BMP9induced fibrogenic response [10]. However, further studies should be performed to determine if Erk1/2 activation induced by BMP9 is an indirect effect of Smad phosphorylation. This is not the first time that BMP9 is related to ECM protein synthesis, as it was previously shown that this cytokine stimulates the synthesis of the ECM components Col2A1 and aggrecan in human multipotent mesenchymal cells [12]. Moreover, BMP9 regulates Col9a1, Col9a3 and Col1a synthesis in forebrain cholinergic neurons [13]. In endothelial cells, BMP9 regulates CTGF expression through a Yes-associated protein (YAP)dependent mechanism [14]. BMP9 and BMP10 stimulate the expression of fibronectin and collagen I in cultured endothelial cells [15]. Recently, Bailly et al. have demonstrated that BMP9 and BMP10 are both necessary for the proper closure of the ductus arteriosus in newborns; one of the mechanisms involved in this process is BMP9promoted ECM deposition [15].
Zhu et al. have discovered 4-fold increased levels of BMP9 in the serum of dialysis-patients with chronic kidney disease with respect to pre-dialysis patients [17]. It will be interesting to elucidate whether there is any correlation between BMP9 levels and kidney fibrosis. Previous work from our laboratory showed that the predominant form of endoglin (L-endoglin), which is now considered as a receptor required for BMP9-induced signalling [7,18] promotes renal tubuleinterstitial fibrosis after the unilateral ureteral obstruction experimental model through Smad1/5 and Smad2/3 activation. The possible involvement of BMP9 in these mechanisms should be addressed.
Moreover, the study of the role of BMP9 in liver fibrosis is of undoubted relevance for different reasons: First, BMP9 shows its major expression in the liver in physiological conditions. Second, during the development of fibrosis, ALK1 activates Id1 through the Smad1 pathway promoting the transdifferentiation of hepatic stellate cells into myofibroblasts [19]. Moreover, BMP9 induces EMT in hepatocellular carcinoma cells [20]. Finally, another important target of BMP9 in liver cells is hepdicin, a polypeptide involved in the inhibition of iron absorption and recycling, being these processes associated with liver fibrosis and cirrhosis [21].
Further studies will be needed for the complete characterization of BMP9 as a regulator of tissue fibrosis and to its validation as a good therapeutic target. The deeper knowledge of the complex and large TGF-β superfamily is widening as new members and their roles are being identified. New studies will be necessary to integrate the interaction of all these molecules that share receptor complexes and intracellular Smads.