Statistics from Altmetric.com
We read with interest the article by Breitkopf-Heinlein et al,1 investigating the role of bone morphogenetic protein (BMP)-9 in liver regeneration and fibrosis. In mice, the authors demonstrate that BMP-9, a member of the transforming growth factor-β family, is primarily produced by hepatic stellate cells (HSCs) and is a profibrogenic factor. Inhibition or knockout of BMP-9 reduced liver fibrosis, while the recombinant protein increased the proliferation and migration of human HSCs. However, as shown in figure 1H and I of that study, rhBMP-9 did not upregulate fibrosis markers (eg, collagen I, fibronectin, platelet-derived growth factor B and α smooth muscle actin). While BMP-9 expression increased during the activation of primary human HSCs, the authors were unable to demonstrate a correlation between human hepatic BMP-9 expression and liver fibrosis stage from a publicly available microarray dataset.
We analysed multiple publicly available microarray datasets; consistent with their findings, differential expression of BMP-9 mRNA was not observed between healthy and cirrhotic liver (GSE6764, GSE89377 and GSE36411, data not shown), while another dataset (GSE77627, figure 1A) revealed downregulation rather than the upregulation of BMP-9 in a cirrhotic liver. Hepatic BMP-9 expression was also downregulated in subjects with NASH compared with obese individuals with a healthy liver (GSE48452, figure 1B) and in alcoholic liver cirrhosis compared with mild acute alcoholic hepatitis (GSE94417, figure 1C). Analysis of BMP-9 in two other datasets that included multiple stages of liver pathology leading to hepatocellular cancer (HCC) also revealed a downregulation of BMP-9 (GSE89377, figure 1D; GSE6764, data not shown). BMP-10, the heterodimeric partner of BMP-9, was downregulated in the NASH cohort (figure 1B). Interestingly, Alk1, the principal type I receptor for BMP-9, was upregulated in two of the liver cirrhosis cohorts (GSE36411, data not shown; GSE77627, figure 1A), consistent with the findings of Breitkopf-Heinlein et al, while ACVR2B, the most likely type II receptor for BMP-9, was downregulated in two of the cirrhosis cohorts (GSE6764 and GSE36411, data not shown).
We examined mRNA expression of BMP-9 and its downstream targets, inhibitor of DNA binding (ID) 1, ID2, ID3, ID4 and hepcidin antimicrobial peptide (HAMP) in human HCC. BMP-9 and ID1 (figure 2A and B), and ID2, ID3, ID4 and HAMP (data not shown) were downregulated in HCC compared with normal liver in The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) dataset. Consistently, qRT-PCR in our patient samples revealed reduced BMP-9 and ID1 in HCC compared with paired non-tumour liver (figure 2C and D). BMP-9 target genes were downregulated even though Alk1 was upregulated in the TCGA-LIHC dataset (data not shown) suggesting that BMP-9 downregulation inhibited the pathway despite upregulation of its receptor. Finally, lower BMP-9 expression correlated with reduced overall and disease-free survival in the TCGA-LICH dataset (figure 2E and F).
The role of BMP-9 in cancer is not well understood. A reduced BMP-9 expression is observed not only in HCC but in cholangiocarcinoma, lung adenocarcinoma and lung squamous cell carcinoma in the TCGA dataset. In HCC cells, there are contradicting reports of oncogenic2 3 and tumour suppressor roles for BMP-9.4 As observed in the data presented in this letter, the linear trend of decline in BMP-9 expression from normal liver to various precancerous conditions and in HCC, the reduction of BMP-9 and its major downstream target ID1 in HCC compared with normal liver, as well as the association between BMP-9 expression and survival suggests a tumour suppressive role for BMP-9 in HCC. In sum, in humans, BMP-9 may have a tumour suppressor rather than a dominant profibrogenic role. Further studies are warranted to clarify these findings.
Contributors MJ and LQ: study concept and design. K-JK and SDWB: experimental data generation. MJ, K-JK, SDWB and LQ: data analysis. MJ and LQ: manuscript writing. LQ and JG: critical review of the manuscript. JG: overall guidance of the project.
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.