|Home | About | Journals | Submit | Contact Us | Français|
In contrast to common genomic amplifications that support cancer cell growth by rewiring intracellular signaling, VEGFA amplification drives tumor cell proliferation via the tumor microenvironment. VEGFA amplification is present in a subset of mouse and human hepatocellular carcinomas (HCCs) that appear to be particularly sensitive to sorafenib treatment, indicating its potential value as a biomarker for HCC treatment.
Genomic amplifications (also called amplicons) are known to play different roles in various aspects of tumor development and progression by increasing the expression of proto-oncogenes encoded in the amplified regions. Amplifications may occur any time during tumor progression,1 and cells harboring amplifications that provide a selective advantage form clonal tumors. This makes identification of amplicons appealing as they may contain genes responsible for driving tumorigenesis.
We applied array comparative genomic hybridization (aCGH), a widely-used method for identification of recurrent tumor amplicons, to hepatocellular carcinoma (HCC) from Mdr2 deficient mice, a mouse model for inflammation-induced HCC.2 Among several amplifications and deletions,3 we focused on a recurrent amplification in Chr17qB3 harboring the VEGFA gene for 2 main reasons: (1) A syntenic amplification (Chr6p21) was previously reported in several different human tumor types, including HCC;4 (2) It was previously shown that VEGFA promotes hepatocyte proliferation through induction of hepatocyte growth factor (HGF) in the sinusoidal endothelium.5
Validating the aCGH results, we found Chr17qB3 amplification in 14% of 93 Mdr2−/− HCCs analyzed, close to the frequency of 7–10% reported previously in human HCC.4,6-7 Tumors bearing this amplicon showed increased mRNA expression levels of several genes residing on this amplicon, including VEGFA. We also found increased VEGFA protein levels in VEGFA amplified tumor lysates and plasma samples. Amplicon-bearing mouse HCCs had more blood vessels, higher proliferation indices, and increased macrophage counts, probably of the protumorigenic M2 phenotype.
Previous studies have shown that upon systemic elevation of VEGFA levels or under regenerative conditions, secreted VEGFA induces expression of HGF in the liver sinusoidal endothelium.5 Another study showed that constitutively increased expression of VEGFA leads to a massive recruitment of monocytes.8 Taking both observations together, we postulated that the VEGFA amplicon leads to recruitment of cells that express HGF. Similar to previous reports,5 we also demonstrated that HCC transformed hepatocytes do not express the VEGFA receptors Flt or Kdr, yet expression of the HGF receptor c-Met is markedly higher in hepatocytes than in liver residing macrophages. We thus portray a unique tumor cell–microenvironment interaction in which one cell type secretes a molecule that activates the other cell type, yet by itself is inert.
These observations point to a three-fold protumorigenic activity of the VEGFA amplicon: (1) Increased angiogenesis, a well-known role of VEGFA in cancer; (2) Recruitment of macrophages carrying the protumorigenic M2 phenotype; (3) Induction of reciprocal HGF secretion by cells in the microenvironment that enhances cancer cell proliferation (Fig. 1). The protumorigenic features associated with this amplicon render it a potential tumor Achilles’ heel as VEGFA inhibition would abolish all these properties concomitantly.
To test this hypothesis we treated Mdr2−/− mice with VEGFA blocking agents at the age of 14–18 months, when most of the mice bear tumors. We injected mice with adenoviral vectors expressing soluble VEGFA receptor, sacrificed them 10 d later, and analyzed the tumors for presence of the VEGFA amplicon. We also measured proliferative indices of treated tumors and found decreased proliferation only in the group harboring the VEGFA amplicon.3 As predicted by our hypothesis, this was accompanied by downregulation of HGF in amplicon-positive treated tumors. To test whether VEGFA overexpression alone is sufficient to instigate this protumorigenic crosstalk, we transduced a human HCC cell line (Hep3B) with a lentiviral vector overexpressing VEGFA. When grown in vitro the transduced cells showed no proliferative advantage, but upon xenotransplantation to immunodeficient mice they manifested faster growth, increased proliferation, and higher HGF levels. By purifying macrophage and endothelium cells from these tumor explants we found that HGF expression was restricted to the macrophage fraction of the tumor. Taken together, our data implicate a role for VEGFA overexpression in a tumor promoting circuit via induction of macrophage-derived HGF, which in turn boosts cancer cell proliferation.
Sorafenib is the only drug that shows some clinical advantage in patients with advanced HCC.9 Sorafenib is a small molecule multikinase inhibitor, the targets of which include BRAF, CRAF, PDGFR2, c-KIT, and the VEGFA receptors. As sorafenib inhibits VEGFA signaling, we tested its effect on Mdr2−/− murine HCCs and human Hep3B xenografts. Sorafenib reproduced the inhibitory effects of soluble VEGFR; therefore, despite its relatively broad target spectrum, sorafenib appeared to have a selective antitumor effect in tumors harboring VEGFA overexpression.
The demonstration that VEGFA amplification could predict the response to anti-VEGFA treatments in mice prompted us to study its predictive power in human HCCs. This study was hampered by the fact that, unlike most other tumors, many patients with advanced HCC are treated with sorafenib without obtaining a tissue diagnosis. Therefore, we collaborated with colleagues from 3 medical centers—University Hospital Heidelberg, Hannover Medical School, and University Hospital Basel—to obtain resected HCC specimens from patients that were mostly treated with sorafenib following recurrence. Consistent with the mouse data, we found that sorafenib treatment is associated with much better survival in VEGFA amplicon-positive tumors than in either VEGFA amplicon-negative patients or in amplicon-positive patients who did not receive sorafenib. These data suggest that the presence of VEGFA amplification can predict the clinical response to sorafenib in patients with HCC.
To conclude, we provide an example of use of a mouse model to identify a possibly useful tumor biomarker for human cancer therapy. A recent Phase III study of sorafenib as an adjuvant treatment following HCC surgery did not show a survival advantage in all patients. It is, however, possible that classification of HCC patients according to the presence or absence of VEGFA amplification might reveal a differential beneficial effect of sorafenib.10 In addition, our study calls for prospective testing of VEGFA amplification as a biomarker for patient response to sorafenib. Chromosomal gains are fairly common in human cancer and may be the prevailing cancer driver aberrations in certain types of cancer. Mouse tumors harboring chromosomal gains may therefore be very useful for predicting therapeutic efficacy in human cancer, especially when the overall response in an unselected cancer patient population is relatively low.
No potential conflicts of interest were disclosed.
Research leading to these results received funding from the European Research Council Grant Agreements 281738 to E.P. and 294390 to Y.B-N.; the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation (AMRF); the DKFZ-MOST cooperation; and the Israel Science Foundation (ISF) Centers of Excellence (1779/11).