Novel Antibody Targets Glypican-3 in Liver Cancer
New treatments for patients with liver cancer, the third most common cause of cancer-related death, are desperately needed. Hepatocellular carcinoma (HCC) is the most common type of liver cancer, and HCC tumors are particularly insensitive to chemotherapy. Surgery is the standard treatment for HCCs caught early, but only about a third of cases are identified at this stage. Antibody therapy offers a potential alternative for treating later-stage tumors.
One attractive antibody target is glypican-3 (GPC3), a cell surface associated protein that is highly overexpressed on HCC cells but not on normal liver tissue. While the exact function of GPC3 is not well understood, the protein seems to play a role in regulating cell growth. Mitchell Ho, Ph.D., in CCR’s Laboratory of Molecular Biology, and his colleagues decided to further investigate the role of GPC3 in HCC and to develop GPC3-specific antibodies.
The researchers began their studies by generating three small hairpin RNAs (shRNAs) designed to target GPC3 to see whether its loss would affect HCC cell growth. In two HCC cell lines, shRNAs 1 and 2 reduced the expression of GPC3 by more than 90 percent while shRNA 3 inhibited GPC3 levels by only 20 percent. Similar to their levels of GPC3 knock-down, shRNAs 1 and 2 greatly reduced the growth of both cell lines, and shRNA 3 moderately reduced their growth, supporting the idea that GPC3 signaling plays a critical role in controlling HCC proliferation.
To identify potential GPC3 targeting antibody domains, the investigators screened a phage library of heavy-chain variable domains for their ability to bind recombinant full-length GPC3 proteins. The domain library was previously described by Dimiter Dimitrov, Ph.D. in CCR’s Nanobiology Program. The advantage of using single domains is that they can interact with small pockets or other regions on the GPC3 surface that full-length antibodies cannot recognize. After four rounds of phage panning, Mingqian Feng, Ph.D., a Postdoctoral Fellow in the Ho lab, identified four potential interacting domains. Since the HN3 domain had the strongest association with GPC3, the scientists used this domain for their subsequent studies.
The researchers cloned the HN3 domain into a vector that fused the domain with the constant region of a human antibody and stably expressed the fusion antibody in a cell line for large scale production. To see whether the HN3 antibody could recognize native GPC3, the investigators treated HCC or GPC3-negative cell lines with the HN3 antibody. Only cells that expressed GPC3 bound to the antibody with high affinity.
The scientists then examined the GPC3 domain recognized by HN3. The antibody could bind wild type GPC3 as well as a GPC3 core protein lacking heparan sulfate chains. The HN3 antibody failed to bind N-terminal or C-terminal GPC3 fragments expressed alone but recognized both when the fragments were co-expressed, suggesting that HN3 binds a domain formed by both the N- and C-termini. In support of this idea, HN3 did not recognize denatured GPC3 in Western blots but could immunoprecipitate native GPC3 from HCC cell lines and stain HCC patient tissue samples.
To see what effect HN3 antibody binding had on cell growth, the researchers treated four HCC cell lines with varying amounts of the antibody. The HN3 antibody significantly reduced cell growth in three of the four lines at 0.1μM and reduced the growth of the fourth cell line at 1.0μM. A cell line that normally lacks GPC3 was not affected by HN3 antibody treatment. Likewise, the HN3 antibody did not slow the growth of HCC cells that had their GPC3 expression knocked-down by shRNA 2.
The investigators determined that in HCC cells HN3 antibody treatment and GPC3 knock-down caused cell cycle arrest in G1. To understand how HN3 caused growth arrest, the scientists examined several signaling pathways associated with proliferation in four HCC cell lines. HN3 treatment decreased pro-growth phospho-Erk and phospho-Akt levels in all four cell lines. Levels of phosphorylated yap, an inactive form of this Hippo pathway member, increased in three of the cell lines in response to HN3. Expression of cyclin D1, a target of yap, decreased in all four cell lines. Similar changes were observed with GPC3 knock-down, except that phospho-Erk and -Akt only decreased in one cell line. Because Erk and Akt can affect Hippo signaling, these results suggested that yap may play a role in HN3-mediated cell cycle arrest.
To test this idea, the researchers over-expressed constitutively-active yap or knocked down yap levels in an HCC cell line and evaluated HN3’s ability to induce growth arrest. Cells over-expressing active yap had higher proliferation rates and were insensitive to treatment with HN3 antibody. Conversely, cells lacking yap had much lower levels of cell growth. Adding HN3 antibody did not change their growth rate. These studies indicate that yap is an important downstream target of the HN3 antibody and normal GPC3 signaling.
Finally, the researchers tested their antibody in an animal model by treating mice bearing HCC cell line-derived tumors with HN3 antibody or a control twice a week. HN3 antibody treatment significantly reduced the size of tumors from two different HCC cell lines. The scientists also detected increases in phospho-yap levels and decreases in cyclin D1 and phospho-Erk in treated tumors. The investigators noted that they observed no antibody-related toxicities and suggested that the HN3 antibody should undergo further testing as a potential therapeutic for patients with liver cancer.Summary Posted: 03/2013
Feng M, Gao W, Wang R, Chen W, Man YG, Figg WD, Wang XW, Dimitrov DS, Ho M. Therapeutically targeting glypican-3 via a conformation-specific single-domain antibody in hepatocellular carcinoma. PNAS. PubMed Link