Immunotoxin Targeting Glypican-3 Effective against Liver Tumors
Mitchell Ho, Ph.D., and colleagues generated a new immunotoxin, HN3-PE38, that recognizes GPC3 for liver cancer therapy. Administration of HN3-PE38 induces regression of liver tumor xenografts in mice by dual mechanisms: inactivation of cancer signaling (Wnt/Yap) via the antibody and inhibition of protein synthesis via the toxin. This study establishes GPC3 as a promising target for immunotoxin-based liver cancer therapy.
Worldwide, liver cancer is the fifth most common malignancy, and hepatocellular carcinoma (HCC) makes up about two-thirds of those cases. Unfortunately, surgery is still the most effective treatment for HCC but is only available to patients diagnosed at early stages. HCC is known for being particularly drug resistant, sorafenib being the only FDA-approved chemotherapy currently available. Thus, new strategies for treating the disease are urgently needed.
Mitchell Ho, Ph.D., of CCR’s Laboratory of Molecular Biology, and his colleagues began trying to develop novel treatments for patients with HCC by designing antibodies to bind glypican-3 (GPC3), a protein expressed by HCC but not adult normal tissue and that seems to play a role in the Wnt and Yap signaling pathways. One antibody, YP7, binds the GPC3 C-terminus. Another, HN3, binds to a core GPC3 domain, and a third, HS20, recognizes the heparin sulfates on GPC3 and other glypican family proteins. All three demonstrated antitumor activity in vivo, but none could induce liver tumor regression. To increase the potency of their antibodies, the researchers decided to generate immunotoxins by fusing a portion of the antibodies to a 38 kDa fragment of the Pseudomonas exotoxin (PE38), which inhibits protein synthesis.
To be a valid immunotoxin target, the researchers first had to ensure that GPC3 is expressed at sufficient levels on the cell surface and that GPC3 is internalized regularly so that the toxin can be effective. After examining surface staining in a panel of liver cancer cell lines, they found strong GPC3 expression on the HCC line Hep3B and the heptoblastoma line HepG2. The investigators next monitored the internalization of fluorescently-labeled YP7. In the cell lines tested, including Hep3B and HepG2, they found that the amount of GPC3 internalized was two- to fourfold higher than the amount on the cell surface indicating efficient internalization.
The scientists then determined the binding affinities of the immunotoxins on both purified protein and cells. They decided to focus on HN3-PE38 and YP7-PE38 because of their GPC3 specificity. Using enzyme-linked immunosorbent assay and flow cytometry, they found that YP7-PE38 had a fourfold higher affinity for GPC3 protein and a greater than 20-fold higher affinity for cells expressing GPC3 than HN3-PE38. The researchers next examined the cytotoxicity of the immunotoxins. Surprisingly, HN3-PE38 was more effective than YP7-PE38 even though the latter had a higher binding affinity. To ensure that the higher activity of HN3-PE38 was due to GPC3 binding, the investigators assessed cytotoxicity in cells lacking GPC3. Encouragingly, HN3-PE38 was less potent against GPC3 knockdown cells.
The research team next investigated why HN3-PE38 is more potent than YP7-PE38. To exclude the activity of the toxin, the scientists generated a mutant that eliminated its catalytic activity. The immunotoxins incorporating the mutated PE38 showed similar binding affinities to the unmutated versions but were unable to inhibit protein synthesis. The YP7-PE38 mutant had no activity while the HN3-PE38 mutant retained some. These results suggest the antibody fragment is important for the stronger efficacy of the HN3-PE38 immunotoxin.
Since GPC3 is known to be involved in Wnt signaling, the researchers examined Wnt activity using a reporter gene. Intriguingly, the HN3-PE38 mutant could inhibit Wnt signaling, but the YP7-PE38 mutant had no effect. The investigators next assessed the effect of the immunotoxins on Yap signaling. It is currently unclear in mammals how Yap is triggered by cell surface molecules. However, the scientists found that Wnt3a could activate Yap in liver cell lines and demonstrated that the HN3-PE38 mutant could inhibit Wnt3a-induced Yap activity. The YP7-P38 mutant could not. With Yap knockdown Hep3B cells became 50-fold more sensitive to the HN3-PE38 mutant but only two-fold more sensitive to the YP7-PE38 mutant, and the difference between the two disappeared over several days of Yap knockdown. These results show that Yap and GPC3 cooperate and that binding of HN3-PE38 reduced both canonical Wnt and Wnt3a-Yap signaling.
The researchers next examined the effects of the immunotoxins in vivo. First they evaluated the drugs’ toxicities in nude mice. The mice tolerated up to 0.8 mg/kg of HN3-PE38, but YP7-PE38 was lethal at doses above 0.3mg/kg. Then the investigators tested the antitumor activity of the immunotoxins on established subcutaneous Hep3B tumors. At 0.2mg/kg tumors of the mice receiving HN3-PE38 were significantly smaller than those receiving YP7-PE38. The response to HN3-PE38 was dose dependent, and mice receiving 0.8mg/kg exhibited tumor regression and had the best survival. Mice treated with this dose did lose weight during treatment, but their weights returned to normal once treatment ended. Immunohistochemistry of the HN3-PE38 treated tumors showed reduced nuclear β-catenin and phosphorylated Yap, supporting the scientists’ observations in cell lines.
Finally, the researchers wanted to see if they could enhance the activity of the immunotoxins by combining them with chemotherapy. They initially looked at combinations with sorafenib but saw no improved antitumor activity. The investigators next tested a panel of drugs with distinct mechanisms and found that the topoisomerase I inhibitor irinotecan was the most potent against HepG2 xenografts. They tested the combination of HN3-PE38 and irinotecan on Hep3B and HepG2 tumors. In the Hep3B model, the combination caused a modest reduction in tumor size, but growth restarted once treatment was stopped. In the HepG2 model, the combination resulted in synergistic tumor regression. Both tumor weight and levels of serum alpha-fetoprotein, a liver cancer marker, decreased with the combination. Toxicology studies suggested that the combined treatment was well tolerated.
Together these data indicate that GPC3 is a valid therapeutic target in liver cancer and suggest that immunotoxins, particularly those with dual mechanisms of action such as HN3-PE38, are a promising treatment for patients with this deadly cancer.Summary Posted: 03/2015
Gao W, Tang Z, Zhang Y, Feng M, Qian M, Dimitrov DS, and Ho M. Immunotoxin targeting glypican-3 regresses liver cancer via dual inhibition of Wnt signaling and protein synthesis. Nature Communications. March 11, 2015 PubMed Link