Therapeutic options for hepatocellular carcinoma (HCC) remain limited particularly in advanced stages. Immunotherapy with NK- or T-cell augmenting therapies to date has yielded some early promising results [24
] but the low affinity of endogenous tumor-specific T-cell receptors and the immunosuppressive milieu of the tumor microenvironment represent barriers to effectively harnessing the power of the endogenous immune system to control cancer. Yeast-derived scFv offer many advantageous properties for the development of anti-tumor biologics. scFv are inexpensive to produce, easily modifiable e.g. biotinylation [28
], and facile for subsequent cloning in cis
with diagnostic or effector domains.
Identification of an appropriate tumor–associated antigen is an obviously essential requirement for scFv development. Glypican-3 (GPC3), a heparan-sulfate proteoglycan, has recently been identified as a highly specific, membrane-associated tumor antigen found in 49-100% of HCC [2
]. GPC3 is not expressed (or is expressed very focally [29
]) in non-tumorous cirrhotic liver tissue [7
] and expression of GPC3 in other normal tissues appears limited [7
]. GPC3 modulates the effect of growth factors such as IGF-2, BMP-7 and FGF-2 on hepatoma cells [11
] and may recruit M2 tumor-promoting macrophages to the HCC microenvironment [13
]. Emerging evidence also suggests that inhibition of glypican-3 function via knockdown [14
] or competition [12
] has a profound negative effect on HCC proliferation. Expression on the cell surface makes GPC3 an attractive target for antibody-directed therapy. Another group has shown that a murine anti-hGPC3 antibody induces antibody-dependent cytotoxicity that manifests an anti-tumor effect in a xenograft animal model of hepatocellular carcinoma [31
]; this antibody has subsequently been humanized [18
] and is entering human clinical trials. Thus, available evidence suggests that glypican-3 is a rational target for humoral and potentially chimeric immunotherapy for HCC.
In this study, we utilized the paired display/secretion yeast system to isolate five candidate scFv with affinity in the range from 5.0 – 110.9 nM that each demonstrate specificity for binding the surface of glypican-3-expressing cell lines. scFv binding was significantly reduced after specific knockdown of glypican-3. The paired yeast display/secretion system minimizes post-translational and conformational changes in the conversion from displayed to soluble scFvs, a property that allows for consistency during the high throughput screening and validation process [20
]. scFv specificity to the naturally processed glypican-3 protein at physiological conditions was critical given complex post-translational modifications of glypican-3. We utilized increasingly physiological screening criteria to select scFv candidates for further evaluation. Dramatic differences of scFv binding between wild-type and glypican-3-knockdown HepG2 in cell culture conditions confirmed not only the specificity of scFv binding but also the capacity to bind to naturally-processed cell surface glypican-3 in situ. In separate work, we are currently validating a chimeric antigen receptor to redirect T-cells against glypican-3-expressing targets using our 3E11 scFv.
Not surprisingly, scFv had no direct positive or negative impact on cellular proliferation unlike that demonstrated by soluble glypican-3 [12
]. The relatively small size of scFv (27 kD) makes competitive inhibition of growth factor binding unlikely. We did not include agonism or antagonism in our screening strategy, and thus lack of agonist and antagonistic effect is not unexpected. Conjugation of our scFv to cellular cytotoxins will be explored as a potential therapeutic application of the scFv technology.