Introduction: Elevated Glioma-associated Oncogene Homolog-1 (Gli1) protein expression, which is a downstream target of the Hedgehog (Hh) pathway, is an independent poor prognostic factor for serous ovarian cancer1; inhibition of the Hh-signalling pathway in a Gli1-overexpressed high-grade serous ovarian cancer (HGSOC) patient-derived xenograft (PDX) showed significant inhibition of tumour growth particularly in combination with chemotherapy2. However, early phase clinical trials of Vismodegib, a potent Hh inhibitor, failed to show promise, despite 13.5% of advanced HGSOC enrolled in a phase-II vismodegib maintenance trial showing Hh-ligand overexpression3,4. We identified one out of 15 HGSOC PDX demonstrating Hh-overexpression in the corresponding human baseline sample. We explored the efficacy of vismodegib therapy in this Hh-overexpressing HGSOC PDX model.
Method: A cell line was generated from a well-characterised PDX with Indian Hh-overexpression and homozygous deletion of the TSC1 gene. The cell line identity was validated by p53-PCR specific for that HGSOC. In vitro cell viability was assessed in response to vismodegib. In parallel, this PDX model was randomized in vivo to treatment with vismodegib alone, vismodegib in combination with cisplatin, cisplatin alone or vehicle.
Results: The in vitro drug assay confirmed that the Hh-high cell line was vismodegib sensitive with an EC50 of 3.5 µM in comparison with a PDX-derived cell line from a HGSOC, which did not over-express Hh-ligand and was refractory to vismodegib (35 µM). However no response was observed to single agent vismodegib in vivo and, paradoxically, in combination, vismodegib mostly abrogated the effect of the chemotherapeutic agent cisplatin.
Conclusion: Lack of Vismodegib efficacy in Hh-overexpressing solid tumours has been reported to involve crosstalk between Hh-signalling and non-canonical activation of Gli1 by mTOR-signaling5. As this HGSOC PDX harbors bi-allelic deletion of TSC1, which could drive mTOR pathway activity6, we hypothesized that inhibiting both Hh and mTOR may be required in our current PDX model and this is being assessed. This may have important impact for the use of Hh pathway inhibitors in the clinic.