Pancreatic ductal adenocarcinoma (PDA) is among the most lethal human cancers, in part because it is insensitive to many chemotherapeutic drugs. Studying a mouse model of PDA that is refractory to the clinically used drug gemcitabine, we found that the tumors in this model were poorly perfused and poorly vascularized, properties that are shared with human PDA. We tested whether the delivery and efficacy of gemcitabine in the mice could be improved by coadministration of IPI-926, a drug that depletes tumor-associated stromal tissue by inhibiting the † To whom correspondence should be addressed. david.tuveson@cancer.org.uk.

Several Hsp90 (heat shock protein 90) inhibitors are currently under clinical evaluation as anticancer agents. However, the correlation between the duration and magnitude of Hsp90 inhibition and the downstream effects on client protein degradation and cancer cell growth inhibition has not been thoroughly investigated. To investigate the relationship between Hsp90 inhibition and cellular effects, we developed a method that measures drug occupancy on Hsp90 after treatment with the Hsp90 inhibitor IPI-504 in living cells and in tumor xenografts. In cells, we find the level of Hsp90 occupancy to be directly correlated with cell growth inhibition. At the molecular level, the relationship between Hsp90 occupancy and Hsp90 client protein degradation was examined for different client proteins. For sensitive Hsp90 clients (e.g. HER2 (human epidermal growth factor receptor 2), client protein levels directly mirror Hsp90 occupancy at all time points after IPI-504 administration. For insensitive client proteins, we find that protein abundance matches Hsp90 occupancy only after prolonged incubation with drug. Additionally, we investigate the correlation between plasma pharmacokinetics (PK), tumor PK, pharmacodynamics (PD) (client protein degradation), tumor growth inhibition, and Hsp90 occupancy in a xenograft model of human cancer. Our results indicate Hsp90 occupancy to be a better predictor of PD than either plasma PK or tumor PK. In the nonsmall cell lung cancer xenograft model studied, a linear correlation between Hsp90 occupancy and tumor growth inhibition was found. This novel binding assay was evaluated both in vitro and in vivo and could be used as a pharmacodynamic readout in the clinic.

A requisite step for canonical Hedgehog (Hh) pathway activation by Sonic Hedgehog (Shh) ligand is accumulation of Smoothened (Smo) to the primary cilium (PC). Activation of the Hh pathway has been implicated in a broad range of cancers, and several Smo antagonists are being assessed clinically, one of which is approved for the treatment of advanced basal cell carcinoma. Recent reports demonstrate that various Smo antagonists differentially impact Smo localization to the PC while still exerting inhibitory activity. In contrast to other synthetic small molecule Smo antagonists, the natural product cyclopamine binds to and promotes ciliary accumulation of Smo and “primes” cells for Hh pathway hyper-responsiveness after compound withdrawal. We compared the properties of IPI-926, a semi-synthetic cyclopamine analog, to cyclopamine with regard to potency, ciliary Smo accumulation, and Hh pathway activity after compound withdrawal. Like cyclopamine, IPI-926 promoted accumulation of Smo to the PC. However, in contrast to cyclopamine, IPI-926 treatment did not prime cells for hyper-responsiveness to Shh stimulation after compound withdrawal, but instead demonstrated continuous inhibition of signaling. By comparing the levels of drug-induced ciliary Smo accumulation with the degree of Hh pathway activity after compound withdrawal, we propose that a critical threshold of ciliary Smo is necessary for “priming” activity to occur. This “priming” appears achievable with cyclopamine, but not IPI-926, and is cell-line dependent. Additionally, IPI-926 activity was evaluated in a murine tumor xenograft model and a pharmacokinetic/pharmacodynamic relationship was examined to assess for in vivo evidence of Hh pathway hyper-responsiveness. Plasma concentrations of IPI-926 correlated with the degree and duration of Hh pathway suppression, and pathway activity did not exceed baseline levels out to 96 hours post dose. The overall findings suggest that IPI-926 possesses unique biophysical and pharmacological properties that result in Hh pathway inhibition in a manner that differentiates it from cyclopamine.

1. IPI-926 is a novel semisynthetic cyclopamine derivative that is a potent and selective Smoothened inhibitor that blocks the hedgehog signal transduction pathway. 2. The in vivo clearance of IPI-926 is low in mouse and dog and moderate in monkey. The volume of distribution is high across species. Oral bioavailability ranges from moderate in monkey to high in mouse and dog. Predicted human clearance using simple allometry is low (24 L h(-1)), predicted volume of distribution is high (469 L) and predicted half-life is long (20 h). 3. IPI-926 is highly bound to plasma proteins and has minimal interaction with human α-1-acid glycoprotein. 4. In vitro metabolic stability ranges from stable to moderately stable. Twelve oxidative metabolites were detected in mouse, rat, dog, monkey and human liver microsome incubations and none were unique to human. 5. IPI-926 is not a potent reversible inhibitor of CYP1A2, 2C8, 2C9 or 3A4 (testosterone). IPI-926 is a moderate inhibitor of CYP2C19, 2D6 and 3A4 (midazolam) with KI values of 19, 16 and 4.5 µM, respectively. IPI-926 is both a substrate and inhibitor (IC50 = 1.9 µM) of P-glycoprotein. 6. In summary, IPI-926 has desirable pre-clinical absorption, distribution, metabolism and excretion properties.

Ductal pancreatic tumors are unusually resistant to chemotherapy, exhibiting primary resistance to each of the numerous regimens tested to date. These tumors are also unusually desmoplastic and harbor a sparse and inefficient vasculature. The resulting lack of perfusion appears to play a role in the inefficient delivery of numerous chemotherapeutic agents to the parenchyma of pancreatic tumors. Recent evidence has implicated the sonic hedgehog pathway in promoting desmoplasia through a paracrine signaling mechanism active in pancreatic tumors of both humans and the Kras/p53/PdxCre (KPC) mouse model. We sought to evaluate the dependence of stromal desmoplasia on Hedgehog (Hh) pathway signaling and, by extension, assess the effects of Hh pathway inhibition on drug delivery and chemoresistance. Using IPI-926, a semisynthetic inhibitor of the Smoothened protein, we found that stromal contribution was markedly diminished within 10 days of inhibition of the Hh pathway. Paradoxically, the depletion of stromal cells was coincident with an increase in microvessel density, despite previous data indicating a pro-angiogenic role for the Hh pathway. These non-cell autonomous changes resulted in an increase in the delivery of small molecules to KPC pancreatic tumors and an increase in apoptosis when administered in combination with gemcitabine. Ultimately, mice treated with both IPI-926 and gemcitabine had a significant extension of survival and a decrease in the incidence of liver metastases. We conclude that the Hh pathway plays an important role in the maintenance of pancreatic tumor stroma and that this contributes to the primary chemoresistance of pancreatic tumors. These data also support the clinical evaluation of agents that target the stroma in pancreatic cancer. Citation Information: Clin Cancer Res 2010;16(14 Suppl):PR2.