A combined, rational approach towards inhibition of the MEK-ERK and mTOR pathways in pancreatic ductal adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) will be diagnosed in an estimated 42,470 Americans in 2009.¹ Almost all of these patients will die of this cancer, despite present surgical and medical intervention, with a five-year survival rate estimated to be less than 5%.¹ Hence, PDAC is truly a disease in which no effective treatment is available, underscoring the acute need for novel approaches to therapy.

Since close to 100% of all human PDAC specimens feature an oncogenic point mutation in Kras, many investigators have speculated that the downstream components of the mitogen activated protein kinase (MAPK) pathways would be effective targets for therapy. The MAPK pathways are complex, featuring many points of crosstalk with other pathways and divergence. Hence pharmacologic inhibition of a single component of the pathway, despite a proximal position in the cascade, may not lead to complete arrest of tumor cell growth in patients.

In this issue of Cancer Biology and Therapy, Chang and colleagues⁴ have taken a rational approach towards the augmentation of downstream oncogenic Kras action. The extracellular regulated kinase (ERK) pathway has been shown to be mis-regulated in most PDAC cells, likely due to a Kras mutation. Numerous studies have documented that the ERK pathway is involved in many processes during tumorigenesis, including unchecked proliferation. It is thus not surprising that the ERK pathway interacts with multiple other pathways downstream. One such point of crosstalk is with the mammalian target of rapamycin (mTOR) pathway, which is upregulated in PDAC cells.⁵ It is thought that the multiple points of interaction of the two pathways allow for the discrete pairing of cell cycle activation, by the ERK pathway, and regulation of protein translation, by the mTOR pathway. Chang and colleagues hypothesized that the combined targeting of the ERK and mTOR pathways would lead to a synergistic inhibition of protein synthesis and hence proliferation in PDAC cells. To test this hypothesis, the investigators performed experiments with rapamycin and AZD6244, a potent small molecule inhibitor of MEK, which forms a complex with ERK to mediate downstream effects (Fig. 1).

Figure 1.

Schematized diagram of the pharmacologic inhibition of the convergence of the meK-erK and mtor pathways performed in Chang et al.⁴ Colored boxes indicate site of drug action. tSC2, tuberous sclerosis complex 2.

Indeed, combined treatment with rapamycin and AZD6244 led to enhanced effects in all variables tested in vitro. As expected, dephosphorylation of ERK1/2 as well as the mTOR targets p70 (S6 kinase) and the S6 ribosomal protein, an important regulator of mRNA translation, was found to be enhanced with the dual treatment as compared to each treatment alone. Cell cycle analysis also revealed an enhancement of S-phase inhibition with combination treatment.

To study the effect of this drug combination in a more physiologic context, the investigators next performed experiments in a long-standing xenograft model of pancreatic cancer. After subcutaneous transplantation of BxPC-3 and MIAPaCa-2 cells into the flanks of NOD/SCID mice, a treatment trial was performed. Tumor growth was inhibited to a greater degree with the rapamycin/AZD6244 combination as compared to each alone. Further, the degree of apoptosis was enhanced with double therapy as was the previously measured indices of dual pathway inhibition in this model. Taken together, Chang and colleagues provide robust, cogent evidence for further testing of combined ERK and mTOR pathway inhibition as a basis for therapy for PDAC.

However, the authors’ impressive work needs to be balanced with a major caveat. The xenograft model of PDAC has been thoroughly exploited over many years, since its use by pioneers such as Courtenay⁶ in the 1970s. It was thought to be able to yield more accurate information regarding the behavior of PDAC cells within a physiologic context. As such, many investigators have employed these models to test candidate compounds to treat PDAC; hundreds of drug treatment trials for PDAC have been performed in the subcutaneous xenograft model, with the vast majority yielding positive results since the 1980s.⁷ And yet, none of these treatments have been translated into effective therapy for human PDAC.

A recent publication by the Tuveson group points to the microenvironment as the reason for this disconnect.⁸ As most pathologists can attest, human PDAC specimens are relatively avascular compared to other carcinomas. Further, due to a robust desmoplastic response, PDAC lesions tend to be insulated from the circulation by stromal and inflammatory cells. The theorized result of these phenomena is the impairment of drug delivery to the pancreatic tumor itself. Based on Olive et al. subcutaneous xenografts, and, to a lesser degree, orthotopic xenografts result in histology different from human PDAC—transplanted lesions tend to be more vascularized. These results could explain why gemcitabine works so effectively in xenograft models of PDAC, but so horribly in the clinic.

How can we model PDAC more effectively? Olive et al. were able to show that transgenic mouse models resulting in spontaneous PDAC closely recapitulated the human avascular microenvironment.⁸ Further, these tumors did not respond to gemcitabine treatment as intra-tumor concentrations were minimal due to impaired delivery of blood-borne drugs to the cancer. These data suggest that transgenic mouse models of PDAC would yield more accurate information regarding the efficacy of treatment compared to xenograft models.

The work of Chang et al. is an example of a rational approach to developing a potential therapeutic regimen for PDAC. Their experiments were well-planned and thoroughly carried out. However, as can be said about studies preceding theirs, how can we be sure that these results will translate to humans? Moreover, how can we justify continued testing of candidate compounds in humans based on efficacy in PDAC xenograft models, when so many have failed? In this author’s opinion, it is time to change the paradigm of candidate drug testing for PDAC. The interposition of a treatment trial in transgenic mouse models of PDAC between testing in vitro and before early phase clinical trials in humans may be in order. Such an approach may yield not only more accurate information on the biology of pancreatic cancer, but also greater efficiency in identifying a potential cure for this lethal disease.