Modeling Resistance to Pathway-Targeted Therapy in Ovarian Cancer

Abstract

A detailed understanding of the biochemical pathways that are responsible for cancer initiation and maintenance is critical to designing targeted cancer therapy. Although we have accumulated knowledge about individual molecular changes that underlie cancer development, we are still learning how multiple biochemical pathways cooperate in cancer. This cooperation and cross-talk between redundant biochemical pathways appear to be the main reasons for the failure of therapeutic agents that are designed to interfere with a specific molecular target. In order to simulate the cooperation of several biochemical pathways in cancer development, we have engineered mouse ovarian cancer cell lines and tumors with different combinations of defined genetic alterations. We have used this system to determine the functional contributions of individual pathways that are necessary for cell proliferation and tumor maintenance, as well as to test the molecular mechanisms of tumor resistance to pathway-targeted therapy.

Progress in the development of effective therapies is hindered by a lack of understanding about the complexities of biochemical pathways that are required for tumor maintenance and the absence of defined experimental systems that simulate altered biochemical processes that occur in cancer initiation and progression. We have designed a system in which multiple genetic alterations can be introduced simultaneously or sequentially into mouse ovarian surface epithelial cells.^1,2 We have recently used this system to test the effectiveness of mammalian target of rapamycin (mTOR) inhibition when several major pathways, such as c-myc, Ras, and Akt, are activated in the tumor cells.³

mTOR inhibitors specifically target mTOR, a downstream mediator in the phosphatidylinositol 3-kinase (PI3K)/Akt pathway, which plays a critical role in regulating cell growth and proliferation. Most of the upstream and downstream components of this pathway are directly implicated in tumor initiation and progression. Moreover, mTOR receives input from several biochemical pathways that are altered in cancer cells. Thus, tumors with a number of distinct molecular changes should, in theory, be sensitive to mTOR inhibition. mTOR inhibitors, such as rapamycin, CCI-779, RAD001 and AP23576, have been tested as suppressors of tumor growth in preclinical models^4–14 and are currently tested as anti-tumor agents in several clinical trials. These agents inhibit the activity of mTOR by forming a complex with the FK binding protein 12 (FKBP-12), which in turn binds to mTOR. This association results in the inactivation of the ribosomal protein S6 kinase (S6K) and the hypophosphorylation of the eukaryotic initiation factor 4E binding protein (4E-BP1). Hypophosphorylated 4E-BP1 associates with the eukaryotic initiation factor 4E (eIF4E),^15,16 thereby inhibiting mRNA translation. Hypersensitivity to mTOR inhibitors may be induced by the loss of PTEN phosphatase^7,8 or p53 tumor suppressor function,¹⁷ or amplification of the GLI oncogene.¹⁸ Although in most preclinical models mTOR inhibitors were potent suppressors of tumor growth, the failure to see robust responses in clinical trials and some preclinical models^11–13 taught us that multiple redundant pathways in cancer are capable of overcoming mTOR inhibition.

Using ovarian cancer cell lines with defined combinations of alterations in p53, c-myc, Akt, K-ras, H-ras, and Her-2 genes, we demonstrated that rapamycin effectively inhibits the growth of cells and tumors that rely on Akt signaling for proliferation and growth. However, cells and tumors in which Akt signaling is not the driving force in proliferation are resistant to rapamycin. We then introduced additional genetic alterations to the rapamycin-resistant and rapamycin-sensitive cell lines. First, we explored whether rapamycin-resistant cells become sensitive to rapamycin when they are transduced with constitutively activated Akt. In contrast to the hypothesis that high levels of Akt sensitize tumor cells to rapamycin inhibition,^7,8,14,19 we demonstrated that the introduction of activated Akt to the rapamycin-resistant cells does not render the cells sensitive to rapamycin if they can utilize alternative pathways for survival and proliferation. These results indicate that mTOR inhibitors may be effective in a subset of tumors that depend on Akt activity for survival, however, not effective in all tumors that exhibit Akt activation.³ This is a subtle, but important, point that brings into question the selection of patients for clinical trials that is currently based on the detection of activated Akt and/or the loss of PTEN expression.

Second, we explored whether the introduction of alternative survival and proliferation pathways induces resistance in rapamycin-sensitive cell lines. Since the Ras and Akt signaling pathways are known to cooperate in facilitating global translation²⁰ and tumor growth,^1,21 we explored the ability of oncogenic Ras to modulate the sensitivity of tumor cells to rapamycin.³ We demonstrated that the rapamycin-sensitive tumors develop resistance to rapamycin when presented with alternative survival pathways, such as the Ras signaling pathway. Signal transduction downstream of oncogenic Ras is mediated through several effectors: the serine/threonine kinase Raf activates the mitogen-activated extracellular kinase (MEK)/extracellular signal-regulated kinase (ERK) pathway; PI3K activates several target proteins including the Akt kinase and the GTPase Rac; the nucleotide exchange factor Ral-GDS activates the Ras-related protein Ral; and the phospholipase Cε pathway activates protein kinase C and induces calcium mobilization from intracellular stores.²² The exact functional contributions of these individual Ras-activated pathways to cell cycle deregulation and cellular transformation remain unclear.

In order to determine whether the MEK/ERK pathway plays a role in cell resistance to rapamycin inhibition, we treated the rapamycin-resistant cells with the MEK inhibitor PD98059 alone or in combination with rapamycin. We showed that cell lines that are resistant to the individual inhibitors respond to the combination of rapamycin and PD98059 by significantly reduced cell proliferation.³ The cell proliferation was not completely inhibited, indicating that Ras pathways other than the MEK/ERK pathway may play a role in the resistance. We did, however, determine that the MEK/ERK pathway is responsible for the epithelial-mesenchymal transition (EMT) in ovarian surface epithelial cells, which typically display epithelial morphology. The introduction of H-ras induced spindle-shaped morphology in these cells (Fig. 1). The spindle-shaped morphology was converted to an epithelial morphology in the presence of the MEK inhibitor PD98059, but not in the presence of the mTOR inhibitor rapamycin (Fig. 1).

Figure 1.

Epithelial-mesenchymal transition is induced by the MEK/ERK signaling pathway in mouse ovarian carcinoma cells. The T22 mouse ovarian carcinoma cell line was engineered to contain genetic alterations in the p53, c-myc and Akt oncogenes. The T22 cells have an epithelial morphology and accumulate crystal violet dye. The introduction of oncogenic H-ras into the T22 cells results in epithelial-mesenchymal transition and a reduced accumulation of crystal-violet dye. Treatment of the T22+H-ras cells with vehicle or 100 ng/ml rapamycin for 48 hours does not change the mesenchymal morphology of the cells. However, the mesenchymal cells revert to an epithelial morphology upon treatment with 50 μM of the MEK inhibitor PD98059.

Defining how the Ras and Akt pathways converge to facilitate translation and tumor progression will be crucial for the development of effective targeted therapy. It is likely that when Ras and Akt pathways are activated simultaneously, they increase expression of key proteins that control cell proliferation and growth, such as cyclin D1 and c-myc. This provides cells with redundant proliferation and growth signals that are exploited when one of the pathways is inhibited. Oncogenic Ras and Akt signals also phosphorylate and inactivate 4E-BP1. 4E-BP1 competes with the eIF4G protein for an overlapping binding site on eIF4E. In the phosphorylated state, the affinity of 4E-BP1 for eIF4E decreases, releasing eIF4E to interact with eIF4G and facilitating translational initiation. Ras is thought to be the foremost signaling pathway that regulates the levels of the cyclin-dependent kinase inhibitor p21. p21 is a short-lived protein, which upon accumulation, associates with cyclin D1 and blocks apoptosis induced by various stimuli. Ras-induced accumulation of p21 has been attributed to both transcriptional activation and post-transcriptional inhibition of proteasome-mediated degradation, as well as p53-dependent and -independent mechanisms.^23–30 Inhibition of mTOR by rapamycin is thought to trigger a potentially lethal response that can be prevented by elevating the levels of p21. Thus, we explored whether any of these proteins are affected when oncogenic H-ras is introduced into cell lines that contain alterations in the p53, c-myc and Akt genes. We determined that the introduction of H-ras results in increased phosphorylation of 4E-BP1 and increased protein levels of cyclin D1 and p21 (Fig. 2). These changes existed even in the presence of rapamycin, indicating that mTOR inactivation is not sufficient to inhibit the synergistic activity of Ras and Akt (Fig. 2). Therefore, the development of agents that lower cyclin D1 and p21 protein levels could be a promising approach to sensitizing tumor cells to mTOR inhibitors.³¹ Several other factors, such as optimal dose, scheduling, and combination with other therapies, may significantly improve the efficacy of the targeted therapy. For example, a combination of mTOR inhibitors and other therapies has been used successfully to overcome drug resistance to cytotoxic chemotherapeutic agents and to sensitize cells to apoptosis.^13,32

Figure 2.

Oncogenic H-ras signaling contributes to the phosphorylation of the 4E-BP1 protein and the accumulation of the cyclin D1 and p21 proteins. These effects cannot be reversed by rapamycin treatment.

However, our data indicate that mTOR inhibitors may not be effective in all tumors that exhibit Akt activation, but effective in a subset of tumors that depend on Akt activity for survival. Tumors with alternative survival pathways may require the inactivation of multiple individual pathways for effective treatment. These results have significant implications for the use of pathway-targeted therapy in advanced human cancers, which typically display numerous genetic alterations. A better characterization of the aberrantly activated pathways in tumors from individual patients may be necessary for successful treatment. An important avenue for future studies will be the development of appropriate high-throughput molecular diagnostic assays for identifying predictive markers for tumors that are likely to respond to mTOR inhibitors.

ABBREVIATIONS

4E-BP1

eukaryotic initiation factor 4E binding protein

eIF4E

eukaryotic initiation factor 4E

EMT

epithelial-mesenchymal transition

ERK

extracellular signal-regulated kinase

mTOR

mammalian target of rapamycin

MEK

mitogen-activated extracellular kinase

PI3K

phosphatidylinositol 3-kinase

S6K

S6 kinase

FKBP-12

FK binding protein 12