The Pathways of Tomorrow: What Does the Future Hold for the Treatment of Colorectal Cancer?

June 1, 2007

By John J. Marshall, MD
Michael Morse, MD

Publication

Article

The currently available therapies for colorectal cancer have led to a significant increase in survival, but the majority of patients with advanced disease progress and eventually die of their disease. This is a particularly frustrating scenario when a patient experiences a complete remission, only to recur with refractory disease.

Introduction

One explanation is that that the currently available drugs-fluoropyrimidines, oxaliplatin (Eloxatin), irinotecan (CPT-11, Camptosar), bevacizumab (Avastin), cetuximab (Erbitux), and panitumumab (Vectibix)-target pathways necessary for growth of the majority of tumor cells, but do not address the tumor cells truly responsible for sustaining the malignant process. Recent evidence suggests that a subset of cells within a tumor have characteristics similar to stem cells. These cells have been termed tumor stem cells or tumor initiating cells.

Tumor Stem Cell Pathways

Although tumor stem cells constitute a minority (<1%) of the tumor bulk and are slowly dividing, they nonetheless give rise to both self-renewing progeny that sustain the stem cell population and proliferating progeny that lead to the more differentiated cells comprising the bulk of the tumor. Indeed, undifferentiated cells expressing CD133+ (prominin-1) and capable of tumor formation and maintenance have been identified in colorectal cancers.¹Unfortunately, these cells are relatively resistant to most therapeutic modalities, particularly chemotherapy and radiotherapy. Although resistance to chemotherapy and radiotherapy has not been well studied in colorectal cancer yet, resistance has been observed in melanoma and glioma stem cells.²

The reasons for this resistance may be due to activated pathways unique to “stem cells,” such as the Wnt, Notch, and Hedgehog (Hh) pathways, or may be due to more efficient repair of DNA damage or enhanced drug efflux compared with the remaining tumor cells. Preliminary studies in animal models suggest that targeting these tumor stem cell pathways either directly or in combination with chemotherapy or radiotherapy leads to enhanced tumor destruction. This module will look at these tumor stem cell pathways and how targeting these pathways may lead to a new paradigm in anticancer therapy for the majority of malignancies, including colorectal cancer.

Wnt Pathway

The Wnt pathway (Figure 1), critical for embryogenesis, is important for decisions regarding differentiation of progenitor cells, and supports proliferation of neoplastic tissue. Simply put, activation of the Wnt pathway, usually by mutations in proteins within the pathway, leads to increased transcription of genes (such as Myc) important for growth, proliferation, differentiation, apoptosis, genetic stability, migration, and angiogenesis. More specifically, the Wnt pathway regulates the ability of Î²-catenin to accumulate in the nucleus and to bind to a member of the DNA-binding T-cell factor/lymphoid enhancer factor (Tcf/Lef) family in a complex. This complex converts Tcf/Lef family members (Tcf1, Lef, Tcf3, and Tcf4) from transcriptional repressors to activators of target genes, including c-Myc and cyclin D1.³

Click to enlarge

In the absence of Wnt signals, Î²-catenin is bound by the tumor suppressors adenomatous polyposis coli (APC) and Axin/Axin2 in a destruction complex along with kinases, which sequentially phosphorylate serine and threonine residues at the N terminus of Î²-catenin,^4,5 with the results that Î²-catenin is targeted for ubiquitination and proteosomal degradation^6,7 and free levels remain low. In this scenario, the Tcf/Lef family acts in the nucleus as a transcriptional repressor. In contrast, canonical Wnt signaling occurs when one of a family of Wnt proteins interacts with a Frizzled family member and a low-density lipid receptor family member (eg, Frizzled-Lrp5/6) and disheveled (DVL) becomes hyperphosphorylated. The signal is then transduced to the APC-GSK3beta-axin-beta-catenin multiprotein complex, which subsequently dissociates. As a result, the destruction complex is inhibited and Î²-catenin accumulates in the cytoplasm, enters the nucleus, and binds to a member of the Tcf/Lef family^8,9 resulting in target gene activation.

In the normal intestine, Wnt signals near the bottom of crypts are crucial for the maintenance of the undifferentiated progenitors.¹⁰ In contrast, activating mutations in the Wnt pathway are found in the majority of intestinal tumors, particularly mutations in APC (allelic loss or inactivating somatic mutations) or Axin2 or Î²-catenin itself.¹¹These mutations ultimately lead to the ability of Î²-catenin to accumulate and activate target genes. For example, mutations in the domain of APC that binds Î²-catenin result in increased levels of nuclear Î²-catenin. Accumulation also occurs by mutations in the serine/threonine residues of Î²-catenin where phosphorylation necessary for its recognition and degradation occurs.

Inactivating Axin2 mutations prevent normal functioning of the destruction complex resulting in high levels of Î²-catenin. The result of Î²-catenin accumulation is activation of a number of genes that lead to polyp and adenoma formation, a prerequisite for most colorectal cancers.

It is also apparent that this pathway is important for existent colorectal cancers because inhibiting the pathway by a number of means (eg, antisense against mutant Î²-catenin) results in decreased tumor growth rate in xenograft models.¹² This has led to a search for drugs targeting the Wnt pathway.¹³ (See Table 1.)

Click to enlarge

Available drugs, such as nonsteroidal anti-inflammatory drugs (NSAIDs), and vitamins (A and D) may have activity against the Wnt/Î²-catenin pathway. Antibodies against Wnt and Frizzled members are in development and may have activity in animal models despite mutated downstream molecules.¹⁴ Studies have identified small molecules that disrupt Tcf-4/Î²-catenin complexes using a high-throughput in vitro screen of natural products,^15,16 although none of these small molecules had “druggable” features. Because this is a protein-protein interaction, it is considerably more difficult to develop small molecule inhibitors that have specificity.

Others have suggested targeting the genes activated by the Wnt pathway, such as c-myc, cyclin D1, CD44, c-MYB, peroxisome proliferator-activated receptor-Î´ (PPARÎ´), COX2, and matrix metalloproteinase 7 (MMP7). Indeed, reduced transcription of genes regulated by the Tcf transcription factor has been demonstrated in colorectal cancer cells treated with celecoxib (Celebrex) and other nonsteroidals such as resveratrol. Currently, there is a clinical trial “to define the actions of resveratrol on the Wnt signaling pathway in... patients with colon cancer [who] receive treatment with resveratrol.”¹⁷

Thiazolidinediones, peroxisome proliferator-activated receptor-gamma ligands currently used for diabetes control, inhibit growth and metastasis of human colon cancer cells by inducing shifts in the localization of Î²-catenin and promoting differentiation of the cells. Curis and Genentech are codeveloping a Wnt pathway antagonist that acts upstream of Î²-catenin. Given the importance of the Wnt pathway to colorectal cancer, more novel approaches to targeting this pathway will be developed.

Notch Pathway

The Notch pathway is critical for the decision points for progenitor differentiation¹⁸ and mutations have been linked conclusively with T-cell acute lymphoblastic leukemias (T-ALL).¹⁹ In the simplest terms, activation of the Notch pathway leads to proliferation of progenitor cells. More specifically, the Notch genes encode transmembrane receptors (Notch 1-4) that interact with transmembrane ligands (Delta-like-1, -3, -4 and Jagged-1 and -2) on adjacent cells. Ligand-receptor binding prompts a series of proteolytic cleavages and posttranslational modifications to Notch (Figure 2), resulting in cleavage of the Notch intracellular domain (NICD) of the Notch receptor within the lipid bilayer by a multiprotein Î³-secretase complex. The NICD subsequently enters the nucleus where it binds CBP-1(RBP-J), converting it from a transcriptional repressor to a transcriptional activator. This results in upregulation of downstream genes such as the repressor Hes1, which in turn represses expression of downstream genes such as neurogenin (ngn), Achaete-Scute, and Math1.

Click to enlarge

While it might not be immediately clear how this pathway might be related to malignancies, Hes1 appears to skew the transit amplifying cells derived from progenitor cells in crypts towards enterocytes, and Math1 skews cells towards the secretory phenotype (ie, goblet, Paneth, enteroendocrine).^20,21 It is therefore thought that Notch signaling is essential within the crypt compartment to maintain the undifferentiated, proliferative state of the crypt progenitors.

Notch pathway intermediaries have been identified in neoplasms in animal models. For example, Notch pathway components and target genes (such as Hes1) are found in adenomas that spontaneously occur in APC mutant mice, such as the multiple intestinal neoplasia (min) mice.^22,23 This suggests that the Wnt and Notch pathways may be active in adenomas. When the Notch pathway was inhibited by a Î³-secretase inhibitor in APC min mice, Math1 expression was induced in adenomas and led to the conversion of proliferative adenoma cells into postmitotic goblet cells.²² A phase I clinical trial for the Î³-secretase inhibitor MK0752 has been initiated for relapsed or refractory T-ALL patients and advanced breast cancers.²⁴ (See Table 2.)

A challenge will be to determine how to use these inhibitors without affecting normal intestinal processes. In rodents, Î³-secretase inhibitors lead to increases in the number of mucosecreting goblet cells.^{22, 25-27} Moreover, proliferation of intestinal epithelium entirely halted with high-dose in vivo treatment.²²

Click to enlarge

Another approach to targeting this pathway may involve other modifying molecules. (See Table 2.) For example, it was recently reported that IKK (inhibitor of nuclear factor ÎºB [NFÎºB] kinase) in an NFÎºB-independentrole in colorectal cancer, is aberrantlyactivated, leading to transcriptional activation of a variety of Notchtargets, including Hes1, Hes5, and Herp2. Inhibition of IKK activity with BAY11-7082resulted in down regulation of these genes and led to tumor apoptosis.²⁸

Hedgehog (Hh) Pathway

The Hedgehog (Hh) signaling pathway is vital for embryonic development and is also involved in the control of normal stem cell proliferation in adult tissues. Activation of the pathway in tumors ultimately results in upregulation of genes that support proliferation of tumor stem cells. Dysregulation/mutation of the pathway intermediaries is linked to basal cell carcinoma and medulloblastoma.²⁹

Three Hh homologues with different spatial and temporal distribution patterns have been identified in humans: Sonic hedgehog (SHh), Indian hedgehog (IHh), and Desert hedgehog (DHh).

The Hh signaling cascade is initiated by Hh binding to Patched 1 (PTCH1) on the target cell. In the absence of the Hh ligand, PTCH1 represses the activity of Smoothened (SMOH), a G-protein coupled receptor (GPCR)-like receptor, presumably by preventing its localization to the cell surface.³⁰ (See Figure 3.) Binding of Hh ligands to PTCH1 causes inhibition of SMOH to be relieved and the Hh signal is transmitted from SMOH through a not completely understood protein complex that results in the appearance of activator forms of the protein GLI that then regulates the expression of cyclins, cyclin dependent kinases, angiogenic factors,³¹ and FOXM1 (forkhead box M1), which may play a role in cell proliferation.

Click to enlarge

Duouard and colleagues³² demonstrated greater SHh gene transcription in human colorectal adenocarcinoma compared with normal tissue and SHh activation was associated with downstream activation of GLI and FOXM1 transcription. The classic smoothened inhibitor is cyclopamine, but it is not a good drug candidate because of its low affinity for SMOH.

Click to enlarge

A number of other compounds are under development for targeting the Hh pathway, including Hh-neutralizing antibodies, forskolin, small molecule inhibitors, GLI antisense, and siRNA.33,34 (See Table 3.) The furthest in development appears to be the benzimidazole derivative, Hh-Antag691, also called HhAntag, which directly binds to SMOH and inhibits the Hh pathway.34 Such drugs have had considerable preliminary activity in preclinical models of medulloblastoma. A phase I study of a Smoothened inhibitor was initiated by Genentech and Curis in January 2007.

Conclusions

The pathways currently targeted in colorectal cancer are, in general, found on the progeny of tumor stem cells. It is becoming increasingly clear that tumor stem cells themselves (or their microenvironments) will need to be directly targeted and interest is increasing in targeting pathways necessary for stem cell function and survival.

Although some drugs may be applied first to tumor sites other than colorectal cancer (such as Hedgehog inhibitors for medulloblastoma), these drugs are expected to be used against colorectal cancer shortly thereafter. It is envisioned that in the future, these stem cell targeting drugs will be used in combination with cytotoxic chemotherapy in order to destroy stem cells as well as their progeny. It will be important to understand the effects on normal stem cells and tissues, such as the bowel, with rapid cell turnover.