Several agents targeting the hedgehog pathway are under preclinical and clinical evaluation. In this article, we describe various classes of these agents and the evidence supporting their anticancer activity.
Usha Malhotra, MD
Assistant Professor, Department of Medicine, Roswell
Park Cancer Institute, Buffalo, NY;
Hedgehog, a crucial signaling pathway involved in embryonic development and patterning, is postulated to be active during carcinogenesis. In a number of cancers, this pathway is activated either through mutations in genes encoding for protein components or by overproduction of ligands. Several agents targeting this pathway are under preclinical and clinical evaluation. In this article, we describe various classes of these agents and the evidence supporting their anticancer activity.Hedgehog (Hh) is an embryonic developmental pathway initially described inDrosophila, a genus of small flies, as a critical component of signal transduction during organogenesis.1Studies have confirmed the highly conserved nature of this pathway in other species including vertebrates.2,3The key components of this complex pathway in vertebrates include the Hh ligands, Sonic hedgehog (Shh), Indian hedgehog, and Desert hedgehog; 12-transmembrane receptor patched homologue 1 (ptch1); 7-transmembrane activator protein smoothened (Smo); a cytoplasmic inhibitory protein called suppressor of fused (Sufu); and a family of zinc-finger transcription factors called glioma-associated oncogene homologue (Gli).
The Hh pathway is inactive under normal physiologic conditions in most differentiated tissues. Activation in response to various stimuli, including tissue injury and inflammation, occurs in a liganddependent manner. Binding of a Hh ligand to ptch1 causes de-repression of Smo, leading to dissociation of the Sufu-Gli complex, followed by nuclear translocation of Gli, and subsequently, translation of Gli-associated genes involved in cell proliferation, migration, cell cycle checkpoints, apoptosis, vascularization, etc (Figure). A comprehensive description of the components of this pathway and their function in different species are beyond the scope of this review and are described in detail in a number of published reviews.4-6The first evidence linking the Hh pathway with cancer was described in Gorlin’s syndrome-associated cancers. These cancers include basal cell carcinoma and medulloblastoma, and are associated with mutations in Hh pathway components. The majority of basal cell carcinomas (up to 70%) exhibit mutations in ptch or Smo, and up to 25% of medulloblastomas are also associated with mutations leading to constitutive signaling of the Hh pathway, and hence, continuous activation of downstream events leading to uncontrolled cancer cell growth.7,8
In most other solid tumors, the mode of activation of this pathway has been postulated to be liganddependent in an autocrine or paracrine manner. In the autocrine model, the ligand is secreted and pathway activation occurs in the same tumor cell. The paracrine model involves secretion of the ligand by tumor cells and activation of downstream machinery in the stromal compartment.9The stromal compartment then sends a feedback activation signal to tumor cells. Additionally, a reverse paracrine mode of action has been described in hematologic malignancies such as B-cell lymphoma and multiple myeloma, wherein the ligand is secreted by the stromal cells and helps to maintain the viability and proliferation of tumor cells.9
The precise role of the Hh pathway in cancer is still not clearly elucidated; however, aberrant Hh signaling has been implicated in up to 25% of cancers.10Based on downstream genes activated through Hh pathway signaling, this pathway could potentially be involved at different stages of carcinogenesis, including initiation of cancer, tumor progression, and development of metastases. Additionally, this pathway has also been proposed to be involved in maintenance of cancer stem cells, quiescent cancer cells that retain the ability to self-replicate, self-renew, and differentiate to diverse lineages.11Retrospective studies associating this pathway with radiation resistance provide an additional potential avenue that warrants further exploration and provides a rationale to combine Hh inhibition with chemoradiation.In parallel to scientific evidence demonstrating the Hh pathway’s involvement in cancer, there has been great progress in the development and evolution of various compounds targeting this pathway, including Smo antagonists, Gli transcriptional activity inhibitors, and direct Hh ligand inhibitors (Table). Additionally, there is evidence supporting Hh inhibition by the antifungal agent itraconazole as well as by arsenic trioxide.
1a: Inactive pathway: Canonical Hedgehog (Hh) pathway is inactive in the absence of ligands or activating mutations. 1b: Constitutive pathway activation: Upstream mutations and release of inhibition leads to continuous activation of downstream machinery. 1c: Ligand-dependent activation: Binding of Hh ligands releases inhibition of Smo followed by downstream signaling. 1d: Mechanisms of ligand-dependent signaling based on tumor-stromal interaction.
Most of the agents that are in advanced clinical phase testing are Smo inhibitors. Recent data have revealed limitations associated with utilization of Smo inhibitors, including development of resistance after initial response and inactivity in a number of cancers that do not harbor activating mutations in Hh pathway components. This has led to investigation of a wider category of agents directed at components of the Hh signaling pathway that are upstream and downstream of Smo.
The earliest compound shown to have activity against the Hh pathway was cyclopamine, an alkaloid present in the corn lily that inhibits Smo.12Poor oral bioavailability, limited potency, and unfavorable pharmacokinetics due to low solubility in an acid environment were major limitations for further development of cyclopamine. This led to development of a number of other derivatives and synthetic compounds targeting Smo (Table). The only molecule in this class currently approved by the FDA for clinical use is vismodegib (Erivedge, formerly GDC 0449). It is an orally active small molecule that binds to Smo and has been evaluated extensively in preclinical as well as clinical studies for various cancers.
The most affirmative data have been shown in cancers that harbor Hh pathway-activating mutations, in particular, basal cell carcinoma, and to a lesser extent, medulloblastoma. In its initial evaluation in a phase I trial, vismodegib was associated with remarkable activity in basal cell carcinoma, which led to inclusion of a dose expansion cohort for basal cell carcinoma only.13As a part of this study, 33 additional patients with advanced basal cell carcinoma were enrolled; an objective response was seen in 18 patients. Absence of any incremental increases in plasma steady state concentration on dose escalation, along with demonstration of clinical efficacy at the lowest initial starting dosage of 150 mg orally once daily, led to this dosage being finalized for phase II evaluation.
Following this, a phase II study was initiated to assess efficacy of vismodegib in patients with advanced basal cell carcinoma.14A disease control rate as high as 86% was reported in this trial, leading to the FDA approval of vismodegib in advanced basal cell carcinoma in January 201215 at a dosage of 150 mg orally daily.
Major toxicities reported in the phase II study were graded as mild to moderate and included muscle spasms, alopecia, dysguesia, fatigue, nausea, and loss of appetite. At the data cut-off point for analysis, 30% to 45 % of patients had discontinued treatment for reasons other than progression of disease, with median duration of treatment of about 10 months. This highlights the fact that high incidence of these toxicities was a major factor for discontinuation of therapy. These toxicities are on-target effects and are related to the role of Hh signaling during normal physiologic processes, and hence, a class effect of Smo inhibitors that is further substantiated by similar toxicity data with other agents.16-18
Another challenge associated with the use of vismodegib has been limited duration of response due to the development of resistance. Both in medulloblastoma and basal cell carcinoma harboring Hh pathway-activating mutations, resistance after initial response has been observed, potentially as the result of acquired tumor-specific Smo mutations that prevent vismodegib from binding to Smo protein. Other postulated mechanisms of resistance include amplification of downstream targets such as Gli and cyclin D1, and activation of other pathways such as PI3 kinase. Thus, strategies to prevent the emergence of resistance, as well as alternative therapies to treat resistant or nonresponsive disease, are clearly needed.
In addition to its efficacy as a therapeutic agent for basal cell carcinoma, there is evidence supporting vismodegib as a potential preventive agent in basal cell nevus syndrome.19However, there is considerable toxicity after long-term use, and so it is essential to define an appropriate target group that would derive benefit from such intervention.
With respect to the role of Smo inhibition in other Hh ligand-dependent cancers, single-agent responses have not been observed, leading to utilization of Smo inhibitors in combination with cytotoxic chemotherapy. In addition to vismodegib, other Smo inhibitors are in various phases of development. Among them, IPI-926 (saridegib) has advanced to phase II trials. Recently, a phase II trial utilizing saridegib in combination with gemcitabine for pancreatic cancer was stopped due to lack of efficacy.20Other agents currently in phase I/II trials are listed in the Table. Most of these studies are currently ongoing or recently completed with results being eagerly awaited.
Phase of Development
Approved for basal cell carcinoma Phase I/II for multiple cancer types (NCT00636610, NCT00982592, NCT00887159, NCT01154452, NCT01774253, NCT01537107 NCT01601184)
Phase Ib and II for multiple cancer types (NCT00670189, NCT00909402, NCT00884546, NCT00927875)
Phase I for solid tumors (NCT00761696)
Phase I for solid tumors (NCT00880308)
Phase I for CML (NCT00953758)
Phase I for solid tumors (NCT01106508)
Phase I for nonhematologic malignancies (NCT01204073)
Sonic Hh ligand inhibitor
Utilization of these agents in ligand-driven cancers that show Hh pathway activation has been particularly challenging, as activation seen in tumor specimens may not translate to response, emphasizing the need for predictive biomarkers.
Gli InhibitorsAs already discussed, Smo inhibitors are associated with several limitations, including the development of resistance and the inefficacy of such agents when Hh pathway activation occurs due to events downstream of Smo. The effort to expand development of compounds directed at other components of the Hh pathway has led to identification of agents directed at Gli. Two such agents, GANT-58 and GANT-61, are currently being studied in preclinical tumor models.21GANT-61 is a small molecule being evaluated in colon cancer cell lines and was shown to interfere with binding of Gli1 and Gli2 to promoter regions of its target genes.22In addition, four new molecules, HPI (Hedgehog pathway inhibitor) 1-4, are also being evaluated as potential Gli antagonists.
Hh Ligand Inhibition
Another potential therapeutic approach, especially relevant in cancers with paracrine activation of the Hh pathway, is direct inhibition of the Hh ligands. Robotnikinin, a small molecule that binds directly to the amino terminal of the Sh ligand, has recently been identified and is undergoing preclinical evaluation.23
Other Classes: Arsenic Trioxide, Itraconazole
Arsenic trioxide has recently been shown to have an inhibitory effect on the Hh pathway by directly binding to Gli1/2 and interfering with transcriptional activity. Some preclinical studies also support interference with the accumulation of Gli2 in primary cilium by arsenic trioxide.24
The antifungal drug itraconazole has been shown to inhibit Hh pathway activation by Smo antagonism. Preclinical studies support its potential role as a Hh pathway-directed agent,25but currently no active clinical studies are ongoing to evaluate its role in Hh pathway-dependent cancers.There is evidence to support activation of the Hh pathway in a number of cancers in addition to basal cell carcinoma and medulloblastoma. The limitations of agents targeting tumors with activating mutations of the pathway components include short duration of response and the development of resistance underscoring the need to evaluate alternative strategies for targeting this pathway. On the other hand, the precise therapeutic niche for targeting this pathway in cancers supporting ligand-dependent activation still remains to be defined.
The early-stage development of several agents targeting the Hh pathway underscores the potential for rapid translation into clinical practice. However, in the absence of accurate predictive markers, the challenge remains to characterize a population of patients who would benefit from these agents.
Dr. Malhotra has no conflicts of interest to report.