Cilengitide

Abstract

Integrins are critical intermediaries in a wide spectrum of cancer cell activities and thus represent a highly attractive target in oncology therapy. Nonetheless, successful exploitation of anti-integrin therapeutics has proven challenging to date for cancer patients. In this review, we will focus on cilengitide, an RGD pentapeptide inhibitor of α V integrins. Although several integrin inhibitors are under clinical evaluation, cilengitide is the most clinically advanced and is emerging as a prototype for this class of anticancer therapy. A foundation of encouraging preclinical studies led to a well-designed clinical development plan that culminated in a pivotal phase III study of cilengitide in combination with radiation therapy and temozolomide chemotherapy for newly diagnosed glioblastoma patients. Accrual to this study recently completed, while phase II studies of cilengitide are ongoing for head and neck cancer as well as lung cancer. Important future considerations for cilengitide and other integrin-targeting agents will likely include the identification of optimal combinatorial regimens and the delineation of biomarkers associated with efficacy.

Introduction

Integrins are heterodimeric transmembrane receptors that strategically bridge between cancer cells as well as between cancer cells and other important cellular and noncellular components in the tumor microenvironment. Specific ligand binding is determined by α and β subunit pairings that define each of the 24 members of the integrin family. Binding of an integrin with its ligand leads to recruitment of key signaling and adapter proteins into focal adhesion complexes that localize within the cell membrane. Focal adhesion complexes in turn trigger cellular signaling including the NF-κB,^1,2 PI3/Akt,³ SRC,⁴ and ras/MAPK kinase^5,6 pathways that regulate survival, proliferation, invasion, and angiogenesis.

Many cell types in the tumor microenvironment express several different integrins including endothelial cells, pericytes, infiltrating myeloid cells, monocytes, fibroblasts, and bone marrow– derived precursor cells.⁷ Furthermore, some tumor cells also express integ- rins directly including glioblastoma (GBM).^8-12 In addition, many integrin ligands are abundantly expressed in the tumor microenvironment.^13-15

One of the better detailed actions of integrins is their crucial role in regulating cell motility.¹⁶ Integrin binding to extracellular ligands provides directional traction support for cell movement. Simulta- neously, integrin focal adhesion complexes activate remodeling of the intracellular cytoskeleton to direct cytoplasmic flow. Thus, effective integrin blockade may offer a promising approach to inhibit metastatic capability of cancer cells as well as the inherent infiltrative nature of some tumors, such as GBM.

Integrins also play an important role in mediating angiogenesis.¹⁷ Integrin inhibitors such as cilengitide (EMD 121974, Merck, Darmstadt, Germany) inhibit angiogenesis in the rabbit cornea retina and chicken chorioallantoic membrane models.^18-20 The precise mechanism of their antiangiogenic activity has not been clearly delineated; however, they can block proliferation and induce apoptosis of endothelial cells as well as differentiation of human endothelial precursor cells.^21-23 Of note, a recent study demonstrated that very low levels of RGD-mimetic integrin antagonists may enhance angiogenesis and promote tumor growth.²⁴ Such paradoxical activity has not been observed with the clinical application of integrin antagonists including cilengitide to date, but nonetheless, monitoring carefully for such activity in ongoing and future clinical trials will elucidate whether these laboratory findings translate into meaningful clinical measures.

Insights into the role of integrins in several additional important aspects of tumor biology are emerging. For example, integrin activation contributes to regulating several aspects of the host tumor response including the recruitment of pericytes to stabilize tumor vasculature and the infiltration of myeloid cells, bone marrow–derived precursor cells, fibroblasts, and platelets into the tumor microenvironment.^7,25-27 In addition, integrins are upregulated in some stem cell populations and are physiologically active in these cells,^28-31 including those associated with GBM.³² Furthermore, integrins have also recently been shown to mediate hypoxia response in some tumors including GBM.⁸ The influence of integrins on different facets of tumor biology may vary depending on tumor site. For example, recent studies demonstrate that the activation status of the αvβ3 integrin regulates angiogenesis and cell growth of tumor metastases in the brain but not in the mammary fat pad.³³ In summary, efficient targeting of integrins offers the potential to detrimentally impact several important aspects of tumor biology including intracellular cell signaling, migration, angiogenesis, and the host tumor response. Although integrin inhibitors may have direct antitumor capabilities, their optimal therapeutic potential may lie in combination with cytotoxic, antiangiogenic, or tumor-targeting agents.

Cilengitide Overview

There are 3 main classes of integrin-targeting agents under clinical evaluation including monoclonal antibodies, peptidomimetics, and oral small-molecule inhibitors. Cilengitide, the most advanced integrin inhibitor in clinical development, is a cyclized pentapeptide peptidomimetic designed to compete for the arginine–glycine–aspartic acid (RGD) peptide sequence that regulates integrin-ligand binding. Specifically, cilengitide selectively and potently blocks the ligation of the αvβ3 and αvβ5 integrins to provisional matrix proteins such as vitronectin, fibronectin, fibrinogen, von Willebrand factor, osteopontin, and others.^18,34,35 These integrins play a significant role in tissue remodeling and angiogenesis and contribute to the growth and malignancy of a wide range of cancers.⁷ Previous studies have revealed that α V integrin antagonists not only disrupt tumor angiogenesis but in some cases have a direct impact on the growth and malignant properties of tumor cells themselves.⁴

Preclinical Studies

Initial preclinical studies demonstrated single-agent activity of cilengitide against melanoma³⁶ and orthotopic brain tumor xenografts.^12,37 Subsequent studies revealed that cilengitide can augment the therapeutic benefit associated with cytotoxic agents including chemotherapy and radiation therapy in GBM tumor models.^38-41 More recent preclinical studies demonstrate that the antitumor activity of cilengitide is independent of the activity of the ubiquitous DNA repair enzyme methylguanine methyltransferase (MGMT),⁴² a key mediator of GBM sensitivity to temozolomide.⁴³

Phase I Studies

Cilengitide has been evaluated in a series of phase I and II studies among both recurrent and newly diagnosed GBM patients (Table 1). In an initial phase I study conducted among patients with recurrent solid tumors, no dose-limiting toxicities (DLTs) were observed following administration of cilengitide over 1 hour, twice weekly. In this study, the maximum tolerated dose (MTD) was not established despite increasing the dose from 30 mg/m² to 1,600 mg/m². Although no radiographic responses were observed, 3 patients achieved stable disease.⁴⁴ A second phase I study conducted among patients with advanced solid tumors evaluated cilengitide dosing administered between 120 mg/m² and 2,400 mg/m² on a twice-weekly schedule, and again, no DLTs were observed, and the MTD was not determined.⁴⁵

Table 1.

Clinical Trials Evaluating Cilengitide among Malignant Glioma Patients

Study	Population	Phase	Design	No. of patients	Cilengitide dose (twice weekly)	Efficacy	Reference
NABTT 9911	Recurrent adult MG	I	Dose escalation	51	200-2,400 mg/m2	CR, n = 2; PR, n = 3; SD, n = 16	Nabors et al.46
PBTC-012	Refractory pediatric primary CNS tumors	I	Dose escalation	33	120-2,400 mg/m2	CR, n = 1; SD > 22 weeks, n = 3	MacDonald et al.47
EMD 009	Recurrent GBM	II	Randomized	81	500 mg v. 2,000 mg	PR, n = 7; PFS-6, 10%-15%	Reardon et al.48
NABTC 0302	Recurrent GBM	II	Randomized, perioperative	30	500 mg v. 2,000 mg	PFS-6, 12%	Gilbert et al.61
EMD 010	Newly diagnosed adult GBM	II	Single arm	52	500 mg	Median OS, 16.1 months	Stupp et al.50
NABTT 0306	Newly diagnosed adult GBM	I/II	Randomized; safety run-in	112	500 mg v. 2,000 mg	OS-12, 79.5%	Nabors et al.54
CECIL	Newly diagnosed adult GBM; unmethylated MGMT	II	Randomized	108	2,000 mg	NA	Ongoing
CENTRIC	Newly diagnosed adult GBM; unmethylated MGMT	III	Randomized	504	2,000 mg	NA	Ongoing
CORE	Newly diagnosed adult GBM; unmethylated MGMT	I/II	Randomized	240	2,000 mg 2-5 times per week during XRT; then 2,000 mg twice a week	NA	Ongoing
ABTC-0903	Recurrent GBM	Ib	Single arm	52	2,000 mg	NA	Ongoing

Note: ABTC = Adult Brain Tumor Consortium; CNS = central nervous system; CR = complete response; GBM = glioblastoma; MG = malignant glioma; MGMT = methylguanine methyltransferase; NA = not available; NABTC = North American Brain Tumor Consortium; NABTT = New Agents Brain Tumor Therapy Cooperative Group; OS = overall survival; OS-12 = overall survival rate at 12 months; PBTC = Pediatric Brain Tumor Consortium; PFS-6 = progression-free survival rate at 6 months; PR = partial response; SD = stable disease; XRT = radiation therapy.

Based on a hypothetical concern that patients with primary malignant brain tumors may be more at risk for intracranial hemorrhage or stroke following therapy with agents capable of inhibiting angiogenesis including integrin inhibitors, a separate phase I study was performed in this patient population.⁴⁶ Of note, no instances of intracranial hemorrhage or stroke were observed. Dose-limiting toxicities were observed but were not consistent and varied by type and dose level of therapy. Once again, an MTD was not established. Among 51 patients treated in this study, 2 complete responses (CRs) and 4 partial responses (PRs) were observed. In addition, 6 patients achieved stable disease for over 6 months. Correlative imaging studies suggested that decreased perfusion was associated with higher cilengitide dosing.

In addition, 31 pediatric patients with recurrent primary brain tumors were treated in a separate phase I dose escalation study.⁴⁷ Three patients treated at the highest dose level (2,400 mg/m²) experienced intracranial hemorrhage in this study; however, 2 hemorrhages were asymptomatic (grade 1), and it was unclear whether these events were attributable to cilengitide or to underlying tumor growth. In this study, 1 patient with recurrent GBM achieved a CR, and 2 additional patients had stable disease.

A consistent finding across the dose escalation phase I studies conducted with cilengitide was that cilengitide could be safely administered to solid tumor and central nervous system tumor patients. Consistent DLTs were not observed, nor did any of these studies define an MTD.

Phase II Studies

Following the encouraging results achieved in the phase I studies, a phase II program for cilengitide focused on GBM was initiated with several goals, including 1) to evaluate the antitumor activity of cilengitide among both recurrent and newly diagnosed GBM patients, 2) to gather further safety data associated with this agent in this patient population, and 3) to evaluate intratumoral cilengitide penetration and activity among GBM patients pretreated with cilengitide prior to scheduled surgical resection. In addition, given that an MTD had not been established and that evidence of antitumor activity had been observed across a range of cilengitide doses evaluated in phase I studies, phase II studies also were designed to explore intermediate-low and intermediate-high dose levels.

The first phase II study evaluated adult GBM patients at first recurrence treated with twice-weekly cilengitide administered at either 500 mg or 2,000 mg per dose.⁴⁸ The primary end point was progression-free survival at 6 months (PFS-6). Neither study cohort experienced significant adverse events, and no patients developed intracranial hemorrhage. Linear cilengitide pharmacokinetics were observed in both plasma and cerebrospinal fluid. Although both treatment arms demonstrated evidence of antitumor activity, patients treated at the 2,000-mg dose level achieved superior rates of radiographic response, PFS-6, and overall survival (OS).

A second phase II study among recurrent malignant glioma patients incorporated a design to evaluate intratumoral cilengitide levels in pretreated patients scheduled for tumor debulking.⁴⁹ Patients enrolled in this North American Brain Tumor Consortium study (NABTC-0302) received 3 cilengitide doses administered at either 500 mg or 2,000 mg prior to surgical resection, and cilengitide dosing was resumed following postoperative recovery. No significant toxicities were observed among the 30 patients treated in this study, and specifically, no episodes of perioperative bleeding or wound dehiscence occurred. Despite a 24-hour interval between the third cilengitide dose and tumor harvesting, average intratumoral cilengitide concentrations were 400 and 1,190 ng/g for patients treated with the 500-mg and 2,000-mg cilengitide doses, respectively, and the tumor:plasma ratio was 1.83:4.17.

Given the modest evidence of antitumor activity achieved by single-agent cilengitide in phase II studies conducted in recurrent GBM patients, a single-arm phase II study was performed primarily in Europe among newly diagnosed GBM patients.⁵⁰ Further rationale for this study was derived from preclinical data, demonstrating that cilengitide can augment the antitumor activity of radiation therapy (XRT) in an orthotopic GBM model.³⁸ The phase II study evaluated the addition of 500 mg of cilengitide administered twice weekly during the 6-week period of radiation/daily temozolomide followed by 6 cycles of adjuvant, postradiation temozolomide chemotherapy. Median follow-up of the 52 patients enrolled in this study was 34 months, and no significant toxicity attributable to the addition of cilengitide was observed. Evaluation of the metabolism of both cilengitide and temozolomide revealed that pharmacokinetic findings for each drug were consistent with historical data, suggesting that neither agent impacted the metabolism of the other. Overall outcome compared favorably to historical results. Specifically, PFS-6 rate was 69%, and median OS rate was 16.1 months. Of note, 45 patients (87%) in this study had tumor material available for analysis of the DNA repair protein MGMT, which is upregulated based on the lack of promoter methylation in approximately 60% of newly diag- nosed GBM patients.⁵¹ MGMT promoter methylation was detected in 23 (51%) of these patients. Interestingly, patients with MGMT-methylated tumors achieved a longer median PFS rate (P < 0.001) and median OS rate (P = 0.022) with the addition of cilengitide compared to patients with unmethylated MGMT tumors. Importantly, patients with methylated MGMT who received cilengitide had a better outcome than historical MGMT-methylated patients treated with standard temozolomide chemoradiotherapy.^52,53 Thus, this phase II study suggested that although the addition of cilengitide modestly improves the outcome for newly diagnosed GBM patients undergoing treatment with temozolomide chemoradiotherapy, patients with MGMT- methylated tumors benefit preferentially. The precise mechanism associated with this beneficial effect has not been elucidated. However, one hypothetical possibility is that the antiangiogenic effect of cilengitide may lead to improved temozolomide delivery, which in turn could contribute to greater cytotoxicity among tumors with enhanced temozolomide sensitivity due to low MGMT expression. Nonetheless, 2 important factors should be considered when interpreting the results of this study. First, all patients in this study received 500 mg of cilengitide and thus were treated with a relatively low cilengitide dose. Second, adjuvant temozolomide was discontinued after 6 cycles, and nearly all patients also discontinued cilengitide dosing at this time. It is unclear whether a higher cilengitide dose or a prolonged cilengitide dosing schedule beyond 6 months may have led to greater antitumor activity.

A second parallel single-arm phase II study among newly diagnosed GBM patients was conducted by the New Agents in Brain Tumor Therapy (NABTT) cooperative group.⁵⁴ Following a safety lead-in that demonstrated that twice-weekly cilengitide doses up to 2,000 mg were well tolerated during XRT/temozolomide, patients were randomized to receive either 500 mg or 2,000 mg of cilengitide during XRT/temozolomide and then during monthly cycles of temozolomide after XRT. Importantly, patients were allowed to continue both monthly cycles of temozolomide as well as cilengitide for more than 6 months following the completion of XRT. Recently reported preliminary results revealed a median OS rate of 21.9 months for all 112 patients treated in this study, and 82% were alive at 1 year. These findings also compare favorably to historical data. Breakdown of outcome based on cilengitide dose (500 mg v. 2,000 mg) and by MGMT expression determined by immunohistochemistry will be forthcoming.

Two additional important studies have recently completed accrual among newly diagnosed GBM patients. The first study (CENTRIC; clinicaltrials.gov identifier: NCT00689221) is a registration, randomized phase III study of cilengitide for newly diagnosed GBM patients with MGMT promoter–methylated tumors. The primary end point of this multinational study is a median OS rate with stratification for geographic region and Radiation Therapy Oncology Group recursive partitioning analysis classification.⁵⁵ Secondary end points include PFS and safety.

The second study (CORE; clinicaltrials.gov identifier: NCT00813943) that recently completed accrual was a phase I/II dose escalation study of cilengitide plus standard temozolomide chemoradiotherapy among newly diagnosed GBM patients with tumors that are unmethylated at the MGMT promoter. All patients received 2,000 mg of cilengitide, but dosing increased from twice weekly to 3 times per week and then 5 days per week during XRT only among successive cohorts of patients enrolled to the phase I component of the study. Of note, no significant DLTs were observed during the phase I aspect of the trial, including among patients who received 2,000 mg of cilengitide up to 5 times per week during XRT. The phase II aspect of this study randomized patients to receive standard temozolomide chemoradiotherapy alone versus the same regimen with cilengitide administered at 2,000 mg either twice weekly or 5 times per week during XRT/temozolomide. Of note, all patients in the latter 2 arms continue to receive 2,000 mg of cilengitide twice weekly during subsequent monthly cycles of adjuvant temozolomide.

Other Indications

Although GBM is the lead indication for cilengitide clinical development, additional studies in head and neck cancer as well as lung cancer are underway. EMD 121974-013 (clinicaltrials.gov identifier: NCT00842712) is an open-label, randomized phase I/II study designed to evaluate the safety and efficacy of cilengitide in combination with cisplatin, 5-fluorouracil, and cetuximab for patients with recurrent/metastatic small cell head and neck cancer. EMD 121974-014 (clinicaltrials.gov identifier: NCT00705016) is an open-label, randomized, controlled multicenter phase II study designed to evaluate the efficacy and safety of cilengitide in combination with cisplatin and either vinorelbine or gemcitabine as first-line therapy for patients with advanced non–small cell lung cancer.

Future Developments

Two critical factors will likely prove to be major influences on the ultimate success of integrin-targeting agents for cancer patients. First, correlative biomarkers predictive of benefit from an integrin inhibitor are critically needed. Such biomarkers may derive from an analysis of archival tumor material, imaging studies, and/or circulating factors. First, such markers may identify a priori patients who are most likely to respond to integrin antagonists, or conversely, those who are least likely to benefit. For example, one could postulate that patients with higher levels of a specific integrin or its respective ligand may be most likely to benefit from an inhibitor that specifically targets a given integrin. Subsequent clinical trials of a specific integrin antagonist could be designed to evaluate outcome among a patient subpopulation based on increased target integrin expression. Expression of integrin target molecules may be evaluated directly from tumor samples or may also be assessed via imaging approaches. For example, ¹⁸F-labeled glycosylated Arg-Gly-Asp (RGD) peptide has been shown to detect αvβ3 integrin expression using positron emission tomography in suspected or recurrent GBM tumors obtained from patients undergoing surgical debulking.⁵⁶ A similar approach using [64Cu]DOTA-cyclo-RGD predicted αvβ3 integrin–mediated migration and proliferation of U87MG GBM tumors undergoing treatment with the Src inhibitor dasatinib.⁵⁷ In similar experiments, superparamagnetic polymeric micelles encoded with cyclic ligand (RGDfK) were able to detect αvβ3 integrin–expressing tumors and tumor microvasculature in murine xenografts.⁵⁸

A second critical factor affecting the overall development of integrin inhibitors in the oncology arena will be their successful pairing with other antitumor agents. Preclinical data are emerging confirming enhanced antitumor activity when integrin-targeting agents are combined with traditional cytotoxic agents. However, a recent publication highlights that such an effect with XRT may incorporate important schedule dependence.³⁸ Based on the myriad activities regulated by integrins in tumors, multiple novel classes of cancer therapeutics may warrant evaluation in combination with integrin- targeting agents. For example, based on their antiangiogenic activity, the combination of integrin antagonists with direct VEGF or VEGFR inhibitors may provide greater antiangiogenic effects, which in turn could lead to enhanced antitumor efficacy. In addition, antiangiogenic agents can increase invasion and migration of cancer cells.^59,60 Based on the role of integrins regulating cell migration and invasion, integrin inhibitors may be able to diminish this adaptive response following antiangiogenic therapy. For these reasons, studies are planned or ongoing to combine cilengitide with either the VEGFR2 inhibitor cediranib (clinicaltrials.gov identifier: NCT00979862) or bevacizumab, a humanized monoclonal antibody against VEGF. A substantial attribute of the integrin-targeting agents evaluated to date including cilengitide is their lack of consistent toxicity or adverse events. This important attribute may also likely be an important factor as novel combinatorial regimens incorporating integrin-targeting agents are developed for clinical application.

Conclusion

Integrins are highly attractive therapeutic targets in oncology because of their key role in tumor cell migration, angiogenesis, and dysregulated cell signaling, coupled with an emerging role in modulating the host tumor response. Although clinical experience with integrin antagonists to date confirms their excellent safety and tolerability, translation of successful integrin targeting into meaningful clinical benefit for cancer patients has nonetheless proven to be a difficult challenge. Cilengitide, an RGD-mimetic cyclicized pentapeptide, is the most advanced integrin inhibitor in oncology development. Preclinical and early phase I and II studies demonstrated modest proof-of-concept antitumor activity, and a phase III registration study for newly diagnosed GBM patients recently completed accrual, while phase II studies in other solid tumor indications are ongoing. Further optimization of the antitumor benefit of cilengitide and other integrin inhibitors will likely include the identification of predictive biomarkers of antitumor activity as well as their integration into multimodality therapeutic regimens.