On the TRAIL Toward Death Receptor–Based Cancer Therapeutics

Abstract

Identifying targeted agents capable of selectively inducing death of cancer cells, while sparing normal host cells, has been the ultimate goal of cancer therapeutic drug development. In the late 1980s, monoclonal antibodies (mAbs) that were found to have antitumor activity via induction of tumor cell apoptosis were developed by several laboratories. These mAbs were found to ligate a tumor necrosis factor (TNF) receptor family member named Fas (APO-1/CD95), and investigation into the biochemical mechanism by which anti-Fas Abs induce cell death has revealed incredible detail regarding the apoptotic machinery. Unfortunately, agonistic antimurine Fas mAbs, when tested in preclinical models in vivo, were extremely toxic, because of rapid induction of hepatocyte apoptosis. In addition, a critical role for Fas/FasL interactions in the regulation of T-cell homeostasis was uncovered through genetic analyses of the autoimmune-prone mouse strains Lpr and Gld. It was, therefore, not possible to move nonspecific Fas-based therapeutics into the clinical arena, although engagement of Fas on tumor cells by Fas ligand expressed by specific CD8 T cells can promote tumor cell death in an antigen-specific fashion, and alterations in Fas signaling pathways in cancer have been reported to contribute to tumor cell resistance to immunemediated destruction. These observations support the notion that Fas engagement still may be an important component of antitumor immune effector mechanisms in patients. An additional TNF family member that also could promote apoptosis of tumor cells, called TNF-related apoptosis-inducing ligand (TRAIL), was subsequently identified. TRAIL can interact with five distinct receptors, two of which are death receptors (TRAIL-R1/DR4 and TRAIL-R2/DR5). Ligation of these receptors initiates a caspase-driven apoptotic pathway that is similar to what is induced through Fas. In contrast to FasL-deficient mice, TRAIL-deficient mice generated by gene targeting do not show global perturbation in T-cell development and homeostasis. Preclinical in vivo experiments of tumor growth in TRAIL-deficient mice (or in mice with endogenous TRAIL blocked) indicated a critical role in immune surveillance, because increased susceptibility to tumor formation— both spontaneously and in response to carcinogens—was observed. Constitutive expression of TRAIL receptors has been observed in a wide variety of human tumor types. These observations thus provided a sound rationale for the development of therapeutic agents that engage or mimic the TRAIL pathway on cancer cells. Several approaches have been explored to target TRAIL receptors therapeutically. One avenue of research has been the use of trimeric TRAIL itself as a recombinant natural ligand. Preclinical mouse models have shown effective antitumor activity in vivo, without significant toxicity. However, clinical development was stalled when hepatotoxicity was observed with some TRAIL preparations. It was subsequently determined that hepatocyte apoptosis occurred only with polyhistidine-tagged versions of recombinant TRAIL, a manufacturing artifact that could be avoided, which then reopened the door toward clinical application. In addition to TRAIL itself, an alternative therapeutic strategy is based on agonistic mAbs that specifically target the TRAIL receptors DR4 or DR5. The existence of decoy receptors that can bind TRAIL, yet not deliver a death signal, suggests a potential disadvantage of using the recombinant ligand and indicates a theoretical advantage to the use of specific anti-DR4 or -DR5 mAbs. All of these strategies are currently in clinical testing. In this issue of the Journal of Clinical Oncology, Tolcher et al report a phase I clinical trial of a fully human mAb against human TRAIL-R1/DR4. A careful dose escalation was performed, in which patients first received a single dose of mAb, which was an important step for safety concerns, given the potential for hepatotoxicity. The dose and frequency of administration were increased to 10 mg/kg every 14 days. A terminal half-life (t1/2 ) of 18 days was observed, suggesting that dosing every 2 to 3 weeks is a reasonable schedule. The concentration of anti-TRAIL-R1 mAb that is effective for tumor apoptosis in vitro is approximately 1 g/mL, and all dose levels above 1 mg/kg achieved this concentration as a trough level in this phase I study. Importantly, no grade 3 elevations in transaminases and no bilirubin elevation clearly attributable to the drug were observed, although several patients did show grade 1 to 2 transaminase elevation, and two patients with pre-existing liver metastases with elevated transaminase levels at baseline did increase to grade 3. Therefore, although the overwhelming hepatotoxicity previously observed with anti-Fas mAbs in murine studies did not occur, these detectable effects on the liver from this trial warrant caution in moving forward in combination with other anticancer agents having hepatic effects, or in the context of patients receiving hepatotoxic drugs for other indications. There were no global effects on lymphocyte subsets seen in this study. However, recent work has indicated a clear role for the activity of TRAIL in determining whether CD4 T-cell help for JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 25 NUMBER 11 APRIL 1