Combined modality therapy with TRAIL or agonistic death receptor antibodies

Abstract

Molecularly targeted therapies, such as antibodies and small molecule inhibitors have emerged as an important breakthrough in the treatment of many human cancers. One targeted therapy under development is tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) due to its ability to induce apoptosis in a variety of human cancer cell lines and xenografts, while lacking toxicity in most normal cells. TRAIL and apoptosis-inducing agonistic antibodies to the TRAIL death receptors have been the subject of many preclinical and clinical studies in the past decade. However, the sensitivity of individual cancer cell lines of a particular tumor type to these agents varies from highly sensitive to resistant. Various chemotherapy agents have been shown to enhance the apoptosis-inducing capacity of TRAIL receptor-targeted therapies and induce sensitization of TRAIL-resistant cells. This review provides an overview of the mechanisms associated with chemotherapy enhancement of TRAIL receptor-targeted therapies including modulation of the apoptotic (death receptor expression, FLIP, and Bcl-2 or inhibitors of apoptosis (IAP) families) as well as cell signaling (NFκB, Akt, p53) pathways. These mechanisms will be important in establishing effective combinations to pursue clinically and in determining relevant targets for future cancer therapies.

Introduction

Cancer occurs in various tissues of the body with uncontrolled growth and metastasis formation. Depending on the site and type of cancer, treatment will consist of surgical resection, chemotherapy and radiation therapy. The development of molecularly targeted therapies consisting of antibodies and small molecule inhibitors has revolutionized cancer therapy with selective agents that provide favorable and non-overlapping toxicity profiles. Since its discovery in 1995, tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) or Apo2 ligand has been investigated as a cancer therapeutic agent. TRAIL induces apoptosis in many human tumor cell lines and tumor xenografts, but not in normal cells.1–4 It has been widely reported that tumor cell killing is increased by combination treatment with drugs. Different classes of drugs sensitize cancer cells to TRAIL and TRAIL receptor agonists by a variety of cellular mechanisms. This review will provide an update on optimizing TRAIL or TRAIL antibody agonists as cancer therapeutics alone and in combination with current clinically used drugs and discuss the cellular mechanisms of enhanced efficacy.

TRAIL and Receptors

TRAIL is a member of the tumor necrosis factor (TNF) superfamily, which currently comprises nineteen type II transmembrane proteins with an intracellular N-terminus. TRAIL contains a conserved TNF homology domain at its C-terminus and is associated with immune system function and homeostasis, similar to many other family members.5 TRAIL exists naturally on the surface of immune cells capable of inducing apoptosis (i.e., natural killer cells) or may be proteolytically cleaved to release the extracellular domain.3,4,6 Cellular and soluble TRAIL form a homotrimer stabilized by a zinc atom and bind to receptors, inducing stable receptor trimers.3,4,7 Six members of the TNF receptor (TNFR) superfamily form a subset known as death receptors (DR), which are characterized by an intracellular death domain.8 TNFR1, which binds to TNF, and Fas/CD95, which binds to Fas ligand, have been examined for their role in immune system function and induction of apoptosis.8,9 Death receptor 4 (DR4, TRAIL-R1) and death receptor 5 (DR5, TRAIL-R2) have been identified to bind with TRAIL. DR4 and DR5 have the capacity to induce apoptotic signaling after TRAIL ligand binding and are the targets of developing cancer therapies. Three additional members of the TNFR superfamily have been identified that bind to TRAIL (Fig. 1).10 Decoy receptor 1 (DcR1, TRAIL-R3) and decoy receptor 2 (DcR2, TRAIL-R4) bind TRAIL but fail to elicit an apoptotic response. A fifth soluble receptor, osteoprotegerin (OPG), also fails to mediate apoptosis.

DR4 was first identified11 via sequence homology to the TNFR-1 death domain (DD), a characteristic motif amongst the apoptotic-inducing members of the TNFR superfamily. DR5 was identified by a similar approach.11–15 These receptors are type I transmembrane proteins with two cysteine-rich domains extracellularly and an intracellular death domain, which acts as a site for protein–protein interactions involved in the apoptotic signaling cascade. Overexpression of apoptosis-inducing death receptors, DR4 and DR5, can induce ligand-independent apoptosis via receptor homo- or hetero-oligomerization.12–16 The first decoy receptor, DcR1, has two cysteine-rich extracellular domains and a putative hydrophobic region, but lacks an intracellular domain and instead has a glycosyl-phosphoinositol membrane anchor.12–14,16 This is consistent with the lack of apoptotic signaling and TRAIL induced cytotoxicity in cells overexpressing DcR1. The second decoy receptor, DcR2, has two cysteine-rich extracellular domains and a hydrophobic transmembrane region, but only a partial intracellular DD.17,18 The truncated intracellular domain lacks the ability to induce apoptosis, but has been shown to induce nuclear factor-kappaB (NFκB) activation when the receptor is overexpressed in some systems,17 but not in others.18,19 DcR2 may also produce anti-apoptotic signaling by activation of NFκB.17 The binding of TRAIL to DcR1 and DcR2 may decrease the amount bound to death receptors.20 The fifth receptor, OPG, is a soluble protein first identified by binding to RANKL/TRANCE, but later found to also bind TRAIL.21 Unlike the other receptors, OPG has four cysteine-rich domains but is a soluble receptor lacking transmembrane and cytoplasmic regions. The C-terminal region of OPG has two homologous DD and a heparin binding domain.22 OPG displays the weakest affinity for TRAIL of the five receptors at physiologic temperatures,23 and its relevance is unclear.

The physiologic role of TRAIL has not been fully elucidated; however some insight has been gained. TRAIL may be important in natural killer cell function, virus and tumor cell immune surveillance, autoimmune disease development and airway remodeling and inflammation. TRAIL expression has been shown to be induced by interferon (IFN) in neutrophils,24 natural killer cells25 and monocytes,26 which may be important in TRAIL-mediated modulation of the immune system.

Mechanisms of Apoptosis by TRAIL Binding to DR4/DR5

TRAIL-induced apoptosis commences with the activation of DR4 or DR5 by ligand binding and receptor trimerization to stimulate the extrinsic and intrinsic apoptosis pathways (Fig. 2). The apoptotic cascade is initiated by the assembly of a death-inducing signaling complex (DISC) with the recruitment of Fas-associating protein with death domain (FADD), an adaptor protein between the death receptor and initiator caspases-8 or 10. The DD of trimerized receptors interacts with a homologous domain within FADD, by which caspase-8 is then recruited via interactions between death effector domains (DED).15,27–32 Caspase-8 is cleaved through autocatalytic processing to produce active subunits.33 The p55 and p52 pro-caspases are cleaved into p43, p41 and p12 fragments. Active p18 and p10 are formed in a second cleavage stage.34 The activity of caspase-8 may also be positively or negatively regulated by ubiquitinated as summarized by Gonzalvez and Ashkenazi.8 In the extrinsic apoptotic pathway, the active caspase-8 subunits interact directly with downstream effector caspases, such as capase-3 or 7, to cleave and activate them. Caspase-3 is then able to cleave many downstream substrates, such as poly (ADP-ribose) polymerase (PARP) and DNA fragmentation factor (DFF), to initiate apoptosis.34

In some tumor cell lines, TRAIL activates the intrinsic apoptotic pathway, which occurs when active caspase-8 cleaves Bid (Bcl-2 inhibitory BH3-domain-containing protein), a Bcl-2 family member. Truncated Bid migrates to the mitochondrial membrane where it stimulates the oligomerization of Bak and Bax. Upon activation, Bax undergoes a conformational change and translocates to the mitochondrial membrane where homooligomers form. Bak exists as an outer mitochondrial membrane protein and forms homo-dimers, trimers and tetramers following activation.35 Next, permeabilization of the outer mitochondrial membrane occurs, allowing release of mitochondrial proteins, including cytochrome c and Smac/DIABLO (second mitochondria-derived activator of caspase/direct inhibitor of apoptosis binding protein with low pI). In the cytosol, Smac/DIABLO interacts with X-linked inhibitor of apoptosis (XIAP) to release caspase-9 and caspase-3 from XIAP inhibition.34 Cytochrome c binds with Apaf-1, dATP and caspase-9 to form the apoptosome where caspase-9 is activated. Active caspase-9 cleaves caspase-3, which then cleaves a variety of substrates to initiate apoptosis.34,36 Crosstalk has been shown to exist between the extrinsic and intrinsic apoptotic pathways, suggesting TRAIL may activate both pathways.

TRAIL and Agonistic Antibodies to TRAIL Receptors as Cancer Therapeutics

TRAIL is promising as a cancer therapeutic agent showing efficacy against tumor cells without the toxicities to normal cells associated with other TNF family members. TNF and Fas ligand both induce cytotoxicity against tumor cells, but in murine models TNF induces a lethal inflammatory response and Fas ligand results in severe hepatotoxicity.37 Early reports indicated certain preparations of recombinant TRAIL also produced hepatotoxicity in vitro.38 A different recombinant form of TRAIL lacking sequence modifications to amino acids 114–281 and with the addition of a modified leucine zipper produced tumor cell apoptotic activity in vitro and tumor growth inhibition in vivo without hepatotoxicity.1,39 Nonhuman primate studies did not reveal any organ or systemic toxicities despite binding to primate receptors with an affinity similar to the human receptor. High doses of TRAIL have been administered and well tolerated in nude mice, rats, cynomolgus monkeys and chimpanzees, but display rapid whole body clearance and short plasma half-lives (3–5 minutes in rodents and 24–31 minutes in non-human primates).1 The relevance of the short half-life to efficacy is still to be determined in clinical trials, which are currently underway. In Phase I studies, no dose-limiting toxicities have been reported, and out of 32 patients, 17 had stable disease and there was one patient with a partial response.40

TRAIL has shown variable cytotoxic activity against a broad spectrum of human tumor cell lines, including breast, colon, lung, pancreatic, prostate, renal and thyroid carcinoma, glioma, multiple myeloma and leukemia.41 However, certain cell lines or tumor types exhibit TRAIL resistance. Many TRAIL and chemotherapy combinations act synergistically against a variety of tumor cell lines and can reverse resistance to either agent (Table 1).37 Most of the current clinically used chemotherapy agents have been shown to enhance TRAIL-mediated apoptosis, including cisplatin, doxorubicin, 5-fluorouracil (5-FU) and camptothecin (CPT-11).42 To demonstrate different classes of drugs are capable of producing increased cytotoxicity against non-small cell lung carcinoma cells in combination with TRAIL receptor-targeted therapies, we evaluated TRA-8 cytotoxicity in combination with various chemotherapy agents. Figure 3 shows the activity of doxorubicin, bortezomib and docetaxel in combination with TRA-8 against the A549 lung cancer cell line. These results indicate that each of these chemotherapy agents is capable of sensitizing cells to TRA-8 in a synergistic manner. All three drugs interacted with TRA-8 in a significantly synergistic manner (combination index <1). Doxorubicin is classified as a topoisomerase II inhibitor, docetaxel as a microtubule stabilizer and bortezomib as a proteasome inhibitor, yet each interacts with TRA-8 in the A549 lung cancer cells. As will be described later in greater detail, this may occur through modulation of the intracellular regulatory components of the apoptotic cascade and other cell signaling pathways. Table 1 provides a summary of chemotherapy agents reported to enhance TRAIL or death receptor antibody efficacy and the apoptotic regulatory proteins the combinations modulate.

Tumor cell resistance to TRAIL-induced apoptosis may be due to the expression of decoy receptors on the cell surface. For this reason, agonistic antibodies may have greater therapeutic potential due to specific targeting of the death receptors without decoy receptor binding, in addition to a longer plasma half-life.42 There has been an immense effort both in academia and the pharmaceutical industry to develop antibodies to TRAIL death receptors. Notable examples currently in clinical trial include: Humanized TRA-8 anti-DR5 (Tigatuzumab, CS-1008) from Daiichi-Sankyo;43–45 fully human antibodies against DR4 (Mapatumumab, HGS-ETR1, TRM-1) or DR5 (Lexatumumab, HGS-ETR2) from Human Genome Sciences; human anti-DR5 (Conatumumab, AMG 655) from Amgen;45,46 and human anti-DR5 antibody (4H6, Apomab) from Genentech Inc.42

TRA-8, a murine antibody to DR5, produced significant tumor growth inhibition of 2LMP breast cancer xenografts and TRA-8 combined with doxorubicin or paclitaxel produced greater tumor inhibition than any agent alone.47 The interaction between doxorubicin and TRA-8 was shown to be synergistic in vivo and was further enhanced by the addition of ⁶⁰Co radiation therapy. TRA-8 was shown to activate apoptotic pathways and its efficacy was enhanced by doxorubicin similar to what has been observed with TRAIL (Fig. 4). Combination treatment of breast cancer cells with TRAIL or TRA-8 and doxorubicin resulted in activation of caspases, cleavage of Bid and PARP (Fig. 5). Also, there was a reduction in XIAP levels to a varying degree in different cell lines.48 Efficacy of TRA-8 has been observed against breast, cervical, ovarian, pancreatic, glioma and colon cancer cell lines in vitro and in vivo in tumor xenograft models, which was enhanced by combination treatment with chemotherapy drugs.42,47,49–54 In an ex vivo assay of primary ovarian cancer, 79% (15 of 19) of patient tumor specimens demonstrated sensitivity to TRA-8 treatment in a dose-dependent manner associated with the induction of apoptosis.50 A Phase I trial with a humanized version of TRA-8 (Tigatuzumab, CS-1008) has been completed without any dose limiting toxicity and 7 of 17 patients had stable disease.44

Apomab, an additional agonistic DR5 antibody in development, was shown in combination with chemotherapy (paclitaxel and carboplatin) to significantly inhibit tumor growth and prolong survival in mice with orthotopic NCI-H460 lung tumor xenografts.55 In preclinical studies, treatment with mapatumumab, an agonistic antibody to DR4, inhibited the growth of colon, non-small cell lung and renal tumor xenografts in vivo and was shown to induce activation of caspases-3, 8 and 9 in vitro. When combined with 5-FU, CPT-11 or topotecan, mapatumumab produced greater anti-tumor efficacy against colon carcinoma xenografts than any agent alone.56 Mapatumumab has been shown to have a terminal plasma half-life of 6.9–8.7 days in mice. Mapatumumab and lexatumumab, an antibody against DR5, were shown individually to inhibit COLO205 colon cancer xenograft growth in vivo, whereas lexatumumab demonstrated greater growth inhibition with more tumor regressions.57 Mapatumumab and lexatumumab also showed apoptotic activity against 67 and 70% of 27 primary lymphoma samples, respectively.58

Phase I clinical trials have shown mapatumumab and lexatumumab antibodies to be well tolerated with grade 3 toxicity in a small number of patients.59,60 Mapatumumab Phase I clinical trials established that the antibody can be administered safely with no significant hematologic toxicity. Two out of eleven patients had grade 3 elevations of liver function tests, although each had elevated transaminases at baseline. Antibody plasma concentrations comparable to efficacious concentrations in preclinical mouse models were attainable with 10 mg/kg dosing in humans with trough concentrations greater than 1 µg/mL.60 A Phase II trial of mapatumumab in advanced non-small cell lung cancer patients who had received prior chemotherapy demonstrated 10 mg/kg was well tolerated, but no patients responded. Nine of 32 patients had stable disease for a minimum of 4 weeks.61 However, a recent Phase II trial reported no improvement in response rate or progression-free survival with the addition of mapatumumab to paclitaxel and carboplatin in non-small cell lung cancer patients.62 Another Phase II trial in patients with non-Hodgkin's lymphoma reported one complete response, two partial responses and 12 patients had stable disease. Two serious adverse events (shingles or fever) were reported and may have been related to treatment.63 The investigators concluded that higher doses of mapatumumab and future trials with combination chemotherapy are warranted.61 In Phase I trials, lexatumumab was also well tolerated and 12 of 37 patients had stable disease. A maximum tolerated dose of 10 mg/kg was determined as dose limiting toxicities occurred in 3 of 7 patients treated with 20 mg/kg.59 Additional Phase I trials have been reported and Phase II trials are planned. Important to note is that the majority of the patients in the Phase I trials have previously failed treatment and had disease progression on chemotherapy regimens. Therefore, stable disease and a small percentage of patients with partial and complete responses is promising.

The potential of TRAIL targeted therapies lies in their ability to enhance the tumor cytotoxicity of existing chemotherapy or antibody regimens. TRAIL has been combined with rituximab for the treatment of non-Hodgkin's lymphoma;63,64 mapatumumab was used in combination with gemcitabine and cisplatin;65 and lexatumumab was used in combination with gemcitabine, pemetrexed, doxorubicin or FOLFIRI (a combination of leucovorin, 5-FU and irinotecan).66 In each of these trials, preliminary reports suggest that each agent can be safely administered to patients in combination with chemotherapy or antibody regimens. Each of these examples mapatumumab and lexatumumab demonstrated the clinical applicability and promise of TRAIL receptor agonistic antibodies in the treatment of human cancer. As Phase II clinical trials of these targeted therapies combined with chemotherapy continue and are reported, the clinical utility of these therapies will become more apparent.

Determinants of Sensitivity

As described above, TRAIL and agonistic antibodies to the TRAIL death receptors have apoptosis-inducing activity against a variety of human cancer cell types both in vitro and in vivo. However, approximately one-third of human tumor cells are resistant to TRAIL treatment and an additional one-third have only a moderate response.42 Resistance can occur at various points in the apoptotic pathway or in other cellular signaling pathways. A variety of apoptosis regulatory molecules, including death and decoy receptors, FLIP, XIAP and Bcl-X_L and signaling pathways, including NFκB and Akt, have been associated with modulating resistance (Table 2).67 The mechanism of resistance may be a delicate balance between levels of pro- and anti-apoptotic molecules within the cells. It is likely that synergistic effects between drugs and TRAIL or death receptor antibody agonists are achieved by modulation of one or more of these apoptotic regulatory proteins or signaling pathways. A better understanding of these mechanisms will aid in the development of cancer therapeutics with combination therapies to tip the balance towards apoptosis.

Receptor expression.

TRAIL and its receptors are expressed in a variety of tissues, unlike other TNF superfamily members that show more specific expression patterns. For example, Fas ligand is primarily found in stimulated T cells.11 TRAIL is expressed throughout various parts of the adult human body, including spleen, prostate, thymus, ovary, small intestine, colon, peripheral blood leukocytes, heart, lung, skeletal muscle and kidney.3 The wide expression of TRAIL suggests it is nontoxic to normal cells. Researchers originally hypothesized that relative expression of death and decoy receptors would predict sensitivity of cells to TRAIL.12,17 However, results showed that in many instances basal receptor expression and TRAIL sensitivity did not correlate. No relation of sensitivity and DR4, DR5 and DcR1 expression was shown amongst eleven breast cancer lines68 or in Jurkat leukemia cells.69 A lack of correlation was also reported in multiple pancreatic cancer cell lines70,71 and six hepatocellular carcinoma cell lines.72

However, some chemotherapy agents and radiation have been shown to increase the expression of DR4 and/or DR5, and along with other factors may contribute to TRAIL sensitization. For example, doxorubicin and etoposide have been shown to upregulate levels of DR4 and DR5 and synergize with TRAIL.73,74 DNA damaging chemotherapy agents, including doxorubicin and etoposide, and radiation induce DR5 gene expression via a p53-dependent mechanism.75–79 Takimoto and El-Deiry80 and Liu et al.81 identified intronic p53 binding sites within the DR5 and DR4 genes, respectively. In addition, NFκB has been shown to have binding sites within the DR4 promoter region and intron 1 of the DR5 gene.82,83 Regulation of NFκB by overexpression of active NFκB subunits (cRel, p65) or by etoposide have been shown to increase expression of both receptors.74,83,84 Along with activation of the DR5 promoter, DR5 expression may be subject to transcriptional repression by Yin Yang 1 (YY1) for which a binding site has been suggested within the DR5 promoter.85 Baritaki et al.85 reported that treatment of PC-3 prostate cancer cells with cisplatin, etoposide, doxorubicin or vincristine increased DR5 expression, decreased YY1 expression and sensitized cells to TRAIL-induced apoptosis. A reduction in YY1 levels by siRNA also increased DR5 expression and TRAIL-induced apoptosis. The reduction in YY1 and subsequent increases in DR5 by etoposide were correlated to a decrease in NFκB activity. Later studies showed that a proteasome inhibitor NPI-0052 and a nitric oxide donor DETANONOate sensitized tumor cells to TRAIL-induced with a similar reduction in NFκB activity, decreased YY1 and increased DR5 expression.86,87 Another molecule proposed to regulate the transcription of DR5 is Sp1. A putative binding site within the DR5 promoter for transcription factor Sp1 was identified by Yoshida et al.82,88 Histone deacetylase inhibitors (sodium butyrate in HCT116 colon cancer cells and sodium butyrate and trichostatin A in 786-O renal cell carcinoma cells) were shown to increase the mRNA and protein levels of DR5, which correlated with an increase in apoptosis and caspase activity. Further analysis using mutations within the Sp1 binding sites demonstrated Sp-1 was involved in the increased DR5 expression.89,90 These studies demonstrate the variety of mechanisms and chemotherapeutic agents that can modulate death receptor expression and subsequently sensitize cells to death receptor-modulated apoptosis.

Another means of modulating DR5 expression on the surface of tumor cells by chemotherapy agents is by upregulating ceramide to form ceramide-rich membrane rafts to cluster DR5 and enhance DISC formation.91 Therefore, basal death receptor expression may not predict sensitivity to TRAIL-targeted therapies, but increased death receptor expression on cancer cells by chemotherapy may play a role in sensitization. Another important concept in TRAIL death receptor function is internalization following ligand binding.8 DR4 and DR5 have been shown to undergo dynamin-dependent clathrin-mediated endocytosis upon TRAIL binding, but blockade of internalization by dominant negative dynamin enhanced TRAIL-induced apoptosis.92 Other mechanisms of receptor internalization also exist and the overall impact on TRAIL activity remains unknown. Post-translational modifications of the death receptors have also been associated with TRAIL sensitivity.8,93 Ashkenazi and colleagues found that expression of O-glycosyltransferase GALNT14 mRNA correlated with TRAIL sensitivity of 119 human cancer cell lines using genome-wide profiling. O-glycosylation of DR4 and DR5 promoted clustering of death receptors and DISC formation. When O-glycosylation was inhibited, death receptor complexing and caspase-8 association within the DISC were reduced.93 This post-translational modification of death receptors and correlation to sensitivity may provide a useful biomarker for response in future clinical trials.

Another point to consider in regards to death receptor-targeted therapy is the relative contribution of each receptor to the induction of apoptosis. In certain types of cancer it has been reported either DR4 or DR5 is predominantly responsible for the apoptotic response. Kelley et al.94 used TRAIL variants, which preferentially bind to either DR4 or DR5, to show a greater contribution of DR5 to induction of apoptosis in Colo205 and Colo320 colon cancer cells and MDA-MB-231 breast cancer cells. In a panel of 12 glioma cell lines, DR5 binding antibodies produced cytotoxicity against 8 cell lines, while all were resistant to a DR4 antibody.95 Conversely, mapatumumab (a DR4 antibody) produced greater cytotoxicity than lexatumumab (a DR5 antibody) in 9 of 13 pancreatic cancer cell lines.96 In these studies, mapatumumab produced synergistic cytotoxicity in combination with XIAP inhibitors, while less combination effect was seen with lexatumumab. Primary pancreatic carcinoma cells also were more sensitive to maptumumab.96 Additional studies have highlighted the dominance of DR4-mediated apoptosis in pancreatic cancer and chronic lymphocytic leukemia.97,98 The relevance of these studies to the clinical application of TRAIL receptor-targeted therapies remains to be determined.

FLIP.

FLIP is structurally related to caspase-8 and multiple splice mRNA variants are produced, but the cellular short form (cFLIP_S) and long form (cFLIP_L) are most commonly detected with each having two DED domains similar to those within FADD and caspase-8.99 cFLIP_L also contains a pseudo-caspase domain, which lacks critical cysteine residues required for caspase activation. FLIP can be recruited during DISC formation (Fig. 2) to inhibit the apoptotic cascade by binding to FADD or caspase-8 by DED-DED interactions.67 FLIP has been indicated as important in the progression of cancer. For example, Ryu et al.100 showed overexpression of cFLIP_L in colonic adenocarcinomas compared to matched normal tissues. FLIP has also been identified in the development of drug and TRAIL resistance in human cancers. FLIP levels were higher in three TRAIL-resistant melanoma cell lines compared to five sensitive lines and actinomycin D treatment of one resistant cell line reduced FLIP levels and significantly sensitized cells to TRAIL.101 A variety of chemotherapy agents have been shown to reduce FLIP levels and enhance susceptibility to TRAIL-induced apoptosis in diverse types of human cancers. For instance, combination treatment with doxorubicin and TRAIL produced tumor growth inhibition of PC3 prostate cancer xenografts and reduced tumoral FLIP levels.102 PPAR-γ ligands103 and synthetic triterpenoids104,105 have also been shown to reduce FLIP and sensitize tumor cells to TRAIL-induced apoptosis. In human multiple myeloma cells, an increased FLIP to procaspase-8 ratio was present in TRAIL resistant cells. Treatment with cyclohexamide, bisindolymalemide (a protein kinase C inhibitor) or FLIP oligonucleotides resulted in the reversal of resistance.106 Therefore, FLIP may be an important modulator of TRAIL-resistance in a variety of human tumors, and many agents that reduce FLIP levels enhance TRAIL efficacy. However, other investigators have failed to show any correlation between FLIP levels and TRAIL-resistance and attribute it to other intracellular factors. For example, no relationship between TRAIL susceptibility and FLIP expression was detected in a panel of 28 melanoma cell lines,107 6 lung cancer lines108 or 13 glioma cell lines.109

Bcl-2 family.

The balance between pro- and anti-apoptotic activities of the Bcl-2 family of proteins also regulates sensitivity to TRAIL and other therapies (Table 3). This family consists of at least 20 proteins, all of which contain one or more conserved Bcl-2 homology (BH) domains.110,111 Many anti-apoptotic members have been identified, including: Bcl-2, Bcl-X_L, Bcl-w, Bfl-1 and Mcl-1. These proteins contain a hydrophobic groove containing residues of their BH1, BH2 and BH3 regions and a hydrophobic C-terminal domain that allows them to target intracellular membranes. The Bax family and the BH3-only family comprise two pro-apoptotic groups. Bax family members have BH1, BH2 and BH3 protein domains similar to the anti-apoptotic proteins, but their C-terminal domain occludes the hydrophobic groove until a conformation change occurs with apoptotic signals. The BH3-only proteins have a short BH3 region and act as internal sensors for damage and antagonize the anti-apoptotic Bcl-2 members. Both Bax and BH3-only pro-apoptotic molecules must be present to produce apoptosis.110

Bcl-2, Bcl-X_L, Bcl-w and Mcl-1 strongly inhibit apoptosis in response to many cytotoxic agents in a variety of cell types and overexpression of Bcl-2 or Bcl-X_L has been reported to confer resistance to TRAIL in a variety of tumor cells.110,112 For example, Fulda et al.112 reported Bcl-2 overexpression protected neuroblastoma, glioblastoma and breast cancer cells from TRAIL-induced apoptosis. Cleavage of caspase-3, 7, 8 and 9 was reduced, as well as decreased processing of their substrates PARP, DFF45 and XIAP. Protection against TRAIL cytotoxicity was also demonstrated by Bcl-2 overexpression in lung113 and colon cancer cells.114 The expression of Bcl-X_L in three pancreatic cancer cell lines was associated with TRAIL resistance.115 Expression of Mcl-1, a more recently described Bcl-2 family member, has also been correlated to TRAIL resistance in cancer cells and knock-down of Mcl-1 levels by various methods, such as small-interfering RNA, sensitized cancer cells to TRAIL-induced apoptosis.116–119

Decreased expression of pro-apoptotic Bax family proteins has also been implicated in TRAIL resistance. TRAIL induced cytochrome c release and apoptosis in Bax or Bak knockout murine embryonic fibroblasts, but not in the double knockout cells, suggesting that in these cells Bax and Bak may provide some compensation for each other.36 In HCT116 colon carcinoma cells, Bax-deficient cells were TRAIL resistant and lacked cleavage of caspase-9, -7 and PARP; however TRAIL sensitivity was restored with camptothecin and etoposide pretreatment which produced an increase in Bak and DR5 expression.120 TRAIL in combination with 5-FU121 or ionizing irradiation122 synergistically induced apoptosis in Bax expressing prostate cancer cells, while cells without Bax were resistant to TRAIL-induced apoptosis in combination with either agent. Han et al.123 reported that resistance to TRAIL cytotoxicity in Bax and Bak deficient Jurkat leukemia cells could be reversed with adenoviral transduction of the Bax gene, but not Bak. These reports indicate that the loss of pro-apoptotic proteins, especially Bax, may be important in the resistance of cancer cells to TRAIL-induced apoptosis.

TRAIL has been combined with a variety of other agents to overcome resistance by modification of the Bcl-2 family of proteins.124 Ray and Almasan124 reported that TRAIL combined with CPT-11 increased Bax and reduced Bcl-X_L expression in prostate cancer cells in vitro; whereas in vivo, they induced increased intratumoral Bak and Bcl-X_S expression and decreased Bcl-w and Bcl-X_L. Bortezomib, a proteasome inhibitor, was shown to decrease Bcl-2 and Bcl-X_L in glioblastoma multiforme cells in vitro and enhance TRAIL-induced cytotoxicity.125 Two TRAIL-resistant colon cancer cell lines produced by Zhu et al.126 were sensitized by bortezomib or MG-132, another proteasome inhibitor, which resulted in increased expression of DR5 and Bik a BH-3 only pro-apoptotic protein. Promising new agents under investigation for combination treatment with TRAIL are small molecule Bcl-2 inhibitors. HA14-1, a Bcl-2 inhibitor, combined with TRAIL resulted in increased apoptosis in Bcl-2 overexpressing TRAIL-resistant SW480 colon carcinoma cells.114 CEM leukemia cells were sensitized to TRAIL by low concentrations of HA14-1 and BH3I-2′ another Bcl-2 inhibitor.127 Bcl-2 siRNA treatment enhanced TRAIL-induced apoptosis in A375 melanoma cancer cells.128 Gossypol, a cottonseed oil extract, has also shown BH3-mimetic properties and sensitized lung and esophageal cancer cells to TRAIL with an increase in apoptosis.129 Another Bcl-2 small molecule inhibitor, ABT-737, was combined with TRAIL to increase cytotoxicity against certain renal, lung and prostate cancer cell lines.130 ABT-737 was also shown to be effective in enhancing TRAIL cytotoxicity against the human pancreatic cell lines PANC-1 and BxPC-3. Mechanistic studies revealed the combination produced greater activation of apoptosis via disassociations of the pro-apoptotic Bcl-2 family members from the anti-apoptotic members to favor apoptosis.131 These strategies highlight the importance of the Bcl-2 family of proteins in TRAIL-induced apoptosis.

IAPs and Smac/DIABLO.

Inhibitors of apoptosis (IAP) proteins are a highly effective cellular means of blocking the apoptotic cascade through interactions with caspases or Smac/DIABLO. Each member of the IAP family is characterized by one to three tandem repeats of a baculoviral IAP repeat (BIR)-binding domain which allow for binding to specific caspases or pro-apoptotic molecules. Many family members have been identified, including cIAP1, cIAP2, XIAP, survivin, BRUCE and NAIP.67 Certain members also have RING domains that allow them to act as ubiquitin E3 ligases to initiate the degradation of target proteins following attachment of ubiquitin molecules.132 XIAP blocks the activity of effector caspase-3 and 7 and prevents the activation of caspase-9 by direct interactions.133,134 Other IAPs function by binding to pro-apoptotic molecules such as Smac/DIABLO, which is a mitochondrial protein released along with cytochrome c following mitochondrial membrane depolarization by certain apoptotic stimuli (Fig. 2). Smac/DIABLO associates with IAPs to inhibit their anti-caspase activity and progression of the apoptotic cascade may be related to the balance of pro- and anti-apoptotic molecules. Many IAPs have been associated with chemotherapy and TRAIL resistance.135–140 XIAP and survivin have been most extensively described to play a major role in TRAIL resistance.67

XIAP appears to be the most potent caspase inhibitor in the family and functions by direct binding to caspases and by serving as ubiquitin-protein ligase for active caspase-3 to promote its degradation.141 Disruption of the XIAP gene in human colon cancer cells was shown to enhance their sensitivity to TRAIL suggesting that XIAP is an important modulator of TRAIL-induced apoptosis.136 Various methods have been used to reduce XIAP protein or messenger RNA levels to reverse TRAIL resistance. In our own studies, doxorubicin decreased XIAP protein levels to a varying extent in breast cancer cell lines.48 Flavopiridol, a cyclin-dependent kinase inhibitor and TRAIL synergistically increased apoptosis in human leukemia cells with reductions in XIAP.142 RNA interference targeting XIAP was used in combination with TRAIL to induce apoptosis in pancreatic cells in vitro and in vivo where the combination induced the regression of PancTu1 tumor xenografts.143 A small molecule Smac/DIABLO mimetic, which binds to XIAP with strong affinity, was shown to synergize with TRAIL or the anti-DR5 antibody HGS-ETR2 (Lexatumumab) against ovarian cancer cells and with TRAIL against breast cancer cell lines.144,145 The modulation of Smac/DIABLO and XIAP may provide future clinical benefit as development of other mimetics continues.

Survivin has a dual function; it inhibits caspase-9 activation within the apoptosome and it has a role in microtubule stability during mitosis that functions in cell cycle progression.146 Li et al.137 showed lower survivin expression in four TRAIL sensitive lines compared to seven more TRAIL resistant uveal melanoma cell lines. Topotecan produced a decrease in survivin and an increase in DR4 and DR5 levels in prostate cancer cells while also increasing susceptibility to TRAIL.147 Decreased survivin expression and TRAIL sensitization of breast cancer cells was also noted with PPAR-γ agonists.148 Survivin antisense RNA has been shown to reverse TRAIL resistance in two uveal melanoma cell lines.137 siRNA-mediated downregulation of XIAP and survivin also have been used to sensitize melanoma and renal cell carcinoma cells to TRAIL.128

Nuclear factor-kappaB signaling.

The nuclear factor-kappa B (NFκB) family members are transcription factors, including cRel, RelA (p65), RelB, p50 (NFκB1) and p52 (NFκB2).149 Each has a conserved Rel homology domain and together form more than ten homo- and heterodimer complexes. Most NFκB dimers interact with the majority of κB DNA binding sites with high affinity; however some interact preferentially with other promoters and can elicit transcription with varied efficiencies.149 NFκB proteins are ubiquitously expressed in cells and their activity is regulated by the inhibitor of κB (IκB) family of proteins (Fig. 6). IκB proteins block nuclear localization signals of functional NFκB dimers by binding to dimerization domains and sequestering the dimers in the cytoplasm. Upon exposure to a NFκB-inducing stimulus, IκB kinase (IKK) complexes are activated and IκB proteins are phosphorylated at serine residues. Following phosphorylation, IκB is ubiquitinated at lysine residues and degraded by the proteasome, which releases active NFκB to translocate to the nucleus. Once active NFκB dimers are located in the nucleus, they can induce transcription of a variety of target genes. NFκB complexes may have a pro- or anti-apoptotic function.41,149 Anti-apoptotic targets include cIAP1/2, XIAP, TRAF1/2, Bfl-1, Bcl-X_L, DcR3 and FLIP. Some investigators report pro-apoptotic NFκB activity with the induction of gene transcription and protein expression of DR4, DR5 and TRAIL. However, the balance between pro- and anti-apoptotic signaling requires further study.41,149

TRAIL activation of NFκB signaling activity is complex and may occur through DR4, DR5 and DcR2 signaling (Fig. 6). TRAIL induces NFκB signaling via recruitment of receptor-interacting protein (RIP), a serine threonine kinase, by FADD within the DISC.19 RIP, along with TNF receptor associated factor 2 (TRAF2), stimulates members of the IκB kinase (IKK) complex, NFκB-inducing kinase (NIK) and IKKα/β (IKK1/2),150 which lead to IκB degradation and release of active NFκB dimers. Recruitment of RIP is enhanced when cells are pretreated with a caspase inhibitor.19 Proteolytically active caspase-8 cleaves RIP to form a dominant negative fragment, which blocks the NFκB pathway. Therefore when the apoptotic cascade is activated, NFκB activity is diminished in a caspase-sensitive manner.149

The pro-survival or pro-apoptotic function of NFκB signaling within cells may be dependent on the relative abundance of the various NFκB proteins. Researchers report differences in transcriptional activity of the cRel and RelA proteins. Ravi et al.84 reported that wild-type and RelA double knockout mouse fibroblasts were sensitive to TRAIL-induced apoptosis, but cRel knockout cells were resistant. Forced expression of cRel was shown to enhance sensitivity to TRAIL and increase levels of DR4 and DR5, which could be blocked by IκB expression. RelA expression reduced TRAIL cytotoxicity and increased Bcl-X_L levels. Chen and colleagues151 found that RelA overexpression in MDA-MB-231 breast cancer cells reduced expression of caspase-8, DR4 and DR5 expression, while an increase in cIAP1/2 protected cells from TRAIL-mediated apoptosis. Overexpression of cRel amplified TRAIL-induced apoptosis with an increase in DR4, DR5 and Bcl-X_S and reduced cIAP1/2 and survivin. Therefore, NFκB may enhance or hinder apoptosis depending on the permutations of subunits and dimers present in cells.

In many types of human cancer cells, reductions in NFκB anti-apoptotic activity enhance the cytotoxic response to TRAIL. NFκB was shown to be induced by TRAIL treatment in hepatoma cells with activation of IKK and degradation of IκB, while NFκB inhibition increased TRAIL-induced cytotoxicity.152 Proteasome inhibitors are promising modulators of the NFκB pathway, mainly by reducing IκB degradation. Mitsiades et al.153 used bortezomib (PS-341/Velcade), a proteasome inhibitor, to enhance TRAIL-mediated apoptosis in multiple myeloma cells. Bortezomib and geldanamycin, a heat shock protein 90 inhibitor, were shown to synergistically block NFκB activity in TRAIL resistant pancreatic cancer cells. The combination also reduced expression of Bcl-X_L, Bcl-2, cIAP1 and cyclin D and reversed resistance to TRAIL.154 Interferon-α155 and curcumin, a plant extract,156 are additional agents that restore cancer cell sensitivity to TRAIL by inhibiting NFκB activity. In TRA-8 resistant BT-474 cells, 24 or 48 h exposure to doxorubicin produced a dramatic decrease in expression of IκBα, compared to untreated control cells, while cells treated with a combination of TRA-8 and doxorubicin had a greater reduction in IκBα protein levels (Fig. 7). A reduction in IκBα generally indicates activation of NFκB signaling. The expression of the active subunits of the NFκB complex determines whether its function is primarily pro- or anti-apoptotic. The NFκB subunit, p65, showed a modest reduction following 3 h of TRA-8 and 24 h of doxorubicin treatment. However, combination treatment substantially reduced p65 levels after 24 h TRA-8 and 48 h doxorubicin exposure. These results indicate that despite a decrease in IκBα, NFκB signaling may be reduced by doxorubicin treatment in breast cancer cell lines. However, blockade of NFκB signaling via inhibition of translocation of NFκB subunits into the nucleus by SN50 or knockdown of p65 by siRNA failed to sensitize BT-474 cells to TRA-8 (Fig. 8). These results show that blockade of only NFκB signaling may not be sufficient to enhance sensitivity to TRAIL receptor-targeted therapies.

PI3K and Akt.

Phosphatidylinositol-3 kinase (PI3K) is a major regulator of receptor tyrosine kinase and G protein-coupled receptor activity. Upon stimulus with growth factors of these various receptors, PI3K phosphorylates the plasma membrane phospholipid, phosphatidylinositol-4,5 bisphosphate (PIP2) to phosphatidylinositol-3,4,5 trisphosphate (PIP3).157 One critical downstream effector of PI3K is the serine/threonine kinase Akt. Negative regulation of the PI3K/Akt pathway is primarily by PTEN (phosphatase and tensin homologue deleted on chromosome 10) activity. PTEN dephosphorylates PIP3 to PIP2, which reduces PI3K and Akt activity.158 Akt exists in mammalian cells as three isoforms (Akt1, Akt2 and Akt3). Akt is recruited to the plasma membrane where PIP3 binding induces a conformational change uncovering phosphorylation sites within Akt. Next, 3′-phosphoinositide-dependent kinase 1 (PDK1) phosphorylates Akt and stabilizes its active conformation. Akt has many downstream targets, especially mediators of cell proliferation and cell survival.158 Akt activation promotes cell proliferation through inhibition of glycogen synthase kinase-3, which leads to increased cyclin D expression and cell cycle progression. Akt also phosphorylates p21/Waf1 and p27/Kip2 to prevent their nuclear translocation and anti-proliferative effects.158 Anti-apoptotic effects of Akt include phosphorylation of Bad, which prevents it from inactivating Bcl-X_L and blocks cytochrome c release. Akt may also phosphorylate caspase-9 to prohibit its activation. The forkhead transcription factor family is also inactivated by means of phosphorylation by Akt to inhibit its transcription of pro-apoptotic genes. Akt phosphorylates murine double minute-2 protein (MDM2) to enhance p53 degradation and inhibit apoptosis. Akt stimulates the NFκB pathway by activation of IKK to increase IκB degradation, allowing NFκB to induce the expression of a variety of anti-apoptotic proteins.

PI3K, Akt and PTEN have important roles in cancer cell survival and resistance to cell death by many agents, including TRAIL. PTEN is one of the more frequently mutated or deleted tumor suppressors in human tumors.158 Loss of PTEN expression leads to an increase in PIP3 levels resulting in constitutively activated Akt. This has been reported in thyroid, breast, colon, prostate and other tumors. LNCaP prostate cancer cells are reported to be TRAIL resistant due to lack of active PTEN and presence of constitutively active Akt, which may be overcome by PI3K inhibitors159,160 or dominant negative Akt.159 Restoration of active PTEN expression in LNCaP cells by an adenoviral vector sensitized cells to TNF and TRAIL-induced apoptosis in a FADD-dependent manner.161 Amongst six human gastric cancer cell lines, the most TRAIL-resistant line, SNU-216, exhibited the highest level of Akt activity and FLIP_S expression. LY294002, a PI3K inhibitor, was able to decrease both Akt and FLIP and sensitize cells to TRAIL-mediated apoptosis. Furthermore, sensitive cells could be made resistant by overexpression of constitutively active Akt.162,163 In five non-small cell lung cancer cell lines, expression of phospho-Akt inversely correlated with TRAIL sensitivity.164 Akt blocked Bid cleavage and the intrinsic pathway of apoptosis in TRAIL-resistant cells; additionally, PI3K inhibitors, dominant negative Akt expression or PTEN transfection sensitized resistant H1155 lung cancer cells to TRAIL. Conventional chemotherapy agents, including paclitaxel and cisplatin, enhanced TRAIL-mediated apoptosis in SKRC-49 renal cell carcinoma cells by ceramide formation, which produced Akt inactivation.165 Measurements of basal phospho-Akt levels, the active form, in 2LMP and BT-474 breast cancer cells revealed phospho-Akt activity in BT-474 cells with no detection of phospho-Akt (Ser473) in 2LMP cells (Fig. 9). In BT-474 cells, phospho-Akt was reduced by treatment with a combination of TRA-8 and doxorubicin. These results suggest that Akt may contribute to the resistance of BT-474 cells. To further determine the importance of Akt signaling, chemical inhibitors of the pathway were used to interrupt Akt signaling by a variety of mechanisms. BT-474 cells were pretreated with a PI3K inhibitor, LY294002 or an Akt inhibitor, 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate, for 24 h prior to the addition TRA-8 antibody for an additional 24 h. Neither agent combined with TRA-8 increased cytotoxicity (Fig. 10). These results indicate that doxorubicin in combination with TRA-8 modulated Akt expression in BT-474 cells, but this modulation alone was not the mechanism responsible for increased cytotoxicity after combination treatment. Therefore, the PI3K/Akt pathway may be important in some human tumor cell lines, but not all. Nonetheless, Akt may play a role in cellular resistance to TRAIL-therapy in certain human cancer cell types and modifying the PI3K/Akt pathway in cancer cells may identify new targets to reverse TRAIL resistance.

Growth factors may affect TRAIL response via modulation of the PI3K/Akt pathway. Insulin-like growth factor-1 (IGF-1) decreased TRAIL-induced cytotoxicity in multiple myeloma cells166 and thyroid carcinoma cells167 while enhancing Akt and NFκB activity with upregulation of FLIP, survivin, cIAP1 and XIAP levels. Epidermal growth factor (EGF) protected MDA-MB-231 breast cancer cells and human embryonic kidney cells HEK 293 from TRAIL-mediated apoptosis via activation of Akt and blockade of cytochrome c release.168 This activation of the Akt pathway by EGF may be exploited for combination therapy with TRAIL. Shrader et al.169 reported that gefitinib, an EGFR inhibitor, in combination with TRAIL induced increased apoptosis by decreasing active Akt and XIAP levels. Thus, some growth factors modulate TRAIL sensitivity through anti-apoptotic signaling and inhibition of growth factor signaling may increase sensitivity.

p53.

p53 and other members of the p53 pathway can have a profound effect on cancer cells by regulating apoptosis and the cell cycle, while playing an important role in chemotherapy-induced sensitization to the TRAIL ligand and agonistic antibodies. For example, antiproliferative effects of doxorubicin and SN-38, the CPT-11 active metabolite, have been related to the association of p21, a p53 target, with DNA leading to the degradation of c-myc and cdc25A and inducing cell cycle arrest.170 However, p53 deficiency or mutations are very common amongst human cancers and are frequently associated with resistance to chemotherapy.171 Full-length p53 homologous protein, p73, contains a transactivation domain similar to p53 and may induce p53 targets, such as p21 and GADD45, resulting in apoptosis.172,173 Expression of certain isoforms of these proteins may compensate for mutated p53, as p73 activation may induce apoptosis and cell cycle arrest in response to DNA damage.172,173 Pharmacological agents that activate the p53 pathway and induce apoptosis in p53 null or mutant tumor cells are currently in development. A small molecule library was screened in cells lacking p53 to show induction of p53 activity, such as increased p21 and DR5 expression. In vivo tumor xenograft models showed a reduction in tumor weight and increased p53 transcriptional activity with three different p53 stimulating small-molecules.171 Weinmann et al.174 identified a “p53 rescue compound P53R3” by screening a compound library for p53 binding. P53R3 was reported to induce some p53 target genes, including the upregulation of DR5 mRNA, protein and surface expression. P53R3 sensitized nine of twelve glioma cells lines to TRAIL-induced apoptosis. With the large number of pro- and anti-apoptotic proteins regulated by p53, these and similar agents may have importance in the reversal of resistance to TRAIL-based therapies and chemotherapy as well.

Autophagy.

Several studies have the shown that TRAIL induces autophagy in certain cell lines.175–178 Autophagy is a cellular process of recycling macromolecules activated by cellular stress that can either result in lysosome-mediated cell death or cytoprotection.175,179 Han et al.175 reported that HCT116 colon cancer cells overexpressing FLIP did not undergo apoptosis upon treatment with TRAIL, but rather an autophagic response with an increase in Beclin-1 and the presence of autophagosomes (double membrane-bound vesicles that fuse with lysosomes and are characteristic of autophagy).180 Knock-down of Beclin-1 and UVRAG (another autophagy related protein) sensitized these cells to TRAIL-induced apoptosis. In the wild-type HCT116 cells, ∼40% of cells did not undergo apoptosis with TRAIL treatment alone, but were sensitized by Beclin-1 knockdown. Similar studies in TRAIL-resistant Bax^-/- HCT116 cells, RKO colon cancer cells, cisplatin-resistant MCF7 and etoposide-resistant MDA-MB-231 breast cancer cells, and U251 and LN229 glioma cell lines showed sensitization to TRAIL-induced apoptosis with Beclin-1 siRNA treatment.175,176,181 The cellular switch between apoptosis and autophagy has been related to the activity of caspase-8 and the activation of the mitochondrial apoptotic pathway.175,177 These studies suggest that novel and existing therapeutic agents which induce autophagy may be useful in sensitizing apoptosis-deficient cancer cells to TRAIL-induced apoptosis.182

Therapeutic Potential of TRAIL and Agonistic Death Receptor Antibodies in Combination Therapy

Resistance to chemotherapy or radiation is a common problem for many cancer patients, and some tumor cells are resistant to TRAIL-induced apoptosis. TRAIL or antibodies targeted to TRAIL death receptors have been shown to interact with various chemotherapeutic agents to sensitize cells in an additive to synergistic manner. The mechanisms of sensitization include induction of increased cell surface death receptor expression or increased activation of the intrinsic or extrinsic apoptotic pathways via modulation of apoptotic regulatory proteins. As previously described, many therapeutic agents sensitize cancer cells to TRAIL-induced apoptosis by modulation of the various apoptotic regulatory proteins. Many classes of chemotherapy agents are used for the treatment of cancer (Table 1) and have been shown to enhance the efficacy of TRAIL and death receptor agonistic antibodies. With such a large variety of drugs sensitizing cancer cells to TRAIL receptor-targeted therapies, further study is needed to determine if sensitization occurs via similar mechanisms for drugs with very different primary mechanisms of action.

Murine in vivo tumor xenograft models have been used to investigate the efficacy of TRAIL and drug or radiation combination treatment on tumor growth inhibition. TRAIL with either 5-FU or CPT-11 produced greater anti-tumor effects than either agent alone against primary human colon cancer samples implanted into SCID mice. TRAIL and CPT-11 combination treatment achieved complete tumor regression in 50% of animals.183 In an orthotopic NCI-H460 lung cancer model, TRAIL combined with paclitaxel and carboplatin significantly inhibited tumor growth and increased 90 day survivial.184 These examples encompass only a small fraction of studies describing the in vivo effects of TRAIL or death receptor agonistic antibodies in combination with chemotherapy in a variety of tumor types.1,63 A recently published review by Ashkenazi and Herbst63 provides a summary of chemotherapy agents used in combination with TRAIL in multiple preclinical in vivo models of human carcinomas.

In addition to chemotherapy, radiation has also been shown to increase the efficacy of TRAIL. Breast, lung, colorectal and head and neck cancer cell lines were treated in vitro with TRAIL plus irradiation resulting in synergistic induction of apoptosis in five of six tumor cell lines and increased DR5 expression in four cell lines.185 Chinnaiyan et al.78 reported a p53-dependent synergistic effect of TRAIL and radiation against breast cancer cell lines and tumor regression of MCF-7 tumor xenografts. Sequential treatment with radiation followed by TRAIL 24 h later synergistically inhibited PC-3 prostate and MCF-7 breast tumor xenograft growth and increased survival in nude mice with caspase-3 activation detected in both models.79,186 Recently, X-irradiation in combination with TRAIL was shown to synergistically inhibit the growth of MKN45 and MKN28 human gastric cancer xenografts. Caspase-3 activation was shown by combination treatment in normoxic and hypoxic regions of the tumors.187 These studies highlight the potential for TRAIL-based therapies in combination with standard therapeutic agents for cancer treatment.

Newer agents currently at various stages of clinical development also show promise for combination treatment with TRAIL and death receptor antibodies, including proteasome inhibitors (bortezomib),125,153,188–191 histone deacetylase inhibitors (HDAC),192–197 heat shock protein 90 (HSP90),198,199 inhibitors and small molecule apoptotic modulators.131,145,200 TRAIL death receptor therapies may have an impact on a variety of cancer types. Beneficial combinations may include already approved drugs and newer agents currently under development.

Disclosure of Potential Conflicts of Interest

Dr. Buchsbaum has intellectual property interests related to the TRA-8 anti-DR5 antibody.

Figures and Tables

Figure 1 TRAIL receptors. Ligand binds to cysteine-rich domains. The death receptors have death domains that provide interactions with intercellular proteins to induce apoptotic signaling. Figure based on Almodovar et al.10

Figure 2 The death receptor induced extrinsic and intrinsic apoptotic pathways. Each pathway begins with caspase-8 activation. Extrinsic pathway proceeds with the direct activation of caspase-3 by activated caspase-8. Intrinsic or mitochondrial pathway activation involves the cleavage of Bid to activate Bcl-2 family members to depolarize the mitochondrial membrane and release cytochrome c and Smac/DIABLO into the cytosol. Once released these molecules interact with Apaf-1 to activate caspase-9, which activates effector caspase-3. In some cells, the extrinsic pathway is reported to be sufficient for TRAIL-induced apoptosis, while in other cells both pathways are activated. Chemotherapy drugs may increase activation of the intrinsic pathway to enhance TRAIL receptor targeted therapies.

Figure 3 TRA-8 induced cytotoxicity against A549 lung carcinoma cells was enhanced by chemotherapy pretreatment. Cells (1,000 per well) were plated in 96 well plates and incubated at 37°C for 24 h. Then cells were treated with doxorubicin, docetaxel or bortezomib for 24 h prior to the addition of TRA-8. Cell viability was determined via ATPLite assay after 24 h of TRA-8 treatment. ATP levels are reported relative to untreated control cells and represent the mean and SD of quadruplicate samples from three independent experiments.

Figure 4 TRA-8 or TRAIL induced cytotoxicity was enhanced by doxorubicin pretreatment. Cells (1,000 per well) were plated in 96 well plates and incubated at 37°C for 24 h. Then cells were treated with DOX for 24 h prior to the addition of TRA-8 or TRAIL. Cell viability was determined via ATPLite assay after 24 h of TRA-8 or TRAIL treatment. ATP levels are reported relative to untreated control cells and represent the mean and SD of quadruplicate samples from three independent experiments.

Figure 5 Caspase-3, -8, -9 and PARP cleavage and decreased Bid and XIAP levels were induced by TRA-8 or TRAIL in sensitize cells and with combination treatment in resistant cells. 2LMP and BT-474 cells were pretreated for 24 h with DOX (50 and 5,000 nM, respectively) before the addition of TRA-8 or TRAIL (125 and 1,000 ng/ml; TRA-8, 0.84 and 6.7 nM; TRAIL, 5 and 40 nM) for 3 h. Whole cell lysates were analyzed by western blot analysis using antibodies specific for the identified protein: caspase-3 (Stressgen); caspase-8, PARP (BD PharMingen); caspase-9, Bid (cell signaling) and XIAP (R&D Systems). Actin (Sigma) was used as a loading control.

Figure 6 NFκB pathway activation via death receptor signaling. TRAIL has been reported to induce the canonical NFκB signaling pathway, which involves the inactivation of IκB subunits by the IKK complex to release active NFκB homodimers or heterodimers. Generally, these active NFκB dimers translocate to the nucleus and induce gene transcription. NFκB may have different functions depending on the NFκB subunits present and their binding partners. Chen et al.151 reported the p65 subunit to be anti-apoptotic and cRel to be pro-apoptotic.

Figure 7 Expression of IκBα or p65 after exposure to TRA-8 and doxorubicin. BT-474 cells were treated with DOX (5 µM for 24 h samples and 3 µM for 48 h samples) and TRA-8 (1,000 ng/ml) prior to western blot analysis with antibodies to IκBα (Cell Signaling) and p65 (Santa Cruz).

Figure 8 Inhibition of NFκB did not enhance cytotoxicity of TRA-8 against BT-474 breast cancer cells. Cells were pretreated with the NFκB translocation inhibitor SN50 or p65 specific siRNA for 24 h prior to the addition of TRA-8. Cell viability was determined via ATPLite assay after 24 h of TRA-8 treatment.

Figure 9 Akt activity in breast cancer cell lines. 2LMP and BT-474 cells were treated with TRA-8 (125 and 1,000 ng/ml) for 3 h following pretreatment with DOX (50 and 5,000 nM) for 24 h. Phospho-Akt (Ser473) and Akt were detected with specific primary antibodies (cell signaling) in whole cell lysates.

Figure 10 Pretreatment with Akt inhibitors did not enhance cytotoxicity of TRA-8 against BT-474 breast cancer cells. Cells were treated with the PI3K inhibitor, LY294002 or an Akt inhibitor, 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate, for 24 h prior to the addition of TRA-8. Cell viability was determined via ATPLite assay after 24 h of TRA-8 treatment.

Table 1 Chemotherapeutic agents that interact with TRAIL and mechanisms of sensitization

Family	Mode of action	Examples	Modulation of apoptotic regulatory proteins
Anthracycline Antibiotics	Topoisomerase II inhibition and DNA intercalation	Doxorubicin, Epirubicin, Actinomycin	↑ DR4/5; ↓ FLIP; ↓ XIAP; ↑ p21; ↑ Apaf-1
Camptotheca	Topoisomerase I inhibition and DNA strand breaks	Camptothecin, Topotecan, CPT-11	↑ DR4/5; ↑ Bax, ↓ Bcl-XL; ↓ FLIP; ↓ survivin; ↓ Akt activity; ↑ p21
Platinum	DNA alkylating agents	Carboplatin, Cisplatin, Oxaliplatin	↑ DR4/5; ↓ FLIP
Podophyllum	Topoisomerase II inhibition and cell cycle arrest	Etoposide	↑ DR4/5
Proteasome Inhibitors	Disruption of protein degradation via inhibition of the 20S proteasome	Bortezomib	↑ DR5; ↓ Bcl-2, Bcl-XL; ↑ Bik; ↓ cIAP1; ↓ NFκB activity; ↑ p21
Pyrimidines	Inhibit enzymes of nucleic acid synthesis and incorporate into RNA and DNA	5-Fluorouracil, Gemcitabine	↓ FLIP; ↑ caspase-8
Taxanes	Microtubule stabilization and mitotic arrest	Paclitaxel, Docetaxel	↑ DR4/5; ↓ Akt activity

Table 2 Intracellular proteins and signaling pathways associated with resistance to TRAIL and TRAIL receptor-targeted therapies

Protein	Resistance mechanism	Agents that modulate
Death Receptors	No relation to basal expression, may be post-translational modification	Doxorubicin, etoposide (increase expression)
FLIP	Increased FLIPS or FLIPL	Doxorubicin, cyclohexamide, PPAR-γ ligands
Bcl-2 Family	Increased expression of anti-apoptotic (Bcl-2, Bcl-XL, Mcl-1) or decreased expression of pro-apoptotic (Bax, Bak) proteins	CPT-11, bortezomib, Bcl-2 inhibitors
IAP Family	Increased XIAP or survivin	Doxorubicin, topotecan, flavopiridol, Smac mimetics
NFκB	Constitutive or increased pro-apoptotic signaling	Bortezomib, geldanamycin, IFNγ, curcumin
Akt	Constitutive anti-apoptotic activity	Cisplatin, paclitaxel, gefitinib
p53	Deficiency or mutated	DNA-damaging agents, small molecule p53 activators

Table 3 Bcl-2 proteins

Anti-apoptotic		Pro-apoptotic
Bcl-2 family		Bax family		BH3-only
Bcl-2	Bcl-XL	Bax Bcl-Xs	Bak Bok	Bid	Bik	Bim
Bcl-w	Bfl-1			Bad	Bmf	Blk
Mcl-1				Noxa	Puma

Acknowledgements

This work was supported in part by DOD grant #W81XWH-06-1-0706, DOD grant #BCDAMD17-00-1-0119 and NIH Breast Cancer SPORE 5P50CA089019-08.