**1. Introduction**

200 Advances in Cancer Therapy

Matlin, A.J.; Clark, F. & Smith, C.W. (2005). Understanding alternative splicing: towards a

Mola, G. et.al (2007). Exonization of Alu-generated splice variants in the survivin gene of

Nowak, S.J. Corces, V.G. (2000). Phosphorylation of histone H3 correlates with

Salz, W. et.al (2005). A survivin gene signature predicts aggressive tumor behavior. *Cancer* 

Sampath, J. Pelus, L.M. (2007). Alternative splice variants of survivin as potential targets in

Skotheim, R.I. Nees, M. (2007). Alternative splicing in cancer: noise, functional, or

Sohn, D.M. et.al (2006). Expression of survivin and clinical correlation in patients with breast

Song, Z. Wu, M. (2005). Identification of a novel nucleolar localization signal and a

Takizawa, B.T. et.al (2007). Downregulation of survivin is associated with reductions in TNF

Wei, Y. et.al (1999). Phosphorylation of histone H3 is required for proper chromosome

Wheatley, S.P. et.al (2004). Aurora-B phosphorylation in vitro identifies a residue of survivin

degradation signal in Survivin-deltaEx3: a potential link between nucleolus and

receptors' mRNA and protein and alterations in nuclear factor kappa B signaling in

that is essential for its localization and binding to inner centromere protein

human and non-human primates. *J Mol Biol*, Vol.366:1055-1063.

transcriptionally active loci. *Genes Dev*, Vol.14:3003-3013.

cellular code. *Nat Rev Mol Cell Biol*, Vol.6:386-398.

cancer. *Curr Drug Discov Technol*, Vol.4:174-191.

cancer. *Biomed Pharmacother*, Vol.60:289-292.

systematic? *Int J Biochem Cell Biol*, Vol.39:1432-1449.

protein degradation. *Oncogene*, Vol.24:2723-2734.

urothelial cancer cells. *Cancer Invest*, Vol.25:678-684.

condensation and segregation. *Cell*, Vol.97:99-109.

(INCENP) in vivo. *J Biol Chem*, Vol.279:5655-5660.

*Res*, Vol.65:3531-3534.

One fundamental problem of most malignancies, including those of haematological origin, is the development of multiple mechanisms of resistance, which progressively reduce or suppress the therapeutic efficacy of conventional radio-chemotherapy. In recent years novel compounds have been identified to overcome this major hurdle. Among these, TNF-related apoptosis-inducing ligand (TRAIL) generated considerable enthusiasm for its anticancer therapeutic effectiveness, selectively targeted to a wide range of cancer cells without affecting cells derived from normal tissues and organs. A number of preliminary studies sustain the use of TRAIL-Receptors agonistic antibodies (TRAs) instead of rTRAIL (recombinant TRAIL) in the treatment of tumour cells protected from rTRAIL-induced apoptosis by the expression of cell surface decoy receptors. Although the early clinical trials are promising and well tolerated, the efficacy of these novel approaches is restricted to patients with TRAIL-sensitive tumours. In addition, TRAIL-Rs loss or mutations that can often occur in neoplastic diseases can compromise expected therapeutic results. To overcome TRAIL resistance, novel strategies based on the combination of TRAIL with radiochemotherapy, or with proteasome or histone deacetylase or NF-B inhibitors have been developed. In light of this complex background, this chapter will discuss the current knowledge of the signalling pathways leading to TRAIL resistance and promising targeted therapies in the treatment of haematological malignancies.

### **2. The TRAIL/TRAIL-Rs system as a novel avenue in anticancer treatment**

In recent years novel compounds have been identified to overcome emergence of cancer cells resistance to conventional radio-chemotherapy. Among these, TRAIL generated considerable enthusiasm for its anticancer therapeutic effectiveness, selectively targeted to a wide range of cancer cells without affecting cells derived from normal tissues and organs. TRAIL, also known as Apo-2 Ligand (Apo-2L) (Pitti et al., 1996), is a member of the TNF super-family of cytokines including structurally related proteins that play important roles in regulating cell death, immune response and inflammation. The story of TRAIL begins when a new member of the TNF super-family capable of inducing apoptosis was identified and characterized by virtue of its sequence homology to CD95/Fas/Apo1L (FasL) and TNF (Wiley et al., 1995).

Signalling Pathways Leading to TRAIL Resistance 203

death domain (FADD) and the apoptosis initiator caspase-8 are recruited to TRAIL-R1 and/or TRAIL-R2 (Sprick et al., 2000). Although initial studies have attributed a central role to caspase-8 in mediating the apoptotic signal of TRAIL, subsequent studies have demonstrated that apoptosis can be triggered independently through TRAIL-R1 or TRAIL-R2 and proteolytic activation of effector caspases either directly by apical caspase-8 or -10 (Kischkel et al., 2001) and/or indirectly through Apaf-1-mediated activation of caspase-9 (Green, 2000). Similarly to CD95L, the response to TRAIL is cell-type specific and might be characterized by two distinct cell death pathways (LeBlanc & Ashkenazi, 2003). In the type I extrinsic pathway, extrinsic signals lead to the activation of large amounts of caspase-8 and to the rapid cleavage of executioner caspase-3 prior to loss of mitochondria trans-membrane potential. As a consequence, in type I cells (including leukaemia cells) Bcl-2 over-expression blocks the mitochondrial changes associated with cell death but does not prevent apoptosis that occurs upon death receptors activation. Recent studies have suggested that death receptor induction of apoptosis may depend on the degree of receptor aggregation/multimerization, which may, in turn, depend on the concentration of the death ligand, its form (i.e. soluble versus membrane-bound), the relative DR expression on the cell surface and the array of growth factors and cytokines to which the cells are exposed (Abdulghani & El-Deiry, 2010; Mellier et al., 2010). In the type II intrinsic pathway of apoptosis, intrinsic signals, like DNA damage, growth factor withdrawal or cytokine deprivation, affect the function of Bcl-2 family members (Roos & Kaina, 2006). In fact, in type II cells the extrinsic pathway activated by death receptors is ineffective to recruit, at the DISC level, enough caspase-8 to activate effector caspases. However, through homotypic aggregation at the DISC, caspase-8 is stabilized in an active form and released into the cytosol, where it cleaves its target proteins, most notably the pro-apoptotic Bcl-2 homology domain (BH3)-only protein Bid (BH3-interacting-domain death agonist) (Kelley & Ashkenazi, 2004), thus connecting the "intrinsic" mitochondrial pathway to the "extrinsic" DR pathway (Sprick & Walczak, 2004). In turn, truncated Bid (tBid) is able to bind antiapoptotic Bcl-2 family members like Bcl-2, Bcl-XL, Bcl-W and A1 allowing the pro-apoptotic Bcl-2 family members Bax and Bak to engage the mitochondria and induce the release of mitochondrial cytochrome c and Smac (second mitochondria-derived activator of caspases)/DIABLO into the cytosol, where these latter factors promote caspase activation. Actually, cytochrome c forms the "apoptosome" complex with the adaptor protein Apaf-1 resulting in the activation of the apoptosis-initiating protease caspase-9, which then stimulates effector caspases (Green, 2000). Instead, Smac/DIABLO binds to inhibitors of apoptosis proteins (IAPs), preventing their negative-regulatory binding to caspase-9 and -3 and then augmenting apoptosis induction (Salvesen & Abrams, 2004). In this scenario, oncogenic mutations affecting molecules involved in the intrinsic mitochondrial pathway might cause resistance emergence in type II cells, while mutations in the DR pathway could

confer resistance to DR-dependent apoptosis especially in type I cells (Fig. 1).

It has been demonstrated that the expression pattern of the two killer receptors is broad and partly overlapping, suggesting that they may serve as an alternate or "backup" system, allowing the immune system to control aberrant cells even if one of the two receptors had failed (Greil et al., 2003). Although further investigations are needed to assess differences between DR signalling and regulation, some interesting observations have been reported so far. For example, it has been shown that TRAIL-R1 is activated both by the soluble and the membrane-bound form of the ligand (MacFarlane et al., 2005) whereas TRAIL-R2 is activated by cross-linked soluble and membrane-bound TRAIL ligand but not by the soluble non-cross-

TRAIL/Apo-2L consists of 281-291 amino acids in the human and murine forms, respectively, which share 65% amino acid identity. Like other members of the TNF family, TRAIL is a type II membrane protein having an intracellular amino-terminal portion, an internal transmembrane domain, and an external carboxyl-terminus that forms a soluble molecule upon proteolytic cleavage (Mariani & Krammer, 1998). To facilitate biological studies, an epitopetagged soluble form of TRAIL was constructed and identified with SDS-PAGE with an apparent molecular weight of 28 kDa (Wiley et al., 1995). Gel filtration analysis of the purified soluble TRAIL disclosed that the native molecule was multimeric in solution with a size of ~80 kDa. Since then, a lot of rTRAIL preparations have been obtained and commercialized so that the variety of techniques of construction/purification employed may justify some data inconsistencies (LeBlanc & Ashkenazi, 2003). Both TRAIL and FasL exist as full-length membrane-bound molecules and as shorter soluble forms (Almasan & Ashkenazi, 2003) and induce apoptosis in a wide variety of transformed cell lines of diverse origin (Walczak & Krammer, 2000), although the membrane-bound form with a greater efficiency (Schneider et al., 1997). What immediately distinguished TRAIL both from FasL and TNF was its ability to induce apoptosis in various continuous cell lines and primary tumour cells, displaying minimal or no toxicity on most normal cells and tissues (Ashkenazi & Dixit, 1999). Significant levels of TRAIL transcripts have been detected in many human tissues and expressed constitutively in some cell lines (Wiley et al., 1995; Pitti et al., 1996), suggesting that TRAIL, unlike FasL, must not be cytotoxic to most tissues *in vivo*. Actually, accumulating evidences indicate that TRAIL plays important roles in the immune response to viruses, self-antigens and allergens and in the immune surveillance of tumours and metastases (Falschlehner et al., 2009).

#### **2.1 The TRAIL receptor family and regulation**

The biological effects induced by TRAIL are mediated by interactions with cell surface TRAIL-Receptors (TRAIL-Rs). Several studies have demonstrated an extreme complexity of TRAIL-Rs expression and function (Mellier et al., 2010; Mahmood & Shukla, 2010)). In fact, at least five TRAIL receptors belonging to the apoptosis-inducing TNF-receptor family have been described so far. TRAIL-R1 (death receptor DR4; TNFRSF10a) (Pan et al., 1997a) and TRAIL-R2 (DR5; TNFRSF10b) (Pan et al. 1997b) transduce apoptotic (Ozören & El-Deiry, 2003; Huang & Sheikh, 2007) as well as non-apoptotic signals (Di Pietro & Zauli, 2004; Park et al., 2005) upon TRAIL binding, while TRAIL-R3 (DcR1; TNFRSF10c) (Pan et al., 1997b) and TRAIL-R4 (DcR2; TNFRSF10d) (Marsters et al., 1997) as well as osteoprotegerin (OPG; TRAIL-R5; TNFRSF11b) (Emery et al., 1998) are homologous to TRAIL-R1 and TRAIL-R2 in their cysteine-rich extracellular domain but they lack the intracellular death domain (DD) and apoptosis inducing capability (Almasan & Ashkenazi, 2003).

#### **2.1.1 TRAIL death receptors**

TRAIL-R1 (Pan et al., 1997a) and TRAIL-R2 (Pan et al. 1997b) are type I trans-membrane proteins exposing the N-terminal TRAIL-binding domain and exerting pro-apoptotic signals through the cytoplasmic death domain. The cytoplasmic domain of TRAIL-R1/R2 shares a significant homology to the DD of different death receptors (DRs), such as CD95 and TNF-R1. Upon TRAIL binding to appropriate cognate receptors (TRAIL-R1 and TRAIL-R2), there is aggregation of TRAIL-Rs on the cell surface followed by the activation of both extrinsic and intrinsic intracellular death signalling pathways (Cretney et al., 2007). Shortly after addition of the ligand, the death-inducing signalling complex (DISC) is assembled (Kischkel et al., 1995). TRAIL DISC resembles that of Fas since the adaptor protein Fas associated

TRAIL/Apo-2L consists of 281-291 amino acids in the human and murine forms, respectively, which share 65% amino acid identity. Like other members of the TNF family, TRAIL is a type II membrane protein having an intracellular amino-terminal portion, an internal transmembrane domain, and an external carboxyl-terminus that forms a soluble molecule upon proteolytic cleavage (Mariani & Krammer, 1998). To facilitate biological studies, an epitopetagged soluble form of TRAIL was constructed and identified with SDS-PAGE with an apparent molecular weight of 28 kDa (Wiley et al., 1995). Gel filtration analysis of the purified soluble TRAIL disclosed that the native molecule was multimeric in solution with a size of ~80 kDa. Since then, a lot of rTRAIL preparations have been obtained and commercialized so that the variety of techniques of construction/purification employed may justify some data inconsistencies (LeBlanc & Ashkenazi, 2003). Both TRAIL and FasL exist as full-length membrane-bound molecules and as shorter soluble forms (Almasan & Ashkenazi, 2003) and induce apoptosis in a wide variety of transformed cell lines of diverse origin (Walczak & Krammer, 2000), although the membrane-bound form with a greater efficiency (Schneider et al., 1997). What immediately distinguished TRAIL both from FasL and TNF was its ability to induce apoptosis in various continuous cell lines and primary tumour cells, displaying minimal or no toxicity on most normal cells and tissues (Ashkenazi & Dixit, 1999). Significant levels of TRAIL transcripts have been detected in many human tissues and expressed constitutively in some cell lines (Wiley et al., 1995; Pitti et al., 1996), suggesting that TRAIL, unlike FasL, must not be cytotoxic to most tissues *in vivo*. Actually, accumulating evidences indicate that TRAIL plays important roles in the immune response to viruses, self-antigens and allergens and in the immune surveillance of tumours and metastases (Falschlehner et al., 2009).

The biological effects induced by TRAIL are mediated by interactions with cell surface TRAIL-Receptors (TRAIL-Rs). Several studies have demonstrated an extreme complexity of TRAIL-Rs expression and function (Mellier et al., 2010; Mahmood & Shukla, 2010)). In fact, at least five TRAIL receptors belonging to the apoptosis-inducing TNF-receptor family have been described so far. TRAIL-R1 (death receptor DR4; TNFRSF10a) (Pan et al., 1997a) and TRAIL-R2 (DR5; TNFRSF10b) (Pan et al. 1997b) transduce apoptotic (Ozören & El-Deiry, 2003; Huang & Sheikh, 2007) as well as non-apoptotic signals (Di Pietro & Zauli, 2004; Park et al., 2005) upon TRAIL binding, while TRAIL-R3 (DcR1; TNFRSF10c) (Pan et al., 1997b) and TRAIL-R4 (DcR2; TNFRSF10d) (Marsters et al., 1997) as well as osteoprotegerin (OPG; TRAIL-R5; TNFRSF11b) (Emery et al., 1998) are homologous to TRAIL-R1 and TRAIL-R2 in their cysteine-rich extracellular domain but they lack the intracellular death domain (DD)

TRAIL-R1 (Pan et al., 1997a) and TRAIL-R2 (Pan et al. 1997b) are type I trans-membrane proteins exposing the N-terminal TRAIL-binding domain and exerting pro-apoptotic signals through the cytoplasmic death domain. The cytoplasmic domain of TRAIL-R1/R2 shares a significant homology to the DD of different death receptors (DRs), such as CD95 and TNF-R1. Upon TRAIL binding to appropriate cognate receptors (TRAIL-R1 and TRAIL-R2), there is aggregation of TRAIL-Rs on the cell surface followed by the activation of both extrinsic and intrinsic intracellular death signalling pathways (Cretney et al., 2007). Shortly after addition of the ligand, the death-inducing signalling complex (DISC) is assembled (Kischkel et al., 1995). TRAIL DISC resembles that of Fas since the adaptor protein Fas associated

**2.1 The TRAIL receptor family and regulation** 

**2.1.1 TRAIL death receptors** 

and apoptosis inducing capability (Almasan & Ashkenazi, 2003).

death domain (FADD) and the apoptosis initiator caspase-8 are recruited to TRAIL-R1 and/or TRAIL-R2 (Sprick et al., 2000). Although initial studies have attributed a central role to caspase-8 in mediating the apoptotic signal of TRAIL, subsequent studies have demonstrated that apoptosis can be triggered independently through TRAIL-R1 or TRAIL-R2 and proteolytic activation of effector caspases either directly by apical caspase-8 or -10 (Kischkel et al., 2001) and/or indirectly through Apaf-1-mediated activation of caspase-9 (Green, 2000). Similarly to CD95L, the response to TRAIL is cell-type specific and might be characterized by two distinct cell death pathways (LeBlanc & Ashkenazi, 2003). In the type I extrinsic pathway, extrinsic signals lead to the activation of large amounts of caspase-8 and to the rapid cleavage of executioner caspase-3 prior to loss of mitochondria trans-membrane potential. As a consequence, in type I cells (including leukaemia cells) Bcl-2 over-expression blocks the mitochondrial changes associated with cell death but does not prevent apoptosis that occurs upon death receptors activation. Recent studies have suggested that death receptor induction of apoptosis may depend on the degree of receptor aggregation/multimerization, which may, in turn, depend on the concentration of the death ligand, its form (i.e. soluble versus membrane-bound), the relative DR expression on the cell surface and the array of growth factors and cytokines to which the cells are exposed (Abdulghani & El-Deiry, 2010; Mellier et al., 2010). In the type II intrinsic pathway of apoptosis, intrinsic signals, like DNA damage, growth factor withdrawal or cytokine deprivation, affect the function of Bcl-2 family members (Roos & Kaina, 2006). In fact, in type II cells the extrinsic pathway activated by death receptors is ineffective to recruit, at the DISC level, enough caspase-8 to activate effector caspases. However, through homotypic aggregation at the DISC, caspase-8 is stabilized in an active form and released into the cytosol, where it cleaves its target proteins, most notably the pro-apoptotic Bcl-2 homology domain (BH3)-only protein Bid (BH3-interacting-domain death agonist) (Kelley & Ashkenazi, 2004), thus connecting the "intrinsic" mitochondrial pathway to the "extrinsic" DR pathway (Sprick & Walczak, 2004). In turn, truncated Bid (tBid) is able to bind antiapoptotic Bcl-2 family members like Bcl-2, Bcl-XL, Bcl-W and A1 allowing the pro-apoptotic Bcl-2 family members Bax and Bak to engage the mitochondria and induce the release of mitochondrial cytochrome c and Smac (second mitochondria-derived activator of caspases)/DIABLO into the cytosol, where these latter factors promote caspase activation. Actually, cytochrome c forms the "apoptosome" complex with the adaptor protein Apaf-1 resulting in the activation of the apoptosis-initiating protease caspase-9, which then stimulates effector caspases (Green, 2000). Instead, Smac/DIABLO binds to inhibitors of apoptosis proteins (IAPs), preventing their negative-regulatory binding to caspase-9 and -3 and then augmenting apoptosis induction (Salvesen & Abrams, 2004). In this scenario, oncogenic mutations affecting molecules involved in the intrinsic mitochondrial pathway might cause resistance emergence in type II cells, while mutations in the DR pathway could confer resistance to DR-dependent apoptosis especially in type I cells (Fig. 1).

It has been demonstrated that the expression pattern of the two killer receptors is broad and partly overlapping, suggesting that they may serve as an alternate or "backup" system, allowing the immune system to control aberrant cells even if one of the two receptors had failed (Greil et al., 2003). Although further investigations are needed to assess differences between DR signalling and regulation, some interesting observations have been reported so far. For example, it has been shown that TRAIL-R1 is activated both by the soluble and the membrane-bound form of the ligand (MacFarlane et al., 2005) whereas TRAIL-R2 is activated by cross-linked soluble and membrane-bound TRAIL ligand but not by the soluble non-cross-

Signalling Pathways Leading to TRAIL Resistance 205

lines to TRAIL is highly variable, with some cell lines demonstrating a marked resistance. In this respect, the large majority of primary haematological tumours are TRAIL-resistant, basically due to the activation of anti-apoptotic signalling pathways (such as Akt and NF- B), over-expression of anti-apoptotic proteins (such as FLIP, Bcl-2, X-IAP), reduced expression of TRAIL death receptors or increased expression of decoy receptors (Testa, 2010;

Defects in either of the different molecules involved in TRAIL signalling can lead to TRAIL resistance. With regard to the relationship between TRAIL-receptors expression and TRAILresistance, the data reported in the literature are often contradictory (Russo et al., 2010). An interesting characteristic in the family of DRs is that normal cells are TRAIL resistant, but the molecular basis for TRAIL tumour selectivity is still unclear. In fact, as many chemotherapeutic drugs, TRAIL is not universally active against tumour cell lines, especially primary tumour cells, even expressing DRs on their surface. Dysfunctions of TRAIL-R1/R2 due to oncogenic mutations have been found in different tumours and in different cancer patients (breast, lung, head and neck cancer and non-Hodgkin lymphoma) (Lee et al., 1999; 2001). In particular, most tumours mutations map the intracellular domain of TRAIL-R2 that binds the adaptor protein FADD (Shin et al., 2001), known for being essential together with caspase-8 for the DISC assembly. It has been reported that posttranslational modifications, such as O-glycosylation, at the receptor level are essential for TRAIL-R1/R2 full functionality (Wagner et al., 2007), since protein glycosylation could enhance ligand-mediated receptor clustering. Thus, the glycosylation status of TRAIL-

R1/R2 has been proposed as a marker of TRAIL sensitivity (Russo et al., 2010).

TRAIL resistance has also been linked to the presence of decoy receptors, although their physiological role as well as their impact on normal and cancer cells signalling is still poorly understood. It has been demonstrated that under experimental conditions decoy receptors over-expression could sequester TRAIL, decreasing the functional binding to TRAIL-R1/R2 and attenuating the apoptotic signalling, but experiments under physiological conditions are still missing (Russo et al., 2010). Initially, the preferential expression of TRAIL-R3/-R4 mRNA in normal cells, including peripheral blood lymphocytes, spleen and thymus, was related to the absence of TRAIL cytotoxicity in normal cells, but subsequent studies, using specific monoclonal antibodies (MoAbs), demonstrated that TRAIL sensitivity was not correlated with the relative expression of TRAIL death or decoy receptors (Griffith et al., 1999). It is now accepted that the simple expression of a death or decoy receptor is not an

Another molecular mechanism responsible for TRAIL resistance emergence is considered the constitutive activation of pro-survival pathways, including Akt pathway. Akt, also known as protein kinase B (PKB), is a serine/threonine kinase that acts as a transducer of many functions initiated by the growth factor receptors that activate phosphatidylinositol 3 kinase (PI3K). In this respect, Akt and PTEN (phosphatase and tensin homologue deleted on chromosome 10) constitutive phosphorylation has been linked to TRAIL resistance of acute lymphoblastic leukaemia (ALL) cell lines (Didaa et al., 2008) and acute myeloid leukaemia (AML) patient blasts (Martelli et al., 2006). Interestingly, TRAIL itself can paradoxically

Mellier et al., 2010) (Fig. 1).

essential feature for apoptosis sensitivity.

**3.2 PI3K/Akt, MAPK, c-FLIP** 

**3.1 TRAIL-receptors** 

linked ligand (Wajant et al., 2001). We previously found a selective radiation-induced upregulation of TRAIL-R1 in different cell lines of haematological origin that sensitized cells to TRAIL cytotoxic activity (Di Pietro et al., 2001), whereas other investigators have suggested a key role for TRAIL-R2 in p53-dependent apoptosis in response to DNA damage both *in vitro* and *in vivo* (Burns et al., 2001). Similarly, TRAIL-R2 was up-regulated in B-CLL (chronic lymphocytic leukaemia) cells in response to the small molecule Nutlin-3 in a p53-dependent manner (Coll-Mulet et al., 2006). By means of receptor-selective TRAIL mutant ligands, MacFarlane et al. (2005) demonstrated that CLL cells signal to apoptosis primarily through TRAIL-R1, whereas cross-linked agonistic TRAIL-R2 antibodies facilitate signalling via TRAIL-R2 (Natoni et al., 2007). Other authors have shown different responses of DRs in their cytoplasmic domains that may account for the differences in the activation of these receptors (Thomas et al., 2004). These authors postulated that the binding of ligand and agonistic antibody to the extracellular domain exposes the FADD-binding region differently in the cytoplasmic domain of TRAIL-R1 and TRAIL–R2 to enhance caspase-8 binding and cleavage while promoting recruitment of ancillary proteins. In addition, one more recent paper indicates TRAIL-R2 as a mediator of anoikis of colorectal carcinoma cell lines through the preferential recruitment of the "extrinsic" pathway of apoptosis (Laguinge et al., 2008).

#### **2.1.2 TRAIL decoy receptors**

Unlike TRAIL death receptors, TRAIL-R3 and TRAIL-R4 have been originally proposed as "decoy" receptors able to inhibit apoptosis by sequestering TRAIL from the death-inducing TRAIL-Rs or by aggregating with TRAIL death receptors upon binding to the trimeric ligand (Ashkenazi & Dixit, 1999). Nevertheless, it has been demonstrated in primary human CD8+ T cells that the inhibition of TRAIL-induced apoptosis by TRAIL-R4 critically depends on its ligand-independent association with TRAIL-R2 via the NH2-terminal preligand assembly domain overlapping the first partial cysteine-rich domain of both receptors (Clancy et al., 2005). In addition, it has been shown that expression of TRAIL-R1 and/or TRAIL-R2 was necessary but not always sufficient to mediate apoptosis, while expression of TRAIL-R3 and/or TRAIL-R4 often did not correlate with normal and tumour cells resistance to TRAIL effects (Di Pietro & Zauli, 2004). In fact, a number of evidences, based on the use of TRAIL-R agonists rather than over-expression models, have pointed at the primary role of intracellular mechanisms in controlling TRAIL resistance in a number of cell types (Griffith & Linch, 1998; Leverkus et al., 2000), thus cutting down the importance of control at decoy receptors level, whose ability to modulate TRAIL-mediated apoptosis is still controversial. In this regard, particularly intriguing is the role of soluble OPG (Corallini et al., 2008; Secchiero & Zauli, 2008). OPG was initially characterized for its ability to inhibit RANKLstimulated osteoclastogenesis (Emery et al., 1998), but more recent studies highlighted the capability of OPG to counteract the pro-apoptotic activity of TRAIL in a variety of neoplastic cell types, at least *in vitro* (Schaefer et al., 2007). Of note, the interplay between OPG, RANKL and TRAIL is an important issue in bone and bone marrow biology as well as in the physiopathology of haematological malignancies and in particular of multiple myeloma (MM) (Secchiero & Zauli, 2008).

#### **3. Molecular mechanisms of TRAIL resistance**

The response of individual leukaemia cell lines to TRAIL may depend on which of proapoptotic or pro-survival pathways is dominant. In fact, the sensitivity of leukaemia cell lines to TRAIL is highly variable, with some cell lines demonstrating a marked resistance. In this respect, the large majority of primary haematological tumours are TRAIL-resistant, basically due to the activation of anti-apoptotic signalling pathways (such as Akt and NF- B), over-expression of anti-apoptotic proteins (such as FLIP, Bcl-2, X-IAP), reduced expression of TRAIL death receptors or increased expression of decoy receptors (Testa, 2010; Mellier et al., 2010) (Fig. 1).

#### **3.1 TRAIL-receptors**

204 Advances in Cancer Therapy

linked ligand (Wajant et al., 2001). We previously found a selective radiation-induced upregulation of TRAIL-R1 in different cell lines of haematological origin that sensitized cells to TRAIL cytotoxic activity (Di Pietro et al., 2001), whereas other investigators have suggested a key role for TRAIL-R2 in p53-dependent apoptosis in response to DNA damage both *in vitro* and *in vivo* (Burns et al., 2001). Similarly, TRAIL-R2 was up-regulated in B-CLL (chronic lymphocytic leukaemia) cells in response to the small molecule Nutlin-3 in a p53-dependent manner (Coll-Mulet et al., 2006). By means of receptor-selective TRAIL mutant ligands, MacFarlane et al. (2005) demonstrated that CLL cells signal to apoptosis primarily through TRAIL-R1, whereas cross-linked agonistic TRAIL-R2 antibodies facilitate signalling via TRAIL-R2 (Natoni et al., 2007). Other authors have shown different responses of DRs in their cytoplasmic domains that may account for the differences in the activation of these receptors (Thomas et al., 2004). These authors postulated that the binding of ligand and agonistic antibody to the extracellular domain exposes the FADD-binding region differently in the cytoplasmic domain of TRAIL-R1 and TRAIL–R2 to enhance caspase-8 binding and cleavage while promoting recruitment of ancillary proteins. In addition, one more recent paper indicates TRAIL-R2 as a mediator of anoikis of colorectal carcinoma cell lines through the

preferential recruitment of the "extrinsic" pathway of apoptosis (Laguinge et al., 2008).

Unlike TRAIL death receptors, TRAIL-R3 and TRAIL-R4 have been originally proposed as "decoy" receptors able to inhibit apoptosis by sequestering TRAIL from the death-inducing TRAIL-Rs or by aggregating with TRAIL death receptors upon binding to the trimeric ligand (Ashkenazi & Dixit, 1999). Nevertheless, it has been demonstrated in primary human CD8+ T cells that the inhibition of TRAIL-induced apoptosis by TRAIL-R4 critically depends on its ligand-independent association with TRAIL-R2 via the NH2-terminal preligand assembly domain overlapping the first partial cysteine-rich domain of both receptors (Clancy et al., 2005). In addition, it has been shown that expression of TRAIL-R1 and/or TRAIL-R2 was necessary but not always sufficient to mediate apoptosis, while expression of TRAIL-R3 and/or TRAIL-R4 often did not correlate with normal and tumour cells resistance to TRAIL effects (Di Pietro & Zauli, 2004). In fact, a number of evidences, based on the use of TRAIL-R agonists rather than over-expression models, have pointed at the primary role of intracellular mechanisms in controlling TRAIL resistance in a number of cell types (Griffith & Linch, 1998; Leverkus et al., 2000), thus cutting down the importance of control at decoy receptors level, whose ability to modulate TRAIL-mediated apoptosis is still controversial. In this regard, particularly intriguing is the role of soluble OPG (Corallini et al., 2008; Secchiero & Zauli, 2008). OPG was initially characterized for its ability to inhibit RANKLstimulated osteoclastogenesis (Emery et al., 1998), but more recent studies highlighted the capability of OPG to counteract the pro-apoptotic activity of TRAIL in a variety of neoplastic cell types, at least *in vitro* (Schaefer et al., 2007). Of note, the interplay between OPG, RANKL and TRAIL is an important issue in bone and bone marrow biology as well as in the physiopathology of haematological malignancies and in particular of multiple myeloma

The response of individual leukaemia cell lines to TRAIL may depend on which of proapoptotic or pro-survival pathways is dominant. In fact, the sensitivity of leukaemia cell

**2.1.2 TRAIL decoy receptors** 

(MM) (Secchiero & Zauli, 2008).

**3. Molecular mechanisms of TRAIL resistance** 

Defects in either of the different molecules involved in TRAIL signalling can lead to TRAIL resistance. With regard to the relationship between TRAIL-receptors expression and TRAILresistance, the data reported in the literature are often contradictory (Russo et al., 2010). An interesting characteristic in the family of DRs is that normal cells are TRAIL resistant, but the molecular basis for TRAIL tumour selectivity is still unclear. In fact, as many chemotherapeutic drugs, TRAIL is not universally active against tumour cell lines, especially primary tumour cells, even expressing DRs on their surface. Dysfunctions of TRAIL-R1/R2 due to oncogenic mutations have been found in different tumours and in different cancer patients (breast, lung, head and neck cancer and non-Hodgkin lymphoma) (Lee et al., 1999; 2001). In particular, most tumours mutations map the intracellular domain of TRAIL-R2 that binds the adaptor protein FADD (Shin et al., 2001), known for being essential together with caspase-8 for the DISC assembly. It has been reported that posttranslational modifications, such as O-glycosylation, at the receptor level are essential for TRAIL-R1/R2 full functionality (Wagner et al., 2007), since protein glycosylation could enhance ligand-mediated receptor clustering. Thus, the glycosylation status of TRAIL-R1/R2 has been proposed as a marker of TRAIL sensitivity (Russo et al., 2010).

TRAIL resistance has also been linked to the presence of decoy receptors, although their physiological role as well as their impact on normal and cancer cells signalling is still poorly understood. It has been demonstrated that under experimental conditions decoy receptors over-expression could sequester TRAIL, decreasing the functional binding to TRAIL-R1/R2 and attenuating the apoptotic signalling, but experiments under physiological conditions are still missing (Russo et al., 2010). Initially, the preferential expression of TRAIL-R3/-R4 mRNA in normal cells, including peripheral blood lymphocytes, spleen and thymus, was related to the absence of TRAIL cytotoxicity in normal cells, but subsequent studies, using specific monoclonal antibodies (MoAbs), demonstrated that TRAIL sensitivity was not correlated with the relative expression of TRAIL death or decoy receptors (Griffith et al., 1999). It is now accepted that the simple expression of a death or decoy receptor is not an essential feature for apoptosis sensitivity.

#### **3.2 PI3K/Akt, MAPK, c-FLIP**

Another molecular mechanism responsible for TRAIL resistance emergence is considered the constitutive activation of pro-survival pathways, including Akt pathway. Akt, also known as protein kinase B (PKB), is a serine/threonine kinase that acts as a transducer of many functions initiated by the growth factor receptors that activate phosphatidylinositol 3 kinase (PI3K). In this respect, Akt and PTEN (phosphatase and tensin homologue deleted on chromosome 10) constitutive phosphorylation has been linked to TRAIL resistance of acute lymphoblastic leukaemia (ALL) cell lines (Didaa et al., 2008) and acute myeloid leukaemia (AML) patient blasts (Martelli et al., 2006). Interestingly, TRAIL itself can paradoxically

Signalling Pathways Leading to TRAIL Resistance 207

Baltimore, 1996). In response to many stimuli, such as TNF, lipopolysaccharide (LPS) or interleukin-1 (IL-1), IB kinase (IKK) is activated and can phosphorylate IBs, which, in turn, can be poly-ubiquitinated and rapidly degraded by the proteasome, allowing the release of sequestered NF-B. After its translocation into the nucleus, NF-B is able to activate its target genes, which, depending on the physiological circumstances (Barkett & Gilmore, 1999), can mediate cell survival or apoptosis. It has been reported, for example, that the TRAIL-mediated recruitment of apical caspase-8/-10 was able to induce the simultaneous activation of both effector caspases (-3, -6, -7) and of NF-B pathway in TRAIL-sensitive myeloid leukaemia cells (Secchiero et al., 2002). As a consequence, the TRAIL cytotoxic/cytostatic activity mediated by effector caspases was reduced by the concomitant pro-survival effect exerted by NF-B. Interestingly, NF-B activation was causally linked to the induction of maturation of the surviving leukaemia cells along the monocytic pathway (Secchiero et al., 2002). Moreover, NF- B activation was paralleled by the absence of degradation and by the nuclear translocation of I in Jurkat T leukaemia cell lines sensitive to the cytotoxic action of TRAIL (Zauli et al., 2005), whereas in TRAIL-resistant primary human erythroblasts NF-B activation was concomitant with IB cytoplasmic localization (unpublished observations). The dual function of NF-B, as an inhibitor or activator of apoptosis, would depend on the relative levels of RelA and c-Rel subunits (Chen et al., 2003). In fact, over-expression of RelA or a transcriptionaldeficient mutant of c-Rel inhibits TRAIL-induced apoptosis in mouse embryonic fibroblasts,

whereas depletion of RelA increases cytokine-induced apoptosis (Chen et al., 2003).

inactivate Akt as well as other anti-apoptotic molecules (Milani et al., 2003).

cancer patients to TRAIL-based therapy.

All these findings underline the complexity of the TRAIL-mediated intracellular signals, which simultaneously activate pro-apoptotic and anti-apoptotic pathways. The fate of individual malignant cells would depend on which of these pathways prevail within the cell. These effects should be carefully evaluated in the individual assessment of eligibility of

Among the anti-apoptotic genes up-regulated by NF-B are included cellular inhibitors of apoptosis proteins 1 and 2 (c-IAP1 and c-IAP2), TRAF1 and TRAF2, c-FLIP and Bcl-XL (Wang et al., 1998). Some of these (e.g. survivin, X-IAP, Bcl-2, Bcl-XL) have been shown to be associated with poor prognosis in AML (Tamm et al., 2000; Paydas et al. 2003). Overexpression of Bcl-2, Bcl-XL, or Mcl-1, loss of Bax or Bak function, increased expression of IAPs and reduced release of Smac/DIABLO from the mitochondria to the cytosol are all events resulting in TRAIL resistance in type II cancer cells (Vogler et al., 2008)(Fig. 1) . Another important mechanism through which haematological malignancies can escape TRAIL cytotoxicity is inactivation of the intracellular pro-apoptotic pathways (Ashkenazi et al., 2008). This allows malignant cells not only to escape from TRAIL-induced apoptosis, but also to take advantage of the pro-survival signals induced by TRAIL, which paradoxically may act as a survival cytokine (Fig. 1). In this respect, a possible interplay between the Akt and caspase pathways has been already described in several cell systems (Cardone et al., 1998; Jones et al., 2002; Milani et al., 2003). The picture emerging from these studies is that, when survival signals dominate, Akt impairs the activation of the apical caspases, by directly phosphorylating caspase-9 (Cardone et al., 1998) or by inhibiting the recruitment of procaspase-8/-10 to the DISC (Jones et al., 2002). On the other hand, when pro-apoptotic signals prevail, apical caspase-8/-10 activates downstream caspases, which cleave and

**3.4 IAP, Bcl-2, caspases** 

activate Akt and downstream targets, like CREB/ATF transcription factors, in leukaemia cells sensitive to its cytotoxic action, loosing to some extent its pro-apoptotic effect (Zauli et al., 2005; Caravatta et al., 2008). Constitutively active Akt is an important regulator of TRAIL sensitivity in prostate cancer (Chen et al., 2001) and protects HL60 leukaemia cells from TRAIL-induced apoptosis by activating the transcriptional factor NF-B and up-regulating c-FLIP (cellular FADD-like interleukin-1beta-converting enzyme-inhibitory protein) (Bortul et al., 2003). c-FLIP acts as an important intracellular inhibitor of TRAIL sensitivity and delivers growth signals by activating NF-B and ERK signalling pathways (Almasan & Ashkenazi, 2003). It is structurally related to caspase-8 (but is devoid of enzymatic activity) since it possesses two death domains that facilitate the binding to the death domain of FADD, thereby preventing association of caspase-8 with the DISC. Over-expression of c-FLIP correlates with TRAIL resistance in several types of cancer, especially in type I cells (MacFarlane et al., 2002). Although it is considered as one of the non-apoptotic NF-B target genes, downwards the anti-apoptotic PKB/Akt and MAPK pathways, it has not been confirmed as a molecular switch between life and death in the same cell (Park et al., 2005). More controversial is the TRAIL-R1/R2-induced activation of PKB/Akt and MAP kinases (Falschlehner et al., 2007). The three key enzymes of MAPK pathway, i.e. ERK, p38-MAPK and JNK, have been often associated with the anti-apoptotic function of TRAIL receptors. Paradoxically, pro-apoptotic effects connect TRAIL and MAPK pathways in different cell models (Frese et al., 2003; Jurewicz et al., 2006). The existence of a death domain alternative DISC complex has also been suggested to explain the TRAIL-dependent activation of MAPK pathways. FADD, caspase-8, RIP (receptor-interacting protein) and TRAF2 (TNF receptorassociated factor 2) might form a cytoplasmic complex upon TRAIL stimulation leading to p38 and JNK activation (Lin et al. 2000). In HUVECs (human umbilical vein endothelial cells) (Zauli et al., 2003) and synovial fibroblasts (Morel et al., 2005), TRAIL has a direct effect on cell survival and proliferation stimulating the PI3K-dependent phosphorylation and activation of PKB/Akt kinase without activation of NF-B. Of note, the activation of an ERK-dependent pathway has been linked to TRAIL-induced maturation of erythroid cells (Secchiero et al., 2004a).

#### **3.3 NF-B/ IB**

A number of studies of different groups of investigators, including ourselves, have outlined the importance of NF-B activation in determining the resistance/susceptibility of target cells to TRAIL cytotoxicity (Ehrhardt et al., 2003; Zauli et al., 2005), possibly by modulating c-FLIP levels (Bortul et al., 2003). Mounting experimental evidences highlight in TRAIL-resistant cells the activation of NF-B following engagement of TRAIL-R1, -R2, or -R4 (Zauli et al. 2005; Henson et al., 2008). TRAIL activation of NF-B is mediated via TRADD (TNF-R1-associated death domain protein), TRAF2 and RIP and occurs independently of caspase-8/-10 activation (Mühlenbeck et al., 1998; MacFarlane, 2003) (Fig. 1). Importantly, the level of NF-B activation has been related to resistance of leukaemia (Ehrhardt et al., 2003) and neuroblastoma cell lines (Yang & Thiele, 2003) to TRAIL cytotoxicity and its aberrant activation has been involved in promoting tumour migration and dissemination. These findings are consistent with the pleiotropic activity of NF-B transcription factors, which are implicated in the control of cell survival and tumorigenesis (Rayet & Gelinas, 1999). Activation and regulation of Rel/NF-B proteins are tightly controlled by IB proteins, which mask the nuclear localization signal (NLS) of NF-B family members, thereby preventing their nuclear translocation (Baeuerle &

activate Akt and downstream targets, like CREB/ATF transcription factors, in leukaemia cells sensitive to its cytotoxic action, loosing to some extent its pro-apoptotic effect (Zauli et al., 2005; Caravatta et al., 2008). Constitutively active Akt is an important regulator of TRAIL sensitivity in prostate cancer (Chen et al., 2001) and protects HL60 leukaemia cells from TRAIL-induced apoptosis by activating the transcriptional factor NF-B and up-regulating c-FLIP (cellular FADD-like interleukin-1beta-converting enzyme-inhibitory protein) (Bortul et al., 2003). c-FLIP acts as an important intracellular inhibitor of TRAIL sensitivity and delivers growth signals by activating NF-B and ERK signalling pathways (Almasan & Ashkenazi, 2003). It is structurally related to caspase-8 (but is devoid of enzymatic activity) since it possesses two death domains that facilitate the binding to the death domain of FADD, thereby preventing association of caspase-8 with the DISC. Over-expression of c-FLIP correlates with TRAIL resistance in several types of cancer, especially in type I cells (MacFarlane et al., 2002). Although it is considered as one of the non-apoptotic NF-B target genes, downwards the anti-apoptotic PKB/Akt and MAPK pathways, it has not been confirmed as a molecular switch between life and death in the same cell (Park et al., 2005). More controversial is the TRAIL-R1/R2-induced activation of PKB/Akt and MAP kinases (Falschlehner et al., 2007). The three key enzymes of MAPK pathway, i.e. ERK, p38-MAPK and JNK, have been often associated with the anti-apoptotic function of TRAIL receptors. Paradoxically, pro-apoptotic effects connect TRAIL and MAPK pathways in different cell models (Frese et al., 2003; Jurewicz et al., 2006). The existence of a death domain alternative DISC complex has also been suggested to explain the TRAIL-dependent activation of MAPK pathways. FADD, caspase-8, RIP (receptor-interacting protein) and TRAF2 (TNF receptorassociated factor 2) might form a cytoplasmic complex upon TRAIL stimulation leading to p38 and JNK activation (Lin et al. 2000). In HUVECs (human umbilical vein endothelial cells) (Zauli et al., 2003) and synovial fibroblasts (Morel et al., 2005), TRAIL has a direct effect on cell survival and proliferation stimulating the PI3K-dependent phosphorylation and activation of PKB/Akt kinase without activation of NF-B. Of note, the activation of an ERK-dependent pathway has been linked to TRAIL-induced maturation of erythroid cells

A number of studies of different groups of investigators, including ourselves, have outlined the importance of NF-B activation in determining the resistance/susceptibility of target cells to TRAIL cytotoxicity (Ehrhardt et al., 2003; Zauli et al., 2005), possibly by modulating c-FLIP levels (Bortul et al., 2003). Mounting experimental evidences highlight in TRAIL-resistant cells the activation of NF-B following engagement of TRAIL-R1, -R2, or -R4 (Zauli et al. 2005; Henson et al., 2008). TRAIL activation of NF-B is mediated via TRADD (TNF-R1-associated death domain protein), TRAF2 and RIP and occurs independently of caspase-8/-10 activation (Mühlenbeck et al., 1998; MacFarlane, 2003) (Fig. 1). Importantly, the level of NF-B activation has been related to resistance of leukaemia (Ehrhardt et al., 2003) and neuroblastoma cell lines (Yang & Thiele, 2003) to TRAIL cytotoxicity and its aberrant activation has been involved in promoting tumour migration and dissemination. These findings are consistent with the pleiotropic activity of NF-B transcription factors, which are implicated in the control of cell survival and tumorigenesis (Rayet & Gelinas, 1999). Activation and regulation of Rel/NF-B proteins are tightly controlled by IB proteins, which mask the nuclear localization signal (NLS) of NF-B family members, thereby preventing their nuclear translocation (Baeuerle &

(Secchiero et al., 2004a).

**3.3 NF-B/ IB** 

Baltimore, 1996). In response to many stimuli, such as TNF, lipopolysaccharide (LPS) or interleukin-1 (IL-1), IB kinase (IKK) is activated and can phosphorylate IBs, which, in turn, can be poly-ubiquitinated and rapidly degraded by the proteasome, allowing the release of sequestered NF-B. After its translocation into the nucleus, NF-B is able to activate its target genes, which, depending on the physiological circumstances (Barkett & Gilmore, 1999), can mediate cell survival or apoptosis. It has been reported, for example, that the TRAIL-mediated recruitment of apical caspase-8/-10 was able to induce the simultaneous activation of both effector caspases (-3, -6, -7) and of NF-B pathway in TRAIL-sensitive myeloid leukaemia cells (Secchiero et al., 2002). As a consequence, the TRAIL cytotoxic/cytostatic activity mediated by effector caspases was reduced by the concomitant pro-survival effect exerted by NF-B. Interestingly, NF-B activation was causally linked to the induction of maturation of the surviving leukaemia cells along the monocytic pathway (Secchiero et al., 2002). Moreover, NF- B activation was paralleled by the absence of degradation and by the nuclear translocation of I in Jurkat T leukaemia cell lines sensitive to the cytotoxic action of TRAIL (Zauli et al., 2005), whereas in TRAIL-resistant primary human erythroblasts NF-B activation was concomitant with IB cytoplasmic localization (unpublished observations). The dual function of NF-B, as an inhibitor or activator of apoptosis, would depend on the relative levels of RelA and c-Rel subunits (Chen et al., 2003). In fact, over-expression of RelA or a transcriptionaldeficient mutant of c-Rel inhibits TRAIL-induced apoptosis in mouse embryonic fibroblasts, whereas depletion of RelA increases cytokine-induced apoptosis (Chen et al., 2003).

#### **3.4 IAP, Bcl-2, caspases**

Among the anti-apoptotic genes up-regulated by NF-B are included cellular inhibitors of apoptosis proteins 1 and 2 (c-IAP1 and c-IAP2), TRAF1 and TRAF2, c-FLIP and Bcl-XL (Wang et al., 1998). Some of these (e.g. survivin, X-IAP, Bcl-2, Bcl-XL) have been shown to be associated with poor prognosis in AML (Tamm et al., 2000; Paydas et al. 2003). Overexpression of Bcl-2, Bcl-XL, or Mcl-1, loss of Bax or Bak function, increased expression of IAPs and reduced release of Smac/DIABLO from the mitochondria to the cytosol are all events resulting in TRAIL resistance in type II cancer cells (Vogler et al., 2008)(Fig. 1) . Another important mechanism through which haematological malignancies can escape TRAIL cytotoxicity is inactivation of the intracellular pro-apoptotic pathways (Ashkenazi et al., 2008). This allows malignant cells not only to escape from TRAIL-induced apoptosis, but also to take advantage of the pro-survival signals induced by TRAIL, which paradoxically may act as a survival cytokine (Fig. 1). In this respect, a possible interplay between the Akt and caspase pathways has been already described in several cell systems (Cardone et al., 1998; Jones et al., 2002; Milani et al., 2003). The picture emerging from these studies is that, when survival signals dominate, Akt impairs the activation of the apical caspases, by directly phosphorylating caspase-9 (Cardone et al., 1998) or by inhibiting the recruitment of procaspase-8/-10 to the DISC (Jones et al., 2002). On the other hand, when pro-apoptotic signals prevail, apical caspase-8/-10 activates downstream caspases, which cleave and inactivate Akt as well as other anti-apoptotic molecules (Milani et al., 2003).

All these findings underline the complexity of the TRAIL-mediated intracellular signals, which simultaneously activate pro-apoptotic and anti-apoptotic pathways. The fate of individual malignant cells would depend on which of these pathways prevail within the cell. These effects should be carefully evaluated in the individual assessment of eligibility of cancer patients to TRAIL-based therapy.

Signalling Pathways Leading to TRAIL Resistance 209

Moreover, TRAIL exerts a variable cytotoxic activity on haematological malignancies (Snell et al., 1997) and synergistically cooperates with: i) chemotherapeutic drugs, such as etoposide, campthotecin-11, doxorubicin, 5-fluorouracil, taxol (Sabatini et al., 2004; Henson et al., 2008); and ii) ionizing radiation (Chinnaiyan et al., 2000; Di Pietro et al., 2001), causing substantial regression or complete ablation of solid (colon and mammary) cancers in animal models. Besides acting as a tumour suppressor *in vivo* in primary tumours, TRAIL could play a substantial role in suppressing tumour metastasis. In fact, it has been observed that this cytokine may partially limit the formation of hepatic metastases of a variety of mouse tumours (Seki et al., 2003). A study performed in TRAIL -/- null mice demonstrated that the incidence of spontaneous lymphoid malignancies was increased by 25% in comparison with control animals (Zerafa et al., 2005), suggesting a crucial role of TRAIL in the immunesurveillance against lymphoid malignancies. Although it is not established whether TRAIL causes liver toxicity in humans (Jo et al., 2000; Lawrence et al., 2001), pre-clinical studies performed in mice and non-human primates indicated that rTRAIL protein promotes potent apoptosis-inducing activity against tumour cells without a relevant systemic toxicity (Walczak et al., 1999). Phase I and phase II clinical trials in patients with advanced solid tumours or non-Hodgkin lymphoma (NHL) appeared to go in the same direction, indicating that both rTRAIL and TRAs are safe and well tolerated (Koschny et al., 2007; Tolcher et al., 2007). Therefore, TRAIL ligand and TRAs are strong candidates for an effective but tolerable treatment of solid cancers, either used alone or in combination with radio-chemotherapy.

The cytotoxic activity of TRAIL has been evaluated in haematological diseases by different groups of investigators, including our research group (Secchiero & Zauli, 2008; Sancilio et al., 2008; Impicciatore et al., 2010). Overall the activity of TRAIL as a single treatment in acute and chronic leukaemia is poor. Unlike the poor outcome of TRAIL treatment in primary AML blasts, continuous cell lines derived from AML display a pronounced sensitivity to the apoptotic action of TRAIL (Snell et al., 1997; Secchiero et al., 2004b). Moreover, when TRAIL is used in combination with chemotherapeutic agents (fludarabine, cytosine arabinoside or daunorubicin) additive or super-additive apoptotic effects are obtained, due to the ability of these agents to activate apical caspase-8/-10 (Jones et al., 2003). In line with these findings, other authors demonstrated that triterpenoids, natural and synthetic compounds with demonstrated anti-tumour activity, induced a substantial increase in cell death in both B-CLL and AML blasts, by inducing a concentrationdependent decrease in the levels of FLIP protein (Suh et al., 2003; Pedersen et al., 2002). A recent report has related the poor response of AML to the simultaneous expression of death and decoy receptors (Inukai et al., 2006), whereas co-expression of death receptors with the decoy receptor TRAIL-R3 resulted in significant shortened overall survival of AML patients (Chamuleau et al., 2011). Another weak point in leukaemia treatment is represented by p53 gene deletions or mutations, that usually occur in less than 15% of AML cases. To augment the poor response of AML to TRAIL cytotoxicity, Secchiero et al. (2007) have recently adopted the strategy to combine rTRAIL with Nutlin-3, a potent non-genotoxic activator of the p53 pathway (Impicciatore et al., 2010). In this investigation Nutlin-3 synergized with TRAIL in inducing apoptosis both in AML cell lines and primary M4-type and M5-type AML blasts, but not in mutp53 AML cells, suggesting that the combined treatment of Nutlin-

3 plus TRAIL might offer a novel therapeutic strategy for wtp53 AML cells.

**4.1 Acute myeloid leukaemia (AML)** 

Fig. 1. Schematic representation of key mechanisms involved in TRAIL resistance of haematological malignancies. Activation of TRAIL-Rs can trigger both death and survival pathways, depending on the cell system and environmental conditions. TRAIL-R1 and TRAIL-R2 can lead to apoptotic cell death through the recruitment of FADD and the following cleavage of caspase-8 and –10. Both DRs together with the "decoy" TRAIL-R4 are also involved in the priming of survival genes through the activation of a) NF-B and JNK pathways triggered by the engagement of TRAF2 and RIP; b) PI3K/Akt and MAPK/ERK1/2 pathways, by means of still unclear mechanisms (highlighted with a question mark). Other mechanisms leading to TRAIL resistance include different caspase or PI3K physiological inhibitors.
