TRAIL signals to apoptosis in chronic lymphocytic leukaemia cells primarily through TRAIL-R1 whereas cross-linked agonistic TRAIL-R2 antibodies facilitate signalling via TRAIL-R2

Authors


Gerald M. Cohen, MRC Toxicology Unit, Hodgkin Building, University of Leicester, PO Box 138, Lancaster Road Leicester, LE1 9HN, UK. E-mail: gmc2@le.ac.uk

Summary

Tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), a member of the TNF family, which is being developed as an anti-tumour agent due to its selective toxicity to tumour cells, induces apoptosis by binding to two membrane-bound receptors, TRAIL-R1 and TRAIL-R2. Clinical trials have been initiated with various preparations of TRAIL as well as agonistic monoclonal antibodies to TRAIL-R1 and TRAIL-R2. Previously we reported that prior treatment of primary chronic lymphocytic leukaemia (CLL) cells with histone deacetylase inhibitors was required to sensitize CLL cells to TRAIL and, using various receptor-selective TRAIL mutant ligands, we demonstrated that CLL cells signalled to apoptosis primarily through TRAIL-R1. Some, but not all, agonistic TRAIL-receptor antibodies require cross-linking in order to induce apoptosis. The present study demonstrated that CLL cells can signal to apoptosis through the TRAIL-R2 receptor, but only after cross-linking of the agonistic TRAIL-R2 antibodies, LBY135 and lexatumumab (HGS-ETR2). In contrast, signalling through TRAIL-R1 by receptor-selective ligands or certain agonistic antibodies, such as mapatumumab (HGS-ETR1), occurs in the absence of cross-linking. These results further highlight important differences in apoptotic signalling triggered through TRAIL-R1 and TRAIL-R2 in primary tumour cells. Such information is clearly important for the rational optimisation of TRAIL therapy in primary lymphoid malignancies, such as CLL.

Chronic lymphocytic leukaemia (CLL) is often considered a disease of failed apoptosis. Apoptosis is a major form of cell death, used to remove unwanted or excess cells. Two major pathways of apoptosis have been described that ultimately result in the activation of a family of caspases that are responsible for the biochemical and morphological changes associated with apoptosis. The intrinsic pathway involves initial mitochondrial perturbation resulting from cellular stress or cytotoxic insults, whereas the extrinsic pathway is triggered by activation of death receptors of the tumour necrosis factor (TNF) family (Sun et al, 1999; Bratton et al, 2000). Ligation of death receptors, such as CD95 (Fas/Apo1), TNF receptor-1 (TNF-R1) and the TNF-related apoptosis-inducing ligand (TRAIL) receptors-1/2 (TRAIL-R1/R2), by their cognate ligands or agonistic antibodies results in receptor aggregation and formation of a death-inducing signalling complex (DISC) resulting in activation of caspase-8, as the apical caspase (Ashkenazi & Dixit, 1998; Peter & Krammer, 2003). TRAIL is being developed as an anti-tumour agent as it induces apoptosis in a wide range of tumour cell lines but not in most normal cells and, in contrast to other members of the TNF family, such as TNF or CD95L, does not appear to cause systemic inflammatory syndrome or liver toxicity, although there has been some controversy concerning its potential hepatotoxicity (Ashkenazi, 2002; Hao et al, 2004; Koschny et al, 2007). TRAIL induces apoptosis by binding to two membrane-bound receptors, TRAIL-R1 (DR4) and TRAIL-R2 (DR5/TRICK2) (Ashkenazi, 2002). We and others have shown that primary cells from patients with CLL and B-cell non-Hodgkin lymphoma, are universally resistant to TRAIL-induced apoptosis (Snell et al, 1997; MacFarlane et al, 2002). This resistance may be overcome by prior treatment with a histone deacetylase inhibitor (HDACi) that acts by facilitating increased formation of the TRAIL DISC (Inoue et al, 2004). Inhibition of HDAC1 and HDAC2 seems to be required for sensitization to TRAIL (Inoue et al, 2006). Based on these studies we proposed that the combination of HDACi and TRAIL agonists could be of therapeutic benefit in CLL as well as other TRAIL-resistant haematological malignancies. To ascertain which form of TRAIL may be optimal in such a combination, we utilized its different forms, including soluble ligands selective for TRAIL-R1 or TRAIL-R2, as well as mapatumumab (HGS-ETR1) and lexatumumab (HGS-ETR2) (Human Genome Sciences, Rockville, MD, USA), selective agonistic antibodies for TRAIL-R1 and TRAIL-R2 respectively (Pukac et al, 2005). These two monoclonal antibodies (mAbs) were derived from single chain variable fragment reagents that were selected for their ability to induce apoptosis via their cognate receptors and do not require cross-linking for efficacy. Both CLL cells and cell lines that signal to apoptosis predominantly through either TRAIL-R1 (Ramos cells) or TRAIL-R2 (Jurkat cells) (MacFarlane et al, 2005a,b) were analysed for their sensitivity to these agents. As mapatumumab and the TRAIL-R1-selective ligands, but not lexatumumab and the TRAIL-R2-selective ligands, induced apoptosis in primary CLL cells after HDACi sensitization, we concluded that CLL cells signalled to apoptosis predominantly through TRAIL-R1 and not TRAIL-R2 despite expressing both receptors at the cell surface (MacFarlane et al, 2005a,b).

More recently we wished to extend these studies to two further TRAIL-R agonistic mAbs. Firstly, LBY135, is a potent agonistic antibody against TRAIL-R2, that exerts significant single agent activity in colon, breast and melanoma mouse xenograft models but in vitro requires cross-linking in order to induce apoptosis (Li et al, 2007). Secondly, DR4-A, a mouse mAb that binds specifically to TRAIL-R1 but which in vitro requires cross-linking to induce apoptosis. The present study demonstrated that LBY135 and DR4-A show an absolute requirement for cross-linking in order to induce caspase-dependent apoptosis. Following cross-linking, LBY135 but not DR4-A induced apoptosis in Jurkat cells whereas both antibodies could induce apoptosis in Ramos cells. Our finding that cross-linking of LBY135 resulted in the induction of apoptosis in CLL cells from some, but not all, patients raised the possibility that signalling to apoptosis through TRAIL-R2 may be facilitated by oligomerization of the receptor following cross-linking of the TRAIL-R2 agonistic antibody. Support for this hypothesis was provided by our finding that cross-linking also sensitised CLL cells to the agonistic mAb, lexatumumab. Our data highlight important differences in the ability to activate apoptosis through TRAIL-R1 and TRAIL-R2 in primary CLL cells.

Materials and methods

Materials

Media and serum were from Life Technologies, Inc. (Paisley, UK). LBY135, LBH589 and DR4-A were all kindly provided by Novartis Pharma (East Hanover, NJ, USA). LBY135 is a specific anti-TRAIL-R2 agonistic antibody, which has been generated by hybridoma technology and functional cell based screening (Li et al, 2007). Mapatumumab (HGS-ETR1) and lexatumumab (HGS-ETR2) were kindly provided by Human Genome Sciences. Human recombinant wild-type TRAIL, TRAIL-R1 (TRAIL.R1-5) and TRAIL-R2 (TRAIL.R2-6)-selective ligands were prepared as previously described (MacFarlane et al, 2005b). Low endotoxin TRAIL (TRAIL-LE) was from Alexis Corporation (Nottingham, UK), depsipeptide was kindly provided by Dr. E. Sausville (NCI, USA) and trichostatin A (TSA) was from Sigma (Poole, UK). Antibodies were sourced as follows: anti-acetylated tubulin mAb (clone 6-11B-1) was from Sigma; anti-acetylated H3 (K9/14) and H4 (K5/8/12/16) polyclonal rabbit antibodies were from Upstate Biotechnology. The antibody to poly (ADP-ribose) polymerase (PARP) (clone C2-10) was from Alexis Corp. (San Diego, CA, USA). F(ab’)2 fragment goat anti-mouse IgG Fcγ fragment specific (used for DR4-A) and F(ab’)2 fragment goat anti-human IgG Fcγ fragment specific (used for LBY135) cross-linking antibodies were from Jackson ImmunoResearch Laboratories, Inc (Baltimore, MD, USA). Anti-caspase-3 and -8 antibodies were from previously described sources (Sun et al, 1999; Inoue et al, 2004).

Lymphocyte purification, cell lines and culture

Blood samples were obtained from CLL patients with patient consent and local ethical committee approval. Samples were collected into lithium-Heparin tubes and CLL cells were purified as described previously (MacFarlane et al, 2002; Inoue et al, 2004). The clinical details of the patients are given in Table I. Purified CLL cells were cultured in RPMI 1640 medium (4–5 × 106 cells/ml) at 37°C in an atmosphere of 5% CO2 and incubated for 16 h with different HDACi, including depsipeptide (10 nmol/l), TSA (0·25 μmol/l) or LBH589 (1–50 nmol/l). CLL cells were also pretreated with depsipeptide or LBH589 at the indicated concentrations for 16 h followed by treatment with different forms of TRAIL or agonistic mAbs to TRAIL-R1 or TRAIL-R2 for a further 4–6 h. Agonistic mAbs to TRAIL-R1 and TRAIL-R2 were used either alone or cross-linked with the appropriate isotype of F(ab’)2 fragment goat anti-IgG Fcγ fragment specific cross-linking antibodies (1:3 ratio of mAbs to cross-linking antibodies, preincubated at room temperature for 30 min). Jurkat T cells (clone E6-1) obtained from European Collection of Cell Cultures (ECACC, Wiltshire, UK) and Ramos cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 5% Glutamax. Ramos and Jurkat cells express predominantly TRAIL-R1 and TRAIL-R2, respectively, on their cell surface (MacFarlane et al, 2005a). Wild-type and TRAIL-R2/DR5-deficient BJAB cells, kindly provided by Dr. A. Thorburn (University of Colorado Health Sciences Center, Aurora, CO, USA), were cultured in RPMI 1640 medium supplemented with 10% FBS and 5% Glutamax (Thomas et al, 2004). Where indicated, cells were pretreated for 30 min with the caspase inhibitor, benzyloxycarbonyl-Val-Ala-Asp (OMe) fluoromethyl ketone (z-VAD.fmk) (Enzyme Systems, CA, USA). Samples were either analysed immediately for apoptosis or pelleted and stored at −80°C for subsequent Western blotting.

Table I.   Clinical data of patients in current study.
PatientSexAgeStageWCCCytogeneticsIGHV statusRxF Res
  1. CLL was staged by the Binet classification.

  2. WCC, White cell count x109/l at diagnosis; Age, age (years) at diagnosis; F res, Fludarabine resistance; IGHV, Immunoglobulin Heavy Chain; U, Unmutated; M, Mutated; nd, not done; N/A, not applicable; t, translocation; del, deletion; Rx, number of different chemotherapy/antibody regimes.

 1M74B234ndM1N/A
 2F59A01811q and 13q delU0N/A
 3F84A029ndM0N/A
 4M77A081·911q delU1N/A
 5M67B113normalU3no
 6F57B7111q and 13 delM8no
 7M59A26·6normalM1N/A
 8F60A5011q delU0N/A
 9F88A13·4ndM2N/A
10F70A020ndM0N/A
11M65A04311q delU0N/A
12M59B60normalM0N/A
13F77A7·5ndU2N/A
14M65A01813q del, trisomy 12 and t(14;18)M0N/A
15F66B41913q delM3YES

Quantification of apoptosis and Western blot analysis

Apoptosis was quantified either by phosphatidylserine (PS) externalisation in the presence of propidium iodide or by loss of mitochondrial membrane potential (ΔΨm) (MacFarlane et al, 2002; Inoue et al, 2004). Samples for Western blot analysis were prepared and caspases and cleaved PARP were detected as previously described (MacFarlane et al, 2002; Inoue et al, 2004). Extraction of histones for the detection of acetylated H3 and H4 were prepared as described (Inoue et al, 2004).

Results

LBY135 but not DR4-A induces apoptosis in Jurkat cells but only when cross-linked

LBY135 (5–500 ng/ml) alone failed to induce apoptosis in Jurkat cells within 6 h, as assessed by externalisation of PS (Fig 1A). However in the presence of a cross-linking antibody, LBY135 (5–500 ng/ml) caused a concentration-dependent induction of apoptosis. These results demonstrated that cross-linking is absolutely required for LBY135 to induce apoptosis in Jurkat cells. On the other hand, DR4-A (5–500 ng/ml) did not induce apoptosis in Jurkat cells at 6 h even in the presence of a cross-linking antibody (Fig 1B). These results are generally in agreement with previous observations that Jurkat cells respond primarily through TRAIL-R2 (Muhlenbeck et al, 2000; Kelley et al, 2005; MacFarlane et al, 2005a,b). LBY135, in the presence of a cross-linker, caused a time-dependent increase in PS externalisation together with a marked decrease in mitochondrial membrane potential (ΔΨm), both of which were almost completely abrogated by the broad spectrum caspase inhibitor z-VAD.fmk (10 μmol/l) (Table II). LBY135 induced a time-dependent processing of caspase-8, the apical caspase in the death-receptor pathway, to its p43 and p41 forms as well as its p18 catalytically active large subunit (Fig 1C, lane 3–4). This was accompanied by processing of caspase-3 to its p20 and p19/17 catalytically active large subunits, which was also catalytically active as PARP was cleaved to its characteristic 85 kDa fragment (Fig 1C, lane 3–4). Pretreatment with z-VAD.fmk (10 μmol/l) completely blocked processing of caspase-8, caspase-3 and PARP cleavage (Fig 1C lane 5). No caspase processing or PARP cleavage was observed in the presence of either LBY135 or the cross-linker alone (Fig 1C lanes 6–7). Taken together these results demonstrated that LBY135, when cross-linked, induced a caspase-mediated apoptosis of Jurkat cells primarily through activation of TRAIL-R2.

Figure 1.

 LBY135 but not DR4-A induces apoptosis in Jurkat cells. Jurkat cells were incubated for 6 h with the indicated concentration of (A) LBY135 or (B) DR4-A alone or in the presence of a cross-linking antibody and apoptosis was assessed by phosphatidylserine (PS) externalisation. Results are expressed as the mean ± SE of at least four independent experiments. (C) Jurkat cells were incubated for the indicated times with LBY135 (100 ng/ml) either alone or in the presence of a cross-linking antibody. Where indicated, cells were pretreated for 30 min with z-VAD.fmk (10 μmol/l) and then incubated with LBY135 in the presence of a cross-linking antibody. Cell lysates were prepared at the indicated times and analysed by Western blotting for the processing of caspase-8, caspase-3 and PARP cleavage.

Table II.   Differential sensitivity of cells to cross-linked LBY135 and DR4A.
Time (h)Cross-linkerJurkatRamos
LBY 135LBY 135DR4-A
% PS%ΔΨm% PS%ΔΨm% PS%ΔΨm
  1. Jurkat and Ramos cells were exposed to either LBY135 or DR4-A for 0–6 h in the presence or absence of a cross-linker as indicated. Apoptosis was then assessed either by phosphatidylserine externalisation or loss in mitochondrial membrane potential (ΔΨm) as described in Materials and methods. Some cells were also incubated for 6 h in the presence of the indicated cross linked antibody and the broad spectrum caspase inhibitor, z-VAD.fmk (10 μmol/l) prior to assessment of apoptosis. The results represent the mean ± SE of at least three separate experiments.

03·2 ± 1·090 ± 0·98·0 ± 0·890 ± 1·68·7 ± 1·990 ± 1·2
 +3·6 ± 0·193 ± 0·67·8 ± 0·790 ± 1·97·6 ± 1·490 ± 1·6
13·4 ± 0·593 ± 0·36·9 ± 1·790 ± 1·96·4 ± 2·090 ± 1·8
 +2·6 ± 0·291 ± 0·67·0 ± 1·990 ± 1·86·9 ± 2·289 ± 2·3
33·9 ± 0·692 ± 0·27·4 ± 1·590 ± 1·57·2 ± 1·590 ± 1·7
 +25 ± 8·476 ± 412 ± 1·784 ± 2·138 ± 5·862 ± 4·1
64·9 ± 0·591 ± 0·88·1 ± 1·190 ± 1·38·8 ± 1·690 ± 1·5
 +54·3 ± 1147 ± 330 ± 7·567 ± 6·259 ± 3·141 ± 2·7
6
z-VAD.fmk
+3·2 ± 0.586 ± 36·9 ± 1·590 ± 1·27·0 ± 1·690 ± 1·1

LBY135 and DR4-A induce apoptosis in Ramos cells

Previously we proposed that TRAIL induces apoptosis in Ramos cells primarily by signalling through TRAIL-R1 (MacFarlane et al, 2005a,b). Exposure of Ramos cells for 6 h to DR4-A (5–500 ng/ml), in the presence but not in the absence of a cross-linking antibody, caused a concentration-dependent induction of apoptosis (Fig 2A), indicating an absolute requirement for cross-linking in order for DR4-A to exert its pro-apoptotic activity in vitro. Furthermore DR4-A exerted its pro-apoptotic activity in Ramos but not in Jurkat cells, consistent with our previous suggestions that these cells respond primarily by signalling through TRAIL-R1 and TRAIL-R2 respectively (MacFarlane et al, 2005a,b). LBY135, in the presence but not in the absence of a cross-linking antibody, caused a concentration-dependent induction of apoptosis in Ramos cells albeit to a lesser extent than DR4-A, with LBY135 (500 ng/ml) inducing a similar level of apoptosis as DR4-A (25 ng/ml) (compare Figs 2A and B). However some of the differing potencies of DR4-A and LBY135 may be due to different efficiencies of the specific cross-linking antibodies used. Both DR4-A (100 ng/ml) and LBY135 (100 ng/ml) induced a time-dependent induction of apoptosis in Ramos cells assessed by PS externalisation and changes in ΔΨm, which were blocked by pretreatment with z-VAD.fmk (10 μmol/l) (Table II). However, the cross-linked DR4-A induced a marked increase in PS externalisation and loss in ΔΨm at 3 h and further at 6 h (Table II), while the cross-linked LBY135 induced these apoptotic changes at 6 h (Table II). DR4-A and LBY135 alone, as well as the cross-linker alone, were completely inactive. DR4-A, in the presence of a cross-linking antibody, induced efficient time-dependent processing of caspase-8, caspase-3 and cleavage of PARP (Fig 2C lanes 3 and 4). Within 3 h, caspase-8 was processed to its p43/41 forms and to its p18 catalytically active large subunit (Fig 2C lane 3). An uncharacterized p30 fragment was also detected, which may possibly represent the unprocessed large and small subunits of caspase-8 derived following cleavage of caspase-8 between its second N-terminal death effector domains and its large subunit. Caspase-3 was again processed to its p20 and p19/17 catalytically active large subunits (Fig 2C lanes 3 and 4). Z-VAD.fmk completely blocked the processing of caspase-8, caspase-3 and cleavage of PARP (Fig 2C lane 5). LBY135, in the presence of a cross-linking antibody, also induced processing of caspases and cleavage of PARP (Fig 2C lanes 10–11), although to a lesser extent than DR4-A commensurate with its lower efficiency in inducing apoptosis in these cells. Again, z-VAD.fmk completely prevented the processing of caspase-8, caspase-3 and cleavage of PARP (Fig 2C lane 12). DR4-A and LBY135 alone as well as the cross-linker alone were completely inactive (Fig 2C lanes 6, 7, 13 and 14). These results clearly indicate that in Ramos cells the cross-linked DR4-A is a more potent apoptotic inducer than the cross-linked LBY135.

Figure 2.

 LBY135 and DR4-A induce apoptosis in Ramos cells. Ramos cells were incubated for 6 h with the indicated concentration of (A) DR4-A or (B) LBY135 alone or in the presence of a cross-linking antibody and apoptosis assessed by phosphatidylserine (PS) externalisation. Results are expressed as the mean ± SE of at least four independent experiments. (C) Ramos cells were incubated for the indicated times with either DR4-A (100 ng/ml) or LBY135 (100 ng/ml) either alone or in the presence of a cross-linker. Where indicated, cells were pretreated for 30 min with z-VAD.fmk (10 μmol/l) and then incubated with cross-linked DR4-A (100 ng/ml). Cell lysates were prepared at the indicated times and analysed by Western blotting for the processing of caspase-8, caspase-3 and PARP cleavage. (D) Wild type (wt) and DR5 null (TRAIL-R2−/−) BJAB cells were incubated for 6 h with the indicated concentration of LBY135 in the presence or absence of a cross-linker. Apoptosis was assessed by PS externalisation. Results are expressed as the mean ± SE of at least three independent experiments.

LBY135 requires TRAIL-R2 to induce apoptosis in BJAB cells

The observation that LBY135 induced apoptosis in Ramos cells was unanticipated as we have previously demonstrated that these cells signalled to apoptosis predominantly through TRAIL-R1, although they express low levels of TRAIL-R2 on the cell surface (MacFarlane et al, 2005a,b). The different sensitivity of Ramos cells to lexatumumab and LBY135 could be due to several possibilities including the different epitopes on TRAIL-R2 recognized by these two mAbs, the requirement of the cross-linker for LBY135 or, alternatively, that cross-linked LBY135 has some pro-apoptotic activity through TRAIL-R1 (DR4). To test the last hypothesis, we tested the ability of LBY135 to induce apoptosis in BJAB cells lacking DR5 (TRAIL-R2) (Thomas et al, 2004). In the presence of a cross-linking antibody, LBY135 (5–500 ng/ml) caused a concentration-dependent induction of apoptosis in wild type but not in BJAB cells lacking TRAIL-R2 (Fig 2D). These results demonstrated that the cross-linked LBY135 requires TRAIL-R2 for its pro-apoptotic activity and exhibits no activity via TRAIL-R1.

LBH589 sensitizes CLL cells to TRAIL-induced apoptosis

Our previous studies had shown that HDACi, such as depsipeptide and TSA sensitize CLL cells to TRAIL-induced apoptosis (Inoue et al, 2004, 2006; MacFarlane et al, 2005a). We wished to examine if a novel HDACi, LBH589, (Qian et al, 2006) also sensitized CLL cells to TRAIL. Firstly we ascertained at what concentration LBH589 would inhibit HDAC activity in CLL cells. Using freshly isolated CLL cells from two individual patients (Table I, Patients 3 and 9), depsipeptide (10 nmol/l) caused an increase in acetylated H3 and H4 but not in acetylated tubulin (Fig 3A, lane 3) compatible with it being a class I HDACi (Inoue et al, 2006). TSA (0·25 μmol/l) caused an increase in both acetylated H3 and H4, as well as an increase in acetylated tubulin (Fig 3A, lane 2) compatible with it being an inhibitor of class I and class II HDACs (Inoue et al, 2006). LBH589 (1–50 nmol/l) caused a concentration-dependent acetylation of histones H3 and H4 as well as a concentration-dependent acetylation of tubulin (Fig 3A, lanes 4–9) supporting the previous suggestion that LBH589 is both a class I and class II HDACi (Qian et al, 2006). Next we examined the ability of LBH589 to sensitize primary CLL cells to TRAIL-induced apoptosis using depsipeptide as a positive control. Depsipeptide (10 nmol/l) sensitized CLL cells to TRAIL-induced apoptosis (Fig 3B) in agreement with our previous findings (Inoue et al, 2004). Pretreatment of CLL cells from a further eight different individuals for 16 h with LBH589 (1–50 nmol/l) resulted in a concentration-dependent sensitization of CLL cells to TRAIL-induced apoptosis (Fig 3B). Furthermore, LBH589 appeared to be equipotent to depsipeptide in its ability to sensitize CLL cells to TRAIL (Fig 3B).

Figure 3.

 LBH589 sensitizes chronic lymphocytic leukaemia (CLL) cells to TRAIL-induced apoptosis. (A) CLL cells from two patients were incubated for 16 h either alone with trichostatin A (TSA, 0·25 μmol/l), depsipeptide (Dep, 10 nmol/l) or the indicated concentration of LBH589. Cells were harvested and cell lysates analysed for inhibition of histone deacetylase activity by accumulation of acetylated tubulin (Ac-tubulin) and acetylated histone H3 or H4 (Ac-H3 or Ac-H4). (B) CLL cells from eight patients were pretreated with depsipeptide (Dep, 10 nmol/l) or LBH589 (1–50 nmol/l) for 16 h then treated for further 4 h with TRAIL (100 ng/ml) and apoptosis assessed by PS externalisation. Results are expressed as the mean ± SE of eight different patients.

Investigation of potential apoptotic activity of LBY135 and DR4-A in CLL cells

We then investigated cells from a further 11 CLL patients to determine their sensitivity to TRAIL, DR4-A and LBY135. Neither TRAIL, DR4-A nor LBY135 alone induced apoptosis in cells from any individual unless the cells had been pretreated with an HDACi. No activity of DR4-A or LBY135 was observed in cells in the absence of a cross-linker. Cells from one patient were resistant to all the treatments and cells from two other patients had a high level of spontaneous apoptosis (≥40 % within 24 h) so leaving eight CLL cases for analysis. CLL cells from these eight patients all responded to the combination of HDACi and TRAIL but appeared to comprise three subsets depending on their response to DR4-A or LBY135. Cells from three patients (Table I, Patients 1, 2 and 7) were sensitized to TRAIL but were not sensitized to DR4-A or LBY135, with a representative example given in Fig 4A. Cells from a further three patients (Table I, Patients 3, 4 and 6) showed a sensitization ranking as follows: TRAIL > DR4-A > LBY135 with a representative example shown in Fig 4B. Cells from the remaining two patients (Table I, Patients 5 and 8) showed a sensitization ranking as follows: TRAIL > LBY135 > DR4-A with a representative example given in Fig 4C. To confirm that this induction of cell death by LBY135 was due to apoptosis, we also examined the processing of caspases in CLL cells from one patient (Patient 8), which were sensitive to LBY135 in the presence of a cross-linker (Fig 4D). Induction of apoptosis assessed by PS externalisation was accompanied by processing of caspase-8 to its p43/41 fragments as well its p18 catalytically active large subunit together with processing of caspase-3 to its catalytically active p19/17 large subunit (Fig 4D lane 8).

Figure 4.

 Marked inter-individual variation in response of chronic lymphocytic leukaemia (CLL) cells to DR4-A and LBY135. CLL cells were pretreated with depsipeptide (Dep, 10 nmol/l) or LBH589 (10 nmol/l) for 16 h then treated for a further 4 h (except for Patient 2 where cells were treated for 6 h) with TRAIL (100 ng/ml) or cross-linked DR4-A or cross-linked LBY135 and apoptosis assessed by PS externalisation. (A) Cells from Patient 2 responded to TRAIL but did not respond to cross-linked DR4-A or LBY135. Cells from Patients 1 and 7 behaved similarly. (B) Cells from Patient 3 responded to TRAIL and DR4-A to a greater extent than to LBY135. Cells from Patients 4 and 6 behaved similarly. (C) Cells from Patient 5 responded to TRAIL and LBY135 to a greater extent than to cross-linked DR4-A. Cells from Patient 8 behaved similarly. (D) Cells from Patient 8 were incubated for 16 h with depsipeptide (10 nmol/l) followed by incubation for a further 4 h with TRAIL (100 ng/ml), cross-linked DR4-A (5 μg/ml) or cross-linked LBY135 (5 μg/ml) and then cell lysates prepared and analysed by Western blotting for the processing of caspase-8 and caspase-3. Apoptosis assessed by PS externalisation is also shown for cells from this patient.

Cross-linking of lexatumumab enables CLL cells to signal to apoptosis

The above results indicated that, under some circumstances, CLL cells may possibly signal through TRAIL-R2 as well as TRAIL-R1. To further investigate this possibility, we examined the ability of lexatumumab (HGS-ETR2), a fully human agonistic TRAIL-R2 antibody, to signal to apoptosis in CLL cells. Previously we had shown that lexatumumab did not signal to apoptosis in CLL cells even when the cells had been pretreated with an HDACi (MacFarlane et al, 2005a,b). In these studies we also compared the effects of soluble forms of TRAIL (100 ng/ml), TRAIL.R1-5, a TRAIL-R1-selective ligand and TRAIL.R2-6, a TRAIL-R2-selective ligand (MacFarlane et al, 2005b). We examined cells from a further nine patients to determine their sensitivity to lexatumumab in the presence and absence of a cross-linker. None of the preparations of TRAIL induced apoptosis in CLL cells unless the cells were pretreated with depsipeptide, an HDACi. Cells from one patient (Table I, Patient 15) were completely resistant to all treatments whereas cells from the other eight patients were all sensitized following pretreatment with depsipeptide, albeit to differing extents (Fig 5). Some inter-individual variation was observed with cells from different patients but some important common themes emerged. Following depsipeptide pretreatment, TRAIL and TRAIL.R1-5, but not TRAIL.R2-6, induced significant levels of apoptosis above the controls in CLL cells (Fig 5). Mapatumumab (ETR1) induced similar levels of apoptosis irrespective of whether or not it was cross-linked (Fig 5). Importantly cross-linking of lexatumumab (ETR2) resulted in an enhancement of apoptosis above lexatumumab alone in cells from 7/8 patients (Fig 5). Some inter-individual variation was also observed in that depsipeptide alone (10 nmol/l) induced apoptosis in cells from some patients possibly because of the slightly longer times of incubation compared to those used in our earlier studies (Inoue et al, 2004; MacFarlane et al, 2005a). Thus, CLL cells can signal to apoptosis via TRAIL-R2 but only following cross-linking of agonistic antibodies to TRAIL-R2.

Figure 5.

 Chronic lymphocytic leukaemia (CLL) cells can signal through TRAIL-R2 following cross-linking of agonistic antibodies. CLL cells were pretreated with depsipeptide (Dep, 10 nmol/l) for 16 h then treated for further 6 h with low endotoxin TRAIL (TRAIL-LE, 100 ng/ml), the TRAIL-R1-selective ligand (R1-5, 500 ng/ml) or the TRAIL-R2 selective ligand (R2-6, 500 ng/ml), Mapatumumab (ETR1)(1 μg/ml) either alone or cross-linked or lexatumumab (ETR2) (1 μg/ml) either alone or cross-linked. Results shown are from cells from seven patients as indicated. The dark black line indicates the mean value for the % apoptotic cells induced by the specific treatment. Cells from Patient 2 (Table I) did not respond to the cross-linking and are omitted from this analysis. Cross-linking alone did not induce any apoptosis in the presence or absence of depsipeptide.

Discussion

Novel therapies are urgently required for the treatment of CLL. Although TRAIL has been reported to be selectively toxic to many tumour compared to normal cells (Ashkenazi, 2002; Koschny et al, 2007), we observed that primary cells from patients with CLL were completely resistant to TRAIL, even to concentrations as high as 2000 ng/ml (MacFarlane et al, 2002). Similarly we and others have shown that in addition to CLL, primary cells from patients with other haematological malignancies including B-cell acute lymphoblastic leukaemia, B-cell non-Hodgkin lymphoma and mantle cell lymphoma are also resistant to TRAIL (Snell et al, 1997; Clodi et al, 2000; MacFarlane et al, 2005b). Subsequently, we demonstrated that primary CLL and mantle cell lymphoma cells could be sensitised to TRAIL by prior exposure to a HDACi (Inoue et al, 2004; MacFarlane et al, 2005b). In this study, it was shown that another HDACi, LBH589, also sensitises CLL cells to TRAIL-induced apoptosis (Fig 3). Nanomolar concentrations of HDACi, such as LBH589, were sufficient to sensitize CLL cells to TRAIL-induced apoptosis whereas much higher concentrations (∼1000-fold) of these HDACi alone, such as LBH589, were required to induce apoptosis in CLL cells (Inoue et al, 2007).

In studies to assess which of the many available forms of TRAIL may be most appropriate for a clinical study in CLL, we discovered that soluble TRAIL appeared to signal to apoptosis primarily by activation of TRAIL-R1 (MacFarlane et al, 2005a,b). The present studies were undertaken partly to confirm and extend our previous work to two novel agonistic Abs, LBY135 and DR4-A, which activate their respective receptors only upon cross-linking. Surprisingly, our new data demonstrate that cross-linking of agonistic Abs to TRAIL-2, such as LBY135 or lexatumumab can enhance apoptosis in cells, which otherwise signal predominantly through TRAIL-R1, such as CLL or Ramos cells (Figs 2, 4 and 5). By contrast, no cross-linking of mapatumumab was required for it to signal to apoptosis in CLL or Ramos cells (Fig 5) (MacFarlane et al, 2005a,b). These results confirm and extend our previous observations where we proposed that TRAIL signals predominantly through TRAIL-R1 and not through TRAIL-R2 in CLL and Ramos cells (MacFarlane et al, 2005a,b). In these initial studies, exposure to lexatumumab in the absence of a cross-linker resulted in the induction of apoptosis in Jurkat but not in Ramos or CLL cells. Based on these observations, together with data using receptor-selective TRAIL mutant ligands, we proposed that Jurkat cells signalled to apoptosis primarily through TRAIL-R2 whereas CLL and Ramos cells signalled to apoptosis primarily through TRAIL-R1 (MacFarlane et al, 2005a,b). However, it is now apparent that cross-linking of agonistic antibodies, particularly mAbs to TRAIL-R2, such as LBY135 and lexatumumab, enables these antibodies to signal to apoptosis in CLL cells. These results indicate that signalling through TRAIL-R1 and TRAIL-R2 is clearly different and suggests that at least in some cell types, such as CLL and Ramos cells, increased oligomerization of TRAIL-R2 may be required in order to facilitate induction of apoptosis by agonistic TRAIL-R2 antibodies. Our data are compatible with previous studies using cells in which the receptors TRAIL-R1 and -R2 were overexpressed, which showed that TRAIL-R1 was sensitive to both apoptosis and nuclear factor (NF)-κB/Jun N-terminal kinase (JNK) induction following exposure to cross-linked or non-cross-linked soluble TRAIL, whereas TRAIL-R2 only responded to cross-linked forms of TRAIL (Muhlenbeck et al, 2000). These results suggested distinct cross-linking requirements of TRAIL-R1 and -R2 for apoptosis and NF-κB induction and for activation of JNK. Furthermore, in support of these conclusions, this group also demonstrated that TRAIL-R1 was activated by both soluble and membrane-bound forms of the ligand, whereas TRAIL-R2 was only activated by membrane-bound TRAIL or an epitope-tagged form of soluble TRAIL cross-linked with a mAb (Wajant et al, 2001).

In the present and our previous studies, we have shown that in response to wild type TRAIL, receptor-selective TRAIL mutants and non-cross linked agonistic antibodies, CLL cells signal predominantly through TRAIL-R1 and not through TRAIL-R2. However, we now show that following cross-linking of certain agonistic TRAIL-R2 antibodies, CLL cells can also signal through TRAIL-R2. In vitro cross-linking can be induced by various reagents including anti-immunoglobulin antibodies or by chemical cross-linkers, whereas in vivo effector molecules or cells responsible for cross-linking are restricted to Fc receptors present on most immune effector cells or the complement component C1q (Chuntharapai et al, 2001; Motoki et al, 2005). Such endogenous cross-linkers may vary in different individuals either due to immunosuppressive therapy or polymorphisms of FcR, therefore in terms of therapy it may be preferable to use an agent that signals via TRAIL receptors but does not require such cross-linking (Cartron et al, 2002; Kono et al, 2002; Motoki et al, 2005). Taken together our data suggest that therapies that target TRAIL-R2 and that rely on efficient cross-linking in vivo may not be optimal as the reliance on activation through Fc receptors may well be comprised in patients with CLL, who are also often receiving concomitant therapy with immunosuppressive agents. In conclusion, we propose that it may still be preferable to treat patients with CLL with a HDACi followed by a receptor-selective TRAIL mutant ligand specific for TRAIL-R1 or an agonistic antibody to TRAIL-R1 that does not require cross-linking in order to induce apoptosis.

Acknowledgements

This work was supported in part by a grant from Novartis. We thank Dr. Andrew Thorburn (University of Colorado Health Sciences Center, Aurora, CO, USA) for the TRAIL-R2/DR5-deficient BJAB cells.

Ancillary