Distinct molecular mechanisms responsible for bortezomib-induced death of therapy-resistant versus -sensitive B-NHL cells

Abstract

Resistance to currently available therapies is a major impediment to the successful treatment of hematological malignancies. Here, we used a model of therapy-resistant B-cell non Hodgkin lymphoma (B-NHL) developed in our laboratory along with primary B-NHL cells to study basic mechanisms of bortezomib activity. In resistant cells and a subset of primary B-NHLs, bortezomib treatment led to stabilization of Bak and subsequent Bak-dependent activation of apoptosis. In contrast to sensitive cells that die strictly by apoptosis, bortezomib was capable of killing resistant cells through activation of apoptosis or caspase-independent mechanism(s) when caspases were pharmacologically inhibited. Our data demonstrate that bortezomib is capable of killing B-NHL cells via multiple mechanisms, regardless of their basal apoptotic potential, and contributes to growing evidence that proteasome inhibitors can act via modulation of B-cell lymphoma 2 (Bcl-2) family proteins. The capacity of bortezomib to act independently of the intrinsic apoptotic threshold of a given B-NHL cell suggests that bortezomib-based therapies could potentially overcome resistance and result in relevant clinical activity in a relapsed/refractory setting.

Figure 1

Therapy-resistant B-NHL cell lines undergo delayed apoptosis when treated with bortezomib. (A) After a 72-hour incubation, bortezomib (25nM) induced the death of an approximately equal (P > .05) percentage of therapy-sensitive (Raji, RL) and therapy-resistant (Raji 2R, Raji 4RH, RL 4RH) cells. Data shown represent the average percent of propidium iodide–positive (PI⁺) cells ± SD from at least 5 independent experiments. Asterisks (*) indicate a significant increase compared with killing by cisplatinum (CDDP; 100μM). (B) Forty-eight hours following incubation with bortezomib (25nM), increased caspase-3/7 activity was detected in sensitive and resistant cells. Data shown are from 1 representative experiment repeated 3 times. Bars represent the average caspase-3/7 activity from quadruplicate wells ± SD. Asterisks (*) indicate a significant (P < .01) increase in caspase-3/7 activity compared with vehicle-treated cells. Double asterisks (**) indicate a significant increase in caspase-3/7 activity compared with vehicle-treated cells and compared with the same treatment in resistant cells. (C) A kinetic analysis of PARP cleavage indicated that sensitive cells undergo apoptosis more rapidly than resistant cells. PARP cleavage was observed following 24 hours of incubation of sensitive cells (Raji shown here) with 25nM bortezomib but not until 48 hours following incubation of resistant cells (Raji 4RH shown here) with 25nM bortezomib. Western blots shown are from a representative experiment that was repeated twice. (D) Analysis of apoptotic markers by flow cytometry confirmed that bortezomib (25nM) induced apoptosis with delayed kinetics in resistant cells (Raji 2R shown here) compared with sensitive cells (Raji shown here). Data shown represent the average ± SD of at least 3 independent experiments. Asterisks (*) indicate a significant (P < .05) difference between sensitive and resistant cells at a given time point. Other sensitive and resistant cell lines exhibited similar kinetic responses to bortezomib treatment (data not shown).

Figure 2

Bak, Bik, and Noxa protein levels increase following bortezomib treatment of resistant cells. (A) Western blot analysis of whole cell lysates collected 24 hours following bortezomib (25nM) treatment of sensitive (Raji, RL) or resistant (Raji 2R, Raji 4RH, RL 4RH) cells demonstrated a striking increase in the expression of Bak, Bik, and Noxa in resistant but not sensitive, cells. Blots shown are from a representative experiment repeated more than 3 times. (B) Kinetic analysis of Bak, Bik, and Noxa expression following bortezomib treatment showed that their induction occurred before caspase activation (indicated by PARP cleavage) in resistant cells (Raji 2R shown here). Induction of Bak, Bik, and Noxa was not seen following bortezomib treatment of sensitive cells (Raji shown here), yet PARP cleavage occurred more rapidly than in resistant cells. Data shown are from a representative experiment repeated twice. Other bortezomib-treated sensitive and resistant cell lines displayed similar kinetics with respect to PARP cleavage along with Bak, Bik, and Noxa expression (data not shown). (C) Resistant cells (Raji 2R shown here) expressing a Bak transgene rapidly undergo apoptosis as indicated by caspase-3/7 activity. Eighteen hours following transfection of pIRES2-EGFP containing either a BAK1, BIK, or PMAIP1 (Noxa) transgene, or empty vector as a control, resistant cells were assayed for caspase-3/7 activity. Data shown are from a representative experiment repeated at least 3 times. Bars represent average caspase-3/7 activity from triplicate wells ± SD. Asterisks (*) denote a significant (P < .01) increase in caspase-3/7 activity compared with vector transfected cells. (D) Western blot confirmed expression of Bak, Bik, and Noxa transgenes. (E) Mutation of the BH3 domain of Bak abrogates its ability to induce apoptosis as indicated by caspase-3/7 activity. Data shown are from a representative experiment repeated twice. Bars represent average caspase-3/7 activity from triplicate wells ± SD. Asterisks (*) denote a significant (P < .01) decrease in caspase-3/7 activity compared with wild-type Bak transfected cells.

Figure 3

Increased Bak expression was observed in patient samples following bortezomib treatment. (A) A visible increase in Bak expression was observed in 3 of 10 primary B-NHL patient samples treated ex vivo with bortezomib (5-10nM). This visible increase in Bak expression correlated with a fold change in Bak of ≥ 1.5-fold. Individual fold change values were normalized to Actin and are indicated next to Bak Western blots. (B) Seven of 10 patient samples treated ex vivo with bortezomib had no visible change in Bak expression. (C) Following the general pattern established by sensitive and resistant cells, primary B-NHL cells that demonstrated no induction of Bak (ΔBak < 1.5-fold) following bortezomib treatment died more rapidly than those in which Bak increased (ΔBak ≥ 1.5-fold). B-NHL cells derived from patient 11 did not follow this general pattern. No pattern of sensitivity to rituximab (D) or adriamycin (E) was noted when the same 10 patient samples were segregated based on the ability of bortezomib to induce Bak expression.

Figure 4

Bak is ubiquitinated and degraded by proteasomes in resistant B-NHL cell lines. (A) Dose-dependent inhibition of proteasome activity in resistant cells correlated with dose-dependent induction of Bak expression. Resistant cells (RL 4RH shown here) were incubated with 0-25nM bortezomib for 24 hours, then assayed for chymotrypsin-like activity of the 26S proteasome. Bak expression within the same cell population was determined by Western blot. Proteasome inhibition occurred at similar doses in other resistant and sensitive cell lines (data not shown). Bak induction correlated with proteasome inhibition in other resistant but not sensitive cell lines (data not shown). Data shown are from a representative experiment repeated 3 times. Points on the proteasome activity graph represent the average of triplicate wells ± SD (B) Increased Bak expression was observed in resistant cells treated with proteasome inhibitors and not other stress-inducing agents. Western blot analysis of whole cell lysates demonstrated increased Bak expression in resistant cells (Raji 2R shown here) treated for 48 hours with the proteasome inhibitors bortezomib (Bort.; 25nM) or MG132 (5μM) but not the standard chemotherapeutic agents cisplatinum (CDDP; 100μM) or adriamycin (ADR; 50μM), or the endoplasmic reticulum (ER) stress-inducers brefeldin A (BfA; 1 μg/mL), tunicamycin (Tunic.; 1 μg/mL), or thapsigargin (Thaps.; 1μM). (C) Incubation of resistant cells with bortezomib did not induce mRNA coding for Bak. Quantitative real-time PCR analysis of RNA extracted from resistant cells (Raji 4RH cells shown here) treated with bortezomib over a 24-hour time course showed no increased expression of BAK1, while Bak protein level within the same cell population increased approximately 2.5-fold. Data shown are from a representative experiment repeated twice. Real-time PCR data were analyzed using the ΔΔCt method and is expressed as fold change compared with time = 0 hours ± SD. (D) Bortezomib treatment of resistant cell lines stabilized expression of Bak protein. Bak stability was determined by standard ³⁵S-methionine pulse-chase experiments. Bak was rapidly degraded in vehicle-treated (DMSO) resistant cells (open triangles) but not sensitive cells (closed circles). Incubation of resistant cells with bortezomib (50nM) stabilized the expression of Bak to a level comparable with untreated sensitive cells. Data shown are from a representative experiment repeated twice. (E) Exogenously expressed Bak appears ubiquitinated in resistant but not sensitive cells. FLAG-tagged Bak was immunoprecipitated using FLAG beads and subjected to Western blot for Bak expression. Higher molecular weight bands were detected in FLAG-IP/Bak Western blots from resistant (Raji 2R shown here) but not sensitive cells (Raji shown here). (F) Bortezomib treatment increases the ubiquitination of Bak in resistant cells. After 24 hours of bortezomib treatment (25nM), endogenous Bak or ubiquitin were immunoprecipitated from lysates of sensitive (RL shown here) or resistant (RL 4RH shown here) cells. Ubiquitin detected in the Bak IP from resistant cells increased to a greater degree following bortezomib treatment than did the ubiquitin detected in the Bak IP from sensitive cells while total ubiquitin levels increased similarly in sensitive and resistant cells treated with bortezomib.

Figure 5

Increased Bak expression is required for bortezomib-induced apoptosis, but not death, of resistant cells. (A) siRNA-mediated knock-down of Bak, Bik, or Noxa was confirmed in whole cell lysates collected 24 hours following bortezomib treatment. (B) Knock-down of Bak significantly inhibited the ability of bortezomib to induce apoptosis in resistant cells (Raji 4RH shown here). Resistant cells transfected with siRNA specific for BAK1, BIK, or PMAIP1 (Noxa), or non-targeting siRNA (ctl) as a control, were treated with bortezomib (20nM) for 24 hours and assayed for caspase-3/7 activity. Data shown are from a representative experiment repeated at least 3 times. Bars indicate average caspase-3/7 activity from triplicate wells ± SD. Asterisks (*) denote a significant decrease in caspase-3/7 activity compared with ctl siRNA transfected cells within the same treatment group. (C) Knock-down of Bak (or Bik or Noxa) does not alter the ability of bortezomib to kill resistant cells (Raji 4RH shown here). Cell death was determined by propidium iodide uptake in the same population of cells from B following 72 hours of incubation with bortezomib (20nM). Data shown are the average of at least 3 independent experiments ± SD. No significant difference (P > .05) in the percentage of propidium iodide–positive (PI⁺) cells was seen comparing BAK1, BIK, or PMAIP1 (Noxa) siRNA transfected to control siRNA transfected cells within the same treatment group.

Figure 6

Caspase inhibition does not inhibit bortezomib-induced death of resistant cells and promotes phenotypic changes reminiscent of autophagy in bortezomib-treated cells. (A) Chemical inhibition of caspase activation inhibited bortezomib-induced cell death of sensitive but not resistant cells. After 72-hour incubations with bortezomib (25nM) alone or combined with the pan-caspase inhibitors zVAD-fmk (50μM) or Q-VD-OPh (5μM), sensitive (Raji, RL) and resistant (Raji 2R, Raji 4RH, RL 4RH) cells were assayed for cell death by propidium iodide uptake. Data shown are averages of at least 3 independent experiments ± SD. Asterisks (*) indicate significant (P < .05) inhibition of bortezomib-induced cell death. Double asterisks (**) indicate significant (P < .05) enhancement of bortezomib-induced cell death. (B) Western blot analysis of whole cell lysates from resistant cells (Raji 2R shown here) treated for 48 hours with bortezomib (25nM) with or without zVAD-fmk (50μM) or Q-VD-OPh (5μM) revealed that addition of pan-caspase inhibitors were sufficient to inhibit bortezomib-induced apoptosis, as demonstrated by inhibition of PARP cleavage.

Figure 7

Bortezomib induces caspase-dependent (▿) or -independent (▾) cell death in primary tumor cells isolated from patients with B-NHL. Tumor cells isolated from patients with (A) follicular lymphoma (n = 8), (B) small lymphocytic lymphoma (SLL, n = 5), (C) germinal center (GCB) diffuse large B-cell lymphoma (DLBCL, n = 2), and (D) non-GCB DLBCL (n = 3) were exposed in vitro to bortezomib (10μM) in the presence or absence of the pan-caspase inhibitor Q-VD-Oph (2.5μM). Viability was determined using the CellTiter Glo luminescent assay after 48 hours of incubation and expressed as percentage of luminescent signal compared with untreated controls. Samples in which caspase-dependent cell death was observed (n = 5) are denoted with an open upside-down triangle (▿) and those where caspase-independent cell death was observed (n = 10) are denoted with a filled upside-down triangle (▾) while those that were found resistant to bortezomib (> 80% viable) have no marks above them. Of interest, in some patient specimens, caspase inhibition further enhanced bortezomib activity (patients 24 and 37).