Bortezomib-mediated 26S proteasome inhibition causes cell-cycle arrest and induces apoptosis in CD-30(+) anaplastic large cell lymphoma

The 26S proteasome has a direct impact on cellular transcription regulation, cell-cycle progression, oncogenesis and apoptosis, since the turnover of the vast majority of intracellular proteins involved in the aforementioned mechanisms is regulated through the ubiquitin–proteasome pathway.1 Proteasome inhibitors have entered single-agent or combination Phase I and Phase II trials in solid tumors and hematologic malignancies because of the high susceptibility of cancer cells to proteasome inhibitor-induced cell death.2 Here, we have investigated the antiproliferative and pro-apoptotic activity of proteasome inhibitor bortezomib in anaplastic large cell lymphomas (ALCLs), and demonstrate that bortezomib potently inhibits the in vitro growth of ALCL cells and induces apoptosis at nanomolar concentrations in a time- and dose-dependent fashion, suggesting that targeting 26S proteasome may represent a novel therapeutic strategy for ALCLs. ALCLs can be subdivided into two clinically significant subtypes based on expression and constitutive activation of anaplastic lymphoma kinase (ALK), a tyrosine kinase that impacts proliferation, drug resistance, apoptosis and cellular transcription of ALCL cells. Activated NPM-ALK causes transcription silencing of p27KIP, via induction of the PI3K/Akt pathway, as well as upregulation of Bcl-xL, Mcl-1, survivin and cyclin D3, through phosphorylation of STAT3 transcription factor. This promotes cell survival over apoptosis in ALCL cells, given that upregulation of p27KIP expression, pharmacological impairment of PI3K/Akt pathway, or inhibition of STAT3 transcription activity has been shown to cause apoptosis of ALCL cells.3 Targeting ALK kinase or interfering with ALK-dependent signaling are promising strategies for the treatment of ALCLs, but have some limitations: ALK-negative ALCLs, which are characterized by a dismal prognosis, also express at high levels both phosphorylated Akt and STAT3, and these two proteins in some cases may be dispensable for NPM-ALK transforming capability. To determine the antiproliferative activity of bortezomib, ALK-positive (KARPAS299, SR786 and SUDHL1) and ALK-negative ALCL cells (FE-PD) were cultivated in the presence or absence of increasing concentrations (0.0005–0.02 M) of the proteasome inhibitor, and viability was assessed by MTT assay at 24 and 48 h. As shown in Figure 1a, in vitro growth of all ALCL cells was strongly inhibited at submicromolar concentrations in a dose- and time-dependent manner, irrespective of NPM-ALK status and activity. Cytotoxicity profile measured after 24 h exposure demonstrated that KARPAS299 cells had a higher IC50 (0.018 M) compared to the other cell lines (SR786, 0.0045 M; SUDHL1, 0.0071 M; FE-PD, 0.0067 M), as confirmed by the degree of cleavage of PARP protein shown in Figure 1b. Bortezomib almost completely inhibited the growth of SR786, SUDHL1 and FE-PD cells at 0.02 M for 24 h, whereas a significant growth inhibition was observed in KARPAS299 cells at later time points. To rule out any difference in the uptake of the drug, proteasome activity of ALCL cells was assessed by measuring the release of 7-amino-4-methylcoumarin (AMC) fluorophore from the proteasome peptide substrate N-succinyl-Leu-Leu-Val-Tyr-AMC. On the basis of the dose-dependent cleavage of PARP described above, a bortezomib concentration of 0.02 M was used to inhibit herein the cellular proteasome activity and to characterize the molecular events responsible for the growth inhibition of SR786 and FE-PD in all the subsequent experiments, whereas 0.1 M bortezomib was used for KARPAS299 and SUDHL1 cells. At these equal toxic concentrations, bortezomib was shown to penetrate cells and cause up to 80% inhibition of proteasome chimotryptic activity after 1 h in all four cell lines (Figure S1A, inset), maintaining such an inhibitory activity for as long as 24 h (Figure S1A). Chymotryptic inhibitory activity of bortezomib was confirmed by the comparable accumulation level of the proteasome substrate -catenin, later on substituted by a 70–75 kDa apoptotic fragment (Figure S1B, arrowheads). Given that bortezomib causes cell-cycle arrest in transformed cells,4 we sought to extend these findings to ALCLs. SR786 and FE-PD cells were exposed to 0.02 M bortezomib, whereas KARPAS299 and SUDHL1 were treated with 0.1 M bortezomib, and cell-cycle profile was evaluated after 0, 8, 16 and 24 h (Figure 2). When maintained in the presence of bortezomib, the percentage of cells in the G2/M phase increased in a time-dependent manner from 14.3–20.6% at time 0 to 16.9–50.2% at 24 h post-treatment, along with an increase of the sub-G1 population (from 0.4–1.4% at time 0 to 10.2–43.4% at 24 h post-treatment), which indicated the ability of bortezomib to induce G2/M cell-cycle arrest and apoptosis. With respect to the intracellular levels of cyclins, treatment with bortezomib confirmed a marked upregulation of G2/M phase control proteins cyclin B1 or cyclin A, whereas intracellular levels of cyclin E, which controls G1–S transition, were unchanged (Figure S2A). The endogenous levels of cell-cycle regulators p21WAF and p27KIP, short-lived proteins regulated through the ubiquitin–proteasome pathway, were also significantly increased in all four cell lines. This occurred with different kinetics, as transient increase of p21WAF peaked at 8 h in SR786 and FE-PD, whereas at 16 h post-treatment in KARPAS299 and SUDHL1 cells. In contrast, the extent of p27KIP accumulation was similar in all four cell lines, and prolonged if compared to p21WAF (Figure S2A). Since low levels of p27KIP transcription are maintained in ALCL cells by Akt,5 we measured Akt steady state upon exposure to bortezomib and found that endogenous Akt was downregulated with kinetics that correlated with the cytotoxicity profile of each ALCL cell line but not with p27KIP accumulation pattern (Figure S2A). To assess whether depletion of Akt, p21WAF and p27KIP proteins were related to induction of apoptosis, ALCL cells were exposed to equal toxic concentrations of bortezomib for 16 h in the presence of the broad-range caspase inhibitor z-VAD-fmk. As expected, co-administration of z-VAD-fmk prevented the disappearance of Akt, p21WAF and p27KIP from the whole cell extracts of bortezomib-treated SR786 and FE-PD, whereas it did not affect p53, which is not a caspase target protein (Figure S2B). According to the downregulation kinetics of these proteins, this was not observed in KARPAS299 and SUDHL1 cells at this time point. We next investigated the processing and activation of initiator and of effector caspases. As shown in Figure 3, inhibition of proteasome activity in SR786 and FE-PD caused a time-dependent processing and activation of caspase-3, along with cleavage of apical caspase-8 and -9 (Figure 3, cleaved caspase-3, Cl. C-3; Figure S2C). We observed that proteolysis of caspase-8 and -9 did not precede caspase-3 activation, and all were found to happen with similar kinetics, along with PARP cleavage. Nevertheless, despite a less complete activation of the apoptotic machinery, shown by limited processing of pro-caspases-3, -8 and -9, the accumulation of active caspase-3 subunits p19/17 was also observed in bortezomib-treated KARPAS299 and SUDHL1 cells, and this was accompanied by the generation of C-terminal 89 kDa PARP fragment. Being XIAP (X-linked inhibitor of apoptosis) both regulator and target of caspase enzymes in cells undergoing apoptosis, we also measured the steady state of the protein as function of bortezomib exposure time in ALCL cells. As expected, intracellular XIAP was downregulated in drug-treated lymphoma cells (Figure S2C), and such a downregulation was prevented in the presence of the pan-caspase inhibitor z-VAD-fmk (data not shown). Next, we assessed the mitochondrial involvement in bortezomib-induced apoptosis, by evaluating the steady state of pro-apoptotic and anti-apoptotic factors involved in mitochondria integrity. Whereas no consistent changes were observed for Bak protein, pro-apoptotic Bax, which is thought to cause cytocrome-c release from mitochondria, was found to relocate into the mitochondrial fraction (Figure 4a). Because Bax activation occurs through the exposure of its buried BH3 domain at the N-terminus, followed by its oligomerization at the outer mitochondrial membrane, we confirmed Bax activation by immunostaining bortezomib-treated ALCL cells with the specific anti-Bax monoclonal antibody clone 6A7 that recognizes the exposed N-terminus portion of the protein. As shown in Figure 4b, most of bortezomib-treated SR786 and FE-PD cells were apoptotic (panels d and h) and positive for activated Bax (panels c and g), unlike KARPAS299 and SUDHL1 cells in which both Bax labeling and nuclear morphology were not affected by bortezomib at this time point (panels a and b and e and f, respectively). Nevertheless, unlike PARP cleavage, Bax activation was not prevented by the broad-range caspase inhibitor z-VAD-fmk (data not shown). Accordingly, bortezomib promoted the translocation of cytochrome-c from the membrane enriched to the cytoplasmic fraction of ALCL cells in a time-dependent manner (Figure 4a), and the release of such a death inducer occurred upstream of caspases activation, as cytoplasmic relocation of cytochrome c proceeded even in the presence of z-VAD-fmk inhibitor (Figure S3A). Disruption of the mitochondrial membrane potential was further assayed in KARPAS299 and FE-PD cells treated for 12 and 24 h with bortezomib by using as mitochondrial activity marker a fluorescent-tagged lipophilic cation (MitoLight™), which accumulates in the mitochondria of living cells in a membrane potential-dependent manner (red fluorescence), whereas it remains in the cytoplasm of apoptotic cells in its monomeric form (green fluorescence). As shown by the ratio of green to red fluorescence measurements (Figure S3B, 530/590 nm), bortezomib caused a marked loss of membrane permeability in FE-PD cells in a time-dependent manner, which was at least in part prevented by z-VAD-fmk treatment. Yet, and consistent with the recent observation that caspase-3 may act on permeabilized mitochondria to disrupt further m and the permeability of the inner membrane,6 changes in membrane potential of mitochondria were modest in bortezomib-treated KARPAS299 cells, in which activation of the apoptotic program has been shown to be delayed. Among the anti-apoptotic members of the Bcl-2 family, Mcl-1 is consistently expressed in ALK-positive cell lines and tumors, and moderately in the ALK-negative lymphoma cells.7 Mcl-1 promotes cell survival in lymphoma and myeloma cells by binding to and inhibiting pro-apoptotic Bak activity, and the accumulation of Mcl-1 exerts a protective role in cancer cells exposed to proteasome inhibitors.8 We observed a significant accumulation of Mcl-1 in both cytosol and enriched-mitochondrial fraction in the presence of bortezomib (Figure S4A), owing to the inhibition of its ubiquitin-dependent proteasomal degradation, but this was followed by a delayed downregulation perhaps through the activity of executioner caspase proteases. Interestingly, time-dependent downregulation of Mcl-1 paralleled caspase-3 activation described previously, and when ALCL cells were exposed to bortezomib for 24 h, Mcl-1 was completely degraded unless the cells were pre-treated with the broad-range caspase-inhibitor z-VAD-fmk (Figure S4B, SR786 and FE-PD). Consistently, Mcl-1 associated to Bak in the mitochondria of untreated ALCL cells, but this complex was disrupted when the cells were cultivated in the presence of bortezomib, unless z-VAD-fmk caspase inhibitor was added before bortezomib (Figure S4C, FE-PD). These data support the hypothesis that late Mcl-1 downregulation observed in ALCL cell lines occurs downstream of caspase activation, and the inability of the proteasome inhibitor bortezomib to affect steady state and function of Mcl-1 in SUDHL1 or KARPAS299 cells would depend on the level of activated caspases. In conclusion, this study underlines for the first time the effectiveness of bortezomib to induce, at clinically achievable concentrations, growth inhibition and apoptosis in ALCL cells in vitro, regardless of the expression and activity of NPM-ALK protein, and suggests that targeting 26S proteasome may represent a novel therapeutic strategy for ALCLs, especially for those that do not express ALK and are associated with worse prognosis.