Rituximab Modifies the Cisplatin-mitochondrial Signaling Pathway, Resulting in Apoptosis in Cisplatin-resistant Non-Hodgkin’s Lymphoma¹

Patients suffering from HIV who develop NHL³ face a dismal rate of survival. ARL patients experience relapse of tumor cells resistant to conventional chemotherapy. Treatment of ARL patients with therapeutic drugs results in median survival times of 4.5–16 months (1, 2, 3, 4). The chimeric anti-CD20 antibody, rituximab, has been shown to have an antitumor effect when administered as a single agent to patients with a variety of lymphoma types (5). Studies show, however, that 50% of patients are nonresponsive to this therapy (6). Using rituximab in combination with chemotherapeutic drugs results in a higher sensitivity to therapy in B-cell NHL lines and thus allows drugs to be used at lower concentrations than otherwise necessary (7, 8, 9). Clinically, this approach could result in lower systemic toxicity and myelosuppression along with an extended rate of survival.

We have shown in previous studies that rituximab can sensitize drug-resistant NHL and ARL cell lines to the cytotoxic and apoptotic effects of chemotherapeutic drugs (7, 8, 9, 10). Upon rituximab treatment, there was selective down-regulation of the Bcl-2 protein in the 2F7 and 10C9 ARL cell lines (8). The overexpression of Bcl-2 was dependent on the autocrine/paracrine loops of endogenous IL-10 secreted by the tumor cells. IL-10 accumulated in the tumor environment over time and progressively induced activation of the transcription factor STAT3, which was responsible for the overexpression of Bcl-2 (11). In the presence of rituximab, IL-10 secretion and STAT3 activation were down-regulated, and Bcl-2 expression was inhibited. Furthermore, inhibition by rituximab and anti-IL-10 antibody was shown to sensitize cells to chemotherapeutic drugs.

The mechanism by which rituximab sensitizes these tumors to drug-mediated apoptosis is not known. We hypothesized that Bcl-2 and other proapoptotic or antiapoptotic factors in the ARL cells were modulated by CD20 signaling via rituximab treatment and led to a drug-sensitive phenotype. To obtain a better understanding of how the internal apoptotic machinery in ARL responded to rituximab and drugs, we analyzed several factors involved in the apoptotic cascade in the 2F7 cell line upon cisplatin and rituximab treatment.

The resistance to apoptotic stimuli, such as drug-mediated cytotoxicity, is often due to the inability of cells to carry out the signal transduction ultimately leading to cell death (12). This may be due to insufficient expression of signaling pathway proteins, the overexpression of protective factors, or mutations in apoptotic proteins, such as p53. Drugs have been shown to down-regulate the levels of antiapoptotic proteins or induce higher expression of proapoptotic proteins (13, 14, 15, 16). This phenomenon illustrates the possibility that therapeutic drugs may not result in the toxicity of tumor cells but nonetheless possess the ability to alter protein expression in a way that allows additional agents to induce cell death at much lower thresholds.

Cell death resulting from apoptotic stimuli is dependent on a number of protein-protein interactions and intercellular events (17, 18). A hallmark of apoptosis is the activation of the caspase cascade. Several studies have shown that UV irradiation, chemotherapeutic drugs, and death-inducing surface receptors initiate the activation of caspases, which ultimately results in the cells’ demise, leading to internal cellular degradation, membrane blebbing, nuclear condensation, and DNA fragmentation. Two major pathways have been identified in the signal transduction of caspases leading to cell death. The type I pathway, studied primarily in death receptor signaling, involves the activation of caspase-8 by death-inducing signaling complexes formed by receptor trimerization and adapter proteins. Activated caspase-8 leads to the subsequent cleavage and activation of caspase-3, which is responsible for the targeting of cellular substrates downstream. The type II pathway varies from the type I pathway in that caspase-8 targets the Bcl-2 family member protein Bid. Bid is cleaved by caspase-8, whereupon its truncated form translocates to the mitochondrial membrane and initiates the mobilization of cytochrome c into the cytoplasm from the mitochondrial intermembrane space. This efflux of cytochrome c from the mitochondria is regulated by various members of the Bcl-2 family. The amount of protein of each family member can have a significant impact on cytochrome c release, depending on the initiator stimulus. Cytoplasmic cytochrome c complexes with pro-caspase-9 and Apaf-1 to form the apoptosome. The apoptosome formation perpetuates pro-caspase-9 cleavage to the active caspase-9 form. Caspase-9 targets caspase-3 for cleavage, in a manner similar to that of caspase-8 within the type I pathway, and proceeds with the aforementioned events involved in programmed cell death.

A central question regarding the phenomenon of drug sensitization by rituximab involves the role that drugs play in the death initiator signal. Cisplatin specifically has been shown to induce apoptosis by (a) induction of FasL on the surface of tumor cells, resulting in killing of the Fas-sensitive subpopulation; (b) direct activation of caspase-8; and (c) generation of ROS, which culminates in the destabilization of mitochondria. The objective of this study was to investigate the mechanism by which rituximab sensitizes ARL to cisplatin-induced apoptosis and the role that cisplatin plays in the initiation of the apoptotic signal. We theorize that each agent, rituximab and cisplatin, induces signaling or invokes changes in the apoptotic signaling pathway that, by themselves, are not sufficient to result in cell death. However, in combination, the effects of the two agents may act in a complementary manner, converging on one pathway or using both, and eventually synergize to culminate in tumor cell apoptosis. We also sought to identify which pathway was predominant in the sensitization of these tumor cells to apoptosis by rituximab.

The CD20-positive, Burkitt’s lymphoma cell line 2F7 was established from a patient suffering from AIDS and was generously donated by Dr. Otoniel Martinez-Maza (Jonsson Comprehensive Cancer Center, Los Angeles, CA). RPMI 1640 (Life Technologies, Inc., Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (Gemini, Calabasas, CA) was used for all cell culture. RPMI 1640 supplemented with fetal bovine serum is referred to hereafter as complete medium. Cell culture was kept at 37°C and in 5% atmospheric CO₂.

Rituximab is a monoclonal antibody specific for human CD20. It is of mouse origin and genetically engineered to be chimeric, possessing a F(ab′)₂ mouse fragment linked to a human Fc segment. Rituximab was generated by adjoining the murine antihuman CD20 antibody IDEC-2B8 variable regions to the human IgG1κ constant regions. Rituximab antibody was kindly provided by Dr. Christos Emmanouilides (Department of Medicine, University of California Los Angeles School of Medicine, Los Angeles, CA).

To determine apoptosis, cells were stained with PI (19). Tumor cells (2 × 10⁶) were treated with complete medium, 20 μg/ml rituximab, 1 μg/ml cisplatin, or a combination of rituximab and cisplatin for 3, 6, 12, and 24 h. Cells were collected and then washed twice in PBS/0.1% BSA. Cells were subsequently permeabilized by resuspension in 500 μl of cold 75% ethanol and incubated at −20°C for 1 h. Cells were then washed twice as described before and resuspended in 100 μl of PI solution (100 μg/ml PI and 50 μg/ml RNase). Cells were incubated for 30 min at room temperature while protected from light. After PI staining, 1 ml of PBS was added to each sample, and then the samples were analyzed by flow cytometry with an Epics-XL MCL flow cytometer (Coulter Inc., Miami, FL).

To determine whether cell death in 2F7 cells was, at least in part, due to induction/up-regulation of Fas or FasL, RT-PCR was used to detect Fas and FasL transcriptional regulation. Tumor cells (1 × 10⁶) were treated in 12-well plates (Costar, Cambridge, MA) with complete medium, 20 μg/ml rituximab, 1 μg/ml cisplatin, or a combination of rituximab and cisplatin for 3, 6, 12, and 24 h at 37°C. Total RNA was extracted after incubation using the single-step guanidinium thiocyanate-chloroform method with STAT 60 reagent (Tel-Test B, Inc., Friendswood, TX). One μg of total RNA was reverse-transcribed to single-stranded cDNA for 1 h at 42°C using Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.). The RT-PCR reaction mixture was comprised of 20 μm random hexamer primers, 10 μm DTT, 125 μm each deoxynucleotide triphosphate, and 4 μl of 5× first-strand buffer (Life Technologies, Inc.).

Amplification conditions for Fas were as follows: 30 cycles; denaturation at 95°C for 45 s; annealing at 60°C for 60 s; and extension at 72°C for 3 min. Amplification conditions for FasL were as follows: 35 cycles; denaturation at 95°C for 45 s; annealing at 52°C for 60 s; and extension at 72°C for 3 min. All amplifications were performed with a DNA ThermoCycler 480 (Perkin-Elmer, Norwalk, CT) and analyzed on 1% agarose (Sigma Chemical Co.) gels in TBE [89 mm Tris base, 89 mm boric acid, and 2 mm EDTA, (pH 8.0)].

Sequence-specific primers used in this study included: (a) Fas receptor upstream (5′-ATG-CTG-GGC-ATC-TGG-ACC-CT-3′) and downstream (5′-GCC-ATG-TCC-TTC-ATC-ACA-CAA-3′) primers; (b) FasL upstream (5′-CAG-CTC-TTC-CAG-CTG-CAG-AAG-G -3′) and downstream (5′-AGA-TTC-CTC-AAA-ATT-CAT-CAG-AGA-GAG-3′) primers; and (c) glyceraldehyde-3-phosphate dehydrogenase upstream (5′-GAA-CAT-CAT-CCC-TGC-CTC-TAC-TG-3′) and downstream (5′-CTT-GCT-GTA-GCC-AAA-TTC-GTT-G-3′) primers.

Cells were stained with DiOC₆(3) to quantitate mitochondrial membrane potential (20). Tumor cells (2 × 10⁶) were seeded in 12-well plates (Costar) and grown in complete medium, 20 μg/ml rituximab, 1 μg/ml cisplatin, or a combination of rituximab and cisplatin for 3, 6, 12, and 24 h. After each incubation, 50 μl of 40 μm DiOC₆(3) (Molecular Probes, Inc., Eugene, OR), a mitochondria-specific dye used to detect membrane depolarization, were added to each well and incubated for 30 min at 37°C. Cells were washed twice in PBS/0.1% BSA. After washing, 500 μl of PBS/0.1% BSA were added to each sample, and then the samples were analyzed by flow cytometry with an Epics-XL MCL flow cytometer (Coulter Inc.).

Cells (2 × 10⁶) were seeded in 12-well culture plates (Costar) and grown in 1 ml of complete medium, 20 μg/ml rituximab, 1 μg/ml cisplatin, or a combination of rituximab and cisplatin for 3, 6, 12, and 24 h. Thirty min before each time point, 1.5 μl of 5 mm 2′,7′-dichlorodihydrofluorescein-diacetate (Molecular Probes, Inc.), a ROS-sensitive compound whose degradation by intracellular peroxides results in the generation of a fluorescent breakdown by-product (21), were added to each sample and incubated at 37°C for 30 min. Cells were collected, resuspended in 500 μl of PBS/0.1% BSA, and then analyzed by flow cytometry with an Epics-XL MCL flow cytometer (Coulter Inc.). The untreated cells were used to set the baseline between 0% and 2% to account for background peroxide generation obtained through normal cellular metabolism. Peroxide build-up due to rituximab or cisplatin treatment increased fluorescence on a per-cell basis and was measured as a shift to a higher fluorescence intensity on the mean channel fluorescence axis.

Cells were seeded in cell culture petri dishes (Corning Inc., Corning, NY) at a concentration of 10⁶ cells/ml and treated with complete medium, 20 μg/ml rituximab, 1 μg/ml cisplatin, or a combination of both rituximab and cisplatin. Cells were incubated at 37°C for 3, 6, 12, and 24 h. Cells were washed in PBS/0.5 mm EDTA, pelleted for 5 min at 200 × g (Marathon 3200R microcentrifuge; Fisher Scientific), and lysed with ice-cold immunoprecipitation assay buffer (1% NP40, 0.1% SDS, 0.5% deoxycholic acid, and 1× PBS) supplemented with protease inhibitor mixture tablets (Boehringer Mannheim). Lysates were transferred to microcentrifuge tubes (Fisher Scientific), sheered using 1-ml insulin syringes (Becton Dickinson, Franklin Lakes, NJ), and frozen at −80°C. Lysates were thawed on ice and subsequently centrifuged at 16,000 × g at 4°C for 20 min. Protein quantitation was performed using the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA). Sample buffer [6.2 mm Tris (pH 6.8), 2.3% SDS, 5% mercaptoethanol, 10% glycerol, and 0.02% bromphenol blue] was added to the lysates at a 1:1 volume. Samples were boiled for 10 min and stored at −20°C.

Lysates were run on polyacrylamide electrophoresis gels and transferred to Hybond enhanced chemiluminescence nitrocellulose membrane (Amersham, Arlington Heights, IL). Primary antibodies for protein analysis were used to detect Apaf-1 (Chemicon International, Temecula, CA), actin (Chemicon International), Bax (Santa Cruz Biotechnology, Santa Cruz, CA), Bcl-2 (Dako Corp., Carpinteria, CA), Bid (PharMingen, San Diego, CA), caspase-3 (PharMingen), caspase-8 (Cell Signaling Technology, Beverly, MA), caspase-9 (Calbiochem-Novabiochem Corp., San Diego, CA), cIAP-1 (Trevigen Inc., Gaithersburg, MD), cIAP-2 (Trevigen Inc.), cytochrome c (PharMingen), FLIP (Upstate Biotechnology, Lake Placid, NY), p53 (Santa Cruz Biotechnology), PARP (PharMingen) and XIAP (Trevigen Inc.). Secondary antimouse and antirabbit antibodies were purchased from New England BioLabs, Inc. (Beverly, MA). Nitrocellulose blots were developed with LumiGlo (New England Biolabs, Inc.).

To isolate mitochondria-free cytosolic fractions (22), cells were treated for 3, 6, 12, and 24 h in cell culture Petri dishes (Corning Inc.) with rituximab (20 μg/ml), cisplatin (1 μg/ml), a combination of rituximab and cisplatin, or complete medium alone. At the specified time points, cells were collected and washed twice with cold PBS/0.5 mm EDTA. After pelleting the cells for 5 min at 200 × g in a Marathon 3200R centrifuge (Fisher Scientific), cells were resuspended in 2 volumes of homogenization buffer (250 mm sucrose, 20 mm HEPES, 10 mm KCl, 1.5 mm MgCl₂, 1 mm sodium EDTA, 1 mm sodium EGTA, and 1 mm DTT) supplemented with protease inhibitor mixture tablets (Boehringer Mannheim). After 30 min of incubation on ice for cell swelling, cells were lysed using a 2-ml glass dounce. Lysates were transferred to microcentrifuge tubes and centrifuged twice for 5 min at 2500 × g to remove nuclei and unbroken cells. Lysates were then transferred to clean microcentrifuge tubes and centrifuged at 16,000 × g for 30 min to remove all mitochondria. Lysates were filtered sequentially through 0.2 and 0.1 μm Ultrafree MC filter columns. The Bio-Rad protein assay (Bio-Rad Laboratories) was used to quantify protein concentration.

We assessed the apoptotic effect of rituximab and cisplatin on 2F7 tumor cells, as well as the effect of their combination, by PI incorporation. In DNA fragmentation assays performed at 3, 6, 12, and 24 h, no significant apoptosis was shown to occur in cells treated with rituximab or cisplatin alone (Table 1). However, combination rituximab-cisplatin treatment demonstrated that by 24 h, 45% of cells were sensitized and induced to undergo cell death (Fig. 1), as compared with 4.62% and 8.6% of cells by rituximab treatment and cisplatin treatment, respectively. Apoptosis did not occur by 12 h in cells treated with the combination therapy, suggesting that cell death was induced through the slower mitochondria-dependent apoptosis pathway.

Induction of FasL in tumor cells has been shown to result in fratricide, whereby the up-regulation of FasL triggers the Fas receptor on neighboring cells and induces cell death among the Fas-sensitive cells within the population (23, 24). To determine whether the induction of cell death in rituximab-cisplatin-treated 2F7 cells was due, in part, to regulation of Fas or FasL expression, RT-PCR was performed, and transcriptional regulation was analyzed. 2F7 tumor cells were shown to be positive for Fas expression (Fig. 2). However, neither rituximab, cisplatin, nor a combination of the two caused a change in the regulation of Fas transcription. Moreover, cells displayed a negative phenotype for FasL expression. Induction of FasL was not seen in any treated sample population, including cells treated with the combination treatment. Therefore, the apoptosis observed in cells sensitized by rituximab to cisplatin does not involve Fas-FasL interaction.

The mitochondrion is a central player in the signaling of apoptosis (25, 26). Upon receiving an apoptotic stimulus, a cell’s mitochondrion releases its cytochrome c content into the cytosol, which complexes with pro-caspase-9 and Apaf-1 (27) and initiates the caspase cascade, leading to cell death. ROS generated by cisplatin are known to disrupt mitochondrial integrity and also initiate apoptosis (28, 29, 30). Treatment of 2F7 tumor cells with cisplatin, both alone and in combination with rituximab, induced high levels of ROS (Table 2). The highest levels of ROS were seen at 24 h (Fig. 3). However, accumulation of ROS was not sufficient to induce cell death because tumor cells remained resistant to cisplatin treatment, as seen by PI staining (Fig. 1).

Mitochondrial depolarization is associated with cytochrome c release and caspase activation. DiOC₆(3) staining was used to determine whether rituximab, cisplatin, or combination therapy interfered with mitochondrial membrane stability by measuring membrane potential (ΔΨ_m). Time course experiments in which 2F7 tumor cells were treated with rituximab, cisplatin, or combination therapy revealed that neither agent alone was sufficient to depolarize the mitochondrial membrane potential (Table 3). However, combination therapy was shown to induce depolarization in 65.3% of the tumor cell population at 24 h, as compared with 1.02% of the tumor cell population for rituximab and 7.1% of the tumor cell population for cisplatin (Fig. 4). Combination therapy had no effect by 12 h. Mitochondrial membrane depolarization thus requires the two-signal process provided by rituximab and cisplatin and occurs with slow kinetics.

When treated with rituximab alone or cisplatin alone, 2F7 tumor cells do not release cytochrome c into their cytosol (Fig. 5). Only in the presence of both rituximab and cisplatin do the tumor cells allow the release of cytochrome c by the mitochondria, which was witnessed at 24 h.

After 24 h of treatment with rituximab (20 μg/ml) or cisplatin (1 μg/ml), the 2F7 tumor cells remained resistant to caspase activation, as determined by caspase cleavage using Western immunoblotting (Fig. 6). When treated with the rituximab and cisplatin combination, tumor cells remained refractory to caspase activation at 12 h but exhibited caspase-9 and caspase-3 cleavage by 24 h. By 24 h, the cells are also seen to cleave the PARP protein. Caspase-8, the function of which resides upstream of caspase-9 and caspase-3, was not activated even at 24 h. FLIP, an antiapoptotic inhibitor of caspase-8 activation that has been shown to be sensitive to cisplatin treatment (15, 31), was not changed by treatment with either agent alone or the combination (Fig. 5). The inactivity of caspase-8 and the activation of caspase-9 and caspase-3 cleavage imply that the induction of apoptosis in these cells by rituximab-mediated cisplatin cytotoxicity is caspase-8 independent and that the death signal is mediated through the mitochondria-dependent caspase cascade and not the mitochondria-independent type I pathway.

Of note, Apaf-1 levels are decreased in cells undergoing combination treatment for 24 h (Fig. 5), implying that the apoptotic signal induced by cytochrome c must have occurred some time before the 24 h time point.

In the presence of rituximab, the 2F7 tumor cells do not up-regulate the level of Bcl-2 protein expression (Fig. 7). Moreover, in the presence of combination treatment, Bcl-2 remains expressed at baseline levels even after 24 h.

Bid is a Bcl-2 family member that is constitutively cytosolic until activation of caspase-8 results in its cleavage. Once cleaved, it is translocated to the mitochondrial membrane and induces release of cytochrome c by facilitating the aggregation of Bax. Neither Bid nor Bax was seen to undergo changes in protein expression upon rituximab, cisplatin, or combination treatment conditions (Fig. 7). Likewise, p53, a regulator of drug-mediated apoptosis that exerts its effect largely through the up-regulation of Bax, was not changed after 24 h of rituximab, cisplatin, or combination treatment (Fig. 7).

One method by which cells inhibit the apoptotic cascade is through the expression of IAPs (32, 33, 34). These proteins bind to the apoptosome complex (cytochrome c, Apaf-1, and caspase-9) and block activation of caspase-9, which prevents the cells from undergoing apoptosis. In particular, XIAP protein stability is sensitive to the effects of cisplatin (35). Regulation of IAPs by a single agent would imply that a particular treatment, although not cytotoxic, could reduce the block inhibiting the apoptotic signal and sensitize cells to the cytotoxic effects of a second agent. We analyzed by protein immunoblotting whether either rituximab, cisplatin, or combination treatment could regulate the expression of IAPs in 2F7 tumor cells. Treatment with the combination of rituximab and cisplatin resulted in a slight decrease of all IAPs after 24 h (Fig. 8). However, the slight decrease of IAPs does not appear to be a point within the pathway that is responsible for allowing the apoptotic signal to continue because neither the rituximab sensitizing agent nor cisplatin was shown to effect their protein expression at any time point when used alone. Because the decrease of IAPs is minimal and seen only in the combination treatment at 24 h, their down-regulation is, in all probability, a consequence and not a facilitator of cell death.

The present findings delineate a mechanism by which rituximab immunotherapy sensitizes cisplatin-resistant ARL tumor cells to the cytotoxic apoptotic effect of cisplatin. The findings show that rituximab-mediated down-regulation of Bcl-2 results in a more drug-sensitive phenotype and that the signaling pathway involved to induce cell death by cisplatin is through the mitochondria-dependent apoptosis pathway.

The means by which cisplatin induces apoptosis in a variety of tumor types are still not entirely understood. Cisplatin has been shown to be an apoptosis-inducing drug that can trigger the death receptor-mediated apoptosis pathway. To initiate the death receptor-mediated pathway, cisplatin has been shown to induce Fas/FasL expression and sequentially activate death-inducing signaling complex formation, caspase-8, and caspase-3 (23, 24, 36). The tumor cell death induced by cisplatin potentially involves the autocrine/paracrine expression of FasL. Contrary to the previous findings, treatment with cisplatin alone on 2F7 tumor cells did not induce Fas/FasL expression or caspase-8 cleavage. Also, the death receptor-mediated pathway can direct the death signal to the mitochondria by a truncated form of Bid generated by active caspase-8. Our results indicated that no cleaved Bid was generated, suggesting that minimal caspase-8 was activated and that the death receptor-mediated pathway is not involved.

Cisplatin is also capable of initiating the mitochondrial death pathway directly and inducing tumor cell death (37, 38). In the mitochondrial pathway, the instability of mitochondria leads to the redistribution of cytochrome c into the cytoplasm, which initiates the formation of the apoptosome and the sequential activation of caspase-9 and caspase-3. Although the downstream events following the release of cytochrome c from the mitochondria have been characterized, the early signal-initiating events that cause the instability of mitochondria by cisplatin are poorly understood. Treatment of ovarian tumor cells with cisplatin causes the production of ROS within the mitochondria. The use of the ROS inhibitors phenoxan and butylated hydroxyanisole blocked cisplatin-mediated cytotoxicity, indicating that these ROS are involved, at least in part, in the cytotoxic effect of cisplatin (39). Production of ROS within the mitochondria is known to initiate the release of cytochrome c into the cytosol (30). In the ARL tumor cell system studied here, a substantial amount of ROS was produced by cisplatin. However, the high level of ROS generated by cisplatin treatment was not sufficient to induce downstream mitochondrial events, such as mitochondrial membrane depolarization, cytochrome c release, or activation of caspases. These results suggest that one of the major resistance mechanisms of the 2F7 tumor cells to apoptosis lies in the protection of mitochondria from activation of ROS generated by cisplatin. The antiapoptotic Bcl-2 members then become likely factors responsible for such resistance. One of the mechanisms by which Bcl-2 has been shown to protect against apoptosis is by preventing oxidative damage (40, 41). Overexpression of Bcl-2 may block the oxidative stress caused by cisplatin and inhibit its toxicity, similar to the ROS inhibitors. Moreover, using rituximab to inhibit Bcl-2 overexpression may allow damage by ROS to result in the release of cytochrome c and eventually in cell death.

Activation of caspase-3, independent of caspase-8, by cisplatin has also been reported. The activation of caspase-3 is mediated through a p53-dependent pathway, whereby DNA damage induces p53 translocation to the nucleus and initiation of Bax transcription. De novo Bax protein is then localized to the mitochondria, heterodimerizes with and neutralizes Bcl-2, and thus allows the cell to undergo apoptosis. Upon treatment with cisplatin, however, the 2F7 lymphoma cells did not induce the up-regulation of p53 or Bax protein. Nor did cisplatin treatment alone induce caspase-3 activation. Furthermore, cisplatin did not alter the expression of any of the pro- or antiapoptotic factors we examined.

Our previous work has characterized a mechanism by which endogenously secreted IL-10 from ARL tumor cells induces overexpression of Bcl-2 (8) through activation of the STAT3 transcription factor (11). This phenomenon was abrogated in the presence of rituximab antibody through disruption of IL-10 transcription. Rituximab is shown here to down-regulate the overexpression of the Bcl-2 protein in these cells. However, none of the other proteins analyzed in this study exhibited a regulatory effect by the rituximab antibody within 24 h. Moreover, when treated with both rituximab and cisplatin, the levels of Bcl-2 were further inhibited, implying that the additional effect on Bcl-2 by cisplatin in combination therapy is another reason for the reversal of drug resistance. The Bcl-2 protein is known to protect tumor cells from a variety of apoptotic stimuli, including chemotherapeutic drugs (42). Many drugs that induce cell death do so through a pathway involving the mitochondria-dependent caspase cascade (12). Activation of this cascade is blocked in many cases by the overexpression of Bcl-2 (43, 44), which is largely concentrated on the mitochondria. Therefore, it is conceivable that the down-regulation of Bcl-2 expression by rituximab in these ARL tumor cells releases the block to the apoptotic signal induced by cisplatin. Once the block is released, the cells resume the caspase cascade and undergo cell death.

Experimental evidence shown here demonstrates that by 6 h, nearly half of the tumor cell population has accumulated intracellular ROS when treated with cisplatin. Likewise, rituximab is shown to inhibit Bcl-2 overexpression at the same time point. However, neither is sufficient alone to induce cell death because the 2F7 tumor cells undergo apoptosis only when they are sensitized by rituximab to cisplatin-mediated cytotoxicity. In the combination treatment where apoptosis is observed, overexpression of Bcl-2 is not detected due to the inhibitory effect of rituximab. However, in the cisplatin only-treated cells, where we do not detect apoptosis, Bcl-2 is overexpressed by 6 h. Therefore, ROS-mediated cell death through disruption of mitochondrial integrity necessitates the inhibition of Bcl-2 accumulation on the mitochondria by rituximab. Moreover, the combination does not show caspase initiation until after 12 h. Time kinetics reveal that, by 24 h, the combination of rituximab and cisplatin induces changes in mitochondrial membrane potential, cytochrome c release, caspase-9 and caspase-3 activation, PARP cleavage, the degradation of IAPs and Apaf-1, and DNA fragmentation. Taken together, this evidence indicates that cisplatin cannot induce apoptosis by the time the tumor cells induce higher levels of Bcl-2. We therefore presume that Bcl-2 is indeed serving as the primary block to apoptosis by cisplatin in 2F7 tumor cells.

Based on our findings, chemotherapeutic drugs fail to result in the onset of apoptosis in ARL tumor cells due to a block within the mitochondria-dependent apoptotic pathway. Evidence provided here implicates overexpression of the antiapoptotic protein Bcl-2 as a major component of this block (Fig. 9). Upon CD20 signaling by rituximab, IL-10 secretion and STAT3 activation are inhibited, thus preventing overexpression of the Bcl-2 protein. Abrogation of Bcl-2 overexpression by rituximab allows the accumulation of ROS by cisplatin treatment to induce mitochondrial depolarization and cytochrome c release. After cytochrome c complexes with caspase-9 and Apaf-1, the caspase-9 protein is activated and results in the activation of caspase-3, cleavage of PARP, DNA fragmentation, and subsequent cell death. Treatment of tumor cells with rituximab in all likelihood causes a multitude of effects. Our findings therefore do not exclude the possibility that rituximab alters additional methods of cisplatin resistance, such as drug uptake, cisplatin-DNA adduct formation, interaction with detoxifying intracellular thiols, cAMP- or DNA-dependent protein kinase activity, or DNA repair mechanisms (45, 46, 47, 48, 49).

Patients suffering from ARL have few treatment options. Standard therapy involves chemotherapeutic drug regimens that garner little success. This study shows that drug-resistant ARL tumor cells can be sensitized by rituximab therapy to drug-mediated cell death and delineates the intracellular events involved. This and previous studies (7, 8), justify the use of rituximab treatment in NHL patients with AIDS complications. Future studies to delineate additional factors involved in drug resistance, the contribution of cytotoxic drugs, and the sensitization mechanisms in non-ARL cell lines are currently under way.

We thank Stephanie Louie for assisting with preparation of the manuscript.