Radiation Therapy with Tositumomab (B1) Anti-CD20 Monoclonal Antibody Initiates Extracellular Signal-Regulated Kinase/Mitogen-Activated Protein Kinase–Dependent Cell Death that Overcomes Resistance to Apoptosis

Although most non–Hodgkin's lymphomas initially respond well to conventional treatment, the majority of patients with advanced disease will still relapse (1). Therefore, more effective treatments are being sought, and the introduction of the anti-CD20 monoclonal antibody (mAb) rituximab has significantly improved patient outcome in non–Hodgkin's lymphomas. As a single agent, rituximab results in a 46% response rate in patients with relapsed low-grade lymphomas with complete responses occurring in around 6% (2). However, when rituximab is combined with chemotherapy, greatly enhanced response rates, durations of remissions, and improvements in survival are observed (3). An alternative approach to chemoimmunotherapy is radioimmunotherapy, wherein radioisotopes have been conjugated to mAb. Radioimmunotherapy using anti-CD20 mAb radioimmunoconjugates has also led to impressive clinical results in relapsed indolent B-cell malignancies with high overall (80-97%) and complete (30-75%) response rates (4–6). To date, two distinct anti-CD20 radioimmunotherapy approaches have been approved by the U.S. Food and Drug Administration for the treatment of relapsed indolent B-cell malignancies, namely ⁹⁰Y ibritumomab tiuxetan (Zevalin) and ¹³¹I tositumomab (Bexxar). Despite these clinical successes, the underlying cellular mechanisms behind these impressive clinical responses are poorly understood and commonly attributed simply to efficient targeting of systemic radiation by the mAb (7).

Previous studies have suggested that several mechanisms might be involved in the therapeutic action of anti-CD20 mAb, including antibody-dependent cellular cytotoxicity, complement-dependent cytotoxicity, and the induction of growth arrest or cell death (reviewed in ref. 8). How anti-CD20 mAb and radiation treatment (RT) might combine to increase therapeutic efficacy are poorly understood. Furthermore, it is now emerging that fine specificity differences exist in the epitope binding site of different anti-CD20 mAb, which may determine their biological activity (9).

Our group has previously shown that anti-CD20 mAb may be subdivided as either rituximab-like (type I) or tositumomab-like (type II) according to their linked activity in a number of in vitro assays. For example, rituximab and other type I mAb redistribute CD20 into Triton X-100 insoluble membrane rafts, correlating with their ability to engage complement effectively and cause complement-dependent cytotoxicity (10). In contrast, type II mAb, such as tositumomab, do not redistribute CD20 into Triton X-100 rafts but are more potent at inducing homotypic adhesion and cell death of target cells (11). Importantly, these differences seem to translate to the in vivo mechanisms used by these mAb, at least in xenograft tumor models (12).

Binding of anti-CD20 mAb to lymphoma cells in vitro has been shown to induce modest levels of cell death, presumably via signaling through the CD20 molecule (13). A range of signaling events may be induced after ligation of CD20 (reviewed in ref. 8), including activation of the mitogen-activated protein kinase (MAPK) cascade (14). The Ras-Raf–mitogen-activated protein/extracellular signal-regulated kinase (ERK) kinase (MEK)–ERK1/2 pathway is an evolutionary conserved pathway that is involved in the control of many fundamental cellular processes, including cell proliferation, survival, differentiation, cell death, motility, and metabolism (reviewed by ref. 15). Although commonly thought of as a component of proliferation and survival pathways, ERK1/2 signaling has also been associated with apoptotic signaling in immature B-cell lymphoma (16) and diffuse large B-cell lymphoma cells (17). Similarly, it has been suggested that the overall balance of MAPK activity may determine B-cell fate depending on the kinetics of activation and maturation state of the cell (18, 19).

Intriguingly, the MAPK cascade is also triggered by RT and other DNA-damaging agents (20, 21) where it may be involved in signaling for DNA double-strand break repair by homologous recombination (22). Given that both anti-CD20 mAb and RT have been shown to trigger MAPK activation and that MAPK activation can result in diverse biological outcomes, we have therefore investigated the effect of combining RT with anti-CD20 mAb on this cell signaling pathway. Specifically, we have attempted to investigate the anti-C20 mAb used in the U.S. Food and Drug Administration–approved radioimmunotherapy modalities, ⁹⁰Y ibritumomab tiuxetan (zevalin) and ¹³¹I tositumomab (Bexxar), by combining rituximab and its murine parent mAb (ibritumomab), as well as tositumomab with RT.

Here, we report that when tositumomab is combined with RT an additive increased level of cell death that is not seen with either rituximab or ibritumomab in combination with RT is observed. We show that MEK/ERK1/2 activity seems critical for this additive cell death. Furthermore, although Bcl-2 overexpression mediated protection of cells against RT alone, it had no effect on the additive tumor cell death seen with tositumomab plus RT. This study provides potentially new clinically relevant insights into the involvement of MEK/ERK1/2 signaling in cell death induced by type II anti-CD20 mAb and RT that seems to bypass the apoptotic machinery.

Cell lines and materials. Human cell lines were obtained from the European Collection of Animal Cell Cultures and maintained in antibiotic-free RPMI 1640 with FCS (Sigma; 10%), glutamine (2 mmol/L; Life Technologies) at 37°C, 5% CO₂. The following therapeutic antibodies were used at 5 μg/mL: tositumomab (anti-CD20; GlaxoSmithKline), rituximab (anti-CD20; Roche), ibritumomab (anti-CD20; IDEC Pharmaceuticals), L243 (anti-HLA DR; American Type Tissue Collection), OKT3 (anti-CD3). F(ab′)₂ fragments of antibodies were kindly prepared by Dr. S.A. Beers and Dr. A. Tutt (Tenovus Laboratory). Diagnostic antibodies used in this study were as follows: MEK1 and MEK2 (Labvision/Neomarkers), phosphorylated ERK1/2 (pERK1/2) and total ERK1/2 (New England Biolabs (UK) Ltd.), Bcl-2-FITC (Dako UK Ltd.), goat anti-rabbit horseradish peroxidase (Sigma), rabbit anti-mouse horseradish peroxidase (Sigma-Aldrich). Inhibitors used in the study were as follows: MEK1/2 inhibitors U0126 (10 μmol/L; Promega), PD98059 (10 μmol/L), or general caspase inhibitor QVD (20 μmol/L; both Calbiochem, Merck Chemicals Ltd.). For experiments involving pharmacologic inhibition of cellular signaling cascades, cells were preincubated in the presence of the relevant inhibitor for 30 min.

Irradiation of cells. Cell lines were irradiated using a Gulmay D3 225 X-ray source using a dose rate of 0.77 Gy/min. Low-dose rate irradiations were carried out using a ⁶⁰Co γ-ray source, as described previously (23).

Transfection of Raji and BL60 cells. Transfection of Raji and BL60 cells with pEF plasmid encoding Bcl-2 vector, kindly donated by Dr. Andreas Strasser and Dr. David Huang, was achieved via electroporation using standard electroporation techniques. Selection with geneticin (1-2 mg/mL) or puromycin (1 μg/mL) was applied 24 to 48 h later. Bcl-2 expression levels were determined by flow cytometric intracellular staining methods using an anti–Bcl-2 antibody (BD PharMingen) according to the manufacturer's instructions.

Silencing of MEK1 and MEK2 with small interfering RNA. Silencer validated small interfering RNA (siRNA) against MEK1 and MEK2 (or scrambled control siRNA) were obtained from Ambion (Europe Ltd.) and used at 100 nmol/L. siRNA was delivered to cells using the Amaxa nucleofection device, nucleofection, kit T, and program G016. The nucleofection efficiency was assessed using the pmaxGFP control vector (supplied with nucleofection kit) and was 86% to 93% for Raji and 55% to 60% for SU-DHL4 cells at the 24-h time point.

Western blotting. Cell lysates were prepared in PhoshoSafe lysis solution (MERCK Biosciences Ltd.) containing protease inhibitors. Protein samples were separated by SDS/PAGE and then electroblotted onto a polyvinylidene difluoride membrane (Hybond; Amersham Pharmacia Biotech). Antibodies were added for either 1 h at room temperature or at 4°C overnight as indicated by the manufacturers. Detection was done with horseradish peroxidase–conjugated secondary antibodies (specific to mouse or rabbit IgG) and enhanced chemiluminescence (Amersham Biosciences).

Detection of cell surface phosphatidylserine exposure: Annexin V/propidium iodide assay. Cells (1 × 10⁵) were washed and resuspended in binding buffer [10 mmol/L HEPES (pH 7.4), 140 mmol/L NaCl, and 2.5 mmol/L CaCl₂], containing 1 μg/mL FITC-Annexin V (BD Biosciences). Propidium iodide (PI; 10μg/mL) was added to the samples to distinguish between early cell death and secondary necrosis. Subsequently, cells were assessed by flow cytometry on a FACscan (BD Biosciences).

Measurement of DNA fragmentation: fluorescence in situ end labeling. The fluorescence variant of terminal dUTP N-end labeling has been done essentially as described previously (24). The reaction mix (100 μL) contained 1 μL (25 units) terminal transferase (Promega), 20 μL 5 × cacodylate buffer (supplied with the enzyme), 1 μL 0.5 mmol/L biotin-16-dUTP (Roche), 3 μL 0.5 mmol/L dTTP, and 75 μL H₂O. Preparations were incubated at 37°C for 1 h with this reaction mix, and the reaction was terminated by washing. The incorporated biotin-dUTP was detected with FITC-conjugated avidin (Vector Laboratories). The proportion of fluorescence in situ end labeling (FISEL)–positive cells was determined using a fluorescence microscope (Zeiss AxioScope). At least 300 cells in at least five viewing fields were observed.

Clonogenic survival assay. To measure clonogenic survival, a serial dilution assay was done using a method similar to that published previously (25). Briefly, each sample was treated with a range of doses of irradiation and seeded into individual wells of a 96-well plate over a range of cell densities (12 wells per cell density, eight cell densities per dose) in 20% conditioned medium. After 14 d, the fraction of wells containing no surviving cells was determined using a light microscope. The cell number that resulted in 30% empty wells was calculated and used to generate a clonogenic survival curve that was normalized according to plating efficiency.

2,3-Bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt cell viability assay. Colorimetric 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt (XTT) cell viability assays were done using the Cell Proliferation kit II (XTT; Roche Diagnostics Ltd.) according to the manufacturer's instructions. Briefly, cells were treated in 96-well plates, and 72 h later, XTT reagent was added. After 4 h of incubation, absorbance was measured with the GENios Multi-Detection Microplate Reader (Tecan UK Ltd.) using 485 nm detection and 680 nm reference filters.

Fluorescence microscopy and image analysis. For immunofluorescent staining, harvested cells were cytospun onto clean poly-l-lysine–coated microscope slides. Samples were fixed in absolute methanol at −20°C for 30 min and rinsed in ice-cold acetone for a few seconds. Slides were washed thrice for 10 min and then incubated with primary antibodies for 60 min at room temperature. Subsequently, slides were washed in TBS (3 × 10 min) and stained using anti-rabbit Alexa fluor 488 SFX kit (Invitrogen Ltd.) according to the manufacturer's protocol. After three washes in TBS/0.05% Tween 20, DNA was counterstained with 7-aminoactinomycin D before mounting in ProLong Gold antifade reagent (Invitrogen Ltd.). Images were obtained using an epifluorescence microscope (Olympus BX51), equipped with a ColorView 12 camera. The amount of pERK1/2 colocalized with DNA was determined using the colocalization module in Image-Pro Plus 6.0 software (MediaCybernetics) using algorithms (26) to generate colocalization coefficients (Cgr). Cgr represents the fraction of green pixels that have a red component and is determined by the equation, Cgr = ΣG_i (coloc)/ΣG_i, wherein G_i is green intensity. At least 50 cells in at least three viewing fields were analyzed per sample.

Statistical analysis. To compare differences between the experimental groups, a two-tailed t test was done using Microsoft Excel or SPSS 13 software (SPSS, Inc.).

Tositumomab in combination with RT induces additive tumor cell death, which is MEK/ERK1/2 dependent. To investigate the potential effect of ERK1/2 on cell death induced by anti-CD20 mAb with and without RT, we assessed the amount of cell death induced in Raji and SUDHL-4 tumor cells by measuring the number of Annexin V/PI–positive cells 24 hours after treatment. Figure 1 shows that treatment with the anti-CD20 mAb tositumomab in combination with 4 Gy RT results in a greater induction of cell death than after tositumomab or RT alone. However, this increased additive tumor cell death was inhibited with the specific MEK1/2 inhibitors U0126 and PD98059 (Fig. 1A-C), but no such inhibition was seen with U0124, a structural analogue of UO126, that lacks MEK-inhibiting properties (Fig. 1C). In contrast, the other anti-CD20 mAb used rituximab and ibritumomab (the murine parent mAb of rituximab), both failed to cause the significant increases in cell death observed in combination with tositumomab and RT (Fig. 1A-D).

Similar results were obtained with BL60 tumor cells; tositumomab and 4 Gy RT alone induced ∼40% death, whereas the combination of the two treatments increased the percentage of Annexin V–positive cells to ∼80%. This additive cell death was again inhibited by preincubation of cells with U0126, which reduced cell death to ∼50%. The decrease in tumor cell death resulted in a highly significant difference between combination of tositumomab plus RT in the presence and absence of the MEK1/2 inhibitor (P < 0.006).

We previously showed that Fc-lacking tositumomab F(ab′)₂ fragments are as efficient at triggering cell death as whole IgG molecules (reviewed in ref. 8). However, Fc-FcγR interactions are also known to be able to influence signaling pathways and in a manner dependent upon mAb isotype. Tositumomab displays a mouse IgG2a Fc region, whereas rituximab and ibritumomab, which share identical variable regions, have human IgG1 and mouse IgG1 Fc regions, respectively. Therefore, to verify that the differing Fc domains were not accounting for the varying potency of these anti-CD20 mAb and to establish whether the effects of tositumomab on cell death were Fc-independent, we did a series of experiments comparing whole IgG and F(ab′)₂ fragments of rituximab or tositumomab and examined their effects with or without RT. These experiments revealed that a similar amount of cell death was induced in Raji cells given either the whole IgG or F(ab′)₂ fragments of tositumomab or rituximab in either the presence or absence of RT and/or U0126 (Fig. 1D). The amount of cell death induced by ibritumomab and rituximab in these different conditions was almost identical. Taken together, these data indicate that the Fc region of these mAb is not important for inducing direct cell killing and that differences in the F(ab′) binding sites account for the different effects of rituximab and tositumomab alone and when examined in combination with RT.

To address whether tositumomab had a similar effect in combination with low-dose rate radiation, we compared the proportion of dead cells induced after irradiating cells at 150 Gy/h (as before) or 0.3 Gy/h with or without anti-CD20 mAb. We observed an enhanced level of cell death in the samples treated with tositumomab plus RT, irrespective of the dose rate applied (Fig. 1E). In fact, these results suggested that lower dose rate was more effective at inducing cell death both alone and in combination with tositumomab.

MEK inhibition reverses the loss of clonogenic survival induced by tositumomab plus RT. Next, we investigated the importance of ERK1/2 signaling on the clonogenic survival of Raji, SU-DHL4, and BL60 cells treated with tositumomab with and without RT. In all three cell lines, tositumomab decreased clonogenic survival by 50% to 80% in the absence of RT. Loss of clonogenic survival was, however, much more prominent when tositumomab was combined with RT (Fig. 2A; data not shown). Addition of U0126 improved the clonogenic survival for cells treated with tositumomab plus RT to that seen with RT alone. No significant differences in clonogenic survival were seen between cells treated with rituximab, RT, or both (Fig. 2A). These data were also confirmed using a 72-hour XTT growth inhibition assay (Fig. 2B). Taken together with the Annexin V assay data, these results indicate that there are fundamental differences in the cell death pathways evoked by these two different types of anti-CD20 mAb when combined with RT.

As DNA damage is known to disrupt the cell cycle, we also assessed the cell cycle distribution of Raji cells 24 hours after 4-Gy irradiation in the presence or absence of anti-CD20 mAb. As expected, a clear arrest in the G₂-M phase was observed after irradiation. However, coincubation with anti-CD20 antibodies did not have any additional effect on the proportion of cells in G₂-M phase (data not shown).

Effect of anti-CD20 mAb and RT on ERK1/2 phosphorylation. To examine the activation status of the MAPK cascade after ligation of CD20 and/or RT, we did Western blot analysis to detect dually phosphorylated pERK1/2 in cells treated with tositumomab or rituximab plus RT over a range of radiation doses and time points. Elevated levels of pERK1/2 were observed 10 min after treatment with rituximab or tositumomab in SU-DHL4 cells. For cells treated with rituximab, levels of pERK1/2 peaked 1 hour after treatment and then subsequently decreased. In contrast, stable and prolonged phosphorylation of ERK1/2 was observed in cells treated with tositumomab, although levels of pERK1/2 never reached those seen after rituximab treatment at 1 hour. The amount of pERK1/2 induced by RT alone was minimal and diminished to background levels after 1 hour. When RT was combined with mAb, the pERK responses were similar but enhanced compared with when either mAb was used alone. Rituximab combined with RT resulted in a maximal pERK1/2 signal 1 hour after treatment and then decreased (Fig. 3A). In contrast, there was a slower but more prolonged level of pERK1/2 when cells were treated with both tositumomab and RT, with the pERK1/2 signal lasting for at least 24 hours. These data indicate that both rituximab and tositumomab are able to stimulate the MAPK cascade, resulting in ERK1/2 activation, but with marked differences in the kinetics of activation. High but transient levels of pERK are observed after rituximab treatment, whereas slower but extremely prolonged levels of pERK are seen in tositumomab-treated cells. Furthermore, this pERK activation is more prominent when anti-CD20 mAb are combined with RT.

Recent data have suggested that the trafficking of activated (phosphorylated) ERK1/2 between the cytoplasm and nucleus, rather than phosphorylation per se, determines its biological activity (27–29). Therefore, we next investigated the spatial distribution of pERK1/2 in cells after the various treatments detailed above, using immunocytochemical techniques and image analysis. Rituximab-induced and tositumomab-induced phosphorylation of ERK1/2 was assessed at 8-hour and 24-hour time points in SU-DHL4 (Fig. 3B) and Raji cells (not shown). After treatment with tositumomab, pERK1/2 signals were observed in the center of individual cells and colocalized with 7-aminoactinomycin D, confirming the translocation of pERK1/2 into the nucleus (Fig. 3B). In contrast, although rituximab was able to trigger the phosphorylation of ERK1/2, it remained cytoplasmic and pERK1/2 was not observed in the nucleus (Fig. 3B). To quantify the degree of colocalization of pERK1/2 and 7-aminoactinomycin D, we did image analysis according to a previously published protocol (28). In brief, we determined the proportion of green pixels (representing pERK1/2) that contained a red component (nucleus). This analysis confirmed our earlier observation that tositumomab, but not rituximab, induced nuclear translocation of pERK (Fig. 3C). RT further enhanced the proportion of pERK1/2 signals colocalized with the nucleus compared with the mAb alone, in particular with tositumomab. Thus, the proportion of colocalized nuclear signals was slightly enhanced in the rituximab plus RT–treated cells compared with either treatment alone but the degree of colocalized nuclear signals was much greater after the tositumomab plus RT combination than after tositumomab or RT alone.

Cell death evoked by tositumomab combined with RT is nonapoptotic. We next investigated the mode of cell death induced with tositumomab with or without RT and began by looking for DNA fragmentation, a classic feature of apoptosis. Initially, we did a fluorescent variant of the terminal dUTP N-end labeling assay (FISEL) on Raji and SU-DHL4 cells to establish whether the cell death induced after tositumomab plus RT involves DNA fragmentation. The proportion of FISEL-positive cells in tositumomab and rituximab–treated samples did not significantly differ from nontreated controls (Fig. 4) in keeping with our previously published data (11). This proportion increased to 10% to 12% 24 hours after RT, in accordance with the suggestion that apoptosis is induced after this treatment. However, we found no difference in the percentage of FISEL-positive cells in samples treated with RT or RT combined with either tositumomab or rituximab. These data indicate that apoptotic DNA degradation does not seem to contribute to tositumomab plus RT–induced cell death, and therefore, classic apoptosis is unlikely to be responsible for the increased cell death observed.

Cell death induced by tositumomab plus RT is not blocked by Bcl-2 overexpression. To further investigate whether the mechanism of cell death induced by tositumomab plus RT involved the mitochondrial cell death machinery, we transfected Raji cells with the antiapoptotic, prosurvival molecule Bcl-2. These cells overexpressed Bcl-2 as shown in Fig. 5A and were, as expected, more resistant to cell death induced by RT alone (Fig. 5B). In contrast, tositumomab was able to evoke similar levels of cell death in both wild-type and Bcl-2–overexpressing cells in agreement with our earlier published data (11). A similar result was observed when tositumomab was combined with RT. Taken together, these data indicate that, although Bcl-2 is capable of blocking apoptosis in these cells, tositumomab in the presence or absence of RT triggers a mode of cell death that is independent of Bcl-2 controlled mitochondrial regulation. Importantly, this independence was not simply due to down-regulation of Bcl-2, as there was an increase in the amount of intracellular Bcl-2 protein after treatment with tositumomab or tositumomab plus RT in both wild-type and Bcl-2–overexpressing Raji cells (Fig. 5A).

Cell death induced by tositumomab is not dependent on caspase activation. Next, we investigated the effect of caspases on cell death triggered by tositumomab and RT by preincubating cells with the pan-caspase inhibitor QVD-OPH and examining cell death over a 24-hour period. In the presence of control antibody OKT3, RT caused a modest level of cell death that was partially blocked by QVD-OPH (Fig. 5C). QVD-OPH treatment had no effect on cell death triggered by tositumomab alone. However, there was a partial reduction in the proportion of dying cells after preincubation with QVD-OPH and treatment with tositumomab plus RT. This inhibition was first apparent after 12 hours but was unable to fully suppress the cell death induced even when QVD-OPH concentrations were increased to 100 μmol/L (data not shown), indicating a high degree of caspase independency (Fig. 5C). In contrast, there was no significant effect of QVD-OPH on rituximab-treated cells with or without RT.

siRNA targeting MEK1 and MEK2 abolishes tositumomab plus RT additive cell death. Finally, to further confirm that the MEK/ERK1/2 pathway is critical for the enhanced cell death observed when tositumomab and RT are combined, we investigated the effect of down-regulating MEK1 or MEK2 using siRNA. Importantly, siRNA to MEK1 or MEK2 were able to down-regulate the expression of their respective targets efficiently for up to 96 hours and with good specificity (MEK1 siRNA did not down-regulate MEK2 and vice versa; Fig. 6A). Furthermore, the reduction in expression of either MEK1 or MEK2 correlated with a reduction in the degree of cell death observed when tositumomab was combined with RT. As with the pharmacologic studies done earlier, the extent of cell death after treatment with RT or tositumomab alone was unchanged.

In this study, we have shown that RT and the type II anti-CD20 mAb tositumomab combine to evoke enhanced levels of cell death compared with either treatment alone and that the MAPK signaling pathway downstream of ERK1/2 is important for this effect. The additive cell death and loss of clonogenic survival seen with tositumomab and RT were reversed with specific MEK inhibitors U0126 and PD98059, as well as with siRNA targeting MEK1 or MEK2. This effect seemed to be specific for tositumomab, as rituximab was unable to trigger prominent cell death either alone or in combination with RT and no significant reduction in clonogenic survival was seen with rituximab.

The cell death triggered by tositumomab was not accompanied by the DNA fragmentation characteristic of classic apoptosis. Furthermore, although overexpression of Bcl-2 mediated resistance to RT-induced apoptosis, it had little or no effect on the death induced by tositumomab alone or with the combination of tositumomab plus RT. Moreover, the increase in intracellular Bcl-2 observed after treatment with tositumomab with or without RT gives further evidence that the cell death evoked by the combination of tositumomab and RT is independent of Bcl-2 controlled mitochondrial regulation.

A number of recent reports have suggested that the activation of the ERK1/2 MAPK cascade may promote nonapoptotic, cytoplasmic forms of cell death (30, 31). Prodeath signaling through the MAPK/ERK1/2 pathway has been previously reported for immature B-cell lymphoma (19) and thymocytes (32). It has also been shown that prolonged ERK1/2 activation promotes reactive oxygen species–induced nonapoptotic cell death (reviewed in ref. 33), whereas transient activation of ERK1/2 protects cells from death (34). In our model system, prolonged activation (at least 24 hours) of ERK1/2 was evident after treatment with tositumomab, coincident with elevated cell death. In contrast, rituximab evoked a more rapid, transient activation of ERK1/2 without any accompanying cell death. Therefore, our data provide further evidence to support the hypothesis that sustained ERK1/2 activation is required for cell death induction through MAPK signaling.

In the present study, the extent of cell death induced after ligation of CD20 seemed to directly correlate with the translocation of active pERK1/2 into the nucleus, which was enhanced by RT. Under physiologic conditions, pERK1/2 is thought to translocate into the nucleus where it is dephosphorylated by MAPK phosphatases, such as the ERK1/2 selective MAPK phosphatase-3 (35). Under oxidative conditions, sustained nuclear accumulation of pERK1/2 can occur (36, 37) and is a major contributor to oxidative toxicity in neurons (38). RT is known to induce the formation of reactive oxygen species and membrane peroxydation that can last for several hours after initial administration (38), thereby providing a potential explanation for its ability to prolong and enhance ERK1/2 phosphorylation and nuclear translocation. In support of this suggestion, radiation-induced free radicals are thought to play a major role in the activation of the ERK1/2 pathway in carcinoma cell lines (20).

The distinct properties of tositumomab and other type II mAb have been reported by us previously (8, 10). Unlike other anti-CD20 mAb, they fail to redistribute CD20 into Triton X-100 insoluble lipid rafts and do not induce potent complement-dependent cytotoxicity of target cells. Instead, tositumomab and other type II mAb are more able to induce cell death (10, 13, 39). Moreover, this cell death was previously found by us and others to occur in the absence of DNA fragmentation and independently of the mitochondrial apoptosis pathway (11, 13, 40). Our current results are in agreement with these previous reports and provide new insights into the existing literature by demonstrating that even small DNA nicking, as detected by FISEL, does not occur during tositumomab-induced cell death. Moreover, we have extended these observations by demonstrating that the increased cell death triggered by the combination of tositumomab with RT is also independent of mitochondrial regulation and DNA fragmentation.

We believe that these data may provide some potentially important mechanistic insights into the impressive results observed in the clinic with targeted radiation combined with tositumomab, albeit that the type and quality of radiation is different from high dose rate external RT. The efficacy of tositumomab has been shown to be greatly enhanced by its conjugation to the radionuclide ¹³¹I resulting in highly significant improvements in complete response rates and relapse-free survival (6). Our data indicate that the increased therapeutic efficacy seen with targeted radiation may relate to the MEK-dependent enhanced cell death induced with the combination of tositumomab plus RT. We have also shown that the combination of tositumomab plus RT is able to produce an increased tumor cell death that is independent of Bcl-2 regulation. Overexpression of Bcl-2 is a frequent and defining feature of follicular lymphoma and is known to inhibit apoptotic cell death. Inhibition of apoptosis by Bcl-2 and other antiapoptotic Bcl-2 family members is considered to be an important factor in making this lymphoma refractory to treatment with conventional chemotherapy (41). In relation to the potential clinical implications of these findings, it is interesting to note that high response rates have been seen with ¹³¹I tositumomab in patients with follicular lymphoma who are refractory to both chemotherapy and rituximab (42). While we have not directly investigated the effects of the radioisotopes ¹³¹I or ⁹⁰Y used in Bexxar and Zevalin, respectively, we have found that low-dose rate external beam RT, more equivalent to that produced during radioimmunotherapy, also increased tumor cell kill in combination with tositumomab. In conclusion, we believe our data may provide new molecular insights into the clinical efficacy of targeted radiation and tositumomab in chemorefractory patients.

No potential conflicts of interest were disclosed.

We thank Waleed Alduaij for his technical assistance.