Inhibition of the p53 E3 Ligase HDM-2 Induces Apoptosis and DNA Damage–Independent p53 Phosphorylation in Mantle Cell Lymphoma

Non–Hodgkin's lymphomas (NHL) are a complex group of lymphoid malignancies showing a wide range of pathobiology and clinical behavior (1). Mantle cell lymphoma (MCL) is a distinct NHL subgroup accounting for 6% to 8% of all lymphoid malignancies, which displays the immunophenotype of a B-cell lymphoma (2, 3). At the molecular level, MCL is characterized by the presence of an 11;14 translocation resulting in the overexpression of cyclin D1. Clinically, MCL presents in elderly patients and follows an aggressive course, with relapse occurring almost invariably despite good initial outcomes to conventional chemotherapies (2). One recent encouraging development has been the approval of the proteasome inhibitor bortezomib (VELCADE) after demonstration of its clinical activity in the relapsed/refractory setting (4–6). This agent has helped to validate the ubiquitin-proteasome pathway as a target for NHL therapy and preclinically resulted in part in accumulation of p53 and downstream targets such as p21^Cip1 (7).

Another mechanism by which p53 turnover can be affected is by targeting p53 ubiquitination, a necessary precursor before its proteasome-mediated degradation. The murine double minute (MDM)-2 protein and its human homologue, HDM-2, are the major E3 ligases responsible for p53 ubiquitination (8). HDM-2 is an oncogene that is overexpressed in a variety of malignancies, including leukemias, sarcomas, and solid tumors (9), and may play a role in NHL as well (10). In addition to its E3 ligase activity, HDM-2 also inhibits p53 function through binding to the NH₂ terminus of p53 and suppressing its transcriptional activation properties (8). Under conditions of cellular stress, the ability of HDM-2 to bind p53 is blocked, preventing HDM-2 from inhibiting p53 function. This results in an increase in p53 levels, cell cycle arrest, induction of apoptosis, and senescence (11, 12).

Small-molecule inhibitors of the interaction between HDM-2 and p53 such as nutlin restore p53 function in tumor cells with wtp53 (13, 14). Nutlin binds in the p53-binding pocket of HDM-2 and displaces p53, resulting in stabilization of p53, p21 expression, cell cycle arrest, apoptosis, and growth inhibition (13). These effects were found to be specific for malignancies with wtp53, did not induce p53 phosphorylation, and had a limited effect on primary cells. Recent work has found that, although nutlin acts through the HDM-2/p53 interaction, other cellular targets have been identified, including the cell cycle regulator E2F1 (15), the androgen receptor (16), hypoxia-inducible factor (17), and the nuclear factor-κB pathway (18). The potential to target the HDM-2/p53 interaction has not been evaluated in models of NHL; however, given the above rationale, its potential utility in MCL was of interest.

Reagents. The HDM-2 inhibitor MI-63 was provided by Ascenta Therapeutics, Inc. Nutlin-3, mechlorethamine, doxorubicin, pifithrin-α, and rapamycin were purchased from Sigma-Aldrich, whereas bortezomib and cisplatin were from the University of North Carolina clinical pharmacy.

Cell culture and patient samples. JVM-2, GRANTA-519, REC-1, Jeko-1, Karpas-422, BL-41, and WSU-NHL cells were purchased from the German Collection of Microorganisms and Cell Cultures, whereas CCRF-CEM were purchased from the American Type Culture Collection. Cells were grown in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (Sigma-Aldrich), 100 units/mL penicillin, and 100 μg/mL streptomycin, except GRANTA-519, which were grown in DMEM (Invitrogen). Patient samples were collected and purified as previously described (19).

Semiquantitative reverse transcription-PCR. Total RNA was isolated using an RNeasy Plus kit (Qiagen) and cDNA was synthesized using Superscript II (Invitrogen). PCR was done for 30 cycles, unless indicated otherwise, for p53 up-regulated modulator of apoptosis (PUMA; Genbank accession number NM_014417) using sense primer 5′-CGGCGGAGACAAGAGGAGCA-3′ and antisense primer 5′-GTGAAGGAGCACCGAGAGGAGA-3′. Primers and conditions for p21^Cip1 and β₂-microglobulin were as previously described (20, 21).

Immunoblot. Protein expression in drug-treated cells was measured by immunoblot analysis done as previously described (21). Antibodies to p53, HDM-2, E2F1, and ubiquitin were purchased from Santa Cruz Biotechnologies, whereas anti–phospho-p53 and anti–poly-(ADP-ribose) polymerase (PARP) antibodies were from Cell Signaling Technologies and anti–β-actin was from Sigma-Aldrich.

Cell cycle analysis. Cells were treated with drug for 24 h, then fixed and stained with propidium iodide (Sigma-Aldrich). Cell cycle data were analyzed on a FACSCalibur (Becton Dickinson) using FlowJo v.6.3.3 (Tree Star, Inc.).

Flow cytometry. Levels of total p53, phospho-Ser15-p53, H2A.X-Ser139, and cleaved caspase-3 were measured by flow cytometry after cells were fixed in paraformaldehyde and then permeabilized in 100% methanol. Antibodies to cleaved caspase-3 and H2A.X-Ser139 were purchased from Cell Signaling Technologies.

ELISA for p21 and phospho-p53-Ser15. Cells were treated with vehicle control, 5 μmol/L MI-63, nutlin, or 0.5 μmol/L doxorubicin. ELISAs were done using a p21 (Assay Designs) or a phospho-p53-Ser15 ELISA kit (Cell Signaling Technology) according to the manufacturer's instructions. p21 levels were expressed in pg/mL calculated from the recommended standard curve, whereas the fold induction in phospho-p53-Ser-15 content was determined from the absorbance values and normalized to untreated controls from triplicate experiments.

Caspase activation assay. Activation of caspases was evaluated in cell lysates as previously described (19). The pan-caspase inhibitor Z-VAD and caspase-8 and caspase-9 inhibitors were purchased from Calbiochem and used at concentrations of 50 μmol/L.

DNA damage assays. Cells were treated as indicated in the text, and DNA damage was measured using the Trevigen Comet Assay Kit. Apurinic and apyrimidinic DNA damage was measured using the DNA Damage Quantification Kit (Biovision).

Cell proliferation assay. The WST-1 reagent (Roche Diagnostics) was used to determine the effects of HDM-2 inhibitors on cell proliferation and their relative IC₅₀ as previously published (19).

Drug synergy assays. To evaluate for the presence of synergistic interactions, the methods of Chou and Talalay were used (22, 23). Briefly, MI-63, doxorubicin, cisplatin, bortezomib, or rapamycin were added to cells, and the IC₅₀ of each drug individually was determined using the WST-1 assay. The calculated IC₅₀ was then designated 1× IC₅₀, and similar determinations were made at 0.25×, 1×, 2×, and 4× the IC₅₀. MI-63 and the other chemotherapy drugs were used in a 1:1 ratio and added simultaneously to the cells for 3 d and WST-1 assays were done. Data were then analyzed using CalcuSyn software (Biosoft), and combination indices were calculated.

MI-63 inhibits the growth of NHL cell lines. The HDM-2 inhibitor MI-63 has been shown to bind HDM-2 in vitro with a K_i of 3 nmol/L and was effective in the activation of p53 resulting in inhibition of growth of epithelial cancer cell lines (24, 25). To evaluate its effect on NHL, cell lines of varying p53 status (Supplementary Table S1) were treated with MI-63 for 3 days, and IC₅₀ values were calculated using a cell viability assay. MCL cell lines with wild-type p53 (wtp53; JVM-2, Granta-519 and REC-1) were sensitive to MI-63, with IC₅₀ values ranging from 0.79 to 1.12 μmol/L (Fig. 1A, left). However, an MCL cell line with mutant p53 (mutp53; Jeko-1) was relatively resistant, with an IC₅₀ of almost 12 μmol/L. A second panel of cells representing different NHL histologies was also studied (Fig. 1A, right), and three of the mutp53 cell lines (CCRF-CEM, BL-41, and KARPAS-422) were again relatively resistant to MI-63. Interestingly, the mutp53 Waldenström's cell line WSU-NHL was sensitive, with an IC₅₀ of 1.95 μmol/L.

HDM-2 inhibition should stabilize p53 and induce up-regulation of HDM-2, because it is itself a p53 target. MCL wtp53 cell lines treated with MI-63 for 24 hours indeed revealed an increase in total p53 levels (Fig. 1B; Supplementary Fig. S1) similar to that seen with doxorubicin, whereas no change was seen in mutp53 Jeko-1 cells. HDM-2 levels increased significantly in MI-63–treated JVM-2 and Granta-519 cells, although only a slight increase was noted in REC-1 cells and no change was seen in Jeko-1 cells. To determine if these findings correlated with triggering of apoptosis, the late-stage apoptosis marker PARP was studied and found to be cleaved in wtp53 MCL lines at concentrations equal to or greater than 5 μmol/L.

We also examined the effect of MI-63 on p53 in the mutp53 lines, and in the two where p53 was detectable, WSU-NHL and CCRF-CEM, total p53 levels did not change significantly (Fig. 1B; Supplementary Fig. S1). No HDM-2 expression was seen in either Jeko-1 or WSU-NHL cells, but in CCRF-CEM cells, MI-63 induced a paradoxical decrease in HDM-2 levels. A noticeable effect was seen on PARP in the CCRF-CEM and WSU-NHL cell lines, where MI-63 at 5 μmol/L and greater induced cleavage that was equivalent to or greater than that seen with doxorubicin. This lack of change in p53 levels with continued induction of PARP cleavage may point to the presence of p53-independent mechanisms of action for HDM-2 inhibitors in these cell backgrounds.

Given the possible relevance of p53-independent mechanisms, it was of interest to determine if any of the mechanism of action of MI-63 was dependent on functional wtp53. Pifithrin-α inhibits p53-dependent transcription and p53-mediated apoptosis (26), and should therefore reverse some of the effects of MI-63 treatment of wtp53 cells. Indeed, Granta-519 cells exposed to pifithrin-α in combination with MI-63 were partially rescued from cell death (Fig. 1C) and a similar effect was seen with nutlin. Thus, p53-dependent mechanisms are important to the mechanism of action of these HDM-2 inhibitors. The incomplete rescue may indicate that another secondary, p53-independent mechanism may also be active or it may indicate incomplete suppression of p53 by pifithrin-α in these experiments.

MI-63 induces up-regulation of p53 target genes. We next examined the ability of MI-63 to up-regulate the downstream p53 target genes p21 and PUMA by evaluating their transcript levels with reverse transcription-PCR (RT-PCR). Both p21 and PUMA increased in response to MI-63 in the wtp53 MCL cell lines (Fig. 2A; Supplementary Fig. S2A). The mutp53 WSU-NHL had a small increase in p21 and PUMA by RT-PCR (Fig. 2A; Supplementary Fig. S2A), but no change was seen in Jeko-1 or CCRF-CEM. p21 protein levels were measured with a quantitative ELISA to determine if this increased transcript level resulted in increased protein abundance. Protein lysates from MCL, WSU-NHL, and CCRF-CEM that had been treated with MI-63 showed that wtp53 MCL cells had an increase in p21 protein, which was especially marked in JVM-2 cells (Fig. 2B). In contrast, mutp53 Jeko-1, WSU-NHL, and CCRF-CEM cells showed little to no baseline p21 expression and no increase with MI-63.

The HDM-2 inhibitor nutlin has previously been shown to induce cell cycle arrest (13, 16, 27, 28); thus, we examined NHL cells treated with MI-63 for a similar effect. Wtp53 MCL cells showed a G₁-S cell cycle arrest after 5 μmol/L MI-63, with REC-1 cells being the most affected (Fig. 2C). Jeko-1 and WSU-NHL mutp53 cells had no evidence of cell cycle arrest, but they did arrest in G₂-M after treatment with doxorubicin (Supplementary Fig. S2B). These results suggested that MI-63 acted similarly to nutlin in its ability to induce p21 and a G₁-S cell cycle arrest.

MI-63 induces caspase-mediated cell death. To further define the mechanism of cell death in MI-63–exposed cells, we measured caspase-3, caspase-8, and caspase-9 activities in MI-63–treated MCL cells. All wtp53 MCL cells had an increase in caspase-3 activity, ranging from a 3-fold increase in REC-1 to an 8-fold increase in JVM-2 (Fig. 3A). This was associated with an increase in caspase-8 activity to some extent in all of these cells, particularly REC-1, and may indicate a death receptor–mediated mechanism of cell death. Caspase-9 activation was more variable and seen predominantly in REC-1, whereas mutp53 Jeko-1 cells showed little if any activation of these caspases. The incubation of the pan-caspase inhibitor Z-VAD with MI-63 resulted in a partial rescue of cell death in REC-1 (Supplementary Fig. S3A), and a similar effect was seen with nutlin. Inhibition of caspase-8 and caspase-9 individually in cells treated with either MI-63 or nutlin showed little to no ability to rescue cell viability.

The time course of caspase-3 induction was also measured in JVM-2 and REC-1 cells. Caspase-3 activity began to increase 3 hours after addition of MI-63, peaked after 12 hours, and then began to decrease, whereas activity in doxorubicin-treated cells continued to increase up to 24 hours in JVM-2 (Fig. 3B). REC-1 cells showed a slower onset of caspase-3 activation in that levels did not increase until 6 hours after addition of MI-63 or doxorubicin, then peaked at 12 hours, and began to decrease at 24 hours for both agents (Supplementary Fig. S3B). Accumulation of p53 was also rapid following MI-63 treatment in both cell lines, with increased p53 levels seen as early as 3 hours, which continued to be elevated above baseline even after 24 hours (Fig. 3C; Supplementary Fig. S3C). We also noticed that the abundance of the cell cycle regulator E2F1 decreased after addition of MI-63 after 3 hours and was lost almost completely by 12 and 24 hours in JVM-2 and REC-1 cells, respectively.

HDM-2 inhibitors induce p53 phosphorylation. Interference with the HDM-2/p53 interaction by nutlins has previously been shown to induce cell death in the absence of genotoxic effects as measured by the phosphorylation status of p53 at serine-15 (13). It was therefore of interest to determine if this was also the case in our MCL models, because it could indicate that adding genotoxic drugs such as alkylating agents to HDM-2 inhibitors may enhance anti-NHL activity. Notably, treatment with both MI-63 and nutlin resulted in the appearance of p53 phosphorylated at serine-15, serine-37, and serine-392 (Fig. 4A), although to a lesser extent than with the known genotoxic agent doxorubicin. Because ataxia telangiectasia mutated protein kinase is responsible for phosphorylation of p53 at serine-15 and serine-37 in response to DNA damage, we determined its expression levels. No change in the abundance of total ataxia telangiectasia mutated was detected after treatment with MI-63 or nutlin nor was there any alteration of expression of activated ataxia telangiectasia mutated, which is phosphorylated at Ser-1981.

To further investigate the phosphorylation of p53, we quantitatively measured phospho-p53-Ser-15 levels by an ELISA. The wtp53 cells JVM-2, GRANTA-519, and REC-1 each had increases in phospho-p53 abundance (Supplementary Fig. S4A), ranging from a 2- to 3-fold increase in REC-1 and Granta-519, to a 14-fold increase in JVM-2. In addition, the time course of p53 phosphorylation was studied and phospho-p53-Ser-15 levels increased as early as 3 hours after exposure to MI-63 (Supplementary Fig. S4B), as was the case for total p53 levels. Levels peaked at 6 to 12 hours after exposure, at which time the first evidence of PARP cleavage was seen.

As p53 phosphorylation is associated with damage to DNA (reviewed in ref. 29), we used the comet assay to determine whether HDM-2 inhibition was inducing DNA damage. JVM-2 cells treated with doxorubicin suffered extensive DNA damage as indicated by the development of large comet tails (Fig. 4B). Although MI-63 and nutlin did not induce comet tails to the same extent, MI-63 nonetheless produced a halo effect around cells, suggesting the presence of early DNA damage that was not seen after vehicle treatment. To further evaluate for the presence of DNA damage, we next studied cells for the presence of apurinic and apyrimidinic (AP) DNA lesions formed during base excision and repair of oxidized, deaminated, or alkylated bases. JVM-2 cells treated with MI-63 were found to contain 2.0 × 10⁵ AP lesions per base pair (Supplementary Fig. S4C), compared with 3.7 × 10⁵ AP lesions for cells treated with the alkylating agent mechlorethamine. Interestingly, no increase in AP lesions was seen in nutlin-treated cells and may indicate an off-target effect for MI-63.

To further characterize the p53 phosphorylation and apparent DNA damage in situ, we combined phospho-Ser15-p53 or total p53 immunostaining with either cleaved caspase-3 or the surrogate DNA double strand break marker phospho-H2AX (30, 31). In JVM-2, neither MI-63 nor nutlin induced a significant increase in H2AX expression compared with vehicle-treated cells, despite inducing strong caspase-3 cleavage (Fig. 4C, top). Moreover, this H2AX expression was not linked to expression of phospho-Ser-15-p53 (Fig. 4C, middle), with only 7% of cells expressing both p53-phospho-Ser-15 and H2AX after treatment with MI-63 or nutlin. In contrast, cisplatin treatment resulted in 46% of cells expressing both p53-phospho-Ser-15 and H2AX. These results were further confirmed by evaluating total p53 and H2AX (Fig. 4C, bottom), which showed that total p53 levels increased in response to MI-63 or nutlin, whereas little to no increase was seen in the number of cells expressing H2AX compared with cisplatin. These findings indicate that low levels of DNA damage occurred in a small percentage of cells after HDM-2 inhibition, but the damage was not significant enough to elicit DNA response pathways as measured by ataxia telangiectasia mutated and H2AX phosphorylation.

MI-63 synergizes with other molecularly targeted therapeutics. Given the finding that HDM-2 inhibition resulted in little to no DNA damage, it was of interest to determine if the activity of these agents in vitro could be augmented by DNA-damaging drugs. We combined MI-63 with either cisplatin or doxorubicin using JVM-2 as a model system. Under some conditions, there seemed to be an additive effect on JVM-2 viability (Fig. 5). Interestingly, interaction analysis failed to show combination indices documenting synergy (Table 1). Two other agents that have recently shown promise against MCL are the targeted drugs bortezomib and rapamycin (32). Enhanced antiproliferative activity was seen with the MI-63 and rapamycin combination and also with MI-63 and bortezomib (Fig. 5). Interaction analysis showed that both rapamycin and bortezomib were synergistic with MI-63 over all IC₅₀ doses, and this synergy was especially marked with rapamycin, whereas the bortezomib synergy was weaker (Table 1).

The effectiveness of combination chemotherapy may in some cases depend on the order of drug administration (33). We therefore examined whether MI-63 would be more or less effective at inhibiting cell proliferation depending on the exposure sequence to other chemotherapeutics. Pretreatment of JVM-2 with the IC₅₀ of MI-63, followed 24 hours later by addition of the IC₅₀ of doxorubicin, cisplatin, bortezomib, or rapamycin, was not more effective at inhibiting proliferation than MI-63 alone (Supplementary Fig. S5). However, pretreatment with the IC₅₀ of the chemotherapeutic for 24 hours followed by the IC₅₀ of MI-63 for 24 hours was far more efficient at inhibiting cell growth. The proteasome inhibitor bortezomib was particularly effective at inhibiting cell growth in this assay. Pretreatment of cells with MI-63 followed by chemotherapy resulted in lower levels of p53 accumulation due to the low IC₅₀ doses used compared with MI-63 alone (Supplementary Fig. S5), whereas when the chemotherapeutic was given first and then followed by MI-63, the ability of MI-63 to induce p53 was restored. Notably, neither doxorubicin nor cisplatin in combination with MI-63 was more effective at inducing p53 expression compared with MI-63 alone when the chemotherapeutic agent was given second. However, when MI-63 was given after the chemotherapeutic agent, p53 accumulation was markedly enhanced. Similar findings were seen with bortezomib and rapamycin, which induced greater expression of p53 only when cells were exposed first to these agents and then later to MI-63 (Supplementary Fig. S5).

MI-63 effects in patient-derived samples. As inhibition of HDM-2 function in MCL cell culture models showed promising efficacy, we evaluated the ability of MI-63 and nutlin to affect normal peripheral blood mononuclear cells and lymphoma cells. Peripheral blood from normal healthy donors and a newly diagnosed patient with blastoid MCL (MCL-1) were purified and treated with a range of doses of MI-63 or nutlin. Both agents inhibited the growth of MCL cells, with IC₅₀ values of 6.4 and 7.6 μmol/L, respectively (Fig. 6A). In contrast, neither agent reached its IC₅₀ against the normal peripheral blood mononuclear cell samples, even at doses up to 100 μmol/L, suggesting a degree of selectivity for malignant cells. Finally, incubation of MCL-1 with MI-63 or nutlin resulted in the accumulation of p53 and an increase in p21 expression (Fig. 6B), even in the absence of a clear increase in HDM-2. Notably, both nutlin and MI-63 were more effective at inducing p53 and p21 than doxorubicin.

MCL is an aggressive B-cell lymphoproliferative disorder in which standard chemotherapeutic regimens have limited efficacy, in that although they can induce remissions these responses are typically short lived (34, 35). The inability to cure MCL has accelerated research into the field of molecularly targeted drugs that may have enhanced efficacy with manageable side-effect profiles. Oncotherapeutics often harness the ability of cells to induce apoptosis in response to DNA damage, but genotoxic agents frequently result in nonspecific side effects with selection of drug-resistant clones. One encouraging development has been the approval of the proteasome inhibitor bortezomib, which functions in part by inducing accumulation of p53 and p21, as well as other downstream p53 targets. In addition, the nongenotoxic activation of apoptosis through exploitation of a functional p53 pathway and the use of HDM-2 inhibitors has been shown to be an effective apoptosis-inducing strategy for solid tumor malignancies (13), as well as hematologic malignancies such as multiple myeloma (36). Thus, our goal was to evaluate the potential activity of such inhibitors as therapeutics in models of lymphoma.

In this work, we have examined the effects of a novel inhibitor of the HDM-2/p53 interaction, MI-63, and compared its efficacy with another agent in this class, nutlin, in a panel of NHL cell lines, normal CD19⁺ B-cells, and an MCL patient sample. MCL lines with wtp53 were sensitive to the effects of MI-63, as were patient-derived lymphoma cells, whereas normal peripheral blood mononuclear cells were much more resistant, supporting the possibility of an acceptable therapeutic index. Wtp53 MCL cell lines treated with HDM-2 inhibitors activated a p53-mediated gene expression program, became arrested at G₁-S, and were induced to undergo programmed cell death.

The HDM-2 inhibitor nutlin was originally shown to induce effects consistent with p53 accumulation, but without any effects on p53 phosphorylation (13). Our data show, however, that both nutlin and MI-63 induce p53 phosphorylation in the MCL models used in our study. These findings are consistent with those of Secchiero and colleagues (37), who reported p53 phosphorylation at Ser-15 in B-cell chronic lymphocytic leukemia samples after exposure to nutlin. Similarly, weak p53 phosphorylation was seen recently in Hodgkin Reed-Sternberg cells (27). Because previous work with nutlins in epithelial cancer cell lines showed no p53 phosphorylation (38), one interpretation is that these effects are specific to lymphoid cells. However, very recent publications have shown the presence of p53 phosphorylation in mouse embryonic fibroblasts at Ser-18 (39), the human homologue of phospho-Ser-15-p53, and in primary kidney cells (40) after treatment with nutlin. Thus, it seems likely that this may prove to be not just a cell type but even a cell line–dependent effect.

Phosphorylation of p53 at serine-15 and serine-37 is typically the result of both genotoxic and nongenotoxic insults. Genotoxic stresses include topoisomerase II inhibitors, nitric oxide, hydrogen peroxide, UV light, and ionizing radiation (29), whereas nongenotoxic insults include severe hypoxia and replicative senescence (29). In the comet assay, we did not observe clear comet tails that would indicate DNA damage and only a small amount of damage was seen with MI-63 in the context of AP lesions. Moreover, this only induced H2AX in a small percentage of cells, suggesting that the phosphorylation changes were largely due to nongenotoxic mechanisms. One possibility is that the increase in phospho-Ser-15-p53 and phospho-Ser-37-p53 levels may be due to an increase in total p53, which is then phosphorylated to the same extent as in inhibitor-naïve cells, but more readily detected due to a greater abundance. Indeed, densitometric scanning of immunoblots was supportive of this hypothesis (data not shown). However, the increase in phospho-Ser-392-p53 is important because this increases p53 tetramerization and its ability to bind its consensus DNA sequence (41), thereby promoting its ability to induce transactivation (42). Further studies will be needed to determine the extent to which these phosphorylation status alterations reflect true changes in p53 function and whether these contribute to the apoptotic process.

HDM-2 inhibition in MCL was more effective in suppressing proliferation in JVM-2, Granta-519, and REC-1 cells, which have wtp53, compared with Jeko-1, which has mutp53. As expected, mutp53 Karpas-422 diffuse large-cell lymphoma cells were among the most resistant to MI-63 and this supports the role of wtp53 as the mechanism responsible for MI-63 in this cell type. A surprising finding was that MI-63 was quite potent against the mutp53 WSU-NHL Waldenström's cell line, and also active, albeit with an intermediate IC₅₀, against mutp53 CCRF-CEM acute lymphoblastic leukemia cells. On the mechanistic side, this suggests that such inhibitors are disrupting HDM-2–mediated pathways that are independent of p53. This idea is supported by our data with pifithrin-α, which was unable to give a complete rescue from nutlin or MI-63 in wtp53 cells. Differential effects of nutlin have been reported in cell lines with mutant and wild-type p53 with respect to the modulation of E2F1 when nutlin was combined with cisplatin (15). Our results in mutp53 cells did not require another drug, however, and we did not observe a decrease in E2F1 in mutp53 cells (data not shown) as we did for the wtp53 cells. We observed an increase in p21 activation by RT-PCR in WSU-NHL cells, which did not result in p21 protein expression but could indicate that p73 may be acting in these cells to induce cell death. In this regard, nutlin has previously been shown to disrupt HDM-2/p73 interactions in wtp53, mutp53, and p53-null cells (43). However, recent studies suggest that nutlin can have pleiotropic effects, including on pathways such as angiogenesis (17), androgen signaling (16), and nuclear factor-κB (18). Thus, a number of candidate pathways will need to be studied to determine those that most contribute to the activity of these agents in mutp53 NHL models. Also of note, MI-63 and nutlin differ in their structures, and although both bind residues Phe-19, Trp-23, and Leu-26 of the p53/HDM-2 interacting domain, MI-63 also captures a fourth residue, Leu-22, which may play an important role in the interaction between HDM-2 and p53 (24, 44, 45). This may allow MI-63 to serve as a more potent inhibitor of the p53/HDM2 interaction compared with nutlin, and may also allow it to extend effects to mutp53 cells. We are in the process of generating cell lines resistant to MI-63, which should aid in identifying at least some of these pathways and may also guide the development of future combination regimens based on these drugs.

From a clinical perspective, the validation of a new target as being of potential interest for lymphoma therapy is encouraging and the current findings support studies of HDM-2 inhibitors in patients with relapsed and/or refractory NHL. Combination chemotherapy is typically a far more effective treatment modality than single agents, however, and we indeed found that combining MI-63 with rapamycin or bortezomib displayed increased synergy compared with conventional chemotherapeutics. Pretreatment with MI-63 was antagonistic to chemotherapy, which is consistent with other published works demonstrating that pretreatment of wtp53 tumors actually protected the cells from chemotherapy-mediated toxicity. However, similar pretreatment in mutp53 cell lines induced greater levels of apoptosis (46). Interestingly, another work has shown that nutlin is protective for kidney cells receiving cisplatin therapy and this protection appears independent of the HDM-2, HDM-4, or p53 status of the cells (40). The potential clinical effectiveness of MI-63 is also bolstered by a recent study of the in vivo form of this drug, MI-219, which was shown to be active in murine xenograft models with an excellent side-effect profile (25). Moreover, MI-219 plasma levels of >13 μmol/L were achieved following a single oral dose of 50 mg/kg in mice (25), which are above those which were needed in our MCL and NHL models. Taken together, these findings support the translation of HDM-2 inhibitors such as MI-63 into the clinic both as single agents and in combination with other chemotherapeutics using a rational sequence of administration that may be influenced by p53 status, for the treatment of lymphoid malignancies.

N. Bruey-Sedano was and D. Yang is an employee of Ascenta Therapeutics, which is developing MI-63 for clinical use.

We thank Drs. George Small and Deborah Kuhn for useful discussions and help with manuscript preparation.