INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The oncoprotein c-Myc, paradoxically, is an inducer of both cell proliferation and cell death, and the levels of Myc and/or the conditions in which it is expressed dictate cell fate (2, 7, 40). Most cancer cells that overexpress Myc, by translocation, amplification, or other means, harness the full growth potential of this oncogene by inactivating the apoptotic effectors of Myc, including the tumor suppressors ARF and p53 (5, 50). ARF is a nucleolar protein that binds to and sequesters Mdm2 (55, 59). Mdm2 is a p53 transcription target (3, 61) that inhibits p53's transactivation functions (37) and ubiquitinates p53 (12), leading to p53 degradation (48). Myc activation induces the sustained expression of both ARF and p53, and this triggers apoptosis; as a consequence, primary ARF- and p53-null hematopoietic and fibroblast cells are impaired in their apoptotic response to Myc (5, 63). Furthermore, deletion of ARF, mutation or deletion of p53, and Mdm2 overexpression occur in 24, 28, and 48%, respectively, of the lymphomas that arise in Eµ-myc transgenic mice (80% overall [5]), and ARF- or p53-null Eµ-myc transgenic mice have a markedly accelerated course of lymphoma (5, 17, 50).

Loss of the antiapoptotic protein Bcl-X_L or Bcl-2 compromises hematopoietic cell survival, whereas loss of ARF or p53 has no effect upon hematopoietic cell development (6, 38, 39, 41, 57). Bax is a proapoptotic Bcl-2 family member whose deletion has modest effects on lymphocyte numbers (22). However, the combined loss of Bax and Bak, another proapoptotic Bcl-2 family member, results in profound defects in both development and lymphocyte homeostasis (24). Bax normally resides in the cytosol of healthy cells, yet it relocalizes and inserts into the outer mitochondrial membrane after stimulation with a variety of apoptotic stimuli (reviewed in reference 9). In turn, this leads to mitochondrial dysfunction with alterations in the permeability transition pore, the release of cytochrome c, and the activation of Apaf-1 and caspases, which cleave intracellular targets required for cell survival (9). The balance of proapoptotic and antiapoptotic Bcl-2 family members regulates the susceptibility of cells to apoptosis (reviewed in reference 23). For example, an excess of Bax can overwhelm the cell and trigger an apoptotic response, whereas the antiapoptotic Bcl-2 family members Bcl-2 and Bcl-X_L inhibit the deleterious effects of Bax (23).

Bcl-2 and Bcl-X_L are overexpressed in many human malignancies (reviewed in reference 46), and Bcl-X_L expression is activated by retroviral insertions in some murine T-cell leukemias and lymphomas (41). Bcl-2 and Myc have been shown to cooperate in transformation (8, 56), and Eµ-myc/Eµ-bcl-2 double transgenic mice develop an aggressive and rapid lymphoma composed of primitive lymphoid cells (54). Although the cooperation between Bcl-2 and Myc and the regulation of Bax by Bcl-2 are well documented, the precise role that Bax plays in Myc functions is less clear.

Bax-deficient mice manifest a modest lymphoid hyperplasia but are not prone to spontaneous tumor development (20, 22). However, mutations that inactivate Bax are found in a subset of human colon adenocarcinomas (45) and some human hematopoietic cancer cell lines (4, 30, 31), and Bax loss cooperates with simian immunodeficiency virus (SV40) large T antigen in transgenic mouse models of cancer (52, 62). Recently, bax has been suggested to be a direct transcriptional target of c-Myc in human tumor cell lines (32). However, it is unclear how Bax influences Myc-induced hematopoietic cell apoptosis and tumorigenesis and whether Bax expression influences the ARF-Mdm2-p53 tumor suppressor pathway. Here we report that, although Myc activation fails to regulate Bax levels in primary murine pre-B cells, Bax-deficient cells are markedly resistant to Myc-induced apoptosis. More importantly, Bax loss accelerates Myc-induced tumorigenesis in Eµ-myc transgenic mice, and the lymphomas arising in Bax-null transgenics selectively lack mutations or deletions of p53. However, Bax-null Eµ-myc transgenics still display the same frequency of alterations of ARF and Mdm2, indicating that Bax is not necessarily a target of ARF or Mdm2 even though both can function with p53 in this tumor suppressor pathway. The results support a model whereby Bax functions as a critical downstream effector of the p53 apoptotic pathway, and thus ARF and Mdm2 must have other mediators important for tumorigenesis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Transgenic and knockout mice. The inbred C57BL/6 Eµ-myc transgenic mouse strain was kindly provided by Alan Harris (Walter & Eliza Hall Institute, Melbourne) and Charles Sidman (University of Cincinnati). Bax-null mice have been previously described and were C57BL/6 × 129/svj (22). The ARF-null (C57BL/6 × 129/svj) and p53-null (C57BL/6 × 129/svev) mice were generously provided by Charles Sherr and Gerard Grosveld, respectively. Eµ-myc transgenics were mated to Bax^+/ mice and the F₁ littermates were then mated to each other to obtain Bax^+/+, Bax^+/, and Bax^/ Eµ-myc transgenics.

Primary B cells. Primary pre-B cell cultures were generated from the bone marrow of 6- to 8-week-old wild-type, Bax-, ARF-, p53-, and ARF/p53-double null mice as previously described (5). Briefly, culture of bone marrow in interleukin-7 (IL-7)-containing medium after 12 to 14 days established >98% pure population of pre-B cells as determined by phenotype analysis using B-cell-specific antibodies and fluorescence-activated cell sorting (FACS). The pre-B cells expressed CD19, B220, and CD24 and were negative for surface immunoglobulin M (IgM) and CD43 irrespective of genotype. IgM⁺/CD19⁺ B cells were sorted from spleens from age- and gender-matched mice: one wild-type and two precancerous Eµ-myc transgenics. All antibodies used for phenotypic analyses were from PharMingen (San Diego, Calif.) or Southern Biotechnology (Birmingham, Ala.).

Virus infection. Virus was produced and used to infect primary pre-B cells as previously described (5). Briefly, MSCV-Myc-ER-IRES-GFP or control MSCV-IRES-GFP virus was cotransfected with helper virus into 293T cells, and live virus was then collected at intervals, pooled, and filtered (49). Viral stocks, MSCV-Myc-ER-IRES-GFP virus or the MSCV-IRES-GFP control virus, were used to infect primary pre-B cells in the presence of 8 µg of Polybrene/ml. Green fluorescent protein (GFP)-positive infected cells were isolated 3 to 4 days postinfection by sterile sorting with a Cytomation MoFlo cell sorter (Fort Collins, Colo.). GFP-positive cells were expanded in IL-7-containing medium and analyzed for levels of Myc-ER (previously referred to as Myc-ER^TM [5, 6]) protein and sensitivity to Myc-induced apoptosis. Myc-ER is a fusion protein of c-Myc linked to a modified estrogen receptor hormone binding domain (25) and is designed to hold Myc-ER in heat shock complexes in the cytosol. Upon addition of 1 µM 4-hydroxytamoxifen (4-HT), which binds to the ER portion of Myc-ER, Myc-ER then translocates to the nucleus and activates transcription (25). As reported elsewhere (5), addition of 4-HT to uninfected or MSCV-IRES-GFP control virus-infected cells had no effect on pre-B cell growth or viability.

Viability and apoptosis assays. Cell viability was determined at specific intervals by trypan blue dye exclusion after the removal of IL-7 or the addition of 1 µM 4-HT (Sigma, St. Louis, Mo.) to the culture medium to activate Myc-ER. For the IL-7 deprivation experiments, wild-type and Bax^/ pre-B cells were washed twice with phosphate-buffered saline and resuspended in medium lacking IL-7 but still containing 10% fetal calf serum. Apoptosis was measured by propidium iodide staining of DNA and quantitation of fragmented (sub-G₁) DNA.

Western blotting. Whole-cell protein extracts from primary pre-B cells or pre-B- or B-cell lymphomas from Eµ-myc transgenic mice were isolated as previously described (5, 63). Briefly, ice-cold lysis buffer (50 mM HEPES, pH 7.5; 150 mM NaCl; 1 mM EDTA; 2.5 mM EGTA; 0.1% Tween 20; 1 mM phenylmethylsulfonyl fluoride; 0.4 U of aprotinin/ml; 1 mM NaF; 10 mM -glycerophosphate; 0.1 mM sodium orthovanadate; 10 µg of leupeptin/ml) was added to cells pellets or small (3- to 5-mm²) tumor chunks. Samples were then subjected twice to sonication for 8 s and centrifuged (4°C, 7 min, 14,000 rpm) to sediment the undissolved cellular material, and then the protein in the supernatant was quantified by using a Bio-Rad Protein Assay (Hercules, Calif.). Equal amounts of protein (200 µg per lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (10%), transferred to nitrocellulose membranes (Protran; Schleicher & Schuell, Dassel, Germany), and blotted with antibodies specific for the p19^ARF (44), p53 (Ab-7), and poly(ADP-ribose) polymerase (PARP; Ab-2) (both from, Calbiochem, La Jolla, Calif.), Mdm2 (C-18; Santa Cruz, Inc., Santa Cruz, Calif.), c-Myc (06-340; Upstate Biotechnology, New York, N.Y.), and Bax (13686E; PharMingen, San Diego, Calif.). Bound immunocomplexes were detected by enhanced chemiluminescence (Amersham, Piscataway, N.J.) or Supersignal (Pierce, Rockford, Ill.).

Southern blotting. Genomic DNA was isolated from lymphomas arising in Bax^+/ and Bax^/ Eµ-myc transgenic mice and digested with AflII or BamHI. Equal amounts of DNA were electrophoretically separated in agarose gels, transferred to nitrocellulose membranes, and then probed with cDNAs coding for ARF (exon 1) (AflII digested) and p53 (exons 2 to 10) and the joining region of immunoglobulin heavy chain (J_H) (both BamHI digested). Genomic DNA isolated from the spleen of a wild-type littermate was used as a control.

Northern blotting. Total RNA was isolated by using TRIzol Reagent according to the manufacturer's directions (Life Technologies, Grand Island, N.Y.) at intervals (0, 1, 3, or 6 h) from primary pre-B cells after addition of 1 µM 4-HT after a 30-min pretreatment with 10 µg of cycloheximide or vehicle control (100% ethyl alcohol)/ml. Northern blotting with 20 µg of total RNA per lane was performed using conventional techniques and probed with the coding portion of murine bax cDNA (kindly provided by John Reed, The Burnham Institute).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Bax loss impairs Myc-induced apoptosis. Bax-null mice have modest increases in B- and T-lymphocyte numbers, presumably due to decreased apoptosis (22). To determine whether Bax-deficient B lymphocytes were intrinsically more resistant to apoptosis than wild-type lymphocytes, we harvested bone marrow cells and expanded pre-B cells in IL-7-containing medium and measured proliferation rates and viability. The Bax-null pre-B cells grew at a rate similar to wild-type pre-B cells (data not shown) and showed no difference in viability (Fig. 1A, open symbols). Moreover, deprivation of IL-7 induced apoptosis in Bax-null pre-B cells with kinetics virtually identical to that of wild-type pre-B cells (Fig. 1A, closed symbols). Similar results were obtained when wild-type and Bax^/ thymocytes were deprived of cytokines (22). Therefore, Bax deficiency does not confer a growth or survival advantage to lymphocytes during ex vivo culture.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Bax does not influence cytokine deprivation-induced apoptosis but does mediate c-Myc-induced apoptosis. (A) Bax+/+ (squares) and Bax/ (triangles) pre-B cells were cultured in medium with (open symbols) or without (solid symbols) IL-7, and their viability was determined at intervals by trypan blue dye exclusion. The data are representative of two independent experiments. (B) 4-HT was added to the indicated primary pre-B-cell cultures to activate Myc-ER, and their viability was determined at intervals thereafter by trypan blue dye exclusion. Apoptosis was confirmed by analysis of subdiploid DNA content after staining with propidium iodide. Steady-state levels of apoptosis in the wild-type primary pre-B cells are indicated at the zero hour time point. (Inset) The protein levels of Myc-ER in Bax / (/) and Bax +/+ (+/+) pre-B cells infected with a retrovirus encoding Myc-ER-GFP and pre-B cells infected with the GFP vector control (V) retrovirus were determined by immunoblotting with an antibody for Myc. The data are the mean of three independent experiments, and error bars represent one standard deviation.

View larger version (33K): [in this window] [in a new window]   FIG. 2.   Myc does not upregulate Bax protein expression in B cells. (A) 4-HT was added to wild-type pre-B-cell cultures infected with the Myc-ER encoding retrovirus to activate Myc-ER. At the indicated intervals cells were collected and protein lysates were made. Equal quantities of protein were assessed by immunoblotting with antibodies specific for Bax, p53, or PARP. The 85- kDa caspase cleavage fragment of PARP is denoted by an asterisk. (B) Equal quantities of protein from FACS sorted IgM+/CD19+ splenic B cells from one wild-type (WT) and two precancerous Eµ-myc transgenics (Tg) were assessed by immunoblotting with an antibody specific for Bax. (C) Total RNA was isolated from Myc-ER-infected wild-type pre-B cells activated with 4-HT for the indicated intervals. Pre-B cells pretreated with cycloheximide are indicated by a plus sign. The expression of bax transcripts was assessed by Northern blot analyses utilizing bax cDNA. (D) 4-HT was added to wild-type, ARF/, p53/, and ARF/p53/ pre-B-cell cultures infected with the Myc-ER-encoding retrovirus to activate Myc-ER. At the indicated intervals, cells were collected and protein lysates were prepared. Equal quantities of protein were evident by immunoblotting with antibodies specific for Bax, and the levels of Myc-ER protein expressed in the four genotypes were equivalent (5).

Bax loss accelerates Myc-induced lymphomagenesis. Loss of ARF or p53 impairs Myc-mediated apoptosis and consequently accelerates Myc-induced lymphomagenesis (5, 50, 63). To test the genetic contribution of Bax to Myc-induced lymphomagenesis, we crossed Eµ-myc transgenic mice onto the Bax-null background. Congenic C57BL/6 Eµ-myc transgenic mice were mated to C57BL/6 × 129/svj Bax^+/ mice, since Bax^/ males and females have fertility problems (22, 43). F₁ littermates were intercrossed to obtain Bax^+/+, Bax^+/, and Bax^/ Eµ-myc transgenic mice, and these transgenics were carefully monitored for disease development. Notably, Bax loss accelerated the course of lymphoma development, decreasing the average age of survival from 21.7 weeks in Bax^+/+ Eµ-myc transgenics to 12.6 weeks in Bax-null Eµ-myc transgenics (Fig. 3). None of the Bax^/ Eµ-myc transgenic mice survived past 30 weeks, whereas 22% (12 of 54) of the Bax^+/+ Eµ-myc transgenics lived longer than 30 weeks. Bax haploinsufficency also accelerated tumor development, since Bax^+/ Eµ-myc transgenic mice had a mean life span of 16.0 weeks (Fig. 3). Therefore, Bax impairs Myc-induced lymphomagenesis. FACS analysis with lymphoid-specific antibodies showed that the Bax-null and Bax^+/ Eµ-myc transgenics develop pre-B- and B-cell lymphoma typical of wild-type Eµ-myc transgenics (data not shown) and not the primitive lymphoid tumor described in the Eµ-myc/Eµ-bcl-2 double-transgenic mice (54). Predictably, Southern blot analysis revealed that all but one of the lymphomas that arose in Bax-null Eµ-myc transgenics were clonal (data not shown), which is characteristic of the lymphomas that develop in Eµ-myc transgenics (1).

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Myc-induced lymphomagenesis is accelerated by Bax loss. The genotypes of the mice are indicated next to the Kaplan-Meier survival curves, and the numbers of mice in each group are denoted by the n values. Vertical lines indicate ages of surviving mice: 0 Bax/, 3 Bax+/, and 10 Bax+/+ mice. The average life spans of Bax+/+, Bax+/, and Bax/ Eµ-myc transgenics were 21.7, 16.0, and 12.6 weeks, respectively. Pre-B- and/or B-cell lymphoma was documented in all of the animals.

Bax^+/

Bax^+/

Bax^+/

Bax^+/

View larger version (56K): [in this window] [in a new window]   FIG. 4.   Western blot analysis of lymphomas arising in Bax+/ and Bax/ Eµ-myc transgenic mice. Levels of Bax (A), p53 (B, top), p19ARF (B, middle), and Mdm2 (B, bottom) protein in whole-cell extracts of tumors from Bax+/ and Bax/ Eµ-myc transgenic mice were assessed by immunoblotting with antibodies specific for each protein. Protein extracts from a tumor arising in a Bax+/+ Eµ-myc transgenic mouse that contains the p92, p90, and p85 Mdm2 isoforms were run to show the location of these isoforms in panel B and as a blotting control for Bax expression in panel A. The asterisk in panel B marks the position of a nonspecific background band detected with the Mdm2 antibody.

Bax loss selectively bypasses p53 mutations that arise during Myc-induced lymphomagenesis. ARF-null animals are tumor prone and spontaneously develop sarcomas and lymphomas within 8 months of age (19). ARF is upregulated after Myc activation, and loss of ARF impairs Myc-induced apoptosis and accelerates lymphomagenesis in Eµ-myc transgenic mice (5, 50). To determine whether the loss of Bax influences tumorigenesis initiated in ARF-deficient mice, we crossed ARF- and Bax-deficient mice, and F₁ mice were then mated to generate ARF/Bax-double null mice. The ARF^/Bax^/ mice were monitored for tumor development and compared with the tumor latency in ARF^/Bax^+/+ littermates. We found that Bax deficiency does not alter the survival of ARF-null mice, since the average age of survival for Bax^+/+ ARF^/ (48.9 weeks, n = 37) and Bax^/ ARF^/ (47.6 weeks, n = 39) mice was essentially equivalent. Therefore, loss of Bax does not influence the overall survival of ARF-null mice. At face value this could suggest that Bax is in the same pathway as ARF. On the other hand, the results could indicate that ARF and Bax exist in separate pathways that function independently of each other. In support of the latter concept, biallelic deletion of ARF occurred in lymphomas from both Bax^/ Eµ-myc transgenics (12%) and Bax^+/ Eµ-myc transgenic mice (20%) (Fig. 5; Table 1), indicating that Bax loss does not prevent inactivation of ARF. The percentage of tumors with ARF deletions was slightly lower than previously reported (5), but this could be due to experimental variation or (more likely) to the different background of this cross from the one previously reported. Indeed, genetic background effects are evident when the differences in average survival of wild-type Eµ-myc transgenics, 22 weeks in the Bax background (Fig. 3) and 33 weeks in the ARF background, are compared (5).

View larger version (69K): [in this window] [in a new window]   FIG. 5.   Southern blot analysis of Bax+/ and Bax/ Eµ-myc lymphomas. AflII and BamHI restriction fragments containing ARF exon 1 and p53 exons 2 to 10, respectively, from genomic DNA isolated from lymphomas arising in Bax+/ (A) and Bax/ (B) Eµ-myc transgenic mice. Genomic DNA from the spleen of a wild-type littermate was used as a control in both panels A and B. Lack of a band denotes biallelic deletion of that gene.

View this table: [in this window] [in a new window]   TABLE 1.   p53, ARF, and Mdm2 protein expression in lymphomas from Bax+/ and Bax/ Eµ-myc transgenics

Bax^+/

Bax^+/

Bax^+/

Bax^+/

Bax^/

Bax^+/

Bax^/

Bax^+/

View larger version (22K): [in this window] [in a new window]   FIG. 6.   Schematic describing the outcome of molecular events that occur in lymphomas in Bax+/+, Bax+/, and Bax/ Eµ-myc transgenic mice. On the left, the majority of lymphomas that arise in Bax+/+ and Bax+/ Eµ-myc transgenic mice have alterations in ARF (deletion, X) or p53 (mutation or deletion, X) and/or Mdm2 (overexpression, ) (Fig. 4 and 5 and Table 1; see also reference 5). On right, a similar frequency of alterations in ARF and Mdm2 still occurs in lymphomas from Bax-null Eµ-myc transgenic mice; however, no p53 mutations or deletions are observed in these lymphomas (Fig. 4 and 5, Table 1). Therefore, Myc targets ARF and Mdm2 but not p53 in the absence of Bax during Myc-induced lymphomagenesis.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

B cells lacking mediators of Myc-induced apoptosis have an accelerated course of lymphoma development in Eµ-myc transgenic mice (5, 17, 50). For example, deleting ARF or p53, both mediators of Myc-induced apoptosis, accelerates lymphomagenesis in Eµ-myc transgenic mice (5, 50). Here, by using in vivo models, we extend these observations to the proapoptotic Bcl-2 family member Bax and show that intersecting apoptotic pathways play a crucial role in Myc-induced lymphomagenesis. Loss of Bax in pre-B cells confers resistance to Myc-induced apoptosis and accelerates pre-B- and B-cell lymphoma development in Eµ-myc transgenic mice. This is consistent with the finding that Bax-null mouse embryo fibroblasts (MEFs) are more resistant to Myc-induced apoptosis (32). Therefore, Bax is a mediator of Myc-induced apoptosis and inhibits Myc-initiated tumorigenesis.

The balance of proapoptotic and antiapoptotic Bcl-2 family members regulates the susceptibility of cells to apoptosis (reviewed in reference 23). An excess of Bax induces cell death, whereas overexpression of Bcl-2 or Bcl-X_L suppresses apoptosis induced by a variety of apoptotic stimuli. In immortal human cells bax has been reported to be a transcriptional target of both p53 (34, 35) and Myc (32), yet we failed to detect any direct or significant increase in Bax expression upon Myc activation in primary murine pre-B cells. Nevertheless, activation of p53 can induce bax and suppress bcl-2 expression in certain cell types (33-35). Myc, on the other hand, upregulates p53 and ARF (5, 63) and suppresses Bcl-X_L expression, and the latter is independent of either p53 or ARF in primary murine hematopoietic cells (6). Thus, Myc activation alone is sufficient to alter the ratio of pro- and antiapoptotic Bcl-2 family members with the net result being an excess of Bax, which leads to apoptosis. Bax loss short-circuits this response and confers survival to cells overexpressing Myc.

p53 is a mediator of Myc-induced apoptosis (5, 63), and Bax plays an important role in p53-dependent apoptosis in some cell types in vitro (28, 35) and in vivo (52, 62). This study confirms and extends these results by linking Myc with p53 and Bax in vivo. Lymphomas arising in Bax-deficient Eµ-myc transgenic mice lack p53 mutations and deletions, whereas 28% of tumors from wild-type Eµ-myc transgenics (5) and 27% of lymphomas from Bax^+/ Eµ-myc transgenic mice sustain mutations or deletions in p53. Therefore, under selective pressure from Myc, Bax-deficient B cells differ from their wild-type counterparts by sustaining wild-type p53 expression. These results imply that during transformation mutation of p53 is unnecessary when Bax is absent, supporting the observations that Bax is downstream from p53 (10). This is consistent with the observation that there is no cooperative effect on the rate of tumorigenesis in mice lacking both p53 and Bax compared to mice deficient in p53 alone (20). Moreover, when the statuses of ARF and p53 were analyzed in lymphomas that arise in Eµ-myc transgenic mice, biallelic deletion of ARF or p53 deletion or mutation occurred in a mutually exclusive fashion (5). Thus, although ARF and p53 can function in the same tumor suppressor pathway, our results demonstrate that this pathway must bifurcate, since only Bax and p53 are in the same Myc-induced pathway (Fig. 6). Indeed, this study provides formal genetic proof that ARF, Mdm2, and p53 have different targets.

One of the most intriguing outcomes of these studies is the finding that that Bax loss influences p53 status independent of ARF (Fig. 6). Bax deficiency did not affect the frequency of ARF deletions in Eµ-myc lymphomas nor tumor latency in ARF-null mice, suggesting that Bax and ARF are not in the same pathway, or that Bax resides in a position in the pathway that does not influence ARF. However, further analysis has demonstrated that Bax loss does affect the tumor spectrum in ARF-null mice (C. M. Eischen and J. L. Cleveland, unpublished data). Thus, there is cooperativity at least at some level, and ARF and Bax must function in separate pathways.

Even though ARF and p53 function in the same tumor suppressor pathway, p53 has been reported to function independently of ARF in certain situations and vice versa. For example, gamma irradiation-induced apoptosis is p53 dependent and still occurs in ARF-null cells (19), and ARF can still induce cell cycle arrest in fibroblasts lacking p53 and Mdm2 (58). Our data from Bax-null Eµ-myc transgenic mice reveals a role for Bax that is dependent on p53 but independent of ARF. Myc thus affects the p53 pathway in two ways: by upstream activation through ARF and by downstream activation of Bax. However, the loss of Bax does seem to disrupt the ARF-Mdm2-p53 pathway by other means. The increased percentage of lymphomas in Bax-null Eµ-myc transgenics that overexpressed ARF (Table 1) suggests that Bax expression could influence the ARF-Mdm2-p53 pathway by somehow targeting proteins that regulate ARF expression, such as Bmi-1 (16), Dmp-1 (13), JunD (60), Tbx2 (15), and Twist (27). Alternatively, Bax status could disrupt the delicate feedback control mechanisms that regulate the expression of ARF, Mdm2, and p53 (47, 53).

Mdm2 is a negative regulator of p53, and we therefore predicted that, since the loss of Bax bypassed requirements for inactivating p53, then the frequency of Mdm2 overexpression would be decreased as well. Surprisingly, Mdm2 was overexpressed in 50% (16 of 32) of all of the lymphomas analyzed, as shown in wild-type Eµ-myc transgenics (5), regardless of Bax status. Mdm2 has many targets other than p53 (e.g., E2F-1, DP-1, p300, and pRb) (reviewed in reference 36) and appears to function independently of the p53 pathway in certain scenarios. For example, Mdm2 is overexpressed in a third of all lymphomas that had mutated or deleted p53 (5), and haploinsufficiency of Mdm2 alters the tumor spectrum in p53-null mice (29). Moreover, Mdm2 transgene overexpression resulted in altered mammary gland development (26) and tumorigenesis (18) in p53^/ mice. Therefore, selection for Mdm2 overexpression in Eµ-myc lymphomas is not influenced by Bax status and is also not necessarily linked to alterations in p53 or ARF (Fig. 6).

Bax does not function as a classic tumor suppressor in Eµ-myc-induced lymphomas, as do ARF and p53 (reviewed in reference 51). However, Bax tumor suppressor function has been reported in some scenarios. First, a Bax frameshift mutation occurs in a subset of colon adenocarcinomas (45), and cells with this mutation display a survival advantage when transplanted into nude mice (14). Second, Bax deficiency accelerates brain and breast tumor development in SV40 large T antigen-transgenic mice (52, 62). Finally, there is increased foci formation in transformation assays by using Bax-deficient MEFs (28). In contrast, Bax is not mutated in any of the Bax^+/+ or Bax^+/ Eµ-myc tumors analyzed. Additionally, there was no loss of heterozygosity of Bax in mammary carcinomas from Bax^+/ C3 (1)/SV40 large T antigen-transgenic mice (52) or in lymphomas from Bax^+/ Eµ-myc transgenics (Fig. 4A). These latter findings are consistent with the observation that Bax is infrequently mutated in most types of human B-cell lymphoma (42). Moreover, Bax-null mice do not spontaneously develop cancer (22), and loss of Bax did not influence the survival of ARF-null mice. Therefore, Bax may function as a classic tumor suppressor under very specific circumstances but acts as a modifier in most situations, such as Myc-induced lymphomas. Our observations also indicate that targets in addition to Bax must contribute to p53-dependent tumor suppression.

It has been known for over a decade that Bcl-2 and Myc can cooperate in transformation. The malignancy in Eµ-myc/Eµ-bcl-2 double-transgenic mice is composed of primitive lymphoid cells (54), rather than the pre-B and/or mature B cells typical of Eµ-myc lymphomas (1). Although loss of Bax also accelerates lymphomagenesis in Eµ-myc transgenics, flow cytometric analysis clearly indicates that these lymphomas are of pre-B- and/or mature B-cell origin. Thus, overexpression of Bcl-2 blocks Myc-induced apoptosis and alters the differentiation of the lymphoid cell, while Bax loss alters the sensitivity to Myc-induced apoptosis without overtly affecting B-cell differentiation. Thus, Bcl-2 and Bax appear to provide different developmental roles when B-cell precursors are forced to proliferate by Myc expression. Although Bcl-2 can inhibit the apoptotic effects of Bax (reviewed in reference 23), Bcl-2 and Bax can regulate apoptosis independently of each other (21). Moreover, Bcl-2 is overexpressed at the same frequency in Bax^/ and Bax^+/+ Eµ-myc lymphomas (Eischen and Cleveland, unpublished). Therefore, it will be interesting to evaluate the status of p53, ARF, and Mdm2 in the tumors from Eµ-myc/Eµ-bcl-2 double transgenics in order to determine whether Bcl-2 and Bax function in common or distinct pathway(s) in relationship to p53.