Endogenous Bcl-x_L is essential for Myc-driven lymphomagenesis in mice

Impaired apoptosis is considered a prerequisite for the development of most, if not all, cancers,^1-3  yet the mechanisms that promote the survival of most nascent malignant cells during the process of neoplastic transformation remain unknown. Apoptosis is regulated by opposing factions of the Bcl-2 family, which include both proteins essential for cell survival and those that drive cell death.^4,5  Despite very similar biochemical functions, the pro-survival family members (Bcl-2, Bcl-x_L, Bcl-w, Mcl-1, and A1) have proven to have essential functions in specific cell types.⁵  The critical initiators of apoptosis are the “BH3-only” proteins (eg, Bim, Puma, Bid), so-called because they share only a single Bcl-2 homology (BH) domain with other family members. They are activated by developmental cues and diverse stress stimuli, including cytokine deprivation, DNA damage and activation of oncogenes, such as c-Myc.^6-10  A second proapoptotic Bcl-2 sub-family, principally represented by the multi-BH domain proteins Bax and Bak, mediates the pivotal downstream step of mitochondrial outer membrane permeabilization (MOMP), which evokes activation of the caspase cascade that demolishes the cell.^4,11  The BH3-only proteins may provoke activation of Bax and Bak by their direct engagement, by sequestering pro-survival relatives or both ways.^4,5,12-14 

Studies using transgenic mice have established that overexpression of Bcl-2^15-18  or its antiapoptotic relatives, including Bcl-x_L¹⁹  and Mcl-1,^20,21  or loss of proapoptotic Bcl-2 family members, such as Bim¹⁰  or Puma,^22-24  contribute to tumor development, particularly in conjunction with mutations that deregulate cell cycle control, such as enforced c-Myc expression. Importantly, lymphomas evoked by combined overexpression of Bcl-2 and c-Myc require sustained Bcl-2 overexpression,²⁵  most likely to counter the apoptosis promoted by c-Myc overexpression under stress conditions, such as limiting supply of growth factors.^26-29  However, because only a proportion of cancers contain cytogenetic alterations (chromosomal translocations or somatically acquired copy number alterations) that directly promote the overexpression of Bcl-2 or one of its homologs,^2,3,30  we surmise that most cells undergoing neoplastic transformation are sustained (at least initially) by the normal expression of endogenous Bcl-2–like antiapoptotic proteins. Identification of the pro-survival Bcl-2 proteins critical for development of a tumor is expected to guide the design of prophylactic and early intervention strategies.

We previously investigated the contribution of endogenous Bcl-2 to lymphoma development³¹  by interbreeding Eμ-myc transgenic mice³²  with Bcl-2–deficient mice^33,34  and using fetal liver-derived stem cells from their offspring to reconstitute lethally irradiated wt mice with an Eμ-myc/bcl-2^−/− (or control genotype) hematopoietic system. In the transgenic pre-malignant B lymphoid cells, Bcl-2 proved critical for the survival of mature (sIg⁺) B cells but largely dispensable for that of pro-B and pre-B cells. Despite the marked deficit in mature pre-leukemic B lymphoid cells, however, the onset and incidence of Eμ-myc pre-B/B lymphoma were unaffected.³¹  This suggested that the tumors arise from B cell precursors, whose survival during neoplastic transformation might be sustained by a Bcl-2 relative. An appealing candidate is Bcl-x_L, because it is highly expressed in early B lineage cells³⁵  and required for their maintenance.³⁶  Indeed, the studies reported here demonstrate that Bcl-x_L is essential for the survival of Myc-driven pre-leukemic cells and for lymphoma development.

Experiments with mice were conducted with approval from The Walter and Eliza Hall Institute Animal Ethics Committee. Eμ-myc transgenic³²  and bcl-x^+/−36  mice have been described. These strains had been backcrossed to the C57BL/6-Ly5.2 genetic background for > 30 or > 10 generations, respectively. The rag-2^−/− mice³⁷  (Taconic) were backcrossed to the C57BL/6-Ly5.1 genetic background for > 15 generations.

Eμ-myc/bcl-x^−/− E12.5 (Ly5.2) embryos were generated as follows: Eμ-myc transgenic males were first crossed with bcl-x^+/− females and their Eμ-myc/bcl-x^+/− male offspring were then mated with bcl-x^+/− females. The day when the vaginal plug was detected was deemed embryonic day 0.5. Embryos from timed matings were harvested at E12.5 and genotyped by PCR on tail-derived DNA. Immediately before injection, single cell suspensions were prepared from fetal livers and ∼ 2 × 10⁶ cells mixed with ∼ 1 × 10⁵rag-2^−/− (C57BL/6-Ly5.1) bone marrow leukocytes and injected into the tail vein of lethally irradiated (2 × 5.5 Gy, at 3-hour intervals) C57BL/6-Ly5.1 mice. Mice were maintained on neomycin sulfate-supplemented drinking water for 14 days after irradiation to prevent infection.

Mice were monitored daily for signs of lymphoma development. Tumor-free survival was defined as the time from lethal irradiation and hematopoietic reconstitution with fetal liver cells to the time the animal was deemed ill by an experienced animal technician. To verify that mice used for pre-leukemic analysis lacked malignant cells, 1 × 10⁶ of their bone marrow and/or spleen cells were transplanted into non-irradiated histocompatible recipient mice, which were then monitored for 90 days for the development of any tumor. Kaplan-Meier survival curves were constructed using GraphPad Prism Version 5.0, and statistical analysis performed using a 1-way ANOVA test (for comparison of 3 cohorts), or log-rank (Mantel-Cox) test (for comparison of 2 cohorts).

Bone marrow (both femora), spleen and lymph nodes (combined mesenteric, inguinal and axillary) were harvested and single cell suspensions prepared. Total cell numbers were determined by trypan blue staining and counting in a hemocytometer. Absolute numbers in each cellular compartment were calculated by multiplying the percentage of a cell type (as determined by FACS analysis) by the total organ cellularity. Donor-derived cell numbers were calculated by multiplying the percentage donor-derived (Ly5.2⁺) cells by the total organ cellularity. Peripheral blood was harvested via retro-orbital bleed or at sacrifice by cardiac puncture and collected into heparinized vessels. Total white blood cells (WBCs), red blood cells (RBCs), the percentage of hematocrit and hemoglobin levels were determined using an Advia 120 blood analyzer equipped with a mouse analysis software module (Bayer).

Before FACS-based immuno-phenotyping, peripheral blood was depleted of red blood cells by incubation (2 × 5 minutes) with red cell lysis buffer (56mM NH₄Cl, 0.1mM EDTA, 12mM NaHCO₃; pH 7.3) at 4°C. To prevent non-specific antibody binding, cells were incubated in the presence of 2.4G2 (anti-FcgRII) antibody plus 2% normal rat serum. Host (Ly5.1⁺) and donor-derived (Ly5.2⁺) cells were discriminated by staining with anti-Ly5.1 (A201.1) and anti-Ly5.2 (5.450.15.2) monoclonal antibodies. Pre-leukemic pro-B (B220⁺c-Kit⁺sIgM⁻sIgD⁻), pre-B (B220⁺c-Kit⁻sIgM⁻sIgD⁻) and sIg⁺ (B220⁺c-Kit⁻sIg⁺) B lymphoid populations were purified by FACS sorting after staining with monoclonal antibodies to: B220 (RA3-6B2), c-Kit (ACK2 or ACK4), IgM (5.1 or 333.12) and IgD (11-26C). Antibodies were produced in our laboratory and conjugated to biotin, FITC (both from Molecular Probes), cyanine 5 (Cy5, Amersham), R-phycoerythin (R-PE) or allophycocyanin (APC, both from Prozyme) according to the manufacturers' instructions. Biotinylated antibodies were detected by secondary staining with FITC-, PE- or Tricolor-coupled streptavidin (Caltag). Dead cells were excluded by staining with propidium iodide (2 μg/mL). Purification of B lymphoid sub-populations was performed by multi-parameter FACS sorting using a MoFlo (Cytomation) or DiVa (BD Biosciences) high-speed flow cytometer.

Bone marrow was harvested from the femora of pre-leukemic mice at 8-12 weeks after reconstitution. Fetal liver donor-derived pro-B (Ly5.2⁺B220⁺c-Kit⁺sIg⁻) and pre-B cells (Ly5.2⁺B220⁺c-Kit⁻sIg⁻) were purified by multi-parameter FACS sorting, fixed in 70% ice-cold ethanol and placed at −20°C overnight. Cells were washed twice in MT-PBS and incubated with 200 μL PI solution (38mM Na₃C₆H₅O₇, pH 7.4, 690nM propidium iodide, 5 μg/mL RNaseA) at 37°C for 20 minutes. DNA content was measured by acquiring an equal number of events (> 10⁴ cells/sample) using a FACScan (BD Biosciences) with linear amplification settings.

FACS-sorted B lymphoid sub-populations were cultured in a humidified incubator (10% CO₂) at 37°C in flat bottom 96-well microtitre plates at 2-5 × 10⁴ cells/100 μL (per time point) in the high glucose version of DMEM supplemented with 250μM L-asparagine, 50μM 2-mercaptoethanol and 10% heat-inactivated FCS (JRH Biosciences). Cells were cultured in simple medium (no added growth factors) to assess the effects of cytokine deprivation. Cells were harvested at the indicated time points and viability determined by staining with PI and FITC-conjugated annexin V followed by flow cytometric analysis using a FACScan analyzer (BD Biosciences).

Because endogenous Bcl-2 proved dispensable for Myc-induced lymphomagenesis³¹  and Bcl-x_L is expressed at relatively high levels in pro-B and pre-B cells,³⁵  the stages from which Εμ-myc lymphomas are thought to arise,^32,38,39  we sought to examine the role of Bcl-x_L in this process. To do so, we first crossed Εμ-myc transgenic³²  mice with bcl-x^+/− mice³⁶  and compared lymphoma development in their progeny (Figure 1). Remarkably, loss of a single bcl-x allele significantly delayed the lymphomagenesis: the median survival rose from 116 days for Εμ-myc mice to 174 days for Εμ-myc/bcl-x^+/− animals (P = .015). In contrast, and consistent with our previous study,³¹  loss of 1 allele of bcl-2 did not significantly affect Εμ-myc–induced lymphomagenesis (Figure 1; P = .303).

To explore the impact of complete loss of endogenous Bcl-x_L in Myc-driven lymphomagenesis, we circumvented the embryonic lethality caused by germ line bcl-x deficiency³⁶  by generating hematopoietic chimeric mice. To this end, lethally irradiated C57BL/6-Ly5.1⁺ (wt) mice were transplanted with hematopoietic stem/progenitor cells from the fetal liver of E12.5 Eμ-myc/bcl-x^−/− (Ly5.2⁺) embryos or, as controls, Eμ-myc/bcl-x^+/+ (hereafter “Eμ-myc”), bcl-x^−/− or wt embryos (Figure 2A). The erythroid deficit caused by Bcl-x_L deficiency³⁶  was overcome by coinjecting hematopoietic stem/progenitor cells from rag-2^−/− (C57BL/6-Ly5.1) mice (Figure 2B), which contribute to erythroid (and myeloid) but not lymphoid development.³⁷  Analysis of the peripheral blood also revealed that the Eμ-myc/bcl-x^−/− fetal liver cells could readily engraft and yield cells of the B lymphoid and other hematopoietic lineages (Figure 2C).

To determine how loss of endogenous Bcl-x_L affected development of Εμ-myc transgenic B cells, we isolated the blood, bone marrow, spleen and lymph nodes of pre-leukemic mice at 8-12 weeks after reconstitution and enumerated pro-B, pre-B and sIg⁺ B cells by flow cytometric analysis (Figure 3). Because deregulated c-Myc expression enhances proliferation of B lymphoid precursors but retards their subsequent differentiation, the bone marrow of Eμ-myc mice displayed a marked increase in pre-B cells (B220⁺c-Kit⁻ sIg⁻) over non-transgenic mice (Figure 3A), but the bone marrow, blood, spleen and lymph nodes all had several-fold fewer sIg⁺ B cells (Figure 3A-D), consistent with previous observations.³⁸  Notably, Bcl-x_L loss significantly reduced the level of Εμ-myc pro-B (P < .05) and pre-B cells (P < .01) in the bone marrow (Figure 3A), as well as the sIg⁺ B cells (P < .01) in the bone marrow, blood, spleen and lymph nodes (Figure 3A-D).

Overexpression of c-Myc enhances the apoptosis of diverse cell types (including lymphocytes) subjected to many stress stimuli, including growth factor deprivation.^26,29  To determine whether the reduction in early B lymphoid cells in Eμ-myc/bcl-x^−/− mice reflected enhanced c-Myc-induced apoptosis, we analyzed their survival when cultured without cytokine support (Figure 4A). As reported,^29,31  Eμ-myc pro-B and pre-B cells died faster than their non-transgenic counterparts (not shown). Notably, Bcl-x_L deficiency provoked significant (P < .05) further acceleration of cell death: by 16 hours of culture, the viability of Eμ-myc/bcl-x^−/− pro-B or pre-B cells was only ∼ 25 to ∼ 30% that of their Eμ-myc counterparts (Figure 4A). Flow cytometric analysis confirmed that the exacerbated cell death in culture of the Myc-driven cells lacking Bcl-xL reflected apoptosis (supplemental Figure 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). As expected, loss of Bcl-x_L did not affect the enhanced cycling of the B lymphoid cells evoked by Εμ-myc transgene expression (Figure 4B). Thus, whereas endogenous Bcl-2 is critical primarily for the survival of pre-leukemic sIg⁺ Εμ-myc B lymphoid cells,³¹  endogenous Bcl-x_L is also essential to maintain their precursors.

To determine how loss of endogenous Bcl-x_L affects lymphomagenesis, we monitored cohorts of the reconstituted mice. Eμ-myc reconstituted mice developed donor-derived (Ly5.2⁺) lymphoma with a median onset of 22 weeks and 40% mortality by 65 weeks (Figure 5), yielding ∼ 75% pre-B (B220⁺sIg⁻) and ∼ 25% sIg⁺ B lymphomas. In striking contrast, none of 25 Eμ-myc/bcl-x^−/− animals monitored for 70 weeks developed a tumor (Figure 5; P < .001), and autopsy after this time revealed no overt signs of lymphoma.

The discovery that the bcl-2 gene, which is commonly translocated in follicular lymphoma,⁴⁰  promotes cell survival¹⁵  and lymphomagenesis in transgenic mice^16-18  engendered the now widely accepted concept that evasion of apoptosis is a critical step in neoplastic transformation.^1-3  Nevertheless, because most tumors do not contain genomic aberrations that directly cause overexpression of a Bcl-2 homolog, it seems likely that the development of such cancers relies on the normal expression of an endogenous Bcl-2–like protein. Although Bcl-2 overexpression greatly accelerates lymphomagenesis in Eμ-myc mice,¹⁶  we found that endogenous Bcl-2 is dispensable for lymphoma development in these cancer-prone animals.³¹  In contrast, we show here that endogenous Bcl-x_L is essential for Eμ-myc–driven lymphomagenesis (Figure 5).

The finding that Bcl-x_L loss but not Bcl-2 deficiency prevented Myc-induced lymphoma development initially appeared surprising because loss of either gene reduced the accumulation of Eμ-myc–expressing pre-leukemic cells. However, whereas loss of either endogenous Bcl-2³¹  or Bcl-x_L (Figure 3) provoked a ∼ 5- to 10-fold drop in pre-leukemic sIg⁺ Eμ-myc B cells, only Bcl-x_L loss substantially reduced the numbers of pre-leukemic pro-B and pre-B cells, sensitized them to apoptosis (Figure 3A) and abrogated lymphomagenesis (Figure 5). We surmise that loss of Bcl-x_L but not of Bcl-2 ablated tumorigenesis primarily because Bcl-x_L³⁵  is expressed more highly than Bcl-2⁴¹  in pro-B and pre-B cells, the cell types that most readily acquire collaborating oncogenic mutations because of their high proliferation rate and the genomic instability associated with Ig gene rearrangement.⁴² 

Two of our findings help to explain how the absence of Bcl-x_L leads to ablation of Myc-induced lymphomagenesis (see model in Figure 6). The first is the approximately 5-fold lower number of pro-B and pre-B cells in the bone marrow of Eμ-myc/bcl-x^−/− than Eμ-myc mice (Figure 3A). This decrease in the likely target cells for transformation almost certainly reflects their greater sensitivity to in vivo stresses, including in particular the apoptotic impetus promoted by Myc under conditions of stress, such as cytokine deprivation (Figure 4A). However, reduction in target cell numbers alone should markedly reduce but not ablate malignant transformation. Hence, the increased apoptosis evident on culture of the transgenic pro-B and pre-B cells lacking Bcl-x_L suggests that the demise of these cells continues during the early stages of transformation, with Myc-driven proliferation balanced by Myc-induced apoptosis (Figure 6A).

Myc-driven apoptosis is understood in outline (Figure 6B). Over-expression of c-Myc increases expression of Bim and Puma, and loss of either of these proapoptotic BH3-only proteins accelerates Εμ-myc–induced lymphomagenesis by reducing Myc-induced apoptosis^10,22-24,43 . Hence, we surmise that in the absence of Bcl-x_L the elevated levels of Bim and Puma induced by c-Myc overwhelm the remaining antiapoptotic Bcl-2 relatives (eg, Bcl-2) in early B lymphoid cells and provoke the activation of Bax and Bak (Figure 6B). The resulting apoptosis may both reduce the target population for transformation and eliminate pre-neoplastic clones otherwise destined for malignancy (Figure 6A). Pertinently, ∼ 20% of human mantle cell lymphomas exhibit bi-allelic BIM deletion⁴⁴  and the c-MYC translocations that drive Burkitt lymphomas are reportedly sometimes accompanied by point mutations in c-MYC that prevent it from increasing Bim expression.⁴³ 

Given that c-Myc is abnormally highly expressed in the great majority of human cancers,⁴⁵  our demonstration that Bcl-x_L is essential for Εμ-myc–induced lymphoma development has important implications for both oncogenesis and anti-cancer therapy. Notably, in many diverse human malignancies the genes for Bcl-x_L (or Mcl-1) are amplified, most often in concert with MYC amplification.³⁰  In addition, Burkitt lymphomas frequently express Bcl-x_L,⁴⁶  and notably one such lymphoma was recently found to bear not only the hallmark MYC translocation but also a translocation of BCL-X that drove its high expression.⁴⁷ 

The BH3 mimetic ABT-737⁴⁸  and the closely related ABT-263 (navitoclax),⁴⁹  which target Bcl-2, Bcl-x_L, and Bcl-w but not Mcl-1, are showing promise as anti-cancer agents.¹⁴  Our results suggest that specific BH3 mimetic drugs could also prevent the development of certain types of tumors by unleashing the proapoptotic impetus of the oncogenic changes they have sustained (eg, deregulated c-Myc expression). Hence, such drugs could contribute to the management of individuals with hereditary predisposition to malignancies and to reduce the likelihood of relapses or secondary malignancies in patients with sporadic cancers. A potential hazard would be induction of immune deficiency, but a recent study showed that conditional deletion of Bcl-x in antigen-activated B cells of mice had no effect on development of germinal centers or memory B cells and little on plasma cells, whereas deletion of Mcl-1 essentially abolished all 3 populations.⁵⁰  Hence, a BH3 mimetic specific for Bcl-x_L could have prophylactic potential for individuals at high risk of developing lymphomas and perhaps other tumors driven by Myc.

The authors thank Drs S. Cory, D. Lo, D. Huang, S. Nutt, P. Bouillet, M. Bath, and L. Tai for mice, reagents, and advice; K. Vella, G. Siciliano, D. Cooper, N. Iannarella, and J. Coughlin for animal care; J. Corbin for automated blood analysis; B. Helbert and C. Young for genotyping; Dr F. Battye, V. Milovac, C. Tarlinton, and J. Garbe for cell sorting; and D. Quilici, T. Nikolaou, and G. Thomas for irradiation.

This work was supported by grants and fellowships from the Cancer Council of Victoria (to P.N.K.), the National Health and Medical Research Council (Program Grant #461221, NHMRC Australia Fellowship), the National Institutes of Health (CA43540), and the Leukemia & Lymphoma Society (SCOR grant #7413). This work was made possible through State Government Operational Infrastructure Support and Australian Government NHMRC IRIISS.

National Institutes of Health

Contribution: P.N.K. planned research, conducted experiments, interpreted data, and wrote manuscript; S.G. and A.R.D.D. conducted experiments, interpreted data, and helped with writing of manuscript; and J.M.A. and A.S. conceived the study, planned experiments, interpreted data, and wrote manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

The current affiliation for P.N.K. is Metabolism Branch, National Cancer Institute/National Institutes of Health, Bethesda, MD.

Correspondence: Andreas Strasser, PhD or Jerry Adams, PhD, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, 3052, Victoria, Australia; e-mail: strasser@wehi.edu.au or adams@wehi.edu.au.