Bcl2 is not required for the development and maintenance of leukemia stem cells in mice

Abstract

The existence of leukemia stem cells (LSCs) responsible for tumor maintenance has been firmly established. Therefore, therapeutic targeting of these LSCs may have a profound impact on cancer eradication. The anti-apoptotic protein Bcl2 has been proposed as a therapeutic target, but its role in LSC biology has not been investigated. In order to understand the role of Bcl2 in LSC generation and maintenance, we have taken advantage of our Sca1-BCRABLp210 mouse model of human chronic myeloid leukemia and bcl2 gene-targeted mice. This study provides genetic evidence that the inhibition of Bcl2 is not critical for the generation, selection or maintenance of the tumor initiating and maintaining cells in mice.

Introduction

The interference with the apoptotic processes leads to tumor development, as it is demonstrated in the cases of p53 losses or Bcl2 overexpression (1). High levels of Bcl2 expression have been found in many cancers, including chronic myeloid leukemia (CML) (2), follicular lymphoma (3), chronic lymphocytic leukemia (4), acute myelogenous leukemia (5), myeloma (6), small-cell lung cancer (7) and melanoma (8). The elimination of Bcl2 rapidly induces apoptosis and remission of differentiated tumor cells (2,9). These findings support the view of Bcl2 as a potential therapeutic target and have paved the way for the development of Bcl2 antagonists for anticancer therapy (10).

In contrast to these findings in differentiated tumor cells, the contribution of Bcl2 to the biology of leukemia stem cells (LSCs) has not been previously investigated. Dissecting the contributions of altered Bcl2 activity to LSC biology should provide valuable mechanistic insights and also help to determine if the presence of this genetic lesion can be used to gain a therapeutic advantage against the neoplastic disease. In this study, we have investigated the role of Bcl2 in LSC biology and obtained unexpected data with important implications for the eradication of LSCs in human cancer.

Materials and methods

Ethics statement

All animal work has been conducted according to relevant national and international guidelines.

Mice and BM transplantation

The Sca1-BCRABLp210 mice (11) and bcl2 gene-targeted mice (12) have both been described previously. Southern blot-based and polymerase chain reaction (PCR)-based genotyping for Sca1-BCRABLp210 (9) and bcl2 (12), respectively, were performed as described. In order to bypass the lethality of bcl2-deficient mice, we used the line 1B of Sca1-BCRABL-p210 mice that develop CML symptoms ∼5 months of age (11). Heterozygous bcl2^+/− mice were bred to Sca1-BCRABLp210 mice to generate compound heterozygotes. F1 animals were crossed to obtain null bcl2^−/− mice heterozygous for Sca1-BCRABLp210 mice.

In order to determine the nature of the leukemogenic cell, Sca1⁺Lin⁻ and Sca1⁻Lin⁺ cells were isolated and highly purified from the bone marrow (BM) of a leukemic primary mouse before or after different treatments or from control mice. The sorting purity of these cells was re-analyzed with the fluorescence-activated cell sorting and determined to be >98%. In each cohort, these cells were injected intravenously into sublethally irradiated (4 Gy) recipient mice. Diseased mice were killed and assessed for leukemia development. Mice that did not develop disease within 20 weeks posttransplantation were killed and tested for engraftment.

Histological analysis

All mice included in this study were subjected to standard necropsy. All major organs were examined under the dissecting microscope, and samples of each organ were processed into paraffin, sectioned and examined histologically. All tissue samples were taken by the pathologist from homogenous and viable portions of the resected sample and fixed within 2–5 min of excision. Hematoxylin- and eosin-stained sections of each tissue were reviewed by a single pathologist (T.F.). For comparative studies, age-matched mice were used.

TUNEL assay

Terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate nick end labeling was performed by using the ‘In situ cell death detection kit’ (Boehringer Mannheim, Spain), essentially following manufacturer's instructions with minor modifications depending upon the specimen preparation. In brief, sections were postfixed for 15 min in 4% paraformaldehyde, were washed twice with phosphate-buffered saline (PBS), and were incubated in ethanol:acetic acid (2:1) for 5 min at –20°C. After two washes in PBS, sections were subjected to a proteinase K digestion (10 mg/ml in 10 mM Tris–HCl, pH 8.0, and 1 mM ethylenediaminetetraacetic acid), were washed twice with PBS, and were counterstained with methyl green.

Analysis and monitoring of disease

Peripheral blood was collected from retro-orbital plexus with a heparinized capillary tube, and total white blood cell and differential counts were performed twice a week. The number of white blood cells was determined with a hemocytometer after lysis of enucleated red blood cells with red cell lysis buffer lysis buffer (0.15 M NH₄Cl, 1 mM KHCO₃, 0.1 mM Na₂-ethylenediaminetetraacetic acid; pH 7.4).

Flow cytometry

Nucleated cells were obtained from total bone marrow (flushing from the long bones), peripheral blood, thymus, liver and spleen. In order to prepare cells for flow cytometry, contaminating red blood cells were lyzed with red cell lysis buffer lysis buffer and the remaining cells were then washed in PBS with 2% fetal calf serum. After staining, all cells were washed once in PBS with 2% fetal calf serum containing 2 mg/ml propidium iodide to allow dead cells to be excluded from both analyses and sorting procedures. Monoclonal antibodies were obtained from Pharmingen(Spain) and included lineage markers (CD45R/B220 for B lineage staining; CD4, CD8 and CD3 for T cell lineage; CD11b and Gr1 for myeloid lineage and TER119 for erythroid lineage) and Sca1 (E13-161.7) for stem cells. Single cell suspensions from the different tissue samples obtained by routine techniques were incubated first with purified anti-mouse CD32/CD16 (Pharmingen) prior to the addition of other antibodies, to block binding via Fc receptors and then with an appropriate dilution of the different antibodies at room temperature or 4°C, respectively. The samples and the data were analyzed in a FACSCalibur using CellQuest software (Becton Dickinson, Spain). Specific fluorescence of fluorescein isothiocyanate and phycoerythrin excited at 488 nm (0.4 W) and 633 nm (30 mW), respectively, as well as known forward and orthogonal light scattering properties of mouse cells were used to establish gates. For each analysis, a total of at least 5000 viable (propidium iodide) cells were assessed.

Cell purification

For cell sorter separation, bone marrow cells were incubated with anti-Sca1 and anti-lineage markers antibodies (CD3, CD4, CD8, B220, TER119, Gr1 and Mac1). Sca1⁺Lin⁻ and Sca1⁻Lin⁺ cells were isolated and highly purified from the BM of leukemic primary mice or control mice by fluorescence-activated cell sorting (FACSVANTAGE; Becton Dickinson). c-kit (CD117) was not used for stem cell isolation as previous studies of human and mouse specimens have described downregulation of c-kit as a feature of leukemia stem cells (11). Sorted cells were then re-analyzed for purity with the fluorescence-activated cell sorting and determined to be >98% (Figure 1).

Fig. 1.

Purity of the sorted cell populations. The sorted cell populations, Sca1⁻Lin⁺ cells (A) and stem cells (B; Sca1⁺Lin⁻) used in this study were re-analyzed by flow cytometry. The expression of the characteristic cell surface markers is shown. Antibody staining and fluorescence-activated cell sorting of the different cell types were performed as described in Materials and Methods.

STI571 (Gleevec, Spain) treatment

STI571 treatment regimens were based on previous pharmacokinetic studies of STI in BCR–ABL tumor-bearing mice (13–15). Effectivity of BCRABLp210 suppression by STI571 was confirmed by assaying survival of Ba/F3 cells expressing BCRABLp210 24 h after STI571 treatment and by treatment of BCR–ABL transgenic mice as previously published (15) (data not shown). For the animal studies, stock solutions of 5 mg/ml and 10 mg/ml were freshly prepared in water, sterile filtered and administered to mice in a volume of 250 μl by gavage twice a day. Mice were started on STI571 or placebo (or the same volume of diluent water) beginning 1 day after leukemia was confirmed (day 0) by means of an STI571 regimen of 50 mg/kg every morning and 100 mg/kg every evening by gavage. STI571 was administered in a volume of 250 ml sterile water by means of straight or curved animal feeding needles. Mice tolerated the therapy well and no interruption of therapy was necessary. Mice were followed clinically three times a week, and periodic peripheral blood counts were obtained by tail vein blood draw as indicated. For the survival analysis portion of this study, the death end point was determined either by spontaneous death of the animal or by elective killing of the animal because of signs of pain or suffering according to established criteria.

Quantitative reverse transcription–PCR

RNA (1 μg) was reverse transcribed by using Advantage RT-for-PCR kit (BD Biosciences, Spain). SYBR Green PCR Master mix (Applied Biosystems, Spain) was used for template amplification with the primers specific for each of the transcripts examined (Table I). PCRs with reverse transcription sample were used as negative controls. Thermocycling for all targets were carried out in 30 μl reaction for 40 cycles in triplicate. Each cycle consisted of 94°C for 15 s, 56°C for 30 s and 72°C for 30 s. Incorporation of the SYBR Green dye into PCR products was monitored in real time with a QPCR Mastercycler epgradientS realplex system (Eppendorf, Spain). SDS system software was used to convert the fluorescent data into threshold cycle (C_t) at which exponential amplification of products begins. The differences in the C_t values (ΔC_t) between the transcript of interest and endogenous control (GAPDH) were used to determine the relative expression of the gene in each sample and the ΔΔC_t method was used to calculate fold expression. To determine correlation between expressions of two genes in the same set of samples, ΔC_t values were used to calculate regression coefficient.

Table I.

PCR oligos used in the quantitative PCR

Gene	Forward oligos	Backward oligos
Bcl-2	5′-GTCGTGACTTCGCAGAGATGT-3′	5′-TCAAAGAGAGGTCGCATGCTG-3′
Bcl-x	5′-CGTAGACAAGGAGATGCAGGT-3′	5′-ACCAGCGGTTGAAGCGCTCCT-3′
Mcl-1	5′-GAGGAAGAGGACGACCTATAC-3′	5′-ATGTCCAGTTTCCGGAGCATG-3′
A1	5′-TGAGCACTATCTTCAGTATGT-3′	5′-GTCATCCAAGTATGACTTCAG-3′
Bim	5′-CTACAGACAGAACCGCAAGGT-3′	5′-CCTGAGACTGTCGTATGGAAG-3′
Puma	5′-GCTGTATCCTGCAGCCTTTGC-3′	5′-ACTCTAAGTGCTGCTGGGCTG-3′
Bad	5′-GAGTGAGCAGGAAGACGCTAG-3′	5′-ACTGTGGCGACTCCGAGTCTC-3′
Bid	5′-GCTCCTTCAACCAAGGAAGAA-3′	5′-TGTCTGGCAATGTTGTGGATG-3′
Bax	5′-CAAGAAGCTGAGCGAGTGTCT-3′	5′-CTTGAGCACCAGTTTGCTAGC-3′
Apaf-1	5′-CTCTAGATGAAGCCATGTCGA-3′	5′-CAAGTCCCAGAGAACGCACAA-3′

Statistical analysis

The χ² test was used to compare leukemia incidence in Sca1-BCRABLp210 bcl2^−/− mice versus control (Sca1-BCRABLp210 bcl2^+/−) mice.

Results and discussion

In order to understand the role of Bcl2 in LSC generation and maintenance, we have taken advantage of our Sca1-BCRABLp210 mouse model of human CML, a paradigmatic stem cell disorder (11,16,17). This model represents an ideal system to study the contribution of LSCs to disease development and maintenance (11,17,18). Similarly, to human LSCs in CML, Bcl2 but not Bcl-X is highly expressed in the LSCs of Sca1-BCRABLp210 mice (11,17–19). Quantitative reverse transcription–PCR of Bcl2 messenger RNA confirmed that Bcl2 was highly expressed in Sca1⁺Lin⁻ cells of Sca1-BCRABLp210 mice (Figure 2). We did not detect variation in the expression levels of other members of the Bcl2 family in Sca1⁺Lin⁻ progenitors of Sca1-BCRABLp210 bcl2^−/− mice (Table II). Thus, we generated and analyzed cohorts of Sca1-BCRABLp210 bcl2^−/− (n = 19) experimental mice together with Sca1-BCRABLp210 bcl2^+/− (n = 20) controls. Surprisingly, when the compound Sca1-BCRABLp210 Bcl2-deficient mice were analyzed, CML onset and severity were not found to be overtly altered in comparison with control Sca1-BCRABLp210 bcl2^+/− mice (Figure 3A), as evidenced by the similar spleen sizes (Sca1-BCRABLp210 bcl2^+/− = 0.34 ± 0.2 g versus Sca1-BCRABLp210 bcl2^−/− = 0.31 ± 0.5 g) and similarly elevated leukocyte cellularity in the peripheral blood (Sca1-BCRABLp210 bcl2^+/− = 47 ± 14 × 10⁶/ml versus Sca1-BCRABLp210 bcl2^−/− = 42 ± 23 × 10⁶/ml). Loss of bcl2 only moderately decreased the survival of Sca1-BCRABLp210 mice, but this can be explained in the context of the known increased mortality of bcl2-knockout animals (12).

Fig. 2.

Quantification of Bcl2 expression in Sca1⁺Lin⁻ cells of Sca1-BCRABLp210 and control mice. Bcl2 messenger RNA expression with reference to GAPDH is shown.

Table II.

Quantification of anti- or pro-apoptotic members of the Bcl2 family in Sca1⁺Lin⁻ cells of Sca1-BCRABLp210 bcl2^+/− and Sca1-BCRABLp210 bcl2^−/− mice by quantitative PCR

Gene	Sca1+Lin− cells of Sca1-BCRABLp210 bcl2+/− micea	Sca1+Lin− cells of Sca1-BCRABLp210 bcl2−/− micea
Bcl-x	0.097 ± 0.035	0.107 ± 0.040
Mcl-1	0.124 ± 0.028	0.117 ± 0.034
A1	0.084 ± 0.025	0.087 ± 0.019
Bim	0.105 ± 0.021	0.111 ± 0.036
Puma	0.135 ± 0.039	0.127 ± 0.037
Bad	0.198 ± 0.045	0.197 ± 0.040
Bid	0.085 ± 0.023	0.077 ± 0.031
Bax	0.077 ± 0.030	0.086 ± 0.023
Apaf-1	0.080 ± 0.036	0.087 ± 0.029

^aPercentage of transcripts with reference to GAPDH is shown. The expression levels of anti- or pro-apoptotic members of the Bcl2 family were expressed at comparable levels suggesting that CML disease in Sca1-BCRABLp210 bcl2^−/− mice appears not to have selected for pro- or anti-apoptotic Bcl-2 family members.

Fig. 3.

Loss of bcl2 is not required for CML development in Sca1-BCRABLp210 mice. (A) Kaplan–Meier survival plots of Sca1-BCRABLp210 bcl2^−/− and Sca1-BCRABLp210 bcl2^+/− mice. The total number of mice analyzed in each group is indicated. Loss of bcl2 moderately decreased the survival of Sca1-BCRABLp210 mice (although statistically significant, P < 0.001). (B) Representative histologic appearance of liver and spleen of diseased Sca1-BCRABLp210 bcl2^+/− and Sca1-BCRABLp210 bcl-2^−/− mice after hematoxylin–eosin staining. Note the organ infiltration by myeloid cells in spleen and the perivascular infiltration of the liver by blasts and mature myeloid cells. (C) Matched sections from bcl2^−/− and Sca1-BCRABLp210 bcl2^−/− spleen were subjected to 4′,6-diamidino-2-phenylindole and terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate nick end labeling assay. The increased apoptosis described previously in bcl2-deficient animals was not observed in Sca1-BCRABLp210 bcl2^−/− mice. (D) Phenotypic characteristics of cells from peripheral blood of Sca1-BCRABLp210 bcl2^+/− and Sca1-BCRABLp210 bcl2^−/− mice as determined by flow cytometry. Myeloid blast cells are present in peripheral blood from leukemic Sca1-BCRABLp210 bcl2^+/− and Sca1-BCRABLp210 bcl2^−/− mice. Note that the presence of blast cells is defined by co-expression of Mac-1 and B220. The color code associates each relevant population in the forward scatter versus side scatter plot with the corresponding stained population, for each staining combination used. Percentages for each population are indicated in the corresponding color in the forward scatter versus side scatter plot.

Histologic analysis revealed similar pathology in moribund Sca1-BCRABLp210 bcl2^−/− and Sca1-BCRABLp210 bcl2^+/− mice. Apoptotic bodies were not seen in histological sections from Sca1-BCRABLp210 bcl2^−/− mice (Figure 3B). Consistent with this interpretation, we did not detect terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate nick end labeling-positive cells in spleen sections from Sca1-BCRABLp210 bcl2^−/− animals (Figure 3C). The myeloid infiltration was further confirmed in the peripheral blood of Sca1-BCRABLp210 bcl2^−/− mice by flow cytometric analysis (Figure 3D). These results indicate that the loss of bcl2 does not interfere with the development of CML in Sca1-BCRABLp210 mice.

Since CML in Sca1-BCRABLp210 mice has been proven (and designed) to be a stem cell disease (11), we next sought to examine if Bcl2 was influencing the nature of the leukemic stem cell in this model. Thus, we sorted Sca1⁺Lin⁻ progenitors and Sca1⁻Lin⁺ cells from either Sca1-BCRABLp210 bcl2^−/− or Sca1-BCRABLp210 bcl2^+/− control mice. Transplantation of purified fractions of cells into sublethally irradiated syngeneic recipient mice was used to assess leukemogenesis in vivo. When irradiated wild-type recipients were reconstituted with Sca1⁺Lin⁻ cells from either Sca1-BCRABLp210 bcl2^−/− or Sca1-BCRABLp210 bcl2^+/− donors, all the mice developed CML that was phenotypically identical to the primary disease (Table III). In contrast, the Sca1⁻Lin⁺ cells were incapable of inducing CML into secondary recipients, even when injected in a 10- or 100-fold higher number than the Sca1⁺Lin⁻ cells (Table III). Overall, these data indicate that the LSCs reside in the Sca1⁺ cell compartment even in the absence of Bcl2 and that Bcl2 is not required for the generation and function of LSCs in Sca1-BCRABL-p210 mice.

Table III.

CML disease of Sca1-BCRABLp210 bcl2^−/− mice is readily transplantable to secondary recipients by LSCs

Sorted cells	Number of transplanted cells	Transplanted animals	Incidence of CML (%)	Latency of disease (days ± SD)
Sca1+Lin− (BM, Sca1-BCRABLp210/Bcl2−/− mice)	10 000	6	100	59 ± 16
	1000	6	100	92 ± 13
Sca1+Lin− (BM, Sca1-BCRABLp210/Bcl2+/− mice)	10 000	6	100	64 ± 17
	1000	6	100	102 ± 21
Sca1−Lin+ (BM, Sca1-BCRABLp210/Bcl2−/− mice)	1 × 105	6	0	NA
	1 × 106	6	0	NA
Sca1−Lin+ (BM, Sca1-BCRABLp210/Bcl2+/− mice)	1 × 105	6	0	NA
	1 × 106	6	0	NA

SD, standard deviation.

In human CML, the LSC pool can be detected indefinitely in Gleevec (STI571)-treated patients, leading to disease relapse upon drug discontinuation (20–25). The high levels of Bcl2 expression in these LSC, which would make them particularly resistant to induction of apoptosis by a variety of agents, have led to the proposal that Gleevec might be able to eradicate this reservoir if combined with Bcl2 antagonists (26,27), providing a potential explanation for the resistance of the LSC cells in Gleevec-treated CML patients. Consequently, we next examined the effect of STI571 treatment in Sca1-BCRABLp210 bcl2^−/− mice. STI571 treatment began 1 day after leukemia was confirmed by peripheral blood analysis. Mice were monitored clinically and by serial peripheral blood count for evidence of leukemia. STI571 did not prolong the survival of Sca1-BCRABLp210 bcl2^−/− mice (Figure 4) and Sca1-BCRABLp210 bcl2^−/− mice treated with STI571 did not demonstrate a marked reduction in white blood cell (Sca1-BCRABLp210 bcl2^+/− = 45 ± 21 × 10⁶/ml versus Sca1-BCRABLp210 bcl2^−/− = 46 ± 27 × 10⁶/ml). This result clearly shows the insensitivity of the bcl2-deficient LSCs to STI571. To further confirm this point, we next examined whether LSCs from STI571-treated Sca1-BCRABLp210 bcl2^−/− mice can propagate CML disease to secondary recipients. Equal numbers of Sca1⁺Lin⁻ cells (5 × 10³) from the BM of STI571-treated animals failing therapy were used to reconstitute sublethally irradiated recipients. All reconstituted animals (n = 6) developed CML disease with a similar average latency (20 days), confirming that persistent Bcl2 expression is not required to maintain the malignant phenotype of CML LSCs.

Fig. 4.

STI571 treatment in diseased Sca1-BCRABLp210 bcl2^−/− mice. STI571 treatment does not modify the survival of Sca1-BCRABLp210 bcl2^−/− mice. Mice were randomized to treatment with either STI571 or placebo in order to study the in vivo efficacy in three independent experiments. The survival curve depicts the percentage of animals alive at the indicated time point. The number of mice in each arm (n) is also shown.

Overall, our work demonstrates for the first time that Bcl2 is not required for the generation and maintenance of LSCs. This evidence that Bcl2 inactivation could not modify LSC is in apparent conflict with previous studies showing that Bcl2 is required for the persistence of tumoral characteristics of differentiated tumor cell (2,9). However, we should consider that the effects of Bcl2 inactivation will depend on the mechanisms by which Bcl2 is actually contributing to the tumorigenic phenotype, which are probably to vary according to the genetic and cellular context. Previous findings have shown that inhibition of apoptosis is required later in tumor progression (2,9). However, our results indicate that the inhibition of Bcl2 is not critical for the generation, selection or maintenance of the tumor initiating and maintaining cells.

Funding

Fondo Europeo de Desarrollo Regional to I.S.G. group; Ministerio de Ciencia e innovación (SAF2009-08803) I.S.G. group; Junta de Castilla y León (CSI13A08, proyecto Biomedicina 2009–2010); Ministerio de Educación y Cicencia-MEC OncoBIO Consolider-Ingenio 2010 (Ref. CSD2007-0017); National Institutes of Health (2R01 CA109335-04A1); Sandra Ibarra Foundation; Group of Excellence Grant (GR15), Junta de Castilla y Leon. Fondo Europeo de Desarrollo Regional to C.C.; Fondo de Investigaciones Sanitarias (PI080164), Junta de Castilla y León (SA060A09, proyecto Biomedicina 2009–2010); Fundación de Investigación Médica MM.

Glossary

Abbreviations

BM

bone marrow

CML

chronic myeloid leukemia

LSC

leukemia stem cell

PBS

phosphate-buffered saline

PCR

polymerase chain reaction