The antiapoptotic protein Bcl-xL is induced in activated T lymphocytes upon costimulation through CD28, 4-1BB, and OX40. Bcl-xL is also highly enriched in memory T lymphocytes. Based on this body of evidence, it was thought that Bcl-xL plays an essential role in the generation of effector and memory T lymphocytes. We report that mice with a conditional deletion of Bcl-x in T lymphocytes develop a normal CD8+ T cell response to Listeria monocytogenes infection. Furthermore, Bcl-x conditional knockout mice exhibit normal T-dependent humoral immune responses. These results indicate that Bcl-x is dispensable for the generation of effector and memory T lymphocytes and suggest that costimulation of T lymphocytes promotes their survival through a Bcl-xL independent mechanism.
Antiapoptotic Bcl-x is an important member of the Bcl-2 gene family. Five isoforms of Bcl-x (Bcl-xL, Bcl-xs, Bcl-xγ, Bcl-xβ, and Bcl-xΔTM) derived from alternative splicing have been identified (1, 2, 3, 4, 5). A role for Bcl-x in regulating cell survival has been demonstrated in Bcl-x-deficient mice. Bcl-x-deficient mice die around embryonic day 13 due to massive apoptosis of immature hemopoietic cells and neurons (6). Furthermore, mice lacking Bcl-x in erythroid cells exhibit hemolytic anemia and profound splenomegaly (7). These genetic studies indicate that Bcl-x promotes cell survival in different host tissues.
Bcl-x also plays an important role in thymocyte development. Bcl-xL is the predominant isoform found in developing T lymphocytes (8, 9). Its expression is tightly controlled. Bcl-xL is highly expressed in CD4+CD8+ double positive (DP)3 thymocytes but is down-regulated in single positive (SP) thymocytes (8, 9). DP thymocytes in Bcl-x−/−/Rag-2−/− chimeric mice (6, 8) or retinoid acid-related orphan receptor RORγ−/− mice that have a dramatically reduced Bcl-xL expression undergo massive apoptosis and develop a defective TCR-α repertoire (10, 11). In addition, recent data demonstrate that dendritic cell survival depends on Bcl-xL (12, 13).
Although there is no direct evidence, many studies have implicated Bcl-xL in the differentiation of effector and memory T lymphocytes. First, Bcl-xL is up-regulated in T lymphocytes activated through TCR and costimulatory molecule CD28 (14). CD28 may promote T lymphocyte survival through Bcl-xL induction because expression of Bcl-xL in CD28-deficient T lymphocytes restores the cell survival defect (15). Furthermore, it has been shown that the Src homology 2 binding domain on CD28 is required for Bcl-xL induction and cell survival (16, 17, 18). Second, signaling through several other costimulatory molecules including OX40 and 4-1BB results in Bcl-xL expression and enhanced T lymphocyte survival (19, 20, 21, 22). Third, Bcl-xL is up-regulated in memory CD4+ T lymphocytes (23), suggesting that up-regulation of Bcl-xL expression may play a critical role in memory T lymphocyte formation.
Bcl-xγ, a cytosolically expressed isoform of Bcl-x, has been reported to play a critical role in CD28-dependent costimulation (5, 24). Unlike the wide tissue distribution of Bcl-xL, the expression of Bcl-xγ is restricted to the T cell compartment and is up-regulated upon TCR and CD28 ligation (5). T lymphocytes lacking Bcl-xγ show defective cell proliferation and cytokine production to CD28-dependent costimulatory signals but normal cell survival after activation (24). These results suggest that Bcl-xγ may perform some unique function in CD28-dependent costimulation of T lymphocytes that cannot be compensated by Bcl-xL.
Given the extensive evidence implicating Bcl-xL in CD28-mediated costimulation, as well as effector and memory T lymphocyte differentiation, it is important to directly determine the role of Bcl-xL in these processes. To address this issue, we have generated a conditional Bcl-x knockout mouse strain with all the isoforms of Bcl-x deleted in T lymphocytes. We examined the effector and memory T lymphocyte development in these mutant mice. To our surprise, we observed normal CD4+ and CD8+ effector and memory T cell development in the absence of Bcl-x expression. These results suggest that costimulatory signal-induced T cell survival can be mediated through a Bcl-x independent mechanism.
Materials and Methods
Generation of Bcl-x conditional knockout mice
To generate Bcl-x conditional knockout mice, genomic fragments from a Bcl-x BAC clone (Roswell Park Cancer Institute) were cloned into the pGKneoF2L2DTA targeting vector so that exon 1 and exon 2 of Bcl-x gene are flanked by two loxP sites. The targeting construct was linearized by NotI and transfected into TC1 embryonic stem cells (25). Embryonic stem cell clones with homologous recombination were identified by PCR and Southern blot. Three embryonic stem cell clones with the correct targeting events were injected into C57BL/6 blastocysts. Chimeric male founder mice (Bcl-xfl/+) were crossed with FLPeR female mice (26) (kindly provided by Dr. P. Soriano, Fred Hutchinson Cancer Research Center, Seattle, WA) to delete the neomycin cassette flanked by two FRT sites. Bcl-xfl/+ mice were then bred with Lck-cre transgenic mice (27) (The Jackson Laboratory) to generate Bcl-xfl/fllck-cre, Bcl-xfl/+lck-cre, and Bcl-xfl/fl mice. The phenotypes of Bcl-xfl/+lck-cre and Bcl-xfl/fl mice are indistinguishable from those of wild-type (Bcl-x+/+) C57BL/6 × 129 mice. These mice were used as controls throughout the experiments. Bcl-2+/− mice (28) were purchased from The Jackson Laboratory. Animal usage was conducted according to protocols approved by the Duke University Institutional Animal Care and Use Committee.
Flow cytometric analysis
Single cell suspension of the thymus, spleen, and lymph nodes were lysed of RBC, incubated with an Fc receptor blocker (2.4G2; eBioscience), followed by biotinylated mAb, PE-streptavidin, FITC-, CyChrome-, or allophycocyanin-labeled mAb on ice for 30 min, and washed with PBS containing 2% FCS. Data for 1–5 × 104 cells were collected on a FACScan flow cytometer (BD Biosciences) and analyzed using CellQuest software. For intracellular IFN-γ or Bcl-2 staining, peptide or anti-CD3/CD28-stimulated T cells were first stained for surface CD8 and CD4 and then fixed with 2% paraformaldehyde in PBS for 30 min on ice. The cells were permeabilized with 0.5% saponin (Sigma-Aldrich) and stained with anti-IFN-γ FITC or anti-Bcl-2 FITC. All Abs, including anti-CD4, CD8, CD25, CD69, CD44, CD62 ligand, were obtained from eBioscience, Biolegend, or BD Pharmingen. Apoptotic cells were determined by annexin V and 7-aminoactinomycin D (7-AAD) staining using an annexin V-PE kit (BD Pharmingen).
Southern, Northern, and Western blot analysis
Southern, Northern, and Western blot analyses were performed as previously described (11). Anti-Bcl-x Ab (clone 2H12) was from BD Pharmingen.
Immunization, Ab titration, and histochemical analysis
Six- to 10-wk-old Bcl-xfl/fllck-cre and Bcl-xfl/+lck-cre (or wild-type) mice were immunized i.p. with 100 μg of DNP-keyhole limpet hemocyanin (KLH; Calbiochem) with or without CFA. Sera were collected at different time points. At day 28, the same dose of DNP-KLH with or without IFA was used in secondary challenge. Isotype-specific anti-DNP Abs were determined by ELISA as previously described (29). The titer was expressed as relative units as compared with a “master” sample. For SRBC immunization, mice were injected i.p. with 1 ml of PBS or PBS containing 1 × 108 SRBC. Seven to 10 days later, sera were collected. Freshly removed spleens were immersed in Tissue-Tek OCT compound (Sakura Finetek) and snap-frozen in dry ice. Tissue sections at 5-μm thickness were stained with biotin anti-B220 (eBioscience) and FITC anti-GL7 (BD Pharmingen) and developed with HRP-Streptavidin (Southern Biotechnology Associates) and anti-FITC alkaline phosphatase (Sigma-Aldrich).
Cell proliferation assays
Peripheral CD4+ or CD8+ T cells purified through positive selection by autoMACS were labeled with CFSE (Molecular Probes) and stimulated with plate-bound anti-CD3 (2C11) (eBioscience) and/or anti-CD28 (clone 37.51) (BioLegend). After 3 days, cell proliferation was determined by FACS analysis of CD4+ or CD8+ T cells. CD4+ recall response was examined by incubating splenocytes from DNP-KLH immunized mice (8 wk after immunization) in the presence of different concentrations of DNP-KLH. T cell-specific proliferation to DNP-KLH was measured by pulsing the cells with [H3]thymidine in the last 8 h of a 48 h culture.
Bacterial infection and CTL assay
The recombinant Listeria monocytogenes strain engineered to secrete chicken OVA (rLmOVA) (30) and pMHC/peptide tetramers were kindly provided by M. Bevan (University of Washington, Seattle, WA). rLmOVA was grown in brain-heart infusion broth supplemented with 5 μg/ml erythromycin. Bacteria were diluted in PBS and injected i.v. at the dose of 2 × 103 CFU for primary infection and 2 × 105 CFU for secondary infection. To examine the CTL activity of infected mice, splenocytes normalized for an equal number of CD8+ T cells were incubated with 51Cr-labeled EL-4 target cells pulsed with 10−7 M SIINFEKL peptide for 6 h at 37°C. The supernatants of the killing assays were collected and counted to determine the amount of 51Cr release. The percentage of specific lysis was calculated as 100 × ((experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm)).
Generation of Bcl-x conditional knockout mice
Bcl-x has five isoforms that are derived from alternate splicing. To generate mice specifically lacking Bcl-x in T lymphocytes, we constructed a targeting vector with exon 1 and exon 2 of Bcl-x flanked by two loxP sites (Fig. 1,A). The flanked exon 2 containing the initiation codon is used by all five isoforms. A neomycin-resistant gene cassette flanked by two FRT sites is located within the two loxP sites (Fig. 1,A). We generated chimeric founder mice by microinjecting three correctly targeted embryonic stem clones into C57BL/6 blastocysts. Male chimeric mice were bred with FLPeR female mice expressing the FLPe recombinase (26) once to delete the neomycin cassette in vivo. The Bcl-x flanked mice (Bcl-xfl/fl) were then bred with Lck-cre transgenic mice to induce specific deletion of this gene in T lymphocytes. Restriction analysis of genomic DNA from total thymocytes of Bcl-xfl/fllck-cre mice revealed that the floxed Bcl-x exons were deleted in >98% of the thymocytes (Fig. 1,B). Northern blot analysis of total thymocytes from wild-type (Bcl-x+/+), Bcl-xfl/fl, and Bcl-xfl/fllck-cre (herein referred as Bcl-x−/−, and Bcl-xfl/+lck-cre as Bcl-x+/−) mice showed that Bcl-xL mRNA was not detected even when the total RNA of Bcl-x−/− thymocytes was overloaded (Fig. 1,C). Furthermore, Western blot analysis demonstrated a complete lack of Bcl-xL protein expression in Bcl-x−/− total thymocytes (Fig. 1 D). These results demonstrate that Bcl-x is efficiently deleted in the thymus as no mRNA or protein of Bcl-x was detected.
T cell development in Bcl-x conditional knockout mice
Mice with Bcl-x specifically deleted in T lymphocytes had normal growth and development. The total thymic cellularity in Bcl-x−/− mice was reduced by 40–50% when compared with age- and sex-matched controls (Fig. 1,E). The number of mature CD4+ and CD8+ T lymphocytes in the spleen of Bcl-x−/− mice was also reduced by ∼40–50% (Fig. 1,E). We examined thymocyte subsets in Bcl-x−/− mice by FACS analysis. The percentages of double negative (DN), DP, and SP cells in the mutant mice were comparable to those of the control mice (Fig. 1,F), in agreement with the results from Bcl-x−/−/Rag-2−/− chimeric mice (6, 8). To test the role of Bcl-xL in promoting thymocyte survival, we cultured total thymocytes in vitro for different times and determined the phenotypes of live cells by excluding 7-AAD+/annexin V+ cells. After 1–2 days of culture, a dramatic reduction in the DP compartment of Bcl-x−/− thymocytes was observed (Fig. 1,F). In contrast, CD4+ and CD8+ SP and DN thymocytes from Bcl-2−/− mice disappeared rapidly in the culture (Fig. 1 F). These results correlate with the expression of Bcl-xL in DP thymocytes and Bcl-2 in DN and SP thymocytes and indicate that Bcl-x plays a critical role in promoting DP thymocyte survival.
Activation and survival of peripheral Bcl-x−/− T cells
As shown in Fig. 1,E, the number of peripheral CD4+ and CD8+ T lymphocytes was reduced by 40–50% in 4- to 8-wk-old Bcl-x−/− mice and was reduced by 30–40% in 10 wk or older mutant mice (data not shown). The reduced peripheral T cell compartment likely reflects a reduced production of thymocytes because Bcl-x is not detectable in resting mature T cells (9, 31). We examined the phenotype of peripheral T cells in Bcl-x−/− mice. Bcl-x−/− CD4+ and CD8+ T cells expressed similar levels of CD25 and CD69 to those on control cells (Fig. 2,A). However, a higher fraction of peripheral Bcl-x−/− T cells expressed CD44high and CD62Llow than that of the controls (Fig. 2 A), suggesting that these cells were undergoing lymphopenic-driven homeostatic proliferation.
Given that Bcl-xL is up-regulated upon TCR stimulation and may be involved in CD28-mediated costimulation and that Bcl-xγ appears to be essential in CD28-mediated costimulation of T cell proliferation and cytokine production, we examined TCR-mediated proliferation of Bcl-x−/− T cells. CFSE-labeled CD4+ or CD8+ T cells were stimulated with plate-bound anti-CD3 and/or anti-CD28. Surprisingly, the proliferation of Bcl-x−/− CD4+ T cells (Fig. 2,B) and CD8+ T cells (data not shown) upon anti-CD3 or anti-CD3 plus anti-CD28 stimulation was comparable to that of control cells. Furthermore, IL-2 production by Bcl-x−/− and control T lymphocytes was similar (Fig. 2 C). These results indicate that TCR/CD28-mediated T cell proliferation and cytokine production were not impaired in the absence of Bcl-x.
The expression of Bcl-xL and Bcl-xγ in activated T lymphocytes suggests that these proteins might promote T cell survival after their activation. Therefore, we examined apoptosis in Bcl-x−/− T cells after anti-CD3/CD28 stimulation. Similar fractions of T cells underwent apoptosis in Bcl-x−/− and control T cells (Fig. 3,A). We further examined whether Bcl-x−/− T cells display a higher sensitivity to activation-induced T cell death. We did not find significant difference in the activation-induced T cell death of Bcl-x−/− and control T cells (Fig. 3 B). Furthermore, activated Bcl-x−/− T cells did not exhibit an increased sensitivity to IL-2 deprivation-induced apoptosis. These results demonstrate that apoptosis induced through TCR stimulation and cytokine deprivation in Bcl-x−/− T cells is not defective and argue against an essential role of Bcl-x, including Bcl-xL and Bcl-xγ, in promoting activated T cell survival.
One possibility for the normal T cell survival in the absence of Bcl-x is that another antiapoptotic protein, Bcl-2, may be up-regulated in resting or stimulated Bcl-x−/− T cells and compensate for the loss of Bcl-x. To test this, we examined Bcl-2 expression in Bcl-x−/− T cells by intracellular staining. As shown in Fig. 3,C, Bcl-2 expression in Bcl-x−/− CD4+ T cells was similar to that in control cells before TCR stimulation and slightly lower than that in control cells after anti-CD3/CD28 stimulation. In contrast, Bcl-2 expression in Bcl-x−/− CD8+ T cells was slightly higher than that in control cells before or after TCR stimulation (Fig. 3 C).
Effector and memory CD8+ T cell response in Bcl-x−/− mice
To examine the role of Bcl-x in the development of CD8+ T effector and memory cells, we used a Listeria monocytogenes infection model (32, 33). We infected Bcl-x−/− and control mice i.v. with 2 × 103 L. monocytogenes expressing chicken OVA. Seven days later, the primary CD8+ T cell response was examined by intracellular cytokine staining of IFN-γ. Bcl-x−/− CD8+ T cells mounted a similar primary response to L. monocytogenes infection to that of control CD8+ T cells as determined by the percentage of IFN-γ+ cells (Fig. 4,A). Furthermore, the cytotoxicity of CD8+ T cells in L. monocytogenes-infected Bcl-x−/− mice was identical with that of control CD8+ T cells (Fig. 4,B). To test the memory CD8+ T cell response in Bcl-x−/− mice, we reinfected the mice with 2 × 105 L. monocytogenes 8 wk after the primary infection. Three days after the second infection, splenic CD8+ T cells were examined for IFN-γ production and CTL activity against OVA. Surprisingly, Bcl-x−/− CD8+ T cells exhibited a similar response to that of control cells upon OVA stimulation (Fig. 4,A). In addition, the CTL activity of Bcl-x−/− CD8+ T cells in the memory response was comparable to that of the controls (Fig. 4,B). The frequency of OVA-specific CD8+ T cells in wild-type and Bcl-x−/− mice was also assessed by pMHC/peptide tetramer staining and the results were similar to IFN-γ staining (data not shown). Although the percentage of OVA-specific T cells in total CD8+ T cells in Bcl-x−/− mice is similar to that in wild-type mice, the absolute numbers of OVA-specific CD8+ T cells in the spleen of Bcl-x−/− mice were slightly lower (but statistically not significant) than those in the control mice (Fig. 4,C). The lowered numbers of OVA-specific CD8+ T cells in the mutant mice are likely due to a lowered thymic production of naive CD8+ T cells (Fig. 1).
To rule out the possibility that some T cells escaped Lck-cre-mediated Bcl-x deletion in the thymus to mount this apparently normal immune response, CD3+ T cells from Bcl-x−/− mice immunized with L. monocytogenes for 3 times were purified by double FACS sorting and analyzed by Southern blot. As shown in Fig. 4 C, we did not detect any signal for the floxed allele of Bcl-x in the purified CD3+ Bcl-x−/− T cells, indicating that the normal T cell effector and memory response in Bcl-x−/− mice was not caused by an incomplete deletion of the Bcl-x allele in the peripheral T cells.
CD4+ Th function in Bcl-x−/− mice
To test the role of Bcl-x in Th function, we immunized Bcl-x−/− and control mice with a T cell-dependent Ag, DNP-KLH and boosted 4 wk later. The titers of anti-DNP specific Abs were determined by ELISA. The production of IgM and IgG was comparable in the primary and secondary responses in both Bcl-x−/− and control mice (Fig. 5, A and B). In addition, Bcl-x−/− mice produced similar levels of IgG1, 2a, 2b, and 3 against DNP, indicating normal Ig class switches. To measure CD4+ T cell memory response, we examined DNP-KLH-induced CD4+ cell proliferation from immunized mice. CD4+ T cell recall response was identical in Bcl-x−/− and control groups (Fig. 5 C). In addition, we did not find any differences in the production of anti-DNP Abs in Bcl-x−/− and control mice received DNP-KLH immunization without adjuvant (data not shown).
To further test CD4+ Th function in the absence of Bcl-x, we immunized Bcl-x−/− mice with another T cell-dependent Ag, SRBC. Seven to 10 days later, Ab production and germinal center formation were examined in these mice. We did not find any difference in SRBC-induced germinal center formation (Fig. 5 D) or anti-SRBC Ab production (data not shown) in Bcl-x−/− and control mice. Taken together, these results demonstrate the CD4+ Th function is not impaired in the absence of Bcl-x.
Our results have addressed a long-standing issue: the role of Bcl-xL in the development of effector and memory T lymphocytes. Because Bcl-xγ, the T cell restricted isoform of Bcl-x, might have a redundant function with Bcl-xL, we chose to delete the Bcl-x exons used by all the isoforms in T lymphocytes. To our surprise, T cells lacking Bcl-x exhibit a normal CD4+ and CD8+ effector and memory differentiation in our model systems.
It is well established that CD28 regulates multiple aspects of T lymphocyte function including activation, proliferation, cytokine production, survival, effector and memory formation (34, 35). Given the fact that CD28 engagement promotes T lymphocyte survival and up-regulates Bcl-xL expression (14, 36, 37), it was assumed that Bcl-xL plays a key role in CD28-mediated function. Indeed, this notion was further supported by a study showing that retrovirus-based delivery of Bcl-xL into CD28−/− T lymphocytes restores the survival defect in these cells (15). However, our results argue against an essential role of Bcl-xL in CD28-mediated function. In contrast to the defective survival, proliferation, and cytokine production in CD28−/− T lymphocytes, we did not find any similar defect in Bcl-x−/− T lymphocytes. Furthermore, it has been shown that CD28 is essential for effector and memory T cell differentiation in studies examining CD8+ response to L. monocytogenes infection and CD4+ helper function in T cell-dependent humoral immune response (38, 39, 40). We have used similar model systems to evaluate effector and memory T cell differentiation in Bcl-x−/− mice. The normal CD4+ and CD8+ effector and memory responses in Bcl-x−/− mice, again in sharp contrast to the defective responses in CD28−/− mice, suggest that CD28-regulated T lymphocyte function is independent of Bcl-x. Alternatively, the defective immune response in CD28−/− mice may be caused by other cellular components rather than T lymphocytes. Our data clearly suggest that the survival defect in CD28−/− T lymphocytes is not solely caused by a lack of Bcl-xL expression, even though enforced expression of Bcl-xL alone corrected the defect.
The apparently normal development of effector and memory T cells in the absence of Bcl-x may be due to a compensatory role of Bcl-2. Bcl-2 and Bcl-x function similarly in promoting cell survival (41). Bcl-2 expression is up-regulated in CD8+ memory T cells whereas Bcl-xL is enriched in day 8 effector T cells in a LCMV infection model (42, 43). We observed a differential expression of Bcl-2 in Bcl-x−/− CD4+ and CD8+ T cells. Bcl-2 is expressed slightly lower in Bcl-x−/− CD4+ T cells although slightly higher in Bcl-x−/− CD8+ T cells than that in control T cells. These results suggest that Bcl-2 may be up-regulated in CD8+ but not CD4+ T cells for the loss of Bcl-x. However, a functional role of Bcl-2 in the development of effector and memory T cells needs to be definitively addressed in T cells lacking both Bcl-x and Bcl-2. In addition, other antiapoptotic pathways such as those mediated by the serine/threonine kinase protein kinase B/Akt and NF-κB as well as Mcl-1 may be able to compensate for Bcl-x deficiency in supporting T cell survival (44, 45)
The phenotypes displayed by Bcl-x−/− T lymphocytes are dramatically different from those of Bcl-xγ−/− T lymphocytes. T lymphocytes from Bcl-xγ−/− mice exhibit defective proliferative and cytokine responses to CD28-dependent costimulation (24). The reason for this discrepancy is unclear. One possibility is that other isoforms of Bcl-x, such as Bcl-xs, in the Bcl-xγ−/− mice become functionally dominant and promote cell death. This is unlikely because Western blot analyses did not detect other isoforms of Bcl-x in thymocytes and activated T cells at the protein level (31). The discrepancy between Bcl-x−/− and Bcl-xγ−/− mice needs to be resolved in future investigation.
We thank Dr. David Draper, Heather Hartig, and Ivan Dzhagalov for critical review of this manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by American Cancer Society Grant RSG-0125201 and National Institutes of Health Grant CA92123.
Abbreviations used in this paper: DP, double positive; SP, single positive; 7-AAD, 7-aminoactinomycin D; KLH, keyhole limpet hemocyanin; DN, double negative.