Nature Immunology
3, 932 - 939 (2002)
Published online: 3 September 2002; | doi:10.1038/ni834
Deficiency in Bak and Bax perturbs thymic selection and lymphoid homeostasisJeffrey C. Rathmell, Tullia Lindsten, Wei-Xing Zong, Ryan M. Cinalli
& Craig B. ThompsonAbramson Family Cancer Research Institute, Departments of Cancer Biology, Medicine and Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA.
Correspondence should be addressed to Craig B. Thompson drt@mail.med.upenn.eduBak and Bax are required and redundant regulators of an intrinsic mitochondrial cell death pathway. To analyze this pathway in T cell development and homeostasis, we reconstituted mice with Bak-/-Bax-/- hematopoietic cells. We found that the development and selection of Bak-/-Bax-/- thymocytes was disrupted, with altered representation of thymic subsets and resistance to both death-by-neglect and antigen receptor−induced apoptosis. Elimination of Bak-/-Bax-/- T cells that responded to endogenous superantigen was also reduced. Despite more efficient early reconstitution and apoptotic resistance of Bak-/-Bax-/- thymocytes, thymic cellularity declined over time. Reduced thymic cellularity resulted from a progressive cessation of thymopoiesis. However, animals developed splenomegaly as a result of accumulated memory T cells that were not deleted after antigen-driven expansion. These data indicate that Bak and Bax are required for thymic selection and peripheral lymphoid homeostasis and suggest that thymopoiesis can be negatively regulated by the accumulation of cells that would normally be eliminated by pro-apoptotic Bcl-2−related genes.Regulation of lymphocyte cell number is critical for proper lymphoid development and homeostasis. Failure to correctly control cell number can lead to insufficient numbers of lymphocytes for protection from foreign pathogens, whereas the accumulation of excess lymphocytes may increase the risk of developing autoimmunity. The number of peripheral lymphocytes in a mature animal remains relatively constant despite periodic expansions in response to infections and immunizations1. A variety of factors influence lymphocyte number, including the rate of production of lymphocyte precursors from hematopoietic stem cells, thymic and bone marrow output of mature lymphocytes as well as peripheral lymphoid expansion and survival. At each point, cell proliferation, differentiation and death are balanced to maintain high lymphocyte production during development and to maintain lymphocyte numbers upon maturity. One mechanism of cell death used to regulate lymphocyte number is death-by-neglect2; cells are deprived of the necessary extrinsic signals that maintain metabolism sufficiently to preserve cell size and survival.
Members of the Bcl-2 family regulate lymphocyte cell death-by-neglect and can have profound effects on lymphocyte cell numbers. Active Bcl-2 and its relatives localize to intracellular membranes and regulate mitochondrial homeostasis. During neglect, expression of the anti-apoptotic Bcl-2 or Bcl-xL proteins can maintain mitochondrial integrity despite cellular atrophy and a reduced supply of electron-transport substrates3. When Bcl-2 or Bcl-xL are overexpressed as transgenes in lymphocytes, small cells accumulate in a gene dose−dependent manner4,
5,
6. Removal of Bcl-2 or Bcl-xL by gene targeting causes a loss of mature or developing lymphocytes, respectively7,
8. In contrast to the increased survival and cell numbers observed when anti-apoptotic Bcl-2 family members are expressed, transgenic expression of the pro-apoptotic Bax in T cells results in decreased numbers of T cells and a high rate of T cell turnover9. Bax and its pro-apoptotic paralog Bak are expressed in inactive forms that can become activated during neglect and initiate cell death by disrupting mitochondrial integrity. This results in the release of cytochrome c and other pro-apoptotic proteins. Expression and activation of the BH3-only Bcl-2 family member Bim can also cause mitochondrial dysfunction and cell death10, and Bim deficiency results in splenomegaly, lymphadenopathy11 and defects in thymic negative selection12.
Despite their ubiquitous expression and potent ability to promote cell death when overexpressed, the pro-apoptotic Bcl-2 family members Bak and Bax show only minor phenotypic changes when removed individually by gene targeting, with a mild lymphoid hyperplasia noted in some Bax-deficient mice13,
14. However, the combined deficiency of Bak and Bax has profound effects on mitochondria-dependent cell deaths, such as death-by-neglect14. Cells from Bak-/-Bax-/- mice are resistant to a variety of death stimuli, including neglect, irradiation, DNA-damaging drugs and apoptosis induced by expression of pro-apoptotic BH3 proteins, such as Bad, Bid, Noxa and Bim15,
16,
17. Because each of these death stimuli acts by causing the loss of mitochondrial homeostasis, these findings indicate that Bak and Bax are required and redundant in initiating mitochondria-induced apoptosis. As a consequence of the inability of Bak-/-Bax-/- cells to undergo mitochondrially induced cell death, cell number regulation is disrupted in Bak-/-Bax-/- mice in multiple tissues14. In the hematopoietic system, Bak-/-Bax-/- mice have increased precursor cell frequencies and accumulate large numbers of peripheral lymphocytes, resulting in splenomegaly and lymphadenopathy. Accumulated lymphocytes display memory-phenotype surface markers, with high CD44 expression and evidence of class switching in B cells.
It is unclear whether the accumulation of lymphocytes in Bak-/-Bax-/- mice is due to failure of apoptosis to eliminate excess stem cells and lymphoid precursor cells, resulting in increased lymphocyte production, or is due to a failure to eliminate peripheral lymphocytes by neglect. To determine how apoptosis pathways affect lymphocyte development and cell number regulation, we analyzed T cell development and homeostasis in the absence of Bak and Bax. Our data indicated that Bak and Bax play critical cell-intrinsic roles in regulating T cell survival. In the absence of the T cell turnover that results from the pro-apoptotic function of Bak and Bax, thymic selection was perturbed, memory T cells accumulated and thymopoiesis was suppressed. Our data suggested that regulation of peripheral, rather than developmental, T cell survival exerts the primary control over T cell numbers in secondary lymphoid organs and that thymopoiesis can be suppressed by the accumulation of cells that were not eliminated by Bak or Bax.
Results
Thymocyte development without Bak and Bax
To determine the role played by Bak and Bax in lymphocyte development, Bak-/-Bax-/-and littermate control mice were analyzed. Despite the splenomegaly observed in Bak-/-Bax-/- animals14, thymic cellularity in Bak-/-Bax-/- mice was similar to that observed in control mice (Fig. 1a). The representation of thymic subsets, however, was altered. In the absence of Bak and Bax, CD4-CD8- double negative (DN) and CD4+CD8- and CD4-CD8+ single positive (SP) thymocytes were increased in number, whereas CD4+CD8+ double positive (DP) thymocytes were decreased (Fig. 1b). The frequency of DP thymocytes in Bak-/-Bax-/- mice was 1−30%, whereas control mice showed an average of 80% DP thymocytes. Because Bak-/-Bax-/- mice have multiple developmental defects14, it was possible that this loss of DP T cells was due to stress-induced apoptosis that resulted from high cortisol expression. When cultured in the presence of the corticosteroid dexamethasone, however, Bak-/-Bax-/- thymocytes were resistant to death (Fig. 1c). This suggested that the decreased numbers of DP cells in Bak-/-Bax-/- mice were not due to stress-induced apoptosis. The increased numbers of DN thymocytes were not due to the increased prevalence of T cell precursors in Bak-/-Bax-/- mice because 77.4  5.8% of the accumulated DN thymocytes were CD3+ small resting cells, compared to 5.1 1.8% in littermate control mice (data not shown).
 | | |
Because Bak-/-Bax-/- mice have abnormalities in multiple cell lineages, defects in nonhematopoietic cells could contribute to the altered thymic development and T cell homeostasis. To determine which defects in Bak-/-Bax-/- mice were intrinsic to the hematopoietic system, bone marrow from Bak-/-Bax-/- and control mice was transferred into lethally irradiated recombination-activating gene 1−deficient (RAG-1-/-) recipient mice. This allowed analysis of reconstituted Bak-/-Bax-/- hematopoietic systems in the context of otherwise normal animals that lacked endogenous lymphocytes. Bone marrow was collected from 8- to 16-week-old donors. Despite the presence of altered thymic phenotypes and splenomegaly in Bak-/-Bax-/- donors, bone marrow from Bak-/-Bax-/- mice reconstituted the lymphoid systems of RAG-1-/- recipients at least as well as marrow from wild-type (WT) mice. Thus, the hematopoietic precursor cells from Bak-/-Bax-/- donors were capable of efficient thymic reconstitution and peripheral T cell accumulation. Three weeks after reconstitution, the highest number of thymocytes was observed in animals reconstituted with Bak-/-Bax-/- bone marrow (Fig. 2a). Splenic cellularity was also higher in mice reconstituted with Bak-/-Bax-/- bone marrow than in control mice 3 weeks after reconstitution (Fig. 2b). Although peripheral T cell reconstitution 3 weeks after reconstitution was incomplete, the thymic production of T cells had been at least as high in mice reconstituted with Bak-/-Bax-/-compared to WT bone marrow: there were 2.2  106 0.6 106 and 1.2 106 0.4 106 CD44loCD62hi T cells, respectively.
| | |
By 9 weeks, Bak-/-Bax-/-−reconstituted animals showed reduced thymic cellularity compared to the 3-week time point. The thymic cellularity of Bak-/-Bax-/-−reconstituted mice was comparable to mice reconstituted with WT and Bak or Bax single-deficient bone marrow (Fig. 2a). In contrast to the reduced thymic cellularity, mice reconstituted with Bak-/-Bax-/- bone marrow developed splenomegaly by 9 weeks after reconstitution: there were 5.3 107 1.3 107 and 1.6 107 0.2 107 splenic T cells in Bak-/-Bax-/-−and WT-reconstituted mice, respectively (Fig. 2b). Thus, in the absence of Bak and Bax the accumulation of lymphocytes occurred only in the peripheral lymphoid organs and resulted from intrinsic defects caused by Bak and Bax−deficiency in the hematopoietic system.
We found that the cellular composition of thymocytes in Bak-/-Bax-/-−reconstituted mice was increasingly altered as lymphoid reconstitution progressed. Analysis of thymic lobes from 3 week−reconstituted mice showed only slightly altered frequencies in the thymic subsets, with the biggest change being a fivefold increase in the frequency of DN thymocytes (Fig. 3a). Nine weeks after reconstitution, however, the representation of thymic subsets was altered (Fig. 3b). Similar to Bak-/-Bax-/- mice, recipient mice that were reconstituted with Bak-/-Bax-/- bone marrow showed an increased percentage of DN and SP thymocytes and decreased percentage of DP thymocytes 9 weeks after reconstitution.
Bak-/-Bax-/- thymocyte apoptosis
Thymocyte apoptosis by corticosteroids, neglect and antigen receptor−induced pathways are critical in shaping normal thymic development and cellularity. To determine whether the cell-intrinsic deficiency of Bak and Bax affected these apoptosis pathways, we analyzed isolated thymocytes from mice reconstituted with bone marrow from Bak-/-Bax-/- and control donors 3 weeks after reconstitution. At this time point, thymic subset profiles were similar and thus provided equivalent starting cell populations. Corticosteroids play a critical role in stress-induced apoptosis of thymocytes and may be important in repertoire selection by sensitizing thymocytes to cell death in the absence of T cell receptor (TCR) ligation18. Bak-/-Bax-/- thymocytes, however, were resistant to death induced by dexamethasone (Fig. 4a).
| | |
Thymocytes that do not properly rearrange both a TCR and a TCR chain18, lack access to cytokines such as interleukin 7 (IL-7)19, or are not positively selected18 undergo death-by-neglect. Splenocytes and fibroblasts from Bak-/-Bax-/- mice can survive in the absence of extrinsic signals14,
17,
20. To determine whether Bak-/-Bax-/- thymocytes were also resistant to death-by-neglect and whether this resistance was intrinsic to the hematopoietic system, thymocytes were cultured for 4 days in the absence of stimulation and their cell viabilities were measured (Fig. 4b). Unlike control thymocytes that progressively underwent apoptosis, Bak-/-Bax-/- thymocytes remained viable.
Negative selection of Bak-/-Bax-/- cells
Thymocytes also undergo apoptosis by negative selection in response to antigen receptor signaling if the signal is of sufficient strength and ligand binds with high avidity18. To examine the role played by Bak and Bax in TCR-induced apoptosis, thymocytes from 3 week−reconstituted mice were cultured for 20 h on tissue culture plates coated with CD3 monoclonal antibody (mAb) at various concentrations and a constant concentration of CD28 mAb. Unlike control thymocytes, Bak-/-Bax-/- thymocytes did not undergo apoptosis when cultured in the presence of high concentrations of CD3 mAb (Fig. 5a). To determine whether Bak or Bax were required for elimination of T cells by negative selection in vivo, elimination of self-reactive T cells to endogenous retroviral superantigens was examined. On the H-2b background of the donor and recipient mice we used here,  60% of CD4+ TCR V 5+ T cells are eliminated relative to the frequency of DP TCR V5+ thymocytes, whereas TCR V6+ T cells are unaffected21,
22,
23. We therefore determined the frequency of V5+ and V6+ DP thymocytes and CD4+ T cells (Fig. 5b). In agreement with published data, we found that the frequency of TCR V5+ Bak+/+Bax+/+ cells decreased by 57 4% (P < 0.001) in CD4+ T cells compared to DP thymocytes23, which indicated the negative selection of V5+ T cells. Mice reconstituted with Bak-/-Bax-/- bone marrow, however, showed equivalent representation of TCR V5+ cells in both DP thymocytes and CD4+ T cells (P > 0.35). T cells bearing TCR V6 were unaffected in either genotype. Together, these data suggested that deficiency in Bak and Bax both prevented the death of cells that failed to receive positively selecting survival signals and disrupted the cell deaths that occured as a consequence of antigen-specific negative selection.
| | |
Phenotype of Bak-/-Bax-/- thymocytes
Prevention of apoptosis in neglect by the expression of Bcl-xL or Bcl-2 does not inhibit cellular atrophy, but rather allows cells to adapt and survive as small metabolically inactive cells6. Consistent with the observation that Bak-/-Bax-/- thymocytes failed to die upon neglect, we found that these thymocytes were smaller than control thymocytes (Fig. 6a). In addition, unlike control thymocytes, high amounts of CD3 and TCR were expressed in substantial fractions of Bak-/-Bax-/- DN and DP thymocytes. For Bak-/-Bax-/-−reconstituted mice, the percentages of CD3+TCR+ cells were 76% and 37.1% for DN and DP thymocytes, respectively, whereas control DN and DP thymocytes were 5.2% and 3.5% CD3+TCR+, respectively (Fig. 6b). The large numbers of Bak-/-Bax-/- DN thymocytes may have represented cells that accumulated due to failed apoptosis upon either unsuccessful TCR rearrangement or selection.
Thymopoiesis in Bak-/-Bax-/-−reconstituted mice
Because Bak-/-Bax-/- thymocytes resisted multiple forms of apoptosis, we assessed whether turnover of Bak-/-Bax-/- thymocytes was sufficiently rapid to explain the development of peripheral lymphocytosis despite the declining numbers of thymocytes as Bak-/-Bax-/-−reconstituted mice aged. To determine the rates of thymopoiesis for Bak-/-Bax-/- and control thymocytes, mice that had been reconstituted for 10 weeks with Bak-/-Bax-/- or Bak+/-Bax+/- bone marrow were injected with the nucleotide analog bromodeoxyuridine (BrdU). Injections were administered each day for 7 days and were used to indicate the rate at which cells had undergone cell division either in the thymus or from precursor cells in the bone marrow. On day 8, thymocytes were analyzed by flow cytometry for the expression of CD4, CD8 and incorporation of BrdU.
Because the most proliferative phases of thymocyte development are within the DN population and in the transition from DN to DP populations24,
25, most thymocytes that incorporate BrdU do so at these stages. As they continue to differentiate into CD4+ and CD8+ thymocytes, these populations become BrdU+. Thus, the fractions of DP, CD4+ and CD8+ cells that contain BrdU are indicative of the rate at which thymocytes are differentiating. In agreement with the determined half-life for thymocyte differentiation of 3−4 days26, the DN and DP thymocytes from control mice were 71.8 3.8% and 68.9 3.0% BrdU+, respectively, after 7 days of treatment (Fig. 7a). CD4+ and CD8+ SP control thymocytes were 36.6 0.5% to 27.6 0.9% BrdU+ after 7 days. In contrast, Bak-/-Bax-/- thymocytes had a much lower percentage of BrdU+ cells in all thymic subsets; this was most notable in the DN T cell subset, in which BrdU incorporation was reduced >90%. These rates of BrdU incorporation indicated the generation of 4.7 107 0.4 107 new thymocytes per week in control mice; in contrast, only 0.7 107 0.1 107 new Bak-/-Bax-/- thymocytes were generated in the same time period.
| | |
To investigate the reduction in thymopoiesis further, the phenotypes of DN thymocytes from Bak-/-Bax-/-−reconstituted mice and control mice were analyzed. Because Bak-/-Bax-/-−reconstituted mice developed a large population of CD3+ DN thymocytes, only CD3- DN thymocytes were analyzed. DN thymocytes differentiate sequentially through DN1 (CD25-CD44+), DN2 (CD25+CD44+), DN3 (CD25+CD44-), and DN4 (CD25-CD44-) populations. Within these subsets, the DN2 and DN4 populations are the most proliferative, as cells are stimulated at these stages by IL-7 and the pre-TCR, respectively24,
25. Compared to control thymocytes, CD3- DN Bak-/-Bax-/- thymocytes had excess DN1 cells and a near complete lack of DN2 cells (Fig. 7b). To determine the fraction of proliferating DN4 populations in mice reconstituted with WT and Bak-/-Bax-/- bone marrow, the forward light scatter of DN4 cells was analyzed (Fig. 7c). Compared to the control DN4 population, the Bak-/-Bax-/- DN4 population showed a 90% reduction in the number of large blasting cells. These findings indicated that the lack of thympoiesis in Bak-/-Bax-/-−reconstituted mice was not due to a deficiency of early thymocyte precursors or an intrinsic inability to differentiate; rather, it was due to a reduction in the proliferative expansion of Bak-/-Bax-/- thymocytes at both the DN2 and DN4 stages of development.
Characteristics of Bak-/-Bax-/- T cells
In the periphery, the excess Bak-/-Bax-/- T cells that accumulate have the phenotype of memory cells14. Nine weeks after reconstitution, CD4+ and CD8+ T cells were represented in normal proportions in the spleens of Bak-/-Bax-/-−reconstituted mice (Fig. 8a). However, the absolute numbers of T cells were substantially higher for both lineages due to the splenomegaly observed in these mice. Unlike splenic T cells 3 weeks after reconstitution—which were comparable to control cells for CD44 and CD62 ligand (CD62L) expression—accumulated T cells 9 weeks after reconstitution were 90.4% CD44hiCD62Llo, compared to 36.6−44.2% CD44hiCD62Llo in WT or either of the Bak or Bax single-deficient control mice (Fig. 8b). These data indicated that Bak-/-Bax-/- T cells were not intrinsically activated and suggested that the splenomegaly in Bak-/-Bax-/- mice was caused by a failure to eliminate cells after antigen clearance in immune responses.
| | |
To determine whether Bak and Bax were required for T cell death after immunization and expansion by antigen, T cell proliferation and elimination were observed after immunization with the superantigen staphylococcal enterotoxin B (SEB). Mice reconstituted with Bak-/-Bax-/- or control bone marrow were immunized with SEB shortly after peripheral T cell reconstitution; their blood was periodically sampled to determine the percentage of CD4+ T cells bearing TCRs with V8, which are specifically activated by SEB, or V6, which are unaffected by SEB and acted as a control (Fig. 9a). We found that the numbers of Bak-/-Bax-/- and Bak+/-Bax+/- V8 CD4+ T cells had expanded markedly by day 4 after immunization. After this point, Bak+/-Bax+/- V8 CD4+ T cells decreased in frequency to day 16 after immunization, where they stabilized at a frequency that was slightly higher than in unimmunized mice. In contrast, Bak-/-Bax-/- V8 CD4+ T cells remained at the expanded frequency in the blood and did not appear to be eliminated. CD4+ T cells bearing V6 TCRs were unaffected in both Bak-/-Bax-/- and control bone marrow−reconstituted mice.
| | |
To ensure that changes in the frequency of CD4+ V8−bearing T cells in blood accurately portrayed changes in cell numbers in lymphoid tissue, unimmunized and SEB-immunized mice whose blood had not been sampled were analyzed 20 days after immunization to determine the number of CD4+ V8 and V6 T cells in the spleens of Bak-/-Bax-/- and control bone marrow−reconstituted animals (Fig. 9b). In support of the results we obtained by blood analysis, Bak+/-Bax+/- V8 CD4+ T cells in SEB-immunized mice were similar in number to unimmunized control animals by 20 days after immunization. However, there was a significant increase in the number of V8 CD4+ T cells in Bak-/-Bax-/-−reconstituted mice after immunization compared to unimmunized Bak-/-Bax-/-−reconstituted mice. V6 CD4+ T cells did not change significantly in number after SEB immunization in either control or Bak-/-Bax-/- bone marrow−reconstituted mice. Thus, elimination of T cells after superantigen-induced expansion in vivo required the presence of Bak or Bax.
Discussion
We found that Bak and Bax deficiency resulted in altered thymocyte development and peripheral T cell homeostasis. This altered thymic subset profile was not due to peripheral lymphocytosis and recolonization or contamination of the thymus with peripheral lymphocytes because, unlike thymocytes that were CD44lo, peripheral T cells were CD44hi. Neither was the increased frequency of DN cells due to the presence of B cells or natural killer (NK) cells, as cells were predominantly CD3+ and lacked the B cell marker B220 or the NK cell marker NK1.1 (data not shown). In addition, CD3+ DN cells remained in the thymus because CD3+ T cells in the periphery were either CD4+ or CD8+ (data not shown).
The failure of Bak-/-Bax-/- thymocytes to die in vitro when subjected to neglect or antigen receptor stimulation suggested that the abnormal thymocytes in Bak-/-Bax-/-−reconstituted mice may have accumulated as a consequence of failed selection. The accumulation of DN Bak-/-Bax-/- T cells may have been due to the survival of cells that did not properly rearrange a complete TCR and/or receive sufficient survival signals to allow progression to the DP stage of differentiation27. The accumulation of CD4+ and CD8+ thymocytes may also have been caused by the inability of cells to undergo apoptosis in response to failed positive or negative selection. Because cells that fail positive selection die as a result of insufficient signals through their antigen receptors or corticosteriods18, failed positive selection represents a form of death-by-neglect. The resistance of Bak-/-Bax-/- thymocytes to death-by-neglect, therefore, suggested that Bak-/-Bax-/- thymocytes were less susceptible to dying when positive selection had failed.
Although the resistance of Bak-/-Bax-/- thymocytes to death in response to antigen receptor ligation suggested a deficiency in negative selection, similar analysis of Bcl-xL−transgenic T cells shows that Bcl-xL expression confers T cell resistance to in vitro antigen receptor−induced death but permits near normal deletion of T cells to in vivo self-antigens4. Similarly, overexpression of Bcl-2 has limited in vivo effects on T cell selection28,
29,
30,
31,
32. In contrast, Bim-deficient mice fail to completely eliminate self-reactive lymphocytes in vivo12. Because Bim requires Bak and Bax to elicit cell death16,
17, a role for Bak and Bax in negative selection is likely. This hypothesis is supported by the inability of endogenous retroviral superantigens to eliminate Bak-/-Bax-/- V5 T cells.
A failure to die through negative selection could account for the accumulation of CD3+ DN thymocytes in the Bak-/-Bax-/-−reconstituted mice, as such cells have been observed when negative selection is impaired33,
34. Although only small numbers of these cells have been reported in Bim-deficient or Bcl-2−transgenic animals12, the complete inhibition of mitochondrially induced death by Bak and Bax deficiency may allow their extensive accumulation. Activation of Bak and Bax is mediated through the action of BH3 proteins. Therefore, the increased numbers of CD3+ DN cells in Bak-/-Bax-/-−reconstituted mice relative to Bim-deficient mice may indicate that other BH3-only proteins, in addition to Bim, are also important in thymic selection.
In the absence of Bak and Bax, mitochondrially based apoptosis is blocked and T cells accumulate in large numbers in peripheral lymphoid organs. Transgenic expression of Bcl-2 and Bcl-xL can also cause T cell accumulation6. Whether Bcl-2 and Bcl-xL affect T cell numbers by increased production and survival of T cell progenitors or by increased survival of peripheral T cells is unclear. Despite minor effects of Bcl-2 overexpression on T cell selection28,
29,
30,
31,
32, no effect on T cell production and thymopoiesis has been reported. We found that turnover of Bak-/-Bax-/- thymocytes, however, was reduced by >85%, even though the numbers of hematopoietic precursors in the bone marrow are markedly increased in Bak-/-Bax-/- mice throughout their lifetimes14. The memory phenotype of the accumulated Bak-/-Bax-/- T cells combined with their failure to be eliminated after immunization with SEB suggested that rather than accumulating as a result of excess T cell production, peripheral T cells accumulated as a result of antigen-induced expansion and failure to be eliminated after antigen clearance.
In contrast to the accumulation of lymphocytes in the periphery, thymic cellularity declined in mice reconstituted with Bak-/-Bax-/- bone marrow as they aged. This reduction in cell number appeared to occur because thymopoiesis declined in Bak-/-Bax-/-−reconstituted mice. Because the generation of peripheral T cells in Bak-/-Bax-/-−reconstituted mice was similar or slightly enhanced compared to control mice, Bak-/-Bax-/- thymocytes must not intrinsically lack the ability to develop properly. Rather, cell-extrinsic effects—such as the accumulation of cells that did not undergo Bak and Bax−dependent elimination—appeared to have inhibited thymopoiesis. Although reduced rates of thymocyte proliferation have been reported in Bcl-2−transgenic mice35,
36,
37, a marked effect on the overall rate of thymopoiesis and thymic output has not been reported. There are two likely reasons this effect has not been observed in Bcl-2− or Bcl-xL−transgenic mice. First, overexpression of a transgene is dependent on the range of activity of an exogenous promoter and therefore may not indicate the function of a gene as clearly as a gene deficiency. Second, because Bak and Bax are required to initiate the mitochondrial death pathway, their deficiency causes a complete block in cell death by this pathway, whereas Bcl-2 or Bcl-xL overexpression or Bim deficiency may offer only partial protection and, thus, not allow sufficient cell accumulation to greatly inhibit thymopoiesis.
Thymopoiesis could be inhibited by the accumulation of cells in the thymus or in the periphery. Accumulating small Bak-/-Bax-/- thymocytes may have limited the availability of ligands required for prothymocytes to initiate or sustain T cell development. However, we found that the thymi of Bak-/-Bax-/-−reconstituted mice could support larger numbers of thymocytes early after reconstitution. In addition, small cells show a reduced capacity to bind and sequester growth factors and lack the ability to capture nutrients6. Thus, it is unlikely that the accumulated Bak-/-Bax-/- thymocytes inhibited normal T cell development by out-competing newly arrived thymic immigrants from the bone marrow. Instead, accumulation of peripheral T cells may lead to feedback inhibition of thymopoiesis. Therefore, the pro-apoptotic functions of Bak and Bax are required to maintain the turnover of T cells, and, in the absence of such turnover, thymopoiesis may be inhibited.
Other genetic models, such as CD25-/- (ref. 38) and CTLA-4-/- (refs. 39,
40,
41) mice, also accumulate peripheral T cells, but no loss of thymopoiesis has been reported. In contrast to Bak-/-Bax-/- T cells that accumulate as resting memory cells, CD25-/- and CTLA-4-/- T cells accumulate in an activated state. This raises the possibility that the accumulation of memory T cells may have different homeostatic effects on thymopoiesis compared to the accumulation of activated T cells that are participating in an ongoing immune response. Such a model may explain why marked T cell expansions that can occur in immune responses do not affect thymopoiesis, yet allow for a feedback mechanism to reduce thymopoiesis when the T cell compartment is fully constituted. Thus, the immune systems of Bak-/-Bax-/-−reconstituted mice resemble those of aged individuals, with accumulation of memory lymphocytes and inhibition of thymopoiesis. Such a feedback mechanism has been suggested by the resumption of thymopoiesis after antiretroviral therapy in T cell−depleted HIV patients42,
43. The mechanism of such a feedback pathway is not yet understood, but will be important to consider in future work aimed at understanding the control of lymphocyte development and homeostasis.
Methods
Mice.
Bak and Bax−deficient mice were bred and maintained by intercrossing Bak+/-Bax+/- mice on a C57BL/6 (129S3 129X1) background, as described14. Mice were 8−16 weeks of age when analyzed or used as donors in bone marrow−transfer experiments. For bone marrow transfers, 8−16-week-old H-2−matched female recipient RAG-1-/- C57BL/6 129S−RAG-1tmMom mice (Jackson Laboratory, Bar Harbor, ME) were lethally irradiated with  -irradiation (9 Gy) 1 day before reconstitution with 1 106−2 106 bone marrow cells from donor mice. Recipient mice were maintained on antibiotic water (Sulfamethoxazole and Trimethoprim Oral Suspension, Hi Tech Pharmcal, Amityville, NY) for the duration of each experiment. Recipient mice were analyzed 3 or 8−10 weeks after reconstitution. Cell numbers in tissues recovered from mice after death were determined by counting cell yields in single-cell suspensions with a hemocytometer. Animal experiments were done according to institutional guidelines.
Cell culture.
Cell suspensions of thymocytes were cultured in vitro in complete RPMI-1640 with 10% fetal calf serum (FCS, Gibco-BRL, Grand Island, NY). Thymocytes were cultured in medium alone, in various concentrations of dexamethasone (Sigma, St. Louis, MO) or on plates coated with CD3 mAb (clone 145-2C11, BD-Pharmingen, San Diego, CA) at various concentrations. In samples cultured in CD3 mAb, plates were also coated with 10  g/ml of CD28 mAb (clone 37.51, BD-Pharmingen). Cell viabilities were determined by flow cytometry with propidium iodide (10 g/ml) exclusion (Molecular Probes; Eugene, OR).
Flow cytometry.
Flow cytometry was done with a FACSCalibur and data were analyzed with Cell Quest software (Becton Dickinson, Mountain View, CA). Antibodies used were all from BD-Pharmingen and included fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)- or cychrome-conjugated CD4 mAb (clone H129.19); FITC-, PE- or allophycocyanin-conjugated CD8 mAb (clone 53-6.7); allophycocyanin-CD62L mAb (clone MEL-14); cychrome-conjugated CD44 mAb (clone IM7); FITC- or PE-conjugated CD3 mAb (clone 145-2C11); FITC-conjugated TCR mAb (clone H57-597); PE- or allophycocyanin-conjugated CD25 mAb (clone PC61); allophycocyanin-conjugated B220 mAb (clone RA3-6B2); and PE-conjugated NK1.1 mAb (clone PK136). TCR Vs were screened with antibodies from the mouse V TCR−screening panel.
BrdU treatment and detection.
Mice were treated daily for 7 days with BrdU (1 mg in 200 l of PBS, Sigma) by intraperitoneal injection 10 weeks after reconstitution with Bak-/-Bax-/- or control bone marrow. On day 8, mice were killed and thymocyte cell suspensions were prepared. BrdU incorporation was detected by flow cytometry with a BrdU Flow Kit, as described by the manufacturer (BD-Pharmingen). Briefly, thymocytes were stained with PE-conjugated CD4 mAb and allophycocyanin-conjugated CD8 mAb, fixed, treated with DNAse and stained with BrdU mAb (clone 3D4) followed by FITC-conjugated mouse IgG1 (clone A85-1). Both mAbs were from BD-Pharmingen.
SEB immunization.
Mice were immunized with 100 g of SEB (Sigma) in 200 l of PBS with 2% FCS by intraperitoneal injection 8 weeks after reconstitution with Bak-/-Bax-/- or control bone marrow. In one experiment, mice were bled through the tail vein immediately before immunization; this was followed by serial blood collections at various time points after immunization to determine the expansion of TCR V8.1+ or TCR V8.2+ and TCR V6+ CD4+ T cells as a fraction of total blood CD4+ T cells. For analysis, red blood cells were lysed and the frequencies of T cells bearing TCR V8 and TCR V6 as a fraction of total CD4+ T cells were determined by flow cytometry: cells were stained with cychrome-conjugated CD4 mAb, PE-conjugated V8.1 or V8.2 mAbs (clone MR5-2, BD-Pharmingen) and FITC-conjugated V6 mAb (clone RR4-7, BD-Pharmingen). In another experiment, mice were not bled but were killed 20 days after immunization with SEB or PBS and cell suspensions of splenocytes were analyzed to determine the absolute numbers of splenic TCR V8.1+ or TCR V8.2+ and TCR V6+ CD4+ T cells. Absolute numbers of TCR V8.1+ or TCR V8.2+ and TCR V6+ CD4+ T cells were determined by multiplying the frequency of these cell populations with the spleen cell numbers determined by hemocytometer.
Received 22 March 2002; Accepted 23 July 2002; Published online: 3 September 2002.
REFERENCES
-
Marrack, P. et al. Homeostasis of TCR+ T cells. Nature Immunol. 1, 107–111 (2000). | Article | PubMed | ISI | ChemPort |
-
Van Parijs, L., Biuckians, A. & Abbas, A.K. Functional roles of Fas and Bcl-2-regulated apoptosis of T lymphocytes. J. Immunol. 160, 2065–2071 (1998). | PubMed | ISI | ChemPort |
-
Vander Heiden, M.G., Chandel, N.S., Schumacker, P.T. & Thompson, C.B. Bcl-xLprevents cell death following growth factor withdrawal by facilitating mitochondrial ATP/ADP exchange. Mol. Cell 3, 159–167 (1999). | Article | PubMed | ISI | ChemPort |
-
Grillot, D.A., Merino, R. & Nunez, G. Bcl-XL displays restricted distribution during T cell development and inhibits multiple forms of apoptosis but not clonal deletion in transgenic mice. J. Exp. Med. 182, 1973–1983 (1995). | Article | PubMed | ISI | ChemPort |
-
Chao, D.T. et al. Bcl-xL and Bcl-2 repress a common pathway of cell death. J. Exp. Med. 182, 821–828 (1995). | Article | PubMed | ISI | ChemPort |
-
Rathmell, J.C., Vander Heiden, M.G., Harris, M.H., Frauwirth, K.A. & Thompson, C.B. In the absence of extrinsic signals, nutrient utilization by lymphocytes is insufficient to maintain either cell size or viability. Mol. Cell 6, 683–692 (2000). | Article | PubMed | ISI | ChemPort |
-
Veis, D.J., Sorenson, C.M., Shutter, J.R. & Korsmeyer, S.J. Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair. Cell 75, 229–240 (1993). | PubMed | ISI | ChemPort |
-
Motoyama, N. et al. Massive cell death of immature hematopoietic cells and neurons in Bcl-x-deficient mice. Science 267, 1506–1510 (1995). | PubMed | ISI | ChemPort |
-
Brady, H.J., Gil-Gomez, G., Kirberg, J. & Berns, A.J. Bax perturbs T cell development and affects cell cycle entry of T cells. EMBO J. 15, 6991–7001 (1996). | PubMed | ISI | ChemPort |
-
O'Connor, L. et al. Bim: a novel member of the Bcl-2 family that promotes apoptosis. EMBO J. 17, 384–395 (1998). | Article | PubMed | ChemPort |
-
Bouillet, P. et al. Proapoptotic Bcl-2 relative Bim required for certain apoptotic responses, leukocyte homeostasis, and to preclude autoimmunity. Science 286, 1735–1738 (1999). | Article | PubMed | ISI | ChemPort |
-
Bouillet, P. et al. BH3-only Bcl-2 family member Bim is required for apoptosis of autoreactive thymocytes. Nature 415, 922–926 (2002). |
|
