B6.H-2Kb−/−Db−/− (DKO) mice have greatly reduced numbers of mature CD8αβ T cells in their periphery. However, these non-class Ia-selected CD8αβ T cells are able to mediate immune responses to a number of pathogens. Approximately 60% of the CD8αβ T cells in the spleen and peripheral lymph nodes of naive DKO mice display a memory (CD44high) phenotype. To investigate the origins of these non-class Ia-selected CD8αβCD44high cells, we traced the phenotype of recent thymic emigrants and found that most were CD44low. We also determined whether their appearance was thymus dependent and found that only a small percentage of non-class Ia-selected CD8αβCD44high cells develop in a thymus-independent pathway. Functionally, CD8αβCD44high cells from DKO mice are able to secrete IFN-γ in response to IL-12 and IL-18 in the absence of cognate Ag. When challenged with anti-CD3 in vivo, nearly half of these cells produce IFN-γ within 3 h. When purified CD8αβCD44high cells from Thy1.2.DKO mice were transferred into Thy1.1 DKO recipients and then challenged with Listeria monocytogenes, an Ag-specific anti-L. monocytogenes response was observed 6 days later. Our data suggest that non-class Ia-selected CD8αβCD44high cells in naive animals can respond rapidly to Ag and play a role in the innate as well as the early phase of the acquired immune response.
Upon encountering foreign microorganisms, naive CD8 T cells differentiate in an Ag-specific manner into effector CD8 T cells (1). After resolving the infection, most of the effector CD8 T cell population undergoes massive contraction by the process of activation-induced apoptosis (2). Only a small population of memory cells is maintained, which can respond rapidly upon re-exposure to the pathogen (3). CD8 memory cells can be defined phenotypically as CD44high. However, they also up-regulate other cell surface markers, such as Ly6C, CD122, CD43, CD132, ICAM-1, LFA-1, and IL-7Rα (CD127) (4, 5, 6, 7). The pattern of gene expression by memory CD8 T cell has been studied. Genes involved in signal transduction, such as members of the p38 and JNK signaling pathways; cell migration, such as CCR2 and CXCR4; cell division, such as cyclin E1 and B1; and effector functions, such as IFN-γ and Fas ligand, were found to be up-regulated (6). Memory CD8 T cells require lower concentrations of Ag and less costimulation than naive CD8 T cells for activation (8). They are able to rapidly produce IFN-γ after restimulation. Unlike naive CD8 T cells, they are maintained in vivo in the absence of MHC class I molecules (9, 10).
CD8 T cells can also acquire a memory phenotype in the absence of foreign Ags by several different mechanisms. For example, when naive CD8 T cells undergo homeostatic proliferation in a lymphopenic environment, this results in the up-regulation of expression of CD44, CD122, Ly6C, as well as other cell surface markers. These T cells that expand under lymphopenic conditions appear indistinguishable from Ag-experienced memory T cells (4, 11). CD8 T cells undergoing homeostatic proliferation are capable of lysing target cells directly in vitro and in vivo and secrete IFN-γ rapidly upon restimulation, similar to memory cells (5). CD8 T cells that develop in athymic mice also express memory markers (12). It has been reported that these extrathymically developed T cells are positively selected by self-Ags and possess a higher activation threshold than naive T cells (13, 14). Class Ib-restricted CD8 T cells have been shown to be positively selected on hemopoietic cells in the thymus and acquire memory cell markers, such as CD44high, at this site (15).
Approximately 20% of CD8 T cells from naive B6 mice display memory markers, and this number increases with age (16). These cells are distinct from conventional memory CD8 T cells in that, upon activation by IL-2, they express high levels of the NK receptor NKG2D and adaptor molecule DAP12 that are normally expressed in activated NK cells (17). IL-2-activated CD8CD44high cells are able to kill syngeneic tumor cells, suggesting that they may play a role in the surveillance of host cells that have been altered through infection or transformation (17). By analogy with H-Y Ag-specific TCR-transgenic mice, it has been suggested that the development of a significant portion of memory phenotype CD8 T cells in naive wild-type mice is driven by the interaction of αβTCR with self-Ags and occurs extrathymically (13, 18, 19).
B6.H-2Kb−/−Db−/− (DKO) mice lack expression of class Ia Kb and Db H chains, but show normal expression of β2-microglobulin and class Ib molecules (20). In the thymus of DKO mice, most mature single-positive CD8αβ T cells are CD44low, similar to those in B6 mice (21). However, most CD8 T cells from DKO spleens have a CD44highCD122highLy6Chigh phenotype characteristic of memory CD8 T cells (22). The CD8 molecule can be expressed either as an αα homodimer or an αβ heterodimer. CD8αα cells are different from CD8αβ cells in a number of phenotypic characteristics. They have been shown to be CD44high and can be either class I or class II restricted (23, 24). However, most CD8 cells in DKO mice are CD8αβ cells (21). Their origin and function have not been widely investigated. Some of these cells could be the result of CD8 T cells being selected in the thymus on hemopoietic cells (15). Alternatively, the cells may be the result of an extrathymic pathway for their development or the result of a postthymic encounter with factors in the periphery.
Although DKO mice have greatly reduced numbers of mature CD8+ T cells, these non-class Ia-selected CD8+ T cells are able to function effectively to mediate immune responses against infectious agents. For example, H2-M3-restricted CD8 T cells can play an important role in the clearance of Listeria monocytogenes (LM)3 infection (22, 25, 26, 27, 28). H2-M3-restricted CD8+ cells have also been found in Mycobacteria tuberculosis-infected mice (29, 30).
LM is a Gram-positive intracellular bacterium that has been used widely as a laboratory model to understand immune responses to intracellular bacterial infections (31). Sterilizing clearance of LM is mediated by CD8 T cells (32). The CD8 T cells that mediate protective immunity include two populations: one is restricted by MHC class Ia molecules, and the other is restricted by MHC class Ib molecules. These two T cell populations respond to bacterial infections with different kinetics and potentially provide distinct contributions to the immune response (22, 33). MHC class Ia-restricted CD8 T cells reach their peak response to LM ∼8 days after i.v. inoculation. In contrast, class Ib H2-M3-restricted CD8 T cells undergo rapid clonal expansion and achieve their peak response 5–6 days after primary infection with LM in both B6 as well as DKO animals (22, 27, 34). Although it has been suggested that anti-LM CD8 T cells may have been primed by previous exposure to commensal organisms in LM-naive animals, the phenotype of the CD8 T cells that respond to this infection is not known (35).
In addition to acquired immunity, CD8 T cells can play a role in innate immunity by secreting IFN-γ in response to cytokines. IFN-γ is produced in most types of infections and plays a key role in innate immune responses against pathogens (36, 37, 38). It is known that NK cells provide an early and necessary source of IFN-γ (37, 39). However, a portion of CD8CD44high memory cells has also been reported to be an early source of IFN-γ, especially in the lymph nodes after LPS injection (38). IFN-γ-secreting CD8 cells can also be found in naive B6 mice 16 h after LM infection (36). This finding suggested that CD8CD44high memory cells may contribute to innate immunity by providing an early non-Ag-specific source of IFN-γ. It also has been reported that IFN-γ is induced rapidly in a small subset of CD8 T cells after anti-CD3 injection. By examining the cells from normal and various MHC-knockout mice, MHC class Ib-restricted and TAP-independent CD8+ cells have been proposed to be an early source of IFN-γ that has a role in polarizing CD4 T cells to become Th1 cells (40).
In this report, we characterize the origin and functional properties of CD8αβCD44high cells in mice lacking class Ia Ags. Our results show that they can play a role in both innate and adaptive immune responses against infection by LM.
Materials and Methods
All mice were bred and maintained at the University of Texas Southwestern Medical Center animal facility. DKO mice were generated as previously described and were a generous gift from F. Lemonnier (Institute Pasteur, Paris, France) (20). DKO were crossed to B6.Thy1.1 mice and the F2 progeny were screened for Thy1.1+DKO animals.
LM 10403 serotype 1 was originally provided by H. G. A. Bouwer (Veterans Affairs Medical Center, Portland, OR). Bacteria are grown on brain-heart infusion agar plates (Difco). Virulent stocks were maintained by repeated passage through B6 mice. For infection, log-phase cultures of LM grown in brain-heart infusion broth were washed twice and diluted in PBS before injection of 2 × 103 bacteria in the lateral tail vein.
Cell lines, cell culture, and reagents
The J774 macrophage cell line (H-2d) was grown in DMEM with 10% FCS, l-glutamine, vitamins, and penicillin/streptomycin. For infection of J774 cells with LM, antibiotic-free medium was used. Culture of mouse splenocytes was conducted in RPMI 1640 supplemented with 10% FCS and recombinant human IL-2. Recombinant murine IL-12 and IL-18 were purchased from PeproTech. Blocking Abs against murine IL-12 and IL-18 were purchased from PeproTech and Medical and Biological Laboratories, respectively, and were used at 1 μg/ml final concentration.
Abs and flow cytometry
The following Abs from BD Biosciences were used for flow cytometry: anti-CD24 (M1/69), anti-CD8α (53-6.7), anti-CD8β (53-5.8), anti-CD44 (IM7), anti-IFN-γ (XMG1.2), anti-CD90.2 (Thy1.2; 53-2.1), anti-CD90.1 (Thy1.1; OX-7), and anti-TCR Vβ segments (TCR Vβ2, -4, -5, -6, -7, -8.1/.2, 8.3, -9, -10, -11, -12, -13, and -14). Secondary streptavidin-conjugated reagents were used to reveal biotinylated primary Abs. Data were acquired using a FACSCalibur flow cytometer and were analyzed using CellQuest software (BD Biosciences).
In vitro IL-12 and IL-18 stimulation
Splenocytes were cultured in 24-well plates at a concentration of 3 × 106/well with 20 U/ml recombinant human IL-2 or with 20 U/ml recombinant human IL-2, 5 ng/ml IL-12, and 10 ng/ml IL-18 overnight. GolgiPlug containing brefeldin A (BD Pharmingen) was added 4 h before harvest of the cultures. IFN-γ-producing CD8 T cells were detected with the Cytofix-Cytoperm kit Plus (BD Pharmingen) according to the manufacturer’s protocol as previously described (41).
In vivo anti-CD3 stimulation
Mice were injected i.p. with 5 μg/mouse NA/LE anti-CD3e (145-2C11) or the same amount of control isotype-matched IgG (G235-2356; BD Pharmingen). Three hours later, splenocytes were harvested and cultured in 24-well plates at a concentration of 3 × 106/well for 4 h. GolgiPlug containing brefeldin A was added 2 h before harvest of the cultures. IFN-γ-producing CD8 T cells were detected as previously described (41).
CD8 T cell purification, adoptive transfer, and stimulation
Single-cell suspensions from the spleens of DKO mice were pooled, and CD8α cells were purified with CD8α magnetic beads from BD Pharmingen according to the manufacturer’s instructions. For separation of CD8αCD44high cells, CD8α magnetic bead-purified CD8+ T cells were stained with anti-CD44-PE. After washing, the CD8α cell population was sorted on the basis of CD44high using a MoFlo high-speed sorter (DakoCytomation). Purified CD8αCD44high cells (2 × 105) from DKO mice were injected i.v. into Thy1.1.DKO recipients and infected the next day with 2 × 103 LM. Six days later, the mice were killed, and the resulting splenocytes were cultured in 24-well plates at a concentration of 3 × 106/well with 3 × 105 LM-infected J774 cells overnight.
Intrathymic FITC injection
Intrathymic injections were performed according to published protocols (42, 43). Briefly, 6-wk-old B6 or B6.DKO mice were anesthetized, and the chest cavities were opened. Ten microliters of FITC solution (1 mg/ml; Sigma-Aldrich) was injected into each thymic lobe with a 30-gauge needle. The chest was then closed with surgical clips. After 16–20 h, the thymus, spleens, and lymph nodes were harvested and stained.
Thymectomy and bone marrow (bm) chimeras
Suction thymectomies were performed at 6–8 wk of age as previously described (44, 45). Completeness of the procedure was confirmed by visual inspection at the time of death. The bm chimeras were prepared by gamma-irradiating (1200 rad) thymectomized or normal B6.DKO recipient mice with a 137Cs source, followed by i.v. injection of 107 T-depleted bm cells from Thy1.1.DKO donors (43). Before injection, T cell depletion was performed with CD4 and CD8 Dynal beads according to the manufacturer’s instructions and was >95% effective. Chimeras were infected with LM 6–8 wk later.
Acquisition of the memory phenotype of CD8 T cells in DKO animals
The goal of these initial experiments was to determine when non-class Ia-selected CD8αβ T cells acquired their memory phenotype. We examined CD44 expression on FITC+ recent thymic emigrants (RTEs; after 16–20 h) in B6 and DKO mice after intrathymic FITC injection. As shown in Fig. 1,A, ∼15% of mature single-positive CD8αβ thymocytes, defined as CD8αβCD24−, were CD44high in DKO mice, similar to those in B6 animals. Sixteen to 20 h after intrathymic FITC injection, we found that ∼23% of the FITC+ CD8αβ RTEs in the spleens were CD44high in DKO mice. This is much lower than resident CD8αβ splenocytes, ∼55% of which were CD44high. This indicates that the acquisition of a high expression of CD44 by CD8αβ cells in the periphery of DKO mice is a postthymic event. Similar results were noted in B6 animals, which is consistent with previous reports that recent thymic emigrants are CD44low (46, 47). Similar results for RTEs were observed in lymph nodes (Fig. 1,A). As a control, we examined the CD44 phenotype for CD4 cells. CD4+ RTEs and resident cells from both B6 and DKO mice have similar low levels of CD44 expression in both spleen and lymph nodes (Fig. 1 B).
Few CD8 T cells arise in thymectomized DKO bm chimeras
It has been reported that CD8 T cells that arise in a thymic-independent environment are CD44high (12). To test the possibility that CD8αβCD44high cells in naive DKO mice developed from a thymic-independent pathway, we generated thymectomized (Tx) bm (Txbm) chimeras by repopulating irradiated Tx DKO animals with bm from Thy1.1.DKO donors. Only 0.11% donor-derived Thy1.1+CD8αβ T cells developed in the spleen of a Txbm chimera (Fig. 2,A). In contrast, 0.37% of donor-derived Thy1.1+CD8αβ T cells developed in the spleen of a thymus-intact bm chimera (Fig. 2,C). Almost all Thy1.1+CD8αβ T cells in the Txbm chimera were CD44high (Fig. 2,B), consistent with previous reports (23). However, <50% of the Thy1.1+CD8αβ T cells from thymus-intact bm chimera were CD44high (Fig. 2,D), similar to that found in normal DKO animals. The data from all chimeras examined are summarized in Fig. 2 E. Taken together with the analysis of RTEs, our data indicate that most CD8αβCD44high cells in DKO mice are thymus dependent and acquire their phenotype in the periphery.
Ability of CD8 T cells from DKO mice to participate in the innate immune response
IL-12 and IL-18 are cytokines that are secreted after infection with LM and other pathogens and can activate cells to secrete IFN-γ during the innate phase of the immune response. Both memory and effector CD8 T cells express receptors for these cytokines and have been shown to secrete IFN-γ both in vitro and in vivo (48, 49). To investigate the ability of non-class Ia-selected CD8αβCD44high cells to contribute to innate immunity, CD8+ T cells from naive DKO mice were cultured overnight with IL-2 (control), or IL-2, IL-12, and IL-18 and tested for their ability to secrete IFN-γ by intracellular cytokine staining (ICS). Culture of CD8+ cells with IL-2 alone did not induce IFN-γ secretion, as expected. However, >50% CD8αβ+CD44high cells secreted IFN-γ in response to IL-12 and IL-18 (Fig. 3). Therefore, the memory phenotype CD8 T cells in naive DKO animals can participate in the innate response similarly to that described for conventional memory cells (38, 48).
Ability of CD8 T cells from DKO mice to respond rapidly to anti-CD3 in vivo
MHC class Ib-restricted, TAP-independent CD8+ cells have been proposed to be an early source of IFN-γ after anti-CD3 injection in vivo (40). It is likely that the non-class Ia-selected CD8αβCD44high cells are responsible for this rapid IFN-γ secretion within the first few hours after stimulation. After 3 h of anti-CD3 in vivo challenge, almost half the CD8αβCD44high cells from DKO mice secreted IFN-γ, compared with ∼20% of CD8αβCD44high from B6 mice (Fig. 4,A). In contrast, only a very small percentage of CD8αβCD44low cells from both B6 and DKO animals secreted IFN-γ (Fig. 4 A). These data indicate that one of the sources of the early burst of IFN-γ is from CD8αβCD44high cells restricted to a non-class Ia MHC molecule(s).
Previous data show a similar Vβ gene usage between DKO and B6 CD8 T cells (50). We also noted that CD8αβCD44high cells in DKO mice have a similar Vβ diversity as those in B6 animals (data not shown). To investigate Vβ usage by this population of IFN-γ-secreting CD8αβCD44high cells, we used a panel of Vβ Abs to analyze their usage and compared this between DKO and B6 animals. As shown in Fig. 4 B, all Vβ-expressing populations of the panel tested from CD8αβCD44high cells in both B6 and DKO mice showed a comparable ability to secrete IFN-γ after anti-CD3 stimulation. There was a higher percentage of IFN-γ-secreting cells in each Vβ subset in DKO vs B6 animals, but there was no evidence of Vβ repertoire skewing in this response.
Ability of thymus-independent CD8αβCD44high cells from DKO animals to respond to LM challenge
It has been previously shown that CD8 T cells from DKO mice respond to infection with LM (22, 27). To test whether the responding CD8 cells arose from a thymus-dependent vs thymus-independent pathway, we challenged Thy1.1.DKO-reconstituted DKO Txbm chimeras and Thy1.1.DKO-reconstituted DKO (thymus-intact) bm chimeras with LM. CD8αβ cells in the Txbm DKO chimera showed minimal expansion 7 days after challenge with LM (compare Figs. 5,A with 2,A). In contrast, in the thymus-intact bm chimera, the CD8αβ T cells expand in response to LM infection (compare Figs. 5,B with 2,C), and more CD8αβ cells display the CD44high phenotype (compare Figs. 5,B with 2,D). The data are summarized in Fig. 5 C, where it is noted that there was a large increase in CD8αβ T cells in normal DKO and thymus-intact bm chimeras 7 days after infection with LM, whereas no expansion was seen in the Txbm chimeras.
Although thymus-independent, non-class Ia-selected CD8 cells did not show expansion after LM challenge, we investigated their ability to secrete IFN-γ in response to LM infection. This was performed by culturing splenocytes from LM-infected Txbm and thymus-intact bm chimeras overnight with LM-infected J774 cells. As previously reported, LM-infected J774 cells induce IFN-γ secretion in CD8 T cells by two pathways. One occurs when J774 cells present LM epitopes to CD8 T cells that induce Ag-specific IFN-γ secretion; the other is induced upon LM infection by secretion of IL-12 and IL-18 that induces non-Ag-specific IFN-γ secretion (see Fig. 3) (36). Approximately 50% of thymus-independent CD8αβ cells from a Txbm chimera secrete IFN-γ in response to LM-infected J774 cells. However, this secretion is blocked by including Abs to IL-12 and IL-18 in the culture, indicating that thymus-independent CD8 cells from DKO mice do not mount a specific primary immune response against this pathogen (Fig. 5,D). Approximately 70% of the CD8 cells from a thymus-intact bm chimera secreted IFN-γ in response to LM-infected J774 macrophages. Unlike the results with T cells from Txbm animals, a substantial proportion of CD8 T cells from the thymus-intact bm chimera secreted IFN-γ in response to LM-infected J774 cells even in the presence of the cytokine-blocking Abs (Fig. 5,E). These data are summarized in Fig. 5 F and indicate that LM-specific CD8 T cells in DKO mice are derived from a thymus-dependent pathway. Thymus-independent cells may only contribute to the innate response against LM infection by secreting IFN-γ in the presence of proinflammatory cytokines, such as IL-12 and IL-18.
Ability of CD8αβCD44high cells from DKO mice to mount an Ag-specific response to LM
As previously reported, CD8 T cells from naive DKO mice play an important role in clearing a primary LM infection (22). Although >50% of the CD8 T cells in DKO mice display a memory phenotype, whether these CD8+CD44high cells are able to initiate an Ag-specific response against infection is not known. To determine whether these cells participate in this response, CD8α cells were purified with CD8α beads and then sorted for the CD44high phenotype (Fig. 6, A–C). The sorted cells were then transferred into Thy1.1.DKO mice and challenged with LM. After 6 days, we detected a small population of donor (Thy1.2+/+) cells in the spleens of recipient animals, and almost all detectable Thy1.2+/+ donor cells were CD8αβ (Fig. 6,D). The cells were cultured with LM-infected J774 cells in vitro in the presence or the absence of cytokine-blocking Abs to determine the potential of both donor and recipient cells to mount a specific response. Neither donor CD8αβCD44high nor recipient CD8αβ cells secreted IFN-γ in response to uninfected J774 cells, as expected (Fig. 6,E). However, ∼80% secrete IFN-γ in response to LM-infected J774 cells. Inclusion of Abs against IL-12 and IL-18 still resulted in a strong response to LM from both donor and recipient cells. To demonstrate the Ag specificity of the anti-LM response, we cultured the splenocytes with two peptides derived from LM that are presented by M3, f-MIGWII and f-MIVTLF, as well as the Kb-restricted peptide, SIINFEKL, as a control. Both donor CD8CD44high and host cells responded to the M3-restricted peptides by secreting IFN-γ, as demonstrated by ICS (Fig. 6 F). These data indicate that CD8CD44high cells from naive DKO mice are able to initiate an Ag-specific response against LM infection. Thus, the non-class Ia-selected CD8αβCD44high cell subset from these animals has the ability to contribute to both innate and adaptive immune responses to LM infection.
In naive young B6 mice, ∼10–20% of CD8 T cells express memory cell markers (17). This proportion of memory phenotype cells is even higher in naive young DKO mice where >50% of CD8αβ T cells are CD44highCD122high. Whether these memory phenotype cells are conventional in that they have had prior exposure to extrinsic Ags is unclear. Although previous studies have provided several possible reasons for the appearance of these cells in both normal as well as DKO animals, their origin and function are still largely unknown.
One explanation for the appearance of memory cells in naive mice is that these animals have been exposed to environmental Ags derived from commensal organisms (35). However, memory phenotype CD4 cells are also found in naive mice maintained under germfree conditions, suggesting that their development is not foreign Ag dependent (51). Because T cells from male anti-H-Y TCR-transgenic mice can develop extrathymically and display memory markers, it is possible that the development of memory phenotype CD8 T cells in naive mice is driven by the interaction of the αβTCR with self-Ags and developed extrathymically (13, 18, 19). However, our finding that only a small percentage of CD8αβ cells develop in Txbm DKO mice compared with thymus-intact bm chimeras suggests that most non-class Ia-selected CD8αβ cells are from the thymus-dependent pathway.
The size of the lymphocyte pool is very important for the adaptive immune system. Homeostatic proliferation functions to maintain peripheral T cell numbers so that if there is a decrease in T cell numbers, naive cells undergo proliferation to fill up the space (52, 53). Naive cells that undergo homeostatic proliferation acquire a memory phenotype and function, in that they produce more IFN-γ than naive T cells in response to specific Ags (54). Neonatal lymphopenia can allow CD8+ thymic emigrants to undergo lymphopenia-induced proliferation during early neonatal life and thus equip the immune system with a set of preactivated CD8+ T cells before any infection. This might contribute to the rapid initiation of some immune responses in the adult (55). There are few CD8 cells in the periphery of DKO mice. It is possible that CD8 RTEs sense this space and undergo homeostatic proliferation to acquire the CD44high phenotype. This CD44high phenotype on CD8 T cells from naive mice is also characteristic of other MHC class I-deficient strains, such as TAP−/− and β2-microglobulin−/− mice, which have very few CD8 cells in their periphery (21). It has also been reported that under conditions of low single-positive cell output from the thymus that peripheral single-positive T cells can compensate by homeostatic expansion (43). Extensive cell apoptosis of CD8αβCD44high cells in DKO mice in vivo may contribute to the lack of CD8 cells in DKO mice (data not shown). Although there are data that suggest that there is a lymphoid compartment that regulates homeostatic proliferation that is not compartmentalized into CD4 and CD8 subsets, it remains possible that CD8 cells sense CD8 space and attempt to fill this compartment (56).
Some class Ib-restricted T cells selected on hemopoietic cells in the thymus display a memory phenotype and may account for this memory phenotype (15). However, we showed that the acquisition of the memory phenotype on most CD8 T cells occurs as a post-thymic event. There are several possible explanations for the acquisition of the memory phenotype in postthymic T cells from DKO mice. One possibility is that these non-class Ia-selected CD8 cells recognize nonpathogenic commensal bacteria and as a result become primed. A previous report demonstrated that animals housed in a conventional colony could mount CTL responses to H2-M3-restricted LM Ags without previous sensitization (57). In our studies, a small percentage of CD8 RTEs acquire the CD44high phenotype very rapidly after exit from the thymus. This small percentage of CD44high cells may represent the H2-M3-restricted cell population that recognized commensal Ags.
A previous study proposed that non-class Ia-selected CD8 T cells may be the early source of IFN-γ after administration of anti-CD3 in vivo (40). We noted that there are more CD44highCD8 cells from DKO mice that secrete IFN-γ at a rapid rate after anti-CD3 stimulation both in vitro (data not shown) and in vivo compared with those in B6 mice. We also observed that ∼80% of the CD8 cells from LM-infected DKO mice secrete IFN-γ after exposure to LM-infected J774 macrophages. This response is in part due to signaling by IL-12 and IL-18, because anti-IL-12 and IL-18 Abs greatly reduced the response. Approximately 20–30% of CD8CD44high cells from naive DKO mice are able to produce IFN-γ after culture with LM-infected J774 macrophages. In this study the secretion is completely blocked by anti-IL-12 and IL-18 (data not shown). Thus, these data suggest that non-class Ia-selected CD8CD44high cells are an early source of both Ag-specific and cytokine-induced IFN-γ secretion during infection and can potentially help bridge the gap between innate and adaptive immunity during intracellular bacteria infection. It is possible that these same non-class Ia-selected CD8CD44high cells play a role in normal B6 animals, although it is possible that this population in normal mice can be selected on either class Ia or non-class Ia Ags.
CD1d-restricted NK T cells that are specific for glycolipid Ags have a CD44high cell surface phenotype, are selected upon interaction with hemopoietic cells, and have been shown to exhibit rapid activation responses (58). The majority of NK T cells expresses a semi-invariant TCR composed of an invariant Vα14-Jα18 chain, associated preferentially with Vβ8.2 or Vβ7 chains in mice (59). However, we found that there was no preferential usage of particular Vβ chains by these non-class Ia-selected CD8CD44high cells that secrete IFN-γ rapidly after anti-CD3 stimulation (60, 61). Thus, non-class Ia-selected, IFN-γ-secreting, CD8αβCD44high cells are heterogeneous cell populations and do not belong to a specific cell subset. This characteristic might be beneficial to the host by mounting a rapid IFN-γ response to a wide range of Ags that can be recognized by conserved non-class Ia molecules.
Non-class Ia-selected CD8 cells provide adaptive immune protection against LM infection (22, 33). The H2-M3-restricted primary response is rapid and precedes that of the class Ia-restricted response (22, 33). Although it is conventionally thought that naive cells are the responding population to foreign Ags during the animals’ first encounter with a pathogen, it is possible that this memory phenotype population also has the potential to respond in an Ag-specific manner. We showed that the transfer of small numbers of naive CD8CD44high cells from DKO donors into recipient animals resulted in an expansion of the transferred cells in vivo that mounted a specific response mediated by CD8αβ cells that arise in a thymus-intact environment. It has been shown that H2-M3-restricted CD8+ T cells display a high degree of peptide cross-reactivity (62, 63). Taken together, this suggests that naive DKO mice contain a population of memory cells, sensitized by commensal organisms, which are specific for pathogenic epitopes, including those displayed by LM. These CD8 T cells could account for the rapid primary response observed.
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 National Institutes of Health Research Grant AI34930 (to J.F.).
Abbreviations used in this paper: LM, Listeria monocytogenes; bm, bone marrow; ICS, intracellular cytokine staining; RTE, recent thymic emigrant; Tx, thymectomized.