|Home | About | Journals | Submit | Contact Us | Français|
Determining the magnitude and kinetics, together with the phenotypic and functional characteristics of responding CD8 T cells, is critical for understanding the regulation of adaptive immunity as well as in evaluating vaccine candidates. Recent technical advances have allowed tracking of some CD8 T cells responding to infection, and a body of information now exists describing phenotypic changes that occur in CD8 T cells of known Ag-specificity during their activation, expansion, and memory generation in inbred mice. Here, we demonstrate that antigen but not inflammation driven changes in expression of CD11a and CD8α can be used to distinguish naïve from Ag-experienced (effector and memory) CD8 T cells after infection or vaccination. Interestingly and in contrast to inbred mice, tracking polyclonal CD8 T cell responses with this approach after bacterial and viral infections revealed substantial discordance in the magnitude and kinetics of CD8 T cell responses in outbred hosts. These data reveal limitations to the use of inbred mouse strains as preclinical models at vaccine development and suggest the same dose of infection or vaccination can lead to substantial differences in the magnitude and timing of Ag-specific CD8 expansion as well in differences in protective memory CD8 T cell numbers in outbred individuals. This concept has direct relevance to development of vaccines in outbred humans.
CD8 T cells recognize and respond to pathogen-encoded peptides displayed by MHC Class I molecules on infected or on Ag-presenting cells (APC). Activated CD8 T cells manifest an array of antimicrobial effector pathways and molecules, which eliminate infected cells by cytolysis or by recruiting and activating other immune cells(1, 2). Pathogen-specific CD8 T cell responses are important goal of vaccination and much attention has been devoted to understanding the properties of effector and memory antigen (Ag)-specific CD8 T cells in experimental models.
Primary CD8 T cell response can be divided into four relatively distinct phases. Initial activation (phase I) is critically dependent on relatively stable interactions between naïve T cells and mature, Ag and costimulatory molecule-expressing, dendritic cells (DC)(1–8). Upon Ag recognition, Ag-specific CD8 T cells undergo substantial proliferative expansion in numbers, acquire effector mechanisms, and disseminate throughout the body. The expansion phase is followed by a contraction (death) phase where 90–95% of Ag-specific CD8 T cells are eliminated. The final phase is the initiation and maintenance of the memory CD8 T cell pool by the cells surviving the contraction phase. Recent studies collectively indicate that the expansion, contraction, and memory phases of the CD8 T cell response are largely if not completely independent of continued Ag-display, suggesting that relatively brief encounter with Ag is sufficient to instigate the full program of CD8 T cell differentiation(9–15). However, in addition to Ag, a variety of signals (inflammatory cytokines) must be integrated by the responding CD8 T cells to generate optimal responses and ensure proper regulation of the CD8 T cell response(5, 7, 16).
Immunological techniques such as tetramers (or multimers) of pathogen-derived peptide-MHC molecules and peptide-induced cytokine secretion (ICS and ELISPOT assays) have enabled the quantification of Ag-specific CD8 T cell responses in various human and animal model systems of infection(17–23). Importantly, our ability to determine the magnitude and kinetic of Ag-specific CD8 T cell responses is dependent on a priori knowledge of specific epitope(s) (antigen) and MHC restriction element(s). Numerous studies with these techniques have enabled a comprehensive evaluation of the phenotypic changes exhibited by responding CD8 T cells. As a result, information describing phenotypic changes that occur in CD8 T cells of known antigen-specificity during their activation, expansion, and subsequent memory CD8 T cell generation is available(2, 24–29). This leads to the exciting possibility that all of the CD8 T cells responding to infection and/or immunization might exhibit similar phenotypic footprints that could be used to characterize the CD8 T cell response without a priori knowledge of specific epitopes or MHC molecules.
Here, we demonstrate that the increased expression of integrin molecule CD11a and downregulation of CD8α on CD8 T cells after various types of infections and/or antigen stimulations can distinguish naïve from antigen-experienced effector and memory CD8 T cells in both lymphoid and non-lymphoid tissues. Thus, these markers can be used to determine the magnitude and kinetics of most, if not all, antigen-specific CD8 T cell responses. In addition, while the magnitude and kinetics of polyclonal CD8 T cell responses in inbred mice were coordinated (albeit different between inbred strains), substantial variability was observed in outbred hosts.
Inbred BALB/c (Thy1.2), and C57BL/6 (B6; Thy1.2), and outbred Swiss Webster mice from obtained from the US National Cancer Institute (NCI) and B6 Thy1.1 mice were obtained from The Jackson Laboratory (JAX). OT-I transgenic Thy1.1 mice were previously described(30). Pathogen-infected mice were housed in the appropriate bio safety conditions. All mice were used at 6–12 weeks of age. All animal protocols followed approved Institutional Animal Care and Use Committee (ACURF) protocols.
The virulent Listeria monocytogenes (vir LM) strain 10403s and attenuated actA-deficient LM strains DP-L1942 and Ova257-expressing (Att LM and Att LM-Ova, respectively) were resistant to streptomycin and were grown, injected i.v., and quantified, as described(31). Colony Forming Units (CFU), as a measure of bacterial clearance after infection, were determined in the spleen and liver on indicated days after infection, as described(31). The Armstrong strain of LCMV was used as described(31).
Splenic dendritic cells (DC) were isolated after subcutaneous injection of B6 mice with 5×106 B16 cells expressing Flt3L (provided by M. Prlic and M. Bevan, University of Washington). When tumors were palpable (5 mm × 5 mm), mice were injected with 2 mg LPS (Sigma) i.v. to mature the DC. Spleens were harvested 16 hours later and were digested with DNase and Collagenase for 20 min at 37°C/7% CO2 with shaking. Spleen pieces were smashed through a nylon cell strainer (70μm) to generate a single cell suspension, RBC were lysed and splenocytes were resuspended in 2 parts of 10% FCS RPMI-1640 to one part B16-Flt3L conditioned medium + rGM-CSF (1000 m/ml) + 2 mM Ova257-264 and incubated 2 hours at 37°C/7% CO2. Spleen cells were washed three times and CD11c+ cells were isolated using anti-CD11c microbeads (Miltyni Biotec). The purity and activation status of DC were determined by staining for CD11c, CD86 and MHC class II. Routinely, greater than 90% pure CD11c+ DC were obtained. DCs were resuspended in saline and injected i.v. at 1×106 per mouse.
Purified naïve OT-I Thy1.1 cells (500 cells per recipient mouse(32)) were transferred into naïve B6 Thy1.2 mice and 1 day later mice were infected with indicated strains of LM or treated with CpG or Poly(I:C). Prior to tissue removal, samples of blood were obtained by retroorbital puncture. Anesthetized mice were then perfused through the left ventricle with PBS. Single cell suspension from liver, lung, spleen, and lymph nodes were washed prior to Thy1.1, CD8, and CD11a staining.
The following monoclonal antibodies with the indicated specificity and with appropriate combinations in fluorochromes were used: IFN-γ (clone XMG1.2, eBioscience), CD8 (53-6.7, eBioscience and BD), Thy1.1 (OX-7, eBioscience), Thy1.2 (53-2.1, BD), CD127 (A7R34, eBioscience), CD62L (MEL-14, eBioscience), CD43 (1B11, BD), CD44 (Pgp-1, eBioscience), CD11a (M17/4, eBioscience), and CD27 (LG.7F9, eBioscience), and isotype controls IgG2a, IgG2b, and IgG1 (clones eBR2a, KLH/G2b-1-2, eBRG1, respectively, eBioscience). Synthetic peptides, which represent defined Listeria monocytogenes LLO91-99, p60217-225, p60449-457 H-2d-restricted epitopes were used, as previously described(31).
Ova257-specific CD8 T cells were also detected by APC-conjugated tetramer complexes as described(31). In the most experiments total number of CD8 T cells responding to infection and/or immunization was determined in the blood or spleen by CD11a vs. CD8 co-staining. For determination of Ag-experienced CD8 T cells in the blood, small samples (~20 μl) were obtained from tail-tip snips at the indicated days post infection.
To test the notion that characteristic phenotypic changes can be used to differentiate recently activated antigen-specific CD8 T cells from naïve non-responding CD8 T cells, C57BL/6 (B6) mice were infected with a sublethal dose (0.1LD50) of attenuated (act A− deficient) recombinant Listeria monocytogenes (Att LM) that expresses the Ova257 epitope (Att LM-Ova(11, 31)) (Figure 1). At day 7 post infection Ova257-specific CD8 T cell responses were detected in the blood using Ova257-specific MHC Class I tetramers (KbOva). The expression of surface markers was analyzed on Ova257-specific and non-specific CD8 T cells (CD8+/KbOvapositive and CD8+/KbOvanegative, respectively) from the same LM-infected mice and compared to the CD8 T cells from non-infected age and sex matched naïve controls (Figure 1A,B). The majority of Ova257-specific effector CD8 T cells downregulated CD127 (IL-7Rα) and CD62L when compared to CD8 T cells from naïve mice (Figure 1B). In contrast, all of the Ova257-specific CD8 T cells expressed high surface levels of CD44, CD43 (detected by 1B11 mAb(33)), and CD11a. Therefore, as previously described(2, 24–29), substantial phenotypic changes were associated with the infection-induced naïve to effector Ag-specific CD8 T cell transition in vivo. Importantly, ~ 20% of CD8+/KbOvanegative T cells from infected mice showed phenotypic changes similar to the changes observed in effector CD8 T cells specific for Ova257 (Figure 1B) suggesting that the total number of CD8 T cells responding to LM infection is substantially higher than the number of Ova257-specific CD8 T cells. In addition, surface CD8α expression on Ova257-specific CD8 T cells is downregulated compared to the expression observed in naïve CD8 T cells (Figure 1A). The downregulation of CD8α is also observed on some non Ova-specific CD8 T cells present in infected mice (Figure 1B) suggesting that the level of CD8α expression in conjunction with upregulation of specific activation markers can be used to enumerate number of effector CD8 T cells responding to infection.
Taken together, these data suggest that Ag-specific CD8 T cells showed characteristic expression patterns of molecules that can be used to determine the magnitude of total effector CD8 T cells responding to infection. A similar approach was recently used to quantify the Ag-specific CD8 T cells early after infection or vaccination in mice and humans(34, 35). As shown here in Figure 1B, multiple phenotypic markers including CD8α, CD127, CD62L, CD43, CD44, and CD11a can distinguish Ag-specific effector CD8 T cells from naïve cells. However, an important unresolved issue is whether Ag-specific CD8 T cells retain phenotypic traits that can be used to quantify their numbers as they progress to memory. To address this, the phenotype of Ova257-specific CD8 T cells was compared to the rest of the CD8 T cells in the same mice 50 days after Att LM-Ova infection (Figure 1C,D). As previously shown in various models of infections and/or vaccinations(2, 24–29), most of the memory CD8 T cells upregulated CD127 and downregulated CD43 expression (Figure 2C). As expected, CD62L upregulation is observed on some but not all Ova257-specific CD8 T cells 50 days after infection(36, 37). Importantly, > 99% of Ova257-specific CD8 T cells showed sustained and elevated expression of CD44 and CD11a. CD11a high expressing cells are also observed in the tetramer negative CD8 T cells suggesting that antigen-experienced memory CD8 T cell responses of unknown specificity are present in the same mice (Figure 2D). Thus, results presented so far suggest that sustained and elevated expression of CD44 and CD11a on known Ag-specific CD8 T cells can be used to define and evaluate the magnitude and kinetics of Ag-experienced CD8 T cell responses after infection. Since a relatively high frequency of CD8 T cells from naïve mice express high levels of CD44 (Figure 1B) we will focus on changes in CD11a surface expression as a potential bio marker that can be used to track Ag-experienced CD8 T cells during the course of infection. Importantly, similar to the CD11ahigh effector Ag-specific CD8 T cells CD8α downregulation is sustained on Ag-experienced CD8 T cells as they progress to memory (Figure 1C,D) and should be used together with CD11a to follow the total CD8 T cells responding to antigenic challenge.
The full spectrum of antigenic epitopes have not been identified for most pathogens, with possible exception of LCMV and vaccinia viruses in B6 mice(34, 38–40). This knowledge gap prevents a complete enumeration of infection-induced CD8 T cell responses and it remains possible that Ag-independent, cytokine driven CD8 T cell expansion (bystander activation) still contributes significantly to the magnitude of CD8 T cell responses(41, 42). Therefore, it is critical to show that all (or the majority) of the host CD8 T cells expressing high levels of CD11a and low levels of CD8α are specific for antigens related to pathogen infection. In order to document the contribution (if any) of antigen-independent inflammation in changes in CD11a and CD8α expression on CD8 T cells, naïve B6 mice were treated with inflammatory cytokine inducing agents such as CpG and Poly(I:C)(31, 42, 43) (Figure 2A). As a positive control additional B6 mice were infected with Att LM. As shown in Figure 2A–C, CpG- or Poly(I:C)- induced inflammation in the absence of antigen(s) did not increase the frequency of CD11ahigh/CD8alow CD8 T cells in the blood (Figure 2A–C) or spleen (data not shown) at day 6 after treatment when compared to non-treated naïve mice. On the other hand, LM infection led to substantial increase in frequency of CD11ahigh/CD8alow CD8 T cells in the blood (17% of all blood cells compared to ~ 0.6% in all other groups of mice). Thus, inflammation in the absence of antigen stimulation did not increase the baseline numbers of polyclonal endogenous CD11ahigh/CD8alow CD8 T cells in naïve mice.
To further address this question and to fully mimic inflammation generated by infection, naïve Thy1.1 TCR-Tg OT-I (Ova257-specific) CD8 T cells were purified and transferred into groups of naïve B6 Thy1.2 mice one day before induction of inflammation in the presence (Att LM-Ova infection) or absence of antigen (treatment with CpG or Poly(I:C), or infection with LM that does not express Ova (Att LM)) (Figure 2D,E). As shown for the endogenous CD8 T cells CpG and/or Poly(I:C) treatments did not increase the frequency of OT-I TCR-Tg CD8 T cells that express high levels of CD11a and low levels of CD8α when compared to a non-treated group of naïve mice (none group) (Figure 2E). More importantly, changes in CD11a and CD8α expression on OT-I CD8 T cells was observed only in mice challenged with Listeria expressing the Ova antigen but not in mice that received the same infection with LM that do not express the antigen (Figure 2E). Taken together, these results show that inflammation in the absence of Ag is not sufficient to modulate the CD11a and CD8α expression on endogenous polyclonal and OT-I TCR-Tg CD8 T cells in naïve mice. Therefore, these data provide evidence that host CD8 T cells expressing high levels of CD11a and low levels of CD8α are specific for pathogen-derived antigens as well as evidence that those CD8 T cells that do not change the expression of CD11a and CD8α are not responding to pathogen-derived antigens.
Recently, we and others have shown that naïve to memory CD8 T cell progression is accelerated after peptide-DC immunization and that the rate of acquisition of memory CD8 T cell phenotype and function is controlled by inflammation present at the time of priming(5, 16). In order to address whether inflammation coupled with Ag (as observed after infection) is required for CD11a upregulation on responding Ag-specific CD8 T cells naïve B6 mice were immunized with Ova257-coated DCs and the expression of CD11a on Ova-specific CD8 T cells analyzed on day 7 after immunization (Figure 2F,G). Despite the priming in low inflammatory environment all of the Ova257-specific CD8 T cells showed changes in the expression of CD11a and CD8α after DC-Ova immunization. Therefore, upregulation of CD11a and downregulation of CD8α expression distinguish naïve from memory Ag-specific CD8 T cells after infection and after priming in a low inflammatory environment.
Some markers of effector Ag-specific CD8 T cells (ex. CD43high, CD62Llow, CD127low) are modulated in memory CD8 T cell populations. In contrast, the CD11ahigh/CD8alow phenotype is preserved at day 50 on all Ova257-specific CD8 T cells after LM infection (Figure 1D). In order to determine that the complete pathogen-specific memory CD8 T cell pool can be defined by high CD11a and low CD8α expression, naïve B6 Thy1.2 mice were immunized with Att LM and 35 or 90 days later CD8 T cells from the spleens were cell-sorted into CD11alow/CD8αhigh and CD11ahigh/CD8αlow subpopulations (Figure 3A). CD11alow CD8 T cells obtained from naïve (non-infected) mice were used as controls (Figure 3). Purity of the sorted populations was >96% for CD11alow and >93% for CD11ahigh before adoptive transfer into naïve Thy1.1 B6 mice (Figure 3B). Recipient mice were challenged with Att LM and the expansion in numbers of Thy1.2 CD8 T cells was determined in the blood and in the spleen at days 5 and 6 post infection (Figure 3C and data not shown). Expansion of Thy1.2 cells was not detected in Thy1.1 mice that received either 2.5×104 or 1.0×105 purified CD11alow CD8 T cells from immune (day 35 or day 90 post primary infection) or naïve mice after LM infection suggesting that the CD11alow CD8 T cell population from immune mice does not contain LM-specific memory CD8 T cells. Importantly, vigorous secondary expansion of Thy1.2 cells was observed in the blood and in the spleen of mice that received lower number (2.5×104 cells/mouse) of CD11ahigh CD8 T cells (Figure 3C). Thus, CD11ahigh/CD8αlow CD8 T cell population in immune mice consists of memory CD8 T cells that are able to respond to a secondary antigen encounter. In addition, these data suggest that CD11a and CD8α can be used as a markers to determine the number of memory CD8 T cells as well as to enrich polyclonal CD8 T cells that form a memory CD8 T cell pool after infection.
The magnitude of CD8 T cell responses of known specificity studied to date depends on the type of infection. In these studies a correlation between the dose of the infection and the level of Ag-specific CD8 T cell expansion has been described(1). Whereas some pathogens elicit robust endogenous CD8 T cell responses that are easily detectable by tetramers and cytokine secretion-based assays, endogenous responses are not always detectable for some pathogens or protein Ag, where the dominant epitopes may not be known. It is also important to note that the number of pathogen encoded proteins that might serve as targets for CD8 T cells by far exceeds the number of proteins that contain described CD8 T cell epitopes.
Listeria monocytogenes (LM) is a Gram positive, intracellular bacterium and multiple LM-derived CD8 T cell antigens (epitopes) have been described and used to probe CD8 T cell immunity in mouse model of infection(44). The entire genome of LM has been sequenced and the existence of more than 2,800 protein coding genes has been suggested(45). Since secreted as well as non-secreted LM proteins prime CD8 T cell responses(46) the number of CD8 T cells responding to LM infection should be substantially higher than the magnitude of CD8 T cell responses directed to known and defined antigens. For the two most often used strains of inbred mice (BALB/c (H-2d) and B6 (H-2b)) multiple LM-derived epitopes have been described. Immunodominant LLO91 as well as additional ‘subdominant’ epitopes (ex. p60217 and p60449) are readily detectable either by tetramer or peptide-stimulated intracellular cytokine staining (ICS) in LM-infected BALB/c mice(44). Interestingly, although numerous CD8 T cell epitopes are described in H-2b B6 mice based on highly sensitive ELISPOT assay(47) none of those can be reliably detected by ICS technique which has a limit of detection of ~ 5×103 cells per spleen (data not shown). Thus, we sought to use changes in CD11a and CD8α expression to determine the magnitude of CD8 T cells responding to LM infection. Naïve BALB/c and B6 mice were infected with 5×106 CFU (> 0.1 LD50; ‘high dose’) or 50-fold lower dose (1×105 CFU per mouse; ‘low dose’) of Att LM and their CD8 T cells analyzed in the blood and spleen on day 7 post infection (Figure 4A,B). After high dose challenge ~ 30% of total CD8 T cells (~ 5×106 cells/spleen) in BALB/c mice upregulated CD11a and downregulated CD8α and thus were specific for LM-derived antigens (Figure 4A,B). Among those, ~ 25% were specific for strongest LM-derived dominant and subdominant H-2d epitopes (LLO91, p60217, and p60449(44)) (data not shown) suggesting that the specificity of the ~75% of CD8 T cells responding to LM in BALB/c mice is currently unknown. Interestingly, despite the absence of strong and easily detectable LM-derived H-2b epitopes, the same dose of LM infection primed substantially higher frequency and total number of CD11ahigh/CD8αlow CD8 T cells in B6 mice when compared to BALB/c mice (Figure 4A,B). ~ 50% of all CD8 T cells (~ 10×106 cells/spleen) showed CD11ahigh/CD8αlow phenotype at day 7 post high dose LM infection of B6 mice. Importantly, the majority of CD11ahigh/CD8αlow CD8 T cells in B6 mice after LM infection showed a effector phenotype (CD127low, CD62Llow, CD44high) (data not shown) suggesting that changes in the expression of CD11a and CD8α can be used to enumerate the total number of effector CD8 T cells responding to infection. Finally, in accordance to recently published study in which CD8 T cell responses specific to all known LCMV-specific epitopes were analyzed(34), ~ 95% of all CD8 T cells (70–90×106 cells/spleen) upregulated CD11a and downregulated the CD8α molecules after LCMV infection in B6 mice (Figure 4A,B). Similar findings were observed in BALB/c mice although the magnitude (frequency and total number) of CD8 T cells responding to LCMV infection was slightly lower than observed in B6 mice (Figure 4A,B). These results suggest that complex bacterial pathogens such as LM can prime robust CD8 T cell responses (approaching 50% of all CD8 T cells in the blood and spleen) and that the magnitude of the expansion might be dependent on the strain of inbred mice used.
To this point we have shown that changes in the expression of CD11a/CD8α can be used to determine the magnitude of CD8 T cell expansion in the blood and spleen of LM and LCMV infected B6 mice. In order to determine if this strategy permits tracking of Ag-experienced CD8 T cells in tissues, naïve B6 mice were seeded with 500 Thy1.1 OT-I TCR-Tg cells prior to Att LM-Ova infection (Figure 5A). At day 7 p.i. the upregulation of CD11a and downregulation of CD8α on responding OT-I CD8 T cells was confirmed in the blood of infected mice (Figure 5B). Importantly, in all organs analyzed most, if not all, of OT-I CD8 T cells were CD11ahigh/CD8αlow (Figure 5C). Interestingly, most of the total CD8 T cell population detected in lung and liver (87 and 77%, respectively) showed increased expression of CD11a while most of the CD8 T cells in LN remained CD11a low (Figure 5C). Taken together, these results show that changes in CD11a and CD8α expression can be used to track antigen-experienced CD8 T cells in various organs and in some of them (ex. liver and lung) the majority of CD8 T cells were specific for bacterial pathogen.
Naive to memory CD8 T cell progression after various types of infections and/or DC immunization is accompanied with sustained changes of CD11a molecule on the surface of responding CD8αlow antigen-specific CD8 T cells (Figures 1–3). All of the data presented so far compare the CD11a and CD8α expression in naïve (non immunized) mice. A small frequency of CD8 T cells in naïve mice showed CD11ahigh/CD8αlow phenotype which make it easier to follow primary polyclonal CD8 T cell responses after antigenic challenge. However, potential translation of these findings to humans will require tracking CD8 T cell responses in individuals with a history of previous antigen encounters and substantially increased representation of CD11ahigh/CD8αlow CD8 T cells. Therefore, it is critical to determine if CD11a and CD8α can be used as a markers for following the kinetics of polyclonal CD8 T cell responses to newly introduced antigens in immune (previously infected) hosts. To address this, naïve B6 mice were infected with Armstrong strain of LCMV and the kinetic of CD11ahigh/CD8αlow CD8 T cell responses was determined in the blood (Figure 6). ~ 7 months after initial challenge those mice, together with naïve (non-infected) controls, were challenged with Att LM and CD11ahigh/CD8αlow expression in CD8 T cells in the blood was determined at the indicated days after infection. Importantly, a similar magnitudes and kinetics of primary LM-specific CD8 T cells responses were observed in naïve or LCMV immune mice that started with ~ 30% of their CD8 T cells expressing CD11ahigh/CD8αlow phenotype (Figure 6). Similar data were obtained in mice primary infected with Att LM and then challenged with LCMV (data not shown). Finally, the magnitude and kinetics of polyclonal CD8 T cell responses can be determined by CD11ahigh/CD8αlow phenotype on CD8 T cells in mice that were challenged multiple times with the same pathogen (homologous prime-boost infections; data not shown). Thus, these data suggests that CD11a and CD8α can be used as a reliable marker that can distinguish naïve from Ag-experienced CD8 T cells at any time point after antigen-encounter(s) as long as the frequency of CD11ahigh/CD8αlow CD8 T cells (‘baseline frequency’) is determined before the challenge.
A substantial body of evidence generated in inbred strains of mice suggested that Ag-specific CD8 T cell responses to various types of infections or immunizations might be programmed and follow similar pathways to expansion, contraction, and memory generation(1). Here we show that CD8 T cell responses can be analyzed without a priori knowledge of their antigen specificity suggesting the possibility that similar approach might be used to track polyclonal CD8 T cell responses in outbred mice after infection. Therefore, the CD11a and CD8α expression on CD8 T cells was determined in the blood of individual naïve B6 (inbred) and Swiss Webster (SW; outbred) mice before and after Att LM infection (Figure 7A,B). At the peak of the primary expansion (as determined in inbred strains of mice) the frequency of CD11ahigh/CD8αlow CD8 T cells was determined in the blood. Interestingly, in contrast to the results obtained in inbred B6 mice, substantial variation in the magnitude of the LM-specific CD8 T cell responses was observed in individual SW mice (Figure 7A,B). While the difference in frequency of CD11ahigh CD8 T cells between inbred B6 mice was 1.1 fold (defined as ‘inside group variability’), ~ 9-fold difference was observed between outbred mice at day 7 pi (Figure 7B). These data suggests that genetic background of the mice controls, at least in part, the magnitude of CD8 T cell expansion to infection.
Since anti-listerial immunity is mediated primarily by memory CD8 T cells and the level of protection is directly proportional to the number of memory CD8 T cells present at the time of re-challenge(1) one group of LM-infected SW mice were analyzed at the memory time point (day 46 post infection) to determine the frequency of CD11ahigh/CD8αlow CD8 T cells (Figure 7C) and their ability to protect after high-dose Listeria secondary challenge. In accordance to the previously published data where the number of memory CD8 T cells is proportional to the magnitude of the primary CD8 T cell expansion(1, 2) substantial differences in CD11ahigh/CD8αlow CD8 T cells were observed in individual LM-immune SW mice (Figure 7C). Importantly, all of the LM-immune outbred mice showed anti-listerial immunity in the spleen and the liver when compared to naïve controls after high-dose LM challenge (Figure 7D). However, the degree of protection (survival and > 98% reduction in bacterial numbers) directly correlated with the memory CD8 T cell numbers.
To extend these results, we examined the magnitude and kinetics of CD8 T cell responses to LCMV in B6 and SW outbred mice. As expected, the magnitude and kinetics of CD8 T cell response were similar in individual inbred B6 mice (Figure 8A,B). Substantial variability in the magnitude of the CD8 T cell response to LCMV challenge at day 8 post infection (peak of the anti-LCMV response in B6 mice(22)) is observed in outbred when compared to inbred mice (inside group variability - 1.04 compared to 3.87-fold) (Figure 8A). However, and in striking contrast to inbred mice, the high-resolution kinetic study of a subset of inbred and outbred mice revealed that the peak of the anti-viral CD8 T cell responses differs substantially between individual outbred mice (Figure 8B). Thus, these data suggest that the same dose of infection (or vaccination) might lead to substantial differences in magnitude and timing of Ag-specific CD8 T cell expansion as well in differences in protective memory CD8 T cells numbers in outbred individuals.
The hallmarks of CD8 T cell-mediated immunity are specificity and memory(1, 2). In recent years tremendous progress has been made in understanding the factors that can influence and regulate naïve to effector to memory CD8 T cell progression after infection and/or vaccination(1). This burst in knowledge was made possible with the development of technologies that can identify and track Ag-specific CD8 T cell responses. Virtually all of the data related to the Ag-specific CD8 T cell responses published to date were generated with inbred strains of mice. The advantage of using genetically identical siblings that are maintained through brother and sister mating is obvious since unlimited numbers of identical mice are available and the reproducibility of the experimental findings (inside a lab and among different labs) is elevated. On the other hand, while the differences (ex. in susceptibility or resistance to infection and predominance of Th1 or Th2 cytokine phenotypes(48–54)) observed among different inbred strains of mice have served as a tool for exploring the mechanisms that control particular biological processes, it is also a reminder that the genetic background of the mice can influence the outcome of the host-pathogen interaction.
Here, we show that CD8 T cell responses to infection or vaccination can be analyzed directly ex vivo in the absence of a priori knowledge of specific antigen(s). Changes in the surface expression of CD11a and CD8α can distinguish naïve from antigen-experienced CD8 T cells after antigenic stimulation. Importantly, antigen recognition by naïve CD8 T cells is required for CD11a upregulation and CD8α downregulation and inflammation in the absence of antigen did not modulate expression of both molecules. Finally, changes in the expression of CD11a and CD8α are stable on primary memory and on CD8 T cells responding to repeated antigen stimulations (homologous secondary and tertiary antigen challenges – data not shown) suggesting that all facets of Ag-specific CD8 T cell responses can be analyzed and determined without a priori knowledge of their antigen specificity.
It should be noted that our ability to clearly distinguish naïve from previously activated Ag-specific CD8 T cells relies on modulation of both CD11a and CD8α expression. Recently, it has been reported that antigen-driven CD8 downregulation on TCR-Tg OT-I cells is transient and that CD8α expression recovers while those cells progress to memory after LM or vaccinia infections(55). However, using a similar approach, but with low number of OT-I or P14 TCR-Tg CD8 T cells transferred, we were able to see distinguishable CD8a downregulation on OT-I and P14 CD8 T cells months after challenge with LM expressing Ova or GP33, respectively (data not shown). In addition, more than 200 days after LCMV infection the majority of endogenous LCMV-specific CD11ahigh CD8 T cells do express lower levels of CD8α (data not shown). Therefore, we would argue that changes in expression of CD11a and CD8α are sustainable on Ag-experienced CD8 T cells for long periods of time after infection and that, more importantly, those markers can be reliably used to determine the frequency of memory CD8 T cells for long periods of time.
The work by Hamilton and Jameson(56) suggested that the expression of CD11a can be modulated to some extent when naïve CD8 T cells were forced to undergo vigorous homeostatic expansion in a lymphopenic environment. The resulting ‘homeostatic memory’ CD8 T cell populations acquire some attributes of ‘true’ memory CD8 T cells (ex. ability to provide protection). It remains to be tested if homeostatic proliferation impairs our ability to detect total CD8 T cell responses to infection. The critical aspect of the method used here to track polyclonal CD8 T cell responses is to determine the number of CD11ahigh/CD8αlow CD8 T cells before and after antigenic challenge. Thus, we are suggesting that the majority (if not all) of CD8 T cells that upregulated CD11a and downregulated CD8α (over the measured ‘baseline’ (before challenge) levels) are indeed antigen (pathogen) specific when analyzed after infection in normal (non lymphophenic) environment.
The data obtained in inbred strains of mice showed that the CD8 T cell responses are coordinated after infection - reaching similar magnitude of the expansion with similar kinetics when analyzed longitudinally in individual inbred hosts. However, in direct comparison of CD8 T cell responses in two most commonly used inbred strains of mice – BALB/c (H-2d) and C57Bl/6 (H-2b), substantial difference in the magnitude of CD8 T cell expansion was observed after identical dose and type (bacteria or virus) of infections. Although none of the Listeria-derived H-2b epitopes can be detected with tetramers or ICS techniques, ~ 50% of all CD8 T cells in B6 mice were specific for Listeria at the peak of the CD8 T cell expansion. The same dose of LM infection primed less vigorous polyclonal CD8 T cell responses (~ 30% of CD8 T cells downregulated CD8α and showed increased CD11a expression) in H-2d BALB/c mice. Even in the situation when strong CD8 T cell epitopes are defined (BALB/c mice and LM infection) their contribution to the overall breadth of CD8 T cell response was only 25% of total CD8 T cells responding to challenge. Therefore, these data suggest that genetic background of the host can influence, at least in part, the CD8 T cell responses to infection and that substantial numbers of CD8 T cells responding to infection in inbred mice are of unknown antigen-specificity.
Interestingly, longitudinal analysis of infection-induced CD8 T cell responses in individual outbred hosts showed substantial variation in the magnitude of CD8 T cell responses. After Listeria infection substantial differences at the memory stage of CD8 T cell responses were also observed and mice that contained higher frequency of CD11ahigh/CD8αlow CD8 T cells were more resistant to high dose LM re-challenge. Therefore, the same type of challenge (vaccination) designed to evoke antigen-specific CD8 T cell responses in outbred hosts might fail to achieve protective memory CD8 T cell numbers in all individuals. In addition, data obtained in LCMV model of infection showed that in contrast to inbred mice substantial differences in kinetics of polyclonal CD8 T cell responses were observed in outbred mice. These data suggest that the total number of CD8 T cells responding to challenge (vaccination) can be assessed only after longitudinal analysis (multiple time points) of CD8 T cell compartment have been performed.
In summary, these studies established a model where polyclonal Ag-specific CD8 T cell responses to acute infections and/or vaccination can be analyzed in outbred populations without a priori knowledge of antigen-specificity and MHC Class I restriction. This model, together with functional assessment of responding CD8 T cell populations, could greatly enhance evaluation of adaptive CD8 T cell responses to infection or vaccination in outbred hosts and, if extended to humans, would likely speed up identification of vaccines that have efficacy for the greatest number of recipients.
We thank Jaime Sabel for technical assistance.
1Supported by start-up funds from Department of Pathology, University of Iowa (VPB) and NIH grants (AI83286 – VPB; AI42767, AI46653, AI50073, and AI59752 - JTH)