|Home | About | Journals | Submit | Contact Us | Français|
There is very limited evidence concerning the phenotype, function, and homing characteristics of memory T (TM) cells elicited by vaccination against intracellular bacteria in humans. Here we studied TM subsets elicited by exposure to Francisella tularensis in humans as a model of immunity to intracellular bacteria. To this end, TM cells were evaluated in 2 groups: (1) subjects immunized with live attenuated tularemia vaccine by skin scarification and (2) tularemia naturally-infected subjects. In both groups the immune responses were mediated by CD4+ and CD8+ effector TM cells, mostly CD45RA−CD62L− and CD45RA+CD62L−. Based on the expression of CD27, integrins α4/β7, and α4/β1, it is likely that some of these TM cells have lytic potential and the ability to enter both mucosal and non-mucosal sites. Thus, regardless of whether by immunization or natural exposure, tularemia antigens elicited a broad spectrum of specific TM subsets with diverse homing characteristics.
One of the hallmarks of immunity against intracellular pathogens is the induction of memory T (TM) cells. This process involves proliferation and differentiation of primed T cells into TM cells with the ability to home to a variety of tissues. Based on the expression of defined surface molecules, TM cells can be sub-divided into two main subsets: central memory T cells (TCM, CD45RA−CD62L+) and effector memory T cells (TEM, CD45RA−CD62L− or TEMRA, CD45RA+CD62L−) [1–3]. The TEM re-circulate primarily through non- lymphoid tissues, whereas TCM re-circulate primarily through lymph nodes, but both subsets are found in the blood . Both subsets are equally good at producing IFN-γ and TNF-α, while IL-2 appears to be produced exclusively by TCM [4, 5]. However, there is significant phenotypic heterogeneity within each subpopulation; CD8+ TEM in mouse and human express a stronger direct ex vivo lytic activity than TCM, whereas CD4+ TEM exhibit stronger effector cytokine production than TCM . The protective potential of each subset is also controversial; whereas studies in mice infected by Lymphocytic choriomeningitis virus (LCMV) show that TCM confer the more effective protective immunity , protection against Sendai virus infection in the lung  requires TEM. Moreover, when comparing TM subsets obtained after vaccination, striking differences were also observed. Studies in humans have shown that proliferating vaccinia-specific T cells display the characteristics of TEM decades after smallpox vaccination , while high rates of proliferating hepatitis B-specific T cells were found to be TCM . Although most of our knowledge of human memory-phenotype CD4+ and CD8+ T cells comes from analyses of lymphocytes from peripheral blood, the population dynamics and functional status of TM subsets in the human blood remain poorly characterized.
Here, we evaluated the PBMC from subjects that were exposed to tularemia antigens by two different routes, 1) immunization with tularemia live attenuated vaccine (F. tularensis Live Vaccine Strain, LVS) by skin scarification, and 2) naturally infected-subjects with primary pneumonic tularemia, as a model to study TM subsets in human blood. This model is particularly relevant because Francisella tularensis (F. tularensis), the causative agent of tularemia, is a CDC Category A threat organism of great public health importance. In addition, tularemia induces vigorous and long-lasting T-cell immune responses in humans. These responses are mediated by both CD4+ and CD8+ T cells and persist for up to 25 years, and are as vigorous as the responses to purified protein derivative (PPD) of Mycobacterium tuberculosis [10–12]. However, to our knowledge, the study of T memory subsets and their homing potential has not been explored in immunity to F. tularensis in either mice or humans. Specifically, we studied ex vivo IFN-γ-secreting and proliferating TM cells in human peripheral blood following LVS immunization by scarification, as well as in tularemia naturally infected-subjects via aerosol by examining their phenotype, function and homing potential characteristics. We found that regardless of the site of exposure, F. tularensis elicited a broad spectrum of specific CD4+ and CD8+ TM with diverse homing characteristics.
Five healthy adult LVS vaccinees (volunteers 34, 36, 37, 40 and 42), three F. tularensis naturally-infected subjects (F0715V1, F0717V1 and F0716V1) and two healthy adult controls (CVD4000#89 and CVD4000#76) participated in this study. The LVS vaccinees involved laboratory personnel at an increased risk of infection with F. tularensis. These subjects were recruited from the U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID), Frederick, Maryland. All vaccinees had positive “takes” (initial formation of small papule which ulcerates). Before blood collections, all volunteers were explained the purpose of this study and signed informed consents. The human experimentation guidelines of the US Department of Health and Human Services and those of the USAMRIID, were followed in the conduct of the present clinical research. The vaccinees donated blood before (day 0) and at days 15, 28 and 60 after immunization. The F. tularensis naturally-infected subjects, aged 45–60 years, were recruited at the Martha’s Vineyard Hospital at Oak Bluffs, Massachusetts. PBMC from these subjects were isolated at EpiVax, Inc., Providence, Rhode Island. Infection with F. tularensis was confirmed in all subjects by serology. Before blood collections, all subjects were explained the purpose of this study and signed informed consent. The human experimentation guidelines of the US Department of Health and Human Services and those of the EpiVax, Inc., were followed in the conduct of the present clinical research. They donated blood 2->25 years after being diagnosed with tularemia (subject FV0715V1: 2–5 years, subjects FV0716V1 and F0717V1: >25 years). The diagnosis was pneumonic tularemia in subjects FV0716V1 and F0717V1 and unknown in subject FV0715V1. PBMC from the CVD4000#89 and CVD4000#76 adult healthy volunteers were collected at the Center for Vaccine Development, U. Maryland, Baltimore, under the Blood Donor CVD 4000 protocol. The human experimentation guidelines of the US Department of Health and Human Services and those of the University of Maryland, Baltimore, were followed in obtaining these specimens. In all cases, PBMC were isolated from the blood of volunteers by density gradient centrifugation and cryopreserved in liquid N2 following standard techniques.
The heated and formalin-killed F. tularensis LVS strain antigenic preparation (NR-74) used in these studies was obtained through the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH. Briefly, the bacterial lawn was fixed in 3% formalin, heat-inactivated in a water bath at 60 °C for 30 min and washed 3 times with PBS. This antigenic preparation was used at a concentration of 107 CFU/ml (CFU determinations were performed before killing). This heated and formalin-killed LVS preparation is the same as the one currently being employed to assess immune responses in phase 1 and 2 clinical trials with a new formulation of the LVS vaccine. A Coxiella particulate antigenic preparation (provided by Baylor College of Medicine, Houston, Texas) and Staphylococcus enterotoxin B (SEB) (Sigma, St. Louis, MO) were used as negative and positive controls, respectively.
The cells were stimulated with LVS (107 CFU/ml concentration) antigen at 37 °C, 5% CO2. Cells stimulated with Coxiella (8.3 ng/ml) or SEB (Sigma, 10 μg/ml) were used as negative and positive controls, respectively. The culture medium consisted of RPMI (Gibco, Grand Island, New York) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamicin, 2 mM L-glutamine, 2.5 mM sodium pyruvate, 10 mM HEPES buffer, 1% non-essential aminoacids, and 2% heat-inactivated AB human serum.
The IFN-γ, IL-2 and TNF-α cytokine production by PBMC was measured using the flow cytometry-based BD Cytometric bead arrays (CBA) human kit assays (Becton-Dickinson – B-D, San Diego, CA) as previously described . Briefly, PBMC were stimulated with LVS and the supernatants harvested after 16–18 hours and kept at −70 °C until assayed. CBA assays were carried-out following the manufacturer’s instructions. Based on linearity, the levels of sensitivity for the various cytokines ranged from 2.5–10 pg/ml. The cut-off for CBA assays was established as the median of the background obtained with the cells stimulated with Coxiella. Results with PBMC obtained from vaccinees were scored positive if after background (Coxiella-stimulated cultures) subtraction the data from at least one point after immunization was greater than two fold compared with those before immunization.
The following mAb to surface molecules were used to stain PBMC: cutaneous lymphocyte antigen (CLA)(clone HECA-452), CCR4 (clone 1G1), CD3 (clone UCHT1), CD4 (clone SK3), CD8 (clone SK1), CD45RA (clone HI100) and CD62L (clone Dreg-56), all from B-D Pharmingen; CD14 (clone RMO52), CD19 (clone J4.119), CD27 (clone 1A4CD27) and CD49d (clone IM1404), all from Beckman-Coulter (Miami, FL); and integrin α4/β7 (clone Act-1)(kindly provided by W. Newman, Leukosite, Cambridge, MA) conjugated to Alexa 647 using an Alexa 647-labeling kit (Molecular probes, Eugene, OR). Monoclonal antibodies or their corresponding isotype controls conjugated to the following fluorochromes were used in these studies: FITC, ECD (Energy Coupled Dye), PE-Cy5, PE-Cy7, Alexa 647, allophycocyanin (APC)-Alexa 700, APC-Cy7, Pacific blue and biotin followed by streptavidin-Pacific Orange (Molecular Probes).
After 16–18 hours of stimulation with LVS, protein transport was blocked by adding BD GolgiStop (B-D Pharmingen) and the cells were incubated for an additional 5 hours at 37°C as previously described [13–15]. Cells were then stained with ethidium monoazide (EMA)(Molecular probes), followed by staining with monoclonal antibodies against surface antigens, fixation/permeabilization by using cytofix/cytoperm solution (B-D Pharmingen) and stained intracellularly for IFN-γ (B-D Pharmingen, clone B27). Cells were then resuspended in fixation buffer (1% formaldehyde) and analyzed as soon as possible by flow cytometry. PBMC from healthy subjects (CVD4000#76 or CVD4000#89) were used as internal controls in experiments using specimens from naturally-infected subjects.
The ability of PBMC to proliferate was measured by intracellular staining with a mAb to Ki67 (PE, clone B56)(B-D Pharmingen), a protein expressed in dividing cells . After 7 days of stimulation with LVS, cells were stained with EMA, followed by staining with mAb against surface antigens, fixation/permeabilization by using cytofix/cytoperm solution (B-D Pharmingen), and stained intracellularly for Ki67 as previously described . Cells were then resuspended in fixation buffer (1% formaldehyde) and analyzed as soon as possible by flow cytometry.
During sample acquisition, 100,000–500,000 events were collected in the forward and side scatter (FS/SS) lymphocyte gate. This large number of gated lymphocyte events was necessary to ensure that a sufficient number of positive cells for a defined subset would be collected for each tube analyzed. A “dump” channel was used to eliminate dead cells (EMA+), as well as macrophages/monocytes (CD14+) and B lymphocytes (CD19+) from analysis. Specimens were included in the analysis if (1) the cell viability was >50% after an overnight incubation (the viability of PBMC after overnight incubation ranged from 60 to 100% [median 91]), and (2) cells were shown to be functionally active as determined by the production of IFN-γ by at least 0.2% CD3+ cells after stimulation with SEB (the frequency of IFN-γ-producing cellsranged from 0.5 to 1.9% for total CD4+ [median 1.6] and from 5.1 to 18.4% for total CD8+ cells [median 9.7]) . The list-mode data files were analyzed using the Win-List 3D software (Verity Software House, Topsham, ME). A response was considered specific if the differential in the number of positive events between experimental (LVS) and negative control (Coxiella) cultures was significantly increased by Chi-square tests. Vaccinees were considered responders if the results from at least one time point after immunization were greater than 2.0 fold from those observed before immunization.
All tests were performed using SigmaStat software (version 3.10, SSPS Science software products, Chicago, IL). Comparisons between groups were performed using One-way ANOVA tests. P values <0.05 were considered significant.
Previous human studies showed that IFN-γ, IL-2 and TNF-α are secreted in response to inactivated F. tularensis [10, 19]. Thus, in the present studies we used IFN-γ, IL-2 and TNF-α production as surrogate markers for CMI responses. PBMC were stimulated with LVS (107 CFU/ml) for 16–18 hours. The LVS concentration of 107 CFU/ml was selected based on preliminary experiments showing that it induced the highest numbers of specific effector cells with the best viability, as indicated by their ability to exclude Trypan blue (data not shown). We selected 16–18 hours to minimize bystander stimulation while maintaining a close parallel with the in vivo cytokine response . After 16–18 hours, supernatants were collected and IFN-γ, IL-2 and TNF-α production analyzed by using flow cytometry-based B-D Cytometric bead arrays (CBA). The cut-off for specific responses to LVS in CBA assays was established as the median of the background obtained when cells were stimulated with the control antigen Coxiella (2.5, 38.7 and 2.5 pg/ml for IL-2, TNF-α and IFN-γ respectively). Vaccinees were considered responders if the results from at least one time point after immunization were greater than 2.0 fold above those observed before immunization. In agreement with previous studies [10, 19], we consistently observed increases in IFN-γ and TNF-α in all 5 subjects. However, substantial increases in IL-2 production were only detected in 3 subjects (Figure 1A). It is also important to note that there was considerable variation in the strength of responses among vaccinees but not among control subjects. This is not surprising since differences in the degree of responses among individuals is a common finding in human studies [13, 15, 21]. Finally, similar results were observed in F. tularensis naturally infected-subjects (Figure 1B and 1C). Having confirmed the presence of CMI responses in these specimens, we next focused our studies on the kinetics of induction of memory T cells.
To determine the kinetics of induction of CD4+ and CD8+ T cells implicated in the induction of the Th1 responses described above, we studied IFN-γ production by intracellular staining in PBMC stimulated with LVS 107 for 16–18 hours. Because cell frequencies were expected to be low and because particulate antigens, like the ones used in this study (LVS), are known to induce non-specific responses, the specificity and sensitivity of the flow cytometric assays were first carefully established. As described in methods, specific responses were defined as those in which the differential in the number of positive cell events by flow cytometry between experimental and unrelated particulate negative control (Coxiella) was significantly higher by Chi-square tests. Cell number is another important parameter to consider when low frequency is expected. To this end, we first determined the median percentage of IFN-γ-producing cell frequency of specimens stimulated with the control Coxiella antigen using PBMC from all 5 LVS vaccinees evaluated at 4 time points (n=20). These frequencies were found to be 0.03% and 0.07% for CD4+ and CD8+ cells, respectively (data not shown). This threshold was used to define the sensitivity of the test. Finally, as a third criteria, the results from at least one time point after immunization needed to be greater than 2.0 fold compared to those recorded before immunization. Similar parameters were also used to define positive proliferative responses as determined by Ki67 staining. The median Ki67+ cell frequencies of specimens (n=20) stimulated with control Coxiella antigen were 1.18% and 2.31% for CD4+ and CD8+ cells, respectively (data not shown). It is also important to note that in positive control cultures (PBMC incubated with SEB) the frequency of IFN-γ-producing cellsranged from 0.5 to 1.9% for total CD4+ (median 1.6) and from 5.1 to 18.4% for total CD8+ cells (median 9.7), and the frequency of Ki67-producing cellsranged from 2.1 to 9.0% for total CD4+ (median 5.8) and from 4.6 to 11.6% total CD8+ cells (median 10.1) (data not shown).
Following immunization and after background subtraction, increases in IFN-γ-secretion were recorded in 2 of the 5 vaccinees (Figure 2). In contrast to the relatively low number of vaccinees with IFN-γ-secreting cell responses, after 7 days of in vitro antigen exposure, tularemia-specific proliferative responses were detected in the PBMC of 4 of the 5 vaccinees. Of note, negative responses for both IFN-γ-secreting cells and proliferative responses were observed in 1 of the 5 vaccinees (vol#42). The simultaneous evaluation of both functions showed that (a) when IFN-γ-secreting cell responses were detectable, they were associated with proliferative responses [vol#34 and vol#36], and (b) 2 out of 5 vaccinees displayed only an expandable pool of memory T cells (i.e., proliferation in the absence of IFN-γ-secreting cell responses; vol#37 and vol#40). It is important to note that proliferative responses were somewhat lower among the group without IFN-γ-secreting cell responses. Both IFN-γ and proliferative responses were observed in CD4+ and CD8+ subsets, with the relative proportions of these responses differing among vaccinees. Peak responses were generally observed at days 28 and 60 after immunization (Figure 2). Taken together, these results suggest that IFN-γ-secreting TM cells and the expandable pool of TM cells had distinct patterns of long-term maintenance.
To investigate the relationship between the observed kinetic patterns in antigen-specific responses and changes in the frequency and phenotype of the TM cell subsets in peripheral blood, a detailed phenotypic analysis using 10-color flow cytometry was carried out in all individuals that responded to LVS at the peak of the response and then these measurements were compared with those recorded before immunization. Using the generally accepted definition to classify the T cells into the three main subsets, naïve (CD45RA+CD62L+), central memory T cells (TCM, CD45RA−CD62L+) and effector memory T cells (TEM, CD45RA−CD62L− or TEMRA, CD45RA+CD62L−) [1–3], we found that that among vaccinees, most IFN-γ-secreting CD4+ and CD8+ exhibited a phenotype consistent with TEM and TEMRA subsets, whereas only a few IFN-γ+ cells exhibited a phenotype consistent with naïve or TCM cells. A representative example of vaccinees’ responses is shown in Figure 3. To determine whether the distribution of these memory subsets was altered in our study population, we examined the ex vivo expression of CD45RA+ and CD62L+ on CD4+ and CD8+ T cell populations. After 16–18 hours of stimulation with Coxiella (negative control), no significant differences were observed in the % of CD4+ and CD8+ TM subsets among PBMC obtained before and after vaccination (Figure 4). Of note, these % were similar to those previously reported in healthy controls [22, 23].
We next investigated which cell subsets from LVS vaccinee’s specimens were preferentially expanded after 7 days of in vitro re-stimulation with LVS. In agreement with IFN-γ-secreting cell results, significant increases in proliferative CD4+ T cell responses were observed in TEM and TEMRA subsets. However, significant increases in proliferative CD8+ T cell responses were only detected in TEMRA subsets (Figure 5). Of note, the volunteers with lower proliferative responses and who did not exhibit increased IFN-γ-secreting cell responses (Vol#37 and Vol#40, Figure 2) exhibited frequencies of proliferating CD8+ TCM subsets comparable to those observed in CD8+ TEM and TEMRA subsets (data not shown). In contrast, the volunteers with high proliferative and IFN-γ responses (Vol#34 and Vol#36, Figure 2) exhibited higher frequencies of proliferating CD4+ and CD8+ TEM and TEMRA subsets than those observed in CD4+ and CD8+ TCM subsets, respectively (data not shown).
We next evaluated whether the increases observed in the TEM and TEMRA subsets in LVS vaccinees were also present in the F. tularensis naturally-infected subjects. Increases in IFN-γ-secreting cell responses were recorded in 2 of the 3 subjects examined (F0715V1 and F0716V1) (Figure 6). The remaining subject (F0717V1) was a non-responder when IFN-γ-secreting responses were evaluated by intracellular staining. As observed with LVS vaccinees, these responses were mediated mostly by TEM and TEMRA cells in both CD4+ and CD8+ subsets. No increases in IFN-γ secreting cells were observed in a healthy subject (CVD4000#76) included as a control in these studies. Unfortunately, due to the limited number of PBMC available from naturally-infected subjects, we were unable to perform proliferative response TM subset analysis in these subjects.
To investigate whether the rise of IFN-γ secreting TEM and TEMRA subsets was correlated to specific surface molecules involved in lymphocyte cytotoxicity or migration into secondary lymphoid tissues, we investigated the expression of CD27, CD49d, integrin α4/β7, CLA and CCR4 surface markers in these tularemia-specific TM cell subsets. Since studies have shown that the lack of CD27 expression levels generally correlate with cytotoxic activity [24, 25], we examined the lytic potential of IFN-γ-secreting CD4+ and CD8+ TEM subsets by measuring their ex vivo expression of CD27. We found that a substantial % of the specific (i.e., IFN-γ+) CD4 and CD8 TEM subsets from the PBMC of both, LVS vaccines and F. tularensis naturally infected subjects do not express CD27 (Figure 7). Moreover, a considerable proportion of these specific IFN-γ+ CD4+ and CD8+ TEM cell populations also expressed CD49d+ (an adhesion molecule that helps direct lymphocytes from circulation into tissues by strengthening lymphocyte adhesion to endothelial cells)  and integrin α4/β7+ (a molecule that mediates lymphocyte migration into intestinal lymphoid tissues) . Interestingly, skin-associated memory T cells as defined by expression of the cutaneous lymphocyte antigen (CLA) [28–30] and the chemokine receptor CCR4 [31, 32], were not observed in either CD4 and/or CD8 TEM subsets from the PBMC of LVS vaccines (Figure 7) or F. tularensis naturally infected subjects (Figure 8).
Similar trends appear to be present in specific IFN-γ+ CD4 and CD8 TEMRA subsets; however, these results are not definitive because of (1) the wide variations in CD49d, integrin α4/β7, CLA and CCR4 expression levels among the TEMRA subsets in PBMC of the subjects evaluated, and (2) the lower numbers of IFN-γ secreting cells in TEMRA subsets, which typically represent a small proportion of cells among PBMC, these results are not definitive. Studies with additional volunteers should establish the validity of these results.
We also investigated F. tularensis-specific proliferating T cells for the expression of molecules required for lymphocyte homing. To this end, we re-stimulated the lymphocytes for 7 days in presence or absence of LVS antigen. After 7 days, the presence of proliferating lymphocytes was measured by intracellular staining with a mAb to Ki67, a protein expressed in dividing cells  in conjunction with the same surface markers used to identify IFN-γ secreting TM subsets (CD27, CD49d, integrin α4/β7, CLA and CCR4). Ten-color flow cytometric analysis showed that in vitro expanded TEM and TEMRA subset cells contained a greater proportion of cells expressing CD49d but lower expression of CD27. A variety of increases were observed in the percentages of CD4+ and CD8+ TEM and TEMRA subsets expressing integrin α4/β7, CLA or CCR4 after in vitro re-stimulation. Representative examples of proliferative responses are shown in Figure 9 (LVS vaccinee #34). Overall, these results demonstrate the heterogeneity of these TM subsets and highlight differences in the homing potential of specific cells that produce cytokines or proliferate in response to F. tularensis antigens.
In this study we evaluated PBMC from subjects who were exposed to tularemia antigens by two different routes, i.e., (1) immunization with the tularemia live attenuated vaccine (F. tularensis Live Vaccine Strain, LVS) by skin scarification, and (2) natural infection with F. tularensis, as models to study the induction of TM subsets in human blood. We observed that the immune responses elicited by LVS vaccination, as well as by natural infection were mediated mostly by TEM and TEMRA cells in both CD4+ and CD8+ subsets. These specific cells were able to produce IFN-γ and proliferate. Remarkably, two of the three naturally infected subjects that showed CMI responses to F. tularensis antigens had been infected >25 years before PBMC were collected (#F0716V1 and #F0717V1). Although we can not exclude that volunteers were subsequently exposed to F. tularensis after their initial infection, these results suggest that CMI to F. tularensis persist for extended periods of time. These results further support previous observations demonstrating long term CMI responses to F. tularensis antigens in humans [10–12]. It is worth noting that IFN-γ responses may be more easily detected by CBA than by flow cytometry. Using CBA, positive IFN-γ responses were detected in 5/5 vaccinees and in 3/3 naturally infected subjects; whereas using flow cytometry, IFN-γ responses were detected in only 2/5 vaccinees and 2/3 naturally-infected subjects. These differences could be due to differences in the levels of sensitivity or the timing of both assays. CBA measures the cumulative levels of cytokines in culture supernatants over 16–18 hours, whilst intracellular staining measures cumulative intracellular levels during the last 5 hours in culture. Interestingly, proliferative responses were relatively lower among the vaccinees without IFN-γ-secretion detected by flow cytometry. These results argue in favor of the possibility that multifunctional cells able to secrete high levels of cytokines, as well as proliferate, may play an important role in the host’s response to tularemia infection. In this regard, previous human study have demonstrated a relationship between the quality of functional T cell responses and the AIDS disease course . This study showed that non-progressors preferentially maintain HIV-specific T cells that are able to mount multiple cytokine/chemokine and degradulation responses on a cell-by-cell basis. Similar results were observed in a human vaccinia virus immunization model . It is also important to highlight that significant increases in proliferative CD8+ T cell responses in PBMC of LVS vaccinees were only detected in TEMRA subsets (Figure 5), indicating either that in vitro stimulation with formalized-LVS is not ideal for optimal CD8+ TEM cell stimulation or that CD4+ TEM cells remain in circulation for longer periods of time. It should be emphasized that proliferative responses may reflect the loss of some T cell subsets (e.g., by apoptosis death, loss of specificity) and their replacement by others able to better expand in vitro in response to antigen, a phenomenon frequently observed in T cell cultures.
Based on the expression of CD27, integrin α4/β7 and integrin α4β1, some of these memory T cells have lytic potential and the ability to enter intestinal mucosal and non-mucosal sites such as the lung. Interestingly, skin-associated memory T cells as defined by expression of the CLA and the chemokine receptor CCR4, were not observed in IFN-γ-secreting cells from LVS vaccinees or from F. tularensis naturally-infected subjects. Of note, a considerable proportion of cells from LVS vaccinees that were found to proliferate to F. tularensis antigens (Fig. 9), expressed CLA and CCR4, which presumably endow them with the potential to home to the skin. We hypothesize that cells that proliferate in response to F. tularensis antigen might represent a heterogeneous population of cells, some of which are endowed with the capability to home to the skin but are not capable of secreting IFN-γ. However, we can not exclude the possibility that the up-regulation of CCR4 and CLA markers in Ki67+ cells represent the selective expansion of a few T cell clones that cross-react with the F. tularensis particulate antigen used in these studies, a phenomenon unlikely to be observed in short incubation periods such as our culture conditions to measure intracellular IFN-γ production. It is widely accepted that a short period of incubation minimizes bystander stimulation. Taken together, these observations appear to indicate that regardless of the site of exposure, F. tularensis elicited a broad spectrum of specific cells with diverse homing characteristics. These results support the contention raised by others that it is possible that there are homing characteristics shared between mucosal and non-mucosal sites . Moreover, recent studies have demonstrated plasticity in the memory T cell homing potential. Migrating memory cells may modify their migration behavior depending on their location at the time of the antigen reencounter . It should be emphasized that, although cells and phenotypes may be present and may even increase in number or functional potential following exposure to F. tularensis antigens, these increases by themselves do not prove that they play a role in protection. However, since human mucosal biopsies are extremely difficult to obtain, the use of peripheral blood cells constitutes a reasonable alternative approach that allowed us to observe increases in defined TM cell subsets that might be directly relevant for human resistance to infection.
In summary, we have shown that TEM and TEMRA cells may play an important role in the host immune response against F. tularensis and that both specific CD4+ and CD8+ subsets are elicited following in vivo exposure to this organism. An unexpected observation was that these effector cells are heterogeneous in their homing characteristics, whether these measurements were performed in PBMC from LVS vaccines or naturally infected subjects. These effector cells may concomitantly express homing markers for mucosal and non-mucosal sites. This observation contributes important information for the development of vaccines that rely on the induction of TM cells.
We are indebted to the volunteers who allowed us to perform this study. We thank Dr. Julie McMurry (EpiVax, Inc.) for providing PBMC specimens from naturally infected-individuals as well as for helping in the discussion of this manuscript. We also thank Dr. Bernadette McConnell and the staff from the Blood Bank of the Maryland University hospital for their help in collecting blood specimens; Mr. Guillermo Sahaniuk and Ms. Regina Harley for excellent technical assistance.
This work was supported, in part, by NIAID, NIH, DHHS federal research contract NO1 AI30028 (Immunology Research Unit (IRU) of the Food and Water Borne Diseases Integrated Research Network (FWD-IRN)) to M. Sztein.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.