The clonal expansion and anatomic location of microbe-specific CD4+ Th cells was studied by tracking the fate of adoptively transferred DO11.10 TCR transgenic T cells specific for OVA peptide 323–339/I-Ad in BALB/c mice infected s.c. with Escherichia coli expressing a MalE-OVA fusion protein. After infection, the DO11.10 T cells accumulated in the T cell-rich paracortical regions of the draining lymph nodes, proliferated there for several days, and then moved into the B cell-rich follicles before they slowly disappeared from the lymph nodes. These changes occurred despite the fact that viable organisms were never found in the lymph nodes. The DO11.10 T cells also accumulated in the s.c. infection site, but about 1 day later than in the draining lymph nodes. Injection of purified MalE-OVA fusion protein alone induced a transient accumulation of DO11.10 T cells in the paracortical regions, but these T cells never entered follicles and the mice did not produce anti-OVA antibodies. The DO11.10 T cells that survived in animals injected with MalE-OVA alone were hyporesponsive to in vitro Ag restimulation and did not produce IL-2 and IFN-γ, whereas DO11.10 T cells from mice infected with MalE-OVA-expressing bacteria produced both lymphokines. These results suggest that Ag-specific T cells are first activated in secondary lymphoid organs following primary bacterial infection and then migrate to the infection site. Furthermore, productive activation of the T cells during the primary response is dependent on bacterial components other than the Ag itself.
CD4+ Th cells are essential for the elimination of microbes during primary infections and for the development of long-term immunity (1). Following primary infection, it is thought that APC within secondary lymphoid tissues internalize microbes, process microbial proteins, and present microbial peptide-MHC complexes on their surfaces (2). Specific T cells are then thought to interact with these APC and become activated via TCR recognition of the microbial peptide-MHC complexes. Once activated, the T cells proliferate in the secondary lymphoid tissues and some of the clonal progeny are thought to migrate to the nonlymphoid site of infection and facilitate clearance of the microbe (3). The end result of this primary response is an expanded population of memory T cells capable of orchestrating the protective secondary immune response via rapid and enhanced production of lymphokines that regulate macrophage, B cell, and cytolytic T cell function (1).
The aforementioned scenario is based in large part on indirect evidence because it has not been feasible to physically trace the in vivo behavior of the few T cells that are specific for the peptide-MHC complexes from a given microbe. One limitation has been the lack of tracking methods with sufficient specificity and sensitivity to physically identify and detect these T cells in the vast sea of T cells with other specificities (4). Another problem has been that the current model of the primary T cell response is based largely on the study of responses to model Ags, not microbial infections (5). We previously addressed the first limitation by developing a model in which a small number of CD4+, OVA peptide 323–339-I-Ad-specific T cells from the DO11.10 TCR transgenic mouse line can be tracked in lymphoid tissues following adoptive transfer into normal BALB/c recipients (6). Here we use this system to study the more physiologic immune response induced by infection of adoptive transfer recipient mice with a recombinant Escherichia coli strain expressing MalE-OVA. Our results provide direct support for the model of T cell activation described above and demonstrate that bacteria have adjuvant properties that are necessary of the generation of functional memory T cells.
Materials and Methods
Bacterial strains and plasmids
All strains used in this study were produced by transformation of E. coli strain DH5α (GIBCO/BRL, Grand Island, NY) with plasmids containing an ampicillin-resistance gene. The pMAL-p2 plasmid was purchased from New England Biolabs (Beverly, MA). The pAC-Neo-OVA plasmid (7) was kindly provided by Dr. Michael Bevan (University of Washington, Seattle, WA). The construction of plasmids used for this study is described below.
Construction of a MalE-OVA fusion gene
The pAc-Neo-OVA plasmid was used as a template for PCR amplification of a segment of the chicken OVA gene. The forward primer for the PCR was 5′-GAATCTAGAGCAGAGAGCCTGAAG-3′ that encodes an XbaI restriction site and amino acids 319 to 324 of chicken OVA, and the reverse primer was 5′-GCCGGATCCTTAAGGGGAAACACAT-3′ (Integrated DNA Technologies, Coralville, IA) that is complementary to the sequence that encodes residues 403 to 406 of chicken OVA plus 13 additional nucleotides. The PCR product DNA was then purified and ligated into the cloning site of the pCR II vector (a TA cloning vector) (Invitragen, Carlsbad, CA) by taking advantage of the A-overhangs present at the ends of the PCR product and the T-overhangs present in the vector. Digestion of this plasmid with XbaI and HindIII produced a DNA fragment that was purified and ligated into the multiple clone site of the pMAL-p2 expression vector, which had been digested with the same pair of restriction enzymes. The resulting plasmid (pMAL-OVA) encodes a fusion protein with MalE at its N terminus and the last 87 amino acids of OVA at the C terminus. This plasmid, or the control pMAL-p2 plasmid (that encodes MalE alone), were introduced into E. coli DH5α by the calcium chloride method (8).
Purification of MalE-OVA
Recombinant E. coli DH5α containing pMAL-OVA or pMAL-p2 were grown in Luria-Bertani broth supplemented with 100 μg/ml ampicillin and 2 mg/ml glucose overnight at 37°C. The overnight culture was diluted 1:100 into fresh Luria-Bertani broth supplemented with ampicillin and glucose and grown at 37°C to 2 × 108 cells/ml. Isopropyl β-d-thiogalactoside (IPTG3; Sigma, St. Louis, MO) was added to the culture to a final concentration of 0.3 mM. The culture was incubated for an additional 2.5 h before the bacteria were harvested by centrifugation at 4000 × g for 20 min. Periplasmic proteins were prepared by the osmotic shock method, and the MalE-OVA fusion protein was enriched on an amylose affinity column according to the protocol provided by the manufacturer (New England Biolabs, Beverly, MA). Residual endotoxin was removed by incubating the eluate from the affinity column with END-X B15 beads (Associates of Cape Cod, Woods Hole, MA). The final endotoxin concentration was determined to be less than 50 ng/ml by the Limulus amebocye lysate pyrochrome assay (Associates of Cape Cod). The purified MalE-OVA protein was then quantitated using a Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA).
Bacteria were induced with IPTG as described above. One million cells were lysed in 2× sample buffer (6% Tris-Cl, pH 6.8, 20% glycerol, 4% SDS, 2% 2-ME), and the lysates were separated on a 12% SDS polyacrylamide gel. Proteins were then transferred to nylon membranes, which were incubated overnight with blocking buffer (PBS with 0.05% Tween 20 and 5% BSA) at 4°C. The membranes were washed with 0.05% Tween 20 in PBS and incubated with a 1:2500 dilution of anti-OVA rabbit serum (Sigma Chemicals) at room temperature for 2 h with continuous shaking. Horseradish peroxidase-conjugated goat anti-rabbit serum (Sigma Chemicals, used at 1:10,000 dilution) was used as secondary Ab. The blot was developed using ECL reagents (Amersham International, Little Chalfont, U.K.)
Mice and adoptive transfer
BALB/c mice were purchased from the National Cancer Institute (Frederick, MD). The DO11.10 TCR transgenic mice (9) were bred in a specific pathogen-free facility according to National Institutes of Health guidelines and screened for transgene expression as previously described (6). These mice have been extensively backcrossed (>15 generations) onto the BALB/c background and are therefore histocompatible with normal BALB/c mice. Adoptive transfer was performed as previously described (6). Briefly, pooled splenocytes and lymph node cells from DO11.10 donors were depleted of CD8+ T cells by Ab plus complement treatment, and the percentage of CD4+, KJ1-26+ T cells present in the surviving population was determined in an aliquot of the cells by two-color flow cytometry as described below. Cells from an unstained aliquot were then injected into unirradiated BALB/c mice such that each mouse received 2.5 × 106 CD4+, KJ1-26+ T cells.
E. coli containing the pMAL-OVA or pMAL-p2 plasmids were induced in vitro with IPTG as described above. Bacterial cells (108) were injected in 0.15 ml of PBS s.c. into three sites on the back of each recipient mouse. In some cases, heat-inactivated bacteria (prepared by incubating bacteria in an 80°C water bath for 20 min) were used. Purified MalE-OVA fusion protein (100 μg/mouse) was also delivered in 0.15 ml of PBS via s.c. injection into three sites on the back. In some experiments an air sac was generated as described by Boyle et al. (10) on the backs of mice by s.c. injection of 0.9 ml of air and 0.1 ml of PBS containing the recombinant bacteria. Where indicated, 5-bromodeoxyuridine (BrdU) was offered to mice in the drinking water (0.8 mg/ml) ad libitum.
Flow cytometric analysis of lymph node cells
Axillary, brachial, and inguinal lymph nodes were harvested at various times after Ag injection, and the transferred DO11.10 T cells were identified by two-color flow cytometry as previously described (6). Briefly, 2 × 106 lymph node cells were incubated on ice with phycoerythrin (PE)-labeled anti-CD4 mAb (PharMingen, San Diego, CA) and biotinylated KJ1-26 mAb, which recognizes the DO11.10 TCR and no other (11). Cells were then washed and incubated with FITC-labeled streptavidin (SA; Caltag, South San Francisco, CA) to detect the KJ1-26 mAb. Twenty-thousand events were then collected for each sample on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA) and analyzed using Lysis II software (Becton Dickinson). DO11.10 cells were identified as CD4+, KJ1-26+ cells. The total number of DO11.10 T cells present at any given time was calculated by multiplying the total number of viable lymph node cells (obtained by counting viable cells) by the percentage of CD4+, KJ1-26+ cells obtained by flow cytometry.
Cell surface markers and BrdU- or propidium iodide-labeled DNA were detected by three-color flow cytometry. Surface markers were detected in cells stained with anti-CD4-PE mAb, biotinylated KJ1-26 mAb, SA-CyChrome, and either FITC-labeled anti-LFA-1, l-selectin, or isotype-matched control mAb (all purchased from PharMingen). BrdU-labeled DNA was detected in lymph node cells stained with anti-CD4-PE mAb, biotinylated KJ1-26 mAb and SA-CyChrome, incubated with 1% paraformaldehyde and 0.01% Tween 20, and stained with FITC-labeled anti-BrdU mAb as described by Carayon and Bord (12). Propidium iodide-labeled DNA was detected in cells stained with anti-CD4-PE mAb, biotinylated KJ1-26 mAb and SA-CyChrome, fixed with 70% ethanol, and stained with propidium iodide. In each case, the CD4+, KJ1-26+ cells were identified on a FACScan flow cytometer, and the fluorescence intensities of 1000 of these cells in the third color were measured.
Brachial lymph nodes were flash frozen in Oct embedding media (Miles, Elkhart, IN). Thin sections (6 μm) were cut on a refrigerated microtome and fixed to glass slides with acetone. Peroxidase-labeled KJ1-26 or anti-B220 mAb were used to detect DO11.10 T cells or B cells, respectively, as previously described (13).
In vitro culture
Lymph node cells (2 × 105) were cultured with 5 × 105 irradiated (3000 rads) BALB/c splenocytes and 5 μM OVA peptide 323–339 in 0.2 ml of complete Eagle’s Hanks’ amino acids medium (Biofluids, Rockville, MD) supplemented with 10% FCS, 2 mM glutamine, 100 U/ml penicillin, 100 U/ml streptomycin, 20 mg/ml gentamicin sulfate, and 5 × 10−5 M 2-ME for 60 h, and then the last 12 h in the presence of 1 μCi of [3H]thymidine. Alternatively, 2 × 104 cells from a short-term T cell line derived from DO11.10 mice were cultured for 72 h with heat-killed bacteria and 104 irradiated BALB/c peritoneal exudate cells as APC, the last 8 h in the presence of [3H]thymidine. The cultures were harvested at the end of the culture period and [3H]thymidine incorporation in DNA was measured by liquid scintillation counting.
Detection of IL-2 and IFN-γ by ELISA
IL-2 and IFN-γ in culture supernatants were measured by capture ELISA using noncompeting mAb pairs according to the protocol provided by the Ab supplier (PharMingen). To measure serum anti-OVA IgG levels, serial dilutions of serum samples were incubated for 2 h at 37°C in EIA/RIA plates (Costar, Cambridge, MA) coated with OVA (20 μg/ml OVA for 3 h and blocked with 1% BSA for 1 h at 37°C). The plates were washed with 0.5% Tween 20 in PBS, and the bound Abs were detected with a 1:5000 dilution of horseradish peroxidase-conjugated goat anti-mouse IgG(H+L) (Pierce, Rockford, IL). O-phenylenediamine (Sigma) was used as the chromagen. The optical densities at 490 nm were determined, and the titer at half of the maximal optical density was calculated for each sample.
Recombinant E. coli expressing a MalE-OVA fusion protein stimulate DO11.10 T cells in vitro
A recombinant E. coli strain was designed to express a fusion protein containing E. coli protein MalE at its N terminus and amino acids 319 to 406 of chicken OVA at its C terminus (Fig. 1,A), under the control of the Ptac promoter. The expression of the fusion protein was confirmed by immunoblot analysis of lysates from IPTG-induced bacteria probed with a polyclonal antiserum containing anti-chicken OVA Abs (Fig. 1,B). Based on densitometric comparison between immunoblots performed with bacterial lysates and purified OVA, the amount of fusion protein in the IPTG-induced recombinant bacteria was estimated to be about 1 μg OVA per 106 bacterial cells. To test whether a peptide-I-Ad complex that could be recognized by DO11.10 T cells was produced from the MalE-OVA fusion protein by Ag processing, peritoneal macrophages were exposed to IPTG-induced recombinant bacteria and then tested in vitro for stimulation of DO11.10 T cells. As shown in Figure 1 C, DO11.10 T cells proliferated in vitro in response to macrophages pulsed with bacteria expressing MalE-OVA but not MalE alone. Furthermore, the proliferation of DO11.10 T cells in response to MalE-OVA-expressing bacteria was blocked by addition of anti-I-Ad, but not anti-I-Ed mAb (data not shown). Together, the results demonstrate that a peptide-I-Ad complex capable of stimulating DO11.10 T cells was generated by APC exposed to recombinant E. coli expressing the MalE-OVA fusion protein.
Characterization of the primary DO11.10 T cell response induced in vivo by E. coli expressing the MalE-OVA fusion protein
The capacity of recombinant E. coli expressing the MalE-OVA fusion protein to stimulate OVA-specific T cells in vivo was assessed with the adoptive transfer method we used previously to characterize the primary T cell response induced by soluble OVA (6). A small number of CD4+, OVA peptide 323–339-I-Ad-specific T cells from the DO11.10 TCR transgenic mouse line were transferred into normal BALB/c recipients. Flow cytometry was then used to identify the transferred T cells following staining of recipient lymph node cells with the KJ1-26 anti-clonotypic mAb, which uniquely binds to the DO11.10 TCR. CD4+, KJ1-26+, DO11.10 T cells were detected in the lymphoid tissues of BALB/c mice previously injected with DO11.10 T cells (Fig. 2,B) but not in normal BALB/c mice that did not receive DO11.10 T cells (Fig. 2,A). A s.c. injection of 108 live E. coli expressing the MalE-OVA fusion protein caused the DO11.10 T cells to become more numerous in the draining lymph nodes (Fig. 2,C) compared with injection of nothing (Fig. 2,B) or control E. coli expressing MalE alone (Fig. 2,D). Analysis of the kinetics of this response revealed that as early as day 3 postinfection, the number of DO11.10 T cells in the draining lymph nodes was significantly higher in mice infected with MalE-OVA-expressing bacteria than in mice infected with control bacteria (Fig. 3). The number of DO11.10 T cells in mice infected with MalE-OVA-expressing bacteria continued to increase to a peak on day 5 and then declined slightly by day 7, and dramatically by day 17.
The majority of the DO11.10 T cells (identified as shown in Fig. 4,A) in mice infected 2 days before with MalE-OVA-expressing E. coli were blasts (Fig. 4,B). In addition, many of the DO11.10 T cells recovered from the lymph nodes of mice infected with MalE-OVA-expressing bacteria possessed the surface phenotype indicative of prior activation (14): l-selectinlow, LFA-1high (Fig. 4, C and D); whereas DO11.10 T cells recovered from mice infected with control bacteria had the surface phenotype of naive T cells: l-selectinhigh, LFA-1low. On day 3 postinfection, significantly more DO11.10 cells were in the S, G2, or M phases of the cell cycle in the lymph nodes of mice infected with MalE-OVA-expressing bacteria than in comparable nodes of mice infected with control bacteria (Fig. 5,A), indicating that the DO11.10 T cells in the former group were proliferating in vivo. This interpretation was supported by BrdU-labeling studies which showed that most of the DO11.10 T cells incorporated BrdU in vivo during the first 5 days after infection with MalE-OVA-expressing bacteria (Fig. 5 B). Taken together, the results demonstrate that the accumulation of DO11.10 T cells in the draining lymph nodes of mice infected with MalE-OVA-expressing bacteria is due to specific activation and proliferation of the transferred T cells.
Ag-activated T cells migrate to the nonlymphoid site of bacterial infection
The ability to physically track Ag-specific T cells allowed us to determine whether the initial activation of naive T cells occurs in the secondary lymphoid tissues or the nonlymphoid site where the Ag enters the body. We addressed this point by using the connective tissue air sac model of bacterial infection (10). Air sacs were produced in the skin on the backs of mice that previously received DO11.10 T cells via the i.v. route. MalE-OVA-expressing or control bacteria were then injected into the air sacs, and the number of DO11.10 T cells and bacteria present within the air sac and the draining lymph nodes was measured. Viable bacteria were present in the air sac immediately after injection (Fig. 6). The number of viable bacteria in the air sac declined slowly after this; however, many viable bacteria were still present 6 days postinfection (Fig. 6). No difference in the rate of clearance of the MalE-OVA-expressing or control bacteria was observed (Fig. 6). From days 1 to 3 postinfection, most of the cells recovered from the air sacs were neutrophils and macrophages as assessed by expression of the Gr-1 and CD11b markers (Fig. 7,A). A small number of cells had the light scatter properties of lymphocytes at this time (R1 in Fig. 7,B), and although a few CD4+, KJ1-26− cells of recipient origin were present in this population, few if any DO11.10 T cells were detected (Fig. 7, C and G). However, by day 4 postinfection a clearly defined population of CD4+, KJ1-26+ DO11.10 T cells was present in the air sacs of mice infected with MalE-OVA-expressing (Fig. 7, D and G), but not control, bacteria (Fig. 7, E and G).
DO11.10 T cells accumulated in the draining lymph nodes about a day earlier than they appeared at the infection site, achieving a near maximal level by day 3 postinfection (Fig. 7,G). This clonal expansion of the DO11.10 T cells in the lymph nodes occurred despite the fact that viable bacteria were not detected in the draining lymph nodes at any time after infection (Fig. 6). Taken together, the data support the possibility that Ag-specific T cells initially proliferate in the secondary lymphoid tissues in response to debris carried from the infection site in the afferent lymph, increase expression of molecules involved in extravasation such as LFA-1, and then migrate to the nonlymphoid infection site.
Injection of purified MalE-OVA fusion protein in the absence of other bacterial components results in DO11.10 T cell hyporesponsiveness
Previous work with this system showed that injection of soluble OVA alone caused the DO11.10 T cells to proliferate transiently in the T cell areas of the draining lymph nodes (6). Most of the cells then disappeared, and although the surviving cells had the surface phenotype of memory cells, they were hyporesponsive to antigenic stimulation (6). In contrast, injection of OVA plus the adjuvants CFA or LPS caused the DO11.10 T cells to proliferate more extensively in the T cell areas and migrate into the B cell-rich follicles (6, 13). Although many of the DO11.10 T cells again disappeared, many did not, and the surviving cells responded rapidly to Ag stimulation by producing lymphokines including IFN-γ as expected for memory cells (13).
Because adjuvants were required for maximal generation of memory T cells when soluble OVA was used as the Ag, it was of interest to determine whether the in vivo stimulation of DO11.10 T cells by the MalE-OVA fusion protein was influenced by the natural adjuvanticity of the bacteria. This was investigated by studying the in vivo behavior of DO11.10 T cells following injection of affinity-purified, endotoxin-free MalE-OVA protein. As shown in Figure 3, s.c. injection of the purified fusion protein caused the DO11.10 cells to accumulate rapidly in the draining lymph nodes to a peak level on day 3. However, most of the cells then disappeared rapidly from the lymph nodes. In addition, the number of DO11.10 T cells present on day 3 in the lymph nodes of mice injected with MalE-OVA alone was much lower than the number present in mice infected with MalE-OVA-expressing bacteria. Furthermore, the DO11.10 T cells did not accumulate in the B cell-rich follicles in response to soluble MalE-OVA (Fig. 8,A) to the same extent as they did in response to MalE-OVA-expressing bacteria (Fig. 8,C). Quantitative image analysis revealed that the density of DO11.10 T cells in the follicles 3 days after injection of MalE-OVA alone (Fig. 8,A) was 14-fold lower than the density of DO11.10 T cells in the follicles of mice injected with MalE-OVA-expressing bacteria (Fig. 8,C). Despite the fact that many of the DO11.10 cells became blasts 3 days after injection of soluble MalE-OVA (Fig. 4,B) and many converted to a “memory” cell phenotype (Fig. 4, C and D), these proliferated poorly (Fig. 9,B) and did not produce IL-2 (Fig. 9,C) or IFN-γ (Fig. 9 D) in response to in vitro stimulation with the OVA peptide at later times. These results indicate that soluble MalE-OVA stimulates a transient and abortive response by the DO11.10 T cells that does not produce functional memory cells.
To assess whether the productive T cell response only occurred when MalE-OVA was physically associated with the bacterial cells, adoptive transfer recipients were injected with a mixture of soluble MalE-OVA protein and bacteria that do not express MalE-OVA. The DO11.10 T cells behaved under these conditions exactly as they did when confronted with MalE-OVA-expressing bacteria: the cells accumulated to high levels in the draining lymph nodes (Fig. 3), entered the follicles (data not shown), and differentiated into IL-2 (Fig. 9,C) and IFN-γ-producing (Fig. 9,D) memory T cells. In addition, recipients injected with soluble MalE-OVA plus control bacteria produced the same amount of anti-OVA Abs as mice infected with MalE-OVA-expressing bacteria (Fig. 9 A). These results demonstrate that the bacteria did not have to express the OVA themselves to facilitate induction of productive immunity.
The experiments presented here were designed to extend our previous studies on the in vivo behavior of Ag-specific CD4+ T cells to a more physiologic situation where the Ag of interest was expressed by a microbe. T cell responses induced by microbes might be expected to differ from those induced by soluble Ags because microbes are a source of replicating Ag and thus would not necessarily disappear rapidly from the body. In addition, microbial products may modify the T cell response, for example, by altering Ag processing or presentation (15) or T cell activation (16). Given these possibilities, it was remarkable to observe that MalE-OVA-expressing bacteria stimulated the same kinetics of clonal expansion and contraction, follicular migration, and differentiation of the Ag-specific T cells stimulated by injection of soluble Ag in CFA or LPS (6, 13). Therefore, it appears that T cell responses to microbial Ags, at least those derived from noninvasive microbes, can be reliably modeled with soluble Ags plus adjuvant.
The simplest explanation for the finding that MalE-OVA stimulated a productive response by Ag-specific T cells when introduced into the body as part of a bacterium but not as a soluble protein is that bacterial components act as immunologic adjuvants (17). We have shown that efficient clonal expansion and follicular migration of Ag-activated T cells only occurs when Ag is delivered with LPS, TNF-α, or IL-1 (13). Thus, it is likely that bacterial components such as LPS stimulate TNF-α and IL-1 production by macrophages, and that these cytokines produce an environment that is conducive to T cell activation (18). This could be accomplished by induction of adhesion and costimulatory molecules (19) or stabilization of peptide-MHC molecules (20) on APC such as dendritic cells, the most abundant class II MHC-bearing APC in the T cell areas of lymphoid tissue (21). An alternative possibility was that the recombinant bacteria provide a replicating source of OVA, which results in prolonged Ag-presentation compared with a single injection of soluble OVA. However, the finding that a mixture of nonreplicating soluble Ag and control bacteria induced full activation of the Ag-specific T cells is strong evidence against this possibility.
Our results suggest that the T cells specific for microbial Ags are initially activated in the secondary lymphoid tissues such as the draining lymph nodes. This supposition is based on the finding that activated Ag-specific T cells accumulated in the draining lymph nodes before accumulating in the s.c. infection site. Earlier work in this system showed that naive DO11.10 T cells are only present in the blood and T cell-rich areas of secondary lymphoid tissues of recipient mice; few if any cells are found in nonlymphoid tissues (22). This restriction argues that naive T cells would have to be activated in lymphoid tissues before migrating to nonlymphoid tissues. The finding that viable bacteria were never found in the lymph nodes suggests that some of the bacteria are killed at the injection site, perhaps by neutrophils, and that debris including MalE-OVA is carried to the lymph nodes via the afferent lymph for presentation by resident APC. Alternatively, immature dendritic cells at the s.c. site may internalize bacterial debris, process MalE-OVA and migrate to the draining lymph node (23, 24, 25). Once activated in the secondary lymphoid tissues, the Ag-specific T cells lost l-selectin and increased expression of LFA-1, changes that would be predicted to limit re-entry into lymph nodes and facilitate entry into inflamed nonlymphoid tissues such as the infection site (26, 27).
The finding that MalE-OVA-expressing and control bacteria were cleared equally well, despite the fact that the mice contained an elevated population of OVA-specific DO11.10 T cells, is consistent with work from other studies that showed that cells other than T cells are the primary effectors of clearance of E. coli (28). However, it is possible that T cells contribute to clearance, but the number of T cells that are specific for E. coli peptide-MHC complexes is not limiting such that the additional DO11.10 T cells made no difference. Although no role for the transferred T cells in clearance of E. coli could be shown here, adoptive transfer of T cells from bacterial Ag-specific TCR transgenic donors should be a powerful tool for studying the in vivo behavior of Ag-specific T cells during infections caused by pathogens such as Salmonella, protection from which, is known to be dependent on CD4+ T cells (29).
We thank K. A. Pape for instruction on the Adobe Photoshop image analysis software and helpful discussion, Dr. Y. Ji for assistance with s.c. air-sac injection, and J. White for expert technical assistance.
This work was supported by National Institutes of Health Grants AI27998 and AI39614.
Abbreviations used in this paper: IPTG, isopropyl β-d-thiogalactoside; PE, phycoerythrin; SA, streptavidin; BrdU, 5-bromodeoxyuridine