Influence of the Route of Infection on Development of T-Cell Receptor β-Chain Repertoires of Reovirus-Specific Cytotoxic T Lymphocytes

Abstract

It is well established that the route of infection affects the nature of the adaptive immune response. However, little is known about the effects of the route of exposure on development of cytotoxic T-lymphocyte (CTL) responses. Alternative antigen-presenting cell populations, tissue-restricted expression of class I major histocompatibility complex-encoded molecules, and unique T-cell receptor (TCR)-bearing cells in mucosal tissues could influence the selection and expansion of responder T cells. This study addresses the question of whether the route of virus infection affects the selection and expansion of subpopulations of virus-specific CTLs. Mice were infected orally or in the hind footpads with reovirus, and the repertoires of TCR β-chains expressed on virus-specific CD8⁺ T cells in Peyer's patches or lymph nodes and spleens were examined. CD8⁺ cells expressing the variable gene segment of the TCR β-chain 6 (Vβ6) expanded in the spleens of mice infected by either route and in CTL lines established from the spleens and draining lymphoid tissues. Adoptively transferred Vβ6⁺ CD8⁺ T cells from orally or parenterally infected donors expanded in reovirus-infected severe combined immunodeficient recipient mice and mediated cytotoxicity ex vivo. Furthermore, recovered Vβ6⁺ cells were enriched for clones utilizing uniform complementarity-determining region 3 (CDR3) lengths. However, sequencing of CDR3β regions from Vβ6⁺ CD8⁺ cells indicated that Jβ gene segment usage is significantly more restricted in CTLs from orally infected mice, suggesting that the route of infection affects selection and/or subsequent expansion of virus-specific CTLs.

Respiratory enteric orphan virus (reovirus) has been used as a model viral pathogen to study virus-host interactions at mucosal surfaces and in the periphery. Reovirus is a nonenveloped, segmented double-stranded RNA virus (31) that replicates and elicits both humoral and cell-mediated immunity following oral or parenteral infection (7, 13, 14, 16, 25). After gaining access to the intestinal tissue via the M cells of the Peyer's patches (PP) (46), reovirus serotype 1 strain Lang (T1/L) infection of immunocompetent mice causes a self-limited disease of enterocytes of the crypts of Lieberkuhn adjacent to the PP of the distal ileum (37). In response, virus-specific cytotoxic T lymphocytes (CTLs) are induced within the PP (25). CTLs migrate via efferent lymphatic vessels to the mesenteric lymph nodes and then through the thoracic duct lymph and the systemic circulation to the spleen (25) or to intestinal mucosal sites, such as the intestinal intraepithelial lymphocyte (IEL) compartment (8, 9, 24). Parenteral infection with reovirus induces virus-specific CTLs in the draining peripheral lymph nodes and spleen (22, 45).

Although the humoral immune response to reovirus is influenced by microenvironmental or cellular factors at the anatomic site of infection (7, 26, 44), it is not known whether intestinal reovirus infection results in a CTL response distinct from that following parenteral infection. Most reovirus-specific CTLs are CD8αβ⁺ TCRαβ⁺ Thy-1⁺ and major histocompatibility complex (MHC) class I restricted (25). In addition, reovirus-specific CD8⁺ CTLs induced in the PP following enteric infection and in the lung following respiratory infection express the unusual cell surface marker called germinal center and T antigen (23, 44). Intratracheal instillation of reovirus elicits an unusual population of cytotoxic CD4⁺ CD8αβ⁺ TCRαβ⁺ T cells (34), suggesting that unique CTL populations might be induced by infection at distinct anatomic locations. Furthermore, enteric reovirus T1/L infection of C3H mice has been shown to elicit CD8αβ⁺ CTL populations expressing Vβ12 and Vβ17, with minor populations expressing Vβ2, Vβ7, Vβ9, and Vβ14 among the IELs (8). These CTL populations were thought to be representative of the CTL response primed in the PP following oral infection with reovirus T1/L (9), although some might have been derived in situ, given the uncertain ontogeny of the IEL.

There are several reasons why CTLs induced following oral infection could be different from those induced parenterally. During enteric reovirus infection, the ingested virions undergo pancreatic chymotryptic proteolysis of the outer coat proteins in the duodenum to yield infectious intermediate subviral particles (3). This processing within the lumen of the small intestine could potentially generate antigenic determinants distinct from those generated in the systemic periphery. Additionally, cleaved viral antigens might be taken up by intestinal absorptive epithelial cells (32, 42), which express a number of classical and nonclassical antigen presentation and costimulatory molecules, such as MHC class I, class II, CD1 and TL (6, 17), and the CD8 binding and activating ligand gp180 (6). These intestinal epithelial cells are thought to act as unconventional antigen-presenting cells to lymphocytes of the intestinal mucosa (20). Evidence also exists that lamina propria dendritic cells underlying the epithelium sample lumenal antigen directly (20) or acquire antigen from endocytosed, apoptotic intestinal epithelial cells (18), possibly acting in the priming of mucosal T-cell populations (29). In addition, Kuckelkorn et al. (21) recently presented evidence that tissue-specific differences exist in proteosome subunit usage that result in tissue-specific differences in processing of MHC class I T-cell epitopes. Thus, the generation of different antigenic determinants and the presence of distinct populations of antigen-presenting cells could influence the priming of the CTL response.

Because of the existence of unique populations of mucosal effector and antigen-presenting cells and the potential that unique epitopes can be generated at mucosal sites, it is reasonable to question whether qualitative differences develop in T-cell-mediated immune responses following mucosal versus systemic infection. Resolving this issue could impact rational vaccine design, particularly for subunit vaccines that could be administered mucosally, and our understanding of oral tolerance, a phenomenon that has been exploited for treating autoimmune disease. We sought to determine if the route of infection influences the virus-specific CTL response by characterizing the T-cell receptor (TCR) β-chain repertoires of CD8⁺ T cells from mice infected orally or parenterally with reovirus. We also assessed the Vβ repertoires and β-chain complementarity-determining region 3 (CDR3β) spectratypes of virus-specific precursor CTLs primed in the PP, as well as those primed in the peripheral lymph nodes, to determine whether the virus-specific CTL populations recruited to the spleen following reovirus infection were representative of all the populations initially induced in the priming sites. Additionally, we adoptively transferred splenic and PP T cells from mice previously infected orally or in the hind footpads into reovirus-infected syngeneic severe combined immunodeficient (SCID) mice to determine what CTL populations expand in the presence of continuous virus restimulation in vivo. We further characterized the responder Vβ6⁺ CTL population by CDR3β length distribution and by sequencing of the CDR3β region of over 90 randomly selected clones. We show that the Vβ6⁺ CD8⁺ CTLs from both orally and parenterally infected mice utilize CDR3β regions of similar lengths. However, analysis of Jβ gene segments revealed a highly statistically significant bias in Jβ gene segments used in TCRs on cells from orally infected mice compared to TCRs from systemically infected mice, suggesting that the route of infection influences repertoire selection and/or expansion of virus-specific CTLs.

MATERIALS AND METHODS

Animals.

Seven- to 8-week-old male C3HeB/FeJ and C3H^SmnPkrc SCID mice were purchased from Jackson Laboratories (Bar Harbor, Maine) and were housed in microisolator cages under specific-pathogen-free conditions. This work was performed under a protocol approved by the West Virginia University Institutional Animal Care and Use Committee.

Virus.

Third-passage reovirus T1/L stocks were grown in L929 cells and purified by 1,1,2-trichloro-1,2,2-trifluoroethane extraction followed by step-wise CsCl gradient centrifugation as previously described (41).

Animal infections.

Orally infected mice received 3 × 10⁷ PFU of reovirus in 50 μl of sterile borate-buffered saline and gelatin by gavage. Hind footpad-infected mice received 1.5 × 10⁷ PFU of reovirus in 10 μl of sterile borate-buffered saline and gelatin in each hind footpad. SCID recipient mice received 3 × 10⁷ PFU of reovirus in 100 μl total of sterile borate-buffered saline and gelatin by intraperitoneal injection.

Preparation and culture of splenic, lymph node, and PP cells.

Single cell suspensions of splenic and popliteal lymph node (PLN) cells were prepared by expressing tissues through sterile nylon mesh. Single-cell suspensions of PP were prepared by mechanical dissociation using sterile 18-gauge needles. Splenocytes were depleted of red blood cells by incubation in a hypotonic solution of NH₄Cl and Tris-HCl. Splenocytes and PLN cells were washed three times in medium consisting of RPMI 1640 (Biowhittaker, Walkersville, Md.) supplemented with 10% fetal bovine serum (FBS; HyClone, Logan, Utah), 10 mM HEPES, 2 mM l-glutamine (Biowhittaker), 100 U of penicillin/ml, 0.1 μg of streptomycin/ml (Biowhittaker), and 50 μM β-mercaptoethanol (TCM).

Generation of reovirus-specific T-cell lines.

CTL lines were generated by suspending cells in TCM to a concentration of 10⁷ cells/ml prior to restimulation with reovirus T1/L at a multiplicity of infection of 1 for 1 h at 25°C. Cells were subsequently resuspended in TCM to a final concentration of 2 × 10⁶ cells/ml and incubated at 37°C in an atmosphere of 5% CO₂ for the first week. Cultures were restimulated at weekly intervals using syngeneic naïve splenic feeder cells previously pulsed for 1 h at 25°C with reovirus T1/L and irradiated with 3,000 rad of gamma radiation. Irradiated feeder cells were cultured with effector splenocytes at a ratio of five feeders to each effector and brought to a final concentration of 2 × 10⁶ cells/ml in fresh TCM containing 5% conditioned medium from concanavalin A-stimulated rat splenocytes and 5% 1 mM α-methylmannoside. Cultures were incubated at 37°C in an atmosphere of 5% CO₂.

Depletion of CD4⁺ or CD8⁺ cells.

CD4⁺ or CD8⁺ T cells were depleted by treatment with anti-mouse CD4 immunoglobulin M (IgM; RL172) or anti-mouse CD8α IgM (3.155), followed by the addition of Low-Tox rabbit complement (Cedarlane, Hornby, Ontario, Canada) to a final concentration of 10% by volume. The efficacy of complement-mediated depletion was determined by flow cytometry.

Enrichment and depletion of Vβ-expressing populations by indirect panning.

Ficoll-Hypaque gradient-enriched cells were treated with sterile phosphate-buffered saline (PBS)-diluted anti-Vβ antibodies. Antibodies used were the following: MR9-4 (mouse anti-mouse Vβ5.1 and -5.2 IgG2a), F23.1 (mouse anti-mouse Vβ8.1, -8.2, and -8.3 IgG), RR3-15 (rat anti-mouse Vβ11 IgG), or KJ23 (mouse anti-mouse Vβ17 IgG), RR4-7 (rat anti-mouse Vβ6 IgG2b), TR310 (rat anti-mouseVβ7 IgG2b), MR11-1 (mouse anti-mouse Vβ12 IgG1), MR12-3 (mouse anti-mouse Vβ13 IgG1), and 14-2 (rat anti-mouse Vβ14 IgM) (Pharmingen, San Diego, Calif.). The cells were washed three times in PBS-5% FBS and resuspended in 4 ml of PBS-5% FBS. Half of the cells from each treatment group were panned by two successive incubations for 45 min at 4°C on sterile 16- by 100-mm polystyrene bacteriology-grade petri dishes (Falcon, Franklin Lakes, N.J.) coated with 10 μg of goat anti-rat or goat anti-mouse IgG (Rockland, Gilbertsville, Pa.)/ml in a 0.15 M Tris buffer (pH 9). Adherent cells were washed three times with PBS-5% FBS before being dislodged with a sterile rubber SP scraper and cultured as above. Depleted and nondepleted cell cultures were assessed for TCR expression by flow cytometry.

Adoptive transfer.

Donor lymphocytes collected from mice that were infected 30 days previously were depleted of surface Ig-positive B cells by three successive incubations for 45 min at 4°C on sterile 16- by 100-mm polystyrene bacteriology-grade petri dishes (Falcon) coated with 100 μg of goat anti-mouse IgG, IgM, and IgA antibody (Rockland)/ml in a 0.15 M Tris buffer (pH 9). A total of 1.5 × 10⁷ nonadherent splenocytes were resuspended in 0.5 ml of sterile PBS and were injected intraperitoneally into reovirus-infected SCID recipient mice.

Cytotoxicity assay.

Assays were performed as previously described (10). Briefly, Ficoll-Hypaque gradient-purified effector CTLs were added in twofold dilutions to duplicate or triplicate wells of 5 × 10³ reovirus-infected or noninfected L929 target cells labeled with 150 μCi of sodium [⁵¹Cr]dichromate (Dupont, Wilmington, Del.) in V-bottom 96-well plates (Nalgene, Rochester, N.Y.) at various effector/target cell ratios. The assay plates were incubated for 4 h at 37°C in an atmosphere of 5% CO₂. Specific lysis of reovirus-infected targets for each effector/target ratio was calculated as described elsewhere (25).

Flow cytometric analysis of effector cells.

Red blood cell-depleted mononuclear splenocytes, PP cells, and PLN cells were stained with phycoerythrin-conjugated anti-CD8α monoclonal antibody (MAb; CALTAG, Burlingame, Calif.) and fluorescein isothiocyanate (FITC)-conjugated MAbs to either CD4 or TCR Vβ (Pharmingen). In some experiments, cells were also stained with a biotinylated anti-Vβ3 MAb (Pharmingen) followed by FITC-avidin (CALTAG). Percentages of cells staining positive for each marker were determined by analysis on a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, N.J.), and results were further analyzed using WinMIDI 2.8 software (J. Trotter, Scripps Research Institute, La Jolla, Calif.).

CDR3β length profile analysis.

Fragment analyses were performed by a modification of the methods originally reported by Pannetier et al. (33). mRNA was isolated from CD8⁺ cells using Qiagen RNAeasy kits as per the manufacturer's instructions. Isolated mRNA was converted to cDNA and subsequently amplified using Vβ6 or Vβ2 primers and a carboxytetramethylrhodamine-conjugated Cβ primer (Integrated DNA Technologies, Coralville, Iowa) using the Qiagen reverse transcription-PCR kit (Qiagen Inc., Valencia, Calif.) as per the manufacturer's instructions. Primer sequences were as follows: Vβ6, 5′-CTCTCACTGTGACATCTGCCC-3′; Vβ2, 5′-TCACTGATACGGAGCTGAGGC-3′; Cβ, 5′-CTTGGGTGGAGTCACATTTCTC-3′ (33). Amplified products were analyzed using an ABI Prism model 377 automated sequencer at the DNA core facility of the University of California—San Francisco, and CDR3β length distribution profiles were analyzed using Genescan 3.0 software (Perkin-Elmer, Norwalk, Conn.).

CDR3β sequencing.

TCR β-chain cDNA derived as above from Vβ6⁺ CD8⁺ T cells was amplified by PCR and subsequently ligated into the pCR 4-TOPO plasmid. Plasmids were electroporated into competent Escherichia coli cells and cloned using the TOPO TA cloning kit (Invitrogen, Carlsbad, Calif.). Insert-containing plasmids were sequenced by automated dye terminator sequencing (University of Pennsylvania DNA Core Facility, Philadelphia, Pa., or Northwoods DNA Inc., Bemidji, Minn.). Sequence data were analyzed using DNA Strider 1.2 software (C. March, Centre d'Etudes de Saclay, Gif sur Yvette, France). Jβ gene segments were identified as previously published sequences (15, 27) and confirmed by BLAST searches in GenBank (accession no. AE000665).

Statistical analysis.

Means and standard deviations were calculated for each Vβ population expressed among all groups of mice and tissues assessed by flow cytometry. Statistically significant differences in Vβ expression were determined by one-way analysis of variance followed by Tukey's test using Sigmastat 2.0 software (Jandel Scientific, Chicago, Ill.). Fischer's exact test was used to compare Jβ gene segment usage in CDR3 regions from orally or systemically infected mice.

RESULTS

Vβ6⁺ cells are slightly expanded in the spleens of orally or systemically infected mice.

The Vβ repertoire of CD8⁺ splenocytes from mice that were infected orally or in the hind footpads with reovirus 10 days previously was assessed by flow cytometry. CD8⁺ splenocytes from footpad-infected mice showed a slight but significantly elevated percentage and absolute number of Vβ6⁺ cells compared to noninfected age-matched mice, while no difference in the percentage or absolute numbers of cells was observed in orally infected mice(http://www.hsc.wvu.edu/micro/cuff/tcrpaper/dns1.pdf). TheVβ repertoire of CD8⁺ T cells in PP and PLN of mice infected orally or in the hind footpads with reovirus was also assessed. No significant differences in the percent expression of any Vβ population of CD8⁺ T cells were noted among mice infected by either route compared to mock-infected controls (http://www.hsc.wvu.edu/micro/cuff/tcrpaper/dns2.pdf).

In vitro restimulation with reovirus expands virus-specific Vβ6⁺ populations of CD8⁺ T cells from the spleens of footpad-infected and orally infected mice.

Vβ6⁺ CD8⁺ cells were enriched in cell lines from PP and PLN from orally or systemically infected mice (Fig. 1). In some experiments, as shown, expansion of Vβ8⁺ cells was observed in in vitro cultures but was not consistent following either oral or systemic infection. CTL lines established from splenocytes from either footpad-infected or orally infected mice consistently showed similar expansion of Vβ6⁺ CD8⁺ T cells, with lesser representation of minor populations such as Vβ7⁺ (Fig. 2). Most other CD8⁺ Vβ populations decreased to low levels by 4 weeks in vitro. Coincident with expansion of Vβ6⁺ cells, in vitro restimulation with reovirus enriched virus-specific CTLs, which mediated strong virus-specific cytotoxicity (Fig. 2, inset).

FIG. 1.

Vβ repertoire of representative cultures of CD8⁺ PLN or PP cells from animals infected with reovirus in the hind footpads or perorally. Shown are the percentages of CD8⁺ cells staining positive with a panel of FITC-anti-Vβ antibodies immediately ex vivo and after 2 weeks of in vitro restimulation with reovirus.

FIG. 2.

Vβ repertoire of CD8⁺ spleen cell lines from infected mice. Representative CTL lines established from the spleens of mice infected with reovirus in the hind footpads or orally were analyzed after 2 weeks (closed bars) and 4 weeks (open bars) in culture. Shown are the percentages of CD8⁺ cells staining positive with a panel of FITC-anti-Vβ antibodies as assessed by flow cytometry. Insets are virus-specific cytotoxicity mediated by the CTL lines at 2 weeks (closed circles) and 4 weeks (open circles) in vitro. Cytotoxicity data are presented as (mean percent lysis of virus-infected L929 target cells) − (percent lysis of noninfected L929 cells) (y axis) at various effector cell/target cell ratios (x axis).

Vβ6⁺ CD8⁺ cells mediate reovirus-specific cytotoxicity.

Splenic CTL lines from mice infected orally or in the hind footpads were depleted of CD4⁺ T cells, and the remaining CD8⁺ cells were fractionated into Vβ6⁺-, Vβ7⁺-, or Vβ8.1⁺-, -8.2⁺-, and -8.3⁺-depleted and enriched fractions by indirect panning. Depletion removed 94 to 99% of the target population, and enrichment yielded up to 63% positive populations. Sorted cell lines were subsequently assessed for reovirus-specific cytotoxicity. The Vβ6-depleted fractions of splenic CTL lines from footpad-infected mice showed diminished reovirus-specific killing, whereas the Vβ6-enriched fractions of splenic CTL lines consistently mediated high levels of reovirus-specific cytotoxicity. In contrast, the Vβ7-enriched fractions did not mediate reovirus-specific killing (Fig. 3A and B). Vβ8-enriched fractions did on rare occasion mediate a low level of killing (Fig. 3B). Additionally, the Vβ6-depleted fractions of splenic CTL lines from orally infected mice consistently showed diminished reovirus-specific killing (Fig. 3C). By contrast, the Vβ6-enriched fractions of splenic CTL lines consistently mediated high levels of reovirus-specific cytotoxicity. No other CD8⁺ T-cell Vβ populations in splenic CTL lines from orally and footpad-infected mice were found to mediate reovirus-specific cytotoxicity, including Vβ5, Vβ9, Vβ10, Vβ11, Vβ12, Vβ13, Vβ14, or Vβ17 (data not shown). Together, these data indicated that Vβ6⁺ cells were the dominant killers in cultures generated from orally or systemically infected mice.

FIG. 3.

Reovirus-specific cytotoxicity by subpopulations of CD8⁺ CTL lines established from orally infected and footpad-infected mice. CTL lines were depleted of CD4⁺ cells and sorted on the basis of Vβ expression prior to assessment of cytotoxicity. Fractions of splenic CTL lines from footpad-infected mice (A and B) and orally infected mice (C) depleted of or enriched for Vβ6⁺, Vβ7⁺, and Vβ8⁺ CD8⁺ T cells are shown. Control groups represent the CTL activity of cell lines not subjected to fractionation. Data are presented as mean percent lysis ± standard deviation of triplicate wells of infected L929 targets minus noninfected L929 targets.

Reovirus-primed CD8⁺ cells from spleen or PP expand in reovirus-infected SCID recipients and are enriched for virus-specific CTLs.

In vitro expansion of virus-specific CTLs could skew the repertoires of responding cells. To assess CTL expansion in vivo, 1.5 × 10⁷ reovirus-immune splenic T cells (approximately 4.6 × 10⁶ CD8⁺ cells) were injected intraperitoneally into congenic reovirus-infected SCID mice. Recipient virus-infected SCID mice were sacrificed at days 8, 10, and 12 posttransfer, and their splenocytes were stained for Vβ expression. A two- to fivefold expansion in the number of transferred Vβ6⁺ CD8⁺ cells in the spleens of infected SCID recipients was observed for recipients of both orally infected and footpad-infected donor cells (Fig. 4). Minor expansion of other populations, including Vβ7, Vβ14, and Vβ17, was inconsistently observed. Spleen cells recovered from infected SCID recipients of both orally infected and footpad-infected donors mediated strong ex vivo cytotoxicity (Fig. 5). Depletion of CD8⁺ cells, but not CD4⁺ cells, eliminated virus-specific killing.

FIG. 4.

Vβ repertoire of CD8⁺ T cells recovered from reovirus-infected SCID recipients. Shown are analyses of splenocytes from donor mice infected orally (A) or in the hind footpads (B). Splenocytes were isolated from recipient reovirus-infected SCID mice at days 8, 10, and 12 posttransfer. Data are presented as mean cell number ± standard deviation. In both experiments, n = 3, 3, and 4 SCID mice for days 8, 10, and 12, respectively.

FIG. 5.

Cytotoxic activity of splenocytes expanded in SCID mice. Cells recovered from reovirus-infected SCID recipients 10 days after adoptive transfer of purified splenic T cells from donor mice previously infected orally (A) or in the hind footpads (B) were assayed for CTL activity. Recovered splenocytes were depleted of either CD4⁺ cells or CD8⁺ cells as described in Materials and Methods. Control fractions were treated with complement alone. Data are presented as mean percent lysis ± standard deviation of triplicate wells of infected L929 targets minus noninfected L929 targets.

In addition to experiments with spleen as donor tissue, PP lymphocytes from orally infected mice or naïve mice were injected intraperitoneally into SCID mice infected with reovirus 4 days previously. Recipient SCID mice were sacrificed at day 10 posttransfer, and the Vβ repertoire of recovered CD8⁺ spleen T cells was analyzed (Fig. 6). Consistent with the results for the adoptive transfer of immune spleen cells, a dramatic expansion in the number of Vβ6⁺ CD8⁺ cells in the spleens of infected SCID recipients was observed. CD8⁺ cells recovered from the spleens of infected SCID recipients of naïve CD8⁺ lymphocytes showed no clear expansion of any Vβ population. Spleen cells recovered from infected SCID recipients of PP lymphocytes from orally infected and naïve donor mice were also assessed at day 10 posttransfer for ex vivo cytotoxicity. Lymphocytes from immune donors mediated strong reovirus-specific ex vivo cytotoxicity. By contrast, lymphocytes from naïve donors recovered from infected SCID recipients did not mediate killing (Fig. 7).

FIG. 6.

Vβ repertoire of CD8⁺ lymphocytes recovered from SCID recipients of PP T cells from orally infected (A) or naïve (B) donor mice. Ten days after adoptive transfer of purified PP T cells, splenocytes were recovered from SCID recipient mice and assessed for the Vβ repertoire among CD8⁺ cells. Data are presented as mean cell number ± standard deviation. (A) n = 5 recipient mice compiled from two separate experiments; (B) n = 3 recipient mice from one experiment.

FIG. 7.

Ex vivo reovirus-specific cytotoxicity mediated by splenocytes recovered from SCID recipient mice. Ten days after adoptive transfer of purified PP T cells from immune or naïve donors, splenocytes were recovered from SCID recipients and assessed for reovirus-specific cytotoxicity. Data are presented as mean percent lysis ± standard deviation of tripicate wells of infected L929 targets minus noninfected targets. Each line represents the response by an individual mouse.

Adoptively transferred, reovirus-primed Vβ6⁺ CD8⁺ cells from reovirus-immune donors show altered CDR3β length profiles upon recovery from reovirus-infected SCID recipients.

Vβ6⁺ cells were preferentially activated following either systemic or oral infection. CDR3 spectratype analysis was performed to assess potential oligoclonal differences within the cytotoxic Vβ6⁺ T cells resulting from oral or systemic infection. Spleen cells recovered at day 10 posttransfer from infected SCID recipients of splenic T cells from footpad- or orally infected donor mice, or PP T cells from orally infected donor mice, were depleted of CD4⁺ T cells and assessed for CDR3β length distribution among Vβ6⁺ or, as a control, Vβ2⁺ cells. Vβ6⁺ CD8⁺ cells recovered from reovirus-infected SCID recipients of cells from both orally infected and hind footpad-infected donors showed consistent and similar alterations in the CDR3β length distribution compared to the input cells (Fig. 8).

FIG. 8.

CDR3β length distributions of Vβ6⁺ and Vβ2⁺ CD8⁺ T cells recovered from the spleens of reovirus-infected SCID recipients of reovirus-immune T cells. Splenic T cells from footpad-infected mice and T cells from the spleens or PP of orally infected donor mice were adoptively transferred into reovirus-infected SCID recipient mice. Splenic mononuclear cells were recovered from recipient mice 10 days posttransfer and were depleted of CD4⁺ cells. Cellular RNA was isolated and used as template for reverse transcription-PCR using Vβ2 or Vβ6 and Cβ primers and assessed as described in Materials and Methods. Data shown are representative CDR3β length distributions for recovered Vβ6 and Vβ2 CD8⁺ cells. Arrows indicate consistent expansions of uniform CDR3β lengths.

The route of infection influences expansion of Jβ gene segments used by CD8⁺ Vβ6⁺ T cells expanded in virus-infected SCID mice.

Analysis of the CDR3 region by spectratype and sequencing has been used to characterize immunodominant clonotypes in the T-cell response to pathogens and autoantigens (35). The highly diverse CDR3 region of the TCR β-chain is encoded by the terminal end of the Vβ gene segment, the Dβ gene segment, and the proximal end of the Jβ gene segment. X-ray crystallographic analysis suggests that the CDR3 region of the β-chain segment binds directly to peptides presented in class I MHC molecules and therefore contributes significantly to the fine specificity of the TCR (11). CD8⁺ spleen cells were isolated from spleen cells from five individual reovirus-infected SCID recipients of T cells from systemically (n = 2) or orally (n = 3) infected mice. TCR CDR3 sequences were amplified using Vβ6- and Cβ-specific primers. CDR3 regions were analyzed in 92 randomly selected cDNA clones. A number of clones bore identical or nearly identical sequences. However, comparison of the frequency of Jβ gene segments used by TCRs on cells initially expanded by oral infection compared to systemic infection revealed a route of infection difference that was highly significant (Fischer's exact test, P < 0.00001) (Table 1). Both routes of infection expanded clones using Jβ1.6 or -2.1. However, oral infection resulted in a substantial percentage of clones that expressed Jβ1.1, which was rarely found in clones from systemically infected mice. Overall, cells stimulated by oral infection and then expanded in SCID mice used a restricted repertoire of Jβ gene segments compared to cells primed by systemic infection.

TABLE 1.

Jβ gene segments among CD8⁺ Vβ6⁺ TCR clones from virus-infected SCID recipients of systemically or orally primed effector cells

Infection	Mouse	Jβ segment
		Total	1.1	1.2	1.3	1.4	1.5	1.6	2.1	2.2	2.3	2.4	2.5	2.6	2.7
Systemic	1	9	1	1	0	0	0	1	2	0	1	0	1	1	1
	2	11	0	3	1	1	0	2	1	1	1	1	0	0	0
	Total (%)	20 (100)	1 (5)	4 (20)	1 (5)	1 (5)	0 (0)	3 (15)	3 (15)	1 (5)	2 (10)	1 (5)	1 (5)	1 (5)	1 (5)
Oral	1	26	22	1	2	0	0	0	0	0	0	0	1	0	0
	2	26	10	1	4	1	0	0	10	0	0	0	0	0	0
	3	20	7	1	0	0	0	11	1	0	0	0	0	0	0
	Total (%)	72 (100a)	39 (54.2)	3 (4.2)	6 (8.4)	1 (1.4)	0 (0)	11 (15.3)	11 (15.3)	0 (0)	0 (0)	0 (0)	1 (1.4)	0 (0)	0 (0)

^aApproximate.

DISCUSSION

The goal of this study was to ascertain whether the route of infection influenced the repertoire of virus-specific CTLs. Vβ6⁺ CD8⁺ T cells from spleen, PP, and PLN consistently expanded following in vivo virus infection and in vitro restimulation with reovirus. CD8⁺ T-cell populations bearing other TCR Vβ elements were observed following several rounds of in vitro restimulation, albeit less consistently than the ubiquitous Vβ6⁺ CD8⁺ T cells. These other CD8⁺ Vβ6⁻ populations included cells bearing Vβ8.1 or Vβ8.2, Vβ8.3, Vβ7, Vβ14, or Vβ13 (in order of frequency of occurrence) and showed no association with the initial route of infection. These Vβ6⁻ cells mediated low or no CTL activity, and so it was not clear whether they were virus specific but not cytotoxic or if they were perhaps artifacts of in vitro culture. Nevertheless, only the Vβ6⁺ CD8⁺ cells consistently mediated virus-specific cytotoxicity, regardless of the initial route of infection.

It was possible that in vitro culture skewed the development of CTL populations that developed from different routes of infection. Therefore, we expanded reovirus-specific CTLs in vivo by transferring splenic T cells, from donors infected either orally or systemically, into virus-infected SCID mice. We reasoned that immune cells would proliferate in virus-infected SCID mice, because previous reports indicated that immune PP T cells adoptively transferred into SCID recipients were able to clear reovirus infection (16). Transferred splenocytes and/or their progeny were recovered from the spleens of reovirus-infected SCID mice in the second week posttransfer and demonstrated high levels of virus-specific cytotoxicity immediately ex vivo. This was surprising, because reovirus-specific CTL activity is typically not detectable directly ex vivo in cells from infected immunocompetent mice but requires in vitro restimulation with virus. On the basis of depletion experiments, it was apparent that recovered CD8⁺ T cells were responsible for the observed virus-specific cytotoxicity, not CD4⁺ CTL or activated NK cells (43). In some adoptive transfer systems, the relative ratio of CD4⁺ to CD8⁺ T cells transferred into T-cell-deficient mice appears to be tightly regulated (36). However, we observed a rapid inversion of the CD4/CD8 ratios of recovered donor cells (data not shown), a result that closely agrees with the report of Zimmermann and Pircher (47), who used vaccinia virus, vesicular stomatitis virus, or lymphocytic choriomeningitis virus. The proliferation of CD8⁺ T cells in our transfer experiments was predominantly due to the activation of memory CTLs, because T cells recovered from the spleens of infected SCID recipients of naïve T cells did not mediate reovirus-specific cytotoxicity ex vivo. Additionally, nonspecific proliferation of donor CD8⁺ T cells in the recipient to fill the otherwise empty T-cell compartment (38), a phenomenon called blind homeostasis (1, 2), did not result in preferential Vβ6⁺ Τ-cell expansion, as neither CD8⁺ spleen cells recovered from the infected SCID recipients of naïve PP cells nor CD8⁺ spleen cells recovered from noninfected SCID recipients of immune donor splenocytes showed any expansion of Vβ6⁺ cells. This system provides a method to expand functional CTLs in vivo and a tool to further analyze the development and function of virus-specific intestinal T cells.

There are indications that reovirus infection in distinct mucosal compartments might prime unique CTL populations. For example, intratracheal instillation with reovirus T1/L primes CTL responses in the draining tracheobronchial and mediastinal lymph nodes, including unique CD4⁺ CD8αβ⁺ TCRαβ⁺ CTLs (34). Additionally, Chen et al. (8) determined that enteric reovirus T1/L infection of C3HeB/FeJ mice induced reovirus-specific CTLs in the intestinal IEL compartment that were Vβ12⁺ or Vβ17⁺, with other apparently virus-responsive CD8⁺ populations expressing either Vβ2, Vβ7, Vβ9, or Vβ14. We demonstrated that the CTL populations recovered from the PP and spleen or the PLN and spleen following enteric and systemic infections were essentially the same, yet distinct, from those that Chen et al. expanded from the IEL. Although evidence exists that reovirus-specific CTLs recovered from the IEL compartment are originally primed in the PP, the intestinal epithelium is also populated by unconventional CD8⁺ T cells of uncertain ontogeny and function which might not be representative of conventional T cells of the peripheral lymphoid tissues. The presence of unusual lymphoid structures, such as cryptopatches (19, 39) and lymphocyte-filled villi (28), postulated by some to provide intestinal sites of extrathymically derived IEL, suggests that T cells produced and primed in situ might provide a source of virus-specific CTLs in addition to those derived from the PP. The relationship between the Vβ6⁺ CTLs we observed and the Vβ12⁺ and Vβ17⁺ CTLs found in the IEL compartment by Chen et al. remains to be determined and is currently under investigation. These observations also highlight the issue that the mice used in this report were conventionally reared and housed under specific-pathogen-free conditions. Numerous studies suggest that the intestinal microflora can influence T-cell responses to antigens, which could result in skewing of intestinal TCR repertoires (12, 40). In contrast to this notion, Bousso et al. showed that the formation of TCR repertoires in peripheral CD8⁺ T cells is similar in conventional and germ-free mice (4). Nevertheless, one could examine the role of bacterial microflora in influencing the TCR repertoire by performing similar experiments in germ-free or antigen-free mice.

Whether or not enteric reovirus infection induces minor populations of CTLs of unconventional ontogeny, the CTL response was overwhelmingly dominated by Vβ6⁺ cells. However, distinct clonal responses might occur within the broad Vβ6⁺ response, dependent on the anatomic site of infection. A more detailed characterization of the responder Vβ6⁺ CD8⁺ T cells in the spleens and priming lymphoid tissues was performed by CDR3β length profile analysis and sequencing. Adoptive transfer of immune cells into infected SCID recipients yielded Vβ6⁺ CD8⁺ CTLs with a consistent expansion of clones utilizing TCR β-chains of the same CDR3β length. This CDR3β length distribution was uniformly found in Vβ6⁺ CD8⁺ CTLs initially primed by either oral or parenteral infection. Inspection of Jβ gene segments used in these Vβ6⁺ TCRs revealed that the route of infection influenced the repertoire of expanded T cells. Both routes of infection expanded cells that used a relatively high frequency (30% total) of Jβ1.6 and Jβ2.1. However, the remaining clones used a significantly restricted number of Jβ gene segments following oral priming.

It is reasonable to conjecture that a broader repertoire of Jβ gene segments results from prolonged antigen presentation, an increased number of epitopes presented, or an increased efficiency of antigen-presenting cells to stimulate T cells of lower antigen affinities. A longer duration or dose of antigen exposure, as might be the case following systemic infection, could lead to stimulation of CTLs that recognize epitopes with lower affinities or subdominant epitopes. Such an effect was previously described by Nelson et al. (30), who compared CTL responses to single-dose exposure to antigen and following repeated antigen exposure in adjuvants. Alternatively, the route of exposure could influence the types of antigen-presenting cells that initiate the response, resulting in different epitopes being efficiently presented. Brookes et al. (5) invoked this hypothesis after finding route-of-exposure-dependent differences in T-cell epitope hierarchy to simian immunodeficiency virus p27. Although it is likely that dendritic cells are primary activators of the immune response, heterogeneity in dendritic cell populations and the unclear phenotype and distribution of these cells in the intestine make this a plausible explanation for the results reported here. An alternative explanation is that mucosal tissues produce distinct epitopes as a result of different antigen presentation. Kuckelkorn et al. (21) described tissue-specific differences in the distribution of α- and β-chain subunits of 20S proteosomes that produce unique peptide fragments. Anatomically distinct tissues could therefore potentially process antigens differently, resulting in unique patterns of epitope display. Thus, induction of responses in the intestine could result in activation of T cells bearing subtle but distinct repertoires of TCRs. Although for reovirus infection these differences do not appear to influence the outcome of infection, our results provide evidence that the route of infection influences the repertoire of responding CTLs. These results are significant when considering rational vaccine design and approaches to manipulate immune responses through oral tolerance regimens.