α₄β₇ independent pathway for CD8⁺ T cell–mediated intestinal immunity to rotavirus

Rotavirus (RV), which replicates exclusively in cells of the small intestine, is the most important cause of severe diarrhea in young children worldwide. Using a mouse model, we show that expression of the intestinal homing integrin α₄β₇ is not essential for CD8⁺ T cells to migrate to the intestine or provide immunity to RV. Mice deficient in β7 expression (β7^–/–) and unable to express α₄β₇ integrin were found to clear RV as quickly as wild-type (wt) animals. Depletion of CD8⁺ T cells in β7^–/– animals prolonged viral shedding, and transfer of immune β7^–/– CD8⁺ T cells into chronically infected Rag-2–deficient mice resolved RV infection as efficiently as wt CD8⁺ T cells. Paradoxically, α₄β₇^hi memory CD8⁺ T cells purified from wt mice that had been orally immunized cleared RV more efficiently than α₄β₇^low CD8⁺ T cells. We explained this apparent contradiction by demonstrating that expression of α₄β₇ on effector CD8⁺ T cells depends upon the site of initial antigen exposure: oral immunization generates RV-specific CD8⁺ T cells primarily of an α₄β₇^hi phenotype, but subcutaneous immunization yields both α₄β₇^hi and α₄β₇^low immune CD8⁺ T cells with anti-RV effector capabilities. Thus, α₄β₇ facilitates normal intestinal immune trafficking to the gut, but it is not required for effective CD8⁺ T cell immunity.

Mucosal immunity provides the first level of defense against foreign antigens. Some mucosal pathogens, like rotavirus (RV) in the gut and respiratory syncytial virus in the respiratory tract, replicate at the site of entry and cause disease by local tissue damage (1). In such mucosal infections, systemic memory cells are frequently unable to prevent clinical symptoms; optimal protective immunity correlates with the presence of effector cells or local antibody at mucosal sites (1, 2).

CD8⁺ T cell responses are a major defense against viral infections at different tissue sites. The ability of lymphocytes to traffic to relevant tissues is critical for an effective immune response and is mediated by homing receptors on effector cells that have cognate ligands at peripheral or mucosal sites (3). Intestinal CD8⁺ T cell immunity, for example, has been specifically correlated with expression of α₄β₇ integrin (4). This integrin and its ligand, mucosal adressin cell adhesion molecule (MAdCAM-1) are known to play an important role in homing of activated lymphocytes to Peyer’s patches and the lamina propria (5–7). Using adoptively transferred immune CD8⁺ T cells in a murine RV (mRV) model, our earlier experiments supported the hypothesis that α₄β₇ integrin expression on CD8⁺ T cells was critical for effective intestinal immunity (8). In that study we demonstrated that α₄β₇^hi but not α₄β₇^low memory CD8⁺ T cells from wild-type (wt) mice that had been orally immunized were able to resolve chronic infection when transferred into Rag-2 deficient mice. It was unclear however, whether the ability of α₄β₇^hi CD8⁺ T cells to resolve RV infection was solely dependent upon α₄β₇ expression or whether it reflected a higher frequency of anti-RV immune CD8⁺ T cells in the α₄β₇^hi population generated following oral immunization with RV.

Determination of the function of β7 integrins has recently been facilitated by development of a β7 gene knockout (β7^–/–) mouse (9–13). These β7^–/– mice, lacking both α₄β₇ and α_Eβ₇ integrins, have dramatically reduced numbers of intestinal lymphocytes (9). Whereas α₄β₇ integrin has been implicated in lymphocyte homing to the intestine, α_Eβ₇ integrin is believed to retain CD8⁺ T cells in the intraepithelial compartment of the intestine (14). Taking advantage of the existence of β7^–/– mice, we set out to investigate the functional properties of CD8⁺ T cells lacking α₄β₇ expression, using an RV intestinal infection model. We specifically sought to determine whether such cells could localize at the site of RV infection and efficiently participate in antiviral immunity.

β7^–/– mice clear primary RV infection with the same efficiency as wt mice. Following oral infection with mRV, the level of RV shedding and anti-RV antibody secretion in the stool was determined from fecal samples collected from β7^–/– or wt (C57BL/6) mice. β7^–/– mice developed significantly less intestinal anti-RV IgA compared to wt mice following oral RV infection (Figure 1a). Despite diminished intestinal humoral immunity, β7^–/– mice were able to resolve primary RV infection at the same rate as wt animals (Figure 1b).

Figure 1

Anti-RV IgA response and virus shedding in the stool of β7^–/– and wt mice following oral infection with RV. Data were combined from six mice in each group. (a) Level of intestinal anti-RV IgA in β7^–/– or wt mice at days 10 and 30 after oral RV infection. Data were determined by ELISA (presented as ng/ml). (b) Stool RV Ag shedding from β7^–/– (open circles) and wt (filled circles) mice following oral infection with mRV. (Total Ag shedding was not significantly different between the two groups of mice.)

To determine whether resolution of primary RV infection in β7^–/– mice was CD8⁺ T-cell dependent, β7^–/– mice were depleted of CD8⁺ T cells using mAb’s (see Methods). Antibody-treated mice exhibited a substantial reduction in the number of CD8⁺ T-cells in the iIEL compartment, confirming successful depletion (Figure 2a). Accordingly, CD8⁺ T cell–depleted mice exhibited significant prolongation of RV shedding compared with nondepleted β7^–/– mice (Figure 2b). The duration of rotaviral Ag shedding was similar to that in CD8⁺ T cell–depleted wt mice (data not shown, and ref. 28). These results indicate that, as in wt mice, CD8⁺ T cells are involved in the timely resolution of primary RV infection in β7^–/– mice.

Figure 2

RV shedding of β7^–/– mice depleted of CD8⁺ T cells with mAb. The effect of anti-CD8 mAb treatment on CD8⁺ T-cell populations and rotaviral shedding is shown. (a) Flow cytometry data indicating CD8⁺ αβ T-cell populations in different tissues of β7^–/– and wt mice that were treated or untreated with anti-CD8 mAb. β7^–/– mice were depleted of CD8⁺ T cells and subsequently infected with RV as described in Methods. At day 25 from the beginning of the experiment, the mice were sacrificed and IEL, MLNs, and spleen cells were purified, stained with anti-CD8 and anti-TCRαβ antibodies, and analyzed by FACS. (b) Stool samples were collected daily following RV infection, and the levels of RV Ag were determined by ELISA. SD is based on four animals per group (the figure represents one out of two experiments performed with similar results).

Immune CD8⁺ T cells resolve RV infection when transferred from β7^–/– mice into chronically infected Rag-2 deficient mice. We used an adoptive transfer approach to evaluate the influence of α₄β₇ expression on the effector function of CD8⁺ T cells. As previously described (29), Rag-2 mice become chronically infected when orally inoculated with mRV (Figure 3a). These mice were used as recipients for CD8⁺ splenic T cells derived from β7^–/– or wt mice that had previously cleared RV. The ability of adoptively transferred CD8⁺ T cells to resolve chronic RV infection was monitored by measuring daily RV shedding in stools of recipient Rag-2 mice. Adoptive transfer of 5 × 10³ immune β7^–/– or wt CD8⁺ T cells into infected Rag-2 mice failed to clear RV infection (Figure 3, b and c); clearance was variable when 5 × 10⁴ immune CD8 ⁺ T cells were transferred (Figure 3, d and e), but all Rag-2 recipient mice given 5 × 10⁵ β7^–/– or wt CD8 ⁺ T cells resolved infection (Figure 3, f and g). No difference in the duration of shedding was observed between Rag-2 mice given β7^–/– or wt immune CD8⁺ T cells at each cell number injected. These findings indicate a quantitative threshold for RV clearance by immune CD8⁺ T cells that was independent of α₄β₇ expression.

Figure 3

Fecal rotaviral Ag shedding in Rag-2 mice chronically infected with RV following adoptive transfer of varying numbers of immune β7^–/– or wt CD8⁺ splenic T cells. Donor mice were immunized orally as described in Methods, and spleen cells from these donor mice were harvested 30 days after immunization. (a) Ag shedding in RV-infected Rag-2 mice not transferred with cells. (b and c) RV Ag shedding of Rag-2 mice transferred with 5 × 10³ CD8⁺ T cells. (d and e) RV Ag shedding of Rag-2 mice transferred with 5 × 10⁴ CD8⁺ T cells. (f and g) RV Ag shedding of Rag-2 mice transferred with 5 × 10⁵ CD8⁺ T cells. Asterisks represent RV Ag shedding in chronically infected Rag-2 mice that did not receive immune cells (a). Open circles represent RV shedding in Rag-2 mice adoptively transferred with β7^–/– CD8⁺ T cells (b, d, and f). Filled circles represent RV shedding in Rag-2 mice adoptively transferred with wt CD8⁺ T cells (panels c, e, and g).

Using flow cytometry, CD19⁺ (B cells) and CD4⁺ cells were not detected in the Rag-2 mice following transfer with immune CD8⁺ T cells, confirming that the donor cells were not contaminated (data not shown). The absence of B cell contamination was further supported by the observation that anti-RV IgA remained undetectable in the stool of Rag-2 mice 30 days after CD8⁺ T-cell transfer (data not shown). Hence we were unable to show a quantitative or qualitative difference in the effector function of β7^–/– versus wt CD8⁺ T cells.

α₄β₇^–/– CD8⁺ T cells can migrate to the IEL compartment of chronically infected Rag-2 mice. To determine whether the α₄β₇ deficient CD8⁺ T cells migrated into the intestinal epithelium, the percentage of CD8⁺ TCRαβ⁺ cells in the iIEL compartment of Rag-2 mice was evaluated at day 30 following adoptive transfer (Figure 4). For comparison, the percentage of TCRαβ⁺ CD8⁺ cells in the spleen and MLNs was also determined since lymphocyte migration to these organs is not dependent on α₄β₇ integrin expression.

Figure 4

Flow cytometric analysis of IEL, MLNs, and spleen cells from RV-infected Rag-2 mice 30 days after transfer of 5 × 10⁵ β7^–/– or wt CD8⁺ T cells. Cells from Rag-2 and wt (C57BL/6) mice served as controls. Cells were double-stained with anti-CD8 FITC and anti-TCRαβ PE or with anti-CD8 FITC and anti-α₄β₇ PE as described in Methods. Percentage of cells in the quadrant of interest is indicated.

As anticipated, wt mice possessed CD8⁺ T cells in all three compartments (iIEL, MLNs, and spleen) while the treated Rag-2 mice had virtually no detectable TCRαβ⁺ cells in any of these compartments. In contrast, on day 30 following RV clearance, CD8⁺ αβ ⁺ donor T cells were detected in iIEL, spleen, and MLNs of Rag-2 mice regardless of whether they expressed α₄β₇ integrin (Figure 4).

A somewhat lower percentage of lymphocytes was recovered from the iIEL compartment of chronically infected Rag-2 mice reconstituted with purified CD8⁺ TCRαβ⁺ cells from β7^–/– mice (14.7%) than from wt mice (34.0%) (Figure 4). The absolute number of cells recovered from iIEL of chronically infected Rag-2 mice at day 30 following adoptive transfer of β7^–/– or wt CD8⁺ T cells was similar (in the range of 2–6 × 10⁶ cells per mouse). When the percentages of CD8⁺ T cells in iIELs from six animals per group were compared (wt versus β7^–/–) the differences were not statistically significant (P > 0.05) (using Student t test) (see Figure 7d).

Figure 7

Percentage of CD8⁺ TCRαβ⁺ cells in IELs, MLNs, and spleens of mice transferred with β7^–/– or wt CD8⁺ T cells at indicated times after transfer. (a) Percentage of CD8⁺ TCRαβ⁺ FITC-labeled cells. Splenocytes from nonimmune β7^–/– or wt mice were labeled in vitro with FITC and 10⁷ cells were injected intravenously into C57BL/6 recipient mice. Three hours following injection of the labeled cells, recipient mice were sacrificed and percentages of CD8⁺ T cells in the indicated tissues were determined by FACS. (b, c, and d) Chronically infected Rag-2 mice were adoptively transferred with 5 × 10⁵ CD8⁺ T cells from RV immune β7^–/– or wt mice. Seven (b), 14 (c), and 30 (d) days after transfer, recipient mice were sacrificed and cells from IEL, MLNs, and spleen were analyzed by FACS. Each group consisted of six animals. Differences in percentages of CD8⁺ T cells between groups (each comprising six animals) were not statistically significant using standard t test (P > 0.05). Note that the scales in a and b are different because fewer CD8⁺ T cells were detected at 3 hours than at 7 days following adoptive transfer.

α₄β₇^–/–(Thy1.2) CD8⁺ T cells can migrate to the intestine when α₄β₇ positive wt (Thy1.1) immune CD8⁺ T cells are present. To test whether migration into the intestinal mucosa of α₄β₇ integrin–deficient CD8⁺ T cells occurred independently of α₄β₇ ligand binding, we examined the competition for iIEL entry between α₄β₇-deficient and wt CD8⁺ T cells. We used wt (Thy1.1) and β7^–/– (Thy1.2) mice as donors. Thirty days following RV infection of wt (Th1.1) or β7^–/– (Th1.2) mice splenic CD8⁺ T cells were affinity-purified and double FACS sorted (purity ≥ 99.8%). A mixture of 5 × 10⁵ β7^–/– and 5 × 10⁵ wt CD8⁺ T cells was adoptively transferred into the same Rag-2 recipients. The recipients were sacrificed at 14 days following adoptive transfer (when viral infection had cleared from the intestine) and the percentages of β7^–/– or wt CD8⁺ T cells in the iIEL were determined using flow cytometry. In all adoptively transferred Rag-2 mice we observed distinct populations of β7^–/– and wt CD8⁺ T cells in both iIEL and MLN compartments (Figure 5, a and b). The percentage of β7^–/– T cells in the iIEL was ten times lower than wt CD8⁺ T cells (Figure 5a). When five times more wt than β7^–/– T cells were transferred, we were still able to detect α₄β₇ deficient CD8⁺ T cells in the iIEL (percentages were from 15- to 20-fold lower than wt T cells, data not shown). These findings suggest that α₄β₇-deficient T cells were able to migrate to iIEL even in the presence of competing wt CD8⁺ T cells, consistent with the notion that they were trafficking into the iIEL via an alternative recognition signal.

Figure 5

FACS analyses of IEL and MLN cells purified from RV-infected Rag-2 mice at day 14 following adoptive transfer of 5 × 10⁵ β7^–/– (Thy1.2) and 5 × 10⁵ wt (Thy1.2) CD8⁺ T cells. IEL and MLN cells from untreated Rag-2 mice served as controls. (a) Percentage of wt (Thy1.1) or β7^–/– (Thy1.2) CD8⁺ T cells in the IEL. (b) Percentage of wt (Thy1.1) or β7^–/– (Thy 1.2) CD8⁺ T cells in the MLNs.

Histological analysis of intestinal tissues from RV-challenged Rag-2 mice in the presence or absence of adoptively transferred β7^–/– or wt CD8⁺ T cells. Immunohistological examination of intestinal tissue from adoptively transferred Rag-2 mice was performed to determine the location of adoptively transferred lymphocytes. RV-infected Rag-2 mice, reconstituted with immune CD8⁺ T cells from β7^–/– or wt mice, were sacrificed at day 30 after adoptive transfer, and sections of ileum were used for histological analysis. Intestinal sections from wt and untreated Rag-2 mice served as controls (Figure 6, a and b). CD3⁺ cells were detected in the iIEL compartment of wt mice (Figure 6a) but, as expected, not in mice that were Rag-2 (Figure 6b). On the other hand, Rag-2 mice that received wt or β7^–/– CD8⁺ T cells demonstrated CD3⁺ cells in both IEL and lamina propria (LP) compartments of the intestine (Figure 6, c and d).

Figure 6

Immunohistological analyses of intestinal tissue sections obtained from the ilea of uninfected wt mice, uninfected Rag-2 mice, and chronically infected Rag-2 mice transferred with 5 × 10⁵ CD8⁺ T cells from RV-immune wt or β7^–/– mice. The samples were stained with biotinylated anti-mouse CD3 followed by streptavidin-conjugated Texas red and were analyzed using confocal microscopy. (a) wt mouse (uninfected). (b) Rag-2 untreated mouse (uninfected). (c) Chronically infected Rag-2 mouse transferred with 5 × 10⁵ wt CD8⁺ T cells. (d) Chronically infected Rag-2 mouse transferred with 5 × 10⁵ β7^–/– CD8⁺ T cells.

The time of reconstitution of the IEL compartment with β7^–/– or wt CD8⁺ T cells correlates with resolution of RV shedding. To follow the migration of the transferred β7^–/– or wt CD8⁺ T cells in chronically infected Rag-2 mice, recipients were sacrificed at different time points after cell transfer and IEL, MLNs, and spleen cells were analyzed by flow cytometry (Figure 7). At 3 hours (Figure 7a) and 3 days (data not shown) after transfer, the highest percentage of transferred CD8⁺ αβ⁺ T cells was found in the spleen (and blood, data not shown). By 7 days after transfer, the highest percentage of transferred immune cells was detected in the MLNs, but none were detected in the iIEL (Figure 7b). Correspondingly, rotaviral replication in the intestine continued during the same period (Figure 3, f and g). By day14, β7^–/– and wt CD8⁺ αβ⁺ T cells were detected in the IEL of the respective recipient Rag-2 mice (Figure 7, c and d). Correspondingly, the mice resolved rotaviral infection. Thus, the presence of CD8⁺ T cells in the IEL compartment correlated temporarily with rotaviral clearance independently of α₄β₇ expression.

The iIEL compartment remained populated with CD8⁺ T cells through 30 days following cell transfer. The percentage of β7^–/– CD8⁺ T cells in the iIEL of recipient Rag-2 mice remained lower relative to the wt CD8⁺ T cells. Differences in percentages of CD8⁺ T cells in iIEL between groups (each composed of six animals) were not statistically significant (P > 0.05), however (Figure 7, c and d).

Oral immunization with live mRV generates RV-specific CD8⁺ T cells in the α₄β₇^hi memory cell population whereas subcutaneous administration of inactivated RRV generates anti-RV CD8⁺ effectors with both α₄β₇^low and α₄β₇^hi phenotypes. The experiments described above demonstrate that α₄β₇ expression is not a requirement for CD8⁺ T cell–mediated immunity against RV in the intestinal mucosa. However, in an earlier study we showed that α₄β₇^hi CD8⁺ splenic T cells from orally infected wt mice resolved chronic RV infection more efficiently in Rag-2 mice than α₄β₇^low cells (8). Given our new findings with β7^–/– mice, we postulated that the reason we had not observed effective anti-rotaviral immunity with α₄β₇^low spleen cells (see Figure 10a and ref. 8) might be due to the presence of fewer RV-specific CD8⁺ T cells in α₄β₇^low versus α₄β₇^hi subsets following oral immunization rather than a low level of intestinal homing integrin expression in the α₄β₇^low population. Therefore we quantified the number of RV-specific CD8⁺ T cells in the splenic populations of donor mice. These were the cells used for adoptive transfer into RV infected Rag-2 mice. We specifically determined the frequency of anti-RV splenic CD8⁺ T cells in α₄β₇^hi and α₄β₇^low populations following oral versus subcutaneous immunization with RV.

Figure 10

Wt mice (C57BL/6) were immunized twice at a 15-day interval either subcutaneously or orally with RV as described. Thirty days following the last immunization, the memory CD8⁺ T cells from the spleen were separated by FACS into α₄β₇^hi or α₄β₇^low fractions and subsequently injected into chronically infected Rag-2 mice. The data represent one of two experiments performed with similar results. (a) Stool Ag shedding of RV-infected Rag-2 mice adoptively transferred with 30,000 α₄β₇^low memory CD8⁺ T cells purified from mice orally immunized with live RV. (b) Virus shedding of RV-infected Rag-2 mice adoptively transferred with 30,000 α₄β₇^low memory CD8⁺ T cells purified from mice systemically immunized with inactivated RV. (c) Stool Ag shedding of RV-infected Rag-2 mice adoptively transferred with 10,000 α₄β₇^hi memory CD8⁺ T cells purified from mice orally immunized with live RV. (d) Stool Ag shedding of RV-infected Rag-2 mice adoptively transferred with 10,000 α₄β₇^hi memory CD8⁺ T cells purified from mice systemically immunized with inactivated RV. Circles indicate RV shedding of Rag-2 mice transferred with CD8⁺ T cells from orally immunized donors. Squares indicate RV shedding of Rag-2 mice transferred with CD8⁺ T cells from systemically primed donors. Filled symbols indicate Ag shedding of Rag-2 mice adoptively transferred with α₄β₇^hi CD8⁺ T cells. Open symbols indicate Ag shedding of Rag-2 mice adoptively transferred with α₄β₇^low CD8⁺ T cells.

Wt mice were immunized orally with live mRV or subcutaneously with inactivated RRV in CFA and IFA (as described in Methods). Systemic immunization was employed with the hope that RV immune CD8⁺ T cells might be induced efficiently in the α₄β₇^low population. For systemic immunizations, inactivated RRV was used because this virus, in contrast to mRV, does not spread to the intestine when administered subcutaneously. To further eliminate this possibility, we chemically treated and UV-inactivated the RRV. Successful inactivation was confirmed by demonstrating lack of viral replication in the inactivated RRV preparations when cultured in vitro (data not shown).

Two approaches were adopted to determine the frequency of RV-specific CD8⁺ T cells in the immunized mice: ELISPOT assay and staining of Ag-specific CD8⁺ T cells using an MHC class I tetramer. Thirty days following immunization, we determined the frequency of IFN-γ SF CD8⁺ T cells (23) from systemically or orally immunized mice after in vitro restimulation. Oral immunization generated 26 IFN-γ SF CD8 ⁺ T cells/10⁴ α₄β₇^hi CD8⁺ T cells, whereas only two IFN-γ SF CD8⁺ T cells were detected in the α₄β₇^low population (Figure 8). On the other hand, systemic administration with inactivated RV generated IFN-γ–producing cells at a frequency of 4 and 10 SF cells/1 × 10⁴ CD8⁺ T cells in the α₄β₇^hi and α₄β₇^low populations, respectively (Figure 8).

Figure 8

IFN-γ ELISPOT assay. RV-responsive cells are mainly in α₄β₇^hi population of orally but not subcutaneously immunized mice. Splenocytes from C57BL/6 mice immunized orally or subcutaneously with RV were enriched for CD8⁺ T cells using affinity column purification. Subsequently, the cells were FACS-sorted into α₄β₇^hi- or α₄β₇^low-expressing CD44⁺ (memory) αβ TCR CD8⁺ T cells. The sorted cells were restimulated in vitro for 72 hours, and the frequency of IFN-γ SFCs was determined by ELISPOT. The figure represents one out of three experiments performed with similar results. Five mice were used per group in each experiment. The SD is based on three replicates per group.

The high ratio of α₄β₇^hi to α₄β₇^low RV-specific CD8⁺ T cells (26:2) was confirmed using class I restricted tetramer staining (Figure 9). A higher percentage of RV-specific CD8⁺ T cells was found in the α₄β₇^hi fraction of orally immunized animals (0.16%), consistent with our previous ELISPOT results, whereas very few RV-specific CD8⁺ T cells were detected in the α₄β₇^low fraction (0.03%). In contrast, both assays demonstrated that subcutaneous immunization generated RV-specific CD8⁺ T cells in both α₄β₇^hi and α₄β₇^low populations (0.23% and 0.22% by tetramer analysis [Figure 8] and 4 ± 2 and 10 ± 4 by ELISPOT analysis [Figure 9]).

Figure 9

Tetramer staining for RV-specific CD8⁺ T cells. RV-responsive cells are mainly in α₄β₇^hi population of orally but not subcutaneously immunized mice. Wt mice (C57BL/6) were immunized twice at a 15-day interval either subcutaneously or orally with RV as described. Thirty days following the last immunization, the splenocytes from immune mice were stained with anti-CD8 CyChrome; anti-CD19, anti-CD13, and anti-CD4 FITC; anti-α₄β₇ PE; and class I tetramer APC. Small lymphocytes were negatively gated on FITC-stained cells and positively gated on anti-CD8 CyChrome–stained cells. The gated cells were analyzed for α₄β₇ PE expression and tetramer-APC binding. Splenocytes from nonimmunized mice were stained simultaneously and were used as negative control. As negative control we also used APC-conjugated class I tetramer specific for irrelevant Ag. The gates for α₄β₇-negative cells and α₄β₇-expressing cells were set using splenocytes from β7^–/– mice. The data represent one out of three experiments performed with similar results. Three mice per group were used in each experiment.

α₄β₇^low CD8⁺ T cells from subcutaneously immunized wt donor mice can eliminate RV infection from chronically infected Rag-2 recipients. We next tested the in vivo ability of α₄β₇^hi and α₄β₇^low CD8⁺ T cells purified from subcutaneously versus orally immunized mice to clear RV when transferred into chronically infected Rag-2 mice. Initially we confirmed our previous studies (8) demonstrating that 10,000 purified α₄β₇^hi CD8⁺ T cells from orally immunized wt mice efficiently resolved chronic RV infection when transferred into Rag-2 mice (Figure 10c), whereas 30,000 α₄β₇^low CD8⁺ T cells (threefold more) from the same donors could not resolve infection (Figure 10a). In contrast, 30,000 α₄β₇^low CD8⁺ T cells transferred from subcutaneously immunized wt mice efficiently resolved RV infection in Rag-2 mice (Figure 10b). Although both α₄β₇^low and α₄β₇^hi cell populations underwent homeostatic expansion in the recipient Rag-2 mice (similar number of cells were recovered from the spleen MLNs and iIEL compartments, data not shown), only α₄β₇^hi CD8⁺T cells were able to resolve RV infection (Figure 10b). These results demonstrate that subcutaneous RV immunization generated protective anti-RV CD8⁺ T cells in the α₄β₇^low population.

This study demonstrates that α₄β₇ integrin expression on memory CD8⁺ T cells is not essential for those cells to migrate to the gut and resolve RV infection. Our results suggest the existence of a β7-independent mechanism of CD8⁺ T cell migration to and/or retention in the intestine. This alternative mechanism appears sufficient to direct and/or retain CD8⁺ TCRαβ⁺ cells to intestinal effector sites in the absence of both α₄β₇ and α_Eβ₇ integrin expression. We also provide evidence that oral immunization with live RV generates Ag-specific CD8⁺ T cells almost exclusively in the α₄β₇^hi population, whereas subcutaneous RV administration induces a more divided CD8⁺ T cell immune response in both α₄β₇^hi and α₄β₇^low cell fractions.

Previous studies have specifically correlated intestinal lymphocyte homing with expression of the α₄β₇ integrin. In vitro and short-term in vivo experiments have shown the importance of α₄β₇ in lymphocyte homing to Peyer’s patches (PP) and intestinal LP (3, 5–7, 30, 31). In vivo experiments demonstrated more efficient homing of α₄β₇^hi versus α₄β₇^low memory T cells to PP (7). A recent report using gene targeting to disrupt β₇ integrins led to drastically impaired formation of the gut associated lymphoid tissue, due to the inability of β₇ integrin deficient lymphocytes to efficiently migrate through the high endothelial venules into this compartment (9).

Here, using β7-deficient animals and intestinal RV infection, we assessed the role played by α₄β₇ integrin expression in promoting CD8⁺ T cell effector function in the gut. Because in mice, as well as in humans, RV replication is localized to the villus tips of the intestinal epithelium, the RV mouse model is particularly helpful in identifying key components of mucosal immune effector function.

Consistent with the observation that β7^–/– mice have a defective gut inductive immune system (9), β7^–/– mice orally infected with RV generated lower levels of intestinal anti-RV IgA compared to wt controls (Figure 1). Despite diminished antiviral humoral responses, β7^–/– mice cleared RV from the intestine at the same time as wt animals (day 7 following infection), suggesting the efficient involvement of class I mediated T-cell immunity. β7^–/– mice depleted of CD8⁺ T cells were able to clear RV days 10 after infection just like CD8⁺ T cell–depleted wt mice (28). The prolongation of RV shedding in CD8⁺ T cell–depleted versus nondepleted β7^–/– mice indicates that β7^–/– CD8⁺ T cells contribute to the timely clearance of RV during primary infection despite the absence of α₄β₇ integrin expression. Although fewer CD8⁺ T cells are found in the iIEL of β7^–/– compared with wt mice (Figure 2a), these cells appear to be sufficient to hasten the elimination of virus from the intestinal epithelium during primary infection (Figure 2b). It is possible that during primary infection β7^–/– CD8⁺ T cells already present in the intestine become primed locally and subsequently clear RV infection.

To directly evaluate the anti-RV function of CD8⁺ T cells lacking α₄β₇ and α_Eβ₇, we used an adoptive transfer mouse model. We used β7^–/– mice as a source of immune CD8⁺ T cells and compared their anti-RV effector function with that of wt CD8⁺ T cells after passive transfer. We injected log order differences (5 × 10³, 5 × 10⁴, and 5 × 10⁵) of β7^–/– or wt CD8⁺ TCRαβ⁺ cells into chronically infected Rag-2 mice and compared the kinetics and efficiency of viral clearance. We could not detect a difference either in the level or duration of RV Ag shedding between chronically infected Rag-2 mice injected with β7^–/– or wt CD8⁺ T cells (Figure 3). Although a small difference in efficiency could have been missed in our analysis, β7^–/– CD8⁺ T cells did not appear to be substantially less efficient than wt cells in mediating antiviral function in the intestine.

We purified iIELs of the recipient Rag-2 mice at different time points following CD8⁺ T cell transfer and analyzed them by flow cytometry to determine whether the injected β7^–/– CD8⁺ T cells were actually migrating to the site of viral replication. CD8⁺ TCRαβ⁺ T cells were found in the iIEL compartment of the recipient Rag-2 mice at time points synchronous with, but not before, virus elimination. At 3 hours, 3 days, and 7 days after cell transfer, CD8⁺ T cells were found in both spleen and MLNs, but, if present in iIEL compartment, their frequency was too low to resolve RV infection or to be detected by flow cytometry (Figure 7). Recovery of CD8⁺ T cells from the iIEL compartment correlated temporally with viral clearance from the intestine. β7^–/– CD8⁺ T cells could still migrate to the intestine even when adoptively transferred simultaneously with wt CD8⁺ T cells into the same recipients (Figure 5). This observation indicated that α₄β₇-deficient CD8⁺ T cells were not blocked by wt CD8⁺ T cells migrating into the intestine and further supports the existence of an α₄β₇-independent mechanism of CD8⁺ T cell migration into the gut epithelium.

Our findings are consistent with a recent report using class I restricted ovalbumin (OVA) transgenic mice and systemic infection with recombinant Vesicular Stomatitis Virus expressing OVA (rVSV OVA) to study the role of the α₄β₇ and α_Eβ₇ integrins in intestinal homing (13). This study also identified some residual ability of β7^–/– cells to migrate to the gut. It was unclear, however, whether the residual ability of β7^–/– memory cells to migrate to effector sites would be sufficient to exert an immune effector function during intestinal infection. In the present study, using a common gut pathogen, we demonstrated that β7^–/– CD8⁺ T cells, when compared with wt CD8⁺ T cells, are able to mediate anti-RV activity in the intestine with similar efficiency during primary infection and adoptive transfer (Figures 1, 2, and 3).

The successful function of β7^–/– CD8⁺ T cells in the gut, as demonstrated here, provides further evidence for redundancy within the immune system and points to the existence of an alternative mechanism of CD8⁺ T cell trafficking to and/or retention at intestinal effector sites. Molecules other than β7 seem likely to be involved in directing CD8⁺ T cells to RV-infected intestinal epithelium. For example, β2 integrins have been suggested as candidates for directing intestinal CD8⁺ T cells (32), a possibility we are currently investigating. This would be consistent with the multistep model of lymphocyte homing in which α₄β₇ integrins initiate gut homing and are thought to have serial but partially overlapping functions in vascular endothelial recognition (33).

Recently it has been demonstrated that intestinal RV infection in young mice can induce expression of various chemokines by intestinal epithelial cells (34). A CC chemokine, thymus expressed chemokine (TECK), and its receptor CCR9 (predominantly expressed on memory α₄β₇⁺, CD4⁺, and CD8⁺ T cells) are selectively expressed in the small intestine, implying a potential role in intestinal immunity (35). It is possible that RV-induced chemokines could additionally activate and recruit β7^–/– CD8⁺ T cells in the intestine, facilitating their local expansion. Future studies will be designed to explore this possibility.

In contrast to the ability of β7^–/– and wt CD8⁺ T cells to resolve RV infection when injected into chronically infected Rag-2 mice, α₄β₇^low CD8⁺ T cells purified from orally immunized immunocompetent mice were much less efficient than α₄β₇^hi CD8⁺ T cells at clearing RV infection (8). Based on our results with β7^–/– mice, we hypothesized that this difference might not reflect a requirement for α₄β₇ expression on memory CD8⁺ T cells, but rather resulted from selective enrichment of RV-specific memory CD8⁺ T cells in the α₄β₇^hi subset after oral immunization. In this study we provided evidence that the α₄β₇^low population of CD8⁺ T cells purified from orally immunized mice was not protective when transferred into RV-infected Rag-2 mice because it contained very few RV-specific immune CD8⁺ T cells (Figures 8 and 9). We were able to increase the frequency of RV-specific CD8⁺ T cells in the α₄β₇^low population by subcutaneous immunization with inactivated RV. Using this immunization route, we observed that 30,000 purified memory α₄β₇^low CD8⁺ TCRαβ⁺ T cells were able to resolve RV infection when injected into chronically infected Rag-2 mice, whereas the opposite was observed after adoptive transfer of α₄β₇^hi CD8⁺ TCRαβ⁺ T cells from orally immunized animals.

To determine the frequency of RV-specific CD8⁺ T cells in the α₄β₇^hi and α₄β₇^low populations of orally and systemically immunized animals, we used two assays: RV (VP7) specific class I restricted tetramer staining and Ag-specific IFN-γ ELISPOT assay. Both approaches demonstrated that oral immunization generated anti-RV CD8⁺ T cell immunity almost exclusively in the α₄β₇^hi population. A substantial increase in the frequency of RV-specific CD8⁺ T cells in the α₄β₇^low fraction was achieved when inactivated RRV was injected systemically. Hence the principle reason α₄β₇^low CD8⁺ T cells from orally immunized mice fail to eradicate chronic RV infections is that they contain few, if any, RV-specific memory cells. These data are consistent with our previous study in humans demonstrating that anti RV CD4⁺ T cell immunity also resides primarily in the α₄β₇^hi population (36). In addition, we have shown that oral immunization of mice with live RV generates RV-specific Ig producing cells (measured by RV-specific ELISPOT) primarily in the α₄β₇^hi B cell populations (37).

It is widely assumed that systemic immunity is best induced by parenteral immunization, whereas mucosal responses are generated as a result of Ag exposure at mucosal surfaces. Our data support this notion. However we also demonstrate that systemic (subcutaneous), but not oral immunization can prime protective anti-CD8⁺ T cell immunity in α₄β₇^low, as well as α₄β₇^hi, populations. Hence, induction of immune α₄β₇^hi and α₄β₇^low CD8⁺ T cells does occur after systemic immunization. On the other hand, oral immunization strongly favors the induction of α₄β₇^hi immune CD8⁺ T cells with little immunity residing in the α₄β₇^low population. Surprisingly, local RV immune effector functions can be mediated by either fraction of CD8⁺ T cells. Whether this is true for other CD8⁺ T cell–mediated local anti-microbial effects remains to be determined. It will also be important to determine if B cells with low or absent α₄β₇ expression can mediate anti-RV effects, since the ability to induce efficient humoral mucosal effector function following systemic immunization would widen the possibility for the design of vaccines against mucosal pathogens.

We thank Stewart L. Cooper for constructive editorial comments and helpful suggestions. We thank Sally Morefield for secretarial assistance. This work was supported by NIH grant R37AI21362, a Veterans Administration Merit Review grant to H.B. Greenberg, and NIH training grant 5 T32 AI07328-12 to N.A. Kuklin.

1. Ahmed, R, Gray, D. Immunological memory and protective immunity: understanding their relation. Science 1996. 272:54-60.
   View this article via: PubMed CrossRef Google Scholar
2. Murphy, B.R. 1996. Immunization against virus disease. In Virology. Volume 1. B.N. Fields, editor. Lippincott-Raven Publishers. Philadelphia, Pennsylvania, USA. 15–59.
   View this article via: PubMed Google Scholar
3. Butcher, EC, Williams, M, Youngman, K, Rott, L, Briskin, M. Lymphocyte trafficking and regional immunity. Adv Immunol 1999. 72:209-253.
   View this article via: PubMed CrossRef Google Scholar
4. Rose, J, Franco, M, Greenberg, H. The immunology of rotavirus infection in the mouse. Adv Virus Res 1998. 51:203-235.
   View this article via: PubMed CrossRef Google Scholar
5. Holzmann, B, McIntyre, BW, Weissman, IL. Identification of a murine Peyer’s patch: specific lymphocyte homing receptor as an integrin molecule with an alpha chain homologous to human VLA-4 alpha. Cell 1989. 56:37-46.
   View this article via: PubMed CrossRef Google Scholar
6. Hamann, A, Andrew, DP, Jablonski, WD, Holzmann, B, Butcher, EC. Role of alpha 4-integrins in lymphocyte homing to mucosal tissues in vivo. J Immunol 1994. 152:3282-3293.
   View this article via: PubMed Google Scholar
7. Williams, MB, Butcher, EC. Homing of naive and memory T lymphocyte subsets to Peyer’s patches, lymph nodes, and spleen. J Immunol 1997. 159:1746-1752.
   View this article via: PubMed Google Scholar
8. Rose, JR, Williams, MB, Rott, LS, Butcher, EC, Greenberg, HB. Expression of the mucosal homing receptor alpha4beta7 correlates with the ability of CD8+ memory T cells to clear rotavirus infection. J Virol 1998. 72:726-730.
   View this article via: PubMed Google Scholar
9. Wagner, N, et al. Critical role for beta7 integrins in formation of the gut-associated lymphoid tissue. Nature 1996. 382:366-370.
   View this article via: PubMed CrossRef Google Scholar
10. Steeber, DA, et al. Efficient lymphocyte migration across high endothelial venules of mouse Peyer’s patches requires overlapping expression of L-selectin and beta7 integrin. J Immunol 1998. 161:6638-6647.
    View this article via: PubMed Google Scholar
11. Wagner, N, et al. L-selectin and beta7 integrin synergistically mediate lymphocyte migration to mesenteric lymph nodes. Eur J Immunol 1998. 28:3832-3839.
    View this article via: PubMed CrossRef Google Scholar
12. Kunkel, EJ, et al. The roles of L-selectin, beta 7 integrins, and P-selectin in leukocyte rolling and adhesion in high endothelial venules of Peyer’s patches. J Immunol 1998. 161:2449-2456.
    View this article via: PubMed Google Scholar
13. Lefrancois, L, et al. The role of beta7 integrins in CD8 T cell trafficking during an antiviral immune response. J Exp Med 1999. 189:1631-1638.
    View this article via: PubMed CrossRef Google Scholar
14. Cepek, KL, et al. Adhesion between epithelial cells and T lymphocytes mediated by E-cadherin and the alpha E beta 7 integrin. Nature 1994. 372:190-193.
    View this article via: PubMed CrossRef Google Scholar
15. Burns, JW, et al. Analyses of homologous rotavirus infection in the mouse model. Virology 1995. 207:143-153.
    View this article via: PubMed CrossRef Google Scholar
16. Hoshino, Y, Wyatt, RG, Greenberg, HB, Flores, J, Kapikian, AZ. Serotypic similarity and diversity of rotaviruses of mammalian and avian origin as studied by plaque-reduction neutralization. J Infect Dis 1984. 149:694-702.
    View this article via: PubMed Google Scholar
17. Groene, WS, Shaw, RD. Psoralen preparation of antigenically intact noninfectious rotavirus particles. J Virol Methods 1992. 38:93-102.
    View this article via: PubMed CrossRef Google Scholar
18. Offit, PA, Clark, HF, Plotkin, SA. Response of mice to rotaviruses of bovine or primate origin assessed by radioimmunoassay, radioimmunoprecipitation, and plaque reduction neutralization. Infect Immun 1983. 42:293-300.
    View this article via: PubMed Google Scholar
19. Kuklin, N, Daheshia, M, Karem, K, Manickan, E, Rouse, BT. Induction of mucosal immunity against herpes simplex virus by plasmid DNA immunization. J Virol 1997. 71:3138-3145.
    View this article via: PubMed Google Scholar
20. Butcher, EC, Weissman, IL. Direct fluorescent labeling of cells with fluorescein or rhodamine isothiocyanate. I. Technical aspects. J Immunol Methods 1980. 37:97-108.
    View this article via: PubMed CrossRef Google Scholar
21. Taguchi, T, et al. Novel function for intestinal intraepithelial lymphocytes. Murine CD3+, gamma/delta TCR+ T cells produce IFN-gamma and IL-5. J Immunol 1991. 147:3736-3744.
    View this article via: PubMed Google Scholar
22. Franco, MA, Greenberg, HB. Role of B cells and cytotoxic T lymphocytes in clearance of and immunity to rotavirus infection in mice. J Virol 1995. 69:7800-7806.
    View this article via: PubMed Google Scholar
23. Kuklin, NA, Daheshia, M, Chun, S, Rouse, BT. Immunomodulation by mucosal gene transfer using TGF-beta DNA. J Clin Invest 1998. 102:438-444.
    View this article via: JCI PubMed CrossRef Google Scholar
24. Nair, S, Zhou, F, Reddy, R, Huang, L, Rouse, BT. Soluble proteins delivered to dendritic cells via pH-sensitive liposomes induce primary cytotoxic T lymphocyte responses in vitro. J Exp Med 1992. 175:609-612.
    View this article via: PubMed CrossRef Google Scholar
25. Altman, JD, et al. Phenotypic analysis of antigen-specific T lymphocytes [erratum 1998, 280:1821]. Science 1996. 274:94-96.
    View this article via: PubMed CrossRef Google Scholar
26. Franco, MA, et al. An immunodominant cytotoxic T cell epitope on the VP7 rotavirus protein overlaps the H2 signal peptide. J Gen Virol 1993. 74:2579-2586.
    View this article via: PubMed CrossRef Google Scholar
27. Garboczi, DN, Hung, DT, Wiley, DC. HLA-A2-peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides. Proc Natl Acad Sci USA 1992. 89:3429-3433.
    View this article via: PubMed CrossRef Google Scholar
28. Franco, MA, Tin, C, Greenberg, HB. CD8+ T cells can mediate almost complete short-term and partial long-term immunity to rotavirus in mice. J Virol 1997. 71:4165-4170.
    View this article via: PubMed Google Scholar
29. Dharakul, T, Rott, L, Greenberg, HB. Recovery from chronic rotavirus infection in mice with severe combined immunodeficiency: virus clearance mediated by adoptive transfer of immune CD8+ T lymphocytes. J Virol 1990. 64:4375-4382.
    View this article via: PubMed Google Scholar
30. Bargatze, RF, Jutila, MA, Butcher, EC. Distinct roles of L-selectin and integrins alpha 4 beta 7 and LFA-1 in lymphocyte homing to Peyer’s patch-HEV in situ: the multistep model confirmed and refined. Immunity 1995. 3:99-108.
    View this article via: PubMed CrossRef Google Scholar
31. Picarella, D, et al. Monoclonal antibodies specific for beta 7 integrin and mucosal addressin cell adhesion molecule-1 (MAdCAM-1) reduce inflammation in the colon of scid mice reconstituted with CD45RBhigh CD4+ T cells. J Immunol 1997. 158:2099-2106.
    View this article via: PubMed Google Scholar
32. Huleatt, JW, Lefrancois, L. Beta2 integrins and ICAM-1 are involved in establishment of the intestinal mucosal T cell compartment. Immunity 1996. 5:263-273.
    View this article via: PubMed CrossRef Google Scholar
33. Butcher, EC, Picker, LJ. Lymphocyte homing and homeostasis. Science 1996. 272:60-66.
    View this article via: PubMed CrossRef Google Scholar
34. Rollo, EE, et al. The epithelial cell response to rotavirus infection. J Immunol 1999. 163:4442-4452.
    View this article via: PubMed Google Scholar
35. Zaballos, A, Gutierrez, J, Varona, R, Ardavin, C, Marquez, G. Cutting edge: identification of the orphan chemokine receptor GPR-9-6 as CCR9, the receptor for the chemokine TECK. J Immunol 1999. 162:5671-5675.
    View this article via: PubMed Google Scholar
36. Rott, LS, et al. Expression of mucosal homing receptor alpha4beta7 by circulating CD4+ cells with memory for intestinal rotavirus. J Clin Invest 1997. 100:1204-1208.
    View this article via: JCI PubMed CrossRef Google Scholar
37. Williams, MB, et al. The memory B cell subset responsible for the secretory IgA response and protective humoral immunity to rotavirus expresses the intestinal homing receptor, alpha4beta7. J Immunol 1998. 161:4227-4235.
    View this article via: PubMed Google Scholar