INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The mucosal immune system provides the first line of defense against pathogens that invade at the wet epithelial surfaces of the body (28). These sites include the gastrointestinal, respiratory, and urogenital tracts as well as other mucosal surfaces of the body (42). The predominance of immunoglobulin A (IgA) antibody (Ab) at mucosal surfaces has allowed the definition of an interrelated humoral immune response that operates among these mucosal sites (45). Thus, immunization at one mucosal site frequently results in the generation of antigen (Ag)-specific IgA Ab at other mucosal sites as opposed to systemic sites (42). The linking of mucosal sites by the B lymphocytes that preferentially recirculate to mucosal areas defines the common humoral mucosal immune system. Because specific IgA Ab at mucosal surfaces has been correlated with protection against a number of mucosally related pathogens (31, 46), knowledge concerning the details of immunity generated at mucosal surfaces is critical for the control of many human and animal pathogens.

Gastrointestinal tract-associated lymphoid tissue (GALT) includes organized lymphoid components consisting of Peyer's patches (PP), regional lymphatics and mesenteric lymph nodes, dispersed lymphoid cells in the epithelial layer (intraepithelial lymphocytes), and the gut lamina propria (9). The entry of Ags into the host and uptake by PP-associated Ag-presenting cells (APC) are mediated in part by M cells found overlying the PP as part of the follicle-associated epithelium (20, 33). The gastrointestinal mucosa provides a formidable barrier to the systemic entry of commensal organisms and pathogens, such as viruses, due to the presence of nonimmune factors and specific immune functions. While some protection is afforded by innate cells such as natural killer (NK) cells, protection from and clearance of infection usually require expansion of Ag-committed B and T cells (12, 43). The development of such specific immunity at the mucosal surface is regulated and potentially involves populations of cells that differ from those found in the systemic circulation (19, 34). An immune response generated in the PP results in the emigration of primed lymphocytes into the lymph and circulatory systems (22, 28). Ultimately, the mucosally primed effector cells home to the mucosal epithelium of the gastrointestinal tract and may also potentially seed the lung and other distal mucosal sites (22, 42, 47).

Mucosally associated lymphoid tissue found in the respiratory tract is referred to as the respiratory tract-associated lymphoid tissue (RALT). There is a paucity of information about the phenotypes and functional potential of the cellular elements that comprise the mucosal immune component of the RALT. Respiratory syncytial virus and influenza virus infections of the respiratory tract have been used to perturb and activate the cellular components of RALT, and these have been analyzed for lymphokine production and APC activity (14, 17, 18). Recently, following the use of reovirus as a stain to identify rat M cells in respiratory tissue that resemble those in the follicle-associated epithelium of PP of the gut (29), reoviruses have been used to infect the mouse respiratory tract and perturb local lymphoid cell subsets (3, 44). Finally, an unusual T-cell phenotype has been described for porcine mucosally associated lymphoid tissueCD4⁺/CD8⁺ double-positive (DP) cells (49, 50). This is an intriguing observation, since we found, as reported in this article, that the draining lymph nodes of the mouse RALT develop a significant population of such cells following acute intratracheal (i.t.) infection. Mucosally associated lymphoid tissue in the respiratory tract in general is referred to as RALT while the bronchus-associated lymphoid tissue in particular is called BALT (6). Lymphatics leading from the RALT eventually drain into the lung-associated lymph nodes, the mediastinal (MD) node, and the tracheobronchial (TB) node, where a local immune response may be generated (14, 17).

A number of studies have examined the nature of immune responses to particular mucosal pathogens, many of which are able to colonize and stimulate specific immunity at one or another particular site. Thus, it has been difficult to examine the extent of cross priming between two different mucosal sites by using the same pathogen. Reovirus 1/Lang (1/L) has been used extensively in our laboratory to study both the cellular and the humoral components of the gut mucosal immune system (5, 26). In the gut, reovirus 1/L induced a B-cell response in PP leading to reovirus 1/L-specific IgA memory cells (26) as well as to an increase in the potential of previously primed B cells of other specificities to express IgA (8). In addition, major histocompatibility complex-restricted CD8⁺ virus-specific precursor and effector cytotoxic T lymphocytes (CTLs) arose in PP and the epithelium of the gut mucosa after enteric reovirus 1/L infection (25, 26). Therefore, reovirus 1/L is an effective gut mucosal immunogen, eliciting a humoral as well as a cellular immune response. Since reovirus has been cultured as isolates from the lung as well as from the gut (40) and has been shown to infect murine respiratory tissues (3, 44), we chose to investigate whether reovirus could serve as an effective respiratory stimulus and elicit an immune response in the RALT.

In this article, we describe a reovirus 1/L-induced murine model of respiratory mucosal immunity that allows investigations of the mucosal immune response generated at respiratory surfaces and of cross priming between the GALT and RALT. We also report that a single i.t. application of reovirus 1/L is capable of generating a unique T-cell subsetCD4⁺/CD8⁺ DP cells in the MD node in an acute reovirus infection. The potential role for this undescribed lymphocyte component within the murine RALT is discussed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. Male C3HeB/FeJ mice weighing 25 to 30 g were purchased from the Jackson Laboratories, Bar Harbor, Maine. Virally primed mice were kept physically isolated from all other experimental and stock mice. Mice 6 to 12 weeks old were used in all experiments. Lymphoid tissues from three mice were pooled for each experiment.

Cell lines. L-929 (H-2^k) fibroblast cells were used in this study. Fetal African green monkey kidney cells (MA-104) were grown as previously described (32).

Culture medium. Complete RPMI medium was RPMI 1640 medium (Gibco Laboratories, Grand Island, N.Y.) supplemented with 10% fetal calf serum (Gibco), 1% L-glutamine (Gibco), 0.01% gentamicin (Gibco), and 1% antibiotic-antimycotic solution (100 U of penicillin/ml, 1% streptomycin, 0.25 µg of amphotericin B [Fungizone]/ml) (Gibco).

Viruses and immunizations. Reovirus 1/L was obtained from C. Cuff, West Virginia University, Morgantown. The mice were anesthetized with an intraperitoneal injection of 0.5 ml (per mouse) of a solution of Avertin (2,2,2-tribromoethanol; Aldrich Chemical Co., Milwaukee, Wis.) (23). The animals were inoculated i.t. with 10⁷ PFU of reovirus 1/L in 50 µl of sterile phosphate-buffered saline (PBS) (Gibco Laboratories). Reovirus 1/L-inoculated mice were kept physically isolated from all other experimental and stock mice. The same dose of 10⁷ PFU of reovirus was used when administered by other routes. Mice were anesthetized with Metofane for intranasal (i.n.) inoculation of 25 µl by instillation with a yellow pipette tip into each nostril. For intraduodenal (i.d.) inoculations, Avertin was used for anesthesia and injection was as described previously (23, 26). Intragastric (i.g.) immunization was by oral intubation and delivery of 100 µl (5).

Titration of infectious virions. TB tree, lung, spleen, liver, and small intestine were dissected, and each was put in 1 ml of complete RPMI medium, followed by homogenization of the tissues. Serial dilutions of the tissue homogenates were used in a standard plaque assay on MA-104 monolayers in a six-well tissue culture dish (Linbro) as previously described (32).

Preparation of cell suspensions. To obtain RALT cells, mice were sacrificed and perfused by direct cardiac injection of PBS until the lungs were visibly whitened. The TB tree and lungs were separated and removed from each mouse; the peripheral tissues were dissected out and discarded. The remaining tissue was transferred into a bottle containing a solution of Dispase (1.5 mg/ml; grade II; Boehringer Mannheim Biochemicals, Indianapolis, Ind.) in Hanks balanced salt solution (HBSS) (Gibco). The mixture was stirred continuously at 37°C for 2 h. Dissociated cells were washed extensively with calcium- and magnesium-free HBSS (Gibco) and then resuspended in complete RPMI medium. MD and TB cells were obtained by mechanical dispersion of the nodes in complete RPMI medium.

Proliferation assay. Lymphocytes were cultured in complete RPMI medium in a 200-µl volume in 96-well round-bottom plates (Corning Glass Works, Corning, N.Y.) and incubated at 37°C in a humidified atmosphere of 5% CO₂ and 95% air. Three replicates of 200-µl samples were pulsed with 1 µCi of [³H]thymidine (specific activity, 40 to 60 Ci/mmol; Amersham, Arlington Heights, Ill.) for 18 h in 96-well round-bottom tissue culture plates (Corning). After harvesting of the cells onto glass fiber filter mats with a multiple harvester (Skatron AS, Lier, Norway), the [³H]thymidine incorporated into newly synthesized DNA was measured by conventional liquid scintillation procedures in a Beckman beta counter.

Two-color analysis and fluorescence-activated cell sorting (FACS) of T cells. For analytical and preparative sorts, cells were incubated with an appropriate dilution of fluorochrome-coupled reagent in PBS for 40 min on ice. The cells were then washed three times and analyzed or sorted on a FACS flow cytometer (Becton Dickinson, Sunnyvale, Calif.). Sorted cells were resuspended in complete RPMI medium. Cells were stained with phycoerythrin (PE)-labeled anti-mouse CD4 (L3T4; Pharmingen, San Diego, Calif.) and fluorescein (FLU)-labeled CD8 (53.6.72; Pharmingen), and we collected cells which were positive for both markers (see reference 27). Cells from various lymphoid sources were analyzed by FACS with PE-conjugated polyclonal goat anti-µ chain or anti- chain (Southern Biotechnological Associates, Birmingham, Ala.) for B cells and monoclonal Abs (MAbs), FLU-anti-CD4 (L3T4; see above), biotinylated anti-CD4 (Pharmingen) followed by PE-avidin (Pharmingen), PE-anti-CD4 (L3T4; see above), and/or FLU-anti-CD8 (53.6.72; see above).

Culture of purified lymphocytes. Lymphocytes purified by FACS were cultured in 96-well round-bottom plates at 2 × 10⁵ cells per well. Virally pulsed (5 × 10⁴) peritoneal exudate stimulator cells (PECs) or splenic cells were used as a source of APC. PECs were obtained from syngeneic mice that had received a single intraperitoneal injection of 1.5 ml of thioglycolate medium without indicator (BBL Microbiology Systems, Cockeysville, Md.) 3 to 4 days before the PECs were harvested by peritoneal lavage with 10 ml of HBSS. Harvested PECs or spleen cells were virally pulsed at a multiplicity of infection (MOI) of 1 for 1 h, during which they were exposed to 1,600 rads of  radiation from a cobalt source. On the following day, concanavalin A (ConA)-conditioned medium at a final concentration of 10% was added, and the cultures were harvested after 6 days. The preparation of ConA-conditioned medium has been previously described (37). Viable lymphocytes were obtained from bulk cultures by Ficoll-Isopaque centrifugation (10).

In vitro generation of virus-specific CTLs from BALT and MD and TB nodes. Conventionally raised C3H mice were immunized by i.t. application of 3 × 10⁷ PFU of reovirus 1/L suspended in PBS. Control mice were subjected to the same surgical procedures but injected with PBS alone. Ten days after priming, single-cell suspensions of BALT, MD, and TB lymphocytes were obtained. For bulk cultures, 2 × 10⁵ lymphocytes/ml were added to a T-25 (Corning) flask in the presence of 5 × 10⁴ virus-pulsed PECs (see above). CTLs were generated from MD nodes 10 days post-i.t. immunization with reovirus 1/L. The lymphocytes from the MD nodes were sorted for cells that were DP for both CD4 and CD8 markers. The sorted lymphocytes were cultured with either irradiated splenic cells pulsed with reovirus to an MOI of 1 or without the irradiated splenic cells. On the following day, ConA-conditioned medium at a final concentration of 10% was added, and the cultures were harvested after 6 days for bulk cultures. Viable lymphocytes were obtained from bulk cultures by Ficoll-Isopaque centrifugation (10).

Cytotoxicity assay. A standard ⁵¹Cr release assay, using various effector/target cell ratios, was used to measure cell-mediated cytotoxicity (7). Assays were performed in 96-well V-bottom plates (Nunc) in 5% CO₂ in an air incubator for 4 h. All assays were performed in triplicate. L cells were infected with reovirus at an MOI of 5 before overnight culture in T-25 flasks in the presence of 200 µCi of Na⁵¹Cr (Amersham). Adherent targets were released by incubation with trypsin-EDTA solution (Gibco).

Fragment culture. Mice were sacrificed on days 3, 7, 11, 14, 21, and 28 postimmunization. RALT organ cultures were established with lungs, TB trees, and MD and TB nodes from the adult mice (see reference 24). Briefly, small pieces of the lung or tracheobronchus were taken and washed extensively in calcium- and magnesium-free HBSS containing 0.1% gentamicin and with complete RPMI medium. The small pieces were cultured in a sterile flat-bottom 24-well plate (Costar) in 1 ml of Kennett's H-Y medium (JRH Biosciences, Lenexa, Kans.) containing 10% fetal calf serum, 1% L-glutamine, 0.01% gentamicin, and 1% antibiotic-antimycotic solution for 7 days under 90% O₂ and 10% CO₂ at 37°C. Culture supernatants were frozen prior to the assay.

Radioimmunoassay (RIA) of culture supernatants. The RIA for reovirus-specific IgA Ab has previously been described (5). To estimate total IgA produced by fragment cultures, RIA plates were coated with polyclonal goat anti-mouse Fab (Jackson ImmunoResearch, West Grove, Pa.) and 20 µl of 1.0 ml of each culture supernatant, or 20 µl of serial dilutions, was applied to the coated plates. After appropriate washing, ¹²⁵I-labeled polyclonal goat anti-mouse IgA (Southern Biotechnological Associates) was used to develop assays for both specific and total IgA. Since we had no IgA MAbs to reovirus, we used a prototypical mouse IgA MAb to the phosphocholine (PC) determinant, TEPC15, to construct standard curves. A series of 20-µl samples of this anti-PC Ab, containing known amounts of Ab, were adsorbed to plates coated with PC-bovine serum albumin. The washed plates were developed with the same ¹²⁵I-labeled anti-IgA used in the assays described above. The standard curve was linear in the range of 0 to 20 ng, and generally, about 1,000 cpm is convertible to 130 ng of IgA per ml in culture fluids.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Comparison of various routes of infection: i.t., i.n., i.d., and i.g. vis-à-vis initial viral growth in GALT and RALT. To develop a model of reovirus infection, conventionally reared C3H mice were immunized with 3 × 10⁷ PFU of reovirus type 1 i.t. Mice were sacrificed on days 3, 6, 11, 14, 21, and 28, and titers of virus (if present) in the lung, trachea (TB tree), liver, spleen, and small intestine were determined. Table 1 shows that lung and trachea (TB tree) had persistently high titers for the first week following infections (10⁷ to 10⁵ PFU/g of tissue), whereas titers were 10- to 100-fold less in the intestine, liver, and spleen (10⁶ to 10⁵ PFU/g of tissue). By day 11, titers of reovirus in the respiratory tree (trachea and lung) were 10⁴ to 10⁵ PFU/g of tissue, whereas virus was cleared from the intestine and only minimal amounts of virus were recovered from liver (10¹ PFU/g of tissue) and spleen (10³ PFU/g of tissue). Virus was recovered from the lungs only on day 14, with clearance occurring by day 21. Our results demonstrate that i.t. administration of reovirus type 1 can provide a substantially local infection in the RALT, with delayed clearance from this site.

Following i.t. infection with reovirus, the highest concentrations of Abs are in RALT and most of this specific response is in the form of IgA Abs. Infection by the i.t. route results in a dramatic increase in specific IgA Abs expressed by RALT in tissue fragment cultures (Fig. 1). Using our radiolabeled anti-IgA, we have determined that a signal of 1,000 cpm is equivalent to 0.13 µg of IgA per ml of culture. Other isotypesIgG1, IgG2, and IgMwere not present at levels appreciably above background when tested on Ag-coated plates. IgA secretion increased in the TB tree and lung tissue until day 11 and then decreased. The TB and MD nodes were markedly enlarged during the virus infection, similar to the substantial increase observed in the PP during reovirus infection of the gut. Reovirus-specific IgA secretion in the TB node also peaked on day 11. Fragment cultures of the GALT following i.t. inoculation showed minimal production of specific IgA Abs, suggesting minimal emigration and lodging of effector cells from RALT and minimal effective stimulation by the relatively low numbers of virus that reached the gut.

View larger version (32K): [in this window] [in a new window]   FIG. 1.   IgA Ab response of RALT following i.t. infection with reoviruses. Conventional C3H mice were immunized i.t. with 3 × 107 PFU/mouse. Sets of two mice per day were sacrificed on days 3, 6, 11, 14, 21, and 28 postimmunization. Fragment cultures of the RALT (MD and TB nodes, lung, and TB tree) were set up. Supernatants of the fragment cultures were assessed for reovirus-specific IgA Abs by RIA. Background cpm from supernatants of tissue fragments from noninfected, control mice were less than 100.

Surface phenotypes from RALT compartments. B-cell, CD8⁺ T-cell, and CD4⁺ T-cell elements could be found in the RALT compartment, as shown in Fig. 2. Interestingly, a subset of CD4⁺/CD8⁺ (DP) T cells transiently appears in the MD and TB nodes of RALT 7 to 10 days after i.t. reovirus infection, as shown in both Fig. 2 and Fig. 3. These cells maintained the DP characteristics even after being in culture for 6 days as shown in Fig. 3D to H. Many of these cells were found to express the / heterodimer of CD8 as shown in Fig. 3I.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Cytofluorometric analysis of murine lymphocytes from the respiratory tract 7 to 10 days after i.t. immunization of conventional C3H mice. Numbers represent the percentages of positive cells within each sector. (Panels: A to C, MD node; D to F, TB nodes; G to I, BALT. FSC, forward scatter; SSC, side scatter.

View larger version (40K): [in this window] [in a new window]   FIG. 3.   Two-color cytofluorometric analysis of MD cells and sorted DP cells. Many murine MD cells were CD4+/CD8+ (A and B) and CD8+ (C). The DP CD4+/CD8+ population in the MD node (D) was purified by FACS (E) from conventional C3H mice immunized i.t. with reovirus (RV) and stimulated for 6 days with irradiated spleen cells (SP) plus reovirus (F), reovirus alone (G), or medium alone (H). Most of these sorted DP cells remained CD8+ after 6 days of culture (I).

Virus-specific cytotoxic activity is found in BALT and MD and TB nodes after i.t. application of reovirus type 1. To investigate CTL responsiveness after mucosal challenge, an in vitro culture system, in which precursor CTLs (pCTLs) could expand into effector CTLs, was used. Ten days after the i.t. application of 10⁶ PFU of reovirus 1/L, significant levels of cytotoxicity were generated in vitro from BALT and MD and TB nodes (Fig. 4). The BALT, MD, and TB cells primed in vivo and then cultured in the presence of irradiated PECs plus reovirus produced substantial levels (55% lysis) of cytotoxicity for infected L cells. Primed BALT, MD, and TB cells cultured without stimulator cells and virus produced no detectable levels of cytotoxicity for either infected or uninfected L cells (data not shown). Therefore, reovirus-specific pCTLs can be generated in vitro from BALT, MD, and TB cells that have been stimulated in vivo with reovirus.

View larger version (11K): [in this window] [in a new window]   FIG. 4.   Cytotoxic activity of BALT (a), MD (b), and TB (c) lymphocytes obtained from C3H mice immunized i.t. with reovirus (RV). The lymphocytes were stimulated in vitro with PECs pulsed with reovirus 1/L and conditioned medium, reovirus 1/L alone, or conditioned medium alone. After 6 days of culture, cytotoxic activity was tested by using uninfected L cells ( and ) or reovirus 1/L-infected L cells ( and ) as targets.

Lymphoproliferation of DP cells. DP cells were restimulated in vitro with either virally pulsed PECs or splenic cells as shown in Fig. 5. In vitro antigenic stimulation of DP cells with virally pulsed splenic cells was twofold higher than that with virally pulsed PECs. The results indicate that the DP cells could be stimulated to proliferate in an Ag-dependent manner.

View larger version (10K): [in this window] [in a new window]   FIG. 5.   Proliferative response of CD4+/CD8+ lymphocytes (DP) to stimulation with virus-pulsed irradiated PECs or spleen cells. MD lymphocytes from reovirus (RV)-immunized mice were stained for CD4+ and CD8+ markers. The DP population was sorted and stimulated in culture. Values for lymphoproliferation represent mean counts per minute of triplicate cultures. Background proliferation was less than 100 cpm in all samples.

Virus-specific cytotoxic activity is found in DP cells after i.t. application of reovirus type 1. To investigate CTL responsiveness after mucosal challenge, an in vitro culture system in which pCTLs could expand into effector CTLs was used. Ten days after i.t. immunization of 10⁶ PFU of reovirus 1/L, significant levels of cytotoxicity were generated in vitro from DP lymphocytes prepared by FACS as shown in Fig. 6.

View larger version (12K): [in this window] [in a new window]   FIG. 6.   Cytotoxic activity of reovirus (RV) 1/L-primed DP lymphocytes. DP lymphocytes obtained from MD nodes of i.t. immunized mice were stimulated in vitro with reovirus and conditioned medium or with conditioned medium alone. The effectors were tested for cytotoxic activity for the following targets: uninfected L cells or reovirus-infected L cells.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we examined the ability of i.t. administered reovirus 1/L to perturb, in an Ag-specific manner, both T- and B-cell subsets that occur in RALT. Since reovirus 1/L can potentially enter the RALT via M cells (29), we postulated that it would be an efficacious mucosal immunogen capable of expanding the numbers of virus-specific T and B cells in this compartment. In this study, we have found that i.t. inoculation of conventionally reared C3H mice with reovirus 1/L results in high virus titers in the lung and TB tree by day 3 postimmunization. Virus was recovered from the lungs on day 14, with a clearance occurring on day 21 as shown in Table 1. The high titers and delayed clearance of infectious reovirus from RALT should facilitate our assessment of the effectiveness of priming at a distant mucosal site in limiting infection of the respiratory tract upon i.t. challenge. These characteristics of respiratory infection with reovirus should also enable us to assess the effectiveness of passively acquired IgA Abs and adoptively transferred subsets of B and T cells, derived from transiently infected gut, in attenuating infection of the respiratory tract upon i.t. challenge. Our data demonstrated that infectious virons persist in the lungs through 21 days after i.t. immunization. This persistence of the virus in the lungs was unexpected, since the gut mucosal route of infection with reovirus 1/L is not associated with such prolonged persistence of infectious virus in the intestine (8). In the case of experimental respiratory infection with influenza virus, infectious virus is cleared from the lung in 7 to 10 days (1, 2), and after similar infection with Sendai virus, infectious virus is cleared from the lung by 10 days postinfection (15, 16).

Table 2 shows the viral titers in proximal and distal tissues, determined 3 days after infection by the i.t., i.d., i.n., and i.g. routes. Day 3 postimmunization was chosen for this comprehensive comparison because our previous studies using i.d. infection showed that maximum infectious virus titers were found in gut tissue at that time following primary infection. Clearly, i.d. inoculation results in the most restricted local infection. No infectious virus was recovered from RALT following i.d. inoculation, offering the possibility for analyzing whether and by what mechanism cross priming of GALT to RALT may occur. Of the two routes most likely to lead to primary infection of the respiratory tract, the i.t. route provided the most locally restricted infection compared with i.n. inoculation. Twenty- to fifty-fold-higher titers of infectious virus were found in the RALT than in the gut after i.t. inoculation. Thus, for this study of the characteristics of host response to RALT infection, we used i.t. infection exclusively.

We found that reovirus 1/L, given i.t., effectively stimulates a virus-specific IgA Ab response in the various lymphoid tissues associated with the respiratory tract (Fig. 1) and also results in hypertrophy of the draining TB and MD lymph nodes which contain prominent CD8⁺ T-cell populations (Fig. 2). The latter lymph node cell populations and those from BALT develop virus-specific pCTLs, which generate effector CTLs upon in vitro restimulation with virus-pulsed, irradiated APC (Fig. 4). CD8⁺ T cells are thought to be the major effector cells in clearance of both influenza virus (1, 4, 41, 48) and respiratory syncytial virus (18) from respiratory tissues, and CTLs are likely to play a similar role in the eventual clearance of reovirus from RALT (Table 1).

A novel population of CD4⁺/CD8⁺ DP lymphocytes was detected by two-color flow cytometry in dispersed cells from lymph nodes draining the respiratory tract (MD and TB nodes) following i.t. inoculation of reovirus (Fig. 2). Insufficient cells were available from dispersed BALT to properly ascertain whether and when such DP cells may also appear at this site. Such CD4⁺/CD8⁺ DP lymphocytes have been reported for pigs, wherein these cells are functionally mature T cells and differ significantly in their phenotype from the immature CD4⁺/CD8⁺ cells found in the porcine thymus (35, 38). Furthermore, CD4⁺/CD8⁺ DP porcine lymphocytes are able to respond to recall viral Ag and, akin to human memory cells, express high levels of 1 integrin (50). All of these properties are typical of a memory cell population. In this study, some functional aspects of murine DP cells have been studied. The data demonstrate directly, for the first time, the presence of CD4⁺/CD8⁺ DP lymphocytes in MD and TB draining nodes of the RALT. The presence of this lymphocyte population has previously been observed for the intraepithelial lymphocyte compartment in the small intestines of mice (30) and in tonsils of pigs (49). Little is known about the tissue origin, differentiation pathway, migratory behavior, and functions of the CD4⁺/CD8⁺ DP T-cell population in healthy animals. Nevertheless, there is evidence that porcine DP T cells are CD4⁺ single-positive (SP) cells that have acquired CD8⁺ after exposure to Ag. It has been demonstrated elsewhere that a significant proportion of porcine lymphoblasts generated during an in vitro response to allogeneic (39), viral (36), or parasitic (11) Ags are CD4⁺/CD8⁺ DP cells, while in the same culture CD4⁺ SP lymphoblasts are very scarce. Furthermore, purified porcine CD4⁺ SP cells give rise to CD4⁺/CD8⁺ DP cells upon in vitro stimulation with recall viral Ag (50). We observed that sorted murine DP lymphocytes of 96 to 98% purity maintained their DP phenotype after 6 days in culture (Fig. 3). Thus, it is unlikely that the DP phenotype detected by FACS was due either to CD4⁺ SP and CD8⁺ SP doublets or to passively acquired CD4 or CD8 Ag. In vitro stimulation with reovirus-pulsed APC did seem to give rise to some SP cells, although the majority remained DP, while incubation in medium alone with or without virus did not. Under similar conditions of culture with virus-pulsed APC, DP cells proliferated (Fig. 5) about as well as did SP cells (data not shown). Recently, it has been shown that activated CD8⁺ T cells function by initiating programmed cell death in the target, either by a perforin-granzyme-mediated process or via the ligation of Fas (CD93) (21). The former is the more rapid and efficient process. Possibly, the acquisition of CD4 surface molecules may enhance the killing of infected cells that also express class II major histocompatibility complex molecules, such as infected mucosal epithelial cells, by the less efficient mechanism.

In summary, i.t. infection of C3H mice with reovirus 1/L provides a useful model not only for study of respiratory mucosal immunity but also for analyses of interactions between the various mucosal sites. Because reovirus 1/L can infect both respiratory and gut epithelium and because there exist parallel responses between the gut and respiratory tract, such as generation of IgA Abs and cytotoxic activity, this model system allows us to study the cross talk between GALT and RALT populations. Understanding the role of CD4⁺/CD8⁺ DP cells and their unique distribution may provide important clues for elucidating the role of such cells in immunity and/or disease processes.