Unique Characteristics of Lacrimal Glands as a Part of Mucosal Immune Network: High Frequency of IgA-Committed B-1 Cells and NK1.1⁺ αβ T Cells

Abstract

purpose. Immunologic characterization of IgA-committed B-1 and B-2 cells, and unique subsets of T cells isolated from the murine lacrimal gland (LG), the primary exocrine tissue for the ocular surface, which is considered to be a part of the mucosal immune system.

methods. Single cells were obtained from LGs of C57BL/6 mice by the enzyme dissociation method using collagenase type IV. Samples underwent flow cytometric analysis to characterize the unique subsets of T and B cells. To test the effectiveness of ocular vaccination, mice were immunized ocularly or nasally with cholera toxin (CT; 10 μg/mouse) suspended in phosphate-buffered saline. Antigen-specific immune responses were determined by isotype and CT-specific enzyme-linked immunosorbent assay (ELISA) and enzyme-linked immunospot (ELISPOT) assay.

results. When mononuclear cells (MC) isolated from LG samples were examined by flow cytometry, approximately 28% of cells were characterized as B220⁺ B cells. Because surface IgA⁺ (sIgA⁺) B cells develop from B-1 and B-2 lineages, it was important to examine which subset of B cells gives rise to LG sIgA⁺ B cells. Examination of the MC isolated from LG samples showed that approximately 4% of cells were sIgA⁺ B cells. Furthermore, nearly all these sIgA⁺ B cells (97.5%) belonged to the B-1 lineage, especially the B-1a cell line (B220^low, CD5⁺). Of the isolated CD3⁺ T cells, 75% were αβ and 25% were γδ T-cell receptor positive. The proportion of NK1.1⁺ αβ T cells was higher (3%) in LG samples than in submandibular gland samples (0.5%). Ocular immunization with CT-induced antigen-specific mucosal (e.g., found in tear-wash and saliva samples) and systemic (e.g., serum) immune responses. The magnitude of antigen-specific antibody responses was comparable to those induced by nasal immunization.

conclusions. These results show that LG contains unique subsets of B (e.g., sIgA⁺ B-1 cells) and T (e.g., NK1.1⁺ αβ T cells) cells. Furthermore, as a part of the mucosal immune barrier, the LG is an important immunologic tissue for the ocular surface.

According to the functional role played in the immune response to mucosally encountered antigens, the mucosal immune system can be classified into two categories: IgA-inductive and IgA-effector sites. Payer’s patches (PP), gut-associated lymphoreticular tissue (GALT), are characterized as IgA inductive sites ¹ in which the induction of IgA-committed B cells occurs. In contrast, the intestinal lamina propria (i-LP) and the epithelial layer of the small intestine are considered to be the main locations of IgA-effector sites. It has been shown that oral or intragastric immunization results in the generation of Th2-type CD4⁺ T cells and surface IgA⁺ (sIgA⁺) B cells in GALT, which leads these lymphocytes to subsequently home into distant IgA-effector tissues ^{2 3} such as the intestinal tract and glandular tissues, including the salivary glands (SG) and lacrimal glands (LG). At these distant IgA-effector sites the lymphocytes induce both antigen-specific IgA antibody–producing cells and secretory IgA (S-IgA) antibodies. According to the accumulated molecular and cellular evidence from the ocular immune system, ^{4 5} the LG is considered to be an immune effector site for the mucosal immune system that protects the ocular surface. 

Mucosal effector sites, such as the lamina propria of the gut, LG and submandibular gland, contain high numbers of plasma cells committed to the secretion of IgA antibody. Constituting a first line of defense against pathologic microorganisms, the dimmer or polymeric forms of these IgA are transported across the epithelium into the gut lumen in the medium of a secretory component. According to the expression of B220, IgM, IgD, Mac-1, CD23, and CD5, mucosal B cells can be divided into two subsets: B-1 cells are B220^low, IgM^high, IgD^low, Mac-1⁺, and CD23⁻, and conventional B-2 cells are B220^high, IgM^low, IgD^high, Mac-1⁻, and CD23⁺. ^{6 7 8 9} The former B-cell subset can be further classified depending on the surface expression of CD5 ^{8 9} : B-1a cells are CD5⁺, and B-1b cells are CD5⁻. In recent separate study, we showed that sIgA⁺ B-1 cells are predominantly found in the mucosal effector tissues such as i-LP and submandibular gland (SMG), whereas conventional sIgA⁺ B-2 cells are located in both inductive (e.g., PP) and effector (e.g., i-LP) tissues. ¹⁰ Of the sIgA⁺ B-1 cells, a dominant fraction of the B-1b, but not B-1a cells was committed to express IgA in the i-LP. ¹⁰ Other studies have also suggested that in the mucosal effector tissues, B-1 cells could be a major supplier of IgA plasma cells. ^{11 12 13} Thus, it is of value to consider the potential contribution of sIgA⁺ B-1 and B-2 cells to the ocular immune system. 

In this study, to investigate the unique and important role, as part of the common mucosal immune network system, played by the ocular immune system, especially in the induction and regulation of the ocular IgA immune response, mononuclear cells (MC) were enzymatically isolated from LG tissue and examined. CD3⁺ cells purified from lacrimal MC were analyzed for the expression of both αβ and γδ T-cell receptors (TCR), for NK1.1, and for CD4 and CD8. B lineage cells were characterized into B-1 and B-2 subsets of sIgA⁺ B cells. We also investigated whether ocular immunization induced mucosal and systemic antigen-specific immune responses. 

Materials and Methods

Mice

C57BL/6 mice (male, 6 to 8 weeks old) were obtained from Charles River Japan (Atsugi, Japan). Experiments were performed using 10 animals per group. The animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Isolation of MC from LG

After exsanguination of mice under anesthesia with 2 mg ketamine (Sigma, St. Louis, MO), the skin of the head was carefully removed. The murine LGs were found both within and outside the orbit, which were ventral and anterior to the ear subcutaneously. ^{14 15 16} Each individual LG was visualized under stereoscopic microscopy (Leica, Heerbrugg, Switzerland) and carefully removed using microsurgical tweezers. In general, two LGs were isolated from each mouse, and a total of at least 20 LGs were used per experiment. The tissues were carefully dissected and transferred to Petri dishes (100 × 15 mm; Falcon 1029; Becton Dickinson, Lincoln Park, NJ) containing RPMI 1640 (GIBCO BRL; Gaithersburg, MD) supplemented with sodium bicarbonate, nonessential amino acids, sodium pyruvate, l-glutamine, penicillin, streptomycin, and gentamicin (incomplete medium). ^{17 18} For the isolation of MC from LG, a modified enzymatic dissociation method was developed according to a previously described protocol. ^{17 18 19 20 21 22} Lacrimal gland tissue was dissected into small fragments and then dissociated into single cells by use of RPMI 1640 containing collagenase type IV (Sigma). ^{17 18 21} After 20-minute incubation with continuous stirring at 37°C, dissociated cells were harvested and washed with incomplete medium. Cells were then resuspended in incomplete medium containing 2% fetal calf serum (FCS). For additional dissociation, the residual tissues were further mixed with fresh medium containing collagenase. This process was performed at least five times. Individual cell fractions were pooled and washed with incomplete medium containing 2% FCS. Cells were then passed through a cotton-glass wool column to remove dead cells, clumps, and tissue debris. The dissociated LG cells were then resuspended in 2 ml RPMI 1640 containing 75% Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden). In addition, 40% Percoll (4 ml) was carefully and sequentially layered on top of the 75% layer in a polystyrene round-bottomed tube (17 × 100 mm, Falcon 2057; Becton Dickinson). After centrifugation (600g) at 25°C for 20 minutes, the interface between the 75% and 40% layers was carefully removed as an enriched lymphocyte fraction. ^{17 18} This procedure provided >97% viable MC with a cell yield of ∼2 × 10⁵ cells (LG)/mouse. 

Analysis of T- and B-Cell Subsets by Flow Cytometry

To characterize the T and B cells from the LG samples, two-color or three-color flow cytometric analysis was performed. ^{17 18 21 22} To stain the different subsets of T and B cells, we used the appropriate fluorescence-conjugated or biotin-conjugated anti-CD3 (145-2C11), anti-L3T4 (anti-CD4; G.K 1.5), anti–Ly-1 (anti-CD5; 53-7.3), anti–Lyt-2 (anti-CD8α; 53.6-72), anti-γδ TCR (UC7-13D5), anti–αβ TCR (H57-597), anti-CD45R/B220 (RA3-6B2), anti-IgA (R5-140), and anti-NK1.1 (PK136) monoclonal antibodies, which were purchased from Pharmingen (San Diego, CA). After the two-color or three-color staining, these samples were subjected to flow cytometric analysis using a FACS Caliber (Becton Dickinson, Sunnyvale, CA). Each sample comprised at least 10⁵ live cells. For control, some samples were incubated with the particular isotype control antibody, and these cells were used to set the lymphocyte gates. Each analysis was performed at least three times to verify the results obtained, and the results were expressed as the mean. 

Ocular Immunization with Cholera Toxin

With 2 μg/mouse per week of cholera toxin (CT) suspended in phosphate-buffered saline (PBS), mice were immunized ocularly ^{23 24 25} or nasally ²⁶ for 5 consecutive weeks. Serum, saliva, and tear-wash samples were obtained at 1-week intervals. Antibodies to CT were measured using a standard enzyme-linked immunosorbent assay (ELISA) with 2 μg/well of CT as coating antigen (see below). One week after the fifth immunization, mice were killed to examine antigen-specific IgM-, IgG-, and IgA-producing cells in spleen, SMG, and LG by ELISPOT assay (see below). 

Analysis of Isotype and Antibody Titers of Antigen-Specific Immunoglobulins by ELISA

Isotype and antibody titers of CT-specific immunoglobulin in tear-wash and saliva samples were determined by ELISA as previously described. ^{2 3 22} Tear-wash samples were collected by washing the eyeball with 100 μl cold PBS. ²⁷ Saliva samples were obtained by the standard method routinely performed by our group. ^{22 26} The 96-well plates (Nunc, Roshilde, Demmark) were coated with an optimal concentration of CT (2 μg/ml) in PBS. Wells were blocked with 200 μl PBS containing 10% normal goat serum (GIBCO BRL) for 2 hours at 37°C. After extensive washing, serial dilutions of tear-wash or saliva samples were added and incubated for 2 hours at 37°C. After incubation and washing, 100 μl of 1:1000 diluted biotinylated goat anti-mouse μ, γ, or α heavy chain–specific antibody (SBA, Birmingham, AL) was added to the wells. The detection solution containing a 1:2000 dilution of horseradish peroxidase–conjugated streptavidin (GIBCO BRL) was added. The plates were incubated at room temperature for 1 hour. After washing, the color was developed at room temperature with 100 μl of 1.1 mM 2,2′-azino-bis (3-ethylbenz-thiazoline-6-sulfonic acid) containing 0.01% H₂O₂. After a 15-minute incubation, the plates were read at an optical density of 414 nm using a microplate reader (Bio-Rad, Hercules, CA). Reactions were terminated by the addition of 50 μl of 10% SDS in 0.01 M citrate-phosphate buffer. 

Enumeration of Antigen-Specific Immunoglobulin–Producing Cells by ELISPOT

To count the numbers of CT-specific IgA-, IgG-, and IgM-producing cells in LG, we used a previously described a ELISPOT assay method. ^{2 3 22 28} After 96-well nitrocellulose base filtration plates (Millititer HA; Millipore, Bedford, MA) were coated with 100 μl of cholera toxin β subunit (CT-B) (2 μg/ml) in PBS and incubated overnight at 4°C, the plates were washed three times with PBS and then blocked with incomplete medium containing 5% FCS for 1 hour. The blocking medium was removed and cell samples in complete medium (incomplete medium with 10% FCS) were added at various concentrations and cultured for 4 hours at 37°C in air with 10% CO₂ and 90% humidity. After incubation, the plates were thoroughly washed with PBS and then with PBS containing Tween solution (0.05%; PBS–TW). To capture antibody-producing cells, 1 μg/ml biotin-labeled affinity-purified goat anti-mouse μ-, γ-, orα -specific antibody (SBA) in PBS–TW containing 2% FCS was added. After overnight incubation at 4°C, the plates were washed three times with PBS–TW, after which an aliquot of 100 μl avidin–peroxidase (Zymed Laboratories, San Francisco, CA) diluted 1/1000 in PBS–TW was added to each well. The plates were incubated in the dark at room temperature for 1 hour. After washing with PBS, the spots were developed with 3-amino-9-ethylcarbazole (Polysciences, Warrington, PA) containing hydrogen peroxide. Under observation through a dissecting microscope, red-brown–colored spots were counted as evidence of antigen-specific antibody-forming cells (AFC). The data were expressed as the mean number of AFC per 10⁵ cells in each experiment. 

Results

Flow Cytometric Analysis of T- and B-Cell Subsets in LG

Using multiple parameter flow cytometry, we initially examined lymphocytes from LG tissue for T-cell subsets, and the expression of TCR from these, and B cells. We found that the samples contained approximately 40% CD3⁺ T cells and 28% B220⁺ B cells (Fig. 1) . When different subsets of T cells in LG were analyzed according to the expression of CD4 and CD8 molecules, the proportion of CD4⁺CD8⁻ T cells was always higher than that of CD4⁻CD8⁺ T cells (Fig. 1) . In lymphocyte preparations, the T-cell population contained approximately 17% CD4⁺CD8⁻ and 11% CD4⁻CD8⁺ (Fig. 1) . 

We also examined TCR expression in different subsets of T cells derived from the LG samples. Twenty-five percent of CD3⁺ T cells were γδ T cells and 75% were αβ T cells (Table 1 and Fig. 1 ). Over 99% of CD4⁺CD8⁻ T cells expressed αβ TCR (Table 1 and Fig. 1 ). On the other hand, it is notable that approximately 22% of CD4⁻CD8⁺ T cells expressed γ and δ heterodimer chains of TCR in addition toαβ TCR–bearing cells (78%; Table 1 and Fig. 1 ). Furthermore, the CD4⁻CD8⁻ T-cell fraction from LG samples expressed γδ TCR. The proportion of NK1.1⁺ αβ cells was higher (3%) in LG samples than those obtained from the SMG (0.5%; Table 2 and Fig. 1 ). 

High Frequency of B-1 Lineage of IgA-Committed B Cells in LG

As described above, when the MC isolated from the LG tissue were examined for the B cells, approximately 28% of cells were B220⁺ B cells (Fig. 1) . Several studies have shown that sIgA⁺ B cells develop from the B-1 and B-2 lineages, ^{10 11 12 29 30 31 32 33} so we thought that it was important to examine which subset of B cells gives rise to sIgA⁺ B cells in the LG. In MC isolated from LG, the proportion of sIgA⁺ B cells was 4% (Figs. 1 and 2) . Of these, the overwhelming majority, with 3.9% of the total cell count, were B-1; B-2 cells accounted for only 0.1% of the total (Fig. 2) . Furthermore, we found that the dominant fraction (97.5%) of sIgA⁺ B-1 cells consisted of B-1a (B220^lowCD5⁺), whereas B-1b cells (B220^highCD5⁻) were present in much lower frequencies (2.5%). 

Induction of CT-Specific IgM, IgG, and IgA in Tear-Wash Samples by Ocular Immunization

The results described above show that samples obtained from LG contain all the immunocompetent cells that are required for the induction and regulation of the IgA immune response. Consequently, our next experiment was intended to examine whether ocular immunization can induce a system-wide antigen-specific immune response in the mucosal immune system. To assess the effectiveness of ocular immunization for the induction of antigen-specific IgA immune responses, a group of mice was immunized with CT via the ocular surface. As a control, a separate group of mice was nasally immunized with CT. The serum of ocularly and nasally immunized mice yielded CT-specific IgM, IgG, and IgA (Fig. 3) . Moreover, we detected high titers of CT-specific IgA antibodies in the tear-wash and saliva samples of mice that had received ocular and nasal immunization (Fig. 4) . In the examination for the IgG isotype in tear-wash and saliva samples, after ocular and nasal immunization, CT-specific IgG were induced. We noted, however, that the proportion of antigen-specific IgG antibodies was lower in the tear samples than in the saliva samples (Fig. 4) . 

Induction of CT-Specific IgA-Producing Cells in LG by Ocular Immunization

After this discovery of CT-specific IgG and IgA antibodies in serum and tear-wash samples following ocular immunization, we thought that it was important to elucidate the presence and frequency of AFC in the relevant tissues of the immunized mice. When splenic MC were analyzed, CT-specific IgM-, IgG-, and IgA-producing cells were found in ocularly and nasally immunized mice. On the other hand, we found that the dominant isotype of antigen-specific immunoglobulin-producing cells in LG and SMG samples was IgA, followed by small numbers of IgG and IgM AFC (Fig. 5) . These findings show that ocular immunization is an effective way of inducing antigen-specific antibody-producing cells (e.g., IgA and IgG) in both the mucosal (e.g., LG and SMG) and systemic (e.g., spleen) compartments of the immune system. 

Discussion

An important finding of the present study was that in LG samples a high frequency of sIgA⁺ B cells exists that are associated with B-1 cells. Found predominantly in peripheral tissues such as the peritoneal and pleural cavities, ³² B-1 cells constitute a distinct B-cell population and functional properties that differ in several ways to conventional B cells (B-2 cells). ^{8 12 31} Research has shown that B-1 cells from the peritoneal cavity home into the i-LP and become IgA plasma cells. ^{12 32 33} In a recent and separate study, we demonstrated that the B-1 lineage of sIgA⁺ B cells are predominantly found in the mucosal effector sites (e.g., i-LP and salivary glands), whereas sIgA⁺ B cells associated with the B-2 lineage reside in mucosal inductive (e.g., GALT or PP) and mucosal effector (e.g., i-LP, nasal passage and salivary glands) tissues. ¹⁰ Thus, the B-1 and B-2 lineages of sIgA⁺ B cells may represent common mucosal immune system (CMIS)–independent (B-1) and –dependent (B-2) IgA-committed B cells, respectively. ¹⁰ The results of this experiment show that in LG, an effector site for the ocular surface, the dominant fraction of sIgA⁺ B cells consists of B-1 cells (Fig. 2) . Among these sIgA⁺ B cells, we found that most had the surface phenotype characteristics of B-1a cells (i.e., B220^low, CD5⁺). Moreover, interleukin-5 (IL-5) has been shown to induce the proliferation and differentiation of B-1 cells into immunoglobulin-producing cells, ^{29 34} including IgA-producing cells. ^{10 30} In consideration of these findings, it is reasonable to suggest, especially on the evidence of B-1a cells that we found, that B-1 cells isolated from LG are a subset of CMIS-independent B cells that have migrated from the peritoneal cavity or other unidentified site and become IgA plasma cells under the influence of Th2 cytokine (e.g., IL-5), which is produced by LG CD4⁺ T cells. 

In addition to T helper (e.g., CD4⁺CD8⁻) and cytotoxic T (e.g., CD4⁻CD8⁺) cells, another interesting finding, as shown in Table 2 , was that the proportion of NK1.1⁺αβ T cells was higher (3%) in LG than in SG tissues (0.5%). A report has described that V_α14 NKT cells show proliferative responses to galactosylceramide (GalCer) and produce large amounts of IL-4 and interferon-γ, and also, upon stimulation with GalCer, kill Yac-1 cells. ³⁰ It has been generally considered that V_α14 NKT cells directly kill target tumor cells by an NK-like mechanism, and there is evidence that these cells, in tumor-bearing mice that have been treated with GalCer, inhibit tumor growth and metastasis. ³⁰ Based on these characteristics, it is suggested that NK1.1⁺ αβ T cells from LG may play an important role as the first line of defense of the ocular surface. 

We also found that the CD3⁺ T cells from LG contained a high proportion of γδ T cells: Approximately 25% of CD3⁺ T cells in LG expressed γδ heterodimer chains of TCR. These γδ T cells are rarely found in systemic lymphoid tissues, whereas mucosa-associated tissues, including the intraepithelial lymphocytes (IEL) of murine small intestine ^{36 37 38 39} and SMG, ^{17 18 22} contain this subset. Although the precise nature and function of these γδ IEL T cells is not well understood, it seems that these T lymphocytes possess cytolytic activity. ^{36 40 41} Furthermore, in separate studies we have found that the γδ T cells from the IEL of mice that had been orally primed with T-cell–dependent (TD) antigen possess the ability to convert oral tolerance to antigen-specific immune responses. ^{38 42} Thus, γδ T cells may play an important role as regulatory T cells that protect (or enhance) CD4⁺ T helper cells for maximum IgA response at IgA-effector sites such as in LG under the presence of oral tolerance. ^{38 42} Furthermore, removal of γδ T cells resulted in the reduction of IgA response in mucosal effector sites (e.g., i-LP and SMG). ⁴³ The presence of γδ T cells in LG might be an essential factor for the maintenance of the high level of IgA antibody production seen in this tissue. 

In this study, we also determined the effectiveness of ocular immunization for the induction of mucosal and systemic immune responses. Ocular administration induced a level of antigen-specific mucosal and systemic immune responses that was comparable to those of nasal immunization. Although the dominant isotype of CT-specific antibody response was IgA, found in mucosa-associated tissue (e.g., LG and SMG; Fig. 5 ), the proportion of antigen-specific IgG antibodies depends on the site of secretion (e.g., the levels are different in tears and saliva). This is possibly because saliva contains, in addition to antibodies locally produced in SMG tissues, serum-derived CT-specific IgG antibodies from crevicular fluids that are present in the oral cavity. Samples derived from SMG tissues yielded CT-specific IgA, but we found no IgG antibody–forming cells (Fig. 5) . On the other hand, antigen-specific IgG antibodies were detected in the serum of ocularly immunized mice. The majority of antigen-specific antibodies in tears are likely to derive from IgA produced locally in LG tissues. We verified that samples derived directly from LG tissues, having a predominance of total IgA antibody–forming cells, had very low numbers of total IgG-producing cells (data not shown). It is generally accepted for humans, rats, mice, and rabbits that LG tissue is a site for IgA rather than IgG. ⁴ Consequently, in tear-wash samples of mice ocularly immunized with CT, the major isotype of antigen-specific antibody was IgA. 

Several animal models have also been used to assess tear antibody response after immunization or infection at the ocular surface. In guinea pigs, only the use of live Chlamydia psittaci induced tear IgA antibodies ^{23 24 25} after ocular immunization. It has also been shown that ocular application is a more effective route for eliciting IgA antibody response in tears than gastrointestinal, subconjunctival, and intraperitoneal immunization. ⁴⁴ Ocular administration to guinea pigs with dead organisms has failed to induce tear antibodies, ²³ although a report suggests that inactivated Chlamydia trachomatis induces tear antibodies in owl monkeys. ⁴⁵ In a cynomolgus monkey model, ocular chlamydial infection induced IgA, IgG, and IgM responses in serum and tears. ⁴⁶ Taken together, these findings imply that ocular immunization is an effective means for the induction of antigen-specific immune responses. 

It is possible, however, that ocularly administered antigen may pass through the nasolacrimal duct and stimulate mucosa-associated lymphoreticular tissues in intestinal (e.g., GALT) and nasopharyngeal (e.g., nasopharyngeal-associated lymphoreticular tissue) tracts and subsequently stimulate the CMIS to induce an antigen-specific IgA response. Alternately, ocularly administrated antigen may directly stimulate lymphocytes that reside in a locally situated site of induction. Here, conjunctiva-associated lymphoreticular tissue (CALT) has been suggested as a potential site of induction for the ocular immune system owing to the histologic characteristics that it shares with GALT. ⁴ Ocularly administrated antigen could trigger immunocompetent cells in CALT to subsequently seed to the LG. ⁵  

Although the precise mechanism by which ocular immunization elicits tear IgA antibody responses remains unknown, our results suggest that LGs contain all the immunocompetent cells that are needed for the production of antigen-specific IgA antibodies after ocular immunization. Moreover, this means of immunization could prove a useful way to stimulate antigen-specific IgA-producing cells in the LG against ocular infection. Bearing in mind the different distribution of B-1 cells, including B-1a and B-1b cells, and B-2 cells in samples from mucosal (e.g., LG, GALT, and i-LP) and systemic (e.g., spleen) tissues, it would be worthwhile to continue examination to elucidate the contribution these different subsets of sIgA⁺ B cells to the induction of antigen-specific mucosal and systemic immune responses to TD, T-cell–independent type 1 (TI-1), and TI-2 antigens introduced by ocular immunization. 

In summary, the present study has added to previously important findings ^{4 5 10 14 15 16 23 24 25} by revealing several unique immunologic features of the lacrimal glands, including a high frequency of sIgA⁺ B-1a cells; the presence γδT cells; the occurrence of NK1.1 αβ T cells; and the effectiveness of ocular immunization for the induction of antigen-specific mucosal IgA and serum IgG antibody responses. Here we have discussed these new findings in the context of the LG being an important part of the immune system, but further investigation, both molecular and cellular, is needed to elucidate any cross-talk mechanisms that may involve the ocular mucosal and the systemic immune system. 

 Supported by grants from the Ministry of Education, Science, Sports and Culture, the Ministry of Health and Welfare, and the Organization for Pharmaceutical Safety and Research, Japan.

 Submitted for publication January 14, 1999; revised June 8, 1999; accepted July 6, 1999.

 Commercial relationships policy: N.

 Corresponding author: Hiroshi Kiyono, Department of Mucosal Immunology, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan.

 

Figure 1.

 

Flow cytometric analysis of T and B cells in LG, according to the expression of CD3, B220, CD5, CD4, CD8, αβ TCR, γδ TCR, IgA, and NK1.1. Two- or three-color immunofluorescence analysis was performed to characterize different subsets of T and B cells. Each analysis was performed at least three times to verify the results. Results represent the values (mean) from 3 separate experiments (10 mice/group).

Table 1.

 

Analysis of γδ TCR and αβ TCR Expression on Different T-Cell Subsets in Murine LG and SMG

T-Cell Subsets	TCR	Proportion of γδ and αβ TCR+ Cells (%)
				LG	SMG
CD3+	γδ	25.0	23.8
	αβ	75.0	76.2
CD4+CD8−	γδ	0–1	0–2
	αβ	99–100	98–100
CD4−CD8+	γδ	22.0	24.8
	αβ	78.0	75.2
CD4−CD8−	γδ	90.0	98–100
	αβ	10.0	0–2

Table 2.

 

Occurrence of NK1.1⁺ αβ⁺ T Cells in Murine LG but not SMG

	Percentage of Positive Cells in MC
		LG	SMG
NK1.1+ αβ TCR	2.9 ± 0.2	0.5 ± 0.2

Figure 2.

 

Flow cytometric analysis of B cells in LG, according to the expression of B220, IgA, and CD5. Two- or three-color immunofluorescence analysis was performed to characterize B cells. Results represent the mean values from three separate experiments (10 mice/group).

Figure 3.

 

Ocular immunization with CT induces antigen-specific antibody responses in the systemic compartment. Levels of serum antibodies by ocular immunization (□) and nasal immunization (▪) were analyzed by ELISA. Results represent the mean values from 3 separate experiments (10 mice/group).

Figure 4.

 

Induction of CT-specific IgG and IgA antibodies in mucosal secretions, tear washings, and saliva by ocular immunization. Levels of secretory antibodies induced by ocular immunization (□) and nasal immunization (▪) were analyzed by ELISA. Results represent the mean values from 3 separate experiments (10 mice/group).

Figure 5.

 

Proportion of antigen-specific IgA, IgG, and IgM AFC in LG, SMG, and spleens of mice ocularly immunized with CT. Numbers of CT-specific AFC by ocular immunization (□) and nasal immunization (▪) were analyzed by ELISPOT assay. Results represent the mean values from 3 separate experiments (10 mice/group).

The authors thank Satoru Kodama and Koichi Iwatani for their great help and advice on our experiments.