The external part of the eye shares mucosa-associated common characteristics and is an obvious entry site for foreign Ags. We assessed the potential of eyedrop vaccination for effective delivery of vaccines against viral or bacterial infection in mice. Both OVA-specific IgG Ab in serum and IgA Ab in mucosal compartments were induced by eyedrops of OVA with cholera toxin (CT). Eyedrop vaccination of influenza A/PR/8 virus (H1N1) induced both influenza virus-specific systemic and mucosal Ab responses and protected mice completely against respiratory infection with influenza A/PR/8 virus. In addition, eyedrop vaccination of attenuated Salmonella vaccine strains induced LPS-specific Ab and complete protection against oral challenge of virulent Salmonella. Unlike with the intranasal route, eyedrop vaccinations did not redirect administered Ag into the CNS in the presence of CT. When mice were vaccinated by eyedrop, even after the occlusion of tear drainage from eye to nose, Ag-specific systemic IgG and mucosal IgA Abs could be induced effectively. Of note, eyedrops with OVA plus CT induced organogenesis of conjunctiva-associated lymphoid tissue and increased microfold cell-like cells on the conjunctiva-associated lymphoid tissue in the nictitating membrane on conjunctiva, the mucosal side of the external eye. On the basis of these findings, we propose that the eyedrop route is an alternative to mucosal routes for administering vaccines.
A vaccine that induces an immune response by fortifying mucosal immunity is an effective way of targeting the pathogen before infection occurs (1, 2). Mucosal vaccination, in contrast to parenteral vaccination, is of particular interest because it can elicit both systemic and mucosal immune responses, mainly secretory IgA (sIgA) Abs, at the very portal of entry of most infectious pathogens (3). Vaccine development has lagged behind the rapidity of disease propagation in the era of global travel. Mucosal vaccination, which is easy to administer and does not require special training, has become a strategy to thwart new pathogen strains before they become pandemic.
The eye mucosa is a possible route for mucosal vaccine because it is an important entry point for environmental Ags and infectious materials occupying most of the external ocular surface (4–6). The conjunctiva, part of the eye mucosa, has immunologic features in common with other mucosal tissues. The conjunctiva has CD8+ T cells in the epithelium, equal proportions of CD4+ and CD8+ T cells, B cells, and mast cells in the lamina propria, and dendritic cells (DCs) and Langerhans cells (7, 8). As such, the conjunctiva is part of the mucosal barrier that is exposed to the external environment and shares many common immunologic features of other mucosal compartments. Previous studies showed successful protection by eyedrop vaccination in avian and bovine models (9, 10). However, the underlying mechanism of the induction of acquired immune responses and the systematic comparison of conjunctival and intranasal (i.n.) routes are not yet elucidated (11–13).
In this study, we assessed whether eyedrop administration on the eye mucosa can induce Ag-specific immunity and protective efficacy. Eyedrop administration of a prototype protein Ag plus cholera toxin (CT) induced a broad range of immune responses in both mucosal and systemic tissues. In addition, eyedrop vaccination of influenza A/PR/8 virus and recombinant Salmonella strains protected mice against respiratory challenge of influenza virus and oral challenge of Salmonella, respectively. In contrast to i.n. administered vaccines, we showed that eyedrop vaccination poses no risk of Ag redirection to the CNS in the presence of CT. On the basis of our findings, we propose that the eye mucosa is a good candidate for mucosal vaccine delivery for inducing protective immunity and theoretically a safe alternative for vaccine delivery targeting viral or bacterial infection.
Materials and Methods
Specific pathogen-free BALB/c mice, aged 6–10 wk, were purchased from Charles River Laboratories (Orient Bio, Sungnam, Korea). CCR6−/− mice on a C57BL/6 background were purchased from The Jackson Laboratory (Bar Harbor, ME). polymeric IgR (pIgR)−/− mice were kindly provided by Dr. Masanobu Nanno (Yakult Central Institute for Microbiological Research, Tokyo, Japan). Dr. Kazuhiko Yamamoto (University of Tokyo, Tokyo, Japan) generously made available the OVA-TCR transgenic mice (DO11.10) on a BALB/c background expressing OVA epitope (323-339), and Dr. Martin Lipp (Max Delbruck Center for Molecular Medicine, Berlin, Germany) generously provided the CCR7−/− mice on a C57BL/6 background. All mice were maintained in the experimental animal facility under specific pathogen-free conditions at the International Vaccine Institute (Seoul, Korea) and received sterilized food (Certified Diet MF; Oriental Yeast, Osaka, Japan) and filtered tap water ad libitum. All animal experiments were approved by the Animal Research Committee of the International Vaccine Institute (Seoul, Korea).
Prior to experimental manipulation, mice were anesthetized by i.p. injection of ketamine (100 mg/kg body weight) and xylazine hydrochloride (10 mg/kg body weight). For conjunctival immunization, 100 μg OVA (Sigma-Aldrich, St. Louis, MO) and 2 μg CT (List Biological Laboratories, Campbell, CA) suspended in 5 μl PBS were dropped weekly for 3 consecutive weeks onto a conjunctival sac by micropipette. In some experiments, mice were immunized with 0.1 × LD50 (500 PFU) live A/PR/8 virus [A/Puerto Rico/8/34 (H1N1)] or recombinant attenuated Salmonella enterica serovar Typhimurium [χ9241 ΔpabA1516 ΔpabB232 ΔasdA16 ΔaraBAD23 ΔrelA198::araCPBADlacI(ATG)TT containing pYA3802; 1 × 108 CFU] (14, 15) suspended in 5 μl PBS.
Serum was obtained by retro-orbital bleeding. Tear-wash samples were obtained by lavaging with 10 μl PBS per eye. Saliva was obtained following i.p. injection of mice with pilocarpine (500 mg/kg body weight; Sigma-Aldrich). Fecal extract was obtained by adding weighed feces to PBS containing 0.1% sodium azide. The feces were mixed by vortexing and centrifuged, and the supernatants were collected for assay. Vaginal wash samples were collected by lavage with 100 μl PBS. After the mice were sacrificed, nasal wash samples were obtained by flushing 100 μl PBS through the anterior (oral) entrance of the nasal passages (NPs) using a pipette.
ELISA for detection of Ag-specific Ab
ELISA plates (Falcon, Franklin Lakes, NJ) were coated with OVA (100 μg/ml) or inactivated A/PR/8 (5 μg/ml) or LPS (1 μg/ml) in PBS and incubated overnight at 4°C. Blocking was done with 1% BSA (Sigma-Aldrich) in PBS, and 2-fold serially diluted samples were applied to plates. HRP-conjugated goat anti-mouse IgG or IgA Ab (Southern Biotechnology Associates, Birmingham, AL) was added to each well and incubated overnight at 4°C. For color development, tetramethylbenzidine solution (Moss, Pasadena, MD) was used. Then, plates were measured at 450 nm on an ELISA reader (Molecular Devices, Sunnyvale, CA) after addition of stopping solution (0.5 N HCl). Endpoint titers of Ag-specific Ab were expressed as reciprocal log2 titers of the last dilution that showed >0.1 absorbance over background levels. For detection of total IgA levels, plates were coated with goat F(ab′)2 anti-mouse Ig, and HRP-conjugated goat anti-mouse IgA Abs (Southern Biotechnology Associates) were used as detection Abs. To detect OVA-specific sIgA Ab levels, plates were coated with OVA (100 μg/ml), and goat anti-pIgR (R&D Systems, Minneapolis, MN) and HRP-conjugated rabbit anti-goat IgG (Southern Biotechnology Associates) were used as detection Abs.
In vitro T cell proliferation assay
Following eyedrop vaccination with a mixture of OVA and CT, DCs were isolated from submandibular lymph nodes (SMLNs), as well as jugular, mediastinal, axillary, inguinal, and iliac lymph nodes (LNs). CD4+KJ1.26+ T cells isolated from DO11.10 mice were labeled with CFSE (Invitrogen, Carlsbad, CA) for 15 min at 37°C and washed several times in PBS. The purified DCs (3 × 104 cells/well) were cocultured with CD4+KJ1.26+ T cells (2 × 105 cells/well) in the presence of OVA peptide (OVA323–339) for 2 d at 37°C. CFSE proliferation in each tissue was analyzed by FACSCalibur (BD Biosciences, Franklin Lakes, NJ). To assess OVA-specific T cell proliferation in vitro, CD4+ T cells isolated from SMLNs of vaccinated mice and CD3+ T cell-depleted splenocytes were prepared from naive mice. CD4+ T cell- (2 × 105 cells/well) and CD3+ T cell-depleted splenocytes (2.5 × 104 or 1.25 × 104 cells/well) were cocultured in 10% FBS containing RPMI 1640 for 3 d in the presence of OVA peptide. [3H]Thymidine incorporation was measured by scintillation counter (Perkin Elmer, Waltham, MA).
To determine whether eyedrop vaccination could redirect Ag into the CNS (16, 17), mice were given acridinium-labeled OVA (40 μg) plus CT (2 μg) or acridium-labeled CT (2 μg; List Biological Laboratories) via the i.n. (5 μl for each nostril) or eye route (5 μl for each eye). Then, lung and olfactory bulb were removed 24 h following immunization, as previously described (17). Tissues were weighed, and 200 μl CellLytic MT lysis buffer (Sigma-Aldrich) was added per 10 mg wet weight tissue. Tissues were homogenized and frozen at −20°C. After thawing, the homogenates were centrifuged at 10,000 × g for 10 min, and the supernatants were tested for light activity using the LMax II384 system (Molecular Devices).
Protection assay against influenza virus A/PR/8/34
At 2 wk after eyedrop immunization with 0.1 × LD50 (500 PFU) live A/PR/8 virus, anesthetized mice were challenged with live A/PR/8 virus suspension (20 × LD50; 1 × 105 PFU) by the i.n. route. Animals were monitored for weight loss and survival every day for 12 d.
Protection assay against the wild-type virulent S. typhimurium (UK-1) strain
Four weeks after eyedrop immunization with recombinant attenuated S. typhimurium vaccine (RASV) strain (1 × 108 CFU), the virulent UK-1 strain (1 × 107 CFU) was given orally for a challenge experiment. Body weight and survival were monitored every day for 8 d.
M cell staining
For the preparation of whole-mount staining, the small nictitating membrane of the conjunctiva of naive or immunized mice was fixed in 4% paraformaldehyde at 4°C for 1 h and dehydrated in glucose solutions (10, 20, and 30%). The specimens were embedded in O.C.T. Compound (Tissue-Tek, Tokyo, Japan) and stored at −80°C until processing. Cryostat sections (5 μm) were fixed in ice-cold acetone and blocked with FcRII/III mAb (BD Pharmingen, San Diego, CA) in PBS. For lectin staining, tissues were stained with 20 μg/μl tetramethylrhodamine isothiocyanate TRITC-conjugated Ulex europaeus agglutinin (UEA-1) (Vector Laboratories, Burlingame, CA) and FITC-conjugated wheat germ agglutinin (WGA). DAPI (Invitrogen) was used for nucleus staining. Sections were studied by confocal laser scanning microscope (Carl Zeiss, Jena, Germany).
Electron microscopic evaluations
Scanning and transmission electron microscopic (TEM) analyses were performed for the characterization of M cells. For scanning electron microscopic analysis, small fragments of the nictitating membrane of the conjunctiva were cleaned of mucus and fixed in 2% glutaraldehyde and 2% paraformaldehyde in PBS containing 100 mM HEPES for 1 h at reverse transcription. After being washed with PBS, specimens were treated with 1% osmium tetroxide for 1 h at reverse transcription and dehydrated in graded ethanol solution. Dehydrated tissues were critical point-dried with CO2, sputter-coated, and observed with a JSM 5410LV scanning electron microscope (JEOL, Tokyo, Japan). For TEM analysis, tissues fixed with 4% paraformaldehyde in PBS were immersed in 0.3% H2O2, diluted in methanol for 30 min to block endogenous peroxidase, incubated with 20 μg/ml UEA-1-HRP in PBS for 1 h, and then stained with 3,3′-diaminobenzidine tetrahydrochloride (DAB). After staining, the tissues were fixed overnight with 2% glutaraldehyde in 100 mM sodium phosphate buffer (pH 7.3) at 4°C. The tissues were subjected to TEM analysis (JEM 1010, JEOL).
Ag uptake in situ
S. typhimurium PhoPc strain transformed with the pKK GFP plasmid and GFP-expressing Yersinia pseudotuberculosis were prepared by the method described (18). GFP-expressing bacteria (5 × 108 CFU) were suspended in 5 μl PBS and inoculated into the conjunctival sac of anesthetized mice and incubated in situ for 10 min. Then, whole conjunctiva, including nictitating membranes, was removed and extensively washed with cold PBS and RPMI medium including gentamicin (50 μg/ml). Conjunctival tissues were fixed in 4% paraformaldehyde and dehydrated in glucose solutions (10, 20, and 30%). The specimens were embedded in O.C.T. compound (Tissue-Tek), and cryostat sections were labeled with UEA-1-TRITC. Whole-mounted segments and frozen sections were processed for confocal microscopy, as described above.
Data and statistical analyses
Data were expressed as the mean ± SD, and statistical analyses were done by the t test (Sigma plot program).
Significant induction of Ag-specific Ab responses by eyedrop administration
To assess the efficacy of eyedrop administration for inducing systemic and mucosal Ab responses, BALB/c mice were immunized three times at 1-wk intervals. At 1 wk after final immunization, the levels of Ag-specific Abs and the numbers of Ab-secreting cells (ASCs) were measured by ELISA and ELISPOT, respectively. Groups of mice given OVA (100 μg/dose) plus CT (2 μg/dose) by eyedrop showed significantly higher levels of OVA-specific IgG and IgA Ab in serum and IgA Ab in mucosal compartments (e.g., tear, nasal, saliva, and vaginal wash) than found in control mice given PBS or OVA alone (Fig. 1A). Levels of OVA-specific IgG and IgA Abs induced by eyedrop were identical to those elicited by i.n. administration with the same dose of OVA plus CT (Fig. 1A). Because some Abs in secretions may originate from plasma by transudation of Ig, ELISPOT assays were carried out to determine the contribution of the local plasma cell pool after eyedrop immunization. OVA-specific ASCs were counted in cell suspensions from spleen, SMLNs, NP, lacrimal gland, conjunctiva, and submandibular gland 2 wk after the final vaccination. Eyedrop immunization with OVA plus CT elicited higher numbers of OVA-specific IgA ASCs in all tested tissues than did PBS (Fig. 1B). We used pIgR−/− mice, in which the transepithelial transport of dimeric IgA is blocked, to measure whether IgA Abs induced by eyedrop are in secretory form (19, 20). As shown in Fig. 1C, OVA-specific IgA Abs were not detectable in tears or in nasal, saliva, and vaginal wash samples of pIgR−/− mice, but they produced identical levels of serum IgG and IgA Abs following eyedrop vaccination. Although some IgA Abs were detected in the nasal wash of pIgR−/− mice, this was expected because of inevitable blood contamination during nasal wash preparation. To exclude the possibility of nonspecific increase of IgA in mucosal secretions, we determined the ratio of OVA-specific sIgA in total IgA Abs (Fig. 1D). These results indicate that most Ag-specific IgA Abs elicited in the mucosal compartments by eyedrop are in the dimeric form of sIgA. Taken together, eyedrop vaccination elicited Ag-specific IgG and sIgA Abs in systemic and mucosal tissues effectively, and the increased Ab levels were similar to those induced by i.n. administration.
Ags drain to SMLNs and IgA response is dependent on CCR6 rather than CCR7 in eyedrop immunization
To determine the anatomic location of draining LNs, where Ag is presented to T cells following eyedrop administration, DCs were harvested from SMLNs and from jugular, mediastinal, axillary, inguinal, mesenteric, or iliac LNs of BALB/c mice 24 h after treatment with a single dose of OVA plus CT. Then, each DC was cocultured with CD4+KJ1.26+ cells isolated from OVA-TCR transgenic mice. Only CD4+KJ1.26+ cells cocultured with DCs isolated from SMLNs showed significant proliferation of CD4+ T cells when compared with the CD4+ T cell alone group. The CD4+KJ1.26+ cells cocultured with DCs from other LNs, including jugular, mediastinal, axillary, inguinal, mesenteric, and iliac, did not show any significant proliferation (Fig. 2A). To address OVA-specific CD4+ T cell proliferation, CD4+ T cells were isolated from SMLNs of naive and eyedrop-vaccinated mice with OVA plus CT three times each week and cocultured with CD3+ T cell-depleted splenocytes from naive mice (Fig. 2B). Higher levels of proliferation were shown in the CD4+ T cells from eyedrop-vaccinated mice than in those from wild-type mice. These OVA-specific proliferations were more enhanced when professional APCs were added (Fig. 2B). Recent studies by ourselves and others indicate that different chemokine–chemokine receptors are involved in eliciting efficient Ag-specific systemic and mucosal immune responses (including T and B cells) by i.n., oral, or sublingual routes (17, 21–23). We thus investigated whether OVA-specific Ab responses could be dependent on CCR6 or CCR7 signaling in eyedrop vaccination by analyzing the OVA-specific Ab titers in both systemic (i.e., serum) and mucosal compartments following eyedrop administration of OVA plus CT using CCR6−/− and CCR7−/− mice. Of note, there were significantly lower levels of OVA-specific IgA Abs in mucosal compartments of CCR6−/− mice than in wild-type and CCR7−/− mice (Fig. 2C). In contrast, no significant differences were found in Ag-specific IgG Abs in sera of all groups of mice (Fig. 2C). These data demonstrate that induction of mucosal IgA Ab responses by eyedrop are tightly regulated in a CCR6-dependent manner, but not by CCR7.
Eyedrop vaccination of live A/PR/8 virus protects mice against lethal challenge with influenza virus
To address the efficacy and safety of the eyedrop route for delivery of live influenza virus vaccine, groups of mice were administered live A/PR/8 virus (0.1 × LD50; 500 PFU) without any boosting. This dose elicited much higher levels of virus-specific IgG Abs in serum and IgA Abs in mucosal secretions (e.g., tears, lung wash, nasal wash, and saliva) of mice given A/PR/8 than was observed in those given PBS (Fig. 3A). Of interest, eyedrop immunization with live A/PR/8 virus resulted in no body weight loss and 100% protection (Fig. 3B) and cleared influenza virus efficiently from the bronchial alveolar lavage fluid (Fig. 3C) against lethal i.n. challenge with A/PR/8 virus (20 × LD50; 1 × 105 PFU). We also compared the body weight changes between eyedrop-vaccinated and i.n.-vaccinated mice given live A/PR/8 virus (0.1 × LD50; 500 PFU). In contrast to the i.n. route, mice given vaccine by eyedrop did not lose body weight (Fig. 3D). Overall, eyedrop vaccination with live A/PR/8 virus is safer than i.n. inoculation and is highly effective in protecting mice against lethal respiratory challenge with influenza virus.
Eyedrop administration of attenuated Salmonella vaccine strain protects mice against lethal challenge with Salmonella
To further investigate the efficacy of eyedrop vaccination against bacterial infection, we immunized mice by a single eyedrop using the RASV strain (14, 15). At 4 wk after immunization, levels of LPS-specific IgG and IgA Abs in the serum and IgA Abs in the mucosal secretions were significantly increased in mice vaccinated by eyedrop with RASV, compared with those given PBS (Fig. 4A). We then challenged groups of vaccinated mice with a lethal dose of virulent Salmonella strain (UK-1, 1 × 107 CFU) by the oral route. Eyedrop administration with RASV resulted in 100% survival without any loss of body weight against lethal oral challenge with Salmonella (Fig. 4B). When these results are considered together, eyedrop vaccination may also be an effective and innovative mucosal vaccine delivery system against enteric infectious bacterial pathogens.
Eyedrop-administered Ag passes into the nasal cavity minimally and does not enter the CNS
The trafficking of vaccine Ag into the CNS following i.n. administration with CT or the heat-labile enterotoxin (LT) as adjuvant raises safety concerns (24, 25) and poses a serious obstacle to the clinical use of this route for vaccine delivery. Thus, we sought to determine if eyedrop vaccination would be similarly limited by Ag trafficking to the CNS. We administered acridinium-labeled OVA plus CT to groups of mice via either the eyedrop or the i.n. route and obtained nasal washes at 30, 60, 120, and 240 min after inoculation. The levels of acridinium in the nasal wash specimens were high in the i.n. group and significantly lower in the eyedrop group at all measurement times (Fig. 5A). At 24 h after inoculation, i.n. administration of acridinium-labeled OVA resulted in the accumulation of Ag in lung and olfactory bulbs (Fig. 5B). In contrast, acridinium-labeled OVA was undetectable in both sites following eyedrop administration, demonstrating that this route did not redirect Ag into the CNS. When the same experiment was performed with acridinium-labeled CT alone, eyedrop administration did not redirect CT into the CNS, whereas i.n. administration did (Fig. 5B). To directly prove the role of eye mucosa in inducing systemic and mucosal immune responses, we blocked the passage of eyedrop Ag into nasal mucosa by suturing four nasolacrimal duct puncta, blocking the draining route from eye to nose. To confirm the blockade of passage into the nasal mucosa, acridinium levels were determined in the NP 24 h following eyedrop of acridinium-labeled OVA plus CT. As expected, significantly less acridinium was found in the sutured mice compared with those not sutured (Fig. 5C). Of note, eyedrop inoculation of sutured mice induced levels of IgA Abs in the serum and mucosal secretions identical to those in the nonsutured mice (Fig. 5D). These results strongly suggest that effective induction of T and B cell responses by eyedrop administration is completely independent of the inductive capacity of the nasal cavity.
Organogenesis of conjunctiva-associated lymphoid tissue and increased M cell- and goblet cell-like cells on the nictitating membrane of eye mucosa after eyedrop immunization
To evaluate the structure of eye mucosa as an Ag delivery site involving protective immunity by eyedrop vaccination, we examined whole-eye specimens by histologic staining and electron microscopic methods. The mucosal surfaces of Ag-treated eyes showed lymphoid tissues in the nictitating membrane (the third eyelid; homologous with the plica semilunaris in humans) (Fig. 6A,a). The development of conjunctiva-associated lymphoid tissue (CALT) during eyedrop immunization was observed in a time-dependent manner (Fig. 6A,b–d). In this tissue, CALT is one of the peripheral lymphoid organs that can contain M cells with Ag sampling function. Scanning electron microscopy demonstrated increased numbers of two different types of cells on the nictitating membrane after eyedrop administration of OVA plus CT (Fig. 6B,c, 6B,d), when compared with PBS-treated mice (Fig. 6B,a, 6B,b). We found epithelial cells having depressed surfaces with irregular and short microvilli (Fig. 6B,d, arrow, and 6B,e) and clusters of cells with more rounded tops and straighter microvilli, which were observed only in conjunctiva goblet cells (Fig. 6B,d, arrowhead, and 6B,f) (26, 27). TEM analysis after the binding of HRP-conjugated UEA-1 Ab (DAB reacted) showed a typical morphology of UEA-1− conjunctival epithelial cells following eyedrop administration of OVA plus CT (Fig. 7A, 7B). Of note, M cells had more irregular and coarse UEA-1+ microvilli (Fig. 7C, 7D) with infiltrating mononuclear cells beneath the cell membrane (Fig. 7C, arrowheads), much like a feature of pocket lymphocytes in intestinal M cells (18, 28). In addition, clustered cells had straighter microvilli, and goblet cells possessed many vesicles in their cytoplasm (Fig. 7E, 7F).
UEA-1+ M cells can sample and internalize the rSalmonella-GFP
Because expression of α (1, 2) fucose is a hallmark of murine M cells, lectin UEA-1 possessing affinity for α (1, 2) fucose is routinely used to detect such cells in mice (29). UEA-1 also reacts to goblet cells; however, these cells possess strong affinity to epithelial cell-specific lectin, such as WGA (18). Using confocal image analysis of whole-mount murine conjunctiva stained with TRITC-conjugated UEA-1 (red) and FITC-labeled WGA (green) Ab, we found UEA-1+WGA− cells, representing M cell-like cells, in the nictitating membrane of mouse conjunctiva (Fig. 8A,a and 8A,c, arrows). Analysis of frozen sections also revealed the presence of UEA-1+WGA− cells in the nictitating membrane of conjunctiva (Fig. 8A,b and 8A,d, arrows). These conjunctival M cell-like cells (UEA-1+WGA−) clearly differed from goblet cells (UEA-1+WGA+) (Fig. 8A, arrowhead). Of note, there were more UEA-1+WGA− cells in the nictitating membrane of conjunctiva of whole-mount tissues (Fig. 8A,c) and in cross-sections (Fig. 8A,d) after eyedrop administration of OVA and CT than were found in PBS-treated mice (Fig. 8A,a, 8A,b). An additional experiment was performed to gauge the ability of UEA-1+ cells in the murine conjunctiva to take up pathogenic microorganisms. Mice were inoculated by eyedrop with S. typhimurium or Y. pseudotuberculosis expressing GFP plasmid. Ten minutes later, sequential immunohistologic analyses of conjunctiva revealed the presence of S. typhimurium in UEA-1+ cells in the whole mount of murine conjunctival epithelium (Fig. 8B,a, 8B,b). In addition, Yersinia-GFP was also specifically adhered to UEA-1+ cells of conjunctival epithelium (Fig. 8B,c, 8B,d). To show the ability of UEA-1+ cells to internalize bacteria, we performed eyedrop administration using rSalmonella-GFP and analyzed the results by vertical section. Of note, rSalmonella-GFP was located in the intracellular region of UEA-1+ cells on the conjunctiva (Fig. 8B e). Taken together, these results indicate that UEA-1+ cells in the nictitating membrane of conjunctiva have the ability to take up and internalize bacteria from the lumen and are involved in the induction of protective immunity via eye mucosa after eyedrop vaccination.
Mucosal vaccination has the advantage of producing both sIgA Abs in mucosal compartments and IgG Abs in serum, in contrast to parenteral vaccines, which induce only serum IgG Abs. Such sIgA Ab responses play an important role in protecting against the invading external pathogen on the mucosal surface. No changes in behavior, weight loss, or local inflammation were observed in mice after eyedrop administration. Our results provide the evidence that eyedrop vaccination induces both mucosal and systemic immunities and that it is protective against virus (i.e., influenza) and bacteria (i.e., Salmonella) infections in mice.
To determine the mechanism of vaccine delivery by a novel route, a crucial step is to demonstrate the Ag delivery pathway by sampling cells and the draining lymphoid organs for Ag presentation. The fact that Ag-specific CD4+ T cell expansion occurred only in the SMLNs after eyedrop application with OVA plus CT (Fig. 2A) suggests that SMLNs target draining LNs during eyedrop vaccination. As a drainage site of Ag from eye, SMLNs were the most commonly mentioned candidate in an earlier study and were suggested as the main priming site of donor Ag in corneal allograft (30). In addition, following posterior chamber injection of Ag, adopted KJ1-26+ cells accumulated primarily in the SMLNs within 3 d (31). Further, cells presenting OVA peptide (OVA323–339) in vivo were found only in the SMLN but not in other LN, spleen, or nasal-associated lymphoid tissue after conjunctival application of OVA together with colonization-factor Ag in the mice (32). Together these findings suggest that the corneal conjunctiva and the posterior eye chamber share the common draining SMLN for induction of immunity or tolerance.
These data also suggest that eyedrop-administered vaccine enters through the eye mucosa, reaches the SMLN, and induces effective immunity against a pathogen. However, because of tear drainage from eye to nose, it is important to show that the eyedrop route does not share the characteristics of the i.n. pathway. Our results indicate that the eyedrop route relies on the chemokine receptor CCR6, contrary to CCR7 dependency in the i.n. or sublingual pathways (Fig. 2B) (21, 33). Eyedrop vaccination did not redirect Ag into CNS (Fig. 5B), as reported for i.n. delivery when CT and LT were used as adjuvant, which was considered a possible reason for nerve damage in one clinical trial (24, 25). Normal levels of Ag-specific Ab responses in serum and mucosal secretions were induced by eyedrop vaccination after suturing four puncta of the nasolacrimal duct (Fig. 5C). Finally, live virus used in 0.1 × LD50 doses for i.n. vaccination caused body weight loss, but eyedrop inoculation had no effect on body weight (Fig. 3B). Therefore, these results imply that eyedrop vaccination is unique and a safer method of mucosal vaccine delivery independent of nasal mucosa.
Nagatake et al. (34) showed that tear duct-associated lymphoid tissues (TALTs), which are located in the lacrimal sac, play a role in the induction of Ag-specific immune response against Ag found on the ocular surface. TALT is the site of ocular Ag uptake by M cell-like cells and also the site for induction of Ag-specific IgA and CD4+ cells after ocular immunization. Thus, it is possible that TALT could be one candidate for inductive eyedrop vaccination. In addition to TALT, our study demonstrated, after Ag application to the conjunctiva by eyedrop vaccination, that lymphoid follicles (i.e., CALT) on the nictitating membranes of murine conjunctiva became sufficiently large to be detectable by microscope (Fig. 6Aa, 6Ad). At steady state, 8% of mice (2 of 25) showed organized CALT in the nictitating membranes of conjunctiva (data not shown). A previous study revealed the existence of follicles in the nictitating membrane of mouse conjunctiva and their characteristic plasticity in terms of size and numbers after conjunctiva OVA challenge (35, 36). Furthermore, Steven et al. (37) also showed induction of CALT in 70% of mice by external application of Chlamydia trachomatis serovar C or a solution of OVA and B subunit of CT. It seems likely that CALT could be another candidate as an inductive site for eyedrop vaccination.
In this study, we also found the existence of M cell-like cells (i.e., UEA-1+WGA− cells) in the CALT of the nictitating membrane of mouse conjunctiva. M cells, a unique epithelial cell type specializing in Ag sampling, have been discovered in follicle-associated epithelium of TALT, GALT, and nasal-associated lymphoid tissue (34, 38, 39). The M cell-like cells found in our study after eyedrop application were depressed from adjacent epithelial cells and had irregular and longer microvilli (Figs. 6B, 7). Others have reported the presence of M cells in conjunctival mucosa in several species, including rabbit, guinea pig, and canine (26, 40–42). M cells in the guinea pig conjunctiva were located in follicle-associated epithelium and showed variably sized microvilli (100 nm–1 μm), in contrast to epithelial cells (350 nm) (26, 43). These characteristics were also found in M cells in rabbit conjunctiva, which had longer and more irregular microvilli than did epithelial cells (41). Further, M cells in the conjunctiva of both guinea pig and rabbit had properties for endocytosis and/or transcytosis. In contrast, canine M cells had shorter and attenuated microvilli compared with M cells from the other two species (40). In our present study, microvilli of M cells in murine conjunctiva were longer and more irregular than those of the adjacent epithelial cells, suggesting more similarity to rabbits and guinea pigs than to canines. Most importantly, UEA-1+WGA− M cell-like cells in the nictitating membrane of mouse conjunctiva are able to sample and internalize Ag, such as S. typhimurium or Y. pseudotuberculosis (Fig. 8). Consequently, the M cell-like cells in the nictitating membrane of conjunctiva might be directly involved in the induction of protective immunity via eye mucosa after eyedrop vaccination.
In summary, this study shows that eye mucosa has a function as an inductive site for mucosal and systemic immunity. Although eyedrop vaccination needs to be elucidated further for usefulness and limitations in applicable populations, it could be an alternative method for induction of immune responses and could be used for protection against influenza virus and Salmonella bacterial infection. In view of the recent progress in our understanding of immunity and tolerance by the oral, i.n., and sublingual routes, eyedrop administration might be considered as a strategy for mucosal vaccination.
Disclosures The authors have no financial conflicts of interest.
This work was supported by a grant from the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare and Family Affairs, Republic of Korea (A084790).
Abbreviations used in this paper:
conjunctiva-associated lymphoid tissue
recombinant attenuated Salmonella typhimurium vaccine
submandibular lymph node
tear duct-associated lymphoid tissue
transmission electron microscope
Ulex europaeus agglutinin
wheat germ agglutinin.