Mucosal Immunity in the Ocular System

I. Introduction

Acquired immunity at the ocular surface is mediated primarily by the mucosal immune system. The linkage of ocular subcompartments to the mucosal immune network is now well established (Mestecky et al., 1978; Montgomery, et al., 1983, Montgomery, et al., 1985; Gregory and Allansmith, 1986). Since no vaccines currently use the mucosal immune system to specifically target ocular diseases, this chapter will review ocular mucosal immunobiology, examine approaches for inducing ocular mucosal immune responses, and identify target organisms for use in potential vaccination against selected ocular infections. Additional information on ocular mucosal immunity, including the impact of ocular infection on the structure and function of this system, is contained in an excellent review (Sullivan, 1994). It is important to note that other aspects of ocular immunobiology do not involve the mucosal immune system. Descriptions of anterior chamber-associated immune deviation and retinal immunology are provided in a number of other reviews (Gery et al., 1986; Streilein, 1987, Streilein, 1990, Streilein, 1993; Rocha et al., 1992; Niederkorn and Ferguson, 1996).

II. Ocular Mucosal Immunobiology

A. Tissues and Cells

The conjunctiva and lacrimal gland are the primary ocular tissues that are considered part of the mucosal immune system. The conjunctiva is a mucous membrane consisting of an outer squamous epithelial cell layer (two to five cells thick) interspersed with goblet cells, a basement membrane, and the underlying connective tissue of the substantia propria (stroma). Figure 1 diagrams the major features of the human conjunctiva. Bulbar conjunctiva overlies the sclera between the corneal limbus and the conjunctival fornix. Multiple sebaceous glands within the deeper stroma contribute to the corneal tear film and the lubrication of the conjunctiva (Fine and Yanoff, 1979). Mast cells have been identified in the conjunctiva of humans, other primates, and several rodents (Allansmith et al., 1977; Fine and Yanoff, 1979). These are particularly prominent in trauma-induced and allergic conjunctivitis. There are no microfold (M) cells in the conjunctiva and the normal conjunctiva is relatively free of lymphocytes. However, similar to gastrointestinal-associated lymphoid tissue (GALT), intraepithelial lymphocytes (IEL) have been identified in human and mouse conjunctiva (Dua et al., 1994; J. A. Whittum-Hudson, unpublished observations). Human conjunctival IEL are T cells and express the mucosal lymphocyte antigen HML-1 and CD8 (Dua et al., 1994). Another study found that the majority of conjunctival T cells expressed αβ TCR (Soukiasian et al., 1992). The murine equivalent IEL is CD3⁺ , CD8⁺ and TCR γδ⁺ (J. A. Whittum-Hudson, unpublished observations). Few T cells and essentially no B cells are found in the substantia propria of normal human, monkey, or mouse conjunctiva (Reacher et al., 1991; Soukiasian et al., 1992; Dua et al., 1994; Whittum-Hudson et al., 1995a). The majority of the substantia propria T lymphocytes are CD8⁺ cells in man and monkey and lie subepithelially. In the lower fornix of human conjunctiva, clusters of unorganized (nonfollicular) lymphocytes may be found (Fine and Yanoff, 1979). Some species such as rabbit and guinea pigs may exhibit true follicles in the lower fornices (Shimada and Silverstein, 1975; Chandler and Gillette, 1983; Sullivan, 1994). These may develop due to persistent exposure to exogenous antigens, or reflect a nonspecific response to the environment.

Figure 1.

Diagrammatic representation of human conjunctiva showing the conjunctival stratified columnar epithelium (E), intraepithelial lymphocytes (IEL), mucous-producing goblet cells (GC), basement membrane (BM), an accessory lacrimal gland of Krause (ALG), as well as lymphatic channels (LC), collagen fibrils (CF), lymphocytes (L), and dendritic cells (DC) within the substantia propria (SP).

(This figure has been adapted from Sacks et al., 1986.)

Lacrimal glands also play a major role in ocular mucosal defense. In humans, lacrimal glands are found in the anterior, superolateral area of the orbit. Smaller accessory lacrimal glands (see Fig. 1) also may be found in the upper and lower conjunctiva. Ducts from the major lacrimal gland drain into the superotemporal conjunctival cul-de-sac, while ducts from accessory glands drain directly through the epithelium of the conjunctiva adjacent to their location. Lacrimal glands of other species are found in different areas both within and outside the orbit. Histological examination shows that the glands are made up of acinar units which are interconnected by ductules and that these interconnected acinar units form the lobes of the gland. Each acinar unit consists of secretory acinar epithelial cells, which are surrounded by a basement membrane, and the glandular ducts are lined with pseudostratified epithelium. Lymphatic channels also are present in the gland, and these drain into the cervical and preauricular lymph nodes. In addition to epithelial cells, a variety of other cell types such as plasma cells, B and T lymphocytes, macrophages, and dendritic cells have been identified in the interstitial connective tissue between the acinar units and ducts. In human, rabbit, rat, and mouse lacrimal glands, IgA plasma cells predominate, although IgG and IgM plasma cells are present and, in human, IgD and IgE plasma cells have been found (Franklin, et al., 1973, Franklin, et al., 1979; Shimada and Silverstein, 1975; Allansmith et al., 1976; Brandtzaeg, et al., 1979, Brandtzaeg, et al., 1987; Gillette et al., 1980; Crago et al., 1984; Damato et al., 1984; McGee and Franklin, 1984; Gudmundsson et al., 1984; Brandtzaeg, 1985; Kett et al., 1986; Wieczorek et al., 1988; Ebersole et al., 1988; Pappo et al., 1988b; Hann et al., 1988; Montgomery et al., 1989). B cells bearing various surface immunoglobulin isotypes also have been found in smaller numbers in humans and rodents (McGee and Franklin, 1984; Pappo et al., 1988b; Wieczorek et al., 1988; Montgomery, et al., 1989, Montgomery, et al., 1990; Pepose et al., 1990). Both CD4⁺ and CD8⁺ T cells also have been identified in human and rat (Ebersole et al., 1988; Gudmundsson et al., 1988; Wieczorek et al., 1988; Pappo et al., 1988b; Pepose et al., 1990; Montgomery, et al., 1989, Montgomery, et al., 1990). Of the CD3⁺ T cells isolated from rat lacrimal gland, 89% bear the aß and 7% express γδ TCR (C. A. Skandera and P. C. Montgomery, unpublished observations). Macrophages have been demonstrated in both human and rat (Wieczorek et al., 1988; Pappo et al., 1988a; Montgomery et al., 1989; Pepose et al., 1990) and Langerhans-type dendritic cells identified in human lacrimal gland (Wieczorek et al., 1988; Pepose et al., 1990). These various cell types contribute a number of biologically active molecules which are important for defense at the ocular surface. The acinar epithelial cells produce lysozyme and lactoferrin which are naturally occurring tear components with bactericidal properties (Franklin et al., 1973; Gillette and Allansmith, 1980; Gillette et al., 1981; McGill et al., 1984). In addition, the plasma cells are the main source of tear S-IgA antibodies (Sullivan and Allansmith, 1984; Peppard and Montgomery, 1987; Franklin, 1989; Sullivan, 1994) which are produced in response to antigenic challenge and are major effector molecules in mucosal defense (Mestecky and McGhee, 1987; Childers et al., 1989; McGhee and Mestecky, 1990).

B. Relationship to the Mucosal Network

Although it has been postulated that the conjunctiva functions both as a mucosal inductive and effector site (Chandler and Gillette, 1983; Franklin and Remus, 1984; Franklin, 1989), the precise relationship of the conjunctiva to the mucosal immune network requires further clarification (Sullivan, 1994). With respect to inductive capabilities, the presence of lymphoid follicles containing T and B cells in rabbit (Shimada and Silverstein, 1975; Franklin et al., 1979; Franklin and Remus, 1984) and, in humans, the presence of Langerhans cells, dendritic cells, and macrophages (Sacks et al., 1986), as well as lymphatic channels (Srinivasan et al., 1990), support the potential to process and perhaps respond to antigen. Normal conjunctival epithelial cells do not express MHC class II; however, the presence of class II⁺ Langerhans cells in the epithelium and dendritic cells in the stroma suggests that antigen presentation may occur in conjunctiva. Further support for the inductive function comes from studies showing that topical application of antigen to the conjunctiva is an effective means of eliciting tear IgA antibody responses (Montgomery, et al., 1984a, Montgomery, et al., 1984b; Peppard et al., 1988). Recent investigations have suggested that nasal-associated lymphoid tissue (NALT) may also function as an inductive site following ocular topical antigen application (Carr et al., submitted; see Section III. A). With the exception of rabbit conjunctiva, which contains substantial numbers of IgA plasma cells and epithelial cells that synthesize secretory component (Franklin, et al., 1973, Franklin, et al., 1979; Liu et al., 1981), the conjunctiva of most species does not display the typical features of a mucosal effector site. However, it should be noted that studies in at least two species, rat and monkey, have documented lymphocyte traffic from gastrointestinal-associated lymphoid tissue (GALT) to conjunctiva (Zhang et al., 1983; Taylor et al., 1985). These data indicate that the functional role of the conjunctiva in the mucosal network may vary significantly between species.

An abundance of evidence now indicates that the lacrimal gland is a functional part of the mucosal immune network. As has been demonstrated for other mucosal effector tissues and described above (see Section II.A), the lacrimal glands of many species contain a predominance of IgA⁺ plasma cells. In addition, lacrimal acinar epithelial cells produce secretory component (Franklin et al., 1973; Sullivan, et al., 1984b, Sullivan, et al., 1990; Gudmundsson et al., 1985; Hann et al., 1991), which functions as part of the polymeric immunoglobulin receptor to mediate transport of IgA into external secretions (Solari and Kraehenbuhl, 1985). Figure 2 diagrams the biosynthetic events within lacrimal gland that lead to the production and transport of IgA into tears. Linkage of the lacrimal gland to the mucosal network occurs via trafficking lymphocyte populations. While both B and T lymphocytes traffic to mucosal tissues including lacrimal gland, B cells have been studied more extensively than T cells. IgA bearing lymphocytes from GALT (e.g., mesenteric lymph nodes) and glandular mucosal tissues have been shown to seed lacrimal glands (Montgomery, et al., 1983, Montgomery, et al., 1985) in a manner similar to other effector sites in the mucosal immune system (Roux et al., 1977; McDermott and Bienenstock, 1979; Weisz-Carrington et al., 1979). Tear IgA antibodies with specificity for oral microbes (Burns et al., 1982; Gregory and Allansmith, 1986, Gregory and Allansmith, 1987) as well as demonstrations that gastrointestinal immunization induced tear IgA antibodies (Mestecky et al., 1978; Montgomery et al., 1983; Gregory and Filler, 1987; Waldman and Bergmann, 1987) have provided further support for this linkage. Figure 3 shows the interrelationship of various compartments of the mucosal network and summarizes events leading to mucosal IgA induction following GALT immunization.

Figure 2.

Diagrammatic representation of the lacrimal gland showing S-IgA biosynthesis. Polymeric IgA (plgA) is synthesized in plasma cells located adjacent to the acinar epithelial cells. pIgA interacts with the polymeric immunoglobulin (poly-Ig) receptor on the basolateral surface of the acinar epithelial cells, is internalized and transported to the apical surface, and is released into the glandular lumen. Prior to release of the S-IgA molecule, the poly-Ig receptor is cleaved, leaving behind the cytoplasmic domain. The remainder of the poly-Ig receptor remains associated with the S-IgA molecule and is designated secretory component (s.c).

(This figure has been reproduced from Franklin and Montgomery (1996) and published courtesy of Mosby.)

Figure 3.

Schematic representation of the major features of the mucosal immune network as they relate to gastrointestinal-associated lymphoid tissue (GALT). Intestinal antigens (Ag) are taken up by microfold (M) cells overlying Peyer’s patches (PP) and are delivered to lymphoid cells in the PP. T cells and IgA committed B cells (●) migrate to the mesenteric lymph nodes (MLN), enter the circulation, and traffic to the lacrimal (LG), salivary (SG), mammary (MG) glands as well as to the lamina propria of the small intestine, the bronchial (BT) and urogenital (UGT) tracts, and liver. B lymphocytes lodging in mucosal tissues differentiate into IgA secreting plasma cells (○), producing pIgA which is transported into external secretions as S-IgA antibody (Ab). In some species the liver directly transports significant quantities of pIgA from the circulation into bile.

(This figure has been reproduced from Franklin and Montgomery (1996) and published courtesy of Mosby.)

The mechanism accounting for the preferential localization or retention of specific lymphoid populations in ocular mucosal tissues is an issue of central importance in understanding the regulation of ocular mucosal immunity. Lymphocyte-high endothelial venule (HEV) interactions control lymphocyte trafficking into organized lymphoid tissue and lymph nodes (Butcher, 1986; Woodruff et al., 1987, Stoolman, 1989; Chin et al., 1991b; Picker and Butcher, 1992; Shimizu et al., 1992; Bevilacqua, 1993) as well as into organized mucosal lymphoid tissues (Jalkanen et al., 1989; Chin et al., 1991a). In general, these interactions involve lymphocyte homing receptors which interact with vascular addressins on HEVs and other specialized endothelial cells. Although little is known regarding the mechanisms involved in cell trafficking to conjunctival tissue, the parameters regulating lymphocyte localization within the lacrimal gland are beginning to be defined. With respect to glandular structures, the control of lymphocyte traffic could occur on two levels: (1) at exit from the vasculature, or (2) within the stroma of the glandular tissue. There are no HEVs in lacrimal gland and no selective interactions with glandular endothelium have yet been demonstrated. Therefore, it appears that lymphocyte entry into lacrimal gland is random (McGee and Franklin, 1984). Current evidence suggests that lymphocyte populations are selectively retained based on interactions that occur within the lacrimal tissue microenvironment. Early studies indicated that B-ceII retention, particularly those committed to IgA production, resulted from a direct interaction with T cells located in lacrimal tissue (Franklin and Shepard, 1990). Recently, an in vitro binding assay has been used to show that circulating lymphocyte populations preferentially adhere to lacrimal gland acinar epithelium (O’Sullivan and Montgomery, 1990). Both B and T lymphocytes participate in this interaction with B cells binding in greater numbers (O’Sullivan et al., 1994a). Studies with biochemical inhibitors and antibodies directed against various adhesion molecules initially suggested the involvement of L-selectin (O’Sullivan, et al., 1994a, O’Sullivan, et al., 1994b). Other data indicated that pretreatment of lymphocytes with certain cytokines (IL-4 and TGFβ) reduced the capacity of those cells to interact with lacrimal gland but not lymph node or Peyer’s patches (Elfaki et al., 1994). These latter data provided the initial support for the involvement of a receptor distinct from those mediating interactions with lymph node and Peyer’s patch HEVs. The characterization of the lymphocyte receptor is still ongoing, but an 85-kDa factor has been identified which specifically inhibits lymphocyte interactions with lacrimal and salivary gland tissues (Montgomery and Liberati, 1995). Current studies are focused on defining the relationship of this molecule to other lymphocyte surface determinants and adhesion molecules, as well as on the identification of the acinar epithelial cell ligand or counter receptor. With respect to the ligand, recent studies using cultured rat lacrimal gland acinar cells suggest that this molecule is shed into the culture medium (O’Sullivan et al., 1995) and has led to the speculation that trafficking lymphocytes may also interact with secreted ligand accessing the interstitial glandular spaces. Although the ligand or counter receptor has not yet been isolated, it appears to be a 20- or 21-kDa glycoprotein (R. Raja and P. C. Montgomery, unpublished observations).

III. Induction of Ocular Mucosal Immune Responses

Regulatory events leading to the induction and expression of ocular mucosal immune responses are not completely defined. However, it is clear that many factors governing immune responses at other mucosal sites (Kiyono et al., 1992; Walker, 1994) have general applicability to mucosal immune responses at the ocular surface. The immunization route is of primary importance with local (e.g., topical application to the eye) and remote site (e.g., oral or gastrointestinal) routes generating tear IgA antibody responses. Combinations of routes and their sequence of administration can also affect antibody induction in tears. Not surprisingly for replicating microbes, tissue tropism and invasive properties influence the outcome of mucosal responses. When dealing with nonreplicating immunogens, particulate antigens are more effective at inducing ocular mucosal immune responses than soluble antigens. Thus far, traditional mucosal adjuvants such as cholera toxin, delivered topically to the eye with antigen, do not appear to enhance ocular antibody responses (Peppard and Montgomery, 1990). Helper T (Th) cells and T-cell products such as cytokines appear to play key roles in regulating ocular mucosal IgA responses, but the direct role of cytotoxic T (Tc) cells in ocular mucosal immunity is currently not clear.

As noted above (see Section II.B), lymphocyte traffic from mucosal inductive to effector sites is a central feature of the ocular mucosal system. Antigen-presenting cells (e.g., macrophages and dendritic cells) are also required, but their participation in ocular mucosal responses has yet to be delineated. The neuroendocrine system also plays a role in regulating ocular mucosal responses affecting both secretory component synthesis in vivo (Sullivan et al., 1984a) and cultured lacrimal gland acinar cells (Kelleher et al., 1991) as well as tear IgA production (Sullivan and Allansmith, 1985; Sullivan, 1988). Neuroendocrine regulation is reviewed in detail elsewhere (Stanisz et al., 1989; Sullivan, 1990, Sullivan, 1994). Although these as well as other factors yet to be identified impact ocular mucosal immune responses, our focus will be on selected parameters relevant to the development of ocular vaccination strategies. Therefore, the effects of immunization route, immune potentiators and delivery vehicles on the induction of ocular mucosal responses, in particular IgA antibody production, will be examined in detail.

A. Immunization Routes

In humans, most investigations have focused on use of the oral or gastrointestinal (GI) immunization route. Ingestion of heat-killed Streptococcus mutans elicited IgA antibodies in tears and saliva, but not serum antibodies (Mestecky et al., 1978). In addition, a second oral immunization markedly increased tear and salivary IgA responses. Other studies using inactivated S. mutans have generally confirmed these initial observations (Gregory et al., 1984; Gregory and Filler, 1987; Czerkinsky et al., 1987). Human tear IgA antibody responses have also been noted following oral immunization with influenza vaccine (Bergmann, et al., 1986, Bergmann, et al., 1987; Waldman and Bergmann, 1987) as well as following intranasal immunization with live rhinovirus (Douglas et al., 1967; Knopf et al., 1970).

The rat model has been employed to directly compare the effectiveness of antigen application to the conjunctiva (ocular topical, ot) and administration by the oral or gastrointestinal (GI) route using the nonreplicating particulate antigen, dinitrophenylated type III pneumococcus (DNP-Pn). These studies have shown: (1) both the immunization route and the sequence of stimulation play a role in tear antibody expression (Montgomery et al., 1984b); (2) long-term repeated ot stimulation leads to an eventual downregulation of tear IgA antibody levels (Montgomery, et al., 1984a, Montgomery, et al., 1984b); (3) the ot route generally yields higher tear IgA antibody levels than GI immunization (Peppard et al., 1988); and (4) coadministration of antigen and cytokines by the ot route enhances tear IgA antibody levels (Pockley and Montgomery, 1991; see Section III.B). The rat model also has been used to study the effect of immunization route on the kinetics of serum and tear antibody responses to live Chlamydia trachomatis (Davidson et al., 1993). This study compared the effectiveness of ot, GI, subconjunctival, and intraperitoneal immunization and showed that ot application was the most effective route for eliciting IgA antibody response in tears. Interestingly, subconjunctival injection of chlamydia yielded a vigorous serum IgG response and minimal tear IgA antibodies, which is consistent with responses to other antigens delivered by this route (Cousins et al., 1991). The explanation for this latter finding is not clear, but subconjunctival delivery may bypass an essential ocular mucosal processing step during antigen uptake or allow live organisms to gain access to the circulation.

Although the precise mechanism by which ot immunization elicits tear IgA antibody responses is not known, it appears that the inductive process could involve at least two pathways: (1) antigen uptake, processing, and triggering of B cells in the conjunctival substantia propria or (2) antigen drainage via the nasolacrimal canal, uptake by NALT M cells, processing and triggering of resident B cells in NALT (Carr et al., submitted). The schematic details of these two pathways are shown in Fig. 4 . Not shown in this figure is the option for antigen drainage via the nasolacrimal canal to access GALT or other alternatives which may be applicable to primates. In humans and monkeys, antigen may reach the accessory lacrimal glands directly or gain access to potential ductal associated lymphoid tissue as has been shown for minor salivary glands (Nair and Schroeder, 1986). However, since it is not clear how antigen would be processed within the acinar units of the accessory lacrimal glands and ductal-associated lymphoid tissue has not been identified for these glandular structures, these latter alternatives appear unlikely.

Figure 4.

Schematic representation of the major components of the ocular mucosal immune system showing possible pathways for antigenic stimulation at the ocular surface. 1. Antigen may gain access to the conjunctival (CONJ) substantia propria where dendritic cells would present it to resident lymphocytes (see Fig. 1). Once triggered, B cells (●) would traffic to the superior cervical lymph node (sCLN) and gain access to lymph. 2. Ag may drain via the nasolacrimal canal (NLC), be taken up by M cells, and gain access to the nasal-associated lymphoid tissue (NALT) where presentation and B-cell triggering would occur. B cells (●) would traffic to the posterior cervical lymph node (pCLN) and gain access to lymph. B cells reaching lymph from either pathway would enter the circulation, traffic to the major lacrimal gland (LG), and differentiate into IgA antibody-secreting plasma cells (○). The IgA would be processed as shown in Fig. 2 and reach the ocular surface via the lacrimal duct (LD).

Several animal models have also been used to evaluate tear antibody responses after immunization or infection at ocular surfaces with chlamydia and other pathogens. In guinea pigs, only ocular immunization with live Chlamydia psittaci (GPIC) induced tear IgA (and in some cases IgG) antibodies (Murray et al., 1973; Watson et al., 1977; Rank and Whittum-Hudson, 1994). Killed organisms administered via ocular topical (Murray et al., 1973) or intraperitoneal routes (Malaty et al., 1981) failed to induce tear antibodies. However, similar studies with human biovars of C. trachomatis in owl monkeys suggested that inactivated chlamydia could indeed induce tear antibodies (MacDonald et al., 1984). In the cynomolgus monkey model, ocular infection with the chlamydial organism induced vigorous IgA, IgG, and IgM responses in both serum and tears (reviewed by Taylor, 1990). Only oral immunization with live L₂ biovar of chlamydia, but not ocular serotypes, successfully induced mucosal and systemic immunity, which partially protected the host from subsequent infectious challenge (Taylor et al., 1987b).

Cynomolgus monkeys develop conjunctival disease similar to that observed in humans (Whittum-Hudson and Taylor, 1984). An intense mononuclear infiltrate appears in infected conjunctiva (Whittum-Hudson, et al., 1986a, Whittum-Hudson, et al., 1986b; Whittum-Hudson and Taylor, 1989) and is composed of high frequencies of chlamydia-specific T and B lymphocytes (Pal et al., 1992). Supporting the importance of regional draining cervical lymph nodes in conjunctival responses to pathogens, essentially identical frequencies of antigen-specific cells were detected in these lymph nodes and the conjunctivae after infection, and in “ocular-immune” animals which were topically challenged with the chlamydial hsp60 antigen (Pal, et al., 1990b, Pal, et al., 1992). Serologic responses to chlamydial hsp60 and hsp70 are vigorous during chronic chlamydial infections in man and animals (Newhall et al., 1982; Pal et al., 1990a), although neither of these proteins delivered by the ot route immunized naive monkeys (Taylor, et al., 1987b, Taylor, et al., 1990). In contrast, ot boosting after systemic or oral immunization with purified or recombinant chlamydial antigens induced vigorous tear IgA and IgG and serum antibodies (Campos et al., 1995; O’Brien et al., 1994). It remains unclear whether the initial dosage of some antigens is too small to induce typical primary secretory immune responses when presented to the mucosal immune system by ot delivery, or if an antigen must replicate and/or be invasive to adequately prime experimental animals. Nevertheless, ocular boosting with purified antigens or infectious ocular challenge of systemically primed animals has demonstrated that anamnestic tear antibody responses occur rapidly for chlamydia (Campos et al., 1995). Combined oral and subcutaneous priming with a chlamydial fusion protein of a major outer membrane protein (MOMP) variable domain incorporated into liposomes also primed for mucosal immune responses (O’Brien et al., 1994).

The rabbit model has also been used to test ocular immune responses to other organisms. Live herpes simplex virus type 1 administered by the ocular route elicited tear IgA, but no IgG or IgM antibodies (Willey et al., 1985). In contrast, application to scarified cornea induced tear IgA, IgG, and IgM responses (Centifanto et al., 1989). Whether these differing results relate to virus strain invasiveness or access to underlying dendritic cells is unclear. Live Staphylococcus aureus administered ot elicited tear IgA and IgG antibodies as did a S. aureus subunit (peptidoglycan-ribitol teichoic acid) vaccine administered by the intradermal or subconjunctival routes with complete Freund’s adjuvant (Mondino, et al., 1987a, Mondino, et al., 1987b, Mondino, et al., 1991). However, given the severe inflammatory properties of Freund’s adjuvant and the restrictions on its use, this latter approach is impractical as a vaccination strategy.

B. Immune Potentiators

Immune potentiators, or adjuvants, have been used to enhance antibody responses in both humans and animal models. Although the mechanisms responsible for response enhancement vary and are often not well defined, adjuvants appear to function by one or a combination of mechanisms: (1) increasing the influx of inflammatory cells; (2) promotion of antigen presentation; and/or (3) enhancing antigen uptake. Consideration of immune potentiators to enhance ocular mucosal responses requires selection of compounds which will not promote ocular inflammation and pathology.

The most widely studied mucosal adjuvant has been cholera toxin (Lycke and Holmgren, 1968) although its effects on ocular immunity are less defined. Enhancement of mucosal responses depends on the binding of the cholera toxin B subunit to the G_M1 ganglioside displayed on the luminal surface of enterocytes such as M cells as well as potential modulatory effects of the toxin on the B, T, and antigen-presenting cells (Dertzbaugh and Elson, 1990; Elson and Dertzbaugh, 1994). Cholera toxin enhances antibody responses when coadministered with many antigens, whereas the B subunit apparently requires conjugation to antigen to function optimally. Oral immunization with cholera toxin induced the appearance of toxin-specific antibody-forming cells in rat conjunctiva (Zhang et al., 1983) and toxin-specific antibodies in tears of monkeys (Campos et al., 1995). However, ot administration of cholera toxin with the DNP-Pn antigen did not enhance tear IgA antibody responses in the rat model (Peppard and Montgomery, 1990), but the effects of oral or ocular administration of antigens conjugated to the B subunit on tear IgA responses have not been fully investigated. Based on recent data suggesting that nasal associated lymphoid tissue can serve as an inductive site for tear IgA antibody responses (Carr et al., submitted) and the ability of the B subunit to serve as an adjuvant when coupled to antigens administered by the intranasal route (Bessen and Fischetti, 1990; Wu and Russell, 1993), this immunization protocol requires rigorous testing. Two other mucosal adjuvants, avridine and muramyl tripeptide, administered topically by the ocular route with DNP-Pn, did not enhance tear IgA antibody responses (Peppard et al., 1988).

Cytokines also appear to exert important regulatory effects on mucosal IgA responses including those at ocular sites. Interleukin (IL)-6 knock-out mice have greatly reduced numbers of IgA producing cells at mucosae and exhibit grossly deficient local antibody responses to mucosal challenge (Ramsay et al., 1994). In addition, transforming growth factor-β (TGFß), IL-5, and IL-6 have been shown to enhance in vitro IgA responses in murine B-cell cultures. TGFβ appears to increase IgA synthesis by inducing B cells to switch to IgA production (Coffman et al., 1989; Ehrhardt et al., 1992), whereas IL-5 and IL-6 induce IgA⁺ B cells to terminally differentiate into IgA producing plasma cells (Coffman et al., 1987; Beagley, et al., 1988, Beagley, et al., 1989; Harriman et al., 1988). IL-5, IL-6, and TGFβ have been shown to specifically enhance IgA production in vitro in a rat lacrimal gland tissue fragment culture system (Pockley and Montgomery, 1990; Rafferty and Montgomery, 1993). Further, IL-5 and IL-6 in combination with antigen have been shown to enhance specific tear IgA antibody responses in rats following ot administration without affecting serum IgG antibody levels (Pockley and Montgomery, 1991). Recent studies indicate that these enhanced tear IgA antibody responses persist following restimulation with antigen in the absence of IL-5 and IL-6 (Montgomery et al., 1994). Although the mechanism(s) underlying this in vivo cytokine-mediated enhancement is not yet known, it is clear that the regulatory roles and therapeutic potential of these and other cytokines to modulate ocular mucosal antibody responses require further investigation.

C. Delivery Vehicles

Although the oral or GI immunization route triggers IgA antibody responses at both intestinal and distal mucosal sites and represents a safe, convenient alternative for vaccine administration, it is often difficult to present intact antigen to GALT-inductive sites. Particulate antigens, including viable organisms, generally appear to be the most effective immunogens, probably due to their ability to survive the pH extremes and enzymatic degradation that occur along the alimentary tract. Since many potential vaccine candidates are microbial components or subunits and are not viable organisms or particulate immunogens, considerable effort has focused on the development of delivery systems. Such delivery systems include liposomes and biodegradable microparticles, as well as recombinant bacterial and viral vectors. Strategies for vehicle development have included not only delivery of intact immunogens and genes that code for relevant antigens, but also systems that promote uptake by relevant cells (e.g., M cells) at GALT inductive sites. Detailed information on antigen delivery systems can be found in several recent reviews (Eldridge et al., 1993; Michalek, et al., 1994, Michalek, et al., 1995). Although many of these delivery systems may be relevant to the induction of mucosal antibody responses owing to their ability to induce vigorous responses at GALT effector sites, information is more limited on the potential to employ these vehicles to induce ocular mucosal immune responses. Based on recent observations indicating that microparti-cle-encapsulated antigens evoke ocular mucosal responses, discussion will be confined to this delivery vehicle.

Poly(lactide-co-glycolide) (PLG) microparticles are biodegradable under physiological conditions and represent an important candidate vehicle for delivery and controlled release of vaccine components (O’Hagan et al., 1991; O’Hagan, 1994). PLG has been used for suture material (Wise et al., 1979) and as a delivery system for drugs (Hutchinson and Furr, 1985). PLG does not evoke a tissue response (Visscher et al., 1987) and has been certified by the FDA for human use. Recently, PLG microparticles have been used for ocular drug delivery without provoking histological or electrophysiological changes to the eye (Montera et al., 1992). In addition, PLG encapsulated antigens induce sustained systemic and secretory immune responses following oral administration (Challacombe et al., 1992; Maloy et al., 1994). Based on these data as well as observations that the cytokines IL-5 and IL-6 enhanced tear IgA antibody responses following topical application to the eye (Pockley and Montgomery, 1991; see Section III.B), PLG microparticles appear to be an appropriate vehicle to deliver antigens and immune potentiators to ocular mucosal inductive sites. Studies indicate that DNP-bovine serum albumin as well as IL-5 and IL-6 can be encapsulated in PLG microparticles without significant loss of activity. In addition, ot delivery of encapsulated antigen and cytokines evokes long-term mucosal antibody responses in tears and distal secretions as well as circulating antibodies in the rat model (Rafferty et al., 1996). Further, oral delivery of a poly(Iactide) encapsulated chlamydia candidate vaccine in mice induced antigen specific serologic responses and conferred partial protection against ocular challenge (Whittum-Hudson et al., 1995b; see Section IV.A). While further studies are required to define the regulatory parameters and events leading to these responses, microparticles clearly are promising candidate delivery vehicles to elicit ocular mucosal immune responses.

IV. Targets for Vaccine Development

The conjunctiva is the most commonly infected tissue of the eye (Syed and Hyndiuk, 1992; Tabbara and Hyndiuk, 1995). Infection may be limited to the conjunctiva only or may be part of a lid or corneal infection. Most conjunctival infections are not vision-threatening, but most notable exceptions are Neisseria gonorrheae and herpes simplex virus. Conjunctivitis associated with tularemia (adults) (Knorr and Weber, 1994) and haemophilus infections (infants) (Syed and Hyndiuk, 1992) are potentially fatal because of systemic dissemination. The profile of pathogens isolated from normal and diseased human conjunctiva may vary depending upon geographic locale (Syed and Hyndiuk, 1992). Similarly, some veterinary pathogens of the conjunctiva/lacrimal gland are geographically confined (Lepper et al., 1993; Rogers et al., 1993). Chronic inflammation of the conjunctiva may lead to scar formation and is termed chronic progressive conjunctival cicatrisation (Bernauer et al., 1993). Cicatricial diseases often progress to involve the lacrimal and meibomian glands. The combination of altered tear film, scar contraction, and corneal trauma from trichiasis may lead to blindness. Although there are noninfectious etiologies for many cicatrical disorders including pemphigoid, chlamydial infections of the conjunctiva are a classic example of the blinding sequelae of infection.

As discussed above and elsewhere (Sullivan, 1994), there are physical and biochemical barriers to help prevent ocular infection. For extracellular ocular bacterial infections, antibodies are a major component of protective immunization. Both antibody and cell-mediated immunity are believed to be important in protection against intracellular pathogens including viruses, chlamydia, and mycobacteria. Antibodies would block infection and clear free organisms, whereas CD8⁺ cytotoxic T cells would clear infected cells and prevent full differentiation and spread from lysed host cells. There is little in vivo evidence from ocular sites that cytotoxic T lymphocytes are important in clearance of infectious pathogens. However, CD4 T cells serve as antigen presenting cells for B-cell responses and both CD4 and CD8 T cells produce cytokines that facilitate B- and T-cell responses. Many bacteria, viruses, and parasites are targets for vaccine development to prevent respiratory, genital tract, or other infections. If protective mucosal immunization is induced at one or more of these latter sites, it is probable that the ocular mucosal surfaces would also be protected.

A. Bacterial Infections

A variety of bacteria can infect ocular tissues. Table I lists some of the bacterial pathogens that are known to cause conjunctivitis. Bacterial conjunctivitis is most commonly caused by S. aureus. Streptococcus sp. are common conjunctival pathogens, except for β-hemolytic streptococci (S. pyogenes), while other members of this group are part of the normal conjunctival flora. These bacteria and Haemophilus influenzae are the most common agents causing dacryocystitis (lacrimal gland infection). Corynebacterium diphtheriae is a rare ocular pathogen, but can cause conjunctival scarring and associated problems. For the most part, bacterial infections of the conjunctiva are self-limiting and nonvision threatening. Neisseria gonorrhoeae is an exception, and this organism is the leading cause of hyperacute conjunctivitis particularly in developing countries (Syed and Hyndiuk, 1992). N. meningitis conjunctivitis, which cannot be distinguished clinically from N. gonorrhoeae, also may be associated with systemic disease. H. influenzae conjunctivitis occurs in children as well as adults, and one could predict a decline in the incidence of this ocular disease now that the Hib vaccine is available. Anaerobic organisms compose another part of the normal flora of the conjunctival sac. As for other infectious pathogens, the incidence of conjunctival infections with abnormal anaerobic organisms is increased in patients with AIDS (Dugel and Rao, 1993; Campos et al., 1994). Similarly, some cases of microsporidial conjunctivitis have been reported in AIDS patients (Lacey et al., 1992; Lowder, 1993; Weber et al., 1993).

Table I.

Major Bacterial Pathogens Causing Conjunctivitis

Organism	Referencesa
Staphylococcus aureus	Holzberg et al., 1992
Neisseria gonorrheae	Syed and Hyndiuk, 1992
Streptococcus sp.	Murphy et al., 1991
Haemophilus influenzae, H. meningitidis	Syed and Hyndiuk, 1992
Borellia burgdorferi	Aaberg, 1989; Zaidman, 1993
Francisella tularensis	Knorr and Weber, 1994
Chlamydia trachomatis	Wills et al., 1987; Taylor et al., 1992; Heggie and Lass, 1994; Brunham and Peeling, 1994;
C. psittaci	Dhingra and Mahajan, 1991
Pseudomonas aeruginosab	King et al., 1988; Brook and Hulburd, 1993
Shigella flexneri, S. sonnetc	Linde et al., 1993
Brucella melitensisd	Young, 1983; Zundel et al., 1992

^aselected primary articles and/or reviews of bacterial ocular pathogens.

^bMay cause conjunctivitis, but usually is associated with corneal infections following abrasion or tissue injury (Hazlett et al., 1976; Ramphal et al., 1981).

^cGuinea pig ocular challenge model is used to test for protective immunity induced by anti-Shigella vaccine, because corneal/ conjunctival epithelium is similar to intestinal epithelium in susceptibility to invasion by Shigella.

^d‘Infects conjunctival mucosa, but is more of a problem for abortion in ewes and goats; however, vaccination via the conjunctiva reduced spontaneous abortions compared to systemic immunization.

Chlamydia trachomatis, an obligate intracellular bacterium, has received a great deal of attention with respect to vaccine development. While chlamydia has gained notoriety in recent years as a major sexually transmitted pathogen, it is, in fact, the leading cause of preventable blindness. Trachoma, resulting from chronic chlamydial infection of the conjunctiva, has been known since ancient times. Studies of patients in several endemic areas worldwide yielded valuable information regarding the epidemiologic, microbiologic, and serologic characteristics of trachoma (Nichols et al., 1973; Grayston et al., 1985; Dawson et al., 1989; Mabey et al., 1992; Taylor et al., 1992). Two major animals models have been used for immunologic studies of chlamydial ocular infections: a guinea pig model which can be infected with C. psittaci causing guinea pig inclusion conjunctivitis (GPIC), and subhuman primates, including owl monkeys and cynomolgus monkeys, which can be infected with several biovars of C. trachomatis, the usual human ocular pathogen (Patton, 1990; Rank and Whittum-Hudson, 1994). More recently, a mouse model of ocular infection was developed using a human ocular biovar (Whittum-Hudson et al., 1995a); this model should aid in immunologic studies which have not been possible with the other outbred species.

Vaccine development for this bacterium has been hampered by several key factors: (1) no protective antigen or epitope(s) has been conclusively identified from the plethora of immunogenic proteins; (2) no chlamydial target antigens have been shown convincingly to be expressed on the surface of infected host cells, and (3) chlamydia-induced disease sequelae are immunemediated, though only the 57-kDa chlamydial heat-shock protein has been associated with immunopatho-genic responses (Watkins et al., 1986; Taylor et al., 1987a; Morrison, et al., 1989a, Morrison, et al., 1989b).

Various permutations of the chlamydial major outer membrane protein (MOMP) have been tested for immunogenicity and their ability to induce protective immunity in experimental animals. Biochemically extracted native MOMP, recombinant MOMP, or its variable domains have been presented by systemic and mucosal routes. In some cases, protection was tested by infectious challenge at the ocular mucosal surface or other sites (genital tract, lung, etc.). Table II lists some of the chlamydial vaccine candidates tested experimentally over the past two decades. In most cases, whole MOMP or subunits were highly immunogenic as measured primarily by serologic responses, but no significant in vivo protection against ocular infection was observed after direct immunization. Studies by several groups (Ward et al., 1986; Stephens, 1990; Mabey et al., 1991; Stagg et al., 1993; Murdin, et al., 1993, Murdin, et al., 1995; Tuffrey, 1994) have identified important MOMP peptide epitopes to which immune responses have been induced. None of these have been tested in ocular infection models as yet, but several have been at least partially protective in chlamydial genital infection models (reviewed by Tuffrey, 1994).

Table II.

Potential Chlamydial Vaccine Antigens Tested in Ocular or Extraocular Models

Antigena	Delivery routeb 1°/2°	Adjuvant/carrierc	References
Whole organism	i.n.		Pal et al., 1994
	Oral		Taylor et al., 1987b
Whole organism extracts	i.m. or s.c.	Oil emulsion	Grayston et al., 1962; Nichols, et al., 1966, Nichols, et al., 1969; Soldati et al., 1971; Schachter, 1985
MOMP (TX 100)	s.c.	CFA	Taylor et al., 1987a
	Oral		Taylor et al., 1988
MOMP (OGP)	s.c.	CFA	Batteiger et al., 1993; Campos et al., 1995
	Oral/ot	CT	Campos et al., 1995
rMOMP subunit	ot		Wong et al., 1990
	Oral/ot	CT-B conjugate	Taylor et al., 1989
MOMP/rMOMP subunits	s.c./Peyer’s patches	Alum	Tuffrey, 1994
rMOMP subunit fusion protein	s.c./s.c. (EB)	Titermax	Allen and Stephens, 1993
rMOMP subunit	Oral/oral ot	CT/liposomes	O’Brien et al., 1994; Whittum-Hudson et al., 1995a
fusion protein	s.c./s.c.
	i.d.	Alum/Salmonella typhimurium	Hayes et al., 1991
rMOMP hybrid	s.c./s.c.	Poliovirus hybrid	Murdin, et al., 1993, Murdin, et al., 1995
hsp 70	s.c. or ot		Taylor et al., 1990
hsp 60	ot		Taylor et al., 1990
	s.c./ot		Rank et al., 1995
Anti-idiotypic antibodies to chlamydial exoglycolipid	s.c./s.c. Oral/oral	Alumina (Maalox) Encapsulated (microspheres)	Whittum-Hudson and Taylor, 1994; Whittum-Hudson, et al., 1995a, Whittum-Hudson, et al., 1995b

^aMOMP, major outer membrane protein; TX100, Triton X-100; OGP, octyl glucosyl pyranoside; rMOMP, recombinant MOMP; hsp, heat shock protein.

^bs.c, subcutaneous; i.n., intranasal; ot, ocular topical; i.d., intradermal; i.m., intramuscular.

^cCFA, complete Freund’s adjuvant; CT, cholera toxin; CT-B, cholera toxin B subunit; IFA, incomplete Freund’s adjuvant; EB, chlamydial elementary bodies.

Other chlamydial antigens such as the chlamydial lipopolysaccharide (LPS) have been shown to be immunogenic but not protective in an ocular infection model (Taylor and Prendergast, 1987). An additional chlamydial antigen, the exoglycolipid antigen (GLXA) (Stuart and MacDonald, 1989), is also expressed by chlamydia and secreted from infected cell in vitro (Stuart, et al., 1991, Stuart, et al., 1994; Wyrick et al., 1994). Monoclonal anti-idiotypic (Id) antibodies to GLXA have been shown to immunize and protect mice against subsequent infectious ocular challenge with a human biovar of chlamydia (Whittum-Hudson et al., 1994). Oral immunization of BALB/c mice with poly(lactide) encapsulated anti-Id antibodies to GLXA was shown recently to be even more protective than the soluble antibody since 25-fold less antibody presented in microspheres induced equivalent or better reduction in microbiologic disease. In both cases, mice developed serum antibodies to the GLXA and these antibodies were neutralizing for in vitro infectivity (Whittum-Hudson et al., 1995b; Whittum-Hudson et al., submitted). This type of vaccine approach has potential value for chlamydial infections at other sites, and against other biovars since the GLXA antigen is genus-specific as opposed to being serovar-specific.

It has been hypothesized by several investigators that protective immunity against chlamydia will require induction of neutralizing antibody in tears and it is assumed that this will be S-IgA. IgA in tears from patients with trachoma passively neutralized chlamydia and reduced disease in eyes of owl monkeys (Nichols et al., 1973; Barenfanger and MacDonald, 1974). This was consistent with observations in guinea pigs ocularly infected with C. psittaci (Orenstein et al., 1973; Malaty et al., 1981). The dilemma with the chlamydial pathogen is that individuals who have had a previous ocular (or other site) infection with chlamydia are not significantly protected against subsequent ocular infections. This lack of protective immunity is seen despite evidence of vigorous antibody responses to many chlamydial antigens in tears and serum and in vitro lymphocyte proliferative responses to whole organism or purified antigens (Newhall et al., 1982; Ward et al., 1986; Whittum-Hudson and Taylor, 1989; Taylor, 1990). Several clinical studies have shown a poor correlation of acquisition of genital chlamydial infection with the presence of neutralizing serum antibodies (Jones and Van der Pol, 1994).

Cytotoxic T cells have not yet been associated directly with ocular protective immunity, nor with the immunopathogenesis of chlamydia ocular infection. It has been argued that the lack of exposure of chlamydial antigens on the surface of infected host cells would prevent them from serving as immunologic targets and may explain the failure to induce T-cell-mediated immunity. However, recent electron microscopic studies of human endometrial cells have shown that chlamydial MOMP, LPS, and the GLXA (Wyrick et al., 1994) all pass from the inclusion body through the cytoplasm to the cell membrane. LPS are GLXA and both known to exit infected cells (Brade et al., 1986; Stuart et al., 1991; Lukacova et al., 1994). In addition, recent studies demonstrated CD8 mediated cytotoxic killing by chlamydia-primed T cells in vitro and some in vivo protection after experimental genital infection (Beatty and Stephens, 1994; Starnbach et al., 1994). Thus, it is possible that chlamydial antigen presentation to CD8⁺ effector T cells may occur in the context of MHC class I.

B. Viral and Parasitic Infections

Table III lists many of the viral and parasitic pathogens known to infect ocular tissues. With respect to viruses, such infections frequently result in conjunctivitis in both children and adults and these highly contagious infections (epidemic keratoconjunctivitis) resolve spontaneously over 2–3 weeks. The most common viral pathogen of the conjunctiva is adenovirus. Several adenovirus types (3, 7, 8, 19) may be associated with conjunctivitis, though corneal involvement is often seen (Murrah, 1988).

Table III.

Major Viral and Parasitic Pathogens of Ocular Tissues

Organisms	Referencesa
Coronavirusb	Wickham et al., 1994
Herpesviruses	Syed and Hyndiuk, 1992; Tabarra and Hyndiuk, 1995
Cytomegalovirus	Huang et al., 1994
Epstein-Barr virus	Pepose, 1994; Jones et al., 1994
Herpes simplex	Holzberg et al., 1992; Pepose, 1994
Adenoviris (Types 3,7,8,19)	Holzberg et al., 1992; Syed and Hyndiuk, 1992
Papillomavirus 16	McDonnell et al., 1992
Picornaviruses (Coxsackie virus A24)	Syed and Hyndiuk, 1992
Simian immunodeficiency virus	Conway et al., 1991
Feline immunodeficiency virus	Callanan et al., 1992
Onchocerca vulva	Taylor, 1994
Rhinosporidium seeberi	Gori and Scasso, 1994

^aSelected primary articles and/or reviews of nonbacterial ocular pathogens.

^bInfects and replicates in cultured rat lacrimal gland acinar cells; but its capacity to infect in vivo is not known.

Ocular herpesvirus infections, though most often affecting the cornea, may cause conjunctivitis during primary infection. However, conjunctivitis is rare during recurrent infections with herpes simplex virus (Holzberg et al., 1992). Herpes zoster ophthalmicus may develop in patients with latent varicella-zoster virus infections, and the conjunctivia is one of several ocular sites most often affected (Liesegang, 1993). Since at present there is no vaccine for herpesviruses, antiviral drugs remain the therapeutic option. There is a large literature concerning anti-herpesvirus vaccine development strategies, and this information is detailed in several reviews (Rouse and Lopez, 1984; Dix, 1987; Mader and Stulting, 1992; Stanberry, 1994). In addition to the fact that herpesviruses are ubiquitous and that the majority of individuals have been exposed previously, vaccine development strategies for ocular and extraocular herpes virus infections must deal with a number of questions: (1) Can recurrence of latent infections be ablated by vaccination or must primary infection be prevented? (2) Will herpesvirus-induced immunosuppression be overcome by appropriate immunization? (3) Will a combination subunit vaccine be required? and (4) Is systemic rather than mucosal immunization required? Table IV lists some recent herpes simplex virus vaccine candidates. A number of viral glycoproteins have been targeted for vaccines (e.g., Burke et al., 1994), although there are some examples of better protection being conferred by vaccination with live organisms (Ghiasi et al., 1995). Studies which directly tested for protection against ocular herpes simplex virus infections have used various viral constructs ranging from single or multiple glycoprotein constructs and deletion mutants delivered with or without adjuvant. Avirulent constructs have been tested in a variety of experimental models and clinical trials. While most immunizations have been via systemic routes (subcutaneous or intraperitoneal), a recent study using the rabbit model of recurrent ocular herpes infection showed that subconjunctival immunization with recombinant gB and gD after establishment of latent infection reduced spontaneous recurrences (Nesburn et al., 1994). Promising results have beenshown in extraocular infection models by several groups. These have included strategies of mucosal vaccination with recombinant adenovirus expressing herpes simplex virus gB (Gallichan et al., 1993), subcutaneous immunization with virus constructs missing an essential glycoprotein (e.g., gH; Farrell et al., 1994), and systemic delivery of synthetic peptides. Other approaches have included use of anti-sense nucleic acids (Cantin et al., 1992) and use of other expression systems (e.g., baculovirus and vaccinia). Recent studies of cytokine responses associated with in vivo or in vitro anti-viral protection have implicated IL-6, IL-10, and TNFα in either recovery from or reduced herpes infection (Chen et al., 1994; Tumpey et al., 1994; Babu et al., 1995). Further investigations are required to distinguish between T-cell-mediated immunopathologic effects (Newell et al., 1989; Hendricks and Tumpey, 1990; Hendricks et al., 1992) and the potential requirement for individual T-cell subsets in vaccine-induced immunity.

Table IV.

Recent Herpes Simplex Virus Vaccine Candidates

Vaccine	Referencesa
Purified or synthetic glycoproteins
gB,.gC	Chan, 1983
gB,.gC	Roberts et al., 1985
gD-1, gD-2	Long et al., 1984
gB-1, gD-1,-2	Dix and Mills, 1985
gD	Ishizaka and Mishkin, 1991; Burke et al., 1994
Baculovirus-expressed glycoproteins
gB, gI, gC, gE, gG, and/or gH	Ghiasi et al., 1995
gB-2, gD-2	Nesburn et al., 1994
Adenovirus-expressed glycoproteins gB	Gallichan et al., 1993
Vaccina virus-expressed glycoproteins	Aurelian et al., 1991;
	Rooney et al., 1991;
	Banks et al., 1994; Fleck et al., 1994
Deletion mutants (gH)	Farrell et al., 1994
Anti-sense nucleic acids	Cantin et al., 1992

^aselected references; see also Rouse and Lopez, 1984; Dix, 1987; Mader and Stulting, 1992; Stanberry, 1994.

Several common parasitic infections may cause conjunctivitis. These include Mulluscum contagiosum, Onchocerca vulva, Verruca vulgaris, and Trypanosoma cruzi. In some cases, the conjunctival mucosa may be the portal of entry (e.g., T. cruzi). The pathologic results of trypanosomiasis include the oculoglandular or oph-thalmoganglionar complex composed of lid edema, granulomatous conjunctivitis, follicles, and inflammation of the lacrimal gland. Parinaud’s oculoglandular syndrome is rare, and may also be caused by additional viral or bacterial agents (e.g., F. tularensis, M. tuberculosis) (Syed and Hynduik, 1992).

Experimental models of conjunctival and lacrimal gland infections have been used for many years both to identify pathogenic and protective immune responses and, more recently, to evaluate vaccine candidates. Recent advances include new rat and rabbit models of adenovirus ocular infections, which should help in evaluation of immune-mediated responses to these pathogens (Tsai et al., 1992; Gordon et al., 1994).

V. Summary

It is clear that events taking place both in conjunctival tissue and lacrimal gland are relevant to ocular mucosal vaccination strategies. While the lacrimal gland functions as a primary effector site for mucosal IgA antibody responses, the precise relationship of the conjunctiva to the mucosal network requires further study. It appears that the conjunctiva plays a role in the induction of mucosal antibody responses as well as serving as an effector site for cell-mediated responses. A variety of factors are involved in the regulation of ocular mucosal antibody responses and many options such as immunization routes, delivery vehicles, and immune potentiators have been investigated. Although the development of appropriate immunization strategies will depend on the properties of each potential vaccine candidate as well as the target ocular pathogen, the appropriate technology is now available to elicit mucosal antibody responses at the ocular surface. The current challenge is to identify and select the vaccine component(s) that can be used to elicit long-term protective responses to ocular pathogens.

Acknowledgments

Studies from our laboratories which are summarized in this chapter were supported in part by NIH Grants EY05133, EY04068, and EY03324, and by Edna Mc-Connell Clark Foundation Grants 10993, 11092, and 13594. The authors thank Linda Hazlett, Nancy O’Sullivan, Robert Prendergast, and Robert Swanborg for their critical assessment of this chapter and Nancy Hynous for secretarial support.