The Corneal Stroma Is Endowed with a Significant Number of Resident Dendritic Cells

Abstract

purpose. Dendritic cells (DCs) comprise a system of highly efficient antigen-presenting cells (APCs) that initiate immune responses. The purpose of this study was to examine the normal stroma for the presence of DCs and other bone marrow (BM)–derived cells.

methods. Normal uninflamed corneas of BALB/c and other murine strains were excised, and immunofluorescence single- and double-staining for multiple markers was performed for examination by confocal microscopy. Corneal buttons were placed in culture and immunocytochemistry and flow cytometry performed.

results. MHC class II⁺CD80⁺CD86⁺ cells were found in the periphery of the anterior normal stroma. These cells were CD45⁺, CD11c⁺CD11b⁺ suggesting a BM-derived and monocytic DC lineage. In a surprising finding, significant numbers of MHC class II⁻CD80⁻CD86⁻ cells were found in the center of the anterior stroma. These cells were also CD45⁺CD11c⁺CD11b⁺ but CD3⁻, GR-1⁻, keratan sulfate⁻, and CD8α⁻, reflecting an immature precursor phenotype of myeloid DC. In addition to DC subsets in the anterior stroma, a CD11c⁻CD11b⁺ population of BM-derived cells was found primarily in the posterior stroma, representing monocytes/macrophages. These cells were rarely present in the anterior third of the normal stroma. Further, CD14⁺ precursor-type DCs were found throughout the stroma. These in vivo findings were not strain specific and were confirmed by immunocytochemistry and flow cytometry analyses of cells derived from corneal explants and by transmission electron microscopy.

conclusions. This study demonstrates that, in addition to the known Langerhans cells in the corneal epithelium, at least three BM-derived cell subsets reside in the normal corneal stroma.

Antigen-presenting cells (APCs) serve as the immune sentinels to the foreign world. In general, APCs can be divided into “professional” and “nonprofessional” types. Although the latter are found among nonlymphoid (e.g., vascular endothelial) cells, professional APCs, such as dendritic cells (DCs), macrophages, and B cells, form an integral part of the immune system and are bone marrow (BM)-derived. DCs are now recognized as essential regulators of both the innate and acquired arms of the immune system. ¹ They comprise a system of highly efficient APCs that initiate immune responses, such as sensitization of naive T cells by major histocompatibility complex (MHC) molecules, rejection of organ transplants, and formation of T-cell–dependent antibodies, ^{1 2} and they may play a role in induction of tolerance. ^{3 4} DCs are also thought to be important in initiating autoimmune disease by efficiently presenting autoantigens to self-reactive T cells that mount autoimmune reactions. ⁵  

DCs, first isolated from lymphoid tissue of mice in 1973 by Steinman and Cohn, ⁶ are a heterogeneous group of leukocytes that include members of different lineages and states of maturation. ^{2 7 8} Some DCs and macrophages are derived from a myeloid lineage, whereas others have a lymphoid lineage. MHC class II (murine Ia; henceforth, Ia)–negative proliferating DC progenitors from the BM give rise to nonproliferating DC precursors in the blood that seed nonlymphoid tissues in a stage referred to as immature DCs. ^{2 9 10} As do all leukocytes, they express CD45 (leukocyte-common antigen), but unlike monocytes or macrophages, they express CD11c, a DC-specific marker. ^{11 12} Immature DCs express negligible amounts of Ia on their surfaces, are able to take up and process antigen, and are localized in the interstitial spaces of many solid organs (heart, liver, and kidney). ^{10 13 14 15} In their immature state, they do not have the requisite accessory signals for T-cell activation, such as CD40, CD80, and CD86, and remain dormant until signals in the extracellular milieu through inflammatory mediators (derived from microbes or distressed bystander cells) induce a rapid change in function, also known as activation or maturation. 

Under nonpathologic circumstances, Langerhans cells (LCs) of the epithelium, a subset of DCs, are thought to be the only cells that constitutively express Ia molecules in the cornea. In the past, studies examining the cornea for APCs largely relied on expression of MHC class II in these cells. Until very recently, reports in the guinea pig, hamster, mouse, and human, led to the dogma that APCs are absent from the central epithelium and the stroma, ^{16 17 18 19 20 21 22} although isolated MHC class II⁺ or CD45⁺ cells had been observed in the normal corneal stroma of various species, mostly in the periphery and in the anterior stroma. ^{23 24 25 26 27 28 29 30 31 32 33 34}  

This paradigm was recently jolted when data from our group demonstrated that the cornea is indeed endowed with resident DCs that are universally MHC class II⁻ but are capable of expressing class II antigen after surgery and migrating to draining lymph nodes (LNs) of allograft hosts. ³⁵ Many of these cells appear to be LC-type DCs that reside in the central corneal epithelium. ³⁶ More recently, Brissette-Storkus et al. ³⁷ have shown that BM-derived cells also reside in the normal uninflamed murine stroma and have identified them as macrophages. ³⁷ However, to date, a systematic study examining the phenotype and distribution of DCs in the corneal stroma has not been performed. The purpose of this study was to extend our preliminary observations ^{35 36} and those of Brissette-Storkus et al. ³⁷ and to characterize more fully the lymphoreticular populations in the corneal stroma. In the current study, we demonstrate that the normal corneal stroma contained large numbers of resident BM-derived cells of different lineages and that these cells were not only macrophages, but also CD11c⁺ DCs. This is the first reported study with these findings. 

Methods

Experimental Animals

Seven- to 14-week-old male BALB/c, C57BL/6, and C3H mice (Taconic Farms, Germantown, NY, or our own breeding facility) were used in these experiments. Most experiments were performed on BALB/c mice, and experiments on other strains were performed only when noted. All protocols were approved by the Schepens Eye Research Institute Animal Care and Use Committee, and all animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Antibodies

The primary antibodies (Abs; all from PharMingen, San Diego, CA, except where noted) used in the immunohistochemical, immunocytochemical, and flow cytometric (FACS) staining procedures, their specificity, and their respective control antibodies (all from PharMingen), are summarized in Table 1 . The secondary antibodies were Cy5-conjugated goat anti-Armenian hamster IgG (PharMingen), rhodamine- and FITC-conjugated goat anti-rat IgG (Santa Cruz Biotechnology, Santa Cruz, CA), and Cy3-conjugated goat anti-mouse. The anti-keratan sulfate antibody and the secondary Cy3 antibody were gifts of J. Wayne Streilein (Schepens Eye Research Institute). 

Immunohistochemical Studies

Normal corneas were excised and the epithelium removed from the stroma by a modification of a technique previously described. ³⁸ Briefly, freshly excised corneas were immersed in phosphate-buffered saline (PBS), containing 20 mM EDTA (Sigma Chemical Co., St. Louis, MO) at 37°C for 1 hour. The epithelium was removed from the underlying stroma with forceps and the stromal layer washed in PBS. Corneal stromas were fixed in acetone for 15 minutes at room temperature (RT) for immunofluorescence staining for CD3, CD8α, CD11b, CD11c, CD14, CD40, CD45, CD80, CD86, GR-1, keratan sulfate, and Ia^d-type class II MHC. The purified CD11c antibody was used, with Cy5 as the secondary antibody. For optimal results, other forms of fixation were tried initially, but fixation with acetone produced the best results. Corneas were than incubated in 2% bovine serum albumin (BSA), diluted in PBS (PBS-BSA) for 15 minutes. To block nonspecific staining, sections were blocked with anti-Fc receptor (FcR) mAb (CD16/CD32) for 30 minutes before they were immunostained with primary antibodies or isotype-matched control antibodies for 2 hours. Afterward, corneal tissues were incubated for 1 hour with a second FITC- or phycoerythrin (PE)-conjugated primary antibody or with secondary antibodies (all diluted for optimal concentrations in PBS-BSA). All staining procedures were performed at RT, and three thorough washings in PBS of 5 minutes each followed every step. Finally, corneas were covered with a mounting medium (Vector, Burlingame, CA) and analyzed by a confocal microscope (Leica TCS 4D; Lasertechnik, Heidelberg, Germany). Central, paracentral, and peripheral areas for each cornea were assessed separately, as described before. ³⁶ At least three different corneas were examined per each double-staining experiment. Representative data are presented in the Results section. To count the number of positively labeled cells in the different areas, multiple z-sections were generated throughout the stroma and stacked. The anterior and posterior stromas were analyzed separately as necessary. Five to eight different fields were analyzed for each specimen, by using a grid, and the numbers were averaged. Student’s t-test was used to compare the number of positively labeled stromal cells in different areas of the stroma. P < 0.05 was considered significant. 

Corneal Stroma Culture

Corneal buttons were excised and placed into a six-well plate, with 10 buttons per well, after the epithelium was removed with forceps. Buttons were cultured in 2.5 mL RPMI-1640 medium with 10% fetal bovine serum (FBS; Hyclone, Salt Lake City, UT), 10 mM HEPES, 0.1 mM nonessential amino acid, 100 U/mL penicillin, 100 μg/mL streptomycin (BioWhittaker, Walkersville, MD), and 1 × 10⁻⁵ M 2-mercaptoethonol (Sigma Chemical Co.) and incubated at 37°C for 3 or 7 days. The nonadherent (dendritic) cells were isolated by centrifuging the culture supernatant, resuspending the cells in PBS, and washing them once in PBS. Adherent cells (enriched for macrophages) were collected by washing the wells with cold PBS and incubation on ice for 30 minutes. After incubation, the adherent cells were physically scraped with a plastic scraper. The final suspension of both nonadherent and adherent cells was filtered through a 70-μm nylon cell strainer to remove corneal fragments and then washed with cold PBS and counted. Nonadherent and adherent cells were analyzed by flow cytometry or used in immunocytochemical studies. 

Immunocytochemistry and Flow Cytometry

Cytospin preparations were made from nonadherent and adherent cells of cultured corneal explants and air dried. The cytospin slides were fixed in chilled acetone for 15 minutes, and cells were stained with anti-CD11c, -CD11b, and -CD45, with relevant isotype control antibodies after FcR blockade. All experiments were conducted three times, independently. 

For flow cytometry, the nonadherent or adherent cells were blocked by anti-FcR mAb (CD16/CD32) before cells were labeled with FITC-conjugated rat anti-mouse CD11b and PE-conjugated hamster anti-mouse CD11c. For isotype control, the cells were labeled with FITC-conjugated rat IgG_2b and PE-conjugated hamster anti-mouse CD11c. Cells were washed and analyzed using a flow cytometer (Epics XL; Coulter, Miami, FL). The analysis was done by gating on CD11c or CD11b positive cells using appropriate isotype and cell culture controls to adjust color compensation and gating parameters. Nonadherent or adherent cells of parallel spleen cell cultures were used as controls to evaluate relative CD11c or CD11b expression. The splenic cultures were established with initially adherent cells from naive BALB/c mice, incubated in culture for 90 minutes, and washed, and 2.5 mL 10% FBS RPMI-1640 medium was added to the cultures. The adherent and nonadherent cells were collected as described earlier, and the harvested cells were treated identically as cells derived from corneal explants. Cultures were incubated for 7 days at 37°C. 

Transmission Electron Microscopy

Freshly excised healthy BALB/c corneas were fixed in Karnovsky solution. After three washes in cacodylate buffer, corneas were postfixed for 1.5 hours in 1% osmium tetroxide in the same buffer. Corneas were washed with H₂O, stained in aqueous 2% uranyl acetate, dehydrated, and embedded in Epon. Corneal sections were cut at 6 nm, and a transmission electron microscope (410 TEM; Philips, Eindhoven, the Netherlands) was used for electron microscopy. 

Results

BM-Derived DCs of Different Maturation Stages in the Normal Corneal Stroma

Immunofluorescense confocal microscopy of wholemount murine BALB/c corneal stromas was performed initially with anti-MHC class II, because constitutive expression of this antigen is thought to be a characteristic feature of DCs. Staining revealed the presence of a significant number of Ia⁺ dendritic-shaped cells in the normal corneal stroma (Fig. 1A) , limited to the periphery, with no Ia⁺ cells in the central or paracentral areas. To determine whether these cells were BM-derived (and hence not keratocytes), we performed double staining of stromas with CD45, a panleukocyte marker, and MHC class II. Our results showed unexpected labeling of cells with CD45 throughout the corneal stroma (Fig. 1B) , with the density decreasing from the limbus toward the center. The MHC class II⁺ cells in the periphery were all CD45⁺ (Fig. 1B) , confirming their BM derivation. 

To characterize the maturation stage of these BM-derived cells in the center and periphery of the cornea, wholemount corneas were double stained with CD45 or CD11c (dendritic cell marker), and CD80 (B7.1), CD86 (B7.2), or CD40, accessory molecules that are expressed on the surface of APCs and promote activation of T cells, indicating the APCs’ state of maturation. ³⁹ Cells stained for CD11c, indicating that they were indeed DCs. Some CD11c⁺ cells in the periphery also expressed CD80 (Fig. 1C) , CD86, and CD40 (data not shown), whereas in the central areas of the cornea, CD11c⁺ cells did not express any of these cell surface markers, as shown by double staining for CD11c and CD80 (Fig. 1D) , nor did they express MHC class II. Results similar to those for CD11c⁺ cells were obtained for CD45⁺ cells when tissues were stained for CD45 and CD80, CD86, or CD40 (data not shown). Studies in C57BL/6 and C3H mice demonstrated the presence of similar cells in the corneal stroma (data not shown). Moreover, staining with the isotype control (instead of primary antibodies) showed no staining for all antibodies, demonstrating the specificity of the staining. 

Heterogeneous Phenotypes of BM-Derived Cells in the Corneal Stroma

To confirm that the CD11c⁺ population and the CD45⁺ population are identical and therefore represent DCs, we double stained corneas with CD11c and CD45. We found that all CD11c⁺ cells were also CD45⁺ (Fig. 2A) , confirming the BM origin of these DCs. However, not all BM-derived cells expressed CD11c. These cells were located more in the posterior stroma, whereas the CD11c⁺ cells were located in the anterior third of the stroma. Double staining with CD11c and CD11b (monocyte/macrophage marker) was performed to define further the lineage of the resident DCs in the stroma. Results showed that all CD11c⁺ DCs also expressed the integrin marker CD11b (Fig. 2B) , confirming that they are myeloid and have a monocytic lineage. The density of these CD11c⁺CD11b⁺ DCs decreased from the limbus (mean, 266 cells/mm²) toward the center of the cornea (mean, 135 cells/mm²). Staining of corneas with GR-1 (neutrophil marker), CD3 (T cell marker), CD8α (lymphoid DC marker), and anti-keratan sulfate (microglia marker), as well as staining with isotype controls were all negative (data not shown). These findings excluded that these cells could represent T cells, neutrophils, microglia (also can express CD11c), or DCs from a lymphoid lineage. Therefore, the CD11c⁺CD11b⁺ have the phenotype of myeloid DCs, and the CD11c⁻CD11b⁺ cells the phenotype of monocytes/macrophages. 

To characterize the large number of resident BM-derived cells further, we stained corneal stromas with CD14, an immature cell surface marker reported to be associated with nondifferentiation of DCs and other cells of the myeloid lineage, and found expression of CD14 on a large number of cells in the stroma. Although CD14⁺ cells were uniformly CD45⁺, the CD11c⁺ cells were CD14^dim or CD14⁻. CD14⁺ cells were all Ia⁻, B7⁻, CD40⁻, GR-1⁻, and CD3⁻. The number of CD14⁺ cells was by far larger than the number of CD11c⁺ or CD11b⁺ cells, indicating that the large number of CD14⁺ cells (Fig. 3) represent a population of undifferentiated monocytic precursor cells distinct from DC and macrophage populations described earlier. Staining of C57BL/6 and C3H mice confirmed similar findings as those found in the BALB/c mice in these strains (data not shown) and staining with isotype-matched control antibodies instead of primary antibodies demonstrated no staining. 

Distribution of BM-Derived Cell Subsets

The presence of CD11c⁺CD11b⁺ DCs and CD11c⁻CD11b⁺ macrophage-type cells led us to explore the location of these cells in the vertical axis of the stroma, because the previous results were from stacked z-sections of all stromal layers and an exact localization was not possible. We evaluated the anterior, middle, and posterior aspects of the stroma and found that CD11c⁺CD11b⁺ cells were located exclusively in the anterior stroma (Fig. 4A) . The midportion of the stroma contained only rare or negligent numbers of cells of either type (Fig. 4B) , whereas the posterior stroma was endowed almost exclusively with monocytes/macrophages of the CD11c⁻CD11b⁺ phenotype (Fig. 4C) . In addition, we found the Ia⁺ cells only in the most anterior layer of the peripheral stroma on maximally 50% of the CD11c⁺ DCs (Fig. 4D) . There was no detectable expression of MHC class II in either the mid (Fig. 4E) or posterior (Fig. 4F) layers of the stroma. The same pattern as that for MHC class II expression was seen for CD80 and CD86 (results not shown). Therefore, expression of MHC class II or costimulatory molecule is restricted to the DCs of the anterior stroma and was not found on the monocytes/macrophages of the posterior stroma. When corneas were stained for CD14, CD14⁺ cells were present in all stromal layers, even in the midportion of the stroma, although to a lesser extent than in other layers (data not shown). 

The results of analyses of cell distribution, density, and expression of costimulatory molecule in the anterior third of the normal stroma are summarized in Figure 5A . CD11c⁺CD11b⁺ DCs decreased in density from the periphery to the center. Nearly one half of the DCs were MHC class II⁺ and B7⁺ (CD80 and CD86, respectively) in the peripheral stroma, and there was no expression of MHC class II or B7 molecules in the central or paracentral areas. The number of cells shown in the y-axis represents the total number of cells that expressed a specific cell surface marker across an area of 1 mm² and depth of 20 μm, hence reflecting a volume of 0.02 mm³. A conceptual model for the distribution of these stromal DCs is suggested in Figure 5B . 

Morphology of Stromal BM-Derived Cells

To evaluate the morphology of these resident BM-derived cells, we evaluated corneas at very high magnification. CD11c⁺CD11b⁺ cells in the anterior stroma had a dendritic morphology with typical dendrites (Fig. 6A) , whereas the CD11c⁻CD11b⁺ cells in the posterior stroma had a morphology (larger cells bodies and fewer dendrites) resembling macrophages (Fig. 6B) . In addition to the immunohistochemical studies, TEM was performed to confirm the presence of BM-derived DCs in the stroma. TEM demonstrated the presence of numerous DCs in the anterior stroma, with long processes, a smooth endoplasmic reticulum, active cytoplasm with multivesicular body–like vacuoles, and indented nuclei with a chromatin pattern typical of DCs (Fig. 6C) . Phagolysosomes and Birbeck granules (characteristic of LCs) were not found on these DCs. In contrast to DCs, keratocytes contained a large elongated nucleus that fills the cell body, and the cytoplasm was not as extensive as in DCs. Slender processes, which may taper to only one to several micrometers, left the nuclear region at all angles. Macrophages, in contrast, could be recognized by their large size, villous surface processes, and cytoplasmic inclusions, which often were very large. They had a darker nucleus and a heterogeneous cytoplasm that often contained a large number of vacuoles and small dense granules. 

In Vitro Characterization of Migratory Corneal Cells

Our in vivo results showed at least two populations of partially differentiated resident BM-derived cells: DCs and macrophages. Both cell types are known to migrate from tissue explants. ⁴⁰ Whereas DCs or LCs are known to be nonadherent in culture and to float into the supernatant, macrophages are known to adhere to plastic. ⁴⁰ To confirm our in situ results in vitro, normal corneas were excised and stripped of their epithelium to deplete LCs. Corneas were than placed in culture, as described earlier. We used standard techniques of harvesting adherent and nonadherent floating cells in culture to harvest these cells and stained them for expression of dendritic and macrophage markers. Nonadherent cells demonstrated ample expression of CD11c (Fig. 7A) , which double stained with CD11b (Fig. 7B) or CD45 (data not shown). Adherent cells failed to express CD11c (Fig. 7C) , but stained mostly for CD11b (Fig. 7D) or CD45 (data not shown). 

To confirm the immunocytochemical results, the harvested nonadherent and adherent cells were subjected to two-color staining (CD11c-PE and CD11b-FITC) for flow cytometric analysis. With gating on CD11b⁺ cells, 91% of the nonadherent (DC-enriched) cells also expressed CD11c (Fig. 7C) , whereas only 24% of the adherent cells expressed both markers. When gating on CD11c⁺ cells, only 2% of the nonadherent cells were CD11c⁻CD11b⁺ (data not shown), supporting our in situ results of a CD11c⁺CD11b⁺ DC population, and a CD11c⁻CD11b⁺ monocyte/macrophage population. 

Discussion

Current dogma holds that DCs are absent in the central cornea and that they are present in small numbers only in the peripheral corneal epithelium. ^{16 17 18 19 20 21 22} This putative absence of resident corneal DCs has been cited as a critical facet of corneal immunity. ^{17 18 20 41} However, we have recently shown that the normal cornea contains MHC class II⁻-resident DCs that are capable of expressing class II antigen after surgery and migrating to draining lymph nodes of allografted hosts, ³⁵ but their distribution in the cornea remains unknown. Recently, we demonstrated that the central normal corneal epithelium contains a significant number of MHC class II⁻LCs. ³⁶ This study provides evidence for the first time that the normal anterior one third of the murine stroma is endowed with a large number of resident BM-derived immature DCs in the center and with more mature DCs in the periphery. In addition, we show a population of monocytes/macrophages in the posterior stroma, which probably corresponds with the cell type recently demonstrated by Brissette-Storkus et al. ³⁷  

It is not surprising that the herein-described DC population was not detected previously. Our studies were largely based on a very sensitive technique of confocal microscopy on wholemount corneal stromas, in addition to TEM and flow cytometric analysis. Previously, investigators have used epithelial sheets or transverse cross-sections of the cornea. ^{16 17 18 19 20 21 22} In our experience, even with confocal microscopy, detection of DCs can be very difficult on cross-sections, because the transection of DCs makes it quite difficult to detect them. This, however, may explain isolated observations of MHC class II⁺ or CD45⁺ cells in the stroma in past studies. ^{23 24 25 26 27 28 29 30 31 32 33 34} Moreover, we used different fixation methods and antibody concentrations for each reagent to optimize our experiments. To eliminate any nonspecific staining, we used Fc blockade, because APCs are known to express FcR. 

In addition to the DCs in the anterior stroma, we observed another population of CD11c⁻CD11b⁺ BM-derived cells in the posterior stroma, phenotypically similar to cells recently found by Brissette-Storkus et al. ³⁷ with a monocyte/macrophage phenotype. To confirm our in situ results of these distinct populations, we characterized these cells in culture by immunocytochemistry and flow cytometry, separating them by using their different migration and adhesion patterns. Both cytospin slide staining and flow cytometry confirmed the presence of nonadherent CD11c⁺CD11b⁺ DCs and CD11c⁻CD11b⁺ adherent macrophages. Similar subsets of resident macrophages and MHC class II⁺ DCs have also been identified in the murine uveal tract, ⁴² the rat ciliary body and iris, ⁴³ the rat choroid, ⁴⁴ and the human retina. ⁴⁵  

There are important functional differences between DCs and macrophages that should be emphasized. Members of the DC family play a pivotal role in the initiation of antigen-specific adaptive immune response and in the induction of tolerance. ^{4 13} DCs are 100 times more potent at initiating and perpetuating secondary immune responses than other APCs, such as macrophages. ^{13 46} Macrophages, however, are professional phagocytes and play a pivotal role as effector cells in cell-mediated immunity and inflammation and in other processes including immune regulation, tissue reorganization, and angiogenesis. ⁴⁷ Activated macrophages express low amounts of MHC class II and thus can play a role in antigen presentation in secondary immune responses, but resident tissue macrophages are in general poorly responsive to activation signals. ⁴⁸  

Ia⁻ DC precursors have been identified in the mouse blood, ⁹ and spleen. ⁴⁶ In addition, a distinct DC precursor that precedes development of mature and immature DCs has recently been characterized that is CD11c⁺CD11b⁺. ⁴⁹ This subset of DC precursors has been suggested to correspond to the human CD1a⁻CD14⁺ DC precursor, ⁵⁰ that coexpresses the CD34 progenitor marker, which was also recently detected in the corneal stroma. ⁵¹ There are two different pathways that lead to the generation of myeloid DCs. DCs can branch off early within the myeloid lineage and then appear as immature DCs, as they express a low level of MHC class II. Alternatively, DCs and monocytes/macrophages can develop from a more differentiated monocyte termed an “indeterminate” cell. ^{52 53} Staining of the stroma for CD14, a myeloid cell surface marker that designates relative immaturity, demonstrated a significant number of cells stained for this marker. These CD14⁺ cells were present throughout all layers of the stroma, both in the periphery and in the center. CD11c⁺CD11b⁺ DCs in the anterior stroma stained CD14^dim, which further confirmed their myeloid lineage. Cells that stained CD14^bright in the stroma, were, however, mostly CD11c⁻, thereby likely belonging to an even less differentiated form of precursor BM-derived cells. The presence of an undifferentiated precursor DC, would be similar to the recent finding of DC precursors in the central nervous system, ⁵⁴ where these cells can be skewed toward a more DC or macrophage-like profile in response to different factors. Thus, in contrast to other organs, where terminally differentiated populations of resident DCs and/or macrophages outnumber colonizing precursors, large numbers of DCs within the cornea (and the central nervous system) remain in an undifferentiated state. Whether these stromal cells represent an additional line of immunologic defense other than the epithelial LCs and whether they play a role in induction of tolerance and the immune-privileged state of these tissues remain to be determined. 

To understand the implications of our data for clinical conditions, we have started preliminary experiments with the human cornea. Initial data have demonstrated the presence of HLA-DR⁺ dendritiform cells in the periphery of the human cornea, when evaluating horizontal sections through corneal flatmounts. We caution, however, that we have not phenotyped these cells as thoroughly as we have in the mouse. The identified immunogenetics of the mouse makes it a perfect model system for studying cellular immunology. Conversely, detailed phenotyping of the human cornea requires extensive experimentation with freshly procured tissues (because placing tissues in culture leads to migration of DCs from the explants) from healthy donors, a difficult task that has not been completed by us. 

The presence of BM-derived leukocytes in the corneal stroma, including APC/DC populations, may have important implications for a variety of pathological responses and immunoinflammatory responses in the ocular anterior segment, including alloimmune, autoimmune, and innate immune responses and wound healing. The constitutive presence of these cells in the cornea focuses attention on the cornea as a participant in immune and inflammatory responses, rather than the stroma being essentially a collagenous tissue that simply responds to the activity of infiltrating cells. For example, in transplantation it has been proposed that, because of the putative absence of resident corneal DCs, sensitization to graft antigens is reliant solely on the indirect pathway of sensitization, which requires the processing of antigens by host APCs. ⁵⁵ Our findings suggest that perhaps, under certain conditions, the activation of these resident corneal DCs could lead to direct presentation of graft antigens to host T cells. Similarly, in wound healing, there is a significant literature ^{56 57 58 59 60} regarding the participation of leukocytes in tissue remodeling, cytokine secretion, and scarring. To date, the pathobiology of stromal wound healing has been related only to stromal keratocyte and matrix responses to infiltrating leukocytic populations. Because it is known that DCs can play an essential role as mediators of innate immunity, ^{1 2} which can participate in secretion of inflammatory cytokines such as interleukin-1, ⁶¹ and given the contribution of inflammatory cytokines to stromal wound healing, it is now critical to evaluate the possible role of these cells in stromal wound healing as well. 

Given the exceptional role of DCs in both innate and adaptive cell-mediated immunity it may be rewarding to investigate and manipulate immune and inflammatory responses at the level of DCs. Better understanding of the mechanisms that lead to maturation of DCs and activation in the cornea may lead to novel approaches in the induction of tolerance in transplantation, autoimmunity, and allergy. Further studies are needed to determine the molecular mechanisms that regulate the maturation of these cells, and their immunobiologic phenotype in stimulating, or tolerizing, T cells generated in response to ocular antigens. 

²  These authors contributed equally to the work and should therefore be regarded as equivalent senior authors.

³  Present address: Department of Ophthalmology and Visual Sciences, University of Louisville, Louisville, KY.

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology in Ft. Lauderdale, Florida, May 2001.

 Supported by Grants K08-EY00363 and R01-EY12963 from the National Institutes of Health, a research grant from the Massachusetts Lions Eye Research Fund, and a William and Mary Greve Special Scholar Award from Research to Prevent Blindness (MRD).

 Submitted for publication August 16, 2002; revised September 25, 2002; accepted September 27, 2002.

 Disclosure: P. Hamrah, None; Y. Liu, None; Q. Zhang, None; M.R. Dana, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Monoclonal Antibodies

Primary Antibody (Clone)	Specificity	Control Antibody
Iad-FITC (39-10-8)	BALB/c mouse class II MHC antigen	Mouse IgG3-FITC
CD3e-FITC (145-2C11)	T-lymphocyte marker	Hamster IgG-FITC
CD8α-FITC (53-6.7)	Lymphoid DC marker	Rat IgG2a-FITC
CD11c (HL3)tissue staining	Integrin DC/LC marker	Hamster IgG
CD11c-PE (HL3)flow cytometry	Integrin DC/LC marker	Hamster IgG-PE
CD11b-FITC (M1/70)	MAC-1, Integrin monocyte/macrophage marker	Rat IgG2b-FITC
CD14/CD14-FITC (rmC5-3)	Immature myeloid marker	Rat IgG1/rat IgG1-FITC
CD40 (3/23)	Accessory molecule/maturation marker for APCs	Rat IgG2a
CD45 (30-F11)	Panleukocyte marker	Rat IgG2b
CD80-FITC (16-10A1)	B7-1, costimulatory molecule	Hamster IgG-FITC
CD86-PE(GL1)	B7-2, costimulatory molecule	Rat IgG2a-PE
GR-1-FITC(RB6-8C5)	Neutrophil marker	Rat IgG2b-FITC
CD16/32 (2.4G2)	FcγIII/II receptor (for Fc-block)

Figure 1.

 

Corneal, unlike limbal, CD45⁺ cells, were MHC class II⁻. Confocal micrographs of MHC class II (Ia)–stained corneal stroma demonstrated Ia⁺ dendrite-shaped cells in the periphery of the cornea (A, bottom right), whereas the central areas did not stain for MHC class II (A, top left). Double staining with CD45 (red) and MHC class II (green) showed that all Ia⁺ cells were BM-derived (yellow) and demonstrated CD45⁺ Ia⁻ dendrite-shaped leukocytes throughout the cornea. The density decreased from the limbus (B, bottom right corner) toward the center of the cornea (B, top left corner). Micrographs of double-stained corneas with CD11c (red) and CD80 (green) showed large numbers of CD11c⁺ cells present in the corneal stroma. In the periphery, these DCs coexpressed costimulatory molecules (C, yellow). However, CD11c⁺ cells (red) did not express the B7 costimulatory molecule, CD80 (green), in the center of uninflamed cornea (D). Total thickness: (A–D) 20 μm. Magnification: (A–C) ×160; (D) ×400.

Figure 2.

 

Myeloid DCs in the corneal stroma. Whole-mounted corneal stromal were double-stained with CD45 (green) and CD11c (red). Confocal micrograph of stacked sections of the anterior stroma showed CD11c⁺ dendritic leukocytes as CD45⁺ (A, yellow). CD11b (green) expression of these CD11c⁺ (red- yellow: double labeled) DCs in a micrograph of stacked sections of the whole stroma provided evidence that they were of a myeloid lineage. In addition CD11c⁻CD11b⁺ cells (green only) were detected (B). Limbus (right), center (left). Total thickness: (A) 30 μm; (B) 50 μm. Magnification, ×160.

Figure 3.

 

The normal corneal stroma contained a significant number of CD14⁺ myeloid cells. Staining of normal whole-mounted corneal stromas with CD14 demonstrated a large number of CD14⁺ cells throughout the corneal stroma, representing undifferentiated precursor DCs. These cells are universally MHC class II⁻ and did not express any costimulatory molecules. Magnification, ×160. Total thickness, 50 μm.

Figure 4.

 

DCs were present in the anterior stroma, whereas monocyte/macrophages were localized in the posterior stroma. Stacked sections of the anterior (A, D), middle (B, E), and posterior stroma (C, F) showed different populations of cells. Corneas double-stained with CD11c (red) and CD11b (green) showed that CD11c⁺CD11b⁺ DCs (yellow) were located exclusively in the anterior stroma (A). The middle part of the stroma harbored a negligible amount of cells (B). The posterior stroma, contained only CD11c⁻CD11b⁺ monocyte/macrophages (C). Double staining with CD11c (red) and MHC class II (green) showed that expression of Ia was limited to the most anterior part of the stromal DCs (D, yellow), whereas the middle and posterior stroma did not stain for MHC class II (E, F). In addition, CD11c⁺ DCs were absent in the posterior stroma (F). Total thickness: (A–F) 20 μm. Magnification: (A–F) ×400.

Figure 5.

 

DC phenotype and density. CD11c⁺CD11b⁺ myeloid DCs were present throughout the stroma, with the density decreasing from the periphery toward the center. The data indicate the presence of a large number of MHC class II⁻, costimulatory molecule⁻ DCs in the central regions of the corneal stroma. Mean ± SE (cells/mm²) from five to eight fields of at least three corneas per staining are presented. The volume represents 0.02 mm³ (A, stromal depth of 20 μm). A conceptual model for BM-derived cells in the stroma shows Ia⁺ B7⁺ mature CD11c⁺ DCs in the stromal periphery and Ia⁻ B7⁻ immature or precursor DCs in the corneal center (B). Monocytes/macrophages are seen in the posterior stroma.

Figure 6.

 

Morphology of cells in the anterior and posterior stroma were different. At higher magnification, CD11c⁺CD11b⁺ cells had a dendritic morphology (A), whereas CD11c⁻CD11b⁺ cells in the posterior stroma had a morphology that resembled macrophages (B). The center of the normal uninflamed corneal stroma contained numerous DCs with long processes (arrows) as seen by TEM (C). Magnification: (A, B) ×1000; (C) ×7500.

Figure 7.

 

Two subsets of BM-derived cells migrated out of cultured corneas. Corneal explants were placed in culture, and the nonadherent and adherent cells harvested and stained for expression of dendritic and macrophage markers. Representative data from 3-day-cultured explants show nonadherent CD11c⁺ cells (A). These CD11c⁺ nonadherent cells double-stained uniformly for CD11b (B). Adherent cells did not express CD11c (C) and were CD11b⁺ (D). In (E) nonadherent cells were stained for CD11b and CD11c and analyzed by flow cytometry (CD11c-PE and CD11b-FITC) gating on CD11b⁺ cells. The histogram shows CD11b⁺ cells that also stained for CD11c.

The authors thank their colleagues at the Schepens Eye Research Institute: Wayne Streilein and Ilene Gipson for helpful discussions, Don Pottle (Confocal Microscopy Unit) for excellent technical assistance, Pat Pearson (Morphology Unit) who provided valuable help in the corneal TEM studies, and Peter Mallen (Graphic Services).