CD34+ Corneal Stromal Cells Are Bone Marrow–Derived and Express Hemopoietic Stem Cell Markers
http://www.100md.com
《干细胞学杂志》
a Department of Ophthalmology, University of Aberdeen, Scotland, United Kingdom;
b Department of Neuroimmunology, Center for Brain Research, Medical University of Vienna, Austria
Key Words. Cornea ? CD34 ? Hemopoietic stem cell ? Dendritic cell ? Leukocytes
Correspondence: John V. Forrester, M.D., Department of Ophthalmology, University of Aberdeen, AB255ZD, Aberdeen, Scotland, United Kingdom. Telephone: 0044-122-455-3782; Fax: 0044-122-455-5955; e-mail: j.forrester@abdn.ac.uk
ABSTRACT
Corneal stem cells have been a topic of interest for several years. The focus has almost exclusively been in connection with regeneration of the corneal epithelium via so-called limbal stem cells. Limbal stem cells are considered to be a population of cells with extensive proliferative and self-renewal capability . Although they are well recognized, definitive markers have not been established. However, a panel of molecular tags for limbal stem cells has become the standard for identification of these cells, including p63, ABCG2 integrin alpha9 beta1, epidermal growth factor receptor (EGFR), K19, enolase-alpha, and CD71 . It is also established that limbal stem cells do not express hemopoietic stem cell (HSC) markers such as CD34 and CD133 . The self-renewal capacity of stromal keratocytes and corneal endothelial cells is less well recognized, at least in humans, and, consequently, the search for stem cells at these sites has not been pursued to any great extent.
The mechanisms for tissue and cell maintenance and renewal during adult life are considered to depend on pluripotent stem cells. Recently, a more general role has been proposed for bone marrow stem cells in tissue regeneration such as in cardiac muscle cells, Purkinje neurons in the brain, and liver cells . In these studies it was suggested that bone marrow cells have the potential to form new tissue cells especially after injury. Hemopoietic cells thus may either transdifferentiate into a fully mature tissue-specific cell or more likely fuse with existing cells to form a multinucleated cell. HSCs are characterized by a set of discrete molecular markers, including CD34, CD133, and Sca-1 (Ly-6A/E) . CD34 is also expressed on vascular endothelium.
Recent studies have suggested that corneal keratocytes express CD34 . The observations were made on single-stained immunohistochemical preparations of human corneal tissues containing stromal cells with the morphology of keratocytes. The cells also expressed the L-selectin ligand, CD62L. No CD34+ cells were found in corneal epithelium or endothelium . In addition, CD34+ cells were very recently found in corneas with Mooren’s ulcer . Intriguingly, culture of stromal keartocytes is associated with loss of the CD34 marker .
In work by others, the corneal stroma in the mouse has also been shown to contain a population of CD45+ leukocytes , and thus the possibility exists that the CD34+ population of stromal cells may in fact be true HSCs and not keratocytes expressing CD34. In the present study, we decided to investigate more closely the resident cells in normal mouse cornea using dual and triple immunostaining, flow cytometric analysis, and short-term cell culture.
MATERIALS AND METHODS
Confocal Microscopy of Mouse Corneal Whole Mounts
Overview of Corneal Cell Populations ? Normal mouse cornea consists of three cellular layers and two interfaces: epithelium, Bowman’s layer, stroma, Descemet’s membrane, and endothelium. The thickness of mouse cornea is approximately 100 μm . The epithelium is comprised of several layers of squamous epithelial cells overlying a layer of small hexagonal, uniformly sized basal cells (Figs. 1A, 1B). In the mouse, the corneal epithelium represents more than one third of normal corneal thickness. The corneal endothelium comprises a single layer of polygonal cells (Fig. 1C).
Figure 1. Confocal microscopy of in situ fixed corneal specimens stained with phalloidin. (A): Surface squamous epithelial cells; note polyhedral shape and tendency to overlap. (B): Basal epithelial cells; note small, uniform size and hexagonal shape. (C): Endothelial monolayer; note less well-defined actin pericellular belt delimiting regular hexagonal shape. (D): Stromal cells demonstrating very large flat stellate structure. (E): For comparison, a phalloidin-stained section of cornea demonstrates extremely thin typical spindle-shaped appearance of the stromal cells. (F): Propidium iodide stain of stromal cells showing nuclear staining only. All images were taken with a x40 objective.
Whole-mount phalloidin-stained preparations of the epithelium-denuded cornea revealed that the stroma contained a dense population of very large flat stellate, scallop-edged cells(Fig. 1D). This contrasted with the characteristic scattered, spindle-shaped cell forms seen in conventional sections of corneal stroma (Fig. 1E). To estimate the full complement of cells in normal mouse corneal stroma, epithelium-denuded corneas were stained with propidium iodide as a general stain for cell nuclei (Fig. 1F). Using confocal microscopy, multiple Z-sections were obtained through the entire stromal thickness and a single image was created. It was thus possible to determine the number of nuclei per unit area (mm2) for the entire thickness of corneal stroma. Our data reveal that there are 3,710 ± 907 cells per mm2 in the normal mouse corneal stroma (Fig. 1F).
Characterization of Corneal Stromal Cells ? We next wished to determine the level of heterogeneity in the stromal cell population, because previous studies have shown that, in addition to keratocytes, there are discrete populations of leukocytes in the stroma . Because keratocytes in the human are reportedly CD34+ , we investigated CD34 antigen expression in the mouse corneal stroma. Single immunostaining showed significant numbers of CD34+ cells (Fig. 2A), which were distributed throughout the cornea. In the periphery, they numbered 122 ± 33/mm2, whereas in the center there were somewhat fewer (88 ± 13/mm2). They thus represented approximately 2.4%–3.3% of the total population of stromal cells depending on location to periphery or center of the cornea. No CD34+ cells were observed in epithelium or on endothelium. The stromal CD34+ cells had a generally rounded, variable morphology, with several fine processes; they measured approximately 20 μm in diameter, which was significantly smaller that the more frequent stellate, scallop-edged stromal cells (compare Figs. 2B and 1D)
Figure 2. Confocal microscopy of epithelium-denuded mouse corneal stroma. (A): Low-power view of CD34+ cells showing distribution of mononuclear, round-edged cells with processes. (B): Higher-power view of CD34+ cells showing fine processes. (C): Low-power view of CD45+ cells in corneal stroma, showing similar distribution to CD34+ cells. (D): CD45+ cell showing rounded cell body with fine processes and punctate staining areas. (E): CD45+ cell showing large cells with extensive dendriform cell processes. (F): Same as (E). Images (A) and (C) were taken with a x20 objective; images (B), (D), (E), and (F), with a x40 objective.
These differences suggested to us that the CD34+ cells in the mouse cornea were not keratocytes but may form part of the CD45+ bone marrow–derived, leukocytic cell populations recently described . Immunostained normal mouse corneas showed the presence of significant numbers of CD45+ cells (Fig. 2C) throughout the central, paracentral, and peripheral parts of the corneal stroma as described previously . Morphologically, there were two different subsets of CD45+ cells: a round compact mononuclear-type cell found throughout the corneal stroma (Fig. 2D) and a second cell type restricted to the peripheral cornea and anterior stroma, which appeared much larger with very long dendriform processes (Figs. 2E, 2F). The number of CD45+ cells in the normal mouse stroma was 215 ± 45/mm2 in the peripheral part and 138 ± 26/mm2 in the central part of the cornea, representing 5.8% and 3.7%, respectively, of the total cell population. Immunostaining of corneal stroma cells with the macrophage integrin marker CD11b showed similar results as with CD45 (173 ± 19 cells/mm2 for peripheral cornea and 164 ± 38 cells/mm2 for central region of the cornea). To investigate the possible presence of dendritic cells in the corneal stroma, corneas were stained for the dendritic cell marker CD11c. Only a few positive cells were found in the peripheral parts of cornea; no CD11c+ cells were detected in the central parts (data not shown).
To characterize additionally the resident stromal cells, normal corneal stromas were double- and triple-stained with antibodies to various leukocyte differentiation markers in combination with phalloidin to visualize F-actin. Simultaneous two-color staining with antibodies to CD45 and CD11b indicated that 100% of CD45+ cells were also CD11b+ (Table 1). As in the single immunostaining studies above, both markers were expressed on two different subsets of cells, distinguished by the expression of MHC class II antigen. MHC class II was coexpressed on 40 ± 10% of CD45+ cells, mostly restricted to the large dendriform cells in the anterior third and the periphery of the corneal stroma (Table 1, Fig. 3A). Only occasional MHC class II+ cells could be found in the central area of the corneal stroma.
Table 1. Coexpression of leukocyte antigens on cells in corneal stroma
Figure 3. Double immunofluorescence staining of corneal CD45+ cell population. (A): Dual staining for CD45 (green) and major histocompatibility complex (MHC) class II (red); note dual-stained cell with prominent dendriform MHC class II positive processes. (B): Dual staining for CD45 (green) and CD34 (red); note that CD34+ cells were compact rounded cells, and all coexpressed CD45. Both images were taken with a x40 objective.
Dual staining of corneal stromal cells for CD34 and CD45 revealed that all CD34+ cells coexpressed CD45+ and that they represented 66% ± 11% of all CD45+ cells (Table 1, Fig. 3B). Furthermore, as predicted from the above data, 100% of the CD34+ cell population coexpressed CD11b+. However, no CD34+ cells expressed MHC class II or CD11c. In addition, less than 10% of CD34+ cells coexpressed the stem cell marker Sca-1 (Table 1).
Most murine dendritic cell populations are CD11c+ . However, there is a small population of dendritic cells recently characterized that are CD11clo or negative . These cells, termed plasmacytoid dendritic cells (pDCs), are B220+ and variably expressive of Gr-1, the neutrophil marker. Accordingly, we examined mouse cornea for the presence of these cells by dual and triple staining for CD45, B220, and Gr-1. A very small population of CD45+ B220+Gr-1+ cells was found predominantly in the peripheral cornea (Fig. 4). Occasional CD45+ cells in this sub-population also either expressed only the B220 marker or the Gr-1 marker, suggesting an intermediate phenotype (Fig. 4).
Figure 4. Plasmacytoid dendritic cells in corneal stroma. Triple staining for CD45 (green), CD45R/B220 (red), and GR-1 (blue). Note triple-positive large cell and second cell expressing only CD45 and B220. Image was taken with a x40 objective.
Flow Cytometry
To confirm the immunohistochemical results, we prepared single-cell suspensions from 70 collagenase-digested naive corneas for flow cytometry analysis as described in Materials and Methods. Two separate experiments were performed using the same technique. Cells were stained for CD34, CD45, and CD11b. To focus on bone marrow–derived cells, we gated on the CD45+ cell population, representing approximately 5% of the total cell population. In both experiments, similar results were obtained, as follows: 68.0% and 61.54% of the CD45+ cell population coexpressed both CD34 and CD11b, respectively. The proportion of CD45+CD11b+CD34– cells represented 22.09% and 21.4%, respectively. Only a small percentage of the CD45+ cell population failed to express CD11b while being CD34+ (3.49% and 9.43%, respectively). In addition, a small percentage of cells expressed CD45 alone (Fig. 5).
Figure 5. Flow cytometry of CD45+ cell population. Cells were isolated from collagenase-digested mouse corneas and gated on CD45+ population, representing approximately 5% of the total cell population. A total of 68 % of CD45+ cells coexpressed both CD34 and CD11b; CD11b+CD34– cells represented 22.09% of the CD45+ population.
Anatomical Relationship Between Bone Marrow–Derived Cells and Tissue-Resident Stromal Cells
Because bone marrow–derived corneal leukocytes comprised more than one type of cell and seemed to be restricted to different regions of the stroma, it was of interest to evaluate the relationship between the CD45+ and CD45– cells in situ. This was performed by dual and triple staining with phalloidin for F-actin to visualize the CD45– cell population and the various leukocyte markers, CD45, CD34, MHC class II, and CD11b. Tissue-resident CD45– stromal cells (presumed keratocytes) represent most corneal cells and staining only with phalloidin and showed the typical morphology: scallop-edged, large stellate cells with multiple processes, connecting between individual cells. The smaller rounded CD34+CD45+ cells seemed to lie in intimate contact with the large stellate CD34–CD45– cells (Fig. 6A) and were distributed throughout the stroma. Most of these cells stained only for F-actin, whereas a small percentage also stained for G-actin, indicating their motile nature (Fig. 7). In contrast, in the anterior stroma, the large CD34–, CD45+ (Fig. 6B), CD11b+ (Fig. 6C), and MHC class II+ (Fig. 6D) cells that are restricted to the anterior and peripheral stroma adopted dendriform-shaped cells, which seemed to be predicated upon by their contact with the F-actin+CD34– stromal keratocytes. Indeed, their intercellular contacts seemed to follow precisely the contours of the keratocyte, presumably thus ensuring maximum cell–cell contact interface.
Figure 6. Anatomical relationship between bone marrow–derived cells and tissue-resident stromal cells. (A): CD34+F-actin+ small round cells attached to the cell bodies of the keratocytes (CD34–F-actin+ cells). (B): CD45+F-actin+ cells associated with CD45–F-actin+ keratocytes. (C): CD11b+ cells adopting a morphology dictated by the spaces between the keratocytes (jigsaw effect). (D): MHC class II+ cells similar to (C). All images were taken with a x40 objective.
Figure 7. Triple staining of cornea showing F-actin+ cells (red), CD45+ cells (blue), and G-actin+ cells (green). Most CD45+ cells were negative for G-actin, but a small proportion (5%) of the round CD45+ cells were G-actin+. These cells are also CD34+ (see Fig. 3B). Image was taken with a x40 objective.
Studies in the GFP–bone marrow reconstituted chimeric rat revealed that there was a clear population of bone marrow–derived cells in the corneal stroma (113 ± 24/mm2) (Fig. 8). There seemed to be two morphological types, one form adopting a spindle-shaped phenotype and the other being more rounded (Fig. 8A). Staining with phalloidin clearly indicated that they were a separate population of cells from the stromal keratocytes (Fig. 8B). They also uniformly expressed CD45+ (Fig. 8A, inset 1). A small percentage of the spindle-shaped cells also expressed MHC class II (Fig. 8A, inset 2). Examination of chimeric rats at different ages indicated that the numbers of CD45+ cells remained constant between 2 months and 1 year after reconstitution (data not shown). No GFP+ cells were found in corneal epithelium or endothelium.
Figure 8. Confocal microscopy corneal stroma of eGFP–bone marrow–reconstituted chimeric rat. (A): Two different types of bone marrow–derived eGFP+ cells are found in the rat corneal stroma: long spindle-shaped cells and a more rounded cell. (B): Staining with phalloidin (red) clearly shows two different cell populations: the eGFP+F-actin+ bone marrow–derived cells and eGFP–F-actin+ stro-mal keratocytes; eGFP+ cells uniformly expressed CD45+ (A, inset 1). A small percentage of the spindle-shaped cells also expressed major histocompatibility complex class II (A, inset 2). All images were taken with a x40 objective. Abbreviation: eGFP, enhanced green fluorescent protein.
Short-Term Cell Culture of Stromal Leukocytes
In explant cultures, resident leukocytes migrate out of tissues during the first hours or days of culture. In an attempt to isolate CD45+ cells from corneal tissue, we prepared small explant cultures of epithelium-denuded corneal stromal tissue in Dulbecco’s modified Eagle’s medium plus 10% fetal bovine serum, as described in Materials and Methods. After 3 days, nonadherent cells were harvested from the supernatant and cytospins were stained for CD34 and CD45 initially. Positively labeled cells for both leukocyte differentiation markers were observed (Figs. 9A, 9B). To confirm the bone marrow origin of the CD34+ cells, we dual stained some cytospin preparations with CD45. All CD34+ cells after short-term cell culture coexpressed CD45 (Fig. 9C). No CD11c+ cells were observed in these preparations, and negative controls did not show any staining (Fig. 9D).
Figure 9. Short-term cell culture of stromal leukocytes. Nonadherent cells harvested from the supernatant from corneal explants express (A) CD34 and (B) CD45. (C): Simultaneous staining for CD34 (red) and CD45 (green) shows double-positive cells; nuclei are visualized with DAPI. (D): Negative control shows only DAPI nuclear staining.
DISCUSSION
We believe that in addition to confirming the presence of the recently described novel resident macrophage in the corneal stroma , this study provides the first description of yet another unusual type of corneal stromal leukocyte, namely HSCs, some of which are in an intermediate stage of differentiation. What remains unclear is the role of HSCs in the corneal stroma. Recent studies have suggested that HSCs provide a means for replenishing tissue cells, whether by fusing with endogenous parenchymal cells or by directly differentiating into tissue cells. At least in the heart, the evidence for the latter process has not been found , and current concepts suggest that HSC fusion with aging tissue cells is a more likely mechanism. To date there is no evidence that such a process occurs in the eye, and for the moment the most apparent role for corneal HSCs is to replenish tissue resident macrophages and pDCs.
ACKNOWLEDGMENTS
Dua HS, Azuara-Blanco A. Limbal stem cells of the corneal epithelium. Surv Ophthalmol 2000;44:415–425.
Chen Z, de Paiva CS, Luo L et al. Characterization of putative stem cell phenotype in human limbal epithelia. STEM CELLS 2004;22:355–366.
Dua HS, Joseph A, Shanmuganathan VA et al. Stem cell differentiation and the effects of deficiency. Eye 2003;17:877–885.
Alvarez-Dolado M, Pardal R, Garcia-Verdugo JM et al. Fusion of bone-marrow–derived cells with Purkinje neurons, cardiomyocytes and hepatocytes. Nature 2003;425:968–973.
Ikehara S. Pluripotent hemopoietic stem cells in mice and humans. Proc Soc Exp Biol Med 2000;223:149–155.
Toti P, Tosi GM, Traversi C et al. CD-34 stromal expression pattern in normal and altered human corneas. Ophthalmology 2002;109:1167–1171.
Joseph A, Hossain P, Jham S et al. Expression of CD34 and L-selectin on human corneal keratocytes. Invest Ophthalmol Vis Sci 2003;44:4689–4692.
Ye J, Chen J, Kim JC et al. Bone marrow-derived cells are present in Mooren’s ulcer. Ophthalmic Res 2004;36:151–155.
Espana EM, Kawakita T, Liu CY et al. CD-34 expression by cultured human keratocytes is downregulated during myofibroblast differentiation induced by TGF-beta1. Invest Ophthalmol Vis Sci 2004;45:2985–2991.
Brissette-Storkus CS, Reynolds SM, Lepisto AJ et al. Identification of a novel macrophage population in the normal mouse corneal stroma. Invest Ophthalmol Vis Sci 2002;43:2264–2271.
Liu Y, Hamrah P, Zhang Q et al. Draining lymph nodes of corneal transplant hosts exhibit evidence for donor major histocompatibility complex (MHC) class II-positive dendritic cells derived from MHC class II-negative grafts. J Exp Med 2002;195:259–268.
Lois C, Hong EJ, Pease S et al. Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 2002;295:868–872.
Schulz D, Iliev ME, Frueh BE et al. In vivo pachymetry in normal eyes of rats, mice and rabbits with the optical low coherence reflectometer. Vision Res 2003;43:723–728.
Hamrah P, Huq SO, Liu Y et al. Corneal immunity is mediated by heterogeneous population of antigen-presenting cells. J Leukoc Biol 2003;74:172–178.
Heath WR, Belz GT, Behrens GM et al. Cross-presentation, dendritic cell subsets, and the generation of immunity to cellular antigens. Immunol Rev 2004;199:9–26.
Espana EM, He H, Kawakita T et al. Human keratocytes cultured on amniotic membrane stroma preserve morphology and express keratocan. Invest Ophthalmol Vis Sci 2003;44:5136–5141.
Guo Y, Lubbert M, Engelhardt M. CD34– hematopoietic stem cells: current concepts and controversies. STEM CELLS 2003;21:15–20.(Magdaléna Sosnováa, Monik)
b Department of Neuroimmunology, Center for Brain Research, Medical University of Vienna, Austria
Key Words. Cornea ? CD34 ? Hemopoietic stem cell ? Dendritic cell ? Leukocytes
Correspondence: John V. Forrester, M.D., Department of Ophthalmology, University of Aberdeen, AB255ZD, Aberdeen, Scotland, United Kingdom. Telephone: 0044-122-455-3782; Fax: 0044-122-455-5955; e-mail: j.forrester@abdn.ac.uk
ABSTRACT
Corneal stem cells have been a topic of interest for several years. The focus has almost exclusively been in connection with regeneration of the corneal epithelium via so-called limbal stem cells. Limbal stem cells are considered to be a population of cells with extensive proliferative and self-renewal capability . Although they are well recognized, definitive markers have not been established. However, a panel of molecular tags for limbal stem cells has become the standard for identification of these cells, including p63, ABCG2 integrin alpha9 beta1, epidermal growth factor receptor (EGFR), K19, enolase-alpha, and CD71 . It is also established that limbal stem cells do not express hemopoietic stem cell (HSC) markers such as CD34 and CD133 . The self-renewal capacity of stromal keratocytes and corneal endothelial cells is less well recognized, at least in humans, and, consequently, the search for stem cells at these sites has not been pursued to any great extent.
The mechanisms for tissue and cell maintenance and renewal during adult life are considered to depend on pluripotent stem cells. Recently, a more general role has been proposed for bone marrow stem cells in tissue regeneration such as in cardiac muscle cells, Purkinje neurons in the brain, and liver cells . In these studies it was suggested that bone marrow cells have the potential to form new tissue cells especially after injury. Hemopoietic cells thus may either transdifferentiate into a fully mature tissue-specific cell or more likely fuse with existing cells to form a multinucleated cell. HSCs are characterized by a set of discrete molecular markers, including CD34, CD133, and Sca-1 (Ly-6A/E) . CD34 is also expressed on vascular endothelium.
Recent studies have suggested that corneal keratocytes express CD34 . The observations were made on single-stained immunohistochemical preparations of human corneal tissues containing stromal cells with the morphology of keratocytes. The cells also expressed the L-selectin ligand, CD62L. No CD34+ cells were found in corneal epithelium or endothelium . In addition, CD34+ cells were very recently found in corneas with Mooren’s ulcer . Intriguingly, culture of stromal keartocytes is associated with loss of the CD34 marker .
In work by others, the corneal stroma in the mouse has also been shown to contain a population of CD45+ leukocytes , and thus the possibility exists that the CD34+ population of stromal cells may in fact be true HSCs and not keratocytes expressing CD34. In the present study, we decided to investigate more closely the resident cells in normal mouse cornea using dual and triple immunostaining, flow cytometric analysis, and short-term cell culture.
MATERIALS AND METHODS
Confocal Microscopy of Mouse Corneal Whole Mounts
Overview of Corneal Cell Populations ? Normal mouse cornea consists of three cellular layers and two interfaces: epithelium, Bowman’s layer, stroma, Descemet’s membrane, and endothelium. The thickness of mouse cornea is approximately 100 μm . The epithelium is comprised of several layers of squamous epithelial cells overlying a layer of small hexagonal, uniformly sized basal cells (Figs. 1A, 1B). In the mouse, the corneal epithelium represents more than one third of normal corneal thickness. The corneal endothelium comprises a single layer of polygonal cells (Fig. 1C).
Figure 1. Confocal microscopy of in situ fixed corneal specimens stained with phalloidin. (A): Surface squamous epithelial cells; note polyhedral shape and tendency to overlap. (B): Basal epithelial cells; note small, uniform size and hexagonal shape. (C): Endothelial monolayer; note less well-defined actin pericellular belt delimiting regular hexagonal shape. (D): Stromal cells demonstrating very large flat stellate structure. (E): For comparison, a phalloidin-stained section of cornea demonstrates extremely thin typical spindle-shaped appearance of the stromal cells. (F): Propidium iodide stain of stromal cells showing nuclear staining only. All images were taken with a x40 objective.
Whole-mount phalloidin-stained preparations of the epithelium-denuded cornea revealed that the stroma contained a dense population of very large flat stellate, scallop-edged cells(Fig. 1D). This contrasted with the characteristic scattered, spindle-shaped cell forms seen in conventional sections of corneal stroma (Fig. 1E). To estimate the full complement of cells in normal mouse corneal stroma, epithelium-denuded corneas were stained with propidium iodide as a general stain for cell nuclei (Fig. 1F). Using confocal microscopy, multiple Z-sections were obtained through the entire stromal thickness and a single image was created. It was thus possible to determine the number of nuclei per unit area (mm2) for the entire thickness of corneal stroma. Our data reveal that there are 3,710 ± 907 cells per mm2 in the normal mouse corneal stroma (Fig. 1F).
Characterization of Corneal Stromal Cells ? We next wished to determine the level of heterogeneity in the stromal cell population, because previous studies have shown that, in addition to keratocytes, there are discrete populations of leukocytes in the stroma . Because keratocytes in the human are reportedly CD34+ , we investigated CD34 antigen expression in the mouse corneal stroma. Single immunostaining showed significant numbers of CD34+ cells (Fig. 2A), which were distributed throughout the cornea. In the periphery, they numbered 122 ± 33/mm2, whereas in the center there were somewhat fewer (88 ± 13/mm2). They thus represented approximately 2.4%–3.3% of the total population of stromal cells depending on location to periphery or center of the cornea. No CD34+ cells were observed in epithelium or on endothelium. The stromal CD34+ cells had a generally rounded, variable morphology, with several fine processes; they measured approximately 20 μm in diameter, which was significantly smaller that the more frequent stellate, scallop-edged stromal cells (compare Figs. 2B and 1D)
Figure 2. Confocal microscopy of epithelium-denuded mouse corneal stroma. (A): Low-power view of CD34+ cells showing distribution of mononuclear, round-edged cells with processes. (B): Higher-power view of CD34+ cells showing fine processes. (C): Low-power view of CD45+ cells in corneal stroma, showing similar distribution to CD34+ cells. (D): CD45+ cell showing rounded cell body with fine processes and punctate staining areas. (E): CD45+ cell showing large cells with extensive dendriform cell processes. (F): Same as (E). Images (A) and (C) were taken with a x20 objective; images (B), (D), (E), and (F), with a x40 objective.
These differences suggested to us that the CD34+ cells in the mouse cornea were not keratocytes but may form part of the CD45+ bone marrow–derived, leukocytic cell populations recently described . Immunostained normal mouse corneas showed the presence of significant numbers of CD45+ cells (Fig. 2C) throughout the central, paracentral, and peripheral parts of the corneal stroma as described previously . Morphologically, there were two different subsets of CD45+ cells: a round compact mononuclear-type cell found throughout the corneal stroma (Fig. 2D) and a second cell type restricted to the peripheral cornea and anterior stroma, which appeared much larger with very long dendriform processes (Figs. 2E, 2F). The number of CD45+ cells in the normal mouse stroma was 215 ± 45/mm2 in the peripheral part and 138 ± 26/mm2 in the central part of the cornea, representing 5.8% and 3.7%, respectively, of the total cell population. Immunostaining of corneal stroma cells with the macrophage integrin marker CD11b showed similar results as with CD45 (173 ± 19 cells/mm2 for peripheral cornea and 164 ± 38 cells/mm2 for central region of the cornea). To investigate the possible presence of dendritic cells in the corneal stroma, corneas were stained for the dendritic cell marker CD11c. Only a few positive cells were found in the peripheral parts of cornea; no CD11c+ cells were detected in the central parts (data not shown).
To characterize additionally the resident stromal cells, normal corneal stromas were double- and triple-stained with antibodies to various leukocyte differentiation markers in combination with phalloidin to visualize F-actin. Simultaneous two-color staining with antibodies to CD45 and CD11b indicated that 100% of CD45+ cells were also CD11b+ (Table 1). As in the single immunostaining studies above, both markers were expressed on two different subsets of cells, distinguished by the expression of MHC class II antigen. MHC class II was coexpressed on 40 ± 10% of CD45+ cells, mostly restricted to the large dendriform cells in the anterior third and the periphery of the corneal stroma (Table 1, Fig. 3A). Only occasional MHC class II+ cells could be found in the central area of the corneal stroma.
Table 1. Coexpression of leukocyte antigens on cells in corneal stroma
Figure 3. Double immunofluorescence staining of corneal CD45+ cell population. (A): Dual staining for CD45 (green) and major histocompatibility complex (MHC) class II (red); note dual-stained cell with prominent dendriform MHC class II positive processes. (B): Dual staining for CD45 (green) and CD34 (red); note that CD34+ cells were compact rounded cells, and all coexpressed CD45. Both images were taken with a x40 objective.
Dual staining of corneal stromal cells for CD34 and CD45 revealed that all CD34+ cells coexpressed CD45+ and that they represented 66% ± 11% of all CD45+ cells (Table 1, Fig. 3B). Furthermore, as predicted from the above data, 100% of the CD34+ cell population coexpressed CD11b+. However, no CD34+ cells expressed MHC class II or CD11c. In addition, less than 10% of CD34+ cells coexpressed the stem cell marker Sca-1 (Table 1).
Most murine dendritic cell populations are CD11c+ . However, there is a small population of dendritic cells recently characterized that are CD11clo or negative . These cells, termed plasmacytoid dendritic cells (pDCs), are B220+ and variably expressive of Gr-1, the neutrophil marker. Accordingly, we examined mouse cornea for the presence of these cells by dual and triple staining for CD45, B220, and Gr-1. A very small population of CD45+ B220+Gr-1+ cells was found predominantly in the peripheral cornea (Fig. 4). Occasional CD45+ cells in this sub-population also either expressed only the B220 marker or the Gr-1 marker, suggesting an intermediate phenotype (Fig. 4).
Figure 4. Plasmacytoid dendritic cells in corneal stroma. Triple staining for CD45 (green), CD45R/B220 (red), and GR-1 (blue). Note triple-positive large cell and second cell expressing only CD45 and B220. Image was taken with a x40 objective.
Flow Cytometry
To confirm the immunohistochemical results, we prepared single-cell suspensions from 70 collagenase-digested naive corneas for flow cytometry analysis as described in Materials and Methods. Two separate experiments were performed using the same technique. Cells were stained for CD34, CD45, and CD11b. To focus on bone marrow–derived cells, we gated on the CD45+ cell population, representing approximately 5% of the total cell population. In both experiments, similar results were obtained, as follows: 68.0% and 61.54% of the CD45+ cell population coexpressed both CD34 and CD11b, respectively. The proportion of CD45+CD11b+CD34– cells represented 22.09% and 21.4%, respectively. Only a small percentage of the CD45+ cell population failed to express CD11b while being CD34+ (3.49% and 9.43%, respectively). In addition, a small percentage of cells expressed CD45 alone (Fig. 5).
Figure 5. Flow cytometry of CD45+ cell population. Cells were isolated from collagenase-digested mouse corneas and gated on CD45+ population, representing approximately 5% of the total cell population. A total of 68 % of CD45+ cells coexpressed both CD34 and CD11b; CD11b+CD34– cells represented 22.09% of the CD45+ population.
Anatomical Relationship Between Bone Marrow–Derived Cells and Tissue-Resident Stromal Cells
Because bone marrow–derived corneal leukocytes comprised more than one type of cell and seemed to be restricted to different regions of the stroma, it was of interest to evaluate the relationship between the CD45+ and CD45– cells in situ. This was performed by dual and triple staining with phalloidin for F-actin to visualize the CD45– cell population and the various leukocyte markers, CD45, CD34, MHC class II, and CD11b. Tissue-resident CD45– stromal cells (presumed keratocytes) represent most corneal cells and staining only with phalloidin and showed the typical morphology: scallop-edged, large stellate cells with multiple processes, connecting between individual cells. The smaller rounded CD34+CD45+ cells seemed to lie in intimate contact with the large stellate CD34–CD45– cells (Fig. 6A) and were distributed throughout the stroma. Most of these cells stained only for F-actin, whereas a small percentage also stained for G-actin, indicating their motile nature (Fig. 7). In contrast, in the anterior stroma, the large CD34–, CD45+ (Fig. 6B), CD11b+ (Fig. 6C), and MHC class II+ (Fig. 6D) cells that are restricted to the anterior and peripheral stroma adopted dendriform-shaped cells, which seemed to be predicated upon by their contact with the F-actin+CD34– stromal keratocytes. Indeed, their intercellular contacts seemed to follow precisely the contours of the keratocyte, presumably thus ensuring maximum cell–cell contact interface.
Figure 6. Anatomical relationship between bone marrow–derived cells and tissue-resident stromal cells. (A): CD34+F-actin+ small round cells attached to the cell bodies of the keratocytes (CD34–F-actin+ cells). (B): CD45+F-actin+ cells associated with CD45–F-actin+ keratocytes. (C): CD11b+ cells adopting a morphology dictated by the spaces between the keratocytes (jigsaw effect). (D): MHC class II+ cells similar to (C). All images were taken with a x40 objective.
Figure 7. Triple staining of cornea showing F-actin+ cells (red), CD45+ cells (blue), and G-actin+ cells (green). Most CD45+ cells were negative for G-actin, but a small proportion (5%) of the round CD45+ cells were G-actin+. These cells are also CD34+ (see Fig. 3B). Image was taken with a x40 objective.
Studies in the GFP–bone marrow reconstituted chimeric rat revealed that there was a clear population of bone marrow–derived cells in the corneal stroma (113 ± 24/mm2) (Fig. 8). There seemed to be two morphological types, one form adopting a spindle-shaped phenotype and the other being more rounded (Fig. 8A). Staining with phalloidin clearly indicated that they were a separate population of cells from the stromal keratocytes (Fig. 8B). They also uniformly expressed CD45+ (Fig. 8A, inset 1). A small percentage of the spindle-shaped cells also expressed MHC class II (Fig. 8A, inset 2). Examination of chimeric rats at different ages indicated that the numbers of CD45+ cells remained constant between 2 months and 1 year after reconstitution (data not shown). No GFP+ cells were found in corneal epithelium or endothelium.
Figure 8. Confocal microscopy corneal stroma of eGFP–bone marrow–reconstituted chimeric rat. (A): Two different types of bone marrow–derived eGFP+ cells are found in the rat corneal stroma: long spindle-shaped cells and a more rounded cell. (B): Staining with phalloidin (red) clearly shows two different cell populations: the eGFP+F-actin+ bone marrow–derived cells and eGFP–F-actin+ stro-mal keratocytes; eGFP+ cells uniformly expressed CD45+ (A, inset 1). A small percentage of the spindle-shaped cells also expressed major histocompatibility complex class II (A, inset 2). All images were taken with a x40 objective. Abbreviation: eGFP, enhanced green fluorescent protein.
Short-Term Cell Culture of Stromal Leukocytes
In explant cultures, resident leukocytes migrate out of tissues during the first hours or days of culture. In an attempt to isolate CD45+ cells from corneal tissue, we prepared small explant cultures of epithelium-denuded corneal stromal tissue in Dulbecco’s modified Eagle’s medium plus 10% fetal bovine serum, as described in Materials and Methods. After 3 days, nonadherent cells were harvested from the supernatant and cytospins were stained for CD34 and CD45 initially. Positively labeled cells for both leukocyte differentiation markers were observed (Figs. 9A, 9B). To confirm the bone marrow origin of the CD34+ cells, we dual stained some cytospin preparations with CD45. All CD34+ cells after short-term cell culture coexpressed CD45 (Fig. 9C). No CD11c+ cells were observed in these preparations, and negative controls did not show any staining (Fig. 9D).
Figure 9. Short-term cell culture of stromal leukocytes. Nonadherent cells harvested from the supernatant from corneal explants express (A) CD34 and (B) CD45. (C): Simultaneous staining for CD34 (red) and CD45 (green) shows double-positive cells; nuclei are visualized with DAPI. (D): Negative control shows only DAPI nuclear staining.
DISCUSSION
We believe that in addition to confirming the presence of the recently described novel resident macrophage in the corneal stroma , this study provides the first description of yet another unusual type of corneal stromal leukocyte, namely HSCs, some of which are in an intermediate stage of differentiation. What remains unclear is the role of HSCs in the corneal stroma. Recent studies have suggested that HSCs provide a means for replenishing tissue cells, whether by fusing with endogenous parenchymal cells or by directly differentiating into tissue cells. At least in the heart, the evidence for the latter process has not been found , and current concepts suggest that HSC fusion with aging tissue cells is a more likely mechanism. To date there is no evidence that such a process occurs in the eye, and for the moment the most apparent role for corneal HSCs is to replenish tissue resident macrophages and pDCs.
ACKNOWLEDGMENTS
Dua HS, Azuara-Blanco A. Limbal stem cells of the corneal epithelium. Surv Ophthalmol 2000;44:415–425.
Chen Z, de Paiva CS, Luo L et al. Characterization of putative stem cell phenotype in human limbal epithelia. STEM CELLS 2004;22:355–366.
Dua HS, Joseph A, Shanmuganathan VA et al. Stem cell differentiation and the effects of deficiency. Eye 2003;17:877–885.
Alvarez-Dolado M, Pardal R, Garcia-Verdugo JM et al. Fusion of bone-marrow–derived cells with Purkinje neurons, cardiomyocytes and hepatocytes. Nature 2003;425:968–973.
Ikehara S. Pluripotent hemopoietic stem cells in mice and humans. Proc Soc Exp Biol Med 2000;223:149–155.
Toti P, Tosi GM, Traversi C et al. CD-34 stromal expression pattern in normal and altered human corneas. Ophthalmology 2002;109:1167–1171.
Joseph A, Hossain P, Jham S et al. Expression of CD34 and L-selectin on human corneal keratocytes. Invest Ophthalmol Vis Sci 2003;44:4689–4692.
Ye J, Chen J, Kim JC et al. Bone marrow-derived cells are present in Mooren’s ulcer. Ophthalmic Res 2004;36:151–155.
Espana EM, Kawakita T, Liu CY et al. CD-34 expression by cultured human keratocytes is downregulated during myofibroblast differentiation induced by TGF-beta1. Invest Ophthalmol Vis Sci 2004;45:2985–2991.
Brissette-Storkus CS, Reynolds SM, Lepisto AJ et al. Identification of a novel macrophage population in the normal mouse corneal stroma. Invest Ophthalmol Vis Sci 2002;43:2264–2271.
Liu Y, Hamrah P, Zhang Q et al. Draining lymph nodes of corneal transplant hosts exhibit evidence for donor major histocompatibility complex (MHC) class II-positive dendritic cells derived from MHC class II-negative grafts. J Exp Med 2002;195:259–268.
Lois C, Hong EJ, Pease S et al. Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 2002;295:868–872.
Schulz D, Iliev ME, Frueh BE et al. In vivo pachymetry in normal eyes of rats, mice and rabbits with the optical low coherence reflectometer. Vision Res 2003;43:723–728.
Hamrah P, Huq SO, Liu Y et al. Corneal immunity is mediated by heterogeneous population of antigen-presenting cells. J Leukoc Biol 2003;74:172–178.
Heath WR, Belz GT, Behrens GM et al. Cross-presentation, dendritic cell subsets, and the generation of immunity to cellular antigens. Immunol Rev 2004;199:9–26.
Espana EM, He H, Kawakita T et al. Human keratocytes cultured on amniotic membrane stroma preserve morphology and express keratocan. Invest Ophthalmol Vis Sci 2003;44:5136–5141.
Guo Y, Lubbert M, Engelhardt M. CD34– hematopoietic stem cells: current concepts and controversies. STEM CELLS 2003;21:15–20.(Magdaléna Sosnováa, Monik)