Characteristics of Bone Marrow–Derived Microglia in the Normal and Injured Retina

Abstract

purpose. To compare the distribution and immunologic characteristics of bone marrow (BM)–derived and resident microglia in the retina.

methods. Mice were irradiated and injected with enhanced green fluorescent protein–positive (EGFP⁺) BM cells. One month to 12 months after BM transplantation, eyes were analyzed histologically for the expression of EGFP and various monocyte/microglia/macrophage markers (Iba-1, F4/80, GS-1, major histocompatibility complex [MHC] class II). N-methyl-N-nitrosourea (MNU) was injected or retinal detachment was created to induce retinal damage.

results. Many BM-derived EGFP⁺ cells were found in the ciliary body and choroid and around the optic nerve in the uninjured eyes. Within the retina, few such cells existed at the retinal margin and juxtapapillary area at 3 to 12 months after BM transplantation. However, after MNU injection, many EGFP⁺ cells were found in the retina adjacent to the retinal vessels, optic nerve, and ciliary body that rapidly spread throughout the retina. Most of them showed morphologic and immunohistochemical features of microglia. By 7 days after MNU injection, EGFP⁺ BM-derived cells occupied approximately 15% of the total Iba-1⁺ retinal microglia. Meanwhile, the proportion of MHC class II⁺ cells was larger among BM-derived (EGFP⁺/Iba-1⁺) than resident (EGFP⁻/Iba-1⁺) microglia. In the eyes with retinal detachment, EGFP⁺/F4/80⁺ cells engrafted exclusively around the detached retina.

conclusions. In response to retinal damage, numerous BM-derived cells migrated to the retina from the ciliary body, optic nerve, and retinal vessels and differentiated into microglia. The higher rate of immunologic activation and the increased specificity to the damaged site appeared to be the characteristic features of BM-derived microglia.

The potential of circulating bone marrow (BM)–derived cells to migrate into various tissues and adopt different cell types has been reported by numerous groups. Many researchers took advantage of BM transplantation to distinguish the population of donor-derived cells that usually enter the target tissues from the peripheral circulation. When the donor-derived cells engraft to the organs, they differentiate into various cell types such as vascular endothelial cells, ^{1 2} myocytes, ^{2 3 4} or pancreatic endocrine cells. ^{5 6} In the brain, BM-derived cells were reported to differentiate into glial cells ^{7 8 9 10} or neurons. ^{11 12 13} However, the physiological role of BM-derived cells in the homeostasis of the normal brain is unclear. Meanwhile, in response to neural damage, a greatly increased number of BM-derived cells migrate to the brain. ^{9 10} Most of these cells appear to differentiate into microglia, tissue macrophages of the central nervous system. ^{9 10} The potential roles of BM-derived cells in the damaged brain include the removal of neurotoxic substances. ¹⁴  

In addition, other groups have shown the migration of circulating BM-derived cells to various parts of the eye, including the retina, ^{15 16 17 18} choroid, ^{17 19 20 21 22} retinal pigment epithelium, ²³ and cornea. ²⁴ Cell types that differentiate from BM-derived cells include microglia/macrophages, vascular endothelial cells, ^{17 19 20 21 22} and retinal pigment epithelial cells. ²³ Although the increased homing of the BM-derived cells to the injury site in nonneural tissues of the eye has been reported, ^{17 19 20 21 22 23 25} little is known about the nature and the role of BM-derived cells in the neural retina. 

The suggested roles of microglia in the retina include secretion of neurotrophic factors, ^{26 27} phagocytosis of cellular debris, ^{27 28 29} and killing of invading microorganisms. ²⁹ However, most of the previous works did not distinguish between the two different categories, resident and BM-derived microglia. Resident cells migrate from hyaloid vessels to the retina during ocular development. ^{30 31 32} These cells are thought to be closely associated with the death and phagocytosis of neurons in retinal histogenesis. ^{30 31 32} BM-derived cells are engrafted to the retina after ocular histogenesis. ^{15 16 17 18} To date, the exact distribution and the role of BM-derived microglia in the retina remain unclear. Similarly, the differences between these two types of microglia are largely unknown. 

In this study, we investigated the role of resident and BM-derived microglia in the retina by comparing the distribution and immunohistochemical characteristics of these cells in the normal and degenerating retina. Our results suggest that BM-derived cells are specifically homed to the eye in response to retinal damage, where they proliferate, differentiate into microglia, and become immunologically active. BM-derived tissue macrophages in the ciliary body and around the optic nerve may serve as a source of microglia in the injured retina. 

Methods

Generation of Chimeric Mice

All animals were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. BM transplantation was conducted as previously reported with minor modifications. ⁴ The BM of 4- to 6-week-old C57BL/6 recipient mice was reconstituted with BM cells from the tibias and femurs of transgenic mice (C57BL/6 background) that ubiquitously expressed enhanced green fluorescent protein (EGFP). ³³ BM cells (1 × 10⁷ cells) were injected intraperitoneally into the recipients 3 to 5 hours after irradiation with x-rays (9 Gy). The eyes of the recipients were protected with lead shields to prevent radiation-induced damage (radiation retinopathy). Four weeks after transplantation, the peripheral blood of chimeric mice was extracted from the tail vein, and the reconstituted BM was assessed. After lysis of the erythrocytes (PharM Lyse; BD PharMingen, San Diego, CA), cell suspensions were analyzed by a fluorescence-activated cell sorter (FACS; FACSCalibur; Becton Dickinson, San Jose, CA). Data were acquired and processed using Becton Dickinson software (CellQuest). 

Histochemical Analyses

Histochemical analyses were performed as previously described with minor modifications. ^{34 35} First, antibodies against F4/80 (a marker for monocytes/microglia/macrophages; 1:500; AbD Serotec, Oxford, UK), ionized calcium-binding adaptor molecule 1 (Iba-1; a marker for monocytes/microglia/macrophages; 1:500; Wako, Tokyo), major histocompatibility complex (MHC) class II (a marker for active antigen-presenting cells; 1:100; BD PharMingen), or lysosome-associated membrane protein 2 (LAMP2; a marker for lysosomes; 1:100; Developmental Studies Hybridoma Bank, Iowa City, IA) were used. Samples were further stained with Alexa 488-, 568- and 594-conjugated antibodies (1:1500; Molecular Probes, Eugene, OR), diamino-2-phenyl-indol (DAPI; 1:1000; Molecular Probes), and tetramethylrhodamin isothiocyanate (TRITC)-labeled Griffonia simplicifolia lectin-1 (GS-1; a marker for vessels and monocytes/microglia/macrophages; 10 mg/L; Sigma-Aldrich, St. Louis, MO). The same concentrations of antibodies were applied to the sections and cilioretinal flatmount specimens. 

Because the same markers—F4/80, Iba-1, and GS-1–stained monocytes, microglia, and macrophages, these cells likely represent overlapping spectra of cells of the same lineage. However, from a practical view, these cells could best often be distinguished based on their morphology and location. To maintain consistency with regard to the naming of these cells throughout this article, we defined monocytes as round cells in the vessels or still attached to the vessel walls, microglia as ramified or ameboid cells in the neural retina and along the retinal vessels, and macrophages as ramified or ameboid cells in the nonneural tissue. 

MNU-Induced Retinal Degeneration

N-methyl-N-nitrosourea (MNU; Sigma-Aldrich), a direct-acting alkylating agent, is preferentially toxic for photoreceptors and causes severe retinal degeneration within approximately 7 days of administration. ³⁶ MNU was dissolved in saline solution containing 0.05% acetic acid and was administered intraperitoneally (60 mg/kg body weight) to the chimeric mice 4 to 5 weeks after BM transplantation. 

Experimental Retinal Detachment

After the chimeric mice were anesthetized with tribromoethanol (0.5 mg/g body weight; Sigma-Aldrich), 1% sodium hyaluronate was injected subretinally into the eyes with a 33-gauge needle 4 to 5 weeks after BM transplantation to induce partial retinal detachment (see 1 2 3 4 Fig. 5A ). The presence or absence of retinal detachment was confirmed by fundus examination before enucleation. Eyes with reattached retinas or incidental cataract were excluded from the study. 

Detection of Proliferating or Apoptotic Cells

To detect proliferating cells, we injected bromodeoxyuridine (BrdU; 150 mg/kg; Sigma-Aldrich) intraperitoneally 2 hours before kill and histologically analyzed the cells with an anti–BrdU antibody (1:500; Oxford Biotechnology, Killington, UK). To detect apoptotic cells in the outer nuclear layer (ONL) of the retina, eye sections were treated using an in situ cell detection kit (Roche, Indianapolis, IN) and the terminal-UDP nick end labeling (TUNEL) technique according to the manufacturer’s instructions. 

Quantification and Statistical Analysis

In Figures 1H and 2F , the numbers of EGFP⁺ cells within fields of 640 μm × 640 μm (the average of two fields from the dorsal and ventral peripheral retina) and 3200 μm × 3200 μm (a single image centered at the optic disc; Fig. 2G ), respectively, were determined and statistically processed from confocal images of cilioretinal flatmounts. In Figure 2F , the TUNEL⁺ cells in the ONL were counted, and an average count was obtained from three independent images per retinal section cut through the dorsal to the ventral meridian. In Figures 3E and 4H , the numbers of EGFP⁺, Iba-1⁺, or MHC class II⁺ cells within 640 μm × 640 μm square fields were determined and averaged from four independent confocal images per flatmount (anterior and posterior retina from dorsal and ventral hemispheres; Fig. 3F ). Cell counts from each flatmount specimen were normalized using the data of the FACS analysis from the peripheral leukocytes of the same animal before the analyses (Figs. 1H 2F 3E 4H) . All experimental data were evaluated by an operator blinded to the treatment conditions. Statistical significance was determined by a paired t-test (Fig. 4H) . 

Results

Spatial and Temporal Distribution of EGFP⁺ Cells in the Uninjured Retina

We studied the distribution of EGFP⁺ BM-derived cells within the eyes of adult wild-type mice that underwent successful BM transplantation. In these animals, the percentage of EGFP⁺ cells among peripheral leukocytes was 58.4% ± 13.0% (mean ± SD). Thirty days after BM transplantation, EGFP⁺ cells were identified in the nonneural tissues of the eye but not in the retina (n = 6; Figs. 1A 1B 1C 1D ). These cells engrafted to the ocular tissues surrounding the retina, such as the retinal pigment epithelial layer (Fig. 1C) , choroid (Fig. 1C) , and ciliary body (Fig. 1D) . EGFP⁺ BM-derived cells were also found at the head of the optic nerve (Fig. 1A) . However, analysis of the transverse section of the optic nerve revealed the presence of these cells around the subarachnoid space but not within the optic nerve (Fig. 1B) , indicating that these EGFP⁺ cells might have traveled from the ventricles of the brain through the cerebrospinal fluid that surrounds the optic nerve. By 3 to 12 months after BM transplantation, rare EGFP⁺ cells were detected within the retina at the juxtapapillary area (data not shown) and at the retinal margin adjacent to the cilioretinal border (Fig. 1E)in cilioretinal flatmounts. All had ramified processes and were immunopositive for Iba-1 and F4/80, consistent with BM-derived microglia (Figs. 1E 1F 1G) . These cells obviously constituted a minute proportion of the microglial population in the uninjured retina (Fig. 1F) . We further investigated the numbers of EGFP⁺ cells in the retina and ciliary body up to 1 year after BM transplantation (Fig. 1H) . A steady increase in the engraftment of EGFP⁺ cells was noted in the peripheral retina and the ciliary body from 3 to 9 months after BM transplantation. This was followed by a decrease in the number of EGFP⁺ cells in the retina at 12 months after BM transplantation. However, at all time points examined, the BM-derived cells occupied a negligible proportion of Iba-1⁺ microglia in the uninjured retina. 

Spatial and Temporal Distribution of EGFP⁺ Cells in the Degenerating Retina

In the brain, a large number of BM-derived cells are observed around the site of neural damage. ^{9 10} To investigate whether retinal damage also enhances the engraftment of BM-derived cells to the retina, we treated chimeric mice with MNU to induce photoreceptor degeneration and analyzed the cilioretinal flatmounts (Fig. 2) . In these animals, the percentage of EGFP⁺ cells among peripheral leukocytes was 77.4 ± 15.5% (mean ± SD). As expected, a large number of EGFP⁺ cells were found within the injured retina. As early as 12 hours after MNU injection, cilioretinal flatmounts showed that ramified EGFP⁺ cells were present within the retina but were restricted at the juxtapapillary (Fig. 2A)and the peripheral (Fig. 2B)retina. By 24 hours after MNU administration, in addition to the areas of the retina, many EGFP⁺ cells were found around the vascular structures (Fig. 2C) . Because the retinal vessels were perfused with a fixative (4% paraformaldehyde) to wash out blood cells before enucleation, those within the vascular structure were adherent cells probably about to migrate to the retinal parenchyma from the peripheral circulation. These EGFP⁺ cells concentrated around the retinal vessels were no longer obvious by 3 days after MNU injection, by which time the cells were scattered in most areas of the retina (Fig. 2D) . Twenty-eight days after MNU injection, EGFP⁺ cells were reduced in number but showed no gross alteration in distribution compared with 3 days after injection (Fig. 2E) . Quantification of the EGFP⁺ cells in the retina revealed that the infiltration of BM-derived cells was maximal 24 hours after MNU injection and decreased gradually thereafter (Fig. 2F) . We found a similar rapid increase in the number of TUNEL⁺ apoptotic cells after MNU administration. However, the number peaked at 3 days and rapidly decreased by 7 days after injection (Fig. 2F) . 

As expected from their characteristic ramified morphology, most of the EGFP⁺ BM-derived cells in the retina were microglia positive for Iba-1 (Figs. 3A 3B 3C) , GS-1 (Figs. 4A 4B 4C) , or F4/80 (data not shown). These cells were indistinguishable from EGFP-negative resident microglia in their morphology. Next, we determined the number of EGFP⁺/Iba-1⁺ cells and calculated the percentages of BM-derived cells among the total Iba-1⁺ retinal microglia. Results indicated that BM-derived EGFP⁺ microglia accounted for approximately 15.5% ± 6.1% (mean ± SD; Fig. 3E ) of the total number of Iba-1⁺ microglia and that the remaining approximately 84.1% ± 6.1% were resident microglia in the retina 7 days after MNU injection. 

Characteristics of BM-Derived Cells in the Degenerating Retina

Twenty-four hours after MNU administration, round EGFP⁺ cells positively stained for the microglia/macrophage marker GS-1 lectin adhered to the retinal vascular lumens (Figs. 4A 4B 4C) . Because such small round cells, probably monocytes, were less frequent in the adjacent retina, these cells likely transformed to ramified cells before or immediately after they entered the retina. We also found that though many adherent cells were proliferating cells that incorporated BrdU, those in the retina became less mitotic once they entered the retina (Figs. 4D 4E 4F) . 

In the adult uninjured retina, a small number of MHC class II⁺ microglia, identified as immunologically active antigen-presenting cells, existed only at the juxtapapillary (2.2 ± 1.5 cells/eye; mean ± SD; n = 6) and peripheral (6.7 ± 4.8 cells/eye) areas of the retina. On the other hand, many immunopositive cells existed in the ciliary body. However, the number of MHC class II⁺ cells in the retina increased greatly after MNU injection (Fig. 4G) . We found that the proportion of microglia immunopositive for MHC class II was approximately 2.6-fold greater among the Iba-1⁺/EGFP⁺ BM-derived microglia (17.5% ± 2.4%; mean ± SEM) compared with Iba-1⁺/EGFP⁻ resident microglia (6.9% ± 1.8%) 7 days after MNU injection (Fig. 4H) . The presence of LAMP2⁺ granular structures in the cytoplasm of many EGFP⁺ cells suggested their role as phagocytes (Fig. 4I) . However, no direct relationship between the in situ distributions of the TUNEL⁺ and EGFP⁺ cells was noted in the histologic sections (Supplementary Fig. S1). 

Recruitment of BM-Derived Cells to the Site of Retinal Detachment

To compare the distribution of BM-derived cells in the injured and uninjured retina within the same eye, we created a small retinal detachment that occupied a restricted area of the retina in chimeric mice (Fig. 5A) . In the histologic sections, EGFP⁺ cells were identified only in and under the detached retina, whereas no such cells were present in the intact retina (Figs. 5B 5C)28 days after the induction of retinal detachment. Within the detached retina, these cells resided mostly in the inner plexiform layer or around the ganglion cell layer and were less frequent in the ONL, where progressive photoreceptor degeneration took place. ³⁷ Virtually all these cells were positive for the microglial marker F4/80. Although EGFP⁺ cells in the retina had a complex and variable morphology consistent with that of microglia, those in the subretinal space were round and large. These cells resembled phagocytic macrophages seen in human and rodent retinal abnormalities. ^{25 38 39} Analysis of the cilioretinal flatmounts revealed that these EGFP⁺ microglia in the retina were found primarily around the vessels, similar to the observation in the early stages of MNU-induced retinal degeneration (Figs. 5D 5E) . 

Discussion

In this study, we aimed to elucidate the nature and roles of circulating BM-derived cells in the retina. Our results showed that there was only a small amount of BM-derived cells in the uninjured normal retina, whereas a large number of them existed in the adjacent ciliary body and around the optic nerve at least up to 12 months after BM transplantation. It was only after the induction of retinal damage that a bulk of BM-derived cells rapidly migrated to the retina. This was in contrast to a previous report that showed the engraftment of numerous BM-derived cells to the uninjured retina of the same C57BL/6J mice 6 months after BM transplantation. ¹⁸ Although exact comparison of the data is difficult because no statistical analysis was provided with regard to the numbers or the percentages of EGFP⁺ and resident microglia, the number of EGFP⁺ microglia in the uninjured retina was clearly smaller in our study than in the previous report. ¹⁸ The obvious difference observed could be partially attributed to differences in the experimental designs. For example, in the previous study, irradiation was conducted without shielding of the eyes or heads, whereas we used lead shields to protect eyes and heads from radiation-induced damage. The amount of radiation used (8–10 Gy) in the previous study could have caused unintentional damage to neurons, including photoreceptors ⁴⁰ and vessels, which might have enhanced the migration of BM-derived cells to the injured retina. 

The presence of BM-derived cells in the ciliary body and the optic nerve has been reported by other groups, ^{17 18} but the potential roles of the cells in the retina are unclear. In the present study, we also found a large number of BM-derived tissue macrophages immunopositive for monocyte/macrophage markers, some with a ramified microglia-like morphology in the ciliary body and around the optic nerve, 30 days after BM transplantation. On the other hand, BM-derived cells were rarely present in the adjacent uninjured retina, though there is no known anatomic barrier between the ciliary body and the peripheral retina. This indicated the presence of factors other than the blood-retinal barrier ^{16 27 41} that prevented retinal infiltration by BM-derived cells in the uninjured eye. However, the effect of this cue could have been partially reversed by the induction of retinal damage resulting in the rapid retinal migration of these cells from the ciliary body and optic nerve as early as 12 hours after MNU injection. This was followed by the adhesion of numerous BM-derived monocytes to the retinal vascular lumen, where they differentiated into microglia and entered the retina. Meanwhile, we could not detect clear evidence of the retinal migration of BM-derived cells from the retinal pigment epithelial layer in cilioretinal flatmounts. Nonetheless, our results indicated that BM-derived tissue macrophages in the nonneural ocular tissues act as the mediators of the earliest phase of retinal damage and the reservoir for retinal microglia. In the eyes with partial retinal detachment, BM-derived cells were homed only to the detached retina. Their high spatial specificity to the injury site suggested the contribution of an in situ factor in the attraction of these cells. At the same time, we found large and round BM-derived cells, likely phagocytic macrophages, ²⁵ under but not within the detached retina. Such large, “plump” morphology of macrophages likely represents the accumulation of digestive compounds within the phagosomes or phagolysosomes as the result of active ingestion of tissue debris. The obvious difference in morphology compared with BM-derived microglia in the retina and the presence of such rare cells in the photoreceptor layer adjacent to the subretinal space implied that the subretinal BM-derived cells might have migrated through the retinal pigment epithelium from the choroidal circulation rather than from the neural retina. ²⁵  

The roles of BM-derived microglia in the homeostasis of the normal retina are unclear. No definitive morphologic distinction could be made between the BM-derived versus resident microglia. However, our results suggest the specific roles of the former in retinal damage. We found that the engraftment of BM-derived microglia in the retina takes place almost only under retinal damage. These cells occupied approximately 15% of the total number of retinal microglia by 7 days after the induction of damage, indicating their significant contribution to the injury-induced microglial response. BM-derived microglia were more prone to immunologic activities, such as antigen presentation, than were resident microglia. A similar finding has been reported in the brain. ⁸ Moreover, we found that BM-derived microglia were positive for a lysosome marker, ⁴² an observation consistent with their involvement in the phagocytosis of tissue debris. Although no clear spatial association with these cells and apoptotic cells was detected with the method used in this study, the result was inconclusive because we did see some apoptotic cells that appeared to be in contact with microglia. Our difficulty in finding such an association could be partially attributed to the limited time microglia physically reside by the apoptotic cells before they ingest them. 

In conclusion, the lack of engraftment of BM-derived cells suggests a minor role of these cells in the maintenance of the uninjured retina. In response to retinal damage, however, these cells migrated to the retina by way of the ciliary body, optic nerve head, and retinal vessels, most of which differentiated into microglia. The higher rate of immunologic activation and the increased specificity to the injury site appear to be the features of BM-derived microglia. Our findings provide an important base for the design of an appropriate gene delivery system using exogenous microglia/macrophages to target the damaged retina. 

 Submitted for publication January 15, 2008; revised April 16, 2008; accepted July 7, 2008.

 Disclosure: H. Kaneko, None; K.M. Nishiguchi, None; M. Nakamura, None; S. Kachi, None; H. Terasaki, 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.

 

Figure 1.

 

Distribution of BM-derived cells in normal mouse eye. (A–D) Histologic sections from uninjured eyes 30 days after BM transplantation. (A) EGFP⁺ cells (filled arrowheads) and EGFP⁻/F4/80⁺ cells (open arrowheads) were present at the optic nerve. (B) A transverse section of the optic nerve showed that EGFP⁺ cells (filled arrowheads) resided only around the optic nerve, whereas EGFP⁻/F4/80⁺ resident cells (open arrowheads) also existed within the optic nerve. (C) Histologic section showing EGFP⁺ cells in the retinal pigment epithelial layer (arrowheads) and choroid (arrows) but not in the retina. (D) EGFP⁺ cells (filled arrowheads) in the ciliary body. (E–G) Cilioretinal flatmount showing BM-derived microglia positive for EGFP (E) and Iba-1 (F) at the retinal margin (arrowheads) 12 months after BM transplantation. (G) Merged image of (E) and (F). (H) Number of EGFP⁺ cells (mean ± SD) in the peripheral retina and ciliary body 1 to 12 months after BM transplantation. Cell counts from each flatmount specimen were normalized with the data of the FACS analysis from the peripheral leukocytes of the same animal before analyses. The largest number of cells was identified at 9 months after BM transplantation. (I) Illustration of the areas in the eye from which images (A) through (D) were obtained. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; IS/OS, inner/outer segment; CB, ciliary body. Scale bar, 100 μm.

Figure 2.

 

Spatial and temporal distribution of EGFP⁺ cells in the degenerating retina. (A, B) Cilioretinal flatmount 12 hours after MNU injection showed the migration of EGFP⁺ cells to the juxtapapillary (A; arrows) and the far peripheral retina (B; arrows). Note that the EGFP⁺ cells were also found in the ciliary body (arrowheads). (C) By 24 hours after injection, a large number of EGFP⁺ cells were found along the retinal vessels (arrows). (D) EGFP⁺ cells were distributed throughout the retina by 3 days after MNU administration. Note that EGFP⁺ cells were no longer concentrated around the vessels. (E) Distribution of EGFP⁺ cells 28 days after the injection. (F) Numbers of EGFP⁺ cells in the cilioretinal flatmounts (mean ± SD) and TUNEL⁺ cells in histologic the sections (mean ± SD) were determined before (n = 6) and 12 hours (n = 6), 24 hours (n = 7), 3 days (n = 6), 7 days (n = 7), and 28 days (n = 6) after MNU injection. Cell counts from each flatmount specimen were normalized with the data of the FACS analysis from the peripheral leukocytes of the same animal before analyses. The largest numbers of EGFP⁺ and TUNEL⁺ positive cells were identified at 24 hours and 3 days after the injection, respectively. (G) Scheme of a cilioretinal flatmount showing the area from which the EGFP⁺ cells were counted for (F). OD, optic disc. CB, ciliary body. Scale bar: (A) 100 μm; (B) 200 μm; (C–E) 300 μm.

Figure 3.

 

Quantification of microglia positive for EGFP and Iba-1. (A–C) Most EGFP⁺ cells (A) were also positive for Iba-1 (B) 7 days after MNU administration. (C) Merged image of (A) and (B). (D) Example of FACS analysis of the peripheral leukocytes from a mouse with chimeric BM. (E) Estimated percentage of BM-derived cells among total Iba-1⁺ microglia. Horizontal bar indicates the mean. Numbers of EGFP⁺ and EGFP⁻ microglia counted from each flatmount specimen were normalized with the data of the FACS analysis from the peripheral leukocytes of the same animal before analysis. (F) Scheme showing the retinal areas (640 μm × 640 μm square fields; anterior and posterior retina from dorsal and ventral hemispheres) from which data related to (E) were obtained. CB, ciliary body. Scale bar, 100 μm.

Figure 4.

 

Characteristics of BM-derived cells in the degenerating retina. (A–F) Analyses of cilioretinal flatmounts 24 hours after MNU administration. Many EGFP⁺ cells (A) found within the vessels were small, round, and positive for GS-1 (B) in the degenerating retina 24 hours after MNU administration. EGFP⁺ cells (D) within the retinal vessels were more likely to be positive for BrdU than those outside the vascular structures (F). (C) is a merged image of (A) and (B), and (F) is a merged image of (D) and (E). (G) Cilioretinal flatmount stained with antibodies for Iba-1 and MHC class II 7 days after MNU administration. (H) Increased percentage of MHC class II⁺ cells was found among BM-derived (EGFP⁺/Iba-1⁺) compared with resident (EGFP⁻/Iba-1⁺) microglia in the retina. Cell counts from each flatmount specimen were normalized with the data of the FACS analysis from the peripheral leukocytes of the same animal before analyses. (I) Some EGFP⁺ cells were positive for the lysosome marker LAMP2. Scale bar: (A, D, G) 100 μm; (I) 10 μm. *P < 0.05.

Figure 5.

 

Recruitment of BM-derived cells to the site of retinal detachment. (A) Scheme illustrating the generation of partial retinal detachment by subretinal injection of 1% sodium hyaluronate with 33-gauge needle. (B, C) EGFP⁺ cells were found within the detached retina, mainly in the GCL and IPL but also in the subretinal space 28 days after the induction of retinal detachment. Note that those in the retina have a ramified morphology with immunopositivity for F4/80. These features of EGFP⁺ cells are consistent with those of microglia. Meanwhile, EGFP⁺ cells in the subretina are strongly positive for F4/80 and are morphologically large, round, and plump, resembling phagocytic macrophages. (B) Magnified image of (C). (D, E) Cilioretinal flatmount 28 days after the induction of retinal detachment showed the EGFP⁺ microglia found along the retinal vessels in the detached retina; this was similar to the observation in the early stages of MNU-induced retinal degeneration. (E) Magnified image of (D). Scale bar: (B) 50 μm; (C–E) 100 μm.

Supplementary Materials

Supplementary Figure S1 - (PDF) 

The authors thank the BioResource Center (RIKEN Tsukuba Research Institute) and Masaru Okabe (University of Osaka) for providing the EGFP mice, and the Developmental Studies Hybridoma Bank for providing the anti-LAMP2 antibody used in this study.