Bone Marrow-Derived Progenitor Cells Contribute to Experimental Choroidal Neovascularization

Abstract

Purpose. The pathogenesis of choroidal neovascularization (CNV) is postulated to be driven by angiogenesis, a process in which the cellular components of the new vessel complex are derived from cells resident within an adjacent preexisting capillary. Recently, an alternative paradigm, termed postnatal vasculogenesis, has been shown to contribute to some forms of neovascularization. In vasculogenesis, the cellular components of the new vessel complex are derived from circulating vascular progenitors from bone marrow. In the current study, transplantation of green fluorescent protein (GFP)-labeled bone marrow and laser-induced CNV were combined to examine the contribution of vasculogenesis to the formation of CNV.

Methods. Ten adult C57BL/6 female mice were used as recipients for bone marrow transplantation. Bone marrow was obtained from three C57BL/6 female mice transgenic for the β-actin promoter GFP. One month after bone marrow transplantation, CNV was induced in recipient mice by making four separate burns in the choroid of each eye with a red diode laser. Four weeks after CNV was induced, eyes of recipient mice were processed for immunohistochemistry to detect GFP and markers for vascular smooth muscle cells (α-smooth muscle actin, desmin, and NG2 chondroitin sulfate proteoglycan), endothelial cells (CD31, BS-1 lectin), or macrophages (F4/80).

Results. GFP-labeled cells represented 17% of the total cell population in the lesion. Many of the GFP-labeled cells were immunoreactive for α-smooth muscle actin (39%), desmin, NG2, CD31 (41%), BS-1 lectin, or F4/80. GFP-labeled cells were morphologically indistinguishable from cells normally present in CNV lesions.

Conclusions. This study is the first to demonstrate that bone marrow-derived progenitor cells are a source of endothelial and smooth musclelike cells in CNV.

Age-related macular degeneration (AMD) is the most important cause of vision loss in elderly patients. The major cause of severe vision loss in AMD is pathologic neovascularization under the retina that produces a lesion called choroidal neovascularization (CNV). ¹ The pathogenesis of CNV is clearly multifactorial, but is generally considered to be driven by angiogenesis. In angiogenesis, the cellular components of the new vessel complex (endothelial cells, smooth muscle cells and other types) are derived from cells resident within adjacent preexisting capillaries. Recently, an alternative paradigm, termed postnatal vasculogenesis, has been shown to contribute to some forms of neovascularization. ^{2 3 4} In vasculogenesis, the cellular components of the new vessel complex are derived in part from bone marrow-derived circulating vascular progenitors, which differentiate into mature endothelial cells or vascular smooth muscle cells in situ. The role of bone marrow-derived cells has been studied during development of ocular vasculature and retinal neovascularization. ^{5 6} The contribution of vasculogenesis to formation of CNV is unknown. 

Conceptually, bone marrow has been shown to contain progenitor cells that demonstrate the capacity to enter the circulation, home into peripheral tissues, and differentiate into parenchymal tissues such as liver, heart, blood vessels, pancreas, muscle, and even neurons. ^{2 3 4 6 7 8 9 10 11} Evidence from several laboratories indicates that many cells in neovascularization, or in other vascular reparative responses after injury, are derived from circulating bone marrow progenitors. ^{2 3 4 6 7 8} Many important questions remain unresolved, including the cell type of origination and physiological significance of these progenitors. For instance, the phenomena reported in studies of bone marrow transplantation may not be true vasculogenesis, but may reflect the recruitment of already committed endothelial and vascular smooth muscle progenitors into proliferating vessels. Also, transdifferentiation of stem cells may reflect somatic cell fusion. 

Operationally, the existence of tissue progenitors is demonstrated by bone marrow transplantation. If an animal is lethally irradiated to destroy its endogenous bone marrow and then receives a bone marrow transplant from a donor containing a ubiquitously expressed marker protein (e.g., GFP), the recipient wild-type animal now has bone marrow reconstituted with fluorescently GFP-labeled donor stem and progenitor cells. If vascular injury is induced 4 to 8 weeks after bone marrow transplantation, the relative contribution of GFP-labeled cells recruited from the bone marrow versus unlabeled cells derived from cells preexisting in the vessel can be determined. 

Using bone marrow transplantation with GFP-labeled progenitors in various models of vascular injury or neovascularization, studies have demonstrated that many endothelial cells and vascular smooth muscle cells are derived from bone marrow. ^{2 3 4 6 7 8} Thus, vascular precursor cells from the bone marrow must circulate and home into the vessel. There, they differentiate into endothelial cells and vascular smooth muscle cells in situ, participating in both normal and pathologic repair responses. 

In the present study, we used bone marrow transplantation of GFP-labeled bone marrow combined with laser-induced CNV to demonstrate that CNV contains abundant GFP labeled endothelial cells and vascular smooth muscle cells. Our results demonstrate that the paradigm of vasculogenesis contributes to the pathophysiology of CNV. 

Material and Methods

Experimental Animals

Mice used in this study were handled in accordance with ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Ten adult C57BL/6 female mice (National Institute of Aging, Bethesda, MD) were used as recipient mice for bone marrow transplantation. Bone marrow was obtained from three C57BL/6 female mice transgenic for the chicken β-actin promoter-GFP and cytomegalovirus enhancer (stock no. 003291; The Jackson Laboratory, Bar Harbor, ME). The femur and tibia were dissected and placed in RPMI 1640 culture medium (Invitrogen-Gibco, Grand Island, NY) containing 2.5% HEPES (1 M) and 1% gentamicin at 4°C. Bone marrow was obtained by slowly flushing medium inside the diaphyseal channel with a syringe through a 27-gauge needle. Bone marrow was homogenized through an 18-gauge needle and filtrated with a nylon filter (70 μm; Spectrum, Houston, TX). Bone marrow donor cells were centrifuged and the pellets resuspended in the medium described earlier. Recipient mice were lethally irradiated (950 cGy) and given 10⁷ nonpurified bone marrow cells intravenously (200 μL). Blood components were allowed to reconstitute for 1 month. The survival rate of mice transplanted with exogenous bone marrow was 100%. By contrast, irradiated mice without exogenous bone marrow died approximately 10 to 14 days after irradiation. One month after bone marrow transplantation, CNV was induced in recipient mice by making four separate choroidal burns in each eye with a red diode laser. ¹² Four weeks later, animals were killed and the eyes removed for experimental analysis. Chimerism of this bone marrow transplantation protocol is approximately 90% (Sen Li, MD, personal communication, February 2003). 

A group of mice (n = 5) was used for visualizing the distribution of GFP-labeled cells in flatmount preparations of the posterior pole of the eye. We removed the anterior segment and the neurosensory retina and made four radial relaxing incisions in the remaining sclera-choroid-RPE complex. The RPE was removed with microsponges (Alcon Laboratories, Fort Worth, TX). Cell nuclei were stained with a mixture of 1 mg/mL digitonin (Sigma-Aldrich, St. Louis, MO) and 0.5 mg/mL propidium iodide (Sigma-Aldrich). GFP and propidium iodide labeling were examined by fluorescence microscope (Axiophot; Carl Zeiss Meditec, Oberkochen, Germany) as described later. As a control, experimental CNV was also induced in age-matched mice (n = 5) that did not receive GFP bone marrow transplantation. 

Immunohistochemistry

Four weeks after CNV was induced, mice (n = 5) were perfused transcardially with phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (10 minutes). Mouse eyes were removed and postfixed for 1 hour. After cryoprotection (sucrose, 10%, 20%, and 30% in PBS), vertical sections of the eyes were cut on a cryostat (14 μm). Sections were washed in PBS (three times, 10 minutes each) and incubated in PBS containing 5% bovine serum albumin and 0.1% Triton (1 hour). Thereafter, sections were incubated overnight in PBS with anti-GFP antibodies conjugated to Alexa 488 (1:1000; Molecular Probes, Eugene, OR) and anti-α smooth muscle actin (αSMA; 1: 1000; Sigma-Aldrich), anti-desmin (1:500; Abcam, Cambridge, UK), anti-NG-2 (1:500; Chemicon, Temecula, CA), anti-CD31 (1:500; Pharmingen, San Diego, CA), or anti-mouse F4/80 (1:500; Serotec, Raleigh, NC). Immunostaining other than for GFP was visualized using Alexa 568-conjugated secondary antibodies (1:500; Molecular Probes). For detection of endothelial cells, sections were also incubated in biotinylated Griffonia simplicifolia lectin (BS-1; 1:1000; overnight; Sigma-Aldrich) followed by Cy3-conjugated extravidin (1:500; Sigma-Aldrich). Cell nuclei were stained with 4′,6′-diamino-2-phenylindole (DAPI; Molecular Probes). Slides were mounted with aqueous mounting medium (Crystal Mount; Biomeda, Foster City, CA) and coverslipped. 

Several results support the specificity of our immunohistochemical approach. First, we skipped incubation with primary antibodies to show that secondary antibodies did not produce any background staining (not shown). Only secondary antibodies against mouse IgGs produced some unspecific staining in the sclera. However, this did not interfere with our immunohistochemical analyses of the choroid. Second, the immunostaining patterns of the primary antibodies were in agreement with the distribution of the respective cell types in the normal eye (e.g., αSMA was present in large vessels in the choroid, CD-31 in retinal blood vessels, and F4/80 in retinal microglia cells). Third, we used several different markers for a particular cell type (e.g., αSMA, desmin, and NG2 for smooth muscle cells; CD-31 and BS-1 for endothelial cells) and they produced very similar staining patterns and incidences in CNV lesions and elsewhere in the eye. 

Data Analysis

Serial cross sections of eyes containing CNV lesions were examined for the presence of GFP-labeled cells and their expression of the different cell markers, with a fluorescence microscope (Axiophot; Carl Zeiss Meditec) and a dual-channel laser scanning confocal microscope (Fluoview; Olympus America Inc., Melville, NY). All images were digitally acquired (Axiovision; Zeiss or Fluoview; Olympus) and recompiled (Photoshop ver. 5.0; Adobe, San Jose, CA). Sections were viewed at 40× and 100× magnification. Filters on the fluorescence microscope for green fluorescence were band-pass (BP) 485 nm (excitation), 510 nm (dichroic mirror), and 515 to 565 nm (emission), for red fluorescence: BP 546 nm (excitation), 580 nm (dichroic mirror), and 575 to 640 nm (emission). 

To document clearly that cell markers were expressed in GFP-labeled cells within CNV lesions, dual-fluorescence scanning confocal microscopy was used. Antibody concentrations and confocal microscope settings were adjusted to avoid fluorescence interference in the colocalization studies. Fluorescence images for the two fluorochromes were acquired sequentially with their specific settings and merged subsequently. For every section examined we determined that no red fluorescence (605 nm emission high pass filter) could be detected using the settings for green fluorescence (i.e., 488 nm excitation with the argon laser) and, conversely, that no green fluorescence (510–550-nm emission band-pass filter) could be detected using the settings for red fluorescence (i.e., 568-nm excitation with the krypton laser). Thin optical sections (1 μm) were examined to ensure that overlapping immunofluorescence was due to colocalization. Images shown are representative. 

For each eye, the anterior and posterior segments were surveyed for accumulation of GFP-labeled cells. Quantitative assessment of GFP cell number was restricted to CNV lesion. The percentage of GFP-positive cells per lesion was calculated as the number of GFP-positive cells divided by the total number of DAPI positive cells × 100. We further determined the proportion of GFP-labeled cells within a lesion that were also immunoreactive for αSMA, desmin, NG2, or CD31 or were labeled with BS-1. Only cells that had a clearly labeled nucleus (DAPI staining) were included in this analysis. Cell numbers for a particular eye were calculated as the average from at least three adjacent sections from at least two separate CNV lesions per eye. The results from five eyes were averaged. Data are presented as the mean ± SEM. Analyses were performed on digitized fluorescence microscopic images (Axiovision software; Carl Zeiss Meditec). We examined the confocal microscopic images with another image-analysis program (Fluoview software; Olympus) and obtained similar results. 

Results

Irradiation did not affect the morphology of CNV and did not cause cataract. In mice that underwent bone marrow transplantation, irradiation did not induce recruitment of progenitors to the normal choroid, suggesting that the ocular dose was low. Furthermore, we observed that the CNV area in bone marrow transplant recipients and control mice was the same (Fig. 1) , indicating that irradiation and bone marrow transplantation did not affect CNV’s severity and morphology. CNV lesions in bone marrow-recipient mice were morphologically indistinguishable from the standard model (discussed later). 

Mice that underwent bone marrow transplantation were examined for the presence of GFP-labeled cells in the normal choroid. Bone marrow cells were not recruited to the normal choroid (Fig. 2) . Nevertheless, some green cells were present in the choroid, and because they immunostained for F4/80, they most probably represented circulating macrophages. We estimated that four to eight GFP-labeled cells were visible in a 1-mm linear portion of choroid in our vertical sections. 

In contrast to the normal choroid, GFP-labeled cells derived from bone marrow were abundant in laser-induced CNV lesions (Figs. 3 4) . In choroidal flatmounts, GFP-positive cells were present within and around the edge of CNV and within the optic nerve head, but were sparse in the normal choroid distant from the CNV (Figs. 3A 3B 3C) . In transverse sections, lesions from bone marrow-recipient mice were indistinguishable from those of control mice (Fig. 3D) . GFP-labeled cells were visible throughout the lesion, surrounding vascular lumens or closely apposed to the retina (Figs. 3E 4) . Most GFP-labeled cells had an elongated shape and were oriented parallel to the retinal surface (Figs. 3E 4C 4D 5) . Many GFP-labeled cells appeared well integrated into the neovascular tissue (described later). In addition, GFP-labeled cells were observed to infiltrate the overlying neurosensory retina, and within the deep layers of the choroid beneath the CNV lesion (Figs. 4A 4B) . We estimated that GFP-labeled cells represented 17% of the total number of cells within the lesion. 

Although the precise cellular composition of human or rodent CNV has not been published, CNV lesions are believed to consist mainly of endothelial and vascular smooth muscle cells. ^{1 13 14} To determine whether GFP-labeled cells had differentiated into these mesenchymal cell types, we colocalized GFP with markers for endothelial cells (BS-1, CD31), or vascular smooth muscle cells (αSMA, desmin, NG2) using thin optical sections (<1 μm) of immunostained tissue with laser scanning confocal microscopy. We found that BS-1, CD31, and αSMA colocalized with GFP-labeled cells in the lesion (Figs. 5 6) . Desmin and NG-2 were also expressed by GFP-labeled cells in the lesion (not shown). These results indicate that bone marrow-derived progenitor cells had acquired phenotypic markers of mature endothelial and vascular smooth muscle cells. We calculated that 41% ± 7% (mean ± SEM) of GFP-labeled cells expressed CD31 and 39% ± 4% expressed αSMA. BS-1 lectin, desmin, and NG2 were expressed in similar proportions (in ∼ 40% of the GFP-labeled cells). In contrast, no bone marrow-derived endothelial cells or vascular smooth muscle cells were observed in normal choroid, suggesting that vascular injury is necessary for induction of vasculogenesis. 

Previous publications have also indicated that macrophages infiltrate into human and experimental CNV, but their identity as blood-derived or tissue-resident macrophages has not been well described. Using colocalization of the monocytic marker F4/80 with GFP (Figs. 5J 5K 5L) , we confirmed that the majority (>90%) of monocytes associated with the CNV were GFP labeled, indicating that they had recently been recruited into the lesion from circulation (i.e., blood-derived macrophages). In general, macrophages were predominant in the choroid under the lesion and in the neurosensory retina over the lesion, although few infiltrated into the central CNV itself (Fig. 7A) . 

GFP-labeled cells were also present in other regions of the eye. A few GFP-labeled cells were present in normal peripheral choroid distant from the CNV, but almost none were seen in the normal retina (Figs. 7B) . Also, some GFP cells were present in the sclera (Figs. 7B 7E) . In the ciliary body and the limbus, however, we found a high density of GFP-labeled cells (Figs. 7C 7F) . In contrast to the vascular markers that colocalized with GFP within the CNV, almost all GFP-labeled cells in regions other than the CNV (e.g., retina, sclera, and limbus) were immunoreactive for F4/80 (Figs. 2 7) , indicating that most recently recruited bone marrow-derived cells in noninjured normal uvea and sclera represented dendritic cells or macrophages. GFP-labeled mononuclear cells represented approximately 50% of the total population of F4/80 mononuclear cells in the limbus, and approximately 25% of the F4/80-positive cells in the ciliary body and sclera. 

Discussion

Bone marrow-derived progenitor cells differentiating into vascular smooth muscle cells and endothelial cells have been demonstrated to contribute to different models of neovascularization or vascular injury (e.g., tumor vascularization, aortic balloon injury, renal glomerulosclerosis, corneal and retinal neovascularization). ^{2 3 4 6 7 8} Our results showed that bone marrow-derived cells contributed approximately 20% of the total cell population in laser-induced CNV. These cells were morphologically indistinguishable from unlabeled resident cells normally present in CNV lesions. Approximately 40% of these bone marrow-derived cells expressed vascular smooth muscle cell markers and 40% expressed endothelial cell markers. Taken together, these findings support our hypothesis that bone marrow-derived progenitors participate in CNV formation. 

Our results also confirm that blood-derived monocytes are recruited into the choroid under CNV and into the neurosensory retina overlying CNV. Because most macrophages were GFP labeled, blood-derived macrophages contributed more than local tissue-resident macrophages. Blood-derived macrophages were present even 1 month after induction of CNV. These findings are consistent with our recent work that shows that depletion of blood-derived macrophages lessens the severity of laser-induced CNV. ¹³ Blood-derived macrophages, or possibly dendritic cells, were also observed in the limbus, ciliary body, normal choroid, and sclera, suggesting a high turnover and recruitment rate of monocyte infiltration into normal limbus, ciliary body, and sclera. 

In human AMD, the cellular composition of CNV has not been well defined. Clearly, endothelial cells are essential in the formation of the lumens of perfused new capillaries. However, CNV are fibrovascular lesions that also consist of vascular smooth muscle cells, poorly differentiated myofibroblastoid cells, proliferating retinal pigment epithelium cells, and inflammatory cells. ^{1 14 15 16} Although laser-induced CNV is not identical with CNV in AMD, this model shares enough biological similarity that preclinical studies of pathogenesis and drug inhibition predict human responses. Thus, we believe that the contribution of bone marrow-derived progenitors to both the endothelial cells and the vascular smooth muscle cells observed in this model is likely to be physiologically relevant. Whether bone marrow-derived progenitor cells participate in CNV in humans remains unknown. Our study constitutes a first step toward understanding what role these cells play in the pathogenesis of CNV. 

Several important technical and conceptual issues deserve consideration in the interpretation of these results. Our experimental design did not evaluate the degree of reconstitution of bone marrow with GFP-labeled cells. Incomplete reconstitution resulting in chimeric bone marrow may vary from 50% to 99% repopulation with transplanted stem cells (Csaky K, personal communication, April 2003), although our bone marrow transplantation technique resulted in more than 90% reconstitution in preliminary data (Sen L, unpublished observation, 2003). Thus, 17% of GFP-labeled cells in the CNV should be considered a minimal estimate. In the ongoing work in which only mice with a documented repopulation of more than 90% were used, data indicate that the percentage of bone marrow-derived endothelial cells in CNV may be even greater than observed in this study (Csaky K, personal communication). Also, mice required irradiation to permit acceptance of the transplanted marrow, and the potential artifact of radiation damage remains unknown. 

Our study did not address the origination and identity of the vascular progenitor cells within the bone marrow, a very controversial issue. At least three hypotheses have been proposed to explain the origin of vascular progenitors: (1) They are derived from a pluripotent stem cell in the bone marrow that maintains plasticity to differentiate into lymphoid, myeloid, mesenchymal, and epithelial precursors; (2) they are derived from a self-renewing but dedicated vascular progenitor within the bone marrow; (3) they are derived from a dedicated progenitor that resides in peripheral tissues, but recirculates to transit through the bone marrow during its lifespan. Addressing these alternatives was beyond the scope of this study, but preliminary data suggest that lineage-negative hematopoietic cells (i.e., a nonlymphoid, nonmyeloid bone marrow subset) can transfer a population of progenitors that become endothelial cells within CNV. 

The physiological importance of postnatal vasculogenesis remains unclear, because no research has demonstrated that the recruited cells or the preexisting resident cells respond differently to angiogenic stimuli within the injured blood vessel. For example, a growing metastasis that requires a new vessel to provide nutrient support cannot distinguish whether the new vessel is derived from mature cells with the preexisting adjacent capillary or from recruited progenitor cells as long as the two populations respond similarly to the neovascular signals. Some laboratories have proposed that bone marrow progenitors may be a source of cells to promote regeneration of damaged vessels in ischemic tissue. ^{17 18}  

However, it is possible that progenitors may also contribute to pathologic responses. Recently, we have shown that much more severe CNV develops in old than in young mice. In preliminary data, we have shown that bone marrow transplantation of old bone marrow can transfer the age-related susceptibility to severe CNV into young recipients (Espinosa-Heidmann, et al. IOVS 2003;44:ARVO E-Abstract 3936). Ongoing work in the laboratory is attempting to determine whether aged vascular progenitors may have acquired abnormalities in function that cause them to contribute to age-related abnormalities in vascular responses to injury. 

²  Contributed equally to the work and therefore should be considered equivalent senior authors.

 Supported by National Eye Institute Grants EY/AI 13318 (SWC) and DC04525-01 (AC).

 Submitted for publication April 11, 2003; revised June 20 and August 11, 2003; accepted August 12, 2003.

 Disclosure: D.G. Espinosa-Heidmann, None; A. Caicedo, None; E.P. Hernandez, None; K.G. Csaky, None; S.W. Cousins, 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.

 

Quantitative analysis of flatmount specimens shows that the size of CNV lesions in the bone marrow transplant-recipient mice (BMT) in the study was not significantly different from that in the age-matched mice (control).

Figure 2.

 

GFP-labeled cells derived from bone marrow were not recruited to the normal, unlesioned choroid. Fluorescence images of vertical sections of the eye show that some GFP-labeled cells present in the choroid and the sclera (A, arrows) were labeled for the macrophage marker F4/80 (red). Colocalization appears yellow. A GFP-labeled cell present in the choriocapillaris (B) was immunoreactive for F4/80 (C). A superimposed transmitted light image shows the localization of this cell (D). SC, sclera; CH, choroid; ONL, outer nuclear layer; RPE, retinal pigment epithelium. Scale bars, 100 μm.

Figure 3.

 

GFP-labeled cells derived from bone marrow are shown invading a CNV lesion. (A–C) Flatmount preparation of the posterior pole of a mouse eye showing migration of GFP-labeled cells into a CNV lesion. The CNV lesion had a high density of cells, as indicated by propidium iodide staining (A, red). GFP-labeled cells (B, green) accumulated within the lesion (yellow in C), but were sparse in other regions. GFP-labeled cells were also present in the optic disc (A, ). (D) Histopathology section of a CNV lesion. Dotted line demarcates the CNV lesion. (E) Vertical section of a CNV lesion showing immigration of GFP-labeled cells. GFP-labeled cells can be seen surrounding a vascular lumen. Scale bars: (A–C) 200 μm; (D) 100 μm; (E) 20 μm.

Figure 4.

 

GFP-labeled cells migrated into the CNV lesion and into the adjacent retina and choroid. (A, B) Confocal fluorescence images of vertical sections of the eye showing accumulation of GFP-labeled cells in a CNV lesion. (B, D) Transmitted light images on which fluorescence images (A, C, respectively) were superimposed. (C, D) GFP-labeled cells had elongated shapes and were closely apposed to the outer nuclear layer (ONL) of the retina. Images in (C) and (D) are higher magnification views of (A) and (B) (rectangle in B). Scale bars: (A, B) 100 μm; (C, D) 20 μm._art;1>

Figure 5.

 

GFP-labeled cells derived from bone marrow expressed markers for vascular smooth muscle cells (A–C), endothelial cells (D–I), or macrophages (J–L) in CNV lesions. Confocal images (thin optical sections, 1 μm) were used to show colocalization of various markers with GFP. (A–C) Colocalization (C, arrows) of αSMA (A) with GFP (B). (D–F) Colocalization (F, arrows) of Griffonia simplicifolia lectin (D, BS-1) with GFP (E). (G–I) Colocalization (I, arrow) of CD31 (G) with GFP (H). (J–L) Colocalization (L, arrow) of F4/80 (J) with GFP (K). Scale bar: 20 μm.

Figure 6.

 

Example of a GFP-labeled cell that was αSMA immunoreactive (arrow). The nucleus was labeled with DAPI (blue) and the cytoplasm was immunoreactive for αSMA (red) and GFP (green). (B) Blue and red fluorescence are superimposed; (C), blue, red, and green fluorescence are superimposed. (C, yellow) Overlap of αSMA and GFP immunostaining. The dense outer nuclear layer (photoreceptor cells) is adjacent to the lesion. Scale bar, 20 μm.

Figure 7.

 

Distribution of GFP-labeled monocytes and macrophages in the eye. Superimposed confocal images of vertical sections immunostained for GFP (green) and F4/80 (red). Colocalization appears yellow. (A) GFP-labeled cells accumulated in a CNV lesion, as well as in the overlying retina (arrow) and underlying choroid. (B) Few GFP-labeled cells were present in the choroid (CH) and sclera (arrow, SC) distant from the lesion. (C) In the limbus (arrow) and ciliary body GFP-labeled cells were present at high densities. GFP-labeled cells in the retina over CNV (D) and sclera (E) and most cells in the limbus (F) expressed F4/80. There were F4/80 immunoreactive cells in the limbus that were not GFP labeled. (E, F) Higher magnification views of (B) and (C), respectively. RPE, retinal pigment epithelium. Scale bars (A–C) 100 μm; (D–F) 20 μm.

The authors thank Stephen D. Roper for access to the fluorescence confocal microscope, Gary Striker for his intellectual contributions, and Kalaiselvi Pannerselvam for excellent technical assistance.