Identification of Photoreceptor Precursors in the Pars Plana During Ocular Development and After Retinal Injury

Identiﬁcation of Photoreceptor Precursors in the Pars Plana during Ocular Development and after Retinal Injury

Koji M. Nishiguchi, Hiroki Kaneko, Makoto Nakamura, Shu Kachi, and Hiroko Terasaki

PURPOSE. To study the distribution and differentiation of pho- ripheral retina or the surrounding ciliary body has not been toreceptor precursors in the ciliary epithelium in mice. identified. Evidence suggests that the ciliary epithelium con- METHODS. Proliferating cells in flat-mount specimens of the tains stem cells that have the potential to proliferate and express markers specific to retinal neurons in mammals,4–7 ciliary body and retina were studied by bromodeoxyuridine 8 (BRDU; 150 mg/kg) labeling in young C57Bl mice. Immunore- including primates. However, the exact potential of these activity to anti-recoverin, rhodopsin, and Pax6 antibodies and cells, including whether they can produce morphologically binding to peanut agglutinin were analyzed histologically to differentiated retinal neurons, remains unknown. In mice, the ciliary body is composed of two distinct ana- assess the distribution and differentiation of photoreceptor 9 progenitors or precursors. Mice injected intraperitoneally with tomic substructures. The anterior aspect of the ciliary body N-methyl-N-nitrosourea (MNU; 60 mg/kg) were also examined. with ciliary processes is known as the pars plicata. The pars plana, a flat circumferential zone not more than 12 to 16 cells RESULTS. Part of the neuroblast layer composed of BrdU-positive wide, forms the posterior part of the ciliary body and connects retinal progenitor cells was identified within the ciliary epithe- the retina and the pars plicata.9 Developmentally, both the lium of the pars plana in continuation of the layer of the ciliary epithelium and the retina share a common origin and peripheral retina during ocular development. In both the cili- arise from the inner layer of the optic cup. However, these two ary epithelium and the retina, the layer size decreased rapidly anatomic structures are thought to eventually follow different and disappeared mostly by postnatal day (P)9. Within the developmental routes because each has a distinct role in the ciliary epithelium of the pars plana, numerous postmitotic rod eye; the retina translates light stimuli into neural signals and and cone photoreceptor precursors were identified that rap- conveys them to the brain, whereas the ciliary epithelium idly differentiated morphologically and decreased in number secretes aqueous humor to maintain the intraocular pressure with ocular development. Rod precursors were no longer seen and integrity of the eye. In young mice, when gross structures in the pars plana at P12, whereas rare presumptive cone of the eye, including the ciliary body, are already formed, a precursors persisted even at P120. An increase in the number layer composed of proliferating retinal progenitor cells—called of presumptive cone precursors (approximately 16-fold) was the neuroblast layer—plays an important role in the devel- identified in the pars plana of adult mice with MNU-induced opment of retinal neurons that migrate to form distinct photoreceptor degeneration. retinal layers. As the maturation of the retina begins around CONCLUSIONS. Rod and cone precursors were identified in the the optic nerve and extends gradually toward the retinal ciliary epithelium of the murine pars plana during ocular de- margin, the neuroblast layer in the peripheral retina, adja- velopment but nearly disappeared after the completion of cent to the pars plana, persists at the late stages of retinal histogenesis. However, in response to retinal injury, an in- development. creased number of presumptive cone precursors was found Here, we report that part of the neuroblast layer resides even in the adult pars plana. (Invest Ophthalmol Vis Sci. 2008; within the ciliary epithelium of the pars plana during postnatal 49:422–428) DOI:10.1167/iovs.07-1008 ocular development in mice. Moreover, cone and rod photoreceptor precursors in various stages of morphologic develop- n fish and amphibians, a circumferential zone in the retinal ment are identified in the pars plana during this period and Imargin, named the ciliary marginal zone (CMZ), produces all after retinal injury, even in adult animals. types of retinal neurons and thereby continuously regenerates 1–3 the retina throughout life. However, in mammals, an analo- METHODS gous zone for retinal neurogenesis or regeneration in the pe- Animals All experimental procedures adhered to the ARVO Statement for the From the Department of Ophthalmology, Nagoya University Grad- Use of Animals in Ophthalmic and Vision Research and to the guide- uate School of Medicine, Nagoya, Japan. lines for the use of animals at Nagoya University School of Medicine. Supported by Grant-in Aid C1659174 for Scientific Research from C57BL/6J mice used were kept under a 12-hour light/12-hour dark the Ministry of Education, Culture, Sports, Science, and Technology of cycle. Japan (MN). Submitted for publication August 5, 2007; revised September 17, 2007; accepted November 19, 2007. BrdU Labeling Disclosure: K.M. Nishiguchi, None; H. Kaneko, None; M. Na- To label cells in the S-phase of the cell cycle, mice were injected kamura, None; S. Kachi, None; H. Terasaki, None intraperitoneally with bromodeoxyuridine (BrdU; Sigma, St. Louis, The publication costs of this article were defrayed in part by page MO) at a dose of 150 mg/kg and were humanely killed after 2 hours so charge payment. This article must therefore be marked “advertise- that eyes could be collected. ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Corresponding author: Koji M. Nishiguchi, Department of Oph- Immunohistochemistry thalmology, Nagoya University Graduate School of Medicine, 65 Tsu- ruma, Showa-ku, Nagoya 466-8550, Japan; Immunohistochemical analyses were conducted as previously described, [email protected]. with slight modification.10 For eye sections, the eyecups were fixed in 4%

Downloaded from jov.arvojournals.org on 09/25/2021 IOVS, January 2008, Vol. 49, No. 1 Photoreceptor Precursors in the Pars Plana 423

paraformaldehyde for 2 hours, followed by cryoprotection in 30% sucrose malian eye, mitotic cells were labeled by intraperitoneal injec- overnight at 4°C. The eyes were embedded in OCT compound (Tissue- tion of BrdU into young mice of various ages 2 hours before Tek; Sakura Finetek Japan Co. Ltd., Tokyo, Japan), and frozen sections enucleation. To improve the accuracy of the interpretations of 12-␮m thick were cut through the dorsal to ventral meridian at Ϫ21°C. the experimental results, we deﬁned retinal progenitors in this After sections were permeabilized with 0.1% Triton X-100 in phosphate- study as mitotic cells with the potential to generate retinal buffered saline (PBS) for 15 minutes, they were incubated in 4 N HCl for neurons and retinal precursors as postmitotic cells with mor-

15 minutes, neutralized with 0.1 M Na2B4O7 for 5 minutes, blocked in 5% phologic evidence of neuronal differentiation not yet inte- goat serum in PBS for 30 minutes, and incubated with primary antibodies grated into the retina. We first studied the distribution and for 1 hour and with secondary antibodies for 1 hour. quantity of BrdU-positive cells in the ciliary epithelium and For flat-mount specimens, after radial incisions were made to flatten peripheral retina during (P6 and P9) and after (P12 and P18) the eyecups, the samples were fixed in 4% paraformaldehyde for 2 histogenesis of the retina using eye sections. At P6, many hours, and the vitreous was carefully removed. The eyecups were BrdU-positive cells were observed in the ciliary epithelium and permeabilized in 0.5% PBST for 4 hours, incubated in 2 N HCl for 45 the retina (Figs. 1a, 1b). These cells formed the presumptive

minutes, and neutralized with 0.1 M Na2B4O7 for 15 minutes. After neuroblast layer in the retina and were most frequently iden- blocking with 5% goat serum in PBS for 1 hour, the samples were tified around the pars plana, close to the retinal margin in the incubated with primary antibodies for 12 hours and with secondary ciliary epithelium. The BrdU-positive cells decreased rapidly by antibody for 9 hour. All procedures were conducted at room temper- P9 and continued to decrease further thereafter (Figs. 1c, 1d). ature unless indicated otherwise. Consequently, when BrdU was administered at P18, cells in the Sections and flat-mounts were stained with primary antibodies for ciliary epithelium and retinal margin were only rarely labeled BrdU (1:1000; Oxford Biotechnology, Oxford, UK), rhodopsin (1:1000; in histologic sections. Chemicon, Temecula, CA), recoverin (1:1000; Chemicon), Pax6 (1: To confirm the findings in eye sections, we repeated the 200; Developmental Studies Hybridoma Bank), and proliferating cell experiment using flat-mounts of the ciliary body and retina nuclear antigen (PCNA; 1:200; BD Biosciences, San Jose, CA) followed from mice ranging in age from P6 to P60. The anatomies of the by combinations of fluorescent dye-conjugated secondary antibodies retina and ciliary body are compared between histologic sec- (Alexa 405, 488, and 568; all at 1:1000; Molecular Probes, Eugene, OR), tions and a flat-mount specimen in Figure 2. Immunohisto- peanut agglutinin (PNA; 1:100; Molecular Probes), and diamidino-2- phenylindole (DAPI; 1:1000; Molecular Probes). Identification of the Retinal Margin and Pars Plana with Confocal Microscope The pars plana was distinguished clearly from the retina in flat-mounts because the following features were observed in immunohistochemical analyses. First, the pars plana had features of both the retina and the pars plicata of the ciliary body. The inner surface of the pars plana was flat, showing a contour similar to that of the retina, whereas in the deeper layers it had discernible folds that appeared to be the contin- uum of the pars plicata. Second, with rhodopsin or recoverin staining, an intervening space was frequently detected between immunopositive cells at the retinal margin and pars plana, particularly in adult mice with toxin-induced retinal degeneration. Third, the retina showed dense staining, forming a plane of rhodopsin- or recoverin-positive cells in contrast to the pars plana, which had much less densely distributed cells aligned circumferentially. Fourth, photoreceptor precursors in the pars plana resided only in the surface layer parallel to the retina, consistent with their location in the nonpigmented ciliary epithelium. Meanwhile, retinal photoreceptors located deep in the outer- most nuclear layer were oriented perpendicularly to the vitreoretinal interface and comprised multiple layers of neurons of the same cell type. Therefore, cells in the pars plana could be selectively visualized by scanning the flat-mounts at the level of the ciliary epithelium with a confocal microscope. Although most of these features were confirmed before obtaining images, not all are presented in every figure to FIGURE 1. Distribution of BrdU-positive cells in the ciliary epithelium highlight the cells of interest. of the pars plana during ocular development in histologic sections. Double arrows indicate the pars plana. (a) At P6, the neuroblast layer N-methyl-N-nitrosourea Injection composed of BrdU-positive retinal progenitors (red) is present in the P60 C57BL/6J mice were intraperitoneally injected with N-methyl-N- eye section. BrdU was injected 2 hours before enucleation (applies to nitrosourea (MNU; 60 mg/kg; Sigma) and were humanely killed at P75 all the data presented in Fig. 1). (b) Enlarged image of (a) overlaid with for histologic analysis. DAPI staining (blue). (c) By P9, the numbers of BrdU-positive cells (red) in both the ciliary epithelium and the retina had decreased markedly. (d) The number of BrdU-positive cells in the entire ciliary RESULTS epithelium and retina 200 ␮m from the cilioretinal margin during and after retinal histogenesis counted from eye sections (mean Ϯ SEM). Neuroblast Layer Encompasses Peripheral Retina The number of BrdU-positive cells in the P6 retina was 106 Ϯ 5, which and Ciliary Epithelium of Pars Plana in is indicated by an off-the-scale bar (asterisk). (d) 8, 7, 7, and 7 animals Postnatal Mouse were used at P6, P9, P12, and P18, respectively. Scale bar, 100 ␮m. NBL, neuroblast layer; CB, ciliary body; Cor, cornea; Scl, sclera; I, iris; To test the hypothesis that retinal progenitors or precursors CPE, ciliary body pigmented epithelium; RPE, retinal pigmented epi- are present in the ciliary epithelium in the developing mam- thelium; Rho, rhodopsin.

Downloaded from jov.arvojournals.org on 09/25/2021 424 Nishiguchi et al. IOVS, January 2008, Vol. 49, No. 1

FIGURE 2. Comparison of ocular structures between sections and a flat-mount. (a) Normal histologic section containing the ciliary body and the retina. (a–c) Upper and lower dashed lines correspond to the anterior and posterior border of the pars plana, respectively. (b) Histologic section of flat-mount specimen. Note that the ciliary processes of the pars plicata are displaced toward the iris. (c) Flat-mount of the ciliary body and the retina stained with anti-recoverin antibody at P6. Note that the retinal margin is demarcated by dense recoverin-immunopositive cells. Fig- ures 3–6 and S1–S3 are all images of flat-mounts and are presented in the same orientation as this but at different magnifications. Scale bar, 100 ␮m.

chemical evaluation of BrdU incorporation in P6 flat-mounts Generation of Photoreceptor Precursors in the revealed a discrete zone of dense BrdU-positive cells consistent Pars Plana during Ocular Development with the neuroblast layer (Fig. 3a), as supported by the presence of cells positive for Pax6, a marker for retinal progenitors, We studied the immunoreactivity for rhodopsin, a marker and BrdU immunoreactivity (Supplementary Fig. S1; all supple- specific for rod photoreceptors, in the ciliary epithelium using mentary figures are online at http://www.iovs.org/cgi/content/ flat-mounts. Within the pars plana, we found radially aligned full/49/1/422/DC1). The BrdU-positive cells were also immuno- postmitotic rhodopsin-positive cells with morphologic features reactive for PCNA, a marker for the G1 to S phase of cell cycle, of rod photoreceptor precursors (Figs. 3b, 3c, 4a–c, 4f). How- which confirms that they are proliferating (data not shown). ever, these cells were found only during retinal development. However, concomitant staining of the flat-mounts with anti– Unlike late rod photoreceptor precursors that express rhodop- rhodopsin antibody revealed that a part of the discrete zone sin predominantly in the outer segment, rod photoreceptor of dense BrdU-positive cells roughly corresponded to the precursors were immunoreactive for rhodopsin over the entire pars plana of the ciliary body, in which numbers of photo- cell surface up to P6 (Figs. 4a, 4b). Many of these cells had one receptor precursors were identified (Fig. 3b). These cells— process with budding at the end, consistent with rod spher- reactive to anti–rhodopsin and recoverin antibodies—never ules, and another with appreciable inner and outer segment- stained with anti–Pax6 antibodies. We failed to identify the like structures (Fig. 4c). However, most of the rhodopsin- presence of other classes of retinal neurons, namely bipolar positive cells were not immunoreactive for recoverin (97.5% Ϯ or ganglion cells, using anti–PKC and Brn3 antibodies, re- 2.7% [mean Ϯ SD]; n ϭ 6; Fig. 4b), a marker for rod and cone spectively (data not shown). photoreceptors.13,14 This was in contrast to photoreceptors in The BrdU-positive cells occupied a significant portion of the retinal margins of the same P6 mice, in which recoverin the ciliary body up to P6 (Fig. 3b), but their numbers immunoreactivity was detected in nearly half the rhodopsin- decreased rapidly by P9, when the neuroblast layer was no positive cells (Fig. 4b). By P9, when retinal histogenesis is longer identified in the retina and pars plana in many eyes almost complete, rod photoreceptor precursors were absent (Fig. 3c). Minor proportions of BrdU-positive cells at P6 and from a large portion of the ciliary epithelium, whereas the most of those after P9 were also rhodopsin positive. Their remaining cells resided within a narrow band along the pars morphology11,12 (Supplementary Fig. S2) and distinct loca- plana (Figs. 3c, 4f). Most of these precursors were immunore- tion—they were present not within the retina or the ciliary active for rhodopsin predominantly in their outer segments epithelium but on the surfaces of these structures, they (90.9% Ϯ 4.4%; n ϭ 5; Fig. 4f), consistent with its expression were associated with the presence of residual vitreous (Sup- pattern in mature rods. The visible outer segments were plementary Fig. S2), and they were scattered in the pars aligned radially, and some bridged the retina and pars plana. plana and pars plicata—indicate that these cells were prob- However, such rod photoreceptor precursors were no longer ably hyalocytes. Meanwhile, BrdU-negative rod photorecep- observed by P12 (n ϭ 19) with the completion of retinal tor precursors had a characteristic morphology and rhodop- histogenesis.15–17 sin expression pattern (described here in detail), resided Meanwhile, at P6, most of the recoverin-positive cells in the within the retina or ciliary epithelium, and were never ciliary epithelium had a uniform appearance; one or two pro- identified in the pars plicata. Taken together, these results cesses were typically shorter than those of rod photoreceptor indicate that the rhodopsin staining in BrdU-positive cells precursors in the retina (91.8% Ϯ 7.7%; n ϭ 5; Figs. 4d, 4e). was probably false positive. At P12, there were only spo- These cells were probably cone photoreceptor precursors. As radic BrdU-positive cells in the ciliary body (Fig. 3d), some we confirmed in some of the cells, one of the processes was of which were also immunoreactive for Pax6 (Figs. 3e, 3f), positive for peanut agglutinin (PNA; Fig. 4d), a marker specific suggesting that they maintained the potential to develop for the inner and outer segments and pedicles of cone photo- into retinal neurons. The BrdU-positive cells continued to receptors.18,19 This is consistent with previous reports indicat- decrease in number. ing that in the developing eye, cone photoreceptor precursors

Downloaded from jov.arvojournals.org on 09/25/2021 IOVS, January 2008, Vol. 49, No. 1 Photoreceptor Precursors in the Pars Plana 425

express recoverin earlier than rod photoreceptor precur- cess (68.1% Ϯ 4.4%; n ϭ 6). Some areas of the pars plana still sors.13,14 Unfortunately, accurate evaluation of cone photore- contained recoverin-positive cells at P18 (Fig. 5a), but by P30 ceptor precursors was difficult because of the relatively weak these cells were nearly absent from the ciliary epithelium (Figs. signal and the high background of PNA staining, particularly 5b, 5c). However, rare cells persisted in the pars plana up to at during retinal development.18,20 By P9, many recoverin-posi- least P120 (data not shown). Most of these persistent recov- tive retinal precursors in the pars plana had developed a longer erin-positive cells had morphology consistent with early cone process (65.1% Ϯ 17.1%; n ϭ 4) and aligned radially (Figs. 4g, photoreceptor precursors. 4h). In addition to the appreciable inner and outer segment structures (Figs. 4h, 4i), some of these long processes showed Photoreceptors in the Pars Plana of Adult Mice PNA-positive terminal budding consistent with the pedicles of cone photoreceptors18–20 (Supplementary Fig. S3). The ob- with MNU-Induced Retinal Degeneration served morphologic features of recoverin-positive cells are in Our findings suggest that the existence of rod and cone pho- close agreement with immature cone precursors found in the toreceptor precursors in the ciliary epithelium of the pars 20 developing retina, which is summarized in Supplementary plana is mostly restricted to a short period during retinal Figure S4. Some of these cells bridged the ciliary body and the development in wild-type mice. In the brain, it is well known retina. At P12, many recoverin-positive cells persisted in the that neural stem cells proliferate in response to neural injury pars plana, most of which had a detectable PNA-positive pro- even in adult mammals.21–23 The same is true for retinal progenitor/stem cells in mice with certain genetic conditions.24,25 To examine whether photoreceptor precursors are identified in the adult ciliary epithelium, perhaps by neurogenesis and neural differentiation in mice with severe retinal injury, flat- mounts of the eyes from toxin-induced retinal degeneration were evaluated. We analyzed the pars plana of P75 C57Bl mice injected with MNU at P60. By 7 days after intraperitoneal injection of MNU, severe photoreceptor degeneration is observed in injected animals.26 In addition to the reappearance of rare rhodopsin-positive rod photoreceptor precursors (data not shown), an increased number of recoverin-positive photoreceptor precursors in the pars plana was seen in MNU- treated mice compared with untreated control mice at P60 (10.7-fold; P ϭ 2.1 ϫ 10Ϫ5) or P75 (16.5-fold; P ϭ 4.3 ϫ 10Ϫ6; Fig. 6). Most of these cells had a prominent process with relatively large budding at the end, simulating cone pedicle. They were aligned randomly, in contrast to cells in the adjacent retina, which were organized radially (Fig. 6a). Many of these cells also had a PNA-positive process. A few recoverin-positive cells had two relatively long processes with a small budding in one of its end resembling rod photoreceptors (Fig. 6a). These cells were rarely found before MNU treatment and appeared to be consistent with newly generated photoreceptor precursors. However, we did not obtain direct evidence that these cells were generated from retinal progenitors after MNU injection, despite

FIGURE 3. Distribution of BrdU-positive cells and rod photoreceptor precursors in the pars plana in flat-mounts. (a) At P6, a distinct layer composed of BrdU-positive cells (red), the neuroblast layer, was identified. (b) Merged image with (a) showing BrdU (red) and rhodopsin (green) labeling. Numerous rhodopsin-positive rod photoreceptor precursors (filled arrowheads) are identified in the pars plana. Note that the retinal margin (open arrowheads) is clearly distinguished by an alteration in the density of rhodopsin-positive cells relative to the adjacent ciliary epithelium. (c) By P9, the pars plana contained spo- radic BrdU-positive cells (red). Note that only the inner/outer segments of the rod photoreceptor precursors stain with anti-rhodopsin antibodies (green; open arrowheads) at this age. (d) Rod photoreceptor precursors were no longer seen in the pars plana at P12. Round to oval cells double-positive for rhodopsin and BrdU (c, d; filled arrowheads) are likely hyalocytes based on their location and morphology (Supple- mentary Fig. S2).11,12 Rod photoreceptor precursors with evidence of morphologic differentiation were never labeled with BrdU. (e, f) Some BrdU-positive cells (red) in the pars plana were also immunoreactive for Pax6 (green; double-positive cells are indicated with filled arrowheads). Densely and evenly distributed postmitotic (BrdU-negative) Pax6-positive cells in the retinal surface are consistent with ganglion/ amacrine cells.31,32 Note that retinal vessels showed false-positive staining (open arrowheads). Scale bar, 100 ␮m. CB, ciliary body; Rho, rhodopsin; PP, pars plana.

Downloaded from jov.arvojournals.org on 09/25/2021 426 Nishiguchi et al. IOVS, January 2008, Vol. 49, No. 1

FIGURE 4. Postmitotic rod and cone photoreceptor precursors in the pars plana in flat-mounts. Double arrows: pars plana. (a) At P6, within the pars plana, rhodopsin-positive (green) rod photoreceptor precursors (filled arrowhead) were found between BrdU-positive cells (red). BrdU was injected 2 hours before enucleation, which applies also to (d). (b) Rod photoreceptor precursors (green; filled arrowhead) in the pars plana were negative for recoverin immunoreactivity (blue), whereas many of those in the retina were positive (open arrowhead). (c) Schematic illustration of a rod photoreceptor precursor labeled with a filled arrowhead in (a, b). (d) At P6, many recoverin-positive cells (blue) with a short PNA-positive process (green; filled arrowhead) were found between BrdU-positive cells (red). (e)A selected area with recoverin-positive cells (green) in both the pars plana (open arrowhead) and the pars plicata (filled arrow) at P6. (f)AtP9, rhodopsin immunoreactivity (green) was localized to the outer segments (filled arrowhead) of rod photoreceptor precursors. (g) At P9, recoverin-positive cells (green) developed a long process in addition to the short process positive for PNA (red). (h) Many recoverin-positive cells (green) with a long process were aligned perpendicularly to the retinal margin. (i) Schematic illustration of a recoverin-positive cell labeled with a filled arrowhead in (h). Scale bar, 50 ␮m. CB, ciliary body; PP, pars plana; Rho, rhodopsin; Rcv, recoverin; IS, inner segment; OS, outer segment.

analyzing the eyes collected at variable times after BrdU development under certain circumstances. In the pars plana of labeling (data not shown). adult mice with MNU-induced retinal degeneration, increased numbers of photoreceptor precursors were found. The morphology of most of these cells was consistent with that of cone DISCUSSION photoreceptor precursors with well-developed cone pedicles, Based on the analyses of flat-mounts and sections from the but they lacked the prominent inner and outer segment struc- murine ciliary body and peripheral retina, it appears that at tures seen frequently at approximately P9 in the pars plana and least part of the ciliary epithelium in the pars plana and neural at P4 in the retina during ocular development20 (Supplemen- retina are derived from the same neuroblast layer from which tary Fig. S3). However, we were able to find little evidence photoreceptor precursors are produced during postnatal ocu- supporting their de novo neurogenesis from retinal progeni- lar development in mice. We could not detect the presence of tors or stem cells with pulse-chase study using BrdU (data not bipolar or ganglion precursors in the ciliary epithelium with shown), possibly because of the adverse effect of BrdU incor- the condition and antibodies used in this study. However, in poration on retinal progenitors, such as the inhibition of dif- the pars plana, presumptive early cone photoreceptor precur- ferentiation or death of these cells.27 Meanwhile, it is possible sors were rarely found after P30. Interestingly, morphologically that retinal precursors in the pars plana arose from cells diffi- mature photoreceptor precursors were not found in the pars cult to detect with BrdU labeling, such as slowly cycling or plana. The observations that some of the photoreceptor pre- quiescent retinal progenitors, postmitotic retinal precursors, a cursors bridged both the retina and the pars plana and that the small number of rapidly dividing stem cells, or even ciliary precursors rapidly disappeared after retinal development raise epithelial cells by transdifferentiation. the possibility that these cells migrate into the peripheral retina The role and distribution of reported multipotential retinal to participate in its histogenesis. It is also possible that these stem cells4,5 remains unknown, but the relatively small number cells die in situ with no virtual significance. Meanwhile, our of photoreceptor precursors identified in the pars plana sug- results suggest that the potential of ciliary epithelium to pro- gests that the physiological significance, if any, of these cells duce photoreceptor precursors may not be limited to ocular with regard to their ultimate contribution to retinal functional

Downloaded from jov.arvojournals.org on 09/25/2021 IOVS, January 2008, Vol. 49, No. 1 Photoreceptor Precursors in the Pars Plana 427

may be limited. This is unlike the adult CMZ in the retinal margin of the lower vertebrates, which contains numerous dividing retinal progenitors capable of replacing large number of various classes of neurons throughout the retina.1–3 In the pars plana, we found postmitotic photoreceptor precursors of late ontogenetic stages that have recently been shown to be an ideal source for photoreceptor transplantation.28 The only other known sources of such cells are the immature retina.28–30 Although fetal retinal tissue transplants yielded photoreceptors consistent with rod and cone photoreceptors,29,30 cell transplants derived from the immature retina developed almost exclusively into rod photoreceptors28 re- sponsible for vision under dim light only. On the other hand, we found that large proportions of photoreceptor precursors generated in the pars plana were presumably of the cone photoreceptor lineage, which is by far the most important class of photoreceptors mediating daylight vision. Therefore, the use of retinal progenitors/precursors from the pars plana may have a signiﬁcant advantage when considering cone photoreceptor cell replacement therapy for treating most types of photoreceptor degeneration. However, its simple clinical ap- plication may be hampered by the relatively small amount of collectable cells. In conclusion, at least a part of the ciliary epithelium in the pars plana is derived from the neuroblast layer. Rod and cone photoreceptor precursors are identiﬁed in the pars plana during murine retinal development and may participate in the histogenesis of the peripheral retina. In adult mice with severe retinal injury, increased numbers of photoreceptor precursors are also observed in the pars plana. These results raise the

FIGURE 6. Photoreceptor precursors in the pars plana of mice with MNU-induced retinal degeneration. (a) At P75, an increased number of recoverin-positive photoreceptor precursors were identified in the pars plana (double arrow) of mice treated with MNU. Most of these cells had a single prominent process with relatively large budding at its end resembling cone pedicle (filled arrowhead). Rare cells had two processes, with small budding at one end simulating rod photoreceptor precursors (open arrowhead). Recoverin-positive cells in the pars plana were aligned randomly, whereas most of those in the peripheral retina were organized radially. Note that recoverin-positive cells at the retinal margin had prominent inner/outer segment structures and were morphologically consistent with mature cone photoreceptors (filled arrow). (b) An increased number of recoverin-positive photoreceptor precursors (mean Ϯ SEM) was observed in the pars plana of P75 wild-type mice treated with MNU at P60 (n ϭ 7) compared with untreated litters at P60 (10.7-fold; n ϭ 6) or P75 (16.5-fold; n ϭ 7). Scale bar, 100 ␮m. CB, ciliary body; Rho, rhodopsin; Rcv, recoverin; WT, wild-type.

possibility that the pars plana may serve as a potential source for photoreceptor replacement therapy. FIGURE 5. Rare recoverin-positive photoreceptor precursors in the pars plana after retinal histogenesis. Double arrows: pars plana. (a) Recoverin-positive cells with a short process (arrowhead) persisted in Acknowledgments the pars plana and the pars plicata at P18. (b) At P30, there were only The authors thank Thaddeus P. Dryja and Motokazu Tsujikawa for helpful rare recoverin-positive retinal precursors in the retina. Most of these cells had irregular to oval cell bodies with a short single process, with discussions and advice regarding this manuscript. Anti–Pax6 monoclonal a small budding at its end morphologically consistent with early cone antibodies were obtained from the Developmental Studies Hybridoma photoreceptor precursors ﬁrst seen in the embryonic retina.20 (c) Bank developed under the auspices of the National Institute of Child There were no recoverin-positive cells in most parts of the pars plana Health and Human Development and maintained by the Department of in P30 mice. Scale bar, 50 ␮m. CB, ciliary body; Rcv, recoverin. Biological Sciences at the University of Iowa (Iowa City, IA).

Downloaded from jov.arvojournals.org on 09/25/2021 428 Nishiguchi et al. IOVS, January 2008, Vol. 49, No. 1

References 17. Young RW. Cell differentiation in the retina of the mouse. Anat Rec. 1985;212:199–205. 1. Moshiri A, Close J, Reh TA. Retinal stem cells and regeneration. Int 18. Blanks JC, Johnson LV. Selective lectin binding of the developing J Dev Biol. 2004;48:1003–1014. mouse retina. J Comp Neurol. 1983;221:31–41. 2. Reh TA, Fischer AJ. Stem cells in the vertebrate retina. Brain 19. Haverkamp S, Grunert U, Wassle H. The synaptic architecture of Behav Evol. 2001;58:296–305. AMPA receptors at the cone pedicle of the primate retina. J Neu- 3. Hitchcock P, Ochocinska M, Sieh A, Otteson D. Persistent and rosci. 2001;21:2488–2500. injury-induced neurogenesis in the vertebrate retina. Prog Retin 20. Fei Y. Development of the cone photoreceptor mosaic in the Eye Res. 2004;23:183–194. mouse retina revealed by fluorescent cones in transgenic mice. 4. Tropepe V, Coles BL, Chiasson BJ, et al. Retinal stem cells in the Mol Vis. 2003;9:31–42. adult mammalian eye. Science. 2000;287:2032–2036. 21. Jin K, Galvan V, Xie L, et al. Enhanced neurogenesis in Alzheimer’s 5. Ahmad I, Tang L, Pham H. Identification of neural progenitors in disease transgenic (PDGF-APPSw, Ind) mice. Proc Natl Acad Sci the adult mammalian eye. Biochem Biophys Res Commun. 2000; USA. 2004;101:13363–13367. 270:517–521. 22. Jin K, Peel AL, Mao XO, et al. Increased hippocampal neurogenesis 6. Haruta M, Kosaka M, Kanegae Y, et al. Induction of photoreceptor- in Alzheimer’s disease. Proc Natl Acad Sci USA. 2004;101:343– specific phenotypes in adult mammalian iris tissue. Nat Neurosci. 347. 2001;4:1163–1164. 23. Felling RJ, Snyder MJ, Romanko MJ, et al. Neural stem/progenitor 7. Akagi T, Mandai M, Ooto S, et al. Otx2 homeobox gene induces cells participate in the regenerative response to perinatal hypoxia/ photoreceptor-specific phenotypes in cells derived from adult iris ischemia. J Neurosci. 2006;26:4359–4369. and ciliary tissue. Invest Ophthalmol Vis Sci. 2004;45:4570–4575. 24. Moshiri A, Reh TA. Persistent progenitors at the retinal margin of 8. Coles BL, Angenieux B, Inoue T, et al. Facile isolation and the ptcϩ/Ϫ mice. J Neurosci. 2004;24:229–237. characterization of human retinal stem cells. Proc Natl Acad Sci 25. Dhomen NS, Balaggan KS, Pearson RA, et al. Absence of chx10 USA. 2004;101:15772–15777. causes neural progenitors to persist in the adult retina. Invest 9. Smith RS, John SWM, Nishina PM, Sundberg JP. Regional anatomy Ophthalmol Vis Sci. 2006;47:386–396. and development. In: Smith RS, ed. Systemic Evaluation of the 26. Yoshizawa K, Nambu H, Yang J, et al. Mechanisms of photorecep- Mouse Eye. Boca Raton, FL: CRC Press, 2002:13–27. tor cell apoptosis induced by N-methyl-N-nitrosourea in Sprague- 10. Nishiguchi KM, Nakamura M, Kaneko H, Kachi S, Terasaki H. The Dawley rats. Lab Invest. 1999;79:1359–1367. role of VEGF and VEGFR2/Flk1 in proliferation of retinal progen- 27. Sekerkova G, Ilijic E, Mugnaini E. Bromodeoxyuridine adminis- itor cells in murine retinal degeneration. Invest Ophthalmol Vis tered during neurogenesis of the projection neurons causes cere- Sci. 2007;48:4315–4320. bellar defects in rat. J Comp Neurol. 2004;470:221–239. 11. Zhu M, Provis JM, Penfold PL. The human hyaloid system: cellular 28. MacLaren RE, Pearson RA, MacNeil A, et al. Retinal repair by phenotypes and inter-relationships. Exp Eye Res. 1999;68:553–563. transplantation of photoreceptor precursors. Nature. 2006;444: 12. Ogawa K. Scanning electron microscopic study of hyalocytes in 203–207. the guinea pig eye. Arch Histol Cytol. 2002;65:263–268. 29. Ghosh F, Bruun A, Ehinger B. Immunohistochemical markers in 13. Stepanik PL, Lerious V, McGinnis JF. Developmental appearance, full-thickness embryonic rabbit retinal transplants. Ophthalmic species and tissue specificity of mouse 23-kDa, a retinal calcium- Res. 1999;31:5–15. binding protein (recoverin). Exp Eye Res. 1993;57:189–197. 30. Aramant RB, Seiler MJ. Transplanted sheets of human retina and 14. Alexiades MR, Cepko CL. Subsets of retinal progenitors display retinal pigment epithelium develop normally in nude rats. Exp Eye temporally regulated and distinct biases in the fates of their prog- Res. 2002;75:115–125. eny. Development. 1997;124:1119–1131. 31. Walther C, Gruss P. Pax-6, a murine paired box gene, is expressed 15. Olney JW. An electron microscopic study of synapse formation, in the developing CNS. Development. 1991;113:1435–1449. receptor outer segment development, and other aspects of devel- 32. de Melo J, Qiu X, Du G, Cristante L, Eisenstat DD. Dlx1, Dlx2, oping mouse retina. Invest Ophthalmol. 1968;7:250–268. Pax6, Brn3b, and Chx10 homeobox gene expression defines the 16. Young RW. Cell proliferation during postnatal development of the retinal ganglion and inner nuclear layers of the developing and retina in the mouse. Brain Res. 1985;353:229–239. adult mouse retina. J Comp Neurol. 2003;461:187–204.

Downloaded from jov.arvojournals.org on 09/25/2021