Progranulin promotes the retinal precursor cell proliferation and the photoreceptor differentiation in the mouse retina

Progranulin (PGRN) is a secreted growth factor associated with embryo development, tissue repair, and inflammation. In a previous study, we showed that adipose-derived stem cell-conditioned medium (ASC-CM) is rich in PGRN. In the present study, we investigated whether PGRN is associated with retinal regeneration in the mammalian retina. We evaluated the effect of ASC-CM using the N-methyl-N-nitrosourea-induced retinal damage model in mice. ASC-CM promoted the differentiation of photoreceptor cells following retinal damage. PGRN increased the number of BrdU+ cells in the outer nuclear layer following retinal damage some of which were Rx (retinal precursor cell marker) positive. PGRN also increased the number of rhodopsin+ photoreceptor cells in primary retinal cell cultures. SU11274, a hepatocyte growth factor (HGF) receptor inhibitor, attenuated the increase. These findings suggest that PGRN may affect the differentiation of retinal precursor cells to photoreceptor cells through the HGF receptor signaling pathway.

PGRN treatment after retinal damage increased the number of BrdU + cells in the ONL. In a previous reported we found that ASC-CM contained a high concentration of PGRN (75 fold ASC-CM equals PGRN 574.15 ng/mL) 15 . To confirm whether the differentiation effect of ASC-CM ( Fig. 1) resulted from PGRN using light-induced retinal damage model, a better model similar to a pathology of retinal degenerative diseases, we investigated whether PGRN promotes the differentiation of retinal precursor cells to retinal photoreceptor cells after retinal damage in vivo. Recent reports have showed that PGRN is associated with muscle regeneration through the regulation of myogenic progenitor cells 21 . Therefore, we investigated the effect of PGRN on retinal regeneration. No BrdU + cells were observed in any retinal layer in an non-injured normal group (Fig. 2B). PGRN treatment after retinal damage increased the number of BrdU + cells in the inner plexiform layer (IPL) and ONL (approximately 4 fold) compared to the control group (Fig. 2B,D). PGRN had no effect on the number of BrdU + cells in GCL and inner nuclear layer (INL). The BrdU + cells in the PGRN group were not rhodopsin positive in spite of their presence in the ONL (Fig. 2B,D). PGRN promoted Rx + retinal precursor cell genesis in the ONL. We observed that PGRN increased the number of BrdU + cells in the ONL. To eliminate the possibility that these BrdU + cells were produced by glial cell proliferation during gliosis, we used double staining with BrdU and glial cell specific markers. No co-staining in the ONL was observed using antibodies to glial fibrillary acidic protein (GFAP), a marker specific for astrocytes, and BrdU in the control or PGRN-treated groups (Supplementary Fig. 2A), with GFAP expression mainly observed in retinal inner layer. Moreover, no co-staining was apparent using antibodies to ionized calcium binding adaptor molecule 1 (Iba-1), a marker specific to microglia and BrdU staining in the ONL in the control or PGRN-treated groups ( Supplementary Fig. 2B) with Iba-1 expression mainly observed in the IPL. This data shows that BrdU + cells are not glial cells in the ONL. We also performed immunostaining using retinal precursor cell markers. Pax6 and Rx (retinal homeobox protein) were selected as appropriate retinal precursor cell markers. Pax6 is a transcription factor which is closely associated with eye development 22 , it is expressed in ganglion, amacrine, and retinal precursor cells 23 . Rx is also associated with retinal development and is expressed in retinal precursor cells 24,25 . Pax6 labeling in the INL was mainly observed in INL because the amacrine cells were labeled ( Supplementary Fig. 2C). Pax6 was not expressed in any cells in the ONL. On the other hand, some Rx + cells were observed in PGRN-treated ONL ( Supplementary Fig. 2C). No Rx + cells were observed in the control (vehicle-treated group) group but co-staining of Rx and BrdU showed that a few BrdU + cells in the ONL were Rx + in the PGRN-treated group ( Supplementary Fig. 2C,D). However, the high background (non-specific signals) was observed in control and PGRN-treated group. Then, we investigated about the localization of Rx mRNA by in situ hybridization (ISH). The staining by ISH and immunofluorescense revealed Rx mRNA was colocalized with Rx protein and BrdU in PGRN-treated ONL, but not control group (Fig. 2C). This suggested that the increase in BrdU + cells in the ONL resulting from PGRN treatment were a few of Rx + retinal precursor cells. Nestin is a marker of neural progenitors. It is reported that nestin is expressed when the injury induces Müller glial neural stem cell-like properties 14 . Nestin expression in PGRN-treated group was not altered compared to the control group ( Supplementary Fig. S3A). Sox2 is a stem cell marker and we observed a few of BrdU and Sox2 double-positive cells in PGRN-treated group ( Supplementary Fig. S3B). Moreover, cone-rod homeobox protein (CRX) indicates the presence of retinal photoreceptor precursor cells 29 , and we investigated whether PGRN increased the CRX expression. Light damage did not generate the expression of CRX as seen the control group. CRX expression was observed in the PGRN-treated group ( Supplementary Fig. S4). These results suggest that PGRN increased the newly-generated retinal precursor cells in ONL. SU11274, an HGFR inhibitor, attenuated the promotion of photoreceptor differentiation by PGRN. We investigated whether HGFR signaling could affect the differentiation induced by PGRN.

PGRN increased rhodopsin
Immunoblots showed that PGRN treatment increased the phosphorylation of HGFR after 5 min. Co-incubation with the HGFR inhibitor, SU11274, attenuated the phosphorylation by PGRN (Fig. 4B). Next, we confirmed the no change in the cell number in control group, PGRN-treated group and PGRN and SU11274 co-treated group ( Supplementary Fig. S5B). Immunostaining results showed that SU11274 at a concentration of 1 μM inhibited the increase of rhodopsin + cells by PGRN (Fig. 4C,D). The increase in rhodopsin expression by PGRN was attenuated by SU11274 treatment (Fig. 4E,F). The addition of SU11274 alone had no effect on rhodopsin expression (Fig. 4E,F).

PGRN loss caused retinal neuron loss.
We measured retinal layer thickness in 8-12 week old PGRN-knockout mice 30 . We first confirmed that PGRN protein was not observed in Grn −/− mice using western blotting and immunohistochemistry (Fig. 5A,B). The ONL thickness was significantly decreased in Grn −/− mice compared to wild-type (WT) mice (Fig. 5C). The ONL thickness of Grn −/− mice at 4 weeks old also tended to be decreased compared to heterozygous PGRN-knockout (Grn +/− ) mice at the same age ( Supplementary Fig. S6A-C). The cell number in the GCL was also decreased in Grn −/− mice ( Supplementary Fig. S8A,B). There were no significant changes in any of the retinal layers in Grn +/− mice ( Fig. 5C and Supplementary Fig. S8). In this period, the cell proliferation was not occurred because of Ki-67 negative in all retinal cell layer of Grn −/− and WT mice ( Supplementary Fig. S7). Western blotting demonstrated a decrease in rhodopsin expression in Grn −/− mice compared to WT mice (Fig. 5D). These results suggest that the retina in Grn −/− mice exhibited abnormal photoreceptor cell development.

Discussion
It has been reported that adult mammals slightly show a limited potential for regeneration of retinal neurons after injury 19 . Brain-derived neurotrophic factor (BDNF) treatment following MNU-induced retinal damage has been shown to promote Müller glial cell proliferation and differentiation to photoreceptor cells 17 . In the present study, we also used an MNU-induced retinal injury model to investigate whether ASC-CM, containing PGRN and the other factors promotes the proliferation of retinal precursor cells and the differentiation to photoreceptor cells because BrdU + cells are increased by the treatment of ASC-CM and PGRN and furthermore BrdU + cells were colocalized with photoreceptor marker or retinal precursor cell marker. These findings suggest that PGRN and ASC-CM were associated with the migration and the proliferation of or the differentiation of retinal photoreceptor precursor cells after retinal damage in mice. In a previous report, we showed that ASC-CM contains a The eyes from 8-day old mice were enucleated and the retinas were dissected. After dissection the retinas were centrifuged with any reagents. The retinal cells were incubated for 20 h after dissociation. After incubation, the medium was changed and vehicle or PGRN (500 ng/mL) was added to the retinal cell culture. After 3 days, reagents were added to the culture. The cells were collected for western blotting (after 4 days) and for immunostaining (after 5 days). number of growth factors (HGF and activin A etc.) in addition to PGRN 15 . Activin A promotes the differentiation of photoreceptor cells in vitro 31 . The deletion of the HGFR gene impairs liver regeneration through a decrease in oval cell migration and hepatocytic differentiation 32 . HGF is also associated with axonal regeneration after optic nerve crush 33 . On the basis of these results and those from previous reports, it appears that ASC-CM containing PGRN, HGF, and activin A may exert multiple effects on retinal precursor cells and promote their differentiation to rhodopsin + photoreceptor cells following retinal damage (Fig. 1B-D). However, PGRN treatment alone did not result in the full regeneration of photoreceptor cells in vivo ( Fig. 2 and Supplementary Figs S2-4). PGRN increased BrdU + cells in the ONL and the very few of these were Rx + retinal precursor cells (Fig. 2). An HGFR inhibitor suppressed the differentiation to photoreceptor cells promoted by PGRN (Fig. 4D-G). Previous reports have shown that PGRN treatment can induce the phosphorylation of HGFR in cultured cell line 15 , which is consistent with the PGRN induced phosphorylation of HGFR found in the present study (Fig. 4B). Zebrafish GrnA (an orthologue of mammalian PGRN) knockdown decreased the protein expression of HGFR and downstream β -catenin 15,21,34 , suggesting that PGRN is closely involved in HGFR signaling. HGFR is associated with oval cell migration 32 and the proliferation and migration of myogenic precursor cells 35 . The activation of the HGFR pathway by PGRN may result in the proliferation and the migration of Rx + retinal precursor cells into the ONL.
PGRN promoted differentiation to rhodopsin + photoreceptor cells and resulted in a decrease in CRX + photoreceptor precursor cells and DCX + neural precursor cells (Fig. 3C-F). Some reports have shown that PGRN may be involved in hepatocyte growth factor receptor (HGFR) and Wnt/β -catenin signaling 10,15,34 , and an association between HGFR and Wnt signaling has been suggested 36,37 . The activation of the Wnt signaling pathway promotes Müller glial cell proliferation and dedifferentiation 14 , whilst inhibition of Wnt signaling promotes neuronal differentiation 38 . On the other hand, Wnt activation increases adult hippocampal neurogenesis by increasing DCX + cells and Tuj1 + mature neurons 13 . However, the association between Wnt signaling and neuronal differentiation remains controversial.
The present study showed that the thickness of the ONL was also decreased in younger Grn −/− mice retina (Fig. 5C,D). A previous report showed that the thickness of the ONL was decreased in 12 months old Grn −/− mice 39 with the accumulation of retinal lipopigments. The decrease in the ONL thickness in Grn −/− mice suggests that endogenous PGRN is essential for retinal photoreceptor cell development. These results show that PGRN may play a key role in photoreceptor cell development. However, PGRN may also be associated with the survival of retinal precursor cell or retinal precursor cell proliferation and migration considering from the result of PGRN  (Figs 2 and 3). PGRN deletion may decrease the ONL thickness through a broader effect. Further investigation should clarify whether PGRN is associated with photoreceptor cell development.
Previous reports have shown that the combination of Wnt and retinoic acid or valproic acid promote differentiation to photoreceptor cells 14 . Valproic acid treatment after lentivirus-mediated expression of Sox2 (sex determining region Y-box 2) promotes the maturation of neurons in the brain 18 . The present findings demonstrate that ASC-CM and PGRN may be associated with the migration and the proliferation of retinal precursor cells and their differentiation to photoreceptor cells (Fig. 6). Retinal regeneration proceeds through two pathways, namely dedifferentiation (of Müller glia) and differentiation (of retinal precursor cell). The combination of factors promoting dedifferentiation and PGRN may encourage the regeneration of photoreceptor cells after injury.

Materials and Methods
Animals. Male adult C57BL/6, ddY mice, female ddY pregnant mice, and neonatal mice (Japan SLC, Hamamatsu, Japan) were maintained under controlled lighting environment (12 h/12 h light/dark cycle). PGRN knockout mice generated by the technique of Kayasuga et al. 30 were obtained from Riken BioResource Center (Tsukuba, Japan) and were backcrossed with C57BL/6 mice. Genotyping was performed as described in the Data Sheet provided by the Riken BioResource Center. All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved and monitored by the Institutional Animal Care and Use Committee of Gifu Pharmaceutical University. ASC isolation, culture and the collection of CM. Murine ASCs were obtained from a C57BL/6-Tg (CAG-EGFP) mouse that ubiquitously expresses enhanced green fluorescent protein (EGFP) as previously reported 15 . Adipose tissue was taken from a subcutaneous site. The inguinal fat pads were removed for ASC culture and the tissue including ASC was obtained as previously described 40 . The fat tissue was cut up with a blade, digested with 0.15% collagenase (Wako Pure Chemical Industries, Ltd., Osaka, Japan), and centrifuged. The cell pellet was re-suspended in 10% fetal bovine serum (FBS: Thermo Scientific, Waltham, MA, USA)/Dulbecco's modified Eagle's medium (DMEM: NacalaiTesque Inc, Kyoto, Japan) and plated onto a 100-mm culture dish. ASCs were maintained in 10% FBS/DMEM, 100 U/mL penicillin (Meiji Seika Pharma Co., Ltd., Tokyo, Japan), and 100 μg/mL streptomycin (Meiji Seika) in a humidified atmosphere of 95% air and 5% carbon dioxide (CO 2 ) at 37 °C. The cells were passaged by trypsinization every 2-3 days with cells from passages 4 to 8 harvested for use in experiments. For the collection of CM, ASCs (4 × 10 5 cells) were cultured in FBS-free DMEM. ASC-CM were collected after 72 h of culture, centrifuged at 300 × g for 5 min and filtered using a 0.22-μm syringe filter. The media were concentrated by centrifugation at 2,600 × g using the Amicon Ultra-15 units (Millipore, Bedford, MA, USA; molecular weight cutoff: 3,000).
Primary retinal cell culture. Retinas from P8 ddY mice were dissected to remove the choroidal vessels and the cells were dissociated by incubating for 20 min in pre-activated papain at 37 °C according to the protocol used in our previous report 41 . Neurobasal medium (Invitrogen, Carlsbad, CA, USA), including ovomucoid (Sigma-Aldrich, St. Louis, MO, USA) and DNase (Sigma-Aldrich) was added to the cells. The cells were then centrifuged at 522 × g for 8 min at room temperature. The pellet was suspended in neurobasal medium including ovomucoid without DNase and re-centrifuged. The cells were then resuspended in neurobasal medium containing L-glutamine, B27 supplement (Invitrogen) and antibiotics. Cells were plated onto poly-D-lysine (sigma)/laminin (corning)-coated 12-well plate at a concentration of 2.0 × 10 6 cells/well and onto glass chamber slides at 1.0 × 10 6 cells/well. After incubation for 20 h, the medium was changed to neurobasal medium containing L-glutamine, B27 minus antioxidants (Invitrogen) and antibiotics. At 1 h prior to PGRN treatment, an HGFR inhibitor, SU11274 (Merck & Co., Whitehouse Station, NY, USA) was added at a final concentration of 1 μM. Recombinant mouse PGRN (R&D systems, minneapolis, MN, USA) was added at a final concentration of 500 ng/mL. The vehicle-treated group was treated medium alone. Three days after the cells had been isolated, the medium was changed with the same additions as outlined above. After 5 days the cells were harvested and used for western blotting or immunostaining.

N-methyl-N-nitrosourea (MNU)-induced retinal damage model in vivo.
Male ddY mice were injected 60 mg/kg MNU (Sigma-Aldrich) by intraperitoneal (i.p.) injection according to our previous procedure 41 . Mice were also injected 50 mg/kg 5-Bromo-2′ -deoxyuridine (BrdU: Sigma-Aldrich) by i.p. injection. After this treatment, mice were injected 150-fold concentrated ASC-CM (2 μL) by intravitreal (i.v.) injection in the left eye under general anesthesia. Mice were anesthetized with 3.0% isoflurane (Merck Animal Health, Boxmeer, The Netherlands) and maintained with 1.5% isoflurane in 70% nitrous oxide and 30% oxygen by using an animal general anesthesia apparatus (Soft Lander; Sin-ei Industry Co., Ltd., Saitama, Japan). For the vehicle-treated (control) group, mice were injected with 150-fold concentrated DMEM (2 μL). At 2 and 4 days after MNU treatment, mice were treated with BrdU and ASC-CM or vehicle similarly. After 5 days, the eyes were enucleated and used for immunohistochemistry.

Light-induced retinal damage model in vivo.
We have previously performed the evaluation using a light-induced retinal degeneration model 15,42,43 . The mice underwent dark adaptations for 24 h, the pupils of the mice were then dilated using 1% cyclopentolate hydrochloride eye drops (Santen, Osaka, Japan) 30 min before exposure to light. Non-anesthetized mice were exposed to 8,000 lx of white fluorescent light (Toshiba, Tokyo, Japan) for 3 h in cages with a reflective interior. Following the light exposure, mice were injected 50 mg/kg BrdU (Sigma-Aldrich) by i.p. injection. After BrdU treatment, in the dark, mice were injected with recombinant mouse PGRN (R&D systems) 250 μg/mL (2 μL) by intravitreal (i.v.) injection in the left eye under general anesthesia. Mice were anesthetized as described above. In the vehicle-treated (control) group, mice were injected with D-PBS (Wako Pure Chemical Industries, Ltd.). The temperature during exposure to light was maintained at 25 °C ± 1.5 °C. The animals were kept in darkness for 24 h after light exposure. The mice were then returned to a normal light/dark cycle. At 2 and 4 days after light exposure mice were treated with BrdU and PGRN or vehicle. After 5 days eyes were enucleated and used for immunohistochemistry. Images were acquired using a confocal microscope (FLUOVIEW FV10i; Olympus, Tokyo, Japan). For quantitative data images were taken 500 μm superior from the optic nerve. The total number of immunoreactive cells was counted within the entire area of the image (211.968 × 211.968 μm). The number was calculated as number/mm. In situ hybridization. Retinas Switzerland)] was used for the making of RNA probes. RNA probes were hydrolyzed by carbonate buffer for 60 min and diluted in hybridization buffer. Retinal sections on slides were pretreated by proteinase K for 4 min. After washing slides, the sections were dried in air at least 1 hour. Sections were hybridized with probes in hybridization buffer overnight at 65 °C in a humidified box. Next, RNase was used for the elimination of unnecessary RNA. Then, blocking was performed with 1X Maleate/0.05% Triton/1 X Denhardt's solution (Sigma-Aldrich) for 2 hours. Sections were incubated with the anti-DIG antibody (1:5000 dilution) (Roche Diagnostics) overnight at room temperature. After the incubation, NBT/BCIP solution (pH 9.5) was used for the color reaction. When color reaction was completed, BrdU and Rx staining was performed according to above method. Mouse Rx primer sequence (5′ -3′ ) Forward: GCTTCTCGCTCGCTGGCCAC Reverse: CTTCCAGCGAGAACTTGTCC In vitro immunostaining. The primary retinal cultures were fixed with 4% paraformaldehyde at room temperature for 15 min. The cells were then incubated with 0.2% Triton X-100 (Bio-Rad Labs) in PBS for 10 min and 50 mM glycine (Wako) in PBS for 15 min. The cells were blocked with 3% goat serum or horse serum (Vector Labs) for 30 min and incubated with the primary antibodies overnight at 4 °C. The cells were then incubated for 1 h with secondary antibodies and counterstained with Hoechst 33342 (Invitrogen). Finally, the cells were mounted in Fluoromount (Diagnostic BioSystems) and imges were taken using a confocal microscope (FLUOVIEW FV10i). For quantitative data, the images were obtained within the entire area of the image (211.968 × 211.968 μm). The number of immunoreactive cells was counted and calculated as the ratio of immunoreactive cells to total cells.

In vivo
The following antibodies were used: mouse anti-rhodopsin ( Western blot analysis. Primary retinal cells or mice retinas were lysed using a buffer (RIPA buffer; Sigma-Aldrich) containing protease (Sigma-Aldrich) and a phosphatase inhibitor cocktail (Sigma-Aldrich). To extract retinal protein, the tissue was homogenized in cell-lysis buffer using a Physcotron homogenizer (Microtec Co., Ltd., Chiba, Japan). The cell lysate was centrifuged at 12,000 × g for 20 min, and the supernatant was used for subsequent experiments. Protein concentration was measured by comparison with a known concentration of BSA using a bicinchoninic acid protein assay kit (Thermo Scientific). A mixture of equal parts of protein and sample buffer with 10% 2-mercaptoethanol (Wako Pure Chemical Industries, Ltd.) was subjected to SDS-PAGE using 5-20% gradient gels (Wako Pure Chemical Industries, Ltd.). The separated proteins were transferred onto a polyvinylidene difluoride membrane (Immobilon-P: Millipore Corporation, Billerica, MA, USA). After blocking for 30 min at room temperature with Block One-P (Nacalai Tesque, Inc., Kyoto, Japan), membranes were washed in 10 mM Tris-buffered saline containing 0.05% Tween 20 and then incubated with the primary antibody overnight at 4 °C. The following primary antibodies were used: mouse anti-rhodopsin (1:1000 dilution: Millipore), sheep anti-PGRN (1:100 dilution: R&D systems), and mouse anti-β -actin (1:2,000 dilution: Sigma-Aldrich). After exposure to the primary antibody, the membranes was incubated with peroxidase goat anti-rabbit, goat anti-mouse or rabbit anti-sheep IgG (Thermo Scientific) as the secondary antibody. The immunoreactive bands were visualized using an ImmunoStar LD (Wako Pure Chemical Industries, Ltd.).
Histological analysis. PGRN-knockout and WT mice eyes were enucleated and fixed in 4% paraformaldehyde for 24 h at 4 °C. Six paraffin-embedded sections (5 μm thickness) cut through the optic disc of each eye were prepared in a standard manner and stained with hematoxylin and eosin. Six sections from each eye were used for the morphometric analysis. Light microscopy images were taken and the thickness of the ONL was measured at 240 μm intervals from the optic disc in a blind manner by H.I. The data was averaged for each eye.