Paired Immunoglobulin-like Receptor B Inhibition in Müller Cells Promotes Neurite Regeneration After Retinal Ganglion Cell Injury in vitro

Abstract

In the central nervous system (CNS), three types of myelin-associated inhibitors (MAIs) have major inhibitory effects on nerve regeneration. They include Nogo-A, myelin-associated glycoprotein, and oligodendrocyte-myelin glycoprotein. MAIs possess two co-receptors, Nogo receptor (NgR) and paired immunoglobulin-like receptor B (PirB). Previous studies have confirmed that the inhibition of NgR only results in a modest increase in regeneration in the CNS; however, the inhibitory effects of PirB with regard to nerve regeneration after binding to MAIs remain controversial. In this study, we demonstrated that PirB is expressed in primary cultures of retinal ganglion cells (RGCs), and the inhibitory effects of the three MAIs on the growth of RGC neurites are not significantly decreased after direct PirB knockdown using adenovirus PirB shRNA. Interestingly, we found that retinal Müller cells expressed PirB and that its knockdown enhanced the regeneration of co-cultured RGC neurites. PirB knockdown also activated the JAK/Stat3 signaling pathway in Müller cells and upregulated ciliary neurotrophic factor levels. These findings indicate that PirB plays a novel role in retinal Müller cells and that its action in these cells may indirectly affect the growth of RGC neurites. The results also reveal that PirB in Müller cells affects RGC neurite regeneration. Our findings provide a novel basis for the use of PirB as a target molecule to promote nerve regeneration.

Electronic supplementary material

The online version of this article (10.1007/s12264-020-00510-w) contains supplementary material, which is available to authorized users.

Introduction

Determining a means of promoting the repair and regeneration of the injured central nervous system (CNS) has long been a research hotspot and a medical challenge. The optic nerve is a unique CNS component that has difficulty regenerating after damage. Active treatment methods for diseases that affect the optic nerve remain scarce, and this lack of options frequently results in visual dysfunction or even blindness in affected individuals [1]. Myelin-associated inhibitors (MAIs), which are abundant in the microenvironment, are responsible for the regeneration difficulties following optic nerve damage. Myelin sheaths typically surround optic nerve axons and are exposed to a large number of MAIs after damage. These MAIs specifically bind to receptors on the axons, resulting in an unstable cytoskeleton for the growth of cone-plate pseudopods and filopodia, ultimately hindering regeneration [2].

Among these MAIs, Nogo-A, myelin-associated glycoprotein (MAG), and oligodendrocyte-myelin glycoprotein (OMgp) play major roles in the inhibition of axon regeneration [3]. In the CNS, they have two known co-receptors, Nogo receptor (NgR) and paired immunoglobulin-like receptor B (PirB) [4–8]. The MAIs facilitate the collapse of neuronal growth cones after they bind to NgRs, ultimately resulting in actin depolymerization in cells. Interestingly, myelin-associated inhibition of NgRs in the context of CNS regeneration is very limited. Indeed, the knockdown of NgR alone does not significantly reduce the inhibitory effects of MAIs on axon growth [9, 10].

PirB is widely expressed in a number of immune cells such as macrophages, B cells, and dendritic cells, and this protein functions as a major histocompatibility complex class I (MHC I) molecule receptor. After binding to MHC I molecules, PirB negatively regulates cells by recruiting the tyrosine protein phosphatase of Src oncogene homologous domains 1 and 2 (SHP-1 and SHP-2) to the intracellular immunoreceptor tyrosine-based inhibition motif segment [11, 12]. In the CNS, PirB is mainly expressed in the axons and synapses of neurons in the cerebral cortex, hippocampus, cerebellum, retina, and optic nerve [13, 14]. After binding to PirB, MAIs in the CNS inhibit axon growth through the PirB-SHP1/SHP2 and PirB-POSH (plenty of SH3s)-myosin IIA pathways [15, 16]. A previous study has demonstrated that antagonizing PirB reduces the inhibitory effect of Nogo-A on axon growth [8]; however, whether PirB promotes axonal regeneration after MAI inhibition in the optic nerve remains unclear.

In the retina, Müller cells are closely associated with retinal ganglion cells (RGCs). Müller cells constitute a “bridge” that provides a link between neurons and the microenvironment [17]. This contact allows retinal neurons to exchange substances with extracellular blood vessels, the vitreous cavity, and the subretinal space [18]. Many nerve growth factors secreted by Müller cells affect RGC axon growth [19–21]. Our preliminary study demonstrated that PirB is also expressed in Müller cells; however, its specific role remains unclear. In addition, it is unknown if PirB affects RGC axon growth.

In this study, we used adenovirus (AD) PirB shRNA to knock down PirB expression in primary RGCs and in Müller cells. Subsequently, we examined the growth of neurites after direct PirB silencing in RGCs, and monitored the indirect effect of the PirB-silenced Müller cells on co-cultured RGC neurites. We further investigated if the indirect effect of PirB in Müller cells was related to the promotion of endogenous ciliary neurotrophic factor (CNTF) expression, and the associated signaling pathways were also explored. The results clarify the role of PirB in Müller cells and provide a basis for studying regeneration after optic nerve damage by targeting PirB in Müller cells.

Experimental Procedures

Experimental Animals

Sprague-Dawley rat pups 1–2 days old were used for primary cell culture. All experimental animal procedures complied with the Association for Research in Vision and Ophthalmology statement and were approved by the Medical Ethics Committee of Daping Hospital, Army Medical University. The animals were purchased from the Animal Center of Daping Hospital, Army Medical University.

Experimental Reagents

The reagents used in this study included fetal bovine serum (Gibco, Newcastle, Australia, Lot No: 10100147), Neurobasal culture medium (Gibco, NY, USA, Lot No: 2085380), B27 (Gibco, NY, USA, Lot No: 2042265), Nogo-A, MAG, OMgp (R&D, MN, USA), Rotenone (Sigma, St. Louis, USA, Lot No: MKBZ2534V), and an Ultra-high-sensitivity ECL kit (Medchem Express, Monmouth Junction, NJ, USA, Cat No: HY-K1005). The antibodies used are listed in Table 1.

Table 1.

List of primary and secondary antibodies used for western blot (WB), immunohistochemistry (IHC), and immunocytochemistry (ICC).

Antibody	Source	Lot/Cat No.	Dilution
Primary antibodies			WB	IHC/ICC
Goat anti-PirB	R&D, MN, USA	AF2754	1:1000	1:100
Rabbit anti-phospho Stat3 Y705	Abcam, Cambridge, UK	ab76315	1:5000
Mouse anti-GAPDH	Abcam	ab8245	1:5000
Rabbit anti-SHP2	Abcam	ab187040	1:5000
Rabbit anti-phospho SHP2 Y542	Abcam	ab17939	1:1000
Rabbit anti-phospho SHP1 Tyr536	ThermoFisher, MA, USA	PA5-36682	1:750
Rabbit anti-SHP1	Abcam	ab32559	1:1000
Rabbit anti-CNTF	Abcam	ab175387	1:1000
Rabbit anti-beta III tubulin	Abcam	ab18207		1:500
Mouse anti-beta III tubulin	Abcam	ab78078		1:500
Rabbit anti-vimentin	Abcam	ab92547		1:500
Mouse anti-glutamine synthetase	Abcam	ab64613		1:200
Mouse anti-CD90/Thy1	Abcam	Ab225		1:500
Mouse anti-Brn-3a	Merck Millipore, Watford, UK	MAB1585		1:50
Secondary antibodies
HRP-labelled anti-mouse IgG	Invitrogen, Poole, UK	G-21040	1:10000
HRP-labelled anti-rabbit IgG	Invitrogen	31460	1:10000
HRP-labelled anti-donkey IgG	R&D	HAF109	1:1000
Alexa488 anti-goat IgG	Invitrogen	A32814		1:300
Alexa488 anti-mouse IgG	Invitrogen	A32723		1:300
Alexa594 anti-rabbit IgG	Invitrogen	A32740		1:300

Viruses and Sequences

The target sequences for Si-r-CNTF and AD PirB shRNA (with GFP tag) were GGCTTACCGTACCTTCCAA (Ribobio Co., Ltd., Guangzhou, China) and GGAGCCGAACTTTATTGTCTCTATA (Hanbio Co., Ltd., Shanghai, China).

RGC Culture

Sprague-Dawley rats aged 1–2 days were euthanized by intraperitoneal injection of 2% pentobarbital. The retinas were separated under a high-power microscope and centrifuged at 1100 ×g for 5 min. Digestion medium (0.2% papain, 10% DNase, and 1% glutamine) was added for digestion in an incubator containing 5% CO₂ at 37°C for 20 min. An appropriate amount of Neurobasal complete medium (Neurobasal medium supplemented with 2% B27, 1% glutamine, and 0.5% penicillin–streptomycin) was added to prepare a cell suspension, which was filtered using a 70-μm cell strainer. The cells were inoculated on plates coated with polylysine (PDL) at a density of 3×10⁵ /mL and cultured in an incubator under 5% CO₂ at 37°C. Cell adherence was observed the next day, and the medium was replaced once for an additional 6–8 days of culture [22, 23].

Müller Cell Culture

Müller cells were extracted according to the previously reported methods [24, 25]. Briefly, retinas were digested in 0.25% trypsin for 15 min, and complete Dulbecco’s modified Eagle’s medium [DMEM; containing 1% penicillin–streptomycin and 10% fetal bovine serum (FBS)] was added. The resulting cell suspension was filtered using a 70-μm cell strainer. Cells at a density of 3×10⁵ cells/mL were grown in medium under 5% CO₂ at 37°C. The next day, the medium was replaced prior to 6–8 days of continuous culture. Passaging was performed when the cell density reached 80%. Subsequent experiments were performed after cells were passaged 3–4 times.

Cellular Immunofluorescence

Cells were fixed in 4% paraformaldehyde for 15 min and then washed three times for 5 min with phosphate-buffered saline (PBS). The cells were permeabilized using 0.1% Triton for 15 min, and then blocked with 10% goat serum for 30 min. Primary antibodies (mouse anti-β-III tubulin, rabbit anti-vimentin, and mouse anti-Brn 3a) were added prior to incubation at 4°C overnight. After three PBS washes, fluorescent secondary antibodies were added and left at room temperature (20–22°C) for 1 h in the dark; this was followed by counterstaining with 4′, 6-diamidino-2-phenylindole (DAPI) for 15 min. Five regions in each well were randomly selected and imaged under a fluorescence microscope (Leica, Weztlar, Germany). The lengths of 20 neurites with adjacent RGCs were measured using Image-Pro plus 6.0 (Media Cybernetics, Maryland, USA) to assess their growth [26, 27]. Cells exhibiting positive vimentin or Brn3a expression were counted to determine the purity of the primary cells. Müller/RGC purity = number of cells positive for vimentin or Brn3a expression / total number of cells positive for DAPI nuclear staining × 100% [28, 29].

Transfection of Primary Cells with AD PirB shRNA

Müller cells and RGCs were seeded into 6- or 24-well plates at 3×10⁵ cells/mL and cultured under 5% CO₂ at 37°C for 24 h. The medium was refreshed and AD PirB shRNA (with GFP tag) was added gradually according to the manufacturer’s instructions to achieve final titers of 10⁸, 10⁷, and 10⁶ plaque-forming units (PFU)/mL. The medium was not replaced during transfection. In addition, AD shRNA was added to the control group, and the blank group was left untreated. Three duplicate wells were used for each group. Cell growth and GFP expression were observed under a microscope at 1, 3, and 7 days after transfection. Müller cells and RGCs were transfected at 10⁶ PFU/mL in all subsequent experiments.

Transfection of Müller Cells Using Si-r-CNTF

Müller cells were passaged 3–4 times and then seeded on a 24-well plate at a density of 3×10⁵ cells/mL. These cells were then transfected with AD PirB shRNA. The next day, empty vector Si-r (control) and Si-r-CNTF were added at final concentrations of 10, 20, 30, 50, and 100 nmol/L. The interference efficiency was assessed, and Si-r-CNTF was used at 50 nmol/L in all subsequent experiments. After 48 h–72 h of transfection, cells were repeatedly transfected with Si-r-CNTF to consolidate the interference effect.

EdU (5-ethynyl-2´-deoxyuridine) Assay of Müller Cells

Müller cells were passaged 3–4 times and then seeded into a 24-well plate at 3×10⁵ cells/mL. These cells were divided into the following groups: blank (without treatment), control (with empty virus), and PirB shRNA (at 10⁸, 10⁷, and 10⁶ PFU/mL). Three duplicate wells were used for each group, and cells were cultured in an incubator under 5% CO₂ at 37°C for 72 h. Based on the kit instructions, 300 μL medium containing EdU A solution was added to each well and left for 2 h. Then, the cells were fixed in 4% paraformaldehyde for 30 min, followed by neutralization with 2 mg/mL glycine for 15 min. The cells were permeabilized using 0.5% Triton at room temperature for 30 min, and this was followed by nuclear staining with F solution for 15 min, three PBS washes for 5 min each, and then imaging.

Inhibition of Primary RGCs by Nogo, MAG, and OMgp

RGCs were divided into the following groups: blank (culture plates coated with 0.1 mg/mL PDL), PBS (plates coated with 0.1 mg/mL PDL+ PBS at an amount equal to that of the MAI group), and MAI (plates coated with 0.1 mg/mL PDL + Nogo, MAG, or OMgp in PBS). The concentrations of MAIs were 0.25, 0.5, 1.0, and 2.0 µg/mL. Three duplicate wells were prepared for 7 days of culture. Then, the culture medium was removed, and β-III tubulin was stained by immunofluorescence to assess the lengths of RGC neurites. In the MAI group, the coating concentrations of Nogo, OMgp, and MAG were 0.5, 0.5, and 1 µg/mL, respectively, in subsequent experiments.

Establishment of the RGC Injury Model

Rotenone is a mitochondrial complex 1 inhibitor that induces neuronal cytotoxicity and is commonly used to prepare injury models for in vitro assays [30]. Primary RGCs were cultured in 24-well plates for 3 days, the medium was then replaced and rotenone (100 nmol/L) was added as previously described [31]. After rotenone injury, the surviving RGCs were detected using an MTT assay (Solarbio, Beijing, China, Lot No:M1020) as directed by the manufacturer. Briefly, blank wells without cells were used for zero-adjustment, and the control and rotenone groups (injuring RGCs for 6 h, 12 h, 24 h, 48 h, and 72 h) were prepared. Three duplicate wells were used for each group, and the culture supernatant was replaced with 360 μL fresh medium. Then, 40 μL MTT solution was added and left for 4 h. After careful removal of the supernatant, 490 μL dimethyl sulfoxide was added to each well to dissolve the formazan crystals. The cells were incubated on a shaker at low speed for 10 min, and the absorbance was measured at 490 nm. RGCs injured for 24 h were selected for subsequent experiments.

Co-culture of Müller Cells and RGCs

Müller cells were passaged, inoculated into transwell chambers, then transfected with AD PirB ShRNA or CNTF siRNA. After 3 days, the transwell chambers were extracted, the medium was removed, and the cells were washed three times with PBS. After adding fresh medium, the cells were transferred into RGC culture plates in the presence of renewed solution for co-culture [27]. Two duplicate wells were used for each group, and the experiment was repeated three times. Isolated Müller cells and RGCs were cultured separately. There was no contact between Müller cells and RGCs except via the culture medium. Using the 24-well plate as an example, Müller cells were cultured in the transwell, and 0.5 mL DMEM containing 10% FBS and 1% penicillin–streptomycin was added. Primary RGCs were cultured outside the transwell in 0.6 mL Neurobasal medium containing 2% B27, 1% glutamine, and 0.5% penicillin–streptomycin. The cells were co-cultured in an incubator under 5% CO₂ at 37°C.

Real-Time PCR

RNA was extracted as previously described [32] and reverse-transcribed into cDNA according to the kit instructions (TAKARA BIO. INC, Beijing, China, Lot No: RR047A). A total of 2 µL of nuclease-free and high-purity water, 5 µL of Mix and oligo (dT) 18 primers, 0.5 µL each of forward and reverse primers, and 2 µL of cDNA were used for the reaction. β-actin was used as an internal reference. The primers were designed based on GenBank and included: PirB, forward AGGATGGAGTGGAGCTGAAC and reverse TGATTGTTTGCTCCTTGGCC; CNTF, forward CTTTCGCAGAGCAAACACCTC and reverse ACTGTGAGAGCTCTTGAAGGAC; and GAPDH, forward GACATCAAGAAGGTGGTGAAGC and reverse TGTCATTGAGAGGACCTGCCAGC. The qPCR conditions were pre-denaturation at 95°C for 10 min, denaturation at 95°C for 15 s (cycled 39 times), annealing at 57°C for 1 min, and extension at 65°C for 5 min. Relative quantification was performed to obtain real-time PCR results.

Western Blot

Immunoblotting was performed as previously described [33]. Briefly, equal amounts of protein (40 µg) were separated by 10% SDS-PAGE and transferred onto polyvinylidene fluoride membranes. After blocking with 5% skimmed milk for 2 h, primary antibodies were added prior to incubation on a shaker at 4°C overnight. Then, secondary antibodies were added and left at room temperature for 1 h. Enhanced chemiluminescence was used for development prior to detection on an Omega Lum G gel imager (Aplegen, Pleasanton, USA). Gray values were analyzed using ImageJ software (National Institutes of Health, Bethesda, Germany), and β-actin or GAPDH was used as an internal reference. The ratio of the gray value of the target strip to that of the internal reference was calculated.

Statistical Analysis

Data were analyzed using SPSS 19.0 (IBM, Armonk, USA). Cell counts are expressed as percentages. Continuous data are expressed as the mean ± SD. Student’s t test or one-way ANOVA was used in cases with homogeneity of variance and normal distribution, and for all other cases, a nonparametric test was used. P < 0.05 was considered statistically significant.

Results

PirB Knockdown in RGCs does not Promote the Growth of Neurites

After primary culture of RGCs for 24 h, the cells were observed by microscopy to be adherent. Some cells aggregated, and a few single cells were scattered. After 7–8 days in culture, RGCs had large numbers of neurites. The RGCs were specifically labelled with Brn-3a, and they accounted for > 85% of all cells (RGC purity = Brn-3a / DAPI × 100%) (Fig. 1A). They also expressed PirB, and this expression occurred primarily in the cell bodies and neurites (Fig. 1B).

Fig. 1.

Effects of RGC PirB knockdown on neurite growth. A Representative images of a primary culture of RGCs, where Brn-3a (green, 488 nm) is primarily localized within the nucleus (blue, DAPI nuclear staining; arrows, non-RGC nuclear staining; arrowheads, RGC-positive cells; scale bars, 100 μm). B Image of co-localization of RGC PirB (green at 488 nm) with β-III tubulin (red at 594 nm) (blue DAPI nuclear staining; scale bars, 250 μm). C MTT assay results showing cytotoxicity of AD PirB shRNA in RGCs. D PirB mRNA levels after RGC transfection with AD PirB shRNA as evaluated by real-time PCR. E Inhibitory effects of Nogo, MAG, and OMgp on RGC neurite growth. F Effects of AD PirB shRNA on RGC neurite growth after pretreatment with PBS, Nogo (0.5 µg/mL), MAG (1.0 µg/mL), and OMgp (0.5 µg/mL). G Representative images of fluorescent staining specific for RGC neurite growth [red (594 nm), β-III tubulin-labeled RGC neurites; green (488 nm), GFP indicating autofluorescence of AD PirB shRNA or empty virus in RGCs; blue, DAPI-labeled nuclei; scale bars, 250 μm]. All data are the mean ± SD. The experiment was repeated three times (n = 9). *P < 0.05, **P < 0.01, one-way ANOVA or Student’s t test.

After transfection with AD, >55% of RGCs were GFP-positive. MTT assays showed no significant difference in cell number from that of the blank (negative control, 100% ± 20%, P = 0.23) and control (empty virus, 90% ± 13.45%, P = 0.62) groups when AD PirB shRNA was transfected at 10⁶ PFU/mL (83.33% ± 12.22%; Fig. 1C). Cytotoxicity against RGCs was enhanced, and the cell number was decreased at the final transfection concentrations of 10⁷ (67.33% ± 7.63%, P = 0.048) and 10⁸ PFU/mL (42% ± 21.66%, P = 0.004) compared to controls (Fig. 1C). Subsequent experiments were performed using a transfection concentration of 10⁶ PFU/mL. As assessed by qPCR (Fig. 1D), the PirB mRNA levels in cells transfected with AD PirB shRNA at 10⁶ PFU/mL (59 ± 18.68) were significantly lower than those of the control groups (108 ± 13.11, P = 0.014).

The inhibitory effects of Nogo, MAG, and OMgp on RGC axonal growth were concentration-dependent. The growth of RGC neurites was significantly inhibited by Nogo and OMgp at 0.5 µg/mL and by MAG at 1 µg/mL (94.33 ± 13.65 µm, P = 0.017; 90.67 ± 21 µm, P = 0.012; 77.33 ± 10.26 µm, P = 0.013) compared to that of the PBS group (138.33 ± 23.63 µm, Fig. 1E). AD PirB shRNA reduced the PirB expression in RGCs but did not significantly promote the regeneration of neurites in RGCs treated with Nogo, MAG, or OMgp compared to the control group transfected with empty virus. In addition, the lengths of the neurites were not significantly different (P > 0.05, Fig. 1F, G).

PirB Knockdown in Müller Cells Increases CNTF Expression and Activates the Janus Kinase/Signal Transducer and Activator Of Transcription (JAK/Stat3) Signaling Pathway

After primary Müller cells were passaged three times, >95% of the cells expressed the Müller cell-specific protein vimentin and glutamine synthetase (Müller cell purity = vimentin / DAPI×100%) (Fig. 2A and Fig. S1). We found for the first time that Müller cells expressed PirB, and this expression occurred primarily in the cell membrane and cytoplasm (Fig. 2B). EdU assays revealed that cell proliferation decreased with increasing transfected virus concentration. There was no significant effect on the proliferation of Müller cells at a viral PirB shRNA titer of 10⁶ PFU/mL (relative to blank at 100%; 79.33% ± 7.02%, P = 0.516). At transfection titers of 10⁷ PFU/mL (53.33% ± 10.02%, P = 0.006) and 10⁸ PFU/mL (35.34% ± 13.5%, P = 0.001), cell proliferation in the AD PirB shRNA groups decreased significantly compared to that of the control (empty virus, 84.67% ± 6.11%, Fig. 2D and Fig. S2).

Fig. 2.

Treatment of primary Müller cells with AD PirB shRNA for knockdown. A Representative image showing fluorescence-based identification of Müller cells with vimentin (green, 488 nm) located primarily in the cytoplasm (blue, DAPI nuclear staining; scale bar, 250 μm). B Images showing co-localization of vimentin and PirB in a Müller cell. Vimentin (red, 594 nm) was primarily found in the cytoplasm, and PirB (green, 488 nm) was located mostly in the cell membrane (blue, DAPI nuclear staining; scale bars, 250 μm). C GFP expression in Müller cells transfected with AD PirB shRNA over time (red, Müller cells labelled with vimentin; green, GFP fluorescence showing positive expression of AD PirB shRNA; blue, DAPI nuclear staining; scale bars, 250 μm). D Effects of AD PirB shRNA transfection on Müller cell proliferation (n = 3). E PirB mRNA expression in Müller cells (n = 3). F PirB protein expression in Müller cells (n = 3). *P < 0.05, **P < 0.01, one-way ANOVA and Student’s t test.

After the cells were transfected with AD at a final titer of 10⁶ PFU/mL for 1 day, scattered green fluorescence (GFP) was observed, and this increased significantly at 3 days (30%–40%), and 7 days (50%–60%) (Fig. 2C). After Müller cells were transfected with viruses for 1 day (0.87 ± 0.16), the PirB mRNA levels were lower than those of the control group (1.18 ± 0.32). The decrease at 3 days was statistically significant (0.54 ± 0.11, P < 0.028), and the lowest level was reached at 7 days (0.20 ± 0.07, P = 0.007) (Fig. 2E). Compared to the level in the control group (0.55 ± 0.02), PirB protein expression significantly decreased to the lowest value at 7 days after Müller cells were transfected with the virus (0.32 ± 0.05, P < 0.05; Fig. 2F).

Meanwhile, mRNA level detection of the three main nerve growth factors secreted by Müller cells (NGF, BDNF, and CNTF) showed that the NGF and CNTF mRNA levels were significantly higher in the AD PirB shRNA group (1.92 ± 0.14, P = 0.041, P = 0.036; 2.17 ± 0.18, P = 0.298, P = 0.018) than in the blank and control groups (0.76 ± 0.10, 1.23 ± 0.14; 0.75 ± 0.1, 1.12 ± 0.09; Fig. 3A). The NGF and BDNF protein levels did not significantly differ (P > 0.05; Fig. 3B, b and D, d); however, the CNTF protein levels in the AD PirB shRNA groups (1.48 ± 0.12) were significantly higher than those of the blank and control groups (0.44 ± 0.017, P < 0.001; 0.67 ± 0.13, P < 0.001; Fig. 3C, c). These results suggested that PirB knockdown in Müller cells mainly results in the increased expression of intracellular CNTF.

Fig. 3.

NGF, CNTF, and BDNF levels in primary Müller cells transfected with AD PirB shRNA. A NGF, CNTF, and BDNF mRNA expression in Müller cells (n = 3). B, b NGF protein expression in Müller cells (n = 3). C, c CNTF protein expression in Müller cells (n = 3). D, d BDNF protein expression in Müller cells (n = 3). *P < 0.05, **P < 0.01, one-way ANOVA and Student’s t test.

To explore the signaling pathways affecting CNTF expression in Müller cells in the context of PirB, we investigated the protein expression levels of signaling molecules downstream of PirB. Western blot analysis showed that the SHP1 and SHP2 protein levels were not significantly different (Fig. 4A, C) and that the P-SHP1 levels were significantly lower (0.36 ± 0.08) than those of the blank and control groups (0.88 ± 0.06, P = 0.031; 0.72 ± 0.15, P = 0.024, Fig. 4B). The P-SHP2 levels were significantly higher (1.01 ± 0.13) than those of the blank and control groups (0.47 ± 0.03, P = 0.038; 0.59 ± 0.06, P = 0.0294, Fig. 4D). In addition, the P-Stat3 levels were significantly higher (0.81 ± 0.12) than those of the blank and control groups (0.46 ± 0.10, P = 0.0218; 0.43 ± 0.09, P = 0.0189, Fig. 4E). These findings suggested that the JAK/Stat3 signaling pathway is activated in Müller cells after transfection with AD PirB shRNA.

Fig. 4.

Levels of intracellular proteins in primary Müller cells transfected with AD PirB shRNA. GAPDH and β-actin served as internal references. The y-axes show the ratios of gray values of target proteins to those of the respective internal references. Each experiment was repeated three times (n = 9). Data are the mean ± SD; *P < 0.05, **P < 0.01, one-way ANOVA and Student’s t test.

PirB Knockdown in Müller Cells Promotes Axonal Growth of Co-cultured RGCs

The transwell system (Corning Incorporated, NY, USA, REF No: 3413) allowed cells inside the chamber to be physically separated from those outside so that only the culture medium connected them (Fig. 5C). Primary RGCs were cultured alone for 1 day, then co-cultured with differently-treated Müller cells for 7 days. Neurites in the blank group with RGCs cultured alone were the shortest, and those of the AD PirB shRNA group were the longest (202.57 ± 23.59 µm, Fig. 5A, B). These differed significantly from the blank (negative control, 166.23 ± 14.86 µm, P = 0.005), sham (181.54 ± 11.13 µm, P = 0.041), and control groups (173.01 ± 22.82 µm, P = 0.006). These results suggested that Müller cells promote the growth of co-cultured RGC neurites, and PirB knockdown further enhances this neurite growth-promotion effect.

Fig. 5.

Effects of Müller cells transfected with AD PirB shRNA on co-cultured RGCs. A Representative images of immunofluorescent neurite staining after co-culture of RGCs with differently-treated Müller cells [green (488 nm), β-III tubulin-labeled RGC neurites; blue, DAPI-labeled RGC nuclei; scale bars, 250 μm; n = 6 for each group]. B Quantification of RGC neurite lengths. Data are the mean ± SD; *P < 0.05, **P < 0.01, one-way ANOVA and Student’s t test. C Diagram of the co-culture of Müller cells and RGCs. Müller cells in the inner chamber; primary RGCs outside of the chamber; the Müller cells and RGCs were not in contact, but the medium was continuous with both.

In Müller Cells with PirB Knockdown, CNTF is a Major Factor That Promotes the Regeneration of RGC Neurites

MTT assays revealed significant cytotoxicity after RGCs were injured for 24 h (61.33% ± 5.03%) by exposure to rotenone at 100 nmol/L compared to that of the control group (100% ± 14%, P = 0.001, Fig. 6A). When injured RGCs were co-cultured with pretreated Müller cells for 4 days, immunofluorescent staining revealed that PirB knockdown in Müller cells promoted the regeneration of neurites in the co-cultured RGCs. In addition, in the siRNA CNTF + AD PirB shRNA group, the regeneration of neurites was significantly inhibited (55.45 ± 15.01 µm) compared with that in the AD PirB shRNA group (118.39 ± 13.57 µm, P = 0.001) and the siRNA CNTF + AD shRNA group (79.86 ± 8.30 µm, P = 0.044) (Fig. 6B, C).

Fig. 6.

Effects of Müller cells transfected with AD PirB shRNA on injured RGCs. A Toxic effects of rotenone on RGCs (control, untreated RGCs). B Neurite growth of injured RGCs co-cultured with Müller cells transfected with AD PirB shRNA (RGCs injured for 24 h by rotenone, Müller cells pretreated with AD PirB shRNA for 72 h, injured RGCs co-cultured with pretreated Müller cells for 4 days, and Si-r CNTF interference in Müller cells for 4 days, n = 6). Data are the mean ± SD. *P < 0.05, **P < 0.01, one-way ANOVA. C Fluorescent staining of injured RGC neurites [green (488 nm), β-III tubulin-labelled RGC neurites; blue, DAPI-labelled RGC nuclei; scale bars, 250 μm; n = 6 per group].

Discussion

MHC-1 in the CNS is either minimally expressed or not expressed in adulthood, so the effects of PirB in the CNS differ from those in the immune system. A previous study showed that NgR and PirB are co-receptors for Nogo-A, MAG, and OMgp; however, PirB inhibition is more effective in promoting axonal growth than NgR suppression [8].

In this study, Nogo-A, MAG, and OMgp were used to inhibit primary RGCs, and AD PirB shRNA was used for PirB knockdown. PirB knockdown did not promote the growth of RGC neurites after treatment with Nogo-A, MAG, or OMgp. These results suggested that PirB plays a limited role in the myelin-associated inhibition of RGCs, and inhibition of PirB is not sufficient to significantly promote the regeneration of neurites. There are two likely reasons for these findings. (1) PirB is not the main MAI receptor in inhibiting axonal growth, and other receptors, such as NgR, may play the main inhibitory role; however, previous findings do not support this notion [9, 10]. (2) In RGCs, there are other unidentified inhibitors that may not act through the PirB pathway, so PirB knockdown cannot completely block their effects.

Although direct PirB knockdown in RGCs did not significantly promote the growth of neurites under MAI inhibition, that does not necessarily mean that PirB does not affect optic nerve regeneration. On the contrary, PirB is very important in optic nerve regeneration. A study by Fujita et al. showed that PirB knockout in mice does not directly promote optic nerve regeneration [15], but optic nerve regeneration is stronger in mice with PirB knockdown than in wild-type animals after treatment with exogenous BDNF, suggesting that PirB plays a role in optic nerve regeneration that may not depend on MAI inhibition and is instead caused by other factors that affect neurite growth. Our study demonstrated for the first time that PirB is expressed in retinal Müller cells and that PirB knockdown increases CNTF expression. This suggests that PirB regulates the expression of NGF in Müller cells. Based on this, we infer that PirB also has important physiological functions in non-neuronal cells.

Müller cells are an important component of the microenvironment, as they are neuroprotective and provide nutritional support for RGCs [34]. Here, we revealed that Müller cells with PirB knockdown not only promoted normal RGC neurite growth, but also induced the growth of rotenone-injured RGC neurites when Müller cells and RGCs were co-cultured. There was no significant protective effect on RGCs, however, suggesting that inhibiting PirB in Müller cells primarily promotes the growth of RGC neurites. In addition, treatment of Müller cells with siRNA-CNTF significantly suppressed neurite regeneration, suggesting that this effect is indirectly influenced by the regulation of CNTF expression.

We also revealed that PirB knockdown in Müller cells activated the JAK/Stat3 signaling pathway, and while P-SHP1 levels decreased significantly, P-SHP2 levels markedly increased. Han et al. [35] reported that SHP-1 and SHP-2 play negative regulatory roles in the JAK/Stat3 pathway, where their downregulation activates JAK/Stat3 signaling. In this study, the P-SHP1 protein levels were significantly down-regulated and the P-SHP2 protein levels were significantly up-regulated. The expression of these two negative regulatory molecules was inconsistent; however, the JAK/Stat3 signaling pathways were activated. Therefore, we speculate that SHP-1 has a stronger effect than SHP-2 in regulating the signaling pathways. A strategy that incorporates the use of an si-SHP1/2 construct could provide a convincing means of testing this hypothesis; however, further research is required.

We found that PirB is expressed in Müller cells and that its knockdown promotes neurite regeneration in co-cultured RGCs. This regenerative effect may be exerted via the PirB-SHP1-JAK/Stat3-CNTF pathway; however, it is unclear if PirB affects the production of other factors in Müller cells. In addition, we used cells from early postnatal animals in which certain intrinsic factors such as cAMP and NMDA are different from those in adult retinal cells. Thus, age-related differences in the role of PirB require further investigation.

In conclusion, these findings reveal that PirB in Müller cells exerts important effects on RGC neurite regeneration, and our results provide a new basis for the use of PirB as a target molecule to promote nerve regeneration.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Conflict of interest

The authors declare that they have no conflict of interest.