Wnt signaling induces neurite outgrowth in mouse retinal ganglion cells

Abstract

Wingless-type (Wnt) signaling pathways mediate axonal growth and remodeling in the embryonic optic nerve, brain and spinal cord. Recent studies demonstrated that the canonical Wnt/β-catenin signaling pathway also induces axonal regeneration after injury in the optic nerve of adult animals. However, the molecular mechanisms of Wnt-mediated axonal growth are not well understood. Additionally, because Wnt signaling is stimulated in neurons as well as neighboring non-neuronal cells, the cell type(s) responsible for Wnt-induced axonal regeneration are not known. The objectives of this study were to investigate potential mechanisms and target cells of Wnt3a stimulated neurite growth using primary retinal ganglion cell (RGC) cultures. We demonstrated that Wnt3a ligand induced dose-dependent increases in average neurite length and number of neurites in RGCs. QPCR analysis of candidate mediators showed that Wnt3a-dependent neurite growth was associated with lower expression of Ripk1 and Ripk3 genes. Additionally, inhibiting Ripk1 signaling with Necrostatin-1s led to increased neurite number per cell but not increased neurite length. Therefore, Ripk signaling may be involved in mediating the effects of Wnt3a on neurite number but Ripk activity does not seem to be required for Wnt3a-dependent regulation of neurite length. This study shows that RGCs are direct cellular targets of Wnt3a-induced axonal growth, and we identified a novel association between Wnt signaling and Rip kinases in neurite formation.

1. Introduction

The canonical Wnt/β-catenin (“Wnt”) pathway is an essential signaling cascade that regulates a wide variety of processes in the developing CNS, including neuronal differentiation, axonal extension and stem/progenitor cell proliferation. Recent studies using animal models of neuronal and axonal injury demonstrated an important role for Wnt signaling in axonal growth and neuronal survival in adult tissue (reviewed in (Garcia et al., 2018)). For example, administration of the Wnt3a ligand induced axonal regeneration after optic nerve crush injury and spinal cord contusion injury in rodents (Yin et al., 2008; Patel et al., 2017). Similarly, in zebrafish spinal cord injury models, Wnt signaling promoted axonal regrowth (Yin et al., 2008; Briona et al., 2015), and over-expression of the Wnt inhibitor Dkk1 inhibited axonal regeneration and neuronal function (Strand et al., 2016). However, the molecular and cellular mechanisms of Wnt-induced neurite and axonal growth are mostly unknown (Garcia et al., 2018).

Wnt/β-catenin signaling is stimulated by Wnt ligands binding to the co-receptors LRP5/6 and Frizzled in the plasma membrane, leading to stabilization of β-catenin and its translocation to the nucleus. In the nucleus, β-catenin forms protein complexes with T-cell factor/lymphoid enhancer factor (TCF/LEF) transcription factors, which directs expression of specific Wnt target genes (Nusse, 2005). Damage to the retinal ganglion cells (RGCs) in the inner most layer is an important cause of vision loss in glaucoma and traumatic optic nerve injury. Evidence in animal models indicates that both neuronal and non-neuronal cells play roles in Wnt-induced axon regeneration after retina and spinal cord injury, but their precise contributions are unknown. Upregulated Wnt signaling in retinal neurons and Muller glia was associated with axonal regeneration and RGC survival (Patel et al., 2017), whereas radial glia and fibroblast-like cells mediated the protective effect of Wnt signaling after spinal cord injury (Briona et al., 2015; Wehner et al., 2017). In contrast, loss of β-catenin within oligodendroyte precursor cells in the damaged optic nerve led to lower gliosis and increased RGC axon regeneration (Rodriguez et al., 2014). Identifying the target cell(s) for Wnt ligands and characterizing mechanisms of action is important for understanding how Wnt regulates axonal regeneration.

Cultures of primary neurons grown in vitro are often used to explore mechanisms of neurite formation and growth and were used here as a reductionist approach to test whether Wnt signaling acts directly on RGCs to induce neurite growth, and to identify intrinsic (RGC-specific) mechanisms. Wnt3a was used to stimulate Wnt signaling because it is endogenously expressed by RGCs and neighboring non-neuronal cells in the retina (Patel et al., 2015a), and induces axonal regrowth from injured RGCs in vivo (Patel et al., 2017). In this study, we asked whether Wnt signaling in RGCs is sufficient to induce neurite growth, determined the effect of Wnt3a stimulation on neurite length and number, and investigated the association of Wnt3a-induced neurite growth with Rip kinases, which are regulators of cell death in RGCs (Dvoriantchikova et al., 2014) and influence axon homeostasis in the spinal cord (Ito et al., 2016).

2. Materials and methods

All procedures involving mice were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Care and Use Committee at the University of Miami. Mice were housed under standard conditions of temperature and humidity, with free access to food and water and a 12-h light/dark cycle. Animals used in our experiments were adult breeding pairs and 10–11 day-old pups. To generate RGC cultures, a two-step immunopanning protocol using positive and negative selection was used to isolate RGCs, as previously described (Dvoriantchikova et al., 2014). Briefly, retinas from wildtype pups were dissociated in papain (16.5 U/mL; Worthington Biochemical Corp, Lakewood, NJ) for 30 min to obtain the cell suspension. Macrophages and endothelial cells were removed from the cell suspension by panning with an anti-macrophage antibody (negative selection) (Accurate Chemical, Westbury, NY). RGCs were then bound to panning plates containing CD90.2/Thy 1.2 hybridoma supernatant (positive selection) and released by incubating with trypsin. RGCs were plated on glass coverslips in 24-well plates at a density of 50,000 cells per well and cultured in serum-free media (Neurobasal/B27 media; Thermo Fisher Scientific, Grand Island, NY). Our previous studies using this preparation method confirmed that the protocol leads to almost pure (> 95%) RGC cultures, as indicated by immunodetection of RGC-specific proteins, qPCR analysis, and retrograde labeling using DiI (Ivanov et al., 2006; Crabb et al., 2010). The RGC cultures were treated with Wnt3a (10–100 ng/ml; R&D Systems, Inc., Minneapolis, MN, USA), PBS (vehicle control) or the Ripk1 inhibitor Necrostatin-1s (Nec-1s, 20 μM; Millipore catalog #504297, Burlington, MA (hereinafter referred to as Nec-1)) diluted in fresh media and incubated for 2 days. The cultures were then processed for immunohistochemistry or gene expression analysis.

For immunocytochemistry, primary RGCs grown on glass coverslips were fixed for 30 min in fresh 4% paraformaldehyde, washed in PBS, then permeabilized in 0.3% Triton X-100 diluted in PBS. The coverslips were blocked in a solution of 2% BSA, 5% goat serum and 0.15% Tween-20 in PBS for 30 min at room temperature, then incubated in primary antibody rabbit anti-βIII-tubulin (Abcam, Cambridge, MA; catalog # ab18207; 1:1000) diluted in blocking solution overnight at 4 °C, followed by three washes in Tween 20/PBS. The coverslips were incubated with secondary antibodies (anti-rabbit Alexa 546, Life Technologies; catalog # A11030; 1:700) for 30 min at room temperature, followed by additional washes in Tween 20/PBS, then were mounted onto glass slides and the cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize the nuclei. The cells were viewed using a Zeiss Axiovert 200 fluorescent microscope and a Leica Microsystems confocal microscope. Controls for the immunostaining included an irrelevant antibody from the same species or omitted the primary antibody. Microscopic and imaging settings were kept constant among positive and negative controls. Neurite length and number per cell were counted from neurite-containing cells in at least 10 random fields per replicate (approximately 100–200 neurites were counted per replicate). Any cells with neurites that could not be traced back to a single soma were excluded from our quantifications. Investigators measuring the neurites were masked to the identity of the treatments in all experiments.

For gene expression analysis, total RNA was isolated from treated primary RGC cultures using the Absolutely RNA Nanoprep Kit (Agilent Technologies, Wilmington, DE), according to the manufacturer’s directions, followed by cDNA synthesis on 1 μg RNA using the SuperScript III First-Strand Synthesis SuperMix (Life Technologies, Carlsbad, CA). The cDNA was diluted 1:4 then 1 μl was used in a volume of 20 μl for the QPCR reactions, which were performed in an Eppendorf Realplex 2 Mastercycler using PowerUp SYBR Green Master Mix (ThermoFisher) and the following specific primers (0.4 μM each) that were designed to cross intron-exon boundaries: mouse Ripk1, forward: 5′ GAAGACAGACCTAGACAGCGG 3′, reverse: 5′ CCAGTAGCTTCACCACTCGAC 3’ (amplifying a region from nucleotides 193–374 in NM_009068.3); mouse Ripk3, forward: 5′ GTGCTACCTACACAGCTTGAAC 3′, reverse: 5′ CCCTCCCTGAAACGTGGAC 3’ (amplifying a region from nucleotides 608–731 in NM_019,955.2). The acidic ribosomal phosphoprotein P0 (ARP) gene was used as a reference gene for normalization, as in Yi et al. (2007). The cycle conditions for the QPCR were recommended by the manufacturer, as follows: 55 C for 2 min; 95 C for 2 min; 95 C for 15 s then 60 C for 60 s, for 40 cycles; 95 C for 15 s; followed by a melting curve analysis from 60 C to 95 C. Primer amplification efficiencies were calculated using 10-fold serial dilutions of each template from 10⁸ to 10⁰ copies; the calculated primer efficiencies are 97.7% for the reference gene ARP, 96.8% for Ripk1 and 97.4% for Ripk3. Relative transcript levels of each gene were calculated using the delta-delta Ct method normalized to a standard cDNA sample and a refernce gene. The PCR products were sequenced by GENEWIZ (South Plainfield, NJ) and demonstrated a 100% match only to the respective mouse Ripk1 and Ripk3 transcripts, demonstrating the specificity of the primers.

Statistical analysis of the data was performed using ANOVA or Student’s t-test with GraphPad Prism. P < 0.05 was considered statistically significant, and data are reported as mean ± standard deviation.

3. Results and discussion

To determine the effect of Wnt3a signaling on neurite formation, primary RGC cultures were incubated with increasing concentrations of Wnt3a ligand using doses previously demonstrated to stimulate canonical Wnt signaling (Yi et al., 2007; Patel et al., 2015b). Wnt3a treatment resulted in significantly greater length and number of βIII-tubulinpositive neurites compared with the saline control (Fig. 1). In the saline-treated RGC cultures, the average neurite length per cell after 48 h was 35.5 ± 12.6 μm, whereas Wnt3a-treated RGCs had an average neurite length of up to 134.8 ± 5.8 μm (p < 0.01) (Fig. 1A). The 100 ng Wnt3a dose frequently led to complex branched neurites over 300 μm (Fig. 1C). Furthermore, Wnt3a increased the number of neurites per cell, up to 5.1 ± 0.45 in the 100 ng Wnt3a dose, compared with 1.6 ± 0.63 for the saline-treated cultures (p < 0.05) (Fig. 1B). Representative images for βIII-tubulin-positive neurites in the Wnt3a-treated and saline-treated RGC cultures are shown in Fig. 1C.

Fig. 1.

Wnt3a induces neurite growth in primary RGC cultures. (A) Wnt3a induced dose-dependent increases in neurite growth. Average neurite length per cell is shown (mean ± SD). Each point is an average of 100–200 neurites.*p = 0.0084; #p < 0.0001. (B) Wnt3a also induced a dose-dependent increase in the average number of neurites per RGC (mean ± SD). Only RGCs with at least 1 neurite were counted. N = 6, @ p=0.043; #p <0.0001 compared to saline control. Each dot is an experimental replicate that represents an average of approximately 100–200 neurites. (C) Representative images showing Wnt3a-induced neurites. The neurites were visualized by immunodetection of βIII-tubulin (green). Extensive neurite growth was observed in Wnt3a-treated RGCs, whereas control-treated RGCs show minimal neurites. A higher magnification image is also shown for 100 ng/ml Wnt3a. Scale bar is 50 μm.

Wnt signaling may induce neurite growth by regulating the activity of genes that influence neurite extension. Ripk1 and Ripk3 were recently shown to contribute to axonal changes in a mouse model of ALS (Ito et al., 2016) and we previously demonstrated that they were associated with retinal damage and RGC death after ischemia-reperfusion injury (Dvoriantchikova et al., 2014). As shown in Fig. 2A, qPCR analysis indicated reduced expression of Ripk1 (37% lower, p = 0.023) and Ripk3 (39% lower, p = 0.0195) in RGCs treated with Wnt3a compared with saline. Due to insufficient amounts of protein that can be obtained from the RGC cultures we were unable to measure expression changes at the protein level; therefore, we investigated the potential involvement of Ripk protein signaling in Wnt3a-induced neurite growth using the Ripk1 inhibitor Necrostatin-1s (Nec-1). RGC cultures were co-treated with Wnt3a and Nec-1 and the neurites were quantified as above. Interestingly, the percent of cells with multiple neurites was significantly higher in cultures with Wnt3a + Nec-1 compared with Wnt3a, higher in cultures treated with Wnt3a alone (Fig. 2B). Additionally, the average number of neurites per cell was 1.53-fold higher in Wnt3a + Nec-1 treated cultures compared with Wnt3a alone (p = 0.0167) (Fig. 2D).

Fig. 2.

The Ripk1 inhibitor Nec-1 enhanced Wnt3a-induced neurite formation. (A) QPCR analysis on primary RGC cultures demonstrated significantly lower expression of Ripk1 and Ripk3 genes in Wnt3a-treated cells (50 ng/ml) compared with control after two days of Wnt3a treatment. (B) RGC cultures co-treated with Wnt3a and Nec-1 (Wnt3a + Nec-1) showed more cells with multiple (> 2 per cell) neurites, whereas more cells treated with Wnt3a alone had only 1 neurite. Percentage of cells at each neurite number ± standard error are shown. Comparisons between Wnt3a + Nec-1 and Wnt3a, and between Nec-1 and Wnt3a are indicated by the symbol above the bars (*p < 0.05, #p < 0.01). N = 4. Each dot is an experimental replicate that represents an average of approximately 100–200 neurites. (C) Representative images of RGC cultures in each treatment group immunostained for βIII-tubulin (green) to visualize the neurites. Scale bar, 50 μm. (D) RGC cultures co-treated with Wnt3a and Nec-1 (Wnt3a + Nec-1) showed 53% higher average number of neurites than Wnt3a treatment alone. (E) The average neurite length per cell was equivalent between Wnt3a + Nec-1 and Wnt3a. Mean ± SD. (F) Cell densities after the different treatments were equivalent, suggesting that Nec-1 did not alter cell survival. Mean ± SD.

The Nec-1 only control showed increased number of neurites per cell and was equivalent to Wnt3a + Nec-1 treatment (Fig. 2D), and both were higher than Wnt3a alone, suggesting a dominant effect of Ripk inhibition on neurite number. In contrast, the average neurite length per cell was not significantly altered by Nec-1 treatment (compare Wnt3a alone with Wnt3a + Nec-1, Fig. 2E). Furthermore, there was not a significant difference in the number of cells among the treatments, suggesting that Nec-1 did not affect cell survival (Fig. 2F).

Therefore, the results of this study demonstrated that Wnt3a-induced neurite growth is an inherent property of RGCs and does not require Wnt signaling from other cell types such as glia. These findings are consistent with previous studies on neurite formation in cultured neurons, such as Wnt7b promoted neurite extension in mouse hippocampal neurons (Rosso et al., 2005), Wnt7a induced neurites in cultured mouse cerebellar granule cells (Lucas and Salinas, 1997) and Wnt3 and Wnt3a induced neurite outgrowth in a β-catenin/TCF4-dependent manner in cultured spinal cord neural cells (David et al., 2010). LiCl, which activates Wnt/β signaling by inhibiting GSK3β kinase activity, induced neurite sprouting in cultured adult spiral ganglion neurons (Shah et al., 2013).

We also observed that inhibiting Ripk1 with Nec-1 increased neurite number but not neurite length. There is extensive literature regarding the specificity of the Nec-1s inhibitor used in this study. For example, Nec-1s was found to be > 1000-fold more selective for Ripk1 than for any other kinase out of 485 human kinases (Christofferson et al., 2012), and Nec-1s does not bind to or inhibit IDO enzyme, further indicating its specificity (Takahashi et al., 2012). However, we did not confirm its specificity in the RGC culture in the current study. A neuritogenesis effect for Nec-1 in our study is consistent with work by Ito that showed that Ripk1 contributed to axonal degeneration in a mouse model of ALS (Ito et al., 2016). The Rip kinases regulate a cell death process known as necroptosis and are induced in conditions where apoptosis is inhibited (Zhou and Yuan, 2014). Our previous data showed that Nec-1 significantly reduced RGC death following ischemic-reperfusion injury, indicating a pro-survival effect in the retina when Ripk1 activity is downregulated (Dvoriantchikova et al., 2014). Therefore, blocking Ripk1 activity with Nec-1 may indirectly lead to neurite formation by suppressing necroptosis-mediated cell death; further, Wnt3a may mediate neurite growth by reducing Ripk1/Ripk3 expression and reducing cell death, which would facilitate RGC survival and allow more time to regrow neurites. However, we did not observe changes in cell numbers, and it is unclear whether changes in neurite number without corresponding changes in neurite length could be caused by altered cell survival. Because we cannot exclude the possibility that our assay may have missed subtle changes in necroptosis by quantifying remaining cells rather than dead cells, a more detailed time course of cell survival with additional cell markers of necroptosis should be examined in future studies to determine whether Ripk1 is related to Wnt-dependent regulation of cell survival.

Further experiments are needed to investigate how Ripk1 inhibition leads to more neurites per cell, for example, by testing whether Ripk1 enhances neuritogenesis or inhibits extension of the main neurite. It will also be important to investigate the relationship between Wnt3a and Rip kinases to determine whether Ripk1 and Ripk3 are downstream from Wnt3a, whether they act independently, and which form of Ripk1 (phosphorylated, unphosphorylated and/or ubiquinated non-degraded) is involved in neurite growth.

In conclusion, Wnt signaling directly induces neurite growth in RGCs, and Ripk1 is identified as a potential modulator of neurite growth. Ripk signaling may be involved in mediating the effects of Wnt3a on neurite number but does not seem to be required for the effect of Wnt3a on neurite length. These findings demonstrate that RGCs are a direct cellular target for Wnt3a and reveal new avenues of investigation into the cellular and molecular mechanisms of Wnt-induced axonal growth.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.exer.2019.03.004.