2.1 Mice

Heterozygous and homozygous Gpr37l1 knock-out mutant, male and female mice and their wild-type littermates were used (Gpr37l1^+/−, Gpr37l1^−/− and Gpr37l1^+/+, RRID:MGI:5512669). All mice were bred from heterozygous crossings, following backcrossing onto a wild-type C57BL/6J background for 10 generations and used as previously described (Marazziti et al., 2013). After weaning, mice were housed by litter of the same sex, 3–5 per cage, and maintained in a temperature-controlled room at 21 ± 2°C, on a 12-hr light–dark cycle (lights on at 07:00 a.m.), with food and water available ad libitum. All animals were born and bred in a specific pathogen-free facility and were subjected to experimental protocols, as reviewed and approved by the Ethical and Scientific Commission of Veterinary Department of the Italian Ministry of Health, according to the ethical and safety rules and guidelines for the use of animals in biomedical research provided by the Italian laws and regulations, in application of the relevant European Union's directives (no. 86/609/EEC and 2010/63/EU).

2.2 Transmission electron microscopy

Adult (3 months) mice organs were fixed by intracardial perfusion with 2% paraformaldehyde (PFA) and 2.5% glutaraldehyde. Brains were postfixed in 2.5% glutaraldehyde overnight at 4°C, cut sagittaly at 300 µm, and processed for electron microscopy as described previously (Doetsch et al., 1997). Briefly, sections were postfixed in 2% osmium tetroxide in 0.1 M phosphate buffer (pH 7.2) for 2 hr. Semi-thin sections, 1.5 µm thick, were cut with a diamond knife and stained with 1% toluidine blue. For the identification of PC, 300 ultrathin (0.05 µm) serial sections were collected in slot grids covered with Formvar resin (Sigma-Aldrich-Merck, Cat# 09823) counterstained with uranyl acetate and lead citrate, and analyzed in a Jeol JEM-1010 electron microscope (JEOL USA Inc.). Pictures were taken with a Gatan MSC 791 CCD digital camera (Gatan) and programmatically adjusted for brightness and contrast using Adobe Photoshop CS software (RRID:SCR_014199).

2.3 Cerebellar astrocyte culture and treatments

Primary astrocytes cultures were prepared from the pooled cerebella of 7–8 Gpr37l1^+/+ or Gpr37l1^−/− male or female mouse pups (WT and KO primary cultures) by a modification of a previously described method (Hatten, 1985). A total of about 150 mouse pups per genotype were used. Each cerebellar pool initially comprised ca. 2–3 × 10⁶ cells of same genotype, to be expanded for use in each experimental analysis. Briefly, cerebella from postnatal day 7 (P7) pups were dissected, cut into small pieces and incubated for 10 min at 37°C, with trypsin (Sigma-Aldrich, Cat# T6763; 10 mg/ml) and deoxyribonuclease I (DNAse I; Sigma-Aldrich, Cat# D4527; 1 mg/ml) in Hanks’ balanced salt solution (HBSS; GE Healthcare Life Sciences, Cat# SH30368.01). Cell suspensions were washed for three times with Dulbecco's Modified Eagle Medium (DMEM; EuroClone, Cat# ECM0728L), containing 10% fetal bovine serum (FBS; Gibco, Cat# 10270106), 2 mM l-glutamine (Gibco, Cat# 25030024), 50 U/ml penicillin-streptomycin (Gibco, Cat# 15070-063 5,000 U/ml), and pre-plated on an uncoated 6-well plate (EuroClone, Cat# ET3006) for 1 hr to remove contaminant fibroblasts. Unattached cells were transferred for 1 hr on poly-l-lysine-coated (0.1 mg/ml; Sigma-Aldrich, Cat# P1399) 24-well plates (0.5 ml/well; Costar, Cat# 3524) or 6-well plates (2 ml/well) to allow astroglial cell attachment. After 1h cells were rinsed several times with 1× Dulbecco's phosphate-buffered saline (DPBS; Gibco, Cat# 14040091) to remove contaminant neurons and cultured at low density to avoid direct cell–cell contact, in DMEM containing 10% FBS serum, at 37°C and 5% CO₂. The medium was changed after 24 hr and adhering cells were cultured for 3–5 days in the same conditions, before splitting with trypsin (0.5 mg/ml) in 1× DPBS and plating on poly-l-lysine-coated culture plates.

Cells at ca. 70% confluence were starved for 24 hr in serum-free DMEM and then treated for 24 hr with the recombinant mouse sonic hedgehog (C25II) N-terminus (Shh-N; 100 nM; R&D System, Cat# 464-SH-200), or the prosaptide ligand TX14(A) (100 nM; Eurogentec, Cat# AS-60248-1), or the Smo inhibitor SANT-1 (100 nM; Calbiochem, Cat# 559303), or with a combination of either 100 nM Shh-N and 100 nM TX14(X) or 100 nM SANT-1 and 100 nM TX14(X). For the latter treatment cells were pre-incubated for 1 hr with SANT-1 prior to the addition of TX14(X). Equal volumes of vehicle alone were added to the untreated cell samples. All treatments were administered with cells maintained in serum-free DMEM at 37°C and 5% CO₂. The above reagents were dissolved as follows: Shh-N in 1× DPBS, pH 7.4; TX14(A) in a binding assay buffer with protease inhibitors (Meyer et al., 2013); SANT-1 in dimethyl sulfoxide (DMSO; Sigma-Aldrich, Cat# D2650).

2.4 Antigens and antibodies used for immunofluorescence labeling

The following primary antibodies were used as specific tissue/cell type markers and have extensive published data supporting their validity (Table 1): aquaporin 4 (Aqp4; 1:50, Santa Cruz, RRID:AB_2059853); ADP-ribosylation factor-like protein 13B (Arl13B; 1:500, UC Davis/NIH NeuroMab Facility, RRID:AB_11000053; 1:1,000, Proteintech, RRID:AB_2060867); caveolin 1 (1:50, Cell Signaling Technology, RRID:AB_2072166); contactin 2 (Cntn2/Tag1; 1:100, R&D Systems, RRID:AB_2245173); glial fibrillary acidic protein (Gfap; 1:500, Millipore, RRID:AB_2109645; 1:500, BD Biosciences Pharmingen, RRID:AB_396365); glial high affinity glutamate transporter (or glutamate aspartate transporter, Glast; 1:100, Novus Biologicals, RRID:AB_839154); Gpr37l1 (1:50, Mab Technologies, RRID:AB_2857918); ionized calcium binding adaptor molecule 1 (Iba1; 1:500, FUJIFILM Wako Chemicals, RRID:AB_839504); oligodendrocyte transcription factor 2 (Olig2; 1:1,000, Abcam, RRID:AB_11205039); Shh-E1 (1:400, Santa Cruz Biotechnology, RRID:AB_10709580); smoothened (Smo; 1:50, Santa Cruz, RRID:AB_2239686); patched 1 (Ptch1; 1:50, Santa Cruz, RRID:AB_2174039); RAB5A, member RAS oncogene family (Rab5a or Rab5; 1:200, Cell Signaling Technology, RRID:AB_2300649); RAB7, member RAS oncogene family (Rab7; 1:200, Cell Signaling Technology, RRID:AB_1904103); Alexa Fluorophore-conjugated secondary antibodies made in donkey (1:500, Invitrogen, specific for: rat, Cat# A11006 RRID:AB_141373; mouse, Cat# A21202 RRID:AB_141607 and Cat# A31570 RRID:AB_2536180; goat Cat# A11055 RRID:AB_2534102 and Cat# A21432 RRID:AB_141788; rabbit, Cat# A21206 RRID:AB_2535792 and Cat# A31572 RRID:AB_162543). The immunostaining of each marker showed the expected pattern of cellular morphology.

2.5 Immunofluorescence labeling, microscopy, and quantitative image analysis

Primary astrocytes from WT and KO cultures were grown on poly-l-lysine-coated glass coverslips in 24-well plates. After treatments, the cells were kept at room temperature and fixed with 4% paraformaldehyde (PFA; Sigma-Aldrich, Cat# 158127) in DPBS for 15 min, permeabilized with 0.1% Triton X-100 (Merck, Cat# 108603) for 10 min and incubated for 1 hr in blocking buffer containing 20% normal donkey serum (Millipore, Cat# S30), 1% bovine serum albumin (BSA; Sigma-Aldrich, Cat# A4503) and 0.05% Tween-20 (Bio-Rad Laboratories, Cat# 1706531). Samples were then incubated overnight at 4°C, with primary antibodies diluted in blocking buffer, followed by washing and incubation with fluorophores-conjugated secondary antibodies. Nuclei were stained with 4,6-diamino-2-phenylindole (DAPI; Molecular Probes, Cat# D1306).

Gpr37l1 immunostaining was performed as above, after fixing the cells with 100% methanol at −20°C for 20 min, permeabilizing with 0.1% Triton X-100 and incubating for 1 hr at room temperature in blocking buffer containing 0.5% BSA, 0.3 M Glycine (Merck, Cat# 104201) and 0.1% Tween-20.

Fluorescence micrographs were acquired with a TCS SP5 laser scanning confocal microscope or with a motorized LMD7000 microscope (Leica Microsystems) using the manufacturer's imaging software. The percentage of Gfap-positive ciliated cells was assessed by fluorescence microscopy-based visualization of Arl13b-stained PC and DAPI-labeled nuclei. The percentage of cells with Smo localized in PC was assessed by counting the total number of Gfap-positive ciliated cells with also Smo-positive cilia. Images were analyzed by ImageJ software (NIH, RRID:SCR_003070). For each sample at least 20 fields were counted. Images were scored blinded before quantification.

The colocalization analysis was performed using the specific threshold plug-in of FiJi-ImageJ software v2.1.0/1.53c (NIH, RRID:SCR_003070). Colocalization of Ptch1 and Rab7 was evaluated upon measurement of Pearson's correlation coefficient, based on the pixel intensity correlation over space. For each multichannel acquired image, the analysis was performed on manually chosen regions of interest (ROI), each selected within one cell. At least 10 ROIs per experimental group with at least three samples per genotype were analyzed. Images were scored blinded before quantification.

2.6 RNA extraction and real-time PCR assay

Total RNA was extracted from WT and KO primary cultures (ca. 1 × 10⁶ cells) or frozen whole brain samples (1 mg) from wild-type, Gpr37l1^−/− and Gpr37^−/− (Marazziti et al., 2004) adult mice, using a RNeasy Plus Mini kit (QIAGEN, Cat# 74134) according to the manufacturer's instructions. Total RNA preparations were reverse-transcribed according to standard procedures. Real-time PCR (RT-PCR) was performed using the following primers: mouse Gpr37l1, forward 5′-GGCAATCTGTCTGTCATGTG-3′ and reverse 5′-CACATGGAATCGGTCTATGC-3′; mouse Gpr37, forward 5′-AGAGCTGGAGCTGTCGCCC-3′ and reverse 5′-GGCCTGCCTTCAATATACA-3′ (Marazziti et al., 2004). Experiments were repeated at least three times, and samples were analyzed in triplicate.

2.7 Cell proliferation assay

Cell proliferation was determined by 5-bromo-2′-deoxyuridine (BrdU, Sigma-Aldrich, Cat# B5002) incorporation experiments. WT and KO cultures were plated on poly-l-lysine-coated glass coverslips in 24-well plates and grown as previously described. Cells at ca. 70% confluence were starved for 24 hr in serum-free DMEM and then treated with either Shh-N or TX14(A) at different concentrations (10, 100, or 200 nM and 10, 100, or 1,000 nM, respectively) or vehicle for 24 hr, in serum-free DMEM, at 37°C and 5% CO₂. Other cell samples were starved for 24 hr with serum-free DMEM and then treated as described above, with either 100 nM Shh-N, 100 nM SANT-1, 100 nM TX14(A) or their combinations, or vehicle for 24 hr, in serum-free DMEM. All treated cells were pulsed with 10 μM BrdU for the last 16 hr during each treatment period and then fixed and co-stained with anti-BrdU and anti-GFAP according to standard protocols (Marazziti et al., 2013).

At least 500 BrdU-positive cells per experimental condition were quantified using the ImageJ software in three or more independent experiments. Experiments were repeated at least three times with different cell culture preparations, and samples were analyzed in triplicate in a blinded manner.

2.8 Filipin staining and analysis of free intracellular cholesterol levels

Filipin staining was applied for analyzing free intracellular cholesterol levels (Muller et al., 1984). Primary astrocytes were grown on glass coverslips, fixed with 4% paraformaldehyde solution in DPBS, incubated for 20 min at room temperature with ammonium chloride (20 mM) in DPBS, and blocked and permeabilized in 1× DPBS, 0.2% Triton X-100 solution. After Gfap immunostaining (1:500, BD Biosciences Pharmingen; RRID:AB_396365), cells were incubated overnight at 4°C, with filipin (50 µg/ml; Sigma-Aldrich, Cat# F4767) in DPBS. Excess filipin was removed by washing with cold DPBS and fluorescence images were immediately captured using a TCS SP5 laser scanning confocal microscope (Leica Microsystems), with ultraviolet excitation around 360 nm and emission detection around 450 nm, using manufacturer's imaging software. All procedures were carried out with complete protection from ambient light. Filipin staining intensity was quantified with Imaris Standard Package for OLME Systems, v7.1 (Bitplane, RRID:SCR_007370). For each acquired image, the analysis was performed on manually selected ROI, each corresponding to one cell. At least 10 ROIs per experimental group with at least three samples per genotype were analyzed. Images were scored blinded before quantification.

2.9 Protein extraction and Western blot analysis

Protein extracts were prepared from WT and KO primary astrocytes, after culture in control conditions or after serum starvation followed by treatment for 24 hr with Shh-N (100 nM) or TX14(A) (100 nM), or from striata of wild-type or Gpr37^−/− adult mice (Marazziti et al., 2004). Cell and tissue samples were dissolved in lysis buffer (20 mM HEPES pH 7.3, 120 mM NaCl, 5 mM EDTA, 10% glycerol, 1% Triton X-100, Roche complete protease inhibitor cocktail) and cleared by centrifugation. The protein content of supernatant was quantified by detergent-compatible assay (Bio-Rad) and protein samples (50 μg) were separated by SDS-PAGE and analyzed by Western blot, according to standard protocols. Protein antigens were immunodetected with the following primary antibodies, which labeled the respective protein bands of expected molecular weight as reported in previous publications: Ptch1 (1:1,000, R&D Systems, RRID:AB_2174045), Gpr37l1 (1:500, Santa Cruz, RRID:AB_10844948); Gpr37 (clone #202, 1:1,000, kindly provided by R. Takahashi and Y. Imai, RIKEN Brain Science Institute, Saitama, Japan); Shh-N19 (1:400, Santa Cruz, RRID:AB_632416); α-tubulin (1:1,000, Sigma-Aldrich, RRID:AB_11204167). Horseradish peroxidase-conjugated secondary antibodies, specific for rat (Amersham Biosciences-GE Healthcare, Cat# NA935 RRID:AB_772207) mouse (Amersham Biosciences-GE Healthcare, Cat# NA931 RRID:AB_772210), goat (Santa Cruz, Cat# sc-2020, RRID:AB_631728) immunoglobulins were used, following the producer's instructions. The blotted membranes were then processed for chemiluminescence detection with an ECL kit (Amersham, Cat# GEHRPN2232) and exposed (Chemidoc XRS+ imager, Bio-Rad, RRID:SCR_014210). The luminescent signal of immunoreactive bands was imaged and quantified with the Image Lab software (Bio-Rad, RRID:SCR_014210). The intensity of each band was normalized to the intensity of the corresponding α-tubulin band. The average values of each experimental group were expressed in arbitrary units, as a ratio to the mean values obtained from the wild-type control group.

2.10 Analysis of mitogen-activated protein kinase phosphorylation levels

Cultured primary astrocytes were starved for 3 hr in serum-free DMEM at 37°C and 5% CO₂ before analysis. About 1 × 10⁶ cells for each sample were rapidly harvested in lysis buffer (62.5 mM Tris·HCl pH 6.8, 25% glycerol, 2% SDS and 0.01% bromophenol blue), sonicated and processed for Western blot analysis, as described above, with polyclonal antibodies specific for mitogen-activated protein kinase 3/1 (Mapk3/1 or Erk1/2 or p44/p42 Mapk; 1:1,000, Cell Signaling Technologies, RRID:AB_330744, Table 1) or Thr202/Tyr204 phosphorylated Mapk3/1 (phospho-Mapk3/1 or phospho-p44/p42 Mapk; 1:1,000, Cell Signaling Technologies, RRID:AB_331646, Table 1). The data shown in the figure are representative of three to five independent experiments. Experimental results were scored blinded to genotype and analyzed by unpaired t test. P values less than 0.05 were considered significant.

2.11 Conditioned medium assay and analysis

Cultures of WT and KO primary astrocytes were seeded into 24-well plates and allowed to reach ca. 80% confluence. Following starvation in serum-free DMEM for 24 hr, conditioned medium (CM) samples were collected. CM from either WT or KO cultures was added to other sets of WT cells, after they had been subjected to serum starvation for 24 hr. A Shh-specific monoclonal antibody (E1; Santa Cruz; RRID:AB_10709580) or anti-mouse immunoglobulins G (Sigma-Aldrich, Cat. #M-9902, RRID:AB_1839970) were added (final concentration 2.5 μg/ml), as indicated. After 24 hr of incubation in serum-free medium at 37°C and 5% CO₂, BrdU incorporation or Western blot assays were performed, as reported above.

Proteins in CM samples (5 ml) from WT or KO cultures were concentrated by Vivaspin Turbo 4 protein concentrator spin columns (Sartorius AG, Cat. #VS04T01), separated by SDS-PAGE and analyzed by Western blot, according to standard protocols.

2.12 TUNEL assay

The presence of apoptotic cells in WT or KO cultures was tested by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay with In Situ Cell Death Detection Kit (Merck, Cat# 12156792910). Fluorescence micrographs were acquired with a LMD7000 motorized microscope.

2.13 Figure composition

Figures were composed in Adobe Illustrator CS (Adobe, RRID:SCR_010279). For representative confocal images, maximum projections of more than 10 z-stacks were generated with Leica Application Suite X (RRID:SCR_013673). Images were adjusted for brightness and contrast using Adobe Photoshop CS software.

2.14 Statistical analysis

GraphPad Prism version 5.0 (Graph Pad, RRID:SCR_002789) was used for statistical analysis of the data. All analyzed samples were obtained from at least seven individual pups for each experimental condition, in agreement with specific sample size analysis for standardized comparison of murine mutant phenotypes, as reported by the reference European Mouse Disease Clinic Consortium (EUMODIC) and International Mouse Phenotyping Consortium (IMPC) (Hrabe de Angelis, 2015; Meehan et al., 2017).

All data are presented as mean ± standard deviation (SD). All experimental data were included in the results. Experimental results were scored blinded and one-way analysis of variance (ANOVA) with Bonferroni post hoc test, was used to determine statistical significance for experiments with three or more experimental groups. The effects of the two studied genotypes on Mapk phosphorylation levels were analyzed by unpaired t test. P values of less than 0.05 were considered significant. Statistical details are reported in each figure legend.

3.1 Primary cilia in mouse cerebellar Bergmann astrocytes

PC-like projections were previously described in mammalian cerebral cortical astrocytes, as well as cultured hippocampal neurons and astrocytes (Berbari et al., 2007; Bishop et al., 2007). The present study originally provides the specific ultrastructural identification and characterization of PC in mouse BG astrocytes upon electron microscopy and morphological analysis of cerebella from wild-type, C57BL/6J congenic adult animals (Figure 1). This confirms previous immunolabeling results (Di Pietro et al., 2017; Marazziti et al., 2013) and highlights the importance of specifically studying the presence and function of PC in BG-derived cultured cerebellar astrocytes.

3.2 Production and characterization of Gpr37l1^+/+ and Gpr37l1^−/− mouse cerebellar primary astrocytes

Ciliogenesis-competent murine cerebellar primary astrocytes were cultured and used to study Gpr37l1’s modulation of Shh-induced mitogenic. Homogeneous monolayers of mouse primary astrocytes were derived as previously described (Yoshimura et al., 2011), from cerebella of P7 Gpr37l1^+/+ and Gpr37l1^−/− littermates (WT and KO astrocytes). All pups originated from crossing of heterozygous animals of the original C57BL/6J congenic strain, which shows increasing cerebellar expression of the Gpr37l1 protein from P5 to P15 (Marazziti et al., 2013). Astrocyte cultures at sub-confluent conditions were subjected to serum starvation, to induce cell-cycling arrest and entering in G0 phase and to stimulate the formation of PC (Kiprilov et al., 2008; Ott & Lippincott-Schwartz, 2012; Zhang et al., 2009). Serum removal was applied to also deplete exogenous mitogens and other soluble effectors in the culture medium (Liu et al., 2018).

Serum-starved cultures were then characterized by immunofluorescence labeling with antibodies against various astrocyte markers, including Gpr37l1, and the PC marker, Arl13b (Figures 2 and S1). About 90% of all cells were positively stained for astrocyte-specific marker, Gfap. All Gfap-positive cells also showed specific Gpr37l1 immunostaining (Figure 2a). There was no difference between WT and KO cultures in the expression of Gfap or the other glial markers, high-affinity glutamate transporter (Glast) and water transporter aquaporin 4 (Aqp4) (Figures 2b and S1a,b). Cells positive for the oligodendrocyte progenitor transcription factor 2 (Olig2) or the postmitotic granule neuron contactin 2 (Cntn2 or Tag1) markers were very rare, while about 10% of cells were positive for the microglial marker, allograft inflammatory factor 1 (or ionized calcium binding adaptor molecule 1, Iba1) (Figure S1c). Two predominant forms of astrocytes were present: stellate and bi-tripolar cells (Mason, 1988), in a proportion of about 85% to 15% in WT and 70% to 30% in KO cultures (Figure S1d). Astrogliosis (Eddleston & Mucke, 1993; Sofroniew, 2015) or astrocyte de-differentiation (Sirko et al., 2013) was not detected in either genotype's samples. TUNEL-positive apoptotic cells were absent in all studied experimental samples.

Arl13b-positive PC were similarly present in about 50% of Gfap-positive cells from both WT and KO cultures (Figure 2b).

RT-PCR assays (Figure 2c) on total RNA extracts from WT or KO whole brain samples and cerebellar astrocyte cultures showed the specific transcription of the Gpr37l1 gene in all WT samples. Its closest homolog, Gpr37, was markedly expressed in WT whole brain samples, but scarcely transcribed in both Gpr37l1^+/+ and Gpr37l1^−/− cultured astrocytes (Jolly et al., 2018; Yang et al., 2016). Consistently, the Gpr37 protein was absent in cerebellar primary astrocytes of both genotypes (Figure 2d).

3.3 Genetic ablation of Gpr37l1 enhances basal and Shh-induced proliferation of cerebellar astrocyte

Specific tests were carried out to determine whether and to what extent Shh could induce cerebellar astrocytes’ mitogenesis and to investigate its modulation upon ablation of Gpr37l1’s function (Di Pietro et al., 2017; Marazziti et al., 2013; Ugbode et al., 2017), or treatment with the Gpr37l1-interacting, prosaposin-derived TX14(A) ligand (Liu et al., 2018; Meyer et al., 2013).

Following serum-starvation WT and KO astrocytes were cultured in the presence of varying concentrations of the biologically active N-terminal fragment of Shh (Shh-N) (Roelink et al., 1995), the Smo synthetic antagonist, SANT-1 (Chen et al., 2012) or TX14(A) and proliferation was measured by BrdU incorporation, as reported in M&M (Figures S2a and 3a). In untreated conditions, about 2%–3% of serum-starved WT cells were in proliferative state, similar to what reported for other mammalian glial cell types (Barca et al., 2007; Cragnolini et al., 2009), while Gpr37l1-deficient cultures showed on average a ca. fourfold higher basal mitotic rates.

Shh-N treatment (100 nM) induced significant increments in the percentage of proliferating WT astrocytes, while null mutant cultures did not display a significantly higher level of proliferation at all tested concentration (Figure S2a). In particular, Shh-N (100 nM) induced a ca. threefold increase in the percentage of proliferating WT astrocytes, compared to unstimulated cells (mean ± SE; WT control: 3.2 ± 0.4; KO control: 10.7 ± 0.8; WT+Shh-N: 9.4 ± 1.1; KO+Shh-N: 12.8 ± 1.1) (Figure 3a). SANT-1 (100 nM) drastically reduced the proportion of proliferating KO cells, to values comparable to those observed in WT untreated samples (mean ± SE; KO+SANT-1:3.8 ± 0.6) (Figure 3a).

Administration of TX14(A) at various concentrations (10–100 nM) specifically stimulated WT astrocytes, inducing mitotic levels similar to those shown by untreated KO cells, which were instead not significantly affected by TX14(A) treatment (mean ± SE; 100 nM TX14(A); WT TX14(A): 9.9 ± 1.3; KO TX14(A): 5.7 ± 1.1) (Figure S2a and 3a). Furthermore, the proliferative levels of WT cultures were significantly incremented upon co-treatment with Shh-N (100 nM), while concomitant delivering of SANT-1 (100 nM) did not significantly alter the prosaptide's effects (mean ± SE; TX14(A) + Shh-N; WT: 10.9 ± 1.8; TX14(A) + SANT-1:7.6 ± 1.1) (Figure 3a). One-way ANOVA with Bonferroni post hoc test was performed and data are shown as mean ± SD (Figure S2a: F_{(236, 13)} = 13.99, n = 14, p < 0.0001; Figure 3a: F_{(140, 11)} = 10.64, n = 12, p < 0.0001).

The levels of MAPK phosphorylation associated with cellular proliferation (Wei & Liu, 2002) were also investigated, in control conditions (Meyer et al., 2013; Okuda et al., 2016) (Figure S2b). Unstimulated KO cultures exhibited increased levels of MAPK phosphorylation, in comparison with WT cultures (**p < 0.009, WT vs. KO, unpaired t test), in agreement with the higher extent of proliferation showed by KO cells in the same experimental conditions (Figures S2 and 3a).

Thus, the constitutive absence of Gpr37l1 membrane receptor's expression upon gene ablation does increase the basal proliferation level of primary murine cerebellar astrocytes. This outcome is associated with the stimulation of the Shh–Ptch1–Smo mitogenic pathway, as it is specifically antagonized by SANT-1 treatment and accompanied by a consistent increase in MAPK phosphorylation. Interestingly, the Gpr37l1–prosaptide interaction, with consequent receptor's internalization from the plasma membrane (Meyer et al., 2013), induces a comparable, significant augment of WT astrocyte's proliferation, in the absence of any experimental treatment with exogenous Shh-N.

3.4 Smo, Ptch1, and Gpr37l1 localize to the primary cilium in a dynamic manner

Hedgehog signaling involves binding to the transmembrane co-receptor Ptch1, relieving its inhibition of Smo, which is consequently translocated into the PC membrane. Smo, Ptch1, and Gpr37l1 membrane localization was hence analyzed by double immunofluorescence labeling. Smo was found colocalized with the PC membrane marker, Arl13b, in a small proportion of untreated WT astrocytes, but the colocalization was strikingly evident in most unstimulated KO cells (Figure 3b,c). Indeed, the percentage of ciliated cells was comparable among samples of both genotypes, in all tested conditions, while in the absence of exogenous stimuli 61.3% of KO cells had Smo localized at PC, compared to 12.5% of WT cells (Figure 3c). Shh-N (100 nM) treatment markedly increased the ciliary localization of Smo, in both WT and KO cells (64.9% and 88.2% of cells, respectively), while TX14(A) had a specific effect on WT astrocytes, only (56.6% of cells with Smo on cilium; Figure 3b,c). The results of one-way ANOVA with Bonferroni post hoc test are shown as mean ± SD (Figure 3c, left panel: F_(42,5) = 1.13, n = 6; right panel: F_{(66, 5)} = 25.75, n = 6, p < 0.0001).

Ptch1 and Gpr37l1 were found localized in correspondence of PC membranes only in unstimulated WT cultures. Addition of Shh-N or TX14(A) resulted in the absence of cilium-localized Gpr37l1 in WT samples and of cilium-localized Ptch1 (or its translocation to the ciliary base) in both WT and KO astrocytes (Figure 4).

Gpr37l1 and Ptch1 have been reported to functionally interact (Marazziti et al., 2013) and therefore their possible co-expression and localization were studied in the absence or presence of Shh-N or TX14(A), by immunofluorescence labeling and Western blotting with specific antibodies.

Upon treatment with either Shh-N or TX14(A), WT astrocytes presented an increased intracellular colocalization of both Gpr37l1 and Ptch1 proteins. On the other hand, in KO cells Ptch1 always had a predominant intracellular location and aggregation, even in untreated conditions (Figure 5a).

Ptch1 protein's expression was significantly increased in KO cultures, in comparison with WT samples, while the above treatments did not significantly affect its expression levels, in cells of both genotypes (Figure 5a,c). The indicated differences were statistically significant as determined by one-way ANOVA followed by Bonferroni post hoc test. Data are shown as mean ± SD (Figure 5c: F_{(29, 5)} = 9.97, n = 6, p < 0.0001).

These findings are in agreement with the reported analysis of cerebellar samples from Gpr37l1 KO pups (Marazziti et al., 2013) and various studies of Ptch1 localization and expression, during Shh-induced mitogenesis (Incardona et al., 2002; Yue et al., 2014). Also, Gpr37l1’s expression levels in WT cultures were not altered by Shh-N or TX14(A) (Figure 5d). No statistically significant difference was determined by one-way ANOVA followed by Bonferroni post hoc test (Figure 5d: F_{(13, 2)} = 0.13, n = 3).

Cholesterol importantly affects Shh signaling cascades (Blassberg & Jacob, 2017; Mann & Beachy, 2004) and Ptch1 is involved in modulating its intracellular concentration (Bidet et al., 2011; Hu & Song, 2019). Filipin staining was performed for the analysis of free intracellular cholesterol distribution (Tabas et al., 1994) (Figure 5a). Unstimulated KO cultures showed a significantly higher extent of cholesterol, compared to untreated WT cells, consistent with the concomitant, marked increment in cilium-associated Smo (Figures 3b,c and 5a,b).

Shh-N treatment significantly increased intracellular cholesterol levels in WT cultures and comparable effects were observed upon administration of TX14(A) (Figure 5a,b; one-way ANOVA followed by Bonferroni post hoc test; data are shown as mean ± SD; F_{(90, 5)} = 4.01, n = 6, p < 0.001). These results are in accordance with studies of mitogenic Shh stimulation enhancing cellular concentration of cholesterol (Bidet et al., 2011; Hu & Song, 2019), thus critically controlling Smo enrichment at ciliary membranes and consequently intracellular signal transduction (Huang et al., 2016; Luchetti et al., 2016; Xiao et al., 2017).

3.5 Gpr37l1 ablation affects Shh-induced Ptch1 trafficking

Translocation of Ptch1 to intracellular endosomal/lysosomal organelles crucially regulates Smo mobilization and the activation of Shh-induced proliferative pathways (Incardona et al., 2002; Karpen et al., 2001). Membrane trafficking of Ptch1 was therefore studied in WT or KO cultures, in the absence or presence of Shh-N, upon co-labeling of caveolin 1, Rab5 or Rab7, as specific markers of plasma membrane lipid rafts, early endosomes or late endosomes, respectively (Figure 6). In untreated conditions, Ptch1 appeared to be similarly distributed with caveolin 1, Rab5, and Rab7, in both WT and KO cells. In the presence of Shh-N, Ptch1 showed a predominant intracellular distribution, in Rab7-positive, late endosomal/prelysosomal compartments (Figure 6a). The analysis of Pearson's correlation coefficient from Ptch1 and Rab7 double-immunofluorescence imaging revealed a significantly higher level of colocalization of the two proteins in KO astrocytes upon Shh-N treatment (Figure 6b; one-way ANOVA followed by Bonferroni post hoc test; data are shown as mean ± SD; F_{(56, 3)} = 9.51, n = 4, p < 0.0001).

3.6 Gpr37l1^−/− primary astrocytes produce Shh

Given the increased proliferative rate of unstimulated KO primary astrocytes, immunofluorescence staining and Western blot analysis were carried out to assess their production and secretion of active Shh. The Shh protein was markedly immunolabeled in serum-starved Gpr37l1 null mutant samples, as well as in KO cell extracts, from both starved and not starved cultures, while control WT samples showed negligible Shh expression (Figure 7a–c).

Secreted Shh molecules were specifically detected in medium samples of KO cultures, which markedly stimulated the proliferation of distinct, testing samples of WT cells. These mitogenic effects were completely abolished in the presence of Shh-specific antibodies (Figure 7c,d; one-way ANOVA followed by Bonferroni post hoc test; data are shown as mean ± SD; F_{(40, 5)} = 12.73; n = 6, p < 0.0001).

Therefore, the genetic ablation of Gpr37l1 results in cerebellar primary astrocytes producing active Shh, which can in turn promote cellular proliferation, upon stimulation of Ptch1 intracellular translocation from the plasma membrane, intracellular cholesterol accumulation, and Smo mobilization to ciliary membrane.