Glio- and neuro-protection by prosaposin is mediated by orphan G-protein coupled receptors GPR37L1 and GPR37

Abstract

Discovery of neuroprotective pathways is one of the major priorities for neuroscience. Astrocytes are natural neuroprotectors and it is likely that brain resilience can be enhanced by mobilizing their protective potential. Among G-protein coupled receptors expressed by astrocytes, two highly related receptors, GPR37L1 and GPR37, are of particular interest. Previous studies suggested that these receptors are activated by a peptide Saposin C and its neuroactive fragments (prosaptide TX14(A)), which were demonstrated to be neuroprotective in various animal models by several groups. However, pairing of Saposin C or prosaptides with GPR37L1/GPR37 has been challenged and presently GPR37L1/GPR37 have regained their orphan status. Here, we demonstrate that in their natural habitat, astrocytes, these receptors mediate a range of effects of TX14(A), including protection from oxidative stress. The Saposin C/GPR37L1/GPR37 pathway is also involved in the neuroprotective effect of astrocytes on neurons subjected to oxidative stress. The action of TX14(A) is at least partially mediated by Gi-proteins and the cAMP-PKA axis. On the other hand, when recombinant GPR37L1 or GPR37 are expressed in HEK293 cells, they are not functional and do not respond to TX14(A), which explains unsuccessful attempts to confirm the ligand-receptor pairing. Therefore, this study identifies GPR37L1/GPR37 as the receptors for TX14(A), and, by extension of Saposin C, and paves the way for the development of neuroprotective therapeutics acting via these receptors.

Keywords: GPR37; GPR37L1; PKA; Saposin C; astrocyte; astroprotection; cAMP; neuroprotection; orphan receptors; prosaptide.

Figure 1

TX14(A) acting on GPR37L1/GPR37 reduces cAMP levels in astrocytes. (a) Layout of the adenoviral vectors for knock‐down of GPR37L1 and GPR37. Each vector allows co‐cistronic expression of three pre‐miRNAs targeting different regions of the target gene. AVV = human adenoviral vectors serotype 5; CMV = human cytomegalovirus promoter; EmGFP = Emerald green fluorescent protein; miR155 = flanking pre‐miRNA sequence derived from miR‐155. (b) Western blot confirms that AVV–CMV–EmGFP–miR155/GPR37L1 and AVV–CMV–EmGFP–miR155/GPR37 (MOI 10) efficiently knock‐down GPR37L1 and GPR37 in astrocytes. AVV–CMV–EmGFP–miR155/negative is a control vector with hairpin sequence relevant to no known vertebrate gene. (c) Concentration–response curves for inhibition of cAMP production by TX14(A) in astrocytes pretreated with 1 μM NKH477. Cells were transduced with AVV–CMV–Glosensor and either AVV–CMV–EmGFP (control), a mixture of AVV–CMV–EmGFP–miR155/GPR37L1 and AVV–CMV–EmGFP–miR155/GPR37 to knock‐down GPR37L1/GPR37, or with AVV–CMV–EmGFP–miR155/negative (n = 4, triplicates). (d) AVV–CMV–Glosensor transduced astrocytes were pretreated with PTX (20 hr, 100 ng/mL). About 100 nM TX14(A)‐induced cAMP reduction in astrocytes was PTX sensitive (n = 12, ***p < .001 vs. indicated group, one‐way ANOVA with Turkey's post hoc analysis). (e): Astrocytes expressing an EPAC‐based cAMP sensor were kept in PSAP‐depleted media overnight and were stimulated with NKH477 (0.5 μM) in the absence or presence of TX14(A) (100 nM). TX14(A) decreased the transient cAMP signal; average of 58 astrocytes from four experiments. (f) Pooled data from (e) shows significantly decreased FRET ratio peaks with TX14(A) (n = 58 = ****p < .0001, paired t test) [Color figure can be viewed at wileyonlinelibrary.com]

Figure 2

GPR37L1 and GPR37 are non‐responsive to prosaptide TX14(A) in PRESTO‐Tango assay in HEK293 cells. PRESTO‐Tango uses clones of numerous human GPCR, C‐terminally tagged with a special signaling element. These receptors need to be expressed in a specially designed clone of HEK293 cells. Agonist binding triggers receptor internalization and eventually leads to expression of luciferase and luminescence. Two concentrations of plasmid DNA were used to express the tagged receptors. GPR37 exhibits strong constitutive activity, especially when using 0.1 μg/μL. Values obtained with 0.1 μg of GPR37 DNA could not be fitted with the regression algorithm of Prism software, hence no line is shown (n = 6) [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3

PSAP/GPR37L1/GPR37–mediated signaling is essential for migration of astrocytes in the scratch assay and the effect of PSAP is mimicked by TX14(A). (a) Astrocytes move into a wound in media containing 5% FBS and in PDM (prosaposin‐depleted media) supplemented with 100 nM TX14(A) but the movement is inhibited in absence of PSAP. Representative images at 0, 24, and 48 hr post scratch. See also Supporting Information Movies SS1–SS3. (b) Dynamics of the relative wound density under conditions shown in a (n = 3, triplicates). (c) Relative wound density 48 hr post scratch for astrocytes incubated in FBS‐containing media, PDM, PDM + TX14(A), PDM + TX14(A) + both knock‐down vectors and PDM + TX14(A) + knock‐down control vector n = 6, triplicates, ***p < .001 vs. indicated group, one‐way ANOVA with Turkey's post hoc analysis). (d): DAPI staining revealed no significant differences in the cell density between astrocytes cultured in media (+FBS), PDM and PDM supplemented with TX14(A) (100 nM; n = 15). (e): The addition of TX14(A) to PDM does not affect the numbers of new astrocytes based on BrdU staining (n = 7, triplicates) [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4

cAMP in astrocytes affects wound closure in the scratch assay. (a) A scratch wound was created in astrocyte monolayers cultured in FBS‐containing media or PDM supplemented with 100 nM TX14(A). Drugs were added and dynamics of the wound closure was monitored for 72 hr (n = 5, triplicates). (b) DAPI staining of astrocytes shows that NKH477 (10 μM), 6‐BenZ‐cAMP (500 μM), and 8‐pCPT‐2′‐O‐Me‐cAMP (1,000 μM) significantly decreased cell numbers as compared to control (n = 4, triplicates). In both cases, one‐way ANOVA with Dunnett's post hoc analysis. **p < .01, ***p < .001 vs. control [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5

TX14(A) acts via GPR37L1 and GPR37 to protect primary astrocytes against toxicity induced by H₂O₂, staurosporine or rotenone. (a) Pre‐exposure to stressors (5 hr) drastically reduce relative wound density recorded at 48 hr in PDM. TX14(A) (100 nM) rescued astrocytes bringing wound density close to normal (compare to Figure 3). GPR37L1/GPR37 knock‐down prevented the protective effect of TX14(A), while the control vector had no effect (n = 6, triplicates, ***p < .001 vs. indicated group). (b) LDH release was used as a measure of cytotoxicity, 24 hr after exposure of astrocytes to oxidative stress. In PDM, manipulation of GPR37L1/GPR37 had no effect. TX14(A) (100 nM) protected them from damage but only when they were expressing GPR37L1/GPR37, (n = 6, triplicates, **p < .01, ***p < .001 vs. control group, for example, PDM groups). One‐way ANOVA with Bonferroni's post hoc analysis [Color figure can be viewed at wileyonlinelibrary.com]

Figure 6

Co‐cultured astrocytes protect cortical neurons against oxidative toxicity partially through GPR37L1/GPR37 signaling in astrocytes. (a) Experimental design: (A—stressed neurons (blue) in absence or presence of astrocytes (green) on a culture insert; B—depletion of PSAP from the media with anti‐PSAP antibodies; C—GPR37L1 and GPR37 knock‐down in astrocytes co‐cultured with neurons; D—GPR37L1 and GPR37 knock‐down in neurons in the co‐culture system. Neurons were treated with H₂O₂ (250 μM) for 1 hr, or staurosporine (100 nM) or rotenone (50 μM) for 2 hr. Stressors were then removed and inserts with astrocytes were introduced. LDH assay was carried out 24 hr later. (b) Astrocytes protect cortical neurons against H₂O₂‐induced stress, the effect saturates at ~75 k astrocytes per co‐culture (n = 5, triplicates, ***p < .001 vs. control [no astrocytes insert], ****p < .0001 vs. groups with less than 50 k astrocytes, one‐way ANOVA with Tukey's post hoc analysis). (c) TX14(A) has a weak protective effect on neurons against oxidative stress (n = 6, triplicates, ***p < .001, ****p < .0001 vs. 50 nM TX14(A) group (effect of 50 nM is not significant, one‐way ANOVA with Bonferoni's post hoc analysis). D: PSAP depletion or GPR37L1/GPR37 knock‐down selectively in astrocytes significantly attenuates the protective effect of astrocytes on neurons pre‐exposed to oxidative stressors (n = 6, triplicates, ***p < .001 vs. indicated group, one‐way ANOVA with Turkey's post hoc analysis) [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7

Working model of the neuroprotective role of astrocytic GPR37L1/GPR37 based on the evidence presented in this study. Damaged neurons release diffusible “SOS” factor(s) which trigger release of PSAP. PSAP acts on GPR37L1 on astrocytes and activates release of diffusible neuroprotective factor(s) [Color figure can be viewed at wileyonlinelibrary.com]