Bicarbonate signalling via G protein-coupled receptor regulates ischaemia-reperfusion injury

Abstract

Homoeostatic regulation of the acid-base balance is essential for cellular functional integrity. However, little is known about the molecular mechanism through which the acid-base balance regulates cellular responses. Here, we report that bicarbonate ions activate a G protein-coupled receptor (GPCR), i.e., GPR30, which leads to G_q-coupled calcium responses. Gpr30-Venus knock-in mice reveal predominant expression of GPR30 in brain mural cells. Primary culture and fresh isolation of brain mural cells demonstrate bicarbonate-induced, GPR30-dependent calcium responses. GPR30-deficient male mice are protected against ischemia-reperfusion injury by a rapid blood flow recovery. Collectively, we identify a bicarbonate-sensing GPCR in brain mural cells that regulates blood flow and ischemia-reperfusion injury. Our results provide a perspective on the modulation of GPR30 signalling in the development of innovative therapies for ischaemic stroke. Moreover, our findings provide perspectives on acid/base sensing GPCRs, concomitantly modulating cellular responses depending on fluctuating ion concentrations under the acid-base homoeostasis.

Fig. 1. Physiological concentration of bicarbonate ions activate GPR30 in vitro.

a–f Calcium mobilisation assay in Fluo-8-loaded MCF-7 (MCF-GPR30, a–e) and HEK293 (HEK-GPR30, f) cells stably expressing human GPR30 (hGPR30). The increase in the intracellular calcium level was evaluated by maximum relative fluorescence units (RFU) minus minimum RFU (max − min). ATP activates endogenous purinergic receptors and serves as a positive control. a Cells treated with vehicle, oestradiol (E_2, 10⁻¹¹–10⁻⁷ M), aldosterone (10⁻¹¹–10⁻⁷ M), and ATP (10⁻⁸–10⁻⁴ M). b Cells treated with vehicle, Dulbecco’s modified Eagle’s medium (DMEM), and ATP (25 µM). c MCF-GPR30 cells were treated with the mixed inorganic solution at various pH values, Minimum Essential Medium (MEM), and DMEM. d The depletion of sodium bicarbonate from the mixed inorganic solution abolished the increase in intracellular calcium levels in MCF-GPR30 cells. e, f Sodium bicarbonate and potassium bicarbonate, but not potassium chloride, increased intracellular calcium levels in MCF-GPR30 (e) and HEK-GPR30 (f) cells in a dose-dependent manner. Statistical analysis: two-tailed unpaired t-test with Dunnett’s correction (a) or Bonferroni’s correction (e and f) after two-way ANOVA. Two-tailed unpaired t-test with Holm-Šídák’s correction (b–d). Data are presented as mean values ± SEM. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 2. Bicarbonate ions activate GPR30 signalling through the G_q family of G proteins.

a TGFα shedding assay using HEK293 cells devoid of G_αq/11/12/13 and transfected with the indicated subtypes of G_α proteins. Blue and magenta bars indicate G_αi and G_αq family candidates, respectively. b, c cAMP assay using HEK293 cells transiently (b) and stably (c) expressing hGPR30. Sodium bicarbonate did not inhibit forskolin (FSK)-dependent cAMP production in either cell line. d Calcium mobilisation assay using Fluo-8-loaded HEK-GPR30. The cells were treated with vehicle, 3.3 mM, and 11 mM NaHCO₃. Bicarbonate-induced intracellular calcium increase was completely abolished by the G_αq/11/14 inhibitor YM-254890 (YM, 1 µM, 45 min) but not by pertussis toxin (PTX, 100 ng/ml, 16 h). e, f HEK293 cells transiently expressing hGPR30 were pretreated with YM-254890 (e) or PTX (f) and then treated with 11 mM NaHCO₃ for the indicated periods. β-actin served as the loading control. g GPR30-dependent phosphorylation of ERK1/2 was evaluated using western blotting. HEK293 cells transiently expressing hGPR30 were treated with vehicle, 11 mM NaHCO₃, or 100 nM E₂ and harvested at 15 and 30 min after stimulation. β-actin serves as a loading control. h, i GPR30-dependent accumulation of inositol phosphates. HEK293 cells transiently expressing hGPR30 (h) and MCF-mGPR30 cells (i) were treated with indicated sodium bicarbonate concentrations. Nonlinear regression (four parameters) was performed. The EC₅₀ values of sodium bicarbonate were 11.40 mM (h) and 12.20 mM (i). Statistical analysis: two-tailed unpaired t-test with Bonferroni’s correction after two-way ANOVA (a–d). Data are presented as mean values ± SEM. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 3. Identification of amino acid residues essential for GPR30 activation by bicarbonate.

a, b The predicted model of human GRP30 was generated based on the active conformation of CC chemokine receptor 5 (PDB:707 F, https://www.rcsb.org/structure/7O7F) in combination with other similar coordinates obtained from GPCRdb (https://gpcrdb.org/). The candidate residues for bicarbonate coordination were selected based on their side chain properties (e.g. positively charged residues: Arg (R), His (H), and Lys (K)) (a) and their conservation across animals (b) and mapped on the model structure. c, d TGFα shedding assay using HEK293 cells transfected with hGPR30 (c) or HA-tagged hGPR30 (d). The mutants H307A, E115A, and Q138A are highlighted in green, red, and blue, respectively. e The candidate residues analysed in our study were mapped on the model structure. f Three residues identified as important for bicarbonate-dependent activation were mapped on the model structure. g Western blotting analysis for HA-tagged hGPR30 mutants (upper panel) and β-actin control (lower panel). h Calcium mobilisation assay using HEK293 cells stably expressing three mutants (E115A, Q139A, and H307A) of hGPR30. Statistical analysis: ^$p < 0.05, ^#p < 0.0001 compared to mock cells using two-tailed unpaired t-test with Bonferroni’s correction after two-way ANOVA (c, d). Two-tailed unpaired t-test with Bonferroni’s correction after two-way ANOVA (h). Data are presented as mean values ± SEM. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 4. Endogenous GPR30 is activated by bicarbonate ions.

Calcium mobilisation by sodium bicarbonate in Fura-2-loaded mouse myoblast C2C12 endogenously expressing GPR30. a, b C2C12 cells (n = 89 cells in one experiment) were stimulated with the indicated concentrations of sodium bicarbonate and 100 µM ATP. Representative image (a) and quantification (b). c Knockdown efficiency (87–90%) of Gpr30 in puromycin (2–8 µg/ml)-resistant stable C2C12 cells was assessed using quantitative RT-PCR. d Puromycin (4 µg/ml)-resistant C2C12 cells stably expressing the control or Gpr30-shRNA were stimulated with the indicated concentrations of sodium bicarbonate. Statistical analysis: two-tailed paired t-test with Bonferroni’s correction after repeated measures one-way ANOVA (b) or two-tailed unpaired t-test with Bonferroni’s correction after two-way ANOVA (c, d). Data are presented as mean values ± SEM. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 5. GPR30 is expressed in brain mural cells.

a Design of the Gpr30-Venus-KI construct. The coding sequence of Gpr30 was replaced in frame with that of Venus. b Confocal microscopy image of a heterozygous Gpr30-Venus-KI (Gpr30^+/Venus) mouse brain. Slices (50 µm) of the brains were stained with an antibody for collagen type IV (Col-IV) and DAPI. Scale bars, 100 µm. c, d Quantitative RT-PCR analyses of marker genes for neurons, astrocytes, oligodendrocytes, oligodendrocyte progenitor cells, pericytes, and endothelial cells. Relative expression levels of the genes in Venus⁺ cells compared with those in the whole brain cortex are shown in (d). e In situ hybridisation analyses of brain sections from wild-type (Gpr30^+/+) mice. Gpr30 mRNA is visualised as magenta dots. Scale bars, 50 µm. f Visualisation of Gpr30 (green), the vascular smooth muscle cell (SMC) marker Acta2 (red), and the mural cell marker Pdgfrb (magenta) using multiple in situ hybridisation. Scale bar, 50 µm. g Visualisation of Gpr30 (green), endothelial marker Pecam1 (red), and pericyte marker Pdgfrb (magenta) using multiple in situ hybridisation. Scale bars, 10 µm. h, i Quantitative RT-PCR analyses of the marker genes for neurons, astrocytes, oligodendrocytes, oligodendrocyte progenitor cells, pericytes, and endothelial cells. CD41⁻CD45⁻CD31⁺CD13⁻ and CD41⁻CD45⁻CD31⁻CD13⁺ microvascular cells were isolated from the brain cortex of Gpr30^+/+ mice. Relative expression levels of the genes compared with those in the whole brain cortex are shown in (i). Statistical analysis: two-tailed unpaired t-test with Holm-Šídák’s correction (i). Data are presented as mean values ± SEM. P values are shown if significant. N.D. indicates not detected. Source data are provided as a Source Data file.

Fig. 6. GPR30 expressed in brain mural cells is activated by bicarbonate.

a–c Calcium mobilisation assay for endogenous GPR30 in primary culture. The primary culture of Venus-positive cells harvested from the brain cortex of Gpr30-heterozygous (a, Gpr30^+/Venus) and Gpr30-null (b, Gpr30^Venus/Venus) mice were subjected to calcium imaging. c Quantification of a) and b) as area under the curve. d–i Calcium mobilisation assay for freshly isolated SMCs (d–f) and pericytes (g–i). Freshly isolated mCherry-positive cells harvested from the cerebral arteries (d, e) and cortex (g, h) of Gpr30-heterozygous (d, g, Gpr30^+/iCre; Rosa26-GCaMP6) and Gpr30-null (e, h, Gpr30^iCre/iCre; Rosa26-GCaMP6) mice were subjected to calcium imaging. f, i Quantification of d and e, g and h, as area under the curve, respectively. Statistical analysis: two-tailed unpaired t-test with Bonferroni’s correction after two-way ANOVA (c, f, i). Data are presented as mean values ± SEM. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 7. GPR30 deficiency protects against ischaemia-reperfusion injury after transient middle cerebral artery occlusion.

a Experimental design of a transient middle cerebral artery occlusion (MCAO) model. b Modified Neurological Severity Score (mNSS) evaluated over time after MCAO. c, d Evaluation of the blood–brain barrier disruption. Representative images (c) and quantification (d) of IgG staining of whole brain sections 3 days after MCAO. Scale bars, 1 mm. e, f Evaluation of infarct volume. Representative images (e) and quantification (f) of cresyl violet staining of whole brain sections 3 days after MCAO. Scale bars, 1 mm. g, h Evaluation of apoptosis using TUNEL staining of whole brain sections 3 days after MCAO. Representative images (g) and quantification (h) of hemibrain sections 3 days after MCAO. The ratio of TUNEL-positive nuclei to total nuclei was calculated on hemibrain sections. TUNEL-positive apoptotic cells were scarcely detectable in GPR30-deficient (Gpr30^Venus/Venus) mice. Scale bars, 50 µm. Statistical analysis: b two-tailed mixed-effects analysis with Bonferroni’s multiple comparison correction. Data are presented as the median ± interquartile range. d, f, h two-tailed unpaired t-test. Data are presented as dot plots with the median. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.

Fig. 8. Rapid recovery of cerebral blood flow in the MCA region upon reperfusion in Gpr30-deficient mice.

a, b Measurement of serum concentration of bicarbonate ions during ischaemia and reperfusion. a Experimental design. The modified Neurological Severity Score (mNSS) was evaluated at 45 min of MCAO, and blood was collected before (−10 min) and after ( + 5 min) reperfusion. b Changes in serum bicarbonate levels in Gpr30^+/+ and Gpr30^+/Venus (left), Gpr30^−/−, Gpr30^−/Venus, and Gpr30^Venus/Venus (middle), and sham-operated (right) mice. A set of data collected at pre- and post-reperfusion from the same mouse are connected with a line (left box). The difference in serum bicarbonate concentration (Δbicarbonate (mM) post − pre) presented as dot plots with mean (right box). c–e Magnetic resonance angiography (MRA) analysis before and after reperfusion. c Experimental design. d Maximal intensity projection images before (top) and after (middle) reperfusion. Delineated MCA and branches after reperfusion are shown (bottom). e Vessel volume of the left MCA and its branches (reperfusion side) as a ratio to that of the right (control side). f–h Measurement of cerebral blood flow using laser Doppler flowmetry. f Experimental design. g Normalised blood flow as a ratio to the basal blood flow. h Area under the curve of the blood flow from 0 to 20 min during the reperfusion phase. Statistical analysis: b two-tailed paired t-test. e, h the main effect in two-way ANOVA. Data are presented as dot plots with median. P values are shown if significant. ns indicates no significant difference. Source data are provided as a Source Data file.