CNS distribution, signalling properties and central effects of G-protein coupled receptor 4

Abstract

Information on the distribution and biology of the G-protein coupled receptor 4 (GPR4) in the brain is limited. It is currently thought that GPR4 couples to G_s proteins and may mediate central respiratory sensitivity to CO₂. Using a knock-in mouse model, abundant GPR4 expression was detected in the cerebrovascular endothelium and neurones of dorsal raphe, retro-trapezoidal nucleus locus coeruleus and lateral septum. A similar distribution was confirmed using RNAscope in situ hybridisation. In HEK293 cells, overexpressing GPR4, it was highly constitutively active at neutral pH with little further increase in cAMP towards acidic pH. The GPR4 antagonist NE 52-QQ57 effectively blocked GPR4-mediated cAMP accumulation (IC₅₀ 26.8 nM in HEK293 cells). In HUVEC which natively express GPR4, physiological acidification (pH 7.4-7.0) resulted in a cAMP increase by ∼55% which was completely prevented by 1 μM NE 52-QQ57. The main extracellular organic acid, l-lactic acid (LL; 1-10 mM), suppressed pH dependent activation of GPR4 in HEK293 and HUVEC cells, suggesting allosteric negative modulation. In unanaesthetised mice and rats, NE 52-QQ57 (20 mg kg^-1) reduced ventilatory response to 5 and 10% CO₂. In anaesthetised rats, systemic administration of NE 52-QQ57 (up to 20 mg kg^-1) had no effect on hemodynamics, cerebral blood flow and blood oxygen level dependent responses. Central administration of NE 52-QQ57 (1 mM) in vagotomised anaesthetised rats did not affect CO₂-induced respiratory responses. Our results indicate that GPR4 is expressed by multiple neuronal populations and endothelium and that its pH sensitivity is affected by level of expression and LL. NE 52-QQ57 blunts hypercapnic response to CO₂ but this effect is absent under anaesthesia, possibly due to the inhibitory effect of LL on GPR4.

Keywords: Antagonist; Distribution; GPR4; Lactate; Modulation; Respiration.

Fig. 1

GPR4 expression in the mouse brain. A-I. Fluorescent immunohistochemistry of GPR4-CRE-STOP-EGFP mouse brain slices using antibodies against EGFP (anti-GFP) and TH (marker of catecholaminergic neurones). Unless indicated otherwise, scale bars are 50μm. A. GPR4 expression in blood vessels was prominent throughout the whole brain (see all panels). B. Strong and wide-spread GPR4 expression in dorsal raphe nucleus neurones.C. GPR4 expression in neurones of the lateral septum. D, E. Higher magnification images from areas indicated in panel C. F. In the locus coeruleus (A6 cell group) GPR4 expression was detected in neurones, some of which were TH-positive (yellow arrows) or TH-negative (white arrows). G. Ventral edge of the medulla oblongata at the level of RTN. A few RTN neurones could be detected in some sections. Less than 10 EGFP-positive cells per animal were found. H. Some of the sections at RTN level were double-stained for TH, but TH-positive cells were EGFP-negative and probably belonged to the C1 cell group. I. Scattered neurones of unclear phenotype were sporadically detected throughout the whole brain. J-Q – Results of RNAscope FISH. J. A blood vessel, scale bar 10μm K. Dorsal raphe L. lateral septum M. Lateral septum, magnified from L, scale bar 10μm N. Locus cortuleus, boundaries highlighted by the dotted line O. C1 cell group P and Q: RTN Green arrows in N and Q – blood vessels aq – aqueduct; CC – corpus collosum; RTN – retrotrapezoid nucleus See Supplement for the large scale versions of some of the images.

Fig. 2

In vitro characterisation of GPR4 activity. A. Increases in cAMP induced by a water-soluble forskolin analogue NKH 477 and detected using Glosensor assay are stable across the physiologically relevant pH range. 20μM NKH 477 was applied to naïve HEK293 cells (not GPR4 transfected). Data represent mean ± s.e.m (n = 4 of triplicates). Differences are not statistically significant (repeated measures one-way ANOVA). B. pH-dependent cAMP accumulation in GPR4 expressing HEK293 cells depends on the quantity of DNA used for transfection and by implication on the level of the expression of GPR4. Data represent mean ± s.e.m. for averages of triplicates (n = 7 of triplicates for 0.1 μg/μl, n = 3 of triplicates for 0.01 and 0.001 μg/μl). High levels of expression shifts the activation curve towards the alkaline range. Activation profile with 0.001 μg/μl DNA leads a profile which resembles response of HUVEC (see panel 2C) which express GPR4 natively. C. cAMP accumulation in HUVEC at different pH. Increase in cAMP without NE 52-QQ57 at pH 7.1 and 6.8 is statistically significant (n = 6 of triplicates, repeated measures one-way ANOVA shows p < 0.01) D. Concentration-response curve for cAMP production evoked using non-selective β-adrenoceptor agonist isoprenaline to illustrate that HUVEC cells can mount a robust cAMP response to a different Gs-coupled GPCR. Isoprenaline activates natively expressed β2 adrenoceptors in a concentration-dependent manner (n = 3 of triplicates). -∞ indicates zero concentration of the drug.

Fig. 3

NE 52-QQ57 is a novel GPR4 antagonist working in a physiological range of pH. A. NE 52-QQ57 (formula above the bar chart) inhibits GPR4-mediated cAMP accumulation but its potency depends on proton concentration (pH). Plotted are mean % changes ± s.e.m. (n = 5 of triplicates) where % change is calculated as (with drug – baseline)/(baseline x 100) for every triplicate. Inhibition by NE 52-QQ57 becomes less effective as pH decreases. Significant differences are shown by repeated measures one-way ANOVA with Tukey's multiple comparisons post hoc test (** - p < 0.01). B. Concentration-response curve for NE 52-QQ57 at pH 7.4. Calculated IC₅₀ is 26.8 nM. Data represent mean ± s.e.m. of triplicates. Log −10 corresponds to zero concentration of the drug.

Fig. 4

LL inhibits cAMP accumulation induced by acidosis in a concentration-dependent manner in (A) HEK293 and (B) HUVEC cells. All values are normalised to control at pH 8.0 for each triplicate measurement. Two-way repeated measures ANOVA revealed a significant effect of pH (p < 0.0001) and LL (p < 0.05) on cAMP levels in HEK293 cells. Post-hoc Sidak's multiple comparisons test shows confirmed significant differences between effects of 0 and 10mM LL (p < 0.05 at pH 7.4 and 6.8, p < 0.01 at pH 7.1) and between 100 μM and 10 mM (p < 0.05), but the difference between 1 and 10mM LL was not significant (p = 0.53 at pH 7.1, p = 0.78 at pH 6.8). In HUVECs, the same analysis revealed a significant effect of pH (p < 0.0001) and LL (p < 0.05) and a significant interaction effect (p < 0.001). The post-hoc test shows significant differences between 0 and 10mM (p < 0.05 at pH 7.1, p < 0.0001 at pH 6.8), between 100 μM and 10 mM (p < 0.05 at pH 7.1, p < 0.0001 at pH 6.8), and between 100 μM and 1 mM (p < 0.05 at pH 6.8).n = 4–5 of triplicates for each data point in (A) and (B), apart from n = 3 for 1μM NE 52-QQ57 curve in HUVEC. 1μM NE 52-QQ57 antagonises pH dependent activation of GPR4, the block is overcome at acidic pH in HEK293 cells.

Fig. 5

NE 52-QQ57 blunts hypercapnic response to 5% and 10% CO2 in unanaesthetised mice (n=10) and rats (n=8). V_T and V_E measured at 5 and 10% CO₂ were significantly reduced by NE 52-QQ57 (20 mg kg⁻¹ i.p.) relative to the vehicle group (25% DMSO).**p < 0.01 and ****p < 0.0001 vs. vehicle, repeated measurement two-way ANOVA followed by Sidak's multiple comparisons test.

Fig. 6

Peripheral or central administration of NE 52-QQ57 has no effect on CO₂-induced respiratory responses. A. Averaged responses (left) and raw traces (right) to illustrate responses to hypercapnia after i.p. administration of NE 52-QQ57 (20 mg kg⁻¹). Di – raw activity of the diaphragm, ʃDi – integrated activity of the diaphragm to illustrate respiration. B. Effect of i.p. administration of NE 52-QQ57 (20 mg kg⁻¹) on baseline diaphragmatic EMG activity. C. Responses to hypercapnia after topical application of NE 52-QQ57 (1mM) on the ventral surface of the medulla oblongata. NE 52-QQ57 had no effect on CO₂-induced responses (two-way ANOVA). PN – activity of the phrenic nerve. D. Effect of direct application of NE 52-QQ57 (1mM) on the ventral surface of the medulla oblongata on baseline phrenic nerve activity. PN Rate – frequency of respiration measured from the ʃPN traces.

Fig. 7

Schematic to illustrate the putative impact of GPR4 expression level on its function. High level of expression (as exemplified by transiently transfected HEK293 cells) leads to production of copious amounts of cAMP at neutral pH (7.4) and makes the system insensitive to physiological acidification. Low, natural level of expression, such as in HUVEC, reveals GPR4 mediated responses within the physiological window (7.4–6.8) but limits the potency of GPR4-mediated signaling via cAMP, making it unsuitable for fast and robust cellular responses.