The Orphan G Protein–Coupled Receptor 3 Modulates Amyloid-Beta Peptide Generation in Neurons

See allHide authors and affiliations

Science  13 Feb 2009:
Vol. 323, Issue 5916, pp. 946-951
DOI: 10.1126/science.1160649


Deposition of the amyloid-β peptide is a pathological hallmark of Alzheimer's disease. A high-throughput functional genomics screen identified G protein–coupled receptor 3 (GPR3), a constitutively active orphan G protein–coupled receptor, as a modulator of amyloid-β production. Overexpression of GPR3 stimulated amyloid-β production, whereas genetic ablation of GPR3 prevented accumulation of the amyloid-β peptide in vitro and in an Alzheimer's disease mouse model. GPR3 expression led to increased formation and cell-surface localization of the mature γ-secretase complex in the absence of an effect on Notch processing. GPR3 is highly expressed in areas of the normal human brain implicated in Alzheimer's disease and is elevated in the sporadic Alzheimer's disease brain. Thus, GPR3 represents a potential therapeutic target for the treatment of Alzheimer's disease.

Accumulation of the amyloid-β peptide (Aβ) is a central pathological feature in the brain of Alzheimer's disease (AD) patients (1). Aβ is generated after sequential cleavage of the β-amyloid precursor protein (APP) by the β-and γ-secretases (24), whereas initial cleavage by the α-secretase within the Aβ sequence precludes Aβ generation. All three secretases are considered to be relevant therapeutic targets for AD. Here, we sought to identify modulators of Aβ production using a high-throughput functional genomics screen, the full-length human FLeXSelect cDNA library (table S1), which is composed of 4217 individual adenoviruses, representing the transcripts of 1905 unique genes encoding potential drug targets (5). Several primary hits affecting Aβ secretion were identified (Fig. 1A) and confirmed using independent viral stocks. Genes previously shown to modulate Aβ levels or APP processing were identified: APP, the serotonin receptor HTR2C (6), and the prostaglandin E2 receptor PTGER2 (7) led to an increase in Aβ1-42 production, whereas BACE2 repressed Aβ production (8, 9). Several other Aβ-modulating targets were identified and confirmed in the human SH-SY5Y neuroblastoma cell line. Secondary assays, including a secreted alkaline phosphatase (SEAP) assay to exclude proteins that affect general secretion or transport mechanisms (fig. S3A), RNA interference (RNAi)–mediated knockdown experiments (Fig. 1D and fig. S2), real-time reverse transcription polymerase chain reaction (RT-PCR) analysis of human brain RNA (tables S4 and S5), and mass spectrometric analysis of immunoprecipitated Aβ peptides (fig. S4), were used to prioritize the targets. Among the candidate targets, G protein–coupled receptor 3 (GPR3) was of particular interest because the GPR3 gene has been mapped to the candidate AD linkage region on chromosome 1p36.1-p34.3 (10, 11), and GPR3 is predominantly expressed in the central nervous system (10, 12). Given that Ad-GPR3 transduction of SH-SY5Y or human embryonic kidney (HEK) 293 APP770 cells led to a robust increase in Aβ1-40 and Aβ1-42 release in a dose-responsive manner (Fig. 1, B and C), we assessed the functional relevance of reduced GPR3 expression on Aβ production using an RNAi-mediated strategy. Specific siRNA duplexes efficiently suppressed Aβ1-42 secretion to 50% below control levels (Fig. 1D). Thus, endogenous GPR3 plays a role in the regulation of secreted Aβ without affecting SEAP release (fig. S3B).

Fig. 1.

GPR3 modulates Aβ generation. (A) Duplicate screening data were plotted against each other and expressed as relative (fold) standard deviation based on the mean of the sample. (B and C) Overexpression of GPR3 increased secretion of Aβ1-40 and Aβ1-42 peptides in a dose-responsive manner in HEK293 APP770 cells. The mean ± SD (n = 4) is shown. (D) Two independent siRNAs (D-003951-01 and D-003951-02) directed against GPR3 suppressed Aβ1-42 secretion to below enhanced GFP (eGFP) and GL2 (luciferase) control levels in HEK293 APP770 cells. siRNAs directed against BACE1 were used as a positive control. The mean ± SD (n = 5) is shown.

To investigate whether GPR3 regulates Aβ secretion through modulation of β-secretase activity, we determined the expression and activity of the β-secretase after Ad-GPR3 transduction. Immunoprecipitation–mass spectrometry analysis did not indicate increased β-secretase activity (fig. S4), and β-secretase expression levels were unchanged after Ad-GPR3 transduction (see below), which indicates that GPR3 modulation of Aβ generation occurs downstream of β-secretase activity. Expression of APP-C99, a direct γ-secretase substrate and immediate precursor of Aβ, together with Ad-GPR3 in primary hippocampal neurons yielded substantial increases in Aβ1-40 and Aβ1-42 release (Fig. 2, A and B), which suggests that GPR3-mediated Aβ generation occurs through modulation of the γ-secretase cleavage of APP without affecting expression of the γ-secretase subunits (fig. S5A). Furthermore, L-685,458, a highly selective γ-secretase inhibitor (13), effectively abolished the increase in secreted Aβ1-40 and Aβ1-42 upon GPR3 overexpression (Fig. 2C and fig. S5B).

Fig. 2.

GPR3 modulates activity of the γ-secretase. Release of (A) Aβ1-40 and (B) Aβ1-42 was measured in cell culture supernatants from hippocampal neurons cotransduced with Ad-APP-C99 and increasing multiplicity of infection (MOI) of either Ad-GFP or Ad-GPR3 by enzyme-linked immunosorbent assay (ELISA). The results are expressed as the mean percentage ± SD of three independent experiments performed in triplicate relative to GFP + C99 (0 MOI). (C) Primary murine neurons were transduced with Ad-APP-C99 and either Ad-GFP or Ad-GPR3 and treated in the presence or absence of the γ-secretase inhibitor L-685,458 before measurement of Aβ1-40 and Aβ1-42 secretion in cell culture supernatant samples by ELISA. Release of both Aβ1-40 and Aβ1-42 levels was clearly increased relative to GFP transduction alone. Incubation with L-685,458 abolished the increase in secreted Aβ1-40 and Aβ1-42. The results are expressed as the mean percentage ± SD of three independent experiments performed in triplicate relative to GFP (control). (D) HEK293 WT cell membrane extracts, after transduction with Ad-GFP or Ad-GPR3, were separated under native conditions by BN-PAGE and analyzed by immunoblot using antibodies that recognize the γ-secretase complex subunits (indicated in the upper left corner of each blot). Ad-GPR3 transduction enhances expression of the mature γ-secretase complex. (E) Immunoblot analysis of cell surface biotinylated proteins. Expression of the transferrin receptor (Tfr) and total cell extracts (Total) were used as controls. (F) HEK293 WT cells, transfected with empty vector (control) or GPR3 cDNA, were subject to sucrose gradient fractionation. Equivalent volumes were assessed by immunoblot using antibodies that recognize the γ-secretase subunits, RT-GPR3, and the organelle-specific markers: caveolin-1 (caveolae) and calnexin (ER).

GPR3 is an orphan G protein–coupled receptor (GPCR), although a putative ligand has been identified (14), and GPR3 constitutively elevates cyclic adenosine monophosphate (cAMP) levels through adenylate cyclase activation (15, 16), which implies that it intrinsically activates the Gprotein Gs. Elevation of intracellular cAMP levels after expression of a constitutively active mutant of the thyroid stimulating hormone receptor (TSHr), another GPCR, failed to stimulate Aβ release in HEK293 APP695 cells (fig. S13C). Similarly, expression of a GsDN mutant or inactivation of G failed to reduce the GPR3-mediated increase in Aβ generation (figs. S13 and S14). We then examined the potential effect of GPR3 expression on the assembly and/or cell surface expression of the γ-secretase complex subunits NCT, PS1-NTF, PS1-CTF, APH-1a, and PEN-2 (17). NCT immunostaining typically revealed three bands using blue native polyacrylamide gel electrophoresis (BN-PAGE) (Fig. 2D). Two bands between 150- and 440-kD represent subcomplexes of NCT–APH-1 and NCT–APH-1–PS1-CTF, respectively. The ∼ 440-kD band represents the mature γ-secretase complex (18). GPR3 overexpression led to an increase in the relative amount of mature γ-secretase complex (Fig. 2D). Furthermore, we observed increased cell surface expression of the γ-secretase subunits after Ad-GPR3 transduction (Fig. 2E).

GPCRs and the γ-secretase complex have been shown to be localized in detergent-resistant membranes (DRMs) (19, 20). Thus, we determined the localization of the γ-secretase subunits and GPR3 by differential flotation after sucrose density gradient centrifugation (Fig. 2F). Notably, the γ-secretase subunits and GPR3 partially co-distributed in low-density fractions 3 and 4. Moreover, GPR3 overexpression resulted in an enrichment of the γ-secretase subunits in low-buoyancy fractions, which suggests that expression of GPR3 leads to increased localization of the γ-secretase complex in DRMs.

The most severe side effects due to absence of γ-secretase activity are caused by deficient Notch signaling (21). Consequently, intramembranous cleavage of Notch was evaluated in HEK293 wild-type (WT) cells using a truncated form of the Notch receptor, Myc-tagged Ad-NΔE, that undergoes cleavage by the γ-secretase, releasing the Notch intracellular cytoplasmic domain (NICD) (22). Using an antibody against cleaved Notch (Val 1744), which recognizes a neoepitope in the NICD after γ-secretase cleavage, we determined that similar levels of the NICD were generated in green fluorescent protein (GFP)– and GPR3-expressing cells (Fig. 3A). The selectivity of GPR3 modulation of γ-secretase activity with regard to APP relative to Notch was confirmed by assessment of the APP intracellular domain (AICD) generation (23). Similar to the NICD, the AICD is generated by γ-secretase cleavage. Cotransduction with GPR3 led to an enhancement of AICD formation (Fig. 3B), which was prevented by the γ-secretase inhibitor L-685,458, demonstrating that release of the AICD, but not the NICD, is modulated by GPR3.

Fig. 3.

GPR3 stimulates an increase in AICD generation but does not have an effect on the S3 cleavage of Notch. (A) Notch processing is unaffected in HEK293 WT cells cotransduced with Ad-NΔE and either Ad-GFP or Ad-GPR3. Ad-GFP and/or Ad-NICD transduction were used as controls. Lactacystin (10 μM) was used to prevent degradation of the NICD, which was visualized with a cleavage-specific antibody (Notch1 Val-1744). GPR3 expression levels were not affected by lactacystin. (B) Cotransduction of HEK293 WT cells with Ad-APP-C99 and Ad-GPR3 led to increased AICD generation. Ad-GFP alone was used as a negative control. The γ-secretase inhibitor L-685,458 abrogated AICD and Aβ generation.

To establish the physiological relevance of Aβ modulation by GPR3, we expressed GPR3 in vivo in an AD mouse model expressing KM670/671NL (Swedish) mutated APP and the L166P mutated PS1 (APP/PS1) (24). We performed stereotactic injection of the purified GPR3 adenoviral vector or a control GFP vector into the hippocampus of 3-month-old APP/PS1 transgenic mice. Mice were killed after 1 week, and the hippocampus was removed and prepared for analysis. The contralateral cerebral hemisphere was not injected and was used as an internal control. Hippocampal expression of GPR3 enhanced Aβ1-40 and Aβ1-42 generation in vivo (Fig. 4A) without affecting γ-secretase expression (fig. S6). We then evaluated the endogenous physiological relevance of GPR3 with regard to APP proteolytic processing using Gpr3+/+, +/–, and –/– primary hippocampal neuronal cultures. Release of Aβ1-40 and Aβ1-42 was substantially reduced in Gpr3–/– relative to Gpr3+/+ and Gpr3+/– neuronal cultures (Fig. 4B). Expression of the γ-secretase subunits and Ad-APP-C99 was indistinguishable among the genotypes (fig. S7); however, mature γ-secretase complex formation was reduced in Gpr3–/– neurons (fig. S8). Interestingly, reconstitution of GPR3 expression was sufficient to restore Aβ release in Gpr3–/– neurons relative to control (Fig. 4C and fig. S9), establishing the contribution of endogenous GPR3 toward Aβ generation.

Fig. 4.

In vivo modulation of GPR3 expression affects Aβ generation in an AD mouse model. (A) Hippocampal concentrations of soluble Aβ1-40 and Aβ1-42 levels after stereotactic injection of either Ad-GFP or Ad-GPR3 were determined by ELISA (**, P = 0.0038; ***, P = 0.0033 by the unpaired Student's t test: n = 12 independent animals per treatment group). (B) Release of Aβ1-40 and Aβ1-42 was considerably reduced in hippocampal neuronal Gpr3–/– cultures relative to the Gpr3+/+ and Gpr3+/– cultures (***, P < 0.001 relative to +/+ or +/–). (C) Release of Aβ1-40 and Aβ1-42 was rescued after Ad-GPR3, but not Ad-GFP, transduction in E17 Gpr3/ hippocampal neurons cotransduced with Ad-APP-C99 (**, P < 0.01 relative to GFP (+/–) Abeta 40; *, P < 0.05 relative to GFP (–/–) Abeta 40, GFP (+/–) Abeta 42, or GFP (–/–) Abeta 42. (D) Hippocampal concentrations of soluble Aβ1-40 and Aβ1-42 levels in APP/PS1 transgenic mice crossed with Gpr3+/+, Gpr3+/, or Gpr3/ mice were determined by ELISA (**, P = 0.003 relative to Gpr3+/+; ***, P = 0.0003 relative to Gpr3+/+ for Aβ1-40 levels, **, P = 0.002 relative to Gpr3+/+; ***, P = 0.001 relative to Gpr3+/+ for Aβ1-42 levels by the unpaired Student's t test: n = 6 independent mice per cross). (E) Immunolocalization of GPR3 in tissue sections from normal post-mortem human brains without neurological disease using an antibody specific for GPR3 (red). Counter nuclear staining (hematoxylin) is shown in blue. (F) Representative immunoblot analysis and densitometric quantification of GPR3 expression in control and AD patient brain samples suggest an elevation of GPR3 in AD patients (**, P = 0.016 by the unpaired Student's t test: n = 14 patients per group).

We then investigated the in vivo consequence of the absence of GPR3 on Aβ generation in an AD mouse model. APP/PS1 transgenic mice were mated with Gpr3–/– mice. Both hemizygosity and complete genetic ablation of GPR3 expression resulted in a dramatic reduction in Aβ1-40 and Aβ1-42 levels (Fig. 4D) in the absence of an effect on γ-secretase expression (fig. S10). Thus, endogenous GPR3 is involved in Aβ generation.

In normal human brain tissue sections, GPR3 is strongly expressed in neurons in the hippocampus, amygdala, cortex, entorhinal cortex, and thalamus, regions of the brain that strongly correlate with the pathogenesis of AD (Fig. 4E) (11, 25). GPR3 expression is also elevated in the brains of sporadic AD patients relative to age-matched controls (Fig. 4F), which provides support for the involvement of GPR3 in AD.

We have defined a role for GPR3 in γ-secretase modulation of Aβ generation in vitro and in vivo. Given that mice deficient in Notch signaling display an embryonic lethal phenotype and die on about embryonic day 9.5 (21) and in view of the apparent absence of a Notch loss-of-function phenotype in Gpr3-deficient mice, which survive to adulthood and have only been reported to exhibit a reproductive defect thus far (26), this correlative in vivo evidence supports the in vitro studies, which suggests that GPR3 does not modulate NICD generation. GPR3 appears to promote complex assembly of the γ-secretase, resulting in increased trafficking of the γ-secretase components and the mature γ-secretase complex to the cell surface and increased localization in DRMs, which eventually leads to an increase in Aβ generation. Activation-induced receptor-mediated endocytosis of the β2-adrenergic receptor, together with the γ-secretase complex, also results in an increase in Aβ production (27), although a role for the β2-adrenergic receptor in Notch processing is not clear.

Thus, the level of expression of GPR3 regulates localization of the γ-secretase complex, thereby affecting the amyloidogenic processing of APP, which suggests that GPR3 is an interesting AD therapeutic target. The finding that GPR3 is involved in assembly of the γ-secretase complex also suggests an intriguing regulatory mechanism for proteolytic activity of the γ-secretase, whereby a pool of inactive intracellular subcomplexes can be mobilized to assemble into mature complexes and redistribute to specific microdomains in the cell membrane to differentially affect the processing of APP and Notch. If similar regulation mechanisms are involved in the cleavage of other γ-secretase substrates, this might provide a level of specificity to the promiscuous proteolytic activities of this intriguing complex.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S14

Tables S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article