Constitutive Coupling of a Naturally Occurring Human Alpha1a-Adrenergic Receptor Genetic Variant to EGFR Transactivation Pathway

Constitutive coupling of a naturally occurring human alpha1a-adrenergic receptor genetic variant to EGFR transactivation pathway

Anush Oganesiana,1, Vladimir Yarov-Yarovoyb, William C. Parksc, and Debra A. Schwinna,d,e

aDepartment of Anesthesiology and Pain Medicine, cDepartment of Medicine, dDepartment of Pharmacology, and eDepartment of Genome Sciences, University of Washington, Seattle, WA 98109; and bDepartment of Physiology and Membrane Biology, University of California, Davis, CA 95616

Edited by Robert J. Lefkowitz, Duke University Medical Center, Howard Hughes Medical Institute, Durham, NC, and approved October 25, 2011 (received for review October 6, 2011) We previously identified a naturally occurring human SNP, G247R, in (HB-EGF), a soluble EGFR ligand generated through extracel- the third intracellular loop of the α1a-adrenergic receptor (α1a-247R) lular proteolytic cleavage of its membrane-anchored form (proHB- and demonstrated that constitutive expression of α1a-247R results EGF) (9). Binding of HB-EGF to EGFR leads to transactivation in twofold increased cell proliferation compared with WT. In the of EGFR and activation of the downstream ERK/MAPK pathway present study we elucidate molecular mechanisms and signal trans- (10). MMP/ADAM-dependent transactivation of EGFR and its duction pathways responsible for increased cell proliferation unique contribution in the development of cardiovascular disorders is an α α α α to 1a-247R, but not 1a-WT, 1b,or 1dAR subtypes. We show that intriguing and important research topic. elevated levels of matrix metalloproteinase-7 (MMP7) and a disinte- Various cardiovascular disorders, such as hypertension and α – grin and metalloproteinase-12 (ADAM12) in 1a-247R expressing heart failure, are associated with polymorphisms in genes that cells are responsible for EGF receptor (EGFR) transactivation, down- regulate the adrenergic system, mostly βARs and α2ARs (11, 12). stream ERK activation, and increased cell proliferation; this pathway We identified nine naturally occurring human SNPs in the α1aAR is confirmed using MMP, EGFR, and ERK inhibitors. We demonstrate and characterized them pharmacologically (13). The G247R that EGFR transactivation and downstream ERK activation depends SNP present in the third intracellular loop of the α1aAR was on increased shedding of heparin-binding EGF. Finally, we demon- originally identified in a patient with severe hypertension. Sev- β strate that knockdown of MMP7 or -arrestin1 by shRNAs results in eral studies suggest association of α AR genetic variants with α 1a attenuation of proliferation of cells expressing 1a-247R. Impor- human disease, and a few report associations between α AR α 1a tantly, accelerated cell proliferation triggered by the 1a-247R is SNPs and hypertension in humans (14, 15). serum- and agonist-independent, providing unique evidence for An important feature of α -247R (247R) is that it confers β 1a constitutive active coupling to the -arrestin1/MMP/EGFR transacti- a proliferative advantage to cells cultured in the absence of agonist fi vation pathway by any G protein-coupled receptor. These ndings stimulation. In this study, we report that the molecular mechanism raise the possibility of a previously unexplored mechanism for sym- for this proliferation is G protein-independent, β-arrestin1–de- pathetically mediated human hypertension triggered by a naturally pendent transactivation of EGFR and activation of the down- occurring human genetic variant. stream ERK pathway induced by elevated levels of MMP7 and ADAM12 with subsequent release of HB-EGF. This unique α α he 1-adrenergic receptors ( 1AR) are G protein-coupled constitutive activation of the MMP7/ADAM12 pathway is pre- Ttransmembrane receptors (GPCRs) that mediate actions of viously undetected for GPCRs and leads to the intriguing hypoth- the sympathetic nervous system through binding of endogenous esis that this may represent a unique mechanism for sympatheti- catecholamines epinephrine or norepinephrine. Three subtypes cally mediated hypertension triggered by a naturally occurring α α α α of 1ARs ( 1a, 1b, 1d) exist in human tissues; upon agonist human genetic variant. stimulation, α1ARs couple predominantly to the Gq/11 family of G proteins. Among the three α1AR subtypes, α1aARs pre- Results dominate in human vascular smooth muscle, particularly in re- Expression of 247R Confers Increased Cell Proliferation. The location α sistant vessels (1). Functional studies implicate 1ARs in human of the G247R substitution in the third intracellular loop is sche- vasoconstriction, hypertension, and myocardial hypertrophy, and matized in Fig. 1, showing a structural model of human α1aAR. To demonstrate an important role in regulating vascular tone (1, 2). expand our previous studies (13), we compared growth rates of Supporting these observations, genetically engineered mice with cells expressing WT or 247R with other α1AR subtypes: α1b and targeted deletion of α ARs have impaired vasopressor activity 1a α1d. Cells expressing WT, α1b,orα1d have comparable growth rates required for maintenance of normal arterial blood pressure (3), but 247R cells exhibit significantly higher proliferation (Fig. 2A). α and 1aAR antagonists lower blood pressure when administered To determine if the elevated growth rate is caused by in- to humans (4). creased affinity to serum growth factors or intrinsic signal- Stress-induced hypertrophy or increased vascular tone is char- transduction pathways activated by 247R, cells expressing each acterized by changes in the structure of blood vessels and the heart. of the α ARs were cultured without serum for 24 or 48 h. Cells fi 1 Speci cally, transactivation of the EGF receptor (EGFR) by expressing 247R exhibit ∼twofold increased proliferation com- GPCRs is one potential underlying mechanism of myocardial hypertrophy (5). Specific mechanisms by which signals are trans-

duced from GPCRs to EGFR and downstream MAPK/ ERK Author contributions: A.O., W.C.P., and D.A.S. designed research; A.O. performed re- cascade are beginning to be unraveled (6). One mechanism by search; V.Y.-Y. contributed new reagents/analytic tools; A.O. and D.A.S. analyzed data; which cross-talk between agonist-activated GPCRs and EGFR and A.O., W.C.P., and D.A.S. wrote the paper. occurs is via proteolysis of latent ligands by speciﬁc metal- The authors declare no conﬂict of interest. loproteinases (MMPs) or a disintegrin and metalloproteinases This article is a PNAS Direct Submission. (ADAMs). MMP2, MMP7, ADAM10, ADAM12, and ADAM17 1To whom correspondence should be addressed. E-mail: [email protected]. are present in arteries and have been implicated in ectodomain This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. shedding of growth factors (7, 8), such as heparin-binding EGF 1073/pnas.1116271108/-/DCSupplemental.

19796–19801 | PNAS | December 6, 2011 | vol. 108 | no. 49 www.pnas.org/cgi/doi/10.1073/pnas.1116271108 Downloaded by guest on September 29, 2021 N-terminus dependent or represents more generalizable phenomena, we TM3 TM5 EXTRACELLULAR established stable H9C2 rat cardiomyoblast cell lines expressing WT or 247R at low receptor densities and evaluated cell pro- TM1 TM2 liferation in the absence of agonist stimulation. Cells expressing 247R exhibit ∼twofold increased proliferation compared with WT or control cells (Fig. S1). These data demonstrate that enhanced proliferation triggered by 247R is generalizable. TM7 TM4 Inhibition of EGFR Activity Reverses Enhanced Cell Proliferation Induced by 247R. To determine how 247R promotes cell pro- INTRACELLULAR TM6 1iL 4iL liferation, we ﬁrst assessed whether an EGFR autocrine pathway

2iL was involved. Activation of EGFR signaling is associated with agonist-induced hypertension and promotes proliferation, func- G247R C-terminus tioning upstream of MAPK (16). EGFR transactivation depends 3iL on the activity of speciﬁc MMPs/ADAMs, which activate the latent

Fig. 1. Transmembrane view of a structural model of the human α1aAR ligands via ectodomain proteolysis (8, 17). Therefore, we evaluated receptor. The α1aAR contains seven transmembrane domains (TM), TM1 to thymidine incorporation into DNA in cells treated with inhibitors TM7, four intracellular loops (iL) colored individually, and C-terminal domain to EGFR (AG1478), MMPs (GM6001), or MEK (PD98028). Cell (C terminus). The mutation located in the third intracellular loop of the proliferation in the presence of serum is elevated ∼twofold in 247R a1aAR results in substitution of glycine (G) G247 with arginine (R). cells compared with WT (Fig. 3A). Treatment of 247R cells with EGFR or MMP inhibitors reduces thymidine incorporation, but α not to the level observed in WT cells (Fig. 3A). pared with all other 1ARs (Fig. 2B). Flow cytometry analysis of Because serum growth factors can activate additional pro- cell-cycle progression of DAPI-labeled cells demonstrates that liferative pathways and partially mask the inhibitory effects of after 24 h in serum-deprived conditions (baseline), 247R cells AG1478 and GM6001, we evaluated thymidine incorporation exhibit a higher proportion of proliferating cells compared with under serum-deprived conditions. Our data demonstrate a simi- WT (Fig. 2C). Cell-cycle analysis at 5 and 24 h in 10% serum in lar increase in thymidine incorporation in 247R cells compared the absence of agonist reveals a substantially increased pro- with WT (Fig. 3 B and C). Importantly, treatment of 247R cells portion of proliferating cells in 247R population compared with with AG1478, GM6001 (Fig. 3B) or PD98059 (Fig. 3C) reduces WT (Fig. 2C). Our ﬁndings demonstrate that expression of 247R thymidine incorporation to levels comparable with that in WT confers a robust cell-cycle progression compared with WT, α1b, or α1dAR independent of serum and agonist stimulation. To determine whether agonist- and serum-independent en- C hanced-proliferation of Rat-1 cells expressing 247R is cell type- A B160 247R 180 247R * 140 160 (cpm) (cpm) * 120 50 -3 -3 3 - 140 (cpm) * -3 A B 120 100 40 * x10 100 80 30 WT * 80 WT 60 60 20 40 40 10 20 20

Thymidine incorporation x 10 0 0 incorporation xThymidine 10 0 WT 247R 247R AG1478 - + - - + - PE ------Thymidine incorporation GM6001 - - + - - + AG1478 - + - - + - PD98059 GM6001 - - + - - +

WT 247R D E 1 2 3 4 5 6 p-ERK C 90 247R

(cpm) ERK

-3 80 70 * AG1478 - + - - + - GM6001 - - + - - + PHARMACOLOGY 60 WT * 50 30 * * F 40 * 25 30 20 * 20 15 * 10 10 Thymidine incorporation x10

0 P-ERK/ERK, PE - + + + - + + + 5 AG1478 - - + - - - + - 0 GM6001 - - - + - - - + 1 2 3 4 5 6 Relative band Relative intensity WT 247R

Fig. 3. Increased ERK activation in 247R cells is reduced to near normal Fig. 2. Constitutive expression of 247R confers unique serum-independent levels by AG1478 (EGFR), GM6001 (MMP), and PD98059 (MEK) inhibitors.

accelerated cell proliferation compared with WT, α1b or α1dAR cells. (A) Cells Thymidine incorporation assays of cells cultured in (A) the presence of 10% expressing various α1ARs were synchronized for 24 h in serum-deprived serum, (B–D) serum-deprived conditions with or without inhibitors, or (D)PE, conditions and cultured in 10% FBS or (B) serum-deprived conditions for 24 as indicated. (E) Western blots of cells cultured in the absence (lanes 1 and 4) or 48 h. Data are mean ± SE (SEM) of four independent experiments. *P < or presence of AG1478 (lanes 2 and 5) or GM6001 (lanes 3 and 6) with p-ERK 0.05. (C) Flow cytometry analysis of cell-cycle progression of synchronized or ERK antibodies. (F) Relative intensity of p-ERK protein bands, as de- WT or 247R cells cultured 5 or 24 h. The percentage of cells in proliferative termined by densitometric analyses normalized to total ERK. Data are S-phase is indicated. mean ± SEM from three independent experiments. *P < 0.05.

Oganesian et al. PNAS | December 6, 2011 | vol. 108 | no. 49 | 19797 Downloaded by guest on September 29, 2021 cells, thus demonstrating that 247R-triggered serum and agonist- findings reveal that 247R triggers agonist-independent (consti- independent proliferation occurs via MMP-dependent trans- tutive) transactivation of the MMP/EGFR pathway. activation of EGFR and downstream activation of ERK. 247R-Triggered EGFR Transactivation Mediated by Enhanced Shedding 247R-Induced Cell Proliferation is Agonist-Independent. We exam- of ProHB-EGF. EGFR transactivation can be induced by various ined the effect of phenylephrine (PE), an α1AR agonist, on cell stimuli including MMP- or ADAM-mediated cleavage/shedding proliferation and EGFR transactivation. With serum depriva- of proHB-EGF. Because the level of soluble HB-EGF in culture tion, PE slightly elevates thymidine incorporation in WT and medium of 247R cells is below the detection limit of the ELISA 247R cells (Fig. 3D). Importantly, treatment of 247R cells with used, and other researchers have reported a similar observation PE and EGFR or MMP inhibitor results in similar reduction of (9), we initially examined Hbegf mRNA levels in WT and 247R cell growth as does treatment with inhibitors without PE (Fig. 3 cells. RT-PCR revealed comparable levels of Hbegf mRNA in B vs. D). These data further confirm that 247R-triggered pro- 247R and WT cells (Fig. 4A). liferation is agonist and serum-independent. Interestingly, EGFR To determine whether cleavage of proHB-EGF is increased in and MMP inhibitors reduce proliferation of WT cells, indicating 247R cells, lysates from serum-deprived 247R or WT cells were fi PE-induced EGFR transactivation (Fig. 3D). Neither EGFR nor evaluated for the level of proHB-EGF. We observe a signi cant reduction (∼70%) in the amount of cell surface HB-EGF on 247R MMP inhibitors affect growth rate of control cells (Fig. S2A). PE fi treatment of WT and 247R cells cultured in the presence of se- cells compared with WT (Fig. 4B). These ndings support the rum significantly inhibits cell growth (Fig. S2B), but has no effect notion that increased shedding of proHB-EGF from 247R cells occurs and stimulates proliferation. These data suggest that re- on control cells (Fig. S2C), consistent with agonist-induced cell- duced levels of HB-EGF in response to 247R expression is not cycle arrest reported for Rat-1 cells expressing WT (18). a result of reduced expression of HB-EGF, but rather because of 247R-Induced Proliferation Is a Result of Activation of ERK Through increased shedding and release of active HB-EGF from 247R cells. Transactivation of EGFR. To determine involvement of ERK in To explore the role of HB-EGF in increased proliferation of 247R-induced proliferation, we immunoblotted lysates from WT 247R cells, WT cells were cultured in conditioned medium from 247R cells with or without immunodepletion of HB-EGF (Fig. or 247R cells cultured in the absence of agonist with phospho- 4C). WT cells cultured in HB-EGF-deprived medium have ERK (p-ERK) antibody (Fig. 3 E and F). Western blotting reveals ∼twofold reduced proliferation compared with medium without substantially increased levels of p-ERK in 247R cells (Fig. 3E, lane depletion. The growth rate was not different when cells were 4) compared with WT (Fig. 3E, lane 1). Treatment with EGFR or fi cultured in WT medium before or after HB-EGF depletion (Fig. general MMP inhibitor signi cantly reduces ERK phosphoryla- 4C). The significant reduction in proHB-EGF on 247R cells, tion in 247R cells (Fig. 3 E and F, lanes 5 and 6 vs. 4). When treated together with reduced thymidine incorporation when WT cells with PE, 247R cells exhibit a similar time-dependent pattern of are grown in HB-EGF–depleted 247R medium, indicate that ERK phosphorylation, again with higher levels of p-ERK com- 247R triggers a chain of intracellular events leading to increased pared with WT (Fig. S2D), thus indicating that the effect of PE on shedding of HB-EGF and subsequent transactivation of EGFR, ERK phosphorylation is not different for both cell types. which in turn activates ERK signaling and cell proliferation. To determine if basal ERK activation in 247R cells is caused by MMP/EGFR transactivation, cells were treated with EGFR MMP7 and ADAM12 mRNA Levels Are Increased in 247R Cells. We or MMP inhibitors, and cell lysates immunoprecipitated (see SI next assessed MMPs/ADAMs that are expressed in human ves- Methods) with EGFR antibody and immunoblotted with phos- sels and may be responsible for enhanced shedding of HB-EGF photyrosine antibody (Fig. S3). Compared with WT, 247R cells from 247R cells. Mmp7 and Adam12 mRNA levels are increased exhibit increased basal EGFR phosphorylation (Fig. S3, lanes 1 in 247R cells (Fig. 4D), but the levels of other MMPs did not and 4), and treatment with inhibitors significantly reduces EGFR differ or were reduced (Mmp2). To confirm involvement of phosphorylation in 247R cells (Fig. S3, lanes 5 and 6 vs. 4). These MMP7 in increased proliferation of 247R cells, we used MMP7-

A B C 120 120

WT 247R 100 100 * 80 80 * HB-EGF * 60 60 * 40 40

GAPDH % of control % of control 20 20

0 Thymidine incorporation, 0 Membrane bound pro-HB-EGF, WT 247R 247R depleted WT depleted media 247R media WT media media

D WT 247R WT 247R Fig. 4. Shedding of HB-EGF is increased in 247R cells and is caused by increased levels of Mmp7 and Adam12.(A) HB- MMP 2 ADAM 12 EGF transcript level in WT or 247R cells as determined by RT- PCR. (B) Levels of proHB-EGF in WT or 247R cells as de- MMP 7 ADAM 17 termined by ELISA. (C) Growth rate of WT cells cultured in normal or HB-EGF–depleted conditioned medium from WT or 247R cells. Data are mean ± SEM from three independent MMP 9 ADAM 10 experiments. (D) Expression of 247R triggers up-regulation of Mmp7 and Adam12 concurrent with reduction in Mmp2. GAPDH GAPDH Representative data from one of three independent RT-PCR experiments presented. *P < 0.05.

19798 | www.pnas.org/cgi/doi/10.1073/pnas.1116271108 Oganesian et al. Downloaded by guest on September 29, 2021 specific or scrambled shRNAs and examined cell proliferation. studies using cells stably expressing WT or 247R (see SI Methods). Four shRNAs were tested and all substantially reduced the lev- The results demonstrate that 247R and WT are constitutively in- els of Mmp7 mRNA (Fig. 5A). However, the more effective ternalized consistent with a previous report for WT receptor (20). shRNA1 and shRNA4 were used in subsequent experiments. Importantly, a distinct colocalization of β-arrestin1with 247R is Knockdown of Mmp7 mRNA with either shRNA significantly observed in perinuclear region of unstimulated cells (−PE), which reduces cell proliferation to near normal levels as determined by is not apparent in WT cells (Fig. 6B, and Fig. S6). Agonist stimu- thymidine incorporation assay (Fig. 5B) or cell counts (Fig. S4). lation of 247R cells (+PE) does not appear to alter colocalization These data demonstrate that MMP7 is responsible for shedding of the receptor and β-arrestin1 (Fig. 6B, and Fig. S6). There is no HB-EGF and increased proliferation of 247R cells. apparent colocalization of 247R and β-arrestin2 (Fig. S7). These data confirm constitutive recruitment of β-arrrestin1 and 247R Enhanced Agonist-Independent 247R Proliferation Is G Protein-Inde- observed in co-IP experiments. pendent but β-arrestin1–Dependent. To test for classic Gq-coupled signaling effects, we investigated cell proliferation in the pres- Discussion α ence or absence of prazosin, an 1AR inverse agonist, or Gi/o In the present study we found that the naturally occurring human inhibitor pertussis toxin (PTX). Neither prazosin nor PTX block genetic variant 247R triggers β-arrestin1/MMP7/EGFR trans- proliferation of WT, 247R, or control cells (Fig. 5 C and D and activation in an agonist-independent manner. The resultant pro- Fig. S5). These data demonstrate that enhanced proliferation of liferative rate of 247R cells is significantly higher than that of cells 247R cells is not mediated by G protein-coupling, further sup- expressing WT, α1bARs, or α1dARs, even in the absence of serum, porting our previous report that the basal levels of IP3 pro- findings supported by flow cytometry analysis of cell-cycle pro- duction are not different in WT or 247R cells (13). gression, thymidine incorporation, and cell counts. Furthermore, β-Arrestins may be involved in agonist-induced MMP-medi- 247R-induced proliferation is normalized upon treatment with ated EGFR transactivation and cell proliferation (19). We MMP or EGFR inhibitors, as well as β-arrestin1 or MMP7-specific assessed if the basal, serum-independent proliferation triggered shRNAs. Although recently agonist-independent constitutive ac- by 247R is β-arrestin–dependent. Efficiency of shRNAs to re- tivation of serotonin receptor was reported in transiently trans- duce expression of β-arrestin1 or β-arrestin2 was determined by Western blots (Fig. 5 E and F). Knockdown of β-arrestin1 but fected HEK293 cells (21), the authors did not examine whether β this finding led to EFGR transactivation and cell proliferation. not -arrestin2 reduces thymidine incorporation in 247R cells β compared with scrambled shRNA transfected cells (Fig. 5G); in Thus, to our knowledge, this report of constitutive -arrestin1/ WT cells, neither had an effect. These data demonstrate that MMP7/EGFR transactivation for any GPCR is unique. 247R-induced proliferation is β-arrestin1–dependent (Fig. 6A). EGFR transactivation through MMP-dependent shedding of HB-EGF by agonist activation of several GPCRs is emerging as Coimmunoprecipitation and Colocalization of 247R and β-Arrestin1. a new, unique mechanism of growth and hypertrophy in vascu- To further confirm that 247R induced proliferation depends on lature. MMPs comprise a family of 24 or more zinc- and calcium- constitutive interaction of 247R with β-arrestin1, we coimmuno- dependent proteases, with studies revealing the role of MMPs precipitated (co-IP) β-arrestin1 from cell lysates expressing WT or and ADAMs in agonist-induced GPCR-mediated hypertension 247R (see SI Methods). The levels of β-arrestin1 is significantly (7, 8). Some proliferative effects of agonist-activated GPCRs are higher in immunoprecipitates from 247R cells compared with WT mediated through EGFR transactivation, often by binding to cells cultured in the absence of PE (Fig. 5 H and I, lanes 3 and 1). shed ligands like HB-EGF (9). ProHB-EGF is proteolytically PE stimulation of 247R or WT cells does not affect this interaction cleaved by specific MMPs and ADAMs to release a soluble (Fig. 5 H and I, lanes 3 and 4, and 1 and 2). These data confirm growth factor that activates EGFR in an autocrine/paracrine constitutive interaction between 247R and β-arrestin1 in 247R manner (8, 10, 22, 23). Once bound by ligands, EGFR stimulates cells. To further validate these findings, we performed confocal activation of downstream ERKs that mediate cell growth and

A B C D Fig. 5. 247R Triggered serum and agonist-independent cell proliferation is β-arrestin1/MMP7-dependent and G protein- independent. (A) Mmp7-specific shRNA1 and shRNA4 significantly inhibit Mmp7 expression; GAPDH was used as an internal control. (B) 247R-Induced cell proliferation is inhibited by Mmp7 shRNAs as determined by thymidine in- −6 corporation assays. Treatment of cells with 10 M prazosin PHARMACOLOGY (C) or 100 ng/mL PTX (D) does not affect cell proliferation. (E) β-Arrestin1 or (F) β-arrestin2–specific shRNAs inhibit the Arrb1 or Arrb2 gene expression, as determined by Western blots; β-actin was used as a loading control. (G) 247R-In- H duced cell proliferation is inhibited by β-arrestin1 but not EG β-arrestin2 shRNAs, as determined by thymidine incorporation assays. Data are from one of three independent experiments, each done in triplicates with mean ± SD; *P < 0.05. (H) β-Arrestin1 constitutively interacts with 247R. IP of WT, 247R, or control Rat-1 cells using anti–HA-tag antibody I and Western blot with β-arrestin1 antibody. Bands in lanes 5 F and 6 represent nonspecific binding of β-arrestin1 to protein A/G Sepharose beads. Representative data from one of three independent experiments are shown. (I) Relative intensity of β-arrestin1 protein bands, as determined by densitometric analyses normalized to the corresponding band intensity of co-IP receptors. Data are mean ± SEM from three independent experiments. **P < 0.01.

Oganesian et al. PNAS | December 6, 2011 | vol. 108 | no. 49 | 19799 Downloaded by guest on September 29, 2021 that the V2 vasopressin receptor activates ERK in a β-arrestin– dependent and G protein-independent manner (27). However, this pathway was not shown to lead to increased DNA synthesis, possibly because of cytosolic retention of ERK in cells. Others report that the NK1R–β-arrestin complex does not retain activated ERK in cytosol, thus allowing nuclear translocation of ERK leading to cell proliferation (29). It is known that GPCRs differentially interact with β-arrestin1 and -2, depending on cell type and specific receptors (30). Differential interaction also was demonstrated in transiently transfected HEK293 cells for α1AR subtypes and β-arrestins (31). β-Arrestin1 and -2 have been shown to differentially interact with the third intracellular loop of GPCRs, such as muscarinic receptor (32). It is therefore plausible that the 247R mutation resulting from substitution of a small nonpolar glycine with a larger polar arginine residue confers conformational change of the third intracellular loop of fi Fig. 6. (A) Initial simpli ed schematic of 247R-induced serum and agonist α1aAR, facilitating its interaction with β-arrestins. Our present independent EGFR transactivation. 247R Triggers β-arrestin1–dependent study reveals that knockdown of β-arrestin1 but not β-arrestin2 activation of MMP7 and ADAM12, which shed proHB-EGF to soluble HB-EGF, reduces 247R-triggerred proliferation to near normal levels, in- – leading to activation of EGFR. Activated EGFR stimulates the Ras-Raf MEK- dicating that agonist and serum-independent proliferation trig- ERK pathway, cell-cycle progression, and cell proliferation. (B) 247R Is colo- β – calized with β-arrestin1. Live WT or 247R cells were labeled with anti-HA gered by 247R is -arrestin1 dependent. Co-IP experiments antibody (0′), incubated with or without PE (60′), fixed and stained with anti- reveal that β-arrestin1 is recruited to 247R independent of ag- arrestin1 antibody and visualized with Cy3 (red) or FITC (green) antibodies. onist treatment. These findings are consistent with constitutive Data are representative of three independent experiments. Colocalization is colocalization of β-arrestin1 with 247R but not with WT, further shown in yellow in the merged images (white arrows). (Scale bars, 5 μm.) confirming constitutive 247R–β-arrestin1 interaction. It has been shown that stably expressed α1aAR-WT constitutively internalized in Rat-1 and HEK293 cells (20, 33), whereas proliferation. In this study we report that 247R cells exhibit sig- there was negligible internalization observed in transiently nificantly lower levels of proHB-EGF than WT cells and shed transfected HEK293 cells (31). α AR-WT spends minimal time more HB-EGF into culture medium, as evidenced by HB-EGF 1a at the cell surface and exhibits predominantly intracellular lo- depletion experiments, resulting in EGFR transactivation and calization in native smooth-muscle cells and hepatocytes (34). increased proliferation. Although EGFR transactivation by some Our data are consistent with these reports and reveal pre- GPCRs, like angiotensin II (17), endothelin (24), or adrenergic (7, 8) receptors occurs only with agonist stimulation, our study is dominantly intracellular localization for both WT and 247R unique in the finding that EGFR is transactivated in 247R cells in stably expressed at near physiological levels in Rat-1 cells. We a serum and agonist-independent manner, and this process is also demonstrate that 247R is constitutively internalized and β MMP7-dependent. This pathway and downstream ERK activation interacts with endogenous -arrestin1 in the perinuclear region. was confirmed by abrogation of increased proliferation by general Failure of prazosin or PTX to inhibit increased proliferation of MMP inhibitor or MMP7-specific shRNAs, as well as EGFR- and 247R cells indicates that 247R-triggered cell proliferation is G MEK-specific inhibitors. PE stimulation of cells cultured in serum protein-independent. To the best of our knowledge, this report had the expected antiproliferative effect on both cell types, con- is unique in demonstrating serum and agonist-independent, β – α sistent with a previous report (18). Nevertheless, the proliferation -arrestin1 dependent coupling of 1aAR genetic variant, or any rate of 247R cells was still ∼twofold higher compared with WT. In GPCR, to MMP7/ADAM12-mediated (through constitutive serum-deprived conditions, PE slightly increases proliferation in shedding of proHB-EGF) transactivation of EGFR and ERK both cell types, and EGFR or MMP inhibitors reduce 247R cell activation and increased proliferation (Fig. 6A). proliferation, similar to that observed in the absence of agonist. In summary, our findings reveal a unique constitutive pathway These data further confirm constitutive agonist-independent EGFR activated by a naturally occurring G247R mutation in the third α transactivation in 247R cells. Our observations that PE-stimulated intracellular loop of human 1aAR, with subsequent activation of proliferation of WT cells is sensitive to EGFR and MMP inhibitors an autocrine loop involving β-arrestin1 and MMP/ADAM-de- reveal agonist-dependent EGFR transactivation in WT cells. In pendent EGFR transactivation in the absence of agonist or se- addition to our key finding of constitutive transactivation by 247R, rum. Constitutive proliferation of 247R cells appears to be here we demonstrate agonist-induced α1a-AR mediated EGFR caused by increased levels of MMP7 and ADAM12 and sub- transactivation, which has not been previously explored. sequent shedding of HB-EGF, which together with reduced Normally, in response to agonist binding, GPCRs couple to G levels of MMP2 that may result in increased collagen deposition, proteins and activate downstream signaling through generation raise the possibility that the presence of α1a-247R in human re- of secondary messengers. GPCR signaling is terminated by resistance vessels might lead to alterations of blood pressure by ceptor phosphorylation and subsequent binding of β-arrestin, narrowing and stiffening blood vessels (35). These findings also fi which classically desensitizes receptor and alters traf cking (25). raise the tantalizing hypothesis that a naturally occurring α1aAR Recent emerging data indicate that β-arrestins can also initiate genetic variant may be mechanistically involved in some forms of signaling in a G protein-independent manner (26, 27). It has human hypertension, and that MMP/ADAM-specific inhibitors been shown that angiotensin receptor AT1a activates ERK by may be effective therapeutic agents for the treatment of some β both pathways in parallel, involving -arrestin2 and G protein- forms of sympathetically mediated human hypertension. mediated signaling (28). β-Arrestin–dependent and G protein- independent transactivation of EGFR, and subsequent ERK Methods activation by β1ARs in response to agonist stimulation, has re- Cell Culture. Rat-1 fibroblasts stably expressing human WT hemagglutinin (HA)

cently been reported to be cardioprotective in mice chronically epitope tagged HA-α1aAR-WT or HA-α1aAR-G247R (referred to as WT and treated with catecholamines (6). In addition, it has been shown 247R) were maintained as previously described (13) (see also SI Methods).

19800 | www.pnas.org/cgi/doi/10.1073/pnas.1116271108 Oganesian et al. Downloaded by guest on September 29, 2021 Cell Cycle Analyses. Cells expressing WT or 247R were serum-starved for 24 h RT-PCR Analysis and Transfection of Cells with shRNAs. Cells were transfected (baseline), followed by culturing in the absence or presence of 10% serum. with gene speciﬁc shRNAs (SA Biosciences) using lipofectamine as described Cell aliquots taken subsequently at 5 to 8 h or 22 to 24 h were labeled with in SI Methods. a DNA-binding dye DAPI and analyzed by ﬂow cytometry, as previously described (36). Results are presented as the percentage of cells in different Detection of Membrane-Bound HB-EGF. To detect the proHB-EGF, the ELISA kit phases of the cell cycle based on their DNA content. from R&D Systems was used. Visualization was performed using SureBlue TMB Microwell Peroxidase (KPL) as a substrate. Molecular Modeling of α1aAR. A structural model of human α1a AR was generated using the Robetta protein structure prediction server (37) (see Cell Growth in HB-EGF–Deprived Medium. Cells expressing WT or 247R were SI Methods). serum deprived for 24 h, and the resultant conditioned medium was incubated in the presence or absence of HB-EGF monoclonal antibodies (R&D Thymidine Incorporation Assay. Cells stably expressing WT or 247R were plated Systems) for 2 h at 4 °C, followed by 2-h incubation with Protein A/G at 2–4 × 104 cells per well in 24- or 12-well plates and cultured in 10% or Sepharose beads (Santa Cruz). Beads were spun and collected medium was 0.5% serum, or in serum-deprived conditions in the presence or absence of added onto WT cells plated in 24-well plates at 2 × 104 cells per well. The inhibitors: 250 nM AG1478, 25 μM GM6001 (Calbiochem) or 10 μM PD98059 medium without deprivation was added into different wells of the same (Cell Signaling Technology). DMSO was used as a control vehicle. In experi- plate. Cells were cultured for 48 or 72 h and the thymidine incorporation − ments with 10 5 Μ PE stimulation, serum-deprived cells were cultured 24 h assay was performed as described above. with or without inhibitors or 10% FBS. Thymidine incorporation assays were performed as previously described (38). Statistical Analysis. Quantitative analyses of p-ERK and phospho-tyrosine (pY), β-arrestin1, and HA-tagged AR levels were performed with Adobe Photoshop Determination of Phosphorylated ERK. Cells stably expressing WT or 247R CS3. Statistical analysis was performed using a two-tailed Student t test. receptors were serum-starved overnight and treated with inhibitors, as described above. In experiments with agonist stimulation, serum starvation was ACKNOWLEDGMENTS. We thank Irina Gradinaru for excellent technical followed by 24-h treatment with inhibitors and10 μM PE. Western blots with support and Dr. D. Morris for helpful discussion of the manuscript. This work p-ERK or ERK antibody were performed as previously described (38). was supported by National Institutes of Health Grant R37HL49103 (to D.A.S.).

1. Rudner XL, et al. (1999) Subtype specific regulation of human vascular α(1)-adrenergic 20. Morris DP, Price RR, Smith MP, Lei B, Schwinn DA (2004) Cellular trafficking of human receptors by vessel bed and age. Circulation 100:2336–2343. alpha1a-adrenergic receptors is continuous and primarily agonist-independent. Mol 2. Autelitano DJ, Woodcock EA (1998) Selective activation of alpha1A-adrenergic re- Pharmacol 66:843–854. ceptors in neonatal cardiac myocytes is sufficient to cause hypertrophy and differ- 21. Labasque M, et al. (2010) Constitutive activity of serotonin 2C receptors at G protein- ential regulation of alpha1-adrenergic receptor subtype mRNAs. J Mol Cell Cardiol 30: independent signaling: Modulation by RNA editing and antidepressants. Mol Phar- – 1515–1523. macol 78:818 826. 3. Rokosh DG, Simpson PC (2002) Knockout of the alpha 1A/C-adrenergic receptor 22. Sunnarborg SW, et al. (2002) Tumor necrosis factor-alpha converting enzyme (TACE) subtype: The alpha 1A/C is expressed in resistance arteries and is required to maintain regulates epidermal growth factor receptor ligand availability. J Biol Chem 277: 12838–12845. arterial blood pressure. Proc Natl Acad Sci USA 99:9474–9479. 23. Goishi K, et al. (1995) Phorbol ester induces the rapid processing of cell surface 4. Zusman RM (2000) The role of alpha 1-blockers in combination therapy for hyper- heparin-binding EGF-like growth factor: Conversion from juxtacrine to paracrine tension. Int J Clin Pract 54:36–40. growth factor activity. Mol Biol Cell 6:967–980. 5. Rockman HA, Koch WJ, Lefkowitz RJ (2002) Seven-transmembrane-spanning re- 24. Daub H, Weiss FU, Wallasch C, Ullrich A (1996) Role of transactivation of the EGF ceptors and heart function. Nature 415:206–212. receptor in signalling by G-protein-coupled receptors. Nature 379:557–560. 6. Noma T, et al. (2007) Beta-arrestin-mediated beta1-adrenergic receptor trans- 25. Lefkowitz RJ, Shenoy SK (2005) Transduction of receptor signals by beta-arrestins. activation of the EGFR confers cardioprotection. J Clin Invest 117:2445–2458. Science 308:512–517. 7. Wang X, et al. (2009) Matrix metalloproteinase-7 and ADAM-12 (a disintegrin and 26. Shenoy SK, et al. (2006) Beta-arrestin-dependent, G protein-independent ERK1/2 ac- fi metalloproteinase-12) de ne a signaling axis in agonist-induced hypertension and tivation by the beta2 adrenergic receptor. J Biol Chem 281:1261–1273. – cardiac hypertrophy. Circulation 119:2480 2489. 27. Charest PG, Oligny-Longpré G, Bonin H, Azzi M, Bouvier M (2007) The V2 vasopressin 8. Hao L, Du M, Lopez-Campistrous A, Fernandez-Patron C (2004) Agonist-induced ac- receptor stimulates ERK1/2 activity independently of heterotrimeric G protein sig- tivation of matrix metalloproteinase-7 promotes vasoconstriction through the epi- nalling. Cell Signal 19:32–41. dermal growth factor-receptor pathway. Circ Res 94:68–76. 28. Wei H, et al. (2003) Independent beta-arrestin 2 and G protein-mediated pathways 9. Zhuang S, Kinsey GR, Rasbach K, Schnellmann RG (2008) Heparin-binding epidermal for angiotensin II activation of extracellular signal-regulated kinases 1 and 2. Proc growth factor and Src family kinases in proliferation of renal epithelial cells. Am J Natl Acad Sci USA 100:10782–10787. Physiol Renal Physiol 294:F459–F468. 29. DeFea KA, et al. (2000) The proliferative and antiapoptotic effects of substance P are 10. Prenzel N, et al. (1999) EGF receptor transactivation by G-protein-coupled receptors facilitated by formation of a β-arrestin-dependent scaffolding complex. Proc Natl requires metalloproteinase cleavage of proHB-EGF. Nature 402:884–888. Acad Sci USA 97:11086–11091. 11. Liggett SB (2010) Pharmacogenomics of beta1-adrenergic receptor polymorphisms in 30. Kuo FT, Lu TL, Fu HW (2006) Opposing effects of β-arrestin1 and β-arrestin2 on acti- heart failure. Heart Fail Clin 6:27–33. vation and degradation of Src induced by protease-activated receptor 1. Cell Signal 12. Flordellis C, Manolis A, Scheinin M, Paris H (2004) Clinical and pharmacological sig- 18:1914–1923. nificance of α2-adrenoceptor polymorphisms in cardiovascular diseases. Int J Cardiol 31. Stanasila L, Abuin L, Dey J, Cotecchia S (2008) Different internalization properties of 97:367–372. the alpha1a- and alpha1b-adrenergic receptor subtypes: The potential role of re- – 13. Lei B, et al. (2005) Novel human alpha1a-adrenoceptor single nucleotide poly- ceptor interaction with beta-arrestins and AP50. Mol Pharmacol 74:562 573.

morphisms alter receptor pharmacology and biological function. Naunyn Schmiede- 32. Wu G, Krupnick JG, Benovic JL, Lanier SM (1997) Interaction of arrestins with in- PHARMACOLOGY bergs Arch Pharmacol 371:229–239. tracellular domains of muscarinic and alpha2-adrenergic receptors. J Biol Chem 272: – 14. Freitas SR, Pereira AC, Floriano MS, Mill JG, Krieger JE (2008) Association of alpha1a- 17836 17842. 33. Chalothorn D, et al. (2002) Differences in the cellular localization and agonist-medi- adrenergic receptor polymorphism and blood pressure phenotypes in the Brazilian ated internalization properties of the alpha(1)-adrenoceptor subtypes. Mol Pharma- population. BMC Cardiovasc Disord 8:40–47. col 61:1008–1016. 15. Gu D, et al. (2006) Association of alpha1A adrenergic receptor gene variants on 34. Pediani JD, et al. (2005) Beta-arrestin-dependent spontaneous alpha1a-adrenoceptor chromosome 8p21 with human stage 2 hypertension. J Hypertens 24:1049–1056. endocytosis causes intracellular transportation of alpha-blockers via recycling com- 16. Nagareddy PR, MacLeod KM, McNeill JH (2010) GPCR agonist-induced transactivation partments. Mol Pharmacol 67:992–1004. of the EGFR upregulates MLC II expression and promotes hypertension in insulin-re- 35. Intengan HD, Schiffrin EL (2000) Structure and mechanical properties of resistance – sistant rats. Cardiovasc Res 87:177 186. arteries in hypertension: Role of adhesion molecules and extracellular matrix deter- 17. Mifune M, et al. (2005) G protein coupling and second messenger generation are minants. Hypertension 36:312–318. indispensable for metalloprotease-dependent, heparin-binding epidermal growth 36. Rostomily RC, et al. (1997) Pleomorphic xanthoastrocytoma: DNA ﬂow cytometry and factor shedding through angiotensin II type-1 receptor. J Biol Chem 280:26592–26599. outcome analysis of 12 patients. Cancer 80:2141–2150. 18. Gonzalez-Cabrera PJ, et al. (2004) Differential regulation of the cell cycle by alpha1- 37. Kim DE, Chivian D, Baker D (2004) Protein structure prediction and analysis using the adrenergic receptor subtypes. Endocrinology 145:5157–5167. Robetta server. Nucleic Acids Res 32(Web Server issue):W526–W531. 19. Fuentes LQ, et al. (2008) Vasopressin up-regulates the expression of growth-related 38. Oganesian A, Armstrong LC, Migliorini MM, Strickland DK, Bornstein P (2008) immediate-early genes via two distinct EGF receptor transactivation pathways. Cell Thrombospondins use the VLDL receptor and a nonapoptotic pathway to inhibit cell Signal 20:1642–1650. division in microvascular endothelial cells. Mol Biol Cell 19:563–571.

Oganesian et al. PNAS | December 6, 2011 | vol. 108 | no. 49 | 19801 Downloaded by guest on September 29, 2021