| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Baxter Laboratory in Genetic Pharmacology, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California, USA
4Correspondence: Baxter Laboratory, CCSR 4415, 269 Campus Dr., Stanford, CA 94305-5175, USA; E-mail: hblau{at}stanford.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: luminescent in vivo imaging • in vivo pharmacology • GPCRs
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Consequently, GPCRs comprise the primary class of therapeutic targets (2
, 3)
; indeed, nearly half of all currently marketed drugs modulate GPCRs (4)
. New functions are continuously being ascribed to "orphan" GPCRs, receptors without known ligands, defining new pharmacological targets for the treatment of diseases. Given the past successes in targeting this class of receptors and the opportunities lying ahead, there is a great need for novel technologies in order to improve the process of agonist and antagonist discovery. In particular, to date no method exists to monitor the activation status of a given GPCR in living animals. As a result, the pharmacologic evaluation of candidate molecules that target GPCRs relies on extrapolations from in vitro data and on indirect in vivo surrogate markers. Here we present a technology that offers two important improvements for the study of GPCRs: 1) a universal, robust, specific, high-throughput cell-based assay for GPCR activity in vitro, and 2) a bioluminescent imaging system for monitoring drug effects repeatedly in real time in live animals.
Existing in vitro methods for assessing GPCR activation suffer from different limitations. Conventionally these assays rely on detecting the activation of downstream cellular pathways through the subsequent production of intracellular second messengers such as calcium, cyclic AMP, and IP3 (5
, 6)
. Although these assays have been the mainstay of GPCR drug discovery, they generate a high rate of false positives due to the promiscuous nature of second messenger signaling, and they are not universal, as each GPCR typically produces a specific second messenger. More recently, assays have been designed to detect the binding of β-arrestin to an activated GPCR (7
8
9)
. After agonist binding, GPCRs are phosphorylated and attain a specific conformation that provides a binding site for β-arrestin. β-Arrestins serve as multifunctional adaptor proteins that direct the recruitment and subsequent scaffolding of cytoplasmic signaling complexes (10
11
12
13)
. Importantly, β-arrestin binding terminates the signals produced by the GPCR and initiates internalization of the bound receptor. This process is common to virtually all GPCRs and thus enables a single assay to be used to screen for activators and inhibitors of this class of receptors.
Here we describe a technology that enables measurement of GPCR activation, the inducible interaction of a GPCR with β-arrestin2 (arrestin) using complementation of the β-galactosidase (β-gal) enzyme, a novel assay we have recently described and extensively employed (14
15
16)
. In a separate report, we use the β-gal assay to measure GPCR internalization (17)
. The β-gal enzyme mutants we use have been optimized to provide high dissociation rates and, hence, minimal affinity, making them ideal for monitoring protein-protein interactions. Specifically, the GPCR was fused to the mutated 
peptide (*) fragment of the β-gal reporter enzyme, and β-arrestin2 to the M15 deletion mutant (the 
peptide). Using the * and mutants, we found that the interaction of the GPCR and arrestin was sufficient to force the interaction of the weakly complementing β-gal fragments resulting in enzyme activity. This system was validated using four different GPCRs yielding appropriate pharmacological responses to ligands and good signal-to-noise ratios.
The high sensitivity and specificity of the system made it possible to assay GPCR activity in live animals in real time by bioluminescence imaging. This was achieved by using a recently described luminescent β-gal substrate (Lugal), a caged luciferin molecule that can be cleaved by firefly luciferase to generate light only after it has been cleaved by β-gal (18
, 19)
. By implanting cells harboring the enzyme complementation arrestin assay into mice, we showed that the effects of systemic administration of agonists and antagonists could be monitored in living animals.
Together with its remarkable sensitivity as a cell-based in vitro system applicable to high-throughput screening, this technology can by used to assess agonist and inhibitor effects over time, in the context of the intact organism in which biodistribution and metabolism are at play, and in the milieu of competing proteins. Hence, this technology overcomes important limitations of existing methods, and could improve the speed and accuracy of future lead identification and optimization efforts.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
* fusions were created by polymerase chain reaction (PCR) amplification from cDNA constructs obtained from the UMR cDNA Resource Center. Angiotensin receptor I (AGTR1) was reverse transcriptase-PCR amplified from mouse RNA. The human β2 adrenergic receptor (β2AR) was derived from the previously described retroviral construct (7)
. The entire coding sequence of the GPCR was included minus the stop codon. The PCR products were cloned into the MfeI-XhoI sites of the MFG-YFP-H31R-IRES CD8 plasmid (16)
to create the GPCR-* retroviral fusions. The β-arrestin2- construct was created by PCR amplification of the β-arrestin2 and inserted using the MfeI-XhoI sites of the pWZL- vector as described (7)
. Firefly luciferase (Fluc) was cloned into the MFG-IRES-CD8 vector. The constitutively active Ras mutant was a generous gift from P. Khavari (Stanford University, Stanford, CA, USA). Isoproterenol, [Arg (8)
]-vasopressin, angiotensin, somatostatin-14 and propranolol were obtained from Sigma-Aldrich (St. Louis, MO, USA).
Cell line creation
The GPCR was fused to the mutated "donor" peptide (*) fragment of the β-galactosidase reporter enzyme, and β-arrestin2 to the M15 "acceptor" deletion mutant (the peptide). The * fragment is an H31R mutant of the wild-type peptide, as described previously (14)
. Due to the deletion of amino acids 11–41, the fragment is completely inactive. When the interaction of two proteins that are fused to the * and fragments is induced, the enzyme fragments complement, resulting in readily detectable increases in β-gal activity (14
15
16
17)
. The GPCR activation assay was designed to cause minimal perturbation of receptor physiology by fusing the GPCR to the smaller of the two fragments, *. Yellow fluorescent protein (YFP) was included in these constructs between the GPCR and * to provide a sortable marker for the selection of cells containing the constructs. β-Arrestin2 was fused to the amino terminus of the fragment (Fig. 1
A).
|
A cell-based system for assaying GPCR activity was generated by first transducing C2C12 cells (murine skeletal myoblast cell line originally subcloned in our laboratory; CRL-1772TM, available from the American Type Culture Collection, Manassas, VA, USA) with the retroviral arrestin- construct, linked to the hygromycin resistance gene via an internal ribosome entry site (IRES), and selection of cells with 1 mg/ml hygromycin. C2C12 cells were grown in Dulbecco modified Eagle medium (DMEM) with 20% FBS and penicillin/streptomycin. For the in vitro experiments, cells were transduced with the GPCR-* and β-arrestin2 retroviral constructs. For the in vivo experiments, cells were transduced with the Ras and Fluc plasmids. Cells were sorted using flow cytometry for CD8 expression to ensure that all cells expressed Fluc. Cells were then transduced with the GPCR* and β-arrestin2 retroviral constructs. Retroviral transduction was accomplished by lipofectamine (Invitrogen, Carlsbad, CA, USA) transfection of the ecotropic retroviral packaging cell line 
nx cells (a generous gift from Garry Nolan, Stanford University, Stanford, CA, USA). After 48 h, the supernatant was applied to the target cells in the presence of 8 µg/ml polybrene. After a 15 min incubation at 37°C, cells were spun at 2000 rpm for 30 min, then returned to a humidified 37°C incubator.
β-Galactosidase in vitro assays
Cells expressing the GPCR-* and β-arrestin2- fusions were seeded at 20,000 cells/well in a 96-well dish overnight. Ligand was added for 1 h. Medium was aspirated, and 50 µl of lysis solution (buffer A) and substrate (Gal-screen) in a 25:1 ratio was added to each well (Tropix, Bedford, MA, USA). The plates were incubated for 60–90 min at room temperature before reading in a TR717 luminometer (Applied Biosystems, Foster City, CA, USA) with an integration time of 1 s. EC50 values were calculated using nonlinear regression with a variable slope sigmoidal dose response algorithm in Prism software (Graphpad, San Diego, CA, USA). To assess the time course of activation, several wells were stimulated in parallel, and β-gal activity was measured at different time points as indicated.
Bioluminescent in vivo imaging
All protocols were approved by the Administrative Panel on Laboratory Animal Care at Stanford University School of Medicine. BALB/c nude mice were obtained from the Stanford University in-house colony. C2C12 myoblasts were harvested and suspended in 0.5% BSA in PBS at a final concentration of 108 cells/ml. Cells carrying the β2AR/arrestin β-gal/Ras/FLuc or the SSTR2/arrestin β-gal/Ras/FLuc constructs were injected s.c. into the backs of mice (4x106 cells/injection; 2 injections/mouse), and allowed to grow into small, palpable tumors over 7 to 14 days.
Imaging was carried out using a Xenogen-100 device (Xenogen, Alameda, CA, USA) as described previously (18)
. Briefly, the system is composed of a light-tight imaging chamber, a digital CCD camera cooled by a cryogenic refrigeration unit and a computer system (Living ImageTM 2.50 software, Xenogen). Imaging of β-gal activity was performed using a modification of our previously described method. Following reconstitution of Beta-Glo (Promega, Madison, WI, USA) in 2000 ml, 5 µl/g body weight of the final solution was injected i.p., and images were continuously acquired up to 1 h (integration time 120–180 s). Data were stored for subsequent off-line analysis, and images acquired 3 to 5 min after Lugal injection were used for analysis unless otherwise indicated.
Between 12 and 23 h after acquisition of baseline images, the β2AR cells were induced by i.p. injection of 6 mg/kg isoproterenol, and bioluminescent imaging using Lugal was performed repeatedly, 1, 8, 24, and 36 h later. To test specific inhibition, 3 mg/kg propranolol was injected i.p. 1 h before isoproterenol injection, and Lugal injection and imaging were performed starting 1 h after isoproterenol injection. To induce the SSTR2 cells, octreotide (0.3 mg/kg) (Novartis, Basel, Switzerland) was injected i.p. 1 h before Lugal injection and bioluminescent imaging.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
15
16
17)
. We sought to enlist this system to develop a cell-based technology to monitor the interactions of GPCRs with arrestin (Fig. 1A
).
C2C12 myoblasts were first transduced with the retroviral arrestin- construct. The resulting cell line expressing the arrestin- fusion protein was transduced with one of the following GPCR * constructs: human SSTR2-*, human β2AR-*, mouse AGTR1-*, or human AVPR2-*. Cells stably transduced to express combinations of these constructs are henceforth referred to as the GPCR/arrestin β-gal complementation system. The sensitivity, specificity and time course of activation of the resulting cell lines carrying the panel of different GPCR-* constructs and arrestin- were characterized and compared in vitro.
Measurement of GPCR activation in cells
To test the sensitivity of the GPCR/arrestin assay, each cell line was plated in a 96-well dish and treated for 60 min with escalating doses of the indicated agonist (Fig. 1B
). β-gal activity was measured using a luminescent substrate. The AGTR1 and SSTR2 cell lines showed an increase in β-gal activity of greater than 15-fold in response to the agonists. The EC50 values of their agonists, 5 nM (angiotensin) and 3 nM (somatostatin), are in good agreement with published radiolabeled ligand binding assays. The AVPR2 cell line exhibited the largest induction of enzyme activity, with a maximum that was 25x the unstimulated value. The EC50 determined for vasopressin using the present method was in the expected range (6 nM). Treatment of the β2AR cell line with isoproterenol resulted in an 8-fold increase in β-gal activity, with an EC50 of 8 nM. The time course of activation did not differ significantly among the four receptors, as expected. Using β-gal activity as a measure of arrestin binding, all cell lines were found to be induced within minutes of agonist addition and to reach 80% of their maximum values after 30 min (Fig. 1C
). Incubation times of up to 2 h can be used to further increase the signal (data not shown). Following removal of the inducer and subsequent washing, the signal returned to baseline within 12 h and could be restimulated to similar maximal levels thereafter (Fig. 1D
).
Conventional cell-based assays that rely on downstream signals are inherently nonspecific and, as a result, need to be performed in strictly defined nonphysiological conditions (typically in serum-free medium). These restrictions preclude applications that would be useful in secondary screening and lead candidate evaluation and optimization, such as in vitro testing of compounds in patient plasma or ex vivo evaluation of serum from animals treated with drug candidates. To verify that the GPCR/arrestin β-gal complementation was indeed specific, we switched the agonists from the list above (isoproterenol, angiotensin, somatostatin and vasopressin) in order to incubate cells with unrelated agonists not known to interact with the specific GPCR. These treatments did not result in increased β-gal activity for any of the receptors, validating the specificity of the assay (data not shown). In summary, the in vitro characterization of all four GPCR cell lines exhibited high sensitivity and specificity, rapid inducibility as well as reversibility, providing evidence that the cell lines constitute useful assays for primary high-throughput cell-based screening and for secondary screening under various conditions.
Imaging of GPCR activation and inhibition in living mice
We have recently described the use of the luminescent substrate Lugal for in vivo bioluminescent imaging of an intact β-gal that possesses relatively high levels of enzymatic activity when conjugated to antibodies or resulting from transgene activation in mice (18)
. This assay relies on the sequential action of two reporter enzymes, β-gal and firefly luciferase (Fluc), making it possible to combine the advantages of both. Lugal, a caged form of D-luciferin, is cleaved by β-gal to generate free D-luciferin that subsequently serves as a substrate for the ubiquitously-expressed Fluc in the final, light-emitting enzymatic step. Thus, luminescence generation is dependent on β-gal activity. To explore the possibility of applying this system to the detection of GPCR activation in living mice, C2C12 cells carrying the β2-adrenoreceptor/arrestin β-gal construct (termed β2AR cells) were transduced with retroviral constructs to express the proto-oncogene ras (to enhance cell survival) and Fluc. At 7 to 14 days after implantation of β2AR/Fluc cells into the backs of Balb/c nude mice, baseline bioluminescence images were acquired by i.p. injection of Lugal, and luminescence was imaged using a Xenogen IVIS-100 system, as described previously (19)
. Between 12 and 23 h later, isoproterenol (6 mg/kg) or vehicle was injected i.p., and imaging was repeated 1, 8, 24, and 36 h later (Fig. 2
A). A robust, 
4-fold increase in luminescence was seen compared with baseline 1 h after stimulation with isoproterenol (Fig. 2B, C
). Luminescence progressively declined thereafter and eventually returned to baseline within 24 to 36 h.
|
To determine whether this system could also be used to assess the pharmacodynamic effects of inhibitors of β2-adrenoreceptors, Balb/c nude mice were injected with the β2AR/Fluc cells (Fig. 3
A). Following acquisition of baseline bioluminescence 7 to 14 days after cell injection, mice were injected with either the β-blocker propranolol (3 mg/kg i.p.) or vehicle. One hour later, all mice were injected with isoproterenol (6 mg/kg) i.p., then imaged 1, 8, and 24 h later by i.p. injection of Lugal. As before, mice pretreated with vehicle alone prior to injection of isoproterenol showed, again, an 4-fold increase in luminescence, but luminescence generation was completely abrogated in mice that had been pretreated with propranolol (Fig. 3B, C
). These experiments showed that the isoproterenol-induced β2-adrenoreceptor activation could be effectively inhibited by systemic administration of a β-blocker and confirmed that the isoproterenol-induced luminescence was specific to β2-adrenoreceptor activation.
|
Finally, we sought to confirm that agonist-induced luminescence in vivo could be applied to other receptors. We engineered new cell lines carrying the SSTR2/arrestin-β-gal construct by retroviral transduction with the Fluc and ras genes (SSTR2/Fluc cells) (Fig. 4
A). Cell injection and bioluminescent imaging using Lugal were performed as described above. Following acquisition of baseline images, the SSTR2-expressing tumors were activated by injection of the small molecule somatostatin analog octreotide (0.3 µg/kg i.p.) or vehicle. An increase in luminescence of 1.9-fold was seen 1 h after octreotide injection. Interestingly, luminescence did not decrease immediately thereafter, but remained at the same level until 8 h after octreotide injection before finally returning to baseline after 24 h (Fig. 4B, C
). This behavior is in line with the relatively long serum half-life of octreotide (120 min), a drug specifically developed to provide longer plasma concentrations than the physiological agonist somatostatin.
|
In summary, the approach of assessing GPCR activation using low-affinity β-gal complementation fragments tethered to specific GPCRs and β-arrestin, in combination with bioluminescent sequential reporter enzyme imaging using the β-gal substrate Lugal, made it possible to image activation and inhibition of GPCR in living mice.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Key factors that ultimately determine the suitability of a compound as a candidate drug (e.g., potency, distribution, protein binding, clearance, and metabolism) can only be estimated by a series of in vitro and in vivo assays in a lengthy, error-prone process that leads to undue pipeline attrition. Clearly, the possibility of directly testing a larger number of candidate molecules at their site of action and in real-time in living subjects would greatly reduce the strain on resources and increase the accuracy of candidate selection. Radiolabeling of drugs is useful in assessing biodistribution and excretion; however, this technology requires special and costly facilities and provides no information about the pharmacodynamic activity of the compound in the physiological environment (21)
.
GPCR activation in a living subject at a given time is the result of a complex interplay of systemic and local agonists and antagonists that cannot be modeled using in vitro methods and must therefore be assessed in vivo. For example, β2 adrenergic receptors are directly activated by the endogenous ligands adrenaline (epinephrine) and noradrenaline (norepinephrine). While their plasma concentrations can be readily measured, their effective local concentration at the receptor or their biological action on target cells expressing the β2 adrenergic receptors cannot be determined in vivo. This picture is further complicated when pharmacological agents are being evaluated (22
, 23)
: 1) β-adrenoreceptor blockers possess intrinsic activity that leads to simultaneous activation and inhibition of the target receptor to various degrees; 2) β-adrenoreceptor blockers can lead to hypotension that triggers the reactive, direct or indirect, up-regulation of endogenous β-adrenoreceptor activators; and 3) the highly dynamic nature of events makes determination of circulating levels of therapeutic and endogenous ligands futile. It is surprising that, given the widespread clinical application of drugs that directly or indirectly interact with the adrenergic system, adrenoreceptor activation cannot be directly measured but, instead, must be inferred from in vitro data and secondary biomarkers (e.g., blood pressure) due to the lack of better monitoring technologies. Clearly, a method that would make it possible to directly determine the activation of a GPCR of interest in living animals would expedite the in vivo testing of drug candidates.
Development of a GPCR activation assay using β-galactosidase complementation that provides sensitive and specific signals
By coupling the and peptide fragments of the low-affinity β-gal complementation system with different GPCRs and arrestin, we have developed a novel assay that can be used for high-sensitivity screening and in vivo luminescent imaging of GPCR activation. Agonist stimulation of the receptor leads to binding of arrestin, which forces the complementation of the β-gal fragments, leading to enzymatic activity. β-gal provides the advantage of enzymatic amplification that can be read out using a wide variety of substrates, making this technology suitable for many different applications. The binding of arrestin to GPCRs occurs independently of the specific G-protein coupling of the receptor and thus provides the opportunity for a single technology to be used to detect activation of the majority of GPCRs. In this work we showed that the activation of Gs- (β2AR and AVPR2) as well as Gi- (SSTR2 and AGTR1) coupled receptors can be assayed using β-gal complementation. All four receptors tested attained a >8-fold signal-to-noise ratio using a chemiluminescent assay that is directly applicable to a high-throughput screening environment.
Luminescent imaging of GPCR activation and inhibition in living mice
The inherent specificity of the GPCR/arrestin-β-gal complementation system and the high signal-to-noise ratio suggested that this technology could yield real-time signals enabling GPCR activation to be monitored by bioluminescence imaging in complex systems in vivo. Luminescent in vivo imaging is increasingly used as a tool to monitor physiological and pathological processes as well as the pharmacological action of candidate drugs in real time (24
, 25)
. However, existing cell-based assays to monitor GPCR activation are not applicable to in vivo imaging for the following reasons: 1) lack of specificity, requiring assays to be performed under strictly defined conditions that do not exist in vivo (e.g., second messenger-based assays); and 2) lack of a suitable readout for whole-body imaging (e.g., assays of protein translocation based on fluorescence). Using cells engineered to express the GPCR/arrestin system implanted into mice, we showed that specific activation of a GPCR by a systemically administered agonist can be measured noninvasively in these animals within hours using bioluminescent imaging. Importantly, the signal can be blocked by pretreatment with an antagonist, demonstrating the specificity of the system and suggesting that this technology can be used to test inhibitors and stimulators of given GPCRs. This technology also takes full advantage of the ease of use of bioluminescent imaging, using Fluc substrates that can be injected i.p. (whereas i.v. administration is required with other imaging methods). Thus, this is currently the only method allowing "relatively high-throughput" imaging, an important feature for drug discovery.
The technology shown here was used as a cell-based system that makes it possible to perform in vitro screening first, followed by in vivo testing of agonists or antagonists using the same cell line. One potential limitation is that transplanted cells do not necessarily reflect the native local microenvironment, possibly altering the accessibility of endogenous or pharmacological modulators. Although this caveat is acknowledged, the pharmacokinetics and pharmacodynamics of the test compound would remain unperturbed in this system. Hence, the influence of modulating factors such as bioavailability, metabolism, plasma half-life, protein binding, or intrinsic effects of inhibitors can be assessed. This system could be further developed by generating transgenic mice expressing the GPCR/arrestin β-gal constructs constitutively or under the control of tissue specific elements, providing expression of the assay system in endogenous cells. Alternatively, direct gene transfer, using plasmids or adeno-associated virus, could be used to integrate the constructs into endogenous tissues.
In summary, by coupling the and peptide fragments of the low-affinity β-galactosidase complementation system with different GPCRs and β-arrestin, we have developed a novel assay that can be used for high-sensitivity screening and in vivo luminescent imaging of GPCR activation and inhibition. Notably, we have shown that the sensitivity and specificity we achieved were sufficient using weakly complemented β-gal derived from protein translocation to allow the monitoring of GPCR activation in the milieu of competing proteins in living mice.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Current address: Cardiovascular Research, Bayer Healthcare AG, Wuppertal, Germany.
3 Current address: DiscoveRx Corp., Fremont, CA, USA.
Received for publication August 14, 2007. Accepted for publication October 3, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  X. Zhao, A. Jones, K. R. Olson, K. Peng, T. Wehrman, A. Park, R. Mallari, D. Nebalasca, S. W. Young, and S.-H. Xiao A Homogeneous Enzyme Fragment Complementation-Based {beta}-Arrestin Translocation Assay for High-Throughput Screening of G-Protein-Coupled Receptors J Biomol Screen, September 1, 2008; 13(8): 737 - 747. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
