PHYSIOLOGICAL REVIEWS   Vol. 78 No. 1 January 1998, pp. 35-52
Copyright ©1998 by the American Physiological Society

G Protein-Coupled Receptors: Functional and Mechanistic Insights Through Altered Gene Expression

DANIEL K. ROHRER AND BRIAN K. KOBILKA

Department of Molecular and Cellular Physiology and Howard Hughes Medical Institute, Stanford University, Stanford, California

I. INTRODUCTION
II. GENETIC APPROACHES TO STUDY G PROTEIN-COUPLED RECEPTORS
    A. Transgenic Techniques
    B. Antisense Techniques
    C. Receptor Knockout Techniques
III. NEW INSIGHTS INTO G PROTEIN-COUPLED RECEPTOR FUNCTION: ILLUSTRATIVE EXAMPLES
    A. Endothelin Receptor
    B. Thrombin Receptor
    C. ₂-Adrenoceptors
    D. ₁-Adrenoceptors
    E. Applying Different Methods to the Same Receptor
    F. Modification of Genes Associated with G Protein-Coupled Receptor Signaling
IV. ADAPTING PHYSIOLOGICAL TECHNIQUES TO THE MOUSE
V. SUMMARY
REFERENCES

	ABSTRACT

Top References

Rohrer, Daniel K., and Brian K. Kobilka. G-Protein Coupled Receptors: Functional and Mechanistic Insights Through Altered Gene Expression. Physiol. Rev. 78: 35-52, 1998.  G protein-coupled receptors (GPCRs) comprise a large and diverse family of molecules that play essential roles in signal transduction. In addition to a constantly expanding pharmacological repertoire, recent advances in the ability to manipulate GPCR expression in vivo have provided another valuable approach in the study of GPCR function and mechanism of action. Current technologies now allow investigators to manipulate GPCR expression in a variety of ways. Graded reductions in GPCR expression can be achieved through antisense strategies or total gene ablation or replacement can be achieved through gene targeting strategies, and exogenous expression of wild-type or mutant GPCR isoforms can be accomplished with transgenic technologies. Both the techniques used to achieve these specific alterations and the consequences of altered expression patterns are reviewed here and discussed in the context of GPCR function and mechanism of action.

	I. INTRODUCTION

Top Next References

The signaling function of a large variety of hormones, neurotransmitters, and even sensory stimuli is mediated through G protein-coupled receptors (GPCRs). The large array of GPCRs, numbering in the hundreds, represent a superfamily of proteins with both structural and functional homology (131). The GPCR superfamily can be subdivided into three major subfamilies: the rhodopsin/-adrenergic, vasoactive intestinal polypeptide (VIP)/secretin, and the metabotropic glutamate receptor families. A GPCR transduces signals across the plasma membrane by interacting with ligand on the extracellular surface and undergoing a conformational change that modifies the function of an associated intracellular G protein. The most highly conserved regions within this superfamily are the seven hydrophobic transmembrane segments. The domains of low sequence conservation include the NH₂ and COOH termini and intervening loop regions; GPCR topography places the NH₂ terminus and three intervening loop regions on the extracellular surface, with three intervening loop regions and the COOH terminus residing intracellularly. In the case of the bioamine receptors, the ligand binding pocket is largely formed by the transmembrane helices, whereas many of the peptide and/or glycoprotein GPCRs also rely on the extracellular loop and NH₂-terminal regions for binding. The poorly conserved intracellular loops and COOH terminus encode the domains that are involved in G protein interaction, cellular trafficking, and agonist-induced desensitization.

In addition to the large diversity of GPCRs, the G proteins with which they interact also constitute a large family of homologous members (22), and the demonstration that some GPCRs can interact with more than one type of G protein reveals the complexity of signaling possibilities (45). The large host of end-stage effectors modulated by G proteins, such as adenylyl cyclases, phospholipases, and ion channels, gives rise to the extraordinary range and subtlety of signaling that can be achieved in this system.

One of the fundamental issues facing investigators studying GPCRs is the ability to dissect out the mechanistic and functional differences between GPCR subtypes, especially among highly conserved members. The ability to do so is critical, for in many cases the need to discriminate between related subtypes can be of therapeutic significance. With a therapeutic approach in mind, there are two interdependent issues commonly addressed. One is to identify the receptor subtype(s) that modulates the physiological process in question. The other is to discover ligands that are selective for that receptor or set of receptors. Their interdependence derives from the fact that while many ligands have aided in the identification and classification of distinct receptor subtypes, distinct GPCR subtypes that have been identified by other means such as cloning drive the search for greater selectivity in ligands. From a therapeutic standpoint then, the ultimate goal is to develop selective agents for each GPCR subtype.

Unfortunately, traditional pharmacological methods are frequently unsuitable or incapable of discriminating between highly related GPCR subtypes, especially in vivo. A striking example of this would be the 14 members of the serotonin receptor family (58), which presumably interact with the same endogenous ligand, but for which only a few selective agents are available. So although subtype-selective drugs are an invaluable resource, in many instances either their relative selectivity is poor, the receptor subfamily size is intractably large, or the predominance of receptors in a given target tissue disproportionately favors occupation of a nonrelevant subtype.

The ability to clone and express the genes encoding GPCRs has been instrumental to our understanding of GPCR function and mechanism of action, and has thus far succeeded in identifying far more GPCRs than there are natural ligands. In some instances, this has led to the identification of so-called "orphan receptors" that possess high sequence conservation to members of the GPCR superfamily yet have no identified ligand(s) (99). Other important properties of GPCR function can be addressed using the techniques of molecular biology as well. The spatial and temporal expression pattern of the receptor in question can be studied in great detail using in situ hybridization, reverse transcription-coupled polymerase chain reaction, and other mRNA quantitation methods. The disadvantage of these approaches is that mRNA studies fail to measure receptor protein levels, which do not necessarily correlate with mRNA levels. Having access to the cloned receptor also makes it possible to transfect and express GPCRs in cultured cells at high receptor densities, thereby permitting more detailed pharmacological analysis, as well as analysis of G protein coupling behavior and second messenger generation. Frequently such studies can be performed in heterologous cell types that are null for that GPCR subtype or any related subtypes, so the information is unequivocal. It should be noted, however, that overexpression of receptors in heterologous cell systems does not always faithfully reproduce normal coupling behavior and frequently cannot provide a physiological readout that intact organs, tissues, or whole animals can provide.

Along with the advent of cloning and expression techniques has come the ability to manipulate gene expression in vivo, which represents perhaps the most powerful method to analyze GPCR function to date. Once a given GPCR gene has been cloned, the technology exists to 1) delete the gene from the entire genome; 2) delete the gene within specific tissues or in an inducible fashion; 3) exogenously express the gene at higher than normal levels or in a tissue-specific manner; 4) express mutant or antisense forms of the receptor, or exchange mutated forms of the receptor gene for wild type counterparts; and 5) use GPCR regulatory sequences in conjunction with reporter genes to monitor the developmental and tissue-specific patterns of expression. These varied approaches now allow investigators to explore many aspects of GPCR function in development and physiology, especially when comparing highly related members. Such techniques also frequently reveal novel or unexpected roles for GPCRs and, in the case of orphan receptors, can be used as an investigative tool to understand their endogenous functions and aid in identification of the natural ligand.

Although such techniques have provided a wealth of information on GPCR function and expression, these approaches are also likely to gain use as tools for drug testing and/or discovery. Thus drugs that are highly subtype selective should have no effect on mice lacking the gene for that specific subtype. A response to the drug would indicate interaction with closely related subtypes or even more distantly related subtypes. Furthermore, the ability to manipulate GPCR expression in the whole animal has been used to create models of human disease (53, 97, 144). Such models can be used to track developmental progress of the disease or to test therapeutic interventions. In either case, the ability to manipulate GPCR expression in the whole animal represents an exciting opportunity to both reveal new insights into GPCR mechanism and function, as well as to provide new therapeutic approaches for the treatment of disease. In this review, we will provide specific examples of how the techniques of in vivo genetic manipulation have been used to study GPCR function and discuss the impact they have had on our understanding of GPCR biology.

	II. GENETIC APPROACHES TO STUDY G PROTEIN-COUPLED RECEPTORS

Top Previous Next References

The ability to stably introduce genetic material into animals has greatly facilitated our understanding of GPCR function and mechanism of action. The most frequent method for generating transgenic animals involves microinjection of DNA into the male pronucleus of a fertilized mouse egg, transfer of microinjected eggs into pseudopregnant females, and normal delivery of mice. Some portion of these animals will integrate the injected DNA into the host genome, and such founder animals (whose cells are frequently mosaic for the transgene) can transmit the integrated product to progeny. Barring any deleterious phenotypic effects, the transgene can then be transmitted to subsequent generations in a Mendelian fashion. Such techniques have also been applied to rats, pigs, goats, and other mammals (103, 104), but by far the most frequent recipient is the mouse, due to its short generation time and ease of use. Other techniques to introduce genes into animals include viral infection (146), direct DNA injection (154), or incorporation of genetically engineered cells into host animals or embryos (57). Each of these approaches has its own particular advantages and disadvantages, as is discussed in section II, A-C, in the context of GPCRs.

The ability to introduce foreign DNA is also accompanied by the ability to specifically manipulate the introduced genetic material. Genes of interest can be fused to any number of tissue-specific, inducible, or globally expressed regulatory elements in an attempt to localize or specify expression patterns of the transgene. The expressed gene can be the wild type, the result of site-specific mutagenesis, or represent a chimeric union between multiple gene products. The transgene can also be expressed in an antisense orientation to suppress endogenous gene expression. Although the transgenic technique provides a great deal of versatility, there are some aspects of gene delivery that are problematic. It can be difficult to control transgene copy number. Trangenes are injected as linear DNA fragments that form concatamers of varying lengths before integration into the chromosomal DNA. Moreover, the site of integration of introduced DNA into the genome is frequently nonspecific, and position effects are not uncommon (111, 160). Such position effects influence transgene expression patterns and can alter the expression of genes near the site of integration. These limitations are frequently addressed by creating and analyzing several independent lines to provide a spectrum of transgene copy numbers and to eliminate the effect of individual mouse lines showing positional effects. The discovery of one such insertional mutant, the chakragati (ckr) mouse, may prove to be a valuable tool in studying GPCR function. This mouse was serendipitously created by standard transgenic mouse technology, and the transgene appears to have interrupted a locus involved in motor coordination and behavior, affecting brain dopamine receptors (34, 117). Such transgenic lines can inadvertently provide new tools in the dissection of the GPCR signaling cascade and the genetic influences that lead to proper receptor expression and brain development.

Several transgenic animals have been made to characterize regulatory control regions within GPCR genes. When fused to reporter genes like lacZ or the gene encoding chloramphenicol acetyltransferase (CAT), these regulatory sequences can be used to map spatial and temporal GPCR expression patterns. Frequently, such studies are useful for very detailed analyses of GPCR expression patterns, especially when binding or receptor autoradiography lack specificity, but they can also be used to identify novel promoter or enhancer elements that regulate tissue or developmental expression patterns.

Regulatory sequences from the 5'-flanking regions of the dopamine D_1A receptor, serotonin (5-HT₂) receptor, follicle-stimulating hormone (FSH) receptor, and rhodopsin have all been used in conjunction with reporter genes to analyze receptor-specific expression patterns in mice (42, 43, 124, 139). In the case of the D_1A receptor, 6.4 kb of 5'-flanking sequence was fused to the lacZ reporter gene, and the pattern of lacZ expression in the brain was analyzed in transgenic mice. As a result of the sensitivity of reporter gene assays and the specificity inherent in the technique (obviating interference from related receptor subtypes), it was possible to demonstrate that within the striatum, lacZ expression was confined to neurons within the matrix, and not in striosomes (124). Promoter regions derived from the 5-HT₂ or FSH receptors fused to the CAT can direct brain- and gonadal-specific expression, respectively (43, 139), whereas various lengths of the bovine rhodopsin promoter fused to lacZ direct rod and cone specific expression (42). As well, GPCR regulatory regions can theoretically provide reagents for cell- or tissue-specific expression of other gene cassettes. An example of such an approach is the expression of phenylethanolamine N-methyltransferase (PNMT) under the control of dopamine -hydroxylase regulatory sequences, which effectively converts noradrenergic nerve terminals to an adrenergic phenotype (74). Although these reagents are not derived from GPCRs themselves, this method can be exploited to study the biosynthetic enzymes involved in catecholamine synthesis, and in turn, GPCR function and behavior. Such studies demonstrate that the transgenic technique can be useful in characterizing receptor expression patterns with high sensitivity and specificity, but it must also be noted that transgene expression patterns can be quite variable when comparing independent lines derived from the same construct. As previously noted, integration site and position effects can influence transgene expression, sometimes producing expression patterns that are ectopic or incomplete in comparison with the endogenous gene (43). Thus, for the analysis of GPCR expression patterns, the traditional transgenic technique is now frequently supplanted by gene targeting techniques (see sect. IIC), which possess higher fidelity for proper expression patterns in vivo.

Overexpression of nonmutated GPCRs in transgenic animals has been useful in characterizing the mechanism(s) of receptor activation, as well as the functional consequences of receptor overexpression. Overexpression of the human ₂-adrenergic receptor (AR) specifically targeted to the hearts of mice using the -myosin heavy chain promoter has profound effects on cardiac function. In the absence of -AR agonist, heart rate, dP/dt_max , and adenylyl cyclase activity are elevated to levels comparable to those observed in wild-type mice after maximal stimulation with isoproterenol. In the ₂-AR transgenics, there is little additional stimulation of these parameters by isoproterenol (96). These results suggest that ₂-ARs can couple to chronotropic and inotropic responses in the heart, and also demonstrate that persistent ₂-AR signaling can occur in the absence of agonist. The same transgenic line was later used as a model to test the theory of inverse agonism, and the results were used to provide support for the two-state model of GPCR activation (13). In this case, the transgenic model provided a convenient platform to test both receptor function and theory.

Atrial-specific overexpression of the cloned human ₁-AR has also been achieved in mice (11) and, similar to the ₂-AR transgenic model, shows evidence of constitutive signaling in the absence of agonist. In contrast to the ₂-AR overexpressors, however, basal and stimulated adenylate cyclase in ₁-transgenics is actually lower than the wild-type counterparts (91) and suggests that ₁- and ₂-AR coupling mechanisms may be quite distinct.

Transgenic mice in which the thyroglobulin promoter was used to direct expression of the adenosine A₂ receptor specifically to the thyroid displayed glandular hyperplasia and severe hyperthyroidism, eventually leading to premature death (81). These data suggested that A₂ receptor overexpression can lead to constitutive signaling and that some naturally occurring thyrotropin receptor mutants (like the A₂ receptor, coupled to G_s) could cause hyperfunctioning thyroid adenomas (112). These mice could be valuable animal models for studying hyperfunctioning adenomas and provide further evidence for the mitogenic properties of some GPCRs (reviewed by van Biesen, Ref. 145).

The creation of transgenic mice bearing engineered mutations in GPCRs has also been used to study GPCR function. A constitutively activated mutant of the _1b-AR whose expression is targeted specifically to the heart using the -myosin heavy chain promoter shows evidence of cardiac hypertrophy, including increased heart weight-to-body weight ratios, increased myocyte cross-sectional area, and elevated diacyglycerol content, a hallmark of ₁-AR activation (97). Interestingly, these animals only overexpress the mutant _1b-receptor by ~3-fold over normal levels, whereas two independent studies of transgenic mice expressing the wild-type _1b-receptor at ~20- to 50-fold above normal levels show no evidence of hypertrophy (W. Koch and H. Rindt, personal communication). The creation of a heritable model of cardiac hypertrophy can be exploited to dissect the molecular pathways of the hypertrophic response, whereas the failure of wild-type _1b-receptor to induce hypertrophy demonstrates subtype-specific differences in basal GPCR coupling behavior (in contrast with ₂-AR, ₁-AR, and A₂ transgenics).

A naturally occurring polymorphism in the human ₂-AR (T164I) has also recently been used to create transgenic mice overexpressing the ₂-AR receptor specifically in the heart, again using the heart-specific -myosin heavy chain promoter (144). Similar to the original ₂-AR transgenic (96), overexpressed wild-type ₂-ARs also result in enhanced cardiovascular function, whereas mice with the overexpressed mutant ₂-AR in comparison display impaired cardiac function (144). These findings are interpreted to suggest that this polymorphism may have pathophysiological consequences in humans. Although both transgenic approaches have taken the ₂-AR out of its normal cellular context, both studies nonetheless provide valuable information regarding GPCR function and coupling behavior.

Transgenic techniques are most often used to overexpress a component of a signaling pathway; however, they can also be used to overexpress modified, nonfunctional proteins that can compete with and suppress the activity of endogenous proteins. This "dominant negative" approach was used to examine the importance of the -adrenergic receptor kinase (-ARK) on cardiac -receptor function by creating mice that overexpressed either -ARK or a -ARK inhibitor (-ARK_i). The -ARK_i is a nonfunctional competitor of -ARK activity derived from the COOH terminus of -ARK itself. The cardiac-specific expression of these genes resulted in attenuated -AR agonist signaling in the case of -ARK and a basal enhancement of cardiac contractility in the case of -ARK_i , elegantly demonstrating the importance of in vivo desensitization mechanisms on heart function in both basal and stimulated states (76).

The application of the transgenic technique has been particularly fruitful in studying the mechanism of rhodopsin signal transduction and the molecular basis of heritable visual disorders. A number of studies have been able to recapitulate the phenotype of the human disease retinitis pigmentosa (RP) by expressing transgenes with analogous mutations to those commonly found in RP patients (20, 41, 57, 83, 105, 106, 110, 114, 119, 134, 151). These studies have effectively documented the aberrant signaling properties and accelerated photoreceptor cell degeneration associated with various rhodopsin mutants. As one would expect, many of these studies show positive correlations between transgene dose and severity of disease (57, 83, 110). Specific mutations in rhodopsin have also been created to examine the role of phosphorylation and desensitization in the visual signaling cascade, providing important information on the role of desensitization in GPCR signaling (41). As well, detailed mechanisms of the retinal degeneration have been investigated, including one study in which chimeric mice were created by aggregation of wild-type embryos with embryos from transgenic mice expressing mutant rhodopsin (57). Despite the chimerism, these studies showed uniform retinal degeneration, suggesting that the retinal degeneration associated with the RP phenotype is not limited to the genetically altered photoreceptor cells. Together with other transgenic studies, these results have allowed investigators to invoke a disease model whereby mutant rhodopsin undergoes abnormal cellular trafficking, triggering accelerated cellular apoptosis.

Although traditional transgenic techniques are widely applicable, in some instances this method of gene delivery is inappropriate, such as when excess copy numbers of the transgene or manipulation of the genome have deleterious consequences. Transgenic strategies for single-copy, site-specific integration of genes have been developed (17), and the advent of inducible transgenes now allows for graded and precise control of transgene expression. Examples of such inducible systems include the Tet-transactivator (35, 71) and ecdysone inducible systems (108). Although such inducible systems have not yet been applied to GPCRs, the ability to regulate the timing and degree of transgene expression will extend the range and power of experiments directed at the analysis of GPCR behavior. This is particularly true when the developmental role of a given GPCR is being studied, since the timing of transgene expression is under direct control by the investigator.

Viral delivery methods can also be used to express foreign genes, and there are a variety of vectors and methods available (10, 47, 126, 153). Although use of viral delivery methods to the study of GPCRs has been carried out (62), such approaches have not been widely exploited. Clearly, the ability to deliver and express GPCR genes at selected times during the development of the animal is advantageous. Attempts to couple such delivery systems with a tissue-specific component is also a worthwhile goal. Most transgenic strategies attempt to direct tissue-specific expression in some fashion, and although some viral infection mechanisms display tissue tropism, tissue-specific delivery is usually not regulatable. Tissue-specific viral delivery has been reported (69), and a promising approach for the future is exemplified by the experiments of Federspiel et al. (32), who showed that tissue-specific expression of the avian leukosis virus (ALV) receptor rendered the expressing tissue susceptible to ALV infection. A transgenic method for circumventing DNA microinjection can be used in early mouse embryos that express the tv-a receptor (the receptor for subgroup A avian leukosis and sarcoma viruses), rendering them susceptible to ALV infection (33). Novel delivery methods such as these will undoubtedly enable researchers in the future to tailor delivery methods to their respective needs.

The use of antisense approaches to reduce or eliminate the expression of specific genes was initially reported over 10 years ago (66). While the basic goal remains the same, the technical repertoire has been expanded and refined to make the technique feasible for in vivo work. In some instances, this approach can provide as much or more information than transgenic or knockout techniques at a substantial savings in cost and time. Furthermore, many of the genetic approaches such as those using transgenic or knockout technology are limited in their animal host range (most frequently using the mouse), whereas in theory, many of the current antisense technologies are applicable to virtually any species and can be applied at various developmental stages. One major disadvantage of the antisense approach, however, is its empirical nature. Optimizing the region and length of antisense homology can be problematic and must be determined by trial and error.

The capacity to reduce expression using antisense can derive from several different mechanisms. Central to all of these is the specific base-pair interactions between the target and the antisense probe. In the case of plasmid or viral expression systems that express antisense RNA, loss of expression can occur when RNA-RNA hybrids are degraded by endogenous double strand-specific ribonucleases (RNases). In the case of antisense inhibition by oligodeoxynucleotides, the RNA-DNA hybrid can be degraded by RNase H. There are also examples of oligonucleotides invading chromatin DNA, forming triple helices, or causing translational arrest [for recent reviews, see Crooke and Bennett (26) or Albert and Morris (4)]. In all of these instances, a critical balance is struck between the extent to which expression is repressed and the specificity for a single gene product. The extent of repression is governed by the abundance of antisense probe relative to the target mRNA and appears to require ~20-fold excess for antisense RNA, and several hundredfold excess for oligonucleotides (4, 52, 66). At extremely high levels of antisense probe, specificity will be lost by cross-hybridization to related targets, although such losses in specificity can be mitigated by designing antisense probes to unique, nonconserved regions of the gene to avoid cross-hybridization.

Early applications of antisense technique frequently used expression vectors to produce antisense transcripts from (usually) large fragments, and although still in use, this technique has been more recently replaced by the use of antisense oligonucleotides. The finding that short synthetic oligonucleotides (15-50 nt) can 1) suppress expression via antisense mechanisms and 2) gain entry to cells in vitro or in vivo without the need for complicated transfection or infection protocols make it an ideal choice for many investigators. The use of more stable phosphorothioate oligonucleotide derivatives to avoid loss of probe from nuclease digestion has become almost standard practice (26, 168). The use of catalytic antisense ribozymes has also gained favor recently as an antisense strategy to reduce protein expression (128).

Direct injection of antisense oligonucleotides into the brain has become a popular tool in the study of GPCR function. There are many examples of its use to delineate the role of specific GPCR subtypes in various behaviors or physiological properties. For instance, when antisense oligonucleotides directed against ₂-adrenoceptor subtypes a and c were injected into the locus ceruleus of the rat, the hypnotic response to the ₂-specific agent dexmedetomidine was specifically blocked only in _2a-antisense-treated rats and unaffected in _2c-treated rats. Furthermore, sensitivity to dexmedetomidine could be recovered in _2a-antisense-treated rats after cessation of treatment (100).

In the case of brain dopamine (D) receptors, three known subtypes have been targeted by antisense strategies, and in virtually all examples, antisense reduction in dopamine receptor expression is limited to the targeted subtype. The D₁ receptor antisense treatment results in decreased D₁ agonist-induced grooming behavior and D₁ agonist-induced rotational behavior of lesioned mice, while leaving unaffected the stereotyped behavior induced by D₂ agonists (163). Likewise, antisense treatment of D₂ receptors in rats (162) or mice (115, 165) results in specific loss of D₂ receptors and reduced D₂ agonist-mediated locomotor and grooming behaviors, with no alterations in D₁ agonist-induced behaviors (162, 165). Effects could also be seen in the absence of agonist treatment on spontaneous locomotor activity and cataleptic behavior (162). Antisense inhibition of D₃ receptors has revealed a possible role for this subtype in the mediation of dopamine synthesis (107). These experiments are particularly informative, since a high degree of specificity could be achieved for each knockout. Interestingly, in these experiments, the reduction in dopamine receptor density varied from 19 to 48%, but in all cases, a phenotype could be observed. This suggests that the antisense technique can provide useful information even when disruption of receptor expression is not complete. Moreover, from a mechanistic standpoint, it also suggests that there is little or no reserve of surplus receptors in dopamine receptor signaling. This is also apparent in gene knockout experiments where heterozygous mice can be distinguished from wild-type mice (8).

The full complement of known opiate receptors (ORs) has also been targeted by antisense strategies. An antisense-directed reduction in brain -, - or µ-subtypes is accompanied by decreases in locomotor hyperactivity and/or antinociceptive effects induced by receptor subtype-specific agonists (2, 12, 21, 101, 140, 141). Similar to the ₂-adrenoceptor and dopamine receptor antisense experiments, antisense effects on opiate receptor expression are reversible after cessation of oligonucleotide delivery. Additionally, antisense inhibition of -ORs has been achieved through an intrathecal administration route, blocking the action of intracerebroventricularly administered -endorphin (140). The power of the antisense approach is exemplified by experiments that demonstrate that antisense oligonucleotide directed against -OR preferentially blocks the antinociceptive effects to a putative ₂-OR agonist, but not a putative ₁-OR agonist, while antisense directed against a conserved region shared by all three known opioid receptors can block the effects of -, µ-, and -OR agonists (including both ₁- and ₂-agonists) (12). These results have been interpreted to suggest that multiple -OR subtypes exist, a conclusion that would be difficult to make on pharmacological grounds alone.

Antisense inhibition of angiotensin II type 1 (AT₁) receptors has also been carried out and has effectively shown that the hypertensive effects of angiotensin II are mediated by AT₁ receptors in the central nervous system (48) as well as by AT₁ receptors in vascular smooth muscle and the adrenal gland (65). In both cases, antisense inhibition of AT₁ expression had preferential effects on spontaneously hypertensive rats, raising the possibility that hereditary or congenital alterations in the renin-angiotensin system resulting in hypertension may be treatable by an antisense approach. Interestingly, Iyer et al. (65) were able to show that one-time, systemic delivery of a retroviral construct expressing AT₁ receptor antisense was able to suppress AT₁ expression and lower blood pressure up to 73 days after delivery. The AT₁ receptor antisense oligonucleotides have also been used to demonstrate the dipsogenic role of this subtype in rats (121). The data gained from antisense approaches to study angiotensin receptors point to the general advantages and disadvantages inherent in oligonucleotide or transgenic delivery strategies. Although oligonucleotide-based methods are titratable and reversible, they must be regularly or continuously administered. Viral or transgenic delivery of antisense, on the other hand, is more persistent, but less regulatable.

In addition to the ability of antisense strategies to control hypertension, recent experiments have also demonstrated a convincing antisense effect on another clinically relevant model, that of bronchial asthma (109). These investigators showed that delivery of aerosolized phosphorothioate antisense oligonucleotides directed against the adenosine A₁ receptor provided significant protection against adenosine and allergen challenges. This antisense treatment was able to significantly reduce adenosine receptor density in airway smooth muscle, suggesting the route of administration is a viable one for antisense treatment and that direct modulation of GPCR expression in the lung can have important functional consequences.

Antisense oligonucleotide strategies have also been applied to the inhibition of the GPCRs for neuropeptide Y (89, 133, 150), serotonin (15), cholecystokinin (CCK)_B/gastrin (116, 147), vasopressin (80), oxytocin (95), corticotrophin-releasing hormone (85), and muscarinic M₁ receptors (161). Notably, the CCK_B/gastrin receptor antisense experiments provided evidence for subtype diversity in this receptor family. Antisense inhibition of this receptor led to a loss of gastrin-induced acid secretion in the stomach but had no effect on CCK-8-induced acid secretion (116).

Despite the broad range and utility of the antisense approach, there are examples of nonideal behavior or alternative mechanisms for antisense action (26). For instance, constant intravenous infusion of neuropeptide Y-Y₁ receptor antisense oligonucleotides resulted in a modest reduction in the vasoconstrictor response to neuropeptide Y in rats, whereas the "sense" oligonucleotide control actually resulted in exaggerated neuropeptide Y responses (133). Likewise, antisense inhibition of vasopressin V₁ receptor resulted in specific deficits in arginine vasopressin-induced social memory tasks and anxiety-related behavior as expected, whereas mRNA levels for this receptor actually increased after antisense treatment. Like the neuropeptide Y-Y₁ antisense experiments, the sense oligonucleotide-treated animals also showed phenotypic changes, distinct from those of random oligonucleotide-treated animals (80).

Although low cost and ease of use are attractive benefits of this technique, they must always be weighed against the potential confounding factors mentioned above. Despite such shortcomings where specificity and degree of knockout can be problematic, antisense techniques remain extremely useful, especially in situations in which germ line deletion is lethal or when specific organs like the brain or lungs are to be targeted. Expression of GPCRs may be particularly amenable to antisense inhibition given the low level at which most subtypes are expressed and the fact that phenotypic changes can be seen with only modest reduction in GPCR levels.

2. Tissue-specific and inducible knockouts

In all of the examples mentioned thus far, gene disruption is carried through the germline; that is, all cells of the body contain the targeted allele through all stages of development. There are many situations that call for either a tissue-specific gene targeting or a developmentally timed targeting event, both of which are now feasible. In the currently applied version of tissue-specific targeting, the initial gene modification involves targeted insertion of loxP (locus of crossing over) recognition sites for the cre (causes recombination) recombinase enzyme, which recognizes these sites and mediates excision of the intervening sequence between the recognition sites. The targeted gene having loxP sites (referred to as the "floxed" locus) is present in all cells of the body throughout development, and the loxP sites are placed such that expression of the floxed gene is unaffected. The tissue-specific targeting is achieved by tissue-specific expression of cre, which will mediate excision of the floxed sequence only in those cells in which cre is expressed. By these means, the gene of interest will remain and be functional in all non-cre-expressing cells, but will be lost in cre-expressing cells. The first in vivo application of this technique was to tissue-specifically target the DNA polymerase  gene (pol) within the T-cell lineage. The original intent of the experiment was to study the role of pol in T-cell gene rearrangement, but whole body knockouts resulted in lethality, so alternative strategies were required to study the role of pol in T cells. The floxed DNA pol was specifically excised in T cells by mating mice carrying a floxed DNA pol to mice that expressed a cre transgene only in T cells (

pol

	III. NEW INSIGHTS INTO G PROTEIN-COUPLED RECEPTOR FUNCTION: ILLUSTRATIVE EXAMPLES

Top Previous Next References

Transgenic and knockout experiments have yielded a great deal of useful information about the function and physiology of GPCRs. In many cases, the results confirm or clarify results obtained using pharmacological approaches. However, there are several examples of knockout or transgenic experiments yielding unexpected or novel findings. These results have revealed new therapeutic targets for some GPCRs and have provided insight into new modes of receptor signaling. An elegant example of the way in which knockout experiments can uncover novel roles for GPCRs is illustrated by the genetic disruption of the endothelin B (ET_B) receptor. Since their discovery in 1988, investigators have primarily focused on the role of endothelins and their receptors in blood pressure regulation and their potential role in the pathophysiology of hypertension, stroke, and heart disease (120). Disruption of the gene encoding the ET_B receptor, however, results in a phenotype identical to the long-known mouse mutant piebald-lethal (s^l), which may have a human correlate in the hereditary form of Hirschprung's disease (55). Surprisingly, the primary phenotypic observation among these ET_B mutants is a perturbation in the function of neural crest-derived cell lineages, manifested in coat color spotting and megacolon. These findings were not predicted based on pharmacological studies and demonstrate the utility of this approach in identifying the role of specific GPCRs in development. The ability to deliver pharmacological agents in utero can be difficult, especially if subtype selectivity is desired, and knockout strategies may in some cases be the only means to definitively ascertain the role of a given GPCR in development or early signaling events.

Disruption of the gene coding for the cloned thrombin receptor revealed several unexpected results. Like the ET_B receptor knockout, mice lacking the cloned thrombin receptor (tr) show developmental abnormalities, resulting in significant lethality around embryonic days 9-10. Perhaps most surprising was the observation that the absence of this receptor had no effect on bleeding time, thrombin-dependent platelet aggregation, or ATP secretion (23). The cloned receptor that was used to engineer this knockout possesses a pharmacological profile that matches quite closely the profile which alters platelet function, so it came as some surprise that these functions were preserved in thrombin receptor knockouts (tr^/). It was also observed that fibroblasts but not platelets from the tr^/ mice were deficient in thrombin receptor signaling. As a result of these findings, the authors concluded that a second thrombin receptor exists and is important in mouse platelet function. This type of study is in many ways typical of the results that can be obtained from knockout experiments: it both reveals a new and unexpected role for the gene in question and provides indirect evidence for the existence of other signaling components acting in the same pathway.

The ₂-adrenergic receptors (₂-AR) consist of three highly homologous family members (a, b, and c). The three ₂-receptor subtypes couple to the same G proteins and have similar pharmacological profiles, making it difficult to determine the functional role of each subtype in vivo. Gene targeting techniques have now been applied to all three ₂-AR subtypes, advancing our knowledge of their function and providing new impetus for developing better subtype-selective agents (87, 88, 90). As outlined in Table 1, the cardiovascular effects of disrupting the genes encoding ₂ARs a, b, and c are quite distinct. From these studies, it has been suggested that the _2a-AR is the primary ₂-AR responsible for the clinically beneficial hypotensive effect of ₂-AR agonists in the control of hypertension, whereas the _2b-AR subtype is largely responsible for the peripheral hypertensive effect of ₂-AR agonists. Although disruption of _2c-ARs failed to produce any hemodynamic effects in mice, other studies provide evidence that this subtype does play a subtle role in modulating other processes (122). Function of _2c-AR was studied in wild types, _2c-AR knockouts, and transgenic mice overexpressing the _2c-AR by approximately threefold under the control of the natural promoter. This study demonstrated that the hypothermic effect of ₂-agonists was dependent on _2c-AR gene dosage, such that knockouts displayed a reduced hypothermic effect and transgenic overexpressors displayed an enhanced hypothermic effect compared with wild types after administration of the ₂-agonist dexmedetomidine. This study also showed an _2c-AR gene dose-dependent effect on levels of the dopamine metabolite homovanillic acid in the brain. The use of knockouts and transgenics to modulate gene dosage has also been effectively applied to study the hemodynamic effects of angiotensinogen (70).

In the case of -ARs, knockout of the ₁-AR was used to show that this subtype is solely responsible for the chronotropic and inotropic effects of nonspecific -AR agonists in the heart (Table 1) (118). Traditionally, it had been believed that both ₁- and ₂-ARs were coupled to these responses in vivo, since both subtypes are expressed in cardiac myocytes and conduction tissue, and both subtypes are known to couple to adenylate cyclase stimulation. The ₁-AR knockout studies showed, however, that the ₂-ARs that were expressed in the heart could neither couple to adenylate cyclase nor mediate chronotropic or inotropic responses. This is somewhat unexpected in light of transgenic experiments mentioned above. In transgenic mice that overexpress ₂-ARs specifically in the heart, the overexpressed ₂-ARs have been shown to couple to adenylyl cyclase stimulation as well as positive chronotropic and inotropic responses. In these transgenic experiments, however, the ₂-ARs are expressed at ~50-200 times their normal level (96, 144). Although these transgenic experiments clearly demonstrate that ₂-ARs can couple to these responses in cardiac myocytes, the ₁-AR knockout experiments demonstrate that they do not do so when expressed at physiological levels. This conclusion is also supported by studies that have shown that cardiac ₂-ARs couple to inhibitory, pertussis toxin-sensitive G proteins (156). The ability to manipulate -AR gene expression in vivo by multiple independent means has thus allowed for a revised model of -AR coupling behavior that was originally based solely on pharmacological data. The results from the ₂-AR transgenic mice also led to the suggestion that gene therapy with -ARs may be a future therapeutic goal for various models of heart failure (96). Although this does not fully agree with the majority of clinical evidence suggesting that -AR antagonists are beneficial in the treatment of heart failure (30, 39, 50), the ability to effect long-lasting physiological changes through defined manipulations of GPCR expression suggests that gene therapy aimed at some component(s) of the GPCR signaling cascade may be a viable therapeutic approach for the treatment of some diseases.

Functional studies on other GPCRs have also benefited from the use of multiple methods to regulate gene expression. In addition to the aforementioned studies on the ₂-ARs, the angiotensin II receptor AT₁ has been targeted by gene disruption as well as antisense oligonucleotides and retroviral antisense expression vectors (6, 64, 65, 93, 121, 132). In all cases, inhibition of AT₁ expression led to decreased blood pressure. More specifically, this could be achieved by either total gene disruption, systemic antisense gene delivery, or intracerebroventricular injection of antisense oligonucleotides. In the case of dopamine receptors, the D₁ , D₂ , and D₃ subtypes have been targeted by antisense oligonucleotides as well as gene knockouts, and although the studies on D₁ and D₃ receptors did not perform analogous tests, both the D₂ receptor knockout as well as D₂ antisense experiments found a motor impairment upon loss of receptor (8, 115, 162, 165). Likewise, either genetic ablation or antisense treatment of the µ-opioid receptor results in loss of morphine-induced analgesia (21, 94, 129). Given the caveats inherent in either transgenic, knockout, or antisense experiments, the ability to combine the results from such disparate approaches provides an opportunity to verify the functional role(s) ascribed to various GPCRs. Our increasing ability to spatially and temporally control the delivery of oligonucleotides, expression of transgenes, or excision of gene fragments will only add to the power and subtlety of these techniques to more clearly define mechanisms of action and function of GPCRs.

The use of transgenic and knockout models targeting the enzymes involved in catecholamine biosynthesis has been particularly informative in dopaminergic and adrenergic signal transduction mechanisms. Mice lacking the tyrosine hydroxylase (TH) gene fail to synthesize any catecholamines (73, 167), whereas mice lacking the gene encoding dopamine -hydroxylase fail to synthesize norepinephrine or epinephrine (137). The ability to mate mice lacking the TH gene with transgenic mice expressing TH only in noradrenergic cells allows for the creation of mice deficient only in dopamine synthesis (166), whereas transgenic mice expressing PNMT under the control of the dopamine -hydroxylase gene promoter convert noradrenergic nerve terminals to adrenergic nerve terminals (74). Likewise, the role of neurotransmitter reuptake in GPCR function has been addressed by the disruption of the gene coding for the dopamine transporter (40).

Transgenic or knockout animals have also been created to alter the expression of other GPCR ligands such as the endothelins (79, 86), angiotensinogen (70), agouti (72, 77, 98, 152), and neuropeptide Y (31). Effectors of signal transduction downstream of GPCRs in the signaling cascade such as G_s (37), -ARK (76, 67), and the cyclic nucleotide-gated cation channel (18) have also been specifically altered, and all have significantly impacted our models of GPCR function and mechanism of action.

The ability to mate mice bearing distinct genetic alterations represents perhaps one of the most intriguing prospects for the future. G protein-coupled receptors with redundant functions can be revealed by creation of double knockouts, and in the case of knockout lethality, gene rescue experiments can be performed by crossing the knockout to mice expressing tissue-specific transgenes. The creation of dopamine-deficient mice, noted above (166), is but one example of how the large array of GPCR-related knockouts and transgenics can be combined to create new models of GPCR signaling and function. These types of studies will allow investigators in the future to apply genetic techniques to the study of GPCR signal transduction that were previously possible only in more classical genetic model systems such as yeast or Drosophila.

	IV. ADAPTING PHYSIOLOGICAL TECHNIQUES TO THE MOUSE

Top Previous Next References

The proliferation of techniques that allow the in vivo manipulation of GPCR gene expression has by necessity been followed by a corresponding proliferation in the methods to study them. Because most of these techniques involve mice or small rodents, this has resulted in the miniaturization of monitoring devices and a concerted effort to better understand mouse physiology. Within the last five years, technical advances have made it possible to catheterize and chronically instrument mice (27, 96), perform echocardiographic techniques in adult (36, 54) and embryonic mouse hearts (46), and obtain magnetic resonance images (84, 127) for detailed analyses. The need for such tools, especially for the analysis of subtle phenotypes or preterm defects, is critical for a comprehensive determination of the changes that result from genetic alteration of GPCRs.

Mice have several experimental advantages over other animal models, such as a short generation time and the availability of a detailed genetic map. There are many issues related to the differences between mice and humans that are not yet well understood. Despite obvious differences, many basic aspects of development, physiology, and GPCR function are well conserved between mice and humans, and in many instances, the development of mouse models to aid in the study of human disorders shows promise for accurate recapitulation of the phenotype and serves as a convenient experimental platform. As mentioned in section IIA, the creation of mouse transgenic lines that overexpress constitutively active adrenergic _1b-receptors or point mutants of the ₂-AR can serve as models for cardiac hypertrophy and attenuated signaling efficacy of ₂-ARs, respectively (96, 144). A mouse knockout model of human familial hypocalciuric hypercalcemia and severe hyperparathyroidism has been created by disruption of the calcium-sensing receptor (Casr), a GPCR homolog, and the phenotype shows remarkable similarities to inherited human disorders which affect the Casr gene (53). Manipulation of other GPCR subtypes such as rhodopsin has also provided valuable tools for the study of human disease, and the promise of such models to further our understanding of disease processes and potential therapies is ever-expanding.

	V. SUMMARY

Top Previous Next References

The ability to manipulate the expression patterns of specific gene products in vivo has led to significant advances in our understanding GPCR function and mechanism of action. In many cases, these techniques have provided information that could not be achieved by standard pharmacological techniques. Additionally, these techniques have also been used to effectively model various human disease processes, and thus provide a convenient experimental platform. Recent advances in the techniques of in vivo gene manipulation promise to advance our knowledge of GPCR function and will undoubtedly play a role in the design of future therapeutic approaches.

  

	REFERENCES

Top Previous

1.   ACCILI, D., C. S. FISHBURN, J. DRAGO, H. STEINER, J. E. LACHOWICZ, B. H. PARK, E. B. GAUDA, E. J. LEE, M. H. COOL, D. R. SIBLEY, C. R. GERFEN, H. WESTPHAL, AND S. FUCHS. A targeted mutation of the D₃ dopamine receptor gene is associated with hyperactivity in mice. Proc. Natl. Acad. Sci. USA 93: 1945-1949, 1996[Abstract/Free Full Text].

2.   ADAMS, J. U., X. CHEN, J. K. DERIEL, M. W. ADLER, AND C. L. LIU. Intracerebroventricular treatment with an antisense oligodeoxynucleotide to kappa-opioid receptors inhibited kappa-agonist-induced analgesia in rats. Brain Res. 667: 129-132, 1994[Medline].

3.   AIBA, A., C. CHEN, K. HERRUP, C. ROSENMUND, C. F. STEVENS, AND S. TONEGAWA. Reduced hippocampal long-term potentiation and context-specific deficit in associative learning in mGluR1 mutant mice. Cell 79: 365-375, 1994[Medline].

4.   ALBERT, P. R., AND S. J. MORRIS. Antisense knockouts: molecular scalpels for the dissection of signal transduction. Trends Pharmacol. Sci. 15: 250-254, 1994[Medline].

5.   ALFIE, M. E., X. P. YANG, F. HESS, AND O. A. CARRETERO. Salt-sensitive hypertension in bradykinin B₂ receptor knockout mice. Biochem. Biophys. Res. Commun. 224: 625-630, 1996[Medline].

6.   AMBUHL, P., R. GYURKO, AND M. I. PHILLIPS. A decrease in angiotensin receptor binding in rat brain nuclei by antisense oligonucleotides to the angiotensin AT₁ receptor. Regul. Pept. 59: 171-182, 1995[Medline].

7.   ASKEW, G. R., T. DOETSCHMAN, AND J. B. LINGREL. Site-directed point mutations in embryonic stem cells: a gene-targeting tag-and-exchange strategy. Mol. Cell. Biol. 13: 4115-4124, 1993[Abstract/Free Full Text].

8.   BAIK, J. H., R. PICETTI, A. SAIARDI, G. THIRIET, A. DIERICH, A. DEPAULIS, M. M. LE, AND E. BORRELLI. Parkinsonian-like locomotor impairment in mice lacking dopamine D₂ receptors. Nature 377: 424-428, 1995[Medline].

9.   BAYNASH, A. G., K. HOSODA, A. GIAID, J. A. RICHARDSON, N. EMOTO, R. E. HAMMER, AND M. YANAGISAWA. Interaction of endothelin-3 with endothelin-B receptor is essential for development of epidermal melanocytes and enteric neurons. Cell 79: 1277-1285, 1994[Medline].

10.   BERGLUND, P., I. TUBULEKAS, AND P. LILJESTROM. Alphaviruses as vectors for gene delivery. Trends Biotechnol. 14: 130-134, 1996[Medline].

11.   BERTIN, B., P. MANSIER, I. MAKEH, P. BRIAND, W. ROSTENE, B. SWYNGHEDAUW, AND A. D. STROSBERG. Specific atrial overexpression of G protein coupled human beta 1 adrenoceptors in transgenic mice. Cardiovasc. Res. 27: 1606-1612, 1993[Abstract/Free Full Text].

12.   BILSKY, E. J., R. N. BERNSTEIN, V. J. HRUBY, R. B. ROTHMAN, J. LAI, AND F. PORRECA. Characterization of antinociception to opioid receptor selective agonists after antisense oligodeoxynucleotide-mediated "knock-down" of opioid receptor in vivo. J. Pharmacol. Exp. Ther. 277: 491-501, 1996[Abstract/Free Full Text].

13.   BOND, R. A., P. LEFF, T. D. JOHNSON, C. A. MILANO, H. A. ROCKMAN, T. R. MCMINN, S. APPARSUNDARAM, M. F. HYEK, T. P. KENAKIN, L. F. ALLEN, AND R. J. LEFKOWITZ. Physiological effects of inverse agonists in transgenic mice with myocardial overexpression of the beta 2-adrenoceptor (see comments). Nature 374: 272-276, 1995[Medline].

14.   BORKOWSKI, J. A., R. W. RANSOM, G. R. SEABROOK, M. TRUMBAUER, H. CHEN, R. G. HILL, C. D. STRADER, AND J. F. HESS. Targeted disruption of a B₂ bradykinin receptor gene in mice eliminates bradykinin action in smooth muscle and neurons. J. Biol. Chem. 270: 13706-13710, 1995[Abstract/Free Full Text].

15.   BOURSON, A., E. BORRONI, R. H. AUSTIN, F. J. MONSMA, AND A. J. SLEIGHT. Determination of the role of the 5-HT₆ receptor in the rat brain: a study using antisense oligonucleotides. J. Pharmacol. Exp. Ther. 274: 173-180, 1995[Abstract/Free Full Text].

16.   BOZIC, C. R., B. LU, U. E. HOPKEN, C. GERARD, AND N. P. GERARD.