| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Departments of Psychiatry (R.J.S., S.W.C.) and Division of Endocrinology (D.D’A.), University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0559
Address all correspondence and requests for reprints to: Randy J. Seeley, Department of Psychiatry, Box 670559, University of Cincinnati, Cincinnati, Ohio 45267-0559. E-mail: seeleyrj{at}email.uc.edu
 |
Introduction |
|---|
 Top Introduction References |
|---|
The discovery of glucagon-like peptide 1 (GLP-1), a gastrointestinal hormone and central neurotransmitter, provides a tangible example of the recent progress in endocrinology/neuroendocrinology. The existence of GLP-1 was inferred from the cloning of a complementary DNA for preproglucagon, making it one of the first peptides discovered on the basis of a nucleic acid sequence (2, 3). Subsequent studies in humans and animals revealed that GLP-1 is secreted from the gut after meals and augments nutrient-stimulated insulin secretion (4). Thorens cloned a GLP-1 receptor from rat pancreatic islets that was subsequently shown to be expressed in the CNS as well as other tissues (5, 6). Eng and colleagues (7) first demonstrated that several peptides derived from the venom of the gila monster serve as GLP-1 receptor agonists and antagonists, providing powerful tools for further study of the role of GLP-1. Most recently the research group led by Drucker (8) at the University of Toronto produced a line of transgenic mice with a targeted deletion of the GLP-1 receptor, a "knock-out" model that this group has employed in a series of interesting and timely studies on the physiologic role of GLP-1.
The initial characterization of the GLP-1 receptor knockout mice focused on glucose homeostasis, the best-established function of GLP-1. Exogenous administration of GLP-1 or GLP-1 receptor agonists is insulinotropic in humans, rodents, and several in vitro models, suggesting at least a pharmacologic effect of GLP-1 on the pancreatic ß-cell. Continuous infusion of the GLP-1 receptor antagonist exendin-(9–39), or immunoneutralization of circulating GLP-1 with specific antibodies, impairs insulin secretion in experimental animals during enteral glucose administration (8, 9, 10). These data indicate that endogenous GLP-1, released during nutrient absorption, has a physiologic role to promote insulin secretion. Consistent with these findings, the GLP-1 receptor knockout mouse has significantly higher circulating blood glucose levels than wild-type controls following intragastric glucose loading (11). The glucose intolerance is most notable in male knockout mice and is likely caused by diminished early insulin secretion in response to glucose absorbed from the gut. The agreement between studies using acute pharmacologic methods with those using receptor knockouts provides compelling evidence for a physiologic role of GLP-1 as a hormonal mediator of postprandial insulin secretion.
Much less certain is the function of GLP-1 as a neurotransmitter although several putative roles have been proposed for this peptide in the CNS. Of great interest is the possibility that GLP-1 contributes to the regulation of food intake. Several agonists of the GLP-1 receptor show prominent effects to decrease food intake when administered directly into the CNS (12, 13, 14), while blocking the activation of central GLP-1 receptors with a peptide antagonist increases food intake (12, 15). Based on these studies using acute pharmacologic approaches, a solid case can be made that GLP-1 has some role in modulating food intake. However, in several careful studies, Drucker and colleagues demonstrated that the GLP-1 receptor knockout mice eat similar amounts and follow nearly the identical growth trajectory as wild-type control mice (16). So, unlike the situation with the regulation of postprandial insulin secretion by GLP-1, the pharmacological data and the knockout data conflict with regard to the role of GLP-1 in the control of food intake. Such disagreement between data derived by inducing acute changes in signaling pathways and those based on permanent gene knockout models has now been recognized in several instances (see Refs. 17, 21). Reconciling these differences is at first glance difficult, but likely to be enlightening, as we discuss below.
In this issue of Endocrinology, the Toronto group reports the results of another set of experiments in their GLP-1 receptor knockout mouse, this time testing the role of GLP-1 to mediate signaling through the hypothalamic-pituitary axis. Again, data from pharmacologic studies suggest that GLP-1 may act as a modulator of this system. Administration of GLP-1 to the third cerebral ventricle of rats activated CRH-containing neurons in the paraventricular nucleus (PVN), caused plasma corticosterone concentrations to rise 4-fold in one study (18), and stimulated lutenizing hormone release in another (19). These in vivo results are supported by studies using neuroendocrine cell lines that demonstrate that GLP-1 stimulates secretion from LH and TSH synthesizing cells (20). The findings of MacLusky et al., using the GLP-1 receptor knockout mice, do not agree with these previous reports. For the most part, the knockout animals are not distinguishable from their wild-type controls in terms of overall reproductive function, PVN CRH mRNA abundance, and basal levels of testosterone, estradiol, progesterone, thyroxine, and corticosterone.
However, despite the relatively normal concentrations of basal corticosterone, there was a suggestion of abnormalities in the hypothalamic-pituitary-adrenal axis in the knockout animals. Compared with wild-type mice, the knockouts had decreased adrenal weights, and although it is not clear whether this was due to less cortical mass, this finding could be consistent with the data in rats suggesting a role for GLP-1 in the activation of the HPA axis. However, MacLusky et al. also observed that the GLP-1 receptor null mice were more prone to elevations of corticosterone when stressed by added handling before the mice were killed, and performed in two psychometric tests as if they had augmented central CRH activity. These latter findings of augmented HPA activity in mice that presumably have no signaling through the GLP-1 receptor contradict the pharmacologic data from rat studies that GLP-1 has a role in activating this system. Thus, analogous to the situation with food intake, experiments with receptor knockout mice provide data that conflict with those derived from rats studied with receptor agonists and antagonists.
Taken alone, one would interpret the data from the GLP-1 receptor null mice as indicating that GLP-1 signaling in the brain is not a critical component of either the food intake or hypothalamic-pituitary regulatory systems. However, being satisfied with that conclusion requires ignoring the data from pharmacologic studies in rats. To their credit, MacLusky et al. do not overstate their results and attempt to make definitive conclusions on this topic. Nonetheless, using their mouse model they have presented rigorous tests of the role of GLP-1 in food intake, and now a solid initial study of the effects of this peptide on hypothalamic-pituitary function. Thus, we are left with the question of how to interpret the conflicting data from agonists and the knockouts? Data from exogenous administration of the agonist supports a role for GLP-1 in the central control of these neuroendocrine systems, whereas the preponderance of the data from the knockout does not.
At one extreme of interpretation would be to challenge the data from studies using agonists as less relevant because of the inherent limitations of this type of research. Pharmacologic experiments always suffer from the potential that the agonist interacts with receptors other than the one of interest and the work on GLP-1 and neuroendocrine control is no exception. Further, CNS administration of GLP-1 suffers from the possibility that the dose chosen produces supraphysiological levels of peptide, and consequently the observed effects do not reflect its normal physiological function. The knockout model addresses both of these problems because targeting of the GLP-1 receptor is quite precise and definitely eliminates the signaling through this system. However, to dismiss the pharmacological data altogether would leave us in the unappealing position of having a well-defined population of GLP-1 containing CNS neurons with no described function. Furthermore, it is far more difficult to dismiss as "dirty," or nonspecific, data obtained using receptor antagonists in conjunction with agonists, as has been done for GLP-1 signaling in the brain.
The other extreme in dealing with the problem of discrepancies between pharmacologic studies and the GLP-1 receptor knockouts is to dismiss the latter as less relevant to normal physiology. The first argument typically made in questioning the validity of a null phenotype in a knockout mouse model is that of "compensation." This argument comes in several flavors, but the most common version is that mice that have by necessity developed without GLP-1 receptors, have increased activity of other systems to ameliorate this deficiency and consequently hide an important role for GLP-1 in neuroendocrine function. These alternative systems might be ones that normally share regulatory functions with the GLP-1 receptor system, or processes that emerge only in the face of the deficiency. While several examples of compensation exist in the knockout literature (21), they have been instances of intracellular rather than network compensation. The argument for compensation for absent CNS signaling through the GLP-1 receptor is difficult to make given the lack of clear precedent for neuroendocrine pathways to substitute for one another. Hence, in the current example of the GLP-1 receptor knockouts, it would seem unwarrranted and unjustified simply to ignore the data from the knockouts because of a potential unknown compensation.
This leaves us with an increasingly common dilemma on just how to interpret null phenotypes in knockout models in instances where the gene deletion is expected to cause abnormalities based on pharmacologic experiments. Given that, for CNS signaling by GLP-1, neither set of data can be logically discounted, there remain several possibilities. First, there may be important species differences in the role of GLP-1 in controlling neuroendocrine function in mice compared with rats. While this explanation is scientifically unappealing without good evidence to support it, it is a question that could be readily addressed by repeating pharmacological experiments with GLP-1 in mice to confirm the findings from rats. Second is that the receptor targeted for disruption in these experiments is not the only one that interacts with GLP-1, and that another system exists in the CNS to mediate the effects of GLP-1 agonists. While at present there are no data directly supporting the presence of an alternate GLP-1 receptor, there is certainly a precedent for multiple receptors existing to mediate the effects of regulatory peptides. Third, it is possible that a neuroendocrine phenotype does exist in the knockout mice that is not apparent under the conditions that have been studied. This possibility is especially relevant to neuroendocrine systems, where complex multidetermined feedback controls are typical. It is possible that alternative experimental paradigms applied to the GLP-1 receptor knockouts—measuring the TSH response to blockade of thyroid hormone release, or determining corticosterone after different methods of external stress—could reconcile the data between this model and rats studied pharmacologically. Finally, to determine the role of endogenous GLP-1 in the control of neuroendocrine systems, several independent methods for producing relatively acute reductions in GLP-1 signaling such as pharmacological antagonists, antisense oligonucleotides or specific antibodies could be used. If such experiments reveal a role for endogenous GLP-1 in these neuroendocrine systems, it would provide rationale to hunt for compensatory mechanisms in these knockout mice. Discovering such a mechanism would change the way we think about the regulation of complex processes and interpret the results of gene deletion experiments. Alternatively, if augmented acute experiments do not reveal a role for endogenous GLP-1 in the regulation of food intake or hypothalamic-pituitary function, then alternative actions for GLP-1 in the CNS will need to be sought.
The important point here is that within endocrine and neuroendocrine research it is likely that situations such as the one presented here where pharmacological data disagree with genetic knockout data will be confronted repeatedly. Unfortunately, at present there is no general solution or formula that guides the resolution of such conflicts. Each research problem is unique and brings with it a separate set of interpretive issues that will require empiric resolution. The devil always has been, and will continue to be, in the details and can only be answered by further investigation. Drucker and his colleagues at the University of Toronto have made important contributions by taking one well-conceived genetic model and seeking to describe it phenotypically on a number of different fronts. However, in the end, transgenic technology alone will not always provide easy direct answers in these complex neuroendocrine systems. Nonetheless, the impetus to reconcile differences between knockout and pharmacologic experiments is likely to provide new information and potentially open up new lines of inquiry. Not surprisingly, scientific progress will come not from any single technique no matter how powerful. Rather, progress will come the way it has always come, from creative investigators doing incisive and careful experiments with methods that they find appropriate for their scientific questions. Like many other powerful research tools, knockout mice raise even more questions than they answer. Isn’t that just the way it ought to be?
Received November 30, 1999.
|
References |
|---|
|
TopIntroductionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. E. Mayo, L. J. Miller, D. Bataille, S. Dalle, B. Goke, B. Thorens, and D. J. Drucker International Union of Pharmacology. XXXV. The Glucagon Receptor Family Pharmacol. Rev., March 1, 2003; 55(1): 167 - 194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Baggio, F. Adatia, T. Bock, P. L. Brubaker, and D. J. Drucker Sustained Expression of Exendin-4 Does Not Perturb Glucose Homeostasis, beta -Cell Mass, or Food Intake in Metallothionein-Preproexendin Transgenic Mice J. Biol. Chem., October 27, 2000; 275(44): 34471 - 34477. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Lovshin, J. Estall, B. Yusta, T. J. Brown, and D. J. Drucker Glucagon-like Peptide (GLP)-2 Action in the Murine Central Nervous System Is Enhanced by Elimination of GLP-1 Receptor Signaling J. Biol. Chem., June 8, 2001; 276(24): 21489 - 21499. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
