Mechanisms of Hormone Action
David G. Gardner MD
Robert A. Nissenson PhD
Hormones produce their biologic effects through interaction with high-affinity receptors which are, in turn, linked to one or more effector systems within the cell. These effectors involve many different components of the cell's metabolic machinery, ranging from ion transport at the cell surface to stimulation of the nuclear transcriptional apparatus. Steroids and thyroid hormones exert their effects in the cell nucleus, although regulatory activity in the extranuclear compartment has also been documented. Peptide hormones and neurotransmitters, on the other hand, trigger a plethora of signaling activity in the cytoplasmic and membrane compartments while at the same time exerting parallel effects on the transcriptional apparatus. The discussion that follows will focus on the primary signaling systems employed by selected hormonal agonists and attempt to identify examples where aberrant signaling results in human disease.
The biologic activity of individual hormones is dependent upon their interactions with specific high-affinity receptors on the surfaces or in the cytoplasm of target cells. The receptors, in turn, are linked to signaling effector systems responsible for generating the observed biologic response. Receptors therefore convey not only specificity of the response (ie, cells lacking receptors lack responsiveness to the hormone) but also the means for activating the effector mechanism. In general, receptors for the peptide hormones and neurotransmitters are aligned on the cell surface while those for the steroid hormones, thyroid hormone, and vitamin D are found in the cytoplasmic or nuclear compartments.
Interactions between the hormone ligand and its receptor are governed by the laws of mass action:
where [H] is the hormone concentration, [R] is the receptor concentration, [HR] is the concentration of the hormone-receptor complex, and k+1 and k-1 are the rate constants for [HR] formation and dissociation respectively. Thus, at equilibrium,
where KD is the equilibrium dissociation constant which defines the affinity of the hormone-receptor interaction (ie, the lower the dissociation constant, the higher the affinity). Assuming that total receptor concentration Ro = [HR] + [R], this equation can be rearranged to give
This is the Scatchard equation and states that when bound over free ligand (ie, [HR]/[H]) is plotted against
bound ligand (ie, [HR]), the slope of the line is defined by -1/KD, the y-intercept by Ro/KD and the x-intercept by Ro (Figure 3-1). When [HR] = Ro/2, [H] = KD; therefore, the KD is also the concentration of hormone [H] at which one-half of the available receptors are occupied. Thus, knowledge of bound and free ligand concentrations, which can be determined experimentally, provides information regarding the affinity of the receptor for its ligand and the total concentration of receptor in the preparation.
Agents that bind to receptors with high affinity are classified as either agonists or antagonists based on the functional outcome of this receptor-ligand interaction. Agonists are ligands that trigger the effector mechanisms and produce biologic effects. Antagonists bind to the receptor but do not activate the effector mechanisms. Since they occupy receptor and block association with the agonist, they antagonize the functional activity of the latter. Partial agonists bind to the receptor but possess limited ability to activate the effector mechanisms. In different circumstances, partial agonists demonstrate variable biologic activity. For example, when employed alone, they may display weak activating activity, whereas their use together with a full agonist may lead to inhibition of function since the latter is displaced from the receptor molecule by a ligand with lower intrinsic activity.
Figure 3-1. Ligand saturation (A) and Scatchard analysis (B) of a hypothetical hormone receptor interaction. KD represents the dissociation constant; Ro the total receptor concentration; [HR] and [H] the bound and free ligand, respectively. Note in A that the KD is the concentration [H] at which half of available receptors are occupied.
In some systems, receptors are available in a surplus which is severalfold higher than that required to elicit a maximal biologic response. Such spare receptor systems, though they superficially appear redundant, are designed to rectify a mismatch between low circulating ligand levels and a relatively low affinity ligand-receptor interaction. Thus, by increasing the number of available receptors, the system is guaranteed a sufficient number of liganded receptor units to activate downstream effector systems fully, despite operating at subsaturating levels of ligand.
NEUROTRANSMITTER & PEPTIDE HORMONE RECEPTORS
As mentioned above, neurotransmitter and peptide hormones interact predominantly with receptors scattered on the plasma membrane at the cell surface. These receptors fall into four major groups (Table 3-1). The first includes the so-called serpentine or “seven-transmembrane-domain” receptors. These receptors each contain an amino terminal extracellular domain followed by seven hydrophobic amino acid segments, each of which is believed to span the membrane bilayer (Figure 3-2). The seventh of these, in turn, is followed by a hydrophilic carboxyl terminal domain that resides within the cytoplasmic compartment. As a group, they share a dependence on the G protein transducers (see below) to execute many of their biologic effects. A second group includes the single-transmembrane domain receptors that harbor intrinsic tyrosine kinase activity. This includes the insulin, IGF, and EGF receptors. A third group, which is functionally similar to the second group, is characterized by a large extracellular binding domain followed by a single membrane-spanning segment and a cytoplasmic tail. These receptors do not possess intrinsic tyrosine kinase activity but appear to function through interaction with soluble transducer molecules which do possess such activity.
Prolactin and growth hormone are included in this group. A fourth group, which includes the natriuretic peptide receptors, operates through activation of a particulate guanylyl cyclase and synthesis of cGMP. The cyclase is covalently attached at the carboxyl terminal of the ligand-binding domain and thus represents an intrinsic part of the receptor molecule.
Table 3-1. Major subdivisions (with examples) of the neurotransmitter-peptide hormone receptor families.1
G PROTEIN-COUPLED RECEPTORS
G protein-coupled receptors constitute a large superfamily of molecules capable of responding to ligands of remarkable structural diversity—ranging from photons to large polypeptide hormones. These receptors share overall structural features, most notably seven membrane-spanning regions connected by intracellular and extracellular loops (Figure 3-2). The receptors are oriented such that the amino terminal domain is extracellular, whereas the carboxyl terminal tail is cytoplasmic. The membrane-spanning segments interact with one another, forming an irregular cylindric bundle around a central cavity within the molecule. G protein-coupled receptors can assume at least two conformations with differing orientations of the membrane-spanning segments relative to one another. One orientation is favored in the absence of an agonist ligand, and in this
orientation the receptor does not activate a G protein (inactive conformation). The second orientation is stabilized by the binding of an appropriate agonist ligand, and in this conformation the receptor activates a cognate G protein (active conformation). All G protein-coupled receptors are thought to undergo a similar conformational switch upon agonist binding, producing a structural change in the cytoplasmic domain that promotes G protein activation. Some small agonists like catecholamines are able to enter the cavity formed by the transmembrane segments, thereby directly stabilizing the active receptor conformation. Other agonists, such as large polypeptide hormones, bind primarily to the extracellular domain of their G protein-coupled receptors. This indirectly results in movement of the transmembrane region of the receptor and stabilization of the active receptor conformation.
Figure 3-2. Structural schematics of different classes of membrane-associated hormone receptors. Representative ligands are presented in parentheses. (EGF, epidermal growth factor; GH, growth hormone; ANP, atrial natriuretic peptide.)
Heritable mutations in a variety of G protein-coupled receptors are known to be associated with clinical disease. Loss-of-function phenotypes result from mutations that eliminate one or both receptor alleles or that result in the synthesis of signaling-defective receptors. Gain-of-function phenotypes generally result from point mutations that produce receptors which are constitutively active (ie, stably assume the active receptor conformation even in the absence of an agonist ligand). Examples of such G protein-coupled receptor disorders relevant to endocrinology are described below and discussed in greater detail elsewhere in this book.
G PROTEIN TRANSDUCERS
G protein-coupled receptors initiate intracellular signaling by activating one (or in some cases multiple) G proteins. G proteins are a family of heterotrimeric proteins that regulate the activity of effector molecules (eg, enzymes, ion channels) (Table 3-2), resulting ultimately in biologic responses. The identity of a G protein is defined by the nature of its α subunit, which is largely responsible for effector activation. The major G proteins involved in hormone action (and their actions on effectors) are Gs (stimulation of adenylyl cyclase), Gi (inhibition of adenylyl cyclase; regulation of calcium and potassium channels), and Gq/11 (stimulation of phospholipase Cβ). The β and γ subunits of G proteins are tightly associated with one another and function as a dimer. In some cases, the βγ subunit dimer also regulates effector function.
G proteins are noncovalently tethered to the plasma membrane and are thus proximate to their cognate receptors and to their effector targets. The basis for specificity in receptor-G protein interactions has not been fully defined. It is likely that specific structural determinants presented by the cytoplasmic loops of the G protein-coupled receptor determine the identity of the G proteins that is activated. It is the nature of the α subunit of the G protein that is critical for receptor recognition. There are about a dozen different G protein α subunits, and hundreds of distinct G protein-coupled receptors. Thus, it is clear that a particular G protein is activated by a large number of different receptors. For example, Gs is activated by receptors for ligands as diverse as β-adrenergic catecholamines and large polypeptide hormones such as LH. LH is thereby able to stimulate adenylyl cyclase and raise intracellular levels of cAMP in cells that express LH receptors (eg, Leydig cells of the testis).
Table 3-2. G protein subunits selectively interact with specific receptor and effector mechanisms.
Figure 3-3 is a schematic representation of the molecular events associated with activation of G proteins by G protein-coupled receptors. In the basal, inactive state, the G protein is an intact heterotrimer with guanosine diphosphate (GDP) bound to the α subunit. Agonist binding to a G protein-coupled receptor promotes the physical interaction between the receptor and its cognate G protein. This produces a conformational change in the G protein, resulting in the dissociation of GDP. This in turn allows the binding of GTP (which is present at much higher concentration in cells than is GDP) to the α subunit. Dissociation of the βγ-bound and GTP bound α subunits then occurs, allowing these subunits to activate their effector targets. Dissociation of the hormone-receptor complex also occurs. The duration of activation is determined by the intrinsic GTPase activity of the G protein α subunit. Hydrolysis of GTP to GDP terminates the activity and promotes reassociation of the βγ subunits, returning the system to the basal state.
Figure 3-3. G protein-mediated signal transduction. α and β/γ subunits of a representative G protein are depicted. (See text for details.) (R, hormone receptor; H, hormonal ligand; E, effector.)
G PROTEIN DISORDERS
Two bacterial toxins are capable of covalently modifying specific G protein α subunits, thereby altering their functional activity. Cholera toxin is a protein that binds to receptors present on all cells, resulting in the internalization of the enzymatic subunit of the toxin. The toxin enzyme is an ADP-ribosyl transferase that transfers ADP-ribose from NAD to an acceptor site (Arg201) on the α subunit of Gs. This covalent modification greatly inhibits the GTPase activity of αs, enhancing the activation of adenylyl cyclase by extending the duration of the active GTP-bound form of the G protein. Even in the absence of an active G protein-coupled receptor, GDP dissociates (albeit very slowly) from the G protein. Thus, cholera toxin will eventually activate adenylyl cyclase activity even without agonist binding to a G protein-coupled receptor. The result is a large and sustained activation of adenylyl cyclase. When this occurs in intestinal epithelial cells, the massive increase in cAMP results in the increased water and salt secretion characteristic of cholera.
Pertussis toxin is also an ADP-ribosyl transferase. However, in this case, the substrates are the α subunits of different G proteins, most notably Gi and Go. The ADP-ribose moiety is transferred to a cysteine residue near the carboxyl terminal of the α subunit, a region required for interaction with activated G protein-coupled receptors. Once ADP-ribosylated by pertussis toxin, these G proteins are no longer able to interact with activated receptors and are thus stuck in an inactive (GDP-bound) conformation. Inhibition of receptor-mediated activation of Gi and Go accounts for many of the clinical manifestations of pertussis infection.
Genetic mutations in G protein α subunits are seen in a number of human diseases. Acquired, activating mutations in αs can produce a variety of phenotypes depending on the site of expression of the mutant protein. In McCune-Albright syndrome, the mutation occurs in a subset of neural crest cells during embryogenesis. All of the descendants of these cells, including certain osteoblasts, melanocytes, and ovarian or testicular cells, express the mutant protein. The result is a form of genetic mosaicism in which the consequence of unregulated production of cAMP in particular tissues is evident (ie, the progressive bone disorder polyostotic fibrous dysplasia, abnormal skin pigmentation referred to as café au lait spots, gonadotropin-independent precocious puberty). In cells where cAMP is linked to cell proliferation (eg, thyrotropes, somatotropes), a subset of patients with benign tumors have been shown to have acquired activating mutations in αs. Activating mutations in one of the Gi proteins that is coupled to cell proliferation, αi2, have been reported in a subset of adrenal and ovarian tumors.
Loss-of-function mutations in αs are associated with the hereditary disorder pseudohypoparathyroidism type Ia (PHP-Ia). This disorder, first described by Fuller Albright,
is the first documented example of a human disease attributable to target cell resistance to a hormone. Affected patients display biochemical features of hypoparathyroidism (eg, hypocalcemia, hyperphosphatemia) but have markedly increased circulating levels of parathyroid hormone (PTH) and display target cell resistance to PTH. Many hormone receptors couple to adenylyl cyclase via Gs, yet patients with PHP-Ia generally display only subtle defects in responsiveness to other hormones (eg, TSH, LH). The explanation for this lies in the fascinating genetics of this disorder. In brief, affected patients have one normal and one mutated αs allele. The mutated allele fails to produce an active form of the protein. Tissues in these patients are expected to express about 50% of the normal level of αs, a level sufficient to support signaling to adenylyl cyclase. However, in certain tissues, the αs gene is subject to genetic imprinting such that the paternal allele is expressed poorly or not at all. For individuals harboring inactivating mutations, if it is the paternal allele that is mutated, all cells will express about 50% of the normal level of αs (derived from the normal maternal allele). However, if the mutation is on the maternal allele, then the cells where paternal imprinting occurs will express low levels or no αs. One of the major sites of this paternal imprinting is in the proximal renal tubule, an important target tissue for the physiologic actions of PTH. This accounts for the clinical resistance to PTH seen in PHP-Ia and accounts also for the fact that only a subset of patients with haploinsufficiency of αs are resistant to PTH. Interestingly, essentially all patients with haploinsufficiency of αs display Albright's hereditary osteodystrophy, a developmental disorder with phenotypic manifestations affecting a variety of tissues. This indicates that even a partial loss of adenylyl cyclase signaling is incompatible with normal development.
Numerous effectors have been linked to the G protein-coupled receptors. A number of these are presented in Table 3-2. There are a great many other G proteins—not dealt with here—that are coupled to physical or biochemical stimuli but have very limited involvement in hormone action. As discussed above, adenylyl cyclase, perhaps the best-studied of the group, is activated by Gs (Figure 3-4). This activation results in a transient increase in intracellular cAMP levels. cAMP binds to the inhibitory regulatory subunit of inactive protein kinase A (PKA) and promotes its dissociation from the complex, thereby permitting enhanced activity of the catalytic subunit. The latter phosphorylates a variety of cellular substrates, among them the hepatic phosphorylase kinase that initiates the enzymatic cascade which results in enhanced glycogenolysis and the nuclear transcription factor CREB (cAMP response element binding protein), which mediates many of the known transcriptional responses to cAMP (and to some extent calcium) in the nuclear compartment. Other transcription factors are also known to be phosphorylated by PKA.
Phospholipase C beta (PLCβ) is a second effector system that has been studied extensively. The enzyme is activated through Gq-mediated transduction of signals generated by a wide array of hormone-receptor complexes, including those for angiotensin II, α-adrenergic agonists, and endothelin. Activation of the enzyme leads to cleavage of phosphoinositol 4,5-bisphosphate in the plasma membrane to generate inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (Figure 3-5). The former interacts with a specific receptor present on the endoplasmic reticulum membrane to promote release of Ca2+ into the cytoplasmic compartment. The increased calcium, in turn, may activate protein kinases, promote secretion, or foster contractile activity. Depletion of intracellular calcium pools by IP3 results in enhanced uptake of calcium across the plasma membrane (perhaps through generation of IP4 [1,3,4,5-tetrakisphosphate]), thereby activating a second, albeit indirect, signaling mechanism that serves to increase intracellular calcium levels even further. Diacylglycerol (DAG) functions as an activator of protein kinase C (PKC) within the cell. Several different isoenzymatic forms of PKC (eg, α, β, γ) exist in a given cell type. A number of these are calcium-dependent, a property which, given the IP3 activity mentioned above, provides the opportunity for a synergistic interaction of the two signaling pathways driven by PLCβ activity. However, not all protein kinase C activity derives from the breakdown of PIP2 substrate. Metabolism of phosphatidylcholine by PLCPC leads to the generation of phosphocholine and DAG. This latter pathway is believed to be responsible for the more protracted elevations in PKC activity seen following exposure to agonist. Other phospholipases may also be important in hormone-dependent signaling. Phospholipase D employs phosphatidylcholine as a substrate to generate choline and phosphatidic acid. The latter may serve as a precursor for subsequent DAG formation. As with PLCPC above, no IP3 is generated as a consequence of this reaction. Phospholipase A2 triggers release of arachidonic acid, a precursor of prostaglandins, leukotrienes, endoperoxides, and thromboxanes, all signaling molecules in their own right. The relative contribution of these other phospholipases to hormone-mediated signal transduction and the role of the idiosyncratic products (phosphocholine, phosphatidic acid, etc) in conveying regulatory information remains an area of active research.
Activation of effectors by G protein-coupled receptors is subject to regulatory mechanisms that prevent overstimulation of cells by an agonist ligand. At the
level of the receptor, two regulatory events are known to occur. One is desensitization, wherein initial stimulation of a receptor by its agonists leads to a loss of the ability of the receptor to subsequently elicit G protein activation. This is shown schematically in Figure 3-6 for the β-adrenergic receptor, and a similar regulatory mechanism exists for many G protein-coupled receptors. Agonist binding to the receptor produces G protein activation and results also in activation of a kinase that phosphorylates the cytoplasmic domain of the receptor. By virtue of this phosphorylation, the receptor acquires high affinity for a member of the arrestin family of proteins. The name “arrestin” derives from the observation that the receptor is no longer capable of interacting with a G protein when arrestin is bound. Thus, the phosphorylated receptor becomes uncoupled from its G protein, preventing signaling to the effector. The receptor remains inactive until a phosphatase acts to restore the receptor to its unphosphorylated state. Many G protein-coupled receptors are also susceptible to agonist-induced down-regulation, resulting in a reduced level of cell surface receptors following exposure of cells to an agonist. This can result from agonist-induced internalization of receptors, followed by trafficking of receptors to lysosomes where degradation occurs. In addition, chronic exposure of cells to an agonist may result in signaling events that suppress the biosynthesis of new receptors, thereby lowering steady state receptor levels. Together, these regulatory events ensure that the cell is protected from excessive stimulation in the presence of sustained high levels of an agonist.
Figure 3-4. β-Adrenergic receptor signaling in the cytoplasmic and nuclear compartments. The cAMP response element binding protein (CREB) is depicted bound to a consensus CRE in the basal state. Phosphorylation of this protein leads to activation of the juxtaposed core transcriptional machinery.
G PROTEIN-COUPLED RECEPTORS & HUMAN DISEASE
Mutations in the genes encoding G protein-coupled receptors are being increasingly recognized as important in the pathogenesis of endocrine disorders. Loss-of-function mutations generally need to be homozygous (or compound heterozygous) in order to result in a significant
disease phenotype. This is probably due to the fact that most cells have a complement of receptors which exceeds what is needed for maximal cellular response (“spare receptors”). Thus, a 50% reduction in the amount of a cell surface receptor may have little influence on the ability of a target cell to respond. However, in some situations, haploinsufficiency of a G protein-coupled receptor can produce a clinical phenotype. For instance, heterozygous loss-of-function mutations in the G protein-coupled calcium-sensing receptor results in the dominant disorder familial hypocalciuric hypercalcemia due to mild dysregulation of PTH secretion and renal calcium handling. Homozygous loss of function of the calcium-sensing receptor results in severe neonatal hyperparathyroidism due to the loss of the ability of plasma calcium to suppress PTH secretion. Syndromes of hormone resistance have also been reported in patients lacking expression of functional G protein-coupled receptors for vasopressin, ACTH, and TSH. Loss of functional expression of the PTH receptor results in Blomstrand chondrodysplasia, a disorder that is lethal due to the inability of PTH-related protein (a PTH receptor agonist) to promote normal cartilage development.
Figure 3-5. PLCβ-coupled receptor signaling in the cytoplasmic and nuclear compartments. (PLC, phospholipase; PC, phosphatidylcholine; DAG, diacylglycerol; PKC, protein kinase C.)
Mutations that render G protein-coupled receptors constitutively-active (in the absence of an agonist ligand) are seen in a number of endocrine disorders. Generally
speaking, such mutations produce a disease phenotype resembling that seen with excessive levels of the corresponding hormone agonist. Thus, activating mutations in the TSH produce neonatal thyrotoxicosis, and activating mutations in the LH receptor result in pseudoprecocious puberty or testotoxicosis. Activating mutations in the PTH receptor result in Jansen's metaphysial chondrodysplasia, a disorder characterized by hypercalcemia and increased bone resorption (mimicking the effects of excess PTH on bone) and delayed cartilage differentiation (mimicking the effects of excess PTH-related protein on cartilage). Molecular analysis of G protein-coupled receptors has revealed that point mutations, in addition to producing constitutive activity, can alter the specificity of ligand binding or the ability of the receptor to become desensitized. It is almost certain that such mutations will be found to provide the basis for some perhaps more subtle endocrinopathies.
Figure 3-6. Kinase-dependent desensitization of the ligand-receptor complex. Schema shown is that for the β-adrenergic receptor, but similar systems probably exist for other types of G protein-linked receptors. (βARK identifies the β-adrenergic receptor kinase; PKA, protein kinase A; ACa, active adenylyl cyclase; ACi, inactive adenylyl cyclase.)
GROWTH FACTOR RECEPTORS
The growth factor receptors differ from those described above both structurally and functionally. Unlike the G protein-associated receptors, these proteins span the membrane only once and acquire their signaling ability, at least in part, through activation of tyrosine kinase activity, which is intrinsic to the individual receptor molecules. The insulin and IGF receptors fall within this group as do those for the autocrine or paracrine regulators platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and epidermal growth factor (EGF). Signaling is initiated by the association of ligand (eg, insulin) with the receptor's extracellular domain (Figure 3-7) and subsequent receptor dimerization. This results in phosphorylation of tyrosines both on the receptor itself as well as on nonreceptor substrates. It is assumed that phosphorylation of these substrates results in a cascade of activation events, similar to those described for the G protein-coupled systems, which contribute to perturbations in the intracellular phenotype. The autophosphorylation of the receptor molecules themselves has been studied extensively and provided some intriguing insights into the mechanisms that underlie signal transduction by this group of proteins.
Tyrosine phosphorylation takes place at specific locations in the receptor molecule. Once phosphorylated, these sites associate, in highly specific fashion, with a
variety of accessory proteins that possess independent signaling capability. These include phospholipase Cγ (PLCγ), phosphoinositol (PI) 3′ kinase, GTPase-activating protein (GAP), and growth factor receptor-bound protein-2 (GRB2). These interactions are fostered by the presence of highly conserved type 2 src homology (based on sequence homology to the src proto-oncogene) domains (SH2) in each of the accessory molecules. Each individual SH2 domain
displays specificity for the contextual amino acids surrounding the phosphotyrosine residues in the receptor molecule. In the PDGF receptor, for example, the SH2 domain of PLCγ associates selectively with Tyr977 and Tyr989 while that of PI 3′ kinase associates with Tyr708 and Tyr719. Thus, diversity of response is controlled by contextual sequences around individual phosphotyrosine residues that determine the types of accessory proteins which will be brought into the signaling complex. These protein-protein interactions may provide a means of directly activating the signaling molecule in question, perhaps through a change in steric conformation. Alternatively, they may facilitate the sequestration of these accessory proteins in or near the plasma membrane compartment, in close proximity to key substrates (eg, membrane lipids in the case of PLCγ) or other important regulatory proteins.
Figure 3-7. Signaling by tyrosine kinase-containing growth factor receptor. Receptors depicted here as monomers for simplicity; typically dimerization of receptors follows association with ligand. Autophosphorylation of one or more critically positioned tyrosine residues in the receptor leads to association with accessory proteins or effectors through SH2 domains present on the latter. In some cases an SH3 domain present on the same protein leads to recruitment of yet other proteins leading to further complex assembly.
While some of these associations trigger immediate signaling events, others (eg, GRB2) may act largely to provide the scaffolding needed to construct a more complex signaling apparatus (Figure 3-8). In the case of GRB2, another accessory protein (son-of-sevenless; SOS) associates with the receptor-GRB2 complex through a type 3 src homology (SH3) domain present in the latter. This domain recognizes a sequence of proline-rich amino acids present in the SOS protein. SOS, in turn, facilitates assembly of the Ras-Raf complex which permits activation of downstream effectors like mitogen-activated protein kinase (MAPK) kinase (MEK). This latter kinase, which possesses both serine-threonine and tyrosine kinase activity, activates the p42 and p44 MAPKs (also called extracellular signal-regulated kinases; ERKs). ERK acts upon a variety of substrates within the cell, including the RSK kinases, which, in turn, phosphorylate the ribosomal S6 protein and thereby stimulates protein synthesis. These phosphorylation reactions (and their amplification in those instances where the MAPK substrate is a kinase itself) often lead to protean changes in the phenotype of the target cells.
The liganded growth factor receptors, including the insulin receptor, may also signal through the phosphoinositide 3-OH kinase (PI-3K). SH2 domains of the p85 regulatory subunit of PI-3K associate with the growth factor receptor through specific phosphotyrosine residues (Tyr740and Tyr751 in the PDGF receptor) in a manner similar to that described above for GRB2 (Figure 3-9). This leads to activation of the p110 catalytic subunit of PI-3K and increased production of phos-phatidylinositol-3,
4,5-trisphosphate (PIP3) and phos-phatidylinositol-3,4-bisphosphate (PI[3,4]P2). These latter molecules sequester protein kinase B (also known as Akt) at the cell membrane through association with the plekstrin homology domains in the amino terminal of the kinase molecule. This in turn leads to phosphorylation of PKB at two separate sites (Thr308 in the active kinase domain and Ser473 in the carboxyl terminal tail) by PIP3-dependent kinases (PDK1 and PDK2). These phosphorylations result in activation of PKB. Downstream targets of activated PKB (eg, following insulin stimulation) include 6-phosphofructo-2-kinase (increased activity), glycogen synthase kinase-3 (decreased activity), the insulin-responsive glucose transporter GLUT 4 (translocation and increased activity) and p70 S6 kinase (increased activity). This leads to increased glycolysis, increased glycogen synthesis, increased glucose transport, and increased protein synthesis, respectively. There is also a growing body of evidence suggesting that PKB may protect cells from programmed cell death through phosphorylation of key proteins in the apoptotic pathway.
Figure 3-8. Growth factor-dependent pathway. Assembly of the components involved in the ras/raf/MEK/MAPK and PI-3 K/PKB signaling mechanisms.
It has been reported that G protein-coupled receptors may also activate the Raf-MEK-ERK cascade, though in this case the signal traffics through a non-receptor protein tyrosine kinase (NRPTK such as Src and Fyn) rather than the traditional growth factor receptor-linked tyrosine kinases. The details of the mechanism are incompletely understood, but it appears to require the participation of β-arrestin (see above) as an adaptor molecule linking the G protein receptor to the NRPTK. Interestingly, this implies that β-arrestin, which terminates coupling between the receptor and G protein, actually promotes coupling between the desensitized receptor and downstream effectors traditionally associated with growth factor-dependent activation.
These include the receptors for a variety of cytokines, erythropoietin, colony-stimulating factor, growth hormone, and prolactin. These receptors have a single internal hydrophobic stretch of amino acids, suggesting that they span the membrane but once (Figure 3-9). Interestingly, alternative splicing of the GH receptor gene primary transcript results in a foreshortened “receptor”that lacks the membrane anchor and carboxyl terminal domain of the protein. This “receptor” is secreted and serves to bind GH in the extracellular space (eg, circulating plasma). Unlike the growth factor receptors described above, GH receptors lack a tyrosine kinase domain. Their mechanism of action is not perfectly understood but appears to involve the participation of signaling intermediates, like JAK2, a protein that possesses intrinsic tyrosine kinase activity. The association of JAK2 with the liganded GH receptor presumably provokes a conformational change in JAK2 and activation of its tyrosine kinase catalytic activity. This, in turn, triggers downstream signaling events, including activation of transcription factors (eg, STAT), and stimulation of MAPK and the S6 kinase (RSK) activity.
Figure 3-9. Signaling by the growth hormone receptor. Different portions of a single growth hormone molecule associate with homologous regions of two independent growth hormone receptor (GHR) molecules. This is believed to lead to the recruitment of the accessory protein JAK2 and activation of downstream effectors.
GUANYLYL CYCLASE-LINKED RECEPTORS
Activation of guanylyl cyclase-dependent signaling cascades can occur through two independent mechanisms. The first involves activation of the soluble guanylyl cyclase, a heme-containing enzyme that is activated by the gas nitric oxide (NO) generated in the same or neighboring cells. NO is produced by the enzyme nitric oxide synthase. NO synthase exists as three different
isozymes in selected body tissues. Constitutive forms of NO synthase are produced in endothelial (NOS-3) and neuronal (NOS-1) cells. The endothelial enzyme possesses binding sites for FAD and FMN as well as calcium and appears to require calcium for optimal activity. Agents like bradykinin and acetylcholine, which interact with receptors on the surface of endothelial cells and increase intracellular calcium levels, trigger an increase in constitutive NO synthase activity with consequent generation of NO and activation of soluble guanylyl cyclase activity in neighboring vascular smooth muscle cells (Figure 3-10). Thus, in this instance, the cGMP-dependent vasodilatory activity of acetylcholine requires sequential waves of signaling activity in two different cell types to realize the ultimate physiologic effect.
The inducible (i) form of NO synthase (NOS-2) is found predominantly in inflammatory cells of the immune system although it has also been reported to be present in smooth muscle cells of the vascular wall. Unlike the endothelial form of NO synthase, expression of iNO synthase is low in the basal state. Treatment of cells with a variety of cytokines triggers an increase in new iNO synthase synthesis (hence, the inducible component of iNO synthase activity), probably through activation of specific cis elements in the iNO synthase promoter. Thus, hormones, cytokines, or growth factors with the capacity for induction of iNO synthase activity may direct at least a portion of their signaling activity through a cGMP-dependent pathway.
A third mechanism for increasing cGMP levels within target cells involves the activation of particulate guanylyl cyclases. From an endocrine standpoint, this involves predominantly the natriuretic peptide receptors (NPR). NPR-A is a single-transmembrane-domain receptor (about 130 kDa) with a large extracellular domain that provides ligand recognition and binding. This is followed by a hydrophobic transmembrane domain and a large intracellular domain which harbors the signaling function. The amino terminal portion of this intracellular region contains a kinase-like ATP-binding domain that is involved in regulating cyclase activity while the carboxyl terminal domain contains the catalytic core of the particulate guanylyl cyclase. It is believed that association of ligand with the extracellular domain leads to a conformational change in the receptor that arrests the tonic inhibitory control of the kinase-like domain and permits activation of guanylyl cyclase activity. NPR-B, the product of a separate gene, has a similar topology and a relatively high level of sequence homology to the NPR-A gene product; however, while NPR-A responds predominantly to the cardiac atrial natriuretic peptide (ANP), NPR-B is activated by the C-type NP (CNP), a peptide found in the central nervous system, endothelium, and reproductive tissues but not in the heart. Thus, segregated expression of the ligand and its cognate receptor convey a high level of response specificity to these two systems despite the fact that they share a common final effector mechanism.
NUCLEAR ACTION OF PEPTIDE HORMONES
Although the initial targets of peptide hormone receptor signaling appear to be confined to the cytoplasm, it is clear that these receptors can also have profound effects on nuclear transcriptional activity. They accomplish this through the same mechanisms they employ to regulate enzymatic activity in the cytoplasmic compartment (eg, through activation of kinases and phosphatases). In this case, however, the ultimate targets are transcription factors that govern the expression of target genes. Examples include hormonal activation of c-jun and c-fos, nuclear transcription factors which make up the heterodimeric AP-1 complex. This complex has been shown to alter the expression of a wide variety of eukaryotic genes through association with a specific recognition element, termed the phorbol ester (TPA)-response element (TRE), present within the DNA sequence of their respective promoters. Other growth factor receptors that employ the MAPK-dependent signaling mechanism appear to target the serum response factor (SRF) and its associated ternary complex proteins. Posttranslational modification of these transcription factors is believed to amplify the signal that traffics from this complex, when associated with the cognate serum response element (SRE), to the core transcriptional apparatus. cAMP-dependent activation of protein kinase A results in the phosphorylation of a nuclear protein CREB (cAMP response element binding protein) at Ser119, an event which results in enhanced transcriptional activity of closely positioned promoters. The latter requires the participation of an intermediate CREB-binding protein (CBP). CBP is a coactivator molecule that functionally tethers CREB to proteins of the core transcriptional machinery. Interestingly, CBP may also play a similar role in nuclear receptor signaling (see below). Growth hormone is known to induce the phosphorylation of an 84 kDa and a 97 kDa protein in target cells. These proteins have been shown to associate with the sis-inducible element (SIE) in the c-fos promoter and to play a role in signaling cytokine activity which traffics through this element. It remains to be demonstrated that these proteins play a similar functional role in mediating GH-dependent effects. Several recent studies have provided evidence suggesting that a number of peptide hormones and growth
factors may bind to high-affinity receptors in the cell nucleus. The role these receptors play—if any—in contributing to the signaling profile of these peptides remains undefined.
Figure 3-10. Signaling through the endothelial (e) and inducible (i) nitric oxide synthases (NOS) in the vascular wall. Activation of eNOS in the endothelial cell or iNOS in the vascular smooth muscle cell leads to an increase in NO and stimulation of soluble guanylyl cyclase (GC) activity. Subsequent elevations in cGMP activate cGMP-dependent protein kinase (PKG) and promote vasorelaxation.
The nuclear receptors, which include those for the glucocorticoids, mineralocorticoids, androgens, progesterone, estrogens, thyroid hormone, and vitamin D, differ from the receptors of the surface membrane described above in that they are soluble receptors with a proclivity for employing transcriptional regulation as a means of promoting their biologic effects. Thus, though some receptors are compartmentalized in the cytoplasm (eg, glucocorticoid receptor) while others are confined to the nucleus (eg, thyroid hormone receptor), they all operate within the nuclear chromatin to initiate the signaling cascade. These receptors can be grouped into two major subtypes based on shared structural and functional properties. The first, the steroid receptor family, includes the prototypical glucocorticoid receptor (GR) and the receptors for mineralocorticoids (MR), androgens (AR), and progesterone (PR). The second, the thyroid receptor family, includes the thyroid hormone receptor (TR), estrogen (ER), retinoic acid (RAR and RXR), peroxisome proliferator-activated receptor (PPAR), and vitamin D (VDR) receptors. In addition, there are more than 100 so-called orphan receptors that bear structural homology to members of the extended nuclear receptor family. For most of these the “ligand” is unknown, and their functional roles in the regulation of gene expression have yet to be determined.
STEROID RECEPTOR FAMILY
Steroid receptors (ie, GR, MR, AR, and PR), under basal conditions, exist as cytoplasmic, multimeric complexes that include the heat shock proteins hsp 90, hsp 70, and hsp 56. The estrogen receptor (ER), though demonstrating similar association with heat shock proteins, is largely confined to the nuclear compartment. Association of the steroid ligand with the receptor results in dissociation of the heat shock proteins. This in turn exposes a nuclear translocation signal previously buried in the receptor structure and initiates transport of the receptor to the nucleus, where it associates with the hormone response element (Figure 3-11).
Figure 3-11. Signaling through the steroid receptor complex. Similar mechanisms are employed by members of the TR gene family, though most of the latter are concentrated in the nuclear compartment and are not associated with the heat shock protein complex prior to binding ligand. (meG, methyl guanosine.)
Each of the family members has been cloned and sequenced, and crystallographic structures have been obtained for many of them. Consequently, we know a great deal about their structure and function (Figure 3-12. Each has an extended amino terminal domain of varying length and limited sequence homology to other
family members. In at least some receptors, this region, which has been termed AF-1, is believed to participate in the transactivation function through which the individual receptors promote increased gene transcription. Significant variability in the length of the amino terminal regions of the different receptors suggests potential differences in their respective mechanisms for transcriptional regulation. The amino terminal is followed by a basic region that has a high degree of homology to similarly positioned regions in both the steroid and thyroid receptor gene families. This basic region encodes two zinc finger motifs (Figure 3-13) which have been shown to establish contacts in the major groove of the cognate DNA recognition element (see below). Based on crystallographic data collected for the DNA binding region of the GR, we know that the amino acid sequence lying between the first and second fingers (ie, recognition helix) is responsible for establishing specific contacts with the DNA. The second finger provides the stabilizing contacts that increase the affinity of the receptor for DNA. The DNA binding region also harbors amino acid residues that contribute to the dimerization of monomers contiguously arrayed on the DNA recognition element. Following the basic region is the carboxyl terminal domain of the protein. This domain is responsible for binding of the relevant ligand, receptor dimerization or heterodimerization, and association with the heat shock proteins. It also contributes to the ligand-dependent transactivation function (incorporated in a subdomain termed AF-2) that drives transcriptional activity. Interestingly, in selected cases, nonligands have been shown to be capable of activating steroid receptors. Dopamine activates the progesterone receptor and increases PR-dependent transcriptional activity, probably through a phosphorylation event, which elicits a conformational change similar to that produced by the association of the receptor with progesterone.
The DNA-binding regions of these receptors contact DNA through a canonical hormone recognition element (HRE) which is described in Table 3-3. Interestingly, each receptor in the individual subfamily binds to the same recognition element with high affinity. Thus, specificity of hormone action must be established either by contextual DNA sequence lying outside the recognition element or by other, nonreceptor DNA-protein interactions positioned in close proximity to the element. Interestingly, the GR, as well as some other nuclear receptors (eg, ER), are capable of binding to DNA sequence lacking the classic HRE. Originally described in the mouse proliferin gene promoter, these composite elements associate with heterologous complexes containing GR, as well as components of the AP-1 transcription factor complex (ie, c-junand c-fos), and display unique regulatory activity at the level of contiguously positioned promoters. One such composite element, for example, directs very specific transcriptional effects depending on whether the GR or the MR is included in the complex.
Figure 3-12. Structural schematic of a representative steroid receptor molecule. Separate designations are given to the amino terminal (NH2), DNA-binding (DBD), and ligand-binding (LBD) domains. Functional activity associated with each of these individual domains, as determined by mutagenesis studies, are indicated below.
Several steroids, particularly the glucocorticoids and estrogens, have been reported to have independent effects on the stability of target gene transcripts. At this point it is unclear what role the hormone receptors play in this process and whether transcript stabilization is tied mechanistically to the enhancement of transcriptional activity.
THYROID RECEPTOR FAMILY
Included in this group are the TR, RAR, RXR, ER, PPAR, and VDR. They share a high degree of homology to the proto-oncogene c-erbA and high affinity for a common DNA recognition site (Table 3-3). With the exception of the ER, they do not associate with the heat shock proteins, and they are constitutively bound to chromatin in the cell nucleus. Specificity of binding for each of the individual receptors is, once again, probably conferred by a contextual sequence surrounding this element, the orientation of the elements (eg, direct repeats or inverted repeats or palindromes), the polarity (ie, 5′ in contrast to 3′ position on two successive repeats), and the number and nature of the spacing nucleotides separating the repeats.
The estrogen receptor binds to its RE as a homo-dimer, while the VDR, RAR, RXR, and TR prefer binding as heterodimers. The nature of the heterodimeric partners has provided some intriguing insights into the
biology of these receptors. The most prevalent TR-associated partners appear to be the retinoid X receptors. These latter receptors, which as homodimers form high-affinity associations with 9-cis-retinoic acid, also form heterodimeric complexes in the unliganded state with VDR and RAR. In individual cases where it has been examined, heterodimerization with RXR amplifies both the DNA binding and the functional activity of these other receptors. Thus, the ability to form such heterodimeric complexes may add significantly to the flexibility and potency of these hormone receptor systems in regulating gene expression. Interestingly, the positioning (5′ versus 3′) of the participant proteins on the RE is important in determining the functional outcome of the association. In most of those situations linked to transcriptional activation, RXR seems to prefer the upstream (5′) position in the dimeric complex. Thus, diversity of response is engendered by the selection of recognition elements (eg, monomeric versus dimeric versus oligomeric sites) and by the choice and positioning of the dimeric partner (eg, homodimer versus heterodimer) where applicable.
The crystallographic structures of the ligand-binding domains (LBDs) of several members of the thyroid receptor family have been described. These include the dimeric unliganded RXRα, monomeric liganded RARγ, monomeric liganded TRα, dimeric agonist (ie, estradiol)- and antagonist (ie, raloxifene)-liganded ERα, liganded VDR, and liganded PPARγ. Each LBD displays a common folding pattern with 12 alpha helices (numbered by convention H1-H12) and a conserved β turn. Some variability exists in that there is no H2 in RARγ and a short H2′ helix is present in PPARγ, but the overall structural configuration is preserved. The dimeric interface is formed through interaction of amino acids located in helices 7-10, with the strongest influence exerted by H10. These interactions appear to be important for both homo- as well as heterodimeric interactions. Binding of ligand has been shown to occur through what has been termed a “mousetrap” mechanism. In the unliganded state, H12, which contains the carboxyl terminal activation domain AF-2, is displaced away from the ligand-binding pocket (seeFigure 3-14). Association of agonist ligand (eg, estradiol in the case of the ER) with the hydrophobic core of the receptor leads to a repositioning of H12 over the ligand-binding cavity, where it stabilizes receptor-ligand interactions and closes the “mousetrap.” Binding of an antagonist ligand
such as raloxifene, which because of its structure engenders steric hindrance in the ligand-binding pocket, prevents closure of H12 into the normal agonist position. Instead, H12 folds into an alternative location between H4 and H3, a conformation that suppresses the activation function of the receptor (see below).
Figure 3-13. Schema of the two zinc fingers, together with coordinated zinc ion, which make up the DNA-binding domain of the glucocorticoid receptor (amino acids are numbered relative to the full length receptor). Shaded regions denote two alpha helical structures which are oriented perpendicularly to one another in the receptor molecule. The first of these, the recognition helix, makes contact with bases in the major groove of the DNA. Large arrows identify amino acids which contact specific bases in the glucocorticoid response element (GRE). Lighter arrows identify amino acids that confer specificity for the GRE; selective substitutions at these positions can shift receptor specificity to other response elements. Dots identify amino acids making specific contacts with the phosphate backbone of DNA. (Modified from Luisi BF et al. Reprinted, with permission, from Nature 1991;352:498. Copyright Š 1991 by Macmillan Magazines Ltd.)
Table 3-3. DNA recognition elements for major classes of nuclear hormone receptors.1
The mechanistic underpinnings of transcriptional regulation by the nuclear receptors have been partially elucidated (Figure 3-15). In the unliganded state, the receptor dimers are associated with a macromolecular complex containing the repressor proteins N-CoR or SMRT, a transcriptional corepressor Sin3, and a histone deacetylase RPD3. N-CoR and SMRT each use two independent IDs (receptor interaction domains) to associate with the nuclear receptors (one repressor: two receptors). Histone acetylation is typically associated with activation of gene transcription (presumably reflecting decompaction of chromatin surrounding the transcriptional unit), so the presence of histone deacetylase activity in the complex promotes a transcriptionally quiescent state. Addition of ligand leads to a change in receptor conformation that no longer favors interaction with the repressor and promotes both ATP-dependent chromatin remodeling and assembly of an activator complex containing p160 coactivator proteins
(eg, SRC-1, GRIP-1, or P/CIP) and, secondarily, the CREB-binding protein (CBP) and the histone acetylase P/CAF. The net accrual of histone acetylase activity (CBP and P/CIP as well as PCAF possess acetylase activity) leads to acetylation of chromatin proteins (eg, histones) as well as components of the core transcriptional machinery, resulting in chromatin decompaction and a net increase in transcriptional activity. Interaction of the nuclear receptors with the coactivators in this complex takes place through LXXLL motifs (where L = leucine and X = any amino acid) present in the coactivator proteins. Each coactivator may have several of these motifs, which preferentially associate with different nuclear receptors, other transcription factors, or other coactivators. This allows for a degree of selectivity in terms of which regulatory proteins are incorporated
in the complex. Notably, a recent structural analysis showed that a 13-amino-acid peptide, containing an LXXLL motif from the GRIP-1 protein, interacts with TRβ through a hydrophobic cleft generated by helices 3, 4, and 12 (including AF-2) in the receptor protein. This is the same cleft that is occupied by helix 12, which harbors an LXXLL motif, in the raloxifene-bound ERα. This suggests that the antagonist in the latter instance acquires its activation-blocking properties by repositioning helix 12 in a manner that leads to displacement of the coactivator protein from this groove (see above). SRC also interacts with the AF-1 domain, suggesting a potential mechanism for maximizing synergistic activity between AF-1 and AF-2 domains in the receptors.
Figure 3-14. Three-dimensional structures for the agonist- and antagonist-occupied ERα LBD. Panel A: Orthogonal views of the agonist diethylstilbestrol-ERα LBD-NR Box II peptide complex. Coactivator peptide and LBD are presented as ribbon diagrams. Peptide is colored medium blue, helix (H) 12 (ERα residues 538-546) is colored dark blue. Helices 3, 4, and 5 are colored light blue. Diethylstilbestrol is depicted in a space-filling format.
Figure 3-14. Panel B: Orthogonal views of the antagonist 4-hydroxytamoxifen-ERα LBD complex. Color scheme is the same as in panel A. 4-Hydroxytamoxifen is shown in dark gray in a space-filling format. Note that NR Box II is absent in this structure.
Figure 3-14. Panel C: Schematic representation of the mechanism underlying agonist-dependent activation of nuclear hormone receptor. In the presence of agonist, helix 12 (the terminal helix in the LBD) folds across the ligand-binding pocket, stabilizing ligand-receptor interaction and promoting a conformation conducive for coactivator association. In the presence of antagonist, steric hindrance precludes folding of helix 12 across the ligand-binding pocket. Instead, it positions itself in the region typically occupied by the coactivator, thereby blocking the activation function of the receptor. (Reprinted, with permission, from Shiau A et al: The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 1998;95:927. Copyright Š 1998 by Cell Press.)
Figure 3-15. Interaction of corepressor (top) versus coactivator (bottom) molecules with the ligand binding domain of a representative nuclear receptor. (See text for details.) The temporal order of p 160 versus DRIP/TRAP binding remains undetermined.
CBP is thought to function as a pivotal component of the nuclear receptor regulatory complex. While the p160 class of coactivators interact directly with the nuclear receptors, CBP associates primarily with the p160 coactivators, thereby establishing an indirect link to the receptors. As noted above, CBP appears to function as a central integrator of transcriptional regulatory signals from multiple pathways, including the cAMP-dependent activation of the transcription factor CREB. Recent evidence suggests that an additional level of regulatory control may be involved in selectively amplifying nuclear receptor-dependent transcriptional activity. An enzyme called CARM1 (coactivator-associated arginine methyltransferase 1) associates with CBP and methylates the protein. This results in a reduction in CREB-dependent gene activation and, secondarily, an increase in nuclear receptor-dependent gene transcription. This switching mechanism effectively refocuses the transcriptional machinery on expression of nuclear receptor-dependent gene expression.
More recently, another family of coactivator complexes have been identified as playing an important role in nuclear receptor signaling. The human TR-associated protein (TRAP) and vitamin D receptor-interacting protein (DRIP) complexes are the best-characterized to date. These complexes, which contain in the neighborhood of 25 individual proteins, are thought to serve as a functional bridge between the liganded nuclear receptor bound to DNA and the general transcription factors (eg, TBP, TFIIB, RNA polymerase II, and TAFs) involved in formation of the preinitiation complex. The TRAP220 subunit appears to establish the relevant contacts with the nuclear receptors in promoting this assembly. Their role vis-ŕ-vis the p160 coactivators alluded to above remains undefined; however, it has been suggested that they succeed the p160 coactivator complex in binding with liganded nuclear receptors positioned on target gene promoters, establish the requisite structural and functional connections with the core transcriptional machinery, and initiate mRNA synthesis. It has also been suggested that acetylation of one of the key nuclear receptor (NR) binding motifs (LXXLL) on the SRC coactivator by CBP leads to dissociation of SRC from the nuclear receptors, thereby allowing access for assembly of the TRAP/DRIP complex (Figure 3-15).
While the glucocorticoid receptor is encoded by a single gene, there are two genes for the TR (α and β). TR α1 and TR β1 appear to be the dominant forms of TR in the body. Although considerable overlap exists in their tissue distribution, TR α1 is enriched in skeletal muscle, brown fat, and the central nervous system, while TR β1 is found in the liver, kidney, and central nervous system. They are believed to signal most of the developmental and thermogenic effects of thyroid hormone in the whole animal. TR β2, a splice variant of the TR β gene, is found in the rodent pituitary gland, where it may subserve a specific regulatory function (eg, control of TSH secretion). TR α2, an alternatively spliced product of the TR α gene, lacks the hormone binding domain at the carboxyl terminal of the molecule and thus is not a true thyroid hormone receptor. While under certain experimental conditions TR α2 can block the activity of other members of the TR family, its physiologic role, if one exists, remains undefined. Similar heterogeneity exists in the retinoid receptor family. There are three isoforms for both the RXR and RAR. Collectively, these receptors are thought to play an important role in morphogenesis, but the function of the individual isoforms remains only partially understood.
NONGENOMIC EFFECTS OF THE STEROID HORMONES
While steroids exert most of their biologic activity through direct genomic effects, there are several lines of evidence suggesting that this does not provide a complete picture of steroid hormone action. There are several examples which, for kinetic or experimental reasons, do not fit the classic paradigm associated with a transcriptional regulatory mechanism. Included within this group are the early water imbibition of the uterus associated with estrogen administration, the rapid suppression of ACTH secretion following steroid administration, the modulation of oocyte maturation by progesterone, and regulation of calcium channel function by 1,25-(OH)2 vitamin D. Recent studies have demonstrated the presence of conventional estrogen receptors on the plasma membrane of target cells. The relationship of these receptors to their nuclear counterparts and their role in signaling estrogen-dependent activity (genomic versus nongenomic) is being actively investigated. While we still do not completely understand the mechanisms underlying these nongenomic effects, their
potential importance in mediating steroid or thyroid hormone action may, in selected instances, approach that of their more conventional genomic counterparts.
Neurosteroids represent another class of nontraditional hormonal agonists with unique biologic activity. Some of these are native steroids (eg, progesterone), while others are conjugated derivatives or metabolites of the native steroids (eg, dihydroprogesterone). These agonists have been identified in the central nervous system and in some instances shown to have potent biologic activity. It is believed that they operate through interaction with the receptor for γ-aminobutyric acid, a molecule that increases neuronal membrane conductance to chloride ion. This has the net effect of hyperpolarizing the cellular membrane and suppressing neuronal excitability. Interactions that promote receptor activity would be predicted to produce sedative-hypnotic effects in the whole animal, while inhibitory interactions would be expected to lead to a state of central nervous system excitation.
STEROID & THYROID HORMONE RECEPTOR RESISTANCE SYNDROMES
Heritable defects in these receptors have been linked to the pathophysiology of a number of hormone resistance syndromes. These syndromes are characterized by a clinical phenotype suggesting hormone deficiency, by elevated levels of the circulating hormone ligand, and increased (or inappropriately detectable) levels of the relevant trophic regulatory hormone (eg, ACTH, TSH, FSH, or LH). Point mutations in the zinc fingers of the DNA-binding domain as well the ligand-binding domain of the vitamin D receptor leads to a form of vitamin D-dependent rickets (type II) characterized by typical rachitic bone lesions, secondary hyperparathyroidism, and alopecia. It is inherited as an autosomal recessive disorder. Molecular defects scattered along the full length of the androgen receptor, though concentrated in the ligand-binding domain, have been linked to syndromes characterized by varying degrees of androgen resistance ranging from infertility to the full-blown testicular feminization syndrome. Clinical severity, in this case, is thought to be related to the severity of the functional impairment which the mutation imposes on the receptor. Since the androgen receptor is located on the X chromosome, these disorders are inherited in an X-linked fashion. Defects in the glucocorticoid receptor are less common, perhaps reflecting the life-threatening nature of derangements in this system. However, mutations have been identified that impact negatively on receptor function. Clinical presentations in these cases have been dominated by signs and symptoms referable to glucocorticoid deficiency (eg, fatigue, asthenia) and adrenal androgen (eg, hirsutism and sexual precocity) and mineralocorticoid (low renin hypertension) overproduction. This presumably results from defective steroid-mediated suppression of ACTH secretion and adrenal hyperplasia as the former rises in a futile attempt to restore glucocorticoid activity at the periphery. Resistance to thyroid hormone has been linked to a large number of mutations scattered along the full length of the β form of the receptor, although, once again, there is a concentration of mutations in the ligand-binding domain, particularly along the rim of the coactivator binding pocket. No mutations in the α form of the receptor have been linked to a hormone-resistant phenotype. The clinical presentation of thyroid hormone resistance extends from the more typical mild attention deficit syndromes to full-blown hypothyroidism with impaired growth. Different target tissues harboring the mutant receptors display variable sensitivity to thyroid hormone, with some tissues (eg, pituitary) displaying profound resistance and others (eg, heart) responding in a fashion suggesting hyperstimulation with thyroid hormone (ie, thyrotoxicosis). These syndromes are rather unique in that they are inherited as autosomal dominant disorders, presumably reflecting the ability of the mutated receptors to interfere with receptors produced from the normal allele, either by binding to the RE with higher affinity than the wild-type receptors and precluding access of the latter to target genes or by forming inactive heterodimers with the wild-type receptor proteins. Defects in the estrogen receptor are rare, perhaps reflecting the critical role estrogens play in regulating lipoprotein metabolism. However, one male patient has been described who harbors a mutation within the ligand-binding domain of the estrogen receptor. His clinical presentation was characterized by infertility as well as osteopenia, suggesting important roles for estrogens in the maintenance of spermatogenesis as well as bone growth even in male subjects. A syndrome of mineralocorticoid resistance, or pseudohypoaldosteronism, has been described in a number of independent kindreds. Pseudohypoaldosteronism type I is characterized by neonatal renal salt wasting, dehydration, hypotension, hyperkalemia, and hyperchloremic metabolic acidosis despite the presence of elevated aldosterone levels. Heterozygous mutations in the MR are responsible for a milder form of the disease which is inherited in an autosomal dominant pattern. A more severe form of the disease, inherited in an autosomal recessive pattern, appears to be due to loss-of-function mutations in genes encoding subunits of the amiloride-sensitive epithelial sodium channel. Of equivalent interest is the recent identification of an activating mineralocorticoid receptor mutation (Ser810-to-Leu810). This mutation results in severe early-onset hypertension that is markedly exacerbated by pregnancy. The mutation leads to constitutive activation of the
MR and alters the specificity of ligand binding such that traditional MR antagonists, like progesterone, function as partial agonists. This latter property presumably accounts for the dramatic increase in blood pressure during pregnancy.
G Protein-Coupled Receptors
Caron MG, Lefkowitz RJ: Catecholamine receptors: Structure, function and regulation. Recent Prog Horm Res 1993; 48:277.
Clark AJL, Weber A: Molecular insights into inherited ACTH resistance syndromes. Trends Endocrinol Metab 1994;5:209.
Farfel Z, Bourne HR, Iiri T: The expanding spectrum of G protein diseases. N Engl J Med 1999;340:1012.
Iiri T et al: Rapid GDP release from Gsα in patients with gain and loss of endocrine function. Nature 1994;371:164.
Spiegel AM (editor): G Proteins, Receptors and Disease. Humana Press, 1998.
Weinstein LS et al: Endocrine manifestations of stimulatory G protein alpha-subunit mutations and the role of genomic imprinting. Endocr Rev 2001;22:675.
Asaoka Y et al: Protein kinase C, calcium and phospholipid degradation. Trends Biochem Sci 1992;17:414.
Balla T, Catt KJ: Phosphoinositides and calcium signaling. Trends Endocrinol Metab 1994;5:250.
Liscovitch M: Crosstalk among multiple signal-activated phospholipases. Trends Biochem Sci 1992;17:393.
Meyer TE, Habener JF: Cyclic adenosine 3′,5′-monophosphate response element binding protein (CREB) and related transcription-activation deoxyribonucleic acid-binding proteins. Endocr Rev 1993;14:269.
Shikama N, Lyon J, LaThangue NB: The p300/CBP family: integrating signals with transcription factors and chromatin. Trends Cell Biol 1997;7:230.
Sterweis PC, Smrcka AV: Regulation of phospholipase C by G proteins. Trends Biochem Sci 1992;17:502.
Tyrosine Kinase-Coupled and Cytokine Receptors
Carter-Su C, Smit LS: Signaling via JAK tyrosine kinases: Growth hormone receptor as a model system. Recent Prog Horm Res 1998;53:61.
Downward J: Mechanisms and consequences of activation of protein kinase B/Akt. Curr Opin Cell Biol 1998;10:262.
Mussachchio A, Wilmanns M, Saraste M: Structure and function of the SH3 domain. Prog Biophys Molec Biol 1994;61:283.
Nishida E, Gotoh Y: The MAP kinase cascade is essential for diverse signal transduction pathways. Trends Biochem Sci 1993;18:128.
Pazin MJ, Williams LT: Triggering signaling cascades by receptor tyrosine kinases. Trends Biochem Sci 1992;17:374.
Pelech SL, Sanghera JS: Mitogen-activated protein kinases: versatile transducers for cell signaling. Trends Biochem Sci 1992; 17:233.
Roupas P, Herington AC: Postreceptor signaling mechanisms for growth hormone. Trends Endocrinol Metab 1994;5:154.
Guanylyl Cyclase-Linked Receptors
Drewett JG, Garbers DL: The family of guanylyl cyclase receptors and their ligands. Endocr Rev 1994;15:135.
Sessa WC: The nitric oxide synthase family of proteins. J Vasc Res 1994;31:131.
Freedman LP, Luisi BF: On the mechanism of DNA binding by nuclear hormone receptors: A structural and functional perspective. J Cell Biochem 1993;51:140.
Geller DS et al: Mutations in the mineralocorticoid receptor gene cause autosomal dominant pseudohypoaldosteronism type I. Nat Genet 1998;19:279.
Ito M, Roeder RG: The TRAP/SMCC/Mediator complex and thyroid hormone receptor function. Trends Endocrinol Metab 2001;12:127.
Lee KC, Kraus WL: Nuclear receptors, coactivators and chromatin: new approaches, new insights. Trends Endocrinol Metab 2001;12:191.
Moras D, Gronemeyer H: The nuclear receptor ligand binding domain: Structure and function. Curr Opin Cell Biol 1998; 10:384.
Torchia J, Glass C, Rosenfeld MG: Co-activators and co-repressors in the integration of transcriptional responses. Curr Opin Cell Biol 1998;10:373.
Wehling M: Nongenomic actions of steroid hormones. Trends Endocrinol Metab 1994;5:347.
Zhang J, Lazar MA: The mechanism of action of thyroid hormones. Annu Rev Physiol 2000;62:439.