| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, University of Texas Health Science Center at San Antonio, San Antonio Texas 78229-3900
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
ALDOSTERONE REGULATES GENE...
ALDOSTERONE-INDUCED PROTEINS
AN INTEGRATED MODEL
REFERENCES
|
|---|
The systemic actions of aldosterone are well documented; however, in comparison, our understanding of the cellular and molecular mechanisms by which aldosterone orchestrates these actions is rudimentary. Aldosterone exerts most of its physiological actions by modifying gene expression. It is now apparent that aldosterone represses almost as many genes as it induces. Several aldosterone-sensitive genes, including serum and glucocorticoid-inducible kinase (sgk) and small, monomeric Kirsten Ras GTP-binding protein (Ki-ras) have recently been identified. The molecular mechanisms and elements bestowing corticosteroid sensitivity on these and many other genes are becoming clear. Induction of Ki-Ras and Sgk is necessary and sufficient for some portion of aldosterone action in epithelia. These two signaling factors are components of a converging pathway with phosphatidylinositol 3-kinase positioned between them that enables both stabilizing the epithelial Na+ channel (ENaC) in the open state as well as increasing the number of ENaC in the apical membrane. This aldosterone-induced signaling pathway contains many potential sites for feedback regulation and cross talk from other cascades and potentially impinges directly on the activity of transport proteins and/or cellular differentiation to modify electrolyte transport.
mineralocorticoid; Sgk; Ki-Ras; corticosteroid hormone-induced factor; sodium-potassium-adenosinetriphosphatase; NEDD4; epithelial sodium channel; epithelial; hypertension; transport; sodium absorption; potassium secretion
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
ALDOSTERONE REGULATES GENE...
ALDOSTERONE-INDUCED PROTEINS
AN INTEGRATED MODEL
REFERENCES
|
|---|
THE ADRENAL CORTICOSTEROID HORMONE aldosterone plays a pivotal role in homeostasis. This is particularly apparent when one considers that dysfunctional regulation of aldosterone secretion and inappropriate activity of aldosterone effectors are involved in many human diseases associated with electrolyte and fluid imbalance.1 In addition, aldosterone has attracted much attention lately as a possible mediator of pathological heart remodeling (129). Tissues targeted by aldosterone include cardiac fibroblasts and myocytes, neurons and their support cells, smooth muscle, endothelial cells, and adipose tissue (see Refs. 58, 70, 176, and 177 for further reference). Electrically tight epithelial monolayers, such as the renal distal tubule and collecting duct system, distal colon, and those in salivary glands, are considered classic aldosterone target tissues. Aldosterone action in these tissues, as well as in toad bladder and frog skin, has been the focus of much important investigation over the last half-century. Thus the natriferic and kaliuretic effects of aldosterone have long been established. In this regard, aldosterone, acting as a mineralocorticoid, targets epithelial cells to increase Na+ (re)absorption and K+ secretion. It is striking that all known forms of Mendelian hypertension in humans result from aberrant regulation of aldosterone or its downstream effectors that implement aldosterone's signal in sodium homeostasis (reviewed in Refs. 68 and 97). The interested reader has many excellent contemporary review articles available for a discussion of the effects of aldosterone at the systemic, tissue, cellular, and molecular levels and for a history of important findings in this field (51, 52, 58, 70, 176, 177). The scope of the present review is to summarize and discuss the results of numerous recent studies investigating the cellular and molecular mechanisms of aldosterone action. The review focuses primarily on the intracellular signal transduction pathways initiated by aldosterone and the relationship of aldosterone-sensitive signaling factors to each other and their final effectors, including the luminal epithelial Na+ channel (ENaC), secretory K+ channel, and serosal Na+/K+-ATPase. Only advances in our understanding of the classic genomic mechanism of aldosterone action in epithelial cells are considered. Because several comprehensive reviews have been published recently in this journal (12, 51, 139, 177), background concepts underpinning regulation of ENaC and other aldosterone-sensitive effectors, as well as mineralocorticoid specificity and the actions and regulation of nuclear receptors and their accessory proteins, are explicitly not (re-)covered in great detail.
| |
ALDOSTERONE REGULATES GENE EXPRESSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
ALDOSTERONE REGULATES GENE...
ALDOSTERONE-INDUCED PROTEINS
AN INTEGRATED MODEL
REFERENCES
|
|---|
Aldosterone and other adrenal corticosteroids exert many of their physiological actions through modulation of gene expression. This results in a substantial lag period (~0.5-1 h) preceding overt changes in cellular activity. Similar to other steroids, aldosterone also affects cellular activity through faster (<1 min), nongenomic actions mediated presumably by distinct plasma membrane/cytosolic receptors. Nongenomic actions, while important, are for the most part beyond the focus of the present review. (The interested reader is referred to Refs. 51, 57, 70, 137, and 188 for further development of this topic.)
Molecular Mechanisms of Action
Inhibitors of transcription and translation, as well as other experimental maneuvers, have been employed to definitively demonstrate that induction of gene expression is necessary, in part, for aldosterone action in epithelia (reviewed in Refs. 51, 137, 176, and 177). However, the contrary view, that gene repression is needed, in part, for aldosterone action, has been more difficult to demonstrate. This has arisen primarily because genes negatively influenced by aldosterone have not been well described in the past. Recently, a number of aldosterone-repressed transcripts have been identified (135, 154). It is now apparent that aldosterone represses almost as many genes as it induces. It follows, then, that aldosterone-sensitive gene repression plays an important but yet undefined role in the final cellular response.Aldosterone, as well as other corticosteroid hormones, binds cytosolic
steroid receptors that translocate to the nucleus in a ligand-dependent
manner (reviewed in Refs. 51, 57, and
58). Once in the nucleus, corticosteroid receptors function as
transcription factors through interaction with genomic DNA at a
steroid-response element (SRE; consensus sequence AGAACAnnnTGTTCT).
Aldosterone is a ligand for two distinct but similar types of nuclear
receptors: the mineralocorticoid receptor (MR) and the glucocorticoid
receptor (
-isoform; GR). Glucocorticoids, as well, are ligands
for these receptors. It is well recognized that circulating levels of
glucocorticoids are in excess of those of mineralocorticoids by 100- to
1,000-fold. In contrast to GR, which is ubiquitous, MR expression is
confined primarily to epithelial cells. Thus circulating levels of
corticosteroids and receptor expression patterns often define the
cellular specificity of the glucocorticoid response mediated by GR. The
cellular mechanism(s) specifying a mineralocorticoid response are more
complex, considering that all cells expressing MR also express GR
(reviewed in Refs. 51, 57, and
58).
Corticosteroids have similar affinities for both MR and GR, although,
on- and off-rates, ligand-receptor half-lives, association with
accessory proteins, and trans-activating potential differ. The ligand-independent GR
acts as a dominant negative regulator of
both GR and MR (7). The physiological significance of
GR to aldosterone signaling is unclear at this time. Both MR and GR
form homo- and heterodimers with each other and other accessory factors. It has been speculated that unique receptor complexes may
target distinct cis-acting elements (99, 172).
Although it is commonly accepted that MR and GR are capable of
mediating trans-activation through a common SRE, the concept
that these receptors also target unique cis-acting elements
remains quite controversial.
A novel MR splice variant that lacks a ligand-binding domain (MR
5,6)
has recently been described (196). MR5,6 is highly expressed (relative to other tissues) in the kidney. MR5,6 forms dimers with both MR and GR, with dimers binding DNA at the SRE and
trans-activating in a ligand-independent manner. Thus
MR5,6 appears to have the opposite function of GR. It has been
hypothesized that this novel MR variant plays a role in defining
receptor specificity; however, the role of MR5,6 in a
mineralocorticoid-specific response remains somewhat ambiguous,
considering that it potentiates both MR and GR
trans-activating ability to a similar degree. Several MR
mRNA species, some of which encode unique proteins, have now been
identified in various mammalian tissues (23, 90, 93, 195,
201). It is likely that all of these transcripts arise from
differential splicing or alternative transcription start sites within a
common gene. A good understanding of the physiological significance of
each type of MR isoform remains elusive at this time. However, if the
transcripts encoding unique proteins result in MR isoforms with
different specificities for corticosteroids, accessory proteins, and
SRE with distinct trans-acting capabilities, then they could
play a role in defining specificity. Distinct promotors also have been
reported to drive MR expression in a tissue-specific manner
(93). These promotors enable MR levels in disparate
tissues to be varied independently of each other. In addition, nuclear
corticosteroid receptors are modified at the posttranslation level.
Indeed, cAMP signaling via protein kinase A leads to phosphorylation of
MR, with phosphorylated receptors having greater
trans-activating potential compared with unmodified receptors (108). Thus further investigation of the
function, posttranslational regulation, and transcriptional control of
MR isoforms is needed to gain more precise insight into the
physiological role played by each during a mineralocorticoid response.
Two distinct molecular mechanisms as depicted in Fig.
1 are widely accepted to define the
actions of nuclear receptors on gene expression: 1) the
classic mechanisms involving trans-activation and
trans-repression via interaction with cognate DNA-binding sites, such as SRE and 2) mechanisms of transcription
interference and synergy mediated by protein-protein interactions
between corticosteroid receptors and other trans-acting
factors. In contrast to direct trans-activation and
repression, the corticosteroid receptor does not need to physically
bind DNA in the latter cases, although this mechanism does impinge on
gene expression. The relevance of these mechanisms of action to
aldosterone signaling is considered below.
|
Trans-activation. Both MR and GR modulate gene expression through the canonical pentadecamer SRE, with the monomers of the dimeric receptor complex binding each "half-site" (hSRE). The finding that MR and GR have distinct effects on cognition and learning mediated by the hippocampus supports the possibility of unique MR and GR cis-acting elements or that unique protein-protein interactions involving these distinct receptors convey different signals (43, 44). MR and GR also have distinguishable actions in the colon (9, 27, 39, 181). Further support for unique cis-acting elements comes from the study of model systems. In one model system, MR has only 5% of the trans-acting activity of GR on the mouse mammary tumor virus promotor in CV-1 cells (4, 140). However, these results may just as easily reflect findings that trans-activation in response to MR and GR is ligand dependent and influenced differentially by numerous accessory factors (51, 137). The most direct support for MR and GR serving distinct functions comes from mice genetically engineered to be devoid of one receptor type or the other or to express a receptor mutant with a partial loss of function (16, 17, 40, 72, 78, 134; see Trans-repression). From these animals, it is clear that GR influences lung maturation, response to stress, metabolism, and inflammation as well as immune responses. MR, on the other hand, primarily mediates a mineralocorticoid response in epithelia to maintain electrolyte and fluid balance.
The cis-acting elements responsive to corticosteroid receptors for many aldosterone-induced genes have now been identified. A corticosteroid-sensitive gene encodes serum and glucocorticoid-inducible kinase (Sgk; 36, 45, 104, 186, 187). Presently, three isoforms of Sgk (Sgk1, Sgk2, and Sgk3) are recognized, of which only Sgk1 is corticosteroid sensitive (reviewed in Ref. 124). Thus, for simplicity, Sgk1 is referred to as Sgk. The promotor region of the sgk gene contains a classic but imperfect pentadecameric cis-acting SRE (AGGACAgaaTGTTCT; 104, 187). In mammary epithelia, this element is trans-activated in response to dexamethasone signaling (104, 187). Indirect evidence from numerous epithelia, including the renal distal nephron and distal colon, infers that this cis-element is also responsive to aldosterone via MR (18, 28, 36, 45, 101, 151). That glucocorticoids and aldosterone via GR and MR trans-activate sgk in distinct tissues through a common SRE suggests that both types of steroids may initiate a cellular response mediated by a common signaling pathway or one that contains many common signaling elements. Thus GR and MR likely have, at least partially, overlapping functions in epithelia expressing both receptors. This possibility, however, remains quite controversial. Functional redundancy is supported, nevertheless, by findings showing that aldosterone via MR and glucocorticoids via GR induce expression of many of the same genes and that the widely expressed GR is sometimes capable of complementing MR dysfunction or eliciting a mineralocorticoid response, especially in disease states (22, 37, 53, 58, 75, 97, 137). A corticosteroid-sensitive gene also encodes the-subunit of ENaC
(42, 114). A classic but slightly imperfect SRE
(AGAACAgaaTGTCCT) was recently identified as sufficient and
necessary for trans-activation of the human and rat -ENaC
promotors in renal, lung, salivary, and colonic epithelial cell lines
(98, 114, 142, 184, 197). This cis-acting
element is responsive to MR and GR, with no overt specificity to either
glucocorticoids or aldosterone. Taken together, these findings are
consistent with those showing -ENAC to be an aldosterone-induced
transcript in renal epithelia but a glucocorticoid-induced transcript
in lung epithelia (42, 114, 142).
The gene encoding the 1-subunit of the
Na+/K+-ATPase is also sensitive to
corticosteroids (24, 52, 88, 128). The nucleic acid
sequence AGTCACAGGAGGCACTCTGAGAGCA located in the 5'-flanking region of
the human Na+/K+-ATPase 1 gene
functions as a cis-acting element responsive to both MR and
GR (89). Although this sequence is distinct from the
common SRE thought to mediate most GR and MR signaling, there are some
similarities. It is believed that this newly identified SRE is
influenced by GR/MR through a classic mechanism.
The genes encoding many small, monomeric GTP-binding proteins in the
Ras family, such as Ha-Ras and Ki-Ras, are controlled by
corticosteroids via regulation of transcription (127, 149, 154,
165). A recently identified family of GTP-binding proteins (DexRas) that share much sequence identity and many functional domains
with Ras are also regulated at the level of transcription by
glucocorticoids (81, 173). It is unclear at this time
whether aldosterone affects DexRas levels and whether this protein
plays a prominent role in mineralocorticoid signaling; however, DexRas is expressed in the kidney, where it is positively regulated by glucocorticoids (81). The cis-acting element
modulating dexras expression has not yet been identified.
However, more is known about those modulating Ha-ras and
Ki-ras. Several cis-acting elements within the
human, mouse, and rat Ha-ras and Ki-ras genes
responsive to glucocorticoids have now been identified (121, 127,
149, 165). Although it is clear that the glucocorticoid-GR
complex trans-activates ras expression in
epidermal and mammary epithelial cells through these elements, it
remains to be determined whether these elements function in other
epithelia and whether aldosterone also modulates transcription at these
sites. The promiscuity between GR and MR signaling, however, suggests
that aldosterone likely affects Ras expression through these elements.
Indeed, work from the Verrey laboratory (154, 155) on
amphibian distal nephron epithelia demonstrates that aldosterone
signaling through MR and/or GR results in a similar increase in
Ki-ras transcript levels. Unpublished data from our
laboratory and from that of Fuller (Fuller P, personal communication)
show Ki-ras to be regulated at the level of transcription by
aldosterone via MR in rodent heart and colon, respectively. These
findings are also consistent with these cis-acting elements
playing an important role during a mineralocorticoid response in epithelia.
Genes encoding 
-fibrinogen, vitellogenin, and A-Raf are regulated at
the level of transcription by glucocorticoids (3, 95, 136, 145,
166, 183). The effect of aldosterone on these genes is not
presently known, and no information suggests that they are important to
mineralocorticoid action; however, much may be learned about
aldosterone signaling by considering the glucocorticoid paradigm. The
cis-acting elements mediating glucocorticoid responsiveness
for vitellogenin and -fibrinogen are similar to those described for
Ha-ras and Ki-ras (3, 96, 117, 118, 127,
149, 165). They are hSRE (containing variants of either AGAACA
or TGTTCT) that lie near consensus sequences defining half- and full
sites for other transcription factors and/or accessory proteins. As
indicated in Fig. 1, the GR, after forming heterodimers with other
trans-acting factors/accessory proteins, which themselves bind DNA, binds to these hSREs. Recent work demonstrates that the
accessory protein XGRAF (Xenopus laevis GR
accessory factor) forms a dimer with GR, with the GR-XGRAF complex
binding DNA at both a hSRE and the nearby cognate XGRAF DNA-binding
site (GAGTTAA; Refs. 96, 117, and 118).
Binding of the GR-XGRAF heterodimer leads to
trans-activation of the -fibrinogen gene through a
classic mechanism. It is possible that such a molecular mechanism,
albeit with different accessory factors and/or other
trans-acting factors, is utilized by GR to activate
ras and vitellogenin. Although aldosterone clearly induces
ras expression via MR in some tissues (155; Meszaros JG and
Stockand JD, unpublished observations; Fuller P, personal communication), it is not known whether MR is capable of
employing such a molecular mechanism of action involving hSREs. There
are three isoforms of Raf kinase: A-Raf, B-Raf, and C-Raf [also called Raf-1 (103)]. B- and C-Raf phosphorylate
mitogen-activated protein kinase (MAPK)/extracellular signal-regulated
kinase (ERK) kinase (MEK) to activate the MAPK-ERK cascade. The
function of A-Raf is less well documented. The promotor regions of the
human and mouse A-raf genes contain functional SREs that are
trans-activated by the dexamethasone-GR complex in HeLa
cells (94, 95). The functional cis-element
responsive to GR in this gene is two hSREs situated near each other and
close to a cis-acting element responsive to the thyroid
receptor. A-raf is expressed predominately in urogenital tissues, including the kidney (164). It is exciting that
A-Raf is one primary effector of Ras signaling proteins, including
Ki-Ras, which is encoded by a recently identified aldosterone-induced gene (154, 161).
Trans-repression. Glucocorticoids via GR repress expression of a host of genes. Recent studies also have identified several aldosterone-repressed genes and proteins (92, 135, 154). Compared with induction, less is known about the molecular mechanisms of direct gene repression in response to corticosteroids. Almost nothing is known about MR-mediated repression, including a complete paucity of information regarding cis-acting elements and accessory proteins involved in this action. The term "negative"-SRE (nSRE) is used here to define the cis-acting element that directly mediates trans-repression in response to glucocorticoid signaling via GR.
GR interacts directly with the 5'-ATTTTTGTCAATGGACAAGTCATAAGAA-3' nSRE sequence in the promotor region of the corticotropin-releasing hormone gene to trans-repress cAMP-activated expression of this gene (105, 106). A cis-acting element in the promotor of the prolactin gene that is bound by ligand-activated GR also has been identified (141). This element exerts positive tonic regulation of prolactin and heterologous promotors in the absence of liganded GR. Tonic regulation is suppressed in the presence of ligand-activated GR. In a study involving both in vivo and in vitro work, Burke and colleagues (30) mapped the cis-acting element bound by GR during trans-repression of the bovine vasopressin gene. The nSRE in the gene encoding proopiomelanocortin (POMC; 5'-GGAAGGTCACGTCCA-3') has also been identified (35, 48, 49). Compared with SRE involved in trans-activation, which bind GR dimers, the nSRE in POMC simultaneously binds a GR homodimer and monomer in a cooperative manner. It is not presently known how the novel arrangement of three GR around this nSRE affects transcription machinery and other factors to repress expression. The effects of MR on these nSREs have not been studied, and thus it remains to be determined whether MR is capable of modulating expression via these sorts of nSREs. As of the writing of this review, no nSRE definitively responsive to MR has been reported, and it is unclear whether one actually exists. The recent identification of a number of putative aldosterone-repressed genes and proteins (92, 135, 154) points to possible candidates worthy of further investigation in this regard. Another consideration worthy of further study is that gene repression in response to aldosterone does not directly involve binding to nSRE. This alternative mechanism could take one of two forms: 1) a secondary response dependent on the primary induction of factors that ultimately negatively influence transcription of the repressed genes and 2) a response dependent on protein-protein interactions between MR and other accessory or transcription factors that then negatively influence expression without MR actually binding DNA. It has become apparent in the last couple of years that many genes repressed by glucocorticoids in fact do not contain an nSRE directly bound by ligand-activated GR. It is accepted that these genes are repressed by direct protein-protein interactions between ligand-activated GR and other trans-acting factors, which bind DNA in the promotor region of the repressed genes. Indeed, it is now believed that most corticosteroid gene repression actually results from a molecular mechanism involving transcription interference mediated by protein-protein interactions and not direct trans-repression of expression (reviewed in Ref. 78). Also being reevaluated is whether such protein-protein interactions actually play a more prominent role in trans-activation than was initially thought (134).Protein-protein interactions during MR and GR modulation of gene
expression.
The -casein gene is induced by GR independently of SRE
(163). The mechanism of action here entails GR
"tethering" through a protein-protein interaction to signal
transducer and activator of transcription-5 (Stat-5), with the latter
factor associating with its respective cis-element in the
promotor region of the -casein gene (see Fig. 1). Induction of
-casein in response to glucocorticoids, then, is actually dependent
on formation of GR-Stat-5 heterodimers that bind to Stat-5 response
elements without ligand-activated GR actually interacting with its
cognate DNA-binding site. GR forms an equivalent type of complex with
the transcription factor organic cation transporter-2
(130) and has also been reported to associate with Stat-3
(200). It is not known whether MR also binds these and/or
other such tethering factors, and thus such a mechanism of action
involving transcription synergy for MR is purely speculative at this
time. If MR has limited interaction with these tethering factors or
interacts with distinct ones, this mechanism might then distinguish
between GR and MR responses.
B on
transcription via direct protein-protein interactions (reviewed in
Refs. 78 and 79). This action of GR is rapid and
independent of induction of GR-responsive genes. Such cross talk
between GR and AP-1 and NF-B plays a critical role in the
anti-inflammatory and immunosuppressive functions of glucocorticoids. A
possible role for such a mechanism of action during a mineralocorticoid response remains to be determined, but because signaling through GR and
MR has much in common, it merits further consideration.
The functional consequence of protein-protein interactions involving GR
in vivo have been clarified by the use of a transgenic mouse model
containing mutant GR incapable of forming dimers (GRdim),
which thus cannot bind and trans-activate at SRE (72,
78). Also, trans-repression of POMC and prolactin via
nSRE is compromised in GRdim/dim mice, but these mice
retain the ability for ligand-activated GR to interfere via
protein-protein interactions with transcription initiated by other
trans-acting factors (134). Indeed, the ability of activated GR via AP-1 to interfere with trans-activation
of genes encoding collagenase and gelatinase is retained in
GRdim/dim mice. The ability of GR to interfere with
NF-B activity also is retained in GRdim/dim mice.
Distinct activation-deficient GR mutants, as well, retain their ability
to interfere with AP-1 activity, repress interleukin-2 production and
c-myc expression, and induce apoptosis in
lymphocytes (74). It is presently unclear whether MR has
such an extraordinary ability to mediate some of its cellular actions
via protein-protein interactions involved in transcription interference
and/or synergy.
The promotor regions of the aldosterone-sensitive genes sgk,
Ki-ras, and -ENaC all contain consensus AP-1 sites
(104, 114, 127). A possible role for these
cis-acting elements in the context of corticosteroid
regulation has not been directly investigated. In a series of elegant
studies, the laboratory of Ann (98, 184, 197, 198) clearly
showed that GR-mediated trans-activation of -ENaC is
suppressed by an interference mechanism involving activation of the
MAPK signaling cascade. Importantly, the MAPK cascade can ultimately
lead to activation of the AP-1 complex, with the MAPK-activated trans-acting factor c-Jun forming homo- and heterodimers
with c-Fos to target AP-1 sites. It is provocative that GR and
c-Jun/c-Fos form low-affinity protein-protein interactions capable of
mutually interfering with each other's function (reviewed in Refs.
78 and 79). Although such protein-protein interactions
clearly abrogate both GR and AP-1 trans-activation
(193), it is unclear whether coassociation of GR with AP-1
components compromises DNA binding or directly interferes with their
respective trans-acting potential. Thus one could evoke a
dynamic mechanism for regulation of -ENaC, ras, and
sgk involving both classic trans-activation and
transcription interference by GR. It will be interesting to learn
whether MR also mutually interferes with AP-1 elements via protein-protein interactions.
The promotor region of the glucocorticoid-induced gene encoding
phosphoenolpyruvate carboxykinase (PEPCK) contains both AP-1 and NF-B response elements, as well as classic SRE
(183). In contrast to glucocorticoids, AP-1 signaling
decreases PEPCK expression. Similarly, NF-B represses induction of
PEPCK by glucocorticoids. The intereference of both AP-1 and NF-B
with GR is mutually antagonistic (87, 111-113).
NF-B interferes with trans-activation of PEPCK by GR
independently of the former protein's interaction with its cognate
DNA-binding element. Thus it is not unreasonable to suggest that a
system involving GR trans-activation through SRE- and
GR-mediated interference of AP-1 and NF-B repression combines to
control PEPCK levels. As mentioned immediately above, it is unclear
whether MR affects AP-1, but MR clearly interferes with NF-B
trans-activation via protein-protein interactions
(87). The role of NF-B with regard to mutual
interference with MR/GR during a mineralocorticoid response remains unexplored.
The laboratory of Firestone (104) recently identified for
the first time a novel mechanism of reciprocal interference between GR
and the p53 protein. Induction of sgk by GR signaling is
abrogated by simultaneous activation of p53. p53 interferes with GR via protein-protein interactions by preventing its interaction with its
cognate SRE. Substantiating this mechanism are findings that p53
interferes with glucocorticoid-induction of the AGP gene, which
contains a SRE but no p53 binding site in its promotor region (10). Because sgk is clearly induced by
aldosterone via MR in some epithelia capable of Na+
reabsorption (36, 119), it is possible that interference
from p53 may play an important but as yet unexplored regulatory role modulating this and other functions of epithelia influenced by Sgk
during a mineralocorticoid response.
Although not directly or extensively studied, abundant circumstantial
evidence, such as that described immediately above, suggests that
transcription interference and synergy via MR and/or GR may play a role
in mineralocorticoid signaling. Thus further study of this novel aspect
of aldosterone signaling is warranted.
| |
ALDOSTERONE-INDUCED PROTEINS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
ALDOSTERONE REGULATES GENE...
ALDOSTERONE-INDUCED PROTEINS
AN INTEGRATED MODEL
REFERENCES
|
|---|
Significant advances clearly have been made in the last couple of years in identifying aldosterone-induced transcripts and proteins, as well as in determining the molecular mechanism for modulation of aldosterone-sensitive genes. Similarly, our general understanding of the function of aldosterone-induced proteins during a mineralocorticoid response is also becoming clearer. However, much remains to be learned about the mechanism by which aldosterone-induced proteins exert their final control over cellular activity. This is particularly evident when one considers that very little detailed information exists about how aldosterone-induced proteins specifically interact and regulate their respective effectors in epithelia. Our understanding of the cellular signaling cascade transducing the actions of aldosterone at the nucleus outward toward the plasma membrane, where its final effectors are located, is also very limited. These areas of aldosterone signaling are the primary focus of most contemporary research investigating the mineralocorticoid response.
Temporal Actions of Aldosterone
The genomic actions of aldosterone are traditionally divided into an early and late phase (reviewed in Refs. 176-178). This division is somewhat arbitrary, and it is unclear whether and/or how the two phases are related. Moreover, the physiological significance of each discrete phase has never been definitively quantified. However, it has been suggested that the later phase of aldosterone action sets the capacity of transport epithelia for solute and water (re)absorption, and thus this phase may be considered a trophic or chronic response (159, 176, 178). At a superficial level, the genes affected by aldosterone during the later phase, those encoding transport proteins and proteins involved in energy metabolism for example, appear to support this contention. The early phase, in comparison, is predicted to respond to acute changes in salt and water balance to allow for more rapid responses to movements away from homeostasis. That aldosterone primarily affects transcription of signaling factors, such as Sgk and Ki-Ras, during the early phase is consistent with such a mechanism. Because the in vivo effects of aldosterone have been almost exclusively studied at extremes, little information exists regarding dose-dependent actions and threshold effects during either phase in the integrated system. In addition, the significance of time-dependent aldosterone responses in whole animal studies is often obscured by experimental limitations.The early phase of aldosterone action is most often demarcated as the period where Na+ transport is increased without an accompanying increase in the levels of the transport proteins involved in this action. This phase directly follows the 0.5- to 1.0-h latent period required for changes in gene expression and proceeds for 2-4 h. The second phase, then, is classified as that following this period and is associated with a further or sustained increase in transport accompanied by increases in the number of transport proteins, such as ENaC and the Na+/K+-ATPase (reviewed in Refs. 52 and 176-178). Both phases clearly have an absolute requirement on induction/repression of gene expression. For the early phase, aldosterone action is considered to be exclusively mediated through a primary effect on gene expression. In contrast, the later phase results from both primary and secondary effects on gene expression. Both phases also involve regulation at the level of posttranslation. The distinction between these phases demonstrates that aldosterone induces signaling proteins during the early phase that result in activation via posttranslational control of existing proteins involved in transport. These early signaling factors potentially could also lead into the later phase by stimulating a second round of gene expression to increase production and guarantee proper regulation of transport proteins.
Early actions of aldosterone. The initial actions of aldosterone with respect to increasing Na+ (re)absorption and K+ secretion are mediated by existing transport proteins that are targeted by diffusible signaling factors, the expression of which is regulated at the level of transcription by steroids. The limiting step in both transcellular Na+ and K+ movement are the activities of the apical ion channels mediating (re)absorption and secretion, respectively. Although the serosal Na+/K+-ATPase is essential to the maintenance of the electrochemical forces necessary for transport, there is enough inherent capacity within pump number and activity to ensure a constant gradient favoring Na+ (re)absorption and K+ secretion (reviewed in Refs. 52, 60, and 176). Thus the activity of the Na+/K+-ATPase is not typically considered to be limiting during the early phase of aldosterone action. Considering this, then, there are ultimately only two ways aldosterone can enhance transport: 1) by increasing the open probability of apical ion channels and 2) by increasing the number of active ion channels in the luminal membrane. There is considerable and often conflicting evidence supporting both mechanisms of action (47, 56, 59, 67, 73, 80, 84, 85, 189). Importantly, an aldosterone-induced change in channel-gating kinetics and number during the early phase of action is not necessarily mutually exclusive. In fact, as described below, Sgk is presently believed to increase ENaC number in the apical membrane, whereas Ki-RasA is believed to influence channel gating, with both happening during the early phase.
Late actions of aldosterone.
The late actions of aldosterone are primarily trophic. During this
phase, there is a clear increase in the amount of ENaC protein within
the cell, as well as at the apical membrane (100, 107,
189). However, controversy surrounds which subunits of the
heteromultimeric ENaC are induced by aldosterone, with some studies
showing increases in -ENaC levels (100, 107, 110, 114)
and some increases in -ENaC levels (158, 189), and yet others reporting increases in both (42, 47). Moreover,
controversy surrounds the idea that expression and insertion into the
apical membrane of ENaC subunits are discordantly and independently
regulated by aldosterone (67, 100, 107, 139, 189). Similar
to ENaC, the number of Na+/K+-ATPase pumps and
of the secretory, apical K+ channel (ROMK) increase in
response to steroids (180). Indeed, aldosterone
definitively induces, at the level of transcription, expression of the
gene encoding the 1- and possibly
1-subunits of the Na+/K+-ATPase
(24, 88, 89), as well as inducing expression of corticosteroid hormone-induced factor (CHIF; see below), a protein that
shares much similarity with the -subunit of the
Na+/K+-ATPase. In addition to increases in
these transport proteins, enzymes essential to energy metabolism
increase (52, 176, 178). All three actions combine to
establish a cell programmed for prolonged ion transport. Thus the
programming associated with the trophic action of aldosterone could in
a general sense be considered further differentiation of these
epithelial cells. It is provocative, as described further below, that
many of the early signals of aldosterone are also associated with
cellular growth and differentiation.
Function of Aldosterone-Induced/Regulated Proteins
As described above, aldosterone influences expression of a broad pool of genes, and its actions are pleiotropic, involving regulation of several distinct end-effectors through the amalgamation of diverse intermediaries and signaling inputs. A definitive understanding of the roles played by most aldosterone-induced proteins is presently lacking; however, it is accepted that aldosterone signaling is quite complex with much convergence and divergence onto key signaling factors that program the intended cellular response. The two signaling factors induced by aldosterone that have recently garnered much attention as key mediators in a mineralocorticoid response are Sgk and Ki-RasA.2 Figure 2 shows a working model of aldosterone signaling that includes some of the known aldosterone-induced proteins and their activators and effectors.
|
Function of Sgk. Of the recently identified aldosterone-induced genes, sgk has received the most attention (reviewed in Refs. 120, 124, and 125). This gene was established originally in fibroblasts and mammary epithelia as an immediate-early gene induced by glucocorticoids at the level of transcription independently of de novo protein synthesis (186, 187). Subsequently, the Pearce (36) and Naray-Fejes-Toth (119) laboratories simultaneously identified sgk as a primary aldosterone-induced gene in amphibian and mammalian renal epithelia. This transcript also is strongly induced by corticosteroids throughout the gastrointestinal tract but not the lungs (28, 151). From these findings, it is clear that a more complete understanding of how corticosteroids target induction of sgk in kidney and gut but not lung will yield insight into the cellular/molecular mechanisms bestowing steroid specificity. The sgk transcript is also commonly expressed in many nonepithelial tissues in a corticosteroid-insensitive manner, suggesting that some sgk expression must be constitutive (36, 187). Corticosteroids, through both GR and MR, increase sgk levels within 15-30 min of treatment, with levels peaking after 1-2 h and subsequently tending to pretreatment values soon afterward. Induction by aldosterone of Sgk protein in renal epithelia follows a similar time course, with protein levels increasing within 30 min, peaking by 6 h, and returning to pretreatment levels by 24 h (36).
Sgk is a serine/threonine kinase that shares much homology with protein kinase B (PKB)/Akt kinases and phosphorylates at a consensus sequence (RXRXXS/T; optimal site KKRNRRLSVA) similar to that targeted by PKB/Akt, protein kinase C, and p90 ribosomal protein S6 kinase (86, 122). Sgk itself is a phosphoprotein, with phosphorylation by PDK-1 being required for Sgk activity (86, 122). Interestingly, aldosterone increases both absolute and phospho-Sgk levels in renal A6 cells (185), suggesting that, in addition to inducing sgk expression, aldosterone activates a converging signaling cascade that ensures proper phosphorylation of Sgk. The lipid/protein kinase PI3K is upstream of PDK-1 in the PI3K signaling cascade and is required for activation of PDK-1 and, subsequently, Sgk, by insulin, insulin-like growth factor (IGF)-1, and other stimuli (122). Thus Sgk is a constituent of the PI3K signaling cascade positioned downstream of PDK-1 in parallel with PKB/Akt (77, 86, 175). In renal A6 epithelia, inhibition of PI3K attenuates aldosterone-induced increases in Na+ transport and in the active (phosphorylated) but not absolute levels of Sgk (185). This maneuver also blocks activation of Sgk and Na+ transport by insulin in the same cells. These observations are consistent with aldosterone-stimulating Sgk activation by control of sgk transcription and posttranslational modification. It is exciting that, in A6 cells, aldosterone stimulates PI3K activity independently of inducing PI3K expression (20, 21) and induces expression of Ki-RasA (154, 161), which is an upstream activator of PI3K (55, 192). Thus, as discussed further in this and the subsequent subsection, two primary aldosterone-induced genes encode signaling factors Sgk and Ki-RasA, which belong to a converging signaling cascade. Also exciting are results from a recent proteomic analysis of A6 cells treated with aldosterone that identified the-isoform of the catalytic subunit of protein phosphatase 2A
(PP2A; PP2A) as being decreased fourfold 3 h after steroid treatment (Stockand JD, unpublished observations). PP2A
dephosphorylates and inactivates Sgk in in vitro experiments
(122). All of these observations strongly suggest that Sgk
is a critical factor in the aldosterone-signaling cascade initiated in
epithelia. Moreover, they suggest that there are multiple sites at
which aldosterone affects the active levels of Sgk to fine-tune a final response.
Sgk localizes to both the cytosol and nucleus and plays a role in
preventing apoptosis and promoting cell proliferation (2, 31, 65, 115, 122). It is unclear how such an action could ultimately lead to increased transport. Moreover, effector proteins phosphorylated by Sgk and those that physically interact with Sgk for
the most part remain unspecified. Zhang et al. (199) recently identified B-Raf as an effector of Sgk, with phosphorylation by Sgk negatively regulating this protein. It is intriguing that some
isoforms of Raf are induced by corticosteroids (95) and are activated by Ki-RasA, which itself is an aldosterone-induced protein (154, 161). Could Sgk inhibition of Raf direct
Ki-RasA signaling to its alternative first effector, PI3K (55,
138, 192), with subsequent positive feedback onto Sgk itself via
PDK-1 (see Fig. 2)? In addition to B-Raf, the Forkhead family member FKHRL1, a proapoptotic transcription factor, is targeted by Sgk (29). A role for negative regulation of FKHRL1 by Sgk
during a mineralocorticoid response is presently unclear. The ubiquitin ligase NEDD4 directly interacts with the COOH termini of ENaC subunits
at the PY motif and targets this channel for degradation (156,
157). Snyder and colleagues (153) recently showed
that Sgk directly interacts with NEDD4-2 to decrease the latter
protein's function and binding to -ENaC. One consequence is that
ENaC levels increase in the luminal membrane in response to Sgk
signaling via suppression of normal retrieval pathways. Debonneville
and colleagues have observed similar results
(42a). However, conflicting data showing that
mutations within ENaC that destroy the NEDD4-binding site and other
domains targeting retrieval do not interfere with Sgk activation of
this channel in oocytes have also been reported (45, 101,
179). Resolution of this apparent controversy is paramount to
understanding Sgk actions on ENaC.
The only other known activator of Sgk besides PDK-1 (86)
that directly interacts with this kinase is Big MAPK-1 (BMK-1; a newly
described member of the MAPK/ERK family, also termed ERK5; 71). BMK-1
activates Sgk via phosphorylation at a site distinct from that targeted
by PDK-1. Thus BMK-1 and PDK-1 may serve similar functions with respect
to Sgk. Because BMK-1 is required for growth factor-induced cell
proliferation controlling entry into the S phase of the cell cycle, it
is presently unclear whether and how BMK-1-Sgk interactions relate to
induction of Na+ transport by aldosterone in polarized,
differentiated epithelia.
Bolstering a direct role for Sgk in mediating transport are findings
that Sgk physically interacts with the COOH termini of both - and
-ENaC, when the latter proteins are expressed as glutathione
S-transferase fusion proteins (185). These
domains are similar to those targeted by NEDD4 (see paragraph earlier in this subsection), and thus it is possible that NEDD4 acts as a
linker coupling Sgk to ENaC. Moreover, several laboratories have shown
that overexpression of Sgk with ENaC in the heterologous X. laevis oocyte expression system leads to activation of the channel
(36, 45, 101, 119, 151, 179). Enthusiasm for these experiments must be tempered, however, for they were performed in a
nonpolar, non-aldosterone-sensitive, nonepithelial cell only after
chronic Sgk and ENaC overexpression. This heterologous system obviously
is quite different from the in vivo situation of aldosterone-stimulated Na+ transport across polarized, differentiated epithelial
cells. Moreover, the effects of Sgk on ENaC when overexpressed in
oocytes may reflect nonspecific actions for overexpression of this
kinase with the cystic fibrosis transmembrane conductance regulator
(179), as well as certain voltage-gated K+
channels (91) also leads to activation of these latter
channels in the oocyte. In addition to regulating ion
channels, preliminary evidence from oocytes suggests that Sgk also
promotes activation of the Na+/K+-ATPase and
Na+-K+-Cl
cotransporter BSC-1 via
promoting insertion into the plasma membrane when these proteins are
chronically cooverexpressed with the kinase (194). It is
not immediately clear how a "specific" mediator of a
mineralocorticoid response could activate such diverse channel types
and transporters, which localize to distinct membranes in polarized
epithelia. One possible explanation is that Sgk affects membrane
trafficking via a general mechanism. Indeed, this idea is presently the
most favored explanation of Sgk action on transport (124),
but this again raises the specter of how specificity is determined with
such a mechanism of action. Studies of Sgk's effects on ENaC when both
are overexpressed in the oocyte confirm that, in this system, Sgk does
increase the number of ENaC within the plasma membrane independently of
overt effects on the gating kinetics of this channel (45, 101,
179). An alternative to Sgk directly affecting membrane cycling
is that this kinase establishes a cellular program, in part, through
antiapoptotic signals that induce a transport phenotype without
directly affecting transport proteins themselves. In fact, indirect
actions or secondary effects of Sgk on gene expression cannot be
excluded from mediating this kinase's action on transport in any of
the studies performed thus far.
Initial thoughts about the mechanism of Sgk's action during the early
phase of aldosterone signaling included the notion that it directly
phosphorylated ENaC to dynamically regulate ion channel insertion into
and/or retrieval from the apical membrane. Because experimental
manipulations and the introduction of mutations in ENaC that delete
signals which promote retrieval from the membrane to the lysosomal
and/or proteosomal compartments do not abrogate Sgk actions on the
channel in oocytes (38, 45, 151), it is likely that this
kinase regulates insertion. A report demonstrating that ENaC in a
heterologous system was phosphorylated by protein kinase C
(152), which has a consensus sequence similar to that targeted by Sgk (see paragraph earlier in this subsection), seemed to
bolster the idea that Sgk directly affected the channel to promote
insertion into the plasma membrane. However, experiments directly
testing this notion have been negative (45), and attempts to phosphorylate native as well as in vitro transcribed/translated ENaC
with active, purified, recombinant Sgk have been unsuccessful to date
(38, 160). However, the same preparation of active Sgk
phosphorylates at least seven proteins in A6 cell lysate in a
dose-dependent manner (160). This is strong evidence that
direct phosphorylation of ENaC by Sgk does not play a significant role in aldosterone signaling and thus that Sgk-to-ENaC signaling must be
indirect. Further identification of Sgk substrates in epithelia is
anxiously awaited and should better define how Sgk signals to ENaC.
When the studies performed in epithelia to determine the mechanism of
Sgk action on transport, as well as those investigating the actions of
Sgk on cell proliferation and apoptosis, are considered, a
logical suggestion is that Sgk is a pleiotropic signaling factor capable of programming several distinct cellular responses to include
regulation of membrane trafficking. This notion is supported by
findings showing that Sgk resides at distinct cellular locales in
proliferative compared with differentiated cells (2, 65) and that corticosteroids regulate expression at the level of
transcription of other signaling factors, such as Ras, Raf, and
p21waf1/cip1, also important to the regulation
of cell cycle progression (34, 41, 94, 95, 121, 127, 148, 149,
154).
It must be mentioned here that, although extensively studied in the
context of being a specific mediator of mineralocorticoid action, a
preponderance of the evidence suggesting that Sgk actually serves this
function in epithelia is indirect. Only with more direct studies
performed in whole animals and on native epithelial cells will Sgk be
definitively established as a specific mediator of aldosterone action.
Such definitive studies are presently ongoing in several different
laboratories. Indeed, preliminary studies from Lang and colleagues
(Lang F, personal communication) described linkage between the
sgk locus and hypertension in humans. Moreover, preliminary findings from the Kuhl laboratory (191) appear
to conclusively demonstrate for the first time that Sgk plays a pivotal role in the mineralocorticoid response. These investigators generated a
mouse model containing sgk1 lacking a functional kinase
domain. These mice show significantly decreased prenatal viability,
which may be related to loss of the proliferative and antiapoptotic actions of Sgk. Surviving 
/ mice show no gross functional
abnormalities, and histology is normal in all organs assayed, including
the gut and kidney. Moreover, these mice have normal Na+
and K+ metabolism when maintained on a normal diet but
present with inappropriate Na+ wasting when stressed with a
low-Na+ diet. The relatively mild phenotype of these mice,
compared with those of ENaC (76) and MR (16,
17) knockout mice, is thought to suggest that Sgk1-independent
signaling must also regulate Na+ metabolism. One
possibility is that Sgk2 and/or Sgk3 isoforms, which are constitutively
expressed in renal epithelia independently of corticosteroids and
stimulate ENaC-mediated Na+ transport in heterologous
systems (reviewed in Ref. 124), in part complement loss of
Sgk1. Thus there may be functional redundancy between Sgk isoforms.
Liddle's syndrome shows inappropriately high levels of Na+
reabsorption in the distal nephron due to gain-of-function mutations in
ENaC that result in an inability for proper retrieval of this channel
from the apical membrane (69, 143, 144). It is striking that this avid Na+ reabsorption happens in the complete
absence of effective aldosterone levels, suggesting that, in part, a
mineralocorticoid response is independent of actions on channel
insertion, which presumably is the domain of Sgk. Thus it is
obvious, when the phenotypes of Sgk knockout mice and Liddle's
syndrome are considered, that other signals in addition to Sgk must be
important for mediating a mineralocorticoid response.
Function of Ki-RasA. The small, monomeric GTP-binding protein Ki-RasA, like Sgk, is encoded by an aldosterone-induced transcript and plays a pivotal role in mediating aldosterone action in renal epithelia. Spindler and colleagues (154) were the first to identify Ki-ras as an aldosterone-induced gene. In particular, this group identified the A splice variant of Ki-ras as sensitive to aldosterone. There are four homologous Ras proteins: Ha-Ras, N-Ras, Ki-RasA, and Ki-RasB. The latter two result from splice variants encoded by a common gene (reviewed in Ref. 8). Induction of Ki-rasA is a primary action of aldosterone that is independent of de novo protein synthesis. Previously, the Shekhar (127, 149) and Pelling laboratories (121, 165) showed that both the Ha-ras and Ki-ras genes were induced by glucocorticoids. It is unclear why aldosterone does not induce Ha-ras expression in epithelia capable of vectorial Na+ (re)absorption, but this appears to be a common finding among investigators studying aldosterone action (1, 154, 161; Stockand JD, unpublished observations; Fuller P, personal communication). Further study of this apparent controversy may shed light on aldosterone/glucocorticoid and MR/GR specificity. Subsequent to the initial study by Spindler and colleagues (154), Stockand et al. (161) and Spindler et al. (155) showed that aldosterone increased Ki-RasA protein levels. Aldosterone induces Ki-RasA during the early phase of action, with levels rising as early as 30 min after treatment. Controversy has surrounded the importance of Ki-RasA to aldosterone signaling. This controversy centers on concerns that aldosterone-dependent induction of Ki-rasA does not appear common to all epithelia capable of a mineralocorticoid response. Aldosterone via MR preferentially increases Ki-RasA transcript and protein levels in amphibian distal nephron (155), rodent colon (Fuller P, personal communication) and heart (Stockand JD and Meszaros JG, unpublished observations) and via GR in amphibian renal A6 cells (154, 161). Corticosteroids via GR, in addition, increase Ha-ras and Ki-ras levels in epidermal (165) and mammary epithelia (127, 149). In contrast to these findings, the Brown (131) and Verrey laboratories (personal communication) have been unable to detect the effects of aldosterone on Ki-ras in mouse and rat kidney. Robert-Nicoud and colleagues (135), using a distinctly molecular strategy, were also unable to identify sgk and Ki-ras as being induced by aldosterone in a collecting duct cell line. As pointed out by this group, such negative findings likely reflect, in part, the low expression levels of these transcripts in differentiated cells, which make them difficult to identify much less quantify. Resolution of this apparent conflict, however, is essential for distinguishing whether Ki-RasA plays a fundamental and universal role in aldosterone signaling or whether it is critical to steroid signaling only in select tissues and particular species. Nonetheless, in cells where aldosterone does increase Ki-RasA, this protein plays an important role in mediating the mineralocorticoid response.
Mastroberardino and colleagues (109) showed in the heterologous X. laevis oocyte expression system that overexpression of constitutively active Ki-RasA with ENaC has conflicting actions on the ion channel, both stabilizing the open probability and decreasing the number of channels in the plasma membrane. The effect on channel number resulted from nonspecific actions of activated Ki-RasA on induction of oocyte maturation. Although exciting, interpretation of these results was somewhat limited for they were performed in a heterologous system utilizing a nonpolarized, nonepithelial cell type not responsive to aldosterone or capable of regulated solute transport, and only after chronic overexpression of both ENaC and activated Ki-RasA proteins. We demonstrated in a subsequent study that induction of Ki-RasA during the early phase is necessary and sufficient for some part of aldosterone's action on Na+ transport in polarized renal A6 epithelial cells (161). In this study, as well as in a follow-up study (1), Ras was shown to be critical for stabilization of ENaC in the open state. The molecular mechanism by which Ras does this remains elusive. Indirect evidence from several other studies (11, 162), however, suggests that Ras must be in close proximity to ENaC and that the Ras-to-ENaC signal mediating changes in gating kinetics is, at least partially, membrane delimited (reviewed in Ref. 159). Thus direct interaction between Ras and ENaC leading to channel stabilization is not unexpected. Alternatively, Ras could signal to ENaC through effector proteins that are also localized to the plasma membrane. Although aldosterone-induced Ki-RasA clearly affects ENaC gating during the early phase of action, its action on channel number is less clear, but Ras signaling is known to activate PI3K in numerous cell types (55, 138, 192) and possibly Sgk in A6 epithelia (160), both of which affect ENaC number (20, 21, 123, 124). During states of chronic and unrestricted activation of Ras and its downstream effector cascades, specifically the MAPK cascade, ENaC levels are ultimately decreased in both native epithelia and heterologous systems (98, 109, 184, 198). Aldosterone likely does more than just increase Ki-RasA levels. In addition, Ki-RasA activates several distinct effector cascades, including the MAPK and PI3K cascades, which differentially affect ENaC (21, 98, 123, 185, 198). Thus the decrease in ENaC levels resulting from prolonged and unrestricted activation of Ras and MAPK signaling not unexpectedly may reveal a potential, classic negative-feedback pathway that could temper avid Na+ reabsorption in the continued presence of aldosterone. Preliminary results presented in abstract form (160) indeed demonstrate that MAPK signaling downstream of Ki-RasA does not play a role in the positive actions of aldosterone but that Ras-dependent PI3K signaling does. As noted above, Ras and MAPK signaling decrease-ENaC expression through a mechanism of
transcription interference with corticosteroid signaling (98,
184, 198). Thus ENaC is potentially both positively (via direct
actions) and negatively (via secondary actions mediated by Ras
MAPK
signaling) regulated by aldosterone at the level of transcription. With
this signaling cascade, the primary actions of aldosterone on induction
of Ki-RasA and ENaC would favor increased Na+ transport
during the early and late phases, respectively, with continued Ki-RasA
signaling eventually leading, via the MAPK cascade and, possibly, AP-1
interference with corticosteroid signaling, to a secondary suppression
of ENaC expression during the later phase, which would then temper
Na+ reabsorption.
Function of PI3K during an aldosterone and insulin response. PI3K, a multimeric enzyme containing both catalytic and regulatory subunits, is important to both aldosterone- and insulin-dependent actions on epithelia (20, 21, 123, 133). However, PI3K is not an aldosterone-induced protein, but its activity is increased by both aldosterone and insulin in renal epithelial cells. Blockade of PI3K impedes both the early and late phases of aldosterone actions, with PI3K apparently promoting/protecting ENaC levels in the apical membrane. Similar effects are observed when PI3K is inhibited during insulin induction of Na+ transport. Thus PI3K is either permissive for Na+ transport or it is common to both the aldosterone- and insulin-signaling pathways that culminate in increased Na+ transport (see Fig. 2). If the latter scenario is true, then PI3K activity must be directly and continuously linked to ENaC activity for, when this kinase is inhibited, sustained Na+ transport is quickly diminished even in the continued presence of aldosterone and insulin. This observation, in conjunction with those showing that aldosterone-sensitive Sgk levels rise within 1-2 h and then return by 4-6 h to pretreatment levels in the continued presence of aldosterone (see above), seems to indicate that PI3K plays a role in addition to that of activating Sgk during chronic aldosterone signaling. At this time, the Sgk-independent role for PI3K is unclear. It is possible, however, that some epithelia have enough basal Sgk1 and/or Sgk2 and -3 expression to enable constant PI3K signaling to ENaC through this pathway.
In contrast to aldosterone, which results in an increase in transport only after a latent period required for gene expression, insulin quickly (within minutes) stimulates transport independently of gene modulation, utilizing a signaling cascade that is initiated at the plasma membrane. Insulin is recognized to stimulate PI3K through a transduction pathway involving several adapter proteins that ultimately couple PI3K activity to allosteric changes in the insulin receptor. How aldosterone induces PI3K activity and where this kinase fits into the aldosterone transduction pathway remain less clear. One possibility that is consistent with all present findings is that aldosterone-dependent induction of Ki-RasA leads to activation of PI3K (see Fig. 2). PI3K is a well-known first effector of Ras proteins, including Ki-RasA (55, 192). With such mechanisms of PI3K activation, insulin would quickly affect the kinase, whereas aldosterone would only affect the kinase after the latent period required for Ki-RasA transcription and translation. In addition to aldosterone and insulin, vasopressin has been reported to stimulate Na+ transport in A6 cells through a mechanism dependent, in part, on PI3K (50). Thus PI3K may be a focal point where insulin, vasopressin, and aldosterone signal transduction converge to activate a common cascade directed toward ENaC and/or the Na+/K+-ATPase.Function of CHIF. CHIF was first identified as a corticosteroid-induced gene in rat colon (6). CHIF is expressed in epithelia of the distal colon and nephron (6, 33, 181). This protein is localized primarily to the basolateral membrane, where it presumably interacts with its final effector to stimulate transport (150). Corticosteroids via MR but not GR regulate CHIF expression at the level of transcription in colon but not kidney (27, 33, 181, 182). Induction of CHIF in the colon in response to corticosteroids is a primary action independent of de novo protein synthesis, with CHIF levels increasing as early as 1 h after steroid treatment (27). In contrast to the colon, in the kidney, CHIF expression is modulated by aldosterone at the level of translation and/or posttranslation (150). Thus CHIF expression must be regulated differently by aldosterone in the colon and kidney. How aldosterone modulates CHIF translation in the kidney remains a mystery, but identification of the cis-acting element and molecular mechanism regulating CHIF expression in the colon is excitedly awaited. Because this element is responsive to MR but not GR, identifying it in the gut will likely yield great insight into the mechanisms defining corticosteroid hormone receptor specificity.
CHIF is a member of the newly identified FXYD protein family as defined by Sweadner and Rael (167). CHIF, similar to other FXYD proteins, is a transmembrane regulator of ion channels and other transport proteins (6, 147). The-subunit of the
Na+/K+-ATPase is also a FXYD protein. It is
unclear whether family membranes can substitute functionally for one
another; however, this remains a distinct possibility and may readily
explain the role played by CHIF during a mineralocorticoid response.
The -subunit is well known to modulate the activity of the
Na+/K+-ATPase (5, 14, 171). Recent
reports showing that CHIF associates with the pump in renal tissue and
stimulates pump activity in the heterologous oocyte expression system
suggest that CHIF may serve a similar function (13, 194).
Although CHIF is proposed to modulate primarily K+
homeostasis, it likely also plays an important role in Na+
homeostasis, considering that its primary target appears to be the
serosal Na+/K+-ATPase. By activating the
serosal pump, CHIF likely serves a supportive function in maintaining
cellular electrochemical gradients favorable for luminal
Na+ absorption and K+ secretion but has little
role in directly modulating the limiting steps of
transcellular transport. This mechanism is consistent with findings
showing that, during the early phase of aldosterone action, preexisting
Na+/K+-ATPase pumps are activated in response
to the genomic actions of this steroid (128).
Function of glucocorticoid-induced leucine zipper protein.
Robert-Nicoud et al. (135), using serial analysis of gene
expression, identified in an immortalized mouse principal cell line,
which contains MR, the glucocorticoid-induced leucine zipper protein
(GILZ) as being encoded by an aldosterone-induced gene (135). GILZ transcript levels were increased by
aldosterone within 30 min, indicating that it is an early signal. GILZ
belongs to the transforming growth factor--stimulated clone 22 (TSC-22)/DSIP-immunoreactive leucine zipper protein/bunched
(TSC-22/DIP/bun) family of proteins originally thought to be
transcription factors but recently reconsidered to serve an as yet
undetermined function. How GILZ could affect proteins involved
in transport remains unclear; however, it is becoming clear that
leucine zippers, in addition to affecting transcription, are also
capable of modulating ion channel activity in a dynamic manner
(reviewed in Ref. 102). While serving this function,
leucine zippers act as adapter proteins to recruit kinases and
phosphatases into a macromolecular complex with the regulated ion
channel. Interestingly, using a distinct genomics approach, we also
identified TSC-22 with two different probes as being encoded by a
corticosteroid-induced transcript in the M-1 cell line that is
increased within 3 h of steroid treatment (Stockand JD and Eaton
DC, unpublished observations).
Function of Aldosterone-Induced Transport Proteins
As discussed above (see Late actions of aldosterone), aldosterone induces expression of ENaC, ROMK, and the Na+/K+-ATPase pump at the level of transcription, with this induction being part of the trophic, late actions of aldosterone. The functional consequences of these effects of aldosterone with respect to a mineralocorticoid response are straightforward. In addition to these transport proteins, aldosterone induces expression of the luminal Na+/H+ exchanger (NHE3) in the proximal but not distal portion of the colon (39) and the luminal, thiazide-sensitive Na+-Cl cotransporter (NCC) in the distal
renal tubule (83). Presesntly, it is unclear whether
induction of NHE3 and NCC are primary responses to aldosterone, but
both actions appear to be associated with the trophic effects of this
steroid. Again, the functional consequences of these actions with
respect to a mineralocorticoid response are straightforward. Increases
in NHE3 and NCC translate into the sustained elevation of
electroneutral Na+ (re)absorption in the colon and kidney
associated with volume contraction (54, 174).
| |
AN INTEGRATED MODEL |
|---|
|
TOP
ABSTRACT
INTRODUCTION
ALDOSTERONE REGULATES GENE...
ALDOSTERONE-INDUCED PROTEINS
AN INTEGRATED MODEL
REFERENCES
|
|---|
It is hard to overlook the fact that there is possibly a linear
signaling relationship between aldosterone-induced Ki-RasA and Sgk,
with PI3K positioned between these factors. In addition, other
corticosteroid-regulated proteins, such as PP2A and A-Raf, potentially influence or are influenced by these factors. When the
common nature of these factors to signaling cascades that control
cellular growth, apoptosis and differentiation is considered, one generalized view of the actions of aldosterone on epithelia is that
this steroid programs the cell to "differentiate" more toward a
Na+-reabsorbing state and that Ki-RasA and Sgk are merely
the early messengers of this signal. Adding further support for such a
generalized mechanism is that corticosteroids, via the initiation of a
complex interaction between Ras and PI3K signaling, are known to induce functional polarity and promote formation of tight junctions and transepithelial resistances in mammary epithelia (32,
190). Moreover, signaling through MR promotes differentiation of
brown adipocytes (126), and PI3K is a central switch
directing tubulogenesis of epithelial cells (82). This
generalized cascade would contain multiple converging and diverging
pathways, exerting pleiotropic effects on epithelia. Shown in Fig. 2 is
one possible signaling cascade that includes many of the known
aldosterone-regulated proteins. What is clear in this idealized cascade
are the many sites for possible cross talk and feedback. For instance,
Ki-RasA can activate PI3K, which in turn can activate PDK1, which then activates Sgk. Ki-RasA also activates the MAPK cascade via stimulation of Raf. Active Sgk is a negative regulator of B-Raf through
phosphorylation (199). Could Sgk possibly also regulate
the corticosteroid-induced gene A-Raf and other Raf proteins, such as
c-Raf, that are known effectors of aldosterone-induced Ki-RasA?
PKB/Akt, which shares much homology with Sgk, phosphorylates a similar
consensus site and is positioned at the same site as Sgk in the PI3K
signaling cascade, inhibits both B- and c-Raf via phosphorylation
(66, 77, 138, 202). An additional site of possible cross
talk regulation is between MAPK and Sgk, for MAPK signaling in response
to stimulation of Raf induces expression of sgk in
fibroblasts (116). Here again is a possible
positive-feedback signal with aldosterone-induced Ki-RasA activating
Raf and the MAPK cascade to further ensure sgk expression,
which subsequently directs Ki-RasA signaling to the PI3K cascade by
Akt-mediated inactivation of Raf. Thus Sgk expression and
proper phosphorylation could, in part, be influenced by Ki-RasA
signaling. Numerous other points of possible cross talk exist in this
cascade. For instance, aldosterone-induced Ki-RasA via MAPK signaling
has been shown in A6 cells to lead to induction of MAPK phosphatase-1
(MKP-1) (160), which is a negative-feedback regulator of
MAPK. Indeed, activation of MKP-1 directs RasRaf signaling to
cascades other than the MAPK cascade (148). MAPK, in
addition, is a negative regulator of the GTP exchange factors, such as
SOS, that stimulate Ras activity (25, 26, 55).
Interestingly, PI3K via PDK-1 on the other hand, positively influences
GTP exchange factors to prolong Ras signaling (46, 132).
Thus, if all of these points of cross talk hold true in
corticosteroid-sensitive epithelia, then aldosterone induction of
Ki-RasA simultaneously with Sgk would preferentially lead to RasPI3KSgk signaling with inhibited MAPK signaling. Such a system of transduction is consistent with all the present literature describing the positive actions of Sgk and Ki-RasA, as well as the
negative actions of prolonged MAPK signaling on transport.
| |
ACKNOWLEDGEMENTS |
|---|
Drs. A. Firulli, K. Hamilton, and R. T. Worrell are recognized for critical evaluation of this article.
