|
|
||||||||
National Heart, Lung, and Blood Institute, Bethesda, Maryland
ABSTRACT I. INTRODUCTION II. PERTURBING EFFECTS OF HYPEROSMOLALITY A. Cell Cycle Arrest B. Apoptosis C. DNA Damage D. Oxidative Stress E. Inhibition of Transcription and Translation F. Mitochondrial Depolarization III. HIGH NaCl: HYPERTONICITY A. Many Cellular Components and Functions Are Affected by Hypertonicity B. Cytoskeletal Rearrangement in Response to Hypertonicity IV. ORGANIC OSMOLYTES ACCUMULATE IN RESPONSE TO HYPERTONICITY A. Sorbitol B. Betaine C. Inositol D. Taurine E. GPC V. ACTIVATION OF THE TRANSCRIPTION FACTOR TonEBP/OREBP BY HYPERTONICITY A. TonEBP/OREBP mRNA and Protein Abundance B. Dimerization of TonEBP/OREBP C. Phosphorylation of TonEBP/OREBP D. TonEBP/OREBP Nuclear Localization E. TonEBP/OREBP Transactivational Activity F. Interactions Between Proteins That Regulate TonEBP/OREBP G. Inhibition of TonEBP/OREBP H. DNA Breaks, ROS, and Cytoskeletal Perturbations as Regulators of TonEBP/OREBP VI. ADAPTATION TO CHRONIC HYPERTONICITY VII. SENSORS AND SIGNALERS OF HYPERTONICITY VIII. HIGH UREA A. Cellular Components and Functions Affected by High Urea B. Counteraction of Urea Effects by GPC C. Mutually Protective Effects of High Urea and High NaCl IX. WHY IS OSMOTIC TOLERANCE OF RENAL MEDULLARY CELLS GREATER IN VIVO THAN IN CULTURE? X. OSMOPROTECTION BY GENERAL STRESS PROTEINS IN CELL CULTURE AND IN VIVO A. Gadd45 B. p53 C. Heat Shock Proteins D. Heme oxygenase 1 XI. SUMMARY AND CONCLUSIONS GRANTS ACKNOWLEDGMENTS REFERENCES
| ABSTRACT |
|---|
|
|
|---|
| I. INTRODUCTION |
|---|
|
|
|---|
|
|
| II. PERTURBING EFFECTS OF HYPEROSMOLALITY |
|---|
|
|
|---|
Acute elevation of NaCl or urea produces rapid arrest in all phases of the cell cycle. The cells in G1 do not proceed to initiate DNA replication. Those in S stop replicating their DNA. Those in G2 do not divide. The duration of the arrest depends on how high the NaCl or urea is. The arrest is shortest in G1 and S, so cells accumulate in G2 (198). By the time that the arrest ends, the cells are sufficiently adapted so that can they proliferate through many passages at the elevated osmolality. The rate of proliferation of adapted cells is directly related to the osmolality (198). When NaCl is elevated to 400–500 mosmol/kgH2O, proliferation is the same as at 300 mosmol/kgH2O, but at higher levels it decreases progressively. Finally, there is a limit as to how high a concentration the cells can withstand. Beyond this they die, as discussed in section IIB. Cell cycle arrest provides time for cells to adapt to higher osmolality by accumulation of compatible and counteracting organic osmolytes (see sects. IV and VIIIB) and increased expression of heat shock proteins (see sect. X). In the absence of such protective adaptation, the cells die.
DNA damage is a well-known cause of cell cycle arrest. High NaCl increases the number of DNA breaks in cells (see sect. IIC), and some of the same signals that arrest the cell cycle in response to DNA damage from other causes also play a role in high NaCl-induced cell cycle arrest. Delay in G1 and S depends on activation of p53 following both high NaCl (72) and ionizing radiation (IR) (100). In the case of high NaCl, the delay in S is sensitive to caffeine, implying involvement of ATM/ATR kinases (72), which are targets of p53 (100). Delay in G2 depends on activation of p38 following both high NaCl (74) and ultraviolet (UV) radiation (28).
Despite these similarities, there are major differences between the cell cycle delay caused by high NaCl and that from other causes. DNA breaks from the other causes are repaired during the cell cycle arrest, but those caused by high NaCl are not (see sect. IIC). Furthermore, the mechanisms differ. Phosphorylation of Chk1 is important for the cell cycle arrest induced by the DNA damage from other stresses (173), but high NaCl does not cause phosphorylation of Chk1 and, in fact, prevents its phosphorylation in response to DNA damage caused by UV or IR (73). Also, inhibition of Chk1 by UCN-01 prevents the cell cycle delay caused by other DNA damaging agents, but not the cell cycle delay caused by high NaCl (73). The question remains as to why the increased DNA breaks that persist in cells adapted to high NaCl do not activate cell cycle checkpoints. Possibly, they are modified so that they are invisible to the DNA damage sensors or the DNA damage response is inhibited.
The cell cycle arrest induced by high urea closely resembles that induced by high NaCl (198), but less is understood about its mechanism. Urea does not activate some of the DNA damage response proteins involved in the cell cycle arrest induced by high NaCl, such as p53 (72) and p38 (149, 251). High urea increases phosphorylation of ATM in HEK293 cells (124), but we do not know whether this is involved in the accompanying cell cycle arrest.
Cell death results when acute hypertonicity exceeds a certain level. The level differs between cell lines, but is rather sharply defined in each. Above that level, massive cell death occurs within several hours. The dying cells exhibit classical features of apoptosis (25, 94, 198, 257). DNA condenses and apoptotic bodies appear, the DNA becomes fragmented with appearance of a "DNA ladder" in gel electrophoresis, the broken DNA ends are labeled in TUNEL assays, and phosphatidylserine becomes exposed at the cell surface. There are two basic pathways for apoptosis: an intrinsic (mitochondrial) pathway and an extrinsic (death receptor) pathway (132). There is evidence favoring both of these pathways as a cause of hypertonicity-dependent apoptosis. In favor of the intrinsic pathway is that hypertonicity causes a rapid mitochondrial dysfunction (see sect. IIF) that precedes apoptosis. Thus acutely exposing mIMCD3 cells to a lethal level of osmolality (700 mosmol/kgH2O, NaCl added) rapidly depolarizes their mitochondria, decreases mitochondrial Bcl2/Bax ratio, and increases cell ADP/ATP ratio (197). However, the mitochondria do not release cytochrome c, and although their Bcl2/Bax ratio decreases, it also decreases at 500 mosmol/kgH2O without any subsequent apoptosis. In favor of the extrinsic pathway is that acute exposure of HeLa cells to 900 mosmol/kgH2O (sorbitol added) causes clustering and internalization of tumor necrosis factor-
(TNF-
) receptors (253). Also, increasing osmolality to 500 mosmol/kgH2O by adding NaCl causes apoptosis in macrophages (U937) and B lymphocytes (LCL721), accompanied by increased production of TNF-
, suggesting that they produce their own death ligand and self-destruct (162). Neutralization of the TNF-
by a specific antibody protects the U937 cells from apoptosis, but it does not protect the LCL721 cells, indicating that the autogenous TNF-
could be triggering the apoptosis in the former, but not in the latter. In another example, increasing the osmolality bathing primary rat hepatocytes to 405 mosmol/kgH2O (NaCl added) causes trafficking of CD95 to the plasma membrane, accompanied by caspase-8 and -3 activation; however, this only amounts to a priming for apoptosis because full-blown apoptosis (DNA fragmentation and phosphatidylserine redistribution) does not occur unless exogenous CD95 ligand is also added (244). We conclude that, since hypertonicity has been observed to trigger components of both the intrinsic and extrinsic pathway to apoptosis in one type of cell or another, it is premature to conclude that either pathway is always involved.
The fact that both hypertonicity and apoptosis cause cells to shrink has led to recognition that the processes are linked. Thus apoptotic volume decrease (AVD) is an early event in apoptosis. It precedes cytochrome c release from mitochondria, and it is maintained until the cell fragments (184, 224). Inhibition of AVD prevents apoptosis from going to completion (184, 224). Regulatory volume increase (RVI) is the usual initial response to cell shrinkage by hypertonicity (311). During RVI, electrolytes and water are taken up into the cell, restoring its volume. An ion carrier (NKCC1) and ion exchangers (NHEs) are activated in RVI (Table 1 and Ref. 311). Although RVI occurs in cells shrunken by hypertonicity, it is impaired in cells undergoing an AVD initiated either by a death receptor ligand (Fas) or an by an inducer of mitochondria-mediated apoptosis (staurosporine) (185). Given that AVD is an early event in apoptosis and that RVI does not occur, it seemed possible that cell shrinkage induced by hypertonicity, if uncompensated by RVI, might by itself cause apoptosis (24, 25). This idea is supported by a number of observations. RVI is impaired in T lymphocytes, and they undergo apoptosis at a much lower level of tonicity than do cells proficient in RVI (25). CHO (PC120) cells, which are deficient in RVI because they lack NHE-1, are particularly susceptible to hypertonicity-induced apoptosis. Reconstituting them with NHE-1 restores RVI and protects them (185). Finally, inhibition of NHE by amiloride sensitizes HeLa (185) and renal tubular epithelial (RTC) cells (319) to hypertonicity-induced apoptosis. Taken together, these observations indicate that hypertonicity-induced shrinkage, if uncompensated by RVI, can cause apoptosis. What is lacking at this point is an understanding of how cell shrinkage activates the executioners of apoptosis. Although hypertonicity shrinks cells immediately, several hours elapse before the final stages of apoptosis are reached (25, 88, 185, 197). Interestingly, when cells are exposed to levels of hypertonicity that would be deadly if prolonged, returning the cells to normotonicity after a brief time (0.5–1 h) prevents apoptosis (24; Dmitrieva NI and Burg MB, unpublished results). Thus there must some additional processes between the initiation of hypertonicity and the point of no return from apoptosis. Those processes apparently include activation of Rac and p38, followed by phosphorylation and nuclear translocation of p53 (88).
Acute elevation of osmolality from 300 to 500–600 mosmol/kgH2O by adding NaCl to cells causes DNA strand breaks in mIMCD3 cells (73, 153), accompanied by cell cycle arrest in all phases of the cell cycle (see sect. IIA). The cell cycle arrest that is associated with most other causes of DNA damage provides a lull in DNA replication during which the DNA can be repaired (345). Thus it seemed logical that, during the transient cell cycle arrest that follows acute elevation of NaCl, the DNA breaks would be repaired. Surprisingly, they are not. Numerous DNA breaks persist in cells that have adapted to high NaCl and reentered the cell cycle (75). Despite these DNA breaks, the adapted cells proliferate rapidly, appear normal, and exhibit little, if any, apoptosis (42, 75, 259). Even more surprising, there are numerous DNA breaks in cells in the mouse renal inner medulla as long as its interstitial NaCl concentration remains at its normal high level (75).
High urea causes a different sort of DNA damage. Neither acute (153, 153) nor chronic (75) exposure of cells to urea causes DNA strand breaks. Furthermore, addition of urea does not significantly increase the number of breaks caused by high NaCl (75), implying that NaCl, alone, is responsible for the DNA breaks in the normal renal inner medulla. However, high urea does produce oxidative DNA damage, namely, 8-oxoguanine lesions (337). If 8-oxoguanine lesions are not repaired, they produce mutations during DNA replication (104). Thus renal medullary cells have DNA breaks caused by high NaCl and oxidative lesions caused by high urea. Nevertheless, the cells survive and function in vivo and, following adaptation, proliferate rapidly in culture.
Normally, DNA breaks attract and activate a protein complex that rapidly repairs the DNA. If the repair fails, the cells die by apoptosis. Cells deficient in components of the DNA repair system are very sensitive to extrinsically caused DNA damage and may not be viable because they cannot repair DNA breaks that occur during normal replication or transcription (173, 183, 320, 349). Remarkably, DNA breaks induced by high NaCl persist as long as NaCl remains high. Yet, despite the DNA breaks, apoptosis is not evident, and the cells proliferate rapidly in culture (75). Reducing NaCl back to normotonic levels in culture (73) and in vivo (75) results in rapid repair of the breaks, implying that high NaCl inhibits DNA repair. Direct proof for this came from measuring the rate of repair of luciferase reporter plasmids damaged by UV before they were transfected into cells. The repair is much faster at 300 mosmol/kgH2O than when NaCl is added to 500 mosmol/kgH2O (73, 75).
Although it is clear that hypertonicity induces DNA damage, the molecular basis is poorly understood. Hypertonicity increases expression or activity of several DNA damage response proteins including Gadd45 (see sect. XA), p53 (see sect. XB), ATM (see sect. V, D and E), and chk2 (264). Other DNA damage response proteins are not affected. Thus other genotoxic agents that induce DNA breaks cause phosphorylation of Chk1, signaling cell cycle arrest (27, 122), but high NaCl does not (73, 75, 264). The effect of high NaCl on some of the other DNA damage response proteins is controversial. Thus Mre11 exonuclease is a nuclear protein that normally binds rapidly to DNA breaks, beginning the process of repair (64), and histone H2AX becomes phosphorylated at DNA double-stranded breaks caused by a number of agents (90, 250, 309), both being prerequisite for successful DNA repair (45, 229). In some studies, high NaCl decreased the chromatin-bound fraction of Mre11 (76) and caused it to move from the nucleus to the cytoplasm, where it is unavailable to initiate DNA repair (73, 75). Also, phosphorylation of histone H2AX was not induced (73, 75). These aspects of the DNA damage response became activated immediately when NaCl was reduced. When osmolality was lowered from 500 to 300 mosmol/kgH2O, Mre11 moved back into the nucleus, histone H2AX and Chk1 became phosphorylated, and the DNA breaks disappeared within a few hours. These results are consistent with the idea that high NaCl inhibits those particular components of DNA damage response or that the response to high NaCl modifies or camouflages DNA breaks so that they are not recognized by the usual DNA damage response system. However, in another study, Mre11 stayed in the nucleus after NaCl was increased and histone H2AX became phosphorylated (264). Why the results differ remains unresolved. The authors of the latter study (264) speculated that the transient H2AX phosphorylation that they observed indicated repair of any DNA double-strand breaks, but they did not actually measure the nature or extent of DNA damage. In summary, direct measurements demonstrate that high NaCl increases the number of DNA breaks both in cell culture and in vivo. However, the exact nature of the breaks, the mechanism by which they are induced, and the details of the cellular response remain to be settled.
The DNA-protein kinase (DNA-PK) complex, which contains Ku86, Ku70, and DNA-PKcs, contributes to repair of DNA double-strand breaks by nonhomologous end joining (NHEJ) (193, 222, 274). When cells that lack Ku86 are exposed to high NaCl, they have aberrant mitosis, slower growth, and increased chromatin fragmentation compared with wild type cells (76).
Elevation of NaCl or urea causes oxidative stress (328, 337, 340, 347, 348, 352) with elevation of ROS (328, 337, 348). ROS can modify DNA bases by forming 8-oxoguanine (104) and can carbonylate proteins (170). Mutations result when DNA containing 8-oxoguanine lesions is replicated (104). Carbonylation affects protein stability and function, notably inhibiting enzyme activity (279). In cell culture, carbonylation occurs within 5 min and with addition of as little as 5 mM urea, which is the normal level in plasma. Since carbonylation increases as urea is raised over the range found in uremia (337), urea apparently can contribute directly to the carbonylation found in that condition. Strikingly, a high level of protein carbonylation is present in vivo in inner medullary cells (337), which are normally exposed to high NaCl and urea. Carbonylated proteins are not repaired. Instead, they undergo accelerated degradation by proteasomes and are replaced (169). At this point we do not know what the functional consequences are of urea- and NaCl-induced oxidation of renal medullary proteins, but it is of obvious interest to find out. Oxygen tension is low in the renal inner medulla because its countercurrent exchange system limits the access of oxygen. Glycolysis predominates there (9). The low oxygen tension and respiration may minimize the oxidative stress in this tissue.
ROS can be destructive, but they also have important signaling functions (reviewed in Ref. 87). Several observations support a role for ROS in signaling osmoprotective responses: 1) antioxidants that reduce superoxide suppress high NaCl-induced increases of both TonEBP/OREBP transcriptional activity and expression of BGT1 mRNA (see sect. VH)(347, 348). 2) The antioxidant N-acetyl-L-cysteine (NAC) prevents urea-induced increase of mRNA and protein abundance of the oxidative stress-responsive transcription factor Gadd153/CHOP (340). 3) NAC also prevents both high NaCl-induced phosphorylation of ERK1/2 and p38 and increase of COX-2 protein expression (328). 4) Lowering ROS production by mitochondrial inhibitors reduces high NaCl-induced increase of COX-2 (328).
E. Inhibition of Transcription and Translation
High NaCl reversibly inhibits synthesis of DNA, RNA, and protein in cultured cells (248). Synthesis recovers within 10 min if NaCl is lowered (248) and more slowly as cells adapt to sustained hypertonicity (235). The effect on DNA synthesis has already been discussed in section IIA. That on transcription and translation is discussed in what follows.
Hypertonicity rapidly and reversibly inhibits RNA synthesis in HeLa cells (248). Increasing osmolality to 450 mosmol/kgH2O by adding NaCl inhibits synthesis of heterogeneous nuclear RNA (mRNA precursor), whereas higher concentrations of NaCl are necessary to inhibit synthesis of ribosomal and transfer RNA (232).
Hypertonicity also profoundly inhibits protein synthesis (248, 254). The inhibition is independent of the solute added (NaCl, KCl, or sucrose), occurs within several minutes, and reverses completely when normotonicity is restored. Inhibition was observed in HeLa cells (254), in chicken embryo fibroblasts (235), and in Madin-Darby canine kidney (MDCK) cells (58). Translation of globin mRNA by rabbit reticulocyte lysate in vitro is inhibited when osmolality is increased by adding NaCl, KCl, CH3CO2Na, or CH3CO2K, but not by adding the compatible osmolytes betaine or myo-inositol (26). Taken together, these results suggest that hypertonicity-induced increase of intracellular inorganic ions inhibits translation, but the inhibition abates as the inorganic ions are replaced by compatible organic osmolytes (see sect. IV) (26). Inhibition of translation could be responsible for the cell cycle arrest. The onset and ending of the cell cycle arrest (72) coincide temporally with the changes in protein synthesis (248, 254). A possible link is that cyclins are translationally regulated.
Hypertonicity reduces translation by inhibiting polypeptide chain initiation (248, 254, 313), whereas elongation and termination of those polypeptides already initiated generally proceed normally (254). Although hypertonicity inhibits translation of most cellular proteins, translation of some, such as heat-shock proteins, actually increases (see sect. XC) (209). The mechanism by which hypertonicity inhibits translation is not well understood, but there are some suggestive findings. Initiation factor eIF2
contributes to inhibition of translation by hypertonicity in reticulocytes and fetal liver nucleated erythroid progenitors (181). In those cells hypertonicity causes autophosphorylation and activation of heme-regulated inhibitor (HRI), which in turn phosphorylates and activates eIF2
. The activated eIF2
inhibits translation (181). Other unidentified eIF2
kinases may also contribute (181). The lengths of untranslated regions (UTRs) in mRNAs and their secondary structure may also play a role. With the use of the rat preproinsulin II-coding sequence linked to the 5'-untranslated sequence of the message that encodes viral SV40, insertion of a hairpin barely affects translation in normotonic medium, but adding NaCl to increase osmolality by 240–300 mosmol/kgH2O virtually abolishes translation (148). When a 136-nucleotide unstructured sequence is placed just upstream from the hairpin, translation in hypertonic medium is restored. Thus leader length and secondary structure can have opposing effects on translational efficiency under hypertonic conditions (148).
F. Mitochondrial Depolarization
Increasing osmolality by adding NaCl perturbs mitochondria in mIMCD3 (197) and Vero (60) cells, although the effects differ in detail. In the mIMCD3 cells, increasing osmolality to 700 mosmol/kgH2O causes mitochondrial depolarization, but increasing it to only 500 mosmol/kgH2O does not. Nuclei become condensed and caspase-9 activity increases at 700 mosmol, indicative of apoptosis, but this does not occur at 500 mosmol/kgH2O. Electron microscopy does not reveal any mitochondrial structural changes at either of the elevated osmolalities. Although NADH exits the mitochondria at 700 (but not at 500) mosmol/kgH2O, cytochrome c remains in the mitochondria at all osmolalities. In the Vero cells, high NaCl causes rapid mitochondrial depolarization, as in mIMCD3, but the effect occurs at a much lower osmolality. Increasing the osmolality by as little as 60 mosmol/kgH2O depolarizes mitochondria in the Vero cells, whereas increasing it by 200 mosmol/kgH2O does not in mIMCD3 cells. At 600 mosmol/kgH2O mitochondria rapidly fragment in Vero cells, but the effect reverses when NaCl is lowered. In contrast, the mitochondria remain intact in mIMCD3 cells. High NaCl does not cause cytochrome c to exit from the mitochondria in either cell type. Given these results, it is not clear whether mitochondrial depolarization is a rapid response of viable cells to hypertonicity (Vero cells), or whether it is an early event in apoptosis (mIMCD3 cells). Possible mechanisms for the depolarization are reviewed in Reference 197. We conclude that high NaCl greatly perturbs mitochondria, but the significance of this effect and its mechanism remain uncertain.
| III. HIGH NaCl: HYPERTONICITY |
|---|
|
|
|---|
High NaCl is hypertonic, and NaCl is the principal determinant of tonicity of the extracellular fluid. Although cell membranes are permeable to Na, the Na+-K+-ATPase transports Na out of cells, which maintains the normal cell volume. In examining the effects of high NaCl, however, we also have to keep in mind that in addition to raising tonicity, high Na and Cl may also have specific effects. For example, elevated chloride specifically increases the expression of
-Na+-K+-ATPase (43). Thus, although cellular changes caused by high NaCl may generally be due to hypertonicity, this is not necessarily so. For that reason, poorly permeating organic solutes like sorbitol, mannitol, or raffinose are often used to produce hypertonicity (Table 1). Of course, these also may have specific effects. Therefore, to establish definitively that a particular effect is caused by hypertonicity, hyperosmolality should be tested adding more than one different solute.
A. Many Cellular Components and Functions Are Affected by Hypertonicity
Table 1 lists
200 cellular components that change with osmolality, including mRNAs and proteins identified by Northern analysis, ribonuclease protection, RT-PCR, or Western analysis. The number of constituents that are affected by osmolality is too great to consider them all individually in detail. Also, caution should be exercised in interpreting the findings of those studies in which osmolality is increased by 300 mosmol/kgH2O or more. That degree of hyperosmolality can lead to apoptosis (198, 257), so viability of cells is questionable under those circumstances. The different categories affected by tonicity (Table 1) include ion and water channels, transporters, heat shock proteins, other stress proteins, cellular and extracellular structural proteins, transcription factors and coactivators, kinases, phosphatases, other signaling proteins, enzymes, hormones, receptors, growth-related proteins, second messengers, and organic osmolytes. Determining the significance of any particular change requires additional information since it could represent an osmoprotective response, but it could also simply represent a dysfunction caused by hypertonicity. Many of these changes were discovered when an investigator added hypertonicity to the list of different stresses being applied to a component that initially was of interest for reasons unrelated to tonicity.
Screening methods have also been used to identify mRNAs and proteins affected by hypertonicity. The effect of hypertonicity on the abundance of cellular mRNAs in mIMCD3 cells was examined by using microarrays (294). When 200 mM mannitol or 100 mM NaCl is added for 6 h,
12% of
12,000 transcripts are downregulated by mannitol and NaCl, and
4% of transcripts are upregulated. As might be expected, the genes identified in this manner generally agree with those identified by other techniques (Table 1). The effect of hypertonicity on cellular proteins was examined by proteomic techniques. In one study (305) mIMCD3 cells were examined by two-dimensional gel electrophoresis (2D-GE) combined with MALDI-TOF/TOF mass spectrometry. Raising osmolality by 300 mosmol/kgH2O (NaCl added) increases abundance of some proteins and decreases others. Several molecular chaperones are induced, such as
B-crystallin, two Hsp70 isoforms, and the osmotic stress protein (Osp94). Also, some isoforms of the translation elongation factors eEF2 and eEF1A1 are induced. With TLAH cells (70), osmolality was increased by 300 mosmol/kgH2O for 35 h by adding NaCl. Proteins were examined by 2D-GE and MALDI-MS to determine changes in abundance; 25 proteins increase and 15 proteins decrease. Besides aldose reductase, many other metabolic enzymes like glutathione S-transferase, malate dehydrogenase, lactate dehydrogenase,
-enolase, glyceraldehyde-3-phosphate dehydrogenase, and triose-phosphate isomerase increase. The cytoskeleton proteins and cytoskeleton-associated proteins vimentin, cytokeratin, tropomyosin 4, and annexins I, II, and V also increase, whereas tubulin and tropomyosins 1, 2, and 3 decrease. Vimentin, cytokeratin, and tropomyosin remained elevated in TAHL-NaCl cell lines adapted to high NaCl, as determined by Western analysis. The heat shock proteins
-crystallin chain B, HSP70, and HSP90 and endoplasmic reticulum stress proteins like glucose-regulated proteins (GRP78, GRP94, and GRP96), calreticulin, and protein-disulfide isomerase decrease.
Since we do not have space to discuss all of these changes in detail, we will restrict ourselves to some that are known to be involved in osmoregulation and osmoprotection, including cellular structural proteins (see sect. IIIB), organic osmolytes (see sect. IV), stress proteins (see sect. X), the transcription factor TonEBP/OREBP (see sect. V), and the molecules that signal its activation (see sect. VII).
B. Cytoskeletal Rearrangement in Response to Hypertonicity
Hypertonicity induces rapid F-actin polymerization and remodeling of the actin cytoskeleton (35, 68, 322). However, the pattern of remodeling varies between cell types. In native rat MTAL cells (35) and glial cells (207), hypertonicity induces redistribution of F-actin from the cortical ring to a diffuse network of actin bundles, whereas in fibroblasts and HL60 cells the cortical actin ring becomes denser (68, 110). The assembly is induced by the combined action of small GTPases (Rac and Cdc42), an actin binding protein (cortactin), and a regulator of F-actin nucleation (the Arp2/3 complex), all of which form a complex with actin (68). Hypertonicity also causes rapid movement of myosin IIB to the cortical region, just beneath the plasma membrane, dependent in part on Rho kinase and p38 (230).
Integrins are a large family of heterodimeric transmembrane glycoproteins that attach cells to extracellular matrix proteins of the basement membrane. Integrin
1
1 is a major collagen-binding receptor that is highly expressed in renal tubular cells. Hypertonicity increases integrin
1 mRNA and protein in MDCK cells (269). Increased expression of this integrin is involved in hypertonicity-induced activation of the osmoprotective transcription factor TonEBP/OREBP (see sect. VH). Furthermore, inactivation of integrin-substratum interactions by using the integrin-binding peptide GRGDTP inhibits the hypertonicity-induced increase in glutamine transport that is mediated by phosphatidylinositol 3-kinase (PI3K) in skeletal muscle (180).
Phosphatidylinositol phosphate 5-kinase (PIP5K)
kinase (322) contributes to hypertonicity-induced cytoskeletal rearrangement. Its lipid kinase activity catalyzes phosphorylation of inositol 4-phosphate (Ins-4-P) on the D5 position of the inositol ring, producing inositol 4,5-bisphosphate (Ins-4,5-P2). PIP5KIs are predominantly cytosolic proteins that are recruited to plasma and organelle membranes through association with small GTPases such as Rho and Rac. Their lipid kinase activity is augmented by these associations and by integrin signaling. Hypertonicity increases Ins-4,5-P2. siRNA knock down of PIP5K
blocks both hypertonicity-induced increase of Ins-4,5-P2 and cytoskeletal rearrangement. PIP5KI
, being constitutively phosphorylated, is activated by Ser/Thr dephosphorylation. Hypertonicity reduces phosphorylation of PIP5K
and increases its lipid kinase activity. The phosphatase inhibitor calyculin A decreases basal PIP5KI
activity and blocks PIP5KI
activation by hypertonicity. Thus hypertonicity induces a phosphatase that activates PIP5KI
lipid kinase activity by dephosphorylating it, and the resultant increase of Ins-4,5-P2 contributes to the ensuing cytoskeletal rearrangement.
The exact role of the cytoskeleton in osmoregulation remains speculative (77). It could function as a cell volume sensor, it could offer mechanical protection against the deleterious effects of excessive shrinkage, and it could be involved in signal transmission from osmosensors (whatever they may be) to osmoeffectors. A role in signal transmission is supported by evidence that hypertonicity induces assembly of an actin-associated protein complex containing a scaffolding protein (OSM), a GTPase (Rac), and members of a mitogen-activated protein (MAP) kinase cascade (MEKK3, MKK3, and p38) (303). p38 is a MAP kinase that contributes to signaling hypertonicity-induced activation of the osmoprotective transcription factor TonEBP/OREBP (see sect. VE). Activation of p38 by hypertonicity is inhibited if any of the components of the OSM complex is knocked out (303). These findings associate the actin cytoskeleton with a signaling molecule (p38) that contributes to osmoprotective transcription. However, the exact role of the actin cytoskeleton in the system remains to be established. Are tonicity-dependent actin cytoskeletal changes necessary to activate p38 or is the mere presence of the cytoskeleton sufficient? Also, p38 can be activated in many ways and convey many different signals. Is the OSM system the only one involved in that activation of p38 that contributes to TonEBP/OREBP activity? Is it the most important one? An added complexity is that cytoskeletal remodeling depends on p38 in MTAL cells (35), implying that there could be positive feedback.
| IV. ORGANIC OSMOLYTES ACCUMULATE IN RESPONSE TO HYPERTONICITY |
|---|
|
|
|---|
Over a period of hours, however, compatible organic osmolytes accumulate in cells, gradually reducing the ionic strength back towards that in the normotonic state, while maintaining cell volume (Figs. 1 and 2). The benefit is that compatible organic osmolytes perturb less than do inorganic ions (325) because they stabilize proteins by interacting favorably with the protein backbone (282). The principal compatible organic osmolytes in renal medullary cells are sorbitol, glycine betaine (betaine), myo-inositol (inositol), taurine, and glycerophosphocholine (GPC) (8, 215). We previously reviewed (95) the evidence that hypertonicity increases the concentration of these organic osmolytes in renal cells both in vivo and in cell culture, and the mechanisms by which they are accumulated. To save space, we briefly summarize that earlier information and concentrate, instead, on more recent studies.
|
Sorbitol (95) is synthesized from glucose, catalyzed by aldose reductase (AR), and is metabolized to fructose, catalyzed by sorbitol dehydrogenase. Hypertonicity raises AR activity by increasing transcription of its gene, resulting in elevation of its mRNA and protein. Sorbitol dehydrogenase activity, on the other hand, is low in the renal cells that were examined and does not change significantly with tonicity. Furthermore, although hypertonicity increases AR mRNA in inner medullary collecting ducts (IMCDs) in vitro, it does not alter sorbitol dehydrogenase mRNA (106). Therefore, hypertonicity-induced increase of sorbitol apparently is entirely due to increased transcription of AR, leading to increased AR activity. TonEBP/OREBP is the transcription factor responsible for the hypertonicity-induced increase of AR. We consider osmotic regulation of TonEBP/OREBP in detail in section V.
AR is activated by other stimuli besides hypertonicity, and there are consequences of its activation besides osmoprotective synthesis of sorbitol. Although detailed consideration of other aspects of AR regulation and activity are beyond the scope of this review, it is worth noting that some of them are related to oxidative stress, which we have already seen is associated with osmotic stress (see sect. IID). For example, lipid aldehydes generated endogenously during the process of degradation of lipid peroxides contribute to cell damage associated with oxidative stress. 4-Hydroxynonenal (HNE), which is one of the most abundant and toxic lipid aldehydes, is efficiently detoxified by AR (278). AR mRNA and protein are elevated in areas of tissue destruction in inflamed arteries, colocalized with high levels of HNE, and in vitro exposure of mononuclear cells to HNE is sufficient to increase AR protein abundance (246). Similarly, HNE treatment of rat vascular smooth muscle cells induces AR mRNA (277). Oxidative stress can also activate AR posttranslationally. ROS, formed in the ischemic heart, activate AR by modifying its cysteine residues to sulfenic acids (133). Thus hypertonicity-induced increase of AR activity may affect and be affected by both the osmotic stress itself and the accompanying oxidative stress.
AR activation is not necessarily protective. AR activity is high in diabetes. This activity has been implicated in the microvascular complications of that disorder, and, as a result, there are major efforts to develop and use inhibitors of AR to minimize those complications (51). Furthermore, AR is activated during myocardial ischemic injury, and inhibitors of AR or sorbitol dehydrogenase protect rat hearts from ischemia-reperfusion injury (121). Adverse effects of increased traffic through the AR pathway have been attributed to elevation of the cytosolic NADH/NAD+ ratio.
Betaine (95) is synthesized from choline both in liver and kidney. However, unlike sorbitol, hypertonicity does not significantly affect the rate of its synthesis. Rather, hypertonicity increases the number of transporters that carry it into the cells from the extracellular fluid. Betaine/GABA transporter (BGT1) is the osmoregulated betaine transporter (323). It couples transport of betaine to that of Cl and Na. Gradients of those ions energize the transport, providing a driving force that can concentrate betaine from micromolar levels in the plasma to millimolar levels within the cells. Hypertonicity elevates the transcription of the BGT1 gene, followed by increase in its mRNA and transport activity (302). Thus tonicity regulates betaine transporter activity by affecting BGT1 transcription. TonEBP/OREBP (see sect. V) is the transcription factor responsible for this effect.
BGT1 is regulated not only by transcription, but also by plasma membrane insertion (139). The relatively small amount of BGT1 that exists under normotonic conditions is mainly in the cytoplasm. Following hypertonicity, BGT1 is seen mainly in basolateral plasma membranes. Addition of the microtubule inhibitor nocodazole causes the additional BGT1 to remain in the cytoplasm, even following hypertonicity (139), and elevation of calcium causes internalization of BGT1, possibly mediated by PKC (138). Also, EGFP-BGT1, expressed in fibroblasts that lack endogenous BGT1, moves into plasma membranes following hypertonicity (139). Then, transport is upregulated without change in total abundance of BGT1, only its location. Interestingly, membrane insertion of existing EGFP-BGT is delayed for 3 h after tonicity increases, consistent with the need to synthesize accessory proteins for membrane insertion.
Inositol (95) is synthesized by renal cells. Nevertheless, the inositol accumulated following hypertonicity is transported into the cells, not synthesized. In the absence of extracellular inositol, none is accumulated. Hypertonicity increases inositol transport by increasing the number of inositol transporters. Following its cloning, the transporter was named SMIT (sodium myo-inositol transporter)(160). SMIT couples transport of inositol to that of Na in a 1:2 ratio (109). The Na gradient energizes the transport, providing a driving force that can concentrate inositol from micromolar levels in the plasma to millimolar levels with the cells. Hypertonicity elevates the transcription of the SMIT gene, followed by increase in its mRNA and transport activity (324). Thus tonicity regulates inositol transporter activity by affecting SMIT transcription. TonEBP/OREBP (see sect. V) is the transcription factor responsible for this effect.
Taurine content is regulated by tonicity in renal cells in vivo and in cell culture. Thus taurine in rat renal inner medullas is approximately twice as great when the rats are infused with high NaCl solutions (presumed to increase medullary interstitial NaCl) than when they are infused with low NaCl solutions (215). Taurine in MDCK cells doubles when tonicity is increased by adding raffinose (302). Accumulation of taurine could be due to increased synthesis or increased transport into cells. Of the two possibilities, hypertonicity-induced accumulation of taurine appears to be mainly via transport, rather than synthesis. The original evidence for transport was that hypertonicity (raffinose added) doubles taurine uptake into MDCK cells across their basolateral membranes (301). The taurine transporter that is involved was cloned by expression in Xenopus oocytes and named TauT (300). It is Na and Cl dependent, and hypertonicity increases its mRNA abundance in MDCK cells. In HepG2 cells, hypertonicity (high NaCl or sucrose) increases transcription of the TauT gene, the abundance of TauT mRNA, and the abundance of TauT protein, resulting in increased taurine uptake (126). This hypertonicity-induced increase in TauT transcription is regulated by the transcription factor TonEBP/OREBP (see sect. V) via ORE/TonE sites in the 5'-flanking region of the TauT gene (126). A possible role for Ca2+/calmodulin (CaM) kinase II in hypertonicity-induced activation of TauT was inferred from the action of its inhibitor, KN-93, in Caco-2 cells (261).
Taurine synthesis is from cysteine, catalyzed by the sequential actions of cysteine dioxygenase (CDO), which gives rise to cysteinesulfinate, and cysteinesulfinate decarboxylase (CSD), which decarboxylates cysteinesulfinate to hypotaurine. The synthesis is mainly in liver, from which it is released to the general circulation, a major portion being taken up by the kidneys and excreted. Although the enzymes for taurine synthesis are present in renal medullas, hypertonicity increases TauT mRNA much more in rat renal medullas than it does CDO and CSD mRNAs (22).
Taurine has many other sites of action besides the renal medulla (261). It regulates heart rhythm, contractile function, blood pressure, platelet aggregation, neuronal excitability, body temperature, learning, motor behavior, food consumption, eye sight, sperm motility, cell proliferation and viability, energy metabolism, and bile acid synthesis. At least some of these effects have been attributed to its effects on membranes and on protein phosphorylation. Thus it is conceivable that taurine that is accumulated by cells in response to hypertonicity affects them in other ways besides serving as a compatible organic osmolyte.
GPC (95) is abundant in cells of renal inner medullas, and it accumulates in renal cells in tissue culture in response to either high NaCl or high urea. It is synthesized from phosphatidylcholine (PC), then is broken down into choline and
-glycerophosphate. The high NaCl-induced increase of GPC was studied both in isolated renal inner medullary collecting duct cells (14, 15) and in various cell cultures (156, 157, 332). In MDCK cells, high NaCl increases GPC synthesis and also reduces its degradation (156, 157). The synthesis from PC is catalyzed by a phospholipase activity that is increased by high NaCl. Similarly, phospholipase activity increases in inner medullas of rats when they are thirsted (331). Neuropathy target esterase (NTE) is a phospholipase B that catalyzes production of GPC from PC (333). In mIMCD3 cells (91), high NaCl raises NTE mRNA, beginning within 8 h, and NTE protein within 16 h. Di-isopropyl fluorophosphate (DFP), which inhibits NTE esterase activity, reduces the GPC accumulation, as does an siRNA that specifically decreases NTE protein abundance. In vivo, NTE protein expression is higher in the renal inner medulla than in the cortex. The lower renal inner medullary interstitial NaCl concentration that occurs chronically in ClCK1 –/– mice and acutely in normal mice given furosemide is associated with lower NTE mRNA and protein. Furthermore, the 20-h half-life of NTE mRNA is unaffected by high NaCl, but knock down of the transcription factor TonEBP/OREBP (see sect. V) by a specific siRNA inhibits the high NaCl-induced increase of NTE mRNA, showing that high NaCl elevates NTE mRNA through TONEBP/OREBP-mediated increase in its transcription. Transcriptional activity of TonEBP/OREBP requires binding to a specific DNA element, ORE/TonE (83, 287), whose consensus sequence is NGGAAAWDHMC(N) (83). Between bp –900 and –1 in the 5'-flanking region of NTE there are four DNA elements that fit this consensus. Thus high NaCl increases transcription of NTE, mediated by TonEBP/OREBP, and the resultant increase of NTE activity contributes to increased production and accumulation of GPC in mammalian renal cells in tissue culture and in vivo.
In addition, high NaCl also reduces degradation of GPC by decreasing the activity of GPC-choline phosphodiesterase (GPC-PDE), the enzyme that degrades GPC to choline and
-glycerophosphate (15, 157, 332), and the reduced degradation of GPC contributes to its increased abundance. The GPD-PDE that is involved is Gdpd5 (unpublished observations), which is a mouse homolog of bacterial glycerol diesterases. High NaCl reduces Gdpd5 mRNA abundance. In mIMCD3 cells at 300 mosmol/kgH2O, reducing Gdpd5 abundance with a specific siRNA increases GPC. High NaCl reduces Gdpd5 mRNA by destabilizing it. In short, high NaCl inhibits degradation of GPC by decreasing the abundance of Gdpd5, a GPC-PDE that catalyzes breakdown of GPC.
One compatible organic osmolyte can substitute for another in protecting cells from hypertonicity. For example, addition of betaine to the medium, and its resultant uptake by the cells, largely replaces the decrease in sorbitol caused by AR inhibitors and restores the viability of the cells (205). Specific reduction of one organic osmolyte generally causes a compensating increase of the others (204). This is readily understandable, since the same transcription factor, TonEBP/OREBP (see sect. V), regulates betaine and inositol transporters and the enzymes that catalyze synthesis of sorbitol and GPC. GPC is exceptional in that its degradation is also regulated (see above). It is of interest that the degradation of GPC is also affected by the abundance of the other osmolytes. Thus increasing the level of betaine and inositol in MDCK cells reduces GPC by stimulating its degradation (158). Other factors have also been reported to affect GPC. In MDCK cells, caffeine increases GPC through ryanodine-inhibitable increase of cellular calcium and activation of phospholipase A2 (140).
| V. ACTIVATION OF THE TRANSCRIPTION FACTOR TonEBP/OREBP BY HYPERTONICITY |
|---|
|
|
|---|
B (NF
B) and nuclear factor of activated T-cells (NFAT). Each TonEBP/OREBP target gene contains in its regulatory regions at least one DNA consensus motif called an osmotic response element (ORE) (83–85, 144, 346) or tonicity-responsive enhancer (TonE) (126, 187, 200, 245, 287). As described in section IV, TonEBP/OREBP transactivates several genes that are responsible for tonicity-dependent accumulation of organic osmolytes including sorbitol, glycine betaine, inositol, taurine, and GPC (Fig. 2) (30, 91). TonEBP/OREBP also transactivates genes coding for heat shock protein 70 (HSP70) (317), vasopressin-activated urea transporters, UT-A1 (216) and possibly UT-A2 (161), as well as aquaporin-2 (AQP2) (113, 114). As described in section XC, HSP70 is a protein chaperone and is osmoprotective against high NaCl and high urea. UT-A1 and UT-A2 recycle urea, producing high concentrations in the renal inner medulla (255), which contributes to the corticomedullary osmotic gradient necessary for concentrating urine. AQP2 increases water permeability of the renal collecting duct, which increases urine concentration. Mice that overexpress a dominant negative form of TonEBP/OREBP in renal collecting tubule cells have impaired urinary concentrating ability, probably due to reduced expression of AQP-2 and UT-A (161).
TonEBP/OREBP knock-out or dominant negative expression in vivo has major renal and extrarenal consequences. Disruption of both TonEBP/OREBP alleles in mice results in embryonic lethality (101, 176). The few surviving homozygous TonEBP/OREBP null-mice have profound, progressive atrophy of the kidney medulla, coupled with impaired activation of the TonEBP/OREBP target genes AR, BGT1, and SMIT (176). Heterozygous mice exhibit defects in adaptive immunity, evidenced by lymphoid hypocellularity and impaired antigen-specific antibody response (101). Additionally, in culture, TonEBP/OREBP –/+ lymphocytes proliferate more slowly in hypertonic, but not normotonic media (101). Mice that overexpress a dominant negative form of TonEBP/OREBP in the lens develop nuclear cataracts soon after birth (308). Developing fiber cells in these lenses exhibit a physiological state similar to that produced by hypertonicity, including DNA strand breaks, activation of p53, and induction of checkpoint kinase (308).
In the following sections we analyze how the transcriptional activity of TonEBP/OREBP is controlled and the complex network by which hypertonicity signals the activation.
A. TonEBP/OREBP mRNA and Protein Abundance
Hypertonicity increases TonEBP/OREBP mRNA in MDCK (316), HeLa (143), and mouse inner medullary collecting duct (mIMCD3) cells (39). The increase is transient, peaking at 4–12 h depending on cell type. TonEBP/OREBP mRNA rises because of a transient increase in its stability, mediated by elements within its 5'-UTR (39). In mIMCD3 cells, TonEBP/OREBP mRNA is stabilized for 6 h after addition of NaCl, as measured by determining the rate of mRNA decrease after inhibiting transcription with actinomycin D (39). Stabilization by high NaCl is sufficient to explain the increase of TonEBP/OREBP mRNA without any change in transcription (39). Following increased mRNA abundance, TonEBP/OREBP protein increases. In combination with tonicity-dependent nuclear localization, discussed below, the elevated nuclear abundance of TonEBP/OREBP increases TonEBP/OREBP protein binding to OREs. The hypertonicity-induced increase in TonEBP/OREBP protein abundance and synthesis rate is approximately equal to the increase in mRNA abundance (316). Hypertonicity does not affect the rate of TonEBP/OREBP protein degradation (316).
B. Dimerization of TonEBP/OREBP
TonEBP/OREBP is a constitutive dimer, and dimerization is required for several aspects of its hypertonicity-dependent activation. The NH2 terminus of TonEBP/OREBP contains two dimerization domains (177). TonEBP/OREBP forms stable dimers (or higher order oligomers) in solution (178) independent of tonicity (177). Dimerization of TonEBP/OREBP is required for DNA binding (177). In binding to DNA, TonEBP/OREBP forms a homodimer that completely encircles its DNA target (284). Constitutive dimerization of TonEBP/OREBP is also required for its "proper" phosphorylation when NaCl is high (166). Dimerization also is required for TonEBP/OREBP transactivation of downstream genes (177).
C. Phosphorylation of TonEBP/OREBP
TonEBP/OREBP becomes phosphorylated on serine and tyrosine residues within 30 min of addition of NaCl (65). Possible phosphorylation sites in TonEBP/OREBP isoform "c" (1,531 amino acids) are its 216 serines, 15 tyrosines, and 111 threonines. The exact role of this tonicity-dependent TonEBP/OREBP phosphorylation remains unclear because there has been no direct identification of the amino acids that are phosphorylated. However, single and multisite phosphorylation are known to regulate other transcription factors at multiple levels, including nuclear localization, DNA binding, and transactivation.
Serine/threonine phosphorylation is involved in TonEBP/OREBP binding to its DNA element. The binding, as measured using nuclear extracts and electromobility shift assays with an ORE probe, is eliminated by treatment with calf intestinal phosphatase (CIP), a general phosphatase, and PP1, a serine/threonine phosphatase, but not by treatment with PTPase, a phosphotyrosine phosphatase (2). Nevertheless, binding of immunoprecipitated TonEBP/OREBP is not prevented by CIP, indicating that its DNA binding does not require the actual phosphorylation of TonEBP/OREBP itself (65).
Phosphorylation generally regulates transactivation 1) by recruiting coactivator proteins that contact the general transcriptional machinery, 2) as a requirement for maximal transcriptional activity, or 3) as a requirement for the formation of stable homodimer-DNA complexes. High NaCl increases phosphorylation of TonEBP/OREBP within amino acids 548–1531, as indicated by phosphatase-inhibitable decrease of migration of that fragment in gels (86). TonEBP/OREBP amino acids 872–1271 contain a transactivation domain (TAD) that is activated by high NaCl (86). Serine/threonine kinase inhibitors reduce this high NaCl-dependent transactivational activity of TonEBP/OREBP (86), which could be due either to a direct effect on the phosphorylation state of TonEBP/OREBP, itself, or it could be mediated by a signaling partner or a transcriptional cofactor. An indirect effect is plausible, since no net tonicity-dependent change in phosphorylation of the TonEBP/OREBP TAD was detected (165), while a direct effect is supported by site-directed mutation of putative phosphorylation sites (124), as discussed below.
D. TonEBP/OREBP Nuclear Localization
To bind to their cognate DNA elements, transcription factors must enter the nucleus. Proteins the size of TonEBP/OREBP are too large (>50 kDa) to diffuse freely in and out through nuclear pores. Such large molecules require a nuclear localization sequence (NLS) to mediate binding to importin. The importin-cargo complex enters the nucleus by active transport through nuclear pores. There is a similar requirement for a nuclear export sequence (NES) and binding to exportin for large proteins to exit from the nucleus. The distribution of TonEBP/OREBP in the cell is tonicity dependent. In MDCK cells, HeLa cells and Jurkat cells at normotonicity (
300 mosmol/kgH2O), TonEBP/OREBP is distributed between cytoplasm and nucleus, but at 450–500 mosmol/kgH2O, most of the TonEBP/OREBP is in the nucleus (143, 177, 201, 296). At 135–250 mosmol/kgH2O in MDCK or HeLa cells, TonEBP/OREBP is mostly in the cytoplasm (296, 316), correlated with a decrease in its activity (86, 316). There is indirect evidence that TonEBP/OREBP may interact with a cytoplasmic binding protein. MG-132, a 26S proteasome inhibitor, blocks tonicity-dependent nuclear localization of TonEBP/OREBP in MDCK cells (318). This calls to mind the regulation of NF
B (63). In the cytoplasm, I
B binds to NF
B blocking NF
B nuclear import. During activation of NF
B, I
B becomes phosphorylated, then degraded in the proteasome, releasing NF
B to move into the nucleus. Although a similar system could exist for regulation of TonEBP/OREBP, no TonEBP/OREBP cytoplasmic binding protein analogous to I
B has been identified. MG-132 also blocks NaCl-induced nuclear localization in HepG2 cells, but not in COS-7 cells, indicating that the effect is cell type dependent (338). Cyclosporin A has a similar effect in MDCK cells (267), consistent with calcineurin dependence of nuclear localization.
Like the native protein, a recombinant TonEBP/OREBP expressing NH2-terminal amino acids 1–547 localizes to the nucleus when NaCl is increased, indicating that the COOH terminus is not required for translocation (338). The NH2 terminus of TonEBP/OREBP contains a series of amino acids (amino acids 199–216 of isoform "c") that fit with a consensus bipartite NLS (143, 179) in that there are two clusters of basic amino acids in tandem. However, the NLS of TonEBP/OREBP is actually monopartite. Only the first cluster of basic amino acids is necessary for its nuclear import (296). The NH2 terminus of TonEBP/OREBP additionally contains two motifs that function in nuclear export (296). The NES at amino acids 8–15 (isoform "c") is a CRM1-responsive protein domain. Accordingly, leptomycin B, an inhibitor of the exportin CRM1, blocks nucleocytoplasmic shuttling of TonEBP/OREBP at 300 mosmol/kgH2O (296). Disruption of this NES increases TonEBP/OREBP in the nucleus at 300 mosmol/kgH2O, but hypotonicity can still cause nuclear export of this mutant, indicating that this NES is not responsible for nuclear export under hypotonic conditions (296). An auxiliary export domain (AED) is found at amino acids 132–156 (isoform "c") and is functional in nuclear export of TonEBP/OREBP under both normotonic and hypotonic conditions. Deletion of the AED causes nuclear retention of TonEBP/OREBP regardless of tonicity (296). However, the AED does not affect subcellular localization when fused to green fluorescent protein, indicating it does not function independently (296).
Activation of ataxia telangiectasia-mutated (ATM) kinase by high NaCl contributes to nuclear translocation of TonEBP/OREBP (338). In cells lacking functional ATM, nuclear translocation of native TonEBP/OREBP and of a recombinant NH2-terminal construct (amino acids 1–547, which contains the NLS as well as the NES and AED) is reduced, and reconstituting the kinase in these cells restores the translocation (338). Thus ATM contributes to high NaCl-induced nuclear translocation of TonEBP/OREBP through interaction at sites within amino acids 1–547. The effect of ATM could be direct or it could be mediated by another factor in the signaling pathway that activates TonEBP/OREBP. Since transactivation domains can affect nucleocytoplasmic shuttling (238), it is possible that that ATM also modulates translocation via sites in TonEBP/OREBP-548–1531, which contains a tonicity-dependent transactivation domain as well as putative ATM consensus phosphorylation sites (86, 124).
Other kinases such as Fyn and PI-3K class 1A, which also are known to stimulate TonEBP/OREBP-dependent transcription, do not affect NaCl-dependent nuclear localization of TonEBP/OREBP (123, 142), but, as discussed below, each of these kinases does target the hypertonicity-sensitive TonEBP/OREBP transactivation domain.
E. TonEBP/OREBP Transactivational Activity
In addition to motifs that regulate nuclear localization and DNA binding, transcription factors also contain transactivation domains (TADs), which interact with and recruit a specific assemblage of proteins that forms an enhanceosome, and with the basal transcriptional complex which includes RNA polymerase. Transactivation of target genes by transcription factors is regulated by DNA binding, by association with cofactors, by heterologous transcription factors, and by posttranslational modifications. Transactivation of target genes by TonEBP/OREBP varies directly with extracellular NaCl concentration (86). TonEBP/OREBP has a high NaCl-dependent TAD within amino acids 1039–1249 (isoform "c") (86, 165). Additional TonEBP/OREBP TADs are located within amino acids 1–76 and 1363–1476, but these are not sensitive to tonicity (165). TonEBP/OREBP also has two modulation domains (amino acids 618–820 and 889–955) that potentiate the activity of the TADs but are not independently active in stimulating transcription. One of the modulation domains is responsive to tonicity (amino acids 618–820), and all of the domains act synergistically to transactivate target genes in response to hypertonicity (165).
Many regulators contribute to tonicity-dependent increase in TonEBP/OREBP transactivation of its target genes. Each is necessary for full TonEBP/OREBP activity, but no one of them is sufficient.
cAMP-dependent kinase (PKA) is a serine-threonine kinase that regulates many cellular processes, most commonly in response to increased intracellular cAMP. However, high NaCl activates PKAc without increasing intracellular cAMP (82). Chemical inhibition of PKAc or transfection with a PKAc dominant negative construct reduces hypertonicity-induced TonEBP/OREBP transcriptional activity, transactivation activity, and resultant accumulation of AR and BGT1 mRNAs (82). Conversely, overexpression of wild-type PKAc increases TonEBP/OREBP transactivation activity (82). Also, PKAc and TonEBP/OREBP reciprocally coimmunoprecipitate each other, indicative of physical association of the proteins (82).
p38 is a mitogen-activated protein kinase (MAPK). Hypertonicity activates p38 (111), contributing to increased TonEBP/OREBP activity (142, 210, 265). Thus inhibition of p38 by chemicals (142, 210, 265) or by transfection of a dominant negative construct (142) reduces 1) hypertonicity-induced activation of TonEBP/OREBP (142, 210), 2) the increase in its transactivational activity (142), and 3) the resultant increase of mRNA abundance of its transcriptional targets, AR and BGT1 (142, 265). The pathway for activation of p38 MAPK by osmotic stress is cell type-specific and can involve different combinations of the MAPK kinases, MKK3 and MKK6, as well as autoactivation (134). Hypertonicity causes MEKK3 activation (303) by autophosphorylation (89) and also activates MKK3 and MKK6 (62, 194). p38, itself, becomes autoactivated in response to hypertonicity, promoted by association with transforming growth factor-
-activated protein kinase 1-binding protein (TAB1) (134). Furthermore, some of the cellular p38 is contained in a multienzyme complex, attached to the scaffolding protein, OSM. Hypertonicity activates p38 in the complex, associated with hypertonicity-induced cytoskeletal rearrangement (see sect. IIIB) (303). The complex attached to OSM also includes actin, a GTPase (Rac), MEKK3, and MKK3 (303). Overexpression of dominant negative Rac or siRNA knockdown of OSM or MEKK3 reduces the hypertonicity-induced phosphorylation and activation of p38 (303). TonEBP/OREBP activity was not addressed in these studies. Although linkage of TonEBP/OREBP activity to Rac and OSM remains conjectural, activation of MEKK3 does stimulate TonEBP/OREBP (226). Overexpression of recombinant active MEKK3 in MDCK cells increases TonEBP/OREBP transcriptional activity and the mRNA abundance of the TonEBP/OREBP target gene BGT1 (226). Additionally, siRNA-mediated depletion of native MEKK3 in HEK293 cells reduces BGT1 mRNA (226). In other experiments overexpression of constitutively active MKK3 or MKK6 did not affect TonEBP/OREBP expression or function (177), and dominant negative MKK3 did not affect TonEBP/OREBP transcriptional activity (154). Thus there is some uncertainty as to the composition of the MAPK kinase signaling module that mediates hypertonicity-dependent TonEBP/OREBP activation via p38. As is true for all other factors known to regulate TonEBP/OREBP activity, inhibition of p38 MAPK only partially reduces the hypertonicity-induced increase of TonEBP/OREBP activity (142). Activation of p38 by hypertonicity also apparently depends on the accompanying hypertonicity-induced increase of ROS (see sect. IID). Thus the antioxidant NAC prevents high NaCl-induced phosphorylation of p38 in mIMCD-K2 cells (328).
Fyn is a member of the SRC family of nonreceptor, cytoplasmic protein tyrosine kinases. Regulation of SRC kinases involves both phosphorylation and dephosphorylation. Inactive SRC kinases are constitutively phosphorylated at a COOH-terminal tyrosine. Upon dephosphorylation of the COOH-terminal tyrosine, SRC kinases undergo a conformational change that exposes the catalytic domain. Autophosphorylation of a regulatory tyrosine within the catalytic domain is required for full activation. Various SRC tyrosine kinases are involved in diverse cellular responses to hypertonicity that include RVI, reorganization of cytoskeletal proteins, and mechanotransduction (53). Several studies have investigated the roles of SRC kinases in TonEBP/OREBP activation. Herbimycin, an SRC kinase inhibitor, decreases high NaCl-induced TonEBP/OREBP-dependent transcription, transactivational activity, and upregulation of TonEBP/OREBP target gene mRNAs (86). Among SRC family kinases, Fyn and Yes are activated by high NaCl (136, 243). Several lines of evidence show that Fyn stimulates TonEBP/OREBP transactivation of its target genes (142). Fyn null cells have lower TonEBP/OREBP transcriptional activity than wild-type cells and show reduced hypertonic induction of the TonEBP/OREBP target gene AR (142). Furthermore, both pharmacological and dominant negative-based inhibition of Fyn decrease TonEBP/OREBP-dependent transcription (142). Fyn is activated by the cell shrinkage that results from hypertonicity, rather than an increase in ionic strength (136). SRC kinase activation of cell surface receptors is of interest as a possible osmotic sensor upstream of TonEBP/OREBP. Hypertonicity activates cell surface epidermal growth factor receptor (EGFR) tyrosine kinase (253), apparently dependent on the SRC kinase Yes. Hyperosmotic stress signals an oxidative stress-dependent Yes kinase activation, followed by rapid Yes/EGFR association (243). However, no link between Yes or EGFR and TonEBP/OREBP activation has, as yet, been established. Fyn, which does contribute to activation of TonEBP/OREBP (142), does not associate with EGFR during hyperosmotic stress (243).
ATM is a serine/threonine protein kinase that is activated by intermolecular autophosphorylation on serine-1981 following DNA breakage. Several lines of evidence indicate that the activating signal is change in higher-order chromatin structure caused by breakage of the phosphodiester backbone, rather than the DNA breaks per se. Once activated, ATM phosphorylates several proteins including p53, NBS1, BRCA1, and SMC1, that initiate cell cycle arrest, induce repair of DNA double-strand breaks, and enhance cell survival. As previously discussed in section IIC, high NaCl causes DNA strand breaks in cell culture and in vivo. High NaCl also deranges chromatin structure (152), presumably associated with the DNA breaks. That led to the hypothesis that high NaCl might activate ATM and stimulate TonEBP/OREBP transactivation of its target genes (124). Pharmacological inhibition of ATM reduces high NaCl-dependent increase in TonEBP/OREBP-mediated transcription and the resultant increase in BGT1 mRNA abundance (124). Additionally, reconstitution of AT cells, which do not express function ATM, with wild-type ATM increases high NaCl-dependent TonEBP/OREBP transcriptional and transactivational activity. In contrast, reconstitution with inactive ATM that is mutated at S1981 has no effect. ATM and TonEBP/OREBP reciprocally coimmunoprecipitate, and ATM is present along with TonEBP/OREBP at ORE sites in vitro, indicative of physical association (124). ATM kinase activity is necessary for full activation of TonEBP/OREBP, but it is not clear whether ATM directly phosphorylates TonEBP/OREBP or some intermediate protein. TonEBP/OREBP has three serines in putative ATM consensus phosphorylation motifs within COOH-terminal amino acids 548–1531 (124). As discussed above, high NaCl increases the activity of the TAD that resides in this region and causes a posttranslational modification consistent with increased phosphorylation (86). Mutation of serine-1247 in TonEBP/OREBP reduces its transcriptional activity in HEK293 cells, but does not in AT cells (which lack functional ATM), unless the AT cells are reconstituted with functional ATM (124). These results support ATM-mediated phosphorylation of TonEBP/OREBP. However, the conclusion remains questionable as long as there is no direct evidence linking any change in phosphorylation of a specific amino acid in TonEBP/OREBP accompanying hypertonicity-induced increase of TonEBP/OREBP transactivational activity (165). Multiple agents activate ATM without activating TonEBP/OREBP (124), including urea, UV, IR, and low NaCl (11, 124). Furthermore, high urea (293) or low NaCl (86) actually decrease TonEBP/OREBP activation. Thus ATM is necessary for full TonEBP/OREBP activity, but activation of ATM, alone, is not sufficient to induce it.
PI3Ks are intracellular lipid kinases that phosphorylate the 3'-hydroxyl moiety of phosphatidylinositol and phosphoinositides. Many different signaling proteins are activated by binding to the lipid products of PI3Ks. Three classes of PI3Ks are recognized, based on sequence homology and lipid substrate specificity. Class IA (PI3K-IA), which generates phosphatidylinositol 3,4,5-trisphosphate (PIP3) from phosphatidylinositol 4,5-bisphosphate (PIP2), is involved in activation of TonEBP/OREBP (78). Elevation of NaCl results in phosphorylation of tyrosine-508 of the PI3K-IA p85 regulatory subunit and activates the p110
catalytic subunit, resulting in increased production of PIP3 (123). Knockdown of PI3K-IA by a dominant negative mutant of p85
reduces high NaCl-induced transcriptional and transactivational activity of TonEBP/OREBP, and knockdown by an siRNA against p110
reduces TonEBP/OREBP-driven transcription (123). These effects of PI3K-IA are mediated by ATM (123). The interaction between PI3K-IA and ATM is discussed below. Neither the dominant negative- or siRNA-mediated knockdown of PI3K-IA affects high NaCl-induced increase in TonEBP/OREBP protein abundance (123). As with other kinases, inhibition of PI3K-IA does not completely eliminate high NaCl-induced activation TonEBP/OREBP. It is necessary for full activation of TonEBP/OREBP, but is not in itself sufficient.
F. Interactions Between Proteins That Regulate TonEBP/OREBP
To date, we are certain of only two regulatory proteins whose actions on TonEBP/OREBP are interrelated. The effect of PI3K-IA on TonEBP/OREBP is mediated through ATM, since ATM activation, as measured by S1981 phosphorylation in response to high NaCl, is abrogated by expression of dominant negative acting PI3K-1A (123). Additionally, the effects of inhibiting PI3K-IA and ATM are equivalent and are not additive. A PI3K-1A interaction with p38 interaction also was investigated but yielded negative results, indicating that high NaCl activation of p38 is not mediated by PI3K-1A (123). Similarly, p38 and Fyn act independently, since inhibition or deficiency of one does not affect activation of the other, and their effects on TonEBP/OREBP are additive (142). Presumably, other interactions among proteins that regulate TonEBP/OREBP remain to be determined.
RNA helicase A (RHA) is a nuclear RNA-associated protein that unwinds DNA and RNA. RHA inhibits activation of TonEBP/OREBP, independent of its helicase activity (59). It associates with TonEBP/OREBP in vitro and in vivo as demonstrated by reciprocal pull down. The NH2 and COOH termini of RHA bind TonEBP/OREBP at the NH2-terminal E'F loop, which is the region of TonEBP/OREBP that forms a secondary dimer interface involved in encirclement of DNA. The association of RHA with TonEBP/OREBP is independent of DNA binding and TonEBP/OREBP dimerization, and it does not involve any of the TonEBP/OREBP transactivation domains. Hypertonicity reduces binding of RHA to TonEBP/OREBP, and overexpression of RHA partially reduces TonEBP/OREBP transcriptional activity. Thus hypertonicity apparently decreases RHA-TonEBP/OREBP binding, which releases TonEBP/OREBP from inhibition by RHA.
H. DNA Breaks, ROS, and Cytoskeletal Perturbations as Regulators of TonEBP/OREBP
DNA breaks increase in response to high NaCl both in cell culture and in vivo (see sect. IIC). The DNA breaks, by causing change in higher-order chromatin structure, activate PI3K-IA and ATM, which initiates cell cycle arrest and chromosomal repair and promotes cell survival. PI3K-IA and ATM also stimulate TonEBP/OREBP (see sect. VE), thus linking TonEBP/OREBP activation to high NaCl-induced DNA breaks.
ROS convey signals from cytokines, hormones, and ions to downstream effectors, including transcription factors (87). High NaCl increases intracellular ROS in renal cells in culture (328, 337, 348) and in the renal medulla (337). Renal papillary ROS come from mitochondria (352), as do the NaCl-induced ROS that contribute to activation of TonEBP/OREBP (347, 348). Reduction of ROS by antioxidants decreases high NaCl-induced TonEBP/OREBP activation. The antioxidants reduce high NaCl-induced TonEBP/OREBP transcriptional activity and the mRNA abundance of its target gene BGT1. Mitochondrial ROS stimulate NaCl-induced TonEBP/OREBP transactivating activity, but do not affect its nuclear localization (347, 348). Similar to ATM activation, the increased ROS are necessary for full high NaCl-induced activation of TonEBP/OREBP, but elevation of ROS, by itself, does not activate TonEBP/OREBP (124, 348).
Actin cytoskeleton reorganization, which is induced by hypertonicity (see sect. IIIB), contributes to activation of TonEBP/OREBP. Hypertonicity induces the phosphorylation and/or redistribution of myosin, mediated by activation of the Rho and Rho/ROK pathway (69). Hypertonicity also stimulates the Rho family GTPase, Rac (68) which associates with OSM, actin, MEKK3, and MKK3 to activate p38 (303), providing a possible link to activation of TonEBP/OREBP. Another possible connection between cytoskeletal rearrangement and TonEBP/OREBP activation is through the increased mitochondrial ROS production that occurs in response to actin cytoskeleton remodeling, mediated by integrins and activation of Rac and RhoA (4, 314).
Integrins are extracellular matrix receptors that link the internal actin cytoskeleton to the extracellular matrix and thus mediate mechanotransduction. Hypertonicity induces mRNA and protein expression of the cell adhesion molecule
1-integrin in MDCK cells (see sect. IIIB).
1
1-Integrin is a collagen-binding receptor that is highly expressed in the kidney. In response to dehydration, inner medullas of
1
1-integrin null mice have reduced accumulation of the organic osmolytes, sorbitol, betaine, and inositol as well as mRNA abundance of the inositol transporter SMIT and the sorbitol synthesizing enzyme AR (202). Furthermore, primary cultures of inner medullary collecting duct cells that lack
1
1-integrin have impaired hypertonic induction of TonEBP/OREBP protein (202). Thus
1
1-integrin is involved in hypertonicity-induced activation of TonEBP/OREBP.
| VI. ADAPTATION TO CHRONIC HYPERTONICITY |
|---|
|
|
|---|
| VII. SENSORS AND SIGNALERS OF HYPERTONICITY |
|---|
|
|
|---|
| VIII. HIGH UREA |
|---|
|
|
|---|
A. Cellular Components and Functions Affected by High Urea
High urea increases expression and activity of transcription factors, enzymes (particularly kinases), and other signaling proteins (Table 1), but it also decreases the expression and activity of others, an effect in some cases opposite to that of hypertonicity (Table 1).
Many aspects of high urea-induced signaling are reminiscent of those observed in response to peptide mitogens (292). Multiple receptor tyrosine kinase effector pathways are activated by urea in renal medullary cells, including 1) Shc/Grb2/SOS/Ras/Raf/MAPK/(MEK)/ERK/Elk-1 (52, 341), 2) PLC-
/Ins(1,4,5)P3 (52), and 3) PI3K/Akt/p70 S6 kinase (341) pathways, as illustrated in Figure 1 of Reference 292. An initial step is epidermal growth factor (EGF) receptor transactivation through an autocrine mechanism. Enzymic cleavage by a metalloproteinase results in ectodomain shedding that liberates heparin-binding EGF as a soluble peptide agonist (343). This catalyzes the observed tyrosine phosphorylation of the adapter protein Shc (341). Shc recruits a second adapter protein, Grb2 (341). The Grb2 interaction partner and guanine nucleotide exchange factor, SOS, also becomes tyrosine phosphorylated, associated with rapid activation of the small G protein Ras (291). Subsequent events are activation of Raf (292), MEK (329), ERK (52), and the transcription factor Elk-1 (52), resulting in increased transcription of immediate-early genes, including Egr-1 (54, 56). Interestingly, mitogenesis in response to urea treatment was not observed in the mIMCD3 cells that were used for most of these studies (292). Nevertheless, indices of proliferation in response to urea treatment have been observed in other urea-responsive renal epithelial cell lines such as MDCK, LLC-PK1, and NRK-52E (55, 56, 174). Thus this urea-induced mitogenic signaling pathway has been thoroughly documented, but its physiological role remains unclear.
Similarly, in microarray studies of mIMCD3 cells, the response to high urea was closer to that of the growth factor EGF than to hypertonic stress (294). High urea decreases abundance of many mRNAs. Approximately 6% of the
12,000 transcripts were downregulated, whereas only
0.5–0.8% were upregulated. mRNA expression of only 21 genes was upregulated threefold or more (294). Prominent among them was a large increase (27-fold) of activating transcription factor 3 (ATF3) mRNA that was accompanied by a large increase in its protein abundance (294). ATF3 is a member of the mammalian activation transcription factor/cAMP responsive element-binding (CREB) protein family of transcription factors. It is induced by a large number of other stresses, including endoplasmic reticulum stress and amino acid starvation (130), and it contributes to the accompanying induction of Gadd153. High urea also increases expression of Gadd153 (340), as well as of the cytoprotective enzyme heme oxygenase-1 (38, 290), both dependent on urea-induced oxidative stress (290, 340) and on Ras (291).
B. Counteraction of Urea Effects by GPC
Urea is sufficiently high in renal medullas to perturb macromolecules (see introduction to sect. VIII), yet the cells survive and function. Methylamine organic osmolytes protect cells from high urea by counteracting its effects (325). Animal tissues containing high levels of urea generally also contain high levels of methylamines (325). Most studies of counteraction have involved examining the effects of combinations of urea and methylamines on structure and function of macromolecules in vitro. Optimal counteraction generally occurs when the molar ratio of urea to methylamine is 2:1, which is the ratio in many tissues (325). The methylamine organic osmolytes in renal medullas are betaine and GPC (95). Both can counteract the effects of urea (31, 33). Nevertheless, although high urea increases GPC in tissue culture, it reduces betaine (95). Also, the level of GPC in renal medullas correlates with the level of urea, but betaine does not (95). Thus, although both betaine and GPC effectively counteract the effects of urea, renal cells apparently prefer GPC (Fig. 3). The reason for this preference is speculative. In so far as accumulation of GPC in response to high urea involves reducing its degradation (157, 332), the energetic cost may be less than that of transporting betaine into the cells against a high concentration gradient. Alternatively, GPC might have some unrecognized additional beneficial effect that betaine lacks.
|
C. Mutually Protective Effects of High Urea and High NaCl
A high concentration of either NaCl or urea can perturb and even kill cells (see sect. IIB). However, their effects are not additive, and one can protect against the other. There are numerous examples of such mutual protection, as follows.
Adding either NaCl or urea to a total osmolality of 700 mosmol/kgH2O kills >90% of the mIMCD3 cells by apoptosis within 24 h. However, the cells are more tolerant of NaCl and urea when they are added in combination than when they are added alone; 40% of cells survive equiosmolal addition of NaCl plus urea to 900 mosmol/kgH2O. Added NaCl causes a rapid decrease in RNA and protein synthesis, which parallels the later apoptosis, whereas lethal doses of urea have little or no effect on these biosynthetic processes. The combination of NaCl plus urea results in an increase of Hsp70 expression that correlates with survival. Thus high NaCl and urea activate independent and complementary cellular programs that in combination enhance osmotolerance of mIMCD3 cells (257). In somewhat similar studies (198), elevating NaCl or urea reduced the number of proliferating mIMCD3 cells by slowing the transit through the S phase, by cell cycle delay in the G2M and G1, and by inducing apoptotic cell death. The effects were not additive in that concentrations of NaCl and urea elevated in combination had no greater effect than the individual concentrations added alone.
High urea protects against high NaCl and vice versa, but different mechanisms seem to be involved. Pretreatment with urea protects mIMCD3 cells from high NaCl-induced apoptosis (339). The signals induced by high urea are similar to those induced by peptide mitogens, and urea is mitogenic, at least in some cell lines (see sect. VIIIA). Importantly, the peptide mitogens EGF and insulin-like growth factor have protective effects comparable to that of urea. Also, urea activates EGF receptor kinase (see sect. VA), and preventing this action negates the protective effect of urea against high NaCl (343). Thus the protective effect of high urea against high NaCl apparently is related to activation of the EGF receptor by urea.
Pretreatment with NaCl protects MDCK cells from high urea-induced apoptosis. The experiments were on cells conditioned by increasing osmolality from 290 to 600 mosmol/kgH2O by adding NaCl. Upon subsequent exposure to an additional 600 mM urea in the medium for 24 h, 90% of the osmotically conditioned cells but only 15% of nonconditioned cells survive. The effect was attributed to protection against urea by HSP72, which is strongly induced by high NaCl, but much less so by high urea (219). In the absence of hypertonic stress, survival rates of MDCK cells exposed for 24 h to 600 mM urea is improved from 14 to 43% by inducible overexpression of HSP72 (218). Thus the protective effect of high NaCl against high urea apparently is related to induction of HSP72 by high NaCl.
| IX. WHY IS OSMOTIC TOLERANCE OF RENAL MEDULLARY CELLS GREATER IN VIVO THAN IN CULTURE? |
|---|
|
|
|---|
One factor is rate of proliferation and the associated DNA replication. Proliferation is rapid in continuous lines but is very slow in vivo (335). Proliferation in vivo was measured by counting cells that express proliferating cell nuclear antigen (PCNA). It is very low in the rat renal inner medulla (335). The effect of proliferation on osmotic tolerance was studied by comparing a continuous line of renal inner medullary cells (mIMCD3) to early passage cells (p2mIME). Even when they are confluent, 20% of mIMCD3 cells are in the S phase, replicating DNA, whereas only 5% of p2mIME cells are (335). This is pertinent, since cells in S phase are particularly susceptible to the apoptosis caused by high NaCl (72). Starting from a baseline of 300 mosmol/kgH2O, few mIMCD3 cells survive acute increase of NaCl or urea above 600 mosmol/kgH2O. In contrast, most p2mIME cells survive NaCl added to 1,100 mosmol/kgH2O and of urea added to 700 mosmol/kgH2O. Similarly, when NaCl and urea are added in combination, mIMCD3 cells have increased apoptosis and decreased survival above 900 mosmol/kgH2O, whereas p2mIME cells survive well up to 1,300 mosmol/kgH2O; 1,300 mosmol/kgH2O is still less than inner medullary cells tolerate in vivo, which suggested the existence of other factors.
Possible additional factors that are present in vivo and might account for the greater osmotic tolerance were tested in p2mIME cells (40). Several of the factors did not affect their osmotic tolerance. The ineffective factors include hormones (insulin-like growth factor I, EGF, and deamino-8-D-arginine vasopressin), growth on porous supports, and presence of matrix proteins (collagen I, collagen IV, fibronectin, laminin, or fibrillar collagen I). The one factor that does make a major difference is the time course of the osmotic change. Greater survival had previously been observed following a slow increase in osmolality. Thus increasing NaCl by 50 mosmol/kgH2O from passage to passage conditions mIMCD3 cells to proliferate at 900 mosmol/kgH2O, which is higher than if the change is made acutely (42). Also, an initial small elevation of NaCl enhances survival of mIMCD3 after a later larger increase (168). However, the time course in these experiments was days or weeks, whereas inner medullary cells in vivo survive large increases of osmolality over a period of hours. During those hours, however, the increase is gradual, not an abrupt single step as in most tissue culture experiments. Under normal conditions, NaCl and urea change more or less in concert in renal inner medullary interstitial fluid. When osmolality is increased from 640 to 1,640 mosmol/kgH2O by addition of NaCl and urea in a single step, only 30% of p2mIME cells survive for 24 h. However, when the same increase is made linearly over 20 h, 89% of the cells remain viable 24 h later. Thus gradual changes in osmolality allow cells to survive much greater changes than do the step changes routinely used in cell culture experiments. A possible factor in this difference is that when the rate of change of osmolality is slow enough, isovolumetric regulation (IVR) occurs, i.e., no measurable change occurs in cell volume, despite large cumulative changes in osmolality (175, 208). In contrast, a step increase stresses cells more by rapidly shrinking them (175).
The more gradual, linear increase in osmolality is accompanied by a greater increase in mRNA expression of many genes known to be osmoprotective (38). For example, when with p2mIME cells NaCl and urea are increased in combination from total osmolality of 640 to 1,640 mosmol/kgH2O, a linear increase raises TonEBP/OREBP and AR mRNA, but a step increase reduces them. Heat shock proteins 70.1 (HSP70.1) and HSP70.3 increase more with a linear than a step increase, and osmotic stress protein 94 (OSP94) and heme oxygenase-1 (HO-1) increase with a linear, but decrease with a step increase. mRNAs for known urea-responsive proteins, GADD153 and Egr-1, increase with both a step and linear increase. A step increase in urea alone reduces mRNAs, similar to the combination of NaCl and urea, but a step increase in NaCl alone does not. Thus the poorer survival of p2mIME cells with a step than with linear increase in NaCl and urea is accounted for, at least in part, by urea-induced suppression of protective genes, particularly HSP70.
| X. OSMOPROTECTION BY GENERAL STRESS PROTEINS IN CELL CULTURE AND IN VIVO |
|---|
|
|
|---|
Acute elevation of NaCl increases mRNA and protein abundance of the growth arrest- and DNA damage-inducible protein Gadd45 in mIMCD3 cells (46, 155). Gadd45 was originally identified as a transcript that increases following numerous different stresses. Later, two additional Gadd45 genes were identified. The three Gadd45 genes are now called Gadd45a (the original Gadd45), Gadd45b, and Gadd45g. The Gadd45 proteins are multifunctional molecules with important roles in cell differentiation, apoptosis, negative growth control, DNA repair, genome stability, immunity, and cell survival. Possible consequences of their induction by high NaCl were initially examined by overexpressing recombinant Gadd45a, Gadd45b, or Gadd45g in mIMCD3 cells. Following acute increase of NaCl to a moderate level, each overexpressed Gadd inhibits mitosis and promotes G2/M arrest, and, in addition, Gadd45g overexpression strongly potentiates apoptosis (186). At a higher level of NaCl, Gadd45a/b overexpression transiently increases the early indicators of apoptosis, but inhibits the later stages. From these observations, it was concluded that increased abundance of Gadd45 isoforms, when NaCl is high, contributes to low mitotic index and protection of genomic integrity in cells of the renal inner medulla (186).
However, in these studies the baseline osmolality was 300 mosmol/kgH2O, whereas the lowest normal osmolality of renal medullary interstitial fluid is
600 mosmol/kgH2O. Also, mIMCD3 cells are immortalized by expression of SV40, so they proliferate rapidly and continue to proliferate even after they become confluent, whereas renal inner medullary cells proliferate very slowly in vivo. When, in subsequent studies, these differences were taken into account, no effect of the Gadd45s on osmotic tolerance was found. The later studies used early passage mouse inner medullary epithelial (p2mIME) cells instead of mIMCD3. p2mIME cells tolerate much higher osmolality than mIMCD3 cells and, not being immortalized, proliferate more slowly (335). With the use of p2mIME cells (37), elevation of the baseline osmolality of 640 mosmol/kgH2O to 1,040–1,640 mosmol/kgH2O by adding NaCl and/or urea increases both Gadd45b and Gadd45g mRNA, but lowers Gadd45a mRNA. Thus, under conditions similar to those in the inner medulla, Gadd45a does not respond to elevations of NaCl or urea, but Gadd45b and Gadd45g do. Even for Gadd45b and Gadd45g, however, their expression was not found to affect osmotic tolerance of the cells. p2mIME cells from Gadd45b–/–, Gadd45g–/–, and Gadd45bg–/– mice survive elevation of NaCl (or urea) essentially as well as do wild-type cells. Furthermore, the morphology of inner medullas from Gadd45abg–/– mice is normal, and they produce concentrated urine. Thus, although overexpression of Gadd45s affects osmotically stressed mIMCD3 cells, their absence does not seem to affect osmotic tolerance of slowly proliferating p2mIME cells or to perturb inner medullary epithelial cells in vivo.
p53 is best known as a tumor suppressor, but it has numerous additional functions. Accumulation and activation of p53 are promoted by DNA-damaging agents, which result in cell cycle delay during which the DNA is repaired. When the DNA is not successfully repaired, p53 activity can cause apoptosis. High NaCl increases the number of DNA strand breaks (see sect. IIC) and increases p53 protein abundance, phosphorylation on S15, and transcriptional activity (71). High NaCl is most harmful to cells when they are replicating DNA in the S phase of their cell cycle (72). In agreement with the results with other stress responses, accumulation and activation of p53 in mIMCD3 cells restrict DNA replication and reduce high NaCl-induced apoptosis in mIMCD3 cells (71, 72). The result is similar in p2mIME cells. After they become confluent, p2mIME cells from p53–/– mice proliferate and replicate DNA more rapidly than those from p53+/+ mice, resulting in less tolerance for high NaCl (37).
Since the increase in p53 activity that occurs in mIMCD3 cells, when NaCl is elevated, protects the cells from apoptosis (71, 72), it seemed likely that p53 might also be important for osmotic tolerance of renal inner medullary cells in vivo. However, this does not seem to be the case. Urinary concentrating ability of p53–/– mice is normal, as is the histology of their inner medullas (37). Therefore, although p53 expression is important for osmotic tolerance of mIMCD3 cells and even of p2mIME cells that are studied under conditions similar to those in the renal inner medulla, lack of p53 apparently is not deleterious to renal inner medullary cells in vivo.
Heat shock proteins (HSPs) are a group of highly conserved proteins that are expressed constitutively and/or induced by stress. Constitutively expressed HSPs participate in protein folding and assembly, elimination of misfolded proteins, and stabilization of newly synthesized protein in various intracellular compartments. Expression of inducible HSPs is evoked by various stresses, many of which denature proteins. The HSPs enhance cellular survival and aid cellular recovery by acting as molecular chaperones that limit or correct damage to proteins. Also, HSPs prevent apoptosis by inhibition of JNK activation, prevention of cytochrome c release from mitochondria, and disruption of apoptosomes. HSPs are abundant in renal medullas. High NaCl increases expression of HSPs, protecting cells from damaging effects of high urea (218, 219).
The abundance of several of the inducible HSPs increases along the corticomedullary osmotic gradient, including HSP25,
B-crystallin, HSP70, HSP110, and Osp90 (18, 258). Since hypertonicity increases expression of these heat shock proteins in cell culture (Table 1), and dehydration increases inner medullary HSP110 and Osp94 (258), high NaCl in the renal inner medullary interstitial fluid likely accounts for the great abundance of HSPs there.
Activity of TonEBP/OREBP mediates hypertonicity-induced increase in expression of one of the isoforms of HSP70. In mice there are two HSP70 genes that have identical open reading frames but that differ in their regulation. The HSP70.1 gene contains ORE/TonEs in its 5'-flanking region, and its expression is increased both by hypertonicity and heat shock in mIMCD3 (317), mouse embryo fibroblast (MEF) (272), and BALB/c3T3 (288) cells and by high NaCl in p2mIME (second passage mouse renal inner medullary) cells (38). HSP70.3 does not contain ORE/TonEs, and its expression is increased by heat, but not by hypertonicity, in mIMCD3 (317), MEF (272), and BALB/c3T3 (288) cell lines. In contrast, high NaCl does increase HSP70.3 expression in p2mIME cells (38), which apparently is independent of TonEBP/OREBP. Expression of both hsp70-2 (the human homolog of HSP70.1) and hsp70-1 (the human homolog of HSP70.3) is increased by heat, but only hsp70-2 is increased by hypertonicity in HEK293T and human embryonic fibroblast cells (38). The human hsp70-2 promoter region contains three ORE sites, and mutation of one of these ORE sites (at –135) completely abolishes induction by hypertonicity.
HSP70.1 expression helps cells survive high NaCl both in culture and in vivo. Embryonic fibroblasts from hsp70.1–/– mice have reduced tolerance for high NaCl. After preliminary treatment with elevated NaCl (100 mM added), hsp70.1+/+ MEFs survive a subsequent addition of 600 mM of NaCl, but no hsp70.1–/– cells survive under the same conditions (272). High NaCl elevates caspase-9 and caspase-3 activity in the hsp70.1–/– cells and kills them by apoptosis (164). Renal inner medullary cells of the hsp70.1–/– mice are also more susceptible to salt stress in vivo. The mice were given drinking water that contained 3% NaCl for 30 days. Under those conditions, the renal medullas of hsp70.1–/– mice contain many apoptotic (TUNEL-positive) cells, whereas wild-type mice have only a few (272).
Heme oxygenase 1 (HO-1) is induced by diverse stresses. It protects cells by its antioxidant, antiapoptotic, and anti-inflammatory actions (206). In the kidney, there is a gradient of HO-1 mRNA and protein, highest in the renal inner medulla and lowest in the renal cortex (351). HO-1 is induced by either high NaCl or urea in mIMCD3 (290) and p2mIME cells (38). Inhibition of HO-1 by the HO-1 inhibitor ZnDPBG reduces tolerance of mIMCD3 cells for high NaCl, but does not affect tolerance for high urea. A step increase of NaCl and urea in combination reduces HO-1 mRNA in p2mIME cells, but a linear increase raises HO-1 mRNA (38). This difference in HO-1 expression could contribute to the greater tolerance of p2mIME cells for a gradual increase than for a step increase of NaCl and urea in combination.
| XI. SUMMARY AND CONCLUSIONS |
|---|
|
|
|---|
| GRANTS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
|
|
|---|
| REFERENCES |
|---|
|
|
|---|
. Am J Physiol Cell Physiol 292: C1280–C1290, 2007.This article has been cited by other articles:
![]() |
S. C. Chow, L. Y. Ching, A. M. F. Wong, and C. K. C. Wong Cloning and regulation of expression of the Na+-Cl--taurine transporter in gill cells of freshwater Japanese eels J. Exp. Biol., October 15, 2009; 212(20): 3205 - 3210. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Z. Chen, X. Zhu, H.-Q. Sun, Y. S. Mao, Y. Wei, M. Yamamoto, and H. L. Yin Oxidative Stress Decreases Phosphatidylinositol 4,5-Bisphosphate Levels by Deactivating Phosphatidylinositol- 4-phosphate 5-Kinase {beta} in a Syk-dependent Manner J. Biol. Chem., August 28, 2009; 284(35): 23743 - 23753. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Gallazzini and M. B. Burg What's New About Osmotic Regulation of Glycerophosphocholine Physiology, August 1, 2009; 24(4): 245 - 249. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. K. Klose, G. L. Boulianne, R. M. Robertson, and H. L. Atwood Role of ATP-Dependent Calcium Regulation in Modulation of Drosophila Synaptic Thermotolerance J Neurophysiol, August 1, 2009; 102(2): 901 - 913. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. A. Lanaspa, A. Andres-Hernando, C. J. Rivard, Y. Dai, N. Li, and T. Berl ZAC1 Is Up-regulated by Hypertonicity and Decreases Sorbitol Dehydrogenase Expression, Allowing Accumulation of Sorbitol in Kidney Cells J. Biol. Chem., July 24, 2009; 284(30): 19974 - 19981. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Hao, H. Zhao, Z. Darzynkiewicz, S. Battula, and N. R. Ferreri Expression and function of NFAT5 in medullary thick ascending limb (mTAL) cells Am J Physiol Renal Physiol, June 1, 2009; 296(6): F1494 - F1503. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Aramburu and C. Lopez-Rodriguez Brx Shines a Light on the Route from Hyperosmolarity to NFAT5 Sci. Signal., April 7, 2009; 2(65): pe20 - pe20. [Abstract] [Full Text] [PDF] |
||||
![]() |
U. Hasler Controlled aquaporin-2 expression in the hypertonic environment Am J Physiol Cell Physiol, April 1, 2009; 296(4): C641 - C653. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Koivusalo, A. Kapus, and S. Grinstein Sensors, Transducers, and Effectors That Regulate Cell Size and Shape J. Biol. Chem., March 13, 2009; 284(11): 6595 - 6599. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. C. P. Thirone, P. Speight, M. Zulys, O. D. Rotstein, K. Szaszi, S. F. Pedersen, and A. Kapus Hyperosmotic stress induces Rho/Rho kinase/LIM kinase-mediated cofilin phosphorylation in tubular cells: key role in the osmotically triggered F-actin response Am J Physiol Cell Physiol, March 1, 2009; 296(3): C463 - C475. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. K. Hoffmann, I. H. Lambert, and S. F. Pedersen Physiology of Cell Volume Regulation in Vertebrates Physiol Rev, January 1, 2009; 89(1): 193 - 277. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. P. Choe and K. Strange Genome-wide RNAi screen and in vivo protein aggregation reporters identify degradation of damaged proteins as an essential hypertonic stress response Am J Physiol Cell Physiol, December 1, 2008; 295(6): C1488 - C1498. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. I. Dmitrieva and M. B. Burg Analysis of DNA breaks, DNA damage response, and apoptosis produced by high NaCl Am J Physiol Renal Physiol, December 1, 2008; 295(6): F1678 - F1688. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Inagaki, E. Omori, J.-Y. Kim, Y. Komatsu, G. Scott, M. K. Ray, G. Yamada, K. Matsumoto, Y. Mishina, and J. Ninomiya-Tsuji TAK1-binding Protein 1, TAB1, Mediates Osmotic Stress-induced TAK1 Activation but Is Dispensable for TAK1-mediated Cytokine Signaling J. Biol. Chem., November 28, 2008; 283(48): 33080 - 33086. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Gallazzini, J. D. Ferraris, and M. B. Burg GDPD5 is a glycerophosphocholine phosphodiesterase that osmotically regulates the osmoprotective organic osmolyte GPC PNAS, August 5, 2008; 105(31): 11026 - 11031. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Xu, C. C. L. Wong, E. H. Y. Tong, S. S. M. Chung, J. R. Yates III, Y. Yin, and B. C. B. Ko Phosphorylation by Casein Kinase 1 Regulates Tonicity-induced Osmotic Response Element-binding Protein/Tonicity Enhancer-binding Protein Nucleocytoplasmic Trafficking J. Biol. Chem., June 20, 2008; 283(25): 17624 - 17634. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Stuible, N. Dube, and M. L. Tremblay PTP1B Regulates Cortactin Tyrosine Phosphorylation by Targeting Tyr446 J. Biol. Chem., June 6, 2008; 283(23): 15740 - 15746. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Yang, J. M. Reece, J. Cho, C. D. Bortner, and S. B. Shears The Nucleolus Exhibits an Osmotically Regulated Gatekeeping Activity That Controls the Spatial Dynamics and Functions of Nucleolin J. Biol. Chem., April 25, 2008; 283(17): 11823 - 11831. [Abstract] [Full Text] [PDF] |
||||
![]() |
X. Zhou, J. D. Ferraris, N. I. Dmitrieva, Y. Liu, and M. B. Burg MKP-1 inhibits high NaCl-induced activation of p38 but does not inhibit the activation of TonEBP/OREBP: Opposite roles of p38{alpha} and p38{delta} PNAS, April 8, 2008; 105(14): 5620 - 5625. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. B. Burg and J. D. Ferraris Intracellular Organic Osmolytes: Function and Regulation J. Biol. Chem., March 21, 2008; 283(12): 7309 - 7313. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. C. Bartolo and J. A. Donald The effect of water deprivation on the tonicity responsive enhancer binding protein (TonEBP) and TonEBP-regulated genes in the kidney of the Spinifex hopping mouse, Notomys alexis J. Exp. Biol., March 15, 2008; 211(6): 852 - 859. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |