Osmolality of the mammalian renal medulla is high because of the operation of the urinary concentrating mechanism. To understand molecular events during the early phase of cellular adaptation to hypertonicity, we performed comprehensive searches for genes induced in response to hypertonicity using a cell line (mIMCD3) derived from the inner medullary collecting duct of mouse kidney. PCR-based subtractive hybridization of cDNA pools and cDNA microarray analysis were used. We report 12 genes whose mRNA expression is significantly increased within 4 h after exposure to hypertonicity. The increase in mRNA expression was the result of increased transcription. Many are either stress response genes or growth regulatory genes, supporting the notion that hypertonicity evokes the stress response and growth regulation in cells. Experiments using inhibitors revealed that mitogen-activated protein kinases were commonly involved in signaling for the induction of genes by hypertonicity. Tyrosine kinases and phosphatidylinositol 3-kinase also play a significant role. Signaling pathways for stimulation of transcription appeared quite diverse in that each gene was sensitive to different combinations of inhibitors.
- renal medulla
- p38 mitogen-activated protein kinase
- complimentary deoxyribonuclease microarray analysis
osmolality of themammalian kidney medulla is very high. In rat kidney medulla, the osmolality routinely exceeds 3,000 mosmol/kgH2O depending on the animal's hydration status. The high osmolality provides the driving force for water reabsorption and urinary concentration, an important function of the kidney for proper maintenance of blood pressure (25). There are two principal solutes in the interstitium of the renal medulla (NaCl and urea; see Ref.5). Hyperosmolal NaCl and urea both induce apoptosis in a dose-dependent manner (37, 47). However, the underlying mechanisms appear different. Exposure of a cell to hyperosmolal salt (hypertonicity) results in an immediate increase in ionic strength inside the cell and double-stranded DNA breaks (29). As a result, the classical response to double-stranded DNA breaks is induced: apoptosis on one hand and activation of ATM kinase and induction and activation of p53, which in turn opposes apoptosis and arrests cell cycle progression (16), on the other hand. The balance of the two opposing pathways is determined by the degree of hypertonicity. In contrast, hyperosmolar urea does not increase cellular ionic strength nor causes double-stranded DNA breaks (29).
Cells in the renal medulla adapt to hypertonicity by accumulating organic solutes called “compatible osmolytes” or “organic osmolytes,” such as betaine, taurine, sorbitol, andmyo-inositol (21). The accumulation of compatible osmolytes reduces the stress of hypertonicity by lowering cellular ionic strength resulting from osmotic replacement (5). The cellular accumulation of compatible osmolytes is orchestrated in large part by a transcription factor named tonicity-responsive enhancer binding protein (TonEBP, also called NFAT5; see Ref. 39). TonEBP is stimulated by hypertonicity and, in turn, stimulates transcription of genes that encode the Na+-myo-inositol cotransporter (SMIT; see Ref.46), the Na+-Cl−-betaine cotransporter (BGT1; see Ref. 38), and aldose reductase (AR; see Ref. 27), which are responsible for the cellular accumulation of myo-inositol, betaine, and sorbitol, respectively. Emerging data suggest that hypertonicity is also a signal for tissue-specific gene expression. The vasopressin-regulated urea transporter (UT-A) is exclusively expressed in the renal medulla and plays a key role in accumulation of urea (4). TonEBP also stimulates transcription of UT-A (42). Thus hypertonicity induces a specific set of gene expression that determines the phenotype of the renal medulla and allows cells to overcome the stress of hypertonicity.
The sensors and signaling pathways to TonEBP are not known. Because it takes several hours for induction of TonEBP in response to hypertonicity (58), there might be early genes required for this process. In addition, we are interested in uncovering the network of genes that are activated by hypertonicity. To explore these questions, we performed comprehensive searches for genes in which mRNA expression was increased immediately after exposure to hypertonicity. We report 12 genes in which transcription is stimulated within 4 h after cells are exposed to hypertonicity. Experiments with inhibitors indicate that multiple signaling pathways with a variety of protein and lipid kinases are involved in the stimulation of gene transcription.
MATERIALS AND METHODS
Cell cultures and reagents.
mIMCD3 cells were established from the inner medullary collecting ducts of a transgenic mouse expressing SV40 large T-antigen (43). The cells of passage number fewer than 20 were maintained on plastic tissue culture dishes in DMEM-F-12 (1:1) supplemented with 10% FBS and 2% penicillin-streptomycin. For hyperosmolality experiments, confluent cells were exposed to control isoosmolar medium or hyperosmolar medium supplemented with (in mM) 100 NaCl, 200 raffinose, or 200 urea. To study the effects of different stresses, cells were exposed for 2.5 h to heat (42°C for 1.5 h followed by a 1-h recovery period) or for 4 h to various concentrations of H2O2. For inhibitor experiments, cells were initially incubated with inhibitors in isotonic medium for 1 h and switched to isotonic or hypertonic medium (100 mM NaCl) containing the same concentration of inhibitors. Cells were further incubated for either 1 or 4 h depending on the genes studied. Each inhibitor was dissolved in DMSO, and an appropriate volume of this solution was added to the medium. The largest volume of solvent used to dissolve inhibitors was added to the medium of controls. The final concentration of DMSO in the medium was <0.1%. Actinomycin D was purchased from Sigma (St. Louis, MO). Cycloheximide, SB-203580, genistein, and LY-294002 were purchased from Calbiochem (San Diego, CA). U-0126 was purchased from Promega (Madison, WI).
mIMCD3 cells were cultured for 1 or 4 h in either isotonic or hypertonic (100 mM NaCl added) medium. Total RNA was isolated using Trizol reagent (Life Technologies, Rockville, MD), and poly(A) RNA was affinity purified on an Oligotex column (Qiagen, Valencia, CA). The poly(A) RNA were reverse transcribed, and suppressive subtractive hybridization (15) was performed to enrich cDNAs overrepresented in the hypertonic sample compared with matched isotonic sample using the PCR-Select cDNA Subtractive Hybridization kit (Clontech, Palo Alto, CA). The enriched cDNAs were recovered by PCR and cloned into pCRII TOPO vector using the T/A cloning kit (Invitrogen, Carlsbad, CA). The selected cDNA clones were screened for false positives according to the manufacturer's instructions, and the final cDNAs were further analyzed by Northern blot analysis and DNA sequencing. Nucleic and amino acid homology searches were performed using the BLAST program.
cDNA microarray analysis.
With the use of a commercial service (Genome Systems, St. Louis, MO), the same pairs of poly(A) RNA used above were also analyzed to identify genes that are differentially expressed. Poly(A) RNA from isotonic and hypertonic cells were reverse transcribed separately with 5′-Cy3- and Cy5-labeled random nanomers. The labeled probes were then hybridized to a mouse GEM1 Microarray containing 8,700 cDNAs. The results were analyzed with the GEM tool 2.4 software. The cDNA clones for those genes whose expression was higher in hypertonic cells were purchased and used for Northern blot analyses.
Northern blot analysis.
An equal amount of RNA (10 μg) as determined by ultraviolet (UV) absorbance was separated on an 1% agarose gel containing 2.2 M formaldehyde and was transferred to a nitrocellulose membrane. Membranes were then hybridized overnight with 32P-labeled cDNA probes. After washing for 20 min at 60°C in 0.5× saline-sodium citrate and 0.1% SDS, radioactivity was visualized and quantified using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). To demonstrate equal loading of RNA, ethidium bromide staining of the blot (shown in Figs. 1 and2) and Northern analysis with glyceraldehyde-3-phosphate dehydrogenase probe (data not shown) were performed with the same results.
Identification of mRNAs upregulated in response to hypertonicity using suppressive PCR subtractive hybridization and cDNA microarray analysis.
In an effort to identify genes that are stimulated early in response to hypertonicity, mIMCD3 cells were cultured in hypertonic and control isotonic medium for 1 and 4 h before poly(A) RNA was isolated. The mRNA species enriched in the hypertonic cells was identified using two different methods [suppressive PCR subtractive hybridization (SSH) and cDNA microarray analysis] as described inmaterials and methods. SSH is an “open” system that does not require any knowledge of genes or their sequences, whereas the microarray analysis is a “closed” system that uses available gene sequences, mostly expressed sequence tags (ESTs).
From the initial enriched cDNA libraries of SSH, we obtained 150 clones from the 1-h library and 100 clones from the 4-h library. Two rounds of selections, including Northern blot analysis of RNA from isotonic and hypertonic cells, yielded 12 true positive clones from the 1-h library and 16 clones from the 4-h library. The selected cDNA clones were sequenced and BLAST matched for their identity. Some were previously reported sequences (see Table 1), whereas others were not. Of note, 12 out of the 16 4-h clones were partial cDNAs of mouse AR. Because the induction of AR by hypertonicity has been characterized extensively (21, 27), the AR clones were excluded from the subsequent experiments. Three clones out of the 12 1-h clones are partial cDNAs for amino acid transport system A2 (ATA2). Thus a total of 15 genes were identified from SSH.
Aliquots from the two pairs of isotonic and hypertonic poly(A) RNA used for SSH were also used for cDNA microarray analysis. The microarrays contained 8,700 mouse EST clones. Based on analysis of the data with the GEM tool software, EST clones whose expression increased 1.5-fold and higher in hypertonic cells were selected, and Northern blot analyses were performed to confirm the increase in mRNA abundance. Six clones were verified from the 1-h sample and 14 clones from the 4-h sample. There was no overlap with the SSH clones except AR. In summary, using SSH and cDNA microarray analysis, we identified 34 genes in which mRNA was upregulated in response to hypertonicity. Function was known in 13-AR plus those in Table 1 that were chosen for further analysis.
Characterization of mRNA induction by hypertonicity.
To characterize the induction of mRNA species identified above, we performed Northern blot analysis (Fig. 1, A andB). First, the time course of mRNA induction in response to hypertonicity was examined (Fig. 1 A). The kinetics of mRNA induction was consistent with the data obtained by SSH and the microarray analysis (Table 1). Expression of all the genes obtained from the 1-h sample was induced at 1 h. Among the 1-h genes, expression of ATA2, thrombospondin-1 (TSP-1), human liver DnaJ-like protein (HLJ1), and inhibitory factor-κB-α (IκB-α) remained elevated at 4 h. Expression of other genes [growth-arrest and DNA damage-inducible (GADD) 34, CYR61, and TIS11b] returned to the basal (isotonic) level, suggesting that their function may be specific for the acute-phase adaptive process during hypertonic stress. Expression of the 4-h genes peaked at 4 h except tissue plasminogen activator (t-PA) and p8.
To investigate the nature of the osmotic stimulus responsible for the induction of gene expression, we examined the effects of urea (a permeable solute) and raffinose (effectively nonpermeable solute like NaCl). As shown in Fig. 1 B, expression of all 12 genes was induced by raffinose but not by urea. Therefore, like TonEBP (58), these genes respond to hypertonicity rather than hyperosmolality.
Effects of other stresses.
Expression of many stress proteins is induced by more than one kind of stress. For instance, expression of several heat shock proteins is induced by hypertonicity and heat (44, 48). As shown in Fig. 2 A, seven genes were induced by heat and four genes were induced by H2O2 (i.e., oxidant stress) in a dose-dependent manner. GADD34, IκB-α, and TIS11b were induced by both heat and H2O2, indicating that they are general stress response genes.
Effects of actinomycin D and cycloheximide.
Because mRNA accumulation is a function of gene transcription and mRNA stability, we asked whether transcription played a role in the accumulation of mRNA in response to hypertonicity. In addition, the possibility that the upregulated genes are immediate early genes was investigated since the induction of these genes occurred within 4 h. By definition, transcription of immediate early genes is stimulated rapidly and is not dependent on de novo protein synthesis. Confluent mIMCD3 cells were treated with either actinomycin D (5 μg/ml) or cycloheximide (10 μg/ml) for 1 h before being switched to hypertonic medium for 1 or 4 h. Actinomycin D inhibited mRNA induction of all of the 1- and 4-h genes by hypertonicity (45–95% reduction in mRNA expression; Fig. 3,lane 2). These data indicate that transcription rather than increased mRNA stability is the key step for mRNA accumulation of the upregulated genes in response to hypertonicity.
As shown in the Fig. 3, lane 3, cycloheximide completely abolished the induction of t-PA, p8, and 3-O-sulfotransferase-1 (3-OST-1) mRNA (85–94% reduction in mRNA abundance in hypertonic condition). This indicates that de novo protein synthesis is required for the induction of t-PA, 3-OST-1, and p8 genes. The induction of four-and-a-half LIM-only protein 2 (FHL2) mRNA was also inhibited significantly by cycloheximide (∼40% reduction in the mRNA abundance). On the other hand, the mRNA expression levels of all of the 1-h genes and ephrin receptor A2 (EphR A2) gene in cells under hypertonic conditions were not inhibited by cycloheximide, indicating that protein synthesis is not required for the induction of these genes by hypertonicity. Therefore, by definition, these genes are immediate early genes for hypertonicity.
Effects of kinase inhibitors.
Exposure of cells to hypertonicity leads to immediate and vigorous activation of a number of kinases, including the three major families of mitogen-activated protein kinase (MAPK), tyrosine kinases, and phosphatidylinositol 3-kinase (PI3K; see Refs. 6,28, 54, 61). We used inhibitors to investigate the role of the kinases in the same way we used actinomycin D and cycloheximide (see above). A given inhibitor was scored as inhibitory (see Fig. 5, bottom) only if it exhibited a statistically significant inhibition on both absolute expression of mRNA in the hypertonic condition (Fig.4) and a degree of induction of mRNA expression (Fig. 5).
As shown in Figs. 4 and 5, lane 2, SB-203580, an inhibitor of p38 MAPK, inhibited the induction of all genes except IκB-α. In fact, none of the inhibitors affected the abundance of IκB-α mRNA. Genistein, a broad-spectrum tyrosine kinase inhibitor, significantly inhibited induction of GADD34, TIS11b, EphR A2, and 3-OST-1 (lane 3). LY-294002, an inhibitor of PI3K, inhibited the induction of HLJ1, p8, and 3-OST-1 (lane 4). In addition, LY-294002 enhanced the mRNA expression of GADD34 and TIS11b in the isotonic condition. This resulted in a statistically significant reduction in the degree of induction without a reduction of absolute mRNA expression in the hypertonic condition (Fig. 5). This implies that PI3K may negatively regulate the expression of GADD34 and TIS11b in the isotonic condition. U-0126, an inhibitor of mitogen/extracellular signal-regulated kinase (MEK) 1, which is an upstream kinase of extracellular/signal-regulated kinase, inhibited the induction of CYR61, t-PA, p8, EphR A2, FHL2, and 3-OST-1 (lane 5). Thus induction of all of the 4-h genes was significantly inhibited by the MEK1 inhibitor U-0126. Collectively, these results suggest that p38 MAPK is involved most widely in gene induction by hypertonicity. Tyrosine kinase and PI3K also play a significant role. For many genes (Fig. 5, bottom), more than one protein kinase is involved, suggesting that multiple kinases may be part of a signaling pathway.
Using SSH and cDNA microarray analysis, we identified 34 genes in which transcripts were upregulated early when cells are exposed to hypertonic medium. Only those genes (Table 1) in which function has been described, except AR, were selected and characterized. Many of them are either general stress response genes or growth regulatory genes, in line with other reports that hypertonicity results in the stress response and perturbation of cell growth (16, 29, 37,47). It is likely that the increase in mRNA abundance results in a corresponding increase in protein abundance. In case of TonEBP, a threefold increase in mRNA abundance in response to hypertonicity leads to a threefold increase in synthesis of TonEBP (58).
Eight of the selected genes, including all of those induced at 1 h after exposure to hypertonicity, are immediate early genes in that their induction by hypertonicity was not prevented by cycloheximide, an inhibitor of protein synthesis. Interestingly, many of them (GADD34, CYR61, TIS11b, HLJ1, IκB-α, and FHL2) were induced by cycloheximide in isotonic conditions (Fig. 3). Induction of these genes by hypertonicity might be caused by inhibition of protein synthesis per se because hypertonicity is known to inhibit the overall rate of protein synthesis (11).
Induction of the ATA2 gene by hypertonicity.
Three cDNA clones identified from SSH were the mouse ortholog of human (22) and rat (53, 60) ATA2 (or SAT2). ATA2 is ubiquitously expressed in mammalian tissues, including kidney and brain. It mediates Na+-coupled cellular uptake of short-chain neutral amino acids such as alanine, serine, proline, and glutamine. Its induction in response to hypertonicity and its putative role as a compatible osmolyte transporter have been described in the literature (10, 12, 59). The induction and shut off of ATA2 are much faster (Fig. 1 A) than those of SMIT, BGT, and AR (21, 58). This observation is consistent with the hypothesis (12) that the “regulatory volume increase” that occurs immediately after exposure to hypertonicity is mediated by the accumulation of neutral amino acids. Subsequently, neutral amino acids are replaced by inositol, betaine, and sorbitol as the activity of SMIT, BGT1, and AR increases. In addition, its robust expression in brain (22, 53, 60) raises the possibility that ATA2 contributes to the maintenance of brain cell volume during hypernatremia.
Induction of stress response and growth regulatory genes by hypertonicity.
During the early phase of exposure to hypertonicity, the structure and function of proteins are perturbed (44). It has been postulated that heat shock proteins protect the cell by stabilizing protein conformation until compatible osmolytes accumulate to sufficient levels (44, 48). Hypertonicity induces the expression of many heat shock proteins, including Hsp70, αB-crystallin, Hsp110, and Osp94 (44, 48). There is a renal corticomedullary gradient of Hsp70, Hsp25, and Osp94 expression that parallels the osmotic gradient (40). In this regard, our discovery of the early induction of HLJ1 by hypertonicity is highly significant. HLJ1 is a member of the Hsp40 family and is expressed ubiquitously, including in the kidney (23). The members of the Hsp40 family function in association with Hsp70 molecular chaperones to facilitate protein folding (26). As cochaperones, they recruit Hsp70 partners and augment their binding to their protein substrates by accelerating the ATP hydrolysis step of the chaperone cycle. The simultaneous induction of Hsp40-Hsp70 functional partners reinforces the notion that hypertonicity is a form of stress that denatures proteins.
Hypertonicity as a stress is further reinforced by GADD34 induction. GADD34 becomes the third reported member of the GADD proteins induced by hypertonicity in addition to GADD45 and 153 (30). The five GADD genes were originally isolated as UV-inducible transcripts in Chinese hamster ovary cells (19). Later it was established that they are general stress proteins induced by different stresses, including alkylating agents, hypoxia, starvation, and others (18). Although their names suggest a role in growth arrest, the precise role of these proteins is not clear (24).
IκBs are a family of inhibitory subunits of nuclear factor (NF)-κB that play a key role in inflammation (20). Heat shock induction of IκB-α mRNA (Fig. 2 A) is probably the result of heat shock factor binding to the promoter region (57) like that of Hsp70. This is associated with mitigated activation of NF-κB in response to inflammatory stimuli (56, 57). It would be very interesting to see if hypertonicity also negatively modulates activation of NF-κB. At any rate, IκB-α appears to be a general stress response protein in that it is also induced by oxidant stress (Fig. 2 B).
The p8 gene was originally cloned from a rat pancreatic cDNA library based on its overexpression in pancreatic acinar cells during the acute phase of pancreatitis (36). Although its function is not known, studies using cultured cells suggested that p8 is a mitogen (52). Its role in cellular adaptation to hypertonicity remains to be determined.
CYR61 is a secreted cysteine-rich and heparin binding protein that associates with the extracellular matrix (ECM) and cell surface (32). Upon its secretion, CYR61 functions as a regulatory ECM signaling molecule that mediates cell-cell and cell-ECM interactions and promotes cell proliferation, adhesions, and differentiation. TSP-1 is the first identified member of the thrombospondin family of ECM proteins that participates in cell-cell and cell-matrix communication (1). It is induced rapidly by stressful conditions such as hypoxia and growth factors, including platelet-derived growth factor, transforming growth factor-β, and basic fibroblast growth factor. TIS11b is a member of the TIS11 gene family of immediate early genes, which are rapidly and transiently induced by phorbol esters and polypeptide mitogens in a wide variety of tissues (3, 55). TIS11b may modulate growth or survival pathways in cells imposed by external stimuli (50).
Induction of the genes involved in anticoagulation and fibrinolysis by hypertonicity.
Heparan sulfate d-glucosaminyl 3-OST-1 converts nonanticoagulant heparan sulfate to anticoagulant heparan sulfate by transferring a sulfate group to the 3-OH position of glucosamine residues (34). The anticoagulant heparan sulfate binds to and activates antithrombin, which is an important natural anticoagulant. t-PA is a serine protease that converts the blood zymogen plasminogen to more active plasmin and degrades the fibrin clot. In HeLa and human umbilical cord endothelial cells, t-PA mRNA expression is stimulated by hypertonicity (33), as in mIMCD3 cells reported here. It is possible that 3-OST-1 and t-PA are TonEBP target genes like SMIT and BGT1 because their induction is slow (Fig. 1 A) and sensitive to cycloheximide.
Other genes induced by hypertonicity.
A member of the Eph family of receptor tyrosine kinases, the EphR A2, is widely expressed in the central and peripheral nervous system (17). The ephrin receptors and their ligands play a role in axon guidance, cell migration, and boundary formation between groups of cells during development of the nervous system. FHL2 is a member of the FHL protein family (9). The hallmark of the FHL proteins is the presence of four LIM domains and a LIM half-motif located at the amino terminus of the protein. The LIM domain is a cysteine-rich, double-zinc motif involved in protein-protein interaction (2) and regulation of transcription (13). The physiological implication of the induction of EphR A2 and FHL2 by hypertonicity is not known. Because both are also induced by heat shock (Fig. 2), they may be general stress response genes.
Involvement of multiple signaling pathways in gene induction by hypertonicity.
In 11 of the 12 genes studied, gene induction by hypertonicity was inhibited by SB-203580 (Figs. 4 and 5). This is significant in that p38 MAPK is implicated in adaptation to hypertonicity from yeast to mammals. The yeast ortholog of p38 MAPK, named Hog1p, plays a key role in adaptation to hypertonicity (8). Hog1p is part of the protein kinase cascade (35) that relays the two osmosensors in the plasma membrane to the early phase ofGPD1 stimulation (45). Gpd1p is a key enzyme in biosynthesis of glycerol, the major compatible osmolyte of yeast (8). In mammalian cells, SB-203580 inhibits TonE-mediated induction of BGT1, SMIT, and HSP70 (14, 41, 49), indicating that p38 MAPK is involved in signaling to TonEBP. Therefore, the p38 MAPK-sensitive genes reported here are candidate early genes in the signaling pathways to TonEBP.
In contrast to SB-203580, other inhibitors inhibited the gene induction only in distinct subsets of genes (Figs. 4 and 5). With the exception of GADD34 and TIS11b, a different combination of inhibitors was effective for each gene, indicating a unique signaling mechanism. It should be noted that SB-203580 inhibits other protein kinases and signaling molecules in addition to p38 MAPK (7, 31). Thus the genes that are sensitive to SB-203580 do not necessarily share the same signaling pathways. Related to this, hypertonicity causes double-stranded breaks of DNA and, as a result, activates a number of DNA damage-dependent protein kinases such as ATM and DNA-dependent protein kinase (16, 29). Because LY-294002 and wortmannin also inhibit these kinases (51), their role should be considered for the effects of LY-294002 in addition to that of PI3K. Collectively, these data indicate that signaling pathways for stimulation of gene transcription in response to hypertonicity are quite diverse, involving a variety of protein and lipid kinases.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-42479 (to H. M. Kwon), National Research Service Award DK-9960 (to O. Nahm), and a fellowship from the Juvenile Diabetes Foundation International (3-1999-727; to S. K. Woo).
↵* O. Nahm and S. K. Woo contributed equally to this work.
Address for reprint requests and other correspondence: H. M. Kwon, 963 Ross Bldg., 720 Rutland Ave., Baltirmore, MD 21205 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published September 21, 2001; 10.1152/ajpcell.00267.2001
- Copyright © 2002 the American Physiological Society