p38 Kinase Activity Is Essential for Osmotic Induction of mRNAs for HSP70 and Transporter for Organic Solute Betaine in Madin-Darby Canine Kidney Cells*

In renal cells, hypertonicity induces genes for heat shock proteins (HSP70, αB-crystallin), as well as enzymes and transporters directly involved in the metabolism and transport of protective organic osmolytes. While heat shock proteins are induced by many stresses including osmotic stress, the induction of the osmolytes genes appears to be specific to osmotic stress. These two adaptive mechanisms allow kidney cells to survive and function in the hypertonic environment that exists on routine basis in kidney medulla. In mammalian cells, hypertonicity induces three mitogen-activated protein kinase pathways: ERK (extracellular regulated kinase), JNK (Jun N-terminal kinase), and p38. ERK activation by osmotic stress is a consistent finding in many cells, but it is not essential for transcriptional regulation of mRNA for transporter of organic osmolyte betaine. While the growth of yeast cells on NaCl-supplemented medium is dependent on HOG1 pathway, it is still unclear which pathway mediates the adaptation to osmotic stress in mammalian cells. Here, we show that inhibition of p38 kinase activity, using the specific inhibitor SB203580 (4-(fluorophenyl)-2-(4-methylsulfonylphenyl)-5-(4-pyridyl) imidazole), abolishes the hypertonicity-mediated induction of mRNAs for HSP70 and betaine transporter in Madin-Darby canine kidney cells. The inhibition is dose-dependent and correlates with thein situ activity of native p38 kinase, determined as MAPKAPK-2 activity in cell extracts. As reported previously, the activities of ERK-1 and -2 were not affected by SB203580, but surprisingly, inhibition of native p38 kinase activity correlates with up-regulation of native JNK-1 activity in osmotically stressed cells. p38 mRNA is induced by hypertonic stress and is attenuated with p38 kinase inhibition. We also find that thermal induction of HSP70 mRNA is not affected by p38 kinase inhibition. Such findings suggest that p38 kinase activity is essential for the induction of genes involved in the adaptation of mammalian cells to osmotic stress and that the increased activity of JNK-1 during p38 kinase inhibition is consistent with regulation of JNK-1 by p38 kinase in osmotically stressed cells. In addition, the transduction pathways mediating HSP70 mRNA induction by different stresses appear to be divergent; osmotic induction of HSP70 is p38 kinase-dependent, while thermal induction is not.

During water deprivation, the extracellular osmolality in the mammalian renal medulla can exceed 3000 mosmol/kg of H 2 O. While such an osmolality is incompatible with the survival of cells from other organs, kidney cells tolerate it well. The survival of kidney medullary cells in hypertonic environment is essential for the generation of the concentrating gradient, which is key for maintenance of body solute and water homeostasis.
Many organisms, including bacteria, yeast, plants, and animals, adapt to sustained hyperosmotic stress by accumulating osmotically active organic solutes (compatible organic osmolytes) (1). These compounds do not perturb cellular macromolecules and are preferentially accumulated over inorganic salts (1). Extracellular hyperosmolality exerted by a non-permeable solute (e.g. NaCl or raffinose, but not urea) has a hypertonic effect, i.e. it shrinks cells and increases intracellular potassium and sodium concentrations (ionic strength). There is evidence that increased intracellular ionic strength is among the initial signals for the induction of genes responsible for organic osmolytes accumulation (osmoprotective genes) (2,3). How the increased intracellular ionic strength ultimately induces transcription of these genes remains to be determined. The molecular mechanisms involved in this adaptive process in mammalian cells have been extensively studied in two renal epithelial cell lines, Madin-Darby canine kidney (MDCK) 1 and PAP-HT25 (rabbit renal papillary epithelium) (4). In these cells, hypertonicity induces the transcription of genes that encode proteins (specific enzymes and transporters) directly involved in the metabolism and transport of organic osmolytes such as sorbitol (2,5,6), betaine (7), and inositol (8). The induction of these genes by hypertonic stress appears to be relatively specific, since it is not produced by other stresses such as heat shock (9). Hypertonicity also induces mRNA for heat shock protein 70 (HSP70), and it is proposed that the induction of heat shock proteins during osmotic stress protects intracellular macromolecules from the harmful effects of ele-vated intracellular ionic strength until the cell accumulates the appropriate level of organic osmolytes (9,10).
Recently, it was reported that the mitogen-activated protein kinase (MAPK) HOG1 and its activator (MAPK kinase), PBS2, are involved in osmosensing signal transduction pathway in yeast, and that cells that carry mutations in these genes fail to grow in medium supplemented with NaCl (11). In mammalian cells, activation by osmotic stress of three MAPK pathways has been shown so far, JNK (also known as stress-activated protein kinase (SAPK)) (12,13), ERK (13,14) and p38, a lipopolysaccharide-induced MAPK in mammalian cells (HOG1 homologue), which complements HOG1 gene mutant yeast cells that fail to grow otherwise in hypertonic medium (15).
While demonstration of MAPKK and ERK activation by osmotic stress in MDCK cells has been shown (14), ERK activity does not appear to be essential for either transcriptional regulation of the betaine transporter gene (16) or the stimulation of inositol uptake (17) during osmotic stress. Furthermore, despite evidence linking N-terminal Jun phosphorylation to osmosensing transduction pathway proximal to gene activation (12), it is unclear which MAPK transduction pathway is essential for osmotically driven gene transcription in mammalian cells.
Here we show that the p38 kinase inhibitor SB203580 blocks the osmotically driven induction of mRNAs for HSP70 and BGT1, the betaine transporter. The effect is dose-dependent and correlates with in situ p38 kinase activity. Consistent with previous reports (18,19), ERK-1 and -2 activities were not inhibited by SB203580, but surprisingly, inhibition of p38 kinase in hypertonically stressed cells correlates with marked up-regulation of JNK-1 activity. In addition, we find that the thermal induction of HSP70 mRNA is not dependent on the activity of p38 kinase, suggesting divergent pathways for induction of HSP70 gene by different stresses.

EXPERIMENTAL PROCEDURES
Cell Culture-Confluent MDCK cells (American Type Culture Collection, Rockville, MD), grown on polystyrene (Corning, NY) dishes were used at passages 65-78. Cells were grown in serum-free, previously defined medium (20, 21) (315 mosmol/kg of H 2 O), containing 120 M myoinositol (inositol) and no betaine or taurine. In some experiments, the media were made hyperosmotic by the addition of 200 mosmol of NaCl/kg of H 2 O (final osmolality of 515-525 mosmol/kg of H 2 O). All cultures were maintained in 5% CO 2 , 95% air at 37°C. Heat shock was induced by placing cells in 42°C incubator for predetermined periods of time.
Northern Blot Analysis-Total RNA was isolated using RNAzol (22), and poly(A) ϩ RNA was isolated with oligo(dT) columns (Collaborative Biomedical Products), as described previously (6). Electrophoresis was performed by loading equal amounts of poly(A) ϩ RNA per lane on 1% agarose, 2.2 M formaldehyde gel, followed by transfer to GeenScreen membrane (NEN Life Science Products) (6). Dog betaine transporter cDNA was obtained by NotI, MluI digestion of pBGT1 (23) (a generous gift from Drs. Kwon and Handler, Johns Hopkins School of Medicine, Baltimore, MD). Human HSP70 insert (2.3 kb) was obtained by BamHI, HindIII digestion of pAT153 (American Type Culture Collection). Human (2 kb) full-length ␤-actin cDNA was purchased from CLONTECH. Inserts were labeled with [␣-32 P]dCTP (Random Primed DNA labeling kit, Boehringer Mannheim) for use as probes. Probes are hybridized to the blots overnight at 42°C in a solution containing 40% formamide, 5 ϫ SSC (0.75 M NaCl, 75 mM trisodium citrate, pH 7), 5 ϫ Denhardt's solution (0.5% (w/v) polyvinylpyrrolidone, 0.5% (w/v) Ficoll 400), 0.5% SDS, 250 g/ml salmon sperm DNA, 10 mM Tris, pH 7.5, and 10% dextran sulfate. The blots were then washed under high stringency at 65°C as follows: for 30 min twice in 3 ϫ SSC, 0.5% SDS; 1 h in 3 ϫ SSC, 0.5% SDS; 30 min twice in 0.3 ϫ SSC, 0.5% SDS. A 35-nucleotide p38-specific antisense oligonucleotide probe was designed, based on published data and used as probe. The sequence of the p38-specific probe is 5Ј-TGG TCT GTA CCA GGA AAC AAT GTT CTT CCA GTC AA-3Ј and corresponds to bases 857-891 of human p38 kinase (19) (GenBank accession no. L35264). Sequence specificity of the probe was determined using BLAST search of the National Center for Biotechnol-ogy Information (NCBI) data bases, aided by Baylor College of Medicine Molecular Biology Computational Resource Center (MBCR) services. End-labeled oligonucleotide was hybridized to blots overnight at 42°C in a solution as above but containing 100 mg/ml salmon sperm DNA and no formamide. These blots were then washed under high stringency at 42°C as follows: for 30 min in 5 ϫ SSC, 0.5% SDS; 1 h in 2 ϫ SSC, 0.5% SDS; 30 min in 0.5 ϫ SSC, 0.5% SDS; and 30 min in 0.1 ϫ SSC, 0.5% SDS. The BGT1 probe hybridizes in MDCK cells to 2.8-and 3.4-kb bands that behave similarly under hypertonic conditions. HSP70 and ␤-actin probes hybridize to 2.8-and 2.2-kb bands, respectively. Autoradiographs were prepared using X-Omat AR film (Eastman Kodak Co.) with an intensifying screen. Relative band intensities were determined by scanning laser densitometry (Ultroscan, LKB Biotechnology). mRNA abundance was determined relative to that of cells maintained in isotonic medium, after normalization of band intensities to ␤-actin bands for each respective time point. For consistency of BGT1 data analysis, the 2.8-kb bands were compared. mRNA abundance for HSP70 was determined relative to that of non-heat-stressed cells.
Kinase Assays-p38 kinase activity was determined as per Pombo et al. (24) and Moriguchi et al. (12) with slight modifications. Briefly, experimental medium was removed and cells were washed using icecold phosphate-buffered saline of equal osmolality. Dishes were kept on ice, and cells were lysed in buffer containing: 20 mM Tris, pH 7.4, 137 mM NaCl, 2 mM EDTA, 1% Triton X-100, 25 mM ␤-glycerophosphate, 1 mM sodium orthovanadate, 2 mM sodium pyrophosphate, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 1 g/ml leupeptin.Cell lysate was cleared of large particles by 10-min centrifugation at 15,000 ϫ g. After immobilizing anti-p38 kinase antibody (p38 antibodies from Santa Cruz Biotechnology, Santa Cruz, CA; p38b antibodies were a generous gift from J. Han, Scripps Institute, La Jolla, CA) with Pansorbin (Calbiochem, La Jolla, CA), cell lysate (100 g of protein) was added, and p38 kinase was immunoprecipitated. The immunoprecipitate was washed five times in lysis buffer, followed by five additional washes in kinase buffer. p38 kinase activity was determined by in vitro labeling of activating transcription factor 2 (Santa Cruz Biotechnology) with 32 P, in a buffer containing: 25 mM Hepes, pH 7.4, 25 mM ␤-glycerophosphate, 25 mM MgCl 2 , 2 mM dithiothreitol, 0.1 mM sodium vanadate, 25 M ATP. Following 20 min of incubation at 30°C, kinase reaction was stopped by 2-min centrifugation at 12,000 ϫ g. Supernatant was suspended in Laemmli sample buffer (25), and resolved on a 10% SDS-PAGE (polyacrylamide gel electrophoresis). Gels were dried and autoradiographs taken. JNK-1 activity was determined using 1-79 N-terminal amino acids of c-Jun (Santa Cruz Biotechnology) as substrate and anti-JNK-1 antibodies (26) (Santa Cruz Biotechnology) for immunoprecipitation. For ERK-1 and -2 assay, the above procedure was used except for the following. Anti-ERK antibodies (Upstate Biotechnology Inc.; the antibody cross-reacts with ERK-1 and ERK-2) were used for immunoprecipitation, protein A/G-agarose was used to immobilize the antibodies, myelin basic protein (Sigma) was used as substrate, and the reaction supernatant was resolved on 15% SDS-PAGE.
MAPKAP kinase-2 activity was determined as per Krump et al. (27) with slight modification; cell lysate was obtained as described above. MAPKAP kinase-2 was immunoprecipitated using anti-MAPKAP kinase 2 antibodies (Stress Gen. Biotechnologies Corp., Victoria, British Columbia, Canada), and protein A/G-agarose (Santa Cruz Biotechnology) was used to immobilize the antibody. Immunoprecipitate was suspended in 25 l of kinase buffer (20 mM MgCl 2 , 1 mM sodium vanadate, 5 mM NaF, 20 mM ␤-glycerophosphate, 20 mM p-nitrophenylphosphate, 2 mM dithiothreitol, 20 mM ATP, 5 Ci of [␥-32 P]ATP, 50 mM HEPES, pH 7.4), and kinase reaction was started by the addition of 3 g of HSP27 substrate (Stress Gen.). Following 20 min of incubation at 30°C, the reaction was stopped by 2-min centrifugation at 12,000 ϫ g. Supernatant was suspended in Laemmli sample buffer, boiled for 2 min, and analyzed on 15% SDS-PAGE. The pellet was processed for Western blotting as described below.
Protein Quantitation-p38 protein was quantified using a modification of the method of Laemmli et al. (25). Briefly, equal amounts of protein were loaded and run on 12% reducing SDS-PAGE. Proteins were transferred at room temperature, 250 mA, onto Hybond™-ECL membrane (Amersham Life Sciences) in a buffer containing 10% methanol, 0.03% SDS, 25 mM Tris, 52 mM glycine, pH 8.3. Blots were blocked for 15 min with 5% dried milk in TBST (20 mM Tris, pH 7.6, 137 mM NaCl, 0.05% Tween 20), and incubated for 45 min at room temperature, with 1:100 (1 g/ml) dilution of polyclonal rabbit anti-p38 antibody (Santa Cruz Biotechnology), in TBST containing 5% milk. After a 15min wash in TBST, blots were incubated with 1:1000 dilution of horseradish peroxidase-conjugated secondary antibody in TBST buffer containing 5% milk. Protein was visualized using ECL detection system (Amersham Life Sciences) and quantified by densitometric analysis of the resultant bands.
For quantitation of MAPKAPK-2, the pelete from the kinase assay (as described above) was suspended in Laemmli sample buffer (25), boiled for 2 min, and run on 12% reducing SDS-PAGE. Proteins were transferred at room temperature, 250 mA for 1 h, onto Hybond™-ECL membrane (Amersham Life Sciences) in a buffer containing, 10% methanol, 0.1% SDS, 25 mM Tris, 52 mM glycine, pH 8.3. Blots were blocked for 1 h with 5% dried milk in PBST (100 mM NaPO 4 , pH 7.5, 300 mM NaCl, 0.1% Tween 20), and incubated overnight at 4°C, with 2 g/ml polyclonal anti-rabbit MAPKAPK-2 (Upstate Biotechnology Inc.), in PBST containing 5% milk. After a 15-min wash in PBST, blots were incubated for 1 h at room temperature, with 1:1000 dilution of horseradish peroxidase-conjugated secondary antibody in PBST buffer containing 5% milk. Protein was visualized using an ECL detection system (Amersham Life Sciences) and quantified by densitometric analysis of the resultant bands.

SB203580 Inhibits Hypertonicity-induced Betaine
Transporter (BGT1) and HSP70 mRNAs-As mentioned above, hypertonicity shrinks cells and increases intracellular ionic strength. The rise in intracellular ionic strength is believed to be among the initial signals for induction of osmoprotective genes (2,3). To investigate the involvement of p38 kinase in mediating osmotically driven transcription of these genes, MDCK cells were exposed for 16 h to hypertonic medium, in the absence or presence of increasing concentrations of SB203580, a potent inhibitor of p38 kinase, and the abundance of BGT1 mRNA was measured. NaCl increases the abundance of BGT1 mRNA 8-fold above that of cells in isotonic medium (Fig. 1). SB203580 attenuates the hypertonic induction of the mRNA in a dose-dependent manner with IC 50 of 35-40 M and complete inhibition at 100 M.
In addition to induction of osmoprotective genes, hypertonicity was shown to induce heat shock proteins ␣B-crystallin and HSP70 (10,28). Therefore, we examined whether p38 kinase is involved in hypertonic induction of HSP70. Hypertonicity increases the abundance of HSP70 mRNA 6-fold above that of cells in isotonic medium (Fig. 2). As seen with BGT1, hypertonic induction of HSP70 mRNA is inhibited by SB203580 with similar IC 50 . We conclude that p38 kinase activity is essential for hypertonic induction of BGT1 mRNA, a representative gene involved in the accumulation of compatible organic solute be-taine in MDCK cells. In addition, p38 kinase activity appears to be essential for hypertonic induction of HSP70 mRNA as well. These findings suggest that p38 kinase is essential for the adaptation of kidney cells to osmotic stress, either by induction of heat shock proteins or of osmoprotective genes.
SB203580 Inhibits Native p38 Kinase Activity, but Not ERK-1 or -2 in Osmotically Stressed Cells-This inhibition correlates with up-regulation of native JNK-1 activity. The pyridinyl imidazole compounds to which SB203580 belongs demonstrate a highly specific and potent inhibitory activity against p38 kinase. In earlier studies, these compounds showed no inhibitory activity against ERK-1 and -2, JNK, MAPKactivated protein kinase-2 (MAPKAPK-2), MAPKK, protein kinase C, calmodulin-dependent protein kinase, or cAMP-dependent protein kinase (18,19). In fact, the binding specificity of these compounds was utilized to identify and affinity-purify p38 kinase (19). Recently, a second p38 kinase, p38b, was identified (18). Both kinases are equally inhibited by pyridinyl imidazole and display near identical response to tumor necrosis factor, interleukin-1, epidermal growth factor, phorbol-12-myristate-13-acetate, and extracellular stresses such as UV irradiation, H 2 O 2 , osmotic stress, and arsenate. Since the inhibitory effect of the emidazol compounds on p38 kinase is reversible (29), in vitro examination of p38 kinase activity may not reflect its in situ activity. To examine the specificity of SB203580 on the activity of p38 kinase in our system, MDCK cells were placed for 16 h in hypertonic medium in the presence of 0, 10, 25, 50, or 100 M SB203580, and cell lysates were prepared for determination of native MAPKAPK-2 activity. p38 kinase is the only known activator of MAPKAPK-2, and the activity of MAPKAPK-2 is dependent on its phosphorylation by p38 kinase (27). Therefore, the in vitro activity of immunoprecipitated MAPKAPK-2 reflects the in situ activity of p38 kinase. MAPKAPK-2 activity declines with increasing medium concentrations of SB203580, with an IC 50 of 35-40 M (Fig. 3,  A and B). The kinetics of MAPKAPK-2 activity is similar to the kinetics of hypertonic induction of HSP70 and BGT1 mRNAs in the presence of SB203580 (see Figs. 1 and 2). To exclude the possibility of change in MAPKAPK-2 or p38 kinase protein abundance, as a potential cause for the change in p38 kinase activity, Western blot analysis was performed on cell lysates. Cell treatment with SB203580 for 16 h does not affect the abundance of MAPKAPK-2 or p38 kinase protein (Fig. 3C). It is concluded that SB203580 exerts specific inhibitory effect on p38 kinase in MDCK cells. The decline in p38 kinase activity does not result from diminished protein. Rather, it results from inhibition of p38 kinase enzymatic activity. However, the concentration of SB203580 that is required for p38 kinase inhibition in situ in these cells is higher than that reported for hematopoietic cells (18,19,30). These data are consistent with a role for p38 kinase in the induction of BGT1 and HSP70 genes during osmotic stress.
Of interest, JNK-1 activity in the hypertonically stressed cells is equal to that of non-stressed cells. Inhibition of native p38 kinase activity in the hypertonically stressed cells correlates with 8 -10-fold increase in native JNK-1 activity (Fig. 4). No change is seen in the activities of ERK-1 and -2 with medium SB203580 concentrations up to 100 M (Fig. 4). The data suggest that, in hypertonically stressed cells, p38 kinase may have an inhibitory effect on JNK-1, either directly or indirectly.
SB203580 Does Not Affect Thermal Induction of HSP70 mRNA-It was previously reported that HSP70 is induced by the stress of hypertonicity, and accumulation of organic osmolytes by the cell attenuates the thermal and hypertonic induction of HSP70 mRNA (9). While heat shock proteins are induced in response to general stresses (ischemia, arsenate, heat stress, osmotic stress, heavy metals, alcohol, and amino acid analogues) (9,31,32), the osmoprotective genes, such as the betaine transporter, are specifically induced by osmotic stress (9). The inhibition of hypertonic induction of HSP70 and betaine transporter mRNAs by SB203580 presents the following question. Is the SB203580-sensitive pathway specific for induction of genes following osmotic stress, or does it mediate gene expression during exposure of kidney cells to general stresses? SB203580-treated (0, 50, and 100 M) MDCK cells were heat-shocked (42°C for up to 3 h), and HSP70 mRNA abundance was examined. As seen in Fig. 5, thermal induction of HSP70 mRNA is not affected by SB203580 at concentrations sufficient to block the hypertonic induction of HSP70 completely. These results suggest that HSP70 (and perhaps other heat shock proteins) induction is mediated through more than one pathway; hypertonic induction of HSP70 is mediated by an SB203580-sensitive pathway, similar to the pathway that me- diates hypertonic induction of betaine transporter mRNA, while thermal induction of HSP70 mRNA is not SB203580-sensitive.
p38 Kinase mRNA Is Induced by Hypertonic Stress; This Induction Is SB203580-sensitive-Regulatory proteins can modulate the abundance of their own mRNAs directly or indirectly (33). If a positive loop existed between p38 kinase activity and the regulation of its own mRNA, it may be possible to demonstrate down-regulation of p38 kinase mRNA with inhibition of kinase activity. MDCK cells were exposed for 16 h to hypertonic medium, in the absence or presence of increasing concentrations of SB203580. As seen in Fig. 6, a 4.2-kb band is detected with a p38-specific oligonucleotide probe. The mRNA is induced 1.5-2-fold by hypertonic stress. This induction is markedly attenuated (2.5-fold, relative to hypertonically stressed cells) in the presence of SB203580 at concentrations greater than 50 M. A 5-kb mRNA is detected with p38bspecific antisense oligonucleotide probe that behaves similarly (data not shown). We conclude that p38 is up-regulated by hypertonicity at the mRNA level and that p38 kinase positively regulates the abundance of its own mRNA. The decline in mRNA abundance without concomitant change in p38 kinase protein after 16 h of SB203580 treatment (as shown in Fig. 3C), suggests that the half-life of the protein may exceed 16 h. DISCUSSION Our findings demonstrate that p38 kinase activity is essential for the hypertonic induction of mRNAs for HSP70 and betaine transporter BGT1, but not for thermal induction of HSP70 mRNA. In addition, JNK-1 activity may not be required for HSP70 and BGT1 mRNAs induction under hypertonic conditions. These findings represent the first direct evidence linking p38 kinase to regulation of genes involved in the adaptation to osmotic stress in mammalian cells. Betaine transporter is a representative osmoprotective gene. HSP70 is a representative heat shock protein, one of the major heat shock proteins expressed in the kidney, and has been shown to play a major role in early stages of adaptation of kidney cells to osmotic stress (9). Both adaptive processes, the induction of heat shock proteins and the accumulation of compatible organic solutes, are essential for the survival of kidney cells in hyperosmotic environment. Hence, these findings not only are important for the understanding of the molecular physiology of the adaptation to osmotic stress in kidney cells, but also may offer clues to disease states involving solute and water handling by the kidney. In addition, these findings may provide insight into osmotic stress adaptation in the brain, since brain cells behave similar to kidney cells under hyperosmotic conditions (34). The low JNK-1 activity we observe after 16 h of exposure to hypertonic medium when p38 kinase is not inhibited is consistent with the low JNK-1 activity that was found in outer medulla slices of rat kidney after osmotic stress (35). The activation of JNK-1 concomitant with p38 inhibition is interesting and suggests that p38 kinase may have an inhibitory effect on JNK-1 in hypertonically stressed MDCK cells, either directly or indirectly. The significance of these findings remains to be determined.
Recent reports suggest involvement of JNK in apoptotic signals (36 -39); ERK inhibits, whereas JNK mediates, cytokineinduced apoptosis (36). Hypertonic stress induces cell cycle arrest (40,41), inhibits general mRNA and protein synthesis (40,41), and induces DNase I-hypersensitive sites (42), consistent with the existence of apoptotic signals during such stress. As kidney medulla cells are exposed to variable extracellular tonicity, the existence of an un-opposed apoptotic signal would obviously lead to cell death immediately upon exposure of the medulla to the first cycle of hypertonicity. To avoid uniform cell death, apoptosis must be regulated. Since ERK, JNK, and p38 kinases are induced during osmotic stress (12)(13)(14)(15), and ERK may not be involved in the induction of osmoprotective genes (16), it is proposed that the ERK pathway may mediate growth signals and opposes the apoptotic signals that are induced during osmotic stress. Whether JNK-1 mediates these apoptotic signals remains to be determined.
A recent report, based on inhibition of MKK3 (one of the upstream activators of p38 kinase) in rabbit papillary epithelial cells, suggested that p38 kinase may not be required for induction of osmoprotective genes in mammalian cells (46). Kidney cells express at least p38 and p38b (18). Both kinases are activated equally by stimuli such as cytokines and environmental stresses including osmotic stress, yet differ in their upstream activators; p38 kinase is activated in parallel by MKK3, MKK4, and MKK6, while p38␤ is activated predomi- nantly by MKK6 (18). Therefore, inhibition of MKK3, may not affect the function of p38␤ kinase. As these pathways might be redundant in function, if both kinases are not inhibited concomitantly, one might erroneously conclude that elements of p38 kinase pathway are not necessary for induction of osmoprotective genes in mammalian cells (see Scheme 1). Hence, the availability of a specific inhibitor of p38 kinases offers a screening tool for identification of relevant functions. The identity of the exact p38 kinase(s) responsible for osmoprotective genes induction remains to be determined.
The regulation of heat shock response is complex, and our results provide a hint to even greater complexity in its regulation. Heat shock is a well known activator of p38 kinase, which mediates phosphorylation of HSP25/27 (30). As p38 is not required for induction of HSP70 during heat stress, it is possible that thermal activation of p38 kinase serves distinct functions that may not be related to induction of HSP70. The finding of divergent pathways mediating thermal and osmotic induction of HSP70 is intriguing but not surprising. It is amply reasonable that induction of proteins that are required for the adaptive response to osmotic stress be mediated by a specific pathway if specificity is important. Since some of the proteins that are induced during osmotic stress, such as heat shock proteins, may also be required for other cellular functions, their regulation may require separate, function-specific pathways.