INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CELLS OF THE RENAL INNER medulla are continually exposed to an elevated concentration of both NaCl and urea in vivo. NaCl (11, 13, 22) and urea (5) both activate the extracellular signal-regulated kinase (ERK) mitogen-activated protein kinases (MAPKs) in renal epithelial cells, but in other respects the signaling pathways activated by these stimuli are distinct. Hypertonicity, acting through a putative tonicity-responsive transcription factor (17), activates transcription of genes whose protein products are involved in the synthesis or transport of organic osmolytes (reviewed in Ref. 3). Although little is known about the signaling events involved, this process has been reported to be either dependent on (9, 20) or independent of (12) activation of the stress-responsive p38 MAPK in different models. In contrast, urea, in cells of the renal inner medulla, activates hallmarks of a receptor tyrosine kinase-mediated pathway including transcription of immediate-early genes in an ERK- and Elk-1-dependent fashion (5, 6, 8) and activation of phospholipase C- with resultant generation of inositol 1,4,5-trisphosphate (7). The ability of these two disparate signaling pathways to converge upon ERK activation prompted us to investigate the mechanism of urea- and hypertonicity-inducible ERK activation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture and treatment. mIMCD3 inner medullary collecting duct cells were grown and passaged as previously described. mDCT distal convoluted tubule (kindly supplied by P. Friedman, Dartmouth Medical School), MDCK, HEK293, COS-7, and NIH3T3 cells were maintained in DMEM-F12 supplemented with 10% fetal bovine serum and passaged after confluence. Supplementation of medium with hyperosmotic stressors was performed in situ through addition of an aliquot of concentrated solute in sterile water. Additional control experiments were performed via complete medium change to medium that had previously been supplemented with the relevant solute. For experiments combining addition of solute with dilution of the hyperosmotic medium (see Fig. 12), appropriate volumes of sterile water were premixed with added solute (sorbitol or urea) prior to addition to the well such that both stimuli were applied simultaneously. Pretreatment with PD-98059 and SB-203580 (at the indicated concentrations) was performed for 30 min.

Immunoblotting. Whole cell detergent lysates were prepared and subjected to electrophoresis as previously described (6). Anti-phospho-ERK, anti-phospho-p38, anti-phospho-MEK1/2, anti-phospho-MKK4, anti-phospho-MKK3/6 (all from New England BioLabs), anti-MEK1 (Santa Cruz Laboratories), and anti-MEK2 (Santa Cruz Laboratories) antibodies were used for immunoblotting or immunoprecipitation according to the manufacturer's directions. MEK1 and MEK2 were discriminated on the basis of approximate molecular mass; however, control experiments were also performed with immunoprecipitation with isoform-specific antibody followed by phospho-specific antibody immunoblotting to confirm the identity of the differentially migrating bands. In the short time courses examined for the present study, there was no appreciable change in total ERK or MAPK/ERK kinase (MEK) abundance as assessed by anti-ERK and anti-MEK immunoblotting in the setting of urea or NaCl treatment. On the basis of molecular mass standards, immunoreactive bands migrated at the following approximate molecular masses: anti-P-ERK, 42 and 44 kDa; anti-P-MEK, doublet at 57 kDa; anti-P-p38, 40 kDa; anti-P-MKK4, 51 kDa; and anti-P-MKK3/MKK6, 43 and 45 kDa (identity of MKK3 and MKK6 was assigned on the basis of relative molecular mass). Unless indicated, depicted immunoblots are representative of at least two separate experiments.

Where data are depicted graphically, immunoblots from between three and seven separate experiments were quantitated via scanning densitometry using an Epson ES-1000C scanner and Image software (National Institutes of Health). For experiments describing the effects of PD-98059 upon urea- and NaCl-inducible ERK and MEK phosphorylation (see Figs. 4 and 7), 8 µg of protein from NaCl-treated cells and 4 µg of protein from urea-treated cells was used in an effort to generate approximately equivalent maximal signal intensity and thereby minimize confounding by film sensitivity and linearity. For determination of MEK isoform phosphorylation, area under the complex curve generated densitometrically was mathematically reduced into peaks corresponding to MEK1 and MEK2 phosphorylation. In all lanes examined, anti-P-MEK immunoreactivity corresponding to MEK1 exceeded that of MEK2 (e.g., Fig. 7B). On the basis of approximate interpeak distance (distance between MEK1 and MEK2 arrowheads in Fig. 7, A and B) and with the assumption of a near-normal distribution for each peak, it was estimated that the contribution of MEK2 anti-P-MEK immunoreactivity to the principal peak (corresponding primarily to MEK1) was relatively small (see below). Therefore, a vertical line was scribed from the apex of the MEK1 peak (maximum y value) perpendicular to the baseline (x-axis), separating the area under the curve into components A and B (see Fig. 7B). The degree of MEK1 phosphorylation was approximated as (2 × B) and that for MEK2 phosphorylation, A  B. Although this method potentially introduced a modest but consistent inaccuracy (a very minor component of the area of component B was potentially contributed by MEK2), it was superior to arbitrarily dividing the complex curve into components corresponding to the MEK1 peak and the MEK2 peak, particularly at relatively low levels of MEK2 phosphorylation.

MEK activity. MEK activity was assessed via immune complex kinase assay as previously described (21). Briefly, mIMCD3 monolayers were washed with ice-cold PBS, lysed in situ with lysis buffer (in mM: 50 HEPES, 150 NaCl, 1.5 MgCl₂, 1 EGTA, 10 NaF, pH 7.5, 800 µl/P100 dish) supplemented with 10% glycerol, 1% Triton X-100, 1% phenylmethylsulfonyl fluoride, and 10 µg/ml aprotinin for 15 min at 0°C. Lysate was scraped from the plate, centrifuged at 4°C for 5 min in a microcentrifuge to pellet insoluble debris, and subjected to immunoprecipitation with anti-MEK1 or -MEK2 (2 µg/sample) and Protein A/G beads (25 µl/sample). Immunoprecipitates were washed twice with lysis buffer and once with incomplete kinase buffer (in mM: 50 Tris, pH 7.4, 10 MgCl₂, and 1 dithiothreitol), and then resuspended with 20 µl of complete kinase buffer [incomplete kinase buffer + 50 µM ATP + kinase-inactive ERK substrate (Santa Cruz Laboratories; 1 µg/sample) + [-³²P]ATP (1/40 volume)]. Kinase reaction was performed at 30°C for 40 min and terminated with one volume of 2× Laemmli sample buffer. Radiolabeled immunoprecipitates were resolved via SDS-PAGE and exposed for autoradiography. Densitometry was quantitated as described above.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Urea activates ERKs more potently and earlier than does NaCl in mIMCD3 cells. We have shown that urea activates Egr-1 transcription in a largely ERK-dependent fashion in renal medullary cells (5). To determine the relevance of this effect at lower concentrations of urea, dose-response studies were performed. Urea in concentrations as low as 20 mM activated ERK (p44 and p42), as measured by phospho-specific antibody immunoblotting (Fig. 1). The half-maximal effect occurred at ~40 mM (data not shown), consistent with that observed in the context of urea-inducible immediate early gene transcription (8). The effect of urea (200 mM) upon ERK activation was maximal at 5 and 10 min and was essentially undetectable by 60 min (Fig. 2). In contrast, treatment of cells with an equiosmolar dose of NaCl (100 mM) revealed minimal activation at 5 min. NaCl exerted its maximal effect at 15 min of treatment, and this degree of activation was considerably less than that of urea. NaCl-inducible ERK activation was still evident at 120 min, the last time point examined.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Dose response of urea effect upon extracellular signal-regulated kinase (ERK) (p44 and p42) phosphorylation. Anti-phospho-ERK immunoblot of detergent lysates of mIMCD3 inner medullary collecting duct cell monolayers treated with the indicated concentration of urea for 10 min. C, control.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Time course of urea- and NaCl-inducible ERK phosphorylation. Anti-phospho-ERK immunoblot of detergent lysates of mIMCD3 monolayers treated for the indicated intervals with equiosmolar urea (200 mM) or NaCl (100 mM).

Urea-inducible ERK activation is not universal. In marked contrast to observations with mIMCD3 cells, urea and NaCl (each at 200 mosmol/kgH₂O) failed to activate ERK in two renal nonmedullary cell lines, HEK293 and COS-7 cells (Fig. 3). These cell lines were originally examined because of the readiness with which they may be transfected and with the expectation that they might provide a tractable model for examining urea-inducible signaling. An extremely modest response to urea in the murine mDCT renal cell line could not be excluded. Only in the MDCK cell line was a robust response to urea observed. In the nonrenal NIH3T3 cell line, urea induction was absent. In all cell lines examined except for mIMCD3 and MDCK, however, the effect of phorbol ester (a positive control for ERK activation) markedly exceeded that of urea (Fig. 3). Therefore, in only these two lines was a potent activation in response to urea observed.

View larger version (65K): [in this window] [in a new window]   Fig. 3.   Cell type specificity of urea effect. Effect of no treatment (C), urea (U, 200 mM), NaCl (100 mM), or 12-O-tetradecanoylphorbol-13-acetate (TPA, 100 nM) for 10 min upon ERK phosphorylation in mIMCD3, MDCK, mDCT, HEK293, COS-7, and 3T3 cells.

Differential sensitivity of urea- and NaCl-inducible ERK phosphorylation to PD-98059. Because of the discrepant kinetics between urea- and NaCl-inducible ERK activation, the role of upstream ERK activators was examined. It was observed that NaCl was less potent than urea at activating ERK phosphorylation; therefore, approximately equipotent doses of urea (200 mM) and NaCl (200 mM/400 mosmol/kgH₂O) were established densitometrically and used for further comparison. At each given time point of solute treatment, urea treatment was much less sensitive to the MEK inhibitor PD-98059 than was NaCl treatment (Fig. 4). At 5 min of NaCl treatment, the IC₅₀ for PD-98059 was ~1.3 µM, whereas at 5 min of urea treatment it was ~20 µM. Similarly, at 15 min of NaCl treatment, the IC₅₀ for PD-98059 was ~8 µM, whereas the effect of even 100 µM PD-98059 upon urea-inducible ERK activation was modest (25% inhibition). There was a tendency for very high concentrations of PD-98059 (100 µM) to activate ERK in other treatment groups. This differential PD-98059 sensitivity of urea- and NaCl-inducible ERK phosphorylation was also evident when urea and NaCl of equivalent osmolarity were applied (Fig. 4C). Hypotonicity-inducible ERK phosphorylation was also highly sensitive to PD-98059, analogous with hypertonicity (data not shown). There appeared to be lot-to-lot variability in the potency of PD-98059; for each experiment, the effects of PD-98059 upon both NaCl and urea were evaluated in parallel.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Differential sensitivity of urea- and NaCl-inducible ERK activation to PD-98059. Inhibitory effect of 30-min pretreatment with increasing concentration of PD-98059 upon ERK phosphorylation in response to urea treatment (200 mM for 15 or 5 min) or NaCl treatment (200 mM/400 mosmol/kgH2O for 15 or 5 min) in mIMCD3 cells. Representative anti-phospho-ERK immunoblots are depicted (A), as are pooled densitometric data from 4-7 independent experiments (B). In C, the effect of either 200 or 400 mosmol/kgH2O urea or NaCl (for 15 min), in presence (+) or absence () of pretreatment with PD-98059 (30 µM), is depicted with respect to ERK phosphorylation.

Differential MEK activation. The only PD-98059-sensitive mechanism of ERK activation described involves the ERK activators, MEK1 and MEK2. Interestingly, MEK2 is reported to be ~25-fold less sensitive to PD-98059 than MEK1 (1). It was therefore initially hypothesized that urea-inducible ERK activation would be more highly MEK2 dependent than NaCl-inducible ERK activation. To examine this hypothesis, the ability of urea and NaCl to activate MEK isoforms was evaluated via immune complex kinase assay. Anti-MEK1 or anti-MEK2 immunoprecipitates prepared from control and solute-treated mIMCD3 cells were assayed for ERK-directed (MEK) kinase activity. Although abundant MEK activity was evident in anti-MEK1 immunoprecipitates, MEK1 activity was only increased by ~25% following both urea and NaCl treatment (Fig. 5A). This was a consequence of the high basal level of MEK-like kinase activity evident in the anti-MEK1 immunoprecipitates under control conditions (Fig. 5B), but not evident in anti-P-MEK immunoblots (see below). ERK-directed (MEK-like) kinase activity in the anti-MEK2 immunoprecipitates was in general less than that of the anti-MEK1 immunoprecipitates (Fig. 5B) but was increased fourfold (relative to control) in response to urea treatment and threefold in response to NaCl treatment. (Less constitutive ERK-directed kinase activity was evident in the anti-MEK2 immunoprecipitates.) This modest differential effect between urea and NaCl was preserved in all experiments; however, the magnitude of this difference was highly variable between experiments and did not achieve statistical significance. Although the differential PD-98059 sensitivity of urea- and NaCl-inducible ERK activation did not appear to be a consequence of differential activation of MEK isoforms, the relative activities of MEK2 with respect to MEK1 could not be established in this native system because of possible differential affinity of the respective antibodies influencing efficiency of immunoprecipitation.

View larger version (34K): [in this window] [in a new window]   Fig. 5.   MEK activation by urea and NaCl. A: effect of urea (200 mM) and NaCl (200 mM/400 mosmol/kgH2O) upon MEK1 and MEK2 activity as quantitated by ERK-directed (MEK) kinase activity in immune complex kinase assay following anti-MEK1 or anti-MEK2 immunoprecipitation; n = 4 separate experiments. * P < 0.05 with respect to control treatment for each isoform. B: autoradiograph from representative kinase assay of lysates prepared from control, urea-treated (U), and NaCl-treated (N) cells and immunoprecipitated with anti-MEK1 (lanes 1-3) or anti-MEK2 (lanes 4-6). Solid arrowhead, phosphorylated MEK substrate (kinase-inactive ERK).

To address this issue, the relative efficacy of the anti-MEK antibodies in immunoprecipitating their respective targets was evaluated. The anti-P-MEK immunoblotting model system employed also provided correlative data regarding the relative magnitudes of solute-inducible MEK activation. Under control conditions, no phosphorylated MEK was detectable in mIMCD3 cell lysates (Fig. 6, lane 7). Following both urea and NaCl treatment, an immunoreactive doublet was evident (Fig. 6, lanes 8 and 9). The absence of detectable signal under control conditions prevented determination of multifold inductions for both MEK1 and MEK2 phosphorylation. Although it appeared that de novo MEK1 phosphorylation exceeded that of MEK2 following treatment with either solute, equal avidity of the antigen-antibody (anti-P-MEK) interaction for each isoform could not be assumed. Lysates from mIMCD3 cells were then immunoprecipitated with either anti-MEK1 (Fig. 6, top) or anti-MEK2 (Fig. 6, bottom) antibody, and the resultant immunoprecipitates as well as the immunodepleted lysates were examined by anti-P-MEK immunoblotting and compared with total (naive) lysate. Immunoprecipitation with anti-MEK1 substantially immunodepleted the more rapidly migrating band of the doublet and markedly enriched this species in the immunoprecipitate. Immunoprecipitation with anti-MEK2 antibody was relatively ineffective at immunodepleting the more slowly migrating band; however, it specifically resulted in the appearance of only this species in the immunoprecipitates. It was concluded from these studies that: 1) urea and NaCl increased phosphorylation of both MEK1 and MEK2 and to a greater extent than would be anticipated by the immune complex kinase assay; 2) MEK1 represented the more rapidly and MEK2 the more slowly migrating components of the doublet (consistent with the slightly greater molecular mass of the latter); and 3) anti-MEK1 immunoprecipitated MEK1 specifically and almost quantitatively, whereas anti-MEK2 immunoprecipitated MEK2 specifically but not quantitatively.

View larger version (48K): [in this window] [in a new window]   Fig. 6.   MEK phosphorylation. Anti-phospho-MEK1/2 immunoblot of whole cell detergent lysates prepared from control-treated cells (C) or cells treated with urea (U; 200 mM for 5 min) or NaCl (N; 200 mM/400 mosmol/kgH2O for 5 min). Replicate treated wells were immunoprecipitated with anti-MEK1 (top) or anti-MEK2 (bottom) prior to resolution for anti-phospho-MEK immunoblotting. Following immunoprecipitation, immunodepleted (lanes 1-3) and immunoprecipitated (lanes 4-6) fractions from each condition were electrophoresed for comparison with each other and total (naive) lysate (lanes 7-9). Equal microgram quantities of total lysate and immunodepleted lysate were used for comparison; immunoprecipitates could not be similarly controlled, because of enrichment of the precipitating antibody. IP, antibody used for immunoprecipitation. Solid arrowhead denotes immunoglobulin light chain (nonspecific band); shaded arrowhead denotes MEK2; and open arrowhead denotes MEK1.

The ability of urea and NaCl to activate MEK isoforms was examined in the presence of the MEK inhibitor, PD-98059. The specific mechanism of action of PD-98059 has been the subject of controversy; although described as a MEK inhibitor, other data have suggested that it may function as a direct inhibitor of Raf action. We therefore sought to determine whether the differential PD-98059 sensitivity observed with ERK activation would be recapitulated at the level of MEK phosphorylation. Phospho-MEK1 and phospho-MEK2 nearly comigrated, as is evident in Fig. 6. When abundance of both phosphorylated isoforms was considered in aggregate and integrated densitometrically, urea-inducible MEK phosphorylation was substantially less PD-98059-sensitive than was NaCl-inducible MEK phosphorylation (e.g., Fig. 7A). We employed a simple deconvolution method to reduce the complex curve generated after densitometric scanning of the phospho-MEK doublet, thereby permitting evaluation of the effect of PD-98059 upon both MEK1 and MEK2 phosphorylation (Fig. 7B; see METHODS). The identity of the upper (more slowly migrating) band as MEK2 had previously been established via immunoprecipitation and immunodepletion analysis (Fig. 6). Both urea-inducible MEK1 (Fig. 7C) and urea-inducible MEK2 (Fig. 7D) phosphorylation were substantially less sensitive to PD-98059 than were the corresponding phosphorylation events induced by hypertonicity (NaCl). As little as 1 µM PD-98059 significantly reduced NaCl-inducible MEK1 and MEK2 phosphorylation. This differential PD-98059 sensitivity paralleled that observed with ERK phosphorylation. In addition, neither MEK isoform was markedly inhibited by PD-98059 in the context of urea treatment, mitigating against a contribution of differential MEK isoform activation in response to urea vs. NaCl.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Effect of PD-98059 upon MEK phosphorylation. A: representative anti-P-MEK immunoblot of mIMCD3 cells pretreated with PD-98059 (0-30 µM) prior to urea (200 mM for 5 min) or NaCl (200 mM/400 mosmol/kgH2O for 5 min). B: example of complex ("double-peaked") curve generated by densitometric analysis of the anti-P-MEK immunoreactive doublet, comprised of superimposed P-MEK1 (open arrowhead) and P-MEK2 (shaded arrowhead) peaks. Vertical line, scribed from the apex of the primary (MEK1) peak to the baseline, divides the area under the curve into components A and B, which are used to estimate areas under the P-MEK1 and P-MEK2 curves according to the indicated formulas (see METHODS). C and D: effect of pretreatment with indicated concentrations of PD-98059 upon urea-inducible and NaCl-inducible phosphorylation of MEK1 (C) and MEK2 (D).

To further substantiate this interpretation, the effect of the newly described MEK inhibitor, U-0126 (10), was examined. U-0126 is at least an order of magnitude more potent than PD-98059 and, more importantly, does not appear to exhibit appreciable isoform selectivity with respect to MEK1 and MEK2 [IC₅₀ (in vitro) of 0.07 and 0.06 µM, respectively] (10). Similar to data acquired with PD-98059, urea-inducible ERK activation was far less sensitive to this compound than was NaCl-inducible ERK activation (Fig. 8). It therefore appeared highly unlikely that the differential PD-98059 sensitivity of urea- and NaCl-inducible ERK activation was a consequence of selective activation of either MEK1 or MEK2. The ability of urea to activate other MAPK kinases (MAPKKs) was then evaluated.

View larger version (30K): [in this window] [in a new window]   Fig. 8.   Effect of the non-isoform-selective MEK inhibitor, U-0126, upon solute-inducible ERK activation. Anti-P-ERK immunoblot of lysates prepared from cells pretreated for 30 min with 0, 0.3, 1, or 10 µM U-0126 prior to the indicated solute treatments, demonstrating differential sensitivity.

Phosphorylation of other MAPKKs. MAPKKs for the non-ERK members of the MAPK family include MKK4/SEK1, the principal SAPK/JNK activator, and MKK3 and MKK6, principal MAPKKs for p38. The ability of urea and NaCl to induce phosphorylation of each of these kinases was investigated as a correlate of the effect of these solutes upon MEK1 and MEK2 activation. MKK6 activation was detectable under control conditions and was modestly activated by NaCl treatment (Fig. 9); activation by NaCl was more evident at 30 min than at 5 min. The effect of urea, on an equiosmolar basis, was less pronounced than that of NaCl. MKK3 phosphorylation was absent under control conditions and, in general, paralleled that of MKK6. Urea-inducible MKK3 phosphorylation could not be demonstrated. MKK4, in contrast, exhibited readily detectable phosphorylation under basal conditions and was not reproducibly activated in the present context by any of these conditions. Treatment with higher urea concentrations decreased MKK4 phosphorylation.

View larger version (44K): [in this window] [in a new window]   Fig. 9.   Effect of urea and NaCl upon other MAPKKs. Anti-phospho-MKK3/MKK6 and anti-phospho-MKK4 (SEK1) immunoblots of lysates prepared from mIMCD3 cells treated for the indicated durations with the indicated concentrations of NaCl or urea.

Effect of the p38 inhibitor, SB-203580. A possible role for another PD-98059-insensitive kinase in urea signaling to ERK activation was suggested by the above data. Recently p38 has been implicated in stress signaling to ERK activation (16), as well as stress signaling to immediate-early gene expression (15). In addition, an interrelationship between p38 signaling and JNK signaling has also been suggested (20). Therefore, the ability of the p38 inhibitor, SB-203580 (14), to abrogate the urea effect was examined using immunoblots detecting ERK and MEK phosphorylation. SB-203580 (50 µM) substantially inhibited the ability of NaCl to activate this signaling cascade (Fig. 10). This is within the dose range reported to be required for inhibition of p38 effect in other cells of renal epithelial origin (20). In contrast, the effect of SB-203580 upon urea signaling was much more modest at the levels of both MEK1/2 and ERK phosphorylation. We and others have shown that the effect of urea upon p38 activation is modest at best, compared with that of NaCl, in in vitro kinase assay (2, 23). We examined p38 phosphorylation as a potentially more sensitive index of activation and confirmed the observations of Berl et al. (2). Specifically, NaCl markedly increased p38 phosphorylation as early as 5 min and in concentrations as low as 100 mM (the lowest concentration examined; Fig. 11). The effect of NaCl persisted until at least 30 min of treatment. For a given degree of hyperosmolarity, urea treatment exerted a much more modest effect than did NaCl upon p38 phosphorylation, and this effect was most pronounced at 5 min rather than 30 min of treatment.

View larger version (42K): [in this window] [in a new window]   Fig. 10.   Effect of SB-203580 upon ERK and MEK phosphorylation. Effect of urea (200 mM for 5 and 15 min) and NaCl (200 mM/400 mosmol/kgH2O for 5 and 15 min) on ERK and MEK phosphorylation as determined by anti-phospho-ERK and anti-phospho-MEK immunoblotting, respectively, in absence or presence of pretreatment with SB-203580 (50 µM for 30 min).

View larger version (25K): [in this window] [in a new window]   Fig. 11.   Urea- and NaCl-inducible p38 phosphorylation. Effect of indicated solute treatments on p38 phosphorylation in anti-phospho-p38 immunoblotting.

Effect of urea upon ERK activation is likely volume independent. Hypertonicity-inducible ERK activation is presumed to be mediated through resultant cell shrinkage. To investigate a role for cell volume in urea-inducible vs. sorbitol-inducible ERK activation, mIMCD3 cells were exposed to 100 mM solute for 10 min. In parallel wells, medium was diluted to achieve the indicated intermediate osmolarities (375, 350, 325, and 300 mosmol/kgH₂O; see Fig. 12), whereas concentration of the supplemented solute was maintained at 100 mM. (Sorbitol rather than NaCl was used as the hypertonic stressor because NaCl constituted the principal osmotically active solute in the medium, and hence, dilution of applied hypertonic NaCl would essentially restore the medium to near its original composition.) As anticipated, hypertonic sorbitol (100 mM) enhanced ERK phosphorylation at 10 min of treatment. This effect was inhibited by normalization of final osmolality, consistent with an effect dependent upon cell volume. In marked contrast, the effect of urea treatment was independent of final medium osmolality. Because urea inclusion in isosmotic media can result in cell swelling, and because we and others have shown that cell swelling can independently result in ERK activation, as an additional control, we pharmacologically assessed the potential contribution of cell swelling to the invariant ERK activation observed in the presence of urea and the various final medium osmolarities depicted in Fig. 10. As described above, cell swelling-inducible ERK phosphorylation, like NaCl-inducible ERK phosphorylation (Fig. 4), but in contrast to urea-inducible ERK phosphorylation, is highly sensitive to PD-98059. The effect of PD-98059 upon ERK activation in the presence of urea treatment was independent of the degree of normalization of medium osmolarity. Under all conditions, ERK phosphorylation was equivalently inhibited (data not shown). Had urea-inducible ERK phosphorylation been mediated under any of these conditions of diminishing total osmolarity by a component of cell swelling, then these conditions should have exhibited a much more pronounced sensitivity to PD-98059. Therefore, urea-inducible ERK phosphorylation is likely not mediated by cell shrinkage, and the inability of normalization of medium osmolarity to abrogate the effect of urea is likely not attributable to a component of swelling-inducible ERK activation.

View larger version (23K): [in this window] [in a new window]   Fig. 12.   Isotonic exposure to solutes. Effect of medium supplementation (for 10 min) with sorbitol (100 mM) or urea (100 mM) in absence ("400") or presence of medium dilution with sterile water to achieve the indicated final osmolarity ("375," "350," "325," "300"). Final osmolarities (estimated in mosmol/kgH2O) are depicted above the gel.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We sought to investigate the mechanism underlying urea-inducible ERK activation. The unusual ability of cells of the inner medulla to respond to urea treatment with ERK activation may reflect a uniquely differentiated phenotype acquired by cells perpetually bathed in urea in vivo. This effect of urea exceeded that of hypertonicity and even the potent ERK activator and positive control, phorbol ester. Other data have suggested the presence of a specific urea-inducible signaling cascade in cells of this lineage including the ability of urea to activate immediate-early gene transcription (6, 8) and hallmarks of receptor tyrosine kinase activation (7).

An additional striking finding is the marked differential sensitivity of urea- and NaCl-inducible ERK phosphorylation to the MEK inhibitor, PD-98059. This differential sensitivity could be consistent with differential activation of each of the two principal ERK kinases, MEK1 and MEK2, whose IC₅₀ values with respect to this inhibitor are reportedly 2 µM and 50 µM, respectively (1). This possibility was investigated in detail. The following lines of evidence mitigate against this explanation: 1) although both urea and NaCl treatment increased MEK2 activity to a greater extent than MEK1 activity in immune complex kinase assays (as a consequence of the high basal level of ERK-directed kinase activity in the anti-MEK1 immunoprecipitates), MEK1 was likely the principal operative kinase in response to either solute in both immune complex kinase assay (Fig. 5) and anti-P-MEK immunoblotting (Fig. 6); 2) phosphorylation of both MEK1 and MEK2 was inhibited to a much lesser extent by PD-98059 in the context of urea treatment than in the context of NaCl treatment (Fig. 7); and 3) urea-inducible ERK activation was also substantially less sensitive to the non-isoform-specific MEK inhibitor, U-0126, than was NaCl-inducible ERK activation (Fig. 8). Activation of a coimmunoprecipitating phosphatase activity (e.g., Ref. 4) or kinase activity could be differentially regulated by the two solutes, influencing the results of the immune complex kinase assay. Supporting this interpretation is the high basal level of ERK-directed kinase activity evident in anti-MEK1 immunoprecipitates prepared from control-treated cells (Fig. 5), when contrasted with the complete absence of basal MEK1 or MEK2 phosphorylation observed in the setting of anti-P-MEK immunoblotting (e.g., Fig. 6). Alternatively, a known or novel non-MEK1/2 upstream ERK activator may be preferentially activated by urea. An examination of the other principal MAPKK activities (MKK3, MKK4, and MKK6), however, failed to identify such a candidate. Regardless of the molecular mechanism, the significant differential PD-98059 (and U-0126) sensitivity may have implications for physiological responses to this and related pharmacological agents and represents, to our knowledge, the first report wherein relatively PD-98059-sensitive and -insensitive ERK activators are described in parallel in the same model.

Data in Fig. 4 clearly indicate that, for a given time point, urea-inducible ERK activation is considerably less sensitive to PD-98059 than is NaCl-inducible ERK activation. Although the data in Fig. 4 appear to depict dissimilar IC₅₀ values for PD-98059 for different solute treatments, an alternative explanation may also be operative. PD-98059 exhibits a reasonably "sigmoidal" inhibitory dose-response in the contexts of urea treatment at 5 min and NaCl treatment at both 5 and 15 min, therefore the apparent IC₅₀ values (concentration of PD-98059 corresponding to a 50% inhibition in ERK activation/phosphorylation) may approximate the true IC₅₀ for this inhibitor. In the case of urea treatment for 15 min, in contrast, no amount of PD-98059 inhibits ERK activation greater than 25%; therefore, IC₅₀ is difficult to interpret. If both PD-98059-sensitive and PD-98059-insensitive mechanisms of ERK activation are operative simultaneously, then the IC₅₀ will be overestimated; no amount of inhibitor will produce complete inhibition. Only the effect of NaCl at 5 min of treatment is fully inhibitable by PD-98059. It also remains possible that the relatively MEK-independent pathway of ERK activation in response to urea is an alternate, or redundant, pathway only operative in the context of MEK1 and MEK2 inhibition; this possibility could not be experimentally addressed. Regardless of the interpretation, however, these data suggest fundamental differences in the mechanisms of urea and NaCl signaling to ERK activation. In addition, in the case of curves U200/5' and N400/15' (Fig. 4B), the dose-response relationship may also be confounded by the potentially activating effect of high doses (100 µM) of PD-98059. Interestingly, for a given solute stimulus, the PD-98059 sensitivity of solute-inducible ERK activation is time dependent (Fig. 4B). Whether this reflects differential activation of upstream signaling events or, less likely, inhibitor inactivation in solution or intracellularly remains unclear.

The inability to quantitatively establish the relative contributions of MEK1 and MEK2 in urea-inducible ERK activation in this native system precipitated an investigation of other potential upstream activating events. A recent report described the unexpected ability of p38 to function upstream of ERK activation in a specific stress-responsive model (16). Similarly, we show here that NaCl-inducible, but not urea-inducible, ERK phosphorylation is highly sensitive to pharmacological inhibition of p38. In contrast to the observations of Ludwig et al. (16), however, p38-dependent ERK activation in the present model required only 5 min, whereas it required at least 60 min in the model of arsenite stress. In addition, a role for p38 has been suggested in stress (e.g., arsenite) induction of immediate-early gene expression, based largely upon inhibitor and overexpression data (15). In this model, as well, however, stress-inducible p38 activation required at least 15-30 min to become evident (15). For p38 activation to precede ERK activation in the model of urea stress, it would need to be evident by 5 min of treatment, when ERK activation is maximal. This is precisely what is observed in the present model (Fig. 11). Interestingly, p38 has also recently been implicated in some renal models in the downstream events initiated by hypertonic (but not urea-associated) stress including expression of genes encoding heat shock proteins and an osmolyte transporter (9, 20). Originally, we reported that urea treatment failed to appreciably induce p38 activation in mIMCD3 cells. Although modest upregulation was detectable at only 5 min of urea treatment, this increase failed to achieve statistical significance (23). Berl et al. (2) reported that urea modestly but reproducibly increased p38 activity, relative to NaCl in this cell line. Using the more sensitive anti-phospho-p38 immunoblot assay, we confirmed their observation. Nonetheless, the ability of urea to activate this kinase is much more modest than that of hypertonic stressors.

We present evidence suggesting that cell shrinkage does not mediate the effect of urea upon ERK activation. Whereas the ability of hypertonic sorbitol to induce ERK phosphorylation could be completely abolished by concomitant medium dilution, the inability of medium dilution to abrogate the effect of hyperosmotic urea upon ERK activation argues against a volume-mediated component. Cell swelling, like cell shrinkage, may result in ERK activation (18, 19). Inclusion of urea in isosmotic medium may induce cell swelling. Therefore, it could not be excluded that a small component of the urea effect is attributable to a change in volume. Nonetheless, among the broad range of medium dilutions employed in the present study, none influenced the effect of urea upon ERK activation. In addition, swelling-inducible ERK activation, like that induced by NaCl, is far more sensitive to PD-98059 than is urea-inducible ERK activation. At none of the intermediate dilutions depicted in Fig. 12 was the degree of ERK phosphorylation more sensitive to PD-98059 than was urea treatment alone. Therefore, it is highly unlikely that swelling-associated ERK phosphorylation accounted for the inability of medium dilution to abrogate the effect of urea.

In aggregate, these studies indicate that among cell lines examined, urea potently induces ERK phosphorylation only in the mIMCD3 and MDCK cell lines. In addition, this effect is accompanied by enhanced MEK1 and MEK2 activation, and is likely independent of changes in cell volume. The relative PD-98059 insensitivity of urea-inducible (in contrast to NaCl-inducible) ERK activation is not a consequence of differential MEK1/2 isoform activation.