Although urea is considered to be a cell stressor even in renal medullary cells perpetually exposed to this solute in vivo by virtue of the renal concentrating mechanism, aspects of urea signaling resemble that of a peptide mitogen. Urea was compared with epidermal growth factor and hypertonic NaCl or hypertonic mannitol using a large-scale expression array-based approach. The expression profile in response to urea stress more closely resembled that of EGF treatment than hypertonic stress, as determined by hierarchical cluster analysis; the effect of urea+NaCl was equidistant from that of either solute applied individually. Among the most highly urea- and hypertonicity-responsive transcripts were genes that had previously been shown to be responsive to these solutes, validating this approach. Increased expression of the activating transcription factor 3 by urea was newly detected via expression array and confirmed via immunoblot analysis. Earlier, we noted an abrogation of tonicity-dependent gene regulation by urea, primarily in a transient transfection-based model (Tian W and Cohen DM. Am J Physiol Renal Physiol 280: F904–F912, 2001). Here we applied K-means cluster analysis to demonstrate that the genes most profoundly up- or downregulated by hypertonic stress were partially restored toward basal levels in the presence of urea pretreatment. These global expression data are consistent with our earlier biochemical studies suggesting that urea affords cytoprotection in this context. In the aggregate, these data strongly support the hypothesis that the urea effect in renal medullary cells resembles that of a peptide mitogen in terms of the adaptive program of gene expression and in terms of cytoprotection from hypertonicity.
- expression array
- tonicity enhancer binding protein
- activating transcription factor 3
- mIMCD3 cells
- epidermal growth factor
cells of the mammalian kidney medulla are exposed to both hypertonic stress and urea stress by virtue of the renal concentrating mechanism. Hypertonic stress upregulates expression of a variety of gene products, including those coding for enzymes involved in osmolyte synthesis and transport proteins responsible for cellular uptake of osmolytes or their precursors (14). The best studied mechanism of hypertonic gene regulation requires the tonicity enhancer element and its interaction with an osmotic stress-responsive member of the NFAT family (14), tonicity enhancer binding protein (TonEBP). Hypertonic stress also upregulates expression of molecular chaperones of the heat shock protein (HSP) family (8, 29), perhaps through the same mechanism (27).
Urea exposure resembles a cellular stressor in some respects, although urea reputedly is readily membrane permeant and hence does not engender protracted changes in cell volume (12). Urea modestly activates the stress-responsive p38 and jun NH2-terminal kinase (JNK) mitogen-activated protein kinases (MAPKs) (1,37). In addition, urea treatment is associated with increased oxidative stress (40) and the oxidative stress-dependent expression of the stress-responsive transcription factor Gadd153 (40) and the cytoprotective enzyme heme oxygenase-1 (30). Furthermore, in renal epithelial cells urea may adversely affect the cell cycle and cell viability, much like hypertonic stress (9, 22, 25, 26).
In other respects, however, the effect of urea stress more closely resembles that of a peptide mitogen. In various renal epithelial cell models, urea activated the Ras-Raf-MAPK/extracellular signal-regulated kinase (ERK; MEK)-ERK-Elk-1 pathway, ultimately resulting in transcription and expression of immediate-early gene transcription factors (2, 7, 31). Urea also activated other receptor tyrosine kinase effectors including Shc, with recruitment of Grb2 (41); phospholipase C-γ, with production of inositol 1,4,5-trisphosphate (6); and phosphatidylinositol-3-kinase, with subsequent activation of Akt (41). Urea may also exert a proproliferative effect on renal epithelial cells (4, 5, 19) in marked contrast to the effect of hypertonic stress. Furthermore, urea pretreatment affords cytoprotection from the adverse consequences of hypertonic stress (33, 39), much like a peptide mitogen (21,39).
We used an expression array-based strategy permitting simultaneous examination of ∼12,000 murine gene products in the murine inner medullary collecting duct mIMCD3 cell line to establish whether the genetic program in response to urea treatment more closely resembled that of hypertonic stress or a peptide mitogen. Unexpectedly, yet consistent with our earlier signaling data, the effect of urea was more akin to that of epidermal growth factor (EGF). In agreement with a putative cytoprotective role for urea, pretreatment with this solute resulted in restoration toward basal levels of transcripts exhibiting the highest degree of up- or downregulation by hypertonicity.
Cell culture conditions and reagents.
mIMCD3 cells were maintained and passaged in DMEM/F-12 medium (Life Technologies) supplemented with 10% fetal bovine serum (JRH Biosciences) as described (3). Solute (200 mM urea, 200 mM mannitol, 100 mM NaCl, or 200 mM urea × 30-min pretreatment, followed by 100 mM NaCl) or EGF (100 nM) was applied to cells for 6 h before harvest. The urea pretreatment interval of 30 min was selected on the basis of prior experiments (39); cells remained in the presence of urea for the duration of the subsequent NaCl exposure.
RNA preparation, sample labeling, array hybridization, and array data processing.
Total RNA was prepared from cells using the RNeasy Midi Kit (Qiagen) in accordance with the manufacturer's direction and then labeled as described in the GeneChip Expression Analysis Technical Manual, rev. 3 (Affymetrix, Santa Clara, CA). The labeling is performed in two steps. In the first step, mRNA is converted to double-stranded cDNA using Superscript RT (GIBCO BRL) and an oligo-dT primer linked to a T7 RNA polymerase binding site sequence. In the second step, the cDNA is converted to labeled cRNA (the target) using T7 RNA polymerase in the presence of biotinylated UTP and CTP (Enzo Diagnostics). This step also provides a linear amplification of the labeled material. After removal of free nucleotides, the target is fragmented at 95°C in the presence of a high magnesium concentration to an average size of ∼50–100 nucleotides.
The fragmented material was combined with control oligomer (used for grid alignment during image processing) and control cRNAs for theEscherichia coli biotin synthesis genes BioB (1.5 pM), BioC (5 pM), and BioD (25 pM) and for the P1 bacteriophage cre gene (100 pM; all 4 were used for concentration and linearity assessment) in hybridization solution [100 mM 2-(N-morpholino)ethanesulfonic acid, 1 M Na+concentration, 20 mM EDTA, 0.01% Tween 20, 0.1 mg/ml herring sperm DNA, 0.5 mg/ml acetylated BSA]. All solutions and hybridization assay procedures were as described in the GeneChip Expression Analysis Technical Manual, rev. 3 (Affymetrix). Labeled cRNA (10 μg) was hybridized with the Murine Genome U74A GeneChip array (Affymetrix), representing 12,000 full-length genes and expressed sequence tag (EST) clusters. These chip sets were free from earlier errors reported by the manufacturer. Hybridization occurred overnight at 45°C, followed by washing, staining with streptavidin-phycoerythrin (SAPE; Molecular Probes), signal amplification with biotinylated anti-streptavidin antibody (Vector Laboratories), and a final SAPE-staining step. The distribution of fluorescent material on the array was determined using a confocal laser scanner (GeneArray Scanner, Affymetrix). We previously performed a series of related preliminary experiments using commercially available nylon membrane-based arrays spotted with a modest number of transcripts (∼1,200) preselected to encode stress-responsive gene products (Clontech); these data were scored manually from autoradiographs and PhosphorImager files (32). All mRNA samples subjected to the present (Affymetrix) analysis were obtained exclusively for the present investigations; in addition, all treatments were performed in parallel-treated cells for a particular experiment.
Cells (confluent mIMCD3 monolayers) were serum deprived for 24 h, treated as indicated, and subjected to a combination detergent-based and mechanical lysis as described by Dmitrieva et al. (11). Lysates were resolved via SDS-PAGE, transferred to polyvinylidene difluoride membranes, immunoblotted with anti-ATF3 (sc-188, Santa Cruz Biotechnology) and horseradish peroxidase-conjugated anti-rabbit secondary antibody (Pierce), and subjected to enhanced chemiluminescence (NEN).
We analyzed the gene expression data of ∼12,000 transcripts from 11 Murine Genome U74A chips: 6 chips from the initial experiment (control, urea, NaCl, EGF, mannitol, urea+NaCl) and 5 chips from the second experiment (control, urea, NaCl, EGF, urea+NaCl). Affymetrix Microarray Analysis Suite Version 4.0 (MAS 4.0) was used to determine the presence and differential expression of a transcript. MAS 4.0 measures the presence or absence of a transcript by average difference (AD) and absolute call (AC), whereas the differential expression is measured by difference call (DC) and fold-change (FC). AC has three designations: absent (A), marginally present (MP) and present (P). DC has five designations: decreased (D), marginally decreased (MD), no change (NC), marginally increased (MI), and increased (I). We performed a total of 13 comparative analyses: 5 analyses comparing the control with each of the treatment conditions in the initial experiment, 4 analyses comparing the control with each of the treatment conditions in the second experiment, and 4 analyses comparing the same conditions between the 2 experiments. For subsequent analyses, we excluded 4,057 transcripts that were consistently absent (absolute call = A) from all 11 chips. Because the number of replicate observations was necessarily small (i.e., 2 replications/gene except for the mannitol condition), we did not carry out statistical significance testing. Instead, we utilized a cluster analysis as a means of exploring data and identifying samples and/or genes with a similar expression profile. Hierarchical clustering was performed to identify samples that have a similar differential expression pattern across genes, whereas K-means clustering was performed to identify genes that have a similar differential expression profile across conditions. For the hierarchical clustering, we used the Euclidean distance and single linkage method. The following steps were taken to identify reproducibly up- or downregulated transcripts: 1) inclusion of genes exhibiting a DC of I or D for replicate comparisons; 2) assignment of I or MI and a >2-fold change for both replicate conditions → upregulated; 3) assignment of D or MD and less than a −2-fold change for both replicate conditions → downregulated; and4) assignment in all other cases → no change. In addition, we mapped FC < −30 to −30 and FC > 30 to 30 to minimize the effect of outliers in the cluster analysis. Where averaged data are presented (e.g., see Figs. 2 and 4), actual FC was used; in this case, SE averaged 16% for a representative sampling (the 30 most urea-responsive transcripts) with a range of 0–56% of the mean. SE was essentially independent of FC in this sampling. All analyses were performed using the Statistical Analysis System (version 6.0) and Statistica software.
Hypertonic stress, whether from mannitol or NaCl, affected expression of the greatest percentage of genes (Fig.1). In the depicted experiment, ∼12% of the ∼12,000 transcripts studied were downregulated by mannitol and NaCl, whereas only ∼6% were downregulated by urea stress; very few were downregulated in response to EGF treatment. The highest percentage of genes was upregulated by hypertonic stress as well: ∼4% of transcripts were upregulated by mannitol and by NaCl, whereas only 0.8 and 0.5% were upregulated by urea and EGF, respectively. Interestingly, the effect of urea+NaCl, despite reflecting a doubling of the increment in osmolarity, was approximately midway between the effects of the individual solutes with respect to the number of regulated genes. The net effect was that urea pretreatment increased the percentage of transcripts unaffected by hypertonic stress from 83 to 90%. Urea pretreatment decreased tonicity-inducible transcripts by 40% in one experiment and by 59% in the second; urea decreased tonicity-repressible transcripts by 36% in one experiment and by 12% in the second. This “normalizing” effect of urea is explored in greater detail below. Of note, urea pretreatment was selected for study, rather than simultaneous application of the two solutes, because our previous studies addressing the putative cytoprotective effects of urea were performed with this model (33, 39). While potentially less reflective of renal medullary physiology in vivo, this approach is widely applied in cytoprotection studies and studies with pharmacological inhibitors.
At 6 h of treatment, a number of genes were reproducibly upregulated by urea stress in mIMCD3 cells. The most highly regulated named genes are depicted in Fig. 2; a larger number of genes were more modestly regulated. Although most regulated genes were named and not anonymous ESTs, the most strongly regulated gene product was an anonymous EST (AA690218; average FC = 34). The next most highly regulated gene product was LRG-21, also known as ATF3. This transcription factor was rapidly and markedly induced at the protein level in response to urea stress in mIMCD3 cells (Fig. 3). The transcription factor chop-10 (also known as Gadd153) was the third most urea-responsive gene product; we previously identified this gene as being urea responsive through a candidate gene approach (40). Other transcription factors known to be urea responsive, including the zinc-finger protein Egr-1 (3) and the transcription factor c-fos, were also detected in our analysis, further validating these results.
A larger number of genes were regulated by hypertonic stress than by urea stress in the mIMCD3 model. Named genes upregulated more than sevenfold are depicted in Fig. 4; far more were regulated to a lesser extent. Of the listed genes, at least five are known to be osmotically responsive in renal cells, including HSP70 (8), osmotic stress protein (OSP)94 (18), aquaporin-1 (16, 17), HSP105 (18), and cyclooxygenase-2 (36, 38), further validating the experimental approach and results.
Concordance among experiments varied with the experimental condition applied. The response to hypertonicity was the most reproducible. Of genes upregulated by hypertonic stress in the first experiment, 84% were upregulated in the second experiment (data not shown). It must be recalled that the threshold for upregulation was established as a twofold change, and much of the discrepancy included genes exhibiting more modest levels of induction. Of genes downregulated by hypertonic stress in the first experiment, 55% were downregulated in the follow-up experiment. Again, this apparent discordance was absent among the most highly regulated genes. The concordance for urea-responsive genes was slightly less: of genes upregulated by urea in the first experiment, 69% were upregulated in the subsequent experiment; and of the genes downregulated by urea in the first experiment, 52% were downregulated in the subsequent experiment. Concordance, however, was much higher among the more dramatically up- or downregulated genes.
We next sought to establish the relationship among the different treatments in mIMCD3 cells to determine whether the effect of urea more closely resembled that of a peptide mitogen or a hypertonic stressor. Hierarchical cluster analysis was performed for each of the two experiments and for the average FC from two comparative analyses; results were similar except that mannitol treatment was excluded from the second set of experiments. The hierarchical cluster analysis permits identification of treatments exhibiting similar effects on each of the individual gene products. The dendrogram in Fig.5 indicates that, in terms of fold-induction, urea treatment more closely resembled that of EGF treatment than it did either hypertonic stressor. Similar data were obtained when the absolute difference for each gene product was used rather than fold-induction. The effect of urea+NaCl was equidistant from that of urea or NaCl alone.
The influence of urea pretreatment on tonicity-dependent gene regulation was assessed on a global scale. We had earlier shown that urea pretreatment blunted tonicity-dependent transcription in a primarily transient transfection-based model system (33). K-means cluster analysis was applied to the entire data set for each experiment and to averaged data from both experiments to identify statistical clusters of genes exhibiting coregulation across all treatment groups, with the goal of identifying functional relationships. For the averaged data, analysis was based on FC relative to control under each of 4 conditions (+EGF, +urea, +NaCl, urea+NaCl); 10-cluster analysis was selected (Fig.6 A), and the composition of resultant clusters is depicted in Fig. 6 B. The large majority of genes segregated into clusters exhibiting little regulation under any experimental condition (e.g., clusters 1,2, 5, and 8). The most conspicuous findings concerned hypertonic stress and the modulating effect of urea pretreatment. Genes comprising one cluster, cluster 10 (109 genes), were markedly upregulated in response to hypertonic stress (either NaCl, as shown, or mannitol, not shown; mean FC ∼14) but not to urea; importantly, expression of these transcripts was restored toward basal levels by urea pretreatment (mean FC ∼8). In similar fashion, genes comprising another cluster, cluster 3 (128 genes), were profoundly downregulated by hypertonicity (mean FC about −14) and restored toward basal levels of expression by urea pretreatment (mean FC about −4). Urea alone exerted little effect on expression of these transcripts. Expression of genes comprisingcluster 9 (180 genes) was suppressed by urea treatment, relative to control (mean FC about −9) but not by NaCl treatment. Expression of these genes was restored toward normal levels by concurrent urea and NaCl treatment (mean FC about −1). EGF treatment, however, modestly suppressed expression of this cluster as well. There was also a small cluster of genes (cluster 4; 225 genes) exhibiting downregulation only in the presence of both urea+NaCl. When K-means clustering was performed, not with averaged data from each of the two experiments but with expression data from each treatment group from each of the two individual experiments (for a total of 8 groups segregated into 10 clusters), similar findings were observed (not shown). Restated, among the smaller, more highly regulated gene clusters, there was a high degree of concordance between these two analyses. For example, for cluster 3 (the NaCl-sensitive genes restored by urea), fully 112 of 128 genes mapped to this cluster by the former analysis (using averaged data) were similarly mapped to this cluster in the latter analysis (using all data from both experiments).
Two-way joining analysis was used to graphically represent the effect of each stimulus on the expression level of each gene in the predominantly NaCl-influenced clusters. Figure7, A and B, graphically depicts the effect of urea pretreatment on expression of constituents of clusters 3 and 10, respectively. The overall pattern of expression in the setting of urea pretreatment more closely resembles basal expression (shown as red in Fig.7 A and yellow in Fig. 7 B; see legends) than does NaCl treatment in the absence of urea pretreatment.
The identities of named genes in clusters 3 and10 are shown in Fig. 8. A large proportion of genes in cluster 3 were anonymous ESTs, hence the limited list; the tabulation is nearly identical to that of the most highly tonicity-responsive genes depicted in Fig. 4. Tonicity-dependent downregulation of nearly all of these genes was blunted by urea pretreatment. Cluster 10 (Fig.9; truncated at 50 genes) includes several genes known to be osmotically responsive, such as cyclooxygenase-2, HSP70, OSP94, and aquaporin; all are returned toward basal levels by urea pretreatment. Tonicity-dependent expression of a small number of genes (Fig. 9, bottom) is not abrogated by urea pretreatment (e.g., Fos-like antigen-1 and Delta transcription factor); these genes segregated into this cluster on the basis of similarities under other treatment conditions (e.g., EGF).
Earlier data suggested that urea stress activated a receptor tyrosine kinase-like signaling pathway. In the mIMCD3 and related renal epithelial cell models, urea activated the Ras-Raf-MEK-ERK-Elk-1 pathway, culminating in immediate-early gene transcription and expression (2, 7, 31). Urea also activated other effector arms of a receptor tyrosine kinase including activation of Shc and recruitment of Grb2 (41); activation of phospholipase C-γ, the relatively receptor tyrosine kinase-specific phospholipase, and production of inositol trisphosphate (6); and activation of phosphatidylinositol-3-kinase and Akt (41). In addition, data from several groups support a role for urea in either complete or incomplete mitogenesis of renal epithelial cells (4,5, 19). Nonetheless, the identity of the urea-sensing molecule has remained elusive, and the specificity of these urea-dependent signaling events to the activation of a putative upstream receptor tyrosine kinase remains to be established.
In addition to its role as an activator of receptor tyrosine kinase effectors, other data have suggested that, at the concentrations present in the renal medulla, urea functions as a cell stressor. Urea is a potent denaturant of proteins (35) and nucleic acids (10). Urea increases expression of the stress-responsive transcription factor Gadd153 (40) and the cytoprotective enzyme heme oxygenase-1 (30) in an oxidative stress-dependent fashion. Urea also modestly activates the stress-responsive JNK and p38 mitogen-activated protein kinases (1, 37).
Because of this dual nature, we sought to establish whether urea signaling genuinely resembled activation of a receptor tyrosine kinase in a global fashion. On the basis of K-means analysis incorporating expression information from >12,000 gene products, it was clear that urea signaling more closely resembled that of EGF than that of hypertonicity, whether achieved with NaCl or mannitol. This unbiased assessment corroborated our biochemical findings in renal epithelial cell models of signaling by urea stress (33, 39).
We recently hypothesized that urea pretreatment affords cytoprotection from the adverse effects of hypertonicity in renal epithelial cells (33, 39). This view arose from observations that urea pretreatment partially blocked hypertonicity-inducible apoptosis as measured via activation of caspase-3 and appearance of annexin-V on the outer leaflet of cell membranes (39). In addition to these biochemical indices, urea also improved cell viability in response to hypertonic stress, as determined via exclusion of the fluorescent DNA-intercalating dye ethidium D-1 (data not shown). These properties were reminiscent of the mitogen-dependent cytoprotection observed in a nonrenal model of osmotic shock (21). Urea also blocked tonicity-inducible (TonEBP-dependent) expression of the osmotic effector gene aldose reductase and blocked the increase in TonEBP expression that accompanies osmotic shock (33). The present model system permitted an examination of the effect of urea on tonicity-dependent gene expression in an unbiased fashion; consistent with our previous hypothesis, a substantial subset of tonicity-regulated genes (including nearly all of the most profoundly regulated genes) were less dramatically regulated if urea treatment preceded the hypertonic stress (Figs. 8 and 9). Rather than representing a maladaptive phenomenon, in light of the enhanced cell survival seen with urea pretreatment (39), we previously reasoned that the cytoprotective effect of urea obviated the need for maximal up- (or down-) regulation of these putative adaptive or protective gene products (33).
The role that EGF signaling in response to urea exposure may play in the renal medulla in vivo is unclear. Although this tissue is not mitotically active, not all growth factor effector pathways exclusively subserve this end. It is possible that activation of a subset of receptor tyrosine kinase effectors affords protection from the adverse effects of hypertonic stress in vivo, as has been observed in vitro (Figs. 8 and 9) (21, 39).
Not all have agreed that urea affords protection from NaCl stress, although a consensus is emerging that simultaneous application of the solutes is superior. Santos et al. (26) showed that the combination of NaCl and urea was better tolerated than either solute applied in isolation in the mIMCD3 model. Neuhofer et al. (25) observed that exposure to hypertonic stress was protective of subsequent urea stress in the MDCK cell model. Michea et al. (22) concluded that the effects of NaCl and urea, when each was applied at 200 mosmol/kgH2O, were not additive in terms of growth inhibition in an indirect assay of cell viability [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay]. Nonetheless, the 400 mosmol/kgH2O aggregate osmolarity (urea+NaCl) was enormously better tolerated than even 350 mosmol/kgH2O of either solute applied in isolation by this assay (22). Most recently, Colmont et al. (9) noted that urea sensitized mIMCD3 cells to heat shock-induced apoptosis, whereas this phenomenon was blunted by hypertonic NaCl.
The number of genes upregulated by urea stress in the present analysis, in contrast to that upregulated by hypertonic stress, is quite small; a substantial percentage of these genes (∼9/21 in Fig. 2), however, encode transcription factors. It is likely that examination of time points beyond 6 h will be required to fully elucidate the genetic program engendered by urea stress in cells of the renal medulla.
Although the expression array approach is a relatively new one, we believe that the present data are valid for the following reasons. First, a subset of the prospectively identified regulated genes [e.g., HSP70 (8), OSP94 (18), Egr-1 (3), and Gadd153 (40)] were previously shown to be inducible by the corresponding stress. Second, there was a high degree of concordance between the two models of hypertonic stress (mannitol inducible and NaCl inducible) with respect to gene expression. Third, there was a high degree of concordance between individual experiments with respect to the most highly regulated gene products for a given treatment group. Fourth, with respect to urea-dependent cytoprotection from hypertonic stress, there was considerable agreement between the K-means cluster analysis data (with respect to cluster 10 in Figs. 6 and 9) and data derived manually from the rank-order list of tonicity-inducible genes (Fig. 4). Fifth, a previously undescribed urea-responsive gene, ATF-3, was prospectively identified and subsequently confirmed at the level of immunoblot analysis.
In contrast to this global perspective, little but speculation can be derived from the individual identities of genes responsive to any conditions, as this was not the primary objective of these investigations. Nonetheless, as discussed above, several urea-responsive and tonicity-responsive gene products identified through the present analysis were previously identified in this context. A subset of the tonicity-responsive transcripts encode proteins with important immune response functions; whether this relates to the nature of the tonicity-dependent transcription factor TonEBP as a member of the immunomodulatory NFAT family (20, 24), to the structural and functional similarity of TonEBP to members of the nuclear factor-κB family (15, 28), or to a role for TonEBP in lymphocyte activation (34) remains to be established. With respect to genes upregulated in response to hypertonic stress, however, caution is required when data at the mRNA level alone are interpreted. Hypertonic stress inhibits protein translation (e.g., Ref. 8), and pharmacological inhibitors of protein translation nonspecifically “superinduce” mRNAs encoding various immediate-early gene products such as transcription factors (13, 23). Therefore, it is likely that hypertonic stress results in some indiscriminate upregulation of immediate-early gene transcripts that fail to undergo subsequent translation and hence are of questionable physiological significance. Importantly, the same caveat does not hold true for urea stress, which does not globally inhibit translation (8). For example, the immediate-early gene transcription factor ATF-3, which was dramatically upregulated in response to urea stress in these studies (Fig. 2), was also markedly upregulated at the protein level (Fig. 3). Nonetheless, at least a subset of the tonicity-responsive genes identified herein were previously shown to be upregulated at the protein level by others; expression of these genes, like many others, was suppressed by urea pretreatment.
In the aggregate, these data indicate that the genetic program in response to urea stress in renal medullary cells more closely resembles that of a peptide mitogen than a hypertonic stressor. In addition, the putative cytoprotective role of urea, another hallmark of a peptide mitogen, is supported by its ability to restore toward basal levels of expression genes most profoundly up- or downregulated in response to hypertonicity.
The authors thank Chris Harrington of the Oregon Health Sciences University (OHSU) Gene Microarray Shared Resource Affymetrix Microarray Core and Edwin Quick, Yi-Ching Hsieh, and Motomi Mori of the OHSU Cancer Institute's Biostatistics and Bioinformatics Shared Resource (NIH 5P30 CA69533–04) for data management and analysis.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-52494 and by the Department of Veterans Affairs.
Address for reprint requests and other correspondence: D. M. Cohen, Mailcode PP262, Oregon Health Sciences Univ., 3314 S.W. US Veterans Hospital Rd., Portland, OR (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.
March 12, 2002;10.1152/ajprenal.00031.2002
- Copyright © 2002 the American Physiological Society