Oxidative signaling in renal epithelium: Critical role of cytosolic phospholipase A₂ and p38^SAPK

Abstract

Previous studies from this laboratory have demonstrated a critical role of cytosolic phospholipase A₂ (cPLA₂) and arachidonic acid in angiotensin II (Ang II) AT2 receptor-mediated signal transduction in renal epithelium. In primary proximal tubular epithelial cells exposed to hydrogen peroxide (H₂O₂), both the selective cPLA₂ inhibitors and the cPLA₂ antisense oligonucleotides significantly attenuated H₂O₂-induced arachidonic acid liberation and activation of p38^SAPK, ERK1/2, and Akt1. This H₂O₂-induced kinase activation was significantly attenuated by a Src kinase inhibitor PP2, or by transient transfection of carboxyl-terminal Src kinase (CSK) that maintained Src in the dormant form. Under basal conditions, Src coimmunoprecipitated with epidermal growth factor receptor (EGFR), while H₂O₂ increased EGFR phosphorylation in the complex. We observed that inhibition of EGFR kinase activity with AG1478 significantly attenuated H₂O₂-induced p38^SAPK and ERK1/2 activation, but did not inhibit Akt1 activation. Furthermore, it seems that p38^SAPK is upstream of ERK1/2 and Akt1, since a p38^SAPK inhibitor SB203580 significantly blocked H₂O₂-induced activation of ERK1/2 and Akt1. Interestingly, overexpression of the dominant-negative p38^SAPK isoform α inhibited ERK1/2 but not Akt1 activation. Our observations demonstrate that in these nontransformed cells, activation of cPLA₂ is a converging point for oxidative stress and Ang II, which share common downstream signaling mechanisms including Src and EGFR. In addition, p38^SAPK provides a positive input to both growth and antiapoptotic signaling pathways induced by acute oxidative stress.

Introduction

Oxidative stress has a major contributory role in the pathophysiology of various forms of renal disease. As a stable reactive oxygen species (ROS), H₂O₂ has been shown to be the main initiator of proteinuria associated with acute glomerulonephritis caused by anti-glomerular basement membrane antibody, complements, and phorbol ester acetate [1-3]. Exposure of freshly isolated rat renal cortical cells to H₂O₂ produced significant cell death [4]. H₂O₂ also increased the transepithelial electrical conductance of cultured MDCK cells, as the result of an increase of the paracellular pathway permeability associated with cytoskeleton reorganization, particularly in the area of cell-to-cell junctions [5]. It is clear that at least a portion of ROS is generated by infiltrated phagocytes in disease states [6,7]; however, the mechanisms of de novo ROS production and oxidative signaling pathways within the renal epithelium have not been clarified.

Oxidative stress has been linked to activation of members of the MAP kinase superfamily in various cell types. ERK1/2 are ubiquitously expressed and generally considered to be responsible for mediating the mitogenic and differential effects of growth factors, while p46/p54^SAPK and p38^SAPK respond to environmental stress and play an important role in apoptosis [8]. Particularly, p38^SAPK mediates Ang II- and ROS-induced hypotrophy which was only partially dependent on ERK activation in vascular smooth muscle cells [9]. Moreover, p38^SAPK has been found to inhibit apoptosis, presumably via Akt1 activation by forming a PDK2-like complex upon stimulation [10]. Therefore, we hypothesize that p38^SAPK plays an essential role in facilitating survival in renal epithelium in the presence of oxidative stress.

Previous studies from this laboratory and others have shown that Ang II activates ERK1/2 in primary proximal tubular cells. Transactivation of the EGFR tyrosine kinase is an upstream element of shc/grb2/sos-mediated ras activation in our system [11,12]. Recently, we have demonstrated that arachidonic acid and/or its metabolites mediate all downstream effects of Ang II in activation of members of the MAPK, i.e., the ERK1/2, p46/p54^SAPK, and p38^SAPK [13,14]. Arachidonic acid also induces ROS production in these proximal tubular epithelial cells [15], which provides mechanism for endogenous ROS generation [16]. However, ERK, but not p38^SAPK, was suggested to be responsible for Ang II-induced and ROS-mediated hypertrophy in MCT and LLCPK1 cells, cell lines with proximal tubular phenotype [17]. Therefore, the current studies were designed to determine the regulation of oxidative signaling in primary proximal tubular cells based on the hypothesis that arachidonic acid is a common second messenger for oxidative stress and Ang II receptor activation.

Materials and methods

Materials

Cell culture medium, additives, and LipofectAmine for transfection were from Gibco BRL (Gaithersburg, MD). [³H] Arachidonic acid was obtained from Dupont/NEN (Boston, MA). Rabbit polyclonal anti-phospho antibodies recognizing activated p38^SAPK, ERK1/2, Akt1, and cPLA₂ were from Cell Signaling (Beverly, MA). Polyclonal anti-p38^SAPK, anti-ERK1/2, anti-Akt1, anti-p110α, and anti-cPLA₂ antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-c-Src and anti-EGFR goat polyclonal, anti-ras, anti-vSrc, anti-phosphotyrosine, and anti-phospho-EGFR monoclonal antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). SB203580, PD98059, LY294002, arachidonyltrifluoromethyl ketone (AACOCF₃), methyl arachidonyl fluorophosphate (MAFP), oleyloxyethyl phosphorylcholine (OPC), and PP2 were from CalBiochem (San Diego, CA).

Cell culture

Proximal tubule epithelial cells were isolated from New Zealand White rabbits as previous described [15]. They were maintained for 9–14 days in modified DMEM:F12 (1:1) media supplemented with 5% fetal calf serum, 5 μg/ml insulin, 5 μg/ml transferrin, 0.5 μM hydrocortisone, 350 μg/ml l-glutamine, 100 unit/ml penicillin, and 100 μg/ml streptomycin. Subcon-fluent monolayers of first-passage cells were employed for experiments.

Arachidonic acid liberation

Cells grown on 12-well plates to subconfluence were labeled with 0.5 μCi/ml/well [³H]arachidonic acid for 3 h prior to treatment. They were washed 3 times with DMEM medium containing 1 mg/ml fatty acid-free bovine serum albumin to remove free labels. After treatment of cells, the medium was removed and the released arachidonic acid was determined by scintillation counting. Each data point is the average from at least three wells.

Rabbit cPLA₂ antisense

The amount of 5 μM of each of the two phosphorothioate modified antisense oligonucleotides was introduced to cells in serum-free medium 24 h prior to experiment, with control cells treated with corresponding sense oligonucleotides. The antisense oligonucleotides were: 5′-GATCTATAAATGACATTTTGGT-3′ and 5′-ATCCGGGGTGTCAAGCATGTCA-3′. They were designed to target the sequences 1–22 and 114–135 that contain the two possible translation starting codons.

Immunoblot and immunoprecipitation

Cells were serum-deprived for 20-24 h prior to experiments. Following experimental treatments, the cells were washed twice with ice-cold Dubecco's phosphate-buffered saline. They were then lysed on ice with lysis buffer (50 mM Tris, pH 7.2, 1% (vol/vol) Triton X-100, 1 mM Na₃VO₄, 1 mM EGTA, 0.2 mM phenylmethanesulfonyl fluoride, 25 μg/ml leupeptin, and 10 μg/ml aprotinin). The samples were centrifuged at 14,000g for 10 min. Protein content in the supernatants was determined by the BCA assay according to the manufacturer's instructions (Pierce, Rockford, IL). Fifteen micrograms of total cell lysate protein was subjected to SDS-PAGE and then transferred to a polyvinylidene difluoride membrane by electroblotting at 200 mA for 1.5 h. The membrane was immunoblotted according to Western blot protocol furnished by the manufacturer, and the immunoreactive proteins were detected by enhanced chemiluminescence (ECL) (Amersham Pharmacia Biotech Inc., Piscataway, NJ). All experiments had at least one membrane reprobed with antibodies recognizing nonphosphorylated kinases to confirm equal protein loading. The exposure autoradiograph was analyzed by the OS-Scan image analysis system to obtain the densitometry data. For immuno-precipitation, each cell lysate containing 1 mg protein was mixed with 10 μg antibody and 40 μl protein A + G agarose bead suspension. The bead pellet was washed 3 times and then heated in Laemmli's sample buffer. The supernatant was resolved by 10% SDS-PAGE. The associated protein was detected by immunoblot.

Src GST-SH2 fusion protein and pull-down experiment

Src SH2 sequence (amino acids 140–250) was retrieved from chicken c-Src cDNA plasmid by PCR and cloned into pGEX-2T vector and transformed into BL21 Escherichia coli. The sequence was verified by sequencing both strands at Cleveland Genomic LT with ABI PRISM, Model 377. Standard protocol (Pharmacia) was employed to prepare the GST fusion protein conjugated with glutathione Sepharose 4B beads. The preparation was washed and resuspended to 1:1 slurry in PBS containing 0.1% azide, 0.1% bovine serum albumin. Twenty microliters GST fusion protein slurry was mixed with 1 mg cell lysate prepared in SDS/RIPA buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 0.25% sodium deoxycholate, 0.1% Nonidet P-40, 100 μM NaVO₄,1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 25 μg/ml leupeptin, 0.1% SDS) at 4°C for 2 h. The samples were centrifuged and washed for 3 times in lysis buffer to remove nonspecific bound protein. The final pellets were boiled in Laemmli's sample buffer and resolved by 8% SDS-PAGE. The associated EGF receptor was detected by anti-EGF receptor immunoblot.

cDNA transient transfection

Cells grown on 6-well plates were transfected using LipofectAmine 2000 with 2 μg/well of CSK plasmid or p38^SAPK α (TGY → AGF) plasmid or control plasmid for mock transfection. These cells were employed 48–72 h following transfection.

Data analysis

The obtained densitometry arbitrary units varied from blot to blot depending on the exposure time of ECL. Therefore, data were converted to percentage of the resting value (set as 100%). The basal effect of a putative inhibitor was first compared with resting value, and then the effect on H₂O₂-stimulated kinase activation was calculated as (change of kinase activity in the presence of inhibitor/change of kinase activity in the absence of inhibitor) × 100. The significance was determined by the paired two-tailed Student t test, with p ≤ 0.05 designated as significant. The effects of putative inhibitors or antisense on arachidonic acid release were calculated as above, except that the raw data are the mean of released radioactive counts (percentage of total) from three wells.

Results

We used H₂O₂ as the model molecule to initiate intracellular oxidative stress, as this molecule mediates renal pathologies and is readily permeable to the plasma membrane. We have previously employed immune complex kinase assay and demonstrated in primary proximal tubular epithelial cells that H₂O₂ dose dependently activates p46/p54^SAPK [15]. In the current studies, activation of p38^SAPK, ERK1/2, and Akt1 was determined by immunoblot with anti-phospho-p38^SAPK (Thr180/Tyr182), anti-phospho-ERK1/2 (Thr202/Tyr204), or anti-phospho-Akt1 (Ser473) antibodies which recognize the activated enzymes. Therefore, we refer to the phosphorylation as enzyme activation in the text. We have chosen 0.5 mM H₂O₂ to stimulate cells according to a dose-dependent activation of Akt1 in preliminary studies (not shown). As shown in Fig. 1, when cells were exposed to 0.5 mM H₂O₂, the activation of p38^SAPK peaked at 5 min (8.9 ± 5-fold) and subsided, which was followed by a second activation that peaked at 60 min. At 2 and 4 h, p38^SAPK activation was 118 ± 37 and 51 ± 16% of the 60 min value, respectively. The activation of ERK1, ERK2, and Akt1 was significant at 5 min (2.5 ± 0-fold, 2.3 ± 0-fold, and 7.8 ± 2-fold, respectively) and was sustained for 1 h. During this time frame, there was no change of protein expression level (bottom panel) for p38^SAPK.

Fig. 1.

Kinetics of kinase activation: 0.5 mM H₂O₂ was added to quiescent cells for the time period indicated. Fifteen micrograms of each cell lysate was resolved by 10% SDS-PAGE. Activation of p38^SAPK, ERK1/2, and Akt1 was detected by immunoblotting with anti-phospho antibodies that recognize the active forms of the kinases. p38^SAPK protein level and equal protein loading are shown at the bottom.

Our previous studies demonstrated that cPLA₂ activation is critical for Ang II-induced MAP kinase activation in these cells. We decided to investigate if cPLA₂ serves as a converging point of oxidative stress- and growth hormone-induced signaling. We observed that H₂O₂ time dependently induced cPLA₂ activation, by phosphorylation at Ser505[18], in these proximal tubular cells (Fig. 2A). We then measured H₂O₂–induced arachidonic acid liberation from [³H]arachidonic acid-labeled cells. A significant 135 ± 6% increase was observed within 5 min, and reached 230 ± 15% of the control level at 60 min (Fig. 2B). Our experiments have revealed that the presence of 3 mM EGTA or pretreatment of cells with a nonspecific PLA₂ inhibitor mepacrine significantly blocked above H₂O₂-induced arachidonic acid liberation (not shown). We then pretreated cells with two structurally unrelated specific cPLA₂ inhibitors arachidonyl trifluoromethyl ketone and methyl arachidonyl fluorophosphate [19-21]. Both inhibitors blocked H₂O₂-induced p38^SAPK, ERK1/2, and Akt1 activation (Fig. 2C). AACOCF₃ also blocked H₂O₂-induced arachidonic acid release by 66–80% in two independent experiments (data not shown). In contrast, H₂O₂-induced activation of kinases was not inhibited by oleyloxyethyl phosphorylcholine, an inhibitor of secretory PLA₂ (Fig. 2C).

Fig. 2.

(A) Time course of H₂O₂-induced cPLA₂ activation: 0.5 mM H₂O₂ was added to quiescent cells for the time period indicated. Fifteen micrograms of each cell lysate was resolved by 7.5% SDS-PAGE. Activation of cPLA₂ was detected by immunoblotting with anti-phospho-cPLA₂ antibody (top panel). The same blot was stripped and reprobed with control anti-cPLA₂ antibody showing equal loading (lower panel). (B) Time course of H₂O₂-induced arachidonic acid release. Cells were labeled with [³H]arachidonic acid and were treated with 0.5 mM H₂O₂ for the time period indicated. Aliquots of medium were counted in the scintillation fluid and calculated as a percentage of radioactivity released from cells exposed to medium alone for the same length of time. Arachidonic acid release was calculated from at least 3 determinations of two similar experiments (*p < 0.05, **p < 0.01, ***p < 0.001). (C) Effects of PLA₂ inhibitors on H₂O₂-induced p38^SAPK, ERK1/2, and Akt1 phosphorylation. Quiescent cells were incubated with or without 10 μM AACOCF₃,10 μM MAFP, or 10 μM OPC for 30 min followed by exposure to 0.5 mM H₂O₂ for 5 min. The cell lysates were subjected to immunoblot analysis as described in Fig. 1 legend. The same blot was stripped and reprobed with control anti-p38^SAPK, anti-ERK1/2, and anti-Akt1 antibodies showing equal loading (lower panels).

To further investigate the role of cPLA₂ in the signal transduction, we cloned the full-length rabbit cPLA₂ cDNA (GenBank Accession No. AF204923) by RT-PCR of total RNA extracted from heart tissue of New Zealand White rabbit. We verified the presence of cPLA₂ in rabbit renal epithelial cells. As shown in Fig. 3A, a 999-bp fragment was amplified from renal epithelial cell total RNA employing the primers designed with the heart tissue cPLA₂ sequence. With introduction of the cPLA₂ antisense oligonucleotides that target the translation starting sites of cPLA₂, we observed a decreased expression of cPLA₂ protein by 50% as compared to cells treated with sense oligonucleotides (Fig. 3B). As expected, expression of another protein, Akt1, which is also involved in the H₂O₂ signaling, was not affected by this antisense treatment. The antisense treatment attenuated arachidonic acid release induced by H₂O₂ and Ang II to 23 ± 7 and 31 ± 22%, respectively, as compared to those treated with sense oligonucleotides (Fig. 3C). Furthermore, introduction of the cPLA₂ antisense oligonucleotides significantly attenuated activation of p38^SAPK, ERK1/2, and Akt1 to 37 ± 27, 12 ± 8 (ERK1), 19 ± 14 (ERK2), and 66 ± 14%, respectively, as compared to levels in cells exposed to sense oligonucleotides (Fig. 3D). Therefore, we have demonstrated that H₂O₂-induced kinase activation is mediated by a cPLA₂-dependent release of arachidonic acid in these cells. It indicates that oxidative stress (represented by H₂O₂) and growth hormones (represented by Ang II) share a common signaling mechanism in these cells.

Fig. 3.

(A) Amplification of cPLA₂ by RT-PCR. Total RNA was extracted using a Qiagen kit and was reverse-transcribed (RT) with random hexamer primers. The primers used for PCR were 5′-GGAGATCACATTGATGGATGC-3′ (forward) and 5′-TCTTCCTCCATTGTGGAACC-3′ (reverse). (B) Protein expression. The cPLA₂ protein expression was detected by immunoblot of 15 μg lysate from sense or antisense oligonucleotide-treated cells employing anti-mouse cPLA₂ antisera. Mouse cPLA₂ was employed as standard. Membrane of above blot was stripped followed by immunoblotting with anti-Akt1 antibody. (C) Effects of cPLA₂ antisense on H₂O₂- and Ang II-induced arachidonic acid release. Cells treated with sense or antisense oligonucleotides were labeled with [³H]arachidonic acid and treated with 0.5 mM H₂O₂ or 1 μM Ang II for 5 min. Arachidonic acid release was determined as described in Fig. 2B legend and the results are presented as the percentage of released arachidonic acid of total label. The mean of two experiments is shown with 3 replicates in each experiment. (D) Effects of cPLA₂ antisense on H₂O₂-induced kinase activation. Cells treated with cPLA₂ sense or antisense oligonucleotides were exposed to 0.5 mM H₂O₂ for 5 min. Kinase activities were determined as described in Fig. 1 legend (*p < 0.05, **p < 0.01, ***p < 0.001).

We have previously demonstrated that arachidonic acid mimics Ang II in activation of Src tyrosine kinase, induction of Src and EGFR association via SH2 domain, transactivation of the EGFR tyrosine kinase, activation of small GTPase ras, and activation of downstream ERK1/2 [11,12,14,22]. Experiments were designed to test whether Src and EGFR tyrosine kinases are responsible for mediating H₂O₂-induced signaling in these epithelial cells. As shown in Fig. 4A, a specific Src-family kinase inhibitor PP2 significantly attenuated p38^SAPK activation to 38 ± 14%, blocked ERK1/2 activation to 2 ± 3 and 1 ± 2%, and decreased Akt1 activation to 79 ± 6%, at a submaximum dose of 5 μM (effective range 1–50 μM, not shown). Alternatively, transient transfection of CSK phosphorylates Tyr529 in Src C-terminus and keeps the Src SH2 domain intramolecularly bond [22,23]. This manipulation also significantly attenuated H₂O₂-induced ERK1/2 and Akt1 activation to 59 ± 5, 51 ± 1, and 24 ± 19%, respectively. Immunoprecipitation with anti-v-Src antibody and blotting with anti-total EGFR antibody indicated a constitutive association of Src and EGFR in these cells (Fig. 4B, middle panel). However, H₂O₂ induced tyrosine phosphorylation of the Src-associated EGFR, which peaked at 10 min (Fig. 4B, top panel). In parallel, H₂O₂ induced the association of phosphorylated EGFR with ras, suggesting induction of downstream ERK1/2 signaling as seen with Ang II stimulation. The induction of Src-phospho-EGFR association was confirmed with Western blot analysis using anti-phospho-EGFR antibody (Fig. 4B), as well as with in vitro pull-down experiments employing immobilized Src-SH2 protein (Fig. 4C). Further, an EGFR kinase inhibitor AG1478 [23] dose dependently attenuated H₂O₂–induced kinase activation (Fig. 5A). But AG1478 was obviously more effective in blocking ERK1/2 activation (5 ± 6 and 5 ± 7%, respectively) than p38^SAPK (62 ± 28%) and Akt1 (67 ± 37%) activation (Fig. 5B). Taken together, our data confirmed the existence of Src/EGFR-dependent ERK1/2 pathway induced by acute oxidative stress.

Fig. 4.

Assessment of the role of Src. (A) Cells pretreated with 5 μM PP2 for 30 min or transfected with Csk plasmid were stimulated with 0.5 mM H₂O₂ for 5 min. Kinase activities were determined as described in Fig. 1 legend (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Cells were lysed following treatment with 0.5 mM H₂O₂ for the time period indicated. One milligram total protein was immunoprecipitated with either anti-ras (left) or anti-v-Src (right) monoclonal antibody followed by blotting with (a) anti-phosphotyrosine antibody (shown are bands that resolve at 180 kDa); (b) anti-EGF receptor antibody; and (c) anti-c-Src antibody; std, molecular weight standard; Lys, lysate of EGF-treated A431 cell as positive control. (C) Cells were lysed following treatment with 0.5 mM H₂O₂ for 0-15 min. Lysates containing 1 mg total protein were immunoprecipitated with anti-c-Src goat antibody followed by blotting with anti-phospho-EGFR monoclonal antibody. (D) Association of the SH2 domain of Src with the EGF receptor. Cells were lysed following treatment with 0.5 mM H₂O₂ for the time period indicated. Cell lysates were mixed with GST-SH2 fusion protein conjugated with agarose beads. The pull-down products were resolved by SDS-PAGE and immunoblotted with anti-EGF receptor antibody.

Fig. 5.

Assessment of the role of EGFR. (A) Cells were exposed to 0–1000 nM AG1478 for 30 min prior to stimulation with 0.5 mM H₂O₂ for 5 min. (B) Cells pretreated with 100 nM AG1478 for 30 min were stimulated with 0.5 mM H₂O₂ for 5 min. Kinase activities were detected as described in Fig. 1 legend (*p < 0.05, **p < 0.01, ***p < 0.001).

The pyridinyl imidazole compound SB203580 has been reported to inhibit p38^SAPK isoforms [24]. Pretreatment of renal proximal tubular cells with SB203580 dose dependently inhibited H₂O₂-induced early and sustained p38^SAPK activation at 5 and 30 min, respectively, while ERK1/2 and Akt1 activation was attenuated as well (Fig. 6A). Fig. 6B shows that 10 μM SB203580 pretreatment decreased 5 min H₂O₂ stimulation of p38^SAPK to 14 ± 7% as compared to values in the absence of the inhibitor, while activation of ERK1/2 and Akt1 was decreased to 3 ± 3, 0 ± 0, and 25 ± 16%, respectively. In addition, we employed the dominant-negative p38^SAPK α (TGY → AGF) plasmid to transiently transfect the cells. Interestingly, expression of the dominant-negative p38^SAPKαAGF only significantly attenuated p38^SAPK and ERK1/2 activation following a 30 min H₂O₂ exposure (51 ± 16, 64 ± 6, and 69 ± 7%, respectively). Akt1 activation was not affected. With these data, we conclude that p38^SAPK differentially regulates the ERK1/2 and Akt1 pathways in renal epithelial cells.

Fig. 6.

Assessment of the role of p38^SAPK. (A) Cells were incubated with 0–10 μM SB203580 for 20 min prior to stimulation with 0.5 mM H₂O₂ for 5 and 30 min, respectively. (B) Cells were pretreated with 10 μM SB203580 followed by stimulation with H₂O₂ for 5 min, or transfected with the dominant-negative p38αAGF plasmid followed by stimulation with H₂O₂ for 30 min. Kinase activities were determined as described in Fig. 1 legend (*p < 0.05, **p < 0.01, ***p < 0.001).

The relationship between the ERK1/2 and Akt1 pathways was next examined. As shown in Fig. 7A, the specific ERK1/2 pathway inhibitor PD98059 did not affect p38^SAPK activation (72 ± 27%), while it completely blocked ERK1/2 activation (3 ± 2 and 2 ± 3%, respectively). PD98059 also attenuated Akt1 activity to 58 ± 28%. On the other hand, the specific inhibitor of the PI3-kinase/Akt1 pathway LY20094 completely blocked Akt1 activation (1 ± 1%) and decreased ERK1/2 activation to 24 ± 41 and 25 ± 22%, respectively. By contrast, LY20094 significantly potentiated p38^SAPK activation to 199 ± 50%. With immunoprecipitation employing anti-ras antibody, we observed that H₂O₂ stimulated association of the PI3-kinase subunit p110α with ras, which peaked at 10 min. The association of p110α with Src was also found increased in a similar pattern (Fig. 7B).

Fig. 7.

Assessment of the crosstalk between the pathways. (A) Cells were incubated with 30 μM PD98059 or 50 μM LY294002 for 20 min prior to stimulation with 0.5 mM H₂O₂ for 5 min. Kinase activities were detected as described in Fig. 1 legend (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Cells were lysed following treatment with 0.5 mM H₂O₂ for the time period indicated. One milligram total protein was immunoprecipitated with either anti-ras or anti-v-Src monoclonal antibody followed by immunoblotting with anti-p110α antibody. Two protein bands were seen in the blots and in positive control lysate (not shown).

Discussion

The current work establishes a signal transduction pathway whereby H₂O₂ induced activation of cPLA₂, Src, EGFR, and p38^SAPK that link two downstream kinases, i.e., ERK1/2 and Akt1 in kidney epithelial cells. The conclusion that cPLA₂ is a key upstream enzyme which mediates the oxidative signaling is based on observations that (1) cPLA₂ is activated by H₂O₂; (2) specific cPLA₂ inhibitors (AACOCF₃ and MAFP) and the cPLA₂ antisense oligonucleotides blocked H₂O₂-induced arachidonic acid release and activation of all kinases in the signaling pathway; and (3) other PLA₂ family members such as the secretory PLA₂ play a negligible role as we could not detect their existence in these cells by immunoblot, and inhibitors such as OPC [25] and sPLA2-IIA Inhibitor I (CalBiochem) [26] were not effective (L.D. Alexander, unpublished observations). To our knowledge this is the first study in nontransformed cells to show that oxidative stress induces cPLA₂ activation and shares a common signaling mechanism of Ang II.

Arachidonic acid is a polyunsaturated fatty acid with significant importance in membrane structural integrity and function [27]. It is also a regulator of oxidative status documented by its role in activation of NADPH oxidase to generate ROS in the epithelial cells used herein and other cell types [15,28,29]. Recent studies demonstrate that exposure of these renal epithelial cells to exogenous arachidonic acid and/or its metabolites differentially activate ERK1/2, p46/p54^SAPK, and p38^SAPK [13]. On the other hand, both ERK2 and p38^SAPK have been shown to activate cPLA₂ in various cell systems [18,30,31]. Therefore, our work provides a potential mechanism for feed forward of the oxidative signaling at multiple levels. That is, a combined effect of arachidonic acid and H₂O₂ could yield significant damage to the kidney.

The Src family of nonreceptor tyrosine kinases have been reported to be activated by ROS in various cell types [32,33]. Environmental stress, including ROS, has also been shown to induce MAP kinase pathway activation via activation of receptor tyrosine kinases [34,35]. A number of mechanisms have been proposed for ROS-induced EGFR activation [36]. In some cell lines, Src is responsible for direct activation of the EGFR through phosphorylating essential tyrosine residues on EGFR [37,38]. In renal epithelium, Src tyrosine kinase activation and transactivation of EGFR were found to activate the ERK1/2 pathway in responding to Ang II or arachidonic acid stimulation [11,14]. A physical association of Src and EGFR was observed in these cells under basal conditions. However, H₂O₂ induces Src-bond EGFR phosphorylation, that may be via new formation through the Src SH2 domain. It seems that Src-dependent oxidative signaling in these cells is complex. Inhibition of Src activation with PP2 or CSK transient transfection significantly decreased kinase activation, however, with different profiles. PP2 was more efficient in ERK1/2 inhibition and CSK was more efficient in Akt1 inhibition (Fig. 4A). It suggests that the Src and EGFR kinase activities are essential for ERK1/2 activation, while Akt1 might be regulated by conformational change of Src or perhaps Src association with proximal Akt1 effectors, rather than being an EGFR kinase target.

It is interesting that H₂O₂ induces p38^SAPK activation with different kinetics (Fig. 1). At present we are unsure whether the biphasic activation of p38^SAPK is a result of activation of two different isoforms of p38^SAPK or a consequence of simultaneous activation of protein phosphatases that dephosphorylates and inactivates p38^SAPK after the immediate activation phase. Protein tyrosine phosphatase 2C has been reported to directly interact with p38^SAPK [39], and MKP-1 is another potential candidate [40]. It is worth noting that the p38^SAPK isoforms have been reported to differentially regulate cell functions such as cardiac myocyte hypertrophy and apoptosis [41]. In addition, a p38^SAPKα activation mechanism that depends on interaction with transforming growth factor-β-activated protein kinase 1, rather than MEK, has been elucidated [42]. In human kidney, β and α isoforms of the p38^SAPK were identified by Northern blot [43,44]. Indeed, we have amplified two p38^SAPK -related fragments from the rabbit epithelial cells employing RT-PCR (data not shown). The current data favor activation of distinct p38^SAPK isoforms for the following reasons: (1) although SB203580 blocked both ERK1/2 and Akt1 activation which suggests that both kinases are regulated by p38^SAPK, the dominant-negative p38α construct blocked the delayed phase of p38^SAPK and the ERK1/2 pathway rather than the Akt1 pathway; (2) inhibition of the EGF receptor kinase with AG1478 or the Src kinase with PP2 blocked ERK1/2 activation profoundly, but had moderate effect on Akt1 activation. As a matter of fact, in all our experiments, Akt1 inhibition would not be obvious if p38^SAPK inhibition were not deep or complete. Therefore, we propose that the two activation phases of p38^SAPK may have distinct physiological roles. Although the current study focused more on the acute phase of oxidative stress, it is important to further study the mechanism of p38^SAPK activation starting with cloning the p38^SAPK isoforms.

There are apparent interactions of the ERK1/2 and Akt1 pathways because the inhibitors of either pathway were partially effective on each other. We realize that what we observed here could be cell type specific. For example, in opossum kidney proximal tubular cells, both kinases mediate the proliferating effect of lysophosphatidic acid without converging to each other [45]. However, PI3-kinase-dependent ERK activation has been shown to mediate insulin-induced protein synthesis in renal epithelial cells[46]. At the upstream level, ras-dependent activation of PI3-kinase has been elucidated both in vivo and in vitro [47,48]. H₂O₂ induces association of the p110α subunit of PI3-kinase with ras and Src in the renal epithelial cells. It is then possible that Src, EGFR, ras, and PI3-kinase form a dynamic protein complex that is regulated by acute oxidative input. These data also reveal the complexity of the signaling pathways and suggest, again, regulation at multiple levels.

Thus in summary, our work provides the novel observation that in renal epithelial cells, cPLA₂ activation and arachidonic acid release are critical mediators of oxidative stress signaling in a manner similar to Ang II AT2 receptor signaling. The current studies also demonstrated that p38^SAPK is an essential oxidative mediator, as it transduces the signal to at least two downstream pathways, i.e., ERK1/2 and Akt1. It extends our previous observations wherein arachidonic acid-dependent ROS generation via NADPH oxidase was responsible for activation of p46/p54^SAPK in these same kidney epithelial cells [15]. Together, the data provide a signaling paradigm for ROS-induced renal pathology.

Abbreviations

H₂O₂

hydrogen peroxide

ROS

reactive oxygen species

Ang II

angiotensin II

cPLA₂

cytosolic phospholipase A₂

EGFR

epidermal growth factor receptor

SAPK

stress-activated protein kinase

CSK

Src carboxyl-terminal kinase

SDS-PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

AACOCF₃

arachidonyltrifluoromethyl ketone

MAFP

methyl arachidonyl fluorophosphates

OPC

oleyloxyethyl phosphorylcholine