Regulation of Mesangial Cell Hexokinase Activity and Expression by Heparin-binding Epidermal Growth Factor-like Growth Factor

Heparin-binding epidermal growth factor -like growth factor (HB-EGF) expression and hexokinase (HK) activity are increased in various pathologic renal conditions. Although the mitogenic properties of HB-EGF have been well characterized, its effects on glucose (Glc) metabolism have not. We therefore examined the possibility that HB-EGF might regulate HK activity and expression in glomerular mesangial cells, which constitute the principal renal cell type affected by a variety of pathologic conditions. Protein kinase C (PKC)-dependent classic mitogen-activated protein kinase (MAPK) pathway activation has been associated with increased HK activity in this cell type, so we also examined dependence upon these signaling intermediates. HB-EGF (≥10 nm) increased total HK activity over 50% within 12–24 h, an effect mimicked by other EGF receptor agonists, but not by IGF-1 or elevated Glc. EGF receptor and classic MAPK pathway antagonists prevented this increase, as did general inhibitors of gene transcription and protein synthesis. Both HB-EGF and phorbol esters activated the classic MAPK pathway, albeit via PKC-independent and PKC-dependent mechanisms, respectively. Both stimuli were associated with increased HK activity, selectively increased HKII isoform expression, and increased Glc metabolism via both the glycolytic-tricarboxylic acid cycle route and the pentose phosphate pathway. HB-EGF thus constitutes a novel regulator of mesangial cell HK activity and Glc metabolism. HKII is the principal regulated isoform in these cells, as it is in insulin-sensitive peripheral tissues, such as muscle. However, the uniform requirement for classic MAPK pathway activation distinguishes HKII regulation in mesangial cells from that observed in muscle. These findings suggest a novel mechanism whereby growth factors may couple metabolism to glomerular injury.

Hexokinases (HKs) 1 play a central role in cellular glucose (Glc) uptake and utilization and are of fundamental importance to all cells. By catalyzing the first committed step of Glc metabolism, the phosphorylation of Glc to yield Glc-6-phosphate (Glc-6-P), HKs both maintain the concentration gradient permitting facilitated Glc uptake and initiate all major pathways of Glc metabolism. Three high affinity HK isoforms (HKI-III) are expressed in mammalian kidneys, but the specific intrarenal expression patterns, regulation, and relative importance of individual isoforms have not been fully addressed. Interestingly, the extant descriptions of altered HK activity in the adult kidney have been largely restricted to pathological conditions associated with renal functional and/or structural abnormalities (1)(2)(3)(4). Thus, as the principal renal cell type affected by a variety of pathologic conditions, including diabetic glomerulopathy and many forms of glomerulonephritis, the regulation of HK expression and activity in glomerular mesangial cells may have significant pathophysiologic, as well as physiologic, relevance. We have previously demonstrated novel regulation of HK activity, vis-à -vis total cellular Glc phosphorylating capacity, in cultured mesangial cells exposed to phorbol esters or thrombin (5,6), but the relative contributions of individual HK isoforms have not yet been defined. Induction of mesangial cell HK activity by these agents requires signal transduction via the classic mitogen-activated protein kinase (MAPK) pathway, which is activated by a wide variety of stimuli in this cell type, including growth factors and cytokines (7). We have therefore sought to examine other known activators of the classic MAPK pathway for the ability to increase HK activity, as well as the expression of individual HK isoforms, in mesangial cells.
Heparin-binding epidermal growth factor-like growth factor (HB-EGF) is a recently described EGF family member that is structurally and functionally homologous to EGF (8). This growth factor is expressed in the kidney (9,10) and activates the ErbB1/HER1 EGF receptor (EGFR), which is strongly expressed by mesangial cells (11,12). HB-EGF is one of the most potent known mesangial cell mitogens (13) and is increased in a variety of renal pathologic conditions (14 -18). Local generation of HB-EGF by both glomerular epithelial and mesangial cells suggests the possibility of important intraglomerular autocrine and/or paracrine actions (13)(14)(15). However, the specific physiologic and pathophysiologic consequences of HB-EGF stimulation are largely undefined, and no effects on Glc metabolism have been described in any cell type that we are aware of. We therefore examined the influence of HB-EGF on HK activity and Glc metabolism in cultured murine mesangial cells. In addition, we sought to integrate our findings with previous demonstrations of regulated mesangial cell HK activity and to identify both HK isoforms and signaling intermediates that contribute to HK induction in this cell type.
Cell Culture-Mycoplasma-free SV40 MES 13 (murine mesangial) cells, which exhibit biochemical and morphologic features of normal mesangial cells in culture (5,19), were obtained from the American Type Culture Collection (Manassas, VA) at passage 27. Cells were maintained in Dulbecco's modified Eagle's medium/F12 medium (3:1) containing 6 mM Glc and supplemented with both 14 mM HEPES, pH 7.4, and 5% fetal bovine serum. Cell monolayers were routinely grown to confluence in a humidified 37°C/5% CO 2 incubator before testing, and all experiments were performed between passages 30 and 40 to minimize the effects of phenotypic variation in continuous culture. Where appropriate, cells were serum-deprived for 16 -24 h prior to and during testing. When inhibitors were employed, cells were typically pretreated with inhibitor alone for at least 30 -60 min. before testing. Identification of mesangial cell HK isoforms was performed in parallel using both SV40 MES 13 cells and cells cultured from glomerular explants of adult male Sprague-Dawley rats (provided by Dr. Ashok K. Singh, Hektoen Institute, Chicago, IL). The preparation and maintenance of the latter cells has been detailed previously (20). In selected experiments, we also examined GLUT1-overexpressing MCGT1 (rat mesangial) cells and their ␤-galactosidase-expressing MCLacZ counterparts (the gift of Dr. Charles Heilig, Johns Hopkins University, Baltimore, MD). These cells have been characterized previously and were maintained as described (21).
Hexokinase Activity Assays-HK activity was measured as the total Glc-phosphorylating capacity of whole cell extracts using a standard G6PDH-coupled spectrophotometric assay as described previously (5,6). Protein content was assayed by the method of Bradford (22) using purified bovine ␥-globulin as a reference standard. Specific HK activity was routinely expressed as units/g protein, where 1 unit corresponds to that level of enzyme activity resulting in the phosphorylation of 1 mol of Glc/min at 25°C. To facilitate comparisons, results were also expressed as percent activity relative to unstimulated time-paired control cells. A modification of this procedure was also employed to visualize HK activity attributable to individual HK isoforms following nondenaturing cellulose acetate electrophoresis (23-25) in a Zip Zone chamber (Helena Laboratories; Beaumont, TX). In situ detection of resolved HK activities was performed as described by Wilson (25). Lysates from cells overexpressing a heterologous HKI transgene were routinely analyzed in parallel as a reference standard. HKI overexpression in these cells was accomplished by adenoviral gene transfer as detailed below.
Ectopic Hexokinase Expression-The recombinant adenoviral expression vector used to overexpress the HKI isoform (rAd⅐HKI) has been characterized previously (26,27) and was the generous gift of Dr. Christopher Newgard (UT Southwestern Medical Center, Dallas, TX). In preliminary experiments, a ␤-galactosidase-expressing control vector (rAd⅐LacZ) obtained from the same source exhibited uniform transfection efficiency in this cell type. rAd⅐HKI viral stocks were routinely titrated to increase total HK activity in transfected cell monolayers approximately 2-to 3-fold within 24 h.
Lysate Preparation and Immunoblot Analysis-Whole cell lysates were prepared in cold 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerolphosphate, 1 mM Na 3 VO 4 , 1% (v/v) Triton X-100, 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 20 mM Tris⅐HCl, pH 7.5, before electrophoretic resolution via SDS-PAGE and transfer to nitrocellulose for immunoblotting. Blots were routinely stained with 0.1% (w/v) Ponceau S in 5% acetic acid to confirm both uniformity of gel loading and transfer between samples. Immunoblot analysis was performed as described previously (5). The rabbit polyclonal antisera ␣HK2-C18 was used to probe for HKII, and major findings were confirmed using either the goat polyclonal antisera ␣HK2-C14 or the rabbit polyclonal antisera ␣HK2-C9. HeLa cell lysates and/or recombinant human HKII (generously provided by Richard L. Printz and Daryl K. Granner, Vanderbilt University) were used as positive controls in all HKII immunoblot analysis. HKIII detection was accomplished using the monoclonal antibody C7C3 (29), and major results were confirmed using rabbit polyclonal antisera directed against human HKIII (␣HK3-C9). HKI immunoblotting was performed with goat (␣HK1-N19) or rabbit (␣HK1-C10 and ␣HK1-C11) polyclonal antisera. Brain lysates of murine, rat, or bovine origin were routinely employed as positive controls in HKI immunoblot analysis. With the exception of ␣HK2-C18, which could often be used at dilutions as low as 1:20,000, immunoblotting with all antibody preparations was routinely performed using a 1:1000 dilution of neat serum (or neat ascites, in the case of C7C3). Specificity was evaluated by parallel analysis substituting preimmune sera for, or omitting immune sera from, individual preparations. Specific protein bands were visualized and quantitated using a commercially available chemiluminescent detection system (Phototope, New England BioLabs) and NIH Image 1.62 software for Macintosh computers (National Institutes of Health, Bethesda, MD) as described previously (5,6).
Immunocytochemistry-Immunocytochemical analysis was performed on fixed SV40 MES 13 cells cultured on Lab Tek II TM multichamber glass slides (Nunc, Naperville, IL). Where appropriate, cells were washed with phosphate-buffered saline (PBS) before fixation in 4% (w/v) paraformaldehyde for 5-10 min. at 25°C. Fixed cells were then thoroughly washed with PBS before incubation in blocking and permeabilization buffer (PBS supplemented with 5% (v/v) goat serum, 0.3% (v/v) Triton X-100, and 1% (w/v) bovine serum albumin for 2 h at 25°C. Following the addition of primary antibodies (typically at a 1:100 -1:2000 dilution), cells were incubated overnight at 4°C and washed with PBS containing 0.3% (v/v) Triton X-100 and 1% (w/v) bovine serum albumin before incubation with matched secondary antibodies in the same buffer (10 g/ml for 1 h at 25°C). Samples were then thoroughly washed with PBS before mounting in Vectashield (Vector Laboratories, Burlingame, CA) or Cytoseal XYL (Stephens Scientific, Riverdale, NJ). Fluorescent immunocytochemical localization was achieved using a Ziess LSM 410 confocal laser scanning microscopy system (Carl Zeiss, Thornwood, NY) equipped with an Ar-Kr laser and Zeiss LSM4 software for Microsoft Windows NT. Postacquisition digital image processing was routinely performed using PhotoShop 5.5 software for Macintosh computers (Adobe Systems, San Jose, CA). Image acquisition parameters and postacquisition image processing were uniformly identical for a given set of paired samples. Preimmune sera or irrelevant antibodies were routinely employed as negative controls, and major results were confirmed using at least two independent antibody preparations.
Analysis of EGFR Tyrosine Phosphorylation-Immobilized ␣EGFR/ sc-03 or PY20 antibodies were employed to immunoprecipitate total EGFR and tyrosine-phosphorylated EGFR, respectively, from whole cell lysates. Immunoprecipitates were then probed in parallel with both ␣EGFR/sc-03 and PY20 or RC20, and quantitative immunoblot analysis was employed to estimate specific EGFR phosphorylation by a modification of the method of Riese et al. (30). Major results were also confirmed by standard EGFR and phospho-EGFR immunoblots using the same antibodies.
Analysis of ERK1/2 Phosphorylation and in Vitro Immunocomplex ERK1/2 Activity Assays-Specific ERK1/2 phosphorylation was evaluated by parallel assessment of ERK2 and phospho-ERK1/2 content of whole cell lysates via quantitative immunoblot analysis as described previously (5,6). ERK1/2 phosphotransferase activity was measured in whole cell lysates by a standard in vitro immunocomplex activity assay as described previously (5,6). In brief, activated ERK1/2 immunoprecipitates were prepared from cell lysates 1-60 min following stimulation and were analyzed for the ability to specifically serine-phosphorylate a chimeric Elk-1 fusion protein in vitro.
Glucose Utilization and Lactate Production Assays-Glc utilization and lactate production were assayed as the net disappearance of Glc and the net accumulation of lactate in the culture medium, respectively, as described previously (5,6). All measures of medium Glc and lactate content were performed in the presence of non-limiting concentrations of Glc (initial concentration ϳ6 mM) and under conditions of linear net Glc utilization and lactate accumulation.
Glucose Oxidation Assays-The oxidation of 14 C-radiolabeled Glc was monitored as described previously (31)(32)(33), albeit in the absence of artificial electron acceptors, using a simple modification of a classical CO 2 capture assay (34). All manipulations were performed at 37°C unless otherwise noted. In these experiments, individual cell monolayers (ϳ55 cm 2 ) were washed with divalent cation-deficient PBS, pH 7.4, before detachment and disaggregation by limited trypsinization (0.05% (w/v) trypsin, 0.5 mM EDTA) for ϳ3 min. in the same buffer. After pelleting at 250 ϫ g for 5 min., cells were resuspended in oxidation medium consisting of Dulbecco's modified Eagle's medium base (Sigma) containing 5.5 mM Glc and modified by both the equimolar substitution of NaCl for NaHCO 3 and supplementation with 14 mM HEPES, pH 7.4. Cell numbers in these suspensions were estimated using an improved Neubauer hemocytometer (Hausser Scientific) and adjusted to ϳ10 7 cells/ml. Aliquots (2 ml) were then transferred to 25-ml Erlenmeyer flasks equipped with center wells designed specifically for CO 2 capture experiments (Kimble/Kontes, Vineland, NJ). All incubations were ini-tiated by the direct addition of 1 Ci [1-14 C]Glc or [6-14 C]Glc (specific activity ϳ56.0 mCi/mmol) to the medium. Flasks were then sealed and continuously agitated at 40 rpm on a temperature-controlled orbital shaker. Reactions were terminated at appropriate time points by the addition of 0.5 ml of 2 N H 2 SO 4 to the medium via hypodermic needle. Liberated 14 CO 2 was trapped overnight in 0.2 ml of Carbo-Sorb E that was introduced into the empty center well in identical fashion following termination of the reaction. After dissolution in Permafluor Eϩ, trapped 14 CO 2 content was assayed in individual samples using a Tri-Carb 1600TR Liquid Scintillation Analyzer (Packard). In preliminary experiments, 14 CO 2 generation from 14 C-radiolabeled Glc was found to be linear for at least 3-4 h under these conditions, so all subsequent experiments were performed in this timeframe. Paired treated and untreated cells, analyzed in duplicate, were uniformly evaluated in parallel to facilitate direct comparisons, and cell-free control preparations were employed to correct for nonspecific background counts.
Other Assays-Total cellular reduced GSH content was assessed spectrophotometrically using a commercially available assay kit (Calbiochem) per the manufacturer's recommendations. PKC phosphotransferase activity was assayed as described previously (5).
Statistical Analysis-All data were expressed as the means Ϯ S.E. for at least three independent measures. Statistical comparisons were made by two-tailed paired t-testing or analysis of variance, where appropriate, using a significance level of 95% and StatView 5.0.1 software for Macintosh computers (SAS Institute, Cary, NC). Fig. 1, recombinant human HB-EGF increased total HK activity over 50% within 12 h (Fig. 1B) and at concentrations as low as 10 nM, (Fig. 1A). HK activity was still maximally elevated at 24 h, after which activity declined (data not shown).

Exogenous HB-EGF Increases Total HK Activity in a Concentration-and Time-dependent Manner-As shown in
Other EGF Receptor Ligands, but not IGF-1, Mimic the Effect of HB-EGF on HK Activity-To examine whether other EGFR agonists were capable of mimicking the effect of HB-EGF on HK activity, we first tested the ability of purified murine EGF to increase total HK activity in cultured murine mesangial cells. As shown in Fig. 1A, EGF fully mimicked the effect of recombinant human HB-EGF in this regard, albeit with a lower apparent EC 50 (0.3 nM versus 1.3 nM). Other EGFR ligands expressed by or with actions in the kidney were similarly tested for this ability. In three consecutive experiments, 20 nM TGF-␣, 20 nM AR, and 10 nM BTC increased total HK activity at 24 h by 78 Ϯ 6% (p Ͻ 0.01), 85 Ϯ 7% (p Ͻ 0.01), and 88 Ϯ 9% (p Ͻ 0.02), respectively. Paired control cells treated with 200 nM PMA or 10 nM HB-EGF exhibited corresponding increases of 91 Ϯ 7% (p Ͻ 0.01) and 71 Ϯ 13% (p Ͻ 0.04), respectively. Both exogenous HB-EGF and IGF-1 increase [ 3 H]thymidine uptake by cultured SV40 MES 13 cells. 2 Thus, the inability of IGF-1 to mimic the effect of HB-EGF suggests that increased HK activity is not a general mitogenic effect in these cells. This inter-2 A. K. Singh and R. B. Robey, unpublished observations. FIG. 1. HB-EGF increases mesangial cell HK activity in a concentration-and time-dependent manner. A, when tested at physiologic concentrations (Յ10 nM), HB-EGF and EGF, but not IGF-1, significantly increased HK activity at 24 h (*, p Ͻ 0.05). B, HB-EGF (10 nM) was capable of increasing total HK activity within 6 -8 h, and maximal increases of over 50% were observed within 12-24 h of stimulation (*, p Ͻ 0.05). IGF-1 (100 nM) failed to mimic these effects over the same time period when tested in parallel (data not shown). All data are presented as the means Ϯ S.E. for at least four independent measures.
Tyrosine Kinase Inhibition Prevents Increased HK Activity following HB-EGF Treatment-As shown in Fig. 2A, the phytoestrogen genistein, a general inhibitor of tyrosine kinases (35), completely prevented the increase in HK activity following HB-EGF treatment. This effect was maximal at genistein concentrations Ն25 M (apparent IC 50 2 M) and was not accompanied by a corresponding change in basal HK activity ( Fig. 2A). The tyrphostin PD153035, an EGFR-selective tyrosine kinase inhibitor (36 -38), also prevented increased HK activity without affecting basal activity (Fig. 2B). As reported previously, vehicle alone (Me 2 SO) had no independent effect on total HK activity in these cells at the concentrations employed (5).
EGFR Ligands, but Not Phorbol Esters, Acutely Increase EGFR Tyrosine Phosphorylation-To directly address EGFR activation, we evaluated the tyrosine phosphorylation status of immunoprecipitable EGFR in both unstimulated cells and in cells stimulated with either phorbol esters or EGFR ligands, e.g. EGF and HB-EGF. As shown in Fig. 3, EGFR agonists, but not phorbol esters, markedly increased tyrosine phosphorylation of the receptor within 5 min, consistent with immediate receptor activation. EGFR agonists uniformly increased receptor autophosphorylation at concentrations shown to increase total HK activity. In contrast, phorbol esters, which also increase HK activity in these cells, did not mimic this effect. These results were also confirmed by conventional anti-phosphotyrosine and phospho-EGFR immunoblot analysis of whole cell lysates (data not shown).
The Effects of HB-EGF and PMA on HK Activity Are Not Additive-Phorbol esters and thrombin increase mesangial cell HK activity via a common mechanism involving immediate PKC and classic MAPK pathway (Raf/MEK/ERK) activation (5,6). A similar time course for HK induction by HB-EGF is compatible with the hypothesis that all three agents share a common mechanism of induction. We therefore evaluated HB-EGF for the ability to augment phorbol ester-induced HK activity. In three consecutive experiments, 10 nM HB-EGF increased total HK activity by 85 Ϯ 18% (p Ͻ 0.05). The addition of 1 M PMA, which increases mesangial cell HK activity over 50% when administered alone (5), had no additional effect as the combination of these agents increased HK activity by 88 Ϯ 28% (p Ͻ 0.05; p ϭ 0.92 versus HB-EGF alone). Since the effect of HB-EGF was clearly not additive to that of PMA, a common mechanism of induction is suggested.
MEK Inhibition by PD98059 Prevents Increased HK Activity following HB-EGF Stimulation-The specific MEK1/2 inhibitor PD98059 is capable of blocking the induction of HK activity by both phorbol esters and thrombin (5,6). We therefore tested the ability of this inhibitor to prevent HK induction by HB-EGF. As shown in Fig. 4, PD98059 inhibited HB-EGF-inducible HK activity in a concentration-dependent manner. The effect of HB-EGF at 24 h was almost completely inhibited by PD98059 at concentrations Ն25 M (apparent IC 50 3 M). Basal HK activity in unstimulated cells is only slightly affected by PD98059 at these concentrations (5), suggesting that PD98059 does not directly inhibit HK activity.
HB-EGF Increases MEK-dependent ERK Activation-To further evaluate the involvement of the classic MAPK pathway, we directly tested the ability of HB-EGF to activate ERK1/2. As depicted in Fig. 5, HB-EGF increased ERK1/2 activity within 1 min, and maximal activation was observed within 5 min, after which ERK1/2 activity began to decline, albeit not to baseline during the 60 min examined. ERK1/2 activation was accompanied by parallel changes in ERK1/2 phosphorylation (Fig. 6A) and was fully inhibited by pretreatment with PD98059 (Fig. 5).
Phorbol Esters and HB-EGF Activate the Classic MAPK Pathway via PKC-dependent and PKC-independent Mechanisms, Respectively-Induction of mesangial cell HK activity by both phorbol esters and thrombin requires PKC activation (5,6). Since it has been reported that EGFR activation does not stimulate PKC in this cell type (39), we examined the dependence of HB-EGF's effects on PKC activation. Cellular PKC depletion by prolonged (24 h) antecedent exposure to 1 M PMA completely prevented increased ERK1/2 phosphorylation (Fig.   FIG. 2. Both general and EGFR-selective tyrosine kinase inhibitors prevent increased mesangial cell HK activity following HB-EGF treatment. Both genistein (A), a general inhibitor of tyrosine kinases, and the tyrphostin PD153035 (B), an EGFR-selective tyrosine kinase inhibitor, completely prevented increased HK activity following HB-EGF treatment. Neither inhibitor had a corresponding effect on basal HK activity at the concentrations employed for these studies. All data are presented as the means Ϯ S.E. for at least three independent measures (*, p Ͻ 0.05 versus unstimulated control cells and †, p Ͻ 0.05 versus HB-EGF-stimulated cells in the absence of inhibitors).

FIG. 3. EGFR ligands, but not phorbol esters, increase mesangial cell EGFR tyrosine phosphorylation.
A, whole cell lysates (1 mg of total protein) prepared from both unstimulated cells (Unstim) and cells stimulated with PMA (1 M), EGF (10 nM), or HB-EGF (10 nM) for 5 min were immunoprecipitated with phosphotyrosine-specific antibodies (IP: PY20). Immunoblot analysis of PY20 immunoprecipitates detected both EGFR (IB: ␣EGFR) and a tyrosine-phosphorylated species (IB: PY20) of identical size in lysates prepared from both EGF-and HB-EGF-stimulated cells, but not from either PMA-stimulated or unstimulated control cells. B, when EGFR immunoprecipitates (IP: ␣EGFR) of the same samples were immunoblotted in parallel using phosphotyrosine-specific antibodies (IB: PY20), similar results were obtained. The representative experiments depicted above were repeated at least three times with identical results. 6A) and activity (Fig. 6B) following PMA, but not HB-EGF, treatment. Consistent with these findings, the general PKCselective inhibitor bisindolylmaleimide I also prevented the effects of phorbol esters, but not HB-EGF (data not shown). These findings are compatible with our previous observations and are also consistent with the hypothesis that HB-EGF and PMA share a requirement for classic MAPK pathway activation, but not PKC activation, in increasing HK activity.
HB-EGF Increases Both Net Glc Utilization and Lactate Accumulation by Cultured Mesangial Cells-To examine the functional consequences of increased cellular Glc phosphoryl-ating capacity following HB-EGF treatment, we monitored both net Glc disappearance and net lactate accumulation in the culture medium of mesangial cells grown in the presence or absence of HB-EGF. As shown in Fig. 7B, 10 nM HB-EGF increased both Glc utilization and lactate accumulation more than 50% above basal levels, and these changes corresponded temporally to maximal HB-EGF-induced changes in HK activity. PMA, tested in parallel at a concentration of 1 M, produced similar changes. The net rate of Glc utilization was uniformly linear during these studies (Fig. 7A).
HB-EGF Increases Glc Oxidation via Both the Mitochondrial Tricarboxylic Acid Cycle and the Cytosolic PPP-14 CO 2 generation from [6-14 C]Glc ( 14 C(6)O 2 ) can be taken as a general index of Glc oxidation via the tricarboxylic acid cycle (31). 14 CO 2 production from [1-14 C]Glc ( 14 C(1)O 2 ) will similarly reflect tricarboxylic acid cycle flux, but will also reflect flux through the PPP. Thus, differences in 14 CO 2 yields from these differentially radiolabeled hexoses ( (5,6). To evaluate HB-EGF stimulation of HK activity for similar transcriptional dependence, we tested the ability of the general transcriptional inhibitors DRB and actinomycin D to attenuate this response. As shown in Table I, DRB inhibited HB-EGF-induced HK activity in a concentrationdependent manner, with 20 M DRB inhibiting the effect of 10 nM HB-EGF by over 75% at 24 h. Transcriptional inhibition by 1 g/ml actinomycin D or inhibition of protein translation by 10 g/ml CHX completely prevented this increase, suggesting that the effect of HB-EGF requires both ongoing gene expression and de novo protein synthesis. In contrast, basal HK activity was largely unaffected, and the modest observed decreases in basal activity (Ͻ20%) were not accompanied by changes in cellular trypan blue exclusion (data not shown), consistent with our previous reports of limited dependence upon ongoing gene expression for maintenance of basal HK activity (5,6).

HB-EGF Induction of HK Activity Requires Both Ongoing Gene Transcription and de Novo Protein Synthesis-We previously demonstrated that general inhibitors of gene transcription and protein translation prevent increased HK activity by activators of PKC
Elevated Ambient Glc Does Not Mimic the Effect of Phorbol Esters on HK Activity-Not all activators of PKC are capable of increasing mesangial cell HK activity (5), suggesting that HK induction by phorbol esters (5) and thrombin (6) represent specific, rather than general, responses to PKC activation. Since elevated ambient Glc is known to acutely increase PKC activity in this cell type (40), we also examined elevated Glc for the ability to increase total HK activity in these cells. In four independent experiments, total HK activity in SV40 MES 13 cells exposed to 16 mM or 26 mM Glc for 48 h (26 Ϯ 6 and 25 Ϯ 6 units/g protein, respectively) was not significantly different from that observed in control cells maintained in 6 mM Glc (27 Ϯ 3 units/g protein). Cell growth was not altered by elevated Glc in these experiments, but Glc-6-P content increased over 40% in cells maintained in 26 mM Glc for 24 -48 h (data not shown). Interestingly, HB-EGF increased cellular GSH content nearly 40% within 24 h (66 Ϯ 12 versus 48 Ϯ 9 mol/g protein; p Ͻ 0.02), whereas parallel exposure to 26 mM Glc   FIG. 4. MEK1/2 inhibition by PD98059 prevents HB-EGF-inducible HK activity. The specific MEK1/2 inhibitor PD98059 blocked increased HK activity following 10 nM HB-EGF treatment in a concentration-dependent manner (apparent IC 50 3 M). Basal HK activity is not similarly affected by PD98059 at the concentrations employed (5).

FIG. 5. HB-EGF rapidly activates ERK1/2 via a MEK1/2dependent mechanism.
HB-EGF increased ERK1/2 activity within 1 min as assessed by an in vitro immunocomplex kinase activity assay (5,6). ERK1/2 activation was maximal within 5 min and persisted for at least 1 h. This increase in ERK1/2 activity was completely abrogated by antecedent exposure to the MEK1/2-selective inhibitor PD98059. The representative experiment depicted above was repeated at least three times with identical results.
FIG. 6. HB-EGF activation of mesangial cell ERK1/2 involves a PKC-independent mechanism. A, total ERK2 and phospho-ERK1/2 were detected and quantitated in parallel by immunoblot analysis of lysates prepared from unstimulated control cells or stimulated cells exposed to either 10 nM HB-EGF or 1 M PMA. Both HB-EGF and PMA were capable of increasing the specific phosphorylation of ERK1/2 within 5 min. The effects of PMA, but not HB-EGF, were inhibited by antecedent PKC depletion. B, total ERK1/2 activity was also assayed in whole cell lysates as the ability of activated ERK1/2 immunoprecipitates to phosphorylate an Elk-1 fusion protein in vitro (Elk-1-P). Both 10 nM HB-EGF and 1 M PMA were capable of increasing ERK1/2 activity within 5 min, and changes in ERK1/2 activity uniformly paralleled changes in ERK1/2 phosphorylation. Representative experiments, each repeated at least four times with identical results, are depicted.
alone had no independent effect on GSH content (51 Ϯ 11 mol/g protein; p ϭ 0.58). PKC activity increased within 1 h of exposure to elevated Glc but returned to basal levels within 24 h (data not shown), as reported previously in this cell type (40). To exclude a rate-limiting effect of Glc transport in these experiments, we also examined GLUT1-overexpressing MCGT1 rat mesangial cells (21) in parallel. In three independent experiments, total HK activity in MCGT1 cells grown in 18 or 28 mM Glc for 48 h (25 Ϯ 1 and 24 Ϯ 1 units/g protein, respectively) was not significantly different from that observed in cells maintained in normal growth medium containing 8 mM Glc (25 Ϯ 2 units/g protein) or in ␤-galactosidase-expressing MCLacZ control cells (21) maintained in 8, 18, or 28 mM Glc for the same time period (25 Ϯ 2, 24 Ϯ 2, and 24 Ϯ 2 units/g protein, respectively).
Cultured Mesangial Cells Express All Three Renal HK Isoforms-To better understand the contributions of individual HK isoforms, fresh mesangial cell lysates were electrophoretically resolved in cellulose acetate and analyzed for Glc-ATP phosphotransferase activity. Resolved HK activities were visualized colorimetrically in situ via a coupled reaction with G6PDH in an agarose overlay in the presence of the redoxsensitive dye nitro blue tetrazolium (23,25). As shown in Fig.  8A, two prominent bands of activity corresponding to HKI and HKII were identified. Ectopic HKI expression was employed to confirm the relative positions of bands of activity, and cells overexpressing HKI uniformly showed an increase in the slower, but not the faster, migrating band of activity, suggesting identity between HKI and the former band. Neither band was observed in the absence of ATP and/or Glc (data not shown), suggesting specificity for HK activity. Although HKIII activity could not be resolved in these experiments, contributions by this isoform cannot be excluded on the basis of this assay alone, as HKIII is substrate-inhibitable at very low ambient Glc concentrations and is frequently not detectable by this method (24, 41, 42). Moreover, Wilson has demonstrated that this technique is not universally applicable for the detection and quantitation of individual isoform activities (25). We therefore performed parallel immunoblot analysis of cell lysates using isoform-specific antibodies directed against HKI, HKII, and HKIII. These studies confirmed the presence of all three ϳ100 kDa isoforms in SV40 MES 13 cells (Fig. 8B). HKI-specific antibodies identified a single protein band of appropriate size that co-migrated with the corresponding HKI band in immunoblots of mouse or rat brain lysates, and identical results were observed with rat mesangial cells cultured from glomerular explants (data not shown). Similarly, both HKII-and HKIII-specific antibodies recognized protein bands of the appropriate size. In the case of HKII, this band uniformly co-migrated with a recombinant poly-His-tagged HKII fusion protein. None of the antibodies directed against HKII or HKIII were cross-reactive with HKI of brain origin. Similarly, antibodies directed against HKI and HKIII did not cross-react with recombinant HKII. Identical results for these isoforms were also obtained in rat mesangial cells (data not shown). As depicted in Fig. 8C, immunocytochemical analysis of individual HK isoforms validated the results of the immunoblot analysis above and revealed distinct intracellular distribution patterns that are compatible with previous reports (29).
HB-EGF and PMA Selectively Increase HKII Expression-To examine the possibility that increased HK expression could contribute to increased cellular Glc phosphorylating capacity following growth factor or phorbol ester treatment (5, 6), we examined individual HK isoform abundances at 24 h when maximal activity was still observed. As shown in Fig. 9, both 1 M PMA and 10 nM HB-EGF increased SV40 MES 13 cell HKII protein content nearly 3-fold at 24 h. Similar changes were not observed for HKI, HKIII, or the ubiquitous facilitative Glc transporter isoform GLUT1 (data not shown), suggesting specificity for HKII. Similar results were obtained in SV40 MES 13 cells at 18 h and in cultured rat mesangial cells at 24 h (data not shown). In each case, increased HKII protein expression FIG. 7. HB-EGF stimulates mesangial cell Glc metabolism. HB-EGF (10 nM) increased both net Glc utilization and net lactate accumulation by cultured mesangial cells. These changes were observed between 12 and 24 h following HB-EGF stimulation and temporally correlated with maximal stimulation of HK activity. A, net cellular Glc utilization, measured as medium Glc disappearance, was uniformly linear (r 2 Ͼ 0.9, where r ϭ correlation coefficient for linear regression analysis) for at least 6 h with an initial Glc concentration of 6 mM. Antecedent exposure to 10 nM HB-EGF for 12-24 h was also associated with an increase in the net rate of Glc utilization. Data from a representative experiment, repeated at least twice with identical results, are depicted. Similar results were also obtained in the presence of 16 -26 mM Glc (data not shown). B, both 10 nM HB-EGF and 1 M PMA increased net lactate accumulation in parallel with net Glc utilization, and the consistent relationship between Glc utilization and lactate accumulation was not suggestive of metabolic uncoupling by either phorbol esters or EGFR agonists. All data are presented as the means Ϯ S.E. for at least six independent experiments.

TABLE I Inhibition of HB-EGF-inducible HK activity by general inhibitors of gene transcription and protein translation
The number of individual measures used to calculate the means Ϯ S.E. are indicated in parentheses. *, p Ͻ 0.05 vs. control cells, and †, p Ͻ 0.05 vs. HB-EGF-stimulated cells.

No inhibitors
10 M DRB 20 M DRB 1 g/ml Act D 10 g/ml CHX was completely prevented by the MEK-selective inhibitor PD98059 (data not shown), consistent with a role for the classic MAPK pathway in this induction. DISCUSSION We have previously demonstrated that both phorbol esters and thrombin increase HK activity and Glc metabolism in cultured mesangial cells (5,6). These effects were found to share a common requirement for PKC and classic MAPK pathway activation and were dependent upon both ongoing gene expression and de novo protein synthesis. Increased HK expression was suggested, but the specific HK isoforms responsible for these changes were not identified. In the present work, we have shown that EGFR ligands, such as HB-EGF, are similarly capable of inducing mesangial cell HK activity, and we have provided evidence to support the contention that increases in HKII isoform abundance are largely responsible for these changes. The associated increases in Glc metabolism suggest that our observations are biologically relevant. HK induction by HB-EGF was found to require classic MAPK pathway but not PKC activation, suggesting that the classic MAPK pathway serves a common downstream effector function for both PKC-dependent and PKC-independent mechanisms of induction. The nonadditive effects of HB-EGF and PMA on total HK activity are compatible with this interpretation, as is their common ability to increase HKII abundance. Our studies specifically address the role of EGFR activation and signaling in the modulation of mesangial cell HK abundance and activity, effects that have not previously been described in any cell type that we are aware of. They do not, however, address the relative contributions of HB-EGF versus other EGFR ligands, e.g. EGF, TGF-␣, AR, or BTC, that are expressed by or have actions in the kidney (43)(44)(45).
Both HKI and HKIII expression have been reported within glomeruli, although the specific cell types expressing these isoforms have not been identified (29,46). We have recently reported, in preliminary form, the glomerular expression of HKII (47), but the specific intrarenal and intraglomerular localization of this expression has also not been defined. In the present work, we have clearly demonstrated for the first time that cultured mesangial cells express all three renal HK isoforms. We have also shown that both phorbol esters and HB-EGF selectively increase HKII abundance in these cells. The demonstrated changes suggest a role for regulated HKII gene expression in mediating the corresponding increases in both HK activity and Glc metabolism. Our findings are thus consistent with the hypothesis that HKII constitutes the principal inducible HK isoform in mesangial cells. FIG. 8. Cultured mesangial cells express all three 100-kDa mammalian HK isoforms. A, following nondenaturing electrophoretic resolution in cellulose acetate, zones of individual HK isoform activity were detected by in situ chromophore generation coupled to the HK-, Glc-, and ATP-dependent reduction of NADP by G6PDH. The representative assay depicted above was performed in the presence of 0.5 mM Glc, and the relative positions of the anode (ϩ) and cathode (Ϫ) are marked on each lane. Similar results were obtained in the presence of 3.7 mM Glc. To help establish the identity of individual bands of activity, a heterologous HKI transgene was overexpressed in SV40 MES 13 cells and analyzed in parallel. Although activity attributable to HKIII could not be visualized or resolved by this method, contributions by this isoform cannot be excluded on the basis of this test alone (see "Results"). B, immunoblot analysis of whole cell lysates using ␣HK1-N19, ␣HK2-C18, and C7C3 revealed the presence of HKI, HKII, and HKIII isoforms, respectively. These results were confirmed by parallel testing with independent HK isoform-specific antisera (data not shown). The relative positions of ϳ83-kDa and ϳ175-kDa molecular size markers are indicated. C, the presence of all three 100-kDa isoforms were also confirmed via in situ immunocytochemistry using the same antibody preparations employed in B. Representative paired results from a single experiment are depicted. Irrelevant IgG control antibodies were substituted for HK-specific antisera as a negative control in this experiment, which was repeated at least four times with identical results.
FIG. 9. Both phorbol esters and HB-EGF selectively increase mesangial cell HKII abundance. A, SV40 MES 13 cells were found to express an ␣HK2-C18-immunoreactive protein ϳ100 kDa in size that increased in abundance within 24 h of exposure to either 1 M PMA or 10 nM HB-EGF. B, densitometric quantitation of this specific HKII immunoreactivity revealed an ϳ3-fold increase in HKII protein abundance at 24 h (*, p ϭ 0.05; n ϭ 6), and similar results were obtained at 18 h (data not shown). These findings were verified using independent HKII-specific ␣HK2-C14 antisera (A, inset).
In muscle and adipose tissue, which together account for the bulk of peripheral Glc uptake and utilization, HKII expression is regulated by insulin (48). In rat myotubes, this regulation does not require classic MAPK pathway activation (49). It is therefore of considerable interest that increased HK activity and HKII isoform expression in mesangial cells is dependent upon this pathway. It is of additional interest that the major regulators of HKII expression in muscle and adipose tissue, e.g. insulin, IGF-1, adrenomimetics, and glucorticoids, do not appear to contribute substantively to the regulation of mesangial cell HK activity. 3 These regulatory differences may have an important pathologic correlate in diabetes, where the renal cortex is characterized by a state of Glc overutilization and increased HK activity that contrasts markedly with the Glc underutilization and decreased Glc phosphorylating capacity observed in other end-organ targets of disease such as muscle and adipose tissue (2, 50 -53). As the principal renal cell type affected by diabetes, mesangial cells provide a suitable model to address the underlying mechanisms and functional significance of these differences. The inability of the principal hallmarks of diabetes, i.e. altered insulin and Glc, to directly affect HK activity in these cells suggests independent mechanisms of coupling increased renal HK activity to the diabetic state. HB-EGF-regulated HK activity is particularly interesting in this regard, because this growth factor has been implicated in the pathogenesis of a number of diabetic end-organ derangements, including renal disease (10) and macroangiopathy (54). Renal HB-EGF expression increases within the first week of experimental diabetes and is maximally induced by 4 -6 weeks (10), which is fully compatible with a role for HB-EGF in the metabolic changes associated with diabetes in the renal cortex. Although not directly addressed in the present work, it would be attractive to speculate that fundamental differences in the regulation of HK activity and HKII expression by factors such as HB-EGF may contribute to these directionally opposite metabolic responses in different end-organ targets of disease.
Our findings also have pathophysiological implications outside the context of diabetes that warrant mention. HB-EGF has been implicated in the pathogenesis of glomerular injury of diverse etiologies, and increased glomerular HB-EGF expression has been reported in anti-Thy-1.1 mesangial proliferative glomerulonephritis (14,16), acute puromycin aminonucleoside nephrosis (15), passive Heymann nephritis (15), anti-glomerular basement membrane glomerulonephritis (18), and experimental focal glomerular sclerosis (17). A recent study also reported increased glomerular HB-EGF expression in human biopsy material taken from patients with glomerular diseases as diverse as lupus nephritis, membranoproliferative glomerulonephritis, IgA nephropathy, and Schönlein-Henoch purpura (16). In the case of puromycin aminonucleoside nephrosis (15), increased glomerular HB-EGF expression may explain, in part, the increased glomerular HK activity reported by Dubach and Recant in this model over forty years ago (1). Although the functional significance of these changes have not been defined and pathogenetic contributions cannot presently be excluded, recent demonstrations of reparative and cytoprotective effects of HB-EGF and other EGFR ligands in both renal (55,56) and other tissue (57,58) injury models suggest that these growth factors, and HB-EGF in particular, may also play an important role in the response to, rather than the mediation of, acute glomerular injury (59).
Adaptive roles in cell injury have been proposed for HK activity in both pulmonary cells (24,41) and glomerular mesangial cells (5, 6) on the basis of intrinsic regulatory features. Although not directly tested in either model, the present findings are compatible with such a hypothesis. 14 CO 2 yields from [1-14 C]Glc in mesangial cells were comparable in magnitude to that previously reported for isolated white adipose cells (60), suggesting a substantial capacity for Glc oxidation. The ability of HB-EGF to increase 14 CO 2 generation from both [1-14 C]Glc and [6-14 C]Glc is consistent with increased oxidation via the mitochondrial tricarboxylic acid cycle. Although flux through this pathway may be associated with increased reactive oxygen species generation, the parallel, and proportionately larger, increase in the differential 14 CO 2 yield from [1-14 C]Glc and [6-14 C]Glc ( 14 C(1)O 2 Ϫ 14 C(6)O 2 ) also suggests increased PPP flux, which could serve to buffer the cellular redox status against oxidant stress. A parallel increase in cellular GSH content is compatible with this interpretation and suggests at least one mechanism whereby increased HK activity could contribute to adaptive responses to cell injury. Corresponding increases in net lactate accumulation also suggest increased glycolytic flux accompanying increased oxidative tricarboxylic acid cycle activity. HKs are known to bind mitochondria and directly couple Glc metabolism to oxidative phosphorylation (61)(62)(63), which may facilitate the coordination of glycolytic and oxidative Glc metabolism within the cell (62). We have recently demonstrated that this interaction also promotes mitochondrial viability and cell survival and contributes to the antiapoptotic effects of growth factors in fibroblasts (64). Moreover, we have found that increased HK activity can mimic the salutary effects of growth factors in preventing acute oxidant-induced cell death in renal epithelial cells (56). Since oxidative stress is a common pathologic feature of many forms of glomerular injury, these findings suggest additional potential adaptive functions for HKs. Given the central role of Glc in cellular energy metabolism, increased HK activity also has specific energetic implications. In principle, HKs are ideally positioned 3 R. B. Robey, unpublished observations. The regulation of mesangial cell HK activity by EGFR ligands involves both novel and common regulatory effectors and culminates in selectively increased HKII isoform expression. The interrelationships between other known effectors (5,6) and their relevance to the present work are depicted schematically above. Both EGFR agonists, e.g. EGF and HB-EGF, and activators of PKC, e.g. phorbol esters and thrombin, are capable of increasing mesangial cell HK activity. The depicted PKC-dependent and PKC-independent mechanisms both involve classic MAPK pathway activation and increased HKII isoform expression. The influences of general tyrosine kinase inhibition (genistein), EGFR inhibition (PD135035), classic MAPK pathway inhibition (PD98059), inhibition of gene transcription (DRB, Act D), and inhibition of protein synthesis (CHX) are also shown.
to influence not only the magnitude but also the direction of metabolic Glc flux, and we have presented evidence in support of this contention. In general, increased HK activity would also be predicted to decrease free intracellular Glc levels and shunt Glc flux away from the polyol pathway, which have both been implicated in the pathogenesis of diabetic nephropathy. Taken together, our findings are clearly relevant to the understanding of growth factor action and metabolic coupling to injury in mesangial cells. Although presently speculative, they also suggest a number of nonexclusive mechanisms whereby HKs may mediate adaptive functions attributable to growth factors in this cell type.
In conclusion, our demonstration of ERK-dependent regulation of HKII expression represents a previously unrecognized mechanism of HKII regulation that is relevant to the understanding of metabolic responses to both glomerular injury and growth factor stimulation. Why HKII expression in mesangial cells exhibits regulatory behavior that is markedly different from that observed in insulin-sensitive peripheral tissues, e.g. muscle or adipose, represents an extremely important question and suggests cell-specific determinants. However, additional studies will be required to elucidate both the molecular basis and functional significance of these differences.