Downregulation of Protein Kinase Cδ Activity Enhances Endothelial Cell Adaptation to Hypoxia
Background—Although protein kinase C (PKC) has been implicated in ischemic cell death, the role of individual PKC isoenzymes in the response of endothelial cells (ECs) to hypoxia is unknown.
Methods and Results—To test the effect of hypoxia on the activity of individual PKC isoenzymes, human ECs were exposed to 95% N2 with 5% CO2 for 24 hours. This severe hypoxia reduced PKCδ specific activity in both human umbilical vein ECs (HUVECs) and a HUVEC-derived EC line (ECVs) significantly (80.5±5.7% and 55.5±8.6% of normoxia controls, respectively); the activities of PKCα and PKCε were unchanged. The protein levels of PKCα, PKCδ, and PKCε were unchanged by hypoxia. To determine whether PKCδ downregulation by hypoxia was linked to EC function, ECVs in which PKCδ was stably overexpressed (PKCδ-ECs) were exposed to hypoxia. A significant increase in cell death was observed in PKCδ-ECs compared with controls (5.8±0.6% versus 2.3±0.4% at 24 hours, 13.2±1.2% versus 4.1±0.4% at 48 hours, P<0.05) during hypoxia. Neither the DNA laddering assay nor TUNEL staining revealed an increase in apoptosis of PKCδ-ECs exposed to hypoxia, suggesting a hypoxia-induced increase in nonapoptotic cell death of PKCδ-ECs. Inhibition of NO synthase with NG-monomethyl-l-arginine (L-NMMA) affected neither the decline in PKCδ activity nor the EC death induced by hypoxia.
Conclusions—PKCδ activity is decreased by hypoxia by a mechanism that does not involve NO synthase; this downregulation appears to enhance EC survival during hypoxia by decreasing nonapoptotic cell death.
Subjecting the endothelium to hypoxia increases vascular permeability, thrombogenicity, leukocyte adhesion, and the production of proinflammatory cytokines and impairs the control of vasomotor tone.1 2 3 These effects of hypoxia are sustained in vivo even after reoxygenation, a phenomenon that may be explained by decreased NO and cAMP levels.3 Hypoxia induces cell proliferation under certain experimental conditions,4 5 but severe hypoxia results in cell death.6 7 8
In various cell types, the protein kinase C (PKC) family members have been found to be important mediators of hypoxia-induced changes.4 5 9 Previously, PKC has been reported to be involved in hypoxia-mediated cell proliferation in both mesangial cells4 and smooth muscle cells derived from the pulmonary artery5 and in enhanced monocyte migration into endothelial cells (ECs).9 Increased release of platelet-activating factor during hypoxia may upregulate PKC activity, which in turn might increase phosphorylation of platelet and endothelial cell adhesion molecule in ECs and thus enhance monocyte migration.9 PKC also mediates various aspects of EC function; however, whether PKC mediates or protects against hypoxia-induced EC death is not understood. Only limited information is available about the role of individual PKC isoenzymes in hypoxia; in the majority of studies, activators or inhibitors of PKC with nonselective, or at best unknown, effects on specific isoenzymes were used to investigate the role of PKC in hypoxia. In cardiomyocytes, PKCα and PKCε translocate from soluble to particulate fractions of the cell with hypoxia; conversely, PKCδ dissociates from the particulate to the soluble portions under similar conditions, although the total amount of intracellular protein corresponding to all 3 individual PKC isoenzymes remains constant.7 PKCε may prevent hypoxia-induced cell death by triggering a preconditioning mechanism6 in which a hypoxic injury can be attenuated by a preceding short period of exposure to reduced oxygen tension.10 11 Whether hypoxia alters the activity of selective PKC isoenzymes in ECs is unknown.
In this study, we found that the activity of PKCδ was specifically suppressed by hypoxia, whereas that of PKCα and PKCε were not. To assess whether PKCδ suppression was required for ECs to survive during hypoxia, we tested the effects of hypoxia on PKCδ-overexpressing ECs. We also investigated the role of NO and heat-shock proteins (HSPs), which protect various cell types from hypoxia-induced injuries.12 13
Cell Cultures, Hypoxia, and Other Materials
Human umbilical vein ECs (HUVECs) were cultured as previously described.14 Immortalized human ECs (ECVs, obtained from American Type Culture Collection) were cultured in M-199 (Gibco BRL Products) supplemented with 10% FBS (Gemini Bio Products Inc) and antibiotics (100 U/mL penicillin, 100 g/mL streptomycin) on 100-mm dishes. To expose ECs to hypoxia, subconfluent ECs in a 100-mm dish were incubated in either 5 mL of reduced-serum medium containing 5% newborn calf serum without endothelial cell growth supplements (for HUVECs) or 4 mL of serum-free medium (for ECVs) and placed in an airtight culture chamber. The chamber was ventilated with 95% N2 and 5% CO2 for 15 minutes, then switched to a closed circuit to maintain the N2 gas mixture in the chamber and incubated at 37°C for 24 to 48 hours. ECs incubated in the same amount of serum-free medium in normal oxygen conditions served as controls for each experiment. For experiments in which an NO synthase inhibitor was used, NG-monomethyl-l-arginine (L-NMMA, Sigma Chemical Co) was added to the medium at a final concentration of 250 μmol/L to 1 mmol/L. The full-length cDNA of PKCδ was cloned into the pcDNA3 mammalian expression vector (Invitrogen) by use of the EcoRI restriction endonucleotide enzyme site. The construct was transfected into ECs by the lipofectin method (Life Technologies Inc). ECs that stably expressed the empty vector or PKCδ were selected by resistance to neomycin.
Immunoblot Analysis and Translocation of PKC Isoenzymes
Cell lysates were prepared by addition of 2 mL of lysis buffer (PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 57.4 mol/L phenylmethylsulfonyl fluoride, 2 μg/mL aprotinin, 4.2 mol/L leupeptin) per 1×107 cells. Immunoblots were performed as previously described.15 Polyclonal antibodies were used to immunoblot PKCα, PKCε, and HSP27 (Santa Cruz Biotechnology Inc), and mouse monoclonal antibodies were used to immunoblot PKCδ, HSP70, and HSP90 (Transduction Laboratories). To determine intracellular distribution of PKC isoenzymes, the particulate and cytosolic fractions of ECs were separately collected by differential centrifugation,16 and the percentages of individual PKC isoenzymes in each fraction were calculated as previously reported.16
Kinase Activity Assay
Subconfluent monolayers of ECs (in 100-mm plates) were treated with trypsin, counted, and then washed with PBS. ECs (4.0×106 for HUVECs or 2.0×106 for ECVs) were resuspended in 1 mL of ice-cold lysis buffer for 10 minutes and then were homogenized by repeated aspiration through a 21-gauge needle. Cell debris was removed by centrifugation at 3500 rpm at 4°C for 15 minutes.
To determine the specific activity of PKCα, PKCδ, and PKCε, each individual PKC isoenzyme was immunoprecipitated from the whole-cell lysate with the same anti-PKC specific antibody as used for immunoblottings, and the kinase assay was carried out according to methods described previously.15 In the experiments that assessed PKCδ and PKCε, the reaction mixture did not contain additional calcium acetate. The presence of individual PKC isoenzymes was confirmed by immunoblotting with an anti-PKC isoenzyme specific antibody. PKC activity was normalized to the cell number and expressed as the percentage of simultaneously measured PKC activity of ECs cultured under normoxic conditions.
Cell Proliferation Analysis
EC growth was determined by counting the cells with a hemocytometer under ×50 magnification. Subconfluent cells were seeded at 2.0×105 per 16-mm plate in 1 mL of medium. At 24 hours after seeding, the ECs were exposed to either hypoxia or normoxia with 0.3 mL serum-free medium. After a 24- to 48-hour incubation period, ECs were harvested for counting.
Trypan Blue Exclusion Analysis
After exposure to hypoxia, ECs in 100-mm dishes were gently treated with trypsin, suspended, and mixed with the same volume of 0.4% trypan blue solution (Gibco BRL Products). Percentages of viable cells were evaluated under the field of a light microscope and normalized to the total cell number in the field.
DNA Ladder Detection Assay
ECs were harvested, washed with PBS, digested with 1 mL of lysis buffer [10 mmol/L Tris (pH 8.0), 100 mmol/L NaCl, 25 mmol/L EDTA, 0.5% SDS, 0.1 mg/mL protease K (Gibco BRL Products)] overnight at 37°C. Genomic DNA was precipitated with isopropanol after phenol chloroform precipitation. Equal quantities of each sample (20 to 30 μg) were subjected to electrophoresis on 1.25% agarose gels containing 0.5 μg/mL ethidium bromide.
Terminal Deoxynucleotidyltransferase Nick-End Labeling
Subconfluent ECs in 100-mm dishes were treated with trypsin and washed with PBS. Approximately 1×106 ECs were fixed with 1 mL of 4% paraformaldehyde in PBS for 10 minutes, then centrifuged and resuspended with 80% ethanol for 24 hours. ECs were placed on slides and air-dried overnight. Terminal deoxynucleotidyltransferase (TdT) nick-end labeling (TUNEL) assay was performed on slides with a Trevigen TACS 2 TdT (TBL) kit (Trevigen). For each slide, the number of TUNEL-positive EC nuclei in a 0.0625-mm2 area was scored in 12 randomly chosen high-power fields with a grid (×400). The number of TUNEL-positive nuclei was normalized to the number of total cells counted.
All data are presented as mean±SEM. The effects of different time points and cell lines were compared by ANOVA. Multigroup comparison was carried out with Bonferroni-modified t tests. The comparisons between normoxia and hypoxia in the same cell lines were performed with an unpaired Student’s t test. Probability values <0.05 were accepted as statistically significant.
Individual PKC Isoenzyme Activity of ECs in Hypoxia
Figure 1⇓ shows individual PKC activity normalized to normoxic ECs under similar conditions after 24 hours of hypoxia in HUVECs (A) and ECVs (B). The activity of PKCα did not change significantly with hypoxia; however, PKCδ specific activity markedly decreased to 80.5±5.7% and 55.5±8.6% of that in normoxic ECs in HUVECs (n=5, P<0.05 versus normoxic controls) and HUVEC-derived ECVs (n=8, P<0.01 versus normoxic controls), respectively. Although PKCε expression was detected in both HUVECs and ECVs by immunoblotting, its activity could be measured only in ECVs, in which PKCε activity was unchanged by 24 hours of hypoxia.
For comparison, we also assessed activation of individual PKC isoenzymes by their ability to translocate as detected by cell fractionation and immunoblotting. The expression of PKCδ in cytosolic fractions increased significantly with 24 hours of hypoxia in both HUVECs (Figure 2A⇓) and ECVs (Figure 2B⇓) compared with normoxic controls. Neither PKCα nor PKCε translocated significantly. These changes could be correlated with the changes in enzymatic activity of individual PKC isoenzymes seen in hypoxia (Figure 1⇑). Conversely, the total protein levels of PKCα, PKCδ, and PKCε did not change over a 24-hour period of hypoxia by immunoblotting (Figure 3⇓). Therefore, the decrease in PKCδ activity in hypoxia resulted from posttranslational mechanisms, including translocation of intracellular distribution.
Mechanism of Hypoxia-Induced EC Death in HUVECs and ECVs
We examined whether hypoxia injures ECs in a similar fashion in both HUVECs and ECVs. Forty-eight hours of hypoxia induced total cell death in 11.8±0.4% of HUVECs (n=4) and 5.9±1.0% of ECVs (n=6). This degree of hypoxia also provoked apoptotic ECs in 2.7±0.3% of HUVECs (n=5) and 1.1±0.2% of ECVs (n=5). The magnitude of EC death by hypoxia in ECVs was half that of HUVECs; however, we found that the large proportion of EC death was due to nonapoptotic cell death in both HUVECs and ECVs. Furthermore, HUVECs and ECVs both demonstrated a similar ratio of apoptotic ECs to total EC death (23% in HUVECs and 19% in ECVs). Thus, although it is a cell line, ECVs demonstrated characteristics of EC death by hypoxia that were similar to those of HUVECs.
EC Proliferation in Hypoxia
We examined the effect of hypoxia on ECV proliferation. The number of ECVs incubated in serum-free medium increased over a 2-day period in normoxia in wild-type and vector controls and also in PKCδ-overexpressing ECVs (PKCδ-ECs), but the number of ECVs subjected to 2 days of hypoxia dropped significantly compared with normoxic ECVs. This decrease was more severe in PKCδ-ECs, as shown in Figure 4⇓. This suggests that PKCδ-ECs are more susceptible to hypoxia-induced cell death.
Effect of PKCδ Overexpression on Hypoxia-Induced EC Death
To examine whether PKCδ suppression by hypoxia affected ECV survival during hypoxia, we used trypan blue staining to determine the number of ECVs that died after 2 days of hypoxia. PKCδ-ECs demonstrated a significant increase of EC death compared with vector controls at 24 and 48 hours of hypoxia (5.8±1.8% versus 2.3±1.2%, P<0.05, at 24 hours of hypoxia and 13.2±1.2% versus 4.1±0.4%, P<0.05, at 48 hours of hypoxia, Figure 5A⇓). PKCδ-ECs showed a somewhat higher baseline cell death compared with vector controls at 24 hours of serum-free normoxia; however, baseline cell death was comparable to that seen at 48 hours of normoxia (2.7±0.4% versus 2.7±0.3%, P=NS). Our results suggest that PKCδ suppression during hypoxia may contribute to EC survival.
Effect of PKCδ Overexpression on Hypoxia-Induced EC Apoptosis
To investigate whether this increase of cell death in PKCδ-ECs during hypoxia was due to apoptotic cell death, we assessed the number of TUNEL-positive cells after 24 and 48 hours of hypoxia. A small population of ECVs underwent apoptosis at either 24 or 48 hours of hypoxia (Figure 5B⇑), as previously discussed. PKCδ-ECs did not demonstrate an increase in apoptotic cell death above that seen with vector controls at either 24 or 48 hours of hypoxia (0.7±0.1% versus 1.1±0.2% and 1.6±0.3% versus 1.1±0.1%, respectively). To confirm these findings, a DNA laddering assay was performed in ECVs subjected to up to 48 hours of hypoxia. As shown in Figure 6⇓, no laddering was seen in either normoxia- or hypoxia-treated ECVs in either wild-type cells, vector controls, or PKCδ-ECs. Although it is possible that neither assay is sufficiently sensitive to detect apoptosis, these results suggest that PKCδ suppression by hypoxia enhances cell survival mainly by preventing nonapoptotic cell death rather than apoptosis.
Effect of NO Synthase Inhibition and HSP Expression on Hypoxia-Induced PKC Specific Activity and Hypoxia-Induced EC Death
Because we have reported that PKCδ activity can be downregulated by activation of NO synthase in ECs stimulated by vascular endothelial growth factor,17 we tested whether the hypoxia-induced decrease of PKCδ activity was mediated by NO synthase. In such a case, NO synthase inhibition might exacerbate hypoxia-induced EC death by reversing PKCδ suppression. No reversal of hypoxia-induced PKCδ suppression was seen, however, in the presence of L-NMMA in concentrations ranging from 250 μmol/L to 1 mmol/L [58.9±2.4% of normoxic controls at 250 μmol/L (n=4), 41.1±10.0% at 500 μmol/L (n=4), and 56.4±10.3% at 1 mmol/L (n=6)]. Thus, our data suggest that the hypoxia-induced decrease in PKCδ activity is mediated through a pathway that does not involve NO synthase. To determine whether NO inhibition might affect ECV death during hypoxia by a mechanism independent from PKCδ activity, the NO inhibitor L-NMMA (500 μmol/L) was added. Such an addition did not further increase ECV death assessed by trypan blue exclusion, nor did it enhance apoptotic cell death in wild-type ECVs at either 24 or 48 hours of hypoxia (Figure 7⇓). This suggests that inhibition of NO does not increase ECV death during hypoxia by a mechanism independent from that resulting from sustained expression of PKCδ. Similarly, immunoblots of HSP27, HSP70, and HSP90, the latter of which, in particular, has recently been identified as a coactivator of NO synthase,18 revealed that these proteins were not upregulated by 24 hours of hypoxia in either wild-type or vector controls or in PKCδ-ECs, although 60 minutes of 42°C heat exposure to ECVs increased expression of these HSPs in these ECs (Figure 8⇓). Furthermore, HSP27, HSP70, and HSP90 were not upregulated in HUVECs by 24 hours of hypoxia (data not shown). Thus, induction of these HSPs is not an essential function of EC adaptation to hypoxia.
The major finding of our study is that the specific activity of PKCδ is decreased by hypoxia in human ECs in both primary cultured human ECs (HUVECs) and human EC lines (ECVs). After we confirmed that ECVs preserved the characteristic EC injury pattern of HUVECs by hypoxia, we overexpressed PKCδ in ECVs to investigate further the physiological role of decreased PKCδ activity in ECs. The enhancement of hypoxia-induced EC death by overexpression of PKCδ activity, which prevents the hypoxia-induced decrease in its activity, suggests that the decrease in PKCδ activity favors EC survival during hypoxia by preventing nonapoptotic cell death. The fact that PKCα and PKCε activities were not altered by hypoxia as determined by either enzymatic assay or translocation study suggests that these effects are specific for PKCδ.
The time course of the decrease in PKCδ isoenzyme activity in hypoxic ECs is consistent with the time course of the loss of the association of this isoenzyme with the membrane of cardiomyocytes subjected to hypoxia.7 In that experiment, as in ours, the protein levels of each PKC isoenzyme were not altered by 24 hours of hypoxia. Thus, the decrease of PKCδ activity as a consequence of hypoxia appears to result from a posttranslational process. Our translocation studies further support this finding. Thus, it seems that the individual PKC isoenzymes can be altered in the physiological or pathological setting by posttranslational mechanisms.
This study also demonstrates that overexpression of PKCδ in ECs increases nonapoptotic cell death but is not associated with an increase in hypoxia-induced apoptosis of ECs over that seen in control ECs. The observation is in contrast to that seen in human myeloid cancer cells,19 in which PKCδ activation is required for ionizing irradiation to induce apoptosis. The increase in nonapoptotic cell death is seen predominantly in the presence of hypoxia; only a minimal increase is seen in the PKCδ-ECs not exposed to hypoxia. Thus, it is likely that the decrease in PKCδ by hypoxia is a defense mechanism that preserves ECs by preventing nonapoptotic cell death.
The mechanism of how PKCδ overexpression increases nonapoptotic cell death in the presence of hypoxia is not clear. PKCδ activates Ras-dependent signals, including AP1/Jun and MAP kinase.20 Thus, in wild-type ECs, suppression of PKCδ by hypoxia may be beneficial by decreasing the activation of the MAP cascade, which in turn decreases fos and jun,21 which may directly elicit changes in the expression of other genes that promote cell death. Also, increased PKCδ activity both causes ECs to enter the S phase inappropriately and slows the exit of ECs from the S phase.22 It is possible that growth signals induced by hypoxia,4 5 together with the accumulation of ECs in the S phase, exert an additive or synergistic effect to initiate cell death pathways, even though neither insult is itself sufficient to cause cell death. Neither the decrease in PKCδ activity nor the increase in EC death by hypoxia was mediated by NO synthase. NO can decrease activity after other stimuli17 and thus might be expected to enhance EC survival in hypoxia via suppression of PKCδ. NO can also downregulate the oxygen demand of cells23 and thus may be beneficial for cell survival during hypoxia. In our study, however, we revealed that PKCδ suppression by hypoxia is not mediated by NO synthase, nor does inhibition of NO further increase EC death or apoptotic cell death, suggesting that NO is not crucial for ECs to survive under hypoxia.
HSPs have been reported to protect against ischemia- or hypoxia-induced injury in various cell types.12 13 The potential functions of HSPs include suppression of proinflammatory cytokines, reduction of the oxidative burst, NO-mediated prevention of apoptosis, and collagen synthesis.24 Among HSPs, HSP70, HSP90, and HSP27 exert a “molecular chaperone” function that is proposed to assist in the assembly or repair of newly synthesized or damaged proteins. These functions may protect cells from ischemia- or hypoxia-induced injury. We assessed whether these proteins were involved in the protective effect of decreased PKCδ activity on hypoxia-induced injury. The protein levels of HSP70 and HSP90 do not seem to be determinants of whether PKCδ suppression enhances EC survival during hypoxia in our study. It has been reported previously that hypoxia induces new categories of proteins, called “hypoxia-associated proteins,” in ECs.1 These proteins differ from HSPs, which are more commonly induced by hypoxia in other types of cells. Recently, 1 of these proteins was identified as a glycolytic enzyme, glyceraldehyde-3-phosphate dehydrogenase (GAPDH).1 In ECs, these proteins may be more important in the cellular adaptation to hypoxia than are HSPs. The roles of HSPs other than HSP70, HSP90, and HSP27 and that of such hypoxia-associated proteins in hypoxia-induced EC death remain undefined.
In conclusion, hypoxia decreases PKCδ specific activity in ECs. This decrease in PKCδ activity may enhance EC survival during hypoxia by preventing nonapoptotic cell death. Because the activities of PKCα and PKCε do not change in ECs exposed to hypoxia, our investigation raises the possibility that a selective inhibitor of PKCδ may prevent hypoxia-induced EC injury.
This study was supported by grants HL-51043 and HL-47032 from the National Heart, Lung, and Blood Institute.
Reprint requests to J. Anthony Ware, MD, Albert Einstein College of Medicine, Forchheimer Building, G-46, 1300 Morris Park Ave, Bronx, NY 10461.
- Received April 7, 1999.
- Revision received June 9, 1999.
- Accepted June 17, 1999.
- Copyright © 1999 by American Heart Association
Sahai A, Mei C, Pattison TA, Tannen RL. Chronic hypoxia induces proliferation of cultured mesangial cells: role of calcium and protein kinase C. Am J Physiol. 1997;273:F954–F960.
Xu Y, Stenmark KR, Das M, Walchak SJ, Ruff LJ, Dempsey EC. Pulmonary artery smooth muscle cells from chronically hypoxic neonatal calves retain fetal-like and acquire new growth properties. Am J Physiol. 1997;273:L234–L245.
Gray MO, Karliner JS, Mochly-Rosen D. A selective ε-protein kinase C antagonist inhibits protection of cardiac myocytes from hypoxia-induced cell death. J Biol Chem. 1997;272:30945–30951.
Kalra VK, Shen Y, Sultana C, Rattan V. Hypoxia induces PECAM-1 phosphorylation and transendothelial migration of monocytes. Am J Physiol. 1996;271:H2025–H2034.
Kent KC, Mii S, Harrington EO, Chang JD, Mallette S, Ware JA. Requirement for protein kinase C activation in basic fibroblast growth factor-induced human endothelial cell proliferation. Circ Res. 1995;77:231–238.
Tang S, Morgan KG, Parker C, Ware JA. Requirement for protein kinase Cθ for cell cycle progression and formation of actin stress fibers and filopodia in vascular endothelial cells. J Biol Chem. 1997;272:28704–28711.
Shizukuda Y, Tang S, Yokota R, Ware JA. Vascular endothelial growth factor-induced endothelial cell migration and proliferation depend on a nitric oxide-mediated decrease in protein kinase Cδ activity. Circ Res. 1999;85:247–256.
Emoto Y, Manome Y, Meinhardt G, Kisaki H, Kharbanda S, Robertson M, Ghayur T, Wong WW, Kamen R, Weichselbaum R, Kufe D. Proteolytic activation of protein kinase C δ by an ICE-like protease in apoptotic cells. EMBO J. 1995;24:6148–6156.
Webster KA, Discher DJ, Bishopric NH. Induction and nuclear accumulation of fos and jun proto-oncogenes in hypoxic cardiac myocytes. J Biol Chem. 1993;268:16852–16858.
Harrington EO, Loffler J, Nelson PR, Kent KC, Simons M, Ware JA. Enhancement of migration by protein kinase Cα and inhibition of proliferation and cell cycle progression by protein kinase Cδ in capillary endothelial cells. J Biol Chem. 1997;272:7390–7393.
Xie Y-W, Shen W, Zhao G, Xu X, Wolin MS, Hintze TH. Role of endothelium-derived nitric oxide in the modulation of canine myocardial mitochondrial respiration in vitro: implication for the development of heart failure. Circ Res. 1996;79:381–397.
Benjamin IJ, McMillan R. Stress (heat shock) proteins: molecular chaperones in cardiovascular biology and disease. Circ Res. 1998;83:117–132.