Protein kinase C-beta and oxygen deprivation. A novel Egr-1-dependent pathway for fibrin deposition in hypoxemic vasculature.

Fibrin deposition is a salient feature of hypoxemic vasculature and results from induction of tissue factor. Such tissue factor expression in an oxygen deficient environment is driven by the transcription factor Early Growth Response (Egr)-1. Using homozygous null mice for the protein kinase C beta-isoform gene (PKCbeta null), PKCbeta is shown to be upstream of Egr-1 in this oxygen deprivation-mediated pathway for triggering procoagulant events. Whereas wild-type mice exposed to hypoxia (6%) displayed a robust increase in tissue factor transcripts and antigen, and vascular fibrin deposition, PKCbeta null animals showed a markedly blunted response. Consistent with a central role for Egr-1 in hypoxia-induced expression of tissue factor, PKCbeta null mice subjected to oxygen deprivation displayed at most a minor elevation in Egr-1 transcripts, antigen, and intensity of the gel shift band by electrophoretic mobility shift assay, compared with normoxic animals. These data firmly establish PKCbeta as a trigger for events leading to induction of Egr-1 and tissue factor under hypoxic conditions, and provide insight into a biologic cascade whereby oxygen deprivation recruits targets of PKCbeta and Egr-1, thereby amplifying the cellular response.

Oxygen deprivation is a frequently encountered physiologic stress accompanying disorders of the lung and cardiovascular system, as well as consequent to high altitude exposure. The cellular response to hypoxia is complex and involves a range of mechanisms, some occurring within minutes of oxygen deprivation, whereas others reset multistep biosynthetic and physiologic programs (1)(2)(3)(4)(5)(6)(7)(8)(9). For example, hypoxia-induced translocation of P-selectin to the endothelial cell surface, thereby promoting adherence of leukocytes at sites of ischemic stress, occurs within minutes and is due to a transient rise in cytosolic calcium (4). The best characterized mechanism triggering biosynthetic adaptation to oxygen deprivation involves the transcriptional regulator hypoxia-inducible factor-1 (HIF-1) 1 (2,3,10). Activation of HIF-1 at low ambient oxygen tension results in expression of an array of genes, which, in a highly integrated manner, redirects metabolic and other cellular mechanisms enhancing cell survival in the hypoxic environment (2,3,7,10). HIF-1 increases expression of the noninsulin-dependent glucose transporter (GLUT1) (11) and multiple glycolytic enzymes (12), thereby promoting glucose uptake and glycolysis, as the efficacy of aerobic respiration is diminished due to limited oxygen availability. In a highly complementary manner, increased levels of erythropoietin and vascular endothelial growth factor expand oxygen carrying capacity of the blood and, over time, enhance ingrowth of vessels to sites of hypoxemia, respectively (2,3,13). The positive impact of these changes for cellular and organ homeostasis is evident from the lethal phenotype of homozygous null mice for either vascular endothelial growth factor or HIF-1␣ (14 -17).
Another facet of the biosynthetic response to hypoxia concerns expression of oxygen-regulated proteins (ORPs), which occurs within the first 48 h of hypoxia (8). The ORPs comprise a group of polypeptides whose members overlap with glucoseregulated proteins (GRPs), such as GRP78 and GRP94 (18), well known for their role as chaperones in the endoplasmic reticulum (ER). Studies of ORP150, a recently described member of this group cloned from hypoxic astrocytes (19), also places it in the ER; enhanced expression of ORP150 in response to severe oxygen deprivation promotes cell survival (20). These data indicate that ER stress, likely due to accumulation of incompletely folded/misfolded proteins in the ER, as energy depletion and changes in protein biosynthesis impair normal protein processing, is an important feature of the intracellular milieu in hypoxia.
Our recent studies have highlighted a quite distinct pathway triggered by oxygen deprivation. Hypoxia causes transcriptional activation of tissue factor, the key procoagulant cofactor that initiates the coagulation mechanism resulting, ultimately, in vascular fibrin formation (6,21,22). Increased tissue factor in hypoxemic vasculature is especially evident in mononuclear phagocytes (MPs) (6,21). Further analysis of this pathway showed that induction of tissue factor expression resulted from up-regulation of the transcription factor Early Growth Response (Egr)-1 (22). For example, homozygous Egr-1 null mice subjected to hypoxia did not display either enhanced tissue factor expression or vascular fibrin deposition. As fibrin accumulation in vessels subject to hypoxemia could have far-reach-ing pathophysiological consequences (23,24) and increased levels of Egr-1 could recruit multiple other target genes, we have performed further studies to elucidate mechanisms underlying transcriptional activation of Egr-1 in response to oxygen deprivation. Our first studies using cultured cell lines suggested that events leading to Egr-1 expression could be traced back to protein kinase C isoform ␤II (PKC␤II) (22). Because of the complex array of PKC isoforms and their overlapping properties, as well as the likelihood that results in established cell lines subject to transient transfection with multiple expression constructs would not faithfully model all aspects of the in vivo milieu, we have turned to experiments in genetically manipulated mice. In this regard, mice deficient in the PKC␤ gene (since PKC␤I and PKC␤II isoforms are encoded by the same gene, these mice are deficient in both) have been produced by homologous recombination (termed PKC␤ null mice) (25). The phenotype of these animals includes immunologic abnormalities, impaired humoral immune responses and reduced cellular responses to B cells, similar to those observed in X-linked immunodeficiency, although survival, reproductive status, and other general properties of these mice appear to be intact. Using PKC␤ null animals, we have established a central role for PKC␤ in the pathway triggered by hypoxia leading to activation of Egr-1 and tissue factor expression, and resulting in fibrin deposition in lung vasculature. These studies delineate a new facet of the response to oxygen deprivation, distinct from HIF-1, which is likely to initiate a complex pattern of biosynthetic changes in hypoxic/hypoxemic environments.

Induction of Hypoxia in Vivo and in
Vitro-Experiments employing mice subjected to hypoxia were performed according to protocols approved by the Institutional Animal Care and Use Committee at Columbia University, in accordance with the Association for the Accreditation of Laboratory Animal Care guidelines. Homozygous PKC␤ null mice were prepared as described (25). Both PKC␤ null and wild-type controls (12-15 weeks old) were in a similar mixed background (129xC57BL6). However, since the 129 strain actually comprises a collection of genetically heterogeneous substrains, it is more than likely that the control mice (which were most likely derived from a different strain wild-type embryonic stem cell than the PKC␤ null mice) are not genetically identical to the knockout mice. Mice were subjected to normobaric hypoxia (n ϭ 5 per experimental condition unless indicated otherwise) for the indicated times by the regulated addition of nitrogen to a chamber equipped with circulating fans, carbon dioxide and ammonia elimination systems, and an on-line oxygen sensor (Horiba Ltd., Kyoto, Japan) (6). The environment within the chamber (including temperature and humidity) was regulated by a custom-built interface (K ϩ K Interface Inc., New York, NY), which used a computer-driven environmental control program. Mice placed in the chamber in their usual cages were allowed free access to food and water, and the system parameters were adjusted to a final oxygen concentration of 5.5-6.5%. Mice exposed to hypoxia were tachypneic and had reduced activity, compared with normoxic counterparts, but there was no mortality during the experimental period. At the indicated times, animals were sacrificed, and tissues were studied as described below.
Cells were subjected to hypoxia using an environmental chamber (Coy Laboratory Products, Ann Arbor, MI), which maintained a controlled temperature (37°C), an humidified atmosphere with carbon dioxide (5%) and the balance made up of nitrogen. Use of this chamber for cultured cells has been described previously (6). The pH of the medium remained unchanged throughout experiments; pO 2 in the medium was Ϸ12-14 torr (oxygen leached continuously from the tissue culture plasticware during the course of the experiments). The glucose concentration fell by less than 10 -20% during the longest experimental time points (4 h; this degree of glucose depletion does not induce changes in gene expression described below). Exposure of cultured MPs to hypoxia (see below) did not alter cell viability, as assessed by: (i) lack of increased release of intracellular markers, such as lactate dehydrogenase; (ii) continued exclusion of trypan blue; (iii) continued protein synthesis (see below); and (iv) continued cell adherence to the substrate. Cells subjected to hypoxia were exposed to medium pre-equilibrated with the hypoxic gas mixture just prior to placement in the environ-mental chamber. Thus, cultures were immediately immersed in the oxygen-deprived environment at the time of medium change/placement in the chamber.
Analysis of Egr-1, ERK1/2, Tissue Factor, Fibrin, and GLUT1 in Genetically Manipulated Mice-Following hypoxia, mice were sacrificed and tissue was processed immediately. For Northern analysis, tissue was cut into small pieces, immersed in Trizol (Life Technologies, Inc.), and homogenized, and total RNA was extracted and subjected to electrophoresis (0.8% agarose). RNA was transferred to Duralon-UV membranes (Stratagene), and membranes were then hybridized with 32 P-labeled cDNA probe for mouse Egr-1 (26), tissue factor (6), or GLUT1 (27). Blots were also hybridized with 32 P-labeled ␤-actin as an internal control for RNA loading.
For Western blotting of Egr-1, nuclear extracts were prepared (see below), and subjected to SDS-PAGE (7.5%; nonreduced). Proteins in the gel were transferred electrophoretically to nitrocellulose membranes, and immunoblotting was performed with rabbit anti-Egr-1 IgG (Santa Cruz Biotechnology) according to the Blotto procedure (28). Sites of primary antibody binding were visualized with horseradish peroxidase-conjugated goat anti-rabbit IgG (Amersham International, Buckinghamshire, United Kingdom). Final detection of immunoreactive bands was performed using the enhanced chemiluminescent Western blotting system (Amersham). A similar strategy was utilized to detect ERK1/2, except that an extract of total cell protein was employed, and anti-MAPK IgG (reactive with ERK1/2 regardless of phosphorylation status; Zymed Laboratories Inc.) or anti-active MAPK IgG (1:5000 dilution; the latter selective for phosphorylated ERK1/2; Promega) was used as the primary antibody (29,30).
For immunoblotting of tissue samples for fibrin, lung was harvested from animals treated with heparin (10 units/g, resulting in an activated partial thromboplastin time Ͼ300 s) prior to sacrifice (21). Tissue was placed in buffer (0.05 M Tris, 0.15 M NaCl, 500 units/ml heparin; final pH 7.6) on ice and homogenized. Plasmin digestion was performed by a modification of the method of Francis (31) by adding human plasmin (0.32 units/ml, Sigma) and incubating the sample for 6 h at 37°C. This was followed by addition of more plasmin (0.32 units/ml) and additional incubation for 2 h, followed by centrifugation (2,300 ϫ g for 15 min), and aspiration of the supernatant. Samples of plasmin-treated hypoxic and normoxic lung homogenates (200 g of total protein) were boiled in reducing SDS-sample buffer for 5 min and subjected to SDS-PAGE (10%; reduced). Samples were electrophoretically transferred to nitrocellulose, and blots were reacted with rabbit anti-fibrin antibody made to ␥-␥ chain cross-links (21,32), followed by affinity-purified peroxidase-conjugated anti-rabbit IgG (Sigma). Final detection of the bands was as above.
The electrophoretic mobility gel shift assay was performed on nuclear extracts prepared immediately after harvest of the hypoxic lung by the method of Dignam et al. (33). Doublestranded oligonucleotide probes for Egr (Santa Cruz Biotechnology) were 5Ј end-labeled with [ 32 P]ATP (3,000 Ci/mmol) using T4 polynucleotide kinase and standard procedures. Binding reactions were performed as described (34), and samples (5 g of protein in each lane) were loaded directly onto nondenaturing polyacrylamide/bisacrylamide (6%) gels prepared in 0.5ϫ TBE (45 mM Tris, 45 mM boric acid, 1 mM EDTA). Gels were pre-run for 20 min before samples were loaded, and electrophoresis was performed at room temperature for 1.5-2 h at 200 V. For competition studies, an 100-fold molar excess of unlabeled probes for either Egr (Santa Cruz Biotechnology) or Sp1 (Promega) was added.
Studies with Cultured MPs-Resident peritoneal MPs were harvested from PKC␤ null and wild-type mice by lavage with ice-cold medium (35). Cells were maintained in culture in complete medium (RPMI 1640 with fetal calf serum (10%)) for 24 h and then serum-starved for the next 24 h prior to placement in the hypoxic environment. Then, cultured MPs were subjected to hypoxia for 4 h or were exposed to lipopolysaccharide (LPS; 1 g/ml; Escherichia coli serotype 026B6; Sigma) for 6 h or to phorbol 12-myristate 13-acetate (50 ng/ml; Sigma) for 1 h. RT-PCR for tissue factor transcripts was performed using the following primers: 5Ј -TTGATGTGGAAGAAGGAGTA-3Ј (sense) and 5Ј-TGGAATCAAAGCGTTAGC-3Ј (antisense). The size of the amplified DNA fragment was 416 base pairs, and it was also identified by Southern blotting with a 32 P-labeled fragment from the murine cDNA for tissue factor (322 base pairs), which did not contain the region included in the tissue factor primers. Other experiments were performed with a line of rat alveolar macrophages (NR8383) obtained from ATCC (Manassas, VA) and grown in Ham's F12K containing 15% heat-inactivated fetal calf serum. NR8383 cells were incubated with the specific PKC␤ inhibitor LY379196 (36,37) alone, or in the presence of LPS or an hypoxic environment. Total RNA was isolated and either subjected to Northern analysis using 32 Plabeled cDNA probes for Egr-1, GLUT1, or ␤-actin. Where indicated, autoradiograms were scanned with a Hewlett-Packard ScanJet IICX linked to a Macintosh G3 computer, and images were quantitated and compared using Molecular Analyst software (Bio-Rad).

Expression of Egr-1 and Activation of ERK1/2 in Hypoxic
Lung: Effect of PKC␤ Deletion-In a previous study, hypoxia was shown to induce Egr-1 expression and Egr-1-dependent transcription of tissue factor (22). In order to assess whether PKC␤ might participate in upstream events leading to Egr-1 up-regulation, PKC␤ null mice were subjected to oxygen deprivation and Egr-1 expression was compared with that observed in age-and strain-matched controls. Wild-type mice demonstrated strong up-regulation of transcripts for Egr-1 mRNA in the lung within 30 min of exposure to the environment with 6% oxygen (Fig. 1A, lanes 1 and 2; Ϸ16-fold increase in intensity of the band in samples from hypoxic lung compared with normoxic controls). In contrast, PKC␤ null mice subjected to hypoxia for the same time showed only a weak increase in Egr-1 transcripts (Fig. 1A, lanes 3 and 4; Ϸ1.6-fold increase in hypoxia compared with normoxia). Immunoblotting of nuclear extracts from hypoxic lung confirmed that increased amounts of Egr-1 antigen were present in samples from wild-type mice after 30 min, compared with normoxic controls (Fig. 1B, lanes  1 and 2). In fact, the rapidity of the rise in Egr-1 antigen FIG. 1. Hypoxia-mediated induction of Egr-1 and Egr-1 DNA binding activity in hypoxic murine lung: effect of PKC␤ deletion. A, wild-type (ϩ/ϩ) and PKC␤ null (Ϫ/Ϫ) mice were exposed to hypoxia (H; 6% oxygen) or normoxia (N) for 30 min, sacrificed, and total RNA was isolated from the lung, and subjected to Northern analysis (10 g/lane) with 32 P-labeled cDNA probes for Egr-1 and ␤-actin. B, immunoblotting with anti-Egr-1 IgG was performed on nuclear extracts from lungs of mice exposed to hypoxia (H) or normoxia (N) for 30 min. Each lane was loaded with 10 g of total protein. C-H, immunostaining for Egr-1 in lungs of normoxic (C) or hypoxic (D) wild-type mice. Adjacent sections of lung from wild-type animals exposed to hypoxia were stained with antibody to Egr-1 (E) and F4/80 (F). Panels G and H show immunostaining for Egr-1 in lungs from normoxic (G) and hypoxic (H) PKC␤ null mice. In C, D, G, and H, marker bar ϭ 5 m. In E and F, marker bar ϭ 2 m. The arrows in E and F point to the same areas/cells in adjacent sections. I, electrophoretic mobility gel shift assay with 32 P-labeled consensus Egr oligonucleotide probe and nuclear extracts from lungs of wild-type or PKC␤ null mice exposed to hypoxia or normoxia (5 g/lane). FP indicates a lane with only free probe.
suggests that even post-translational mechanisms related to protein stabilization might also be involved. Analogous to the lower levels of Egr-1 mRNA in PKC␤ null mice, the knockouts showed an immunoreactive band of low intensity after exposure to hypoxia (Fig. 1B, lanes 3 and 4), versus the robust response in wild-type mice. In hypoxic lung from wild-type mice, Egr-1 antigen was especially evident in smooth muscle cells and in mononuclear phagocytes ( Fig. 1D; panel C shows Egr-1 immunostaining in the normoxic control). Staining of adjacent sections with F4/80 demonstrated colocalization of this murine monocyte marker (Fig. 1F) in cells expressing Egr-1 (Fig. 1E). The PKC␤ null mice showed only low levels of Egr-1 antigen in lungs of normoxic and hypoxic mice (Fig. 1, G  and H). Gel shift analysis of nuclear extracts from lungs of wild-type mice using 32 P-labeled consensus oligonucleotide probe for Egr demonstrated a gel shift band whose intensity increased strongly in response to hypoxia (Fig. 1I, lanes 2 and  3). Nuclear binding activity was sequence-specific, based on competition with excess unlabeled Egr probe, but not unlabeled oligonucleotide probe for Sp1 (data not shown). When PKC␤ null mice were exposed to hypoxia, the intensity of the gel shift band from nuclear extracts of lung increased minimally compared with normoxia (Fig. 1I, lanes 4 and 5). These data demonstrate that deletion of the PKC␤ gene strongly inhibits signaling events in the lung leading to induction of Egr-1. The small residual increase in Egr-1 in PKC␤ null mice was not reproducible (often being similar to normoxic controls) and, thus, was difficult to characterize further.
In a cultured monocyte-like cell line, we had observed previously that hypoxia caused translocation of PKC␤II to the membranous fraction and autophosphorylation, and closely correlated with events leading to Egr-1 expression by a pathway that involved activation of MAP kinases ERK1/2 (22). This led us to examine whether activation of ERK1/2 occurred in hy-poxic lung, and if this would be altered in PKC␤ null mice. Wild-type mice were exposed to hypoxia for 10 min, and lung homogenates were prepared for immunoblotting with antibody specific for phosphorylated ERK1/2 (29,30). A strong increase in intensity of immunoreactive material with mass Ϸ42/44 kDa, corresponding to the migration of phospho-ERK1/2, was observed in hypoxic, versus normoxic, samples ( Fig. 2A, lanes 2  and 1, respectively). Similar experiments performed with lung homogenates from PKC␤ null mice showed only low levels of phosphorylated ERK1/2 antigen in normoxic and hypoxic mice ( Fig. 2A, lanes 3 and 4). The latter result was obtained after 10 min of hypoxia (a time point when strong ERK1/2 activation was observed in wild-type mice); experiments performed at shorter and longer incubation times using PKC␤ null mice did not show increased phospho-ERK1/2, indicating the apparent absence of ERK1/2 activation, rather than just an altered time course (data not shown). Panel B displays the presence of similar levels of total ERK1/2 antigen in lung harvested from wild-type and PKC␤ null mice under normoxic and hypoxic conditions. Immunostaining was performed on lung tissue from normoxic and hypoxic mice with antibody to phosphorylated ERK1/2. Compared with normoxic controls (Fig. 2C), immunoreactive material was found lungs of wild-type hypoxic mice (Fig. 2D) in a distribution corresponding to vascular smooth muscle cells (main panel in D) and mononuclear phagocytes (inset to D). Studies on lung tissue from PKC␤ null animals demonstrated little immunoreactivity in lungs from either normoxic and hypoxic animals (Fig. 2, E and F). Thus, deletion of PKC␤ prevents an hypoxia-mediated signaling pathway in which ERK1/2 activation/phosphorylation is tied to up-regulation of Egr-1.
Expression of Tissue Factor and Induction of Fibrin Formation in Hypoxic Lung: Effect of PKC␤ Deletion-Wild-type mice exposed to hypoxia display activation of Egr-1, which, in turn,

FIG. 2. Hypoxia-mediated activation of ERK1/2 in hypoxic murine lung: effect of PKC␤ deletion.
A and B, wild-type (ϩ/ϩ) and PKC␤ null (Ϫ/Ϫ) mice were exposed to hypoxia (H; 6% oxygen) or normoxia (N) for 10 min, sacrificed, and protein extracts from the lung were prepared and subjected to SDS-PAGE (10%; 50 g of protein/lane)/immunoblotting with anti-active MAPK IgG (A) or anti-MAPK IgG (B; this antibody detects all ERK1/2 regardless of activation state). C-F, immunostaining using antibody to phosphorylated ERK1/2 in lungs from normoxic (C) or hypoxic (D) wild-type mice. Panels E and F show immunostaining for activated ERK1/2 in normoxic (E) and hypoxic (F) PKC␤ null mice. In C-F, marker bar ϭ 5 m. In the inset to panel D, marker bar ϭ 1 m. triggers increased transcription of the tissue factor gene (6). Thus, wild-type mice subjected to oxygen deprivation showed elevated levels of steady-state tissue factor mRNA (Fig. 3A, lane 2; Ϸ20-fold increase comparing lanes 2 and 1), which was followed by enhanced expression of tissue factor antigen in the lung (Fig. 3C), versus normoxic controls (Fig. 3B). Further studies showed the distribution of tissue factor antigen in lungs from oxygen-deprived animals to overlap that observed for hypoxia-induced Egr-1 expression and activated ERK1/2, being found principally in smooth muscle cells and MPs (data not shown). When the same protocol was followed with PKC␤ null mice, animals subjected to hypoxia showed only a slight increase in tissue factor transcripts on Northern blots (Fig. 3A,  lane 4; Ϸ2.2-fold increase comparing lanes 3 and 4) and virtually no elevation of immunoreactive tissue factor in the lung (Fig. 3E), compared with PKC␤ null mice maintained in normoxia (Figs. 3, A, lane 3, and D). As with levels of Egr-1 in hypoxic PKC␤ null mice, the residual increase in tissue factor was variable and at extremely low levels, making it difficult to analyze in detail. The key role of tissue factor in fibrin deposition in hypoxic lung was displayed by immunoblotting plasmindigests of lung extracts using an antibody specific for a neoepitope in fibrin (21). Lung from wild-type mice subjected to hypoxia displayed a strong band immunoreactive with antifibrin antibody, which was only weakly stained in normoxic controls (Fig. 4A, lanes 4/5). Consistent with these results, immunostaining for fibrin in hypoxic lung tissue from wildtype mice displayed deposits of fibrin immunoreactive material ( Fig. 4C), versus their absence in normoxic lung (Fig. 4B). In parallel with low levels of tissue factor in hypoxic PKC␤ null mice, there was no evidence of fibrin epitopes in lung tissue harvested from these animals after exposure to oxygen deprivation by immunoblotting or immunostaining (Fig. 4, panels A (lane 5) and E; note that panels A (lane 4) and D display results in normoxic PKC␤ null mice). Our data are consistent with a cause-effect relationship for induction of Egr-1 and tissue factor in hypoxic lung and the appearance of fibrin deposits, and indicate that PKC␤ isoforms are upstream of these events.

Specificity of Suppressed Gene Expression in PKC␤ Null
Mice: Hypoxia and LPS-Although PKC␤ null mice demonstrated inhibition of Egr-1 and tissue factor gene activation following exposure to hypoxia, this did not reflect a general suppression of gene expression. For example, systemic infusion of lipopolysaccharide into PKC␤ null mice resulted in a strong increase in Egr-1 and tissue factor transcripts in total RNA harvested from lung (Fig. 5, A and B, respectively, compare  lanes 3 and 4). The latter response was comparable to that observed in wild-type mice (Fig. 5, A and B, compare lanes 1  and 2). Also, other pathways regulating the cellular response to hypoxia, such as HIF-1, remained intact. Up-regulation of the noninsulin-dependent glucose transporter (GLUT-1) is a well known feature of the protective response to oxygen deprivation, which depends, in large part, on HIF-1 (11). Hypoxic PKC␤ null mice displayed strong induction of GLUT-1 transcripts in the lung (Fig. 5C, lane 4; lane 3 shows normoxic PKC␤ null control), which was comparable to that observed in wild-type an-FIG. 3. Hypoxia-mediated induction of tissue factor: effect of PKC␤ deletion. A, wild-type (ϩ/ϩ) and PKC␤ null (Ϫ/Ϫ) mice were exposed to hypoxia (H; 6% oxygen) or normoxia (N) for 4 h and sacrificed, and total RNA was isolated from the lung and subjected to Northern analysis (20 g/lane) with 32 P-labeled cDNA probes for murine tissue factor (TF) and ␤-actin. B-E, immunostaining using anti-tissue factor IgG in lungs of normoxic (B) or hypoxic (C) wild-type mice. The arrows in panel C point to macrophages immunoreactive with anti-TF IgG. Panels D and E show immunostaining for tissue factor in lungs from normoxic (D) and hypoxic (E) PKC␤ null mice. Marker bar ϭ 2 m.
imals subjected to the same protocol (Fig. 5C, lane 2; lane 1 shows the normoxic wild-type control). In terms of steady-state levels of GLUT-1 transcripts, there was Ϸ7.6-fold increase in wild-type mice and Ϸ7.3-fold increase in PKC␤ null mice, comparing hypoxia and normoxia (Fig. 5C).
Macrophage Expression of Tissue Factor in Response to Oxygen Deprivation: Effect of PKC␤ Deletion-One critical site in hypoxic lung for activation of gene expression in the hypoxiatriggered Egr-1-tissue factor pathway is the MP (22). It was important to perform additional experiments to be certain that the observed inhibitory effect in PKC␤ null mice reflected events at the level of MPs, rather than complex intercellular compensatory mechanisms possibly operative within the PKC␤ null phenotype (as has been observed with other knockouts). For example, mice in whom the Interleukin 6 gene had been deleted displayed high levels of tumor necrosis factor-␣, the latter contributing to the response of these mice to a range of stimuli (38,39). Resident peritoneal macrophages (also termed MPs) were obtained by lavage from wild-type and PKC␤ null mice. Following exposure of MPs from wild-type mice to hypoxia (pO 2 Ϸ14 torr) for 4 h, induction of tissue factor transcripts was observed by RT-PCR (Fig. 6A, lane 3; lane 2 shows normoxic control). The identity of this amplicon was confirmed by Southern blotting with a 32 P-labeled tissue factor probe (Fig.  6B, lane 3). In contrast, similar experiments with resident peritoneal macrophages from PKC␤ null mice displayed virtually no increase in tissue factor transcripts (Fig. 6, A and B,  lane 7; lane 6 shows the normoxic control). As in our studies using the intact animal, induction of tissue factor in response to other stimuli, such as phorbol ester and lipopolysaccharide, remained intact in the PKC␤ null mice; both of these stimuli caused an increase in transcripts in MPs from wild-type (Fig. 6,  A and B, lanes 4 and 5) and PKC␤ null mice (Fig. 6, A and B,  lanes 8 and 9).
Hypoxia-induced Expression of Egr-1 in Alveolar Macrophages: Effect of PKC␤ Inhibition-The coagulant properties of alveolar macrophages are likely to be relevant to the procoagulant environment of hypoxemic lung. In a previous study with monocyte-like U937 cells, hypoxia was shown to induce mem-brane translocation and autophosphorylation of PKC␤II, compared with lack of such changes in PKC isoforms ␣ and ⑀ (22). Furthermore, transient transfection studies demonstrated that expression of dominant-negative PKC␤II selectively suppressed hypoxia-mediated activation of Egr-1 and tissue factor transcription (22). In order to determine the applicability of these results to macrophages in the lung, we turned to a line of cultured alveolar macrophages (NR8383) and assessed the effect of hypoxia on induction of Egr-1. The role of PKC␤ was studied using LY379196, a selective PKC␤ inhibitor, although it has a similar K i for the ␤1 and ␤2 isoforms (36,37). Exposure of normoxic macrophages to LY379196 was without effect on Egr-1 expression (data not shown). In the absence of inhibitor, hypoxic macrophages showed a strong increase in Egr-1 transcripts in hypoxia (Fig. 7A, lane 2; Ϸ10.6-fold increase), compared with cells maintained in normoxia (Fig. 7A, lane 1). However, in the presence of the inhibitor, there was no appar- FIG. 6. Expression of tissue factor by peritoneal macrophages exposed to LPS, phorbol 12-myristate 13-acetate, or hypoxia: effect of PKC␤. Resident peritoneal macrophages were isolated from age-and strain-matched wild-type control (ϩ/ϩ, lanes 2-5) and PKC␤ null mice (Ϫ/Ϫ, lanes 6 -9). Cells were exposed to normoxia (N) or hypoxia (H) for 4 h (lanes 2 and 3 or lanes 6 and 7). In other experiments, cultures were maintained in normoxia and exposed to phorbol myristate acetate (50 ng/ml; lanes 4 and 8 labeled P) for 1 h or to LPS (1 g/ml; lanes 5 and 9 labeled L) for 6 h. Following each of these treatments, RNA was isolated, and RT-PCR was performed with primers for murine tissue factor (TF) or ␤-actin. Amplicons were visualized by ethidium bromide staining, TF (A) or ␤-actin (C), or were transferred to membranes and hybridized with a 32 P-labeled cDNA for murine tissue factor (B). Lane 1 shows the migration of molecular weight markers (100-base pair DNA ladder; Promega).

FIG. 7.
Hypoxia-mediated expression of Egr-1 and GLUT1: effect of PKC␤ inhibition by LY379196. A, cultured rat alveolar macrophages (NR8383) were preincubated with LY379196 (0.2 M) for 60 min, and were then incubated under normoxic (N) or hypoxic (pO 2 Ϸ12-14 torr; H) conditions for 30 min. Then, total RNA was isolated and Northern analysis was performed (10 g of RNA/lane) using 32 Plabeled cDNA for Egr-1 or ␤-actin. Note that controls in which LY379196 was incubated with normoxic NR8383 cells showed no differences in Egr-1 transcripts. B, cultured alveolar macrophages were treated as in A (above) except that the incubation time in hypoxia was 4 h, and Northern blotting was performed with 32 P-labeled cDNA for GLUT1 and ␤-actin. In other experiments (C), mice were exposed to hypoxia (H; 6% oxygen) or normoxia (N) for 4 h, sacrificed, and total RNA was isolated from the lung, and subjected to Northern analysis (15 g/lane) with 32 P-labeled cDNA probes for GLUT1 and ␤-actin. ent increase in Egr-1 transcripts in the hypoxic cells (Fig. 7A,  lane 3). The apparent specificity of the inhibitor for the hypoxia-induced pathway under study was consistent with the observation that when the alveolar macrophage cell line was exposed to hypoxia, induction of GLUT-1 mRNA in hypoxic macrophages was maintained in the presence of LY379196 (Fig. 7B). DISCUSSION Isoforms of PKC have been linked to many physiologic and pathophysiologic processes (40 -44). Recent studies have particularly focussed attention on the ␤-isoforms because of their potential role in cardiovascular dysfunction, especially as relates to endothelial cells and cardiac myocytes (45)(46)(47)(48)(49)(50)(51)(52)(53). Most studies analyzing PKC␤ isoforms have relied on inhibitors that show apparent specificity for ␤I/␤II versus other members of the PKC family (44, 45, 54 -59). Alternative experimental approaches have employed overexpression of wild-type PKC␤II or inducible expression of a constitutively active variant in cardiac myocytes (46,47). Another means to dissect the contribution of PKC␤ is use of PKC␤ null mice. The advantage of this approach is that selective ablation of the PKC␤ gene would be not be predicted to interfere with other PKC family members or to impact on downstream targets of PKC␤. Of course, the possibility that compensatory mechanisms are recruited in the absence of PKC␤ must be kept in mind, as in any experiment with knockout mice. Previously described properties of PKC␤ null mice made them ideal for our studies, as their growth, reproductive capacity, and survival are intact (25). Furthermore, although their phenotype remains to be fully defined, based on their characteristics reported to date (involving mainly immunologic defects which would not be predicted to affect phenomena in our study a priori), it appears that multiple physiologic responses, such as induction of GLUT1 consequent to hypoxia and expression of tissue factor following LPS, remain intact.
Studies reported herein using PKC␤ null mice delineate a central role for PKC␤ in hypoxia-mediated induction of Egr-1 transcription, which leads to increased expression of tissue factor. Our previous studies with cultured monocyte-like cell lines demonstrated that oxygen deprivation triggered PKC␤II activation within minutes, and was followed by a pathway including Raf, mitogen-activated protein kinase/extracellular signal-regulated kinase kinase, and ERK1/2 (22). Subsequent ERK1/2-induced activation of elk-1, resulting in formation of a complex with serum response factor, promotes transcription of Egr-1, and leads directly to tissue factor transcription. The current results in PKC␤ null mice confirm the broad outlines of this pathway in vivo, and indicate a cause-effect relationship between PKC␤ activation in hypoxemic lung and subsequent events leading to tissue factor induction: 1) enhanced expression of Egr-1 and tissue factor observed in response to oxygen deprivation in wild-type mice was not observed to any appreciable extent in PKC␤ null animals; 2) hypoxia-mediated activation of ERK1/2, key contributors to the pathway linking PKC␤ activation to up-regulation of Egr-1 transcription, was also not seen in PKC␤ null mice; and 3) levels of tissue factor antigen remained low and pulmonary vascular fibrin deposition was not observed in hypoxic PKC␤ null mice. These data provide strong support for a pathway in which activation of PKC␤ triggers events leading to accumulation of fibrin in pulmonary vasculature. This response occurs rapidly and could potentially have long term effects on vascular properties by several mechanisms; deposited fibrin could block blood flow and trigger vascular remodeling directly, thrombin formed within the vessel could activate thrombin receptors on cells within the vessel wall (60), and other target genes of Egr-1 and PKC␤ could be expressed.
The biologic significance of the PKC␤-Egr-1-tissue factor pathway induced by hypoxia will require considerably more study to delineate. However, it is clear that these hypoxiamediated events can be differentiated from HIF-1-induced responses. Whereas hypoxia-induced up-regulation of Egr-1 and tissue factor was strongly suppressed in PKC␤ null mice, oxygen deprivation led to comparable induction of GLUT-1 (which is largely due to HIF-1) (11) in PKC␤ null and wild-type mice. This is consistent with our previous observation that a cultured hepatoma line unable to form active HIF-1, due to deficient ARNT/HIF-1␤ function, demonstrated comparable induction of Egr-1 transcripts, versus that in wild-type controls, when exposed to hypoxia (22). Furthermore, expression of tissue factor in response to other stimuli, such as phorbol ester and LPS, also remained intact in PKC␤ null mice. In this context, the adaptive advantage of a pathway causing induction of local vascular fibrin formation remains uncertain; though it could allow sequestration of hypoxemic areas from nonischemic tissues, the possibility that subsequent obstructive thrombus formation might prevent blood flow leading to necrosis also exists. Future studies examining the range of responses triggered by PKC␤ and Egr-1 activation in hypoxemic tissues will be required to appreciate the impact of these events on the vascular adaptation to hypoxemia. Another important issue concerns the specificity of this pathway for different cells and in different tissues. Our studies in hypoxemic lung have shown strong up-regulation of Egr-1 and tissue factor at the antigen level in smooth muscle cells and MPs. However, in view of the presence of Egr-1 in virtually every cell, and the similar potential of a wide range of cells to express tissue factor, it will be essential to know why certain cells are especially susceptible to hypoxiainduced responses in this pathway. Nonetheless, the contribution of the current work is to firmly establish that PKC␤ has an integral role in an hypoxia-induced pathway leading to activation of MAP kinases, and transcription of Egr-1 and tissue factor. These observations form a strong foundation for future studies to further analyze in detail mechanisms underlying this pathway in vitro and physiologic consequences for such effector mechanisms in vivo.