Tissue Factor Expression in Vital Organs during Murine Traumatic Shock : Role of Transcription Factors AP-1 and NF-κB

Male Sprague-Dawley rats weighing 175–225 g were anesthetized with sodium pentobarbital (60 mg/kg, intraperitoneally) before any experimental procedures. The studies were carried out in accordance with the National Institute of Health's Guidelines on the Use of Laboratory Animals and were approved by the Thomas Jefferson University Institutional Animal Care and Use Committee. Traumatic shock was induced in pentobarbital-anesthetized rats by whole-body trauma administered in a Noble-Collip drum apparatus. Anesthetized rats were subjected to a total of 550 revolutions at a rate of 60 rpm. A carotid artery was cannulated for the determination of arterial blood pressure to confirm the occurrence of traumatic shock, according to established procedures. 22All trauma rats underwent autopsy to confirm the presence of gross evidence of traumatic injury (i.e. , bowel ischemia, serosanguinous ascites, splanchnic vascular engorgement) to the splanchnic viscera. Rats were excluded from the study if these findings were not observed to a significant degree. Less than 10% of rats (i.e. , 2 of 24) did not meet these criteria. Noble-Collip drum trauma results in a mortality rate of 90% within 5 h. 22After 30, 60, and 120 min, the lungs, small intestine, and liver were removed for immediate total RNA extraction. Animals were randomly divided into experimental groups consisting of rats subjected either to trauma or to a sham-trauma protocol. Sham-trauma rats were anesthetized, cannulated, and monitored at the same time as traumatized rats.

Total RNA was extracted from the lungs, small intestine, and liver using the acid guanidium-phenol-chloroform extraction method described by Chomczynski and Sacchi. 27The lungs, small intestine, and liver (200–400 mg pieces) were minced on ice and homogenized in a Polytron tissue homogenizer with 4 ml guanidine isothiocyanate solution (Ultra Pure; Gibco/BRL, Gaithersburg, MD). RNA was precipitated with equal volumes of 2-propanol. RNA pellets were washed with 75% ethanol and dissolved in nuclease-free water.

We prefer nuclease protection over methods such as Northern blot analysis for a number of reasons. 28The ribonuclease protection assay is superior for the detection of low-abundance RNAs contained in total RNA, as described previously. The sensitivity of the ribonuclease protection assay comes from the use of a complementary, in vitro  radiolabeled transcript probe specific for the target mRNA. The probe and total RNA are hybridized in solution, after which the mixture is treated with ribonuclease to degrade all remaining single-stranded (unhybridized) RNA. The hybridized portion of the probe will be protected from digestion and can be visualized after electrophoresis by autoradiography of the mixture in a gel. The plasmid for synthesis of the rat TF probe for ribonuclease protection assays was created by synthesis of a complementary DNA (cDNA) using rat lung total RNA and oligodeoxythymidine followed by cloning. Rat TF cDNA was amplified using forward primer (5′-CTACACTGTTCAGATAAGCGATAG-3′) and reverse primer (5′-GTCCACCCAAATCCCACTTTCCTG-3′). The 1,077-bp polymerase chain reaction (PCR) fragment of rat TF, containing nucleotides 243–1319 (GenBank accession No. U07619; Bethesda, MD) was cloned using a polymerase chain reaction 2.1-TOPO Cloning Kit (gift from Invitrogen Corp., Carlsbad, CA). For ribonuclease protection assays an Eco  RI–Hind  III fragment of rat TF was subcloned in the pBluescript II KS⁺vector (Stratagene, La Jolla, CA). The plasmid, containing fragment of TF cDNA, was digested with Xba I and used as a template for in vitro  transcription of a 559-base radiolabeled antisense probe that contained a 487-base protected fragment using T3 RNA polymerase (Boehringer Mannheim, Indianapolis, IN) in the presence of ³²P-UTP (Amersham Corp., Arlington Heights, IL). All constructs, used in this investigation, were verified by sequencing the insert in the plasmid. The rat TF sequence was found to be 100% identical to the published sequences (GenBank accession No. U07619).

For ribonuclease protection assay, 2 μg total RNA was used for detection of β-actin mRNA, 10 μg was used for detection of TF mRNA in the lung, and 50 μg was used for detection of TF mRNA in the liver and small intestine according to previously described procedures. 29Expression of mRNA was quantified using storage phosphor technology (Molecular Dynamics, Sunnyvale, CA). Intensity of each TF mRNA band was normalized for β-actin mRNA level (β-actin plasmid obtained from Ambion, Austin, TX).

Tissue factor activity was measured in rat plasma and lung cytoplasm (postmitochondrial supernatant) using a TF activity assay kit (America Diagnostica, Greenwich, CT). This assay is designed for detection of human TF that activates a chromogenic enzyme cleaved from de novo  factor VIIa/TF (personal oral communication with Dr. R. Hart, America Diagnostica, August 1999, and package insert). This assay is specific for TF; however, rat TF is capable of activating human plasma. 30,31 

Nuclear extracts were prepared from lungs of sham-operated control or trauma rats as described by Andrews and Faller. 32Binding reactions for electrophoretic mobility shift assays (EMSA) of the transcription factors AP-1, SP-1, and NF-κB were carried out according to previously described procedures. 21A 30-bp fragment (−226 to −197) of 5′-region of the rat TF gene was used to analyze AP-1 binding. 31,33This fragment of the TF gene (5′-TGGTGTTGAATCACGGTTGAGTCACCCCTT-3′) contains two (distal and proximal) AP-1 sites [AP-1 (D+P)]. It was created by annealing synthetic oligonucleotides: forward (5′-TGGTGTTGAATCACGGTTGAG-3′) and reverse (3′-GCCAACTCAGTGGGGAA-5′). Three mutant forms of this fragment were used also for analysis of AP-1 binding: mutant 1, containing native proximal and mutated (TGA CGAC ) distal AP-1 site [AP-1 (P)]; mutant 2, containing native distal and mutated (TGAG GGC ) proximal site [AP-1 (D)]; and mutant 3, containing mutated AP-1 sites (both). The mutated nucleotides are shown in bold. The NF-κB 22-bp probe (5′-GGGTCTCGGAGTTTCCTACGGG-3′; from −191 to −170 of mouse TF gene) was created by annealing forward (5′-GGGTCTCGGAGTT-3′) and reverse (3′-GAGCCTCAAAGGATGCCC-5′) oligonucleotides. For creation of a mutant form of NF-κB, the same forward oligonucleotide and a mutated reverse oligonucleotide (3′-GAGCCTCAA TCT  ATGCCC-5′) were used. The SP-1 probe (5′-GGGGCCC ATT  GGGGCGGGGAGTCGCGA-3′; from −107 to −81 of rat TF gene) was created by annealing of forward (5′-GGGGCCC ATT  GGGGCGG-3′) and reverse (3′-CCCCCGCCCCTCAGCGCT-5′) oligonucleotides. This probe, with mutations (C ATT  GGGGCG) in the Egr-1 site was used for analysis of SP-1 binding. A fragment containing mutations in Egr-1 and SP-1 sites was used as a control for nonspecific binding. The radiolabel was introduced at both termini with α[³²P]-deoxycytidine triphosphate in the presence of three other cold deoxynucleotide triphosphates. To show the specificity for binding to AP-1, NF-κB, and SP-1 elements, we used mutant probes and antibodies against cRel (C)-G, and NF-κB p65 (C-20) X (both antibodies were obtained from Santa Cruz Biotechnology, Santa Cruz, CA). Nuclear extracts from lungs of control and trauma rats reacted with SP-1 oligonucleotide as further control.

All values are represented as the mean ± SEM. Data comparing all organ TF gene expression percent changes were subjected to analysis of variance followed by the Dunnett method for evaluation of differences between groups. Comparison of individual organ trauma to control TF mRNA was subjected to the Student t  test for paired observations. Values of P < 0.05 were considered statistically significant.

Traumatized rats exhibited mean arterial blood pressures of 56.3 ± 2.8 mmHg 2 h after trauma, whereas the mean arterial blood pressures in sham-trauma rats were 148.8 ± 9.0 mmHg at this time. To verify that the anesthetized rats experienced traumatic shock, intestinal myeloperoxidase (an enzymatic marker for infiltrated leukocytes) and plasma-free amino nitrogen (a marker for plasma proteolysis) activities were determined 30, 60, and 120 min after trauma. All rats subjected to Noble-Collip drum trauma developed signs of shock. Gross evidence of traumatic injury (i.e. , bowel ischemia, serosanguinous ascites, splanchnic vascular engorgement) to the splanchnic viscera was found in all rats used in this investigation of trauma rats.

Tissue factor concentration and its TF mRNA expression at the transcriptional level were investigated in trauma and sham-operated control rats. TF activity was measured in plasma and lung extracts 30, 60, and 120 min after trauma. Plasma TF activity was very low in control rats and increased up to 14-fold at 30 and 60 min and 2 h (P < 0.001) after trauma (fig. 1A). In the lung, TF activity did not change at 30 min but increased at 1 and 2 h after trauma compared with sham-operated control rats (fig. 1B).

Tissue factor mRNA expression was also assessed in trauma and sham-operated control rats. There was a marked increase in TF mRNA signal in the lungs and liver 2 h after Noble-Collip drum trauma, compared with the same organs from sham-operated control rats (fig. 2). Control experiments measuring the levels of β-actin transcripts indicate similar mRNA levels in these organs of sham-operated control animals and trauma rats (fig. 2). Quantification of the effect of trauma on TF mRNA expression in the lungs, liver, and small intestine 30, 60, and 120 min after Noble-Collip drum trauma is shown in figures 3 A–C. These data show that TF mRNA slightly increased in the lungs (+34%;P < 0.01), small intestine (+40%;P < 0.01), and liver (+24%;P < 0.05) 1 h after trauma and significantly increased in the lungs and liver (+134 and +139%, P < 0.001, respectively) 2 h after Noble-Collip drum trauma, compared with sham-operated anesthetized rats. Much smaller increases in TF mRNA expression were observed in the small intestine 2 h after trauma (+47%;P < 0.001, fig. 3C). Although there was an increase in mRNA in the small intestine, there were no significant changes in TF gene expression in the lung and liver 30 min after trauma, compared with control sham-operated animals (figs. 3A and B).

To investigate the possible molecular mechanisms of the regulation of TF gene expression in the lung during traumatic shock, we performed electrophoretic mobility shift assays using lung nuclear extracts and probes specific for the regulatory region of the rat TF gene. Because the 5′-region of rat TF gene contains two, closely spaced AP-1 elements, an NF-κB (cRel/p65)–binding motif and an SP-1 sequence, nuclear proteins were incubated with AP-1 (D+P), AP-1 (D), AP-1 (P), NF-κB, and SP-1 probes. As shown in figure 4, lung nuclear proteins were bound significantly with the SP-1 probe, but this binding did not change with trauma. In the presence of a 100-fold excess of the unlabeled SP-1 probe, binding of the labeled SP-1 probe was inhibited by competition (data not shown). Sequence-specific binding with a 30-bp AP-1 (D+P) probe containing distal (D) and proximal (P) AP-1 sites was strongly increased in lung nuclear extract from rats subjected to Noble-Collip drum trauma. Binding to this probe was inhibited by competition in the presence of a 100-fold excess of the unlabeled probe. Reasoning that distal or proximal AP-1 sequence elements may be responsible for induction of binding, we used two mutant AP-1 probes for gel-shift analysis. One probe, AP-1 (P), containing mutated distal and native proximal AP-1 sequences, was not bound by nuclear proteins of lung extracts from either control or trauma rats (data not shown).

The AP-1 (D) probe, containing mutated proximal and native distal AP-1 binding motif, bound nuclear proteins from the lung in control and trauma rats (fig. 5). These results showed an increase in binding of nuclear proteins with the AP-1 (D) probe in trauma rats compared with sham-operated control rats when equivalent amounts of protein extracts were used. AP-1 binding was completely inhibited by competition of a 100-fold excess of the unlabeled AP-1 probe (fig. 5). Sequence-specific complex binding to AP-1 (D) domains were enhanced (+395%;P <0.001) in the lung 2 h after trauma (fig. 6). A double mutated AP-1 probe containing mutations in both AP-1 binding motifs was not bound to nuclear proteins from the lung of sham-operated control rats and rats subjected to Noble-Collip drum trauma (data not shown).

The TF-specific NF-κB probe was bound to nuclear proteins from rat lungs of sham-operated control and trauma rats (fig. 7). This NF-κB (cRel/p65) probe exhibits a relatively lower level of rat-lung nuclear protein binding than do p50/p52 (P-selectin specific) or p50/p65 probes (data not shown). Binding of lung nuclear proteins from rats subjected to Noble-Collip drum trauma to the NF-κB probe was increased when equivalent amounts of protein extracts were used (binding with SP-1 probe;fig. 7). Quantification of the effect of trauma on TF-specific NF-κB binding of lung nuclear proteins from control and trauma rats 2 h after Noble-Collip drum trauma showed an increase in binding in trauma compared with sham-operated control (+168%;P <0.001;fig. 6). Specificity of NF-κB regulatory protein binding to the TF-specific NF-κB motif was studied using a probe containing a mutated NF-κB regulatory element and by competitive binding. The labeled mutant NF-κB probe was not bound by lung nuclear proteins from control or trauma rats (fig. 7). The unlabeled NF-κB probe in 100-fold excess was completely inhibited from binding to lung nuclear proteins from control and trauma rats. Binding of lung nuclear proteins from control and trauma rats with a labeled NF-κB probe was inhibited by an antibody against cRel (data not shown).

Traumatic shock is a multifactorial disorder that generates various molecular and cellular mediators. 34Among these are biologically active lipids, cytokines, endotoxins, and oxygen-derived free radicals. This results in a generalized inflammatory event that is similar to the “systemic inflammatory response syndrome.”35The primary insult, however, is not caused by sepsis. 22,35Recently, attention has been focused on leukocyte endothelial interaction and the factors that regulate leukocyte adhesion. 22Significant nonbacterial inflammatory events occur immediately after Noble-Collip drum trauma that directly affect leukocyte–endothelium cell interaction and lead to endothelial dysfunction. In our traumatic shock model, the time course of endothelial dysfunction occurred within 15–30 min and exhibited reduced endothelium-derived nitric oxide, translocation of P-selectin from the endothelium to the cell surface, and increased leukocyte rolling and adherence. 21,22One humoral substance that may contribute to increased leukocyte–endothelium interaction is TF. The rapid translocation of P-selectin protein from the Weibel-Palade bodies of vascular endothelial cells and from the α granules of activated platelets may help to explain the increase in TF we observed within the circulation and in various tissues.

Soft tissue trauma, such as that occurring in Noble-Collip drum trauma, results in significantly increased plasma and lung TF activity and up-regulation of TF mRNA expression at the transcriptional level in the lung, the liver, and, to a lesser extent, the small intestine. Plasma TF activity was very low in sham-operated control rats but increased dramatically in trauma rats. Baseline lung TF activity was higher in control rats relative to plasma and increased after trauma to a lesser extent relative to plasma. Similar results in TF activity were observed in hepatic ischemic–reperfusion injury of the rat. 24 

Recently, we showed that P-selectin mRNA expression is significantly increased in the lung during trauma and that this rapid increase is caused by increased binding of P-selectin–specific NF-κB heterodimers (p50/p52) to the P-selectin promoter. 21A parallel increase in TF and P-selectin mRNA levels in Noble-Collip drum trauma rats suggests that there is a positive correlation between P-selectin and TF gene expression. These results, are consistent with the data of Celi et al. , 13which show that the expression of TF is, at least in part, P-selectin dependent. During tissue injury, P-selectin expression on platelets and stimulated endothelium mediates the binding of leukocytes to the vessel wall and the accumulation of leukocytes within a thrombus. 5Furthermore, blockade of P-selectin by specific anti–P-selectin antibodies, nitric oxide, or ligand competitors, inhibits the accumulation of leukocytes and the deposition of fibrin within a thrombus. 5,36–39Purified P-selectin induced TF expression on mononuclear leukocytes in a dose-dependent manner by 50- to 100-fold. 5,13P-selectin also increased TF mRNA in monocytes and the expression of TF antigen on the monocyte surface. TF expression and procoagulant activity of mononuclear cells were inhibited if monoclonal P-selectin antibody was added to cells before P-selectin exposure. 13The binding of P-selectin to monocytes in the area of vascular injury may initiate thrombosis through induction of TF expression. 13 

In this study, we also analyzed binding of lung nuclear proteins from sham-operated control and trauma rats with AP-1, Egr-1, and NF-κB probes specific for the regulatory region of the rat TF gene using electrophoretic mobility shift assays. Sequence-specific complex binding to AP-1 and NF-κB domains was shown to be significantly enhanced in the lung 2 h after trauma. Because the 5′-region of the rat TF gene contains two, closely spaced AP-1 elements, we incubated the nuclear proteins with different AP-1 probes. Both AP-1 sites include core sequence elements, TGAATCA (distal) and TGAGTCA (proximal), that resemble the consensus binding site of the AP-1 family of transcription factors. Oeth et al.  10demonstrated that the two AP-1 sites bind c-Fos/c-Jun heterodimers in unstimulated and lipopolysaccharide-stimulated human monocytic cells. However, Groupp and Donovan-Peluso 40demonstrated that lipopolysaccharide induces a change in the AP-1 binding from JunD/Fos in control THP-1 monocytes to c-Jun/Fos and JunD/Fos. Furthermore, Bierhaus et al.  16showed that the distal AP-1 site of the human TF gene was inducible by Jun homodimers, but the proximal AP-1 site was inducible by Jun/Fos heterodimers during tumor necrosis factor α–mediated induction in bovine aortic endothelial cells. It is possible that the molecular mechanism of induction of TF expression may be dependent on the specific inducer and cell type. Reasoning that the distal or the proximal AP-1 sequence element may be responsible for induction of binding, we used two AP-1 probes. One, AP-1 (P), containing mutated distal and native proximal AP-1 sequence, was not bound to nuclear proteins from the lung in control and trauma rats. The AP-1 (D) probe, containing mutated proximal and native distal AP-1 binding motif, was bound by nuclear proteins from the lung of sham-operated control rats and rats subjected to Noble-Collip drum trauma. These results clearly showed that the distal AP-1 site is responsible for induction of nuclear protein binding and, possibly, induction of TF gene expression in the lung in trauma. Bierhaus et al.  16reported that the proximal site participates in tumor necrosis factor α–mediated induction of endothelial TF gene expression. At the distal AP-1 site, only minor induction of binding activity was observed during tumor necrosis factor α–mediated induction. However, there are data that maximal lipopolysaccharide induction of the TF promoter in human monocytic cells necessitated both the AP-1 site and the κB site within the lipopolysaccharide response element. 10,41Therefore, it appears that the incipient stimulus is important in determining how the two AP-1 TF binding sites are activated and bound.

Activation of TF gene expression by lipopolysaccharide may also be mediated by IκBα because agents that block the nuclear translocation of c-Rel/p65 heterodimers by preventing IκBα phosphorylation or proteolytic degradation inhibits lipopolysaccharide-induced TF gene transcription in monocytic cells. 41 

The demonstration of increased AP-1 and NF-κB binding offers a potential explanation for the increase in TF mRNA expression in trauma. We found significant increases in the binding of the transcription factors, AP-1 and NF-κB, from lung nuclear extracts of trauma rats consistent with this possibility.

In conclusion, our study provides strong evidence that TF mRNA expression and TF activity are increased in the lung, liver, and small intestine during traumatic shock in rats. This enhanced TF activity and mRNA expression, occurring within 1 to 2 h, was accompanied by induction of P-selectin gene expression at the transcriptional level and may play a significant pathophysiologic role in the early thrombosis of severe trauma. We also demonstrated by electrophoretic mobility shift assays using lung nuclear extracts that the AP-1 and NF-κB binding to the TF specific promoter is induced after trauma and may be important for TF gene expression in trauma conditions. Further investigations of the mechanisms by which trauma affects TF expression and the biologic significance of trauma-induced alteration in TF expression are needed. Specifically, the relation among TF, leukocyte–endothelium cell interaction, and modulators of cell adhesion related to shock warrants further study.

The authors thank Barry Campbell for technical assistance.