and Septic Vascular Dysfunction Signaling in the Pathogenesis of Septic Shock B k A Pivotal Role of Endothelial-Specific NF-LiuJianqiang

Although the role of NF- (cid:1) B in the pathogenesis of sepsis and septic shock has been extensively studied, little is known about the causative contribution of endothelial-intrinsic NF- (cid:1) B to these pathological processes. In this study, we used transgenic (TG) mice (on FVB genetic background) that conditionally overexpress the NF- (cid:1) B inhibitor, mutant I- (cid:1) B (cid:2) , selectively on endothelium and their transgene-negative littermates (wild type (WT)) to deﬁne the causative role of endothelial-speciﬁc NF- (cid:1) B signaling in septic shock and septic vascular dysfunction. In WT mice, LPS challenge caused systemic hypotension, a signiﬁcantly blunted vasoconstrictor response to norepinephrine, and an impaired endothelium-dependent vasodilator response to acetylcholine, concomitant with a markedly increased aortic inducible NO synthase expression, signiﬁcantly elevated plasma and aortic levels of nitrite/nitrate, increased aortic TNF- (cid:2) expression, and decreased aortic endothelial NO synthase (eNOS) expression. In TG mice whose endothelial NF- (cid:1) B was selectively blocked, LPS caused signiﬁcantly less hypotension and no impairments in vasoconstrictor and endothelium-dependent vasodilator responses, associated with signiﬁcantly reduced aortic inducible NO synthase expression, decreased plasma and aortic levels of nitrite/nitrate, reduced aortic TNF- (cid:2) expression, and increased aortic eNOS expression. TNF- (cid:2) knockout mice prevented LPS-induced eNOS down-regulation. WT mice subjected to cecal ligation and puncture showed signiﬁcant systemic hypotension, which was prevented in TG mice. Our data show that selective blockade of endothelial-intrinsic NF- (cid:1) B pathway is sufﬁcient to abrogate the cascades of molecular events that lead to septic shock and septic vascular dysfunction, demonstrating a pivotal role of endothelial-speciﬁc NF- (cid:1) B signaling in the pathogenesis of septic shock and septic vascular dysfunction. The Journal of Immunology, 2009, 183: 4031–4038. The EC-TG mice display endothelial-restricted blockade of NF- (cid:1) B pathway (17) and enable us to selectively inhibit endothelial NF- (cid:1) B activation in vivo under physiological setting. A preliminary study using those mice showed that endothelial NF- (cid:1) B blockade partially reversed endotoxemic hypotension (17). The current study extends our preliminary study by exam-ining the effects of selective blockade of endothelial NF- (cid:1) B pathway on the cascades of molecular events that lead to septic shock and septic vascular dysfunction in LPS and cecal ligation and puncture (CLP) models of sepsis. We demonstrated that blockade of endothe-lial-speciﬁc NF- (cid:1) B signaling is sufﬁcient to abrogate the molecular cascades leading to septic vascular dysfunction. Our data deﬁne the mechanistic role of endothelial-intrinsic NF- (cid:1) B in the pathogenesis of septic shock and septic vascular dysfunction, and provide new in-sights into the molecular mechanisms of sepsis and septic shock. Endothelial NF- (cid:1) B blockade inhibited TNF- (cid:2) and iNOS expression in smooth muscle cells mice that underwent CLP also showed a signiﬁcant systemic hypotension. Consistent with the changes in physiological parameters, WT-LPS mice showed a markedly increased aortic iNOS protein expression, elevated plasma and aortic levels of nitrite/nitrate, the stable end prod-ucts of NO, increased aortic TNF- (cid:2) protein expression, and decreased aortic eNOS protein expression. In sharp contrast to WT-LPS mice, TG-LPS mice with endothelial-selective NF- (cid:1) B blockade displayed no signiﬁcant hypotension, a normal vasoconstrictor response to NE, no impairment in the endothelium-dependent vasodilator response to Ach, reduced aortic iNOS expression, decreased plasma and aortic levels of nitrite/nitrate, reduced aortic TNF- (cid:2) expression, and increased aortic eNOS expression. TG-CLP mice were also prevented from systemic hypotension. Taken together with our previous demonstration that the EC-TG mice express the NF- (cid:1) B inhibitor, I- (cid:1) B (cid:2) mt, only on endothelium, and display endothelial-selective blockade of NF- (cid:1) B activation (17), these results illustrate that blockade of endothelial-intrinsic NF- (cid:1) B pathway mitigates the cas-cade of molecular events that lead to septic shock and septic vascular dysfunction, implying that endothelial-speciﬁc NF- (cid:1) B signaling plays a pivotal role in the development of septic shock and septic vascular dysfunction.

S eptic shock and septic vascular dysfunction are characterized by systemic hypotension, persistent vasodilatation, hyporesponsiveness to vasoconstrictors, and impairment of endothelium-dependent vasodilator response. The mechanisms of septic shock and septic vascular dysfunction are complex and multiple (1)(2)(3). One well-established mechanism is the activation of NF-B pathway. Bacterial pathogens and their products, such as LPS, activate NF-B, which causes inducible NO synthase (iNOS) 4 expression (2, 3), leading to an excessive production of NO. NO released subsequently causes vasodilatation, vascular hyporeactivity, and hypotension by activating soluble guanylyl cyclase-dependent mechanism (4 -7). NF-B activation mediates the expression of numerous cytokines, which lead to further activation of NF-B, amplifying, and perpetuating the inflammatory response (2). LPS and cytokines, such as TNF-␣, down-regulate endothelial NO synthase (eNOS) expression (8,9), which is believed to be an important molecular event underlying the impaired endothelium-dependent vasodilator response (10,11). Animal studies have demonstrated that inhibition of NF-B activation inhibits multiple inflammatory gene expression (2,12), reverses systemic hypotension (2,(13)(14)(15), corrects myocardial dysfunction (16), and prevents the impairment of endothelium-dependent vasodilatation (2,14).
The pathogenic role of NF-B activation in septic shock and septic vascular dysfunction is unquestionable. However, the contribution of individual cell-specific NF-B to these pathological processes is significantly less clear. The pathophysiology of sepsis and septic shock involves complex cell-cell and mediator-mediator interactions (1,2). Emerging evidence suggests that different cell-intrinsic NF-B may play distinct role in the pathophysiology of sepsis (17)(18)(19). Elucidation of the contribution of individual cell-specific NF-B to the complicated pathological process of septic shock will help to better understand the pathologic mechanisms of septic shock.
The vascular endothelium is considered an important mechanism regulating vascular tone and vascular homeostasis (4,20). In response to physiological stimuli and hemodynamic forces, endothelial cells (ECs) release contracting and relaxing factors, including endothelin, NO, and prostacyclin. These factors alter vascular tone by directly causing vasoconstriction or vasodilatation, and by synergistically or counteractively interacting with neurotransmitters, enhancing or diminishing the neurally induced vascular contraction or relaxation (4,20). Endothelium-derived NO inhibits adrenergic neural contraction (4), and mediates the vasodilator response to acetylcholine, the cholinergic neural transmitter (4), as well as the vasodilator responses to vasoactive humoral substances (4).
The central role of NF-B in septic pathologies and the predominant influence of endothelium on vascular homeostasis suggest that activation of endothelial-specific NF-B signaling may play an essential role in the development of septic shock and septic vascular dysfunction. However, the causative role of endothelialintrinsic NF-B in septic shock has not been studied. A number of studies have examined endothelial NF-B activation and its role in endothelial inflammation (21)(22)(23)(24). However, those studies were performed with cultured ECs, and have not addressed the pathogenic role of endothelial NF-B activation in septic shock and septic vascular dysfunction. Several previous studies, including our own, have shown that inhibition of NF-B activation using chemical NF-B inhibitors alleviates septic shock and septic vascular dysfunction (13)(14)(15). However, those inhibitors inhibit NF-B activation in all cell types, and may have effects that are not related to NF-B inhibition. The causal contribution of endothelial-intrinsic NF-B to the pathogenesis of septic shock and septic vascular dysfunction has never been examined, due to lack of an investigative tool.
Recently, we have created and characterized transgenic (TG) mice designated as EC-TG mice, in which a mutant I-B␣ (I-B␣mt), a specific inhibitor of NF-B, is expressed in ECs under the control of the tetracycline gene regulatory system (17). The EC-TG mice display endothelial-restricted blockade of NF-B pathway (17) and enable us to selectively inhibit endothelial NF-B activation in vivo under physiological setting. A preliminary study using those mice showed that endothelial NF-B blockade partially reversed endotoxemic hypotension (17). The current study extends our preliminary study by examining the effects of selective blockade of endothelial NF-B pathway on the cascades of molecular events that lead to septic shock and septic vascular dysfunction in LPS and cecal ligation and puncture (CLP) models of sepsis. We demonstrated that blockade of endothelial-specific NF-B signaling is sufficient to abrogate the molecular cascades leading to septic vascular dysfunction. Our data define the mechanistic role of endothelial-intrinsic NF-B in the pathogenesis of septic shock and septic vascular dysfunction, and provide new insights into the molecular mechanisms of sepsis and septic shock.

Animal groups
The generation and characterization of the EC-TG mice that conditionally overexpress I-B␣mt selectively on endothelium have been previously described (17). In this study, we used this mouse strain to define the causative contribution of endothelial-intrinsic NF-B to septic shock and septic vascular dysfunction. We studied eight groups of mice (8 -10 wk, on FVB genetic background), as follows: transgene-negative control or sham (wild type (WT)-Con, WT-sham), transgene-negative LPS or CLP (WT-LPS, WT-CLP), TG control or sham (TG-Con, TG-sham), and TG LPS or CLP (TG-LPS, TG-CLP). We also studied four groups of mice on B6129S genetic background (from The Jackson Laboratory; stock numbers, WT mice, 101045, TNF-␣ knockout (KO), 003008), as follows: WT-Con, WT-LPS, TNF-␣ KO control (TNF-␣-KO-Con), and TNF-␣-KO-LPS. All animal experiments were approved by the institutional animal care and use committee and complied with National Institutes of Health Guide.

Measurement of systemic blood pressure
Mice were anesthetized with tribromoethanol (300 mg/kg, i.p.), intubated, and ventilated with a mouse ventilator, as we have previously described (13). We chose to use tribromoethanol as anesthetics because it causes less cardiovascular depression (25). A microcannula was inserted into carotid artery for continuously monitoring systemic blood pressure. Mouse body temperature was kept constant with a servo-controlled electronic blanket and intraanal thermal probe. After a 30-min equilibration period and measurement of basal blood pressure, mice were injected with saline or LPS (Escherichia coli 0111:B4, 2.5 mg/kg, i.p.). Systemic blood pressure was recorded for 4 h, and mean arterial blood pressure (MBP) was calculated. In a separate set of experiments, mice were injected with saline or LPS (10 mg/kg, i.p.). At 24 h after saline or LPS injection, systemic blood pressure was recorded, as described above.
For the CLP model, mice were anesthetized and cannulated at 18 h after operation, and systemic blood pressure was recorded, as described above.

Assessment of vascular reactivity in vivo
Mice were anesthetized and cannulated at 5.5 h after saline or LPS (10 mg/kg, i.p.) injection. Because basal blood pressure influences vascular reactivity, mice that had low initial MBP were resuscitated with 6% dextran in 7.5% NaCl during the equilibration period to ensure a comparable baseline MBP among all groups. Following the measurement of baseline MBP, dose-response relationship to ␣-adrenergic receptor agonist, norepinephrine (NE; 30, 100, and 300 ng/kg, i.v. bolus injection), to the endothelium-dependent vasodilator, acetylcholine (Ach; 60, 200, and 600 ng/ kg, i.v. bolus injection), or to the endothelium-independent vasodilator, sodium nitroprusside (SNP; 60, 200, and 600 ng/kg, i.v. bolus injection), was recorded in three separate sets of experiments. The maximum increase or decrease in MBP elicited by each dose of NE, Ach, or SNP was calculated and compared.

Assessment of vasoreactivity in isolated mesenteric vascular bed
At 6 h after saline or LPS (10 mg/kg, i.p.) injection, the mesenteric vascular bed was isolated, as previously described (26), and perfused with oxygenated physiological salt solution at a constant flow rate of 200 l/min. Because perfusion flow rate is constant, changes in perfusion pressure represent changes in vascular resistance. Following a 40-min equilibration period, dose response to NE (30, 100, and 300 ng) was recorded. To study vasodilator response, perfusion pressure was elevated by ϳ60 mmHg by perfusing the mesenteric vascular beds with 100 M NE. After a sustained elevation in perfusion pressure was achieved, dose response to Ach or SNP (1, 10, and 100 ng) was recorded in two separate sets of experiments. Ach or SNP was injected into the perfusion circuit immediately proximal to the mesenteric artery. The maximal increase or decrease in perfusion pressure caused by each dose of NE, Ach, or SNP was calculated and compared.

CLP model of sepsis
Mice in sham groups were subjected to sham, and in CLP groups subjected to CLP operation, as we have previously described (17), using an 18-gauge needle. At 18 h postoperation, mice were cannulated for systemic blood pressure measurement, as described above.

Western blot
Aortic levels of iNOS and eNOS proteins were determined by Western blot, as we previously described (17), using Abs against iNOS, eNOS, and actin (all from Santa Cruz Biotechnology).

Assays for plasma and aortic levels of nitrite/nitrate, and aortic level of TNF-␣
Plasma and aortic levels of nitrite/nitrate, the stable metabolic product of NO, were measured using nitrite/nitrate assay kit (Cayman Chemical). Aortic level of TNF-␣ was determined using ELISA kit (eBioscience).

Immunohistochemical staining
Cryosections (6 m) were prepared from aorta of each group of mice at 6 h after saline or LPS injection, fixed with paraformaldehyde, permeabilized, blocked with blocking solution, and incubated with rabbit anti-TNF-␣ Ab (Abcam) or rabbit anti-iNOS Ab (Santa Cruz Biotechnology) overnight at 4°C. Specific binding was detected with biotinylated secondary Ab-HRP complexes using VECTASTAIN Elite ABC kits (Vector Laboratories). Ag-Ab complexes were visualized using 3Ј,3Ј-diaminobenzidine (Vector Laboratories). Sections were counterstained with hematoxylin, mounted, and viewed under light microscope.

Statistical analysis
Data were expressed as mean Ϯ SEM, and analyzed using ANOVA or Kruskal-Wallis rank test, followed by Holm-Sidak method or Dunnett's test for post hoc analysis. The null hypothesis was rejected at 5% level.

Selective blockade of endothelial NF-B reduced systemic hypotension
To define the role of endothelial intrinsic NF-B signaling in septic shock, WT-Con and TG-Con were injected with saline, and WT-LPS and TG-LPS mice were injected with LPS. MBP was monitored for 4 h. Baseline MBP was identical in the four groups of mice and decreased slightly over time, most likely due to loss of blood or body fluids (Fig. 1A). At 3 and 4 h post-LPS, WT-LPS mice showed a marked drop in MBP, which was significantly less in TG-LPS (Fig. 1A). The effect of endothelial NF-B blockade on the development of septic shock was further examined at late time points in both LPS and CLP models of sepsis. Compared with control or sham group of mice, WT-LPS or WT-CLP mice showed a significant drop in MBP at 24 h after LPS injection (Fig. 1B) or at 18 h after CLP operation (Fig. 1C), which was significantly attenuated or prevented in TG-LPS or TG-CLP mice (Fig. 1, B and C). These results unveil an important role for endothelial-specific NF-B signaling in the development of septic shock.

Selective blockade of endothelial NF-B restored vasoconstrictor response to NE
A major feature of septic vascular dysfunction is the repressed vasoconstrictor response to catecholamine. We have therefore examined whether endothelial-selective NF-B blockade alters the pressor response to NE in control and endotoxemic mice. Because basal vascular tone affects vascular reactivity, mice with low initial MBP were resuscitated to ensure comparable levels of MBP among all groups of mice before starting the NE trial. Baseline MBP was 94 Ϯ 1, 91 Ϯ 2, 93 Ϯ 1, and 93 Ϯ 1 mmHg for WT-Con, WT-LPS, TG-Con, and TG-LPS group, respectively. As expected, NE caused a dose-dependent elevation in MBP. Compared with that in WT-Con and TG-Con mice, the NE-elicited elevation in MBP decreased significantly in WT-LPS at all three doses ( Fig. 2A). In contrast, the NEelicited elevation in MBP in TG-LPS mice was comparable to that in WT-Con and TG-Con mice, and was significantly higher than that in WT-LPS mice ( Fig. 2A).
In vivo vasoreactivity is affected by systemic factors such as cardiac output and reflex. To avoid these effects, we further as-sessed the NE response in isolated perfused mesenteric vascular beds. Because the vascular bed was perfused at constant flow rate, changes in perfusion pressure represent alteration in vascular resistance. Mesenteric perfusion pressure was comparable among the four groups of mice at baseline, but was elevated by NE injection in a dose-dependent manner (Fig. 2B). Compared with WT-Con and TG-Con mice, the NE-elicited elevation in mesenteric perfusion pressure was greatly attenuated in WT-LPS mice (Fig. 2B), but was not affected in TG-LPS mice (Fig. 2B). These results indicate that blockade of endothelial-intrinsic NF-B abrogates LPSinduced repression of the vasoconstrictor response to NE, suggesting that activation of endothelial-intrinsic NF-B pathway plays an important role in the development of vascular hyporesponsiveness to NE in endotoxemic mice.

Selective blockade of endothelial NF-B prevented the impairment of endothelium-dependent vasodilator response
We next examined whether endothelial-selective NF-B blockade prevents the impairment of endothelium-dependent vasodilator response, another major feature of septic vascular dysfunction. Mice were injected with saline or LPS, cannulated, and resuscitated, as described above. Baseline MBP was comparable among the four groups of mice before initiation of Ach or SNP trial. Both Ach (endothelium-dependent vasodilator) and SNP (endothelium-independent vasodilator) caused dose-dependent drop in MBP (Fig. 3). Compared with WT-Con and TG-Con mice, WT-LPS mice displayed a significantly blunted vasodilator response to Ach at all three doses (Fig. 3A). In contrast, TG-LPS mice showed an Ach response that was identical with that of WT-Con and TG-Con mice, and was significantly bigger than that of WT-LP mice (Fig.  3A). The four groups of mice showed a similar vasodilator response to SNP (Fig. 3B).
Likewise, Ach and SNP caused dose-dependent drop in perfusion pressure in the isolated perfused mesenteric vascular bed (Fig.  4). The drops in perfusion pressure evoked by the three doses of Ach, but not by the three doses of SNP, were significantly less in WT-LPS mice as compared with that of WT-Con and TG-Con mice (Fig. 4). In TG-LPS mice, Ach or SNP caused a drop in mesenteric perfusion pressure that was similar to that in WT-Con and TG-Con mice, and the Ach-mediated response was significantly bigger than that of WT-LPS mice (Fig. 4). Overall, these results illustrate that selective blockade of endothelial NF-B pathway prevents the LPS-induced impairment of endothelium-dependent vasodilator response to Ach, implying that activation of endothelial-intrinsic NF-B plays an important role in the impairment of endothelium-dependent vasodilator response during endotoxemia.

Selective blockade of endothelial NF-B reduced iNOS expression
LPS causes systemic hypotension and vascular hyporeactivity by inducing iNOS expression, resulting in overproduction of NO, which causes hypotension and blunts vasoconstrictor response (5-7). To investigate the mechanism through which en-dothelial-selective NF-B blockade restores systemic MBP and vascular reactivity, we determined aortic iNOS protein expression in WT and TG mice. Fig. 5A consists of Western blot photographs showing that endothelial-selective NF-B blockade inhibited LPS-induced iNOS protein expression in aortae. The iNOS bands were quantified using densitometry and summarized in Fig. 5B. Aortic level of iNOS protein was negligible in WT-Con and TG-Con mice, increased markedly in WT-LPS mice, but reduced by ϳ73% in TG-LPS mice, as compared with that in WT-LPS mice (Fig. 5).
Endothelial-selective NF-B blockade also inhibited iNOS activity, as indicated by the reduced plasma and aortic levels of nitrite/nitrate in TG-LPS mice (Fig. 6). Compared with WT-Con and TG-Con mice, WT-LPS mice showed a ϳ5-fold increase in plasma level and 6-fold increase in aortic level of nitrite/nitrate, which were reduced by ϳ64 and 52% in TG-LPS mice (Fig. 6). Thus, blockade of endothelial-intrinsic NF-B

Selective blockade of endothelial NF-B prevented LPS down-regulation of eNOS expression
Ach causes vasodilatation by activating eNOS, resulting in the release of endothelium-derived NO (10,11). We have therefore compared aortic levels of eNOS protein expression between WT-LPS and TG-LPS mice. Fig. 7A consists of Western blot photographs showing that endothelial-selective NF-B blockade abrogated LPS-induced eNOS protein down-regulation. Fig. 7B summarized the densitometry quantification of the eNOS bands. Consistent with impaired vasodilator response to Ach, WT-LPS mice showed a significantly reduced aortic expression of eNOS protein, as compared with WT-Con and TG-Con mice (Fig. 7). Aortic level of eNOS protein in TG-LPS mice was identical with that of WT-Con and TG-Con mice (Fig. 7). These results suggest that activation of endothelial-specific NF-B contributes to the LPS-induced down-regulation of eNOS expres-sion, which accounts for the impairment of endothelium-dependent vasodilator response to Ach in endotoxemic mice.

TNF-␣ is required for LPS-induced eNOS down-regulation
To establish a link between activation of endothelial NF-B signaling and eNOS down-regulation, we examined the effect of endothelial NF-B blockade on aortic TNF-␣ expression, and determined the role of TNF-␣ in LPS-induced eNOS downregulation in our mouse model. Compared with WT-Con and TG-Con mice, WT-LPS mice displayed a 10-fold increase in aortic level of TNF-␣ protein, which was reduced by ϳ51% in TG-LPS mice (Fig. 8). To define the role of TNF-␣ in LPSinduced eNOS down-regulation, we compared aortic levels of eNOS protein expression between TNF-␣-KO mice and their genetic background-matched WT mice at 6 h after LPS injection. Compared with that in WT-Con and TNF-␣-KO-Con mice, aortic eNOS expression was significantly down-regulated in WT-LPS mice, which was prevented in TNF-␣-KO-LPS mice (Fig. 9). This result indicates that TNF-␣ plays an obligatory role in LPS-induced eNOS down-regulation. Collectively, these results suggest that activation of endothelial NF-B signaling leads to eNOS down-regulation via TNF-␣-dependent mechanisms.

Endothelial NF-B blockade inhibited TNF-␣ and iNOS expression in smooth muscle cells
We next performed immunohistochemical staining of aortic sections from WT and TG mice to identify the cells whose levels of iNOS and TNF-␣ expression were inhibited by endothelial-selective NF-B blockade. TNF-␣-and iNOS-positive cells were not detectable in aortic sections from WT-Con and TG-Con mice (Fig.  10, A, B, E, and F), markedly increased in aortic sections from WT-LPS mice (Fig. 10, C and G), and significantly reduced in aortic sections from TG-LPS mice (Fig. 10, D and H). Strong TNF-␣ and iNOS staining was localized to ECs and vascular smooth muscle cells (VSMCs) in aorta of WT-LPS mice (Fig. 10,  C and G). In TG-LPS mice, TNF-␣and iNOS-positive ECs were barely detected, and TNF-␣-and iNOS-positive VSMCs were significantly reduced (Fig. 10, D and H). This result illustrates that endothelial-selective NF-B blockade diminishes EC TNF-␣ and iNOS expression, and also inhibits VSMC TNF-␣ and iNOS expression.

Discussion
The major finding of this study is that endothelial-specific NF-B signaling plays a pivotal role in the pathogenesis of septic shock and septic vascular dysfunction. Challenge of WT mice with LPS resulted in a marked drop in systemic MBP at both early and late time points, a significantly blunted vasoconstrictor response to NE, and an impaired endothelium-dependent vasodilator response to Ach. WT mice that underwent CLP also showed a significant systemic hypotension. Consistent with the changes in physiological parameters, WT-LPS mice showed a markedly increased aortic iNOS protein expression, elevated plasma and aortic levels of nitrite/nitrate, the stable end products of NO, increased aortic TNF-␣ protein expression, and decreased aortic eNOS protein expression. In sharp contrast to WT-LPS mice, TG-LPS mice with endothelial-selective NF-B blockade displayed no significant hypotension, a normal vasoconstrictor response to NE, no impairment in the endotheliumdependent vasodilator response to Ach, reduced aortic iNOS expression, decreased plasma and aortic levels of nitrite/nitrate, reduced aortic TNF-␣ expression, and increased aortic eNOS expression. TG-CLP mice were also prevented from systemic hypotension. Taken together with our previous demonstration that the EC-TG mice express the NF-B inhibitor, I-B␣mt, only on endothelium, and display endothelial-selective blockade of NF-B activation (17), these results illustrate that blockade of endothelial-intrinsic NF-B pathway mitigates the cascade of molecular events that lead to septic shock and septic vascular dysfunction, implying that endothelial-specific NF-B signaling plays a pivotal role in the development of septic shock and septic vascular dysfunction.
Several groups, including our own, have shown that inhibition of NF-B activation reduces systemic hypotension and restores or partially restores vascular reactivity in mice or rats subjected to endotoxemia or multimicrobial sepsis (13)(14)(15). However, those studies used nonspecific chemical NF-B inhibitors, which may have nonspecific effects. More importantly, those studies inhibited NF-B in all cell types, and did not address the function of cell-specific NF-B in the pathological processes. This study is the first to define the specific contribution of endothelial-intrinsic NF-B signaling to septic shock and septic vascular dysfunction, and thus provides novel insights into the molecular mechanisms of septic shock and septic vascular dysfunction.
It is well documented that LPS or sepsis causes systemic hypotension and vascular hyporesponsiveness by activating NF-B-mediated iNOS expression (2,3), leading to an excessive production of NO. The marked inhibition of LPS-induced iNOS expression and activity by selective blockade of endothelial NF-B pathway with a concomitant abrogation of LPSinduced hypotension and vascular hyporeactivity supports the notion that endothelial NF-B blockade alleviates septic hypotension and septic vascular dysfunction by inhibiting NF-Bmediated iNOS expression.
We showed that LPS impaired endothelium-dependent vasodilator response in WT-LPS, but not in TG-LPS mice. The preservation of endothelium-dependent vasodilator response in TG-LPS mice could be explained by the prevention of TNF-␣-mediated eNOS down-regulation. First, TNF-␣ is a NF-B-regulated gene product (2, 3), and blockade of endothelial NF-B significantly inhibited LPS-induced TNF-␣ expression in TG-LPS mice. Second, TNF-␣ is known to down-regulate eNOS expression (8). We demonstrated in the current study that TNF-␣ mediates LPS-induced eNOS down-regulation. Third, it is well documented that Ach elicits vasodilatation by stimulating eNOS-mediated endothelial NO release (4,10,11). Fourth, we demonstrated in this study that blockade of endothelial NF-B concomitantly inhibited the LPS-induced TNF-␣ up-regulation and eNOS down-regulation in vascular tissue, and restored endothelium-dependent vasodilatation. Inhibition of endothelial NF-B activation suppresses local inflammation within ECs, which could reduce the production of reactive oxygen species and increase NO bioavailability (27). This may also contribute to the preserved endothelium-dependent vasodilator response to Ach in TG-LPS mice.
Physiologically, vascular tone is mainly influenced by the physical and biochemical properties of VSMCs. NF-B pathway in smooth muscle cells is not inhibited in our EC-TG mice (17). The effectiveness of endothelial-restricted NF-B inhibition in correcting septic vascular dysfunction and in inhibiting aortic iNOS expression is somewhat unexpected. Endothelialselective inhibition of NF-B activation restored systemic MBP to 80% of control level, completely abrogated LPS-induced repression of vasoconstrictor responses to NE, and inhibited LPSinduced aortic iNOS protein expression by ϳ73%. These results highlight the importance of endothelium-and endothelial-specific NF-B signaling in septic shock and septic vascular dysfunction. These results suggest that cross-talk between ECs and VSMCs may be an important mechanism regulating inflammatory response within vascular wall during sepsis. In support of this speculation, we demonstrated in this study that selective blockade of endothelial NF-B pathway not only diminished LPS-induced TNF-␣ and iNOS expression in ECs, but also significantly inhibited TNF-␣ and iNOS expression in VSMCs. Further studies to elucidate the mechanisms underlying the paracrine interactions between ECs and VSMCs will help to better understand the molecular mechanisms of vascular wall inflammation during sepsis and other pathological conditions.
Our current results are in good agreement with our previous studies demonstrating that endothelial selective NF-B blockade ameliorated multiple organ injury and improved survival in septic mice (17). Our results are also consistent with two recent reports showing that endothelial-specific NF-B suppression attenuated hypertension-induced renal damage (28) and high fat diet-induced atherosclerosis (29). It is reported that blockade of endothelial NF-B pathway enhanced LPS-induced endothelial permeability (30). This result does not necessarily contradict our findings, because different models (conventional vs conditional TG mice) were used in the two studies. The two mouse models have different basal endothelial barrier integrity and, therefore, display different response to LPS in term of alteration in endothelial permeability (17,30). Nevertheless, both studies illustrate a critical role of endothelial NF-B in controlling endothelial barrier integrity and function.
Our findings have therapeutic implications. The central roles of NF-B in systemic inflammation and in septic pathology indicate that NF-B is an ideal target for therapeutic intervention (2,(12)(13)(14)(15)(16). However, NF-B inhibition impairs host defense mechanism and causes immune suppression (31,32). Consequently, the beneficial anti-inflammatory effect can be offset by the detrimental proinfectious effect of NF-B inhibition, leading to unaltered or even worse outcome. To overcome this problem, we need to develop novel approaches that can selectively inhibit NF-B-mediated inflammatory and injurious responses (detrimental) without significantly interfering with the NF-B-mediated immune and host defense responses (ben-eficial). Our current and previous demonstrations that selective blockade of NF-B-driven inflammatory response within endothelium ameliorated septic multiple organ injury (17), abrogated septic vascular dysfunction, and improved survival, but had no effect on bacterial clearance capacity (17) provide experimental basis for endothelial-selective NF-B inhibition as an innovative strategy to develop sepsis therapies.