Hepatic Scavenger Receptor BI Protects Against Polymicrobial-induced Sepsis through Promoting LPS Clearance in Mice*

Background: We utilized Scarb1I179N mice, a model deficient in hepatic SR-BI, to determine the role of hepatic SR-BI in sepsis. Results: Upon cecal ligation and puncture (CLP), Scarb1I179N mice had a 3.5-fold increase in fatality associated with an impaired LPS clearance. Conclusion: Hepatic SR-BI protects against sepsis through promoting LPS clearance. Significance: Promoting hepatic SR-BI-mediated LPS clearance may provide a therapeutic approach for sepsis. Recent studies revealed that scavenger receptor BI (SR-BI or Scarb1) plays a critical protective role in sepsis. However, the mechanisms underlying this protection remain largely unknown. In this study, using Scarb1I179N mice, a mouse model specifically deficient in hepatic SR-BI, we report that hepatic SR-BI protects against cecal ligation and puncture (CLP)-induced sepsis as shown by 75% fatality in Scarb1I179N mice, but only 21% fatality in C57BL/6J control mice. The increase in fatality in Scarb1I179N mice was associated with an exacerbated inflammatory cytokine production. Further study demonstrated that hepatic SR-BI exerts its protection against sepsis through its role in promoting LPS clearance without affecting the inflammatory response in macrophages, the glucocorticoid production in adrenal glands, the leukocyte recruitment to peritoneum or the bacterial clearance in liver. Our findings reveal hepatic SR-BI as a critical protective factor in sepsis and point out that promoting hepatic SR-BI-mediated LPS clearance may provide a therapeutic approach for sepsis.

Sepsis is a major health issue, which claims over 215,000 lives and costs $16.7 billion per year in the United States (1)(2)(3). The death rate of sepsis is high, exceeding 30%, because of the poor understanding of the disease (4). Identifying molecules involved in sepsis, especially endogenous protective modulators, is of great importance not only in understanding the mechanisms, but also in providing new insights for efficient therapies.
Recent studies revealed an important function of SR-BI, namely, protection against sepsis, as shown by a marked increase in fatality in mice deficient in SR-BI upon lipopolysaccharides (LPS) or cecal ligation and puncture (CLP) challenge (48 -51). Available evidence suggests that SR-BI may exert its protection via multiple mechanisms. Macrophage SR-BI suppresses inflammatory response by modulating LPS-TLR4 signaling in macrophages, which contributes to protection against septic animal death (50,52); Adrenal SR-BI is a key determinant of inducible glucocorticoid (GC) generation in response to stress (49,50,53), and SR-BI null mice are completely deficient in inducible GC generation in CLP-induced sepsis (50); A number of studies suggest that SR-BI may provide protection against sepsis through detoxifying LPS. SR-BI binds to LPS and mediates the uptake of LPS in vitro (54), and SR-BI-null mice display impaired LPS clearance in circulation in LPS-induced endotoxemia or CLP-induced sepsis (49,50). Liver is the most important organ for LPS detoxification. In sepsis, the majority of LPS target to Kupffer cells and hepatocytes (55)(56)(57). However, little is known about how the liver detoxifies LPS. It has been speculated that Kupffer cells and hepatocytes participate in uptake and clearance of LPS via a receptor-mediated mech-anism. Given the abundant expression of SR-BI in hepatocytes (9) and the role of SR-BI in LPS uptake and clearance, we hypothesized that hepatic SR-BI provides protection against sepsis by promoting LPS clearance.
Huby et al. developed loxp-floxed SR-BI mice but the mice exhibited hypo-phenotypes as shown by a marked reduction in SR-BI expression in all tissues (58). Overexpression of SR-BI by adenoviral vector is a useful and widely used animal model to elucidate the function of hepatic SR-BI in regulating HDL metabolism (20,59,60). Unfortunately, adenovirus induces host immune response, which may alter the outcomes of sepsis (61). This presents a barrier to use these animal models for septic study. Stylianou et al. recently reported an interesting hepatic specific SR-BI deficient mouse model, Scarb1 I179N mice (62). The Scarb1 I179N mutant mice had a 90% decrease in hepatic SR-BI protein expression, resulting in a 170% increase in plasma cholesterol concentrations compared with C57BL/6J control mice. However, the Scarb1 I179N mice had normal SR-BI expression in non-hepatic tissues such as adrenal gland and ovary (62). In this study, we utilized this unique animal model to assess the role of hepatic SR-BI in sepsis. We demonstrate that hepatic SR-BI protects against CLP-induced septic death by promoting LPS clearance. Our findings reveal hepatic SR-BI as a critical protective factor in sepsis and point out that promoting hepatic SR-BI-mediated LPS clearance may provide a therapeutic approach for sepsis.

EXPERIMENTAL PROCEDURES
Materials-Anti-SR-BI serum was custom made by Sigma-Genosys using a 15 amino acid-peptide derived from the C terminus of human SR-BI. LPS (Escherichia coli serotype K12) was from InvivoGen. The endotoxin low FBS was from Hyclone. ELISA kits for quantifying TNF-␣ and IL-6 were from eBioscience. The alanine aminotransferase (ALT) kit was from BQ Kits, Inc. (San Diego, CA). The competitive ELISA kit for quantifying mouse corticosterone was from Cayman. The ELISA kit for quantifying LPS was from Charles River.
Mice Specifically Deficient in Hepatic SR-BI (Scarb1 I179N )-Scarb1 I179N mice in C57BL/6J background were obtained from the Jackson Laboratory. The mice have an I179N point mutation resulting in a 90% decrease in liver SR-BI expression, but the mice have normal SR-BI expression in non-liver tissues (62). Thus, this is a unique animal model to elucidate the role of hepatic SR-BI in sepsis. C57BL/6J mice were used as controls. The animals were fed with a standard laboratory rodent diet. Animal care and experiments were approved by the Institutional Animal Care and Use Committee of the University of Kentucky.
PCR Genotyping of Scarb1 I179N Mice-The SR-BI I179N mutation is caused by a T to A transversion which eliminates a BstYI site. Utilizing this character, we developed a simple PCR genotyping method to replace TaqMan SNP genotyping (62). Forward primer: 5Ј-GCC TGA GCT CTT TGC CTG AA; reverse primer: 5Ј-CCC AAC AAA CAG GCC AAA. PCR was run for 35 cycles using standard conditions. Upon BstYI digestion of the PCR product, a single 294 bp band for Scarb1 I179N mice, and 96 bp and 198 bp bands for C57BL/6J controls were obtained.
Analysis of Leukocyte Recruitment to Peritoneum-leukocyte recruitment to peritoneal cavity was analyzed by flow cytometry as described (63)(64)(65). Briefly, 8 -10-week old Scarb1 I179N and C57BL/6J mice were treated with/without CLP for 6 h and the peritoneal fluids were collected. Peritoneal neutrophils (PMN, CD11b hi Ly6C hi ) and inflammatory monocytes (IM, CD11b int Ly6C hi ) were gated by CD11b, and Ly6C expression on CD45 ϩ cells. Ly6G expression was confirmed to be high in gated PMNs and low in gated IMs.
Biochemical Assays-10 -12-week-old mice were euthanized by CO 2 inhalation 6 and 20 h following CLP. The blood was obtained by cardiac puncture and stored at Ϫ80°C. Serum ALT levels were quantified with the ALT kit and were used as an indicator for liver damage; the serum TNF-␣, IL-6, corticosterone, and LPS levels were quantified with corresponding ELISA kits.
Quantification of Cytokine Generation in LPS-stimulated Macrophages-Bone marrow-derived macrophages were cultured as described previously (50). Briefly, bone marrow cells were cultured in a 12-well plate at 1.5 ϫ 10 6 cells/well in RPMI1640 medium containing 20% FBS and 15% supernatant of L929 cell culture for 5 days. For measurement of cytokines, the cells were incubated in PBS for 4 h and treated with LPS at 0.5 ng/ml for 20 h in RPMI1640 medium containing 20% FBS. Cytokines in the culture supernatant were quantified with corresponding ELISA kit. Of note, LPS from E. coli k12 strain, is much more potent than LPS from other commonly used strains with respect to the activity of stimulating inflammatory cytokine production.
Assay for Bacterial Load-Bacterial load in liver and spleen was analyzed as previously described (50).
Assay for SR-BI Expression-Tissue or cells were homogenized in lysis buffer containing 1% proteinase inhibitor mixture (Sigma), and subjected to Western blot analysis against SR-BI as described previously (50). The expression of SR-BI was quantified with Fuji LAS-4000 and normalized to actin expression.
Statistical Analysis-The survival assay was analyzed by Log-Rank x 2 test using SAS software. Significance in experiments comparing two groups was determined by 2-tailed Student's t test. Significance in experiments comparing more than two groups was evaluated by One Way ANOVA, followed by post hoc analysis using Tukey's test. Means were considered different at p Ͻ 0.05.

Hepatic SR-BI Protects against CLP-induced Septic Death-
The Scarb1 I179N mutation is caused by a T to A transversion in exon 4 which eliminates a BstYI site (R/GATCY). Based on this character, we designed a PCR method to genotype the mutant mice. As shown in Fig. 1A, PCR generated a 294-bp product. BstYI digestion of the PCR product yielded a single 294 bp band for Scarb1 I179N mice, but 96 bp and 198 bp bands for C57BL/6J controls. The Scarb1 I179N mice had a 90% decrease in SR-BI expression in the liver (Fig. 1B) and a 150% increase in serum cholesterol concentrations (Fig. 1C), which was consistent with the previous report (62). Thus, the Scarb1 I179N mice presented a unique animal model to elucidate the role of hepatic SR-BI in sepsis. As shown in Fig. 1D, CLP induced 21.4% fatality in C57BL/6J control mice, but 75% fatality in Scarb1 I179N mice, indicating that hepatic SR-BI provides significant protection against CLP-induced septic death. CLP-induced liver injury was assessed by measuring serum ALT levels. Compared with wild type controls, Scarb1 I179N mice had a significant increase in serum ALT levels at 6 and 20 h following CLP (Fig. 1E). These findings indicate that hepatic SR-BI protects against polymicrobial sepsis.
Exacerbated Innate Immune Response in Scarb1 I179N Mice during Sepsis-To understand why Scarb1 I179N mice were susceptible to septic death, we assessed inflammatory cytokine production at 6 and 20 h following CLP. Upon CLP, C57BL/6J control mice exhibited a typical acute phase response as shown by a rapid and strong induction of TNF-␣ and IL-6 in the early stage of sepsis (6 h) and significant decreases in TNF-␣ and IL-6 levels by 20 h (Fig. 2, A and B). Scarb1 I179N mice also had a rapid and strong induction of TNF-␣ and IL-6 in the early stage of sepsis (6 h); however, Scarb1 I179N mice displayed uncontrolled inflammatory cytokine generation as shown by high serum concentrations of TNF-␣ and IL-6 20 h after CLP (Fig. 2, A and B).
Hepatic SR-BI Is Required for LPS Clearance in Sepsis-Macrophages are one of the major types of cells that generate inflammatory cytokines during sepsis. SR-BI is moderately expressed in macrophages and recent studies demonstrated that macrophage SR-BI suppresses inflammatory response of macrophages to LPS (50,52). To test whether a deficiency of hepatic SR-BI affects inflammatory response in macrophages, we utilized bone marrow-derived macrophages from Scarb1 I179N and control mice. Upon stimulation by 0.5 ng/ml of LPS for 20 h, macrophages from Scarb1 I179N mice produced similar levels of TNF-␣ and IL-6 compared with macrophages from the control mice (Fig. 3, A and B).
We then looked for an alternative explanation for hepatic SR-BI's protective role in sepsis. Endotoxemia is a hallmark of sepsis. LPS released by Gram-negative bacteria is a major stimulator for systemic inflammatory cytokine production. We speculated that the exacerbated inflammatory cytokine production observed in Scarb1 I179N mice is caused by elevated LPS in the circulation. To test this, we measured serum LPS levels in CLP-challenged mice. Scarb1 I179N mice had a 3-fold increase in serum LPS levels compared with wild type controls 6 and 20 h following CLP (Fig. 4A), suggesting that hepatic SR-BI is responsible for LPS clearance in sepsis.
SR-BI is abundantly expressed in adrenal glands. Adrenal SR-BI mediates the intracellular uptake of cholesterol ester from HDL, which provides a substrate for glucocorticoid syn- Scarb1 I179N (n ϭ 12) and C57BL/6J mice (n ϭ 14) were treated with CLP and survival was observed for 7 days. The data were expressed as the percentage of mice surviving at indicated times, and survival was analyzed by Log-Rank x 2 test. E, CLP induced liver injury. Scarb1 I179N and C57BL/6J mice were treated with CLP for the indicated times, and liver injury was assessed by measuring serum ALT levels. n ϭ 6 -11 each group with duplicate measurements, mean Ϯ S.E. FIGURE 2. Exacerbated inflammatory cytokine production in Scarb1 I179N mice in sepsis. Scarb1 I179N and C57BL/6J mice were treated with/without CLP for the indicated times, and the serum concentrations of TNF-␣ (A) and IL-6 (B) were quantified. n ϭ 6 -11 each group with duplicate measurements, mean Ϯ S.E. thesis (9,12,49,66,67). Given the strong effects of glucocorticoids in regulating inflammatory response, we tested whether the exacerbated inflammatory cytokine generation is caused by a disruption in adrenal glucocorticoid production. As shown in Fig. 4B, CLP induced a significant increase in corticosterone generation in both Scarb1 I179N mice and wild type mice 6 and 20 h following CLP. Unexpectedly, CLP-Scarb1 I179N mice actually had a moderate increase in corticosterone concentrations compared with CLP-wild type mice. It is unlikely that the exacerbated inflammatory cytokine generation in Scarb1 I179N mice is caused by altered adrenal glucocorticoid production.
SR-BI is capable of binding Gram-negative bacteria and facilitating the uptake of Gram-negative bacteria in vitro (54,68). This raises a possibility that SR-BI might play a role in bacterial clearance in sepsis. Kupffer cells in the liver constitute a major pool of macrophages in the body. To determine whether hepatic SR-BI affects bacterial uptake by macrophages during sepsis, we isolated total DNA from the liver and quantified bacteria with qPCR using bacteria specific primers. No significant difference in the number of bacteria was observed between Scarb1 I179N mice and C57BL/6J control mice (Fig. 4C). We also quantified living bacteria in spleen, no significant difference in the number of living bacteria was observed between Scarb1 I179N mice and C57BL/6J mice (1.5 ϫ 10 5 Ϯ 2.0 ϫ 10 5 /g tissue in Scarb1 I179N mice versus 1.2 ϫ 10 5 Ϯ 2.5 ϫ 10 5 /g tissue in C57BL/6J mice).
SR-BI is a well-established HDL receptor. Hepatic SR-BI mediates the uptake of cholesterol ester from HDL, which plays a critical role in regulating HDL metabolism. Stylianou et al. reported that Scarb1 I179N mice display moderately larger HDL particles. We had similar observations (data not shown). HDL regulates the expression of adhesion molecules (69), and a deficiency of HDL has been shown to impair the recruitment of leukocytes to peritoneal cavity following CLP (65). To elucidate if hepatic SR-BI affects leukocyte recruitment in sepsis, we quantified peritoneal neutrophils and inflammatory monocytes with flow cytometry in Scarb1 I179N mice and C57BL/6J mice. Almost no peritoneal neutrophils and inflammatory monocytes were observed in non-CLP animals (Fig. 5A). Six hours following CLP, a significant number of neutrophils and inflammatory monocytes were recruited to the peritoneum (Fig. 5B) in both Scarb1 I179N and C57BL/6J mice, but no difference in leukocyte recruitment was observed between these two strains (Fig. 5C).

Role of Hepatic SR-BI in Gram-positive Bacteria-induced
Sepsis-The above findings suggest that hepatic SR-BI provides protection primarily through its role in promoting LPS clearance. To determine whether hepatic SR-BI provides protection other than LPS clearance, we infected Scarb1 I179N mice and C57BL/6J mice with Gram-positive bacteria S aureus, and monitored survival for 7 days. A moderate decrease in survival was observed in Scarb1 I179N mice compare with C57BL/6J mice (p ϭ 0.68) (Fig. 6).

DISCUSSION
Using Scarb1 I179N mice, a unique hepatic specific SR-BI deficient mouse model, we demonstrate that hepatic SR-BI is a critical protective factor in sepsis. Scarb1 I179N mice had a 3-fold increase in CLP-induced septic fatality associated with exacerbated inflammatory cytokine production compared with wild type control mice.
To understand the mechanisms by which hepatic SR-BI protects against sepsis, we tested five possible effects of hepatic SR-BI: 1) on LPS clearance. SR-BI has been shown to bind LPS and mediate the uptake of LPS in vitro (54), and SR-BI-null mice display impaired LPS clearance in circulation in LPS-induced endotoxemia or CLP-induced sepsis (49,50). These findings suggest that SR-BI plays a critical role in LPS clearance. Given that majority of LPS target to the liver in sepsis (55)(56)(57) and SR-BI is abundantly expressed in the liver, it is plausible that hepatic SR-BI functions to clear LPS. In this study, by measuring serum LPS levels, we found that Scarb1 I179N mice had a 3-fold increase in serum LPS concentrations compared with wild type control mice; 2) on macrophage inflammatory response. We found that macrophages from Scarb1 I179N and wild type control mice exhibit similar inflammatory cytokine production in response to LPS, suggesting that a deficiency of hepatic SR-BI does not alter the macrophage inflammatory response; 3) on adrenal steroidogenesis. We found that Scarb1 I179N mice have a moderate increase in corticosterone production compared with wild type control mice in sepsis. Given the potent suppressive effects of glucocorticoid on inflammation, it is unlikely that the increase in corticosterone production contributes to the elevated inflammatory cytokine production; 4) on bacterial clearance. We found no difference in the number of liver bacteria between Scarb1 I179N mice and wild type mice; 5) on leukocyte recruitment. Given that HDL regulates leukocyte recruitment, we assessed neutrophils and inflammatory monocytes in the peritoneum in response to CLP. We found no difference in the number of peritoneal neutrophils or inflammatory monocytes between Scarb1 I179N mice and wild type mice.
While LPS-induced inflammatory cell damage has been considered as a major cause of septic death, the failure of anti-LPS monoclonal antibodies in clinical trials questions the contribution of LPS to septic death. An issue associated with LPS neutralization is that the binding of antibody to LPS sequesters LPS, which delays LPS clearance (70), suggesting that simple neutralization of LPS may have limited effect on detoxifying LPS. It is worth noting that, different from neutralizing LPS, SR-BI detoxifies LPS through mediating the intracellular uptake of LPS and promoting LPS clearance (49,50,54). Com- bined with the observations that SR-BI-null mice are susceptible to LPS-induced endotoxic animal death (48,49), SR-BI is abundantly expressed in liver (9), liver is a major organ for LPS clearance (55)(56)(57) and a deficiency of hepatic SR-BI leads to impaired LPS clearance, we get our conclusion that hepatic SR-BI-mediated LPS clearance contributes to protection against sepsis. To further clarify this issue, we administered polymyxin B (PMB, a potent LPS neutralizer (71), 0.6 mg/kg body weight, intraperitoneal) to Scarb 1I179N and C57BL/6J mice 2h post CLP challenge. PMB profoundly inhibited IL-6 production in Scarb 1I179N mice but not in B6 mice 18 h post CLP, compared with mice without PMB treatment (0.47 Ϯ 0.08 ng/ml in CLP-Scarb1 I179N mice with PMB treatment versus 25.08 Ϯ 8.32 ng/ml in CLP-Scarb1 I179N mice without PMB treatment; 5.22 Ϯ 4.66 ng/ml in C57BL/6J mice with PMB treatment versus 2.66 Ϯ 1.21 ng/ml in C57BL/6J mice without PMB treatment). We only observed a slight increase in survival in PMB-treated septic Scarb 1I179N (33.3% survival in CLP-Scarb1 I179N mice with PMB treatment versus 25% survival in CLP-Scarb1 I179N mice without PMB treatment). The PMB treatment actually decreased the survival in septic B6 mice (33.3% survival in CLP-C57BL/6J mice with PMB treatment versus 78.6% survival in CLP-C57BL/6J mice without PMB treatment). These suggest that simple neutralization of LPS by LPS neutralizer cannot provide efficient protection against septic death and promoting LPS clearance is a more efficient way for LPS detoxification.
Our earlier study indicated that mice with whole body SR-BI deficiency are susceptible to CLP-induced septic death (50). While both whole body SR-BI-null mice and the hepatic SR-BIdeficient mice are susceptible to CLP-induced sepsis, we have noticed an interesting difference in inflammatory cytokine generation between these two animal models. Compared with strain-marched wild type control mice, Scarb1 I179N mice had a rapid induction of TNF-␣ and IL-6 in the early stage of sepsis (6 FIGURE 4. Impaired LPS clearance in Scarb1 I179N mice in sepsis. Scarb1 I179N and C57BL/6J mice were treated with CLP for 6 and 20 h, and the sera were collected for measurements of LPS (A) and corticosterone (B); the number of bacteria in the liver was analyzed by quantitative PCR (C). n ϭ 6 -11 each group with duplicate measurements, mean Ϯ S.E. FIGURE 5. Deficiency of hepatic SR-BI does not affect CLP-induced leukocyte recruitment to peritoneal cavity. Scarb1 I179N and C57BL/6J mice were treated with/without CLP for 6 h, and the peritoneal fluid was collected. Peritoneal neutrophils (PMN, CD11b hi Ly6C hi ) and inflammatory monocytes (IM, CD11b int Ly6C hi ) were gated by CD11b, and Ly6C expression on CD45 ϩ cells. A, non-CLP Scarb1 I179N and C57BL/6J mice. Almost no neutrophils and inflammatory monocytes were observed in the peritoneum of non-CLP mice. B and C, CLP Scarb1 I179N and C57BL/6J mice. Similar percentage of PMNs and IMs in CD45 ϩ cells was detected in Scarb1 I179N and C57BL/6J mice. C, similar number of recruited PMNs and IMs was detected in CLP Scarb1 I179N and C57BL/6J mice. n ϭ 7 per group, mean Ϯ S.E. FIGURE 6. Deficiency of hepatic SR-BI slightly affects Gram-positive bacteria-induced sepsis. Scarb1 I179N (n ϭ 17) and C57BL/6J (n ϭ 16) mice were intraperitoneal injected with 7 ϫ 10 8 CFU/mouse Staphylococcus aureus, and survival was monitored for 7 days. The data were expressed as the percentage of mice surviving at indicated times, and survival was analyzed by Log-Rank x 2 test. h), and uncontrolled inflammatory cytokine generation in the late stage of sepsis (20 h). Meanwhile, whole body SR-BI-null mice displayed delayed inflammatory cytokine generation in the early stage of sepsis and uncontrolled inflammatory cytokine generation in the late stage of sepsis (50); interestingly, the difference in inflammatory cytokine generation patterns is in line with the LPS levels in the circulation. Scarb1 I179N mice had elevated serum LPS in both the early and late stage of sepsis; but, the whole body SR-BI-null mice had lower serum LPS in the early stage of sepsis and elevated serum LPS in the late stage of sepsis (50). These findings suggest that hepatic SR-BI is responsible for LPS removal and a defect in hepatic SR-BI-mediated LPS removal leads to endotoxemia, which likely accounts for exacerbated inflammatory cytokine production observed in Scarb1 I179N mice; These findings also imply that non-hepatic SR-BI may be required for the recruitment of LPS to circulation, and a deficiency of non-hepatic SR-BI leads to a delayed inflammatory response in the early stage of sepsis. Endothelial cells have been shown to uptake LPS in a receptordependent manner (72,73). Interestingly, earlier studies showed that endothelial cells express high levels of SR-BI (23,24,74). It is likely that endothelial SR-BI is responsible for recruitment of LPS to circulation. Further study is warranted to test these speculations.
An earlier study pointed out that hepatic SR-BI expression is suppressed by LPS and by inflammatory cytokines such as TNF-␣ (75). Sepsis induces profound increases in circulating LPS and inflammatory cytokines, which likely causes the downregulation of hepatic SR-BI. Indeed, we found that the hepatic SR-BI expression is down-regulated in C57BL/6J mice 20 h post CLP (data not shown). Combined with our finding that hepatic SR-BI is required for LPS clearance, these observations suggest that promoting hepatic SR-BI expression may provide a therapeutic approach for sepsis. Further study is needed to test this speculation.
In summary, using the unique hepatic specific SR-BI-deficient animal model, we demonstrate that hepatic SR-BI provides protection against sepsis through promoting LPS clearance. Given that the expression of SR-BI is down-regulated in sepsis, we propose that promoting hepatic SR-BI-mediated LPS removal may provide a novel therapeutic strategy for sepsis.