Corticosteroid Therapy Benefits Septic Mice With Adrenal Insufficiency But Harms Septic Mice Without Adrenal Insufficiency

Abstract

Objectives:

Corticosteroid therapy is frequently used in septic patients given the rationale that there is an increased demand for corticosteroid in sepsis, and up to 60% of severe septic patients experience adrenal insufficiency. However, the efficacy of corticosteroid therapy and whether the therapy should be based on the results of adrenal function testing are highly controversial. The lack of an adrenal insufficiency animal model and our poor understanding of the pathogenesis caused by adrenal insufficiency present significant barriers to address this long-standing clinical issue.

Design:

Prospective experimental study.

Setting:

University laboratory.

Subjects:

Scavenger receptor BI null and adrenal-specific scavenger receptor BI null mice.

Interventions:

Sepsis was induced by cecal ligation and puncture.

Measurements and Main Results:

Using scavenger receptor BI^−/− mice as the first relative adrenal insufficiency animal model, we found that corticosteroid therapy significantly improved the survival in cecal ligation and puncture–treated scavenger receptor BI^−/− mice but causes more septic death in wild-type mice. We identified a corticosteroid cocktail that provides effective protection 18 hours post cecal ligation and puncture; using adrenal-specific scavenger receptor BI^−/− mice as an inducible corticosteroid–deficient animal model, we found that inducible corticosteroid specifically suppresses interleukin-6 production without affecting tumor necrosis factor-α, nitric oxide, and interleukin-10 production. We further found that inducible corticosteroid does not induce peripheral lymphocyte apoptosis but promotes phagocytic activity of macrophages and neutrophils.

Conclusions:

This study demonstrates that corticosteroid treatment benefits mice with adrenal insufficiency but harms mice without adrenal insufficiency. This study also reveals that inducible corticosteroid has both immunosuppressive and immunopermissive properties, suppressing interleukin-6 production, promoting phagocytosis of immune effector cells, but not inducing peripheral lymphocyte apoptosis. These findings support our hypothesis that corticosteroid is an effective therapy for a subgroup of septic patients with adrenal insufficiency but harms septic patients without adrenal insufficiency and encourage further efforts to test this hypothesis in clinic.

A hallmark of sepsis is the profound production of inducible corticosteroid (iCS) in response to septic stress. Given the rationale that there is an increased demand for corticosteroid in septic patients and 25–60% of severe septic patients experience adrenal insufficiency due to impaired corticosteroid production or signaling (1-4), low-dose corticosteroid is often used for septic shock patients. However, the efficacy of corticosteroid therapy for adult septic patients and whether the therapy should be based on the results of adrenal function testing are highly controversial (5). The French trial showed a reduction in mortality following corticosteroid therapy in septic shock patients with relative adrenal insufficiency (6); by contrast, the Corticosteroid Therapy of Septic Shock trial failed to show a mortality benefit (7). Numerous efforts have been made but the results are inconclusive (8-11). Nevertheless, the debate is still ongoing and recent studies show that early corticosteroid therapy may improve survival in septic shock patients (12, 13) or in sicker septic shock patients (14, 15).

Due to weak clinical evidence, the Survival Sepsis Campaign Guidelines recommend the use of hydrocortisone for septic shock patients at grade 2C (5). Despite the weak recommendation, data from the Survival Sepsis Campaign show that 8,992 out of 17,847 qualified septic patients received corticosteroid therapy, among which more than 30% of the patients received corticosteroid therapy after or despite the publication of the 2008 Guidelines (the 2012 Guidelines are essentially the same as the 2008 Guidelines regarding corticosteroid therapy [16]). A more recent study at the Emory Center for Critical Care indicated that 64% septic patients had hydrocortisone ordered when they were receiving two or more vasopressors (17). These data indicate that corticosteroid therapy is frequently used for septic patients even without clear evidence whether the treatment is beneficial or deleterious, which points out an urgent need to determine the efficacy of corticosteroid therapy.

There are three types of adrenal insufficiency in sepsis, defined by low plasma cortisol level (absolute), impaired iCS production in response to stress, which is diagnosed by a delta total cortisol of less than 9 μg/dL post–adrenocorticotropic hormone (ACTH) stimulation (relative), and impaired cellular corticosteroid signaling in response to corticosteroid (corticosteroid resistance) (18). Relative adrenal insufficiency is the most common type in septic patients. The French trial showed that corticosteroid therapy benefits septic patients with relative adrenal insufficiency (6). However, the CORTICUS trial failed to confirm a mortality benefit (7). The heterogeneity of septic patients and the technical difficulties in identifying adrenal insufficiency present a significant barrier to evaluate the efficacy of corticosteroid therapy in septic patients (19-21). Given these limitations, it is necessary to evaluate corticosteroid therapy in an adrenal insufficiency animal model to provide a proof of concept whether corticosteroid therapy is beneficial or deleterious.

Scavenger receptor BI (SR-BI or Scarb1) is a well-characterized high-density lipoprotein (HDL) receptor and plays a key role in reverse cholesterol transport (22, 23). Recent studies in our laboratory and others demonstrated that SR-BI is a critical protective factor in sepsis (24-31). SR-BI is most abundantly expressed in adrenal glands, mediating intracellular cholesterol uptake from HDL for corticosteroid synthesis (22, 32). SR-BI^−/− mice lack iCS production in response to ACTH stimulation or in stressed conditions including cecal ligation and puncture (CLP)-induced sepsis but have roughly normal plasma corticosterone at physiologic conditions (25, 29, 33, 34). These present SR-BI^−/− mice a unique relative adrenal insufficiency animal model to determine the efficacy of corticosteroid therapy and to assess the pathogenesis caused by relative adrenal insufficiency.

In this study, we used SR-BI^−/− mice as the first relative adrenal insufficiency animal model to test our conceptual hypothesis that iCS is protective in sepsis and corticosteroid therapy benefits septic mice with adrenal insufficiency. We demonstrate that corticosteroid supplementation significantly prevents CLP-induced septic death in SR-BI^−/− mice but causes more septic death in wild-type mice. Our study suggests that it may be necessary to determine the status of adrenal insufficiency and only prescribe corticosteroid for septic patients with adrenal insufficiency. Our study also reveals that iCS has both immunosuppressive and immunopermissive properties, not simply acting as a potent immunosuppressive molecule.

MATERIALS AND METHODS

SR-BI^+/− mice in 129×B6 background were from the Jackson Laboratory. SR-BI^−/− and SR-BI^+/+ littermates were generated by breeding SR-BI^+/− mice. The animals were fed a standard laboratory diet and kept with a 10/14 hours light/dark cycle and used at 10–14 weeks old. Animal care and experiments were approved by the Institutional Animal Care and Use Committee of the University of Kentucky. Please see detailed Materials and Methods in the Supplemental Materials (Supplemental Digital Content 1, http://links.lww.com/CCM/B390).

RESULTS

Corticosteroid Supplementation Benefits Septic Mice With Adrenal Insufficiency

SR-BI^−/− mice have normal corticosterone level at physiologic conditions but lack inducible corticosterone production in response to ACTH stimulation or septic stress (25, 29). Thus, SR-BI^−/− mice present a useful relative adrenal insufficiency animal model to determine whether corticosteroid therapy is beneficial or deleterious. We first supplemented corticosterone, the major endogenous corticosteroid in rodents, to CLP-treated SR-BI^−/− mice. We found that corticosterone treatment changed survival from 38% to 80% (p = 0.059) (Fig. 1A). We then elucidated the effect of hydrocortisone—the corticosteroid currently used in septic patients—and found that hydrocortisone treatment changed survival from 38% to 70% (p = 0.073) (Fig. 1B). Considering that septic stress induces a group of corticosteroids which may work together, we speculated that a corticosteroid cocktail may provide better protection. Indeed, we identified a novel corticosteroid cocktail containing a mixture of 3 corticosteroids with varied potency and half-lives and demonstrated that the corticosteroid cocktail worked effectively improving survival from 38% to 81% (p = 0.003) (Fig. 1C).

Figure 1.

Supplementation of corticosteroid protects against cecal ligation and puncture (CLP)–induced septic death in mice with adrenal insufficiency. CLP (23-G needle, half ligation) was conducted on scavenger receptor BI (SR-BI)^−/− mice. Immediately following CLP, the mice were treated with/without 200 μg corticosterone (A), 100 μg hydrocortisone (B), or a corticosteroid cocktail (C) containing 100 μg hydrocortisone, 20 μg 6δ-methylprednisolone, and 25 ng fludrocortisone (doses given per 25 g body weight, s.c.). Survival was monitored for 7 days. Data are expressed as the percentage of mice surviving at the indicated time points and analyzed by the log-rank χ² test.

Corticosteroid Supplementation Harms Septic Mice Without Adrenal Insufficiency

We further assessed the effect of corticosteroid supplementation on wild-type mice. In contrast to SR-BI^−/− mice, supplementation of corticosterone to CLP-treated SR-BI^+/+ mice not only failed rescuing the mice but caused significantly more animal death, as shown by a decrease in survival from 81% to 40% (p = 0.012) (Fig. 2A). Supplementation of corticosteroid cocktail also caused significantly more animal death (p < 0.001) (Fig. 2A). These data indicate that corticosteroid supplementation harms septic mice without adrenal insufficiency. To understand the underlying cause, we analyzed the effects of corticosteroid supplementation on adrenal SR-BI expression, endogenous corticosterone production, interleukin (IL)-6 level, and bacterial load. Supplementation of corticosteroid cocktail to CLP-SR-BI^+/+ mice caused a moderate decrease in adrenal SR-BI expression (Fig. 2, B and C), a significant decrease in endogenous corticosterone (Fig. 2D) and IL-6 (Fig. 2E) production at 4 hours post CLP, and a marked increase in the bacterial load in circulation (Fig. 1F) and peritoneum (Fig. 2G) 20 hours post CLP. These data suggest that the exogenous corticosteroid disrupts endogenous corticosteroid generation and further suppresses IL-6 production, which may contribute to an increase in bacterial load and animal death.

Figure 2.

Supplementation of corticosteroid (CS) causes more septic death in mice without adrenal insufficiency. A, Survival analysis. Cecal ligation and puncture (CLP) (22-G needle, full ligation) was conducted on scavenger receptor BI (SR-BI)^+/+ mice. Immediately following CLP, the mice were treated with/without 200 μg corticosterone or the CS cocktail containing 100 μg hydrocortisone, 20 μg 6δ-methylprednisolone, and 25 ng fludrocortisone acetate. Survival was monitored for 5 days. Data are expressed as the percentage of mice surviving at the indicated time points and analyzed by the log-rank χ² test. B–G, CLP (22-G needle, full ligation) was conducted on SR-BI^+/+ mice. Immediately following CLP, the mice were treated with/without the CS cocktail and the blood and tissues were harvested. The adrenal glands were analyzed for SR-BI expression with Western blot (B) and normalized to actin expression (C). n = 6 in each group. Plasma corticosterone (D) and interleukin (IL)-6 (E) levels and the bacteria number in blood (F) or peritoneum (G) were quantified. n = 9 in each group. Data represent means ± sem.

Corticosteroid Therapy Provides Protection Post CLP in SR-BI^−/− Mice

To determine whether corticosteroid cocktail can be used as a sepsis therapy, we evaluated the effect of corticosteroid cocktail post CLP. As shown in Figure 3A, supplementation of a single dose of corticosteroid cocktail to SR-BI^−/− mice 2 hours post CLP did not show effective protection (p = 0.14). When we looked at the survival of the first 3 days and 3–7 days, we noticed that the single dose of corticosteroid cocktail treatment resulted in 73% survival (44% in control group) in the first 3 days (p = 0.10) but lost its protection at 3–7 days (p = 0.907) (Supplemental Fig. S1, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). This suggests that the loss of protection might be caused by unable to maintaining corticosteroid levels by a single dose of corticosteroid cocktail. Thus, we modified the regimen of corticosteroid treatment by administering the corticosteroid cocktail 2 hours post CLP and then administering the corticosteroid cocktail daily at ½ doses for 4 days. This treatment regimen significantly improved the survival of SR-BI^−/− mice (p = 0.017) (Fig. 3B). We further evaluated the effect of corticosteroid cocktail 18 hours post CLP. Supplementation of the corticosteroid cocktail 18 hours post CLP and then daily at ½ doses for 4 days significantly improved the survival from 31% to 67% (p = 0.043) (Fig. 3C).

Figure 3.

Supplementation of corticosteroid (CS) post cecal ligation and puncture (CLP) protects against septic death in mice with adrenal insufficiency. CLP (23-G needle, half ligation) was conducted on scavenger receptor BI (SR-BI)^−/− mice. Two hours or 18 hr following CLP, the mice were treated with/without a CS cocktail containing 100 μg hydrocortisone, 20 μg 6δ-methylprednisolone, and 25 ng fludrocortisone acetate. A, Treated 2 hr post CLP with one dose of the CS cocktail. B, Treated 2 hr post CLP with the CS cocktail and then further treated daily with ½ dose of the CS cocktail for 4 days. C, Treated 18 hr post CLP with the CS cocktail and then further treated daily with ½ dose of the CS cocktail for 4 days. Survival was monitored for 7 days. Data are expressed as the percentage of mice surviving at the indicated time points and analyzed by the log-rank χ² test.

Pathogenesis Caused by Relative Adrenal Insufficiency

A lack of understanding about the pathogenesis caused by adrenal insufficiency presents a significant barrier for corticosteroid therapy. Considering that SR-BI plays multiple protective roles in sepsis, for example, the liver SR-BI promotes the clearance of lipopolysaccharide (LPS) (23), to specifically assess the pathogenesis caused by relative adrenal insufficiency, we generated adrenal-specific SR-BI^−/− mice by adrenal transplantation following the method described by Van Eck group (34). Six weeks post transplantation, the adrenal gland developed well (Fig. 4A; Supplemental Fig. S2, Supplemental Digital Content 1, http://links.lww.com/CCM/B390), which is consistent with the previous report (34). Upon ACTH stimulation or CLP, the adrenal SR-BI^+/+ mice displaced a marked increase in corticosterone, but the adrenal-specific SR-BI^−/− mice failed to generate this inducible corticosterone (Fig. 4, B and C). Of note, the adrenal-specific SR-BI^−/− mice did not show a significant difference in plasma corticosterone concentration compared with the control mice at physiologic conditions, confirming the adrenal-specific SR-BI^−/− mice as a useful model of relative adrenal insufficiency.

Figure 4.

Mice with adrenal insufficiency are susceptible to cecal ligation and puncture (CLP)–induced septic death and kidney injury. Adrenal gland from 9-day-old scavenger receptor BI (SR-BI)^+/+ or SR-BI^−/− mice was transplanted to adrenalectomized adult SR-BI^+/+ mice to generate adrenal-specific SR-BI^+/+ or SR-BI^−/− mice. Six weeks after the transplantation, the transplanted adrenal developed well under the capsule of kidney as shown by the morphology and hematoxylin and eosin staining (A). B, Plasma corticosterone levels 1 hr post adrenocorticotropic hormone (ACTH) stimulation. C–H, The mice were subjected to CLP (21-G needle, half ligation). C, The plasma corticosterone concentrations at indicated times (n = 7–10). D, Seven-day survival analysis. Data are expressed as the percentage of mice surviving at the indicated time points and analyzed by the log-rank χ⁷ test (n = 18 for adrenal SR-BI^+/+ and n = 13 for adrenal SR-BI^−/− mice). E, Body temperature at 18 hr post CLP (n = 7–10). F, Plasma alanine aminotransferase (ALT) concentrations as a marker for liver injury. G, Lung wet/dry (W/D) ratios as an indicator for lung injury. H, Blood urea nitrogen (BUN) levels as a marker for kidney damage. n = 6–10 for F–H. Data represent means ± sem.

Mice with adrenal insufficiency are susceptible to CLP-induced septic death and kidney injury. As shown in Figure 4D, CLP induced a 76.9% fatality in adrenal-specific SR-BI^−/− mice compared with a 16.7% fatality in adrenal SR-BI^+/+ control mice (p < 0.001). Low body temperature is an indicator of septic shock (35). As shown in Figure 4E, 18 hours post CLP, most of the adrenal-specific SR-BI^−/− mice had a body temperature below 30°C, and the average body temperature was significantly lower in adrenal-specific SR-BI^−/− mice than in adrenal SR-BI^+/+ control mice (p = 0.004). Organ injury is a hallmark of sepsis. To understand why mice with adrenal insufficiency are susceptible to CLP-induced sepsis death, we looked at organ injury by assessing liver, lung, and kidney damages in CLP. As shown in Figure 4, F and G, no difference in alanine aminotransferase level or lung wet/dry ratio was observed between adrenal-specific SR-BI^−/− mice and adrenal SR-BI⁺⁷⁺ control mice. However, the adrenal-specific SR-BI^−/− mice had a 2.4-fold increase in plasma blood urea nitrogen concentrations 18 hours post CLP compared with the control mice (Fig. 4H). Collectively, these observations indicate that adrenal insufficiency causes kidney injury and susceptibility to CLP-induced septic shock and death.

Mice with adrenal insufficiency have aberrant inflammatory response in sepsis; inflammatory cytokines are largely responsible for organ injury. Corticosteroid is known as a potent “anti-inflammatory molecule” (36). To investigate the effect of adrenal insufficiency on inflammatory response, we quantified plasma cytokine and nitrogen oxides (NOx), or NOx (NO and NO₂), levels 4 and 18 hours post CLP. Unexpectedly, we did not observe a significant difference in tumor necrosis factor (TNF)-α or NOx level between adrenal-specific SR-BI^−/− mice and wild-type controls (Fig. 5, A and B). Interestingly, we found a significant increase in IL-6 level in adrenal-specific SR-BI^−/− mice compared with wild-type controls, and the IL-6 level in the adrenal-specific SR-BI^−/− mice remained high even at 18 hours post CLP (Fig. 5C). We also looked at plasma IL-10, an anti-inflammatory cytokine. We did not found a difference in IL-10 concentration between adrenal-specific SR-BI^−/− mice and wild-type controls. We further analyzed the correlation between the plasma corticosterone and the cytokine levels. We found a significant negative correlation between plasma corticosterone and IL-6 concentrations (p < 0.05; Supplemental Table S1, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). These data indicate that adrenal insufficiency causes uncontrolled IL-6 production but does not affect TNF-α, NOx, or IL-10 production during sepsis.

Figure 5.

Mice with adrenal insufficiency have aberrant inflammatory response in sepsis. Cecal ligation and puncture (CLP) (21-G needle, half ligation) was conducted on adrenal-specific scavenger receptor BI (SR-BI)^+/+ or SR-BI^−/− mice. Plasma was harvested by cardiac puncture at 0, 4, and 18 hr post CLP, and the concentrations of tumor necrosis factor (TNF)-α (A), nitrogen oxides (NOx) (B), interleukin (IL)-6 (C), and IL-10 (D) were quantified. n = 6–10. Data represent means ± sem. *p < 0.05.

Mice with adrenal insufficiency have impaired phagocytosis in sepsis. The phagocytic activity of phagocytes is impaired in septic patients, which is associated with a high mortality. Thus, we evaluated the phagocytosis of fluorescein isothiocyanate (FITC)-labeled Escherichia coli. Compared with the control mice, the adrenal-specific SR-BI^−/− mice had a 60% lower percentage of E. coli⁺ monocytes (CD115⁺CD11b⁺Ly6C⁺; 1.53% ± 0.61% vs 3.85% ± 0.85%; p = 0.064) and an unchanged percentage of E. coli⁺ neutrophils (Gr-1⁺CD11b⁺; 43.78% ± 4.86% vs 42.36% ± 7.10%; p = 0.873) in blood (Fig. 6, A and B). The capability of phagocytosis was evaluated by the mean fluorescence intensity (MFI) of E. coli in the phagocytes. We found that the adrenal-specific SR-BI^−/− mice displayed a 39% lower MFI of E. coli in monocytes (19.3 ± 3.1 × 10³ vs 31.6 ± 3.3 × 10³; p = 0.027) and a 55% lower MFI of E. coli in neutrophils (9.2 ± 2.7 × 10³ vs 20.1 ± 1.5 × 10³; p = 0.011) than the control mice (Fig. 6, A and B). We also looked at the phagocytosis of FITC-labeled E. coli in spleen. The adrenal-specific SR-BI^−/− mice had a 54% lower percentage of E. coli⁺ macrophages (F4/80⁺; p = 0.131), a 40% lower percentage of E. coli+ neutrophils (Gr1⁺CD11b⁺; p = 0.101), and a 25% lower percentage of E. coli+ dendritic cells (CD3⁻CD11c⁺; p = 0.254), but the differences did not reach statistical significance (Supplemental Fig. S3A, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). There was no significant difference in MFI of E. coli in splenic macrophages and neutrophils in the adrenal-specific SR-BI^−/− mice (Supplemental Fig. S3B, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). These findings indicate that adrenal insufficiency impairs the phagocytic activity of phagocytes.

Figure 6.

Mice with adrenal insufficiency have impaired phagocytosis, impeded activation of lymphocytes, but do not affect lymphocyte apoptosis in sepsis. Cecal ligation and puncture (CLP) (21-G needle, half ligation) was conducted on adrenal-specific scavenger receptor BI (SR-BI)^−/− and SR-BI^+/+ mice. For analysis of phagocytic activity of phagocytes, the mice were IV injected with 1 × 10⁸ heat-killed fluorescein isothiocyanate (FITC)-Escherichia coli 17 hr post CLP, and 1 hr later, the mean fluorescence intensity (MFI) of E. coli in monocytes (A) and neutrophils (B) in blood was analyzed with flow cytometry. n = 4–5. For analysis of activation of splenic lymphocytes, the spleens were harvested 18 hr post CLP and analyzed using CD69 as marker (C). For analysis of lymphocyte apoptosis, the splenic lymphocytes were analyzed by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining (D). n = 7–9. Data represent means ± sem.

Mice with adrenal insufficiency have impeded activation of lymphocytes in sepsis. Using CD69 as an activation marker, we analyzed the effect of adrenal insufficiency on the activation of lymphocytes in spleen. The adrenal-specific SR-BI^−/− and wild-type control mice showed a similar percentage of activated T cells (CD69⁺CD3⁺; 2.68% ± 0.16% vs 2.74% ± 0.25%; p = 0.837) and activated B cells (CD69⁺CD19⁺; 2.78% ± 0.17% vs 3.35% ± 0.33%; p = 0.182) in total lymphocytes under physiologic conditions. Interestingly, at 18 hours post CLP, we observed a two- to three-fold increase in the percentage of activated B and T cells in the wild-type control mice (data not shown), but the adrenal-specific SR-BI^−/− mice displayed a significantly lower percentage of activated T cells (Fig. 6C).

Adrenal insufficiency does not affect peripheral lymphocyte apoptosis. Corticosteroid is well known to induce lymphocyte apoptosis in vitro (37, 38); however, whether iCS is responsible for inducing peripheral lymphocyte apoptosis in sepsis remains to be validated. Adrenal-specific SR-BI^−/− mice present a unique animal model to test this speculation. We analyzed lymphocyte apoptosis in spleen 18 hours post CLP. Unexpectedly, we did not observe a difference between adrenal-specific SR-BI^−/− mice and control mice in the number of splenocytes, the percentage of lymphocytes, or the percentage of T cells or B cells in lymphocytes in the spleen (Supplemental Fig. S4, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). We then analyzed apoptotic lymphocytes in the spleen by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining. There was no difference in the percentage of TUNEL⁺ lymphocytes, TUNEL⁺ T cells, or TUNEL⁺ B cells in adrenal-specific SR-BI^−/− mice compared with wild-type mice (Fig. 6D).

We also tested the effect of corticosteroid treatment on adrenal-specific SR-BI^−/− mice. As shown in Supplemental Figure S5 (Supplemental Digital Content 1, http://links.lww.com/CCM/B390), supplementation of corticosteroid cocktail to adrenal-specific SR-BI^−/− mice suppressed plasma IL-6 levels at 4 hours post CLP and provided significant protection against CLP-induced septic death compared with the adrenal-specific SR-BI^−/− mice treated without corticosteroid.

DISCUSSION

Corticosteroid is synthesized in adrenal glands using cholesterol as substrate. The intracellular cholesterol comes from three sources, by uptake of cholesteryl ester through HDL/SR-BI or low-density lipoprotein (LDL)/low-density lipoprotein receptor (LDLr) pathway or de novo cholesterol synthesis. SR-BI^−/− mice have normal corticosterone level at physiologic conditions but fail to generate iCS in response to septic stress or ACTH stimulation (25, 29). Interestingly, the mice have higher expression of adrenal LDLr, 3-hydroxy-3-methylglutaryl-coenzyme A reductase (key modulator of cholesterol synthesis) (29). These observations suggest that the LDL/LDLr pathway or de novo cholesterol synthesis can provide sufficient cholesterol for corticosteroid production at physiologic conditions, but the HDL/SR-BI pathway is required for massive iCS production in stressed conditions.

In this preclinical study, using SR-BI^−/− mice as an adrenal insufficiency animal model, we demonstrated for the first time that corticosteroid therapy benefits septic mice with adrenal insufficiency but harms septic mice without adrenal insufficiency. Importantly, we demonstrated that supplementation of corticosteroid cocktail 18 hours post CLP still provides effective protection, suggesting that corticosteroid cocktail might be used as a therapy for septic patients with adrenal insufficiency. Our study reveals that iCS has both immunosuppressive and immunopermissive properties.

Implications for Corticosteroid Therapy

This preclinical study provides a number of insights that may be of translational potential: 1) although the current Survival Sepsis Campaign Guidelines do not emphasize the diagnosis of adrenal insufficiency before the administration of corticosteroid (5), our data suggest that adrenal insufficiency is a key indicator for the use of corticosteroid therapy, and we should avoid using corticosteroid for septic patients without adrenal insufficiency; 2) when we compared corticosteroid treatment immediately following CLP to 2 hours post CLP, we found the survival changed from 81% to 53% (p = 0.072; Supplemental Fig. S6, Supplemental Digital Content 1, http://links.lww.com/CCM/B390), although the data were not statistically significant due to a small number of mice used. This finding suggests that we should apply corticosteroid once adrenal insufficiency is identified, not wait until irreversible hypotension has developed. In support of this notion, a recent study showed that early corticosteroid therapy improves survival in septic shock patients (12, 13); 3) this study also supports the use of corticosteroid cocktail. Sepsis induces a group of corticosteroids which may work together. The French trial included fludrocortisone, but the CORTICUS trial did not (6, 7). This suggests that mineralocorticoid may be beneficial, although this speculation remains to be verified; the corticosteroid is dynamically produced and maintains at high levels. When we treated the SR-BI^−/− mice 2 hours post CLP, we found that a single dose of corticosteroid cocktail treatment resulted in a survival of 73% compared with 44% in control group in the first 3 days but no protection after 3 days (Supplemental Fig. S1, Supplemental Digital Content 1, http://links.lww.com/CCM/B390). Considering that the corticosteroid cocktail has a corticosteroid half-life of 10–30 hours, this suggests that a single dose of corticosteroid may not be enough to maintain corticosteroid levels post CLP. Indeed, we found that administering the corticosteroid cocktail daily at ½ doses for 4 more days was required for improving the survival when the first corticosteroid treatment was given 2 or 18 hours post CLP. Thus, we think that the use of corticosteroid cocktail may be more efficient in maintaining the corticosteroid levels compared with corticosterone (~20 min) or hydrocortisone levels (~10 hr) (39, 40).

Collectively, this study suggests that it is critical to determine the status of adrenal insufficiency and only prescribe corticosteroid to septic patients with adrenal insufficiency. This study further suggests that trailing the composition and dose regimen of corticosteroid cocktail might provide more efficient protection against sepsis.

Pathogenesis Caused by Relative Adrenal Insufficiency

During sepsis, the production of iCS is markedly induced in response to an increased demand for corticosteroid; however, the septic patients with relative adrenal insufficiency lose the capability to generate this iCS. While corticosteroid is generally considered to be a potent anti-inflammatory reagent and this concept is well supported by its wide and effective use for the treatment of a variety of inflammatory diseases (36), the pathogenesis caused by adrenal insufficiency and whether iCS still acts as an immunosuppressive molecule remain to be tested in the context of sepsis. Using adrenal-specific SR-BI^−/− mice as an adrenal insufficiency animal model, we found that these mice displayed aberrant inflammatory response in sepsis. Interestingly, compared with the control mice, the adrenal-specific SR-BI^−/− mice displaced a significant increase in IL-6 production at 4 and 18 hours post CLP, without a significant difference in the generation of TNF-α, NOx, or IL-10. These observations suggest that iCS may not universally act as an anti-inflammatory molecule; rather it may selectively suppress inflammatory cytokine production, particularly IL-6.

Sepsis features a massive depletion of peripheral lymphocytes, which causes immunosuppression and is considered as a risk factor in sepsis (41). Synthetic corticosteroid is well known for inducing lymphocyte apoptosis (37, 38). Unexpectedly, we found that there is no significant difference in splenic lymphocyte apoptosis between adrenal-specific SR-BI^−/− mice and wild-type control mice although the wild-type control mice generated high levels of iCS in response to CLP. Of note, we recently reported that SR-BI-mediated iCS is responsible for inducing thymocyte apoptosis in sepsis (26). Thus, these data indicate that the endogenous iCS does not induce peripheral lymphocyte apoptosis, and the thymocytes and peripheral lymphocytes undergo apoptosis with different mechanisms in sepsis. Interestingly, we found that iCS is required for the activation of T lymphocytes.

Several in vitro studies showed that corticosteroid has opposing effects on macrophage function dependent on the concentrations of corticosteroid (42, 43) and can increase the phagocytic activity of cultured human monocytes through regulating genes involved in the phagocytosis (44). Our study revealed that iCS in sepsis significantly enhances the phagocytic activity of monocytes and neutrophils in CLP-induced sepsis. Thus, we for the first time provide in vivo evidence that iCS is supportive to the functions of phagocytes in sepsis.

Collectively, we identified a number of unique functions of iCS in sepsis, such as suppressing IL-6 production without affecting TNF-α and NOx, promoting phagocytosis of immune effector cells, but not inducing peripheral lymphocyte apoptosis. These findings indicate that iCS has both immunosuppressive and immunopermissive properties, not simply a potent immunosuppressive molecule as we thought.

Numerous efforts have been made to determine the role of corticosteroid in sepsis. Animals with adrenalectomy are completely deficient in corticosteroid production so that it can be considered as absolute adrenal insufficiency model. Adrenalectomy renders animals susceptible to sepsis (45), which supports the idea that corticosteroid should be provided for septic patients with absolute adrenal insufficiency. Using an endotoxemia model, Cai et al (29) showed that pretreatment of SR-BI^−/− mice with corticosterone 8 hours prior to LPS protects the mice from LPS-induced endotoxic death. However, no such protection was observed in CLP-challenged SR-BI^−/− mice following the same regimen (25). These paradoxical observations can be explained that the corticosterone pretreatment suppresses cytokine production so that it rescues SR-BI^−/− mice from LPS-induced death since inflammatory cytokine is a major cause of fatality in LPS model; however, for CLP model, the inhibition of cytokine production by corticosterone pretreatment compromises innate immunity to fight against infections so that it fails to rescue SR-BI^−/− mice; a recent study by Leelahavanichkul et al (46) administered high dose of dexamethasone to SR-BI^−/− and C57BL/6 mice 24 hours prior to CLP and found less SR-BI^−/− mice death compared with C57BL/6 mice. This observation was likely caused by problematic experimental approaches (27). The pretreatment of the C57BL/6 mice with dexamethasone completely abolished adrenal SR-BI expression (47), which generates adrenal-specific SR-BI knockout C57BL/6 mice. Thus, the less SR-BI^−/− mice death was likely caused by the difference in strain background between the SR-BI^−/− (129/B6 mixed background) and wild-type control mice (B6 background) (27). Gilbert et al (31) recently assessed the protective function of adrenal SR-BI using Sf1CreSR-BI^fl/fl conditional knockout mice (hypomSR-BI^ΔAC) in B6 background and showed that the hypomSR-BI^ΔAC mice lack inducible corticosterone production upon CLP challenge, and the mice are susceptible to CLP-induced animal death. These findings are consistent with our observations. They showed that the hypomSR-BI^ΔAC mice have a marked increase in TNF-α, IL-6, and IL-10 levels but a significant decrease in NOx levels compared with control mice upon LPS challenge. However, in our study, we found that the adrenal-specific SR-BI^−/− mice have uncontrolled IL-6 production upon CLP challenge but only have minor changes in TNF-α, IL-10, and NOx levels. It is worth noting that the SR-BI^fl/fl mice display hypomorphism with a 90% decrease in SR-BI expression globally (48), which may contribute to the differential inflammatory response between the hypomSR-BI^ΔAC mice and our adrenal-specific SR-BI^−/− mice. For example, a decrease in SR-BI expression in the liver impaired LPS clearance, which promotes inflammatory response (29, 31). In support of this speculation, an earlier study showed that mice with a dimerization-deficient glucocorticoid receptor (GR^dim) display uncontrolled IL-6 production but has minor effects on TNF-α production in response to LPS (49).

A limitation of the adrenal transplantation model is the cutoff of the communication between the preganglionic sympathetic neuron and the chromaffin cell in the adrenal medulla, which is responsible for the secretion of catecholamine vasopressors (50). The lack of hemodynamic effects of catecholamine may make the mice more susceptible to sepsis. However, the control mice also experienced the same disrupted neuron communication. Therefore, this limitation may not affect the results.

CONCLUSIONS

In this preclinical study, we demonstrate that corticosteroid treatment benefits septic mice with adrenal insufficiency but harms septic mice without adrenal insufficiency. Importantly, we demonstrate that supplementation of the corticosteroid cocktail 18 hours post CLP still provides effective protection. This study also reveals that iCS has both immunosuppressive and immunopermissive properties, suppressing IL-6 production and promoting phagocytosis of immune effector cells. These findings support our hypothesis that the corticosteroid cocktail can be used an effective therapy for a subgroup of septic patients with adrenal insufficiency but harms septic patients without adrenal insufficiency. We hope that this study will encourage more efforts to improve the diagnosis of relative adrenal insufficiency, to test our hypothesis in septic patients, and to tailor the corticosteroid components for better corticosteroid therapy.