Insulin-like growth factor-1 improves survival in sepsis via enhanced hepatic bacterial clearance

Abstract

Rationale: Both insulin-like growth factor (IGF)-1 and bacterial clearance by Kupffer cells are significantly reduced in severe sepsis. Kupffer cell apoptosis is triggered by tumor necrosis factor (TNF)-alpha and activation of the PI-3 kinase pathway prevents TNF-induced Kupffer cell death.

Objectives: We evaluated if the marked decline in IGF-1 is related to bacterial clearance in sepsis.

Methods: Sepsis was induced in C57BL/6 mice by intratracheal inoculation with Pseudomonas aeruginosa (strain PA103). Some mice received IGF-1 24 mg/kg either before infection or 12 hours after infection. In vitro studies were performed using the clonal Kupffer cell line KC13-2.

Measurements and main results: Sepsis resulted in decreased levels of IGF-1. In vitro studies with KC13-2 cells demonstrated that IGF-1 protected Kupffer cells against TNF-alpha-induced apoptosis by activating the PI-3 kinase pathway and stabilizing the inhibitor of apoptosis protein, XIAP. In the animal model, pretreatment with IGF-1 decreased hepatic TNF-alpha and IL-6, improved hepatic bacterial clearance as demonstrated by real-time polymerase chain reaction with primers specific for P. aeruginosa, and improved survival in severe sepsis. Moreover, we rescued mice from severe sepsis by IGF-1 treatment 12 hours after infection.

Conclusions: These studies show that the decline in IGF-1 levels in sepsis is related to bacterial clearance and that replacement of IGF-1 in a murine model of sepsis improves overall survival.

Figure 1.

Severe sepsis results in decreased serum IGF-1 and blood glucose levels. (A) Serum levels of IGF-1 were measured in human subjects with severe sepsis 24 hours after the clinical recognition of organ dysfunction. Six healthy volunteers served as control subjects. Mann-Whitney test demonstrates a significant reduction in serum IGF-1 levels in septic subjects compared with controls (*P < 0.05). (B) C57BL/6 mice received intratracheal inoculation of PA103 at either 5 × 10³ cfu (mild sepsis) or 5 × 10⁴ cfu (severe sepsis). Sham control animals received an equivalent volume of saline. At predetermined time points, animals were killed according to animal care guidelines. IGF-1 was measured in mouse serum via ELISA. Analysis of variance (ANOVA) followed by Bonferroni's test for multiple comparisons demonstrates that serum IGF-1 levels are decreased in both mild and severe sepsis compared with sham controls at 4 hours (*P < 0.01). However, by 24 hours (a time when bacterial clearance is impaired), IGF-1 levels in severe sepsis remain reduced compared with both mild sepsis and sham controls (**P < 0.05). Graph reflects the mean and SD of n > 4 animals in each group. (C) Blood glucose was measured using a hand-held glucometer. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that blood glucose levels are reduced in animals after severe sepsis at 24 hours (*P<0.05) compared with both mice with mild sepsis and sham control animals. Graph reflects the mean and SD of n > 4 animals in each group. Solid bars, sham control; hatched bars, mild sepsis; cross-hatched bars, severe sepsis.

Figure 2.

IGF-1 treatment decreases the duration of hepatic inflammation in severe sepsis. (A) C57BL/6 mice were inoculated with PA103 intratracheally to induce severe sepsis. A subset of mice was treated with IGF-1 24 mg/kg subcutaneously either just before infection or 12 hours after the onset of infection. At 24 hours after infection, mice were killed according to animal care guidelines. Serum IGF-1 levels were measured by ELISA. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that treatment with IGF-1 either before or 12 hours after the onset of infection resulted in an increase in serum IGF-1 levels compared with animals treated with severe sepsis alone (*P < 0.05). The level of IGF-1 after treatment (before infection or 12 h after infection) remained lower than that seen in the control animals (P < 0.05). Graph reflects the mean and SD of n > 4 animals in each group. (B) Blood glucose was measured using a hand-held glucometer. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that treatment with IGF-1 either before or 12 hours after the onset of infection resulted in an increase in blood glucose levels compared with animals treated with severe sepsis alone (*P<0.05). Graph reflects the mean and SD of n > 4 animals in each group. (C) After 24 hours of infection, livers were harvested and homogenized as described. TNF-α and IL-6 were measured in liver lysates via ELISA. Results were normalized per gram of liver tissue. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that treatment with IGF-1 either before or after the onset of infection resulted in a decrease in TNF-α and IL-6 at 24 hours compared with animals treated with severe sepsis alone (*P < 0.05). Graphs reflect mean and SD of n > 4 animals in each group.

Figure 3.

Treatment with IGF-1 improves bacterial clearance in severe sepsis. (A) C57BL/6 mice were inoculated with PA103 intratracheally to induce severe sepsis. A subset of mice was treated with IGF-1 24 mg/kg subcutaneously either just before infection or 12 hours after the onset of infection. At 24 hours after infection, mice were killed according to animal care guidelines. Bacterial load was measured in blood drawn from the portal vein (PV) (solid bars) and the right ventricle (RV) (hatched bars). In severe sepsis, there is no difference between bacterial load in the PV and RV at 24 hours, suggesting impaired hepatic bacterial clearance. Treatment with IGF-1 either before or after the onset of infection resulted in a decrease in total bacterial load. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that animals that received IGF-1 had more bacteria in the PV compared with the RV at 24 hours (*P < 0.05), suggesting effective clearance of bacteria by the liver. Graph reflects mean and SD of the log transformation of bacterial load of n > 4 animals in each group. (B) Serum alanine aminotransferase (ALT) was measured 24 hours after the onset of infection. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates a significant increase in serum ALT in mice with severe sepsis compared with control animals (*P < 0.001). There was a reduction in ALT in animals treated with IGF-1 at the time of infection or 12 hours after the onset of infection compared with severe sepsis alone (P < 0.01 and P < 0.05, respectively). Graph reflects mean and SD of n > 4 animals in each group. (C) Caspase-3/7 activity was measured in liver homogenates 24 hours after infection. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates an increase in caspase-3/7 activity compared with control (*P < 0.001). There was a reduction in caspase activity in animals treated with IGF-1 at the time of infection or 12 hours after the onset of infection compared with severe sepsis alone (P < 0.001). Graph reflects mean and SD of n > 4 animals in each group.

Figure 4.

IGF-1 improves Kupffer cell survival and function in vitro. (A) KC13-2 cells were treated with tumor necrosis factor (TNF)-α, IGF-1, or both under serum-free conditions. IGF-1 treatment was initiated 30 minutes before incubation with TNF-α. Cells were incubated for 6 hours and viability was measured by either propidium iodide staining (left panel) or ATP assay (right panel). ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that treatment with IGF-1 decreases TNF-α–induced KC13-2 cell death compared with TNF-α alone (*P < 0.01). Graph reflects the mean and SD of three separate experiments. (B) Western blot demonstrates that cells treated with both TNF-α and IGF-1 have increased phospho-Akt, increased XIAP, and decreased cleaved caspase-3 compared with cells treated with TNF-α alone. β-Actin is included as a loading control. Figure is representative of three separate experiments. (C) Caspase activity was measured in KC13-2 cells after treatment with TNF-α, IGF-1, or both. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates a significant increase in caspase activity with TNF-α. This was decreased to levels consistent with serum-free control cells by pretreatment with IGF-1 (*P < 0.001) compared with cells treated with TNF-α alone. (D) KC13-2 cells were pretreated with either IGF-1 or saline 30 minutes before incubation with PA103, TNF-α, or both. After 6 hours, cells and supernatants were harvested and bacterial DNA was isolated. Quantitative real-time polymerase chain reaction with primers specific for P. aeruginosa was performed to determine residual bacterial load. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that pretreatment with IGF-1 before PA013 and TNF-α decreased the bacterial burden compared with cells treated with PA103 and TNF-α alone (*P < 0.05). Solid bars, control; hatched bars, IFG-1. ND = none detected. Graph reflects the mean and SD of the log transformation and is representative of three separate experiments. Control: serum-free media; Control (serum): media containing serum.

Figure 4.

IGF-1 improves Kupffer cell survival and function in vitro. (A) KC13-2 cells were treated with tumor necrosis factor (TNF)-α, IGF-1, or both under serum-free conditions. IGF-1 treatment was initiated 30 minutes before incubation with TNF-α. Cells were incubated for 6 hours and viability was measured by either propidium iodide staining (left panel) or ATP assay (right panel). ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that treatment with IGF-1 decreases TNF-α–induced KC13-2 cell death compared with TNF-α alone (*P < 0.01). Graph reflects the mean and SD of three separate experiments. (B) Western blot demonstrates that cells treated with both TNF-α and IGF-1 have increased phospho-Akt, increased XIAP, and decreased cleaved caspase-3 compared with cells treated with TNF-α alone. β-Actin is included as a loading control. Figure is representative of three separate experiments. (C) Caspase activity was measured in KC13-2 cells after treatment with TNF-α, IGF-1, or both. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates a significant increase in caspase activity with TNF-α. This was decreased to levels consistent with serum-free control cells by pretreatment with IGF-1 (*P < 0.001) compared with cells treated with TNF-α alone. (D) KC13-2 cells were pretreated with either IGF-1 or saline 30 minutes before incubation with PA103, TNF-α, or both. After 6 hours, cells and supernatants were harvested and bacterial DNA was isolated. Quantitative real-time polymerase chain reaction with primers specific for P. aeruginosa was performed to determine residual bacterial load. ANOVA followed by Bonferroni's test for multiple comparisons demonstrates that pretreatment with IGF-1 before PA013 and TNF-α decreased the bacterial burden compared with cells treated with PA103 and TNF-α alone (*P < 0.05). Solid bars, control; hatched bars, IFG-1. ND = none detected. Graph reflects the mean and SD of the log transformation and is representative of three separate experiments. Control: serum-free media; Control (serum): media containing serum.

Figure 5.

Treatment with IGF-1 improves survival in a murine model of sepsis. C57BL/6 mice were inoculated with PA103 intratracheally to induce severe sepsis. A subset of mice was treated with IGF-1 24 mg/kg subcutaneously either just before infection or 12 hours after the onset of infection. Temperature was used as a surrogate endpoint for mortality. Treatment with IGF-1 just before infection with the severe sepsis model resulted in decreased mortality compared with severe sepsis alone out to 48 hours by log-rank test (*P = 0.036). Treatment with IGF-1 12 hours after the onset of infection with the severe sepsis model resulted in decreased mortality compared with severe sepsis alone out to 48 hours by log-rank test (**P = 0.044). Solid squares, severe sepsis; solid triangles, IGF-1 at time 0; open inverted triangles, IFG-1 at 12 hours.