Defective monocyte oxidative burst predicts infection in alcoholic hepatitis and is associated with reduced expression of NADPH oxidase

Abstract

Objective: In order to explain the increased susceptibility to serious infection in alcoholic hepatitis, we evaluated monocyte phagocytosis, aberrations of associated signalling pathways and their reversibility, and whether phagocytic defects could predict subsequent infection.

Design: Monocytes were identified from blood samples of 42 patients with severe alcoholic hepatitis using monoclonal antibody to CD14. Phagocytosis and monocyte oxidative burst (MOB) were measured ex vivo using flow cytometry, luminometry and bacterial killing assays. Defects were related to the subsequent development of infection. Intracellular signalling pathways were investigated using western blotting and PCR. Interferon-γ (IFN-γ) was evaluated for its therapeutic potential in reversing phagocytic defects. Paired longitudinal samples were used to evaluate the effect of in vivo prednisolone therapy.

Results: MOB, production of superoxide and bacterial killing in response to Escherichia coli were markedly impaired in patients with alcoholic hepatitis. Pretreatment MOB predicted development of infection within two weeks with sensitivity and specificity that were superior to available clinical markers. Accordingly, defective MOB was associated with death at 28 and 90 days. Expression of the gp91 ^phox subunit of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase was reduced in patients with alcoholic hepatitis demonstrating defective MOB. Monocytes were refractory to IFN-γ stimulation and showed high levels of a negative regulator of cytokine signalling, suppressor of cytokine signalling-1. MOB was unaffected by 7 days in vivo prednisolone therapy.

Conclusions: Monocyte oxidative burst and bacterial killing is impaired in alcoholic hepatitis while bacterial uptake by phagocytosis is preserved. Defective MOB is associated with reduced expression of NADPH oxidase in these patients and predicts the development of infection and death.

Keywords: ALCOHOLIC LIVER DISEASE; BACTERIAL INFECTION; IMMUNOLOGY IN HEPATOLOGY.

Figure 1

Uptake of bacteria by phagocytosis was similar between severe alcoholic hepatitis (SAH) and healthy monocytes. (A–C) Scavenger receptor expression was equivalent on SAH monocytes. (D) FcγR expression was increased in SAH compared with healthy control (HC) and CLD; (E) phagocytosis is preserved in SAH monocytes.

Figure 2

Monocyte oxidative burst (MOB) and bacterial killing is impaired in patients with severe alcoholic hepatitis (SAH). (A) The production of reactive oxygen species (ROS) at rest was similar between SAH and healthy control (HC) monocytes; (B) MOB in response to Escherichia coli is impaired in SAH compared with HC and CLD; (C) impaired MOB is also seen in cirrhotic patients who were actively drinking at the time of sampling compared with abstinent cirrhotic patients; (D) impaired MOB corresponded to a reduction in the production of superoxide radicals in response to E. coli from SAH monocytes; (E) killing of phagocytosed bacteria is reduced in SAH compared with HC monocytes; (F) far more E. coli were enumerated from the lysate (intracellular fraction) of SAH monocytes compared with the respective supernatant (extracellular fraction).

Figure 3

Impaired monocyte oxidative burst (MOB) predicts the subsequent development of infection within 2 weeks. (A) Prior prescription of intravenous antibiotics did not affect severe alcoholic hepatitis (SAH) MOB; (B) patients with SAH who developed infection within 2 weeks of sampling had a lower pretreatment MOB compared with patients who did not develop infection; (C) patients who developed infection within 90 days of sampling had a lower pretreatment MOB than patients who did not develop infection.

Figure 4

Monocyte oxidative burst (MOB) <50th centile (MOB defect) can identify patients likely to develop infection more accurately than conventional clinical markers of infection. (A) Cut-off values at the 50th and 25th centiles of severe alcoholic hepatitis (SAH) MOB and their association with the development of infection within the subsequent 2 weeks; (B) OR of MOB <50th centile (MOB defect) for predicting infection within 2 weeks; (C) receiver operating characteristic curve (ROC) curve for MOB and the development of infection within the subsequent 2 weeks; area under ROC (AUROC) is 0.86 (p<0.0001). For comparison, the ROC curves for C-reactive protein (CRP), white cell count (WCC) and procalcitonin in predicting the subsequent development of infection are included (AUROC 0.73, 0.75 and 0.72; p=0.019, 0.007 and 0.02, respectively).

Figure 5

Patients with severe alcoholic hepatitis (SAH) with monocyte oxidative burst (MOB) defect are more susceptible to catalase-positive organisms: (A) organisms grown by culture from patients with SAH recruited to the study; (B) catalase status of organisms grown; (C) susceptibility of patients with SAH with MOB defect to catalase-positive organisms compared with patients with SAH without MOB defect.

Figure 6

Ex vivo monocyte oxidative burst (MOB) is not affected by patient therapy with prednisolone. (A). Evolution of MOB between days 0 and 7 without prednisolone therapy; (B) evolution of MOB between days 0 and 7 with prednisolone therapy.

Figure 7

Reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase is associated with monocyte oxidative burst (MOB) defect and aberrant Janus Kinase (JAK)-signal transducer and activator of transcription (STAT) signalling with resistance to exogenous interferon (IFN)-γ. (A) Patients with severe alcoholic hepatitis (SAH) with MOB defect (SAH+MOB) have diminished levels of the gp91^phox subunit of the NADPH oxidase complex compared with patients with SAH without MOB defect (SAH-MOB); (B) reduced gene expression of gp91^phox after IFN-γ stimulation in SAH+MOB, but not SAH-MOB, monocytes; (C) reduced gene expression of STAT-1 after IFN-γ stimulation in all SAH monocytes; (D) resting levels of phosphorylated STAT-1 are reduced in all SAH monocytes with or without MOB defect; (E) impaired activation of STAT-1 in response to IFN-γ in all SAH monocytes; (F) suppressor of cytokine signalling-1 (SOCS-1) protein is present at high levels in all SAH monocytes compared with healthy control (HC); (G) exogenous IFN-γ stimulation does not improve MOB in vitro.