Immunosuppression in acutely decompensated cirrhosis is mediated by prostaglandin E2

Journal name:
Nature Medicine
Volume:
20,
Pages:
518–523
Year published:
DOI:
doi:10.1038/nm.3516
Received
Accepted
Published online

Liver disease is one of the leading causes of death worldwide1. Patients with cirrhosis display an increased predisposition to and mortality from infection due to multimodal defects in the innate immune system2, 3, 4; however, the causative mechanism has remained elusive. We present evidence that the cyclooxygenase (COX)-derived eicosanoid prostaglandin E2 (PGE2) drives cirrhosis-associated immunosuppression. We observed elevated circulating concentrations (more than seven times as high as in healthy volunteers) of PGE2 in patients with acute decompensation of cirrhosis. Plasma from these and patients with end-stage liver disease (ESLD) suppressed macrophage proinflammatory cytokine secretion and bacterial killing in vitro in a PGE2-dependent manner via the prostanoid type E receptor-2 (EP2), effects not seen with plasma from patients with stable cirrhosis (Child-Pugh score grade A). Albumin, which reduces PGE2 bioavailability, was decreased in the serum of patients with acute decompensation or ESLD (<30 mg/dl) and appears to have a role in modulating PGE2-mediated immune dysfunction. In vivo administration of human albumin solution to these patients significantly improved the plasma-induced impairment of macrophage proinflammatory cytokine production in vitro. Two mouse models of liver injury (bile duct ligation and carbon tetrachloride) also exhibited elevated PGE2, reduced circulating albumin concentrations and EP2-mediated immunosuppression. Treatment with COX inhibitors or albumin restored immune competence and survival following infection with group B Streptococcus. Taken together, human albumin solution infusions may be used to reduce circulating PGE2 levels, attenuating immune suppression and reducing the risk of infection in patients with acutely decompensated cirrhosis or ESLD.

At a glance

Figures

  1. Elevated PGE2 in plasma of patients admitted to hospital with acute decompensation is immunosuppressive.
    Figure 1: Elevated PGE2 in plasma of patients admitted to hospital with acute decompensation is immunosuppressive.

    (a) Left, ESI/LC-MS/MS analysis of PGE2 in plasma of patients with acutely decompensated cirrhosis (AD; obtained on day 1 or 2 of hospital admission) and healthy volunteers (HV). Right, original trace shown for abundance of PGE2 in AD plasma with internal standard (top), HV plasma (middle) and AD plasma (bottom). (b) TNF-α release from human MDMs stimulated with LPS in the presence of PGF2α (1 ng/ml), PGE2 (0.1 ng/ml), 5-HETE (0.5 ng/ml), 12-HETE (1.5 ng/ml) or 15-HETE (1 ng/ml) (experiments on cells from n = 7 healthy volunteers in duplicate). ('Cells' refers to macrophages untreated with LPS and/~Figure 1Elevated PGE2 in plasma of patients admitted to hospital with acute decompensation is immunosuppressive. (a) Left, ESI/LC-MS/MS analysis of PGE2 in plasma of patients with acutely decompensated cirrhosis (AD; obtained on day 1 or 2 of hospital admission) and healthy volunteers (HV). Right, original trace shown for abundance of PGE2 in AD plasma with internal standard (top), HV plasma (middle) and AD plasma (bottom). (b) TNF-α release from human MDMs stimulated with LPS in the presence of PGF2α (1 ng/ml), PGE2 (0.1 ng/ml), 5-HETE (0.5 ng/ml), 12-HETE (1.5 ng/ml) or 15-HETE (1 ng/ml) (experiments on cells from n = 7 healthy volunteers in duplicate). ('Cells' refers to macrophages untreated with LPS and/or plasma.) (c) TNF-α release from human MDMs stimulated with LPS in the presence of plasma from patients with AD (day 1 or 2 of admission) and healthy volunteers with or without the PGE2 EP1–3 antagonist AH6809 (50 μM or 300 μM) (plasma from n = 35 patients used, with experiments done in duplicate). (d) Human MDM bacterial killing in the presence of plasma from patients with AD (day 1 or 2 of admission) or healthy volunteers with or without AH6809 (50 μM) (n = 8 HV, n = 9 HV + AH6809, n = 9 AD and n = 9 AD + AH6809). (e,f) TNF-α (e) and IL-10 (f) release from human MDMs stimulated with LPS in the presence of plasma from patients with AD (day 2–6 (D2–D6) of admission) and healthy volunteers with or without AH6809 (AH, 50 μM) (for AD plasma, n = 13 patients on day 2, n = 6 on day 3, n = 4 on day 4 and n = 3 on days 5 and 6). Data are represented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, analysis of variance (ANOVA).

  2. Plasma from patients with ESLD but not from those with stable cirrhosis demonstrates PGE2-mediated immunosuppression.
    Figure 2: Plasma from patients with ESLD but not from those with stable cirrhosis demonstrates PGE2-mediated immunosuppression.

    (a) TNF-α release from human MDMs stimulated with LPS in the presence of plasma from patients awaiting liver transplant (ESLD), sampled 24 weeks apart, and from healthy volunteers with and without AH6809 (n = 13 patients). ('Cells' refers to macrophages untreated with LPS and/or plasma.) (b) TNF-α release from human MDMs stimulated with LPS in the presence of plasma from outpatients with stable cirrhosis (Child-Pugh score grade A, n = 16) or noncirrhotic liver disease (n = 5) and healthy volunteers (HV) with and without AH6809 (50 μm). (c) rtPCR of COX-2 and COX-1 expression in PBMCs isolated from patients with AD and healthy volunteers (n = 5 per group). AU, arbitrary units normalized according to the housekeeping gene RPS20, which encodes ribosomal protein 20. (d) Plasma PGE2 concentrations in naive (healthy) mice, sham mice and mice with BDL- or CCL4-induced liver injury (n = 3 naive; n = 4 sham (for BDL) and n = 7 BDL; n = 3 sham (for CCL4) and n = 7 CCL4). (e,f) Immunohistochemical staining for COX-2 in liver Kupffer cells (e) and alveolar macrophages (f). Scale bars, 50 μm. Data are represented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA.

  3. Inhibiting PGE2 restores bacterial killing and survival following bacterial infection in mouse models of liver injury.
    Figure 3: Inhibiting PGE2 restores bacterial killing and survival following bacterial infection in mouse models of liver injury.

    (a) TNF-α release from peritoneal macrophages from naive mice stimulated with LPS in the presence of plasma from naive mice or BDL mice with or without administration of indomethacin, SKF525A, baicalein or L-NAME before blood sampling (n = 11 cells alone and n = 13 cells + LPS; n = 9 naive mice plasma; n = 8 BDL mice plasma; n = 4 plasma from indomethacin-treated BDLs; n = 6 plasma from SKF525A-treated BDLs; n = 6 plasma from L-NAME–treated BDLs; n = 4 plasma from baicalein-treated BDLs). (b) TNF-α release from peritoneal macrophages from naive mice stimulated with LPS in the presence of plasma from naive mice or CCL4 mice with or without indomethacin before blood sampling or AH6809 (300 μM) added in vitro to cell culture (n = 4 naive mice plasma; n = 4 plasma from indomethacin-treated mice; n = 10 naive mice plasma with AH6809 added in vitro; n = 11 CCL4 plasma; n = 9 plasma from CCL4 mice treated with indomethacin; n = 13 CCL4 plasma with AH6809 added in vitro). (c,d) Blood bacterial counts 3 h following either i.p. (c) or i.v. (d) GBS administration to sham or BDL mice (with or without indomethacin pretreatment) (i.p. GBS n = 21 sham, n = 16 BDL and n = 16 BDL + indomethacin; i.v. GBS n = 6 sham, n = 8 BDL and n = 11 BDL + indomethacin). CFU, colony-forming units. (e) Kaplan-Meier survival curves of sham and BDL mice (with or without indomethacin pretreatment) following i.p. GBS administration. (f,g) H&E stains of liver from CCL4 mice treated with (f) or without (g) celecoxib for 5 d. Scale bars, 50 μm. Data are represented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA.

  4. PGE2-mediated immunosuppression by AD plasma is improved by albumin.
    Figure 4: PGE2-mediated immunosuppression by AD plasma is improved by albumin.

    (a) Correlation between LPS-stimulated human MDM TNF-α synthesis in the presence of AD plasma and that plasma samples' PGE2 concentration divided by its albumin concentration. (b) LPS-stimulated MDM TNF-α synthesis in the presence of AD plasma samples compared with that in the same AD plasma samples with >99% purified human serum albumin added in vitro to restore mean levels to 40 mg/dl (n = 19 patients, experiments in duplicate). (c) LPS-stimulated MDM TNF-α synthesis in the presence of increasing concentrations of PGE2 with or without 40 mg/dl of >99% purified human serum albumin. (d) LPS-stimulated MDM TNF-α synthesis in the presence of AD plasma samples from the most immunosuppressed (ADMIS) and the least immunosuppressed (ADLIS) patients compared with the same AD plasma samples with >99% purified human serum albumin added in vitro to restore mean levels to 40 mg/dl. (e) 3 h blood bacterial counts from BDL mice given 0.5 ml of 20% human albumin solution (HAS) subcutaneously (s.c.) compared to normal saline (0.9%) before i.p. GBS administration n = 7 mice per group saline; n = 10 mice per group albumin). (f) PGE2 plasma concentrations 3 h after i.p. GBS administration from BDL mice given either 20% HAS or normal saline before bacterial administration (n = 5). (g) LPS-stimulated human MDM TNF-α synthesis in the presence of AD plasma samples from patients before (AD pre albumin) and after (AD post albumin) infusion of 20% HAS (median 200 ml) (n = 6) or plasma samples taken 1 d apart in patients not treated with 20% HAS (AD control). (h) LPS-stimulated human MDM TNF-α synthesis in the presence of healthy volunteer or AD plasma samples in patients administered 20% HAS on days 1 (1.5 g/kg) and 3 (1 g/kg) after admission with samples obtained on day 1 of admission and at follow-up, with one sample taken between days 8 and 10 and a second between days 30 and 60 after admission (n = 10 at day 1 and n = 8 at days 8–10 and 30–60, as two lost to follow-up; n = 13 HV). Data are represented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA.

References

  1. Lim, Y.S. & Kim, W.R. The global impact of hepatic fibrosis and end-stage liver disease. Clin. Liver Dis. 12, 733746 (2008).
  2. Fierer, J. & Finley, F. Deficient serum bactericidal activity against Escherichia coli in patients with cirrhosis of the liver. J. Clin. Invest. 63, 912921 (1979).
  3. Fernández, J. et al. Bacterial infections in cirrhosis: epidemiological changes with invasive procedures and norfloxacin prophylaxis. Hepatology 35, 140148 (2002).
  4. Moreau, R. et al. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology 144, 14261437 (2013).
  5. Fernández, J. et al. Prevalence and risk factors of infections by multiresistant bacteria in cirrhosis: a prospective study. Hepatology 55, 15511561 (2012).
  6. O'Brien, A.J., Welch, C.A., Singer, M. & Harrison, D.A. Prevalence and outcome of cirrhosis patients admitted to UK intensive care: a comparison against dialysis-dependent chronic renal failure patients. Intensive Care Med. 38, 9911000 (2012).
  7. Hassner, A. et al. Impaired monocyte function in liver cirrhosis. Br. Med. J. (Clin. Res. Ed.) 282, 12621263 (1981).
  8. Rajkovic, I.A. & Williams, R. Abnormalities of neutrophil phagocytosis, intracellular killing and metabolic activity in alcoholic cirrhosis and hepatitis. Hepatology 6, 252262 (1986).
  9. Shawcross, D.L. et al. Ammonia impairs neutrophil phagocytic function in liver disease. Hepatology 48, 12021212 (2008).
  10. Fagiuoli, S. et al. Management of infections in cirrhotic patients: report of a consensus conference. Dig. Liver Dis. 46, 204212 (2014).
  11. Scher, J.U. & Pillinger, M.H. The anti-inflammatory effects of prostaglandins. J. Investig. Med. 57, 703708 (2009).
  12. Kalinski, P. Regulation of immune responses by prostaglandin E2. J. Immunol. 188, 2128 (2012).
  13. Fullerton, J.N., O'Brien, A.J. & Gilroy, D.W. Pathways mediating resolution of inflammation: when enough is too much. J. Pathol. 231, 820 (2013).
  14. Fullerton, J.N., O'Brien, A.J. & Gilroy, D.W. Lipid mediators in immune dysfunction after severe inflammation. Trends Immunol. 35, 1221 (2014).
  15. Serezani, C.H. et al. Prostaglandin E2 suppresses bacterial killing in alveolar macrophages by inhibiting NADPH oxidase. Am. J. Respir. Cell Mol. Biol. 37, 562570 (2007).
  16. Bourdonnay, E., Serezani, C.H., Aronoff, D.M. & Peters-Golden, M. Regulation of alveolar macrophage p40phox: hierarchy of activating kinases and their inhibition by PGE2. J. Leukoc. Biol. 92, 219231 (2012).
  17. Aronoff, D.M., Canetti, C. & Peters-Golden, M. Prostaglandin E2 inhibits alveolar macrophage phagocytosis through an E-prostanoid 2 receptor-mediated increase in intracellular cyclic AMP. J. Immunol. 173, 559565 (2004).
  18. Medeiros, A.I., Serezani, C.H., Lee, S.P. & Peters-Golden, M. Efferocytosis impairs pulmonary macrophage and lung antibacterial function via PGE2/EP2 signaling. J. Exp. Med. 206, 6168 (2009).
  19. Bozyk, P.D. & Moore, B.B. Prostaglandin E2 and the pathogenesis of pulmonary fibrosis. Am. J. Respir. Cell Mol. Biol. 45, 445452 (2011).
  20. Yang, J., Petersen, C.E., Ha, C.E. & Bhagavan, N.V. Structural insights into human serum albumin-mediated prostaglandin catalysis. Protein Sci. 11, 538545 (2002).
  21. Wasmuth, H.E. et al. Patients with acute on chronic liver failure display “sepsis-like” immune paralysis. J. Hepatol. 42, 195201 (2005).
  22. Goss, J.A., Mangino, M.J. & Flye, M.W. Prostaglandin E2 production during hepatic regeneration downregulates Kupffer cell IL-6 production. Ann. Surg. 215, 553559, discussion 559–560 (1992).
  23. McAnulty, R.J., Hernandez-Rodriguez, N.A., Mutsaers, S.E., Coker, R.K. & Laurent, G.J. Indomethacin suppresses the anti-proliferative effects of transforming growth factor-beta isoforms on fibroblast cell cultures. Biochem. J. 321, 639643 (1997).
  24. Runyon, B.A. Introduction to the revised American Association for the Study of Liver Diseases Practice Guideline management of adult patients with ascites due to cirrhosis 2012. Hepatology 57, 16511653 (2013).
  25. Alves de Mattos, A. Current indications for the use of albumin in the treatment of cirrhosis. Ann. Hepatol. 10 (suppl. 1), S15S20 (2011).
  26. Mookerjee, R.P. et al. Neutrophil dysfunction in alcoholic hepatitis superimposed on cirrhosis is reversible and predicts the outcome. Hepatology 46, 831840 (2007).
  27. Garcia-Martinez, R. et al. Albumin: pathophysiologic basis of its role in the treatment of cirrhosis and its complications. Hepatology 58, 18361846 (2013).
  28. Romanelli, R.G. et al. Long-term albumin infusion improves survival in patients with cirrhosis and ascites: an unblinded randomized trial. World J. Gastroenterol. 12, 14031407 (2006).
  29. Georgiev, P. et al. Characterization of time-related changes after experimental bile duct ligation. Br. J. Surg. 95, 646656 (2008).
  30. Domenicali, M. et al. A novel model of CCl4-induced cirrhosis with ascites in the mouse. J. Hepatol. 51, 991999 (2009).

Download references

Author information

Affiliations

  1. Centre for Clinical Pharmacology and Therapeutics, Division of Medicine, University College London, London, UK.

    • Alastair J O'Brien,
    • James N Fullerton,
    • Grace Auld,
    • Sarah James,
    • Justine Newson,
    • Effie Karra &
    • Derek W Gilroy
  2. Manchester Pharmacy School, Faculty of Medical and Human Sciences, University of Manchester, Manchester, UK.

    • Karen A Massey &
    • Anna Nicolaou
  3. Division of Medicine, University College London, London, UK.

    • Gavin Sewell
  4. Department of Histopathology, University College London Hospitals, London, UK.

    • Alison Winstanley
  5. Liver Unit, Centre for Digestive Diseases, Blizard Institute of Cell and Molecular Science, Queen Mary University of London, London, UK.

    • William Alazawi
  6. Hospital Clínic de Barcelona, Servicio de Hepatología, Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain.

    • Rita Garcia-Martinez
  7. Centro de Investigacion Biomédica en Red de Enfermedades Hepaticas y Digestivas, Instituto de Salud Carlos III, Madrid, Spain.

    • Joan Cordoba

Contributions

D.W.G. and A.J.O. conceived of the idea, and A.J.O. carried out the work. D.W.G. and A.J.O. cowrote the paper, and J.N.F. edited. J.N.F., G.S., J.N., S.J., E.K. and G.A. carried out biochemical assays, and W.A. (Royal London Hospitals), J.C. and R.G.-M. (both from ALFAE and MACHT clinical trials) supplied clinical samples. K.A.M. and A.N. carried out ESI/LC-MS/MS analysis, and A.W. carried out histological analysis.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Author details

Supplementary information

PDF files

  1. Supplementary Text and Figures (828 KB)

    Supplementary Figures 1–5 and Supplementary Tables 1–4

Additional data