Beneficial role of G-CSF in Acute-on-Chronic Liver Failure: effects on liver regeneration, inflammation/immunoparalysis or both?


  • Thierry Gustot

    1. Department of Gastroenterology and Hepato-Pancreatology, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium
    2. Laboratory of Experimental Gastroenterology, Université Libre de Bruxelles, Brussels, Belgium
    Search for more papers by this author

In the current issue of Liver International, Khanam and colleagues try to better characterize mechanisms of action of granulocyte colony stimulating factor (G-CSF) explaining its potential beneficial effect in a severe form of decompensation of cirrhosis requiring frequently intensive care unit (ICU) admission, called Acute-on-Chronic Liver Failure (ACLF). These investigational results came from analysis of samples of patients of a randomized controlled trial (RCT) comparing G-CSF and placebo in 47 patients with ACLF published in Gastroenterology in 2012 [1].

There is increasing interest in ACLF. ICU management of critically ill cirrhotic patients becomes an essential part of Hepatology. The estimated number of ICU admissions for cirrhosis in United States is in excess of 26 000 per year with a cost of $3 billion [2]. Unfortunately at this time, no universal definition for ACLF is accepted. Based on consensus opinion of experts of the Asian Pacific Association for the Study of the Liver (APASL), ACLF has been defined as an acute hepatic insult manifesting as jaundice (total bilirubin ≥5 mg/dl) and coagulopathy (international normalized ratio ≥1.5), complicated within 4 weeks by ascites and/or encephalopathy in a patient with chronic liver disease [3]. This definition has been developed on a theoretical rather than experimental basis. When APASL criteria were met, the overall 30-day mortality was 25% [4]. Interestingly, organ failure defined by the classical SOFA score contributed highly to 30-day mortality rate of ACLF (58% in the organ failure group compared to 8% in patients without). In a recent large multicenter European study (CANONIC), diagnostic criteria for ACLF were established on the presence of organ failures defined by the CLIF-SOFA score and high 28-day mortality rate [5]. Based on these criteria, patients with ACLF had a 28-day and 90-day mortality of 34% and 51% respectively. These data suggest a reversibility of this syndrome in about one half of the cases. In the present study, Khanam and colleagues used APASL criteria to define their ACLF patients.

The precipitating events of this syndrome vary widely among countries. In Asian countries, Hepatitis B virus (HBV) reactivation is a frequent trigger of ACLF. On the other hand, in Western countries, bacterial infections and alcoholic hepatitis (AH) are the most identifiable precipitating event but in 40%, no precipitating event is discovered. Patients with severe AH defined by a modified discriminant function higher than 32 and not treated by corticosteroids has a 28-day mortality rate of 34% and is consistent with the outcome of ACLF patients [6]. In the current study, the most common aetiology of ACLF was AH (57%) followed by HBV (21%).

Hepatocyte regeneration is crucial to restore liver function in ACLF but is impaired in cirrhosis [7]. After liver injury, bone marrow-derived circulating pluripotent cells contribute to hepatocyte regeneration by providing cytokines and growth factors that promote liver repair more than being precursors of new hepatocytes [8]. In animal models of acute liver injury, the administration of G-CSF induced mobilization of bone marrow cells and improved liver repair and survival [9]. In patients with AH, subcutaneous injections of G-CSF (10 μg/kg daily for 5 days) induced a mobilization of bone marrow CD34+ cells, expression of hepatocyte growth factor and proliferation of hepatic progenitor cells (HPC) [10]. In a larger population of decompensated alcoholic cirrhosis (81% with AH), the combination of G-CSF injections and bone marrow mononuclear cell transplantation (bone marrow aspiration at day 5 with mononuclear cell re-infusion in hepatic artery) was not better than placebo in terms of survival, proliferation of HPC or improvement of liver function [11]. On the other hand, based on the present RCT, a longer protocol of G-CSF administration (5 μg/kg daily for 5 days and then every 3 days during 1 month) significantly improved 60-day survival, liver function and prevented sepsis, hepatorenal syndrome and hepatic encephalopathy [1]. These discrepancies might be also explained by differences in ACLF aetiologies and in applied standard medical management (corticosteroids vs. pentoxifylline for AH).

A common feature of ACLF is an intensive systemic inflammatory response characterized by high leukocyte count and plasma C-reactive protein level and correlated with mortality [5]. Triggers of this inflammation are largely unknown. Bacterial infection might explain a part of it, but the association between severity of ACLF and systemic inflammation is also observed in non-infected ACLF patients. Infection might be an under-recognized cause or inflammation might be ‘sterile’ and drive by mediators released by injured liver (danger-associated molecular patterns).

Beside this pro-inflammatory profile, patients with ACLF display a ‘sepsis-like’ immune paralysis characterized by a down-regulation of HLA-DR expression on circulating monocytes and decreased secretion of pro-inflammatory cytokines (TNF-α) after lipopolysaccharide stimulation [12]. This feature seems to be conserved among the different aetiologies of ACLF. Neutrophil dysfunction is also observed in patients with an alcoholic hepatitis superimposed on cirrhosis [13]. Recently, an adaptative immune dysfunction with an increase of proportion of regulator T cells compared to conventional CD4+ T cells is described in HBV-related ACLF [14]. Other key players of host response against microbes are myeloid and plasmacytoid dendritic cells (mDCs and pDCs). A decrease of functional mDCs is observed in HBV-related ACLF and is associated with poor outcome [15]. This multiple immune defect could contribute to the increased risk of infection in this subgroup of patients.

In the present study of Khanam and colleagues, G-CSF administration increased circulating and intrahepatic mDCs, pDCs and T cells associated with a reduction of IFN-γ-producing CD8 T cells. Interestingly, it seems that survival and improvement of liver function was associated with this increase. The impact of these increases on liver injury is largely unknown. The enhanced DC population in the liver could affect the cytokine/chemokine environment and improve liver regeneration. Some experiments suggest that increased DC number could accelerate liver restoration through regulating immunity [16]. In tissue injury, resident macrophages and DCs avoid excessive immunopathology and orchestrate repair by becoming growth factor-producing cells [17] [18].

Another potential mechanism of action of G-CSF is the prevention of sepsis by acting on immune dysfunction of ACLF. Beside the prevention of this serious complication, the maintenance of a ‘sterile’ liver injury could facilitate the healing process and liver function recovery. This phenomenon is well known for wound healing. Indeed, sterile wound care is a validated therapeutic strategy to limit the inflammatory response and to enforce healing [19].

We are at the start of understanding of ACLF pathophysiology. As a result of discrepancies of RCT results, we cannot give a generalized recommendation about G-CSF administration in ACLF but mechanistic explorations of pilot RCT in this specific clinical situation are essential tools to discover new hidden sides of liver injury, hepatocyte regeneration and immune dysfunction and potentially new more specific therapeutic targets.

“If we value the pursuit of knowledge, we must be free to follow wherever that search may lead us.” Adlai E. Stevenson Jr.