Secretory leukocyte protease inhibitor: A pivotal mediator of anti-inflammatory responses in acetaminophen-induced acute liver failure


  • Potential conflict of interest: Nothing to report.

  • Supported by the Medical Research Council (MRC) Clinician Scientist Award (to CGA); EASL Sheila Sherlock Physican Scientist Award (to CGA); Rosetrees Charitable Trust.


Acetaminophen-induced acute liver failure (AALF) is characterized both by activation of innate immune responses and susceptibility to sepsis. Circulating monocytes and hepatic macrophages are central mediators of inflammatory responses and tissue repair processes during human AALF. Secretory leukocyte protease inhibitor (SLPI) modulates monocyte/macrophage function through inhibition of nuclear factor kappa B (NF-κB) signaling. The aims of this study were to establish the role of SLPI in AALF. Circulating levels of SLPI, monocyte cluster of differentiation 163 (CD163), human leukocyte antigen-DR (HLA-DR), and lipopolysaccharide (LPS)-stimulated levels of NF-κBp65, tumor necrosis factor alpha (TNF-α) and interleukin (IL)-6 were determined in patients with AALF, chronic liver disease, and healthy controls. Immunohistochemistry and multispectral imaging of AALF explant tissue determined the cellular sources of SLPI and hepatic macrophage phenotype. The phenotype and function of monocytes and macrophages was determined following culture with recombinant human (rh)-SLPI, liver homogenates, and plasma derived from AALF patients in the presence and absence of antihuman (α)SLPI. Hepatic and circulatory concentrations of SLPI were elevated in AALF and immunohistochemistry revealed SLPI expression in biliary epithelial cells and within hepatic macrophages (h-mψ) in areas of necrosis. H-mψ and circulating monocytes in AALF exhibited an anti-inflammatory phenotype and functional characteristics; typified by reductions in NF-κBp65, TNF-α, and IL-6 and preserved IL-10 secretion following LPS challenge. Culture of healthy monocytes with AALF liver homogenates, plasma, or rhSLPI induced monocytes with strikingly similar anti-inflammatory characteristics which were reversed by inhibiting the activity of SLPI. Conclusion: SLPI is a pivotal mediator of anti-inflammatory responses in AALF through modulation of monocyte/macrophage function, which may account for the susceptibility to sepsis in AALF. (Hepatology 2014;59:1564-1576)


acetaminophen-induced acute liver failure


cluster of differentiation 163


human leukocyte antigen-DR


hepatic macrophages






nonacetaminophen-induced acute liver failure


secretory leukocyte protease inhibitor


transforming growth factor-beta1


tumor necrosis factor-alpha

Acute liver failure (ALF) is characterized by overwhelming hepatocyte death and activation of innate immune responses.[1, 2] Drug-induced liver injury and in particular acetaminophen-induced ALF (AALF) is responsible for most cases. Secondary infection is a common complication and is a major contributor to mortality.[3, 4] It has been postulated that ALF leads to immunoparesis due to impairment of innate immune responses.

Secretory leukocyte protease inhibitor (SLPI) is a multifaceted protein, produced by epithelial and immune cells, that plays an important role in tissue inflammation.[5] Although initially coined for its antiprotease activity, its importance in the resolution of inflammation and tissue repair processes has been recently highlighted through its ability to stimulate epithelial cell proliferation[6-8] and modulation of macrophage function. Following injury, SLPI is a key anti-inflammatory mediator in the tissue microenvironment that directly inhibits macrophage nuclear factor kappa B (NF-κB)-dependent proinflammatory responses.[9-11] Furthermore, macrophages secrete SLPI following phagocytosis of apoptotic cells and in response to anti-inflammatory cues such as interleukin (IL)-6 and IL-10.9 More recently, SLPI has been shown to play an immunoregulatory role in human and experimental models of systemic sepsis, inhibiting both innate and adaptive immune responses to microbial stimuli.[12-14]

A marked increase in hepatic macrophages (h-mψ) is consistently observed in both experimental liver injury and human ALF. Rodent models demonstrate that h-mψ contribute to both propagation and resolution of acute liver injury.[15-18] We have recently published data implicating both circulating monocytes and h-mψ in resolution of inflammation and tissue repair processes during human AALF.[19, 20]

In view of the immunoregulatory role of SLPI during inflammation, we sought to determine whether SLPI could account for the phenotypic and functional changes in circulating monocytes that we have previously reported in human AALF, in order to establish whether it could account for the increased predisposition to sepsis encountered in these patients.

Patients and Methods

Patient Recruitment

Seventy-three AALF and 25 nonacetaminophen-induced ALF (NAALF) patients were recruited into the study within 24 hours following admission to the Liver Intensive Care Unit. AALF patients were stratified into patients who carried a poor prognosis (AALF-P [n = 40]: patients who died without undergoing orthotopic liver transplantation [OLT]; those who underwent OLT) and those who survived with medical management alone (AALF-S [n = 33]). Fifteen consecutive inpatients with chronic liver disease (CLD) and 24 healthy volunteers served as pathological and healthy controls (HC), respectively. Exclusion criteria were: age <18 or >65 years, neoplasia, immunosuppressive therapy. The presence of infection was defined as positive microbiological cultures, radiological evidence of pulmonary infiltrates, and clinical evidence of infection. AALF patients were identified for emergency transplantation according to the King's College Hospital (KCH) criteria.[21] The study was approved by the King's College Hospital Ethics Committee (LREC 04LG03). Assent was obtained by the patients' nominated next of kin if they were unable to give informed consent themselves.

Clinical, Hematological, and Biochemical Parameters

White cell count (monocyte, neutrophil, lymphocyte [count × 109/L]) was determined in AALF patients using a hematological analyzer (Siemens Advia 2120, Berks, UK). International normalized prothrombin ratio (INR), liver and renal function tests, arterial lactate, and clinical and physiological variables were prospectively entered into a database at the time of sampling.

Phenotyping of Ex Vivo Fresh Blood Monocyte and Measurement of Cytokine Responses to Lipopolysaccharide (LPS)

Flow cytometry was performed on fresh blood monocytes using monoclonal antibodies against cluster of differentiation (CD)14, CD16, human leukocyte antigen-DR (HLA-DR), and CD163 (BD Biosciences, Oxford, UK) as previously described.[20] Levels of TNF-α and IL-6 were determined in supernatants following a 6-hour incubation of peripheral blood mononuclear cells (PBMC) with 100 ng/mL LPS from Escherichia coli (serotype 0111:B; SBioscience, Nottingham, UK).


Ex vivo monocytes were stimulated with LPS and assessed for phosphorylation of NF-κBp65 using the phosphoflow technique (BD Phosflow Protocol for human whole blood samples; BD Biosciences) (Supporting Methods section 1.1).

Determination of SLPI Levels

Circulating SLPI levels were assessed in sera samples in: AALF (n = 73), NAALF (n = 25) patients within 24 hours of admission and 15 healthy controls (HC). In 35 AALF patients SLPI levels were measured within 24 hours (day 1) and 72 hours (day 3) following admission. Regional concentrations of SLPI (hepatic vein [HV], portal vein [PV]) were determined in six AALF patients at the time of OLT. Liver tissue homogenate obtained (as previously described[19]) from AALF (n = 7) and pathological control liver tissue (n = 8) was assessed for SLPI levels using enzyme-linked immunosorbent assay (ELISA (R&D Systems, Abingdon, UK).

Liver Samples and Immunohistochemistry

Explant liver tissue was obtained in 10 patients undergoing OLT due to AALF as previously described.[19] Liver tissue obtained during OLT for chronic liver disease (n = 5) and hepatic resection margins of colorectal malignancies (n = 5) served as pathological and normal control tissue, respectively. Tissue samples were taken for diagnostic histological examination and formalin-fixed, paraffin-embedded (FFPE) using a standard procedure. Snap-frozen liver sections were concomitantly obtained and stored in liquid nitrogen. Using immunohistochemistry on FFPE tissue, the numbers of SLPI and CD163-expressing cells were assessed. Hepatic macrophage phenotype (SLPI+/CD68+; SLPI/CD163+; CD68+/CD163+/HLA-DR+) and proliferating SLPI-expressing biliary epithelial cells (SLPI+/Ki67+) were studied using double and triple-staining immunohistochemistry (Supporting Methods 1.2-1.5).

Immunostains were analyzed by a liver histopathologist (A.Q.) who was blinded to the clinical outcome data. A cell count was carried out using an eyepiece graticule as described previously.[19] Colocalization and enumeration of SLPI and CD163+-expressing macrophages was performed using Nuance multispectral imaging technology (Supporting Methods 1.6).

Monocyte Isolation and Phenotyping

CD14+ cells were isolated from peripheral blood mononuclear cells (PBMCs) using CD14-specific magnetic beads (Miltenyi Biotec, Surrey, UK) for cell culture experiments. Following 48 hours in culture, monocytes were harvested and stained for cell surface markers CD14, CD16, CD36, CD163, CD206, and HLA-DR (eBioscience, Hatfield, UK) using flow cytometry using Flowlogic software (Invai Technologies) (Supporting Methods 1.7-1.9). Monocyte-derived mψ were generated as described in Supporting Methods 1.11.

Effect of SLPI on Monocyte Phenotype and Function

In order to examine the effect of SLPI on healthy monocytes and mψ, the cells were cultured in the presence of increasing concentrations of recombinant human (rh)-SLPI (R&D Systems, Abingdon, UK) (0, 0.1, and 0.5 μg/mL). Following 48 hours of culture, cells were harvested and analyzed for their surface phenotype and function (Supporting Methods 1.10).

Effect of Blocking SLPI in Liver Tissue Homogenates and Sera Samples on Monocyte Phenotype and Function

Healthy monocytes were cultured in complete media (CM) alone, CM plus AALF (ALF), or normal liver (NL) tissue homogenates.[22] In a separate series of experiments, monocytes and macrophages were incubated for 48 hours in 25% human sera samples derived from healthy donors (n = 5) or AALF (n = 20) patients (Supporting Methods 1.8). In selected experiments, tissue lysates and plasma samples were preincubated with increasing concentrations of antihuman SLPI neutralizing antibody (α-SLPI) (R&D Systems) (0, 1, 5 μg/mL) for 1 hour at room temperature before addition to the monocyte cultures (Supporting Methods 1.10).

Phagocytosis Assays

Following exposure to liver tissue homogenates and rhSLPI, the ability of monocytes to phagocytose E. coli was assessed using a modified flow cytometry-based phagotest protocol and phRodo assays (Supporting Methods 1.12).

ELISA and MSD Multiplex Cytokine Detection System

Collected supernatants were tested for cytokine secretion using ELISA (R&D Systems) and the Meso Scale Discovery (MSD) Human Th1/Th2 10 plex, Human Proinflammatory 4-Plex II, and transforming growth factor beta1 (TGF-β1) kit detection system (Gaithersburg, MD) (Supporting Methods 1.13).

Statistical Analysis

To identify differences between groups, nonparametric analysis was used (Mann-Whitney U, Kruskal-Wallis, Wilcoxon Rank tests). The results are expressed as median and interquartile range (IQR). All analyses were performed using GraphPad Prism 6 statistical software package (San Diego, CA).


Patient Characteristics

There was no significant difference in median ages of AALF patients when compared to HC, while CLD patients were significantly older. The number of circulating monocytes was significantly reduced in all AALF patients when compared to CLD patients and HC (P < 0.01). AALF patients had significantly higher biochemical and physiological indices of acute liver injury compared to CLD patients (Supporting Results, Table 1).

Circulating Monocytes Have Anti-inflammatory Characteristics in AALF

We assessed the frequency (Supporting Results 2.1) of the three major peripheral blood monocyte subsets (CD14+/CD16−, CD14++/CD16+, and CD14low/CD16+ cells) and their expression of HLA-DR and CD163, phenotypic markers reflecting monocyte anti-inflammatory responses to microbial challenge, in 40 AALF, 15 CLD patients, and 17 HC (within 24 hours of admission). In AALF patients, HLA-DR expression was reduced in all three monocyte subsets when compared to CLD and HC. In AALF patients, CD163 expression was elevated in both CD16+ monocyte subsets but reduced in CD16− monocytes when compared to CLD patients and HC (Fig. 1A-C).

Figure 1.

Phenotypic and functional characteristics of circulating monocytes in AALF patients. The expression of HLA-DR and CD163 was determined using flow cytometry and quantified as MFI in AALF patients (n = 40), compensated CLD (n = 15) patients, and normal HC (n = 17) in (A) CD14+CD16−, (B) CD14++CD16+, and (C) CD14lowCD16+ monocyte subsets. (D,E) NF-κBp65 expression in fresh blood using phosphoflow was determined in CD14+ monocytes at baseline (medium only, blue histogram) and following stimulation with 100 ng/mL LPS for 60 minutes (orange histogram). CD14+ monocytes were gated on the basis of CD14+ expression (neutrophils were excluded using CD62L expression). Representative flow cytometry profiles of NF-κBp65 expression in an HC (top row) and a patient with AALF (bottom row). (E) Levels of NF-κBp65 phosphoprotein were measured in HC (n = 10) and AALF (n = 10) patients following LPS stimulation (100 ng/mL). Results are expressed as a ratio of activation (MFI medium + LPS stimulated NF-κBp65/MFI medium alone NF-κBp65). (F) Levels of IL-6 and TNF-α (pg/mL) were measured following LPS stimulation (100 ng/mL) in PBMCs from AALF (n = 11) and HC (n = 10) after 6 hours culture period. SSC: side scatter, FCS: forward scatter.

Endotoxin Tolerant Monocytes in AALF

Following LPS challenge, levels of NF-κB signal transduction pathway phosphoprotein-NF-κBp65, and secretion of NF-κB-dependent proinflammatory mediators (TNF-α, IL-6) were determined in 11 AALF patients and 10 HC. At baseline, ex vivo monocyte NF-κBp65 was significantly elevated in AALF patients compared to HC (MFI: 339.5 [222.3-465.5] versus 197 [136.5-226.0]; P = 0.01). Following LPS stimulation, there was a significant reduction in the ratio of activation of NF-κBp65 (0.88 [0.70-0.99] versus 1.59 [1.5-1.99]; P = 0.008), TNF-α (465.5 pg/mL [252-1144] versus 2358 [1325-3324]; P = 0.003) and IL-6 (327 pg/mL [109-792] versus 2544 [1762-3406]; P = 0.006) secretion in AALF patients compared to HC (Fig. 1D-F).

Circulating SLPI Levels Are Elevated in ALF and Are Associated With Infection

On admission, median circulating levels of SLPI were markedly elevated in AALF and NAALF patients when compared to HC (Fig. 2A). No significant differences were detected between NAALF-P and NAALF-S patients (60964 [45849-79224] versus 51194 pg/mL [21927-62991]; P = 0.2).

Figure 2.

SLPI levels in AALF and NAALF. Circulating concentrations of SLPI (pg/mL) were determined by ELISA in: (A) plasma samples from AALF (n = 73) and NAALF (n = 25), within 24 hours of admission, and HC (n = 15); (B) sequential samples in 35 AALF patients, within 24 hours (day 1) and 72 hours (day 3) following admission compared to HC (n = 15); (C) liver tissue homogenates derived from explanted AALF tissue (n = 7) and pathological controls (n = 8); (D) Regional SLPI concentrations (portal vein [PV]; hepatic vein [HV]) at the time of OLT in six AALF patients. (E) Regional and systemic monocyte NF-κBp65 phosphoprotein, expression before (baseline) and after LPS stimulation (100 ng/mL) for 1 hour, in three AALF patients at the time of OLT. Results are expressed as MFI.

When subdivided according to clinical outcome, admission encephalopathy score, creatinine, INR, platelet count, arterial pH, and Model for Endstage Liver Disease (MELD) score, SLPI levels were significantly higher in AALF-P compared to AALF-S patients (Supporting Table 2). On sequential analysis, peak SLPI levels were detected at day 1 versus day 3 of admission. Compared to HC, AALF patients had significant elevations of circulating SLPI levels at both days 1 and 3 following admission (Fig. 2B).

AALF patients were also subdivided according to whether they developed infection (AALF-I) or remained infection-free (AALF-IF) during the first week following admission. While AALF-IF patients had a significant reduction in circulating SLPI levels at day 3 versus day 1 (72,421 [46,442-80,769] versus 58,586 pg/mL [41356-62784]; P = 0.03), levels in AALF-I remained unchanged (59,398 [46609-72434] versus 61470 pg/mL [40626-91611]; P = 0.8).

Elevated Circulating of SLPI Are Derived From the Inflamed Liver in AALF

As shown in Fig. 2C, whole liver tissue levels SLPI were significantly elevated compared to pathological controls (442 [313-737] versus 116 pg/mL [79-288]; P = 0.003). We measured regional levels of SLPI in six AALF patients at the time of OLT. A transhepatic gradient was demonstrated where higher concentrations of SLPI (72,130 [55,530-88,065] versus 60,410 pg/mL [44,177-67,410]; P = 0.03) were detected in the hepatic vein than in the portal vein (Fig. 2D). In three AALF patients, LPS-induced NF-κBp65 activation was reduced in monocytes in the hepatic vein compared to the portal vein at the time of OLT (Fig. 2E).

Using single immunostaining, SLPI expression was not detected in normal liver tissue but was detected in biliary epithelial cells in cirrhotic liver tissue. On examination of AALF explant tissue, there was marked SLPI expression in biliary epithelial cells and within large mononuclear cells within areas of necrosis; particularly at the boundary between necrotic and nonnecrotic parenchyma (Fig. 3A). Using double immunostaining, evidence of proliferative (ki67+) activity in SLPI-expressing biliary epithelial cells was detected in large bile ducts (representative example, Fig. 3A).

Figure 3.

SLPI expression in AALF and pathological control liver tissue. (A) (i) SLPI-expressing cells (red) were not detected in normal liver tissue (200× magnification); (ii) SLPI expression was detected in biliary epithelial cells in cirrhotic tissue (100× magnification); In AALF explant tissue, SLPI expression was seen in large mononuclear cells within areas of centrilobular necrosis (iii), particularly at the boundary between necrotic and nonnecrotic parenchyma (indicated by arrowheads, magnification = 100×) and in biliary epithelial cells. Of note, many biliary epithelial cells staining for SLPI were in cycle (Ki67+ [brown]/SLPI [red]) biliary epithelial cells) (iv-v). (B) Representative example of double epitope immunostaining and multispectral imaging (Nuance) used to determine colocalization of SLPI (red) within CD68+ hepatic macrophages (brown) in an area of centrilobular necrosis (200× magnification) and within a single macrophage (inset). Red chromogen signal for SLPI displayed digitally as red, brown chromogen signal for CD68 displayed as green, and the composite resulting image displaying colocalization of SLPI and CD68. (C) Representative example of multispectral imaging to colocalize and enumerate SLPI+CD68+ hepatic macrophages, shown in yellow on the pseudofluorescent composite image, within the same area as displayed in (B).

In line with previous data,[19] CD68+ hepatic macrophages (h-mψ) were abundant and concentrated within areas of centrilobular necrosis compared to pathological control liver (Fig. 3A,B). Double immunostaining for CD68 and SLPI (Fig. 3B) revealed the expression of SLPI in h-mψ (CD68+/SLPI+) within areas of centrilobular necrosis. Using multispectral imaging, the median percent of SLPI expressing h-mψ (CD68+/SLPI+) in 10 high-power fields was 65% (IQR 47-71) within areas of centrilobular necrosis (representative example shown in Fig. 3C).

Hepatic Macrophages Are Characterized by an Anti-inflammatory Phenotype in AALF

H-mψ expressing the anti-inflammatory marker CD163 were increased in AALF compared to pathological control tissue (Fig. 4A). In Fig. 4B,C, multispectral image analysis of triple immunostained (CD68, CD163, HLA-DR) AALF tissue revealed a higher percentage of h-mψ possessing the anti-inflammatory phenotype (CD68+CD163+HLA-DR+) in periportal regions compared to more central and perivenular areas of necrosis (86% [90-90] versus 61% [51.5-67]; P = 0.002).

Figure 4.

Anti-inflammatory hepatic macrophage phenotype in AALF. (A) Enumeration of CD163-positive macrophages in pathological control (Path control; n = 4) and AALF tissue (n = 5) in centrilobular areas and portal tracts. Representative single epitope immune for CD163 (red chromagen; 20× magnification) in pathological control and AALF tissue in portal tracts and centrilobular areas are shown below graph. (B) Representative example of triple epitope immunostaining and multispectral imaging (Nuance) used to immunophenotype hepatic macrophages: CD68+ (brown), HLA-DR (purple), and CD163 (red) within an area of necrosis (RGB; 10× magnification [inset 40× magnification]). Multispectral imaging of interface between portal tract and necrotic area (40× magnification; R sided panels) are displayed in pseudofluorescence: CD68-red; HLA-DR-purple; CD163-green and composite image of CD163 and HLA-DR colocalization shown in yellow (coloc). (C) Representative example of multispectral imaging to illustrate the periportal distribution of CD68+CD163+HLA-DR+ hepatic macrophages (and within single macrophages [inset]), shown in yellow on the pseudofluorescent composite image, within the same area as displayed in (B).

SLPI Induces an Anti-inflammatory Monocyte/Macrophage Phenotype Similar to AALF

We assessed the effects of recombinant human (rh)SLPI on the phenotype and function of normal monocytes (Fig. 5A-C). Consistent with previous reports,[24-26] SLPI-treated monocytes were significantly reduced in LPS-induced secretion of proinflammatory cytokines (TNF-α, IL-6), while levels of anti-inflammatory cytokines (IL-10, TGF-β1) were unaffected. Compared to untreated cells, rhSLPI induced a dose-dependent, selective increase in the surface expression of CD163 in CD14+CD16+ monocytes (29% [26-29.5] versus 51% [44-63]; P = 0.03), whereas no significant changes in CD163 expression in CD14+CD16− monocytes were detected (Supporting Results Fig. 11). Phagocytosis of E. coli (mean fluorescence intensity [MFI]) was similar between untreated and rhSLPI-treated (0.1 and 0.5 μg/mL) monocytes (9,329 [7,889-10,769] versus 8,835 [7,330-10,339] versus 9,307 [7,775-10,838], respectively; P = 0.1).

Figure 5.

Effects of SLPI on the phenotype and function of monocytes. (A,B) The effects of recombinant human SLPI (rhSLPI) (0.1 μg/mL [0.1], 0.5 μg/mL [0.5]) on the phenotype of healthy monocyte was assessed by flow cytometry and expressed as percentage of positive cells. Representative dotplots (A) of CD163, CD206, and HLA-DR percentage expression in untreated monocytes (no SLPI) (top row), SLPI treated monocytes ([0.1] and [0.5], middle and bottom row, respectively). (C) LPS-induced cytokine secretion was determined (in triplicate) after 48 hours of culture (n = 5 independent experiments). SSC-A: side scatter area. *P < 0.05; **P < 0.01; ***P < 0.001.

Hepatic and Systemic Inflammatory Microenvironment Induces Anti-inflammatory Monocyte Responses

We postulated that the hepatic inflammatory microenvironment in AALF modulated the phenotype and function of circulating monocytes similar to what we have documented ex vivo. We therefore cultured CD14+ monocytes, derived from HC, with homogenates derived from five AALF, five NL tissues and compared to culture in medium (CM) alone. As shown in Fig. 6A, significant increases were detected in CD163, CD206, and CD36 expression in monocytes cultured in the presence of homogenates derived from ALF, NL tissue compared to CM. Compared to NL and CM, culture in AALF homogenates resulted in a significant increase in the frequency of CD14+CD16+CD163+ monocytes and reduced TNF-α and IL-6 production following LPS stimulation (Fig. 6A,B). No significant difference in phagocytosis of FITC-labeled E. coli was detected in monocytes cultured in CM compared to NL and AALF homogenates (61% [59-79] versus 63% [56-83] versus 73% [61-88.5], respectively; P = 0.5).

Figure 6.

Effects of the hepatic inflammatory microenvironment on monocyte phenotype and function. CD14+ isolated monocytes derived from healthy volunteers were incubated for 48 hours with culture media alone (CM), liver homogenates from pathological control (CM+NL [n = 5]), and AALF explant liver tissue (CM+ALF [n = 5]). Cells were harvested at 48 hours for (A) phenotypic analysis and (B) TNF-α and IL-6 levels following 100 ng/mL LPS stimulation. (C-E) Effects of blocking SLPI (using α-SLPI at 1 μg/mL) on (D) phenotype and (E) LPS-induced secretion of TNF-α, IL-6, and IL-10 in monocytes incubated with AALF liver homogenates were assessed after a 48-hour culture period. Representative dotplots (C) of CD163 percentage expression in monocytes cultured for 48 hours in liver tissue homogenates from AALF patient for 48 hours in the absence (top row) or presence (bottom row) of α-SLPI. SSC-A: side scatter area. * P < 0.05; **P < 0.01; ***P < 0.001.

We also examined whether the systemic inflammatory microenvironment in ALF modulated the ability of CD14+ isolated monocytes and monocyte derived-macrophages (mψ) to produce TNF-α, IL-6, and IL-10 following LPS challenge (Fig. 7A-D). Culture of CD14+ monocytes in ALF sera markedly reduced secretion of TNF-α (2,887 [2,631-2,962] versus 176 pg/mL [71-626]; P < 0.0001), IL-6 (2800 [1751-3436] versus 30 pg/mL [16.5-72.5]; P < 0.0001), whereas no changes in IL-10 secretion were detected (26.5 [9.5-107] versus 58 pg/mL [28.5-277.5]; P = 0.1). Compared to HC sera, culture of mψ in ALF sera reduced TNF-α secretion (240 [92-512] versus 1.5 pg/mL [0-114]; P < 0.0001), while increasing IL-10 secretion (12 [0-62] versus 1165 pg/mL [569-2144]; P = 0.03). No significant changes in IL-6 secretion were detected (669 [567-853] versus 454 [276-687]; P = 0.1).

Figure 7.

Effects of the circulatory inflammatory microenvironment on monocyte/macrophage inflammatory cytokine secretion. The effects of incubating plasma from patients with AALF (n = 20) and healthy normal volunteers (n = 5) on LPS-induced levels of TNF-α and IL-10 in CD14+ monocytes (A,B) and monocyte derived-macrophages (C,D) after a 48-hour culture period; levels of LPS derived TNF-α (A,C) and IL-10 (B,D) were measured following preincubating AALF plasma with α-SLPI (ALF+αSLPI at 5 μg/mL) for 60 minutes at room temperature.

Blocking the Activity of SLPI Partially Reverses the Anti-inflammatory Monocyte Phenotype in AALF

In order to evaluate whether the elevated hepatic levels of SLPI could account for the changes in monocyte phenotype and function, we cultured CD14+ monocytes in AALF homogenates with and without blocking SLPI (1 μg/mL human recombinant anti-(α)SLPI). As shown in Fig. 6C-E, the administration of αSLPI significantly reversed the increase in CD14+CD16+CD163+ monocytes (33 [16-53] versus 58% [20-80]; P = 0.04) and the reduction in LPS-stimulated TNF-α (957 [583-1269] versus 1220 pg/mL [867-1476]; P = 0.04), whereas IL-6 and IL-10 secretion remained unaffected.

In order to assess the effects of elevated circulating SLPI levels on monocyte function, we cultured both monocytes and monocyte derived-macrophages in ALF plasma with and without αSLPI (5 μg/mL). Following LPS challenge, the administration of αSLPI significantly reversed the reduction in TNF-α secretion in monocyte (71 [0-176] versus 419 pg/mL [124-682]; P = 0.001) and mψ (1.5 [0-114] versus 175 pg/mL [141-356]; P < 0.001), whereas no significant changes in IL-10 secretion were detected following αSLPI (Fig. 7A-D). No differences in TNF-α and IL-10 secretion were detected following incubation of monocytes with plasma derived from healthy volunteers with αSLPI (Supporting Fig. 12).


Our data demonstrate marked elevations in circulating levels of SLPI in patients with both acetaminophen and nonacetaminophen-induced ALF, with higher SLPI levels detected in AALF patients with a greater severity of acute liver injury and an adverse clinical outcome. In NAALF, however, SLPI levels did not predict outcome, which likely reflects the smaller number of patients, heterogeneous nature, and clinical presentation of this patient group.

Peak circulating SLPI levels were detected on admission to our unit, suggesting its release is an early event in the evolution of acute liver injury and in keeping with its role in inhibiting tissue proteases released following tissue injury.[5] As seen in other inflammatory conditions,[12, 27] we propose that SLPI production from the inflamed liver is of sufficient magnitude to “spill-over” into the systemic circulation. This is evidenced by elevated intrahepatic concentrations and a significant SLPI gradient across the liver, with peak concentrations detected in the hepatic vein at the time of OLT.

Although the early increase in circulating levels of SLPI is likely to initially reflect a homeostatic response to the acute liver injury, we detect persistently elevated SLPI levels AALF patients up to 3 days following admission, which are associated with an increased frequency of infection. These findings are of relevance given reports showing that elevated SLPI levels were detrimental in patients suffering from systemic sepsis[12-14] where it inhibits proinflammatory responses and impairs monocyte driven T-cell proliferation and T-helper cell (Th)1 responses.[28] The utility of SLPI as a biomarker of infection and disease severity requires investigation in larger groups of patients suffering from ALF of acetaminophen and nonacetaminophen etiology.

Our data reveal that SLPI treatment recapitulates the functional and phenotypic characteristics of circulating monocytes from patients with AALF.[29, 30] Following ex vivo LPS challenge, AALF monocytes show reductions in NF-κBp65 expression, TNF-α, and IL-6 secretion. These data are in keeping with published reports demonstrating the ability of SLPI to directly inhibit monocyte proinflammatory responses following LPS challenge through the inhibition of the NF-κB signaling pathway.[24, 25, 31] Our in vitro data also demonstrate that SLPI induces anti-inflammatory monocyte phenotype that closely resembles that seen in AALF, where both SLPI-treated and AALF monocytes are characterized by significant increases in the anti-inflammatory marker CD163 on the surface of CD16-positive monocyte subsets. CD163 is a cell surface marker that is up-regulated in M2-polarized monocytes/macrophage populations[32] which attenuate proinflammatory responses and promote of resolution of inflammation and tissue repair processes.[33-36] Based on our findings, we hypothesize that the increased circulating concentrations of SLPI produced from the inflamed liver induce the anti-inflammatory monocyte phenotype characterized by attenuated proinflammatory responses to microbial challenge.

To address whether SLPI induced the anti-inflammatory monocyte phenotype of AALF, monocytes/macrophages were incubated with liver homogenates and plasma derived from AALF patients. Exposing healthy monocytes to intrahepatic and circulatory microenvironments induced an anti-inflammatory phenotype, typified by increased CD163 expression and attenuated proinflammatory (TNF-α, IL-6), while preserving/augmenting anti-inflammatory (IL-10) cytokine following LPS stimulation. Crucially, by blocking the activity of SLPI within liver homogenates and plasma, we were able to reverse the anti-inflammatory monocyte phenotype induced in AALF, thereby proving that SLPI is a key mediator in the induction of monocyte dysfunction in AALF. Given that SLPI did not fully reproduce the AALF phenotype, with respect to HLA-DR down-regulation, it is likely that other mediators, such as IL-10,[1] play a role in modulating monocyte function in AALF. Further studies are required to explore which other mediators also modulate monocyte function in AALF.

In response to tissue injury, SLPI serves to trigger resolution of inflammation by enhancing macrophage uptake of necrotic debris, apoptotic cells, and secretion of anti-inflammatory mediators.[9] Within the inflamed liver, we show that biliary epithelial cells and h-mψ within areas of hepatic necrosis are the cellular sources of elevated tissue concentrations of SLPI that is in keeping with other inflammatory pathologies.[5] Given the ability of anti-inflammatory mediators to induce macrophage SLPI production,[37] the increased concentrations of IL-6 and IL-10 within areas of hepatic necrosis may serve as the microenvironmental cues to promote SLPI expression in h-mψ.[19]

Our data also suggest that SLPI may also directly influence the phenotype of h-mψ in a paracrine fashion. We show an expansion of h-mψ typified by an anti-inflammatory phenotype (CD163+HLA-DR+) bearing striking similarities with the in vitro effects of SLPI described. The fact that CD163+HLA-DR+ h-mψ are predominately located in periportal areas indicates that secretion of SLPI from cholangiocytes may modulate their function and phenotype in AALF.

Taken together, we postulate that SLPI directly modulates the function of hepatic macrophages, offsetting excessive tissue destructive processes and initiating the clearance of necrotic debris and cells. Furthermore, given the role of SLPI in tissue repair processes and regenerative responses,[6-8, 38] the expression of SLPI in proliferating biliary epithelial cells suggests that it may also promote hepatic regeneration responses following acute liver injury. Functional and phenotypic analyses of freshly isolated h-mψ are warranted in order to dissect the role of SLPI in resolution of inflammation and tissue repair/regenerative responses in ALF.

Interventional strategies using SLPI to ameliorate the severity of tissue injury have previously been tested in inflammatory lung pathologies[39-42] and therefore may represent a promising therapeutic agent to ameliorate the severity of liver injury in patients with AALF. On the other hand, our data demonstrate significant effects of SLPI on suppression of peripheral monocyte function. As such, blocking the effects of SLPI could be a therapeutic option aimed at restoring peripheral immune function and decreasing the incidence of nosocomial sepsis in AALF. The balance of these two potential therapeutic interventional strategies need to be further investigated in experimental models to determine their impact and appropriate timeline of application in order to promote liver repair processes and, subsequently, to improve peripheral immune function without exacerbation of liver injury.

Figure 8 summarizes the postulated mechanisms through which SLPI promotes anti-inflammatory responses and offers a mechanistic explanation as to why circulating monocytes and hepatic macrophages acquire anti-inflammatory phenotypic and functional characteristics in AALF. SLPI is produced by epithelial cells and macrophages from the inflamed liver in order to offset overwhelming hepatic necrosis. It is, however, the excessive hepatic production of this anti-inflammatory mediator that is of sufficient magnitude to “spill-over” into the systemic circulation and render circulating monocytes hyporesponsive to microbial challenge and therefore increase the susceptibility to infection that is encountered during AALF.

Figure 8.

Postulated role of SLPI in monocyte dysfunction in AALF. During AALF, SLPI is released from the liver by biliary epithelial cells and hepatic macrophages in areas of centrilobular necrosis. The elevated levels SLPI likely reflect tissue responses to overwhelming hepatic necrosis and are postulated to promote the resolution of inflammation through (i) inhibition of tissue proteases and (ii) induction of an anti-inflammatory hepatic macrophage population through both autocrine (1) and paracrine (2) mechanisms. SLPI has an autocrine effect on macrophages within areas of necrosis, down-regulating NFκB-regulated proinflammatory cytokines while promoting anti-inflammatory/regenerative cytokines (e.g., IL-6, IL-10) and phagocytosis. (2) SLPI produced by the biliary epithelium may exert paracrine effects on resident and newly infiltrating macrophages, at the interface between necrotic and viable tissue. Here SLPI may contribute to the enhanced expression of anti-inflammatory marker CD163, along with other mediators, such as IL-6 and IL-10 released from necrotic tissue. However, hepatic production of SLPI from the inflamed is of sufficient magnitude to spill over in the systemic circulation where it modulates the function of circulating monocytes. SLPI induces an anti-inflammatory monocyte phenotype characterized by: (i) increased expression of anti-inflammatory marker, CD163, on CD16+ monocytes; (ii) reduced production of NF-κB-dependent proinflammatory cytokines (IL-6, TNFα), and (iii) preserved anti-inflammatory cytokine (IL-10) following microbial challenge. These functional changes in circulating monocytes could account for the immunoparesis observed and marked susceptibility to infection observed in patients with AALF.


We thank the Imperial National Institute of Health Research Biomedical Research Centre for infrastructure support, Medical Research Council (MRC), European Association for the Study of the Liver (EASL), Rosetrees Charitable Trust, and the King's College Hospital Research & Development for ongoing funding support. We also thank Mr. C. Starling, Department of Liver Histpathology at King's College Hospital, for technical assistance.