Portal de Periódicos da CAPES

Serum Apoptosis Markers in Acute Liver Failure: A Pilot Study

2007; Elsevier BV; Volume: 5; Issue: 12 Linguagem: Inglês

10.1016/j.cgh.2007.08.007

1542-7714

Anna Rutherford, Linda S. Hynan, Carolina B.S. Borges, David G. Forcione, Jason T. Blackard, Wenyu Lin, April Gorman, Obaid S. Shaikh, Adrian Reuben, Edwyn M Harrison, K. Rajender Reddy, William M. Le, Raymond T. Chung,

Hepatitis B Virus Studies

Background & Aims: We sought to determine whether circulating apoptotic markers are altered in acute liver failure (ALF), differ with etiology, or predict clinical outcome in this condition. Methods: Serum levels of soluble Fas (sFas), tumor necrosis factor–alpha (TNF-α), hepatocyte growth factor (HGF), and interleukin-6 (IL-6) were measured in 67 acute liver failure patients, as well as controls. In a subset of the groups, we measured serum M-30 antigen, an exposed neoepitope from caspase cleavage. We also assessed M-30 immunoreactivity in liver tissue of ALF patients and controls. Results: Median levels for TNF-α, HGF, IL-6, and M-30 antigen were at least 10-fold greater in ALF than in hepatitis C virus or normal controls (P < .0001). Median day 1 sFas, day 3 sFas, and day 1 HGF levels varied according to etiology of acute liver failure (P = .004, P = .011, and P = .019, respectively), with values for drug-induced liver injury and acetaminophen-related ALF higher than other etiologies. Median M-30 antigen levels were significantly higher in patients who were transplanted and/or died (2183 U/L) than spontaneous survivors (1004 U/L) (P = .026). M-30 immunoreactivity in liver tissue was significantly greater in ALF patients than HCV controls (P = .004). Conclusions: TNF-α, HGF, IL-6, and M-30 antigen were significantly elevated in ALF. High levels of sFas and HGF might help to confirm a diagnosis of drug-induced liver injury or acetaminophen-related ALF. Higher levels of M-30 antigen are associated with poor clinical outcomes in ALF. Background & Aims: We sought to determine whether circulating apoptotic markers are altered in acute liver failure (ALF), differ with etiology, or predict clinical outcome in this condition. Methods: Serum levels of soluble Fas (sFas), tumor necrosis factor–alpha (TNF-α), hepatocyte growth factor (HGF), and interleukin-6 (IL-6) were measured in 67 acute liver failure patients, as well as controls. In a subset of the groups, we measured serum M-30 antigen, an exposed neoepitope from caspase cleavage. We also assessed M-30 immunoreactivity in liver tissue of ALF patients and controls. Results: Median levels for TNF-α, HGF, IL-6, and M-30 antigen were at least 10-fold greater in ALF than in hepatitis C virus or normal controls (P < .0001). Median day 1 sFas, day 3 sFas, and day 1 HGF levels varied according to etiology of acute liver failure (P = .004, P = .011, and P = .019, respectively), with values for drug-induced liver injury and acetaminophen-related ALF higher than other etiologies. Median M-30 antigen levels were significantly higher in patients who were transplanted and/or died (2183 U/L) than spontaneous survivors (1004 U/L) (P = .026). M-30 immunoreactivity in liver tissue was significantly greater in ALF patients than HCV controls (P = .004). Conclusions: TNF-α, HGF, IL-6, and M-30 antigen were significantly elevated in ALF. High levels of sFas and HGF might help to confirm a diagnosis of drug-induced liver injury or acetaminophen-related ALF. Higher levels of M-30 antigen are associated with poor clinical outcomes in ALF. Acute liver failure (ALF) is characterized by widespread hepatocyte death in excess of regeneration. Apoptotic cell death is an important contributor to the organ failure common to ALF, even for etiologies thought traditionally to involve mainly hepatocyte necrosis.1Kass G.E. Macanas-Pirard P. Lee P.C. et al.The role of apoptosis in acetaminophen-induced liver injury.Ann N Y Acad Sci. 2003; 1010: 557-559Crossref PubMed Scopus (31) Google Scholar, 2El-Hassan H. Anwar K. Macanas-Pirard P. et al.Involvement of mitochondria in acetaminophen-induced apoptosis and hepatic injuryL Roles of cytochrome c, Bax, Bid and caspases.Toxicol Appl Pharmacol. 2003; 181: 118-129Crossref Scopus (130) Google Scholar Recent research has clarified the processes intrinsic to apoptosis, which makes the study of these pathways possible. The chief protagonists in apoptosis are the Fas/Fas ligand pathway, mitochondrial cytochrome C release, and the tumor necrosis factor–alpha (TNF-α)–mediated death pathway. All of these pathways lead to activation of caspases, cellular proteases that induce degradation of DNA and membrane perturbations that are characteristic of apoptotic cell death. It is believed that pro-apoptotic proteins work in dynamic opposition to antiapoptotic proteins to cause apoptosis. Because of the rarity of ALF, no comprehensive human studies of apoptotic markers or their potential as predictors of outcome have been conducted to date. In mice, agonistic antibodies to the critical apoptosis-inducing membrane receptor Fas (CD95) cause rapid development of fulminant apoptotic hepatic failure.3Galle P.R. Hofmann W.J. Walczak H. et al.Involvement of the CD95 (APO-1/Fas) receptor and ligand in liver damage.J Exp Med. 1995; 182: 1223-1230Crossref PubMed Scopus (678) Google Scholar This process can be abrogated experimentally by the antiapoptotic molecule bcl-2,4Lacronique V. Mignon A. Fabre M. et al.Bcl-2 protects from lethal hepatic apoptosis induced by an anti-Fas antibody in mice.Nat Med. 1996; 2: 80-86Crossref PubMed Scopus (349) Google Scholar as well as by inhibitors of the effector caspase-8,5Bajt M.L. Lawson J.A. Vonderfecht S.L. et al.Protection against Fas receptor-mediated apoptosis in hepatocytes and nonparenchymal cells by a caspase-8 inhibitor in vivo: evidence for a postmitochondrial processing of caspase-8.Toxicol Sci. 2000; 58: 109-117Crossref PubMed Scopus (112) Google Scholar and hepatocyte growth factor (HGF),6Kosai K. Matsumoto K. Nagata S. et al.Abrogation of Fas-induced fulminant hepatic failure in mice by hepatocyte growth factor.Biochem Biophys Res Commun. 1998; 244: 683-690Crossref PubMed Scopus (185) Google Scholar which induces production of the antiapoptotic bcl-xL as well as blockade of caspase-3. In viral liver disease, Fas mediates hepatocyte death in animal models of HBV liver disease and is overexpressed in hepatocytes from patients with chronic viral hepatitis.7Song le H. Binh V.Q. Duy D.N. et al.Variations in the serum concentrations of soluble Fas and soluble Fas ligand in Vietnamese patients infected with hepatitis B virus.J Med Virol. 2004; 73: 244-249Crossref PubMed Scopus (24) Google Scholar In humans, circulating soluble Fas (sFas) ligand, which activates cell death by binding cell surface Fas, was observed to be elevated in a small Japanese study of acute hepatitis and fulminant hepatitis.8Ryo K. Kamogawa Y. Ikeda I. et al.Significance of Fas antigen-mediated apoptosis in human fulminant hepatic failure.Am J Gastroenterol. 2000; 95: 2047-2055Crossref PubMed Google Scholar Toxic liver injury, caused by drugs for example, is likely to induce apoptosis through direct mitochondrial injury induced by accumulation of reactive oxygen intermediates. In addition, experimental evidence from animal models suggests that in alcoholic liver disease9Tagami A. Ohnishi H. Moriwaki H. et al.Fas-mediated apoptosis in acute alcoholic hepatitis.Hepatogastroenterology. 2003; 50: 443-448PubMed Google Scholar and fulminant Wilson's disease,10Strand S. Hofmann W.J. Grambihler A. et al.Hepatic failure and liver cell damage in acute Wilson's disease involve CD95 (APO-1/Fas) mediated apoptosis.Nat Med. 1998; 4: 588-593Crossref PubMed Scopus (232) Google Scholar elevated Fas ligand expressed in hepatocytes might induce "fratricidal" apoptosis. In a study of 24 patients with ALF, serum sFas was significantly elevated in ALF, and levels were greater in patients with ALF caused by acetaminophen than in those with ALF caused by viral hepatitis.11Tagami A. Ohnishi H. Hughes R.D. Increased serum soluble Fas in patients with acute liver failure due to paracetamol overdose.Hepatogastroenterology. 2003; 50: 742-745PubMed Google Scholar A critical role for apoptosis in mediating acetaminophen-induced liver injury has been suggested by data showing that inhibitors of caspases protect mice from acetaminophen-induced hepatic injury and ultimately necrosis in vivo by blocking the activation of the mitochondrial pathway of apoptosis.2El-Hassan H. Anwar K. Macanas-Pirard P. et al.Involvement of mitochondria in acetaminophen-induced apoptosis and hepatic injuryL Roles of cytochrome c, Bax, Bid and caspases.Toxicol Appl Pharmacol. 2003; 181: 118-129Crossref Scopus (130) Google Scholar Thus, in addition to the necrotic injury attributable to a range of toxins, including acetaminophen, there is evidence that significant apoptotic cell injury is operative in these etiologies. Recently, Bantel et al12Bantel H. Lugering A. Heidemann J. et al.Detection of apoptotic caspase activation in sera from patients with chronic HCV infection is associated with fibrotic liver injury.Hepatology. 2004; 40: 1078-1087Crossref PubMed Scopus (224) Google Scholar measured circulating apoptotic activity in patients with chronic HCV by using an enzyme-linked immunosorbent assay (ELISA) that selectively recognizes an exposed neoepitope after cleavage of the caspase substrate cytokeratin-18 called M-30 antigen. This protein was selected because it is abundant in hepatocytes, thereby providing potential specificity for the source. Bantel et al found that in chronic HCV, in contrast to conventional surrogate markers such as serum aminotransferases, detection of caspase activity is a more sensitive method for detecting early apoptotic liver injury. Although it has been shown that the processing of caspases to their active fragments and cleavage of cytokeratin-18 occur in acetaminophen-induced liver injury,1Kass G.E. Macanas-Pirard P. Lee P.C. et al.The role of apoptosis in acetaminophen-induced liver injury.Ann N Y Acad Sci. 2003; 1010: 557-559Crossref PubMed Scopus (31) Google Scholar these substrates have never been evaluated in ALF patients. In this study, we measured serum levels of 4 apoptosis-associated markers and M-30 antigen in ALF patients, as well as chronic HCV and healthy controls. We hypothesized that serum levels of these markers would be significantly elevated in ALF patients as compared with controls, and that serum levels would differ on the basis of etiology of ALF and outcome. We investigated sera from 67 ALF patients, randomly selected within a particular etiology on the basis of availability of serum samples, from the Acute Liver Failure Study Group (ALFSG) data and serum bank and 2 control groups: 68 patients with chronic HCV from Massachusetts General Hospital and 20 healthy patients from Massachusetts General Hospital. The HCV patients were randomly selected serum samples from observational cohorts conducted at Massachusetts General Hospital. They were HCV antibody–positive and had positive HCV RNA. Of the 68 HCV controls, 35% had been treated, and 61% were known to be genotype 1. The healthy controls were human immunodeficiency virus–negative, HCV-negative individuals enrolled in a study to determine HLA types in the general population. All the control specimens were obtained with informed consent under a protocol approved by the Massachusetts General Hospital Institutional Review Board. All samples were stored at −80° within 2 hours until ready for use, and the investigators were blinded to diagnosis and outcome. Diagnoses for the etiology of the ALF patients were determined by the site investigator by using standard criteria. Since January 1, 1998, the ALFSG has prospectively collected demographic, clinical, laboratory, and outcome information as well as serum on all subjects meeting entry criteria for ALF at the 23 academic centers participating in the ALFSG. Eligible patients had an international normalized ratio (INR) >1.5, evidence of hepatic encephalopathy, and presented within 26 weeks of illness onset without apparent chronic liver disease. Because subjects were encephalopathic (by definition), written informed consent was obtained from their legal next of kin. Data were collected on an admission case report form and subsequently on a separate study outcome form, with outcome defined as liver transplantation, discharge, or death 3 weeks after admission.13Ostapowicz G. Fontana R.J. Schiodt F.V. et al.Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States.Ann Intern Med. 2002; 137: 947-954Crossref PubMed Scopus (1762) Google Scholar The study was performed according to the institutional review board guidelines of each of the 23 centers of the ALFSG and under authorization of the ALFSG ancillary studies committee. For quantitative measurement of the 4 apoptosis associated proteins, sFas, TNF-α, HGF, and interleukin-6 (IL-6), we used cytokine-specific Quantikine ELISA kits (R&D Systems Inc, Minneapolis, MN). The minimal detectable dose for each assay was sFas (<20 pg/mL), TNF-α (<1.6 pg/mL), HGF (<40 pg/mL), and IL-6 (<0.70 pg/mL). Any value lower than the minimal detectable dose was given a value of 0 pg/mL in the analysis. For ALF patients' sera, samples from day 1 and day 3 of study enrollment were examined. All assays were performed under blinded identifier conditions. Assays were initially performed with no dilution, and then samples with values out of the linear range of the ELISA standards were repeated with dilutions at up to 1:100 as necessary, and a mean value was calculated for each sample. For a subset of 52 patients with varying etiologies of ALF as well as 20 controls who had available sera, serum was quantitatively measured for the apoptosis-associated CK18Asp396 (M30) proteolytic neoepitope with the M30-Apoptosense ELISA (PEVIVA; Alexis, Gruwald, Germany). The minimal detectable dose for this assay was <30 U/L. Assays were performed on day 1 sera where available (22/52), or if day 1 sera were unavailable, day 2 (9/52) or day 3 sera (21/52) were used. Dilutions of 1:3 and 1:10 were performed in each sample because of the initial high values for the ALF patients, and a mean value was calculated for each sample. Histologically, we assessed apoptosis by staining paraffin-embedded sections from liver explants or liver biopsies of 12 of 67 ALF patients and 6 chronic HCV controls from our institution with M-30 mouse monoclonal antibody (M-30 CytoDeath; Peviva, Bromma, Sweden) according to the manufacturer's instructions. Briefly, the histologic sections were incubated with the primary antibody at room temperature for 2 hours, diluted 1:100 PBS containing 1% bovine serum albumin. For the negative control samples, the primary antibody was replaced with PBS. The slides were then washed 3 times in PBS and incubated for 1 hour with an anti-Fc biotinylated secondary antibody and, after another washing, were incubated for 30 minutes with a solution containing a trivalent complex constituted by avidin-biotin-peroxidase (Vectastain Elite ABC Kit; Vector Laboratories, Burlingame, CA). Development was carried out with 3,3′-diaminobenzidine (Liquid DAB+; DAKO, Glostrup, Denmark) for 4 minutes. Quantitative evaluation of M-30 cytoplasmic staining was performed by counting apoptotic cells/100 hepatocytes in 20 randomly selected high-power fields, and results were expressed as a percentage of the total number of cells counted (apoptotic cells/100 hepatocytes). Comparisons between the median values of the 2 groups were performed with a Wilcoxon rank sum test. Data are presented as medians and ranges, except where noted. Statistical analyses comparing the median levels of the sFas, TNF-α, HGF, and IL-6 (all pg/mL) among ALF, HCV, and healthy controls, among ALF patients with different etiologies, and for different outcomes among ALF patients (death, spontaneous survival, transplanted alive, and transplanted dead) were performed with a Kruskal-Wallis test. The t test was used to compare mean levels of day 1 sFas, TNF-α, HGF, and IL-6 by outcome (transplant/death vs spontaneous survival), whereas the Wilcoxon rank sum test was used to compare the median M-30 antigen level by outcome (transplant/death vs spontaneous survival), and a Kruskal-Wallis test was used to compare median M-30 antigen level by etiology of ALF. Spearman correlations were performed for day 1 and day 3 cytokine values (sFas, TNF-α, HGF, IL-6, M-30 antigen) and other known clinical markers of severity including admission ALT, lactate, sodium, INR, and maximum coma grade. All analyses were two-tailed. P values of less than .05 were considered to indicate statistical significance. All analyses were performed with SAS 9.1.3 (SAS Institute Inc, Cary, NC) or SPSS (SPSS Inc, Chicago, IL). Table 1 depicts the demographic characteristics of the ALF patients grouped by etiology for cytokine analysis, as well as a comparison to the characteristics of the entire ALFSG cohort enrolled during the same time period. The groups were not further compared by etiology because the cytokine analysis was designed to include a certain number of each major category of etiology of ALF; therefore, they were not expected to match the overall cohort distribution of etiology. The groups differed in terms of median admission bilirubin, which was much higher in the cytokine group than in the overall cohort (18.2 vs 7.5 mg/dL, P = .0001). This difference might be explained by the greater percentages of indeterminate and drug-induced liver injury cases included in the cytokine analysis, which tend to have higher admission bilirubin than acetaminophen cases.12Bantel H. Lugering A. Heidemann J. et al.Detection of apoptotic caspase activation in sera from patients with chronic HCV infection is associated with fibrotic liver injury.Hepatology. 2004; 40: 1078-1087Crossref PubMed Scopus (224) Google Scholar For the chronic HCV controls, the median age was 50 years (range, 28–71), and 28 of 68 (41%) were female. For the healthy control group, the median age was 30 years (range, 24–57), and 10 of 20 (50%) were female.Table 1Baseline Characteristics of Selected Cytokine Patients Compared With Overall CohortCytokine cohort (n = 67)Overall ALFSG cohort (n = 642)P valueMedian age (y) (range)39 (17–76)38 (15-81).751Female (%)40 (61.2)439 (67.3).231Etiology (%) Acetaminophen17 (25.4)295 (46) Viral16 (23.9)67 (10.4)N/A Drug-induced liver injury12 (17.9)75 (11.7) Wilson's disease6 (8.9)9 (1.4) Indeterminate16 (23.9)87 (13.6)Outcome (%) Overall death from ALF27 (40.3)210 (32.7).210 Transplanted22 (32.8)162 (25.2).177 Spontaneous survival23 (34.3)288 (44.9).098Median admission coma grade (range)2 (1–4)2 (1–4).726Median admission ALT (U/L) (range)1474 (11–13,100)2057 (3–19,826).073Median admission bilirubin (mg/dL) (range)18.2 (1–60)7.5 (0.3–63.3).0001Median admission INR (range)2.5 (1.4–10.3)2.8 (1.2–44.4).087NOTE. The demographic characteristics of the patients grouped by etiology for cytokine analysis are depicted, as well as a comparison to the characteristics of the entire ALFSG cohort enrolled during the same time period. The cytokine group was designed to include a balanced number from each major ALF etiologic category; therefore, they did not match the overall cohort etiology distribution. The groups differed in terms of median admission bilirubin, which was much higher in the cytokine group than in the overall cohort (18.2 vs 7.5 mg/dL, P = .0001). Open table in a new tab NOTE. The demographic characteristics of the patients grouped by etiology for cytokine analysis are depicted, as well as a comparison to the characteristics of the entire ALFSG cohort enrolled during the same time period. The cytokine group was designed to include a balanced number from each major ALF etiologic category; therefore, they did not match the overall cohort etiology distribution. The groups differed in terms of median admission bilirubin, which was much higher in the cytokine group than in the overall cohort (18.2 vs 7.5 mg/dL, P = .0001). Median levels of TNF-α were 10.6 pg/mL (range 0–119 pg/mL) for ALF patients, 0 pg/mL (range 0–119 pg/mL) for HCV controls, and 0 pg/mL for healthy controls. Median levels of HGF were 14,675 pg/mL (range 2500–96,575 pg/mL) for ALF patients, 1586 pg/mL (range 126–8447 pg/mL) for HCV controls, and 335 pg/mL (range 153–2375 pg/mL) for healthy controls. Median levels of IL-6 were 105.4 pg/mL (range 0–1571 pg/mL) for ALF patients, 6.4 pg/mL (range 2–105 pg/mL) for HCV controls, and 2.1 pg/mL (range 0–20 pg/mL) for healthy controls. Therefore, median values for TNF-α, HGF, and IL-6 were at least 10-fold higher in patients with ALF compared with those with chronic HCV or healthy controls (P < .0001). In contrast, median sFas levels were more elevated in chronic HCV (26,200 pg/mL, range 0–81,500 pg/mL) compared with ALF patients (14,038 pg/mL, range 3250–78,807 pg/mL) or healthy controls (7300 pg/mL, range 2461–12,382 pg/mL) (P < .0001) (Figure 1). Day 1 and day 3 sFas levels were significantly different (P = .004 and P = .011, respectively) according to the underlying etiology of ALF, with values for ALF caused by idiosyncratic drug-induced liver injury (DILI) higher than those for viral hepatitis or indeterminate causes (Figure 2). Median day 1 sFas levels in patients with ALF caused by acetaminophen (19,731 pg/mL, range 5250–41,500 pg/mL) approached those observed with DILI (25,885 pg/mL, range 5654–78,808 pg/mL) but were not significantly different compared with the other etiologies studied. Median day 1 HGF was highest in DILI patients (26,275 pg/mL, range 3678–96,575 pg/mL) compared with acetaminophen (15,975 pg/mL, range 7300–57,625 pg/mL), indeterminate causes (9587 pg/mL, range 3200–63,075 pg/mL), or viral hepatitis (12,962 pg/mL, range 2500–22,575 pg/mL) (P = .019) (Figure 2). For sFas as well as day 1 HGF, median levels for Wilson's disease were similar to the values of DILI, and the differences between them were not significant. Median day 1 sFas was significantly higher in patients who died (19,654 pg/mL, range 4310–78,808 pg/mL) than in those who survived (13,700 pg/mL, range 5250–41,885 pg/mL) or underwent liver transplantation (9192 pg/mL, range 3250–27,192 pg/mL) (P = .043). Day 3 HGF and TNF-α levels were also significantly higher in patients who died than for all other outcomes (P = .032 and P = .039, respectively). Within the entire group of ALF and controls, TNF-α, HGF, and IL-6 were all highly intercorrelated (minimum rho > 0.60 and P < .0001). Correlations of day 1 and day 3 sFas, TNF-α, HGF, and IL-6 levels with admission ALT, INR, serum lactate, serum sodium, and maximum coma grade were performed. A significant correlation was seen between day 3 HGF, day 3 TNF-α, day 3 IL-6, and maximum coma grade (P = .0012, P = .0191, and P = .0002, respectively). The mean levels of M-30 antigen for healthy controls (120.84 ± 37.0 U/L) and chronic HCV (517 ± 645.4 U/L) patients were in strong agreement with those measured previously by Bantel et al12Bantel H. Lugering A. Heidemann J. et al.Detection of apoptotic caspase activation in sera from patients with chronic HCV infection is associated with fibrotic liver injury.Hepatology. 2004; 40: 1078-1087Crossref PubMed Scopus (224) Google Scholar (173 ± 9 and 498 ± 58 U/L, respectively). We report the mean M-30 antigen values for this comparison, because the original study by Bantel et al reported mean values; however, the rest of our data is reported as median values. Median M-30 antigen levels were significantly higher in all ALF patients (1686 U/L, range 230–15,590 U/L) than in healthy controls (105 U/L, range 86–206 U/L) or chronic HCV patients (288 U/L, range 92–2265 U/L) (P < .0001) (Figure 3A). Median levels of M-30 antigen differed among different etiologies of ALF: acetaminophen (1217 U/L), viral hepatitis (2345 U/L), DILI (1441 U/L), Wilson's disease (5392 U/L), and indeterminate (1991 U/L), but these differences were not statistically significant, given the small sample sizes (P = .241). However, median M-30 antigen levels were considerably higher in patients who were transplanted and/or died (2183 U/L, range 536–15,590 U/L) than in those who spontaneously survived (1004 U/L, range 230–15,590 U/L) (P = .026) (Figure 3B). Comparisons between M-30 antigen and admission ALT, INR, serum lactate, serum sodium, maximum coma grade, and MELD score revealed significant correlation between M-30 antigen levels and maximum coma grade (P = .013) and MELD score (rho = 0.304, P = .03). Histologically, we assessed apoptosis by immunohistochemical staining with M-30 antibody in 12 of the 67 ALF patients who had available biopsies (all underwent transplantation or died) and compared this with 6 cases of chronic HCV as controls. For the ALF patients, the median value of apoptotic hepatocytes, expressed by M-30 immunoreactivity, was 13.7 apoptotic cells/100 hepatocytes (range, 4.1–24.3 apoptotic cells/100 hepatocytes) compared with 1.4 apoptotic cells/100 hepatocytes (range, 0.2–2.4 apoptotic cells/100 hepatocytes) for HCV controls, P = .004. Figure 4 shows an example of M-30 staining in an ALF case in comparison to an HCV control. In this pilot study, we found that the activity of the apoptosis-associated proteins TNF-α, HGF, IL-6, and M-30 antigen was significantly higher in ALF patients than in either chronic HCV or healthy controls, confirming the dramatic activation of apoptosis in the cell death attending ALF. Of note, the strong correlation between apoptotic and antiapoptotic markers suggests that activation of apoptosis triggers compensatory changes in the antiapoptotic markers as well. The markedly elevated levels of sFas that we saw in chronic HCV patients has been previously described in association with a decrease in responsiveness to interferon (IFN) treatment,14Lapinski T.W. Kowalczuk O. Prokopowicz D. et al.Serum concentration of sFas and sFasL in healthy HbsAG carriers, chronic viral hepatitis B and C patients.World J Gastroenterol. 2004; 24: 3650-3653Google Scholar as well as increased severity of HCV histologic grade.15Kakiuchi Y. Yuki N. Iyoda K. et al.Circulating soluble Fas levels in patients with hepatitis C virus infection and interferon therapy.J Gastroenterol. 2004; 12: 1189-1195Crossref Scopus (6) Google Scholar Although it has traditionally been asserted, on the basis of previous studies of small ALF cohorts,11Tagami A. Ohnishi H. Hughes R.D. Increased serum soluble Fas in patients with acute liver failure due to paracetamol overdose.Hepatogastroenterology. 2003; 50: 742-745PubMed Google Scholar that differing etiologies of ALF result in differing proportions of apoptosis to necrosis, we now demonstrate variable degrees of apoptotic activity both within and between different ALF etiologies. In particular, high early levels of sFas and HGF appear to be characteristic of DILI as well as most acetaminophen ALF cases. These early differences might help distinguish DILI from other etiologies of ALF and could help to confirm the diagnosis, especially when toxicology screens and patient history are unrevealing or not obtainable. Median levels of sFas, TNF-α, and HGF were all significantly higher in patients who died of ALF, suggesting these cytokines might be useful in developing aids to predict outcome. Interestingly, we did not show that the antiapoptotic proteins HGF and IL-6 were significantly lower in patients with worse outcomes, again suggesting that compensatory increases in antiapoptotic markers occur in all patients in the face of overwhelming apoptosis, regardless of clinical outcome. Given that we observed diffuse activation of apoptosis throughout the ALFSG cohort tested, we therefore asked whether a more specific measure of the end products of apoptosis might correlate with ALF outcomes. Caspase cleavage, predominantly by caspase-3 in hepatocytes, is the final common executioner step in the apoptotic pathway. Already previously suggested as a very sensitive indicator of apoptosis in chronic HCV patients, the measured level of the M-30 antigen is not only dramatically elevated in ALF patients but also appears to be a significant predictor of death or need for liver transplantation versus spontaneous survival. The sensitivity of this marker suggests it has high potential to be an early, predictive tool of outcome in ALF. There was a significant correlation between day 3 TNF-α, HGF, IL-6, and M-30 antigen levels and maximum coma grade levels in these patients. M-30 antigen also appeared to be moderately correlated with admission MELD score. In contrast, serum cytokine levels were not significantly correlated with other markers of disease severity in ALF such as ALT, INR, lactate, and serum sodium. These findings suggest that serum cytokines might have unique potential for prediction of outcome and could be used in conjunction with other outcome variables to enhance development of a prediction rule for outcome in ALF. However, in this pilot study, we were not able to incorporate serum cytokine levels into a clinical prediction rule, in conjunction with other clinical and serum markers, to predict outcome in ALF because of the small number of patients in this study. Further study of larger numbers of ALF patients is clearly warranted. Finally, our finding of significantly increased M-30 immunostaining in liver tissue from ALF patients as compared with HCV controls confirms that serum M-30 antigen is accompanied by profound hepatocyte apoptosis. Specifically, the staining confirms that cytokeratin-18 degradation products originate predominantly from the liver and not from other epithelial cells affected by multiple organ failure. We were unable to statistically confirm correlation between histologic M-30 immunoreactivity and serum M-30 antigen levels; this is likely due to the small number of patients for whom histology was available. This pilot study represents the largest effort to date to characterize levels of apoptosis-related proteins for different etiologies and outcomes of ALF. However, we were still limited by the relatively small number of ALF patients examined. Assays were not run in duplicate because of the limited supply of serum and the need for significant dilution for many samples, which might have decreased the accuracy of each specific measurement. Although randomly selected, the patients used for the analysis differed from the overall ALFSG cohort characteristics in terms of median admission bilirubin, although they were very similar in outcome. This difference is likely due to the selected overall distribution of the etiologies for the cytokine analysis not reflecting the overall distribution of etiologies within the ALFSG cohort. In summary, we confirm that apoptosis plays a significant role in the cell death that occurs in ALF across all etiologies examined. The serum apoptosis markers sFas and HGF appear to be most markedly elevated in ALF caused by DILI and acetaminophen, implying that apoptosis is the pivotal feature in drug-related liver injury. Our findings also suggest a scientific rationale for treatment trials with agents such as caspase inhibitors that block apoptotic pathways, and these markers might offer a specific biologic means of assessing treatment responses. Finally, the ability to predict outcome early in the clinical course of ALF patients through the use of a sensitive marker of hepatic apoptosis justifies a larger study of the predictive value of M-30 antigen levels in ALF. The authors thank the ALFSG site coordinators and nurses for their support. U. S. Acute Liver Failure Study Group, 1998-2007: William M. Lee (principal investigator), Julie Polson, Linda Hynan, Joan Reisch, Carla Pezzia, Joe Webster, University of Texas Southwestern Medical Center, Dallas, TX; Anne Larson, University of Washington, Seattle, WA; Tim Davern, University of California, San Francisco, CA; Paul Martin, Mount Sinai School of Medicine, New York, NY; Timothy Macashland, University of Nebraska, Omaha, NE; J. Eileen Hay, Mayo Clinic, Rochester, MN; Nataline G. Murray, Baylor University Medical Center, Dallas, TX; Obaid Shakil Shaikh, University of Pittsburgh Medical Center, Pittsburgh, PA; Andres Blei, Northwestern University Medical School, Chicago, IL; Kenneth Ingram, Atif Zaman, Oregon Health Sciences University, Portland, OR; Steven Han, University of California, Los Angeles, CA; Robert Fontana, University of Michigan Medical Center, Ann Arbor, MI; Brendan McGuire, University of Alabama at Birmingham, Birmingham, AL; Raymond T. Chung, Anna Rutherford, Massachusetts General Hospital, Boston, MA; Alastair Smith, Duke University Medical, Durham, NC; Michael L. Schilsky, New York Presbyterian Hospital, New York, NY; Adrian Reuben, Medical University of South Carolina, Charleston, SC; Andres Riera, Santiago Munoz, Albert Einstein Medical Center, Philadelphia, PA; Mical Campbell, Rajender Reddy, University of Pennsylvania, Philadelphia, PA; Todd Stravitz, Virginia Commonwealth University, Richmond, VA; Lorenzo Rossaro, University of California, Davis, CA; Mayo Clinic, Jacksonville, Jacksonville, FL; Tarek Hassanein, University of California San Diego, San Diego, CA.