Journal of Critical Care, Vol 16, No 2 (June), 2001: pp 59-63 59

Purpose: The Fas ligand (FasL)/Fas system is an apop-

tosis induction system that plays an important role in

homeostasis and biophylaxis. We measured tumor

necrosis factor alpha (TNF-�), soluble FasL (sFasL), and

soluble Fas (sFas) in patients with acute hepatic fail-

ure to determine the relation between such failure and

apoptosis.

Materials and Methods: We assayed 21 blood samples

from patients with acute hepatic failure and 8 from

patients with sepsis but without acute hepatic failure.

Serum TNF-�, sFas, and sFasL levels were determined

by enzyme-linked immunosorbent assay.

Results: sFasL levels were significantly higher in the

patients with acute hepatic failure than in the patients

with sepsis (0.68 � 0.42 ng/mL vs. 0 ng/mL, P � .0001).

No significant differences were observed in sFas lev-

els between the two groups. A significant correlation

was observed between TNF-� and sFas levels (r �0.657, P � .0008); a negative correlation was observed

between TNF-� and sFasL levels (r � 0.454, P � .038).

Conclusions: Our results suggest that pathologic ag-

gravation of acute hepatic failure are related to

changes in the FasL/Fas system and that TNF-� and

sFasL, in particular, may play hepatoprotective roles.

Copyright © 2001 by W.B. Saunders Company

Soluble Fas and Soluble Fas Ligand Levels in Patients With AcuteHepatic Failure

Hajime Nakae, Kyoko Narita, and Shigeatsu Endo

FAS ANTIGEN belongs to the tumor necrosisfactor (TNF) receptor family and is often ex-

pressed in thymus, ovary, heart, kidney, and liver.1-3

Fas ligand (FasL) is a protein that is specificallybound to Fas antigen and has been discovered inthe membrane of cytotoxic T lymphocytes.4-6

When FasL comes into contact with Fas, apoptosisis induced in target cells. TNF-�, which is similarin structure to FasL, is produced by inflammatorycells, such as macrophages, monocytes, lympho-cytes, and fibroblasts, subsequent to inflammationand infection.

Apoptosis has been shown to play an importantrole in liver damage associated with hepatitis andother liver diseases as well as in the natural deathof cultured hepatocytes.2,7 TNF-� and anti-Fas an-tibody, when administered in vivo, cause extensivehepatocyte apoptosis8,9; when added to hepatocyteculture, they also induce apoptosis. Multiple sub-stances are involved in the development of liverfailure, but apoptosis of liver cells via the FasL/Fassystem is considered the most important mecha-nism in this process. Thus we measured TNF-�,soluble FasL (sFasL), and soluble Fas (sFas) in patients with acute liver failure to better under-stand the relation between acute liver failure andapoptosis.

MATERIALS AND METHODS

Patients

Five patients (1 man and 4 women, aged 54 to 75 years) withacute hepatic failure who were referred to our intensive care unitin 1999 and underwent plasma exchange comprised the studygroup. Three of these patients had fulminant hepatic failure dueto hepatitis, and 2 had postoperative hepatic failure due to acute

pancreatitis; all 5 patients died within 59 days (13 to 59 days)after the exchange. Fulminant hepatic failure was diagnosedwhen hepatic failure developed in a previously healthy individ-ual within 8 weeks of the onset of liver disease, a grade 2 orhigher coma ensued, and a prothrombin time �40% of normalwas observed. Postoperative hepatic failure was defined as acutesevere liver dysfunction with a total serum bilirubin of �10mg/dL and an ensuing grade 2 or higher coma that developedwithin 1 month of surgery.10,11

Eight patients (4 men and 4 women; aged 52 to 68 years)with sepsis who did not suffer from acute hepatic failure com-prised a control group. Diagnosis of sepsis was based on thecriteria of the College of Chest Physicians/Society of CriticalCare Medicine.12,13Each patient received appropriate conserv-ative or supportive medical treatments.

Informed consent of patients or responsible family memberswas obtained before beginning the study.

Blood Collection

Blood samples for assays of serum TNF-�, sFas, and sFasLin patients with acute hepatic failure were collected at the startof each plasmapheresis session into endotoxin-free heparinizedblood-specimen tubes. Blood samples in patients with sepsiswere collected when patients were diagnosed. They were im-mediately centrifuged at 1,000 g for 5 minutes and stored at�80°C until use.

From the Department of Emergency and Critical Care Med-icine, Akita University School of Medicine, Akita, Japan; andthe Critical Care and Emergency Center, Iwate Medical Uni-versity, Morioka, Japan.

This study was supported in part by grant 10470311 from theMinistry of Education, Science, and Culture of Japan.

Address reprint requests to Hajime Nakae, MD, Departmentof Emergency and Critical Care Medicine, Akita UniversitySchool of Medicine, 1-1-1 Hondo, Akita 010-8543, Japan.

Copyright © 2001 by W.B. Saunders Company0883-9441/01/1602-0003$35.00/0doi:10.1053/jcrc.2001.25470

Assays

Serum TNF-� levels were determined by enzyme-linked im-munosorbent assay (ELISA) (Biosource Europe SA, Fleurus,Belgium). The detection limit was 3 pg/mL, below which falllevels of normal/healthy persons. sFas and sFasL levels wereboth determined by ELISA (Fujirebio Inc., Tokyo, Japan); as-say sensitivity was 0.5 ng/mL and 0.1 ng/mL, respectively. Nor-mal levels are known to be below these limits.

We assayed 21 blood samples from the acute hepatic failurepatients and 8 from the sepsis patients.

Statistical Analysis

Data are expressed as mean � SD. Differences were evalu-ated statistically with Student’s t test, and data were tested forassociation by calculating Pearson’s correlation coefficients. AP value of less than .05 was considered significant.

RESULTS

No significant differences were observed in sFaslevels between patients with acute hepatic failure(4.9 � 2.2 ng/mL) and patients with sepsis (6.6 �2.8 ng/mL) (P �.1011). However, sFasL levelswere significantly higher in the acute hepatic fail-ure patients (0.68 � 0.42 ng/ml) than in the sepsispatients (0 ng/mL) (P �.0001) (Fig 1).

A significant correlation was observed betweenTNF-� (52.7 � 35.6 pg/mL) and sFas levels in pa-tients with acute hepatic failure (r � 0.657,P �.0008) (Fig 2), and a negative correlation was ob-

served between TNF-� and sFasL levels in patientswith acute hepatic failure (r � �0.454,P � .038)(Fig 3). However, no correlation was observed, be-tween sFas and sFasL levels in this patient group(r � �0.140,P � .5486).

DISCUSSION

Fas and TNF receptor 1 (TNFR-I), two cytokinereceptors present on the cell surface, may bind totheir respective physiologic ligands or agonistic spe-cific antibodies, thereby inducing apoptosis.14,15Fasis a type I membrane protein with a molecularweight of 45 kDa and a structure characteristic ofTNF/nerve growth factor (NGF) superfamily re-ceptors in the extracellular region. It is expressedparticularly by activated lymphocytes, human T-celllymphoma/leukemia virus, human immunodefi-ciency virus, and Epstein-Barr virus, in transformedlymphocytes, and in the liver, an organ consistingof nonlymphocytic tissues.1,2,16In contrast, FasL isa type II membrane protein with a molecular weightof 40 kDa and belongs to the TNF family.17 FasLis expressed in activated T cells, especially in CD8�T cells and CD4� T helper 1 cells, and it may playan important role in cytotoxicity induced by cyto-toxic T lymphocytes.18,19

Fas is a receptor of FasL, which is cut from the

60 NAKAE, NARITA, AND ENDO

Fig 1. sFas and sFasL levels in patients with acute hepatic failure and sepsis. No significant differences are observed in sFas

levels between Groups 1 and 2. sFasL levels are significantly higher in Group 1 than in Group 2. Bar, mean � SD: Group 1, pa-

tients with acute hepatic failure; Group 2, patients with sepsis.

cell membrane through the action of metallopro-tease, is released into the blood in soluble form, andacts as a cytotoxic humoral factor in organs far fromthe production cells. Mice to which anti-Fas anti-bodies were given intraperitoneally were reported todevelop acute liver failure and die within a few hoursafter apoptosis of hepatocytes.2 This is pathologi-cally similar to fulminant hepatitis observed clini-cally following viral hepatitis. Using transgenic miceexpressing hepatitis B (HB) antigens of hepatitis Bvirus, Kondo et al20 demonstrated that cytotoxic Tlymphocytes (CTL) clones exhibited Fas-dependentcytotoxicity when the clones destroyed liver cellspresenting HB antigens. They showed also that he-patocellular damage induced by CTL could be pre-vented by adding solubilized Fas to transgenic micethat express HB antigens. In a mouse model of

lipopolysaccharide-induced liver damage, neutral-ization of FasL is reported to prevent fatality.20,21

Together, findings suggest that FasL plays an im-portant role in the development of hepatitis. Patho-logic findings for the Councilmann body, which ischaracteristic of viral hepatitis, are similar to find-ings for apoptotic cells or nuclear fragments. Fasantigens are also reported to be significantly ex-pressed in liver tissues from patients with chronichepatitis C and B.22,23

Fas is a membrane protein, but its soluble typeis produced by alternate splicing, and sFas inhibitsthe binding of FasL to Fas, negatively controllingapoptosis.24 In this study, patients with acute liverfailure had higher sFas levels than did patients withsepsis, but the difference was not. These levels re-flect the degrees of Fas expression in liver tissueand suggest that a negative feedback mechanisminhibiting severe apoptosis operates in the presenceof liver failure. Processing of sFasL from mem-brane to soluble types is reported to control Fas-mediated apoptosis negatively because sFasL is lesscytotoxic than membrane-type FasL.19

Even if fulminant hepatitis and postoperative he-patic failure were diagnosed acute hepatic failureall together, the severity of illness was thought tobe different. The severity of illness in sepsis wasalso thought to be different according to the patho-physiology. Therefore, there may be the possiblebias effects of multiple samples in a given patient.Considering these points, our patients with acuteliver failure displayed significantly higher sFasLlevels than did the sepsis patients, may reflect theamount of Fas expressed in liver tissue. However,the negative association and negative control ofapoptosis possibly resulted in consumption, al-though no significant difference in sFas was noted.TNF-� is produced as a membrane protein cut bymetalloprotease and is involved in inflammatoryand immunologic reactions as a soluble cyto-kine.25-28 TNF binds to receptors TNFR-I andTNFR-II to form a trimer.29Ando et al30 suggestedthat, in addition to the Fas-mediated antigen-specific cytotoxic mechanism of cytotoxic T lym-phocytes, TNF-� secreted from cytotoxic T lym-phocytes is involved in cytotoxicity in bystanderliver cells. However, Yamada et al31 reported thatTNF signaling via TNFR-I initiated liver regener-ation, and Takehara et al32 suggested that TNF-�acts on the protective factor against Fas-mediatedhepatocyte apoptosis in hepatectomized mice.

sFas AND sFasL IN ACUTE HEPATIC FAILURE 61

Fig 2. Relation between TNF-� and sFas levels. A signifi-

cant correlation is observed. (Y � 10.586X � 0.832, n � 21,

r � 0.657, P �.0008)

Fig 3. Relation between TNF-� and sFasL levels. A nega-

tive correlation is observed. (Y � �38.73X � 78.894, n � 21,

r � �0.454, P �.038)

This study reveals significant positive correla-tion between TNF-� and sFas and a significantnegative correlation between TNF-� and sFasL.Multiple molecular groups may act in the devel-opment of acute liver failure, but results suggestthat the apoptosis of hepatocytes through FasL/Fasassociated with TNF-� acts as an important mech-anism of development. It is not clear why levelsof sFasL, but not sFas, were significantly high inthe patients with acute hepatic failure in compar-ison to the patients with sepsis. TNF-� may act astimulant to produce sFasL in the prevention ofliver injury.

Two hypotheses regarding the relation betweenacute liver failure and apoptosis are as follows:(1) that apoptosis inhibits severe inflammatory re-

actions, and (2) that apoptosis accelerates acuteliver failure. However, it is more reasonable thatapoptosis, unlike necrosis, plays a role in inhibit-ing severe inflammatory reactions when it is in-duced by an intrinsic mechanism. In this case, thescenario may be that it directly results in the deathof excited cells releasing cytokines, such as TNF-�, and that it indirectly prevents further inflamma-tory reactions by inducing early death in the targetcells.

It was unclear in this study whether apoptosisoccurs via a mechanism of the first hypothesis, thesecond, or both, but it does appear that apoptosisis closely associated with the pathology of acuteliver failure and the living body regulate apoptosticreactions through FasL/Fas substances.

62 NAKAE, NARITA, AND ENDO

REFERENCES

1. Galle PR, Hofmann WJ, Walczak H, et al: Involvement ofthe CD95 (APO-1/Fas) receptor and ligand in liver damage. J Exp Med 182:1223-1230, 1995

2. Ogasawara J, Watanabe-Fukunaga R, Adachi M, et al:Lethal effect of the anti-Fas antibody in mice. Nature 364:806-809, 1993

3. Watanabe-Fukunaga R, Brannan CI, Itoh N, et al: ThecDNA structure, expression, and chromosomal assignment ofthe mouse Fas antigen. J Immunol 148:1274-1279, 1992

4. Kagi D, Vignaux F, Ledermann B, et al: Fas and perforinpathways as major mechanisms of T cell-mediated cytotoxicity.Science 265:528-530, 1994

5. Lowin B, Hahne M, Mattmann C, et al: Cytolytic T-cellcytotoxicity is mediated through perforin and Fas lytic path-ways. Nature 370:650-652, 1994

6. Suda T, Okazaki T, Naito Y, et al: Expression of the Fasligand in cells of T cell lineage. J Immunol 154:3806-3813, 1995

7. Leist M, Gantner F, Bohlinger I, et al: Tumor necrosis factor-induced hepatocyte apoptosis precedes liver failure in ex-perimental murine shock models. Am J Pathol 146:1220-1223,1995

8. Ni R, Tomita Y, Matsuda K, et al: Fas-mediated apoptosisin primary cultured mouse hepatocytes. Exp Cell Res 215:332-337, 1994

9. Shinagawa T, Yoshioka K, Kakumu S, et al: Apoptosis incultured rat hepatocytes: The effects of tumour necrosis factoralpha and interferon gamma. J Pathol 165:247-253, 1991

10. Matsubara S, Okabe K, Ouchi K, et al: Continuous re-moval of middle molecules by hemofiltration in patients withacute liver failure. Crit Care Med 18:1331-1338, 1990

11. Sekiyama KD, Yoshiba M, Thomson AW: Circulatingproinflammatory cytokines (IL-1�, TNF-�, and IL-6) and IL-1receptor antagonist (IL-1Ra) in fulminant hepatic failure andacute hepatitis. Clin Exp Immunol 98:71-77, 1994

12. American College of Chest Physicians/Society of Criti-cal Care Medicine Consensus Conference: Definitions for sep-sis and organ failure and guidelines for the use of innovativetherapies in sepsis. Crit Care Med 20:864-874, 1992

13. Bone RC, Balk RA, Cerra FB, et al: Definitions for sep-

sis and organ failure and guilelines for the use of innovativetherapies in sepsis. American College of Chest Physicians/So-ciety of Critical Care Medicine. Chest 101:1644-1655, 1992

14. Nagata S: Apoptosis by death factor. Cell 88:355-365,1997

15. Nagata S, Golstein P: The Fas death factor. Science267:1449-1456, 1995

16. Itoh N, Yonehara S, Ishii A, et al: The polypeptide en-coded by the cDNA for human cell surface antigen Fas can me-diate apoptosis. Cell 66:233-243, 1991

17. Suda T, Takahashi T, Golstein P, et al: Molecular cloningand expression of the Fas ligand, a novel member of the tumornecrosis factor family. Cell 75:1169-1178, 1993

18. Mariani SM, Matiba B, Baumler C, et al: Regulation ofcell surface APO-1/Fas (CD95) ligand expression by metallo-proteases. Eur J Immunol 25:2303-2307, 1995

19. Tanaka M, Suda T, Haze K, et al: Fas ligand in humanserum. Nat Med 2:317-322, 1996

20. Kondo T, Suda T, Fukuyama H, et al: Essential roles ofthe Fas ligand in the development of hepatitis. Nat Med 3:409-413, 1997

21. Adachi M, Suematsu S, Kondo T, et al: Targeted muta-tion in the Fas gene causes hyperplasia in peripheral lymphoidorgans and liver. Nat Genet 11:294-300, 1995

22. Hiramatsu N, Hayashi N, Katayama K, et al: Immuno-histochemical detection of Fas antigen in liver tissue of patientswith chronic hepatitis C. Hepatology 19:1354-1359, 1994

23. Iio S, Hayashi N, Mita E, et al: Serum levels of soluble Fasantigen in chronic hepatitis C patients. J Hepatol 29:517-523, 1998

24. Cheng J, Zhou T, Liu C, et al: Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Sci-ence 263:1759-1762, 1994

25. Endo S, Inada K, Yamada Y, et al: Functional modifica-tion of vascular endothelial cells by cytokines during septicshock. Res Commun Molec Pathol Pharmacol 94:23-38, 1996

26. Nakae H, Endo S, Inada K, et al: Plasma concentrationsof type II phospholipase A2, cytokines and eicosanoids in pa-tients with burns. Burns 21:422-426, 1995

27. Nakae H, Endo S, Inada K, et al: Bound and soluble ad-

hesion molecule and cytokine levels in patients with severeburns. Burns 26:139-144, 2000

28. Yamada Y, Endo S, Inada K, et al: Tumor necrosis fac-tor-� and tumor necrosis factor receptor I, II levels in patientswith severe burns. Burns 26:239-244, 2000

29. Smith CA, Farrah T, Goodwin RG: The TNF receptor su-perfamily of cellular and viral proteins: Activation, costimula-tion, and death. Cell 76:959-962, 1994

30. Ando K, Hiroishi K, Kaneko T, et al: Perforin, Fas/Fasligand, and TNF-alpha pathways as specific and bystander

killing mechanisms of hepatitis C virus-specific human CTL. J Immunol 158:5283-5291, 1997

31. Yamada Y, Kirillova I, Preschon JJ, et al: Initiation ofliver growth by tumor necrosis factor: Deficient liver regenera-tion in mice lacking type I tumor necrosis factor receptor. ProcNatl Acad Sci USA 94:1441-1446, 1997

32. Takehara T, Hayashi N, Mita E, et al: Delayed Fas-mediated hepatocyte apoptosis during liver regeneration in mice:Hepatoprotective role of TNF alpha. Hepatology 27:1643-1651,1998

sFas AND sFasL IN ACUTE HEPATIC FAILURE 63