JOURNAL OF INTERFERON AND CYTOKINE RESEARCH 18:1069-1075 (1998)Mary Ann Liebert, Inc.

Tumor Necrosis Factor-a Production by HumanHepatoma Cell Lines Is Resistant to Drugs

That Are Inhibitory to MacrophagesMEIKE WÖRDEMANN, JOACHIM FANDREY, and WOLFGANG JELKMANN

ABSTRACT

Little is known about the potential of immunomodulatory agents to lower tumor necrosis factor-a (TNF-a)synthesis in tissues of nonmonocytic origin. We studied effects of diverse drugs on the formation of im-munoreactive TNF-a in the human hepatoma cell lines HepG2 and Hep3B, in which TNF-a production was

induced by treatment (3 h incubation periods) with interleukin-1/3 (IL-1/3, 300 pg/ml) or phorbol myristateacetate (PMA, 100 nmol/l). TNF-a production in IL-l/3-stimulated or PMA-stimulated hepatocyte cultureswas not altered following the addition of dihydrocortisone (< 1 /ng/ml), dibutyryl-cAMP (db-cAMP, < 100/¿mol/1), adenosine (< 1 mmol/1), thalidomide (< 25 fig/ml), or cyclosporine (< 300 ng/ml). TNF-a produc-tion was inhibited by taurolidine (> 300 ¡ng/ml), but this inhibition was associated with reduced cell viabil-ity. Pentoxifylline (1 mg/ml) did not influence PMA-induced TNF-a production, but it augmented IL-l/3-in-duced TNF-a production. Measurements of TNF-a mRNA by RT-PCR indicated that pentoxifylline exertedits effect posttranscriptionally. Additional studies with PMA-treated human whole blood cultures confirmedthat pentoxifylline, db-cAMP, and adenosine reduced TNF-a production by leukocytes. These results providefirst evidence to assume cell type-specific effects of immunomodulatory drugs on TFN-a synthesis, which maybe relevant with respect to their clinical application.

INTRODUCTION

Tumor necrosis FACTOR-a (TNF-a) is a pleiotropic cy-tokine that induces metabolic and inflammatory responses

of virtually all types of tissues, including hematopoietic cellsand leukocytes, endothelial cells, fibroblasts, and epithelialcells. The main stimuli of TNF-a production are bacteriallipopolysaccharides (LPS), parasites, viruses, malignant cells,and various cytokines. The action of TNF-a is basically bene-ficial in protecting the organism from infections and malig-nancies. However, overwhelming levels of circulating TNF-ahave been implicated in the pathogenesis of severe diseases,such as sepsis and septic shock, autoimmune processes, AIDS,and cancer cachexia.<12) Therefore, clinical trials have been car-ried out on the administration of anti-TNF-a antibodies and on

drugs that suppress TNF-a synthesis.(3) The agents that can

downregulate the production of TNF-a are highly diverse intheir structure and mechanism of action. They include gluco-corticoids, phosphodiesterase inhibitors, adenosine agonists,taurine analogs, and cyclosporine/2'4'5'

Monocytes and resident macrophages have been consideredthe primary sources of TNF-a. However, recent studies haveclearly shown that the TNF-a gene can be expressed by a va-

riety of nonphagocytic tissues. Apart from tumors, these includeconnective tissue, smooth muscle, and renal epithelium.*2' We(6)and others(7) have demonstrated TNF-a mRNA formation andTNF-a secretion in cultures of the human hepatoma cell lineHepG2 on stimulation with interleukin-1/3 (IL-Iß) or phorbolmyristate acetate (PMA). TNF-a gene expression has also beendemonstrated in the human hepatoma line HUH7,(7> the hu-man/rat hepatocytes coculture line HH25,<7) and primary rat he-patocyte cultures.(8) Further evidence for TNF-a production byhepatocytes has been provided by in situ hybridization of TNF-a mRNA and by immunohistochemical localization of TNF-ain murine liveri9' and in liver sections of patients chronicallyinfected with hepatitis B or hepatitis C virus.(10)

The potential of drugs to inhibit TNF-a synthesis may betissue specific. Thus, we investigated the effects of dihydro-cortisone, pentoxifylline, thalidomide, cyclosporine, and tauro-lidine on TNF-a production in IL-1 ^-stimulated and PMA-

Institute of Physiology, Medical University, Luebeck, Germany.

1069

1070

stimulated HepG2 cultures. This hepatoblastoma line was usedfor study instead of primary hepatocyte cultures because HepG2cells are a well-established cell culture model for regulated he-patic protein synthesis/7,11,12' and contaminating macrophagescan blur results derived from primary hepatocyte cultures.HepG2 cells exhibit the morphologic and functional character-istics of normal hepatocytes. They do not produce hepatitis virusantigen and are not tumorigenic when transplanted in nudemice/13' For the sake of comparison, drug effects were also in-vestigated in the human hepatoma cell line Hep3B and in wholeblood cultures as an established in vitro model for study ofTNF-a synthesis/14,15'

MATERIALS AND METHODS

Cell culturesCells of the human lines HepG2 and Hep3B were obtained

from the American Type Culture Collection (ATCC No. HB8065 and 8064). They were grown in RPMI 1640 medium(Biochrom, Berlin, Germany) supplemented with 10% fetalbovine serum (FBS) (Sigma, Deisenhofen, Germany). The cul-ture medium was renewed two or three times a week and on

the day before the experiments. Confluent cultures in 24-welldishes (Nunc, Wiesbaden, Germany) were incubated in a hu-midified atmosphere of 5% CO2 in air (Heraeus incubators,Hanau, Germany) with the respective agents at 37 °C for 3 hunless otherwise noted. The medium was then removed, andaliquots were stored at

—

20°C for assay of immunoreactiveTNF-a. The concentration of TNF-a was related to cellular pro-tein, which was measured in cell lysates of washed cultures bymeans of a protein microdetermination kit based on the phenolreagent method (Sigma). Cytotoxicity was assessed by the col-orimetric tetrazolium salt/formazan method in confluent 96-well dishes (Nunc)/16) The assay is based on the reduction of3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide(MTT, Sigma) to purple formazan by dehydrogenases in livingcells.

Whole blood cultures were prepared with freshly drawnblood from healthy male volunteers, which was collected intoheparinized endotoxin-free tubes (Monovettes, Sarstedt, Nuern-brecht, Germany)/14,15' Blood was diluted 1:5 in RPMI 1640medium (final volume 500 pi) and incubated in 24-well dishesin the presence of the compounds to be tested for stimulationor inhibition of TNF-a synthesis. After incubation in a humid-ified atmosphere (5% CO2 in air) at 37°C for 6 h, the mediumwas aspirated and centrifuged at 3000g for 1 min. Incubationswere carried out in quadruplicate. Supernatants were frozen at-20°C for assay of TNF-a.

Assay of immunoreactive TNF-a

An immunoenzymometric assay was applied for the quanti-tative measurement of TNF-a in culture medium (TNF-a EA-SIA, Medgenix, Fleuras, Belgium). The assay is based on an

oligoclonal system in which a blend of monoclonal antibodies(mAb) directed against distinct epitopes of human TNF-a isused. The lower detection limit of the assay was 3 pg/ml. Theintraasssay and interassay coefficients of variance were < 5%and < 10%, respectively.

WÖRDEMANN ET AL.

1 TNF-a mRNA quantitationI

At the end of the incubation period, cells were washed with1 ml ice-cold phosphate-buffered saline (PBS), pH 7.4, andsolubilized with 700 pi guanidinium thiocyanate solution (4mol/1 guanidinium thiocyanate in PBS with 0.1 mol/1 /3-mer-captoethanol). RNA was extracted using the acid phenolmethod/17' RNA was redissolved in diethylpyrocarbonate-treated water, and its concentration was determined by mea-

surement of the OD at 260 nm. Total RNA (1 pg) was usedin the reverse transcription (RT) reaction with 0.5 pg of oligo-dT (MWG Biotech, Ebersberg, Germany) and 200 U of M-MLV reverse transcriptase (Promega, Heidelberg, Germany)according to the manufacturer's instructions. Polymerasechain reaction (PCR) for TNF-a was performed with upstreamprimer 5'-CTC TGG CCC AGG CAG TCA GA-3' and down-stream primer 5'-GGC GTT TGG GAA GGT TGG AT-3' inPCR buffer (50 mmol/1 Tris-HCl, pH 8.3, 50 mmol/1 KC1, 1.5

1 mmol/1 MgCl2, 0.01 vol% gelatin), 200 pmol dNTPs' (Promega), 20 pmol of each primer, and 0.2 U Taq polymerase1 (Gibco) for 30 cycles (temperature profile 94°C 1 min, 57°C' 1.5 min, 72°C 3 min). The resulting PCR fragment had a size-

of519bp.1 Competitive PCR was performed exactly as described/6'' Briefly, the competitor was constructed according to Siebert

and Larrick/18' Using the primers for PCR under exactly thesame conditions as for cDNA amplification, a 358 bp PCR

1 product was obtained. For quantitation of cDNA, equal1 amounts (10 pi) of the RT reaction (diluted 1:10 in water); and of solutions with increasing concentrations of competi--

tor DNA (0.004 amol//ri [amol = 10"18 mol] to 0.4 amol//ri)' were coamplified in the same reaction tube. PCR products1 were run on a 3% agarose gel and made visible by ethidium

bromide (0.5 pg/ml) staining. PCR products derived fromcDNA have a length of 519 bp, whereas the amplified com-

t petitor lights up as a band of 358 bp. The equilibrium of the; intensity of the bands, that is, the competitor concentrationI at which the same amount of cDNA was put into the reac-

tion, was determined by visual inspection and measurement1 of the band intensity by means of the E.A.S.Y. software> (Herolab, Wiesloch, Germany). Results are expressed as

TNF-a mRNA in amol//xg total RNA, taking into consider-' ation the efficiency of the reverse transcription reaction (i.e.,* 35%)/19'1

1 Chemical and pharmacologie agentsHuman recombinant IL-1/3 (Ciba-Geigy, Basel, Switzer-

land), PMA, and bacterial LPS (from Escherichia coli,serotype 0.111 :B4, Sigma) were used as potential stimula-tors of TNF-a production. The following agents were usedas potential inhibitors of TNF-a production: dihydro-cortisone, N6,2'-0-dibutyryladenosine 3':5'-cyclic mono-

phosphate (db-cAMP), pentoxifylline, adenosine, N6-1 cyclopentyladenosine (CPA, Ai receptor agonist), 5'-N-eth-s ylcarboxamidoadenosine (NECA, A2 receptor agonist) (alls from Sigma), thalidomide (kindly provided by Grünenthal,s Stolberg, Germany), taurolidine (kindly provided byb Geistlich AG, Wolhusen, Germany), and cyclosporine (San-

doz, Basel, Switzerland).

TNF-a PRODUCTION BY HEPATOCYTES 1071

Statistical analysisResults are given as mean values ± standard deviation (SD).

Dunnett's test was used to calculate significant differences be-tween a control mean value and several treatment means of sep-arate hepatocyte cultures. Friedman's nonparametric repeatedmeasures test followed by Dunn's multiple comparisons testwas used to calculate significant differences in matched wholeblood cultures. P was considered significant at the 5% level.

RESULTS

Figure 1 shows that the addition of IL-Iß dose-dependentlystimulated the production of TNF-a in HepG2 and Hep3B cul-tures. The maximum TNF-a concentration in the culturemedium was reached after 3 h, indicating that the stimulationof TNF-a production was transient (Fig. 2). The addition ofPMA to the cells also resulted in a dose-dependent stimulationof TNF-a production. The range of the dose-response curve

was very narrow, with no effect at 30 nmol/1 and a maximumeffect at > 100 nmol/1 PMA. LPS (tested up to 1 pg/ml) didnot induce TNF-a synthesis in hepatocyte cultures. Based on

these findings, subsequent studies on the effects of im-munomodulating drugs were carried out in cells stimulated witheither 300 pg/ml IL-Iß or 100 nmol/1 PMA for 3 h.

Apart from IL-Iß and PMA, none of the compounds usedin this study induced TNF-a production when added alone to

HepG2 or Hep3B cultures. In addition, the following com-

pounds failed to influence the 3 h rates of the production of

TNFa(pg/mg protein)1000

100 300 1000 5000IL-1ß (pg/ml)

FIG. 1. Dose dependency of TNF-a production in HepG2(circles) and Hep3B (triangles) cultures treated with IL-Iß for3 h. TNF-a concentrations in culture medium are related to cel-lular protein (mean ± SD of four parallel cultures). *P < 0.05compared with untreated controls, Dunnett's test.

TNFa(pg / mg protein)

700 A

600

500

400 H

300

200 H

100 H

*

1 2 3Incubation period (h)

FIG. 2. Time course of TNF-a accumulation in HepG2 cul-tures treated with IL-Iß (300 pg/ml) (mean ± SD of four par-allel cultures). *P < 0.05 compared with 0 h, Dunnett's test.

TNF-a in IL-l/3-stimulated or PMA-stimulated cultures: dihy-drocortisone (tested up to 1 pg/ml, acutely added or followinga 3 h preincubation period), db-cAMP (up to 100 pmolft),thalidomide (up to 25 pg/ml), cyclosporine (up to 300 ng/ml),adenosine, and its receptor-specific analogs NCPA and NECA(all tested up to 1 mmol/1).

Taurolidine produced a dose-dependent decrease in the rateof TNF-a production in IL-l/3-treated HepG2 cultures, whichwas significant at 300 pg/ml. However, this drug was cytotoxic(50% and 100% cell killing at 500 and 1000 pg/ml taurolidine,respectively). Note that none of the other compounds used was

toxic at the doses used, when tested by MTT assay.Pentoxifylline did not inhibit TNF-a production in IL-1/3-

treated or PMA-treated hepatoma cells. Instead, at the highestdose tested (1000 pg/ml), pentoxifylline enhanced the effect ofIL-Iß (Fig. 3). Adenosine did not reverse the increase in pen-toxifylline-treated cells (Fig. 4). Pentoxifylline did not increaseTNF-a production in PMA-treated cells.

In a qualitative PCR, HepG2 and Hep3B cells showed a

marked increase in TNF-a mRNA levels when treated with IL-1/3 (300 pg/ml) for 3 h. The addition of pentoxifylline (1000pg/ml) did not further increase TNF-a mRNA levels. TNF-amRNA was quantitated in HepG2 cells as described in Materi-als and Methods. Control cells contained very little TNF-amRNA (0.45 ±0.18 amol//¿g total RNA) (n = 4), whereascells treated with IL-1/3 alone or together with pentoxifyllineexpressed 33.4 ± 9.7 amol TNF-a mRNA/pg total RNA (n =

4 for each treatment).Because several of the compounds that are administered clin-

ically to suppress TNF-a production were ineffective in hepa-

1072 WÖRDEMANN ET AL.

TNFa(pg / mg protein)1400 -i

1200

1000

800

600 H

400

200 -\

100 1000 Pentoxifylline(pg/ ml)

FIG. 3. Effects of pentoxifylline on 3 h rates of the production of TNF-a in HepG2 cultures stimulated with 300 pg/ml IL-1/3(mean ± SD of four parallel cultures). *P < 0.05 compared with cultures treated with IL-1/3 alone, Dunnett's test.

TNFa(pg/mg protein)

1600 -\

1200 H

800

400

IL-1ß IL-lß+ POF(1 mg/ml)

IL-1ß IL-1ß IL-1ß IL-1ß+ POF (100 ug/ml)+ POF (500 ug/ml) + POF (1 mg/ml)

+ Adenosine + Adenosine + Adenosine + Adenosine

FIG. 4. Effects of pentoxifylline (POF) on 3 h rates of the production of TNF-a in HepG2 cultures treated with IL-1/3 (300pg/ml) alone or in combination with 1 mmol/1 adenosine (mean ± SD of four parallel cultures). *P < 0.05 compared with cul-tures treated with IL-1/3 alone, Dunnett's test.

TNF-a PRODUCTION BY HEPATOCYTES 1073

TNFa (pg/ml)2000

Control +IL-1 ß + LPS +PMA

FIG. 5. Effects of IL-1/3 (300 pg/ml), LPS (1 pg/ml), andPMA (100 nmol/1) on the 6 h rates of production of TNF-a inhuman whole blood cultures (20% blood in medium RPMI1640) (mean ± SD of blood samples from five differentdonors). *P < 0.05 compared with untreated control cultures,Dunn's test.

TNFa (pg/ml)1200

tocyte cell cultures, comparative experiments were carried outwith whole blood cultures. Incubation periods of 6 h were re-

quired for significant TNF-a production after the addition ofPMA, IL-1/3, or LPS (Fig. 5). Pentoxifylline, adenosine, anddb-cAMP proved to inhibit TNF-a production in PMA-treatedwhole blood cultures (Fig. 6).

DISCUSSION

Evidence has been accumulated to postulate that all cell typescan express the TNF-a gene under appropriate conditions/2'This report shows that human hepatoma cultures of the linesHepG2 and Hep3B secrete TNF-a on stimulation with IL-1/3or PMA. TNF-a production by Hep3B cells is a novel obser-vation, but it is in line with results from studies with other hu-man hepatocyte cell lines/6,7,20' Taken together with the ex vivodemonstration of TNF-a and its mRNA in murine(9' and hu-man liver/10' it seems likely that hepatocytes contribute to TNF-a synthesis in systemic and local inflammatory reactions.

In view of clinical trials of the administration of drugs be-lieved to lower TNF-a formation/2-5' it was of interest to studyeffects of these drags on TNF-a production in hepatocyte cul-tures. In vitro TNF-a synthesis by macrophages is inhibited byvarious compounds, including glucocorticoids/21,22' pentoxi-fylline/21"24' cAMP/25' adenosine/26,27' thalidomide/28' andcyclosporine/22' In HepG2 and Hep3B cultures, however, none

1000

800-^

600

400 H

200

PMAalone

FIG. 6. Effects of immunomodulatory drugs on the 6 h rates of the production of TNF-a in human whole blood cultures stim-ulated with PMA (100 nmol/1). POF, pentoxifylline 1 mg/ml; DHC, dihydrocortisone 1 pg/ml; Ado, adenosine 1 mmol/1; db-cAMP, 100 /xmol/l; Thali, thalidomide 25 pg/ml (mean ± SD of blood samples from five different donors). *P < 0.05 com-

pared with cultures treated only with PMA, Dunn's test.

1074 WÖRDEMANN ET AL.

of these compounds suppressed TNF-a production, induced bytreatment with IL-Iß or PMA. On the other hand, our con-comitant studies with freshly drawn human blood confirmedthe suppressive effects of pentoxifylline, adenosine, and db-cAMP on TNF-a production by leukocytes. Our failure to in-hibit TNF-a synthesis by dihydrocortisone and thalidomide inwhole blood cultures is not in conflict with previous reports.Glucocorticoids inhibit only when used in extremely high con-centrations/22' Thalidomide can be either agonistic or antago-nistic to stimulators of TNF-a synthesis depending on the typeof leukocytes under study/28-30'

Based on the observation that thalidomide enhances thePMA-induced production of TNF-a in human leukemia celllines to a different degree, Nishimura et al/29' have proposedthat the effects of PMA are cell type specific. This concept isextended in the present studies using human hepatocyte cul-tures and various immunomodulatory agents. The most strik-ing tissue specificity was seen with respect to the action of pen-toxifylline. Although this drug proved to inhibit TNF-aproduction in human whole blood cultures, high doses of pen-toxifylline stimulated TNF-a production in hepatocyte cultures.Possibly, tissue-specific differences at the site of the inductionof transcription of the TNF-a gene account for the fact that theadministration of pentoxifylline in humans resulted in signifi-cantly lowered plasma TNF-a levels in some studies'31' but notin others/32'

The biochemical mechanism of the action of pentoxifyllineis poorly understood. Xanthine derivatives generally exert theireffects via modulation of cAMP metabolism/33,34' At high con-

centrations, they inhibit cAMP-phosphodiesterase. At low con-

centrations, they act as adenosine antagonists and thus inhibit(Ai receptor) or stimulate (A2 receptor) adenylate cyclase ac-

tivity. As pentoxifylline, adenosine, and db-cAMP all inhibitedPMA-induced TNF-a production in whole blood cultores,cAMP could be the common mediator of this inhibition. In con-trast, in hepatocyte cultures, TNF-a production must be con-trolled via a different pathway because db-cAMP and adeno-sine were without effect, whereas pentoxifylline stimulatedhere. In addition, exogenous adenosine did not antagonize theeffect of pentoxifylline. It has been reported that pentoxifyllinecan inhibit protein kinase C (PKC)/35' However, pentoxifyllinedid not antagonize the effect of the PKC-activating phorbol es-ter, PMA, on TNF-a synthesis in hepatocyte cultures.

In conclusion, our results show that immunomodulatorydrugs reduced TNF-a synthesis in HepG2 and Hep3B cells lesspotently than in leukocytes. Moreover, pentoxifylline enhancedIL-1/3 production in hepatocyte cultures. It is of interest to studythe tissue specificity of the effects of immunosuppressive com-

pounds in other cells capable of TNF-a synthesis/2' such as fi-broblasts, smooth muscle cells, and astrocytes.

REFERENCES

1. TRACY, K.J., and CERAMI, A. (1994). Tumor necrosis factor: a

pleiotropic cytokine and therapeutic target. Annu. Rev. Med. 45,491-503.

2. AGGARWAL, B.B., and NATARAJAN, K. (1996). Tumor necro-sis factors: developments during the last decade. Eur. CytokineNetw. 7, 93-124.

3. BEUTLER, B., and KRUYS, V. (1995). Lipopolysaccharide sig-nal transduction, regulation of tumor necrosis factor biosynthesis,and signaling by tumor necrosis factor itself. J. Cardiovasc. Phar-macol. 25 (Suppl. 2), S1-S8.

4. KLAUSNER, J.D., FREEDMAN, V.H., and KAPLAN, G. (1996).Thalidomide as an anti-TNF-a inhibitor: implications for clinicaluse. Clin. Immunol. Immunopafhol. 81, 219-223.

5. BEDROSIAN, I., SOFIA, R.D., WOLFF, S.M., and DINA-RELLO, C.A. (1991). Taurolidine, an analogue of the amino acidtaurine, suppresses interleukin 1 and tumor necrosis factor syn-thesis in human peripheral blood mononuclear cells. Cytokine 3,568-575.

6. FREDE, S., FANDREY, J., and JELKMANN, W. (1996). Inter-leukin Iß and phorbol ester induce tumour necrosis factor a pro-duction in a hepatic cell line (HepG2). Pflügers Arch. Eur. J. Phys-iol. 431, 923-927.

7. ROWELL, D.L., ECKMANN, L., DWTNELL, M.B., CARPEN-TER, S.P., RAUCY, J.L., YANG, S.-K., and KAGNOFF, M.F.(1997). Human hepatocytes express an array of proinflammatorycytokines after agonist stimulation or bacterial invasion. Am. J.Physiol. 273, G322-G332.

8. HORBACH, M., GERBER, E., and KAHL, R. (1997). Influenceof acetaminophen treatment and hydrogen peroxide treatment on

the release of a CTNC-related protein and TNF-a from rat hepato-cyte cultures. Toxicology 121, 117-126.

9. HUNT, J.S., CHEN, H.-L., HU, X.-L., CHEN, T.-Y., and MOR-RISON, D.C. (1992). Tumor necrosis factor-a gene expression inthe tissues of normal mice. Cytokine 4, 340-346.

10. GONZÁLEZ-AMARO, R., GARCÍA-MONZÓN, C, GARCÍA-BUEY, L., MORENO-OTERO, R., ALONSO, J.L., YAGÜE, E.,PTVEL, J.P., LÓPEZ-CARBRERA, M„ FERNÁNDEZ-RUIZ, E.,and SÁNCHEZ-MADRID, F. (1994). Induction of tumor necrosisfactor a production by human hepatocytes in chronic viral hepati-tis. J. Exp. Med. 179, 841-848.

11. DARLINGTON, G.J., WILSON, D.R., and LACHMAN, L.B.(1986). Monocyte-conditioned medium, interleukin-1, and tumornecrosis factor stimulate the acute phase response in human he-patoma cells in vitro. J. Cell Biol. 103, 787-793.

12. KELLY, J.H., and DARLINGTON, G.J. (1989). Modulation of theliver specific phenotype in the human hepatoblastoma line HepG2.In Vitro Cell. Dev. Biol. 25, 217-222.

13. KNOWLES, B.B., HOWE, C.C., and ADEN, D.P. (1980). Humanhepatocellular carcinoma cell lines secrete the major plasma pro-teins and hepatitis B surface antigen. Science 209, 497^199.

14. WILSON, B.M.G., SEVERN, A., RAPSON, N.T., CHANA, J.,and HOPKINS, P. (1991). A convenient human whole blood cul-ture system for studying the regulation of tumor necrosis factor re-lease by bacterial lipopolysaccharide. J. Immunol. Methods 139,233-240.

15. DE GROÓTE, D., ZANGERLE, P.F., GEVAERT, Y., FAS-SOTTE, M.F., BEGUIN, Y., NOIZAT-PTRENNE, F., PIRENNE,J., GATHY, R., LOPEZ, M., DEHART, I., IGOT, D., BAUDRI-HAYE, M., DELACROTX, D., and FRANCHIMONT, P. (1992).Direct stimulation of cytokines (¡L-Iß, TNFa, TL-6, TL-2, JFNyand GM-CSF) in whole blood. I. Comparison with isolated PBMCstimulation. Cytokine 4, 239-248.

16. HANSEN, M.B., NIELSEN, S.E., and BERG, K. (1989). Re-ex-amination and further development of a precise and rapid dyemethod for measuring cell growth/cell kill. J. Immunol. Methods119, 203-210.

17. CHOMCZYNSKI, P., and SACCHI, N. (1987). Single-step methodof RNA isolation by acid guanidinium thiocyanate-phenol-chloro-form extraction. Anal. Biochem. 162, 156-159.

18. STEBERT, P.D., and LARRICK, J.W. (1992). Competitive PCR.Nature 359, 557-558.

19. FANDREY, I., and BUNN, H.F. (1993). In vivo and in vitro reg-

TNF-a: PRODUCTION BY HEPATOCYTES

ulation of erythropoietin mRNA: measurement by competitivepolymerase chain reaction. Blood 81, 617-623.

20. FREDE, S., DEETJEN, C, SCHOBERSBERGER, W., and HOFF-MANN, G. (1998). Granulocyte-colony stimulating factor inhibitsTNFa production in a human hepatoma cell line. Pflügers Arch.Eur. J. Physiol. 436, 233-237.

21. HAN, J., THOMPSON, P., and BEUTLER, B. (1990). Dexa-methasone and pentoxifylline inhibit endotoxin-induced cachectin/tumor necrosis factor synthesis at separate points in the signalingpathway. J. Exp. Med. 172, 391-394.

22. FUNK, J.O., ERNST, M., SCHÖNHARTTNG, M.M., and ZABEL,P. (1995). Pentoxifylline exerts synergistic immunomodulatory ef-fects in combination with dexamethasone or cyclosporin A. Int. J.Immunopharmacol. 17, 1007-1016.

23. DOHERTY, G.M., JENSEN, J.C., ALEXANDER, H.R., BU-RESH, CM., and NORTON, J.A. (1991). Pentoxifylline suppres-sion of tumor necrosis factor gene transcription. Surgery 110,192-198.

24. VAN FURTH, A.M., VERHARD-SEUMONSBERGEN, E.M.,VAN FURTH, R., and LANGERMANS, J.A.M. (1997). Effect oflisofylline and pentoxifylline on the bacterial-stimulated produc-tion of TNF-a, TU-Iß and IL-10 by human leucocytes. Immunol-ogy 91, 193-196.

25. OH-ISHI, S., UTSUNOMIYA, I., YAMAMOTO, T., KOMURO,Y., and HARÁ, Y. (1996). Effects of prostaglandins and cyclicAMP on cytokine production in rat leukocytes. Eur. J. Pharmacol.300, 255-259.

26. SAJJADI, F.G., TAKABAYASHI, K, FOSTER, A.C.,DOMINGO, R.C., and FTRESTEIN, G.S. (1996). Inhibition ofTNF-a expression by adenosine. J. Immunol. 156, 3435-3442.

27. McWHTNNEY, CD., DUDLEY, M.W., BOWLTN, TL., PEET,N.P., SCHOOK, L., BRADSHAW, M., DE, M., BORCHERDTNG,D.R., and EDWARDS, C.K., Tfl. (1996). Activation of adenosineA3 receptors on macrophages inhibits tumor necrosis factor-a. Eur.J. Pharmacol. 310, 209-216.

28. SAMPAIO, E.P., SARNO, E.N., GALTLLY, R., COHN, Z.A., andKAPLAN, G. (1991). Thalidomide selectively inhibits tumornecrosis factor a production by stimulated human monocytes. J.Exp. Med. 173, 699-703.

29. NISHIMURA, K., HASHIMOTO, Y., and IWASAKI, S. (1994).

1075

Enhancement of phorbol ester-induced production of tumor necro-sis factor a by thalidomide. Biochem. Biophys. Res. Commun. 199,455^*60.

30. SHANNON, E.J., and SANDOVAL, F. (1996). Thalidomide canbe either agonistic or antagonistic to LPS evoked synthesis of TNF-a by mononuclear cells. Immunopharmacol Immunotoxicol. 18,59-72.

31. ZABEL, P., SCHÖNHARTTNG, M.M., WOLTER, D.T., andSCHADE, U.F. (1989). Oxpentifylline in endotoxaemia. Lancet 2,1474-1477.

32. WALLIS, R.S., NSUBUGA, P., WHALEN, C, MUGERWA,R.D., OKWERA, A., OETTE, D., JACKSON, J.B., JOHNSON,J.L., and ELLNER, J.J. (1996). Pentoxifylline therapy in humanimmunodeficiency virus-seropositive persons with tuberculosis: a

randomized, controlled trial. J. Infect. Dis. 174, 727-733.33. SCHWABE, U., UKENA, D., and LOHSE, M.J. (1985). Xanthine

derivatives as antagonists at Ai and A2 adenosine receptors.Naunyn-Schmiedebergs Arch. Pharmacol. 330, 212-221.

34. PELLEG, A., and PORTER, R.S. (1990). The pharmacology ofadenosine. Pharmacofherapy 10, 157-174.

35. BISWAS, D.K., AHLERS, CM., DEZUBE, B.J., and PARDEE,A.B. (1994). Pentoxifylline and other protein kinase C inhibitorsdown-regulate HTV-LTR NF-kB induced gene expressin. Mol.Med. 1, 31-43.

Address reprint requests to:Dr. Wolfgang JelkmannInstitute of Physiology

Medical University of LuebeckRatzeburger Allee 160

D-23538 LuebeckGermany

Tel: 0049-451-500 4150Fax: 0049-451-500 4151

E-mail: jelkmann@physio,mu-luebeck. de

Received 29 July 1998/Accepted 27 August 1998

This article has been cited by:

1. Jen-Eing Jeng, Huey-Ru Tsai, Lee-Yea Chuang, Jung-Fa Tsai, Zu-Yau Lin, Min-Yuh Hsieh, Shin-Chern Chen, Wan-LungChuang, Liang-Yen Wang, Ming-Lung Yu, Chia-Yen Dai, Jan-Gowth Chang. 2009. Independent and Additive InteractiveEffects Among Tumor Necrosis Factor-α Polymorphisms, Substance Use Habits, and Chronic Hepatitis B and Hepatitis CVirus Infection on Risk for Hepatocellular Carcinoma. Medicine 88, 349-357. [CrossRef]

2. Helen Burmester, Eva-Maria Wolber, Patricia Freitag, Joachim Fandrey, Dr. Wolfgang Jelkmann. 2005. ThrombopoietinProduction in Wild-Type and Interleukin-6 Knockout Mice with Acute Inflammation. Journal of Interferon & CytokineResearch 25:7, 407-413. [Abstract] [Full Text PDF] [Full Text PDF with Links]

3. Takeya Tsutsumi, Tetsuro Suzuki, Kyoji Moriya, Hiroshi Yotsuyanagi, Yoshizumi Shintani, Hajime Fujie, YoshiharuMatsuura, Satoshi Kimura, Kazuhiko Koike, Tatsuo Miyamura. 2002. Alteration of Intrahepatic Cytokine Expression andAP-1 Activation in Transgenic Mice Expressing Hepatitis C Virus Core Protein. Virology 304, 415-424. [CrossRef]

4. K. SUZUKI, M. YANAGI, A. MORI-AOKI, E. MORIYAMA, K. J. ISHII, L. D. KOHN. 2002. Transfection of single-stranded hepatitis A virus RNA activates MHC class I pathway. Clinical & Experimental Immunology 127, 234-242. [CrossRef]

5. D. Śladowski, A. Kinsner, I. Langezaal, S. Kay, S. Coecke. 2001. Activation of the complement system as an indicator ofpyrogenic reaction to lipopolysaccharide (LPS). Toxicology in Vitro 15, 339-342. [CrossRef]