- Email: [email protected]
TNF GENE EXPRESSION IN MONOCYTES OF LOW AND HIGH RESPONDER INDIVIDUALS
TNF GENE EXPRESSION IN MONOCYTES OF LOW AND HIGH RESPONDER INDIVIDUALS Winfried Schraut, Petra Wendelgass, Julia C. Calzada-Wack, Marion Frankenberger, H. W. Lo¨ms Ziegler-Heitbrock Individuals with a consistently lower immune response may be more susceptible to infection but less prone to autoimmune disease or severe sepsis. The molecular mechanisms determining the low responder status are, however, unclear. We have screened 60 male donors for tumour necrosis factor (TNF) protein levels after stimulation of monocytes with lipopolysaccharide (LPS). Among these we identified three donors each that consistently had a level of less than 20% (low responders; LR) or of more than 80% (high responders; HR) of the maximum response seen in this population. Northern blot analysis of TNF mRNA after LPS stimulation revealed lower transcript levels in LR. Half life determination after actinomycin D blockade showed a similar decay rate for LR and HR and after blockade of degradation by cycloheximide treatment mRNA levels increased but LR remained lower compared to HR. These data indicate that the lower TNF mRNA levels in LR are not due to a more rapid mRNA degradation but rather a result of lower transcription. Transcripts for interleukin 6 (IL-6) were also low in LPS-stimulated monocytes of LR. Because expression of the LPS receptor CD14 was similar in LR and HR monocytes, our data suggest that differences in signal transduction account for the LR and HR phenotype. 7 1997 Academic Press Limited
Cytokines are crucial to immune defence because they determine the type and extent of response. One prominent producer of cytokines is the monocyte/ macrophage that can release excessive amounts of mediators like tumour necrosis factor (TNF) and interleukin 6 (IL-6) after being challenged with lipopolysaccharide (LPS) from Gram-negative bacteria. The amount of cytokines produced may vary with the individual with some showing high, and other’s low levels of production. The extent of cytokine production may determine the outcome of inflammatory diseases. An individual with a low level of production (low responder, LR) may, for instance, be less prone to autoimmune disease, to transplant rejection or to severe sepsis. On the other hand, a LR may have difficulty in coping with intracellular parasites like leishmania. Hence, it is clear that the responder status may have clinical implications. The molecular mechanisms From the Institute for Immunology, University of Munich, Goethestr. 31, 80336 Mu¨nchen, Germany Correspondence to: H.W.L. Ziegler-Heitbrock, Institute for Immunology, University of Munich, Goethestr. 31, 80336 Mu¨nchen Received 11 October 1995; revised and accepted for publication 24 June 1996 7 1997 Academic Press Limited 1043-4666/97/030206 + 6 $25.00/0/ck960155 KEY WORDS: Monocyte TNF/LPS 206
governing the different levels of cytokine expression in LR and high responders (HR) have, however, not been analysed and it is unclear whether the degree of expression is controlled at the level of receptors, of transcription, of mRNA degradation, translation or protein export. In the present report the authors have addressed this issue and demonstrated in selected LR and HR that the responder status is controlled at the transcription level.
RESULTS When screening 60 male donors for LPSstimulated TNF production, we detected three donors who produced consistently low and three who produced consistently high levels of this cytokine. As shown in Figure 1, immunoreactive TNF detected by enzyme-linked immunoabsorbent assay (ELISA) in cell free plasma of LPS-stimulated whole blood was below 4 ng/ml in the low responders (LR) and above 13 ng/ml in the high responders (HR). These low and high levels of TNF protein production were observed on 3 or more occasions over a period of more than 3 months (not shown). Exchange of plasma did not change this pattern. As shown in Figure 2, addition of plasma from either LR or HR to blood cells after repeated washing did not CYTOKINE, Vol. 9, No. 3 (March), 1997: pp 206–211
Low and high responses for TNF/ 207
25
A
20
15
10
5 Immunoreactive TNF (ng/ml)
affect the LR or HR phenotype of the cells. These data suggest that plasma factors like, for instance, soluble TNF receptors are not responsible for the responder status. Rather LR monocytes appear to produce less TNF. Therefore, TNF mRNA levels in LPS-stimulated monocytes have been studied. Figure 3 demonstrates that the two LR tested had low levels of LPS-induced mRNA, while TNF transcripts were high in the three HR at 1 h post LPS. One might speculate that LR have a slower kinetic of transcription with a peak at a later point in time. Therefore time course studies were performed. Results in Figure 4 demonstrate high levels of TNF transcripts in HR already at 0.5 h, and this signal is almost completely gone at 5 h. In LR TNF mRNA was detectable at 0.5 h as well, but levels were low at this and at all other time points. Hence, the LR phenotype does not appear to be due to a shift in the kinetics. TNF mRNA levels may be lower in LR because of lower transcription or because of more rapid
0 Cells : LR Plasma : LR
LR HR
HR HR
HR LR
HR HR
HR LR
Exp. 1 25
B
20
15 15
Immunoreactive TNF (ng/ml)
10
5
10
LR HR Exp. 2
Figure 2.
5
Figure 1.
0 Cells : LR Plasma : LR
TNF production in low and high responders
Heparinized whole blood samples were stimulated with LPS (1 mg/ml) for 4 h and TNF was determined by ELISA. The pattern shown was observed repeatedly on two or three occasions over a period of more than 3 months. In LR1 (6) TNF values were 22 2 4% of the respective HR, in LR2 (x) they were 23 2 5% and in LR3 (w) they were 7 2 1.5%.
The responder status is independent of plasma factors
Heparinized blood from low and high responders was separated into cells and plasma, the cells were washed once with culture medium and cells and plasma were recombined as given in the figure. After LPS stimulation for 4 h at 1 mg/ml supernatant plasma was tested for TNF by ELISA. Given are results from two HR–LR combinations (A, B). Similar results were obtained in separate experiments using a TNF bioassay as a read-out system.
degradation. To resolve this question, we have blocked degradation by using treatment with the protein synthesis inhibitor cycloheximide (CHX). Such treatment enhanced TNF transcripts in LR at all points in time (Fig. 5), but in the HR CHX resulted in much higher levels. The authors next studied mRNA degradation directly by blockade of further transcription with actinomycin D followed by determination of mRNA half life. TNF mRNA production continued
208 / Schraut et al.
LPS (1µg/ml)
–
CYTOKINE, Vol. 9, No. 3 (March, 1997: 206–211)
+
–
+
–
+
–
+
–
+
LR 0 0.5 1 2 4 6
HR 0 0.5 1 2 4 6
(h)
0.5 1 2 4 6
0.5 1 2 4 6
(h)
LPS
TNF transcripts in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for 2 h, adherent cells were recovered and isolated RNA (10 mg/lane) was analysed by Northern blot with TNF specific oligonucleotides (lanes 1–4 = 2 low responders, lanes 5–10 = 3 high responders). The 3rd low responder was found to have low TNF mRNA in an independent experiment (not shown).
somewhat after addition of actinomycin D in the low responder because the signal was higher at 45 min. compared to 30 min, while no further increase of mRNA was apparent in the HR (Fig. 6A). From 45 min on TNF mRNA decayed rapidly in both LR and HR to low levels at 2 h. Densitometry and calculation of half life revealed a t1/2 of 27 min for the HR and 24 min for the LR (Fig. 6B). Taken together, the lower levels of LPS induced TNF mRNA in LR monocytes appears not to be due to more rapid degradation, but rather due to reduced transcription. These data would suggest that signal transduction must be different between LR and HR. In studies not shown, we found no differences in cell surface expression of the CD14 LPS receptor. Furthermore, no difference could be detected in binding of nuclear factors C/EBP and NF-kB to motifs in the human TNF promoter. Finally, we asked whether expression of other LPS-induced cytokines was also different between LR and HR. Northern blot analysis of LPS induced IL-6 mRNA revealed low transcription levels in two LR and high levels in two HR, while hybridization of the same blots for the house-keeping gene glyceraldehyde-
LR 0 0.5 1 2
4
5
HR 0 0.5 1 2
4
5
LPS+ CHx
Figure 5. Effect of cycloheximide on TNF transcripts in low and high responders Mononuclear cells were stimulated with LPS (1 mg/ml) for various periods of time with and without addition of cycloheximide (CHX) at 10 mg/ml. Adherent cells were recovered, RNA was isolated and analysed for TNF transcripts by Northern blotting. The same pattern was seen in two additional pairs of low and high responders. Transcript levels estimated by densitometry were in average of the 3 experiments 2.5 2 0.6 higher in HR without and 1.8 2 0.3 with CHX at the 1 h time point.
A
(h)
Time course of TNF mRNA in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for various periods of time and adherent cells were recovered and RNA was analysed for TNF by Northern blotting. The same pattern was seen in a second pair of LR and HR.
HR ACT D
0 30 45 60 90 120 0 30 45 60 90 120
Figure 6. Figure 4.
LR ACT D
TNF mRNA (arbitrary densitometry units)
Figure 3.
200
B
(min)
t1/2 = 27 min
150 100
50 40 30
t1/2 = 27 min
20
15
30 60 Time (min)
120
Half life of TNF mRNA in low and high responders
(A) Mononuclear cells were stimulated with LPS (1 mg/ml) and after 30 min. Actinomycin D (ActD) was added at 10 mg/ml. Adherent cells were recovered, RNA isolated and analysed for TNF transcripts by Northern blotting. (B) Densitometry was performed for time points 15, 30 and 60 min post ActD (=45, 60, 90 min in A). The same pattern was seen with a second pair of low and high responders.
Low and high responses for TNF/ 209
LR
HR
LR
HR
IL-6
GAPDH
Figure 7.
IL-6 transcripts in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for 2 h adherent cells were recovered and isolated RNA (10 mg/lane) was analysed by Northern blotting for IL-6 transcripts.
3-phosphate dehydrogenase (GAPDH) gave similar levels (Fig. 7).
DISCUSSION TNF was originally defined based on its antitumour activity, but it later turned out to be a major pro-inflammatory cytokine that affects almost every cell of the body.1 TNF was shown to be required for an efficient defence in many infectious diseases including enterobacterial, mycobacterial and leishmanial infections.2–6 On the other hand, TNF may have detrimental effects because excessive production of this cytokine has been associated with severe sepsis7 with graft-versus-hostdisease8 and with fibrosis.9 The extent of TNF production may therefore be crucial to the outcome of such diseases in that an individual that produces only low levels of this cytokine, i.e. a low responder may have difficulty in clearing mycobacteria. On the other hand such an individual may be protected against severe sepsis.
In previous studies such low responders have been identified in the healthy population10–14 and this responder status has been linked to certain major histocompatibility complex (MHC)-Class II types. The molecular mechanisms that control the low responder status have however, not been elucidated. The authors have addressed this issue by analysing the LPS-induced TNF production by blood monocyte in healthy donors. The study was restricted to male donors in order to avoid the possible influence of fluctuating hormone levels during the menstrual cycle in females15,16. Among 60 male donors we identified three each that exhibit constantly low and high levels of TNF protein production. Because mixing experiments revealed no influence of plasma factors on the levels of TNF protein we analysed the expression of TNF mRNA. When comparing mRNA levels of LR and HR we found definite lower levels of cytokine mRNA in LR at 2 h after LPS and this was confirmed in kinetic studies. A lower prevalence of TNF mRNA may be due to a more rapid decay by mRNA degrading enzymes. If this were the case then blockade of such degrading enzymes would be expected to lead to similar levels of TNF mRNA. In studies with CHX blockade a pronounced increase of TNF mRNA in LR was seen but the same occurred in HR leading to even higher levels of TNF transcription. Hence, it appears that the lower levels of mRNA in LR is not due to more rapid degradation. This conclusion is supported by studies on half life of mRNA that was determined following blockade of TNF mRNA systems in actinomycin D which showed a similar rate of decay. These results clearly suggest that LR have a lower level of transcription of the TNF gene. LPS-induced TNF gene expression is dependent on binding of LPS to the phosphatidyl-imositol-linked CD14 cell surface receptor17 followed by activation of tyrosine kinases18 and mobilization of transcription factors like NF-kB.19 In studies not shown we found no difference in expression of the CD14 receptor molecule between LR and HR. In addition, comparison of binding activity of transcription factors NF-kB and (CAAT enhancer binding protein (C/EBP), both of which have been shown to control TNF transcription20–23 revealed no difference between the groups. Hence, the factors that determine the lower transcription in LR still have to be determined. Studies on polymorphism of the TNF promoter revealed TNF-1 and TNF-2 alleles that show either a G or A nucleotide at position −308.24 Reporter constructs with the TNF-2 allele were shown to lead to a lower level of transactivation when transfected into cell lines.25
210 / Schraut et al.
Detailed analysis in transfected monocytic and T cell lines did not, however, reveal a difference in reportergene activity and allele specific transcript quantification demonstrated that both alleles are equally expressed in peripheral blood mononuclear cells (PBMC).26,27 In the present study, the genotype of the LR and HR donors is not known. A difference in the TNF promoter is, however, unlikely to explain our findings because expression of another gene i.e. IL-6 was also lower in LR. This is similar to observations of Molvig et al. that reported coordinately low levels of TNF, IL-1 and PGE2 in LR.10 Hence, it appears that a more general mechanism governs the response of monocytes in LR and HR. Low levels of TNF have also been linked to human lentocyte antigen (HLA) HLA-DR214 while another study demonstrated that both HLA-DR2 and HLA-DR4 are required for a low response.12 These findings were extended to patients with systemic lupus erythematodes and here DR expression and low TNF was linked to a higher frequency of nephritis.13 Hence, there appears to be a link between HLA type and low levels of TNF gene expression. It is unlikely that this is due to a linkage disequilibrium of the class II and the TNF locus because the low responder status extends to other cytokines encoded on other chromosomes (refs 10, 12 and this study). Taken together it remains still unclear which factors govern the lower responder status. Our studies have, however, clearly demonstrated that the responder status with respect to TNF is controlled at the level of gene transcription, a finding that will allow for a more precise analysis of gene expression in future studies on LR and HR.
MATERIALS AND METHODS Donors Apparently healthy volunteer blood donors (male, age 18–44) were recruited from lab personnel, students, relatives and friends. Heparinized blood was drawn and either used directly for whole blood stimulation or after isolation of mononuclear cells by ficoll-hypaque density gradient separation.
Cell stimulation Whole blood or isolated PBMC were stimulated with LPS (S. minnesota (L62 61, Sigma, Muenchen, Germany) at 1 mg/ml. Supernatant plasma or culture medium was harvested at 4 h. For mixing experiments whole blood was spun, supernatant plasma harvested, the cells were washed once with RPMI 1640 and were admixed with autologous or heterologous plasma before stimulation.
TNF assay TNF was determined by ELISA essentially as described.28 The assay is based on the use of antibody 199
CYTOKINE, Vol. 9, No. 3 (March, 1997: 206–211)
for coating (kindly provided by Dr T. Subkowski, BASF-Knoll, Ludwigshafen, Germany), and antibody 195-peroxidase conjugate for detection (kindly provided by Dr J. Endl, Boehringer-Mannheim, Penzberg, Germany).
Northern blot analysis For Northern blot analysis, cells were lysed with guanidine isothiocyanate, and RNA was purified over CsCl according to Chirgwin et al.28 RNA (10 mg/lane) was applied to formaldehyde-agarose gels, and after electrophoresis and blotting, the blots were hybridized with synthetic oligonucleotides that were 32P-labelled by poly(A) tailing. The oligonucleotides used were as follows: for TNF, 5'-GGTCTGGTAGGAGACGGCGATGCGGCTGATG-GTGTGGGTGAGGAGCACATGG-GTGGAGGG-3'; for IL-6,5'GGCT GTCTGTGTGGGGCGGCTACATCTTTGGAATCTC-CTGGGGGTACTGG-3'. Equal loading was controlled by analysis of 18 and 28S rRNA in ethidium bromide-stained gels and by GAPDH control hybridization. For analysis of mRNA stability cells were stimulated with LPS for 30 min and actinomycin D was added ((114666 Calbiochem, La Jolla, CA) (HLA) at 10 mg/ml. For treatment with cycloheximide ((C7698 Sigma) cells were admixed with 10 mg/ml of the drug before stimulation.
Acknowledgements This work was supported by Thyssen-Stiftung, Wilhelm Sander Stiftung and Deutsche Forschungsgemeinschaft (SFB 217). We also acknowledge the repeated blood donations by the LR and RH individuals.
REFERENCES 1. Ja¨a¨ttela¨ M (1991) Biology of disease. Biologic activities and mechanisms of action of tumour necrosis factor-alpha/cachectin. Lab Invest 64:724–742. 2. Nakano Y, Onozuka K, Terada Y, Shinomiya H, Nakano M (1990) Protective effect of recombinant tumour necrosis factor-a in murine salmonellosis. J Immunol 144:1935–1941. 3. Kindler V, Sappino A-P, Grau GE, Piguet P-F, Vassalli P (1989) The inducing role of tumour necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 56:731–740. 4. Flynn JL, Goldstein MM, Chan J, Triebold KJ, Pfeffer K, Lowenstein CJ, Schreiber R, Mak TW, Bloom BR (1995) Tumour necrosis factor-a is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2:561–572. 5. Pfeffer K, Matsuyama T, Ku¨ndig TM, Wakeham A, Kishihara K, Shahinian A, Wiegmann K, Ohashi PS, Kro¨nke M, Mak TW (1993) Mice deficient for the 55 kd tumour necrosis receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73:457–467. 6. Rothe J, Lesslauer W, Lo¨tscher H, Lang Y, Koebel P, Ko¨ntgen F, Althage A, Zinkernagel R, Steinmetz M, Bluethmann H (1993) Mice lacking the tumour necrosis factor receptor 1 are resistant to TF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes. Nature 364:798–802. 7. Tracey KJ, Fong Y, Hesse DG, Manogue KR, Lee AT, Kuo GC. Lowry SF, Cerami A (1987) Anti-cachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteraemia. Nature 330:662–664.
Low and high responses for TNF/ 211 8. Piguet P-F, Grau GE, Allet B, Vassalli P (1987) Tumour necrosis factor/cachectin is an effector of skin and gut lesions of the acute phase of graft-vs.-host disease. J Exp Med 166:1280–1289. 9. Piguet PF, Collart MA, Grau GE, Sappino A-P, Vassalli P (1990) Requirement of tumour necrosis factor for development of silica-induced pulmonary fibrosis, Nature [Lond] 344:245–247. 10. Molvig J, Baek L, Christensen P, Manogue KR, Vlassara H, Platz P, Nielsen LS, Svejgaard A, Nerup J (1988) Endotoxinstimulated human monocyte secretion of Interleukin 1, tumour necrosis factor alpha, and prostaglandin E2 shows stable interindividual differences. Scand J Immunol 27:705–716. 11. Endres S, Cannon JG, Ghorbani R, Dempsey RA, Sisson SD, Lonnemann G, Van der Meer JWM, Wolff SM, Dinarello CA (1989) In vitro production of IL 1b, IL 1a, TNF and IL2 in healthy subjects: distribution, effect of cyclooxygenase inhibition and evidence of independent gene regulation. Eur J. Immunol 19:2327–2333. 12. Santamaria P, Gehrz RC, Bryan MK, Barbosa JJ (1989) Involvement of class II MHC molecules in the LPS-induction of IL-1/TNF secretions by human monocytes. J Immunol 143:913–922. 13. Jacob CO, Fronek Z, Lewis GD, Koo M, Hensen JA, McDevitt HO (1990) Heritable major histocompatibility complex class II-associated differences in production of tumour necrosis factor a: relevance to genetic predisposition to systemic lupus erythematosus. Proc Natl Acad Sci USA 87:1233–1237. 14. Bendtzen K, Morling N, Fomsgaard A, Svenson M, Jakobsen B, Odum N, Svejgaard A (1988) Assocation between HLA-DR2 and production of tumour necrosis factor a and Interleukin 1 by mononuclear cells activated by lipopolysaccharide. Scand J Immunol 28:599–606. 15. Loy RA, Lonkides JA, Polan ML (1992) Ovarian steroids modulate human monocyte TNF alpha mRNA levels in cultured human peripheral monocytes. Fertility and Sterility 58:733–739. 16. Angstwurm MWA, Ga¨rtner R, Ziegler-Heitbrock HWL (1977) Cyclic plasma IL-6 levels during normal menstrual cycle. Cytokine 9: in press. 17. Wright SD, Ramos RA, Robias PS, Ulevitch RJ, Mathison JC (1990) CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249:1431–1433. 18. Stefanova´ I, Horejsı´ V, Ansotegui IJ, Knapp W, Stockinger
H (1991) GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science 254:1016–1019. 19. Mu¨ller JM, Ziegler-Heitbrock HWL, Ba¨uerle P (1993) Nuclear factor kappa B, a mediator of lipopolysaccharide effects. Immunobiol 187:233–256. 20. Shakov AN, Collart MA, Vassalli P, Nedospasov SA, Jongeneel CV (1990) kB-type enhancers are involved in LPS-mediated transcriptional activation of the TNF-a gene in primary macrophages, J Exp Med 171:35–47. 21. Ziegler-Heitbrock HWL, Sternsdorf T, Liese J, Belohradsky B, Weber C, Wedel A, Schreck R, Ba¨uerle P, Stro¨bel M (1993) Pyrrolidine dithiocarbamate inhibits NF-kB mobilization and TNF production in human monocytes. J Immunol 151:6986–6993. 22. Pope RM, Leutz A, Ness SA (1994) C/EBP beta regulation of the tumour necrosis factor alpha gene. J Clin Invest 94:1449–1455. 23. Wilson AG, de Vries N, Poclot F, di Giovine FS, van der Putte LBA, Duff GW (1993) An allelic polymorphism within the human tumor necrosis factor a promoter region is strongly associated with HLA A1, B8 and DR3 alleles. J Exp Med 177:557–560. 24. Wedel, A, Sulski G, Ziegler-Heitbrock HWL (1996) CCAAT/Enhancer binding protein in involved in the expression of the tumour necrosis factor gene in human monocytes. Cytokine:335– 341. 25. Wilson AG, Symons JA, McDowell TL, di Giovine FS, Duff GW (1994) Effects of a tumour necrosis factor (TNFa) promoter base transition on transcriptional activity Br J Rheumatol Suppl 33:89. 26. Stuber F, Udalova IA, Book M, Drutskaya LN, Kuprash DV, Turetskaya RL, Schade FU, Nedospasov SA (1996) −308 Tumour necrosis factor (TNF) polymorphism is not associated with survival in severe sepsis and is unrelated to lipopolysaccharide inducibility of the human TNF promoter. J Inflamm 46:42–50. 27. Brinkman BMN, Zuijdgeest D, Kaijzel EL, Breedveld FC, Verweij CL (1996) relevance of the tumour necrosis factor alpha (TNFa) −308 promoter polymorphism in TNFa gene regulation. J Inflamm 46:32–41. 28. Mo¨ller A, Emling F, Blohm D, Schlick E, Schollmeier K (1990) Monoclonal antibodies to human tumour necrosis factor alpha: in vitro and in vivo application. Cytokine 2:162–169. 29. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18 5294–5299.
Cytokines are crucial to immune defence because they determine the type and extent of response. One prominent producer of cytokines is the monocyte/ macrophage that can release excessive amounts of mediators like tumour necrosis factor (TNF) and interleukin 6 (IL-6) after being challenged with lipopolysaccharide (LPS) from Gram-negative bacteria. The amount of cytokines produced may vary with the individual with some showing high, and other’s low levels of production. The extent of cytokine production may determine the outcome of inflammatory diseases. An individual with a low level of production (low responder, LR) may, for instance, be less prone to autoimmune disease, to transplant rejection or to severe sepsis. On the other hand, a LR may have difficulty in coping with intracellular parasites like leishmania. Hence, it is clear that the responder status may have clinical implications. The molecular mechanisms From the Institute for Immunology, University of Munich, Goethestr. 31, 80336 Mu¨nchen, Germany Correspondence to: H.W.L. Ziegler-Heitbrock, Institute for Immunology, University of Munich, Goethestr. 31, 80336 Mu¨nchen Received 11 October 1995; revised and accepted for publication 24 June 1996 7 1997 Academic Press Limited 1043-4666/97/030206 + 6 $25.00/0/ck960155 KEY WORDS: Monocyte TNF/LPS 206
governing the different levels of cytokine expression in LR and high responders (HR) have, however, not been analysed and it is unclear whether the degree of expression is controlled at the level of receptors, of transcription, of mRNA degradation, translation or protein export. In the present report the authors have addressed this issue and demonstrated in selected LR and HR that the responder status is controlled at the transcription level.
RESULTS When screening 60 male donors for LPSstimulated TNF production, we detected three donors who produced consistently low and three who produced consistently high levels of this cytokine. As shown in Figure 1, immunoreactive TNF detected by enzyme-linked immunoabsorbent assay (ELISA) in cell free plasma of LPS-stimulated whole blood was below 4 ng/ml in the low responders (LR) and above 13 ng/ml in the high responders (HR). These low and high levels of TNF protein production were observed on 3 or more occasions over a period of more than 3 months (not shown). Exchange of plasma did not change this pattern. As shown in Figure 2, addition of plasma from either LR or HR to blood cells after repeated washing did not CYTOKINE, Vol. 9, No. 3 (March), 1997: pp 206–211
Low and high responses for TNF/ 207
25
A
20
15
10
5 Immunoreactive TNF (ng/ml)
affect the LR or HR phenotype of the cells. These data suggest that plasma factors like, for instance, soluble TNF receptors are not responsible for the responder status. Rather LR monocytes appear to produce less TNF. Therefore, TNF mRNA levels in LPS-stimulated monocytes have been studied. Figure 3 demonstrates that the two LR tested had low levels of LPS-induced mRNA, while TNF transcripts were high in the three HR at 1 h post LPS. One might speculate that LR have a slower kinetic of transcription with a peak at a later point in time. Therefore time course studies were performed. Results in Figure 4 demonstrate high levels of TNF transcripts in HR already at 0.5 h, and this signal is almost completely gone at 5 h. In LR TNF mRNA was detectable at 0.5 h as well, but levels were low at this and at all other time points. Hence, the LR phenotype does not appear to be due to a shift in the kinetics. TNF mRNA levels may be lower in LR because of lower transcription or because of more rapid
0 Cells : LR Plasma : LR
LR HR
HR HR
HR LR
HR HR
HR LR
Exp. 1 25
B
20
15 15
Immunoreactive TNF (ng/ml)
10
5
10
LR HR Exp. 2
Figure 2.
5
Figure 1.
0 Cells : LR Plasma : LR
TNF production in low and high responders
Heparinized whole blood samples were stimulated with LPS (1 mg/ml) for 4 h and TNF was determined by ELISA. The pattern shown was observed repeatedly on two or three occasions over a period of more than 3 months. In LR1 (6) TNF values were 22 2 4% of the respective HR, in LR2 (x) they were 23 2 5% and in LR3 (w) they were 7 2 1.5%.
The responder status is independent of plasma factors
Heparinized blood from low and high responders was separated into cells and plasma, the cells were washed once with culture medium and cells and plasma were recombined as given in the figure. After LPS stimulation for 4 h at 1 mg/ml supernatant plasma was tested for TNF by ELISA. Given are results from two HR–LR combinations (A, B). Similar results were obtained in separate experiments using a TNF bioassay as a read-out system.
degradation. To resolve this question, we have blocked degradation by using treatment with the protein synthesis inhibitor cycloheximide (CHX). Such treatment enhanced TNF transcripts in LR at all points in time (Fig. 5), but in the HR CHX resulted in much higher levels. The authors next studied mRNA degradation directly by blockade of further transcription with actinomycin D followed by determination of mRNA half life. TNF mRNA production continued
208 / Schraut et al.
LPS (1µg/ml)
–
CYTOKINE, Vol. 9, No. 3 (March, 1997: 206–211)
+
–
+
–
+
–
+
–
+
LR 0 0.5 1 2 4 6
HR 0 0.5 1 2 4 6
(h)
0.5 1 2 4 6
0.5 1 2 4 6
(h)
LPS
TNF transcripts in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for 2 h, adherent cells were recovered and isolated RNA (10 mg/lane) was analysed by Northern blot with TNF specific oligonucleotides (lanes 1–4 = 2 low responders, lanes 5–10 = 3 high responders). The 3rd low responder was found to have low TNF mRNA in an independent experiment (not shown).
somewhat after addition of actinomycin D in the low responder because the signal was higher at 45 min. compared to 30 min, while no further increase of mRNA was apparent in the HR (Fig. 6A). From 45 min on TNF mRNA decayed rapidly in both LR and HR to low levels at 2 h. Densitometry and calculation of half life revealed a t1/2 of 27 min for the HR and 24 min for the LR (Fig. 6B). Taken together, the lower levels of LPS induced TNF mRNA in LR monocytes appears not to be due to more rapid degradation, but rather due to reduced transcription. These data would suggest that signal transduction must be different between LR and HR. In studies not shown, we found no differences in cell surface expression of the CD14 LPS receptor. Furthermore, no difference could be detected in binding of nuclear factors C/EBP and NF-kB to motifs in the human TNF promoter. Finally, we asked whether expression of other LPS-induced cytokines was also different between LR and HR. Northern blot analysis of LPS induced IL-6 mRNA revealed low transcription levels in two LR and high levels in two HR, while hybridization of the same blots for the house-keeping gene glyceraldehyde-
LR 0 0.5 1 2
4
5
HR 0 0.5 1 2
4
5
LPS+ CHx
Figure 5. Effect of cycloheximide on TNF transcripts in low and high responders Mononuclear cells were stimulated with LPS (1 mg/ml) for various periods of time with and without addition of cycloheximide (CHX) at 10 mg/ml. Adherent cells were recovered, RNA was isolated and analysed for TNF transcripts by Northern blotting. The same pattern was seen in two additional pairs of low and high responders. Transcript levels estimated by densitometry were in average of the 3 experiments 2.5 2 0.6 higher in HR without and 1.8 2 0.3 with CHX at the 1 h time point.
A
(h)
Time course of TNF mRNA in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for various periods of time and adherent cells were recovered and RNA was analysed for TNF by Northern blotting. The same pattern was seen in a second pair of LR and HR.
HR ACT D
0 30 45 60 90 120 0 30 45 60 90 120
Figure 6. Figure 4.
LR ACT D
TNF mRNA (arbitrary densitometry units)
Figure 3.
200
B
(min)
t1/2 = 27 min
150 100
50 40 30
t1/2 = 27 min
20
15
30 60 Time (min)
120
Half life of TNF mRNA in low and high responders
(A) Mononuclear cells were stimulated with LPS (1 mg/ml) and after 30 min. Actinomycin D (ActD) was added at 10 mg/ml. Adherent cells were recovered, RNA isolated and analysed for TNF transcripts by Northern blotting. (B) Densitometry was performed for time points 15, 30 and 60 min post ActD (=45, 60, 90 min in A). The same pattern was seen with a second pair of low and high responders.
Low and high responses for TNF/ 209
LR
HR
LR
HR
IL-6
GAPDH
Figure 7.
IL-6 transcripts in low and high responders
Mononuclear cells were stimulated with LPS (1 mg/ml) for 2 h adherent cells were recovered and isolated RNA (10 mg/lane) was analysed by Northern blotting for IL-6 transcripts.
3-phosphate dehydrogenase (GAPDH) gave similar levels (Fig. 7).
DISCUSSION TNF was originally defined based on its antitumour activity, but it later turned out to be a major pro-inflammatory cytokine that affects almost every cell of the body.1 TNF was shown to be required for an efficient defence in many infectious diseases including enterobacterial, mycobacterial and leishmanial infections.2–6 On the other hand, TNF may have detrimental effects because excessive production of this cytokine has been associated with severe sepsis7 with graft-versus-hostdisease8 and with fibrosis.9 The extent of TNF production may therefore be crucial to the outcome of such diseases in that an individual that produces only low levels of this cytokine, i.e. a low responder may have difficulty in clearing mycobacteria. On the other hand such an individual may be protected against severe sepsis.
In previous studies such low responders have been identified in the healthy population10–14 and this responder status has been linked to certain major histocompatibility complex (MHC)-Class II types. The molecular mechanisms that control the low responder status have however, not been elucidated. The authors have addressed this issue by analysing the LPS-induced TNF production by blood monocyte in healthy donors. The study was restricted to male donors in order to avoid the possible influence of fluctuating hormone levels during the menstrual cycle in females15,16. Among 60 male donors we identified three each that exhibit constantly low and high levels of TNF protein production. Because mixing experiments revealed no influence of plasma factors on the levels of TNF protein we analysed the expression of TNF mRNA. When comparing mRNA levels of LR and HR we found definite lower levels of cytokine mRNA in LR at 2 h after LPS and this was confirmed in kinetic studies. A lower prevalence of TNF mRNA may be due to a more rapid decay by mRNA degrading enzymes. If this were the case then blockade of such degrading enzymes would be expected to lead to similar levels of TNF mRNA. In studies with CHX blockade a pronounced increase of TNF mRNA in LR was seen but the same occurred in HR leading to even higher levels of TNF transcription. Hence, it appears that the lower levels of mRNA in LR is not due to more rapid degradation. This conclusion is supported by studies on half life of mRNA that was determined following blockade of TNF mRNA systems in actinomycin D which showed a similar rate of decay. These results clearly suggest that LR have a lower level of transcription of the TNF gene. LPS-induced TNF gene expression is dependent on binding of LPS to the phosphatidyl-imositol-linked CD14 cell surface receptor17 followed by activation of tyrosine kinases18 and mobilization of transcription factors like NF-kB.19 In studies not shown we found no difference in expression of the CD14 receptor molecule between LR and HR. In addition, comparison of binding activity of transcription factors NF-kB and (CAAT enhancer binding protein (C/EBP), both of which have been shown to control TNF transcription20–23 revealed no difference between the groups. Hence, the factors that determine the lower transcription in LR still have to be determined. Studies on polymorphism of the TNF promoter revealed TNF-1 and TNF-2 alleles that show either a G or A nucleotide at position −308.24 Reporter constructs with the TNF-2 allele were shown to lead to a lower level of transactivation when transfected into cell lines.25
210 / Schraut et al.
Detailed analysis in transfected monocytic and T cell lines did not, however, reveal a difference in reportergene activity and allele specific transcript quantification demonstrated that both alleles are equally expressed in peripheral blood mononuclear cells (PBMC).26,27 In the present study, the genotype of the LR and HR donors is not known. A difference in the TNF promoter is, however, unlikely to explain our findings because expression of another gene i.e. IL-6 was also lower in LR. This is similar to observations of Molvig et al. that reported coordinately low levels of TNF, IL-1 and PGE2 in LR.10 Hence, it appears that a more general mechanism governs the response of monocytes in LR and HR. Low levels of TNF have also been linked to human lentocyte antigen (HLA) HLA-DR214 while another study demonstrated that both HLA-DR2 and HLA-DR4 are required for a low response.12 These findings were extended to patients with systemic lupus erythematodes and here DR expression and low TNF was linked to a higher frequency of nephritis.13 Hence, there appears to be a link between HLA type and low levels of TNF gene expression. It is unlikely that this is due to a linkage disequilibrium of the class II and the TNF locus because the low responder status extends to other cytokines encoded on other chromosomes (refs 10, 12 and this study). Taken together it remains still unclear which factors govern the lower responder status. Our studies have, however, clearly demonstrated that the responder status with respect to TNF is controlled at the level of gene transcription, a finding that will allow for a more precise analysis of gene expression in future studies on LR and HR.
MATERIALS AND METHODS Donors Apparently healthy volunteer blood donors (male, age 18–44) were recruited from lab personnel, students, relatives and friends. Heparinized blood was drawn and either used directly for whole blood stimulation or after isolation of mononuclear cells by ficoll-hypaque density gradient separation.
Cell stimulation Whole blood or isolated PBMC were stimulated with LPS (S. minnesota (L62 61, Sigma, Muenchen, Germany) at 1 mg/ml. Supernatant plasma or culture medium was harvested at 4 h. For mixing experiments whole blood was spun, supernatant plasma harvested, the cells were washed once with RPMI 1640 and were admixed with autologous or heterologous plasma before stimulation.
TNF assay TNF was determined by ELISA essentially as described.28 The assay is based on the use of antibody 199
CYTOKINE, Vol. 9, No. 3 (March, 1997: 206–211)
for coating (kindly provided by Dr T. Subkowski, BASF-Knoll, Ludwigshafen, Germany), and antibody 195-peroxidase conjugate for detection (kindly provided by Dr J. Endl, Boehringer-Mannheim, Penzberg, Germany).
Northern blot analysis For Northern blot analysis, cells were lysed with guanidine isothiocyanate, and RNA was purified over CsCl according to Chirgwin et al.28 RNA (10 mg/lane) was applied to formaldehyde-agarose gels, and after electrophoresis and blotting, the blots were hybridized with synthetic oligonucleotides that were 32P-labelled by poly(A) tailing. The oligonucleotides used were as follows: for TNF, 5'-GGTCTGGTAGGAGACGGCGATGCGGCTGATG-GTGTGGGTGAGGAGCACATGG-GTGGAGGG-3'; for IL-6,5'GGCT GTCTGTGTGGGGCGGCTACATCTTTGGAATCTC-CTGGGGGTACTGG-3'. Equal loading was controlled by analysis of 18 and 28S rRNA in ethidium bromide-stained gels and by GAPDH control hybridization. For analysis of mRNA stability cells were stimulated with LPS for 30 min and actinomycin D was added ((114666 Calbiochem, La Jolla, CA) (HLA) at 10 mg/ml. For treatment with cycloheximide ((C7698 Sigma) cells were admixed with 10 mg/ml of the drug before stimulation.
Acknowledgements This work was supported by Thyssen-Stiftung, Wilhelm Sander Stiftung and Deutsche Forschungsgemeinschaft (SFB 217). We also acknowledge the repeated blood donations by the LR and RH individuals.
REFERENCES 1. Ja¨a¨ttela¨ M (1991) Biology of disease. Biologic activities and mechanisms of action of tumour necrosis factor-alpha/cachectin. Lab Invest 64:724–742. 2. Nakano Y, Onozuka K, Terada Y, Shinomiya H, Nakano M (1990) Protective effect of recombinant tumour necrosis factor-a in murine salmonellosis. J Immunol 144:1935–1941. 3. Kindler V, Sappino A-P, Grau GE, Piguet P-F, Vassalli P (1989) The inducing role of tumour necrosis factor in the development of bactericidal granulomas during BCG infection. Cell 56:731–740. 4. Flynn JL, Goldstein MM, Chan J, Triebold KJ, Pfeffer K, Lowenstein CJ, Schreiber R, Mak TW, Bloom BR (1995) Tumour necrosis factor-a is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2:561–572. 5. Pfeffer K, Matsuyama T, Ku¨ndig TM, Wakeham A, Kishihara K, Shahinian A, Wiegmann K, Ohashi PS, Kro¨nke M, Mak TW (1993) Mice deficient for the 55 kd tumour necrosis receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73:457–467. 6. Rothe J, Lesslauer W, Lo¨tscher H, Lang Y, Koebel P, Ko¨ntgen F, Althage A, Zinkernagel R, Steinmetz M, Bluethmann H (1993) Mice lacking the tumour necrosis factor receptor 1 are resistant to TF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes. Nature 364:798–802. 7. Tracey KJ, Fong Y, Hesse DG, Manogue KR, Lee AT, Kuo GC. Lowry SF, Cerami A (1987) Anti-cachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteraemia. Nature 330:662–664.
Low and high responses for TNF/ 211 8. Piguet P-F, Grau GE, Allet B, Vassalli P (1987) Tumour necrosis factor/cachectin is an effector of skin and gut lesions of the acute phase of graft-vs.-host disease. J Exp Med 166:1280–1289. 9. Piguet PF, Collart MA, Grau GE, Sappino A-P, Vassalli P (1990) Requirement of tumour necrosis factor for development of silica-induced pulmonary fibrosis, Nature [Lond] 344:245–247. 10. Molvig J, Baek L, Christensen P, Manogue KR, Vlassara H, Platz P, Nielsen LS, Svejgaard A, Nerup J (1988) Endotoxinstimulated human monocyte secretion of Interleukin 1, tumour necrosis factor alpha, and prostaglandin E2 shows stable interindividual differences. Scand J Immunol 27:705–716. 11. Endres S, Cannon JG, Ghorbani R, Dempsey RA, Sisson SD, Lonnemann G, Van der Meer JWM, Wolff SM, Dinarello CA (1989) In vitro production of IL 1b, IL 1a, TNF and IL2 in healthy subjects: distribution, effect of cyclooxygenase inhibition and evidence of independent gene regulation. Eur J. Immunol 19:2327–2333. 12. Santamaria P, Gehrz RC, Bryan MK, Barbosa JJ (1989) Involvement of class II MHC molecules in the LPS-induction of IL-1/TNF secretions by human monocytes. J Immunol 143:913–922. 13. Jacob CO, Fronek Z, Lewis GD, Koo M, Hensen JA, McDevitt HO (1990) Heritable major histocompatibility complex class II-associated differences in production of tumour necrosis factor a: relevance to genetic predisposition to systemic lupus erythematosus. Proc Natl Acad Sci USA 87:1233–1237. 14. Bendtzen K, Morling N, Fomsgaard A, Svenson M, Jakobsen B, Odum N, Svejgaard A (1988) Assocation between HLA-DR2 and production of tumour necrosis factor a and Interleukin 1 by mononuclear cells activated by lipopolysaccharide. Scand J Immunol 28:599–606. 15. Loy RA, Lonkides JA, Polan ML (1992) Ovarian steroids modulate human monocyte TNF alpha mRNA levels in cultured human peripheral monocytes. Fertility and Sterility 58:733–739. 16. Angstwurm MWA, Ga¨rtner R, Ziegler-Heitbrock HWL (1977) Cyclic plasma IL-6 levels during normal menstrual cycle. Cytokine 9: in press. 17. Wright SD, Ramos RA, Robias PS, Ulevitch RJ, Mathison JC (1990) CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249:1431–1433. 18. Stefanova´ I, Horejsı´ V, Ansotegui IJ, Knapp W, Stockinger
H (1991) GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science 254:1016–1019. 19. Mu¨ller JM, Ziegler-Heitbrock HWL, Ba¨uerle P (1993) Nuclear factor kappa B, a mediator of lipopolysaccharide effects. Immunobiol 187:233–256. 20. Shakov AN, Collart MA, Vassalli P, Nedospasov SA, Jongeneel CV (1990) kB-type enhancers are involved in LPS-mediated transcriptional activation of the TNF-a gene in primary macrophages, J Exp Med 171:35–47. 21. Ziegler-Heitbrock HWL, Sternsdorf T, Liese J, Belohradsky B, Weber C, Wedel A, Schreck R, Ba¨uerle P, Stro¨bel M (1993) Pyrrolidine dithiocarbamate inhibits NF-kB mobilization and TNF production in human monocytes. J Immunol 151:6986–6993. 22. Pope RM, Leutz A, Ness SA (1994) C/EBP beta regulation of the tumour necrosis factor alpha gene. J Clin Invest 94:1449–1455. 23. Wilson AG, de Vries N, Poclot F, di Giovine FS, van der Putte LBA, Duff GW (1993) An allelic polymorphism within the human tumor necrosis factor a promoter region is strongly associated with HLA A1, B8 and DR3 alleles. J Exp Med 177:557–560. 24. Wedel, A, Sulski G, Ziegler-Heitbrock HWL (1996) CCAAT/Enhancer binding protein in involved in the expression of the tumour necrosis factor gene in human monocytes. Cytokine:335– 341. 25. Wilson AG, Symons JA, McDowell TL, di Giovine FS, Duff GW (1994) Effects of a tumour necrosis factor (TNFa) promoter base transition on transcriptional activity Br J Rheumatol Suppl 33:89. 26. Stuber F, Udalova IA, Book M, Drutskaya LN, Kuprash DV, Turetskaya RL, Schade FU, Nedospasov SA (1996) −308 Tumour necrosis factor (TNF) polymorphism is not associated with survival in severe sepsis and is unrelated to lipopolysaccharide inducibility of the human TNF promoter. J Inflamm 46:42–50. 27. Brinkman BMN, Zuijdgeest D, Kaijzel EL, Breedveld FC, Verweij CL (1996) relevance of the tumour necrosis factor alpha (TNFa) −308 promoter polymorphism in TNFa gene regulation. J Inflamm 46:32–41. 28. Mo¨ller A, Emling F, Blohm D, Schlick E, Schollmeier K (1990) Monoclonal antibodies to human tumour necrosis factor alpha: in vitro and in vivo application. Cytokine 2:162–169. 29. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18 5294–5299.