- Email: [email protected]
Inhibition of human TNFα and LT in cell-free extracts and in cell culture by antisense oligonucleotides
Biochi~ic~a ELSEVIER
et Biophysica t~ta Bioc himica et Biophysica Acta 1317 (1996) 168 - 174
Inhibition of human TNFa and LT in cell-free extracts and in cell culture by antisense oligonucleotides Christian Lefebvre d'Hellencourt
a,*,
Ltna Diaw b, Pascale Comillet
a,
Moncef Guenounou
a
a Laboratoire d'H~matologie et Laboratoire de Biologie des Cytokines, C.H.R. Robert Debr~, rue Alexis Carrel, 51092 Reims cedex, France b Unit~ d'Immuno-H~.matologie et d'lmmuno-Pathologie, lnstitut Pasteur, 25 rue du docteur Roux, 75724 Paris cedex 15, France
Received 22 August 1996; accepted 10 September 1996
Abstract Antisense oligonucleotides (ODN), complementary to mRNA of human tumor necrosis factor a (TNFa) and lymphotoxin (LT) were tested for their ability to inhibit TNFs. TNFs production was studied in cell-free systems including wheat germ extract (WGE) and rabbit reticulocyte lysate (RRL). All ODN were effective in WGE at low concentration (0.2 /xM), except those targeted to the 3' region of TNFa mRNA. A short ODN complementary to a common region between TNFa and LT inhibited both TNFs. In contrast, high ODN concentration (50 /xM) was needed to inhibit LT mRNA translation in RRL, whereas no clear inhibition of TNFa was observed unless RNase H was added to the translation mixture. ODN effects on TNFs production by stimulated cell line in culture were also investigated. Three ODN - - one located in the Y-untranslated region, one spanning the AUG initiation codon and one downstream of this AUG - were the most effective sequences to decrease TNFa production. Two ODN targeted to the AUG initiation codon of LT were also able to inhibit its production. In conclusion we confirm the role of RNase H in cell free systems, and we found that there is no correlation between ODN efficiency in a cell-free system nor in cell culture. Efficient ODN could be used for in vitro investigation of the role of TNFa and LT in mechanism in which they are involved. Keywords: Antisense oligonucleotide; Cytokine; TNF; Lymphotoxin; Cell free extract
1. Introduction Tumor necrosis factor a ( T N F a ) and lymphotoxin or tumor necrosis factor /3 (LT, L T a , or TNF/3) are two closely related cytokines involved in many regulatory activities [1,2]. Almost all cells of the immune system (including macrophages, B and T lymphocytes, natural killer cells, and mast cells) and cells that are not part of the immune system (e.g., smooth muscle cells, astrocytes, granulosa cells, epithelial cells, breast and ovarian tumor cells) can produce T N F a [3]. LT, in contrast, is mainly secreted by lymphocytes [4]. T N F a and LT are each
Abbreviations: TNF, tumor necrosis factor; LT, lymphotoxin; PMA, phorbol 12-myristate 13-acetate; PHA, phytohemagglutinin; ODN, antisense oligonucleotides; ELISA, enzyme-linked immunosorbent assay; BMV, brome mosaic virus; WGE, wheat germ extract; RRL, rabbit reticulocyte lysate; FCS, fetal calf serum; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TCA, trichloroacetic acid * Corresponding author. Present address: INSERM U298, CHU, 4 rue Larrey, 49033 Angers cedex 01, France. Fax: + 33 2 41354729. E-mail: hellenco @infobiogen.fr
encoded by unique genes and share approximately 30 percent amino acid homology [5]. They have very similar but not identical biological activities [6]. In a number of pathology such as septic shock, rheumatoid arthritis, allograft rejection, parasitic diseases, meningococcal disease and AIDS a high level of T N F a is released in the plasma [7]. T N F a and LT are involved in inflammatory processes, in pathologies like multiple sclerosis or AIDS dementia [8-11], and may also induce proliferation as autocrine growth factor for B cell lines [12-14] and for cytotrophoblastic cell [15]. Thus it may be of interest to develop specific inhibitor of TNFs to investigate their role in such disorders and to propose new therapies; this can be achieved by antisense strategy. Antisense oligodeoxyribonucleotides (ODN) have been used to block translation of protein in cell-free systems, in cultured cells, and in vivo [16,17]. Numerous studies demonstrate the regulation of cytokine production by this methodology [18], but to our knowledge ODN have neither yet been applied to regulate T N F a and LT production in cell-free systems, nor to regulate LT in cell culture.
0925-4439/96//$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PH S 0 9 2 5 - 4 4 3 9 ( 9 6 ) 0 0 0 5 1-8
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174 5'
AUG
TNFtx mRNA
'I'002 T052 TI02 T152 'I"168 T187 AB T027 '1"077 T127 T180 T188 T183 T195
3'
5'
T920
AUG
L066
LT mRNA
L081
Untranslatedregion
~
169
AB m
Coding region
Fig. 1. Schematic illustration of location of the oligonucleotides.
2. Materials and methods
washed with 75% ethanol, and reconstituted in water. An aliquot of each sample was labelled at the 5'-end using T4 polynucleotide kinase and [32p]ATP and analysed for homogeneity on a 20% polyacrylamide gel. Concentration of oligonucleotide solution was determined spectrophotometrically. Schematic localization and sequence of oligonucleotide are shown in Fig. 1 and in Table 1. ODN T180, T183 and T188 have been previously described [12,19,20]. Position 183 to 185 and 81 to 83 correspond to the AUG codon of human TNFc~ and LT respectively.
2.1. Reagents
2.3. Plasmids
In the present study, the action of different ODN on the synthesis of TNFc~ and LT was investigated, both in cell free system and in cell culture. The importance of RNase H was confirmed in cell free extract and the possibility to predict AS efficiency in cell culture, by calculated secondary structure or by using cell free extract was also studied.
Phorbol 12-myristate 13-acetate (PMA), phytohemagglutinin (PHA-M), and calcium ionophore A23187 were purchased from Sigma Chemicals (Saint Quentin FaUavier, France) and used at 10 ng/ml, 10 /xg/ml and 0.5 /xM, respectively.
2.2. Oligonucleotides Unmodified oligonucleotides were synthesized by Genset (Paris, France) on an automated solid-phase synthesizer (Applied Biosystems) using phosphoramidite chemistry. Oligonucleotides were ethanol precipitated twice,
The construction PSP65-hTNF1 containing the TNF~ cDNA in SP6 orientation was obtained from the Laboratory of molecular biology-plasmid collection University of Gent [21]. LT cDNA clone LT#11 is a generous gift from N. Matsuyama [22]. The EcoRI fragment of LT#11 was subcloned into pGEM4Z plasmid (Promega, Charbonni~res, France). The construction containing LT cDNA in a SP6 orientation was designated pGEM4Z-hLT1.
2.4. In vitro synthesis of capped transcripts Plasmids pSP65-hTNF1 and pGEM4Z-hLT1 linearized, respectively, at the unique NcoI site and HindlII, were
Table 1 Sequence of oligonucleotides Oligonucleotide
Length
Sequence (5' --, 3')
Target site
Target
T002
l8
GAGGGAGCGTCTGCTGGC
T027 T052 T077
18 18 18
5'-untranslated 5'-untranslated 5'-untranslated 5'-untranslated
T 102 T 127 T152
18 18 18
TAGCTGGTCCTCTGCTGT GGTCTGTAGTrGCTTCTC GCGTCTGAGGGTTGI-ITI GCCTGGCAGCTTGTCAGG
T168 * T180 T183 * T187 T 188 T 195 AB * T920 L066
18 15 18 15 17 18 13 18 18
TCATGGTGTCCTTTCCAG TTCAGTGCTCATGGT
hTNFa hTNFa hTNFot hTNFa hTNFa hTNFo~ hTNFa hTNFot hTNFa
L081 MT *
18 20
ACGT'I'CAGGTGGTGTCA T TCATGGTGTCTITrCTGGAG
TCAGTATGTGAGAGGAAG AGGGGAGAGAGGGTGAAG
CATGCTTTCAGTGCTCAT TCATGC I I I C A G T G C GGATCATGCTTI'CAGTG CCACGTCCCGGATCATGC ACCTGGGAGTAGA ATI'GGGGCAGGGGAGGCG CA TGGGGAGAACCCAGGC
5'-untranslated
5'-untranslated 5'-untranslated AUG codon AUG codon AUG codon coding region E 1 coding region E 1 coding region E 1 coding region E4 3' untranslated AUG codon AUG codon AUG codon
hTNFot hTNFot hTNFa hTNFct hTNFot/hLT hTNFa hLT hLT mTNFt~
Underlining indicates sequence complementary to the initiation codon. El, exon 1; E,4, exon 4; h, human; m, murin. * Sense and/or random sequence of these oligonucleotide were also synthesized and used as control.
170
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174
used as a template for run-off transcription. In vitro transcription was carried out essentially as recommended by the supplier (Promega). Briefly 5 /zg of linearized plasmid were transcribe in 10/~1 of 5 × buffer (200 mM Tris-HC1, pH 7.5; 30 mM MgC12; 10 mM spermidine; 50 mM NaC1), 50 U of RNasin TM, 0.5 A260 units of m7(GS)ppp(5')G cap structure (Boerhinger Mannheim, Meylan, France), 250 /zM GTP, 500 /zM of the 3 other NTPs (Promega) and 40 U of SP6 polymerase. After 1 h at 37°C, the mixture was incubated with DNase I 15 min at 37°C and was phenol-chloroform extracted. Unincorporated nucleotides were removed by Sephadex G-50 columms (Boehringer Mannheim). 2.5. Translation in cell-free system
Nuclease-treated wheat germ extract (WGE) and rabbit reticulocyte lysate (RRL) were purchased from Promega. Cell-free translation was carried out essentially as recommended by the supplier. 70 ng of in vitro synthesized capped transcript solution was used per 10 /~1 of translational mixture, and 0.150 MBq of [35S]methionine (37 TBq/mmol, Amersham, Les Ulis, France) was used for labelling. Brome mosaic virus (BMV) RNA was used as control of translation. The reaction was run 1 h at 25°C for WGE and 1 h at 30°C for RRL. The 35S-labelled translation products were analyzed either by trichloroacetic acid (TCA) precipitation [23] or by electrophoretic separation on 15% SDS-polyacrylamide slab gels (SDS-PAGE) [24], and autoradiography. Autoradiographs were quantified using laser densitometry. 2.6. Cell culture
Lymphoblastic T (Jurkat) and monocytic (U937) cell lines were maintained in RPMI 1640 supplemented with L-glutamine, penicillin-streptomycin and 10% heat inactivated fetal calf serum (FCS) (Gibco BRL, Eragny, France). After three washes, cells were resuspended at 5 × 104 cells in 0.1 ml of culture medium and plated in 96 well plates. Various concentrations of oligonucleotide (5 /zM-100 /xM) were added for a 4-h incubation period, stimuli were then added, and the cells reincubated at 37°C and in 5% CO 2. Supernatants were collected after 6 h and 24 h for U937 and Jurkat culture respectively and stored at - 8 0 ° C until assayed for cytokine production. Cell viability was monitored by trypan blue exclusion.
ml). The cells were then incubated at 37°C for 18 h, and extensively washed afterwards in Hanks' balanced salt solution. The remaining adherent cells were stained for 10 min with crystal violet (0.2% in 2% ethanol), washed with water and 0,1 ml of 1% SDS was added to each well to solubilize the stained cells. The absorbance of each well was read at 550 nm. The percentage of cytotoxicity was defined as the relative absorbance of the tested sample versus the absorbance of the control (medium only) wells. %cytotoxicity = 1 -
( absorbance samples ) absorbance control × 100
The cytotoxicity titer (U/ml) was defined as the reciprocal dilution of the test sample inducing 50% of cytotoxicity. Recombinant TNF was used as a positive control. In some experiments, the level of TNFa and LT were determined alternatively by using commercialized kits. TNFa ELISA was purchased from Medgenix (Fleurus, Belgium) and LT ELISA was purchased from R & D Systems Europe (Abingdon, UK).
3. Results 3.1. Effects o f oligonucleotides on TNFs translation in cell-free extracts 3.1.1. Translation in wheat germ extract
In vitro synthesized, capped transcripts of human TNF o r / a n d human LT were translated in WGE in the presence or absence of oligonucleotides. We found that ODN targeted to the 5', AUG, or coding region successfully arrested translation of TNFa, whereas an ODN in the 3' untranslated region (ODN T920) and control sequences (ODN T168S, ABS and MTS) had no inhibitory effects. Furthermore, ODN in the AUG region of LT (ODN L066 and L081) were also effective at inhibiting LT translation. All inhibitory ODN displayed similar efficiency and about 0.2 /xM led to a 50% decrease of translation as determined by TCA precipitation and SDS-PAGE (Fig. 2A). Inhibition of translation was specific, because when the two RNA were co-translated in presence of ODN it was possible to inhibit one of them without modifying the other (at 0.5 and 1 /xM) (Fig. 2B). We also tested a short ODN (ODN AB) which is common to TNFa and LT; we observed that both TNFs were inhibited (at 1 and 5 /zM), whereas the sense sequence (ABS) or an unrelated sequence (MTS) had no effect on TNFs translation (Fig. 3).
2.7. TNFs detection 3.1.2. Translation in rabbit reticulocyte lysate
TNF assays were performed using the sensitive L929 cells as describe by Fish and Gifford [25]. Briefly, 5 × 10 4 L-929 cells/well suspended in 0.1 ml RPMI 1640 containing 10% FCS were seeded in 96-well microtitre plates. Actinomycin D (Sigma) at 4 /xg/ml final concentration was added prior to the addition of the sample dilutions (0.1
The effects of ODN on translation of TNFa in RRL were variable, and no significant inhibition was observed until addition of RNase H (20 U / m l ) in the reaction mixture as previously described [26]. Thus, after RNase H addition, a similar pattern as in WGE was obtained, and the dose of ODN necessary for 50% of inhibition (ID
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174 1
2
3
4
5
6
7
8
9
I0
1
2
171
3
4
5
6
7
8
9
3
4
5
6
7
8
9
TNFot LT~
B I00_
75 ~
LT
25
2
0 3
4
5
6
7
8
9 10
Fig. 2. Inhibitionof TNFa and LT translationin WGE. Samples were analysedon SDS-PAGEand autoradiographed as indicatedin Section2. A: Lane 1: rio RNA, lane 2: BMV mRNA, lane 3: LT mRNA, lane 4-6: LT mRNA+ ODN L081 at 0.1 txM (lane 4), 0.5 /xM (lane 5) and 1 txM (lane 6), lane 7: TNFa mRNA, lane 8-10: TNFa mRNA + ODN T168 at 0.1 txM (lane 8), 0.5 ~M (lane 9) and 1 /xM (lane 10). Similar inhibitionof translationwas observed with ODN T102, T127, T152, T168, T180, T183, T187, T188, T195 for TNFa and with L066 for LT. Densitometrywas expressed as percent control (lane 3 for LT and lane 7 for TNFo~). B: Lane 1: no RNA, lane 2: BMV mRNA, lane 3: LT and TNFa mRNA, lane 4-6: LT and TNFa mRNA + ODN L081 at 0.1 NM (lane 4), 0.5 txM (lane 5) and 1 IxM (lane 6), lane 7-9: LT and TNFa mRNA+ ODN T168 at 0.1 ~M (lane 7), 0.5 IxM (lane 8) and 1 ~M (lane 9). Densitometrywas expressed as percent control (lane 3). Typical autoradiographrepresentativeof three differentexperiments.
1 2
3
4
5
6
7
8
9 10 11 12 13 14
TNFct -~------ LT
50%) was also about 0.2 /xM (Fig. 4B). Besides, we found a difference with LT; its production was inhibited by both LT ODN at high concentration ( 2 0 - 8 0 /zM) even without RNase H addition, and ID 50% was about 40/xM (Fig. 4A). This LT inhibition did not result from a general effect on in vitro translation, as these ODNs did not induce any decrease of BMV protein synthesis at ODN concentrations up to 80 /zM (Fig. 4A).
3.2. Effects of oligonucleotides on stimulated cells
125.
iI1111,I1[11 2
.
.
3
.
4
.
.
5
.
6
R •
m
7
8
.
.
.
.
.
.
.
9 10 11 12 13 14
LT TNF~
Fig. 3. Inhibition of both TNFs translation in WGE. Lane 1, no RNA; lane 2, BMV mRNA; lane 3, LT mRNA; lane 4, TNFa mRNA; lane 5, LT and TNFa mRNA; lane 6-9, LT and TNFa m R N A + O D N AB at 0.1 /~M (lane 6); 0.5 /xM (lane 7); 1 /xM (lane 8) and 5 /zM (lane 9); lanes 10-13, LT and TNFa m R N A + O D N ABS at 0.1 /xM (lane 10); 0.5 /zM (lane 11); 1 /xM (lane 12) and 5 /~M (lane 13); lane 14, LT and TNFa mRNA + ODN MTS at 5/xM. AB is a short ODN complementary to TNFot and LT. ABS is the corresponding sense ODN. MTS is an unrelevant oligonucleotide (sense sequence of MT). Densitometry was expressed as percent control (lanes 3, 4 and 5). Typical autoradiograph representative of three different experiments.
Serum presence in culture media was described as being a major source of oligonucleotide degradation [27]. When the serum was heat inactivated 30 min at 56°C, oligonucleotide was rapidly degraded and no more intact oligonucleotide remained after 4 h of incubation at 37°C. In contrast, heat inactivation of FCS for 1 h at 65°C suppressed nuclease activity, and oligonucleotide were stable up to 48 h (data not shown). In all cell culture experiments, medium was supplemented with FCS which has been heat inactivated for 1 h at 65°C. Two cell lines were chosen in order to study T N F a and LT production in presence of ODN. We found that at least three T N F a ODN and two LT ODN had an inhibitory activity. On the one hand, U937 which is a monocytic cell line produced T N F a when stimulated with PMA + calcium ionophore. An ODN located in the 5'-untranslated region (ODN T027), another surrounding the initiating A U G (ODN T180) and ODN T195 which is in the coding
172
C. Lefebvre d'Hellencourt et al. /Biochimica et Biophysica Acta 1317 (1996) 168-174
100
region (ODN T168 and T183), or in the coding region (T188) were also effective, but exhibited a less pronounced activity (30 to 45% at 40/xM). At least, the other ODN in 5' region (ODN T052, T077), one just downstream AUG (ODN T187), one in the 3' region (ODN T920) and control oligonucleotides induced only slight modification in TNFa release ( < 25% inhibition at 40 IzM) (Fig. 5A). On the other hand, the T cell line Jurkat produced LT when stimulated with PMA + PHA. Two ODN located in the translation initiation codon were tested. ODN L066 and ODN L081 were both effective in LT inhibition (65 to 75% of inhibition at 100 /xM), whereas control sequence had no effect ( < 20% of inhibition at 100 /zM) (Fig. 5B). It should be noted that as it was previously described for collagen [28], we found that there was variation in the values obtained with the same ODN.
A 75
.=
50
.=, 25
10
20
40
80
ODN concentration (I.tM) 100 -
4. D i s c u s s i o n
75 50 25O i
0.1
0.5
1
5
10
ODN concentration (I.tM) Fig. 4. Inhibition of TNFa and LT translation in RRL (A) and in RRL with RNase H (B). Percentage of inhibition was determined from TCA precipitations and counting. Percentage of inhibition of TNFa in the presence of ODN T168 (A), of LT in the presence of ODN L081 ( O ) and of BMV in the presence of ODN L081 (©). One representative experiments out of three identically performed experiments is shown.
region, were most effective in inhibiting the TNFa production (55 to 75% of inhibition at 40/zM). ODN targeted to the 5' region (ODN T002 and T102), spanning the AUG 100 ]
In this study, we firstly investigated ODN effects on TNFa and LT production in cell free systems. In order to assess the role of RNase H we used in vitro translation systems with either a high or low RNase H activity. In WGE all the ODN in the coding or in the 5' leader region had a similar inhibition efficiency, probably due to the high RNase H activity. Indeed in RRL, which has a poor RNase H activity, none of the ODN gave a significative inhibition of TNFa, unless RNase H was added. Conversely, a decrease of LT production in RRL was achieved with ODN even without addition of RNase H, but with higher ODN concentration than in presence of RNase H. This is in agreement with a report in which globin was entirely inhibited by 100 /zM of ODN in RRL [29]. However the reason for the difference between ODN efficiency on TNFa and LT production in RRL remain unclear, but could be partly due to differences in the secondary (or even tertiary) structure of LT and TNF RNA in the translational mixture and thus differences in accessibility.
[] 10uM 120uM
8ot .4ou.
A
[] 25BM l II 50l.tM
B1
i
~ 4o 2
Fig. 5. ODN Inhibition of TNFa and LT production by stimulated cells. The experiment was performed as explained in Section 2. A. ODN inhibition of TNFa release by stimulated U937 cells. ODN were added for a 4-h incubation period, PMA + calcium ionophore were then added, and the cells reincubated for a 6-h period. B. ODN inhibition of LT release by stimulated Jurkat cells. ODN were added for a 4-h incubation period, PMA + PHA were then added, and the ceils reincubated for a 24-h period. Each value represents the mean of three separate experiments + SEM.
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174
We also investigated ODN efficiency at cellular level. Among the panel of ODN tested (Fig. 1), we obtained different results depending on the localization. For TNFa, one ODN located downstream of the AUG initiation codon, another one spanning the translation initiation codon and one more in the 5'-untranslated region were the most efficient to inhibit this cytokine production. Furthermore, three sequences used in this work have been previously described for inhibiting TNFa in non tumorigen human B cells (ODN T180) [12], in LPS stimulated monocytes (T183) [19] and in cytotrophoblastic cells (ODN T188) [20]. In our model only T180 ODN show an important inhibitory activity, while the two others were less efficient. The difference of the cellular type used in these studies were probably responsible for the discrepancies of ODN efficiency. Antisense stategy have also been used to demonstrate an autocrine growth regulator activity of TNFa in a murine culture of bone marrow cells [30]. Concerning LT, two ODN spanning the translation initiation codon were tested for their ability to inhibit this protein release by stimulated Jurkat cells. Both of them inhibited LT production, but a higher ODN concentration than for TNFa was needed to observe an inhibition. This is consistent with a recent report in which, interferon Y, another cytokine produced by T lymphocyte was inhibited by specific ODN at concentration up to 250 /xM [31]. The TNF production were monitored by a TNF bioassay using L929 cells sensitized by actinomycin D. It has been recently reported that oligonucleotide could interact with actinomycin D without sequence specificity and leads to a protective effect [32]. However in our hand, the addition of oligonucleotide (up to 200 /zM) in known concentration of TNF did not modify the response of L929 (data not shown). This can be explain by the use of a higher concentration of actinomycin D (4 /zg/ml vs. 1 /xg/ml). Furthermore some supernatant of ODN treated cells were also tested in ELISA, and in these experiments the percentage of inhibition were comparable with the bioassay result (data not shown). It is difficult to predict ODN efficiency, even if secondary structure of RNA may be calculated [33]. Some results supported the hypothesis that single-stranded loop structures were more promising targets for ODN than double-stranded sequences [34,35]. In contrast, for TNFa, we found no apparent correlation between ODN efficacy and predicted mRNA secondary structure. This is in agreement with a recent report concerning inhibition of human collagen by ODN [28]. The two LT ODN we had tested were essentially targeted to a calculated single strand loop structure, and thus it is not possible to conclude on a relation between structure and efficiency for LT. Our results demonstrate that it was possible to inhibit both TNFs with a unique ODN in cell free systems, however before use of this T N F a / L T ODN in cell culture or in vivo, and because of its short size (13 bases), it would be probably necessary to increase affinity of
173
TNFa/LT ODN by modification, like 2' O-alkyloligoribonucleotides. We can also conclude that there is no correlation between the predicted secondary structure of RNA, ODN efficiency in cell free extract, and ODN efficiency in cell culture. However, it is likely that cell-free production of TNFs could be of interest to evaluate new ODN modification, in a simple way. Moreover ODN showing an inhibitory activity could be used to study in vitro further the role of TNFs in pathological situations or in immunological mechanisms.
Acknowledgements We would like to thank Dr. J.J. Toulm6 for helpful criticism and advice, Dr. W. Fiers for providing pSP65hTNF1 and Dr. N. Matsuyama for providing clone LT#11. This work is supported by INSERM (Grant no. 930504). C. Lefebvre d'Hellencourt is a fellow of Institut de Recherche M6dicale Andr6 Demonchy.
References [1] Old, L.J. (1987) Nature 326, 330-331. [2] Browning, J. and Ribolini, A. (1989) J. Immunol. 143, 1859-1867. [3] Aggarwal, B.B. (1990) Cytokines and Lipocortins in Inflammation and Differentiation, pp. 375-384, Wiley-Liss, New York. [4] Paul, N.L. and Ruddle, N.H. (1988) Ann. Rev. Immunol. 6, 407-438. [5] Pennica, D. and Goeddel, D.V. (1987) Lymphokines 13, 163-180. [6] Aggarwal, B.B. (1990) Prog. Clin. Biol. Res. 349, 375-384. [7] Tracey, K.J. (1994) in: The Cytokine Handbook (Thomson, A., ed.), pp. 289-304, Academic Press, New York. [8] McLaughlin, P.J. (1992) Biochem. Inflammat. 18, 211-223. [9] Jaattela, M. (1991) Lab. Invest. 64, 724-742. [10] Ruddle, N.H. and Homer, R. (1988) Prog. Allergy 40, 162-182. [11] Selmaj, K., Raine, C.S., Cannella, B. and Brosnan, C.F. (1991) J. Clin. Invest. 87, 949-954. [12] Abken, H., Fluck, J. and Willecke, K. (1992) J. Immunol. 149, 2785-2794. [13] Seregina, T.M., Mekshenkov, M.I., Turetskaya, R.L. and Nedospasov, S.A. (1989) Mol. Immunol. 26, 339-342. [14] Gibbons, D.L., Rowe, M., Feldman, M. and Brennan, F.M. (1994) Eur. J. Immunol. 24, 1879-1885. [15] Yang, Y., Yelavarthi, K.K., Chen, H.L., Pace, J.L., Terranova, P.F. and Hunt, J.S. (1993) J. Immunol. 150, 5614-5624. [16] Uhlmann, E. and Peyman, A. (1990) Chem. Rev. 90, 543-584. [17] H~l~ne, C. and Toulm~, J.J. (1990) Biochim. Biophys. Acta. 1049, 99-125. [18] Lefebvre d'Hellencourt, C., Diaw, L. and Guenounou, M. (1995) Eur. Cytokine Netw. 6, 7-19. [19] Zucali, J.R. and Moreb, J. (1990) In Progress in Leukocyte Biology, Vol. 10A (Oppenheim, J.J., Powanda, M.C., Kluger M.J. and Dinarello, C.A., eds.), pp. 315-320, Wiley-Liss, New York. [20] Yang, Y., Yelavarthi, K.K., Chen, H.L., Pace, J.L., Terranova, P.F. and Hunt, J.S. (1993) J. Immunol. 150, 5614-5624. [21] Muller, R., Marmenout, A. and Fiers, W. (1986) FEBS Lett. 197, 99-104. [22] Matsuyama, N., Okawa, N., Tsukii, Y., Endo, T. and Kaji, A. (1992) FEBS Lett. 302, 141-144. [23] Verspieren, P., Loreau, N., Thuong, N.T., Shire, D. and Toulm6, J.J. (1990) Nucl. Acids Res. 18, 4711-4717.
174 [24] [25] [26] [27]
C. Lefebvre d'Hellencourt et aL /Biochimica et Biophysica Acta 1317 (1996) 168-174
Laemmli, U.K. (1970) Nature 277, 680-685. Fish, H. and Gifford, G.E. (1983) J. Immunol. Methods. 57, 311-325. Minshull, J. and Hunt, T. (1986) Nucl. Acids Res. 14, 6433-6451. Verspieren, P., Cornelissen, A.W.C.A., Thuong, N.T., H~l~ne, C. and Toulm~, J.J. (1987) Gene 61,307-315. [28] Laptev, A.V., Lu, Z., Colige, A. and Prockop, D.J. (1994) Biochemistry 33, 11033- 11039. [29] Cazenave, C., Stein, C.A., Loreau, N., Thuong, N.T., Neckers, L.M., Subasinghe, C., H~l~ne, C., Cohen, J.S. and Toulm~, J.J. (1989) Nucl. Acids Res. 11, 4255-4273.
[30] Witsell, A.L. and Schook, L.B. (1992) Proc. Natl. Acad. Sci. 89, 4754-4758. [31] Boeve, M.A. and De Ley, M. (1994) J. Leukoc. Biol. 55, 169-174. [32] Stull, R.A., Zon, G. and Szoka, F.S. (1993) Antisense Res. Dev. 3, 295-300. [33] Zucker, M. (1989) Methods Enzymol. 180, 262-288. [34] Lima, W.F., Monia, B.P., Ecker, D.J. and Freier, S.M. (1992) Biochemistry. 31, 12055- 12061. [35] Thierry, A.R., Rahman, A. and Dritschilo, A. (1993) Biochem. Biophys. Res. Commun.190, 952-960.
et Biophysica t~ta Bioc himica et Biophysica Acta 1317 (1996) 168 - 174
Inhibition of human TNFa and LT in cell-free extracts and in cell culture by antisense oligonucleotides Christian Lefebvre d'Hellencourt
a,*,
Ltna Diaw b, Pascale Comillet
a,
Moncef Guenounou
a
a Laboratoire d'H~matologie et Laboratoire de Biologie des Cytokines, C.H.R. Robert Debr~, rue Alexis Carrel, 51092 Reims cedex, France b Unit~ d'Immuno-H~.matologie et d'lmmuno-Pathologie, lnstitut Pasteur, 25 rue du docteur Roux, 75724 Paris cedex 15, France
Received 22 August 1996; accepted 10 September 1996
Abstract Antisense oligonucleotides (ODN), complementary to mRNA of human tumor necrosis factor a (TNFa) and lymphotoxin (LT) were tested for their ability to inhibit TNFs. TNFs production was studied in cell-free systems including wheat germ extract (WGE) and rabbit reticulocyte lysate (RRL). All ODN were effective in WGE at low concentration (0.2 /xM), except those targeted to the 3' region of TNFa mRNA. A short ODN complementary to a common region between TNFa and LT inhibited both TNFs. In contrast, high ODN concentration (50 /xM) was needed to inhibit LT mRNA translation in RRL, whereas no clear inhibition of TNFa was observed unless RNase H was added to the translation mixture. ODN effects on TNFs production by stimulated cell line in culture were also investigated. Three ODN - - one located in the Y-untranslated region, one spanning the AUG initiation codon and one downstream of this AUG - were the most effective sequences to decrease TNFa production. Two ODN targeted to the AUG initiation codon of LT were also able to inhibit its production. In conclusion we confirm the role of RNase H in cell free systems, and we found that there is no correlation between ODN efficiency in a cell-free system nor in cell culture. Efficient ODN could be used for in vitro investigation of the role of TNFa and LT in mechanism in which they are involved. Keywords: Antisense oligonucleotide; Cytokine; TNF; Lymphotoxin; Cell free extract
1. Introduction Tumor necrosis factor a ( T N F a ) and lymphotoxin or tumor necrosis factor /3 (LT, L T a , or TNF/3) are two closely related cytokines involved in many regulatory activities [1,2]. Almost all cells of the immune system (including macrophages, B and T lymphocytes, natural killer cells, and mast cells) and cells that are not part of the immune system (e.g., smooth muscle cells, astrocytes, granulosa cells, epithelial cells, breast and ovarian tumor cells) can produce T N F a [3]. LT, in contrast, is mainly secreted by lymphocytes [4]. T N F a and LT are each
Abbreviations: TNF, tumor necrosis factor; LT, lymphotoxin; PMA, phorbol 12-myristate 13-acetate; PHA, phytohemagglutinin; ODN, antisense oligonucleotides; ELISA, enzyme-linked immunosorbent assay; BMV, brome mosaic virus; WGE, wheat germ extract; RRL, rabbit reticulocyte lysate; FCS, fetal calf serum; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TCA, trichloroacetic acid * Corresponding author. Present address: INSERM U298, CHU, 4 rue Larrey, 49033 Angers cedex 01, France. Fax: + 33 2 41354729. E-mail: hellenco @infobiogen.fr
encoded by unique genes and share approximately 30 percent amino acid homology [5]. They have very similar but not identical biological activities [6]. In a number of pathology such as septic shock, rheumatoid arthritis, allograft rejection, parasitic diseases, meningococcal disease and AIDS a high level of T N F a is released in the plasma [7]. T N F a and LT are involved in inflammatory processes, in pathologies like multiple sclerosis or AIDS dementia [8-11], and may also induce proliferation as autocrine growth factor for B cell lines [12-14] and for cytotrophoblastic cell [15]. Thus it may be of interest to develop specific inhibitor of TNFs to investigate their role in such disorders and to propose new therapies; this can be achieved by antisense strategy. Antisense oligodeoxyribonucleotides (ODN) have been used to block translation of protein in cell-free systems, in cultured cells, and in vivo [16,17]. Numerous studies demonstrate the regulation of cytokine production by this methodology [18], but to our knowledge ODN have neither yet been applied to regulate T N F a and LT production in cell-free systems, nor to regulate LT in cell culture.
0925-4439/96//$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PH S 0 9 2 5 - 4 4 3 9 ( 9 6 ) 0 0 0 5 1-8
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174 5'
AUG
TNFtx mRNA
'I'002 T052 TI02 T152 'I"168 T187 AB T027 '1"077 T127 T180 T188 T183 T195
3'
5'
T920
AUG
L066
LT mRNA
L081
Untranslatedregion
~
169
AB m
Coding region
Fig. 1. Schematic illustration of location of the oligonucleotides.
2. Materials and methods
washed with 75% ethanol, and reconstituted in water. An aliquot of each sample was labelled at the 5'-end using T4 polynucleotide kinase and [32p]ATP and analysed for homogeneity on a 20% polyacrylamide gel. Concentration of oligonucleotide solution was determined spectrophotometrically. Schematic localization and sequence of oligonucleotide are shown in Fig. 1 and in Table 1. ODN T180, T183 and T188 have been previously described [12,19,20]. Position 183 to 185 and 81 to 83 correspond to the AUG codon of human TNFc~ and LT respectively.
2.1. Reagents
2.3. Plasmids
In the present study, the action of different ODN on the synthesis of TNFc~ and LT was investigated, both in cell free system and in cell culture. The importance of RNase H was confirmed in cell free extract and the possibility to predict AS efficiency in cell culture, by calculated secondary structure or by using cell free extract was also studied.
Phorbol 12-myristate 13-acetate (PMA), phytohemagglutinin (PHA-M), and calcium ionophore A23187 were purchased from Sigma Chemicals (Saint Quentin FaUavier, France) and used at 10 ng/ml, 10 /xg/ml and 0.5 /xM, respectively.
2.2. Oligonucleotides Unmodified oligonucleotides were synthesized by Genset (Paris, France) on an automated solid-phase synthesizer (Applied Biosystems) using phosphoramidite chemistry. Oligonucleotides were ethanol precipitated twice,
The construction PSP65-hTNF1 containing the TNF~ cDNA in SP6 orientation was obtained from the Laboratory of molecular biology-plasmid collection University of Gent [21]. LT cDNA clone LT#11 is a generous gift from N. Matsuyama [22]. The EcoRI fragment of LT#11 was subcloned into pGEM4Z plasmid (Promega, Charbonni~res, France). The construction containing LT cDNA in a SP6 orientation was designated pGEM4Z-hLT1.
2.4. In vitro synthesis of capped transcripts Plasmids pSP65-hTNF1 and pGEM4Z-hLT1 linearized, respectively, at the unique NcoI site and HindlII, were
Table 1 Sequence of oligonucleotides Oligonucleotide
Length
Sequence (5' --, 3')
Target site
Target
T002
l8
GAGGGAGCGTCTGCTGGC
T027 T052 T077
18 18 18
5'-untranslated 5'-untranslated 5'-untranslated 5'-untranslated
T 102 T 127 T152
18 18 18
TAGCTGGTCCTCTGCTGT GGTCTGTAGTrGCTTCTC GCGTCTGAGGGTTGI-ITI GCCTGGCAGCTTGTCAGG
T168 * T180 T183 * T187 T 188 T 195 AB * T920 L066
18 15 18 15 17 18 13 18 18
TCATGGTGTCCTTTCCAG TTCAGTGCTCATGGT
hTNFa hTNFa hTNFot hTNFa hTNFa hTNFo~ hTNFa hTNFot hTNFa
L081 MT *
18 20
ACGT'I'CAGGTGGTGTCA T TCATGGTGTCTITrCTGGAG
TCAGTATGTGAGAGGAAG AGGGGAGAGAGGGTGAAG
CATGCTTTCAGTGCTCAT TCATGC I I I C A G T G C GGATCATGCTTI'CAGTG CCACGTCCCGGATCATGC ACCTGGGAGTAGA ATI'GGGGCAGGGGAGGCG CA TGGGGAGAACCCAGGC
5'-untranslated
5'-untranslated 5'-untranslated AUG codon AUG codon AUG codon coding region E 1 coding region E 1 coding region E 1 coding region E4 3' untranslated AUG codon AUG codon AUG codon
hTNFot hTNFot hTNFa hTNFct hTNFot/hLT hTNFa hLT hLT mTNFt~
Underlining indicates sequence complementary to the initiation codon. El, exon 1; E,4, exon 4; h, human; m, murin. * Sense and/or random sequence of these oligonucleotide were also synthesized and used as control.
170
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174
used as a template for run-off transcription. In vitro transcription was carried out essentially as recommended by the supplier (Promega). Briefly 5 /zg of linearized plasmid were transcribe in 10/~1 of 5 × buffer (200 mM Tris-HC1, pH 7.5; 30 mM MgC12; 10 mM spermidine; 50 mM NaC1), 50 U of RNasin TM, 0.5 A260 units of m7(GS)ppp(5')G cap structure (Boerhinger Mannheim, Meylan, France), 250 /zM GTP, 500 /zM of the 3 other NTPs (Promega) and 40 U of SP6 polymerase. After 1 h at 37°C, the mixture was incubated with DNase I 15 min at 37°C and was phenol-chloroform extracted. Unincorporated nucleotides were removed by Sephadex G-50 columms (Boehringer Mannheim). 2.5. Translation in cell-free system
Nuclease-treated wheat germ extract (WGE) and rabbit reticulocyte lysate (RRL) were purchased from Promega. Cell-free translation was carried out essentially as recommended by the supplier. 70 ng of in vitro synthesized capped transcript solution was used per 10 /~1 of translational mixture, and 0.150 MBq of [35S]methionine (37 TBq/mmol, Amersham, Les Ulis, France) was used for labelling. Brome mosaic virus (BMV) RNA was used as control of translation. The reaction was run 1 h at 25°C for WGE and 1 h at 30°C for RRL. The 35S-labelled translation products were analyzed either by trichloroacetic acid (TCA) precipitation [23] or by electrophoretic separation on 15% SDS-polyacrylamide slab gels (SDS-PAGE) [24], and autoradiography. Autoradiographs were quantified using laser densitometry. 2.6. Cell culture
Lymphoblastic T (Jurkat) and monocytic (U937) cell lines were maintained in RPMI 1640 supplemented with L-glutamine, penicillin-streptomycin and 10% heat inactivated fetal calf serum (FCS) (Gibco BRL, Eragny, France). After three washes, cells were resuspended at 5 × 104 cells in 0.1 ml of culture medium and plated in 96 well plates. Various concentrations of oligonucleotide (5 /zM-100 /xM) were added for a 4-h incubation period, stimuli were then added, and the cells reincubated at 37°C and in 5% CO 2. Supernatants were collected after 6 h and 24 h for U937 and Jurkat culture respectively and stored at - 8 0 ° C until assayed for cytokine production. Cell viability was monitored by trypan blue exclusion.
ml). The cells were then incubated at 37°C for 18 h, and extensively washed afterwards in Hanks' balanced salt solution. The remaining adherent cells were stained for 10 min with crystal violet (0.2% in 2% ethanol), washed with water and 0,1 ml of 1% SDS was added to each well to solubilize the stained cells. The absorbance of each well was read at 550 nm. The percentage of cytotoxicity was defined as the relative absorbance of the tested sample versus the absorbance of the control (medium only) wells. %cytotoxicity = 1 -
( absorbance samples ) absorbance control × 100
The cytotoxicity titer (U/ml) was defined as the reciprocal dilution of the test sample inducing 50% of cytotoxicity. Recombinant TNF was used as a positive control. In some experiments, the level of TNFa and LT were determined alternatively by using commercialized kits. TNFa ELISA was purchased from Medgenix (Fleurus, Belgium) and LT ELISA was purchased from R & D Systems Europe (Abingdon, UK).
3. Results 3.1. Effects o f oligonucleotides on TNFs translation in cell-free extracts 3.1.1. Translation in wheat germ extract
In vitro synthesized, capped transcripts of human TNF o r / a n d human LT were translated in WGE in the presence or absence of oligonucleotides. We found that ODN targeted to the 5', AUG, or coding region successfully arrested translation of TNFa, whereas an ODN in the 3' untranslated region (ODN T920) and control sequences (ODN T168S, ABS and MTS) had no inhibitory effects. Furthermore, ODN in the AUG region of LT (ODN L066 and L081) were also effective at inhibiting LT translation. All inhibitory ODN displayed similar efficiency and about 0.2 /xM led to a 50% decrease of translation as determined by TCA precipitation and SDS-PAGE (Fig. 2A). Inhibition of translation was specific, because when the two RNA were co-translated in presence of ODN it was possible to inhibit one of them without modifying the other (at 0.5 and 1 /xM) (Fig. 2B). We also tested a short ODN (ODN AB) which is common to TNFa and LT; we observed that both TNFs were inhibited (at 1 and 5 /zM), whereas the sense sequence (ABS) or an unrelated sequence (MTS) had no effect on TNFs translation (Fig. 3).
2.7. TNFs detection 3.1.2. Translation in rabbit reticulocyte lysate
TNF assays were performed using the sensitive L929 cells as describe by Fish and Gifford [25]. Briefly, 5 × 10 4 L-929 cells/well suspended in 0.1 ml RPMI 1640 containing 10% FCS were seeded in 96-well microtitre plates. Actinomycin D (Sigma) at 4 /xg/ml final concentration was added prior to the addition of the sample dilutions (0.1
The effects of ODN on translation of TNFa in RRL were variable, and no significant inhibition was observed until addition of RNase H (20 U / m l ) in the reaction mixture as previously described [26]. Thus, after RNase H addition, a similar pattern as in WGE was obtained, and the dose of ODN necessary for 50% of inhibition (ID
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174 1
2
3
4
5
6
7
8
9
I0
1
2
171
3
4
5
6
7
8
9
3
4
5
6
7
8
9
TNFot LT~
B I00_
75 ~
LT
25
2
0 3
4
5
6
7
8
9 10
Fig. 2. Inhibitionof TNFa and LT translationin WGE. Samples were analysedon SDS-PAGEand autoradiographed as indicatedin Section2. A: Lane 1: rio RNA, lane 2: BMV mRNA, lane 3: LT mRNA, lane 4-6: LT mRNA+ ODN L081 at 0.1 txM (lane 4), 0.5 /xM (lane 5) and 1 txM (lane 6), lane 7: TNFa mRNA, lane 8-10: TNFa mRNA + ODN T168 at 0.1 txM (lane 8), 0.5 ~M (lane 9) and 1 /xM (lane 10). Similar inhibitionof translationwas observed with ODN T102, T127, T152, T168, T180, T183, T187, T188, T195 for TNFa and with L066 for LT. Densitometrywas expressed as percent control (lane 3 for LT and lane 7 for TNFo~). B: Lane 1: no RNA, lane 2: BMV mRNA, lane 3: LT and TNFa mRNA, lane 4-6: LT and TNFa mRNA + ODN L081 at 0.1 NM (lane 4), 0.5 txM (lane 5) and 1 IxM (lane 6), lane 7-9: LT and TNFa mRNA+ ODN T168 at 0.1 ~M (lane 7), 0.5 IxM (lane 8) and 1 ~M (lane 9). Densitometrywas expressed as percent control (lane 3). Typical autoradiographrepresentativeof three differentexperiments.
1 2
3
4
5
6
7
8
9 10 11 12 13 14
TNFct -~------ LT
50%) was also about 0.2 /xM (Fig. 4B). Besides, we found a difference with LT; its production was inhibited by both LT ODN at high concentration ( 2 0 - 8 0 /zM) even without RNase H addition, and ID 50% was about 40/xM (Fig. 4A). This LT inhibition did not result from a general effect on in vitro translation, as these ODNs did not induce any decrease of BMV protein synthesis at ODN concentrations up to 80 /zM (Fig. 4A).
3.2. Effects of oligonucleotides on stimulated cells
125.
iI1111,I1[11 2
.
.
3
.
4
.
.
5
.
6
R •
m
7
8
.
.
.
.
.
.
.
9 10 11 12 13 14
LT TNF~
Fig. 3. Inhibition of both TNFs translation in WGE. Lane 1, no RNA; lane 2, BMV mRNA; lane 3, LT mRNA; lane 4, TNFa mRNA; lane 5, LT and TNFa mRNA; lane 6-9, LT and TNFa m R N A + O D N AB at 0.1 /~M (lane 6); 0.5 /xM (lane 7); 1 /xM (lane 8) and 5 /zM (lane 9); lanes 10-13, LT and TNFa m R N A + O D N ABS at 0.1 /xM (lane 10); 0.5 /zM (lane 11); 1 /xM (lane 12) and 5 /~M (lane 13); lane 14, LT and TNFa mRNA + ODN MTS at 5/xM. AB is a short ODN complementary to TNFot and LT. ABS is the corresponding sense ODN. MTS is an unrelevant oligonucleotide (sense sequence of MT). Densitometry was expressed as percent control (lanes 3, 4 and 5). Typical autoradiograph representative of three different experiments.
Serum presence in culture media was described as being a major source of oligonucleotide degradation [27]. When the serum was heat inactivated 30 min at 56°C, oligonucleotide was rapidly degraded and no more intact oligonucleotide remained after 4 h of incubation at 37°C. In contrast, heat inactivation of FCS for 1 h at 65°C suppressed nuclease activity, and oligonucleotide were stable up to 48 h (data not shown). In all cell culture experiments, medium was supplemented with FCS which has been heat inactivated for 1 h at 65°C. Two cell lines were chosen in order to study T N F a and LT production in presence of ODN. We found that at least three T N F a ODN and two LT ODN had an inhibitory activity. On the one hand, U937 which is a monocytic cell line produced T N F a when stimulated with PMA + calcium ionophore. An ODN located in the 5'-untranslated region (ODN T027), another surrounding the initiating A U G (ODN T180) and ODN T195 which is in the coding
172
C. Lefebvre d'Hellencourt et al. /Biochimica et Biophysica Acta 1317 (1996) 168-174
100
region (ODN T168 and T183), or in the coding region (T188) were also effective, but exhibited a less pronounced activity (30 to 45% at 40/xM). At least, the other ODN in 5' region (ODN T052, T077), one just downstream AUG (ODN T187), one in the 3' region (ODN T920) and control oligonucleotides induced only slight modification in TNFa release ( < 25% inhibition at 40 IzM) (Fig. 5A). On the other hand, the T cell line Jurkat produced LT when stimulated with PMA + PHA. Two ODN located in the translation initiation codon were tested. ODN L066 and ODN L081 were both effective in LT inhibition (65 to 75% of inhibition at 100 /xM), whereas control sequence had no effect ( < 20% of inhibition at 100 /zM) (Fig. 5B). It should be noted that as it was previously described for collagen [28], we found that there was variation in the values obtained with the same ODN.
A 75
.=
50
.=, 25
10
20
40
80
ODN concentration (I.tM) 100 -
4. D i s c u s s i o n
75 50 25O i
0.1
0.5
1
5
10
ODN concentration (I.tM) Fig. 4. Inhibition of TNFa and LT translation in RRL (A) and in RRL with RNase H (B). Percentage of inhibition was determined from TCA precipitations and counting. Percentage of inhibition of TNFa in the presence of ODN T168 (A), of LT in the presence of ODN L081 ( O ) and of BMV in the presence of ODN L081 (©). One representative experiments out of three identically performed experiments is shown.
region, were most effective in inhibiting the TNFa production (55 to 75% of inhibition at 40/zM). ODN targeted to the 5' region (ODN T002 and T102), spanning the AUG 100 ]
In this study, we firstly investigated ODN effects on TNFa and LT production in cell free systems. In order to assess the role of RNase H we used in vitro translation systems with either a high or low RNase H activity. In WGE all the ODN in the coding or in the 5' leader region had a similar inhibition efficiency, probably due to the high RNase H activity. Indeed in RRL, which has a poor RNase H activity, none of the ODN gave a significative inhibition of TNFa, unless RNase H was added. Conversely, a decrease of LT production in RRL was achieved with ODN even without addition of RNase H, but with higher ODN concentration than in presence of RNase H. This is in agreement with a report in which globin was entirely inhibited by 100 /zM of ODN in RRL [29]. However the reason for the difference between ODN efficiency on TNFa and LT production in RRL remain unclear, but could be partly due to differences in the secondary (or even tertiary) structure of LT and TNF RNA in the translational mixture and thus differences in accessibility.
[] 10uM 120uM
8ot .4ou.
A
[] 25BM l II 50l.tM
B1
i
~ 4o 2
Fig. 5. ODN Inhibition of TNFa and LT production by stimulated cells. The experiment was performed as explained in Section 2. A. ODN inhibition of TNFa release by stimulated U937 cells. ODN were added for a 4-h incubation period, PMA + calcium ionophore were then added, and the cells reincubated for a 6-h period. B. ODN inhibition of LT release by stimulated Jurkat cells. ODN were added for a 4-h incubation period, PMA + PHA were then added, and the ceils reincubated for a 24-h period. Each value represents the mean of three separate experiments + SEM.
C. Lefebvre d'Hellencourt et al. / Biochimica et Biophysica Acta 1317 (1996) 168-174
We also investigated ODN efficiency at cellular level. Among the panel of ODN tested (Fig. 1), we obtained different results depending on the localization. For TNFa, one ODN located downstream of the AUG initiation codon, another one spanning the translation initiation codon and one more in the 5'-untranslated region were the most efficient to inhibit this cytokine production. Furthermore, three sequences used in this work have been previously described for inhibiting TNFa in non tumorigen human B cells (ODN T180) [12], in LPS stimulated monocytes (T183) [19] and in cytotrophoblastic cells (ODN T188) [20]. In our model only T180 ODN show an important inhibitory activity, while the two others were less efficient. The difference of the cellular type used in these studies were probably responsible for the discrepancies of ODN efficiency. Antisense stategy have also been used to demonstrate an autocrine growth regulator activity of TNFa in a murine culture of bone marrow cells [30]. Concerning LT, two ODN spanning the translation initiation codon were tested for their ability to inhibit this protein release by stimulated Jurkat cells. Both of them inhibited LT production, but a higher ODN concentration than for TNFa was needed to observe an inhibition. This is consistent with a recent report in which, interferon Y, another cytokine produced by T lymphocyte was inhibited by specific ODN at concentration up to 250 /xM [31]. The TNF production were monitored by a TNF bioassay using L929 cells sensitized by actinomycin D. It has been recently reported that oligonucleotide could interact with actinomycin D without sequence specificity and leads to a protective effect [32]. However in our hand, the addition of oligonucleotide (up to 200 /zM) in known concentration of TNF did not modify the response of L929 (data not shown). This can be explain by the use of a higher concentration of actinomycin D (4 /zg/ml vs. 1 /xg/ml). Furthermore some supernatant of ODN treated cells were also tested in ELISA, and in these experiments the percentage of inhibition were comparable with the bioassay result (data not shown). It is difficult to predict ODN efficiency, even if secondary structure of RNA may be calculated [33]. Some results supported the hypothesis that single-stranded loop structures were more promising targets for ODN than double-stranded sequences [34,35]. In contrast, for TNFa, we found no apparent correlation between ODN efficacy and predicted mRNA secondary structure. This is in agreement with a recent report concerning inhibition of human collagen by ODN [28]. The two LT ODN we had tested were essentially targeted to a calculated single strand loop structure, and thus it is not possible to conclude on a relation between structure and efficiency for LT. Our results demonstrate that it was possible to inhibit both TNFs with a unique ODN in cell free systems, however before use of this T N F a / L T ODN in cell culture or in vivo, and because of its short size (13 bases), it would be probably necessary to increase affinity of
173
TNFa/LT ODN by modification, like 2' O-alkyloligoribonucleotides. We can also conclude that there is no correlation between the predicted secondary structure of RNA, ODN efficiency in cell free extract, and ODN efficiency in cell culture. However, it is likely that cell-free production of TNFs could be of interest to evaluate new ODN modification, in a simple way. Moreover ODN showing an inhibitory activity could be used to study in vitro further the role of TNFs in pathological situations or in immunological mechanisms.
Acknowledgements We would like to thank Dr. J.J. Toulm6 for helpful criticism and advice, Dr. W. Fiers for providing pSP65hTNF1 and Dr. N. Matsuyama for providing clone LT#11. This work is supported by INSERM (Grant no. 930504). C. Lefebvre d'Hellencourt is a fellow of Institut de Recherche M6dicale Andr6 Demonchy.
References [1] Old, L.J. (1987) Nature 326, 330-331. [2] Browning, J. and Ribolini, A. (1989) J. Immunol. 143, 1859-1867. [3] Aggarwal, B.B. (1990) Cytokines and Lipocortins in Inflammation and Differentiation, pp. 375-384, Wiley-Liss, New York. [4] Paul, N.L. and Ruddle, N.H. (1988) Ann. Rev. Immunol. 6, 407-438. [5] Pennica, D. and Goeddel, D.V. (1987) Lymphokines 13, 163-180. [6] Aggarwal, B.B. (1990) Prog. Clin. Biol. Res. 349, 375-384. [7] Tracey, K.J. (1994) in: The Cytokine Handbook (Thomson, A., ed.), pp. 289-304, Academic Press, New York. [8] McLaughlin, P.J. (1992) Biochem. Inflammat. 18, 211-223. [9] Jaattela, M. (1991) Lab. Invest. 64, 724-742. [10] Ruddle, N.H. and Homer, R. (1988) Prog. Allergy 40, 162-182. [11] Selmaj, K., Raine, C.S., Cannella, B. and Brosnan, C.F. (1991) J. Clin. Invest. 87, 949-954. [12] Abken, H., Fluck, J. and Willecke, K. (1992) J. Immunol. 149, 2785-2794. [13] Seregina, T.M., Mekshenkov, M.I., Turetskaya, R.L. and Nedospasov, S.A. (1989) Mol. Immunol. 26, 339-342. [14] Gibbons, D.L., Rowe, M., Feldman, M. and Brennan, F.M. (1994) Eur. J. Immunol. 24, 1879-1885. [15] Yang, Y., Yelavarthi, K.K., Chen, H.L., Pace, J.L., Terranova, P.F. and Hunt, J.S. (1993) J. Immunol. 150, 5614-5624. [16] Uhlmann, E. and Peyman, A. (1990) Chem. Rev. 90, 543-584. [17] H~l~ne, C. and Toulm~, J.J. (1990) Biochim. Biophys. Acta. 1049, 99-125. [18] Lefebvre d'Hellencourt, C., Diaw, L. and Guenounou, M. (1995) Eur. Cytokine Netw. 6, 7-19. [19] Zucali, J.R. and Moreb, J. (1990) In Progress in Leukocyte Biology, Vol. 10A (Oppenheim, J.J., Powanda, M.C., Kluger M.J. and Dinarello, C.A., eds.), pp. 315-320, Wiley-Liss, New York. [20] Yang, Y., Yelavarthi, K.K., Chen, H.L., Pace, J.L., Terranova, P.F. and Hunt, J.S. (1993) J. Immunol. 150, 5614-5624. [21] Muller, R., Marmenout, A. and Fiers, W. (1986) FEBS Lett. 197, 99-104. [22] Matsuyama, N., Okawa, N., Tsukii, Y., Endo, T. and Kaji, A. (1992) FEBS Lett. 302, 141-144. [23] Verspieren, P., Loreau, N., Thuong, N.T., Shire, D. and Toulm6, J.J. (1990) Nucl. Acids Res. 18, 4711-4717.
174 [24] [25] [26] [27]
C. Lefebvre d'Hellencourt et aL /Biochimica et Biophysica Acta 1317 (1996) 168-174
Laemmli, U.K. (1970) Nature 277, 680-685. Fish, H. and Gifford, G.E. (1983) J. Immunol. Methods. 57, 311-325. Minshull, J. and Hunt, T. (1986) Nucl. Acids Res. 14, 6433-6451. Verspieren, P., Cornelissen, A.W.C.A., Thuong, N.T., H~l~ne, C. and Toulm~, J.J. (1987) Gene 61,307-315. [28] Laptev, A.V., Lu, Z., Colige, A. and Prockop, D.J. (1994) Biochemistry 33, 11033- 11039. [29] Cazenave, C., Stein, C.A., Loreau, N., Thuong, N.T., Neckers, L.M., Subasinghe, C., H~l~ne, C., Cohen, J.S. and Toulm~, J.J. (1989) Nucl. Acids Res. 11, 4255-4273.
[30] Witsell, A.L. and Schook, L.B. (1992) Proc. Natl. Acad. Sci. 89, 4754-4758. [31] Boeve, M.A. and De Ley, M. (1994) J. Leukoc. Biol. 55, 169-174. [32] Stull, R.A., Zon, G. and Szoka, F.S. (1993) Antisense Res. Dev. 3, 295-300. [33] Zucker, M. (1989) Methods Enzymol. 180, 262-288. [34] Lima, W.F., Monia, B.P., Ecker, D.J. and Freier, S.M. (1992) Biochemistry. 31, 12055- 12061. [35] Thierry, A.R., Rahman, A. and Dritschilo, A. (1993) Biochem. Biophys. Res. Commun.190, 952-960.