J. Mol. Biol. (1992) 223, 831-835

Preformed Ribozyme Destroys Tumour Necrosis Factor mRNA in Human Cells Mouldy ‘Institute

Sioud1v2t, Jacob B. Natvig’

and Bystein

Fmre3

of Immunology and Rheumatology, University of Oslo Fr. Qramsgt 1, N-0172 Oslo 1, Norway

2Public Health Research Institute, 455 First Avenue New York, NY 10016, U.S.A. 3The Rheumatism Hospital, Oslo, Norway (Received 8 July 1991; accepted 13 November 1991) Maintaining RNA stabi1it.y is a major problem in the delivery of preformed inhibitory RNA to target cells. In this study, we delivered a hammerhead ribozyme directed against tumour necrosis factor a into human promyelocytic leukaemia cells by cationic liposome-mediated transfection. Delivering a ribozyme in this manner reduced by 90% and 85% tumour necrosis factor ~1 mRNA and protein, respectively. A modified ribozyme wit,h a bacteriophage T7 transcription terminator at its 3’ end was more stable than one lacking this sequence. This indicates that ribozyme stability can be improved by the addition of terminal sequences expected to protect against cellular nucleases.

Keywords: ribozyme;

stability;

tumour

necrosis factor

CI

showing that a ribozyme designed to cleave the integrase gene of HIV-l is effective when transcribed from a plasmid in Escherichia coli. Integrase RNA was eliminated and integrase protein synthesis was blocked (Sioud & Drlica, 1991). Since ribozymes are effective in vivo, we are now addressing problems of ribozyme stability and delivery. As a first step to interfere with tumour necrosis factor M: (TNFaS) gene expression we have used cationic liposome-mediated transfection (Malone et al., 1989) to deliver a ribozyme directed against TNFu into human promyelocytic leukaemia cells (HL60) and peripheral blood mononuclear cells (PBMNC). TNFcr plays an important role in many inflammatory rheumatic diseases (Shinmei et al., 1989), and it modulates the expression of several proteins, including the class I antigens of the major histocompatibility complex (MHC) and cytokines such as interleukin 1 and interleukin 6 (Beutler & Cerami, 1988, 1989). TNFa also appears to be necessary for normal immune responses, but. large quantities of it can produce destructive effects such as those seen in rheumatoid arthritis (Brennan et al.,

The discovery of RNA molecules that possess enzymatic, self-cleaving activity (ribozymes) has provided a new way to artificially control gene expression (Forster & Symons, 1987; Cech, 1987). Ribozymes have been designed that contain nearly all of the sequences required for cleavage. For the hammerhead type the target RNA needs to contain only the sequence GUX, with cleavage occurring 3’ from X (X = A, C or U: Haseloff & Gerlach, 1988). The high specificity and limited target requirement give these catalytic RNA molecules the potential for inhibiting viral pathogens and for regulating specific gene expression by interfering with transcription in a highly specific manner (Uhlenbeck, 1987; Haseloff & Gerlach, 1988). Several reports indicate that the hammerhead type of ribozyme functions in living cells. Cotten & Birnstiel (1989) and Cameron & Jennings (1989) reported ribozyme-mediated destruction and lowering of specific gene expression in Xenopus Zaevis oocytes and monkey (COSl) cells, respectively. Sarver et al. (1990) showed that a ribozyme directed against HIV-l gag RNA reduced p24 antigen expression in CD4+ HeLa cells. Recently, we extended this line of study to bacterial cells by

$ Abbreviations used: TNFa, tumour necrosis factor a; PBMNC, peripheral blood mononuclear cells; MHC, major histocompatability complex; PMA, phorbol

t Author to whom all correspondence should be addressed at. address I.

12-myristate

13.acetate;

ConA,

concavalin

A.

831 0022-2836/92/040831HE

$03.00/O

0 1992 Academic Press Limited

M. Sioud et al.

032

Rlbozyme

A

U

3~

G~‘J

G

c

A

G

Cat.3ytlC LGAGUACUAAGAUGAUCU Gucc domain cUUCUACUAGA cAGGAG CkW@S c” I G +c we 393 AG UAA A A GC c AU TNFa RNA L G” G C AG c

U

c

374 c u ‘c G c G “2 UGll c A u G

YJ c”

T7

transcr~mx

Targeted ribozymes cut TNFcl RNA in vitro A hammerhead ribozyme (Haseloff & Gerlach. 1988) designed to cleave the TNFcl RNA is shown in

termrnator

AUAACCCCUUGGGGCCUCUA UUUyGGGAGUUCUGGG

CA

U U C

A

c3. (0

1989). In addition, TNFa is the cytokine responsible for the induction of HIV-l expression in ACH-2 cells (Rosenberg & Fauci, 1990). TNFa induces the production of cellular factors that bind to the NF-KB enhancer elements within the viral long terminal repeat sequences and thereby activat’es HIV-1 expression.

)

eO.5

kb

(C)l

Figure 1. (a) Base-pairing of the catalytic strand of ribozyme A with the TNFcr RNA template. Ribozyme A is composed of the conserved ribozyme sequence as described by Haseloff & Gerlach (1988); the 5’ and 3’ flanking sequences complementary to the TNFa RNA (nucleotides 374 and 393: see Pennica et al., 1984 for

numbering) and bacteriophage T7 transcription terminat’or with CU mispair (0) (Rosenberg et al., 1987). Ribozyme B is identical to A except that it lacks the T7 transcription terminator. Antisense RNA is identical to ribozyme A except that it has a single guanosine nucleotide in place of the catalytic domain. The anti-TNFa hammerhead catalytic gene and antisense RNA control were made as described by Sioud & Drlica (1991). Briefly. 2 overlapping half oligonucleotides containing the sequences of a bacteriophage T7 RNA polymerase promoter, the 5’ and 3’ recognition sequences of the ribozyme, the catalytic domain and the T7 transcription terminator were synthesized (we have introduced a XbaI restriction site between the T7 terminator and the 3’ end of the ribozyme, and PvuII and XhoI sites at the 5’ and 3’ ends of the ribozyme sequences, respectively), hybridized and then extended with the Klenow fragment of DNA polymerase. Following the extension, DNA was extracted with phenol, precipitated with ethanol, gel purified and then cloned into a SmaI cleaved pUC 18 vector. The sequences of the overlapping primers (Public Health Research Institut,e. New York 10016. NY) used as follows: (1) Ribozyme primers: SAACAGCTGTAATACGACTCACTATAGAGTACTAA GATGATCTCTGATGAGTCCGTGAGGACGAAACTGC 3’ and 5’ TTCTCGAGAAAAAACCCTCAAGACCCGTTTAGAGG CCCCABGGGGTTATGTCTAGACCAGGCAGTTTCGTCC 3’. (2) Antisense primers: 5’AACAGCTGTAATACGACTCACTATAGAGTACTAA GATGATCTGACTGCCTGGTCTAG 3’ and; 5’ TTCTCGAGAAAAAACCCTCAAGACCCGTTTAGAGG CCCCAAGGGGTTATGTCTAGACCAGCA 3’. (b) In vitro transcription of ribozymes A, B and antisense. Ribozymes and antisense RNA were transcribed with T7 RNA polymerase from PAGE-purified template DNA fragments cleaved from recombinant plasmids as described by Uhlenbeck (1987). RNA was labelled internally with [a3’P]CTP during transcription. Transcription was primed with 7-methyl guanosine (5’) triphospho (5’) guanosine in all cases. Transcripts were treated with DNAse (RNAse-free), extracted with phenol, precipitated with ethanol and then analyzed by electrophoresis in a 15 y0 polyacrylamide gel containing 7 M-urea. The lengths of ribozymes A, B and antisense are 97, 49 and 76 nucleotides, respectively. (c) Cleavage of TNFa RNA by ribozyme in vitro. PBMNC cells were stimulated with PMA and ConA to express TNFa, whole cell RNA was extracted and the RNA (20 pg) was incubated with 1 pg of either ribozyme or antisense RNA for 60 min at 50°C. RNA species were then separated by gel electrophoresis and TNFa RNA was identified by Northern blotting. kb, 103 bases.

Communications

833

Figure l(a). Its in vitro activities were studied using as target total RNA extracted from PBMNC after stimulation of TNFu gene transcription by phorbol I 2-myristate 13-acetate (PMA) and concanavalin A (ConA) as described by English et al. (1991). Two ribozymes generated in in vitro transcription (Fig. l(b)) were examined. Ribozyme A contained a bact,eriophage T7 terminator at its 3’ end while ribozyme B did not. Ribozyme-mediated RNA cleavage was assayed by gel electrophoresis and Northern blot hybridization using TNFcl probe. The TNFa RNA fragments cut by ribozymes A and B are shown in Figure l(c). Both ribozymes cleaved t.he approximately 1800 nucleotide-long target RNA into fragments of 1420 and 380 nucleotides. The sizes of the TNFcl fragments produced by the ribozymes were consistent with the location of predicted site for cleavage. Thus both ribozymes cleave the TXFcr target in vitro in the presence of unrelated RNAs.

Stability of ribozyme,s A and B in living cells Since the cell membrane presents a substantial barrier to the entry of highly charged, high molecular weight molecules, delivering them to the cytoplasm is a major task. To overcome this, transfection techniques such as cationic liposomemediated t,ransfection (Malone et aE., 1989), electroporation (Callis et al., 1987) and microinjection (Rosa et al., 1989) have been developed. Since the liposome-based method appears to be the most versat’ile, we tested its ability to deliver enough functional ribozyme to successfully cleave TNFcx RNA. We first measured the efficiency of RNA transfection in HI,60 cells, as determined by measurement intact 32P-labelled RNA. cell-associated of Following transfection with ribozymes, cells were washed with Hank’s buffered salt solution (GIBCO). Total RNA was prepared from cells and the RNA species were separated by gel electrophoresis. The radioactivity contained in the ribozyme RNA bands was then determined. The results indicate that the radioa&ve bands varied from 2 to 4% of the initial RNA added to the liposomes during a transfection period of between 8 and 20 hours. This corresponded to a delivery of approximately 300,000 molecules of ribozyme A per cell. We then compared the effect of bacteriophage T7 transcription terminator on RNA stability by cotransfecting HL60 cells with ribozymes A and B. Total RNA was extracted and analyzed by electrophoresis in 15 y. (w/v) polyacrylamide gels containing 7 M-Urea. While ribozyme A could be detected more than 72 hours post transfection, the amount of ribozyme B progressively declined (Fig. Z(a). The radioactivity contained in each band was then determined, and the results were expressed as the percentage of the radioactivity at zero time. Figure 2(c) shows that, ribozyme B decays more rapidly than ribozyme A. The residual radioactivity

Figure 2. Ribozyme stability following transfection. Ten million human HL60 cells (ATCC CCL 240), growing in log phase in RPM1 1640 supplemented with 20% (v/v) fetal calf serum (FCS), were used for RNA transfection. Cells were washed twice with serum-free medium. A drop (5 ml) of serum-free medium was added to polystyrene tubes followed by 35 pg of lipofectin (Bethesda Research Laboratories), 10 pg of carrier RNA (E. coli tRNA), 3 x lo6 disints/min of 32P-labelled capped ribozyme A, B or antisense RNA (5 pg). The mixture was immediately mixed. The cells were resuspended in a mixture of serumfree medium/lipofectin/RNA/carrier RNA and returned to the incubator for 20 h. Following transfection cells were washed 3 times with Hank’s buffered saline solution and the returned to the incubator with RPM1 supplemented with 20% FCS. Cells (106) were harvested at the times indicated above each lane, and total RNA was prepared and analyzed by 15% polyacrylamide gel with 7 M-Urea. The RNA samples used for transfection are indicated at the top of the Figure. Ribozyme B alone serves as a marker to indicate its position in co-transfection experiments. (b) Analysis of nuclear (N) and cytoplasmic (C) RNA. A sample (50 PM) of labelled. capped ribozyme A was used to transfect HL60 cells for 20 h. Cells were washed 3 times, and the nuclear and cytoplasmic RNAs were prepared and analyzed by gel electrophoresis. For preparation of cytoplasmic and nuclear fractions, the cells were homogenized in 10 mM-Tris.HCl (pH 7.5). 5 mMKCI, 140 mM-N&l, 5 mM-dithiothreitol and 0.4% (w/v) Nonidet P40 for 10 min at 4”C, and the nuclei. were collected by centrifugation at 800 g for 5 min. RNA in the

supernatant

fluid was precipitated

plasmic fraction. by Chomczynski

and saved as the cyto-

The nuclei were processed as described t Sacchi (1987) for total RNA prepara-

tion. The arrow indicates

the position

of ribozyme

A

monomer. (c) RNA quantification. The amount of radioactivity in the ribozyme bands shown in (a) was determined and expressed as a percentage of the radioactivity present immediately after 20 h transfection time. Et, Ribozyme A: +, ribozyme B.

M. Sioud et al.

for ribozyme A and B 72 hours post transfection was 57 o/0 and IS%, respectively. The stability of the antisense RNA control (ribozyme A lacking the catalytic domain) is similar to ribozyme A (data not shown). Thus, the addition of a bacteriophage T7 terminator to the 3’ end of a ribozyme increases its stability. The compartmentalisation of ribozyme A in HL60 cells was also studied by analysis of cytoplasmic and nuclear RNAs. As shown in Figure 2(b) (lanes N and C), ribozyme A is preferentially located in the nucleus. Destruction

of endogenous TNFa ribozymes

RNA

in vivo by

We next asked whether ribozyme A could eliminate its target following RNA transfection. A preliminary time-course study of TNFa RNA synthesis in HL60 cells indicated that TNFa RNA could be detected after two hours of stimulation by PMA and ConA, reaching maximal expression after four to six hours. Cells were transfected with ribozymes, stimulated with PMA and ConA for six hours, and total RNA was extracted and analyzed by Northern blotting using a TNFa probe. Since the TNFa and actin mRNA have approximately the same electrophoretic mobility, the same blot was hybridized with the actin probe after stripping. Data derived from densitometric scans of underexposed film indicated that the TNFa signal was reduced by 40% (Fig. 3(a), antisense lane) and 90% (Fig. 3(a), ribozyme A lane). In addition, we used a radioimmunoassay to measure TNFa protein in the culture medium. This system has major advantages over bio-assays in that it is specific for TNFa. The data indicate that HL60 stimulated with PMA and ConA secrete as much as 1000 fmol of TNF-a per ml but only 150 and 400 when cells are transfeeted with ribozyme A and antisense, respectively (Fig. 3(c)). Since the effect of antisense RNA is less than that of ribozyme A and since both ribozyme A and the antisense RNA were present inside the cells at similar concentrations (data not shown), we suggest that part of the activity of ribozyme A is due to its ability to cleave RNA. Ribozyme-mediated RNA cleavage in vivo has been observed in eukaryotic cells (Saxena & Ackerman, 1990). In this case no cleavage products were seen. As suggested previously (Cotten & Birnstiel, 1989; Sarver et al., 1990; Sioud & Drlica, 1991) the products of ribozyme-mediated cleavage are probably degraded by cellular nucleases (the 5’ fragment, although capped at its 5’ end, is expected to be rapidly degraded by 3’+5’ exonucleases). We used the same conditions to deliver ribozyme A to PBMNC. As can be seen from Figure 3(b) and (c), ribozyme A reduced by 80% the amount of TNFa RNA and by 70% the protein, raising the possibility that liposome-mediated RNA transfection offers a way to deliver ribozymes to a wide variety of cell types. During this analysis, we noted that ribozyme B appears to induce a similar reduc-

I

2

3

4

5

Figure 3. Ribozyme activity in. viva. Ribozymes and antisense RNA activities in HL60 cells (a) were analyzed after a transfection period (20 h). Following transfection with ribozyme A or antisense RNA, cells were stimulated for 6 h to express TNFa. RNA was extracted, separated by gel electrophoresis through a 1.2 % (w/v) agarose formaldehyde gel, and detected by Northern blotting with a radioactive probe for the TNFa gene. After hybridization with TNFa probe, the filter was stripped and then hybridized with an actin probe (British Biotechnology Limited). In the case of PBMNC (b), cells were separated (Sioud et al.. 1990) and washed 4 times with Hanks’ buffered saline solution and 3 times with serum free-medium. Cells (106) were transfected and processed as HL60 cells. Lanes 1 and 4, controls (transfected only with carrier RNA); lane 2; antisense RNA; lanes 3 and 5, ribozyme A. This autoradiogram was overexposed to display the TNFa signal in ribozyme A lanes. (c) Radioimmunoassay of TNFa protein. The amount, of TNFa protein present in the media was determined using the TNFa [1251] assay system (Amersham). Lanes 1 to 5 correspond to lanes 1 to 5 in Fig. 3(a) and (1,). respectively.

t.ion in TNF-a mRNA as ribozyme A (data not shown). Thus, t’he addition of T7 transcription terminator may decrease the specific activity of ribozyme A. This is the first report, to our knowledge, showing that, cationic liposome-mediated transfection can be used to deliver a functional ribozyme to human cells. Tn addition, our data indicate that ribozymes are more effect,ive inhibitors of gene expression than RNA having only antisense activity. We reached a similar conclusion with bacterial cells (Sioud & Drlica, 1991). Finally the experiments that we have described should be useful for evaluating modifications of ribozyme structure that are expected to improve catalytic activity. We have already learned that

a T7 transcription

terminator

confers stability

Communications on the ribozyme. We are now preparing to test other hairpin configurations to generate greater stability and activity. We are grateful to Dr Karl Drlica for a generous gift of the ribozyme and antisense primers. We thank Heidi Johnsen for technical assistance and Drs Karl Drlica, Morten Harboe, Sanjay Tyagi, Li Shan and Jens Kjeldsen-Kragh for critical reading of the manuscript and helpful discussion. This work was supported by a grant from the Norwegian Women’s Health Organisation, The Grethe Harbitz Legacy, The Leon and Norma Hess Foundation for Rheumatology Research and by Hafslund-Nycomed, Oslo, Norway.

References Beutler, B. & Cerami, A. (1988). Tumour necrosis, a common cachexia, shock, and inflammation: mediator. Annu. Rev. Biochem. 57, 505-518. Beutler, B. & Cerami, A. (1989). The biology of cachectin/ TNF. A primary mediator of the host response. Annu. Rev. Immunol. 7, 625-655. Brennan, F. M., Chantry, D., Jackson, A., Maini, R. N. & Feldmann, M. (1989). Inhibitory effect of TNFcr antibodies on synovial cell interleukin-1 production in rheumatoid arthritis. Lancet, ii, 244-247. Callis. J., Fromm, M. t Walbot, V. (1987). Expression of mRNA electroporated into plant and animal cells. Nucl. Acids Res. 15, 5823-5831. Cameron, F. & Jennings, P. (1989). Specific gene suppression by engineered ribozymes in monkey cells. Proc. Nat. Acad. Sci., U.S.A. 86, 9139-9143. Cech, T. (1987). The chemistry of self-splicing RNA and RNA enzymes. Science, 236, 1532-1539. Chomczynski, P. & Sacchi, N. (1987). Single step method of RNA isolation by acid guanidinium thiocyanate phenol chloroform extraction. Anal. Biochem. 162, 156-160. Cotten, M. & Birnstiel, M. L. (1989). Ribozyme mediated destruction of RNA in vivo. EMBO J. 8, 3861-3866. English, B. K., Weaver, W. M., & Wilson, B. C. (1991). Differential regulation of lymphotoxin and tumor necrosis factor genes in human T lymphocytes. J. Biol. Chem. 236, 7108-7113. Forster. A. C. & Symons, R. H. (1987). Self-cleavage of

835

plus and minus RNAs of a virusoid and a structural model for the active sites. Cell, 49, 211-220. Haseloff, J. & Gerlach, W. (1988). Simple RNA enzymes with new and highly specific endoribonuclease activities. Nature (London), 334, 585-591. Malone, W. R., Felgner, P. L. t Verma. I. M. (1989). Cationic liposome-mediated RNA transfection. Proc. Nat. Acad. Sci., U.S.A. 86, 6077-6081. Pennica, D., Nedwin, G. E., Hayflick, J. S.. Seeburg, P. H., Derynck, R., Palladino, M. A., Kohr, W. J., Aggarwal, B. B. & Goeddel, D. V. (1984). Human tumour necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature (London), 312, 724-729. Rosa, P., Weiss, U., Pepperkok, R., Ansorge, W.. Niehrs, C., Stelzer, E. & Huttner, W. (1989). An antibody against secretogranin I (chromogranin B) is packaged into secretory granules. J. Cell Biol. 109. 17-34. Rosenberg, A., Lade, B., Chui, D., Lin, S., Dunn, J. & Studier, W. (1987). Vectors for selective expression of cloned DNAs by T7 RNA polymerase. Gene, 56, 125-135. Z. F. & Fauci, A. Rosenberg, s. (1990). Immunopathogenic mechanisms of HIV infection: cytokine induction of HIV expression. Immunol. Today, 11, 176-180. Saxena, S. K. & Ackerman, E. J. (1990). Ribozymes correctly cleave a model substrate and endogenous RNA in vivo. J. Biol. Chem. 265. 17106-17109. Sarver, N., Cantin, E., Chang, P., Zaia, J.. Ladne, P., Stephens, D. & Rossi, J. (1990). Ribozymes as potential anti-HIV-l therapeutic agents. Science, 247, 1222-1225. Shinmei, M., Masuda, K., Kikuchi, T. & Shimomura, P. (1989). Interleukin 1, tumour necrosis factor, and interleukin 6 as mediators of cartilage destruction. Sem. Arth. Rheum. 18, (suppl. 1). 27-32. Sioud, M. & Drlica. K. (1991). Prevention of human immunodeficiency virus type 1 integrase expression in Escherichia coli by a ribozyme. Proc. -Vat. Acad. Sci., U.S.A. 88, 7303-7307. Sioud, M., Kjeldsen-Kragh, J., Quayle, A.. Kalvenes, C., Waalen, K., Fclrre, 0. & Natvig, J. B. (1990). The V6 gene usage by freshly isolated T lymphocytes from synovial fluids in rheumatoid synovitis: a preliminary report. &and. J. Immunol. 31, 415-421. Uhlenbeck. 0. (1987). A small catalytic oligoribonucleotide. Nature (London), 328. 596-600.

Edited by J. Karn