5′-,3′-Inverted Thymidine-modified Antisense Oligodeoxynucleotide Targeting Midkine

Oligodeoxynucleotides modified at both 5′- and 3′-ends with inverted thymidine (5′-,3′-inverted T) were introduced as new reagents for antisense strategies. These modifications were performed to make the oligodeoxynucleotides resistant to nucleases. The effectiveness of these oligodeoxynucleotides was evaluated in terms of inhibition of synthesis of midkine (MK), a heparin-binding growth factor, and consequent inhibition of growth of CMT-93 mouse rectal carcinoma cells. 5′-,3′-Inverted T antisense MK suppressed synthesis of MK by CMT-93 cells and their growth in culture. Furthermore, 5′-,3′-inverted T oligodeoxynucleotides exhibited less cytotoxicity and better stability than phosphorothioate oligodeoxynucleotides. When 5′-,3′-inverted T antisense MK was mixed with atelocollagen, and injected into CMT-93 tumors pregrown in nude mice, tumor growth was markedly suppressed as compared with tumors injected with sense controls. The suppressive effect of 5′-,3′-inverted T antisense MK on tumor growth was stronger than that of phosphorothioate antisense MK. These findings indicated the usefulness of inverted thymidine-modified antisense oligodeoxynucleotides as a new reagent instead of phosphorothioate-modified oligodeoxynucleotides.


Oligodeoxynucleotides modified at both 5-and 3ends with inverted thymidine (5-,3-inverted T) were introduced as new reagents for antisense strategies.
These modifications were performed to make the oligodeoxynucleotides resistant to nucleases. The effectiveness of these oligodeoxynucleotides was evaluated in terms of inhibition of synthesis of midkine (MK), a heparin-binding growth factor, and consequent inhibition of growth of CMT-93 mouse rectal carcinoma cells.

5-,3-Inverted T antisense MK suppressed synthesis of MK by CMT-93 cells and their growth in culture. Furthermore, 5-,3-inverted T oligodeoxynucleotides exhibited less cytotoxicity and better stability than phosphorothioate oligodeoxynucleotides. When 5-,3-inverted T antisense MK was mixed with atelocollagen, and injected into CMT-93 tumors pregrown in nude mice, tu-
mor growth was markedly suppressed as compared with tumors injected with sense controls. The suppressive effect of 5-,3-inverted T antisense MK on tumor growth was stronger than that of phosphorothioate antisense MK. These findings indicated the usefulness of inverted thymidine-modified antisense oligodeoxynucleotides as a new reagent instead of phosphorothioate-modified oligodeoxynucleotides.
Oligonucleotides have long been recognized to have huge potential as agents for turning off the expression of specific proteins, in most cases working by inducing degradation of the mRNA encoding the protein (1,2). The possible therapeutic use of oligonucleotides as effective gene regulatory agents in antisense and antigene approaches has kindled further interest in the development of oligonucleotide analogs (1,3). Rapid degradation of the "natural" phosphodiester (PO) 1 backbone oligonucleotides by nucleases (4,5) necessitated chemical modification of the PO backbone. Chemical modifications such as meth-ylphosphonate (6,7), phosphorothioate (PS) (8,9), and phosphoramidate (10) oligonucleotides have been introduced to make the oligonucleotides stable to degradative enzymes in serum (11). Among these chemical modifications, PS-modified oligonucleotides are most frequently used because of their ease of manufacture, low cost, and resistance to nucleases. However, PS-modified oligonucleotides have been shown to have toxic side effects in both cells in culture and in animals (12)(13)(14)(15).
Recently, several approaches have been employed to overcome this problem (16 -20). The present study was performed to develop a new modification for production of antisense oligodeoxynucleotides (ODNs), which remain capable of inducing degradation of the target mRNA, are stable in serum, and exhibit less cytotoxicity. For this purpose, we adopted antisense ODNs modified with inverted thymidine at both 5Ј-and 3Ј-ends, designated as 5Ј-,3Ј-inverted T antisense ODNs, according to recent reports (21,22) on DNA enzymes modified with inverted thymidine at the 3Ј-end. Inverted thymidine modification at 5Ј-or 3Ј-ends is aimed to protect the ODNs against exonuclease attack. The utility of the new modification was evaluated by monitoring anti-cancer activity of midkine (MK) antisense ODN produced by the proposed method.
MK, a heparin-binding growth factor, is a 13-kDa protein rich in basic amino acids and cysteine (23,24). MK is overexpressed in a variety of tumors, such as esophageal, gastric, colon, pancreatic, hepatocellular, lung, breast, and urinary bladder carcinomas (25)(26)(27)(28)(29)(30), neuroblastoma (28) and Wilms' tumor (25), and in normal adult tissue, its expression is usually low or undetectable (25,31). These findings suggested that MK may be a suitable target for cancer therapy. Recently, we established PS-modified antisense ODN targeting MK, which blocked the growth of mouse rectal carcinoma cells (CMT-93) in vitro and in vivo, and indicated its possible usefulness for cancer therapy (32). Thus, the effects of 5Ј-,3Ј-inverted T-modified ODNs and PS-modified ODNs were compared.
The underlined Ts in the above sequences indicate inverted thymidine. Uppercase sequences are PO, and lowercase sequences are PS. We also used ODNs labeled with fluorescein isothiocyanate (FITC) at each 5Ј-end, which were synthesized and purified in a manner identical to the unlabeled one.
Cell Culture Conditions and Transfection of ODNs-CMT-93 cells (American Type Culture Collection, Manassas, VA) derived from mouse rectal carcinoma were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% heat-inactivated fetal bovine serum (FBS) at 37°C in a humidified atmosphere of 5% CO 2 . Cells (3 ϫ 10 5 /35-mm tissue culture dish) were seeded and transfected for 3 h with ODNs in serumfree medium using LipofectAMINE PLUS (Invitrogen) as described previously (32). The transfection complex contained 5 nmol of ODNs and 6.8 nmol of lipids. Then, 1 ml of DMEM with 10% FBS was added, and incubation was continued for 4 h. The medium was then replaced with fresh DMEM containing insulin (10 mg/ml), transferrin (5.5 mg/ ml), sodium serenite (6.7 ng/ml), and heparin (20 g/ml). After 16 h of incubation, conditioned medium was collected for analysis.
Determination of the Size of Complex of ODNs and Liposomes-The transfection complex (32) which contained ODNs, the plus reagent, and the LipofectAMINE reagent was constructed using FITC-labeled ODNs, and the complex was observed by confocal microscopy system (MRC 1024, Bio-Rad), and the diameter of the FITC-labeled particle was determined with the aid of image analyzer (Bio-Rad).
Determination of the Amount of ODNs Complexed with Liposomes-ODNs (5 ϫ 10 Ϫ11 mol) were labeled at the 5Ј-end by T4 polynucleotide kinase (Takara Shuzo Co., Ltd.) and [␥-32 P]ATP (6000 Ci/mmol, Amersham Biosciences) in 50 l of reaction mixture as described previously (34). The reaction mixture was placed in the upper chamber of Microcon YM3 (cut off value 3,000 Da, Millipore Corp.). After adding 200 l of MilliQ water, it was centrifuged (4,000 ϫ g, at room temperature), and the procedure was repeated 4 times. Material remaining in the upper chamber was regarded as 32 P-labeled ODNs.
Five nmol of unlabeled ODNs was mixed with 25 pmol of 32 P-labeled ODNs and complexed with LipofectAMINE PLUS, which contained 6.8 nmol of lipids as described before (32). It was then placed in Slide-A-Lyzer dialysis cassette (cut-off value 10,000 Da, Pierce) and dialyzed against 2 liters of MilliQ water for 2 h. MilliQ water was changed twice, and radioactivity remaining in the cassette was determined. Proliferation Analysis-ODN-transfected CMT-93 cells were plated in 24-well plates in DMEM with 10% FBS at a density of 3 ϫ 10 4 cells per well. Three hours later, the medium was changed to serum-free medium, and the cells were cultured for up to 5 days. Cell proliferation was monitored using a cell counting kit (Dojin, Kumamoto, Japan).
Cytotoxicity Analysis-Cells (2 ϫ 10 5 /well in a 24-well plate) were seeded in DMEM with 10% FBS and cultured overnight. Then, 5Ј-,3Јinverted T SEN or PS SEN was added to cultures. The final concentrations of ODNs were 0, 50, 100, and 200 M as indicated in Fig. 3. Twenty four and 48 h later, cell survival was monitored using a cell counting kit as described above.
Uptake and Intracellular Localization of the Inverted T-modified ODN-Cells were seeded in Lab-Tek Chamber Slides (Nalge Nunc International, Naperville, IL) coated with Permanox and transfected with FITC-labeled various ODNs. After 3 h of incubation, the cells were washed with PBS, fixed with 4% paraformaldehyde at room temperature, and mounted using a ProLong Antifade Kit (Molecular Probes). Cells were photographed using a confocal microscope (Bio-Rad MRC 1024).
Western Blotting Analysis-Proteins in conditioned media were separated by electrophoresis on 15% SDS-PAGE gels and transferred electrophoretically onto nitrocellulose membranes (35). Blots were blocked with 5% nonfat dried milk and incubated with rabbit anti-MK antibody (35) and horseradish peroxidase-conjugated goat anti-rabbit IgG. Protein bands were visualized using an enhanced chemiluminescence detection kit (Amersham Bioscience). Quantitative analysis of the blots was performed with an imaging densitometer.
Tumor Therapy-A total of 1.5 ϫ 10 6 untransfected CMT-93 cells were subcutaneously inoculated in 0.3 ml of serum-free medium through a 24-gauge needle into both lower flanks of 8-week-old athymic nude mice obtained from SLC (Tokyo, Japan). After 7-9 days when tumors reached an average volume of ϳ60 mm 3 , the tumor-bearing nude mice were randomly divided into five different treatment groups (5Ј-,3Ј-inverted T AS, 5Ј-,3Ј-inverted T SEN, 3Ј-inverted T AS, PS AS, and PS SEN). All ODNs were dissolved in DMEM and mixed with an 5Ј-,3Ј-Inverted Thymidine-modified Antisense ODN equal volume of atelocollagen (final concentration of atelocollagen, 1.75%) kept at 4°C and injected directly into the tumor region as reported previously (32). The final concentration of PS SEN and AS was 50 M and that of 5Ј-,3Ј-inverted T AS and SEN and 3Ј-inverted T AS was 200 M. Each therapeutic reagent (50 l) was injected into the tumors every 2 weeks after the first injection. Animal experiments were performed in compliance with the guidelines of the Institute for Laboratory Animal Research, Nagoya University School of Medicine. Tumor diameters were measured at regular intervals with digital calipers, and tumor volume in mm 3 was calculated by the formula: volume ϭ (width) 2 ϫ length/2 (36). Data are presented as means Ϯ S.E.
Animal Experiment with Cationic Liposome as a Delivery Reagent-Cationic liposome (LipofectAMINE PLUS) as a delivery reagent was examined in tumor therapy experiment. One hundred l of 5Ј-,3Јinverted T AS stock solution (1 mM) and the plus reagent (50 l) were mixed in DMEM (100 l). After immediate mixing with a Vortex mixer and standing at room temperature for 15 min, the LipofectAMINE reagent (25 l) in DMEM (225 l) was added, and the mixture was left at room temperature for 15 min. Fifty l of the mixture was injected into the preformed tumor. On day 14, another injection was done, and tumor diameters were monitored as described above.
Stability of 5Ј-,3Ј-Inverted T-modified AS in FBS-Each ODN was incubated in 5% FBS at 37°C. Aliquots of the reaction were removed at different time intervals for electrophoretic analysis. Nuclease reactions were stopped by adding formamide gel loading buffer to each sample and heating at 95°C for 10 min (33). All samples were then run on 20% polyacrylamide gels containing 7 M urea (Bio-Rad) and visualized by staining with a SYBR Gold Nucleic Acid Gel Stain kit according to the manufacturer's instructions (Molecular Probes). All reagents for PAGE were nuclease-free grade.
To test the stability of ODNs complexed with liposome, each ODN was at first mixed with cationic liposome reagent (LipofectAMINE PLUS). Thus, 2.5 l of each ODN stock solution (1 mM) and the plus reagent (5 l) were mixed in MilliQ water (52.5 l) in a small sterile tube. After immediate mixing with a Vortex mixer and standing at room temperature for 15 min, the LipofectAMINE reagent (2 l) in MilliQ water (50 l) was added, and mixture was left at room temperature for 15 min. 5.9 l of FBS was added (final 5%) to the transfection complex (112 l), and the mixture was incubated at 37°C.
For calculation of the half-life of each ODN in serum, all bands were analyzed with an imaging densitometer, GelDoc 1000 system (Bio-Rad). 75% atelocollagen were subcutaneously injected into both lower flanks of 8-week-old male athymic nude mice. One and 7 days after injection, blood was taken from the infraorbital vein and after allowing to clot for 1 day (4°C), and sera were analyzed using HITACHI Clinical Analyzer 7170 by SRL, Inc. (Nagoya, Japan). Seven days after injection, mice were killed, and gross anatomy of organs was examined.
Statistical Analysis-The data were analyzed using the Mann-Whitney U test, and probability values less than 0.05 were considered to indicate significant differences.

Effects of 5Ј-,3Ј-Inverted T AS on CMT-93 Carcinoma Cells in
Culture-We examined whether 5Ј-,3Ј-inverted T AS can decrease MK secretion in CMT-93 rectal carcinoma cells. 5Ј-,3Ј-Inverted T AS transfected with the aid of a cationic liposome reagent (LipofectAMINE PLUS) suppressed synthesis and secretion of MK (Fig. 2A, lane 1), whereas 5Ј-,3Ј-inverted T SEN and 5Ј-,3Ј-inverted T REV showed no effects ( Fig. 2A, lanes 2  and 3). Densitometric analysis of the blots revealed that 5 M 5Ј-,3Ј-inverted T AS decreased MK production to 9% of that in controls ( Fig. 2A, lane 8). We also observed that 3Ј-inverted T AS and PS AS exhibited almost identical effects after transfection ( Fig. 2A, lanes 4 and 5). These inhibitory effects on MK production by 5Ј-,3Ј-inverted T AS and 3Ј-inverted T AS were dose-dependent (Fig. 2B).
We also examined whether the nature of the complex formed between the cationic lipid delivery agent and ODNs were dif- To measure the amount of ODNs bound to liposomes, we relied on the observation that ODNs complexed with liposomes stay within the dialysis tube after dialysis, whereas free ODNs are dialyzed out (34). Indeed, without LipofectAMINE reagents, only 2% of 32 P-labeled ODNs stayed in the dialysis apparatus after the procedure described under "Experimental Procedures." However, around 70 -80% of 32 P-labeled ODNs were in the dialysis apparatus in the presence of the Lipo-fectAMINE reagents. Actual values were as follows: 82.2% for 5Ј-,3Ј-inverted T AS, 73.0% for 3Ј-inverted T AS, and 72.7% for PS AS. Therefore, the efficiency of the complex formation was not significantly different between different ODNs. In the transfection complex the amount of ODN was 5 nmol, of which about 75% formed complex with liposomes, and the amount of lipids was 6.8 nmol. Thus, the ratio of ODNs to lipids in the liposome was about 1:1.8 Cytotoxicity of 5Ј-,3Ј-Inverted T ODNs and PS ODNs-When a high dose of PS SEN was added to cultures of CMT-93 cells, it showed noticeable cytotoxic effects on survival of the cells in a dose-and time-dependent manner (Fig. 3B). On the other hand, 5Ј-,3Ј-inverted T SEN showed almost no cytotoxicity at least up to 48 h (Fig. 3A).
Cellular Uptake of ODNs-The cellular uptake and distribution of 5Ј-,3Ј-inverted T SEN and PS SEN were examined by confocal microscopy. As shown in Fig. 4 5Ј-,3Ј-Inverted Thymidine-modified Antisense ODN intensity in the cell, excluding the possibility that the relative difference of uptake affected the result. The presence or absence of FITC did not change their effects on MK production in the case of 5Ј-,3Ј-inverted T AS, 5Ј-,3Ј-inverted T SEN (Fig. 2C), and 3Ј-inverted T AS (data not shown).
Stability of 5Ј-,3Ј-Inverted T-modified AS in FBS-ODNs with the natural PO backbone are digested by nucleases in less than 5 min in vivo, making them unsuitable for therapeutic use (4,5). PS-modified ODNs are considerably more stable in vivo (6,37). Any modified ODN that would be useful as an antisense agent should show reasonable stability against nucleases as well as acceptable hybridization with the target mRNA. Thus, we studied the stability of ODNs in DMEM supplemented with 5% heat-inactivated FBS. Fig. 5 shows the stability of the ODNs under these conditions. 5Ј-,3Ј-Non-inverted T AS was quickly digested by nucleases in serum, consistent with the reported data on PO-ODN (4). At 24 h, only a faint band of intact ODN was present (Fig. 5D). Unexpectedly, PS-modified ODN was not stable in serum, as indicated in Fig. 5C. In contrast to the above observations, 5Ј-,3Ј-inverted T AS and 3Ј-inverted T AS were strongly resistant to nucleases in serum (Fig. 5, A and B). In the case of 5Ј-,3Ј-inverted T AS, even though intact 5Ј-,3Ј-inverted T AS disappeared by 24 h, the produced fragment, probably 3Ј-inverted T AS, remained intact for up to 96 h. The half-lives of each ODN determined by densitometric analyses were as follows: 5Ј-,3Ј-inverted T AS, 30 h; 3Ј-inverted T AS, 110 h; PS AS, 10 h; 5Ј-,3Ј-non-inverted T AS, 5 h.
We also examined the stability of the complex of ODNs and LipofectAMINE PLUS to serum nucleases (Fig. 5). The complex formation to liposome generally increased the stability of ODNs, and more than half 5Ј-,3Ј-inverted T AS and most 3Јinverted T AS remained as the original one even after 96 h. As compared with them, PS AS and non-inverted T AS were still much more unstable.
Growth Inhibitory Effects of ODNs on CMT-93 Cells-Transfection of 5Ј-,3Ј-inverted T AS into CMT-93 cells inhibited their growth especially 3-5 days after addition (Fig. 6). However, 5Ј-,3Ј-inverted T SEN showed no such effects (Fig. 6). We also observed a similar inhibitory effect of 3Ј-inverted T AS. 5Ј-,3Ј-Inverted Thymidine-modified Antisense ODN Treatment of Preformed Tumors by MK Antisense ODN-We then injected preformed tumors with 5Ј-,3Ј-inverted T AS. Untreated CMT-93 cells were inoculated into nude mice as described previously (32). Nine days after inoculation when a palpable tumor was formed, 5Ј-,3Ј-inverted T AS, 3Ј-inverted T AS, PS AS, 5Ј-,3Ј-inverted T SEN, or PS SEN, which were premixed with atelocollagen, was injected into the tumor, and the injection was repeated every 14 days. 5Ј-,3Ј-Inverted T AS markedly suppressed tumor growth as compared with 5Ј-,3Јinverted T SEN and PS SEN (p Ͻ 0.001) (Fig. 7A). 3Ј-Inverted T AS also significantly suppressed tumor growth as compared with 5Ј-,3Ј-inverted T SEN (p Ͻ 0.001). PS AS also suppressed tumor growth, as reported previously (Ref. 32, Fig. 7A). As tumor volume at the initial injection in the present experiment (Fig. 7A) was larger than that in the previous report, the effect of PS AS was slightly less as compared with that in the previous report (32). We found that the tumor-suppressive effect of 5Ј-,3Ј-inverted T AS was greater than that of 3Ј-inverted T AS and PS AS (p Ͻ 0.01).
In the above experiment, tumor growth was monitored by determining tumor volume estimated with digital calipers. At the end of the study (41 days after initiation of tumor therapy), all animals were sacrificed, and tumor weight was determined. 5Ј-,3Ј-Inverted T AS significantly suppressed the increase of tumor weight as compared with controls (Fig. 7B).
We used atelocollagen as the carrier of ODNs, because liposomes do not consistently give good delivery in vivo (38). Indeed, 5Ј-,3Ј-inverted T AS showed much weaker effects, when LipofectAMINE PLUS was used as a carrier (Fig. 8).
Injection of 5Ј-,3Ј-inverted T AS or PS AS in atelocollagen at the dose used for tumor therapy did not show apparent toxicity to nude mice. Thus, serum levels of creatinine, blood urea nitrogen, aspartate aminotransferase, and alanine aminotransferase were not different between control animals and those injected with the above ODNs in atelocollagen (data not shown). Gross anatomy of organs was also unaffected. DISCUSSION Antisense therapeutics using synthetic oligonucleotides are currently being evaluated in clinical trials for cancer, inflammation, and viral diseases (12). Most antisense ODNs under development for clinical application contain a PS backbone, whereas side effects and toxicities were reported to be induced by PS-modified ODNs (12). In primates, the primary acute side effects are associated with complement activation (13, 39 -41) and systemic effects due to accumulation of high concentrations of PS-modified ODNs in the kidneys (39,42). In rodents, the primary side effect is immune stimulation characterized by lymphoid hyperplasia and mononuclear cell infiltrates in multiple tissues (14). At extraordinarily high doses (15-50 times the planned clinical doses), hepatocellular (Ͼ50 mg/kg) and renal tubular degeneration (Ͼ100 mg/kg) are evident in rodents (14,15,43,44). In the liver, Kupffer cell hyperplasia and exacerbation of background foci of inflammatory cells were observed as typical effects induced by PS-modified ODNs (12), although the mechanism of the liver toxicity was not clear.
To overcome the toxic effects of PS-modified ODNs described above, 5Ј-,3Ј-inverted T antisense ODNs were devised as new reagents to suppress gene expression, and their effectiveness was confirmed by suppression of MK synthesis, leading to suppression of growth of CMT-93 rectal carcinoma in nude mice. 5Ј-,3Ј-Inverted T MK antisense ODN indeed exhibited less toxicity than the PS MK antisense ODN. Thus, we could administer larger amounts of 5Ј-,3Ј-inverted T MK antisense ODN and obtain better therapeutic effects than with the use of PS MK antisense ODN.
In the presence of serum, 20-mer 5Ј-,3Ј-inverted T AS was converted to a slightly shorter oligonucleotide, probably a 19mer, and remained stable. On the other hand, the 19-mer 3Ј-inverted T AS was stable in serum. These results suggested that 5Ј-inverted T of 5Ј-,3Ј-inverted T AS was susceptible to an exonuclease in the serum, whereas the 3Ј-inverted T was resistant to the nuclease. Consistent with these observations, 5Ј-,3Ј-inverted T AS and 3Ј-inverted T AS exhibited similar growth inhibitory activity to CMT-93 rectal carcinoma cells. However, 5Ј-,3Ј-inverted T AS exhibited stronger antitumor activity in nude mice than 3Ј-inverted T AS. Exonucleases in the tumor tissue probably degraded these AS in a manner different from that in serum.
Recently, various new strategies have been introduced to produce antisense DNA, e.g. replacement of the sugar-phosphate backbone to 2-aminoethyl glycine (16,45,46), replacement of PO linkages with amide linkages (18,19), and morpholino-oligonucleotides (20). Comparative evaluation of the 5Ј-,3Јinverted T AS and these new reagents should be performed in future studies.
Nucleotide-mediated therapeutics require effective delivery systems. Atelocollagen is produced by elimination of antigenic telopeptides attached to both ends of the collagen with pepsin (47). Thus, it is neither antigenic nor toxic in animals. In addition, atelocollagen is liquid at 4°C and a gel at 37°C. Satisfactory delivery of plasmid DNA (47) and PS-modified ODNs (32) 2 via atelocollagen was recently achieved. The present study shows that atelocollagen is also suitable for delivering another DNA structure, 5Ј-,3Ј-inverted T ODN, and supports the usefulness of atelocollagen in delivery of various DNA compounds.
The antisense strategy is a hopeful new approach to cure malignant tumors (48,49). Molecules with anti-apoptotic activity are frequent targets of such antisense therapy (48). As MK is also known to have anti-apoptotic activity (50,51) and is overexpressed in a number of carcinomas (25)(26)(27)(28)(29)(30), antisense MK ODNs in the form of 5Ј-,3Ј-inverted T or other forms are excellent candidates for testing for clinical effectiveness in cancer therapy.