A ''combination oligonucleotide'' antisense strategy to ... - Nature

3 downloads 63 Views 256KB Size Report
Cancer Centre, 790 Commissioners Rd., London, Ontario, Canada ... mines 2000, were from Invitrogen Canada (Burlington, ..... 1999;127:1777–1786. 8.
Cancer Gene Therapy (2003) 10, 278–286 All rights reserved 0929-1903/03 $25.00

r 2003 Nature Publishing Group

www.nature.com/cgt

A ‘‘combination oligonucleotide’’ antisense strategy to downregulate thymidylate synthase and decrease tumor cell growth and drug resistance Randal W Berg,1 Peter J Ferguson,1,2 Mark D Vincent,1,3 and D James Koropatnick1,2,3,4,5 1

Cancer Research Laboratories, London Regional Cancer Centre, London, Ontario, Canada; 2Departments of Pharmacology and Physiology; 3Oncology; 4Pathology; and 5Microbiology and Immunology, The University of Western Ontario, London, Ontario, Canada.

Thymidylate synthase (TS) catalyzes de novo production of thymidylate for DNA synthesis and cell proliferation. As such, TS has been a target of antitumor chemotherapy for many years. Our laboratory has identified several antisense oligodeoxynucleotides (ODNs) that downregulate TS mRNA and protein, inhibit cell proliferation, and sensitize cells to TS-directed chemotherapeutic drugs. Based on our observation that targeting distinct regions of the TS mRNA with a variety of antisense molecules resulted in differential effects on TS mRNA levels, it was hypothesized that use of multiple ODNs targeting distinct noncontiguous regions would result in synergistic or antagonistic interactions. In this study, we report that some combinations of TS antisense ODNs were more effective at reducing TS mRNA abundance and inhibiting cell proliferation than the individual ODNs used alone. However, in contrast to the effects on cell proliferation, the enhanced sensitivity to anti-TS chemotherapeutic drugs (i.e., raltitrexed and 5fluorodeoxyuridine) that is achieved by treatment with individual ODNs was not further augmented by combined ODN treatment. This suggests that ODNs targeting TS mRNA inhibit an alternative function of TS mRNA or protein, distinct from thymidylate production. The results are evidence that the novel use of multiple antisense ODNs that target different regions of the same mRNA represents a general strategy to improve antisense effectiveness. Cancer Gene Therapy (2003) 10, 278–286. doi:10.1038/sj.cgt.7700566 Keywords: thymidylate synthase; antisense oligodeoxynucleotide; novel therapy; cancer chemotherapy; combination treatment

hymidylate synthase (TS) is an essential enzyme in de novo production of thymidylate and, because of its T crucial role in DNA synthesis and cell proliferation, has 1

been an important target for cancer chemotherapy for many years.2,3 The TS inhibitors 5-fluorouracil (5-FU) and raltitrexed (ZD1694, Tomudexs) have become integral drugs in standard treatments for colorectal cancer.4 Although reasonably successful in clinical treatment of human tumors, use of these drugs is limited by their toxicity to nontumor cells and development of treatment resistance. These limitations drive the search for alternative treatments. Antisense oligodeoxynucleotides (ODNs) are synthetic nucleic acid molecules designed to hybridize by Watson– Crick base-pairing with specific mRNA sequences to inhibit translation, target mRNA for degradation by ribonuclease H, and reduce cellular levels of target proteins. Several antisense ODNs targeting a variety of molecules have antiproliferative effects against tumor cells in vitro and in vivo,5 and several have demonstrated Received March 15, 2002. Address correspondence and reprint requests to: Dr D James Koropatnick, Cancer Research Laboratories, The London Regional Cancer Centre, 790 Commissioners Rd., London, Ontario, Canada N6A 4L6. E-mail: [email protected]

antitumor activity and limited toxicity in Phase I clinical trials.6 Previous work from our laboratory has shown that treatment of HeLa cells in vitro with antisense ODN 83 targeting human TS effectively reduces TS mRNA and protein levels, and increases sensitivity to the TS-targeting chemotherapeutic drugs raltitrexed, 5-FU, and 5-fluorodeoxyuridine (5-FUdR).7 These results have been recapitulated using human HT-29 colon carcinoma cells, both in vitro and in immunocompromised mice (Berg et al, manuscript in preparation).8 It is anticipated that a therapy utilizing TS antisense ODNs in combination with anti-TS chemotherapeutic drugs will be clinically useful. To our knowledge, few studies have examined the use of combinations of antisense ODNs designed to reduce cellular levels of one or more specific mRNA targets. In some cases where multiple mRNAs were targeted, cooperative effects were not clearly demonstrated.9,10 On the other hand, combinations of anti-sense ODNs targeting several members of the epi-dermal growth factor family demonstrated more-than-additive inhibition of colony formation by colon carcinoma cells in vitro.11 In addition, a combination therapy using antisense ODNs targeting bcr-abl and c-myc was more effective than either treatment alone at delaying tumor growth and prolonging survival in a chronic myelogenous leukemia model in vitro and in vivo.12,13

Two antisense oligos are better than one RW Berg et al

In MCF-7 and HeLa cells, expression of antisense RNA molecules or treatment with antisense ODNs targeting the TS mRNA translation start site increases TS gene transcription.14 Thus, targeting different regions of a specific mRNA can result in upregulation or downregulation of expression. Targeting multiple noncontiguous sequences in the same mRNA is a novel antisense approach with potential to enhance efficacy. In this report, we examine the effects of treatment with combinations of antisense ODNs targeting TS mRNA, with and without TS-targeting chemotherapeutic drugs, on the proliferation of HeLa cells in vitro. The data indicate that use of combined multiple ODNs is more effective than use of individual ODNs in reducing target mRNA abundance to inhibit proliferation, and is equivalent with respect to sensitizing human tumor cells to TS targeting drugs.

Methods

Cell culture and chemicals HeLa cells were obtained from the American Type Culture Collection (Manassas, VA) and grown in Dulbecco’s modified Eagle medium containing 10% fetal bovine serum at 371C in a humidified atmosphere of 5% CO2. All tissue culture reagents, including LipofectAmines 2000, were from Invitrogen Canada (Burlington, ON, Canada). Raltitrexed (AstraZeneca, Macclesfield, UK) was dissolved in 0.1 M sodium bicarbonate, and 5FUdR (Sigma, St Louis, MO) was dissolved in water, and diluted in serum-free medium prior to use.

amounts of ODN and LipofectAmine 2000 were premixed for 15 min in serum-free medium. Complete medium was added to yield a 2  transfection mix, and 2 ml of this mixture was added to each flask. Cytotoxic drugs were added to the flasks 4 h after ODN treatment, in 100 ml volumes. In some experiments, a 1  transfection mix was prepared, and the culture medium on the cells was replaced with 2 ml of the 1  mixture. In this case, cytotoxic drugs were diluted in growth medium to 2  final concentration and added in 2 ml aliquots after 4 h; therefore, the concentration of ODN was reduced at that time. The ODN concentrations reported indicate the initial ODN concentration, representing the 4-h pretreatment values. Essentially similar results were obtained with each method. Cells were counted using an electronic particle counter (Beckman Coulter, Hialeah, FL) on the day of treatment, and 4 days later. Proliferation is expressed as a percentage of that in control flasks using the formula: 100  (experimental final cell number  initial cell number)/ (control final cell number – initial cell number). IC50 values were interpolated from plotted data. Previous studies have shown that, under these conditions, the control scrambled ODN 32 has no effect on TS mRNA or protein levels, or cell proliferation.7

RNA preparation and analysis

Fully phosphorothioated ODNs with 20 -methoxy–ethoxy modification on the six nucleotides at both the 50 - and 30 ends were generously provided by Dr N Dean (ISIS Pharmaceuticals, Carlsbad, CA). ODN 83 (50 GCCAGTGGCAACATCCTTAA-30 ), ODN 502 (50 CCAGCCCAACCCCTAAAGAC-30 ), and ODN 504 (50 -ACTCAGCTCCCTCAGATTTG-30 ) are complementary to nucleotides 1184–1203, 1081–1100, and 1436– 1455, respectively, in the 30 untranslated region of human TS, while ODN 494 (50 -AGCATTTGTGGATCCCTTGA-30 ), 495 (50 -GGCATCCCAGATTTTCACTC-30 ), and ODN 501 (50 -TTGGATGCGGATTGTACCCT-30 ), are complementary to nucleotides 380–399, 419–438, and 1002–1021, respectively, within the protein coding region. The scrambled control ODN 32 (50 -ATGCGCCAACGGTTCCTAAA-30 ) has the same base composition as ODN 83, in random order, and is not complementary to any region of human TS. The ODNs were diluted in Milli-Q purified water, and concentrations calculated based on spectrophotometric absorbance readings.

For isolation of RNA, cells were plated at 1  106 cells per 75-cm2 flask in 5 ml of medium. On the following day, ODNs (600 nM) were mixed with 6 mg/ml LipofectAmine 2000 in serum-free medium for 15 min at room temperature to yield a 6  transfection mix. ODN : lipid mixture (1 ml) was added to each 75-cm2 flask to yield final concentrations of 100 nM ODN and 1 mg/ml lipid. RNA was prepared from HeLa cells using the TRIzol reagent (Invitrogen, Canada), and quantified using a spectrophotometer. For reverse-transcription polymerase chain reaction (RT-PCR), cDNA was prepared from 1 mg of total RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen, Canada) and random hexamer primers. Of the cDNA produced, 2% was used as template for PCR with the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers GAP-for (50 TATTGGGCGCCTGGTCACCA-30 ) and GAP-rev (50 CCACCTTCTTGATGTCATCA-30 ), or the TS primers TS-for (50 -TTTTGGAGGAGTTGCTGTGG-30 ) and TSrev (50 -TGTGCATCTCCCAAAGTGTG-30 ). PCR cycling parameters were: 3 min at 941C; followed by 24 cycles of 30 s at 941C, 30 s at 581C, 45 s at 721C; and a 7 min 721C extension. Products were resolved on 1.5% agarose gels and stained with ethidium bromide. Quantitation of images captured using the ImageMASTER VDS gel documentation system (Amersham Pharmacia Biotech) was done with ImageQuant version 5.1 (Molecular Dynamics).

Antisense ODN and chemotherapeutic drug treatments

TS protein quantitation

Oligodeoxynucleotides

4

2

HeLa cells were plated at 7.5  10 cells per 25-cm flask in 2 ml of medium. On the following day, the required

HeLa cells express levels of TS protein that are not readily detectable by Western blot techniques using the available

Cancer Gene Therapy

279

Two antisense oligos are better than one RW Berg et al

280

antibodies (R Berg and P Ferguson, unpublished observations). Therefore, TS was quantitated using a 5FUdR monophosphate (5-FdUMP) binding assay,15 as previously described.7 This sensitive assay reliably detects TS at levels as low as 10 fmol/mg total protein.16,17

Flow cytometry Cells were treated with 200 nM ODNs and 2 mg/ml lipid, in the same manner as described for RNA preparation, and collected by trypsinization at 48 h after ODN treatment. Cells were centrifuged for 5 min at 500 g, resuspended in phosphate-buffered saline, centrifuged, fixed in 75% ethanol for 15 min at room temperature, centrifuged, and washed again. The cells were stained with propidium iodide (0.02 mg/ml in phosphate-buffered saline with 0.1% (vol/vol) Triton X-100 and 0.2 mg/ml deoxyribonuclease-free ribonuclease A) and analyzed on a Beckman Coulter XL-MCL flow cytometer. At least 50,000 single cells were analyzed for each condition, and the distribution of cells in G0/G1, S, and G2/M cell cycle phases was calculated using MultiCycle software (Version 3.0, Phoenix Flow Systems, San Diego, CA).

Statistical analysis Statistical differences were determined using Student’s ttest. A Po.05 was considered significant. All experiments were performed at least three times.

Results

ODN combinations enhance antisense downregulation of TS To test the hypothesis that combinations of antisense ODNs might be more effective than single ODNs, several pairs of antisense ODNs targeting human TS mRNA were analyzed for their effects on HeLa cells in vitro. TS antisense ODN 83, used alone, effectively inhibits HeLa cell proliferation compared to the scrambled control ODN 32 (Table 1).7 In initial experiments, ODN 83 (50 nM) was combined with equimolar amounts of one of a panel of antisense ODNs targeting different regions of TS mRNA, and the results compared to those obtained when ODN 83 was combined with control ODN 32 (an ODN that does not target TS mRNA or any other known human sequences). Certain partner ODNs enhanced the antiproliferative response while others did not (Table 1); subsequent experiments focused on treatment with ODN 83 in combination with ODN 501 or 504. HeLa cells were treated with these pairs of ODNs, and RT-PCR analysis performed to measure reductions in TS mRNA level. Treatment with combinations of TS antisense ODNs 83+501 and ODNs 83+504 resulted in enhanced reductions of TS mRNA levels, compared to treatment with each ODN used as a single agent (Fig 1a and b). Decreased TS message levels were reflected in reduced TS protein and activity levels, measured by [63H]5-FdUMP binding, as early 24 h after treatment with ODNs 83+501 (Fig 1c). By 48 h post-treatment, the TS

Cancer Gene Therapy

Table 1 Enhancement of the antiproliferative activity of ODN 83 by combination treatment with additional TS antisense ODNs

ODN mixa 32 alone 83+32 83+494 83+495 83+501 83+502 83+504

Cell numberb (SEM) 9.65 4.94 2.82 3.77 1.67 5.06 4.26

(0.34) (0.91) (0.51) (0.58) (0.28) (0.55) (0.89)

Inhibition of cell proliferationc (relative to ODN 32-treated)

Inhibition in combination with ODN 83 (% increased)

55 80 69 94 54 63

45 25 69 3 14

a ODN 32 alone was used at 100 nM; otherwise, mixtures were 50 nM of each ODN. b Number of cells (  105) per 25-cm2 flask after 4 days of ODN treatment (average of three flasks in a representative experiment). Starting cell number (day 0) was 91,300. c Percent inhibition, calculated using the formula: 100100  (experimental final cell number – initial cell number)/(control final cell number – initial cell number) d Calculated as a percentage of the inhibition achieved with ODN 83+32, using the formula: 100  (combination antisense ODN – ODN 83+32)/(ODN 83+32).

level in cells treated with the ODN combinations was equivalent to those treated with the individual ODNs but still significantly lower than in cells treated with the control scrambled ODN 32. Thus, TS antisense ODN combinations downregulate both TS mRNA and protein, and do so more effectively than individual ODNs at an early time following treatment.

Antisense ODN combinations exhibit enhanced antiproliferative activity We previously showed that inhibition of HeLa cell proliferation by ODN 83 treatment was accompanied by transient G2/M cell cycle arrest.8 To assess the ability of ODN combinations to induce cell cycle arrest, a flow cytometric analysis was used. Figure 2 shows that a 48-h treatment with ODN 83 alone or with the combination of ODNs 83+501 induced G2/M arrest to a similar extent, but that ODN 501 alone had no significant effect on the cell cycle profile. ODN 501 is representative of several other TS antisense ODNs that, when used as single-agent treatments, had little or no dose-dependent effects on cell cycle or cell proliferation, but effectively enhanced the cytotoxicity of raltitrexed and 5-FUdR (P Ferguson and R Berg, unpublished observations). The ODN combinations were then examined for the ability to inhibit HeLa cell proliferation over a range of ODN concentrations. The combinations of ODNs 83+501 and ODNs 83+504 enhanced inhibition of HeLa cell proliferation, compared to the effect of exposure to each of the ODNs used individually (Fig 3a and b). The combination of ODNs 501+504 was equivalent to ODN 504 alone, and the triple mixture was no more effective than either double mixture (data not shown).

Two antisense oligos are better than one RW Berg et al

281

a

Time of Treatment: 24 hours 48 hours Total ODN Concentration: 25 50 25 50 25 50 25 50 50 50 25 50 25 50 25 50 25 50 50 50 ODN 32 + + + + ODN 83 + + + + + + + + ODN 501 + + + + + + + ODN 504 + + + + + GAPDH TS Day 1

Day 2

b 25 nM

32

ODN treatment

50 nM 83 501 504 83+501

* *

83+504 0

0.2

* *

0.4

0.6

0.8

1

1.2

0

0.2

0.4

0.6

0.8

1

1.2

TS mRNA:GAPDH mRNA ratio Day 1

c

Day 2

25 nM

32

ODN treatment

50 nM 83 501 504 83+501

*

83+504

† 0

0.2

0.4

0.6

0.8

1

0

0.2

0.4

0.6

0.8

1

TS protein (pmol 5-FdUMP binding/mg protein) Figure 1 Quantitation of TS and GAPDH mRNA levels and TS protein levels in HeLa cells treated with ODN combinations. HeLa cells were treated with the indicated ODNs individually (25 or 50 nM) or in equimolar combinations (25 nM each ODN) for 24 or 48 h. RNA was extracted, and RT-PCR was performed and quantitated as described in Methods. (a) Ethidium-bromide-stained gel shows GAPDH and TS PCR products from a representative experiment. (b) Quantitation of gel shown in (a) and others showing reductions in TS mRNA following various ODN treatments. (c) HeLa cells were treated for 24 or 48 h with 25 or 50 nM of individual ODNs or with ODN combinations (25 nM each ODN), and TS protein levels were quantitated by 5-FdUMP binding as described in Methods. Asterisks indicate significant differences (Po.05) compared to cells treated with individual ODNs (50 nM). A dagger indicates a significant difference only compared to cells treated with ODN 504 alone.

Cancer Gene Therapy

Two antisense oligos are better than one RW Berg et al

c

ODN 32

b ODN 83

G0/G1 57.6

G0/G1 32.9

S 21.2 G2/M 21.2

S 31.6 G2/M 35.5

ODN 501

d

ODN 83+501

G0/G1 62.5

G0/G1 38.2

S 18.6 G2/M 18.9

S 23.1 G2/M 38.7

Figure 2 Flow cytometric analysis of cell cycle profiles in HeLa cells treated with ODN combinations. HeLa cells were treated with 200 nM of (a) ODN 32, (b) ODN 83 or (c) ODN 501; or (d) with 100 nM each of ODNs 83 plus 501 for 48 h, and collected for analysis by flow cytometry as described in Methods. Insets show the fraction of cells in G0/G1, S, and G2/M, determined using MultiCycle software.

Furthermore, varying the ratio of ODN83 : partner ODN from 1 : 7 to 7 : 1 did not further augment the enhanced activity (data not shown).

Antisense ODN combinations chemosensitize HeLa cells to anti-TS drugs The capacity of combinations of TS antisense ODNs to inhibit proliferation more effectively than individual ODNs raised the possibility that ODN combinations would similarly be more effective in enhancing tumor cell sensitivity to TS-targeting drugs. Therefore, the ability of these combinations of ODNs to enhance the sensitivity of HeLa cells to the TS inhibitors raltitrexed and 5-FUdR was examined. Compared to treatment with the control scrambled ODN 32, treatment with the combinations of ODNs 83+501 and ODNs 83+504 increased the cytotoxicity of raltitrexed (Fig 4a) and 5-FUdR (Fig 4b). However, the degree of enhancement was not different from that achieved by treatment with ODN 83 alone. We hypothesized that the observed lack of differential chemosensitizing effects between individual and combination ODNs might be related to the ODN concentration at which testing was done. That is, if cells were treated with lower ODN concentrations, differences might become apparent. The combination ODN strategy was therefore assessed for efficacy in sensitizing HeLa cells to 5-FUdR cytotoxicity using lower concentrations of ODNs. Sensitization to 5-FUdR, achieved using an equimolar combination of ODNs 83+501 (12.5 nM each), was not significantly different from that achieved with either of the individual ODNs used alone at 25 nM (Fig 5a). Similarly, the drug sensitivity of cells treated with the combination of ODNs 83+501 (6.25 nM each)

Cancer Gene Therapy

a

120 100

ODN 501 ODN 83 ODN 83+501

80 †

60 40

*

20

*

0 0

20

40

* 60

80

100

120

Antisense ODN concentration (nM)

b

120

Proliferation (percent of ODN 32 control)

a

Proliferation (percent of ODN 32 control)

282

100

ODN 504 ODN 83 ODN 83+504

80 60

*

40 20

*

0 0

20

40

60

80

100

120

Antisense ODN concentration (nM) Figure 3 Combinations of ODNs inhibit proliferation of HeLa cells more effectively than individual ODNs. HeLa cells were treated with TS antisense ODNs 83, 501, and 504 individually or in a 1 : 1 ratio to yield the combined concentrations indicated, and counted 4 days later. Total ODN concentration in each treatment was adjusted to 100 nM using the control scrambled ODN 32. Shown are the mean proliferation (7SD) relative to cells treated with 100 nM ODN 32 for three flasks from representative experiments. Asterisks indicate significant decreases in proliferation (Po.05, Student’s t test) when compared to cells treated with either antisense ODN alone. A dagger indicates a significant decrease in proliferation only when compared to cells treated with ODN 83 alone.

was equivalent to cells treated with ODN 83 alone (12.5 nM) (Fig 5b). At still lower ODN concentrations (i.e., 1 and 2 nM), treatment with individual or combinations of TS antisense ODNs did not enhance sensitivity to 5-FUdR; the dose–response curves were indistinguishable from control ODN 32-treated cells (data not shown). In further control experiments to examine the dose– response following individual ODN treatment, HeLa cells were treated with 12.5–50 nM ODN 83, followed by various doses of raltitrexed or 5-FUdR. Treatment with ODN 83 at each of these concentrations effectively sensitized HeLa cells to the cytotoxicity of raltitrexed, and while the cytotoxicity of 5-FUdR was increased at 12.5 nM, maximal chemosensitivity was achieved with 25 nM ODN 83 (Fig 6). In contrast, only the 50 nM dose of ODN 83 significantly inhibited cell proliferation in the absence of chemotherapeutic drug in these experiments, by approximately 20% (data not shown).

Two antisense oligos are better than one RW Berg et al

283

a ODN treatment

ODN 83

ODN 501 12.5 nM

ODNs 83+501

25 nM 0

20

40

60

80

Relative 5-FUdR IC50 (percent of ODN 32)

Proliferation (percent of control without drug)

b

120 ODN 32 ODN 83 (6.25) ODN 83 (12.5) ODN 501 ODN 83+501

100 80 60 40 20 0 0

1

2

3

4

5

6

5-FUdR (nM)

Figure 4 Sensitivity of HeLa cells to raltitrexed and 5-FUdR following treatment with combinations of TS antisense ODNs. (a) HeLa cells were treated with 100 nM (total) of the indicated ODNs for 4 h, then the indicated concentration of raltitrexed was added, and the cells were counted 4 days later. Shown are the mean proliferation (7SD) relative to cells treated with each ODN mixture in the absence of raltitrexed for three flasks from a representative experiment. (b) HeLa cells were treated with the indicated total concentration of ODN for 4 h, then with varying concentrations of 5-FUdR. Shown are the average of 5-FUdR IC50 values from at least three experiments.

Figure 5 Lower concentrations of ODN combinations cannot sensitize HeLa cells to 5-FUdR, compared to ODN 83 alone. (a) HeLa cells were treated for 4 h with ODN 83 or ODN 501 at 12.5 or 25 nM, or in combination (12.5 nM each ODN), followed by 0–10 nM 5-FUdR. The IC50 of 5-FUdR for cells treated with antisense ODNs, relative to cells treated with ODN 32, is shown (average of three experiments 7 SD). (b) Cells were treated with 6.25 nM of ODN 83, 12.5 nM of ODN 83 or ODN 501, or 6.25 nM each of ODNs 83+501 for 4 h. Total ODN concentration was equalized at 50 nM using control ODN 32. The indicated concentration of 5-FUdR was added, and cells counted 4 days later. Average relative proliferation (7SD) from three flasks from a representative experiment is shown.

Discussion

Targeting different regions of TS mRNA with antisense ODNs has diverse physiological consequences, not always in accord with antisense-mediated reduction in TS mRNA. TS antisense ODN 83 (targeting the 30 -untranslated region) downregulates TS mRNA and protein, inhibits cell proliferation and sensitizes cells to drugs that inhibit TS enzyme activity.7,8 In contrast, antisense targeting the translation start site increases TS gene transcription.14 Furthermore, certain TS antisense ODNs that downregulate TS mRNA to a similar extent as ODN 83 increased drug sensitivity without affecting cell proliferation on their own (P Ferguson and R Berg, unpublished observations). Differences in the biological consequences following antisense targeting of different regions of the same mRNA led us to predict that combinations of TS antisense ODNs, each targeting distinct regions of TS mRNA, would inhibit or enhance one or more of their downstream effects. We report here

that treatment of HeLa cells with some combinations of antisense ODNs targeting human TS mRNA resulted in enhanced reductions in TS mRNA levels and enhanced inhibition of proliferation, compared with the ODNs used individually. This is the first report of improved antisense effects through use of multiple antisense ODNs against the same target mRNA. TS expression is regulated at multiple transcriptional and post-transcriptional levels. TS antisense ODNs 83, 501, and 504, used individually, reduced TS mRNA and protein levels to a similar extent. In response to 24 h treatment with combined ODNs 83+501, reduction in TS protein levels was greater than that achieved by treatment with either of the ODNs alone, an effect that was matched by decreased TS mRNA levels (Fig 1). Therefore, there was a more-than-additive antisense effect of this combination at both the mRNA and protein level. However, 48 h post-treatment, there was no enhanced downregulation of

Cancer Gene Therapy

Two antisense oligos are better than one RW Berg et al

284 Proliferation (percent of control without drug)

a

120 100 80

without ODN 83 12.5 nM ODN 83

60

25.0 nM ODN 83

40

50.0 nM ODN 83

20 0 0

0.5

1

1.5

2

2.5

b

120

Proliferation (percent of control without drug)

Raltitrexed concentration (nM)

100 80 60 40 20 0 0

0.5

1

1.5

2

2.5

5-FUdR concentration (nM) Figure 6 Enhanced sensitivity of HeLa cells to lower doses of raltitrexed and 5-FUdR following treatment with various concentrations of ODN 83. HeLa cells were treated with the indicated concentration of ODN 83, mixed with ODN 32 to total 50 nM, for 4 h. The indicated concentration of (a) raltitrexed or (b) 5-FUdR was added, and the cells were counted 4 days later. Shown is the mean proliferation (7SD) relative to cells treated with each concentration of ODN 83 in the absence of drug for three flasks from a representative experiment.

TS protein in spite of continued decrease in TS mRNA. The combination of ODNs 83+504 enhanced reduction of TS mRNA, similar to the combination of ODNs 83+501, but, unlike 83+501, the 83+504 combination did not enhance TS protein downregulation: the effect was only additive. This suggests that different combinations of antisense ODNs affected TS expression at different control points. Presumably mRNA stability and translation, and possibly protein stability, were affected by the 83+501 combination. On the other hand, the 83+504 combination enhanced only events regulating mRNA levels (for example, mRNA degradation) without enhancing effects on TS mRNA translation or TS protein stability. These differential effects of simultaneously targeting different TS mRNA regions suggest a role for TS mRNA in regulating expression at those different levels. The translation start site, through its interaction with TS protein, has been suggested to have an autoinhibitory role.18,19 The potential for other TS mRNA regions to regulate TS expression has only been revealed

Cancer Gene Therapy

by antisense targeting, and is currently being explored. Individual or combination ODN treatments, in all cases, enhanced sensitivity to TS-targeting drugs to the same degree, in accord with reduction in TS protein levels, regardless of the enhanced reduction in TS mRNA. This suggests, not unexpectedly, that TS protein (but not TS mRNA) is the essential factor mediating drug sensitivity. The correlation between enhanced reduction in TS mRNA levels mediated by combined ODNs, and enhanced antiproliferative effects, suggested that TS mRNA may mediate a function relevant to proliferation, irrespective of TS protein. However, while treatment of HeLa cells with individual ODN 83 or ODN 501 reduced TS mRNA and protein levels, and sensitized cells to 5-FU and raltitrexed to similar extents, ODN 501 treatment had no effect on cell cycle or proliferation, whereas ODN 83 induced G2/M arrest and inhibited cell proliferation. In fact, the TS mRNA region complementary to ODN 83 is the only one we have been able to target to induce G2/M arrest, although targeting several others inhibited cell proliferation (data not shown). Interestingly, ODN 83 targets a region of exceptionally low potential for secondary structure in TS mRNA (Fig 7), and this may be of significance with respect to potential physiological interactions of TS mRNA with other molecules. Our model of downstream events induced by ODNs includes TS mRNA degradation leading to reduced TS protein levels, and consequent increased sensitivity to TS-targeting drugs. We hypothesize that targeting with ODN 83 also disrupts a function of TS mRNA distinct from its direct role in the production of TS protein. This potential function might be exerted through interactions of the ODN 83 target sequence with other macromolecules. We are currently defining the molecular pathway mediating G2/M arrest following ODN 83 treatment. G2/M arrest may not be essential for ODN 83 to exert its antiproliferative effect. Combined treatment with ODNs 83+501 induced G2/M arrest similar to ODN 83 alone (Fig 2), but was more potent at inhibiting proliferation (Fig 3). Therefore, these two events were not necessarily associated. Caution should be used in interpreting these results: the method used to assess proliferation did not distinguish between effects on cell cycle and cell death. Although we have shown that ODN 83 treatment does not induce apoptosis in HeLa cells,8 treatment with ODNs 501, 504 or combinations of ODNs might induce cell death and contribute to reduced cell numbers. Alternative explanations of our observations exist. RNase H-mediated cleavage of alternate mRNA targets and aptameric effects of TS antisense ODNs on TS or other unidentified proteins are formally possible. Global analysis of gene expression changes following ODN administration using cDNA and oligonucleotide microarrays is one strategy to assess the former possibility. To address potential aptameric effects, we are identifying proteins that interact with these ODNs in cultured cells and cell lysates. Finally, to dissect cytostatic and cytotoxic effects, we are measuring clonogenic survival of cells treated with ODNs (alone and in combinations), with and

Two antisense oligos are better than one RW Berg et al

285

Figure 7 Predicted single-stranded regions in TS mRNA. Predicted foldings of TS mRNA were obtained using mfold version 3.1 (http:// www.bioinfo.rpi.edu/applications/mfold/).20,21 The plot shows the number of foldings (out of 29 unique foldings obtained) in which each base is single-stranded, averaged over a 51-base window. In the upper diagram, thin horizontal lines represent the 50 - and 30 -untranslated regions, while the thick portion is the coding region flanked by the translation start site and stop codon (boxes). Vertical arrows indicate the location of sequences complementary to TS antisense ODNs, while vertical bars indicate the exon boundaries in the TS gene.

without raltitrexed and 5-FUdR. This is an important measure of potential clinical utility of combination therapy since the preferred effect of therapy is to eliminate malignant cells rather than reduce proliferation. Our previous results suggested that penetration of antisense ODNs into a solid tumor mass might be an impediment to successful therapy.8 Distribution of fluorescein-labelled phosphorothioate oligonucleotides into human A549 lung carcinoma explants in nude mice also indicated poor uptake into cells in the tumor interior.22 Combining ODNs to enhance antitumor effects may be effective in minimizing this limitation. Furthermore, our observation that low ODN concentrations (o10 nM) sensitize HeLa cells to drugs in vitro suggests that even moderate entry of ODNs into solid tumors may be sufficient to chemosensitize them. Therapies incorporating antisense ODNs to downregulate TS could enhance the effects of lower doses of cytotoxic drugs (e.g., raltitrexed or 5-FUdR) and reduce the danger of systemic toxicity. We are aware of only two reports of multiple antisense ODNs, each targeting different mRNAs, having morethan-additive effects. Antisense ODNs targeting cripto, amphiregulin, and transforming growth factor a mRNAs were used in combinations in vitro to inhibit growth of colon tumor cells.11 Simultaneous treatment with antisense ODNs targeting BCR-ABL and c-myc reduced tumor growth in in vitro and in vivo.12,13 The capacity of two adjacent antisense ODNs targeting HIV-1 gag mRNA to increase RNAse H activity has also been assessed.23 The present report is the first to describe the

effects of combined ODNs targeting the same mRNA on tumor cell proliferation and drug sensitivity. Antisense ODNs are vulnerable to the criticism that they have not fulfilled their potential as single agents in treatment of disease.6,24,25 Our rationale in targeting TS with antisense ODNs is that reduction in TS mRNA and protein has two effects: reduced proliferation as a consequence of decreased thymidylate production, analogous to the effect of treatment with conventional TStargeting chemotherapeutic drugs; and sensitization of tumor cells (including drug-resistant and TS-overexpressing cells) to existing and novel TS-targeting drugs. Combinations of TS antisense ODNs used in conjunction with TS-targeting drugs (e.g., ODNs 83+501 with raltitrexed) would be expected to have a greater overall antitumor effect than individual antisense ODNs. Decreases in tumor cell numbers would be achieved by both antiproliferative effects and enhancement of TS-targeting drug action. The results presented here suggest that use of multiple antisense ODNs, with or without traditional drugs in combination treatment, may improve the efficacy of antisense therapy. Acknowledgments

This work was supported by funds awarded to MV and JK by Zeneca Pharma Canada, Ltd, and Imperial Oil, Ltd. We thank Charlene Stirling, Helen Guenther, and Natalia Naraine for excellent technical assistance, and Mike Keeney (London Health Sciences Centre) for flow

Cancer Gene Therapy

Two antisense oligos are better than one RW Berg et al

286

cytometry analysis. We also thank Dr Nicholas Dean (ISIS Pharmaceuticals, Carlsbad, CA) for supplying ODNs, and Astra Zeneca, Inc. (Macclesfield, UK) for supplying raltitrexed (Tomudexs).

References 1. Carreras CW, Santi DV. The catalytic mechanism and structure of thymidylate synthase. Annu Rev Biochem. 1995;64:721–762. 2. Danenberg PV. Thymidylate synthetase—a target enzyme in cancer chemotherapy. Biochim Biophys Acta. 1977;473: 73–92. 3. Danenberg PV, Malli H, Swenson S. Thymidylate synthase inhibitors. Semin Oncol. 1999;26:621–631. 4. Papamichael D. The use of thymidylate synthase inhibitors in the treatment of advanced colorectal cancer: current status. Oncologist. 1999;4:478–487. 5. Smith JB, Wickstrom E. Preclinical antisense DNA therapy of cancer in mice. Methods Enzymol. 2000;314:537–580. 6. Gewirtz AM. Oligonucleotide therapeutics: a step forward. J Clin Oncol. 2000;18:1809–1811. 7. Ferguson PJ, Collins O, Dean NM, et al. Antisense downregulation of thymidylate synthase to suppress growth and enhance cytotoxicity of 5-FUdR, 5-FU and Tomudex in HeLa cells. Br J Pharmacol. 1999;127:1777–1786. 8. Berg RW, Werner M, Ferguson PJ, et al. Tumor growth inhibition in vivo and G2/M cell cycle arrest induced by antisense oligodeoxynucleotide targeting thymidylate synthase. J Pharmacol Exp Ther. 2001;298:477–484. 9. Sato N, Mizumoto K, Maehara N, et al. Enhancement of drug-induced apoptosis by antisense oligodeoxynucleotides targeted against Mdm2 and p21WAF1/CIP1. Anticancer Res. 2000;20:837–842. 10. Traidej M, Chen L, Yu D, Agrawal S, Chen J. The roles of E6-AP and MDM2 in p53 regulation in human papillomavirus-positive cervical cancer cells. Antisense Nucleic Acid Drug Dev. 2000;10:17–27. 11. Normanno N, Bianco C, Damiano V, et al. Growth inhibition of human colon carcinoma cells by combinations of anti-epidermal growth factor-related growth factor antisense oligonucleotides. Clin Cancer Res. 1996;2: 601–609. 12. Skorski T, Nieborowska-Skorska M, Campbell K, et al. Leukemia treatment in severe combined immunodeficiency mice by antisense oligodeoxynucleotides targeting cooperating oncogenes. J Exp Med. 1995;182:1645–1653.

Cancer Gene Therapy

13. Skorski T, Nieborowska-Skorska M, Wlodarski P, et al. Antisense oligodeoxynucleotide combination therapy of primary chronic myelogenous leukemia blast crisis in SCID mice. Blood. 1996;88:1005–1012. 14. DeMoor JM, Vincent MD, Collins OM, Koropatnick J. Antisense nucleic acids targeted to the thymidylate synthase (TS) mRNA translation start site stimulate TS gene transcription. Exp Cell Res. 1998;243:11–21. 15. Spears CP, Gustavsson BG. Methods for thymidylate synthase pharmacodynamics: serial biopsy, free and total TS, FdUMP and dUMP, and H4PteGlu and CH2H4PteGlu assays. Adv Exp Med Biol. 1988;244:97–106. 16. Hoganson DK, Williams AW, Berger SH. Isolation and characterization of a thymidylate synthase-deficient human colon tumor cell line. Biochem Pharmacol. 1999;58:1529– 1537. 17. Omura K, Morishita M, Kawakami K, et al. Quantification of thymidylate synthase gene expression in human gastrointestinal carcinoma tissues using competitive PCR. Hepatogastroenterology. 1999;46:985–990. 18. Chu E, Koeller DM, Casey JL, et al. Autoregulation of human thymidylate synthase messenger RNA translation by thymidylate synthase. Proc Natl Acad Sci USA. 1991;88:8977–8981. 19. Chu E, Voeller DM, Jones KL, et al. Identification of a thymidylate synthase ribonucleoprotein complex in human colon cancer cells. Mol Cell Biol. 1994;14:207–213. 20. Mathews DH, Sabina J, Zuker M, et al. Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure. J Mol Biol. 1999;288:911– 940. 21. Zuker M, Mathews DH, Turner DH. Algorithms and thermodynamics for RNA secondary structure prediction: a practical guide. In: Barciszewski J, Clark BFC, (eds). RNA Biochemistry and Biotechnology. New York, NY: Kluwer Academic Publishers;1999:pp. 11–43. 22. Ho SP, Scully MS, Krauthauser CM, et al. Antisense oligonucleotides selectively regulate aspartyl beta-hydroxylase and its truncated protein isoform in vitro but distribute poorly into A549 tumors in vivo. J Pharmacol Exp Ther. 2002;302:795–803. 23. Kandimalla ER, Manning A, Lathan C, Byrn RA, Agrawal S. Design, biochemical, biophysical and biological properties of cooperative antisense oligonucleotides. Nucl Acids Res. 1995;23:3578–3584. 24. Gewirtz AM, Stein CA, Glazer PM. Facilitating oligonucleotide delivery: helping antisense deliver on its promise. Proc Natl Acad Sci USA. 1996;93:3161–3163. 25. Lebedeva I, Stein C. Antisense oligonucleotides: promise and reality. Annu Rev Pharmacol Toxicol. 2001;41:403–419.