Antisense Therapy for Cancer

REVIEWS

ANTISENSE THERAPY FOR CANCER

Martin E. Gleave* and Brett P. Monia‡ Abstract | Improved understanding of the molecular mechanisms that mediate cancer progression and therapeutic resistance has identified many therapeutic gene targets that regulate apoptosis, proliferation and cell signalling. Antisense oligonucleotides offer one approach to target genes involved in cancer progression, especially those that are not amenable to small-molecule or antibody inhibition. Better chemical modifications of antisense oligonucleotides increase resistance to nuclease digestion, prolong tissue half-lives and improve scheduling. Indeed, recent clinical trials confirm the ability of this class of drugs to significantly suppress target-gene expression. The current status and future directions of several antisense drugs that have potential clinical use in cancer are reviewed.

1 PHOSPHOROTHIOATE Zamecnik and Stephenson ushered in the era of anti- nucleotide sequences of cancer-relevant genes offer the BACKBONE sense therapeutics when they reported that an oligo- possibility to rapidly design ASO or short interfering The non-bridging phosphoryl nucleotide complementary to the 3′ end of the Rous RNA (siRNA) duplexes for loss-of-function studies and oxygen of each nucleotide in an sarcoma virus could block viral replication in chicken preclinical proof of principle. Automated DNA synthe- oligomer is replaced with sulphur, which increases fibroblasts. An antisense oligonucleotide (ASO) is a sis and advances in the field of nucleic-acid chemistry resistance to nuclease digestion single-stranded, chemically modified DNA-like mol- have facilitated progress in the use of this technology and prolongs tissue half-life. ecule that is 17–22 nucleotides in length and designed for target validation and therapy. Moreover, improved to be complementary to a selected gene’s mRNA and chemical modifications of the PHOSPHOROTHIOATE thereby specifically inhibit expression of that gene BACKBONE increase resistance to nuclease digestion, (FIG. 1). Therefore, the level of the target protein is prolong tissue half-lives and improve scheduling. The reduced by blocking translation, and by altering the focus of this review is to address recent progress in the subsequent cascades regulating cellular proliferation, use of systemically delivered ASO drugs as targeted differentiation, homeostasis and apoptosis. These therapeutics in oncology, and will not cover all targets agents promise increased specificity for malignant in preclinical development or drugs in clinical trials cells and a more favourable side-effect profile owing without recent data that have been covered in earlier *The Prostate Centre at to well-defined modes of action. The specificity of the reviews2–4. Particular emphasis will be placed on inter- Vancouver General antisense approach is based on the fidelity of Watson– preting recent Phase III trial failures of several ASO Hospital, and Division of Urology, University of Crick hybridization and on estimates that a particular drugs, and highlighting recent advances that promise British Columbia D9, 2733 sequence of 17 bases in DNA occurs only once within to overcome the hurdles that confront nucleotide-based Heather Street, Vancouver, the human genome. therapeutics on their way to successful clinical applica- British Columbia, Canada, Over the last two decades, with the significant tion for oncology. Indeed, recent clinical trials confirm V5Z 355 increase in understanding of the biological basis for the ability of this class of drugs to significantly suppress ‡ISIS Pharmaceuticals, 1896 Rutherford Court, cancer progression, fuelled in part from the develop- target-gene expression in cancerous tissues. Carlsbad, California 92008, ment of high-throughput genomic, transcriptomic and USA. proteomic technologies, a plethora of therapeutic tar- Mechanism of antisense action Correspondence to M.E.G. gets have been identified. Many of these gene products Several mechanisms that explain how ASO inhibit e-mail: [email protected] are not easily amenable to inhibition by small molecules translation have been proposed (FIG. 1). The most doi:10.1038/nrc1631 or antibodies, and so strategies to inhibit these targets at accepted involves the formation of an mRNA–ASO Published online 20 May 2005 the gene-expression level are attractive. Indeed, known duplex formed through Watson–Crick binding, leading

REVIEWS

Summary and higher affinity translates into higher gene- repressing activity12,13. Unmodified phosphodiester • Antisense oligonucleotides (ASOs) offer one approach to target genes involved in oligonucleotides have no clinical use because they cancer progression, particularly those that are not amenable to small-molecule or distribute poorly in vivo and are rapidly degraded by antibody inhibition. endogenous nucleases, resulting in extremely short • ASOs inhibit translation through a mechanism that involves the formation of an in vivo half-lives. Several useful chemical modifica- mRNA–ASO duplex, leading to RNase-H-mediated cleavage of the target mRNA. tions of ASO backbones have yielded compounds • Several useful modifications of ASO backbones have yielded compounds that show that display good tissue distribution and increased good tissue distribution and increased resistance to nuclease digestion. resistance to nuclease digestion while retaining • ASO drugs are evolving through improved chemical modifications to prolong potent hybridization and RNase-H activation (BOX 2). in vivo half-life, increase potency and reduce toxicity. Like antibody-based therapeutics, which evolved to • The most promising targets for antisense therapy are those that become become a clinically useful therapeutic class through upregulated during tumorigenesis and several of these, including BCL2, protein years of optimization, ASO drugs are evolving kinase Cα, clusterin, X-linked inhibitors of apoptosis and survivin, are currently in through improved chemical modifications to prolong or have finished early-phase clinical trials. in vivo half-life, improve tissue distribution, increase • A disappointing lack of clinical efficacy for some ASOs indicates that challenges potency and reduce toxicity. Additional important remain. However, the advanced chemistry incorporated into the second-generation factors — such as improved understanding of the ASOs has significant promise for the future. relative importance of the therapeutic target, optimization of dose and scheduling, enrichment of trials • A recently completed prostate cancer pre-surgery trial provides proof of concept with tumours that are most sensitive to inhibition that the second-generation 2’-MOE OGX-011 can potently suppress the target protein clusterin in humans. of the relevant target, and rational use in combination regimens — will also increase the likelihood of success. to RNase-H-mediated cleavage of the target mRNA5–7. Other proposed mechanisms include prevention of Antisense targets in cancer mRNA transport, modulation or inhibition of splic- Elucidation of the pathogenic role of target genes asso- ing, translational arrest, and formation of a triple helix ciated with tumour progression continues to produce through ASO binding to duplex genomic DNA resulting a growing list of antisense gene candidates. The most in inhibition of transcription. Additionally, some ASOs, promising for antisense therapy are those targets that especially those containing a CpG motifs or strings of G become upregulated during and are causally related to might also possess immunostimulatory activity or other cancer progression and therapeutic resistance, and are OFFTARGET toxicities8. More recently, the use of double- not otherwise amenable to inhibition with antibodies stranded 21-nucleotide siRNAs that specifically suppress or small molecules. ASOs have been reported to spe- the expression of homologous genes has been described9. cifically inhibit expression of many different genes and Double-stranded siRNA also blocks gene expression delay tumour progression in many preclinical models. by a process involving Watson–Crick hybridization ASO drugs that target the APOPTOTIC RHEOSTAT, interfere to a targeted RNA, resulting in post-transcriptional with signalling pathways involved in cell proliferation gene silencing. These siRNAs are incorporated into a and growth, or target the tumour’s microvasculature, multiprotein RNA-induced silencing complex, leaving are particularly promising in combination with con- the antisense strand to guide this complex to its homol- ventional anticancer treatments. Several ASO drugs ogous mRNA targets for endonucleolytic cleavage10. in clinical development are discussed below and are Although siRNAs could offer superior potency and spe- listed in TABLE 1. cificity, few chemical modifications have been identified that support systemic in vivo activity11. Furthermore, the The BCL2 family. The BCL2 gene is the prototype of a toxicology and commercial cost of siRNA-based drugs class of oncogenes that contributes to neoplastic pro- remain unresolved issues. Therefore, whether or not gression by enhancing cell survival through inhibition the power of RNA interference can be harnessed for of apoptosis. Initially identified in follicular lymphoma the development of safe and gene-specific therapeutics due to the characteristic t14;18 translocation14, BCL2 remains to be determined, so this review will focus on is a mitochondrial-membrane protein that functions ASO anticancer drugs. to heterodimerize with BAX and other pro-apoptotic

OFF TARGET regulators, thereby inhibiting both the release of cyto- A non-specific effect of a drug Design and synthesis of ASOs chrome c from the mitochondria and the subsequent or antisense oligonucleotide The first step in identifying useful ASO sequences activation of the apoptotic cascade15. The selective that is different from its for target inhibition is to identify those sites on the and competitive dimerization between pairs of these characterized specific effect on its known target molecule. mRNA that are both unique to that particular gene antagonists and agonists of the BCL2 family of proteins and accessible to hybridization BOX 1. Next, an determines how a cell responds to an apoptotic signal. APOPTOTIC RHEOSTAT ASO must be synthesized in a form that distributes Many studies implicate overexpression of BCL2 with Describes the dynamic effectively and resists degradation in vivo, and yet is treatment resistance15–20, highlighting this gene as a interactions between pro- survival and pro-death signals easy to administer and is well tolerated in patients target to enhance chemotherapy-induced apoptosis. within a cell that regulate BOX 2. The affinity of ASOs to the target mRNA Several BCL2 ASO’s have reported good hormone programmed cell death. is a measure of stability of the nucleic-acid hybrid, or chemosensitization activity in various preclinical

REVIEWS

DNA Nucleus

Inhibits transcription 5′-cap formation Transcription a polyadenylation

Pre-mRNA Pre-mRNA

Inhibit 5′-cap Splicing RNase H formation modulation

Exon 1 Exon 2 Exon 3 AAAA

Splicing b Inhibition of mRNA Splicing polyadenylation

dsRNase

mRNA Exon 1 Exon 2 Exon 3 AAAA

Translation Nucleus

c Cytoplasm d Inhibit translation RNase H mRNA AAAA c Exon 1 Exon 2 Exon 3 d Antisense Ribosome dsRNase/RNAi Ribozymes Polypeptide Figure 1 | Mechanisms of antisense action on target genes. Several mechanisms of antisense action have been determined for antisense drugs (shown in green). These include inhibition of transcription (a), inhibition of splicing and inhibition of mRNA maturation by prevention of 5′-cap formation and of polyadenylation (b), and inhibition of ribosomal readthrough (c). Most importantly, antisense drugs hybridized with target RNA have been shown to serve as substrates for RNase H enzymes (shown in red; b,d) that cleave the RNA in DNA–RNA hybrids or RNA–RNA hybrids (as in RNAi) and decrease mRNA levels of the target gene. So, there is a growing repertoire of potential mechanisms of action that can be optimized through the addition of specific chemical modifications.

models20–23. G3139, also referred to as oblimersen continuous i.v infusion of G3139 in combination with sodium or Genasense (Genta), is a first generation full-dose dacarbazine (DTIC) reduced BCL2 protein 18-mer phosphorothioate ASO that is complimentary levels in serial melanoma biopsies, and this pharma- to the first six codons of the initiating sequence of the codynamic activity was associated with significant human BCL2 mRNA. G3139 moved into many clinical clinical responses26. Transient thrombocytopaenia at trials based on promising activity in preclinical models 12 mg/kg/day was dose-limiting in patients who also of many cancers21,23. received full-dose DTIC treatment. An international, In 1997, the first clinical study evaluating G3139 Phase III, randomized trial was recently completed TABLE 1 enrolled 21 patients with non-Hodgkin’s in patients with advanced melanoma using a 5-day lymphoma and treated them with a subcutaneous pretreatment regimen of 7 mg/kg/day G3139 admin- infusion of G3139 as a single agent24. Local inflamma- istered by continuous i.v. infusion followed by DTIC tion at the infusion site was the most common side at 1000 mg/m2. Recent analysis on an INTENTTOTREAT THROMBOCYTOPAENIA effect observed, while the maximum-tolerated dose (ITT) basis reported a median survival of 9.1 months Low circulating platelet count 2 (<125 ×109/L), which leads to a was determined to be 147.2 mg/m /day and the dose- in patients with G3139 plus dacarbazine, compared reduction in the clotting limiting toxicity was THROMBOCYTOPAENIA. One complete with 7.9 months for dacarbazine alone (P = 0.184). In efficiency of the blood. response and two minor responses were observed, but a PERPROTOCOL ANALYSIS of 480 patients that completed only half of the evaluable patients had measurable a minimum follow-up of 12 months, the addition of INTENTTOTREAT ANALYSIS decreases in BCL2 protein levels following treatment G3139 resulted in a median survival of 10.1 months, Patients are analysed according to the treatment group they with G3139, as measured by western blot analysis of compared with 8.1 months for dacarbazine alone 25 were randomized to, as opposed protein lysates from patient lymphocytes . (P = 0.035) (see Genta in the Online links box). In to the treatment actually Subsequent clinical trials of G3139 used continuous addition, for the ITT population (N = 771), median received. intravenous (i.v.) infusions necessitated by the short tis- progression-free survival was significantly longer sue half-life of first-generation phosphorothioate ASO, (78 versus 49 days, P <0.001) in patients treated with PERPROTOCOL ANALYSIS A priori planned analysis of most often in combination with a cytotoxic agent. In a G3139 plus dacarbazine compared with patients results. Phase I/II trial in patients with advanced melanoma, treated with dacarbazine alone. Response rates were

REVIEWS

Box 1 | Designing optimal antisense sequences in this trial for patients receiving G3139 plus dexamethasome versus dexamethasome alone. The median The AUG start site is often accessible for oligonucleotide hybridization and antisense time to progression was similar for both treatments oligonucleotides (ASOs) that target this codon have been used successfully for and no difference was observed in response rates or 5,127 various target genes, although other sites can prove even more effective . To toxicity between the groups. facilitate the process of selecting target sites for antisense action, a range of in vitro Several trials in hormone-refractory prostate can- techniques have been used, many of these are combinatorial approaches based on cer demonstrated that standard doses of docetaxel or annealing reactions with arrays of antisense species128 and/or monitoring the mitoxantrone could be delivered with G3139 without accessibility of target structures by RNase-H mapping129. In addition, computational apparent increased toxicity29–31. A recently reported attempts to predict the secondary structure and folding pattern of mRNAs have been described and target screening using computer programs, such as ‘mfold’ or Phase II trial in men with metastatic hormone-refrac- ‘RNAstructure’, could prove to be a valid strategy to select effective antisense tory prostate cancer combined G3139 (7 mg/kg/day 2 sequences130. While this approach ignores other parameters that affect antisense for 7 days) with docetaxel (75 mg/m i.v. on day 6), efficacy, such as the three-dimensional structure of RNA in vivo or the accessibility of repeated every 21 days until progression or toxic- the target site for RNase H, it might significantly reduce the number of ASOs needed ity. Partial responses were noted in 4 of 12 patients for screening. By comparing sequences of the most effective versus less effective or (33%) with measurable disease and a >50% reduction ineffective ASOs reported in the literature, the tetranucleotide motif TCCC was in prostate-specific antigen was measured in 15 of identified in 20 of the 42 most effective antisense sequences128. So, in addition to the 27 (56%) patients31. Based on these Phase II data, a techniques described above, the prediction of target sites based on this motif might randomized Phase III trial was scheduled to begin further assist in the design of antisense therapeutics. in 2005, but with the negative results from trials in melanoma and myeloma, and with the subsequent breakup of the Aventis/Genta partnership, this trial also higher in patients receiving G3139 (11.7% versus has been put on hold. Issues persist about the dos- 6.8%, P = 0.019). However, this drug failed to obtain ing and regimen of this first-generation ASO, and Food and Drug Administration approval because whether 6 days of 7 mg/kg/day treatment is enough overall survival of the ITT population was not signifi- to suppress the target sufficiently. cantly prolonged, and the improvements in time to

progression were not deemed clinically significant. Dual targeting of BCL2 and BCL-XL. BCL-XL is Results of a Phase III trial of G3139 in combina- another anti-apoptotic BCL2 family member. In

tion with fludarabine and cyclophosphamide (Flu/ tumours in which BCL2 and BCL-XL are coexpressed, Cy) in patients with advanced chronic lymphocytic it is difficult to predict which of the two proteins is leukaemia (CLL) were recently presented27. In this more crucial for survival, and some tumour cells trial, 120 patients were randomized to receive Flu/Cy have been reported to switch expression from BCL2

with or without G3139 (3 mg/kg/day continuous i.v. to BCL-XL (REF. 32). BCL-XL ASOs have been reported day 1–7). It is unclear how this dose of G3139 was to induce apoptosis in various tumour cells and determined, and it is significantly lower than doses sensitize tumour cells to chemotherapy33–36. While

used in subsequent trials. Patients in this trial had BCL-XL ASOs marginally increased chemosensitiv- failed on previous therapy and the primary end point ity and delayed androgen-independent progression

of this trial was complete response plus nodular par- of prostate cancer xenografts, combined BCL-XL plus tial response. Nineteen patients (16%) treated with BCL2 ASOs acted synergistically to further enhance G3139 plus Flu/Cy achieved the primary end point, chemotherapy beyond that of either agent alone37. compared with 8 patients (7%) treated with Flu/Cy These findings illustrate that combinatorial regi- alone (P = 0.039). However, the overall response rate mens that inhibit two or more specific gene targets was similar between the treatment groups when par- can produce additive effects and provide the basis for tial responders were included. G3139 was generally identifying additional anti-apoptotic genes that might well tolerated, with specific adverse events associated serve as targets. with G3139 treatment being nausea, fever, fatigue, Simultaneous downregulation of both BCL2 and 27 back pain and a cytokine-release reaction . BCL-XL protein expression by a single bispecific The results of a Phase III trial of G3139 in combina- ASO has been accomplished by taking advantage tion with high-dose dexamethasome in patients with of the similarity of specific regions of the BCL2

refractory multiple myeloma has also been recently and BCL-XL mRNAs. Inhibition of both BCL2 and 28 disclosed . In this trial, 224 patients were rand- BCL-XL expression with a novel bispecific ASO omized to receive standard therapy using high-dose with complete sequence identity to BCL2 and three

dexamethasome with or without 7 mg/kg/day G3139. mismatches to BCL-XL inhibits expression of both

The primary objective of the study was to evaluate BCL2 and BCL-XL mRNA in tumour cells and is a 2′OMETHOXYETHYL whether the addition of G3139 would significantly potent inducer of apoptosis in several tumour cell 38 MODIFICATIONS increase the time to progression, with monitoring of types . This bispecific ASO was also recently tested Modifications of the ribose secondary end points that included objective response as a second-generation 2′OMETHOXYTHEYL (2′-MOE) at the 2′-position with an and overall clinical benefit. Before entering the study, MODIFIED ASO and found to be a potent inhibitor of electronegative substitute patients in both groups had received extensive pre- BCL2, BCL-X and MCL-1, and induced apoptosis such as 2′-O-methyl or 2′-O- L methoxy-ethyl group vious treatment with corticosteroids (for example, in human prostate PC3 cells. Future development (illustrated in Box 2). dexamethasome). No clinical benefit was observed plans for this particular ASO are unknown.

REVIEWS

Protein kinase Cα. Protein kinase C (PKC) Box 2 | Chemical modifications to optimize in vivo pharmacokinetics belongs to a class of serine/threonine kinases that First-generation phosphorothioate modifications use a substitution of a non-bridging fine-tune numerous intracellular responses aris- phosphoryl oxygen of DNA with sulphur, increasing resistance to nuclease digestion and ing from G-protein-coupled receptors, receptors prolonging half-life (see figure, part a). In vitro, liposomes or viral vectors are required with tyrosine-kinase activity and non-receptor to efficiently transfect cells with phosphorothioate antisense oligonucleotides (ASOs), tyrosine kinases39. Increased PKC expression has but transfection does not seem necessary in vivo, as the inhibition of the expression of been implicated in oncogenesis, tumour pro- the target can be accomplished with ASOs alone77,131. Serum and tissue half-lives of gression and the multidrug-resistance pheno- phosphorothioate ASOs are less than 2 and 4 hours, respectively, and minimal amounts type40–42. PKCα inhibitors can block prolifera- 132,133 of full-length ASOs can be found in tissues after 48 hours . Because of this, for tion, affect the growth and survival of tumours, clinical trials continuous or frequent intravenous infusions have been used to administer promote apoptosis, and sensitize tumour cells to 25,54,133 conventional phosphorothioate ASOs . As a class, phosphorothioate ASOs are well chemotherapeutic agents43,44. A 20-mer first-gen- tolerated, mostly with non-sequence-specific toxicity attributable to the 133–135 eration phosphorothioate ASO targeting human phosphorothioate backbone and the polyanionic nature of these compounds . PKCα (ISIS 3521, also referred to as LY900003 Considerable effort has been made to generate second-generation ASOs that have and Affinitak; Isis Pharmaceuticals) was identi- improved stability and efficacy by chemically modifying the phosphodiester-linkage, the fied that potently inhibits PKCα expression in a heterocycle or the sugar (see figure, part b). So far, modifications made to the ribose, and in particular, the 2′-position, seem to be the most important in improving the wide range of human tumour types in vitro in an 45 pharmacokinetic characteristics of second-generation phosphorothioate ASOs136. isoform-specific manner . Systemic administration 2′-modification with an electronegative substitute such as 2′-O-methyl or 2′-O-methoxy- of ISIS 3521 to nude mice bearing subcutaneously α ethyl (MOE) group137, or the use of a 2′-O,4’-C-methylene bridge (locked nucleic acid; implanted human tumours suppresses PKC levels LNA)138 confer an RNA-like C3’-endo conformation to the oligonucleotide that greatly in tumour tissue and inhibits tumour growth against increases affinity. While the RNA-like conformation improves binding, it also abrogates numerous human tumour cell lines, including U-87 the oligonucleotide’s ability to activate RNase H, a crucial aspect of ASO activity. This glioblastoma, MDA-MB-231 breast carcinoma, MIA dilemma has been addressed by generating chimeric oligonucleotides with 2′-modified PACA-2 pancreatic carcinoma, and Calu-1 lung nucleotides placed only at the ends of the oligonucleotide, thereby leaving a central carcinoma46–50. Based on the biological evidence RNase-compatible DNA gap. These advanced chemistry ‘gapmers’ (for example, implicating PKCs in human tumorigenesis, and the OGX-011) possess favourable physicochemical, biochemical and pharmacokinetic preclinical activity of ISIS 3521, clinical studies were properties136, with increased RNA affinity10 and reduced non-specific immune initiated for its use as a cancer treatment. 133 stimulation . Most importantly, improved resistance against nuclease-mediated In Phase I clinical studies of ISIS 3521 REFS 51,52, metabolism, resulting in significantly improved tissue half-life in vivo, enables a longer responses were noted in two patients with low-grade 77,136 duration of action and a more patient-friendly intermittent dosing schedule . ASO lymphoma and one patient with ovarian cancer. Dose- with third generation chemical modifications such as peptide nucleic acids (c) promise limiting toxicities were moderate fatigue, and the recom- further improvements in the future. mended Phase II dose was set at 2 mg/kg/day. Toxicity

a Antisense b Antisense c Antisense at this dose appeared mild, with one of six patients 1st generation chemistry 2nd generation chemistries 3rd generation chemistries developing moderate thrombocytopaenia, and another • Nuclease resistant • Greater potency due to • Further improvement in potency? mild LEUKOPAENIA. A similar Phase I clinical study using • Acceptable tissue enhanced affinity for RNA • Improved tissue distribution? distribution • Increased stability • Further reductions in toxicity? ISIS 5132, a first-generation phosphorothioate ASO tar- • RNase H activity • Decreased toxicity • Oral bioavailability? geted to the mitogen-activated protein kinase signalling Phosphorothioate 2-O-methoxyethyl (MOE) Morpholina protein RAF1, increased doses to 5.0 mg/kg/day without oligodeoxynucleotide reaching a maximum tolerated dose53. Differences in O A O apparent tolerability of ISIS 3521 and ISIS 5132 are not B O B O clear, and might not even be real, as they might reflect HO O non-drug-related adverse events, or differences in OO N H3C O OCH patient population. Nevertheless, based in part on these 3 N P – O observations, a recommended Phase II dose for ISIS S P O H3C O O B O– 3521 and ISIS 5132 was set at 2 mg/kg/day and these drugs were compared in a National Cancer Institute of N Canada randomized Phase II study54. Scheduling for both ASOs was 2 mg/kg/day for 21 days by continuous 2-O-methyl Peptide nucleic acid (PNA) O i.v. followed by a 7-day rest period. Overall, treatment B NH O was well tolerated, and although some patients had B stable disease, no biochemical or objective responses N O OCH3 O were observed. The first Phase I/II combination therapy trial Locked nucleic acid (LNA) NH O B evaluated ISIS 3521 at 2 mg/kg/day by continuous i.v. O O N infusion on days 0–14, with carboplatin and paclitaxel B O 175 mg/m2 administered on day 4, in patients with stage IIIB or IV non-small-cell lung cancer (NSCLC)55. O O O Minimal toxicity was observed, which consisted mostly –S P O of NEUTROPAENIA and thrombocytopaenia, leading to O treatment delays in less than 15% of patients. In 48

REVIEWS

Table 1 | Antisense molecules in late-stage preclinical or clinical development Target Compound Company/ Phase of Tumour type investigator development BCL2 G3139 (Oblimersen, Genta I–III Melanoma, CLL, MM, Genesense) HRPC Clusterin OGX-011 OncoGeneX II HRPC, breast, NSCLC PKCα ISIS 3521 (Affinitak) Lilly/Isis III NSCLC Methyltransferase MG98 Methylgene II Metastatic renal cancer; head and neck cancer Ribonucleotide GTI-2501 Lorus II HRPC reductase Survivin LY2181308 Lilly/Isis I Solid tumours MYB LR3001 Genta I CML

XIAP AEG35156 Aegera I Solid tumours HSP27 OGX-427 OncoGeneX I HRPC STAT3 ISIS 345794 Isis Preclinical Xenografts (broad, MM) eIF4E LY2275796 Lilly/Isis Preclinical Xenografts (broad) MDM2 GEM240 Hybridon Preclinical Xenografts (broad) IGFBP2 and OGX-225 OncoGeneX Preclinical HRPC, breast, gliomas IGFBP5 MCL1 ISIS 20408 Isis Preclinical Xenografts (broad) Androgen receptor as750/15 Eder et al.166 Preclinical HRPC PKA GEM 231 Hybridon/NCI Preclinical Xenografts (broad) CLL, chronic lymphocytic leukaemia; CML, chronic myelogenous leukaemia; eIF4E, eukaryotic translation initiation factor 4E; HRPC, hormone-refractory prostate cancer; HSP27, heat-shock protein 27; IGFBP2,5, insulin growth factor binding proteins 2 and 5; MCL1, myeloid cell leukemia sequence 1; MM, multiple myeloma; NSCLC, non-small-cell lung cancer; PKA, protein kinase A; PKCα, protein kinase Cα; STAT3, signal transducer and activator of transcription 3; XIAP, X-linked inhibitor of apoptosis.

evaluable patients receiving a median 6 cycles, the who did not (P = 0.048). Negative results were also response rate was 48%, with 2% (1 patient) obtain- obtained in another Phase III NSCLC trial testing ing a complete response and 46% (22 patients) partial ISIS 3521 in combination with gemcitabine and cis- responses. The median time to progression and the platin (see ISIS Pharmaceuticals in Online links box). median survival were 6.3 months and 15.9 months, Therapy was well tolerated, but median survival was respectively. A second Phase I/II combination trial roughly 10 months in both groups. tested ISIS 3521 at 2 mg/kg/day by continuous i.v. (days Several factors could account for the apparent 0–14), with cisplatin at 80 mg/m2 (day 0) and gemcit- lack of clinical efficacy for ISIS 3521. First, inhibition abine at 1,000 mg/m2 (day 0 and 7) in 44 evaluable, of PKCα might not produce a large enough effect chemotherapy-naive patients with advanced NSCLC. on tumour growth in this population. Inhibition of Toxicity was moderate but included thrombocytopae- PKCα might only result in a cytostatic effect, which nia, neutropaenia, anaemia, fatigue, dehydration, sep- the study design would not detect. Second, because sis and neutropaenic fever56. In the updated analysis measuring target levels in tumours is difficult in this of the trial, the response rate was 37%, including one patient population, it is unknown whether the target complete remission and 11 partial remissions (see ISIS was expressed in most study subjects. Third, inhibit- Pharmaceuticals in Online links box). ing a single molecular target might be insufficient Based on these Phase II data, two large rand- to exert a clinically detectable effect beyond what

LEUKOPAENIA omized Phase III trials were initiated as first-line can be achieved with chemotherapy alone. Finally, Low circulating total white treatment in patients with NSCLC. The first enrolled the dose of ASO used might not have been the blood cell count, often 600 patients using ISIS 3521 in combination with optimal biological dose and was not potent enough associated with anticancer carboplatin and paclitaxel in patients with stage to inhibit PKCα expression sufficiently in tumour drugs, which can predispose the host to infections. IV NSCLC, and the results were disappointing. No under these treatment conditions. Use of a second- difference was observed in time to progression or generation ASO or higher doses might have produced NEUTROPAENIA overall survival between groups. There were, how- better activity. Low circulating neutrophil cell ever, indications of antitumour activity, as patients count (<2 x109/L), which is often associated with anticancer who completed the prescribed course of therapy Clusterin. The clusterin (CLU) gene on chromosome drugs and can predispose the (6 cycles) receiving ISIS 3521 had a median survival of 8 encodes a cytoprotective chaperone protein also host to infections. 17.4 months compared with 14.3 months in patients known as testosterone repressed prostate message 2

REVIEWS

ab because of the second-generation chemistry and long tissue half-life of OGX-011, more relaxed dosing schedules are possible while maintaining the biological efficacy of target inhibition. Two Phase I trials have evaluated OGX-011 given weekly by i.v. infusion as a single agent or in combination with docetaxel83. The single-agent study has a unique design in that patients with localized prostate cancer are treated with the OGX-011 before radical prostatectomy, allowing for dose-dependent correla- tions for CLU expression and tissue concentrations. Surrogate tissues for markers of biological effect Figure 2 | Immunostaining of OGX-011 antisense (CLU expression in peripheral-blood mononuclear oligonucleotide drug distribution in human prostate cells and serum clusterin) are also being assessed tissue. An antibody raised against the 2 -O-methoxytheyl ʹ and can be correlated with those effects found (2′-MOE) backbone of OGX-011 enabled the immunohistochemical staining (brown) of resected human in target tissue. Therefore, the pre-surgery study prostate tissue to verify that the drug had reached its target. design allows for the determination of an optimal a and b show control and treated tissue respectively. biologically effective dose and tissue drug levels in addition to the usual parameters of toxicity. Patients with localized prostate cancer with high-risk features or sulfated glycoprotein 2. CLU is associated with and candidates for prostatectomy were enrolled to several diseases including neuroblastoma57, lym- this dose-escalation trial. OGX-011 was given by i.v. phoma58, prostate59, breast60, urothelial61 and infusion over 2 hours at a starting dose of 40 mg on renal-cell carcinomas62, Alzhiemer’s disease63 and days 1, 3, 5, 8, 15, 22, and 29. Androgen-deprivation nephrotoxic injury64. Clusterin levels increase dur- therapy was started on day 1 and prostatectomy was ing castration-induced apoptosis in rat prostate performed on days 30–36. Twenty-five patients were epithelial cells65, and in androgen-dependent pros- enrolled to 6 cohorts and doses of OGX-011 up to tate cancer models66–68. In human prostate cancer, 640 mg were delivered FIG 2. clusterin levels are low or absent in most untreated This Phase I trial demonstrates that OGX-011 is hormone-naive tissues, but increase significantly well tolerated and potently inhibits CLU expression within weeks after neoadjuvant hormone therapy69. in prostate cancers. The Phase II dose for OGX-011 Because clusterin binds to a wide range of biological is 640 mg based on pharmacokinetic and pharmaco- ligands70,71, and is regulated by transcription factor dynamic parameters, and is significantly higher than heat-shock factor 1 REF. 72, an emerging view sug- doses selected for G3139 and ISIS 3521 reviewed gests that clusterin functions in a similar way to above, highlighting a potential cause for the lack of heat-shock proteins and is able to chaperone and efficacy of these agents. Phase II studies of OGX-011 stabilize conformations of proteins during times of in combination with chemotherapy have begun in cell stress. Indeed, clusterin is substantially more patients with prostate, breast and lung cancers. potent than other heat-shock proteins (HSPs) at inhibiting stress-induced protein precipitation73. Survivin. The inhibitor of apoptosis (IAP) gene fam- Experimental and clinical studies associate clus- ily encodes proteins that protect cells from undergo- terin with the development of hormone and drug ing apoptosis through, at least in part, inhibition of resistance, and clusterin has been shown to sup- caspases, which are key effector proteins of apoptotic press apoptotic cell death in response to androgen cell death84–87. Survivin is a member of the IAP gene withdrawal, chemotherapy or radiation66,74–76. family, and has an important role in both cell division An antisense inhibitor targeting the translation- and apoptosis inhibition84,88,89. Survivin is expressed initiation site of human CLU mRNA (OGX-011) was at a high level in a wide range of human cancer types, developed by OncoGeneX Technologies. OGX-011 including lung, colon, pancreas, breast and pros- has 2′-MOE modifications to the four bases on either tate cancers88,90. However, survivin is generally not end of the 21-mer phosphorothioate backbone77. Such expressed in normal tissue, with expression limited ‘gapmer’ modifications maintain the improved tissue to a few cell types including angiogenic endothelium, pharmacokinetic profile of the second-generation thymus, testis, activated T cells and intestinal epi- chemistry but preserves the high affinity of the mol- thelium crypts. Survivin expression levels correlate ecule for target mRNA and the recruitment of RNase with lower apoptotic index in tumour cells and poor H necessary for activity (BOX. 2). In preclinical models prognosis in cancer patients, and serial analysis of of prostate cancer, OGX-011 improved the efficacy of gene expression studies have indicated that survivin chemotherapy, radiation, and androgen withdrawal is the fourth most common gene that is uniformly by inhibiting expression of clusterin and enhancing expressed in cancer cells but not in normal tis- the apoptotic response66,75–77. Additional preclinical sues91. Furthermore, overexpression of survivin in activity was reported in lung78, renal-cell79, urothelial80 tumour cells inhibits chemotherapy-induced (for and breast81 carcinomas and sarcoma82. Furthermore, example, paclitaxel or etoposide), BAX-induced and

REVIEWS

FAS-induced apoptosis, and expression of dominant- STAT3. The signal transducer and activator of tran- negative mutants of survivin induces apoptosis in scription (STAT) factors function as downstream tumour cell lines92. Taken together, these observa- effectors of many cytokine and growth-factor recep- tions make survivin a very attractive target for novel tors. Following specific receptor stimulation and cancer therapy. dimerization, activation of the Janus tyrosine kinases A second-generation 2′-MOE ASO has been or SRC family members results in the phosphoryla- identified (LY2181308) that potently and specifically tion and activation of STAT family members101. downregulates survivin expression in a broad range of Once activated, STATs dimerize and translocate human cancer cells, including lung, colon, pancreas, to the nucleus, where they bind to specific DNA- breast and prostate cells93,94. Survivin inhibition in regulatory elements, resulting in the regulation of tumour cells by LY2181308 results in caspase-3- gene expression. dependent apoptosis, cell-cycle arrest in the G2/M A crucial role for STAT3 in malignant transfor- phase, the formation of multinucleated cells, and in the mation was first proposed in studies demonstrating sensitization of tumour cells to chemotherapy-induced constitutive activation of STAT3 in oncogene-trans- apoptosis93–95. Moreover, LY2181308 produces potent formed cells102. Since then, an abundance of stud- antitumour activity against a broad range of tumour ies have presented strong evidence that persistent types in human tumour xenograft models (see ISIS STAT3 signalling activity participates in malignant Pharmaceuticals in Online links box). Antitumour transformation, resulting in the increased expression activity displayed by LY2181308 in these models is oli- of genes associated with proliferation, cell survival, gonucleotide-sequence specific, and is associated with angiogenesis103–105 and inhibition of the production reduced survivin levels in tumours and the production of inflammatory signals, thereby facilitating evasion of multinucleated tumour cells. Based on these impres- of the immune system by tumour cells106. Based on sive preclinical results, LY2181308 has been selected this evidence, as well as the difficulties associated for clinical development and Phase I studies have been with discovering small-molecule inhibitors against initiated against a broad range of human cancers. transcription factors, antisense strategies are underway to target STAT3 as a novel approach to treat XIAP. X-linked IAP (XIAP) overexpression inhibits human cancers. apoptosis arising from chemotherapy, radiation and Screening of second-generation 2′-MOE ASOs growth-factor deprivation. By inhibiting caspase-3 against human STAT3 identified a highly potent and and caspase-9 activity, XIAP suppresses apoptosis selective ASO that inhibits STAT3 expression in cells triggered by intrinsic (mitochondrial-related) as well (ISIS 345794). Treatment with STAT3 ASOs results in as extrinsic (death-receptor-related) insults96,97. As reduced STAT3 levels in human tumour cells in vitro described with BCL2 and clusterin ASOs above, the and in vivo. Reduction in STAT3 expression promotes XIAP knockdown using ASOs in cancer cells under tumour-cell death and increases sensitivity to chemo- stress and primed for apoptosis (for example, when therapeutic agents in a range of tumour types in vitro challenged by chemotherapeutic agents), enhances and human xenograft models in vivo, including mul- pro-apoptotic signals and tips the apoptotic rheostat tiple myeloma, melanoma, lymphoma and prostate towards death. XIAP is highly expressed in many tumour107,108. Based on these studies, ISIS 345794 has tumour cell lines, and in glioblastoma, acute myeloid been selected for clinical development and initiation leukaemia (AML) and prostate, pancreatic, gastric of Phase I studies is expected in the near future for and colorectal tumours98. In AML, this overexpres- the treatment of multiple myeloma, lymphoma and sion has been associated with poor clinical outcome98. other forms of cancer. AEG35156/GEM640 (Aegera Therapeutics) is a 19-mer ASO targeted to human XIAP mRNA that HSP27. Many components of survival and apoptotic incorporates 2′-O-methyl chemistry with a phos- pathways are regulated by molecular chaperones, phorothioate backbone. Numerous in vitro and in vivo such as HSPs. HSPs are a family of highly conserved preclinical proof-of-concept studies have demon- proteins whose expression is induced by signals such strated that inhibition of XIAP protein expression by as hyperthermia, oxidative stress, activation of the AEG35156/GEM640 enhances the antitumour activity FAS death receptor, and cytotoxic drugs109–111. HSPs of chemotherapy in several xenograft models99,100. have attracted attention as new therapeutic targets A Phase I dose-escalation tolerability study of for cancer, especially since the discovery and charac- AEG35156/GEM640 as a single agent is currently terization of geldanamycin as an inhibitor of HSP90 underway in the United Kingdom as a 7-day continu- REFS 112,113 and the targeting of CLU76–81, the product ous i.v. infusion in patients with advanced tumours. of which has a small HSP-like function. A second Phase I trial is planned to begin in Canada HSP27 is an ATP-independent molecular chap- in 2005 as a dose-escalation tolerability study erone that is highly induced during stress responses of AEG35156/GEM640 in combination with and can form oligomers to increase affinity for client docetaxel. It is unknown whether tissue half-life proteins, preventing their precipitation and aggre- and surrogate markers of target suppression will be gation. The cytoprotective effects of HSP27 result quantified to identify the optimal biological dose from its chaperone function, direct interference of from these trials. caspase activation, modulation of oxidative stress,

REVIEWS

Box 3 | Future potential ASOs for the treatment of cancer that targets the human translation-initiation site were reported to potently inhibit HSP27 expression in human Among the more promising preclinical antisense oligonucleotide (ASO) agents for prostate PC3 cells with increased caspase-3 cleavage and prostate cancer are those that target the insulin-like growth factor (IGF)-binding apoptosis, and 87% suppression of cell growth121. HSP27 proteins (IGFBPs). IGF1 has an important role in the pathophysiology of prostatic ASO and siRNA also enhanced paclitaxel chemosensi- disease and its activity is regulated by various factors in the microenvironment, 139,140 tivity in vitro, while in vivo, systemic administration of including the IGFBPs . In different physiological contexts, IGFBPs can either HSP27 ASO in athymic mice decreased PC3 tumour increase or decrease IGF signalling. Increased levels of IGFBP5 potentiate IGF1 progression and enhanced paclitaxel chemosensitivity. signalling in prostate cells after androgen ablation, and ASOs attenuate castration- Overexpression of HSP27 in human prostate LNCaP induced increases in IGFBP5 and delayed outgrowth of androgen-independent tumours141,142. IGFBP2 levels also increase in prostate tumours after castration and cells caused these normally androgen-dependent cells IGFBP2 ASO decreased the growth of LNCaP prostate cells in vitro and in vivo. to become androgen independent and more resistant Increased IGFBP5 and IGFBP2 levels after androgen ablation represent adaptive to cytotoxic chemotherapy. These findings indicate that responses to potentiate IGF1-mediated survival and mitogenesis143. Use of ASO to increased levels of HSP27 after androgen withdrawal target IGFBP-modulation of IGF signalling is undergoing further study, and a provide a cytoprotective role during development bispecific second-generation 2′-MOE ASO that simultaneously suppresses both of androgen independence and that ASO-induced IGFBP5 and IGFBP2 mRNA is under development for clinical applications by silencing can increase apoptosis and delay tumour OncoGenex Technologies. progression. A second-generation MOE gapmer ASO Another promising target, MCL1, is a member of the BCL2 protein family, and like targeting HSP27 (OGX-427, OncoGeneX Technologies) other BCL2 family members, possesses BCL2 homology domain BH1-3 and is is planned to enter clinical trials in solid cancers and strongly associated with the suppression of tumour apoptosis and the promotion of multiple myeloma in 2006. malignancy144. Although the molecular mechanisms by which MCL1 expression inhibits apoptosis remain to be elucidated, MCL1 heterodimerizes with proteins Other ASOs showing promising preclinical activity. known to promote apoptosis (for example, BAX), thereby neutralizing their pro- Many other ASOs are currently undergoing preclini- 145 apoptotic activity . MCL1 is overexpressed in many human tumour specimens, and cal testing as potential cancer treatments and it is 146–150 promotes resistance to chemotherapy-induced apoptosis . A second-generation beyond the scope of this review to list all of them. 2′-MOE ASO (ISIS 20408) that potently inhibits MCL1 expression in a range of The more promising agents are discussed in BOX 3. human tumour cell types has been identified146,151–153. Treatment of tumour cells in vitro promotes apoptosis and sensitizes tumour cells to chemotherapy-induced Conclusions apoptosis, and in vivo administration delays the growth of a range of human Antisense inhibition of relevant genes involved in xenograft models, including models of gastric and colon cancer, and of sarcoma and cancer progression remains a promising area for melanoma146,151–153. Preclinical evaluation of ISIS 20408 continues with the goal of initiating clinical trials for cancer in the future. therapeutic development. ASO technology has Additional currently validated targets for antisense therapy in preclinical studies quickly moved from preclinical models to test- include GEM240 (Hybridon), which targets the p53 regulator MDM2 REFS 154156, ing in the clinic. Challenges remain to optimize HRAS (ISIS 2503)157, methyltransferase (MG98, MethylGene Inc)158, BCR–ABL159, tissue exposure, cellular uptake, and demonstra- HSP70 REF. 160, MYB161, PKA (GEM 231)162, RAF (ISIS 5132)163,164, p53 REF. 165 and tion of mechanism and antitumour activity in the the androgen receptor166. clinic. Lack of a success in recent randomized trials in lung cancer, CLL, myeloma and melanoma has dampened enthusiasm for ASO therapeutics. and regulation of the cytoskeleton114,115. Higher levels However, the improved chemistry applied to ASOs of HSP27 are commonly detected in many cancers like the second-generation MOE gapmers, offer including breast116, ovarian117, glial118, prostate68,119 and significant advantages for weekly dosing regimens, other tumours120, and are associated with metastasis, increased potency and higher tissue concentra- poor prognosis and resistance to chemotherapy or tions. As is true for clinical development of all radiation121,122. Moreover, HSP27 expression is targeted therapies, crucial issues include the early induced by hormone or chemotherapy and inhib- determination of optimal biological dose, ensuring its treatment-induced apoptosis through multiple that the target is relevant in the patient population mechanisms121–126. being studied, and the rational use of combination As an important regulator of cell survival and treat- strategies with study designs that will yield unam- ment stress at many different points along the apoptotic biguous end points. Addressing these issues early pathway, HSP27 is now recognized as an important on will allow optimal use of these agents clinically therapeutic target. Recently, HSP27 ASO and siRNA and will best ensure success in Phase III trials.

1. Zamecnik, P. C. & Stephenson, M. L. Inhibition of Rous 3. Dean, N. M. & Bennett, C. F. Antisense oligonucleotide- 7. Galarneau, A., Min, K. L., Mangos, M. M. & Damha, M. J. sarcoma virus replication and cell transformation by a based therapeutics for cancer. Oncogene 22, 9087–9096 Assay for evaluating ribonuclease H-mediated specific oligodeoxynucleotide. Proc. Natl Acad. Sci. USA (2003). degradation of RNA-antisense oligonucleotide 75, 280–284 (1978). 4. Orr, R. M. & Monia, B. P. Antisense therapy for cancer. duplexes. Methods Mol. Biol. 288, 65–80 First report on the use of antisense oligodeoxy- Investig. Drugs 1, 199–205 (1998). (2005). nucleotides to suppress gene activity in cell culture. 5. Crooke, S. T. Molecular mechanisms of antisense drugs: 8. Carpentier, A. F., Chen, L., Maltonti, F. & 2. Gleave, M., Miyake, H., Zangemeister-Wittke, U. & RNase H. Antisense Nucleic Acid Drug Dev. 8, 133–134 (1998). Delattre, J. Y. Oligodeoxynucleotides containing Jansen, B. Antisense therapy: current status in prostate 6. Wu, H. et al. Determination of the role of the human RNase CpG motifs can induce rejection of a neuroblastoma cancer and other malignancies. Cancer Metastasis Rev. H1 in the pharmacology of DNA-like antisense drugs. in mice. Cancer Res. 59, 5429–5432 21, 79–92 (2002). J. Biol. Chem. 279, 17181–17189 (2004). (1999).

REVIEWS

9. Fire, A. et al. Potent and specific genetic interference by bcl-2 antisense (genasense) and docetaxel (D) in hormone 52. Yuen, A. R. et al. Phase I study of an antisense double-stranded RNA in Caenorhabditis elegans. Nature refractory prostate cancer. Proc. Am. Soc. Clin. Oncol. 20, oligonucleotide to protein kinase C-α (ISIS 3521/CGP 391, 806–811 (1998). A119 (2001). 64128A) in patients with cancer. Clin. Cancer Res. 5, First report on the use of double-stranded siRNA 31. Chi, K. N., Murray, R. N. & Gleave, M. E. A phase II study of 3357–3363 (1999). duplexes to suppress gene activity in animal cells. oblimersen sodium (G3139) and docetaxel (D) in patients 53. Cunningham, C. C. et al. A phase I trial of c-Raf kinase 10. Matzke, M. A. & Birchler, J. A. RNAi-mediated pathways in (pts) with metastatic hormone-refractory prostate cancer antisense oligonucleotide ISIS 5132 administered as a the nucleus. Nature Rev. Genet. 6, 24–35 (2005). (HRPC). Proc. Am. Soc. Clin. Oncol. 22, 393 (2003). continuous intravenous infusion in patients with advanced 11. Soutschek, J. et al. Therapeutic silencing of an 32. Han, Z. et al. Isolation and characterization of an cancer. Clin. Cancer Res. 6, 1626–1631 (2000). endogenous gene by systemic administration of modified apoptosis-resistant variant of human leukemia HL-60 cells 54. Tolcher, A. W. et al. A randomized phase II and siRNAs. Nature 432, 173–178 (2004). that has switched expression from Bcl-2 to Bcl-xL. Cancer pharmacokinetic study of the antisense oligonucleotides The first demonstration of siRNA activity in Res. 56, 1621–1628 (1996). ISIS 3521 and ISIS 5132 in patients with hormone- mammals following administration through routes 33. Leech, S. H. et al. Induction of apoptosis in lung-cancer refractory prostate cancer. Clin. Cancer Res. 8, 2530–2535 other than the digestive tract. cells following bcl-xL anti-sense treatment. Int. J. Cancer (2002). 12. Monia, B. P. et al. Evaluation of 2′-modified 86, 570–576 (2000). 55. Yuen, A. R. et al. Phase I/II trial of ISIS 3521, an antisense oligonucleotides containing 2′-deoxy gaps as antisense 34. Simoes-Wust, A. P. et al. Bcl-xl antisense treatment inhibitor of PKC-α, with carboplatin and paclitaxel in non- inhibitors of gene expression. J. Biol. Chem. 268, 14514– induces apoptosis in breast carcinoma cells. Int. J. Cancer small cell lung cancer. Proc. Am. Soc. Clin. Oncol. A1234 14522 (1993). 87, 582–590 (2000). (2001). This report demonstrates that RNase H is a key 35. Lebedeva, I., Rando, R., Ojwang, J., Cossum, P. & 56. Ritch, P. et al. Phase I/II trial of ISIS 3521/LY900003, an α terminating mechanism by which antisense drugs Stein, C. A. Bcl-xL in prostate cancer cells: effects of antisense inhibitor of PKC- , with cisplatin and exert their effects on gene expression. This report overexpression and down-regulation on chemosensitivity. gemcitabine in advanced non-small cell lung cancer. Proc. also demonstrates the use of 2’-modified chimeric Cancer Res. 60, 6052–6060 (2000). Am. Soc. Clin. Oncol. 21, 309a (2002). antisense oligonucleotides in cells. 36. Leung, S., Miyake, H., Zellweger, T., Tolcher, A. & 57. Cervellera, M. et al. Direct transactivation of the anti- 13. Shen, L. et al. Evaluation of C-5 propynyl pyrimidine- Gleave, M.E. Synergistic chemosensitization and inhibition apoptotic gene apolipoprotein J (clusterin) by B-MYB. containing oligonucleotides in vitro and in vivo. Antisense of progression to androgen independence by antisense J. Biol. Chem. 275, 21055–21060 (2000). Nucleic Acid Drug Dev. 13, 129–142 (2003). Bcl-2 oligodeoxynucleotide and paclitaxel in the LNCaP 58. Wellmann, A. et al. Detection of differentially expressed 14. Tsujimoto, Y. & Croce, C.M. Analysis of the structure, prostate tumor model. Int. J. Cancer 91, 846–850 (2001). genes in lymphomas using cDNA arrays: identification of transcripts, and protein products of bcl-2, the gene 37. Miyake, H., Monia, B. P. & Gleave, M. E. Inhibition of clusterin as a new diagnostic marker for anaplastic large- involved in human follicular lymphoma. Proc. Natl Acad. progression to androgen-independence by combined cell lymphomas. Blood 96, 398–404 (2000).

Sci. USA 83, 5214–5218 (1986). adjuvant treatment with antisense BCL-XL and antisense 59. Steinberg, J. et al. Intracellular levels of SGP-2 (Clusterin) 15. Reed, J. C. Bcl-2 and the regulation of programmed cell Bcl-2 oligonucleotides plus taxol after castration in the correlate with tumor grade in prostate cancer. Clin. Cancer death. J. Cell Biol. 124, 1–6 (1994). Shionogi tumor model. Int. J. Cancer 86, 855–862 (2000). Res. 3, 1707–1711 (1997). 16. Miyashita, T. & Reed, J.C. Bcl-2 oncoprotein blocks 38. Gautschi, O. et al. Activity of a novel bcl-2/bcl-xL-bispecific 60. Redondo, M. et al. Overexpression of clusterin in human chemotherapy-induced apoptosis in a human leukemia antisense oligonucleotide against tumors of diverse breast carcinoma. Am. J. Pathol. 157, 393–399 (2000). cell line. Blood 81, 151–157 (1993). histologic origins. J. Natl Cancer Inst. 93, 463–471 61. Miyake, H., Gleave, M., Kamidono, S. & Hara, I. 17. McDonnell, T. J. et al. Expression of the protooncogene (2001). Overexpression of clusterin in transitional cell carcinoma of bcl-2 in the prostate and its association with emergence of 39. Newton, A. C. Regulation of protein kinase C. Curr. Opin. the bladder is related to disease progression and androgen-independent prostate cancer. Cancer Res. 52, Cell Biol. 9, 161–167 (1997). recurrence. Urology 59, 150–154 (2002). 6940–6944 (1992). 40. Liu, B., Maher, R. J., Hannun, Y. A., Porter, A. T. & 62. Parczyk, K., Pilarsky, C., Rachel, U. & Koch-Brandt, C. 18. Kyprianou, N., King, E. D., Bradbury, D. & Rhee, J. G. Honn, K. V. 12(S)-HETE enhancement of prostate tumor Gp80 (clusterin; TRPM-2) mRNA level is enhanced in bcl-2 over-expression delays radiation-induced apoptosis cell invasion: selective role of PKCα. J. Natl Cancer Inst. human renal clear cell carcinomas. J. Cancer Res. Clin. without affecting the clonogenic survival of human prostate 86, 1145–1151 (1994). Oncol. 120, 186–188 (1994). cancer cells. Int. J. Cancer 70, 341–348 (1997). 41. Ways, D. K. et al. MCF-7 breast cancer cells transfected 63. Calero, M. et al. Apolipoprotein J (clusterin) and Alzheimer’s 19. Raffo, A. J. et al. Overexpression of bcl-2 protects prostate with protein kinase C-α exhibit altered expression of other disease. Microsc. Res. Tech. 50, 305–315 (2000). cancer cells from apoptosis in vitro and confers resistance protein kinase C isoforms and display a more aggressive 64. Rosenberg, M. E. & Silkensen, J. Clusterin and the kidney. to androgen depletion in vivo. Cancer Res. 55, 4438–4445 neoplastic phenotype. J. Clin. Invest. 95, 1906–1915 Exp. Nephrol. 3, 9–14 (1995). (1995). (1995). 65. Montpetit, M. L., Lawless, K. R. & Tenniswood, M. 20. Miyake, H., Tolcher, A. & Gleave, M. E. Antisense Bcl-2 42. Gill, P. K., Gescher, A. & Gant, T. W. Regulation of MDR1 Androgen-repressed messages in the rat ventral prostate. oligodeoxynucleotides inhibit progression to androgen- promoter activity in human breast carcinoma cells by Prostate 8, 25–36 (1986). independence after castration in the Shionogi tumor protein kinase C isozymes α and τ. Eur. J. Biochem. 268, 66. Miyake, H., Nelson, C., Rennie, P. S. & Gleave, M. E. model. Cancer Res. 59, 4030–4034 (1999). 4151–4157 (2001). Testosterone-repressed prostate message-2 is an 21. Gleave, M. et al. Progression to androgen independence is 43. Swannie, H. C. & Kaye, S. B. Protein kinase C inhibitors. antiapoptotic gene involved in progression to androgen delayed by adjuvant treatment with antisense Bcl-2 Curr. Oncol. Rep. 4, 37–46 (2002). independence in prostate cancer. Cancer Res. 60, 170–176 oligodeoxynucleotides after castration in the LNCaP prostate 44. Isonishi, S., Ohkawa, K., Tanaka, T. & Howell, S. B. (2000). tumor model. Clin. Cancer Res. 5, 2891–2898 (1999). Depletion of protein kinase C (PKC) by 12-O- Reported that increased levels of clusterin after 22. Zangemeister-Wittke, U. et al. A novel bispecific antisense tetradecanoylphorbol-13-acetate (TPA) enhances platinum hormone therapy conferred therapeutic resistance, oligonucleotide inhibiting both bcl-2 and bcl-xL expression drug sensitivity in human ovarian carcinoma cells. Br. J. identifying clusterin for the first time as a efficiently induces apoptosis in tumor cells. Clin. Cancer Cancer 82, 34–38 (2000). cytoprotective protein and therapeutic target. Res. 6, 2547–2555 (2000). 45. Shen, L., Dean, N. M. & Glazer, R. I. Induction of p53- 67. Kyprianou, N., English, H. F. & Isaacs, J. T. Programmed 23. Jansen, B. et al. bcl-2 antisense therapy chemosensitizes dependent, insulin-like growth factor-binding protein-3- cell death during regression of PC-82 human prostate human melanoma in SCID mice. Nature Med. 4, 232–234 mediated apoptosis in glioblastoma multiforme cells by a cancer following androgen ablation. Cancer Res. 50, (1998). protein kinase Cα antisense oligonucleotide. Mol. 3748–3753 (1990). 24. Webb, A. et al. BCL-2 antisense therapy in patients with Pharmacol. 55, 396–402 (1999). 68. Bubendorf, L. et al. Hormone therapy failure in human non-Hodgkin lymphoma. Lancet 349, 1137–1141 (1997). 46. Wang, X. Y., Repasky, E. & Liu, H. T. Antisense inhibition of prostate cancer: analysis by complementary DNA and tissue This Phase I/IIa trial is the first report of the activity protein kinase Cα reverses the transformed phenotype in microarrays. J. Natl Cancer Inst. 91, 1758–1764 (1999). of BCL2 antisense in cancer patients. human lung carcinoma cells. Exp. Cell Res. 250, 253–263 69. July, L. V. et al. Clusterin expression is significantly 25. Waters, J. S. et al. Phase I clinical and pharmacokinetic (1999). enhanced in prostate cancer cells following androgen study of bcl-2 antisense oligonucleotide therapy in patients 47. Dean, N. M. & McKay, R. Inhibition of protein kinase Cα withdrawal therapy. Prostate 50, 179–188 (2002). with non-Hodgkin’s lymphoma. J. Clin. Oncol. 18, 1812– expression in mice after systemic administration of 70. Koch-Brandt, C. & Morgans, C. Clusterin: a role in cell 1823 (2000). phosphorothioate antisense oligodeoxynucleotides. Proc. survival in the face of apoptosis? Prog. Mol. Subcell. Biol. 26. Jansen, B. et al. Chemosensitisation of malignant Natl Acad. Sci. USA 91, 11762–11766 (1994). 16, 130–149 (1996). melanoma by BCL2 antisense therapy. Lancet 356, 1728– Demonstrated for the first time that antisense 71. Wilson, M. R. & Easterbrook-Smith, S. B. Clusterin is a 1733 (2000). oligonucleotides can inhibit gene expression in a secreted mammalian chaperone. Trends Biochem. Sci. This Phase I/IIa trial was the first to report reduced highly specific manner when administered 25, 95–98 (2000). BCL2 protein levels in serial melanoma biopsies and parenterally to normal mice. 72. Michel, D., Chatelain, G., North, S. & Brun, G. Stress- higher-than-expected response rates using G3139 48. Yazaki, T. et al. Treatment of glioblastoma U-87 by induced transcription of the clusterin/apoJ gene. Biochem. plus DTIC in a small cohort of patients with systemic administration of an antisense protein kinase C-α J. 328, 45–50 (1997). metastatic melanoma. phosphorothioate oligodeoxynucleotide. Mol. Pharmacol. 73. Humphreys, D. T., Carver, J. A., Easterbrook-Smith, S. B. 27. Rai, K. R. & Moore, J. O. Phase 3 randomized trial of 50, 236–242 (1996). & Wilson, M. R. Clusterin has chaperone-like activity fludarabine/cyclophosphamide chemotherapy with or 49. Geiger, T., Muller, M., Dean, N. M. & Fabbro, D. Antitumor similar to that of small heat shock proteins. J. Biol. Chem. without oblimersen sodium (Bcl-2 antisense; genasense; activity of a PKC-α antisense oligonucleotide in 274, 6875–6881 (1999). G3139) for patients with relapsed or refractory chronic combination with standard chemotherapeutic agents 74. Sensibar, J. A. et al. Prevention of cell death induced by lymphocytic leukemia (CLL). Blood 104, 100a (2004). against various human tumors transplanted into nude tumor necrosis factor alpha in LNCaP cells by 28. Chanan-Khan, A. Bcl-2 antisense therapy in B-cell mice. Anticancer Drug Des. 13, 35–45 (1998). overexpression of sulfated glycoprotein-2 (clusterin). malignancies. Blood Rev. 19, 213–221 (2005). 50. Dennis, J. U. et al. Human melanoma metastasis is Cancer Res. 55, 2431–2437 (1995). 29. Chi, K. N. et al. A phase I dose-finding study of combined inhibited following ex vivo treatment with an antisense 75. Zellweger, T. et al. Enhanced radiation sensitivity in treatment with an antisense Bcl-2 oligonucleotide oligonucleotide to protein kinase C-α. Cancer Lett. 128, prostate cancer by inhibition of the cell survival protein (Genasense) and mitoxantrone in patients with metastatic 65–70 (1998). clusterin. Clin. Cancer Res. 8, 3276–3284 (2002). hormone-refractory prostate cancer. Clin. Cancer Res. 7, 51. Nemunaitis, J. et al. Phase I evaluation of ISIS 3521, an 76. Miyake, H., Chi, K. N. & Gleave, M. E. Antisense TRPM-2 3920–3927 (2001). antisense oligodeoxynucleotide to protein kinase C-α, in oligodeoxynucleotides chemosensitize human androgen- 30. De Bono, J. S., Rowinsky, E. K. & Kuhn, J. Phase I patients with advanced cancer. J. Clin. Oncol. 17, 3586– independent PC-3 prostate cancer cells both in vitro and pharmacokinetic (PK) and pharmacodynamic (PD) trial of 3595 (1999). in vivo. Clin. Cancer Res. 6, 1655–1663 (2000).

REVIEWS

77. Zellweger, T. et al. Antitumor activity of antisense clusterin apoptosis in chemoresistant human ovarian cancer cells. 127. Ho, S. P. et al. Mapping of RNA accessible sites for oligonucleotides is improved in vitro and in vivo by Cancer Res. 60, 5659–5666 (2000). antisense experiments with oligonucleotide libraries. incorporation of 2′-O-(2-methoxy)ethyl chemistry. 101. Darnell, J. E. Jr., Kerr, I. M. & Stark, G. R. Jak-STAT Nature Biotechnol. 16, 59–63 (1998). J. Pharmacol. Exp. Ther. 298, 934–940 (2001). pathways and transcriptional activation in response to 128. Matveeva, O., Felden, B., Audlin, S., Gesteland, R. F. & A preclinical pharmacology paper that demonstrated IFNs and other extracellular signaling proteins. Science Atkins, J. F. A rapid in vitro method for obtaining RNA that tissue half-life and target suppression was 264, 1415–1421 (1994). accessibility patterns for complementary DNA probes: superior using a second-generation 2’-MOE OGX- 102. Yu, C. L. et al. Enhanced DNA-binding activity of a Stat3- correlation with an intracellular pattern and known RNA 011 compared to unmodified phosphorthioate ASO. related protein in cells transformed by the Src oncoprotein. structures. Nucleic Acids Res. 25, 5010–5016 (1997). These data provided the preclinical proof of principle Science 269, 81–83 (1995). 129. Stull, R. A., Taylor, L. A. & Szoka, F. C. Jr. Predicting to proceed into human trials using OGX-011. 103. Yu, H. & Jove, R. The STATs of cancer — new molecular antisense oligonucleotide inhibitory efficacy: a 78. July, L. V. et al. Nucleotide-based therapies targeting targets come of age. Nature Rev. Cancer 4, 97–105 computational approach using histograms and clusterin chemosensitize human lung adenocarcinoma cells (2004). thermodynamic indices. Nucleic Acids Res. 20, 3501– both in vitro and in vivo. Mol. Cancer Ther. 3, 223–232 104. Buettner, R., Mora, L. B. & Jove, R. Activated STAT 3508 (1992). (2004). signaling in human tumors provides novel molecular 130. Butler, M. et al. Specific inhibition of PTEN expression 79. Zellweger, T. et al. Chemosensitization of human renal cell targets for therapeutic intervention. Clin. Cancer Res. 8, reverses hyperglycemia in diabetic mice. Diabetes 51, cancer using antisense oligonucleotides targeting the 945–954 (2002). 1028–1034 (2002). antiapoptotic gene clusterin. Neoplasia 3, 360–367 (2001). 105. Mora, L. B. et al. Constitutive activation of Stat3 in human 131. Geary, R. S. et al. Pharmacokinetic properties of 2′-O-(2- 80. Miyake, H., Hara, I., Kamidono, S. & Gleave, M. E. prostate tumors and cell lines: direct inhibition of Stat3 methoxyethyl)-modified oligonucleotide analogs in rats. Synergistic chemsensitization and inhibition of tumor growth signaling induces apoptosis of prostate cancer cells. J. Pharmacol. Exp. Ther. 296, 890–897 (2001). and metastasis by the antisense oligodeoxynucleotide Cancer Res. 62, 6659–6666 (2002). Describes the pharmacokinetic attributes of second- targeting clusterin gene in a human bladder cancer model. 106. Cheng, F. et al. A critical role for Stat3 signaling in immune generation 2’-MOE antisense oligonucleotides in Clin. Cancer Res. 7, 4245–4252 (2001). tolerance. Immunity 19, 425–436 (2003). animals. 81. Biroccio, A., D’Angelo, C., Gleave, M. & Gonos, E. S. 107. Epling-Burnette, P. K. et al. Inhibition of STAT3 signaling 132. Henry, S. P., Geary, R. S., Yu, R. & Levin, A. A. Drug Antisense clusterin oligodeoxynucleotides increase the leads to apoptosis of leukemic large granular lymphocytes properties of second-generation antisense response of HER-2 gene amplified breast cancer cells to and decreased Mcl-1 expression. J. Clin. Invest. 107, oligonucleotides: how do they measure up to their trastuzumab. J. Cell Physiol. 31 Jan 2005 (doi:10.1002/ 351–362 (2001). predecessors? Curr. Opin. Investig. Drugs 2, 1444–1449 jcp.20295). 108. Turkson, J. & Jove, R. STAT proteins: novel molecular (2001). 82. Trougakos, I. P., So, A., Jansen, B., Gleave, M. E. & targets for cancer drug discovery. Oncogene 19, 133. Henry, S. P., Monteith, D., Bennett, F. & Levin, A. A. Gonos, E. S. Silencing expression of the clusterin/ 6613–6626 (2000). Toxicological and pharmacokinetic properties of chemically apolipoprotein J gene in human cancer cells using small 109. Garrido, C. et al. HSP27 and HSP70: potentially modified antisense oligonucleotide inhibitors of PKC-α and interfering RNA induces spontaneous apoptosis, reduced oncogenic apoptosis inhibitors. Cell Cycle 2, 579–584 C-raf kinase. Anticancer Drug Des. 12, 409–420 (1997). growth ability, and cell sensitization to genotoxic and (2003). 134. Levin, A. A. A review of the issues in the pharmacokinetics oxidative stress. Cancer Res. 64, 1834–1842 (2004). 110. Concannon, C. G., Gorman, A. M. & Samali, A. On the role and toxicology of phosphorothioate antisense 83. Chi, K. et al. A phase I pharmacokinetic (PK) and of Hsp27 in regulating apoptosis. Apoptosis 8, 61–70 oligonucleotides. Biochim. Biophys. Acta 1489, 69–84 pharmacodynamic (PD) study of OGX–011, a (2003). (1999). 2′methoxyethyl phosphorothioate antisense to clusterin, in 111. Ciocca, D. R., Oesterreich, S., Chamness, G. C., McGuire, 135. Henry, S. P., Monteith, D. & Levin, A. A. Antisense patients with prostate cancer prior to radical W. L. & Fuqua, S. A. Biological and clinical implications of oligonucleotide inhibitors for the treatment of cancer: 2. prostatectomy. J. Clin. Oncol. 22, 3033 (2004). heat shock protein 27,000 (Hsp27): a review. J. Natl Toxicological properties of phosphorothioate 84. Li, F. et al. Control of apoptosis and mitotic spindle Cancer Inst. 85, 1558–1570 (1993). oligodeoxynucleotides. Anticancer Drug Des. 12, 395–408 checkpoint by survivin. Nature 396, 580–584 (1998). 112. Neckers, L., Schulte, T. W. & Mimnaugh, E. Geldanamycin (1997). 85. Schimmer, A. D. Inhibitor of apoptosis proteins: translating as a potential anti-cancer agent: its molecular target and 136. Geary, R. S. et al. Pharmacokinetics of a tumor necrosis basic knowledge into clinical practice. Cancer Res. 64, biochemical activity. Invest. New Drugs 17, 361–373 factor-α phosphorothioate 2′-O-(2-methoxyethyl) modified 7183–7190 (2004). (1999). antisense oligonucleotide: comparison across species. 86. Okada, H. & Mak, T. W. Pathways of apoptotic and non- 113. Ramanathan, R. K., Trump, D. L. & Eiseman, J. L. A phase Drug Metab. Dispos. 31, 1419–1428 (2003). apoptotic death in tumour cells. Nature Rev. Cancer 4, I pharmacokinetic (PK) and pharmacodynamic (PD) trial of 137. Cook, P. Second generation antisense oligonucleotides: 592–603 (2004). weekly 17-allylamino-17 demethoxygeldanamycin 2′–modifications. Annu. Rep. Med. Chem. 33, 313 (1998). 87. LaCasse, E. C., Baird, S., Korneluk, R. G. & (17AAG, NSC–704057) in patients with advanced tumors. 138. Vester, B. & Wengel, J. LNA (locked nucleic acid): high- MacKenzie, A. E. The inhibitors of apoptosis (IAPs) and J. Clin. Oncol. 22, 3031 (2004). affinity targeting of complementary RNA and DNA. their emerging role in cancer. Oncogene 17, 3247–3259 114. Liang, P. & MacRae, T. H. Molecular chaperones and the Biochemistry 43, 13233–13241 (2004). (1998). cytoskeleton. J. Cell Sci. 110, 1431–1440 (1997). 139. Pollak, M. N., Schernhammer, E. S. & Hankinson, S. E. 88. Ambrosini, G., Adida, C. & Altieri, D. C. A novel anti- 115. Parcellier, A., Gurbuxani, S., Schmitt, E., Solary, E. & Insulin-like growth factors and neoplasia. Nature Rev. apoptosis gene, survivin, expressed in cancer and Garrido, C. Heat shock proteins, cellular chaperones that Cancer 4, 505–518 (2004). lymphoma. Nature Med. 3, 917–921 (1997). modulate mitochondrial cell death pathways. Biochem. Excellent contemporary review on the IGF1 axis in 89. Tamm, I. et al. IAP-family protein survivin inhibits caspase Biophys. Res. Commun. 304, 505–512 (2003). cancer. activity and apoptosis induced by Fas (CD95), Bax, 116. Conroy, S. E. & Latchman, D. S. Do heat shock proteins 140. Jones, J. I. & Clemmons, D. R. Insulin-like growth factors caspases, and anticancer drugs. Cancer Res. 58, have a role in breast cancer? Br. J. Cancer 74, 717–721 and their binding proteins: biological actions. Endocr. Rev. 5315–5320 (1998). (1996). 16, 3–34 (1995). 90. Lu, C. D., Altieri, D. C. & Tanigawa, N. Expression of a 117. Arts, H. J. et al. Heat-shock-protein-27 (hsp27) expression 141. Miyake, H., Nelson, C., Rennie, P. S. & Gleave, M. E. novel antiapoptosis gene, survivin, correlated with tumor in ovarian carcinoma: relation in response to Overexpression of insulin-like growth factor binding cell apoptosis and p53 accumulation in gastric chemotherapy and prognosis. Int. J. Cancer 84, 234–238 protein-5 helps accelerate progression to androgen- carcinomas. Cancer Res. 58, 1808–1812 (1998). (1999). independence in the human prostate LNCaP tumor model 91. Lal, A. et al. A public database for gene expression in 118. Zhang, R., Tremblay, T. L., McDermid, A., Thibault, P. & through activation of phosphatidylinositol 3′-kinase human cancers. Cancer Res. 59, 5403–5407 (1999). Stanimirovic, D. Identification of differentially expressed pathway. Endocrinology 141, 2257–2265 (2000). 92. Altieri, D. C. Survivin, versatile modulation of cell division proteins in human glioblastoma cell lines and tumors. Glia 142. Miyake, H., Pollak, M. & Gleave, M. E. Castration-induced and apoptosis in cancer. Oncogene 22, 8581–8589 42, 194–208 (2003). up-regulation of insulin-like growth factor binding protein-5 (2003). 119. Cornford, P. A. et al. Heat shock protein expression potentiates insulin-like growth factor-I activity and 93. Li, F. et al. Pleiotropic cell-division defects and apoptosis independently predicts clinical outcome in prostate cancer. accelerates progression to androgen independence in induced by interference with survivin function. Nature Cell Cancer Res. 60, 7099–7105 (2000). prostate cancer models. Cancer Res. 60, 3058–3064 Biol. 1, 461–466 (1999). 120. Bruey, J. M. et al. Differential regulation of HSP27 (2000). 94. Chen, J. et al. Down-regulation of survivin by antisense oligomerization in tumor cells grown in vitro and in vivo. 143. Kiyama, S. et al. Castration-induced increases in insulin- oligonucleotides increases apoptosis, inhibits cytokinesis Oncogene 19, 4855–4863 (2000). like growth factor-binding protein 2 promotes proliferation and anchorage-independent growth. Neoplasia 2, 121. Rocchi, P. et al. Heat shock protein 27 increases after of androgen-independent human prostate LNCaP tumors. 235–241 (2000). androgen ablation and plays a cytoprotective role in Cancer Res. 63, 3575–3584 (2003). 95. Ansell, S. M. et al. Inhibition of survivin expression hormone-refractory prostate cancer. Cancer Res. 64, 144. Bae, J., Leo, C. P., Hsu, S. Y. & Hsueh, A. J. MCL-1S, a suppresses the growth of aggressive non-Hodgkin’s 6595–6602 (2004). splicing variant of the antiapoptotic BCL-2 family member lymphoma. Leukemia 18, 616–623 (2004). 122. Garrido, C. et al. HSP27 as a mediator of confluence- MCL-1, encodes a proapoptotic protein possessing only 96. Deveraux, Q. L., Takahashi, R., Salvesen, G. S. & dependent resistance to cell death induced by anticancer the BH3 domain. J. Biol. Chem. 275, 25255–25261 Reed, J. C. X-linked IAP is a direct inhibitor of cell-death drugs. Cancer Res. 57, 2661–2667 (1997). (2000). proteases. Nature 388, 300–304 (1997). 123. Vargas-Roig, L. M., Gago, F. E., Tello, O., Aznar, J. C. & 145. Akgul, C., Turner, P. C., White, M. R. & Edwards, S. W. 97. Silke, J. et al. The anti-apoptotic activity of XIAP is retained Ciocca, D. R. Heat shock protein expression and drug Functional analysis of the human MCL-1 gene. Cell Mol. upon mutation of both the caspase 3- and caspase 9- resistance in breast cancer patients treated with induction Life Sci. 57, 684–691 (2000). interacting sites. J. Cell Biol. 157, 115–124 (2002). chemotherapy. Int. J. Cancer 79, 468–475 (1998). 146. Derenne, S. et al. Antisense strategy shows that Mcl-1

98. Salvesen, G. S. & Duckett, C. S. IAP proteins: blocking the 124. Garrido, C. et al. HSP27 inhibits cytochrome c-dependent rather than Bcl-2 or Bcl-xL is an essential survival protein of road to death’s door. Nature Rev. Mol. Cell Biol. 3, activation of procaspase-9. FASEB J. 13, 2061–2070 human myeloma cells. Blood 100, 194–199 (2002). 401–410 (2002). (1999). 147. Zhou, P., Qian, L., Kozopas, K. M. & Craig, R. W. Mcl-1, a 99. Hu, Y. et al. Antisense oligonucleotides targeting XIAP 125. Paul, C. et al. Hsp27 as a negative regulator of Bcl-2 family member, delays the death of hematopoietic induce apoptosis and enhance chemotherapeutic activity cytochrome C release. Mol. Cell. Biol. 22, 816–834 (2002). cells under a variety of apoptosis-inducing conditions. against human lung cancer cells in vitro and in vivo. Clin. 126. Concannon, C. G., Orrenius, S. & Samali, A. Hsp27 Blood 89, 630–643 (1997). Cancer Res. 9, 2826–2836 (2003). inhibits cytochrome c-mediated caspase activation by 148. Miyamoto, Y. et al. Immunohistochemical analysis of Bcl-2, 100. Sasaki, H., Sheng, Y., Kotsuji, F. & Tsang, B. K. Down- sequestering both pro-caspase-3 and cytochrome c. Bax, Bcl-X, and Mcl-1 expression in pancreatic cancers. regulation of X-linked inhibitor of apoptosis protein induces Gene Expr. 9, 195–201 (2001). Oncology 56, 73–82 (1999).

REVIEWS

149. Chung, T. K. et al. Expression of apoptotic regulators and 157. Adjei, A. A. et al. A phase I trial of ISIS 2503, an antisense 164. Stevenson, J. P. et al. Phase I clinical/pharmacokinetic and their significance in cervical cancer. Cancer Lett. 180, 63–68 inhibitor of H-ras, in combination with gemcitabine in pharmacodynamic trial of the c-raf-1 antisense (2002). patients with advanced cancer. Clin. Cancer Res. 9, oligonucleotide ISIS 5132 (CGP 69846A). J. Clin. Oncol. 150. Cho-Vega, J. H. et al. MCL-1 expression in B-cell non- 115–123 (2003). 17, 2227–2236 (1999). Hodgkin’s lymphomas. Hum. Pathol. 35, 1095–1100 158. Davis, A. J. et al. Phase I and pharmacologic study of the 165. Bishop, M. R. et al. Phase I trial of an antisense (2004) human DNA methyltransferase antisense oligonucleotide OL(1)p53 in hematologic malignancies. 151. Heere-Ress, E. et al. Bcl-X(L) is a chemoresistance oligodeoxynucleotide MG98 given as a 21-day continuous J. Clin. Oncol. 14, 1320–1326 (1996). factor in human melanoma cells that can be inhibited infusion every 4 weeks. Invest. New Drugs 21, 85–97 166. Eder, I. E. et al. Inhibition of LNCaP prostate tumor growth in by antisense therapy. Int. J. Cancer 99, 29–34 (2003). vivo by an antisense oligonucleotide directed against the human (2002). 159. de Fabritiis, P. et al. BCR–ABL antisense androgen receptor. Cancer Gene Ther. 9, 117–125 (2002). 152. Thallinger, C. et al. Mcl-1 is a novel therapeutic target for oligodeoxynucleotide in vitro purging and autologous bone human sarcoma: synergistic inhibition of human sarcoma marrow transplantation for patients with chronic Competing interests statement xenotransplants by a combination of mcl-1 antisense myelogenous leukemia in advanced phase. Blood 91, The authors declare no competing financial interests. oligonucleotides with low-dose cyclophosphamide. Clin. 3156–3162 (1998). Cancer Res. 10, 4185–4191 (2004). 160. Nylandsted, J. et al. Selective depletion of heat shock

153. Wacheck, V. et al. Bcl-xL antisense oligonucleotides protein 70 (Hsp70) activates a tumor-specific death Online links radiosensitise colon cancer cells. Br. J. Cancer 89, program that is independent of caspases and bypasses 1352–1357 (2003). Bcl-2. Proc. Natl Acad. Sci. USA 97, 7871–7876 DATABASES 154. Sigalas, I., Calvert, A. H., Anderson, J. J., Neal, D. E. & (2000). The following terms in this article are linked online to: Lunec, J. Alternatively spliced mdm2 transcripts with loss 161. Ratajczak, M. Z. et al. Acute- and chronic-phase chronic Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. of p53 binding domain sequences: transforming ability and myelogenous leukemia colony-forming units are highly fcgi?db=gene frequent detection in human cancer. Nature Med. 2, sensitive to the growth inhibitory effects of c-myb BCL2 | BCL-XL | caspase-3 | CLU | HSP27 | PKCα | STAT3 | 912–917 (1996). antisense oligodeoxynucleotides. Blood 79, 1956–1961 XIAP 155. Wang, H., Yu, D., Agrawal, S. & Zhang, R. Experimental (1992). National Cancer Institute: http://cancer.gov/ therapy of human prostate cancer by inhibiting MDM2 162. Tortora, G. et al. Oral antisense that targets protein kinase chronic lymphocytic leukaemia | melanoma | multiple myeloma | expression with novel mixed-backbone antisense A cooperates with taxol and inhibits tumor growth, non-Hodgkin’s lymphoma | non-small-cell lung cancer | oligonucleotides: in vitro and in vivo activities and angiogenesis, and growth factor production. Clin. Cancer prostate cancer mechanisms. Prostate 54, 194–205 (2003). Res. 6, 2506–2512 (2000). 156. Zhang, Z., Li, M., Wang, H., Agrawal, S. & Zhang, R. 163. Monia, B. P., Johnston, J. F., Geiger, T., Muller, M. & FURTHER INFORMATION Antisense therapy targeting MDM2 oncogene in prostate Fabbro, D. Antitumor activity of a osphorothioate Genta: www.genta.com/genta/home1.html cancer: effects on proliferation, apoptosis, multiple gene antisense oligodeoxynucleotide targeted against ISIS Pharmaceuticals: www.isispharm.com/index.html expression, and chemotherapy. Proc. Natl Acad. Sci. USA C-raf kinase. Nature Med. 2, 668–675 OncoGenex Technologies: www.oncogenex.ca 100, 11636–11641 (2003). (1996). Access to this interactive links box is free online