Oral Administration of a Novel Taxane, an Antisense Oligonucleotide Targeting Protein Kinase A, and the Epidermal Growth Factor Receptor Inhibitor Iressa Causes Cooperative Antitumor and Antiangiogenic Activity¹

A novel approach to cancer therapy is the integration of new cytotoxic agents with multiple selective inhibitors of relevant signaling pathways. Such a strategy indicates that it may be possible to use agents that cooperate at low doses and/or through simple administration modalities to increase tumor targeting and patient compliance and reduce toxicity. Novel inhibitors of cAMP-dependent protein kinase (PKA)³ and EGFR may satisfy these requirements.

PKA, which plays a key role in the control of cell growth and differentiation of mammalian cells, exists in two distinct isoforms, type I (PKAI) and type II (PKAII), which differ only in their regulatory subunits (RI and RII, respectively; Ref. 1). PKAI is mechanistically involved in cell proliferation and neoplastic transformation; is required for G₁-S cell cycle transition and for the transduction of mitogenic signals from different growth factors, including TGFα and EGF; and is overexpressed in most human cancers, correlating with worse clinicopathological features in several tumor types (reviewed in Ref. 2). We have demonstrated that PKAI, through its RIα subunit, structural interacts with the ligand-activated EGFR and cooperates in the propagation of the mitogenic signals originated by different growth factors and hormones (2, 3). For these reasons, PKAI is considered a relevant target for therapeutic intervention (2). Down-regulation of PKAI by different pharmacological tools, including antisense oligodeoxynucleotides targeting its RIα subunit (AS-PKAI) causes growth arrest and differentiation in a variety of cancer cell lines in vitro and in nude mice (1, 2, 4, 5). An improvement of these effects has recently been obtained with a novel class of MBOs, which exhibit better pharmacokinetic and toxicological properties in vivo (6). We have shown that the AS-PKAI MBOs have antitumor effects in different human cancer xenografts, synergizing with different classes of cytotoxic drugs (2, 7). A Phase I clinical trial in cancer patients with negligible toxicity has recently been completed with a hybrid DNA/RNA MBO AS-PKAI containing 2-O-methyl modifications (GEM 231; Ref. 8). We and others have recently shown that this novel AS-PKAI also exhibits a good pharmacokinetic profile and inhibits tumor growth after oral administration, potentiating Taxol activity (9, 10).

EGFR is a main transducer of mitogenic signals involved in cancer pathogenesis and progression and is considered an important target for anticancer therapy (11). Several compounds that block ligand-induced EGFR activation have been developed. ZD1839 (Iressa) is a p.o. active quinazoline derivative that selectively inhibits the EGFR tyrosine kinase and is under clinical development in cancer patients. We and others have shown that ZD1839 inhibits the growth of a variety of human cancer cell lines in vitro, achieving a supra-additive effect when combined with different cytotoxic drugs and inducing apoptotic cell death (12, 13). Moreover, ZD1839 shows antitumor activity and prolongs survival in vivo in nude mice when given in combination with different drugs (12, 13).

On the basis of the molecular interactions between PKAI and EGFR, we have developed a therapeutic strategy that provides a combined blockade of these two signaling pathways by different selective inhibitors; we have demonstrated the ability of these inhibitors to target the respective kinases both in vitro and in vivo in nude mice models (2, 10, 14). Moreover, we have shown the ability of both PKAI and EGFR inhibitors to potentiate the cytotoxic activity of drugs (2, 13). In particular, we have reported a marked cooperative effect of different PKAI inhibitors with taxanes (2, 7, 10). Additional evidence of the possible interaction among PKAI, EGFR, and targets affected by taxanes is provided by our recent demonstration that a combination of AS-PKAI with the anti-EGFR MAb C225 and docetaxel cooperatively inhibited the in vitro growth of a human breast cancer cell line (15), providing a potential basis for the development of a novel therapeutic strategy.

On this basis, in the present study we evaluated whether a cooperative effect could be obtained by combining the hybrid MBO AS-PKAI and ZD1839, which are both p.o. active, with IDN5109 [13-(N-boc-β-isobutylisoserinyl)-14-hydroxybaccatin-1,14-carbonate], a novel paclitaxel analogue from a new series of 14-OH-DAB derivatives, which has shown potent antitumor activity against a variety of human cancer types, including multidrug-resistant cells (16, 17), as well as after oral administration (18, 19).

We describe the effects of these drugs on tumor growth, angiogenesis, and apoptosis of human cancer, in vitro and in vivo, following oral administration.

Clinical-grade ZD1839 and IDN5109 were provided by AstraZeneca (Macclesfield, United Kingdom) and Indena S.p.A. (Milan, Italy), respectively. MBOs were kindly provided by Dr. Sudhir Agrawal (Hybridon, Inc., Cambridge, MA). The AS-PKAI is a hybrid oligonucleotide, containing 2′-O-methyl-ribonucleosides, targeted against NH₂-terminal codons 8–13 of the RIα regulatory subunit of PKA (5). The control was a scrambled MBO obtained by mixing all four nucleosides at each position. The sequence, structure, and purification methods were as published previously (10, 20).

GEO colon, OVCAR-3 ovarian, and ZR-75-1 breast human cancer cell lines were obtained from the American Type Culture Collection (Rockville, MD). GEO, OVCAR-3 and ZR-75-1 cells were maintained in DMEM supplemented with 10% heat-inactivated fetal bovine serum, 20 mm HEPES (pH 7.4), 100 IU/ml penicillin, 100 g/ml streptomycin, and 4 mm glutamine (ICN, Irvine, United Kingdom) in a humidified 95% air-5% CO₂ atmosphere at 37°C.

On day 0, cells (10⁴ cells/well) were suspended in 0.5 ml of 0.3% Difco Noble agar (Difco, Detroit, MI) supplemented with complete culture medium. This suspension was layered over 0.5 ml of 0.8% agar-medium base layer in 24-well cluster dishes (Becton Dickinson, Lincoln Park, NJ) and treated with the following concentrations of drugs: 0.1 and 0.5 μm AS-PKAI (on days 0, 1, and 2); 0.5, 1, and 5 ng/ml IDN5109 (on day 0); and 0.05, 0.1 and 0.5 μm ZD1839 (on days 1, 2, and 3). After 10–14 days, cells were stained with nitroblue tetrazolium (Sigma), and colonies larger than 0.05 mm were counted as described previously (4).

The induction of apoptosis was determined by the Cell Death Detection ELISA Plus Kit (Roche Molecular Biochemicals, Mannheim, Germany), which detects cytosolic histone-associated DNA fragments. Briefly, GEO, OVCAR-3 or ZR-75-1 cells (5 × 10⁴ cells/dish) were seeded into 35-mm dishes. After treatment with 0.1 μm AS-PKAI (days 0, 1, and 2) and/or 5 ng/ml IDN5109 (day 0) or 0.05 μm ZD1839 (days 1, 2, and 3), on day 5 cells were washed once with PBS, and 0.5 ml of lysis buffer were added. After a 30-min incubation, the supernatant was recovered and assayed for DNA fragments as recommended by the manufacturer. Each treatment was performed in quadruplicate. The total number of cells was measured with an hemocytometer in additional plates receiving an identical treatment. The absorbance values at 405 nm were normalized for cell number, and the ratio of the absorbance of treated cells to the absorbance of untreated cells was considered the apoptotic index, expressed in arbitrary units.

Total cell lysates were obtained from homogenized GEO tumor specimens. The protein extracts were resolved by 7.5% SDS-PAGE and probed with an antihuman VEGF MAb (Santa Cruz Biotechnology, Santa Cruz, CA); immunoreactive proteins were visualized by enhanced chemiluminescence (Amersham International, United Kingdom), as described previously (3).

The concentration of VEGF in the CM obtained from GEO cells was measured using commercially available sandwich ELISA kits and according to manufacturers’ instructions. The ELISA kit for VEGF was purchased from R&D Systems Inc. (Minneapolis, MN). GEO cells were plated in 60-mm dishes (Becton Dickinson) and treated for 4 days with 0.1 μm AS-PKAI, 5 ng/ml IDN5109, and 0.1 μm ZD1839 alone and in combination. Assays were performed using 24-h collected serum-free CM.

Female Balb/cAnNCrlBR athymic (nu+/nu+) mice (5–6 weeks old) were purchased from Charles River Laboratories (Milan, Italy). The research protocol was approved, and mice were maintained in accordance with institutional guidelines of the University of Naples Animal Care and Use Committee. Mice were acclimatized to the University of Naples Medical School Animal Facility for 1 week prior to receiving injections of cancer cells. GEO human colon cancer cells (10⁷) were resuspended in 200 liters of Matrigel (Collaborative Biomedical Products, Bedford, MA) and injected s.c. in mice. After 7 days, when well-established tumors of ∼0.2 cm³ were detected, mice were randomized to receive different treatments. Groups of 10 mice were treated with p.o. AS-PKAI and/or ZD1839 daily on days 7–11, 14–18, 21–25, and 28–32 at doses of 10 and 150 mg/kg, respectively, and/or with 60 mg/kg IDN5109 every 4th day, on days 7, 11, 14, and 18. Tumor volume was measured using the formula π/6 × larger diameter × (smaller diameter)², as reported previously (14). Two mice were sacrificed at day 21 for biochemical and immunohistochemical analyses.

New blood vessels were detected as described previously (21, 22), with a MAb against the human factor VIII-related antigen (Dako, Milan, Italy) at 1:50 dilution and stained with a standard immunoperoxidase method (Vectastain ABC kit; Vector). Each slide was first scanned at low power (magnification, ×10–100), and the area with the highest number of new vessels was identified (hot spot) and then scanned at ×250 magnification (0.37 mm²). Stained blood vessels were counted in each of five different fields. For individual tumors, microvessel count was scored by averaging the counts for the five fields. All analyses were performed in a blinded fashion.

We evaluated the antitumor activity in vitro of the novel taxane IDN5109 on the soft agar growth of a variety of human cell lines, including GEO colon, OVCAR-3 ovarian, and ZR-75-1 breast cancer cells. As shown in Fig. 1, IDN5109 caused dose-dependent growth inhibition of all tested cells, with an IC₅₀ demonstrating higher sensitivity for ZR-75-1 and OVCAR-3 compared with GEO colon cells.

When we combined the AS-PKAI with different doses of IDN5109, a supra-additive effect was observed in all cell lines, particularly with lower doses of AS-PKAI (Fig. 2). In contrast, no cooperative effect was observed when the IDN5109 was combined with an oligonucleotide with a scrambled sequence.

We previously demonstrated the growth-inhibitory activities of AS-PKAI and ZD1839 as single agents on GEO, OVCAR-3, and ZR-75-1 cells (7, 13). We evaluated the effect of combinations of different doses of the two agents on the soft agar growth of these human cancer cells. Fig. 3 shows a clearly supra-additive inhibitory effect on GEO, whereas a cooperative effect with low doses and a mostly additive effect with higher doses were observed for OVCAR-3 and ZR-75-1 cells, probably because ZD1839 is extremely effective as a single agent in ovarian and breast cells. Unlike the AS-PKAI, the combination of a scrambled-sequence oligonucleotide with ZD1839 caused, at most, an additive effect in all cell lines. We then combined different doses of the taxane IDN5109 with ZD1839. A clearly cooperative growth-inhibitory effect was observed at all tested doses in the three cell lines (Fig. 4).

On the basis of our former hypothesis of potential cooperativity between the three drugs (18), we combined the three agents together, obtaining a dramatic synergistic inhibitory effect on the soft agar growth of GEO, OVCAR-3, and ZR-75-1 cancer cells. In fact, low doses that alone induced up to 10–15% inhibition of colony formation caused almost complete suppression of growth (Fig. 5).

We studied the effect of the combination of these agents on the induction of apoptosis in vitro. We used suboptimal doses of AS-PKAI and ZD1839 (unable to induce apoptosis) and a dose of IDN5109 that caused up to 3-fold induction of apoptosis compared with untreated cells. When any two drugs were combined, a mostly additive effect was observed (Fig. 6). A clearly supra-additive effect was achieved in all cell lines tested when the three drugs were combined (Fig. 6), in agreement with the marked antiproliferative effect observed with the same combination.

The endogenous levels of the angiogenic growth factor VEGF were measured in the CM of GEO cells. Untreated GEO cells secreted ∼150 pg VEGF/10⁶ cells. Treatment with either AS-PKAI or ZD1839 alone caused ∼25% inhibition of VEGF secretion, whereas IDN5109 was ineffective. Combination of AS-PKAI with either ZD1839 or IDN5109 caused ∼50 and 30% inhibition of VEGF secretion, respectively, whereas IDN5109 and ZD1839 together caused ∼40% inhibition. A cooperative inhibitory effect of almost 75% was observed when the three agents were used together, reducing the VEGF secretion to 36 pg/10⁶ cells.

We investigated the antitumor activity of AS-PKAI, ZD1839, and IDN5109 administered p.o. alone and in combination in nude mice bearing GEO colon cancer xenografts. When established GEO tumors of ∼0.2 cm³ were detectable, groups of 10 mice were treated with AS-PKAI, IDN5109, and ZD1839 alone and in combination. Two mice were sacrificed on day 21 for biochemical and histochemical analyses; tumor growth studies were therefore performed on the remaining eight mice. As shown in Fig. 7, within ∼5 weeks, GEO tumors reached a size not compatible with normal life in all untreated mice, and treatment with the IDN5109 had only a small inhibitory effect. On the other hand, AS-PKAI and ZD1839 caused a similar inhibition and delay of tumor growth. However, shortly after the end of treatment, these tumors resumed a growth rate similar to that in untreated mice. When we treated the animals with two agents in combination, a remarkably similar effect was obtained with any two drugs tested, causing a significant inhibition of growth. In fact, treated tumors grew very slowly for almost 3 weeks after withdrawal of treatment; they then resumed a faster growth rate (Fig. 7). As shown in Table 1, at this time point, no more than 25% of animals were tumor free. At the end of the experiment, 10 weeks after tumor cell injection, all animals were still alive in all groups treated with two drugs in combination, but only one of the eight mice in each of the groups treated with the AS-PKAI in combination with either IDN5109 or ZD1839 was tumor free after pathology examination. A dramatic and sustained inhibitory effect was obtained by the combination of three drugs together. In fact, no tumor growth was observed for 3 weeks after treatment withdrawal, and a modest increase in tumor size was recorded at the end of the experiment, 10 weeks after tumor cell injection (Fig. 7). Moreover, at this time point, pathological evaluation showed that five of eight mice were still tumor free (Table 1). The combined treatment with any two drugs or with all three drugs together was well tolerated; no weight loss or other signs of acute or delayed toxicity were observed.

In the present study, we quantified by immunohistochemistry the tumor-induced vascularization as microvessel count in the most intense areas of neovascularization, using an anti-factor VIII-related antigen MAb (21). At day 21, two tumor samples were analyzed for each group of mice. As reported in Fig. 7, their volumes were similar to that measured at the same time point in the remaining eight mice. Each single drug caused a significant inhibition of vessel formation compared with samples from mice untreated or treated with the scrambled MBO. As shown by staining, vessel formation in GEO tumors was further inhibited by treatment with any two drugs in combination and completely suppressed with the three drugs in combination (Fig. 8,A). In the same tumor samples, we examined VEGF expression by Western blot (Fig. 8,B). We observed a marked inhibition of VEGF expression in animals treated with ZD1839, whereas AS-PKAI or IDN5109 had only a moderate inhibitory effect. Approximately 70 and 80% inhibition was obtained when the AS-PKAI was combined with either IDN5109 or ZD1839, respectively, whereas IDN5109 and ZD1839 in combination caused almost 90% inhibition. Complete suppression of VEGF expression was observed in the specimens removed from the animals treated with the three agents in combination (Fig. 8 B).

In recent years, several molecular targets relevant for cancer cell growth have been identified, and consequently, different selective inhibitors have been developed. The tyrosine kinase growth factors receptors, such as EGFR, and the protein kinases that transduce intracellular signaling, such as PKA, play an important and complex role in the control of neoplastic growth, apoptosis, and angiogenesis (2). We have previously demonstrated that PKAI and EGFR have structural and functional interaction and that a combined blockade of these two pathways may have therapeutic implications (2, 3, 14).

We have also demonstrated a cooperative effect of taxanes with PKAI and EGFR inhibitors, implying the possibility that these cytotoxic drugs may affect related molecular targets (2, 7, 13, 15).

To translate this hypothesis in an experimental setting, in this study we used three novel agents with peculiar properties, including oral activity: a hybrid DNA/RNA MBO, AS-PKAI; the selective EGFR tyrosine kinase inhibitor ZD1839; and the novel taxane IDN5109. All of these agents are under clinical evaluation in cancer patients.

In this study we demonstrated, in different human cancer cell types, that the combined use of AS-PKAI, ZD1839, and IDN5109 has a cooperative growth-inhibitory effect, achieving maximal activity when the three agents are used together, even at very low doses. Moreover, the in vitro antitumor effect is paralleled by increased induction of apoptosis and by inhibition of VEGF secretion by these drugs in combination, supporting the evidence of a role for PKAI and EGFR in the apoptotic and angiogenic pathways.

We translated these results in vivo in nude mice bearing GEO xenografts, using oral administration for all of the agents tested. We used moderately effective doses of AS-PKAI and ZD1839 and a suboptimal dose of IDN5109 to determine any cooperative effect among the drugs used in combination. We showed that any two drugs used in combination have a similar additive effect, supporting the hypothesis of a functional interaction among all of the targets specifically blocked by these agents. A marked cooperative effect was obtained when the three agents were used together. In fact, no tumor growth was observed at the end of the experiment, on day 70 after tumor injection, and several weeks after treatment withdrawal. Moreover, >50% of mice were still tumor free at this time point.

In recent years, a large body of evidence has emphasized the key role of tumor-induced neovascularization in neoplastic development, progression, and metastasis (23). EGFR may provide a major contribution to the development of neoangiogenesis (24, 25). In fact, EGFR ligand TGFα has been identified as a positive regulator of angiogenesis because it is secreted by cancer cells to stimulate normal endothelial cell growth through paracrine mechanisms (26). Moreover, EGF and TGFα can up-regulate the production of VEGF in human cancer cells at the transcriptional level (27, 28). We have previously shown that the anti-EGFR MAb C225 is able to inhibit the expression of TGFα, VEGF, and basic fibroblast growth factor in human GEO colon cancer cells and that it has a marked cooperative effect with an antisense oligonucleotide targeted to human VEGF (14, 29). More recently, we have demonstrated that ZD1839 is able to inhibit the expression of TGFα, basic fibroblast growth factor, and VEGF in a large variety of human cancer cell lines, including GEO, OVCAR-3, and ZR-75-1, and that this effect is independent of the EGFR expression levels (30).

We have shown that PKAI is also involved in angiogenesis because its selective inhibitor 8-Cl-cAMP or oral AS-PKAI inhibits the production of angiogenic growth factors (31). The antiangiogenic properties of taxanes have also been demonstrated (32).

In this study we found a cooperative effect of AS-PKAI, ZD1839, and IDN5109 in combination on the in vitro secretion of VEGF. More importantly, we demonstrated in nude mice that any two drugs used at very low doses in combination reduce the VEGF expression in treated specimens, whereas the three drugs together caused complete suppression. A parallel effect was obtained for the vessel count, suggesting a direct correlation between the two results and supporting the hypothesis of a direct interference of the respective drugs on the angiogenic signaling pathways.

The possibility of a direct interplay of the different drugs on common targets is further supported by an intriguing recent study showing the direct association of PKAI with microtubules and its role in the organization of the mitotic spindle (33), thus providing a potential explanation for the potent interaction of PKAI inhibitors and taxanes.

A recently proposed effective therapeutic strategy, defined metronomic dosing, is based on the administration of continuous low doses of cytotoxic drugs to target tumor angiogenesis and cell proliferation without toxicity (34, 35). Such a strategy implies that the blockade of mitogenic and angiogenic molecules may play a critical role in the treatment and long-term control of cancer. In agreement with this approach, we have here shown that the use of nontoxic doses of different selective agents targeting molecules involved in mitogenic signaling, angiogenesis, and apoptosis can inhibit tumor growth and angiogenesis, thereby inducing apoptosis.

Ryan and Chabner (36) recently proposed the receptor tyrosine kinases as the most rational and promising targets for drug development and that the ultimate value of their selective inhibitors is through integration with cytotoxic drugs and/or radiotherapy. Our results not only confirm the relevance of such a strategy, but also demonstrate that further development and benefit may be provided by agents also active by oral administration. Because all of the agents are under clinical development, we believe that this study provides a strong rationale to translate this feasible therapeutic strategy into a clinical setting.

We acknowledge the excellent technical assistance of Gaetano Borriello.