Over recent years the fields of antiangiogenesis and vascular targeting have shown a rapid expansion, with many agents entering clinical trial.1, 2 Two main classes of drugs have been identified as having antivascular activity, that is, they cause damage to existing tumour blood vessels, which leads subsequently to secondary cell death and tumour necrosis. The first of these are compounds related to flavone acetic acid, including 5,6-dimethylxanthenone-4-acetic acid (DMXAA), which are believed to cause vascular injury via the production of tumour necrosis factor α (TNFα).3, 4, 5 The second group of drugs are the tubulin-binding agents, of which combretastatin A4 disodium phosphate (CA4P) may be considered the lead compound with antivascular activity. In the Phase I clinical trials completed recently, magnetic resonance imaging (MRI) techniques demonstrated changes consistent with tumour blood flow reduction.6, 7 CA4P has direct effects on the endothelial cytoskeleton, causing a change in cell shape,8, 9, 10, 11 which is associated with a rapid increase in the permeability of endothelial cell monolayers.11 These effects are likely to be important in the early blood flow shutdown measured in vivo.12
In preclinical studies, single non-toxic doses of CA4P have been shown to produce a rapid reduction in blood flow followed by large-scale necrosis in a range of transplantable rodent tumours and human tumour xenografts,8, 9, 13, 14 although tumour sensitivity is variable.15, 16, 17 Even in the most sensitive of tumours, the 2–3 logs of cell-kill that can be achieved by a single dose of drug, is insufficient to have a major effect on tumour growth and surviving cells at the tumour periphery rapidly repopulate the tumour mass.18
To achieve an effective tumour therapy it is likely that combretastatin will need to be combined with conventional therapy, which will target those cells that survive the drug treatment. The viable tumour rim would be expected to be well vascularised and oxygenated as well as accessible to other drugs. Several studies have already demonstrated an enhanced tumour response when CA4P is combined with either radiotherapy13, 16, 18, 19 or chemotherapy, using cisplatin, Taxol, doxorubicin and 5-FU.18, 20, 21 Both fractionated radiation treatments and radioimmunotherapy have been combined successfully with CA4P to produce curative treatments.18, 22 Little attention has been paid to optimising the combretastatin treatment, however; experimental studies have tended to concentrate on the effects of single drug doses. We have compared the therapeutic effectiveness of single and repeated dose schedules of CA4P in a transplanted murine mammary tumour, the CaNT.
The relevance of long established transplanted murine tumour lines to primary human cancers has been raised frequently, as has the use of the subcutaneous implantation site. It is important to demonstrate the efficacy of any new treatment regime in a range of different tumour models. Human tumour xenografts are now employed commonly, as is orthotopic implantation. Spontaneously arising tumours that show more heterogeneous characteristics in terms of growth and treatment response should resemble human cancers more closely than other models. Comparative studies have also been carried out using a spontaneous breast carcinoma model.
MATERIAL AND METHODS
Tumours and drug treatments
The murine adenocarcinoma NT (CaNT), was implanted subcutaneously onto the rear dorsum of 10–16 week-old female CBA/Gy f TO mice, by injecting 0.05 ml of a crude cell suspension prepared by mechanical dissociation of an excised tumour from a donor animal. Tumours were selected for treatment when the geometric mean diameter (GMD) reached 5–6.5 mm (150–300 mg), approximately 3–4 weeks after implantation.
The T138 mouse strain has an inherited predisposition to develop mammary tumours. Between 5–18 months of age, >90% of females develop tumours, which are predominantly well-differentiated glandular or acinar adenocarcinomas. Adenosquamous carcinomas and undifferentiated tumours can also develop occasionally. Tumours show a range of growth rates (4.5–20 days volume doubling time at approximately 6 mm GMD; mean = 10.2 ± 0.9 days). As for the transplanted model, tumours were selected at a size of 5–6.5 mm GMD. All animal experiments were carried out in full compliance with government regulations and UKCCCR guidelines on animal welfare and were approved by the local Ethical Review Committee. Disodium combretastatin A4 3-O-phosphate (CA4P) (Oxigene Inc., Watertown, MA) was dissolved in 0.9% saline at appropriate concentrations to allow each dose to be injected i.p. in 0.1 ml/10 g body weight.
Tumour growth delay
After treatment tumours were measured in 3 orthogonal diameters 2 or 3 times a week and mean growth curves were produced for each group of animals. For experiments involving the CaNT, dose groups comprised 5–6 animals, whereas for the T138, which shows a more variable response to treatment, 10–15 animals were used.
Excision assay and tumour cell growth in vitro
In vivo cell survival was assessed by an in vitro colony-forming assay, as described previously.23 Tumours were excised 18–24 hr after CA4P injection. Two tumours were combined for each data point, weighed, minced with scissors and then disaggregated for 1 hr at 37°C in an enzyme cocktail of 1 mg/ml protease, 0.5 mg/ml DNAase and 0.5 mg/ml collagenase (Sigma-Aldrich Company Ltd., Dorset, UK). After digestion, samples were passed through a 19G needle and a 35 μm filter to obtain a single cell suspension. Haemocytometer counts of trypan-blue excluding cells were made and known numbers of viable cells added to a feeder layer of heavily irradiated V79-379A Chinese hamster cells.
After 7–10 days incubation, cells were fixed, stained with methylene blue and macroscopic colonies were counted. The data were calculated as surviving fraction per gram of tumour, which is a product of relative surviving fraction and relative cell yield per gram of tumour. Each experimental group was repeated at least 3 times so that at least 6 tumours contribute to each data point.
Perfused vascular volume
Functional vascular volume was assessed in control tumours and at selected times after drug treatment using Hoechst 33342 (Sigma-Aldrich Company Ltd., Dorset, UK), as reported previously.24 The fluorescent DNA-binding dye was dissolved in 0.9% saline at 6.25 mg/ml and injected i.v. at a dose of 10 mg/kg. Tumours were excised and frozen 1 min later. Sections were cut at 3 levels and observed under UV illumination. Perfused vessels were identified by their fluorescent outlines and vascular volumes were determined using a random point scoring system based on that described by Chalkley.25 At least 100 fields were scored at each of the 3 tumour levels and the results for treated tumours were expressed as a percentage of control values.
Statistical analysis of data
Cell survival and vascular volume data are expressed as mean values ±1 SEM for groups of 3–10 tumours. Statistical comparisons were carried out using Student's t-test; p < 0.05 was considered significant.
We have demonstrated previously extensive tumour necrosis after a single dose of CA4P and a dose-dependent reduction in tumour cell survival when CaNT tumours are excised 18–24 hours after treatment.18 We compared the effect on clonogenic cell survival of 200 mg/kg CA4P given either as a single i.p. injection or as 2 smaller doses separated in time by up to 6 hr (Fig. 1). Two doses of 100 mg/kg, separated by 1–6 hr, produced more cell killing than the single dose. The effect was most pronounced when the interval between doses was 2–4 hr (p < 0.05). Unequal doses, where either dose was less than 100 mg/kg, were less effective.
When this treatment regime was applied to a growth delay assay, splitting the dose again enhanced the tumour response (Fig. 2). This was apparent in the CaNT when 200 mg/kg CA4P was compared to 2 doses of 100 mg/kg separated by 3 hr and also in the spontaneous T138 tumours. In the latter, more resistant tumour, higher drug doses were employed and 2 doses of 250 mg/kg CA4P produced a measurable response whereas a single dose of 500 mg/kg had no effect on tumour growth.
The effect of dose splitting on the tumour vascular response was also investigated by measuring the perfused vascular volumes of tumours 24 hr after treatment (Fig. 3). The drug doses chosen were lower than those used for the growth delay experiments, to limit the vascular damage induced by the single doses. In both tumour types, giving 2 drug doses separated by 3 hr reduced the fraction of perfused vessels by more than a factor of 2, compared to the single dose treatment (p < 0.05).
Whereas large single doses of drug have produced minimal effects on tumour growth, repeated daily or twice daily dosing resulted in substantial growth delays (Fig. 4). In the CaNT, 500 mg/kg CA4P administered as 2 doses of 25 mg/kg per day (6-hr interval), 5 days a week for 2 weeks, was significantly more effective at slowing tumour growth than either a single dose or 10 daily doses of 50 mg/kg, also delivered over 2 weeks. In the T138 the same dosing regimes were employed, however, in this tumour model both repeated dosing schedules were equally effective. No additional benefit was achieved by giving 2 doses per day. Previous studies have shown that twice-daily injections of saline for 2 weeks have no measurable effect on tumor growth (Hill, unpublished data). Toxicity, as measured by weight loss, was greater for the twice-daily treatments, with a loss of 7 and 6% body weight measured for the CaNT and T138 animals respectively, at the end of 2 weeks of dosing. This compared to a 3% weight loss for the daily dosing group of T138s and a 3% weight gain for the untreated animals.
To gain further insight into the effects of single and multiple dose schedules, we have also examined the histology of CaNT tumours as a function of time after dosing. Figure 5a–d shows the histological appearance of CaNT tumours 1, 3 and 4 days after a single dose of 100 mg/kg CA4P, demonstrating early necrosis followed by rapid regrowth. Figure 5e,f shows the histology of the CaNT 5 days into, and at the end of, a multiple dose schedule. By the end of treatment regrowth is visible from the viable rim, although the centre of the tumour remains necrotic.
The data presented demonstrate that it is possible to achieve a significant growth delay using CA4P as a single agent when the drug is given in a repeated dose schedule. This is true both for the CA4P-sensitive CaNT and for the more variably-responsive T138 spontaneous tumour. The more sensitive cell survival assay showed that splitting the dose into 2 equal fractions separated by 2–4 hr increased tumour cell kill by at least a factor of 2 in the CaNT.
The mechanism behind the increased effectiveness of the repeated dose schedules is not clear. The vascular volume data, however, indicated a greater than 2-fold reduction in the fraction of perfused tumour vessels 24 hr after a split-dose treatment in both tumour types (Fig. 3), suggesting a greater degree of direct damage to the tumour vasculature. Previous studies18 have shown that the greatest reduction in functional vascular volume is measured between 3–6 hr after a single dose of 100 mg/kg CA4P to the CaNT, with some evidence of recovery by 24 hr. Although full time-course studies have not been carried out at other doses in this tumour model, a similar time-dependent reduction in blood flow has been recorded in other rodent tumours using a variety of drug doses and different methods of measurement.14, 26 Thus, it might be assumed that blood flow does not change greatly between 3–6 hr after treatment, the intervals used for the vascular volume and growth studies presented. In the CaNT, 25 mg/kg CA4P reduces the perfused vascular volume to approximately 65% of control values at 6 hr, with almost complete recovery (to 80–90%) by 24 hr (Hill, unpublished data). A second drug treatment delivered 3–6 hr after the first would therefore appear to have a proportionally greater effect and prevent blood flow recovery (vascular volume reduced to 20% of control at 24 hr). Although the mechanism responsible for this effect has not been investigated it could involve “self-trapping” of drug within the tumour tissue, thereby increasing drug exposure. In addition, CA4P treatment is known to reduce NO production in tumour cells several hours after drug administration (Parkins, unpublished observations). Because NOS inhibition can enhance the action of vascular damaging agents such as PDT and CA4P,15, 27 such an effect could also contribute to the enhancement of activity of a second dose of the drug.
As well as causing direct vascular damage, CA4P has also been shown to have antiproliferative and cytotoxic effects in both endothelial and tumour cells in culture.8, 10, 28, 29 The published in-vitro data would suggest that such effects only occur at higher levels of drug exposure than those achieved after a single dose of 25 mg/kg in the mouse, which is approximately equivalent to the maximum tolerated dose achieved clinically.8, 28, 30 Our own in vivo data would support this conclusion for direct tumour cell cytotoxicity.18 We reported previously that at doses below 100 mg/kg, the cell killing achieved by CA4P was primarily manifest as a reduction in the number of cells recovered from a treated CaNT tumour; the cells had lysed between treatment and tumour excision 18 hr later. This early cell death is characteristic of a vascular-mediated response. At higher drug doses a greater effect was apparent on the survival of the recovered cells, reflecting a directly cytotoxic effect toward the tumour cells. In repeated dosing schedules however, direct cytotoxicity and anti-proliferative effects on both tumour and endothelial cells may become more important, because the area-under-the-curve exposure is likely to be increased.12
Vinblastine, another tubulin-binding agent that shows antivascular activity at high drug doses,31, 32, 33 has been shown recently to be antiangiogenic when continuous very low, non-cytotoxic doses are employed either in vitro or in vivo.34, 35 A number of other conventional chemotherapeutic agents have also been reported to have antiangiogenic activity when given in repeated daily or even weekly low-dose schedules.36, 37, 38, 39, 40 It seems likely therefore that by giving repeated doses of CA4P, as well as the direct vascular damaging/blood flow modifying effects suggested by the vascular volume data presented for the split dose treatment, an additional anti-proliferative effect may be exerted on the surviving endothelial and tumour cells at the tumour periphery. This would serve to inhibit the neovascularisation that is necessary for tumour repopulation and regrowth. Histologically, expansion of the peripheral rim of tumour cells is apparent by 72 hr after a single dose of CA4P in the CaNT (Fig. 5). Using a schedule of repeated doses, tumours appear almost entirely necrotic throughout the treatment period, despite continuing to increase in size (Fig. 5e,f).
The increased efficacy of the twice-daily dosing over daily treatments in the CaNT may reflect a greater component of direct damage to the blood vessels. No such difference in efficacy was noted in the T138 tumour system however; both schedules were equally effective. This may relate to the greater resistance to CA4P-induced vascular damage measured in this model (Fig. 3), such that the dose of 25 mg/kg twice daily was too low to induce significant blood-flow modifying effects. Although these spontaneous tumours are more heterogeneous in their growth and response to treatment than the transplanted tumours, they are predominantly well-vascularised and slow growing and appear to contain more mature vessels that stain positively for α-smooth muscle actin (Hill, unpublished data). This may account for their relative resistance to CA4P. If these tumours are indeed more representative of the response to human tumours, it suggests that a daily low-dosing schedule may still offer greater efficacy than widely separated larger doses.
A significant growth response has also been reported for 2 human non-small cell lung cancer xenografts after daily injections of 50 mg/kg CA4P continued for 21 days.41 This response was interpreted as being predominantly induced by antivascular rather than anti-tumour cell effects, on the basis that microvessel density (MVD) was similar in treated and control tumours. It was argued that the preferential death of tumour cells might be expected to result in an increase in the MVD. These data are consistent with a predominantly antiproliferative effect on the tumour endothelium.
Preclinical data published recently indicate that the maximum tolerated dose (MTD) of CA4P was unchanged whether given as a single dose or as 2, 4 or 8 equal doses.42 In our study, the greatest growth delay was measured after 4 doses of 200 mg/kg. Although no therapeutic data are available, Phase I clinical studies also indicate a similar MTD for 5-daily doses repeated every 21 days as for a single weekly dose continued for 3 weeks (Chaplin, personal communication).30 In both studies magnetic resonance imaging showed evidence of a decrease in tumour blood flow.6, 7
In summary, CA4P has a significant effect on the growth of both a transplanted and a spontaneous murine tumour model when delivered in a frequent low-dose treatment schedule, whereas a single large dose has no measurable effect. It is suggested that both the blood flow modifying and antiproliferative effects of CA4P may be involved in the response to twice-daily dosing, whereas daily dosing may produce an enhanced antiproliferative effect on both tumour and endothelial cell compartments. The data suggest that to maximise antitumour activity, CA4P should be given in a frequent-dose treatment schedule, rather than as a small number of large doses that allow early neovascularisation and tumour repopulation between doses.
D.J.C. has declared a financial interest in the company whose product (CA4P) was studied in the work presented in this article.