- Email: [email protected]
Combretastatin A-4 and hyperthermia;a potent combination for the treatment of solid tumors
Radiotherapy and Oncology 60 (2001) 147±154
www.elsevier.com/locate/radonline
Combretastatin A-4 and hyperthermia; a potent combination for the treatment of solid tumors Hans P. Eikesdal a,*, Rolf Bjerkvig b, Olav Mella a, Olav Dahl a a
b
Department of Oncology, University of Bergen, Haukeland University Hospital, 5021 Bergen, Norway Department of Anatomy and Cell Biology, University of Bergen, Aarstadveien 19, 5009 Bergen, Norway Received 2 February 2000; received in revised form 8 August 2000; accepted 30 October 2000
Abstract Background and purpose: Attacking tumor vasculature is a promising approach for the treatment of solid tumors. The tubulin inhibitor combretastatin A-4 disodium phosphate (CA-4) is a new vascular targeting drug which displays a low toxicity pro®le. We wanted to investigate how CA-4 in¯uences tumor perfusion in the BT4An rat glioma and how the vascular targeting properties of CA-4 could be exploited to augment hyperthermic damage towards tumor vasculature. Material and methods: We used the 86RbCl extraction technique to assess how CA-4 in¯uences tumor perfusion, and the tumor endothelium was examined for morphological changes induced by the drug. We combined CA-4 (50 mg/kg i.p.) with hyperthermia (448C, 60 min) at different time intervals to evaluate how therapy should be designed to affect tumor growth, and we studied the tumors histologically to assess tissue viability. Results: We found that CA-4 induced a profound, but transient reduction in tumor perfusion 3±6 h postinjection. If hyperthermia was administered 3±6 h after injecting CA-4, massive hemorrhagic necrosis developed, and tumor response was signi®cantly enhanced compared to simultaneous administration of the two treatment modalities (P , 0:005). CA-4 alone had no in¯uence on tumor growth and failed to disrupt the vasculature of the BT4An solid tumors. Interestingly though, a mild endothelial edema was observed in some tumor areas 3 h after injecting CA-4. Conclusions: We conclude that the combination of CA-4 and hyperthermia is a potent therapeutic option for BT4An tumors, but the selection of adequate time intervals between CA-4 and hyperthermia are imperative to obtain tumor response. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Combretastatin A-4 disodium phosphate; Hyperthermia; Tumor perfusion; Histology; BT4An
1. Introduction The failure of chemotherapy is a common obstacle for the successful treatment of solid tumors. The genetic instability in malignant tumors leads to tumor cell heterogeneity where drug resistance commonly emerges due to a selection of surviving cellular subpopulations. Attempts to block the mechanisms leading to drug resistance, such as by inhibiting the P-glycoprotein, have proven insuf®cient [33], and new strategies have therefore evolved to circumvent acquired drug resistance. As ®rst suggested by Judah Folkman [10], attacking the neovascularization process is an alternative approach for the treatment of solid tumors, and recent reports have disclosed a potential therapeutic effect of anti-angiogenic therapy since proliferating endothelial
* Corresponding author.
cells within the tumors are less capable of developing drug resistance [3,26]. As opposed to inhibiting the outgrowth of new tumor capillaries, therapeutic strategies may also be designed to destroy established tumor blood vessels; i.e. angiotoxic therapy. It has recently been shown that various drugs which inhibit tubulin polymerization, also reduce tumor blood ¯ow, and a new tubulin inhibitor, combretastatin A-4 disodium phosphate (CA-4), demonstrated profound angiotoxic properties in solid tumors [2,5,7]. Hyperthermia is another treatment modality which effectively destructs tumor vasculature, but suf®cient thermal doses are dif®cult to obtain in human tumors due to reactive hyperemia causing thermal washout [25]. Combining hyperthermia in vivo with ¯ow decreasing compounds like sodium nitroprusside (SNOP), have led to encouraging results, increasing tumor temperatures and prolonging tumor growth delay [17,28]. Applying CA-4 before hyperthermia, could presumably mediate simi-
0167-8140/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(00)00318-2
148
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
lar effects as SNOP without the need for systemic hypotension. Additionally, the angiotoxic properties of CA-4 might enhance thermal damage to tumor blood vessels. In the s.c. BT4An rat glioma, we used the 86RbCl extraction technique [29] to assess how CA-4 in¯uences tumor blood ¯ow. We thereafter conducted a tumor response study to evaluate the combination of the drug with hyperthermia at different time intervals. To disclose some of the potential interactions of CA-4 and hyperthermia, we also investigated changes in tumor morphology after treatment. 2. Materials and methods 2.1. Animals and tumor Rats of both sexes of the inbred strain BD IX/HanFoss, mean weight 200 g (SD 87), were used. They were given water and standard pelleted diet ad libitum. Rats were caged at 228C in a night/day cycle 12/12 h. The BT4An tumor is an aggressive glioma which can give rise to lymphatic and haematogenous metastases when grown subcutaneously (s.c.) [23]. For each experiment a single tumor-bearing rat was sacri®ced and tumor pieces measuring 1±2 mm were transplanted to the dorsum of the right hind leg through an incision 5 mm proximal to the tumor bed. The animals in the tumor response study were treated 9±11 days after transplantation at a mean tumor volume of 188 mm 3 (SD 27). For the tumor perfusion study animals with a mean tumor volume of 143 mm 3 (SD 69) were used, and in the tumor thermometry study animals had a mean tumor volume of 213 mm 3 (SD 64). 2.2. Anesthesia During interventional procedures and hyperthermia, the animals were kept anesthetized with s.c. injections of fentanyl (0.05 mg/kg)/¯uanisone (2.5 mg/kg)/midazolam (1.25 mg/kg) (Hypnorm Dormicum) [9]. Normal body temperature in anesthetized animals that did not receive hyperthermia was maintained by the use of a thermostatically controlled incubator and a heating pad (Harvard Apparatus 50-7061, Kent, England). 2.3. CA-4 administration CA-4 was kindly supplied by Oxigene Inc., Lund, Sweden. It was dissolved in 0.9% NaCl to a ®nal concentration of 50 mg/ml. CA-4 was administered i.p. at a maximum tolerated dose of 50 mg/kg [8]. 2.4. Hyperthermia technique The rats were placed in plastic jigs. The tumor-bearing foot was ®xed and immersed 1.5±2 cm below the surface in a circulating, precision controlled waterbath. The waterbath temperature was 44 ^ 0:18C during the tumor response
study, and the treatment time was 60 min. This results in mean tumor temperatures of 41:2±41:88C when hyperthermia is administered alone [8]. To reduce systemic hyperthermia, the rats were cooled by fanning room air through perforations in the jigs [24]. 2.5. Organ perfusion study Organ perfusion was measured at various intervals after drug injection by the 86RbCl extraction technique. Tissue radioactivity 90 s after an i.v. injection of 86RbCl can be used to calculate relative blood ¯ow as a proportion of cardiac output [34]. Brie¯y, after inserting a catheter into the left femoral vein, eight rats per group were injected with a volume of 0.2 ml 86RbCl (Amersham International Biotech UK Limited, Little Chalfont, England) (speci®c activity 0.5±10 mCi/mg and 50 mCi/ml). The rats were sacri®ced 90 s later by an overdose of sodium pentobarbital (100 mg/kg i.v.). Tumor material and organs were excised immediately, weighed and counted for radioactivity (Packard Cobra II Auto Gamma Meriden, CT, USA). The percentage of injected radioactivity/g of tissue was used to calculate relative changes in organ perfusion after CA-4 administration. The selection of an anesthetic regimen is a potential bias when studying tissue perfusion. The anesthetic, Hypnorm Dormicum, employed in the present study lowers the blood pressure and increases the cardiac output and the heart rate in rats, and it has been speculated that the drug reduces total peripheral resistance [17,31]. Such hemodynamic changes might affect the organ perfusion, and to circumvent this problem, we therefore anesthetized all animals included in the organ perfusion study and calculated relative changes in perfusion as a percentage of the control group. 2.6. Tumor response study The rats were strati®ed by tumor size and randomized to one of the following nine treatment groups: controls, hyperthermia alone, CA-4 alone, hyperthermia 3 h before CA-4 (HT3C), CA-4 immediately before hyperthermia (C0HT), CA-4 3, 6, 12 or 24 h before hyperthermia (C3HT/C6HT/C12HT/C24HT). The experiment was run on 3 consecutive days with the same tumor passage. All animals included in the tumor response study were given Hypnorm Dormicum to preclude data skewness due to the anesthetic regimen. 2.7. Normal tissue toxicity During the tumor response study, side effects, including weight changes, were evaluated 48 h after treatment and then twice weekly. 2.8. Tumor evaluation Tumors were measured twice weekly by vernier callipers, and tumor volume calculated using the modi®ed ellipsoid
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
formula: p=6 £ A £ B2 , where A is the longer axis and B is the axis perpendicular to A. The time to reach four times the initial tumor volume was de®ned as tumor growth time (TGT). Partial regression (PR) was de®ned as a greater than 50% volume reduction observed at least twice and lasting for at least 1 week. Complete regression (CR) was de®ned as no visible or palpable sign of tumor observed at least twice. Cure was a complete regression lasting 120 days. 2.9. Tumor morphology Four rats in each of the following treatment groups; controls, CA-4, HT, C0HT, C3HT and C6HT (mean tumor volume 145 mm 3), were sacri®ced 15 h after either the injection of CA-4 (50 mg/kg i.p.) or the termination of hyperthermia. The tumors were taken out, coded as to treatment group and ®xed immediately in 4% buffered formalin. Sections taken from the mid portion of paraf®n-embedded tumors were stained with haematoxylin and eosin. The coded specimens were examined by two separate investigators using a Zeiss microscope (Zeiss Axiophot, Germany) and joint conclusions were made subsequently. The observers were blinded as to treatment group. 2.10. Immunostaining of tumor vasculature Six tumors were collected 3 h after injecting CA-4 (50 mg/kg i.p.), ®xed in formalin and embedded in paraf®n according to standard procedures. Similarly six tumors were also excised 3 h after sham treatment (0.9% NaCl 1 ml/kg i.p.). The tumors were coded as to treatment group, and the mean tumor volume was 131 mm 3. Serial 5 mm sections were deparaf®nized and rehydrated in decreasing concentrations of ethanol. Subsequent incubation steps were performed at room temperature under humidity control, and the slides were washed between each step with Dulbecco's phosphate-buffered saline (DPBS, Sigma, St. Louis, MO, USA). After a 10 min exposure to 0.1% protease, the sections were incubated for 3 £ 150 s with a peroxidase blocking solution (Dako, Glostrup, Denmark). To reduce background staining, the sections were washed with a 1:10 dilution of swine serum in DPBS, before incubating them 25 min with a 1:800 dilution of a rabbit anti-human von Willebrand factor (Dako). Thereafter the sections were exposed 25 min to a 1:500 dilution of a biotinylated antirabbit IgG antibody (Dako), before adding streptavidin complexed with biotinylated peroxidase (25 min, Dako) and a diaminobenzidine solution (3 £ 5 min, Dako). All sections were counterstained with haematoxylin for 60 s before mounting. The specimens were then observed by two separate investigators that were blinded as to treatment group using a Zeiss microscope and 630 £ magni®cation (oil immersion), and joint conclusions were made subsequently.
149
2.11. Tumor thermometry study In a separate experiment we gave increasing doses of tumor hyperthermia to animals not previously injected with CA-4 in order to reproduce the tumor bed temperatures previously measured in rats exposed to C3HT treatment [8]. We thus wanted to assess whether the increased tumor response achieved by separating CA-4 and hyperthermia 3 h was due to an augmentation of vascular damage or solely to an increased heat deposit in the tumors. The waterbath temperature was increased in a step-wise manner to obtain comparable temperature pro®les (Fig. 1). Temperatures were monitored by a copper constantan thermocouple inserted into the tumor basis parallel to the foot axis, and the temperatures were continuously registered in an HLM82 digital thermometer (Ellab, Copenhagen, Denmark) with a precision of ^0:18C. Tumor bed temperatures were measured as previous investigations have shown this to be the coolest part of the tumor during waterbath hyperthermia [22]. 2.12. Statistics Organ perfusion data and tumor growth time in the different treatment groups were compared by the non-parametric Mann±Whitney test. The regrowth curves of individual tumors in the C0HT and C3HT-group were log-transformed and analyzed by the slopes of the straight portions, using linear regression, and the slopes were thereafter compared statistically by the Mann±Whitney test. 2.13. Ethics The experiments and procedures described were approved by the local responsible laboratory animal science specialist under the surveillance of the Norwegian Animal Research Authority. The experiments were therefore conducted in accordance with the laws and regulations controlling experiments on live animals in Norway.
Fig. 1. Comparison of tumor temperatures obtained during hyperthermia with or without previous injection of CA-4 (50 mg/kg i.p.). Values are means (^SEM) of six tumors.
150
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
3.2. Tumor response study
3. Results
Growth rates of tumors in the different treatment groups are shown in Fig. 4. Tumor response data are shown in Table 1. One of the animals in the C3HT-group was euthanized before reaching the de®ned tumor volume. This was due to malnutrition causing enlarged incisor teeth. The animal was excluded from analysis. CA-4 alone had no signi®cant in¯uence on tumor growth whereas hyperthermia alone gave a small but signi®cant tumor response (P 0:02). There were no signi®cant differences in tumor growth between hyperthermia alone and the C0HT/C12HT/ C24HT combinations. Hyperthermia given 3 or 6 h after CA-4, resulted in a profound enhancement in tumor response compared to simultaneous administration (P 0:001 and P 0:005, respectively). Two animals in the C3HT-group and two animals in the C6HT-group had local tumor control. The slopes of the log-transformed regrowth curves of individual tumors in the C3HT-group were signi®cantly lower than tumors in the C0HT-group (P 0:005).
3.1. Organ perfusion study
3.3. Normal tissue toxicity
Drug induced changes in relative tumor perfusion are illustrated in Fig. 2, expressed as a function of control values. We observed a delayed, but signi®cant reduction in tumor blood ¯ow between 3 and 6 h after drug injection (P 0:002 and P 0:02, respectively), and by 12 h full recovery was seen in tumors exposed to CA-4. Organ perfusion had increased signi®cantly in the left kidney, the heart and the lungs by 3 h (P 0:006, P 0:03 and P 0:02), and a signi®cant reduction in tissue perfusion was found in the skin overlying the tumor (P 0:009, Fig. 3). By 24 h the relative changes in organ perfusion were all suspended (data not shown).
CA-4 enhanced the local side effects of hyperthermia (Table 1). Local skin reactions after combined treatment were mild and resorption of foot edema was seen for all cases within 4 days. CA-4 50 mg/kg alone caused a maximum weight loss of 6%, and the systemic CA-4 toxicity was not increased by adding hyperthermia. Although CA-4 100 mg/kg and 75 mg/kg i.p./i.v. caused massive diarrhea in our BD IX-strain [8], 50 mg/kg i.p. produced only slight diarrhea.
Fig. 2. Time-dependent changes in tumor perfusion after injecting CA-4 50 mg/kg i.p. Values are means (^SEM) of eight tumors. ( 86RbCl extraction technique).
Fig. 3. Relative changes in organ perfusion 3 h after injecting CA-4 50 mg/kg i.p. as compared to sham treated animals. Values are means (^SEM) of eight rats. ( 86RbCl extraction technique).
3.4. Tumor morphology The BT4An rat glioma consisted of pleomorphic atypical cells with numerous mitotic ®gures. The tumors were well vascularized with only scarce areas of necrosis. Tumors
Fig. 4. Effect of CA-4 (C; 50 mg/kg i.p.), hyperthermia (HT; 448C, 60 min) or different combination schedules on the growth of BT4An tumors. The numbers given between characters are hours between treatment modalities. Values are means (^SEM) of 11 tumors.
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
151
Table 1 Effect of timing and sequence of waterbath hyperthermia and CA-4 on BD IX rats with BT4An tumors Tumor effects a Group
b
Controls C HT HT3C C0HT C3HT C6HT C12HT C24HT a b c
Toxicity
No.
Mean TGT (days) Median TGT (days)
PR
CR
Cures
Minimal relative weight Deaths Foot edema (Day 1) Diarrhea (Day 1)
11 11 11 11 11 10 11 11 11
10.5 (0.49) 12.5 (1.13) 13.9 (1.02) 10.8 (1.70) 13.1 (1.52) 83.9 (23.8) 52.3 (17.5) 28.7 (12.3) 10.7 (1.55)
0 0 0 1 0 8 9 2 0
0 0 0 0 0 2 2 0 0
0 0 0 0 0 2 2 0 0
0.99 0.94 0.98 0.90 0.96 0.92 0.90 0.94 0.92
10.9 13.1 12.6 10.0 14.5 46.7 28.4 16.7 8.66
0 0 0 0 0 1c 0 0 0
0 0 2 4 6 10 9 3 5
0 2 0 1 1 2 1 0 3
TGT, tumor growth time; SEM, given in parenthesis; PR, partial regression; CR, complete regression. The numbers given between characters for the different groups are hours between hyperthermia (HT; 448C, 60 min) and CA-4 (C; 50 mg/kg i.p.). One animal was euthanized day 33 (see text).
exposed to CA-4 15 h previously had small areas of apoptotic ®gures, slight necrosis and vascular stasis, but most of the tumor tissue still looked viable with numerous mitoses and functional blood vessels (Fig. 5a). Waterbath hyperthermia (448C and 60 min) resulted in spotted areas of necrosis and bleeding and a generalized tissue edema. Although most tumor cells looked viable, few mitotic ®gures were seen. Tumors exposed to CA-4 and hyperthermia simultaneously still had some small areas of viable looking tissue, but large areas of necrosis and bleeding dominated. No mitotic ®gures were seen. Tumors exposed to CA-4 3 or 6 h before hyperthermia had a comparable morphology. The entire central part of the tumor was necrotic with extensive bleeding (Fig. 5b). A thin rim of presumably viable tissue was seen in the most peripheral and basal parts of the tumors.
were taken out 15 h after heating and processed for haematoxylin and eosin-staining. These tumors had a morphology
3.5. Immunostaining of tumor vasculature Tumors exposed to CA-4 (50 mg/kg) 3 h previously had heterogeneous changes in the neovasculature. Some tumor areas had an endothelial morphology similar to sham treated animals, while in other tumor parts, the endothelial cells appeared contracted and swollen, and invading leucocytes surrounded the blood vessels (Fig. 6b). 3.6. Tumor thermometry studies If CA-4 was omitted, preliminary studies demonstrated that waterbath temperatures between 45 and 478C were needed to give measurements comparable to C3HT-tumors. Whereas 45.2 and 45.58C resulted in too low tumor heat deposits, a waterbath temperature of 46.08C for 60 min gave measurements comparable to C3HT-tumors (Fig. 1). We observed during the thermometry-studies that tumor temperatures followed the water temperatures for the initial 10 min before compensatory mechanisms commenced. Using the data from Fig. 1 we administered waterbath hyperthermia at 46 ^ 0:18C for 60 min to the tumors of three rats not previously injected with CA-4. The tumors
Fig. 5. Histological sections ( £ 200) from BT4An tumors exposed to (a) CA-4 alone (50 mg/kg i.p.) or (b) CA-4 3 h before hyperthermia (448C, 60 min), C3HT. Tumors were excised 15 h after treatment and stained with haematoxylin and eosin. CA-4 alone produced small areas of apoptosis and vascular stasis, whereas the C3HT combination induced massive hemorrhagic necrosis. Scale bars: 320 mm.
152
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
Fig. 6. Histological sections ( £ 630, oil immersion) from BT4An tumors exposed to (a) sham treatment or (b) CA-4 (50 mg/kg i.p.) 3 h previously. Tumor vasculature was stained using anti-von Willebrand factor (peroxidase-conjugated). CA-4 induced a mild endothelial edema and contraction of endothelial cells in some tumor blood vessels. Scale bars: 150 mm.
similar to tumors given C3HT or C6HT treatment, with extensive central necrosis and areas of hemorrhage. 4. Discussion We demonstrate in this study how CA-4 can be used to sensitize the BT4An tumor for hyperthermia, and we argue that improvements in tumor response, after adding CA-4 to hyperthermia, can be explained by a combined destruction of tumor vasculature. We also present data indicating that the cytotoxic compound CA-4 induces morphological changes in the tumor endothelium. Using the 86RbCl extraction technique, we found a delayed and transient reduction in tumor perfusion after injecting CA-4 (Fig. 2). Such reversible changes in tumor blood ¯ow has been reported previously [21], but the data differ from measurements in the murine adenocarcinoma CaNT, where CA-4 induced a persisting shutdown of tumor blood ¯ow [4]. Currently, clinical hyperthermia is hampered by problems in reaching the desired tumor
temperatures, and this is partly due to limitations in equipment and high perfusion rates in human tumors [25]. CA-4 could potentially inhibit thermal washout, and thereby facilitate higher tumor heat deposits, also in human solid malignancies. Alternative vasoactive agents such as ¯avone acetic acid (FAA) and SNOP have proven successful in reducing the blood ¯ow or thermal washout of animal solid tumors, but only at doses approaching the maximum tolerated (i.e. FAA) [13] or requiring general anesthesia (i.e. SNOP) [17,28]. Whereas waterbath hyperthermia at 468C for 60 min is too toxic for the surrounding normal tissue [6], similar tumor temperatures can be obtained by the combined therapy of hyperthermia (448C for 60 min) and CA-4 (50 mg/kg), without major side effects (Table 1). The application of CA-4 as a modi®er of tumor blood ¯ow thus is an interesting therapeutic implementation. The organ perfusion data (Fig. 3) clearly indicates a systemic effect of CA-4. In a prior study we found a 12.8% transient increase in the mean arterial blood pressure after administering the drug to BD IX rats [8], and similar ®ndings have also been reported elsewhere [32]. Using laser Doppler ¯owmetry, a 20% reduction in s.c. blood ¯ow was recorded after CA-4 injection [8], and the current estimate of a 10-15% non-signi®cant reduction in skin blood ¯ow, using the 86RbCl method, suggests a similar response. The BT4An tumor was transplanted to the foot dorsum and presumably established its nutritional supply from blood vessels in the subdermis. A decrease in peripheral blood supply to the skin would therefore secondarily reduce tumor blood ¯ow as well. The observed 45-50% decrease in tumor blood ¯ow exceeds however the change in skin perfusion, and this additional decrease might be explained by a tumor speci®c angiotoxicity mediated by CA-4 within the tumor. CA-4 was originally described as a strong inhibitor of tubulin polymerization, and it exhibited cytotoxicity against a variety of tumor cell lines [27]. CA-4 also inhibited HUVEC (human umbilical vein endothelial cell) proliferation and induced apoptosis of endothelial cells in vitro [7,15]. Due to its inhibition of tubulin polymerization, the endothelial cytoskeleton is a possible target for CA-4 [11,27]. The endothelial toxicity exhibited by colchicine, another tubulin disrupting agent, was suggested as a possible mediator of anti-tumor activity already in 1945 [20], and numerous reports have since con®rmed the vascular targeting properties of tubulin inhibitors in solid tumors [1,7,12,16,30]. In a recent study, vinblastine was shown to induce endothelial edema as well as cellular detachment in tumor blood vessels [16], and various other antineoplastic agents exhibit similar vascular toxicity in normal tissues [18]. Using light microscopy and vascular immunostaining, we assessed the endothelial morphology 3 h after injecting CA-4, and in some tumor areas the endothelial cytoplasm appeared swollen compared to sham treated tumors (Fig. 6). Although this apparent alteration indicates how CA-4 might target tumor vasculature, accurate measurements were
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
prevented by the low resolution of light microscopy. These preliminary ®ndings need therefore to be extended, qualitatively as well as quantitatively, using high resolution techniques, and the issue will be addressed in future studies. If the angiotoxicity of CA-4 was of major importance for the tumor response in BT4An tumors, one would expect more vascular destruction in tumors exposed to the combination of drug and hyperthermia than in tumors given a similar heat dose without using the drug. By increasing the waterbath temperature to 468C we obtained the same intratumoral temperatures as in animals subjected to CA-4 3 h before hyperthermia at 448C (Fig. 1), and we found similar changes in tumor morphology. The reduction in tumor blood ¯ow, causing a higher heat deposit, thus seems more important than the angiotoxicity when combining CA-4 at the maximum tolerated dose of 50 mg/kg with hyperthermia in rats. Vascular disruption and extensive hemorrhagic necrosis has been observed in various murine tumor models after injecting 10% of the maximum tolerated dose of CA-4 (100 mg/kg) [7,19]. In BD IX rats we found a maximum tolerated dose of 50 mg/kg [8], and this dosage failed to in¯ict vascular damage in the BT4An rat glioma (Fig. 5a). The therapeutic window for CA-4 thus is a lot narrower in rats than in mice, and the observed preservation of tumor blood vessels could be due to a subtherapeutic drug exposure in the BT4An tumors as indicated in similar tumor studies [11]. The morphology studies indicated, however, that the vasculature of BT4An tumors was targeted when combining the drug with heating (Fig. 5b). We found massive necrosis and bleeding in the tumor tissue after administering CA-4 and hyperthermia together, and these changes are probably due to destruction of neovasculature. The signi®cantly slower regrowth of tumors treated with hyperthermia 3 h after CA-4 also suggests vascular damage as a major cause of tumor response [14]. Yet viable tissue remained in the peripheral parts of the tumor, and most animals had failure of local tumor control. The addition of cytotoxic drugs like cisplatin might result in a therapeutic gain in this setting as the combination of hyperthermia and cisplatin better targets the well vascularized tumor periphery [22]. Indeed, a recent study demonstrated enhancement of tumor response in CaNT tumours when CA-4 was combined with either fractionated irradiation or cisplatin [4]. We conclude that CA-4 can be used to sensitize solid tumors for hyperthermia if the time interval is adequate. The combination of CA-4 and heat yields angiotoxic effects in the BT4An tumor which effectively retard tumor growth. Acknowledgements The authors greatfully acknowledge the technical assistance of Froydis Rykkja. We are also very grateful to Michael Horsman and Arne Kirkebo for valuable advice
153
on the use of the 86RbCl method. This work was supported by The Norwegian Cancer Society. Combretastatin A-4 disodium phosphate was kindly supplied by Oxigene Inc., Lund, Sweden. References [1] Baguley BC, Holdaway KM, Thomsen LL, Zhuang L, Zwi LJ. Inhibition of growth of colon 38 adenocarcinoma by vinblastine and colchicine: evidence for a vascular mechanism. Eur J Cancer 1991;27:482±487. [2] Beauregard DA, Thelwall PE, Chaplin DJ, Hill SA, Adams GE, Brindle KM. Magnetic resonance imaging and spectroscopy of combretastatin A-4 prodrug-induced disruption of tumour perfusion and energetic status. Br J Cancer 1998;77:1761±1767. [3] Boehm T, Folkman J, Browder T, O'Reilly MS. Antiangiogenic therapy of experimental cancer does not induce acquired drug resistance. Nature 1997;390(6658):404±407. [4] Chaplin DJ, Pettit GR, Hill SA. Anti-vascular approaches to solid tumour therapy: evaluation of combretastatin A-4 phosphate. Anticancer Res 1999;19:189±196. [5] Chaplin DJ, Pettit GR, Parkins CS, Hill SA. Antivascular approaches to solid tumour therapy: evaluation of tubulin binding agents. Br J Cancer 1996;74:S86±S88. [6] Dahl O. Effect of hyperthermia on a neurogenic rat cell line (BT4A) in vivo. Acta Radiol Oncol 1982;21:67±77. [7] Dark GG, Hill SA, Prise VE, Tozer GM, Pettit GR, Chaplin DJ. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer Res 1997;57:1829±1834. [8] Eikesdal HP, Schem BC, Mella O, Dahl O. The new tubulin-inhibitor combretastatin A-4 enhances thermal damage in the BT4An rat glioma. Int J Radiat Oncol Biol Phys 2000;46:645±652. [9] Flecknell P, Mitchell M. Midazolam and fentanyl-¯uanisone: assessment of anesthetic effect in laboratory rodents and rabbits. Lab Anim 1984;18:143±146. [10] Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971;285:1182±1186. [11] Grosios K, Holwell S, McGown A, Pettit G, Bibby M. In vivo and in vitro evaluation of combretastatin A-4 and its sodium phosphate prodrug. Br J Cancer 1999;81:1318±1327. [12] Hill SA, Lonergan SJ, Denekamp J, Chaplin DJ. Vinca alkaloids: antivascular effects in a murine tumour. Eur J Cancer 1993;29A:1320± 1324. [13] Hill SA, Sampson LE, Chaplin DJ. Anti-vascular approaches to solid tumour therapy: evaluation of vinblastine and ¯avone acetic acid. Int J Cancer 1995;63:119±123. [14] Hill SA, Smith KA, Denekamp J. Stromal sensitivity to radiation and hyperthermia. Br J Cancer 1987;56:383±388. [15] Iyer S, Chaplin DJ, Rosenthal DS, Boulares AH, Li LY, Smulson ME. Induction of apoptosis in proliferating human endothelial cells by the tumor-speci®c antiangiogenesis agent combretastatin A-4. Cancer Res 1998;58:4510±4514. [16] Klement G, Baruchel S, Rak J, et al. Continuous low-dose therapy with vinblastine and VEGF receptor-2 antibody induces sustained tumor regression without overt toxicity. J Clin Invest 2000;105:R15±R24. [17] Krossnes BK, Mella O, Dahl O. Use of the vasodilator sodium nitroprusside during local hyperthermia: effects on tumor temperature and tumor response in a rat tumor model. Int J Radiat Oncol Biol Phys 1996;36:403±415. [18] Lazo JS. Endothelial injury caused by antineoplastic agents. Biochem Pharmacol 1986;35:1919±1923. [19] Li L, Rojiani A, Siemann DW. Targeting the tumor vasculature with combretastatin A-4 disodium phosphate: effects on radiation therapy. Int J Radiat Oncol Biol Phys 1998;42:899±903.
154
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
[20] Ludford RJ. Colchicine in the experimental chemotherapy of cancer. J Natl Cancer Inst 1945;6:89±101. [21] Maxwell RJ, Nielsen FU, Breidahl T, Stùdkilde-Jùrgensen H, Horsman MR. Effects of combretastatin on murine tumours monitored by 31P MRS, 1H MRS and 1H MRI. Int J Radiat Oncol Biol Phys 1998;42:891±894. [22] Mella O. Combined hyperthermia and cis-diamminedichloroplatinum in BD IX rats with transplanted BT4A tumours. Int J Hyperthermia 1985;1:171±183. [23] Mella O, Bjerkvig R, Schem B, Dahl O, Laerum O. A cerebral glioma model for experimental therapy and in vivo invasion studies in syngenic BD IX rats. J Neuro-Oncol 1990;9:93±104. [24] Mella O, Eriksen R, Dahl O, Laerum O. Acute systemic toxicity of combined cis-diamminedichloroplatinum and hyperthermia in the rat. Eur J Cancer Clin Oncol 1987;23:365±373. [25] Olch AJ, Kaiser LR, Silberman AW, Storm FK, Graham LS, Morton DL. Blood ¯ow in human tumors during hyperthermia therapy: demonstration of vasoregulation and an applicable physiological model. J Surg Oncol 1983;23:125±132. [26] O'Reilly MS, Boehm T, Shing Y, et al. Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 1997;88:277±285. [27] Pettit GR, Singh SB, Hamel E, Lin CM, Alberts DS, Garcia-Kendall D. Isolation and structure of the strong cell growth and tubulin inhibitor combretastatin A-4. Experientia 1989;45:209±211.
[28] Prescott D, Samulski T, Dewhirst M, et al. Use of nitroprusside to increase tissue temperature during local hyperthermia in normal and tumor-bearing dogs. Int J Radiat Oncol Biol Phys 1992;23:377±385. [29] Sapirstein LA. Regional blood ¯ow by fractional distribution of indicators. Am J Physiol 1958;193:161±168. [30] Schirner M, Hoffmann J, Menrad A, Schneider MR. Antiangiogenic chemotherapeutic agents: characterization in comparison to their tumor growth inhibition in human renal cell carcinoma models. Clin Cancer Res 1998;4:1331±1336. Ê byholm FE, Lekven J. [31] Skolleborg KC, GroÈnbech JE, Grong K, A Distribution of cardiac output during pentobarbital versus midazolam/fentanyl/¯uanisone anaesthesia in the rat. Lab Anim 1990;24:221±227. [32] Tozer GM, Prise VE, Wilson J, et al. Combretastatin A-4 phosphate as a tumor vascular-targeting agent: early effects in tumors and normal tissues. Cancer Res 1999;59:1626±1634. [33] Wood L, Palmer M, Hewitt J, et al. Results of a phase III, doubleblind, placebo-controlled trial of megestrol acetate modulation of Pglycoprotein-mediated drug resistance in the ®rst-line management of small-cell lung carcinoma. Br J Cancer 1998;77:627±631. [34] Zanelli GD, Fowler JF. The measurement of blood perfusion in experimental tumors by uptake of 86Rb. Cancer Res 1974;34:1451± 1456.
www.elsevier.com/locate/radonline
Combretastatin A-4 and hyperthermia; a potent combination for the treatment of solid tumors Hans P. Eikesdal a,*, Rolf Bjerkvig b, Olav Mella a, Olav Dahl a a
b
Department of Oncology, University of Bergen, Haukeland University Hospital, 5021 Bergen, Norway Department of Anatomy and Cell Biology, University of Bergen, Aarstadveien 19, 5009 Bergen, Norway Received 2 February 2000; received in revised form 8 August 2000; accepted 30 October 2000
Abstract Background and purpose: Attacking tumor vasculature is a promising approach for the treatment of solid tumors. The tubulin inhibitor combretastatin A-4 disodium phosphate (CA-4) is a new vascular targeting drug which displays a low toxicity pro®le. We wanted to investigate how CA-4 in¯uences tumor perfusion in the BT4An rat glioma and how the vascular targeting properties of CA-4 could be exploited to augment hyperthermic damage towards tumor vasculature. Material and methods: We used the 86RbCl extraction technique to assess how CA-4 in¯uences tumor perfusion, and the tumor endothelium was examined for morphological changes induced by the drug. We combined CA-4 (50 mg/kg i.p.) with hyperthermia (448C, 60 min) at different time intervals to evaluate how therapy should be designed to affect tumor growth, and we studied the tumors histologically to assess tissue viability. Results: We found that CA-4 induced a profound, but transient reduction in tumor perfusion 3±6 h postinjection. If hyperthermia was administered 3±6 h after injecting CA-4, massive hemorrhagic necrosis developed, and tumor response was signi®cantly enhanced compared to simultaneous administration of the two treatment modalities (P , 0:005). CA-4 alone had no in¯uence on tumor growth and failed to disrupt the vasculature of the BT4An solid tumors. Interestingly though, a mild endothelial edema was observed in some tumor areas 3 h after injecting CA-4. Conclusions: We conclude that the combination of CA-4 and hyperthermia is a potent therapeutic option for BT4An tumors, but the selection of adequate time intervals between CA-4 and hyperthermia are imperative to obtain tumor response. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Combretastatin A-4 disodium phosphate; Hyperthermia; Tumor perfusion; Histology; BT4An
1. Introduction The failure of chemotherapy is a common obstacle for the successful treatment of solid tumors. The genetic instability in malignant tumors leads to tumor cell heterogeneity where drug resistance commonly emerges due to a selection of surviving cellular subpopulations. Attempts to block the mechanisms leading to drug resistance, such as by inhibiting the P-glycoprotein, have proven insuf®cient [33], and new strategies have therefore evolved to circumvent acquired drug resistance. As ®rst suggested by Judah Folkman [10], attacking the neovascularization process is an alternative approach for the treatment of solid tumors, and recent reports have disclosed a potential therapeutic effect of anti-angiogenic therapy since proliferating endothelial
* Corresponding author.
cells within the tumors are less capable of developing drug resistance [3,26]. As opposed to inhibiting the outgrowth of new tumor capillaries, therapeutic strategies may also be designed to destroy established tumor blood vessels; i.e. angiotoxic therapy. It has recently been shown that various drugs which inhibit tubulin polymerization, also reduce tumor blood ¯ow, and a new tubulin inhibitor, combretastatin A-4 disodium phosphate (CA-4), demonstrated profound angiotoxic properties in solid tumors [2,5,7]. Hyperthermia is another treatment modality which effectively destructs tumor vasculature, but suf®cient thermal doses are dif®cult to obtain in human tumors due to reactive hyperemia causing thermal washout [25]. Combining hyperthermia in vivo with ¯ow decreasing compounds like sodium nitroprusside (SNOP), have led to encouraging results, increasing tumor temperatures and prolonging tumor growth delay [17,28]. Applying CA-4 before hyperthermia, could presumably mediate simi-
0167-8140/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(00)00318-2
148
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
lar effects as SNOP without the need for systemic hypotension. Additionally, the angiotoxic properties of CA-4 might enhance thermal damage to tumor blood vessels. In the s.c. BT4An rat glioma, we used the 86RbCl extraction technique [29] to assess how CA-4 in¯uences tumor blood ¯ow. We thereafter conducted a tumor response study to evaluate the combination of the drug with hyperthermia at different time intervals. To disclose some of the potential interactions of CA-4 and hyperthermia, we also investigated changes in tumor morphology after treatment. 2. Materials and methods 2.1. Animals and tumor Rats of both sexes of the inbred strain BD IX/HanFoss, mean weight 200 g (SD 87), were used. They were given water and standard pelleted diet ad libitum. Rats were caged at 228C in a night/day cycle 12/12 h. The BT4An tumor is an aggressive glioma which can give rise to lymphatic and haematogenous metastases when grown subcutaneously (s.c.) [23]. For each experiment a single tumor-bearing rat was sacri®ced and tumor pieces measuring 1±2 mm were transplanted to the dorsum of the right hind leg through an incision 5 mm proximal to the tumor bed. The animals in the tumor response study were treated 9±11 days after transplantation at a mean tumor volume of 188 mm 3 (SD 27). For the tumor perfusion study animals with a mean tumor volume of 143 mm 3 (SD 69) were used, and in the tumor thermometry study animals had a mean tumor volume of 213 mm 3 (SD 64). 2.2. Anesthesia During interventional procedures and hyperthermia, the animals were kept anesthetized with s.c. injections of fentanyl (0.05 mg/kg)/¯uanisone (2.5 mg/kg)/midazolam (1.25 mg/kg) (Hypnorm Dormicum) [9]. Normal body temperature in anesthetized animals that did not receive hyperthermia was maintained by the use of a thermostatically controlled incubator and a heating pad (Harvard Apparatus 50-7061, Kent, England). 2.3. CA-4 administration CA-4 was kindly supplied by Oxigene Inc., Lund, Sweden. It was dissolved in 0.9% NaCl to a ®nal concentration of 50 mg/ml. CA-4 was administered i.p. at a maximum tolerated dose of 50 mg/kg [8]. 2.4. Hyperthermia technique The rats were placed in plastic jigs. The tumor-bearing foot was ®xed and immersed 1.5±2 cm below the surface in a circulating, precision controlled waterbath. The waterbath temperature was 44 ^ 0:18C during the tumor response
study, and the treatment time was 60 min. This results in mean tumor temperatures of 41:2±41:88C when hyperthermia is administered alone [8]. To reduce systemic hyperthermia, the rats were cooled by fanning room air through perforations in the jigs [24]. 2.5. Organ perfusion study Organ perfusion was measured at various intervals after drug injection by the 86RbCl extraction technique. Tissue radioactivity 90 s after an i.v. injection of 86RbCl can be used to calculate relative blood ¯ow as a proportion of cardiac output [34]. Brie¯y, after inserting a catheter into the left femoral vein, eight rats per group were injected with a volume of 0.2 ml 86RbCl (Amersham International Biotech UK Limited, Little Chalfont, England) (speci®c activity 0.5±10 mCi/mg and 50 mCi/ml). The rats were sacri®ced 90 s later by an overdose of sodium pentobarbital (100 mg/kg i.v.). Tumor material and organs were excised immediately, weighed and counted for radioactivity (Packard Cobra II Auto Gamma Meriden, CT, USA). The percentage of injected radioactivity/g of tissue was used to calculate relative changes in organ perfusion after CA-4 administration. The selection of an anesthetic regimen is a potential bias when studying tissue perfusion. The anesthetic, Hypnorm Dormicum, employed in the present study lowers the blood pressure and increases the cardiac output and the heart rate in rats, and it has been speculated that the drug reduces total peripheral resistance [17,31]. Such hemodynamic changes might affect the organ perfusion, and to circumvent this problem, we therefore anesthetized all animals included in the organ perfusion study and calculated relative changes in perfusion as a percentage of the control group. 2.6. Tumor response study The rats were strati®ed by tumor size and randomized to one of the following nine treatment groups: controls, hyperthermia alone, CA-4 alone, hyperthermia 3 h before CA-4 (HT3C), CA-4 immediately before hyperthermia (C0HT), CA-4 3, 6, 12 or 24 h before hyperthermia (C3HT/C6HT/C12HT/C24HT). The experiment was run on 3 consecutive days with the same tumor passage. All animals included in the tumor response study were given Hypnorm Dormicum to preclude data skewness due to the anesthetic regimen. 2.7. Normal tissue toxicity During the tumor response study, side effects, including weight changes, were evaluated 48 h after treatment and then twice weekly. 2.8. Tumor evaluation Tumors were measured twice weekly by vernier callipers, and tumor volume calculated using the modi®ed ellipsoid
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
formula: p=6 £ A £ B2 , where A is the longer axis and B is the axis perpendicular to A. The time to reach four times the initial tumor volume was de®ned as tumor growth time (TGT). Partial regression (PR) was de®ned as a greater than 50% volume reduction observed at least twice and lasting for at least 1 week. Complete regression (CR) was de®ned as no visible or palpable sign of tumor observed at least twice. Cure was a complete regression lasting 120 days. 2.9. Tumor morphology Four rats in each of the following treatment groups; controls, CA-4, HT, C0HT, C3HT and C6HT (mean tumor volume 145 mm 3), were sacri®ced 15 h after either the injection of CA-4 (50 mg/kg i.p.) or the termination of hyperthermia. The tumors were taken out, coded as to treatment group and ®xed immediately in 4% buffered formalin. Sections taken from the mid portion of paraf®n-embedded tumors were stained with haematoxylin and eosin. The coded specimens were examined by two separate investigators using a Zeiss microscope (Zeiss Axiophot, Germany) and joint conclusions were made subsequently. The observers were blinded as to treatment group. 2.10. Immunostaining of tumor vasculature Six tumors were collected 3 h after injecting CA-4 (50 mg/kg i.p.), ®xed in formalin and embedded in paraf®n according to standard procedures. Similarly six tumors were also excised 3 h after sham treatment (0.9% NaCl 1 ml/kg i.p.). The tumors were coded as to treatment group, and the mean tumor volume was 131 mm 3. Serial 5 mm sections were deparaf®nized and rehydrated in decreasing concentrations of ethanol. Subsequent incubation steps were performed at room temperature under humidity control, and the slides were washed between each step with Dulbecco's phosphate-buffered saline (DPBS, Sigma, St. Louis, MO, USA). After a 10 min exposure to 0.1% protease, the sections were incubated for 3 £ 150 s with a peroxidase blocking solution (Dako, Glostrup, Denmark). To reduce background staining, the sections were washed with a 1:10 dilution of swine serum in DPBS, before incubating them 25 min with a 1:800 dilution of a rabbit anti-human von Willebrand factor (Dako). Thereafter the sections were exposed 25 min to a 1:500 dilution of a biotinylated antirabbit IgG antibody (Dako), before adding streptavidin complexed with biotinylated peroxidase (25 min, Dako) and a diaminobenzidine solution (3 £ 5 min, Dako). All sections were counterstained with haematoxylin for 60 s before mounting. The specimens were then observed by two separate investigators that were blinded as to treatment group using a Zeiss microscope and 630 £ magni®cation (oil immersion), and joint conclusions were made subsequently.
149
2.11. Tumor thermometry study In a separate experiment we gave increasing doses of tumor hyperthermia to animals not previously injected with CA-4 in order to reproduce the tumor bed temperatures previously measured in rats exposed to C3HT treatment [8]. We thus wanted to assess whether the increased tumor response achieved by separating CA-4 and hyperthermia 3 h was due to an augmentation of vascular damage or solely to an increased heat deposit in the tumors. The waterbath temperature was increased in a step-wise manner to obtain comparable temperature pro®les (Fig. 1). Temperatures were monitored by a copper constantan thermocouple inserted into the tumor basis parallel to the foot axis, and the temperatures were continuously registered in an HLM82 digital thermometer (Ellab, Copenhagen, Denmark) with a precision of ^0:18C. Tumor bed temperatures were measured as previous investigations have shown this to be the coolest part of the tumor during waterbath hyperthermia [22]. 2.12. Statistics Organ perfusion data and tumor growth time in the different treatment groups were compared by the non-parametric Mann±Whitney test. The regrowth curves of individual tumors in the C0HT and C3HT-group were log-transformed and analyzed by the slopes of the straight portions, using linear regression, and the slopes were thereafter compared statistically by the Mann±Whitney test. 2.13. Ethics The experiments and procedures described were approved by the local responsible laboratory animal science specialist under the surveillance of the Norwegian Animal Research Authority. The experiments were therefore conducted in accordance with the laws and regulations controlling experiments on live animals in Norway.
Fig. 1. Comparison of tumor temperatures obtained during hyperthermia with or without previous injection of CA-4 (50 mg/kg i.p.). Values are means (^SEM) of six tumors.
150
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
3.2. Tumor response study
3. Results
Growth rates of tumors in the different treatment groups are shown in Fig. 4. Tumor response data are shown in Table 1. One of the animals in the C3HT-group was euthanized before reaching the de®ned tumor volume. This was due to malnutrition causing enlarged incisor teeth. The animal was excluded from analysis. CA-4 alone had no signi®cant in¯uence on tumor growth whereas hyperthermia alone gave a small but signi®cant tumor response (P 0:02). There were no signi®cant differences in tumor growth between hyperthermia alone and the C0HT/C12HT/ C24HT combinations. Hyperthermia given 3 or 6 h after CA-4, resulted in a profound enhancement in tumor response compared to simultaneous administration (P 0:001 and P 0:005, respectively). Two animals in the C3HT-group and two animals in the C6HT-group had local tumor control. The slopes of the log-transformed regrowth curves of individual tumors in the C3HT-group were signi®cantly lower than tumors in the C0HT-group (P 0:005).
3.1. Organ perfusion study
3.3. Normal tissue toxicity
Drug induced changes in relative tumor perfusion are illustrated in Fig. 2, expressed as a function of control values. We observed a delayed, but signi®cant reduction in tumor blood ¯ow between 3 and 6 h after drug injection (P 0:002 and P 0:02, respectively), and by 12 h full recovery was seen in tumors exposed to CA-4. Organ perfusion had increased signi®cantly in the left kidney, the heart and the lungs by 3 h (P 0:006, P 0:03 and P 0:02), and a signi®cant reduction in tissue perfusion was found in the skin overlying the tumor (P 0:009, Fig. 3). By 24 h the relative changes in organ perfusion were all suspended (data not shown).
CA-4 enhanced the local side effects of hyperthermia (Table 1). Local skin reactions after combined treatment were mild and resorption of foot edema was seen for all cases within 4 days. CA-4 50 mg/kg alone caused a maximum weight loss of 6%, and the systemic CA-4 toxicity was not increased by adding hyperthermia. Although CA-4 100 mg/kg and 75 mg/kg i.p./i.v. caused massive diarrhea in our BD IX-strain [8], 50 mg/kg i.p. produced only slight diarrhea.
Fig. 2. Time-dependent changes in tumor perfusion after injecting CA-4 50 mg/kg i.p. Values are means (^SEM) of eight tumors. ( 86RbCl extraction technique).
Fig. 3. Relative changes in organ perfusion 3 h after injecting CA-4 50 mg/kg i.p. as compared to sham treated animals. Values are means (^SEM) of eight rats. ( 86RbCl extraction technique).
3.4. Tumor morphology The BT4An rat glioma consisted of pleomorphic atypical cells with numerous mitotic ®gures. The tumors were well vascularized with only scarce areas of necrosis. Tumors
Fig. 4. Effect of CA-4 (C; 50 mg/kg i.p.), hyperthermia (HT; 448C, 60 min) or different combination schedules on the growth of BT4An tumors. The numbers given between characters are hours between treatment modalities. Values are means (^SEM) of 11 tumors.
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
151
Table 1 Effect of timing and sequence of waterbath hyperthermia and CA-4 on BD IX rats with BT4An tumors Tumor effects a Group
b
Controls C HT HT3C C0HT C3HT C6HT C12HT C24HT a b c
Toxicity
No.
Mean TGT (days) Median TGT (days)
PR
CR
Cures
Minimal relative weight Deaths Foot edema (Day 1) Diarrhea (Day 1)
11 11 11 11 11 10 11 11 11
10.5 (0.49) 12.5 (1.13) 13.9 (1.02) 10.8 (1.70) 13.1 (1.52) 83.9 (23.8) 52.3 (17.5) 28.7 (12.3) 10.7 (1.55)
0 0 0 1 0 8 9 2 0
0 0 0 0 0 2 2 0 0
0 0 0 0 0 2 2 0 0
0.99 0.94 0.98 0.90 0.96 0.92 0.90 0.94 0.92
10.9 13.1 12.6 10.0 14.5 46.7 28.4 16.7 8.66
0 0 0 0 0 1c 0 0 0
0 0 2 4 6 10 9 3 5
0 2 0 1 1 2 1 0 3
TGT, tumor growth time; SEM, given in parenthesis; PR, partial regression; CR, complete regression. The numbers given between characters for the different groups are hours between hyperthermia (HT; 448C, 60 min) and CA-4 (C; 50 mg/kg i.p.). One animal was euthanized day 33 (see text).
exposed to CA-4 15 h previously had small areas of apoptotic ®gures, slight necrosis and vascular stasis, but most of the tumor tissue still looked viable with numerous mitoses and functional blood vessels (Fig. 5a). Waterbath hyperthermia (448C and 60 min) resulted in spotted areas of necrosis and bleeding and a generalized tissue edema. Although most tumor cells looked viable, few mitotic ®gures were seen. Tumors exposed to CA-4 and hyperthermia simultaneously still had some small areas of viable looking tissue, but large areas of necrosis and bleeding dominated. No mitotic ®gures were seen. Tumors exposed to CA-4 3 or 6 h before hyperthermia had a comparable morphology. The entire central part of the tumor was necrotic with extensive bleeding (Fig. 5b). A thin rim of presumably viable tissue was seen in the most peripheral and basal parts of the tumors.
were taken out 15 h after heating and processed for haematoxylin and eosin-staining. These tumors had a morphology
3.5. Immunostaining of tumor vasculature Tumors exposed to CA-4 (50 mg/kg) 3 h previously had heterogeneous changes in the neovasculature. Some tumor areas had an endothelial morphology similar to sham treated animals, while in other tumor parts, the endothelial cells appeared contracted and swollen, and invading leucocytes surrounded the blood vessels (Fig. 6b). 3.6. Tumor thermometry studies If CA-4 was omitted, preliminary studies demonstrated that waterbath temperatures between 45 and 478C were needed to give measurements comparable to C3HT-tumors. Whereas 45.2 and 45.58C resulted in too low tumor heat deposits, a waterbath temperature of 46.08C for 60 min gave measurements comparable to C3HT-tumors (Fig. 1). We observed during the thermometry-studies that tumor temperatures followed the water temperatures for the initial 10 min before compensatory mechanisms commenced. Using the data from Fig. 1 we administered waterbath hyperthermia at 46 ^ 0:18C for 60 min to the tumors of three rats not previously injected with CA-4. The tumors
Fig. 5. Histological sections ( £ 200) from BT4An tumors exposed to (a) CA-4 alone (50 mg/kg i.p.) or (b) CA-4 3 h before hyperthermia (448C, 60 min), C3HT. Tumors were excised 15 h after treatment and stained with haematoxylin and eosin. CA-4 alone produced small areas of apoptosis and vascular stasis, whereas the C3HT combination induced massive hemorrhagic necrosis. Scale bars: 320 mm.
152
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
Fig. 6. Histological sections ( £ 630, oil immersion) from BT4An tumors exposed to (a) sham treatment or (b) CA-4 (50 mg/kg i.p.) 3 h previously. Tumor vasculature was stained using anti-von Willebrand factor (peroxidase-conjugated). CA-4 induced a mild endothelial edema and contraction of endothelial cells in some tumor blood vessels. Scale bars: 150 mm.
similar to tumors given C3HT or C6HT treatment, with extensive central necrosis and areas of hemorrhage. 4. Discussion We demonstrate in this study how CA-4 can be used to sensitize the BT4An tumor for hyperthermia, and we argue that improvements in tumor response, after adding CA-4 to hyperthermia, can be explained by a combined destruction of tumor vasculature. We also present data indicating that the cytotoxic compound CA-4 induces morphological changes in the tumor endothelium. Using the 86RbCl extraction technique, we found a delayed and transient reduction in tumor perfusion after injecting CA-4 (Fig. 2). Such reversible changes in tumor blood ¯ow has been reported previously [21], but the data differ from measurements in the murine adenocarcinoma CaNT, where CA-4 induced a persisting shutdown of tumor blood ¯ow [4]. Currently, clinical hyperthermia is hampered by problems in reaching the desired tumor
temperatures, and this is partly due to limitations in equipment and high perfusion rates in human tumors [25]. CA-4 could potentially inhibit thermal washout, and thereby facilitate higher tumor heat deposits, also in human solid malignancies. Alternative vasoactive agents such as ¯avone acetic acid (FAA) and SNOP have proven successful in reducing the blood ¯ow or thermal washout of animal solid tumors, but only at doses approaching the maximum tolerated (i.e. FAA) [13] or requiring general anesthesia (i.e. SNOP) [17,28]. Whereas waterbath hyperthermia at 468C for 60 min is too toxic for the surrounding normal tissue [6], similar tumor temperatures can be obtained by the combined therapy of hyperthermia (448C for 60 min) and CA-4 (50 mg/kg), without major side effects (Table 1). The application of CA-4 as a modi®er of tumor blood ¯ow thus is an interesting therapeutic implementation. The organ perfusion data (Fig. 3) clearly indicates a systemic effect of CA-4. In a prior study we found a 12.8% transient increase in the mean arterial blood pressure after administering the drug to BD IX rats [8], and similar ®ndings have also been reported elsewhere [32]. Using laser Doppler ¯owmetry, a 20% reduction in s.c. blood ¯ow was recorded after CA-4 injection [8], and the current estimate of a 10-15% non-signi®cant reduction in skin blood ¯ow, using the 86RbCl method, suggests a similar response. The BT4An tumor was transplanted to the foot dorsum and presumably established its nutritional supply from blood vessels in the subdermis. A decrease in peripheral blood supply to the skin would therefore secondarily reduce tumor blood ¯ow as well. The observed 45-50% decrease in tumor blood ¯ow exceeds however the change in skin perfusion, and this additional decrease might be explained by a tumor speci®c angiotoxicity mediated by CA-4 within the tumor. CA-4 was originally described as a strong inhibitor of tubulin polymerization, and it exhibited cytotoxicity against a variety of tumor cell lines [27]. CA-4 also inhibited HUVEC (human umbilical vein endothelial cell) proliferation and induced apoptosis of endothelial cells in vitro [7,15]. Due to its inhibition of tubulin polymerization, the endothelial cytoskeleton is a possible target for CA-4 [11,27]. The endothelial toxicity exhibited by colchicine, another tubulin disrupting agent, was suggested as a possible mediator of anti-tumor activity already in 1945 [20], and numerous reports have since con®rmed the vascular targeting properties of tubulin inhibitors in solid tumors [1,7,12,16,30]. In a recent study, vinblastine was shown to induce endothelial edema as well as cellular detachment in tumor blood vessels [16], and various other antineoplastic agents exhibit similar vascular toxicity in normal tissues [18]. Using light microscopy and vascular immunostaining, we assessed the endothelial morphology 3 h after injecting CA-4, and in some tumor areas the endothelial cytoplasm appeared swollen compared to sham treated tumors (Fig. 6). Although this apparent alteration indicates how CA-4 might target tumor vasculature, accurate measurements were
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
prevented by the low resolution of light microscopy. These preliminary ®ndings need therefore to be extended, qualitatively as well as quantitatively, using high resolution techniques, and the issue will be addressed in future studies. If the angiotoxicity of CA-4 was of major importance for the tumor response in BT4An tumors, one would expect more vascular destruction in tumors exposed to the combination of drug and hyperthermia than in tumors given a similar heat dose without using the drug. By increasing the waterbath temperature to 468C we obtained the same intratumoral temperatures as in animals subjected to CA-4 3 h before hyperthermia at 448C (Fig. 1), and we found similar changes in tumor morphology. The reduction in tumor blood ¯ow, causing a higher heat deposit, thus seems more important than the angiotoxicity when combining CA-4 at the maximum tolerated dose of 50 mg/kg with hyperthermia in rats. Vascular disruption and extensive hemorrhagic necrosis has been observed in various murine tumor models after injecting 10% of the maximum tolerated dose of CA-4 (100 mg/kg) [7,19]. In BD IX rats we found a maximum tolerated dose of 50 mg/kg [8], and this dosage failed to in¯ict vascular damage in the BT4An rat glioma (Fig. 5a). The therapeutic window for CA-4 thus is a lot narrower in rats than in mice, and the observed preservation of tumor blood vessels could be due to a subtherapeutic drug exposure in the BT4An tumors as indicated in similar tumor studies [11]. The morphology studies indicated, however, that the vasculature of BT4An tumors was targeted when combining the drug with heating (Fig. 5b). We found massive necrosis and bleeding in the tumor tissue after administering CA-4 and hyperthermia together, and these changes are probably due to destruction of neovasculature. The signi®cantly slower regrowth of tumors treated with hyperthermia 3 h after CA-4 also suggests vascular damage as a major cause of tumor response [14]. Yet viable tissue remained in the peripheral parts of the tumor, and most animals had failure of local tumor control. The addition of cytotoxic drugs like cisplatin might result in a therapeutic gain in this setting as the combination of hyperthermia and cisplatin better targets the well vascularized tumor periphery [22]. Indeed, a recent study demonstrated enhancement of tumor response in CaNT tumours when CA-4 was combined with either fractionated irradiation or cisplatin [4]. We conclude that CA-4 can be used to sensitize solid tumors for hyperthermia if the time interval is adequate. The combination of CA-4 and heat yields angiotoxic effects in the BT4An tumor which effectively retard tumor growth. Acknowledgements The authors greatfully acknowledge the technical assistance of Froydis Rykkja. We are also very grateful to Michael Horsman and Arne Kirkebo for valuable advice
153
on the use of the 86RbCl method. This work was supported by The Norwegian Cancer Society. Combretastatin A-4 disodium phosphate was kindly supplied by Oxigene Inc., Lund, Sweden. References [1] Baguley BC, Holdaway KM, Thomsen LL, Zhuang L, Zwi LJ. Inhibition of growth of colon 38 adenocarcinoma by vinblastine and colchicine: evidence for a vascular mechanism. Eur J Cancer 1991;27:482±487. [2] Beauregard DA, Thelwall PE, Chaplin DJ, Hill SA, Adams GE, Brindle KM. Magnetic resonance imaging and spectroscopy of combretastatin A-4 prodrug-induced disruption of tumour perfusion and energetic status. Br J Cancer 1998;77:1761±1767. [3] Boehm T, Folkman J, Browder T, O'Reilly MS. Antiangiogenic therapy of experimental cancer does not induce acquired drug resistance. Nature 1997;390(6658):404±407. [4] Chaplin DJ, Pettit GR, Hill SA. Anti-vascular approaches to solid tumour therapy: evaluation of combretastatin A-4 phosphate. Anticancer Res 1999;19:189±196. [5] Chaplin DJ, Pettit GR, Parkins CS, Hill SA. Antivascular approaches to solid tumour therapy: evaluation of tubulin binding agents. Br J Cancer 1996;74:S86±S88. [6] Dahl O. Effect of hyperthermia on a neurogenic rat cell line (BT4A) in vivo. Acta Radiol Oncol 1982;21:67±77. [7] Dark GG, Hill SA, Prise VE, Tozer GM, Pettit GR, Chaplin DJ. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer Res 1997;57:1829±1834. [8] Eikesdal HP, Schem BC, Mella O, Dahl O. The new tubulin-inhibitor combretastatin A-4 enhances thermal damage in the BT4An rat glioma. Int J Radiat Oncol Biol Phys 2000;46:645±652. [9] Flecknell P, Mitchell M. Midazolam and fentanyl-¯uanisone: assessment of anesthetic effect in laboratory rodents and rabbits. Lab Anim 1984;18:143±146. [10] Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971;285:1182±1186. [11] Grosios K, Holwell S, McGown A, Pettit G, Bibby M. In vivo and in vitro evaluation of combretastatin A-4 and its sodium phosphate prodrug. Br J Cancer 1999;81:1318±1327. [12] Hill SA, Lonergan SJ, Denekamp J, Chaplin DJ. Vinca alkaloids: antivascular effects in a murine tumour. Eur J Cancer 1993;29A:1320± 1324. [13] Hill SA, Sampson LE, Chaplin DJ. Anti-vascular approaches to solid tumour therapy: evaluation of vinblastine and ¯avone acetic acid. Int J Cancer 1995;63:119±123. [14] Hill SA, Smith KA, Denekamp J. Stromal sensitivity to radiation and hyperthermia. Br J Cancer 1987;56:383±388. [15] Iyer S, Chaplin DJ, Rosenthal DS, Boulares AH, Li LY, Smulson ME. Induction of apoptosis in proliferating human endothelial cells by the tumor-speci®c antiangiogenesis agent combretastatin A-4. Cancer Res 1998;58:4510±4514. [16] Klement G, Baruchel S, Rak J, et al. Continuous low-dose therapy with vinblastine and VEGF receptor-2 antibody induces sustained tumor regression without overt toxicity. J Clin Invest 2000;105:R15±R24. [17] Krossnes BK, Mella O, Dahl O. Use of the vasodilator sodium nitroprusside during local hyperthermia: effects on tumor temperature and tumor response in a rat tumor model. Int J Radiat Oncol Biol Phys 1996;36:403±415. [18] Lazo JS. Endothelial injury caused by antineoplastic agents. Biochem Pharmacol 1986;35:1919±1923. [19] Li L, Rojiani A, Siemann DW. Targeting the tumor vasculature with combretastatin A-4 disodium phosphate: effects on radiation therapy. Int J Radiat Oncol Biol Phys 1998;42:899±903.
154
H.P. Eikesdal et al. / Radiotherapy and Oncology 60 (2001) 147±154
[20] Ludford RJ. Colchicine in the experimental chemotherapy of cancer. J Natl Cancer Inst 1945;6:89±101. [21] Maxwell RJ, Nielsen FU, Breidahl T, Stùdkilde-Jùrgensen H, Horsman MR. Effects of combretastatin on murine tumours monitored by 31P MRS, 1H MRS and 1H MRI. Int J Radiat Oncol Biol Phys 1998;42:891±894. [22] Mella O. Combined hyperthermia and cis-diamminedichloroplatinum in BD IX rats with transplanted BT4A tumours. Int J Hyperthermia 1985;1:171±183. [23] Mella O, Bjerkvig R, Schem B, Dahl O, Laerum O. A cerebral glioma model for experimental therapy and in vivo invasion studies in syngenic BD IX rats. J Neuro-Oncol 1990;9:93±104. [24] Mella O, Eriksen R, Dahl O, Laerum O. Acute systemic toxicity of combined cis-diamminedichloroplatinum and hyperthermia in the rat. Eur J Cancer Clin Oncol 1987;23:365±373. [25] Olch AJ, Kaiser LR, Silberman AW, Storm FK, Graham LS, Morton DL. Blood ¯ow in human tumors during hyperthermia therapy: demonstration of vasoregulation and an applicable physiological model. J Surg Oncol 1983;23:125±132. [26] O'Reilly MS, Boehm T, Shing Y, et al. Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 1997;88:277±285. [27] Pettit GR, Singh SB, Hamel E, Lin CM, Alberts DS, Garcia-Kendall D. Isolation and structure of the strong cell growth and tubulin inhibitor combretastatin A-4. Experientia 1989;45:209±211.
[28] Prescott D, Samulski T, Dewhirst M, et al. Use of nitroprusside to increase tissue temperature during local hyperthermia in normal and tumor-bearing dogs. Int J Radiat Oncol Biol Phys 1992;23:377±385. [29] Sapirstein LA. Regional blood ¯ow by fractional distribution of indicators. Am J Physiol 1958;193:161±168. [30] Schirner M, Hoffmann J, Menrad A, Schneider MR. Antiangiogenic chemotherapeutic agents: characterization in comparison to their tumor growth inhibition in human renal cell carcinoma models. Clin Cancer Res 1998;4:1331±1336. Ê byholm FE, Lekven J. [31] Skolleborg KC, GroÈnbech JE, Grong K, A Distribution of cardiac output during pentobarbital versus midazolam/fentanyl/¯uanisone anaesthesia in the rat. Lab Anim 1990;24:221±227. [32] Tozer GM, Prise VE, Wilson J, et al. Combretastatin A-4 phosphate as a tumor vascular-targeting agent: early effects in tumors and normal tissues. Cancer Res 1999;59:1626±1634. [33] Wood L, Palmer M, Hewitt J, et al. Results of a phase III, doubleblind, placebo-controlled trial of megestrol acetate modulation of Pglycoprotein-mediated drug resistance in the ®rst-line management of small-cell lung carcinoma. Br J Cancer 1998;77:627±631. [34] Zanelli GD, Fowler JF. The measurement of blood perfusion in experimental tumors by uptake of 86Rb. Cancer Res 1974;34:1451± 1456.