Combining Antiangiogenic Agents with Metronomic Chemotherapy Enhances Efficacy against Latestage Pancreatic Islet Carcinomas in Mice

G. BERGERS*^ AND D. HANAHANft^ *Department of Neurological Surgery and fDepartment of Biochemistry & Biophysics, Comprehensive fDiabetes and fCancer Centers, University of California, San Francisco, California 94143

Angiogenesis, the formation of new blood vessels, is essential for most tumors to expand (Hanahan and Folk-man 1996). Conversely, inhibition of this complex process can demonstrably restrict tumor growth and even elicit tumor shrinkage or regression in animal models (Folkman 2000). Thus, it is not surprising that widespread efforts in industry and academia have focused on the generation of antiangiogenic drugs that target the endothelial cells, the structural constituents of the vascular system. Antiangiogenic therapy takes advantage of the fact that tumor endothelial cells are chronically activated to proliferate and migrate in the course of assembling a neovasculature, in contrast to the largely quiescent endothelium found in the vessels of normal organs. Most of the current angiogenesis inhibitors were identified by their ability to block proliferation or migration of cultured tumor endothelial cells, typically resulting in en-dothelial cell apoptosis.

The concept of targeting the tumor vasculature as a means to disrupt tumors contrasts with conventional treatment strategies, which typically seek to kill as many proliferating tumor cells as possible by treating tumor-bearing individuals with maximal tolerated doses (MTD) of ionizing radiation or cytotoxic agents that damage DNA or disrupt microtubule functions during mitosis (Kerbel et al. 2002). Such harsh treatment protocols must be alternated with treatment-free periods to permit recovery of normal cells such as bone marrow progenitors and gut mucosal cells. Although often initially efficacious in the form of clinical remissions, such MTD therapies typically elicit recurrent tumors that are insensitive to that therapeutic.

Recently, a new twist on chemotherapeutic trial design has been developed by two groups (Browder et al. 2000; Klement et al. 2000; Kerbel et al. 2002), who discovered that chemotherapeutics, when given in much different dosing schedules, can act as antiangiogenic agents, even in tumors that are resistant to the particular drug. Both the Folkman and Kerbel laboratories showed that by altering the dosing regimen to one of regular inoculations without rest periods at lower (1/3-1/10 of MTD) concentrations, so-called 'metronomic' or antiangiogenic chemotherapy dosing, traditional cytotoxic drugs could produce antian-giogenic effects in xenotransplant models, even against drug-resistant tumors (Browder et al. 2000; Hanahan et al. 2000; Klement et al. 2000). How is such an effect possible? What had been historically ignored with chemotherapeutic strategies was the possibility that proliferating tumor endothelial cells could be killed by cytotoxic drugs. Cytotoxic killing was not evident because their slower doubling times and intact DNA damage response system rendered angiogenic endothelial cells relatively insensitive to episodic high-dose treatment, since they could arrest growth and repair the damage. By increasing the frequency and demanding regularity of treatment, the slowly proliferating endothelial cells could not so readily repair themselves, as they apparently do during the traditional rest periods, resulting in endothelial apop-tosis and impaired angiogenesis. Elimination of the rest periods obligated reduction in the doses to avoid myelo-suppression, which had the added effect of reducing (in some regimens) other toxic side effects. Kerbel and colleagues further demonstrated that the efficacy of a metronomic chemotherapy could be enhanced by combining it with vascular endothelial growth factor (VEGF) receptor-2-blocking antibodies (VEGFRI). The rationale was that the effects of metronomic chemotherapy on activated en-dothelial cells might be selectively amplified in the presence of drugs that target survival mechanisms of en-dothelial cells (Klement et al. 2000). VEGF is known both to stimulate angiogenesis and to prevent endothelial cell apoptosis (Ferrara and Alitalo 1999). Browder and colleagues showed that metronomic chemotherapy could also be beneficially combined with another antiangio-genic agent, TNP-470. Both groups argued that the en-dothelial cells, being genetically stable, would not prove so susceptible to acquired drug resistance, thus providing a means to finesse the otherwise inevitable drug resistance and relapse that result from chemotherapy.

Motivated by these results, we have asked how metronomic chemotherapy performs in combination with targeted antiangiogenic agents in a genetically engineered mouse model (GEM) of multistep pancreatic islet carcinogenesis, the Rip1Tag2 mouse model. We present here trials of a cytotoxic drug (vinblastine or cyclophos-phamide) in a low-dose metronomic regimen, both alone and in combination with a VEGF receptor inhibitor, SU5416, or a broad-spectrum matrix metalloprotease (MMP) inhibitor, BB-94. These combinatorial trials were motivated by both previously published and new data indicating that each class of agent (metronomic chemother-apeutic, VEGF-R inhibitor, and MMP inhibitor) was efficacious when used to treat early-stage disease, but poorly effective in treating well-established solid tumors in this model. Our postulate was that combinatorial therapy might provide an inroad into such otherwise refractory large tumors, and the results support the potential of this strategy for treating end-stage disease.

THE RIP1TAG2 MOUSE MODEL OF MULTISTAGE CARCINOGENESIS ALLOWS DRUG EVALUATION AT DISCRETE STAGES OF TUMOR PROGRESSION

Translation of preclinical results using animal models to clinical trials have typically proved challenging, both in extrapolating experimental design and in interpreting the predictive value of the animal data. In particular, traditional xenograft mouse models for experimental therapeutic trials of anticancer agents have a spotty history in their predictive value. Much has been attributed to the differences in pharmacokinetics consequent to distinctive mouse and human metabolism. Another profound difference, often unmentioned, is that traditional xenotrans-plant models involve cultured tumor cells that are inoculated into different sites, most frequently subcutaneous, where tumor cells assemble into nodules and grow. In contrast, GEM carrying oncogenes or disruptions in tumor suppressors can develop tumors de novo, much like human cancers, originating out of once-normal cells in their natural tissue microenvironments, typically in multistep pathways. These models are providing new insight into mechanisms of carcinogenesis, and some are already demonstrating potential value in evaluating targeted therapies for different types of cancer. Among such GEM, the Rip1Tag2 model (Hanahan 1985) has proved to be particularly amenable toward investigation of the efficacy of candidate drugs and treatment regimens at distinct stages of cancer progression.

Rip1Tag2 mice express the SV40 T-antigen oncogenes under the control of the rat insulin promoter in the p cells populating approximately 400 islets of Langerhans in the pancreas; the mice consequently develop islet cell carcinomas in a multistep pathway. Although oncogene expression ensues in the embryo, morphologically normal islets are formed. Aberrant hyperproliferation starts at 3-4 weeks of age, producing hyperplastic and dysplastic islets, initially with a quiescent vasculature. Then, at 5-7 weeks of age, angiogenesis is switched on in a subset of these dysplastic islets from which solid tumors, encapsulated or invasive, arise as soon as the mice reach an age of about 9-10 weeks (Fig. 1). Most of the transgenic mice die between 13 and 14 weeks of age as a combined consequence of tumor burden and hyperinsulinemia (Fig. 1).

ANTIANGIOGENIC AGENTS SHOW STAGE-SPECIFIC EFFICACY DEPENDENT ON DISEASE PROGRESSION

We have demonstrated the utility of the Rip1Tag2 model as a platform for experimental therapeutic trials (Parangi et al. 1996; Bergers et al. 1999, 2000), taking advantage of the multistage pathway of tumorigenesis to establish three distinctive trial protocols. These trials are designed to investigate whether specific inhibitors can (1) prevent the angiogenic switch in premalignant lesions (prevention trial, 'PT'), (2) intervene in the rapid expansion of small tumors (intervention trial, 'IT'), or (3) regress or stabilize bulky, end-stage tumors (regression trial, 'RT') (Fig. 1). In the PT, treatment commences at 5 weeks of age, when angiogenic switching in hyperplastic islets begins, and continues to 10.5 weeks. We then physically isolate and count angiogenic islets and nascent tumors, which can be distinguished from nonangiogenic islets by their hemorrhagic nature, which produces red-colored nodules. In the IT, treatment is started later, at 10 weeks, when solid tumors are just forming, and continues until 13.5 weeks of age. At this time mice are moribund with substantial tumor burden. Tumors are dissected and tumor number and cumulative volumes (burden) are determined. The same procedure is undertaken in the RT,

Prevention Trial Intervention Trial Regression Trial 5-10.5 wks 10-13.5wks 12-16wks

Figure 1. Overview of tumor progression and experimental therapeutic trials in the Rip1Tag2 model of pancreatic islet carcinogenesis. Expression of the SV40-T antigens in pancreatic ß cells of the islet of Langerhans in Rip1Tag2 mice elicits a multistage pathway beginning as normal islets that progress to hyperplasic islets, dysplasic islets, angiogenic (dysplastic) islets, and then to islet tumors. Based on the synchronicity of progression as a function of age, three therapeutic trials were designed to target distinct stages of disease progression. The three trials aimed to prevent angiogenesis in early lesions (prevention), to reduce growth of small tumors (intervention) or to regress bulky end-stage tumors (regression).

only the trial starts at 12 weeks of age, when mice carry a substantial tumor burden and continues until the mice die at 13.5-14 weeks, or until a defined endpoint of 16 weeks, if the treatment is able to extend life span.

In the course of evaluating a series of candidate angiogenesis inhibitors using these three trial designs, a provocative concept emerged, which is that certain an-tiangiogenic inhibitors were found to be most effective in treating early-to-mid-stage disease, whereas others were more effective on late-stage disease (Bergers et al. 1999). Among the inhibitors of the first group, two classes of synthetic antiangiogenic agents are shown in Figure 2: (1) SU5416, an inhibitor of the two receptor tyrosine kinases, VEGF-R1 and -R2, that signal in response to the angiogenic growth factor VEGF-A and (2) two broad-spectrum metalloprotease inhibitors (MMPI), BB-94 (batimastat, British Biotech.) and BAY 12-9566 (Bayer Corporation), which inhibit a variety of matrix metalloproteinases, including MMP-2, MMP-9, MMP-3, and MMP-8 (Sternlicht and Bergers 2000). We previously reported on BB-94 in all three trials (Bergers et al. 1999) and on SU-5416 in prevention and intervention trials (Bergers et al. 2000). The regression trials of SU5416 are presented here for the first time, as are all three trials of BAY 12-9566. Regimen and dose schedules are summarized in Table 1.

The trials of SU5416 demonstrated that this agent was highly effective in blocking the initial angiogenic switch induced in the previously quiescent islet capillaries (Fig. 2A, PT) and at impairing the explosive growth of nascent tumors (Fig. 2A, IT) indicative of a crucial role for VEGF signaling in islet tumorigenesis (Fig. 2A, PT). Consistent with this conclusion, we recently showed that targeted deletion of the VEGF-A ligand gene in the pancreatic p cells resulted in an almost complete block of angiogenic switching in progenitor hyperplastic/dysplastic islets (In-oue et al. 2002). Furthermore, tumor growth was severely affected in mice whose oncogene-expressing islets lacked VEGF. Most tumors were small with an impaired angio-genic phenotype, containing avascular, necrotic cores, with viable tumor cells present only in a thin layer at the periphery. These results support the notion that tumor growth indeed is angiogenesis-dependent. One question that could not be answered by the cell-type-specific geneknockout approach is whether VEGF signaling remains a critical factor at late stages of disease progression. Pharmacological intervention, in contrast, has allowed this question to be addressed. When SU5416 therapy was initiated against early-stage tumors in an IT, it was effective, reducing tumor burden by ~70%. In marked contrast, this VEGF receptor inhibitor had little effect on end-stage tumors in a RT (Fig. 2A, RT). Although SU5416 was able to achieve a modest extension of life span (from 14 to16 weeks), it was not able to elicit regression of tumor mass, or even to produce 'stable disease'.

We similarly tested the two MMPIs, BB-94 and BAY12-9566, at distinct stages of tumorigenesis, and found that each produced an efficacy profile similar to SU5416. Both were effective at reducing the incidence of angiogenic switching by ~50% (Fig. 2B, PT), and each restrained growth of small solid tumors by ~70-80%

(Fig. 2B, IT) As with SU5416, neither MMPI was able to regress bulky end-stage tumors (Fig. 2B, RT), a result that is likely relevant to the experiences of MMP-I in clinical trials. Three MMPI, including a relative of BB-94 (Marimastat) and BAY 12-9566 were withdrawn from clinical trials due to disappointing responses; (Sternlicht and Bergers 2000; Coussens et al. 2002).

Notably, the similar stage-specific efficacy patterns seen with a VEGFR inhibitor and two MMPI is likely not coincidental. MMP-9, a target protease of these MMPI, is now appreciated to be a regulator of VEGF-A bioavail-ability in the pancreatic islets. The data support a model whereby MMP-9 produced by leucocytes infiltrating dys-plastic islets is released and activated concomitant with the angiogenic switch, thereby mobilizing constitutively expressed but physically sequestered VEGFA, which in turn triggers an angiogenic response in the islet capillaries (Bergers et al. 2000). MMP-9 deficient-Rip1-Tag2 mice, as well as mice that are treated early with BB-94 show a similarly reduced frequency of angiogenic switching and comparable reductions in tumor growth. Interestingly, although MMPs were originally identified as molecules that affect invasion, neither genetic nor functional ablation of MMPs prevented the eventual his-tologic progression to invasive carcinomas.

In summary, both a VEGF-RI and two MMPIs were efficacious in treating premalignant disease and small tumors, but none was able to regress bulky end-stage tumors. This later stage is the situation, however, in which clinical trials are typically performed, and the state where the need for better cancer therapies is particularly acute. Most patients who enter clinical trials have bulky disease, having failed multiple conventional therapies. Our results predict that MMPI as well as both VEGF and VEGF-R inhibitors will not prove generally efficacious against well-established solid tumors, if one accepts the proposition that the Rip1Tag2 model can serve as a general prototype for therapeutic strategies that target the tumor vas-culature. Certainly the results with this model are consistent with the poor responses of MMPI in the clinic and predict that VEGF/VEGFR inhibitors will similarly prove most effective against early-stage disease in human trials. Notably, the VEGFR inhibitor we tested here, SU5416, has recently been withdrawn from clinical trials (http://www.pharmaciaoncology.com/popup.asp?ptype= ct_SU5416&link=CT) based on a number of unelabo-rated considerations, most likely including poor responses against late-stage disease.

Given the imperative to develop drugs that are effective against late-stage tumors, an increasingly common strategy is to combine experimental agents such as these with conventional cytotoxic drugs delivered in MTD regimens. Indeed, such trials were ongoing both with BAY 12-9655 and SU5416, apparently without great benefit. We became interested in further evaluating the prospects of metronomic chemotherapy, based on the results and considerations discussed above, and describe below trials of metronomic chemotherapy, involving two cytotoxic drugs delivered alone, and in combination with the an-giogenesis inhibitors just discussed.

Figure 2. The VEGF-R inhibitor SU5416 and two MMP inhibitors,s BB-94 and BAY12-9566, show stage-specific efficacy in the RIP1Tag2 model of carcinogenesis. Rip1Tag2 mice were treated at distinct stages of disease progression (Prevention, Intervention, Regression Trials). The average numbers ± s.e.m. of angiogenic islets of control (PBS) and treated mice are shown in the prevention trial, whereas average tumor burden ± s.e.m. of control and treated animals is shown in the intervention and regression trial. Cohorts of 5-21 animals were used. Statistical analysis was performed with a two-tailed, unpaired Mann-Whitney test, comparing experimental groups to PBS-injected control mice. Tumor burdens of experimental groups in the regression trial were compared to the tumor burden of 12-week-old Rip1Tag2 mice (Ti). p values less than 0.1 are considered statistically significant. (A) SU5416 was very efficacious in the prevention and intervention trial but had only modest effects on tumor reduction in the regression trial (p values of SU5416 PT= 0.0001; SU5416 IT= 0.0124; SU5416 RT= 0.1807 ). (B) The two MMPIs have a comparable efficacy profile to SU5416, being efficacious in early but not late stages (p-values of BB-94 PT = 0.0007, BAY12-9566=0.021; BB-94 IT= 0.0002, BAY12-9566 IT= 0.0003; BB-94 RT= 0.2029; BAY 12-9566 RT= 0.34).

Figure 2. The VEGF-R inhibitor SU5416 and two MMP inhibitors,s BB-94 and BAY12-9566, show stage-specific efficacy in the RIP1Tag2 model of carcinogenesis. Rip1Tag2 mice were treated at distinct stages of disease progression (Prevention, Intervention, Regression Trials). The average numbers ± s.e.m. of angiogenic islets of control (PBS) and treated mice are shown in the prevention trial, whereas average tumor burden ± s.e.m. of control and treated animals is shown in the intervention and regression trial. Cohorts of 5-21 animals were used. Statistical analysis was performed with a two-tailed, unpaired Mann-Whitney test, comparing experimental groups to PBS-injected control mice. Tumor burdens of experimental groups in the regression trial were compared to the tumor burden of 12-week-old Rip1Tag2 mice (Ti). p values less than 0.1 are considered statistically significant. (A) SU5416 was very efficacious in the prevention and intervention trial but had only modest effects on tumor reduction in the regression trial (p values of SU5416 PT= 0.0001; SU5416 IT= 0.0124; SU5416 RT= 0.1807 ). (B) The two MMPIs have a comparable efficacy profile to SU5416, being efficacious in early but not late stages (p-values of BB-94 PT = 0.0007, BAY12-9566=0.021; BB-94 IT= 0.0002, BAY12-9566 IT= 0.0003; BB-94 RT= 0.2029; BAY 12-9566 RT= 0.34).

Table 1: Cytotoxic and Antiangiogenic Dosing Regimens

Drug Dosing and Administration Route Reference

Drug Dosing and Administration Route Reference

Table 1: Cytotoxic and Antiangiogenic Dosing Regimens

Vinblastine MD

0.75 mg/m2

i.p.

bolus

Klement et al. (2000)

1.5 mg/m2

i.p.

3rd day

Cyclophosphamide

10 mg/kg

drinking water

Q.D.

Man et. al. (2002)

(CTX) MD

SU5416

50-100 mg/kg

s.c.

Q.O.D.

Bergers et al. (2000)

BAY-12-9566

100 mg/kg

p.o.

Q.D.

Bayer Corp., (pers. comm.)

BB-94

30 mg/kg

i.p.

Q.D.

Bergers et al. (2000)

RipTag2 mice were treated with one of two cytotoxic drugs and/or one of two angiogenesis inhibitors as indicated above. Treated and control mice were sacrificed at the end of each trial and pancreases collected for gross and/or histopathology, in particular, quantitation of the angiogenic switching frequency or tumor burden. (MD, metronomic dosing; i.p., intraperitoneal; s.c., subcutaneous; p.o., oral; Q.D., daily). To evaluate a prevention trial (5-10.5-wk treatment), islets were isolated by retrograde perfusion and collagenase P digestion of the pancreas,and sorted under darkfield illumination in a dissecting microscope; angiogenic islets were scored by their red flecked color, and counted. For intervention trials (10-13.5-wk treatment) and regression trials (12-16-wk treatment), tumors were physically excised using fine surgery tools and a dissecting microscope. Tumor burden of a mouse was calculated by summing the tumor volume (volume [mm3] = 0.52 x [width]2 x [length] of every pancreatic tumor >1 mm in diameter. The details of the methods can be found in Bergers et al. (1999, 2000).

RipTag2 mice were treated with one of two cytotoxic drugs and/or one of two angiogenesis inhibitors as indicated above. Treated and control mice were sacrificed at the end of each trial and pancreases collected for gross and/or histopathology, in particular, quantitation of the angiogenic switching frequency or tumor burden. (MD, metronomic dosing; i.p., intraperitoneal; s.c., subcutaneous; p.o., oral; Q.D., daily). To evaluate a prevention trial (5-10.5-wk treatment), islets were isolated by retrograde perfusion and collagenase P digestion of the pancreas,and sorted under darkfield illumination in a dissecting microscope; angiogenic islets were scored by their red flecked color, and counted. For intervention trials (10-13.5-wk treatment) and regression trials (12-16-wk treatment), tumors were physically excised using fine surgery tools and a dissecting microscope. Tumor burden of a mouse was calculated by summing the tumor volume (volume [mm3] = 0.52 x [width]2 x [length] of every pancreatic tumor >1 mm in diameter. The details of the methods can be found in Bergers et al. (1999, 2000).

ASSESSING THE PROSPECTS OF METRONOMIC CHEMOTHERAPY AGAINST END-STAGE DISEASE

We sought to improve the efficacy of the MMP-I BB-94 and the VEGFR-I SU5416 in late-stage pancreatic islet tumors by combining them with the cytotoxic agents vinblastine and cyclophosphamide (CTX) delivered in metronomic chemotherapy regimens. We asked whether a combinatorial therapy was able to reduce tumor burden and extend life span in a regression trial. Thus, we treated 12-week old mice that bear substantial initial tumor burden (Ti, Fig. 3A,B) and will live no longer than another 2 weeks, and assessed survival and tumor burden. The trials were arbitrarily ended after 4 weeks of treatment (at 16 weeks) unless the mice died sooner, as is the case for untreated controls. The cytotoxic drugs were supplied in a metronomic regime alone, or in combination with SU5416 or BB-94. The trial results are summarized in Figure 3 (A and B); Table 1 indicates regimen and dosing schedules. The single-agent trials indicate little or no efficacy for vinblastine, whereas CTX had demonstrable efficacy when delivered continuously in the drinking water. Development of the oral dosing protocol for mice by the Kerbel group was motivated by the routine oral delivery of CTX in humans (Kerbel et al. 2002); this oral dosing protocol in mice has demonstrated efficacy both in traditional xenotransplant models and in the Rip1-Tag2 model (Man et al. 2002).

As shown and discussed above, neither BB-94 nor SU5416 was able to stop continuing tumor growth in a regression trial as a single agent. In contrast, the combination trials with vinblastine showed benefit, producing an apparent condition of stable disease, where continuing growth was restricted, but the tumors did not regress. Remarkably, the combination of oral CTX with both drugs reduced tumor burden by ~50-60% compared to the initial tumor burden at 12 weeks, indicative of tumor regression (Figs. 4 and 5). CTX alone had potential to stabilize disease in some individuals, but 50% of the mice were dead at the end of the regression trial. Thus, the metronomic regimen of oral CTX produced clear and convincing combinatorial efficacy, given that none of the three agents (CTX, BB-94, or SU-5416) was able to significantly regress late-stage tumors alone.

We also assessed survival in each cohort of treated or control mice. Metronomic regimens involving CTX or vinblastine alone resulted in about 50% survival at 16 weeks, whereas all untreated controls were already dead by this time (Figs. 4 and 5). BB-94 treatment, either alone or in combination with vinblastine, also resulted in a median survival rate of 50% at 16 weeks (Fig. 5A). In contrast, combined administration of CTX and BB-94 significantly improved survival and reduced tumor growth in comparison to single treatments, and therefore proved to be the best combination of all variations tested here (Fig. 5B). Interestingly, tumor burden and survival did not always correlate. Although SU5416 + CTX combination therapy reduced tumor burden, it did not improve survival compared to SU5416 treatment as a single agent (Fig. 4B). The other combination, of SU5416 with vin-blastine, caused substantial toxicity, leaving only 30% of the mice alive at 16 weeks (Fig. 4A), and thus was worse than either single agent alone. We and other investigators have noticed that SU5416 causes hemorrhage formation in the lungs due to uncharacterized side effects that were seemingly exacerbated in the presence of vinblastine. These results suggest that the benefits of combinatorial therapies involving cytotoxic drugs delivered in metronomic regimes may be combination- and regimen-specific, because we have been able to improve efficacy of an MMPI but not of SU-5416, when considering survival as the factor.

CONCLUSIONS

We have shown using a prototypical model of multistage tumorigenesis that combinatorial therapies involving antiangiogenic agents and metronomic chemotherapy (itself demonstrably antiangiogenic) can produce tumor regression and increased survival in mice with late-stage pancreatic islet carcinomas. These results are encouraging, given our sobering observation that none of the single agents (metronomic chemotherapeutic, VEGF-RI, or MMPI) was effective alone against late-stage disease,

Figure 3. Metronomic chemotherapy enhances the efficacy of VEGF-RI and MMPI in a late-stage regression trial in Rip1Tag2 mice. Starting at 12 weeks, mice with substantial tumor burden were treated for 4 weeks with SU5416 (A) or BB-94 (B), alone, or in combination with metronomic, low-dose cyclophosphamide (CTX) or Vinblastine (Vinbl). Average tumor burden ± s.e.m. of 12-week-old control mice (Ti) and 16-wk old treated animals are shown in the regression trial. Cohorts of 5-21 animals were used. Statistical analysis was performed with a two-tailed, unpaired Mann-Whitney test comparing experimental groups to the initial tumor burden of 12-week-old Rip1Tag2 mice (Ti). p values less than 0.1 are considered statistically significant. SU5416 (A) or BB-94 (B) alone did not stop tumor growth, whereas SU5416 or BB-94 + vinblastine produced "stable disease," and SU5416 or BB94 + CTX elicited tumor regression (p-values of SU5416= 0.999; SU5416/CTX= 0.048; SU5416/Vinbl.= 0.7573; BB-94= 0.2029; BB-94/CTX= 0.0026, BB-94/Vinbl.= 0.5113).

Figure 3. Metronomic chemotherapy enhances the efficacy of VEGF-RI and MMPI in a late-stage regression trial in Rip1Tag2 mice. Starting at 12 weeks, mice with substantial tumor burden were treated for 4 weeks with SU5416 (A) or BB-94 (B), alone, or in combination with metronomic, low-dose cyclophosphamide (CTX) or Vinblastine (Vinbl). Average tumor burden ± s.e.m. of 12-week-old control mice (Ti) and 16-wk old treated animals are shown in the regression trial. Cohorts of 5-21 animals were used. Statistical analysis was performed with a two-tailed, unpaired Mann-Whitney test comparing experimental groups to the initial tumor burden of 12-week-old Rip1Tag2 mice (Ti). p values less than 0.1 are considered statistically significant. SU5416 (A) or BB-94 (B) alone did not stop tumor growth, whereas SU5416 or BB-94 + vinblastine produced "stable disease," and SU5416 or BB94 + CTX elicited tumor regression (p-values of SU5416= 0.999; SU5416/CTX= 0.048; SU5416/Vinbl.= 0.7573; BB-94= 0.2029; BB-94/CTX= 0.0026, BB-94/Vinbl.= 0.5113).

congruent with disappointing clinical trials with the latter two classes of agents. Thus, we are hopeful that similar benefits could be realized in the clinic, in particular by combining MMPI or VEGF-RI with low-dose metronomic chemotherapy rendering effective drugs otherwise destined for failure. Although we have not yet comprehensively evaluated MTD regimens in similar combination with MMPI or VEGF-RI, we suspect that metronomic low-dose regimens will have demonstrable benefit over traditional regimes, particularly when combined with other agents intended to target endothelial cells and thereby imhibit angiogenesis. There is a further rationale in favor of metronomic dosing designs in combination with MMPIs. It has recently been shown that pharmacological inhibition (or genetic ablation) of MMP-9 impairs recovery from high-dose chemotherapy, by restricting mobilization of bone marrow progenitors (Heissig et al. 2002); repopulation of the hematopoeitic system from the bone marrow is crucial for resolving myleosuppression and obviating susceptibility to acute infection consequent to high-dose chemotherapy. Interestingly, the combination of metronomic low-dose CTX with the MMPI BB-94

Figure 4. Kaplan-Meyer survival curves in Rip1Tag2 mice submitted to the VEGFRI SU5416 alone or in combination with metronomic chemotherapy. Survival curves of Rip1Tag2 mice treated with SU5416 alone or in combination with vinblastine (A) or CTX (B) are shown over a period of 32 days in the regression trial. All untreated mice were dead 2 weeks after the trial started. Cohorts of 6-18 mice were used. There was no survival benefit of SU 5416 combined with either Vinbl. (A) or CTX (B) in the 4-week treatment period.

Figure 4. Kaplan-Meyer survival curves in Rip1Tag2 mice submitted to the VEGFRI SU5416 alone or in combination with metronomic chemotherapy. Survival curves of Rip1Tag2 mice treated with SU5416 alone or in combination with vinblastine (A) or CTX (B) are shown over a period of 32 days in the regression trial. All untreated mice were dead 2 weeks after the trial started. Cohorts of 6-18 mice were used. There was no survival benefit of SU 5416 combined with either Vinbl. (A) or CTX (B) in the 4-week treatment period.

had the best survival benefit in our study, suggesting that this regimen was not impairing homeostasis of the hematopoeitic system.

The results presented here give further encouragement to combinatorial trial designs if one recalls that we demonstrated enhanced efficacy in a relatively short treatment period of 4 weeks, in a very stringent multifocal model, where physiological stresses of hyperinsuline-mia exacerbate local effects of tumor burden. Rip1Tag2 mice become hyperinsulinemic at endstage, and if the tumors do not stabilize and regress fast enough, the mice can develop acute hypoglycemic shock and die. This may explain why a subset of Rip1Tag2 mice in every treatment cohort typically drop out near the beginning of ev ery regression trial. Taking the unusual characteristic of this model into consideration, we expect that metronomic low-dose CTX combined with VEGF/VEGFR inhibitors will prove to confer better survival if tumor-bearing mice are treated for longer periods of time, particularly for other types of cancer where the physiological burden of hyperinsulinemia is not a factor, or perhaps in islet carcinomas when hyperinsulinemia is managed pharmacologically. This argument is supported by the findings of the Kerbel group, who initially observed continuing tumor growth in trials involving oral CTX plus an anti-VEGFR antibody in xenotransplant mice, before disease stabilization ensued after ~60 days of treatment (Man et al. 2002). Furthermore, other VEGFR-I may not show some of the

Figure 5. Kaplan-Meyer survival curves in Rip1Tag2 mice submitted to the MMPI BB-94 alone or in combination with metronomic chemotherapy. Survival curves of Rip1Tag2 mice treated with BB-94 alone or in combination with Vinblastine (A) or CTX (B) are shown over a period of 32 days in the regression trial. All untreated mice were dead 2 weeks after the trial started. Cohorts of 6-12 mice were used. BB-94 combined with Vinblastine (A) had no survival benefit, but when combined with CTX (B), survival was significantly improved.

Figure 5. Kaplan-Meyer survival curves in Rip1Tag2 mice submitted to the MMPI BB-94 alone or in combination with metronomic chemotherapy. Survival curves of Rip1Tag2 mice treated with BB-94 alone or in combination with Vinblastine (A) or CTX (B) are shown over a period of 32 days in the regression trial. All untreated mice were dead 2 weeks after the trial started. Cohorts of 6-12 mice were used. BB-94 combined with Vinblastine (A) had no survival benefit, but when combined with CTX (B), survival was significantly improved.

toxicities associated with SU5416, some of which are suspected to reflect its particular chemistry rather than its actions against the VEGF receptors. In view of these considerations, we conclude there is good reason to continue investigating combinations of metronomic low-dose chemotherapy with VEGF/VEGFR inhibitors.

In summary, the results presented herein support the notion that low-dose metronomic chemotherapy in appropriate combinations can substantially improve current treatment modalities and thereby render otherwise poorly effective angiogenesis inhibitors demonstrably beneficial. We demonstrate for the first time that MMPIs can be efficacious in late-stage disease when used in conjunction with chemotherapy if the metronomic dosing schedule is applied. Thus, our data beg the proposition that MMPIs could be resurrected clinically as cancer therapeutics if they were tested in combination with metronomic chemotherapy; in contrast, we have reservations about MTD chemotherapy and suggest it will not demonstrate such combinatorial efficacy and low toxicity against well-established human tumors. We hope this proposition will be tested as well. Beyond the potential for improving efficacy of MMPIs, the work reported here, as well as the previous reports from the Folkman and Kerbel labs (Browder et al. 2000; Klement et al. 2000, 2002; Man et al. 2002), indicates that other classes of angiogenesis inhibitors will benefit from combination with metronomic, low-dose cytototic therapy. Perhaps such combination therapies will prove able to destabilize and slowly regress well-established human tumors much as they can in the de novo islet carcinomas of the pancreas in Rip1Tag2 mice, producing a chronic, manageable disease in some cases, and potentially resolving tumors altogether in others. More generally, one might ask whether other distinct classes of targeted therapies (e.g. the receptor tyrosine ki-nase inhibitors Herceptin and Aressa) that are currently used with conventional MTD chemotherapy might also be beneficially combined with metronomic chemotherapy, thereby targeting both tumor cells and their supporting vasculature. Indeed, Herceptin has recently been reported to be antiangiogenic (Izumi et al. 2002). Such possibilities deserve further consideration, as modified trial designs involving the combination of new target-selective drugs with metronomic low-dose cytotoxic therapy might provide inroads into otherwise intractable disease states.

ACKNOWLEDGMENTS

We thank SUGEN Inc., South San Francisco, California, for SU5416; British Biotech Pharmaceuticals, Oxford, United Kingdom, for BB-94 (batimastat); and Bayer Corp. West Haven, Connecticut, for BAY-12-9655. We also thank E. Soliven and J. Imperio for excellent technical assistance, and A. McMillan for help with statistical analysis. This work was funded by grants from the National Cancer Institute and by Bill Bowes, whom we thank for his visionary support and encouragement. Correspondence may be addressed to either author.

REFERENCES

Bergers G., Javaherian K., Lo K.-M., Folkman J., and Hanahan D. 1999. Effects of angiogenesis inhibitors on multistage carcinogenesis in mice. Science 284: 808.

Bergers G., Brekken R., McMahon J., Vu T., Itoh T., Tamaki K., Tanzawa K., Thorpe P., Itohara S., Werb Z., and Hanahan D. 2000. Matrix Metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat. Cell Biol. 2: 737.

Browder T., Butterfield C.E., Kräling B.M., Shi B., Marshall B., O'Reilly M.S., and Folkman J. 2000. Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drug-resistant cancer. Cancer Res. 60: 1878.

Coussens L.M., Fingleton B., and Matrisian M. 2002. Matrix metalloproteinase inhibitors and cancer: Trials and tribulations. Science 295: 2387.

Ferrara N. and Alitalo K. 1999. Clinical applications of angiogenic growth factors and their inhibitors. Nat. Med. 5: 1359.

Folkman J. 2000. Tumor angiogenesis. In Cancer medicine, p. 2546. 5th edition (ed. J.F. Holland et al.), B.C. Decker, Hamilton, Ontario.

Folkman J. 2000. 'Tumor angiogenesis', 5th ed., in H. e. al., (ed.), Cancer Medicine, p. 2546. B.C. Decker, Hamilton, Ontario.

Hanahan D. 1985. Heritable formation of pancreatic ß-cell tumors in transgenic mice harboring recombinant insulin/simian virus 40 oncogenes. Nature 315: 115.

Hanahan D., Bergers G., and Bergsland E. 2000. Less is more, regularly: Metronomic dosing of cytotoxic drugs can target tumor angiogenesis in mice. J. Clin. Invest. 105: 1045.

Hanahan D. and Folkman J. 1996. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86: 353.

Heissig B., Hattori K., Dias S., Friedrich M., Ferris B., Hackett N., Crystal R., Besmer P., Lyden D., Moore M., Werb Z., and Rafii S. 2002. Recruitment of stem and progenitor cells from the bone marrow niche requires mmp-9 mediated release of kit-ligand. Cell 109: 625.

Inoue M., Hager J., Ferrara N., Miller D., Basilico C., Gerber H.-P., and Hanahan D. 2002. VEGF- has a critical non-redundant role in angiogenic switching and pancreatic ß-cell carcinogen-esis. Cancer Cell 1: 193.

Izumi Y., Xu L., di Tomaso E., Fukumura D., and Jain R. 2002. Tumour biology: Herceptin acts as an antiangiogenic cocktail. Nature 416: 279.

Kerbel R.S., Klement G., Pritchard K.I., and Kamen B. 2002. Continuous low-dose anti-angiogenic/metronomic chemotherapy: From the research laboratory into oncologic clinic. Ann. Oncol. 13: 12.

Klement G., Baruchel S., Rak J., Man S., Clark K., Hicklin D.J., Bohlen P., and Kerbel R.S. 2000. Chronically sustained regressions of human tumor xenografts in the absence of overt toxicity by continuous low-dose vinblastine and anti-VEGF receptor-2 antibody therapy. J. Clinical Invest. 105: R15.

Klement G., Huang P., Mayer B., Green S.K., Man S., Bohlen P., Hicklion D., and Kerbel R.S. 2002. Differences in therapeutic indexes of combination metronomic chemotherapy and an anti-VEGFR-2 antibody in multidrug-resistant human breast cancer xenografts. Clin. Cancer Res. 8: 221.

Man S., Bocci G., Francia G., Green S.K., Jothy S., Bergers G., Hanahan D., Bohlien P., Hicklin D., and Kerbel R.S. 2002. Anti-tumor effects in mice low-dose (metronomic) cyclophos-phamide administered continously through the drinking water. Cancer Res. 62: 2731.

Parangi S., O'Reilly M., Christofori G., Holmgren L., Grosfeld J., Folkman J., and Hanahan D. 1996. Antiangiogenic therapy of transgenic mice impairs de novo tumor growth. Proc. Natl. Acad. Sci. 93: 2002.

Sternlicht M. and Bergers G. 2000. Matrix metalloproteinases as emerging targets in cancer: Status and prospects. Emerg. Ther. Targets 4: 609.

Diabetes Sustenance

Diabetes Sustenance

Get All The Support And Guidance You Need To Be A Success At Dealing With Diabetes The Healthy Way. This Book Is One Of The Most Valuable Resources In The World When It Comes To Learning How Nutritional Supplements Can Control Sugar Levels.

Get My Free Ebook


Post a comment