Vascular targeting by EndoTAG™-1 enhances therapeutic efficacy of conventional chemotherapy in lung and pancreatic cancer

Abstract

Cationic lipid complexed paclitaxel (EndoTAG™-1) is a novel vascular targeting agent for the treatment of cancer. Here, the aim was to investigate intratumoral drug distribution after EndoTAG™-1 therapy and analyze the impact of EndoTAG™-1 scheduling on antitumoral efficacy. The therapeutic effect of EndoTAG™-1 in combination with conventional gemcitabine or cisplatin therapy was evaluated in L3.6pl orthotopic pancreatic cancer and a subcutaneous Lewis lung (LLC-1) carcinoma model. Oregon Green paclitaxel encapsulated in cationic liposomes in combination with intravital fluorescence microscopy clearly exhibited delivery of the drug by EndoTAG™-1 to the tumor endothelium, whereas Oregon Green paclitaxel dissolved in cremophor displayed an interstitial distribution pattern. The therapeutic efficacy of EndoTAG™-1 was critically dependent on the application schedule with best therapeutic results using a metronomic rather than a maximum tolerated dose application sequence. The combination of EndoTAG™-1 therapy and cytotoxic chemotherapy significantly enhanced antitumoral efficacy in both tumor models. Interestingly, only EndoTAG™-1 in combination with gemcitabine was able to inhibit the incidence of metastasis in pancreatic cancer. In conclusion, vascular targeting tumor therapy by EndoTAG™-1 combined with standard small molecular chemotherapy results in markedly enhanced antitumoral efficacy. Therefore, this combination represents a promising novel strategy for clinical cancer therapy.

Angiogenesis, the formation of new blood vessels from the endothelium of the existing vasculature, is fundamental in tumor growth, progression and metastasis.1 The complex network of tumor blood microvessels guarantees an adequate supply of tumor cells with nutrients and oxygen and provides efficient drainage of metabolites.

Therapeutic strategies that target and disrupt already formed vessel networks of growing tumors are therefore actively pursued. In contrast with the antiangiogenesis approach, the aim of vascular targeting is to destroy the established tumor vasculature thus causing a rapid and extensive decrease in tumor blood flow, followed by secondary tumor cell death.2, 3 Ligand-directed vascular targeting agents and small molecular tubulin-binding agents have been successfully developed to induce a vascular shutdown of tumor microvessels.4 In addition to these compounds, drug delivery systems are of considerable interest in realizing such a new therapeutic concept: cationic lipid complexes have been described to target angiogenic endothelial cells in tumor preferentially.5–7 This property potentially enables selective delivery of cytotoxic drugs to tumor endothelial cells and thus vascular targeting chemotherapy. We have previously shown that vascular targeting therapy can be realized by paclitaxel or camptothecin encapsulated in cationic lipid complexes.8, 9 Treatment with these liposomal compounds significantly retarded primary tumor growth and delayed metastatic disease by an antivascular mechanism, as shown by intravital microscopy10 or DCE-MRI,11 respectively.

Although antiangiogenic inhibitors and vascular targeting agents can regress primary tumor growth and inhibit tumor metastasis in experimental tumor models, antivascular tumor therapy as monotherapy has failed to provide convincing results in clinical trials. To date, antiangiogenic drugs and vascular targeting agents cannot eradicate tumors completely, and remarkable antitumoral effects can be achieved only by combining antivascular tumor therapy with conventional therapy directly targeting the tumor cell compartment.12

In the present study, we therefore evaluated the therapeutic effect of vascular targeting chemotherapy by EndoTAG™-1 in combination with conventional cytotoxic chemotherapy in experimental lung and pancreatic cancer. As dosing and scheduling of antivascular drugs may critically affect antitumoral efficiency when applied in combination with cytotoxic chemotherapy, we first investigated the impact of EndoTAG™-1 scheduling on antitumoral efficiency, tumor vascularity and tumor cell apoptosis. Moreover, we proved tumor endothelial cells to be the postulated therapeutic target of EndoTAG™-1 therapy by intravital analysis of drug distribution within tumor tissue. In subsequent studies, the antitumoral efficiency of EndoTAG™-1 therapy in combination with conventional chemotherapy was investigated in 2 different tumor models. For treatment of subcutaneous Lewis lung carcinoma (LLC-1) vascular targeting chemotherapy by EndoTAG™-1 was combined with cisplatin therapy. Orthotopically grown human L3.6pl pancreatic tumors were treated by EndoTAG™-1 in combination with gemcitabine.

Material and Methods

EndoTAG™-1 synthesis

1,2 dioleoyl-3-trimethylammonium-propane (DOTAP) and 1,2 dioleoyl-sn-glycero-3-phosphocholine (DOPC) (Avanti Polar Lipids, Alabaster, AL) were used for preparation of cationic liposomes. Paclitaxel (Natural Pharmaceuticals, Beverly, MA) was used for encapsulation into cationic liposomes without further purification. Liposome production was performed on the basis of standard procedures as described in the literature. Briefly, DOTAP, DOPC and paclitaxel were dissolved in ethanol in a molar ratio of 50/45/5. The concentrated ethanol solution was injected under stirring into 10% trehalose solution (Ferro Pfanstiehl Laboratories, Waukegan, IL) to obtain a suspension of multilamellar liposomes. The multilamellar liposomes were extruded through polycarbonate membranes (GE Osmonics, Minnetonka, MN), pore size 200 nm, and the resulting monodisperse monolamellar liposomes were sterile filtrated using MilliPak filters, 220 nm (Millipore, Molsheim, France). The liposomes were put into in glass vials and lyophilized using a Christ Epsilon 2-12D lyophilizer (Osterode, Gemany). Thirty minutes before intravenous injection lyophilized EndoTAG-1 was resuspended in aqua injectabile (B. Braun, Melsungen, Germany). For analysis of intratumoral drug distribution 1% Oregon Green 488 paclitaxel (Invitrogen, Karlsruhe, Germany) was encapsulated in cationic lipid complexes as described above. The fluorescent label on Oregon Green 488 paclitaxel is attached by derivatization of the 7-β-hydroxy group of native paclitaxel, a strategy that permits selective binding of the probe to microtubules. To investigate intratumoral distribution of cationic lipid complexes in L3.6pl pancreatic tumors cationic lipid complexes were spiked by 5% Rhodamin-DOPE (Avanti Polar Lipids, Alabaster, AL) as described in detail earlier.

Animals and tumor models

After approval by the local ethics committee, experiments were performed on male C57/BL6 mice (25–32 g b.w.), male athymic nude mice (NCr-nu, 20–25 g b.w.) or male Syrian golden hamsters (60–70 g b.w.), respectively. All experiments were performed in accordance with the UKCCCR ‘Guidelines for the welfare of animals in experimental neoplasia’.13 All surgical procedures were performed under anaesthesia with ketamine (100 mg/kg b.w. i.p., Ketavet®; Parke-Davis, Berlin, Germany) and xylazine (10 mg/kg b.w. i.p., Rompun®; Bayer, Leverkusen, Germany).

To investigate therapeutic effects on s.c. tumor growth, 8 × 10⁵ cells of the Lewis lung carcinoma (LLC-1) were s.c. injected into the mid-dorsal region of C57/BL6 mice. tumor volumes were calculated from the longer (l) and shorter (w) perpendicular axes and the height (h) of each tumor nodule measured by a calliper. Body weight of animals was measured every other day.

To investigate the antitumoral efficacy in an orthotopic pancreatic cancer model, L3.6pl human pancreatic cancer cells were injected orthotopically in male athymic nude mice (NCr-nu) as described previously.14 Briefly, a small, left abdominal flank incision was made and the spleen was exteriorized. tumor cells (8 × 10⁵ in 40 μl saline) were injected into the subcapsular region of the pancreas just beneath the spleen. Pancreatic tumors were allowed to grow in this orthotopic location for 7 days before the start of treatment. Mice were sacrificed on day 30 after tumor cell injection. Excised pancreatic tumors were weighed and measured. The tumor volume was then calculated using the formula V = π × a × b × / 6, where a, b and c represent the length, width and height of the mass. In addition, metastatic L3.6pl tumor growth was evaluated. For metastases in the liver, macroscopically visible tumor nodules (>1 mm) were noted on the liver surface. Furthermore, enlarged regional (celiac and para-aortic) lymph nodes were recorded. Liver and lymph node tissue were excised and processed to confirm metastases by histology.

Intratumoral distribution of cationic lipid complexes in L3.6pl tumors

Fourteen days after orthotopic injection of L3.6pl tumor cells 4 mg total lipid/kg b.w. of rhodamine-labeled cationic lipid complexes were injected via the lateral tail vein (n = 3). To investigate binding of cationic lipid complexes during early and late tumor angiogenesis further experiments were performed on day 23 after tumor cell injection (i.v. injection of 10 mg/kg rhodamine-labeled cationic lipid complexes, n = 3). Twenty minutes after liposome injection 4 mg/kg b.w. of FITC-labeled lectin from Lycopersicon esculentum (1 mg/ml; Vector Laboratories, San Francisco, CA) were injected as a vascular marker. After FITC-lectin was allowed to circulate for 3 min, the animals were perfused intracardially with 37°C warm 1% paraformaldehyde in 10 mM phosphate buffered saline (PFA-PBS) at a physiological pressure (∼120 mm Hg) for 2 min. tumors were removed and postfixed in 1% PFA-PBS at 4°C for at least 4 hr before leaving them in 20% sucrose in PBS at 4°C overnight for cryoprotection. The tumors were frozen on dry ice in ethanol and stored at −80°C. Frozen sections (20 μm thickness) were collected on warm polylysine-coated slides (Menzel, Braunschweig, Germany) and immediately refrozen. Sections were covered with Vectashield (Vector Laboratories, Peterborough, UK) under coverslips and examined using a Zeiss Axiophot 2 fluorescence microscope (Zeiss, Göttingen, Germany) with specific filtersets for rhodamine and FITC.

Analysis of intratumoral drug distribution

Investigation of intratumoral drug distribution was performed by intravital fluorescence microscopy of A-MEL-3 tumors implanted in the hamster dorsal skinfold chamber, with preparation as described in detail elsewhere.10 Oregon Green 488 paclitaxel encapsulated in cationic lipid complexes was injected i.v. yielding an effective Oregon Green paclitaxel dose of 0.053 mg/kg b.w. (n = 3). Intratumoral distribution of the drug delivered by cationic lipid complexes was compared to i.v. injection of Oregon Green 488 paclitaxel dissolved in cremophor (n = 3, Oregon Green 488 paclitaxel dose: 0.78 mg/kg b.w.). For intravital fluorescence microscopy the awake, chamber-bearing hamster was immobilized in a Perspex tube on a specially designed stage (Effenberger, Munich, Germany) under a modified Zeiss microscope (Axiotech Vario; Zeiss, Goettingen, Germany). Selective observation of the flurorescent paclitaxel was carried out using epi-illumination with a 100 W mercury lamp and a selective fluorescence filter block (Zeiss, Goettingen, Germany). Images were acquired by a SIT video camera (C2400-08; Hamamatsu, Herrsching, Germany) and recorded by a digital image analysis system (KS400, Zeiss, Germany). To confirm results obtained by intravital fluorescence microscopy distribution of Oregon Green 488 paclitaxel delivered by cationic lipid complexes was confirmed by fluorescence microscopy of s.c. LLC-1 carcinomas. 2% Oregon Green paclitaxel encapsulated in cationic lipid complexes was i.v. injected into C57/Bl6 bearing s.c. LLC-1 carcinomas (tumor volume: 500–700 mm³). 30, 60 and 90 min after injection animals were perfused intracardially as described above. Texas-red labeled lectin (1 mg/ml; Vector Laboratories, San Francisco, CA) and the perfusion marker H3342 (15 mg/kg b.w.; Sigma-Aldrich, Deisenhofen, Germany) were intravenously injected before intracardial perfusion. Frozen sections (15 μm thickness) were collected on warm polylysine-coated slides and immediately refrozen. Fluorescence images were acquired by a Zeiss AxioCam MRm camera (AxioCAM MRm; Axio Vison 4.7; Zeiss, Goettingen, Germany).

Experimental protocol and treatment schedules

In the first set of experiments the impact of EndoTAG-1 scheduling on antitumoral effects was tested in s.c. LLC-1 carcinomas. Seven days after tumor cell injection animals were randomly assigned to 4 experimental groups: Control animals received i.v. injection of trehalose (300 μl) tree times a week (days 7, 9, 11, 14, 16, 18 after tumor cell injection). Treated animals received i.v. injection of EndoTAG™-1 yielding a total paclitaxel dose of 15 mg/kg b.w. weekly. The total dose was either injected once weekly (group 1: 15 mg/kg b.w./week; days 7 and 14) or divided into 3 or 5 applications, respectively); (group 2: 3 × 5 mg/kg b.w./week; days 7, 9, 11, 14, 16, 18; group 3: 5 × 3 mg/kg b.w./week). Twenty-four hours after the last treatment tumors were harvested for further histological examination.

To investigate antitumoral efficiency of combined therapy C57/Bl6 mice were treated by EndoTAG™-1 (3 × 5 mg/kg b.w./week; days 10, 12, 14, 17, 19, 21) and i.p. injection of cisplatin (2 × 2.5 mg/kg b.w./week; days 11, 13, 18, 20; Platinex®, Bristol-Myers Squibb, New York, NY). In L3.6pl pancreatic tumors, a combination therapy of i.v. EndoTAG™-1 injection (3 × 5 mg/kg b.w./week) and i.p. injection of gemcitabine (Gemzar 2 × 100 mg/kg b.w./week) was performed. Control animals received i.v. injection of 10% trehalose 3 times a week, EndoTAG™-1 monotherapy or gemcitabine monotherapy, respectively. Treatment was started on day 7 after tumor cell injection. Treatment was continued for three weeks and animals were sacrificed on day 19 following the start of treatment.

Immunohistochemistry

Tissue samples were either fixed in 4% paraformaldehyde and embedded into paraffin for H&E staining or frozen on dry ice for immunohistological analysis. Sections of tumor tissue with a 5-μm thickness were taken from all samples. For examination of tumor microvessel density, immunohistochemical staining for CD31 was performed. Cryosections were fixed in cold ethanol (10 min) and washed with PBS. The sections were incubated with the primary goat-anti-mouse CD31/platelet/endothelial cell adhesion molecule-1 antibody (1:250; Santa Cruz Biotechnology, Santa Cruz, CA) overnight at 4°C and rinsed with PBS. Sections were then incubated with the biotinylated secondary donkey-anti-goat antibody (1:200; Santa Cruz Biotechnology) for 1 hr at ambient temperature. Positive reactions were visualized by incubating the slides with avidin-biotin for 1 hr, followed by incubation with 3-amino-9-ethylcarbazole for an additional 30 min. The immunostained sections were counterstained with hemalaun, rinsed with distilled water, and mounted with Ultra Mount (Dako). For quantitative analysis, tissue sections were examined using a Zeiss Axiophot 2 microscope fitted with a 20× objective (Zeiss, Göttingen, Germany) combined with digital image analysis (Zeiss, KS400, Oberkochen, Germany). Specific CD31 staining was obtained after interactively setting an RGB threshold which depicts only specific CD31 staining. The vessel density was calculated as the percentage of the total area of CD31-positive pixels related to the total image area. A minimum of 15 slides from separate areas of each tumor were used for analysis.

TUNEL-positive cells were visualized using Fast Red as a chromogen (in situ cell detection kit; Roche, Mannheim, Germany). Sections were counterstained by haematoxylin.

Statistical analysis

All results are given as mean ± standard error of the mean (SEM). Statistical analysis of the data was performed using the Mann–Whitney rank sum with a Bonferroni-Holmes correction for multiple comparisons (Sigmastat; SPSS, Chicago, IL). Statistical analysis of tumor metastasis was performed using the Fisher exact test. P values less than 0.05 were considered significant.

Results

Cationic lipid complexes selectively target tumor microvessels in L3.6pl tumors

We hypothesized enhanced antitumoral efficiency of conventional chemotherapy by cationic liposome-based vascular targeting chemotherapy. Therefore, we first investigated whether the tumor vascular compartment represents the therapeutic target of EndoTAG™-1-based vascular targeting therapy in the applied orthotopic L3.6pl pancreatic cancer model. Systemic application of cationic lipid complexes revealed a strong accumulation within the tumor during early (d14) and late tumorigenesis (d23). The accumulation of cationic liposomes was highly restricted to tumor microvessels as shown by the extensive colocalization of liposomal fluorescence with the vascular FITC-labeled lectin (Fig. 1).

Figure 1. Tumor vascular targeting of rhodamine-labeled cationic lipid complexes in orthotopically grown L3.6pl pancreatic cancer. During both early and late tumorigenesis preferential binding of cationic liposomes to tumor microvessels compared to normal pancreatic microvessels was observed. Microvessels were stained by i.v. injection of FITC-lectin.

Drug delivery by EndoTAG™-1 is targeted to the tumor vascular compartment

In addition to the analysis of cationic liposome distribution within tumor tissue, the distribution of paclitaxel delivered by cationic lipid complexes was investigated in comparison to paclitaxel dissolved in cremophor. To analyze drug distribution by intravital fluorescence microscopy and confocal microscopy Oregon Green 488 paclitaxel was used as fluorescent tracer. In vitro Oregon Green 488 paclitaxel clearly retained highly specific binding to the tubulin cytoskeleton of HUVECs (Fig. 2a). Oregon Green 488 paclitaxel dissolved in cremophor displayed a homogenous extravascular drug distribution in tumor tissue (Figs. 2b and 2c) 60 min after intravenous injection. In contrast, a homogenous labelling of tumor endothelium was found only if Oregon Green 488 paclitaxel was delivered by EndoTAG™ technology (Fig. 2d). The therapeutic target was directed from the extravascular to the vascular tumor compartment. Moreover, enhanced uptake of Oregon Green 488 paclitaxel in tumor microvessels was found compared to drug uptake in normal surrounding host tissue (Fig. 2e). Notably, the injected dose of Oregon Green 488 paclitaxel dissolved in cremophor was approximately 15-fold compared to Oregon Green 488 paclitaxel encapsulated in cationic liposomes to ensure a reliable signal to tissue autofluorescence ratio. Drug targeting of tumor microvessels by Oregon Green 488 paclitaxel delivered by cationic lipid complexes was also confirmed by fluorescence microscopy of subcutaneous LLC-1 carcinomas following intravenous EndoTAG™-1 injection loaded by Oregon Green 488 paclitaxel (Fig. 2f).

Figure 2. Oregon Green 488 paclitaxel was used as fluorescent tracer to analyze the therapeutic target of EndoTAG™-1. In vitro Oregon Green 488 paclitaxel clearly retained highly specific binding to the tubulin cytoskeleton of HUVECs (Fig. 2a). Intratumoral distribution of Oregon Green 488 paclitaxel dissolved in cremophor (Figs. 2b and 2c) was compared by Oregon Green 488 paclitaxel delivered by EndoTAG™ technology (Figs. 2d–f). The therapeutic target of paclitaxel was directed to the vascular tumor compartment by encapsulation into cationic lipid complexes (Figs. 2d and 2f). Distribution of Oregon Green 488 paclitaxel 30 min after i.v. injection of EndoTAG-1 in liver (g), lung (h) and small intestine (i). Texas-red lectin was i.v. injected as a vascular dye, nuclei were labeled by i.v. injection of H3342.

In addition to tumor tissue the distribution of Oregon Green 488 paclitaxel delivered by cationic lipid complexes was investigated in normal organs (Figs. 2g–i). In liver tissue, an accumulation of Oregon Green paclitaxel was found around portal tracts at 30 min after injection (Fig. 2g). A clear colocalization with the vessel wall of portal veins or small hepatic arteries was not observed. At 90 min Oregon Green 488 paclitaxel appeared to be taken up in part by hepatocytes. In lung tissue Oregon Green 488 fluorescence was obviously detectable at 30 min (Fig. 2h). Green fluorescence was in part colocalized with Texas-red lectin staining. In small intestine, partial colocalization of Oregon Green 488 paclitaxel was restricted to mesenteric microvessels, whereas in microvessels of the mucosa uptake of Oregon Green 488 paclitaxel was not observed (Fig. 2i). In the kidney a scattered signal increase was observed corresponding to afferent and efferent microvessels. Interestingly, no Oregon Green 488 paclitaxel signal was observed in glomerula.

Efficiency of EndoTAG™-1 therapy is scheduling dependent

To achieve optimal scheduling of EndoTAG™-1 therapy for subsequent combination therapy studies, the impact of EndoTAG™-1 scheduling on s.c. LLC-1 tumor growth was observed (Fig. 3a). EndoTAG™-1 therapy clearly delayed tumor growth dependent on the drug scheduling applied. All animals received an effective dose of 15 mg/kg b.w. weekly. However, markedly enhanced therapeutic results were observed if the weekly dose was split into 3 or 5 applications, respectively. Whereas the maximum tolerated single dose of 15 mg/kg b.w. did not induce a significant reduction of tumor growth compared to controls, dose sequencing to 3 × 5 mg/kg or 5 × 3 mg/kg, respectively, significantly delayed tumor growth. Improved therapeutic efficiency by drug sequencing was associated with a decrease in tumor microvessel density (Fig. 3c). A significant reduction of tumor microvessel density by EndoTAG™-1 therapy was observed only if the weekly dose was split into 3 × 5 mg/kg or 5 × 3 mg/kg. Reduced tumor microvessel density was associated with an increase in apoptotic/necrotic tumor cells as shown by quantitative analysis of TUNEL-positive tumor cells (Fig. 3d). On the basis of the result described above, an EndoTAG™-1 scheduling of 3 × 5 mg/kg b.w. weekly was selected for subsequent combination therapy studies. Increasing the EndoTAG™-1 dosage over 15 mg/kg b.w. per week induced toxic side effects as indicated by a decrease of body weight. Therefore, the weekly dose of EndoTAG™-1 therapy was limited to 15 mg/kg b.w.

Figure 3. Impact of EndoTAG™-1 scheduling on antitumoral efficiency (a), body weight (b), tumor microvessel density (c) and TUNEL-positive tumor cells (d). Experiments were performed on s.c. LLC-1 carcinomas grown in C57/Bl6 mice. *p < 0.05 vs. controls.

EndoTAG™-1 therapy increases antitumoral efficiency of conventional chemotherapy

To test the effect of antivascular tumor therapy combined with cytotoxic chemotherapy, cisplatin or gemcitabine, respectively, were added to EndoTAG™-1 treatment. In subcutaneous LLC-1 carcinomas combined treatment of EndoTAG™-1 and cisplatin was superior to both monotherapies, resulting in a remarkable inhibition of primary tumor growth (Fig. 4a).

Figure 4. Antitumoral effects of EndoTAG™-1 therapy in combination with standard chemotherapy. Tumor growth (Fig. 4a) and microvessel density (Fig. 4b) of s.c. LLC-1 carcinomas treated by EndoTAG™-1 in combination with cisplatin. Tumor volumes of orthotopically grown L3.6pl pancreatic tumors after EndoTAG™-1/gemcitabine combination therapy (Figs. 4c and 4d). *p < 0.05 vs. trehalose; ^#vs. gemcitabine; ^§vs. EndoTAG™-1.

Histological analysis (H&E) of trehalose-treated subcutaneous LLC-1 tumors (Fig. 5a) revealed dense tumor cells with low tumor stroma content. In contrast, small insular necrotic tumor areas were detectable following cisplatin treatment (Fig. 5b). In EndoTAG™-1-treated tumors, ring-shaped clusters of necrotic tumor cells were found around tumor microvessels (Fig. 5c) indicating malnutrition of tumor cells by an antivascular mechanism. Finally, after the combination therapy of anti- tumor cell-directed cisplatin and vascular targeting EndoTAG™-1 therapy large necrotic tumor areas were found (Fig. 5d).

Figure 5. H&E staining of subcutaneous LLC-1 carcinomas. Trehalose-treated controls (a), cisplatin monotherapy (b), EndoTAG™-1 monotherapy (c), EndoTAG™-1/cisplatin combination therapy (d). Scale bar: 100 μm.

Cisplatin monotherapy did not induce a significant reduction of tumor microvessel density (Fig. 4b). In contrast, microvessel density was reduced by EndoTAG™-1 monotherapy as well as by the combined treatment regime. However, adding cisplatin therapy to EndoTAG™-1 treatment did not induce further reduction of tumor microvessel density.

In pancreatic cancer, EndoTAG™-1 therapy was combined by gemcitabine treatment. In orthotopically grown human L3.6pl pancreatic tumors EndoTAG™-1 monotherapy induced a significant inhibition of primary tumor growth compared to trehalose treated controls (Fig. 4c). The antitumoral effect by EndoTAG™-1 yielding a weekly paclitaxel dose of 3 × 5 mg/kg b.w. was equal to gemcitabine monotherapy in a dose of 100 mg/kg b.w. twice weekly. EndoTAG™-1 treatment combined by standard gemcitabine therapy induced additive antitumoral effects. Compared to trehalose treated controls primary tumor volume was markedly reduced by 78%. In further experiments, dose effects of gemcitabine were investigated. Dose reduction of gemcitabine monotherapy from 100 mg/kg b.w. to 50 mg/kg b.w. was associated with a significantly reduced antitumoral efficiency (Fig. 4d). Whereas high dose monotherapy induced a reduction of primary tumor weight by 57% compared to trehalose treated controls, low dose monotherapy (50 mg/kg b.w.) reduced tumor weight by less than 30%. Interestingly, adding EndoTAG™-1 treatment to gemcitabine therapy completely equalized the reduced antitumoral efficiency of low dose gemcitabine treatment and antitumoral effects tended to become supra-additive. In both groups primary tumor weight was reduced by ∼80% compared to trehalose treated controls.

In addition to primary tumor growth the incidence of lymph node and liver metastasis and peritoneal carcinosis was investigated (Table 1). Compared to trehalose treated controls gemcitabine treatment did not induce a significant reduction of lymph node or liver metastasis. EndoTAG™-1 monotherapy did not inhibit lymph node metastasis. However, no liver metastasis or peritoneal carcinosis were found in EndoTAG™-1 treated animals. Following combination therapy lymph node metastasis was significantly reduced compared to trehalose and gemcitabine monotherapy and no liver metastasis or peritoneal carcinosis was found in response to gemcitabine/EndoTAG™-1 combination therapy.

Table 1. Primary Tumor Growth and Tumor Metastasis of Human L3.6PL Pancreatic Tumors Grown Orthotypically in Nude Mice

Discussion

Cationic lipid complexes have been shown to target angiogenic microvessels of solid tumors preferentially.5, 6 These lipid complexes therefore appeared to be promising drug carriers directing chemotherapeutic compounds to the tumor endothelium to realize vascular targeting tumor therapy.15, 16 This novel therapeutic strategy was first realized by the synthesis of EndoTAG™-1 (formerly known as Lipopac/MBT-0206) comprising paclitaxel encapsulated in cationic lipid complexes. In earlier studies EndoTAG™-1 was shown to increase significantly the antitumoral efficacy of the drug in different experimental tumor models.8, 10, 17 Subsequent experiments investigated the mode of action and provided convincing evidence of tumor vascular targeting by EndoTAG™-1: impairment of tumor microvascular perfusion in response to EndoTAG™-1 therapy has been demonstrated by intravital microscopy and DCE-MRI,10, 11 respectively. Moreover, tumor endothelial cell apoptosis and intratumoral thrombosis induced by EndoTAG™-1 therapy indeed supported vascular targeting as the underlying mechanism.17, 18 In addition to paclitaxel, doxorubin, 5-FU and camptothecin have, meanwhile, been successfully encapsulated in cationic lipid complexes to realize vascular targeting therapy in preclinical animal models.9, 19, 20

However, the clinical success of vascular targeting and antiangiogenic agents depends upon potential combination with conventional therapies. To date, antivascular drugs can not eradicate tumors completely, and remarkable antitumoral effects can be achieved in the clinical situation only by combining antivascular tumor therapy with conventional cytotoxic radio- or chemotherapy. The benefit sought from the combination of EndoTAG™-1 and conventional chemotherapy is mainly a complementary mode of action between the two therapies: EndoTAG™-1 acting primarily on the tumor vasculature and the chemotherapy mainly affecting proliferating tumor cells. In this way an effective ‘two-compartment’ tumor therapy can be realized affecting both the tumor cell and tumor vascular compartment within the tumor stroma. Moreover, the assumption that most advanced solid tumors derive their growth advantage from a compensatory cross talk among receptors within the tumor stroma and tumor cell compartment provides a further argument supporting a combined targeting approach.

To verify that the tumor vascular compartment is the target of EndoTAG™-1 therapy, the intratumoral distribution of the encapsulated drug paclitaxel has now been analyzed in vivo in the present investigation in addition to cationic liposome distribution. Intravenously injected fluorescently spiked cationic lipid complexes clearly exhibited tumor vascular targeting properties in human L3.6pl orthotopic tumor grafts (Fig. 1). This observation is consistent with results obtained in A-MEL-3 melanomas and LLC-1 carcinomas following systemic injection of EndoTAG™-1 or camptothecin encapsulated in cationic lipid complexes, respectively. Although several different studies have proven the charge depended tumor vascular targeting effect of cationic liposomes,5, 6, 21 only a small number of investigations have addressed the important question, to which tumor compartment the encapsulated drug is delivered. By intravital microscopic analysis of Oregon Green 488 paclitaxel encapsulated in cationic liposomes we now can clearly demonstrate that paclitaxel when loaded in cationic liposomes is delivered preferentially to the tumor endothelium. Therefore, the tumor vascular compartment is proven to be the therapeutic target of EndoTAG™-1 and not the interstitial matrix, where the drug is known to accumulate when paclitaxel dissolved in cremophor is systemically injected. In accordance with our results Campbell et al. have demonstrated preferential delivery of doxorubicin to the tumor vascular compartment when doxorubicin hydrochloride was loaded in peg-modified cationic liposomes.20, 16 In addition to tumor microvessel fluorescence analysis of healthy organs did show delivery of paclitaxel to lung and liver tissue, in which cationic liposomes have recently been shown to be metabolized following systemic injection.22 However, in contrast to homogenous and intensive paclitaxel uptake in angiogenic tumor endothelial cells, paclitaxel uptake related to endothelial cell surface area is markedly reduced in quiescent endothelial cells of healthy organs.

To take advantage of the potential benefits of a combination therapy, issues of dosing, timing and sequencing are critical to take into consideration. Moreover, to prevent a possible negative inhibition of vascular targeting therapy and chemotherapy the chemotherapeutic drug should be critically selected with regard to the molecular target. For instance, potential negative interfering effects have been reported when microtubule-stabilizing taxanes are combined with microtubule-destabilizing vascular disruptive agents (ZD6126).23, 24 To rule out a negative interfering effect in the present therapeutic approach, EndoTAG™-1 vascular targeting therapy was combined with cisplatin or gemcitabine therapy, neither of which affect microtubules of endothelial cells.

Until now, to our knowledge, no experimental results have been published that investigate the impact of drug scheduling on the antitumoral efficacy of cationic liposomal formulations. In the present study, the therapeutic effects of EndoTAG™-1 were clearly dependent on drug scheduling. Interestingly, maximum tolerated dose treatment was less efficient compared to a rather metronomic drug sequencing comprising 3 to 5 intravenous applications a week at a lower dose. This effect can be explained by the endothelial cell turnover time in solid tumors. The potential minimal doubling time of the tumor endothelium has been shown to be approximately 2.5 days in solid mouse tumors.25 Therefore the high endothelial cell turnover can compensate antivascular effects with EndoTAG™-1 therapy given only once weekly, which is in accordance with an almost unchanged tumor microvessel density in response to EndoTAG™-1 therapy given every 7 days. The antiproliferative and apoptotic effect of paclitaxel on endothelial cells is cell cycle dependent and mediated by a G₂-M arrest.26, 27 Interestingly, ultra low concentrations of paclitaxcel (<10 nmol/l) induce cytostatic effects in endothelial cells by the initiation of a mitochondrial apoptotic signalling pathway.27 Therefore, it might be more advantageous to realize a rather metronomic scheduling in order to maintain a minimum effective drug concentration at the tumor endothelium for an extended period of time. If the drug sequencing is estimated by the endothelial turnover time, an improved therapeutic effect can be expected by a drug application every 2–3 days, which is consistent with results obtained in the present investigation.

In both tumor models the combination therapy of EndoTAG™-1 and conventional chemotherapy markedly enhanced antitumoral effects. In pancreatic cancer gemcitabine combined with EndoTAG™-1 was able to inhibit tumor metastasis in addition to primary tumor growth. However, in the present study, we did not investigate, whether EndoTAG™-1 in combination with chemotherapy is capable to induce regression of already existing metastases, which would be an important point in the palliative clinical treatment of advanced pancreatic cancer.

Interestingly, when EndoTAG™-1 therapy was added to gemcitabine treatment, the gemcitabine dose could be markedly reduced (upto 50%), while maintaining similar antitumoral efficacy. These results prove our hypothesis that vascular targeting tumor therapy with a cationic liposomal formulation can be effectively combined with conventional chemotherapy. In addition to the simultaneous targeting of 2 tumor compartments, a direct interaction between both treatment modalities could further explain an additive therapeutic effect: tumor vascular targeting by EndoTAG™-1 has been shown to increase tumor microvessel permeability as quantified by DCE-MRI11 or intravital microscopy.28 Transvascular transport is known to be a prerequisite for good antitumoral effects of small molecular chemotherapeutic drugs. Therefore, modifying the blood- tumor barrier by increasing tumor vessel permeability might be an important mechanism for enhanced antitumoral efficiency in response to EndoTAG™-1 combination therapy. Moreover, a decreased interstitial fluid pressure in response to EndoTAG™-1 therapy as is known for other vascular targeting agents, could contribute to an increased influx of small molecular therapeutic substances.

The preclinical results may in the future be confirmed in a clinical phase II study (CT4001), investigating the safety and efficacy of EndoTAG™-1 in combination with standard gemcitabine treatment in patients with locally advanced and/or metastatic pancreatic cancer.29 Two hundred patients have already been enrolled in this phase II study and preliminary data confirm a favorable safety profile of EndoTAG™-1 in combination with gemcitabine treatment.30 Moreover, the study has shown promising preliminary therapeutic results as the median overall survival was increased by EndoTAG™-1 combination therapy compared to gemcitabine standard monotherapy.30

In conclusion, vascular targeting tumor therapy by EndoTAG™-1 can effectively be combined with standard small molecular chemotherapy resulting in markedly enhanced antitumoral efficacy. Thus, we propose antivascular tumor therapy by EndoTAG™-1 in combination with conventional chemotherapy to be a promising novel strategy for clinical cancer treatment.