Contributing Factors of Temozolomide Resistance in MCF-7 Tumor Xenograft Models

Abstract

Vasculature mediated drug resistance in tumors was studied in female SCID mice bearing wild type MCF-7 and adriamycin resistant MCF-7/ADR xenograft using temozolomide (TMZ). A strong tumor growth inhibitory effect of TMZ treatment was observed in MCF-7 tumors during the initial treatment phase with subsequent relapse, but not in MCF-7/ADR tumors. Non-invasive MRI measurements of tumor vascular volume and vascular permeability-surface area product (PS) demonstrated significant reduction of PS in long-term treated MCF-7, but not in MCF-7/ADR tumors. O⁶-Methylguanine-DNA methyltransferase (MGMT) mRNA, and VEGF expression was analyzed using real-time RT-PCR and ELISA, respectively. No significant changes in MGMT mRNA and VEGF expression were observed in either MCF-7 or MCF-7/ADR tumors. However, in vitro incubation of MCF-7 cells with TMZ did induce the expression of MGMT mRNA. In addition, p53 and p21 levels were scored by immunoblotting. Exposure of cells to TMZ did not affect either the p21 or the p53 expression in both MCF-7 and MCF-7/ADR cells. The absence of these molecular responses to TMZ treatment in MCF-7 tumors in vivo supports the possibility that the onset of cancer drug resistance is associated with reduced PS, which can decrease delivery of the drug to cancer cells.

INTRODUCTION

Changes in the physiological environments of solid tumors can play a key role in the development of tumor drug resistance. In addition to molecular mechanisms such as the evolution of multi-drug resistant proteins and enzymes, poor delivery of anticancer drugs to the tumor also contributes to failure of cancer chemotherapy ^1,2 by preventing efficient accumulation of the cytotoxic agent at therapeutic concentrations in the tumor. For example, doxorubicin treatment led to overexpression of the vascular endothelial growth factor (VEGF), and locally increased vascular permeability, which in turn enhanced drug accumulation in the same location, i.e., this positive feedback loop led to a heterogeneous distribution of the drug within the tumor.³ Similar phenomenon is also observed with the distribution of a low molecular weight agent ¹³C-labeled phenylacetate ⁴ and ¹³C-labeled temozolomide ([¹³C]TMZ)⁵ within tumors using magnetic resonance imaging (MRI). Conventional chemotherapy can have antiangiogenic properties that affect tumor vasculature either by directly damaging endothelial cells or by suppression of VEGF.⁶ Suppression of the tumor vasculature can in turn restrict the access of therapeutic agents to the tumor. The distinct contribution of physiological barriers to drug resistance from other factors such as expression of drug efflux pumps and DNA repair proteins in drug resistance needs to be evaluated for development of improved therapeutic strategies.

In this study, dynamic MRI using a macromolecular contrast agent albumin-GdDTPA was used to evaluate tumor vasculature. Water-soluble macromolecules such as albumin and dextran are known to be retained in the systemic circulation for a long time,⁷ and accumulate in the tumor due to enhanced permeability and retention (EPR) effect.⁸ Dynamic MRI of albumin-GdDTPA was used to quantitatively determine parameters of tumor vasculature, vascular volume (VV) and permeability-surface area product (PS).⁹ Tumor areas with high VV and PS are generally highly accessible to the drug due to efficient blood supply and highly permeable vasculature. We measured VV and PS in MCF-7 and MCF-7/ADR adriamycin resistant breast cancer models subjected to chemotherapy with temozolomide (TMZ). TMZ is generally used against malignant brain tumors;¹⁰ however for the purpose of this study, we have used breast cancer models that are also susceptible to TMZ,¹¹ and lack the blood-brain (or blood-tumor) barrier that can complicate interpretation of experimental results. MCF-7 human breast carcinoma cells were used in this study because TMZ is also established in preclinical studies using breast cancer models.^5,11 Since TMZ is not a substrate for P-glycoproteins ¹² that are one of the major drug efflux pumps, contribution of P-glycoprotein to TMZ resistance can be excluded.

The intratumoral concentrations of the drug were also directly measured by liquid chromatography with mass spectrometric detection (LC/MS/MS) at the end points of TMZ chemotherapy. ELISA and real-time RT-PCR were performed to analyze VEGF protein and DNA repair protein MGMT mRNA expression, respectively. Along with the direct cytotoxic effect of chemotherapy on endothelial cells, another potential mechanism that contributes to poor drug delivery can be suppression of tumor VEGF expression by the drug. Therefore, determination of VEGF expression is important for better understanding of physiological barriers. In addition, we explored the effect of TMZ on p53 that is involved in regulating apoptosis and functions as a tumor suppressor to verify another possible cellular mechanism of TMZ resistance. Using a combination of these approaches, this study has elucidated the relationship among anti-cancer activity, tumor vascular function, and drug resistance during the course of TMZ chemotherapy.

MATERIALS AND METHODS

Materials

Temodar^® capsules (Major component: temozolomide; TMZ) were obtained from Schering-Plough Co. (Kenilworth, NJ, U.S.A.). TMZ was purchased from LKT Laboratories, Inc. (St. Paul, MN, U.S.A.). For in vivo TMZ therapy, the content of the capsule was transferred to a glass tube, and mixed well before application. 17β-Estradiol pellets (0.18 mg/pellet, 60-day release) were purchased from Innovative Research of America (Sarasota, FL, USA). Albumin-GdDTPA was prepared as previously described.⁹ MGMT sense (15-mer) and MGMT antisense (17-mer) primers were purchased from Integrated DNA Technologies, Inc. (Coralville, IA, U.S.A.). TRIzol^® Reagent and SuperScript^™ First-Strand Synthesis System for RT-PCR were purchased from Invitrogen (Carlsbad, CA, U.S.A.). A mouse monoclonal antibody for p53 (sc-126) and a rabbit polyclonal antibody for p21 (sc-756) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, U.S.A.). A horseradish peroxidase (HRP) linked whole anti-mouse IgG and a HRP linked whole anti-rabbit IgG were from GE Healthcare (Bucks, U.K.). Protein molecular weight marker was from Fermentas Inc. (Hanover, MD, U.S.A.). All other chemicals were of reagent grade and were obtained commercially.

Animals and cell lines

Severe combined immunodeficient (SCID) mice (female, five weeks old, approximately 18 g body weight) were purchased from NCI (Bethesda, MD, USA). All mice were maintained on sterilized food and water. All animal experiments in the present study were compliant with the Guidelines for Animal Experimentation of The Johns Hopkins University School of Medicine.

Two well-established human breast carcinoma cell lines MCF-7 and its adriamycin resistant variant MCF-7/ADR were used in this study. The MCF-7 cell line was originally obtained from ATCC (Manassas, VA, U.S.A.), and MCF-7/ADR cells were developed by Dr. M.P. Gamcsik (currently Department of Medicine, Duke University Medical Center). MCF-7/ADR cells selected for resistance to a particular chemotherapeutic agent generally demonstrate more pronounced MDR phenotype in comparison to stable clones of these cells obtained by transfecting with MDR1/P-glycoproteins,¹³ and possess mutations in p53.¹⁴ Both cell lines were maintained in Eagle’s minimum essential medium (EMEM) with 1% penicillin, streptomycin and 10% fetal bovine serum at 37°C with 5% CO₂. For MCF-7/ADR cells, the growth medium was supplemented with 1.5 μM of adriamycin every medium change. One half of a 17β-estradiol pellet was implanted subcutaneously into the back of mouse, and cells were orthotopically inoculated (3 × 10⁶ in 0.05 ml of Hanks’ balanced salt solution) into the mammary fat pad of mice 48 h post-implantation of the estrogen pellet.

In vivo TMZ therapy

TMZ therapy was started when tumors become palpable (n = 6 in each group). The suspension of the contents of a Temodar^® capsule in water (10 mg/ml of TMZ in solution) was administered intragastrically to MCF-7-bearing mice fasted over 24 h at a dose of 100 mg eq. TMZ/kg on 2 consecutive days per week, i.e., at 200 mg eq. TMZ/kg/week, and chemotherapy was repeated for five weeks (1,000 mg eq. TMZ/kg in total) as shown in Figure 1. Control mice received purified water that was administered according to the schedule in the same volumes. Tumor volumes were monitored for 60 days, and the tumor volume at day 1 was used as the initial volume. The ratio of the tumor volume to the initial volume was adopted as an index of tumor growth inhibitory effect of TMZ. Tumor volumes were determined from the length (l, mm) of the longest axis and the width (w, mm) of the vertical axis of the tumor measured with a slide caliper. Tumor volume (V, mm³) was calculated using the equation:

$$V=\frac{(l\times w^2\times \pi )}{6}$$

Figure 1.

Schedule of temozolomide chemotherapy used in this study. This schedule was followed for measuring tumor shrinkage, tumor vascular volume and permeability surface area, intratumoral TMZ concentration, and in vivo VEGF and DNA repair protein, MGMT measurement studies.

Changes in body weight of each mouse were monitored during the course of the study to evaluate possible toxic effects of the therapy.

Measurement of VV and PS

Each group consisted of five animals, and MR measurement of VV and PS was carried out as previously described.⁹ Briefly, each anesthetized mouse was immobilized in a 2 cm solenoid coil in a plastic cradle, and the tail vein was catheterized before placing the animal in the magnet. Body temperature was maintained at 37°C by heat generated from a pad circulating with warm water. Imaging studies were carried out on a Bruker horizontal bore 4.7T spectrometer (Bruker Biospin GmbH, Rheinstetten, Germany) equipped with shielded gradients. Saturation-recovery T₁ images of 8 slices (slice thickness of 1 mm) were acquired with three relaxation delays of 100 ms, 500 ms, and 1 s with an in-plane spatial resolution of 0.250 mm (128 × 64 matrix zerofilled to 128 × 128, field of view = 32 mm, number of scan = 8). An M₀ map with a recovery delay of 7 s was acquired before administration of 0.2 ml of albumin-GdDTPA solution at a dose of 500 mg/kg. Acquisition of T₁ recovery maps was performed before, and repeated six times every 5 min starting from 3 min post-injection of a contrast agent. Quantitative T₁ relaxation maps were reconstructed from data sets for the maps for three different relaxation times and the M₀ data set on a pixel-by-pixel basis. T₁ relaxation times of the blood was measured at the end of the study using blood samples collected from the tail vein of the animal. Measured VV were corrected for permeability of the vessels.

TMZ concentration in the tumor and plasma

The analysis of TMZ concentrations in human plasma has been performed using HPLC ^15–17 or LC/MS/MS.¹⁸ In this study, we analyzed TMZ concentration in the tumor and in mouse plasma using a similarly adapted LC/MS/MS protocol. Since TMZ reaches peak concentrations in plasma approximately 1 h after p.o. administration,¹⁹ tumors and plasma samples were collected at 1.2 h post day 2 and day 30 administration and stored at −80°C until analysis. Sample preparation involved a single protein precipitation step of the addition of 50 μL of mouse plasma or a 200 mg/ml tissue homogenate diluted 1 to 10 in human plasma with 0.15 ml acetonitrile. Separation of the compounds of interest, including the internal standard (d8)-gefitinib, was achieved on a Waters X-Terra^™ C₁₈ (50 × 2.1 mm i.d., 3.5 μm) analytical column using a mobile phase consisting of acetonitrile/10 mM ammonium acetate (80:20, v/v) containing 0.1% formic acid and isocratic flow at 0.15 ml/min for 3 minutes. The samples were monitored by tandem-mass spectrometry with electrospray positive ionization. Linear calibration curves were generated over the range of 1 to 200 μg/ml for the mouse plasma samples and 0.1 to 10 μg/ml for tissue samples with values for the coefficient of determination greater than 0.99. Precision and accuracy were well within the generally accepted criteria for analytical methods (<15%) for values obtained within the same day and between days.

In vitro VEGF expression and O6-methylguanine-DNA methyl-transferase (MGMT) mRNA expression

MCF-7 and MCF-7/ADR cultured cells were treated with 10, 50 or 100 μg/ml TMZ. Media containing TMZ was changed every other day. All cells were split once a week, and this process was continued for three weeks.

VEGF expression

Media were taken from each flask when the cells were split at week 1 and 3, and the media together with 0.5% of a protease inhibitor were stored at −80°C until analysis. VEGF was quantified by the Quantikine (R&D Systems, Minneapolis, MN) enzyme-linked immunosorbent assay (ELISA). ELISA was carried out according to manufacturer’s instructions. The experiments were repeated in triplicate.

MGMT mRNA expression

The expression of MGMT mRNA was analyzed by real-time PCR amplification. Total RNA from each cell line was isolated using TRIzol^® reagent, and cDNA synthesized using SuperScript^™ First-Strand cDNA synthesis kit. 36B4 ribosomal mRNA was amplified as control. Before proceeding with real time PCR, the MGMT expression was confirmed using a conventional RT-PCR. A real-time PCR was performed in a 96 well plate using iCycler Thermal Cycler (Bio-Rad Laboratories, Hercules, CA, U.S.A.) with the following parameters: 94°C for 30 sec, 61°C for 30 sec, and 72°C for 30 sec for 45 cycles for MGMT, and 94°C for 30 sec, 57°C for 30 sec, and 72°C for 30 sec for 45 cycles for 36B4. Relative MGMT mRNA expression of TMZ treated groups to nontreated ones was calculated as a multiplicative factor by subtracting the number of cycles for treated group from that of control. The data were considered to be significantly different when the multiplicative factor was above two.

In vivo VEGF proteins and MGMT mRNA expression in the tumor

TMZ was administered intragastrically into the MCF-7-bearing mice according to the schedule shown in Figure 1. After 5-weeks treatment, tumors were resected, and stored at −80°C until analysis. Each tumor sample was divided in half, and VEGF and MGMT were quantified using ELISA and real-time RT-PCR, respectively, as described above.

Western blot analysis of p53 and p21 protein expression

MCF-7 and MCF-7/ADR cultured cells with 80–90% confluency were treated with 50 or 100 μg/ml TMZ for 2 h. For positive controls, proteins were extracted from the cells 3–4 h or 11–12 h following γ-irradiation (10 Gy) with or without TMZ treatment. After quantification, proteins were run on 4–15% gradient SDS-PAGE (Bio-Rad Laboratories) gels and immunoprobed with p53 and p21 antibodies. Results were visualized by HRP detection kit (Pierce Biotechnology Inc., Rockford, IL, U.S.A.). A mouse monoclonal antibody for p53 and a rabbit polyclonal antibody for p21 were used at 1:500 and 1:250 dilutions, respectively. Secondary antibodies were used at 1:2000 dilution. Antibody against β-actin was used (1:10000 dilution) as loading control. Samples were loaded at 30 μg per lane except that 10 μg of samples were loaded at each lane for β-actin detection.

Data analysis and statistical analysis

MR data were analyzed using in-house software written in the IDL programming environment (Interactive Data Language^™, Research Systems Inc., Boulder, CO, U.S.A.). Computer-assisted visualization was performed using Amira 3.1 (TGS Inc., San Diego, CA, U.S.A.). Three-dimensional images were drawn for all 5 animals in each group, and the images presented in this paper were selected as a representative sample of five animals.

Statistical analysis of intergroup differences in means was performed using the one-way or two-way repeated measure ANOVA. The data were considered to be significantly different when the p value was less than 0.05.

Results

TMZ therapy against MCF-7 tumor xenografts

The schedule for TMZ chemotherapy used in this study is illustrated in Figure 1. TMZ inhibited tumor growth of MCF-7 tumor xenografts while no growth inhibition was observed in the drug resistant phenotype (Fig. 2). In MCF-7 tumors, the volumes of the group treated with TMZ decreased over the first several days, and then gradually increased for the next several weeks. However, significant difference in tumor volumes compared to controls was still observed (p = 0.007). Loss of body weight and some mortality during the course of therapy were observed in both groups treated with TMZ (1/6 for MCF-7, 2/6 for MCF-7/ADR).

Figure 2.

Tumor growth inhibitory effect of temozolomide against MCF-7- and MCF-7/ADR-bearing mice. Along the horizontal axis, Day 1 represents the day of the first administration of temozolomide. The vertical axis represents tumor volume normalized to tumor volume at Day 1. TMZ was administered intragastrically at a dose of 100 mg eq TMZ/kg. Each group contains 6 animals, and the tumor volumes were measured two to three times per week for 60 days. In the MCF-7 control and MCF-7 temozolomide-treated groups, one mouse died at Day 46 and 39, respectively. In the MCF-7/ADR temozolomide-treated group, two mice died at Day 16 and 39.

VV and PS in MCF-7 tumor xenografts

In untreated MCF-7 xenografts, tumor vascular parameters VV and PS (averaged over total tumor volume for all tumors) were 2.93 ± 1.15 μl/g and 0.50 ± 0.12 μl/g·min, respectively. Figure 3A presents changes in the distribution of VV and PS in a representative MCF-7 tumor under TMZ chemotherapy. The distribution of VV was not significantly changed by TMZ (p = 0.368), while a significant change in PS was observed during TMZ chemotherapy (p = 0.039). In addition, almost no distribution of the contrast agent was detected within the central part of the tumor.

Figure 3.

Changes in vascular volume (VV) and permeability surface area (PS) in the tumor with five-weeks temozolomide chemotherapy against (A) MCF-7-bearing mice and (B) MCF-7/ADR-bearing mice. Gray-scale images represent mouse body and tumor xenografts. Red and green images represent the distributions of VV and PS, respectively. Each image in the same week shows the same mouse. The images in each week are delineated as a representative sample of five mice. Distribution of VV and PS in each animal had similar pattern in a given week. The VV (μl/g) and PS (μl/g×min) values at 0 wk were expressed as a mean ± S.D. of five animals. The values at 1 and 5 wk were expressed the differences from the previous time points as a mean ± S.D. of five animals.

VV and PS in MCF-7/ADR tumor xenografts

In untreated MCF-7/ADR xenografts, averaged tumor VV and PS were 0.79 ± 1.09 μl/g and 0.11 ± 0.16 μl/g·min, respectively. VV and PS in untreated MCF-7/ADR tumors were significantly lower than those in untreated MCF-7 tumors (p = 0.017 for VV; p = 0.002 for PS). Changes in the distribution of VV and PS in MCF-7/ADR tumors caused by TMZ chemotherapy are illustrated in Figure 3B. Both VV and PS in the drug resistant tumors were not significantly changed by TMZ treatment (p = 0.104 for VV; p = 0.819 for PS). As with MCF-7 xenografts, low values of VV and PS were observed in the central part of the tumor.

TMZ concentration in the tumor and in plasma

The retention time of TMZ and the internal standard, d₈-ZD1839, were 1.4 ± 0.3 min and 1.3 ± 0.3 min, respectively. The concentration of TMZ in the tumor during TMZ chemotherapy analyzed by LC/MS/MS is shown in Table 1. No significant changes between the drug concentrations at week 1 and week 5 of the treatment were observed in wild type and drug resistant tumors (p = 0.517 for MCF-7; p = 0.087 for MCF-7/ADR). Table 1 also gives TMZ concentration in plasma. Similar to the tumor tissue samples, no significant changes were observed in both groups (p = 0.733 for MCF-7; p = 0.592 for MCF-7/ADR).

Table 1.

Concentration of temozolomide in the tumor and in plasma analyzed by LC/MS/MS

cell lines	TMZ treatment	tumor (μg/g tissue)	Plasma (μg/ml)
MCF-7	1-wk	0.023 ± 0.006	38.4 ± 11.4
	5-wk	0.020 ± 0.003	35.4 ± 19.8
MCF-7/ADR	1-wk	0.029 ± 0.013	26.2 ± 6.1
	5-wk	0.019 ± 0.006	21.6 ± 13.1

Each value is expressed as a mean ± S.D. of five animals.

VEGF expression after TMZ treatment

TMZ treatment induced VEGF levels in the MCF-7 tumors and repressed VEGF levels in MCF-7/ADR tumors (Fig. 4). However, in vitro, the VEGF levels in both cell lines treated with TMZ decreased after 1 week, followed by an increase within three weeks of exposure of drug at different drug concentrations (data not shown). In both in vivo and in vitro experiments, no significant differences were observed amongst each group (p = 0.205 for MCF-7 in vivo; p = 0.394 for MCF-7/ADR in vivo; p > 0.05 for both cell lines at all drug concentrations in vitro).

Figure 4.

In vivo VEGF expression level of MCF-7 and MCF-7/ADR xenografts in mice following TMZ treatment for five weeks. VEGF was detected using ELISA. Open and closed columns represent control and TMZ-treated xenografts, respectively. The columns are expressed as a mean ± S.E. of three animals for control, of four animals for MCF-7 and five animals for MCF-7/ADR.

MGMT mRNA expression levels After TMZ treatment

The basal MGMT mRNA expression level in nontreated MCF-7/ADR cells and xenografts was significantly higher than that in nontreated MCF-7 cells and xenografts, respectively (p = 0.002 for in vitro; p = 0.028 for in vivo). No significant effects of TMZ chemotherapy on MGMT mRNA expression levels were observed in vivo in both MCF-7 and drug resistant xenografts (Table 2, p = 0.469 for wild-type; P = 0.991 for MDR) while in vitro incubation of 50 and 100 μg/ml of TMZ induced the expression of MGMT mRNA in MCF-7 cells but not in MCF-7/ADR cells (data not shown).

Table 2.

Effect of TMZ on MGMT mRNA expression in MCF-7 xenografts analyzed by real-time RT-PCR

cell lines/Xenografts	Animals	36B4-MgMt †	Multiplicative factors ‡
MCF-7	Control	−4.29
	TMZ treated	−3.94	1.27
MCF-7/ADR	Control §	−3.09
	TMZ treated	−3.08	1.00

^†36B4 was used for control primer.

^‡Multiplicative factors represent how many times MGMT expression increased or decreased in comparison with control group (average). Only changes (>2) are considered significant.

^§MGMT mRNA expression levels in non-treated MCF-7/ADR cells were significantly higher than that in non-treated MCF-7 cells (p < 0.05).

Effect of TMZ on p53 and p21 levels in MCF-7 cell lines

Immunoblotting assays for p53 and p21 expression levels in MCF-7/ADR cells showed a different pattern from that of MCF-7 cells (Fig. 5). In MCF-7 cells, both p53 and p21 levels were nearly unchanged following TMZ treatment. There was no p21 expression in MCF-7/ADR cells regardless of TMZ treatment and γ-irradiation.

Figure 5.

Western blot assay for p53 and p21 proteins in MCF-7 (A) and MCF-7/ADR (B) cells. Lane 1, no treatment; lane 2, 3–4 h incubation after γ-Irradiation (10 Gy); lane 3, 11–12 h incubation after γ -Irradiation (10 Gy); lane 4, TMZ (50 μg/ml) exposure for 2 h; lane 5, TMZ (50 μg/ml) exposure for 2 h, then 3–4 h incubation after γ -Irradiation (10 Gy); lane 6, TMZ (50 μg/ml) exposure for 2 h, then 11–12 h incubation after γ -Irradiation (10 Gy); lane 7, TMZ (100 μg/ml) exposure for 2 h; lane 8, TMZ (100 μg/ml) exposure for 2 h, then 3–4 h incubation after γ -Irradiation (10 Gy); lane 9, TMZ (100 μg/ml) exposure for 2 h, then 11–12 h incubation after γ -Irradiation (10 Gy). Samples were loaded at 10 or 30 μg per lane for β -actin or p53/p21 detection.

DISCUSSION

In the current study, we have used TMZ (MW 194.15) as a model drug to study contribution of vascular drug resistance in a solid tumor. TMZ is a relatively new imidazotetrazine alkylating agent with anticancer activity against glioblastoma and anaplastic astrocytoma.¹⁰ TMZ is known to be a prodrug of 5-(3-methyltriazen-1-yl) imidazole-4-carboxamide (MTIC), and is easily degraded at physiological pH, and converted to its active form, MTIC.^16,20 The cytotoxicity of MTIC is primarily due to alkylation at the O⁶ position of guanine with an additional alkylation that also occurs at the N⁷ position. The O⁶-methyl adduct can be removed by the DNA repair protein O⁶-methylguanine-DNA methyltransferase (MGMT) and this is a major pathway of resistance to treatment with methylating agents such as TMZ, and prevents the formation of O⁶-methylguanine: thymine mispairs.^11,12,21 In vivo TMZ chemotherapy did not induce MGMT mRNA expression in MCF-7 xenografts, although above 50 μg/ml of TMZ induced the expression of MGMT in MCF-7 cells in vitro. These different patterns of MGMT expression in vivo and in vitro are most likely due to significantly lower intratumoral TMZ concentrations detected in vivo compared to in vitro. According to the LC/MS/MS study, in vivo TMZ concentrations obtained with high-dose therapy were at least 100-fold lower than concentrations of TMZ used in the in vitro study. Therefore, MGMT may not significantly contribute to TMZ resistance in the in vivo study.

With respect to the schedule for TMZ therapy, the anticancer activity of TMZ is affected by dosing schedules,^19,22,23 and diet.²⁴ In clinical use, the initial dose is 150–200 mg/m² orally once daily for five consecutive days per 28-day treatment cycle.²⁵ The administration dose used in mice was determined from the relationship between body surface area and body weight for mice ^26,27 based on the human clinical dose. The initial success of TMZ therapy is typically observed in many solid tumor models treated with different anti-cancer drugs.²⁸ Cellular resistance mechanisms such as overexpressed MGMT and nonfunctional p53 expression resulted in the minimal therapeutic effects of TMZ in MCF-7/ADR xenografts. On the other hand, both irradiation and TMZ therapy resulted in a robust activation of p53 in MCF-7 cells. Since TMZ is not a potential substrate for P-glycoprotein transporters, the failure of TMZ chemotherapy in MCF-7/ADR-bearing mice can possibly be attributed to the intrinsically high MGMT mRNA expression and a nonfunctional p53 expression rather than P-glycoprotein overexpression.

On the other hand, tumor VV and PS are also important quantitative parameters of tumor vasculature. VV and PS in untreated MCF-7 xenografts were consistent with earlier reports.^29,30 Repetitive TMZ chemotherapy resulted in a significant reduction in tumor PS only in MCF-7 xenografts, indicating that TMZ induces decrease in tumor vascular permeability, which can contribute to failure of chemotherapy in this model. No significant VV reduction was observed in treated MCF-7 tumor xenografts. A possible explanation of this observation can be that host-derived tumor vasculature can be different in different tumor types and dependent on the tumor growth rate and location even within the same host.³¹ Drug delivery to the tumor depends on tumor vascularization and blood supply;³² therefore tumor vascular parameters can present physiological barrier to drug delivery in tumors. In previous studies, the intratumoral distribution of ¹³C-labeled TMZ ([¹³C]TMZ) was measured non-invasively by MRI, and a highly heterogeneous distribution of [¹³C]TMZ was observed in the MCF-7 tumor model.⁵ In addition, the status of the tumor blood supply was assessed by gadolinium enhanced dynamic MRI, and a corresponding heterogeneous distribution of the contrast uptake was detected in the tumors.⁵ The restricted delivery of [¹³C]TMZ to the central part of the tumor observed in these experiments was consistent with the previous report ⁴ as well as with general observations.³³ Our findings indicate that the significant reduction in permeability in the long-term treated MCF-7 xenografts together with a highly heterogeneous distribution of VV and PS with little contribution from the central part of the treated tumors resulted in a physiological barrier leading to restricted access of the drug molecules to the cancer cells. VEGF is one of the key factors that can control tumor vascular volume and permeability.³⁴ Our studies demonstrated that despite the significant decrease in PS in MCF-7 tumors, TMZ did not affect VEGF expression in either cells in vitro or xenografts in vivo. Further, although TMZ concentrations in the MCF-7 xenografts were comparable by LS/MS/MS analysis following TMZ treatment for 1 and 5 weeks, TMZ that reached tumor cells following 5^th-week treatment might be lower than that following 1^st-week treatment. A normalization effect might explain the discrepancy of this result; normalized vessels are less leaky to macromolecules ³⁵ such as albumin-GdDTPA used for the measurement of VV and PS in this study. Although TMZ is a small molecule, the normalization may have also resulted in an increasingly heterogeneous intratumoral distribution of TMZ as observed by Minko et al,³ which contributed to the resistance to TMZ therapy. We are currently exploring the potential mechanistic cause for the decrease in PS in MCF-7 tumors following TMZ treatment and the relationship with drug resistance.

Based on immunoblotting assay and a previous report,¹⁴ another possible mechanism of failure of TMZ chemotherapy against MCF-7/ADR xenografts include nonfunctional p53 in MCF-7/ADR cells. On the other hand, in MCF-7 cells TMZ did not affect p21 expression. This finding indicates that p53 function was also not affected by TMZ. It is conceivable that in vivo TMZ resistance in MCF-7 and MCF-7/ADR xenografts was acquired through different mechanisms, and the latter cells are drug resistant most probably due to hereditary component.³⁶ Again, our in vivo data indicate the possibility that factors such as physiological barriers may be important contributors to the resistance to TMZ therapy in MCF-7 tumor xenografts that are initially amenable to TMZ treatment.

In summary, TMZ had minimal effects on the expression of MGMT mRNA in MCF-7 xenografts in mice in accordance with a recent report that some tumors emerge resistance to alkylating agents independently of either their MGMT level or their mismatch repair status,³⁷ which is probably due to very low concentration of TMZ distributed to the tumor in vivo. In vivo dynamic MRI studies revealed that repetitive TMZ treatment significantly reduced PS in MCF-7 xenografts. These findings lead to the conclusion that failure of TMZ chemotherapy in tumor-bearing mice may be attributed to two different factors; (1) less permeability in the tumor acquired by repetitive TMZ treatment, and (2) nonfunctional p53 expression in tumors preselected for drug resistance. Therefore, heterogeneous distribution of vascular function parameters such as PS and their reduction that lead to limited delivery of TMZ can be an important determinant of drug resistance.

ABBREVIATIONS

LC/MS/MS

liquid chromatography with mass spectrometric detection

MCF-7/ADR

MCF-7 adriamycin-resistant phenotype

MGMT

O⁶-methylguanine-DNA methyltransferase

PS

permeability-surface area products

TMZ

temozolomide

VV

vascular volume