© The Author 2006. Published by Oxford University Press.
ARTICLE |
Tumor Vascular Permeability, Accumulation, and Penetration of Macromolecular Drug Carriers
Affiliations of authors: Department of Biomedical Engineering (MRD, WL, CRM, FY, AC), Department of Radiation Oncology (MWD), Duke University, Durham, NC
Correspondence to: Ashutosh Chilkoti, PhD, Box 90281, Durham, NC 27708 (e-mail: chilkoti{at}duke.edu).
 |
ABSTRACT |
|---|
 TopNotes Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Background: Delivery of anticancer therapeutic agents to solid tumors is problematic. Macromolecular drug carriers are an attractive alternative drug delivery method because they appear to target tumors and have limited toxicity in normal tissues. We investigated how molecular weight influences the accumulation of a model macromolecular drug carrier, dextran covalently linked to a fluorophore, in tumors. Methods: We used dextrans with molecular weights from 3.3 kDa to 2 MDa. Vascular permeability, accumulation, and three-dimensional penetration of these dextrans were simultaneously measured in solid tumors via a dorsal skin fold window chamber, intravital laser-scanning confocal microscopy, and custom image analysis. Results: Increasing the molecular weight of dextran statistically significantly reduced its vascular permeability by approximately two orders of magnitude (i.e., from 154 x 10–7 cm/s, 95% confidence interval [CI] = 134 to 174 x 10–7 cm/s, for 3.3-kDa dextran to 1.7 x 10–7 cm/s, 95% CI = 0.7 to 2.6 x 10–7 cm/s for 2-MDa dextran; P<.001, two-sided Kruskal–Wallis test) but increased its plasma half-life, which provided ample time for extravasation (i.e., to enter tumor tissue from the vasculature). Tumor accumulation was maximal for dextrans with molecular weights between 40 and 70 kDa. Dextrans of 3.3 and 10 kDa penetrated deeply (greater than 35 µm) and homogeneously into tumor tissue from the vessel wall. After a 30-minute period, a high concentration was observed only approximately 15 µm from the vessel wall for 40- to 70-kDa dextrans and only 5 µm for 2-MDa dextrans. Conclusions: Increasing the molecular weight of dextran statistically significantly reduced its tumor vascular permeability. Dextrans of 40 and 70 kDa had the highest accumulation in solid tumors but were largely concentrated near the vascular surface.
|
INTRODUCTION |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The clinical treatment of solid tumors could be improved by controlling the pharmacologic properties of anticancer therapeutic agents to deliver a greater dose to the tumor; with conventional drugs, this dose is typically limited by toxic side effects in normal tissues (1–3). In 1906, Paul Ehrlich established the concept of drug delivery (4) by proposing a carrier that would "bring therapeutically active groups to the organ in question." The goal of drug delivery is to increase the concentration of a therapeutic agent in the tumor while limiting systemic exposure. Numerous drug delivery technologies have been developed to accomplish this goal, including liposomes (5), micelles (6), antibody-directed enzyme-prodrug therapy (7), photodynamic therapy (8), affinity targeting (9), and macromolecular drug carriers (10,11).
Many of these drug delivery approaches take advantage of the unique pathophysiology of tumor vasculature. As early as the 1920s, researchers using a transparent chamber and injectable dye techniques found that, in contrast to normal tissue, tumors contain a high density of abnormal blood vessels that are dilated and poorly differentiated, with chaotic architecture and aberrant branching (12–15). Subsequently, various functions of tumor vasculature were found to be impaired, such as a higher vascular permeability than normal vessels; these impaired functions contributed to the higher concentration of plasma proteins detected in tumor tissues than in normal tissues (16–25). This phenomenon was elucidated by Maeda and colleagues (26–28) [and reviewed by Seymour (29)], who described it as the enhanced permeability and retention effect, which results from a combination of the increased permeability of tumor blood vessels and the decreased rate of clearance caused by the lack of functional lymphatic vessels in the tumor, and results in the increased accumulation of macromolecules in tumors. These findings support the use of macromolecules in tumor diagnosis and in therapy as drug carriers because they passively accumulate in solid tumors after intravenous administration.
A macromolecular drug carrier is typically composed of a macromolecule covalently linked to a therapeutic agent, and it targets solid tumors either passively (via its molecular weight and charge) or actively (via a specific affinity [e.g., an antibody] or stimulus) (11,30,31). In addition to the enhanced permeability and retention effect, macromolecular drug carriers have a longer plasma half-life, reduced toxicity in normal tissue, and higher activity against multiple drug-resistant cell lines than typical chemotherapeutic agents, and they have the ability to increase the solubility of poorly soluble drugs in plasma (10,11). Because of these characteristics, macromolecular drug carriers coupled to a low molecular weight drug often have higher anticancer efficacy than the low molecular weight drug alone (10,11). In general, the concentration of a macromolecule in tumors depends on two sets of parameters—one set that increases the accumulation of the macromolecule in tumors (such as perfusion, vascularity, vascular permeability, plasma half-life, and tumor-specific binding) and the other set that limits tumor localization (such as clearance through a vascular or lymphatic route) (1,32). For passively targeted macromolecules, only permeability, plasma half-life, and clearance are variable; these variables should depend on the molecular weight and charge of the drug carrier. In this study, we focused on the influence of the molecular weight of anionic macromolecules; the effect of charge has been well characterized elsewhere (33).
The antitumor effect of a macromolecular drug carrier coupled to a drug depends on its accumulation in the tumor and its spatial distribution within the tumor. The molecular weight of macromolecules strongly influences their plasma half-life (34–36), biodistribution (33,35–40), and normal vascular permeability (41). For proteins between 25 and 160 kDa, the permeability of tumor vasculature, in contrast to that of normal vasculature, was found to be weakly dependent on molecular weight (42); however, the molecular weights examined in that study did not include the full molecular weight range of potential macromolecular drug carriers. Furthermore, the effective interstitial diffusion coefficient in a tumor depends on molecular weight (43), but to date the three-dimensional penetration of macromolecules into the tumor interstitium from the vascular surface on a micrometer scale has not been reported.
We investigated the relationship between molecular weight and tumor vascular permeability for a set of model macromolecular drug carriers with a clinically relevant range of molecular weights and quantified the three-dimensional penetration of such macromolecular drug carriers from the vascular surface into the tumor interstitium. We hypothesized that there is an optimal molecular weight at which the macromolecular drug carrier's tumor vascular permeability, plasma half-life, and vascular or lymphatic clearance result in the highest accumulation in solid tumors. To test this hypothesis, we used the dorsal skin fold window chamber model, intravital laser-scanning confocal microscopy, and customized image analysis software to simultaneously assess the influence of the molecular weight of a model macromolecular drug carrier on vascular permeability, accumulation, and penetration in solid tumors.
|
MATERIALS AND METHODS |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Cell Line
Human squamous cell carcinoma (FaDu) cells were maintained as monolayers in tissue culture flasks containing minimal essential medium supplemented with Earle's salts, L-glutamine (292 mg/L), 10% heat-inactivated fetal bovine serum, penicillin (100 U/mL), streptomycin (100 µg/mL), and amphotericin B (0.25 µg/mL) (Gibco, Carlsbad, CA). Cultures were incubated at 37 °C with 5% CO2 in air.
Dorsal Skin Fold Window Chamber
All animal experiments were performed in accordance with Duke University's institutional animal care and use committee guidelines. Nude mice (BALB/c nu/nu) were anesthetized by intraperitoneal injection with a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg) and then prepared for window chamber implantation as described (44). A circular incision of 1 cm in diameter was made in a dorsal skin fold, over which a titanium chamber was surgically implanted. A single-cell suspension of FaDu cells was injected into the opposing fold of skin (1 x 104 cells in 5 µL of RPMI 1640 medium from Gibco). A circular glass coverslip was placed over the incision, through which the tumor and its associated vasculature were later visualized. Imaging studies were performed 7–11 days after cell injection, when tumors were 2–3 mm in diameter.
Image Acquisition
Nude mice (BALB/c nu/nu) with implanted dorsal fold window chambers were anesthetized with sodium pentobarbital (80 mg/kg intraperitoneally) and positioned laterally recumbent on a heated custom-designed microscope stage. The tail vein was cannulated for intravenous administration of a model macromolecular drug carrier or anesthesia, if required. The window chamber was fixed to the microscope stage to limit three-dimensional translation (in x, y, and z directions, where z corresponds to the axial direction). Fluorescence images were acquired with a LSM 510 laser-scanning confocal microscope (Zeiss, Jena, Germany) with a two-channel method in which one channel was used to define the tumor vasculature and the other channel was used to image the distribution of the model macromolecular drug carrier. Images were acquired with an in-plane resolution of 512 by 512 pixels with a field of view of 460 by 460 µm and an optical slice thickness of 7.1 µm.
The tumor vasculature was defined by injecting 1 mg of rhodamine-labeled 2-MDa dextran (Molecular Probes, Eugene, Oregon) and collecting a z stack (multiple images separated only in the z direction) through about 100 µm of tumor tissue with a z-step interval that was equal to half of the optical slice thickness. A single image in the middle (z direction) of the previously collected vascular volume was then acquired every 5 seconds for 30 minutes. After approximately 20 seconds of collecting background images, 0.5 mg of the model macromolecular drug carrier (i.e., fluorescein-labeled anionic dextran of the indicated molecular weight or bovine serum albumin labeled with Alexa Fluor 488, both from Molecular Probes) was injected. Dextrans with molecular weights of 3.3, 10, 40, and 70 kDa and 2 MDa were examined in this study.
Tumor Compartment Image Analysis
Three-dimensional vascular masks were created with the 3D for LSM program (Zeiss) and a five-step image processing algorithm [steps: 1) Gaussian filter, 2) segment, 3) scrap, 4) close, and 5) fill holes] to define a continuous tumor vasculature from the initial vascular volume (rhodamine channel). Image analysis was performed with custom-designed MATLAB software (MathWorks, Natick, MA) on 16-bit images in a TIFF format. Vascular and extravascular fluorescence intensities were determined by applying a motion-corrected vascular mask to the images though time, subtracting the background intensity, and normalizing the data by the maximum vascular intensity (expressed as %Vmax). This normalization procedure permitted the averaging of data from multiple animals and removed the confounding effects of tissue absorption because only a single slice was analyzed. Mice were removed from the accumulation study if the tumor vascular network did not behave as a consistent rigid body through the entire course of the experiment (14% of the mice).
Tumor Vascular Permeability Image Analysis
We developed a novel method to measure permeability by monitoring the fluorescence intensity in the vascular and extravascular space and applying the appropriate geometric considerations. The rate of solute transport (Js) across a blood vessel wall is described by the Kedem–Katchalsky equation shown below.
 | (1) |
The first term in Equation 1 represents the rate of solute transport due to convection, where Jv is the rate of fluid flow, 
is the average molar concentration in the vessel wall, and 
f is the filtration reflection coefficient. The second term describes the rate of diffusive transport, where P is the microvascular permeability, S is the surface area of the endothelium, and 
C is the concentration difference across the vessel wall. Because the direction and magnitude of convection across tumor vessels were unknown in our preparation, we ignored the convective term and rearrange Equation 1 for apparent permeability (Papp) to reflect the fact that there may be an unknown influence of convection (41,45).
 | (2) |
To relate fluorescence intensity to a concentration difference across the vessel wall, the available volume must be determined in both the vascular and extravascular compartments. In the vascular space, the available volume is determined by the hematocrit (HCT), which has previously been shown to be approximately 20% in tumor vessels for a similar preparation (46). The available volume fraction in the extravascular compartment (Kav) is a function of molecular weight (47,48). To determine Kav, we assumed that equilibrium partitioning occurred between the vascular and extravascular compartments at a quasi-steady state during the last minute of the experiment and that the ratio of extravascular fluorescence intensity to the corrected vascular intensity was calculated for each of the dextran molecular weights less than or equal to 40 kDa and serum albumin. The Kav for the 70-kDa and 2-MDa dextrans are approximately equal (47) and were determined from a similar set of 70-kDa data except at 1 hour at which quasi-steady state was appropriate. The Papp was determined with the following equation,
 | (3) |
where Ve is the volume of the extravascular space, and Ie and Iv are the extravascular and vascular fluorescence intensity respectively. Ve and S were determined with the 3D for LSM program on five slices centered on the location at which the rapidly scanned data was collected. The Papp was calculated at each individual time point, and the reported value is the average value of Papp from three to 20 time points directly after vascular mixing was complete. The molecular size, expressed as the hydrodynamic radius, was determine with the Stokes–Einstein equation and reported values of the diffusion coefficient (49,50).
Tumor Penetration Image Analysis
Tumor interstitial penetration was defined as the distance that a macromolecular drug carrier moved away from the source at a vascular surface. To determine tumor penetration, the three-dimensional vascular mask was converted into a three-dimensional distance map that encoded the distance from each voxel in the vascular mask to its nearest vascular surface. The distance map was made into 100 consecutive distance bins of equal size and applied to the rapidly scanned images through time. A distance of zero corresponded to the vascular intensity, and every other distance reported was measured to the middle of each distance bin. Boundary effects were removed by truncating the dataset by 40 voxels on a side (in the x–y plane) and then analyzing a slice in the middle (z direction) of the vascular volume. Penetration data were displayed as color-coded contour maps that were Gaussian-filtered over a three by three pixel area for three iterations to remove noise and to aid in image clarity.
Compartment Model
Accumulation of macromolecular drug carriers in the extravascular space was determined for the duration of the treatment by using a four-compartment model. The administration of the macromolecules was modeled as a bolus injection into a central blood compartment that was allowed to exchange with both normal and tumor extravascular compartments and to be eliminated. The tumor extravascular space was modeled as both a free and bound population of macromolecular drug carriers. The central blood compartment was fit to the vascular data, and both tumor extravascular compartments were fit to the extravascular data with SAAM II software (SAAM Institute, Inc., Seattle, WA). The area under the curve (AUC) was calculated for both the vascular and tumor extravascular compartments with a Rosenbrock numerical integrator. The simulation was performed until the extravascular intensity was less than 0.1 %Vmax, and doubling this simulation time increased the predicted AUC by less than 0.4%.
Statistical Analysis
The apparent permeabilities of dextrans with various molecular weights were reported as medians with 95% confidence intervals (CIs) and compared by the Kruskal–Wallis test and Mann–Whitney U post hoc test. Correlations between the apparent permeability and the molecular weight were tested with a Spearman rank test. Nonparametric tests were performed with Statview software (SAS Institute, Cary, NC). Other data were presented as means with 95% confidence intervals, and statistical comparisons were made with the Bonferroni t test (Bonferroni-corrected P values are displayed in the text). P values of less than .05 were considered statistically significant. All statistical tests were two-sided.
|
RESULTS |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Images of Vasculature and Macromolecule Tumor Distribution
To compare normal and tumor vasculature, fluorescein-labeled 2-MDa dextran was administered intravenously to visualize the blood vessels with intravital laser-scanning confocal microscopy. Representative images of normal and tumor vasculature are shown in Fig. 1. In normal vasculature, blood vessels were oriented parallel to one another (Fig. 1, A). In tumor vasculature, however, blood vessels were chaotically organized, with uneven and dilated vessels that appeared to initiate angiogenic sprouting (Fig. 1, B, and Supplemental Movie 1, available at: http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue5). A good example of the pathologic nature of tumor vasculature is illustrated by the loop connecting four separate vessels located in the middle of Fig. 1, B; this structure is not characteristic of normal healthy vascular architecture.
|
The distributions of model macromolecular drug carriers (i.e., fluorescein-labeled dextran) and tumor vasculature were visualized by intravital laser-scanning confocal microscopy. The model macromolecular drug carriers (green) of various molecular weights and tumor vasculature (reflected by rhodamine-labeled 2-MDa dextran, red) are shown in Fig. 2 and Supplemental Movies 2–4 (available at: http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue5). These images qualitatively present a time-dependent distribution of macromolecules in tumor tissue. The 3.3-kDa dextran rapidly extravasated from the tumor blood vessel, indicating a high apparent permeability, but had a short plasma half-life and was rapidly cleared. The 70-kDa dextran accumulated more persistently with a moderate apparent permeability and plasma half-life. The longest plasma half-life was observed with the 2-MDa dextran, but the limited accumulation indicated a low apparent permeability. Areas of focal hyperpermeability were identified by the bright green fluorescence associated with the tumor vasculature after 15–25 minutes in the images of 2-MDa dextran.
|
Molecular Weight Dependence of Apparent Permeability
The tumor's apparent permeability was calculated with Equation 3 for each dextran molecular weight and bovine serum albumin. The dependence of the tumor's apparent permeability on the molecular size of the dextran, expressed as the hydrodynamic radius, is shown in Fig. 3. The apparent permeability decreased approximately two orders of magnitude as the molecular weight of the dextran increased from 3.3 kDa to 2 MDa (i.e., from 154 x 10–7 cm/s for 3.3-kDa dextran to 1.7 x 10–7 cm/s for 2-MDa dextran; difference = 152.3 x 10–7 cm/s, 95% CI = 127.3 to 177.3 x 10–7 cm/s) and was negatively correlated with molecular weight (correlation coefficient = –.8 and P<.001; Spearman rank test). Molecular weight statistically significantly affected the value of apparent permeability (e.g., for 3.3-kDa dextran, 154 x 10–7 cm/s, 95% CI = 134 to 174 x 10–7 cm/s; for 40-kDa dextran, 9.5 x 10–7 cm/s, 95% CI = 7.5 to 11.4 x 10–7 cm/s; and for 2-MDa dextran, 1.7 x 10–7 cm/s, 95% CI = 0.7 to 2.6x10–7 cm/s; P<.001, Kruskal–Wallis test). The apparent permeability of dextran at each molecular weight investigated was statistically significantly different from every other dextran molecular weight, ranging from P = .021 to P = .011 for every comparison (Mann–Whitney U test) except for the difference in apparent permeability between the 40- and 70-kDa dextrans, which was not statistically significant (P = .273). Thus, for an identical concentration gradient across the vessel wall, the decreased apparent permeability for dextrans with high molecular weights should slow the rate of transport for drug carriers into the tumor interstitium (Equation 2). The median apparent permeability of serum albumin, which was used as a standard, was found to be 4.9 x 10–7 cm/s (95% CI = 3.8 to 5.9 x 10–7 cm/s).
|
Vascular Pharmacokinetics and Extravascular Accumulation of Macromolecules
The vascular clearance and extravascular accumulation of dextrans with various molecular weights was determined by quantitative image analysis and is summarized in Fig. 4. The plasma pharmacokinetics (Fig. 4, A) exhibited a characteristic distribution and elimination response for macromolecules, which is well described by a biexponential decay, and an increased plasma half-life for dextrans with higher molecular weights (34–36). The rate of accumulation of dextrans in the extravascular compartment (Fig. 4, B) was most rapid for dextrans with low molecular weights (<40 kDa), but such dextrans were also rapidly cleared from the tumor extravascular compartment and mirrored the plasma pharmacokinetics after about 10 minutes. Despite the lower apparent permeability and rate of extravasation for dextrans with higher molecular weights, accumulation of these dextrans appeared more durable and sustained because of their reduced rate of clearance and prolonged plasma half-life, which gave these larger macromolecules ample time to extravasate.
|
The goal of cancer therapy is to kill the greatest number of cancer cells, and the number of cells killed should be proportional to the concentration of the toxic drugs to which cancer cells are exposed. This exposure is best described by the AUC in the extravascular compartment because the extravascular AUC represents the cumulative exposure of cancers cells to the drug. Data in Fig. 4 were fit with a compartment model to predict the AUC in vascular and extravascular compartments for time periods longer than the 30-minute time course of the experiment to represent the entire exposure that a patient may experience. The compartment model accurately captured the data for dextrans of all the molecular weights investigated (Fig. 5). The predicted AUC in the vascular and extravascular compartments along with the apparent permeability for each dextran are shown in Table 1. The vascular AUC statistically significantly increased with the molecular weight of the dextran (P<.004, Bonferroni t test), as would be expected from the longer plasma half-life of larger molecules (34–36), except for the 40- and 70-kDa dextrans (P>.05, Bonferroni t test). The extravascular AUC (i.e., the exposure of cancer cells to macromolecular drug carriers) was modest for dextrans with a low molecular weight (<40 kDa) (e.g., for 3.3-kDa dextran, 887 %Vmax x minute, 95% CI = 881 to 893 %Vmax x minute), but it increased statistically significantly by approximately threefold when the molecular weight of the dextran used was increased from 10 kDa (846 %Vmax x minute) to 40 kDa (2398 %Vmax x minute) (difference = 1552 %Vmax x minute, 95% CI = 1476 to 1628 %Vmax x minute; P<.001, Bonferroni t test), despite a statistically significant decrease in apparent permeability from 10-kDa dextran (32 x 10–7 cm/s) to 40-kDa dextran (9.5 x 10–7 cm/s) (difference = 22.5 x 10 cm/s, 95% CI = 15.1 to 29.9 x 10–7 cm/s; P = .014, Mann–Whitney U test). The extravascular AUC was maximal for dextrans with molecular weights of 40 and 70 kDa and then decreased for 2-MDa dextran (2068 %Vmax x minute, 95% CI = 2015 to 2120 %Vmax x minute), probably because of the marked reduction in the apparent permeability of 2-MDa dextran (1.7 x 10–7 cm/s, 95% CI = 0.7 to 2.6 x 10–7 cm/s).
|
|
Penetration of Macromolecules From the Blood Vessel Wall into the Tumor
The three-dimensional penetration of macromolecules into the tumor interstitium from the vascular surface over a 30-minute period was determined by quantitative image analysis and is summarized in Fig. 6. Dextrans of 3.3 and 10 kDa rapidly penetrated deep into the tumor tissue (>35 µm) and were homogeneously distributed throughout the extravascular compartment. However, 10-kDa dextran required slightly more time than 3.3-kDa dextran to reach the same level of deep penetration. In contrast, 70-kDa dextran was heterogeneously distributed in the extravascular compartment throughout the entire 30-minute experimental period, with a substantial concentration of approximately 30 %Vmax within only 15 µm of the vascular surface. The interstitial penetration of 40-kDa dextran was similar to that of 70-kDa dextran (data not shown). Substantial penetration of 2-MDa dextran was detected within only 5 µm of the vascular surface after 30 minutes. We expect 2-MDa dextran to continue to penetrate into the tumor tissue after the 30-minute experimental period because of its long plasma half-life; however, 70-kDa dextran appeared to have essentially reached its maximal penetration at 30 minutes because the contour lines of the plot shown in Fig. 6 were horizontal at 30 minutes.
|
|
DISCUSSION |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
In this study, we chose dextran as a model macromolecule because it has long been studied as a macromolecular drug carrier (51). We found that the tumor's apparent permeability to dextran statistically significantly decreased as the molecular weight of dextran increased (range = 3.3 kDa to 2 MDa), which would slow the rate of extravasation for larger dextrans. Despite the decreased apparent permeability of dextrans with a higher molecular weight, the greatest tumor accumulation was found for dextrans with molecular weights of 40 and 70 kDa. Although the larger dextrans displayed the highest accumulation in tumors, they penetrated a relatively short distance into the tumor compared with lower molecular weight dextrans and were highly concentrated only near the vascular surface.
Tumor Vascular Permeability
The vascular apparent permeability of serum albumin was used as a standard. The value that we obtained for serum albumin agreed with previously reported values for apparent tumor vascular permeability that were obtained with intravital microscopy techniques (52–54). These results therefore support the validity of our new method of measuring tumor vascular apparent permeability. However, the statistically significant dependence of apparent permeability on molecular weight, as shown in Fig. 3, is in contrast to previous findings by Yuan et al. (42), who did not find a statistically significant dependence on molecular weight. The discrepancy between their results and our results may arise from different methods of calculating apparent permeability and different experimental designs. Yuan et al. used proteins with a narrower molecular weight range of between 25 and 150 kDa and one liposome to determine their apparent tumor vascular permeability, whereas we used the dextrans. The proteins used by Yuan et al. also had various charges and conformations that may have increased the variability in their measurements and thus reduced their ability to detect a statistically significant dependence of apparent tumor vascular permeability on molecular weight.
The molecular weight dependence of the apparent tumor vascular permeability (Fig. 3) was qualitatively different from the permeability of normal capillaries. The permeability of normal capillaries rapidly decreased for molecules of increasing hydrodynamic radius up to 3.6 nm, which is the size of serum albumin, and then slowly decreased for molecules with a larger hydrodynamic radius (41). In contrast, the apparent permeability of tumor vasculature continuously decreased from a hydrodynamic radius of 1.6 to 25 nm, without a pronounced lower threshold (Fig. 3), suggesting that transport across the tumor vascular wall was less tightly regulated than transport across normal vasculature. Tumor vasculature has large intercellular openings between the endothelial cells (55) that are not present in normal continuous vasculature and contain vesiculo-vacuolar organelles in the tumor endothelium (56). These ultrastructural differences between normal and tumor vasculature probably explain the increased vascular permeability observed in tumors.
Accumulation of Macromolecules in Tumors
The vascular AUC increased as the molecular weight of the dextran increased, as shown in Table 1, which we expected from the longer plasma half-life of larger molecules (34–36). The power of the enhanced permeability and retention effect was demonstrated by the maximal extravascular AUC observed for the 40- and 70-kDa dextrans. The 40- and 70-kDa dextrans performed better than their lower molecular weight counterparts because they combined a moderate apparent permeability with a greater vascular AUC and with a slower rate of clearance (26). The 2-MDa dextran had such low apparent permeability that, although the vascular AUC was substantially higher, the extravascular accumulation was limited.
Our study had several limitations. The imaging studies were only performed for 30 minutes, which was not long enough to capture the entire time course of tumor accumulation for the larger dextrans. To gain data for the entire time course of tumor accumulation (i.e., the time-course duration, defined as the time required for the tumor extravascular concentration to decrease to less than 0.1 %Vmax), we applied commonly used compartmental modeling techniques to our experimental data. It would be preferable to capture the entire time course experimentally, but the high spatial and temporal resolution of our imaging technique should have increased the accuracy of the compartment model. The noninvasive imaging technique that we used also had the disadvantage of collecting dextran concentrations only in a tumor and cannot simultaneously report values in any other organ.
Previous investigations (36–38,40) of the influence of molecular weight on tumor accumulation and biodistribution used non–image-based radiolabeling techniques that had the advantage of acquiring the concentration of macromolecules in many organs, which may indicate the degree of systemic toxicity. However, our compartment model focused only on a tumor and predicted maximal accumulation for dextrans with molecular weights between 40 and 70 kDa (i.e., a diameter of 11.2–14.6 nm), whereas Tabata et al. (38) reported that a 60-nm drug carrier yielded maximal accumulation in tumors in a single 3-hour measurement. In our study, 2-MDa dextran (diameter = 50 nm) reached its highest concentration in a tumor at 3 hours. This result is consistent with their findings (38) and reinforces the accuracy of our technique.
Penetration of Macromolecules Into Tumors
The interstitial transport of molecules is characteristically described by their effective interstitial diffusion coefficient, which decreases as molecular weight is increased (43). Although an effective interstitial diffusion coefficient is used to determine the rate of transport through the extracellular matrix, other factors such as pore interconnectedness may also limit penetration of macromolecules more than the effective interstitial diffusion coefficient would predict (48). Moreover, macromolecular therapeutic agents must often bind to a receptor to induce their therapeutic effect (1,11). This binding may further limit penetration through the "binding site barrier" (57,58). The methods used in this study may better quantify the influence of extracellular matrix structure and binding on the distribution of therapeutic agents for numerous conditions because it provides the actual three-dimensional tissue distribution, instead of a prediction based on an effective interstitial diffusion coefficient.
Implications for Macromolecular Drug Delivery
Therapeutic implications of accumulation and spatial distribution of macromolecular anticancer agents in a tumor are currently unknown because, before the development of the methods used in this study, it was not possible to simultaneously visualize and quantify both events in vivo. We believe that our results are important because they provide guidelines to optimize the use of macromolecular drug carriers. In traditional chemotherapy and systemic radionuclide therapy, the goal is to deliver a high cumulative exposure of drugs to all tumor cells. For this purpose, our results indicate that dextrans with a molecular weight between 40 and 70 kDa are the optimal drug carriers, because they would provide the highest tumor cell exposure to drug, expressed as the extravascular AUC. However, drug carriers larger than 10 kDa were concentrated within approximately 15 µm of the blood vessel wall, which caused the extravascular AUC to underestimate drug exposure of tumor cells close to the vascular surface and overestimate drug exposure far from the vascular surface. This limited exposure of cancer cells far from the vascular surface to drug may be unfavorable for tumor therapy because these cells would receive a lower dose of drug than cells close to the vascular surface.
Despite the limited penetration of dextrans larger than 10 kDa, a therapeutic agent linked to such a dextran would be localized near the vascular surface, where cancer cells proliferate more rapidly, compared with those located farther away from the vascular surface (59), and thus these cells should be inherently more sensitive to common anticancer agents. This higher concentration of therapeutic agents in the vicinity of more sensitive cancer cells may lead to a greater overall therapeutic efficacy of that agent. The distribution profile near the vascular surface may also generate a greater antivascular effect by delivering a higher dose to tumor endothelium. Indeed, our results suggest that the improved tumor therapy observed with antiangiogenic macromolecule conjugates, as previously reported (60,61), may be caused, in part, by the higher concentration of macromolecular drug carriers near the tumor blood vessel wall.
In addition to the accumulation and spatial distribution of drug carriers in tumors, the state of the drug must also be considered (3). For most macromolecular drug carriers, the drug is covalently linked to the polymer through an enzymatically or hydrolytically labile bond that must be broken to liberate the therapeutically active drug (10,62,63). If the drug carrier is larger than 10 kDa and hence localized near the vascular surface, then the drug itself could be released from its carrier when it is near the vascular surface so that the drug could penetrate much deeper into tumor tissue. This may in fact be the mechanism by which liposomal drug carriers act, by first concentrating in the tumor tissue over 1–3 days and then slowly releasing the drug over the next 1–2 weeks, so that the drug penetrates deep into tumor tissue down a diffusion gradient (64,65). Furthermore, the accumulation of liposomes in a tumor differs from that of the macromolecules examined in this study because the plasma half-life of stealth liposomes is approximately 2 days (64), which would allow the drug to accumulate over a much longer time even though vascular permeability may be less for a liposome than for a macromolecular drug carrier. As we observed for macromolecular drug carriers, the optimum liposome size to balance plasma half-life, tumor vascular permeability, and vascular or lymphatic clearance to maximize drug accumulation in tumors has been reported to be 100 nm (64).
To select an optimum molecular weight for a macromolecular drug carrier, additional factors also need to be considered, including properties of the macromolecule and the structure of the tumor tissue, because both affect the transport of macromolecules. There is simply no perfect molecular weight. For example, scaling of the molecular size (i.e., hydrodynamic radius) with molecular weight will depend on the specific macromolecule (49,66), and the conformation and/or flexibility of a macromolecule will influence its permeability (29). Furthermore, regardless of molecular weight, some macromolecules have an innate affinity for specific organs, such as the intrinsic localization of the polysaccharide pullulan to the liver (35). Finally, the vascular permeability and effective interstitial diffusion coefficient of a macromolecule depend on tumor type and its anatomical location (43,67).
In conclusion, we found that the tumor apparent permeability decreased statistically significantly with increasing molecular weight (range = 3.3 kDa to 2 MDa), thereby slowing the rate of extravasation for larger macromolecules. Despite the decreased apparent permeability of larger macromolecules, the reduced rate of clearance and greater vascular AUC resulted in the greatest tumor accumulation for dextrans with a molecular weight between 40 and 70 kDa, which is in the range of clinically available and successful macromolecular drug carriers (11). Although the larger macromolecules displayed the highest accumulation in tumors, they penetrated only a relatively short distance into the tumor and were mostly concentrated near the vascular surface. The findings of this study are important because they can be used to optimize delivery of all macromolecular therapeutic agents, including cytokines, antibodies, and antiangiogenic drugs (9,60,61,68) that are gaining importance as therapeutic agents (1,11).
|
NOTES |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
This research was supported by National Institutes of Health grant R01 EB00188-01 to A. Chilkoti and R01 CA47245 to M. W. Dewhirst.
We thank Dr. Bruce Klitzman, Dr. George Truskey, and Dr. David Katz for their useful discussions on macromolecule transport.
|
REFERENCES |
|---|
|
TopNotesAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
(1) Jain RK. Delivery of molecular and cellular medicine to solid tumors. Adv Drug Deliv Rev 2001;46:149–68.[CrossRef][ISI][Medline]
(2) Moses MA, Brem H, Langer R. Advancing the field of drug delivery: taking aim at cancer. Cancer Cell 2003;4:337–41.[CrossRef][ISI][Medline]
(3) Allen TM, Cullis PR. Drug delivery systems: entering the mainstream. Science 2004;303:1818–22.
(4) Ehrlich P. Collected studies on immunity. 1st ed. New York (NY): John Wiley and Sons; 1906.
(5) Torchilin VP. Recent advances with liposomes as pharmaceutical carriers. Nat Rev Drug Discov 2005;4:145–60.[CrossRef][ISI][Medline]
(6) Kataoka K, Harada A, Nagasaki Y. Block copolymer micelles for drug delivery: design, characterization and biological significance. Adv Drug Deliv Rev 2001;47:113–31.[CrossRef][ISI][Medline]
(7) Senter PD, Springer CJ. Selective activation of anticancer prodrugs by monoclonal antibody-enzyme conjugates. Adv Drug Deliv Rev 2001;53:247–64.[CrossRef][ISI][Medline]
(8) Vrouenraets MB, Visser GWM, Snow GB, van Dongen G. Basic principles, applications in oncology and improved selectivity of photodynamic therapy. Anticancer Res 2003;23:505–22.[ISI][Medline]
(9) Allen TM. Ligand-targeted therapeutics in anticancer therapy. Nat Rev Cancer 2002;2:750–63.[CrossRef][ISI][Medline]
(10) Kopecek J, Kopeckova P, Minko T, Lu Z. HPMA copolymer-anticancer drug conjugates: design, activity, and mechanism of action. Eur J Pharm Biopharm 2000;50:61–81.[CrossRef][ISI][Medline]
(11) Duncan R. The dawning era of polymer therapeutics. Nat Rev Drug Discov 2003;2:347–60.[CrossRef][ISI][Medline]
(12) Lewis WH. The vascular pattern of tumors. Bull Johns Hopkins Hosp 1927;41:156–62.[ISI]
(13) Sandison JC. Observations on the growth of blood vessels as seen in the transparent chamber introduced into the rabbit's ear. Am J Anat 1928;41:475–96.[CrossRef][ISI]
(14) Ide AG, Baker NH, Warren SL. Vascularization of the Brown-Pearce rabbit epithelioma transplant as seen in the transparent ear chamber. Am J Roentgenol 1939;42:891–9.
(15) Algire GH, Chalkley HW, Legallais FY, Park HD. Vascular reactions of normal and malignant tissues in vivo: vascular reactions of mice to wounds and to normal and neoplastic transplants. J Natl Cancer Inst 1945;6:73–85.[ISI]
(16) Duran-Reynals F. Studies on the localization of dyes and foreign proteins in normal and malignant tissue. Am J Cancer 1939;35:98–107.
(17) Babson AL, Winnick T. Protein transfer in tumor-bearing rats. Cancer Res 1954;14:606–11.
(18) Busch H, Greene HSN. Studies on the metabolism of plasma proteins in tumor-bearing rats. Yale J Biol Med 1955;27:339–49.[Medline]
(19) Dewey WC. Vascular-extravascular exchange of plasma proteins in the rat. Am J Physiol 1959;197:423–31.
(20) Song CW, Levitt SH. Quantitative study of the vascularity in Walker carcinoma 256. Cancer Res 1971;31:587–9.
(21) Underwood JCE, Carr I. The ultrastructure and permeability characteristics of the blood vessel of a transplantable rat sarcoma. J Pathol 1972;107:157–66.[CrossRef][ISI][Medline]
(22) Peterson H-I, Appelgren KL. Experimental studies on the uptake and retention of labeled proteins in a rat tumor. Eur J Cancer 1973;9:543–7.[ISI][Medline]
(23) Heuser LS, Miller FN. Differential macromolecular leakage from the vasculature of tumors. Cancer 1986;57:461–4.[CrossRef][ISI][Medline]
(24) Gerlowski LE, Jain RK. Microvascular permeability of normal and neoplastic tissues. Microvasc Res 1986;31:288–305.[CrossRef][ISI][Medline]
(25) Dvorak HF, Nagy JA, Dvorak JT, Dvorak AM. Identification and characterization of the blood vessels of solid tumors that are leaky to circulating macromolecules. Am J Pathol 1988;133:95–109.[Abstract]
(26) Matsumura Y, Maeda H. A new concept for macromolecular therapeutics in cancer chemotherapy: mechanism of tumoritropic accumulation of proteins and the antitumor agent smancs. Cancer Res 1986;46:6387–92.[ISI][Medline]
(27) Maeda H, Matsumura Y. Tumoritropic and lymphotropic principles of macromolecular drugs. Crit Rev Ther Drug Carrier Syst 1989;6:193–210.[ISI][Medline]
(28) Maeda H, Seymour LW, Miyamoto Y. Conjugates of anticancer agents and polymers: advantages of macromolecular therapeutics in vivo. Bioconjug Chem 1992;3:351–62.[CrossRef][ISI][Medline]
(29) Seymour LW. Passive tumor targeting of soluble macromolecules and drug conjugates. Crit Rev Ther Drug Carrier Syst 1992;9:135–87.[ISI][Medline]
(30) Ringsdorf H. Structure and properties of pharmacologically active polymer. J Polymer Sci Polymer Symp 1975;51:135–53.
(31) Tomlinson E. Passive and active vectoring with microparticles: localization and drug release. J Control Release 1985;2:385–91.[CrossRef]
(32) Sung C, Youle RJ, Dedrick RL. Pharmacokinetic analysis of immunotoxin uptake in solid tumors: role of plasma kinetics, capillary-permeability, and binding. Cancer Res 1990;50:7382–92.
(33) Tabata Y, Kawai T, Murakami Y, Ikada Y. Electric charge influence of dextran derivatives on their tumor accumulation after intravenous injection. Drug Deliv 1997;4:213–21.
(34) Seymour LW, Duncan R, Strohalm J, Kopecek J. Effect of molecular-weight (Mw) of N-(2-hydroxypropyl)methacrylamide copolymers on body distribution and rate of excretion after subcutaneous, intraperitoneal, and intravenous administration to rats. J Biomed Mater Res 1987;21:1341–58.[CrossRef][ISI][Medline]
(35) Yamaoka T, Tabata T, Ikada Y. Body distribution profile of polysaccharides after intravenous administration. Drug Deliv 1993;1:75–82.[Medline]
(36) Takakura Y, Hashida M. Macromolecular carrier systems for targeted drug delivery: pharmacokinetic considerations on biodistribution. Pharm Res 1996;13:820–31.[CrossRef][Medline]
(37) Murakami Y, Tabata Y, Ikada Y. Tumor accumulation of poly(ethylene glycol) with different molecular weights after intravenous injection. Drug Deliv 1997;4:23–31.[Medline]
(38) Tabata T, Murakami Y, Ikada Y. Tumor accumulation of poly(vinyl alcohol) of different sizes after intravenous injection. J Control Release 1998;50:123–33.[CrossRef][ISI][Medline]
(39) Noguchi Y, Wu J, Duncan R, Strohalm J, Ulbrich K, Akaike T, et al. Early phase tumor accumulation of macromolecules: a great difference in clearance rate between tumor and normal tissues. Jpn J Cancer Res 1998;89:307–14.
(40) Seymour LW, Miyamoto Y, Maeda H, Brereton M, Strohalm J, Ulbrich K, et al. Influence of molecular-weight on passive tumor accumulation of a soluble macromolecular drug carrier. Eur J Cancer 1995;31A:766–70.
(41) Michel CC, Curry FE. Microvascular permeability. Physiol Rev 1999;79:703–61.
(42) Yuan F, Dellian M, Fukumura D, Leunig M, Berk DA, Torchilin VP, et al. Vascular permeability in a human tumor xenograft: molecular size dependence and cutoff size. Cancer Res 1995;55:3752–6.
(43) Pluen A, Boucher Y, Ramanujan S, McKee TD, Gohongi T, di Tomaso E, et al. Role of tumor-host interactions in interstitial diffusion of macromolecules: cranial vs. subcutaneous tumors. Proc Natl Acad Sci U S A 2001;98:4628–33.
(44) Papenfuss HD, Gross JF, Intaglietta M, Treese FA. Transparent access chamber for the rat dorsal skin fold. Microvasc Res 1979;18:311–8.[CrossRef][ISI][Medline]
(45) Truskey GA, Yuan F, Katz DF. Transport phenomena in biological systems. Upper Saddle River (NJ): Prentice Hall; 2004.
(46) Brizel DM, Klitzman B, Cook JM, Edwards J, Rosner G, Dewhirst MW. A comparison of tumor and normal tissue microvascular hematocrits and red-cell fluxes in a rat window chamber model. Int J Radiat Oncol Biol Phys 1993;25:269–76.[ISI][Medline]
(47) Krol A, Maresca J, Dewhirst MW, Yuan F. Available volume fraction of macromolecules in the extravascular space of a fibrosarcoma: Implications for drug delivery. Cancer Res 1999;59:4136–41.
(48) Yuan F, Krol A, Tong S. Available space and extracellular transport of macromolecules: effects of pore size and connectedness. Ann Biomed Eng 2001;29:1150–8.[CrossRef][ISI][Medline]
(49) Berk DA, Yuan F, Leunig M, Jain RK. Fluorescence photobleaching with spatial Fourier-analysis: measurement of diffusion in light-scattering media. Biophys J 1993;65:2428–36.[Medline]
(50) Lebrun L, Junter GA. Diffusion of sucrose and dextran through agar gel membranes. Enzyme Microb Technol 1993;15:1057–62.[CrossRef][Medline]
(51) Bernstein A, Hurwitz E, Maron R, Arnon R, Sela M, Wilchek M. Higher antitumor efficacy of daunomycin when linked to dextran: in vivo and in vitro studies. J Natl Cancer Inst 1978;60:379–84.[ISI][Medline]
(52) Wu NZ, Klitzman B, Rosner G, Needham D, Dewhirst MW. Measurement of material extravasation in microvascular networks using fluorescence video-microscopy. Microvasc Res 1993;46:231–53.[CrossRef][ISI][Medline]
(53) Yuan F, Leunig M, Berk DA, Jain RK. Microvascular permeability of albumin, vascular surface-area, and vascular volume measured in human adenocarcinoma LS174T using dorsal chamber in SCID mice. Microvasc Res 1993;45:269–89.[CrossRef][ISI][Medline]
(54) Brown EB, Campbell RB, Tsuzuki Y, Xu L, Carmeliet P, Fukumura D, et al. In vivo measurement of gene expression, angiogenesis and physiological function in tumors using multiphoton laser scanning microscopy [erratum in Nat Med 2001;7:1069]. Nat Med 2001;7:864–8.[CrossRef][ISI][Medline]
(55) Hashizume H, Baluk P, Morikawa S, McLean JW, Thurston G, Roberge S, et al. Openings between defective endothelial cells explain tumor vessel leakiness. Am J Pathol 2000;156:1363–80.
(56) Dvorak AM, Feng D. The vesiculo-vacuolar organelle (VVO): a new endothelial cell permeability organelle. J Histochem Cytochem 2001;49:419–31.
(57) Fujimori K, Covell DG, Fletcher JE, Weinstein JN. A modeling analysis of monoclonal antibody percolation through tumors: a binding-site barrier. J Nucl Med 1990;31:1191–8.
(58) Juweid M, Neumann R, Paik C, Perezbacete MJ, Sato J, Vanosdol W, et al. Micropharmacology of monoclonal antibodies in solid tumors: direct experimental evidence for a binding-site barrier. Cancer Res 1992;52:5144–53.
(59) Wijffels K, Kaanders J, Marres HAM, Bussink J, Peters HPW, Rijken P, et al. Patterns of proliferation related to vasculature in human head-and-neck carcinomas before and after transplantation in nude mice. Int J Radiat Oncol Biol Phys 2001;51:1346–53.[Medline]
(60) Satchi-Fainaro R, Mamluk R, Wang L, Short SM, Nagy JA, Feng D, et al. Inhibition of vessel permeability by TNP-470 and its polymer conjugate, caplostatin. Cancer Cell 2005;7:251–61.[CrossRef][ISI][Medline]
(61) Satchi-Fainaro R, Puder M, Davies JW, Tran HT, Sampson DA, Greene AK, et al. Targeting angiogenesis with a conjugate of HPMA copolymer and TNP-470. Nat Med 2004;10:255–61.[CrossRef][ISI][Medline]
(62) Ulbrich K, Subr V, Strohalm J, Plocova D, Jelinkova M, Rihova B. Polymeric drugs based on conjugates of synthetic and natural macromolecules. I. Synthesis and physico-chemical characterization. J Control Release 2000;64:63–79.[CrossRef][ISI][Medline]
(63) Ulbrich K, Etrych T, Chytil P, Jelinkova M, Rihova B. HPMA copolymers with pH-controlled release of doxorubicin: in vitro cytotoxicity and in vivo antitumor activity. J Control Release 2003;87:33–47.[CrossRef][ISI][Medline]
(64) Charrois GJR, Allen TM. Rate of biodistribution of STEALTH liposomes to tumor and skin: influence of liposome diameter and implications for toxicity and therapeutic activity. Biochim Biophys Acta 2003;1609:102–8.[Medline]
(65) Charrois GJR, Allen TM. Drug release rate influences the pharmacokinetics, biodistribution, therapeutic activity, and toxicity of PEGylated liposomal doxorubicin formulations in murine breast cancer. Biochim Biophys Acta 2004;1663:167–77.[Medline]
(66) Burchard W, Schmidt M, Stockmayer WH. Information on polydispersity and branching from combined quasi-elastic and integrated scattering. Macromolecules 1980;13:1265–72.[CrossRef]
(67) Hobbs SK, Monsky WL, Yuan F, Roberts WG, Griffith L, Torchilin VP, et al. Regulation of transport pathways in tumor vessels: role of tumor type and microenvironment. Proc Natl Acad Sci U S A 1998;95:4607–12.
(68) Harris JM, Chess RB. Effect of PEGylation on pharmaceuticals. Nat Rev Drug Discov 2003;2:214–21.[CrossRef][ISI][Medline]
Manuscript received June 30, 2005; revised December 20, 2005; accepted January 10, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  O. Tredan, C. M. Galmarini, K. Patel, and I. F. Tannock Drug Resistance and the Solid Tumor Microenvironment J Natl Cancer Inst, October 3, 2007; 99(19): 1441 - 1454. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Dreher and A. Chilkoti Toward a Systems Engineering Approach to Cancer Drug Delivery J Natl Cancer Inst, July 4, 2007; 99(13): 983 - 985. [Full Text] [PDF]
|
|