Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Gliomas are a tragic form of brain tumor that inevitably leads to death of the patient. Surgical debulking of the tumor and radiotherapy extend patient survival by only 6 to 12 months (Radhakrishnan et al., 1994), whereas chemotherapy has generally been even less successful. Despite the development of new chemotherapeutic agents, it is generally recognized that the blood-brain tumor barrier (BBTB) limits exposure of the tumor to many potentially efficacious chemotherapeutic drugs (Black, 1995).

Several different approaches have been explored for increasing permeability of the BBTB so that higher concentrations of chemotherapeutics might be achieved in the tumor. The majority of these approaches, including infusion of hyperosmotic mannitol (Neuwelt and Rapoport, 1984; Kroll and Neuwelt 1998), as well as infusion of leukotrienes (Black and Hoff, 1985; Black et al., 1990, 1994), histamine (Inamura et al., 1994a), and bradykinin (Inamura and Black, 1994; Black, 1995), require intra-arterial administration directly into the tumor vasculature. More recently, Cereport (RMP-7) was developed as a pharmaceutical analog of bradykinin, possessing greater selectivity for the constitutively expressed cerebrovascular bradykinin B₂ receptor and a measurably longer half-life than bradykinin (Doctrow et al., 1994; Straub et al., 1994; Bartus et al., 1996b). Cereport increases permeability of the blood-brain barrier (BBB) by transiently disengaging the tight junctions comprising the barrier (Sanovich et al., 1995). Cereport is unique among BBTB-modulating agents in that it is effective after both intracarotid and i.v. administration, although as expected, dose-response comparisons demonstrate higher doses are required with i.v. administration to achieve comparable effects (Bartus et al., 1996b; Elliott et al., 1996b; Emerich et al., 1999). The opportunity to enhance delivery of chemotherapeutic agents to brain tumors via i.v. dosing protocols offers the clear advantages of greater simplicity, safety, and cost-effectiveness.

We performed a series of experiments using an RG2 rat glioma model to address several issues fundamental to the use of receptor-mediated modulation of the BBTB to increase delivery of chemotherapeutics for treating brain tumors and to the continued development of Cereport as the first drug of this class. In the initial experiment, the ability of i.v. Cereport to enhance delivery of the chemotherapeutic agents, carboplatin and paclitaxel, was compared directly using a previously defined optimal dose (Bartus et al., 1996b; Elliott et al., 1996b). Paclitaxel was selected as an interesting, relatively new chemotherapeutic agent with promise for treating brain tumors. Despite its relative lipophilicity, paclitaxel does not normally achieve high concentrations in brain tumors (Heimans et al., 1994; Schinkel et al., 1996), and no information yet exists on the ability of Cereport to enhance its delivery across the BBTB. Carboplatin is a logical choice to be combined with Cereport for treating gliomas because it: 1) is hydrophilic (and therefore would benefit from the increased aqueous pathway created by Cereport across the BBTB), 2) has demonstrated cytotoxic effects against glioma both in vitro and in vivo, 3) has demonstrated dose-escalation effects (supporting the concept that increasing the concentration of carboplatin in the tumor will provide a greater chemotherapeutic effect), and 4) has an acceptable pharmacokinetic profile (i.e., a plasma half-life >1 h with relatively little protein binding during that time). A separate experiment was performed to establish that Cereport actually increases levels of the cytotoxic platinum moiety in the tumor and brain surrounding tumor (BST) and to determine whether the increases achieved with Cereport persist beyond the time of BBTB modulation. Next, the enhanced uptake achieved with carboplatin after Cereport administration was studied in detail using a novel, high spatial resolution analysis of quantitative autoradiography within the tumor area and BST. This analysis provided new and unexpected information on the effects of Cereport within micro-subregions of the tumor and BST and revealed that the most marked effects of Cereport occur in areas of the glioma and BST that are normally relatively impermeable. Finally, the ability of i.v. Cereport to enhance the survival benefit of carboplatin was evaluated in this treatment-resistant rat glioma model. This experiment demonstrated significant effects when a Cereport dose was used that produced consistent and robust increases in carboplatin uptake and is estimated to increase plasma concentrations to the level of the K_i of Cereport but not when a lower dose was used (which produces only modest increases in uptake and plasma levels that are below the K_i of Cereport. Collectively, the data presented offer several new and important findings regarding the use of Cereport to increase delivery of chemotherapeutic agents to treat brain tumors and offer additional insight into the emerging field of receptor-mediated modulation of BBB permeability for treating brain tumors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects. Male Fischer rats (total n = 183, 170-220 g; Taconic Farms, Germantown, NY) were used in these studies. The rats were housed in pairs in polypropylene cages with free access to food and water. All procedures were reviewed and approved by Alkermes' Animal Care and Use Committee and were conducted in a manner that met or exceeded National Institutes of Health standards.

Tumor Cell Implantation. Rat glioma (RG2) cells were maintained and implanted as described previously (Bartus et al., 1996a; Elliott et al., 1996b). Briefly, rats were anesthetized with a solution containing ketamine (33.33 mg/ml), xylazine (1.0 mg/ml), and acepromazine (1.66 mg/ml) and placed in a stereotaxic instrument. RG2 cells (5 × 10⁴cells/5 µl) were injected unilaterally into the striatum using a stereotaxic-mounted 10-µl Hamilton syringe with a 22-gauge needle at the following coordinates: A-P (+2.0 mm), L (+3.0 mm), and V (6.5 mm) (Pellegrino et al., 1986).

Drug Administration and Physiological Monitoring during Uptake Studies. Studies to quantify the uptake of radiolabeled carboplatin or paclitaxel into glioma were conducted 8 days after tumor implantation. Cannulas were placed in the jugular vein and into both femoral arteries under urethane anesthesia (1.8 g/kg; i.p.). The jugular cannula provided the means to infuse the drugs, whereas the femoral cannula provided a means for measuring physiological parameters and collecting the blood used to calculate the uptake constant, K_i.

Scintillation Studies. To compare the ability of Cereport to enhance uptake of carboplatin and paclitaxel into tumor and BST, animals received infusions of [¹⁴C]carboplatin or [³H]paclitaxel as described above. At the end of drug administration, rats were decapitated, brains were removed, and the tumor was dissected free. Equal amounts of tissue were taken from the ipsilateral and contralateral cortices and the contralateral striatum. Tissue samples were weighed and incubated overnight at 40°C in 1.0 ml of Soluene (Packard Instrument Co., Inc., Meriden, CT). The following day, 10 ml of Hionic Fluor (Packard Instrument Co., Inc.) was added, and radioactivity (nanocuries per gram) was computed using scintillation counts. For determination of carboplatin and paclitaxel uptake into tumor tissue, arterial blood was withdrawn into PE90 tubing at a constant rate (0.004 ml/min) throughout the period after administration of the radiolabel by using a peristaltic pump. Once collected, the blood was removed from the tubing and prepared for scintillation counting. Four groups of animals were used: 1) carboplatin and saline, n = 8; 2) carboplatin and Cereport, n = 8; 3) paclitaxel and saline, n = 18; and 4) paclitaxel and Cereport, n = 14.

Histology and Autoradiography. For autoradiographic determinations of carboplatin uptake into tumor tissue, animals received carboplatin infusions combined with either saline (n = 7) or Cereport (n = 9) as discussed above. At the end of each drug administration, the rats were decapitated, and their brains were rapidly removed and stored at 20°C until they were sectioned. The brains were cut at 20-µm intervals on a cryostat (16°C) throughout the length of the tumor, and the sections were thaw-mounted onto glass microscope slides. Using standard autoradiographic techniques, the slides were apposed to radiosensitive film (Kodak Biomax MR-1) with ¹⁴C calibration standards (0.002 to 3.58 µCi/g; American Radiolabeled Chemicals, St. Louis, MO) for 1 week and then developed. The slides were then removed and stained with hematoxylin and eosin (H&E) to verify tumor placement.

K_i Calculations. The unidirectional transfer constant, K_i, was calculated as described earlier (Ohno et al., 1978; Ziylan et al., 1988; Inamura and Black, 1994; Inamura et al., 1994a). The K_i value represents the rate of [¹⁴C]carboplatin or [³H]paclitaxel taken up into tissue. The values from the autoradiographic film were quantified as nanocuries per gram of tissue; radioactivity within tumor and brain tissue were defined as Ct and Cbr, respectively. Blood volumes of both tumor and brain were derived from previously published work and were calculated to be 9.4 and 3.7 µl/g, respectively (Inamura et al., 1994b). Cbl was defined as the whole blood radioactivity (nanocuries per milliliter) at the time of sacrifice. Cbt was defined as the total blood radioactivity (nanocuries per milliliter per minute), collected throughout the infusion period by continuously sampling blood (T) via a cannula implanted into the femoral artery. The formula for calculating the K_i is as follows:

Quantitative Determination of Carboplatin Uptake in Tumor, BST, and Normal Brain. Quantitative analysis of the regional radioactivity in the brain sections was performed using an image analysis system (ImagePro Plus; Media Cybernetics, Silver Spring, MD). Individual coronal brain sections stained with H&E were digitized to define the exact tumor location and boundary. This coronal section was then overlaid on the identical autoradiographic film image, and the image was digitized. Using the H&E-stained section to define the tumor boundary, the total radioactivity within it and in a 1-mm ribbon of the BST immediately surrounding but clearly outside the tumor were measured. Ipsilateral brain tissue was defined as cortical tissue at least 2 mm from the tumor border but within the ipsilateral hemisphere. Contralateral tissue was defined as striatal or cortical tissue on the opposite hemisphere of the brain. To convert the film images into units of radioactivity, the optical densities of images produced by the ¹⁴C standards were determined, and a standard curve relating optical density to tissue radioactivity was generated for each film. Based on this curve, the amount of radioactivity (nanocuries per gram) within the various brain regions was computed. For each animal, the mean total radioactivity was calculated from three separate coronal sections containing the largest cross-sectional area of the tumor.

High Spatial Resolution Analysis of Uptake in Tumor and BST. To provide a more detailed analysis of the effects of Cereport within the tumor and BST, a method was developed to quantify uptake of radiolabeled carboplatin with extremely sensitive spatial resolution (4.68-µm² area). This was accomplished by separately evaluating each individual pixel comprising the digitized autoradiographic images. The level of radiolabeled carboplatin from each individual pixel within the tumor and BST was computed and placed within bins representing varying levels of radioactivity (e.g., 0-10, 11-20, 21-30 nCi/g, etc.). This permitted a high spatial resolution profile to be constructed, reflecting both the extent of permeability in the tumor and BST as well as its spatial variability. By comparing the shape of these profiles under vehicle and Cereport conditions, it should be possible to gain greater insight into the uptake effects achieved with Cereport, and in particular, where within the tumor and BST Cereport may be exerting its greatest effects.

Platinum Determinations and Measurement of Residence Time in Tumor. To confirm that the increases in carboplatin levels achieved by Cereport reflect increases in platinum in the tumor and to determine whether the increases remain elevated over time, a separate study was performed using unlabeled (i.e., cold) carboplatin. Two groups were given i.v. carboplatin (10 mg/kg) only, and two received i.v. carboplatin plus Cereport (9.0 µg/kg) as described earlier for the uptake studies. One carboplatin group (n = 12) and one carboplatin plus Cereport group (n = 12) were sacrificed immediately after the drug infusion (corresponding to the timing of the uptake studies), whereas the two other groups were sacrificed 2 h later (n = 12/group).

Survival Studies. To determine whether the enhanced uptake of carboplatin into glioma produced by Cereport prolongs survival, animals received i.v. infusions of carboplatin and/or Cereport/vehicle on days 7 and 9 postglioma implantation (i.e., 1 day before and 1 day after the days that the uptake studies were performed).

Statistics. The effects of Cereport on uptake of carboplatin into tumor, BST, and normal brain were compared in rats using a one-way ANOVA (JMP; SAS Institute Inc., Cary, NC). Data from both the survival studies and the high spatial resolution analysis of carboplatin uptake (i.e., pixel analysis) were analyzed using ². Minimal statistical significance in all cases was defined as P < .01 with all two-way comparisons employing two-tailed tests.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Experiment 1: Effect of i.v. Cereport on Uptake of Carboplatin and Paclitaxel. Using scintillation counting, it was determined that Cereport produced a significant and robust increase in uptake of carboplatin into tumor (P < .0001; Fig. 1). In nontumor brain regions, only nonsignificant trends were seen (data not shown; P > .05). No effect of Cereport on uptake of paclitaxel was seen in tumor or any brain region (P > .10; Fig. 1).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Comparison of differences in uptake of [14C]carboplatin and [3H]paclitaxel, achieved with i.v. Cereport (9.0 µg/kg) versus saline vehicle as determined by scintillation counts (Ki represents the unidirectional transfer constant of carboplatin from blood into brain tumor or tissue). Note that although a greater than 2-fold increase in uptake of carboplatin was achieved in tumor, no increase in uptake of paclitaxel was observed. These data are consistent with the action of Cereport on the tight junctions of the BBB, which provides a transient, aqueous pathway from the vasculature into the brain. The effects of Cereport on carboplatin uptake in nontumor brain were much less robust than those achieved in tumor (P > .05; data not shown). Error bars represent ±S.E.M.

Experiment 2: Retention of Elevated Platinum Levels in Tumor with Cereport. Atomic absorption spectrophotometry confirmed that the increase in carboplatin uptake achieved with Cereport reflects an elevation in platinum levels within the tumor. The platinum levels from the rats sacrificed at the end of the Cereport infusion (i.e., 5 min after the end of the carboplatin infusion) were over 2-fold higher than those from the vehicle controls (Fig. 2; P < .001). The 2-fold difference achieved by Cereport persisted at the 2 h time point, although apparent clearance of carboplatin from the interstitial brain fluid produced a significant decrease in signal in both groups.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Comparison of platinum levels in tumor as measured by atomic absorption spectrophotometry. Left panel, comparison of the level of platinum achieved with Cereport versus saline vehicle, measured immediately after the Cereport infusion was terminated (and therefore 5 min after the termination of the carboplatin infusion; see dosing methods in Materials and Methods). Right panel, comparison of the effects of Cereport versus vehicle 2 h and 5 min after the carboplatin infusion ended. Note that despite the drop in absolute carboplatin levels in the tumor over time, the 2-fold increase in carboplatin levels achieved with Cereport persisted. Error bars represent ±S.E.M.

Experiment 3: High Spatial Resolution Analysis of Carboplatin Uptake into Tumor and BST. Using quantitative autoradiography of the entire tumor and BST, it was determined that Cereport increased uptake of carboplatin by over 2-fold in each (130%, P < .001; Fig. 3). In brain regions not associated with the glioma, nonsignificant trends were observed (P > .05; Fig. 3). Collectively, these data are consistent with previous reports of the effects of Cereport on uptake of carboplatin in rat glioma models (Inamura et al., 1994b; Elliott et al., 1996a,b; Matsukado et al., 1996, 1998) and in human glioma patients (Warkne et al., 1995; Ford et al., 1996; Black et al., 1997).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effect of i.v. Cereport (9.0 µg/kg) on uptake of [14C]carboplatin as measured by quantitative autoradiography. The Ki for the entire area of tumor, BST, and nontumor brain regions (including contralateral striatum, ipsilateral cortex, and contralateral cortex) is depicted. Note the significant and selective effects of Cereport in the tumor and BST with much less robust and consistent effects in the nontumor brain regions. Error bars represent ±S.E.M.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   High spatial resolution profile of carboplatin uptake within tumor. Top, graphs depict the proportion of tumor displaying varying degrees of radiolabeled carboplatin. Each vertical bar represents progressively greater degrees of permeability (i.e., progressively greater amounts of radiolabeled carboplatin). Note the heterogeneity in permeability displayed under both saline vehicle (left) and Cereport (right) conditions. Note also that Cereport both significantly reduced the proportion of tumor that was relatively impermeable (i.e., 0-20 nCi/g) and also increased the proportion of tumor where carboplatin levels were relatively high (i.e., >50 nCi/g). Bottom, graph presents the same data in the form of cumulative frequency (i.e., percentage of pixels within the tumor containing increasingly greater amounts of radiolabeled carboplatin) plotted as a function of the level of radioactive carboplatin. Note the significant drop in the frequency of low nanocurie-level pixels with Cereport together with a substantially higher proportion of higher nanocurie-level pixels. See Materials and Methods for methodological details. This figure depicts the same raw data that are averaged in Fig. 3, Tumor. Error bars represent ±S.E.M.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Direct comparison of the high spatial resolution analysis of carboplatin uptake into tumor achieved with Cereport versus a simulated, across-the-board 130% increase in vehicle scores (this model was chosen because 130% represents the mean difference in uptake between Cereport and saline vehicle achieved in this study; see Fig. 3). Note that the patterns of uptake (plotting the proportion of pixels as a function of [14C]carboplatin levels) are markedly different (P < .001), reflecting the fact that Cereport both selectively increases uptake in relatively impermeable areas within the tumor (i.e., those with <20 nCi/g [14C]carboplatin) as well as selectively increases the proportion of tumor with relatively high levels of carboplatin (i.e., >60 nCi/g). Error bars represent ±S.E.M.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   High spatial resolution profile of carboplatin uptake within BST. The analysis is the same as that used for tumor shown in Fig. 4. Note that the BST under vehicle conditions tends to be somewhat less permeable than the tumor (i.e., greater proportion of area represented by low nanocurie levels). Note also that the effects of Cereport on BST appear to be similar to those in tumor with a dramatic reduction in the proportion of areas that are relatively impermeable to radiolabeled carboplatin. This figure depicts that same raw data that are averaged in Fig. 3, BST. Error bars represent ±S.E.M.

Experiment 4: Increased Survival after Cereport and Carboplatin. Animals treated with saline exhibited a median survival of 19 days and a maximum survival of 24 days. Rats given Cereport only (9.0 µg/kg) displayed a nearly identical survival curve to vehicle-treated rats (Fig. 7). Carboplatin (10 mg/kg, given on days 7 and 9) significantly enhanced both median and maximum survival (Fig. 7), increasing each by approximately 10 days (P < .05). The low dose of Cereport (i.e., 3.0 µg/kg) in combination with carboplatin failed to further increase survival beyond that achieved with carboplatin alone. However, the higher dose of Cereport (9.0 µg/kg) produced a marked increase in survival, increasing median survival to 36 days and maximum survival to 60 days (in each instance, nearly a 2-fold increase over carboplatin alone; P < .01).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Kaplan Meier survival curves for rats implanted with RG2 gliomas and treated 7 and 9 days later with various combinations of carboplatin, Cereport, and/or saline vehicle. Note that animals given either saline vehicle or Cereport show very similar survival functions. Those given carboplatin (10 mg/kg per each treatment) or carboplatin plus a low dose of Cereport (3.0 µg/kg) survived significantly longer than those given saline vehicle but were not significantly different from each other. However, when carboplatin was combined with a higher dose of Cereport (9.0 µg/kg), a significant increase in median and maximum survival was achieved relative to the same dose of carboplatin alone or carboplatin plus a low dose of Cereport.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The experiments reported here establish several novel and important points, demonstrating that: 1) the increased uptake of radiolabeled carboplatin reported previously reflects a genuine increase in platinum levels in the tumor and BST; 2) the residence time for the elevated platinum lasts at least 2 h; 3) Cereport is particularly effective in eliminating those portions of the BBTB (in both tumor and BST) that are normally impermeable to carboplatin; 4) contrary to the effects achieved with hydrophilic carboplatin, lipophilic paclitaxel does not benefit from Cereport despite the fact that it also does not easily penetrate the BBB; and 5) i.v. doses of Cereport that increase carboplatin uptake and were estimated to achieve plasma levels approximating the K_i of Cereport significantly enhance survival when combined with carboplatin; a lower dose of Cereport, which produces marginal uptake effects and Cereport plasma levels, had no effect on survival when combined with carboplatin.

These findings raise several important implications. The 2-fold increase in platinum levels observed (using atomic absorption spectrophotometry) near the end of the Cereport infusion closely mirrored the increase in [¹⁴C]carboplatin obtained via scintillation and autoradiography. This establishes for the first time that the increased uptake of radiolabel seen in these and other studies indeed reflects increased levels of active chemotherapeutic moiety. It also establishes an effect using therapeutic (as opposed to trace) levels of carboplatin. Importantly, the 2-fold advantage in platinum levels achieved with Cereport persisted for at least 2 h, providing the first demonstration that the increased levels achieved with Cereport persist beyond the transient opening of the BBTB, perhaps providing enough residence time to invoke a cytotoxic effect on the tumor cells.

As in previous studies using this and similar glioma models (Inamura et al., 1994b; Bartus et al., 1996a,b; Elliott et al. 1996b,c; Matsukado et al., 1996, Black et al., 1997; Emerich et al., 1999) as well as human glioma patients (Ford et al., 1996; Black et al., 1997; Cloughesy et al., 1999), the increased uptake achieved with Cereport was found to be much greater and more consistent in tumor and BST than in brain tissue distal to tumor. In this study, the greater than 2-fold increases in carboplatin achieved with Cereport in the tumor and BST are contrasted with only marginal increases in other nontumor brain regions.

We used a more detailed and novel, high spatial resolution analysis of the uptake within tumor and BST to further characterize the Cereport-induced changes. This analysis revealed that the BBTB in this model is normally highly heterogeneous (similar to that reported for human gliomas). Interestingly, Cereport did not simply uniformly enhance the uptake of carboplatin to produce the >2-fold change, because with Cereport, almost no portion of any region remained relatively impermeable to carboplatin. Additionally, a greater proportion of highly permeable areas was generated. Thus, the effect of Cereport is not manifested as an indiscriminate doubling of carboplatin levels because the shape of the uptake profile was clearly modified (Figs. 4 and 6). This is most obvious when the pattern of uptake with Cereport in tumor is directly compared with that of a simulated profile achieved by increasing all the raw scores from the vehicle group by 130% (i.e., the difference in overall uptake between Cereport and vehicle; Fig. 5). Of particular interest is the change in the pattern of carboplatin delivery to normally impermeable areas of the tumor and BST, which is significantly different with Cereport from what would be expected from a simple 130% increase in carboplatin uptake. In certain subareas of the tumor, the increased uptake achieved with Cereport is severalfold greater than that achieved with vehicle. This change in the topographic uptake profile with Cereport, involving dramatic differences in certain aspects of carboplatin uptake, may help explain the marked increase in survival associated with the 2-fold increase in overall carboplatin levels with Cereport. Further research will be required to confirm the therapeutic importance of this modified uptake profile as well as to determine the means by which Cereport changes the distribution of carboplatin within tumor and BST. It is surprising that although Cereport is much less effective in nonpermeable normal brain, it is able to substantially reduce areas of impermeability within the tumor and BST. This finding suggests that the relative selectivity Cereport displays for tumor and tumor-associated brain tissue is not simply a function of the tumor tissue being more leaky (see Bartus, 1999 for more detailed discussion of alternative explanations).

The dose comparison of Cereport in the survival studies provides new and important information. It provides the first evidence for a therapeutic benefit in brain tumors with any i.v. modification of the BBTB. The higher dose of Cereport (9.0 µg/kg) was selected because it falls well within the optimal portion of the i.v. dose-response function established previously (Bartus et al., 1996b; Elliott et al., 1996b; Emerich et al., 1999), and it was estimated to produce Cereport plasma levels that range from 25 to 50 nM during the 10-min infusion (i.e., within the range of the K_i of Cereport for stimulating bradykinin B₂ receptors) (Doctrow et al., 1994; Bartus et al., 1996b). Importantly, this dose enhanced the survival benefit of carboplatin by greater than 2-fold. In contrast, 3.0 µg/kg falls within the Cereport dose range where the carboplatin uptake effects are marginal (Bartus et al., 1996b; Elliott et al., 1996b) and the Cereport plasma levels are estimated to range from 8 to 16 nM during the 10-min infusion (i.e., on the extreme low end of the K_i of Cereport) (Doctrow et al., 1994; Bartus et al., 1996b). Interestingly, this dose did not affect survival when combined with carboplatin. The dose-related survival effects reported here, therefore, provide further insight into the plasma concentrations of Cereport required to achieve significant biological efficacy. Using the identical pharmacokinetic methods to model Cereport plasma levels in human patients, it is estimated that doses in the range of 1200 to 1500 ng/kg are likely required to achieve similarly effective plasma levels of Cereport in humans. However, initial phase II studies of Cereport combined with carboplatin used a Cereport dose of only 300 ng/kg (Prados et al., 1997; Gregor et al., 1999), which would produce Cereport plasma levels substantially below those that these data argue are necessary to achieve efficacy in gliomas. For these reasons, higher Cereport doses are now being studied in combination with carboplatin in brain tumor patients.

The lack of significant uptake effects with the lipophilic drug, paclitaxel, are perhaps as interesting as the positive effects achieved with carboplatin. The lack of any increase in paclitaxel uptake with Cereport is reminiscent of previous data with BCNU (another lipophilic chemotherapeutic drug) (Bartus et al., 1996a). Together, these data are consistent with the action of Cereport on the tight junctions of the BBB (Sanovich et al., 1995), whereby Cereport creates an aqueous pathway for water-soluble agents to diffuse from the blood, between the endothelial cells of the BBB, and into the brain. Because lipophilic agents diffuse into the brain through the lipid phase of the membranes of endothelial cells, they neither need nor are aided by the transient aqueous corridor created by Cereport. Interestingly, despite its lipophilic nature, paclitaxel does not reach high concentrations within brain or brain tumors, in part because it is rapidly transported back to the luminal side of the cerebral vessels by the P-glycoprotein system (Schinkel et al., 1996). The inability of Cereport to increase paclitaxel concentrations in the tumor provides circumstantial evidence that Cereport does not inhibit the P-glycoprotein transport system at least within the confines of these dosing parameters.

Cereport was developed to enhance the effects of concomitantly administered chemotherapeutic agents that cannot gain access to brain tumors because of the tight junctions of the BBTB. As a treatment option for brain tumors, it is unique in that it has no cytotoxic activity by itself. To be optimally effective, a delivery agent like Cereport must satisfy several important goals. First, the amount of chemotherapeutic delivered to the tumor and BST should be significantly increased. Secondly, the distribution of the chemotherapeutic within the tumor and BST should be improved so that no area of the tumor or BST can escape the chemotherapeutic agent. Third, the time that the chemotherapeutic is elevated should be increased. Ideally, all of these effects should occur simultaneously and selectively in tumor-associated tissue (i.e., tumor and BST) with little or no effect in normal brain distal to tumor. These data demonstrate that i.v. Cereport achieves all of these goals with carboplatin in the RG2 rat glioma model, making it the only BBTB drug delivery method to date to do so. Moreover, when the identical dosing paradigm is used (substituting therapeutic concentrations of cold carboplatin for radiolabeled carboplatin), a significant 2-fold increase in survival is achieved. These data, therefore, offer clear evidence that modulating cerebrovascular bradykinin activity with selective B₂ agonists such as Cereport represents a novel and potentially valuable means of enhancing the chemotherapeutic effects of carboplatin in gliomas and perhaps other types of brain tumors.