Local Delivery of Poly Lactic-co-glycolic Acid Microspheres Containing Imatinib Mesylate Inhibits Intracranial Xenograft Glioma Growth Free

Glioblastoma multiforme is the most common subtype of primary brain tumor. The long-term survival of these patients remains poor despite a multitude of treatments including surgical resection, followed by radiotherapy and chemotherapy (1, 2). As the challenge of treating glioblastoma multiformes is enormous, new drugs and targets are constantly being investigated. Previous studies indicate that the platelet-derived growth factor (PDGF) receptor (PDGFR) autocrine loop contributes significantly to the pathogenesis and the angiogenesis process associated with malignant gliomas (3–7), and direct inhibition of PDGFR leads to tumor growth arrest (8). Imatinib mesylate (STI571) is a tyrosine kinase inhibitor of the 2-phenylaminopyrimidine class that targets the activated Abl oncoprotein and certain members of the subgroup III receptor tyrosine kinase family including the receptors for PDGF (3–5). Imatinib mesylate is now the standard of care for patients with chronic myelogneous leukemia in which the abl pathway is activated. However, the large amounts of imatinib mesylate that are needed to reach therapeutic concentrations in the brain support the need for a local delivery of imatinib mesylate to the tumor site. Moreover, local continuous treatment can eliminate the need of daily administrations and prolong treatment efficiency.

We have previously reported the development of poly (lactic-co-glycolic) acid (PLGA) microspheres as a local delivery system for endogenous inhibitors of angiogenesis in glioblastoma multiforme (9). PLGA is a Food and Drug Administration–approved polymer with biodegradability properties that can be designed to control the release kinetics of a drug by varying the polymer composition (10, 11). In the present study the optimal protocol for imatinib mesylate–loaded microspheres was established. The microspheres were then characterized and studied in vitro and in vivo, and their biological efficacy was evaluated in human and mouse glioblastoma mice models. Our results show that a single local injection of microspheres loaded with a low concentration of imatinib leads to a significant reduction in an orthotopic glioblastoma multiforme growth.

Preparation of imatinib mesylate–loaded microspheres. Microspheres loaded with imatinib mesylate (provided by Novartis, SW) were prepared using the double emulsion solvent extraction method with slight modifications (9). Imatinib mesylate was loaded in three different formulated particles made of 200 mg PLGA50:50 (RG502; Boehringer Ingelheim), PLGA75:25, or PLGA85:15 lactic to glycolic acids ratio (20,000 kDa; Birmingham polymers Inc.). Briefly, polymers were dissolved in 0.5 mL of dichloromethane. A predetermined solution of imatinib mesylate (1 mg, 0.5 w/w) was added to the dissolved polymer and the solution was homogenized using ultra-turax (type DI-18 IKA; Staufen) for 1 min leading to the formation of the first emulsion (W/O). Polyvinyl alcohol of 85 to 89 kDa (Sigma-Aldrich Chemical) saturated with dichloromethane was rapidly added to the first W/O emulsion, and the solution was homogenized again for 20 s. The resulting multiple W/O/W were mixed for 5 min and a volume of 50 mL of 0.1% w/v aqueous polyvinyl alcohol containing 5% (v/v) 2-propanol solution was added. After 30 min of extensive stirring the microspheres were centrifuged, washed three times, and lyophilized. Empty microspheres were prepared using the same conditions without the addition of imatinib mesylate.

Size and morphologic studies using scanning electron microscopy and Coulter counter. Scanning electron microscopy (JSM 5400; Jeol) was used to evaluate the shape and surface morphology of the imatinib mesylate–loaded PLGA microspheres. Particle size distribution was analyzed using a Coulter LS 230 particle size analyzer (Beckman Coulter). Samples were prepared by resuspending 10 mg of the microspheres in distilled water. The results were reported as microsphere diameter determined by percent volume distribution, and were analyzed by a model for ideal spheres.

Imatinib mesylate loading efficiency in PLGA microspheres. For all the experiments, the amount of loaded imatinib mesylate per unit weight of microspheres was determined as follows: Fractions of 10 mg microspheres loaded with the imatinib mesylate were digested overnight with 0.1N NaOH, to increase PLGA hydrolysis rate. UV emission (optical density 260 nm) was used to determine total imatinib mesylate loaded in the microspheres, and the amount of imatinib released from the microspheres at different time points. The loading efficiency was obtained by calculating the percent of total imatinib mesylate loaded in the microspheres, divided by the initial imatinib added during the preparation of the microspheres.

In vitro release of imatinib mesylate from PLGA microspheres. For the in vitro release studies, microspheres were incubated in PBS at pH 7.3 and maintained in a shaking incubator at 37°C as previously described (9). Microspheres were recovered from the released media by centrifugation, and the media were replaced. This procedure was repeated every 2 d for 30 d. The concentration of imatinib mesylate in the released media was quantitatively determined using spectrophotometric measurements at optical density (260 nm) and by high performance liquid chromatography with an Agilent 1100-series instrument (Agilent Technologies) using an Eclipse XDB-C18 column (5 mm, 4.6 mmL 50 mm; Agilent Technologies) equipped with a photodiode array detector. The isocratic elution was done with acidic H₂O (0.1% formic acid) and acetonitrile as the mobile phase at a flow rate of 1 mL/min. Compounds were identified by comparison of retention times and UV-visible spectra to those of the appropriate standards.

The percent of the released drug of the total drug in the microspheres was calculated for every sample from the different incubation times and presented as a cumulative curve.

Cell culture. Human glioblastoma cell line U87-MG was obtained from the American Type Culture Collection. The murine glioblastoma cell line GL261 was obtained from the National Cancer Institute-Frederick Cancer Research. U87-MG and GL261 were cultured in DMEM or á-MEM (Invitrogen) supplemented with 10% fetal bovine serum, 2 mmol/L L-glutamine, 2 mmol/L nonessential amino acids, 2 mmol/L sodium pyruvate, and 100 U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL fungizone. Cells were maintained in humidified atmosphere containing 5% CO₂ at 37°C.

Effect of imatinib mesylate on cell viability. The effect of imatinib mesylate on glioma cell viability was determined using a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H tetrazolium bromide (MTT) colorimetric assay (Sigma-Aldrich). Briefly, U87-MG or GL261 cells were seeded in 96-well plates at a density of 9,000 cells/well. After 24 h, imatinib mesylate was added at increasing concentrations (1-10 μmol/L) and incubated for an additional 48 h. In a second set of experiments, 150 μL of microsphere release media were taken at different time points and added to wells seeded with 9,000 cells/well. MTT was then added to each well at a final concentration of 0.5 mg/mL and the incubation was continued for an additional 3 h. To terminate the studies, 125 μL HCl 0.1N in isopropanol were added and samples were read at 570 nm using an automatic microtiter plate reader.

For Western blot analysis; 2 × 10⁶ cells were incubated with or without imatinib mesylate (1,10 μmol/L), washed with ice-cold PBS, and lysed with lysis buffer. Exrtracts were centrifuged (10 min, 10,000 g) and protein content was determined using microBCA assay. Cellular proteins were separated by 7.5% SDS-PAGE and blotted onto a nitrocellulose membrane, which was probed with a primary antibody for phosophorylated PDGFR (pPDGFR)-B or PDGFR-B antibodies (Santa Cruz Biotechnology).

Flow cytometry analysis. Cell cycle fractions were determined by staining the cells with propidium iodide (Sigma-Aldrich). Apoptosis and necrosis were determined by propidium iodide and Annexin-V-Fluos (Roche Molecular Biochemicals) as previously described (12). For cell death analysis, U87-MG and GL261 cells were treated with increasing imatinib mesylate concentrations (1, 5, and 10 μmol/L) for 24 h and were analyzed by FACS (Becton Dickinson FACScan flow cytometer). The percentage of necrotic cells (propidium iodide–positive annexin-positive), and apoptotic cells (propidium iodide-annexin-positive, above 10¹) were determined. For the cell-cycle studies, imatinib mesylate was incubated with cells overnight, followed by FACS analysis. Data were collected using CellQuest software (BD Biosciences) and analyzed using WinMDI 2.8 software.

Immunostaining of tumor cells. U87-MG and GL261 cells were seeded on chamber slides with or without 5 μmol/L imatinib overnight, and fixed with 4% paraformaldehyde for 20 min at room temperature. Cells were blocked with 3% rabbit serum and were incubated with primary antibody: anti-p-PDGFR-B (1:100; Santa Cruz Biotechnology, sc-12907). Immunostaining was carried out using the Vectastain Elite ABC kit (Vector Laboratories) and the detection was done by 3,3'-diaminobenzidine chromogen (Vector Lab) or by FITC-conjugated secondary antibodies (1:200; Molecular Probe Inc.). Micrographs of the 3,3'-diaminobenzidine-stained cells were taken by Nikon microscopy and analyzed with image analysis software (Lucia; Nikon Imaging).

Effect of imatinib mesylate released from microspheres on s.c. glioma tumors. The in vivo effect of imatinib mesylate released from PLGA microspheres was studied in two glioma models: human U87-MG glioblastoma cells in Swiss nude mice, and GL261 murine glioblastoma cells in C57/BL mice. Mice (4- to 6-wk old; Harlan Lab) were inoculated s.c with 1 × 10⁶ GL261 cells or 5 × 10⁶ U87-MG. When the tumors reached the size of 60 mm³ (GL261) or 200 mm³ (U87-MG), the animals were randomly divided into two groups of eight animals each. The first group received a single s.c injection of a 1:1 mixture of 10 mg PLGA50:50 and PLGA85:15 microspheres containing a total of 1.25 mg imatinib mesylate/mice. In the second group, each mouse received a single s.c injection of empty PLGA microspheres of the same composition and ratio. Tumor growth was measured transcutaneously with a caliper every 2 d for 30 d, and tumor volume was calculated (13). All mice were sacrificed 30 d after microspheres injection, and tumor weight and volume were measured. These studies were repeated twice. In addition, all major organs, including liver, kidney, and spleen, were harvested and assayed for any signs of toxicity by H&E staining.

In vivo intracranial model: microspheres degradation, peptide distribution and glioma growth inhibition studies. The therapeutic efficacy of imatinib mesylate–loaded microspheres was examined in an in vivo intracranial model. Briefly, Swiss nude male mice (n = 20) 6- to 8-wk old (Charles River) were anesthetized and stereotactically inoculated with U87-MG cells (1.2 × 10 ⁵ cells in 3 μL of PBS) via a 30-gauge Hamilton syringe into the left forebrain (2.5 mm lateral and 1 mm anterior to bregma, at a 2.5 mm depth from the skull surface). Five days after tumor inoculation, the animals were randomized into two groups (n = 10). Mice were injected via a 22-gauge Hamilton syringe, intratumorally with 1.5 mg PLGA microspheres containing 225 μg imatinib mesylate or empty PLGA microspheres using the same burr hole and stereotactic coordinates. Implantation of microspheres was made at two sites: directly into the tumor center (7.5 μL) and 0.5 mm below the lower poles of the tumor (7.5 μL) along the same trajectory of tumor inoculation. At the end of the experiment, tumors were harvested for immunohistochemistry, and imatinib mesylate levels were analyzed in the tumor specimen using mass spectrometry analysis (LC-MS/MS) from Thermo Fisher Scientific. The LC-MS/MS system was monitored by Xcalibur software (Thermo Fisher Scientific).

Magnetic resonance imaging and tumor volume experiments. Two weeks after intracranial injection of microspheres, animals were imaged by magnetic resonance imaging (MRI) on a Bruker 4.7 T system (Bruker). Tumor volumes were estimated using Gd-enhanced T1 weighted spin-echo images, from which three-dimensional (3D) renderings of the tumors were generated with in-house 3D software (14, 15).

After MRI, animals were perfused with 4% paraformaldehyde under deep anesthesia and brains were processed as previously described (14). Tissue was stained with 4', 6-diamidino-2-phenylindole or H&E as per standard protocol. Intracranial distribution of microspheres was assessed using fluorescent microscopy. Tumor volumes were estimated using the formula for ellipsoid and expressed as a mean ± SE as previously described (14).

Tumor immunohistochemistry. Immunohistochemistry was carried out using the Vectastain Elite ABC kit (Vector Laboratories) as previously described (9). Primary antibodies included anti-CD31 (1:100; BD Biosciences PharMingen) for blood vessel density, anti-Ki67 nuclear antigen (1:100; DAKO) for proliferation, PDGFR-B and p-PDGFR-B (1:100; Santa Cruz Biotechnology, sc-339, sc-12909) for the levels of the pan and phosphorylated receptor; and anticleaved caspase 3 (1:100; Cell Signaling Technology) for the detection of apoptosis. Negative control slides were obtained by omitting the primary antibody. Image analyses software (Lucia; Imaging Laboratory) was used for the quantification of staining. The analysis of tissues was done for at least 10 randomly chosen fields per section in 3 to 5 sections (×200 and ×400). The blood vessel density and apoptotic indices were defined as the percentage of positively stained cells of 100 nuclei from five randomly chosen high-power fields.

Statistics. Results are presented as mean ± SD or as percentage of mean of control ± SD. Comparisons between groups were made by using a two-tailed t-test. In the intracranial model, significant differences in tumor volume, microvessel density, proliferation, and apoptosis index were determined using the Mann-Whitney U test. Values of P < 0.05 were considered statistically significant.

In vitro release of imatinib mesylate from PLGA microspheres. Microspheres prepared from all PLGA copolymer ratios exhibited a spherical, intact surface area (Fig. 1A and B). The microspheres had a mean size of 33.83 μm in diameter as obtained by a Coulter counter and a ζ potential ranging between −14 to −20 mV (Fig. 1C). The release profile of imatinib mesylate was characterized by a continuous release for 17 to 35 days depending on polymer composition (Fig. 1D). Both PLGA50:50 and PLGA75:25 showed a significant initial burst release of 81% and 63%, respectively, during the first three days of release. The release rate of imatinib mesylate from PLGA85:15 microspheres was significantly slower than from PLGA50:50 and PLGA75:25 microspheres. The loading efficiency of imatinib mesylate in PLGA75:25 was 15.3 ± 0.2 with a release period of 17 days, which was significantly lower than in PLGA50:50 (54.6 ± 0.5 for 17 days) and PLGA85:15 microspheres (57.1 ± 0.8 for 30 days).

Effect of imatinib mesylate on cell viability. The effect of imatinib mesylate on glioblastoma cell viability was determined using the MTT colorimetric assay (Fig. 2). Increasing concentrations of free imatinib mesylate (1-20 μmol/L) led to a dose-dependent decrease in cell viability (Fig. 2A). The GL261 cells were more sensitive to imatinib mesylate than were U87-MG cells in the range of 1 to 10 μmol/L. A 1 μmol/L concentration of imatinib mesylate led to a 70% decrease in GL261 viability and a 54% decrease in U87-MG viability (when compared with untreated controls). A larger decrease in cell viability was obtained when 5 or 10 μmol/L of imatinib mesylate were added to GL261 (94% decrease) and U87-MG (88% decrease). Cell viability studies were also done on GL261 cells with imatinib mesylate released from the microspheres to assure its biological activity post entrapment (Fig. 2B). As seen in Fig. 2B, imatinib mesylate released from the PLGA50:50 or PLGA85:15 microspheres retained its biological activity. Imatinib mesylate released from PLGA50:50 microspheres on days 1 to 12 significantly inhibited GL261 viability by 57% to 65% (inhibition levels). Imatinib mesylate released from PLGA85:15 microspheres on days 1 to 20 inhibited GL261 viability by 73% to 77%. Control samples taken from empty microspheres did not affect cell viability (data not shown).

Imatinib mesylate effect on growth arrest of glioblastoma cell lines. To determine whether the inhibitory effect of imatinib mesylate on glioblastoma cells was via cell cycle arrest in addition to cell death, we studied the effect of imatinib mesylate on cell-cycle fractions. Figure 3 shows FACS analysis of the propidium iodide graphs and the quantification of the percentage of cells in the G₀-G₁ phase. A statistically significant induction of G₀-G₁ growth arrest occurred when using imatinib mesylate concentrations of 5 and 10 μmol/L for GL261 and from 1 to 10 μmol/L for U87-MG (Fig. 3A, B, and Supplementary Data). We also evaluated the effect of imatinib mesylate released from the microspheres on cell cycle arrest. As can be seen, imatinib affected the cell-cycle after its release from PLGA50:50 and PLGA85:15 microspheres (Fig. 3A and B). A significant increase in G₀-G₁ cell-cycle arrest was obtained by imatinib mesylate released after 1, 7, and 14 days.

The effect of imatinib mesylate on glioblastoma cell death. FACS analysis was used to study the effect of imatinib mesylate on glioblastoma cell death (necrosis and apoptosis). As seen in Fig. 3C and D, imatinib mesylate induced U87 cell death more by cell necrosis than by apoptosis. A concentration of 10 μmol/L of imatinib mesylate led to a 61% necrotic cell death and 29% apoptotic cells when compared with nontreated cells.

Imatinib mesylate effect on PDGFR phosphorylation. In order to study the effect of imatinib mesylate on PDGFR activation, the phosphorylation of this receptor was detected in GL261 (Fig. 4A) and U87-MG (data not shown) cells by immunohistochemistry using 3,3′-diaminobenzidine chromogen and Western blot (data not shown). Similar immunostaining results were obtained when the detection was carried out by secondary FITC antibody (Fig. 4B). Cells that reacted with FITC antibodies were also analyzed by FACS (data not shown). A significant decrease in the level of pPDGFR-B was observed in the presence of 1 μmol/L of imatinib mesylate. This was also supported by Western blot analysis of GL261 cell extracts after imatinib mesylate treatment. The level of pPDGFR-B per cell was evaluated by image analysis software (Lucia; Nikon Imaging) and the results are presented in percent relatively to untreated cells (Fig. 4C and D). As can be seen in Fig. 4C and D, a significant decrease in pPDGFR-B level was obtained by the addition of 1 μmol/L and 5 μmol/L of imatinib mesylate for both GL261 and U87-MG cells. This level was further decreased when using 10 μmol/L of imatinib mesylate. Moreover, imatinib mesylate that was released after 1, 3, and 14 days from PLGA50:50 and PLGA85:15 microspheres remained active and also reduced pPDGFR-B levels in both cell types (Fig. 4C and D). Similar results were obtained when carrying out Western blots on the treated cells (data not shown).

In vivo effects of imatinib mesylate–loaded PLGA microspheres on the growth of s.c. murine and human glioma tumors. For the s.c in vivo efficacy studies, PLGA50:50 and PLGA85:15 microspheres (1:1 ratio) loaded with imatinib mesylate were implanted into two mice models bearing s.c glioma tumors: GL261 and U87-MG. This ratio is based on preliminary in vivo data in which different PL:GA particles loaded with imatinib mesylate were tested (data not shown) and on the in vitro release kinetics and loading efficiency. The growth of U87-MG glioma xenografts was significantly inhibited by a single dose of microspheres containing a total of 1.25 mg mouse (Fig. 5A). Tumor volume and weight were suppressed by 88% and 77%, respectively, 22 days after injection. Similar results were obtained with microspheres containing imatinib mesylate, which were injected into C57/BL mice bearing GL261 tumors. Imatinib mesylate–loaded microspheres led to a significant inhibition (79%) in GL261 tumor growth 30 days after microspheres injection (Fig. 5A).

Immunohistochemistry for therapeutic efficacy indices was done on imatinib and control s.c U87-MG and GL261 tumors (Fig. 5B, C). Quantification of the immunohistochemistry results shows a significant decrease in the CD-31 angiogenic index (U87-MG, 79% ± 21; GL261, 45% ± 30,), KI67 proliferation index (U87-MG, 82% ± 24; GL261, 54% ± 26), and pPDGFR (U87-MG, 69% ± 31; GL261, 42% ± 22) in the treated mice (Fig. 5D). No significant change (U87-MG, 91% ± 58; GL261, 92% ± 30) was observed for the total level of PDGFR in both tumor models (Fig. 5D). No toxicity was observed in harvested organs (liver, kidney, brain, heart, muscle, spleen) in response to the PLGA–imatinib mesylate treatment.

In vivo intracranial model: human glioma growth inhibition. To assess the therapeutic efficacy of imatinib mesylate microspheres administered intracranially, we monitored tumor volume by MRI followed by histologic analysis. According to the MRI results, imatinib-loaded microspheres significantly inhibited tumor growth 14 days post injection by 79% (Fig. 6A, B, and C). These findings were confirmed by tumor volume measurements done on tumors harvested from the mice 14 days after microsphere injection (Supplementary Data). As seen from Fig. 6A, B, C, and Supplementary Data, a single injection of imatinib mesylate–loaded microspheres resulted in a 79% reduction of tumor volume compared with empty microspheres (10.51 ± 2.47 in imatinib mesylate–loaded microspheres and 50.42 ± 15.68 in empty microspheres; P < 0.01). Mass spectrometric analyses done on brains bearing tumors at the termination of the experiments revealed that >50% of imatinib mesylate was still present in the brain tumor tissue.

The immunohistochemical analysis (Fig. 6D) showed a significant (3.63-fold) increase in tumor cell apoptosis (0.78 + 0.06 for empty and 2.84 + 0.59 for imatinib mesylate microspheres). There was no significant decrease in microvessel count, (22.8 ± 0.72 in empty and 20.9 ± 1.08 in imatinib mesylate microspheres) or proliferation index (26.1 ± 1.69 in imatinib mesylate microspheres and 35.4 ± 2.32 in empty microspheres; data not shown).

The activity and overexpression of PDGF receptors through PDGF autocrine signaling plays a critical role in glioma tumor growth and development (16, 17). Studies have shown that autocrine stimulation of cell proliferation by PDGF was sufficient for glioma formation in mice. Injection of recombinant PDGF-B retrovirus into the mouse brain led to the formation of gliomas (18). Therefore, using inhibitors that block this PDGF stimulation may serve as a potential therapy for glioma. Imatinib mesylate has been shown to inhibit activation of PDGFR by phosphylation and thus the growth of different glioblastoma cell lines such as U87-MG, GL261, U343, and C6 (7, 19–22). Moreover, imatinib mesylate given by i.p and oral administration inhibited the growth of glioblastoma xenografts in nude mice and improved animal survival (20). These preclinical observations led to the initiation of clinical trials on glioblastoma using imatinib mesylate given orally as a monotherapy or in a combination with hydroxyurea (8, 23). In the latest phase I/II study of imatinib mesylate for recurrent gliomas, the patients were administrated with relatively large amounts of imatinib mesylate (orally): 600 mg or 800 mg daily (24, 25). However, this treatment led to disappointing minimal therapeutic achievements and the high drug doses led to significant toxicity and side effects (24, 25). Another problem raised when using imatinib mesylate for glioma therapy is its distribution in the brain. In mice it has been shown that imatinib mesylate was able to penetrate the blood-brain barrier and affected the phosphorylation of PDGFR in the brain (21); this effect was enhanced, however, when the animals were treated with imatinib mesylate followed by radiation therapy. In addition, Haiqing et al. (26) showed that imatinib mesylate is a substrate of p-glycoprotein, which plays an important role in limiting the distribution of imibinib to the central nervous system (26).

The current study aims to improve the therapeutic efficiency of imatinib mesylate for malignant gliomas using a controlled delivery system based on local injection of microspheres directly to the tumor site. By using a local drug delivery system we can target the anticancer drug in the brain, prolong its administration, reduce the amount of drug needed (comparing with systemic administration), and thereby reduce possible toxicity and achieve long-term tumor inhibition. The microsphere drug delivery system, which was designed for these studies, is composed of PLGA copolymers. By varying the comonomers ratio, i.e using PLGA50:50, PLGA75:25, or PLGA85:15, the time course of drug release can range from weeks to months (27).

The in vitro studies revealed that the release kinetics of imatinib mesylate depends on the copolymer ratio. PLGA50:50 and PLGA75:25 exhibited a shorter time period of release than PLGA85:15 microspheres. The different ratios of lactic and glycolic acids monomers of PLGA affect microsphere degradation rate and drug release kinetics. The higher hydrophilicity of PLGA50:50 microspheres enables more extensive water penetration to the polymeric matrix and consequently significantly more drug release during the first days of microsphere incubation in aqueous buffer. In addition, PLGA75:25 microspheres had lower drug loading when compared with the other polymers.

To further elucidate the mechanism by which imatinib mesylate affects malignant glioma growth, the effect of imatinib mesylate on glioblastoma cell viability and cell death (GL261 and U87) was examined. Imatinib mesylate significantly decreased glioblastoma cell viability. In addition, FACS analysis showed that necrosis was the dominant mechanism of cell death rather than apoptosis, although low levels of apoptosis were also detected as previously shown (20, 21).

Our results show that imatinib mesylate also led to a significant cell-cycle arrest in the G₀-G₁ phase (P < 0.001). These data support the results of Kilic et al. (20) that showed U87-MG cell-cycle arrest in G₀-G₁ when using 10 μmol/L of imatinib mesylate. Taken together these results show the retention of the biological activity released from the microspheres for up to 14 days. PDGFR is an attractive target for glioblastoma therapy because PDGF and PDGFR are overexpressed and act in an autocrine loop that promotes glioblastoma proliferation and survival (16). Therefore, it was important to evaluate the inhibitory effect of imatinib mesylate on PDGFR stimulation and activation, which occurs by phosphorylation of the receptor by tyrosine kinases. Our immunostaining, FACS analysis, and Western blot analysis, indicated that the phosphorylated form of PDGFR-B was decreased by imatinib mesylate in both the U87-MG and GL261 cell lines. Most importantly, imatinib mesylate released from both PLGA microspheres (50:50 and 85:15) remained active up to two weeks, and pPDGFR-B levels of GL261 cells treated with imatinib mesylate were reduced compared with the control.

Taken together, all the in vitro results clearly show that the PLGA preparation procedure does not affect the biological activity of imatinib mesylate. Inhibition of cell viability, cell growth, and PDGFR phosphorylation was observed with imatinib released from the microspheres. We then conducted in vivo experiments in two s.c malignant glioma tumor models: the U87-MG xenograft model and GL261 syngenic model. Preliminary in vivo studies done with PLGA particles composed of 50:50 and 85:15 PL to GA ratio and loaded with imatinib mesylate indicated that PLGA50:50 lead to smaller tumor inhibition than PLGA85:15. This is probably due to the faster release of imatinib mesylate when compared with 85:15 (data not shown). Therefore, taking these data and the in vitro release data, we decided to use a mixture of microspheres that contains PLGA50:50 and PLGA85:25 microspheres in a 1:1 ratio. Such a combination may lead to optimal release kinetics of imatinib mesylate, characterized by a significant initial release of imatinib mesylate due to the PLGA50:50 microspheres, followed by slow release due to the PLGA85:15 microspheres. A single injection of a mixture of microspheres (PLGA50:50 and PLGA85:15) containing a total of 1.25 mg imatinib mesylate per mouse significantly inhibited tumor growth, and reduced s.c. tumor volume. The immunohistological analysis of both s.c tumor models showed that imatinib mesylate reduced cell proliferation, angiogenesis, and PDGFR phosphorylation, Therefore, it is possible that tumor growth inhibition occurred due to the inhibitory effect of imatinib mesylate on the phosphorylation of PDGFR, thus leading to a decrease in cell proliferation in the brain.

The long-term efficiency of the imatinib mesylate–loaded microspheres shows the significant advantage of this system over systemic treatment described in the literature. Kilic et al. (20) showed that a two-daily dose of total 50 mg/kg/day (1.5 mg mouse/day) given i.p every day for 30 days led to an 88% inhibition of s.c U87 tumor growth.

Finally, the efficacy of the PLGA–imatinib mesylate system was shown intracranially. From our study it is evident that a single injection of microspheres loaded with only 225 μg imatinib mesylate per mouse led to a significant inhibition of tumor 14 days post treatment as shown by MRI. The inhibition of tumor growth was further supported by the H&E staining of the harvested brains. Mass spectroscopy done on the harvested brains at the end of the experiment revealed that >50% of imatinib mesylate was still detectable in the tissue. This finding indicates that the release rate of imatinib mesylate is significantly slower in vivo compared with the in vitro data which showed that >80% of imatinib mesylateis is released after 2 weeks. These data support our previous finding with PF-4/CTF, a 23–amino acid peptide (28) and unpublished data that show a significant slower rate of degradation of microspheres (50% are still intact 6 weeks post injection) when implanted intracranially. The slower release rate can be attributed to the slower degradation of the microspheres in vivo, due to low water availability in the tissue, compared with the in vitro conditions. In addition, different proteins, which are present in the tissue environment, may absorb to microspheres' surface and slow the release of the imatinib mesylate from the microspheres.

As with the s.c. treatment, the amount of imatinib mesylate used for the intracranial studies is significantly lower than the doses given orally by Kilic et al. (20) for the inhibition of intracranial U87 tumors (50 mg/kg/day, which is about 45 mg/mouse for total treatment of 30 days). In our study we used only 225 μg of imatinib mesylate per mouse.

The significant inhibitory effect of such a low dose of imatinib mesylate indicates the ability of imatinib mesylate released from the microspheres to diffuse throughout the tumor area because the microspheres were injected near the tumor site for s.c. tumor and in the tumor bed for intracranial tumors. No toxicity to normal brain or internal organ such as the liver and spleen was observed. From our data we can conclude that imatinib mesylate–loaded PLGA microspheres can be used as an efficient therapeutic delivery device for brain tumors. The in vitro and in vivo data suggest that the microspheres maintain the biological activity of imatinib mesylate for longer periods and lead to a continuous release of an active drug. Moreover, this system offers an improved alternative for the current therapies by reducing the dose of drug required to reach therapeutic efficacy by using single administration as opposed to multiple doses. In addition, in the future this system may be used in combination with other anticancer drugs or conventional therapies in order to improve therapeutic efficacy and increase patient survival time.

No potential conflicts of interest were disclosed.