Combining a T9/9L glioma vaccine, expressing the membrane form of M-CSF, with a systemic antiangiogenic drug-based therapy theoretically targeted toward growth factor receptors within the tumor’s vasculature successfully treated >90% of the rats bearing 7-day-old intracranial T9/9L gliomas. The antiangiogenic drugs included (Z)-3-[4-(dimethylamino)benzylidenyl]indolin-2-one (a platelet-derived growth factor receptor β and a fibroblast growth factor receptor 1 kinase inhibitor) and oxindole (a vascular endothelial growth factor receptor 2 kinase inhibitor). A total of 20–40% of the animals treated with the antiangiogenic drugs alone survived, while all nontreated controls and tumor vaccine-treated rats died within 40 days. In vitro, these drugs inhibited endothelial cells from proliferating in response to the angiogenic factors produced by T9/9L glioma cells and prevented endothelial cell tubulogenesis. FITC-labeled tomato lectin staining demonstrated fewer and constricted blood vessels within the intracranial tumor after drug therapy. Magnetic resonance imaging demonstrated that the intracranial T9 glioma grew much slower in the presence of these antiangiogenic drugs. These drugs did not affect in vitro glioma cell growth nor T cell mitogenesis. Histological analysis revealed that the tumor destruction occurred at the margins of the tumor, where there was a heavy lymphocytic infiltrate. Real-time PCR showed more IL-2-specific mRNA was present within the gliomas in the vaccinated rats treated with the drugs. Animals that rejected the established T9/9L glioma by the combination therapy proved immune against an intracranial rechallenge by T9/9L glioma, but showed no resistance to an unrelated MADB106 breast cancer.
Glioblastoma multiforme (GBM)3 are lethal brain tumors in humans, with survival times of less than 1 year. Traditionally, debulking surgery followed by aggressive rounds of chemotherapy and radiation is used in treating patients with GBM. Even though greater than 50% of the tumor is removed or killed, no lasting and effective antitumor immunity is generated, and the patients usually die within 1 year (1) because some tumor cells escaped those therapies. These failures have led many to believe that an aggressive combination of standard therapies along with other biologically based therapies is needed to successfully treat this cancer.
The advantage of generating an immune response toward the cancer cells is that the immunized T cells can now seek out and destroy the remaining tumor cells that may have localized to sites that were inaccessible to the traditional treatments. Immunotherapy has been used to successfully treat a few glioma patients (2, 3, 4, 5, 6). These successes stimulated further interest in using immunotherapy in conjunction with other strategies to treat these tumors.
We previously observed that T9 (also known as 9L) gliosarcoma cells expressing the extracellular membrane form of M-CSF (mM-CSF), but not the secreted form of M-CSF, were killed by macrophages in vitro (7, 8). The macrophages physically attached themselves to the mM-CSF-transduced T9/9L tumor cells, delivering the necessary tumoricidal signals to the macrophage against the mM-CSF-bearing T9/9L cells. When mM-CSF-expressing T9/9L glial tumor cells were implanted into the brains of normal rats, progressive tumor growth did not occur and the surviving rats were immunized against the parental T9/9L tumor cells (9). When injected s.c., mM-CSF-transduced glioma cells were rapidly killed through a paraptosis-dependent pathway (10, 11), which stimulated long-lasting immunity and produced cross-protective immunity against other unrelated glioma cell lines (12). Rats with 3-day-old, established intracranial T9/9L gliomas survived (83% survival) when treated with a series of s.c. immunizations with mM-CSF-transduced T9/9L cells (T9-C2) (12). Therapeutic vaccination with mM-CSF-transduced tumor cells proved somewhat more effective than a IL-4-transduced tumor vaccine using 9L done under identical conditions with 3-day-old established 9L/T9 gliomas (13). Nevertheless, these types of therapeutic vaccinations were ineffective in treating 7-day-old tumors. Thus, therapeutic vaccination using cytokine-transduced tumor cells is only effective against very small tumors.
Angiogenesis controls the new blood supply routes into the tumor via the host’s endothelial cells and the supportive pericytes. Attacking the tumors’ vasculature at this level is perhaps easier than delivering drugs that kill every tumor cell directly, because the endothelial cells contact the blood-borne drugs before reaching the tumor cells. Growth factors (GF) such as platelet-derived GF (PDGF), epidermal GF, fibroblast GF (FGF), vascular endothelial GF (VEGF), IL-8, placenta GF (PLGF), and schwannoma-derived GF (SDGF) all play roles in glial tumor growth (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26). Many gliomas either make these GF as paracrine factors used to recruit the endothelial cells or overexpress GFR and use them as autocrine GF. Many gliomas have mutated receptors that are constitutively activated, and thereby elicit constant mitogenic stimuli. One class of antiangiogenic drugs is chemical inhibitors of the various intracellular GFR kinases. These drugs theoretically target the tyrosine kinase’s autocatalytic phosphorylation sites of the receptor. Antiangiogenic drugs have increased specificity for individual GFR, are generally tolerated quite well, and fail to display any toxicity even at high doses (17, 18, 21, 22, 26). These inhibitors have low molecular weights (<500 Da), and are lipid soluble, so they easily penetrate the cell, where they can inactivate the ATP transfer region of the receptor and thereby prevent phosphorylation of those tyrosine residues.
There are two major limitations of antiangiogenesis therapy. First, this fine specificity of GF-specific receptors may not be effective in clinical trials, because many advanced tumors use multiple GF-dependent pathways for angiogenesis (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28). Thus, the use of more promiscuous GFR kinase inhibitors may be better than using potent very specific inhibitors. Second, once the tumor is controlled, the hosts are not cured, because the tumor cells remain as a small avascular nest of the malignant cells. These cancerous cells then reform tumors and spread once the angiogeneic inhibitor is withheld (29). Therefore, the use of an angiogenic inhibitor along with a tumor vaccine might prove to be the key therapeutic modalities needed to cure the host of its established tumor. This assumes that the antiangiogenic drugs do not interfere with the antitumor immune responses that are being generated by the tumor vaccine.
In this study, we hypothesize that combining a tumor vaccine regimen, which is effective at controlling small tumors, with GFR-based antiangiogenic drugs, which slow the growth of large tumors at the endothelial cell level, would produce better responses than any single therapy. Vaccination with the whole glioma tumor cells genetically engineered to make mM-CSF generated an effective immune response that attacked the remaining glioma tumor cells, while the antiangiogenic drugs that target the endothelial cells prevented tumor growth by limiting blood vessel growth. Our results show that these two distinct therapies can synergize together and successfully treat larger tumors than either therapy could do alone.
Materials and Methods
Cell lines and cell culture
The rat T9/9L glioma cells were grown in RPMI 1640 medium (Sigma-Aldrich) supplemented with 5–10% FBS (Gemini Bio-Products) and 0.1% antibiotics in a humidified atmosphere of 5% CO2 at 37°C. MS1 endothelial cells were purchased from the American Type Culture Collection. These cells were cultured in DMEM medium (Sigma-Aldrich) with 5% FBS. HUVEC were obtained from P. Carpenter (University of California, Irvine, CA). All culture supplies were screened and selected on the basis of being endotoxin-free (under the limits of detection by Limulus amebocyte lysate assay). All cells were determined to be mycoplasma-free, as determined by using Stratagene’s PCR detection kit.
T9/9L glioma cells (1 × 106 cells in 10 ml) were grown in serum-free medium for 3 days. The GF were analyzed using a mouse VEGF-specific ELISA (R&D Systems) and a human PDGF BB-specific ELISA (R&D Systems).
Splenocytes from immunized rats were cultured in a mixed lymphocyte tumor reaction at a 7.5:1 (lymphocytes:T9/9L tumor cell) ratio for 48 h at 37°C. The IL-2 in the supernatant was measured using the rat IL-2-specific ELISA purchased from BioSource International.
Rat splenocytes (500,000 cells/well) responding to 2.5 μg/ml Con A (Sigma-Aldrich) or T9/9L glioma cells (25,000 cells/well) were cultured in the presence of the various concentrations of the drugs or the DMSO vehicle. To detect the various GF produced by the T9 glioma cells, a confluent plate of T9 cells (5 × 106 cells) was cultured for 1 wk in serum-free medium (Mediatech). The supernatant was sterile filtered, and a 25% solution was added to the MS1 cells (30,000 cells/well) in 96-well microtiter plates in the presence or absence of the antiangiogenic drugs.
Two days later, the growth of these proliferating cells was assessed by adding in 1 μCi/well [3H]thymidine (Amersham Biosciences) overnight. The cultures were harvested with a Ph.D. Harvestor (Cambridge Technology) and prepared for liquid scintillation counting. Data were normalized to the data generated with the DMSO vehicle alone; this was done to control for any effects that were due to the nonspecific effects of the DMSO solvent.
Matrigel was purchased from BD Biosciences. Endothelial cells (MS1 or HUVEC) were cultured on the matrigel at 5 × 105 cells/ml for 30 min. The (Z)-3-[4-(dimethylamino)benzylidenyl]indolin-2-one (DMBI) or oxindole was added and allowed to incubate for 16 h. The scoring was 0 for individual well separated cells; 1, cells aligning themselves; 2, capillary tubes forming without sprouts; 3, sprouting of new capillary tubes; 4, closed polygons forming; and 5, complex mesh structures. Five randomly picked fields were viewed and scored; these scores were then compared with those of the DMSO vehicle controls.
Female Fischer-344 rats (5–6 wk old) were obtained from Charles River Laboratories. All animals received humane care in compliance with American Association for Accreditation of Laboratory Animal Care, and the study was conducted according to protocols approved by the Veterans Affairs Medical Center-Long Beach Subcommittee of Animal Studies. This committee did permit us to perform two major surgical procedures on a limited number of animals. Routine screening in our vivarium over the last several years has been negative for various rodent-specific pathogens (viruses and bacteria).
In vivo angiogenesis/intradermal tumor growth
Rats were injected with 106 T9/9L parental cells intradermally; the animals were injected daily with either oxindole or DMBI. After 10 days of tumor growth, the tumors were measured by length, width, and height. Tumor volume was calculated by the calculation: height × width × length × π/6. The animals were euthanized, and the tumors were excised. The tumors were homogenized in 1 ml of PBS. The hemogloblin content of the homogenate was measured with the spectrophotometer at 540 nm by using Drabkin’s reagent kit (Sigma-Aldrich), according to the manufacturer’s direction.
Intracranial tumor implantation
Animals were anesthetized by an i.p. injection of ketamine (75 mg/kg) and xylazine (7.5 mg/ml). The scalp was shaved and sanitized with betadiene, and a 15-mm incision was made with a scalpel over the cranial midline. A hole in the skull was drilled at 3 mm to the right of the sagittal suture and l mm posterior to the coronal suture. T9/9L glioma cells were washed twice in endotoxin-free PBS (Invitrogen Life Technologies), and 10 μl of a 106 cell/ml cell suspension was injected at a depth of 3 mm using a Hamilton syringe over a 30-s time interval. The hole in the skull was sealed with sterile bone wax, and the incision was closed with surgical staples.
After the T9/9L tumor was allowed to establish itself for 7 days, treatment was initiated. Some rats were vaccinated s.c. with cloned mM-CSF-expressing T9/9L glioma cells (T9-C2) (total of 3 × 105 cells) in the four quadrants on days 8, 9, 10, 13, 17, 20, 23, and 27, except the first dosage was double that (6 × 105) on day 8. Some rats received the drugs (oxindole or DMBI) i.p., 1 mg/rat/day for 2 wk (days 8–22). Other rats were given both the vaccination and the antiangiogenic drugs according to the schedule listed above. Significant differences in animal survival were determined by a one-tailed Fisher’s exact test. For tissue culture studies, Student’s t tests were done. A p value <0.05 was considered significant for all studies.
Immunohistology of intracranial tumor vascularization
Before euthanasia, the rats containing the 7-day-old established T9/9L glioma (treated for 6 days with either DMBI or oxindole) were anesthetized with ketamine/xylazine, as described above. The rats were injected with 0.4 ml of FITC-labeled (2 mg/ml) tomato lectin (Lycopersicon esculentum) (Vector Laboratories) via the carotid vein. Five minutes later, the rats were killed and the brains were removed and processed as frozen thin sections. H&E staining was done of every tenth slide to identify where the tumor was. Slides that possessed the glioma were subsequently analyzed. Tissues containing the glioma were stained with a 1/50 dilution of the rabbit anti-von-Willebrand Ab (DakoCytomation) for 1 h. The tissue was washed and then incubated 1 h with a 1/200 dilution of the goat anti-rabbit secondary Ab Texas Red conjugated (Vector Laboratories). Tissue samples were washed and mounted with Prolong Gold antifade reagent (Molecular Probes). Samples were imaged and analyzed using Nikon two lasers (HeNe and Argon) PCM 2000 Confocal System on the Eclipse E800 microscope. Yellow images were the result of displaying the two channels green and red fluorescence probes of the same sample at the same time by using simultaneous multichannel confocal imaging technique. Fluorescence data were analyzed using the C-imaging software from Compix.
Magnetic resonance (MR) imaging (MRI)
A separate group of animals (n = 15) was implanted with T9/9L glioma cells, as described above. After implantation, these animals were randomly grouped into three groups: untreated controls, oxindole treatment, and DMBI treatment. At 7, 14, and 21 days after implantation, animals underwent noninvasive MRI. Animals were treated for 14 days starting on day 7. For MRI, animals were lightly anesthetized with isoflurane and placed in a plastic stereotactic device to minimize motion. MR data sets were acquired on a 4.7T Bruker (Bruker) using a quadrature radiofrequency coil (Morris Instruments). A multiecho T2-weighted sequence was used with the following parameters: repetition time (TR)/echo time (TE) = 2720/20 ms, matrix size = 256 × 256, field of view = 3 cm with 2 averages, and a total of 20 slices (1 mm thick) provided coverage of the entire brain. In addition, a contrast enhanced T1-weighted image sequence as also acquired to provide supportive information about the location and extent of the tumor. Omniscan (Gd; Amersham Health) was injected via a tail vein catheter (3 ml vol injected with a motorized syringe pump over 12 min).
Volumetric analysis of MR data was performed using a three-dimensional rendering software package, Amira (Mercury Computer System). Briefly, MR data were imported, and cerebrum and tumor volumes were outlined on all 20 slices. Amira then computes and renders the respective volumes. Volume data were exported to SigmaPlot for graphing and SigmaStat for statistical analysis (ANOVA followed by post hoc t tests; significance p < 0.05).
Weighed intracranial tumor samples (7-day-old established intracranial T9/9L tumors were treated for 7 days) were taken and homogenized in RIPA buffer (Santa Cruz Biotechnology) containing protease and phosphatase inhibitors. After the homogenization, the samples were incubated on ice for 30 min. The samples were microfuged for 30 min at 15,000 × g. The supernatants were saved, protein determinations were taken, and equal protein amounts were immunoprecipitated with the primary Ab for 1 h at 4°C. Anti-phospho-PDGFβ1 (SC12909R), anti-VEGFR2 (SC16628R), and anti-phosphotyrosine (PY-20) Abs were purchased from Santa Cruz Biotechnology. Anti-FGFR1 Ab (GR21) was bought from Oncogene Research Products. The Ab lysate mix was then mixed with protein A/G beads overnight at 4°C. The immunoprecipitate was microfuged down and then washed three times in RIPA buffer. The samples were boiled and loaded onto precast 7.5% acrylamide SDS gel (Bio-Rad) and transferred onto a nitrocellulose membrane. The individual blots were incubated using either the phospho-VEGFR2 or phospho-PDGFRβ1 Abs or the anti-FGFR1 Ab, and washed, and the secondary Ab conjugated with HRP. ECL (Amersham) was then used to detect the bound Ab-HRP conjugate. The blot was then exposed to x-ray film and the film was developed.
Abs and flow cytometry
T9/9L glioma or MS1 endothelial cells were phenotyped for their expression of either PDGFRβ or VEGFR2 using Abs directed at the extracellular domains of these rodent GFR. Anti-VEGFR2 mAb was purchased from Chemicon International; the polyclonal rabbit anti-PDGFRβ Ab was obtained from Calbiochem; and the FITC-labeled secondary Abs were obtained from Vector Laboratories. Ten thousand cells were analyzed on the Coulter XL flow cytometer.
Real-time PCR analysis
Total RNA was isolated from the intracranial tumors using TRIzol reagent (Sigma-Aldrich). Any possible DNA contamination in the sample was eliminated by 5 U/sample RNase free DNase I digestion (Boehringer Mannheim) incubating 30 min at 37°C. The DNase I was inactivated at 96°C for 15 min. cDNA was synthesized using iScript cDNA synthesis kit (Bio-Rad) containing equivalent amounts of total RNA (1 μg/sample). Real-time PCR was performed on an iCycler iQ detection system (Bio-Rad). The real-time PCR was conducted by using Brilliant SYBR Green kit (Stratagene). The thermal profile was 95°C for 15 min, followed by 40 cycles of 95°C for 15 s and 58°C for 30 s, finally holding at 4°C. The IL-2 forward primer, 5′-GCC TCC TAC TTA TAA CAC ACA-3′; reverse, 5′-CCT TGG GGC TTA CAA AAA GAA-3′. The 18S rRNA forward primer, 5′-CAG GAT TGA CAG ATT GAT AGC TCT T-3′; reverse, 5′-GAG TCT CGT TCG TTA TCG GAA TTA A-3′. Samples were run in triplicate, and a reaction without cDNA was used to establish baseline fluorescence levels either for IL-2 or the housekeeping gene. The fluorescent signal was plotted vs cycle number, and the threshold cycle (CT) was determined when the cycle number at which an increase above background fluorescence could be reliably detected. Each PCR run also included nontemplate controls containing all reagents except cDNA. These controls generated a CT greater than 40 in all experiments. After cycling, a melting curve was produced by slow denaturation of the PCR end products to validate the specificity of amplification. The quantity of mRNA was calculated by normalizing the CT level of IL-2 to the CT of the housekeeping gene 18S ribosomal RNA, according to the following formula: the mean 18S CT was subtracted from the mean IL-2 CT level. This result represents the change in CT (ΔCT). This ΔCT is specific and can be compared with the ΔCT of a calibration sample. The subtraction of the ΔCT of 18S rRNA from the ΔCT of IL-2 is referred to as ΔΔCT. The relative quantification of expression of IL-2 was determined by 2−ΔΔCT, as described by Pfaffl (32).
DMBI and oxindole only inhibit endothelial cell function
In vitro concentrations from 0.125 to 1.0 μg/ml DMBI and oxindole failed to inhibit the proliferation of rat splenocytes responding to the T cell mitogen, Con A (Fig. 1,A). Neither drug inhibited the growth of T9/9L glioma cells (Fig. 1 B).
T9/9L gliomas produce proangiogenic factors such as VEGF, PLGF, SDGF, and IL-8, also known as cytokine-induced neutrophil chemoattractants (10, 24, 25), which stimulate endothelial cell growth. DMBI and oxindole from 0.25 to 1.0 μg/ml both significantly inhibited MS1 endothelial cells (Fig. 1,C) from proliferating in vitro in response to the soluble GF released by T9/9L glioma cells by culturing these cells in serum-free medium for 7 days (183 pg/ml VEGF and 44 pg/ml PDGF). DMBI and oxindole proved more potent in inhibiting endothelial cell tube and sprout formation on matrigel than in proliferation assays. A representative micrograph of the untreated MS1 endothelial cells growing on the matrigel showing typical endothelial tube structure is shown in Fig. 2,A. Drug-treated MS1 cells showing lack of tube structure are shown in Fig. 2,B (DMBI), while oxindole-treated MS1 cells are shown in Fig. 2,C. The drugs maintained the endothelial cells as isolated single cells and prevented their reorganization into tubules, which are thought to be an in vitro correlate of capillary formation. This inhibition was not as a result of cell death, because washing out the drugs from the 1-day cultures allowed these cells to spontaneously form tubules. The dose-dependent kinetics of DMBI and oxindole on another set of the MS1 cells is shown in Fig. 2 D. Identical inhibitory results were seen using HUVEC with the same doses of DMBI and oxindole (C. Samathanam and M. Jadus, unpublished results). Thus, we concluded that endothelial cells are more sensitive to the effects of DMBI and oxindole than are the glioma and T cells.
Combined antiangiogenic therapy and a tumor vaccine significantly improve the survival of rats with established intracranial tumors
T9/9L glioma cells were surgically implanted into syngeneic rat brains and established for 7 days. Afterward, a vaccination regimen that consisted of eight injections with the mM-CSF-transduced T9/9L cells (3 × 105 cells of the T9-C2 clone/injection) given in four quadrants s.c. over a 3-wk period was used. This approach maximized peripheral lymphoid stimulation of the cervical and axillary lymph nodes. This vaccination regimen was supplemented with daily injections (1 mg/100–120 g rat i.p.) of antiangiogenic drugs for the next 2 wk. By rough calculations, we believe we can briefly achieve the drug concentrations at 1 μg/ml for a short time, which would be theoretically effective against the endothelial cells, as seen in Fig. 2,D. Fig. 3 shows the survival patterns of this experiment. All nontreated animals and rats that received the tumor vaccine alone died within 40 days. Animals treated with either DMBI or oxindole alone showed 40 and 20% 3-mo survival, respectively. When both therapies (T9-C2 + DMBI or T9-C2 + oxindole) were used together, there was a synergistic effect in which >92% of the animals survived long-term (p < 0.001). These conditions proved optimal. Reducing the number of injection sites or reducing the dose of cells or drugs was not as effective as the protocol described above.
Intraperitoneal administration of the antiangiogenic drugs specifically inhibited the receptors within the intracranial glioma
Animals undergoing combined therapy were euthanized at 13 days (7 days for establishing the tumor followed with three vaccinations and 6 days of drug therapy), and the intracranial gliomas were extracted. The GFR were analyzed with Western blotting techniques to detect phosphorylated VEGFR2 (FLK-1), PDGFRβ, and FGFR1 to determine whether the administered drugs were inactivating the GFR within the intracranial glioma. Fig. 4 showed that the GFR found within the untreated control tumors were in the activated, phosphorylated forms, as were the tumors in the rats vaccinated with the T9-C2 cells. Oxindole completely inhibited the phosphorylation of the VEGFR2, while it failed to inhibit the phosphorylation of PDGFRβ. DMBI-treated rats displayed a reduced phosphorylation signal of the PDGFRβ and FGFR1. The overall inhibition of the FGFR1 by DMBI therapy was not as strong as that seen with PDGFRβ. In one rat, DMBI therapy showed a mild reduction in the VEGFR2 signal, suggesting that the DMBI might have nonspecific effects. From these experiments, we conclude that systemic administration of the drugs is preventing proper signal transduction of these GFR within the intracranial glioma.
We tried to elucidate the source of the phosphorylated GFR found in the previous Western blots. We performed flow cytometry of the MS1 endothelial cells and T9 glioma cells using Abs directed at the extracellular domains of the PDGFRβ and the VEGFR2. Fig. 5 shows that the MS1 cells displayed both PDGFRβ (Fig. 5,A) and the VEGFR2 (Fig. 5,B). In contrast, the T9 gliomas failed to show any of these receptors above that expression of the isotypic controls (Fig. 5, C and D). Even though these flow cytometric profiles of the T9 cells are derived from tissue culture sources, we believe it is unlikely that in vivo the T9 cells spontaneously acquired these receptors. Thus, the source of the phosphorylated receptors seen in Fig. 4 is probably coming from the endothelial cells recruited into the tumor.
DMBI and oxindole slow intracranial glioma tumor growth as shown by volumetric MRI
MRI provided a noninvasive method to monitor intracranial glioma tumor growth for 21 days after implantation. The cerebrum volumes (excluding the cerebellum) were not significantly different between animals in the three groups over the 21-day imaging period (Fig. 6). There was a 2.6% increase in cerebrum volume (7 day, 1.09 ± 0.041; 21 day, 1.12 ± 0.02 cm3, mean ± SEM). In contrast, there was a highly significant increase in tumor volume with time in control animals (p < 0.001). Tumor volume at 7 days was 0.002 ± 0.0006, which increased to 0.027 ± 0.017 at 14 days, followed by a dramatic increase to 0.177 ± 0.055 cm3 at 21 days. In contrast, tumor growth was not significant in oxindole- or DMBI-treated animals over the 21-day time course. ANOVA analysis indicated a significant difference in tumor volumes between control and oxindole- and DMBI-treated animals at 21 days (p < 0.05). As can be seen in Fig. 6 D, post hoc analysis demonstrated that the oxindole-treated animals’ tumor volume was significantly decreased compared with control (p < 0.04), whereas the tumor volume in DMBI-treated animals was not significant. Thus, in vivo treatment of intracranial tumors with oxindole, and less so, DMBI, significantly retards tumor growth.
Immunohistology reveals that DMBI and oxindole treatment results in constricted blood vessels within the intracranial tumors
The MRI data showed intracranial tumor growth was being inhibited by the DMBI and oxindole. Studies were done to assess the changes within the vasculature of the 7-day-old established intracranial T9/9L tumors (treated for 6 days) induced by the DMBI and oxindole. FITC-labeled tomato lectin was injected i.v. into the rats 5 min before euthanasia. The circulating lectin binds to the endothelial cells within the lumen of the blood vessel (33). Fig. 7,A shows that there was less fluorescent signal within the rats treated with either the DMBI or oxindole. The top row shows comparable low ×10 magnifications of the active tumor, while the bottom row is photographed at ×40. The number of blood vessels looked slightly less than the control. The blood vessels of the DMBI-treated rats also appeared remarkably constricted and were shorter in length when compared with the nontreated controls. We quantitated the amount of lectin fluorescence by two methods. Both total FITC fluorescence intensity and total FITC fluorescence area were measured (Fig. 7 B). Using both methods, DMBI and oxindole significantly showed less intensity. No such constrictions were seen within the blood vessels in the brain tissue that did not have the tumor. Higher magnification of the oxindole-treated rats showed a disoriented and disrupted pattern of the blood vessels, especially with the oxindole treatment.
To correlate the relationship of the green FITC lectin in the blood lumen and endothelial cells, we stained the tissue with an endothelial cell-specific Ab to von Willebrand factor and visualized with a Texas Red-conjugated secondary Ab. Fig. 8, A–C, shows the images of the green fluorescence channel for the FITC lectin probe of either untreated, DMBI-treated, or oxindole-treated rats, respectively. Fig. 8, D–F, shows the red fluorescence channel images of the labeled endothelial cells from the von Willebrand Ab of the same tissue. Fig. 8, G–I, is the result of the simultaneous representation of the green channel colocalized with red channel.
To verify that fewer blood vessels were forming within the T9 tumor as a result of the drug treatment, we established intradermal T9 tumors and treated the rats with the drugs in an in vivo angiogenesis assay. After 10 days of treatment with the drugs, the intradermal tumors were then excised and the hemoglobin content was determined. Fig. 9,A shows the size of the intradermal T9 cells, while Fig. 9 B shows the hemoglobin content was significantly lower in those rats that were treated with either DMBI or oxindole. From this experiment, we would conclude that DMBI was slightly more antiangiogenic than oxindole within the intradermal site.
Histology reveals that more lymphocytes are entering the tumor, and antitumor effects are located at the periphery of the growing tumor
The experimental procedures were repeated (7-day-old established intracranial tumor, followed by 6 days of drug therapy and T9-C2 vaccination), and the brains were removed. Paraffin-embedded thin sections of the brain were stained with H&E. Fig. 10,A shows the growth of the T9/9L cells at the expanding edge of the tumor. The tumor is growing as a healthy palisading tumor with a minimal lymphocytic infiltration. Fig. 10,B shows the tumor derived from a rat treated with both the mM-CSF tumor vaccine with oxindole treatment. The tumor cells are dying, and numerous lymphocytes are invading the site. In the center, there is a nest of viable tumor cells. At the right side of this micrograph are numerous RBCs, suggesting that the vasculature in this vicinity recently ruptured. In the rats treated with the vaccine and DMBI we saw a similar effect (Fig. 10 C). With both drugs, some healthy tumor is still present, suggesting that this therapeutic response is actively working after 6 days of drug therapy. In general, the antitumor effects of the tumor vaccine combined with either oxindole or DMBI appear to occur at the margins of the growing tumor, which is presumably at the site of new blood vessel formation.
Splenocytes derived from rats immunized with the mM-CSF T9/9L glioma and the antiangiogenic drugs produce IL-2 upon stimulation in a mixed lymphocyte tumor reaction
The same rats used in the Western blots (Fig. 4), along with another set of rats, had their spleens removed, and the splenocytes were used in a mixed lymphocyte T9/9L tumor reaction to determine the levels of IL-2 produced by the immunized lymphocytes. The IL-2 as detected by ELISA is displayed in Fig. 11. The splenocytes from naive rats and those rats bearing the intracranial T9/9L glioma without any treatment (n = 4) made small amounts of IL-2 (20–113 pg/ml). The splenocytes from the T9-C2-immunized rats (n = 5) made 485 pg/ml IL-2. The splenocytes from oxindole + T9-C2-treated rats (n = 5) made more IL-2 (818 pg/ml) than from those rats (n = 5) treated with DMBI + T9-C2 (464 pg/ml) or T9-C2-vaccinated rats (n = 5), although this trend was not statistically significant. We concluded that neither DMBI nor oxindole adversely affected the immune system from responding to the mM-CSF tumor vaccine, which would be a prerequisite for obtaining synergistic effects.
The lymphocytes within the intracranial tumor are making more IL-2 mRNA as a result of combined therapy
The results of Fig. 11 suggested that systemic immunity stimulated by the mM-CSF tumor vaccine was slightly enhanced by the oxindole and DMBI. But to determine whether there was increased local intracranial immunity, we used real-time PCR to quantitate the amount of IL-2-specific mRNA within the brain tumor. By the ΔΔCT method, the intracranial lymphocytes from rats vaccinated cotreated with the T9-C2 cells with either oxindole or DMBI produced 4.9 times and 5.7 times, respectively, as much IL-2 mRNA as the lymphocytes produced by the rats that only received the T9-C2 tumor vaccine. Thus, there appears to be increased immunological activity within the local tumor site from the rats supplemented with the antiangiogenic drugs. These data are therefore consistent with the histology, suggesting more activated lymphocytes are now present within the intracranial glioma, as a result of the treatment with the antiangiogenic drugs.
Animals that survived the initial T9/9L glioma cells were specifically immune toward T9 glioma cells
When the surviving animals from Fig. 3 (>3 mo after the initial T9 challenge) from the initial therapy (T9-C2 + DMBI and T9-C2 + oxindole) were rechallenged in the brain with T9/9L glioma cells, two-thirds of the animals survived (Fig. 12,A). In a repeated experiment using a different set of rats (>3 mo after the initial T9/9L challenge), these rats all succumbed to an unrelated syngeneic MADB106 breast cancer (Fig. 12 B). Therefore, as a result of this combined therapy, the rats were specifically immunized against the T9/9L glioma.
GBM remain lethal brain cancers that are refractory to most standard oncological treatments. Current thinking is that more aggressive therapies combined with a rational, biological-based therapy may significantly improve survival. Immunotherapy has shown some occasional successes against human glioma (2, 3, 4, 5, 6); it has stimulated further research into ways to improve its efficacy.
We showed earlier that mM-CSF-transduced glioma cells (rat T9 and human U251) within 4 h of being injected s.c. into rodents are spontaneously killed through a swelling process called paraptosis (10, 11). During this tumoricidal process, strong inflammatory reactions were seen at the injection site. These responses included: polymorphonuclear cell and monocyte/macrophage accumulations, evidence of inducible NO synthase and peroxynitrite formation, and elevated levels of the heat shock proteins 60 and 70 and gp96. As a result of this inflammation, strong tumor-specific immunity in both CD4+ and CD8+ subsets occurred using a number of tumors, including gliomas (9, 10, 12), hepatomas (34), and breast cancers (35). In the rat T9/9L glioma model, this vaccination lead to rejection of small 1- to 3-day-old established intracranial T9/9L glioma (12). We reasoned that to achieve therapeutic results that are relevant to patients diagnosed with GBM, we must combine this tumor vaccine with another complementary therapy that slows tumor growth. By slowing tumor growth, this allows more time for the activated lymphocytes to enter the tumor and begin killing the malignant cells.
Antiangiogenic therapeutic approaches show promise in treating some human cancers by attacking the cancer at the level of the tumor vasculature (36). One limitation of this method is that once tumor regressions occur, the drug must be constantly administered to keep the residual tumor in a small avascular state. Usually tumor immunity does not occur, because the tumor cells are dying of apoptosis, also called the silent death. Apoptotic cells can be used to vaccinate animals when fed directly to dendritic cells in vitro before using these cells to vaccinate the hosts (38). But apoptotic cells are usually not the optimal way of stimulating an immune response (10, 37, 38, 39, 40, 41, 42, 43, 44), when compared with necrotic cells producing the danger signals.
We thought that by using a combination therapy of immunotherapy with an antiangiogenic approach, we could treat larger tumors than either therapy when used alone. In this study, we successfully combined a T9/9L tumor vaccine expressing mM-CSF with antiangiogenic drugs (DMBI (PDGFRβ and a FGFR1 kinase inhibitor) and oxindole (VEGFR2, flk-1, kinase inhibitor)) to treat 7-day-old intracranial T9/9L gliomas (Fig. 3). The antiangiogenic drugs by themselves allowed 20–40% of the animals to survive past 4 mo. In vitro, these drugs inhibited the endothelial cell formation of tubes and sprouts (Fig. 2). T9/9L gliomas are known to produce a number of GF that can stimulate endothelial cell growth; i.e., SDGF, PLGF, IL-8, VEGF, and PDGF (the last two GFs are reported in this work). These two drugs also inhibited endothelial cells from proliferating in response to the angiogenic factors produced by T9 glioma cells (Fig. 1 C), suggesting that the drugs may have some common inhibiting pathway at the level of signal transduction or downstream sites. Therefore, these drugs may be more broadly reactive than initially reported (30, 31), because they inhibit endothelial cell function from responding to a number of proangiogenic factors. DMBI and oxindole had no direct effect upon glioma cell growth, suggesting that the drugs were more potent in inhibiting endothelial cell function when compared with the T9 glioma.
By using a FITC-labeled tomato lectin, which stains the lumens of blood vessels, we saw fewer blood vessels, and these blood vessels appeared constricted (Figs. 7 and 8). This last finding suggests that DMBI and oxindole were restricting the flow of blood into the growing tumor. By noninvasive MRI (Fig. 6), we saw that the growth of the intracranial T9 glioma was inhibited by these drugs. Additionally, using an in vivo angiogenesis assay (Fig. 9), the antiangiogenic drugs inhibited intradermal tumor growth as well as reducing the amount of RBC found within these tumors. Although DMBI looked like it had better antiangiogenic effects, by inhibiting hemoglobin content within intradermal sites, both drugs had relatively similar effects in the final survival of rats with intracranial tumors when combined with the tumor vaccine.
Oxindole and DMBI (from 1 to 0.25 μg/ml) did not inhibit in vitro T cell mitogenesis in response to Con A, suggesting that the endothelial cells were more sensitive to their effects than were the T cells or T9/9L tumors. Animals that rejected the established T9/9L glioma using the combination therapy proved immune against an intracranial rechallenge by T9/9L glioma (Fig. 12,A), but were not immune to an unrelated MADB106 breast cancer (Fig. 12,B). This suggests that stimulation of the immune response was not compromised by either DMBI or oxindole. Upon in vitro restimulation, the lymphocytes primed in vivo made comparable levels of IL-2, although the IL-2 made by the lymphocytes from the vaccinated/oxindole-treated rats was slightly greater than those splenocytes obtained from vaccinated/DMBI-treated rats. VEGF does interfere with dendritic cell functions (45, 46), so inhibiting VEGF signaling on the dendritic cells could result in better APC function that enhances T cell responses. By real-time PCR, we found more IL-2-specific mRNA within the intracranial tumors of those vaccinated rats that received the antiangiogenic drugs than those rats not given the drugs. This confirms that more activated lymphocytes are entering into the tumor, as we saw in Fig. 10.
Another key feature for making these synergistic responses possible is that the antiangiogenic drug must be compatible with the immune responses. Besides being an antiangiogenic drug for myeloma (47) and glioma (17), thalidomide also suppresses T cell-mediated graft-vs-host disease in bone marrow-transplanted patients (48). In preliminary studies, we found that thalidomide failed to enhance survival of the rats with intracranial gliomas (negative data not shown). Thus, antiangiogenic drugs such as thalidomide may not be the optimal antiangiogenic drug to be used with a tumor vaccine, due to its immunosuppressive nature.
The histology that we have observed with combined therapies seems to occur at the margins of the growing tumor, which probably is the site of newly forming blood vessels of the advancing tumor. This would tend to confirm our in vitro studies (Figs. 1 and 2), in which the drugs inhibited endothelial cell growth and formation of tubes and sprouts. We speculate that the antiangiogenic drugs disrupt the blood supply by inhibiting endothelial cells at the GFR level (Fig. 4) that support the formation of new tumor vasculature. Many tumor-derived GF stimulate the production of several cell adhesion molecules (49, 50, 51, 52, 53). By inhibiting these GF signal transduction pathways, focal adhesion kinase is also inhibited. This disrupts cellular adhesion to the basement membrane by these endothelial cells, making these cells less adherent, thereby increasing vascular leakage. Contrast-enhanced MRI supports this concept (data not shown). This increased permeability then allows more lymphocytes to infiltrate the tumor (Fig. 10, B and C) via diapedesis through the less adherent endothelial cells. We observed that more IL-2-specific mRNA (4.9–5.7 times more) was produced within the intracranial tumors that received both therapies, as opposed to a single treatment, suggesting better immunological responses occurred within the intracranial tumor. By our proposed mechanism, the tumor grows at a slower rate, allowing more time for the lymphocytes to attack the tumor.
Immunotherapy has recently been combined with antiangiogenic therapies to treat established s.c. tumors. Sun et al. (54) used plasmids containing the angiostatin and B7.1 genes to obtain some regressions of small s.c. EL4 tumors. Cuadros et al. (55) used soluble VEGFR1 receptors combined with dendritic cells pulsed with killed tumor cells supplemented with IL-2 and anti-Ox40 Abs to obtain regressions of 7-day-old her2/neu+ breast cancers. However, these therapies did not completely eliminate the tumor, and they failed to show that any lasting immunity was produced as a result of their combined therapy. Our work demonstrates that lasting immunity does occur after the intracranial tumor growth is eliminated. This lasting immunity is critical for eliminating minimal residual tumor that has not been eliminated by other therapies, and it may prove to be beneficial in any complete tumor regression.
In summary, our current study supports the hypothesis that it is possible to combine immunotherapy with an antiangiogenic approach via a systemic vaccination route to treat rats with large well-formed intracranial gliomas. These two therapies proved complementary to each other and allowed a long-lasting immune response to occur. This combination therapeutic approach may lead toward better therapies of patients with GBM than treatment with any single agent.
The authors have no financial conflict of interest.
We thank Dr. Ronald C. Kim for his work in cutting the rat brains, and Dr. Bita Behjatatnia for photographing the histological sections. We also thank Michael Robbins for his assistance with the MRI volumetric analyses.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This study was funded in part from grants obtained from Veterans Affairs Medical Center (to M.R.J. and H.T.W.), National Institutes of Health (Grant CA77802 R01), and Avon Breast Cancer Foundation via University of California (Irvine, CA) Cancer Research Program and Chiron (Emeryville, CA). The Neuro-Imaging was supported by a National Aeronautics and Space Administration Cooperative Agreement to Loma Linda University (NCC9-149).
Abbreviations used in this paper: GBM, glioblastoma multiforme; CT, threshold cycle; ΔCT, change in CT; DMBI, (Z)-3-[4-(dimethylamino)benzylidenyl]indolin-2-one; GF, growth factor; FGF, fibroblast GF; mM-CSF, membrane form of M-CSF; MR, magnetic resonance; MRI, MR imaging; PDGF, platelet-derived GF; PLGF, placenta GF; SDGF, schwannoma-derived GF; VEGF, vascular endothelial GF.