Differential effects of nitric oxide on blood–brain barrier integrity and cerebral blood flow in intracerebral C6 gliomas

Neuro-Oncology, Feb 2011

Nitric oxide (NO) signaling in tumors and endothelial cells regulates vascular permeability and blood flow and therefore influences tumor uptake and response to therapeutic compounds. As delivery and efficacy of chemotherapy is impaired in CNS neoplasms due to a partially intact blood–brain barrier (BBB), we studied the effects of NO released by the short-acting NO donor disodium 1-[2-(carboxylato)pyrrolidin-1-yl]diazen-1-ium-1,2-diolate methanolate (PROLI/NO) on BBB integrity and blood flow in C6 gliomas using [14C]-aminoisobutyric acid (AIB) and [14C]-iodoantipyrine quantitative autoradiography. PROLI/NO selectively increased intratumoral uptake of [14C]AIB and [14C]sucrose when given as a 3-minute intracarotid infusion or a 15-minute i.v. infusion (AIB: tumor, K1 = 68.7 ± 3.2 vs 24.9 ± 0.9 µL g−1 min−1, P < .0001; sucrose, K1 = 16.9 ± 0.9 vs 11.5 ± 0.9 µL g−1 min−1, P = .0007). This effect was achieved without significant changes in cerebral and tumor blood flow or arterial blood pressure, which indicates that the effect on vascular permeability is independent of changes in vascular tone induced by NO. This effect was mediated by activation of the NO/3',5'-cyclic guanosine monophosphate (cGMP) pathway, as it was blocked by guanylate cyclase inhibition by LY83583 and reproduced by the delivery of 8-bromoguanosine 5'-monophosphate or inhibition of cGMP degradation by the phosphodiesterase inhibitor zaprinast. Inhibition of inducible NO synthase by aminoguanidine or cyclooxygenase inhibition by indometacin or dexamethasone did not reduce the blood–tumor barrier (BTB) response to PROLI/NO. PROLI/NO, and perhaps other NO-donating compounds, can be used to selectively increase BTB permeability in gliomas through the NO/cGMP pathway at doses that do not cause unwanted vasodilatory changes in blood flow and that do not affect the systemic circulation.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://neuro-oncology.oxfordjournals.org/content/13/2/203.full.pdf

Differential effects of nitric oxide on blood–brain barrier integrity and cerebral blood flow in intracerebral C6 gliomas

Astrid Weyerbrock () 0 1 2 3 4 Stuart Walbridge 0 1 2 3 4 Joseph E. Saavedra 0 1 2 3 4 Larry K. Keefer 0 1 2 3 4 Edward H. Oldfield 0 1 2 3 4 0 Bethesda , Maryland (A.W., S.W. , E.H.O.); Laboratory of Comparative Carcinogenesis , National Cancer 1 Institutes of Health, National Institute of Neurological Disorders and Stroke , Surgical Neurology Branch 2 Department of Neurosurgery, University Medical Center Freiburg , Freiburg, Germany (A.W.); National 3 Frederick, Maryland (J.E.S.); Department of Neurosurgery, University of Virginia , Charlottesville, Virginia 4 Institute at Frederick , Frederick , Maryland (L.K.K.); Intramural Research Support Program , SAIC Frederick - Nitric oxide (NO) signaling in tumors and endothelial cells regulates vascular permeability and blood flow and therefore influences tumor uptake and response to therapeutic compounds. As delivery and efficacy of chemotherapy is impaired in CNS neoplasms due to a partially intact blood brain barrier (BBB), we studied the effects of NO released by the short-acting NO donor disodium 1-[2-(carboxylato)pyrrolidin-1-yl]diazen-1-ium1,2-diolate methanolate (PROLI/NO) on BBB integrity and blood flow in C6 gliomas using [14C]-aminoisobutyric acid (AIB) and [14C]-iodoantipyrine quantitative autoradiography. PROLI/NO selectively increased intratumoral uptake of [14C]AIB and [14C]sucrose when given as a 3-minute intracarotid infusion or a 15-minute i.v. infusion (AIB: tumor, K1 5 68.7 + 3.2 vs 24.9 + 0.9 mL g21 min21, P < .0001; sucrose, K1 5 16.9 + 0.9 vs 11.5 + 0.9 mL g21 min21, P 5 .0007). This effect was achieved without significant changes in cerebral and tumor blood flow or arterial blood pressure, which indicates that the effect on vascular permeability is independent of changes in vascular tone induced by NO. This effect was mediated by activation of the NO/3,5-cyclic guanosine monophosphate (cGMP) pathway, as it was blocked by guanylate cyclase inhibition by LY83583 and reproduced by the delivery of 8-bromoguanosine 5-monophosphate or inhibition of cGMP degradation by the phosphodiesterase inhibitor zaprinast. Inhibition of inducible NO Received February 9, 2010; accepted September 15, 2010. synthase by aminoguanidine or cyclooxygenase inhibition by indometacin or dexamethasone did not reduce the blood tumor barrier (BTB) response to PROLI/NO. PROLI/NO, and perhaps other NOdonating compounds, can be used to selectively increase BTB permeability in gliomas through the NO/cGMP pathway at doses that do not cause unwanted vasodilatory changes in blood flow and that do not affect the systemic circulation. Nbiology via effects on tumor growth, migration, itric oxide (NO) plays a crucial role in cancer invasion, angiogenesis, tumor blood flow, and vascular permeability.1 3 NO is synthesized by NO synthases (NOS) that are ubiquitously expressed in malignant tumors, including gliomas.4 The NO effects in tumors depend on the activity and localization of NOS, concentration and duration of NO exposure, and cellular sensitivity to NO. NO promotes tumor angiogenesis, maintains tumor blood flow, and increases vascular permeability in tumors, effects that can be blocked using NOS inhibitors or NO scavengers.5 8 Enhanced NOS expression, predominantly of inducible NOS (iNOS), but also of endothelial NOS (eNOS) and neuronal NOS (nNOS), was found in CNS tumors and in tumor endothelial cells and appears to correlate with the degree of malignancy.4 NO is produced by various sources in tumors, including stromal cells and endothelial cells, and triggers multiple signaling pathways in gliomas. Therapeutic strategies to modulate NO signaling might be promising adjuncts to glioma treatment. Some strategies focus on reducing NO signaling to shut down tumor angiogenesis, tumor blood flow, vascular permeability, and tumor growth. Other approaches aim at increasing NO signaling to alter tumor physiology to increase drug delivery into tumors, to induce NO-mediated tumor cell killing, or to induce chemo- or radiosensitization.9,10 Bradykinin and its analogs activate the NO/ 3,5-cyclic guanosine monophosphate (cGMP) signaling pathway via the bradykinin-2 (B2) receptor, leading to increases in blood brain barrier (BBB) permeability and to increased delivery of chemotherapeutic agents into gliomas in rodents.11,12 However, no increased efficacy of carboplatin in combination with the bradykinin analog RMP-7 was observed in a phase II trial of patients with recurrent malignant gliomas.13 Reasons for the lack of efficacy may have been insufficient delivery of carboplatin into the gliomas due to the inhomogeneous expression of B2 receptors in malignant gliomas, desensitization of the receptors, or low dosing of the BBB permeabilizer. To circumvent these delivery issues, exogenous NO can be administered using NO-releasing substances of the diazeniumdiolate group. These substances generate bioactive NO in physiological fluids spontaneously in a controlled fashion with reliable half-lives ranging from 2 s to 20 h.14 Short-acting NO donors, such as disodium 1-[2-(carboxylato)pyrrolidin-1-yl]diazen1-ium-1,2-diolate methanolate (PROLI/NO), are highly vasoactive, whereas NO donors with a longer half-life can be activated for NO release enzymatically for targeting NO to specific cells at high doses to induce growth inhibition and apoptosis.15 PROLI/NO has been investigated as a vasodilator for cerebral vasospasm in subarachnoid hemorrhage, for cerebral ischemia, for pulmonary hypertension, and as an inhibitor of endothelial hyperplasia.16 19 Our group demonstrated tumor-selective opening of the BBB by PROLI/NO and its metabolite sodium nitrite and increased efficacy of carboplatin chemotherapy in the C6 rat glioma model.20 As NO triggers multiple pathways through cGMP and/or through S-nitrosylation of proteins, it is not clear how the effect of exogenous NO released by PROLI/NO on the tumor and tumor vasculature is mediated. Although peripheral microvascular permeability is influenced by a variety of cellular processes, including increase in intracellular calcium concentration, activation of adenylate cyclase, guanylate cyclase, or release of classical inflammatory mediators such as cytokines, excitatory amino acids, or metabolites of arachidonic acid, the potential role of these pathways underlying changes in permeability of the BBB in response to exogenous NO remains uncertain.21 Furthermore, it is yet unknown if the increased uptake of radiolabeled tracers and chemotherapeutic drugs into a tumor is solely due to opening of the blood tumor barrier (BTB) or if PROLI/NO also affects tumor blood flow. The experiments described here elucidate the mechanism of the effect of PROLI/NO on the BBB by blocking various second messenger pathways, assess the effect of PROLI/NO on regional cerebral blood flow (CBF), and establish the appropriate and safe infusion parameters for this approach. Materials and Methods Materials The diazeniumdiolate NO donor PROLI/NO (MW 251) was isolated by exposing methanolic solutions of L-proline and sodium methoxide to gaseous NO as described by Saavedra et al.19 The extent of NO release in pH 7.4 buffer at 378C is 2 mol/mol of PROLI/NO with a half-life of 1.8 seconds. A 10-mM stock solution of PROLI/NO was prepared from the powdered form by dissolving it in ice-cold 0.1-M sodium hydroxide (pH 10.5) and storing it at 2208C. The stock solution showed no evidence of decomposition over an observation period of .3 weeks using this formula and the appropriate storage conditions. Dilution with 0.01 N NaOH in normal saline to the desired concentration (1026 M) was performed immediately prior to use, and the solution was kept on wet ice during the experiment. To confirm stability, a 6.8-mM solution of PROLI/NO in 0.01 N NaOH in normal saline was stored at 48C for 2 weeks. The decay was monitored by UV at 250 nm, and the original optical density (OD) at t 0 was measured at 0.52 (1 7.76 mM21 cm21). After the solution was stored at 48C for 1 week, the OD decreased to 0.48 (29%), and after 2 weeks, it was measured at 0.45 (215%). PROLI/NO was infused over 3 or 15 minutes in 0.1 M sodium hydroxide as the vehicle, and no toxic side effects were reported in our study and previously.19 Dexamethasone (methylprednisolone, MW 392.5), a synthetic glucocorticoid that acts as an anti-inflammatory inhibitor of iNOS and cyclooxygenase (COX), was purchased from Fujiwara Inc., and LY-83583 (6-phenylamino)-5,8 quinolinedione (MW 250.3), a soluble guanylate cyclase (sGC) inhibitor, was purchased from ICN Biomedicals. Zaprinast (1,4-dihydro-5-[2propoxyphenyl]-7H-1,2,3-triazolo[4,5-d]pyrimidine-7one; MW 271.3), a cyclic GMP-specific phosphodiesterase inhibitor, aminoguanidine (aminoguanidine hemisulfate; MW 123.1), a selective iNOS inhibitor, and the phosphodiesterase-resistant cGMP analog 8-bromoGMP (8-bromoguanosine 5-monophosphate; MW 442.1) were purchased from Sigma Chemicals. Indometacin (Indocin i.v., indometacin sodium trihydrate, MW 433.8), a nonselective inhibitor of COX-1 and -2, enzymes that participate in prostaglandin synthesis from arachidonic acid, was purchased from Merck & Co. The [14C]-radiolabeled tracers aminoisobutyric acid ([14C]AIB, MW 103.1), [14C] sucrose (MW 372), and [14C]-iodoantipyrine ([14C]IAP; MW 314.1) were produced by American Radiolabeled Chemicals. Tumor Induction and Animal Preparation The study was conducted in accordance with the National Institutes of Health (NIH) guidelines on the use of animals in research and was approved by the Animal Care and Use Committee of the National Institute of Neurological Disorders and Stroke. C6 cells obtained from the American Type Culture Collection were cultivated in Dulbeccos modified Eagles medium supplemented with 10% fetal calf serum, glutamine, penicillin, and streptomycin at 378C and 5% CO2. One hundred and five male Sprague Dawley rats weighing 300 350 g (n 5 per group) were used to evaluate BBB permeability and regional CBF (rCBF) in various regions of interest (ROIs) in the brain and in the tumors. Intracerebral C6 gliomas were induced in a standard fashion by stereotactic inoculaton of 105 C6 cells into the right caudate nucleus as described elsewhere.20 BBB Permeability and Blood Flow Studies Experimental procedureTen days after stereotactic tumor inoculation, the rats were anesthetized with isoflurane, and both femoral arteries and 1 femoral vein were cannulated. For intra-arterial delivery, a PE-50 catheter was inserted retrogradely into the right external carotid artery with the catheter tip at the common carotid artery bifurcation. Body temperature of 378C was maintained using a heating blanket, and mean arterial blood pressure (MABP) and pulse were monitored continuously via the femoral artery. The MABP at the beginning and the end of the infusion period and after the 15-minute experimental procedure was analyzed. Quantitative autoradiography (QAR) for the assessment of vascular permeability was performed by determining the blood-to-tissue transfer constant K1 as described earlier.22 The permeability tracers [14C]AIB (MW 100) and [14C]sucrose (MW 372) were given as an i.v. bolus (80 mCi) during or at the end of the infusion of the test substances. Arterial blood samples were taken at selected intervals, and plasma radioactivity was determined by liquid scintillation counting with appropriately quenched [14C] standards. After 15 minutes, the rats were decapitated, and the brains were rapidly removed and frozen in isopentane on dry ice. CBF-related tissue radioactivity was measured by the [14C]IAP technique.23,24 100 mCi [14C]IAP were infused i.v. over 35 seconds at incremental volumes to produce a rising arterial concentration of tracer to prevent the equilibration of rapidly perfused tissues with the arterial blood during the period of measurement. Timed samples of arterial blood were collected from a freeflowing femoral artery catheter every 5 seconds. [14C]IAP concentrations were determined in 20 mL of plasma from each sample. One arterial blood sample was taken 5 seconds after sacrifice to correct for dead space within the catheter and transit time in the catheter. The brains were removed within 1 minute after sacrifice and snap-frozen. rCBF F was calculated using an equation developed by Kety, which includes the tissue concentration of [14C]IAP at a given time, the concentration of tracer in arterial blood at a given time, the rate of blood flow per unit mass of tissue, and a tissue-blood partition coefficient as described by Sakurada et al.24 The tissue-blood partition coefficient was determined to be 0.8 for the rat. rCBF (F ) is expressed as mL 100 g21 min21. Tissue sections were serially sliced in 20 mm sections throughout the whole tumor area on a cryostat for histology and QAR. Brain sections were exposed on film for 10 days for the permeability studies and for 5 days for the blood flow study along with tissuecalibrated [14C]methyl methacrylate standards. Corresponding histological sections were stained with hematoxylin and eosin (H&E) for identification of corresponding anatomical structures. The average diameter and the bidimensional tumor area (mm2) were measured for each section. The tissue radioactivity concentration (nCi/g) was used to express the blood-to-tissue transfer constant K1 (mL g21 min21) assuming a unidirectional transfer of the [14C]labeled tracers. K1 was calculated by dividing the tissue radioactivity concentration at the end of the experiment by the integral of the plasma radioactivity concentration from 0 to 15 minutes. Vascular permeability and rCBF were determined in 8 ROIs: tumor center (T), whole tumor, brain adjacent to tumor (BAT), ipsilateral cortex (C), ipsilateral white matter (WM), contralateral tumor area, contralateral cortex, and contralateral white matter. Representative data of the 4 most relevant regionsT, BAT, C, and WMare summarized here. BAT is defined as a 0.5-mm peritumor zone around the blue H&E-stained tumor in rodents.25 Autoradiographs and the corresponding histological sections were digitized and analyzed using NIH imaging software. Study Groups PROLI/NO was studied at a concentration of 1026 M because this dose has been shown to disrupt the BTB to [14C]-labeled tracers with a molecular weight between 100 and 70 000 without systemic side effects.20 Different routes and durations of delivery (intracarotid [ICA] vs i.v.; 3 vs 15 minutes) were assessed using [14C]AIB and [14C]sucrose as tracers. The permeabilizing effect of PROLI/NO was studied in comparison with saline (NaCl 0.9%) controls. The effect on BBB permeability to [14C]AIB of certain compounds directly or in combination with PROLI/NO was studied using the following: (i) the cGMP-specific phosphodiesterase inhibitor zaprinast (20 mg/kg, ICA, 15 minutes), which blocks degradation of cGMP by phosphodiesterase leading to prolonged increased cGMP levels; (ii) the sGC inhibitor LY83583 (1 mg/kg, i.v., 15 minutes), which prevents binding of NO to sGC, preventing the intracellular conversion of guanosine triphosphate to cGMP, thereby blocking cGMP formation; (iii) the potent inhibitor of iNOS aminoguanidine (200 mg/kg, ICA, 3 minutes), preventing the formation of endogenous NO from L-arginine; (iv) the nonselective inhibitor of COX-1 and -2 indometacin (7 mg/kg, i.v., 3 minutes), blocking prostaglandin formation which might play a role in enhanced tumor permeability26; (v) the antiinflammatory glucocorticoid dexamethasone (10 mg/ kg, ICA, 3 minutes), which is the standard drug to counteract edema in brain tumors and inhibits induction of various inflammatory genes including iNOS and COX-227; and (vi) the cell-permeable cGMP analog 8-bromo-GMP (1023 M, i.v., 3 minutes), which is mostly resistant to degradation by phosphodiesterases and preferentially activates cGMP-dependent protein kinases, such as Ca2+-ATPase, which is involved in vasorelaxation. It has been shown to increase microvascular permeability in the cerebral cortex.28 Statistical Analysis Measurements in the defined ROI were performed on 5 representative sections from different tumor areas of each animal. Measurements per section were done in triplicates. Results are expressed as the mean + SEM of all these data. Analysis of the permeability and blood flow data was performed by analysis of variance and post hoc tests. Probability (P) values of ,.05 were considered significant. MABP at various time points (0 minute, end of infusion, and end of experiment prior to sacrifice) were assessed by the paired t-test. BBB Permeability Studies BBB permeability and blood flow studies were performed on day 10 when tumors had a mean diameter of 20 + 0.85 mm2 (median 17.4 mm2); 1026 M PROLI/NO increased the [14C]AIB and [14C]sucrose uptakes by the tumor when given as a 3-minute ICA infusion or a 15-minute i.v. infusion (P , .0001, Figs 1A and B and 2A and B). After a 3-minute ICA infusion of 1026 M PROLI/NO, the influx constant K1 in tumor was 68.7 + 3.2 mL g21 min21 for [14C]AIB (saline: 24.9 + 0.9; P , .0001) and 16.9 + 0.9 mL g21 min21 for [14C]sucrose (saline: 11.5 + 1; P .0007). ICA PROLI/ NO administered over 3 minutes was more effective than i.v. PROLI/NO over 15 minutes (P , .0001). Increased uptake of [14C]AIB and [14C]sucrose was also observed to a lesser degree in the brain around the tumor (BAT, P , .0001). There was also a moderate increase in [14C]sucrose uptake in cortex and white matter after ICA PROLI/NO (P .003). To elucidate the mechanism underlying the permeability increase at the BTB observed after PROLI/ NO administration, the autoradiography studies with [14C]AIB were replicated after pretreatment with compounds affecting various intracellular signaling pathways involved in mediating or modifying NO effects in cells. Blockade of degradation of cGMP by zaprinast induced a significant increase in BTB permeability to [14C]AIB (P , .0001; Fig. 3). Combined infusion of PROLI/NO and zaprinast did not result in an additional increase in tracer uptake, but caused severe prolonged hypotension (Fig. 4). In addition, 2 rats suffered intratumoral hemorrhage. Inhibition of the NO/cGMP pathway by pretreatment with LY83583 significantly blocked increased tracer uptake into the tumor after PROLI/NO to levels even lower than the saline control (P , .0001). 8-Bromo-GMP induced a tumorselective barrier disruption similar to PROLI/NO. Blocking iNOS with aminoguanidine or dexamethasone did not alter increased tracer uptake, indicating that iNOS activation is not required when NO is Fig. 3. K1 values for [14C]AIB (mL g21 min21) in tumor (T), brain around tumor (BAT), cortex (C), and white matter (WM) after saline (NaCl 0.9%), PROLI/NO (1026 M), zaprinast (20 mg/kg), and LY83583 (1 mg/kg) alone or in combination with PROLI/NO and 8-bromo cGMP (1023 M). The values are expressed as the mean + SEM. P , .05 or P , .0001 compared with saline infusion (* or **) or with PROLI/NO (# or ##). released by PROLI/NO. Pretreatment with dexamethasone or indometacin, which both block prostaglandin synthesis by inhibition of COX, did not reduce increased BTB permeability induced by PROLI/NO (Fig. 5). Concomitant exposure to PROLI/NO and aminoguanidine, indometacin, or dexamethasone increased BTB permeability to [14C]AIB. The permeability changes observed in the tumor were also observed to a lesser degree in the brain around the tumor, but not in tumor-free cortex or white matter. Blood Flow Studies When using [14C]IAP QAR to measure rCBF, no significant change in tumor blood flow occurred after 1026 M PROLI/NO which, at the same dose and infusion interval, produced significant BTB disruption (Figs 2C and D and 6). rCBF in all other brain regions was also not changed after PROLI/NO infusion in comparison with saline controls. Vital Signs PROLI/NO at a dose of 1026 M did not cause significant changes in MABP at any infusion rate or in any route. In contrast, zaprinast alone or in combination with PROLI/NO caused a significant drop in the mean arterial pressure (P , .001), which did not recover until the end of the experiment. None of the other compounds affected the systemic circulation (Fig. 4). Fig. 4. Mean arterial blood pressure (MABP) measured before and after infusion of drug or vehicle as well as at the end of the experiment. Values are expressed in mm Hg and given as mean + SEM. *P , .05 and **P .0001. Fig. 5. K1 values for [14C]AIB (mL g21 min21) in tumor (T), brain around tumor (BAT), cortex (C), and white matter (WM) after saline (NaCl 0.9%), PROLI/NO (1026 M), aminoguanidine (200 mg/kg), indometacin (7 mg/kg), and dexamethasone (10 mg/kg) alone or in combination with PROLI/NO. The values are expressed as the mean + SEM. P , .05 or P , .0001 compared with saline infusion (* or **) or with PROLI/NO (# or ##). PROLI/NO increases selective intratumoral uptake of [14C]AIB and [14C]sucrose after ICA infusion over 3 minutes and after i.v. infusion over 15 minutes at a dose that had no hemodynamic side effects as described in this study, as well as previously.20 As the transport of water-soluble compounds across the BBB and into tumors is strictly dependent on molecular weight, uptake of [14C]AIB (MW 103 D) into C6 gliomas was Fig. 6. rCBF (F ) values for [14C]IAP (mL 100 g21 min21) in tumor (T), brain around tumor (BAT), cortex (C), and white matter (WM) after a 3-minute ICA infusion of saline (NaCl 0.9%) or 1026 M PROLI/NO. The values are expressed as the mean + SEM. significantly higher than that of [14C]sucrose (MW 372 D) with K1 values of 24.9 + 0.9 and 11.5 + 1 mL g21 min21, respectively. We showed earlier that ICA and i.v. infusion of PROLI/NO are equally effective when a 3-minute infusion period is used. When the same dose is given i.v. over 15 minutes, increased tumor microvessel permeability to [14C]AIB and [14C]sucrose occurred, but to a lesser degree, possibly due to rapid inactivation of NO in plasma by reactive species. The selectivity of the NO effect on the BTB is based on unique characteristics that differentiate it from the normal BBB, eg, changes in the molecular structure of tight junctions, overexpression or loss of receptors, ion channels, and enzymes in tumor tissue and tumor microvessels such as eNOS and iNOS.4,29 This higher expression of iNOS correlates with the degree of malignancy, angiogenesis, and microvessel density in human gliomas.30 Overexpression of NOS isoforms has been confirmed in C6 gliomas.8 This permeability effect induced by PROLI/NO appears to be mediated directly by activation of the cyclic GMP pathway, as it can be significantly blocked by the sGC inhibitor LY83583 preventing cGMP generation in response to NO. Involvement of the cGMP pathway is also confirmed by the fact that the phosphodiesterase inhibitor zaprinast showed a tumor-selective BBB comparable to PROLI/NO by blocking cGMP degradation. Although PROLI/NO and zaprinast alone were effective BBB permeabilizers, use of them in combination was not synergistic. In contrast, combining these highly vasoactive compounds produced detrimental systemic effects leading to severe hypotension in all animals and intratumoral hemorrhage in 2 animals. Similar observations were made previously with high doses of PROLI/NO (1022 M), which were poorly tolerated by the animals as the strong vasodilatory effect of NO dominated over its effect on vascular permeability. These observations indicate, however, that an NO effect on the BBB can be achieved independent of, and at lower doses than, the systemic vasodilatory effect. The PROLI/NO effect on the BTB was reproduced by intravascular delivery of the cGMP analog 8-bromo-GMP, confirming the assumption that NO activation of the cGMP pathway appears to be the predominant mechanism in regulating the BTB. It has been shown previously that 8-bromo-GMP can also modify microvascular permeability in a normal brain when applied locally to the cortex in a cranial window model and that ICA delivery of the lipid-soluble cGMP analog dibutyryl cGMP (db cGMP) enhances the transport of albumin across the BBB.28 Increased cGMP and NO levels detected in brain tumors, compared with brain endothelial cells and astrocytes, account for the increased leakiness of the BTB and higher tumor blood flow compared with the normal brain.31 Increased tumor microvessel permeability by PROLI/ NO appears to be mediated through the same pathway as bradykinin-induced disruption of the BTB, as both can be modified by blocking or activating the cGMP pathway.32 Since NO is a membrane-permeable mediator that directly activates sGC and other second messenger pathways, it acts independently of surface receptor expression such as bradykinin or NOS. Thus, the expression and activation of NOS in the tumor or in tumor endothelial cells is irrelevant when using an exogenous NO source such as PROLI/NO, as these donors release copious amounts of NO in comparison with iNOS. Blocking iNOS with aminoguanidine did not alter BBB disruption in response to PROLI/NO, since it does not depend on iNOS expression. Further, aminoguanidine itself did not modify BTB permeability to [14C]AIB in our study. PROLI/NO is quickly hydrolyzed to NO once it enters the blood stream of the animal. Maximum caution was taken to maintain the stability of the compound prior to injection by increasing the pH and decreasing temperature. The stability of PROLI/NO in injectable dosing solutions was thoroughly investigated by Waterhouse et al.33 Products of PROLI/NO hydrolysis are mainly nitrite that represents as NOs primary oxidation product at physiological pH and proline, and N-nitrosoproline. Considering NOs short half-life at physiological pH and 378C, we cannot prove whether NO itself or its metabolite nitrite is the active agent opening the BBB. We could show earlier that equimolar concentrations of sodium nitrite had similar disrupting effects on the BTB as PROLI/NO and that proline was inefficient.20 In addition to the oxidation of NO to nitrite, nitrite reduction to NO by deoxyhemoglobin is also observed in the circulation.34 The duration of the BBB opening appears to be between 15 and 30 minutes. Using bradykinin as a permeabilizer, which activates the NO/cGMP second messenger system, Sugita et al.32 showed that the increase in BTB permeability observed after 15 minutes is almost completely abolished after 30 minutes. Performing QAR 30 minutes after PROLI/NO infusion in 2 rats, we also did not see any increased uptake of tracer into the tumor, indicating a closure of the BTB (data not shown). NO regulates BBB permeability during hypertension and in response to inflammatory mediators such as bradykinin, histamine, cytokines, or prostaglandins through a variety of pathways.21 Outside the CNS, NO predominantly derived from eNOS and iNOS controls microvascular permeability and vascular tone in solid tumors. Low levels of NO produced by eNOS might maintain barrier integrity, and NO from iNOS activates signaling pathways that lead to barrier dysfunction. Alterations in iNOS expression and increased release of NO result in increased permeability, which can be inhibited by scavenging NO or by blocking NOS, bradykinin receptors, or COX.7,26 In other pathological conditions with increased BBB permeability, such as meningitis, it is possible to ameliorate BBB dysfunction by iNOS inhibition using i.v. delivery of aminoguanidine, but not by pharmacological inhibition of prostanglandin synthesis.35 In our study, inhibition of COX-2 by indometacin did not change [14C]AIB uptake into C6 gliomas and did not reduce the permeability response to PROLI/NO. Other groups report a reduction in BBB permeability by COX inhibition and blocking the degradation of tight junction proteins by matrix metalloproteinases in bacterial meningitis and cerebral ischemia.36 In experimental colon carcinoma, tumor vascular permeability can be reduced by treatment with a COX-2 inhibitor.37 Although modulation of arachidonic acid metabolism influences peripheral vascular permeability, it does not seem to be involved in the NO-induced opening of the BTB after delivery of PROLI/NO. Dexamethasone is the mainstay of edema treatment in patients with malignant gliomas. It has been described previously that blood-to-tumor transport of [14C]AIB in rat C6 glioma is significantly reduced by pretreatment with dexamethasone.38 In contrast, Molnar et al.39 did not observe any effect of dexamethasone on permeability or on tumor blood flow by double-labeled QAR in the RG-2 glioma model. The exact mechanism by which dexamethasone repairs a disrupted BBB is still unknown. Heiss et al.40 suggested inhibition of vascular endothelial growth factor/vascular permeability factor and involvement of the glucocorticoid receptor. Another potential mechanism to explain the effects of dexamethasone to repair BBB dysfunction is through the inhibition of COX-2.41 Portnow et al.42 described an equal effect of dexamethasone and a COX-2 inhibitor (SC-236) in controlling peritumoral edema in 9 L gliosarcomas. In our study, the permeabilizing effect of PROLI/NO was not significantly antagonized by dexamethasone. Gu et al.43 showed a significant reduction in baseline BTB permeability by dexamethasone in C6 gliomas, but no influence on the bradykinin-mediated BTB permeability increase. Bradykinin-induced BTB increase is mediated through adenosine-sensitive potassium (K(ATP) channels, which are overexpressed in gliomas.44 The increase in the K(ATP) channel expression in C6 gliomas in response to dexamethasone might explain the inability of dexamethasone to inhibit the bradykinin-induced permeability increase. The C6 tumor model was chosen for this study because it is well established and was used frequently in earlier autoradiography studies.38 The variability in capillary permeability and blood flow between tumor models has been described in detail.45 Using a standard tumor model, such as the C6 glioma, allows comparison of our data with the literature with reference to baseline BBB permeability and blood flow and comparison of the efficacy of NO donor therapy with other methods of BBB disruption. In spite of its immunogenicity, the C6 glioma model continues to be used in tumor biology studies, but not in survival studies.46 The impact of the inflammatory response to immunogenic tumor cells on BTB permeability appears to be negligible, as inhibition of inflammatory pathways did not reduce tracer uptake in the control group and after PROLI/NO treatment. Selectively enhancing tumor blood flow can be used as an approach to enhance the efficacy of cancer treatments. As tumor blood vessels are already dilated under the influence of NO generated by endothelial eNOS and tumoral iNOS, it was not clear if exogenous delivery of high doses of NO would further increase tumor blood flow in comparison with CBF. We did not observe increased tumor blood flow in response to PROLI/NO and hence we can assume that the increased tracer uptake is a result of the NO effect on the BTB and not a result of blood flow changes. The variability in intratumoral blood flow, as expressed by our studies to be a large standard error, is explained by the great differences of blood flow in the tumor center and the more vascularized tumor periphery, which have been described earlier.23,45 It is important to note that BTB disruption was achieved without increased permeability in the normal brain or changes in CBF. This might be an advantage compared with osmotic BBB disruption with mannitol, which causes a much stronger permeability effect in the normal brain than in the tumor and poses the risk of side effects in patients.47 Although some groups try to block NOs actions to modulate tumor blood flow in an attempt to maintain vasodilator tone using NOS inhibitors, our strategy uses exogenous NO delivery by PROLI/NO to exploit NO effects on tumor vasculature to increase the delivery of antitumor drugs into brain tumors.48 We reported previously that PROLI/NO could be safely used to enhance BTB permeability to carboplatin in a rat C6 glioma model, an effect that led to a significant tumor response and long-term survival in 40% of the rats.20 Understanding the BBB and its biochemical regulation and establishing the mechanism by which NO exerts its action in tumor and tumor endothelial cells and in the microvessels in the vicinity of the infiltrative edge of the tumor will be increasingly important as efforts continue to improve anticancer drug delivery to brain tumors. The results of this study corroborate our idea to use NO released by NONOates to further modulate tumor microvessel leakiness to increase the selective delivery of drugs into tumors. Improvements in the NO donor design also suggest investigating other antitumor properties of NO in gliomas, such as NO effects on tumor growth, metastasis, response to chemo- and radiotherapy, invasion, migration, and tumor angiogenesis. Conflict of interest statement. None declared. This project has been funded with federal funds from the National Cancer Institute, National Institutes of Health, under contract HHSN261200800001E, and by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.


This is a preview of a remote PDF: https://neuro-oncology.oxfordjournals.org/content/13/2/203.full.pdf

Astrid Weyerbrock, Stuart Walbridge, Joseph E. Saavedra, Larry K. Keefer, Edward H. Oldfield. Differential effects of nitric oxide on blood–brain barrier integrity and cerebral blood flow in intracerebral C6 gliomas, Neuro-Oncology, 2011, 203-211, DOI: 10.1093/neuonc/noq161