Journal of Neuro-Oncology 10: 13-25, 1991. t~) 1991 Kluwer Academic Publishers. Printed in the Netherlands.

Laboratory Investigation

Increased capillary permeability in rat brain induced by factors secreted by cultured C6 glioma cells: role in peritumoral brain edema*

Takanori Ohnishi ~, Peter B. Sher 1, Jerome B. Posner 1, z and William R. Shapiro 1, 2

1George C. Cotzias Laboratory of Neuro-Oncology, Department of Neurology, Memorial Sloan-Kettering Cancer Center,"2Department of Neurology, CorneU University Medical College, New York, New York 10021, USA

Key words: edema, tumors, capillary permeability, blood-brain barrier Summary To investigate whether brain tumors secrete a factor(s) responsible for peritumoral brain edema, we studied the effect of conditioned medium from cultured C6 glioma cells on rat brain capillary permeability. Three different fractions of conditioned medium were obtained. SUP-N was a culture supernatant incubated 4 hours in serum-free medium. SUP-C was the 60--100 fold concentrated fraction obtained by dialysisconcentration of SUP-N; it contained 950/z g/ml of protein > 10 k-daltons from 3 × 108 cells. SUP-L was a water-dispersible lipid fraction from SUP-N; the major components of SUP-L were neutral lipids and free fatty acids. The supernatant fractions and their corresponding control solutions were infused into normal rat brain, and capillary permeability was determined using quantitative autoradiography by measuring the unidirectional entry constant, K (/z 1/g. min), of ~4C-alpha-aminoisobutyric acid (14C-AIB) into brain tissue. SUP-C and SUP-L significantly increased capillary permeability of normal brain; the effect of SUP-C was more intense and extensive than that of SUP-L. The highest mean K value (Kmax) of SUP-C was 10.83 ___0.99 and that of the control was 2.53 + 0.22 (p < 0.001). The Kmax of SUP-L was 5.61 + 0.23 and that of the control was 2.67 + 0.36 (p < 0.01). A time-course study after infusion of SUP-C demonstrated that more than 1.5 hours is required for the supernatant fraction to open the barrier and that the effect of SUP-C was reversible. The increase of capillary permeability induced by SUP-C was significantly inhibited by pretreatment of rats with dexamethasone (10 mg/kg, ip) 1 hour before intracerebral infusion of SUP-C (Kmax (untreated): 8.30 + 0.82, Kmax (treated): 1.33 + 0.64, p < 0.001). These results indicate that experimental brain tumors secrete at least two different diffusible factors responsible for capillary endothelial leakage in normal brain. One is a protein of molecular weight greater than 10 k-daltons, whose effect is inhibited by glucocorticoids, and the other is a waterdispersible lipid.

Introduction Although it is known that peritumoral brain edema is vasogenic, i.e., characterized by increased permeability of brain capillary endothelial cells, the etiology and molecular mechanism(s) underlying

edema formation are poorly understood. Investigation of the microvasculature of human and experimental brain tumors has revealed striking structural alterations of the tumor capillary vessels, including fenestrations, and increased number of pinocytotic vesicles [1-4]. It is likely that these

*Presented in part at the meeting of the American Association for Cancer Research, Atlanta, GA, May 1987.

14 leaky tumor vessels are responsible for the increased capillary permeability within brain tumor tissues. Our quantitative autoradiography studies in experimental brain tumors with C6 gliomas [5], demonstrated that not only were the tumor capillaries altered, capillary permeability in brain adjacentto-tumor and brain surrounding tumor, in which there are usually few or no tumor cells, was also increased. Similar observations have been made with metastatic Walker 256 [6], RG-2 [7] and RT-9 brain tumors [8]. In each case, the capillaries surrounding the brain tumor were modified. One possible source for this effect would be that the tumor produces a diffusible factor(s) that alters capillary permeability of normal brain vessels. Senger et al. [9, 10] have identified a vascular permeability factor from ascites and a variety of other tumor cells. They emphasized that tumor stroma plays an essential role in the growth of all solid tumors. Major components of tumor stroma (e.g., fibrin), moreover, are derived from host plasma protein, requiring that the vasculature's normally low permeability to such proteins must increase substantially [11]. An increase in the tumor microvasculature enables the tumor to grow beyond 1 to 2 mm in diameter [12]. There is accompanying increased permeability with leakage of plasma protein essential to the generation of the stroma. This is critical for the growth of brain tumors because plasma proteins do not cross an intact blood-brain barrier. Investigation and identification of vascular permeability factors in brain tumors may lead not only to an understanding of the molecular mechanisms involved in the formation of peritumoral brain edema, but also to the development of new therapeutic strategies. Recently, Bruce et al. [13] and Criscuolo et al. [14] described a vascular permeability factor produced in the culture by human malignant gliomas. The material increased the permeability of guinea pig dermis capillaries (Miles assay). The effect was blocked by dexamethasone, either incubated with the tumor cultures or given to the guinea pigs before the injection of the factor. Characterization of the factor demonstrated that it had a molecular

weight of at least 30,000 Daltons, it was acid-stable, moderately heat-labile, inactivated by proteolytic enzymes and was cross-antigenic with the permeability factor described by Senger et al. [9, 10]. The effect of dexamethasone on improving cerebral edema has been recognized since the early studies of Galicich et al. [15]. Numerous studies have so far failed to elucidate the mechanism of such action. We have reported a time-course study of the effects of glucocorticoids on capillary permeability in brain tumor-bearing animals [16]. Dexamethasone reduced capillary permeability in the tumor within 1 hour of administration, but its effect continued for at least 12 hours, and by that time had also reduced capillary permeability in the brain adj acent-to-tumor. Theses data, along with the results of Bruce et al. [13] and Criscuolo et al. [14], suggest that the effect of dexamethasone may be on the permeability factors themselves, either on their production or on their action. In the present study, we searched for vascular permeability factors in conditioned medium from cultured C6 glioma cells. We identified the factors by their effects on capillary permeability of normal rat brain after intracerebral infusion. Capillary permeability was measured by determining the unidirectional entry constant of 14C-alpha-aminoisobutyric acid (14C-AIB) using quantitative autoradiography. With this technique, regional changes in capillary permeability can be quantitatively evaluated and correlated with histological sections. The effects of dexamethasone on the action of these factors was also determined.

Materials and methods

Preparation o f supernatants f r o m tumor cells in tissue culture

C6 glioma cells were purchased from the American Type Culture Collection (Rockville, MD) and maintained in tissue culture with McCoy's 5A medium supplemented with 10% fetal calf serum at 37°C in a humidified atmosphere containing 5% CO2 and 95% air. When the cells became con-

15 fluent, they were harvested with 0.05% trypsin and 0.02% EDTA in Hanks balanced salt solution and washed twice with cold phosphate buffer saline (PBS). The cells were then resuspended in Eagles minimum essential medium (EMEM) without serum and plated in 150cm 2 tissue culture flasks (Corning) at a concentration of 5 x 106 viable cells/ ml. Viability was determined with a 0.4% trypan blue dye exclusion method. Four hours after incubation at 37°C, the conditioned medium was collected and centrifuged at 5,000 x g for 20 min at 4° C. The supernatant was then passed through a 0.22/zm filter (Millex-GV, Millipore) to ensure complete removal of cells and cell debris. This supernatant fraction was named SUP-N. From the SUP-N fraction two additional fractions were prepared. SUP-N was aliquoted and one portion was concentrated by a negative pressure dialysis-concentrator (Bio-Molecular Dynamics, Beaverton, OR) with a dialysis membrane of 10,000 molecular weight cut-off size. The dialysis/ concentration was performed against PBS with Ca 2+ and Mg 2+ at 4° C. This concentrated fraction was named SUP-C. Protein contents of SUP-N and SUP-C were determined by the method of Bradford [17]. The yield of protein with this method was estimated to be 90%. The remaining portion of SUP-N was lyophilized and distilled; water, methanol, and chloroform (0.8 : 2 : 1) were added in that order. Total lipid was extracted from this solution and sonieated in PBS with Ca 2+ and Mg2+. This water-dispersible lipid fraction was named SUP-L. The preparation of SUP-L was performed under a stream of nitrogen. Lipid chromatographic analysis of SUP-L was carried out by two-dimensional thin layer chromatography using silica gel G plates (Analtech, Newark, DE) with chloroform-methanol-28% ammonium hydroxide (65 : 25 : 5) for the first development and chloroform-methanol-acetone-acetic acid-water (50 : 10 : 20 : 10 : 5) for the second development [18]. Iodine vapor was used to detect the separated lipids on the plate. All samples were stored at - 80°C before use. To obtain control samples for each fraction, EMEM was concentrated by a negative pressure dialysis-concentrator as a control for SUP-C and a water-dispersible lipid

solution was prepared from EMEM as a control for SUP-L. The pH of these samples was 7.2-7.4.

Intracerebral infusion of supernatant fractions Male Wistar rats (Charles River Breeding Laboratories, Wilmington, MA) weighing 325-350g were anesthetized with NzO-Oy-ethrane (70: 30: 5), and maintained on reduced ethrane. The rat's head was fixed in a stereotaxic apparatus (Model 900, David Kopf Instruments, Tujunga, CA) and the skull was exposed by a midline scalp incision. Burr holes were made 3 mm to either side of the sagittal suture and l m m anterior to the bregma. 30-gauge needles were inserted bilaterally through these holes to a depth of 6 mm into the brain. These points corresponded to right and left caudate nucleus-putamen [19]. The needles were connected to PE-10 polyethylene tubing (Clay Adams, Parsippany, N J). The supernatant fraction was infused through the right needle, and the control solution was infused through the left needle. The infusion was performed at a rate of 1/zl/min for 50 min using a Harvard pump (Model 944, Millis, MA). After infusion, the animals were allowed to awaken from anesthesia. In a second experiment, the timecourse of the effect of intracerebral infusion of SUP-C was examined by changing the time interval between the initiation of infusion and the adminstration of 14C-AIB. Because corticosteroid hormones are used to treat clinical cerebral edema, in a third experiment the effect of dexamethasone, 10 mg/kg, given intraperitoneally to rats 1 hour before starting the intracerebral infusion of the SUP-C fraction was determined.

Experimental procedure of 14C-AIB study Two hours after the animals were allowed to recover from the anesthesia, they were re-anesthetized with NyO-Oz-ethrane, and the right femoral artery was cannulated with PE-50 polyethylene tubing (Clay Adams, Parsippany, NJ) and the right femoral vein with V3 Vinyl tubing (Bolab Inc., Lake

16 Havasu City, AZ). The animals were mounted on a warming block equipped with a thermoregulator, allowed to awaken from anesthesia and permitted to recover for at least 2 hours before the 14C-AIB study. The 14-C-AIB was purchased as alpha-[1-14C]aminoisobutyric acid (50 mCi/mmol; New England Nuclear, Boston, MA). Radiochemical purity was determined by thin layer chromatography using silica gel G plates (Analtech, Newark, DE) with N-propanol and ammonium hydroxide (70: 30) and found to be greater than 99%. Immediately before administration, the pH of the 14C-AIB was adjusted to 7.4 with 1 N sodium hydroxide. Six hours after starting intracerebral infusion, 1 ml of 14C-AIB (100/.~Ci/ml) was injected into the femoral vein as a bolus. Timed blood samples were collected from the femoral artery on the following schedule: 0, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 5, 10, 15 min, for a total of 12 samples. Throughout the experiment, arterial blood pressure and body temperature were monitored and maintained at 126 + 0.7 mm Hg and 37.2 + 0.1 ° C, respectively. Other physiological measurements, which did not change significantly during the experiments, were as follows: pH, 7.45_+ 0.03; pO2, 94.2_+ 4.1mmHg; pCO2, 36.2 _+ 2.8mmHg. All animals were killed by decapitation 15 min after 14C-AIB administration, and the brains were dissected within 1 min and placed in Freon-12 (DuPont, Wilmington, DE) on dry ice. The frozen brains were mounted on planchettes with M-1 embedding matrix (Lipshaw, Detroit, MI) and stored at - 80° C. The collected blood samples were immediately centrifuged and 20ml of plasma was weighed in a scintillation vial. ACS II (Amersham Corp, Arlington Heights, IL) scintillation counting solution was added to each vial and plasma radioactivity was measured with a Beckman liquid scintillation counter (Model LS3801, Beckman Instruments, Fullerton, CA).

Quantitative autoradiography The techniques for this procedure have been previously described [5]. Briefly, the frozen brains were

cut 20 mm thick on a cryostat (Hacker Instruments, Inc., Fairfield, NJ). Adjacent brain sections were either mounted on coverslips and immediately dried at 60 ° C for autoradiography, or mounted on glass slides and stained with hematoxylin-eosin. The sections on the coverslips were exposed to x-rat film (SB-5, Eastman Kodak Co., Rochester, NY) along with 14C-methylmethacrylate standards (Amersham Corp) previously calibrated to 20-mm brain sections of known radioactivity. After 10 days of exposure, the films were developed and the radioactivity was measured as follows. Individual coronal brain sections on the autoradiograms were digitized along with the standards using an Eikonixscan image digitizer (Model 785, Eikonix Corp., Bedford, MA). Histological sections were digitized in the same way, and used to align the adjacent autoradiograms. To convert the autoradiograms into units of radioactivity, the optical densities of images of 14C standards were determined and a standard curve relating optical density to tissue radioactivity was generated for each film. Based on this curve, the optical density of each brain image could be converted to radioactivity values (nCi/g of tissue). Capillary permeability, expressed as a unidirectional blood-to-brain transfer constant (K) of lacAIB, was calculated according to the following equation: K = ( C i ( T ) - Vp. Cp(T))/ fT Cp(t) dt where Ci(T) is the tissue concentration (nCi/g) of a4C-AIB at the end of the experimental time T (15 min), Cp is the concentrationn of 14C-AIB in the arterial blood plasma (nCi/ml), and Vp is the plasma volume (ml/g) in the region of brain. The value of Vp for the caudate nucleus-putamen was assumed to be 0.005 ml/g. [20]. K has units of/zl/g • min. This relationship assumes no brain-to-blood backflux of AIB during the experimental period. Based on the resolution of the optical density measurements above x-ray film sensitivity background, the reliable lower limit for K was 0.4/zl/g • min.

17

Measurement of altered capillary permeability after intracerebral infusion For each animal, the histological section demonstrating the largest needle track was digitized. Three consecutive adjacent autoradiograms were aligned to the histological section using the computer for the alignment. Each needle track on the histological section was traced on the screen and sequential rings were drawn around the trace of the track every 0.2 mm (as shown in Fig. 3 upper right corner). The K values were determined in the areas contained within each ring for each of the three autoradiographic sections; the mean of the K value and its variance were calculated for each ring. The mean K values were plotted as a function of distance from the needle tract. Care was taken to exclude ventricular and cortical surfaces from the measurements because such surfaces often contain artifactually intense radioactivity. From the curves of K values versus distance, the average of the three highest mean K values was determined and d e s i g n a t e d Kma x. A parameter of distance related to the infusion was defined a s Dmax, the maximum distance (in mm) from the needle track at which the mean K values associated with the infused test materials were greater than two standard devia-

tions from the mean of the controls in the contralateral cerebral hemisphere.

Results

Characterization of supernatants from Co glioma cells Three batches of culture supernatants were prepared from the same culture passage (passage 56) of C6 glioma cells by thawing the cells before each experiment. The average cell viability 4 hours after incubation in serum-free MEM was 95%. The rate of floating cells was 0.02% of the total number of plated cells. The first batch of 3 × 108 cells was used for the infusion study (SUP-N, n = 3; SUP-C, n = 6) and for the dexamethasone-treatment study (SUP-C, n = 6). The second batch of 3 x 108 cells was used for the SUP-L (n = 3) infusion study. The third batch of 1 × 108 cells was used for the timecourse study. The characteristics of the fractions for the capillary permeability infusion study are listed in Table 1. SUP-C was concentrated 60-fold relative to SUP-N. The protein concentration of SUP-N was 15/zg/ml and of SUP-C 950/zg/ml. SUP-L was con-

Table 1. Characterization of C6 glioma supernatants and their effect on capillary permeability of normal rat brain. SUP-N was the culture supernatant from 3 x 108 cells incubated 4 hours in serum-free medium. SUP-C was a dialysis-concentrated fraction and SUP-L a water-dispersible lipid fraction, both of SUP-N. For the permeability studies, the supernatant infusion was given over 50 min into the rat's right cerebral hemisphere, the control solution into the left cerebral hemisphere. 14C-AIB was administered iv 6 hours after the start of the infusion and capillary permeability was then determined Infusion materials

SUP-N supernatant (n = 3) control SUP-C supernatant (n = 6) control SUP-L supernatant (n = 3) control

Concentration relative to SUP-N

Protein concentration

-

15/~g/ml

× 60

950~g/ml

x 80

Lipids b

Capillary permeability a Kmax (/~l/g min)

Dmax (mm)

2.86 + 2.24_+ 10.83 __ 2.53 _+ 5.61 _+ 2.67__

0.97 + 0.18

0.21 0.27 0.99 c 0.22 0.23 d 0.36

3.86 _+ 0.03 1.30 __ 0.12

a Capillary permeability is expressed as Kma~and Dmax as defined in the text. Values are mean + SEM. b Included primarily neutral lipids and free fatty acids, along with minor amounts of phosphatidylethanolamine, phosphatidylcholine and phosphatidylserine. Significantly different from control, p < 0.001 (t-test). d Significantly different from control, p < 0.01 (t-test).

18

19 centrated 80-fold relative to SUP-N. SUP-L was a water-dispersible lipid solution, whose major components were neutral lipids and free fatty acids. SUP-L also included phosphatidylethanolamine, phosphatidylcholine and phosphatidylserine as minor components. For the time-course study with the lower cell inoculum (see below), SUP-C was concentrated 100-fold and had a protein concentration of 750/zg/

needle track became edematous, with infiltration of neutrophils, mostly polymorphonuclear leukocytes; the number of leukocytes was increased in vessels near the needle track (Fig. 2A). In SUP-Linfused brain, there was no obvious edema and many fewer neutrophils were seen around the tip of the needle track (Fig. 2B)

ml.

Quantification of altered capillary permeability after intracerebral infusion of C6 supernatants

The effect of Co supernatants of brain and capillary permeability

To evaluate quantitatively the effect of the infusion materials on capillary permeability, we determined mean K values in sequential 'rings' around the needle track and these mean K values were plotted as a function of distance from the needle track. Figure 3 shows an illustration of these sequential rings around the needle track and the diffusion curves of the three different supernatant fractions; Kma~ and Dmaxwere determined from these curves (see Methods). The values of Kmax and Dma x for the three different supernatant fractions (SUP-N, SUP-C and SUP-L) are presented in Table 1 and compared graphically in Fig. 4. The effect of SUP-N was not significantly different from that of the control, but SUP-C and SUP-L showed significantly different K .... s compared to those of the corresponding control (p < 0.001 for SUP-C, p < 0.01 for SUP-L; paired t-test). Although it is not possible to compare exactly the effects of SUP-C with those of SUP-L, it may be estimated that the Kmax of SUP-C was 2.8 times as great as that of SUP-L and the Dmax of SUP-C was 3 times as great as that of SUP-L, because the concentration was almost the same in volume when these two fractions were prepared from SUP-N.

Figure 1 depicts computer-generated images of typical coronal sections of autoradiograms after each supernatant fraction (right side) and the corresponding control solution (left side) were infused into normal rat brain. In no case where the control solution was infused was the K value increased. SUP-N did not show a significant effect on capillary permeability, although there was a slight rise in K values just around the needle track (Fig. 1A). The distribution patterns of K values in SUP-C and SUP-L-infused brains were different. SUP-C altered the capillary permeability most intensely and extensively of the three supernatant fractions (Fig. 1B), while SUP-L produced less increased uptake of 14C-AIB (Fig. 1C). In addition to the high K values around the needle track, SUP-C-infused brain had high K values over a wide area far from the needle track, while the effect of SUP-L infusion was localized to the area surrounding the tip of the needle, forming a ring- or horseshoe-like pattern of K values on the autoradiogram. The microscopic appearances of brain tissue after infusion of SUP-C and SUP-L also showed characteristic differences (Fig. 2). Brains infused with SUP-N or control solutions demonstrated only mechanical destruction around the needle track. In SUP-C-infused brain, tissues around the

Time-course study in C6 SUP-C infusion The time-course of the effect of intracerebral in-


-

_P /

m < W

15SUP-C

129G3-

Pc W 0_

0

;

I

I

I

n~

5

15.

n < 0

12-

SUP-L

9-

30

oo

o:s

l:O

l:S

2:o

DISTANCE FROM NEEDLE TRACK

4.b

(mm)

Fig. 3. Diffusion curves of transfer constant (K) for AIB in rat brains 6 hours after infusion with C6 glioma supernatants SUP-N, SUP-C and SUP-L. Mean K values in the sequential 'rings' around the needle track were calculated and plotted against the distance from the needle track. At the upper right corner, sequential 'rings' around the needle track ('N') are illustrated.

iting the reaction by which a factor in SUP-C may affect capillary endothelial cells of the brain. At present, a few properties of the protein factor in SUP-C fraction have been defined [28]. The factor was sensitive to heating at 70°C for 40 min and more than 70% of the activity of SUP-C was lost after such treatment. On the other hand, storage of the SUP-C fraction at - 8 0 ° C for three months did not change the activity. When membrane ultrafiltration (YM5, Amicon, Lexington, MA), which permits rapid and simultaneous dialysis and concentration, was used for the preparation of SUP-C, the yield of protein was very low (2030%) and the SUP-C fraction obtained showed almost no effect on capillary permeability. This means that either the amount of factor included in the SUP-C fraction was very small or surface denaturation of the factor occurs in this method. It is possible that the SUP-C permeability factor

22 12 --~

.-- 12

__

Kmax SUPERNATANT 10-

"~"

ITmTT] Dmax SUPERNATANT

10

8-

8

E c~ ::::L

E *

6-

|

6

!

-

× u E ,¢,

E ,,._.,

4-

2 -t

°

E c~ 4

I~

~

I

~lllllllll

I

~

0

I- 2

0 SUP-N

SUP-C

SUP-L

Fig. 4. Effects of C6 glioma supernatants on capillary permeability. Capillary permeability is expressed as K ~ , the highest K value of the rings, and Dmax, the distance (in mm) from the needle track at which the mean K values associated with the infused test materials were greater than two standard deviations from those of the controls in the contralateral cerebral hemisphere. Bars: mean + SEM for three separate experiments for SUP-N and SUP-L and six experiments for SUP-C. Kmaxsignificantly different from the control at * p < 0.01, • * p < 0.001 (paired t-test).

is the same as that described by Bruce et al. [13] and Criscuolo et al. [14], and if so, probably also relates to that described by Senger et al. [9, 10], because both share cross-antigenicity. However, there is at least one difference between SUP-C and the factor of Bruce et al. While dexamethasone blocked the in vivo effect of both the Bruce factor [13] and of SUP-C, the in vitro effect differed. We incubated C6 glioma cells with dexamethasone for 4 hours, and found that the SUP-C fraction obtained from the culture medium produced a greater rise in K values than that without such treatment [28]. In contrast, Bruce et al. [13] found that incubation of their glioma cells with dexamethasone reduced the permeability factor. Recently, some correlation of tumor plasminogen activator (TPA) with peritumoral brain edema was indicated in the relationship between TPA levels in specimens of human brain tumors and the amount of peritumoral edema shown in computerized tomography [29]. The factor in SUP-C from C6 gliomas, however, seems to be different from TPA, even if TPA is included in the SUP-C frac-

tion. The production of TPA in cultured cells is markedly inhibited by dexamethasone [29]. This means that the effect of SUP-C obtained after the tumor cells are treated with dexamethasone should be less than that obtained from untreated cells, if the factor in SUP-C is TPA. It has been shown that some proteases such as trypsin, pronase and collagenase induce increased permeability of the blood-brain barrier when infused into the rat brain ventricle [30, 31]. Among these, collagenase has the most conspicuous effect on capillary permeability. A number of tumors have been shown to produce collagenase [32], which has been implicated in tumor invasion and metastasis by digestion of the extracellular matrix and basement membrane [33]. Gazendam et al. [34] demonstrated that intraventricular infusion of collagenase increased capillary permeability to Evans blue or horseradish peroxidase without altering the structures in the brain parenchyma. However, the effect of collagenase was more prominent at 24 and 48 hours after infusion than at 6 hours and pathological changes were still demon-

23 5-

E 2 E

X U

E 1 .,,/

10 TIME BETWEEN

20 INFUSION

310 AND A I B - S T U D Y

410

510

(hrs)

Fig. 5. T i m e course of increased capillary permeability after intracerebral infusion of S U P - C supernatant fraction. Capillary permeability is expressed as the difference in Kmax between S U P - C supernatant and the control. Bars: m e a n _ range for time points: 1.5 hours, n = 3; 6 hours, n = 2; 24 hours, n = 3; 48 hours, n = 2.

strable after 4 and 5 days. Thus, the time-course of the collagenase effect is very different from that of SUP-C seen in our study. Whether the factor in SUP-C has proteolytic activity is the subject of a future study. The mechanism of action of SUP-C in increasing capillary permeability remains to be elucidated. A reaction with a lag-time, as occurs with SUP-C (Fig. 5), generally excludes a direct effect of the causative material. Interleukin 1 (IL-1) induces platelet activating factor (PAF) production by human endothelial cells after a lag-phase of 1-2 hours, reaching a maximum after 4-6 hours [35], suggesting the activation of an intermediate biochemical reaction between IL-1 stimulation and PAF production. As in the case of IL-1, SUP-C may increase capillary permeability of brain endothelial cells through the production of an intermediate biochemical substance. It is possible, in fact, that SUP-C induces an inflammatory response, which may be part of its mechanism of action in increasing vascular permeability, or may be an epiphenomenon. While an inflammatory response was present in the brain tissue around the needle track of the

SUP-C infused animals, a lessor leukocyte infiltrate also occurred with SUP-L and even in the controls. In part, at least, this inflammation was associated with tissue destruction due to the needle insertion. Furthermore, the effect of SUP-C on permeability was not a result of the presence of non-specific denatured protein, because, as noted above, heating SUP-C to 70 ° C markedly retarded the permeability effect. Future studies should define the mechanism of action of these factors. Blood-brain barrier is a highly organized cellular machinery for strictly maintaining the environment of the central nervous system under normal physiological conditions. Brain tumors disturb the barrier and result in the formation of peritumoral brain edema, which substantially contributes to neurological morbidity and sometimes mortality in patients with malignant gliomas and metastatic brain tumors. Our present study shows that both morphological changes and functional alteration of capillary permeability through the secretion of vascular permeability factors from tumor cells can be responsible for the disruption of the barrier in normal brain tissue. They may also be responsible for such

24 10

10-

seas Cancer Fellowship of the Foundation for Promotion of Cancer Research, and of a grant from the Charles H. Revson Foundation. Present address: Department of Neurosurgery, Osaka University Medical School, Osaka, Japan.

UNTREATED TREATED

8-

"Z E

E

6-

6

4

4-

E E E

2-

I

0

Kmax

Dmax

Fig. 6. Effect of dexamethasone pretreatment (10 mg/kg, ip) on the SUP-C induced increase of capillary permeability of the rat brain. The values of Kin= and Dm~ are shown as the difference of those values between supernatant and control. Bars: mean + SEM for six experiments for untreated group and three for dexamethasone-treated group. * significantly different from untreated group, p < 0.001 (unpaired t-test).

changes in tumor capillaries themselves. While these vascular permeability factors ultimately produce the pathological condition of brain edema, the associated barrier disruption may be essential for tumor growth because the entry of elements for the tumor matrix is normally restricted by the blood-brain barrier. Recently, we have determined that two human glioma cell lines also produce a protein factor similar in effect to that seen in SUP-C from C6 glioma cells, while normal human glial cells do not [28]. Characterization of these factors may present a new modality in the treatment of malignant brain tumors.

Acknowledgements This work was supported by Grants CA 39208 and CA 08748 from the National Institutes of Health. Dr. Ohnishi is the recipient of Japanese Over-

References 1. Waggener JD, Beggs JL: Vasculature of neural neoplasms. In: Thompson RA, JR Green (eds) Neoplasia in the Central Nervous System. Advances in Neurology, Vol 15, Raven Press, New York, 1976, pp 27--49 2. Long DM: Capillary ultrastructure and blood brain barrier in human malignant brain tumors. J Neurosurg 32: 127-144, 1970 3. Cox D J, Pilkington GJ, Lantos PL: The fine structure of blood vessels in ethylnitrosourea-induced tumors of the rat nervous system, with special reference to the breakdown of the blood-brain barrier. Br J Exp Patho157: 419-430, 1976 4. Dean BR, Lantos PL: The vasculature of experimental brain tumors. J Neurol Sci 49: 55-66, 1981 5. Hiesiger EM, Voorhies RM, Busier MA, Lipschutz LE, Posner JB, Shapiro WR: Opening the blood-brain and blood-tumor barriers in experimental rat brain tumors. The effect of intracarotid hyperosmolar mannitol on capillary permeability and blood flow. Ann Neurol 19: 50-59, 1986 6. Blasberg RG, Shapiro WR, Molnar P, Patlak CS, Fenstermacher JD: Local blood-to-tissue transport in Walker 256 metastatic brain tumors. J Neuro-Oncol 2: 205-218, 1984 7. Groothuis DR, Fischer JM, Pasternak JF, Blasberg RG, Vick NA, Bigner DD: Regional measurement of blood-totissue transport in experimental RG-2 rat gliomas. Cancer Res 43: 3368-3373, 1983 8. Molnar P, Blasberg RG, Horowitz M, Smith B, Fenstermacher J: Regional blood-to-tissue transport in RT-9 brain tumors. J Neurosurg 58: 874-884, 1983 9. Senger DR, Galli SJ, Dvorak AM, Perruzzi CA, Harvey VS, Dvorak HF: Tumor cells secrete a vascular permeability factor that promotes accumulation of ascites fluid. Science 219: 983-985, 1983 10. Senger DR, Perruzzi CA, Feder J, Dvorak HF: A highly conserved vascular permeability factor secreted by a variety of human and rodent tumor cell lines. Cancer Res 46: 5629-5632, 1986 11. Dvorak HF: Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing. N Engl J Med 315: 1650-1659, 1986 12. Folkman J: Tumor angiogenesis. Adv Cancer Res 19: 331358, 1974 13. Bruce JN, Criscuolo GR, Merrill MJ, Moquin RR, Blacklock JB, Oldfield EH: Vascular permeability induced by protein product of malignant brain tumors: inhibition by dexamethasone. J Neurosurg 67: 880-884, 1987

25 14. Criscuolo GR, Merrill MJ, Oldfield EF: Further characterization of malignant glioma-derived vascular permeability factor. J Neurosurg 69: 254-262, 1988 15. Galicich JH, French LA, Melby JC: Use of dexamethasone in the treatment of cerebral edema associated with brain tumors. Lancet 81: 46-53, 1961 16. Shapiro WR, Hiesiger EM, Cooney GA, Lipschutz LE, Basler GA, Posner JB: Temporal effects of dexamethasone on capillary permeability of rat C6 glioma. J Neuro-Oncol 8: 197-204, 1990 17. Bradford MM: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248-254, 1976 18. Rouser G, Kritchevsky G, Yamamoto A: Column chromatographic and associated procedures for separation and determination of phosphatides and glycolipids. In: Marinetti GV (ed) Lipid Chromatographic Analysis. Marcel Dekker Inc, New York, 1976, pp 713-776 19. Pellegrino LJ, Pellegrino AS, Cushman AJ: A stereotaxic atlas of the rat brain. 2nd ed Plenum Press, New York, 1979 20. Blasberg RG, Fenstermacher JD, Patlak CS: Transport of ct-aminoisobutyric acid across brain capillary and cellular membranes. J Cereb Blood Flow Metab 3: 8-32, 1983 21. Rapoport SI, Fredericks WR, Ohno K, Pettigrew KD: Quantitative aspects of reversible osmotic opening of the blood-brain barrier. Am J Physio1238: 421-431, 1980 22. Long DM, Hartmann JF, French LA: The response of human cerebral edema to glucosteroid administration. Neurology (Minneap) 16: 521-528, 1966 23. Matsuoka Y, Hossmann KA: Corticosteroid therapy of experimental tumor oedema. Neurosurg Rev 4: 185-190, 1981 24. Yamada K, Ushio Y, Hayakawa T, Arita N, Yamada N, Mogami H: Effects of methylprednisolone on peritumoral brain edema. A quantitative autoradiographic study. J Neurosurg 59: 612-619, 1983 25. Nakagawa H, Groothuis DR, Owens ES, Fenstermacher JD, Patlak CS, Blasberg RG: Dexamethasone effects on [12SI]albumin distribution in experimental RG-2 gliomas and adjacent brain. J Cereb Blood Flow Metab 7: 687-701, 1987 26. Jarden JO, Dhawan V, Poltorak A, Posner JB, Rottenberg DA: Positron emission tomographic measurement of blood-to-brain and blood-to-tumor transport of SZRb: the

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

effect of dexamethasone and whole brain radiation therapy. Ann Neurol 18: 636-646, 1985 Yu ZY, Wrange O, Boethius J, Hatam A, Granholm L, Gustafsson JA: A study of glucocorticoid receptors in intracranial tumors. J Neurosurg 55: 757-760, 1981 Ohnishi T, Sher PB, Posner JB, Shapiro WR: Capillary permeability factor secreted by malignant brain tumors. J Neurosurg 72: 245-251, 1990 Quindlen EA, Bucher AP: Correlation of tumor plasminogen activator with peritumoral cerebral edema. A CT and biochemical study. J Neurosurg 66: 729-733, 1987 Roblin R, Young PL: Dexamethasone regulation of plasminogen activator in embryonic and tumor-derived human cells. Cancer Res 40: 2706-2713, 1980 Robert AM, Godeau G: Action of proteolytic and glycolytic enzymes on the permeability of the blood-brain barrier. Biomedicine 21: 36-39, 1974 Robert AM, Godeau G, Miskulin M, Moati F: Mechanism of action of collagenase on the permeability of the bloodbrain barrier. Neurochem Res 2: 449-455, 1977 Woolley DE, Tethow LC, Mooney CJ, Evanson JM: Human collagenase and its extracellular inhibitors in relation to tumor invasiveness. In: Strauli P, Barrett AJ, Baici A (eds) Proteinases and Tumor Invasion, Raven Press, New York, 1980, pp 97-115 Liotta LA, Abe S, Robey PG, Martin GR: Preferential digestion of basement membrane collagen by an enzyme derived from a metastatic murine tumor. Proc Natl Acad Sci USA 76: 2268-2272, 1979 Gazendam J, Houthoff HJ, Huitema S, Go KG: Cerebral edema formation and blood-brain barrier impairment by intraventricular collagenase infusion. In: Go JG, Baethmann A (eds) Recent Progress in the Study and Therapy of Brain Edema. Plenum Press, New York, 1984, pp 159-173 Bussolino F, Breviario F, Tetta C, Aglietta M, Mantovani A, Dejana E: Interleukin 1 stimulates platelet-activating factor production in cultured human endothelial cells. J Clin Invest 77: 2027-2033, 1986

Address for offprints: W.R. Shapiro, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, 350 W. Thomas Road, Phoenix AZ 85013, USA

Increased capillary permeability in rat brain induced by factors secreted by cultured C6 glioma cells: role in peritumoral brain edema.

To investigate whether brain tumors secrete a factor(s) responsible for peritumoral brain edema, we studied the effect of conditioned medium from cult...
983KB Sizes 0 Downloads 0 Views