Effects of methylprednisolone on peritumoral brain edema

J Neurosurg 59:612-619, 1983 Effects of methylprednisolone on peritumoral brain edema A quantitative autoradiographic study KAZUO YAMADA, M.D., YUKITAKA USHIO, M.D., TORU HAYAKAWA, M.D., NORIO ARITA, M.D., NORIKO YAMADA, B.S., AND HEITARO MOGAMI, M.D. Department of Neurosurgery, Osaka University Medical School, Osaka, Japan Peritumoral brain edema was produced by intracerebral transplantation of Walker 256 tumor in rats. Local cerebral blood flow (LCBF), local cerebral glucose utilization (LCGU), and capillary permeability were studied in untreated and methylprednisolone-treated rats by quantitative autoradiography. In the untreated group, LCBF and LCGU were widely depressed in the cortex and deep structures of the hemisphere ipsilateral to the tumor. In the methylprednisolone-treated animals, LCBF and LCGU were significantly better than in the untreated animals. Capillary permeability was highly increased in the viable part of the tumor in the untreated animals. In the methylprednisolone-treated group, capillary permeability of the tumor was significantly lower than that in the untreated group. These results may suggest that increase in capillary permeability of the tumor is the major source for edema fluid production, and that methylprednisolone improves brain edema by decreasing capillary permeability of the tumor. Decrease in edema fluid formation may result in restoration of blood flow and glucose metabolism in the adjacent brain tissue, and may improve clinical symptoms and signs. KEY WORDS 9 brain edema 9 experimental tumor 9 methylprednisolone 9 steroid 9 local cerebral blood flow 9 glucose metabolism 9 blood-brain barrier G LUCOCORTICOIDS are known to improve peritumoral brain edema. Several previous reports have discussed the mechanism of steroid effects on brain edema induced by brain tumors. 6'~3a7'18"21 The exact mechanism still remains uncertain, however. In the present report, local cerebral flood flow (LCBF), local cerebral glucose utilization (LCGU), and capillary permeability in the rat brain-tumor model were studied by quantitative autoradiography. With this technique~ regional changes of these physiological parameters were detected precisely, with correlation to the histological sections. Moreover, effects of methylprednisolone on these physiological parameters were analyzed to investigate a possible role of steriods on resolution of brain edema. Tumor Transplantation Materials and Methods Walker 256 tumor has been maintained in our laboratory by serial subcutaneous transplantation. The tumor was removed aseptically, and a 1-cu mm pellet was made with scissors in Earle's basic medium. Adult female Wistar rats were anesthetized with ketamine hydrochloride and placed in a stereotaxic head holder. A burr hole was made with a dental drill on the left parietal bone 2 mm behind the coronal suture and 4 mm laterally from the midline. The dura and surface of the small cortex were cauterized with a bipolar cautery. A 1-cu mm pellet of the tumor was transplanted to the subcortex. The burr hole was filled with adhesive (Aron-alpha*) and the scalp was closed. These rats were randomly divided into two groups. One group of animals was treated with methylprednisolone succinate (15 mg/kg intraperitoneauy, once a day) from the 5th to the 10th day after tumor inoculation. The other group received no treatment and served as a control. Each group consisted of 15 tumor-beating rats. These were studied autoradiographically on the 10th day after tumor inoculation. Five animals from each of these two groups and a normal control animal were placed into three separate groups for the measurement of LCBF, LCGU, and capillary permeability, respectively. * Adhesive manufactured by Toa-Gosei Chemical Co. Ltd., Tokyo, Japan. 612 J. Neurosurg. / Volume 59 / October. 1983

Methylprednisolone and peritumoral brain edema Experimental Procedure Local CBF, LCGU, and capillary permeability were measured by methods similar to those described in previous reports. 19'2~ Under halothane anesthesia, the right femoral artery and vein were catheterized, and the animals were immobilized in a physiological position by a loosely fitting plaster cast below the thorax, including their hind quarters. They were allowed to recover from anesthesia for at least 1 hour. Arterial blood gases and blood pressure were measured and found to be within normal range. For measurements of LCBF, LCGU, and capillary permeability, carbon-14 (~4C)-iodoantipyrine (75 ~Ci/ kg), ~4C-deoxyglucose (100 #Ci/kg), and ~4C-alpha-aminoisobutyric acid (AIB: 100 t~ci/kg) were used as the tracer substances, respectively. Each tracer was injected into a femoral vein at a constant rate for 1 minute. For LCBF measurement, arterial blood samplings were made every 10 seconds, and animals were sacrificed at 1 minute after starting the injection. For LCGU measurement, blood samplings were made every 1 to 10 minutes, and animals were sacrificed at 45 minutes after injection. Capillary permeability was measured by sampling blood every 1 minute, and animals were sacrificed at 10 minutes after injection. Blood levels of tracers were measured with a liquid scintillation counter. For LCGU measurements, the blood glucose level was measured by the glucose oxidase method. Each brain was removed and cut 60 um thick in a cryostat.t Brain sections and a set of ~4C-methylmethacrylate standard plates were attached to the Kodak x-ray film (SB-5). These standards were precalibrated for their autoradiographic equivalence to the ~4C concentration of brain sections (60 um thick). After 2 weeks of exposure, x-ray films were developed, and the densitometric measurement of the autoradiographs was made with a densitometer.~ A calibration curve between optical density and tissue 14C concentration for each film was obtained by measuring optical density of the standard plates. The representative brain sections were stained with hematoxylin and eosin for histological observation. The size of the tumor used for autoradiography was measured by the average-end-area method as described previously. ~6 Calculation of LCBF, LCGU. and Capillary Permeability Local CBF was calculated according to the equation developed by Kety 5 and applied to the rat using laciodoantipyrine by Sakurada, et al. ~4 Measurement of LCGU was made by the equation developed by Sokoloft, et al. ~5 However, lumped constant and rate constant in the brain-tumor tissue remain unknown. To t Cryostat manufactured by South London Equipment Co. Ltd., London, England. Densitometer (Sakura PDA-15) manufactured by Konishiroku Co. Ltd., Tokyo, Japan. circumvent this problem, relative glucose utilization (RGU) was used as originally described by Blasberg, et al.' The equation is the following: RGU = Ci(T) fg C*p/Cp(t) dt' where Ci(T) is the ~4C activity in the tumor at time T (45 minutes for this experiment), and C*p and Cp are the arterial plasma concentrations of ~4C-deoxyglucose and glucose, respectively. Capillary permeability was expressed as a unidirectional blood-to-brain transfer constant (Ki), which can be calculated according to the equation developed by Ohno, et al. 8 A desk-top computer was used for calculation of these physiological parameters. Results Within 10 days after tumor inoculation, the tumor grew to more than 5 mm in diameter. Shift of the midline structures and peritumoral brain edema were evident in the lesioned hemisphere. The tumor could be divided into three parts: the central necrosis; the viable part: and the periphery. Histological examination revealed numerous capillary vessels in the viable part of the tumor. No significant differences in the size of the tumor used for autoradiography were noted between methylprednisolone-treated animals and untreated controls (Table 1). Local Cerebral Blood Flow In untreated animals, LCBF was significantly reduced throughout the cortex of the lesioned hemisphere. The amount of reduction was about 50% of that in the normal animals on average (Table 2). The maximum reduction was noted in the auditory cortex which was close to the tumor (58% reduction). Significant decrease in LCBF was also evident in the caudateputamen, thalamus, and corpus callosum of the lesioned hemisphere. Moreover, reduction of LCBF to some extent was noted in the cortex and deep structures of the contralateral hemisphere. In the methylprednisolone-treated group, LCBF reduction was less significant than that in untreated animals (Table 2). When TABLE 1 (~rrected volumes of the tumors used for autoradiography Study Corrected Volume (cu mm)* Differ- Untreated Treated ence LCBF measurement LCGU measurement 161.6 _+ 17.5 129.6 _+ 41.4 138.9 + 57.0 186.5 _+ 54.9 NS NS capillary permeability 130.0 + 52.0 129.7 + 34.2 NS measurement * Mean values _+ standard deviation of results in five rats in each group. "Treated" rats were treated with methylprednisolone. NS = no statistically significant difference as compared to untreated animals by t-test. LCBF = local cerebral blood flow: LCGU = local cerebral glucose utilization. J. Neurosurg. / Volume 59 / October, 1983 63 3

K. Yamada, et al. FIG. 1. Carbon-14 iodoantipyrine autoradiographs of the untreated (left) and methylprednisolone-treated (right) animals. Structure TABLE 2 Effects of methylprednisolone on local cerebral blood flow* Ipsilateral Hemisphere Contralateral Hemisphere Untreated Treated Untreated Treated Normal Control olfactory cortex 112 + 18"* 120 + 18w 144 + 24 152 + 7 150 9 frontal cortex 75 + 14"* 122 _+ 18**tt 148 + 17 154 _+ 6 159 + 8 sensorimotor cortex 99 13"* 161 + 35w 185 _+ 13 205 35 205 + 12 parietal cortex 85 + 32** 153 + 10**tt 168 8w 182 + 14 t 183 9 auditory cortex 97 + 19"* 178 +_ 13**it 186 13"* 225 19t-t 230 8 visual cortex 101 + 24** 169 + 18w 171 + 22 188 + 18 190 7 caudate-putamen 77 28** 88 + 28w 108 _+ 16w 118 18 130 _+ 10 thalamus(medial) 82 + 5** 113 _.+ 1 l**-tt 125 + 8w 130 + 8 135 _ 8 hypothalamus 80 + 6 86 + 7 86 _+ 6 87 6 87 + 5 corpus callosum 33 _+ 7** 38 + 4** 46 _+ 7 48 _ 3 47 + 4 * Values are ml/100 gm/min, mean standard deviation of results of five rats per group. "Treated" rats were treated with methylprednisolone. Statistical significance by t-test, different from normal controls at: **p < 0.01; w < 0.05; different from corresponding untreated group at: ttp< 0.01; tp < 0.05. 614 J. Neurosurg. / Volume 59 / October, 1983

Methylprednisolone and peritumoral brain edema FIG. 2. Macroscopic appearance of the untreated (left) and methylprednisolone-treated (right) animals. the untreated and methylprednisolone-treated groups were compared, LCBF was significantly better throughout the cortex and thalamus of the lesioned hemisphere in the latter group (p < 0.01). Some improvement in LCBF was also noted in the parietal and auditory cortex of the contralateral hemisphere. Representative autoradiographs and macroscopic appearance are shown in Figs. 1 and 2. In the tumor tissue, blood flow was different depending upon the part of the tumor. The maximum blood flow was noted in the viable part of the tumor, whereas the central necrosis area and periphery of the tumor had low blood flow (Fig. 3). No significant differences in the tumor blood flow were noted between untreated and methyl prednisolone-treated animals (Fig. 3). Local Cerebral Glucose Utilization Untreated animals showed significantly lower LCGU values than normal controls throughout the cortex of the both hemispheres (Table 3). Significant reduction was also evident in the deep structures of the lesioned hemisphere such as caudate-putamen, thalamus, hypothalamus, and corpus callosum. Maximum reduction was noted in the caudate-putamen complex of the lesioned hemisphere, where LCGU was only 44% of the normal value. Methylprednisolone-treated animals showed significantly better LCGU values than untreated animals in the cortex and deep structures of both hemispheres (Table 3). Relative glucose utilization was maximum in the viable part of the tumor (Fig. 3). The periphery had lower RGU values, and there was almost no RGU activity in the central necrosis. Methylprednisolone reduced RGU significantly only in the viable part of the tumor. Representative autoradiographs of untreated and methylprednisolone-treated animals were shown in Fig. 4 and the macroscopic appearance in Fig. 5. Capillary Permeability Capillary permeability of the untreated animals was markedly increased in the viable part and periphery of the tumor (Figs. 6, 7, and 8). Brain tissue immediately FIG. 3. Local blood flow and relative glucose utilization of intracerebrally implanted Walker 256 tumor. The viable part of the tumor had high blood flow and glucose utilization, whereas the necrotic center had almost no activity. No differences were noted between the untreated and methylprednisolone-treated groups except for the relative glucose utilization in the viable part of the tumor. adjacent to the tumor also showed some increase in capillary permeability. Due to methylprednisolone treatment, about 40% reduction of capillary permeability was noted in the viable part and periphery of the tumor (Fig. 6). However, no effect of methylprednisolone was noted on capillary permeability of the intraventricular tumor (Figs. 7 and 8). Although the adjacent brain tissue also showed significant reduction of capillary permeability, the amount of reduction was much less than that in the tumor tissue. No effects of steroids J. Neurosurg. / Volume 59 / October, 1983 615

K. Yamada, et al. FIG. 4. Carbon-14 deoxyglucose autoradiographs of untreated (left) and methylprednisolone-treated (right) animals. Structure TABLE 3 Effects of methylprednisolone on local cerebral glucose utilization* Ipsilateral Hemisphere Contralateral Hemisphere Untreated Treated Untreated Treated Normal Control olfactory cortex 85 --- 9** 120 23it 111 12w 128 + 21 126 _+ 8 frontal cortex 72 + 17"* 104 + 16w 119 6w 125 8 127 7 sensorimotor cortex 73 + 12"* 101 _+ 8**it 120 7w 129 _+ 7t 130 8 parietal cortex 86 + 12"* 104 _+ 8**t 115 _ 4** 129 + 7**tt 148 _ 7 auditory cortex 88 + 8** 105 + 7**tt 128 _ 7** 138 + 5w 148 + 7 visual cortex 103 8** 131 8it 129 _+ 9w 143 + lit 140 + 8 caudate-putamen 49 _+ 5** 96 8w 101 _+ 8w 111 + 8? 112 + 10 thalamus (medial) 77 + 18"* 92 9w 102 + 13 106 + 6 106 + 7 hypothalamus 76 10w 93 + 8?? 85 + 10 95 11 96 + 9 corpus callosum 28 + 4** 41 7?? 35 + 4w 45 6?? 44 + 5 * Values are #mol/100 gm/min, mean standard deviation of results of five rats per group. "Treated" rats were treated with methylprednisolone. Statistical significance by t-test, different from normal control at: **p < 0.01; w < 0.05; different from corresponding untreated group at: ttp< 0.01; tp < 0.05. 616 J. Neurosurg. / Volume 59 / October, 1983

Methylprednisolone and peritumoral brain edema FIG. 5. Macroscopic appearance of the untreated (lej~) and methylprednisolone-treated (right) animals. on capillary permeability was noted in the distant "normal" brain tissue (Fig. 6). Discussion Although several previous reports have discussed LCBF and LCGU in the edematous brain tissue caused by various lesions, 34'~~ only a few authors have commented on these physiological parameters in peritumoral brain edema. 3~4 Moreover, nothing has been published about the effects of steroids on LCBF and LCGU of the peritumoral brain tissue, in spite of its definite clinical effects on brain tumor patients. In the present study, we tried to analyze the changes of blood flow and metabolism in the peritumoral brain tissue. As we indicated, LCGU and LCBF were widely depressed in the hemisphere ipsilateral to the tumor. This may be the result of widespread depression of the functional state of the brain induced by peritumoral brain edema. In cold lesion, similar widespread depression of LCGU in the lesioned hemisphere was reported by Pappius.l~ However, LCBF was not changed significantly in the cold injury, indicating that intensity and nature of the edema are somewhat different from pentumoral edema. In the tumor-induced brain edema, the edema fluid is thought to be derived from extravasation of the plasma through leaky tumor vessels, and plasma may extend to the adjacent brain tissue. 2 This view has been clearly proven by the autoradiographic observation of capillary permeability. The viable part and periphery of the tumor showed marked increase in capillary permeability, whereas the adjacent brain tissue was less permeable. This strongly suggests the view that the source of edema fluid may be derived from ultrafiltration of the plasma in the tumor as indicated above. In the sensorimotor, parietal, and auditory cortices close to the tumor, maximum reduction of LCBF and LCGU was noted. This may be partly due to direct mechanical compression and chemical effects of the tumor, and partly due to accumulation of the edema fluid in the vicinity of the tumor. FiG. 6. Capillary permeability of the tumor and adjacent brain tissue in the untreated and methylprednisolone-treated animals. In the untreated animals, the viable part of the tumor had a maximum increase in capillary permeability, and the periphery of the tumor and adjacent brain tissue also showed some increase in capillary permeability. In the methylprednisolone-treated animals, capillary permeability was significantly lower in the viable part and periphery" of the tumor, and brain adjacent to the tumor. It may be worth noting that in the peritumoral brain, edema reduction of LCBF was more prominent than that of LCGU. Pappius, et al., ~~ have reported similar results in the edematous state following osmotic opening of the blood-brain barrier. In their experiment, LCBF was reduced about 60%, whereas LCGU only decreased 30%. These facts may suggest the similarity of the edematous state in the peritumoral tissue and osmotic opening of the blood-brain barrier. In those J. Neurosurg. / Volume 59 / October, 1983 617

K. Yamada, et al. FIG. 7. Autoradiographs after injection of t4c-alpha-aminoisobutyric acid indicating regional capillary permeability in untreated (left) and methylprednisolone4reated (right) animals. Methytprednisolone reduced capillary permeability of the intraparenchymal tumors but had no effects on that of intraventricular tumors. FIG. 8. Macroscopic appearance of the untreated (left) and methylprednisolone-treated (right) animals. 618 J, Neurosurg. / Volume 59 / October, 1983

Methylprednisolone and peritumoral brain edema states of edema, LCBF was more susceptible to change than LCGU. This is completely different from cold lesions in which LCGU was more susceptiblej 1 There are no data for a clear explanation of the uncoupling between CBF and metabolism. Experimental and clinical studies have indicated that peritumoral brain edema could be reduced by steroid treatment. 7~8 Improvement of clinical symptoms and signs might result from reduction of edema fluid formation in the leaky tumor vessels. 2 In our previous study, as we observed that water and sodium contents in the peritumoral brain tissue were much reduced by steroid treatment, and we suggested that steroids might reduce edema fluid formation in the tumor tissue. The present study has further supported this view, and clearly indicated that methylprednisolone in high doses reduces capillary permeability in the tumor and adjacent brain tissue. The exact mechanism by which methylprednisolone reduces capillary permeability remains uncertain. Recently, Yu, et al., 2~ reported that steroid receptors are present in primary and metastatic brain tumors. They suggested that the first step for an anti-edema effect of steroids is to form a steroid-receptor complex in the tumor tissues. This complex may have some effects on capillary permeability of the tumor tissue. Anti-inflammatory effects of steroids may also stabilize capillary vessels and may reduce capillary permeability. 9 Since methylprednisolone reduces formation of edema fluid in tumor tissue, the amount of edema fluid in the adjacent brain tissue may diminish and tissue volume subsequently decrease. This may reduce mass effect and improve microcirculation of the edematous brain tissue, resulting in the improvement of LCBF. Restoration of microcirculation may improve glucose metabolism of the neurons and glial cells, and may improve LCGU. This restoration in blood flow and glucose metabolism may closely correlate with improvement of clinical symptoms and signs in patients with brain tumors. Acknowledgments We thank Mr. K. Sakakibara for his help in the radioactive assay and Miss R. Fujita for her secretarial assistance. References 1. Blasberg RG, Groothuis D, Molnar P: Application of quantitative autoradiographic measurements in experimental brain tumor models. Semin Neurol 1:203-221, 1981 2. Ehrenkranz JRL, Posner JB: Adrenocorticosteroid hormones, in Weiss L, Gilbert HA, Posner JB (eds) Brain Metastasis. Boston: GK Hall, 1980, pp 340-363 3. Hossmann KA, Bloink M: Blood flow and regulation of blood flow in experimental peritumoral edema. Stroke 12:211-217, 1981 4. Hossmann KA, Niebuhr I, Tamura M: Blood flow and metabolism in the rat brain during experimental tumor development. Aeta Neurol Stand 60 (Suppl 72):576-577, 1979 5. Kety SS: Measurement of local flow by the exchange of an inert, diffusible substance. Methods Med Res 8: 228-236, 1960 6. Long DM, Hartmann TF, French LA: The response of human cerebral oedema to glucosteroid administration: an electron microscopic study. Neurology 16:521-528, 1966 7. Matsuoka Y, Hossmann KA: Corticosteroid therapy of experimental tumour oedema. Neurosurg Rev 4:185-190, 1981 8. Ohno K, Pettigrew KD, Rapoport SI: Lower limits of cerebrovascular permeability to nonelectrolytes in the conscious rat. Am J Physioi 235:H299-H307, 1978 9. Ortega BD, Demopoulos HB, Ransohoff J: Effects of antioxidants on experimental cold-induced cerebral edema, in Reulen H J, Schi~rmann K (eds): Steroids and Brain Edema. Berlin/Heidelberg/New York: Springer- Verlag, 1972, pp 167-176 10. Pappius HM: Dexamethasone and local cerebral glucose utilization in freeze-traumatized rat brain. Ann Neurol 12:157-162, 1982 11. Pappius HM: Local cerebral glucose utilization in thermally traumatized rat brain. Ann Neuroi 9:484-491, 1981 12. Pappius HM, Savaki HE, Fieschi C, et al: Osmotic opening of the blood-brain barrier and local cerebral glucose utilization. Ann Neurol 5:211-219, 1979 13. Reulen H J, Hadjidimos A, Sch/irmann K: The effect of dexamethasone on water and electrolyte content and on rcbf in perifocal brain edema in man, in Reulen H J, Schtirmann K (eds): Steroids and Brain Edema. Berlin/Heidelberg/New York: Springer-Verlag, 1972, pp 239-252 14. Sakurada O, Kennedy C, Jehle J, et al: Measurement of local cerebral blood flow with iodo[14c]antipyrine. Am J Physiol 3:H59-H66, 1978 15. Sokoloff L, Reivich M, Kennedy C, et al: The [14C]- deoxyglucose method for the measurement of local cerebral glucose utilization. Theory, procedure and normal values in the conscious and anesthetized albino rat. J Neurochem 28:897-916, 1977 16. Sundt TM Jr, Waltz AG: Experimental cerebral infarction: retro-orbital, extradural approach for occluding the middle cerebral artery. Mayo Clin Proc 41:159-168, 1966 17. Weinstein JD, Toy FJ, Jaffe ME, et al: The effect of dexamethasone on brain edema in patients with metastatic brain tumors. Neurology 23:121-129, 1973 18. Yamada K, Bremer AM, West CR: Effects of dexamethasone on tumor-induced brain edema and its distribution in the brain of monkeys. J Neurosurg 50:361-367, 1979 19. Yamada K, Hayakawa T, Ushio Y, et al: Regional blood flow and capillary permeability in the ethylnitrosoureainduced rat glioma. J Neurosurg 55:922-928, 1981 20. Yamada K, Ushio Y, Hayakawa T, et al: Quantitative autoradiographic measurements of blood-brain barrier permeability in the rat glioma model. J Neurosurg 57: 394-398, 1982 21. Yu ZY, Wrange O, Borthius J, et al: A study of glucocorticoid receptors in intracranial tumors. J Neurosurg 55: 757-760, 1981 Manuscript received January, 19, 1983. Accepted in final form April 21, 1983. This work was supported in part by a grant-in-aid for fundamental scientific research from the Ministry of Education and for cancer research from the Ministry of Health and Welfare. Address reprint requests to: Kazuo Yamada, M.D., Department of Neurosurgery, Osaka University Medical School, 1-1-50 Fukushima, Fukushima, Osaka 553, Japan. J. Neurosurg. / Volume 59 / October, 1983 619