Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

Journal of Neuro-Oncology 36: 285–291, 1998. 1998 Kluwer Academic Publishers. Printed in the Netherlands.

Clinical Study

Cerebral edema associated with meningiomas: the role of peritumoral braintissue

Rui Vaz,1 Nuno Borges, Celso Cruz,1 and Isabel AzevedoInstitute of Pharmacology and Therapeutics and 1Department of Neurology and Neurosurgery, MedicalFaculty, Hospital S. Joao, Porto, Portugal

Key words: blood-brain barrier, brain edema, meningiomas, peritumoral vessels

Summary

We undertook a morphological study of small pieces of peritumoral brain tissue removed from seven patientswith meningiomas submitted to surgery. All patients had cerebral edema, as shown by preoperative C.T. andN.M.R.. Control specimens were obtained from five patients undergoing ventriculo-peritoneal shunt. The tissuefragments were fixed in glutaraldehyde-osmium and embedded in Epon. In semi-thin sections observed underlight microscopy peritumoral endothelial cells exhibited voluminous cytoplasm and nucleus. Morphometricalevaluation confirmed that these endothelial cell nuclei were significantly larger than controls. Under the elec-tron microscope those cells showed nuclei rich in euchromatin and cytoplasm rich in pinocytotic vesicles.

The morphological changes observed suggest a process of dedifferentiation of brain peritumoral capillarycells and are compatible with an increase in permeability. Both events, which may be due to diffusion of atumoral vascular permeability factor, favour the hypothesis that peritumoral brain tissue contributes to ede-ma fluid that accumulates around meningiomas.

Introduction

Meningiomas are frequently associated with brainedema. However, the mechanisms underlying theformation of brain edema, in patients with meningio-mas, are far from being fully understood [1]. Manytumoral characteristics have been considered to ac-count for it but no clear-cut relationship was demon-strated between edema and tumor size, location, his-tology, degree of cellularity or vascularity, vascularcompression or sex hormone receptors [1–10].

Indeed the main and still unanswered questionabout this subject refers to the source of the edema.The role of tumoral neovessels and capillaries ofmeningioma peritumoral brain tissue in the forma-tion of edema has been considered but not yet clar-ified [5, 11, 12]. Recently we described a reversion ofphenotype of capillaries around glioblastomas to

that of peripheral vessels suggesting that those ves-sels may participate in the formation of brain ede-ma [13]. Therefore the main purpose of our studywas to investigate the putative involvement of per-itumoral capillary vessels in the production of brainedema in patients with meningiomas.

Materials and methods

Patient population, imaging and surgery

Biopsy specimens of peritumoral brain tissue wereobtained at operation from seven patients who un-derwent initial surgery for their tumors. AdequateC.T. and N.M.R. scans were used to assess size andlocation of the tumor and the degree of edema. Clin-ical data are summarized in Table 1. Patients were

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

286

Table 1. Clinical data of evaluated patients

Patient Age Sex Location Size Edema Histology

1 44 F convexity moderate moderate angioblastic2 59 M sphen.wing large moderate angioblastic3 43 F planum sphe large mild transitional4 41 F falx. large moderate transitional5 64 F sagit. sinus large mild transitional6 44 F convexity large mild transitional7 67 M convexity large severe meningotelial

Tumor size: using the greatest diameter; small < 2 cm; moderate = 2–4.5 cm; large > 4.5 cm.Cerebral edema: mild (just perifocal), moderate (more extensive than mild but less than severe) or severe (hemispheric spread of hypo-density) on the basis of the extension of diminished density area around the tumor, according to Gilbert (1983), Maiuri (1987) and Benzel(1988).

Figure 1. Light micrographies of sections of peritumoral white matter (A) and gray matter (B). * – edema fluid; arrows – extravasated redblood cells; E – endothelial cells. Bars = 10 µm.

operated between January 1993 and April 1994, se-lected from all those admitted during that periodwith meningiomas (17 patients). The only criterionfor inclusion in the study was the presence of edemaaround the tumor seen on M.R.I. Peritumoral edemawas graded as mild, moderate and severe independ-ently by two neuroradiologists of the Department onthe basis of the extension of the area of diminished

density around the tumor. Control specimens wereobtained from five patients undergoing ventriculo-peritoneal shunt for ‘communicating’ hydrocepha-lus. All patients were under dexamethasone treat-ment (4 mg four times daily) after C.T. scan and re-ceived no other medication. All operations were per-formed under total I.V. anesthesia with propofol +sufentanyl and controlled ventilation. There were no

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

287

Figure 2. Ultrastructural micrography of a section of peritumoral brain tissue (white matter). * – edema fluid; E – endothelial cells.Bar = 1 = µm.

per-operative abnormalities. All tumors were re-moved by microsurgical technique and a sample ofedematous brain tissue not contused by surgery wastaken in the tumor-brain junction where dissectionwas more difficult due to stronger adhesion. All pa-tients had uncomplicated post-operative courses andrecovered from neurological deficits present at ad-mission. The protocol had been previously approvedby the Ethics Committee of Hospital S. Joao.

Morphological study

Tissue samples were fixed by immersion in 3% glut-araldehyde in 0.1 mol/l cacodylate buffer, pH 7.3, at4 °C for 90 min. The tissue was then washed over-night in cacodylate buffer with 0.6 mol/l sucrose at4 °C, post-fixed in a 2% solution of osmium tetrox-ide in 0.1 mol/l cacodylate buffer pH 7.3, at 4 °C, for1 h, dehydrated with ethanol and propylene oxideand embedded in Epon 812 (Taab, Aldermaston,Berkshire, U.K.). Semi-thin sections (1 µm) wereobtained from five blocks taken at random fromeach case with a LKB Ultrotome, stained with 1%toluidine blue in 1% borax and examined by lightmicroscopy. Camera lucida drawings of all capillaryendothelial cell nuclei in each section were made

under a final amplification of × 860. The largest dia-meter of each nucleus and the axis perpendicular toit were measured on the paper drawings. These twovalues were then used to calculate the mean dia-meter (D) for each nucleus. All the morphometricaldeterminations were made by the same person inpreparations identified by a code number, themeaning of which was unknown to the observer un-til all measurements were concluded.

Thin sections were cut in a LKB Ultrotomestained with uranyl acetate and lead citrate and ob-served at the electron microscope (Jeol JEM 100Cx, Jeol, Tokyo) at 80 kV.

Statistics

The results are expressed as arithmetic means ±SEM. Student’s t test was applied to the differencebetween means (unpaired observations) and thedifference considered significant if p < 0.05.

Results

Although edema accumulated predominantly inwhite matter (Figure 1A and 2) we could also find

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

288

Figure 3. Mean diameter (D) of capillary endothelial cell nucleiin brain tissue from control and meningioma patients. * Statisti-cally higher than the control, p < 0.001 (n = 160 each).

Figure 4. Ultrastructural micrography of a section of peritumoral brain tissue (gray matter). E – endothelial cells; N – nucleus rich ineuchromatin (Eu) and nucleoli (arrows); arrowheads – tight junctions. Bar = 1 µm.

some increase in the extracellular fluid within thecortex (Figure 1B). Extravasated red blood cellswere occasionally observed (Figure 1).

The alterations seen in endothelial cells of gray

and white matter microvessels were similar. Insemi-thin sections observed under light microscopyperitumoral capillary endothelial cells exhibitedvoluminous cytoplasm and nucleus (Figure 1). Mor-phometrical evaluation confirmed that capillary en-dothelial cell nuclei from meningioma patientswere significantly larger than controls – mean dia-meter 5.5 ± 0.2 µm vs 4.3 ± 0.1 µm in controls, n =160 each, p < 0.001 (Figure 3).

Under the electron microscope, capillary endo-thelial cells from meningioma patients differedfrom controls mainly by showing nuclei rich in eu-chromatin and nucleoli (Figure 4) and cytoplasmrich in pinocytotic vesicles (Figure 5). Tight junc-tions were intact (Figure 4 and 5). Nervous cells andfibers appeared reasonably well preserved, where-as podocytes accumulated edema fluid and were oc-casionally disrupted.

Discussion

It is well known that edema associated with menin-giomas is of the vasogenic type, i.e., due to an in-crease in blood-brain barrier permeability [5, 14–17]. The entities responsible for the production ofedema can be the newly formed microvessels asso-

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

289

Figure 5. Ultrastructural micrography of a section of peritumoral brain tissue (gray matter). E – endothelial cells; arrows – pinocytoticvesicles; arrowheads – tight junction. Bar = 1 µm.

ciated with tumoral angiogenesis or the previouslynormal cerebral capillaries altered by some mean.

The common observation of absence of contrastenhancement around meningiomas on C.T. andN.M.R. scans [1, 17, 18] associated with the ultra-structural abnormalities (gap junctions, fenestratedendothelium, abundant pinocytotic vesicles) pre-sent in tumor capillaries [3, 5, 19–22] led to the pro-posal that these vessels were responsible for theedema. However meningiomas arise extracerebral-ly and are separated from the white matter by lep-tomeninges and cortex [10, 12, 15, 16]. So there wasthe question of how fluid originating within the tu-mor could reach the white matter. The proposed ex-planation was that growing tumor damaged lepto-meninges and cortex establishing continuity be-tween tumor and white matter and therefore allow-ing direct transmission of edema fluid [1, 10, 12, 20,22].

Anyhow, mechanical factors resulting from pres-sure of tumor on surrounding tissue are not enoughto explain edema in the majority of cases [3, 6, 10, 11,20, 23], specially in small meningiomas often associ-ated with marked edema [7, 19, 24]. Moreover, Con-stantini et al. [11] demonstrated a complete lack ofcorrelation between peritumoral edema and tumorwater content. All these facts led several authors to

propose the involvement of peritumoral brain tis-sue in the production of edema [5, 11, 23, 25]: me-ningioma cells could release chemical mediators ca-pable of promoting blood-brain barrier lesions andperhaps other alterations in microvasculature [3, 9,22, 24, 26]. We recently verified that capillariesaround glioblastomas present morphologicalchanges compatible with a reversion of phenotypeand increase in permeability [13].

Therefore we decided to evaluate the tumor-brain interface in patients with meningiomas and toinvestigate if abnormalities of the blood-brain bar-rier were also present in peritumoral brain tissue.The diminished density on C.T. and M.R.I. corre-sponded in all cases to edema formation, no cystsbeing found around the tumor. Sample location wasselected from areas where edema was more evidenton C.T. and M.R.I. The morphological changes ob-served and here described were the same in biopsiesfrom all cases, regardless of localization and hist-ological type of the tumor, or degree of edema.

Just a few previous studies evaluated peritumoralbrain tissue from patients with meningiomas. Ste-vens et al. [10], in four cases, and Philippon et al.[22], in eight, confirmed extracellular fluid accumu-lation, similarly to what we found. Ide et al. [20], infourteen cases, described gliotic changes with loss

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

290

of neurons, edema and cystic changes in adheredbrain tissue, but they didn’t identify the tissue asgray or white matter. Ohno et al. [12], using lightmicroscopy, described cortical thinning or com-plete loss of the cortex under the tumor, whereasGilbert et al. [5], in a single case, using electron mi-croscopy, found edema in the white matter butsparing the cortex and suggested that fluid origi-nated within the white matter. Some of our biop-sies of adjacent brain tissue also included gray mat-ter, where we could also see edema, although lessabundant than in the white matter. Moreover, Gil-bert et al. [5] were the only authors to study endo-thelial cells, concluding that endothelial cells werenormal except for the existence of abundant pino-cytotic vesicles. Beyond the increase in the densityof pinocytotic vesicles, our results provide evi-dence for the existence of other changes in peri-tumoral endothelial cells. The larger nuclei rich ineuchromatin and nucleoli indicate an increase inthe synthetic activity of the cells. All these changesare compatible with a process of dedifferentiationof peritumoral endothelial cells in patients withmeningiomas and brain edema, quite similarly towhat has been previously described in patientswith glioblastomas [13].

This dedifferentiation, i.e., the acquisition of aperipheral vessel phenotype, is compatible with theincrease in permeability due to leukotrienes [27, 28]and with the decrease in permeability to corticoste-roids [29]. As a matter of fact, neither leukotrienesseem to be able to increase normal blood-brain bar-rier permeability [30, 31] nor are corticosteroids ef-fective in reducing that permeability in infarction ortrauma, situations where dedifferentiation of endo-thelial cells has not been described [32].

So our results suggest that peritumoral microves-sels may contribute to the formation of cerebraledema around meningiomas. The presence ofmorphological changes in capillary cells which arenot in direct contact with tumor cells may be due tothe action of one or several diffusible factors capa-ble of acting at some distance from the tumor mass[11, 14].

Conclusions

Peritumoral capillary vessels are most probably in-volved in the production of edema around menin-giomas. The morphological changes of peritumoralcapillary endothelial cells are compatible with anincrease in permeability and with dedifferentiation,i.e., reversion of phenotype of these cells. Bothevents may be due to diffusion of a tumoral perme-ability factor.

Acknowledgements

The authors gratefully acknowledge the technicalassistance of M. Luisa Vasques. Supported byPRAXIS XXI (PSAU/P/SAU/53/96) do Ministérioda Ciéncia e Tecnologia.

References

1. Go KG, Wilmink JT, Molenaar WM: Peritumoral brain ede-ma associated with meningiomas. Neurosurgery 23: 175–179,1988

2. Benzel EC, Gelder F: Correlation between sex hormonesbinding and peritumoral edema in intracranial meningio-mas. Neurosurgery 22(2): 169–174, 1988

3. Bradac BD, Ferszt R, Bender A, Schorner W: Peritumoraledema in meningiomas – a radiological and histologicalstudy. Neuroradiology 28: 304–312, 1986

4. Brandis A. Mirzai S, Tatagiba M, Walter GF, Samii M,Ostertag H: Immunohistochemical detection of female sexhormone receptors in meningiomas: correlation with clin-ical and histological features. Neurosurgery 33(2): 212–218,1993

5. Gilbert JJ, Paulseth JE, Coates RK, Malott D: Cerebral ede-ma associated with meningiomas. Neurosurgery 12(6): 599–605, 1983

6. Inamura T, Nishio S, Takeshira I, Fujiwara S, Fukui M: Peri-tumoral brain edema in meningiomas – Influence of vascu-lar supply on its development. Neurosurgery 31(2): 179–185,1992

7. Maiuri F, Gangemi M, Cirillo S, Delehaye L, Gallicchio B,Carendente M, Giamundo A: Cerebral edema associatedwith meningiomas. Surg Neurol 27: 64–68, 1987

8. Meixensberg J, Caffier H, Naumann M, Hofmann E: Sexhormone binding and peritumoral oedema in meningiomas:is there a correlation? Acta Neurochir (Wien) 115: 98–102,1992

Please indicate author’s corrections in blue, setting errors in red

141428 NEON ART.NO 796-96 (708) ORD.NO 234708.Z

291

9. Smith HP, Challa VR, Moody DM, Kelly DL: Biological fea-tures of meningiomas that determine the production of ce-rebral edema. Neurosurgery 8(4): 428–433, 1981

10. Stevens JM, Ruiz JS, Kendall BE: Observations on peri-tumoral oedema in meningioma. Part II: Mechanism ofoedema production. Neuroradiology 25: 125–131, 1983

11. Constantini S, Tamir J, Gomori M, Shohami E: Tumorprostaglandin levels correlate with edema around supra-tentorial meningiomas. Neurosurgery 33(2): 204–211, 1993

12. Ohno K, Matsushima Y, Aoyagi M, Ikeda J, Suzuki R,Ichimura K, Tamaki M, Hirakawa K: Peritumoral cerebraledema in meningiomas: the role of tumor brain interface.Clin Neurol Neurosurg 94: 291–295, 1992

13. Vaz R, Borges N, Sarmento A, Azevedo I: Reversion of phe-notype of endothelial cells in brain tissue around glioblasto-mas. J Neuro-Oncology 27: 127–132, 1996

14. Del Maestro, Megyesi JF, Farrell CL: Mechanism of tumorassociated edema: a review. Can J Neurol Sci 17: 177–183,1990

15. Go KG, Kamman RL, Wilmink JT, Mooyaart EL: A studyon peritumoral brain edema around meningiomas by CTand MRI scanning. Acta Neurochir (Wien) 125: 41–46, 1993

16. Lindley JG, Challa VR, Kelly Jr DL: Meningioma and brainedema, in Al-Mefty (ed): Meningiomas. New York: RavenPress 1991, pp 59–73

17. Stevens JM, Ruiz JS, Kendall BE: Observations on peri-tumoral oedema in meningioma. Part I: distribution, spreadand resolution of vasogenic oedema seen on computed to-mography. Neuroradiology 25: 71–80, 1983

18. Trittmacher S, Traupe H, Schmid A: Pre and post operativechanges in brain tissue surrounding a meningioma. Neuro-surgery 22(5): 882–885, 1988

19. Challa VR, Moody DM, Marshall RB, Kelly Jr DL: The vas-cular component in meningioma associated with severe ce-rebral edema. Neurosurgery 7: 363–368, 1980

20. Ide M, Jimbo M, Kubo O, Yamamoto M, Imanaga H: Peri-tumoral brain edema associated with meningioma – histo-logical study of the tumor margin and surrounding brain.Neurol Med Chir (Tokyo) 32: 65–71, 1992

21. Long DM: Vascular ultrastructure in human meningiomaand schwannomas. J Neurosurg 38: 409–419, 1973

22. Philippon J, Foncin JF, Grob R, Srour A, Poisson M, Pertui-set BF: Cerebral edema associated with meningioma: pos-sible role of a secretory-excretory phenomenon. Neurosur-gery 14(3): 295–301, 1984

23. Vries J, Wakhloo AK: Cerebral oedema associated withWHO-I, WHO-II and WHO-III meningiomas: correlationof clinical, computed tomographic, operative and histolog-ical findings. Acta Neurochir (Wien) 125: 34–40, 1993

24. Quindlen EA, Bucher AP: Correlation of tumor plasmino-gen activator with peritumoral cerebral edema: J Neurosurg66: 729–733, 1987

25. Fine M, Brazis P, Palacios E, Neri G: Computed tomographyof sphenoid wing meningiomas: tumor location related todistal edema. Surg Neurol 13: 385–390, 1980

26. Cooper C, Jones HG, Weller RO, Walker V: Production ofprostaglandins and thromboxane by isolated cells from in-tracranial tumors. J Neurol Neurosurg Psych 47: 579–584,1984

27. Baba T, Chio C-C, Black KL: The effect of 5-lipoxygenaseinhibition on blood-brain barrier permeability in experi-mental brain tumors. J Neurosurg 77: 403–406, 1992

28. Black KL, Hoff JT, McGillicudy JE, Gebarski SS: Increasedleukotrine C4 and vasogenic edema surrounding brain tu-mors in humans. Ann Neurol 19: 592–595, 1986

29. Haynes Jr RC: Adrenocorticotropic hormone; adrenocorti-cal steroids and their synthetic analogs; inhibitors of the syn-thesis and actions of adrenocortical hormones. In: Good-man Gilman A, Rall TW, Nies AS, Taylor P (eds) The phar-macological basis of therapeutics, Maxwell Macmillan In-ternational-Editors, 8th ed 1991, pp 1431–1462

30. Unterberg A, Schmidt W, Wahl M, Baethmann A: Role ofleukotrienes as mediator compounds in brain edema. Adv inNeurol 52: 211–214, 1990

31. Unterberg A, Schmidt W, Wahl M, Ellis EF, Marmarou A,Baethmann A: Evidence against leukotrienes as mediatorsof brain edema. J Neurosurg 74: 773–780, 1991

32. Cervos-Navarro J, Lafuente JV: Traumatic brain injuries:structural changes. J Neurol Sci 103: S3–S14, 1991

Address for offprints: Isabel Azevedo, Institute of Pharmacologyand Therapeutics, Faculty of Medicine, 4200 Porto, Portugal