Ependymoma

* We report an unusual case of ependymoma with pigmentation, a phenomenon that has only been described in a few cases, to our knowledge. This tumor occurred in the fourth ventricle of a 45-year-old man. It showed the typical histologic appearance of ependymoma with perivascular pseudorosettes and rare ependymal rosettes. Some tumor cells contained brown cytoplasmic pigment, which was shown histochemically to represent a mixture of lipofuscin and neuromelanin. The pigment was positive for acid-fast and periodic acid-Schiff stains and was also focally positive for Masson-Fontana and Schmorl stains (bleached by pretreatment with potassium permanganate). In addition, some other tumor cells showed a signet ring morphology as a result of prominent intracytoplasmic vacuolation. Immunohistochemically, all the tumor cells expressed glial fibrillary acidic protein, and rare pigmented tumor cells also expressed HMB-45. Ultrastructural examination showed irregularly shaped heterogeneous electron-dense bodies corresponding to the pigment, and the cytoplasmic vacuoles were formed by dilatation of intracytoplasmic lumens lined by microvilli. Since lipofuscin production can occur in normal ependymal cells and neuromelanin has been suggested to be a melanized form of lipofuscin, it is not surprising that these 2 pigments can be found in ependymoma. In all the previously reported cases, the pigment was shown to represent melanin only. In our case, the HMB-45 positivity in rare tumor cells indicated that there might also be a minor melanin component in the pigment in addition to lipofuscin and neuromelanin.

(Arch Pathol Lab Med. 2003;127:872-875)

Pigmented primary intracranial tumors other than meningeal melanocytomas and melanomas are rare. Intracytoplasmic pigment has been described infrequently in astrocytoma,1 ganglioglioma,2 ependymoma and subependymoma,3-5 choroid plexus papilloma and carcinoma 6-8 medulloblastoma,9 central neurocytoma,10 and schwannoma.11 The nature of the pigment has been shown to be either neuromelanin1,6,8,10 or melanin,2-4,7,9,11 with lipofuscin identified in addition to neuromelanin in some cases.8 We report a pigmented ependymoma in which the pigment consisted of a mixture of lipofuscin and neuromelanin, with suggestion of a minor melanin component.

REPORT OF A CASE

A 45-year-old man with a history of systemic hypertension presented with a 1-year history of headache of gradual onset. Blurring of vision and unsteady gait were also noticed in the recent 2-month period. Magnetic resonance imaging showed a contrast-enhancing midline cerebellar tumor involving the fourth ventricle, associated with mild obstructive hydrocephalus. The vermis was displaced superiorly. The tumor measured 64 x 35 x 27 mm and extended from the aqueduct to the Cl level of the spinal cord. There was no definite plane between the tumor and the floor of the fourth ventricle. Suboccipital craniotomy with tumor resection was attempted, but only subtotal excision could be achieved because of intraoperative bradycardia. During the operation, a brownish vascular tumor was identified in the cerebellum, and the tumor extended to the right cerebellopontine angle. The operation was complicated by lower cranial nerve palsy with marked dysphagia and aspiration pneumonia in the post-- operative period, and percutaneous enterogastrostomy was subsequently performed because of persistent dysphagia. Radiotherapy was given to treat residual tumor, and the patient was alive with residual disease 9 months after the operation.

MATERIALS AND METHODS

Cytology smears (squash preparation) and sections prepared from formalin-fixed paraffin-embedded tissue were stained with hematoxylin-eosin. The nature of the pigment in the tumor was further investigated by long Ziehl Neelsen, periodic acid-Schiff, S&mort Masson-Fontana, and Prussian blue stains. For Schmorl and Masson-Fontana stains, the effect of bleaching by potassium permanganate solution (10 minutes) was also studied.

Immunohistochemical stain for glial fibrillary acidic protein (dilution 1:200, Dako, Glostrup, Denmark) was performed on paraffin section using labeled streptavidin-biotin peroxidase technique with an automated immunostainer (Ventana, Tucson, Ariz). Immunohistochemical stain for HMB-45 (dilution 1:200, Dako) was performed manually on paraffin section using alkaline phosphatase anti-alkaline phosphatase technique (Dako). Antigen retrieval was effected by pressure cooking in EDTA buffer, pH 8.0, for 2.5 minutes.

Tissue was retrieved from paraffin block for ultrastructural study. After dewaxing in xylene and ethanol, the tissue was fixed in 2.5% glutaraldehyde followed by 1% osmium tetroxide. The tissue was embedded in Araldite 502 epoxy resin, and a semithin section stained with toluidine blue and an ultrathin section stained with uranyl acetate and lead citrate were prepared. The ultrathin section was examined using a Philips CM100 transmission electron microscope (Eindhoven, The Netherlands).

RESULTS

In the cytology smears, the tumor cells were arranged in small clusters and papillae. They had ovoid nuclei, a small amount of lightly eosinophilic cytoplasm, and fibrillary cytoplasmic processes (Figure 1, a). The chromatin was fine and evenly distributed, and occasionally cells had small distinct nucleoli. Occasional tumor cells had abundant brown pigment (Figure 1, b), and some other tumor cells contained a prominent single intracytoplasmic vacuole resulting in a signet ring appearance (Figure 1, c). The findings in the paraffin sections paralleled those observed in the cytology, with pigmented tumor cells and cytoplasmic vacuolation identified in focal areas (Figure 2, a). Perivascular pseudorosettes (Figure 2, b) and rare ependymal rosettes typical of ependymoma were present. The brown pigment was positive for long Ziehl Neelsen (ie, acid-fast) and periodic add-Schiff and focally positive for Masson-Fontana and Schmorl stains. The Masson-Fontana and Schmorl staining was abolished after pretreatment with potassium permanganate solution for 10 minutes. The pigment was negative for Prussian blue stain. The staining characteristics indicated that the pigment was a mixture of lipofuscin and neuromelanin.

Immunohistochemically, all tumor cells strongly expressed glial fibrillary acidic protein, and rare pigmented tumor cells were HMB-45 positive. Although ultrastructural examination was limited by suboptimal tissue preservation, the ependymal nature of the tumor cells was confirmed (Figure 3). Many tumor cells contained small intracytoplasmic lumens lined by microvilli (thin arrow) and long serpentine zonula adherens-type cell junctions (thick arrow). The cytoplasmic vacuoles discerned under light microscopy corresponded to markedly dilated intracytoplasmic lumens, with microvilli still focally found. Occasional tumor cells contained irregular-shaped heterogeneous electron-dense bodies (average 700 run; range 290-- 2000 nm) within the cytoplasm, corresponding to the pigment identified under light microscopy. It was difficult to determine whether these bodies were membranebound because of suboptimal tissue preservation. Premelan somes were not seen.

COMMENT

Pigments are normally found in certain parts of the central nervous system: neuromelanin in the neurons of substantia nigra and locus ceruleus; lipofuscin in the neurons in precentral gyrus, nuclei of cranial nerves, red nucleus, part of the thalamus, globus pallidus, inferior olives, and dentate nucleus; and sometimes melanin in the meninges. The staining characteristics of these 3 pigments are shown in Table 1.10,12 Unlike melanin, which is derived from tyrosine or tyrosine-containing precursors through the enzyme tyrosinase in melanosomes, neuromelanin is believed to be a product of auto-oxidation of catecholamine precursors.13 Lipofuscin, on the other hand, is a "wear-- and-tear" pigment that is derived from oxidation of lipids or lipoprotein sources. Neuromelanin and lipofuscin are closely linked histochemically, and interconvertability has been demonstrated experimentally; it has been suggested that neuromelanin is a melanized form of lipofuscin.14

Pigmented ependymomas are rare, with only 4 cases (including one subependymoma) having been previously reported in the literature (Table 2).3-5 After reviewing their illustrations and the clinical findings, we have determined that the so-called melanotic ependymoma with distant metastases reported by Panyathanya et al15 is probably a malignant melanoma. The pigment in all of these reported cases had the staining characteristics of melanin (MassonFontana positive and not acid-fast).3,4 Structures described as consistent with melanosomes were identified ultrastructurally, but there were no premelanosomes.4 These cases were reported as melanotic ependymomas. The authors speculated that this phenomenon might be related to the melanogenic potential of the primitive neuroepithehum, which is capable of developing into the pigmented cells, as in the outer wall of the optic cup.3,4 In contrast, the pigment in our case consisted of a mixture of lipofuscin and neuromelanin, in that it was acid-fast and periodic acid-Schiff-positive as well as Masson-Fontana- and Schmorl-positive in some cells. The positivity for MassonFontana and Schmorl was abolished after pretreatment with potassium permanganate for 10 minutes. Since this duration of bleaching is not adequate to bleach lipofuscin, the pigment is thus neuromelanin. 12 As lipofuscin production is known to occur in normal ependymal lining,16 it is not surprising that lipofuscin and the closely related neuromelanin can be identified in ependymomas. Choroid plexus is embryologically related to ependyma. It is of interest that pigmented choroid plexus tumors have also been reported. The pigment in some of these cases is similar to ours, consisting of neuromelanin with or without lipofuscin.6,8 The occurrence of rare HMB-45-positive tumor cells in our case is intriguing. HMB-45 is a monoclonal antibody that detects a premelanosome glycoprotein. Its expression indicates that there may be a minor component of melanin in the pigment in our case, which cannot be discriminated histochemically.

A wide variety of primary brain tumors may rarely contain pigment, but they usually should not be confused with the more common meningeal melanocytic tumors or metastasis from an occult melanoma. In cases in which the pigment is identified only focally, care should also be taken not to mistake entrapped pigmented nonneoplastic cells (eg, neurons in substantia nigra or cranial nerve nuclei) as pigmented tumor cells.

Another interesting finding in this case is the presence of intracytoplasmic vacuoles in the tumor cells. Such vacuoles have been reported to be a characteristic but uncommon feature of ependymoma.17,18 Although the resulting signet ring appearance may mimic metastatic adenocarcinoma, the identification of these vacuoles in a glial tumor is a clue to ependymal differentiation. The vacuolation is the consequence of distention of intracytoplasmic lumens. Two types of intracytoplasmic lumens have been described in ependymoma ultrastructurally: the more common type is lined by microvilli with or without rare cilia, and the other type is without microvilli".17,18; the latter was speculated to originate from organelles like rough or smooth endoplasmic reticulum or mitochondria.17

We thank John K. C. Chart, MBBS, for reviewing the case and the manuscript.

References

1. Vajtai I, Yonekawa Y, Schauble B, Paulus W. Melanotic astrocytoma Acta Neuropathol (Berl). 1996;91:549-553.

2. Soffer D, Lach B, Constantini S. Melanotic cerebral ganglioglioma: evidence for melanogenesis in neoplastic astrocytes. Acta Neuropathol (Berl). 1992;83: 315-323.

3. Rosenblum MK, Erlandson RA, Aleksic SN, Budzilovich GN. Melanotic ependymoma and subependymoma. Am I Surg Pathol. 1990;14:729-736.

4. McCloskey JJ, Parker JC Jr, Brooks WH, Blacker HM. Melanin as a component of cerebral gliomas: the melanotic cerebral ependymoma. Cancer. 1976; 37:2373-2379.

5. Russell DS, Rubinstein Li. Pathology of Tumours of the Nervous System. 5th ed. Baltimore, Md: Williams & Wilkins; 1989.

6. Boesel CP, Suhan JR A pigmented choroid plexus carcinoma: histochemical and ultrastructural studies. J Neuropathol Exp Neurol. 1979;38:177-186.

7. Lana-Peixoto MA, Lagos J, Silbert SW. Primary pigmented carcinoma of the choroid plexus: a light and electron microscopic study. J Neurosurg. 1977;47: 442-450.

8. Reimund EL, Sitton JE, Harkin JC. Pigmented choroid plexus papilloma. Arch Pathol Lab Med. 1990;114:902-905.

9. Boesel CP, Suhan JP, Sayers MP. Melanotic medulloblastoma: report of a case with ultrastructural findings. J Neuropathol Exp Neurol. 1978;37:531-543. 10. Ng TH, Wong AY, Boadle R, Compton JS. Pigmented central neurocytoma:

case report and literature review. Am J Surg Pathol. 1999;23:1136-1140.

11. Miller RT, Sarikaya H, Sos A. Melanotic schwannoma of the acoustic nerve. Arch Pathol Lab Med. 1986;110:153-154.

12. Bancroft DJ, Cook HC. Manual of Histologic Techniques and Their Diagnostic Application. Edinburgh, United Kingdom: Churchill Livingstone; 1994. 13. Graham DG. On the origin and significance of neuromelanin. Arch Pathol Lab Med. 1979;103:359-362.

14. Barden H. The histochemical relationship of neuromelanin and lipofuscin. J Neuropathol Exp Neurol. 1969;28:419-441.

15. Panyathanya R, Chantarakul N. Melanotic ependymoma with distant metastases. I Med Assoc Thai. 1982;65:454-458.

16. Friede R. The relation of the formation of lipofuscin to the distribution of oxidative enzymes in the human brain. Acta Neuropathol (Berl). 1962;2:113125.

17. Hirato J, Nakazato Y, lijima M, et al. An unusual variant of ependymoma with extensive tumor cell vacuolization. Acta Neuropathol (Bed). 1997;93:310316.

18. Zuppan CW, Mierau GW, Weeks DA. Ependymoma with signet-ring cells. Ultrastruct Pathol. 1994;18:43-46.

Alexander Chak Lam Chan, MBBS, FRCPA; Luen Cheung Ho, MBChB, MRCP, FRCPA; William Wai Lun Yip, MBChB; Fung Ching Cheung, MBBS, FRCS

Accepted for publication February 10, 2003.

From the Departments of Pathology (Drs Chan and Ho) and Neurosurgery (Dr Cheung), Queen Elizabeth Hospital, Hong Kong; and the Department of Pathology, Alice Ho Miu Ling Nethersole Hospital, Hong Kong (Dr Yip).

Reprints: Alexander C. L. Chan, MBBS, FRCPA, Department of Pathology, Queen Elizabeth Hospital, 30 Gascoigne Rd, Kowloon, Hong Kong (e-mail: chancl@ha.org.hk).

Copyright College of American Pathologists Jul 2003
Provided by ProQuest Information and Learning Company. All rights Reserved