Sage Journals: Discover world-class research

Abstract

This study examined immunohistochemical staining patterns for several meningioma variants involving either the brain or spinal cord of dogs. Formalin-fixed, paraffin-embedded tissue from 15 tumors was obtained. The selected tumor group included seven meningothelial, three transitional, two malignant (anaplastic), one myxoid, one papillary, and one osteomatous meningiomas. Tumors were evaluated for reactivity to the following six immunohistochemical markers: vimentin, pancytokeratin, glial fibrillary acidic protein (GFAP), S100, neuron-specific enolase (NSE), and synaptophysin. Vimentin expression was detected in all meningiomas, and 14 of 15 tumors demonstrated intense vimentin staining in more than 50% of the neoplastic cells. Pancytokeratin expression was present in 11 of 15 neoplasms; however, positive staining frequently was focal and often involved a small percentage of the neoplastic cells. GFAP expression was detected in a single, anaplastic meningioma. Although expression of NSE and S100 was detected in 12 of 25 meningiomas, the intensity of the staining and the percentage of positive neoplastic cells was highly variable. Synaptophysin was uniformly negative. These results will help to establish immunohistochemical profiles for meningiomas that will improve our ability to correctly differentiate these neoplasms of meningeal origin from central nervous system tumors originating from other sites.

Keywords

Cytokeratin dogs GFAP immunohistochemistry meningioma NSE S100 synaptophysin vimentin

Meningiomas are a common neoplasm of the central nervous system (CNS) in both human and veterinary medicine. They occur most frequently external to the dorsal cerebrum near the falx, and other common locations include the tentorium cerebelli, other external surfaces of the brain, and the interventricular region near the choroid plexus of the lateral and third ventricles.19,45 Meningiomas arise from the arachnoid cells (also termed meningothelial cells, meningocytes, or meningoblasts) located within arachnoid villi.4,54 These tumors are histologically diverse and can undergo both mesenchymal and epithelial differentiation, likely as a consequence of their development from the multipotential arachnoid cells.4,19,20

Normal arachnoid cells have the ability to perform many diverse and important functions within the leptomeninges. In some instances, they form intertwining cytoplasmic membranes that have a protective role and function similarly to epithelial and mesothelial cells.4 Arachnoid cells may also be involved in the fibrous response of leptomeninges to injury, and they have the ability to form several structurally important glycoproteins (fibronectin, laminin, collagen type IV, and procollagen type III). In addition, arachnoid cells may have a secretory role related to circulation of cerebrospinal fluid and are considered to be part of the mononuclear phagocyte system.4 Light and electron microscopy coupled with immunohistochemical staining have been utilized to determine the immunophenotype of arachnoid cells. Several human studies have described positive staining for vimentin, desmoplakin, epithelial membrane antigen (EMA), and some cytokeratins.2,21,29

Ultrastructural analyses have revealed that nonneoplastic arachnoid cells are joined through desmosomal attachments and hemidesmosome-like junctions, as well as gap, intermediate, and tight junctions.19,28,54 All of these structures also have been identified in meningiomas, although the number of cell junctions typically is decreased.54 An ultrastructural study by Mennel et al.30 concluded that interdigitating membranes, desmosomes, and intermediate filaments are found in all meningiomas, irrespective of the predominant morphological pattern.

In view of the wide array of functions performed by arachnoid cells and their ability to develop structural features of both mesenchyme and epithelia, it is not surprising that meningiomas exhibit highly variable morphologic patterns. The World Health Organization (WHO) classification for human tumors currently contains 15 distinct meningeal variants, which are separated into three grades (Table 1).27 The classification of meningiomas in veterinary medicine is not as elaborate and well defined. Many of the human categories and terms are utilized, but no universal classification system has been established. The most recent classification by Koestner et al.25 describes nine histologic patterns in domestic animals that parallel some of the variants included in the human classification system (Table 2). A list of patterns common to both human and veterinary medicine includes meningothelial (meningotheliomatous), fibrous, transitional, psammomatous, angiomatous, papillary, and anaplastic tumors.45

Table 1

World Health Organization (WHO) classification of human meningiomas.²⁷

WHO Grade I	WHO Grade II	WHO Grade III
Meningothelial	Atypical	Papillary
Fibrous (fibroblastic)	Clear cell	Rhabdoi
Transitional (mixed)	Chordoid	Anaplastic (malignant)
Psammomatous
Angiomatous
Microcystic
Secretory
Lymphoplasmacyte rich
Metaplastic

Table 2

World Health Organization (WHO) classification of meningiomas in domestic animals.²⁵

WHO Classification
Meningotheliomatous
Fibrous (fibroblastic)
Transitional (mixed)
Psammomatous
Angiomatous (angioblastic)
Papillary
Granular cell
Myxoid
Anaplastic (malignant)

In addition to these general categories, other variants have been described in the veterinary literature and mixtures of the patterns are common.44 Additional meningeal variants that have been documented include microcystic, atypical, lipomatous, myxoid, granular cell, angioblastic, endotheliomatous, and osteomatous tumors.11,25,40,45,48 According to WHO classification for human tumors, the lipomatous and osteomatous variants would be included in the metaplastic meningioma category.9,27 Although the terms angioblastic and angiomatous often are used synonymously in veterinary medicine, human researchers have concluded that neoplasms previously described as angioblastic meningiomas are actually hemangiopericytomas.17 Angiomatous refers to an actual meningeal variant, whereas the term angioblastic is now considered obsolete. The same trend for angioblastic meningiomas possibly exists in veterinary medicine, but this concept has yet to be investigated. Myxoid meningioma is a rare variant that has been described in both the human and veterinary literature. Although it has been included in the recent classification of meningiomas in domestic animals, myxoid meningioma is not currently included in human classification schemes.3,13,22,45,48 An evaluation of canine intracranial tumors by Patnaik et al.36 revealed that the meningotheliomatous (meningothelial) and transitional types are the most commonly diagnosed.

Vimentin is the predominant intermediate filament of human meningiomas, and continued investigation into the immunophenotype of these neoplasms also has revealed significant epithelial membrane antigen expression as well as desmoplakin and cytokeratin expression in many of the histologic subclassifications.5,12,26,31,43

Although meningiomas are a common neoplasm of dogs, particularly older dogs, few immunohistochemical studies on canine meningiomas have been reported in the literature.41 This study describes the staining patterns of canine meningiomas for six immunohistochemical markers that are commonly utilized to diagnose neoplasms of the CNS.

Materials and Methods

Twenty-nine canine meningiomas diagnosed at the Texas A&M University Veterinary Medical Teaching Hospital between 1993 and 2000 were evaluated, and 15 tumors were chosen for analysis. Only tumors that were unequivocally diagnosed as a meningioma were selected. In addition, several tumor blocks contained bone and had been previously decalcified. Because this process is known to interfere with antigen detection, these tumors were excluded from the study. The neoplasms were then screened with vimentin to determine if the preservation of the archived specimens was sufficient for detection and analysis of antigen expression. Tumors with poor vimentin labeling were excluded. The age of the tissue blocks selected for the study ranged from several months to 7 years; however, the majority had been stored at room temperature for less than 3 years. Tumors were classified morphologically with hematoxylin and eosin (HE)–stained sections and the WHO classification system for human meningiomas.27

A panel of six immunohistochemical markers that included vimentin, pancytokeratin, S100, neuron-specific enolase (NSE), glial fibrillary acidic protein (GFAP), and synaptophysin was selected (Table 3). All tissues were routinely processed with 10% neutral buffered formalin and embedded in paraffin. Five-micrometer–thick sections from the tissue blocks were placed onto poly l-lysine–coated slides. After air drying, tissue sections were deparaffinized and rehydrated before immunostaining. Antigen retrieval was performed either by enzyme pretreatment or by heating the tissue sections in a microwave oven before immunostaining. Microwave antigen retrieval was performed by continuous heating for 6 minutes at 20% power in an 800-watt oven. Both vimentin and synaptophysin required microwave pretreatment with citrate buffer pH 6.0. Enzyme pretreatment with proteinase K (Dako, Carpinteria, CA) for 2 minutes at room temperature was performed for pancytokeratin. S100, GFAP, and NSE did not require antigen retrieval. For all immunostains, endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 5 minutes at room temperature. A protein block (Signet Pathology Systems, Dedham MA) also was performed for all immunostains according to package instructions. Those slides requiring antigen microwave pretreatment received avidin–biotin blocking reagents. All primary antibodies were directed against human antigens except for GFAP, which was raised against bovine GFAP, and detection of the bound primary antibody was achieved by using a commercial kit from Signet Pathology Systems. Antibodies were diluted with Primary Antibody Diluent (Signet), and the presence of the relevant antigens was detected with diaminobenzidine (DAB, Signet) according to package instructions. Multitissue positive controls were prepared from canine tissues obtained at Texas A&M University. Negative controls were prepared by using antisera that do not cross-react with canine antigens.

Table 3

Antibodies used for immunohistochemistry.

Antibody	Dilution	Immunogen	Host	Type∗
Vimentin, V9 clone	1:100	Human	Mouse	Monoclonal
Pancytokeratin, MNF116 clone	1:1,000	Human	Mouse	Monoclonal
Glial fibrillary acidic protein	1:1,000	Bovine	Rabbit	Polyclonal
S100	1:1,000	Human	Rabbit	Polyclonal
Neuron-specific enolase	1:1,000	Human	Mouse	Monoclonal
Synaptophysin	1:20	Human	Mouse	Monoclonal

∗

All antibodies were obtained from Dako.

Immunohistochemical results were categorized based on the number of positively stained tumor cells as follows: − = no tumor cells stained; (+) = less than 10% or isolated positive tumor cells; + = 10–50% of the tumor cells stained positive; ++ = 50–90% of the tumor cells stained positive; and +++ = greater than 90% of the tumor cells stained positive. Staining intensity was reported as either light, moderate, or high. Positive results were further evaluated for the presence of focal or diffuse staining patterns, and the location of the stain either in the cytoplasm, nucleus, or on the cell surface.

Results

Several staining trends for meningiomas were observed. Table 4 provides a detailed list of the staining characteristics determined for each meningioma.

Table 4

Reactivity against antibodies.∗

Tumor type	Breed	Age (years)	Sex	Location	Vimentin	Pancytokeratin	GFAP	S100	NSE	Synaptophysin
Meningothelial	Miniature Schnauzer	12	FS	Ventral cerebrum	+	+	−	++	+	−
Meningotheial	Labrador	11	FS	Ventral cerebrum	++	−	−	−	+	−
Meningothelial	Miniature Dachshund	11	MN	Forebrain	+++	+ focal	−	+	+	−
Meningothelial	Mixed breed	14	MN	Pyriform lobe	+++	(+)	−	+	+	−
Meningothelial	Mixed breed	9	MN	Hindbrain	+++	+	−	++	−	−
Meningothelial	Shih Tzu	11	F	Forebrain	++	+ focal	−	+	+	−
Meningothelial	Shetland Sheepdog	13	FS	Forebrain	+++	(+)	−	++	(+)	−
Transitional	Labrador Retriever	1	M	Forebrain	+++	(+)	−	(+)	++	−
Transitional	Golden Retriever	10	MN	Forebrain	+++	−	−	(+)	−	−
Transitional	Mixed breed	11	MN	Forebrain	+++	−	−	+	+	−
Malignant/anaplastic	Shetland Sheepdog	8	F	Cervical spinal cord	++	+ focal	−	−	++	−
Malignant/anaplastic	Miniature Schnauzer	10	MN	Cervical spinal cord	+++	(+) focal	+	+	(+)	−
Myxoid	Golden Retriever	13	FS	Forebrain	+++	+ focal	−	−	−	−
Papillary	Spitz	10	MN	Cervical spinal cord	+++	−	−	(+)	+	−
Osteomatous	Fox Terrier	13	MN	Lumbar spinal cord	+++	(+) focal	−	+ focal	+	−

∗

GFAP = glial fibrillary acidic protein; NSE = neuron-specific enolase; F = female; S = spayed; M = male; N = neutered. +++ = more than 90% of tumor cells positive; ++ = 50–90% of tumor cells positive; + = 10–50% of tumor cells positive; (+) = <10% or isolated positive tumor cells; − = no tumor cell reactivity.

Positive staining for vimentin was identified in all tumors, regardless of subclassification, and 11 (73%) of 15 tumors displayed positivity in greater than 90% of the neoplastic cells. Only one tumor contained less than 50% positive cells. In all tumors, cells positive for vimentin displayed diffuse cytoplasmic staining, and the intensity was generally very high. Figures 1–3 illustrate vimentin staining from a meningothelial, anaplastic, and transitional meningioma, respectively.

Fig. 1.

Meningothelial meningioma. High-intensity, diffuse cytoplasmic staining for vimentin in nearly 100% of the neoplastic cells. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 2.

Anaplastic meningioma. High-intensity, diffuse cytoplasmic staining for vimentin in a majority of the neoplastic cells. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 3.

Transitional meningioma. High-intensity, diffuse cytoplasmic staining for vimentin throughout the cellular whorls typical of a transitional meningioma. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Pancytokeratin staining showed a much higher degree of variability than did vimentin. Four tumors failed to display a positive reaction for cytokeratin. No tumor contained greater than 50% positive cells, and the majority of the tumors had only 5–15% positive cells. Despite the lower percentages, the intensity of the cytokeratin stain was typically high. Positive cells were often randomly scattered throughout the sections, and they frequently formed small, multifocal aggregates (Figs. 4–6).

Fig. 4.

Meningothelial meningioma. Focal area of pancytokeratin positivity displaying a membranous staining pattern. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 5.

Meningothelial meningioma. Scattered cells demonstrating intense pancytokeratin reactivity are irregularly identified throughout the neoplasm. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 6.

Anaplastic meningioma. A focal area of cells staining intensely for pancytokeratin. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

The two most common meningeal variants in this study, meningothelial and transitional, appeared to have differences in cytokeratin expression. Cytokeratin staining was detected in six (86%) of seven of the meningothelial meningiomas, whereas one of three meningiomas contained positive cells. In addition, the number of positive cells in this transitional tumor was very low (<5%). The only other variant that failed to display cytokeratin positivity was the papillary meningioma (n = 1). The staining pattern for cytokeratin of all tumors was membranous with the exception of one meningothelial meningioma that exhibited a diffuse, punctate, cytoplasmic pattern.

GFAP expression was noted in only one anaplastic meningioma. The staining pattern was diffusely cytoplasmic, and the intensity varied from moderate to high. Positive cells were present in focal to locally extensive aggregates, in which approximately 30–40% of the neoplastic cells displayed GFAP expression (Fig. 7).

Fig. 7.

Anaplastic meningioma. A focal area of oval, neoplastic cells displaying intense, cytoplasmic staining for GFAP. Arrows denote several GFAP-positive stellate cells indicative of reactive astrocytes. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Positive staining for S100 was detected in 12 (80%) of 15 tumors. The intensity of the staining and the distribution of positive cells was highly variable. Although the number of positive cells was 70% in one tumor, most of the tumors contained 5–25% positive cells. The staining intensity was typically light to moderate. The pattern of S100 expression was variable. In some tumors, individual positive cells were scattered throughout the sections, whereas other tumors contained focal to locally extensive aggregates of positive cells. Granular staining in both the nucleus and cytoplasm was commonly noted for this marker (Figs. 8 9).

Fig. 8.

Anaplastic meningioma. Moderate to highly intense, granular staining for S100 is detected in both the cytoplasm and the nucleus. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 10 μm.

Fig. 9.

Meningothelial meningioma. Weak to moderately intense, granular staining for S100 is detected in both the cytoplasm and the nucleus. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 5 μm.

Immunostaining for NSE was highly variable in both distribution and intensity. Twelve (80%) of fifteen tumors demonstrated NSE expression. Similar to S100, the intensity of the staining was inconsistent, with the majority of the neoplasms displaying only a light to moderate level of positivity. Tumor sections contained a range of 5–60% positive cells; however, in most cases, the number of positive cells varied from 5 to 30%. Staining was typically cytoplasmic and granular (Figs. 10, 11). An interesting feature of NSE expression was the occurrence of positive cells in a bandlike fashion that occasionally centered around vessels or connective tissue cells (Fig. 12). These areas were located within the body of the tumor and not attributed to edge artifact.

Fig. 10.

Transitional meningioma. Diffuse, moderately intense, granular, cytoplasmic staining for NSE. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 11.

Anaplastic meningioma. Diffuse, granular cytoplasmic staining of variable intensity for NSE. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 20 μm.

Fig. 12.

Meningothelial meningioma. This section illustrates the bandlike pattern of NSE-positive cells that was characteristic of several tumors. Streptavidin–biotin complex, DAB–horseradish peroxidase, Mayer's hematoxylin counterstain. Bar = 40 μm.

Synaptophysin stained normal canine neural tissue, but it was uniformly negative in all tumors.

Discussion

Information on the immunophenotype of CNS neoplasms, including meningiomas, in veterinary medicine is extremely limited. Currently, the majority of the information regarding immunohistochemical phenotyping of CNS tumors has been derived from case reports involving only one or two animals. Table 5 summarizes previous studies and case reports that describe the immunophenotype of canine meningiomas. The present study was designed to elicit a more complete understanding of the immunohistochemical staining characteristics for six different morphologic patterns found in canine meningiomas and to determine similarities between immunohistochemical staining patterns in human meningiomas.

Table 5

Previous reports describing the immunophenotype of canine meningiomas.∗

			Percent of Tumors Demonstrating Positivity
Author, Year	No. of Cases	Type of Preparation	Vimentin	Cytokeratins	NSE	GFAP	S100	Tumor Type
Ribas et al., 1991 41	50	FF/PE	79	42	—	—	—	Variable†
Schulman et al., 1992 42	1	NA	100	0	—	—	—	Papillary
Van Winkle et al., 1994 48	4	FF/PE	100	50 (focal)	0	—	—	Myxoid
Fernández et al., 1995 10	1	NA	0	0	0	0	0	Papillary
Yeomans, 2000 55	1	NA	—	0	0	0	0	Meningothelial

∗

NSE = neuron-specific enolase; GFAP = glial fibrillary acidic protein; FF/PE = formalin-fixed/paraffin-embedded; NA = not available.

†

Multiple tumor variants were included in the study.

Numerous reports describe the immunophenotype of human meningiomas. The majority of these studies were performed retrospectively on formalin-fixed, paraffin-embedded tissues, and the most commonly evaluated immunohistochemical markers include vimentin, cytokeratin, GFAP, S100, NSE, and EMA. Several additional studies also have utilized frozen tissues, cytological specimens, and cultured cells for immunophenotyping. The results of these investigations are summarized in Table 6.

Table 6

Previous reports describing the immunophenotype of human meningiomas.∗

			Percent of Tumors Demonstrating Positivity
Author, Year	No. of Cases	Type of Preparation	Vimentin	Cytokeratins	NSE	GFAP	S100	Comments†
Tascos et al., 1982 46	18	FF/PE	—	—	—	0	—
Schwechheimer et al., 1984 43	10	Frozen	100	0	—	0	—
Halliday et al., 1985 12	13	FF/PE	100	0	—	—	—
Theaker et al., 1986 47	13	FF/PE	100	38	—	—	100
Meis et al., 1986 28	50	FF/PE	18	24	—	0	8
Vick et al., 1987 50	2	Cyto	—	—	—	0	—
Ng et al., 1987 32	29	FF/PE	—	69	—	—	—
Holden et al., 1987 16	14	AF/PE	100	86	—	—	—
Hitchcock and Morris, 1987 15	50	FF/PE	100	56	86	0	78
Mennel et al., 1988 30	39	AF/PE or FF/PE	99	9	—	3	6
Jitawi et al., 1988 18	19	NA	—	—	84	—	42
Radley et al., 1989 39	25	FF/PE	100	32	—	—	36
Chen and Zhang, 1989 6	16	Cyto	—	—	—	0	—
Ng et al., 1989 33	9	Fixed	100	0	—	0	0	Microcystic
Artlich and Schmidt, 1990 1	70	FF/PE	100	0–33	67–91	—	0–88
Mennel et al., 1990 29	12	Culture	100	0	0	0	75
Perez-Guiones et al., 1992 37	26	NA	100	19	—	0	19
Cruz-Sanchez et al., 1992 7	19	FF/PE	100	32	—	—	—
Ng and Wong, 1993 34	50	Frozen	98	50	—	—	—
Kobayashi, 1993 23	33	Cyto	100	—	—	0	0
Kobayashi et al., 1995 24	2	Cyto	100	50	0	0	50	Anaplastic
Probst-Cousin et al., 1997 38	31	FF/PE	84	100	—	—	33	Secretory

∗

NSE = neuron-specific enolase; GFAP = glial fibrillary acidic protein; AF/PE = alcohol-fixed/paraffin-embedded; FF/PE = formalin-fixed/paraffin-embedded; cyto = impression smears or cerebrospinal fluid; NA = not available.

†

Study conducted on a specific meningeal variant.

Vimentin is the most widely distributed intermediate filament found in mesenchymal tissues, and vimentin immunostaining is commonly used to delineate tissues of mesenchymal origin.35 The high level of vimentin expression detected in 100% of the meningiomas in the present study closely paralleled the results obtained from human studies. With the exception of a single report that revealed vimentin staining in only 9 (18%) of 50 tumors, all investigators detected strong vimentin staining in nearly 100% of the tumors examined.28 The conflicting report was obtained from one of the oldest studies, and less than optimal tissue fixation for immunohistochemistry may explain the unusually low level of vimentin staining.

A single study in the veterinary literature reported vimentin staining in 26 (79%) of 33 canine meningiomas.41 Two case reports involving papillary meningiomas had conflicting results. One tumor was vimentin positive, whereas the other did not show vimentin expression.10,42 Vimentin staining also was reported previously in microscopic sections from four myxoid meningiomas and cytologic impression smears from meningiomas in which the variant was not described.48,49

In the human literature, numerous studies have characterized the degree of epithelial differentiation in meningiomas by use of several markers, primarily cytokeratins and EMA. Human studies have consistently reported a high level of EMA expression.1,32,34,39 Unfortunately, effective antibodies for EMA currently have not been established in veterinary medicine. Cytokeratin expression displays a higher degree of variability than EMA expression in human meningiomas. Studies have reported a range of 0–100% positivity, but some of this variation may reflect the type of meningiomas that predominated. For example, a study performed entirely on secretory meningiomas concluded that all 31 tumors expressed cytokeratin.38 Another factor that researchers believe has contributed to the conflicting reports of cytokeratin staining in human meningiomas is the inconsistent use of different individual cytokeratin markers. Overall, 25% of human meningiomas have been estimated to have focal cytokeratin expression.9

In the present study, the staining pattern noted with the pancytokeratin antibody varied between different types of canine meningiomas. Although transitional meningiomas apparently are less likely to express cytokeratin, it is difficult to make a strong conclusion because of the low number of tumors examined (n = 3). A multifocal staining pattern involving less than 50% of the tumor cells was a common feature of cytokeratin expression. The only other comprehensive study evaluating the immunohistochemistry of canine meningiomas demonstrated that 42% (14/33) of the tumors had cytokeratin staining that was frequently focal in distribution.41 In a report detailing four myxoid meningiomas, two tumors displayed focal cytokeratin positivity.48 Immunocytochemical analysis performed on impression smears of meningiomas, in which the variant was not reported, also demonstrated cytokeratin staining in focal cell clusters.49 Thus, our data indicate that many cells in canine meningiomas will express cytokeratins in addition to vimentin.

GFAP is the characteristic intermediate filament of both normal and reactive glial cells, and GFAP production frequently is maintained in glial neoplasms.35 Although GFAP intermediate filaments are not considered a normal structural component of meningeal cells, GFAP expression has been rarely reported in human meningiomas44,52 and other nonglial tissues. Because reactive astrocytes are often associated with neoplasms, it may be difficult to determine whether GFAP-positive cells are infiltrating astrocytes or a true component of the tumor. Several convincing reports have described GFAP positivity in chorvp.39-3-311d and papillary meningiomas in the human literature.5,44,52

The present study also revealed strong evidence of GFAP expression in a malignant meningioma. Although reactive astrocytes were present, the majority of the GFAP-positive cells resembled neoplastic cells and not reactive astrocytes. Based on morphologic features of the neoplasm that were strongly indicative of a meningioma and concurrent cytokeratin expression, it was considered unlikely that this tumor represented an anaplastic glioma. Currently, the basis for GFAP expression in human meningiomas has not been established. This highly anaplastic tumor likely represents an aberration, and we postulate that GFAP expression will be a rare occurrence in meningiomas.

Originally, the S100 protein was found exclusively in neural cells, but subsequent investigations have demonstrated that many nonneural tissues, both neoplastic and nonneoplastic, react with antibodies to the S100 protein.35 Immunohistochemical evaluation of S100 protein in human meningiomas has demonstrated variable positivity that is rarely a prominent feature of the tumor.4,9,27 The staining pattern is inconsistent, and scattered foci of positive cells are often identified. One study described the intensity of S100 staining in several meningeal variants as weak.14 Staining for S100 may occur diffusely in both the cytoplasm and the nucleus.29,38,39 In the present study, very similar results for S100 were obtained. Although S100 often was detected, the intensity was rarely marked, and the positive cells often were scattered in a multifocal and irregular pattern throughout the tumor sections. Both nuclear and cytoplasmic staining were observed. Consequently, this marker may be of ancillary assistance in ruling out certain neoplasms, but will not help to establish a firm diagnosis of meningioma by itself.

Few investigators have attempted to evaluate the presence of NSE in human meningiomas, and significant discrepancy exists between the studies that have examined NSE expression. NSE preferentially localizes to neurons and endocrine cells of the amine precursor uptake and decarboxylation system, and this is frequently used to identify neural and neuroendocrine tumors. Despite the known specificities, NSE also has been identified in many other normal, reactive, and neoplastic tissues.51 Two studies of human meningiomas found no NSE expression, whereas three others found NSE staining in 67–91% of the tumors evaluated. The two studies that reported no staining were performed on either cultured cells or cytologic samples from a total of 14 meningiomas, whereas the other three studies included formalin-fixed, paraffin-embedded samples from 139 tumors.1,15,18,24,29 The differences in NSE expression may be attributable to the type of sample used for analysis; however, given the large percentage of the 139 tumors that demonstrated NSE staining, it is likely that NSE positivity is a real feature of human meningiomas.

In the canine meningiomas, NSE-positive cells seemed to be a consistent feature; however, the number of positive cells in each tumor was typically low (approximately 5–30%) and staining was often light to moderate and not a prominent feature. The significance of the bandlike pattern of positivity identified in some tumors could not be determined.

Synaptophysin, a highly specific marker for presynaptic vesicles found in neural and neuroendocrine tissues, is utilized to diagnose primary neural tumors (e.g., medulloblastoma and ganglioneuroma) in the brain and spinal cord.35 In addition, synaptophysin occasionally is detected in oligodendrogliomas.8,53 All of the canine meningiomas in this study were negative, a feature consistent with the absence of synaptophysin staining in reported cases of human meningiomas.

In summary, the immunohistochemical phenotypes identified for major morphologic variants of canine meningiomas are very similar to those described for human meningiomas. Vimentin is the predominant intermediate filament, and strong vimentin expression in a majority of neoplastic cells is expected. Cytokeratin staining is variable but common, and frequently focal. Although some meningiomas will fail to express cytokeratin, the presence of dual positivity for vimentin and cytokeratin is a useful diagnostic feature that helps distinguish meningiomas from other tumors of the CNS. In general, GFAP staining is indicative of a glial tumor, but one meningeal neoplasm expressed GFAP. Consequently, GFAP expression does not unequivocally exclude the possibility of a meningioma. S100 and NSE staining often are reported for meningiomas; however, they are rarely the predominant feature, and the presence or absence of these markers cannot be utilized to confirm the diagnosis of meningioma. A lack of synaptophysin expression is expected for all meningeal variants. The results of this study suggest that canine meningiomas have antigenic expression that is comparable to their human counterparts, and the immunoreactivity profiles provide a basis for further study of this disease.

References

1 Artlich

Schmidt

: Immunohistochemical profile of meningiomas and their histological subtypes. Hum Pathol 21:843–849, 1990

2 Assi

Declich

Iacobellis

Cozzi

Tonnarelli

: Secretory meningioma, a rare meningioma subtype with characteristic glandular differentiation: an histological and immunohistochemical study of 9 cases. Adv Clin Pathol 3:47–53, 1999

3 Bégin

: Myxoid meningioma. Ultrastruct Pathol 14: 367–374, 1990

4 Bruner

Tien

Enterline

: Tumors of the meninges and related tissues. In: Russell and Rubinstein's Pathology of the Nervous System, ed. Bigner

McLendon

Burner

, 6th ed., pp. 67–139. Oxford University Press, New York, NY, 1998

5 Budka

: Non-glial specificities of immunocytochemistry for the glial fibrillary acidic protein (GFAP). Acta Neuropathol 72:43–54, 1986

6 Chen

Zhang

: Use of monoclonal antibodies to glial fibrillary acidic protein in the cytologic diagnosis of brain tumors. Acta Cytol 33:922–928, 1989

7 Cruz-Sanchez

Garcia-Bachs

Rossi

Rodriguez-Prados

Ferrer

Coakham

Ferreres

Figols

Palacin

: Epithelial differentiation in gliomas, meningiomas and choroid plexus papillomas. Virchows Arch B Cell Pathol Incl Mol Pathol 62:25–34, 1992

8 Dehghani

Schachenmayr

Laun

Korf

: Prognostic implication of histopathological, immunohisto-chemical and clinical features of oligodendrogliomas: a study of 89 cases. Acta Neuropathol 95:493–504, 1998

9 Ellison

Love

Chimelli

Harding

Lowe

Love

Roberts

Vinters

: Neuropathology, pp. 43.1–43.14. Mosby International Ltd, London, UK, 1998

10.

10 Fernández

Añor

Gómez

Borràs

Rios

Pumarola

: Papillary meningioma in the caudal fossa of a dog. Prog Vet Neurol 6:85–87, 1995

11.

11 Fingeroth

Prata

Patnaik

: Spinal meningiomas in dogs: 13 cases. J Am Vet Med Assoc 191:720–726, 1987

12.

12 Halliday

Yeger

Duwe

Phillips

: Intermediate filaments in meningiomas. J Neuropathol Exp Neurol 44:617–623, 1985

13.

13 Harrison

Rose

: Myxoid meningioma: histochemistry and electron microscopy. Acta Neuropathol 68:80–82, 1985

14.

14 Hayashi

Hoshida

Horie

Takahashi

Taguchi

Sonobe

Ohtsuki

Akagi

: Immunohistochemical study on the distribution of α and β subunits of S-100 protein in brain tumors. Acta Neuropathol 81:657–663, 1991

15.

15 Hitchcock

Morris

: Immunocytochemistry of intracranial meningiomas. J Neuro-Oncol 5:357–368, 1987

16.

16 Holden

Dolman

Churg

: Immunohistochemistry of meningiomas including the angioblastic type. J Neuropathol Exp Neurol 6:50–56, 1987

17.

17 Jääskeläinen

Louis

Paulus

Haltia

: Haemangiopericytoma. In: Pathology & Genetics: Tumours of the Nervous System, ed. Kleihues

Cavenee

, 2nd ed., pp. 190–192. International Agency for Research on Cancer, Lyon, France, 2000

18.

18 Jitawi

Cochran

Cancilla

Wen

: The expression of S-100 protein and neuron-specific enolase in meningiomas. Dis Markers 6:109–117, 1988

19.

19 Johnson

: Genesis and pathology of tumors of the nervous system. Semin Vet Med Surg (Small Anim) 5: 210–222, 1990

20.

20 Kepes

: Presidential address: the histopathology of meningiomas. A reflection of origins and expected behavior?. J Neuropathol Exp Neurol 45:95–107, 1986

21.

21 Kida

Yamashima

Kubota

Ito

Yamamoto

: A light and electron microscopic and immunohistochemical study of human arachnoid villi. J Neurosurg 69:429–435, 1988

22.

22 Kimura

Matsumae

Tsutsumi

: Pericellular deposition of basement membrane material in myxoid meningioma: immunohistochemical evidence for unbalanced production of type IV collagen and laminin. Pathol Int 48:53–57, 1998

23.

23 Kobayashi

: Meningioma, neurilemmoma and astrocytoma specimens obtained with the squash method for cytodiagnosis: a cytologic and immunochemical study. Acta Cytol 37:913–922, 1993

24.

24 Kobayashi

Haba

Hirakawa

Yamadori

Miki

Ohmori

: Cytology and immunohistochemistry of anaplastic meningiomas in squash preparations. Acta Cytol 39:118–124, 1995

25.

25 Koestner

Bilzer

Fatzer

Schulman

Summers

Van Winkle

: World Health Organization: Histological Classification of Tumors of the Nervous System of Domestic Animals, pp. 27–29. Armed Forces Institute of Pathology, Washington, DC, 1999

26.

26 Lantos

Vandenberg

Kleihues

: Tumours of the nervous system. In: Greenfield's Neuropathology, ed. Graham

Lantos

, 6th ed., pp. 583–879. Oxford University Press, New York, NY, 1997

27.

27 Louis

Scheithauer

Budka

von Deimling

Kepes

: Meningiomas. In: Pathology & Genetics: Tumours of the Nervous System, ed. Kleihues

Cavenee

, 2nd ed., pp. 176–184. International Agency for Research on Cancer, Lyon, France, 2000

28.

28 Meis

Ordóñez

Bruner

: Meningiomas: an immunohistochemical study of 50 cases. Arch Pathol Lab Med 110:934–937, 1986

29.

29 Mennel

Berweiler-Nippert

Lorenz

Geissel

: Routine immunohistochemical characterization of short-term in vitro explants from human intracranial tumours. Zentralbl Allg Pathol 136:671–678, 1990

30.

30 Mennel

Zinngrebe

Berweiler-Nippert

Lorenz

: Meningiomas: histogenesis and classification. Zentralbl Allg Pathol 134:27–40, 1988

31.

31 Murphy

Chen

George

: Establishment and characterization of a human leptomeningeal cell line. J Neurosci Res 30:475–483, 1991

32.

32 Ng

Tse

CCH

STH

: Meningiomas and arachnoid cells: an immunohistochemical study of epithelial markers. Pathology 19:253–257, 1987

33.

33 Ng

Tse

CCH

STH

: Microcystic meningiomas—an unusual morphological variant of meningiomas. Histopathology 14:1–9, 1989

34.

34 Ng

Wong

ATC

: Expression of epithelial and extra-cellular matrix protein markers in meningiomas. Histopathology 22:113–125, 1993

35.

35 Osborn

Domagala

: Immunocytochemistry. In: Comprehensive Cytopathology, ed. Bibbo

, 2nd ed., pp. 1033–1074. WB Saunders, Philadelphia, PA, 1997

36.

36 Patnaik

Kay

Hurvitz

: Intracranial meningioma: a comparative pathologic study of 28 dogs. Vet Pathol 23:369–373, 1986

37.

37 Perez-Guiones

Cerda-Nicolas

Piquer

Barcia-Marino

: Meningiomas: immunohistochemical analysis of 26 cases. Arch Neurobiol 55:43–49, 1992

38.

38 Probst-Cousin

Villagran-Lillo

Lahl

Bergmann

Schmid

Gullotta

: Secretory meningioma: clinical, histologic, and immunohistochemical findings in 31 cases. Cancer 79:2003–2015, 1997

39.

39 Radley

di Sant'agnese

Eskin

Wilbur

: Epithelial differentiation in meningiomas: an immuno-histochemical, histochemical and ultrastructural study—with review of the literature. Am J Clin Pathol 92:266–272, 1989

40.

40 Raskin

: An atypical spinal meningioma in a dog. Vet Pathol 21:538–540, 1984

41.

41 Ribas

Carpenter

Mena

Schulman

Lee

: Central nervous system meningioma in the dog: a review of 50 cases. J Neuropathol Exp Neurol 50:373, 1991 [Abstract]

42.

42 Schulman

Carpenter

Ribas

Brum

: Cystic papillary meningioma in the sella turcica of a dog. J Am Vet Med Assoc 200:67–69, 1992

43.

43 Schwechheimer

Kartenbeck

Moll

Franke

: Vimentin filament–desmosome cytoskeleton of diverse types of human meningiomas. Lab Invest 51:584–591, 1984

44.

44 Su

Ono

Tanaka

Takahashi

: An unusual meningioma variant with glial fibrillary acidic protein expression. Acta Neuropathol 94:499–503, 1997

45.

45 Summers

Cummings

de Lahunta

: Veterinary Neuropathology, pp. 355–362. Mosby-Yearbook, Inc, St. Louis, MO, 1995

46.

46 Tascos

Parr

Gonatas

: Immunocytochemical study of the glial fibrillary acidic protein in human neoplasms of the central nervous system. Hum Pathol 13: 454–458, 1982

47.

47 Theaker

Gatter

Esiri

Fleming

: Epithelial membrane antigen and cytokeratin expression by meningiomas: an immunohistological study. J Clin Pathol 39:435–439, 1986

48.

48 Van Winkle

Steinberg

DeCarlo

Dayrell-Hart

Steinberg

Smith

Summers

: Myxoid meningiomas of the rostral cervical spinal cord and caudal fossa in four dogs. Vet Pathol 31:468–471, 1994

49.

49 Vernau

Higgins

Bollen

Jimenez

Anderson

Koblik

LeCouteur

: Primary canine and feline nervous system tumors: intraoperative diagnosis using the smear technique. Vet Pathol 38:47–57, 2001

50.

50 Vick

Wikstrand

Bullard

Kemshead

Coakham

Schlom

Johnston

Bigner

: The use of a panel of monoclonal antibodies in the evaluation of cytologic specimens from the central nervous system. Acta Cytol 31:815–824, 1987

51.

51 Vinores

Herman

Rubinstein

: Electron-immunocytochemical localization of neuron-specific enolase in cytoplasm and on membranes of primary and metastatic cerebral tumours and on glial filaments of glioma cells. Histopathology 10:891–908, 1986

52.

52 Wanschitz

Schmidbauer

Maier

Rössler

Vorkapic

Budka

: Suprasellar meningioma with expression of glial fibrillary acidic protein: a peculiar variant. Acta Neuropathol 90:539–544, 1995

53.

53 Wolf

Buslei

Blümcke

Wiestler

Pietsch

: Neural antigens in oligodendrogliomas and dysembryoplastic neuroepithelial tumors. Acta Neuropathol 94: 436–443, 1997

54.

54 Yamashima

Kida

Yamamoto

: Ultrastructural comparison of arachnoid villi and meningiomas in man. Mod Pathol 1:224–234, 1988

55.

55 Yeomans

: Extensive spinal meningioma in a young dog. J Comp Pathol 122:303–306, 2000