Anatomic Pathology / APOPTOSIS AND BCL-2 IN MENINGIOMAS
Apoptotic Activity and bcl-2 Immunoreactivity in
Meningiomas
Association With Grade and Outcome
Caroline M. Abramovich, MD, and Richard A. Prayson, MD
Key Words: Apoptosis; bcl-2 Expression; Meningioma; MIB-1 antibody; Tumor recurrence
Abstract
We retrospectively evaluated 90 meningiomas for
bcl-2 expression, apoptosis counts (per 10 high-power
fields [HPF]), MIB-1 labeling indices (LI), and mitosis
counts (per 10 HPF). Characteristics were as follows:
37 low-grade (benign) meningiomas: mean apoptosis
count, 1.2; MIB-1 LI, 1.0; mitosis count, 0.1; and bcl-2
positivity, 40%; 29 atypical meningiomas: apoptosis
count, 3.3; MIB-1 LI, 5.5; mitosis count, 2.2; and bcl-2
positivity, 62%; 24 malignant meningiomas: apoptosis
count, 6.5; MIB-1 LI, 12.0; mitosis count, 6.0; and bcl-
2 positivity, 67%. By univariate analysis, MIB-1 LI,
apoptosis and mitosis counts, and tumor grade were
associated significantly with death due to tumor; by
multivariate analysis, only mitosis count was
independently associated with death due to tumor.
We compared similar data for 27 patients with
nonrecurrent tumors and 32 patients with recurrent
meningiomas. Histologic sections from the initially
resected tumor and from the most recent recurrence
were reviewed. Only the apoptosis count was
significantly higher by univariate analysis in the initial
resection specimens from tumors that ultimately
recurred vs nonrecurrent tumors. Expression of bcl-2,
MIB-1 LI, and mitosis count did not correlate with
recurrence. By multivariate analysis, only extent of
surgical resection was associated significantly with
tumor recurrence. Although bcl-2 immunostaining was
not associated statistically with outcome, bcl-2
positivity was more common in atypical and malignant
meningiomas than in low-grade tumors.
84 Am J Clin Pathol 2000;114:84-92
Although most meningiomas behave in a benign fashion,
a subset may recur locally, invade brain parenchyma, or,
rarely, metastasize. Unfortunately, there are no precise and
universally accepted histologic criteria established for
grading these neoplasms or for predicting tumor behavior.
The most recent revision of the World Health Organization
(WHO) classification separates meningiomas into 3 grades:
benign (grade I), atypical (grade II), and malignant (grade
III).1 Atypical meningioma is vaguely defined as a tumor with
several of the following histologic features: frequent mitoses,
hypercellularity, cells with increased nuclear/cytoplasmic
ratios, prominent nucleoli, sheet-like growth, and areas of
“spontaneous” or geographic necrosis. Malignant/anaplastic
meningioma is characterized by features of frank anaplasia
that are above and beyond those found in atypical tumors.
This classification leaves much room for individual interpre-
tation but reflects the well-known limitations of histologic
features for consistently predicting tumor behavior.
Numerous investigators have studied the histopathologic
features and the relationship of particular features of menin-
giomas with tumor recurrence and ultimate clinical outcome.2-
10 The proliferative activity of the tumor in terms of mitosis
count seems to be one of the more important features associ-
ated with poor prognosis.8,10 Additional studies have exam-
ined the potential role of cell proliferation markers, such as
MIB-1, in predicting tumor behavior.11-24 The MIB-1 labeling
index (LI) has been shown to increase generally with
increasing histologic tumor grade.11,14,15,25 Although the MIB-
1 LI has been associated significantly with tumor recurrence
in some studies,14,17,20-22 other studies have not supported this
finding.16,19,25
Apoptosis is regulated by several genes, one of which
is the bcl-2 proto-oncogene. The bcl-2 gene is responsible
© American Society of Clinical Pathologists

for the synthesis of bcl-2 protein, a 26-kd protein involved
in the inhibition of apoptosis and promotion of cell
survival. The bcl-2 protein works in conjunction with p53,
bax, and c-myc proteins in the regulation of apoptotic
activity.26 Several studies have examined the expression of
bcl-2 in tumors outside the central nervous system in which
bcl-2 may have some prognostic significance. 27-33 The
association of bcl-2 expression and prognosis in tumors of
the central nervous system and, in particular, meningiomas
is not clearly defined.
We retrospectively evaluated apoptotic activity, expres-
sion of bcl-2, mitotic activity, and MIB-1 labeling indices in
histologically benign, atypical, and malignant meningiomas,
as well as between nonrecurrent and recurrent tumors. Corre-
lation with tumor grade and clinical outcome in terms of
recurrence and death due to tumor is presented.
Materials and Methods
The pathology files at the Cleveland Clinic Foundation,
Cleveland, OH, were searched for meningiomas diagnosed
through 1997. All cases of recurrent meningiomas were
included if the initial tumor and the most recent (or only)
recurrence were resected at the Cleveland Clinic Foundation
and slides from both resections were available for review.
Recurrence, for purposes of the study, was defined as tumor
regrowth requiring a second surgical procedure for removal
of tumor. Nonrecurrent tumors were included only as long as
there was a long clinical follow-up period, with a minimum
follow-up of 88 months. A number of nonrecurrent menin-
giomas approximately equal to the number of recurrent
tumors was desired, so only nonrecurrent tumors resected
between January 1986 and December 1988 were included
for study. A total of 59 cases met the criteria and included 27
cases with nonrecurrent tumors and 32 with recurrent menin-
giomas. Additional benign and atypical meningiomas were
selected such that comparable numbers of benign, atypical,
and malignant meningiomas were included for evaluation. A
total of 90 tumors, a subset of the total number of menin-
giomas encountered at Cleveland Clinic Foundation, were
selected (37 benign, 29 atypical, 24 malignant). None of the
meningiomas were of the clear cell, papillary, or rhabdoid
types. Hemangiopericytomas were not included for study.
All available microscopic slides (range, 1-14) were
reviewed for each case. Routine histologic sections were cut
4-µm thick from formalin- or Hollandes-fixed, paraffin-
embedded tissue.
Tumors were graded as benign, atypical, or malignant
using, in part, criteria set forth in the most recent revision of
the WHO classification of brain tumors (1993).1 Benign
meningiomas were characterized by the presence of only 1 or
© American Society of Clinical Pathologists
Anatomic Pathology / ORIGINAL ARTICLE
none of the worrisome histologic features outlined by the
WHO for atypical meningioma (nuclear pleomorphism,
prominent nucleoli, mitoses, necrosis, and loss of architectural
pattern). Atypical meningiomas were defined by the presence
of 2 or more of the listed features. Any tumor that exhibited
invasion of brain parenchyma or was associated with clinical
metastasis was classified as a malignant meningioma.
Mitosis counts were recorded as the single highest
number of mitotic figures (MF) per 10 high-power fields
(HPF), evaluated in the most proliferative area of the tumor.
Counts were performed using a Nikon (Melville, NY)
binocular microscope with a wide field 10× ocular and 40×
objective, producing a high-power microscopic field with a
calculated area of 0.17 mm2 (diameter, 0.47 mm). Apop-
tosis counts were performed away from areas of geographic
tumor cell necrosis, if present, and were recorded as the
highest number of apoptotic bodies (AB) per 10 HPF.
Apoptosis counts were performed in a fashion similar to
that used for mitosis counts. Apoptotic bodies were identi-
fied as shrunken cells with pyknotic nuclei and eosinophilic
cytoplasm as described previously.34
Immunohistochemical staining with MIB-1 antibody
(1:10 dilution, AMAC, Westbrook, ME), and staining for bcl-
2 (Ventana Medical Systems, Tucson, AZ) was performed
using a microwave processing procedure and an avidin-
biotinylated immunoperoxidase method.35,36 Appropriate posi-
tive and negative controls were performed. MIB-1 and bcl-2
immunostaining were performed on 1 representative block of
tumor. The MIB-1 LI was recorded as the percentage of posi-
tively staining tumor cell nuclei in 1,000 tumor cell nuclei
evaluated. The determinations were made on high magnifica-
tion in the areas with the most immunostaining. Staining for
bcl-2 was recorded as positive or negative.
Clinical information for the study patients has been
reported previously11 and was obtained through review of
medical records, including radiographic and operative
reports and discharge summaries. Specifically noted were the
age of the patient at the time of the initial resection, sex, type
of surgery performed (gross total vs subtotal), number of
recurrences, interval to recurrence, adjuvant therapy, devel-
opment of metastases, most recent follow-up, and other
pertinent medical history.
Comparisons of apoptosis counts and mitosis counts
between benign, atypical, and malignant meningiomas were
completed by using the Dunn test. MIB-1 LIs were compared
between tumor grades by using the Kruskal-Wallis test and the
Wilcoxon 2-sample test. Positivity for bcl-2 was compared
between tumor grades by using the Cochrane-Mantel-Haens-
zel test. For testing associations of unfavorable outcome desig-
nated as death due to tumor with apoptosis count, mitosis
count, and MIB-1 LI, the Wald chi-square test was used.
Univariately significant associations then were tested in a
Am J Clin Pathol 2000;114:84-92 85
Abramovich and Prayson / APOPTOSIS AND BCL-2 IN MENINGIOMAS
multivariate model. The Cochrane-Mantel-Haenszel test was
used for evaluating association of bcl-2 expression and death
due to tumor.
Since the 2 groups of recurrent tumors were from the
same patients and, therefore, were paired groups, compar-
isons of apoptosis count, mitosis count, and MIB-1 LI were
made using the Wilcoxon signed-rank test and of bcl-2
expression using the sign test. Similar comparisons of apop-
tosis count, mitosis count, and MIB-1 LI between the nonre-
current group and the group of recurrent tumors from the
initial resection were performed using the Wilcoxon rank-
sum test, and for bcl-2 expression, the Cochrane-Mantel-
Haenszel test. To adjust for multiple comparisons, a Bonfer-
roni correction set the significance level at P = .0167.
Univariately significant associations then were tested in a
multivariate model.
Results
Benign, Atypical, and Malignant Meningiomas
Patients with benign tumors (n = 37) ranged in age from
30 to 80 years (mean, 54 years). Thirty-two patients (86%)
were women and 5 patients (14%) were men. Twenty-three
(62%) of the benign meningiomas were treated initially by
gross total resection, 10 (27%) by subtotal resection, and the
remainder (4 [11%]) by biopsy alone or by an unknown extent
of resection. Six patients received adjuvant radiation therapy.
Thirty-two patients (86%) were alive during a mean follow-up
interval of 109 months (range, 37-196 months). One patient
(3%) died with residual tumor at a postoperative interval of
129 months. Four patients (11%) died without evidence of
residual tumor or had unknown tumor status at postoperative
intervals of 101 to 202 months (median, 127 months).
Twenty-nine tumors were diagnosed as atypical menin-
gioma. The patients in this group ranged in age from 35 to
83 years (mean, 61 years) and included 17 women and 12
men. Gross total resection was performed as initial surgery
in 22 patients (76%). Five patients (17%) underwent subtotal
resection of the tumor. Information about extent of resection
was unavailable for 2 patients. Three patients received post-
operative radiation therapy after tumor recurrence. Twenty-
four patients (83%) were alive at postoperative intervals of 2
to 156 months (mean, 54 months). Two patients (7%) were
dead of tumor at 36 and 42 months. Three patients (10%)
died without residual tumor or with unknown tumor status at
0.5 to 84 months after surgery (median, 18 months).
Patients with malignant meningiomas (n = 24) ranged in
age from 18 to 82 years (mean, 59 years) and included 14
women and 10 men. Thirteen patients (54%) were treated
initially by subtotal resection, and 8 patients (33%) by gross
86 Am J Clin Pathol 2000;114:84-92
total resection. The extent of resection was unknown in 3
cases. Ten patients received postoperative radiation therapy
and 2 patients, postoperative chemotherapy. In 5 patients,
metastases developed to bone (n = 2), skin (n = 2), lung (n =
2), kidney (n = 1), and liver (n = 1). Fifteen patients (62%)
were alive at postoperative intervals of 3 to 348 months
(mean, 71 months). Five patients (21%) died of tumor at
postoperative intervals of 0.5 to 25 months (median, 24
months). Two patients (8%) died without residual tumor or
with unknown tumor status at 25 and 412 months after
surgery. Two patients were lost to follow-up.
Benign meningiomas demonstrated a generally low
level of proliferative activity as measured by mitosis counts
and MIB-1 LI. Single highest mitosis counts ranged from 0
to 1 MF/10 HPF (mean, 0.1 MF/10 HPF). MIB-1 LIs in
benign tumors ranged from 0.0 to 5.5 (mean, 1.0). In atypical
meningiomas, mitosis counts ranged from 0 to 19 MF/10
HPF (mean, 2.2 MF/10 HPF) and were significantly higher
than counts obtained in benign tumors (P < .001). MIB-1 LIs
were 0.1-32.5 (mean, 5.5) for atypical meningiomas ❚Image
1❚, and also were significantly higher than indices seen in
benign tumors (P < .0001). Malignant meningiomas exhib-
ited the highest proliferative activity, with mitosis counts of 0
to 20 MF/10 HPF (mean, 6.0 MF/10 HPF) ❚Image 2❚ and
MIB-1 LIs of 0.3 to 32.5 (mean, 12.0). Mitosis counts (P =
.01) and MIB-1 LIs (P = .0012) were significantly higher in
malignant than in atypical meningiomas.
Mitosis counts (P = .0035) and MIB-1 LIs (P = .013)
were significantly higher in meningiomas in patients who
died of tumor compared with those in patients who were
alive at the time of follow-up. Meningiomas in patients who
were alive (n = 72) showed mitosis counts of 0 to 20 MF/10
HPF (mean, 1.6 MF/10 HPF) and MIB-1 LIs of 0.0 to 32.5
(mean, 4.5). In patients who died of tumor (n = 8), mitosis
counts were 0 to 19 MF/10 HPF (median, 5.5 MF/10 HPF),
and MIB-1 LIs were 1.0 to 32.5 (median, 8.5).
Apoptosis counts ❚Image 3❚ were associated signifi-
cantly with tumor grade (P < .001). Benign tumors generally
exhibited lower apoptosis counts (range, 0-8 AB/10 HPF;
mean, 1.2 AB/10 HPF) than atypical tumors (range, 0-14
AB/10 HPF; mean, 3.3 AB/10 HPF), and malignant menin-
giomas generally demonstrated the highest apoptosis counts
(range, 0-16 AB/10 HPF; mean, 6.5 AB/10 HPF). Apoptosis
counts also were associated significantly with death due to
tumor (P = .0039). Meningiomas in patients who were alive
at follow-up had apoptosis counts of 0 to 14 AB/10 HPF
(mean, 2.6 AB/10 HPF), while those in patients who died of
tumor demonstrated a higher median value (range, 0-13
AB/10 HPF; median, 7.0 AB/10 HPF). Positivity for bcl-2
❚Image 4❚, however, did not correlate significantly with apop-
tosis counts, tumor grade, or death due to tumor. A summary
of results for proliferative and apoptotic indices for benign,
© American Society of Clinical Pathologists

❚Image 1❚ Immunostaining in an atypical meningioma with
an MIB-1 labeling index of approximately 30% (MIB-1, orig-
inal magnification ×480).
❚Image 3❚ Apoptotic bodies are characterized by
eosinophilic cytoplasm and fragmented pyknotic nuclei
(H&E, original magnification ×1,000).
atypical, and malignant meningiomas is given in ❚Table 1❚.
❚Table 2❚ presents proliferative and apoptotic indices for
patients who were alive at follow-up and in those who died
of tumor.
Variables that were associated significantly with death
due to tumor were tested in a multivariate model using a
significance level of .05. The extent of surgical resection and
tumor grade were included in this analysis. Mitotic activity
was the only variable found to have a strong independent asso-
ciation with death from tumor. All other variables became
© American Society of Clinical Pathologists
Anatomic Pathology / ORIGINAL ARTICLE
❚Image 2❚ Mitotic figures in a malignant meningioma.
Mitotic activity was significantly increased in higher-grade
meningiomas (H&E, original magnification ×900).
❚Image 4❚ Focal bcl-2 positivity in an atypical meningioma.
(original magnification ×1,000).
nonsignificant. Finally, determination of a ratio of mitosis
count over apoptosis count that might have correlated with
death was attempted, but no correlation was found.
Recurrent vs Nonrecurrent Meningiomas
Thirty-two patients had recurrent meningiomas with
ages at the time of initial surgery ranging from 30 to 83 years
(mean, 55 years). Twenty-six patients were women, and 6
were men. Seventeen patients (53%) underwent initial
subtotal resection, 10 (31%) gross total resection, and 1 (3%)
Am J Clin Pathol 2000;114:84-92 87
Abramovich and Prayson / APOPTOSIS AND BCL-2 IN MENINGIOMAS

❚Table 1❚
Apoptotic Activity, bcl-2 Expression, Mitosis Counts, and MIB-1 LI: Comparison With Histologic Grade

Benign Atypical Malignant Benign vs Atypical vs

(n = 37) (n = 29) (n = 24) Atypical Malignant

Apoptosis count (AB/10 HPF)

Mean (range) 1.2 (0-8) 3.3 (0-14) 6.5 (0-16) (n = 21) < .001 < .001*
No. (%) bcl-2 positive 14 (40) (n = 35) 18 (62) 14 (67) (n = 21) .059 .57†
Mitosis count (MF/10 HPF)
Mean (range) 0.1 (0-1) 2.2 (0-19) 6.0 (0-20) (n = 23) < .001 .01*
MIB-1 labeling index (%)
Mean (range) 1.0 (0-5.5) 5.5 (0.1-32.5) 12.0 (0.3-32.5) (n = 22) < .0001 .0012‡

AB, apoptotic bodies; HPF, high-power field; MF, mitotic figures.

* Comparisons made using the Dunn test. P values of .025 or less are considered statistically significant.
† Cochrane-Mantel-Haenszel test. P values of .0167 or less are considered statistically significant.
‡ Kruskal-Wallis test and Wilcoxon 2-sample test. A Bonferroni correction set the significance level at .025.

❚Table 2❚
Apoptotic Activity, bcl-2 Expression, Mitosis Counts, and MIB-1 LI: Association With Death Due to Tumor

Alive (n = 72) Dead of Tumor (n = 8) P*

Apoptosis count (AB/10 HPF)

Mean (range) 2.6 (0-14) (n = 70) 7.0† (0-13) .0039‡
No. (%) bcl-2 positive 38 (56) (n = 68) 6 (75) .062§
Mitosis count (MF/10 HPF)
Mean (range) 1.6 (0-20) (n = 71) 5.5† (0-19) .0035‡
MIB-1 labeling index (%)
Mean (range) 4.5 (0.0-32.5) (n = 71) 8.5† (1.0-32.5) .013‡

AB, apoptotic bodies; HPF, high-power field; MF, mitotic figures.

* P values less than or equal to .05 are considered statistically significant.
† Median.
‡ Wald chi-square test (logistic regression).
§ Cochrane-Mantel-Haenszel test.

biopsy alone. In 4 patients, the extent of initial surgery was
unclear from the available information. Two patients
received adjuvant chemotherapy, and 15 received adjuvant
radiation therapy. All patients developed tumor recurrence
requiring additional surgical resection. The number of recur-
rences, including metastases, ranged from 1 to 5 (mean, 1.4),
with the interval to the first or only recurrence ranging from
5 to 183 months (mean, 55.4 months). The interval between
the initial resection of tumor to the most recent or only recur-
rence ranged from 5 to 360 months (mean, 73.2 months).
Twenty-one patients (66%) were alive at postoperative inter-
vals of 27 to 348 months (mean, 116 months). Six patients
(19%) died of tumor at 19 to 129 months after surgery
(mean, 44 months). Five patients (16%) died without
residual tumor or with unknown tumor status at 84 to 412
months (mean, 190 months) after surgical resection.
The 27 patients with nonrecurrent tumors ranged in age
from 18 to 80 years (mean, 56 years) and included 21
women and 6 men. Twenty-five patients (92%) underwent
initial gross total resection. One patient (4%) underwent a
subtotal resection, and the extent of tumor resection in
another patient was not evident from the operative note.
None of the patients received adjuvant chemotherapy or radi-
ation therapy. None of the patients developed tumor recur-
rence during a minimum postoperative interval of 88 months.
Twenty-six patients (96%) were alive at postoperative inter-
vals of 88 to 124 months (mean, 109 months). No patients
died due to their tumor. One patient (4%) died of unrelated
causes at 101 months after surgical resection.
Mitosis counts recorded for the nonrecurrent menin-
giomas were low (range, 0-1 MF/10 HPF; mean, 0.1 MF/10
HPF). Although mitosis counts observed in both groups of
recurrent tumors were generally higher than those of the
nonrecurrent tumors, the differences were not statistically
significant. Likewise, MIB-1 LIs were not significantly
higher in the initial resections of recurrent tumors compared
with those of the nonrecurrent meningiomas and were some-
what lower in recurrent tumors compared with the corre-
sponding initial resection.
Apoptosis counts were significantly higher (P = .013) in
initial resections of tumors that ultimately recurred (range, 0-
13 AB/10 HPF; mean, 2.8 AB/10 HPF) compared with those
of nonrecurrent tumors (range, 0-3 AB/10 HPF; mean, 0.9,
AB/10 HPF). Apoptosis counts were similar between the 2

88 Am J Clin Pathol 2000;114:84-92 © American Society of Clinical Pathologists


groups of recurrent tumors. Meningiomas demonstrating bcl-
2 positivity were observed in each study group, and no
significant differences were noted between recurrent and
nonrecurrent tumors. ❚Table 3❚ gives the proliferative and
apoptotic indices observed for nonrecurrent and recurrent
meningiomas.
Mitosis and apoptosis counts, along with extent of
surgical resection and overall tumor grade, were tested in a
multivariate model with recurrence as the outcome. Extent of
surgical resection was the only significant variable indepen-
dently associated with tumor recurrence with an odds ratio of
44.2 (95% confidence interval, 4.6-427.7). This wide confi-
dence interval may be due to the relatively small sample but,
regardless, reflects the highly variable risk of recurrence
associated with incomplete surgical resection.
Discussion
It is well known that the clinical behavior of menin-
giomas cannot be predicted consistently using histopatho-
logic features alone. A number of studies have examined the
expression of proliferative markers in meningiomas in the
hope of identifying the tumors at increased risk of recur-
rence. Fewer studies have investigated the role of apoptosis
and expression of bcl-2 in determining the biologic behavior
of meningiomas.
The MIB-1 antibody recognizes the Ki-67 antigen, which
is a nuclear protein expressed only during the active phases of
the cell cycle. Although a few studies have failed to demon-
strate a correlation of tumor grade and MIB-1 LI, most studies
have, including the present study. This is consistent with the
apparent importance of mitotic activity in determining tumor
grade and the assertion by Perry et al10 that mitosis counts are
probably the most important histologic feature predicting
❚Table 3❚
Apoptotic Activity, bcl-2 Expression, and Proliferative Indices in Recurrent and Nonrecurrent Meningiomas
Recurrent
Nonrecurrent Initial Resection
(n = 27) (n = 32)
Apoptosis count (AB/10 HPF)
Mean (range) 0.9 (0-3) 2.8 (0-13)
No. (%) bcl-2 positive 12 (48) (n = 25) 13 (43) (n = 30)
Mitosis count (MF/10 HPF)
Mean (range) 0.1 (0-1) 1.3 (0-8)
MIB-1 labeling index (%)
Mean (range) 1.5 (0.0-8.3) 5.4 (0.0-32.5)
AB, apoptotic bodies; HPF, high-power fields; MF, mitotic figures.
* P values less than or equal to .0167 were considered statistically significant.
† P values for comparisons between nonrecurrent and initially (i) resected recurrent tumors were determined by using a Wilcoxon rank-sum test except for comparison of bcl-2
positivity, which was determined by using a Cochrane-Mantel-Haenszel test.
‡ P values for comparisons between the 2 groups of recurrent tumors (initially resected [i] and recurrent [r]) were determined by using a Wilcoxon signed-rank test except for
comparison of bcl-2 positivity, which was determined by using a sign test.
© American Society of Clinical Pathologists
Anatomic Pathology / ORIGINAL ARTICLE
biologic tumor behavior. In the study by Perry et al10 of 581
patients with meningiomas, one of the most significant
parameters associated with tumor recurrence was a mitotic
rate of at least 4 MF/10 HPF. That study was followed by
another demonstrating MIB-1 LI to be significantly associated
with decreased recurrence-free survival.22
However, there are a few limitations in using MIB-1 LI
to predict clinical behavior in a given tumor. One problem is
that there is considerable overlap with regard to MIB-1 LI
ranges between tumor grades. In general, a high MIB-1 LI
probably indicates a tumor that will behave in a more aggres-
sive fashion, but the issue of how high is too high is prob-
lematic. Another potential drawback is that the amount of
MIB-1 immunostaining can be variable in different regions
of a tumor, so sampling becomes an important issue. Finally,
laboratories may generate different labeling indices related to
differences in technique, making comparison of MIB-1 LIs
from institution to institution difficult.
In the present study, mitotic activity was the only vari-
able independently associated with death due to tumor.
Mitoses were not, however, independently associated with
tumor recurrence. Extent of surgical resection was the only
independent predictor of tumor recurrence. The tumors in the
nonrecurrent and recurrent study groups were not all resected
to the same extent, and this may explain the lack of a signifi-
cant association in this study between mitotic activity and
recurrence that was found in the study by Perry et al.10
The process of apoptosis, or programmed cell death, has
been studied only recently in tumors of the central nervous
system. Patsouris and colleagues37 studied 26 brain tumors,
including a varied number meningiomas, astrocytomas, and
oligodendrogliomas, with respect to apoptotic indices as
determined by an in situ end-labeling method. They found
that the meningiomas (n = 7) exhibited a higher mean apop-
tosis index (1.129%) than did the WHO grade II (n = 4) and
P*
Recurrence Nonrecurrent vs Recurrent (i) vs
(n = 32) Recurrent (i)† Recurrent (r)‡
3.0 (0-19) (n = 30) .013 .44
14 (47) (n = 30) .97 1.00
2.0 (0-17) (n = 29) .039 .24
3.5 (0.0-23.7) .18 .44
Am J Clin Pathol 2000;114:84-92 89
Abramovich and Prayson / APOPTOSIS AND BCL-2 IN MENINGIOMAS
III (n = 8) gliomas (1.097% and 0.699%, respectively).
Glioblastoma multiforme (n = 5), however, demonstrated a
higher apoptosis index (1.649%) than either meningiomas or
the lower-grade gliomas. Ellison et al38 compared apoptotic
indices among 16 low-grade fibrillary astrocytomas, 19
anaplastic astrocytomas, and 46 glioblastomas and found that
apoptotic indices increased with increasing histologic grade,
with median apoptotic indices of 0.0%, 0.072%, and
0.120%, respectively.
Few studies have focused on apoptotic indices specifi-
cally in meningiomas. Ng and Chen26 found that the apop-
totic index as measured in the terminal deoxynucleotidyl
transferase–mediated deoxyuridine triphosphate-biotin
nick-end labeling (TUNEL) assay was significantly higher
in 12 atypical/malignant meningiomas (apoptosis index,
0.12%) compared with 39 benign meningiomas (apoptosis
index, 0.023%). Although Maier and colleagues25 found
that apoptotic indices (determined by in situ tailing)
increased with histologic grade of meningioma, the results
were not statistically significant.
In our study, apoptosis counts increased significantly
with increasing histologic grade (P < .001) and were signifi-
cantly associated with death due to tumor (P = .0039) in
univariate analysis but not in multivariate analysis. They also
were significantly higher in primary resections of tumors that
later recurred than in nonrecurrent tumors (P = .013) in
univariate analysis. This association became nonsignificant
in multivariate analysis in which extent of surgical resection
became the only independent variable associated with tumor
recurrence. In the present study, similar to MIB-1 LI and
mitosis counts, there was a great deal of overlap with respect
to apoptosis counts observed between tumor grades and
between patients who were alive and those who died of
tumor. Less overlap was seen between recurrent and nonre-
current tumors, with none of the nonrecurrent meningiomas
having apoptosis counts higher than 3 AB/10 HPF.
It has been shown that the detection of apoptosis by
counting apoptotic bodies on H&E-stained slides is as accu-
rate as the TUNEL assay in evaluating apoptotic activity.39
The choice of method often is influenced by economic
resources. However, the detection of apoptotic bodies and
the interpretation of apoptosis assays, such as the TUNEL
assay, may be affected by delays in tissue fixation and by
lengthy strong fixation. In formalin-fixed autopsy tissue, a
higher false-positive rate of apoptotic body detection has
been reported owing to intense staining of postmortem mate-
rial with H&E.40
Tateyama and colleagues41 noted that a 2-hour delay in
tissue fixation was the maximum period allowed for accu-
rate assessment of apoptotic activity by the TUNEL assay.
Longer delays in fixation resulted in false-positive cells,
while increased length of fixation had no apparent effect on
90 Am J Clin Pathol 2000;114:84-92
apoptotic activity as measured by the TUNEL assay. 41
Others, however, have reported long fixation time to be
associated with false-negative results with the in situ end-
labeling method.42
The correlation of bcl-2 expression with prognosis in
tumors of the central nervous system and, in particular,
meningiomas, is unclear. Hara and associates43 examined
bcl-2 and bax expression in a variety of brain neoplasms
and found no correlation between expression of either of
these 2 regulatory proteins and tumor grade. In the study of
astrocytomas by Ellison et al,38 bcl-2 expression was found
in 44% of fibrillary astrocytomas, 42% of anaplastic astro-
cytomas, and 28% of glioblastomas. There was no definite
correlation of bcl-2 positivity and apoptosis index in this
set of tumors. Reactive astrocytes also were reported to
stain positively for bcl-2, thus complicating the matter.
Nakasu and colleagues44 examined bcl-2 staining in 140
varied brain tumors and reported more frequent immuno-
staining in low-grade astrocytomas than in high-grade
tumors, and almost half of the recurrent gliomas and
medulloblastomas showed increased protein expression
compared with tissue from initial resections.
Regarding bcl-2 expression in meningiomas, Karami-
topoulou and colleagues,13 in a study of 60 meningiomas,
compared bcl-2 expression between histologic tumor grades
and found bcl-2 positivity (defined as positive staining in
10% or more of the tumor cells) in 22% of the benign, 20%
of the atypical, and 46% of the malignant meningiomas.
When they compared bcl-2 expression between benign
meningiomas that did not recur and benign tumors that did,
they found that the benign meningiomas that ultimately
recurred were significantly more often positive for bcl-2
expression than were tumors that did not recur.
In our study, bcl-2 expression increased slightly with
increasing tumor grade, but the difference was not statisti-
cally significant. Positivity for bcl-2 was not associated
with tumor recurrence or death due to tumor. These results
are similar to those from other studies. Konstantinidou et
al45 reported no association between bcl-2 expression and
either tumor grade or recurrence in their study of 52 totally
resected meningiomas. Most of the atypical/malignant
meningiomas (64%) included in their study were negative
for bcl-2 immunostaining altogether, and the benign tumors
demonstrated a slightly higher mean bcl-2 labeling index
relative to the higher-grade tumors.
It is interesting that benign meningiomas express bcl-2
to any degree since bcl-2 inhibits apoptosis and intuitively
would lead to increased cell survival and tumor progression.
Perhaps the bcl-2 that is detected by immunostaining repre-
sents a dysfunctional protein, or there is accumulation of
protein that is produced constitutively at low levels that accu-
mulates as a result of decreased degradation. Apoptosis is a
© American Society of Clinical Pathologists

complex regulated process in which many other proteins
are involved, and bcl-2 expression alone is probably insuffi-
cient for predicting tumor behavior.
Mosnier and coworkers46 studied 39 meningiomas,
including 32 benign, 6 atypical, and 1 malignant tumor, and
reported bcl-2 positivity primarily in the spindle cell compo-
nent of transitional and fibroblastic meningiomas. This result
is in contrast to that of Hara et al, 43 who found bcl-2
immunoreactivity only in meningotheliomatous menin-
giomas and not in fibroblastic tumors. Mosnier and
colleagues46 also found no correlation between bcl-2 expres-
sion and either Ki-67 LI or sex hormone receptor status.
They compared bcl-2 expression in meningiomas evaluated
by immunohistochemical staining with those evaluated by
the Western blot technique and found that Western blot
detected bcl-2 positivity in 32 (82%) of the 39 meningiomas,
while positive immunostaining was present in only 17 (44%)
of the tumors.46
Expression of bcl-2 in our study was detected by
immunohistochemistry, and the staining in most cases was
focal, involving fewer than 10% of the tumor cells. It is
possible that our study underestimates the true number of bcl-
2–positive cells in meningiomas and that Western blot is a
more sensitive technique, but one that requires fresh tissue.
One might expect apoptotic activity as determined by
apoptosis counts or TUNEL assay to be inversely related to
bcl-2 expression, given that bcl-2 is a protein involved in the
inhibition of apoptosis. We found no correlation between
apoptosis counts and bcl-2 positivity, similar to other investi-
gators. Ng and Chen26 studied 51 meningiomas with regard to
apoptotic activity as measured by the TUNEL assay and
compared apoptotic indices with p53, c-myc, and bcl-2
expression determined by immunohistochemical staining.
They found no relationship between the apoptosis index and
expression of any of these 3 apoptosis-related regulatory
proteins. Our study also failed to demonstrate a statistically
significant association between proliferative indices (mitosis
counts and MIB-1 LI) with either apoptotic body count or bcl-
2 positivity. This result is similar to that of Mosnier and
colleagues,46 who reported no correlation of Ki-67 index with
bcl-2 expression. Karamitopoulou and coworkers13 reported
higher MIB-1 LIs in meningiomas that also expressed bcl-2,
although this was not statistically significant.
Apoptosis is a complex process of cell death regulated
by a number of genes, only one of which is bcl-2. Poor clin-
ical outcome, defined by death or tumor recurrence, is asso-
ciated with meningiomas characterized by increased mitosis
counts and incomplete surgical resection. Apoptosis counts,
MIB-1 LI, and bcl-2 positivity were not independently asso-
ciated with adverse outcome in the present study.
From the Department of Anatomic Pathology, Cleveland Clinic
Foundation, Cleveland, OH.
© American Society of Clinical Pathologists
Anatomic Pathology / ORIGINAL ARTICLE
Address correspondence to Dr Prayson: Dept of Anatomic
Pathology (L25), Cleveland Clinic Foundation, 9500 Euclid Ave,
Cleveland, OH 44195.
Acknowledgment: We thank Brett Larive, MS, for assistance
with the statistical analysis of the data.
References
1. Kleihues P, Burger PC, Scheithauer BW. The new WHO
classification of brain tumours. Brain Pathol. 1993;3:255-268.
2. Chen WYK, Liu HC. Atypical (anaplastic) meningioma:
relationship between histologic features and recurrence—a
clinicopathologic study. Clin Neuropathol. 1990;9:74-81.
3. de la Monte SM, Flickinger J, Linggood RM.
Histopathologic features predicting recurrence of
meningiomas following subtotal resection. Am J Surg Pathol.
1986;10:836-843.
4. Jääskeläinen J. Seemingly complete removal of histologically
benign intracranial meningioma: late recurrence rate and
factors predicting recurrence in 657 patients. Surg Neurol.
1986;26:461-469.
5. Jääskeläinen J, Haltia M, Servo A. Atypical and anaplastic
meningiomas: radiology, surgery, radiotherapy, and outcome.
Surg Neurol. 1986;25:233-242.
6. Mahmood A, Caccamo DV, Tomecek FJ, et al. Atypical and
malignant meningiomas: a clinicopathologic review.
Neurosurgery. 1993;33:955-963.
7. McLean CA, Jolley D, Cukier E, et al. Atypical and
malignant meningiomas: importance of micronecrosis as a
prognostic indicator. Histopathology. 1993;23:349-353.
8. Miller DC. Predicting recurrence of intracranial
meningiomas: a multivariate clinicopathologic model-
interim report of the New York University Medical Center
meningioma project. Neurosurg Clin N Am. 1994;5:193-200.
9. Perry A, Scheithauer BW, Stafford SL, et al. “Malignancy”
in meningiomas: a clinicopathologic study of 116 patients,
with grading implications. Cancer. 1999;85:2046-2056.
10. Perry A, Stafford SL, Scheithauer BW, et al. Meningioma
grading: an analysis of histologic parameters. Am J Surg
Pathol. 1997;21:1455-1465.
11. Abramovich CM, Prayson RA. MIB-1 labeling indices in
benign, aggressive, and malignant meningiomas: a study of
90 tumors. Hum Pathol. 1998;29:1420-1427.
12. Karamitopoulou E, Perentes E, Diamantis I, et al. Ki-67
immunoreactivity in human central nervous system tumors:
a study with MIB-1 monoclonal antibody on archival
material. Acta Neuropathol (Berl). 1994;87:47-54.
13. Karamitopoulou E, Perentes E, Tolnay M, et al. Prognostic
significance of MIB-1, p53, and bcl-2 immunoreactivity in
meningiomas. Hum Pathol. 1998;29: 140-145.
14. Kolles H, Niedermayer I, Schmitt C, et al. Triple approach
for diagnosis and grading of meningiomas: histology,
morphometry of Ki-67/Feulgen stainings, and cytogenetics.
Acta Neurochir (Wien). 1995;137:174-181.
15. Langford LA, Cooksley CS, DeMonte F. Comparison of
MIB-1 (Ki-67) antigen and bromodeoxyuridine proliferation
indices in meningiomas. Hum Pathol. 1996;27:350-354.
16. Madsen C, Schroder HD. Ki-67 immunoreactivity in
meningiomas: determination of the proliferative potential of
meningiomas using the monoclonal antibody Ki-67. Clin
Neuropathol. 1997;16:137-142.
Am J Clin Pathol 2000;114:84-92 91
Abramovich and Prayson / APOPTOSIS AND BCL-2 IN MENINGIOMAS
17. Matsuno A, Fujimaki T, Sasaki T, et al. Clinical and
histopathological analysis of proliferative potentials of
recurrent and non-recurrent meningiomas. Acta Neuropathol
(Berl). 1996;91:504-510.
18. Matsuno A, Nagashima T, Matsuura R, et al. Correlation
between MIB-1 staining index and the immunoreactivity of
p53 protein in recurrent and nonrecurrent meningiomas.
Am J Clin Pathol.1996;106:776-781.
19. Moller M-L, Braendstrup O. No prediction of recurrence of
meningiomas by PCNA and Ki-67 immunohistochemistry. J
Neurooncol. 1997;34:241-246.
20. Niedermayer I, Kolles H, Zang KD, et al. Characterization of
intermediate type (WHO “atypical”) meningiomas. Clin
Neuropathol. 1996;15:330-336.
21. Ohta M, Iwaki T, Kitamoto T, et al. MIB-1 staining index
and scoring of histologic features in meningiomas. Cancer.
1994;74:3176-3189.
22. Perry A, Stafford SL, Scheithauer BW, et al. The prognostic
significance of MIB-1, p53, and DNA flow cytometry in
completely resected primary meningiomas. Cancer.
1998;82:2262-2269.
23. Prayson RA. Malignant meningioma: a clinicopathologic
study of 23 patients including MIB1 and p53
immunohistochemistry. Am J Clin Pathol. 1996;105:719-726.
24. Shibuya M, Ito S, Miwa T, et al. Proliferative potential of
brain tumors. Cancer. 1993;71:199-206.
25. Maier H, Wanschitz J, Sedivy R, et al. Proliferation and
DNA fragmentation in meningioma subtypes. Neuropathol
Appl Neurobiol. 1997;23:496-506.
26. Ng H-K, Chen L. Apoptosis is associated with atypical or
malignant change in meningiomas: an in situ labelling and
immunohistochemical study. Histopathology. 1998;33:64-70.
27. Bubendorf L, Sauter G, Moch H, et al. Prognostic
significance of bcl-2 in clinically localized prostate cancer.
Am J Pathol. 1996;148:1557-1565.
28. Chan WK, Mole MM, Levison DA, et al. Nuclear and
cytoplasmic bcl-2 expression in endometrial hyperplasia and
adenocarcinoma. J Pathol. 1995;177:241-246.
29. Hellemans P, Dam PA, Weyler J, et al. Prognostic value of
bcl-2 expression in invasive breast cancer. Br J Cancer.
1995;72:354-360.
30. Joensuu H, Phylkkänen L, Toikkanen S. Bcl-2 protein
expression and long-term survival in breast cancer. Am J
Pathol. 1994;145:1191-1198.
31. Lauwers GY, Scott GV, Karpeh MS. Immunohistochemical
evaluation of bcl-2 protein expression in gastric
adenocarcinomas. Cancer. 1995;75:2209-2213.
92 Am J Clin Pathol 2000;114:84-92
32. Paraf F, Gogusev J, Chretien Y, et al. Expression of the bcl-2
oncoprotein in renal cell tumors. J Pathol. 1995;177:
247-252.
33. Pezzella F, Turley H, Kuzu I, et al. Bcl-2 protein in non–small
cell lung carcinoma. N Engl J Med. 1993;329:690-694.
34. Biscotti CV, Hart WR. Apoptotic bodies: a consistent
morphologic feature of endocervical adenocarcinoma in situ.
Am J Surg Pathol. 1998;22:434-439.
35. Goldblum JR, Shannon R, Kaldjian EP, et al. Immunohisto-
chemical assessment of proliferative activity in adreno-
cortical neoplasms. Mod Pathol. 1993;6:663-668.
36. Tubbs RR, Sheibani K. Immunohistology of lymphoprolif-
erative disorders. Semin Diagn Pathol. 1984;1:272-284.
37. Patsouris E, Davaki P, Kapranos N, et al. A study of
apoptosis in brain tumors by in situ end-labeling method.
Clin Neuropathol. 1996;15:337-341.
38. Ellison DW, Steart PV, Gatter KC, et al. Apoptosis in
cerebral astrocytic tumours and its relationship to expression
of the bcl-2 and p53 proteins. Neuropathol Appl Neurobiol.
1995;21:352-361.
39. Drachenberg CB, Ioffe OB, Papadimitriou JC. Progressive
increase of apoptosis in prostatic intraepithelial neoplasia
and carcinoma. Arch Pathol Lab Med. 1997;121:54-58.
40. Kashihara K. Detection of apoptosis using hematoxylin-
eosin [letter]. Arch Pathol Lab Med. 1998;122:493.
41. Tateyama H, Tada T, Hattori H, et al. Effects of prefixation
and fixation times on apoptosis detection by in situ end-
labeling of fragmented DNA. Arch Pathol Lab Med.
1998;122:252-255.
42. Davison FD, Groves M, Scaravilli F. The effects of formalin
fixation on the detection of apoptosis in human brain by in
situ end-labeling of DNA. Histochem J. 1995;27:983-988.
43. Hara A, Hirose Y, Yoshimi N, et al. Expression of bax and
bcl-2 proteins, regulators of programmed cell death, in
human brain tumors. Neurol Res. 1997;19:623-628.
44. Nakasu S, Nakasu Y, Nioka H, et al. Bcl-2 protein
expression in tumors of the central nervous system. Acta
Neuropathol (Berl). 1994;88:520-526.
45. Konstantinidou AE, Pavlopoulos PM, Patsouris E, et al.
Expression of apoptotic and proliferation markers in
meningiomas. J Pathol.1998;186:325-330.
46. Mosnier J-F, Perret AG, Brunon J, et al. Expression of the
bcl-2 oncoprotein in meningiomas. Am J Clin Pathol.
1996;106:652-659.
© American Society of Clinical Pathologists