Bcl-2 immunoreactive cells with immature neuronal phenotype exist in the nonepileptic adult brain

Abstract

Bcl-2, a cell death suppressor protein, is expressed during brain development but is largely down-regulated in the adult central nervous system. We previously reported strong expression of bcl-2 in small, "oligodendrocyte-like" cells (OLC) found in glioneuronal hamartias. These hamartias are microscopic cell rests found in temporal lobe resections from patients with intractable epilepsy and are considered a form of cerebral microdysgenesis. However, a causative relationship between these rests and seizures is not clear. We now report the identification, lineage characterization, and postnatal ontogeny of hamartia-like cell rests in temporal lobes of nonepileptic humans. Postmortem temporal lobes from 28 patients without history of neurologic disease (mean age = 53 years; range = 20 to 83 years) were studied. Microscopic cellular aggregates containing immature-appearing, bcl-2-immunoreactive cells (BIC) (identical to OLC) were observed in 23 of 28 (82%) temporal lobes from nonepileptic individuals. BIC were strongly immunoreactive for neuronal-specific class III beta tubulin, neuronal nuclear antigen, and MAP-2, but were consistently negative for neurofilament proteins and Ki67. Such cells were localized to subventricular regions of the caudal amygdala and often extended into the adjacent subcortical white matter and periamygdaloid cortex. BIC became less abundant with advancing age. These findings suggest that hamartia-like rests containing immature postmitotic neurons are normally present in the human brain and that glioneuronal hamartias may not always represent a maldevelopmental lesion associated with epilepsy.

Full text

113
Journal of Neuropathology and Experimental Neurology Vol. 59, No. 2
Copyright
q
2000 by the American Association of Neuropathologists February, 2000
pp. 113–119
Bcl-2 Immunoreactive Cells With Immature Neuronal Phenotype Exist in the Nonepileptic Adult
Human Brain
A
NTHONY
T. Y
ACHNIS
, MD, S
TEVEN
N. R
OPER
, MD, A
LIX
L
OVE
, BS, J
ASON
T. F
ANCEY
, MD,
AND
D
AVID
M
UIR
,P
H
D
Abstract. Bcl-2, a cell death suppressor protein, is expressed during brain development but is largely down-regulated in the
adult central nervous system. We previously reported strong expression of bcl-2 in small, ‘‘oligodendrocyte-like’’ cells (OLC)
found in glioneuronal hamartias. These hamartias are microscopic cell rests found in temporal lobe resections from patients
with intractable epilepsy and are considered a form of cerebral microdysgenesis. However, a causative relationship between
these rests and seizures is not clear. We now report the identification, lineage characterization, and postnatal ontogeny of
hamartia-like cell rests in temporal lobes of nonepileptic humans. Postmortem temporal lobes from 28 patients without history
of neurologic disease (mean age
5
53 years; range
5
20 to 83 years) were studied. Microscopic cellular aggregatescontaining
immature-appearing, bcl-2-immunoreactive cells (BIC) (identical to OLC) were observed in 23 of 28 (82%) temporal lobes
from nonepileptic individuals. BIC were strongly immunoreactive for neuronal-specific class III
b
tubulin, neuronal nuclear
antigen, and MAP-2, but were consistently negative for neurofilament proteins and Ki67. Such cells were localized to sub-
ventricular regions of the caudal amygdala and often extended into the adjacent subcortical white matter and periamygdaloid
cortex. BIC became less abundant with advancing age. These findings suggest that hamartia-like rests containing immature
postmitotic neurons are normally present in the human brain and that glioneuronal hamartias may not always represent a
maldevelopmental lesion associated with epilepsy.
Key Words: Amygdala; Hamartia; Neurogenesis; Tubulin; Nuclear antigen.
INTRODUCTION
Bcl-2, the prototype of a family of cell death regulatory
genes, encodes a 26 kDa intracellular membrane-associ-
ated polypeptide that promotes neuronal survival in a
number of experimental systems (1–8). It is expressed
primarily during brain development in primitive neuro-
epithelium and in early, postmitotic young neurons and
then diminishes to nondetectable levels in most postnatal
CNS neurons (9–13). Instead, most mature CNS neurons
strongly express bcl-x
L
, which appears to be an important
negative modulator of cell death in the developing and
mature CNS (14, 15).
In a previous report, we described a population of im-
mature-appearing, strongly bcl-2-immunoreactive cells
(BIC) in temporal lobe resections from adult patients with
intractable epilepsy (12). BIC are identical to the ‘‘oli-
godendrocyte-like cells’’ (OLC) found in so-called ‘‘gli-
oneuronal hamartias.’’ The latter have been defined by
Wolf and colleagues (16) as ‘‘. . . well to poorly circum-
scribed, microscopic cell clusters that are composed of
randomly oriented mature neurons admixed with small
cells which contain round, hyperchromatic nuclei and,
often, clear perinuclear cytoplasmic haloes.’’ While his-
tologic appearances suggested an oligodendroglial line-
age, BIC often contain a small but prominent basophilic
From the Departments of Pathology, Immunology, and Laboratory
Medicine (AY, AL, JF), Neurological Surgery (SR), Pediatrics and Neu-
roscience (DM), University of Florida Brain Institute and College of
Medicine, Gainesville, Florida.
Correspondence to: Dr. Anthony T. Yachnis, Department of Pathology
and Laboratory Medicine, University of Florida College of Medicine,
1600 S. W. Archer Rd., Room 3110, P.O. Box 100275, Health Science
Center, Gainesville, FL 32610.
nucleolus and express a fetal form of the neural cell ad-
hesion molecule (fetal NCAM) (16) and bcl-2 (12) (oli-
godendroglia are negative for these antigens). ‘‘Hamar-
tias’’ with such cells were found in 58% of glioneuronal
malformative ‘‘lesions’’ associated with intractable epi-
lepsy (16). It has been suggested that BIC, which are
often found near the temporal horn of the lateral ventricle
(12), represent immature neural cells that failed to mi-
grate appropriately from the fetal germinal matrix (16,
17). To date, these cells have been considered to be path-
ologic; representing a form of microdysgenesis associated
with temporal lobe epilepsy (12, 16, 17).
We now report the identification and immunohisto-
chemical characterization of ‘‘hamartias’’ containing im-
mature-appearing BIC/OLC in postmortem brains of non-
epileptic individuals who died without clinical or
pathologic evidence of neurologic disease. The presence
of such cell collections in the normal brain suggests that
glioneuronal hamartias might not have a direct causative
role in temporal lobe epilepsy and that immature neural
cells are more abundant in the adult human brain than
previously thought.
MATERIALS AND METHODS
Tissue Procurement and Processing
Tissues were collected and processed according to protocols
approved by the Institutional Review Board of the University
of Florida. Postmortem adult human brains from 28 patients
ranging in age from 20 to 83 years (mean
5
53 years) were
studied. In addition, the brains of 2 neonates (gestational ages
5
35 and 39 weeks) and a 3-year-old child were examined.
There was no evidence of epilepsy or any other neurologic dis-
ease by review of the clinical record. The brains showed no
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

114
YACHNIS ET AL
J Neuropathol Exp Neurol, Vol 59, February, 2000
significant pathologic changes by routine gross and microscopic
postmortem examination by a neuropathologist (AY).
Particular attention was focused on the region of the caudal
amygdala at or near the level of the mammillary bodies since
hamartias containing immature-appearing BIC were previously
identified in this location in temporal lobe resections from pa-
tients with intractable epilepsy (12). Other areas examined in
selected cases included middle and rostral levels of the amyg-
dala, hippocampus at or near the level of the lateral geniculate
nucleus, basal forebrain including olfactory tubercle, olfactory
bulbs and tracts, and multiple periventricular sites including
head of caudate, anterior thalamus, pulvinar, midbrain, pons,
and medulla. Formalin-fixed, paraffin-embedded tissue sections
from each of the above regions were studied by routine H&E
histology and by immunohistochemistry. Ten serial sections of
each tissue block were typically prepared. The first section was
stained with H&E and subsequent sections were immunostained
for each of the 8 antigens described below. The tenth section
served as a negative control (primary antibody excluded).
Antibodies
Bcl-2 was probed using a well-characterized anti-bcl-2 pep-
tide monoclonal antibody (clone 124, DAKO, Carpinteria, CA)
which was raised against a synthetic peptide corresponding to
amino acids 41–54 of human bcl-2 protein (18). This peptide
spans a region of the protein that is poorly conserved among
members of the bcl-2 family thus reducing the chance of cross-
reactivity. Antigen specificity was previously determined by
Western immunoblotting of postmortem human temporal lobe
tissue (19). A monoclonal antibody that recognizes neuron-spe-
cific class III beta tubulin (clone TuJ1) was obtained from Re-
search Diagnostics Inc. (Flanders, NJ). MAP-2 (microtubule as-
sociated protein-2), neurofilament (clone RMdO20), and Ki67
(MIB1; clone 7B11) were purchased from Zymed (South San
Francisco, CA). The sources of other antibodies used in this
study are as follows: GFAP (glial fibrillary acidic protein;
DAKO), Neu-N (neuronal nuclear antigen; Chemicon, Teme-
cula, CA), and GAP-43 (growth-associated protein-43; Boeh-
ringer Mannheim, Indianapolis, IN).
Immunohistochemistry
Following deparaffinization, endogenous peroxidase activity
was quenched with 0.3% aqueous H
2
O
2
for 5 min. Antigen re-
trieval was accomplished by 15 min immersion in boiling 0.01
M citrate buffer (pH 6) according to a previously described
protocol (20). The avidin-biotin-peroxidase technique was per-
formed according to an established protocol (DAKO LSAB 2
Kit). Primary antibody reactions consisted of overnight incu-
bations at 4
8
C. PBS (pH 7.4) containing 0.3% triton X-100
(PBST) with 1% BSA was used for diluting all antibodies and
PBST was used for all washes. Sections were lightly counter-
stained with hematoxylin. Normal adult human tonsil was used
as the positive control tissue for bcl-2 immunostaining, while
normal adult cerebellum or hippocampus was the positive con-
trol for class III beta tubulin, neurofilament protein, MAP-2,
GAP-43, GFAP and Neu-N. Omission of primary antibody was
the negative control in routine studies.
To confirm the specificity of the bcl-2 monoclonal antibody
in tissue sections, the primary antibody was preincubated with
a 10-fold excess of immunizing peptide prior to immunohisto-
chemical staining. Preadsorption abolished the immunoreactiv-
ity of bcl-2 in human tonsil and control human fetal brain tis-
sue. Preincubation of the bcl-2 antibody with a bcl-x
L
peptide
did not block immunohistochemical staining (12, 13).
Statistical Analysis
The relative abundance of BIC (‘‘BIC score’’) was estimated
in a semiquantitative fashion for each of the 28 adult patients
as follows: (0) no BIC identified; (1) rare clusters of BIC or
rare, individual cells; (2) scattered clusters of BIC; (3) abundant
BIC. To determine if there was an age-related decline in BIC
prevalence, the patients were split into 3 age groups: 20–44
years old (n
5
9), 45–65 years old (n
5
8), and
.
65 years old
(n
5
11). The mean BIC score for each age group was calcu-
lated. Data were analyzed by the Kruskal-Wallis test; a non-
parametric equivalent of the one-way analysis of variance by
ranks. The average postmortem interval for each of the above
groups was 14 hours (20–44 years old); 12 hours (45–65 years
old); and 13 hours (
.
65 years old).
RESULTS
Identification of Hamartia-like BIC in the Adult Human
Amygdala
Microscopic, hamartia-like cell collections containing
BIC were identified in 23 of 28 (82%) of the nonepileptic
adults studied. Such collections consisted of small cells
with round to slightly oval, hyperchromatic nuclei and
indistinct cytoplasmic boundaries on routine H&E stain-
ing (Fig. 1A). The nuclei of these cells often contained
a single, small basophilic nucleolus (Fig. 1A). Such col-
lections consisted of closely packed cell clusters or were
more loosely dispersed. The small round cells were often
associated with mature neurons (Fig. 1A, B). Bcl-2 was
strongly immunoreactive in the perinuclear cytoplasm of
the small round cells (Fig. 1B). Bcl-2 immunostaining
revealed that these cells had fine processes that extended
into the surrounding neuropil (Fig. 1B). Occasionally, a
suggestion of bipolar morphology was apparent in some
BIC. Although most BIC processes were oriented in a
haphazard fashion, some processes were aligned in par-
allel bundles. Mature neurons associated with BIC were
typically negative or, rarely, very weakly reactive for
bcl-2.
The ependyma showed weak to moderate levels of bcl-
2 immunoreactivity. In the hippocampus, only the epen-
dyma and choroid plexus were weakly to moderately im-
munoreactive for bcl-2. Neuronal elements of the
hippocampus (including the dentate gyrus) showed no de-
tectable immunoreactivity for bcl-2, consistent with prior
observations (12). Bcl-2 staining of tissue sections was
blocked by preincubation of the primary antibody in a
10-fold excess of immunizing peptide.
Antigen Expression and Proliferative Potential of BIC
The immunophenotype of BIC was evaluated by im-
munohistochemistry. Strong immunoreactivity for the
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

115
IMMATURE NEURONS OF ADULT BRAIN
J Neuropathol Exp Neurol, Vol 59, February, 2000
neuronal nuclear antigen (Neu-N) was identified in most
of the small round BIC, although others were negative
(Fig. 1C). The perinuclear cytoplasm and cell processes
of BIC were strongly immunoreactive for neuron-specific
class III
b
tubulin (Fig. 1D) and MAP-2. GAP-43 was
diffusely immunoreactive in the neuropil of the amygdala
and other mature gray matter structures, but definitive
staining of the perinuclear cytoplasm or processes of BIC
could not be confirmed. BIC were negative for neurofil-
ament proteins, and GFAP.
As we previously showed in temporal lobe resections
from patients with epilepsy (12), BIC of hamartia-like
cell collections in nonepileptic individuals were consis-
tently negative for Ki67 (MIB1), suggesting they are
postmitotic or quiescent. However, rare, individual Ki67-
immunoreactive cells were identified in the immediate
subependymal region. Such cells appeared immature,
contained little discernible cytoplasm, and, in contrast to
BIC, had irregular, hyperchromatic nuclei without nucle-
oli.
Postnatal Ontogeny of BIC in the Human Amygdala
In term infants, abundant bcl-2 immunoreactivity was
identified in neuroepithelial cells of the ventrolateral re-
gion of the amygdala and adjacent periamygdalar region
including the corticoamygdaloid transition area (21) (Fig.
2A, B). BIC were arranged as a discrete band of imma-
ture cells in the subventricular zone while primitive cells
of the ventricular zone were negative for bcl-2 (Fig. 2A,
B). Conspicuous clusters of BIC were still present in the
subventricular zone and in the deep layers of the peria-
mygdalar cortex in young adults (Fig. 2C, D). Occasion-
ally, some groups of BIC were arranged in a palisading
pattern (Fig. 2D). Although such clusters were identified
in all patients under 40 years of age, they were particu-
larly abundant in patients less than 30-years of age. BIC
were less prominent in individuals 50 years of age and
older, often existing as small, individual cell clusters in
the subventricular gray matter (Fig. 2E, F). Four of 17
individuals greater than 50 years of age did not have
detectable BIC in the material studied while only 1 of 10
patients less than 50 years of age had no detectable BIC.
(The 1 BIC-negative individual was 49 years old.) Inter-
estingly, all 5 of the BIC-negative cases were males. Ap-
parent absence of BIC could not be attributed to poor
fixation or failed antigen retrieval since the other antigens
tested stained appropriately. Presence of BIC in these cas-
es cannot be entirely excluded since serial sections of the
entire amygdaloid region were not studied.
To determine if there was a significant age-related de-
cline in BIC during adulthood, the patients were divided
into 3 age groups: 20–44, 45–65, and
.
65 years old. A
mean ‘‘BIC score’’ for each age group was calculated
and the data were analyzed by a nonparametric method.
This demonstrated a statistically significant reduction in
these cells with age (Fig. 3). These observations suggest
that BIC are abundant in the amygdala and periamyg-
daloid cortex during infancy, persist into adulthood, and
are gradually reduced throughout adult life.
Anatomic Location of BIC
BIC were mainly identified in the subventricular region
of the caudal amygdala adjacent to the intermediate and
parvocellular divisions of the basal nucleus (22); the lat-
ter being referred to by Gloor as the paralaminar nucleus
(21). Most BIC were located in the gray matter of the
amygdalar and periamygdalar subventricular zone and
were not in direct contact with the ependymal lining.
However, a few discrete clusters of such cells were oc-
casionally situated directly beneath or within the epen-
dymal lining (Fig. 2D). Microscopic, BIC-containing cell
clusters were often observed near the periphery of the
medial and rostral amygdala, but usually in an inferior or
medial location near the rostral aspect of the ventricle.
With bcl-2 staining, these cells could be identified near
the gray-white junction and in the deep layers of the per-
iamygdaloid cortex; including the corticoamygdaloid
transition area (21). BIC were often observed along ru-
dimentary collections of ependymal cells that extend
from the rostromedial angle of the temporal horn of the
lateral ventricle between the amygdala and corticoam-
ygdaloid transition area.
BIC were not detected in the following regions of the
adult human brain: hippocampus (including dentate nu-
cleus), basal forebrain (including olfactory tubercle), ol-
factory bulbs and tracts, and caudate nucleus, anterior
thalamus, pulvinar, and periventricular regions of the
midbrain, pons, and medulla. Clusters of small round im-
mature-appearing cells were identified along the amyg-
dalofugal pathway in the basal forebrain, but these were
consistently negative for bcl-2.
DISCUSSION
We have identified BIC of the human amygdala and
periamygdalar cortex that are similar if not identical to
‘‘oligodendrocyte-like cells’’ (OLC) of glioneuronal ha-
martias that have been described in resected temporal
lobes from patients with intractable epilepsy (12, 16, 17).
We previously reported that BIC/OLC could be found in
the region of the intermediate and parvocellular divisions
of the basal amygdaloid nucleus in seizure patients with
pathologically confirmed Ammon’s horn sclerosis (12).
The identification of BIC in the same locations in non-
seizure patients suggests that these immature-appearing
cells do not always play a primary role in epileptogenesis
associated with mesial temporal sclerosis. Furthermore, a
histologic finding of BIC-containing glioneuronal hamar-
tias in the caudal amygdala and/or periamygdalar cortex
of an otherwise typical case of Ammon’s horn sclerosis
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

116
YACHNIS ET AL
J Neuropathol Exp Neurol, Vol 59, February, 2000
Fig. 1. Morphology and cell lineage of bcl-2-immunoreactive cells (BIC). A: H&E-stained section of amygdala from a 35-
year-old nonepileptic individual. This microscopic cell cluster is composed of small round immature-appearing cells with ‘‘oli-
godendrocyte-like’’ perinuclear haloes. Some of the small cells have basophilic nucleoli (small arrow). Mature neurons are
intermixed with the immature-appearing cells (large arrow). The histologic appearance is similar, if not identical, to glioneuronal
hamartias identified in some temporal lobe resections from patients with epilepsy. Original magnification:
3
1200. B: Bcl-2
immunostained section from the same patient as in ‘‘B’’ showing strong immunoreactivity of the perinuclear cytoplasm and
delicate cell processes of the immature-appearing cells (small arrow). Mature neurons admixed with the small cells are bcl-2-
negative (large arrow). Original magnification:
3
1200. C: Loose hamartia-like cluster of cells immunostained for the neuronal
nuclear antigen (Neu-N). Many, but not all, of the small cells show strong nuclear staining. Several mature neurons (bottom right
of center and left side of photo) also are immunostained. Original magnification:
3
800. D: Cells corresponding to BIC show
strong cytoplasmic immunoreactivity for neuron-specific class III
b
tubulin. Abundant cell processes of the neuropil are also
immunoreactive. Original magnification:
3
800.
should probably not suggest presence of ‘‘dual patholo-
gy.’’
The apparent abundance of BIC in some temporal lobe
resections might be explained by the fact that patients
undergoing epilepsy surgery are typically young to mid-
dle-aged adults. Our current study showed that BIC are
more conspicuous in young (nonepileptic) individuals
and that the relative abundance of these cells is reduced
with age. Several factors could explain why BIC have
not been identified in prior retrospective studies in which
autopsy control temporal lobes have been compared with
epilepsy-associated temporal lobe pathology (12, 16, 17):
1) The caudal amygdala is not usually sampled in the
course of routine neuropathologic examination. The re-
gion of the caudal amygdala was not studied in nonepi-
leptic adult controls in our previous developmental in-
vestigation (12) and this area was not specifically
mentioned as having been examined in nonseizure autop-
sy controls in the studies of Wolf et al (16) and Kasper
et al (17). 2) The paucity of BIC in older individuals
makes their identification difficult without bcl-2 staining.
BIC could have been easily missed in several elderly pa-
tients in our study had immunohistochemistry not been
performed. Also, the apparent absence of BIC in 5 cases
presented herein could be due to insufficient sampling as
serial sections of the entire amygdala and periamygdaloid
region was not performed. 3) The precise anatomic lo-
cation of hamartias in many temporal lobe resection spec-
imens is often difficult to determine due to poor orien-
tation or fragmentation of the tissue in addition to the
rather nondescript cytoarchitecture of the amygdala and
periamygdaloid cortex.
An association of glioneuronal hamartias and gangliog-
liomas has been reported (23). We observed 2 cases (one
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

117
IMMATURE NEURONS OF ADULT BRAIN
J Neuropathol Exp Neurol, Vol 59, February, 2000
Fig. 2. Postnatal ontogeny BIC in the human amygdala. A: Mesial temporal lobe of 39-week gestational age human neonate
at a caudal amygdala (Am) level. A prominent layer of bcl-2-immunoreactivity is present in the subventricular zone of the basal
amygdala (small arrows) adjacent to the temporal horn of the lateral ventricle (V). BIC extend into the corticoamygdaloid transition
area (21) (large arrow) and periamygdalar cortex. Original magnification:
3
100. B: High magnification of (A) showing bcl-2-
immunoreactive cells of the subventricular zone. Note that primitive cells of the ventricular zone (VZ) are negative for bcl-2.
Original magnification:
3
800. C: Section from caudal amygdala (Am) of 21-year-old female without neurologic disease showing
BIC adjacent to the ventricle (arrow). BIC extend medially along the periphery of the amygdala towards the corticoamygdaloid
transition area (arrowheads). Original magnification:
3
100. D: Higher magnification of (C) showing abundant subependymal
(small arrow) and subventricular (large arrows) BIC. The latter are arranged in a palisading pattern. Original magnification:
3
400.
E: A rare, focal subependymal ‘‘hamartia-like’’ cluster of BIC in the caudal amygdala of a 61-year-old man without neurologic
disease. This was the only collection of BIC identified in this case. Original magnification:
3
200. F: High magnification of focal
BIC from (E). The ependymal lining is at the bottom of the figure. Original magnification:
3
800.
reported previously in reference 12, the other unpub-
lished) in which a temporal lobe ganglioglioma occurred
with an apparent excess of BIC. This suggested that ab-
normal differentiation of BIC into mature neurons might
be involved the pathogenesis of some gangliogliomas.
Although both tumors were reportedly from the region of
the uncus, we could not determine the exact location
of these neoplasms or their relationship to the caudal
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

118
YACHNIS ET AL
J Neuropathol Exp Neurol, Vol 59, February, 2000
Fig. 3. Age-related decline in BIC during adulthood. The
relative abundance of BIC was expressed as a ‘‘BIC score’’ for
each of the 28 adult patients as follows: (0) no BIC identified;
(1) rare clusters of BIC, or rare individual cells; (2) scattered
clusters of BIC; (3) abundant BIC. Patients were divided into
3 age groups: 20–44 (n
5
9; mean BIC score
5
2.6), 45–65
(n
5
8; mean BIC score
5
1.25), and
.
65 years old (n
5
11;
mean BIC score
5
0.73). Data were analyzed by the Kruskal-
Wallis test; a nonparametric equivalent of the one-way analysis
of variance by ranks. This showed a statistically significant re-
duction in BIC with age (p
,
.001).
amygdala since both cases were lesionectomies. These
tumors arose in individuals who were less than 10 years
of age and our current results suggest that BIC are par-
ticularly abundant in the temporal lobes of children.
However, the possibility that some cases of temporal lobe
lesions may be associated with qualitative or quantitative
abnormalities of BIC is not ruled out by the current work.
Wolf et al (16) have shown that the OLC of glioneu-
ronal hamartias from temporal lobe resections, which we
propose are identical to BIC of nonsiezure individuals,
are immunoreactive for fetal NCAM. Our findings that
BIC are strongly immunoreactive for Neu-N, class III
b
tubulin, and MAP-2 support a neuronal lineage for these
cells. Neu-N has been detected in early postmigratory
neurons of the human fetal brain as early as 8 weeks
gestational age (24). Similarly, class III
b
tubulin is de-
velopmentally regulated in the human brain (25) and is
one of the earliest antigens to appear in postmitotic neu-
rons (26, 27). Since the expression of neurofilament pro-
teins (NFP) generally correlates with a more mature neu-
ronal phenotype, the negativity for NFP displayed by BIC
further supports an early neuronal lineage for these cells.
The strong bcl-2 immunoreactivity of BIC is itself
reminiscent of the strong expression of this antigen in
early postmigratory young neurons in the fetal human
neocortex (12). Specifically, between 20 and 40 weeks of
gestation, bcl-2 immunoreactivity is progressively lost in
the deeper layers of the cortex, which contains neurons
that have migrated into the cortical plate earlier in de-
velopment. During this period, bcl-2 is persistently ex-
pressed in superficial layers (especially layer 2) which
contain less mature, postmitotic neurons that migrated
more recently into the cortex. Bcl-2 down-regulation co-
incides with periods of synaptogenesis and neuronal dif-
ferentiation in the developing human neocortex (12) and
spinal cord (13). The periventricular location of BIC fur-
ther supports the idea that these cells remained near their
site of origin (16), persisted into adulthood, and perhaps
exist in a vestigial or resting stage. Another possibility,
suggested by our finding of rare Ki67-immunoreactive
cells in the ependymal region, is that there is limited pro-
duction of BIC in the adult.
The pattern and distribution of BIC that we have de-
scribed in nonepileptic humans is similar to that de-
scribed for immature, bcl-2-immunoreactive cells in the
amygdala and piriform cortex of the adult squirrel mon-
key (Saimiri scureus) (28, 29). These immature cells dis-
played features of 2 types of neurons: type ‘‘A’’ (imma-
ture) and type ‘‘B’’ (differentiating). Type ‘‘A’’ neurons
were intensely immunoreactive for bcl-2 (28, 29) and
were morphologically similar to the immature-appearing
BIC of the normal human amygdala (present study) and
to OLC observed in ‘‘hamartias’’ that in human temporal
lobe resections (12, 16, 17, 28). Type ‘‘A’’ cells of the
primate became less numerous with age. In contrast, type
‘‘B’’ neurons were larger than type ‘‘A’’ cells, were less
intensely reactive for bcl-2, and became more numerous
with age. These results suggested that the smaller, in-
tensely bcl-2-immunoreactive cells might have differen-
tiated into a more mature neuronal phenotype during nor-
mal aging. Although cells similar to type ‘‘B’’ neurons
were not evident in our study, we did observe a reduction
in BIC (similar to type ‘‘A’’ cells) during normal aging.
Type ‘‘A’’ BIC were more widespread in limbic structures
of the squirrel monkey than in the human. Although im-
mature-appearing clusters of small cells were observed
along the limbic-olfactory pathway in our study, such
cells were negative for bcl-2. The reasons for these in-
terspecies differences are not clear. One possibility is that
several different subtypes of immature neurons exist nor-
mally in the limbic system and not all subtypes express
detectable amounts of bcl-2 in the human. Nevertheless,
the primate model provides further evidence that an im-
mature population of neurons exists in the adult amygdala
and periamygdaloid cortex and that these cells may de-
velop into mature, fully differentiated CNS neurons.
Taken together, our findings provide evidence that cells
with features of immature, postmitotic neurons exist in
the adult human temporal lobe and that BIC may repre-
sent an intermediate stage of neuronal development be-
tween proliferating progenitor and mature neuron. Fur-
ther studies are needed to determine if BIC have a
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from

119
IMMATURE NEURONS OF ADULT BRAIN
J Neuropathol Exp Neurol, Vol 59, February, 2000
physiologic function, if their production or differentiation
is affected by disease processes, or if such cells could be
manipulated to produce functional restoration of the dam-
aged nervous system.
AKNOWLEDGMENTS
The authors wish to thank Ms. Elaine Dooley for outstanding tech-
nical support. These studies were supported by the University of Florida
College Incentive Fund and the Howard Hughes Medical Institute Re-
search Resources Program of the University of Florida College of Med-
icine.
REFERENCES
1. Davies AM. The bcl-2 family of proteins and the regulation of
neuronal survival. TINS 1995;18:355–58
2. Reed JC. Double identity for proteins of the bcl-2 family. Nature
1997;387:773–76
3. Allsopp TE, Wyatt S, Paterson HF, Davies AM. The protooncogene
bcl-2 can selectively rescue neurotrophic factor-dependent neurons
from apoptosis. Cell 1993;73:295–307
4. Zhoung L-T, Sarafian T, Kane DJ, et al. Bcl-2 inhibits death of
central neural cells induced by multiple agents. Proc Natl Acad Sci
1993;90:4533–37
5. Martinou JC, Dubois-Dauphin M, Staple JK, et al. Overexpression
of bcl-2 in transgenic mice protects neurons from naturally occuring
cell death and experimental ischemia. Neuron 1994;13:1017–30
6. Farlie PG, Dringen R, Rees SM, Kannouralis G, Bernard O. Bcl-2
transgene expression can protect neurons against developmental
and induced cell death. Proc Natl Acad Sci 1995;92:4397–4401
7. Dubois-Dauphin M, Frakowski H, Tsujimoto Y, Huarte J, Martinou
J-C. Neonatal motoneurons overexpressing the bcl-2 protooncogene
in transgenic mice are protected from axotomy-induced cell death.
Proc Natl Acad Sci 1994;91:3309–13
8. Raghupathi R, Fernandez SC, Murai H, et al. Bcl-2 overexpression
attenuates cortical cell loss after traumatic brain injury. J Cerebr
Blood Flow Metabol 1998;18:1259–69
9. Merry DE, Veis DJ, Hickey WF, Korsmeyer SJ. Bcl-2 protein ex-
pression is widespread in the developing nervous system and re-
tained in the adult peripheral nervous system. Development
1994;120:301–11
10. Veis-Novack D, Korsmeyer SJ. Bcl-2 protein expression during
murine development. Am J Pathol 1994;145:61–73
11. Shimohama S, Fujimoto S, Sumida Y, Tanino H. Differential ex-
pression of rat brain bcl-2 family proteins in development and ag-
ing. Biochem Biophys Res Com 1998;252:92–96
12. Yachnis AT, Powell SZ, Olmsted JJ, Eskin TA. Distinct neurode-
velopmental patterns of bcl-2 and bcl-x expression are altered in
glioneuronal hamartias of the human temporal lobe. J Neuropathol
Exp Neurol 1997;56:186–98
13. Yachnis AT, Giovanini MA, Eskin TA, Reier PJ, Anderson DK.
Developmental patterns of bcl-2 and bcl-x polypeptide expression
in the human spinal cord. Exp Neurol 1998;150:82–97
14. Motoyama N, Wang F, Roth KA, et al. Massive cell death of im-
mature hematopoetic cells and neurons in Bcl-x-deficient mice. Sci-
ence 1995;267:1506–10
15. Parsadanian AS, Cheng Y, Keller-Peck CR, Holtzman DM, Snider
WD. Bcl-xL is an antiapoptotic regulator for postnatal CNS neu-
rons. J Neurosci 1998;18:1009–19
16. Wolf HK, Wellmer J, Muller MB, Wiestler OD, Hufnagel A, Pietsch
T. Glioneuronal malformative lesions and dysembryoplastic neu-
roepithelial tumors in patients with chronic pharmacoresistant epi-
lepsies. J Neuropathol Exp Neurol 1995;54:245–54
17. Kasper BS, Stefan H, Buchfelder M, Paulus W. Temporal lobe mi-
crodysgenesis in epilepsy versus control brains. J Neuropathol Exp
Neurol 1999;58:22–28
18. Pezzella F, Tse AG, Cordell JL, Pulford KAF, Gatter KC, Mason
DY. Expression of the bcl-2 oncogene protein is not specific for the
t(14;18) chromosomal translocation. Am J Pathol 1990;137:225–32
19. Yachnis AT, Fancey JT, Eskin TA. Bcl-2 immunostaining identifies
‘‘hamartias’’ of the caudal amygdala in non-epileptic patients. (Ab-
stract) J Neuropathol Exp Neurol 1998;57:514
20. Yachnis AT, Trojanowski JQ. Studies of childhood brain tumors
using immunohistochemistry and microwave technology: Method-
ological considerations. J Neurosci Meth 1994;55:191–200
21. Gloor P. The temporal lobe and limbic system. New York: Oxford
University Press, 1997:591–721
22. Sorvari H, Soininen H, Paljarvi L, Karkola K, Pitkanen A. Distri-
bution of parvalbumin- immunoreactive fibers in the human amyg-
daloid complex. J Comp Neurol 1995;185–212
23. Wolf HK, Muller MB, Spanle M, Zentner J, Schramm J, Wiestler
OD. Ganglioglioma: A detailed histopathological and immunohis-
tochemical analysis of 61 cases. Acta Neuropathol 1994;88:166–73
24. Sarnat HB, Nochlin D, Born DE. Neuronal nuclear antigen (Neu-
N): A marker of neuronal maturation in early human fetal nervous
system. Brain Dev 1998;20:88–94
25. Katsetos CD, Frankfurter A, Chrstakos S, Mancall EL, Vlachos IN,
Urich H. Differential localization of class III beta-tubulin isotype
and calbindin-D28k defines distinct neuronal types in developing
human cerebellar cortex. J Neuropathol Exp Neurol 1993;52:655–66
26. Lee MK, Tuttle JB, Rebhun LI, Cleveland DW, Frankfurter A. The
expression and posttranslational modification of a neuron-specific
beta-tubulin isotype during chick embryogenesis. Cell Motil Cy-
toskel 1990;17:118–32
27. Easter SS, Ross LS, Frankfurter A. Initial tract formation in the
mouse brain. J Neurosci 1993;13:285–99
28. Bernier PJ, Parent A. Bcl-2 as a marker of neuronal immaturity in
postnatal primate brain. J Neurosci 1998;18:2486–97
29. Bernier PJ, Parent A. The anti-apoptosis bcl-2 proto-oncogene is
preferentially expressed in limbic structures of the primate brain.
Neuroscience 1998;82:635–40
Received August 5, 1999
Revision received November 3, 1999
Accepted November 4, 1999
by guest on July 6, 2016http://jnen.oxfordjournals.org/Downloaded from