-
Neuron, Vol. 15, 649-661, September, 1995, Copyright © 1995 by
Cell Press
Role of BCL-2 in the Survival and Function of Developing and
Mature Sympathetic Neurons
Laura J. S. Greenlund,* Stanley J. Korsmeyer,~r and Eugene M.
Johnson, Jr.* *Department of Neurology Department of Molecular
Biology and Pharmacology tHoward Hughes Medical Institute
Washington University School of Medicine St. Louis, Missouri
63110
Summary
Sympathetic neurons, when placed in culture during the period of
naturally occurring cell death, will die by apoptosis when deprived
of nerve growth factor (NGF). In this system, the mRNA levels of
the BCL-2 family members decrease after NGF deprivation and during
apoptosis. Sympathetic neurons from BCL-2-deficient mice died more
rapidly after NGF deprivation than neu- rons from wild-type
littermates. Sympathetic neurons of adult animals are relatively
independent of NGF for survival. If sympathetic neurons are
maintained in vitro for several weeks, loss of acute trophic factor
dependence develops with a time course similar to that seen in the
intact animal. Examination of neurons from BCL-2-deficient mice
showed that BCL-2 expression is not required for the development of
trophic factor independence. Therefore, BCL-2 is an important regu-
lator of the survival of sympathetic neurons after NGF deprivation
during the period of naturally occurring programmed neuronal death,
but BCL-2 is not involved in the development of trophic factor
independence in mature sympathetic neurons.
Introduction
The development of the nervous system begins with the birth of
about twice as many neurons as are finally present in the adult. In
the sympathetic nervous system, the imma- ture neuronal precursors
are not dependent on nerve growth factor (NGF) for survival
(Coughlin and Collins, 1985; Birren et al., 1993; DiCicco-Bloom et
al., 1993). Sym- pathetic neurons then become acutely dependent on
NGF for survival, and neurons that do not receive sufficient neu-
rotrophic factor undergo programmed neuronal death. In the rat, the
period of NGF dependence begins about em- bryonic day 19.5 (Birren
et al., 1993). Following the period of programmed neuronal death,
trophic factor depen- dence decreases progressively, so that adult
sympathetic neurons are much tess acutely dependent on NGF for
survival, but still respond to NGF with increased growth and will
die only upon prolonged periods of trophic factor deprivation
(Gorin and Johnson, 1980). This develop- mental process allows for
selecting the appropriate num- ber of neurons to match the target
size in the developing animal, for ridding the nervous system of
inappropriate connections (Oppenheim, 1991), and later, for
stabilizing
the nervous system of the adult animal. Since most neu- rons are
irreplaceable in the adult, it is imperative that they survive
injury, such as peripheral nerve axotomy, so that regeneration can
occur.
The death of sympathetic neurons after NGF deprivation (Martin
et al., 1988; Deckwerth and Johnson, 1993; Ed- wards and Tolkovsky,
1994) is characterized by changes that are apoptotic (Kerr et al.,
1972), including fragmenta- tion of DNA into oligonucleosomes,
shrinkage of the cell with preservation of the organelles, and
blebbing in the cytoplasm. In many cases, apoptosis appears to be
an active process wherein the expression of certain genes is
required for cell death, because in sympathetic neurons (Martin et
al., 1988) and in other cell types (Tata, 1966; Pratt and Greene,
1976; Cohen and Duke, 1984), the pro- cess can be blocked by
inhibitors of RNA and protein syn- thesis.
BCL-2, a mitochondrial and perinuclear membrane pro- tein
(Hockenbery et al., 1990; Monaghan et al., 1992; Ja- cobson et al.,
1993), and its family members BAX (Oltvai et al., 1993) and BCL-X
(long and short forms; Boise et al., 1993) modulate the sensitivity
of cells to death. BCL-2 overexpression in the interleukin-3
(IL-3)-dependent B cell line, FL5.12, blocks death induced by
trophic factor depri- vation; however, if, in addition, BAX is
overexpressed, the death repressor activity of BCL-2 is
antagonized. If BAX is overexpressed independently, death induced
by IL-3 deprivation is accelerated (Oltvai et al., 1993). The
BCL-2- related protein, BCL-X, is produced in a long or a short
form by alternative mRNA splicing. FL5.12 cells are pro- tected
from death induced by trophic factor deprivation if BCL-XL (the
long form of BCL-X) is overexpressed. In contrast, the short form
of BCL-X, BCL-Xs, acts much like bax and can antagonize the
protective effect of BCL-2 overexpression (Boise et al., 1993).
Dimerization is important for the function of BCL-2; BCL-2 can
form homodimers with itself or can form hetero- dimers with BAX
(Oltvai et al., 1993). Oltvai et al. (1993) propose that
BCL-2/BCL-2 homodimers protect cells from death, and that BAX
titrates the levels of these homodi- mere, rendering cells more
susceptible to death. In addi- tion, BCL-2 binds to a number of
other proteins including R-Ras (Fernandez-Sarabia and Bischoff,
1993); interac- tion with ras signal transduction pathways may be
one mode of BCL-2 action. At least three other proteins have been
identified that bind to BCL-2, including Nip1, Nip2, and Nip3 (Boyd
et al., 1994). Nip1 shows homology with the
Ca2÷/calmodulin-dependent phosphodiesterases, Nip2 shows homology
with the human GTPase-activating pro- tein RhoGAP, and Nip3 is a
novel protein. The interaction of BCL-2 with Nip1, Nip2, and Nip3
may be another means to modify intracellular signaling processes
(Boyd et al., 1994).
Although all of the mechanisms by which BCL-2 protects cells
from death remain to be elucidated, a number of functions have been
suggested. Two groups have shown
-
Neuron 650
that BCL-2 has antioxidant properties. Hockenbery et al. (1993)
have reported that BCL-2 overexpression in a T cell line blocks
lipid peroxidation induced by dexamethasone treatment. Kane et al.
(1993) have demonstrated that BCL-2 overexpression in a
hypothalamic neural cell line inhibits the generation of reactive
oxygen species after glutathione depletion with diethylmaleate.
BCL-2 may also alter Ca 2+ fluxes. Lam et al. (1994) examined a
lymphocyte cell line treated with thapsigargin, which inhibits the
endo- plasmic reticulum Ca 2+ pump. Cells that overexpressed BCL-2
released much less Ca 2÷ into the cytoplasm, indicat- ing that
BCL-2 reduces Ca 2÷ efflux through the endoplas- mic reticulum
membrane.
In addition, genetic evidence from the nematode Caenor- habditis
elegans suggests that the worm homolog of BCL- 2, ced-9 (Hengartner
and Horvitz, 1994), functions to block, either directly or
indirectly, the function of the puta- tive cysteine protease,
ced-3. Functional ced-3, a protein that shares homology with the
human cysteine protease IL-lJ~-converting enzyme (Yuan et al.,
1993) and with hu- man ich-1 (Wang et al., 1994), is required for
programmed cell death in C. elegans (Yuan and Horvitz, 1990, 1992);
however, increased ced-9 activity overrides ced-3 activity and
blocks cell death (Hengartner et al., 1992). The human Bcl-2 gene
can functionally replace C. elegans ced-9 to block cell death in
the worm (Vaux et al., 1992; Hengartner and Horvitz, 1994). Thus,
BCL-2 either directly or indirectly blocks the action of ced-3.
The developmental expression patterns of Bcl-2 in the CNS have
been analyzed by a number of groups. Data at both the RNA and
protein levels indicate that Bcl-2 is expressed at its highest
levels in the prenatal brain, but postnatal and adult animals,
including humans, express lower levels of Bcl-2 in the CNS
(Abe-Dohmae et al., 1993; Castren et al., 1994; Ferrer et al.,
1994; Merry et al., 1994). Very few studies have analyzed the
expression of Bcl-2 in the PNS. Merry et al. (1994) reported that
BCL-2 protein expression in the prenatal mouse dorsal root ganglion
(DRG) is retained into adulthood and that postnatal supe- rior
cervical ganglion (SCG) neurons also express BCL-2 protein even
into adulthood. Martinou et al. (1994b) dem- onstrated that SCGs
and DRGs from 10 week human fe- tuses also express BCL-2 protein.
Currently, data are sparse regarding Bcl-2 expression during
development of the PNS or the importance of this expression in
regulating neuronal survival.
Overexpression of BCL-2 in trigeminal neurons (Alisopp et al.,
1993) or sympathetic ganglion neurons (Garcia et al., 1992;
Greenlund et al., 1995) delays apoptosis induced by trophic factor
deprivation. In contrast, ciliary ganglion neurons are not
protected by BCL-2 overexpression from apoptosis induced by ciliary
neurotrophic factor (CNTF) deprivation, suggesting that more than
one cell death pathway exists within PNS neurons (AIIsopp et al.,
1993). PC12 cells, a sympathetic neuron-like cell line, stably
transfected with Bcl-2 are also more resistant to apoptosis in
response to various insults (Batistatou et al., 1993; Mah et al.,
1993). Endogenous expression of BCL-2 in sympa- thetic neurons
clearly occurs, and overexpression of
BCL-2 renders the neurons more resistant to death. How- ever,
very little evidence is apparent that BCL-2 is im- portant for the
normal regulation of sympathetic neuronal survival, either in the
embryo, during the period of naturally occurring cell death, or in
the adult, when neurons have become more resistant to trophic
factor deprivation-in- duced death.
We used an in vitro model (Martin et al., 1988) for the
developmental programmed neuronal death of rat and mouse
sympathetic neurons to examine the mRNA levels of the Bcl-2 family
members during apoptosis. We hypoth- esized that immature
sympathetic neurons may initiate apoptosis after NGF deprivation
either by up-regulating the expression of bax or BcI-Xs or by
down-regulating the expression of Bcl-2 or Bcl-XL. We also examined
sympa- thetic neurons from BCL-2-deficient mice to determine
whether BCL-2 is a regulator of neuronal survival after trophic
factor deprivation.
An in vitro model for the loss of acute trophic factor
dependence that occurs with maturation in sympathetic neurons was
used to examine the importance of BCL-2 in the development of this
independence. We hypothesized that mature neurons could
down-regulate the expression of bax and BcI-Xs or up-regulate the
expression of Bcl-2 and Bcl-XL to become more resistant to
apoptosis. Sympa- thetic neurons from BCL-2-deficient animals were
used to examine whether BCL-2 expression is important for the
development of trophic factor independence.
Lastly, we used microinjection to overexpress BCL-2 in
sympathetic neurons to determine the time course of survival after
NGF deprivation and to examine whether BCL-2 overexpression could
mimic the effects of NGF as evidenced by blocking the early fall in
protein synthesis rates that occurs after NGF deprivation of
sympathetic neurons (Deckwerth and Johnson, 1993).
Results
Immunostaining Shows BCL-2 Protein Is Present in Both Neurons
and Nonneuronal Cells BCL-2 is localized in the mitochondria of
lymphoid cells (Hockenbery et al., 1990) and in the perinuclear
mem- brane and endoplasmic reticulum of lymphoid and other cell
types (Chen-Levy et al., 1989; Monaghan et al., 1992; Jacobson et
al., 1993). Merry et al. (1994), using the mono- clonal anti-BCL-2
antibody 3F11, have shown that in vivo both neurons and supporting
cells in the SCG of postnatal day 20 (P20) mice express BCL-2. The
pattern of staining in neurons is heterogeneous, and nuclear
staining is ob- served. We first determined how the pattern of
BCL-2 ex- pression in the in vitro system compared to the pattern
of BCL-2 expression in vivo. In this system, cultures of
dissociated rat SCG cells are prepared the day before birth of the
animal, just before the period of naturally occurring neuronal
death in the rat SCG (Wright et al., 1983). For immunostaining,
dissociated SCG cells that were not pre- plated to remove
nonneuronal cells (see Experimental Procedures for details) were
used so that Schwann cells and fibroblasts would be more numerous
and could be
-
Endogenous BCL-2 Regulates Neuronal Survival 651
D E F 4
Figure 1. Immunohistochemistry for BCL-2 on SCG Neurons and
Supporting Cells in Culture Photomicrographs illustrating the
pattern of BCL-2 expression in vitro for SCG neurons, Schwann
cells, and fibrobiasts. After 1 week in culture, SCG neurons and
supporting cells were stained with the 3F11 antibody or a control
hamster monoclonal antibody. (A-C) Neurons maintained in the
presence of NGF and incubated with the control antibody show no
staining (A). Neurons stained with the 3Fll anti-BCL-2 antibody
show both cytoplasmic and nuclear staining with exclusion of the
nucleoli (B). After 15 hr of NGF deprivation, neurons incubated
with the 3F11 antibody show significantly less intense staining in
both the nucleus and cytoplasm (C). (D-F) Lower power
photomicrographs of in vitro SCG cells, each showing two fields of
cells. Neurons (open arrows) and nonneuronal cells incubated with a
control hamster monoclonal antibody show no staining (D). In the
presence of NGF, Schwann cells (small arrows), fibroblasts (large
arrows), and neurons stained with 3Fl l (E). Neurons deprived of
NGF for 15 hr show much less intense staining with 3Fl l (F);
however, nonneuronal cells stained with the same intensity as in
the presence of NGF. Bars, 15 I~m.
clearly visualized. Rat neuronal cultures that had been
maintained in NGF for I week were stained with the 3F11 antibody.
Immunohistochemistry showed light staining in both the neurons and
nonneuronal support ing cells. The staining in the neurons was both
cytoplasmic and nuclear; however, the nucleoli appeared to be
excluded (Figures 1A and 1B). Both Schwann cells and fibroblasts
(Figures 1D and 1 E) were also positive for BCL-2; however,
staining was exclusively cytoplasmic in these cells. The intensity
of staining was similar in the neurons and support ing cells,
suggesting that BCL-2 protein levels are similar. The neu- ronal
staining appeared uniform in the cultures; this is the only notable
difference between the in vitro pattern and the in vivo pattern
described by Merry et al., which shows heterogeneous neuronal
staining. The in vitro system, therefore, closely mimics the in
vivo situation.
To determine how the BCL-2 protein levels or localiza- tion
changed after NGF deprivation, sympathet ic neurons were deprived
of NGF for 15 hr and stained. By 15 hr after NGF deprivation, DNA
fragmentat ion begins (Deckwerth and Johnson, 1993) and peak levels
of gene induction
occur for several apoptosis-associated genes (see below)
(Freeman et ai., 1994; Estus et al., 1994). After 15 hr of NGF
deprivation, the intensity of BCL-2 staining in the neurons was
significantly reduced, and the nuclei no longer appeared to stain
(Figure 1C). This is consistent with the observation that neuronal
protein synthesis rates decline rapidly after trophic factor
deprivation (Deckwerth and Johnson, 1993). No reduction in the
intensity of stain- ing was seen in the nonneuronal cells after NGF
depriva- tion (Figure 1F). These results suggest the possibil ity
that loss of BCL-2 protein, subsequent to NGF deprivation, plays a
role in the neuron reaching the point that it under- goes
apoptosis.
Expression of the BCL-2 Family Members Is Decreased during
Apoptosis Since apoptosis of sympathet ic neurons can be com.-
pletely blocked by the addition of protein synthesis or RNA
synthesis inhibitors, the expression of certain genes is probably
important for apoptosis to proceed. Recently, several genes have
been identif ied by RT-PCR that are
-
Neuron 652
m o
O o
O.
m o
C O o
=o
o.
Tyrosine Hydroxylase
0 5 10 15 20 25 30 48 (hrs)
0 5 10 I5 20 25 30 48
Time after NGF Deprivation (hrs)
BAX
0 10 15 25 40 (hrs)
BCL-2
0 5 10 15 20 25 35 4 8 (hrs)
t20
c-
o o
m o
i -
0 o
P O=
0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 8
Time after NGF Deprivation (hrs)
BCL-XL
0 5 10 15 20 25 3 0 4 8 6 7 (hrs)
Figure 2. RT-PCR Analysis of mRNAs for BCL-2 Family Members
during Neuronal Apoptosis In three independent experiments,
preplated SCG neurons, after 1 week in culture, were deprived of
NGF for the indicated number of hours, mRNA was isolated, and cDNA
was pre- pared. Messages were analyzed by PCR that included
radiolabeled dCTP. Products were separated by SDS-PAGE and
visualized by au- toradiography; band intensity was quantified by
Phosphorimager analysis. A representative autoradiograph is shown
for each message, and the mean and SEM for band intensity are
plotted for the three experiments (single aster- isk, p < .05;
double asterisk, p < .01; ANOVA with Dunnett multiple
comparisons test).
o 10 15 20 25 30 48 0 5 10 15 20 25 30 48
Time after NGF Deprivation (hrs) Time after NGF Deprivation
(hrs)
markedly~ induced at the transcriptional level in dying sym-
pathetic neurons (Freeman et al., 1994; Estus et al., 1994). These
include, among others, the G 1 to S transition regula- tor, cyclin
D1, several Fos- and Jun-family transcription factors, and the
transcription factor NGFI-A. We hypothe- sized that the BCL-2
family members may be regulators of the apoptotic process in
sympathetic neurons. To evaluate gene expression in a small number
of cultured sympa- thetic neurons, we used the semiquantitative
RT-PCR assay characterized by Estus et al. (1994) and Freeman et
al. (1994). A group of individual preplated cultures, each
containing the same number of neurons, was prepared at a single
time. These neurons were maintained in NGF for 6 days and then
deprived of NGF for a specified number of hours before mRNA was
harvested. We began the anal- ysis of the BCL-2 family members by
determining the num- ber of PCR cycles and the amount of input cDNA
required for each set of primers to yield a linear amount of
radiola- beled PCR product (data not shown). Using these condi-
tions, the relative mRNA abundance of the BCL-2 family members
during apoptosis was analyzed. In three inde- pendent experiments,
analysis of cDNA harvested from neuronal cultures at given times
after NGF deprivation showed that RNA levels for all Bcl-2 family
members de- clined after NGF deprivation (Figure 2). The general
pat- tern was for levels to be maintained for the first 10-15 hr
and then decline. BcI-Xs was near the limit of detection in the
linear range of amplification, so reliable quantified
results could not be obtained. By 48 hr after deprivation, most
neurons were dead and no tyrosine hydroxyiase message was detected;
however, significant levels of Bcl-2 message, in particular,
remained detectable. This is in agreement with the immunostaining
for BCL-2 that showed a significant amount of the protein in the
cultured nonneuronal cells. These nonneuronal cells are not sensi-
tive to NGF deprivation and are the only cells left in the culture
after 48 hr of NGF deprivation. It is noteworthy that the majority
of messages analyzed after NGF deprivation of sympathetic neurons
fall within similar time courses (Freeman et al., 1994; Estus et
al, 1994; Greenlund et al., 1995). This suggests that both
decreased RNA synthesis and active RNA degradation may be important
for neu- ronal apoptosis (Freeman et al., 1994; Estus et al., 1994)
since under normal circumstances the mRNAs would be expected to
have different half-lives. The decreased neu- ronal synthesis of
the protective proteins, BCL-2 and BCL- XL, during apoptosis may
increase the susceptibility of neurons to death and contribute to
making the process irreversible.
Sympathetic Neurons from BCL-2-Deficient Mice Die More Rapidly
Than Neurons from Wild-Type Mice To determine unequivocally whether
BCL-2 is important in regulating neuronal survival during the
period of naturally occurring cell death, we examined neuronal
cultures from
-
Endogenous BCL-2 Regulates Neuronal Survival 653
W 100~
0 im
Q Z 80
01 i m >
• - 60 > Im
(n
C
U I , , 0 L
40'
20'
---IWild Type n-5 ---O Heterozygote n-7 ---~ Knockout n=8
0 , ~ 0 10 2 0 30 4 0
H o u r s a f t e r N G F D e p r i v a t i o n
Figure 3. TimeCourseof Apoptosis in Neurons from
BCL-2-Deficient, Heterozygous, and Wild-Type Mice After 1 week in
culture, sympathetic neurons from wild-type (n = 5 animals),
heterozygous (n = 7 animals), or BCL-2-deficient (n = 8 animals)
mice were either maintained in NGF or deprived of NGF for 24 or 32
hr. The live neurons in each culture were counted, and this number
was normalized to the NGF-maintained control cultures pre- pared
from the same mouse. After 24 hr of NGF deprivation, signifi-
cantly fewer neurons are alive in cultures prepared from BCL-2-
deficient mice and heterozygous mice compared with cultures
prepared from wild-type littermates. The data show the mean and SEM
for each set of animals (single asterisk, p < .05; double
asterisk, p < .01 ; triple asterisk, p < .001; ANOVA with
Bonferroni multiple comparisons test).
mice that lack BCL-2 (Veis et al., 1993). BCL-2-deficient mice
are the result of a mating between 2 heterozygous animals. Litters
produced by this mating have wild-type, heterozygous, and
BCL-2-deficient animals in them. These animals were sacrificed at
P1 or P2 and dissected; tail clips were taken at the time of
dissection for later geno- typing. Since the genotype of each
animal was not known at the time of dissection, pairs of SCGs were
removed from each animal, individually dissociated, and then main-
tained separately for 1 week in the presence of NGF. At the time of
dissociation, there was no statistical difference in the number of
SCG neurons recovered from animals of the three genotypes (p >
.1, ANOVA); however, after 1 week in vitro in the presence of NGF,
SCG cultures from BCL-2-deficient mice contained significantly
fewer neurons than cultures from wild-type littermates (p < .05,
ANOVA with Bonferroni multiple comparisons test). At 1 week in
vitro, SCG cultures were deprived of NGF for 24 or 32 hr, then
fixed and stained; the number of neurons was then counted. Crystal
violet positivity and cellular mor- phology were used as criteria
for neuronal viability (Deck-
werth and Johnson, 1993; Franklin et al., 1995). At least two
cultures were counted for every t ime point for each animal
indicated. The data in Figure 3 show that BCL-2- deficient neurons
(n = 8 animals) die more quickly than wild-type neurons (n = 5
animals), with heterozygotes (n = 7 animals) having an intermediate
rate. Although the BCL-2-deficient animals do not show gross
defects in the development of the nervous system (Veis et al.,
1993), the data demonstrate that BCL-2 is an important regulator of
neuronal survival under conditions of complete trophic factor
deprivation.
BCL-2 Family Expression and Maturation-Acquired Trophic Factor
Independence To determine whether the BCL-2 family members are im-
portant in the decreased acuteness of trophic factor de- pendence
associated with maturation, we began by com- paring the mRNA levels
of the BCL-2 family members from young neurons, which were still
sensitive to NGF depriva- tion, with those of mature neurons.
Sympathetic neurons that are maintained for 4 weeks in the presence
of NGF show dramatic decreases in trophic factor dependence, much
as they would in vivo. This in vitro model of matura- tion provides
a controlled situation that allows a careful analysis of changes in
gene expression as maturation pro- ceeds, in order that 1- and
4-week-old neurons could be directly compared and that the number
of neurons being analyzed was constant, a set of individual
cultures was prepared at one time and divided into two groups. One
group of cultures was maintained in NGF for I week before
harvesting RNA; the second group of cultures was main- tained in
NGF for 4 weeks. The mature neuronal cultures were deprived of NGF
for specific times before RNA was harvested. Figure 4 is a
representative experiment and shows that the RNA messages of most
of the BCL-2 family members are not significantly altered as
neurons de- crease in trophic factor dependence. The most notable
change is the decrease in Bcl-2 mRNA in 4 week in vitro neurons
compared with I week in vitro neurons. In contrast to young neurons
(see Figure 2), mRNA abundance in mature neurons does not decrease
dramatically after NGF deprivation. That Bcl-2 mRNA is decreased
and that the expression of the other Bcl-2 family members is
unaltered as trophic factor independence develops indicates that
modulation of the expression of these genes is not the molecular
mechanism of trophic factor independence. In addition, the data
show that in older neurons, no longer acutelydependent on trophic
factor, NGFdeprivation does not result in decreased mRNA abundance.
This further correlates the apparent degradation and decreased
levels of mRNA with apoptosis.
BCL-2 Protein Is Not Up-Regulated during Maturation and the
Acquisition of Trophic Factor Independence Using in situ
hybridization in combination with immunohis- tochemistry, Kondo et
al. (1992) and Chleq-Deschamps et al. (1993) have shown that BCL-2
expression may be regulated at the translational level. For
example, germinal center lymphocytes express large amounts of BCL-2
mes-
-
Neuron 654
bcl-2
1 wk neurons 4 wk neurons
0 0 7 15 25 48
bax
*--- bcl-2(23 cycles)
* - - tyr. h yd. (17 cycles)
/ ~ X / \ 0 0 7 15 25 48
~ * ~ batx(23 cycles)
* - - tyr. hyd. (17 cycles)
bcl-XL
0 0 7 15 25 48 ~ ~ ~ '~ ~ ' - -bcI -XL(23cyc les)
* - - tyr, hyd.(17 cycles)
bcI-Xs
0 0 7 15 25 48 ,~ ~ ~ ~ .~ * - - bcI-Xs(23 cycles)
~ ~ : ~ *~ t y r . hyd.(17 cycles)
Figure 4. RT-PCR Analysis of BCL-2 Family mRNAs in Mature
Neurons
Neurons were maintained in culture for either 1 week or 4 weeks.
The 4 week cultures were deprived of NGF for the indicated number
of hours, mRNA was isolated from neurons and cDNA was prepared.
SDS-PAGE was used to separate radiolabeled PCR products, and
autoradiography was used to view the products. The number of
amplifi- cation cycles done for each message is indicated. Note
that there is little or no decrease in the abundance of tyrosine
hydroxylase message or in the messages for the BCL-2 family members
after NGF depri- vation.
sage but little detectable BCL-2 protein, and mantle lym-
phocytes express both message and protein (Kondo et al., 1992). To
examine whether BCL-2 protein was increased even though the mRNA
was decreased, neurons main- tained 1 or 4 weeks in NGF were
immunosta ined for BCL-2 protein. The intensity of staining was
very similar in the two sets of cultures (data not shown), suggest
ing that the amount of BCL-2 protein did not change with maturat
ion. In addit ion, immunoprec ip i ta t ion fo l lowed by Western
blot- t ing was used to determine whether BCL-2 protein in- creased
as neurons matured. We conf i rmed that the anti- mouse BCL-2 ant
ibody, 3 F l l , immunoprec ip i ta ted rat BCL-2 by isolat ing
both rat and mouse thymocytes, immu- noprecipi tat ing with 3F11,
and separat ing by SDS-PAGE. A Western blot of thymocyte lysates
from both species showed a 26 kDa band (data not shown). To compare
the amount of BCL-2 in NGF-sensit ive neurons with amounts in
mature neurons, cultures contain ing the same number
t,n u) c c 0 0 L- L .
P- t -
(1) '¢" (~ (9
LL
30 kd " " " ¢
21.5 kd ; ! ̧ ii¸¸¸!i! iiii
IgG bcl-2
Figure 5. Western Blot Comparing BCL-2 Protein in Young and Ma-
ture Sympathetic Neurons Neuronal cultures were maintained in NGF
for either 1 week or 4 weeks before harvesting. Following
immunoprecipitation with 3F11 antibody, samples were separated by
SDS-PAGE and then transferred. The Western blot shows an
immunoprecipitation containing the 3F11 anti- body alone and
immunoprecipitations containing 3Fl l antibody in combination with
1 week or 4 week neuronal lysates. Both the 1 week and 4 week
neuronal lysates show a single 26 kDa BCL-2 band of similar
intensity relative to the 3Fl l band. This was repeated with
similar results.
of preplated neurons (the equivalent of 19 rats/culture) were
prepared at a single t ime and divided into two sets. The first set
of cultures was maintained in NGF for 1 week before harvest ing,
and the second set was mainta ined in NGF for 4 weeks before
harvesting. An immunoprec ip i ta- tion and Western blot with 3 F l
l of both sets of lysates showed a 26 kDa BCL-2 band of s imi lar
or decreased intensity (Figure 5) in older neurons. This indicates
that an increase in BCL-2 protein is not a factor in the loss of
acute trophic factor dependence.
Neurons from BCL-2-Deficient Mice Become Less Acutely Trophic
Factor Dependent with Maturation Although BCL-2 levels are not
increased during matura- tion, it is possible that the basal
expression levels are nec- essary and suff icient for the loss of
acute trophic factor dependence. To examine this possibil ity,
individual pairs of gangl ia from BCL-2-deficient, heterozygous,
and wild- type mice were independent ly dissociated and mainta ined
separately in NGF for 3 weeks. Cultures were then de- pr ived of
NGF for 72 hr, f ixed, and stained. Neurons were counted using
crystal v iolet posit ivity and cell morpho logy as criteria for
viabi l i ty (Figure 6). At least two cultures per animal indicated
were counted for each t ime point. No neuronal death occurred in
any of the cultures after 3 days of NGF depr ivat ion, regardless
of genotype. Therefore, BCL-2 expression inf luences neuronal
survival in re- sponse to t rophic factor depr ivat ion at the t
ime of natural ly occurr ing p rogrammed neuronal death but is not
required for the loss of acute trophic factor dependence in sympa-
thetic neurons.
-
Endogenous BCL-2 Regulates Neuronal Survival 655
100; w
O L - -
80' Z ¢D
J
--> 60 > = _
40 O O ¢_
20
--l~Wild Type n-4 ~Heterozygote n-4
Knockout n-5
0 0 2'0 4'0 6'0 8'0 Hours after NGF Deprivation
Figure 6. AnalysisofTrophicFactorlndependencein BCL-2-Deficient
Mouse Neuronal Cultures After 3 weeks in culture, sympathetic
neurons from wild-type (n = 4 animals), heterozygous (n = 4
animals), or BCL-2-deficient (n = 5 animals) mice were either
maintained in NGF or deprived of NGF for 72 hr. The live neurons in
each culture were counted, and this number was normalized to the
NGF-maintained control culture prepared from the same mouse. The
data show the mean and SEM (p > .05 for all comparisons; ANOVA
with Bonferroni multiple comparisons test).
Overexpression of BCL-2 in Sympathetic Neurons Delays Death but
Does Not Preserve Cellular Function Overexpression of BCL-2 has
been reported to inhibit or prevent the death of many cell types in
response to a num- ber of insults (Hockenbery et at., 1990; Garcia
et al., 1992; Allsopp et al., 1993; Batistatou et al., 1993; Kane
et al., 1993; Zhong et al., 1993). However, closer examination of
the data suggests that BCL-2 does not block death com- pletely, but
rather slows the inevitable process. To exam- ine whether
overexpression of BCL-2 would block or just delay the death of
sympathetic neurons, we microinjected a human Bcl-2 expression
vector or a control lacZ expres- sion vector into the nucleus of
cultured sympathetic neu- rons. Immunohistochemistry for human
BCL-2 or X-Gal histochemistry for lacZ 24 hr after injection
confirmed that 96% (n = 26) of the neurons injected with the Bcl-2
expres- sion vector and 90% (n = 37) of the neurons injected with
the lacZ vector were positive for the particular protein. To
examine the effect on survival, neurons were injected with the
Bcl-2 or lacZ expression vectors and then maintained in NGF for 24
hr to allow time for sufficient gene expres- sion. Neurons were
then deprived of NGF, and the number of viable, injected neurons
was determined at 24, 48, and 72 hr after NGF deprivation. In five
experiments, the num- ber of viable neurons was significantly
increased by over-
expression of BCL-2 but not by lacZ expression, although the
Bcl-2 vector-injected neurons died over a period of days (Figure
7A).
To explore further why the protection imparted by BCL-2 was
temporary, we examined whether BCL-2 would block the loss of
cellular functions after NGF deprivation. The rate of neuronal
protein synthesis, as assayed by [3~S]meth- ionine incorporation,
decreases rapidly, to 20% of NGF control in 16-18 hr after NGF
deprivation; this decrease can largely be prevented by K ÷
depolarization or increased cAMP (T. L. Deckwerth, unpublished
data), either of which prevents apoptosis of the neurons (Koike et
al., 1989; Ry- del and Greene, 1988). We tested whether BCL-2
overex- pression would similarly block this fall in protein
synthesis. We developed an in situ protein synthesis assay that al-
lowed the comparison of [35S]methionine/cysteine incor- poration in
individual neurons. Using this assay, neurons that were injected
and expressing the foreign gene product were compared with
uninjected neighboring neurons. Cells were injected with the Bcl-2
or lacZ expression vec- tors, maintained in NGF-containing medium
for 24 hr, and then deprived of NGF. After 14 hr of NGF
deprivation, neuronal cultures were pulsed for 4 hr with
[3~S]methio- nine/cysteine. After exposure to emulsion, the number
of grains overlaying neurons was counted under dark-field
microscopy (Figures 7B and 7C). When injected neurons, maintained
in NGF, were compared with uninjected neigh- boring cells, the
number of grains overlaying the neurons did not differ (lowest p
> .2, Student's t test; Figure 7C), irrespective of the vector
injected. Thus, the rate of [3sS]methionine/cysteine incorporation
and overall rates of protein synthesis were unaffected by the
injection pro- cess. Neurons deprived of NGF (18 hr total) had, on
aver- age, 79% fewer silver grains over them than those main-
tained in NGF (Figure 7C). In NGF-deprived cultures, neurons
overexpressing BCL-2 (Figure 7B) or LacZ (data not shown) had no
difference in the number of grains over- laying them compared with
uninjected neighboring neu- rons (lowest p > .4, Student's t
test; Figure 7C).. Therefore, the dramatic decrease in protein
synthesis that occurs with NGF deprivation is not prevented by
BCL-2 overexpres- sion. Because this crucial cellular function is
not main- tained by the expression of BCL-2, it is not surprising
that injected neurons do not remain viable for a prolonged pe-
riod. The failure of BCL-2 to maintain protein synthesis rates
indicates that, although the loss of viability is re- tarded, BCL-2
does not prevent the loss of cellular function associated with NGF
deprivation-induced death in sympa- thetic neurons.
Discussion
Overexpression of BCL-2 in neurons or neuron-like cell lines
protects cells from apoptosis in response to a number of
death-inducing stimuli (Garcia et al., 1992; AIIsopp et al., 1993;
Mah et al., 1993). Recently, Martinou et al. (1994a) have produced
transgenic mice that overexpress BCL-2 in the nervous system. These
mice have a reduced number of neuronal deaths in both the retina
and facial nucleus during the period of programmed neuronal
death,
-
Neuron 656
0 0 2 0 4 0 6"0 8 '0
bcl-2 Iz
Hours After NGF Deprivation
B
C 1 2 C
I,I.
Z 10( +
¢o
2 8¢
Z
6C
O 0 4C
0 0 3 2e a .
T i !
Ave Per Total Neuron Grains
+NGF 61 104& BCL-2 +NG~ 56 927
-NGF 14 390 ]BCL-2 -NG~ 11 321
+NGF BCL-2 -NGF BCL-2 +NGF -NGF
phase cont rast f l u o r e s c e n c e In situ
bcl-2 +NGF
bcl-2 -NGF
Figure 7. BCL-20verexpression Delays Neuronal Apoptosis but Does
Not Block the Fall in Protein Synthesis Rates That Occurs after NGF
Deprivation
(A) After 5-6 days in culture, sympathetic neurons were injected
with the Bcl-2 (n = 300) or lacZ (n = 382) expression vector and
maintained in NGF for 24 hr. All cultures were then deprived of NGF
for the indicated number of hours. The mean and SEM for five
experiments are shown. (B) The left column shows photomicrographs
of phase-contrast images of injected (large arrows) and noninjected
(small arrows) neurons in the presence and absence of NGF. The
middle column shows fluorescent images of the same fields of
neurons; injected neurons were identified by the presence of the
fluorescent label within the cell (identified by large arrows). The
right column shows the in situ protein synthesis assay with silver
grains overlaying neurons. Note that expression of BCL-2 had no
effect on methionine incorporation in either the presence of
absence of NGF. Bar, 25 pm. (C) Quantification of grains over
injected versus uninjected neurons in the presence or absence of
NGF. There is no statistical difference in the number of grains
overlaying neurons overexpressing BCL-2 (hatched bars; n = 17) or
uninjected neurons (closed bars; n = 17) in the presence of NGF (p
> .2, Student's t test). The inset shows the average number of
grains counted per neuron for each condition and the total number
of grains counted. The data are normalized to the NGF-maintained
cells, injected or uninjected. After a total of 18 hr of NGF
deprivation, the number of grains over neurons is decreased by an
average of 79%. BCL-2-overexpressing neurons (n = 28) show the same
decrease in grain density as uninjected neighboring neurons (n =
28). The data are the mean grain counts and SEM of all neurons in
two independent experiments.
-
Endogenous BCL-2 Regulates Neuronal Survival 657
as well as reduced neuronal death after experimental stroke.
Although BCL-2 overexpression clearly affects neuronal survival,
the importance of endogenous BCL-2 has not been assessed. We
examined, in a well-char- acterized in vitro system, the
physiological importance of endogenous BCL-2 in regulation of
neuronal survival, function, and maturation.
BCL-2 and Maturation-Induced Trophic Factor Independence Using
RT-PCR, we determined whether the loss of acute trophic factor
dependence in mature neurons was corre- lated with changes in BCL-2
family member gene expres- sion. The mRNA levels for most Bc/-2
family members were not significantly altered, and there was a
decrease in Bc/-2 message abundance with the development of trophic
factor independence. In addition, immunostaining and Western blots
show no increase in BCL-2 protein with mat- uration. Examination of
neurons from BCL,2-deficient mice demonstrates that BCL-2
expression was not re- quired for the loss of acute trophic factor
dependence. Thymocyte (Moore et al., 1994) and B cell (Merino et
al., 1994) development each involve stage-specific regulation of
BCL-2 expression. Pro-B cells express high levels of BCL-2, but
pre-B and immature B cells express only low levels. Those B cells
that reach maturity express high lev- els of BCL-2 (Merino et al.,
1994), which is important for maintaining memory B cells for
prolonged periods (NuSez et al., 1991). The susceptibility of B
cells at various stages of development to apoptosis-inducing
stimuli, such as glu- cocorticoid treatment, correlates positively
with the levels of BCL-2 expression (Merino et al., 1994). A
similar devel- opmental process in which the BCL-2 family members
reg- ulate the sensitivity of sympathetic neurons to NGF depri-
vation does not appear to occur. It is noteworthy that in mature
neurons the decrease in mRNA abundance after NGF deprivation does
not occur, indicating that the initia- tion of degradative events
after NGF deprivation is greatly slowed in mature neurons. The
molecular mechanism un- derlying loss of acute trophic factor
dependence has yet to be elucidated. Understanding the mechanisms
of resis- tance to apoptosis in the absence of trophic factor may
prove to be important in understanding both apoptosis and disease
states in which neurons are inappropriately susceptible to
death.
BCL-2 and Neuronal Survival during the Period of Programmed
Neuronal Death Again by using RT-PCR, we examined mRNA for the
BCL-2 family members in neurons undergoing apoptosis. Bcl-2 and
Bcl-Xt_ mRNA levels decreased after 15 hr of NGF deprivation; at 15
hr, neurons can be rescued from death by addition of NGF (Deckwerth
and Johnson, 1993). Thus, this fall in rnRNA levels was not caused
by the loss of neurons and most likely reflects active RNA
degradation combined with decreased RNA synthesis. Immunostain- ing
of NGF-deprived neurons also showed a significant decrease in BCL-2
protein by 15 hr after NGF deprivation. The reported half-life for
BCL-2 protein is 10 hr (Merino et al., 1994); since protein
synthesis rates decline rapidly
after NGF deprivation, at least half of the BCL-2 protein would
be expected to be gone by 15 hr after NGF depriva- tion, an
observation consistent with the immunohisto- chemical data. The
decreased levels of BCL-2 and BCL-X, after NGF deprivation, coupled
with increased levels of cyclin D1 and several Fos/Jun family
transcription factors, may be important in creating a cellular
environment that irreversibly commits a neuron to die. Experiments
with neurons from BCL-2-deficient mice demonstrate that elim-
ination of BCL-2 expression significantly increased the rate at
which neurons die in response to trophic factor deprivation.
Similarly, when Nakayama et al. (1993) and Veis et al. (1993)
evaluated the effect of the death-inducing stimuli, dexamethasone
and ~, irradiation, on T cells of BCL-2 homozygous mutant mice
(chimeras and complete knockouts, respectively), they observed an
increased sus- ceptibility to death. The BCL-2-deficient mice do
not show gross developmental defects In the nervous system (Veis et
al., 1993). This may be explained by the redundant ac- tion of
BCL-XL or similar molecules. In addition, in this in vitro
paradigm, neurons were deprived completely of trophic factor. Under
these extreme conditions, the lack of BCL-2, or the presence of
only a single copy, increases the rate of death and may increase
the susceptibility to death induced by a variety of insults. The in
vitro conditions used in these experiments may be more akin to
axotomy or other injury paradigms in which neurons are more com-
pletely deprived of trophic factor. During development, neurons
receive some target-derived or other trophic sup- port and are not
completely deprived of trophic factor; under these conditions,
molecules in addition to BCL-2, such as cell-cell/matrix contacts,
and electrical activity may all function to support neuronal
viability. Under more extreme conditions, the lack of one or more
of these could lead to increased susceptibility to death. Based on
the in vitro results, one would predict that ganglia of BCL-2-
deficient mice would have more neuronal death in re- sponse to
axotomy or target removal. That endogenous BCL-2 levels decrease
shortly before neurons are irrevers- ibly committed to die and that
neurons from BCL-2- deficient mice die much more rapidly suggest
that a de- crease in endogenous neuronal BCL-2 levels increases the
susceptibility of neurons to death, and as such, this decrease is
an important factor in the neuron becoming committed to die. These
observations support previous speculations (Estus et al., 1994)
that the "program" in pro- grammed neuronal death involves not only
increased ex- pression of certain genes but also decreased
expression of others.
Where Does BCL-2 Act to Block Programmed Neuronal Death?
Overexpression of BCL-2 in sympathetic neurons signifi- cantly
delayed, but did not block, apoptosis after NGF deprivation; nor
did it block the early fall in protein syn- thesis rates that
occurs in NGF deprivation-induced apoptosis. This is in contrast to
other agents, K + depolar- ization or increased cAMP, that maintain
viability after tro- phic factor deprivation. Since protein
synthesis rates are not maintained by BCL-2, the expression of
BCL-2 itself
-
Neuron 658
would be expected to fall, and BCL-2 would no longer be
overexpressed. This loss of BCL-2 protein, and the global decrease
in cel lular protein synthesis, contr ibutes to mak- ing BCL-2 only
a temporary saving agent.
Raft and col leagues (Jacobsen et al., 1994) propose that
apoptosis is a mult istep process consist ing of an activation
phase, an effector phase, and a degradat ion phase. We would
suggest a somewhat modif ied scheme of pro- g rammed cell death
involving an act ivat ion phase, fol- lowed by a propagat ion
phase, a terminal apoptot ic phase, and, at least in vivo, a
degradat ion phase. In sympathet ic neurons, the act ivat ion phase
of NGF depr iva t ion- induced death involves the expected
dephosphory la t ion of the TrkA receptor (C. Sanz-Rodr iguez,
unpubl ished data), de- creased MAP kinase act iv i ty (D. J.
Creedon, submitted), and the increased format ion of react ive
oxygen species (Greenlund et al., 1995). These events t r igger
processes in
the propagat ion phase that include posit ive and negat ive
regulat ion of gene expression and other biochemical changes within
the cell. RNA levels for genes including c-jun, c-fos, and cyclin
D1 are markedly increased, whi le RNA levels for Bcl-2 and Bcl-XL
are decreased during the propagat ion phase. A port ion of the
negat ive regulat ion of RNA levels may be exp la ined by an act
ive degradat ion of RNA. In addit ion, the synthesis of new protein
and RNA is dramatical ly decreased during the propagat ion phase.
The terminal apoptot ic phase includes events that abso- lutely
commit a neuron to die, i.e., the f ragmentat ion of DNA into o l
igonucleosomal f ragments and, potential ly, the act ivat ion of a
ced-3-1ike cyste ine protease (Yuan et al., 1993).
To what extent BCL-2 is capable of reversing the events
associated with the act ivat ion and propagat ion phases out l ined
above is not known. Our data suggest that, since BCL-2
overexpression does not block the fall in neuronal protein
synthesis rates (even though it c lear ly prolongs viability),
BCL-2 may be an effect ive b locker of the terminal apoptot ic
phase during which kil l ing actual ly takes place but not of
events in the act ivat ion or propagat ion phase. Albrecht et al.
(1994) observed a similar phenomenon in analyzing BCL-2 protect ion
from tumor necrosis factor (TNF)- induced apoptosis. TNF t reatment
normal ly stimu-
lates the translocat ion of NF-KB to the nucleus of L929 cells
before apoptot ic changes are evident. Al though L929 cells that
overexpress BCL-2 are protected from death induced by TNF, NF-KB is
still t ranslocated after TNF treat- ment. Whether overexpression
of BCL-2 in sympathet ic neurons would block the al terat ions in
gene expression of the propagat ion phase is not known. However,
the abil i ty of BCL-2 to block apoptosis induced by a myr iad of
stimuli, including those showing no lag phase and no requirement
for ongoing macromolecu lar synthesis (e.g., serum depri- vat ion
of PC12 cells; Bat istatou et al., 1993), is most con- sistent with
an act ion at, or immediate ly prior to, the termi- nal apoptot ic
phase. This may include a direct or indirect interact ion of BCL-2
with a ced-3-1ike molecule. Under- standing the events in this
process will provide means not only for prolonging neuronal viabi l
i ty in injury or disease but also for preserving neuronal
function.
Experimental Procedures
Rat Sympathetic Neuronal Culture Primary cultures of SCG neurons
were prepared by a modification (Martin et al., 1988) of the method
of Johnson and Argiro (1983). For preplating, dissociated SCG cells
were plated on untreated 60 or 100 mm Falcon (Becton Dickinson
Labware, Lincoln Park, N J) tissue cul- ture dishes for 90 min,
enough time for most of the nonneuronal cells to attach to the
plastic; however, neurons require a collagen (or other) substratum
for attachment and remained floating. The floating neu- ronal cells
were carefully washed off the plastic dish, filtered through Nitex
(size 3-20/14; Tetko Inc., Gaithersburg, MD) to remove aggre- gates
of neurons, and then plated. Neurons were plated in the center of
collagen-coated 35 mm dishes and maintained in NGF-containing
medium (AM50). AMS0 was Eagle minimal with Earle's salts (MEM; Life
Technologies, Gaithersburg, MD), with the addition of 50 ng/ml NGF
(prepared by the method of Bocchini and Angeletti [1969]), 10%
fetal bovine serum (Hyclone, Logan, UT), 2 mM L-glutamine, 100 p.g/
ml penicillin, 100 p.g/ml streptomycin, 20 ~M fluorodeoxyuridine
(an antimitotic), and 20 ~M uridine. Neurons were deprived of NGF
by incubation in the same medium, but without NGF (AM0) and with
goat polyclonal anti-mouse NGF antiserum added.
Immunostaining To stain for endogenous BCL-2, neurons were fixed
in fresh 4O/o para- formaldehyde in phosphate-buffered saline (PBS)
at room temperature for 30 rain, permeabilized for 20 min in
Tris-buffered saline (TBS) con- taining 0.1% Triton X-100, 1% BSA,
and 1% normal goat serum, and then incubated with the primary
antibody, 3F11 hamster monoclonal anti-mouse BCL-2 (1:20; 50 i~g/ml
final concentration) (Merry et al., 1994), or a control 6C8 hamster
monoclonal anti-human BCL-2 anti- body (1:20; 50 ~g/ml final
concentration) (Hockenbery et al., 1990) diluted in the
permeabilization solution. Incubation with the primary antibody was
for 1.5 hr at room temperature. Samples were rinsed twice and then
washed three times for 10 rain in permeabilization solution. The
biotinylated goat anti-hamster secondary antibody (South- ern
Biotechnology, Birmingham, AL) was diluted 1:100 (5 t~g/ml final
concentration) in permeabilization solution and incubated for 30
min at room temperature with the samples. After three 10 rain
washes in PBS, samples were incubated for 30 rain in ABC reagent
(avidin-HRP; Vector Laboratories, Burlingame, CA). After three 15
rain washes in PBS, samples were stained for 5-7 min with
3,3'-diaminobenzidine (Vector Laboratories). To detect
overexpressed human BCL-2, neu- rons were stained as described
above, but the primary antibody was 1:100 (10 t~g/ml) 6C8 hamster
monoclonal anti-human BCL-2. The biotinylated secondary antibody
and staining procedures were the same as those described above.
cDNA Preparation This method and its validation have been
described previously (Estus et al., 1994). Primary cultures
(-25,000 neurons/dish) that had been preplated on plastic were
maintained in AM50 for 6 days. Cultures were deprived of NGF for
the indicated times, and then RNA was harvested. Polyadenylated RNA
was isolated using an oligo(dT)- cellulose mRNA purification kit as
directed by the manufacturer (QuickPrep Micro kit, Pharmacia,
Piscataway, N J). Half of the poly(A) RNA was converted to cDNA by
reverse transcription with Moloney murine leukemia virus reverse
transcriptase (Superscript, Life Tech- nologies) with random
hexamers (16 I~M) as primers. The 30 p.I reaction contained 50 mM
Tris (pH 8.3), 40 mM KCI, 6 mM MgCI2, 1 mM dithi- othreitol, 500
I~M dATP, 500 p.M dTTP, 500 I~M dCTP, 500 I~M dGTP, and 20 U of
RNasin (Prornega, Madison, WI). After 10 min at 20°C, the samples
were incubated for 50 min at 42°C; the reaction was terminated by
adding 70 p.I of water and heating to 94°C for 5 rain.
PCR Analysis For a more detailed description of this method, see
Freeman et al., 1994; Estus et al., 1994; or Greenlund et al.,
1995. Oligonucleotide primers were synthesized by the Washington
University Protein Chem- istry Laboratory. Reactions for PCR
amplification of specific cDNAs were prepared on ice. Each reaction
contained 50 I~M dCTP, 100 p.M dGTP, 100 I~M dATP, 100 ~M dTTP, 15
p.Ci of [~-32P]dCTP (3000 Ci/
-
Endogenous BCL-2 Regulates Neuronal Survival 659
mmol), 1.5 mM MgCI~, 50 mM KCI, 10 mM Tris (pH 9.0), 0.1% Triton
X-100, 1 mM each primer, 1 U of Taq polymerase (Life Technologies),
and 1% of the cDNA synthesized in the reverse transcription
reaction. Each reaction was run for cycles of 1 min at 94°C, 1 min
at 55°C, and 2 min at 72°C in a Perkin-Elmer Cetus (Norwalk, CT)
thermocycler. After amplification, products were separated on a 10%
polyacrylamide gel, which was dried and visualized with ImageQuant
software on a Phosphorlmager (Molecular Dynamics, Sunnyvale, CA).
The Bcl-2 forward primer sequence was 5'-CTTTGTGGAACTGTACGGCCC-
CAGCATGCG-3'; the reverse was 5'-ACAGCCTGCAGCTTTGTTT-
CATGGTACATC-3' (the pair generated a 231 bp fragment). The bax
forward primer sequence was 5'-GGGAATTCTGGAGCTGCAGAG- GATGATT-3';
the reverse was 5'-GCGGATCCAAGTTGCCATCAG- CAAACAT-3' (the pair
generated a 96 bp fragment). The rat Bcl-X forward primer was
5'-AGGCTGGCGATGAGTTTGAA-3'; the reverse was
5'-CGGCTCTCGGCTGCTGCATT-3' (the pair generated a 337 bp fragment
for the long-form cDNA and a 150 bp fragment for the short-form
cDNA). The tyrosine hydroxylase forward primer was
5'-TTCAGAAGGGCCGTCTCAGA-3'; the reverse was 5'-CCGCTGCT-
GCTGCTGCAGCT-3' (the pair generated a 129 bp fragment). The Bcl-2
and bax primers were based on mouse sequence, so the prod- ucts
were subcloned and sequenced to confirm their identity. The Bcl-X
primers were based on rat sequence and generated only two products,
both of the expected size. The decrease in message abun- dance
after NGF deprivation was tested in at least three independent sets
of cDNA for all messages. The analysis of message abundance with
maturation was tested in four sets of cDNA for BcI-X (five indepen-
dent experiments), two sets for Bcl-2 (five independent
experiments), and one set for bax (two independent
experiments).
Preparation of SCG Cultures and Genotyping from BCL-2-Deficient
Litters Mice (Veis et al., 1993) were dissected on P1 or P2. The
geneotype of each animal was unknown at the time of dissection, so
the pairs of ganglia from each animal were individually dissociated
and plated on collagen-coated glass chamber slides. At least two
cultures were prepared for every data point for each animal. The
final cultures con- tained - 1000 neurons/well. The medium and
culture conditions were the same as those described above for rat
neuronal cultures. Geneo- typing was done by digesting the tail of
each animal overnight at 55°C in 250 I11 of a solution containing
50 mM KCI, 10 mM Tris-HCI, 1 mM EDTA (pH 8.0), 0.5% SDS, and 2
mg/ml proteinase K (Boehringer Mannheim, Indianapolis, IN). Samples
were centrifuged to remove large debris, and 150 ~1 of the
supernatant was transferred to fresh tubes. After 250 Id of a
solution containing 100 mM NaCI, 10 mM Tris- HCI (pH 8.0), 1 mM
EDTA (pH 8.0), 0.5% SDS, and 1 mg/ml RNase A (Boehringer Mannheim)
was added to the tail digest samples, they were incubated at 37°C
for 1 hr. DNA was extracted twice with phenol/ chloroform and once
with chloroform and was then precipitated with ethanol. The
resulting DNA was diluted 1:10 and 1 I11 was used in a PCR reaction
to determine geneotype. Reactions for PCR amplification of specific
cDNAs were prepared on ice. Each reaction contained 50 I~M dCTP,
100 I~M dGTP, 100 pM dATP, 100 ~M dTTP, 1.5 mM MgCI2, 50 mM KCI, 10
mM Tris (pH 9.0), 0.1% Triton X-100, 1 mM each primer, 1 U of Taq
polymerase (Life Technologies), and 1 i11 of the dituted tail DNA.
Each reaction was run for 35 cycles of 1 min at 94°C, 1 min at
55°C, and 2 rain at 72°C. The primers used to detect Bcl-2 were the
same as those stated above. The forward primer for the detection of
the neomycin resistance gene was: 5'-GGATCGGCCATTGAACAA- GATG-3';
reverse primer was 5'-CCGGGCGCCCCTGCGCTGACAGC- 3' (the pair
generated a 141 bp fragment).
Crystal Violet Staining and Neuronal Counting for Transgenic
Mouse Experiments This method is established as a reliable assay of
viability (Deckwerth and Johnson, 1993; Franklin et al., 1995).
Neuronal cultures on glass chamber slides were fixed with freshly
prepared 4% paraformaldehyde in PBS overnight at 4°C, stained with
1% crystal violet (EM Diagnostic, Gibbstown, NY), destained in
water, dehydrated in increasing ethanol concentrations, transferred
to toluene, and mounted in a toluene- based mounting solution
(Pro-Texx; Baxter Diagnostics, Deerfield, IL). Neurons were scored
as viable if they had a defined cellular outline
and visible nucleolus. Slides were coded and counts were done by
an independent observer.
Immunoprecipitation and Western Blots Preplated cultures of rat
SCG neurons were prepared as described. The equivalent of 38
ganglia (3.7 x 10 s preplated neurons) were plated per 60 mm dish,
and one dish was used per immunoprecipitation. After 1 or 4 weeks
in culture, neurons were lysed in 250 i~1 of buffer containing 1%
Triton X-100 (Sigma Chemicals, St. Louis, MO), 0.15 M NaCI, 10 mM
Tris (pH 7.4), 50 p.g/mt phenylmethylsulfonyl fluoride (Boehringer
Mannheim), 10 i~g/ml aprotinin (Boehringer Mannheim), and 10 I~g/ml
leupeptin (Boehringer Mannheim). The plates were incu- bated on ice
for 30 min followed by scrapping to maximize recovery. Lysates were
centrifuged at 13,000 rpm for 5 min to remove large cellular debris
and nuclei. For each immunoprecipitation, 3 ~1 of 1 mg/ ml 3F11
antibody was used. Samples were incubated for 1 hr at 4°C on a
rocker with the antibody. Protein G Sepharose in PBS (50 i11) was
added to each sample followed by another 1 hr incubation at 4°C on
a rocker. The beads were then washed five times in the lysis buffer
to reduce nonspecific binding. On the last wash, all buffer was
removed and reducing sample buffer containing 50 mM Tris-CI (pH
6.8), 100 mM dithiothreitol, 2% SDS, 0.1% bromphenol blue, and 10%
glycerol was added to each reaction. Samples were boiled and loaded
onto a 15% SDS-polyacrylamide gel. After electrophoresis, gels were
trans- ferred to Immobilon-P (Millipore, Bedford, MA). Blots were
blocked overnight in TBS plus 5% nonfat dry milk, washed in TBS
plus 0,05% Tween-20 (TBST), and then incubated with a 1:1000 (1
t~g/mt final concentration) dilution of 3F11 antibody in TBST for
45 min at room temperature. After thorough washing, blots were
incubated with 1: 1000 (0.5 t~g/ml) biotinylated goat anti-hamster
antibody (Southern Biotechnology) in TBST for 30 min at room
temperatu re. After washing, blots were incubated with 1:5000
avidin-HRP (Zymed, San Francisco, CA) in TBST for 15 rain at room
temperature. Blots were developed by enhanced chemiluminescence
according to the manufacturer's in- structions (Amersham, Arlington
Heights, IL).
Construction and Injection of Expression Vectors The base
expression vector was constructed by cloning into pUC19 a 695 bp
hCMV IE-1 promoter (620 bp upstream of the transcription start site
and 75 bp downstream; Boschart et al., 1985) and a 355 bp portion
of the mouse protamine-1 gene (Kleene et al., 1985) including the
polyadenylation signal. Specific expression vectors were con-
structed by cloning cDNAs into the BamHI site between the promoter
and polyadenylation signal of the base vector. The Bcl-2 cDNA was a
1.9 kb fragment, clone #58 (Seto et al., 1988), containing the
entire coding region for human BCL-2. DNAwas prepared by column
purifica- tion (Qiagen, Chatsworth, CA) and resuspended at 0.18
mg/ml in ster- ile-filtered, deionized water. For injections, DNA
was combined 1:1 with a dye solution containing 8 mg/ml rhodamine
dextran, 100 mM KCI, and 10 mM K3PO4, so that the final DNA
concentration injected was 0.09 mg/ml. Neuronal culture medium was
changed to Leibovitz's L-15 (Life Technologies), and injections
were made into the nucleus of each neuron. Concentrations of DNA
above 0.5 mg/ml produced a delay in apoptosis irrespective of the
cDNA in the vector. We demon- strated by the in situ protein
synthesis assay (see below) that injection of DNA at concentrations
as low as 0.2 mg/ml occasionally caused an inhibition of protein
synthesis in the presence of NGF, which pre- sumably explains the
delay in death. Experiments were scored by an independent
observer.
In Situ Protein Synthesis Assay This method has been described
previously (Greenlund et al., 1995). In brief, neurons were
injected with solutions containing 0.09 mg/ml expression vector
DNA, 2.25 mg/ml Cy3-1abeled donkey anti-sheep antibody as a fixable
tracer (Chemicon, Temecula, CA), 50 mM KCI, and 5 mM K3PO4 and then
maintained in AM50 for 24 hr. Cultures were deprived of NGF for 14
hr before beginning a 4 hr labeling period. During the labeling
period, neurons were incubated at 35°C in Eagle medium with Earle's
salts and 10% fetal bovine serum containing 10 ~M unlabeled
L-methionine and cysteine, 10 ~Ci/ml Tran[3SS]-Iabel (ICN, Irving,
CA), and 50 ng/ml NGF or polyclonal goat anti-NGF anti- body. After
labeling, cultures were washed and fixed in 4% paraformal-
-
Neuron 66O
dehyde. Cells were dehydrated and dipped in Kodak emulsion
(Kodak, Rochester, NY). After an overnight exposure, dishes were
developed in D-19 developer (Kodak), fixed, and rinsed. Grains
lying within the borders of the cell body of each neuron were
counted by an indepen- dent observer under dark-field
microscopy.
Acknowledgments
All correspondence should be addressed to E. M J, We thank P. A.
Lampe, J. L. Colombo, and P, A. K. Osborne for their expert
technical assistance and the Johnson Lab members for reviewing the
manu- script, This work was supported by the Renald McDonald
Children's Foundation, the Ataxia Telangiectasia Children's
Project, and the Washington University Alzheimer's Disease Research
Center (P50- AG05681).
The costs of publication of this article were defrayed in part
by the payment of page charges. This article must therefore be
hereby marked "advertisement" in accordance with 18 USC Section
1734 solely to indicate this fact.
Received February 8, 1995; revised May 16, 1995.
References
Abe-Dohmae, S., Harada, N., Yamada, K., and Tanaka, R. (1993).
Bcl-2 gene is highly expressed during neurogenesis in the central
nervous system. Biochem. Biophys. Res. Commun. 191,915-921.
Albrecht, H., Tschopp, J., and Jongeneel, C. V. (1994). Bcl-2
protects from oxidative damage and apoptetic cell death without
interfering with activation of NF-KB by TNF. FEBS Lett. 351,
45-48.
Allsopp, T. E., Wyatt, S., Paterson, H. F., and Davies, A. M.
(1993). The proto-oncogene bcl-2 can selectively rescue
neurotrophic factor- dependent neurons from apoptosis. Cell 73,
295-307.
Batistatou, A., Merry, D. E,. Korsmeyer, S. J., and Greene, L.
A. (1993). Bcl-2 affects survival but not neuronal differentiation
of PC12 cells. J. Neurosci. 13, 4422-4428.
Birren, S. J., Liching, L., and Anderson, D. J. (1993).
Sympathetic neuroblasts undergo a developmental switch in trophic
dependence. Development 119, 597-610. Bocchini, V., and Angeletti,
P. U. (1969). The nerve growth factor: purification as a
30,000-molecular-weight protein. Prec. Natl. Acad. Sci. USA 64,
787-794.
Boise, L. H., Gonz~llez-Garc(a, M., Postema, C. E., Ding, L+,
Lindsten, T., Turka, L. A., Mao, X., Nu,iez, G., and Thompson, C+
B. (1993). bcl-x, a bcl-2-related gene that functions as a dominant
regulator of apoptotic cell death. Cell 74, 597-608.
Boschart, M., Weber, F., Jahn, G., Dorsch-Hasler, K.,
Fleckenstein, B., and Schaffner, W. (1985). A very strong enhancer
is located up- stream of an immediate early gene of human
cytomegalovirus. Cell 41,521-530.
Boyd, J. M., Malstrom, S., Subramanian, T., Venkatesh, L. K.,
Schaeper, U., Elangevan, B., D'Sa-Eipper, C., and Chinnadurai, G.
(1994). Adenovirus E1B 19 kDa and bcl-2 proteins interact with a
com- mon set of cellular proteins. Cell 79, 341-351.
Castr~n, E., Ohga, Y., Berzaghi, M. P., Tzimagiorgis, G.,
Thoenen, H., and Lindhoim, D. (1994). bcl-2 messenger RNA is
localized in neurons of the developing and adult rat brain.
Neuroscience 61,165- 177.
Chen-Levy, Z., Nourse, J., and Cleary, M. L. (1989). The bcl-2
candi- date proto-oncogene product is a 24-kilodalton
integral-membrane pro- tein highly expressed in lymphoid cell lines
and lymphomas carrying the t(14;t8)translocation. Mol. Cell. Biol.
9, 701-710.
Chleq-Deschamps, C. M., LeBrun, D. P., Huie, P., Besnier, D. P.,
Warnke, R. A+, Sibley, R. K., and Cleary, M. L. (1993).
Topographical dissociation ofbcl-2 messenger RNA and protein
expression in human lymphoid tissues. Blood 81,293-298.
Cohen, J. J., and Duke, R. C. (1984). Glucocorticoid activation
of a calcium-dependent endonuclease in thymocyte nuclei leads to
cell death. J. Immunol+ 132, 38-42.
Coughlin, M. D., and Collins, M. B. (1985). Nerve growth factor-
independent development of embryonic mouse sympathetic neurons in
dissociated cell culture. Dev. Biol. 110, 392-401.
Deckwerth, T. L., and Johnson, E. M., Jr. (1993). Temporal
analysis of events associated with programmed cell death
(apoptosis) of sympa- thetic neurons deprived of nerve growth
factor. J. Cell Biol. 123, 1207- 1222.
DiCicco-Bloom, E., Friedman, W. J., and Black, I. B. (1993).
NT-3 stimulates sympathetic neuroblast proliferation by promoting
precur- sor survival. Neuron 11, 1101-1111.
Edwards, S. N., and Tolkovsky, A. M. (1994). Characterization of
apoptosis in cultured rat sympathetic neurons after nerve growth
factor withdrawal. J. Cell Biol. 124, 537-546.
Estus, S., Zaks, W. J., Freeman, R. S., Gruda, M., Bravo, R.,
and Johnson, E. M., Jr. (1994). Altered gene expression in neurons
during programmed cell death: identification of c-jun as necessary
for neu- ronal apoptosis. J. Cell Biol. 127, 1717-1727.
Fernandez-Sarabia, M. J,, and Bischoff, J. R. (1993). Bcl-2
associates with the ras-related protein R-ras p23. Nature 366,
274-275.
Ferrer, I., Tortosa, A., Condom, E., Blanco, R., Macaya, A., and
Planas, A. (1994). Increased expression of bcl-2 immunereactivity
in the devel- oping cerebral cortex of the rat. Neurosci. Lett.
179, 13-16.
Franklin, J. L., Sanz-Rodriguez, C., Juhasz, A., Deckwerth, T.
L, and Johnson, E. M., Jr. (1995). Chronic depolarization prevents
pro- grammed death of sympathetic neurons in vitro but does not
support growth: requirement for calcium influx but not trk
acUvation. J. Neu- rosci. 15, 843-664.
Freeman, R. S., Estus, S., and Johnson, E. M., Jr. (1994).
Analysis of cell cycle-related gene expression in postrnitetic
neurons: selective induction of Cyclin D1 during programmed cell
death. Neuron 12,343- 355.
Garcia, I., Martinou, I., Tsujimoto, Y., and Martinou, J. C.
(1992). Pre- vention of programmed cell death of sympathetic
neurons by the bcl-2 proto-oncogene. Science 258, 302-304.
Gorin, P. D., and Johnson, E. M., Jr. (1980). Effects of
exposure to nerve growth factor antibodies on the developing
nervous system of the rat: an experimental autoimmune approach.
Dev. Biol. 80, 313- 323.
Greenlund, L. J. S., Deckwerth, T. L., and Johnson, E. M., Jr.
(t995). Superoxide dismutase delays neuronal apoptosis: a role for
reactive oxygen species in programmed neuronal death. Neuron 14,
303-315.
Hengartner, M. O., and Horvitz, H. R. (1994). C. elegans cell
survival gene ced-9 encodes a functional homolog of the mammalian
proto- oncogene bcl-2. Cell 76, 665-676.
Hengartner, M. O., Ellis., R. E., and Horvitz, H. R. (1992).
Caenorhab- ditis elegans gene ced-9 protects cells from programmed
cell death. Nature 356, 494-499.
Hockenbery, D. M., Nut~ez, G., Milliman, C., Schreiber, R. D.,
and Korsmeyer, S. J. (1990). Bcl-2 is an inner mitochondrial
membrane protein that blocks programmed cell death. Nature 348,
334-336.
Hockenbery, D. M., Oltvai, Z. N., Yin, X. M., Milliman, C. L.,
and Kors- meyer, S. J. (1993). Bcl-2 functions in an antioxidant
pathway to pre- vent apoptosis. Cell 75, 241-251.
Jacobsen, M. D., Burne, J. F., King, M. P., Miyashita, T., Reed,
J. C., and Raft, M. C. (1993). Bci-2 blocks apoptosis in cells
lacking mitochondrial DNA. Nature 361, 365-369.
Jacobsen, M. D., Burne, J. F., and Raff, M. C. (1994).
Programmed cell death and Bcl-2 protection in the absence of a
nucleus. EMBO J. 13, 1899-1910.
Johnson, M. I., and Argiro, V. (1983). Techniques in the tissue
culture of rat sympathetic neurons. Meth. Enzymol. 103,
334-347.
Kane, D. J., Sarafian, T. A., Anton, R., Hahn, H., Gralla, E.
B., Valen- tine, J. S., Ord, T., and Bredesen, D. E. (1993). Bcl-2
inhibition of neural death: decreased generation of reactive oxygen
species. Science 262, 1274-1277.
Kerr, J. F. R., Wyllie, A. H., and Currie, A. R+ (1972).
Apoptosis: a basic biological phenomenon with wide-ranging
implications in tissue
-
Endogenous BCL-2 Regulates Neuronal Survival 661
kinetics. Br. J. Cancer 26, 239-257.
Kleene, K. C., Distel, R. J., and Hecht, N. B. (1985).
Nucleotide se- quence of a cDNA clone encoding mouse protamine 1.
Biochemistry 24, 719-722.
Koike, T., Martin, D. P., and Johnson, E. M, Jr. (1989). Role of
Ca 2+ channels in the ability of membrane depolarization to prevent
neuronal death induced by trophic-factor deprivation: evidence that
levels of internal Ca 2+ determine nerve growth factor dependence
of sympa- thetic ganglion cells. Proc. Natl. Acad. Sci. USA 86,
6421-6425.
Kondo, E., Nakamura, S., Onoue, H., Matsuo, Y., Yoshino, T.,
Aoki, H., Hayashi, K., Takahashi, K., Minowada, J., Nomura, S., and
Akagi, T. (1992). Detection of bcl-2 protein and bcl-2 messenger
RNA in nor- mal and neoplastic lymphoid tissues by
immunohistochemistry and in situ hybridization. Blood 80,
2044-2051.
Lam, M., Dubyak, G., Chen, L., NuSez, G., Miesfeld, R. L., and
Distel- horst, C. W. (1994). Evidence that bcl-2 represses
apoptosis by regulat- ing endoplasmic reticulum-associated Ca 2+
fluxes. Proc. Natl. Acd. Sci. USA 91, 6569-6573.
Mah, S. P., Zhong, L. T., Liu, Y., Roghani, A., Edwards, R. H.,
and Bredesen, D. E. (1993). The protooncogene bcl-2 inhibits
apoptosis in PC12 cells. J. Neurochem. 60, 1183-1186.
Martin, D. P., Schmidt, R. E., DiStefano, P. S., Lowry, O. H.,
Carter, J. G., and Johnson, E. M., Jr. (1988). Inhibitors of
protein synthesis and RNA synthesis prevent neuronal death caused
by nerve growth factor deprivation. J. Cell Biol. 106, 829-844.
Martinou, J.-C., Dubois-Dauphin, M., Staple, J. K., Rodriguez,
I., Frankowski, H., Missotten, M., Albertini, P., Talabot, D.,
Catsicas, S., Pietra, C., and Huarte, J. (1994a). Overexpression of
BCL-2 in transgenic mice protects neurons from naturally occurring
cell death and experimental ischemia. Neuron 13, 1017-1030.
Martinou, J.-C., Frankowski, H., Missotten, M., Martinou, I.,
Potier, L., and Dubois-Dauphin, M. (1994b). Bcl-2 and neuronal
selection during development of the nervous system. J. Physiol.
(Paris) 88, 209-211.
Merino, R., Ding, L., Veis, D. J., Korsmeyer, S. J., and NuSez,
G. (1994). Developmental regulation of the Bcl-2 protein and
susceptibility to cell death in B lymphocytes. EMBO J. 13,
683-691.
Merry, D. E., Veis, D. J., Hickey, W. F., and Korsmeyer, S. J.
(1994). Bcl-2 protein expression is widespread in the developing
nervous sys- tem and retained in the adult PNS. Development 120,
301-311.
Monaghan, P., Robertson, D., Amos, T. A. S., Dyer, M. J. S.,
Mason, D. Y., and Greaves, M. F. (1992). Ultrastructural
localization of bcl-2 protein. J. Histochem. Cytochem. 40,
1819-1825.
Moore, N. C., Anderson, G., Williams, G. T., Owen, J. J., and
Jenkin- son, E. J. (1994). Developmental regulation of bcl-2
expression in the thymus. Immunology 81, 115-119.
Nakayama, K., Nakayama, K., Negishi, l., Kuida, K., Shikai, Y.,
Louie, M. C., Fields, L. E., Lucas, P. J., Stewart, V., AIt, F. W.,
and Loh, D.Y. (1993). Disappearance of the lymphoid system in Bcl-2
homozygous mutant chimeric mice. Science 261, 1584-1588.
NuSez, G., Hockenbery, D., McDonnell, T. J., Sorensen, C. M.,
and Korsmeyer, S. J. (1991). Bcl-2 maintains B cell memory. Nature
353, 71-73.
Oltvai, Z. N., Milliman, C. L., and Korsmeyer, S. J. (1993).
Bct-2 hetero- dimerizes in vivo with a conserved homolog, Bax, that
accelerates programmed cell death. Cell 74, 609-619.
Oppenheim, R. W. (1991). Cell death during development of the
ner- vous system. Annu. Rev. Neurosci. 14, 453-501.
Pratt, R. M., and Greene, R. M. (1976). Inhibition of palatal
epithelial cell death by altered protein synthesis. Dev. Biol. 54,
135-145.
Rydel, R. E., and Greene, L. A. (1988). cAM P analogs promote
survival and neurite outgrowth in cultures of rat sympathetic and
sensory neu- rons independently of nerve growth factor. Proc. Natl.
Acad. Sci. USA 85, 1257-1261.
Seto, M., Jaeger, U., Hockett, R. D., Graninger, W,, Bennett,
S., Gold- man, P., and Korsmeyer, S. J. (1988). Alternative
promoters and ex- ons, somatic mutation and deregulation of the
bcl-2-1g fusion gene in lymphoma. EMBO J. 7, 123-131.
Tata, J. R. (1966). Requirment for RNA and protein synthesis for
in- duced regression of the tadpole tail in organ culture. Dev.
Biol. 13, 77-94.
Vaux, D. L., Weissman, I. L., and Kim, S. K. (1992). Prevention
of programmed cell death in Caenorhabditis elegans by human bcl-2.
Science 258, 1955-1957.
Veis, D. J., Sorenson, C. M, Shutter, J. R., and Korsmeyer, S.
J. (1993). Bcl-2-deficient-mice demonstrate fulminant lymphoid
apoptosis, polycystic kidneys, and hypopigmented hair. Cell 75,
229-240.
Wang, L,. Miura, M., Bergeron, L., Zhu, H., and Yuan, J. (1994).
Ich-1, an Ice/ced-3-related gene, encodes both positive and
negative regula- tors of programmed cell death. Cell 78,
739-750.
Wright, L. L., Cunningham, T. J., and Smolen, A. J. (1983).
Develop- mental neuron death in the rat superior cervical
sympathetic ganglion: cell counts and ultrastructure. J.
Neurocytol. 12, 727-738.
Yuan, J. Y., and Horvitz, H. R. (1990). The Caenorhabditis
elegans genes ced-3 and ced-4 act cell autonomously to cause
programmed cell death. Dev. Biol. 138, 33-41.
Yuan, J. Y., and Horvitz, H. R. (1992). The Caenorhabditis
elegans cell death gene ced-4 encodes a novel protein and is
expressed during the period of extensive programmed cell death.
Development 116, 309-320.
Yuan, J. Y., Shaham, S., Ledoux, S., Ellis, H. M., and Horvitz,
H. R. (1993). The C. elegans cell death gene ced-3 encodes a
protein similar to mammalian interleukin-l~-converting enzyme. Cell
75, 641-652.
Zhong, L. T., Kane, D. J., and Bredesen, D. E. (1993). bcl-2
blocks glutamate toxicity in neural cell lines. Mol, Brain Res. 19,
353-355.