Brain Development: Normal Processes and the Effects of Alcohol ...

92 NORMA L DEVELOPMENT
(b) axon s reac h th e wron g plac e withi n th e targe t
area, (c ) axons project t o the wron g target , (d ) ther e
are simply too many axons terminating within the target,
or (e) no target exists. One exampl e o f the latter is
the damage d trigeminal-somatosensor y system ; tran -
section o f the infraorbita l nerv e causes a los s of neurons
i n th e trigemina l brai n ste m nucle i an d th e
ventrobasal nucleu s o f th e thalamu s (Klei n e t al. ,
1988; Mille r e t al., 1991 ; Wait e e t al. , 1992 ; Miller ,
1999; Sugimot o et al., 1999 ; Baldi et al., 2000).
It is important to note that programmed neurona l
death i s a phenomenon mainl y confined to developing
invertebrates . As the ter m implies , neuron s ca n
die a s a consequence o f a genetic program. The bes t
examples are provided by Caenorhabditis elegans (see
Apoptosis, below). Thus, programmed cell death must
be distinguishe d fro m naturall y occurrin g neurona l
death, whic h principall y represents target-dependen t
neuronal death .
Many developin g neuron s di e soon afte r the y ar e
generated (se e Chapter 5) , perhaps a s a consequence
of an error in DNA replication. Thi s adaptive mechanism
allows an organism t o protect itsel f from basin g
its circuitr y on a cadr e o f neurons wit h mutate d o r
damaged DNA .
EXPERIMENTAL MODEL SYSTEMS
This chapte r consider s in vitro and i n viv o studies of
neuronal death. Eac h experimenta l approac h ha s its
own advantages and limitations. A major advantage of
in vitr o studie s is that a homogeneou s populatio n o f
cells i s addressed . Thi s settin g i s crucia l t o studie s
of the induction of neuronal death by xenobiotics, because
cells within the brain differ widel y in their susceptibility
(Watts et al., 2005). Additionally, the entir e
population is likely to respond t o a manipulation i n a
similar manne r an d develo p synchronously . Thi s
means that small changes in transcript and protein expression
o r post-translational modification s ar e mor e
likely to be detectable. Disadvantage s t o this method
are tha t cultur e condition s ca n influenc e th e path -
way^) activated , an d th e model ma y not b e physio -
logically relevant . I n contrast , i n vivo studie s us e a
highly heterogenous syste m in which ther e ar e mul -
tiple cel l type s i n divers e developmenta l stage s a t
a give n time . Thi s tempora l asynchron y make s i t
harder to differentiate transcrip t and protei n change s
from a background of ontogenetic noise . In addition,
changes tha t ar e see n ma y be difficul t o r impossibl e
to interpret . O f course , th e majo r advantag e t o thi s
method i s that it is the more physiologically and clini -
cally relevant model .
Direct evidenc e o f neurona l deat h i s difficult t o
amass because th e degenerativ e process ca n be rapid
and the debris can be removed quickly (Reddien and
Horvitz, 2004) . Thus, cells can expres s death-related
proteins (e.g. , activ e caspas e 3 an d bax ) for a lon g
time befor e the y di e (e.g. , Chen g an d Zochodne ,
2003), and no t al l cells rely on caspas e as an effecto r
of cell deat h (se e Caspase-Independen t Cel l Death ,
below). Anothe r widel y used metho d o f identifyin g
dying cell s i s termina l deoxynucleotidy l transferas e
mediated deoxyuridin e nic k end labelin g (TUNEL) ,
in which free , adenylate d ends of degenerating DN A
are labeled . Unfortunately , TUNE L ca n generat e
false positive results; in at least one mode l syste m the
amount o f TUNEL exceeds cel l los s (Gordo n e t al.,
2002).
The onl y incontrovertible proo f of neuronal deat h
is documentation o f a decrease i n neuronal number s
in a longitudinal, population-base d study . Final neu -
ronal numbers ar e a function of three developmenta l
events: generation, migration , and death. A study performed
at a single time-point in a mature animal cannot
determin e th e relativ e contribution s o f thes e
developmental events . A s a resul t o f th e potentia l
of generatin g fals e positiv e results , a mathematica l
method wa s devised to determine the number s of dying
neuron s (Miller , 2003). The grea t advantag e of
this approac h i s that th e estimat e o f cell deat h de -
pends on the documented cel l numbers and the proliferative
activity of the population .
MODES O F CELL DEAT H
The mode s o f cell deat h includ e apoptosis , necrosis ,
and autophag y (e.g. , Gree n an d Reed, 1998 ; Lock -
shin an d Zakeri , 2004) . Apoptosi s i s a highl y con -
trolled mechanism , durin g whic h specifi c transcrip t
and protein change s occur . Apoptosis has been calle d
cellular suicide , a s it only affects th e cel l undergoin g
it. I n contrast , durin g necrosis , cell s an d thei r or -
ganelles swel l and burst , releasing enzymes int o th e
local environment . A resul t o f thi s typ e o f deat h i s
damage to neighboring cells, thus causing a cluster of
dying cells. Autophagy involves lysosymal engulfmen t
of damaged cell s o r cell fragments . In man y model s