Journal of Neurochemistry, 2004, 91, 1284–1291 doi:10.1111/j.1471-4159.2004.02812.

S100B protein is released by in vitro trauma and reduces delayed

neuronal injury

Karen A. Willoughby,* Andrea Kleindienst,  Christian Müller,à Tao Chen,* Judith K. Muir*
and Earl F. Ellis*
*Department of Pharmacology and Toxicology, School of Medicine, Virginia Commonwealth University, Richmond, Virginia, USA
 Department of Neurosurgery, School of Medicine, Virginia Commonwealth University, Richmond, Virginia, USA
àInstitute of Laboratory Medicine and Pathobiochemistry, Charite, Humboldt University, Berlin, Germany

Abstract S100-beta with further release by 24 and 48 h. Adding 10 or

S100B protein in brain is produced primarily by astrocytes,
has been used as a marker for brain injury and has also been
shown to be neurotrophic and neuroprotective. Using a well
characterized in vitro model of brain cell trauma, we examined
the potential role of exogenous S100B in preventing delayed
neuronal injury. Neuronal plus glial cultures were grown on a
deformable Silastic membrane and then subjected to strain
(stretch) injury produced by a 50 ms displacement of the
membrane. We have previously shown that this injury causes
an immediate, but transient, nuclear uptake of the fluorescent
dye propidium iodide by astrocytes and a 24–48 h delayed
uptake by neurons. Strain injury caused immediate release of
100 nM S100B to injured cultures at 15 s, 6 h or 24 h after
injury reduced delayed neuronal injury measured at 48 h.
Exogenous S100B was present in the cultures through 48 h.
These studies directly demonstrate the release and neuro-
protective role of S100B after traumatic injury and that, unlike
most receptor antagonists used for the treatment of trauma,
S100B is neuroprotective when given at later, more thera-
peutically relevant time points.
Keywords: cell culture, in vitro models, neurons, neuropro-
tection, S100B protein, traumatic brain injury.
J. Neurochem. (2004) 91, 1284–1291.

S100B is a member of a group of Ca2+-binding proteins that
have been implicated in the Ca2+-dependent regulation of a
number of cellular functions including protein phosphoryla-
tion, enzyme activity, cell proliferation and differentiation,
cytoskeletal dynamics, intracellular Ca2+ homeostasis and
protection from oxidation damage (Rothermundt et al. 2003;
Van Eldik and Wainwright 2003). One of the earliest known
sources of S100B is brain, where S100B has been localized
mainly to astrocytes (Moore 1965; Boyes et al. 1986). There
are numerous reports indicating that S100B is released after
human brain trauma and the levels are positively correlated
with the degree of injury and negatively correlated with
outcome (Raabe et al. 1999; Leviton and Dammann 2002;
Van Eldik and Wainwright 2003). While increased levels of
S100B are associated with injury, it is uncertain if S100B
actually contributes to dysfunction. It has been hypothesized
that following injury when S100B levels increase S100B may
act in a positive paracrine manner since it has been shown to
stimulate neurite outgrowth (Van Eldik and Wainwright
2003). Recently Slemmer et al. (2002) have shown that strain
injury of mouse hippocampal cell cultures releases S100B,
and Kleindienst et al. (2004) have shown that infusion of
S100B into rats with experimental traumatic brain injury
(TBI) improves post-traumatic cognitive recovery.
The purpose of the current study was to determine whether
tissue cultured neuronal plus glial cells grown on a
deformable Silastic membrane release S100B over time after
strain (stretch) injury and whether S100B added to cultures
after injury affects known markers of delayed neuronal
injury. We report that trauma produces immediate and
prolonged release of S100B, and that S100B added to
neuronal plus glial cultures at 6 or 24 h after injury reduces
Received May 4, 2004; revised manuscript received 21 July, 2004;
accepted 9 August, 2004.
Address correspondence and reprint requests to Earl F. Ellis,
Department of Pharmacology and Toxicology, School of Medicine,
Virginia Commonwealth University, PO Box 980613, Richmond, VA
23298–0613, USA. E-mail: eellis@mail2.vcu.edu
Abbreviations used : PrI, propidium iodide; Rh123, Rhodamine 1,2,3;
TBI, traumatic brain injury.

1284
2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291

delayed neuronal injury by one-half. These findings have
potentially positive implications for the treatment of human
traumatic brain injury, a condition wherein virtually no
therapeutic agents administered at a therapeutically relevant
time after injury have been proven to be effective.
Materials and methods
Materials
This study was conducted using an in vitro method involving culture
plates with wells having flexible Silastic bottoms (Flex Plates,
Flexcell International, Hillsborough, NC, USA), and cells cultured
from rat pups obtained from Zivic Laboratories (Allison Park, PA,
USA). Cells were injured as described below using a commercially
available model 94 A Cell Injury Controller (Custom Design and
Fabrication, Virginia Commonwealth University, Richmond, VA,
USA).
S100B was obtained from Calbiochem (San Diego, CA, USA).
Trypsin was obtained from Sigma (St. Louis, MO, USA), and
DMEM and MEM from Mediatech (Herndon, VA, USA). Fetal
bovine serum was purchased from Hyclone Laboratories (Logan,
UT, USA) and horse serum from Gibco (Grand Island, NY, USA).
Triton-X-100 came from Fisher Scientific (Columbia, MD, USA).
Propidium iodide (PrI) was from Sigma (St. Louis, MO, USA).
S100-beta measurement was performed by chemiluminescence
immunoassay using a commercially available kit (LIAISON
Sangtec100, Sangtec Medical AB, Sweden) on an automated
immunoanalyzer (LIAISON analyzer, DiaSorin Deutschland, Ger-
many). S100B refers to a homodimer of S100-beta polypeptides. As
the LIAISON system is not specific for dimer versus monomer, its
analyte is referred to as S100-beta rather than S100B when reporting
release of the endogenous compound (Fig. 2).
Neuronal plus glial cultures
Mixed cultures of neuronal plus glial cells were prepared from 1- to
2-day-old Sprague–Dawley rat pups, as described previously
(McKinney et al. 1996). In brief, brains were removed aseptically
and placed in sterile dissecting saline. Cortices were dissected from
the brain and cleaned of white matter and meninges, leaving the
neocortex. The neocortex was placed in 5 mL dissecting saline
which contained 0.125% trypsin for 10 min and transferred to a tube
containing 5 mL of Dulbecco’s modified Eagle’s medium supple-
mented with 20 mM glucose, 10% fetal bovine serum, 100 U/mL
penicillin, 100 lg/mL streptomycin, and 2 mM L-glutamine, subse-
quently referred to as plating medium. The tissue was washed, the
supernatant removed, and 5 mL of culture medium was added for a
second wash. After washing, the supernatant was removed and the
tissue fragments were dissociated by trituration with a glass pipette
in 5 mL of culture medium. The suspension was centrifuged for
10 min at 200 · g. The supernatant was removed and 5 mL of
culture medium was added. The suspension was triturated, filtered
with a 70-lm nylon cell strainer, diluted with DMEM, and counted
with a hemocytometer. One-milliliter aliquots of the cell suspension
containing 1 · 106 cells were seeded into each well of a collagen
coated Flex Plate. Cells were cultured in a 5% CO2 incubator at
37C. One day after plating, the plating medium was removed and
replaced with 1 mL fresh plating medium. Three days later, and

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
S100B and in vitro neuronal trauma 1285
every three days thereafter, 0.5 mL medium was removed from each
well and replaced with 0.5 mL neuronal growth medium. Neuronal
growth medium contained minimal essential medium, 10 mM
glucose, penicillin 100 U/mL, streptomycin 100 lg/mL, and 5%
horse serum. Cells were used for experiments within 13–17 days
after removal from the rat. Glial cells adhere to the Silastic
membrane substrate and neurons adhere to the glial cells. Neuronal
processes and interconnections between individual neuronal cells
and neuronal cell clusters were readily apparent.
Cell injury
Neuronal plus glial cultures were injured as previously described
(Ellis et al. 1995). In brief, the Silastic membrane of the Flex Plate
well is rapidly and transiently deformed by a 50-ms pulse of
compressed gas, which deforms the Silastic membrane and adherent
cells to varying degrees controlled by the pulse pressure. The extent
of cell injury, as measured by nuclear uptake of propidium iodide
and release of lactate dehydrogenase, is dependent on the degree of
Silastic membrane displacement (Ellis et al. 1995; McKinney et al.
1996; Weber et al. 1999). Based on our previous work we have
arbitrarily defined three levels of cell injury, mild, moderate, and
severe, produced by displacing the Silastic membrane on which the
cells are grown by 5.5, 6.5, and 7.5 mm, respectively. These degrees
of membrane displacement result in a biaxial strain or ‘stretch’ of
31, 38, and 54%, respectively (Ellis et al. 1995). This range of cell
stretch has been shown to be relevant to what would occur in
humans after rotational acceleration/deceleration injury (Schreiber
et al. 1995). The current injury experiments were all performed
using a 7.5-mm membrane displacement because we have shown
that delayed neuronal injury is more robust at 7.5 mm than at the
two lower degrees of strain (Weber et al. 1999). Correlation of
degree of strain produced in models to strain-induced deficits in
humans has shown that greater than 30% stretch produces persistent
behavioral deficits (Schreiber et al. 1995). It is important to note that
with the cell injury model, as currently used, there is no hypoxia
superimposed on the injured cells.
Assessment of neuronal and astrocyte injury by propidium
iodide uptake
Injury was assessed by measuring nuclear uptake of propidium
iodide (PrI) as previously described (Jones and Senft 1985; Ellis
et al. 1995). PrI (1%) is excluded from cells with intact plasma
membranes but is rapidly taken up by cells with damaged
membranes, staining the nucleus a fluorescent red. Neuronal plus
glial cultures were returned to the incubator immediately after
injury, PrI added at 48 h post injury, and the cells were incubated for
15 min. Fluorescent-stained nuclei were visualized using a Zeiss
epifluorescence microscope with a total magnification of 400 · and
counted using a computer-driven imaging program (Signal Analy-
tics, Vienna, VA, USA).
In the current study we distinguished between astrocytes, neurons
and other cells using the following information and procedures, as
previously reported (Muir et al. 2002). In our previous work we
examined normal and injured cultures stained with propidium iodide
and with immunocytochemical approaches for detecting glial
fibrillary acid protein (GFAP) of astrocytes and microtubule-
associated protein (MAP-2) in neurons using confocal microscopy
with 3D imaging including the capacity to rotate the images.
1286 K. A. Willoughby et al.
Confocal side views showed that the astrocytes attach to and grow
on the Silastic membrane. Neurons grow on top of the astrocytes
and their processes course over the astrocytes. When viewed from
above, astrocyte and neuronal nuclei are at different depths of focus
using regular light or fluorescent microscopy. Thus one can
distinguish astrocytes from neurons by changing the depth of focus
when counting injured propidium iodide-retaining nuclei, as well as
by nuclear shape, round for neurons and oval for astrocytes, when
viewed from above.
Neurons in our neuronal plus glial cultures grow predominantly
in clusters, but also can be found growing individually. With MAP-2
staining and examining the side view with confocal microscopy the
cell bodies of the neuronal clusters appear like clusters of
‘mushrooms’ growing on a mound. The neuronal cell bodies are
represented by the mushroom caps. The neuronal nuclei are oriented
with their axis perpendicular to the surface of the plate. Thus when
viewed from above they appear round. The mound below, on which
the cell bodies are found, consists largely of neurites with some
astrocytic processes which course upward toward the center of the
mound and interdigitate with the neurites. With delayed neuronal
injury and propidium iodide staining, and when examined from the
side using confocal microscopy, the injured neuronal cell nuclei
appear clustered above the GFAP staining astrocytes. Thus, based
on time of PrI uptake after injury, astrocytes immediate and
transient, and, neurons delayed, as well as the depth of focus and
nuclear shape when viewed from above, we are able to distinguish
between neurons and astrocytes.
It might be suggested that the PrI staining nuclei we are counting
are those of microglia. However, we have shown in a previous study
that pure microglial cultures subjected to mild to severe biaxial
strain injury do not stain with PrI at any time after injury
(Rzigalinski et al. 1999).
For counting nuclei of injured cells (Figs 3 and 4), PrI-stained
cells were quantified using a computer-driven imaging program
coupled to a CCD camera as previously described (Ellis et al. 1995;
McKinney et al. 1996). The cells counted in each well consisted of
five fields. A field in the approximate center of the well is randomly
selected and examined, as well as four random fields above, below
and to each side of the center field. All fields are chosen in a blind
manner. No difference in degree of injury has ever been found in the
center versus the four surrounding fields. Thus the average of the
five fields is used for that well.
Following microscopic examination of the cultures, the buffer
was removed and the neuronal plus glial cells were solubilized with
NaOH. Total protein content was determined by the method of
Lowry et al. (1951). Delayed neuronal injury results are expressed
as the number of injured cells per milligram of cell protein.
Percent of neurons and glia in normal cultures
The quantification of percent of neurons versus glial cells in normal,
uninjured neuronal plus glial cultures was performed on 2-week-old
cell preparations grown on Silastic membranes of a six-well Flex
Plate. Propidium iodide (1%) was added to each well; after a 15-min
incubation at 37C 10 lL of 10% Triton-X-100 was added and the
cell nuclei were counted using a Zeiss epifluorescence microscope
and imaging system, as described above. Triton permeabilizes all
cells so both neuronal and glial nuclei take up PrI. Neuronal and
glial cells were distinguished as described above. Three random

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
fields from six different culture wells were counted and averaged.
Three, versus the normal five, fields were captured because the
Triton caused the cells to begin to lift if more time was spent on
capturing additional fields on the computer.
S100B
A concentrated stock of S100B was made in PBS buffer and stored
at 20C. Working dilutions of S100B were 1 lM in PBS and aliquots
of this working solution were added to make the cell culture S100B
concentration 10 nM (200 ng/mL) or 100 nM. Higher, non-physio-
logical S100B concentrations were not used because micromolar
concentrations are associated with increased astrocyte NO produc-
tion and neuronal death (Hu et al. 1997).
S100-beta analysis
S100-beta measurement was performed by chemiluminescence
immunoassay on an automated immunoanalyzer. Sensitivity for
the assays was 0.02 ng/mL. The Sangtec 100 assay linearly
measures S100-beta > 0.02 and < 200 ng/mL (Missler et al. 2000).
S100-beta is a well conserved protein within species and antibodies
react with S100-beta from a wide variety of mammalian species
(Rothoerl et al. 2000). The total extracellular medium (1 mL) was
carefully removed from each well, in order not to disturb the cells,
and immediately frozen. Thus any S100-beta in the cells was not
sampled. The cells were removed from the wells for protein
analysis, as described above. Each Flexplate well was used for
S100-beta sampling only once. At the conclusion of the experi-
mental series the S100-beta samples were shipped on dry ice to
Germany for analysis. Special care was taken to track the transport
of the samples in order to ensure that when the samples arrived at
the laboratory of Dr Müller in Germany they were still frozen. The
samples were marked in such a way that analysis was blind as to the
experimental condition.
Statistical analysis
Data are mean ± SEM values. Statistical significance was estab-
lished by ANOVA followed by Fisher’s protected least significant
difference test using StatView. A difference was considered
significant at p < 0.05.
Results
Percent of neurons versus glial cells in culture
While we have worked extensively with cortical neuronal
plus glial cultures derived from neonatal rats we have never
characterized the relative percent of glia and neurons in these
gray matter-derived cultures grown in our laboratory. Upon
normal microscopic examination of our cultures, astrocytes
form a confluent monolayer and neurons grow individually
or in clusters on top of the astrocytes. Because of the
confluent nature of the astrocytes one might surmise that
astrocytes, which form S100B, are the majority of the cells.
However, as shown in Fig. 1, the relative proportion of
neurons and astrocytes in our experimental cultures was
approximately equal.

Fig. 1 Characterization of the percent of neurons versus astrocytes in
uninjured neuronal plus glial cultures. The cultures examined con-
tained approximately equal proportions of neurons and astrocytes.
Values are mean ± SEM, n ¼ 3, three different cell preparations.
Three fields from each well were counted and averaged (six wells/
experiment). The percent of each cell type was not significantly dif-
ferent as determined by Fisher’s PLSD test.
Fig. 2 Release of S100-beta after severe stretch injury (7.5 mm
membrane displacement) of neuronal plus glial cultures. S100-beta
release occurred immediately after injury and increased with time.
Values are means ± SEM. The SEM for no injury is too small to display
graphically. n ¼ 6–8 wells; (a) p < 0.05 versus no injury; (b) p < 0.05
versus 15 s.
Stretch (strain) injury releases S100-beta
Figure 2 shows that S100-beta release into the culture
medium occurs immediately after injury and is further
increased in media at 24 and 48 h after injury. The amount of
S100-beta released in the 15 s after injury is approximately
equal to that which occurs in each of the two subsequent 24-h
periods. However, when one calculates the percent increase
in S100-beta released by 15 min, compared with uninjured
control, the percent increase immediately after injury is far
greater than subsequent percent increases with time.
The effect of S100B on delayed neuronal injury
Previous studies have shown that whereas astrocytes in pure
or neuronal plus glial cultures are injured and immediately
take up PrI, they regain their capacity to exclude PrI by 24

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
S100B and in vitro neuronal trauma 1287
and 48 h after injury. We observe virtually no PrI in
astrocytes at 48 h after injury (Ellis et al. 1995; McKinney
et al. 1996). In an opposite manner neuronal cells in
neuronal plus glial cultures show minimal PrI uptake or loss
of mitochondrial membrane potential immediately after
injury whereas they display increased membrane permeab-
ility and PrI uptake at 24 and 48 h after injury (Weber et al.
1999; Ahmed et al. 2000). We determined whether S100B,
which has been reported to reduce neuronal injury (Barger
et al. 1995; Brewton et al. 2001), could affect delayed
neuronal PrI uptake in stretch injured neuronal plus glial
cultures. Figure 3 demonstrates that when 10 or 100 nM
S100B is added to uninjured cultures and neuronal PrI uptake
examined at 48 h after addition of S100B the S100B has no
effect. Figure 3 also shows when PrI uptake is examined at
48 h after injury there is a several-fold increase in PrI uptake
in neurons. When 10 or 100 nM S100B was added to the
cultures immediately after injury and again at 24 h after
injury, delayed neuronal uptake of PrI was significantly
reduced.
While addition of S100B immediately after and at 24 h
postinjury affects delayed neuronal injury, it does not answer
the potentially important therapeutic question of its efficacy
when given long after injury. Therefore we performed
another series of experiments in which S100B was not given
to injured cells until 6 h and/or 24 h after injury and delayed
uptake of PrI measured at 48 h postinjury. As depicted in
Fig. 4, S100B had no effect on sham injured cells. Again, as
in Fig. 3, injury produced a multifold increase in delayed
neuronal injury which was reduced by approximately 50%
by S100B given at 6 or 24 h after injury. These experiments
clearly show the efficacy of S100B when administered at a
therapeutically feasible time postinjury.
Fig. 3 The effect of exogenous S100B given immediately and at 24 h
after injury on delayed injury of neurons in stretch-injured neuronal
plus glial cultures. Uninjured cells were given S100B immediately and
at 24 h after sham injury and PrI uptake examined at 48 h. Treating
with S100B immediately after injury plus at 24 h after injury signifi-
cantly decreased delayed neuronal injury as measured by neuronal
nuclear uptake of propidium iodide at 48 h. Values are mean ± SEM.
(a) p < 0.05 versus uninjured; (b) p < 0.05 versus injured without
S100B.
1288 K. A. Willoughby et al.
Fig. 4 The effect of delayed administration of 10 or 100 nM S100B on
delayed neuronal injury in stretch-injured neuronal plus glial cultures.
S100B when given at 6, 24 or 6 + 24 h after injury decreased delayed
neuronal injury by one-half as indicated by propidium iodide uptake at
48 h postinjury. Values are the mean ± SEM. n ¼ 6; (a) p < 0.05
versus uninjured control; (b) p ¼ 0.05 versus injury without S100B.
Persistence of exogenous S100B in cultures
Finally, since the exogenous S100B was effective we wanted
to determine its presence in our culture medium when added
before and at various times after injury and sampled from the
medium. As shown in Fig. 5 the exogenous S100B was
present throughout the maximum 48-h time period. This
implies no matter when the S100B is added its action does
not wane because of complete metabolism by normal or
Fig. 5 Persistence of added S100B in culture media from stretch-in-
jured neuronal plus glial cultures. Exogenous S100B was evident
immediately after injury through 48 h postinjury. In the two groups
receiving S100B at 15 s and 24 h after injury, the 24 h medium sample
was taken after the addition of the 24 h dose of S100B. Values are
means ± SEM. The number of culture wells is shown in each bar. Note
that the S100-beta levels reported in this figure are many fold greater
than that found in injured cultures not receiving exogenous S100B
(maximum 240 ng/mg protein at 48 h postinjury, see Fig. 2). While
cultures receiving exogenous S100B had levels above the linear
portion of the assay, cultures receiving 100 nM S100B always had
significantly more S100-beta than those receiving 10 nM, although not
always ten-fold greater. Note that any S100B located in or on the cells
would not be removed for S100-beta analysis.

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
injured cells. It should be noted that the levels of exogenous
S100B shown in Fig. 5 are many-fold higher than the
maximum release of endogenous S100-beta by injured cells
(Fig. 2). Thus while the test for the S100-beta analyte in
cultures receiving exogenous S100B did not yield linear
results because S100B was above the linear portion of our
assay, it us capable of showing the persistence of exogenous
S100B, with all values for 100 nM S100B being higher than
the 10 nM dose.
Discussion
Studies employing in vivo animal models of TBI have proven
useful in understanding some of the pathologic and patho-
physiologic cascades that are initiated by brain injury.
Perhaps most notably in recent years of studying traumatic
brain injury (TBI) is the sequelae of events initiated by post-
traumatic release of glutamate. While experimental pretrau-
ma treatment with ionotropic glutamate receptor antagonists
have shown protective effects, post-traumatic administration
of NMDA receptor antagonists in human clinical trials has
yielded disappointing results (Narayan et al. 2002; Royo
et al. 2003).
Because of the complexity of studying in vivo experimen-
tal TBI, we have turned to a reductionist tissue culture
approach in which individual cells and their interaction and
response to trauma may be more closely monitored and
influenced. Our approach utilizes inducing rapid cell strain
(stretch), a known component of in vivo TBI. Our work and
that of others using a similar approach has yielded informa-
tion on cellular events of TBI that would not be attainable
in vivo.
Recently, we have begun to focus on how factors which
are neurotrophic might affect the outcome of injury when
released in vivo or given exogenously at times beyond the
initial cascade of post-traumatic events, when treatment of
human TBI is more realistic. As reviewed thoroughly by
Donato (1999), one of the calcium-binding proteins, S100B,
was initially found in brain and is produced primarily by glial
cells. Its intracellular concentration in glial cells has been
estimated to be approximately 10 lM (Ciccarelli et al. 1999).
S100B release from cultured astrocytes can be enhanced by
adenosine and glutamate in a rapid receptor-mediated-like
fashion (Ciccarelli et al. 1999). S100B has been shown to be
neurotrophic in that over time it stimulates neurite extension
(Kligman and Marshak 1985; Van Eldik et al. 1988), as well
as glial cell proliferation (Selinfreund et al. 1991). Exogen-
ous S100B stimulates an immediate increase in intracellular
free calcium concentration in cultured neuronal plus glial
cells by stimulating hydrolysis of phosphoinositides and
release of intracellular calcium stores as well as by allowing
influx of extracellular Ca2+ (Barger and Van Eldik 1992).
S100B has been used as a marker of neuronal injury
in vivo and its increase is positively correlated with the

severity of neuronal trauma and negatively correlated with
outcome (Leviton and Dammann 2002). In a comprehensive
review Rothermundt et al. (2003) also discusses how S100B
has been positively associated with neurodegenerative
diseases such as Alzheimer’s disease as well as neuroinflam-
mation and reports and hypothesizes various mechanisms by
which S100B may exert reparative or destructive forces in
the brain. As S100B is neurotrophic (Kligman and Marshak
1985; Van Eldik et al. 1988) and it has been, for example,
shown to protect in vitro hippocampal neurons from damage
caused by glucose deprivation (Barger et al. 1995), stimulate
serotonergic nerve growth (Azmitia et al. 1990) and reverse
colchicine-induced cytoskeletal collapse (Brewton et al.
2001), there has been extensive discussion and research
suggesting that S100B may represent a reparative factor
released in response to neuronal injury (Van Eldik and
Wainwright 2003). Therefore, because S100B is a neuro-
trophic factor, is found in glial conditioned media, which is
often used in culture to promote neuronal health, and acts
long-term, we studied its role on normal neuronal membrane
integrity, its release from strain- (stretch-) injured neuronal
plus glial cultures and the capacity of exogenous S100B to
affect delayed neuronal injury.
In normal uninjured neuronal plus glial cultures there is a
measurable level of S100-beta in the growth medium
(Fig. 2). However, at 15 s after injury there is a dramatic
rise in S100-beta, indicating cell release of preformed S100-
beta. Recently Slemmer et al. (2002) have also shown release
of S100-beta from mouse hippocampal neuronal plus glial
cultures using an injury procedure identical to ours. Figure 2
also shows that the level of medium S100-beta continues to
rise through 24 and 48 h. The immediate post-traumatic rise
in S100-beta is similar to the immediate rise in medium LDH
and glutamate we have observed in similarly injured pure
astrocyte cultures (Ellis et al. 1995; Ahmed et al. 2002) or
glutamate release in pure neuronal cultures (Rzigalinski et al.
1998). However glutamate and LDH release peaks within
minutes after injury whereas in the current studies S100-beta
continues to rise, suggesting initiation of synthesis and
release. Because the astrocytes directly and firmly adhere to
the stretched Silastic membrane substrate, the S100-beta
release is likely the result of at least two possible processes.
First, direct cell injury and loss of membrane integrity. Our
electron microscopic studies of stretch-injured astrocyte
cultures show that following this degree of stretch injury
astrocyte structural integrity in many of the cultured cells is
severely compromised (Ellis et al. 1995). Also, immediately
after injury astrocyte nuclei in pure astrocyte cultures or in
astrocytes of neuronal plus glial cultures take up the
fluorescent dye propidium iodide which is impermeant to
most uninjured cells but binds readily to nuclei of cells with
compromised membrane integrity (Ellis et al. 1995). Second,
as we have shown that stretch-injured astrocytes release ATP,
and likely its breakdown products (Ahmed et al. 1998), as

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
S100B and in vitro neuronal trauma 1289
well as glutamate (Ahmed et al. 2002), the immediate release
of S100-beta might also be in part due to astrocyte receptor
activation, as suggested by Ciccarelli et al. (1999).
The finding of increased S100-beta release at 24 and 48 h
is intriguing because our previous studies show that, despite
initial astrocyte structural injury that might appear fatal to
these cells, by 24 and 48 h astrocytes repair and regain their
ability to exclude propidium iodide. This may imply that the
continued production of S100-beta at 24- and 48-h postinjury
is a longer-term metabolic response to injury. The additional
S100-beta release at 24 and 48 h may be a response to
promote healing in injured neurons or astrocytes. In fact we
have recently shown that stretch injury causes an increase in
astrocyte GFAP immunoreactivity at 24 but less so at 48 h
after injury (Floyd et al. 2004).
The delayed neuronal injury observed in our model in our
experience has not been pharmacologically preventable
except when agents are given before or seconds to minutes
after injury (Chen et al. 2004). Our in vitro studies have
demonstrated that delayed neuronal injury is in large part
initiated by NMDA receptor occupation occurring immedi-
ately after trauma. Additionally, studies of prolonged alter-
ation in GABA and AMPA synaptic function as well as
stretch-induced delayed depolarization (Tavalin et al. 1995,
1997; Goforth et al. 1999; Kao et al. 2004) support imme-
diate NMDA receptor activation in the initiation of these
phenomena. Thus, we have turned our studies to determining
whether delayed administration of S100B can intervene in
the cascade of events initiated by immediate post-traumatic
glutamate release which leads to delayed neuronal injury.
In a study aimed at a more clinically feasible time for
delivery of S100B we show that postponing S100B admin-
istration to 6 or 24 h after injury appears to reduce delayed
neuronal injury and uptake of PrI. Thus, the effect of delayed
post-traumatic administration of S100B is unlike the need for
immediate post-traumatic administration of NMDA receptor
antagonists.
The exact mechanism by which S100B reduced delayed
neuronal death is uncertain, but includes several possibilities
and evokes several questions. Because S100B is neurotroph-
ic it may, in overly simplified terms, stimulate anabolic
reparative protein synthetic machinery within the cell. S100B
has been shown to reverse colchicine-induced cytoskeletal
collapse in neuroblastoma cells, implying a potential to
correct structural defects (Brewton et al. 2001).
Another possible mechanism for S100B’s positive effect is
by affecting compromised bioenergetics in injured cells.
Barger et al. (1995) showed that exogenous S100B preven-
ted glucose deprivation-induced loss of neuronal mitochond-
rial function, as indicated by decreased mitochondrial
sequestration of Rhodamine 1,2,3 (Rh123). Exactly how
S100B reduced loss of mitochondrial function in the studies
of Barger et al. (1995) is uncertain, but we have previously
performed studies which support the potential importance of
1290 K. A. Willoughby et al.
the astrocytic response in restoring compromised neuronal
energetics (Ahmed et al. 1998). In stretch-injured pure
astrocyte cultures we found that astrocyte mitochondrial
function, as indicated by mitochondrial Rh123 sequestration,
was severely compromised at 15 min after stretch injury, but
normal at 24 h after injury, again supporting a repair of
astrocytes. When studies were performed at the same level of
injury as used in the current studies in pure neuronal cultures,
neuronal mitochondrial function was depressed by 15 min
and did not recover by 24 h. However, when neuronal
mitochondrial function was examined in neurons of neuronal
plus glial cultures it was suppressed similarly to that in
astrocytes at 15 min after injury, but importantly the neuronal
mitochondrial function recovered to normal by 24 h, similar
to astrocytes. These studies further support an important
potential reparative effect of astrocytes, and potentially
astrocyte-produced S100B.
Brain S100 proteins were first reported in 1965 and in the
almost 40 years since their discovery many investigators
have contributed to our knowledge of their structure and
function (Donato 1999). While S100B is thought of by many
as a mere indicator of brain injury we are intrigued by its
reported neuroprotective effect. The recent study of Kleind-
ienst et al. (2004) has provided more substantial evidence
that S100B may be neuroprotective when infused into
traumatically injured rat brains. Despite being studied for
many years, multiple questions about S100B need to be
answered with respect to traumatic brain injury, e.g. what are
their effects on intracellular calcium dynamics, cytoskeletal
structure, bioenergetics, neuronal necrosis or apoptosis and
long-term in vivo post-traumatic outcome. These questions
form the basis of our ongoing investigations.
Acknowledgements
We wish to thank Sallie Holt for her excellent assistance. This work
was supported by grants NS 27214, HL 57869 and HS 07288 from
the National Institutes of Health.
References
Ahmed S. M., Rzigalinski B. A. and Ellis E. F. (1998) Changes in
mitochondrial membrane potential in stretch-injured astrocytes and
neurons. J. Neurotrauma 15, 854.
Ahmed S. M., Rzigalinski B. A., Willoughby K. A., Sitterding H. A. and
Ellis E. F. (2000) Stretch-induced injury alters mitochondrial
membrane potential and cellular ATP in cultured astrocytes and
neurons. J. Neurochem. 74, 1951–1960.
Ahmed S. M., Weber J. T., Liang S., Willoughby K. A., Sitterding H. A.,
Rzigalinski B. and Ellis E. F. (2002) NMDA receptor activation
contributes to a portion of the decreased mitochondrial membrane
potential and elevated intracellular free calcium in stretch-injured
neurons. J. Neurotrauma 19, 1619–1629.
Azmitia E. C., Dolan K. and Whitaker-Azmitia P. M. (1990) S-100B but
not NGF, EGF, insulin or calmodulin is a CNS serotonergic growth
factor. Brain Res. 516, 354–356.

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
Barger S. W. and Van Eldik L. J. (1992) S100b stimulates calcium fluxes
in glial and neuronal cells. J. Biol. Chem. 267, 9689–9694.
Barger S. W., Van Eldik L. J. and Mattson M. P. (1995) S100b protects
hippocampal neurons from damage induced by glucose depriva-
tion. Brain Res. 677, 167–170.
Boyes B. E., Kim S. U., Lee V. and Sung S. C. (1986) Immunohisto-
chemical co-localization of S100B and the glial fibrillary acidic
protein in rat brain. Neuroscience 17, 857–865.
Brewton L. S., Haddad L. and Azmitia E. C. (2001) Colchicine-induced
cytoskeletal collapse and apoptosis in N-18 neuroblastoma cultures
is rapidly reversed by applied S-100b. Brain Res. 912, 9–16.
Chen T., Willoughby K. A. and Ellis E. F. (2004) Group I metabotropic
receptor antagonism blocks depletion of calcium stores and reduces
potentiated capacitative calcium entry in strain-injured neurons and
astrocytes. J. Neurotrauma 21, 271–281.
Ciccarelli R., Di Iorio P., Bruno V., Battaglia G., D’Alimonte I.,
D’Onofrio M., Nicoletti F. and Caciagli F. (1999) Activation of
A(1) adenosine or mGlu3 metabotropic glutamate receptors
enhances the release of nerve growth factor and S-100beta protein
from cultured astrocytes. Glia 27, 275–281.
Donato R. (1999) Functional roles of S100 proteins, calcium-binding
proteins of the ER-hand type. Biochim. Biophys. Acta 1450, 191–
231.
Ellis E. F., McKinney J. S., Willoughby K. A., Liang S. and Povlishock
J. T. (1995) A new model for rapid stretch-induced injury of cells
in culture: characterization of the model using astrocytes. J. Neu-
rotrauma 12, 325–339.
Floyd C. L., Rzigalinski B., Sitterding H. A., Willoughby K. A. and Ellis
E. F. (2004) Antagonism of metabotropic group I receptors and
PLC attenuates injury-induced increases in inositol(1,4,5)-tris-
phosphate in cultured astrocytes. J. Neurotrauma 21, 205–216.
Goforth P. B., Ellis E. F. and Satin L. S. (1999) Enhancement of AMPA-
mediated current following traumatic injury in cortical neurons.
J. Neurosci. 19, 7367–7374.
Hu J., Ferreira A. and Van Eldik L. J. (1997) S100b induces neuronal
cell death through nitric oxide release from astrocytes. J. Neuro-
chem. 69, 2294–2301.
Jones K. H. and Senft A. J. (1985) An improved method to determine
cell viability by simultaneous staining with fluorescein diacetate-
propidium iodide. J. Histochem. Cytochem. 33, 77–79.
Kao C.-Q., Goforth P. B., Ellis E. F. and Satin L. S. (2004) Potentiation
of GABAA currents after mechanical injury of cortical neurons.
J. Neurotrauma 21, 251–270.
Kleindienst A., Rice A. C., Harvey H. B., Müller C., Hamm R. J., Gaab
M. and Bullock M. R. (2004) Intraventricular infusion of the
neurotrophic protein S100B improves cognitive recovery after fluid
percussion injury in the rat. J. Neurotrauma 21, 541–547.
Kligman D. and Marshak D. R. (1985) Purification and characterization
of a neurite extension factor from bovine brain. Proc. Natl Acad.
Sci. USA 1985, 7136–7139.
Leviton A. and Dammann O. (2002) Brain damage markers in children:
neurobiological and clinical aspects. Acta Paediatr. 91, 9–13.
Lowry O. H., Rosebrough N. J., Farr A. L. and Randall R. J. (1951)
Protein measurement with the folin phenol reagent. J. Biol. Chem.
193, 265–275.
McKinney J. S., Willoughby K. A., Liang S. and Ellis E. F. (1996)
Stretch-induced injury of cultured neuronal, glial, and endothelial
cells: effect of polyethylene glycol-conjugated superoxide dismu-
tase. Stroke 27, 934–940.
Missler U., Wiesmann M., Ehlermann P., Tronnier M., Nötzold A.,
Steinmeier E. and Wood W. G. (2000) Validation and comparison
of two solid-phase immunoassays for the quanitification of S-100B
in human blood. Clin. Chem. 46, 993–996.

Moore B. W. (1965) A soluble protein characteristic of the nervous
system. Biochem. Biophys. Res. Commun. 19, 739–744.
Muir J. K., Willoughby K. A. and Ellis E. F. (2002) A confocal
microscopic examination of the effects of strain (stretch) on cul-
tured neurons and astrocytes. J. Neurotrauma 19, 1361.
Narayan R. K., Michel M. E., Ansell B. et al. (2002) Clinical trials in
head injury. J. Neurotrauma 19, 503–557.
Raabe A., Grolms C. and Seifert V. (1999) Serum markers of brain
damage and outcome prediction in patients after severe head injury.
Br. J. Neurosurg. 13, 56–59.
Rothermundt M., Peters M., Prehn J. H. and Arolt V. (2003) S100B in brain
damage and neurodegeneration. Microsc. Res. Tech. 60, 614–632.
Rothoerl R. D., Brawanski A. and Woertgen C. (2000) S-100B protein
serum levels after controlled cortical impact injury in the rat. Acta
Neurochir. 142, 199–203.
Royo N. C., Shimizu S., Schouten J. W., Stover J. F. and McIntosh T.
(2003) Pharmacology of traumatic brain injury. Curr. Opin.
Pharmacol. 3, 27–32.
Rzigalinski B., Weber J. T., Willoughby K. A., Liang S., Woodward J. J.
and Ellis E. F. (1998) Glutamate release and enhanced receptor
sensitivity in stretch-injured neurons. J. Neurotrauma 15, 894.
Rzigalinski B. A., Willoughby K. A., Sitterding H. A. and Ellis E. F.
(1999) A diffusible factor released from injured astrocytes enhan-
ces agonist-stimulated [Ca2+]i elevation and arachidonic acid
release in microglia. J. Neurotrauma 16, 962.
Schreiber D., Gennarelli T. A. and Meaney D. F. (1995) Proceedings of
the 1995 International Research Conference on Biomechanics of

2004 International Society for Neurochemistry, J. Neurochem. (2004) 91, 1284–1291
S100B and in vitro neuronal trauma 1291
Impact, pp. 233–244. International Research Council on Bioki-
netics of Impact, Brunen, Switzerland.
Selinfreund R. H., Barger S. W., Pledger W. J. and Van Eldik L. J. (1991)
Neurotrophic protein S100b stimulated glial cell proliferation.
Proc. Natl Acad. Sci. USA 88, 3554–3558.
Slemmer J. E., Matser E. J. T., De Zeeuw C. I. and Weber J. T. (2002)
Repeated mild injury causes cumulative damage to hippocampal
cells. Brain 125, 2699–2709.
Tavalin S. J., Ellis E. F. and Satin L. S. (1995) Mechanical perturbation
of cultured cortical neurons reveals a stretch-induced delayed
depolarization. J. Neurophys. 74, 2767–2773.
Tavalin S. J., Ellis E. F. and Satin L. S. (1997) Inhibition of the elec-
trogenic Na pump underlies delayed depolarization of cortical
neurons after mechanical injury or glutamate. J. Neurophys. 77,
632–638.
Van Eldik L. J. and Wainwright M. S. (2003) The Janus face of glial-
derived S100B: beneficial and detrimental functions in the brain.
Restor. Neurol. Neurosci. 21, 97–108.
Van Eldik L. J., Staecker J. L. and Winningham-Major F. (1988) Syn-
thesis and expression of a gene coding for the calcium-modulated
protein S100b and designed for cassette-based, site directed mut-
agenesis. J. Biol. Chem. 263, 7830–7837.
Weber J. T., Rzigalinski B. A., Willoughby K. A., Moore S. F. and Ellis
E. F. (1999) Alterations in calcium-mediated signal transduction
after traumatic injury of cortical neurons. Cell Calcium 26, 289–
299.