Social isolation alters neuroinflammatory response
Kate Karelinaa, Greg J. Normanb, Ning Zhangb, John S. Morrisb, Haiyan Penga, and A. Courtney DeVriesa,b,c,1
Departments ofaNeuroscience andbPsychology andcInstitute of Behavioral Medicine Research, Ohio State University, 29 Psychology Building,
1835 Neil Avenue, Columbus, OH 43210
Edited by William T. Greenough, University of Illinois, Urbana, IL, and approved February 24, 2009 (received for review October 24, 2008)
Social isolation has dramatic long-term physiological and psycho-
logical consequences; however, the mechanisms by which social
isolation influences disease outcome are largely unknown. The
purpose of the present study was to investigate the effects of
social isolation on neuronal damage, neuroinflammation, and
functional outcome after focal cerebral ischemia. Male mice were
socially isolated (housed individually) or pair housed with an
ovariectomized female before induction of stroke, via transient
intraluminal middle cerebral artery occlusion (MCAO), or SHAM
surgery. In these experiments, peri-ischemic social isolation de-
creases poststroke survival rate and exacerbates infarct size and
edema development. The social influence on ischemic damage is
accompanied by an altered neuroinflammatory response; specifi-
cally, central interleukin-6 (IL-6) signaling is down-regulated,
whereas peripheral IL-6 is up-regulated, in isolated relative to
socially housed mice. In addition, intracerebroventricular injection
of an IL-6 neutralizing antibody (10 ng) eliminates social housing
differences in measures of ischemic outcome. Taken together,
these data suggest that central IL-6 is an important mediator of
social influences on stroke outcome.
focal cerebral ischemia ? neuroinflammation
supportive are associated with improved health, whereas per-
ceived social isolation and stressful social interactions can be
detrimental to health. Within the clinical literature, low per-
ceived social support and social isolation predict the onset of
depression, as well as increased morbidity and mortality from
cardiovascular and cerebrovascular disease (1–4). Despite grow-
risk factors for cerebrovascular disease, little is known regarding
the mechanisms through which psychosocial factors influence
stroke pathogenesis. The health benefits of social interaction in
humans are typically attributed to improved health behaviors
such as decreased smoking, decreased alcohol consumption,
better nutrition, or better medical compliance, which in turn
improve cerebrovascular health (5). However, both social isola-
tion and perceived lack of social support are predictive of disease
outcome independent of health behaviors (6, 7). Furthermore,
the negative effects of social isolation on stroke and cardiac
arrest outcome can be reproduced in mice, and the data suggest
that socially isolated and socially housed mice mount a quanti-
tatively different pathophysiological response to ischemic dam-
age (8, 9).
Inflammatory processes have a fundamental role in the patho-
physiology of ischemic injury. Indeed, chronic and acute infec-
tion, as well as low-grade systemic inflammation [i.e., elevated
serum C-reactive protein (CRP)], are predictive of future
strokes, as well as death from stroke and cardiac arrest (10–14).
CRP is an acute phase protein that increases substantially in
response to proinflammatory cytokine release and as such is
used clinically as an index of chronic low-grade inflammation
(15). Importantly, emerging evidence indicates a relationship
between the social environment and systemic inflammation (16,
ocial interaction is an important modulator of both mental
and physical health. Social relationships perceived as being
17), and in otherwise healthy humans, low social integration is
associated with increased CRP concentrations (18, 19). Further,
socially isolated mice exhibit increased intraischemic serum CRP
concentrations relative to socially housed animals after experi-
mental stroke (8). Although a direct causative role for CRP on
the extent of ischemic injury has not been established, both the
clinical (11, 16–18) and animal (8) data provide evidence of a
strong correlation between social factors and the inflammatory
response typically associated with ischemic injury.
The goal of the current study was to examine the influence of
social housing on stroke outcome. Specifically, poststroke cyto-
kine expression, edema formation, infarct development, and
functional recovery were compared in socially housed and
Social Isolation Influences Poststroke Survival and Ischemic Damage.
Housing condition was a strong determinant of poststroke
survival rate and ischemic damage. Following middle cerebral
artery occlusion (MCAO) only 40% of socially isolated mice
survived 7 days, compared with 100% of socially housed mice
(U ? 20.00, P ? 0.05, r ? 0.63), which limits interpreting the day
7 infarct and behavior data as being truly representative of the
2 experimental groups. However, it is interesting to note that the
4 surviving mice in the socially isolated group were similar in
infarct size and behavior to the socially housed group on
poststroke day 7 (P ? 0.05; although we caution that this
comparison suffers from the statistical limitation inherent in
having a small sample size in one of the experimental groups).
To address the issue of differential long-term survival, all
remaining measurements were made 24–72 h after initiation of
reperfusion, when survival rates were not statistically different
between groups (90% socially isolated and 100% socially housed
on day 3; U ? 45.00, P ? 0.05, r ? 0.22). Social isolation
exacerbated infarct volume at 24 and 72 h (24 h, t20? 1.738,
P ? 0.05; ?2? 0.12; 72 h, t20? 2.568, P ? 0.05, ?2? 0.11) (Fig.
1A). Social isolation also significantly exacerbated cerebral
edema 48 h after MCAO; socially isolated animals experienced
a 2-fold increase in edema relative to socially housed animals
(t9? 1.801, P ? 0.05, ?2? 0.16) (Fig. 1B).
Within the open field, there were no effects of social housing
on locomotor activity or exploratory behavior measured 24 h
before surgery (all P ? 0.05). A 2-factor ANOVA (factors
were surgery and housing condition) revealed an effect of
surgery on rearing behavior 72 h post-MCAO or SHAM surgery
(F1,26? 28.61, P ? 0.05). After MCAO, mice reared significantly
Author contributions: K.K. and A.C.D. designed research; K.K., G.J.N., N.Z., J.S.M., and H.P.
performed research; K.K. and A.C.D. analyzed data; and K.K. and A.C.D. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed at: Departments of Psychology and Neu-
roscience and Institute of Behavioral Medicine Research, Ohio State University, 51 Psy-
chology Building, 1835 Neil Avenue, Columbus, OH 43210. Email: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
www.pnas.org?cgi?doi?10.1073?pnas.0810737106 PNAS ?
April 7, 2009 ?
vol. 106 ?
no. 14 ?
less than SHAM; however, there were no effects of housing
condition on total locomotor activity or exploratory behavior
(all P ? 0.05) [supporting information (SI) Fig. S1 and Table S1].
Additionally, there were no social housing differences in open
field central tendency (a measure of anxiety-like behavior) at the
presurgical or postsurgerical time points (P ? 0.05). Further,
during rearing in a cylinder, there was no significant effect of
housing condition on contralateral paw use pre- or postsurgery
(all P ? 0.05).
There were no significant housing effects on body mass
(F1,52? 2.189, P ? 0.05, ?2? 0.04), body temperature during
surgery (F1,52? 0.038, P ? 0.05, ?2? 0.0004), or neuroscore
(F1,52 ? 2.901, P ? 0.05, ?2? 0.05) across or within the
Social Isolation Alters the Neuroinflammatory Response to Stroke.
Poststroke gene expression of macrophage antigen complex-1
(MAC-1), a pattern recognition complement receptor protein
expressed on macrophage-lineage cells (F1,96? 5.699, P ? 0.05,
?2? 0.05), and glial fibrillary acidic protein (GFAP), an
intermediate filament protein that is up-regulated in astrocytes
following injury (F1,92 ? 5.519, P ? 0.05, ?2? 0.05), were
significantly elevated in the ipsilateral (ischemic) relative to the
contralateral (nonischemic) hemisphere across both time points
after MCAO (Fig. S2 and Table S2). At the 12 h time point,
there was a main effect of housing within the striatum on MAC-1
(F1,10? 8.709, P ? 0.05, ?2? 0.46) and GFAP (F1,15? 6.63,
P ? 0.05, ?2? 0.31) gene expression, and a post hoc analysis
revealed that both glial markers were significantly elevated in
socially isolated animals relative to socially housed animals (P ?
0.05) (Fig. 2). Cortical gene expression of MAC-1 (F1,15? 0.297,
did not vary significantly by housing conditions (P ? 0.05).
Overall, relative gene expression of proinflammatory cyto-
kines interleukin-1 beta (IL-1?) (F1,146 ? 11.429, P ? 0.05,
?2? .07), tumor necrosis factor alpha (TNF-?) (F1,136? 30.876,
P ? 0.05, ?2? 0.17), and interleukin-6 (IL-6) (F1,127? 15.180,
P ? 0.05, ?2? 0.10), as well as transforming growth factor beta
(TGF-?) (F1,128? 7.886, P ? 0.05, ?2? 0.22) and cyclooxygen-
ase-2 (COX-2) (F1,147 ? 7.773, P ? 0.05, ?2? 0.05) were
significantly up-regulated in the ipsilateral (ischemic) hemi-
sphere relative to the contralateral (nonischemic) hemisphere
across both time points (Fig. S3). Post hoc analyses revealed that
there were no effects of housing on IL-1?, TNF-?, TGF-?, or
COX-2 expression (all P ? 0.05). However, IL-6 gene expression
was significantly lower in socially isolated mice than socially
housed mice at 12 h (striatum, F1,10? 5.689, P ? 0.05, ?2? 0.36).
Further, brain IL-6 protein expression was significantly lower
(cortex: F1,10? 8.711, P ? 0.05, ?2? 0.49), whereas serum IL-6
concentrations were significantly higher in socially isolated
relative to socially housed mice (F1,15 ? 9.297, P ? 0.05,
?2? 0.39) (Fig. 3).
Ischemic Outcome. Treatment with an IL-6 neutralizing antibody
significantly increased infarct volume. A 2-factor ANOVA re-
vealed main effects of treatment (F1,24 ? 16.081, P ? 0.05),
housing (F1,24? 5.057, P ? 0.05), and a treatment by housing
interaction (F1,24? 7.315, P ? 0.05) on infarct volume (?2?
0.32). Among vehicle artificial cerebrospinal fluid (aCSF)
treated mice, a Tukey post hoc analysis revealed that infarct
volume was significantly larger in socially isolated than socially
housed mice (P ? 0.05) but the infarct size was equivalent
between animals in both housing conditions that received IL-6
Further, across both treatment conditions, a 2-factor ANOVA
protein (F1,21? 7.984, P ? 0.05; ?2? 0.28). A Tukey post hoc
revealed that socially isolated mice had significantly higher
housed mice in the vehicle treated group (P ? 0.05). However,
central administration of the IL-6 neutralizing antibody elimi-
nated the difference in circulating IL-6 between socially housed
and isolated mice. Thus, central IL-6 immunoneutralization in
turn eliminated social influences on both postischemic infarct
volume and peripheral IL-6 concentration (Fig. 4).
Poststroke Serum Corticosterone (CORT) Concentrations. A 2-factor
ANOVA revealed a main effect of reperfusion time on CORT
concentration (F1,47 ? 10.975, P ? 0.05; ?2? 0.18). CORT
24 Hours72 Hours
Percent infarct (mean ± SEM)
Edema Index (mean ± SEM)
Percent infarct relative to the contralateral hemisphere is significantly in-
creased in socially isolated mice after 24 h and 72 h of reperfusion. (B) Index
of edema is also significantly increased in socially isolated mice at 48 h of
reperfusion.*, significantly different from socially housed mice, P ? 0.05.
Effect of housing condition on ischemic damage after MCAO. (A)
Relative Gene Expression
(mean ± SEM)
via RT-PCR. Within the ipsilateral hemisphere, both MAC-1 and GFAP are
significantly up-regulated in socially isolated relative to socially housed mice.
Data are presented as gene expression relative to control gene (18S) expres-
sion.*, significantly different from socially housed mice, P ? 0.05.
www.pnas.org?cgi?doi?10.1073?pnas.0810737106Karelina et al.
concentrations were highest at 12 h and decreased significantly
by 24 h. However, there was no effect of housing condition on
CORT concentration at any time point (P ? 0.05) (Table S3).
Social environment influences immune function and disease
outcome (17, 19). However, the mechanisms underlying the
interaction of psychosocial factors and pathophysiology in isch-
emic injury require clarification. Data from the current study
indicate that social housing condition is a strong determinant of
the pathophysiology and long-term survival after experimental
stroke. The survival rate to 7 days after experimental stroke was
100% for socially housed mice, compared with only 40% of
socially isolated mice. The biased distribution in survival may
reflect increased damage in socially isolated animals that con-
sequently did not survive to day 7. Indeed, infarct and edema
analyses at earlier time points indicate significantly greater
ischemic damage in socially isolated mice than socially housed
mice (Fig. 1). These data confirm and extend previous reports
that social isolation potentiates the pathophysiological response
to ischemia (8, 9) and suggest that social isolation contributes to
early differences in the trajectory of ischemic injury development.
A separate cohort of animals was used to determine whether
the increase in infarct size among socially isolated mice was
associated with a difference in the neuroinflammatory response
to MCAO. The neuroinflammatory response is triggered by
activated microglia and astrocytes (i.e., reactive gliosis), as well
as an up-regulation of proinflammatory cytokine release in
response to neuronal damage (20–22). As expected, there was
relative to the contralateral hemisphere after MCAO (Fig. S2).
Importantly, within the ipsilateral hemisphere, gene expression
of both MAC-1 and GFAP was increased in socially isolated
mice relative to socially housed mice (Fig. 2). These data
complement a recent report on social isolation-induced poten-
tiation of neuroinflammatory responses in a model of global
cerebral ischemia (9). The functional role of glia in ischemic
injury is unclear; studies report both neuroprotective and dam-
aging effects of glial products after an ischemic event (23–27).
Although the current study does not indicate a causal relation-
ship between the up-regulated glial markers and infarct volume,
there is evidence that inhibition of microglial activation (via
administration of minocycline) reduces stroke damage (28).
Thus, taken together with increased infarct volume in socially
isolated animals, it is possible that the secondary processes trig-
gered by increased glial activation exacerbate neuronal damage.
We further conducted mRNA gene expression profiles on
several genes that are central to the neuroinflammatory re-
sponse in cerebral ischemia. Key among these genes are the
cytokines IL-1?, TNF?, IL-6, TGF-? and the COX-2 enzyme.
These inflammatory mediators are produced and secreted by
activated glia within hours of ischemic injury and thus contribute
significantly to the extent of neuronal damage after MCAO (21,
Central IL-6 Gene Expression
(mean ± SEM)
Central IL-6 Protein Expression
(pg/mL; mean ± SEM)
Serum IL-6 (pg/mL; Mean ± SEM)
IL-6. In the CNS, striatal IL-6 mRNA gene expression measured via RT-PCR (A)
and cortical protein concentration measured via ELISA (B) are significantly
down-regulated in the ischemic hemisphere of socially isolated relative to
socially housed mice. (C) Serum IL-6 measured via ELISA is up-regulated in
socially isolated mice. Gene and protein expression data in the CNS are
significantly different from socially housed mice, P ? 0.05.
Relative gene expression and protein concentration of poststroke
aCS F IL-6 Antibody
Is ol at ed
Pa ir ed
Pe rc ent In fa rc t (m ea n ± SEM )
aCS F IL-6 Antibody
Serum IL-6 (pg/mL; Mean ± SEM)
treatment with IL-6 neutralizing antibody. (A) Treatment with 10 ng of IL-6
mice and eliminates the effect of social interaction achieved with vehicle
treatment. (B) Serum IL-6 is up-regulated in vehicle-treated socially isolated
antibody eliminates the difference in serum IL-6 concentrations.*, signifi-
cantly different from socially housed mice; #, significantly different from
aCSF-treated mice, P ? 0.05.
Infarct volume and serum IL-6 protein concentrations after ICV
Karelina et al. PNAS ?
April 7, 2009 ?
vol. 106 ?
no. 14 ?
22). Our data indicate that gene expression of IL-1?, TNF-?,
TGF-?, and COX-2 is significantly up-regulated in the ipsilateral
relative to the contralateral hemisphere (Fig. S3), but, contrary
to our initial hypothesis, these inflammatory markers do not
IL-6 signaling is significantly altered by housing conditions; gene
expression of striatal IL-6 is decreased in socially isolated
relative to socially housed mice (Fig. 3A). These data were
confirmed through protein analysis, which also indicated a
decrease in central IL-6 protein expression in socially isolated
mice (Fig. 3B).
Despite conflicting data on the functional role of IL-6 (29–32),
a critical neuroprotective role during an ischemic event (30, 31).
Intracerebroventricular (ICV) administration of IL-6 reduces
infarct size, possibly through a mechanism involving suppressed
excitotoxicity (30, 31). Likewise, blockade of IL-6 signaling
results in increased apoptotic cell death and infarct size, as well
as poor neurological outcome (32). To address a role for central
IL-6 as a mediator of the social housing effects on stroke
outcome, mice were treated with an IL-6 neutralizing antibody
or vehicle aCSF before MCAO. Treatment with the IL-6 anti-
body increased infarct volume in the socially housed group and
eliminated the effect of social housing condition on infarct size
(Fig. 4A). In contrast to reported effects of IL-6 on infarct
volume (31), antibody treatment in our study did not affect
infarct volume of socially isolated mice. One possible explana-
tion for the absence of an effect among socially isolated mice is
that poststroke central gene expression and protein concentra-
tions of IL-6 in isolated mice were similar (or even lower) within
the ischemic compared with the nonischemic hemisphere in our
study; however, IL-6 was significantly elevated in the ischemic
hemisphere of socially housed mice. Thus, the use of neutralizing
antibody may reveal a ‘‘floor effect,’’ whereby IL-6 levels in the
socially isolated mice cannot be further reduced. On the other
hand, preventing the increase in IL-6 signaling via the neutral-
izing antibody potentiated infarct development in socially
In addition to measuring central IL-6 protein levels, we
assessed circulating concentrations of IL-6. Our data indicate
that although central IL-6 is down-regulated (Fig. 3B), periph-
eral levels of IL-6 protein are up-regulated (Fig. 3C) in isolated
relative to socially housed mice. This association between ele-
vated levels of IL-6 and increased infarct size is consistent with
the clinical literature on serum IL-6 concentration and stroke
outcome. Within the clinical literature, elevated peripheral IL-6
is a reliable predictor of stroke occurrence, severity, and mor-
tality (33, 34). The relationship between peripheral IL-6 and
stroke outcome is indicative of an increased proinflammatory
state, largely because of IL-6 mediated signaling of acute phase
to its central actions, peripheral IL-6 is proinflammatory and is
therefore a target of ongoing clinical trials for stroke patients
(36). Data from the current study indicate that social housing
condition influences both the neuroinflammatory and systemic
inflammatory response to stroke. Importantly, both the central
and peripheral IL-6 protein expression assays were performed in
the same cohort of animals. Taken together, an up-regulation of
peripheral IL-6, along with low central IL-6 expression, is
consistent with an altered inflammatory state that contributes to
poorer ischemic outcome in the socially isolated mice. Further,
the increase in serum IL-6 among socially isolated mice is
consistent with a previous report of increased intraischemic
serum CRP concentrations in isolated relative to socially housed
mice (8). Additionally, ICV treatment with the IL-6 antibody
eliminated this group difference in serum IL-6 concentrations
(Fig. 4B). An increase in serum IL-6 likely reflects an increase
in infarct volume that occurred after treatment with the IL-6
antibody. In the current study, serum IL-6 concentrations are
related to infarct size and do not appear to be independently
modulated by social interaction in the postischemic period.
Another physiological system known to contribute to the
extent of ischemic injury is the hypothalamic-pituitary-adrenal
(HPA) axis. The HPA axis functions in part to coordinate the
body’s physiological response to stressors by regulating glucocor-
ticoid release (37). CORT plays an important modulatory role
in ischemic cell death (38–40). After restraint stress, elevated
postischemic serum CORT concentrations influence infarct size
and functional outcome (40); in humans, poststroke cortisol
concentration predicts mortality (41). Because social isolation is
a stressor among several species, including Mus poschiavinus
(42, 43), and is often associated with altered HPA-axis respon-
sivity (44), circulating CORT was measured in the current study
were similar between socially housed and socially isolated mice
at both of these time points, despite a housing difference in
infarct size in the 24 h cohort (Table S3). Although the data from
the current study do not support a role for CORT underlying
housing effects on infarct size, it remains possible that there may
have been group differences in CORT concentration at earlier
time points or that the stress of social isolation may be influ-
encing IL-6 and infarct through a CORT-independent mecha-
nism. Additional research is necessary to identify the upstream
Despite significant differences in infarct size and edema, it was
not apparent through the behavioral testing conducted at 72 h
that a reduction in ischemic damage was associated with a
reduction in behavioral deficits (Fig. S1). One possible expla-
nation is that among socially housed mice, there remained
sufficient damage to surviving neurons that contributed to
functional deficits in these mice. We have previously reported
functional outcome deficits in socially isolated relative to socially
housed mice after stroke (8). However, the behavioral assess-
ments in the previous study were conducted at a later time point,
suggesting that over time, socially housed mice may be better able
to recover from functional deficits than socially isolated mice.
Measures of perceived social isolation or social support are as
powerful, and in some cases more powerful, predictors of
outcome than measures of actual social isolation or support in
clinical studies examining health and well-being (45–47). It is not
possible to differentiate between actual and perceived social
isolation in mice, nor is there a measure in mice that would be
comparable with social support in humans; however, the current
study provides evidence that the presence or absence of a
cohabitating conspecific is sufficient to alter stroke pathogenesis
and outcome. Furthermore, the current study identifies differ-
ential expression of IL-6 as one factor contributing to the
difference in infarct size between socially housed and isolated
mice. Both social isolation and elevated serum IL-6 concentra-
tions are associated with poor outcome in human stroke patients
(1, 33, 34), but whether there is a causal link between these 2
factors in humans, as there appears to be in mice, will need to
social interaction on IL-6 expression in human stroke patients
would also be informative.
In summary, socially isolated mice were less likely to survive
a stroke and had increased infarct volumes and edema compared
with socially housed mice. The increase in ischemic damage
among socially isolated mice was accompanied by an altered
neuroinflammatory response that was consistent with a neuro-
compromising influence of social isolation. Poststroke IL-6
signaling was down-regulated in the CNS and up-regulated in the
periphery among socially isolated mice. Further, treatment with
the IL-6 neutralizing antibody eliminated the effect of social
www.pnas.org?cgi?doi?10.1073?pnas.0810737106Karelina et al.
housing on infarct size. Although numerous reports exist on
neuroinflammatory measures in ischemia, they rarely describe
housing conditions of the experimental mice, making it difficult
to interpret those data independent of social/environmental
influences. To our knowledge, the current study is the first to
investigate the modulation of neuroinflammatory responses by
social housing after experimental stroke. Taken together, these
data support a causal role for IL-6 underlying the increase in
ischemic injury associated with social isolation and provide
evidence that social modulation of immune function can signif-
icantly influence stroke outcome.
Materials and Methods
Animals. Adult male C57/BL6 mice (23–30 g) (Charles River) were maintained
on a 14:10 light/dark cycle in a temperature- and humidity-controlled vivar-
ium. All animals were allowed ad libitum access to food and water. Experi-
mental animals were housed either individually (socially isolated) or with an
ovariectomized female (socially housed) for a period of 2 weeks before
surgery and throughout the reperfusion period. The study was conducted in
accordance with National Institutes of Health guidelines for the care and use
of animals and under protocols approved by the institutional animal care and
The influence of social housing on measures of stroke outcome was assessed
in separate cohorts of mice at 5 different reperfusion periods. In experiment
infarct size at 24 h (pair-MCAO, n ? 10; single-MCAO, n ? 10), 72 h (pair-
MCAO, n ? 13; single-MCAO, n ? 11; pair-SHAM, n ? 6; single-SHAM, n ? 6),
or 7 days of reperfusion [pair-MCAO, n ? 8; pair-SHAM, n ? 10; single-MCAO,
n ? 4 (6 died before sampling); single-SHAM, n ? 10]. Edema was determined
at 48 h, the earliest time point at which secondary damage is observed after
MCAO (pair-MCAO, n ? 6; single-MCAO, n ? 6).
In experiment 2, gene expression of inflammatory markers was measured
in the cortex and striatum after stroke. Tissue was collected from separate
cohorts of animals at 12 and 24 h of reperfusion (pair-MCAO, n ? 6 per time
point; single-MCAO, n ? 6 per time point).
Experiment 3 was designed to test the role of central and peripheral levels
of IL-6 in mediating the effects of social interaction on stroke outcome. In
experiment 3a, blood and tissue were collected at 24 h of reperfusion (pair-
MCAO, n ? 6; single-MCAO, n ? 6) for protein assay. In experiment 3b, mice
were treated with IL-6 neutralizing antibody (10 ng) or vehicle (aCSF) 1 h
before MCAO. Blood and brain tissue were harvested at 24 h of reperfusion
and assessed for infarct volume and circulating IL-6 protein concentration
(pair-MCAO-IL6 antibody, n ? 7; pair-MCAO-aCSF, n ? 6; single-MCAO-IL6
antibody, n ? 7; single-MCAO-aCSF, n ? 7). The ELISA for IL-6 requires a large
amount of blood; among socially housed mice, 2 samples from the IL-6
large to allow the assay. For determination of blood CORT and protein IL-6
concentrations, see SI Materials and Methods.
Surgery. Transient focal cerebral ischemia was induced by MCAO. The mice
were anesthetized with 1.5% isofluorane in oxygen-enriched air provided
through a face mask. Body temperature was maintained at 37 ? 0.5 °C
through the use of a homeothermic blanket system. Briefly, unilateral right
MCAO was achieved by insertion of a 6–0 nylon monofilament into the
internal carotid artery to a point 6 mm beyond the internal carotid-
pterygopalatine artery bifurcation. After 60 min of occlusion, the animal was
reanesthetized and reperfusion was initiated by removal of the filament. For
a detailed description of the MCAO procedure and determination of stroke
volume and edema, see SI Materials and Methods.
Behavioral Testing. Animals in Experiment 1 underwent paw preference and
open field behavioral testing. Both tests were conducted under similar envi-
ronmental conditions (i.e., lighting, temperature, level of background noise,
and time of day) at 24 h before MCAO and again at 72 h of reperfusion.
However, it is important to note that the mouse’s familiarity with the testing
environment was different at baseline testing (their first exposure to the
testing chamber) and postsurgical testing (their second exposure to the
testing chamber), which complicates comparison of behavior across these 2
time points. Thus, emphasis was placed on comparing experimental groups
independently at each time point. Behavioral testing was conducted during
the light phase and scored by an individual who was not aware of group
assignment. The apparatuses were thoroughly cleaned between animals us-
ing a 70% alcohol solution (see SI Materials and Methods).
Real-Time PCR. RT-PCR was conducted at 12 and 24 h of reperfusion after
MCAO. Bilateral samples were dissected from the cortex and striatum, and
and an RNeasy Mini Kit (Qiagen) according to manufacturer’s protocol. Ex-
tracted RNA was suspended in 30 ?L of RNase-free water, and RNA concen-
tration was determined by a spectrophotometer (NanoDrop ND-1000). The
following inventoried primers and probes (Applied Biosystems) were used:
GFAP, MAC-1, interleukins IL-6 and IL-1?, TNF?, COX-2, and TGF-?. A TaqMan
18S rRNA primer and probe set, labeled with VIC dye (Applied Biosystems),
were used as a control gene for relative quantification. Amplification was
performed on an ABI 7000 Sequencing Detection System by using Taqman
Universal PCR master mix. The universal 2-step RT-PCR cycling conditions used
were:50 °Cfor2min,95 °Cfor10min,followedby40cyclesof95 °Cfor15sec
and 60 °C for 1 min.
Intracerebroventricular Cannulation and IL-6 Neutralizing Antibody Injection. A
guide cannula, targeting the left lateral ventricle, was implanted 1 week
before experimental stroke surgery. The mice were anesthetized with
1%–1.5% isofluorane in oxygen-enriched air and were placed in a stereotaxic
apparatus (David Kopf Instruments). An incision was made along the midline
positioned at ?0.02 mm posterior and ?0.95 mm lateral to bregma and
secured with glue. Once the glue was dry, a dummy cannula was inserted into
the guide cannula, and the mice were placed into their home cages for
recovery. The neutralizing antibody to IL-6 (10 ng in 2 ?L vehicle) (zcomR&D
has been used successfully to neutralize IL-6 signaling in mice (48). According
to the manufacturer, this dose is within range of the 50% neutralization dose
Ab, AF-406-NA). The solutions were administered over 30 sec by using a 5-?L
Hamilton syringe. Correct cannula placement was confirmed through cresyl
CORT concentrations, and serum IL-6 protein concentrations were analyzed
via a 2-way ANOVA (factors were surgery and housing), a one-tailed t test
where appropriate (edema), or by using nonparametric statistics (Mann–
via 3-way ANOVA (factors were hemisphere, reperfusion period, and hous-
ing). Further, PCR data were also expressed as a ratio of ipsilateral to con-
tralateral hemisphere (R/L) gene expression and were analyzed via 2-way
ANOVA (factors were reperfusion period and housing). Significant ANOVA
results were followed by a Tukey HSD post hoc test. Behavior was analyzed by
independent 2-way ANOVAs (factors were surgery and housing) at baseline
and 72 h postsurgery because novelty of the testing environment may have
differentially influenced behavior at these 2 time points; for the purpose of
this study, across-group comparisons at the postsurgical time point are more
informative than within group comparisons between baseline and the post-
squared, ?2for parametric data) are reported for all relevant data.
ACKNOWLEDGMENTS. We thank Zachary Weil and James Walton for techni-
cal support and assistance with data analysis and Zachary Weil for critiquing
the manuscript. This work was supported by grants from the American Heart
ship to K.K.), National Institute of Neurological Disorders and Stroke Behav-
ioral Core Grant P30 NS045758 (to A.C.D.), National Institute of Neurological
Disorders and Stroke Grant RO1NS40267–05 (to A.C.D.), and National Heart,
Lung, and Blood Institute Grant RO1HL080249–01 (to A.C.D.).
1. Boden-Albala B, Litwak E, Elkind MS, Rundek T, Sacco RL (2005) Social isolation and
outcomes post stroke. Neurology 64:1888–1892.
2. Lett HS, et al. (2007) Social support and prognosis in patients at increased
psychosocial risk recovering from myocardial infarction. Health Psychol 26:418–
3. Barry LC, Kasl SV, Lichtman J, Vaccarino V, Krumholz HM (2006) Social support and
J Psychosom Res 60:185–193.
4. Ikeda A, et al. (2008) Social support and stroke and coronary heart disease: The JPHC
study cohorts II. Stroke 39:768–775.
Karelina et al.PNAS ?
April 7, 2009 ?
vol. 106 ?
no. 14 ?
5. Cohen S, Lemay EP (2007) Why would social networks be linked to affect and health
practices? Health Psychol 26:410–417.
6. Cacioppo JT, Hawkley LC (2003) Social isolation and health, with an emphasis on
underlying mechanisms. Perspect Biol Med 46:S39–S52.
7. Seeman TE (2000) Health promoting effects of friends and family on health outcomes
in older adults. Am J Health Promot 14:362–370.
9. Weil ZM, et al. (2008) Social isolation potentiates cell death and inflammatory re-
sponses after global ischemia. Mol Psychiatry 13:913–915.
10. Muir KW, Tyrrell P, Sattar N, Warburton E (2007) Inflammation and ischaemic stroke.
Curr Opin Neurol 20:334–342.
11. Everett BM, Kurth T, Buring JE, Ridker PM (2006) The relative strength of C-reactive
protein and lipid levels as determinants of ischemic stroke compared with coronary
heart disease in women. J Am Coll Cardiol 48:2235–2242.
12. Ladenvall C, et al. (2006) Serum C-reactive protein concentration and genotype in
relation to ischemic stroke subtype. Stroke 37:2018–2023.
13. Kuo HK, et al. (2005) Relation of C-reactive protein to stroke, cognitive disorders, and
depression in the general population: Systematic review and meta-analysis. Lancet
14. Spencer SJ, Mouihate A, Pittman QJ (2007) Peripheral inflammation exacerbates
damage after global ischemia independently of temperature and acute brain inflam-
mation. Stroke 38:1570.
15. Dziedzic T (2008) Clinical significance of acute phase reaction in stroke patients. Front
16. Cole SW, et al. (2007) Social regulation of gene expression in human leukocytes.
Genome Biol 8:R189.
inflammation in middle-aged and older adults: The Chicago health, aging, and social
relations study. Psychosom Med 68:376–381.
18. Ford ES, Loucks EB, Berkman LF (2006) Social integration and concentrations of
C-reactive protein among US adults. Ann Epidemiol 16:78–84.
19. Loucks EB, et al. (2006) Social networks and inflammatory markers in the Framingham
Heart Study. J Biosoc Sci 38:835–842.
20. Aschner M (1998) Immune and inflammatory responses in the CNS: Modulation by
astrocytes. Toxicol Lett 102–103:283–287.
21. Huang J, Upadhyay UM, Tamargo RJ (2006) Inflammation in stroke and focal cerebral
ischemia. Surg Neurol 66:232–245.
22. Wang Q, Tang XN, Yenari MA (2007) The inflammatory response in stroke. J Neuro-
23. Lai AY, Todd KG (2006) Microglia in cerebral ischemia: Molecular actions and interac-
tions. Can J Physiol Pharmacol 84:49–59.
24. Watanabe H, Abe H, Takeuchi S, Tanaka R (2000) Protective effect of microglial
conditioning medium on neuronal damage induced by glutamate. Neurosci Lett
25. Trendelenburg G, Dirnagl U (2005) Neuroprotective role of astrocytes in cerebral
ischemia: Focus on ischemic preconditioning. Glia 50:307–320.
26. Neumann J, et al. (2006) Microglia provide neuroprotection after ischemia. FASEB
27. Neumann J, et al. (2008) Microglia cells protect neurons by direct engulfment of
invading neutrophil granulocytes: A new mechanism of CNS immune privilege. J Neu-
28. Nagel S, et al. (2008) Minocycline and hypothermia for reperfusion injury after focal
cerebral ischemia in the rat-Effects on BBB breakdown and MMP expression in the
acute and subacute phase. Brain Res 1188:198–206.
29. Clark WM, et al. (2000) Lack of interleukin-6 expression is not protective against focal
central nervous system ischemia. Stroke 31:1715–1720.
30. Ali C, et al. (2000) Ischemia-induced interleukin-6 as a potential endogenous neuro-
protective cytokine against NMDA receptor-mediated excitotoxicity in the brain.
J Cereb Blood Flow Metab 20:956–966.
31. Loddick SA, Turnbull AV, Rothwell NJ (1998) Cerebral interleukin-6 is neuroprotective
during permanent focal cerebral ischemia in the rat. J Cereb Blood Flow Metab
32. Yamashita T, et al. (2006) Neuroprotection and neurosupplementation in ischaemic
brain. Biochem Soc Trans 34:1310–1312.
33. Smith CJ, et al. (2004) Peak plasma interleukin-6 and other peripheral markers of
stroke severity and long-term outcome. BMC Neurol 4:2.
34. De Simoni MG, et al. (2002) The inflammatory response in cerebral ischemia: Focus on
cytokines in stroke patients. Clin Exp Hypertens 24:535–542.
35. Rost NS, et al. (2001) Plasma concentration of C-reactive protein and risk of ischemic
stroke and transient ischemic attack: The Framingham study. Stroke 32:2575–2579.
36. Shenhar-Tsarfaty S, et al. (2008) Early signaling of inflammation in acute ischemic
stroke: Clinical and rheological implications. Thromb Res 122:167–173.
CNS. Brain Behav Immun 21:259–272.
38. Caso JR, Moro MA, Lorenzo P, Lizasoain I, Leza JC (2007) Involvement of IL-1beta in
acute stress-induced worsening of cerebral ischaemia in rats. Eur Neuropsychophar-
39. Sugo N, et al. (2002) Social stress exacerbates focal cerebral ischemia in mice. Stroke
40. DeVries AC, et al. (2001) Social stress exacerbates stroke outcome by suppressing Bcl-2
expression. Proc Natl Acad Sci USA 98:11824–11828.
41. Marklund N, Peltonen M, Nilsson TK, Olsson T (2004) Low and high circulating cortisol
levels predict mortality and cognitive dysfunction early after stroke. J Intern Med
42. Bartolomucci A (2007) Social stress, immune functions and disease in rodents. Front
43. DeVries AC, Craft TK, Glasper ER, Neigh GN, Alexander JK (2007) 2006 Curt P Richter
award winner: Social influences on stress responses and health. Psychoneuroendocri-
44. Serra M, Pisu MG, Floris I, Biggio G (2005) Social isolation-induced changes in the
hypothalamic-pituitary-adrenal axis in the rat. Stress 8:259–264.
45. Hawthorne G (2008) Perceived social isolation in a community sample: Its prevalence
and correlates with aspects of peoples’ lives. Social Psychiatry and Psychiatric Epide-
46. Uchino BN, Cacioppo JT, Kiecolt-Glaser JK (1996) The relationship between social
support and physiological processes: A review with emphasis on underlying mecha-
nisms and implications for health Psychological Bulletin 119:488–531.
47. Cacioppo JT, et al. (2002) Loneliness and health: Potential mechanisms. Am Psychoso-
matic Soc 64:407–417.
48. Meagher MW, et al. (2007) Interleukin-6 as a mechanism for the adverse effects of
social stress on acute Theiler’s virus infection. Brain Behav Immun 21:1083–1095.
www.pnas.org?cgi?doi?10.1073?pnas.0810737106Karelina et al.