Original Contribution
Combining Growth Factor and Bone Marrow Cell Therapy
Induces Bleeding and Alters Immune Response After
Stroke in Mice
Jan-Kolja Strecker, PhD; Joanna Olk; Maike Hoppen; Burkhard Gess, MD; Kai Diederich, PhD;
Antje Schmidt, MD; Wolf-Rüdiger Schäbitz, MD; Matthias Schilling, MD*; Jens Minnerup, MD*
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Background and Purpose—Bone marrow cell (BMC)–based therapies, either the transplantation of exogenous cells or
stimulation of endogenous cells by growth factors like the granulocyte colony–stimulating factor (G-CSF), are considered
a promising means of treating stroke. In contrast to large preclinical evidence, however, a recent clinical stroke trial on
G-CSF was neutral. We, therefore, aimed to investigate possible synergistic effects of co-administration of G-CSF and
BMCs after experimental stroke in mice to enhance the efficacy compared with single treatments.
Methods—We used an animal model for experimental stroke as paradigm to study possible synergistic effects of­
co-administration of G-CSF and BMCs on the functional outcome and the pathophysiological mechanism.
Results—G-CSF treatment alone led to an improved functional outcome, a reduced infarct volume, increased blood vessel
stabilization, and decreased overall inflammation. Surprisingly, the combination of G-CSF and BMCs abrogated G-CSFs’
beneficial effects and resulted in increased hemorrhagic infarct transformation, altered blood–brain barrier, excessive
astrogliosis, and altered immune cell polarization. These increased rates of infarct bleeding were mainly mediated by elevated
matrix metalloproteinase-9–mediated blood–brain barrier breakdown in G-CSF- and BMCs-treated animals combined with
an increased number of dilated and thus likely more fragile vessels in the subacute phase after cerebral ischemia.
Conclusions—Our results provide new insights into both BMC-based therapies and immune cell biology and help to understand
potential adverse and unexpected side effects. (Stroke. 2016;47:00-00. DOI: 10.1161/STROKEAHA.115.011230.)
Key Words: bone marrow cell transplantation ◼ cerebral ischemia ◼ granulocyte colony-stimulating factor
◼ hemorrhagic transformation ◼ macrophage polarization
S
troke is the third leading cause of death and the most
frequent cause of adulthood disability worldwide.1 Cellbased therapies, either transplantation of exogenous cells or
stimulation of endogenous cells, are considered a promising
means of treating stroke.2–6 Approaches using bone marrow–
derived cells (BMCs) are particularly attractive because BMCs
can be easily obtained and permit autologous transplantation.7
Besides transplantation of exogenous cells, endogenous BMCs
can be mobilized by the granulocyte colony–stimulating factor (G-CSF). A number of studies showed that both BMC and
G-CSF therapies improve neurobehavioral and histological
outcomes in animal models of cerebral ischemia.8,9 BMCs are
assumed to exert their actions by the production of cytokines
rather than by cell replacement.10,11 Cytokines secreted by
BMCs feature acute neuroprotective properties and stimulate
endogenous repair mechanisms, including neurogenesis and
angiogenesis, and support the survival of host neurons.11,12 In
addition to BMC-mediated mechanisms, G-CSF also exerts
direct effects on neurons via neuronal G-CSF receptors.13
However, in contrast to the large preclinical evidence on
G-CSF’s beneficial effects after stroke, a recent clinical stroke
trial on G-CSF was neutral.14 A potential reason might be that
the G-CSF-mediated BMC mobilization occurs too late, even
when G-CSF stimulation is initiated early after stroke onset.
We, therefore, investigated the hypothesis that the combination of G-CSF and BMCs is more effective than either single
treatment because transplantation of exogenous BMCs can
bridge the gap until G-CSF mobilizes endogenous BMCs
into the peripheral circulation (Figure 1A). Vice versa G-CSF
might enhance the efficacy of the exogenous BMC therapy
according to one of its genuine functions, that is, the increase
of survival of BMCs.
Received August 21, 2015; final revision received December 10, 2015; accepted December 31, 2015.
From the Department of Neurology, University of Münster, Albert-Schweitzer-Campus 1, Münster, Germany (J.-K.S., J.O., M.H., B.G., K.D., A.S.,
M.S., J.M.); and EVK Bielefeld, Bethel, Neurologische Klinik, Bielefeld, Germany (W.-R.S.).
Current address for B.G.: Department of Neurology, RWTH Aachen University, Aachen, Germany.
Guest Editor for this article was Eng H. Lo, PhD.
*Drs Schilling and Minnerup are joint senior authors.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.
115.011230/-/DC1.
Correspondence to Jan-Kolja Strecker, PhD, Department of Neurology, University Hospital Münster, Albert-Schweitzer-Campus 1, D-48149 Münster,
Germany. E-mail [email protected]
© 2016 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.115.011230
1
2 Stroke March 2016
Materials and Methods
Animal Care
All animal experiments were conducted in concordance with the local authorities (State Agency for Nature, Environment and Consumer
Protection (LANUV), North Rhine-Westphalia, Germany) according to the Animal Research: Reporting of In Vivo Experiments
(ARRIVE) guidelines.15 Male 16- to 18-week-old C57BL/6J (Charles
River, Germany; n=158) and GFP-transgenic mice (a kind gift by
Prof Masaru Okabe, C57BL/6J–GFP, backcrossed with C57BL/6J
mice at least 10 times; n=48) were used. Mice were maintained in
a pathogen-free laboratory with access to unlimited food and water.
Mice were randomly assigned to 4 groups: (1) placebo; (2) G-CSF
(10 μg/kg); (3) BMC (5×106 cells), and (4) G-CSF+BMC (n=19 each
group/time-point). Surgery, deficit evaluation, and further analyses were done blinded to treatment. We used human G-CSF, which
was demonstrated to exhibit a high specificity to the rodent G-CSF
receptor.
Transient Focal Cerebral Ischemia
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Focal ischemia was induced in 152 C57BL/6J-mice (25–30 g) by
occlusion of the middle cerebral artery (MCA) as previously described.16 Briefly, animals were anesthetized with isoflurane (1.5%
in 30% O2/70% N2O), and the left common carotid artery was separated through a midline neck incision. A silicon-coated monofilament
(0.2±0.1 mm; Doccol) was inserted into the left common carotid artery and advanced distal to the carotid bifurcation for occlusion of
the MCA. After 30 minutes, the suture was withdrawn for reperfusion and the wound closed. Before, during, and after middle cerebral
artery occlusion (MCAO), cerebral blood flow was monitored with
a laser Doppler probe (Periflux 5001; Perimed, Sweden). Shamoperated animals underwent the same procedure without occlusion
of the MCA.
BMC Transplantation and G-CSF Therapy
GFP mice17 were sacrificed, and femur veins were flushed with phosphate-buffered saline. After erythrolysis, BMCs were washed 3× in
phosphate-buffered saline and resuspended at 5×107 cells/mL. BMCtreated mice received 100 μL of the cell suspension (5×106 cells) 90
minutes after occlusion of the MCA by injection into the tail vein.2,18
Mice treated with G-CSF/G-CSF+BMCs received an immediate additional injection of G¬CSF (90 minutes after ischemia induction, 10
μg/kg IP; Figure 1A).
Functional Tests
The rotarod test was used to assess skilled motor performance. Mice
were placed on a Rotarod cylinder (4–50 rpm acceleration in 210
s; TSE-Systems, Germany), and time the animals remained on this
cylinder was measured. Mice were given 3 training sessions before
MCAO. Mean duration was recorded 1 day prior and daily after
MCAO for 7 days. Individual performance was calculated as percent
to pre-MCAO value.
Cylinder Test
The cylinder test was used to assess forelimb activity. Mice were
placed in a transparent cylinder (10 cm diameter) and videotaped for
2 min before MCAO and on day 1 and 7 post MCAO. Forelimb use
was counted, and the first 20 contacts were divided into use of (1)
both, (2) right, or (3) left forelimb.
Tissue Preparation and Infarct Volume
One or 7 days after MCAO, animals were perfused through the left ventricle with 6% hydroxyethyl starch (Fresenius, Germany) for 1 minute
and 4% buffered paraformaldehyde (pH 7.4) for 10 minutes under xylazine/ketamine anesthesia. Brains were removed, fixed (4% paraformaldehyde, 3 hours), immersed in 10% sucrose, frozen, and stored at −80°C.
Figure 1. Functional outcome and infarct volume assessment revealed beneficial effects of granulocyte colony–stimulating factor (G-CSF)
therapy. A, Schematic depicting the experiments rationale and experimental design. B, The rotarod test showed an improved outcome
in G-CSF-treated mice compared with the G-CSF+BMC-treated group on day 2 (P<0.05), day 3 (P<0.05), and day 7 (P<0.01;, Shapiro–
Wilk test for normal distribution, followed by repeated measures analysis of variance [ANOVA]; P<0.0001 and Bonferroni post hoc test,
n=15–19); *P<0.05, **P<0.01 (G-CSF vs GCSF+BMC). C, G-CSF and BMC treatment alone showed a trend toward a more frequent use
of the affected limb compared with placebo and G-CSF+BMC therapy in the cylinder test (ANOVA p=0.06). D, G-CSF-treatment resulted
in smaller infarct volumes compared with the placebo group on day 1 (P=0.019) and on day 7 post MCAO (P=0.0068, 1-way ANOVA with
Bonferroni post hoc test, n=8–10). *P<0.05, **P<0.01. BMC indicates bone marrow cell; and MCAO, middle cerebral artery occlusion.
Strecker et al Growth Factor Combined Bone Marrow Cell Therapy 3
Mice used for mRNA and protein analyses were perfused with 0.9%
NaCl, brains rapidly removed, frozen on dry ice, and stored at −80°C.
Coronal sections were cut with a cryostat (Leica CM 3050, Germany),
mounted on glass slides, and stored at −20°C. To assess ischemic damage, sections were collected in 300 μm intervals, stained with toluidine
blue, and digitized. Infarction was measured with ImageJ (http://imagej.
nih.gov/ij/), and infarct volumes were calculated as described.19
Macroscopic Evaluation of Intracerebral
Hemorrhage
We evaluated the intracerebral hemorrhage on a macroscopic level using
a modified scale according to Henninger et al.20 Intracerebral hemorrhage
was defined as (1) no hemorrhage, (2) hemorrhagic infarction (punctate
or confluent petechiae within the infarcted tissue), and (3) parenchymal
hematoma (large homogenous mass of blood within the infarct).
Gene Expression
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
RNA purification and reverse transcription were done following
the manufacturer’s protocol (RNeasy, Quantitect RT-Kit, Qiagen,
Germany). Primers were purchased from Qiagen: glycerinaldehyde3-phosphat-dehydrogenase; claudin-5; and occludin. Real-time
polymerase chain reactions were performed with ABI PRISM 7700
RT-PCR-System (Applied Biosystems). Standard conditions were
used and measurements done in duplicates.
Protein Preparation and Western Blot
Western blot analyses were done with brain lysates of the infarcted
hemisphere. Lysates were prepared by ultrasonic homogenization
in lysis buffer (50 mmol/L Tris, 150 mmol/L NaCl, 0.5% sodiumdeoxycholate, 1% NP-40, protease inhibitor; Roche, Switzerland).
After centrifugation (12 000g, 20 min, 4°C), lysates were collected
and protein adjusted to 1 mg/mL (BCA Assay, Pierce Chemical).
Lysate (12 μL) was separated on 12.5% SDS-polyacrylamide gels
and transferred to 4% to 20% gradient gels (MiniProtean, BioRad,
Germany). After blocking (5% wt/vol milk powder, TBS-T, 1 hour),
membranes were incubated with antibodies rabbit anti-albumin
(Serotec, Germany; 1:1000), rabbit anti-claudin-5 (Abcam, UK;
1:100), rabbit anti-occludin (Invitrogen; 1:200), and rabbit anti-CD31
(Abcam; 1:100). Detection of antibodies was done with horse-raddish
peroxidase–conjugated anti-rabbit antibody (1:5000, 45 min). Protein
bands were visualized with Lumi-LightPLUS (Roche).
Immunofluorescence Microscopy
Brain sections (bregma 0–1 mm) were pretreated in 3% H2O2/methanol (10 min) and Blocking Reagent (Roche Diagnostics, Germany; 15
min). Primary antibodies (Table I in the online-only Data Supplement)
were applied overnight (4°C). Neuronal nuclei (NeuN) and glial
fibrillary acidic protein (GFAP) were visualized with anti-mouseAlexaFluor594 (Molecular Probes; 45 min, 1:100) and Laminin with
anti-rabbit-AlexaFluor594 (Molecular Probes; 1:100, 4 °C, overnight).
Signal amplification of CD16/32, F4/80, 7/4, matrix metalloproteinase-9 (MMP-9), and CD31 was done with horse-raddish peroxidase–
conjugated streptavidin (DAKO, Denmark; 1:100, 45 min) and biotinyl
tyramide (1:100, 10 min). Visualization was done with streptavidinconjugated dye (Alexa Fluor 594, Molecular Probes, the Netherlands;
1:100, 45 min). 4′6-diamidino-2-phenylindole was applied for nuclear
counterstain (DAPI; Vector). Non-confocal images were taken with
Nikon Eclipse 80i microscope (Nikon GmbH, Germany). Confocal
images were taken with Axiovert microscope (Zeiss, Germany).
Quantification of Astrogliosis and Macrophage
Polarization
GFAP (as astrogliosis marker) and CD16/32 (as exemplary M1 marker) intensity were measured with densitometric analysis technique
described elsewhere.21 Briefly, whole hemisphere images (8 bit) underwent 2 processing steps using ImageJ: background noise was reduced
with a 20 pixel wide rolling-ball subtraction followed by median filter
(2 pixel wide). Remaining pixels between gray values 5 to 105 were
considered GFAP- or CD16/32-positive and summed for each brain.
Signal of placebo-treated mice was set as reference (100%).
Fluorescence-Activated Cell Sorting Analysis
Blood of BMC-transplanted mice was collected from the right atrium
before perfusion. Samples were washed in phosphate-buffered saline
for 3 times, and GFP+-cells were quantified by fluorescence-activated
cell sorting analysis.
MMP-9, High-Mobility-Group-Box-1, and
Interferon-γ In Vivo and In Vitro
MMP-9 concentration was measured using MMP-9 ELISA Kit
(BosterBio), high-mobility-group-box-1 (HMGB1) using HMGB1
ELISA kit (Biozol, Germany), and interferon-γ (IFN-γ) with IFN-γ
ELISA kit (R&D Systems).
MMP-9, HMGB1, and IFN-γ In Vivo
Brain tissue protein concentration was evaluated by applying 5 μg
total protein (n=4 per group) to each well.
MMP-9, HMGB1, and IFN-γ In Vitro
Purified BMCs were seeded in Dulbecco’s modified eagles medium
(Sigma)+10% fetal calf serum (Biochrome, Germany) and PenStrep
(Life Technologies, Germany) on 12-well plates (105 cells/well). After
24 hours, medium was replaced with medium, medium+G-CSF (10
ng/mL), medium+albumin (0.01 mmol), or medium+G-CSF+albumin
(each n=6, albumin was applied to mimic blood–brain barrier [BBB]
breakdown). After 24 hours (37°C), supernatant was collected for
protein evaluation. Cells were trypsinized and counted. Protein
amount was corrected for cell count and presented as ng/103 cells.
Evaluation of Blood Vessel Diameter
Ipsilateral vessel diameter was measured within 3 regions (cortex,
infarcted core, piriform cortex at bregma 0.25 mm; size of each measured area=65.2×103µm2) 7 days after MCAO using ImageJ analysis.
Vessel diameter means were calculated of 10 vessels per area per animal (n=5 per group).
Statistical Analysis
Statistics were done with GraphPad Prism (5.01, GraphPad Software).
Evaluations of intergroup effects were done with analysis of variance
after checking data for normal distribution (Shapiro–Wilk). When a
significant effect could be detected, a Bonferroni post hoc test for
multiple comparisons was applied. For comparison of 2 groups, unpaired 2-tailed student t tests were applied (quantification of immigrated GFP+ cells and fluorescence-activated cell sorting analysis).
Frequency of parenchymal hematoma was evaluated by using the
chi-square test. Results are presented as mean±SEM unless otherwise
stated. A P value of <0.05 was considered significant.
Results
Functional Outcome, Mortality, and Infarct Volume
Quantification
Rectal temperature, body weight, and cerebral blood flow were
not different between groups (see Table II in the online-only
Data Supplement). Four placebo, 1 G-CSF-, 5 BMC-, and 6
G-CSF+BMC-treated mice died during the survey and were
excluded. For evaluation of functional outcomes, each mouse
underwent the Rotarod and cylinder test. Animals treated
with G-CSF alone showed increased running time compared
with the G-CSF+BMC-treated mice (analysis of variance
4 Stroke March 2016
P<0.0001; Figure 1B). Both G-CSF and BMC treatment alone
showed a trend toward a more frequent use of the affected
limb compared with placebo and G-CSF+BMC therapy in
the cylinder test (analysis of variance, P=0.06; Figure 1C).
Furthermore, infarct volumes were reduced in G-CSF-treated
animals on day 1 and day 7 compared with placebo-treated
mice (Figure 1D).
Analysis of Transplanted BMCs
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Fluorescence-activated cell sorting quantification of circulating transplanted BMCs (GFP+) showed no significant changes
in cell counts between the investigated groups (P=0.18;
Figure 2A). Intracerebral evaluation of GFP+ cells showed
an increased number of grafted cells within the brains of
G-CSF+BMC-treated animals (P=0.039; Figure 2A). GFP+
cells were mainly located within the infarcted striatum and
cortex and were increased in G-CSF+BMCs-treated mice
compared with BMC monotherapy. Transplanted cells were
frequently seen in groups of 2 to 3 cells (Figure 2B). GFP+
cells occasionally expressed the microglia/macrophage marker
F4/80 (Figure 2C) and numerous showed ramified shape.
Some F4/80+/GFP+ cells were intermingled with NeuN+ cells,
suggesting phagocytosis of defect neurons (Figure 2D). GFP+
cells were rarely seen within vessels of the contralateral hemisphere. Random evaluation of spleen, liver, and lung showed
numerous GFP+ cells within the respective tissue (not shown).
G-CSF+BMC Combination Induces Hemorrhagic
Transformation and BBB Damage
During the preparation of the coronal brain slices, it came to
our attention that occasionally brains showed an unusual pronounced petechial or even parenchymal intracerebral bleeding.
To our surprise, parenchymal bleeding was only seen in mice
which received BMCs or G-CSF+BMCs (G-CSF+BMC post 7
days: P=0.031; Figure 3A). To investigate mechanisms of hemorrhagic transformation, we performed BBB leakage evaluation via quantification of intracerebral albumin. We measured
increased albumin in G-CSF+BMCs-treated mice on day 1
compared with placebo and G-CSF-treated mice (Figure 3B;
see Figure I in the online-only Data Supplement for whole
blots). To further evaluate BBB integrity, we quantified mRNA
and protein of claudin-5 and occludin. No differences of occludin expression could be found on day 1. On day 7, either
placebo or G-CSF+BMC-treated animals showed a marked
downregulation of occludin compared with the G-CSF- and
BMC-treated group. Protein analyses revealed an increase of
occludin signal in all investigated groups on day 1. Seven days
after MCAO, G-CSF+BMC-treated animals showed reduced
occludin protein compared with the placebo group, indicating
a critical influence of G-CSF+BMC on post ischemic BBB
opening and restructuring (Figure 4A). Claudin-5 expression
revealed no significant changes between the groups, whereas
protein analysis showed elevated claudin-5 protein in BMCand G-CSF+BMC-treated mice compared with the placebo and
G-CSF-treated group (Figure 4B). We next investigated mechanisms of BBB damage. MCAO resulted in pronounced MMP-9
secretion within the infarct. Combination of G-CSF+BMCs
caused increased MMP-9 secretion on day 1 (Figure 4C). The
vast majority of MMP-9+ cells were identified as neurons, endothelial cells, and GFP+ BMCs. In contrast, in vitro studies of
cultured BMCs showed markedly decreased MMP-9 in cells
treated with G-CSF and G-CSF+albumin (Figure 4D). MMP-9
was normalized for cell counts. Proliferation rates compared
Figure 2. G-CSF+BMC combination therapy results in increased immigration of grafted bone marrow cells. A, Fluorescence-activated cell
sorting (FACS) analysis showed no difference of GFP+ cells within the blood of BMC+G-CSF-treated mice (n=3 each). Dot-Blot analysis
of GFP+-cells: G-CSF+BMC treatment resulted in an increased influx of GFP+ cells (heat map of 2 sections of n=8 of each group, day 7;
*P=0.039, t test). B, GFP+ cells were found within the striatum and cortex, frequently seen in groups of 2–3 cells in close vicinity to blood
vessels. C, GFP+ cell expressing macrophage marker F4/80. D, Costaining with neuronal marker NeuN depicting GFP+ cell in close contact
with a neuron. BC indicates blood cells; BMC, bone marrow cell; G-CSF, granulocyte colony–stimulating factor; and NeuN, neuronal nuclei.
Strecker et al Growth Factor Combined Bone Marrow Cell Therapy 5
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 3. Bone marrow cell (BMC) grafting leads
to occasional poststroke parenchymal bleeding.
A, Particularly G-CSF+BMC treatment resulted in
occurrence of parenchymal bleeding (G-CSF+BMC
post 7 days: P=0.031). B, Albumin leakage during
the acute phase in BMC and G-CSF+BMC-treated
animals (analysis of variance [ANOVA], P=0.039;
*P<0.05 [Bonferroni post hoc test]; n=4 each;
mean±SEM, dotted line: sham value=100%±SEM).
G-CSF indicates granulocyte colony–stimulating
factor.
with placebo-treated cells were 138.46% (G-CSF), 111.54%
(albumin), and 208.65% (G-CSF + albumin).
Astrogliosis and Immune Cell Response
Evaluation of GFAP signal intensity indicated increased
astrogliosis in mice treated with G-CSF+BMCs compared
with placebo- and G-CSF-treated animals 7 days after MCAO
(Figure 5A). However, as intracerebral bleeding is known to
cause astrocyte activation,22 we investigated a potential association between bleeding intensity and GFAP level. Such
association could not be observed (Figure II in the online-only
Data Supplement). Because G-CSF+BMC treatment led to an
increased astrocyte activation, equating a more intense inflammatory response, we investigated a potential altered polarization of involved immune cells. Hence, we decided to evaluate
CD16/32 as representative M1-macrophage marker. Indeed,
CD16/32 signal intensity showed a marked upregulation in
BMC- and G-CSF+BMC-treated mice 7 days after MCAO
(Figure 5B). One day after MCAO, CD16/32 signal could
be mainly observed surrounding the infarcted core, whereas
after 7 days, CD16/32+-cells could be seen particularly within
the infarct core and, in case of BMC- and G-CSG+BMCtreated mice, also within the adjacent cortex (Figure 6C).
Quantification of CD16/32+-cells showed no difference
between treatments on day 1. Seven days after MCAO, however, CD16/32+ cell counts were significantly increased in
BMC- and G-CSF+BMC-treated animals (Figure 6D).
HMGB1 and IFN-γ
Because we observed an increased bleeding susceptibility
of BMC- and G-CSF+BMC-treated mice in combination
with increased astrogliosis, we investigated whether these
mechanisms could be mediated by the damage-associated
molecular pattern named HMGB1. HMGB1 is known to be
6 Stroke March 2016
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 4. A, Occludin expression/protein concentration 1 and 7 days after MCAO (mRNA: analysis of variance [ANOVA] P<0.0001; n=5
each. Protein: ANOVA P=0.030; n=4 each). B, Claudin-5 expression/protein concentration 1 and 7 days after MCAO (mRNA: ANOVA
P>0.05; n=5 each. Protein: ANOVA P<0.001; n=4 each). C, Matrix metalloproteinase-9 (MMP-9) levels within the infarct obtained by
ELISA (ANOVA P=0.0211; n=4 each [dotted line: sham value±SEM]). IHC, Autofluorescent erythrocytes within the infarcted striatum in
close vicinity to MMP-9-stained cells. The majority of MMP-9+ cells were identified as neurons, administered GFP+, and endothelial cells.
D, MMP-9 protein concentration in supernatants of cultured bone marrow cells (BMCs). MMP-9 concentration was normalized for cell
numbers in each well (ANOVA P=0.010; n=6 each, data shown as ng/103 cells). Protein and mRNA levels are shown as relative change
compared with sham group. IHC indicates immunohistochemistry; and MCAO, middle cerebral artery occlusion. *P<0.05, **P<0.01,
***P<0.001; Bonferroni post hoc test.
released by reactive astrocytes and to promote neurovascular remodeling during stroke recovery.23 No differences
could be seen between treatment groups on day 1. However,
7 days after MCAO, G-CSF+BMC-treated animals showed
a trend toward a suppressed HMGB1 release (P=0.07) compared with the placebo group (Figure 6A). Concomitantly,
in vitro analysis also showed a reduced HMGB1 release by
BMC after G-CSF+albumin treatment, which could further indicate an altered inflammatory milieu triggered by
the G-CSF stimulated cell graft. We next checked whether
increased astrogliosis, CD16/32+-cells, and altered HMGB1
release led to an altered overall inflammation. We chose
IFN-γ as representative marker because microglia and
immigrating macrophages can be stimulated by IFN-γ to
the pro-inflammatory M1-phenotype.24 G-CSF and, to our
surprise, G-CSF+BMC treatment resulted in a decreased
IFN-γ concentration within the infarcted hemisphere 7 days
after MCAO (Figure 6B). In vitro analysis confirmed the
reduced IFN-γ release in cultured cells after G-CSF and
albumin treatment.
Strecker et al Growth Factor Combined Bone Marrow Cell Therapy 7
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 5. Evaluation of astroglia and CD16/CD32+-cell activation/distribution. A, GFAP immunohistochemistry showed increased astrogliosis in mice treated with G-CSF+BMC (analysis of variance [ANOVA] P=0.0005; n=8–10). B, CD16/CD32+ staining revealed an increased
signal intensity in BMC or G-CSF+BMC-treated mice 7 days after MCAO (black and white inserts: densitometric surface plot [ANOVA
P<0.0001; n=8–9]; dotted line: placebo value=100%±SEM). C, Heat map depicting the distribution pattern of CD16/CD32+ cells and
revealed an increased area of CD16/CD32 expression in BMC- and BMC+G-CSF-treated mice 7 days after MCAO (ANOVA P=0.0052;
n=8–10). D, BMC and G-CSF+BMCs treatment resulted in increased CD16/CD32+ cell counts within the infarcted hemisphere (ANOVA
P<0.0001; n=7–9). *P<0.05, **P<0.01, ***P<0.001; *vs placebo, #vs G-CSF, $vs BMC, Bonferroni post hoc test. BMC indicates bone marrow cell; G-CSF, granulocyte colony–stimulating factor; GFAP, glial fibrillary acidic protein; and MCAO, middle cerebral artery occlusion.
Combination Therapy Increases Vessel Diameters
Within the Ischemic Hemisphere
For further investigation of mechanisms underlying hemorrhagic transformation, we evaluated poststroke vessel integrity.
Mice treated with G-CSF+BMCs showed increased mean blood
vessel diameter compared with placebo and G-CSF-treated animals (Figure 6C). Analysis of CD31, which is mainly expressed
by endothelial cells, showed a diminished reduction of CD31
protein in G-CSF-treated mice compared with placebo, BMCs,
and G-CSF+BMCs groups on day 1 (Figure 6D).
Discussion
The present study confirms that G-CSF treatment results in
smaller infarct volumes and improved functional outcome.13,25
According to our hypothesis, G-CSF co-treatment increased
the number of transplanted BMCs within the ischemic brain
compared with BMC therapy alone. Because G-CSF is
noticeably upregulated as early as 2 hours after ischemia13
and earlier treatment is usually associated with an improved
outcome, we decided to first examine an early treatment initiation with G-CSF. As the combination therapy, in contrast
to our hypothesis, showed detrimental effects, we refrained
from investigating further times of treatment initiation but
evaluated mechanisms underlying the observed deteriorated
outcomes. The neurological deterioration after combination
therapy was frequently accompanied by hemorrhagic infarct
transformation. We identified MMP-9-mediated BBB loss as
reason for the increased susceptibly to intracerebral bleeding.
We also found excessive astrogliosis and altered immune cell
polarization after co-treatment of G-CSF and BMCs.
Several previous studies have shown that grafting of either
mononuclear or mensenchymal bone marrow cells promotes
neuroprotection and angiogenesis.3,18,26,27 Furthermore, it has
been shown that immigrating bone marrow cells are capable
to respond to local epigenetic stimuli and to differentiate
into neural and glial phenotypes.28 In this study, application
8 Stroke March 2016
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Figure 6. A, High-mobility-group-box-1 (HMGB1) protein analysis within the infarcted hemisphere showed no differences between the
investigated groups on day 1 post MCAO and a trend toward decreased HMGB1 in BMC+G-CSF-treated mice compared with placebo.
HMGB1 protein concentration in supernatants of cultured BMCs revealed a decreased HMGB1 secretion by G-CSF+albumin-treated cells
(analysis of variance [ANOVA] P=0.0225; n=6 each). B, G-CSF and G-CSF+BMC treatment led to a reduced interferon-γ (IFN-γ) secretion in
mice subjected to MCAO (post 7 days, ANOVA P=0.0002; n=4 each). Sole G-CSF or albumin treatment resulted in a reduced interferon-γ
(IFN-γ) release by cultured BMC (ANOVA P=0.0058; n=6 each). HMGB1 and IFN-γ protein concentration were normalized for cell numbers
in each well (n=6 wells each, data shown as ng/103 cells). C, G-CSF+BMC-treated mice showed an increased number of dilated blood vessels (ANOVA P<0.001; 30 vessels per animal measured, n=5 per group). D, CD31 levels obtained by Western blot analysis (ANOVA P=0.033;
n=4 each; dotted line: sham value=100%±SEM). *P<0.05, **P<0.01, ***P<0.001; *vs placebo, #vs G-CSF, $vs BMC; Bonferroni post hoc test.
BMC indicates bone marrow cell; G-CSF, granulocyte colony–stimulating factor; and MCAO, middle cerebral artery occlusion.
of BMCs resulted in minor improved functional outcome but
failed to reduce infarct volume. Injected BMCs immigrated
into the ischemic tissue and could be found beyond vessel
structures. However, none of the engrafted cells coexpressed
neuronal markers, likely because of the early investigation
time point. Numerous cells coexpressed the macrophage
marker F4/80 and showed ramified morphology. Additional
G-CSF stimulation resulted in an increase of intracerebral
GFP cells, indicating either G-CSF-mediated enhanced proliferation or survival of grafted cells.
Strecker et al Growth Factor Combined Bone Marrow Cell Therapy 9
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Previous studies revealed a crucial role of bone marrow
cells for the development of hemorrhagic infarct transformation.29,30 Subsets of monocytes/macrophages were demonstrated to be essential for integrity maintenance of the
neurovascular unit after cerebral ischemia.30 Depletion of
circulating monocytes/macrophages or prevention of CCR2dependent recruitment of monocytes caused hemorrhagic
transformation in murine stroke models. The finding of an
increased hemorrhagic transformation after combination of
G-CSF therapy and BMCs is of great importance from a translational perspective because bleeding in the infarct is associated with worse neurological outcomes in stroke patients.
We, therefore, investigated potential underlying mechanisms
of this phenomenon.
Analysis of intracerebral albumin as indicator for BBB
leakage revealed a markedly increased albumin influx in
BMC- and G-CSF+BMC-treated mice already 24 hours
after MCAO. Furthermore, altered claudin-5 and occludin
levels in G-CSF+BMC-treated mice point toward unwanted
side effects, respectively increased BBB damage triggered
by BMC application. Regarding the formation of ischemiainduced BBB damage, MMPs are well-known to contribute to
BBB breakdown by degrading endothelial basal lamina.4,31,32
We, therefore, investigated the secretion of MMP-9 and
found a noticeable increase of intracerebral MMP-9 in
all groups that was markedly elevated in mice treated with
G-CSF+BMC. Immunohistological analysis revealed particularly neurons and endothelial cells as MMP-9 source, confirming the results of an earlier study.33 In addition, grafted BMCs
were also identified as MMP-9 source in vivo. However, our
in vitro analyses of cultured BMCs showed that G-CSF and
G-CSF+albumin treatment led to decreased MMP-9 production. We thus conclude that the adjacent cells trigger either
infiltrating BMCs to an alternate state or that BMCs alter
the cytokine milieu within the injured tissue, resulting in
an increased neuronal and endothelial MMP-9 release. The
idea of a BMC-mediated cytokine release modification is
supported by the finding of a noticeably increased reactive
astrogliosis and altered immune cell polarization toward the
pro-inflammatory M1 phenotype in G-CSF+BMC-treated
animals. Increased astrogliosis has been recently linked to
enhanced angiogenesis via the HMGB1, a protein secreted
into the ischemic tissue by reactive astrocytes after stroke.34
Additionally, HMGB1 induces toll-like-receptor-4–mediated
MMP-9 release and augments inflammation by increasing
the release of tumor necrosis factor-α, interleukin-1β, and
other cytokines.31 To our surprise though, G-CSF+BMC treatment resulted in a decreased HMGB1 and IFN-γ release.
Furthermore, G-CSF+BMC treated mice showed increased
numbers of dilated blood vessels. As MMP-9 secretion is
known to further support neovascularization35 in combination
with reduced angiogenesis-promoting HMGB1, this could
point toward insufficient formation of neovessels or impaired
vessel repair and further increased susceptibility to hemorrhagic transformation after G-CSF+BMC treatment. The idea
of a potential adverse role of the cell graft on BBB integrity is
supported by recent findings of dela Peña et al,36 who showed
that sole application of G-CSF exerts crucial effects on vascular stability, leading to attenuated delayed tissue plasminogen
activator–induced hemorrhagic transformation and enhanced
vasculo- and angiogenesis.
Evaluation of microglia and macrophages showed no differences in cell numbers between the investigated groups but
an enlarged area of activated microglia within the infarct after
G-CSF treatment, despite smaller infarct volumes (see Figure
III in the online-only Data Supplement). This could indicate
a beneficial role of early microglial activation by G-CSF, as
suggested in a current study on spinal cord injury.37 Today
it is suggested that microglia and infiltrating macrophages
show a dynamic response to ischemia by either polarizing
toward an anti-inflammatory and angiogenesis-promoting
(M2) or to a detrimental pro-inflammatory M1 phenotype.38
The reduced CD16/32+ cells in G-CSF-treated mice compared
with the C-CSF+BMC group could indicate a microglial shift
towards less detrimental or even beneficial action during the
acute phase. Cleary, further studies need to address whether
G-CSF and BMCs exercise influence on glial and leukocyte
polarization.
Our data showing increased claudin-5 and occludin protein
in connection with increased BBB leakage in G-CSF+BMCstreated mice is counterintuitive at a first glance. However, in
a previous study, we could show that extended occlusion time
of 45 minutes39 resulted in minor changes of claudin-5 and
occludin protein concentration. These findings were supported
by the results of Stamatovic et al,40 which suggest that claudin-5 and occludin are, in consequence of oxygen shortage,
not decomposed by proteinases, but stored away within endothelial endosomes. This would make both proteins available
for later reestablishment of the BBB. Therefore, our finding
of increased BBB protein could be as a result of demasking of
the unbound proteins within endosomes, making them more
susceptible to antibody staining.
Our finding that transplantation of BMCs abrogates
G-CSFs’ beneficial effects confirm results of our ancestor
study.41 In this previous study, we found that G-CSF+BMC
treatment results in accumulation of endogenous granulocytes
and grafted cells within the spleen, which in turn led to splenic
overload, granulocyte clearance failure, and increase of circulating and intracerebral granulocytes. In the present study, we
also observed increased granulocyte numbers within the ischemic lesion of G-CSF+BMC-treated animals, but not in mice
treated with G-CSF alone. The finding that BMC administration abrogates the beneficial effects of G-CSF after murine
stroke might provide a potential explanation why G-CSF
monotherapy has been beneficial in a number of animal studies, but failed to improve neurological outcome in a clinical
study.14 Mice have lower numbers of circulating granulocytes
compared with humans and G-CSF-induced neutrophilia
is more pronounced in humans compared with mice.14,42,43
Therefore, growth factor–combined cell therapy might
cause excessively increased leukocyte numbers in mice that
resemble effects of G-CSF monotherapy on peripheral white
blood count in humans. The assumption that such excessive
increased peripheral leukocytes are detrimental and abolish
potential neuroprotective effects of G-CSF after stroke is also
supported by a previous study, which found particularly high
doses of G-CSF to be detrimental after cerebral ischemia in
mice.44 Therefore, future preclinical investigations applying
10 Stroke March 2016
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
bone marrow cell–based therapies should consider differences
in blood composition in between rodents and humans.
Our study has strengths and limitations. Strengths are
large animal numbers and multimodal mechanistic experimental approaches. A potential limitation of this work regarding its comparability to other studies using BMCs in animal
stroke studies could be the preparation of the transplanted
cells. Many promising preclinical studies describing beneficial effects during stroke recovery used a plethora of different
subsets, for example, mesenchymal, hematopoietic, or endothelial progenitor cells, which could be a reason why sole injection of BMC resulted only in minor functional improvement.
However, as the preparation needs prior culturing lasting up to
several days, our aim was to study the efficacy of primary cells
prepared directly from the bone marrow, which are available
within hours. In addition, we cannot rule out adverse effects
driven by a potential mismatch of the applied human clinical
grade G-CSF (Neupogen) and the mice-derived BMC, leading to the observed altered immune response. We also assert
that further adjusting of the G-CSF application timing in relation to the cell grafting may reduce the adverse effects of the
combination therapy. A further limitation is based on the fact
that we cannot account for pretreatment stroke severity, which
could potentially differ between groups.
In summary, our results confirm the efficacy of both G-CSF
treatment and BMC monotherapy in experimental stroke.
Contrary to our hypothesis, the combination of G-CSF+BMC
application led to poorer functional outcome and surprisingly
increased intracerebral bleeding. The latter resulted from
G-CSF+BMC co-therapy–mediated upregulation of cerebral
MMP-9 with subsequent BBB damage and altered inflammation. Our findings provide new insights into bone marrow–
based stroke therapies and reveal potential side effects.
Sources of Funding
This study was supported by grants of the Else-Kröner Fresenius
Stiftung (Grant no. 2010_A64) and of the Bundesministerium für
Bildung und Forschung (BMBF, Project MARS, 01GN0980).
Disclosures
Dr Schäbitz is an inventor on a patent claiming the use of granulocyte colony-stimulating factor in stroke treatment. The other authors
report no conflicts of interest. This work forms part of the doctoral
thesis of Joanna Olk.
References
1. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V,
et al. Global and regional mortality from 235 causes of death for 20 age
groups in 1990 and 2010: a systematic analysis for the Global Burden
of Disease Study 2010. Lancet. 2012;380:2095–2128. doi: 10.1016/
S0140-6736(12)61728-0.
2. Fujita Y, Ihara M, Ushiki T, Hirai H, Kizaka-Kondoh S, Hiraoka M, et al.
Early protective effect of bone marrow mononuclear cells against ischemic white matter damage through augmentation of cerebral blood flow.
Stroke. 2010;41:2938–2943. doi: 10.1161/STROKEAHA.110.596379.
3.Moniche F, Gonzalez A, Gonzalez-Marcos JR, Carmona M, Piñero
P, Espigado I, et al. Intra-arterial bone marrow mononuclear cells in
ischemic stroke: a pilot clinical trial. Stroke. 2012;43:2242–2244. doi:
10.1161/STROKEAHA.112.659409.
4.Suofu Y, Clark JF, Broderick JP, Kurosawa Y, Wagner KR, Lu A.
Matrix metalloproteinase-2 or -9 deletions protect against hemorrhagic transformation during early stage of cerebral ischemia and
reperfusion. Neuroscience. 2012;212:180–189. doi: 10.1016/j.
neuroscience.2012.03.036.
5. Wang J, Yu L, Jiang C, Chen M, Ou C, Wang J. Bone marrow mononuclear cells exert long-term neuroprotection in a rat model of ischemic
stroke by promoting arteriogenesis and angiogenesis. Brain Behav
Immun. 2013;34:56–66. doi: 10.1016/j.bbi.2013.07.010.
6. Wang J, Fu X, Jiang C, Yu L, Wang M, Han W, et al. Bone marrow
mononuclear cell transplantation promotes therapeutic angiogenesis via
upregulation of the VEGF-VEGFR2 signaling pathway in a rat model of
vascular dementia. Behav Brain Res. 2014;265:171–180. doi: 10.1016/j.
bbr.2014.02.033.
7. Savitz SI, Misra V, Kasam M, Juneja H, Cox CS Jr, Alderman S, et al.
Intravenous autologous bone marrow mononuclear cells for ischemic
stroke. Ann Neurol. 2011;70:59–69. doi: 10.1002/ana.22458.
8.Minnerup J, Heidrich J, Wellmann J, Rogalewski A, Schneider A,
Schäbitz WR. Meta-analysis of the efficacy of granulocyte-colony
stimulating factor in animal models of focal cerebral ischemia. Stroke.
2008;39:1855–1861. doi: 10.1161/STROKEAHA.107.506816.
9. Sevimli S, Diederich K, Strecker JK, Schilling M, Klocke R, Nikol S, et
al. Endogenous brain protection by granulocyte-colony stimulating factor after ischemic stroke. Exp Neurol. 2009;217:328–335. doi: 10.1016/j.
expneurol.2009.03.018.
10. Brenneman M, Sharma S, Harting M, Strong R, Cox CS Jr, Aronowski
J, et al. Autologous bone marrow mononuclear cells enhance recovery
after acute ischemic stroke in young and middle-aged rats. J Cereb Blood
Flow Metab. 2010;30:140–149. doi: 10.1038/jcbfm.2009.198.
11. He XY, Chen ZZ, Cai YQ, Xu G, Shang JH, Kou SB, et al. Expression
of cytokines in rat brain with focal cerebral ischemia after grafting with
bone marrow stromal cells and endothelial progenitor cells. Cytotherapy.
2011;13:46–53. doi: 10.3109/14653249.2010.510505.
12. Chen J, Zhang ZG, Li Y, Wang L, Xu YX, Gautam SC, et al. Intravenous
administration of human bone marrow stromal cells induces angiogenesis in the ischemic boundary zone after stroke in rats. Circ Res.
2003;92:692–699. doi: 10.1161/01.RES.0000063425.51108.8D.
13. Schneider A, Krüger C, Steigleder T, Weber D, Pitzer C, Laage R,
et al. The hematopoietic factor G-CSF is a neuronal ligand that counteracts programmed cell death and drives neurogenesis. J Clin Invest.
2005;115:2083–2098. doi: 10.1172/JCI23559.
14. Ringelstein EB, Thijs V, Norrving B, Chamorro A, Aichner F, Grond
M, et al; AXIS 2 Investigators. Granulocyte colony-stimulating factor in patients with acute ischemic stroke: results of the AX200 for
Ischemic Stroke trial. Stroke. 2013;44:2681–2687. doi: 10.1161/
STROKEAHA.113.001531.
15. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving
bioscience research reporting: the ARRIVE guidelines for reporting
animal research. PLoS Biol. 2010;8:e1000412. doi: 10.1371/journal.
pbio.1000412.
16.Schilling M, Strecker JK, Schäbitz WR, Ringelstein EB, Kiefer R.
Effects of monocyte chemoattractant protein 1 on blood-borne cell
recruitment after transient focal cerebral ischemia in mice. Neuroscience.
2009;161:806–812. doi: 10.1016/j.neuroscience.2009.04.025.
17.Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y.
‘Green mice’ as a source of ubiquitous green cells. FEBS Lett.
1997;407:313–319.
18.Franco EC, Cardoso MM, Gouvêia A, Pereira A, Gomes-Leal W.
Modulation of microglial activation enhances neuroprotection and
functional recovery derived from bone marrow mononuclear cell transplantation after cortical ischemia. Neurosci Res. 2012;73:122–132. doi:
10.1016/j.neures.2012.03.006.
19. Leach MJ, Swan JH, Eisenthal D, Dopson M, Nobbs M. BW619C89, a
glutamate release inhibitor, protects against focal cerebral ischemic damage. Stroke. 1993;24:1063–1067.
20. Henninger N, Bratane BT, Bastan B, Bouley J, Fisher M. Normobaric
hyperoxia and delayed tPA treatment in a rat embolic stroke model.
J Cereb Blood Flow Metab. 2009;29:119–129. doi: 10.1038/
jcbfm.2008.104.
21. Lee SR, Kim HY, Rogowska J, Zhao BQ, Bhide P, Parent JM, et al.
Involvement of matrix metalloproteinase in neuroblast cell migration
from the subventricular zone after stroke. J Neurosci. 2006;26:3491–
3495. doi: 10.1523/JNEUROSCI.4085-05.2006.
22. Ralay Ranaivo H, Hodge JN, Choi N, Wainwright MS. Albumin induces
upregulation of matrix metalloproteinase-9 in astrocytes via MAPK and
reactive oxygen species-dependent pathways. J Neuroinflammation.
2012;9:68. doi: 10.1186/1742-2094-9-68.
Strecker et al Growth Factor Combined Bone Marrow Cell Therapy 11
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
23. Hayakawa K, Pham LD, Katusic ZS, Arai K, Lo EH. Astrocytic highmobility group box 1 promotes endothelial progenitor cell-mediated
neurovascular remodeling during stroke recovery. Proc Natl Acad Sci U
S A. 2012;109:7505–7510. doi: 10.1073/pnas.1121146109.
24. Orihuela R, McPherson CA, Harry GJ. Microglial M1/M2 polarization
and metabolic states [published online ahead of print March 20, 2015]. Br J
Pharmacol. http://onlinelibrary.wiley.com/doi/10.1111/bph.13139/abstract;
jsessionid=2DC6D353517DE4A0B81D532BD40C8E02.f02t01.
Accessed May 8, 2015. doi: 10.1111/bph.13139.
25. Toth ZE, Leker RR, Shahar T, Pastorino S, Szalayova I, Asemenew B, et
al. The combination of granulocyte colony-stimulating factor and stem
cell factor significantly increases the number of bone marrow-derived
endothelial cells in brains of mice following cerebral ischemia. Blood.
2008;111:5544–5552. doi: 10.1182/blood-2007-10-119073.
26. Hermann DM, Chopp M. Promoting brain remodelling and plasticity for stroke recovery: therapeutic promise and potential pitfalls of
clinical translation. Lancet Neurol. 2012;11:369–380. doi: 10.1016/
S1474-4422(12)70039-X.
27.Honmou O, Onodera R, Sasaki M, Waxman SG, Kocsis JD.
Mesenchymal stem cells: therapeutic outlook for stroke. Trends Mol
Med. 2012;18:292–297. doi: 10.1016/j.molmed.2012.02.003.
28. Baker AH, Sica V, Work LM, Williams-Ignarro S, de Nigris F, Lerman
LO, et al. Brain protection using autologous bone marrow cell, metalloproteinase inhibitors, and metabolic treatment in cerebral ischemia. Proc Natl Acad Sci U S A. 2007;104:3597–3602. doi: 10.1073/
pnas.0611112104.
29. Chen J, Ye X, Yan T, Zhang C, Yang XP, Cui X, et al. Adverse effects
of bone marrow stromal cell treatment of stroke in diabetic rats. Stroke.
2011;42:3551–3558. doi: 10.1161/STROKEAHA.111.627174.
30.Gliem M, Mausberg AK, Lee JI, Simiantonakis I, van Rooijen N,
Hartung HP, et al. Macrophages prevent hemorrhagic infarct transformation in murine stroke models. Ann Neurol. 2012;71:743–752. doi:
10.1002/ana.23529.
31. Qiu J, Xu J, Zheng Y, Wei Y, Zhu X, Lo EH, et al. High-mobility group
box 1 promotes metalloproteinase-9 upregulation through Toll-like
receptor 4 after cerebral ischemia. Stroke. 2010;41:2077–2082. doi:
10.1161/STROKEAHA.110.590463.
32. Shigemori Y, Katayama Y, Mori T, Maeda T, Kawamata T. Matrix metalloproteinase-9 is associated with blood-brain barrier opening and brain
edema formation after cortical contusion in rats. Acta Neurochir Suppl.
2006;96:130–133.
33. Tsubokawa T, Solaroglu I, Yatsushige H, Cahill J, Yata K, Zhang JH.
Cathepsin and calpain inhibitor E64d attenuates matrix metalloproteinase-9 activity after focal cerebral ischemia in rats. Stroke. 2006;37:1888–
1894. doi: 10.1161/01.STR.0000227259.15506.24.
34. Hayakawa K, Pham LD, Katusic ZS, Arai K, Lo EH. Astrocytic highmobility group box 1 promotes endothelial progenitor cell-mediated
neurovascular remodeling during stroke recovery. Proc Natl Acad Sci U
S A. 2012;109:7505–7510. doi: 10.1073/pnas.1121146109.
35. Lei C, Lin S, Zhang C, Tao W, Dong W, Hao Z, et al. Activation
of cerebral recovery by matrix metalloproteinase-9 after intracerebral hemorrhage. Neuroscience. 2013;230:86–93. doi: 10.1016/j.
neuroscience.2012.11.008.
36.dela Peña IC, Yoo A, Tajiri N, Acosta SA, Ji X, Kaneko Y, et al.
Granulocyte colony-stimulating factor attenuates delayed tPA-induced
hemorrhagic transformation in ischemic stroke rats by enhancing angiogenesis and vasculogenesis. J Cereb Blood Flow Metab. 2015;35:338–
346. doi: 10.1038/jcbfm.2014.208.
37. Guo Y, Zhang H, Yang J, Liu S, Bing L, Gao J, et al. Granulocyte colonystimulating factor improves alternative activation of microglia under
microenvironment of spinal cord injury. Neuroscience. 2013;238:1–10.
doi: 10.1016/j.neuroscience.2013.01.047.
38. Hu X, Li P, Guo Y, Wang H, Leak RK, Chen S, et al. Microglia/macrophage polarization dynamics reveal novel mechanism of injury expansion after focal cerebral ischemia. Stroke. 2012;43:3063–3070. doi:
10.1161/STROKEAHA.112.659656.
39. Strecker JK, Minnerup J, Schütte-Nütgen K, Gess B, Schäbitz WR,
Schilling M. Monocyte chemoattractant protein-1-deficiency results in
altered blood-brain barrier breakdown after experimental stroke. Stroke.
2013;44:2536–2544. doi: 10.1161/STROKEAHA.111.000528.
40.Stamatovic SM, Keep RF, Wang MM, Jankovic I, Andjelkovic AV.
Caveolae-mediated internalization of occludin and claudin-5 during
CCL2-induced tight junction remodeling in brain endothelial cells.
J Biol Chem. 2009;284:19053–19066. doi: 10.1074/jbc.M109.000521.
41. Pösel C, Scheibe J, Kranz A, Bothe V, Quente E, Fröhlich W, et al. Bone
marrow cell transplantation time-dependently abolishes efficacy of
granulocyte colony-stimulating factor after stroke in hypertensive rats.
Stroke. 2014;45:2431–2437. doi: 10.1161/STROKEAHA.113.004460.
42. Cox GN, Chlipala EA, Smith DJ, Carlson SJ, Bell SJ, Doherty DH.
Hematopoietic properties of granulocyte colony-stimulating factor/
immunoglobulin (G-CSF/IgG-Fc) fusion proteins in normal and neutropenic rodents. PLoS One. 2014;9:e91990. doi: 10.1371/journal.
pone.0091990.
43.Price TH, Chatta GS, Dale DC. Effect of recombinant granulocyte
colony-stimulating factor on neutrophil kinetics in normal young and
elderly humans. Blood. 1996;88:335–340.
44. Taguchi A, Wen Z, Myojin K, Yoshihara T, Nakagomi T, Nakayama D, et
al. Granulocyte colony-stimulating factor has a negative effect on stroke
outcome in a murine model. Eur J Neurosci. 2007;26:126–133. doi:
10.1111/j.1460-9568.2007.05640.x.
Combining Growth Factor and Bone Marrow Cell Therapy Induces Bleeding and Alters
Immune Response After Stroke in Mice
Jan-Kolja Strecker, Joanna Olk, Maike Hoppen, Burkhard Gess, Kai Diederich, Antje Schmidt,
Wolf-Rüdiger Schäbitz, Matthias Schilling and Jens Minnerup
Downloaded from http://stroke.ahajournals.org/ by guest on June 17, 2017
Stroke. published online February 2, 2016;
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2016 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/early/2016/02/02/STROKEAHA.115.011230
Data Supplement (unedited) at:
http://stroke.ahajournals.org/content/suppl/2016/02/02/STROKEAHA.115.011230.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
SUPPLEMENTAL MATERIAL.
Combining growth factor and bone marrow cell therapy induces bleeding and alters
immune response after stroke in mice
Supplementary figures
Supplementary Fig. I: Western blot analysis of CD31, albumin, claudin-5 and occludin
protein amount within the ipsilateral hemisphere of placebo, G-CSF, BMC and G-CSF +
BMC-treated animals. Whole protein lysates (12 µg per lane) were applied to each lane. Band
intensities were measured using NIH ImageJ. Each bands’ intensity was additionally corrected
using the individual actin-band density. All gels were run at the same time under equal
conditions except for the third occludin gel (far right gel, post 7, G-CSF, BMC, G-CSF +
BMC) which had to be redone.
Supplementary Fig. II: No correlation between astrocyte activation and bleeding severity
could be observed. Image showing bleeding severity on the x-axis and GFAP-signal intensity
obtained by densitometric analysis on the y-axis. (7 days post-MCAO, each group n=8, only
PFA-perfused animals).
Supplementary Fig III: Activation and immigration of innate inflammatory cells.
(A) Quantification of F4/80+-cells, indicating macrophages and microglia, showed no differences
between treatments. However, all groups showed a marked increase of F4/80+-cells on day 7
compared to day 1. Analysis of F4/80+-cell inhabited areas revealed no differences between the
investigated groups (ANOVA p=0.054, n=7-10, Fig IIIa). (B) Quantification of 7/4+-neutrophil
granulocytes showed no differences in cell density, except of a trend towards elevated cell counts
after G-CSF+BMC-treatment on day 1 (p=0.09). Analysis of 7/4+-inhabited areas showed a
reduced area in G-CSF-treated mice on day 1 (ANOVA p=0.0069, n=7-10, Fig.IIIb,
mean±SEM). ***p<0.01; Bonferroni post-hoc test.
Table I: primary antibodies for immunohistochemistry and western-blotting
Dilution,
Antibody
Host species
Manufacturer
Application
7/4*
Rat
1:200 (IHC)
Serotec, Germany
Albumin
Rabbit
1:1000 (WB)
Serotec, Germany
CD16/32*
Rat
1:200 (IHC)
Pharmingen, Germany
CD31*
Rabbit
1:100 (IHC + WB)
Abcam, UK
Claudin-5
Rabbit
1:100 (WB)
Abcam, UK
F4/80*
Rat
1:500 (IHC)
Serotec, Germany
GFAP
Mouse
1:200 (IHC)
Sigma-Aldrich, USA
Laminin
Rabbit
1:100 (IHC)
Sigma-Aldrich, USA
MMP-9
Rabbit
1:100 (IHC)
Abcam, UK
NeuN
Mouse
1:300 (IHC)
Chemicon Int., USA
Occludin
Rabbit
1:100 (WB)
Invitrogen, USA
*denotes applied signal amplification.
Table II: body weight, body temperature and cerebral blood flow during surgery
Weight
[g]
Tpre
[°C]
TMCAO
[°C]
Tpost
[°C]
CBFpre
[%]
CBFMCAO
[%]
CBFpost
[%]
Placebo
27.56±0.80
36.49±0.55
36.64±0.49
36.64±0.49
100±12.33
12.29±2.21
83.23±6.00
G-CSF
27.27±0.79
36.43±0.36
36.53±0.39
36.61±0.35
100±12.39
12.45±2.28
82.35±6.04
BMC
26.66±1.76
36.53±0.34
36.60±0.30
36.63±0.27
100±15.97
12.34±4.68
78.37±10.37
G-CSF+BMC
27.21±0.72
36.49±0.36
36.56±0.28
36.57±0.29
100±13.24
11.23±2.25
83.67±5.18
Body weight of mice prior to occlusion, rectal temperature (T) and cerebral blood flow (CBF)
prior (Tpre, CBFpre), during (TMCAO, CBFMCAO) and after (Tpost, CBFpost) occlusion of the MCA.