The Vasculome of the Mouse Brain
Shuzhen Guo1*, Yiming Zhou1,2, Changhong Xing1, Josephine Lok1,3, Angel T. Som1, MingMing Ning1,4,
Xunming Ji5, Eng H. Lo1,4*
1Neuroprotection Research Laboratory, Departments of Radiology and Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts,
United States of America, 2Broad Institute, Massachusetts Institute of Technology and Harvard Medical School, Boston, Massachusetts, United States of America,
3Department of Pediatrics, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, United States of America, 4Clinical Proteomics Research
Center, Department of Neurology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, United States of America, 5Cerebrovascular Research
Center, XuanWu Hospital, Capital Medical University, Beijing, Peoples Republic of China
The blood vessel is no longer viewed as passive plumbing for the brain. Increasingly, experimental and clinical findings
suggest that cerebral endothelium may possess endocrine and paracrine properties – actively releasing signals into and
receiving signals from the neuronal parenchyma. Hence, metabolically perturbed microvessels may contribute to central
nervous system (CNS) injury and disease. Furthermore, cerebral endothelium can serve as sensors and integrators of CNS
dysfunction, releasing measurable biomarkers into the circulating bloodstream. Here, we define and analyze the concept of
a brain vasculome, i.e. a database of gene expression patterns in cerebral endothelium that can be linked to other databases
and systems of CNS mediators and markers. Endothelial cells were purified from mouse brain, heart and kidney glomeruli.
Total RNA were extracted and profiled on Affymetrix mouse 430 2.0 micro-arrays. Gene expression analysis confirmed that
these brain, heart and glomerular preparations were not contaminated by brain cells (astrocytes, oligodendrocytes, or
neurons), cardiomyocytes or kidney tubular cells respectively. Comparison of the vasculome between brain, heart and
kidney glomeruli showed that endothelial gene expression patterns were highly organ-dependent. Analysis of the brain
vasculome demonstrated that many functionally active networks were present, including cell adhesion, transporter activity,
plasma membrane, leukocyte transmigration, Wnt signaling pathways and angiogenesis. Analysis of representative
genome-wide-association-studies showed that genes linked with Alzheimer’s disease, Parkinson’s disease and stroke were
detected in the brain vasculome. Finally, comparison of our mouse brain vasculome with representative plasma protein
databases demonstrated significant overlap, suggesting that the vasculome may be an important source of circulating
signals in blood. Perturbations in cerebral endothelial function may profoundly affect CNS homeostasis. Mapping and
dissecting the vasculome of the brain in health and disease may provide a novel database for investigating disease
mechanisms, assessing therapeutic targets and exploring new biomarkers for the CNS.
Citation: Guo S, Zhou Y, Xing C, Lok J, Som AT, et al. (2012) The Vasculome of the Mouse Brain. PLoS ONE 7(12): e52665. doi:10.1371/journal.pone.0052665
Editor: Cesar V Borlongan, University of South Florida, United States of America
Received September 25, 2012; Accepted November 20, 2012; Published December 2 , 2012
Copyright: ? 2012 Guo et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by studies funded in part by grants from the NIH (RC2-NS69335, R37-NS37074, R01-NS76694, P01-NS55104), the Claflin award
from Massachusetts General Hospital, the Phyliss and Jerome Lyle Rappaport MGH Research Scholar award, the Beijing Natural Science Foundation, and the China
National Basic Research 973 Program. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: email@example.com (SG); Lo@helix.mgh.harvard.edu (EHL)
In recent years, mechanistic investigations into brain function
and disease have shifted away from a purely ‘‘neurocentric’’ focus
into a more integrative perspective that involves all cell types in the
central nervous system [1,2,3,4]. For example, interactions
between neurons and glia are required for normal neurotransmis-
sion as well as remodeling and recovery after brain injury .
Signals from astrocytes and pericytes provide important regulatory
mechanisms for the blood-brain barrier [6,7,8]. And overlapping
mediators underlie crosstalk between neurogenesis, angiogenesis,
and common patterning of neural and blood vessel architectures
Blood vessels in the brain are no longer viewed as passive or
inert plumbing to simply carry blood flow for oxygen and glucose
delivery. Increasingly, it is now recognized that the cerebral
endothelium provide a rich source of signaling and trophic factors
that influence brain function. Cerebral endothelium can produce
growth factors that promote neurogenesis . Cerebral endo-
thelium can release neuroprotective factors such as brain derived
neurotrophic factor (BDNF) and fibroblast growth factor (FGF)
that defend neurons against a wide range of metabolic and toxic
insults [11,12,13,14]. Conversely, ‘‘sick’’ endothelium can con-
tribute to CNS disease. In diabetes, oxidatively stressed cerebral
endothelium produce lower levels of neurotrophic factors that may
lead to increased neuronal susceptibility to stroke and neurode-
generation . Dysfunctional microvessels and disrupted blood-
brain barrier function have been proposed to worsen neuronal
dysfunction in Alzheimer’s disease and amyotrophic lateral
sclerosis [16,17,18]. Hence, understanding the full functional
profile of cerebral endothelium may be extremely important for
investigations into CNS physiology and pathophysiology.
To date, many studies of brain endothelial gene expression have
been performed. The majority of these efforts primarily focus on
the blood-brain barrier [19,20,21], some of them are from the
microvascular fragments . However, a broader approach that
connects the entire vascular blueprint to brain function and disease
has not been attempted. Here, we propose the concept of a brain
PLOS ONE | www.plosone.org1December 2012 | Volume 7 | Issue 12 | e52665
vasculome, i.e. a systematic mapping of transcriptome profiles of
endothelial cells from brain in comparison with those from two
other major organs, the heart and kidney glomeruli, in order to
potentially reveal differential vascular function at a whole genome
level. Our database here is then dissected to assess the hypothesis
that the brain vasculome may contribute to CNS disease in terms
of mechanisms and circulating biomarkers.
Quality Control of Vasculome Gene Expression
Two levels of quality control were assessed. First, quality and
integrity of RNA samples were tested with standard NanoDrop
and Bioanalyzer approaches to ensure sufficient RNA concentra-
tions, 260/280 ratios, 28 s/18 s ratios and RNA integrity number
(RIN) scores. Second, the quality of microarray hybridization was
also assessed by manually checking the distribution of hybridiza-
tion signals, percentage of positive signals, ratio of 39 to 59 end of
housekeeping genes (GAPDH and b-Actin), and applying princi-
pal component analysis to identify potential outliers. All samples
passed these various checkpoints. For the final data sets, RNA
from n=5 mice were pooled per microarray, and n=3 in-
dependent microarrays were used for each group (see Methods).
Next, we asked whether our vasculome was contaminated by
parenchymal cells. For the brain vasculome, we compared our
data with gene markers for different brain cells in public GEO
GSE13379 datasets [23,24], that contain gene expression profiles
for neurons, astrocytes and oligodendrocytes. This analysis
demonstrated that genes known to be representative of neurons,
astrocytes and oligodendrocytes had extremely low expression
levels (signals ,50,) in our brain vasculome, whereas gene markers
of endothelial cells had much higher expression levels than others
(Table 1). These data suggest that our brain vasculome is
endothelial-specific and not contaminated by surrounding paren-
chymal brain cells. Another check of endothelial purity was
performed using RT-PCR to assess the expression of gene markers
in the vasculome from brain, heart and kidney glomeruli (Table
S1). Compared to corresponding whole organ tissue, each
vasculome had higher expression of endothelial markers ($ 3
fold), whereas gene expression levels were enriched for neuron and
astrocyte markers in whole brain samples; myocyte markers in
heart tissue samples; and kidney tubular markers in kidney tissue
samples, respectively (Figure S1). Taken together, these analyzes
suggest that the various organ vasculomes were not overtly
contaminated with parenchymal genes.
Brain Vasculome Specific Genes and Enriched Pathways
Although the microarrays revealed a large amount of data, we
only focused on genes whose maximal expression values across all
microarrays were greater than 200. Based on these criteria, we
identified 3,557 genes expressed in brain endothelial cells. Next,
we asked whether this brain vasculome differed from patterns
found in our comparative heart and kidney glomerular vascu-
lomes. Applying criteria of p,0.01, fold change $ 4, and maximal
expression value across all samples .200, we identified 318 probes
corresponding to 243 genes found to be highly expressed in brain
endothelial cells, 143 probes corresponding to 110 genes highly
expressed in heart endothelial cells, and 114 probes corresponding
to 81 genes highly expressed in kidney glomerular endothelial cells.
A heat-map analysis demonstrated that each vasculome was highly
organ-specific. Gene expression patterns in the brain vasculome
significantly differed from those in heart or kidney glomeruli
In the brain vasculome, as expected, blood-brain barrier genes
were easily detected. These included occludin (OCLN) and
claudin-5 (CLDN5), two major components of tight junctions in
the blood-brain barrier. Another known feature of cerebral
endothelium is the presence of glutamate receptors that influence
barrier function [25,26,27]. In our datasets, both ionotrophic and
metabtrophic glutamate receptors (Gria2, Gria3, Grin2b and
Grm5 for AMPA2, AMPA3, NMDA2B and mGluR5 respectively)
were specifically expressed in the vasculome from mouse brain but
not heart and kidney glomeruli. Overactivation of glutamate
receptors may cause excitotoxicity in neuronal compartments.
Similarly, activation of NMDA or mGluR5 receptors could also
mediate vascular responses caused by hyperhomocysteinemia and
nitrosative stress in brain endothelial cells [28,29]. The glutamate
transporter Slc1a1 (solute carrier family 1, member 1), known as
EAAC1/EAAT3, was also enriched in the brain vasculome. In
neurons, this transporter plays a key role in regulating synaptic
glutamate kinetics. But vascular functions may also exist, since
brain endothelial cells co-cultured with astrocytes displayed
a polarized brain-to-blood transport of glutamate, suggesting that
transporter responses in brain microvessels may participate in the
regulation of potentially excitotoxic amino acid concentrations
. Overall, the brain vasculome demonstrated many mediators
commonly found in neural systems. For example, semaphorins
comprise a family of factors that control neurite extension and
axon guidance [31,32]. Here, we saw that Sema3C and Sema4D
were enriched in the vasculome of brain compared to heart and
kidney. Both Sema3C and Sema4D have been implicated in
angiogenesis regulation [33,34,35,36], so their involvement in
brain vascular homeostasis may also be important. Finally,
another parallel with neural signaling may also be underscored
by the detection of CamKIIa (calcium/calmodulin-stimulated
protein kinase II alpha) [37,38], indicating that common responses
to intracellular calcium fluxes may occur in both neuronal and
non-neuronal systems in the brain.
Using the Fisher’s exact test to probe GO (Gene Ontology) and
KEGG databases, many signaling and regulatory pathways were
found to be enriched in the brain vasculome (Table 2). These
included transporter activity, cell adhesion molecules (CAMs), and
the Wnt signaling pathway. Besides well-known brain endothelial
transporters such as Glut1 (glucose transporter type 1, or slc2a1,
solute carrier family 2, member 1) and P-gp (multidrug resistence
poly-glycoprotein, or Abcb1a, ATP-binding cassette, sub-family B,
member1A), high levels of transferrin (Tf), transferrin receptor
(TfR) and exporter ferroportin (slc40a1) were also detected in the
brain vasculome. For cell adhesion molecules, NrCAM (neuronal
cell adhesion molecule) appeared to be enriched in the brain
vasculome. NrCAM was originally known as a neuron-specific
gene required for axon guidance and organization of neural
circuitry [39,40,41]. However, NrCAM has recently been
discovered in dermal and umbilical venous endothelium as well,
with potential function in angiogenesis regulation and stress
response in endothelial cells [42,43]. The presence of NrCAM in
our initial draft of the brain vasculome but not the heart or
glomerular vasculome, further suggest close interactions and
potential crosstalk between vascular systems and the organ milieu
they inhabit. Similar enrichment in membrane proteins was found
in the neurexin network. We detected the expression of neurexin
and neuroligin in our mouse brain vasculome. In particular,
neurexin-1 showed high level of expression in the brain vasculome
rather than heart and kidney vasculomes. Once again, expression
of neural-related guidance systems in the brain vasculome suggests
crosstalk and signaling functions between blood vessels and brain
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org2 December 2012 | Volume 7 | Issue 12 | e52665
In addition to physiologic pathways that underlie normal
function, pathophysiologic pathways related to inflammation were
also expressed in the brain vasculome. In the context of brain
injury and neurodegeneration, cytokines and chemokines com-
prise a key network for regulating inflammation. In this vasculome
project, 236 probes were screened for 150 cytokines/chemokines.
Overall, low signals were detected for most cytokines/chemokines
(,50). Applying the criteria of signal intensity .200, only 17
probes for 11 cytokines/chemokines were expressed in the normal
mouse brain vasculome - ccl3, ccl9, ccl27, csf1, cxcl12 (SDF1), kitl,
pdgfb, pglyrp1, ptn, socs7 and tgfb2 (Table 3). Compared to heart
and kidney glomeruli, ccl3 (chemokine (C-C motif) ligand 3), ccl27
(chemokine (C-C motif) ligand 27) and pglyrp1 peptidoglycan
recognition protein 1) appeared to be enriched in the brain
vasculome (Table 3). Examination of the existing literature
suggested that these may be relevant hits. Ccl3 is released by
stimulated brain endothelial cells , and it has been reported
that it may be elevated in brain vessels of Alzheimer’s disease
patients . CCL27 is well known for mediating skin in-
flammation but has also been detected in the brain . Unlike
other chemokines, CCL27 has both secreted and nuclear targeting
forms that directly modulate transcription of many response genes,
thus any involvement of this factor in the brain vasculome could
potentially act as a potent amplifier of inflammation [48,49].
PGLYRPs (or PGRPs, peptidoglycan recognition proteins) have
four isoforms, PGLYRP1-4, that function in antibacterial immu-
Table 1. Differential expression of cell-type specific markers in brain vasculome versus other cell types in brain.
Cell Type SymbolProbe ID Brain vasculome AstrocyteNeuron Oligodendrocyte
astrocyte 2900052N01Rik 1436231_at3.320912.1724 5.5117 5.5370
astrocyteAcsbg11422428_at 3.8742 10.29797.1814 6.9079
astrocyteGfap 1440142_s_at2.964712.4009 7.29917.2095
astrocyte Gjb6 1448397_at3.3301 12.37168.26818.2473
astrocyteTtpa 1427284_a_at3.53219.9466 4.92494.2301
Neuron Crh1457984_at 4.06079.8493 11.797610.5446
Neuron Hs3st21438624_x_at 3.2656 6.0426 10.23908.5537
Neuron Mal21427042_at 3.67879.125811.714710.1006
NeuronNecab11437156_at 2.69247.0119 10.71808.6910
oligodendrocyte Ugt8a1419063_at4.83005.42959.0736 12.6444
endothelialFlt1 1451756_at 10.5391 2.18772.97282.2005
endothelialNos3 1422622_at7.3330 5.66584.47085.0355
endothelialOcln 1448873_at9.9552 2.23252.54222.1702
endothelialPecam11421287_a_at 10.06622.18172.1626 2.1702
endothelialTek1418788_at 11.37762.2534 2.78822.2962
Note: Numbers for log2 signal intensity. Except for brain vasculome, all other data are listed from GSE13379 of GEO (Doyle JP et al. 2008 and Dougherty JD et al. 2010).
Brain vasculome represents mean value of 3 samples, astrocyte represents mean value of 6 samples, neuron represents mean value of 23 samples, oligodendrocyte
represents mean value of 4 samples. Well-established markers for neurons, astrocytes or oligodendrocytes are highly expressed in their corresponding cell types, while
all neuronal,. Astrocytic and oligodendroglial genes have extremely low expression levels (or not detectable) in the brain vasculome. In contrast, endothelial markers are
highly expressed in the vasculome and show low levels in other types of cells.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org3December 2012 | Volume 7 | Issue 12 | e52665
nity and inflammation . PGLYRP1 can bind with the key
stress response proteins such as Hsp70 and S100A4 to trigger
cytotxicity for antibacterial activity [51,52]. Expression pglyrp1 in
the brain have been reported, but its endothelial function is
currently unknown .
Another inflammatory example was found in pathways involved
in leukocyte transendothelial migration (Figure 2A). The brain
vasculome-enriched genes in this pathway included Ncf1 (neutro-
phil cytosolic factor1, or p47 phox), Prkcb (protein kinase C, beta)
and Prkcc (protein kinase C, gamma). Ncf1 is a subunit of
NADPH oxidase, a critical enzyme for ROS production in injured
or diseased vascular systems [54,55]. It was reported that Ncf1
mediated the Abeta42 and RAGE ligation induced ROS pro-
duction and downstream ERK1/2 phosphorylation and cPLA2
(cytosolic phospholipase A2) phosphorylation in cerebral endothe-
lial cells . The PKC family is known to regulate the
phosphorylation and uptake of SLC6 family of neurotransmitter
transporters , also reported to be present in brain endothelial
cells and regulate the blood-brain barrier [58,59,60]. Whether
Prkcb and Prkcc in the brain vasculome contribute to disease
phenomena involved in cerebral ischemia, brain injury and
neurodegeneration remains to be fully elucidated.
Finally, a prominent network that was enriched in the brain
vasculome comprised the Wnt pathway (Figure 2B). Wnt is known
to regulate neuronal stem cells, neurogenesis and neuroplasticity
[61,62,63,64]. But recently, Wnt signaling has been reported to
also participate in the development of CNS vasculature, blood-
brain barrier formation, and the protection of endothelial cells
after injury [65,66,67,68,69]. In our draft of the brain vasculome,
b-cantenin (CTNNB1) was presented in the hub position of the
Wnt protein-protein interaction network, along with brain
endothelial-specific genes Axin2, MAPK10 (mitogen-activated
protein kinase 10) and Lef1 (lymphoid enhancer binding factor
1). Axin2 is a transcriptional target of active Wnt signaling that
also serves to autoregulate and repress the pathway by promoting
b-cantenin degradation ; In the conditional transgenic mice
Figure 1. The vasculome of mouse brain is unique and different from those found in mouse heart and kidney. Heatmap for
visualization of the expression levels of organ-specific endothelial genes across brain, heart and kidney glomeruli. X-axis represents individual
samples and y-axis represents different genes. The expression levels of genes are indexed by color.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org4 December 2012 | Volume 7 | Issue 12 | e52665
Table 2. Enriched pathways detected in the vasculome of mouse brain.
pvaluelog2 odd ratioGO termGO category
1.94E-284.34membrane part Cellular Component
1.09E-274.18membrane Cellular Component
8.93E-274.24intrinsic to membrane Cellular Component
2.21E-254.10 integral to membraneCellular Component
9.93E-095.07cell junction Cellular Component
6.22E-153.47transport Biological Process
7.04E-13 7.41cell-cell signalingBiological Process
5.45E-114.60 ion transportBiological Process
1.14E-072.42 cell communicationBiological Process
1.63E-05 2.22anatomical structure development Biological Process
3.41E-134.40transporter activity Molecular Function
4.81E-12 4.95substrate-specific transmembrane transporter activityMolecular Function
1.71E-11 4.57transmembrane transporter activity Molecular Function
3.99E-104.60 ion transmembrane transporter activityMolecular Function
3.90E-05 2.27 signal transducer activityMolecular Function
pvaluelog2 odd ratioKEGG pathway
2.67E-067.43 Cell adhesion molecules (CAMs)
1.57E-035.15 Leukocyte transendothelial migration
3.50E-0310.95 Alzheimer’s disease
6.24E-033.84Wnt signaling pathway
Note: Analysis based on brain endothelial specific genes in the mouse brain vasculome. These enriched pathways suggest that specific pathways and mechanisms are
selectively enhanced in brain compared to heart and kidney glomerular vasculomes.
Table 3. List of cytokines/chemokines expressed in the vasculome of mouse brain.
log2 signal intensity
Probe IDEntrez ID symboldescriptionbrain heartglomeruli
1419561_at20302Ccl3 chemokine (C-C motif) ligand 3 7.8736 4.96173.6407
1430375_a_at 20301 Ccl27chemokine (C-C motif) ligand 27 8.7402 6.54986.4454
1449184_at 21946Pglyrp1peptidoglycan recognition protein 19.52534.90283.9403
1448823_at20315Cxcl12chemokine (C-X-C motif) ligand 127.67837.64245.0764
1417936_at 20308Ccl9chemokine (C-C motif) ligand 97.95956.87255.3389
1450414_at18591Pdgfbplatelet derived growth factor, B polypeptide8.08187.57717.3091
1460220_a_at 12977Csf1colony stimulating factor 1 8.29819.283010.8936
1426152_a_at 17311Kitlkit ligand8.44088.03397.2815
1439084_at20315Cxcl12chemokine (C-X-C motif) ligand 12 8.44237.79324.9062
1415854_at17311 Kitlkit ligand9.08379.39128.0980
1455402_at192157Socs7suppressor of cytokine signaling 79.56959.26699.7154
1450923_at21808Tgfb2transforming growth factor, beta 29.61537.93577.7813
1448254_at19242Ptn pleiotrophin 10.31946.49709.9998
1417574_at 20315 Cxcl12chemokine (C-X-C motif) ligand 1211.979111.44997.9319
1416211_a_at19242 Ptnpleiotrophin12.37657.9964 11.9745
Note: The first three factors (Ccl3, Ccl27, Pglryp1) are enriched in brain versus heart and kidney glomerular vasculomes.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org5December 2012 | Volume 7 | Issue 12 | e52665
overexpressing b-Catenin, Axin2 is one of the antagonists changed
in the brain . MAPK10 was originally described in neurons
but it was recently reported to also mediate endothelial migration
via eNOS . And Lef1 is the specific transcriptional factor in
the downstream effectors of the Wnt pathway [73,74]. A critical
role of Wnt signaling in cell-cell communication can also be seen
because its central hub b-Catenin also serves in the protein-protein
interaction network for adherens junctions (Figure 2C) for brain
endothelial cells, linking specifically with the brain vasculome
genes of Igf1r (insulin-like growth factor 1 receptor), Tgfbr1
(transforming growth factor, beta receptor 1) and Lef1 in this
Angiogenesis and the Brain Vasculome
In terms of functional networks, angiogenesis should comprise
a central part of any vasculome. Probing the GO database
revealed a dense protein-protein interaction network for angio-
genesis-related genes in the brain vasculome, with hub positions
occupied by b-catenin, Rtn4, HIF-1a, Mapk14, Notch1, Ptk2
(protein tyrosine kinase 2, also called focal adhesion kinase 1) and
Tgfbr2 (Figure 3A). As described in above, b-catenin is highly
expressed in the brain and in the hub positions of other pathways,
connecting angiogenesis with these pathways, including Wnt
pathway and adherens junctions. Also, Rtn4 (also called Nogo) was
highly expressed in the brain. Rtn4 produces 3 isoforms (Nogo-A,
Nogo-B, Nogo-C) that may play overlapping roles in vascular as
well as neuronal systems in the CNS. Nogo-A is a well-
characterized inhibitor of axonal growth and repair , whereas
Nogo-B is already known to be highly expressed in endothelial
cells . Nogo-B regulates vascular homeostasis and remodeling,
in part by controlling endothelial cell migration, macrophage
infiltration, leukocyte transmigration, and overall inflammation
response after tissue ischemia and injury [76,77,78]. Overall,
Nogo-B may be protective since it is lost after injury .
Another angiogeneis gene with high expression levels in the
brain vasculome is Gpx-1 (glutathione peroxidase 1), an in-
tracellular antioxidant enzyme that converts hydrogen peroxide to
water . Various studies with Gpx-1 transgenic mice suggest
that this mediator may be neuroprotective against amyloid toxicity
, Parkinsons-related pathologies , ischemia-reperfusion
, or trauma . These underlying mechanisms of vascular
neuroprotection may broadly include amelioration of cell death,
suppression of astrocyte and microglia activation, preservation of
BBB function, and a reduction of inflammatory infiltration
Figure 2. Protein-protein interaction (PPI) networks in the vasculome of mouse brain. A, PPI network for leukocyte transendothelial
migration. B, PPI network for the WNT signaling pathway. C, PPI network for adherence junctions. The expression levels of genes in the vasculome of
mouse brain are indexed by color.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org6December 2012 | Volume 7 | Issue 12 | e52665
[82,84,85]. Furthermore, Gpx-1 may also contribute to CNS
recovery, in part by interacting with hypoxia inducible factor 1
(HIF-1) and its target genes such as VEGF to regulate the
angiogenesis process for tissue repair . Gpx-1-deficient mice
show decreased recruitment and activation of endothelial pro-
Figure 3. Angiogenesis networks. A, Protein-protein interaction network for angiogenesis in the vasculome of mouse brain (including nearest
neighbors). Circles for genes in angiogenesis and squares for the neighbor genes. The expression levels of genes in the vasculome of mouse brain are
indexed by color. B, Heatmap comparison of expression profiles of genes in the VEGF signaling pathway from the vasculome of mouse brain, heart
and kidney glomeruli. The expression levels of genes are indexed by color.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org7December 2012 | Volume 7 | Issue 12 | e52665
genitor cells after ischemic injury, leading to impaired angiogenesis
and revascularization .
Within the CNS context, VEGF signaling may be especially
important because this mediator may participate in both
angiogenesis as well as neurogenesis [88,89]. Comparison of
VEGF signaling pathways showed that these were highly
conserved across all three vasculomes in brain, heart and kidney
glomeruli (Figure 3B). This may not be surprising since VEGF-
mediated angiogenesis may be commonly required network
regardless of organ systems. However, it is worth noting that
there were two VEGF signaling mediators that appeared to be
specifically expressed in the brain vasculome - Prkcb and Prkcc.
These two signals were also identified in the leukocyte transen-
dothelial migration network of the brain vasculome (see previous
section). Thus, it is possible that particular brain vasculome-
specific components may critically influence how the CNS
responds to injury and disease. Angiogenesis is a physiological
process involving the growth of new blood vessels. This
phenomenon is vital not only for organ development but also for
tissue repair and wound healing. Insofar as the brain vasculome
may be a critical component of CNS plasticity and remodeling,
these angiogenesis networks may represent a rich database to
probe for potential mechanisms and targets for neurorecovery
after stroke, brain injury or neurodegeneration.
Correlation between Brain Vasculome and CNS Disease
Genome-wide association studies (GWAS) provide valuable
information for identifying molecular risk factors and mechanisms
for many diseases . For CNS disorders, however, GWAS may
be complicated by the fact that disease processes operate not only
in neuronal cells but also in other cells from glial and vascular
compartments. In the context of stroke and neurodegeneration,
pathophysiologic mechanisms are increasingly known to take place
in the neurovascular system [1,2,3,4,16]. So we next asked
whether GWAS-defined genes for major CNS diseases could be
found in our initial draft of the mouse brain vasculome. Genes
implicated in Alzheimer’s disease (AD), Parkinson’s disease (PD)
and stroke were compiled from the Database of Genotypes and
Phenotypes (dbGaP) at NCBI. A substantial portion of these
disease genes was expressed in the brain vasculome –41 AD genes,
53 PD genes and 133 stroke genes (Table 4; complete gene list is
provided in Table S2). Representative genes are briefly surveyed
CD2-associated protein (CD2AP), as
an adapter molecule, is mainly studied in kidney glomeruli. It is
highly expressed by podocytes and binds with nephrin to maintain
glomerular slit diaphragm function. Mice lacking CD2AP exhibit
a congenital nephritic syndrome at early age of 3 weeks . In
other tissues, including brain and heart, CD2AP is located in
endothelial or epithelial cells, but the functions of CD2AP in brain
and heart are still unknown .
PAKs (p21-activated kinases), comprising two subfamilies and at
least 6 members (PAK1-6), are serine/threonine protein kinases
that act downstream of Rho family GTPases Cdc42 and Rac.
PAK2 (also known as gamma-PAK), bind with actin and become
activated in response to a variety of stresses, and these responses
have been implicated in regulation of cytoskeletal structure,
apoptosis angiogenesis, vascular integrity and endothelial cell
contraction [93,94,95,96]. PAK2 deletion leads to cerebral
hemorrhage in redhead zebrafish and this defect is rescued by
endothelial-specific expression of PAK2, demonstrating the
important role of PAK2 in brain vessels .PAK2 may also
mediate the VEGF-induced increase of vascular permeability .
In the brain, PAK1-3 was reported to regulate the morphology of
embryonic cortical neurons, whereas inhibiting Pak activity
causing misorientation and branching process of neurons, with
increased numbers of nodes, terminals and length of processes
. In this regard, PAK may represent yet another example of
overlaps between neural and vascular signals in AD pathophys-
Ataxin-1 (ATXN1), is a causative gene for spinocerebellar
ataxia type 1 (SCA1), with mutation of expanded CAG tri-
nucleotide repeats encoding a polyglutamine tract (polyQ) in the
gene . ATXN1 is expressed in both brain and non-neuronal
tissues, and may participate in calcium homeostasis, glutamate
signaling/excitotoxicity, and Notch signaling pathways [100,101]
through the regulation of transcriptional repression and protein
degradation [102,103,104]. In primary neuron cultures, knock-
down of ATXN1 significantly increases Ab40 and Ab42, with
increased APP cleavage by b-secretase; while overexpression of
ATXN1 decreases Ab levels . The role of ATXN1 in
endothelial cells is not presently well understood, so whether
vascular responses in ATXN1 may also affect Ab homeostasis
Angiomotin (AMOT), first identified as a binding protein to
angiostatin, is a transmembrane protein associated with actin.
AMOT controls cell migration and motility, cell polarity, tight
junction formation and angiogenesis, and also plays critical roles in
the tumor suppressor Hippo pathway [106,107,108]. AMOT is
expressed mostly in endothelial cells and in some epithelial cells,
with two protein isoforms, p80 and p130 . The ratio of the
two isoforms may regulate the switch between migration and
stabilization of endothelial cells [110,111]. Most of Amot knockout
mice die between embryonic day E11 and E11.5 and exhibit
severe vascular insufficiency in the intersomitic region as well as
dilated vessels in the brain . Whether AMOT contributes to
dysfunctional remodeling of brain vessels in the face of progressive
Table 4. Expression of disease-related genes in the vasculome of mouse brain.
Alzheimer’s disease Parkinson’s disease stroke
GWAS genes in dbGAP274364 920
GWAS genes in mouse HomolGene 198264643
GWAS genes in mouse M430 2.0 178239 596
GWAS genes in brain vasculome 4153 133
p value 0.0170.0160.00019
odds ratio1.50 1.421.45
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org8 December 2012 | Volume 7 | Issue 12 | e52665
Alzheimer’s neurodegeneration is a hypothesis that remains to be
STK24 (sterile20-like kinase 24, also known as Mst3 (Mamma-
lian sterile 20-like kinase-3)) mediates the axon-promoting effects
of trophic factors, and may help regulate axon regeneration in
damaged neurons [112,113]. Stk24 has also been reported to
regulate cell morphology, migration and apoptosis [114,115,116].
In the context of AD or PD, Stk24 may contribute to neuronal
Tau phosphorylation, neurite outgrowth and synaptic plasticity
modulation by binding with LRRK2 (leucine-rich repeat kinase 2),
the most common genetic cause of PD . Recently, Stk24 has
also been associated with vascular functions. Stk24, when linked
with striatin into a large signaling complex, acts as an essential
downstream effector of CCM signaling during cardiovascular
development. CCM3 is the disease gene for cerebral cavernous
malformations (CCMs), a condition that leads to characteristic
changes in brain capillary architecture resulting in neurologic
deficits, seizures, or stroke . How these vascular effects
interact with neuronal phenomenon remains unclear.
Regulator of G protein signaling
(RGS) proteins form a large family of GTPase-activating proteins
(GAP activity) for heterotrimeric G protein alpha subunits that
negatively regulate G-protein-coupled receptor signaling. RGS2
selectively accelerates the GTPase activity of Gq/11a and Gi/oa
subunits. RGS2 deficiency in mice leads to hypertension and
cardiac hypertrophy . Endothelium-specific deletion of
RGS2 caused endothelial dysfunction with impaired EDHF-
dependent vasodilatation . In the brain, both clinical and
animal models showed that lower RGS2 expression is associated
with anxiety disorders [121,122]. In neurons, RGS2 was reported
to regulate ionic channel function and synaptic plasticity in the
hippocampus [123,124,125,126]. But how RGS2 in brain vessels
interacts with neuronal sequelae in PD remains unknown.
HnRNP U (heterogeneous ribonuclear protein U, also scaffold
attachment facrot A, SFA) is a multi-functional nuclear matrix
protein that has been implicated in multiple inflammatory
pathways [127,128]. Proinflammatory toll-like receptor signaling
can stimulate the translocation of hnRNP U from nuclear to
cytoplasmic compartments, which then allows it to bind and
stabilize mRNA of various proinflammatory cytokines . How
these inflammatory actions affect the brain vasculome in PD
remains to be determined.
RNF114 (RING finger protein 114, also as ZNF313, zinc finger
protein 313), first identified and reported in 2003, is an ubiquitin
binding protein and disease susceptibility gene for psoriasis, an
immune-mediated skin disorder . RNF114 is reported to
regulate a positive feedback loop that enhances pathogenic double-
stranded RNA induced production of type 1 interferon by
modulating RIG-1/MDA5 signaling .
ITSN2 (intersectins 2), a Cdc42 guanine nucleotide exchange
factor (GEF), is a multidomain adaptor/scaffold protein involved
in clatherin- and caveolin-mediated endocytosis, exocytosis, actin
cytoskeleton rearrangement and signal transduction . Several
isoforms of ITSN protein can be assembled from alternative
splicing, including a brain specific isoform . A role of ITSN-
2L in regulating endocytosis within endothelial cells has been
PAK1 belongs to the family of p21 activated kinases. In
neurons, PAK1 is known to regulate migration [135,136], spine
morphogenesis and synapse formation , neuronal polarity
, and hippocampal long-term potentiation . Besides
being a PD GWAS gene, PAK1 may also modulate or bind with
other disease proteins, including Fragile X mental retardation 1
(FMR1) for Fragile X syndrome (FXS), the most commonly
inherited form of mental retardation and autism ; Disrupted-
in-Schizophrenia 1 (DISC1) for schizophrenia ; ALS2/Alsin
for amyotrophic lateral sclerosis (ALS) , and Down syndrome
cell adhesion molecule (DSCAM) . In endothelial cells,
PAK1 may regulate barrier function in different organs [144,145],
and the migration of endothelial cells during angiogenesis .
In the context of inflammation, Pak1 is known to assist the
invasion of Escherichia coli through human brain microvascular
endothelial cells [147,148].
Ubiquitin C-terminal hydrolase 5 (UCHL5), is one of the
proteasome 19S regulatory-particle-associated deubiquitinase.
Inhibiting the activity of UCHL5 leads to cell apoptosis by
altering Bax/Bcl-2 ratios and activating caspase-9 and caspase-3
. Through Rpn13, UCHL5 is recruited in the 26 s
proteasome complex during the deubiquitination process. it is
reported to regulate the degradation of iNOS and IkappaB-alpha
and participated in the process of inflammation and host defense
regulation [150,151]. In the nucleus, UCHL5 is also associated
with human Ino80 chromatin-remodeling complex and kept in
inactive state, and then activated by transient interaction of the
Ino80 complex with proteasome, suggesting that it may cooperate
to regulate transcription or DNA repair . UCHL5 interacts
with Smads and potentially reverse Smurf-mediated degradation;
it also stabilizes type 1 TGF-beta receptor and regulates TFG-beta
signaling . It is possible that these inflammatory phenome-
nons may also be important in the brain vasculome.
TGF-beta signaling is necessary for the development of blood
vessels in many organs including brain and heart. Selective
deletion of TGF-beta in endothelial cells, but not in neural cells,
led to brain-specific vascular pathologies, including intracerebral
hemorrhage . Inactivation of TGF-beta type II receptor
(Tgfbr2) in endothelial cells in mouse embryo resulted in deficient
ventricular separation and haemorrhage from cerebral blood
vessels . At the same time, TGF-beta signaling is also
important for neural cells. In the midbrain, Tgfbr2 ablation results
in ectopic expression of Wnt1/b-Catenin and FGF8, activation of
Wnt target genes for regulating neural stem cell expansion .
These overlapping actions in neuronal and vascular compartments
may allow TGF-beta to play a key role as a PD GWAS gene.
Transcription factor 6 (ATF6) is one of the effectors of
endoplasmic reticulum stress . Both oxidized LDL and
phospholipolyzed LDL can induce endoplasmic reticulum stress in
endothelial cells with ATF6 activation, and this process has been
implicated in the initiation of vascular inflammation with pro-
gression of atherosclerosis. Via endoplasmic reticulum stress,
ATF6 may also regulate responses in angiogenesis and expression
of tight junction proteins [158,159].
BRM (Brahma), ATPase subunit in the chromatin-
remodeling complex SWI/SNF, has important role in gene
regulation, inflammation response, tumorigenesis and embryo
development and differentiation. BRM is preferentially expressed
in brain, liver, fibromuscular stroma and endothelial cell . It
is reported that BRM and Brahma/SWI2-related gene 1(Brg1)
regulate HIF-1 induced gene expression after hypoxia . Two
single nucleotide polymorphisms (SNP) sites were found to be
associated with schizophrenia in a Japanese population. A risk
allele of a missense polymorphism (rs2296212) induced a lower
nuclear localization efficiency of BRM, and risk alleles of intronic
polymorphisms (rs3793490) were associated with low SMARCA2
expression levels in the postmortem prefrontal cortex .
MBD2 belongs to family of methyl-CpG binding domain
(MBD)-containing factors, and mediate epigenetic effects through
gene expression regulation. It has been reported that MBD2 was
induced in hippocampus within few hours post-ischemia and
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org9December 2012 | Volume 7 | Issue 12 | e52665
maintained at high levels for several days . Furthermore,
Mbd2 deficient mice were protected against hind-limb ischemia
evidenced by improved perfusion recovery and increased capillary
and arteriole formation . In vitro experiment also confirmed
that knockdown of MBD2 significantly enhanced angiogenesis and
provided protection against H2O2-induced apoptosis .
NF-kb inhibitor kappa B alpha (NFKBIA, also known as IkB-
alpha) is the major regulator of NF-kB activation. Since NF-kB is
a central factor in the vast network of inflammation pathways, this
stroke GWAS gene is likely to contribute to multiple vascular
responses in the brain. In damaged endothelium, dynamic
response and inhibitory feedback loops exist between the rapid
increase of IkB-alpha and the original NF-kB signal . Links
to oxidative stress and vasoregulation may also be important as
eNOS-derived nitric oxide can be an endogenous inhibitor of NF-
kB activity through IkB-alpha regulation .
WNK1, is a member of novel serine/threonine kinase family,
With-No-K(lysine), with pleiotropic actions. Intronic deletions in
WNK1 gene cause Gordon’s Syndrome, an autosomal dominant,
hypertensive and hyperkalemic disorder . WNK1 poly-
morphisms have also been associated with common essential
hypertension . Mechanistically, the WNK1 to ste20/SPAK/
OSR1 signaling cascade regulates cation-chloride cotransporters
(NKCC1-2), which may be vital for sodium homeostasis regula-
tion, blood pressure response and vascular contractions [168,169].
Endothelial-specific expression of WNK1 is essential for angio-
genesis and heart development in mice, as WNK1 deficiency leads
to cardiovascular developmental defects with smaller chambers
and reduced myocardial trabeculation, together with defective
angiogenesis in both arteries and veins . Overlap with neural
responses may also be important. WNK1 mutations have been
identified as the cause of hereditary sensory and autonomic
neuropathy type II, an early-onset autosomal disease of peripheral
sensory nerves. WNK1 can interact with LINGO-1 (a component
of tripartite receptor complexes) to regulate nogo-induced in-
hibition of neurite extension, through activation of RhoA .
ADD1 is one of three adducin proteins. ADD1 is a well-known
hypertension risk gene. Altered adducin function might cause
hypertension through enhanced constitutive tubular sodium
reabsorption . Polymorphisms of the ADD1 gene are
associated with many physiological responses in hypertensive
individuals as well as healthy subjects. For example, the Trp460
ADD1 allele is associated with higher systolic and diastolic blood
pressure , with increased incidence of peripheral arterial
disease (PAD) and coronary heart disease (CHD) , increased
carotid artery intima-media thickness (IMT) [175,176], increased
risk of stroke , and reduced acetylcholine-stimulated forearm
blood flow (FBF) response via an impaired endothelium-dependent
vasodilation . Again, the study of variants in risk genes
suggested that there are physiological interaction between ADD1
and WNK1-NEDD4L pathways to regulate the renal sodium
handling, blood pressure and antihypertensive responses to drugs
. Furthermore, the overexpression of rat wild type ADD1 in
endothelial cells Increased tube formation in vitro and enhanced
capillary formation in Matrigel implants in vivo, suggesting ADD1
could regulate angiogenesis process .
Among all of these disease genes, there are some with brain
vasculome specificity compared to heart and kidney glomeruli. For
example, the AD disease gene Pllp (plasma membrane proteolipid,
also known as transmembrane 4 superfamily member 11 or
plasmolipin), is a myelin structure protein and mainly expressed in
brain oligodendrocytes and kidney tubular epithelial cells . It
was reported that pllp could form voltage-dependent and K(+)-
selective ion channels in the membrane, or act as entry receptor
for a kind endogenous retrovirous . The expression of Pllp
was signifiacantly reduced in the temporal cortex of patients with
schizophrenia and patients with major depressive disorder,
suggesting its role in the mental disorders [182,183]. The PD
disease gene Foxf1 (forkhead box F1, also known as HFH-8 or
Freac-1), is a developmentally important transcriptional factor.
The deficiency of Foxf1 could cause severe abnormalities in the
development of many organs including lung, liver and gallbladder,
with reduced expression of intergrin-beta3 . As the target of
hedgehog, foxf1 and its target gene Bmp4 mediate the induction of
vasculogenesis  or link hedgehog signaling with Wnt
signaling, to regulate the development of organs . The
expression of foxf1 in endothelial cells has been reported, and may
regulate the inflammation response . For stroke, Apcdd1,
Atp2b2, Axin2, ITIH-5 and Slc1a1 are specifically expressed in
brain vasculome. As previously discussed, Slc1a1 and Axin2 may
be involved in cerebral glutamate handling and vascular de-
velopment and patterning respectively. Apcdd1(adenomatosis
polyposis coli down-regulated 1), a membrane-bound glycopro-
tein, is the target gene of Wnt/b-Catenin signaling pathway
[188,189], also a novel inhibitor to Wnt signaling in a cell-
autonomous manner and acts upstream of b-Catenin .
Apcdd1 has an essential role in hair growth , or regulate
astro-gliogenesis in the brain . ITIH 5 is one of heavy chain
subunits of Inter-alpha-trypsin inhibitors (ITIs), a family of serine
protease inhibitors. ITIHs stabilize the extracellular matrix (ECM)
by interacting with hyaluronic acid, which is a major ECM
component . So far, ITIH molecules have been reported to
play a particulary important role in inflammation and carcino-
genesis . ITIH5 may also be a regulator of human
metabolism, as the expression of ITIH5 in adipose tissue was
increased in obesity, and associated with measures of body size and
metabolism . Hypermethylation in the upstream region of
the promoter-associated CpG island of ITIH5, has been detected
in breast cancer, and associated with adverse clinical outcome,
suggesting ITIH5 as a potential prognostic marker . Atp2b2
is also known as PMCA2 for plasma membrane calcium-
transporting ATPase 2, encoding a plasma membrane Ca2+-
ATPase type 2 pump, which extrudes calcium from the cytosol
into the extracellular space. The mutation of Atp2b2 may cause
deafness and imbalance in mice probably by affecting sensory
transduction in stereocilia as well as neurotransmitter release from
the basolateral membrane . In human primary endothelial
cells, Atp2b2 is found to bind with endogenous eNOS, leading to
the phosphorylation of eNOS and downregulation of its activity;
furthermore, NO production by endothelial cells was significantly
reduced by ectopic expression of Atp2b2 .
Overlap between Brain Vasculome and Plasma Protein
By acting as a sensor and integrator of brain dysfunction,
endothelial cells within the vast network of cerebral microvessels
may represent a critical contributor to CNS biomarkers in
circulating blood . We compared our mouse brain vasculome
with four independent proteomic databases of human plasma
proteins (PMID16041672, PMID16335952, PMID16684767, and
PMID18632595) [199,200,201,202,203,204]. Protein products
corresponding to 754, 1211, 781, and 723 genes respectively,
were detected in the mouse brain vasculome (Table 5; complete
gene list is provided in Table S3). To be more conservative, we
defined a core plasma protein set as the intersection of all 4
databases. This yielded 387 proteins. Out of this core plasma
protein dataset, 100 proteins (25.8%) were expressed in the brain
vasculome. Whether these ‘‘hits’’ from the normal brain
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org 10 December 2012 | Volume 7 | Issue 12 | e52665
vasculome or future analyses of diseased brain vasculomes may
eventually lead to measurable biomarkers in blood remains to be
This study presented initial proof-of-concept for a brain
vasculome. The dense network of microvessels in the brain can
no longer be simply viewed as inert plumbing. Cerebral
endothelium may also be an important source of signaling and
trophic factors that communicate with the brain parenchyma.
Hence, the brain vasculome may offer a critical tool for
investigating how the neurovascular system contributes to the
physiology of normal brain function, the pathophysiology of
stroke, brain injury and neurodegeneration, as well as provide
a database for potential circulating biomarkers that are produced
by endothelium in CNS disorders. Our initial analyses suggest that
the mouse brain vasculome (1) is unique and significantly different
from heart and glomerular vascular systems; (2) is enriched in
many vital signaling networks; (3) includes key elements that may
contribute to CNS disorders; (4) contain many common genes that
have been identified in GWAS databases for stroke, AD and PD;
and (5) show significant overlap with plasma protein databases of
potential biomarkers in circulating blood.
Taken together, this proof-of-concept study suggests that, when
integrated with other genomic and proteomic databases, the brain
vasculome may provide a valuable tool for dissecting disease
mechanisms, assessing new therapeutic targets as well as searching
for new biomarkers in CNS disorders. Nevertheless, there are
several caveats that must be kept in mind. First, there is the
possibility of gene contributions from non-cerebral-endothelial cell
types. Comparisons with other neuronal and glial databases
suggest that this may not be a major problem. But we still can not
unequivocally exclude this potential source of false positives.
Second, although we only focus on endothelial cells in this initial
draft of the vasculome, the neurovascular system obviously
includes perivascular cells such as pericytes and smooth muscle
cells. How the brain vasculome interacts with and is regulated by
these other cells warrant deeper studies. Third, our database is
based on samples prepared from the entire brain cortex in order to
maximize signal-to-noise. But it is likely that the neurovascular
system differs in genomic status and function depending on brain
region. Whether higher resolution maps of the brain vasculome
can be rigorously obtained in the future remains to be determined.
Fourth, our vasculome will not operate in isolation but should
significantly interact with multiple systems in the entire body. Our
data already suggest that vasculome profiles are regulated by the
different milieus of each ‘‘host’’ organ. It is likely that the
vasculome would also interact with circulating blood cells insofar
as genomic signatures in circulating blood are affected by stroke,
trauma and various CNS disorders . Fifth, the current draft
of our brain vasculome is focused only on mRNA, i.e. the
transcriptome. However, other modes of genomic information,
including single-nucleotide polymorphism (SNP), copy-number
variation (CNV), and epigenomics should also be studied and
integrated, in order to obtain a full molecular landscape of the
neurovascular system. Ultimately, proteomic and metabolic maps
of the brain vasculome should also be extremely useful. Finally, the
brain vasculome should be mapped across disease models and
states in stroke, brain trauma and neurodegeneration. The normal
vasculome presented here only provides a physiologic baseline.
Clearly, the vasculome is connected to CNS disease as suggested
by the significant overlaps with many GWAS studies of stroke, AD
and PD. Mapping the brain vasculome in aged and diseased
mouse models may allow us to understand how this system is
pathophysiologically affected by and responds to various triggers of
injury and disease.
In conclusion, this study provided initial proof-of-concept for
a mouse brain vasculome. Mapping and dissecting the full
profile of the brain vasculome in health and disease may
provide a novel database for investigating disease mechanisms,
assessing therapeutic targets and exploring new biomarkers for
Materials and Methods
Preparation of Microvessel Endothelial Cells
Ten week old male C57BLKS/J mice (Jackson Labs) were used.
All experiments were reviewed and approved by a Subcommittee
for Research Animal Care of the Massachusetts General Hospital
IACUC (Institutional Animal Care and Use Committee) and all
these institutionally-approved animal protocols are consistent with
the NIH Guide for the Care and Use of Laboratory Animals. To
measure the vasculome, we extracted endothelial cells from brain,
heart and kidney glomeruli, with modified method from previously
published protocols [206,207]. Briefly, mice were anesthetized by
isofluorane and perfused with 8610‘7 inactivated Dynabeads
diluted in 40 ml of HBSS (Invitrogen). The cerebral cortex, heart
and kidneys were dissected and combined from 5 mice, minced
and digested in Collagenase A at 37uC for 30–40 minutes with
vigorous shaking (2 mg/ml for cortex and heart, 1 mg/ml for
kidney). The digested tissue were mechanically dissociated by
titurating, filtered through a 70 mM cell strainer (Becton Dickinson
Labware, Bedford, MA), and centrifuged at 5006g for 5 minutes
at 4uC. For kidney, materials were further filtered twice with
a 100 mM and a 70 mM cell strainer. Cell pellets from brain cortex
Table 5. Expression of plasma proteins in the vasculome of mouse brain.
Data source plasma protein
plasma proteins expressed in brain
PMID:16041672336575422.4Muthusamy B. et al, 2005
PMID:163359523344 72321.6 Liu T. et al, 2005
PMID:16684767283778127.5Liu T et al, 2006
PMID:186325955776 121121.0Qian WJ. et al, 2008
Note: *core is the intersect of all 4 independent data set. Lists of circulating proteins in human plasma were compiled from 4 different proteomic studies, then each
study was overlapped with the expression profile of the brain vasculome. A core set of 387 proteins were defined as common proteins detected in all 4 human plasma
protein studies. Out of the core set of plasma proteins, 100 proteins were expressed in the brain vasculome.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org11December 2012 | Volume 7 | Issue 12 | e52665
and heart were resuspended in cold HBSS and mounted on
magnetic separator to remove Dynabeads, then supernatant was
collected and centrifuged, and incubated with PECAM-1 coated
Dynabeads (5 ml for each organ from one mouse) for 30 minutes at
4uC with rotation. A magnetic separator was used to recover bead-
bound endothelial cells. Cell pellets from kidney were also
resuspended in HBSS and mounted directly on the magnetic
separator to select glomeruli containing Dynabeads. Purified
glomeruli were further digested with 5 mg/ml of type V
Collagenase (Sigma) at 37uC for 30 minutes with agitation,
magnetically separated, and then Digested glomeruli were
centrifuged, resuspended and incubated with PECAM-1 coated
Dynabeads. After washing all materials with HBSS for 5 times,
recovered endothelial cells from all organs were lysed with buffer
RLT plus for RNA preparation, with RNeasy Micro Plus kit
Real Time PCR
Relative expressions of selected markers for different types of
cells were tested with RT-PCR, with pre-designed primers and
Syber Green system from Bioscience. First strand cDNA was
(Qiagen). Date normalization was performed by quantification
of the endogenous 18S rRNA, and fold change was measured
with 22DDCtmethod. The markers for endothelial cells included
VE-cadherin, PECAM-1 and eNOS. To ensure that our brain
vasculome was not contaminated by parenchymal non-endothe-
lial cells, we also checked markers for astrocytes (Aquaporin-4,
GFAP), markers for neurons (MAP-2, Neurogranin), and
markers for pericytes and smooth muscle cells (smooth muscle
alpha-Actin Acta2), calponin 1 CNN1, desmin, myosin heavy
polypeptide 11 Myh11, transgelin Tagln). For heart and
glomerular preparations, we checked markers for myocytes
(Myh6, NKX 2–5), markers for glomerular podocytes (Nphs-1,
Nphs-2) and markers for kidney tubules (Cadherin-16, Claudin-
Transcriptional Profiling with Microarray
Three RNA samples for each organ were individually
hybridized to Affymetrix GeneChip Mouse Genome 430 2.0
microarrays, after checking the RNA quantity and quality. RNA
concentration was measured by Nanodrop, and the integrity of
RNA was tested with RNA integrity number (RIN) score on
Agilent Bioanalyzer 2010. All samples were used only when
RIN scores were verified to be larger than 7.0. Microarray
hybridization and scanning was performed after amplification
with the NuGEN Ovation WTA Pico kit and fragmentation
and labeling with Encore Biotin Module. Raw expression data
for each chip was summarized and normalized using RMA
algorithm, to allow direct comparison of results obtained among
different chips. The quality of each chip was determined by
manually checking mean values, variances and paired scatter
plots as well as Principal Component Analysis (PCA) plots. All
chips passed the quality check. Among the large amount of
probes/genes, we only focused on genes whose maximal
expression values across all microarrays were great than 200,
while the probes with intensity less than 200 were eliminated for
Identification of Organ Specifically Expressed Genes
The specific genes between two groups were identified based on
both statistical significances, which were determined using SAM
algorithm (a variant of t-test and specifically designed for
microarray data), and fold change. To minimize false positives,
only the genes with maximum expression values across all
microarrays greater than 200 were analyzed here. The genes with
p,0.01 and fold change .4 were considered as specifically
expressed. The combination of p value and fold change threshold
serves to eliminate most false positives, as suggested from a large
microarray study led by FDA . Fisher’s exact test was used to
identify the potential enriched pathways from these brain
endothelial specific genes.
Protein-protein Interaction (PPI) Networks
PPI datasets for human and mouse were downloaded from
BIOGRID database at version 3.1.71. Since there were only
2314 proteins and 4118 interactions in the mouse PPI dataset,
we transformed human PPI information into mouse’s based on
the homolog genes between human and mouse according to
NCBI HomolGene database. The human PPI dataset contained
10121 proteins and 52693 interactions. After combining the
native mouse dataset and the transformed mouse dataset and
deleting repeated records and self-self interaction records,
a mouse PPI network with 9189 proteins and 36073 interactions
was built. Since not all proteins in the networks are expressed in
the endothelial cells under this study, we further shrink the
network to EC-specific PPI (EC-PPI) network by deleting the
proteins that are not expressed
interactions. The EC-PPI contains 4243 proteins and 10825
interactions. The properties of network were calculated with
IGRAPH package in R. The PPI networks were visualized by
Cytoscape software with force-directed layout.
and their corresponding
GWAS and Plasma Protein Databases
Genome-wide- association-studies of disease select the risk
genes for the disease. GWAS-identified disease genes for stroke,
Alzheimer’s disease and Parkinson’s disease were collected
sgap_plus.htm) to analyze the expression of such disease-related
genes in endothelial cells. The expression of human plasma
proteins were also tested in the brain vasculome. Human
plasma proteins determined by proteomics from 4 different
studies were used [199,200,201,202,203,204]. A core set of
human plasma proteins was build with proteins detected in all
of these 4 studies, consisting of 387 individual proteins. It is
worthwhile to notice that GWAS and plasma protein databases
evolve and grow over the time, correlations with our brain
vasculome will have to be continually re-assessed in future
All statistical analyses were performed with the statistics
software R (Version 2.6.2; available from http://www.r-project.
org) and R packages developed by the BioConductor project
(available from http://www.bioconductor.org). Overall, raw
expression data for each chip was summarized and normalized
using RMA algorithm, genes with maximum expression levels
across all microarrays great than 200 were considered for further
analysis. Organ specifically expressed genes were identified using
SAM algorithm; Fisher’s exact test was used to identify the
enriched pathways from these organ specific genes. Only genes
with p,0.01 and fold change .4 were considered as specifically
expressed. The combination of p value and fold change threshold
serves to eliminate most false positives, as validated by a large
microarray study led by FDA . Fisher’s exact test was also
used to test the enrichment of GWAS genes for each disease in the
vasculome of mouse brain.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org12December 2012 | Volume 7 | Issue 12 | e52665
and kidney glomerular endothelial cells. The expression of
different cell type specific genes were tested by RT-PCR, and
compared between endothelial cells and corresponding whole
Purity of isolation protocols for brain, heart
pressed in brain, heart and kidney glomeruli.
List of endothelial genes specifically ex-
expressed in brain vasculome. The label ‘‘brain EC
expressed’’ indicates whether the gene is expressed in brain
vasculome (True) or not (False).
Full list of GWAS disease associated genes
Full list of plasma proteins expressed in brain
Thanks to Francis Luscinskas and Veronica Azcutia Criado (Brigham and
Women’s Hospital and Harvard Medical School, Boston, MA) for helpful
discussions about the isolation of endothelial cells.
Conceived and designed the experiments: SG YZ EHL. Performed the
experiments: SG JL CX ATS. Analyzed the data: YZ EHL SG. Wrote the
paper: SG YZ JL MMN XJ EHL.
1. Gorelick PB, Scuteri A, Black SE, Decarli C, Greenberg SM, et al. (2011)
Vascular contributions to cognitive impairment and dementia: a statement for
healthcare professionals from the american heart association/american stroke
association. Stroke 42: 2672–2713.
2. Guo S, Lo EH (2009) Dysfunctional cell-cell signaling in the neurovascular unit
as a paradigm for central nervous system disease. Stroke 40: S4–7.
3. Lecrux C, Hamel E (2011) The neurovascular unit in brain function and
disease. Acta Physiol (Oxf) 203: 47–59.
4. Zlokovic BV (2011) Neurovascular pathways to neurodegeneration in
Alzheimer’s disease and other disorders. Nat Rev Neurosci 12: 723–738.
5. Attwell D, Buchan AM, Charpak S, Lauritzen M, Macvicar BA, et al. (2010)
Glial and neuronal control of brain blood flow. Nature 468: 232–243.
6. Abbott NJ (2002) Astrocyte-endothelial interactions and blood-brain barrier
permeability. J Anat 200: 629–638.
7. Bell RD, Winkler EA, Singh I, Sagare AP, Deane R, et al. (2012)
Apolipoprotein E controls cerebrovascular integrity via cyclophilin A. Nature
8. Ronaldson PT, Davis TP (2012) Blood-Brain Barrier Integrity and Glial
Support: Mechanisms that can be Targeted for Novel Therapeutic Approaches
in Stroke. Curr Pharm Des 18: 3624–3644.
9. Zacchigna S, Ruiz de Almodovar C, Carmeliet P (2008) Similarities between
angiogenesis and neural development: what small animal models can tell us.
Curr Top Dev Biol 80: 1–55.
10. Shen Q, Goderie SK, Jin L, Karanth N, Sun Y, et al. (2004) Endothelial cells
stimulate self-renewal and expand neurogenesis of neural stem cells. Science
11. Dugas JC, Mandemakers W, Rogers M, Ibrahim A, Daneman R, et al. (2008)
A novel purification method for CNS projection neurons leads to the
identification of brain vascular cells as a source of trophic support for
corticospinal motor neurons. J Neurosci 28: 8294–8305.
12. Guo S, Kim WJ, Lok J, Lee SR, Besancon E, et al. (2008) Neuroprotection via
matrix-trophic coupling between cerebral endothelial cells and neurons. Proc
Natl Acad Sci U S A 105: 7582–7587.
13. Leventhal C, Rafii S, Rafii D, Shahar A, Goldman SA (1999) Endothelial
trophic support of neuronal production and recruitment from the adult
mammalian subependyma. Mol Cell Neurosci 13: 450–464.
14. Guo S, Som AT, Waeber C, Lo EH (2012) Vascular neuroprotection via TrkB-
and Akt-dependent cell survival signaling. J Neurochem 123 Suppl 2: 58–64.
15. Navaratna D, Guo SZ, Hayakawa K, Wang X, Gerhardinger C, et al. (2011)
Decreased cerebrovascular brain-derived neurotrophic factor-mediated neuro-
protection in the diabetic brain. Diabetes 60: 1789–1796.
16. Garbuzova-Davis S, Rodrigues MC, Hernandez-Ontiveros DG, Louis MK,
Willing AE, et al. (2011) Amyotrophic lateral sclerosis: a neurovascular disease.
Brain Res 1398: 113–125.
17. Neuwelt EA, Bauer B, Fahlke C, Fricker G, Iadecola C, et al. (2011) Engaging
neuroscience to advance translational research in brain barrier biology. Nat
Rev Neurosci 12: 169–182.
18. Zlokovic BV (2008) The blood-brain barrier in health and chronic
neurodegenerative disorders. Neuron 57: 178–201.
19. Daneman R, Zhou L, Agalliu D, Cahoy JD, Kaushal A, et al. (2010) The
mouse blood-brain barrier transcriptome: a new resource for understanding the
development and function of brain endothelial cells. PLoS One 5: e13741.
20. Enerson BE, Drewes LR (2006) The rat blood-brain barrier transcriptome.
J Cereb Blood Flow Metab 26: 959–973.
21. Pardridge WM (2007) Blood-brain barrier genomics. Stroke 38: 686–690.
22. Wallgard E, Larsson E, He L, Hellstrom M, Armulik A, et al. (2008)
Identification of a core set of 58 gene transcripts with broad and specific
expression in the microvasculature. Arterioscler Thromb Vasc Biol 28: 1469–
23. Dougherty JD, Schmidt EF, Nakajima M, Heintz N (2010) Analytical
approaches to RNA profiling data for the identification of genes enriched in
specific cells. Nucleic Acids Res 38: 4218–4230.
24. Doyle JP, Dougherty JD, Heiman M, Schmidt EF, Stevens TR, et al. (2008)
Application of a translational profiling approach for the comparative analysis of
CNS cell types. Cell 135: 749–762.
25. Andras IE, Deli MA, Veszelka S, Hayashi K, Hennig B, et al. (2007) The
NMDA and AMPA/KA receptors are involved in glutamate-induced
alterations of occludin expression and phosphorylation in brain endothelial
cells. J Cereb Blood Flow Metab 27: 1431–1443.
26. Parfenova H, Fedinec A, Leffler CW (2003) Ionotropic glutamate receptors in
cerebral microvascular endothelium are functionally linked to heme oxygenase.
J Cereb Blood Flow Metab 23: 190–197.
27. Collard CD, Park KA, Montalto MC, Alapati S, Buras JA, et al. (2002)
Neutrophil-derived glutamate regulates vascular endothelial barrier function.
J Biol Chem 277: 14801–14811.
28. Beard RS Jr, Reynolds JJ, Bearden SE (2011) Hyperhomocysteinemia increases
permeability of the blood-brain barrier by NMDA receptor-dependent
regulation of adherens and tight junctions. Blood 118: 2007–2014.
29. Mayo JN, Beard RS Jr, Price TO, Chen CH, Erickson MA, et al. (2012)
Nitrative stress in cerebral endothelium is mediated by mGluR5 in
hyperhomocysteinemia. J Cereb Blood Flow Metab 32: 825–834.
30. Helms HC, Madelung R, Waagepetersen HS, Nielsen CU, Brodin B (2012) In
vitro evidence for the brain glutamate efflux hypothesis: brain endothelial cells
cocultured with astrocytes display a polarized brain-to-blood transport of
glutamate. Glia 60: 882–893.
31. Pasterkamp RJ, Giger RJ (2009) Semaphorin function in neural plasticity and
disease. Curr Opin Neurobiol 19: 263–274.
32. Kruger RP, Aurandt J, Guan KL (2005) Semaphorins command cells to move.
Nat Rev Mol Cell Biol 6: 789–800.
33. Sakurai A, Doci CL, Gutkind JS (2012) Semaphorin signaling in angiogenesis,
lymphangiogenesis and cancer. Cell Res 22: 23–32.
34. Banu N, Teichman J, Dunlap-Brown M, Villegas G, Tufro A (2006)
Semaphorin 3C regulates endothelial cell function by increasing integrin
activity. FASEB J 20: 2150–2152.
35. Conrotto P, Valdembri D, Corso S, Serini G, Tamagnone L, et al. (2005)
Sema4D induces angiogenesis through Met recruitment by Plexin B1. Blood
36. Arese M, Serini G, Bussolino F (2011) Nervous vascular parallels: axon
guidance and beyond. Int J Dev Biol 55: 439–445.
37. Krizbai IA, Deli MA, Pestenacz A, Siklos L, Szabo CA, et al. (1998) Expression
of glutamate receptors on cultured cerebral endothelial cells. J Neurosci Res 54:
38. Deli MA, Joo F, Krizbai I, Lengyel I, Nunzi MG, et al. (1993) Calcium/
calmodulin-stimulated protein kinase II is present in primary cultures of
cerebral endothelial cells. J Neurochem 60: 1960–1963.
39. Sakurai T (2012) The role of NrCAM in neural development and disorders–
beyond a simple glue in the brain. Mol Cell Neurosci 49: 351–363.
40. Demyanenko GP, Riday TT, Tran TS, Dalal J, Darnell EP, et al. (2011)
NrCAM deletion causes topographic mistargeting of thalamocortical axons to
the visual cortex and disrupts visual acuity. J Neurosci 31: 1545–1558.
41. Moy SS, Nonneman RJ, Young NB, Demyanenko GP, Maness PF (2009)
Impaired sociability and cognitive function in Nrcam-null mice. Behav Brain
Res 205: 123–131.
42. Nadadur SS, Haykal-Coates N, Mudipalli A, Costa DL (2009) Endothelial
effects of emission source particles: acute toxic response gene expression
profiles. Toxicol In Vitro 23: 67–77.
43. Glienke J, Schmitt AO, Pilarsky C, Hinzmann B, Weiss B, et al. (2000)
Differential gene expression by endothelial cells in distinct angiogenic states.
Eur J Biochem 267: 2820–2830.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org13December 2012 | Volume 7 | Issue 12 | e52665
44. Bottos A, Destro E, Rissone A, Graziano S, Cordara G, et al. (2009) The
synaptic proteins neurexins and neuroligins are widely expressed in the vascular
system and contribute to its functions. Proc Natl Acad Sci U S A 106: 20782–
45. Chui R, Dorovini-Zis K (2010) Regulation of CCL2 and CCL3 expression in
human brain endothelial cells by cytokines and lipopolysaccharide.
J Neuroinflammation 7: 1.
46. Tripathy D, Thirumangalakudi L, Grammas P (2007) Expression of
macrophage inflammatory protein 1-alpha is elevated in Alzheimer’s vessels
and is regulated by oxidative stress. J Alzheimers Dis 11: 447–455.
47. Homey B, Alenius H, Muller A, Soto H, Bowman EP, et al. (2002) CCL27-
CCR10 interactions regulate T cell-mediated skin inflammation. Nat Med 8:
48. Baird JW, Nibbs RJ, Komai-Koma M, Connolly JA, Ottersbach K, et al.
(1999) ESkine, a novel beta-chemokine, is differentially spliced to produce
secretable and nuclear targeted isoforms. J Biol Chem 274: 33496–33503.
49. Nibbs RJ, Graham GJ (2003) CCL27/PESKY: a novel paradigm for
chemokine function. Expert Opin Biol Ther 3: 15–22.
50. Liu C, Gelius E, Liu G, Steiner H, Dziarski R (2000) Mammalian
peptidoglycan recognition protein binds peptidoglycan with high affinity, is
expressed in neutrophils, and inhibits bacterial growth. J Biol Chem 275:
51. Dukhanina EA, Romanova EA, Dukhanin AS, Kabanova OD, Lukyanova TI,
et al. (2008) Interactions and possible functional characteristics of Tag7-
S100A4 protein complex. Bull Exp Biol Med 145: 191–193.
52. Yashin DV, Dukhanina EA, Kabanova OD, Romanova EA, Lukyanova TI, et
al. (2011) The heat shock-binding protein (HspBP1) protects cells against the
cytotoxic action of the Tag7-Hsp70 complex. J Biol Chem 286: 10258–10264.
53. Rehman A, Taishi P, Fang J, Majde JA, Krueger JM (2001) The cloning of a rat
peptidoglycan recognition protein (PGRP) and its induction in brain by sleep
deprivation. Cytokine 13: 8–17.
54. Frey RS, Ushio-Fukai M, Malik AB (2009) NADPH oxidase-dependent
signaling in endothelial cells: role in physiology and pathophysiology. Antioxid
Redox Signal 11: 791–810.
55. Chrissobolis S, Faraci FM (2008) The role of oxidative stress and NADPH
oxidase in cerebrovascular disease. Trends Mol Med 14: 495–502.
56. Askarova S, Yang X, Sheng W, Sun GY, Lee JC (2011) Role of Abeta-receptor
for advanced glycation endproducts interaction in oxidative stress and cytosolic
phospholipase A(2) activation in astrocytes and cerebral endothelial cells.
Neuroscience 199: 375–385.
57. Kristensen AS, Andersen J, Jorgensen TN, Sorensen L, Eriksen J, et al. (2011)
SLC6 neurotransmitter transporters: structure, function, and regulation.
Pharmacol Rev 63: 585–640.
58. Krizbai I, Szabo G, Deli M, Maderspach K, Lehel C, et al. (1995) Expression
of protein kinase C family members in the cerebral endothelial cells.
J Neurochem 65: 459–462.
59. Yang T, Roder KE, Bhat GJ, Thekkumkara TJ, Abbruscato TJ (2006) Protein
kinase C family members as a target for regulation of blood-brain barrier
Na,K,2Cl-cotransporter during in vitro stroke conditions and nicotine
exposure. Pharm Res 23: 291–302.
60. Fleegal MA, Hom S, Borg LK, Davis TP (2005) Activation of PKC modulates
blood-brain barrier endothelial cell permeability changes induced by hypoxia
and posthypoxic reoxygenation. Am J Physiol Heart Circ Physiol 289: H2012–
61. Park M, Shen K (2012) WNTs in synapse formation and neuronal circuitry.
EMBO J 31: 2697–2704.
62. Varela-Nallar L, Alfaro IE, Serrano FG, Parodi J, Inestrosa NC (2010)
Wingless-type family member 5A (Wnt-5a) stimulates synaptic differentiation
and function of glutamatergic synapses. Proc Natl Acad Sci U S A 107: 21164–
63. Kalani MY, Cheshier SH, Cord BJ, Bababeygy SR, Vogel H, et al. (2008) Wnt-
mediated self-renewal of neural stem/progenitor cells. Proc Natl Acad Sci U S A
64. Zechner D, Fujita Y, Hulsken J, Muller T, Walther I, et al. (2003) beta-Catenin
signals regulate cell growth and the balance between progenitor cell expansion
and differentiation in the nervous system. Dev Biol 258: 406–418.
65. Tam SJ, Richmond DL, Kaminker JS, Modrusan Z, Martin-McNulty B, et al.
(2012) Death receptors DR6 and TROY regulate brain vascular development.
Dev Cell 22: 403–417.
66. Stenman JM, Rajagopal J, Carroll TJ, Ishibashi M, McMahon J, et al. (2008)
Canonical Wnt signaling regulates organ-specific assembly and differentiation
of CNS vasculature. Science 322: 1247–1250.
67. Liebner S, Corada M, Bangsow T, Babbage J, Taddei A, et al. (2008) Wnt/
beta-catenin signaling controls development of the blood-brain barrier. J Cell
Biol 183: 409–417.
68. Daneman R, Agalliu D, Zhou L, Kuhnert F, Kuo CJ, et al. (2009) Wnt/beta-
catenin signaling is required for CNS, but not non-CNS, angiogenesis. Proc
Natl Acad Sci U S A 106: 641–646.
69. Chong ZZ, Shang YC, Maiese K (2007) Vascular injury during elevated
glucose can be mitigated by erythropoietin and Wnt signaling. Curr Neurovasc
Res 4: 194–204.
70. Fancy SP, Harrington EP, Yuen TJ, Silbereis JC, Zhao C, et al. (2011) Axin2 as
regulatory and therapeutic target in newborn brain injury and remyelination.
Nat Neurosci 14: 1009–1016.
71. Diep DB, Hoen N, Backman M, Machon O, Krauss S (2004) Characterisation
of the Wnt antagonists and their response to conditionally activated Wnt
signalling in the developing mouse forebrain. Brain Res Dev Brain Res 153:
72. Pi X, Wu Y, Ferguson JE 3rd, Portbury AL, Patterson C (2009) SDF-1alpha
stimulates JNK3 activity via eNOS-dependent nitrosylation of MKP7 to
enhance endothelial migration. Proc Natl Acad Sci U S A 106: 5675–5680.
73. Planutiene M, Planutis K, Holcombe RF (2011) Lymphoid enhancer-binding
factor 1, a representative of vertebrate-specific Lef1/Tcf1 sub-family, is a Wnt-
beta-catenin pathway target gene in human endothelial cells which regulates
matrix metalloproteinase-2 expression and promotes endothelial cell invasion.
Vasc Cell 3: 28.
74. Phng LK, Potente M, Leslie JD, Babbage J, Nyqvist D, et al. (2009) Nrarp
coordinates endothelial Notch and Wnt signaling to control vessel density in
angiogenesis. Dev Cell 16: 70–82.
75. GrandPre T, Li S, Strittmatter SM (2002) Nogo-66 receptor antagonist peptide
promotes axonal regeneration. Nature 417: 547–551.
76. Acevedo L, Yu J, Erdjument-Bromage H, Miao RQ, Kim JE, et al. (2004) A
new role for Nogo as a regulator of vascular remodeling. Nat Med 10: 382–388.
77. Di Lorenzo A, Manes TD, Davalos A, Wright PL, Sessa WC (2011)
Endothelial reticulon-4B (Nogo-B) regulates ICAM-1-mediated leukocyte
transmigration and acute inflammation. Blood 117: 2284–2295.
78. Yu J, Fernandez-Hernando C, Suarez Y, Schleicher M, Hao Z, et al. (2009)
Reticulon 4B (Nogo-B) is necessary for macrophage infiltration and tissue
repair. Proc Natl Acad Sci U S A 106: 17511–17516.
79. de Haan JB, Cristiano F, Iannello RC, Kola I (1995) Cu/Zn-superoxide
dismutase and glutathione peroxidase during aging. Biochem Mol Biol Int 35:
80. Crack PJ, Cimdins K, Ali U, Hertzog PJ, Iannello RC (2006) Lack of
glutathione peroxidase-1 exacerbates Abeta-mediated neurotoxicity in cortical
neurons. J Neural Transm 113: 645–657.
81. Aase K, Ernkvist M, Ebarasi L, Jakobsson L, Majumdar A, et al. (2007)
Angiomotin regulates endothelial cell migration during embryonic angiogen-
esis. Genes Dev 21: 2055–2068.
82. Ishibashi N, Prokopenko O, Weisbrot-Lefkowitz M, Reuhl KR, Mirochnitch-
enko O (2002) Glutathione peroxidase inhibits cell death and glial activation
following experimental stroke. Brain Res Mol Brain Res 109: 34–44.
83. Xiong Y, Shie FS, Zhang J, Lee CP, Ho YS (2004) The protective role of
cellular glutathione peroxidase against trauma-induced mitochondrial dysfunc-
tion in the mouse brain. J Stroke Cerebrovasc Dis 13: 129–137.
84. Wong CH, Bozinovski S, Hertzog PJ, Hickey MJ, Crack PJ (2008) Absence of
glutathione peroxidase-1 exacerbates cerebral ischemia-reperfusion injury by
reducing post-ischemic microvascular perfusion. J Neurochem 107: 241–252.
85. Ishibashi N, Prokopenko O, Reuhl KR, Mirochnitchenko O (2002) In-
flammatory response and glutathione peroxidase in a model of stroke.
J Immunol 168: 1926–1933.
86. Tajima M, Kurashima Y, Sugiyama K, Ogura T, Sakagami H (2009) The
redox state of glutathione regulates the hypoxic induction of HIF-1.
Eur J Pharmacol 606: 45–49.
87. Galasso G, Schiekofer S, Sato K, Shibata R, Handy DE, et al. (2006) Impaired
angiogenesis in glutathione peroxidase-1-deficient mice is associated with
endothelial progenitor cell dysfunction. Circ Res 98: 254–261.
88. Greenberg DA, Jin K (2005) From angiogenesis to neuropathology. Nature
89. Sun Y, Jin K, Xie L, Childs J, Mao XO, et al. (2003) VEGF-induced
neuroprotection, neurogenesis, and angiogenesis after focal cerebral ischemia.
J Clin Invest 111: 1843–1851.
90. Gandhi S, Wood NW (2010) Genome-wide association studies: the key to
unlocking neurodegeneration? Nat Neurosci 13: 789–794.
91. Tossidou I, Kardinal C, Peters I, Kriz W, Shaw A, et al. (2007) CD2AP/
CIN85 balance determines receptor tyrosine kinase signaling response in
podocytes. J Biol Chem 282: 7457–7464.
92. Li C, Ruotsalainen V, Tryggvason K, Shaw AS, Miner JH (2000) CD2AP is
expressed with nephrin in developing podocytes and is found widely in mature
kidney and elsewhere. Am J Physiol Renal Physiol 279: F785–792.
93. Bokoch GM (1998) Caspase-mediated activation of PAK2 during apoptosis:
proteolytic kinase activation as a general mechanism of apoptotic signal
transduction? Cell Death Differ 5: 637–645.
94. Buchner DA, Su F, Yamaoka JS, Kamei M, Shavit JA, et al. (2007) pak2a
mutations cause cerebral hemorrhage in redhead zebrafish. Proc Natl Acad
Sci U S A 104: 13996–14001.
95. Goeckeler ZM, Masaracchia RA, Zeng Q, Chew TL, Gallagher P, et al. (2000)
Phosphorylation of myosin light chain kinase by p21-activated kinase PAK2.
J Biol Chem 275: 18366–18374.
96. Zeng Q, Lagunoff D, Masaracchia R, Goeckeler Z, Cote G, et al. (2000)
Endothelial cell retraction is induced by PAK2 monophosphorylation of
myosin II. J Cell Sci 113 (Pt 3): 471–482.
97. Gavard J, Gutkind JS (2006) VEGF controls endothelial-cell permeability by
promoting the beta-arrestin-dependent endocytosis of VE-cadherin. Nat Cell
Biol 8: 1223–1234.
98. Demyanenko GP, Halberstadt AI, Rao RS, Maness PF (2010) CHL1
cooperates with PAK1–3 to regulate morphological differentiation of
embryonic cortical neurons. Neuroscience 165: 107–115.
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org14December 2012 | Volume 7 | Issue 12 | e52665
99. Zoghbi HY, Orr HT (2009) Pathogenic mechanisms of a polyglutamine- Download full-text
mediated neurodegenerative disease, spinocerebellar ataxia type 1. J Biol Chem
100. Matilla-Duenas A, Goold R, Giunti P (2008) Clinical, genetic, molecular, and
pathophysiological insights into spinocerebellar ataxia type 1. Cerebellum 7:
101. Tong X, Gui H, Jin F, Heck BW, Lin P, et al. (2011) Ataxin-1 and Brother of
ataxin-1 are components of the Notch signalling pathway. EMBO Rep 12:
102. Lam YC, Bowman AB, Jafar-Nejad P, Lim J, Richman R, et al. (2006)
ATAXIN-1 interacts with the repressor Capicua in its native complex to cause
SCA1 neuropathology. Cell 127: 1335–1347.
103. Tsai CC, Kao HY, Mitzutani A, Banayo E, Rajan H, et al. (2004) Ataxin 1,
a SCA1 neurodegenerative disorder protein, is functionally linked to the
silencing mediator of retinoid and thyroid hormone receptors. Proc Natl Acad
Sci U S A 101: 4047–4052.
104. Al-Ramahi I, Lam YC, Chen HK, de Gouyon B, Zhang M, et al. (2006) CHIP
protects from the neurotoxicity of expanded and wild-type ataxin-1 and
promotes their ubiquitination and degradation. J Biol Chem 281: 26714–
105. Zhang C, Browne A, Child D, Divito JR, Stevenson JA, et al. (2010) Loss of
function of ATXN1 increases amyloid beta-protein levels by potentiating beta-
secretase processing of beta-amyloid precursor protein. J Biol Chem 285: 8515–
106. Bratt A, Birot O, Sinha I, Veitonmaki N, Aase K, et al. (2005) Angiomotin
regulates endothelial cell-cell junctions and cell motility. J Biol Chem 280:
107. Troyanovsky B, Levchenko T, Mansson G, Matvijenko O, Holmgren L (2001)
Angiomotin: an angiostatin binding protein that regulates endothelial cell
migration and tube formation. J Cell Biol 152: 1247–1254.
108. Zhao B, Li L, Lu Q, Wang LH, Liu CY, et al. (2011) Angiomotin is a novel
Hippo pathway component that inhibits YAP oncoprotein. Genes Dev 25: 51–
109. Ernkvist M, Aase K, Ukomadu C, Wohlschlegel J, Blackman R, et al. (2006)
p130-angiomotin associates to actin and controls endothelial cell shape. FEBS J
110. Ernkvist M, Birot O, Sinha I, Veitonmaki N, Nystrom S, et al. (2008)
Differential roles of p80- and p130-angiomotin in the switch between migration
and stabilization of endothelial cells. Biochim Biophys Acta 1783: 429–437.
111. Roudier E, Chapados N, Decary S, Gineste C, Le Bel C, et al. (2009)
Angiomotin p80/p130 ratio: a new indicator of exercise-induced angiogenic
activity in skeletal muscles from obese and non-obese rats? J Physiol 587: 4105–
112. Lorber B, Howe ML, Benowitz LI, Irwin N (2009) Mst3b, an Ste20-like kinase,
regulates axon regeneration in mature CNS and PNS pathways. Nat Neurosci
113. Irwin N, Li YM, O’Toole JE, Benowitz LI (2006) Mst3b, a purine-sensitive
Ste20-like protein kinase, regulates axon outgrowth. Proc Natl Acad Sci U S A
114. Huang CY, Wu YM, Hsu CY, Lee WS, Lai MD, et al. (2002) Caspase
activation of mammalian sterile 20-like kinase 3 (Mst3). Nuclear translocation
and induction of apoptosis. J Biol Chem 277: 34367–34374.
115. Lu TJ, Lai WY, Huang CY, Hsieh WJ, Yu JS, et al. (2006) Inhibition of cell
migration by autophosphorylated mammalian sterile 20-like kinase 3 (MST3)
involves paxillin and protein-tyrosine phosphatase-PEST. J Biol Chem 281:
116. Stegert MR, Hergovich A, Tamaskovic R, Bichsel SJ, Hemmings BA (2005)
Regulation of NDR protein kinase by hydrophobic motif phosphorylation
mediated by the mammalian Ste20-like kinase MST3. Mol Cell Biol 25:
117. Zach S, Felk S, Gillardon F (2010) Signal transduction protein array analysis
links LRRK2 to Ste20 kinases and PKC zeta that modulate neuronal plasticity.
PLoS One 5: e13191.
118. Zheng X, Xu C, Di Lorenzo A, Kleaveland B, Zou Z, et al. (2010) CCM3
signaling through sterile 20-like kinases plays an essential role during zebrafish
cardiovascular development and cerebral cavernous malformations. J Clin
Invest 120: 2795–2804.
119. Tsang S, Woo AY, Zhu W, Xiao RP (2010) Deregulation of RGS2 in
cardiovascular diseases. Front Biosci (Schol Ed) 2: 547–557.
120. Osei-Owusu P, Sabharwal R, Kaltenbronn KM, Rhee MH, Chapleau MW, et
al. (2012) Regulator of G protein signaling 2 deficiency causes endothelial
dysfunction and impaired endothelium-derived hyperpolarizing factor-mediat-
ed relaxation by dysregulating Gi/o signaling. J Biol Chem 287: 12541–12549.
121. Lifschytz T, Broner EC, Zozulinsky P, Slonimsky A, Eitan R, et al. (2011)
Relationship between Rgs2 gene expression level and anxiety and depression-
like behaviour in a mutant mouse model: serotonergic involvement.
Int J Neuropsychopharmacol: 1–12.
122. Smoller JW, Paulus MP, Fagerness JA, Purcell S, Yamaki LH, et al. (2008)
Influence of RGS2 on anxiety-related temperament, personality, and brain
function. Arch Gen Psychiatry 65: 298–308.
123. Doupnik CA, Davidson N, Lester HA, Kofuji P (1997) RGS proteins
reconstitute the rapid gating kinetics of gbetagamma-activated inwardly
rectifying K+ channels. Proc Natl Acad Sci U S A 94: 10461–10466.
124. Han J, Mark MD, Li X, Xie M, Waka S, et al. (2006) RGS2 determines short-
term synaptic plasticity in hippocampal neurons by regulating Gi/o-mediated
inhibition of presynaptic Ca2+ channels. Neuron 51: 575–586.
125. Hutchison RM, Chidiac P, Leung LS (2009) Hippocampal long-term
potentiation is enhanced in urethane-anesthetized RGS2 knockout mice.
Hippocampus 19: 687–691.
126. Kammermeier PJ, Ikeda SR (1999) Expression of RGS2 alters the coupling of
metabotropic glutamate receptor 1a to M-type K+ and N-type Ca2+ channels.
Neuron 22: 819–829.
127. Xiao R, Tang P, Yang B, Huang J, Zhou Y, et al. (2012) Nuclear matrix factor
hnRNP U/SAF-A exerts a global control of alternative splicing by regulating
U2 snRNP maturation. Mol Cell 45: 656–668.
128. Kukalev A, Nord Y, Palmberg C, Bergman T, Percipalle P (2005) Actin and
hnRNP U cooperate for productive transcription by RNA polymerase II. Nat
Struct Mol Biol 12: 238–244.
129. Zhao W, Wang L, Zhang M, Wang P, Qi J, et al. (2012) Nuclear to cytoplasmic
translocation of heterogeneous nuclear ribonucleoprotein U enhances TLR-
induced proinflammatory cytokine production by stabilizing mRNAs in
macrophages. J Immunol 188: 3179–3187.
130. Capon F, Bijlmakers MJ, Wolf N, Quaranta M, Huffmeier U, et al. (2008)
Identification of ZNF313/RNF114 as a novel psoriasis susceptibility gene.
Hum Mol Genet 17: 1938–1945.
131. Bijlmakers MJ, Kanneganti SK, Barker JN, Trembath RC, Capon F (2011)
Functional analysis of the RNF114 psoriasis susceptibility gene implicates
innate immune responses to double-stranded RNA in disease pathogenesis.
Hum Mol Genet 20: 3129–3137.
132. Tsyba L, Nikolaienko O, Dergai O, Dergai M, Novokhatska O, et al. (2011)
Intersectin multidomain adaptor proteins: regulation of functional diversity.
Gene 473: 67–75.
133. Seifert M, Ampofo C, Mehraein Y, Reichrath J, Welter C (2007) Expression
analysis of human intersectin 2 gene (ITSN2) minor splice variants showing
differential expression in normal human brain. Oncol Rep 17: 1207–1211.
134. Klein IK, Predescu DN, Sharma T, Knezevic I, Malik AB, et al. (2009)
Intersectin-2L regulates caveola endocytosis secondary to Cdc42-mediated
actin polymerization. J Biol Chem 284: 25953–25961.
135. Causeret F, Terao M, Jacobs T, Nishimura YV, Yanagawa Y, et al. (2009) The
p21-activated kinase is required for neuronal migration in the cerebral cortex.
Cereb Cortex 19: 861–875.
136. Schmid RS, Midkiff BR, Kedar VP, Maness PF (2004) Adhesion molecule L1
stimulates neuronal migration through Vav2-Pak1 signaling. Neuroreport 15:
137. Zhang H, Webb DJ, Asmussen H, Niu S, Horwitz AF (2005) A GIT1/PIX/
Rac/PAK signaling module regulates spine morphogenesis and synapse
formation through MLC. J Neurosci 25: 3379–3388.
138. de la Torre-Ubieta L, Gaudilliere B, Yang Y, Ikeuchi Y, Yamada T, et al.
(2010) A FOXO-Pak1 transcriptional pathway controls neuronal polarity.
Genes Dev 24: 799–813.
139. Asrar S, Meng Y, Zhou Z, Todorovski Z, Huang WW, et al. (2009) Regulation
of hippocampal long-term potentiation by p21-activated protein kinase 1
(PAK1). Neuropharmacology 56: 73–80.
140. Hayashi ML, Rao BS, Seo JS, Choi HS, Dolan BM, et al. (2007) Inhibition of
p21-activated kinase rescues symptoms of fragile X syndrome in mice. Proc
Natl Acad Sci U S A 104: 11489–11494.
141. Chen SY, Huang PH, Cheng HJ (2011) Disrupted-in-Schizophrenia 1-
mediated axon guidance involves TRIO-RAC-PAK small GTPase pathway
signaling. Proc Natl Acad Sci U S A 108: 5861–5866.
142. Tudor EL, Perkinton MS, Schmidt A, Ackerley S, Brownlees J, et al. (2005)
ALS2/Alsin regulates Rac-PAK signaling and neurite outgrowth. J Biol Chem
143. Li W, Guan KL (2004) The Down syndrome cell adhesion molecule (DSCAM)
interacts with and activates Pak. J Biol Chem 279: 32824–32831.
144. Ke Y, Lum H, Solaro RJ (2007) Inhibition of endothelial barrier dysfunction by
P21-activated kinase-1. Can J Physiol Pharmacol 85: 281–288.
145. Stockton R, Reutershan J, Scott D, Sanders J, Ley K, et al. (2007) Induction of
vascular permeability: beta PIX and GIT1 scaffold the activation of
extracellular signal-regulated kinase by PAK. Mol Biol Cell 18: 2346–2355.
146. Kanda S, Miyata Y, Kanetake H (2004) Role of focal adhesion formation in
migration and morphogenesis of endothelial cells. Cell Signal 16: 1273–1281.
147. Rudrabhatla RS, Selvaraj SK, Prasadarao NV (2006) Role of Rac1 in
Escherichia coli K1 invasion of human brain microvascular endothelial cells.
Microbes Infect 8: 460–469.
148. Rudrabhatla RS, Sukumaran SK, Bokoch GM, Prasadarao NV (2003)
Modulation of myosin light-chain phosphorylation by p21-activated kinase 1
in Escherichia coli invasion of human brain microvascular endothelial cells.
Infect Immun 71: 2787–2797.
149. D’Arcy P, Brnjic S, Olofsson MH, Fryknas M, Lindsten K, et al. (2011)
Inhibition of proteasome deubiquitinating activity as a new cancer therapy. Nat
Med 17: 1636–1640.
150. Mazumdar T, Gorgun FM, Sha Y, Tyryshkin A, Zeng S, et al. (2010)
Regulation of NF-kappaB activity and inducible nitric oxide synthase by
regulatory particle non-ATPase subunit 13 (Rpn13). Proc Natl Acad Sci U S A
Mapping the Brain Vasculome
PLOS ONE | www.plosone.org 15 December 2012 | Volume 7 | Issue 12 | e52665