RESEARCH ARTICLE

Aberrant Astrocytes Impair Vascular Reactivity in Huntington Disease Han-Yun Hsiao, PhD,1,2 Yu-Chen Chen, MD,1,2 Chien-Hsiang Huang, BS,1,3 Chiao-Chi Chen, PhD,1 Yi-Hua Hsu, PhD,1 Hui-Mei Chen, BS,1 Feng-Lan Chiu, PhD,4 Hung-Chih Kuo, PhD,4 Chen Chang, PhD,1 and Yijuang Chern, PhD1,2 Objective: Huntington disease (HD) is an inherited neurodegenerative disease caused by the mutant huntingtin gene (mHTT), which harbors expanded CAG repeats. We previously reported that the brain vessel density is higher in mice and patients with HD than in controls. The present study determines whether vascular function is altered in HD and characterizes the underlying mechanism. Methods: The brain vessel density and vascular reactivity (VR) to carbogen challenge of HD mice were monitored by 3D DR2-mMRA and blood oxygenation level–dependent (BOLD)/flow-sensitive alternating inversion recovery (FAIR) magnetic resonance imaging (MRI), respectively. The amount of vascular endothelial growth factor (VEGF)-A and the pericyte coverage were determined by immunohistochemistry and enzyme-linked immunosorbent assay in human and mouse brain sections, primary mouse astrocytes and pericytes, and human astrocytes derived from induced pluripotent stem cells. Results: Expression of mHTT in astrocytes and neurons is sufficient to increase the brain vessel density in HD mice. BOLD and FAIR MRI revealed gradually impaired VR to carbogen in HD mice. Astrocytes from HD mice and patients contained more VEGF-A, which triggers proliferation of endothelial cells and may be responsible for the augmented neurovascular changes. Moreover, an astrocytic inflammatory response, which reduces the survival of pericytes through an IjB kinase–dependent pathway, mediates the low pericyte coverage of blood vessels in HD brains. Interpretation: Our findings suggest that the inflammation-prone HD astrocytes provide less pericyte coverage by promoting angiogenesis and reducing the number of pericytes and that these changes can explain the inferior VR in HD mice. The resultant impaired VR might hinder cerebral hemodynamics and increase brain atrophy during HD progression. ANN NEUROL 2015;78:178–192

N

eurovascular abnormalities occur in many brain disorders and neurodegenerative diseases, including Huntington disease (HD).1 HD is an autosomal dominant neurodegenerative disease characterized by chorea, dystonia, cognitive decline, and behavioral dysfunction.2,3 The causative mutation is a CAG trinucleotide expansion in exon 1 of the huntingtin (HTT) gene.4 We and other laboratories reported earlier that brain vessel density is higher in mice and patients with HD than in controls.1,5 The mechanisms underlying the neurovascular abnormalities in and functional consequences of HD pathogenesis are largely unknown.

Here, we first investigated whether vascular reactivity (VR) is altered in HD mouse brains. VR is a sensitive neurovascular functional indicator that reflects the ability of blood vessels to adequately control cerebral blood flow (CBF), vascular dilation, and oxygenation in response to challenges that alter hemodynamics.6,7 Incorporating a gas challenge paradigm, in which the inhalant is switched from air to carbogen,8,9 into blood oxygenation level– dependent (BOLD) magnetic resonance imaging (MRI) causes healthy blood vessels to exhibit more saturated signals because of increased oxygen saturation caused by

View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.24428 Received Dec 16, 2014, and in revised form Mar 2, 2015. Accepted for publication Apr 7, 2015. Address correspondence to Dr Chern, Institute of Biomedical Sciences, Academia Sinica, Nankang, Taipei 115, Taiwan. E-mail: [email protected]. edu.tw or Dr. Chen Chang, Institute of Biomedical Sciences, Academia Sinica, Nankang, Taipei 115, Taiwan. E-mail: [email protected] From the 1Institute of Biomedical Sciences, Academia Sinica; 2Institute of Neuroscience, National Yang-Ming University; 3Institute of Biomedical Engineering, National Taiwan University; and 4Institute of Cellular and Organismic Biology, Academia Sinica, Taipei, Taiwan.

C 2015 American Neurological Association 178 V

Hsiao et al: HD Astrocytes

TABLE. Summary of Demographic Data and Neuropathology of Human Subjects

ID

Race/Gender

Age, yr

Pathology

Cause of Death

PMI, h

Non–HD-102

Caucasian/F

77

Lung disease/HBP/arthritis

ASCVD

8

Non–HD-103

Caucasian/F

79

Depression

Drug overdose

14

Non–HD-104

Caucasian/F

72

Non–Hodgkin lymphomas/ DM/heart disease

Exsanguination (accident)

19

Non–HD-105

Caucasian/M

75

n.d.

HASCVD

16

Non–HD-106

Caucasian/F

42

HBP/DM

HASCVD

4

Non–HD-107

Caucasian/M

49

n.d.

ASCVD

5

Non–HD-108

Caucasian/F

53

HBP/asthma

HASCVD

15

Non–HD-109

Caucasian/M

58

n.d.

n.d.

9

Non–HD-110

Caucasian/F

72

Non–Hodgkin lymphomas/ DM/heart disease

Exsanguination (accident)

19

HD-03

Caucasian/M

57

HD

ASCVD

15

HD-04

Caucasian/M

58

HD

ASCVD

17

HD-05

Caucasian/F

69

HD (IIIII); acute thalamic hemorrhage; remote (cystic) lacunar infarct

COPD

10

HD-106

Caucasian/M

43

HD

Complications of HDa

10

HD-107

Caucasian/M

55

HD

Complications of HDa

10

HD-108

Caucasian/F

61

HD

Complications of HDa

16

Brains sections of HD patients (n 5 6) and non-HD controls (n 5 9) were purchased from the National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders at the University of Maryland. a Detailed cause of death was unknown. ASCVD 5 arteriosclerotic cardiovascular disease; COPD 5 chronic obstructive pulmonary disease; DM 5 diabetes mellitus; F 5 female; HASCVD 5 hypertensive arteriosclerotic cardiovascular disease; HBP 5 high blood pressure; HD 5 Huntington disease; M 5 male; n.d. 5 not determined; PMI 5 postmortem interval.

vasodilation. This gas-challenged BOLD MRI paradigm was used to study the VR integrity in HD mouse brains. Gas-challenged flow-sensitive alternating inversion recovery (FAIR) MRI was also used to measure the changes in CBF, which served as an additional index of VR. Our data suggest that microvessels in the HD mouse brain react less to hemodynamic challenges, such as carbogen, than do microvessels in the non-HD mouse brain. We next characterized the mechanisms underlying the abnormal VR in the HD mouse brain. The neurovascular unit (NVU) is a highly dynamic structure that comprises multiple cell types, including endothelial cells, vascular smooth muscle cells, pericytes, astrocytes, neurons, and microglia.10,11 Endothelial cells are connected through multiple tight junction proteins, which control blood–brain barrier (BBB) permeability.12 Pericytes directly enwrap up to 80% of the capillary wall to August 2015

control the vascular tone.13–15 Astrocytes with end-feet enwrap most of the capillary wall formed by the endothelial cells and pericytes, and the remaining surface of the endothelial cell that is not covered by pericytes.16 During the development and maturation of the brain, astrocytes secrete various trophic factors (eg, vascular endothelial growth factor [VEGF]) and maintain brain homeostasis.17 Previous studies have demonstrated that direct signaling between astrocytes and pericytes is important for the regulation of cerebrovascular functions.18 The expression of mutant huntingtin protein (mHTT) in astrocytes has been reported in mice and patients with HD.19–21 We hypothesized that HD astrocytes would exhibit altered VR in a gas challenge because of increased endothelial cell proliferation via a VEGF-A–dependent pathway and that these changes would compromise the 179

ANNALS

of Neurology

survival of pericytes. The impairment in cerebral vascular regulation may influence the brain’s high metabolic demand and increase neuronal atrophy during the disease progression of HD.

Materials and Methods Materials The mouse VEGF, mouse tumor necrosis factor (TNF)-a, mouse interleukin (IL)-1b, and bromodeoxyuridine (BrdU) were purchased from Sigma-Aldrich (St Louis, MO). The anti– cleaved caspase 3 antibody was purchased from Cell Signaling (Danvers, MA). The anti–glial fibrillary acidic protein (GFAP) antibody was obtained from Merck Millipore (Billerica, MA). The anti-BrdU antibody was obtained from Santa Cruz Biotechnology (Dallas, TX). The anti–collagen-IV, anti-CD31, anti–a smooth muscle actin (aSMA), and anti–VEGF-A antibodies were purchased from Abcam (Cambridge, MA). The anticlaudin5 and anti–ZO-1 antibodies were obtained from Genetex (Irvine, CA). The antidesmin antibody was obtained from Dako (Carpinteria, CA). The IjB kinase (IKK) inhibitor was purchased from Calbiochem (San Diego, CA). Human brains were obtained from the National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders (University of Maryland, Baltimore, MD). The subjects’ demographic data and neuropathology are summarized in the Table.

Animals and Treatments R6/2 (B6CBA-Tg [HDexon1]62Gpb/3J), Hdh150Q (B6.129P2Hdhtm2Detl/J), N171-82Q (B6C3F1/J-Tg[HD82Gln]81Dbo/ J[N171-82Q]), and GFAP-HD (FVB/N-Tg[GFAP-HTT* 160Q]31Xjl/J) mice were originally obtained from the Jackson Laboratory (Bar Harbor, ME). R6/2 mice harbor exon 1 from the human HTT gene driven by the human HTT promoter.22 The offspring were verified using a polymerase chain reaction (PCR) genotyping analysis of the genomic DNA extracted from the tails (primers: 50 -CCGCTCAGGTTCTGCTTTTA-30 and 50 -GGCTGAGGAAGCTGAGGAG-30 for the transgene).22 The number of CAG repeats in R6/2 mice used in the present studies was 239 6 7.8 (mean 6 standard error of the mean [SEM], n 5 80). Hdh150Q mice express the full-length mutant HTT, which contains 150 glutamines.23 Offspring were determined by a PCR genotyping analysis of the genomic DNA (primers: 50 -CCCATTCATTGCCTTGCTG-30 and 50 0 GCGGCTGAGGGGGTTGA-3 ). N171-82Q mice harbor a truncated human HTT cDNA that encodes the first 171 amino acids of HTT proteins with 82 glutamines. Expression of this truncated mHTT protein is driven by a neuron-specific human prion promoter.24 Offspring were verified by a PCR genotyping analysis of the genomic DNA (primers: 50 -GTGGATACCC CCTCCCCCAGCCTAGACC-30 and 50 -GAACTTTCAGC TACCAAGAAAGACCGTGT-30 for the transgene; 50 -CAAA TGTTGCTTGTCTGGTG-30 and 50 -GTCAGTCGAGTG CACAGTTT-30 for the internal control). GFAP-HD mice express a fragment of human HTT that produces the first 208

180

amino acids of HTT and a stretch of 160 glutamines in astrocytes.19 Offspring were verified by a PCR genotyping analysis of the genomic DNA (primers: 50 -ACTCCTTCATAA AGCCCTCGCAT-30 and 50 -TTCACACGGTCTTTCTTGG TAGC-30 ). Male N171-82Q mice were mated with female GFAP-HD mice to generate GFAP-HD/N171-82Q mice. The preparation of lentiviruses carrying dominant negative IKK25 and intrastriatal injection of indicated lentiviruses were described previously.21 To exclude the potential gender effect on the phenotypes observed in the present study, only female mice were employed. Animal studies were carried out using protocols approved by the Academia Sinica Institutional Animal Care and Utilization Committee (Taiwan).

Cell Cultures Human induced pluripotent stem cells (iPSCs) were prepared from 2 HD patients with 43 CAG repeats and 2 normal controls as described previously.26 The purity of HD iPSC-derived astrocyte culture was determined by immunocytochemical staining using an antibody against GFAP. Nearly 95% of the differentiated cells were GFAP positive. Primary astrocytes were isolated from cortexes of postnatal 12-day-old mice.20 At 30 days in vitro (DIV), immunocytochemical staining showed that 99% of the primary astrocytes were GFAP positive. Primary endothelial cells and pericytes were isolated from brains of 10.5-week-old C57BL/6J mice.27,28 Briefly, brains were minced and digested with papain (Sigma-Aldrich) in Dulbecco modified Eagle medium (DMEM)/F12/0.5mM ethylenediaminetetraacetic acid for 45 minutes at 378C, followed by the addition of DNase I (Sigma-Aldrich) for 10 minutes. Samples were subjected to a 30% continuous Percoll gradient procedure and centrifuged at 8,500 3 g for 20 minutes at 48C, and the middle layer was collected. To isolate endothelial cells, cells were cultured with EC medium (DMEM/F12 supplement containing fetal bovine serum [FBS; 15%], heparin [100lg/ml; Sigma-Aldrich], endothelial cell growth supplement [100lg/ml; BD Biosciences, San Jose, CA], penicillin-streptomycin [1%], and puromycin [6lg/ml; Sigma-Aldrich]) on collagen I–coated culture slides. Approximately 95% of the cells were CD311 endothelial cells at 7 DIV. For pericyte preparation, cells were cultured on uncoated 10cm dishes in pericyte medium (DMEM supplemented with 10% FBS, 1% penicillinstreptomycin, and 6lg/ml puromycin). At 14 DIV, about 95% of the cells were desmin-positive pericytes.

Immunohistochemistry Mouse brains, human brains, and primary cells were stained using the indicated antibody (anti–collagen IV, 1:1,000; antiGFAP, 1:2,000; anti–VEGF-A, 1:100; anticlaudin5, 1:100–250; anti-CD31, 1:100-250; antidesmin, 1:500–2,000; anti–cleaved caspase 3, 1:250; anti-aSMA, 1:500; and anti-BrdU, 1:250) as detailed earlier.24 The negative controls were performed using the same conditions, except that no primary antibody was included. The nuclei were stained with Hoechst 33258. The images were collected using a confocal laser scanning microscope

Volume 78, No. 2

Hsiao et al: HD Astrocytes

FIGURE 1: Mutant huntingtin protein–expressing astrocytes and neurons are sufficient to increase the brain vessel density in mice with Huntington disease (HD). (A–F) The vessel density was determined by using immunohistochemistry staining for collagen IV (CO-IV). The total signals of collagen IV were quantified (n 5 4–6 in each group). Scale bars 5 100lm. Data are presented as the mean 6 standard error of the mean. (A, B) R6/2 and their littermate control mice (12 weeks of age), (C, D) Hdh150Q (homozygous) and their littermate control mice (16–20 months of age), (E, F) N171-82Q and their littermate control mice (16–18 weeks of age); *p < 0.05 (Student t test). (G, H) Increased vessels were observed in glial fibrillary acidic protein (GFAP)/N171-82Q mice, but not in N171-82Q or GFAP mice at the age of 5 weeks. (G) Representative images of the 3D DR2-3mMRA. (H) Quantitative analysis of DR2 (n 5 3 in each group). **Specific comparison of vascular reactivity area between HD (R6/2) and wild-type (WT) mice; p < 0.01 (Student t test).

August 2015

181

ANNALS

of Neurology

FIGURE 2: Blood oxygenation level–dependent (BOLD) magnetic resonance imaging (MRI) revealed gradually impaired vascular reactivity (VR) to carbogen challenge in the cortex and striatum of R6/2 mice. (A, C) Representative images (A, the cortex; C, the striatum) from the gas-challenged BOLD MRI. The arrowheads represent extracted vascular signals that show VR. The arrows represent extracted vascular signals that did not show VR. (B, D) Quantification of the area fraction of hypointense signals (n 5 6 in each group; B, the cortex; D, the striatum). **Specific comparison of the VR areas between HD and wild-type (WT) mice; p < 0.01 (Student t test). W 5 weeks old.

(LSM510 or LSM780; Carl Zeiss, G€ottingen, Germany), and analyzed using ImageJ (NIH, Bethesda, MD) or MetaMorph (Molecular Devices, Sunnyvale, CA). To quantify the pericyte

182

coverage, the confocal CD31 or desmin stacks were converted to gray-level images. Signal intensities greater than the mean 1 2 3 the standard deviation were defined as CD311 or desmin1

Volume 78, No. 2

Hsiao et al: HD Astrocytes

FIGURE 3: Flow-sensitive alternating inversion recovery magnetic resonance imaging revealed gradually impaired vascular reactivity (VR) to carbogen in R6/2 mice. (A) Representative cerebral blood flow (CBF) maps. (B) Quantitative analysis of CBF and the increase in CBF induced by carbogen challenge (n 5 6 in each group). *p < 0.05 and **p < 0.01 (Student t test), specific comparison of the VR areas between Huntington disease and wild-type (WT) mice. W 5 weeks old.

signals. The percentage of area covered by CD311 or desmin1 signals out of the total selected area was calculated. To determine the relative expression of VEGF-A in astrocytes, the VEGF-A intensity in the GFAP1 area was quantified and normalized to the number of astrocytes in each image. All immunofluorescence and histological images were processed by investigators blinded to the experimental conditions.

BrdU Incorporation Assay Endothelial cell proliferation was determined using the BrdU incorporation assay. Primary endothelial cells (7–10 DIV) were treated with BrdU (10lM) and the indicated reagent(s) for 3 days. The cells were fixed and stained with an anti-BrdU antibody and an anti-CD31 antibody to identify endothelial cells or an anticlaudin5 antibody for tight junctions as described above. The number of BrdU-positive cells was normalized to the number of nuclei to obtain the proliferation index.

RNA Extraction and Quantitative Reverse Transcription PCR The RNA was purified using the TRIzol reagent (Invitrogen, Carlsbad, CA) and reverse transcribed into cDNA using SuperScript II (Invitrogen) followed by a DNase I treatment (Promega, Madison, WI). Quantitative PCR was performed using an ABI PRISM 7700 Sequence Detection System (Life Technologies, Carlsbad, CA) with the SYBR Green PCR Master Mix (Life Technologies).20 The expression of VEGF-A was

August 2015

determined using the following primers: 50 -CATGCCAAGTG GTCCCAG-30 and 50 -GGTCTCAATCGGACGGCA-30 , and normalized to a reference gene (glyceraldehyde-3-phosphate dehydrogenase).

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis and Western Blotting Sodium dodecyl sulfate polyacrylamide gel electrophoresis and Western blotting were conducted as described elsewhere.21 The immunoreactive bands were stained using a light-emitting nonradioactive method (ECL; PerkinElmer, Waltham, MA).

Enzyme-Linked Immunosorbent Assay The levels of mouse VEGF-A were determined using enzymelinked immunosorbent assay (ELISA) with a DuoSet ELISA Development kit (R&D Systems, Minneapolis, MN) following the manufacturer’s protocol.21

Promoter Assay The VEGF-A promoter construct was a generous gift from Dr Lee-Young Chau (Institute of Biomedical Sciences, Academia Sinica, Taiwan).29 Primary astrocytes (30 DIV) were transfected with the promoter construct (pVEGF-Luc or pGL3; 1lg) plus the internal control (pRL-TK, 1lg) using Lipofectamine 2000 (Invitrogen) for 48 hours.21 The total lysates were collected, and the luciferase activity was measured using the DualLuciferase Reporter Assay System (Promega).20,21

183

ANNALS

of Neurology

TUNEL Assay For the TUNEL (terminal deoxynucleotide transferase–mediated deoxyuridine triphosphate nick-end labeling) assay, we used the DeadEndTM Fluorometric TUNEL System following the manufacturer’s protocol (Promega).21 The localized green fluorescence of apoptotic cells was detected with a Zeiss LSM 780 Confocal Microscope (Zeiss).

was defined as the CBF increase divided by the basal level expressed as a percentage. For data acquisition and quantification of MRI (3D DR2-mMRA, BOLD MRI, and FAIR), drawing of regions of interest was performed using identical standard operating procedures.1,31

Statistical Analysis MRI Experiments and Data Analysis All experiments were performed on a 7T PharmaScan 70/16 MR scanner (Bruker, Ettlingen, Germany). Mice were anesthetized intraperitoneally using 1.5g/kg urethane (Sigma-Aldrich). The 3-dimensional (3D) DR2-mMRA was performed by T2weighted imaging (T2WI) before and after the intravenous injection of iron oxide particle (Industrial Technology Research Institute, Hsinchu, Taiwan) at a dose of 20mgFe/kg.30 T2WI was acquired using a 3D rapid acquisition with relaxation enhancement sequence with the following parameters: repetition time (TR) 5 2,000 milliseconds, effective echo time (TE) 5 80 milliseconds, echo-train length 5 32, matrix size 5 256 3 256 3 96 (zero-filling to 512 3 512 3 192), field of view (FOV) 5 2 3 2 3 1cm3, and averages 5 2. The BOLD MRI was based on T2*-weighted imaging using a 3D gradient-echo sequence with flow compensation and the following parameters: TR 5 100 milliseconds, TE 5 35 milliseconds, flip angle 5 158, matrix size 5 256 3 256 3 96 (zero-filling to 512 3 512 3 192), FOV 5 2 3 2 3 1 cm3, and averages 5 2. For FAIR MRI, we used the inversion recovery fast spin-echo sequence with the following parameters: TR 5 6,000 milliseconds, effective TE 5 12 milliseconds, echo spacing 5 12 milliseconds, echo train length5 4, slice thickness 5 1mm, matrix size 5 256 3 128 (zero-filling to 256 3 256), FOV 5 2.56 3 2.56cm2, inversion time (TI) 5 1,700 milliseconds, and averages 5 2. The T1 value was measured with the TI values of 500, 1,000, 1,500, 2,000, and 5,000 milliseconds. BOLD and FAIR were first acquired under the air inhalant followed by the carbogen inhalant. For 3D DR2-mMRA, the derivation of DR2 maps was conducted as described earlier.1 For BOLD MRI, the cerebral blood vessels that inhaled air or carbogen and appeared as hypointense signals within the cortex were extracted via an adaptive threshold filter with a window size of 12 3 12 using MATLAB (MathWorks, Natick, MA). The threshold was below the median signal intensity in the 12 3 12 matrix. The extracted vascular pixels were calculated in proportion to the total pixels in each selected brain region and defined as the vascular fraction. Ten slices were assessed. The vascular fraction with hypointense signals under air was extracted and defined as all vessels with VR and non-VR. The vascular fraction with hypointense signals under carbogen was extracted and defined as a non-VR area. The VR area was defined as non-VR pixels subtracted from all extracted vessels. For FAIR MRI, the derivation of the CBF maps was according to our previous study.31 The VR was assessed using the level of CBF increase induced by carbogen switched from air in the region of interest. Under the air-inhaling condition, the CBF was defined as the basal level, whereas the VR level

184

The data were expressed as the mean 6 SEM for the triplicate samples. Unless specifically stated, the statistical significance was determined using Student t test or a 1- or 2-way analysis of variance followed by the Student–Newman–Keuls test using Sigmaplot (v3.1; Systat Software, San Jose, CA). Probability values < 0.05 were considered significant.

Results mHTT-Expressing Astrocytes and Neurons Mediate the Alterations in the NVU Structure in HD Mice We previously reported increased vessel density in the caudate nucleus of HD patients and in the striatum and cortex of a transgenic HD mouse model (R6/2) without apparent BBB damage.1 Here, we evaluated whether such increased vessel density is observed in other HD mouse models established using different genetic strategies. Similar to the findings reported in R6/2 mice (12 weeks old, end stage; Fig 1A, B), immunohistochemistry staining for collagen IV (a vascular marker) showed higher vessel density in the brains of a knockin mouse model (HdH150Q; see Fig 1C, D; 16–20 months old, end stage) than in the brains of wild-type (WT) mice. Conversely, the brains of another HD mouse model (N171-82Q) that expresses mHTT only in neurons had normal brain vessel density (see Fig 1E, F; 16–18 weeks old). Thus, non-neuronal cells likely play a major role in the abnormal NVU in HD mouse brains. The abnormal vessel density was investigated further in vivo by applying 3D DR2-mMRA to N171-82Q, GFAP-HD mice19 (in which mHTT is expressed only in astrocytes), and GFAP-HD/N171-82Q mice (in which mHTT is expressed in both neurons and astrocytes, 5 weeks old, early stage, see Fig 1G, H). No difference in DR2 was found between the WT, HD N171-82Q (5week-old), and GFAP-HD (5-week-old) mice, but DR2 was significantly higher in GFAP-HD/N171-82Q mice (p < 0.01). These findings suggest that the mHTTevoked dysfunctions in neurons and astrocytes are sufficient to trigger the increase in vessel density. Vessels in the Brain of HD Mice Function Abnormally We next used BOLD functional MRI to evaluate VR in HD mice (R6/2). As shown in Figure 2A, WT mice that Volume 78, No. 2

FIGURE 4: Astrocytes in brains from Huntington disease (HD) mice and patients contained more vascular endothelial growth factor (VEGF)-A. (A, B) The frontal cortex of HD patients (n 5 6) and age-matched non-HD subjects (n 5 9) were stained immunocytochemically for VEGF-A (green) and glial fibrillary acidic protein (GFAP; an astrocyte marker, red). The scale bar 5 10lm. (C, D) Non-HD control (NC) and HD induced pluripotent stem cell–derived astrocyte cultures were stained immunocytochemically for GFAP (an astrocyte marker, green) and VEGF-A (red). Scale bar 5 50lm. (E, F) Brain sections were collected from 12week-old R6/2 mice and their littermate control mice (n 5 4–6 in each group) to determine the VEGF-A expression level in the cortex through double immunostaining for VEGF-A (red) and GFAP (green). Scale bar 5 10lm. The fluorescence intensity of VEGF-A in GFAP-positive cells in A, C, E was quantified and is presented in B, D, F, respectively. (G, H) Brain tissues were harvested from different HD mouse models (G, 12-week-old R6/2; H, 16–20-month-old Hdh150Q) and their littermate control mice (n 5 4–6 in each group). The level of VEGF-A was measured by enzyme-linked immunosorbent assay. The data are expressed as the mean 6 standard error of the mean. *p < 0.05 (Student t test). WT 5 wild type.

FIGURE 5: Primary astrocytes from R6/2 mice expressed more vascular endothelial growth factor (VEGF)-A, which triggers proliferation of endothelial cells. Primary astrocytes (30 days in vitro [DIV]) were isolated from R6/2 mice and their littermate controls. (A) The level of VEGF-A in the astrocyte-conditioned medium (ACM) was measured by enzyme-linked immunosorbent assay (ELISA). (B) The expression level of VEGF-A mRNA was determined using quantitative reverse transcription polymerase chain reaction. (C) Primary astrocytes were transfected with a VEGF promoter construct, and total lysates were collected 48 hours later to measure the promoter activity. (D, E) Primary endothelial cells (wild type [WT], 7 DIV) were stimulated with the indicated ACM collected from WT or R6/2 astrocytes in the presence or absence of an anti–VEGF-A–neutralizing antibody (VEGF-A nAb) plus bromodeoxyuridine (BrdU; 10lM) for 72 hours. Endothelial cells were fixed and double immunostained for BrdU (green) and CD31 (an endothelial cell marker, red). Nuclei were stained with Hoechst 33258 (blue). Scale bar 5 20lm. (F) Primary endothelial cells (WT, 7 DIV) were stimulated with VEGF-A (10ng/ml) plus BrdU (10lM) for 72 hours. (E, F) The ratio of BrdU-positive cells was calculated, and at least 300 cells were counted in each condition. Con 5 control. (G) Primary astrocytes (30 DIV) were stimulated with or without the indicated cytokine (10ng/ml tumor necrosis factor [TNF]-a plus 10ng/ml interleukin [IL]-1b) in the absence or presence of an IjB kinase (IKK) inhibitor (IKK inhibitor IV, 10lM) for 72 hours. The ACM was collected and the level of VEGF-A was measured by ELISA. Data are expressed as the mean 6 standard error of the mean of 3 independent experiments. *p < 0.05 (A–C and F, Student t test; E and G, 1-way analysis of variance).

FIGURE 6: The inflammation-prone astrocytes in Huntington disease (HD) brains contributed to the low pericyte coverage of blood vessels in HD brains via an IjB kinase (IKK)-dependent pathway. (A, B) Frontal cortexes from HD patients (n 5 6) and age-matched non-HD subjects (n 5 9) were stained immunocytochemically for desmin (red, a pericyte marker) and CD31 (green, an endothelial cell marker). Scale bar 5 20lm. (C, D) The pericyte coverage of R6/2 mice and their littermate controls (12 weeks old, n 5 4–6 in each group) in the cortex was determined using double immunostaining for desmin (red) and CD31 (green). Scale bar 5 100lm. The arterioles and capillaries are indicated by arrows and arrowheads, respectively. (E–G) Primary pericytes (14 DIV) were stimulated with astrocyte-conditioned medium (ACM) collected from wild-type (WT) or R6/2 astrocytes in the presence or absence of an IKK inhibitor (IKK inhibitor IV, 10lM) for 72 hours. (E, F) The pericytes were fixed and double immunostained for cleaved caspase 3 (green) and a smooth muscle actin (aSMA; a pericyte marker, red) after the indicated treatment. The nuclei were stained with Hoechst 33258 (blue). Scale bar 5 20lm. (F) The ratios of cleaved caspase 3–positive cells were determined. (G) A TUNEL assay was performed in pericytes after the indicated treatment. The ratios of TUNELpositive cells were calculated. At least 300 cells were counted for each condition (F, G). The data are the mean 6 standard error of the mean from 3 independent experiments. *p < 0.05 (B and D, Student t test; F and G, 1-way analysis of variance).

ANNALS

of Neurology

inhaled air exhibited numerous hypointense dots (arrowheads) in the cortex, which represent reactive and nonreactive vessels. Between the ages of 5 and 12 weeks, there was no age-related change in the area fraction of the hypointense dots in WT mice. By contrast, the hypointense dots in the cortex of HD mice (R6/2) increased with age during disease progression, and this increase reflected the increased vessel numbers in HD mice (both reactive and nonreactive). During inhalation of carbogen, most hypointense dots became saturated signals in the WT mice at all ages assessed (arrowheads). This response was observed in 92 to 95% of the extracted vascular pixels, which indicated vessels with VR. The remaining pixels (5–8%) were classified as the non-VR areas, which exhibited unaltered hypointensity under carbogen inhalation. Remarkably, the non-VR areas accounted for 5%, 39%, and 49% at 5, 7, and 12 weeks of age, respectively, in HD mice (see Fig 2B). Similar results were found in the striatum of HD mice (see Fig 2C, D). Another interesting alteration in HD mice (R6/2) is altered CBF. As shown in Figure 3, CBF was measured by FAIR MRI in the axial view. During air inhalation, no significant change with age was observed in WT mice, whereas an increasing trend in CBF was observed in HD mice. Importantly, carbogen inhalation increased CBF by 60 to 77% in WT mice and by 53%, 33%, and 25% in 5-, 7-, and 12-week-old HD mice, respectively, above the basal level. We used the induced increases in CBF to estimate the VR. Given that HD mice exhibited a significantly smaller increase in CBF in response to carbogen from the age of 7 weeks compared with WT mice (see Fig 3), these findings suggest that, although the brain vessel density in HD mice was greater, the VR of these increased vessels to hemodynamic challenges (such as carbogen) was impaired in HD mice. Conversely, the VR in N171-82Q mice, which express mHTT only in neurons, was normal compared with WT mice (data not shown). These data support an important role of astrocytes in mediating the neurovascular abnormalities in HD mice. HD Astrocytes Produce More VEGF-A, Which Promotes Angiogenesis and Subsequently Increases the Density of Brain Vessels in HD Mice We hypothesized that astrocytes might mediate neurovascular abnormalities in HD, and the most likely astrocytic factor responsible for this abnormality is VEGF-A, a major angiogenic growth factor.32 Consistent with our hypothesis, GFAP-positive astrocytes in the frontal cortex of HD patients contained more VEGF-A, as monitored by immunofluorescence staining, compared with nonHD subjects (Fig 4). Similar results were found in the 188

caudate nucleus of HD patients. The relative VEGF-A intensity in the caudate nucleus of HD subjects (n 5 6) was statistically higher than that of non-HD subjects (n 5 9; 3.4 6 0.3 and 1.0 6 0.1, respectively; p < 0.05). Human astrocyte-enriched cultures derived from iPSCs of HD patients also contained more VEGF-A compared with that of non-HD controls. Astrocytes in the cortex of HD mice (R6/2) also expressed a higher level of VEGF-A compared with WT mice. Likewise, the relative VEGF-A intensity in the striatum of R6/2 mice (n 5 6) was statistically higher than that of WT mice (2.8 6 0.3 and 1.0 6 0.2, respectively; p < 0.05). ELISAs also showed that the amounts of VEGF-A were higher in the brains of 2 HD mouse models (R6/2, 12 weeks old; Hdh150Q, 16 months old) than in their littermate controls. Consistently, primary astrocytes isolated from R6/2 mice released more VEGF-A into the astrocyteconditioned medium (ACM; Fig 5A). In contrast, primary microglia isolated from HD mice (R6/2) released a similar level of VEGF-A compared with WT primary microglia (data not shown). These data support the critical role of astrocytes, but not microglia, in mediating the neurovascular abnormalities observed in HD mice. Moreover, HD astrocytes produced more VEGF-A transcripts as assessed using quantitative reverse transcription PCR than did WT astrocytes (see Fig 5B). The VEGF-A promoter activity, determined with a reporter assay, was higher in HD astrocytes than in WT astrocytes (see Fig 5C). Collectively, these findings suggest that the astrocytic VEGF-A levels were higher in mice and HD patients compared with their respective controls. The presence of mHTT appears to increase VEGF-A expression at the transcriptional level in astrocytes. To assess the involvement of VEGF-A in the increase in vascular density, we first treated primary mouse endothelial cells with ACM and measured endothelial cell proliferation using a BrdU-incorporation assay. BrdU incorporation was greater in primary endothelial cells treated with R6/2 ACM for 3 days than in cells treated with WT ACM (see Fig 5D, E). Addition of a VEGF-A–neutralizing antibody effectively blocked the increase in endothelial cell proliferation by HD (R6/2) ACM. This suggested that VEGF-A mediates the HD ACM-induced endothelial cell proliferation. Consistently, treating primary endothelial cells with VEGF-A (10ng/ ml) increased endothelial cell proliferation (see Fig 5F). We previously reported that primary R6/2 astrocytes are prone to inflammation because of overactivation of the IKK–nuclear factor jB (NF-jB) pathway.21 Because proinflammatory cytokines increase astrocyte VEGF-A secretion,33 we next tested whether the proinflammatory cytokines IL-1b and TNF-a, and the Volume 78, No. 2

Hsiao et al: HD Astrocytes

FIGURE 7: IjB kinase (IKK) inhibition normalized the increased vessel density and inferior pericyte coverage in brains from R6/ 2 mice. Mice (7 weeks old, n 5 6) were injected intrastriatally and intracortically with a lentivirus carrying dominant negative IKKc mutant (DN-IKK) or irrelevant control protein (green fluorescent protein [GFP]) for 8 weeks. (A, B) Brain tissues were harvested to determine the vessel density using immunohistochemical staining for collagen IV (CO-IV). The total collagen IV signals were quantified in the cortex. Scale bar 5 100lm. (C) The brain tissues were harvested to measure the vascular endothelial growth factor (VEGF)-A level by enzyme-linked immunosorbent assay. (D, E) The brain sections were collected to determine the pericyte coverage in the cortex by double immunostaining for desmin (red, a pericyte marker) and CD31 (green, an endothelial cell marker). The nuclei were stained with Hoechst 33258 (blue). Scale bar 5 100lm. (E) The fluorescence intensity of desmin in CD31-positive cells was quantified. The data are the mean 6 standard error of the mean. *p < 0.05 (1-way analysis of variance). (F) A schematic representation showing the contribution of inflammation-prone astrocytes to low vascular reactivity (VR) in HD brains. The HD astrocytes express elevated VEGF-A levels, which increases primary brain endothelial proliferation. Astrocyte-mediated VEGF-A production can also be increased by inflammatory cytokines (eg, interleukin (IL)-1b and tumor necrosis factor (TNF)-a). In addition, the higher proinflammatory cytokine levels in HD brains reduce pericyte survival. The increased angiogenesis and low pericyte survival yield poor pericyte coverage and impaired VR in the HD brain. WT 5 wild type.

IKK–NF-jB pathway are involved in the increased vascular density in the HD mouse brain. As hypothesized, adding IL-1b (10ng/ml) and TNF-a (10ng/ml) increased VEGF-A release from both WT and R6/2 astrocytes (see Fig 5G). In addition, suppressing IKK eliminated IL-b/ TNF-a–induced VEGF-A release by R6/2 and WT astrocytes (see Fig 5G), which indicated the importance of IKK in proinflammatory cytokine-induced VEGF-A production. Collectively, these observations suggest that proinflammatory cytokines increase VEGF-A production August 2015

by HD astrocytes through an IKK-dependent pathway and thereby increase vessel density in the HD mouse brain. Aberrant HD Astrocytes Reduce the Survival of Pericytes, Which Decreases Pericyte Coverage and Impairs VR Previous studies have shown that impaired VR might result from low pericyte coverage.34 We next evaluated pericyte coverage of blood vessels in the HD mouse 189

ANNALS

of Neurology

brain. We used immunofluorescence staining for desmin and CD31 as markers of pericytes and endothelial cells, respectively.34 Pericyte coverage of the blood vessels of the frontal cortex was lower in HD patients than in nonHD controls (Fig 6A, B). A similar observation was found in the caudate putamen. Pericyte coverage of the blood vessels in HD patients and non-HD subjects was 0.6 6 0.1 (n 5 6) and 1.0 6 0.1 (n 5 9), respectively (p < 0.05). Lower pericyte coverage of brain vessels was also found in the cortex of HD mice (R6/2) compared with WT mice (see Fig 6C, D). Similarly, pericyte coverage of brain vessels in the striatum was lower in R6/2 mice than in WT mice (0.4 6 0.1 and 1.0 6 0.1, respectively, p < 0.05). Similar findings were also observed by immunostaining for another pericyte marker, plateletderived growth factor receptor, which was used to analyze the pericyte coverage in R6/2 mice (0.2 6 0.1) and WT mice (0.7 6 0.1, p < 0.05). These findings suggest that inferior pericyte coverage appears to be an authentic phenotype of HD. We next treated primary pericytes with the indicated ACM for 3 days. Pericyte survival was evaluated using immunohistochemistry to determine the level of cleaved caspase 3 (an apoptotic marker, green) and aSMA (a pericyte marker, red). R6/2 ACM induced much greater caspase 3 activation in pericytes compared with WT ACM (see Fig 6E, F). We also used TUNEL (see Fig 6G) to assess apoptosis of pericytes. In addition to the higher caspase 3 activation level, R6/2 ACM markedly increased the number of TUNEL-positive pericytes compared with WT ACM. Collectively, these findings indicate that R6/2 ACM reduced pericyte survival compared with WT ACM, which suggests that HD astrocytes are detrimental to pericyte survival. Our previous study indicated that an increase in the IKK–NF-jB–mediated inflammatory response in astrocytes contributes to HD pathogenesis. Blockage of IKK reduced the neuronal toxicity, and ameliorated several HD symptoms of R6/2 mice (eg, decrease in neuronal density, impaired motor coordination, and poor cognitive function).21 To assess whether the overactivated IKK–NF-jB pathway in HD astrocytes is involved in the low HD pericyte vessel coverage, primary astrocytes from R6/2 and WT mice were treated with an IKK inhibitor during the ACM preparation. Inhibition of the IKK-mediated pathway eliminated the detrimental effect of R6/2 ACM on cultured pericytes (see Fig 6E–G). Similarly, inhibition of the IKK pathway in the brains of HD mice (R6/2) using lentiviruses with a dominant negative IKKc mutant (HA-DN-IKKc) reduced the elevated VEGF-A level, normalized the vessel density (Fig 7A–C), and reversed the low pericyte coverage in vivo (see Fig 190

7D, E). These data demonstrate that mHTT-expressing astrocytes are detrimental to pericyte survival and that inhibition of the aberrantly activated IKK–NF-jB– signaling pathway in HD astrocytes protects pericytes.

Discussion Hyperperfusion has been reported in HD patients and mice,5,35–39 and is observed as increased vessel density,1,5 cerebral blood volume,40 or CBF.36 The results from our cellular studies indicate that inflammation-prone astrocytes play a key role in the abnormal HD brain NVU. Excessive VEGF-A released by HD astrocytes may be responsible for the augmented neurovascular changes. We found that astrocytes in the brains of HD patients and HD mice (R6/2) expressed higher VEGF-A levels than their controls and that this increase in VEGF-A level was associated with increased proliferation of the primary brain endothelium. In addition, VEGF-A content was higher in human astrocyte-enriched cultures derived from iPSCs from HD patients compared with those from non-HD controls. This astrocyte-mediated VEGF-A production was elevated further by inflammatory conditions via an IKK–NFjB–dependent pathway, which may be related to the increased vessel density in R6/2 mice and HD patients. Consistent with this hypothesis, inhibiting inflammation with HA-DN-IKKc reduced the brain levels of proinflammatory cytokines and VEGF-A, and the brain vessel density. Therefore, the increased inflammatory response likely contributes to the abnormal neurovascular structures in the HD mouse brain. To understand the potential pathogenic impact of this angiogenic feature of HD NVUs, we investigated VR in HD mouse brains. BOLD MRI indicated that HD mice (R6/2) had more non-VR vessels, and FAIR MRI revealed that HD mice (R6/2) had lower VR. The lower pericyte coverage of vessels in the brains of R6/2 mice, as determined by CD31/desmin double immunofluorescence staining, may explain the lower VR detected by MRI. Our data suggest that higher proinflammatory cytokine levels in HD mouse brain may cause apoptosis of pericytes. In addition to the elevated angiogenesis, the lower pericyte survival in HD apparently yields poor pericyte coverage and impaired VR. Inhibition of the inflammatory response by HA-DN-IKKc reduced the VEGF-A level, normalized vessel density and pericyte coverage, and most importantly, delayed the onset of several major HD symptoms (such as motor function impairment and aggregate formation).21 Collectively, impaired cerebrovascular regulation and vasodilatory capacity in the brains of mice and humans with HD may contribute to the HD pathogenesis.

Volume 78, No. 2

Hsiao et al: HD Astrocytes

Neurovascular abnormalities, including either structural or functional alterations, are a common feature of neurodegenerative diseases and brain disorders. The NVU tightly controls the cerebrovascular system. Dysfunction of cells involved in the NVU causes many types of neurovascular abnormalities, including increased tortuosity, vascular permeability, altered vessel density and size, and abnormal vascular perfusion, CBF, and oxygen consumption.41,42 Narrowed vessels and increased vessel density occur in brain tissues from HD patients and mice,1,43 but BBB leakage is not detected.1 Importantly, disturbance in the regulation of CBF is associated with regional atrophy in certain brain areas in HD patients.36,37 Such cerebral vascular abnormalities are generally considered a separate, secondary pathological condition from the neuronal deficits. However, within the notion of the NVU, non-neuronal components may be equally important as the neuronal element. Our results demonstrate that the mHTT-expressing astrocytes play a key role in controlling the aberrant NVU in the brains of HD mice. Intriguingly, damaged neurons greatly exacerbated the astrocyte-mediated angiogenesis in HD mice, suggesting the importance of non– cell-autonomous toxicity of neurons to the development of abnormal NVUs. The mechanism underlying such neuronal contribution is currently unknown and requires further investigation. Disruption of the close associations between the neurons, astrocytes, endothelial cells, and pericytes within the NVU may result in structural or functional alterations and may contribute to HD pathogenesis. Consistent with this hypothesis, neurovascular uncoupling (eg, greater cerebral blood volume with lower brain metabolism) has been reported in R6/2 mice.39 It is essential to consider these various players to generate a more complete picture of the basis of HD. Our study presents the first evidence of impairment of VR in HD. In the literature, impaired VR has been observed in association with stroke, dementia, and Alzheimer disease (AD). In stroke patients, VR to carbon dioxide was lower in the middle cerebral artery perfusion territory on BOLD MRI compared with a normal hemisphere.7 This impaired VR may also indicate the occurrence of ipsilateral ischemic events in patients with carotid artery disease.23 In AD, diffuse yet predominant impairment of VR is observed in the posterior areas.44 Impaired microvascular VR is associated with cognitive function deterioration in AD patients.45 Investigations into VR in HD are in the preliminary stage. The MRI and mechanistic findings from our experiments have established a foundation for using VR as a prognostic HD marker. Future translational studies that bring VR into the clinical setting are warranted to add therapeutic August 2015

value to this simple, semiquantitative, sensitive MRI measure. In summary, our study reveals that impaired VR is likely to be a new HD symptom and may provide important insights into abnormal blood flow regulation in the HD brain. Because proinflammatory cytokines contribute to both increased endothelial cell proliferation and pericyte apoptosis in HD mouse brain, antiinflammatory agents in combination with other clinical interventions will most likely optimize the treatment outcome in HD patients.

Acknowledgment This work was supported by grants from the National Science Council (NSC97-2321-B-001-030, NSC982321-B-001-017, NSC99-2321-B-001-012, NSC1002321-B-001-009, NSC101-2321-B-001-047, Y.C.; NSC102-2321-B-001-062, NSC100-2319-B-001-003-m, NSC 99-3112-B-001-020, C.C.) and Academia Sinica (AS-100-TP2-B02), Taiwan. The construct of HA-DN-IKKc was a kind gift of the late Dr Paul Patterson.

Authorship H.-Y.H., Y.-C.C., and C.-H.H. contributed equally.

Potential Conflicts of Interest Nothing to report.

References 1.

Lin CY, Hsu YH, Lin MH, et al. Neurovascular abnormalities in humans and mice with Huntington’s disease. Exp Neurol 2013; 250C:20–30.

2.

Ross CA, Tabrizi SJ. Huntington’s disease: from molecular pathogenesis to clinical treatment. Lancet Neurol 2011;10:83–98.

3.

Walker FO. Huntington’s disease. Lancet 2007;369:218–228.

4.

Gusella JF, Wexler NS, Conneally PM, et al. A polymorphic DNA marker genetically linked to Huntington’s disease. Nature 1983; 306:234–238.

5.

Franciosi S, Ryu JK, Shim Y, et al. Age-dependent neurovascular abnormalities and altered microglial morphology in the YAC128 mouse model of Huntington disease. Neurobiol Dis 2012;45:438– 449.

6.

Ringelstein EB, Van Eyck S, Mertens I. Evaluation of cerebral vasomotor reactivity by various vasodilating stimuli: comparison of CO2 to acetazolamide. J Cereb Blood Flow Metab 1992;12:162– 168.

7.

Lythgoe DJ, Williams SCR, Cullinane M, Markus HS. Mapping of cerebrovascular reactivity using bold magnetic resonance imaging. Magn Reson Imaging 1999;17:495–502.

8.

Rauscher A, Sedlacik J, Barth M, et al. Noninvasive assessment of vascular architecture and function during modulated blood oxygenation using susceptibility weighted magnetic resonance imaging. Magn Reson Med 2005;54:87–95.

191

ANNALS

of Neurology

9.

Sedlacik J, Kutschbach C, Rauscher A, et al. Investigation of the influence of carbon dioxide concentrations on cerebral physiology by susceptibility-weighted magnetic resonance imaging (SWI). Neuroimage 2008;43:36–43.

10.

Winkler EA, Bell RD, Zlokovic BV. Central nervous system pericytes in health and disease. Nat Neurosci 2011;14:1398–1405.

11.

Chen CC, Chen YC, Hsiao HY, et al. Neurovascular abnormalities in brain disorders: highlights with angiogenesis and magnetic resonance imaging studies. J Biomed Sci 2013;20:47.

28.

Tigges U, Welser-Alves JV, Boroujerdi A, Milner R. A novel and simple method for culturing pericytes from mouse brain. Microvasc Res 2012;84:74–80.

29.

Lin HH, Lai SC, Chau LY. Heme oxygenase-1/carbon monoxide induces vascular endothelial growth factor expression via p38 kinase-dependent activation of Sp1. J Biol Chem 2011;286:3829– 3838.

30.

Nitta T, Hata M, Gotoh S, et al. Size-selective loosening of the blood-brain barrier in claudin-5-deficient mice. J Cell Biol 2003; 161:653–660.

Lin CY, Lin MH, Cheung WM, et al. In vivo cerebromicrovasculatural visualization using 3D DeltaR2-based microscopy of magnetic resonance angiography (3DDeltaR2-mMRA). Neuroimage 2009;45:824–831.

31.

13.

Bonkowski D, Katyshev V, Balabanov RD, et al. The CNS microvascular pericyte: pericyte-astrocyte crosstalk in the regulation of tissue survival. Fluids Barriers CNS 2011;8:8.

Siow TY, Chen CC, Wan N, et al. In vivo evidence of increased nNOS activity in acute MPTP neurotoxicity: a functional pharmacological MRI study. Biomed Res Int 2013;2013:964034.

32.

14.

Peppiatt CM, Howarth C, Mobbs P, Attwell D. Bidirectional control of CNS capillary diameter by pericytes. Nature 2006;443:700– 704.

Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J 1999;13: 9–22.

33.

15.

Winkler EA, Sengillo JD, Bell RD, et al. Blood-spinal cord barrier pericyte reductions contribute to increased capillary permeability. J Cereb Blood Flow Metab 2012;32:1841–1852.

Chiarini A, Whitfield J, Bonafini C, et al. Amyloid-beta(25-35), an amyloid-beta(1-42) surrogate, and proinflammatory cytokines stimulate VEGF-A secretion by cultured, early passage, normoxic adult human cerebral astrocytes. J Alzheimers Dis 2010;21:915–926.

16.

Oberheim NA, Takano T, Han X, et al. Uniquely hominid features of adult human astrocytes. J Neurosci 2009;29:3276–3287.

34.

Bell RD, Winkler EA, Sagare AP, et al. Pericytes control key neurovascular functions and neuronal phenotype in the adult brain and during brain aging. Neuron 2010;68:409–427.

17.

Allaman I, Belanger M, Magistretti PJ. Astrocyte-neuron metabolic relationships: for better and for worse. Trends Neurosci 2011;34: 76–87.

35.

Harris GJ, Codori AM, Lewis RF, et al. Reduced basal ganglia blood flow and volume in pre-symptomatic, gene-tested persons at-risk for Huntington’s disease. Brain 1999;122(pt 9):1667–1678.

18.

Armulik A, Abramsson A, Betsholtz C. Endothelial/pericyte interactions. Circ Res 2005;97:512–523.

36.

19.

Bradford J, Shin JY, Roberts M, et al. Expression of mutant huntingtin in mouse brain astrocytes causes age-dependent neurological symptoms. Proc Natl Acad Sci U S A 2009;106:22480–22485.

Wolf RC, Gron G, Sambataro F, et al. Magnetic resonance perfusion imaging of resting-state cerebral blood flow in preclinical Huntington’s disease. J Cereb Blood Flow Metab 2011;31:1908– 1918.

37.

20.

Chou SY, Weng JY, Lai HL, et al. Expanded-polyglutamine huntingtin protein suppresses the secretion and production of a chemokine (CCL5/RANTES) by astrocytes. J Neurosci 2008;28:3277–3290.

Chen JJ, Salat DH, Rosas HD. Complex relationships between cerebral blood flow and brain atrophy in early Huntington’s disease. Neuroimage 2012;59:1043–1051.

38.

21.

Hsiao HY, Chen YC, Chen HM, et al. A critical role of astrocytemediated nuclear factor-kappaB-dependent inflammation in Huntington’s disease. Hum Mol Genet 2013;22:1826–1842.

Vis JC, Nicholson LF, Faull RL, et al. Connexin expression in Huntington’s diseased human brain. Cell Biol Int 1998;22:837–847.

39.

Mangiarini L, Sathasivam K, Seller M, et al. Exon 1 of the HD gene with an expanded CAG repeat is sufficient to cause a progressive neurological phenotype in transgenic mice. Cell 1996;87: 493–506.

Cepeda-Prado E, Popp S, Khan U, et al. R6/2 Huntington’s disease mice develop early and progressive abnormal brain metabolism and seizures. J Neurosci 2012;32:6456–6467.

40.

Markus H, Cullinane M. Severely impaired cerebrovascular reactivity predicts stroke and TIA risk in patients with carotid artery stenosis and occlusion. Brain 2001;124:457–467.

Hua J, Unschuld PG, Margolis RL, et al. Elevated arteriolar cerebral blood volume in prodromal Huntington’s disease. Mov Disord 2014;29:396–401.

41.

Brown WR, Thore CR. Review: cerebral microvascular pathology in ageing and neurodegeneration. Neuropathol Appl Neurobiol 2011;37:56–74.

42.

Kalaria RN. Vascular basis for brain degeneration: faltering controls and risk factors for dementia. Nutr Rev 2010;68(suppl 2):S74– S87.

12.

22.

23.

24.

Schilling G, Becher MW, Sharp AH, et al. Intranuclear inclusions and neuritic aggregates in transgenic mice expressing a mutant N-terminal fragment of huntingtin. Hum Mol Genet 1999;8:397– 407.

25.

Khoshnan A, Ko J, Watkin EE, et al. Activation of the IkappaB kinase complex and nuclear factor-kappaB contributes to mutant huntingtin neurotoxicity. J Neurosci 2004;24:7999–8008.

43.

Huang YC, Wu YR, Tseng MY, et al. Increased prothrombin, apolipoprotein A-IV, and haptoglobin in the cerebrospinal fluid of patients with Huntington’s disease. PLoS One 2011;6:e15809.

26.

Hsiao HY, Chiu FL, Chen CM, et al. Inhibition of soluble tumor necrosis factor is therapeutic in Huntington’s disease. Hum Mol Genet 2014;23:4328–4344.

44.

Cantin S, Villien M, Moreaud O, et al. Impaired cerebral vasoreactivity to CO2 in Alzheimer’s disease using BOLD fMRI. Neuroimage 2011;58:579–587.

27.

Neng L, Zhang W, Hassan A, et al. Isolation and culture of endothelial cells, pericytes and perivascular resident macrophage-like melanocytes from the young mouse ear. Nat Protoc 2013;8:709–720.

45.

Silvestrini M, Pasqualetti P, Baruffaldi R, et al. Cerebrovascular reactivity and cognitive decline in patients with Alzheimer disease. Stroke 2006;37:1010–1015.

192

Volume 78, No. 2