PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jcbfmJournal of Cerebral Blood Flow & Metabolism
 
J Cereb Blood Flow Metab. 2016 April; 36(4): 794–807.
Published online 2015 September 30. doi:  10.1177/0271678X15606149
PMCID: PMC4821019

Dysfunction of brain pericytes in chronic neuroinflammation

Abstract

Brain pericytes are uniquely positioned within the neurovascular unit to provide support to blood brain barrier (BBB) maintenance. Neurologic conditions, such as HIV-1-associated neurocognitive disorder, are associated with BBB compromise due to chronic inflammation. Little is known about pericyte dysfunction during HIV-1 infection. We found decreased expression of pericyte markers in human brains from HIV-1-infected patients (even those on antiretroviral therapy). Using primary human brain pericytes, we assessed expression of pericyte markers (α1-integrin, α-smooth muscle actin, platelet-derived growth factor-B receptor β, CX-43) and found their downregulation after treatment with tumor necrosis factor-α (TNFα) or interleukin-1 β (IL-1β). Pericyte exposure to virus or cytokines resulted in decreased secretion of factors promoting BBB formation (angiopoietin-1, transforming growth factor-β1) and mRNA for basement membrane components. TNFα and IL-1β enhanced expression of adhesion molecules in pericytes paralleling increased monocyte adhesion to pericytes. Monocyte migration across BBB models composed of human brain endothelial cells and pericytes demonstrated a diminished rate in baseline migration compared to constructs composed only of brain endothelial cells. However, exposure to the relevant chemokine, CCL2, enhanced the magnitude of monocyte migration when compared to BBB models composed of brain endothelial cells only. These data suggest an important role of pericytes in BBB regulation in neuroinflammation.

Keywords: Pericyte, blood brain barrier, monocyte, neuroinflammation, HIV-1

Introduction

Despite antiretroviral therapy (ART) achieving HIV suppression, neurocognitive complications defined as HIV-associated neurocognitive disorders (HAND) continue to be highly prevalent.1 One explanation could be the constant compromise of the blood brain barrier (BBB) driven by chronic inflammation1 documented in HIV-infected individuals with well-controlled virus replication, yet HAND progression.2 Chronic neuroimmune activation is present in ART-treated patients as indicated by elevated levels of inflammatory factors in the cerebrospinal fluid. Brain endothelial cell injury and its underlying mechanisms have been explored recently.2,3 However, changes in pericytes, an important member of the neurovascular unit (NVU) contributing to the BBB integrity, have not been characterized in the context of HIV-1 infection.

Pericytes wrapped around brain microvascular endothelial cells (BMVEC) on the “brain side” provide support for endothelial barrier function.46 Pericytes regulate central nervous system (CNS) blood flow at the level of capillaries providing neuronal coupling with the NVU.7 CNS microvessels have a ratio of pericytes to endothelial cells that is higher when compared to organs with greater permeability.46 Only recently has the important role of pericytes become apparent, including secretion of key barrier-supporting molecules (angiopoietin-1 [Ang-1], transforming growth factor-β1 [TGF-β1]8], production of basement membrane (BM) components (stabilizing BMVEC monolayers),9 and regulation of leukocyte migration across the BMVEC–pericyte barrier.9,10 Pericytes stimulated with IL-1β or TNFα increased production of pro-inflammatory cytokines, chemokines, and adhesion molecules; many secretory factors were critical for HIV-1 neuropathogenesis.11 Changes in expression of α1/α2 integrins (mediating adhesion to BM) have been detected in TNFα-treated pericytes in vitro and in microvessels in an animal model of multiple sclerosis (MS)12 paralleling decreased alpha smooth muscle actin (α-SMA) expression, indicative of a less differentiated phenotype. Much of the role pericytes play during neuroinflammation remains ill defined, particularly how pericytes contribute to the orchestration of leukocyte navigation across endothelial barriers.6,13

Pericyte loss/dysfunction or their diminished attachment to endothelial cells is known to be associated with more “leaky” microvessels in MS,6 Alzheimer’s disease (AD),14 amyotrophic lateral sclerosis (ALS),15 and stroke16 as part of BBB impairment, neuroinflammation, and neuronal demise. We hypothesized that pericyte dysfunction in HIV-1 infection contributes to BBB impairment and HAND.13 We showed a prominent decrease in pericyte coverage in brains of HIV-1-infected patients (even on ART). We profiled phenotypic changes during exposure to HIV-1/relevant inflammatory factors and correlated them with functional stability of the BBB in vitro and pericyte secretion of molecules supporting barrier function. Our data indicated that pericytes under inflammatory conditions switched to a pro-inflammatory phenotype and produced inflammatory factors associated with HIV-1 neuroinflammation.

Materials and methods

Human brain tissues

The current study included a total of 23 HIV-seropositive cases. Four HIV-seronegative patients were chosen as a control group (matched for age, gender, and racial status). Available demographic (age, gender, race/ethnicity) and clinical information (ART status, CD4 count, comorbidity, and neurocognition) were collected from hospital medical records (Figure 1(a)). Study has been approved by the Institutional Review Board of Temple University School of Medicine according to the ethical guidelines of the Helsinki Declaration of 1975 (and as revised in 1983). Macroscopic and microscopic examination of the brains used a standardized protocol with sections from the neocortex (frontal and parietal), basal ganglia, hippocampus, midbrain, pons, medulla, cerebellum, spinal cord, and any grossly evident lesion. Paraffin sections (5 µm) were stained with hematoxylin–eosin. One section from the frontal cortical lobe and the hippocampus from each case was used for the evaluation of neuroinflammation and BBB structure by immunohistochemistry using the following antibodies: CD68 for macrophage/microglia (1:100, Dako, Carpinteria, CA) and Iba-1 (1:100, Wako Chemicals USA, Richmond, VA), Human Leukocyte Antigen-DR (HLA-DR) (1:50, Dako) for microglia activation, HIV-1 p24 antigen (1:10, Dako), CD31 endothelial cell marker (1:50, Cell Marque, Rocklin, CA), platelet derived growth factor (PDGF)-Rβ (1:50, Santa Cruz Biotechnology Inc., Santa Cruz, CA), αSMA (1:50, Ventana Medical Systems Inc., Tucson AZ), and CD13 for pericytes (1:50, R&D Systems, Minneapolis, MN). Primary antibodies were detected by Vectastain Elite Kit (Vector Laboratories, Burlingame, CA) with either 3,3’-diaminobenzidine (DAB) (single stain) or Vector Blue substrate or Vector Red substrate for double stains. Immunostains for inflammatory markers were assessed in blinded fashion using a semiquantitative score protocol: 0—negative staining, 1—mild positivity (<10% of cells), 2—moderate positivity (10–75%), and 3—strong positivity (75%).

Figure 1.
Patient demographic and clinical information (a) and neuropathologic assessment (b).

Image analysis

Immunohistochemistry for CD31 and CD13 were performed on postmortem tissue (frontal cortex, controls, n = 4; HIV-1 without ART, n = 9; HIV-1 with ART, n = 5) as described above. Samples were imaged under 20× objective magnification using a CCD camera (Coolsnap-EZ, Photometrics Inc) configured to an upright microscope (i80 Eclipse, Nikon). Two sets of images per case, which consisted of five images for CD13 and five for CD31 from the same area, were acquired. To determine the degree of CD13 pericyte coverage on the brain vasculature, ratiometrics analysis were performed using automated particle counting from area calculations using Image J 1.48v (NIH) imaging software. The measurements were based on 1.422 × 105 µm2 of total area (field of view). After software calibration, each image was processed as follows: background subtraction (light background with sliding paraboloid) and color segmentation or color threshold. The background subtraction allows for normalization of the image to remove any noise prior to analysis, the color segmentation is needed in order to isolate the DAB/brown from the immunohistochemistry (IHC), thus removing the hematoxylin (blue) counter stain. Color segmentation can also bracket the optical density for the desired targets. Next, the image was converted into a binary (mask-like) image to isolate the objects to be counted. Then the analyze particle function (with sizes of 10 µm2 to infinity and circularity of 0–0.8) was executed. A user, blinded to the category to which the images belonged, ran an executable macro based on the above criteria to obtain the measurements. The total number of the objects counted were added and then averaged across image sets. The results designated as the pericyte coverage, are shown as the average ratio ± standard error the mean (SEM) of CD13 to CD31 in each case and then averaged per group.

Cells

Primary BMVEC were supplied by Michael J. Bernas and Dr. Marlys H. Witte (University of Arizona Tucson, AZ); they were isolated from the temporal cortex of human brain tissue obtained during surgical removal of eleptogenic foci in adult patients and were maintained in culture as previously described.17 Primary human monocytes were obtained from the University of Pennsylvania Human Immunology Core (Philadelphia, PA), maintained as described,17 and used within 24 hr of isolation. Procedures were approved by the Institutional Review Board of Temple University School of Medicine according to the ethical guidelines of the Helsinki Declaration of 1975 (and as revised in 1983). Primary human brain vascular pericytes were purchased from ScienCell Research Laboratories (Carlsbad, CA). We established pericyte culture conditions providing a quiescent, nonproliferating phenotype similar to ones present in the CNS under physiologic conditions. Use of growth medium containing 5% of growth factors slowed pericyte proliferation rate, with a majority of cells in the G1 phase of the cell cycle. These culture conditions had no effect on pericyte viability (data not shown). Pericyte cultures were free of glial or endothelial cells contamination (Supplemental Material, Figure 1). BBB models were assembled as previously described17 with BMVEC alone, pericytes alone, or BMVEC cocultured over pericytes (5:1 seeding ratio of BMVEC:pericytes).

Monocyte adhesion assays

Monocyte adhesion assays were done as previously described.17 BMVEC and/or pericytes were plated on 96-well plates (at a density of 2.5 × 104 BMVEC/well with or without 0.5 × 104 pericytes/well). Briefly, monocytes were fluorescently labeled (2.5 × 105 cells/ml loaded with calcein-AM [Life Technologies] at 1 μM/1 × 106 cells for 45 min). After stimulation with or without recombinant human tumor necrosis factor-α (TNFα, R&D Systems, 20 ng/ml, 4 hr), the endothelial cells were rinsed and fluorescently labeled monocytes (2.5 × 105 cells/well) were added to the endothelial cell monolayers for 15 min at 37[degree celsius]. After adhesion, monolayers were washed and relative fluorescence of the attached monocytes was acquired on a fluorescence plate reader (Synergy 2, BioTek Instruments, Winooski, VT). The data were calculated based on a standard curve derived from fluorescence intensity of a known number of labeled monocytes and results are expressed as percent of monocytes attached to cell monolayers over the total input monocytes.

Transendothelial migration assays

Transendothelial migration assays were done as previously described.17 BMVEC or BMVEC/pericytes were plated on rat-tail collagen type I coated FluoroBlok tinted tissue culture plates (3 µm pores, BD Biosciences, Franklin Lakes, NJ) at a density of 2.5 × 104 BMVEC/insert and 0.5 × 104 pericytes/insert one week prior to use for migration assays. Medium was replaced, cell monolayers were washed, and monocyte chemotactic protein type 1 (MCP-1/CCL2, 30 ng/ml, R&D Systems) was added to the lower chamber. Monocytes were labeled with calcein-AM as described for adhesion assays, washed, placed in the upper chamber, and allowed to migrate for 2 hr at 37[degree celsius]. The number of migrated monocytes was determined with ImageJ software, version 1.43 (NIH, Bethesda, MD) and is expressed for each experimental condition as the mean of triplicate determinations calculated as the number of migrated monocytes divided by the number of migrated monocytes in untreated, no chemoattractant control. Data are presented as percent of input defined as the number of monocytes that migrated to the lower chamber over the total monocytes initially placed in the upper chamber × 100.

Transendothelial electrical resistance (TEER)

TEER was measured in real time using the ECIS system (Applied Biophysics, model 1600R, Troy, NY) as previously described.17 TEER was measured in BMVEC alone, pericytes alone, or BMVEC cocultured over pericytes (5:1 seeding ratio of BMVEC:pericytes). Data are presented as mean resistance ± SEM.

Determination of HIV-1 p24 concentration

Concentration of p24 antigen was measured in HIV-1-containing supernatant with the HIV-1 p24 Antigen Capture Assay (Advanced BioScience Laboratories, Inc., Kensington, MD), according to manufacturer’s instructions. Assuming that an HIV core is composed of 2000 p24 capsid molecules, 1 ng of p24 antigen corresponds to 1.25 × 107 viral particles.18

Enzyme Linked Immunosorbent Assay (ELISA)

TGF-β1, Ang-1, RANTES/CCL5, interferon-inducible protein-10 (IP-10)/CXCL10, and MCP-1/CCL2 were measured in supernatants of pericytes by ELISA. To mimic inflammation or HIV-1 exposure, pericytes were treated for 24 hr with TNFα (75 ng/ml), IL-1β (75 ng/ml) in pericyte medium with 5% growth supplements. After treatments, the conditioned medium was collected for analysis. TGF-β1 and Ang-1 levels were detected using conventional double sandwich ELISA kits from R&D Systems and RANTES, IP-10, and MCP-1 levels were detected with ELISA kits from RayBiotech, Inc. (Norcross, GA), according to the manufacturers’ instructions.

Western blot

Pericytes were treated with the inflammatory cytokines, TNFα (25, 50, or 75 ng/ml) or IL-1β (25, 50, or 75 ng/ml). Cells were lyzed for total protein using 1 × radioimmunoprecipitation assay (RIPA) lysis buffer (Millipore Corp., Billerica, MA) containing protease inhibitor cocktail (Sigma/Aldrich, St. Louis, MO). The bicinchoninic acid (BCA) protein assay (Thermo Scientific, Waltham, MA) was then used to determine the protein content in the lysates. Protein lysates containing 20 µg of protein were mixed with 6 × loading buffer containing dithiothreitol (DTT) and then boiled for 10 min at 95[degree celsius]. The proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (4–20% precast gels) (Thermo Scientific), followed by electrophoretic transfer to nitrocellulose membranes. The primary antibodies were diluted in 1 × Tris buffered saline (TBS)/0.1% Tween 20 and used to detect the following proteins: anti-platelet derived growth factor receptor (PDGFR)-β (diluted 1:500) and anti-actin (diluted 1:1000) from Santa Cruz Biotechnology (Santa Cruz, CA); anti-connexin 43 (diluted 1:500) from Cell Signaling Technologies (Danvers, MA). All primary antibodies were incubated with the membranes overnight at 4[degree celsius] with gentle shaking. Species-specific peroxidase-conjugated secondary antibodies (diluted 1:1000 or 1:2000) (Thermo Scientific) were incubated with membranes for 1 hr at room temperature. Proteins were detected using Supersignal West-Pico chemiluminescent substrate (Thermo Scientific) with a G:Box Chemi HR16 (Syngene, Frederick, MD) gel documentation system.

Flow cytometry

For fluorescence-activated cell sorting (FACS) analysis, pericytes were detached, washed, and resuspended at 1 × 106 per 100 μl in PBS containing 1% bovine serum albumin (BSA) (Sigma/Aldrich). Cells were surface stained with antibodies against human PDGF-Rβ-APC (allophycocyanin), integrin α1-APC (R&D Systems), intracellular adhesion molecule (ICAM)-1-Pacific Blue (Biolegend, San Diego, CA), CD31—Pacific Blue, glial fibrillary acidic protein-phycoerythrin (GFAP-PE), or vascular cell adhesion molecule (VCAM)-1-APC (BD Pharmingen, San Jose, CA) at 4[degree celsius] for 30 min. Cells were then fixed using intracellular (IC) fixation buffer (eBiosciences, San Diego, CA). For intracellular markers, cells were permeabilized after fixation using permeabilization buffer (eBiosciences) following the manufacturer’s instructions and incubated at room temperature for 30 min in antibodies against human α-SMA-PerCP (R&D Systems), NG2-PE (eBiosciences), desmin (Abcam, Cambridge, MA) with the secondary antibody anti-rabbit AlexaFluor 488 (Life Technologies) or 7-aminoactinomycin D (7-AAD) (Life Technologies) viability staining solution (for proliferation assessment, eBiosciences) with 50 µg/ml RNaseA (Qiagen, Valencia, CA). Cytometric acquisition was performed using a BD FACS Canto II flow cytometer and analyzed with FlowJo software (Tree Star, Inc., Ashland, OR).

PCR array and qPCR

Pericytes were treated with or without TNFα, IL-1β (both at 75 ng/ml), or HIV-1 for 24 hr. Total RNA was extracted using RNeasy® Mini kit (Qiagen). Total RNA (2 µg) was converted to cDNA using high-capacity cDNA Reverse Transcription kit (Life Technologies). A PCR-based microarray assay for evaluating the expression of genes involved in inflammatory response and autoimmunity was done using RT2-profiler PCR Array (PAHS-077) (SABiosciences Corp., Frederick, MD).17 Profiling was performed for cDNA from two separate experiments. Data were analyzed using a web-based analysis tool (SA Biosciences/Qiagen). Specific primers and probes for α-SMA, PDGF-Rβ, Cx43, integrin-1α, Fibronectin 1 (FN1), TGF-1β, Ang-1, and Nidogen 1 (Nid1) genes were purchased from Life Technologies and analyses were performed using the StepOnePlus real-time PCR system (Life Technologies). Amplification was analyzed using the ΔΔCt method, using a web-based data analysis tool (SABiosciences/Qiagen), by normalization to housekeeping genes and fold change calculated from the difference between experimental condition and untreated control. Fold changes of <2 or >2 were considered significant. Data presented are triplicate determinations from two independent experiments.

Quantification of F-actin

For immunofluorescence studies, pericytes were seeded on FN-coated Millicell® EZ slides (EMD-Millipore, Billerica, MA), and treated with or without TNFα, IL-1β (both at 75 ng/ml) for 1 hr prior to stimulation with human PDGFβ (25 ng/ml) (Prospec, Nes Ziona, Israel) for 5 hr as described.19 Rho GTPase G-switch™ activator CN04 (1 µg/ml, Cytoskeleton Inc., Denver, CO) was used as a positive control. Cells were fixed, permeabilized, and stained as described.17,20 Three to four confocal images (covering 20–30 cells each) per treatment were taken at 60 × (1024 × 1024 pixel area) using a Nikon Eclipse 80i microscope (Nikon Instruments Inc., Tokyo, Japan) and processed with Adobe Photoshop CS3 software (Adobe Systems, San Jose, CA). Orientation and anisotropy of fibrillar structures per image were measured using ImageJ software with the FibrilTool plug-in and the average stress fiber density in each captured image was used for statistical analysis as described.17

Statistical analysis

Data are expressed as the mean ± SEM of experiments conducted multiple times. Multiple group comparisons were performed by one-way analysis of variance with Dunnet’s post hoc tests (adhesion, migration, FACS, and ELISA). Statistical analyses were performed utilizing Prism v6.0c software (GraphPad Software Inc., San Diego, CA). Differences were considered significant at p < 0.05.

Results

Brain pericyte coverage diminished in HIV-infected patients

We examined brain tissues derived from 23 HIV-positive patients (11 with ART and 12 without treatment) and four seronegative age-matched controls (Figure 1(a)). Neuropathologic findings described in Figure 1 are characteristic for HIV-associated neuropathology in the ART era. Only four cases (17.3%) showed evidence of primary parenchymal HIV-related pathology. “Classical” HIV encephalitis (one case) with multinucleated giant cells and HIV leukoencephalopathy (one case) were found in our cohort (both in the group without ART). Signs of microglial nodular encephalitis were detected in four brains, all but one belonging to the ART naive group. The neuropathological alterations associated with opportunistic infection were mainly found in the patients without ART. Seven cases (30.4%) had nonspecific neuropathological features (hypoxic/ischemic changes and micro infarcts). Four cases (17.3%) had no significant neuropathology (Figure 1(b)).

Patients who were on ART compared to no ART had a similar degree of neuroinflammation. Image analysis revealed a score of 3+ (based on a scale of 0–4+, see criteria described in the Materials and methods section), which equates to a high index of neuroinflammation for patients without ART (CD68: 40% with ART vs. 71.8% without ART) (Supplemental Table 1). However, there was no statistical significance between the two groups (CD68: 2.45 ± 0.25 with ART vs. 2.83 ± 0.10 without ART). Similarly, we found no difference in the prevalence of microglial activation (3+ in the semiquantitative scale) in both groups (Iba-1: 36.3% in ART vs. 41.7% without ART; HLA-DR: 25% with ART vs. 36.3% without ART). In contrast, HIV-naive patients showed minimal microglial activation (Iba-1: 1.33 ± 0.33% in HIV-naive vs. 1.95 ± 0.21% in HIV-positive).

To assess possible changes in pericyte coverage of the BBB, we evaluated expression of pericyte markers (PDGF-Rβ, CD13) together with markers of macrophage/microglia activation (HLA-DR). Control cases (Figure 2(a,,b))b)) showed strong pericyte labeling (blue) and minimal staining for microglial activation marker, HLA-DR (red). There was a significant reduction in pericyte markers in HIV-positive cases with (Figure 2(c,,d))d)) and without ART (Figure 2(eg)) with varying degree of microglial activation. The most prominent changes in HLA-DR staining were seen in HIV-associated encephalitis (HIVE) case (no ART) paralleling diminution of CD13 labeling of pericytes (Figure 2(h,,i)).i)). These data indicated a decrease in pericyte coverage (immunostaining for CD13) in HIV-infected patients with or without ART. Of note, there was no change of α-SMA expression in arterioles or venules (Figure 2(jl)). Quantitative evaluation of CD13 pericyte coverage (as described in Materials and methods section) of microvessels when compared to CD31 (endothelial cell marker) demonstrated ratio of 1 (100% coverage) in control cases (Figure 2(m)). There was a substantial diminution of CD13 staining (percentage of CD31-positive microvessels) in HIV-1 cases as compared to controls (32% without ART and 33% with ART, p  0.0003) (Figure 2(j)). Statistical analysis showed no difference between HIV cases with or without ART. These data indicate that HIV infection can result in diminished pericyte coverage of CNS capillaries.

Figure 2.
Diminished pericyte coverage of BBB in HIV-1-infected patients paralleling macrophage/microglia activation. (a,b) Strong contiguous staining for PDGF-Rβ (a, blue) or CD13 (b, blue) was detected in control brains, while only resting microglia were ...

Diminished BBB support by pericytes in inflammation

The contribution of different cell types of the NVU to the function of the BBB was assessed by employing in vitro models with single cell or coculture configurations. We used commercially available primary human brain vascular pericytes and evaluated cell-specific markers (PDGF-Rβ, α-SMA, NG2, desmin) by immunocytochemistry and/or FACS (Figure 3 and Supplemental Material, Figure 2). Pericyte cultures were negative for staining with glial (GFAP) or endothelial cell (CD31) markers (Supplemental Material, Figure 1). We showed all pericytes expressed PDGF-Rβ, NG2, and only a portion was positive for α-SMA or desmin (~25%). Isolated pericytes express α-SMA, but low serum in the media decreased levels of α-SMA expression. Moreover, the expression varies along the segments of the vascular tree. In mice, pericytes surrounding capillaries with a diameter of less than 10 µm do not show α-SMA, whereas pericytes of larger vessels like arterioles or postcapillary venules are regularly immune positive for α-SMA throughout the brain.21

Figure 3.
Downregulation of pericyte cell markers by cytokines. Decreased expression of αSMA (a) and α1 integrin (b) upon inflammatory insult in quiescent pericytes. Pericytes (grown in 5% medium) were stimulated with TNFα or IL-1β ...

Because pericytes do not proliferate in the CNS under physiologic conditions, we established culture conditions providing a quiescent, nonproliferating phenotype of human brain pericytes similar to that present in the CNS. Using several concentrations of growth promoting factors, we demonstrated that medium containing 5% of growth factors slowed proliferation rate, with a majority of cells being in the G1 phase of the cell cycle and have low levels of α-SMA expression (Supplemental Material, Figure 2(a) to ((c)).c)). These culture conditions had no effect on pericyte viability or levels of cellular marker expression (up to 96 hr in culture) (Supplemental Material, Figure 2(d) to ((ff)).

Diminution of staining for pericyte markers observed in HIV-1-infected human brain tissues and infected “humanized” mice13,22 may be due to pericyte demise or downregulation of cell-specific markers found in inflammatory or other pathologic conditions. Therefore, we assessed alterations of pericyte markers by FACS after exposure to prototypic inflammatory factors, IL-1β, and TNFα, cytokines that are upregulated during HIV-1 CNS infection. There was a significant downregulation of α1 integrin (63–67%) after inflammatory insult suggesting a possible reduction in adhesion to BM and α-SMA (35–47%), suggesting a less-differentiated phenotype and pericyte dysfunction (Figure 3(a,,b)).b)). PDGF secreted by brain endothelium and other CNS cells plays an important role in pericyte survival and function via PDGF-Rβ signaling23 and its downregulation can result in loss of BBB integrity. We investigated PDGF-Rβ expression (by western blot) in whole cell lysates of TNFα- or IL-1β-treated pericytes. Exposure to pro-inflammatory factors resulted in a 50–60% diminution in PDGF-Rβ expression (Figure 3(c)). As gap junctions play a crucial role in pericyte–BMVEC communication, we also assessed the expression of connexin-43 (CX43) and found a concurrent 40–60% loss in cytokine-exposed pericytes (Figure 3(d)). Downregulation of PDGF-Rβ, α-SMA, Cx43, and α1 integrin was at the transcription level. qPCR data showed that gene expression was diminished after exposure to prototypic inflammatory factors, IL-1β, and TNFα, or exposure to HIV-1 (Supplemental material, Figure 3).

Pericytes provide support for brain endothelial function via secretion of several factors (including Ang-1 and TGF-β1). Because inflammation or exposure to HIV-1 can modify essential pericyte functions, we determined whether inflammation affects secretion of Ang-1 and TGF-β1 in pericytes. Treatment with IL-1β or TNFα for 24 hr resulted in a 40–60% decrease in Ang-1 and a 40–50% decrease in TGF-β1 production (Figure 4(a,,b)).b)). Because pericytes produce components of the BM, we assessed changes in gene expression of several BM components (FN, nidogen, laminin) by qPCR in cytokine- or HIV-1-exposed pericytes. FN and nidogen showed a 60% downregulation in pericytes challenged by HIV-1, TNFα, or IL-1β for 24 hr (Figure 4(c,,d)).d)). There was no significant change in laminin expression under the same conditions (data not shown). We also found a significant 25% decrease (p < 0.05) of tissue inhibitor of metalloproteinases-3 (TIMP3, data not shown) in pericytes treated with TNFα or IL-1β. TIMP3 is a key regulator of BM remodeling via inhibition of matrix metalloproteases (MMPs).3 Taken together, these results indicate potential dysfunction of pericytes (expression of cellular markers and receptors and production of BBB-supporting molecules) leading to barrier compromise.

Figure 4.
Exposure of pericytes to TNFα, IL-1β, or HIV-1 infection diminishes pericyte support of endothelial barrier function. Pericytes were incubated with TNFα, IL-1β (75 ng/ml), or HIV-1 (17 or 34 ng/ml) for 24 h. ...

Inflammatory responses of pericytes

Only recently, the role of pericytes in leukocyte migration became apparent.10,24 The particular contribution of pericytes in leukocyte infiltration into different organs may offer a secondary checkpoint or formation of a “secondary” barrier, for immunomodulation and guidance.10,25,26 However, their role in monocyte migration or retention in the perivascular space has not been explored in the CNS. Virtually nothing is known about brain pericyte involvement in mononuclear cell migration associated with HIV-1 infection in the CNS.

In order to study the role of pericytes on barrier function, we measured TEER of BMVEC alone, pericytes alone or cocultured BMVEC/pericytes (Figure 5(a)) before and after the addition of primary human monocytes. There was a significant steady increase of TEER in BMVEC–pericyte cultures, which was 7–8% higher as compared to BMVEC alone; pericytes alone showed low TEER values (Figure 5(a)). Next, we studied whether the presence of pericytes together with BMVEC would affect monocyte adhesion/migration using in vitro BBB models.17 BMVEC alone, pericytes alone, or BMVEC/pericyte cocultures were stimulated with TNFα. Primary human monocytes were placed on the BMVEC, pericytes, or BMVEC/pericyte cocultures to initiate adhesion after all treatments were removed and the medium was changed (Figure 5(b)). TNFα increased monocyte adhesion twofold in BMVEC only and BMVEC/pericyte cocultures. Interestingly, TNFα stimulation enhanced monocyte adhesion to pericytes threefold, indicating that pericytes can engage monocytes and possibly alter their pattern of infiltration into CNS. However, baseline and TNFα-induced adhesion in pericytes was only 35–50% of that seen in BMVEC monolayers (Figure 5(b)).

Figure 5.
Primary human monocyte adhesion to and migration across BMVEC/pericyte constructs.17 (a) TEER was measured in BMVEC alone, pericytes alone, or BMVEC cocultured over pericytes (5:1 seeding ratio of BMVEC:pericytes). Results are presented as percent change ...

Using migration assays in an in vitro BBB model, we tested whether addition of pericytes to endothelial cells changes the rate of monocyte passage across models, using CCL2 as a relevant cytokine that is upregulated in HIV-1 CNS infection.3 Application of CCL2 to the lower chamber of BMVEC only constructs increased monocyte migration sixfold and sevenfold when BMVEC was combined with pericytes. However, the baseline rate was 60% (without CCL2) and 78% (with CCL2 in the bottom well) in dual BBB models versus BMVEC only models, indicating increased “tightness” of the barrier in the presence of pericytes (Figure 5(c)).

Monocyte adhesion and migration require the engagement of integrins (expressed on monocytes) by adhesion molecules. Therefore, we investigated expression of adhesion molecules on pericytes that could engage monocytes, facilitating their adhesion. Stimulation of pericytes with TNFα resulted in a 7- to 10-fold increase of ICAM-1 and four- to sevenfold enhancement of VCAM-1 expression (Figure 5(d,,e)).e)). IL-1β had less of an effect, increasing upregulation two- to fourfold in ICAM-1 and VCAM-1. These results suggest that pericytes could play a role in mononuclear cell infiltration across the BBB and retention in perivascular spaces.

Similar to astrocytes switching their role from supporting/nurturing to pro-inflammatory,3 pericytes could propagate neuroinflammation under certain conditions.27 To evaluate whether inflammatory stimuli upregulate inflammatory factors in pericytes, we profiled the expression of genes commonly involved in the regulation of inflammation. Using a commercial PCR-based array, 84 genes relevant to inflammatory responses were analyzed in untreated pericytes (control) and pericytes activated with TNFα or IL-1β (75 ng/ml, 24 hr), cytokines that are upregulated during HIV-1 CNS infection. The array analyses revealed that 48 and 38 out of 84 pro-inflammatory genes were upregulated more than twofold in TNFα- or IL-1β-treated pericytes, respectively. These genes included a 100- to 150-fold increase for CCL2, a 700- to 3900-fold increase for CCL5, and a 11- to 200-fold increase for CXCL10 (Figure 6(a)). These results were confirmed by protein detection of CCL2, CXCL10, and CCL5 (10- to 200-fold increases) by ELISA in supernatants from the same pericyte cultures (Figure 6(b)–(d)).

Figure 6.
Upregulation of inflammatory molecules in cytokine-activated pericytes. (a) Identification of genes involved in inflammatory responses upregulated by TNFα and IL-1β in pericytes. Pericytes were incubated with TNFα or IL-1β ...

It is known that pericytes can migrate under pathologic conditions22 and therefore their coverage of the BBB is diminished, potentially leading to barrier compromise. Therefore, we evaluated filopodia formation by pericytes, indicative of their migratory phenotype leading to pericyte detachment. Exposure to cytokines that are upregulated during HIV-1 CNS infection, TNFα, or IL-1β, led to an enhanced number of filopodia (Figure 7(a)–(d)), indicating a migratory phenotype potentially resulting in less coverage of the BBB. These data were confirmed by measurement of stress fiber density in pericytes treated with cytokines or Rho GTPase activator (positive control). There was a significant increase in stress fiber density in treated cells that paralleled the results of filopodia formation (Figure 7(e)). In summary, stimulation of pericytes by pro-inflammatory cytokines (relevant to HIV neuropathogenesis) led to a substantial increase in secretion of inflammatory factors, upregulation of adhesion molecules, and a migratory phenotype, all potentially contributing to BBB compromise.

Figure 7.
Inflammatory mediators increasing filopodia formation in pericytes. Representative images of F-actin of pericytes incubated with (c) TNFα, (d) IL-1β (both at 75 ng/ml) for 1 h or untreated pericytes (a). Rho GTPase G-switch™ ...

Discussion

Recent studies have shown nonproductive HIV infection of pericytes28 and their contribution to cell-free virus penetration of the BBB under inflammatory conditions.29 However, comprehensive studies of pericyte functional changes in HIV-1 infection and chronic neuroinflammation that leads to BBB compromise are lacking. Pericyte interactions with the abluminal surface of the brain endothelium are critical in BBB maintenance, and we found a decrease in pericytes in brains of HIV-1-infected patients known to have barrier compromise (previously mainly attributable to endothelial dysfunction).3,30 Reduction in brain pericyte density and detachment from brain endothelial cells leads to BBB dysfunction.6 Decrease in pericyte coverage has been reported in AD, MS, ALS, diabetic microangiopathy, and stroke,1416,31 while very little is known about pericyte changes in brain tissue of HIV-1-infected patients. For the first time, our data indicate diminished coverage and/or downregulation of pericyte markers in patients with HIV infection untreated or on ART, even without HIV-1 encephalitis.22 Our previous work13 suggested that such changes were accompanied by diminution of tight junction (TJ) protein expression and presence of phosphorylated occludin, indicative of BBB compromise.

A decrease in pericyte-specific markers has been suggested to reflect a “less differentiated” phenotype of these cells.32 Diversity of pericyte function along the vascular hierarchy has been reported.7,13,21,33 Additional research into more thoroughly characterizing the heterogeneity of pericyte function along the vascular network is essential for the ultimate understanding of these clearly important cells in the NVU. Indeed, our results indicate that exposure to prototypic inflammatory factors, TNFα, or IL-1β, that are upregulated during HIV-1 CNS infection,1 leads to diminished expression of αSMA or integrin, arguing for changes in the phenotype of pericytes during inflammation. Furthermore, we found downregulation of PDGF-Rβ expression under the same conditions. This receptor possesses important functions in attracting pericytes to brain endothelium, thereby assuring BBB maturation. To recruit pericytes around vessels, brain endothelial cells produce PDGF-B binding to PDGF-Rβ on pericytes.34 Absence or reduction of PDGF-Rβ resulted in attenuated pericyte coverage and a “leaky” BBB.23 The amount of pericytes at the BBB inversely correlated with TJ abnormalities and increased transport across the barrier.

The role of pericytes in BBB formation is underscored by the fact that astrocytes arise later when the BBB is already established due to the presence of pericytes.35 In addition to αSMA and PDGF-Rβ, we also detected attenuation of α1 integrin expression after exposure of pericytes to cytokines that are elevated during HIV-1 infection of CNS. Such change may lead to defective attachment to BM, impaired microvessel stability, and hemorrhage.36 Downregulation of BM protein mRNA (FN, nidogen, but not laminin) found by us in human pericytes after IL-1β, TNFα, or virus exposure could further affect the structure of microvessels during HIV-1 infection or neuroinflammation. Pericytes, along with brain endothelium, synthesize the BM and its abnormalities are known to be present in a number of neurodegenerative and neurovascular diseases.35

Diminished coverage due to pericyte migration cannot be fully excluded. Pericytes migrate during BBB development, angiogenesis in adult brain, and such pathologic conditions as diabetic retinopathy or traumatic brain injury.27 It has been shown that the HIV-1 protein Tat augmented the migratory capacity of pericytes as an additional mechanism of diminished pericyte coverage in HIV infection.22 Migration of pericytes is accompanied by filopodia formation and alterations of the cytoskeleton (stress fibers). Exposure to cytokines enhanced both filipodia and stress fibers in primary human pericytes, indicating their migratory capacity. Of interest, CN04, an activator of RhoA, caused similar changes pointing to GTPase involvement in this process. Pericytes play a significant role in regulating cerebral blood flow (CBF).7,13,33 Towgood et al. recently reported decreased CBF measured by MRI and PET analyses, as well as diminished cerebral metabolic rate of glucose uptake in HIV patients.37 Since functional CBF is actually reduced in HIV patients, additional research into pericyte function during HIV infection could provide a further interpretation of the role of the pericyte as related to control of CBF. We found no changes in expression of αSMA in arterioles in HIV-infected or control brain tissue.

We should acknowledge limitations of pericyte assessment in vitro as their marker expression or functional responses may change under culture conditions. To address such potential critique, we performed thorough expression profiling of pericytes used in this study. They express pericyte markers such as NG-2, PDGF-Rβ, and do not express astrocytic (GFAP) or endothelial (CD31) markers.

The extent of HIV-1-related endothelial activation (reflective of ongoing chronic inflammation) may be associated with altered levels of the vascular growth factors, Ang-1 and -2, recently identified as important regulators of endothelial quiescence and activation.38 Ang-1 secreted by pericytes plays a key role in maintaining vascular stability by binding to endothelial Tie-2, a receptor tyrosine kinase.39 Our results (significant diminution of Ang-1 and TGF-β1 after HIV or cytokine exposure) provide another mechanism of BBB compromise in neuroinflammation. Decreases in Ang-1 and TGF-β1 secretion and diminished pericyte density (PDGF-Rβ down regulation) can reduce barrier stability.

Of note, deregulation of PDGF production in HIV-1-infected macrophages has been extensively studied before, indicating its pro-inflammatory and neuroprotective properties.40 However, changes in the context of BBB impairment due to pericyte dysfunction have not been considered previously.

Recent reports indicated that pericytes are active participants in leukocyte migration across barriers in different organs.9,10,24 Leukocyte–pericyte interaction considerably enhanced their immunosurveillance function during interstitial migration by supporting fast and directed navigation of leukocytes through tissue.10 Pericytes may act as sensory cells that interpret distinct inflammatory signals and send specific pro-inflammatory and pro-survival instructions to recently infiltrating leukocytes.10 Brain pericytes have been linked to the modulation of T cell responses during neuroinflammatory processes in vitro;25,26 however, their role in monocyte migration or retention in the perivascular space has not been explored in the CNS. Pericytes created an additional barrier and allowed leukocyte migration in areas with less coverage, low expression regions of vascular BM for their transmigration.9 It could be related to increased interactions between leukocytes and pericytes. Indeed, we found increases in expression of adhesion molecules as well as adhesion of primary human monocytes to pericytes activated by cytokines. In autopsy brain tissues, we also showed that migrating monocytes accumulated in perivascular spaces (between the abluminal surface of brain endothelium and pericytes).

In order to understand the contribution of pericytes to BBB function, we established coculture models and demonstrated diminished monocyte migration (without inflammatory stimuli) and increased tightness of barrier constructs composed of BMVEC/pericytes as compared to BMVEC alone. However, application of CCL2 to dual models resulted in proportionally more significant migration when compared to BMVEC-only constructs. These results paralleled major increases in mRNAs and protein secretion of inflammatory factors by pericytes stimulated by TNFα or IL-1β. These factors (like CCL2, CXCR4, CCR5) are relevant to HIV-1 infection inside and outside of the CNS.1 Therefore, it appears that pericytes tighten the BBB under physiologic conditions and can enhance inflammatory responses at the BBB during neuroinflammation.

In summary, our results indicate that CNS inflammation can diminish BBB supportive functions of pericytes and promote their pro-inflammatory phenotype. Such changes, which are underappreciated in chronic neuroinflammation, could accelerate neurodegeneration in HIV infection. Importantly, better understanding of pericyte alterations may provide opportunities for new treatments using this cell type as a target.

Supplementary Material

Supplementary material:

Funding

The work performed in the authors’ laboratories is supported by grants from NIMH, MH65151 (YP), NIAAA, AA015913 (YP), NIDA, 2P30DA013429 (CSAR grant), DA031064 (RP), NIH/NINDS, NS086570 (SHR), The Shriners Hospitals for Children 85110-PHI-14 (SHR), and NINDS, NS087385 (SR).

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

YP collected and analyzed data, article writing, revision and final approval, financial support, Jeremy Hill collected data, article revision, and final approval, MZ analyzed data, article revision, and final approval, HD collected data, article revision, and final approval, MW collected data, article revision, and final approval, NLR collected data, article revision, and final approval, RP analyzed data, article revision, and final approval, AM analyzed data, article revision, and final approval, SHR analyzed data, article revision, and final approval, financial support, SR collected and analyzed data, article writing, and revision and final approval, financial support.

Supplementary material

Supplementary material for this paper can be found at http://jcbfm.sagepub.com/content/by/supplemental-data

References

1. Letendre SL, Zheng JC, Kaul M, et al. Chemokines in cerebrospinal fluid correlate with cerebral metabolite patterns in HIV-infected individuals. J Neurovirol 2011; 17: 63–69. [PMC free article] [PubMed]
2. Andras IE, Pu H, Tian J, et al. Signaling mechanisms of HIV-1 Tat-induced alterations of claudin-5 expression in brain endothelial cells. J Cereb Blood Flow Metab 2005; 25: 1159–1170. [PubMed]
3. Persidsky Y, Ramirez SH, Haorah J, et al. Blood-brain barrier: structural components and function under physiologic and pathologic conditions. J Neuroimmune Pharmacol 2006; 1: 223–236. [PubMed]
4. Armulik A, Genove G, Mae M, et al. Pericytes regulate the blood-brain barrier. Nature 2010; 468: 557–561. [PubMed]
5. 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. [PMC free article] [PubMed]
6. Aguilera KY and Brekken RA. Recruitment and retention: factors that affect pericyte migration. Cell Mol Life Sci 2013; 71: 299–309. [PMC free article] [PubMed]
7. Hall CN, Reynell C, Gesslein B, et al. Capillary pericytes regulate cerebral blood flow in health and disease. Nature 2014; 508: 55–60. [PMC free article] [PubMed]
8. Winkler EA, Bell RD, Zlokovic BV. Central nervous system pericytes in health and disease. Nat Neurosci 2011; 14: 1398–1405. [PMC free article] [PubMed]
9. Wang S, Cao C, Chen Z, et al. Pericytes regulate vascular basement membrane remodeling and govern neutrophil extravasation during inflammation. PLoS ONE 2012; 7: e45499. [PMC free article] [PubMed]
10. Stark K, Eckart A, Haidari S, et al. Capillary and arteriolar pericytes attract innate leukocytes exiting through venules and ‘instruct’ them with pattern-recognition and motility programs. Nat Immunol 2013; 14: 41–51. [PubMed]
11. Nehme A, Edelman J. Dexamethasone inhibits high glucose-, TNF-alpha-, and IL-1beta-induced secretion of inflammatory and angiogenic mediators from retinal microvascular pericytes. Invest Ophthalmol Vis Sci 2008; 49: 2030–2038. [PubMed]
12. Tigges U, Boroujerdi A, Welser-Alves JV, et al. TNF-alpha promotes cerebral pericyte remodeling in vitro, via a switch from alpha1 to alpha2 integrins. J Neuroinflammation 2013; 10: 33. [PMC free article] [PubMed]
13. Hill J, Rom S, Ramirez SH, et al. Emerging roles of pericytes in the regulation of the neurovascular unit in health and disease. J Neuroimmune Pharmacol 2014; 9: 591–605. [PMC free article] [PubMed]
14. Sengillo JD, Winkler EA, Walker CT, et al. Deficiency in mural vascular cells coincides with blood-brain barrier disruption in Alzheimer’s disease. Brain Pathol 2012; 23: 303–310. [PMC free article] [PubMed]
15. Winkler EA, Sengillo JD, Sullivan JS, et al. Blood-spinal cord barrier breakdown and pericyte reductions in amyotrophic lateral sclerosis. Acta Neuropathol 2013; 125: 111–120. [PMC free article] [PubMed]
16. Vates GE, Takano T, Zlokovic B, et al. Pericyte constriction after stroke: the jury is still out. Nat Med 2010; 16: 959 author reply 960. [PubMed]
17. Rom S, Zuluaga-Ramirez V, Dykstra H, et al. Poly(ADP-ribose) polymerase-1 inhibition in brain endothelium protects the blood–brain barrier under physiologic and neuroinflammatory conditions. J Cerebr Blood Flow Metabol 2015; 35: 28–36. [PMC free article] [PubMed]
18. Vermeire J, Naessens E, Vanderstraeten H, et al. Quantification of reverse transcriptase activity by real-time PCR as a fast and accurate method for titration of HIV, lenti- and retroviral vectors. PLoS ONE 2012; 7: e50859. [PMC free article] [PubMed]
19. Guijarro-Munoz I, Cuesta AM, Alvarez-Cienfuegos A, et al. The axonal repellent Slit2 inhibits pericyte migration: potential implications in angiogenesis. Exp Cell Res 2012; 318: 371–378. [PubMed]
20. Rom S, Fan S, Reichenbach N, et al. Glycogen synthase kinase 3beta inhibition prevents monocyte migration across brain endothelial cells via Rac1-GTPase suppression and down-regulation of active integrin conformation. Am J Pathol 2012; 181: 1414–1425. [PubMed]
21. Krueger M, Bechmann I. CNS pericytes: concepts, misconceptions, and a way out. Glia 2010; 58: 1–10. [PubMed]
22. Niu F, Yao H, Zhang W, et al. Tat 101-mediated enhancement of brain pericyte migration involves platelet-derived growth factor subunit B homodimer: implications for human immunodeficiency virus-associated neurocognitive disorders. J Neurosci 2014; 34: 11812–11825. [PMC free article] [PubMed]
23. Daneman R, Zhou L, Kebede AA, et al. Pericytes are required for blood-brain barrier integrity during embryogenesis. Nature 2010; 468: 562–566. [PMC free article] [PubMed]
24. Ayres-Sander CE, Lauridsen H, Maier CL, et al. Transendothelial migration enables subsequent transmigration of neutrophils through underlying pericytes. PLoS ONE 2013; 8: e60025. [PMC free article] [PubMed]
25. Balabanov R, Beaumont T, Dore-Duffy P. Role of central nervous system microvascular pericytes in activation of antigen-primed splenic T-lymphocytes. J Neurosci Res 1999; 55: 578–587. [PubMed]
26. Verbeek MM, Westphal JR, Ruiter DJ, et al. T lymphocyte adhesion to human brain pericytes is mediated via very late antigen-4/vascular cell adhesion molecule-1 interactions. J Immunol 1995; 154: 5876–5884. [PubMed]
27. 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. [PMC free article] [PubMed]
28. Nakagawa S, Castro V, Toborek M. Infection of human pericytes by HIV-1 disrupts the integrity of the blood-brain barrier. J Cell Mol Med 2012; 16: 2950–2957. [PMC free article] [PubMed]
29. Dohgu S, Banks WA. Brain pericytes increase the lipopolysaccharide-enhanced transcytosis of HIV-1 free virus across the in vitro blood-brain barrier: evidence for cytokine-mediated pericyte-endothelial cell crosstalk. Fluids Barriers CNS 2013; 10: 23. [PMC free article] [PubMed]
30. Zhong Y, Zhang B, Eum SY, et al. HIV-1 Tat triggers nuclear localization of ZO-1 via Rho signaling and cAMP response element-binding protein activation. J Neurosci 2012; 32: 143–150. [PMC free article] [PubMed]
31. Murakami M. Signaling required for blood vessel maintenance: molecular basis and pathological manifestations. Int J Vasc Med 2012; 2012: 293641. [PMC free article] [PubMed]
32. Thanabalasundaram G, Schneidewind J, Pieper C, et al. The impact of pericytes on the blood-brain barrier integrity depends critically on the pericyte differentiation stage. Int J Biochem Cell Biol 2011; 43: 1284–1293. [PubMed]
33. Hill RA, Tong L, Yuan P, et al. Regional blood flow in the normal and ischemic brain is controlled by arteriolar smooth muscle cell contractility and not by capillary pericytes. Neuron 2015; 87: 95–110. [PMC free article] [PubMed]
34. Gaengel K, Genove G, Armulik A, et al. Endothelial-mural cell signaling in vascular development and angiogenesis. Arterioscler Thromb Vasc Biol 2009; 29: 630–638. [PubMed]
35. Quaegebeur A, Lange C, Carmeliet P. The neurovascular link in health and disease: molecular mechanisms and therapeutic implications. Neuron 2011; 71: 406–424. [PubMed]
36. Abraham S, Kogata N, Fassler R, et al. Integrin beta1 subunit controls mural cell adhesion, spreading, and blood vessel wall stability. Circ Res 2008; 102: 562–570. [PubMed]
37. Towgood KJ, Pitkanen M, Kulasegaram R, et al. Regional cerebral blood flow and FDG uptake in asymptomatic HIV-1 men. Hum Brain Mapp 2013; 34: 2484–2493. [PubMed]
38. Graham SM, Rajwans N, Tapia KA, et al. A prospective study of endothelial activation biomarkers, including plasma angiopoietin-1 and angiopoietin-2, in Kenyan women initiating antiretroviral therapy. BMC Infect Dis 2013; 13: 263. [PMC free article] [PubMed]
39. Fiedler U, Augustin HG. Angiopoietins: a link between angiogenesis and inflammation. Trends Immunol 2006; 27: 552–528. [PubMed]
40. Yao H, Bethel-Brown C, Niu F, et al. Yin and Yang of PDGF-mediated signaling pathway in the context of HIV infection and drug abuse. J Neuroimmune Pharmacol 2013; 9: 161–167. [PMC free article] [PubMed]

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of SAGE Publications