Brain and retinal macro- and microvasculature Li, Youhai
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2018
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Li, Y. (2018). Brain and retinal macro- and microvasculature: Response to ischemic and hyperglycemic stress. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
CHAPTER 3
A novel method to isolate retinal and brain microvessels
from individual rats: Microscopic and molecular biological
characterization and application in hyperglycemic animals
Youhai Lia, Natalia Lapinaa, Nina Weinzierla, Lisbeth Bondeb, Ebbe Boedtkjerb,
Rudolf Schubertc,d, Han Moshagee, Paulus Wohlfartf, Lothar Schillingc,d
aDivision of Neurosurgical Research, Medical Faculty Mannheim,
Heidelberg University, Germany
bDepartment of Biomedicine, Aarhus University, Denmark cCardiovascular Physiology, Center for Biomedicine and Medical Technology
(CBTM), Medical Faculty Mannheim, Heidelberg University, Germany
dEuropean Center of Angioscience (ECAS), Medical Faculty Mannheim,
Heidelberg University, Germany
eDepartment of Gastroenterology and Hepatology and Department of Laboratory
Medicine, University Medical Center Groningen, University of Groningen, The Netherlands
fSanofi Aventis Deutschland GmbH, TA, Diabetes R&D, Industriepark Hoechst,
Germany
Vascular Pharmacology. 2018 Jul 9. pii: S1537-1891(18)30086-7
CHAPTER 3
III
56
Abstract
Alterations in the retinal microvessel (RMV) compartment occurring in systemic disease states such as diabetes may eventually contribute to blindness. To specifically address the pathophysiological role of the microvasculature we developed a new method for RMV bulk isolation from individual rats.
The extraction procedure performed in the cold throughout takes less than one hour. Slight modifications enable isolation of brain microvessels (BMVs) for comparison. Microscopically, RMVs and BMVs consisted mainly of capillaries of good structural integrity. The endothelial cell/pericyte ratio was approximately 1.8 in RMVs and 2.7 in BMVs, well in agreement with data from intact vascular beds. Total RNA extracted from individual rats amounted to approximately 7 ng in RMVs, 50 ng in BMVs, and 155 ng in pial arteries (which were also isolated) with highly preserved integrity throughout. Measurements using microfluidic card methodology revealed segregation of RMVs, BMVs, and pial arteries in distinct clusters based on principal component analysis. In all three vascular compartments endothelial cell-specific markers were significantly enriched. Similarly, pericyte-specific markers displayed accumulation in RMVs, BMVs, and pial arteries, the latter probably reflecting the common ontogenetic origin of pericytes and smooth muscle cells. Isolation of RMVs, BMVs, and pial arteries from rats suffering from 8-weeks hyperglycemia yielded expression patterns of endothelial cell- and pericyte-specific marker genes largely comparable to those obtained in control rats.
Our newly developed protocols allow for selective studies of RMVs from individual rats to characterize reactive pathways, in comparison with the ontogenetically closely related BMVs. Moreover, our protocols with inclusion of pial arteries enable comparative studies of the macro- and microvasculature from the same organ.
III
57
Introduction
Blood vessels differ greatly between organs and even within the same organ according to the position in the vascular bed. Features reflecting special adaptation are evident on the structural, functional, and eventually also on the gene expression level. Prominent examples of specialized microvessels are those of the retina and the brain, which by expressing complex tight junctions between the endothelial cells form complex barriers called the blood-retina barrier (BRB) and the blood-brain barrier (BBB), respectively. The similarities between in the retinal microvessels (RMVs) and brain microvessels (BMVs) are probably related to the ontogenetic relationship including separation of the retina from the diencephalon at an early stage of fetal development and the common vascular supply of the retina and the forebrain from the carotid arteries (1, 2). Despite these similarities, there are distinct differences between RMVs and BMVs including their response to systemic diseases such as diabetes mellitus. Long-lasting and insufficiently treated diabetes results in severe complications in the retina, summarized as diabetic retinopathy, eventually leading to blindness. Comparably severe dysfunction has not yet been reported in the brain, although in a rat model of long-lasting type 1 diabetes, multiple cerebral lesions involving alterations in the cerebrovasculature have been described (3). Similarly, clinical studies suggest a linkage between diabetes and the occurrence of dementia including Alzheimer’s disease, which has been related to cerebrovascular dysfunction and regional hypoperfusion (4).
Diabetic retinopathy has been associated with pronounced alterations in the microvascular compartment (5, 6), but the exact pathophysiological role of vascular dysfunction for the initiation of diabetic retinopathy has yet to be conclusively defined. Against this background, reproducible, fast, and gentle methods for isolation of MVs either from patients or from established animal models are desirable and promising approaches to address the specific microvascular contribution to diabetic complications. Many studies in the field of diabetes are conducted in rats, and injection of streptozotocin is widely used to model type 1 diabetes. However, only a very limited number of protocols for isolation and purification of the retina microvasculature have been described so far with subsequent studies of gene expression not available at all. Thus, we have developed a new method of rat RMVs isolation. The main features of this protocol are that (i) only two eyes are needed making pooling of animals redundant,
III
58
and (ii) the whole isolation procedure can be performed in the cold (≤ 4°C) to ensure maintenance of the structural features allowing extraction of the transcriptome as complete as possible. This protocol was also applied for isolation of BMVs for comparison. Using normal rats, we demonstrated the purity of the microvessel compartments and the suitability for selective vascular gene expression studies by comparing RMVs with full retina tissue and BMVs with full brain tissue. Moreover, we included extra-parenchymal, so-called pial arteries (PAs) to outline the distinction between vessels taken from different positions within the same organ. Finally, the newly developed protocol was applied to hyperglycemic rats in order to confirm its appropriateness for studying the microvasculature of diabetic rats.
Material and Methods
Animals and animal experiments
Prior to starting the experiments were approved by the appropriate authorities (for experiments with untreated control animals: Regierungspraesidium Karlsruhe, for experiment with injection of streptozotocin: Danish Animal Experiments Inspectorate). All experiments were performed in compliance with the relevant laws and institutional guidelines for the care and use of animals in research according to the Directive 2010/63/EU, and reporting conforms to the ARRIVE guidelines (7).
Male Wistar rats were used throughout. The animals used in control experiments were obtained from Janvier Labs (Le Genest-Saint-Isle, France) and allowed acclimatization in the animal facility (at the Medical Faculty Mannheim, Heidelberg University) with free access to food and tap water for at least one week before entering the experiment. The rats made hyperglycemic were 5-6 weeks old when obtained from Taconic Biosciences (Ry, Denmark). After acclimatization for one week in the animal facility at the Department of Biomedicine, Aarhus University they were injected intraperitoneally with a single 60 mg/kg dose of streptozotocin (Sigma-Aldrich, Copenhagen, Denmark) freshly dissolved in cold citrate buffer at pH 4.5. Glucose
III
59
concentration was measured in a drop of blood taken from the tip of the tail with an Accu-Chek Aviva blood glucose meter (Roche Diagnostics, Hvidovre, Denmark). Already 24 hours after the injection and persisting for the subsequent 8 weeks of observation, the average non-fasting blood glucose levels were above 20 mM. During the time in the animal house the rats had free access to food and tap water. Further characteristics of the investigated population of streptozotocin-injected rats have been reported elsewhere (8).
In order to obtain the tissues, the animals were deeply anesthetized with CO2 inhalation
(control rats) or by an intraperitoneal injection of 50 mg/kg pentobarbital (streptozotocin-injected rats) followed by de-capitation. Both eyes were quickly removed and snap frozen in liquid nitrogen. The brain was removed and transferred into ice-cold physiological saline solution for isolation of the PAs making up the Circle of Willis (i.e. middle cerebral artery, anterior cerebral artery, and posterior cerebral artery) and their daughter vessels. Thereafter, the meninges and the choroid plexus were removed. All tissues were snap frozen and maintained at –80°C until use.
Microvessel isolation from retina and brain
For RMVs isolation, the eyes from individual rats were thawed for retina removal. Both retinae were pooled and frozen in phosphate-buffered saline (PBS, for composition see the supplement). A single 50 μm thick cryosection was taken for total RNA extraction from full retina tissue while the remaining tissue block was sectioned at 200 μm thickness (HBM500, Microm, Nussloch, Germany). All steps of the subsequent isolation protocol were performed in the cold (≤ 4°C) throughout. The cryosections were immersed in 3 mL PBS containing 1% dextran (Dextran 70,000, Roth, Karlsruhe, Germany) and mechanically homogenized using a motor-driven homogenizer (Homgen plus, Schuett Biotec, Goettingen, Germany; 60 rpm, 20 upstrokes). The homogenate was transferred onto a density gradient column with PBS containing 31% dextran in the lower phase and 18% dextran in the upper phase. After centrifugation (1300 g, 15 min) the interphase was carefully collected and gently mixed with 20 mL PBS. Finally, the RMVs were captured by filtering through a 60-μm nylon mesh. For BMV isolation, individual hemispheres were serially cut (100 μm thickness). Three individual sections each one taken from the frontal, medial, and occipital part were
III
60
removed and pooled for subsequent total RNA extraction. The remaining sections were pooled for microvessel extraction using the protocol described above with some modifications (for details see the supplement).
Microscopic control of MV purity and stereological measurements
Smears of RMV and BMV extracts obtained from control animals were air-dried on glass slides followed by hematoxylin and eosin staining. The smears were microscopically checked for the presence / lack of tissue elements and tissue remnants adhering to the MVs. Furthermore, in smears from 5 different animals the endothelial cells and pericytes were identified by the shape and the orientation as well as the position of the cell nuclei in the vessel wall and the numbers counted in arbitrarily chosen areas. Counting was done by two observers independently from each other, and the endothelial cell/pericyte ratios were calculated. In addition, stereological
III
61
measurements were performed to determine the distances between neighboring cell nuclei as a measure of cell density using an image analyzing system (CUE-2, Olympus Opticals, Hamburg, Germany).
Gene expression studies
Total RNA was isolated from tissues by a standard RNeasy Plus Micro Kit (QIAGEN). Tissues were homogenized in lysis buffer and passed 10 times through a 30G needle attached to a 1 mL syringe in order to shear the DNA. Further purification of RNA was performed according to the protocol provided by the manufacturer and further detailed in the supplement file. The content and the quality of the total RNA were regularly controlled using either the RNA 6000 Pico kit (RMVs and BMVs, PAs), or Nano kit (retina tissue and brain tissue) (both from Agilent, Waldbronn, Germany). These kits provide a quantitative measure of RNA integrity (expressed as RNA integrity number, RIN), and only samples with a RIN value > 8 were used for reverse transcription real-time semi-quantitative PCR (RT-qPCR) experiments throughout the study.
In order to improve comparability between the samples, the differences in the total RNA contents were adjusted by appropriate dilution with molecular biology grade water to establish concentrations that did not differ more than tenfold among the samples. Within this concentration range, the Ct values are independent of the starting amount of total RNA as shown in preliminary RT-qPCR experiments using tissue samples from two individual rats. After reverse transcription, a pre-amplification step of 14 cycles using a target-specific primer pool was performed followed by RT-qPCR using TaqMan microfluidic card technology (ThermoFisher, Darmstadt, Germany) with a maximum of 40 cycles (further details are given in the supplement).
A matrix of 24 gene expression assays was used. These included 4 sets of primers/probes as reference genes and 17 sets as markers of endothelial cells, pericytes, smooth muscle cells, astrocytes, neurons, and photoreceptors (for the complete list of these assays refer to the supplement section). In addition, 3 assays not representing cell type specific markers were also included, and these assays did not display any difference in the expression levels between full parenchyma and vascular tissue samples.
III
62
Data analysis and StatisticsIn the RT-qPCR experiments, the threshold quantification cycles, Ct values, were
obtained from the manufacturer’s software (Viia7, Thermofisher Scientific, Darmstadt, Germany). These values were analyzed using the ArrayStudio software package (Version 9, Omicsoft Corporation, Research Triangle Park, NC, USA). The levels of gene expression were obtained by normalizing the individual Ct values to the average
reference gene value. These data were clustered and subjected to outlier analysis by principal component analysis (9). Subsequently, one-way ANOVA was performed to determine the genes differentially expressed between the different tissue samples. To adjust for multiple testing, a Benjamini-Hochberg false discovery rate correction was applied (10). Data are given as mean ± SD or in case of the degree of enrichment / decrease of marker genes in MVs as mean value along with the 95% confidence interval. A p value < .05 was considered to indicate significant differences.
Results
Morphometric studies on isolated MVs from control animals
Smears of RMVs and BMVs isolated from 5 individual rats underwent hematoxylin and eosin staining for subsequent histological analysis. These smears did not display signs of gross contamination with non-vascular tissue elements when scrutinized under the microscope. High magnification light microscopy revealed that the vast majority of RMVs and BMVs were capillaries (Fig. 2). Endothelial cells and pericytes could easily be identified by the shape and the orientation as well as the position of the cell nuclei in the vessel wall in accordance to previous descriptions (11). Based on these morphological features, the numbers of endothelial cells and pericytes were counted by two observers (YL, LS) independently from each other in BMV and RMV segments taken from randomly chosen microscopic fields. Calculating the endothelium/pericyte ratio revealed an average value of 1.82 in RMVs (48 microscopic fields from 5 animals) and 2.67 in BMVs (56 microscopic fields from 5 animals). Correlation analysis to check
III
63
for the agreement between both observers yielded a correlation factor r = 0.415 in RMVs (p< .01) and r = 0.389 (p< .01) in BMVs. Additional more sophisticated stereological measurements revealed a shorter distance in RMV compared to BMV segments between neighboring endothelial cells (RMVs: 39.1±0.9 μm, BMVs: 46.5±1.4 μm; p< .001) and pericytes (RMVs: 44.8±1.0. BMVs: 61.2±3.6 μm; p< .001) suggesting a higher cell density in the RMVs wall over BMVs. Altogether, the data strongly indicate that the isolation procedure yielded structurally well preserved RMV and BMV segments, respectively.
Figure 2. Microscopic images demonstrating the purity of rat retina microvessel extracts (panel A) and rat brain microvessel extracts (panel B) at low and high power magnification. Calibration bars represent 50 μm.
A
III
64
Characterization of the total RNA extracted from the different types of tissuesamples
Microscopically, both RMVs and BMVs were assessed as notably pure extracts. We therefore tried to define purity at the level of gene expression as well. To this end, we first checked the suitability of the total RNA extracted from each tissue sample obtained from 8 individual control rats by measuring concentration and integrity using chip-based methodology. The average RIN values were above nine throughout all tissue types and samples indicating well-preserved RNA integrity even in the isolated RMV and BMV extracts. Moreover, for all types of tissue samples the total RNA content and the RIN values did not differ between normo- and hyperglycemic animals (Tab. S2).
Principal component analysis reveals tissue distribution in distinct clusters
The gene expression data obtained from 6 rats were initially checked by principal component analysis taking into account the endothelial cell and pericyte marker genes in order to specifically address the enrichment in RMVs and BMVs. The results shown in figure 3 indicate (i) close proximity of RMVs and BMVs on the one hand and of full tissue samples from retina and brain on the other hand, (ii) that PAs cluster distinctly from the RMV and BMV samples, and (iii) a remarkable degree of data homogeneity for each tissue type studied. Homogeneity is reflected by the fact that the first and second component account for more than 90% of the overall data variation, and it is even more vividly demonstrated by the narrow clustering of the sample types (Fig. 3).
III
65
Figure 3. Principal component analysis of the expression of endothelial cell and pericyte marker genes in different types of tissue obtained from non-diabetic rats (NG, n=6) and 8-weeks diabetic rats (HG, n=4). Both retina and brain microvessels (RMVs, BMVs) form narrow clusters which although lying next to each other, can still be distinguished. In contrast, pial artery (PA) samples are located far away from the BMVs and RMVs. Similar to the RMVs and BMVs, retina tissue samples (Retina) and brain tissue samples (Brain) cluster next to each other yet still being distinguishable. The analysis also shows that the expression of the endothelial cell and pericyte marker genes used in the present study was virtually unaffected by 8 weeks of hyperglycemia.
Purity of the MV and PA compartments is reflected by enrichment of vessel wall-specific and decrease of parenchymal cell marker genes
The set of assays designed for expression analysis contained initially four genes supposed to be suitable reference genes. Among these, the absolute Ct values of
β2-microglobulin (B2M), β-actin (ACTB) and 18S rRNA were very similar throughout all tissue types from individual animals whereas the assay used for detection of elongation factor-1 (EEF1A2) revealed largely different Ct values in the vascular tissue samples
compared with whole retinal tissue and brain tissue. Therefore, this assay was excluded from the analysis leaving three reference genes available for normalization. We employed the mean values of these 3 assays to calculate the ΔCt values for subsequent
III
66
Reasonably, one may expect (i) enrichment of endothelial cell and pericyte markers in the microvascular compartments, (ii) enrichment of endothelial and smooth muscle cell markers in the PAs, and (iii) a decrease of parenchymal cell markers in each vascular compartment versus the respective full tissue samples. In fact, the data obtained in control animals revealed significantly higher expression values for all endothelium- and pericyte-specific genes in the RMV compartment over full tissue retina extracts, and largely comparable results were obtained for BMVs when compared to brain tissue (Fig. 4). These data are taken to indicate microvessel purity on the level of gene expression, fully concurrent with the microscopic results.
In PAs, smooth muscle cell markers were highly enriched compared to BMVs (calponin-1 [CNN1]: >10,000 fold; myosin heavy chain 11 [MYH11]: >1,000 fold) while the endothelial markers were less highly expressed than in BMVs (Fig. 4). Quite unexpectedly, we also found the 3 pericyte markers present in the PA wall with chondroitin sulfate proteoglycan 4 (CSPG4) at an even slightly higher level than in BMVs (Fig. 4). The values of fold enrichment / decrease are listed in table S4.
In contrast to the enrichment of the vessel wall-specific marker genes, nearly all the markers of neuronal cells, glial cells and photoreceptors were heavily decreased in the RMVs compared to retina tissue (Tab. S3). Similarly, in BMVs, glial and neuronal markers were significantly decreased compared to brain tissue, and photoreceptor markers were not detectable at all as it was also the case in brain tissue.
III
67
Figure 4. Gene expression studies to show the quality of retina microvessel (RMV) and brain microvessel (BMV) segments as well as pial arteries (PA) obtained from control rats (n=6). Expression levels are given as 2-∆Ct relating the Ct value of each gene of interest to the average of three reference genes, β2-microglobulin (B2M), β-actin (ACTB), and 18S rRNA. Depicted are mean ± SD.
*p < .05 vs. brain tissue; # p < .05 vs. retina tissue based on one-way ANOVA analysis and subsequent Benjamini-Hochberg false discovery rate correction adjustment for multiple testing. NOS3, endothelial NO synthase; TEK, receptor tyrosin kinase (Tie2); VWF, von Willebrand factor; SLC2A1, solute carrier family 2, member 1 (Glut1); CLDN5, Claudin 5; ACE, angiotensin-converting enzyme; PDGFRB, Platelet-derived growth factor receptor beta; RGS5, regulator of G-protein signaling 5; CSPG4, chondroitin sulfate proteoglycan 4.
III
68
Successful MV and PA isolation in rats exposed to eight weeks hyperglycemia:evaluation based on the marker gene expression pattern
Having established the isolation protocols, we tested whether these would also work in disease conditions. For this purpose, we used hyperglycemic rats 8 weeks after injection of streptozotocin. Microscopically, the RMV and BMV segments appeared as pure as those obtained from control rats, and the content and integrity of the total RNA extracted were fully consistent with the results from control rats. Principal component analysis indicated that the hyperglycemic samples for each tissue type studied clustered closely with the respective control samples (Fig. 2). Furthermore, we found largely comparable levels of enrichment of vessel wall marker genes in RMVs over retinal tissue and in BMVs compared to brain tissue in hyperglycemic and control rats. The results expressed as fold enrichment are shown in table 1.
We also studied the endothelial and pericyte marker gene expression in PAs from hyperglycemic rats using the BMVs for reference (Tab. S4). Some of the endothelial cell markers, most notably NOS3, TEK, and VWF appeared to show some upregulation in hyperglycemia, however, the current data set is not strong enough to allow vigorous statistical analysis.
III
69
Table 1. Fold enrichment of endothelial cell and pericyte marker genes in retinal microvessels (RMVs) and brain microvessels (BMVs) over the respective full tissues in control rats (n=6) and 8-weeks hyperglycemic animals (n=4).
RMVs BMVs
Control Hyperglycemia
Control Hyperglycemia
Endothelial cell maker genes NOS3 19.7 (10.4-37.1) 14.7 (12.3-17.5) 9.5 (6.6-13.5) 11.2 (10.0-12.6) TEK 21.0 (10.7-41.2) (12.7-19.7)15.8 (9.2-37.8)18.7 (13.2-23.7)17.7 VWF 25.8 (19.1-35.0) 25.4 (21.4-30.3) 22.7 (12.3-41.8) 22.6 (14.2-36) SLC2A1 5.0 (3.8-6.6) 5.3 (3.9-7.2) 9.8 (6.5-14.7) 10.8 (8.4-14) CLDN5 64.9 (31.4-134) (38.7-99.7)62.1 (5.0-36.6)13.5 (8.5-15)11.3 ACE 3.4 (1.0-11.4) 1.3 (1.0-1.8) 6.1 (3.4-11.2) 3.5 (2.4-5.1) Pericyte make r genes PDGFRB 15.1 (11.1-20.5) 16.3 (14.2-18.6) 11.3 (6.4-19.8) 10.8 (9.2-12.7) RGS5 15.0 (5.8-39.2) 20.6 (16.2-26.2) 24.6 (12.5-48.5) 23.5 (17.4-31.6) CSPG4 20.9 (10.8-40.2) 13.2 (9.6-18.1) 5.0 (2.7-9.5) 5.5 (3.5-8.7)
Neither of the marker genes studied displayed any significant change in hyperglycemia compared to the normoglycemic control condition. Data are indicated as mean values along with the 95% confidence intervals. NOS3, endothelial NO synthase; TEK, receptor tyrosine kinase (Tie2); VWF, von Willebrand factor; SLC2A1, solute carrier family 2, member 1 (Glut1); CLDN5, Claudin 5; ACE, angiotensin 1 converting enzyme; PDGFRB, Platelet-derived growth factor receptor beta; RGS5, regulator of G-protein signaling 5; CSPG4, chondroitin sulfate proteoglycan 4 (NG2).
III
70
Discussion
The present study introduces a new method of rat RMV extraction. The isolation protocol consists of mechanical steps exclusively, is performed in the cold throughout and can be conducted in clearly < 1 h. It yields a considerable amount of RMV segments, almost exclusively capillaries with high purity and an excellent structural integrity as shown by morphometric measurements. Moreover, these RMVs come along with a minimal degree of RNA degradation making them a highly valid tool for vessel wall-specific gene expression studies. Applying the extraction protocol to individual brain hemispheres yielded comparably pure BMVs thus allowing full comparison of both vascular beds. The microscopic appearance and quality of total RNA were not affected after 8 weeks of hyperglycemia demonstrating the suitability of our extraction method in pathological conditions. Employing microfluidic card methodology, we could demonstrate a significant accumulation of endothelial and pericyte marker genes in the RMV and BMV extracts along with a significant decrease in parenchymal (including glial, neuronal, and in the case of RMVs photoreceptor) marker genes. After 8 weeks of hyperglycemia, the patterns of marker gene accumulation were not grossly altered in RMVs and BMVs.
In previous studies, RMVs have often been obtained from porcine or bovine eyes that provide a relatively high amount of starting tissue. These MV extracts have successfully been used for structural and functional investigations, but the availability of tools for systematic studies at the level of molecular biology, for instance gene expression, is somewhat limited for these species. These limitations do not apply to small laboratory animals such as rats, although use of these animals comes at the expense of a very small amount of starting tissue. Moreover, established disease models, including models of diabetes, are abundant and readily available in rodents. Previously, eyes from several animals have been pooled to compensate for the limited amount of tissue (12, 13). Yet, this runs counter to the 3R guiding principles on the use of animals in research (14) thus calling for better protocols of RMV.
With the development of the new method described in the present study, the purity of all RMV and BMV extracts was screened microscopically to ensure the lack of contamination with parenchymal cells or tissue remnants. The structural integrity of the
III
71
microvessels was checked by morphometric analyses including (i) the ratio of endothelial cells and pericytes in the vascular wall and (ii) the distance between individual cells. Cell nuclei in the segments were identified by their appearance and spatial arrangement with the endothelial cell nuclei typically displaying an oval shape and orientation in the long axis of the vessels. In contrast, the nuclei of pericytes typically show a round shape and they often appear attached to the outside of the vessel wall, in particular at bifurcation sites (11). Counting by two observers independently from each other revealed an endothelial / pericyte ratio of 1.8 in the RMVs, well within the range of 1.3 – 3.5 obtained from measurements of intact vascular beds in mice (15, 16), rats (17, 18) and rhesus monkeys (19). We obtained a somewhat higher ratio of 2.7 in BMVs, which is also in agreement with published data (20). Since we obtained RMVs and BMVs from the same animals with the same protocol, species-related impacts potentially affecting the endothelial cell/pericyte ratio published by different groups (see above) can be ruled out. Thus, our results clearly support notions on a higher relative number of pericytes in RMVs over BMVs. Moreover, we found that the distance between individual pericytes was significantly shorter in RMVs than in BMVs. This is in agreement with data in the literature showing a higher degree of pericyte coverage in RMVs compared to BMVs in rats (21) and monkeys (22). Taken together, the morphometric data strongly suggest that the structural integrity of the RMVs and BMVs is well preserved following our new isolation protocol.
In addition to microscopic inspection, we also aimed at characterizing the isolated microvessels on the gene expression level. Previously used methods of RMV isolation typically include enzymatic digestion and/or prolonged immersion in distilled water at room temperature or even 37°C (11, 23). However, such protocols may carry a high risk of unpredictable degradation of mRNA eventually questioning meaningful studies of gene expression. Therefore, our new method was designed to be performed in the cold throughout. With this approach, there was virtually no RNA degradation discernable. In fact, employing chip-based electrophoretic measurements we obtained RIN values well above 8 throughout all individual samples qualifying them as optimal templates for RT-qPCR studies (24).
Due to the varying amounts of total RNA obtained from the different tissue types, appropriate dilution steps were performed to adjust all samples to the lowest average concentration level, which was found in RMVs. In order to account for the very low
III
72
amounts of total RNA, the sensitivity of the RT-qPCR analysis was enhanced by a pre-amplification step as used successfully in previous studies (9, 25). In preliminary experiments, we checked the effectiveness of the pre-amplification step, and we also confirmed the linearity of the assays used in the microfluidic card measurements by measuring the Ct values over a 4-fold log10 concentration range of total RNA (Fig. S1).
These experiments also showed that the ΔCt values were virtually stable within a
tenfold variation range in the starting amount of total RNA.
In order to characterize the RMV and BMV extracts on the gene expression level we used a panel of cell-specific markers. The endothelial cell markers, which were deliberately chosen to contain proteins with supposedly barrier-related (claudin 5 [CLDN5], glucose transporter 1 [SLC2A1]) and barrier-unrelated (endothelial NO synthase [NOS3], receptor tyrosin kinase [TEK], von Willebrand factor [VWF], and angiotensin I converting enzyme [ACE]) functions as well as pericyte markers, showed largely comparable expression levels displayed as 2-ΔCt values in RMVs and BMVs yet
significantly higher expression levels than in retina and brain tissue. These data indicate pronounced accumulation (when displayed as 2-ΔΔCt values) of these marker genes in
RMVs and BMVs with up to roughly 60-fold enrichment over full tissue samples thus placing further emphasis on their purity. Interestingly, the level of enrichment in BMVs obtained for SLC2A1 and CSPG4 nicely compare with data published by Yousif and coworkers (26), thus providing further support for the quality of the extraction method described here.
Accumulation of endothelial cell and pericyte marker genes in RMVs and BMVs went along with a marked decrease of parenchymal marker genes. With respect to neuronal markers, this decrease was most prominent for synapsin 1 (SYN1) and RNA binding fox-1 homolog 3 (RBFOX3; synonym, neuronal nuclear antigen, NeuN), both of which were present in RMVs and BMVs in trace amounts at best. The photoreceptor marker genes S-antigen visual arrestin (SAG) and retinol binding protein 3 (RBP3) were not at all detectable in brain tissue and BMVs, which is well comprehensible. In RMVs, expression of both mRNA species was significantly decreased compared to full retina tissue, again underscoring the quality of the extraction protocol. When looking at the glial astrocytic marker GFAP, there was a remarkably low degree of depletion in the RMVs and BMVs. This might be related to the presence of astrocytic end-feet that make intimate contact with the capillaries in the eye and brain as demonstrated in
III
73
studies using electron microscopy (6, 27). This intimate contact may hamper complete disruption and separation from the capillary wall leaving residues of glial end-feet attached to the microvessels at the end of the isolation procedure. These structures must be considered a potential source of GFAP mRNA based upon in situ hybridization studies in cultured glial astrocytic cells (28). Thus, the moderate degree of GFAP depletion in RMVs and BMVs does not really challenge the RMV and BMV purity, all the more so since AQP4, another putative marker of glial astrocytic cells, displayed a level of depletion largely comparable with neuronal and photoreceptor markers. In addition to RMVs and BMVs, we also analyzed PAs. As expected, we found high levels of smooth muscle cell markers here exemplified by calponin 1 (CNN1) and myosin heavy chain 11 (MYH11). Furthermore, all endothelial markers were detectable, however, at lower levels than in BMVs. The most pronounced differences were obtained for the glucose transporter SLC2A1 and the tight junction-associated molecule CLDN5. Even for NOS3, we observed a considerably lower level of expression. This is likely related to the relative number of endothelial cells, which might be around 75% in BMVs as deduced from the ratio of endothelial cells to pericytes. In the arterial wall, however, there is a high density of smooth muscle cells in the media (approximately 3 layers in the main trunk of the MCA) and a substantial amount of cells in the adventitial layer (own observations). Thus, the content of NOS3 mRNA per cell is probably much higher in arterial than in capillary endothelial cells, well in agreement with the pivotal role of NOS3 in the regulation of arterial tone. A somewhat unexpected finding was the presence of all 3 putative pericyte marker genes in the PA wall. The reason for this overlap is not yet clear, however, it might be related to the close ontogenetic relationship between pericytes and smooth muscle cells. In fact, quail-chick transplantation experiments have provided evidence that at least a subtype of brain pericytes and arteriolar smooth muscle cells share a common lineage (2, 29).
Alterations or even failure of the retina microvasculature are typical accessory symptoms in systemic disease states such as hypertension and diabetes (30) eventually resulting in blindness. Therefore, we were interested in the applicability of our newly developed extraction protocol in pathological conditions. We chose hyperglycemia induced by injection of streptozotocin, a widely used model of type 1 diabetes. Microscopically, the appearance of the RMV fragments did not differ from those extracted from control rats, and the same was true for the BMVs. Furthermore, total
III
74
RNA content and integrity were not different from controls. Principal component analysis revealed largely comparable levels of marker genes in the RMV and similarly in BMV extracts obtained from 8-weeks hyperglycemic rats. This result must not be taken to indicate that the microvessel compartment is inert to hyperglycemia since we have focused on a limited number of marker genes to characterize purity. These marker genes may not be among the first genes undergoing significant changes of expression in hyperglycemia. In any case, the data indicate that our newly developed method can be successfully applied to studies of pathological conditions. Current studies using methods of whole transcriptome analysis with strictly defined control groups are underway to determine the effects of long-term hyperglycemia in the retina and brain microvasculature.
Our novel protocol of RMV and BMV isolation may also be a worthwhile starting point for cell-type specific follow-up studies. Thus, the microvessel isolation procedure might be followed by enzymatic digestion and subsequent cell sorting. Ideally, this might result in a distinct population of endothelial cells (or pericytes, if desired) of truly retinal and brain microvascular origin. The suitability of these cell populations for subsequent studies would have to be determined first, and this may well be achieved by comparing gene expression patterns with the data provided for intact microvessels in the present study.
Funding sources
This work was supported by a grant from the German Research Foundation (DFG) [grant number IRTG 1874-1 and -2 to RS] and a grant from the Danish Heart Foundation [grant number 14-R97_A5321-22809 to EB]. YL is a graduate student associated to the International Research Training School (IRTG) 1874 (“DiaMiCom”) and received a scholarship under the CIMDS program from the Medical Faculty Mannheim, Heidelberg University. NL is a former graduate student of the IRTG 880 (“Vascular Medicine”) funded by the DFG [grant number GRK880-2/3].
III
75
Acknowledgements
The authors would like to thank Jørgen Andresen for technical assistance with isolation of blood vessels from the diabetic rats. We thank E. Deckert for expert help in cDNA generation and microfluidic card measurements. We also thank Prof. H.-P. Hammes for providing support in the use of the microscopic system for stereological measurements.
Conflicts of interests
Paulus Wohlfart is a full research employee of Sanofi Aventis Deutschland GmbH and supported by this company for pre-clinical research on new approaches in Diabetes. Support by Sanofi-Aventis Deutschland GmbH to Paulus Wohlfart does not alter adherence to international publication policies on sharing data and materials. All other authors declare no conflict of interests.
References
1. Chow RL & Lang RA (2001) Early eye development in vertebrates. Annual review of cell and developmental biology 17:255-296.
2. Etchevers HC, Vincent C, Le Douarin NM, & Couly GF (2001) The cephalic neural crest provides pericytes and smooth muscle cells to all blood vessels of the face and forebrain. Development (Cambridge, England) 128(7):1059-1068.
3. Mukai N, Hori S, & Pomeroy M (1980) Cerebral lesions in rats with streptozotocin-induced diabetes. Acta Neuropathol. 51(1):79-84.
4. Hoscheidt SM, et al. (2017) Insulin resistance is associated with lower arterial blood flow and reduced cortical perfusion in cognitively asymptomatic middle-aged adults. J. Cereb. Blood Flow Metab. 37(6):2249-2261.
5. Hammes HP, Feng Y, Pfister F, & Brownlee M (2011) Diabetic retinopathy: targeting vasoregression. Diabetes 60(1):9-16.
III
76
6. Klaassen I, Van Noorden CJ, & Schlingemann RO (2013) Molecular basis of the innerblood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions. Progress in retinal and eye research 34:19-48.
7. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, & Altman DG (2010) Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS biology 8(6):e1000412.
8. Bonde L, Shokouh P, Jeppesen PB, & Boedtkjer E (2017) Crosstalk between cardiomyocyte-rich perivascular tissue and coronary arteries is reduced in the Zucker Diabetic Fatty rat model of type 2 diabetes mellitus. Acta physiologica (Oxford) 219(1):227-238.
9. Wohlfart P, et al. (2013) Cardioprotective effects of lixisenatide in rat myocardial ischemia-reperfusion injury studies. J. Transl. Med. 11:84.
10. Benjamini Y & Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Statist. Soc. B 57(1):289-300.
11. Dagher Z, et al. (2004) Studies of rat and human retinas predict a role for the polyol pathway in human diabetic retinopathy. Diabetes 53(9):2404-2411.
12. Greenwood J (1992) Characterization of a rat retinal endothelial cell culture and the expression of P-glycoprotein in brain and retinal endothelium in vitro. J. Neuroimmunol. 39(1-2):123-132.
13. Badr GA, Tang J, Ismail-Beigi F, & Kern TS (2000) Diabetes downregulates GLUT1 expression in the retina and its microvessels but not in the cerebral cortex or its microvessels. Diabetes 49(6):1016-1021.
14. Flecknell P (2002) Replacement, reduction and refinement. Altex - Altern. Anim. Exp. 19(2):73-78.
15. Cuthbertson RA & Mandel TE (1986) Anatomy of the mouse retina. Endothelial cell-pericyte ratio and capillary distribution. Invest. Ophthalmol. Vis. Sci. 27(11):1659-1664.
16. Agardh CD, Agardh E, Hultberg B, & Ahren B (2000) Long-standing hyperglycemia in C57BL/6J mice does not affect retinal glutathione levels or endothelial/pericyte ratio in retinal capillaries. J. Diabetes Complications 14(3):146-153.
17. Agardh CD, Agardh E, Zhang H, & Ostenson CG (1997) Altered endothelial/pericyte ratio in Goto-Kakizaki rat retina. J. Diabetes Complications 11(3):158-162.
18. Luo D, Fan Y, & Xu X (2012) The effects of aminoguanidine on retinopathy in STZ-induced diabetic rats. Bioorg. Med. Chem. Lett. 22(13):4386-4390.
19. Kim SY, et al. (2004) Retinopathy in monkeys with spontaneous type 2 diabetes. Invest. Ophthalmol. Vis. Sci. 45(12):4543-4553.
III
77
20. Prakash R, et al. (2012) Enhanced cerebral but not peripheral angiogenesis in the Goto-Kakizaki model of type 2 diabetes involves VEGF and peroxynitrite signaling. Diabetes 61(6):1533-1542.
21. Frank RN, Dutta S, & Mancini MA (1987) Pericyte coverage is greater in the retinal than in the cerebral capillaries of the rat. Invest. Ophthalmol. Vis. Sci. 28(7):1086-1091. 22. Frank RN, Turczyn TJ, & Das A (1990) Pericyte coverage of retinal and cerebral
capillaries. Invest. Ophthalmol. Vis. Sci. 31(6):999-1007.
23. Podesta F, et al. (2000) Bax is increased in the retina of diabetic subjects and is associated with pericyte apoptosis in vivo and in vitro. Am. J. Pathol. 156(3):1025-1032.
24. Fleige S, et al. (2006) Comparison of relative mRNA quantification models and the impact of RNA integrity in quantitative real-time RT-PCR. Biotechnol. Lett. 28(19):1601-1613.
25. Wohlfart P, et al. (2014) Expression patterning reveals retinal inflammation as a minor factor in experimental retinopathy of ZDF rats. Acta Diabetol. 51(4):553-558.
26. Yousif S, Marie-Claire C, Roux F, Scherrmann JM, & Decleves X (2007) Expression of drug transporters at the blood-brain barrier using an optimized isolated rat brain microvessel strategy. Brain Res. 1134(1):1-11.
27. Kacem K, Lacombe P, Seylaz J, & Bonvento G (1998) Structural organization of the perivascular astrocyte endfeet and their relationship with the endothelial glucose transporter: a confocal microscopy study. Glia 23(1):1-10.
28. Medrano S & Steward O (2001) Differential mRNA localization in astroglial cells in culture. J. Comp. Neurol. 430(1):56-71.
29. Korn J, Christ B, & Kurz H (2002) Neuroectodermal origin of brain pericytes and vascular smooth muscle cells. J. Comp. Neurol. 442(1):78-88.
30. Dosso AA, Leuenberger PM, & Rungger-Brandle E (1999) Remodeling of retinal capillaries in the diabetic hypertensive rat. Invest. Ophthalmol. Vis. Sci. 40(10):2405-2410.
III
78
Supplementary materials and methods
Brain microvessel isolation
All steps of the BMV isolation were done in the cold (≤ 4°C) as described for retina microvessel (RMV) isolation. The pooled brain cryosections were filled up with PBS / 1% dextran (composition of the PBS: NaCl, 137 mM; KCl, 2.7 mM; Na2HPO4, 10 mM,
KH2PO4,1.8 mM; pH adjusted to 7.4) to give a total volume of 50 ml and homogenized
using a motor-driven homogenizer (Homgen plus, Schuett Biotec, Goettingen, Germany; 60 rpm, 20 upstrokes) and centrifuged for 10 min (438 g). The pellet was resuspended in 22 ml PBS / 18% dextran and centrifuged again (4400 g, 15 min) to spin down the brain vascular components. This pellet was resuspended in 7 ml PBS / 1% dextran and the mixture loaded onto a density gradient column with PBS containing 31% dextran in the lower phase and 18% dextran in the upper phase. After centrifugation (1300 g, 15 min) the interphase was carefully collected, gently mixed with 20 ml PBS and filtered through a 60 μm nylon mesh to capture the BMVs.
The RMVs and BMVs were taken from the filter and suspended either in 350 μl lysis buffer (RLT buffer, Qiagen, Hilden, Germany) / 3.5μl mercaptoethanol (ME; Sigma-Aldrich, Munich, Germany) (RLT/ME buffer) for subsequent total RNA extraction or in 4% paraformaldehyde (PFA) solution for subsequent hematoxylin/eosin staining and histological analysis.
Quantitative reverse transcriptase polymerase chain reaction (RT-qPCR)
In order to improve comparability between the samples, the differences in the total RNA contents were adjusted by appropriate dilution with molecular biology grade water to establish concentrations that did not differ more than tenfold among the samples. Within this concentration range, the ∆Ct values are independent of the starting
amount of total RNA as shown in preliminary RT-qPCR experiments using tissue samples from two individual rats.
Reverse transcription into cDNA was performed with 10μL total RNA solution using a high capacity kit (cat# 4374966, Life technologies, Darmstadt, Germany). This was followed by a quantitative pre-amplification step consisting of 14 cycles using 4μL
III
79
cDNA solution, 4μL target specific primer pool, and 8μL TaqMan PreAmp MasterMix (cat# 4391128, Life technologies). The resulting pre-amplified solutions were diluted with water and mixed with TaqMan Universal PCR Master Mix (cat# 4369016, cat# 4374966, Life technologies) according to the manufacturer’s instructions. The ports of 384 well microfluidic cards were filled with 100 μl cDNA solution, briefly centrifuged (1,200g for 10 min) and sealed. Each micro-fluidic card contained assays to determine expression of 18 target genes and 4 reference genes which are listed in table S1. Real-time PCR (40 cycles) was performed using a ViiA cycler (Life Technologies).
III
80
Table S1: List of reference and cell specific marker genes along with the assaynumbers used in the microfluidic card measurements.
Gene
identifier Full gene name Assay no.
Reference genes
18S rRNA 18s ribosomal RNA Hs999999901_s1 EEF1a2 Eukaryotic translation elongation factor 1 alpha 2 Rn00561973_m1
ACTB Actin beta Rn00667869_m1
B2M Beta 2 microglobulin Rn00560865_m1
Endothelial cell marker genes
NOS 3 Endothelial NO synthetase Rn02132634_s1 TEK Receptor tyrosine kinase (Tie2) Rn01433337_m1 VWF von Willebrand factor Rn01492158_m1 SLC2A1 Solute carrier family 2, member 1 (GLUT-1) Rn01417099_m1
CLDN5 Claudin 5 Rn01753146_s1
ACE Angiotensin 1 converting enzyme Rn00578401_m1
Pericyte marker genes
PDGFRB Platelet-derived growth factor
receptor beta Rn00709573_m1
CSPG4 Chondroitin sulfate proteoglycan 4
(NG2) Rn00578849_m1
RGS5 Regulator of G-protein signaling 5 Rn00571047_m1 Smooth
muscle cell marker genes
CNN1 Calponin 1 Rn00582058_m1
MYH11 Myosin heavy chain 11 Rn01530321_m1 Glial cell
marker gene
GFAP Glial fibrillary acidic protein Rn01253033_m1
AQP4 Aquaporin 4 Rn01401327_s1
Neuronal cell marker genes
SYN1 Synapsin I Rn00569468_m1
RBFOX33 RNA binding fox-1 homolog 3
(NeuN) Rn01464214_m1
Photoreceptor marker genes
SAG S-antigen visual arrestin Rn00564656_m1 RBP3 Retinol binding protein 3 Rn01450461_m1 The names and abbreviations are used according to the HUGO Gene Nomenclature Committee (HGNC).
III
81
Supplementary results
Test of amplification efficiency
We diluted representative samples of total RNA from full retina and brain tissue and the respective microvessels obtained from 2 animals over a 4-fold log10 range and
checked for the amplification efficiency of the selected assays. The results shown below indicate a good linearity and comparable efficiency for all assays used.
Figure S1. Efficacy of the selected gene assays to quantitatively amplify the marker gene cDNA. Representative samples of full retina and brain tissue and the respective microvessels obtained from control (non-diabetic) animals were adjusted to a RNA concentration of 100 ng/μl and additionally diluted to 10, 1, and 0.1 ng/μl. Original samples and the dilutions were then pre-amplified and subjected to real-time qPCR on a single PCR plate. Shown are the resulting c(t) values for the dilutions and the marker genes versus the lg (2) of the dilution factor (original concentration set to 1). Even genes in highly diluted samples could still be detected in a linear relationship indicating a good primer efficacy close to 100% and a quantitative pre-amplification.
RNA quality assessment
Total RNA extracted from each tissue compartment underwent analysis of RNA content and integrity using chip-based technology (Nano and Pico kits along with the respective chemicals were obtained from Agilent, Waldbronn, Germany). With this approach the RNA integrity is indicated by the so-called RNA integrity number (RIN value) ranging
0 5 1 0 1 5 0 1 0 2 0 3 0 4 0 lg (2 ) d il fa c c( t)
III
82
from 0 to 10. In the present study we used only samples with a RIN value above 8. The results of the total RNA concentration and integrity are shown in table 2.
Expression of maker genes of nonvascular cells
The RMV and BMV extracts showed a marked accumulation of endothelial cell and pericyte marker genes as shown in the main article. We also checked the content of parenchymal cell marker genes (astrocytic glial cells, neuronal cells, and photoreceptor cells) and found a marked decrease of these markers in both, RMVs and BMVs. The results (expressed as ∆∆Ct values and fold decrease) using the respective full tissue
samples as reference are listed in table S3.
Expression of maker genes of pial arteries
In addition to the RMVs and BMVs we also isolated pial arteries from the rats and included these in our study. This allows comparison of the macro- and microvasculature of the same organ, i.e. the brain. The levels of microvessel wall-specific marker genes (i.e. endothelial cells and pericytes) in the PA wall are listed in table 4 indicating the ∆∆Ct values and the respective expression levels using the data obtained from BMVs
as reference.
Table S2. Total RNA content and RNA integrity numbers (RIN) obtained from the different types of tissue studied.
Control Hyperglycemia
Conc. RIN Conc. RIN
RMVs 0.5±0.1 9.3±0.7 0.6±0.3 9.9±0.3
BMVs 3.3±2.9 9.1±0.4 3.0±2.0 9.7±0.1
PAs 11.0±5.3 9.6±0.5 4.4±2.7 9.8±0.2
Retina tissue 26.2±14.0 9.9±0.14 18.4±6.7 10±0
Measurements were performed using Agilent picochip methodology. Not indicated are the values for brain tissue samples which showed a large scatter due to differences in the amounts of starting tissue. BMVs, brain microvessels; PA, pial arteries; RMVs, retina microvessels; Conc., concentration of total RNA given in ng/μl. Given are mean ± SD with n=6 rats in control condition and n=4 rats in hyperglycemia.
III
83
Table S3. Decrease of parenchymal cell markers in retina microvessels (RMVs) and brain microvessels (BMVs) extracts in control rats (n=6).
RMVs BMVs ∆∆Ct expression level (% vs. retina tissue) ∆∆Ct expression level (% vs. retina tissue) Astroc ytic glial mar ke r g enes GFAP 0.43±0.85 (43.0-128) 74 2.15±0.53 (16-32) 22 AQP4 3.14±1.05 (5.0-25.6) 11.3 2.86±0.64 (9-21) 14 Ne urona l c ell mar ke r g ene SYN1 2.02±1.11 24.6 (10.5-57.8) 4.77±0.37 3.7 (2.9-4.7) RBFOX3 n.d. (4) 5.15±0.76 3.0 (1.7-5.3)
Photoreceptor cell marke
r gene SAG 3.09±0.43
11.7
(8.4-16.4) n.d. RBP3 3.85±0.55 (4.5-10.6) 6.9 n.d.
Given are mean values along with SD for the ∆∆Ct values and the 95% confidence interval for the relative expression levels using the data obtained from the respective full tissue samples for reference. n.d., not detectable (number of samples in brackets). GFAP, glial fibrillary acidic protein; AQP4, aquaporin 4; SYN1, synapsin 1; RBFOX3, RNA binding fox-1 homolog 3 (NeuN); SAG, S-antigen visual arrestin; RBP3, Retinol binding protein 3.
III
84
Table S4: Levels of vessel wall marker genes in pial arteries obtained in normoglycemic and 8 weeks hyperglycemic rats.
Control Hyperglycemia ∆∆Ct expression level (% vs. retina tissue) ∆∆Ct expression level (% vs. retina tissue) Endothelial cell marker genes NOS3 1.52 ± 0.45 (24.6-49.4) 34.9 0.33 ± 0.25 79.7 (62.9-101) TEK 2.69 ± 0.39 15.4 (11.4-20,1) 1.68 ± 0.25 31.2 (24.6-39.5) VWF 1.99 ± 0.52 1.9 (1.4-2.6) 1.62± 0.42 32.5 (21.7-48.6) SLC2A1 5.72 ± 0.42 1.9 (1.4-2.6) 5.16 ± 0.42 2.8 (1.9-4.2) CLDN5 4.16 ± 0.67 5.6 (3.3-9.4) 2.69 ± 1.31 15.5 (4.5-54.3) ACE 1.07 ± 0.42 47.7 (34.6-65.7) 0.46± 0.41 72.6 (49.1-107) Pe ricyte ma rk er genes PDGFR B 1.14 ± 0.36 45.3 (34.4-59.7) 1.14± 0.25 45.3 (35.7-57.6) RGS5 4.35 ± 0.38 4.9 (3.7-6.6) 5.06 ± 0.76 3.0 (1.5-6.2) CSPG4 -0.70 ± 0.47 163 (113-234) 0.78 ± 1.34 58.4 (16.2-209)
Given are the ∆∆Ct values (mean ± SD) and the relative expression levels (mean values along with the 95% confidence intervals) using the data obtained from control and hyperglycemic brain microvessels (BMVs), respectively. NOS3, endothelial NO synthetase; TEK, TEK tyrosin-kinase; VWF, von Willebrand factor; SLC2A1, solute carrier family 2, member 1 (GLUT-1); CLDN5, Claudin 5; ACE, angiotensin 1 converting enzyme; PDGFRB, platelet-derived growth factor receptor beta; RGS5, regulator of G-protein signaling 5; CSPG4, chondroitin sulfate proteoglycan 4.
