Transcriptome-wide RNA sequencing analysis of rat skeletal … · 2017-11-16 · Transcriptome-wide...

Transcriptome-wide RNA sequencing analysis of rat skeletal muscle feedarteries. I. Impact of obesity

Nathan T. Jenkins,1 Jaume Padilla,2,3,4 Pamela K. Thorne,5 Jeffrey S. Martin,6 R. Scott Rector,2,7,8

J. Wade Davis,9,10,11 and M. Harold Laughlin4,5,12

1Department of Kinesiology, University of Georgia, Athens, Georgia; 2Department of Nutrition and Exercise Physiology,University of Missouri, Columbia, Missouri; 3Department of Child Health, University of Missouri, Columbia, Missouri;4Dalton Cardiovascular Research Center, University of Missouri, Columbia, Missouri; 5Department of Biomedical Sciences,University of Missouri, Columbia, Missouri; 6Department of Biomedical Sciences, Quinnipiac University, Hamden,Connecticut; 7Research Service, Harry S Truman Memorial Department of Veterans Affairs Medical Center, Columbia,Missouri; 8Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of Missouri, Columbia,Missouri; 9Department of Health Management and Informatics, University of Missouri, Columbia, Missouri; 10Department ofStatistics, University of Missouri, Columbia, Missouri; 11MU Informatics Institute, University of Missouri, Columbia,Missouri; and 12Department of Medical Pharmacology and Physiology, University of Missouri, Columbia, Missouri

Submitted 8 November 2013; accepted in final form 14 January 2014

Jenkins NT, Padilla J, Thorne PK, Martin JS, Rector RS,Davis JW, Laughlin MH. Transcriptome-wide RNA sequencinganalysis of rat skeletal muscle feed arteries. I. Impact of obesity. JAppl Physiol 116: 1017–1032, 2014. First published January 16,2014; doi:10.1152/japplphysiol.01233.2013.—We employed next-generation RNA sequencing (RNA-Seq) technology to determine theinfluence of obesity on global gene expression in skeletal muscle feedarteries. Transcriptional profiles of the gastrocnemius and soleusmuscle feed arteries (GFA and SFA, respectively) and aortic endo-thelial cell-enriched samples from obese Otsuka Long-EvansTokushima Fatty (OLETF) and lean Long-Evans Tokushima Otsuka(LETO) rats were examined. Obesity produced 282 upregulated and133 downregulated genes in SFA and 163 upregulated and 77 down-regulated genes in GFA [false discovery rate (FDR) � 10%] with anoverlap of 93 genes between the arteries. In LETO rats, there were 89upregulated and 114 downregulated genes in the GFA compared withthe SFA. There were 244 upregulated and 275 downregulated genes inOLETF rats (FDR � 10%) in the GFA compared with the SFA, withan overlap of 76 differentially expressed genes common to bothLETO and OLETF rats in both the GFA and SFA. A total of 396transcripts were found to be differentially expressed between LETOand OLETF in aortic endothelial cell-enriched samples. Overall, wefound 1) the existence of heterogeneity in the transcriptional profile ofthe SFA and GFA within healthy LETO rats, 2) that this between-vessel heterogeneity was markedly exacerbated in the hyperphagic,obese OLETF rat, and 3) a greater number of genes whose expressionwas altered by obesity in the SFA compared with the GFA. Also,results indicate that in OLETF rats the GFA takes on a relatively moreproatherogenic phenotype compared with the SFA.

hyperphagia; muscle; blood flow; next-generation sequencing; geneexpression; resistance arteries

THE NEGATIVE EFFECTS OF OBESITY on the skeletal muscle vascu-lature include impaired endothelial function, increased sensi-tivity to vasoconstrictor signals, vascular insulin resistance,and rarefaction (21, 24), ultimately resulting in a diminishedability to match tissue perfusion with local metabolic needs(22, 23). Increasing evidence suggests that these consequencesof obesity are heterogeneous, as not all vascular beds are

uniformly affected. For example, our group recently reportedthat gastrocnemius and soleus feed arteries (GFA and SFA,respectively) are differentially vulnerable to obesity-associatedendothelial dysfunction and are heterogeneous with respect tothe vascular actions of endothelium-dependent vasodilators (3,28). As it is well established that feed arteries serve as a criticalpoint of regulatory control of blood flow to skeletal muscle inrats (80), it is particularly important to improve our under-standing of the effects of obesity and related metabolic dys-regulation on the phenotype of these resistance vessels.

Although there is evidence that obesity negatively affects thefunction of skeletal muscle vascular beds (2, 3, 12, 19, 21–23,43, 63), molecular phenotyping of skeletal muscle resistancevasculature has not, to our knowledge, been performed in acomprehensive fashion. Here we report the results of a globalgene expression analysis of the influence of obesity on GFAand SFA in the Otsuka Long-Evans Tokushima Fatty (OLETF)rat model, an established model of hyperphagia-induced obe-sity and type 2 diabetes (30). We employed next-generationRNA sequencing (RNA-Seq) technology, which, given theenhanced ability of RNA-Seq to analyze the large scale andcomplexity of transcriptomes compared with microarrays (48,77), has emerged as the preferred method of global geneexpression analysis. RNA-Seq also overcomes the inherentbias of microarray analysis (48). Specifically, the design ofmicroarray experiments requires prior knowledge of transcriptsincluded on the array, whereas RNA-Seq interrogates all tran-scripts through direct reads. Moreover, because RNA-Seqprovides the most precise measurement of transcript levelsusing direct sequencing and quantitation compared with othermethods available, the need for confirmation of RNA-Seqresults via quantitative real-time PCR is eliminated (77). Nev-ertheless, RNA-Seq has not yet, to our knowledge, beenapplied to the analysis of transcriptomic effects of obesity onskeletal muscle resistance arteries.

The purpose of the present study was to determine theinfluence of obesity on GFA and SFA global gene expressionin the OLETF rat model with RNA-Seq. We hypothesized thatthe GFA would be more susceptible to obesity than the SFA interms of number of transcripts differentially expressed betweengroups and the magnitude of any differences observed. We also

Address for reprint requests and other correspondence: N. T. Jenkins, Dept.of Kinesiology, 115M Ramsey Center, 330 River Rd., Univ. of Georgia,Athens, GA 30602-6554 (e-mail: [email protected]).

J Appl Physiol 116: 1017–1032, 2014.First published January 16, 2014; doi:10.1152/japplphysiol.01233.2013.

http://www.jappl.org 1017

by 10.220.32.246 on Novem

ber 15, 2017http://jap.physiology.org/

Dow

nloaded from

mailto:[email protected]

http://jap.physiology.org/

sought to determine the influence of obesity on arterial endo-thelial cells, and therefore tested the hypothesis that aorticendothelial cell-enriched samples from OLETF rats woulddisplay a more proatherogenic gene expression profile com-pared with lean Long-Evans Tokushima Otsuka (LETO) rats.This data set was part of a large-scale RNA-Seq experimentinvolving 96 animals in 12 treatment groups and 18 differentvessels. Here we discuss the first research question we soughtto answer as part of that larger experiment, i.e., the effects ofobesity on GFA, SFA, and aortic endothelial cells. Our com-panion article (49a) presents our findings relative to the effectsof exercise training on these vessels in obese OLETF rats.Forthcoming manuscripts will address interactive effects oftraining and metformin on transcriptomic profiles throughoutthe arterial tree.

METHODS

Animals and experimental design. LETO (n � 12) and OLETF(n � 12) rats were obtained at age 4 wk (Japan SLC, Hamamatsu,Shizuoka, Japan). The OLETF rat, characterized by a mutated chole-cystokinin-1 receptor that results in a hyperphagic phenotype, is anestablished model of obesity, insulin resistance, and type 2 diabetes(30). Animals were individually housed in a temperature-controlled(21°C) environment with light from 0600 to 1800 and dark from 1800to 0600. All animals were given ad libitum access to standard chowwith a macronutrient composition of 56% carbohydrate, 17% fat, and27% protein (Formulab 5008, Purina Mills, St. Louis, MO). Rats wereanesthetized at 30–32 wk of age with an intraperitoneal injection ofpentobarbital sodium (50 mg/kg) between 0800 and 0930. Tissueswere then harvested, and the animals were killed by exsanguination.Food was removed from the cages 12 h prior to death. All protocolswere approved by the University of Missouri Animal Care and UseCommittee.

Body weight, body composition, food intake, and citrate synthase.Body weights and food intakes were monitored and recorded on aweekly basis. Weekly food intakes were averaged across ages 20–30

wk. Body composition was assessed by dual-energy X-ray absorpti-ometry (DXA; Hologic QDR-1000, calibrated for rodents) on the dayof death. Omental and retroperitoneal adipose tissue depots were thenremoved and weighed to the nearest 0.01 g. Citrate synthase activitywas measured from whole muscle homogenate of the vastus lateraliswith the spectrophotometric method of Srere (70).

Isolation of skeletal muscle feed arteries and aortic endothelialcells. Immediately after the gastrocnemius-plantaris-soleus musclecomplex was harvested, the muscles were pinned down in a petri dish

Table 1. Animal characteristics

Variable LETO OLETF

BW, g 460.4 � 15.4 685.5 � 13.6*Food intake, g/day 21.5 � 0.3 31.6 � 0.4*Food intake, g·day�1·g BW�1 0.32 � 0.003 0.34 � 0.005*Body fat, % 15.4 � 0.8 34.9 � 1.3*Retroperitoneal adipose tissue mass, g 6.7 � 0.6 46.8 � 2.9*Omental adipose tissue mass, g 0.7 � 0.1 3.2 � 0.5*Epididymal adipose tissue mass, g 7.6 � 0.6 19.5 � 1.0*HW, g 1.4 � 0.03 1.9 � 0.04*HW-to-BW ratio (�103) 3.1 � 0.1 2.8 � 0.1*Citrate synthase VL-red,

�mol·min�1·g�1 24.6 � 5.8 28.6 � 5.2Citrate synthase VL-white,

�mol·min�1·g�1 8.6 � 0.8 7.9 � 1.2Total cholesterol, mg/dl 95.3 � 4.2 149.6 � 7.1*LDL-cholesterol, mg/dl 59.0 � 4.0 41.5 � 7.3HDL-cholesterol, mg/dl 27.1 � 0.7 33.9 � 1.9*Triglycerides, mg/dl 45.6 � 4.5 371.1 � 47.9*NEFA, mmol/l 0.27 � 0.02 0.96 � 0.07*Insulin, ng/ml 2.9 � 0.4 8.4 � 1.6*Glucose, mg/dl 151 � 5.8 303 � 14.8*HOMA-IR index 1.1 � 0.1 6.4 � 1.4*HbA1c, % 5.0 � 0.1 7.2 � 0.4*

Values are means � SE. LETO, Long-Evans Tokushima Otsuka rat;OLETF, Otsuka Long-Evans Tokushima Fatty rat; BW, body weight; HW,heart weight; NEFA, nonesterified fatty acids; HbA1c, glycosylated hemoglo-bin. *Significant difference between groups (P � 0.05).

-80

-60

-40

-20

0

20

40

-80 -60 -40 -20 0 20 40

Fold

Diff

eren

ce in

SFA

(O

LETF

/LET

O)

Fold Difference in GFA (OLETF/LETO)

SFAGFA

Effects of Obesity on SFA and GFA

Fig. 1. Top: number of genes altered by obesity in gastrocnemius and soleusmuscle feed arteries (GFA and SFA, respectively);1 indicates upregulation inOtsuka Long-Evans Tokushima Fatty rat (OLETF) relative to lean Long-EvansTokushima Otsuka rat (LETO), and 2 indicates downregulation in OLETFrelative to LETO. Circle sizes and overlapping area are proportional to thenumber of genes altered. Bottom: between-artery correlation in changes ofgene expression induced by obesity. Obesity produced an overlap of 93 genesbetween GFA and SFA. Each dot represents a gene. Genes depicted in reddisplayed opposite directional changes. The gene that was increased in GFAbut decreased in SFA was Dnpep (top left quadrant), and the gene that wasdecreased in GFA but increased in SFA was Mylpf (bottom right quadrant).Dashed line of identity indicates perfect agreement between the arteries.

1018 Obesity and Vascular Transcriptional Regulation • Jenkins NT et al.

J Appl Physiol • doi:10.1152/japplphysiol.01233.2013 • www.jappl.org

by 10.220.32.246 on Novem


Dow

nloaded from


containing a cold RNA-stabilizing agent (RNAlater; Ambion, Austin,TX). Under the microscope, the SFA and GFA were then dissectedclean of perivascular adipose tissue and excess adventitia as describedpreviously (3, 28, 34, 39, 42, 82, 83). The single GFA supplying themedial head of the gastrocnemius muscle was used for the presentstudy. In our experience, OLETF rats typically have one to three SFA.All SFA were dissected and pooled for the present RNA-Seq analysis.The reader is referred to our recent publication for a visual of theanatomic location and structure of GFA and SFA (28). Aortic endo-thelial cells were isolated by gentle scraping of longitudinally openedaortas as described previously (6, 50, 52, 66). This method of scraping

the luminal surface yields an endothelial cell-enriched sample (50).The aorta was chosen to provide sufficient yield of endothelial cells.Indeed, we have established that we recover �200 ng of total RNAper rat aortic endothelial scrape, which is above the minimum requiredfor RNA-Seq experiments. Samples were kept in RNAlater for 48 h at4°C and then removed from the RNAlater solution and stored at�80°C until analysis.

Blood parameters. Whole blood was collected on the day ofeuthanasia for analysis of glycosylated hemoglobin (HbA1c) by theboronate-affinity high-performance liquid chromatography method(Primus Diagnostics, Kansas City, MO) in the Diabetes Diagnostics

Table 2. Top 20 genes differentially expressed between OLETF and LETO in GFA, sorted by magnitude of fold change

EntrezID Symbol Name FDR Fold

OLETF LETO in GFA

288272 Mis18a MIS18 kinetochore protein homolog A (S. pombe) �0.001 27.9293154 Folr2 Folate receptor 2 (fetal) �0.001 18.6499507 Fam151b Family with sequence similarity 151, member B �0.001 10.9113882 Hemgn Hemogen 0.003 7.7282634 Pxmp4 Peroxisomal membrane protein 4 �0.001 7.394270 Nnat Neuronatin 0.046 5.3

292843 Siglec5 Sialic acid binding Ig-like lectin 5 �0.001 5.329692 Pla2 g2a Phospholipase A2, group IIA (platelets, synovial fluid) �0.001 5.2

690045 Klra17 Killer cell lectin-like receptor, subfamily A, member 17 �0.001 4.9300095 Srebf2 Sterol regulatory element binding transcription factor 2 �0.001 4.4304545 Oasl 2=–5=-oligoadenylate synthetase-like 0.017 4.0246268 Oas1b 2–5 Oligoadenylate synthetase 1B 0.002 3.7

OLETF � LETO in GFA

25086 Cyp2e1 cytochrome P-450, family 2, subfamily e, polypeptide 1 �0.001 �35.7314397 Rin3 Ras and Rab interactor 3 �0.001 �10.9361171 Golga7 Golgin A7 �0.001 �9.8286965 Abra Actin-binding Rho activating protein 0.006 �6.6

89808 Cx3cl1 Chemokine (C-X3-C motif) ligand 1 �0.001 �4.424231 C2 Complement component 2 �0.001 �4.124472 Hspa1a Heat shock 70 kDa protein 1A 0.002 �4.024471 Hspb1 Heat shock protein 1 0.001 �3.9

GFA, gastrocnemius feed artery.

Table 3. Top 20 genes differentially expressed between OLETF and LETO in SFA, sorted by magnitude of fold change


OLETF LETO in SFA

288272 Mis18a MIS18 kinetochore protein homolog A (S. pombe) �0.001 30.8266760 Nalcn Sodium leak channel, nonselective �0.001 20.1689232 Nxnl2 Nucleoredoxin-like 2 0.009 12.3499507 Fam151b Family with sequence similarity 151, member B 0.002 7.3293154 Folr2 Folate receptor 2 (fetal) �0.001 6.5282634 Pxmp4 Peroxisomal membrane protein 4 �0.001 5.8300095 Srebf2 Sterol regulatory element binding transcription factor 2 �0.001 5.5294343 Zfp280b Zinc finger protein 280B �0.001 5.3365841 Slc25a44 Solute carrier family 25, member 44 0.002 4.729142 Vnn1 Vanin 1 0.01 4.683632 Deaf1 DEAF1 transcription factor 0.004 4.4

300968 Uba5 Ubiquitin-like modifier activating enzyme 5 �0.001 4.3

OLETF � LETO in SFA

25086 Cyp2e1 Cytochrome P-450, family 2, subfamily e, polypeptide 1 �0.001 �76.9307100 Calml3 Calmodulin-like 3 0.006 �15.6314397 Rin3 Ras and Rab interactor 3 �0.001 �13.324231 C2 Complement component 2 �0.001 �7.8

361171 Golga7 Golgin A7 �0.001 �6.625708 Ucp3 Uncoupling protein 3 (mitochondrial, proton carrier) 0.065 �6.4

313057 Serinc2 Serine incorporator 2 0.075 �6.325357 Thrsp Thyroid hormone responsive 0.058 �4.8

SFA, soleus feed artery.

1019Obesity and Vascular Transcriptional Regulation • Jenkins NT et al.


by 10.220.32.246 on Novem


Dow

nloaded from


Laboratory at the University of Missouri. Serum samples were pre-pared by centrifugation and stored at �80°C until analysis. Glucose,triglyceride (TG), and nonesterified fatty acid (NEFA) assays wereperformed by a commercial laboratory (Comparative Clinical Pathol-ogy Services, Columbia, MO) on an Olympus AU680 automatedchemistry analyzer (Beckman-Coulter, Brea, CA) using commerciallyavailable assays according to manufacturers’ guidelines. Plasma in-sulin concentrations were determined with a commercially availablerat-specific enzyme-linked immunosorbent assay (Alpco Diagnostics,Salem, NH). Samples were run in duplicate, and manufacturers’controls and calibrators were used according to assay instructions.

RNA extraction. Total RNA isolations were performed with theNucleoMag 96 RNA kit procedure (Clontech part no. 744350.1),which is a single-tube method based on reactive magnetic beadtechnology designed for automated small- or large-scale preparationof highly pure total RNA from tissue or cell samples. All liquidhandling was optimized for use on a Beckman3000 robotic liquidhandler housed within a laminar flow hood (with UV decontamina-tion) designed to ensure a clean room environment for working withmicrotissues, which yield low (pg to ng) amounts of RNA. Briefly,groups of 24 sample vessels were removed from �80°C and imme-diately homogenized for 60–120 s in their own 2-ml microtube withthe Mini-Beadbeater-24 (BioSpec Products) in the presence ofNucleoMag lysis buffer and several miniature chrome-steel (RNasetreated) BBs. Care was taken to get complete microvessel disruption

without extending grinding times to prevent the generation of excessheat. The resulting homogenate was then loaded onto the robot deck,and a digital photo was taken before the sample was transferred into96-deep well microplates. The photo allowed us to have a physicalrecord of each sample ID prior to loading into the microplate foraccurate tracking purposes. This process was repeated four times tocompletely fill a 96-well plate within 10–20 min. The combination ofusing stabilized tissue and immediate homogenization in chaotropicsalt-containing lysis buffer ensured that the RNA was protected fromRNase degradation during tissue disruption. After homogenization,the RNA was bound to RNA beads in the presence of alcohol, and amagnet was used to perform several wash and elution steps in acompletely automated fashion. This method included a DNase diges-tion step ultimately yielding RNA of similar yield and quality fromcolumn-based procedures. Immediately after RNA isolation, pureRNA was transferred to a new 96-well plate and a 5-�l aliquot wastaken into a second plate for RNA quality control. Both plates werestored at �80°C with cryogenic plate seals and placed in secondarycontainment to prevent frost buildup on the plates during storage.

RNA quality control (concentration and integrity). For assessingtotal RNA yield and integrity, tandem Agilent Bioanalyzer 2100instruments were used in combination with the Total RNA 6000 PicoAssay. At the time of this study, the RNA Pico LabChip Kit was theonly platform to give unbiased assessment of RNA integrity (via RIN)and accurate results with extremely low RNA concentrations such as

Top Obesity Responsive Gene Network - GFA

Fig. 2. Top-scoring gene network influencedby obesity in the GFA (score � 76). Nodesrepresent genes/molecules. Shading is in pro-portion to the size of the fold change (red,upregulation; green, downregulation). Whitenodes denote network members that were notrepresented on the array. Gray nodes denotenetwork members that did not reach false dis-covery rate (FDR) � 10%. Solid lines denotedirect relationships.



by 10.220.32.246 on Novem


Dow

nloaded from


those provided by microvessels. The qualitative range for the totalRNA assay is 200–5,000 pg/�l, and the most important advantage ofthis system is the small amount of sample used (1 �l), leaving the restof the RNA for other applications. Typical yields from rat microves-sels were �500–1,000 pg/�l. RNA quality control was performedwith only the aliquot from each isolation plate.

Illumina library preparation (SMARTer amplification and RNA-Seq). Because of the low yields of total RNA from microvessels, totalRNA could not be used directly in traditional Illumina gene expres-sion profiling methods (RNA-Seq) because of the low concentrationof the samples (standard RNA-Seq kits during this project required0.1–1 �g of total RNA). Thus the SMARTer Ultra Low RNA Kit forIllumina Sequencing (Clontech, catalog no. 634935) was utilized forgenerating full-length cDNA transcripts prior to Illumina RNA-Seqlibrary preparation. Briefly, the technology involves SMARTer first-strand cDNA synthesis and purification, utilizing the SMARTer an-chor sequence and poly(A) sequence that serve as universal primingsites for end-to-end generation of single-stranded cDNA, followed bycDNA amplification with LongDistance PCR (LD-PCR) using themanufacturer’s recommended Advantage 2 PCR system (Clontech,catalog no. PT3281-1) containing a novel polymerase and ultrapuredNTPs specifically for Illumina sequencing. Using the concentrationvalues from the Bioanalyzer RNA Pico Assay, we sought to use100–1,000 pg of total RNA as input to the SMARTer 1st cDNAreaction.

After cDNA generation, validation was performed with the Bio-analyzer 2100 High Sensitivity DNA Assay (Agilent, catalog no.5067-4626) for select samples from each plate of 96 samples in orderto accurately size and quantitate DNA up to 12 kb in length, againconsuming minimal sample volumes (1 �l). After 14 cycles ofLD-PCR amplification the final cDNA yields were estimated at�1–10 ng for each microvessel, which is a suitable input amount forlibrary preparation for cDNA/ChIP Seq library preparation. To gen-erate Illumina Paired-End RNAseq libraries, cDNA was fragmentedto �200 bp with the Q700 DNA fragmentation system (QSonica) andthen used directly with the NextFlex DNA preparation kit (BiooScientific, catalog no. 5140-02) with some modifications. Briefly,fragment cDNA was end-repaired and purified with 1.8x SPRI beadsto remove reaction components (Agencourt). The resulting blunt endswere A-tailed in preparation for cohesive ligation to the Illuminaspecific sequencing adapters diluted to 0.6 �M working concentration(NextFlex DNA Adapters, Bioo Scientific, catalog no. 514104). Li-gated DNA was purified twice with 1.0x SPRI to remove adapterdimers and perform gel-free size selection and then amplified through14 cycles of PCR. The final sequencing construct was purified with a1.0x SPRI to remove low-molecular-weight adapter dimer artifacts (ifany), and libraries were validated to contain �330-bp fragments withthe Bioanalyzer 2100 High Sensitivity DNA Assay. Library quanti-tation was performed with the Qubit 2.0 fluorometer and the HighSensitivity DNA assay (Life Tech, catalog no. Q32851).

Top Obesity Responsive Gene Network - SFA

Fig. 3. Top-scoring gene network influencedby obesity in the SFA (score � 66). Nodesrepresent genes/molecules. Shading is in pro-portion to the size of the fold change (red,upregulation; green, downregulation).



by 10.220.32.246 on Novem


Dow

nloaded from


RNA sequencing. By utilizing 48 unique adapter indexes duringlibrary preparation across each plate of 96 libraries created, we wereable to overcome several common technical mistakes. First, it allowedus to account for technical biases through randomization of samplesby vessel type (group) and treatment across each plate of samples.Second, by having a priori knowledge of the sequencing index used toidentify each sample from the sequencing, we were able to use asingle manufacturing lot of adapters that were uniformly diluted andpreplated to ensure similar ligation efficiencies across several plates(hundreds of samples) used in the study. Third, this indexing schemeallowed us to standardize the pooling of several libraries by rowwithin each plate, where equimolar volumes of each sample in a platerow were pooled to a final concentration of 5–10 nM. Altogether, thisapproach prevented inadvertent use of the wrong adapters duringpreparation, randomized the sample and index combinations, andallowed for reduced mixing up of libraries within each sequencingpool. The final pools (50 total) were each loaded on a single lane ofsingle-read 50-base sequencing on the Illumina HiSeq2000 and ulti-mately yielded �175–200 million useable reads per lane (14–17million reads per RNAseq sample). It should be noted that thecombination of the 48 adapters that resulted in 4 pools of 12 indexeswas carefully designed and wet lab tested to be compatible with theHiSeq to ensure maximum sequence yields and to ensure that eachsample was correctly identified by the HiSeq during the index iden-tification steps.

Statistical analysis. The primary analysis of the preprocessedRNA-Seq data was conducted with R/Bioconductor, with other free,open-source programs used for preprocessing. The preprocessingsteps consisted of 1) quality assessment of raw sequenced library data(FASTQ files) with FastQC (http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc), 2) trimming of adapters from each librarywith Cutadapt (40), 3) aligning libraries to the rn4 genome build(http://hgdownload-test.cse.ucsc.edu/goldenPath/rn4/bigZips/) with aseed-and-vote method (37) and saving in SAM format, and 4) forminga read-count matrix at the gene level by read summarization usingfeatureCounts in Rsubread.

Nonspecific filtering of genes prior to statistical testing was carriedout to increase detection power (5), based on the requirement that agene have an expression level greater than 16 counts per million readsmapped (CPM) for at least 9 libraries across the GFA and SFAsamples. For the nonspecific filtering of the aortic endothelial cells,the reader is referred to our companion article (49a). In either case,this CPM cutoff was established empirically based on the point atwhich the ERCC Spike-Ins at different concentrations were no longerdistinguishable for that particular set of data. Library normalization toadjust for differences in library size was then made with the TMMmethod (59) in edgeR (58). This was used in conjunction with aflexible mean-variance modeling and transformation process knownas Voom (35), as part of the limma package (69), a combinationknown as limmaVoom that was recently shown to perform among thebest of available techniques in an comprehensive comparison ofdifferential gene expression methods for RNA-Seq data (56). Withthis framework, normalized expression levels were compared betweenthe groups of interest and significant differences identified by foldchange and adjusted P values. Adjustment to the P values was madeto account for multiple testing with the false discovery rate (FDR)method of Benjamini and Hochberg (4). For analysis of obesity effects(i.e., OLETF vs. LETO within GFA and within SFA) as well asanalysis of vessel effects (i.e., GFA vs. SFA within LETO and withinOLETF), we chose 10% as our FDR threshold for statistical signifi-cance. For analysis of interactive effects of obesity and vessel,computed as (OLETF GFA/LETO GFA)/(OLETF SFA/LETO SFA),we adopted a FDR of 20%, knowing that there is less power to detect aninteractive effect than a main effect (e.g., obesity or vessel). For allcomparisons, a blocking factor on the flow cell was used in the fittedmodel to adjust for systematic flow cell effects, and the robust version oflimma was utilized during model fitting. For between-vessel compari-

sons, we accounted for the dependency due to multiple measurementscoming from the same rat using the duplicateCorrelation option (69) inlimma, but the correlation was negligible (�6%) within the block [rat].As an empirical measure of the FDR, we evaluated what proportion of theidentical ERCC probe/concentration combinations (Set B) appeared inour list of differentially expressed genes. Similarly, we looked at a set of13 putative housekeeping genes derived from a study of more than13,000 rat samples (13) to have another estimate of our FDR. The set ofgenes was Actb, B2m, Gapdh, Gusb, Hprt1, Hmbs, Hsp90b1, Rpl13a,Rps29, Rplp0, Ppia, Tbp, and Tuba1. For both of these sets of controls,we also estimated the fold change of each of the genes as a measure ofthe accuracy of the fold change estimates.

Networks were generated through the use of Ingenuity PathwayAnalysis (Ingenuity Systems, www.ingenuity.com), henceforth IPA,as previously described (52). The full lists of all differentially ex-

LETOOLETF

-10

-5

0

5

10

15

20

-10 -5 0 5 10 15 20

Fold

Diff

eren

ce in

LET

O

(GFA

/SFA

)

Fold Difference in OLETF (GFA/SFA)

Between-Vessel Differences in LETO and OLETF Rats

Fig. 4. Top: number of genes differentially expressed between GFA and SFAin LETO and OLETF rats; 1 indicates upregulation in GFA relative to SFA,and 2 indicates downregulation in GFA relative to SFA. Bottom: between-artery correlation in changes of gene expression induced by obesity. There wasan overlap of 76 genes differentially expressed between GFA and SFA. Eachdot represents a gene. Dashed line of identity indicates perfect agreementbetween the groups.



by 10.220.32.246 on Novem


Dow

nloaded from

http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc

http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc

http://hgdownload-test.cse.ucsc.edu/goldenPath/rn4/bigZips/

http://www.ingenuity.com


pressed genes were uploaded into IPA with Entrez GeneID, foldchange, and adjusted P value. In the case of duplicate mapped IDs, themedian values (fold change and adjusted P value) were used torepresent the single results for that ID. Genes that were not detected,or those that were filtered out (as previously described), were assigneda fold change of 1 and an adjusted P value of 1. An adjusted P value�10% was set to identify molecules whose expression was signifi-cantly differentially regulated, which IPA terms as Network EligibleMolecules (NEMs). Our networks were built using only knowledgefrom experimentally observed relationships contained in the IngenuityKnowledge Base (IKB) and were algorithmically generated to maxi-mize their specific connectivity with each other. Higher networkscores imply a lower probability of finding the observed number of

NEMs in a given network by random chance. Specifically, the scoreis the negative log10 of the P value from Fisher’s exact test applied toa given network. For example, a score of 9 for a network implies a1-in-a-billion chance of obtaining a network containing at least thesame number of NEMs by chance when randomly picking 35 mole-cules from the IKB. A detailed description of the network-generatingalgorithm is provided by Calvano et al. (7).

Graphical representations of the networks were generated withIPA’s Path Designer, which illustrates the relationships betweenmolecules. Molecules are represented as nodes, and the biologicalrelationship between two nodes is represented as an edge (connectingline). All edges are supported by at least one reference from theliterature, from a textbook, or from canonical information stored in the

Table 4. Top 20 genes differentially expressed between GFA and SFA within LETO, sorted by magnitude of fold change


GFA SFA within LETO361765 Ptgis Prostaglandin I2 (prostacyclin) synthase �0.001 12.3

64155 Nalcn Sodium leak channel, nonselective 0.005 9.6304204 Mab21l2 mab-21-like 2 (C. elegans) 0.005 9.3362871 Pnoc Prepronociceptin 0.099 8.6364681 Penk Proenkephalin �0.001 5.1170538 Phex Phosphate regulating endopeptidase homolog, X-linked �0.001 5.0361018 Fhod3 Formin homology 2 domain containing 3 0.005 4.6304078 Fam46b Family with sequence similarity 46, member B 0.094 4.324250 Asb2 Ankyrin repeat and SOCS box-containing 2 0.005 4.2

303313 Sfrp5 Secreted frizzled-related protein 5 0.015 3.6306283 Apod Apolipoprotein D 0.045 3.525527 Aass Aminoadipate-semialdehyde synthase 0.028 3.4

287526 Srpk3 SRSF protein kinase 3 0.011 3.2362245 Prx Periaxin 0.040 2.9

GFA � SFA within LETO

297902 Mylpf Myosin light chain, phosphorylatable, fast skeletal muscle 0.098 �10.4304582 Hemgn Hemogen 0.057 �4.7113882 Anxa8 Annexin A8 0.028 �4.324215 Cbs Cystathionine beta synthase �0.001 �4.1

291015 Prg4 Proteoglycan 4 �0.001 �3.9192181 Esm1 Endothelial cell-specific molecule 1 �0.001 �3.1

Table 5. Top 20 genes differentially expressed between GFA and SFA within OLETF, sorted by magnitude of fold change


GFA SFA within OLETF

680102 Mab21l2 mab-21-like 2 (C. elegans) �0.001 19.3362282 Pck1 Phosphoenolpyruvate carboxykinase 1 (soluble) �0.001 10.7307100 Calml3 Calmodulin-like 3 0.036 9.094270 Nnat Neuronatin 0.01 8.825527 Ptgis Prostaglandin I2 (prostacyclin) synthase �0.001 6.2

300027 Gpihbp1 Glycosylphosphatidylinositol anchored high density lipoprotein binding protein 1 0.063 6.154232 Car3 Carbonic anhydrase 3 0.006 5.6

302920 Rbfox1 RNA binding protein, fox-1 homolog (C. elegans) 1 0.055 5.524638 Pfkfb1 6-Phosphofructo-2-kinase/fructose-2,6-biphosphatase 1 0.058 5.281670 Gpt Glutamic-pyruvate transaminase (alanine aminotransferase) 0.028 4.9

252900 Dgat2 Diacylglycerol O-acyltransferase 2 0.012 4.925357 Thrsp Thyroid hormone responsive 0.046 4.7

500292 Cidec Cell death-inducing DFFA-like effector c 0.025 4.560666 Gpd1 Glycerol-3-phosphate dehydrogenase 1 (soluble) 0.017 4.425086 Cyp2e1 Cytochrome P-450, family 2, subfamily e, polypeptide 1 0.095 4.129171 Aqp7 Aquaporin 7 0.075 4.029397 Ccl11 Chemokine (C-C motif) ligand 11 0.006 4.0

GFA � SFA within OLETF

289104 Prg4 Proteoglycan 4 �0.001 �6.8170913 Abcb1a ATP-binding cassette, sub-family B (MDR/TAP), member 1A 0.02 �5.024215 Atp2b2 ATPase, Ca2 transporting, plasma membrane 2 �0.001 �4.1



by 10.220.32.246 on Novem


Dow

nloaded from


IKB. These are rich, high-information content graphics, with detailsincluded in the figure legends.

Finally, between-group differences for all descriptive variableswere determined by using an independent t-test, for which statisticalsignificance was declared at P � 0.05.

RESULTS

Animal characteristics are summarized in Table 1. For thefive subsequent group comparisons reported involving SFAand GFA the average ERCC Spike-In (Set B, n � 10 probes)empirical FDR was 3.33% at the nominal FDR cutoff of 10%(mean fold � 0.95), while for the putative housekeeping genesthe average was 4.7% (fold � 1.01), results that both stronglysupport the methodology used. For the comparison reportedinvolving aortic endothelial cells the average ERCC Spike-In(Set B, n � 5 probes) empirical FDR was 0% at the nominalFDR cutoff of 10% (mean fold � 0.57), while for the putativehousekeeping genes the average was 10% (fold � 0.985). Itshould be noted that the statistics reported for the putativehousekeeping genes are only based on 11 (SFA and GFA) or

10 (aortic endothelial cells) of the 13 planned housekeepinggenes, because some genes were expressed at low levels anddid not pass the nonspecific filtering criteria. In their entiretythese findings strongly support the methodology used because,on average, the fold changes for these controls are approxi-mately equal to 1 and the empirical FDR is approximatelyequal to the target nominal FDR.

Impact of obesity on GFA and SFA. Figure 1 displays thenumber of differentially expressed genes in GFA and SFAbetween OLETF and LETO rats. Obesity produced 282 up-regulated and 133 downregulated genes in SFA and 163upregulated and 77 downregulated genes in GFA (FDR �10%), with an overlap of 93 genes between the arteries. Of the93 genes altered in both SFA and GFA (overlapping portion ofVenn diagram in Fig. 1), 68 were upregulated and 23 weredownregulated in both vessels. Although the direction ofchange for most genes was the same in both feed arteries, twogenes were oppositely responsive to obesity between the ar-teries, ceruloplasmin (Cp) and aspartyl aminopeptidase (Dn-

Top Network of Genes DifferentiallyExpressed between GFA & SFA - LETO

Fig. 5. Top-scoring gene network of genes dif-ferentially expressed between GFA and SFA inLETO animals (score � 45). Nodes representgenes/molecules. Shading is in proportion to thesize of the fold change (red, upregulation; green,downregulation). Gray nodes denote networkmembers that were not represented on the array.Gray nodes denote network members that did notreach false discovery rate (FDR) � 10%. Solidand dashed lines denote direct and indirect rela-tionships, respectively.



by 10.220.32.246 on Novem


Dow

nloaded from


pep) (Fig. 1). One (Cp) was upregulated in SFA and down-regulated in GFA, whereas the other (Dnpep) was downregu-lated in SFA and upregulated in GFA. Table 2 (GFA) andTable 3 (SFA) provide lists of the top 20 genes differentiallyexpressed between LETO and OLETF groups (sorted by mag-nitude of fold change), and the full lists of differentiallyexpressed genes are provided in Supplemental Tables S1 andS2.1 Figure 2 and Figure 3 illustrate the top-scoring genenetworks influenced by obesity in the GFA and the SFA,respectively. The scores for these gene networks were 76 in theGFA and 66 in the SFA.

Between-vessel differences in LETO and OLETF rats.Figure 4 presents the number of genes differentially expressedbetween GFA and SFA in LETO and OLETF rats. In LETOrats, there were 89 upregulated and 114 downregulated genesin the GFA compared with the SFA. There were 244 upregu-lated and 275 downregulated within OLETF (FDR � 10%) inthe GFA compared with the SFA. Of the 76 differentiallyexpressed genes common to both LETO and OLETF, 25 wereupregulated and 51 were downregulated in GFA compared

with SFA. The directional differences were consistent for eachof these genes (Fig. 4). Table 4 (LETO) and Table 5 (OLETF)provide lists of the top 20 genes differentially expressed betweenGFA and SFA, and the full lists of differentially expressed genesare provided in Supplemental Tables S3 and S4. Figure 5 andFigure 6 illustrate the top-scoring gene networks for between-vessel differences in LETO and OLETF rats, respectively. Thescores for these gene networks were 45 in LETO (Fig. 5) and 43in OLETF (Fig. 6) rats.

Obesity � vessel interaction analysis. Table 6 presents thefull list of genes differentially influenced by obesity betweenthe vessels, i.e., the result of our interaction analysis. Therewere 16 genes identified as upregulated, i.e., for which theeffect of obesity was greater in the GFA than in the SFA, and20 downregulated genes for which the effect of obesity wasgreater in the SFA than in the GFA (FDR � 20%). Thetop-scoring gene network for the interaction analysis is pre-sented in Fig. 7, and the score for this network was 36.

Impact of obesity on aortic endothelial cells. A total of 396transcripts were found to be differentially expressed betweenLETO and OLETF in aortic endothelial cell-enriched samples(FDR � 10%), of which 235 were upregulated in OLETF and161 were downregulated in OLETF compared with LETO.1 Supplemental Material for this article is available online at the Journal

website.

Top Network of Genes DifferentiallyExpressed between GFA & SFA - OLETF

Fig. 6. Top network of genes differentially ex-pressed between GFA and SFA in LETO ani-mals (score � 43). Nodes represent genes/mol-ecules. Shading is in proportion to the size of thefold change (red, upregulation; green, down-regulation). White nodes denote network mem-bers that were not represented in the RNA-seqdatabase. Gray nodes denote network membersthat did not reach FDR � 10%.



by 10.220.32.246 on Novem


Dow

nloaded from


Table 7 provides a list of the top 20 genes differentiallyexpressed between the groups, and the full list is provided inSupplemental Table S5. The top-scoring gene network forthe effects of obesity on aortic endothelial cells is illustratedin Fig. 8, and the score for this network was 38.

DISCUSSION

Our study provides RNA-Seq-based transcriptome assess-ment of skeletal muscle feed arteries of lean and obese rats.Our primary findings are that 1) there is marked heterogeneitybetween the transcriptional profiles of the SFA and GFA,2) this heterogeneity is greater in the hyperphagic, obeseOLETF rat, with the number of differentially expressed genesbetween the vessels being far greater in OLETF than in LETO,3) the number of genes whose expression is altered by obesitywas greater in the SFA compared with the GFA, and 4) in theOLETF rat, the GFA takes on a relatively more proatherogenicphenotype compared with the SFA. Overall, our results indi-cate that obesity induces transcriptional alterations of genesthat have been implicated in development of vascular dysfunc-tion and proatherogenic vascular cell phenotypes. Also, obesityappears to alter expression of gene networks centered on Ubc.

The application of RNA-Seq analysis allowed us to examinethe effects of obesity on the entire transcriptome in a compre-hensive manner instead of focusing on expression of a targetedset of a few genes (e.g., with qPCR) or with a preselected setof transcripts (e.g., with microarray). A significant and unex-pected finding of the present study was that obesity induced agreater number of transcriptional changes in the SFA relativeto the GFA. Given previous evidence from our laboratory thatthe SFA is protected against obesity-induced endothelial dys-function in the OLETF rat model (3, 28), we hypothesized agreater number of transcriptional differences in the GFA com-pared with the SFA. Instead, the present data indicate thatdecreased endothelium-dependent dilation induced by obesitydoes not correlate with the number of differentially expressedgenes.

Between-vessel differences in LETO and OLETF. We as-sume that the list of top genes differentially expressed betweenGFA and SFA within LETO (Table 4) reflects innate differ-ences between these feed arteries in the healthy state. Severalgenes identified seem to reflect developmental differencesbetween the vessels. For example, Mab21l2, Fhod3, Apod, andSrpk3 are all involved in developmental processes (29, 46, 60,

Table 6. Genes differentially altered by obesity between GFA and SFA (“interaction”), sorted by magnitude of fold change

EntrezID Symbol NameGFA

(Fold O/L)SFA

(Fold O/L)

Interaction:(GFA-O/GFA-L)(SFA-O/SFA-L) FDR

Up (GFA-O/GFA-L) (SFA-O/SFA-L)

362282 Pck1 Phosphoenolpyruvate carboxykinase 1 (soluble) 2.8 �3.3 9.3 0.16254249 Cfd Complement factor D (adipsin) 1.1 �3.3 4.0 0.121

312677 Ccdc77 Coiled-coil domain containing 77 1.8 �2.5 3.9 0.12179451 Fabp4 Fatty acid binding protein 4, adipocyte 1.3 �2.5 3.3 0.113

171341 Mgst1 Microsomal glutathione S-transferase 1 1.3 �2.5 3.2 0.17829692 Pla2 g2a Phospholipase A2, group IIA (platelets, synovial fluid) 5.2 1.6 3.2 0.113

311617 Fitm2 Fat storage-inducing transmembrane protein 2 1.3 �2.0 2.6 0.162361256 Svil Supervillin 1.6 �1.3 2.1 0.19524624 Pccb Propionyl CoA carboxylase, beta polypeptide 1.4 �1.4 1.9 0.15724534 Ldhb Lactate dehydrogenase B 1.1 �1.7 1.9 0.17824825 Tf Transferrin 1.4 �1.4 1.9 0.178

246298 Retsat Retinol saturase (all-trans retinol 13,14 reductase) 1.3 �1.4 1.9 0.11360350 Cd14 CD14 molecule 1.3 �1.4 1.8 0.162

364624 Lsm1 LSM1 homolog, U6 small nuclear RNA associated (S. cerevisiae) 1.1 �1.7 1.8 0.178691966 Sqrdl Sulfide quinone reductase-like (yeast) 1.3 �1.3 1.7 0.113301529 Dnpep Aspartyl aminopeptidase 1.2 �1.3 1.5 0.113

Down (GFA-O/GFA-L) � (SFA-O/SFA-L)

364033 Vamp4 Vesicle-associated membrane protein 4 �1.1 1.3 �1.5 0.17825479 Vps35 Vacuolar protein sorting 35 homolog (S. cerevisiae) �1.1 1.4 �1.5 0.113

287765 Ddx5 DEAD (Asp-Glu-Ala-Asp) box helicase 5 1.0 1.5 1.0 0.198295690 Dnajc10 DnaJ (Hsp40) homolog, subfamily C, member 10 �1.3 1.3 �1.6 0.178308069 Slc12a7 Solute carrier family 12 (potassium/chloride transporters), member 7 �1.3 1.3 �1.6 0.134246215 Tpcn1 Two pore segment channel 1 �1.4 1.1 �1.7 0.178

24268 Cp Ceruloplasmin (ferroxidase) �1.3 1.3 �1.8 0.113315608 Ube4a Ubiquitination factor E4A �1.1 1.6 �1.8 0.178289693 Dhx15 DEAH (Asp-Glu-Ala-His) box polypeptide 15 �1.7 1.2 �1.8 0.113361070 Xpo7 Exportin 7 �1.1 1.7 �1.8 0.113360840 Srgap2 SLIT-ROBO Rho GTPase activating protein 2 �1.4 1.3 �1.9 0.178

81747 Pi4kb Phosphatidylinositol 4-kinase, catalytic, beta �1.4 1.3 �1.9 0.17883792 Scd Stearoyl-CoA desaturase (delta-9-desaturase) �1.4 1.4 �2.0 0.113

311224 Ttc17 Tetratricopeptide repeat domain 17 �1.3 1.6 �2.0 0.113314850 Frs2 Fibroblast growth factor receptor substrate 2 �1.4 1.5 �2.1 0.162

29502 Slc20a2 Solute carrier family 20 (phosphate transporter), member 2 �1.7 1.6 �2.5 0.130312516 Gmcl1 Germ cell-less, spermatogenesis associated 1 �1.4 2.1 �2.8 0.121

24943 Plp1 Proteolipid protein 1 �1.3 2.6 �3.1 0.17865153 Ncs1 Neuronal calcium sensor 1 �1.1 3.5 �3.7 0.12178960 Prx Periaxin �2.5 2.1 �5.3 0.121



by 10.220.32.246 on Novem


Dow

nloaded from


62). Periaxin is a myelin-specific protein, suggesting develop-mental differences between the arteries in the degree of inner-vation, consistent with previous evidence of heterogeneousinnervation among skeletal muscle vascular beds (38). Simi-larly, pathway analysis revealed a network of genes related todevelopmental and structural differences between the vesselsin LETO rats (Fig. 5). The network includes molecules relatedto cellular assembly (actins, alpha catenin), vascular develop-ment (Notch), and myosin family genes (Mylpf, Mylk, Myh14,collectively represented by the Mlc node in Fig. 5). Thesedifferentially expressed genes may reflect the greater smoothmuscle content in the GFA compared with the SFA, which alsomay be of developmental origin. In addition to developmentaldifferences, it is important to note that the GFA and SFA differmarkedly with respect to local hemodynamic signals (e.g.,shear stress) owing to differences in soleus and gastrocnemiusmuscle recruitment patterns during standing and walking (i.e.,normal cage activity) in rats (32). As such, these local hemo-dynamic factors, which we and others have shown are impor-tant determinants of vascular function and phenotype (27, 28,

33, 49, 51, 53, 66, 74), could have contributed to differentialtranscriptomic profiles between GFA and SFA in LETO rats.

For the between-artery comparison in OLETF rats, weobserved some of the same intrinsic differences as discussedabove within LETO (e.g., Mab21l2 and Pck1). Thus it appearsthat some developmental/intrinsic differences between thesearteries are unaltered by obesity. Moreover, a major finding ofour study was that the number of genes differentially expressedbetween the vessels within OLETF was �2.5-fold greater thanthat within LETO, indicating greater heterogeneity betweenthese feed arteries in the obese OLETF rat (Fig. 4). The knownfunctions of selected genes listed in Table 5 suggest that theGFA takes on a more inflamed, proatherogenic phenotypecompared with the SFA within the OLETF rat, including Nnat,a novel marker of endothelial inflammation in diabetes (45),Cidec, which is involved in ectopic fat accumulation (47),Gpihbp1, an atherosclerosis-related gene involved in the trans-port of lipoprotein lipase into capillaries (11, 79), and Ccl11,also called eotaxin, an emerging inflammatory marker of car-diovascular diseases in humans (17). In addition to these novel

Interaction Network –(GFA OLETF/GFA LETO)/(SFA OLETF/SFA LETO)

Fig. 7. Top network of genes identified byinteraction analysis (score � 36; see text fordetails on how the interaction was computed).Nodes represent genes/molecules. Shading isin proportion to the size of the fold change(red, upregulation; green, downregulation).White nodes denote network members thatwere not represented in the RNA-seq database.Gray nodes denote network members that didnot reach FDR � 20%.



by 10.220.32.246 on Novem


Dow

nloaded from


genes, we found some “classic” indications of a proatherogenicphenotype in the GFA compared with the SFA of OLETF rats,such as an upregulation of Vcam1, an established marker ofinflamed vascular tissue, and a downregulation of Nos3, theendothelial isoform of nitric oxide synthase (SupplementalTable S4). Similarly, pathway analysis of genes differentiallyexpressed between GFA and SFA of OLETF rats revealed anetwork involving Nos3 and Akt, atheroprotective genes thatwere downregulated in GFA compared with SFA, as well asupregulations in a number of mitochondrial NADH transport-ers linked to mitochondrial complex 1 in the GFA (Fig. 6).Considering the large body of work relating reduced expres-sion and activity of Nos3 and increases in mitochondriallyderived superoxide production to vascular pathologies (64), itseems reasonable to interpret this top-scoring network as evi-dence of a preferential shift of the GFA toward a dysfunctional,prooxidant state compared with the SFA in the obese OLETFrat.

Effects of obesity on GFA and SFA. Both the GFA and theSFA displayed transcriptional alterations in response to hy-perphagia-induced obesity in the OLETF rat. Examination ofthe list of obesity-altered transcripts in the GFA (Table 2 andSupplemental Table S1) revealed a number of genes previouslyimplicated in the development of a dysfunctional/proathero-genic vascular cell phenotype. For example, several genes(Folr2, Hemgn, Pxmp4, Pla2g2a, Klra17, Ly49Q, and Oas1)have been reported to be involved in proinflammatory pro-cesses (20, 26, 36, 55, 76, 78), suggesting a more inflamedphenotype of GFA in OLETF compared with LETO. Notably,the gene Nnat, elevated by more than fivefold with obesity, wasrecently identified as a novel marker of endothelial inflamma-tion in diabetic mice (45). The gene Srebf2, another novelproatherogenic marker, contains an intronic region that en-codes miR33, which suppresses cholesterol transport, ulti-mately leading to increased risk for atherosclerosis (57). Re-

garding obesity-induced downregulated genes in the GFA(Table 2, bottom), several appear to be potent mediators ofantiatherogenic processes. For example, Abra is highly upregu-lated in growing collateral vessels, is responsive to nitric oxide,and is increased in response to atheroprotective shear stress(75). Hspa1a (also known as HSP70) and Hspb1 have well-established antiatherogenic effects, characterized as inhibitoryagainst inflammatory processes such as leukocyte adhesion(16, 54). Finally, the obesity-induced downregulation of com-plement component 2 (C2) in the GFA could be interpreted asa proatherogenic response to obesity, as mice deficient in C2exhibit significant atherosclerosis (67). Together, this profile oftop-responding genes provides evidence that the GFA is ad-versely affected by obesity at the transcriptomic level in theOLETF rat model.

Several of the top obesity-induced upregulated genes in theSFA were also upregulated in the GFA (Mis18a, Srebf2,Pxmp4, Fam151b, Folr2). Overall, the list of top differentiallyexpressed genes presented in Table 3 suggests the presence ofa proatherogenic effect on the SFA as well as the GFA. Forexample, Folr2, Vnn1, and Deaf1 have all been implicatedin atherosclerosis and inflammation (10, 26, 31). Interest-ingly, the SFA and the GFA shared some of the largest-magnitude obesity-induced downregulated genes, includingCyp2e1 and C2.

Pathway analysis (Figs. 2 and 3) revealed in both GFA andSFA that obesity significantly downregulated expression ofgene networks centered on Ubc. Ubc is a ubiquitin gene thatencodes a polyubiquitin precursor. The ubiquitin-proteasomesystem is an intracellular proteolysis system and plays a mul-tifaceted role in vascular cells (14, 71). In endothelial cells areduction in ubiquitin proteasome system activity reduces theactivity of endothelial nitric oxide synthase, and in smoothmuscle cells ubiquitin proteasome inhibition increases sensi-tivity to endoplasmic reticulum stress-mediated cell death and

Table 7. Top genes differentially expressed between OLETF and LETO in aortic endothelial


OLETF LETO in aortic ECs

246074 Scd1 Stearoyl-Coenzyme A desaturase 1 �0.001 14.7500336 Clec1b C-type lectin domain family 1, member B �0.001 6.6117033 Mmp12 Matrix metallopeptidase 12 0.001 6.5113882 Hemgn Hemogen 0.013 4.7282634 Pxmp4 Peroxisomal membrane protein 4 �0.001 4.525608 Lep Leptin 0.015 4.594270 Nnat Neuronatin 0.058 4.0

114511 Emb Embigin 0.001 4.029692 Pla2g2a Phospholipase A2, group IIA (platelets, synovial fluid) �0.001 3.824604 Npy Neuropeptide Y 0.017 3.8

293624 Irf7 Interferon regulatory factor 7 �0.001 3.5315675 Gldn Gliomedin �0.001 3.425424 Ctse Cathepsin E 0.031 3.4

288620 Cct6a chaperonin containing Tcp1, subunit 6A (zeta 1) �0.001 3.3

OLETF � LETO in aortic ECs

24179 Agt Angiotensinogen (serpin peptidase inhibitor, clade A, member 8) �0.001 �28.624186 Alb albumin �0.001 �19.6

307126 Mcm10 Minichromosome maintenance complex component 10 �0.001 �16.4362474 Adhfe1 Alcohol dehydrogenase, iron containing, 1 �0.001 �8.179131 Fabp3 Fatty acid binding protein 3, muscle and heart �0.001 �3.8

619561 Acsf2 acyl-CoA synthetase family member 2 �0.001 �3.3

EC, endothelial cell.



by 10.220.32.246 on Novem


Dow

nloaded from


suppresses the unfolded protein response (14, 71). On the otherhand, inhibition of the ubiquitin-proteasome system has alsobeen shown to reduce neointima thickness and reduce vascularinflammation in atherosclerosis models (14). Thus further in-vestigation into the role of the ubiquitin system and thefunctional significance of reduced expression of this system inSFA and GFA of obese rats is warranted.

Obesity � vessel interaction analysis. As the GFA and SFAhave been shown in the OLETF rat model to be differentiallysusceptible to obesity-induced endothelial dysfunction (3) anddisplay differences in insulin-induced vasomotor reactivity(28), we were particularly interested to know the genes forwhich there were differential effects of obesity between theGFA and the SFA in the present study. Our approach for

examining this was to identify those transcripts that displayedsignificantly different fold differences between the groups inone vessel compared with the other (Table 6). Interestingly,genes involved in endothelial cell permeability [Pla2g2a (76)]and inflammation [Mgst1 (68), Svil (44), Retstat (61), Cd14(72)] were found to be significantly upregulated by obesity inthe GFA but not the SFA. In contrast, notable genes upregu-lated in the SFA but not the GFA included novel markers suchas Cp, which is associated with incident mortality in heartfailure patients (9) and nephropathy onset in type 2 diabetespatients (25), as well as Scd and Dnajc10, both of which havebeen implicated in endoplasmic reticulum stress (8, 41). Fi-nally, the selective upregulation of three key neuronal genes(Plp1, Ncs1, and Prx) in the SFA suggests that neural inner-

Top Obesity Responsive GeneNetwork – Aortic EC Scrapes

Fig. 8. Top-scoring gene network influencedby obesity in aortic endothelial cells (ECs)(score � 38). Nodes represent genes/molecules.Shading is in proportion to the size of the foldchange (red, upregulation; green, downregula-tion). Gray nodes denote network membersthat were not represented on the array. Graynodes denote network members that did notreach FDR � 10%.



by 10.220.32.246 on Novem


Dow

nloaded from


vation or perhaps neural control may be differentially alteredbetween GFA and SFA with obesity, a hypothesis that warrantsfuture research.

Pathway analysis of the gene list presented in Table 6revealed a top-scoring network centered on CD14, with links toproinflammatory signaling molecules such as Jnk, interleukins1 and 12, interferon alpha, P38MAPK, and nuclear factor-�B(Fig. 7). Interestingly, CD14 is widely recognized as a cellsurface antigen on circulating monocytes but has also been de-tected in vascular tissues (72). As part of the receptor complex(along with toll-like receptor 4) for lipopolysaccharide, CD14 isupstream of a host of intracellular proinflammatory events. Ourpresent findings lead us to propose that CD14 and the relateddownstream inflammatory signaling pathways identified in Fig. 7may play roles in the differential effects of obesity between theGFA and the SFA in the OLETF rat model.

Effects of obesity on aortic endothelial cells. Recognizingthat the GFA and SFA analyses were performed on wholevessel homogenates, we also sought to determine the influenceof obesity on a purified population of endothelial cells obtainedfrom aortic scrapes. Overall, the genes identified as differen-tially expressed between LETO and OLETF aortic endothelialcells were consistent with the hypothesized proatherogeniceffects of obesity. Some genes identified as being upregulatedin GFA and/or SFA also appear in Table 7, e.g., Hemgn,Pxmp4, Pla2g2a, and Nnat. Additionally, a number of obesity-induced transcripts in aortic endothelial cells have been previouslyimplicated in proatherogenic processes, including Scd1, previ-ously reported to be elevated in mesenteric arteries in a mousemodel of atherosclerosis (65); Clec1b, a ligand for activation ofinflammatory cells (specifically natural killer cells) (81); Mmp12and leptin, which have well-established roles in atherosclerosisand whose expression has been shown to be upregulated togetherin human abdominal aortic aneurism specimens (73); Irf7, previ-ously found to be a central gene in an IPA network of particulatematter-induced inflammation in endothelial cells (18); Npy, apotent vasoconstrictor involved in pathological vascular remodel-ing (1); and Cct6a, a chaperone protein known to be induced byendoplasmic reticulum stress (15).

The top-scoring gene network in aortic endothelial cells wassuggestive of an obesity-induced shift toward a proinflamma-tory phenotype, as reflected by alterations in a number ofmolecules that converge on the nuclear factor-�B complex(e.g., obesity-induced increases in Stat2, interferon signalingmolecules, and Cxcl16 and a decrease in Sod3). This networktherefore suggests that obesity confers a proatherogenic phe-notype in endothelial cells of large arteries (i.e., the aorta) inaddition to the deleterious effects on gene expression observedin whole vessel homogenates of skeletal muscle feed arteries.

Limitations. A few limitations of our study warrant men-tioning. Regarding our experimental model, the hyperphagicOLETF rat not only is obese but also displays a number ofobesity-related comorbidities (insulin resistance, hypertension,fatty liver disease, etc.) that progressively increase in severityas the animals advance in age. We also did not determinewhether OLETF and LETO rats differed in terms of daily energyexpenditure or cage activity. Thus we cannot rule out that thedifferential gene expression profiles reported here might be par-tially attributable to comorbidities or physical activity. Addition-ally, our aortic endothelial cell gene expression results should not

be interpreted as being representative of endothelial cells from alllarge and small arteries. The aorta was chosen because it is thelargest artery and therefore the most convenient sample for theenrichment of arterial endothelial cells.

Conclusions. In summary, our study indicates that both theGFA and the SFA are susceptible to unfavorable effects ofobesity at the transcriptional level in the OLETF rat model.The SFA displayed a greater number of genes differentiallyexpressed between LETO and OLETF rats, which, in light ofprevious evidence indicating that the SFA largely maintains itsfunctional properties in obese animal models, suggests that theSFA may be more “plastic” in the face of a proatherogenicobese systemic environment. Despite differences in the numberof genes found to be differentially expressed, examination ofthe known functions of genes differentially expressed betweenlean and obese rats indicated a more pronounced shift towarda proinflammatory/prooxidant phenotype in the GFA com-pared with the SFA in obese rats, a finding consistent with ourpreviously published finding that the GFA displays greatersusceptibility to obesity-induced endothelial dysfunction (3).Our study is also consistent with the large body of researchindicating that obesity confers a proatherogenic phenotype inarterial endothelial cells; importantly, our data are the firstRNA-Seq-based characterization of these effects. Finally, it isimportant to emphasize that the majority of differentiallyexpressed genes reported here have little or no known functionin vascular tissues. Therefore, our findings should aid ingenerating new hypotheses that will lead to novel mechanisticinsights into the negative consequences of obesity on skeletalmuscle resistance arteries.

ACKNOWLEDGMENTS

We thank Nicholas Fleming, Eric Gibson, Kelcie Tacchi, and Matt Briel-maier for assisting in the care of the rats and exercise training. Sean Blake(Global Biologics, LLC) performed the RNA extractions and generated theRNA libraries that were submitted to the University of Missouri DNA CoreFacility for high-throughput sequencing services.

GRANTS

This work was supported by National Institutes of Health (NIH) GrantsRO1-HL-036088 (M. H. Laughlin and J. W. Davis) and T32-AR-048523(N. T. Jenkins and J. S. Martin) and Department of Veterans Affairs GrantVHA-CDA2 1299-02 (R. S. Rector). This work was also supported in part withresources and the use of facilities at the Harry S Truman Memorial VeteransHospital in Columbia, MO.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: N.T.J., J.P., R.S.R., J.W.D., and M.H.L. conceptionand design of research; N.T.J., J.P., P.K.T., J.S.M., J.W.D., and M.H.L.performed experiments; N.T.J., J.P., and J.W.D. analyzed data; N.T.J., J.P.,R.S.R., J.W.D., and M.H.L. interpreted results of experiments; N.T.J. andJ.W.D. prepared figures; N.T.J. drafted manuscript; N.T.J., J.P., P.K.T., J.S.M.,R.S.R., J.W.D., and M.H.L. edited and revised manuscript; N.T.J., J.P., P.K.T.,J.S.M., R.S.R., J.W.D., and M.H.L. approved final version of manuscript.

REFERENCES

1. Abe K, Tilan JU, Zukowska Z. NPY and NPY receptors in vascularremodeling. Curr Top Med Chem 7: 1704–1709, 2007.

2. Bakker W, Eringa EC, Sipkema P, van Hinsbergh VW. Endothelialdysfunction and diabetes: roles of hyperglycemia, impaired insulin sig-naling and obesity. Cell Tissue Res 335: 165–189, 2009.



by 10.220.32.246 on Novem


Dow

nloaded from


3. Bender SB, Newcomer SC, Laughlin MH. Differential vulnerability ofskeletal muscle feed arteries to dysfunction in insulin resistance: impact offiber type and daily activity. Am J Physiol Heart Circ Physiol 300:H1434–H1441, 2011.

4. Benjamini Y, Hochberg Y. Controlling the false discovery rate: apractical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289–300, 1995.

5. Bourgon R, Gentleman R, Huber W. Independent filtering increasesdetection power for high-throughput experiments. Proc Natl Acad Sci USA107: 9546–9551, 2010.

6. Bunker AK, Arce-Esquivel AA, Rector RS, Booth FW, Ibdah JA,Laughlin MH. Physical activity maintains aortic endothelium-dependentrelaxation in the obese type 2 diabetic OLETF rat. Am J Physiol HeartCirc Physiol 298: H1889–H1901, 2010.

7. Calvano SE, Xiao W, Richards DR, Felciano RM, Baker HV, Cho RJ,Chen RO, Brownstein BH, Cobb JP, Tschoeke SK, Miller-GrazianoC, Moldawer LL, Mindrinos MN, Davis RW, Tompkins RG, LowrySF. A network-based analysis of systemic inflammation in humans.Nature 437: 1032–1037, 2005.

8. Cunnea PM, Miranda-Vizuete A, Bertoli G, Simmen T, Damdimo-poulos AE, Hermann S, Leinonen S, Huikko MP, Gustafsson JA, SitiaR, Spyrou G. ERdj5, an endoplasmic reticulum (ER)-resident proteincontaining DnaJ and thioredoxin domains, is expressed in secretory cellsor following ER stress. J Biol Chem 278: 1059–1066, 2003.

9. Dadu RT, Dodge R, Nambi V, Virani SS, Hoogeveen RC, Smith NL,Chen F, Pankow JS, Guild C, Tang WH, Boerwinkle E, Hazen SL,Ballantyne CM. Ceruloplasmin and heart failure in the AtherosclerosisRisk in Communities (ARIC) Study. Circ Heart Fail 6: 936–943, 2013.

10. Dammanahalli KJ, Stevens S, Terkeltaub R. Vanin-1 pantetheinasedrives smooth muscle cell activation in post-arterial injury neointimalhyperplasia. PLoS One 7: e39106, 2012.

11. Davies BS, Beigneux AP, Barnes RH, Tu Y 2nd, Gin P, WeinsteinMM, Nobumori C, Nyren R, Goldberg I, Olivecrona G, Bensadoun A,Young SG, Fong LG. GPIHBP1 is responsible for the entry of lipoproteinlipase into capillaries. Cell Metab 12: 42–52, 2010.

12. De Boer MP, Meijer RI, Wijnstok NJ, Jonk AM, Houben AJ, Stehou-wer CD, Smulders YM, Eringa EC, Serne EH. Microvascular dysfunc-tion: a potential mechanism in the pathogenesis of obesity-associatedinsulin resistance and hypertension. Microcirculation 19: 5–18, 2012.

13. de Jonge HJ, Fehrmann RS, de Bont ES, Hofstra RM, Gerbens F,Kamps WA, de Vries EG, van der Zee AG, te Meerman GJ, ter ElstA. Evidence based selection of housekeeping genes. PLoS One 2: e898,2007.

14. Demasi M, Laurindo FR. Physiological and pathological role of theubiquitin-proteasome system in the vascular smooth muscle cell. Cardio-vasc Res 95: 183–193, 2012.

15. Dombroski BA, Nayak RR, Ewens KG, Ankener W, Cheung VG,Spielman RS. Gene expression and genetic variation in response toendoplasmic reticulum stress in human cells. Am J Hum Genet 86:719–729, 2010.

16. Dulin E, Garcia-Barreno P, Guisasola MC. Genetic variations ofHSPA1A, the heat shock protein levels, and risk of atherosclerosis. CellStress Chaperones 17: 507–516, 2012.

17. Emanuele E, Falcone C, D’Angelo A, Minoretti P, Buzzi MP, BertonaM, Geroldi D. Association of plasma eotaxin levels with the presence andextent of angiographic coronary artery disease. Atherosclerosis 186: 140–145, 2006.

18. Erdely A, Antonini JM, Salmen-Muniz R, Liston A, Hulderman T,Simeonova PP, Kashon ML, Li S, Gu JK, Stone S, Chen BT, FrazerDG, Zeidler-Erdely PC. Type I interferon and pattern recognition recep-tor signaling following particulate matter inhalation. Particle Fibre Toxi-col 9: 25, 2012.

19. Eringa EC, Serne EH, Meijer RI, Schalkwijk CG, Houben AJ,Stehouwer CD, Smulders YM, van Hinsbergh VW. Endothelial dys-function in (pre)diabetes: characteristics, causative mechanisms andpathogenic role in type 2 diabetes. Rev Endocr Metab Disord 14: 39–48,2013.

20. Fletcher JM, Jordan MA, Snelgrove SL, Slattery RM, Dufour FD,Kyparissoudis K, Besra GS, Godfrey DI, Baxter AG. Congenic analysisof the NKT cell control gene Nkt2 implicates the peroxisomal proteinPxmp4. J Immunol 181: 3400–3412, 2008.

21. Frisbee JC. Obesity, insulin resistance, and microvessel density. Micro-circulation 14: 289–298, 2007.

22. Frisbee JC. Vascular dysfunction in obesity and insulin resistance. Mi-crocirculation 14: 269–271, 2007.

23. Frisbee JC, Delp MD. Vascular function in the metabolic syndrome andthe effects on skeletal muscle perfusion: lessons from the obese Zucker rat.Essays Biochem 42: 145–161, 2006.

24. Goligorsky MS. Microvascular rarefaction: the decline and fall of bloodvessels. Organogenesis 6: 1–10, 2010.

25. Hellemons ME, Kerschbaum J, Bakker SJ, Neuwirt H, Mayer B,Mayer G, de Zeeuw D, Lambers Heerspink HJ, Rudnicki M. Validityof biomarkers predicting onset or progression of nephropathy in patientswith Type 2 diabetes: a systematic review. Diabet Med 29: 567–577, 2012.

26. Jager NA, Westra J, van Dam GM, Teteloshvili N, Tio RA, Breek JC,Slart RH, Boersma H, Low PS, Bijl M, Zeebregts CJ. Targeted folatereceptor beta fluorescence imaging as a measure of inflammation toestimate vulnerability within human atherosclerotic carotid plaque. J NuclMed 53: 1222–1229, 2012.

27. Jenkins NT, Padilla J, Boyle LJ, Credeur DP, Laughlin MH, Fadel PJ.Disturbed blood flow acutely induces activation and apoptosis of thehuman vascular endothelium. Hypertension 61: 615–621, 2013.

28. Jenkins NT, Padilla J, Martin JS, Crissey JM, Thyfault JP, Rector S,Laughlin MH. Differential vasomotor effects of insulin on gastrocnemiusand soleus feed arteries in the OLETF rat model: role of endothelin-1. ExpPhysiol 99: 267–271, 2014.

29. Kan OM, Takeya R, Abe T, Kitajima N, Nishida M, Tominaga R,Kurose H, Sumimoto H. Mammalian formin Fhod3 plays an essentialrole in cardiogenesis by organizing myofibrillogenesis. Biol Open 1:889–896, 2012.

30. Kawano K, Hirashima T, Mori S, Saitoh Y, Kurosumi M, Natori T.Spontaneous long-term hyperglycemic rat with diabetic complications.Otsuka Long-Evans Tokushima Fatty (OLETF) strain. Diabetes 41: 1422–1428, 1992.

31. Kuttenkeuler D, Pelte N, Ragab A, Gesellchen V, Schneider L, BlassC, Axelsson E, Huber W, Boutros M. A large-scale RNAi screenidentifies Deaf1 as a regulator of innate immune responses in Drosophila.J Innate Immun 2: 181–194, 2010.

32. Laughlin MH, Armstrong RB. Muscular blood flow distribution patternsas a function of running speed in rats. Am J Physiol Heart Circ Physiol243: H296–H306, 1982.

33. Laughlin MH, Newcomer SC, Bender SB. Importance of hemodynamicforces as signals for exercise-induced changes in endothelial cell pheno-type. J Appl Physiol 104: 588–600, 2008.

34. Laughlin MH, Woodman CR, Schrage WG, Gute D, Price EM.Interval sprint training enhances endothelial function and eNOS content insome arteries that perfuse white gastrocnemius muscle. J Appl Physiol 96:233–244, 2004.

35. Law CW, Chen Y, Shi W, Smyth GK. Voom! Precision Weights UnlockLinear Model Analysis Tools for RNA-seq Read Counts, Technical Report1 May 2013. Melbourne, Australia: Bioinformatics Division, Walter andEliza Hall Institute of Medical Research, http://www.statsci.org/smyth/pubs/VoomPreprint.pdf, 2013.

36. Li CY, Zhan YQ, Xu CW, Xu WX, Wang SY, Lv J, Zhou Y, Yue PB,Chen B, Yang XM. EDAG regulates the proliferation and differentiationof hematopoietic cells and resists cell apoptosis through the activation ofnuclear factor-kappa B. Cell Death Differ 11: 1299–1308, 2004.

37. Liao Y, Smyth GK, Shi W. The Subread aligner: fast, accurate andscalable read mapping by seed-and-vote. Nucleic Acids Res 41: e108,2013.

38. Long JB, Segal SS. Quantifying perivascular sympathetic innervation:regional differences in male C57BL/6 mice at 3 and 20 months. J NeurosciMethods 184: 124–128, 2009.

39. Martin JS, Padilla J, Jenkins NT, Crissey JM, Bender SB, Rector RS,Thyfault JP, Laughlin MH. Functional adaptations in the skeletal musclemicrovasculature to endurance and interval sprint training in the type 2diabetic OLETF rat. J Appl Physiol 113: 1223–1232, 2012.

40. Martin M. Cutadapt removes adapter sequences from high-throughputsequencing reads. EMBnet J 17: 10–12, 2011.

41. Masuda M, Ting TC, Levi M, Saunders SJ, Miyazaki-Anzai S, Mi-yazaki M. Activating transcription factor 4 regulates stearate-inducedvascular calcification. J Lipid Res 53: 1543–1552, 2012.

42. McAllister RM, Jasperse JL, Laughlin MH. Nonuniform effects ofendurance exercise training on vasodilation in rat skeletal muscle. J ApplPhysiol 98: 753–761, 2005.

43. Mikus CR, Rector RS, Arce-Esquivel AA, Libla JL, Booth FW, IbdahJA, Laughlin MH, Thyfault JP. Daily physical activity enhances reac-



by 10.220.32.246 on Novem


Dow

nloaded from

http://www.statsci.org/smyth/pubs/VoomPreprint.pdf

http://www.statsci.org/smyth/pubs/VoomPreprint.pdf


tivity to insulin in skeletal muscle arterioles of hyperphagic OtsukaLong-Evans Tokushima Fatty rats. J Appl Physiol 109: 1203–1210, 2010.

44. Morozumi T, Kubota T, Sugita N, Ohsawa Y, Yamazaki K, Yoshie H.Elevated mRNA expression for supervillin and vascular endothelialgrowth factor in human neutrophils stimulated with lipopolysaccharidefrom Porphyromonas gingivalis. J Periodontal Res 36: 160–168, 2001.

45. Mzhavia N, Yu S, Ikeda S, Chu TT, Goldberg I, Dansky HM.Neuronatin: a new inflammation gene expressed on the aortic endotheliumof diabetic mice. Diabetes 57: 2774–2783, 2008.

46. Nakagawa O, Arnold M, Nakagawa M, Hamada H, Shelton JM,Kusano H, Harris TM, Childs G, Campbell KP, Richardson JA,Nishino I, Olson EN. Centronuclear myopathy in mice lacking a novelmuscle-specific protein kinase transcriptionally regulated by MEF2. GenesDev 19: 2066–2077, 2005.

47. Nishino N, Tamori Y, Tateya S, Kawaguchi T, Shibakusa T, Miz-unoya W, Inoue K, Kitazawa R, Kitazawa S, Matsuki Y, Hiramatsu R,Masubuchi S, Omachi A, Kimura K, Saito M, Amo T, Ohta S,Yamaguchi T, Osumi T, Cheng J, Fujimoto T, Nakao H, Nakao K,Aiba A, Okamura H, Fushiki T, Kasuga M. FSP27 contributes toefficient energy storage in murine white adipocytes by promoting the forma-tion of unilocular lipid droplets. J Clin Invest 118: 2808–2821, 2008.

48. Ozsolak F, Milos PM. RNA sequencing: advances, challenges andopportunities. Nat Rev Genet 12: 87–98, 2011.

49. Padilla J, Jenkins NT, Roberts MD, Arce-Esquivel AA, Martin JS,Laughlin MH, Booth FW. Differential changes in vascular mRNA levelsbetween rat iliac and renal arteries produced by cessation of voluntaryrunning. Exp Physiol 98: 337–347, 2013.

49a.Padilla J, Jenkins NT, Thorne PK, Martin JS, Rector RS, Davis JW,Laughlin MH. Transcriptome-wide RNA sequencing analysis of rat skeletalmuscle feed arteries. II. Impact of exercise training in obesity. J Appl Physiol.First published January 9, 2014; doi:10.1152/japplphysiol.01234.2013.

50. Padilla J, Newcomer SC, Simmons GH, Kreutzer KV, Laughlin MH.Long-term exercise training does not alter brachial and femoral arteryvasomotor function and endothelial phenotype in healthy pigs. Am JPhysiol Heart Circ Physiol 299: H379–H385, 2010.

51. Padilla J, Simmons GH, Bender SB, Arce-Esquivel AA, Whyte JJ,Laughlin MH. Vascular effects of exercise: endothelial adaptations be-yond active muscle beds. Physiology 26: 132–145, 2011.

52. Padilla J, Simmons GH, Davis JW, Whyte JJ, Zderic TW, HamiltonMT, Bowles DK, Laughlin MH. Impact of exercise training on endothe-lial transcriptional profiles in healthy swine: a genome-wide microarrayanalysis. Am J Physiol Heart Circ Physiol 301: H555–H564, 2011.

53. Padilla J, Simmons GH, Fadel PJ, Laughlin MH, Joyner MJ, CaseyDP. Impact of aging on conduit artery retrograde and oscillatory shear at restand during exercise: role of nitric oxide. Hypertension 57: 484–489, 2011.

54. Pockley AG, Calderwood SK, Multhoff G. The atheroprotective prop-erties of Hsp70: a role for Hsp70-endothelial interactions? Cell StressChaperones 14: 545–553, 2009.

55. Rahim MM, Tai LH, Troke AD, Mahmoud AB, Abou-Samra E, RoyJG, Mottashed A, Ault N, Corbeil C, Goulet ML, Zein HS, Hamilton-Valensky M, Krystal G, Kerr WG, Toyama-Sorimachi N, Makrigi-annis AP. Ly49Q positively regulates type I IFN production by plasma-cytoid dendritic cells in an immunoreceptor tyrosine-based inhibitorymotif-dependent manner. J Immunol 190: 3994–4004, 2013.

56. Rapaport F, Khanin R, Liang Y, Pirun M, Krek A, Zumbo P, MasonCE, Socci ND, Betel D. Comprehensive evaluation of differential geneexpression analysis methods for RNA-seq data. Genome Biol 14: R95,2013.

57. Rayner KJ, Sheedy FJ, Esau CC, Hussain FN, Temel RE, ParathathS, van Gils JM, Rayner AJ, Chang AN, Suarez Y, Fernandez-Hernando C, Fisher EA, Moore KJ. Antagonism of miR-33 in micepromotes reverse cholesterol transport and regression of atherosclerosis. JClin Invest 121: 2921–2931, 2011.

58. Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductorpackage for differential expression analysis of digital gene expressiondata. Bioinformatics 26: 139–140, 2010.

59. Robinson MD, Oshlack A. A scaling normalization method for differ-ential expression analysis of RNA-seq data. Genome Biol 11: R25, 2010.

60. Saito Y, Kojima T, Takahashi N. Mab21l2 is essential for embryonicheart and liver development. PLoS One 7: e32991, 2012.

61. Sarang Z, Garabuczi E, Joos G, Kiss B, Toth K, Ruhl R, Szondy Z.Macrophages engulfing apoptotic thymocytes produce retinoids to pro-mote selection, differentiation, removal and replacement of double posi-tive thymocytes. Immunobiology 218: 1354–1360, 2013.

62. Sarjeant JM, Lawrie A, Kinnear C, Yablonsky S, Leung W, MassaeliH, Prichett W, Veinot JP, Rassart E, Rabinovitch M. Apolipoprotein Dinhibits platelet-derived growth factor-BB-induced vascular smooth mus-cle cell proliferated by preventing translocation of phosphorylated extra-cellular signal regulated kinase 1/2 to the nucleus. Arterioscler ThrombVasc Biol 23: 2172–2177, 2003.

63. Serne EH, de Jongh RT, Eringa EC, IJzerman RG, Stehouwer CD.Microvascular dysfunction: a potential pathophysiological role in themetabolic syndrome. Hypertension 50: 204–211, 2007.

64. Shen GX. Oxidative stress and diabetic cardiovascular disorders: roles ofmitochondria and NADPH oxidase. Can J Physiol Pharmacol 88: 241–248, 2010.

65. Simeone SM, Li MW, Paradis P, Schiffrin EL. Vascular gene expres-sion in mice overexpressing human endothelin-1 targeted to the endothe-lium. Physiol Genomics 43: 148–160, 2011.

66. Simmons GH, Padilla J, Laughlin MH. Heterogeneity of endothelial cellphenotype within and amongst conduit vessels of the swine vasculature.Exp Physiol 97: 1074–1082, 2012.

67. Sjoholm AG, Jonsson G, Braconier JH, Sturfelt G, Truedsson L.Complement deficiency and disease: an update. Mol Immunol 43: 78–85,2006.

68. Sjostrom M, Jakobsson PJ, Heimburger M, Palmblad J, HaeggstromJZ. Human umbilical vein endothelial cells generate leukotriene C4 viamicrosomal glutathione S-transferase type 2 and express the CysLT(1)receptor. Eur J Biochem 268: 2578–2586, 2001.

69. Smyth GK. Linear models and empirical Bayes methods for assessingdifferential expression in microarray experiments. Stat Appl Genet MolBiol 3: Article 3, 2004.

70. Srere PA. Citrate synthase. Methods Enzymol 13: 3–5, 1969.71. Stangl K, Stangl V. The ubiquitin-proteasome pathway and endothelial

(dys)function. Cardiovasc Res 85: 281–290, 2010.72. Su AI, Wiltshire T, Batalov S, Lapp H, Ching KA, Block D, Zhang J,

Soden R, Hayakawa M, Kreiman G, Cooke MP, Walker JR, Hogen-esch JB. A gene atlas of the mouse and human protein-encoding tran-scriptomes. Proc Natl Acad Sci USA 101: 6062–6067, 2004.

73. Tao M, Yu P, Nguyen BT, Mizrahi B, Savion N, Kolodgie FD, VirmaniR, Hao S, Ozaki CK, Schneiderman J. Locally applied leptin inducesregional aortic wall degeneration preceding aneurysm formation in apoli-poprotein E-deficient mice. Arterioscler Thromb Vasc Biol 33: 311–320,2013.

74. Thijssen DH, Dawson EA, Tinken TM, Cable NT, Green DJ. Retro-grade flow and shear rate acutely impair endothelial function in humans.Hypertension 53: 986–992, 2009.

75. Troidl K, Ruding I, Cai WJ, Mucke Y, Grossekettler L, Piotrowska I,Apfelbeck H, Schierling W, Volger OL, Horrevoets AJ, Grote K,Schmitz-Rixen T, Schaper W, Troidl C. Actin-binding rho activatingprotein (Abra) is essential for fluid shear stress-induced arteriogenesis.Arterioscler Thromb Vasc Biol 29: 2093–2101, 2009.

76. Wang G, Qian P, Xu Z, Zhang J, Wang Y, Cheng S, Cai W, Qian G,Wang C, Decoster MA. Regulatory effects of the JAK3/STAT1 pathwayon the release of secreted phospholipase A2-IIA in microvascular endo-thelial cells of the injured brain. J Neuroinflamm 9: 170, 2012.

77. Wang Z, Gerstein M, Snyder M. RNA-Seq: a revolutionary tool fortranscriptomics. Nat Rev Genet 10: 57–63, 2009.

78. Warke RV, Xhaja K, Martin KJ, Fournier MF, Shaw SK, Brizuela N,de Bosch N, Lapointe D, Ennis FA, Rothman AL, Bosch I. Denguevirus induces novel changes in gene expression of human umbilical veinendothelial cells. J Virol 77: 11822–11832, 2003.

79. Weinstein MM, Yin L, Tu Y, Wang X, Wu X, Castellani LW, WalzemRL, Lusis AJ, Fong LG, Beigneux AP, Young SG. Chylomicronemiaelicits atherosclerosis in mice—brief report. Arterioscler Thromb VascBiol 30: 20–23, 2010.

80. Williams DA, Segal SS. Feed artery role in blood flow control to rathindlimb skeletal muscles. J Physiol 463: 631–646, 1993.

81. Williams KJ, Wilson E, Davidson CL, Aguilar OA, Fu L, Carlyle JR,Burshtyn DN. Poxvirus infection-associated downregulation of C-typelectin-related-b prevents NK cell inhibition by NK receptor protein-1B. JImmunol 188: 4980–4991, 2012.

82. Woodman CR, Price EM, Laughlin MH. Selected Contribution: Agingimpairs nitric oxide and prostacyclin mediation of endothelium-dependentdilation in soleus feed arteries. J Appl Physiol 95: 2164–2170, 2003.

83. Woodman CR, Price EM, Laughlin MH. Shear stress induces eNOSmRNA expression and improves endothelium-dependent dilation in se-nescent soleus muscle feed arteries. J Appl Physiol 98: 940–946, 2005.



by 10.220.32.246 on Novem


Dow

nloaded from

http://dx.doi.org/10.1152/japplphysiol.01234.2013


Transcriptome-wide RNA sequencing analysis of rat skeletal … · 2017-11-16 · Transcriptome-wide...

Documents

Transcript of Transcriptome-wide RNA sequencing analysis of rat skeletal … · 2017-11-16 · Transcriptome-wide...