buttner - JAP 07 - microarray

8/3/2019 buttner - JAP 07 - microarray

1/12


2/12

Exercise affects the gene expression profiles of human white blood cells

Petra Buttner,1 Sandy Mosig,1 Anja Lechtermann,2 Harald Funke,1 and Frank C. Mooren3

1 Institute of Vascular Biology and Medicine, Friedrich-Schiller-University Jena, Jena;2Institute of Sports Medicine, University Hospital Munster, Munster; and 3 Department of

Sports Medicine, Institute of Sports Sciences, Justus-Liebig-University, Giessen, Germany

Submitted 19 January 2006; accepted in final form 5 September 2006

Buttner P, Mosig S, Lechtermann A, Funke H, Mooren FC.Exercise affects the gene expression profiles of human white bloodcells. J Appl Physiol 102: 2636, 2007. First published September 21,2006; doi:10.1152/japplphysiol.00066.2006.White blood cells(WBCs) express tens of thousands of genes, whose expression levelsare modified by genetic and external factors. The purpose of thepresent study was to investigate the effects of acute exercise on geneexpression profiles (GEPs) of WBCs and to identify suitable genesthat may serve as surrogate markers for monitoring exercise andtraining load. Five male participants performed an exhaustive tread-mill test (ET) at 80% of their maximal O2 uptake (VO2 max) and amoderate treadmill test (MT) at 60% VO2 max for exactly the same

time 2 wk later. WBCs were isolated by the erythrocyte lysismethod. GEPs were measured using the Affymetrix GeneChip tech-nology. After scaling, normalization, and filtering, groupwise com-parisons of gene expression intensities were performed, and severalmeasurements were validated by real-time PCR. We found 450 genesupregulated and 150 downregulated (1.5-fold change; ANOVA withBenjamini-Hochberg correction, P 0.05) after ET that were closelyassociated with the gene ontology lists response to stress andinflammatory response. Analysis of mean expression levels afterMT showed that the extent of up- and downregulation was workloaddependent. The genes for the stress (heat shock) proteins HSPA1Aand HSPH1 and for the matrix metalloproteinase MMP-9 showed themost prominent increases, whereas the YES1 oncogene (YES1) andCD160 (BY55) were most strongly reduced. Despite different meth-odological approaches used, the consistency of our results with the

expression data of another study (Connolly PH, Caiozzo VJ, ZaldivarF, Nemet D, Larson J, Hung SP, Heck JD, Hatfield GW, Cooper DM.

J Appl Physiol 97: 14611469, 2004) suggests that expression finger-prints are useful tools for monitoring exercise and training loads andthereby help to avoid training-associated health risks.

inflammation; stress response; microarrays; surrogate marker

REGULAR EXERCISE is of great importance for public health. It hasbeen shown to be effective in the prevention and therapy ofmany widespread diseases in Western civilizations (2, 3, 34),including hypertension, coronary heart disease, and diabetes.

Many aspects of how the prophylactic and therapeutic ef-fects of exercise on chronic diseases are mediated remain

unclear (34, 36). One important way to improve our under-standing of these beneficial effects is the investigation of thecellular and molecular responses to exercise. The use of themicroarray technology thereby enables us to investigate thecomplete gene response to an external factor like exercise (4,8). Since exercise has been shown to be an importantregulator of immune cells and their functions, we havechosen white blood cells (WBCs) as target cells (39). There

is evidence that the exercise stress evokes an inflammation-like reaction of the immune system with the activation of bothproinflammatory and anti-inflammatory pathways (31). Thebalance of both is supposed to be dependent on exerciseintensity and duration. Therefore, acute exhaustive exercise isexpected to transiently decrease the individuals immune com-petence, while moderate exercise has an anti-inflammatoryeffect with improved anti-infectious capabilities (29). More-over, there is growing evidence that the immune system mayserve as an important physiological indicator for a personsindividual ability to recover from workload stresses. This has

led to the hypothesis that the overtraining syndrome, a condi-tion of long-term decrement in performance capacity despitecontinuous training loads, is based on a derangement of cellu-lar immune regulation (18).

Besides the study of underlying signaling mechanisms, thefuture use of microarray technology in exercise physiologymight be of special value for monitoring the athletes trainingprocess. One prerequisite for this intention would be theidentification of robust exercise-induced gene expression pro-files or fingerprints that can be related to exercise intensity ortype, etc. However, the available data in the literature showvery heterogeneous results most likely due to different analyt-ical platforms and experimental procedures (4, 8, 14, 45, 46).

Together, this prompted us to hypothesize that the impact of

the stress factor exercise may be mirrored in the gene expres-sion profiles of leukocytes. We were interested to see whetherexpression fingerprints in these cells reflect exercise intensityand whether a set of exercise-regulated genes could be identi-fied. To address these questions, we examined the transcrip-tional response of the same individuals after moderate andexhaustive treadmill tests. Eventually, results from such studiescould help to improve the design of individualized exerciseprograms.

MATERIALS AND METHODS

Subjects and experimental design. Five male subjects of the Uni-versity of Munsters student population were informed about the

nature, purpose, and potential risks of the study and signed aninformed consent statement approved by the University of MuensterEthics Committee. The anthropometric data of the group were asfollows (means and SD): height 180.0 cm (SD 4.6); weight 71.8 kg(SD 4.0); age 25.4 yr (SD 3.5). Moreover, the subjects participatedpredominantly in leisure time sports such as football and running witha mean training time of6.0 h/wk (SD 2.6). After a general medicalcheckup the subjects were first tested for their maximal oxygen uptake(VO2 max) during a continuous, progressive exercise test on a treadmillergometer (Ergo XELG90 Spezial, Woodway, Weil am Rhein, Ger-

Address for reprint requests and other correspondence: F. C. Mooren, Dept.of Sports Medicine, Institute of Sports Sciences, Justus-Liebig-Univ., Kugel-berg 62, 35394 Giessen, Germany (e-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

J Appl Physiol 102: 2636, 2007.First published September 21, 2006; doi:10.1152/japplphysiol.00066.2006.

8750-7587/07 $8.00 Copyright 2007 the American Physiological Society http://www.jap.org26


3/12

many). The initial velocity was 8 km/h, increasing every 3 min by 2km/h. Respiration parameters were analyzed using Quark b2(Cosmed, Rome, Italy). The mean VO2 max (SD) of the subjects was60.6 ml min1 kg1 (SD 5.3); their mean peak power output wascalculated in accordance to the Magaria formula to be 349.0 W (SD34.4).

About 2 wk later, the participants performed a strenuous treadmillexercise test (ET) at an intensity corresponding to 80% of their

individual VO2 max until exhaustion [mean velocity 12.7 km/h (SD1.7)]. Another 12 wk later the subjects performed a second exercisetest at an intensity corresponding to 60% of VO2 max (moderatetreadmill exercise, MT) for a time identical to that of their strenuousexercise test [mean velocity 9.4 km/h (SD 1.2)]. For both tests bloodsamples were taken before, after, and 1 h after exercise. Bloodsamples for gene expression measurements were prepared immedi-ately following blood withdrawal. RNA isolation and microarraymeasurements were performed only on preexercise samples and thesamples taken 1 h after exercise.

Leukocyte counts and lactate concentration. Blood cell counts,hemoglobin, and hematocrit determinations were performed using asemi-automated hematology analyzer (F-820, Sysmex, Norderstedt,Germany).

Lactate concentration in capillary blood was measured by a pho-

tometric method using a commercially available kit (EKF Diagnostic,Magdeburg, Germany).

RNA, cDNA, and cRNA preparation. WBCs were isolated from 9ml EDTA-blood using the erythrocyte lysis buffer of the QIAampRNA Blood Mini Kit (Qiagen, Hilden, Germany) followed by twowashing steps in PBS containing 2 mmol/l EDTA. Buffers were usedat 4C; tubes were placed on ice. The resulting cell pellet wasresuspended in 20 l ice-cold PBS-EDTA and transferred into 900 lTrizol (Invitrogen, Karlsruhe, Germany) at room temperature wherecells were homogenized by repeated pipetting. RNA samples can bestored in Trizol for weeks (20C) without losses in quality orquantity. In this study the samples were prepared within 2 days afterlysis of the WBCs. RNA was extracted using the phase-lock gel tubemethod (Phase Lock Gel Heavy 1.5 ml; Eppendorf, Hamburg, Ger-many). Subsequent cleaning was done with the RNeasy Mini Kit

(Qiagen), including the optional DNase step. RNA concentrationswere determined with the Nanodrop photometer (NanoDrop, Wil-mington, DE); RNA quality was checked using the Agilent Bioana-lyzer 2100 System (Agilent Technologies, Palo Alto, CA).

cDNA was prepared using the SuperScript double-strand cDNAsynthesis kit (Invitrogen) and a T7-oligo(dT) primer (Affymetrix,Santa Clara, CA) starting with 3 g total RNA as template. In vitrotranscription to generate biotinylated cRNA was done with the ENZOKit (Enzo Life Sciences).

Microarray hybridization. Resulting cRNA was cleaned up withthe RNeasy Mini Kit. RNA fragmentation, hybridization, washing,staining, and scanning were done following the recommendations laiddown in the Affymetrix GeneChip Expression Analysis TechnicalManual. The wash protocol for the Fluidic Station was Midi_Euk2.Samples were hybridized to Affymetrix U133A 2.0 GeneChips. These

arrays represent 18,400 transcripts and variants, including 14,500well-characterized human genes. The arrays were scanned with theGeneChip Scanner 3000; data were recorded with the GCOS 1.1.program (Affymetrix).

Microarray analysis and statistics. GCOS 1.1. was used for back-ground/noise correction and scaling of all signals to a mean signalintensity of 500. Additional scaling, normalization, and filtering stepswere done using GeneSpring 6.1. (Silicon Genetics, Redwood City,CA). For better comparability, median expression values of all 20chips were normalized to 1; expression values of each gene on thesechips were normalized to the median.

Additional data reduction was done using several filter criteria. Allgenes not showing expression values of 50 (raw signal) or higher in atleast five chips were excluded from further analysis. The same was

done with genes that did not show present flags on at least fivechips. These criteria were chosen to allow a change in gene expressionfrom absent to present or from present to absent as a consequence ofthe training procedure. The thus-obtained list of genes was used toidentify changes in gene expression induced by exercise.

Data were grouped according to time point (pre- and postexercise)and extent of physical activity (ET or MT). Each of the four groupscontained data from the same five participants. Groups were compared

using ANOVA with and without correction for multiple testing (usingBenjamini-Hochberg or Bonferroni algorithms). To discover thegenes with the most prominent postexercise changes in gene expres-sion, we used two different strategies. In one strategy, the differencesin gene expression from before and after exercise were determinedwithin each individual; in the other strategy we compared meanexpression levels for each gene calculated from all participants beforeand after exercise. We used a 1.5-fold change as well as a 2-foldchange as the cutoffs. Next the within-individual data gene lists werematched for intersections between these genes (intersection lists).Gene lists emanating from the groupwise comparisons (mean changelists) contained genes whose expression was significantly changed(ANOVA with Benjamini-Hochberg correction; P 0.05).

Hierarchical clustering (using the standard correlation setting) andprincipal component analysis (PCA) were done using GeneSpring 6.1.

PCA is a decomposition technique that produces a set of expressionpatterns known as principal components. Linear combinations of thesepatterns can be assembled to represent the behavior of all of the genesin a given data set. PCA is not a clustering technique, but it is a toolto characterize the most abundant themes or building blocks thatreoccur in many genes of the experiment (GeneSpring technicalmanual). The ANOVA was based on the prefilter list and was carriedout without multiple testing corrections. The list, containing 2,777genes related to exercise, was used for a PCA on conditions [beforetreadmill test (pre-T), MT, ET].

Finally, we matched the lists comprising genes with a 1.5-foldchange and significance levels of P 0.05 (ANOVA with BenjaminiHochberg multiple testing correction) with lists supplied by the geneontology (GO) consortium to sort our genes into functional groups.The gene list belonging to Term GO 0006954 (inflammatory re-

sponse) was imported to GeneSpring because it caught our interestbut was not a part of the standard GO SLIM used in GeneSpring bydefault. All samples were transmitted to the NCBI Gene ExpressionOmnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and stored in thedatabase as GSM82670 to GSM82689. The associated series can befound under the accession number GSE3606.

Quantitative real-time PCR. To validate the relative gene expres-sion data of microarray analysis, we performed quantitative real-timePCR. RNA samples were reversely transcribed using the OmniscriptRT Polymerase kit (Qiagen, Hilden, Germany); 1.2 g RNA and 82.5pmol poly(dT)15-primer (Roche, Mannheim, Germany) were mixed in11.3 l of RNase-free water (Roth, Karlsruhe, Germany) and dena-tured at 72C for 4 min. The RT reaction was carried out in 25 lRNase-free water containing 500 mol/l deoxynucleoside triphos-phates (dNTPs) (Qiagen), 3.3 mol/l poly(dT)15-primer, 0.25 U/l

Omniscript polymerase, 0.5 U/l RNase inhibitor (Roche), the before-mentioned amounts of RNA and primers, 0.01% (wt/vol) BSA(Sigma, Taufkirchen, Germany), and 1 Omniscript buffer at 37Cfor 1 h. Two microliters of cDNA was then used for RT-PCR.

Primers were designed using the Affymetrix U133A 2.0 oligonu-cleotide target sequence and the PRIMER3 program (http://frodo.wi.mit.edu/). They were checked for specificity through blast/insilico PCR using the PUNS (Primer-UniGene Selectivity) program(http://okeylabimac.med.utoronto.ca/cgi-bin/PUNS/). Primer sequencesfor GAPDH (NM_002046) were 5-tcggagtcaacggatttggtcgta-3 and5-atggactgtggtcatgagtccttc-3; the annealing temperature was 66C.MMP-9 (NM_004994) amplification was done using 5-aaagcctatttct-gccaggac-3 and 5-gtggggatttacatggcact-3 as primers. Primers forS100P (NM_005980) amplification were 5-taccaggcttcctgcagagt-3

27EXERCISE AND GENE EXPRESSION

J Appl Physiol VOL 102 JANUARY 2007 www.jap.org


4/12

and 5-ctccagggcatcatttgagt-3; in both reactions an annealing tem-perature of 60C was used. Quantitative RT-PCR was done bycomparison of cycle threshold values for specific genes measured intotal cDNA with those of standards of known copy numbers using theSybrGreen method on a RotorGene 2000 (Corbett Research, Mort-

lake, Australia) cycler. Reaction conditions were 0.5 mol/l for eachprimer, 200 mol/l dNTPs, 1.8 mmol/l MgCl2, 1.5 U Taq polymerase(Platinum Taq; Invitrogen), 0.01% (wt/vol) BSA, 1 reaction buffer(Invitrogen), and 0.16 SybrGreen (Roche) in a final volume of 25l. The PCR program started with a denaturation step at 96C. Fortycycles of denaturation (96C/30 s), annealing (temperature as indi-cated by the specific primer used/30 s), and extension (72C/25 s)were then carried out. Finally, the product melting curves rangingfrom 75C to 99C were measured to exclude the presence ofunspecific by-products.

RESULTS

The mean running time in the treadmill exercise tests was39.0 min (SD 14.8). Heart rate was 182 beats/min (SD 3) at the

end of ET, while after MT it was 142 beats/min (SD 10). Afterboth tests we observed typical exercise intensity-relatedchanges in metabolic parameters and in WBC counts. At theend of the ET, lactate concentrations were significantly en-hanced [3.4 (SD 0.7) vs. 1.1 mmol/l (SD 0.2); P 0.05] whilethey remained unchanged after the MT [1.2 (SD 0.3) vs. 1.3mmol/l (SD 0.2)]. Similarly, the magnitude of the exercise-induced leukocyte change was dependent on exercise intensity,

whereas after both tests an exercise-induced lymphopenia wasobserved (Table 1).

We analyzed whether exercise induces changes in geneexpression profiles (GEPs) of WBCs and if exercise intensity islikewise mirrored. Individual gene chips hybridized with RNAfrom our probands WBCs showed positive expression signals(flagged present) for 11,807 to 13,140 transcripts. A total of

12,911 transcripts passed the prefilter conditions.To get an idea of intraindividual differences in gene expres-

sion, we first compared the gene expression data under restingconditions before the ET test with the data before the MT test.For comparison the mean signal intensities for the pre-ET datawere plotted against those for the pre-MT data (Fig. 1A). Ahigh degree of reproducibility was found for the data from thetwo groups. None of the expressed genes was significantly(ANOVA without correction for multiple tests; data notshown) different between the two data sets. This prompted usto pool all preexercise data for further data analysis. The otherscatterplots in Fig. 1 show a comparison of the pooled preex-ercise gene expression data with data either after the MT (Fig.1B) or after the ET (Fig. 1C). Already this very basic tool ofraw data visualization clearly indicated that regulation ( or2-fold change) was more pronounced after ET than after MT.

To test if exercise tests of different intensities leave geneexpression fingerprints (46), an experiment-specific pattern ofup- or downregulated genes, we performed a hierarchicalcluster analysis. Individual preexercise data and post-ET datawere sorted to different clusters (resting vs. exercise) while thepost-MT data were inconsistently distributed. Two post-MTgene expression profiles, from probands 2 and 3, were groupedtogether with the resting group, while the others were groupedto the ET group (data not shown). Similar results were foundusing PCA. We identified two components, which are visual-ized in Fig. 2, which clearly separate between the before-

exercise GEPs and those obtained after ET. Figure 2 alsoconfirmed the intermediate state of the MT data betweenpreexercise and ET.

To detect genes that might work as indicators of the exerciseworkload, we used a two-stage approach. First we generatedindividual gene lists for all five probands containing all geneschanged more than 1.5-fold within a specific proband. The fivelists were used to create an intersection list (list 1). This list

Table 1. Exercise-induced changes in leukocyte count

Before Test After Test 1 h After Test

WBCMT 5.02 (0.66) 5.36 (1.00) 4.92 (0.68)ET 5.10 (0.92) 7.52* (1.29) 5.88 (1.47)

LymphocytesMT 2.32 (0.38) 2.16 (0.34) 1.58* (0.12)ET 2.32 (0.80) 4.05* (0.93) 1.83 (0.56)

Granulocytes/monocytesMT 2.70 (0.48) 3.20* (0.67) 3.36* (0.71)ET 2.78 (0.41) 3.30 (0.52) 4.18* (2.06)

Values are means (SD). All units of measurement represent numbered cells1000/l. Changes in white blood cell (WBC) counts as well as in lympho-cyte and granulocyte/monocyte fractions after moderate (MT) and exhaustive(ET) treadmill tests. *P 0.05.

Fig. 1. Scatterplots visualizing mean raw gene expression data. A: the variation between 2 measurements of the same 5 individuals [before moderate treadmilltest (pre-MT) vs. before exhaustive treadmill test (pre-ET)]. B: comparison of the mean gene expression levels of all before-exercise data (pre-T) (n 2 5)with those after moderate exercise (post-MT; n 5). C: pre-T vs. after ET (post-ET); blue lines indicate a 2-fold change.

28 EXERCISE AND GENE EXPRESSION



5/12

contained 40 genes that were upregulated and 10 genes that

were downregulated. A second gene list (list 2) was createdcontaining genes whose mean expression levels were signifi-cantly changed (ANOVA with Benjamini-Hochberg correc-tion; P 0.05) and showed an increase or decrease in expres-sion levels of at least 1.5-fold after ET. A total of 450 and 150genes were differentially up- and downregulated, respectively(see Supplementary Tables 1 and 2, contained in the onlineversion of this article). The intersection of gene lists 1 and 2resulted in a final list of potential exercise marker genes, whichcontained 39 upregulated and 7 downregulated transcripts(summarized in Tables 2 and 3, respectively).

On the U133A 2.0 chip many genes are represented byseveral sets of oligonucleotides. Therefore, for all genes that

were part of the final exercise marker gene list, we analyzed the

expression change for all transcripts present on the chip (Ta-bles 2 and 3). For the vast majority of the analyzed genes, weobtained concordant results. When we used the same approachwith a 2.0-fold change as a filter instead of a 1.5-fold change,only one transcript [for heat shock protein 105 kDa (HSPH1)]passed the filter.

The exercise marker gene list contained genes that areinvolved in several different physiological processes. The mostprominent changes were found for stress proteins such asHSPH1 and heat shock 70-kDa protein 1A (HSPA1A). More-over, genes related to inflammatory mediators, transcriptionalregulators, membrane transporters/channels, and molecules in-volved in the breakdown of the extracellular matrix were

Fig. 2. A: visualization of the behavior of allsamples in the context of exercise via hier-archical clustering using standard correlationbased on ANOVA without multiple testingcorrections (2,777 genes) The first clusterincludes all preexercise data together withthe data ofparticipants 2 and 3 after the MT,while all other postexercise data were sortedinto the second cluster. Color codes: red,high expression; yellow, medium expres-sion; green, low expression. B: diagramshows the 2 main principal components cal-culated from 2,777 genes (ANOVA, P 0.05). E, pre-T; , MT; , ET.




6/12

Table 2. Exercise marker genes (upregulated)

AnnotationNumber Description Abbreviation Fold Change

208744_x_at heat shock 105 kDa HSPH1 3.26206976_s_at heat shock 105 kDa HSPH1 2.49

203936_s_at matrix metalloproteinase 9 MMP9 3.07200800_s_at heat shock 70 kDa protein 1A HSPA1A 2.85200799_at heat shock 70 kDa protein 1A HSPA1A 1.86211538_s_at heat shock 70 kDa protein 1A HSPA1A 1.19218865_at hypothetical protein FLJ22390 FLJ22390 2.83

217203_at 2.65216782_at Homo sapiens cDNA: FLJ23026 fis, clone LNG01738, mRNA sequence 2.39218739_at CGI-58 protein CGI-58 2.39

213805_at CGI-58 protein CGI-58 1.87213935_at CGI-58 protein CGI-58 1.32211372_s_at interleukin 1 receptor, type II IL1R2 2.23

205403_at interleukin 1 receptor, type II IL1R2 2.11202497_x_at solute carrier family 2 (facilitated glucose transporter), member 3 SLC2A3 2.21216236_s_at solute carrier family 2 (facilitated glucose transporter), member 3 SLC2A3 2.18202498_s_at solute carrier family 2 (facilitated glucose transporter), member 3 SLC2A3 2.03202499_s_at solute carrier family 2 (facilitated glucose transporter), member 3 SLC2A3 1.93203435_s_at membrane metallo-endopeptidase MME 2.19

203434_s_at membrane metallo-endopeptidase MME 2.32217104_at Homo sapiens mRNA full length insert cDNA clone EUROIMAGE 327506, mRNA sequence 2.16210119_at potassium inwardly-rectifying channel, subfamily J, member 15 KCNJ15 2.16

211806_s_at potassium inwardly-rectifying channel, subfamily J, member 15 KCNJ15 2.05218880_at FOS-like antigen 2 FOSL2 2.12205409_at FOS-like antigen 2 FOSL2 1.90220436_at cell recognition molecule CASPR3 CASPR3 2.04

205645_at RALBP1 associated Eps domain containing 2 REPS2 2.04200881_s_at DnaJ (Hsp40) homolog, subfamily A, member 1 DNAJA1 2.01200880_at DnaJ (Hsp40) homolog, subfamily A, member 1 DNAJA1 1.53218881_s_at hypothetical protein FLJ23306 FLJ23306 2.00202241_at phosphoprotein regulated by mitogenic pathways C8FW 1.94205227_at interleukin 1 receptor accessory protein IL1RAP 1.93

210233_at interleukin 1 receptor accessory protein IL1RAP 1.27204713_s_at coagulation factor V F5 1.92204714_s_at coagulation factor V F5 1.62

209850_s_at CDC42 effector protein CDC42EP2 1.88214014_at CDC42 effector protein CDC42EP2 2.63205016_at transforming growth factor, alpha TGFA 1.87

205015_s_at transforming growth factor, alpha TGFA 2.28211258_s_at transforming growth factor, alpha TGFA 1.53213072_at hypothetical protein BC004544 LOC157542 1.8658780_s_at hypothetical protein FLJ10357 FLJ10357 1.86

220326_s_at hypothetical protein FLJ10358 FLJ10357 1.84201926_s_at decay accelerating factor for complement (CD55, Cromer blood group system) DAF 1.86201925_s_at decay accelerating factor for complement (CD55, Cromer blood group system) DAF 1.64209930_s_at nuclear factor (erythroid-derived 2), 45kDa NFE2 1.85209802_at tumor suppressing subtransferable candidate 3 TSSC3 1.83209803_s_at tumor suppressing subtransferable candidate 3 TSSC3 1.11200939_s_at arginine-glutamic acid dipeptide (RE) repeats RERE 1.83221643_s_at arginine-glutamic acid dipeptide (RE) repeats RERE 1.48200940_s_at arginine-glutamic acid dipeptide (RE) repeats RERE 1.38200938_s_at arginine-glutamic acid dipeptide (RE) repeats RERE 0.72

220404_at PRO0611 [Homo sapiens], mRNA sequence 1.79210484_s_at hypothetical protein MGC31957 MGC31957 1.79

210483_at hypothetical protein MGC31958 MGC31957 1.64209460_at NPD009 protein NPD009 1.72209459_s_at NPD009 protein NPD009 1.42210706_s_at ring finger protein 24 RNF24 1.71

204669_s_at ring finger protein 24 RNF24 1.71215760_s_at KIAA0963 protein KIAA0963 1.68204166_at KIAA0963 protein KIAA0963 1.54205857_at solute carrier family 18 (vesicular monoamine), member 2 SLC18A2 1.61

Genes found to be significantly (ANOVA, Benjamini-Hochberg multiple testing correction) upregulated more than 1.5-fold after exhaustive treadmill exerciseaccording to the annotations present on U133A 2.0 gene chip. These genes are printed in boldface type. Additional oligonucleotides of these genes that are alsopresent on the gene chip are printed in regular type.




7/12

among the potential marker genes. To validate the microarrayexperiments, the expression level changes of three genes,HSPA1A, S100P (part of the mean gene expression list), andMMP-9, were also determined by quantitative real-time PCR(Fig. 3). Data were normalized to the housekeeping proteinGAPDH. The PCR results confirmed an upregulation of theseparticular genes by exercise (Fig. 3).

The upregulated genes (list 2) were highly significantlyassociated with the gene ontology lists stress response (14genes/18 transcripts), heat-shock response (7 genes/11 tran-scripts), and inflammatory response (14 genes/16 transcripts)(Table 4). Exemplarily we imported the intersectional listbetween list 2 and inflammatory response, which containedthe largest number of genes, into PathwayAssist 3.0. Using allavailable information and considering one regulatory elementbetween the genes the program calculated direct interactions

for 8 of 14 genes, which have been shown to exist at the proteinlevel (Fig. 4).When we compared the expression level changes of the

marker genes for exhaustive exercise with those after MT, wefound that after MT the expression level change was muchmore moderate for both up- and downregulated genes (Fig. 5).This effect of exercise intensity on the expression level wasalso confirmed by quantitative PCR for MMP9 and HSPA1A(Fig. 3).

In their recently published study, Connolly and coworkers(4) identified 78 genes that were significantly upregulated by afactor of 1.3 or larger following exercise and a 1-h rest.Interestingly, 75 of these genes (96%) were found to beupregulated in our study, too, of which 43 (56%) showed a

significant (ANOVA; P 0.05) expression increase by afactor of 1.3 or more. None of the three genes from the list ofConnolly et al. (4) whose expression was not increased in ourstudy showed a significant decrease. Moreover, 53 genes werereported to be significantly downregulated in the study ofConnolly et al. Of these, 50 (94%) were also found downregu-lated in our study. Ten of them (19%) reached significance(factor 1.3; ANOVA P 0.05). Only one gene that wasfound to be downregulated by Connolly and coworkers (4),IL-18 receptor accessory protein (IL18RAP), showed an op-posite behavior in our study, where it was significantly in-creased (Fig. 6; for details, see Supplementary Tables 3 and 4,contained in the online version of this article).

DISCUSSION

The present study aimed at the detection in WBCs of specific

patterns of gene expression in response to different exerciseloads. To apply different exercise loads on the subjects twoexercise protocols of different exercise intensities were used.Treadmill running at velocities corresponding to either 80% or60% VO2 max has been used by us recently (25). These proto-cols have been shown to be effective in achieving differentmetabolic and inflammatory responses (25). Likewise, in thepresent study, significantly different values for leukocytes andlactate were obtained.

Differential effects of exercise on gene expression. Basalgene expression data before the MT and the ET tests were quitesimilar, which supports the specificity and the reproducibilityof our method. Performing exercise results in an enhancedvariation of gene expression. Visual inspection of the scatter-grams in Fig. 1 shows that the extent of gene expressionchanges is related to the intensity of exercise. The stress effectsof acute exercise depend on enormous loads of free radicals,electrolyte, and acid-base imbalances (23, 24). Such exercise-induced deviation from homeostatic equilibrium is related toexercise intensity, too. The bodys acute response to exercise ischaracterized by systemic efforts to compensate for thesechallenges and to gain homeostatic recovery (22). This isachieved on different regulatory levels. On the transcriptionallevel it results in the activation of predominantly acute-phaseand stress-related genes. Indeed, in our study acute exerciseaddressed gene ontology lists related to inflammation, cellstress, and apoptosis.

The pathway model for upregulated genes involved in in-flammatory response also confirms an activation of complexinflammatory processes through exhaustive exercise. Eight of14 genes have been found to directly interact at the proteinlevel. The changes we observed are supposed to be more thanjust coincidence and therefore suggest the operation of theseregulatory networks in the WBCs response to exercise.Thereby the global gene response showed remarkable similar-ities within the individuals, which suggests that exercise is ableto leave characteristic fingerprint-like gene expression profilesin leukocytes. This is suggested by both the hierarchical clusteranalysis and the PCA. Both analytical procedures sorted pre-exercise and post-ET data into clearly different groups while

Table 3. Exercise marker genes (downregulated)

AnnotationNumber Description Abbreviation Fold Change

215894_at prostaglandin D2 receptor (DP) PTGDR 1.85215937_at prostaglandin D2 receptor (DP) PTGDR 1.38213995_at mitochondrial ATP synthase regulatory component factor B ATP5S 2.01

206992_s_at mitochondrial ATP synthase regulatory component factor B ATP5S 1.34

206993_at mitochondrial ATP synthase regulatory component factor B ATP5S 0.98212981_s_at PRO2751 ( Homo sapiens), mRNA sequence 2.13

65585_at Unnamed protein product ( Homo sapiens), mRNA sequence 2.14207840_at natural killer cell receptor, immunoglobulin superfamily member BY55 2.17216050_at Homo sapiens cDNA: FLJ20931 fis, clone ADSE01282, mRNA

sequence2.32

202932_at v-yes-1 Yamaguchi sarcoma viral oncogene homolog 1 YES1 2.48202933_s_at v-yes-1 Yamaguchi sarcoma viral oncogene homolog 1 YES1 1.90

Genes found to be significantly (ANOVA, Benjamini-Hochberg multiple testing correction) downregulated more than 1.5-fold after exhaustive treadmillexercise according to the annotations present on U133A 2.0 gene chip. These genes are printed in boldface type. Additional oligonucleotides of these genes thatare also present on the gene chip are printed in regular type.




8/12

the post-MT data were scattered across both groups. Thissuggests that exercise is a major inductor of specific geneexpression patterns. One reason for the scattered distribution ofthe MT samples may be differences in the individual trainingstatus of the participants. This goes along with changes in therelative workload (26). Obviously, for two subjects the load ofmoderate exercise was therefore not effective to induce anexercise gene pattern.

Potential marker genes for acute exercise. Acute exercisehas an effect on a large number of genes in WBCs. When we

used the most rigid conditions for the composition of theexercise marker gene list (the effect had to be present in everyindividual tested and, in a groupwise comparison, these differ-ences had to be statistically significant after correction formultiple testing), our analysis resulted in a final list of 35upregulated genes. Four of them had two significant tran-scripts. The genes from this restricted list could be related to

the following functional classes: stress proteins/stress signal-ing, extracellular matrix, electrolyte and substrate transport,cytokines, and transcription factors.

Among the genes with the highest significant fold changeswere HSPH1 and HSPA1A, which code for stress proteinsbelonging to the HSP105/110 and HSP70 family, respectively(13, 41). These proteins play important roles as molecularchaperones, thereby preventing the stress-induced irreversibleaggregation of denatured proteins; they assist in folding, as-sembly, and translocation of cellular proteins across the mem-brane (10, 13, 40, 44). In leukocytes exercise-induced regula-tion of HSPs on both the transcriptional and the translationallevel has been found (6). Interestingly, the cytoplasmic expres-sion of HSP27 and HSP70 was found to be weaker in trainedsubjects than in untrained, which is in good agreement with ourdata on mRNA levels (7).

A novel finding in circulating WBCs was the exercise-induced upregulation of MMP-9, which breaks down nativetype IV collagen as well as other extracellular matrix mole-cules. There are reports about an increase of intramuscularMMP-9 synthesis and its collagen-degrading activity, as wellas about a release of MMP-9 into muscle interstitial fluid andserum, especially after muscle-damaging eccentric exercise(1517). Therefore, MMP-9 is supposed to be of importancefor tissue remodeling after exercise-induced muscle damage. Incontrast, the role of MMP-9 in WBCs during exercise is lessclearly defined. An attractive hypothesis about the role of

MMP-9 for the hematopoetic system during exercise involvesthe mobilization of hematopoetic progenitor cells (38). Re-cently it could be shown that exercise is an effective stimulusfor the mobilization of hematopoetic progenitor cells (27).Moreover, studies in rhesus monkeys suggest an involvementof MMP-9 in IL-8-induced mobilization of hematopoieticprogenitor cells via cleavage of those matrix molecules thestem cells adhere to (19). In agreement with these findings, wefound the IL-8 receptor- (IL8RA) to be upregulated signifi-cantly (ANOVA; Benjamini-Hochberg). However, it remainsto be shown whether these mechanisms are operative duringexercise, too.

Next, we found genes upregulated that are involved inenergy metabolism and potassium homeostasis, such as

CGI-58 protein (CGI-58), solute carrier family 2 (facilitated

Table 4. Association of exercise-inducedgenes with gene ontology lists

GO SLIMS Term Ove rlap P Value Sample Genes

Response to stress 1.43 1011 AHSA1, STIP1, MAPK14Heat shock response 1.42 109 HSPA1A, HSPA1B, HSPCAInflammatory response 1.70 106 ALOX5, IL1A, IL8RAApoptosis-related function 1.42 106 CFLAR, APAF1, BCL2A1

GO SLIMS terms highly significantly overlaped with genes found upregu-lated more than 1.5-fold (ANOVA, Benjamini Hochberg, P 0.05). See textfor gene descriptions.

Fig. 3. Comparison of microarray data with quantitative real-time PCR (RT-PCR) data (all samples normalized to GAPDH). A: S100P. B: matrix metal-loproteinase-9 (MMP). C: heat shock 70-kDa protein 1A (HSPA1A). ,RT-PCR; F, array data.




9/12

glucose transporter) member 3 (SLC2A3), and the potassiuminwardly-rectifying channel, subfamily J, member 15(KCNJ15) (20, 42). The upregulation of these genes can beregarded as a response to changes in energy requirements andalterations in electrolyte equilibrium (39). Depending on thetype and intensity of exercise, marked increases in plasma

potassium concentration occur. Therefore, exercise challengescellular potassium homeostasis, especially in muscle cellsalthough also in other cell types, which results in well-knownalterations of potassium-regulating membrane proteins (24). Ata first glance it seems surprising that genes are upregulated inlymphocytes that are not doing the actual work. However, arecent study by Zeibig et al. (43) confirmed that the training-induced upregulation of muscle genes involved in oxidativeenergy metabolism is mirrored in lymphocytes, too.

Furthermore, some genes were found upregulated that all arepart of the anti-inflammatory response associated with exercise.This included the release of receptor molecules for inflammatorycytokines such as IL-1 receptor (found significantly altered),which has been shown at the protein level in a number of studies

(4, 37). This included also the membrane metalloendopeptidase(MME), which codes for a glycosylated zinc-dependent mem-brane-bound metalloproteinase (neutral endopeptidase; NEP)known to control the bioavailability of inflammatory neuropep-tides such as substance P, which can be released from sensorynerves, skin, and immune cells (11, 28, 35).

The list of genes that were downregulated by exercisecontained only seven transcripts. Two of these, the v-yes-1Yamaguchi sarcoma viral oncogene homolog 1 (YES) and thenatural killer cell receptor BY55/CD160, should be discussedin more detail. YES is a member of the Src family of tyrosineprotein kinases that function as regulators of the actin-cytoskel-etal structure by acting on the anchoring structures, namely

focal adhesions and focal contacts (9, 21). A recent study byHansen et al. pointed to an exciting relation of YES to heatshock proteins and MMP-9 (12). The overexpression of HSP27induced an upregulation of MMP-9 expression and activity andat the same time downregulated YES expression. Such aconnection seems to exist after exercise stress, too, as sug-

gested by our microarray data.Finally, we found a downregulating effect of acute exerciseon the natural killer cell receptor (BY55; CD160) (1). Nikolovaet al. (30) demonstrated that CD160 behaves as an importantcostimulant of the T-cell receptor, especially in CD28-negativecells with consequences on T-cell expansion and cytotoxicity.In contrast, a decrease in CD160 expression should result in animpaired functional response of at least some lymphocytesubpopulations, which might account partially for the well-known immunosuppressive effects of exhaustive exercise (29).

Gene expression-based training control. Our results showthat the acute exercise response measured in WBC addressesonly a rather limited number of genes significantly and that thegene response is related to exercise intensity. This opens the

perspective for the future to design an exercise chip that mayserve as an innovative tool for monitoring and controllingtraining processes. Knowing the individual response to exer-cise could improve our ability to optimize exercise dosage andrecovery time. One cornerstone down this alley is the identi-fication of suitable genes that are differentially affected byexercise and that show good reproducibility in other microar-ray studies. Connolly et al. (4) used Affymetrix arrays toexamine the impacts of a very similar exercise protocol ongene expression, the same that were used in this study. Thisallows a sufficient comparison of both data sets, which resultsin many similarities. However, our results are more than just aconfirmation of previous data sets because both studies showed

Fig. 4. Pathway model of genes that were found signif-

icantly (ANOVA Benjamini-Hochberg) upregulatedmore than 1.5-fold by ET and that were included in thegene ontology (GO) term inflammatory response(provided by GO consortium); genes with gray shadingare part of the mentioned lists; genes without grayshading are not differentially regulated through exercise[transforming growth factor-1 (TGF- 1)] or flaggedabsent (IL-13). Pathways were calculated using Strat-agene PathwayAssist 3.0. IL8RA, IL-8 receptor-;IL1R1, IL-1 receptor type I; FOS, v-fos FBJ murineosteosarcoma viral oncogene homolog; IL1RN, IL-1receptor antagonist; FPRL1, formyl peptide receptor-like 1; ALOX5, arachidonate 5-lipoxygenase; CXCL1,chemokine (C-X-C motif) ligand 1 (melanoma growthstimulating activity-); IL1A, IL-1.




10/12

substantial methodological differences in sample handling andRNA preparation. This suggests the applicability of this ap-proach for exercise-related purposes as long as the sameanalytical platform is used. Moreover, it is a strong argumentfor the existence of specific exercise-induced gene expressionpatterns and therefore a prerequisite for the design of exercisechips. Using similar cutoffs and statistical approaches (1.3-foldexpression changes; ANOVA with Benjamini-Hochberg cor-rection), both studies revealed 42 genes to be significantlyupregulated and 5 genes downregulated. This common genelist contained some of the genes that were already discussed

above such as HSPA1A, HSPH1, IL1R2, FOSL2, NFE2,PTGDR, and BY55. Additional genes (expression data insupplementary list) were related to environmental stress likehypoxia-inducible factor-1 subunit (HIF1A), heat shock pro-tein-1 90 kDa (SPCA), stress-induced-phosphoprotein 1(STIP1), and vanin1/2 (VNN1/2), to substrate transport-likesolute carrier family 16 (monocarboxylic acid transporters)member 3 (SLC16A3), to apoptosis regulation such as BCL2-related protein A1 (BCL2A1), and to surface receptors such aschemokine (C-C motif) receptor 2 (CCR2), chemokine (C-X3-C motif) receptor 1 (CX3CR1), and CD14 antigen (CD14).

Two other studies investigated the effects of exercise onleukocyte gene expression using custom-made gene chips.

Hilberg et al. (14) used chips with transcripts of 5,000 stressand inflammation relevant genes, while Zieker et al. (45) usedspotted cDNA microarrays containing 277 different genesfocused on inflammation. Since both groups utilized differentmethods in sample processing and used different oligonucleo-tide probes for their research, a comparison of their results withour data is difficult. In addition, Hilberg et al. (14) used a

different time scale, with sampling points 2 and 6 h postexer-cise. A comparison of up- and downregulated genes at either ofthese points with our data revealed a common list of 9 and 2genes, respectively. The list contained genes coding for ara-chidonate 5-lipoxygenase (ALOX5) and ALOX5-activatingprotein, which are related to arachidonic acid metabolism; andfor Casp8 and FADD-like apoptosis regulator (CFLAR), whichlike the above-mentioned BCL2A plays an antiapoptotic role.Other genes were the soluble IL-6 receptor, phosphoinositide3-kinase (PIK3CG), CD16b (FCGR3B), CD14, colony-stimu-lating factor 3 receptor (CSF3R), and leukocyte immunoglob-

Fig. 6. Between-laboratory comparison of exercise-induced gene expressionchanges. Similar test conditions and identical microarrays but different pre-analytical procedures were used in the 2 studies. The graph shows how genesthat have been reported by Connolly et al. to be significantly different in pre-and postexercise samples 1 h past exhaustive exercise (most of them showedan at least 1.3-fold gene expression change) behave in the study presented here.Black boxes indicate genes that were found significantly (ANOVA Ben-jamini-Hochberg) regulated in our data set. A: upregulated genes; B: down-regulated genes. Dashed lines indicate a 1.3-fold change.

Fig. 5. Normalized expression signals of marker genes before exercise (Pre),after moderate exercise (MT), and after excessive exercise (ET). A: the genesupregulated after exercise (39 transcripts representing 35 genes). B: the

downregulated genes (7 genes).




11/12

ulin (Ig)-like receptor subfamily A member 2 (LILRA2). In thestudy by Zieker et al. (45), a group of highly trained enduranceathletes was investigated, who often show a diminished re-sponse of inflammatory parameters to exercise. Moreover,exercise duration (104 min) and sampling points (before andafter exercise) were different compared with our protocol.These might be just a few reasons why we could find only a

limited number of genes demonstrating a similar regulatorypattern as in our study, e.g., CD1c, CD81, CD244, selectin L,intercellular adhesion molecule 2 (ICAM2), mitogen-activatedprotein kinase activating protein kinase 2 (Mapkap K2), andIL-1 receptor antagonist. Another likely reason is the differentmicroarray technology used.

The question remains whether leukocyte gene expressionpatterns can be used as surrogate markers that mirror exerciseand training-induced responses of other tissues such as muscletissue. An example for such a relation is the exercise-inducedmRNA expression of heat shock proteins in muscle, which hasbeen shown to occur depending on type and intensity ofexercise (5, 33). Mahoney et al. (22) found that exercise

differentially affected the expression of genes involved inmetabolism and mitochondrial biogenesis, cell growth, andtranscriptional activation, as well as apoptosis, cell stress, andproteolysis (22). For a number of genes among the three lattergroups, we found identical regulatory patterns in leukocytesafter exercise, e.g., genes involved in cell stress management[e.g., DnaJ (Hsp40)], proteolysis (e.g., MMPs), and apoptosis(e.g., BCL2-related anti-apoptotic proteins). Another exampleis the study by Zeibig et al. (43), which described the nearlyidentical regulation of genes related to oxidative energy me-tabolism in both working muscle cells and nonworking lym-phocytes following a 6-mo training period of cross-countryskiers.

Methodological limitations. The present study presented a

number of results that fit very well with our current under-standing of the organisms responses and adaptations towardsacute exercise. Nonetheless some limitations of our approachshould be addressed. First, the present study suffers from thereduced number of participants as is the case in many microar-ray studies. The high costs of this technology often prevent theinclusion of larger cohorts. When working with small cohortsthe danger of detecting false positives is high. We tried toreduce the number of false positives by using multiple testingcorrections, which, however, increase the number of falsenegatives. False-positive results in microarray analyses mostlystem from the high technical variance of the methodology.When such data are entered into a group classification proce-

dure like the GO class assignment, false positives are likely tobe distributed among the classes at random. The critical num-ber of class assignments necessary for a statistically significantresult is usually only reached by genes that are involved in thesame biological process and not by genes whose expression ismodified by technical insufficiencies.

The fact that most genes that changed expression in ourstudy did this after moderate as well as after exhaustiveexercise speaks against a chance finding. The possibility thatthe expression changes do not stem from exercise but ratherrepresent changes induced during the workup of the leukocytesis also not very likely because this would equally affectsamples from before and after exercise and is thus not likely to

reach significance. In addition, we have observed a stepwisechange in gene expression related to the intensity of exercise.

Second, it might be possible that many of the observedexpression changes relate to changes in the composition of theWBCs. However, most of the genes in our final positive list(after correction for multiple testing) show expression changesthat are bigger than the changes in leukocyte subpopulations.

While there is certainly an important effect of cell populationshifts, it is likely that gene expression is also affected. This isfurther supported by the fact that we and Connolly et al. (4)found similar gene expression patterns despite working ondifferent cell populations. Connolly et al. (4) worked on pe-ripheral blood mononuclear cells isolated by a Ficoll gradientwhile we worked on the total leukocyte fraction. We chose thismethod to guarantee short intervals from blood sampling untilRNA stabilization. Indeed the harmony of the results suggestsa stable reproducibility and applicability of the microarraytechnology for investigating exercise-induced gene expressionchanges. To avoid difficulties in data interpretation and toelucidate the contribution of each subpopulation, further stud-ies are required to utilize subpopulations isolated by cellsorting or by immunoprecipitation techniques.

In summary, our results indicate that exercise is able toaffect leukocyte gene expression in a dose-dependent fashionand to induce specific gene expression profiles that can bedetected successfully by microarray analysis. Moreover, wefound evidence that adaptational training responses known,e.g., from muscle tissues are reflected partially in leukocytes.Further studies are required to verify these findings and toinvestigate these relationships with respect to different types ofexercise, training regimens, etc. Finally, a number of acuteexercise marker genes could be either newly established orconfirmed. They represent potential tools for monitoring exer-cise and training processes in the future.

GRANTS

This work was supported by Bundesinstitut fur Sportwissenschaft BISPVF0407/01/5/2003.

REFERENCES

1. Anumanthan A, Bensussan A, Boumsell L, Christ AD, Blumberg RS,Voss SD, Patel AT, Robertson MJ, Nadler LM, Freeman GJ. Cloningof BY55, a novel Ig superfamily member expressed on NK cells, CTL, andintestinal intraepithelial lymphocytes. J Immunol 161: 27802790, 1998.

2. Barengo NC, Hu G, Lakka TA, Pekkarinen H, Nissinen A, Tuom-ilehto J. Low physical activity as a predictor for total and cardiovasculardisease mortality in middle-aged men and women in Finland. Eur Heart J25: 22042211, 2004.

3. Bassuk SS, Manson JE. Epidemiological evidence for the role of physicalactivity in reducing risk of type 2 diabetes and cardiovascular disease. J Appl Physiol 99: 11931204, 2005.

4. Connolly PH, Caiozzo VJ, Zaldivar F, Nemet D, Larson J, Hung SP,Heck JD, Hatfield GW, Cooper DM. Effects of exercise on geneexpression in human peripheral blood mononuclear cells. J Appl Physiol97: 14611469, 2004.

5. Febbraio MA, Koukoulas I. HSP72 gene expression progressively in-creases in human skeletal muscle during prolonged, exhaustive exercise.

J Appl Physiol 89: 10551060, 2000.6. Fehrenbach E, Niess AM, Schlotz E, Passek F, Dickhuth HH, Northoff

H. Transcriptional and translational regulation of heat shock proteins inleukocytes of endurance runners. J Appl Physiol 89: 704710, 2000.

7. Fehrenbach E, Passek F, Niess AM, Pohla H, Weinstock C, DickhuthHH, Northoff H. HSP expression in human leukocytes is modulated byendurance exercise. Med Sci Sports Exerc 32: 592600, 2000.




12/12

8. Fehrenbach E, Zieker D, Niess AM, Moeller E, Russwurm S, NorthoffH. Microarray technologythe future analyses tool in exercise physiol-ogy? Exerc Immunol Rev 9: 5869, 2003.

9. Gaudreault E, Thompson C, Stankova J, Rola-Pleszczynski M. In-volvement of BLT1 endocytosis and Yes kinase activation in leukotrieneB4-induced neutrophil degranulation. J Immunol 174: 36173625, 2005.

10. Giffard RG, Xu L, Zhao H, Carrico W, Ouyang Y, Qiao Y, SapolskyR, Steinberg G, Hu B, Yenari MA. Chaperones, protein aggregation, andbrain protection from hypoxic/ischemic injury. J Exp Biol 207: 32133220, 2004.

11. de Gouw HW, Diamant Z, Kuijpers EA, Sont JK, Sterk PJ. Role ofneutral endopeptidase in exercise-induced bronchoconstriction in asth-matic subjects. J Appl Physiol 81: 673678, 1996.

12. Hansen RK, Parra I, Hilsenbeck SG, Himelstein B, Fuqua SA. Hsp27-induced MMP-9 expression is influenced by the Src tyrosine proteinkinase yes. Biochem Biophys Res Commun 282: 186193, 2001.

13. Hendrick JP, Hartl FU. Molecular chaperone functions of heat-shockproteins. Annu Rev Biochem 62: 349384, 1993.

14. Hilberg T, Deigner HP, Moller E, Claus RA, Ruryk A, Glaser D,Landre J, Brunkhorst FM, Reinhart K, Gabriel HH, Russwurm S.

Transcription in response to physical stressclues to the molecularmechanisms of exercise-induced asthma. FASEB J 19: 14921499, 2005.

15. Koskinen SO, Heinemeier KM, Olesen JL, Langberg H, Kjaer M.Physical exercise can influence local levels of matrix metalloproteinasesand their inhibitors in tendon-related connective tissue. J Appl Physiol 96:

861864, 2004.16. Koskinen SO, Hoyhtya M, Turpeenniemi-Hujanen T, Martikkala V,Makinen TT, Oksa J, Rintamaki H, Lofberg M, Somer H, Takala TE.Serum concentrations of collagen degrading enzymes and their inhibitorsafter downhill running. Scand J Med Sci Sports 11: 915, 2001.

17. Koskinen SO, Wang W, Ahtikoski AM, Kjaer M, Han XY, Komu-lainen J, Kovanen V, Takala TE. Acute exercise induced changes in ratskeletal muscle mRNAs and proteins regulating type IV collagen content.

Am J Physiol Regul Integr Comp Physiol 280: R1292R1300, 2001.18. Lakier Smith L. Overtraining, excessive exercise, and altered immunity:

is this a T helper-1 versus T helper-2 lymphocyte response? Sports Med33: 347364, 2003.

19. Laterveer L, Lindley IJ, Heemskerk DP, Camps JA, Pauwels EK,Willemze R, Fibbe WE. Rapid mobilization of hematopoietic progenitorcells in Rhesus monkeys by a single intravenous injection of interleukin-8.

Blood 87: 781788, 1996.20. Lefevre C, Jobard F, Caux F, Bouadjar B, Karaduman A, Heilig R,

Lakhdar H, Wollenberg A, Verret JL, Weissenbach J, Ozguc M,Lathrop M, Prudhomme JF, Fischer J. Mutations in CGI-58, the geneencoding a new protein of the esterase/lipase/thioesterase subfamily, inChanarin-Dorfman syndrome. Am J Hum Genet 69: 10021012, 2001.

21. Lynch G, Kohler S, Leser J, Beil M, Garcia-Marin LJ, Lutz MP. Thetyrosine kinase Yes regulates actin structure and secretion during pancre-atic acinar cell damage in rats. Pflugers Arch 447: 445450, 2004.

22. Mahoney DJ, Parise G, Melov S, Safdar A, Tarnopolsky MA. Analysisof global mRNA expression in human skeletal muscle during recoveryfrom endurance exercise. FASEB J 19: 14981500, 2005.

23. McArdle A, Pattwell D, Vasilaki A, Griffiths RD, Jackson MJ. Con-tractile activity-induced oxidative stress: cellular origin and adaptiveresponses. Am J Physiol Cell Physiol 280: C621C627, 2001.

24. Mooren FC. Regulation of intracellular ion composition, and pH. In: Molecular and Cellular Exercise Physiology, edited by Mooren F andVolker K. Champaign, IL: Human Kinetics, 2005.

25. Mooren FC, Bloming D, Lechtermann A, Lerch MM, Volker K.

Lymphocyte apoptosis after exhaustive and moderate exercise. J ApplPhysiol 93: 147155, 2002.

26. Mooren FC, Lechtermann A, Volker K. Exercise-induced apoptosis oflymphocytes depends on training status. Med Sci Sports Exerc 36: 14761483, 2004.

27. Morici G, Zangla D, Santoro A, Pelosi E, Petrucci E, Gioia M,Bonanno A, Profita M, Bellia V, Testa U, Bonsignore MR. Supramaxi-

mal exercise mobilizes hematopoietic progenitors and reticulocytes inathletes. Am J Physiol Regul Integr Comp Physiol 289: R1496R1503,2005.

28. Nadel JA. Peptidase modulation of neurogenic inflammation. In: Neuro-genic Inflammation, edited by Geppetti P and Holzer P. Boca Raton, FL:CRC, 1996.

29. Nieman DC. Exercise immunology: practical applications. Int J SportsMed 8, Suppl 1: S91S100, 1997.

30. Nikolova M, Marie-Cardine A, Boumsell L, Bensussan A. BY55/CD160 acts as a co-receptor in TCR signal transduction of a humancirculating cytotoxic effector T lymphocyte subset lacking CD28 expres-sion. Int Immunol 14: 445455, 2002.

31. Ostrowski K, Rohde T, Asp S, Schjerling P, Pedersen BK. Pro- andanti-inflammatory cytokine balance in strenuous exercise in humans.

J Physiol 515: 287291, 1999.32. Pedersen BK, Hoffman-Goetz L. Exercise and the immune system:

regulation, integration, and adaptation. Physiol Rev 80: 10551081, 2000.33. Puntschart A, Vogt M, Widmer HR, Hoppeler H, Billeter R. Hsp70

expression in human skeletal muscle after exercise. Acta Physiol Scand157: 411417, 1996.

34. Roberts CK, Barnard RJ. Effects of exercise and diet on chronic disease. J Appl Physiol 98: 330, 2005.

35. Scholzen TE, Steinhoff M, Bonaccorsi P, Klein R, Amadesi S, Gep-petti P, Lu B, Gerard NP, Olerud JE, Luger TA, Bunnett NW, Grady

EF, Armstrong CA, Ansel JC. Neutral endopeptidase terminates sub-

stance P-induced inflammation in allergic contact dermatitis. J Immunol166: 12851291, 2001.

36. Stewart KJ. Exercise training and the cardiovascular consequences oftype 2 diabetes and hypertension: plausible mechanisms for improvingcardiovascular health. JAMA 288: 16221633, 2002.

37. Suzuki K, Nakaji S, Yamada M, Totsuka M, Sato K, Sugawara K.Systemic inflammatory response to exhaustive exercise. Cytokine kinetics.

Exerc Immunol Rev 8: 648, 2002.38. Velders GA, Fibbe WE. Involvement of proteases in cytokine-induced

hematopoietic stem cell mobilization. Ann NY Acad Sci 1044: 6069,2005.

39. Wu JV, Krouse ME, Rustagi A, Joo NS, Wine JJ. An inwardlyrectifying potassium channel in apical membrane of Calu-3 cells. J BiolChem 279: 4655846565, 2004.

40. Yamagishi N, Ishihara K, Saito Y, Hatayama T. Hsp105 but not Hsp70family proteins suppress the aggregation of heat-denatured protein in the

presence of ADP. FEBS Lett 555: 390396, 2003.41. Yamagishi N, Nishihori H, Ishihara K, Ohtsuka K, Hatayama T.Modulation of the chaperone activities of Hsc70/Hsp40 by Hsp105 andHsp105. Biochem Biophys Res Commun 272: 850855, 2000.

42. Yamaguchi T, Omatsu N, Matsushita S, Osumi T. CGI-58 interactswith perilipin and is localized to lipid droplets. Possible involvement ofCGI-58 mislocalization in Chanarin-Dorfman syndrome. J Biol Chem 279:3049030497, 2004.

43. Zeibig J, Karlic H, Lohninger A, Damsgaard R, Smekal G. Do bloodcells mimic gene expression profile alterations known to occur in muscularadaptation to endurance training? Eur J Appl Physiol 95: 96104, 2005.

44. Zhang XS, Lue YH, Guo SH, Yuan JX, Hu ZY, Han CS, Hikim AP,Swerdloff RS, Wang C, Liu YX. Expression of HSP105 and HSP60during germ cell apoptosis in the heat-treated testes of adult cynomolgusmonkeys (Macaca fascicularis). Front Biosci 10: 31103121, 2005.

45. Zieker D, Fehrenbach E, Dietzsch J, Fliegner J, Waidmann M, NieseltK, Gebicke-Haerter P, Spanagel R, Simon P, Niess AM, Northoff H.

cDNA-microarray analysis reveals novel candidate genes expressed inhuman peripheral blood following exhaustive exercise. Physiol Genomics23: 287294, 2005.

46. Zieker D, Zieker J, Dietzsch J, Burnet M, Northoff H, Fehrenbach E.cDNA-microarray analysis as a research tool for expression profiling inhuman peripheral blood following exercise. Exerc Immunol Rev 11:8696, 2005.


J Appl Physiol VOL 102 JANUARY 2007 www jap org

buttner - JAP 07 - microarray

Documents

Transcript of buttner - JAP 07 - microarray