cDNA microarray analysis reveals novel candidate genes expressed in human peripheral blood following exhaustive exercise

Derek Zieker1,*, Elvira Fehrenbach1,*, Janko Dietzsch2,*, Judith Fliegner1, Marc Waidmann1, Kay Nieselt2, Peter Gebicke-Haerter4, Rainer Spanagel4, Perikles Simon3, Andreas Michael Niess3 and Hinnak Northoff1

1 Department of Transfusions Medicine, University of Tuebingen
2 Center for Bioinformatics Tuebingen, Department of Information and Cognitive Sciences, University of Tuebingen
3 Department of Sports Medicine, Medical Clinic, University of Tuebingen
4 Central Institute for Mental Health, Department of Psychopharmacology, Mannheim, Germany


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
It is generally accepted that exhausting endurance exercise exhibits strong effects on the immune system. Such effects have been attributed to changes in the cellular composition of peripheral blood as well as to changes in the expression of plausible candidate genes. The list of candidate genes is far from being complete, since this issue has not yet been investigated in a systematic way. In this study, we used a custom-made cDNA microarray focused on inflammation as a screening approach to study gene expression in eight one-half marathon runners before, immediately after, and 24 h after exercise. Significant differential gene expression was verified by quantitative real-time PCR. Linear regression analysis showed that microarray expression analysis of cell type-specific surface molecules reflects the observed individual cellular shifts in peripheral blood cells with high statistical significance. In line with the results of former studies, we observed an upregulation of mitogen-activated protein kinase-activated protein kinase-2 (MAPKAP-K2), L-selectin, and IL-1 receptor antagonist (IL-1ra) after exhaustive exercise. The main results of this study report, for the first time, the downregulation of CD81; the upregulation of thioredoxin, which may play an important part in anti-oxidative defense; and, surprisingly, the downregulation of the anti-carcinogenic gene glutathione-S-transferase-3 (GSTM3) in peripheral blood. The study shows cDNA microarray expression analysis as a reliable systematic instrument to complete the list of candidate genes that may play a role in exhaustive exercise-induced modulation of the immune response.

thioredoxin; CD81; oxidative; marathon


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
HIGH TRAINING AND COMPETITION efforts lead to repeated stress responses in the immune system of athletes. This can cause a transient weakening of the immune system, in contrast to moderate endurance training, which is held to have a positive effect on the immune system (40). The reaction of athletes to exercise is a coordinated response of multiple organ systems (4).

In leukocytes, several different systems are activated or regulated after intensive, exhausting endurance exercise. A variety of genes are known to be involved in this process. These comprise the heat shock proteins as well as the cytokines (10, 15, 55).

After prolonged, exhaustive endurance exercise, an increase in different cytokines, like TNF-{alpha}, IL-1, IL-6, IL-8, and IL-1 receptor antagonist (IL-1ra), were observed in plasma or urine of athletes, with levels falling back to normal on the following day (20, 43, 45, 58, 62). Presently, IL-6, which also has major metabolic functions for the working muscle, is considered to be the major indicator for physical strain in sports medicine (45, 52, 57). Considerably elevated levels of cortisol and adrenaline were suggested to be at least partly responsible for these changes (21, 48). On the other hand, after endurance exercise, ex vivo cytokine production by peripheral blood mononuclear cells was shown to be significantly decreased, probably as a result of counterregulation (63).

Exercise also leads to increased generation of reactive oxygen and nitrogen species (ROS and RNS, respectively) (41). Besides their damaging potential, ROS play a physiological role in cellular signaling and in promoting lymphocyte apoptosis (21, 35).

To date, there is no proven knowledge about the molecular mechanisms by which exercise exerts its effects on the immune system. In humans, the most convenient source available to perform measurements of stress parameters on the cellular and molecular level is peripheral blood. cDNA arrays have evolved into widely used tools for comprehensive analysis of gene expression.

Using spotted cDNA chips, we decided to examine blood samples of one-half marathon runners before and after a run to look for changes in gene expression patterns. The aim, using microarrays, was to find hitherto unknown target genes involved in the reaction of the immune system to exercise. It is hoped that this may finally lead to an improved understanding of physiological and metabolic regulatory pathways that are influenced by endurance exercise.

The genes represented on the cDNA microarray were selected on the basis of up-to-date knowledge about their role in inflammation, apoptosis, stress response, and related pathways. It is also known that exhaustive exercise induces major transit changes in subpopulations of the peripheral blood.


    METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Eight well-trained male athletes [38.9 ± 11.8 yr, body mass index (BMI) 23.6 ± 1.8; see Table 1] performed an official one-half marathon under competition conditions (21.1 km). None of the athletes suffered from acute or chronic diseases or reported intake of medication, including antioxidants and nicotine abuse. The run started at 10:00 AM on a cool and humid December day (1°C) and took place on a hilly and demanding terrain. Blood samples (5 x 2.5-ml whole blood) were drawn in PAXgene blood RNA tubes (Qiagen, Hilden, Germany) at rest before (t0), immediately (up to 15 min) after (t1), and 24 h after (t2) exercise in a sitting position. The individuals had been engaged in specific endurance training for at least 2 yr (52.2 ± 25.5 km/wk, running).


View this table:
[in this window]
[in a new window]
 
Table 1. Anthropometric and training charastatistics of the individual subjects

 
These evacuated PAXgene-tubes were prefilled with a "stop"-reagent, which lyses all cells and stabilizes the RNA immediately, avoiding changes of the cellular RNA profile due to storage.

We compared the gene expression pattern of each athlete before and after exhaustive exercise.

Each subject gave written informed consent before participation in the study. The experimental protocols were approved by the Institute's Human Ethics Committee according to the principles set forth in the Declaration of Helsinki of the World Medical Association.

Preparation of Glass cDNA Microarrays
cDNA microarrays were produced in the Center for Mental Health in Mannheim, Germany. Human clones were purchased from the Ressourcen Zentrum für Genomforschung Library (RZPD) in Berlin, Germany. The final library contained 277 different genes (500–1,500 bp). Insert DNAs of all clones were amplified, each in a 50-µl PCR solution [10x PCR buffer (Qiagen), MgCl2 (Qiagen), Taq polymerase (Qiagen), RZPD vector primers] in 384-well microtiter plates. The plates were incubated for 3 min at 94°C, before 35 cycles of denaturation at 94°C for 75 s, annealing at 59.5°C for 90 s, and elongation at 72°C for 90 s were performed, followed by a final elongation step at 72°C for 10 min. The quality of the PCR products was routinely checked on a 1.5% agarose gel stained with ethidium bromide.

Each PCR sample was concentrated, and the pellet was taken up in 15 µl of 3x SSC plus 1.5 M betaine (Sigma, Munich, Germany) spotting solution. Stratagene's SpotReport-10 Array Validation System included human ß-actin PCR product (3'-sense and 5'-sense) (Stratagene), poly(dA) oligonucleotide (Stratagene), human COT-1 (Stratagene) and cDNA from Universal RNA (Stratagene), 3x SSC buffer, and 1.5 M betaine spotting solution, spotted as positive and negative controls together with the RZPD gene library, in triplicate, on QMT amino slides (75 x 25 mm) (Quantifoil, Jena, Germany). cDNAs were spotted in triplicate with a MicroGrid I spotter (BioRobotics, Cambridge, UK) supplied with 16 solid microspotting pins (BioRobotics). The slides were UV cross-linked with 250 mJ/cm2 and incubated at 80°C for 2 h.

Isolation, Amplification, and Labeling of RNA
Blood was drawn in PAXgene blood RNA tubes (Qiagen). RNA was isolated using the PAXgene blood RNA kit (Qiagen,) including a DNase I treatment, according to the manufacturer's recommendations. The quality of the RNA was checked with Agilent's Lab-on-a-Chip total RNA nanobiosizing assay (Agilent Technologies, Palo Alto, CA). Four micrograms of total RNA from each sample was spiked with Stratagene's SpotReport-10 Array Validation System RNA spikes (amounts per spike varying from 0.0025 to 2.5 ng) and reverse transcribed. Amplification was performed exactly according to the Eberwine procedure (61) using Ambion's MessageAmp kit (Ambion, Austin, TX). After in vitro transcription, the indirect labeling method with aminoallyl was used with equal amounts of all samples. The protocol follows exactly the Institute for Genomic Research (TIGR) protocol from Jeremy Hasseman called "Aminoallyl labeling of RNA for microarrays" (http://pga.tigr.org/sop/M004.pdf), using cyanine-3 (Cy3) and Cy5 ester (Amersham Biosciences, Amersham, UK). For each labeled sample, absorbance was measured at 260 nm, for Cy3 at 550 nm, and for Cy5 at 650 nm, respectively.

Hybridization of Glass cDNA Microarrays
After labeling, the samples were vacuum dried for 3 h. During this time, the cDNA arrays were treated with a prehybridization step according to the Quantifoil prehybridization protocol A (http://www.quantifoil.com), followed by a denaturing step, for arrays spotted with PCR probes [1 x 3 min in boiling distilled water (dH2O; 95–100°C)]. The vacuum-dried Cy3- and Cy5-labeled samples were resuspended together in 10 µl of 10 mM EDTA, incubated at 95°C for 5 min, and followed by incubation at room temperature for another 5 min. Prewarmed Ambion's Hybridization Buffer No. 1 (35 µl) was added to the samples and was incubated for 10 s at 68°C before given on the microarray. Slides were hybridized in hybridization chambers (TeleChem International, Sunnyvale, CA). Hybridization was done at 55°C for 14–16 h in a water bath. After hybridization, each slide was washed with 2x SSC + 0.2% SDS at 30°C for 10 min, 2x SSC at room temperature for 10 min, and 0.2 x SSC at room temperature for 10 min. Slides were dried immediately with N2.

To compare the gene expression pattern of each athlete before and after exhausting exercise, we hybridized each slide with samples from t0 vs. t1 and from t0 vs. t2. A dye swap was done with each investigation.

Scanning, Feature Extraction, and Analysis
Slides were scanned in a microarray scanner (Genetix Limited, Hampshire, UK).

Photomultiplier tube voltage was always set to 100% for both red and green channels. The two resulting green and red images were overlaid using ImaGene 5 (BioDiscovery).

Signal and Feature Extraction
As a first step of feature extraction, all spots that were flagged as poor by the ImaGene 5 software (due to signal foreground or background contamination, shape irregularity, or poor spot quality) were excluded.

The modus of the pixel distribution for the signal foreground and background of each spot was selected as a robust estimator of the raw signal values of both channels (see Fig. 1).



View larger version (59K):
[in this window]
[in a new window]
 
Fig. 1. Flow chart of the normalization procedure (see METHODS, Normalization). The dye-swapped arrays are shown at top. At left, the sample is in the red and the reference in the green channel; at right, the channels are flipped. After subarray-wise Loess correction, the log ratios MxI,MxII of the sample and reference intensities per spot were calculated. M',M'' and A',A'' were gained by averaging over all 3 identically spotted subarrays. Both estimators M' and M'' were used to calculate the finally used estimator for M via a dye-swap correction.

 
Normalization
The normalization was done in three steps (also see Fig. 1). First, a print tipwise Loess correction of the raw values was performed to reduce the spatial dependent intensity bias (6, 9, 64). Then, for each spot the log ratio

and the logarithmic average

of the intensities of the red (Chred) and the green (Chgreen) channels, respectively (6, 13), were calculated as the basis for the next normalization steps. Due to the dye-swap design, one array carries the sample s in the red channel and the reference r in the green channel, and the other array has the sample s in the green channel and the reference r in the red channel. This leads to two M values for each spot

. The use of the log (base 2) transformation has several advantages, like symmetrizing the spot intensity data, stabilizing the variances, and reducing the skewness of skewed distributions (14, 56), and it allows a natural interpretation of the resulting values as the logarithm of the fold change (FC), M = log2(FC). All three replicates (subarrays) of one sample and dye configuration were averaged by taking the arithmetical mean of the three estimators for M',M'' and A',A'' per transcript (see Fig. 1). Finally, the two arrays per sample with flipped dyes were used to execute a dye-swap correction of the log ratios M to remove remaining dye effects as follows (53, 60, 64)

Differentially Expressed Genes
Due to the direct design of the experiment, only log ratios M that lie statistically significantly above or below 0 mark correspond to differentially expressed genes. To investigate this, a one-sample t-test procedure against the mean 0 was used (13). The problem of multiple testing was addressed by controlling the family-wise error rate (FWER) and the false discovery rate (FDR) (13, 17). A gene was called differentially expressed if its Bonferroni-corrected (7) P value was <0.05 (5%). In addition, other correction methods like Benjamini and Hochberg (2), Benjamini and Yekutieli (3), Holm (22), and Sidak (54) were also applied and the results compared. The data discussed in this publication have been deposited in the National Center for Biotechnology Information (NCBI)'s Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE2532. A subset of interesting genes was selected for a further evaluation by real-time PCR.

Software
All parts of the feature extraction and analysis procedures were computationally done inside R 1.8.1 (50) with a mixture of Bioconductor structures (18) and self-written functions.

Quantitative real-time PCR.
Selected cDNAs were quantified by real-time PCR on a LightCycler instrument (Roche, Mannheim, Germany). For PCR-assisted amplification, the LightCycler Fast Start DNA Master SYBR Green I-kit (Roche) was used according to the manufacturer's instructions. Final concentrations of primers and magnesium were optimized for each set of primers (for primer sequences, see Table 2).


View this table:
[in this window]
[in a new window]
 
Table 2. Primer Sequences

 
Specificity of the PCR conditions was assured by determining the melting temperature of the respective amplicon previously shown to correspond to the expected product size by agarose gel analysis.

Relative quantification of gene expression levels was done by running an external standard curve together with the samples of interest. Standards were prepared by serially diluting a post-PCR amplicon of the respective templates. The linear range was determined for each PCR run by blotting the crossing point (CP) value calculated by the LightCycler software against the dilution factor of these standards, thus allowing the quantitative comparison of the samples. For samples within the linear range of the standard curve, the respective value was calculated using regression analysis. The ratio of specific sample to ß-actin values allowed the relative quantification of gene expression.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The average running time of the one-half marathon was 104.75 ± 21.24 min (see Table 1). Cell counts were significantly changed directly after the run and had returned to preexercise levels 24 h later (see Table 3).


View this table:
[in this window]
[in a new window]
 
Table 3. Changes of cell counts before and immediately and 24 h after one-half marathon

 
We found a number of genes that were regulated up or down immediately after the run. Comparison of preexercise cDNA microarray expression levels to 24 h postexercise levels revealed no significance for all transcripts that had been significantly altered immediately after exercise. All genes that were regulated with (P < 1) are listed in Table 4. The one-sample t-test in conjunction with the adjusted P values resulted in a set of six genes, which we refer to as differentially expressed on a stringent statistical level (adjusted P < 0.05). These genes were L-selectin, which was upregulated, and CD81, CD244, glutathione-S-transferase-3 (GSTM3), ICAM2, and integrin-{alpha}x, which were downregulated. Stringent statistics was used to minimize false-positive findings. We are aware of the fact that, by doing so, we increase the rate of false negatives.


View this table:
[in this window]
[in a new window]
 
Table 4. Fold changes and P values of differentially expressed genes detected by analysis of arrays

 
The six highly significant genes (P < 0.05) and four additional genes were selected for an evaluation using real-time PCR. The results are shown in Table 5 and Fig. 2. All of these genes achieved significance in real-time PCR except ICAM2 and CD244.


View this table:
[in this window]
[in a new window]
 
Table 5. Evaluation of differential expression between t0 and t1 by real-time PCR

 


View larger version (11K):
[in this window]
[in a new window]
 
Fig. 2. Box plots of the real-time PCR measurements as log2 values of the fold change between time points t1 (immediately after exercise) and t0 (before exercise). The box whisker plots visualize the minimum (end of the bottom whisker), the first quartile (bottom border of the box), the median (line through the box), the third quartile (top border of the box), and the maximum (end of the top whisker) of the distribution. The separately drawn points are outliers according to the standard definition used inside R (36). Points are regarded as outliers if the minimum of their distance to the first and the third quartile is greater than 1.5 times the interquartile range (IQR = third quartile – second quartile).

 
We performed linear regression analysis to correlate the x-fold expression changes of cell type-specific surface molecules observed by microarray analysis with the observed individual cellular x-fold shifts of peripheral blood cells. As presented in Table 6, differential expression of all relevant specific cell surface markers significantly correlated with the respective x-fold change of cell counts.


View this table:
[in this window]
[in a new window]
 
Table 6. Fold changes of gene expression, fold changes of cell shift, and their correlation

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
During and some hours after exhaustive endurance exercise like marathon, one-half marathon, or triathlon, a number of peripheral immunological parameters have been demonstrated to change significantly (38). These include induction of cytokines and hormones as well as induction of tolerance to pathogenic stimuli and changes in natural killer (NK) cell activity and in adhesion markers in addition to changes in numbers of cells. Changes observed may serve to supply energy for the working muscles or help to curb undue inflammatory reactions to muscular stress, and may also be involved in staleness or in the transient immunosuppression that can occur after exhaustive endurance exercise (63). Usually, most of the changes return back to normal after 24 h (38).

Using an "inflammation-centered" cDNA chip, we intended to find evidence and new indications to further clarify the area. We did not expect to find all exercise-induced alterations of immunological parameters that are described in the literature to come out as significant changes in our experiments, since these changes may occur at different times or even outside peripheral blood (e.g., IL-6). Despite applying very stringent statistics for microarray analysis to avoid false positives, we found interesting candidate genes that showed significant regulation. Using real-time PCR, we confirmed most of the candidate genes. These and some further genes of special interest shall be discussed in detail below, trying to find out possible relevance in the given setting.

The present data reveal up- and downregulated transcripts in peripheral blood cells of one-half marathon runners immediately after the run compared with preexercise levels. Most of these significantly changed mRNAs encode proteins hitherto unknown to be involved in exercise physiology. Nevertheless, these transcripts neatly fit into the context of exercise stress, including oxidative stress and protective proteins [thioredoxin (TRX), GSTM3, IL-1ra], cellular contact mechanisms (CD81, adhesion proteins), signal transduction, and cellular protection [mitogen-activated protein kinase-activated protein kinase-2 (MAPKAP-K2), TRX]. Some of these genes were also shown to be significantly changed in peripheral blood mononuclear cells (PBMC) by Connolly et al. (10) using Affymetrix arrays.

Gene Expression Changes in Concordance with Cellular Shifts
Being aware of the exercise-induced changes of cell counts that may influence mRNA expression levels of whole blood cells, as measured here, we start by discussing those genes in which changes in the microarray-analyzed gene expression were closely paralleled by relevant cellular shifts. This concerns mainly genes for cell surface markers that are differently expressed in individual cell populations of peripheral blood. Sonna et al. (55) already stated that cell type-specific microarray-detected gene expression changes after heat injury may reflect cellular shifts rather than changes within a cell.

In our study it pertains to L-selectin, which was significantly changed, but also to some additional genes like CD1c, CD2, CD3e, CD14, and CD19, which show a strong trend but did not reach significance (see Table 4).

L-Selectin (CD62L) mRNA was upregulated significantly in this study. This pattern corresponds to the situation known from the literature (30), where it was shown by flow cytometric analysis that, during recovery after exercise, the expression of CD62L on granulocytes was significantly increased. However, in this study a 3.4-fold increase in gene expression may well be attributable to the increase in granulocyte numbers (4.3-fold). CD62L is strongly expressed on granulocytes.

CD1c, CD2, CD3e, and CD19 are markers of lymphocytes. All of these were downregulated in nearly the same range as the downshift of lymphocytes, which was ~45% in this study. In contrast to this, CD14 as a monocyte marker, which was upregulated by ~88%, nicely reflected changes in numbers of monocytes, which were increased by 69% after exercise.

The expression of CD244, another surface molecule, may also have been altered due to cellular shifts. CD244 is the NK cell receptor 2B4 and is expressed as coreceptor on human NK cells, on macrophages, and on a subset of CD8+ T cells (36). It is known that numbers of NK cells in peripheral blood increase immediately after exhaustive exercise but can show a fast and sharp drop with high individual variation (39). Further expression studies should show whether CD244 data parallel the kinetics of NK cell numbers.

The finding of a highly significant correlation of the fold changes of cell surface markers with the respective fold changes in cell counts underscores the accuracy of the performed methodology, but nevertheless pinpoints the problem to securely separate expression differences from the bias induced by cellular shifts (see Table 6). This is not only a particular problem of this study (10, 55) but may also be true for microarray studies of solid tissue, for instance, for the frequently performed comparison of tumor with benign tissue, where differences in the presence of immune competent cells may severely bias expression analysis.

Gene Expression Changes Not Attributable to Obvious Cellular Shifts
Cellular communication.
Integrin-{alpha}x (CD11c) and CD81 mRNA were both downregulated significantly in this study. This cannot be attributed to cellular shifts, since CD11c is strongly or moderately expressed in monocytes, macrophages, NK cells and granulocytes, while it is low in T and B cells (26), and CD81 is expressed on polymorphonuclears (PMN), monocytes, and B and T lymphocytes (59). The lymphocytes are the minor population and therefore can hardly explain the strong 50% overall downregulation of gene expression.

Regular moderate exercise leads to a decreased expression of adhesion proteins, especially ß2-integrins (26). It was suggested that the decreased expression of adhesion molecules immediately after exercise negatively influences adhesion and transmigration of the leukocytes into the vessel wall and the surrounding tissue. Shortly after exercise, this may contribute to a reduced defense against infectious agents (37).

To our knowledge, the downregulation of the CD81 gene expression has not yet been demonstrated in context with exhaustive exercise. The protein encoded by this gene is a member of the transmembrane 4 super family (tetraspanins) involved in signal transduction events in the regulation of cell development, activation, growth, and motility. The tetraspanins are common components of exosomes. The CD81 protein may play an important role in the reaction or adaptation to strenuous exercise. As mentioned above, reduced interaction with integrins may contribute to downregulation of undue proinflammatory stimuli resulting from acute muscular (cellular) stress.

Integrins also play a role in the cell contact area during interaction between T cells and antigen presenting cells (APC), which is termed immunological synapse (IS) (33). During formation of IS, the tetraspanin CD81 is redistributed to the contact area of the cells in an antigen-dependent manner and is involved in the antigen-specific B cell-mediated immune response (33). Furthermore, the presence of CD81 at the IS may play an essential role during TH1/TH2 polarization (11). Downregulated CD81 protein may partly contribute to the attenuated cell-mediated immunity and a change of the TH1/TH2 balance in favor of TH2 after exhaustive exercise, rendering the athlete more susceptible to infections. However, CD81 mRNA and corresponding protein changes in individual leukocyte subpopulations still warrant further analyses.

Signal transduction and cellular protection.
In contrast to the cellular communication genes, signal transduction and cellular protection cannot be clearly related to individual subpopulations of peripheral blood. In this study, MAPKAP-K2, TRX, GSTM3, and IL-1ra were changed and confirmed as significantly altered by real-time PCR.

The finding that the gene expression level of MAPKAP-K2 is significantly upregulated in athletes after exhausting endurance exercise has already been shown for the muscle (65).

MAPKAP-K2 is a downstream substrate in the pathway of the MAPK family. The MAPKAP-K2 protein phosphorylates the small heat shock protein 27 (HSP27, HSPB1) (16, 49, 65). It is also known that MAPKAP-K2 is involved in IL-6 synthesis by stabilizing IL-6 mRNA (34). A recent study showed that MAPKAP-K2 is an immediate early gene that is posttranslationally regulated by the mRNA stabilizing protein HuR, which may explain the early and isolated increase in peripheral blood as reported here (32).

We recently showed that inducible nitric oxide synthase (iNOS), IL-6, and HSP27 are upregulated after exercise at mRNA and protein levels (15, 16, 44). It remains to be shown which role the upregulation of MAPKAP-K2 mRNA as demonstrated in this study may play for the posttranscriptional modification of these and possibly other substrates in exercise. Interestingly, hsp27 mRNA (HSPB1) is 1.8-fold upregulated in our study and is included in our list of differentially expressed genes (see Table 4).

The finding of increased TRX mRNA after the one-half marathon is new but does not come entirely unexpected. The TRX protein is a 12-kDa oxidoreductase enzyme containing a dithiol-disulfide active site and is released from various types of human cells, excluding neutrophils (28), in response to oxidative stress (25). It possesses a variety of biological functions including the ability to modulate transcription factor activity (5, 23). The multifunctional redox-active TRX scavenges free radicals, modulates chemotaxis and suppresses leukocyte infiltration into sites of inflammation, and protects cells against TNF-induced cytotoxicity and general oxidative stress (12, 24). The TRX enzyme was found to be elevated in patients with oxidative stress profiles (29, 47). Similarly, heavy exercise is associated with oxidative stress (1, 31, 41, 42). Therefore, the regulation and maintenance of the cellular redox status by intracellular redox-regulating molecules such as TRX become important to maintain tissue homeostasis also after intensive exercise. The TRX enzyme may represent an important anti-oxidative protection mechanism in peripheral blood cells after exercise stress. It is important to mention, however, that ROS generated through cellular metabolism do have positive functions as cellular second messengers through the regulation of numerous signal transduction pathways. The level of ROS needs to be carefully balanced to fulfill their positive tasks without producing undue damage. An upregulation of TRX has already been shown to enable the maintenance of such low and protective ROS levels (19).

Interestingly, another part of the oxidative stress response is GSTM3, which we demonstrate for the first time to be downregulated on gene expression level after endurance exercise. Glutathione transferases (GSTs) form a group of multigene isoenzymes involved in the cellular detoxification of both xenobiotic and endobiotic compounds. More than a dozen genes are known to encode GSTs. In this context, GSTs bind to or catalyze the inactivation of a variety of carcinogens (8). GST gene expression studies in humans have shown that polymorphisms in members of the GST gene family are associated with an increased risk of lung cancer (51). Levels of GST isoform expression have been shown to vary between different tissue and cell types. Within the peripheral blood, expression of GSTM3 mRNA has so far only been demonstrated in patients with acute lymphoblastic leukemia, where its expression was related to good prognosis (27). The role of GSTM3 in exercise immunology remains to be determined.

An additional gene that may also be viewed as part of the cellular protection mechanism is IL-1ra. The main assignment of IL-1ra is the regulation and the control of the effects caused by IL-1ß by a competitive inhibition of the receptor. An overbalance of IL-1ß at the receptor leads to inflammation and tissue destruction. This anti-inflammatory cytokine is known to be increased in serum by prolonged exercise (46, 52, 57). Upregulation of IL-1ra mRNA in our study confirms that IL-1ra as opposed to IL-6 is generated in peripheral blood cells. Changes of IL-1ra mRNA in PBMC after exercise have already been shown (10).

The upregulation of IL-1ra may play an important role for the balance between the proinflammatory and the anti-inflammatory cytokines in the immune system.

In conclusion, we have investigated gene expression in peripheral blood of eight marathon runners using a cDNA microarray focused on inflammatory markers. A number of genes regulated in response to exercise have been identified. We presently have no information about corresponding protein changes. Some of these genes may have been altered as a consequence of cellular shifts. Furthermore, three genes have been found of particular interest that have not yet been described in relation to exercise. These are downregulated CD81, which may attenuate inflammatory reaction; upregulated TRX, which may play an important part in anti-oxidative stress reaction; and downregulated GSTM3, another gene involved in balancing the redox system.

It is beyond the scope of this study to carry out additional investigations on up- or downstream genes. Moreover, correlations of gene transcript abundance with corresponding protein levels will have to be done in the future.

Altogether, the reliability of our microarray results is confirmed by the good reproducibility in real-time PCR measurements. The nice reflection of cellular shifts in some of the transcript changes and the confirmation of several previously known effects strongly support the notion that such microarrays may serve as a suitable screening tool for exercise monitoring. Microarray analysis of separated cell subsets, excluding cell shift-associated changes, should be performed in the future.


    ACKNOWLEDGMENTS
 
We thank the athletes who volunteered to participate in the study.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: D. Zieker, Dept. of Transfusions Medicine, Univ. of Tuebingen, Otfried-Mueller-Str. 4/1, D-72076 Tuebingen, Germany (e-mail : dzieker{at}web.de)

* D. Zieker, E. Fehrenbach, and J. Dietzsch contributed equally to this study. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Alessio HM. Exercise-induced oxidative stress. Med Sci Sports Exerc 25: 218–224, 1993.[ISI][Medline]
  2. Benjamini Y and Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B 57: 289–300, 1995.[ISI]
  3. Benjamini Y and Yekutieli D. The control of the false discovery rate in multiple hypothesis testing under dependency. Annals Stat 29: 1165–1188, 2001.[CrossRef][ISI]
  4. Bernstein D. Exercise assessment of transgenic models of human cardiovascular disease. Physiol Genomics 13: 217–226, 2003.[Abstract/Free Full Text]
  5. Bjornstedt M, Xue J, Huang W, Akesson B, and Holmgren A. The thioredoxin and glutaredoxin systems are efficient electron donors to human plasma glutathione peroxidase. J Biol Chem 269: 29382–29384, 1994.[Abstract/Free Full Text]
  6. Bolstadt BM, Irizzary RA, strand M, and Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19: 185–193, 2003.[Abstract/Free Full Text]
  7. Bonferroni CE. Teoria statistica delle classi e calcolo della probabilita. Pubblicazioni del Instituto Superiore di Scienze Economiche e Commerciali di Firenze 8: 3–62, 1936.
  8. Chanas SA, Jiang Q, McMahon M, McWalter GK, McLellan LI, Elcombe CR, Henderson CJ, Wolf CR, Moffat GJ, Itoh K, Yamamoto M, and Hayes JD. Loss of the Nrf2 transcription factor causes a marked reduction in constitutive and inducible expression of the glutathione S-transferase Gsta1, Gsta2, Gstm1, Gstm2, Gstm3 and Gstm4 genes in the livers of male and female mice. Biochem J 365: 405–416, 2002.[CrossRef][ISI][Medline]
  9. Cleveland WS. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 74: 829–839, 1979.[ISI]
  10. Connolly PH, Caiozzo VJ, Zaldivar F, Nemet D, Larson J, Hung SP, Heck JD, Hatfield GW, and Cooper DM. Effects of exercise on gene expression in human peripheral blood mononuclear cells. J Appl Physiol 97: 1461–1469, 2004.[Abstract/Free Full Text]
  11. Deng J, Dekruyff RH, Freeman GJ, Umetsu DT, and Levy S. Critical role of CD81 in cognate T-B cell interactions leading to Th2 responses. Int Immunol 14: 513–523, 2002.[Abstract/Free Full Text]
  12. Djavaheri-Mergny M, Javelaud D, Wietzerbin J, and Besancon F. NF-{kappa}B activation prevents apoptotic oxidative stress via an increase of both thioredoxin and MnSOD levels in TNF{alpha}-treated Ewing sarcoma cells. FEBS Lett 578: 111–115, 2004.[CrossRef][ISI][Medline]
  13. Dudoit S, Yang YH, Callow MJ, and Speed TP. Statistical methods for identifying genes with differential expression in replicated cDNA microarray experiments. Statistica Sinica 12: 111–139, 2002.[ISI]
  14. Durbin BP, Hardin JS, Hawkins DM, and Rocke DM. A variance-stabilizing function for gene-expression microarray data. Bioinformatics 18: 105–110, 2002.
  15. Fehrenbach E and Niess AM. Role of heat shock proteins in the exercise response. Exerc Immunol Rev 5: 57–77, 1999.[ISI][Medline]
  16. Fehrenbach E, Niess AM, Schlotz E, Passek F, Dickhuth HH, and Northoff H. Transcriptional and translational regulation of heat shock proteins in leukocytes of endurance runners. J Appl Physiol 89: 704–710, 2000.[Abstract/Free Full Text]
  17. Ge Y, Dudoit S, and Speed TP. Resampling-based Multiple Testing for Microarray Data Analysis. Sociedad Española de Estadistica e Investigación Operativa Test 12: 1–77, 2003.
  18. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Ellis B, Gautier L, Ge Y, Gentry J, Hornik K, Hothorn T, Huber W, Iacus S, Irizarry R, Leisch F, Li C, Maechler M, Rossini AJ, Sawitzki G, Smith C, Smyth G, Tierney L, Yang JYH, and Zhang J. Bioconductor: open software development for computational biology and bioinformatics. Genome Biology 5: R80, 2004.[CrossRef][Medline]
  19. Haendeler J, Tischler V, Hoffmann J, Zeiher AM, and Dimmeler S. Low doses of reactive oxygen species protect endothelial cells from apoptosis by increasing thioredoxin-1 expression. FEBS Lett 577: 427–433, 2004.[CrossRef][ISI][Medline]
  20. Hoffman-Goetz L and Pedersen BK. Exercise and the immune system: a model of the stress response? Immunol Today 15: 382–387, 1994.[CrossRef][ISI][Medline]
  21. Hoffman-Goetz L and Quadrilatero J. Treadmill exercise in mice increases intestinal lymphocyte loss via apoptosis. Acta Physiol Scand 179: 289–297, 2003.[CrossRef][ISI][Medline]
  22. Holm S. A simple sequentially rejective multiple test procedure. Scand J Statist 6: 65–70, 1979.[ISI]
  23. Holmgren A and Bjornstedt M. Thioredoxin and thioredoxin reductase. Methods Enzymol 252: 199–208, 1995.[CrossRef][ISI][Medline]
  24. Hoshino T, Nakamura H, Okamoto M, Kato S, Araya S, Nomiyama K, Oizumi K, Young HA, Aizawa H, and Yodoi J. Redox-active protein thioredoxin prevents proinflammatory cytokine- or bleomycin-induced lung injury. Am J Respir Crit Care Med 168: 1075–1083, 2003.[Abstract/Free Full Text]
  25. Jikimoto T, Nishikubo Y, Koshiba M, Kanagawa S, Morinobu S, Morinobu A, Saura R, Mizuno K, Kondo S, Toyokuni S, Nakamura H, Yodoi J, and Kumagai S. Thioredoxin as a biomarker for oxidative stress in patients with rheumatoid arthritis. Mol Immunol 38: 765–772, 2002.[CrossRef][ISI][Medline]
  26. Jordan J, Beneke R, Hutler M, Veith A, Haller H, and Luft FC. Moderate exercise leads to decreased expression of ß1 and ß2 integrins on leucocytes. Eur J Appl Physiol Occup Physiol 76: 192–194, 1997.[CrossRef][ISI][Medline]
  27. Kearns PR, Chrzanowska-Lightowlers ZM, Pieters R, Veerman A, and Hall AG. Mu class glutathione S-transferase mRNA isoform expression in acute lymphoblastic leukaemia. Br J Haematol 120: 80–88, 2003.[CrossRef][ISI][Medline]
  28. Kinnula VL, Soini Y, Kvist-Makela K, Savolainen ER, and Koistinen P. Antioxidant defense mechanisms in human neutrophils. Antioxid Redox Signal 4: 27–34, 2002.[CrossRef][ISI][Medline]
  29. Kumagai S, Jikimoto T, and Saegusa J. Pathological roles of oxidative stress in autoimmune diseases. Rinsho Byori 51: 126–132, 2003.[Medline]
  30. Kurokawa Y, Shinkai S, Torii J, Hino S, and Shek PN. Exercise-induced changes in the expression of surface adhesion molecules on circulating granulocytes and lymphocytes subpopulations. Eur J Appl Physiol Occup Physiol 71: 245–252, 1995.[CrossRef][ISI][Medline]
  31. Liu J, Yeo HC, Overvik DE, Hagen T, Doniger SJ, Chu DW, Brooks GA, and Ames BN. Chronically and acutely exercised rats: biomarkers of oxidative stress and endogenous antioxidants. J Appl Physiol 89: 21–28, 2000.[Abstract/Free Full Text]
  32. Lopez de Silanes I, Zhan M, Lal A, Yang X, and Gorospe M. Identification of a target RNA motif for RNA-binding protein HuR. Proc Natl Acad Sci USA 101: 2987–2992, 2004.[Abstract/Free Full Text]
  33. Mittelbrunn M, Yanez-Mo M, Sancho D, Ursa A, and Sanchez-Madrid F. Cutting edge: dynamic redistribution of tetraspanin CD81 at the central zone of the immune synapse in both T lymphocytes and APC. J Immunol 169: 6691–6695, 2002.[Abstract/Free Full Text]
  34. Miyazawa K, Mori A, Miyata H, Akahane M, Ajisawa Y, and Okudaira H. Regulation of interleukin-1ß-induced interleukin-6 gene expression in human fibroblast-like synoviocytes by p38 mitogen-activated protein kinase. J Biol Chem 273: 24832–24838, 1998.[Abstract/Free Full Text]
  35. Mooren FC, Lechtermann A, and Volker K. Exercise-induced apoptosis of lymphocytes depends on training status. Med Sci Sports Exerc 36: 1476–1483, 2004.[ISI][Medline]
  36. Nakajima H and Colonna M. 2B4: an NK cell activating receptor with unique specificity and signal transduction mechanism. Hum Immunol 61: 39–43, 2000.[CrossRef][ISI][Medline]
  37. Nielsen HG and Lyberg T. Long-distance running modulates the expression of leucocyte and endothelial adhesion molecules. Scand J Immunol 60: 356–362, 2004.[CrossRef][ISI][Medline]
  38. Nieman DC, Berk LS, Simpson-Westerberg M, Arabatzis K, Youngberg S, Tan SA, Lee JW, and Eby WC. Effects of long-endurance running on immune system parameters and lymphocyte function in experienced marathoners. Int J Sports Med 10: 317–323, 1989.[ISI][Medline]
  39. Nieman DC, Miller AR, Henson DA, Warren BJ, Gusewitch G, Johnson RL, Davis JM, Butterworth DE, and Nehlsen-Cannarella SL. Effects of high- vs. moderate-intensity exercise on natural killer cell activity. Med Sci Sports Exerc 25: 1126–1134, 1993.[ISI][Medline]
  40. Nieman DC and Nehlsen-Cannarella SL. The immune response to exercise. Semin Hematol 31: 166–179, 1994.[ISI][Medline]
  41. Niess AM, Dickhuth HH, Northoff H, and Fehrenbach E. Free radicals and oxidative stress in exercise—immunological aspects. Exerc Immunol Rev 5: 22–56, 1999.[ISI][Medline]
  42. Niess AM, Fehrenbach E, Northoff H, and Dickhuth HH. Reactive species and oxidative stress in exercise and training adaptation—an update. Deut Zeitschr Sportmed 53: 345–353, 2002.[ISI]
  43. Niess AM, Passek F, Lorenz I, Schneider EM, Dickhuth HH, Northoff H, and Fehrenbach E. Expression of the antioxidant stress protein heme oxygenase-1 (HO-1) in human leukocytes. Free Radic Biol Med 26: 184–192, 1999.[CrossRef][ISI][Medline]
  44. Niess AM, Sommer M, Schlotz E, Northoff H, Dickhuth HH, and Fehrenbach E. Expression of the inducible nitric oxide synthase (iNOS) in human leukocytes: responses to running exercise. Med Sci Sports Exerc 32: 1220–1225, 2000.[CrossRef][ISI][Medline]
  45. Northoff H, Weinstock C, and Berg A. The cytokine response to strenuous exercise. Int J Sports Med 15, Suppl 3: S167–S171, 1994.[ISI][Medline]
  46. Ostrowski K, Rohde T, Asp S, Schjerling P, and Pedersen BK. Pro- and anti-inflammatory cytokine balance in strenuous exercise in humans. J Physiol 515: 287–291, 1999.[Abstract/Free Full Text]
  47. Park YS, Fujiwara N, Koh YH, Miyamoto Y, Suzuki K, Honke K, and Taniguchi N. Induction of thioredoxin reductase gene expression by peroxynitrite in human umbilical vein endothelial cells. Biol Chem 383: 683–691, 2002.[CrossRef][ISI][Medline]
  48. Pedersen BK and Hoffman-Goetz L. Exercise and the immune system: regulation, integration, and adaptation. Physiol Rev 80: 1055–1081, 2000.[Abstract/Free Full Text]
  49. Polanowska-Grabowska R and Gear AR. Heat-shock proteins and platelet function. Platelets 11: 6–22, 2000.[CrossRef][ISI][Medline]
  50. R Development Core Team. A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing, 2004.
  51. Reszka E and Wasowicz W. Significance of genetic polymorphisms in glutathione S-transferase multigene family and lung cancer risk. Int J Occup Med Environ Health 14: 99–113, 2001.[Medline]
  52. Ronsen O, Lea T, Bahr R, and Pedersen BK. Enhanced plasma IL-6 and IL-1ra responses to repeated vs. single bouts of prolonged cycling in elite athletes. J Appl Physiol 92: 2547–2553, 2002.[Abstract/Free Full Text]
  53. Sanchez-Cabo F, Prokesch A, Thallinger GG, Pieler R, Trajanoskie Z, Butcher PD, Hinds J, Holmes LEA, Campbell SG, Ashe MP, Hubbard S, Cho KH, and Wolkenhauer O. Assessing the efficiency of dye-swap normalization to remove systematic bias from two-color microarray data. Appl Bioinformatics. In press.
  54. Sidak Z. Rectangular confidence regions for the means of multivariate normal distributions. J Am Stat Assoc 62: 626–633, 1967.[ISI]
  55. Sonna LA, Wenger CB, Flinn S, Sheldon HK, Sawka MN, and Lilly CM. Exertional heat injury and gene expression changes: a DNA microarray analysis study. J Appl Physiol 96: 1943–1953, 2004.[Abstract/Free Full Text]
  56. Speed T. Speed Group Microarray Page, "Always log spot intensities and ratios," http://www.stat.berkeley.edu/users/terry/zarray/Html/log.html (2000).
  57. Steensberg A, Fischer CP, Keller C, Moller K, and Pedersen BK. IL-6 enhances plasma IL-1ra, IL-10, and cortisol in humans. Am J Physiol Endocrinol Metab 285: E433–E437, 2003.[Abstract/Free Full Text]
  58. Suzuki K, Nakaji S, Yamada M, Totsuka M, Sato K, and Sugawara K. Systemic inflammatory response to exhaustive exercise. Cytokine Kinetics Exerc Immunol Rev 8: 6–48, 2002.
  59. Tohami T, Drucker L, Radnay J, Shapira H, and Lishner M. Expression of tetraspanins in peripheral blood leukocytes: a comparison between normal and infectious conditions. Tissue Antigens 64: 235–242, 2004.[CrossRef][ISI][Medline]
  60. Tseng G, Oh M, Rohlin L, Liao J, and Wong W. Issues in cDNA microarray analysis: quality filtering, channel normalization, models of variations and assesment of gene effects. Nucleic Acids Res 29: 2549–2557, 2001.[Abstract/Free Full Text]
  61. Van Gelder RN, von Zastrow ME, Yool A, Dement WC, Barchas JD, and Eberwine JH. Amplified RNA synthesized from limited quantities of heterogeneous cDNA. Proc Natl Acad Sci USA 87: 1663–1667, 1990.[Abstract/Free Full Text]
  62. Vikingsson A, Pederson K, and Muller D. Enumeration of IFN-{gamma} producing lymphocytes by flow cytometry and correlation with quantitative measurement of IFN-{gamma}. J Immunol Methods 173: 219–228, 1994.[CrossRef][ISI][Medline]
  63. Weinstock C, Konig D, Harnischmacher R, Keul J, Berg A, and Northoff H. Effect of exhaustive exercise stress on the cytokine response. Med Sci Sports Exerc 29: 345–354, 1997.[ISI][Medline]
  64. Yang YH, Dudoit S, Luu P, and Speed TP. Speed Group Microarray Page, "Normalization for cDNA microarray data," http://www.stat.berkeley.edu/users/terry/zarray/Html/normspie.html (2001).
  65. Yu M, Blomstrand E, Chibalin AV, Krook A, and Zierath JR. Marathon running increases ERK1/2 and p38 MAP kinase signalling to downstream targets in human skeletal muscle. J Physiol 536: 273–282, 2001.[Abstract/Free Full Text]




This Article
Abstract
Full Text (PDF)
All Versions of this Article:
23/3/287    most recent
00096.2005v1
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Zieker, D.
Articles by Northoff, H.
PubMed
PubMed Citation
Articles by Zieker, D.
Articles by Northoff, H.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2005 by the American Physiological Society.