Animals
Twenty-nine male Wistar rats were purchased at the age of 6 weeks from Charles River Laboratories (Sulzfeld, Germany) and were housed under controlled environmental conditions (21°C, 12:12-h light-dark cycle). Food (standard rodent chow from Ssniff; Soest, Germany) and water were given ad libitum. The animals were maintained according to the policy statement with respect to the Declaration of Helsinki. The study design was approved by the Regional Administration of the City of Cologne (Bezirksregierung Köln).

Voluntary running model
Our model was described previously in detail (Matsakas et al., 2004). In brief, after 3 days of acclimatization the animals were randomly divided into two groups: an exercise (n = 20) and a control (n = 9) group. The exercising animals were housed individually in cages with free access to a running wheel to assure stress-free exercise. The exercise training period lasted 12 weeks. During this period, the control group was housed individually in plain cages. The body mass of all animals was monitored weekly.

Tissue collection
After completion of the 12 week period all animals were decapitated at the same time window (9:00-11:00 a. m.). Wheels and food had been removed from the cages 12h earlier to minimize the influence of the last exercise bout and the last feeding on the (molecular) targets of interest. For details of the serum extraction see Matsakas et al., 2004. The whole colon was separated, rinsed with Ringer´s solution and prepared to be free of fat and feaces. Afterwards the colon was turned inside out to scrape off the mucosa. The isolated mucosa was directly put into liquid nitrogen and stored at -80°C for further analysis. Finally, the heart (without the great vessels) was removed and weighed immediately.

Serum IGF-1 analysis
Serum insulin-like growth factor 1 (IGF-1) was detected via enzyme immunoassay (DRG, Marburg, Germany) in an Anthos 2000 photometer (Salzburg, Austria). The sensitivity was about 30 ng ml^-1 and the intra- and inter-assay coefficient of variation were 7.4 and 9.5 %, respectively (Matsakas et al., 2004).

mRNA analysis
Frozen mucosa was homogenized in 2 ml Lysing Matrix D tubes (Q- BIOgene, Irvine, CA) filled with provided 1.4 mm ceramic spheres and additionally filled up with buffer

for cell lysis (Macherey-Nagel, Düren, Germany). For the pulverizing process a FastPrep^® FP120A Instrument (Q-BIOgene, Irvine, CA) was used. Total RNA was isolated using the NucleoSpin^® RNA II-Kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. A DNA digestion step was included. Isolated RNA was diluted in 60 µl of RNAse-free water. RNA concentration was measured photometrically and purity was checked through the ratio of optical density at 260 and 280 nm (Biophotometer Eppendorff; Hamburg, Germany). The quality of the isolated RNA was checked by gel electrophoresis (1 % agarose, formaldehyde containing) through interpretation of 18 S and 28 S bands. Only high quality material was accepted and used for further analysis.

cDNA-synthesis: A mix of 0.5 µg RNA, 8 µl 5x MMLV reaction buffer (Promega, Mannheim, Germany), 6µl dNTP´s (300 µM; Fermentas, St. Leon-Rot, Germany) and an adequate amount of nuclease-free water was prepared to a final volume of 30 µl. The mixture was denatured by incubation at 65°C for 5 minutes. Afterwards, 0.4 µl of random hexamers (0.2 µg·µl^-1; Fermentas, St. Leon-Rot, Germany), 0.63 µl RNAse inhibitor (20 U·µl^-1; Fermentas, St. Leon-Rot, Germany), 1 µl MMLV Reverse Transcriptase (200 U·µl^-1; Promega, Mannheim, Germany) and 7.37 µl nuclease-free water were added. Finally, the samples were incubated at 37°C for one hour followed by 1 minute at 99°C.

Primer design: Highly purified salt-free primers from MWG (Ebersberg, Germany) were used. The final concentration of the primer working solution was 20 µM. The primer sets described in Table 1 were designed based on the LightCycler Probe Design Software, Version 1.0 (Idaho Technology Inc., Salt Lake City, USA). All primer pairs were set at an exon-intron barrier. Optimal annealing temperature (AT) and quantification temperature (QT) were established by using the LightCycler (Roche, Germany). Melting curve analysis and final agarose gel electrophoresis were used to set the experimental values. The quantification temperature was set below the specific melting curve of the product.

Quantitative real-time RT-PCR: For the final gene expression step the SYBR Green I - Kit (Roche, Mannheim, Germany) was used. After denaturation of the cDNA at 65°C for 5 minutes, a standard reaction mix for each sample was prepared: 1 µl cDNA (12.5 ng), 6.4 µl nuclease-free water, 1.2 µl MgCl₂ (final concentration: 4 mM), 0.2 µl of each primer (final concentration: 4 pmol) and 1 µl 10x LCM-reaction mix (Roche, Mannheim, Germany). Samples were analyzed with the LightCycler (Roche, Mannheim, Germany). A standard protocol was used for the specific gene amplification: the initial polymerase activation step was at 95°C for 190 seconds. For further denaturation (95°C, 15 seconds), annealing (AT, 10 seconds), elongation (72°C, 20 seconds), and quantification (QT, 5 seconds) steps the above mentioned AT and QT were used individually for each primer pair. To evaluate the specific amplification a final melting curve analysis (from AT up to 99°C) was added under continuous fluorescence measurements. Relative quantification was performed using sample crossing points as described (Rasmussen, 2001) and analyzed by LightCycler Software 3.5 (Roche, Mannheim, Germany). The method of choice was the "second derivative maximum" method (Rasmussen, 2001). Following data analysis was performed by two different Excel based applications: BestKeeper (Pfaffl et al., 2004) and Rest© (Pfaffl et al., 2002). Using these two programs the data were checked for statistical significance, normality and reliability.

Calculations and statistics
Values are presented as mean ± standard deviation. IGF-1 values were adjusted for body weight by regression analysis. One-way analysis of variance (factor: group) was used to compare the subgroups with the control group in the cases of serum IGF-1 and body mass. A post-hoc test was used to assess significant differences at the p < 0.05 level. mRNA data were analyzed using the BestKeeper program (Pfaffl et al., 2004) based on the calculation of a housekeeping gene index from among several housekeeping genes (AldolaseA, ß-actin, GAPDH). A second program, Rest©, (Pfaffl et al., 2002) was used for transformation of raw gene expression data into a normalized x-fold expression ratio of a target gene compared to the housekeeping genes. The examination of statistical significance (P < 0.05) in this case was done by a Pair Wise Fixed Reallocation Randomisation Test© as described (Pfaffl et al., 2002). Correlation analysis was performed by Pearson product moment correlation test. The level of statistical significance was set at α = 0.05. The statistical analysis was performed with SigmaStat 3.0 (SPSS Inc.) software if not described otherwise.

Body mass
Table 2 shows the final body mass of the animals after the experimental period. For details see Matsakas et al., 2004 and Buehlmeyer et al., 2007.

Running activity
The exercise group was divided into a low (L-EX; <2630 m per night; n = 5), a medium (M-EX; 3000-7460 m per night; n = 10) and a high (H-EX; >8310 m per night; n = 5) exercise volume group. For additional information see Matsakas et al., 2004 and Buehlmeyer et al., 2007.

Serum IGF-1
Serum IGF-1 concentrations (adjusted for body weight) were significantly lower in the L-EX, M-EX and H-EX groups as compared to the control group after the exercise period (Table 2). In addition, there was a negative, yet not significant, trend in the relationship between mean exercise volume and serum IGF-1 concentrations.

Heart mass
For the heart mass data see Buehlmeyer et al., 2007.

Correlations between exercise volume, body mass, heart mass and serum IGF-1
In all rats, heart mass per kg body mass showed a significant negative correlation with serum IGF-1 levels (r = -0,66; p < 0.001) and body mass (r = -0,55). IGF-1 serum levels showed a significant positive correlation with body mass (r = 0.50). Exercise volume revealed a significant negative correlation with body mass (r = -0.50) and a highly significant positive correlation with heart mass per kg body mass (r = 0.77; p < 0.001).

mRNA of IGF-1, IGF-1R and IGF-BP3 in colon mucosa
Figure 1 displays the normalized data of IGF-1, IGF-1R and IGF-BP3 mRNA levels as the ratio of the target to the housekeeping gene by assigning the control group a factor of 1. Even though we could not find any significant differences or correlations, the following results are worth mentioning: First, a linear negative trend between exercise volume and IGF-BP3 mRNA was observed. Second, IGF-1R mRNA tended to be lower in the L-EX and M-EX groups. Third, in comparison to the control group, we observed 1.7 to 3.0 fold higher expression of IGF-1-mRNA in colon mucosa of L-EX and M-EX groups.