| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Toolbox
School of Health Sciences, Deakin University, Burwood, Victoria 3125, Australia
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
gene expression; quantitative RT-PCR; endogenous controls; biological variability; PCR efficiency
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Commonly used housekeeping genes for studies involving mRNA measurements in human skeletal muscle include 18S (10, 12) and 28S ribosomal RNA (rRNA) (4), ß-actin (29), or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (31). It has become apparent that there is no single housekeeping gene that is suitable for all experimental conditions, and a number of papers have been published looking at the viability of using certain genes as housekeeping genes in different systems using various methodologies (11, 13, 14, 27, 28, 33, 34).
Dietary creatine supplementation (CrS) is a widely used intervention in exercise science and, more recently, in clinical studies examining various muscular disease states (for review, see Ref. 32). Many studies have demonstrated that CrS results in an elevated total creatine content in skeletal muscle (9, 20, 26). Importantly, this increase may induce changes in muscle gene expression perhaps by altering the energy state of the cell or by cell volume-mediated processes. For example, changes in both gene and/or protein expression have been recently reported in models involving CrS and resistance training compared with a placebo group (15, 31). In studies examining gene expression, knowledge of the variations in the expression of housekeeping genes under given experimental conditions is required for valid data interpretation. Although real-time RT-PCR requires the use of a housekeeping gene for interpretation of the data, the relative constitutive nature of this gene must first be established, as must the use of raw data values obtained in real-time RT-PCR, before normalizing them to a housekeeping gene. Consequently, the initial aims of the present study were to determine whether a linear response existed for detection of 28S, ß-actin, ß2-microglobulin (ß2M), cyclophilin (CYC), and GAPDH gene expression in human skeletal muscle using raw data values and to characterize the real-time RT-PCR method in terms of intra- and interassay variability. We then aimed to identify genes, which in response to short-term CrS, were expressed with the least variation and to subsequently establish their use as a housekeeping gene in human skeletal muscle. The samples analyzed in this study were part of a larger study, and the protocol included a small number of high-intensity exercise bouts to investigate the potential ergogenic effects of Cr. Finally, to establish the power with which changes could be identified, we performed power analyses for the detection of the genes 28S, ß-actin, ß2M, CYC, and GAPDH using the real-time RT-PCR technique.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Primer designs.
To perform PCR, specific primers were designed for the genes 28S, ß-actin, ß2M, CYC, and GAPDH, using Primer Express software (Applied Biosystems) on sequences attained from GenBank (see Table 1 for details). Searching of the sequences using Blast (http://www.ncbi.nlm.nih.gov/BLAST/) confirmed their specificity. Primers were purchased from GeneWorks (Adelaide, SA, Australia).
Relationship between starting cDNA amounts and PCR results.
To determine the validity of using raw CT values as a measure of starting cDNA concentrations, real-time PCR was performed to measure the genes 28S, ß-actin, ß2M, CYC, and GAPDH. RNA extracted from a single human muscle sample was used to examine the dynamic range of response for a series of dilutions of cDNA generated from the RT step (see Table 2). Linear regression was used to analyze the response of CT vs. the logarithm of the cDNA concentration. Using the slopes of the lines, we calculated the efficiency (E) of target amplification with the equation E = (10-1/slope) - 1 (Ref. 18).
|
Intra-assay variability of RT-PCR.
To determine the intra-assay variability, the coefficient of variance (CV) was determined for samples run in triplicate for detection of ß-actin, ß2M, CYC, and GAPDH. CV values were calculated using the term 100 x (standard deviation/mean). A total of 94 triplicate readings were used over the 4 genes. The CV was determined for the raw CT values, and also for the term 2-CT, which converts the log scale CT value to the linear form.
Interassay variability of RT-PCR.
To determine interassay variability, the CV was calculated for triplicate readings of samples (n = 34), which were run over two separate occasions (Table 3). For the four genes examined, the two occasions were separated by up to 39 days (21 ± 3 days, mean ± SE, see Table 3). The CV was also calculated for the 2-CT values.
|
Study design for the determination of an appropriate housekeeping gene when skeletal muscle Cr content was elevated by oral Cr supplementation.
Healthy, active male subjects, aged 25.4 ± 4.6 yr (mean ± SD) were used in the study, which was approved by Deakin University Human Ethics Committee and was part of a larger study being conducted by us. The study group consisted of habitual meat eaters (n = 4) and non-meat eaters (n = 4). Subjects (n = 8) underwent a double-blind, crossover study involving the ingestion of Cr (Cr, 0.4 g·kg body wt-1·day-1, plus glucose, 0.4 g·kg body wt-1·day-1; CR group) or a placebo (glucose, 0.8 g·kg body wt-1·day-1; placebo group). Treatment order was randomly assigned and was separated by 5 wk. The washout period of intramuscular Cr has previously been shown to be 4 wk (9, 20). Resting muscle biopsies were taken from the vastus lateralis as previously described (26) before the supplementation period (day 0) and following 1 day (day 1) and 5 days (day 5) of supplementation using the percutaneous needle biopsy method modified to include suction. The muscle samples were stored in liquid nitrogen until analyzed. Immediately after each biopsy, two 30-s all-out sprints, separated by 4 min of passive rest, were performed on a cycle ergometer (Excalibur ergometer; Lode, Groningen, The Netherlands). Consequently, the data correspond to the effects of the ingestion of Cr compared with a placebo when both groups perform this exercise.
Total Cr content.
To ensure that all subjects responded to CrS, intramuscular total Cr (TCr) measurements were made. Muscle samples were freeze-dried for 24 h, weighed and powdered, removing any visible connective tissue, and then extracted using 0.5 M perchloric acid and 1 mM EDTA, and neutralized with 2.1 M KHCO3. These extracts were subsequently analyzed for creatine phosphate (CrP) and Cr levels using an enzymatic fluorometric technique (19). TCr was taken as the sum of CrP and Cr.
Examination of housekeeping genes.
Based upon the results of the validity and reliability studies, the protocols for which are outlined above, four genes were chosen to establish their consistency of expression in human skeletal muscle biopsies at 0, 1, and 5 days postsupplementation with Cr or placebo. To reduce intersubject variation, samples were normalized to the day 0 sample of the placebo trial for each subject. For ß-actin, ß2M, CYC, and GAPDH, a 
CT value was calculated for each sample by subtracting the CT value for a given sample from the CT value of the gene being treated as the housekeeping gene (except when this was the same gene). The relative expressions of these genes compared with the basal values were calculated using the expression 2-CT. The efficiency of the RT-PCR for ß2M, ß-actin, and CYC was determined to be the same across a series of dilutions (data not shown), thus justifying their use as housekeeping genes.
Power analyses.
To reveal the differences measurable in the current experiments, power analyses were performed for each gene using the equation
 |
= desired width of confidence interval (CI) in CT; (1 - 
) = 95% CI; (1 - ß) = power; f(,ß) was adopted for a two-tailed test.
Statistics.
BMDP software was used to perform two-way (treatment x time) ANOVAs. TCr results are expressed as means ± SE; 95% CI values, slopes, and y-intercepts were calculated using Prism GraphPad Software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
0.10 µg) or too much (2.0 µg) RNA was added to the RT reaction (data not shown). When the upper and lower concentrations were omitted as outliers, ß-actin, ß2M, CYC, and GAPDH exhibited inverse correlations (r = -0.9230, -0.8787, -0.8130, and -0.7336, respectively) between CT values and the logarithm of the input RNA content (Fig. 3). The efficiencies of the PCR reactions were 1.4, 1.1, 1.3, and 1.2 for ß-actin, ß2M, CYC, and GAPDH, respectively (see table in Fig. 3). Examination of the 95% CI values for the slopes of each gene indicated that the efficiencies were not different.
|
Biological and technical variability.
The technical variation due to the OD260 readings obtained for a series of six aliquots of the same total RNA extract was 0.050 ± 0.003 (mean ± SD; CV, 6.2%; n = 6). This value incorporates a 5.0% variation due to the repeatability of OD260 readings. Once the validity of using raw CT values for ß-actin was determined, the biological variation seen in different extractions from the same piece of muscle was examined. Variability in the raw CT values for replicates of a single sample due to the RT step was 24.0 ± 0.6 (mean ± SD; CV, 2.6%; n = 7; Table 4). Across the samples, variability in the raw CT values obtained was 24.1 ± 1.5 (mean ± SD; CV, 6.2%; n = 7; Table 4), which accounts for both technical (incorporating inconsistencies in OD260 readings) and biological variability. To determine the biological variability, the variation of the two technical steps (RT 2.6% and PCR 1.3% = 1.9%) involved in obtaining a CT value were averaged, then subtracted from the overall variability. Using this procedure, we found that the biological variability accounted for 4.3% of the overall variability (Table 4).
|
|
1 CT value.
|
CT) for each gene, each data point was normalized to the least varying genes (i.e., ß-actin, ß2M, and CYC), as well as to the placebo day 0 CT value. CYC normalized to ß-actin (Fig. 6A) and ß-actin normalized to CYC (Fig. 6D) were not different between the trials. When normalized to ß-actin, ß2M gene expression was different (P < 0.05) between placebo day 0 and day 5 samples (Fig. 6B). There was significant interaction (P < 0.05) for GAPDH (Fig. 6C) when normalized to ß-actin. When normalized to CYC, there were tendencies for main effects for time for ß2M (P = 0.06; Fig. 6E) and GAPDH (P = 0.08; Fig. 6F). When normalized to ß2M, there were no differences in the expression of ß-actin (Fig. 6G), CYC (Fig. 6H), or GAPDH (Fig. 6I).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Linearity of various input cDNA amounts to PCR.
The first experiment investigated the dynamic range of quantitation of RT-PCR using CT values. Following the log conversion of a series of fold dilutions from the same cDNA, a linear response was demonstrated for all genes tested (28S, ß-actin, ß2M, CYC, and GAPDH; Fig. 2). Examination of the 95% CI showed that for all genes the slope of the logarithmic relationship included -3.3, which indicated that the PCR was able to detect twofold differences in input template, reflected by a single CT value. Hence, these data indicated that the amounts of cDNA added to a given PCR run were reflected by differences in CT values for 28S, ß-actin, ß2M, CYC, and GAPDH transcripts. This finding validates the use of the CT values for detection of absolute changes in the amount of input cDNA for these particular transcripts. These data are in agreement with other researchers who have reported the efficiency of real-time RT-PCR using SYBR Green 1 chemistry and various genes in different tissues (6, 21, 24).
Linearity of RT step to PCR.
Linear relationships between the logarithm of various input amounts of RNA into the RT reactions and the CT values of ß-actin, ß2M, CYC, and GAPDH were obtained (Fig. 3), once the lowest and highest input RNA amounts were omitted. The addition of either not enough (e.g., 0.1 µg RNA) or too much (e.g., 2.0 µg RNA) RNA into an RT reaction, optimal for the RT of 1.0 µg RNA, appeared to be inhibitory to the reaction (data not shown). Although the reactions were not 100% efficient for all the genes, the important feature is that since the 95% CI for the slopes of the graphs for each gene overlapped each other (see table in Fig. 3), the efficiency of the AMV reverse transcriptase enzyme was not different between ß-actin, ß2M, CYC, and GAPDH. These data suggest that for samples undergoing RT at the same time, RT efficiency appears to be the same for input RNA amounts between 0.5 to 1.5 µg and that the subsequent analyses for determination of an appropriate housekeeping gene(s) were valid. The response for 28S was different from the other genes examined, since it exhibited a positive relationship between input RNA amounts of 0.5 to 1.5 µg vs. CT values (data not shown). It is not known why 28S responded in this way, although we speculated that the use of the oligo dT primers, which would selectively reverse transcribe RNA species possessing a poly-A tail, would not be suitable for efficient translation of ribosomal RNA (i.e., 28S). To test that possibility, we undertook RT reactions using random hexamers (Applied Biosystems) using various input amounts (0.25 to 1.5 µg RNA) of pooled RNA. In our hands, this also failed to display an inverse linear relationship (Fig. 7), and as such we eliminated the further testing of 28S in the current experiments. We do not know why this was the case; however, it is possible that the relatively high abundance of the ribosomal RNA may somehow impede the detection of the transcripts in human skeletal muscle using SYBR Green 1 chemistry and the sensitive real-time RT-PCR methods. Consequently, the choice of 28S is not appropriate for real-time PCR under the conditions employed by us (i.e., SYBR Green 1 chemistry, oligo dT or random hexamers as cDNA priming oligo, and two-step RT-PCR). It is possible that 28S might be an appropriate housekeeping gene for human skeletal muscle under other specific conditions (e.g., use of gene-specific primers as cDNA priming oligo, use of TaqMan probe with one-step RT-PCR, or greater dilutions of RNA), although these possibilities were not examined in the current study. To our knowledge, the examination of the efficiency of the RT enzyme AMV on high- or low-abundance target genes has not been reported and might warrant further investigation. Interestingly, it has been demonstrated that the efficiency of the RT enzyme, Moloney murine leukemia virus (MMLV), fell from 20% to less than 6% when a low-abundance target template was used (7). Furthermore, it has recently been reported that the choice of cDNA synthesis conditions influences the sensitivity and accuracy of real-time PCR using SYBR Green 1 chemistry (18). Those authors also found that the efficiency is related to both choice of RT primer as well as transcript being investigated (18). As mentioned in the previous section, the use of fivefold dilutions of cDNA and measurement of CT values demonstrated the linearity of the system for determining the input amount of serial dilutions of 28S rRNA. The data obtained using different amounts of RNA input in the RT reaction along with the findings of Lekanne Deprez and colleagues (18) further highlight the importance, emphasized by numerous researchers previously (11, 14, 17, 22, 25), that investigators need to optimize parameters for individual experimental conditions. To our knowledge, this is the first study that has examined the linearity of the RT step to PCR using RNA obtained from skeletal muscle.
|
Biological and technical variability.
The same piece of muscle cut into seven pieces resulted in a CV of 6.2%, which included both biological (within the same sample, 4.3%) and technical error (1.9%; see Table 4). It needs to be emphasized that this total variability is likely to be the minimum, since further intra-individual variation is likely when additional muscle samples from the same subject are analyzed, due variation in parameters such as fiber type (2). Nonetheless, these data highlight the areas of variability present within a single skeletal muscle sample when real-time RT-PCR is the tool of measuring mRNA expression. Although previous researchers have attributed large variability [in uncoupling protein (UCP) 2 and UCP3 gene expression in muscle samples obtained from the same subject] to biological variability, no quantitative data were provided for variation within a sample (3), and to our knowledge the current paper is the first to provide these data. Additionally, the current study did not measure intersubject variability, although CV values of 50–70% have been reported for the expression of UCP2 and UCP3 gene expression in skeletal muscle amongst individuals (3).
Cr supplementation and gene expression.
Only a single study has examined gene expression following CrS (31). In that study it was reported that skeletal muscle myosin heavy chain (MHC) I, IIa, and IIx gene expression was higher following 12 wk of resistance training combined with CrS compared with resistance training and a placebo (31). This increase in MHC gene expression was attributed to a synergistic response between the resistance training and CrS. In terms of proteins, the ingestion of Cr resulted in differential protein expression of the myogenic transcription factors myogenin and MRF4 following a period of immobilization and 10 wk of resistance training (15). Data from these studies indicate that the examination of gene expression consequent to CrS warrants further investigation. It is important to note that outcomes of studies examining the effect of CrS on gene expression are dependent on the knowledge of an appropriate housekeeping gene.
Examination of housekeeping genes using raw CT values.
Following the characterization of real-time RT-PCR, we examined the response of ß-actin, ß2M, CYC, and GAPDH mRNA at days 0, 1, and 5 following ingestion of either CR or placebo interspersed with high-intensity exercise. The CR trial showed increased intramuscular TCr stores at days 1 and 5 compared with day 0, as well compared with the placebo trial, indicating a response to the supplementation period. Initial examination of gene expression by two-way (treatment x time) ANOVA showed no statistical differences for any of the genes (Fig. 5), although there was a trend (P = 0.08) for a main effect for time for GAPDH. Inspection of the 95% CI indicated that GAPDH was the most variable gene, with the remaining genes behaving in a similar manner. The variability in the gene expression of GAPDH would indicate that this is not a good choice of housekeeping gene and that its use may mask differences that are in fact present. This would suggest that ß2M, ß-actin, or CYC could be appropriate housekeeping genes for the present study.
Examination of housekeeping genes using normalized CT values.
Since initial examinations suggested that only GAPDH showed variable expression, ß2M, ß-actin, and CYC were used for the next step of normalizing the 2-CT data to 2-CT values. Normalization of the ß-actin and CYC data to each other showed a similar stability in their expression. Interestingly, when the ß-actin and CYC were normalized to ß2M, their expression did not change (Figs. 6, G and H). Additionally, the expression of ß2M normalized to CYC was not different (Fig. 6E), although the expression of ß2M was different between days 0 and 5 in the placebo trial when normalized to ß-actin (Fig. 6B). The inconsistent expression of ß2M suggests that it would not be an appropriate choice of housekeeping gene compared with ß-actin and CYC. This discrepancy also highlights the importance of the correct choice of housekeeping gene, particularly when only small differences in the order of twofold changes may be present. It could in fact support the suggestion of employing two housekeeping genes (28), as well as using various baselines for normalizing data, as demonstrated by Hortobagyi and colleagues to validate their data (16). The change in ß2M, a gene involved in immune response (1), in the placebo trial at day 5, but not the CR trial, could suggest that this gene is responding to the exercise undertaken during the trial and that this effect is attenuated by the Cr. Examination of this possibility is outside the scope of this paper; however, further investigation is warranted. There were tendencies for main effects for time for GAPDH when normalized to either ß2M or CYC. A significant interaction with no changes detected by post hoc analysis was seen when GAPDH was normalized to ß-actin. Examination of the GAPDH data indicated it was the most variable gene measured in this study. The variable nature of this gene across various interventions and between individuals is highlighted in a recent review (5). As such, its use as a housekeeping gene has been questioned previously, and it would seem that CrS in combination with high-intensity exercise also introduces wide variation in the expression of GAPDH and it would not be recommended as a housekeeping gene for this intervention. Interestingly, the only other studies that have examined the effect of CrS on skeletal muscle gene expression have adopted GAPDH as their endogenous control (31). Justification of using GAPDH due to its constitutive nature was based on cultured cell data (31). Our results point to the possibility that the results may have been affected by large variations in the choice of housekeeping gene in those studies.
Power analyses.
RT-PCR is typically associated with large variability (5), although far less than other methods of mRNA detection. A previous report examining housekeeping genes suggested that for most genes, less than fourfold differences are probably not biologically significant (30); however, it should be added that we currently are unable to say what magnitude of fold difference in gene expression would be required to represent a biologically significant outcome. Our data indicate that, with real-time RT-PCR, approximately a twofold or greater difference in the gene expression of ß-actin, ß2M, CYC, or GAPDH is measurable with 0.8 power and 95% confidence, for n = 8. For 28S a six- to eightfold difference would be required for detection at this power level and sample size.
Conclusion.
We have demonstrated the quantitative and linear detection of amplified products from human skeletal muscle using SYBR Green 1 chemistry and real-time RT-PCR and the variability of triplicate readings using this method. When taking into account all experiments of the present study, it would appear that ß-actin and CYC would be the most valid choices of housekeeping genes for real-time RT-PCR analysis of human skeletal muscle samples exposed to short-term CrS and high-intensity exercise.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by a grant from the School of Health and Behavioural Sciences, Deakin University, Australia.
Additional resources on the World Wide Web may be found at the Entrez Blast web site (http://www.ncbi.nlm.nih.gov/BLAST/) and at the Entrez Nucleotides database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Nucleotide).
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: R. T. Snow, School of Health Sciences, Deakin Univ., Burwood, Victoria 3125, Australia (E-mail: rsnow{at}deakin.edu.au).
10.1152/physiolgenomics.00060.2002.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. Deldicque, P. Atherton, R. Patel, D. Theisen, H. Nielens, M. J. Rennie, and M. Francaux Effects of resistance exercise with and without creatine supplementation on gene expression and cell signaling in human skeletal muscle J Appl Physiol, February 1, 2008; 104(2): 371 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Aughey, K. T. Murphy, S. A. Clark, A. P. Garnham, R. J. Snow, D. Cameron-Smith, J. A. Hawley, and M. J. McKenna Muscle Na+-K+-ATPase activity and isoform adaptations to intense interval exercise and training in well-trained athletes J Appl Physiol, July 1, 2007; 103(1): 39 - 47. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. G. Churchley, V. G. Coffey, D. J. Pedersen, A. Shield, K. A. Carey, D. Cameron-Smith, and J. A. Hawley Influence of preexercise muscle glycogen content on transcriptional activity of metabolic and myogenic genes in well-trained humans J Appl Physiol, April 1, 2007; 102(4): 1604 - 1611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. R. Alfieri, M. A. Bonelli, A. Cavazzoni, M. Brigotti, C. Fumarola, P. Sestili, P. Mozzoni, G. De Palma, A. Mutti, D. Carnicelli, et al. Creatine as a compatible osmolyte in muscle cells exposed to hypertonic stress J. Physiol., October 15, 2006; 576(2): 391 - 401. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. G. Coffey, A. Shield, B. J. Canny, K. A. Carey, D. Cameron-Smith, and J. A. Hawley Interaction of contractile activity and training history on mRNA abundance in skeletal muscle from trained athletes Am J Physiol Endocrinol Metab, May 1, 2006; 290(5): E849 - E855. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. T. Murphy, W. A. Macdonald, M. J. McKenna, and T. Clausen Ionic mechanisms of excitation-induced regulation of Na+-K+-ATPase mRNA expression in isolated rat EDL muscle Am J Physiol Regulatory Integrative Comp Physiol, May 1, 2006; 290(5): R1397 - R1406. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. T. Murphy, A. C. Petersen, C. Goodman, X. Gong, J. A. Leppik, A. P. Garnham, D. Cameron-Smith, R. J. Snow, and M. J. McKenna Prolonged submaximal exercise induces isoform-specific Na+-K+-ATPase mRNA and protein responses in human skeletal muscle Am J Physiol Regulatory Integrative Comp Physiol, February 1, 2006; 290(2): R414 - R424. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Espy, J. R. Uhl, L. M. Sloan, S. P. Buckwalter, M. F. Jones, E. A. Vetter, J. D. C. Yao, N. L. Wengenack, J. E. Rosenblatt, F. R. Cockerill III, et al. Real-Time PCR in Clinical Microbiology: Applications for Routine Laboratory Testing Clin. Microbiol. Rev., January 1, 2006; 19(1): 165 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Liang and B. Ventura Physiological genomics in PG and beyond: July to September 2005 Physiol Genomics, October 17, 2005; 23(2): 119 - 124. [Full Text] [PDF]
|
|
|
|
|
|
M. Fluc |
), ß-actin (
), ß2-microglobulin (ß2M, 
), CYC (•), and GAPDH (
). Values are means with 95% confidence indicated by dashed lines. Bottom: linear regression equations, 95% confidence intervals (CI) for slopes, and efficiencies (E = 10-1/slope - 1) are indicated for each gene.
) trials (n = 8). Values are means ± SE. *P < 0.05 different from CR day 0. #P < 0.05 different from placebo days 0, 1, and 5.
