Animals, clenbuterol treatment and sampling
Twenty-eight healthy, exercised adult Thoroughbred horses (11 mares and 17 geldings; weight range of 470.7 ± 25.0 kg; age range of 2–6 years) were studied. Horses used in this study were a combination of University and client owned horses. This study was approved by the Institutional Animal Care and Use Committee (University and client owned horses) and the Clinical Trials Committee of the University of California, Davis (client owned horses). For client owned horses, written consent was obtained for participation in the study. Prior to and throughout the course of the study, horses were exercised five days a week. The general exercise protocol was meant to simulate the strenuous exercise of race training. The exercise regimen for these horses consists of three days per week on an Equineciser (Centaur Horse Walkers Inc, Mira Loma, CA, USA) (5 min at walk; 30 min trot; 5 min walk) and two days per week on a high speed treadmill (Mustang 2200, Graber AG, Switzerland; Day 1: 5 min @1.6 m/s; 5 min @ 4 m/s; 5 min @ 7 m/s; 5 min @ 1.6 m/s all at 6 % incline. Day 2: 3 min @ 1.6 m/s; 4 min @ 4.0 m/s; 2 min @ 7.0 m/s; 2 min @ 11.0 m/s and 5 min @1.6 m/s all at 3 % incline). All horses were subject to regular fitness testing, including weekly heart rate measurements and calculation of V200 (running velocity that elicited a heart rate 200 bpm) and monthly measurements of end run plasma lactate concentrations, as a means by which to ensure that the fitness level of the horses used in this study were as comparable as possible to the average racehorse.
Prior to commencement of the study, all horses were determined healthy and free of cardiovascular diseases by physical examination, complete blood count and a serum biochemistry panel that included aspartate aminotransferase, creatinine phosphokinase, alkaline phosphatase, total bilirubin, sorbital dehydrogenase, blood urea nitrogen and creatinine. Horses did not receive any other medications for at least two weeks prior to conducting this study.
Twenty-two of the horses received 0.8 μg/kg clenbuterol (Ventipulmin®, Boehringer Ingelheim Vetmedica Inc, St Joseph, MO) PO BID for 30 days and an additional 6 horses received the escalating dosing protocol as described on the manufacturer label for non-responders (0.8 μg/kg, BID × 3 days; 1.6 μg/kg, BID × 3 days; 2.4 μg/kg, BID × 3 days; 3.2 μg/kg, BID for 21 days).
Blood samples were collected at time 0 and at 15, 30, and 45 min, and 1, 1.5, 2, 2.5, 3, 3.5, 4, 5, 6, 8 and 12 h post administration of the first dose and last dose and 18, 24, 36, 48, 60, 72, 84, 96, 108 and 120 h post administration of the last dose. Additionally blood samples were collected every 12 h (immediately prior to administration of each dose (trough concentrations)) throughout the 30-day dosing period. Blood samples were collected by direct venipuncture into EDTA blood tubes (Kendall/Tyco Healthcare, Mansfield MA) and were centrifuged at 3000 × g for 10 min. Plasma was immediately transferred into storage cryovials (Phenix Research Products, Chandler, NC) and stored at −20 ° C until analysis. Drug concentrations were measured by Liquid Chromatography tandem Mass Spectrometry as described previously [13].
Muscle biopsy samples were collected one day prior to administration of the first dose of clenbuterol. Additional samples were collected at 48 h and on days 7, 14, and 28 post administration of the first dose. A final sample was collected one-week post administration of the final dose (35 days post administration of the first dose). Muscle biopsies were collected at the same time on each sampling occasion which was approximately 3–4 h post feeding. Horses were not exercised on the day that muscle biopsy samples were collected. For sample collection, horses were sedated with a combination of xylazine (Lloyd Inc, Shenandoah, IA) and butorphanol (Zoetis, Florham Park, NJ). Lidocaine (Aspen Veterinary Resources Ltd, Liberty MO) was administered subcutaneously over the gluteus muscle. A small area over the superficial gluteal muscle was aseptically prepped and a Bergstrom biopsy needle (Dixons Surgical Instruments, Wickford, Essex) used to collect a small sample (approximately 1 g of tissue). The tissue was transferred to a cryovial containing RNAlater (Qiagen Inc, Valencia, CA) and stored at −20 ° C until processed.
RNA extraction and quality assessment
Muscle biopsy samples were placed in 600 μl of lysis buffer, contaning 2-β- mercaptoethanol (Qiagen Inc, Valencia, CA). The tissue sample was then transferred to MagNa Green Beads tubes (Roche Diagnostics, Mannhein, Germany). Homogenization was performed using a MagNaLyse (Roche Diagnostics, Mannhein, Germany) at 6000 rpm for 30 s followed by a 1.0 min cool down period. Homogenization intervals were continued until pieces of muscle were no longer visible (≤3 times). Total RNA was purified using an miRNeasy Mini kit (RNeasy Mini, Qiagen Inc, Valencia, CA) and following the manufacturer’s instructions. Total RNA integrity was assessed using the Experion Automated Electrophoresis System (Bio-rad, Hercules, CA). Only RNA samples with RIN ≥ 8 and 260/280 ratios between 1.7 and 2.1 were used [14, 15].
Microarray analysis
Equine specific microarrays (EquGene-1.0-st; Affymetrix, Santa Clara, CA), containing 504,603 probes representing 30,559 well-characterized genes were used.
To reduce biological noise as a result of genetic variability, each horse (5 horses for low dose administration and 4 horses for the escalating dosing regimen) was analyzed separately and served as their own control for comparison of baseline samples to day 14 (low dose administration) or day 28 (escalating dose regimen). Five biological replicates per time point were tested. Purified total RNA (5 μg) was used for cDNA synthesis in accordance with the Ambion® WT Expression assay kit (Affymetrix, Santa Clara, CA) manufacturer’s protocol. In vitro transcription was used to incorporate biotin labels using the GeneChip® WT Terminal Labeling system (Affymetrix, Santa Clara, CA) and samples hybridized to the Equine microarray. Arrays were washed and stained on a Fluidics Station 450 (Affymetrix, Santa Clara, CA) and scanned on a GeneChip Scanner 3000 (Affymetrix, Santa Clara, CA) in accordance with manufacturer’s protocols.
The microarrays were evaluated for differential gene expression using Transcriptome Analysis Console (TAC) and for hybridization quality control using Expression Console Software (Affymetrix, Santa Clara, CA). In brief, a total of five Cell Intensity Files were generated per time point, uploaded and normalized under the following conditions: PM (perfect match)-only as a PM intensity adjustment and the Robust Multichip Analysis (RMA) quantification method. For evaluation of the assays performance the number of differentially expressed (DE) genes detected between baseline and day 14 (low dose administration) or day 28 (escalating dose regimen) were assessed. Based on the TAC software user’s manual, genes with mean transformed ratios significantly less than −2 and larger than +2 were considered significantly regulated. Significant DE genes were selected by filtering the genes using an ANOVA (p value < 0.05). A Pearson’s correlation coefficient was used to calculate linear dependence between time point and baseline samples to evaluate the correlation coefficient, where 1 was a positive correlation and 0 was no correlation. For each probeset, expression at day 14 (low dose administration) or day 28 (escalating dose regimen) was compared to expression at baseline in the same horse using a paired t-test. Fold changes and their confidence intervals were calculated by exponentiating (base 2) the mean within-horse difference in expression for each gene and the associated t confidence intervals. P-values were adjusted for multiple testing using the False Discovery Rate (FDR) method of Benjamini and Hochberg [16]. Analyses were conducted using the statistical software environment R, version 3.0.2 (R Core Team, 2013). Gene Ontology biological process of differently expressed genes of interest was performed using the most recent Database for Annotation, Visualization and Integrated Discovery (DAVID) 6.8 Beta (https://david-d.ncifcrf.gov/) with the Entrez Gene ID as the primary identifier and Equus Caballus as the species.
Quantification of mRNAs using Taqman low density arrays (TLDA)
Genes that showed significant changes in expression (greatest fold change and statistical significance) following microarray analysis as well as those that have previously been reported to be affected by clenbuterol administration, were chosen as candidate genes for further study. Microarray data were validated by measuring the levels of specific mRNA in muscle biopsy samples at each time point post clenbuterol administration vs. baseline via Taqman Low Density Arrays (TLDA; Affymetrix, Santa Clara, CA) which are pre-loaded 384 well RT-PCR microfluidic cards.
RNA was diluted to a concentration of 2 ng/μl for cDNA conversion using the QuantiTect Reverse Transcription Kit (Qiagen Inc., Valencia, CA). cDNA was then combined with Taqman Universal Mastermix (Affymetrix, Santa Clara, CA) at a final concentration of 2X and then 100ul of each sample loaded onto the TLDA card.
Primers for candidate genes were designed and manufactured by Life Technologies (Affymetrix, Santa Clara, CA). Reference or “housekeeping” genes were used to evaluate the effect of RNA integrity on the array and qRT-PCR performance. Based on stability across all samples studied (TaqMan Protocol [17]) equine beta 2 microglobulin (β2M) was used as the endogenous control gene to normalize the qRT-PCR data while human 18 s (hs18s) was used as an internal manufacturing control. Each sample was run in quadruplicate with 10 candidate genes per card. The TLDA cards were then run on a QuantStudio™ 12 K Flex Software v1.2.2 (Affymetrix, Santa Clara, CA). Analysis was performed using ExpressionSuite Software 1.0.3 using Singleplex analysis (Affymetrix, Santa Clara, CA).
Statistical analyses using commercially available software (Stata/IC 13.1, StataCorp LP, College Station, TX) were performed to assess significant differences in expression (fold change) between baseline and each time point as well as between the different time points post clenbuterol treatment. Data was analyzed using a mixed effects analysis of variance, with the horse treated as a random effect, and with time as fixed effects. Post-hoc comparisons were performed with a Bonferonni multiple-comparison adjustment to preserve a nominal significance level of 0.05.