Introduction
The mouse models of myocardial infarction (MI) and transverse aortic arch constriction (TAC) are widely used in the study of heart failure (HF), for investigating pathophysiological mechanisms and for improving diagnostic and therapeutic approaches (1). Quantitative polymerase chain reaction (qPCR) has become a crucial technique for studying HF, as it possesses the advantages of improved accuracy, sensitivity and specificity, a larger dynamic range and greater speed as compared with other methods. However, no single reference gene (RG) has been identified that is able to maintain stable expression levels under all conditions. It has been previously suggested that the expression levels of previously well characterized RGs, including GAPDH, ACTB and B2M, can vary markedly across cells, tissues and experimental conditions, emphasizing the requirement to adopt alternative RGs and appropriate strategies for their selection (2–4).
Sundaresan et al (5) indicated that the levels of tubulin and actin are variable during heart failure in TAC mice. Previous research has suggested that the most stable RGs vary between the chambers of the heart in the minipig (6). Koyanagi et al (7) demonstrated that CycA expression was the least variable in a hypertension-induced cardiac dysfunction rat model. Another study demonstrated that ACTB was the most stable RG in heart tissue from murine models of cardiometabolic disease (8). This strategy of RG selection has also been applied to the human myocardium (9).
To determine the appropriate RGs under the condition of HF in mice, the current study utilized geNorm software (Biogazelle, Zwijnaarde, Belgium). geNorm is a bioinformatics tool designed to rank candidate RGs using a normalization factor calculated from the expression level of the candidate RGs in an array of given samples (10).
The aim of the present study was to select a set of RGs to be used for normalizing mRNA expression data obtained by qPCR in two HF mouse models. Eight candidate RGs, including GAPDH, ACTB, B2M, CycA, TBP, PBGD, HTRP 1 and 18S (11–15), were investigated. The normalization strategy used was tested by analyzing the mRNA expression of B-type natriuretic peptide (BNP).
Materials and methods
Ethical approval
All animal experiments conformed to the Guide for the Care and Use of Laboratory Animals, published by the US National Institutes of Health (NIH publication no. 85-23, 1996; Bethesda, MD, USA) and were approved by the Ethics Review Board for Animal Studies of Nanjing Drum Tower Hospital (DTH ERBA 58.01/026B/2011).
HF mouse models
MI model
Adult male C57BL/6 mice (8 weeks old, 18–23 g, n=9) were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China). Mice were anesthetized by intraperitoneal administration of a mixture of ketamine hydrochloride (50 mg/kg; Jiangsu Hengrui Medicine Co., Ltd., Lianyungang, China) and diazepam (2.5 mg/kg; Tianjin Jinyao Amino Acid Co., Ltd., Tianjin, China). To detect response to tactile stimuli, the tips of the whiskers were stroked and a forepaw was pinched slightly using atraumatic forceps. The absence of a response to painful stimuli to the tail was considered indicative of adequate anesthesia. Any reaction or movement was judged as a positive response. Mice were endotracheally intubated and mechanically ventilated using a 16G needle (Finlumen, Inc., Beijing, China), which was inserted ~2 cm behind the tongue of the mice (tidal volume, 6–8 ml/kg; frequency, 120 bpm; respiration ratio, 1:1; Jiangxi Teli Anaesthesia & Respiration Equipment Co., Ltd., Nanchang, China) with supplemental oxygen. The heart was exposed through a thoracotomy on the left at the third or fourth intercostal space, and the left anterior descending coronary artery (LAD) was ligated with a 7-0 silk suture, 1 mm below the tip of the left atrial appendage. MI was confirmed by myocardial blanching and electrocardiographic observations. The mice of the sham group (n=9) underwent a thoracotomy and pericardiotomy without LAD ligation. Prior to sacrifice, the cardiac function of the mice was evaluated by transthoracic ultrasonic cardiogram (UCG) to confirm the successful creation of the HF models. The mice were anesthetized again by ketamine hydrochloride (50 mg/kg) and diazepam (2.5 mg/kg), and sacrificed by intravenous administration of 2 mol/l KCl. The hearts were collected for further analyses.
TAC model
C57BL/6 male mice (8 weeks old, 18–23 g, n=10) were anesthetized as described above. The chest cavity was entered through the second intercostal space at the left upper sternal border via a small incision. Using a light source with two flexible fiber optic arms, the thymus was gently moved from the field of view to expose the aortic arch. The transverse aorta was isolated between the carotid arteries, and then it was constricted by a 7-0 silk suture ligature that was tied firmly against a 27-gauge needle (Becton Dickinson Co., Franklin Lakes, NJ, USA). The latter was promptly removed to yield a constriction of 0.4 mm in diameter. The chest was closed with a 5-0 silk suture (Shanghai Pudong Jinhuan medical products Co., Ltd., Shanghai, China), and the mice were allowed to recover from the anesthesia with an external heat supply to maintain body temperature at 37°C. The surgical preparation was performed without the use of a microscope and lasted 30 min.
UCG
The M-mode measurements of the left ventricular (LV) dimensions were averaged from >3 cycles. LV end-systolic diameters (LVESD) and end-diastolic diameters (LVEDD) were measured. The percentage LV fractional shortening (FS%) and percentage ejection fraction (EF%) were calculated as follows: FS% = (LVEDD-LVESD)/ LVEDD × 100; and EF% = EF% = 100 × [((7.0/(2.4+LVEDD))* LVEDD3 - ((7.0/(2.4+LVESD))*LVESD3)/((7.0/(2.4+LVEDD))* LVEDD3]
Histological examination
Hearts from mice at 4 weeks following MI and 8 weeks following TAC were fixed with formalin, embedded with paraffin (Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd., Beijing, China) and cut into 4 mm slices for histological analysis. The sizes of the hearts were assessed by histological analysis with hematoxylin and eosin staining.
RNA isolation and reverse transcription (RT)
Total RNA was extracted using TRIzol® (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. The ratio of the optical densities at 260 and 280 nm were used to assess the quality and quantity of the total RNA. The left ventricle of the heart was chosen for qPCR analysis. cDNA was prepared from 1 μg total RNA using the PrimeScript RT reagent kit (Takara Biotechnology Co., Ltd., Dalian, China). The Oligo dT Primer from the kit was used for the RT reaction, and the cDNA was stored at −80°C until further use.
qPCR
The sequences of primers used for PCR are listed in Table I. The reactions contained 2× SYBR® Premix Ex Taq, 50× Rox Reference Dye II, forward and reverse primers (200 nM) and 2 μl cDNA template, in a 20 μl reaction volume. The amplification was performed with an initial denaturation step at 95°C for 30 sec, followed by 40 cycles of denaturation at 95°C for 5 sec, annealing and extension at 60°C for 34 sec. Following amplification, the specificity was confirmed by melting-curve analysis of the PCR products. The qPCR was performed using the Applied Biosystems 7500 Real-Time PCR device. All of the samples were run in triplicate.
Analysis of RG expression stability
geNorm is an appropriate selection for the analysis of the stability of RGs. It makes use of the comparative cycle threshold (Ct)-method to validate the stability of RG RNA expression levels. The lowest Ct value was set to a comparative expression level of 1, and the expression levels of the same RGs were relatively quantified in accordance with the 2−ΔΔCt method. The geNorm tool was also used to calculate the stability measures (M) of the RGs. The genes were ranked by their M value, and at each step of the analysis the least stable gene (highest M value) was excluded and M was recalculated. The most stable pair of genes was identified through the reiteration of this procedure in a step-wise manner. The optimal number of genes for normalization was determined as previously described (4), starting from the most stable pair and sequentially adding the other genes (from the most to least stable) until the pair-wise variation reached a set threshold (usually, M-value <1.5 and pair-wise variation <0.15).
Statistical analysis
SPSS software, version 13.0 (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. In all cases, the data were compared by one-way analysis of variance. P<0.05 was considered to indicate a statistically significant difference.
Western blot analysis
Proteins were extracted from the left ventricles of the hearts and assessed by western blotting. The primary antibodies used were rabbit anti-GAPDH poltclonal antibody targeting human, mouse and rat (AP 0066; Bioworld Technology, Inc., St. Louis Park, MN, USA) and rabbit anti-ANF targeting human, mouse and dog (ab91250; Abcam, Cambridge, MA, USA). Phosphate-buffered saline containing 8.0 g NaCl, 0.2 g KCl, 1.44 g Na2HPO4, 0.24 g KH2PO4 and ddH2O to 1000 ml (pH 7.4) was used as a dilution for the antibodies Remote myocardium was homogenized with a Dounce tissue grinder in lysis buffer consisting of: 10 mM Tris/HCl (pH 7.2), 150 mM NaCl, 0.1% SDS, 1% (v/v) Triton X-100, 1% sodium deoxycholate, 5 mM EDTA, 1 mM Na3VO4, 50 mM NaF (all Sigma), 0.2 mM phenymethylsulfonyl fluoride and protease inhibitor cocktail (complete EDTA-free; Roche Diagnostics, Basel, Switzerland). Protein content was determined by a bicinchoninic acid protein assay kit (Pierce Biotechnology, Inc., Rockford, IL, USA). Proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to Immobilon™ polyvinylidene fluoride membranes (EMD Millipore, Billerica, MA, USA).
Results
HF in mouse models following TAC and MI
Four and eight weeks subsequent to MI and TAC, respectively, there was a marked cardiac enlargement (Fig. 1A and B). To confirm the extent of damage to the heart, the expression levels of HF markers were determined (16). The western blot analysis of ANF indicated a reduction of heart function in the HF mouse models (Fig. 1C and D). UCG analysis also revealed left ventricle enlargement, dilatation and systolic dysfunction 4 weeks and 8 weeks subsequent to MI and TAC, respectively (Fig. 1E and F). These findings indicated the successful production of mouse models of HF.
Primer specificity and efficiency
The hearts of mice in the two model groups were resected and analyzed. Following amplification of the primers, the expression levels of the RGs were observed by 1.5% agarose gel electrophoresis (Power Pac 200; Bio-Rad, Hercules, CA, USA). As presented in Fig. 2A, the bands of the RGs, including GAPDH, ACTB, B2M, CycA, TBP, PBGD, HTRP 1 and 18S, were clear. The melting-curve analysis of the eight RGs during the qPCR revealed a single product with a melting temperature in accordance with the expected value (data not shown). The distributions of Ct data among the eight RGs and BNP covered a wide range, from 16 to 34 cycles in the MI, sham and TAC groups (Fig. 2B).
Analysis of RGs using geNorm software
geNorm was used for the analysis of specimens from the sham, MI and TAC groups. A lower M-value indicated a more stable gene expression. All RGs, except HPRT 1 presented an M-value below the theoretical threshold of 1.5 in the sham group. In the sham group, the most stably expressed genes were PBGD and GAPDH. When combined with MI or TAC specimens, the most stable RGs were CycA and GAPDH or B2M and ACTB, respectively. In the MI + TAC + sham group, the most stable RGs were PBGD and ACTB (Table II). In all the tissue subsets, the combination of the two best RGs was identified to be sufficient for gene normalization with pair-wise variation values of ≤0.15 (Fig. 3).
BNP mRNA expression
BNP expression was analyzed to assess whether the normalization strategy had been effectively applied (17). The mRNA levels of BNP were normalized to the subset of selected RGs. As presented in Fig. 4, the relative BNP mRNA levels differed when normalized to each different selected RGs. With the exception of B2M, BNP expression was relatively higher in the MI and TAC models as compared with the sham group when normalized to the expression levels of the remaining seven RGs. qPCR analysis was consistent with geNorm software analysis. The data suggest that neither B2M or HPRT 1 should be used as RGs in the MI model, nor HPRT 1 in the TAC model of HF.
Discussion
The current study aimed to illustrate the variations in the stability of RGs in mouse models of MI and TAC of HF. To the best of our knowledge, the present study is the first to describe the selection of RGs in C57BL/6 mice with two different models of HF. A combination of PBGD and ACTB was determined to be best for use as RGs in mouse MI + TAC + sham models of HF. In addition, a combination of GAPDH and CycA were identified as the most stable RGs for use in the MI model, while B2M and ACTB were most appropriate in the TAC model. The use of HPRT 1 as an RG in C57BL/6 mice was not indicated to be appropriate for use in the mouse models generated in the present study.
The C57BL/6 mouse is widely used to identify potentially important and novel targets for pharmacological or molecular therapy for HF. The mouse HF models not only provide data during the development of the chronic changes in myocardial structure and function, but also quantify the progression of HF and left ventriclar dysfunction. The MI (18,19) and TAC (19,21) mouse models are the most widely and successfully used HF models.
The expression of eight commonly used RGs, including GAPDH (5), ACTB, B2M, CycA (6,22), TBP (6), PBGD (23), HTRP 1 (6) and 18S (1), were analyzed to determine which is the best set for use in the mouse models. The experimental protocol consisted of an LV, which was able to produce a marked remodeling and dilation of the heart as a HF model (24). Substantial differences in gene expression stability, expressed as M-values, were observed between the mouse models. It may be interpreted that during HF, there are different specific physiological and functional changes involved in the different models.
CycA has been used as an RG in the atrium and right ventricle of the minipig HF model (induced by rapid LV pacing for 180 beats/min) (6). However, other studies have indicated that CycA, which is secreted by inflammatory and dying/necrotic cells, was significantly enhanced in patients with inflammatory cardiomyopathy (22) and was able to promote atherosclerosis in mice (25). CycA was determined as one of the most stable RGs in the MI mouse model of HF in the present study, but it was not conclusive whether the mice developed atherosclerosis.
A previous study indicated that GAPDH could not be used as an RG as its expression was highly variable in MI mouse tissues (26). In particular, GAPDH was also demonstrated to be the optimal RG in a rat MI model of HF (2). The results of the present study indicated that GAPDH, which is commonly used for gene expression normalization in MI studies, presented the least expression variability in tissues of the MI model mice, and in combination with CycA created the best set of RGs for the MI mouse model.
ACTB, a cytoskeletal protein, has been indicated to be the most stable gene in heart tissues from murine models of cardiometabolic disease (produced by feeding with high-fat diet) (8). Yperman et al (27) demonstrated that ACTB was not stable in certain types of interstitial cell, thus it cannot be regarded as a reference for gene expression of ovine-derived heart valve interstitial cells in culture. In the current study, ACTB, combined with B2M or PBGD, was observed to be the best RG for the TAC or TAC + MI model.
In order to confirm the results from geNorm, the relative expression of BNP, a 32-amino acid polypeptide secreted by the ventricles of the heart in response to excessive stretching of cardiomyocytes (28), was evaluated. The results displayed a marked difference in the levels of BNP mRNA normalized to the set of the selected RGs. Therefore, the upregulation of BNP mRNA expression was in agreement with the data of the study by Bhalla et al (28), and the comparison revealed that the levels may vary significantly, depending on the RGs used for normalization.
In summary, the present study identified and validated a set of stably expressed RGs for use in mouse MI and TAC models of HF. The results indicated that GAPDH combined with CycA or B2M with ACTB were the best set of RGs in MI or TAC mouse models, respectively. Where MI and TAC mouse models are required simultaneously, it is preferred to use ACTB and PBGD as the RGs. The current study provides important information for RG selection, and will facilitate further investigations into the transcriptome of mice with aberrant heart function.