1 3

Mol Genet GenomicsDOI 10.1007/s00438-014-0877-6

OrIGInal PaPer

Identification and validation of suitable reference genes for RT‑qPCR analysis in mouse testis development

Zu‑Kang Gong · Shuang‑Jie Wang · Yong‑Qi Huang · Rui‑Qiang Zhao · Qi‑Fang Zhu · Wen‑Zhen Lin

received: 16 October 2013 / accepted: 5 June 2014 © Springer-Verlag Berlin Heidelberg 2014

genes used for rT-qPCr normalization and provide crucial information for transcriptional analyses in future studies of gene expression in the developing mouse testis.

Keywords Mouse testis development · reference gene · rT-qPCr · genorm · normFinder · Bestkeeper · Comparative delta Ct method · refFinder

Introduction

The testis is a component of the reproductive and endocrine systems. The primary functions of the testis are to produce sperm (spermatogenesis) and regulate testosterone secretion (steroidogenesis). Spermatogenesis is a complex process involving cell division, differentiation and interaction among cells in the microenvironment of the seminiferous tubules. The developing testis provides an environment that ensures spermatogenesis and fertility (Miles et al. 2013) and any untoward effects on this environment can induce aberrant germ cell development and formation of germ cell tumor precursors (Miles et al. 2013). although a variety of different genes are involved in the processes that regulate spermato-genesis (Phillips et al. 2010), the signaling events involved in testis formation and male germ cell development are still poorly understood (Miles et al. 2013; Chalmel et al. 2013; Migrenne et al. 2012). In order to better understand the phys-iology and pathologies of the testis (including infertility and cancer), many scientists have struggled to pinpoint the events that regulate specific steps of testis development or function or that are involved in pathological events. a first step in identifying these events is to analyze gene expression levels.

although there are numerous techniques that estimate transcript expression levels, reverse transcription-quanti-tative real-time polymerase chain reaction (rT-qPCr) has

Abstract rT-qPCr is a commonly used method for evaluating gene expression; however, its accuracy and reli-ability are dependent upon the choice of appropriate ref-erence gene(s), and there is limited information available on suitable reference gene(s) that can be used in mouse testis at different stages. In this study, using the rT-qPCr method, we investigated the expression variations of six reference genes representing different functional classes (Actb, Gapdh, Ppia, Tbp, Rps29, Hprt1) in mice testis dur-ing embryonic and postnatal development. The expression stabilities of putative reference genes were evaluated using five algorithms: genorm, normFinder, Bestkeeper, the comparative delta Ct method and integrated tool refFinder. analysis of the results showed that Ppia, Gapdh and Actb were identified as the most stable genes and the geomet-ric mean of Ppia, Gapdh and Actb constitutes an appropri-ate normalization factor for gene expression studies. The mrna expression of AT1 as a test gene of interest varied depending upon which of the reference gene(s) was used as an internal control(s). This study suggested that Ppia, Gapdh and Actb are suitable reference genes among the six

Z.-K. Gong and S.-J. Wang contributed equally to this work and should be considered as co-first authors.

Communicated by S. Hohmann.

Z.-K. Gong Guangxi nanning Blood Center, nanning 530007, Guangxi, China

S.-J. Wang · Y.-Q. Huang · r.-Q. Zhao · Q.-F. Zhu · W.-Z. lin (*) Department of Biochemistry and Molecular Biology, School of Preclinical Sciences, Guangxi Medical University, nanning 530021, Guangxi, Chinae-mail: linwenzhen@msn.com

Mol Genet Genomics

1 3

recently been shown to be more efficient, sensitive and reliable than other methods (Stephens et al. 2011; Gebeh et al. 2012). However, there are a variety of problems that affect the use of rT-qPCr, such as rna extraction, pres-ence of contaminating Dna, efficiency of cDna synthe-sis (Chechi et al. 2012). Since PCr exponentially ampli-fies a small amount of cDna to a large quantity of Dna product, very small variations can lead to large errors in the final product and invalidate the gene expression analysis. To deal with these difficulties, researchers usually perform a relative quantification to analyze the expression level of a target gene by normalizing the expression of the target gene to the expression of stably expressed gene(s), com-monly termed reference gene(s) (Zhang et al. 2012). ref-erence genes are typically constitutive endogenous genes, often highly expressed in the cell, that are required for the maintenance of basic cellular function, such as cellular metabolism, cell structure and protein synthesis. Therefore, reference gene expression usually shows minimal variabil-ity and high stability in a multitude of physiological states; otherwise, it would reflect variations in sample preparation and experimental variability (Boda et al. 2009). The most widely used reference genes are β-actin (Actb), glyceral-dehyde-3-phosphate dehydrogenase (Gapdh), 18S and 28S rrnas. However, several studies have shown that no per-fect reference gene exists because the reference gene can be influenced by different tissues, developmental stages, drugs, diseases, etc. (Wong and Medrano 2005; nolan et al. 2006; Hong et al. 2008; Wan et al. 2010; Feng et al. 2013). Dheda et al. (2005) (Bas et al. 2004; Tricarico et al. 2002) demonstrated that using an unstable reference gene can lead to erroneous results. Consequently, it is becoming an essential component of rT-qPCr to systematically evaluate common and novel reference genes derived from genome-wide analyses so as to improve the reliability of published results (Czechowski et al. 2005; Faccioli et al. 2007; Paol-acci et al. 2009). recently, in order to select the most reli-able reference gene(s) for a particular experimental assay, a number of statistical algorithms such as genorm, nor-mFinder, Bestkeeper and comparative delta Ct have been developed to evaluate the expression stabilities of candidate reference genes

With this knowledge, investigators have published some suitable tissue-specific normalizing genes (Cicinnati et al. 2008; Silver et al. 2006; Pombo-Suarez et al. 2008). How-ever, there are currently few data regarding the expres-sion stability of commonly used or novel reference genes in the developing testis of mice. In the current study, we compared the expression of four frequently used refer-ence genes (Actb, Gapdh, Ppia, Hprt1) and two relatively new reference genes (Tbp, Rps29). The most stable refer-ence genes were identified by applying several statistical approaches, such as genorm, normFinder, Bestkeeper and

the comparative delta Ct method. Furthermore, the overall ranking of the six candidate reference genes was deter-mined by an integration tool, refFinder. Finally, validation of reference genes was analyzed by calculating the expres-sion level of angiotensin type 1 receptor (AT1) using differ-ent reference genes and the most stable reference gene(s) as an internal control(s).

Based on our study, then, we expect to identify a panel of fully characterized reference genes that can be used rou-tinely to quantify gene expression in mouse testis develop-ment and to begin to understand how the reference genes affect the expression level of AT1 in the developing testis.

Materials and methods

Tissue collection

In order to compare the stability of the candidate reference genes, a set of experiments was performed on testes from mice at different developmental ages from embryonic day 18 (e18) to 4 months (M) postnatally (three individuals per time point). The experimental protocols were devel-oped in accordance with the regulation for administration of laboratory animals, China, and approved by the ethics Committee in animal experimentation of Guangxi Medi-cal University. Chinese Kun Ming (KM) mice were main-tained in cages in a temperature-controlled room having a 12-h light/12-h dark cycle (8:00 aM lights on/8:00 PM lights off). e1 was identified as the first day when a vaginal plug was observed. Similarly, P1 was identified as the first day when mouse pups were present. Testes from mice were removed at different ages from e18 to 4M (including e18, P1, P7, P14, P18, P22, P26, P30, 2, 3 and 4). The excised tissues were frozen in liquid nitrogen and then stored at −80 °C until rna extraction.

rna extraction and cDna synthesis

Total rna was isolated from 100 mg of testis using TrI-ZOl reagent (Invitrogen, life Technologies, Carlsbad, Ca, USa) according to the manufacturer’s protocol. The yield and purity of isolated rna was determined using a nanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, De, USa). Samples selected for rT-qPCr had a260/a280 ratios between 1.8 and 2.05. The integrity of the rna samples was assessed by electrophoresis on a 1 % agarose gel.

First-strand cDna was synthesized from 1 μg of total rna using M-MlV reverse transcriptase kits (Invitro-gen, life Technologies, Carlsbad, Ca, USa) and ran-dom oligoprimers (TaKara Biotechnology Co., ltd., Dalian, China) in a total volume of 20 μl according to

Mol Genet Genomics

1 3

the manufacturer’s instructions. The cDna was stored at −20 °C and diluted 1:10 with nuclease-free water as a tem-plate in rT-qPCrs.

Primer design

The sequences of six candidate reference genes (Actb, Gapdh, Ppia, Tbp, Rps29, Hprt1) representing different functional classes were obtained from the entrez nucleo-tide database (http://www.nubi.nlm.nih.gov). Primers were designed using Primer express software 3.0 (applied Bio-systems, Foster, Ca, USa) combined with Oligo 6 soft-ware (Molecular Biology Insights, Inc. Cascade, CO, USa). each primer pair spanned an exon to avoid amplifi-cation of genomic Dna. all primers were confirmed using nCBI Blast against all available Dna sequences to ensure their specificity (www.ncbi.nlm.nih.gov/tools/primer-blast). The primer sequences and expected amplicon sizes are given in Table 1.

rT-qPCr analysis

rT-qPCr was performed on an eppendorf Mastercycler ep realplex (eppendorf, Hamburg, Germany) using Fast-Start Universal SYBr Green Master (roche applied Sci-ence, Mannheim, Germany). reactions were carried out in a total reaction volume of 20 μl containing 2 μl of 1:10 diluted template cDna, 10 μl of 2 × SYBr green PCr Master Mix (roche applied Science, Mannheim, Ger-many) and 100 nM of each primer. The following ampli-fication program was used in all PCrs: 95 °C for 10 min, and 40 cycles of 15 s at 95 °C and 1 min at 60 °C. The specificity of each amplified reaction was verified by a dis-sociation curve (melting curve) analysis after 40 cycles, which was carried out by heating the amplicon from 60 to 95 °C. Moreover, the specificity of PCr product was fur-ther confirmed by 2 % agarose gel electrophoresis. each sample was analyzed in triplicate wells, and “no-template controls” (without cDna in the PCr) were included. Data were collected and quantitatively analyzed using realplex software.

For PCr efficiency, a tenfold dilution series was made from a cDna pool of our sample groups (including all samples), ranging from a tenfold dilution to a 100,000-fold dilution. Duplicate rT-qPCrs were carried out for each gene at each dilution. The standard curve was generated by plotting the mean quantification cycle (Cq) values against the logarithm-transformed concentration of the dilution template using realplex software. The PCr efficiency and the squared correlation coefficients (r2) were generated using the slopes of the standard curves. The reaction effi-ciency for each gene assay was calculated according to the following formula: E(%) =

(

10(−1/slope)− 1

)

× 100. Tabl

e 1

Pri

mer

info

rmat

ion

of s

elec

ted

cand

idat

e re

fere

nce

gene

s an

d si

ze o

f r

T-qP

Cr

pro

duct

s

Gen

e sy

mbo

l (ac

cess

ion

num

ber)

Gen

e na

me

Func

tion

Prim

er s

eque

nce

(5′-3

′)a

mpl

icon

leng

th (

bp)

Act

b (n

M_0

0739

3.3)

Bet

a-ac

tin (β

-act

in)

Cyt

oske

leta

l str

uctu

ral p

rote

inC

aT

CC

GTa

aa

Ga

CC

TC

TaT

GC

Ca

aC

171

bp

aT

GG

aG

CC

aC

CG

aT

CC

aC

a

Gap

dh (

nM

_008

084.

2)G

lyce

rald

ehyd

e-3-

phos

phat

e de

hydr

ogen

ase

Gly

coly

tic p

athw

ay e

nzym

eT

GT

GT

CC

GT

CG

TG

Ga

TC

TG

a15

0 bp

TT

GC

TG

TT

Ga

aG

TC

GC

aG

Ga

G

Ppi

a (n

M_0

0890

7.1)

Pept

idyl

prol

yl is

omer

ase

aC

atal

yzin

g th

e ci

s–tr

ans

isom

eriz

atio

n of

pro

-lin

e im

idic

pep

tide

bond

s in

olig

opep

tides

an

d ac

cele

ratin

g th

e fo

ldin

g of

pro

tein

TG

TG

CC

aG

GG

TG

GT

Ga

CT

TT

144

bp

CG

TT

TG

TG

TT

TG

GT

CC

aG

Ca

T

Tbp

(n

M_0

1368

4.3)

TaTa

-box

bin

ding

pro

tein

Gen

eral

rn

a p

olym

eras

e II

tran

scri

ptio

n fa

ctor

aG

Ga

TG

CT

CTa

GG

Ga

aG

aT

175

bp

TG

aa

TaG

GC

TG

TG

Ga

GTa

aG

T

Rps

29 (

nM

_009

093.

2)r

ibos

omal

pro

tein

S29

enc

odes

a r

ibos

omal

pro

tein

that

is a

com

po-

nent

of

the

40S

subu

nit

Ga

GC

CG

aC

TC

GT

TC

CT

TT

156

bp

TG

TT

Ca

GC

CC

GTa

TT

TG

C

Hpr

t1 (

nM

_013

556.

)H

ypox

anth

ine

guan

ine

Met

abol

ic s

alva

ge o

f nu

cleo

tides

CG

TC

GT

Ga

TTa

GC

Ga

TG

aT

G78

bp

Phos

phor

ibos

yl tr

ansf

eras

e I

TC

Ca

aa

TC

CT

CG

GC

aTa

aT

G

Mol Genet Genomics

1 3

Graphpad prism (Graphpad software, Inc. la Jolla, Ca, USa) was used to create box-and-whiskers plots of candi-date reference gene Cq values during the development of the mouse testis. For each gene, Cq values obtained from different developmental stages were pooled together. The boxes corresponding to the Cq values within the 25th and 75th percentiles and the median are indicated by a horizon-tal line, while the whiskers represent the minimum–maxi-mum ranges.

Determination of reference gene expression stability

To assess the stability of candidate reference genes in the development of the mouse testis, we utilized the genorm 3.5 (Vandesompele et al. 2002), normFinder (andersen et al. 2004), Bestkeeper (Pfaffl et al. 2004) and comparative delta Ct (Silver et al. 2006) algorithms. The overall ranking of candidate reference genes was generated according to an integration tool refFinder (http://www.leonxie.com/referencegene.php?type=reference).

For genorm and normFinder, the raw Cq values were transformed into relative quantities using the delta Ct method. genorm software ranks the reference gene based on the gene expression stability measure M, with the low-est M values equaling the most stable expression. Moreo-ver, genorm software calculates the pairwise variation V to determine the number of reference genes required for accu-rate normalization. Commonly, 0.15 is an accepted cutoff, below which an additional reference gene is not required for accuracy normalization. normFinder ranks the set of candidate normalization genes according to their expres-sion stability S for each reference. It also takes into account the intragroup and intergroup variations. The most stable reference genes are those with the least intra- and inter-group variation and that have the lowest stability values.

We used the Bestkeeper applet, an excel-based tool, to estimate the gene expression variation for candidate genes by the calculated standard deviation (SD) and coefficient of variation (CV) using their raw Cq values. Candidate genes with SD values lower than 1 were considered to be stably expressed. The most stably expressed genes were deter-mined based on the correlation coefficient (r) to the Best-keeper Index (BI), which is the geometric mean of Cq val-ues of candidate reference genes, and the P values.

Furthermore, the comparative delta Ct method was also used to estimate the most stable reference genes. By com-paring the relative expression of “pairs of genes” within each stage, this method produced the mean of the SD of each candidate gene. The genes with the lowest SD values were expected to be the most stable genes.

To evaluate the results from the four algorithms, an inte-gration tool refFinder (http://www.leonxie.com/referencegene.php?type=reference) for analyzing reference genes

was used. according to the reference genes ranked by every algorithm from most to least stable, an appropriate weight was assigned to an individual gene and the geometric mean of their weights was calculated for the overall final rank-ing. The gene with the smaller geometric mean was then viewed as the more stable reference gene.

expression levels of AT1 are influenced by the choice of a normalization gene(s)

To assess the validity of reference gene ranking, the expres-sion profile of angiotensin type 1 receptor (AT1) was tested in the developing mouse testis. We normalized AT1 expres-sion to each reference gene using the method described by Pfaffl (Pfaffl et al. 2004). The relative expression ratio (R) was calculated based on the following formula: Ratio =

(Etarget)∆Cqtarget(control−sample)

/(Eref)∆Cqref(control−sample). The

reference sample was obtained from the e18 time point. Since the evaluation of the optimal number of reference genes performed by Genorm identified a normalization fac-tor that included the three best reference genes as the best option, we further carried out a normalization to the nor-malization factor NF3 (the geometric mean of the expres-sion data of Ppia, Gapdh and Actb, the three most stable reference genes identified by Genorm software). Vandes-ompele and his colleagues proposed that this geometric mean was considered to be a unique reference entity. To cal-culate the relative expression ratio (R) of AT1, we used the following equation: Ratio = (Etarget)

∆Cqtarget(control−sample)/

Geomean(Eref)∆Cqref(control−sample).

MIQe guidelines

The minimum information for publication of quantitative real-time PCr experiments (MIQe) guidelines (Bustin et al. 2009) was used to promote implementation of experi-mental consistency and to increase the reliability and integ-rity of the results we obtained.

Results

In order to determine the expression stability of the six candidate reference genes during the development of the mouse testis, we began by verifying the specificity of the products, calculating the PCr amplification efficiencies and analyzing the Cq values. next, genorm, normFinder, Bestkeeper, and comparative delta Ct algorithms, and an integration tool refFinder, were employed to establish the expression stability of candidate genes. Finally, the effects of various reference genes on the gene expression profiles were assessed by calculating the relative expression of aT1 in mouse testis at different chronological ages using each

Mol Genet Genomics

1 3

single reference gene as well as the combinations of best reference genes as recommended by genorm, normFinder, Bestkeeper, comparative ΔCt and the integration tool ref-Finder. a total of 33 cDna samples, including various developmental stages, were analyzed.

Specificity and primer efficiency in rT-qPCrs

The size and specificity of amplicons were confirmed by agarose gel electrophoresis and melting curve analysis. The products of predicted size were amplified using all primer pairs, and there were no primer dimers produced in either cDna or no-template control reactions (Fig. 1). Moreover, a single peak was observed in the dissociation curve using all primer pairs (Fig. 2), indicating that the specific prod-uct was amplified. Furthermore, individual melting curves showed that the temperature of the dissociation curve was above 74 °C, indicating that no primer dimers were pro-duced, because primer dimers normally dissociate at lower temperatures (Fig. 2). In all of the rT-qPCrs, standard curves were included (using cDna serial dilutions) and the amplification efficiencies for the six candidate refer-ence genes ranged from 91 to 101 %. Furthermore, all the squared correlation coefficients (r2) were >0.99. Generally, a PCr efficiency between 90 and 110 % is considered to be acceptable. Therefore, all primers were gene specific, and efficiencies were acceptable for further study. The informa-tion obtained from the rT-qPCr analysis required by the MIQe guidelines is listed in Table 2.

expression levels of candidate reference genes

The expression levels of all six candidate reference genes from 11 different developmental stages of the testis were evaluated using quantification cycles (Cq). The box plot graph of Cq values for all six potential reference genes is shown in Fig. 3. It was clear that the mean Cq values were distributed from lowest in the case of Actb (~16) to highest in the case of Hprt1 (~25) (Fig. 3). The genes could be divided into two groups: a group of genes with a median Cq value below 20 (Actb, Gapdh, Ppia) and one

above 20 (Tbp, Rps29, Hprt1). The range of the Cq val-ues at different developmental stages showed a consider-able variability among the six candidate reference genes (Fig. 3b). It is indicated that a given gene with a narrow range of Cq values is expressed stably in different sam-ples. The fact that Actb and Gapdh possessed Cq values in narrow ranges in all samples showed that Actb and Gapdh were expressed stably in the developing mouse testis (Fig. 3). However, comparison of the raw Cq values was not sufficient to evaluate the expression stability of the candidate reference genes, and this will require more stability verification in order to obtain accurate gene expression data.

expression stability of reference gene candidates

To further evaluate the stability of expression of candidate reference genes, five widely used algorithms were applied to calculate expression stability individually: genorm, nor-mFinder, Bestkeeper, the comparative ΔCt method and an integration tool refFinder.

genorm analysis

For each gene, the raw Cq values were transformed into normalized relative values. The average gene expres-sion stability (M value) of six candidate reference genes was calculated using genorm. The reference genes were ranked based on M values. a lower M value indicated that the gene was more stably expressed. The results evaluated with genorm showed that the ranking from the most stable (lowest M value) to the least stable (highest M value) was as follows: Gapdh | Actb—Ppia—Hprt1—Rps29—Tbp (Fig. 4a; Table 4). The M values of all calculated reference genes were less than the recommended cutoff value of 1.5, suggesting that all of the genes exhibited stable expression during mouse testis development. Moreover, Gapdh and Actb exhibited the lowest M values by successive elimina-tion of the least stable reference genes based on highest M values; hence, Gapdh and Actb showed the greatest stabil-ity of expression.

Fig. 1 agarose gel electropho-resis of rT-qPCr products. The specific products of expected size were amplified using all six primer pairs. no discernible bands were observed in no-template control reactions. Std 100 base-pair marker

Mol Genet Genomics

1 3

In addition, the optimal number of reference genes can be evaluated according to genorm, which calculates pairwise variation values (Vn/n + 1 value) between two sequential normalization factors (nFn and nFn + 1). In our data sets, as the number of genes increased, the varia-tion decreased from 0.218 (V2/3 value) to 0.186 (V4/5 value)

and then increased to 0.263 (V5/6 value) (Fig. 4f). Vandes-ompele et al. (2002) recommended 0.15 as a cutoff value, below which the inclusion of an additional reference gene is not required, and the recommended number of reference genes is given by n. However, the 0.15 value must not be taken as a too strict cutoff, as other studies have reported a

Fig. 2 Dissociation curve analysis of primer-specific amplicons. Dissociation analysis was performed by constant monitoring of fluorescence signal from 60 to 95 °C. all primer pairs a–f amplified signal products with melting temperature above 75 °C

Mol Genet Genomics

1 3

higher cutoff value, depending upon the number of genes and types of samples (Kuijk et al. 2007). The observed trend (of changing V values when using additional genes) can sometimes be equally informative. The lower variable V reflects less variation, so the more useful combination of genes to be used as references is that of the lower pairwise variation. In this study, the V4/5 value showed the lowest V value (V4/5 = 0.186), which indicated that four reference genes were required to obtain an accurate normalization factor. However, it is impractical to use an excessive num-bers of reference genes for normalization, especially when studying a small number of target genes or rare samples (Vandesompele et al. 2002; Gu et al. 2011; Fletcher and Mullins 2010). Moreover, genorm recommends the use of three reference genes as a valid method for an accurate nor-malization strategy, compared with a single non-validated reference gene. Therefore, it has been considered accept-able for the majority of experiments to use the three most

stable reference genes for the calculation of normalization factors (Vandesompele et al. 2002; Gu et al. 2011; Fletcher and Mullins 2010). Considering these factors, we then used Gapdh, Actb and Ppia as stable reference genes in our rT-qPCr experiment, although the value of V4/5 was the mini-mal value.

normFinder analysis

The stability value of each reference gene was also cal-culated using the normFinder algorithm. Genes that are more stably expressed are indicated as lower average expression stability values. From the most to least stable, the rank order of the six genes by normFinder was Ppia—Rps29—Hprt1—Actb—Gapdh—Tbp (Fig. 4b; Table 4). Tbp was identified as the least stable gene by genorm and normFinder. In contrast to genorm, normFinder identified Ppia as the best gene, which was different from the data obtained using genorm. Since statistical algorithms such as genorm and normFinder are distinct, the observed differ-ences between the two programs are to be expected.

Bestkeeper analysis

In order to validate the observations derived from genorm and normFinder, the Bestkeeper applet was used to further cal-culate the gene expression variation of candidate genes. This program was designed based on each candidate gene’s Cq val-ues. The SD, CV, r and corresponding P values are shown in Table 3. From most stable (lowest SD) to least stable (highest

Table 2 Primer-specific amplification efficiencies of the candidate reference gene

Gene symbol Slope Y-intercept r2 efficiency (100 %)

Actb −3.492 30.69 0.998 93

Gapdh −3.478 28.64 0.995 94

Ppia −3.332 27.36 0.994 100

Tbp −3.304 30.55 0.991 101

Rps29 −3.546 32.32 0.991 91

Hprt1 −3.439 33.79 0.994 95

Fig. 3 expression levels of candidate reference genes in all cDna samples. a Values of six candidate reference genes are given as the average rT-qPCr quantification cycle (Cq) values. The boxes showed the Cq values within the 25th and 75th percentiles, and the median is indicated by an horizontal line, while Whiskers represent the

min–max ranges. b Variations of six candidate reference genes was analyzed during the developmental stages with e18–4M. Symbols represent mean ± Se. Points are connected by solid lines to better understand the tendency to increase or decrease during development

Mol Genet Genomics

1 3

Fig. 4 ranking of six candidate reference genes using different algo-rithms and determination of the optimal number of reference genes for normalization by genorm. a The candidate reference genes were ranked by genorm based on their stability parameter M. The lower the M value, the higher the expression stability. b normFinder ranks the reference genes according to calculated stability value. The lower the stability value, the higher the expression stability. c Bestkeeper applet ranks the reference genes based on the coefficient of correla-tion (r) to the Bestkeeper Index (BI), which is the geometric mean of Cq values of candidate reference genes and the P values. d The rank of the candidate reference genes was determined by comparative ΔCt

method. The genes with lowest SD value were suggested to be the most stable gene. e Final ranking of candidate reference genes was evaluated by an integration tool refFinder. f The optimal number of reference genes for normalization was determined by the Genorm software. This software calculates the pairwise variation (Vn/n + 1) between two sequential normalization factors containing an increas-ing number of reference genes. The lower the pairwise variation, the better is the combination of genes for reference. There was lowest value in V4/5, which shows that the five genes are the combination providing the lowest pairwise variation

Mol Genet Genomics

1 3

SD), the rank generated by Bestkeeper was as follows: Gapdh, Actb, Ppia, Hprt1, Rps29, Tbp (Fig. 4c; Table 4). The results were the same as those from genorm analysis. Moreover, the least stable gene (Tbp) evaluated by Bestkeeper was identical to that using genorm or normFinder.

Comparative ΔCt method analysis

Furthermore, the most stable reference genes were also estimated by the comparative ΔCt method. The result was the same as with normFinder. From most to least stable genes, the ranking was Ppia, Rps29, Hprt1, Actb, Gapdh, Tbp (Fig. 4d; Table 4). Tbp was evaluated as the least sta-ble gene, and this was commensurate with the assessments using genorm, normFinder and Bestkeeper.

Final ranking of candidate reference genes

Based on the separation evaluations performed using genorm, normFinder, Bestkeeper and comparative ΔCt analysis, we found that the results were not identical among the four algorithms. In this case, refFinder inte-grated these four computational programs to compare and

rank the tested candidate reference genes. The recom-mended comprehensive rankings were Ppia, Gapdh, Actb, Rps29, Hprt1 and Tbp (Fig. 4e; Table 4). Therefore, Ppia, Gapdh and Actb are presented as the most stable genes dur-ing the various stages of mouse testis development. The geometric mean of these three most stable reference genes would likely be sufficient for use as a normalization factor for gene expression studies in the developing mouse testis.

expression levels of AT1 are influenced by the choice of a normalizer gene(s)

In order to validate the effects of various reference genes on gene expression profiles, we calculated the relative expres-sion of AT1 in testes from mice at different developmental ages from e18 to 4M (including e18, P1, P7, P14, P18, P22, P26, P30, 2M, 3M, 4M), using each single reference gene as well as the recommended combinations of genes from genorm, normFinder, Bestkeeper, comparative ΔCt and the integration tool refFinder. The Cq values for AT1 were between 22 and 27. The relative AT1 expression level was calculated relative to the e18 group and is shown in Fig. 5. It was observed that the mrna expression levels of

Table 3 expression stability evaluated by Bestkeeper

n number of samples, Cq quantification cycle, GM (Cq), the geometric mean of Cq, SD (±Cq) Cq standard deviation, CV (%Cq)variance coefficient expressed as percentage of Cq level

Gapdh Ppia Tbp Hprt1 Rps29 Actb

n 11 11 11 11 11 11

Geo mean (Cq) 18.38 18.71 23.23 25.31 24.54 16.05

ar mean (Cq) 18.38 18.72 23.31 25.32 24.57 16.05

Min (Cq) 17.64 17.66 21.18 24.40 22.98 15.68

Max (Cq) 18.87 19.88 27.06 27.08 27.07 16.67

Min (x-fold) −1.67 −2.07 −4.15 −1.87 −2.95 −1.29

Max (x-fold) 1.41 2.25 14.17 3.42 5.77 1.54

SD (±Cq) 0.23 0.50 1.62 0.75 0.87 0.24

SD (±x-fold) 1.17 1.41 3.06 1.68 1.82 1.18

GM (Cq) 18.38 18.71 23.23 25.31 24.54 16.05

SD (±Cq) 0.23 0.50 1.62 0.75 0.87 0.24

CV (% Cq) 1.26 2.65 6.93 2.96 3.52 1.52

Coeff. of corr. (r) 0.001 0.971 0.937 0.729 0.969 0.271

P value 0.997 0.001 0.001 0.011 0.001 0.420

Table 4 The comprehensive ranking of all six candidate reference genes

ranking genorm normFinder BestKeeper Comparative ΔCt method

recommended comprehensive ranking

1 Gapdh/Actb Ppia Gapdh Ppia Ppia

2 Rps29 Actb Rps29 Gapdh

3 Ppia Hprt1 Ppia Hprt1 Actb

4 Hprt1 Actb Hprt1 Actb Rps29

5 Rps29 Gapdh Rps29 Gapdh Hprt1

6 Tbp Tbp Tbp Tbp Tbp

Mol Genet Genomics

1 3

Fig. 5 expression levels of AT1 in mice testis during the developmental stages with e18–4M. a The gene expression levels of AT1 were normal-ized to the geometric mean of Ppia, Gapdh and actb. The relative expression values were expressed as a ratio to the e18 testis samples and represented as the mean ± SeM. b–g The gene expression levels of AT1 were normalized to different sets of signal reference gene (Ppia, Gapdh, Actb, Rps29, Hprt1, Tbp). The normalized gene expression values were shown for the different stage samples relative to the corre-sponding values at e18 (n = 3). Data were represented as the mean ± SeM

Mol Genet Genomics

1 3

AT1 increased with age using Ppia, Gapdh and the geomet-ric mean of Ppia, Gapdh and Actb. results obtained from correlational analysis also showed that the AT1 increase in expression was correlated with age (P < 0.05). although the relative expression level of AT1 at 4M was decreased using Actb as the reference gene, the relative expression level of AT1 still had a statistically significant correlation with age (r2 = 0.851, P < 0.05). This significance was not observed when using Hprt1, Rps29 or Tbp as the internal control (P > 0.05).

Discussion

rT-qPCr is a powerful tool that is used to accurately detect low-abundance mrna transcripts. In this tech-nique, normalization with stable reference genes is very important for accurate interpretation of variability. However, it is virtually impossible to find a stable refer-ence gene under all of these conditions. Therefore, when researchers perform rT-qPCr to detect mrna transcripts using samples from different cell types or different time points, they usually select the appropriate stable reference genes by various methods for their experimental systems. In the present study, we selected six candidate reference genes representing different classes to determine the most stable reference genes to use for accurate normalization of gene expression in the mouse testis during embryonic and postnatal development. after performing rT-qPCr, we used genorm, normFinder, Bestkeeper, the comparative ΔCt program and the integration tool refFinder to analyze the stability of all six of the candidate reference genes and identified three reference genes (Ppia, Gapdh, Actb) as the most stable.

In this study, genorm calculated the M values of all ref-erence genes that were less than the recommended cutoff value of 1.5, suggesting that all of the genes exhibited sta-ble expression during mouse testis development. Moreover, Gapdh and Actb showed the greatest stability of expression, followed by Ppia. The Bestkeeper method also identified Gapdh, Actb and Ppia as the most stable reference genes, and the ranking from most to least stable was identical to that which we achieved from genorm. Both normFinder and the comparative ΔCt method determined that Ppia was the most stable reference gene, followed by Rps29 and Hprt1. These two methods also provided the same ranking results. all four programs assessed Tbp as the least sta-ble candidate reference gene, so this gene was not recom-mended for use in evaluation of mouse testis development. The results from genorm and the Bestkeeper method were consistent with each other, but they were different from the results acquired from normFinder and comparative ΔCt (which were identical to each other). The final ranking of

the six tested reference genes was identified by the inte-gration tool refFinder. The recommended comprehensive ranking was Ppia, Gapdh, Actb, Rps29, Hprt1 and Tbp.

In the rT-qPCr analysis, there was a trend toward the use of more than one reference gene. as described above, the optimal number of reference genes could be evaluated by calculating pairwise variations between consecutively determined normalization factors of n and n + 1 genes (Vn/Vn + 1) using genorm. The number of genes required for optimal normalization is n when Vn/n + 1 < 0.15, although this threshold should not be viewed too strictly as a cutoff. at times, the trend for changing V values can also be equally informative, and the most useful combi-nation of genes to be used as references is that with the lower pairwise variation. The Vn/n + 1 values showed an unusual trend in our study, with V4/5 providing the low-est value that was close to the cutoff of 0.15. The results from genorm suggested that the optimal number of ref-erence genes was four. However, it is impractical to use an excessive number of reference genes for normalization because of the expense and time involved. Using the three best reference genes is a valid normalization strategy in most cases, and in our hands, this strategy has resulted in more accurate and reliable normalization compared to using a single reference gene. The current study suggests that Ppia, Gapdh, Actb would be three valid reference genes for the analysis of developmental gene expression in mouse testis, although some studies have shown that Gapdh and Actb (Glare et al. 2002; Selvey et al. 2001) are variable and unsuitable references for rna transcrip-tional analysis. Ppia is a gene that catalyzes the cis–trans isomerization of proline imidic peptide bonds in oligo-peptides and accelerates the folding of protein. Gapdh is an enzyme of ~37 kDa that catalyzes an important energy-yielding step in carbohydrate metabolism. additionally, it takes part in apoptosis, membrane trafficking, iron metabolism, nuclear activities and receptor-mediated cell signaling. Actb is also called beta-actin protein and par-ticipates in the composition of cytoskeletal structure. It is a highly conserved protein that is involved in cell motility, structure and integrity (provided by refSeq, Jul 2008).

In order to demonstrate the effects of using various ref-erence genes on gene expression profiles, the expression of AT1 in mouse testis during development from e18 to 4M was calculated using each of six reference genes as well as the geometric mean of Ppia, Gapdh and Actb. AT1 was selected because there is considerable evidence for the local synthesis of components of the renin-angioten-sin system (RAS) in the male reproductive tract, including testis, epididymis and prostate (leung and Sernia 2003). a complete, classic RAS [including renin, an angiotensin-converting enzyme (ACE), angiotensin II and angioten-sin receptor subtypes AT1, AT2] has been found in the

Mol Genet Genomics

1 3

testes (leung and Sernia 2003). It has been shown that AT1 receptors exist in the tails of non-capacitated sperm, and in a head-to-tail distribution in capacitated sperm, and acrosomal exocytosis was induced by Ang II only in capacitated sperm. In addition, Ang II can maintain sperm motility via the AT1 receptor (leung and Sernia 2003). Investigation of the changes in the expression level of AT1 in mouse testis at various developmental stages can aid us in partially understanding the important role that AT1 plays in the regulation of spermatogenesis. The results of our analysis indicated that the expression level of AT1 increased with age regardless of whether a normaliza-tion factor or single stable reference genes (Ppia, Gapdh) were used. We assessed the correlation between the rela-tive expression of AT1 and age, and the relative expression of AT1 (which was calculated by normalization with Ppia, Gapdh, Actb and the geometric mean of Ppia, Gapdh, Actb) was significantly correlated with embryonic/postna-tal age. However, this correlation was not observed when the expression level of AT1 was normalized with Hprt1, Rps29 or Tbp, consistent with a previous report showing that the selection of a reference gene has a distinct influ-ence on the expression of target genes (nolan et al. 2006). although Kanehara et al. (1998) suggested that mrna expression of both AT1 and AT2 subtypes in rat decreased with age as assessed by northern blot analysis, our results indicated by rT-qPCr that the mrna expression of AT1 receptor subtypes in mice increased with age. This differ-ence between the two studies may be attributed to animal strain or reference gene(s) selection.

In conclusion, using rT-qPCr and five algorithms to evaluate reference gene expression stability in mouse tes-tis at different stages, we identified Ppia, Gapdh and Actb as the most suitable reference genes among six candidate reference genes; and the geometric mean of Ppia, Gapdh and Actb was validated as an accurate normalization factor for rT-qPCr normalization. Moreover, the analysis of the expression level of the target gene AT1 in the developing mouse testis emphasized the importance of proper selection of reference genes to ensure accurate analysis of rT-qPCr in such experiments. Our studies provide crucial informa-tion for transcriptional analysis in the developing mouse testis and should assist in improving the overall quality of gene expression data in this organ. This will also aid our understanding of signaling events and the expression of tes-tis-specific genes during testis development.

Acknowledgments This research project was funded by the national natural Science Foundation of China (Grant no. 81160263), the Scientific research Fund of Guangxi education Department (Grant no. 200103YB025) and the Guangxi natural Science Foun-dation Program (Grant no. 2010GXnSFa013162). We thank letPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

References

andersen Cl, Jensen Jl, Ørntoft TF (2004) normalization of real-time quantitative reverse transcription-PCr data: a model-based variance estimation approach to identify genes suited for normal-ization, applied to bladder and colon cancer data sets. Cancer res 64(15):5245–5250

Bas a, Forsberg G, Hammarström S, Hammarström Ml (2004) Util-ity of the Housekeeping genes 18 s rrna, beta-actin and glycer-aldehyde-3-phosphate -dehydrogenase for normalization in real-time quantitative reverse transcriptase-polymerase chain reaction analysis of gene expression in human T-lymphocytes. Scand J Immunol 59(6):566–573

Boda e, Pini a, Hoxha e, Parolisi r, Tempia F (2009) Selection of reference genes for quantitative real-time rT-PCr studies in mouse brain. J Mol neurosci 37(3):238–253

Bustin Sa, Benes V, Garson Ja, Hellemans J, Huggett J, Kubista M, Mueller r, nolan T, Pfaffl MW, Shipley Gl, Vandesompele J, Wittwer CT (2009) The MIQe guidelines: minimum information for publication of quantitative real-time PCr experiments. Clin Chem 55(4):611–622

Chalmel F, lardenois a, Georg I, Barrionuevo F, Demougin P, Jégou B, Scherer G, Primig M (2013) Genome-wide identification of Sox8-, and Sox9-dependent genes during early post-natal testis development in the mouse. andrology 1(2):281–292

Chechi K, Gelinas Y, Mathieu P, Deshaies Y, richard D (2012) Vali-dation of reference genes for the relative quantification of gene expression in human epicardial adipose tissue. PloS One 7(4):e32265

Cicinnati Vr, Shen Q, Sotiropoulos GC, radtke a, Gerken G, Beck-ebaum S (2008) Validation of putative reference genes for gene expression studies in human hepatocellular carcinoma using real-time quantitative rT-PCr. BMC Cancer 8:350

Czechowski T, Stitt M, altmann T, Udvardi MK, Scheible Wr (2005) Genome-wide identification and testing of superior reference genes for transcript normalization in arabidopsis. Plant Physiol 139(1):5–17

Dheda K, Huggett JF, Chang JS, Kim lU, Bustin Sa, Johnson Ma, rook Ga, Zumla a (2005) The implications of using an inappro-priate reference gene for real-time reverse transcription PCr data normalization. anal Biochem 344(1):141–143

Faccioli P, Ciceri GP, Provero P, Stanca aM, Morcia C, Terzi V (2007) a combined strategy of “in silico” transcriptome analysis and web search engine optimization allows an agile identification of reference genes suitable for normalization in gene expression studies. Plant Mol Biol 63(5):679–688

Feng l, Yu Q, li X, ning X, Wang J, Zou J, Zhang l, Wang S, Hu J, Hu X, Bao Z (2013) Identification of reference genes for qrT-PCr analysis in Yesso Scallop Patinopecten yessoensis. PloS One 8(9):e75609

Fletcher Da, Mullins rD (2010) Cell mechanics and the cytoskel-eton. nature 463(7280):485–492

Gebeh aK, Marczylo el, amoako aa, Willets JM, Konje JC (2012) Variation in stability of endogenous reference genes in fallo-pian tubes and endometrium from healthy and ectopic pregnant women. Int J Mol Sci 13(3):2810–2826

Glare eM, Divjak M, Bailey MJ, Walters eH (2002) Beta-actin and GaPDH housekeeping gene expression in asthmatic airways is variable and not suitable for normalising mrna levels. Thorax 57(9):765–770

Gu Yr, li MZ, Zhang K, Chen l, Jiang aa, Wang JY, li XW (2011) evaluation of endogenous control genes for gene expres-sion studies across multiple tissues and in the specific sets of fat- and muscle-type samples of the pig. J anim Breed Genet 128(4):319–325

Mol Genet Genomics

1 3

Hong SY, Seo PJ, Yang MS, Xiang F, Park CM (2008) exploring valid reference genes for gene expression studies in Brachypodium dis-tachyon by real-time PCr. BMC Plant Biol 8:112

Kanehara H, Song K, Hirai K, Ueda H, Shiota n, azuma H, Katsuoka Y, Miyazaki H, Miyazaki M (1998) Involvement of angiotensin II receptor subtypes during testicular development in rats. Int J androl 21(4):186–195

Kuijk eW, du Puy l, van Tol HT, Haagsman HP, Colenbrander B, roelen Ba (2007) Validation Of reference genes for quantitative rT-PCr studies in porcine oocytes and preimplantation embryos. BMC Dev Biol 7:58

leung PS, Sernia C (2003) The renin-angiotensin system and male reproduction: new functions for old hormones. J Mol endocrinol 30(3):263–270

Migrenne S, Moreau e, Pakarinen P, Dierich a, Merlet J, Habert r, racine C (2012) Mouse testis development and function are dif-ferently regulated by follicle-stimulating hormone receptors sign-aling during fetal and prepubertal life. PloS One 7(12):e53257

Miles DC, Wakeling SI, Stringer JM, van den Bergen Ja, Wilhelm D, Sinclair aH, Western PS (2013) Signaling through the TGF beta-activin receptors alK4/5/7 regulates Testis formation and male germ cell development. PloS One 8(1):e54606

nolan T, Hands re, Bustin Sa (2006) Quantification of mrna using real-time rT-PCr. nat Protoc 1(3):1559–1582

Paolacci ar, Tanzarella Oa, Porceddu e, Ciaffi M (2009) Identifica-tion and validation of reference genes for quantitative rT-PCr normalization in wheat. BMC Mol Biol 10:11

Pfaffl MW, Tichopad a, Prgomet C, neuvians TP (2004) Determina-tion of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper–excel-based tool using pair-wise correlations. Biotechnol lett 26(6):509–515

Phillips BT, Gassei K, Orwig Ke (2010) Spermatogonial stem cell regulation and spermatogenesis. Philos Trans r Soc lond B Biol Sci 365(1546):1663–1678

Pombo-Suarez M, Calaza M, Gomez-reino JJ, Gonzalez a (2008) reference genes for normalization of gene expression studies in human osteoarthritic articular cartilage. BMC Mol Biol 9:17

Selvey S, Thompson eW, Matthaei K, lea ra, Irving MG, Griffiths lr (2001) Beta-actin-an unsuitable internal control for rT-PCr. Mol Cell Probes 15(5):307–311

Silver n, Best S, Jiang J, Thein Sl (2006) Selection of housekeeping genes for gene expression studies in human reticulocytes using real-time PCr. BMC Mol Biol 7:33

Stephens aS, Stephens Sr, Morrison na (2011) Internal control genes for quantitative rT-PCr expression analysis in mouse oste-oblasts, osteoclasts and macrophages. BMC res notes 4:410

Tricarico C, Pinzani P, Bianchi S, Paglierani M, Distante V, Pazzagli M, Bustin Sa, Orlando C (2002) Quantitative real-time reverse transcription polymerase chain reaction: normalization to rrna or single housekeeping genes is inappropriate for human tissue biopsies. anal Biochem 309(2):293–300

Vandesompele J, De Preter K, Pattyn F, Poppe B, Van roy n, De Paepe a, Speleman F (2002) accurate normaliza-tion of real-time quantitative rT-PCr data by geometric averaging of multiple internal control genes. Genome Biol 3(7):research0034.1–research0034.11

Wan H, Zhao Z, Qian C, Sui Y, Malik aa, Chen J (2010) Selection of appropriate reference genes for gene expression studies by quan-titative real-time polymerase chain reaction in cucumber. anal Biochem 399(2):257–261

Wong Ml, Medrano JF (2005) real-time PCr for mrna quantita-tion. Biotechniques 39(1):75–85

Zhang Y, Chen D, Smith Ma, Zhang B, Pan X (2012) Selection of reliable reference genes in Caenorhabditis elegans for analysis of nanotoxicity. PloS One 7(3):e31849