Appropriate Real-Time PCR Reference Genes for Fluoride Treatment Studies Performed In Vitro or In Vivo

Abstract

Objective

Quantitative real-time PCR (qPCR) is routinely performed for experiments designed to identify the molecular mechanisms involved in the pathogenesis of dental fluorosis. Expression of reference gene(s) is expected to remain unchanged in fluoride-treated cells or in rodents relative to the corresponding untreated controls. The aim of this study was to select optimal reference genes for fluoride experiments performed in vitro and in vivo.

Design

Five candidate genes were evaluated: B2m, Eef1a1, Gapdh, Hprt and Tbp. For in vitro experiments, LS8 cells derived from mouse enamel organ were treated with 0, 1, 3 and/or 5 mM sodium fluoride (NaF) for 6 or 18 hours followed by RNA isolation. For in vivo experiments, six-week old rats were treated with 0 or 100 ppm fluoride as NaF for six weeks at which time RNA was isolated from enamel organs. RNA from cells and enamel organs were reverse-transcribed and stability of gene expression for the candidate reference genes was evaluated by qPCR in treated versus non-treated samples.

Results

The most stably expressed genes in vitro according to geNorm were B2m and Tbp, and according to Normfinder were Hprt and Gapdh. The most stable genes in vivo were Eef1a1 and Gapdh. Expression of Ddit3, a gene previously shown to be induced by fluoride, was demonstrated to be accurately calculated only when using an optimal reference gene.

Conclusions

This study identifies suitable reference genes for relative quantification of gene expression by qPCR after fluoride treatment both in cultured cells and in the rodent enamel organ.

1. Introduction

Supplementation of drinking water with fluoride is an established method for prevention of caries.¹ The recommended level of water fluoridation in the U.S.A. is 0.7 ppm.² This level of fluoride was determined to present minimal risk for dental fluorosis and reduced the prevalence of caries, which is an important public health concern. In many areas of the world, natural levels of fluoride in drinking water are higher than the levels recommended for prevention of caries. The risk of dental fluorosis increases when the levels of fluoride in drinking water exceed 1.6 ppm.³ In recent years the prevalence of dental fluorosis has increased.⁴ Studies showed that fluoride plasma levels depend on weight, nutrition, growth status, renal function, and genetic background.^5–8 However, the severity of fluorosis is positively associated with the concentration of fluoride within the plasma.^{9, 10} In the United States, prevalence of dental fluorosis among children aged 12–15 increased from 22.6% in 1986–1987 to 40.7% in 1999–2004.¹¹ Approximately a third of the 6–29 year old age group is affected by fluorosis, due to the expansion of water fluoridation and increased availability of additional sources of ingested fluoride such as processed foods.¹²

The mechanism by which high levels of fluoride induce defects in mineralization is the subject of active research. Dental fluorosis causes sub-optimal enamel mineralization, which can manifest as a chalky-white appearance and loss of transparency,^{13, 14} and this may have a negative impact on a child’s quality of life.¹⁵ More severe fluorosis causes brownish enamel discoloration and weakening of the enamel that tends to chip from the teeth. The reduced mechanical properties of fluorosed enamel result in malfunction and increased susceptibility to decay.^{16, 17} Despite its normal thickness, fluorosed enamel is deficient in mineral and contains excessive matrix proteins.^{14, 18–23} Excessive levels of fluoride mainly affect developing unerupted teeth during the maturation stage of enamel development when the fully thick enamel layer begins hardening into its final form.^{18, 24–27} We have previously demonstrated that ameloblasts respond to excess fluoride in part by phosphorylation of eIF2α^{28, 29} and we have recently shown that expression of the proteinase kallikrein-related peptidase-4 (KLK4), responsible for cleaving enamel matrix proteins prior to their removal from the hardening enamel, is inhibited by fluoride treatment.³⁰ Serum fluoride levels in rat drinking water supplemented with 50 ppm fluoride are similar to humans ingesting 2–5 ppm fluoride.³¹ Clearance of fluoride from rodents is more efficient than clearance from humans.¹⁰ In addition, development of the rodent incisor takes approximately 35 days to complete whereas in humans tooth development takes at least three years.³¹ Thus, in humans, ameloblasts are exposed to fluoride present in drinking water for a longer time relative to rodent ameloblasts and this may be a contributing factor to the large difference in sensitivity to fluoride between humans and rodents.

Analysis of gene expression is routinely utilized to investigate fluoride toxicity in vivo and in vitro. Currently, a frequently used technique for gene expression analysis is quantitative real-time polymerase chain reaction (qPCR). The qPCR workflow is comprised of RNA isolation, synthesis of cDNA, selection of primers, followed by the qPCR reaction and analysis of gene expression. Specific limitations were reported for each of these steps. For example, RNA isolation results in variable co-purification of PCR inhibitors, such as polyphenolics or polysaccharides. In addition, there are differences in primer amplification efficiencies.³² In an effort to standardize this workflow, a series of guidelines for design and analysis of qPCR experiments were published by Bustin et al.³³ A common method of analysis for physiological changes in gene expression by qPCR is relative quantification. Relative quantification experiments are designed to compare the expression of treated samples with non-treated samples from the same group. The basic assumption of this design is that the treatment to which the samples were exposed results in differences in gene expression.

Accepted analyses of relative quantification experiments are the ΔΔCt method³⁴ and the primer efficiency method.³⁵ The ΔΔCt method assumes an ideal amplification efficiency of 2, meaning that the quantity of amplicon doubles in each cycle of amplification. In reality, experimental conditions determine the efficiency of amplification. The primer efficiency method takes into account differences in amplification, and relies on estimation of the efficiency from a standard curve. An amplification efficiency value greater than 1.8 is regarded as appropriate.³⁶ In both methods, the expression of a chosen target gene is calculated relative to one or more reference genes. The reference gene or genes are required to be unaffected by the experimental treatment.³⁴ The efficiency of amplification of the reference gene should be similar to that of the target gene. When fluorescence incorporated into the qPCR amplification product rises above background level, the quantification cycle (C_q) is recorded for each sample in the exponential phase of the PCR reaction.

Genes involved in recurrent cell processes, such as amino acid synthesis, are termed “housekeeping genes” and these genes are typically used as reference genes for qPCR. The expression level of housekeeping genes, however, may change depending on the experimental conditions.^37–44 For example, the expression of β-actin (Actb) and glyceraldehyde-3-phosphate dehydrogenase (Gapdh) mRNAs were found to fluctuate upon activation of T-lymphocytes.⁴³ The use of inappropriate reference genes may lead to inaccurate and misleading quantification of target gene expression.^{45, 46} Results from qPCR experiments should accurately reflect gene expression levels and this can be obtained by using a minimum of three biological replicates.⁴⁷

In fluoride experiments, expression levels of reference genes must remain stable regardless of fluoride treatment. The aim of the current study was to identify optimal qPCR reference genes for NaF treatments both in vivo and in vitro. Four commonly-used reference genes were tested in vivo, and five were tested in vitro: beta-2 microglobulin (B2m), transcription factor TATA binding protein (Tbp), hypoxanthine phosphoribosyltransferase (Hprt) and eukaryotic translation elongation factor 1 alpha 1 (Eef1a1). These genes have a variety of cellular functions: B2m is a component of major histocompatibility complex (MHC) class 1, and a cell surface marker for all nucleated cells,⁴⁸ Tbp is a member of the family of TATA-box transcription factors,⁴⁹ Gapdh is a glycolytic enzyme,⁵⁰ Hprt is part of purine synthesis in the salvage pathway,⁵¹ and Eef1a1 functions in the translational machinery.⁵² The data presented here will facilitate accurate and reproducible transcript profiling studies in fluoride treated rats and in enamel organ-derived LS8 cells.

2. Materials and Methods

2.1. Animals

All animals were treated humanely. Sprague–Dawley rats (6-week-old) were purchased from Charles River Laboratories (Wilmington, MA). Animals were euthanized by CO₂ inhalation after 6 weeks of fluoride treatment. Three rats were used in each group. Incisor enamel organs in the maturation stage of enamel development, were separated from the secretory stage by appearance and hardness. A micro-scalpel blade that slices into soft enamel dissects the secretory stage enamel organ whereas hard resistant enamel is in the maturation stage. RNA from the maturation stage was isolated for use in qPCR assays.

2.2. Cell Culture

LS8 cells⁵³ were maintained in alpha minimal essential medium with GlutaMAX supplemented with 10% fetal bovine serum and 1 mM sodium pyruvate (Life Technologies, Grand Island, NY) and 3 × 10⁵ were plated for experiments. NaF (Cat. S299-100, Fisher Scientific, Pittsburgh, PA) was used as indicated. Fluoride concentrations of 1, 3, or 5 mM were used for the LS8 experiments. Rats were provided water containing 0 or 100 ppm fluoride as NaF ad libitum.

2.3. RNA Isolation and Processing

Total RNA was isolated from LS8 cells or rat enamel organs (Direct-zol™ RNA MiniPrep, Zymo Research Corp., Irvine, CA) according to the manufacturer's instructions. Briefly, cells were lysed using the Trizol reagent (Thermo Fisher Scientific, Waltham, MA). Lysates were transferred to Zymo-Spin IIC Columns. Columns were then washed using Direct-zol™ RNA PreWash buffer and RNA was eluted using RNA Wash Buffer. RNA concentration was assessed by use of a spectrophotometer (Nanodrop, Thermo Fisher Scientific, Waltham, MA). A 260/280 ratio of approximately 2.0 indicated high purity RNA. RNA integrity was assessed by gel electrophoresis (Bioanalyzer, Agilent Technologies, Santa Clara, CA). Total RNA (1 μg) was reverse-transcribed into cDNA (Transcriptor First Strand cDNA Synthesis Kit, Roche Diagnostics, Minneapolis, MN). The reaction mixture was incubated for 10 min at 25°C, 120 min at 37°C and 5 sec at 85°C. Primers (Table 1) were designed using Primer Blast.⁵⁴ Melting temperature was limited to 60°C ± 1°C, primer length 18–25 bases, GC content 40–60%, and product size 60–150 base pairs. Primers were designed for B2m, Eef1a1, Gapdh, Hprt, Tbp and Ddit3 (Table 1).

Table 1.

qPCR Primer Sequences

Target gene / Amplicon length	Accession	Sequence / GC content	Efficiency
Mus musculus Beta-2 microglobulin (B2m) / 119 bp	NM_009735.3	Fwd: 5’-GGT CTT TCT GGT GCT TGT CTC-3’ / 52.38 Rev: 5’-CGT AGC AGT TCA GTA TGT TCG G-3’ / 50.00	1.963
Mus musculus Hypoxanthine phosphoribosyltransferase (Hprt) / 142 bp	NM_013556.2	Fwd: 5’-TCA GTC AAC GGG GGA CAT AAA-3’ / 47.62 Rev: 5’-GGG GCT GTA CTG CTT AAC CAG-3’ / 57.14	1.943
Mus musculus TATA box binding protein (Tbp) / 206 bp	NM_013684	Fwd: 5’-CCA ATG ACT CCT ATG ACC CC-3’ / 55 Rev: 5’-GTT GTC CGT GGC TCT CTT ATT C-3’ / 50	1.936
Mus musculus Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) / 63 bp	NM_001289726.1	Fwd: 5’-ACT GGC ATG GCC TTC CG-3’ / 64.71 Rev: 5’-CAG GCG GCA CGT CAG ATC-3’ / 66.67	1.96
Mus musculus Eukaryotic translation elongation factor 1 alpha 1 (Eef1a1) / 98 bp	NM_010106.2	Fwd: 5’-ATT CCG GCA AGT CCA CCA CAA-3’ / 52.38 Rev: 5’-CAT CTC AGC AGC CTC CTT CTC AAA C-3’ / 52.00	1.895
Mus musculus DNA-damage inducible transcript 3 (Ddit3) / 67 bp	NM_001290183.1	Fwd: 5’-CCA CCA CAC CTG AAA GCA GAA-3’ / 52.38 Rev: 5’-AGG TGA AAG GCA GGG ACT CA-3’ / 55.00	1.995
Rattus norvegicus beta-2 microglobulin (B2m) / 79 bp	NM_012512.2	Fwd: 5’-CGTCGTGCTTGCCATTCAGAA-3’ / 52.38 Rev: 5’-GAAGTTGGGCTTCCCATTCTCC-3’ / 54.55	1.938
Rattus norvegicus glyceraldehyde-3-phosphate dehydrogenase (Gapdh) / 89 bp	NM_017008.4	Fwd: 5’-ATGATTCTACCCACGGCAAG-3’ / 50.00 Rev 5’-CTGGAAGATGGTGATGGGTT-3’ / 50.00	1.932
Rattus norvegicus hypoxanthine phosphoribosyltransferase 1 (Hprt1) / 96 bp	NM_012583.2	Fwd: 5’-AATGCAGACTTTGCTTTCCTTGG-3’ / 43.48 Rev: 5’-TATCCAACACTTCGAGAGGTCCTTT-3’ / 44.00	1.978
Rattus norvegicus eukaryotic translation elongation factor 1 alpha 1 (Eef1a1) / 109 bp	NM_1 75838.1	Fwd: 5’-CTCCACTTGGTCGTTTTGCTG-3’ / 52.38 Rev: 5’-GCAGACTTGGTGACTTTGCC-3’ / 55.00	1.876

2.4. qPCR

cDNA was subjected to real-time PCR amplification on a Light Cycler 480 System using SYBRGreen I Mastermix (Roche Diagnostics Corporation, Indianapolis, IN). The PCR protocol started with a heat activation step of 95°C for 10 min. Then, 40 cycles of thermocycling were performed with a denaturation step at 95°C for 30 sec, an annealing step at 60°C for 30 sec and an extension step at 72°C for 15 sec. Fluorescence was measured at the end of each extension step. After amplification, a melting curve was acquired by incubating the product at 95°C for 5 sec, 65°C for 1 min, and then slowly heating to 97°C. Fluorescence was measured through the slow heating phase. Melting curves were used to ensure that only one PCR product was amplified. C_q was measured using the baseline-independent second derivative maximum method.⁵⁵ A standard curve was prepared for each primer set from the C_q values of 5-point 5-fold serial cDNA dilutions. Each dilution was assayed in triplicate. The assay amplification efficiency, also termed primer efficiency, is the derivation of the linear regression of the standard curve (E = 10^(1/−slope)−1). This value represents the increase in the quantity of amplicon in each qPCR cycle. In an ideal reaction the amplification efficiency is 2. No-template controls were negative in all runs. C_q was smaller than 40 for all reactions. ΔC_q was smaller than 0.5 for all replicates.

2.5. qPCR Data Analysis

Mean C_q values for treated and non-treated samples were recorded for each assay. Fold change for each sample and for each target gene was calculated from raw expression values using primer efficiencies from the standard curves.³⁵ The stability of reference genes was calculated from these fold change values using two methods. The first method was based on the geNorm algorithm using the SLqPCR package (version 1.28.0) on R (version 3.0.2).⁵⁶ According to this method, base 2 logarithm for the ratio of fold change values of all pairs of target genes was calculated for all samples. The standard deviation of the result for all treatments for these two target genes was then calculated. An average of the standard deviations for all target gene combinations was calculated and presented as stability (M).^{56, 57} The second method was based on the NormFinder algorithm using a Microsoft Excel Add-In (version 0.953) downloaded at http://moma.dk/normfinder-software.⁵⁷ This method ranks the genes with the smallest intra- and intergroup variation.

3. Results

3.1. Analysis of Reference Gene Expression after 6 h or 18 h Fluoride Treatment

LS8 cells were treated with NaF at 0, 1, 3 or 5 mM for 6 hours (Fig. 1) and at 0, 1 or 3 mM for 18 hours (Fig. 2). RNA was isolated at the respective time points and cDNA was prepared. PCR reactions were performed to identify consistency of expression of the candidate genes at multiple doses. The candidate genes tested were: B2m, Eef1a1, Gapdh, Hprt, and Tbp. Small variations in fold change were observed in the expression of B2m and Tbp, and a higher variation was observed in the expression of Eef1a1 at 6 hours (Fig. 1) and at 18 hours (Fig. 2) relative to untreated cells.

Fig. 1.

Expression of reference genes in LS8 cells after treatment with increasing doses of NaF. Cells were plated and treated with 0, 1, 3, or 5 mM NaF for 6 hours. RNA was isolated and cDNA was prepared. The expression of B2m (A), Tbp (B), Eef1a1 (C), Gapdh (D), and Hprt (E) was evaluated by qPCR. Expression stability was assessed using the geNorm (F) and Normfinder (G) methods, and the most stable genes have a lower M value. Each experiment was performed in triplicate. Fold change for each NaF concentration was calculated relative to untreated cells and error bars represent standard error of the mean (SEM). Eef1a1 had the largest variation in gene expression. C_q, quantification cycle.

Fig. 2.

Expression of reference genes in LS8 cells after treatment with increasing doses of NaF. Cells were plated and treated with 0, 1 or 3 mM NaF for 18 hours. RNA was isolated and cDNA was prepared. The expression of B2m (A), Tbp (B), Eef1a1 (C), Gapdh (D), and Hprt (E) was evaluated by qPCR. Expression stability was assessed using the geNorm (F) and NormFinder (G) methods, and the most stable genes have a lower M value (F). Each experiment was performed in triplicate. Fold change for each NaF concentration was calculated relative to the untreated cells and error bars represent standard error of the mean (SEM). Eef1a1 had the largest variation in gene expression. C_q, quantification cycle.

3.2. Analysis of Reference Gene Expression Stability

Multiple algorithms have been developed to assess stability of reference genes.^56–59 In this study, the geNorm and the Normfinder algorithms were used for analysis of gene stability.⁵⁶ In the 6 hour experiment, B2m and Tbp had the lowest standard deviation (0.03) and therefore the highest stability according to geNorm (Fig. 1F). Eef1a1 had the highest standard deviation (0.37) and the lowest stability (Fig. 1F). After 18 hour fluoride treatment, B2m and Tbp had the lowest standard deviation (0.12) according to the geNorm method (Fig. 2F) and Eef1a1 had the highest standard deviation (0.5). According to Normfinder, in the 6 hour experiment Hprt had the lowest standard deviation (0.06), and Eef1a1 the highest (0.41) (Figs. 1G). By the same method, in the 18 hour experiment, Gapdh had the lowest standard deviation (0.04), and Eef1a1 the highest (0.5) (Fig. 2G).

3.3. Effect of Reference Gene Choice on Quantification of Experimental Gene Expression

DNA-damage inducible transcript 3 (Ddit3, Chop, Gadd153) is a stress response gene that increases its expression with fluoride treatment.²⁸ In order to test whether utilization of different reference genes affects the quantification of target gene expression, qPCR reactions were performed for Ddit3 at 6 and 18 hours after exposure to NaF (Fig. 3). Ddit3 expression was calculated using both the ΔΔCT method³⁴ and the primer efficiency method.³⁵ Using both methods, NaF-treated Ddit3 expression values were calculated relative to Ddit3 expression in non-treated cells. As expected, normalization of Ddit3 using a geometric mean of the three most stable genes resulted in Ddit3 expression values (Fig. 3A,D) that were very different than when calculated relative to the reference gene Eef1a1 (Fig. 3B,E). Normalization of Ddit3 was also performed relative to B2m only, and the result was similar to normalization with three most stable genes (Figs. 3C, F). Results were similar when the efficiency method was used for analysis (Fig. 4). Similar to the ΔΔCt method, normalization to Eef1a1 induced an overestimation of Ddit3 levels in cells that were treated with NaF for 6 or 18 hours.

Fig. 3.

The choice of reference gene determines the calculated expression level of Ddit3 in LS8 cells. Cells were treated with NaF for 6 hours (A–C) or 18 hours (D–F). Ddit3 was normalized to the mean of B2m, Tbp, and Hprt at 6 hours (A) and 18 hours (D). Ddit3 was normalized to Eef1a1 at 6 hours (B) and 18 hours (E). Ddit3 was normalized to B2m at 6 hours (C) and 18 hours (F). Analysis was performed using the ΔΔCT method. Normalization with B2m produced results similar to normalization with a mean of B2m, Tbp, and Hprt, but these results were different from normalization using Eef1a1.

Fig. 4.

Expression of Ddit3 using the primer efficiency method. Cells were treated with NaF for 6 or 18 hours. Ddit3 expression was calculated relative to B2m at 6 hours (A) or 18 hours (C). Ddit3 expression was calculated relative to Eef1a1 at 6 hours (B) or 18 hours (D). The increase in expression of Ddit3 with fluoride treatment is significantly smaller when calculated relative to the more stable gene B2m relative to the less stable Eef1a1 gene.

3.4. Analysis of Reference Gene Expression In Vivo

To test reference gene expression in vivo, drinking water for 6-week-old male Sprague Dawley rats was supplemented with 0 or 100 ppm fluoride as NaF. After 6 weeks, rats were sacrificed and RNA was extracted from enamel organs. Maturation-stage rat incisor enamel organs were dissected and RNA was isolated. qPCR analysis of these samples showed the smallest variation in expression of Gapdh and Eef1a1 (0.18), and a higher variation for Hprt (0.27) (Fig. 5E) according to the geNorm method. By Normfinder, Gapdh had a variation of 0.06 and Eef1a1 had a variation of 0.13 (Fig. 5F).

Fig. 5.

Expression of reference genes in rat enamel organ. The smallest variation with fluoride treatment was observed in the expression of Eef1a1 and the largest was with Hprt. Rat drinking water was supplemented with 0 or 100 ppm fluoride as NaF for 6 weeks. RNA was isolated from maturation stage incisor enamel organs and cDNA was prepared. The expression of B2m (A), Gapdh (B), Hprt (C), and Eef1a1 (D) was evaluated by qPCR. Expression stability was assessed using the geNorm (E) and the Normfinder (F) methods. The most stable genes have a lower M value. Each experiment was performed in triplicate. C_q, quantification cycle. Fold change for NaF treatment was calculated relative to untreated rats and error bars represent standard error of the mean (SEM).

4 Discussion

Relative quantification of genes of interest in qPCR experiments depends on selection of appropriate reference genes. Optimal reference gene expression remains unchanged regardless of experimental conditions. Reference gene normalization greatly improves statistical significance, power and may dramatically reduce sample size. In this study, the goal was to evaluate stability of candidate reference gene expression upon treatment with NaF.

Ameloblast-derived LS8 cells were treated for 6 or 18 hours with NaF and the expression of five potential reference genes was tested: B2m, Eef1a1, Gapdh, Hprt and Tbp. The most stable genes in these experiments were B2m and Tbp by geNorm and Gapdh and Hprt by Normfinder. In order to test the effect of choosing an appropriate gene for normalization, the expression of Ddit3, an ER stress response gene shown to be upregulated by NaF, was assessed in these experiments relative to the B2m and Eef1a1 reference genes. Ddit3 expression levels relative to Eef1a1 were higher than if calculated relative to the more stably expressed B2m gene. Similar results were obtained when using the ΔΔCt or the efficiency methods of analysis. Data in this Ddit3 experiment was normalized to the most stable gene⁵⁶ or to the three most stable genes.⁶⁰ The results in both cases were comparable and therefore the benefit of using multiple reference genes must be considered. For example, a small number of targets of interest may only require one or two reference genes.

In a separate experiment, the expression of B2m, Eef1a1, Gapdh and Hprt was tested in samples from the enamel organs of 6-week-old male Sprague-Dawley rats treated with 100 ppm fluoride. Gapdh and Eef1a1 were the most stable genes in this experiment using both the geNorm and Normfinder algorithms. The differences in gene expression in vitro versus in vivo underscore the importance of testing genes under specific experimental conditions. In the above-described experiments the variability in gene expression of the more stable genes was relatively low. The differences in output of the geNorm and Normfinder algorithms reflect the different statistical approaches to data analysis.⁵⁷ Normfinder uses a mathematical model that describes gene expression values, analyzes the sample subgroups separately, estimates both the intra- and inter-group variation of gene expression, and calculates candidate gene stability. geNorm employs a pairwise comparison, and therefore selects genes with the highest degree of similarity in expression across the sample set. geNorm, however, tends to select genes that are correlated. The NormFinder algorithm on the other hand is less robust in experiments with a small sample size.⁵⁷ Nevertheless, according to both methods the least stable gene observed in vitro was Eef1a1, while the others had a standard deviation lower than the accepted threshold of 0.15.⁵⁶

The differences observed between the in vitro and in vivo experiments likely reflect the differences in the starting material. The cell population in the in vitro experiments was homogenous, and the treatment with fluoride was acute (6 hours or 18 hours). The enamel organ is composed of ameloblasts and connective tissue cells, and the contribution of each subpopulation may explain the differences in gene expression relative to the in vitro experiments. In addition, fluoride treatment in vivo was chronic (3 weeks of fluoride-supplemented drinking water) providing an additional source of variation by eliciting a more complex cellular response.

Given the risk of inducing reference gene expression instability when altering experimental conditions, it is recommended to validate a stable set of reference genes as an initial and an essential step in all qPCR experiments. A recent review presented the software packages that were developed for qPCR analyses.⁶¹

In conclusion, qPCR experiments must first validate the stability of reference gene expression in context of the specified conditions. This study identified stable reference genes for fluoride experiments using an ameloblasts-derived cell line (LS8) and by using rat enamel organ.

Highlights.

Identification of reference genes for fluoride qPCR experiments is proposed.

RNA was isolated from LS8 ameloblast-derived cells and rat enamel organs.

qPCR experiments must first validate the stability of reference gene expression in context of the specified conditions.

Gapdh and Eef1a1 were the most stably expressed genes in enamel organs from fluoride-treated rats.

Abbreviations

AAALAC

Association for Assessment and Accreditation of Laboratory Animal Care International

C_q

quantification cycle

IACUC

Institutional Animal Care Use Committee

qPCR

quantitative real-time PCR