Validation of Housekeeping Genes to Study Human Gingival Stem Cells and Their In Vitro Osteogenic Differentiation Using Real-Time RT-qPCR

Abstract

Gingival stem cells (GSCs) are recently isolated multipotent cells. Their osteogenic capacity has been validated in vitro and may be transferred to human cell therapy for maxillary large bone defects, as they share a neural crest cell origin with jaw bone cells. RT-qPCR is a widely used technique to study gene expression and may help us to follow osteoblast differentiation of GSCs. For accurate results, the choice of reliable housekeeping genes (HKGs) is crucial. The aim of this study was to select the most reliable HKGs for GSCs study and their osteogenic differentiation (dGSCs). The analysis was performed with ten selected HKGs using four algorithms: ΔCt comparative method, GeNorm, BestKeeper, and NormFinder. This study demonstrated that three HKGs, SDHA, ACTB, and B2M, were the most stable to study GSC, whereas TBP, SDHA, and ALAS1 were the most reliable to study dGSCs. The comparison to stem cells of mesenchymal origin (ASCs) showed that SDHA/HPRT1 were the most appropriate for ASCs study. The choice of suitable HKGs for GSCs is important as it gave access to an accurate analysis of osteogenic differentiation. It will allow further study of this interesting stem cells source for future human therapy.

Figure 1

(a), (b), (c), and (d) Alizarin Red S Staining of GSC and ASC after 21 days of osteoblast induction. (a) and (b) GSC and ASC were cultured in proliferation medium. No differentiation is noticed. (c) and (d) GSC and ASC were cultured in osteogenic medium. Calcium mineral deposits confirmed osteoblast differentiation in both stem cells after 21 days of culture. Bar scale = 100 μm. (e) Colony forming units for GSCs. Limiting dilution of GSCs shows their ability to form colonies after 5 days of culture. (f) Agarose gel electrophoresis: total RNA quality. All RNA samples showed absence of degradation and a high degree of integrity. Upper bands: 28S, lower bands: 18S, and lanes 1 to 4: random RNA samples from GSC and ASC.

Figure 2

PCR reaction values and expression levels of the 10 candidate reference genes. (a) Melt peak temperature was determined for each reference gene to confirm the specificity of each primer. (b) Threshold values were manually set for each reference gene to calculate cycle threshold (Ct) values. (c), (d), and (e) Expression levels (Ct values) were determined for all reference genes throughout all the samples: (c) GSCs (n = 8), (d) dGSCs (n = 8), and (e) ASCs (n = 6). The central bars correspond to the mean Ct values; the upper and lower bars represent the standard deviation. These values were used in ΔCt and BestKeeper algorithms.

Figure 3

Ranking of 10 HKG in GSC (a), dGSC (b), and ASC (c) with ΔCt comparative method. ΔCt variability with pairwise comparisons of the complete set of candidate housekeeping genes shown as boxes and whiskers: medians (lines), 25th and 75th percentiles (Boxes). Genes with lowest variability were the most stable.

Figure 4

Gene expression stability and minimal number of genes needed in (a) GSC (n = 8), (b) dGSC (n = 8), and (c) ASC (n = 6), by GeNorm. ((a1), (a2), and (a3)) Average expression stability values (M) for GSC, dGSC, and ASC. A lower M value indicated a more stable expression. ((b1), (b2), and (b3)) Pairwise variation value below 0.16 with the least number of reference candidates used is considered optimal for GSC, dGSC, and ASC.

Figure 5

Ranking of the most stable HKGs with BestKeeper analysis for GSC (a), dGSC (b), and ASC (c). Bars represented the coefficient of correlation r, while error bars represented the standard deviation (SD). Ideal reference gene had low SD and a high r. ^∗ p ≤ 0.05.

Figure 6

Determination of the most stable reference genes with NormFinder For GSC versus dGSC (a) and GSC versus ASC (b). Bars represent intergroup variances, while error bars represented the average of intragroup variances. Ideal reference gene had intergroup variation as close to zero as possible and error bars as small as possible.

Figure 7

Effect of the choice of stable HKGs in the study of four samples of GSC. The relative fold expression (a) and the normalized fold expression of COLL1A1 of 4 samples of GSC. The normalization was performed with SDHA/B2M/ACTB, the most stable genes (b), GAPDH, the most usually used (c), and ALAS1/RPII, the least stable genes (d). Error bars expressed the standard deviation of the mean.

Figure 8

Effect of the choice of stable HKGs in the study of dGSC (n = 6) (a) and ASC (n = 6) (b). The normalization of fold expression of an osteogenic marker, RUNX2, was performed on D0, D7, and D14. As calibrators: for dGSCs: TBP/SDHA the most stable HKGs and RPS18 the least stable one. For ASCs: SDHA/HPRT1 the most stable HKGs and ACTB the least stable one. p values have been obtained with unpaired t-test. ^∗∗∗∗ p ≤ 0.0001, ^∗∗∗ p ≤ 0.001, and ^∗∗ p ≤ 0.01 and “ns” stands for p ≥ 0.05.