br Methods br Results The distribution of individual Ct

Methods
Results The distribution of individual Ct values and existing outliers is depicted in Fig. 1. Notably, we observed Ct values in the range of 16 to 29 (Table 2). Reference genes were minimally less expressed in control BMSCs compared to OA-BMSCs. In numbers, these differences in gene expression (delta Ct values, ∆Ct) range from ∆Ct=−0.010 (ubiquitin) to ∆Ct=−0.476 (polymerase (RNA) II (DNA directed) polypeptide A (POLR2A)). Except for POLR2A no significant differences in gene expression were observed in OA-BMSCs compared to control BMSCs (Fig. 1). The smallest differences in ∆Ct values were detected for ubiquitin (UBC) (−0.01±0.15), beta-actin (ACTB) (−0.03±0.23), and mitochondrially encoded ATP synthase 6 (MT-ATP6) (0.06±0.11). The standard deviations of mean differences of these genes were comparatively high in relation to the corresponding ∆Ct values (Table 2). Calculating the relative fold difference of independent experiments using the ∆∆Ct method (Pfaffl et al., 2004) tyrosine 3-/tryptophan 5-monooxygenase activation protein, zeta polypeptide (YWHAZ) showed the lowest gene expression level and was consequently considered as the relative control for the other 30 genes. MT-ATP6 was the candidate reference gene with the highest relative fold difference normalized to YWHAZ and thus with the highest expression (Fig. 2). Both, TATA box binding protein (TBP) and importin 8 (IPO8) were among the low expressed genes in relation to YWHAZ. The commonly used reference genes ACTB, GAPDH, and beta-2-microglobulin (B2M) were the second, third, and fifth highest expressed, YWHAZ-normalized genes. Using geNorm software package as statistical algorithm we identified IPO8 and TBP as the two most stable genes (M=0.021), followed by cancer susceptibility candidate 3 (CASC3) (M=0.022). In SYN-117 cost to this, the third and second least stable candidate reference genes were B2M and ACTB with stability measures of 0.035 and 0.044, respectively. These were followed by ribosomal protein L37a (RPL37A) gene the standard deviation of which was almost six times higher than its absolute delta Ct value resulting in the lowest gene stability (M=0.060) (Fig. 3A). By the use of NormFinder software – in line with the analysis based on geNorm algorithm – TBP, CASC3, and IPO8 were identified as the three most stable reference genes (Fig. 3B). To validate gene expression stability determined by geNorm and NormFinder the coefficient of variation (CV; percentage on Ct values) was calculated showing that TBP (CV=1.15) and IPO8 (CV=1.23) were the genes with the least expression variability corresponding to the highest stability. RPL37A (CV=4.65) demonstrated the least gene stability. Interestingly, the traditional reference genes ACTB and B2M were associated with the second and third highest CV indicating comparatively low expression stability in the setting investigated. In contrast GAPDH gene stability was middle-rate as compared to all other genes investigated. Table 3 summarizes the ranking of all potential reference genes tested with respect to their delta Ct and fold change expression values obtained by geNorm, NormFinder and CV method, respectively. Furthermore, geNorm was used to determine the minimal number of reference genes by pairwise variation required for an accurate normalization. As demonstrated in Fig. 4 including a third reference gene compared to two genes (V2/3 0.066) results in a small effect on normalization (below the cut-off value of 0.15). Therefore, the two most stable expressed reference genes may be sufficient for qRT-PCR normalization.
Discussion As evidence is revealing a crucial role of BMSCs in the context of OA disease development and progression (Murphy et al., 2002; Lamas et al., 2010) an increasing interest in investigations addressing functional alterations in OA-BMSC at the cellular and subcellular level is arising. To generate accurate and reliable gene expression data the use of suitable reference genes is indispensable (Bustin, 2002; Vandesompele et al., 2002; Dheda et al., 2004). To the authors\' best knowledge no prior study has addressed the validation of reference genes in primary human BMSCs derived from osteoarthritic as compared to healthy donors. Hence, in the present study we investigated the gene expression stability of n=31 potential reference genes in BMSCs isolated from n=13 patients with idiopathic advanced-stage hip osteoarthritis compared to BMSCs derived from n=15 non-diseased individuals.