Inflammation and Starvation Affect Housekeeping Gene Stability in Adipose Mesenchymal Stromal Cells

Abstract

Due to the scientific success of in vitro and in vivo model studies, the interest in using mesenchymal stromal cells (MSCs) for the treatment of orthopaedic conditions is growing. In the context of osteoarthritis (OA), MSCs, and, in particular, those derived from adipose tissues (ASCs), have found broader access to clinical use as active components of minimally manipulated orthobiologics, as well as clinically expanded cell preparations, or to collect their released factors (secretome) for cell-free approaches. In this regard, while both inflammatory priming and starvation are common strategies used to empower cell potency or collect the secretome, respectively, little is known about the possible influence of these approaches on the stability of housekeeping genes (HKGs) for molecular studies able to fingerprint cell phenotype or potency. In this report, the reliability of five commonly used HKGs (ACTB, B2M, GAPDH, HPRT1 and RPLP0) was tested in ASCs cultured under standard protocol after inflammatory priming or starvation. Gene expression data were computed with four different applets able to rank genes depending on their stability in either single or combined conditions. The obtained final ranking suggests that for each treatment, a specific HKG is needed, and that starvation is the condition with the stronger effect on HKGs' stability and, therefore, reliability. The normalization effect of proper HKGs' use was then validated on three genes involved in OA and whose product is released by ASCs. Overall, data presented herein confirm that the choice of the best HKG has to be carefully considered and that each specific condition has to be tested to identify the most reliable candidate.

Keywords: housekeeping genes; inflammation; mesenchymal stromal cells; osteoarthritis; regenerative medicine; secretome; starvation.

Figure 1

ASCs immunophenotype characterization. ASCs were positive for MSCs markers CD44, CD73, CD90 and CD105, and to tissue resident/early passage ASCs marker CD34. CD45 haematological marker was negative. Plots from a representative donor are shown with black lines for unstained samples and coloured lines for respective Ab-stained ones. Radar chart shows the mean of the three independent donors used in the study for each analysed marker. SD was always ≤10%.

Figure 2

Ct values of tested housekeeping genes across all ASCs samples. Box and whiskers plot is shown, with min to max for whiskers.

Figure 3

Principal component analysis (PCA) and heat map of HGK expression values. (A) In PCA, the X- and Y-axes show principal component 1 and principal component 2, which explain 87.2% and 8.6% of the total variance, respectively. (B) In the heat map, positive values mean higher Ct (lower amount), and negative values mean lower Ct (higher amount) with respect to mean values after row centring for each HGK. Both rows and columns were clustered using correlation distance and average linkage. No transformation and no scaling were applied for the dataset.

Figure 4

Gene expression modulation is dependent on HKG choice. (A–F) for IL1β, BMP6 and VEGFA, different amounts between treatments may significantly change using the best (B) or worst (W) HKG for each analysed couple. Values are expressed as fold-change mean ± SD, n = 3 independent cell isolates; § for p-value ≤ 0.1, * for p-value ≤ 0.05, ** for p-value ≤ 0.01, *** for p-value ≤ 0.001.

Figure 5

Comparison of gene expression between the five conditions analysed. Gene expression values for IL1β (A), BMP6 (B) and VEGFA (C) during the different treatments expressed as 2^−ΔCt with respect to the best HKG in Table 4, HPRT1. Values are expressed as fold-change mean ± SD, n = 3 independent cell isolates; ** for p-value ≤ 0.01, *** for p-value ≤ 0.001. For ease of readability, only statistically significant (p-value ≤ 0.05) comparisons are indicated. Where no bar is shown, the absence of significance is meant. For VEGFA, no comparisons reached statistical significance, making them, therefore, all “not significant”.