Epigenetic rejuvenation of mesenchymal stromal cells derived from induced pluripotent stem cells

Abstract

Standardization of mesenchymal stromal cells (MSCs) remains a major obstacle in regenerative medicine. Starting material and culture expansion affect cell preparations and render comparison between studies difficult. In contrast, induced pluripotent stem cells (iPSCs) assimilate toward a ground state and may therefore give rise to more standardized cell preparations. We reprogrammed MSCs into iPSCs, which were subsequently redifferentiated toward MSCs. These iPS-MSCs revealed similar morphology, immunophenotype, in vitro differentiation potential, and gene expression profiles as primary MSCs. However, iPS-MSCs were impaired in suppressing T cell proliferation. DNA methylation (DNAm) profiles of iPSCs maintained donor-specific characteristics, whereas tissue-specific, senescence-associated, and age-related DNAm patterns were erased during reprogramming. iPS-MSCs reacquired senescence-associated DNAm during culture expansion, but they remained rejuvenated with regard to age-related DNAm. Overall, iPS-MSCs are similar to MSCs, but they reveal incomplete reacquisition of immunomodulatory function and MSC-specific DNAm patterns-particularly of DNAm patterns associated with tissue type and aging.

Graphical abstract

Figure 1

Generation of iPS-MSCs (A) Phase contrast images of MSCs, iPSCs, and iPS-MSCs in the course of differentiation either with or without EB formation. Thirty-five days after induction of differentiation, iPS-MSCs revealed similar fibroblastoid morphology as MSCs. (B) Population doublings (PDs) of MSCs and iPS-MSCs within 6 days of culture on gelatin-coated plates (N = 3; n = 3; mean ± SD; ^∗∗∗p < 0.001). (C) iPS-MSCs displayed similar immunophenotypic characteristics as primary MSCs (autofluorescence is indicated in gray). (D) MSCs and iPS-MSCs were differentiated toward adipogenic, osteogenic, or chondrogenic lineages for three weeks and subsequently stained with BODIPY/DAPI, alizarin red, or Alcian blue/PAS, respectively. Controls were simultaneously cultured in normal growth medium, and representative images are presented. (E) In vitro differentiation potential was further assessed by quantitative real-time PCR of adipogenic (ADIPOQ, FABP4), osteogenic (RUNX2, SP7, COL1A1, SPARC), and chondrogenic (SOX9, ACAN, COL2A1) marker genes in MSCs (green) and iPS-MSCs (blue; N = 3; n = 2; mean ± SD; ^∗p < 0.05; ^∗∗p < 0.01 versus nondifferentiated control). See also Figure S1.

Figure 2

Gene Expression Profiles of iPS-MSCs Are Similar to Primary MSCs (A) Hierarchical clustering revealed close relationship of iPS-MSCs and primary MSCs. MSC donor number (“M”) and clone number (“C”) are indicated for iPSCs and iPS-MSCs. Furthermore, passage numbers (“P”) are provided for MSCs and time of redifferentiation (“d”) for iPS-MSCs. (B) Heatmap of pairwise correlation coefficients (R²) demonstrates relationship of iPS-MSCs and MSCs. (C) Pluripotency was assessed by PluriTest analysis (Müller et al., 2011). After differentiation for more than 7 days toward iPS-MSCs, cells were clearly associated with nonpluripotent samples (blue area) and not with pluripotent samples (red area; labeling of samples as in A). (D) MSC marker genes were expressed at similar level in primary MSCs and iPS-MSCs. (E) Number of differentially expressed genes between MSCs, iPSCs, and iPS-MSCs (>2-fold regulation; adjusted p value <0.01; for each cell type, the number of upregulated genes is indicated by color code). (F) Gene ontology analysis of genes that are differentially expressed between MSCs and iPS-MSCs. The most significant categories are depicted. (G) Activity of iPS-MSCs and MSCs on proliferation of stimulated CD4⁺ T cells was assessed by flow cytometry and carboxyfluorescein succinimidyl ester (CFSE) staining. Different T cell:MSC ratios were used and representative histograms are depicted (unstimulated control is indicated in light gray). The percentage of proliferating cells is indicated in each histogram. (H) Quantitative analysis of T cell proliferation assay was performed with percentage of proliferated cells as shown in (G) (MSCs: N = 3; iPS-MSCs: N = 2; mean ± SD; ^∗p < 0.05; ^∗∗p < 0.01). See also Figure S2.

Figure 3

DNAm Profiles of iPS-MSCs (A) Hierarchical clustering of global DNAm profiles. (B) Number of CpGs with differential DNAm between MSCs, iPSCs, and iPS-MSCs (>20% change in DNAm level; adjusted p value <0.01; for each cell type hypermethylated CpGs are indicated by color code). (C) DNAm levels (β values) of CpGs represented in the genes POU5F1 (OCT3/4), NANOG, NT5E (CD73), and ENG (CD105) (TSS1500: 1,500 bp upstream of transcription start site; TSS200: 200 bp upstream of TSS; UTR). (D) Enrichment of differential DNAm of MSCs versus iPS-MSCs in gene regions or in relation to CpG islands (p values were estimated by hypergeometric distribution). See also Figure S3.

Figure 4

Donor-, Tissue-, and Age-Specific DNAm Changes (A) Hierarchical cluster analysis of 1,091 CpGs with highest donor-specific variation in primary MSC preparations (SD > 0.2) (Shao et al., 2013) revealed that iPSCs and iPS-MSCs clustered with their parental cell preparations. This indicates that interindividual DNAm patterns are maintained in iPS-MSCs (^∗cultivated in mTeSR1). (B) Hierarchical cluster analysis of 1,711 CpGs with differential DNAm in MSCs from adipose tissue (AT) and bone marrow (BM; >15% difference in mean methylation level) (Schellenberg et al., 2011) demonstrated that the BM-associated DNAm pattern is erased in iPSCs and not reestablished in iPS-MSCs. (C) The state of cellular senescence was estimated by pyrosequencing analysis of six senescence-associated CpGs (Koch et al., 2012). Predictions of this Epigenetic-Senescence-Signature for cumulative population doublings (cPD) were reversed upon reprogramming into iPSCs and increased again during differentiation toward iPS-MSCs. (D) To estimate the state of cellular senescence in iPS-MSCs we analyzed the frequency of fibroblastoid colony forming units (CFU-f). CFU-f frequency declines continuously in primary BM-MSCs and AT-MSCs (Schellenberg et al., 2012) and the number of CFU-f in iPS-MSCs after 35 days is in line with culture expansion for five passages. (E) Donor age of cell preparations was estimated using a multivariate model based on DNAm of 99 age-related CpGs of blood (Weidner et al., 2014). (F) Alternatively, donor age was predicted using a recently published predictor applicable for different tissues (Horvath, 2013). Overall, epigenetic rejuvenation upon reprogramming into iPSCs is also maintained in iPS-MSCs. See also Figure S4.