Direct modulation of the bone marrow mesenchymal stromal cell compartment by azacitidine enhances healthy hematopoiesis

Abstract

Mesenchymal stromal cells (MSCs) are crucial components of the bone marrow (BM) microenvironment essential for regulating self-renewal, survival, and differentiation of hematopoietic stem/progenitor cells (HSPCs) in the stem cell niche. MSCs are functionally altered in myelodysplastic syndromes (MDS) and exhibit an altered methylome compared with MSCs from healthy controls, thus contributing to disease progression. To determine whether MSCs are amenable to epigenetic therapy and if this affects their function, we examined growth, differentiation, and HSPC-supporting capacity of ex vivo-expanded MSCs from MDS patients in comparison with age-matched healthy controls after direct treatment in vitro with the hypomethylating agent azacitidine (AZA). Strikingly, we find that AZA exerts a direct effect on healthy as well as MDS-derived MSCs such that they favor support of healthy over malignant clonal HSPC expansion in coculture experiments. RNA-sequencing analyses of MSCs identified stromal networks regulated by AZA. Notably, these comprise distinct molecular pathways crucial for HSPC support, foremost extracellular matrix molecules (including collagens) and interferon pathway components. Our study demonstrates that the hypomethylating agent AZA exerts its antileukemic activity in part through a direct effect on the HSPC-supporting BM niche and provides proof of concept for the therapeutic potential of epigenetic treatment of diseased MSCs. In addition, our comprehensive data set of AZA-sensitive gene networks represents a valuable framework to guide future development of targeted epigenetic niche therapy in myeloid malignancies such as MDS and acute myeloid leukemia.

Graphical abstract

Figure 1.

Effects of AZA treatment on stromal support of healthy and MDS HSPCs. (A) Primary healthy or MDS CD34⁺ cells were cultured in serum-free medium in suspension or on EL08-1D2 stromal cells. After 4 days, CD34⁺ cells were harvested and viability was assessed by Annexin V/PI flow cytometry. (B) Primary healthy or MDS CD34⁺ cells were cultured on untreated EL08-1D2 stroma or on stroma treated with 10 µM AZA for 48 hours. After 4 days, CD34⁺ cells were harvested and viability was assessed by Annexin V/PI flow cytometry. Shown is the percentage of viable (Annexin V−/PI−) CD34⁺ cells for each culture condition; each circle represents 1 individual sample (healthy samples, n = 10; MDS samples, n = 10). n.s., not significant.

Figure 2.

AZA treatment of primary MSCs does not alter their viability, morphology, or immunophenotype. Healthy and MDS MSCs were grown to 80% confluency and treated once on day 0 with 10 µM AZA. After 48 hours, stromal layers were washed and fresh medium was added. Untreated MSCs were used as controls. (A) MSC viability was assessed by flow cytometry using Annexin V/PI on day 6. Shown are mean frequencies ± SEM of Annexin V−/PI− cells measured in triplicate for n = 3 healthy and n = 4 MDS MSCs. (B) Cell counts of MSCs were determined by trypan-blue staining at indicated time points. Values represent mean ± SEM of n = 3 healthy and n = 4 MDS MSC cell counts normalized to the cell count before treatment on day 0. (*P = .045 for MDS MSC AZA vs untreated on day 4, all other n.s.) (C) Representative light microscopy images showing MSC morphology at indicated conditions (magnification ×10). (D) Expression of MSC-defining markers was assessed by flow cytometry on day 6.

Figure 3.

AZA treatment of healthy as well as MDS MSCs favors support of healthy HSPCs. (A) Experimental design. Healthy MSCs or MDS MSCs were treated once with 10 µM AZA for 48 hours. MSC layers were washed and healthy or MDS CD34⁺ HSPCs were added. After 4 days, CD34⁺ HSPCs were harvested and progenitor cell activity was assessed by short-term CFU assays. (B) Mean CFU frequencies ± SEM (n = 11 healthy CD34⁺ cells; n = 12 MDS CD34⁺ cells; n = 7 healthy MSCs layers). (C) Mean CFU frequencies ± SEM of n = 11 healthy CD34⁺ cells and n = 12 MDS CD34⁺ cells after culture on n = 7 healthy MSCs were compared with mean CFU frequencies ± SEM of n = 9 healthy CD34⁺ cells and n = 11 MDS CD34⁺ cells after culture on n = 6 MDS MSCs. (D) Mean CFU frequencies ± SEM (n = 9 healthy CD34⁺ cells /n = 6 MDS MSC layers; n = 11 MDS CD34⁺ cells /n = 6 MDS MSC layers). *P < .05; **P < .01; ***P < .001.

Figure 4.

Osteogenic and adipogenic differentiation of MSCs is modulated by AZA in healthy MSCs. (A) Experimental design. Healthy and MDS MSCs were treated with 10 µM AZA on days 0 and 7. Untreated MSCs were used as controls. MSCs were differentiated for 18 days and oil-red-o staining plus Harris-hematoxylin or alizarin-red staining were used to quantify adipogenesis (C) or osteogenesis (B), respectively. Differentiation was ranked according to absent (=0), weak (=1), moderate (=2), or intense (=3) staining. (B-C) Representative images (top panel) and statistical summary (bottom panel) for (B) mean osteogenic potential of n = 5 healthy and n = 5 MDS MSCs ± SEM and (C) mean adipogenic potential of n = 10 healthy and n = 7 MDS MSCs ± SEM. *P < .05; **P < .01; ***P < .001.

Figure 5.

AZA treatment induces a distinct gene-expression pattern in stromal cells. (A-C) RNA-seq analysis of EL08-1D2 cells after 4 days of culture with or without 10 µM AZA added once on day 0. (A) The term “AZA-sensitive genes” comprises all genes that were differentially expressed between AZA-treated and untreated samples according to DESeq2 analysis. The pie diagram shows the percentage of upregulated vs downregulated differentially expressed genes in the AZA-treated cohort compared with untreated cells. (B) Hierarchical clustering and gene-expression heat map across the 500 most differentially expressed genes between untreated and AZA-treated samples. Values are plotted relative to the average of untreated samples. (C) Overrepresentation of significantly differentially expressed genes in KEGG pathways and GO categories were tested by FNEA implemented in the neaGUI R package. Entries are ordered first according to false discovery rate (FDR) and second according to the absolute number of the z score. (D-E) Primary healthy and MDS MSCs were treated with AZA once or left untreated and analyzed after 4 days. (D) TGF-β1 expression was validated by comparative RT-PCR in primary MSCs. Results are shown as fold change compared with untreated control. (E) Protein expression of TGF-β1 and TGF-β receptor 1 as assessed by flow cytometry in healthy or MDS MSCs (−/+) AZA. Sample numbers (corresponding to Table 1) of samples analyzed both by RT-PCR in panel D as well as flow cytometry in panel E are indicated. *P < .05; **P < .01; ***P < .001.

Figure 6.

Integrative analysis reveals AZA-sensitive genes crucial to stromal HSPC support. (A) Gene set enrichment analysis of our RNA-seq data showing significant upregulation of stromal HSPC-supportive genes (published reference gene set) in AZA-treated EL08-1D2 cells. (B) We identified 4 distinct gene coexpression modules where each module was given an arbitrary color (blue, IFN-signaling pathway; brown, ECM receptor interaction; and turquoise, WNT- and PDGF-signaling pathways). Both the blue and the brown modules significantly correlated with AZA treatment (Pearson correlation, P = 2.1e-09 [blue] and 3.4e-08 [brown], respectively) and contained significantly more genes that were significantly regulated by AZA compared with untreated controls (Fisher's exact test, P = 4.22e-10 [blue] and 0.0184 [brown]). ***P < .001. (C) Detailed network of the ECM receptor interaction (brown) module showing significantly enriched biological pathways and their genes. Colors indicate log2-fold change in gene expression in AZA-treated samples vs untreated controls. Hub genes are shown in bold letters. (D) Venn diagram of meta-analysis showing overlap of 21 genes (in red) between AZA-sensitive downregulated EL08-1D2 genes (green) and upregulated genes in MDS MSCs in comparison with healthy controls (blue). The 7 identified collagen genes are listed at the bottom of the panel. NES, normalized enrichment score.

Figure 7.

Transcriptional regulation of HSPC-supportive stromal genes in primary MSCs treated with AZA. Primary healthy and MDS MSCs were treated with AZA (10 µM) once or left untreated. After 4 days, MSCs were harvested and expression of selected stromal HSPC-supportive genes was validated by comparative RT-PCR. (A) Selected hub genes of the brown module (ECM receptor interaction, from Figure 6C). (B) Seven collagens differentially expressed between untreated healthy and MDS MSCs (from Figure 6D). (C) Other stromal HSPC-supportive genes differentially expressed between untreated healthy and MDS MSCs (from Figure 6D, in red). *P < .05; **P < .01; ***P < .001; ****P < .0001.

Figure 7.

Transcriptional regulation of HSPC-supportive stromal genes in primary MSCs treated with AZA. Primary healthy and MDS MSCs were treated with AZA (10 µM) once or left untreated. After 4 days, MSCs were harvested and expression of selected stromal HSPC-supportive genes was validated by comparative RT-PCR. (A) Selected hub genes of the brown module (ECM receptor interaction, from Figure 6C). (B) Seven collagens differentially expressed between untreated healthy and MDS MSCs (from Figure 6D). (C) Other stromal HSPC-supportive genes differentially expressed between untreated healthy and MDS MSCs (from Figure 6D, in red). *P < .05; **P < .01; ***P < .001; ****P < .0001.