Atrial myxomas arise from multipotent cardiac stem cells

Figures

• 

• 

  Figure 1. c-kit expression identifies putative multipotent resident cardiac stem cells within myxoma. (A) Representative confocal microscopy image of c-kitpos cells within the myxoid matrix of a myxoma tumour (c-kit/green fluorescence, chondroitin sulfate/red, nuclei/DAPI/blue). Scale bar = 50 μm. (B) Percentage of total c-kitpos cardiac cells and c-kitpos/CD45pos/CD31pos cardiac cells in normal control (right atria) tissue and in myxoma tumour tissue. (C) Representative confocal microscopy image showing that most of the myxoma c-kitpos cells were also CD45pos/Tryptasepos (c-kit/green, CD45/red, Tryptase/white, nuclei/blue) or (D) CD31pos (c-kit/green, CD31/red, nuclei/DAPI/blue). Scale bar = 50 μm. (E) Representative confocal microscopy image show that despite the majority of c-kitpos cells are CD45pos/CD31pos, few of the c-kitpos cells are CD45neg/CD31neg (white arrow) in myxoma tissue. Scale bar = 50 μm. (F–H) Representative confocal microscopy images of c-kitpos myxoma cells dispersed within myxoid matrix (chondroitin sulfate/red), which express stemness and cardiac progenitor cell transcription factors and respectively Oct-4 (white), Nkx2.5 (white), and Isl-1 (white). Scale bar = 50…

• 

  Figure 2. Identification, expansion/clonogenic potential and phenotipic characterization of myxoma-derived c-kitpos/CD45neg/CD31neg cells in vitro. (A) Flow cytometry dot plots showing the percentage of total c-kitpos, c-kitpos/CD31pos, and c-kitpos/CD45pos cardiac cells population within myxoma cells isolated from myxoma tissue (representative of n = 6 experiments). (B) Representative light microscopy images of a c-kitpos/CD45neg/CD31neg hCSC clone (right) obtained depositing a single c-kitpos/CD45neg/CD31neg hCSC (left) into 96-well Terasaki plates. Scale bar = 400 μm. (C) Graph showing the clonogenic capacity of myxoma-derived c-kitpos/CD45neg/CD31negCSCs compared to normal hCSCs at P8 (representative of n = 6 biological replicates). (D) Cellular growth curve in vitro shows the growth deficit of c-kitpos/CD45neg/CD31neg myxoma-derived CSCs compared to c-kitpos/CD45neg/CD31neg hCSCs at P12. P < 0.05 vs. c-kitpos/CD45neg/CD31neg hCSCs (representative of n = 6 biological replicates). (E) Flow cytometry dot plots showing the phenotypic characterization of c-kitpos/CD45neg/CD31neg myxoma-derived ‘putative' CSCs. Data are presented as mean ± SD.

• 

  Figure 3. c-kitpos/CD45neg/CD31neg myxoma-derived ‘putative' CSCs show multilineage differentiation and myxoid matrix production potential in vitro. (A) Representative confocal microscopy images obtained from cytospin preparations of c-kitpos/CD45neg/CD31neg myxoma-derived CSCs revealed the expression of stemness and cardiac progenitor markers. c-kit (green), Oct-4, Nanog, Bmi-1, Nkx2.5, and Isl-1 (red) (representative of n = 6 biological replicates). Scale bar = 50 μm. (B) Representative light microscopy image of primary (top panel) and tertiary cardiospheres (bottom panel) derived from c-kitpos/CD45neg/CD31neg myxoma CSCs. The number of tertiary cardiospheres is reduced in c-kitpos/CD45neg/CD31neg myxoma-derived CSCs (representative of n = 6 biological replicates). Scale bar = 100 μm. (C) qRT-PCR data (upper panel) showing the expression levels of the main endothelial differentiation genes (CD31and vWf) in normal and myxoma-derived c-kitpos/CD45neg/CD31neg CSCs at baseline (0 day) or after 14 days in endothelial differentiation media and representative confocal microscopy images (bottom panel) showing the expression of vWF in c-kitpos/CD45neg/CD31neg myxoma-derived CSCs at 14 days after…

• 

  Figure 4. Myxoma-derived CSCs show significant and specific differences of their transcriptome when compared with control human CSCs. (A) Heatmaps showing the expression values, as log2 of normalized read-counts (left) and fold-change (FC, right) of RNAs by RNASeq analysis of hCSCs and myxoma-derived (Myx)CSCs. Only RNAs with |FC|≥3 and FDR ≤0.05 are represented. (B) Heatmaps showing the expression values (left) and FC of RNAs involved in selected biological functions (P-value ≤ 0.05) as computed by IPA in the comparison of hCSCs vs. MyxCSCs. (C) (Upper panel) Histogram showing the canonical pathways statistically significant (P-value ≤ 0.05), as defined by IPA, involving RNAs differentially expressed in the comparison of hCSCs vs. MyxCSCs. (Bottom panel) Histogram showing the biological functions statistically significant, as defined by IPA, involving RNAs differentially expressed in the comparison of hCSCs vs. MyxCSCs.

• 

  Figure 5. MicroRNA profile of myxoma-derived CSCs. (A) Heatmaps showing the expression values, expressed as log2 of normalized read-counts (left) and fold-change (right) of miRNAs by RNASeq analysis of hCSCs vs. MyxCSCs. Only miRNAs with |FC| ≥ 2 and FDR ≤0.05 are represented. (B) Boxplots showing the log2 normalized expression read counts of selected miRNAs in hCSCs and MyxCSCs samples. (C) miRNAs-mRNAs targets networks. In green are shown down-regulated miRNAs and respective regulated target mRNAs, while in red are shown up-regulated miRNAs and respective modulated target mRNAs. (D) Histograms showing the pathways where mRNA targets of selected miRNAs are involved (left) and the frequency of miRNAs down-regulated/mRNAs target up-regulated involved in that specific pathway in myxCSCs. (E) Network representation of specific biological functions where selected miRNAs/mRNA targets are involved in myxCSCs.

• 

  Figure 6. miR-126-3p and miR-335-5p dictate, respectively, the expansion/clonogenic deficit and the abortive myogenic differentiation potential of clonogenic myxoma-derived CSCs in vitro. (A) Cellular growth curve in vitro shows the effects of the different modulated (up-regulated by the relative miR mimic or suppressed by the relative miR inhibitor) microRNAs on myxoma-derived CSCs at P12. (B) Cellular growth curve in vitro shows that miR-126-3p suppression by Lenti-miR-126-3p inhibitor-GFP infection increases myxoma-derived CSC proliferation in vitro as compared to control (GFP) lentiviral infection at P12. (C) Graph showing the effects of the different indicated microRNA modulations on the clonogenic capacity of myxoma-derivedCSCs at P12. (D) Heat maps showing qRT-PCR analysis of GATA-4, NKX2.5, MEF2C, MYH6, MYH7, ACTC1 and TNNT2 gene transcripts in differentiating myxoma-derived CSCs infected with the indicated Lenti-MiR or control Lenti-GFP after 14 days in cardiomyogenic media in vitro. Colour scale indicates change in Ct (threshold cycle) relative to the normalized GAPDH control. (E) qRT-PCR data showing the expression levels of the indicated cardiac genes in myxoma-derived CSCs…

• 

  Figure 7. c-kitpos/CD45neg/CD31neg myxoma-derived CSCs seed myxoma tumours in vivo. (A) Schematic representation of the design of the in vivo myxoma tumours seeding study in NOD-SCID immunodeficient mice transplanted with myxoma-derived CSCs (from three different myxoma) at different doses. Human CSCs (hCSCs) isolated from normal cardiac tissues (n = 3) were used as control (control hCSCs). In brief, after tissue dissociation, isolated cardiac cells were depleted of CD45pos and CD31pos cells then were enriched for c-kitpos cardiac cells through MACS immunolabelling. Cells were kept in culture, cloned and engineered to express green fluorescent protein (GFP) by relative lentiviral transduction. GFPpos myxoma-derived CSCs and GFPpos normal human CSCs were transplanted into the quadriceps muscle of NOD-SCID immunodeficient mice at three different cell doses. (B) Representative Alcian Blue-PAS staining of the typical myxoma matrix with scattered cells formed by myxoma CSC transplantation within mouse skeletal muscle. Scale bar = 75 μm. (C) GFPpos myxoma-derived CSCs and GFPpos control hCSCs were transplanted in NOD-SCID immunodeficient mice at three different concentration. Myxoma tumour…