Neonatal and adult cardiac fibroblasts exhibit inherent differences in cardiac regenerative capacity

Abstract

Directly reprogramming fibroblasts into cardiomyocytes improves cardiac function in the infarcted heart. However, the low efficacy of this approach hinders clinical applications. Unlike the adult mammalian heart, the neonatal heart has an intrinsic regenerative capacity. Consequently, we hypothesized that birth imposes fundamental changes in cardiac fibroblasts which limit their regenerative capabilities. In support, we found that reprogramming efficacy in vitro was markedly lower with fibroblasts derived from adult mice versus those derived from neonatal mice. Notably, fibroblasts derived from adult mice expressed significantly higher levels of pro-angiogenic genes. Moreover, under conditions that promote angiogenesis, only fibroblasts derived from adult mice differentiated into tube-like structures. Targeted knockdown screening studies suggested a possible role for the transcription factor Epas1. Epas1 expression was higher in fibroblasts derived from adult mice, and Epas1 knockdown improved reprogramming efficacy in cultured adult cardiac fibroblasts. Promoter activity assays indicated that Epas1 functions as both a transcription repressor and an activator, inhibiting cardiomyocyte genes while activating angiogenic genes. Finally, the addition of an Epas1 targeting siRNA to the reprogramming cocktail markedly improved reprogramming efficacy in vivo with both the number of reprogramming events and cardiac function being markedly improved. Collectively, our results highlight differences between neonatal and adult cardiac fibroblasts and the dual transcriptional activities of Epas1 related to reprogramming efficacy.

Keywords: aging; fibroblast; microRNA (miRNA); myocardial infarction; reprogramming.

Figure 1

Reduced fibroblast to cardiomyocyte reprogramming efficacy in adult cardiac fibroblasts.A, fibroblasts were isolated from neonatal and adult cardiac tissue. After one passage, cells were transfected with either the non-targeting miRNA negmiR or the reprogramming miRNA cocktail miR combo. After 14 days, expression of the indicated cardiomyocyte-specific markers was determined by qPCR. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows the expression values (as Z-scores) of four independent experiments. Raw expression data can be found in Table S1. To determine significance, one-way ANOVA was employed on the raw expression data with Bonferroni post hoc tests to determine significance between groups. Results from one-way ANOVA and Bonferroni post hoc tests are reported in Table S2. Genes for which significant reprogramming (p < 0.05) was observed in neonatal cardiac fibroblasts are shown in bold. B, the expression of the indicated genes in adult and neonatal cardiac fibroblasts (passage 1) was determined by qPCR. Data are represented as a fold change in expression between adult versus neonatal cardiac fibroblasts. N = 3. To determine significance, one-way ANOVA was employed, F (6, 14) = 40.91 p < 0.0001, with Bonferroni post hoc tests to determine significances between groups (∗∗p < 0.01, ∗∗∗p < 0.001). C, adult cardiac fibroblasts were transfected with a control siRNA or a siRNA targeting Cebpd, Epas1, Mmp1, Mmp8, Prrx1, or Sox9. After 3 days, cells were analyzed for the expression of the indicated genes. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows the expression values (as Z-scores) of three independent experiments. Raw expression data can be found in Table S1. To determine significance, two-way independent T-tests were conducted. All genes were successfully targeted (p < 0.05) by their respective siRNAs. D, fibroblasts were isolated from adult cardiac tissue. After one passage, cells were transfected with the indicated combination of miRNAs (negmiR or miR combo) and siRNAs (non-targeting control or a siRNA targeting one of the following: Prx1, Cebpd, Epas1, Sox9, Mmp1b, or Mmp8). After 14 days, expression of the indicated cardiomyocyte-specific markers was determined by qPCR. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows the individual expression values (as Z-scores) of four independent experiments. Raw expression data can be found in Table S1. To determine significance, one-way ANOVAs were performed on raw expression values for each gene individually with Bonferroni post hoc tests to determine the significance between groups. ANOVA and Bonferroni results can be found in Table S2. Genes marked in bold are those for which Epas1 knockdown significantly (p < 0.05) increased expression.

Figure 2

Epas1 is expressed in adult cardiac fibroblasts and inhibits fibroblast to cardiomyocyte reprogramming.A, adult cardiac fibroblasts (1 passage after isolation) were transfected with either a control non-targeting siRNA or a siRNA targeting Epas1. After 3 days, RNA was extracted and analyzed for Epas1 mRNA levels by qPCR. Epas1 expression levels were normalized to the housekeeping gene Gapdh and are shown as a fold change. N = 3. ∗∗∗p < 0.001 (two-way independent t test). B, fibroblasts were isolated from hearts of adult fibroblast lineage tracing mice, Fsp1Cre:tdTomato. In these mice, fibroblasts express the fluorescent protein tdTomato. After one passage, cells were transfected with the indicated combination of miRNAs (negmiR or miR combo) and siRNAs (non-targeting control or a siRNA targeting one of the following: Prx1, Cebpd, Epas1, Sox9, Mmp1b, or Mmp8). After 14 days, cells were fixed and stained with antibodies targeting tdTomato (red) and the cardiomyocyte marker Tnni3 (green). Nuclei were counterstained with DAPI (blue). Representative images show cardiomyocytes derived from fibroblasts and quantification of the number of cardiomyocytes with observable sarcomeres provided. N = 5. ∗p < 0.05 (two-way independent t test between the indicated groups). Scale bar 100 micron. The arrow denotes a positive cell (TdTomato+ Tnni3+). C, cardiac fibroblasts (passage 1) derived from adult mice were transfected with siRNAs (control non-targeting or Epas1 targeting) and miRNAs (non-targeting control negmiR or the reprogramming cocktail miR combo). After 14 days, cells were fixed and stained with an Actn2 antibody (red). Nuclei were counterstained with DAPI (blue). Wide-field representative images are shown (scale bar 50 microns). The number of Actn2+ cells was counted and expressed as a percentage of the total number of cells. N = 4. One-way ANOVA, F (3, 12) = 38.34 p < 0.0001, with Bonferroni post hoc testing was used to determine the significance between groups. Significance to the control group (negmiR + control siRNA) is shown: ns-not significant, ∗∗∗p < 0.001. D, cardiac fibroblasts (passage 1) derived from adult mice were transfected with either a control non-targeting siRNA or a siRNA targeting Epas1. After 14 days, the expression of the indicated genes was determined by qPCR. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows individual expression values (as Z-scores) for three independent experiments. A two-way independent t test was used on the dataset and found no significant difference between the two groups for any gene tested (p > 0.05). E, cardiac fibroblasts (passage 1) derived from adult mice were transfected with siRNAs (control non-targeting or Epas1 targeting) and miRNAs (non-targeting control negmiR or the reprogramming cocktail miR combo). After 14 days, spontaneous calcium oscillations were determined. Representative traces are shown. The number of oscillations per second (Hz) was calculated from five individual traces. One-way ANOVA, F (3, 16) = 42.53 p < 0.0001, and Bonferroni post hoc tests were used to determine significance between the groups. Significances to the control group (negmiR + control siRNA) are shown: ∗p < 0.05, ∗∗∗p < 0.001.

Figure 3

Epas1 is a marker of fibroblast aging.A, fibroblasts were isolated from neonatal and adult cardiac tissue. After one passage, RNA was extracted and analyzed for the expression of Epas1 by qPCR. Epas1 expression values were normalized to the housekeeping gene Gapdh. N = 4. ∗p < 0.05 (two-way independent t test). B, fibroblasts were isolated from neonatal and adult cardiac tissue. After one passage, protein was extracted and analyzed for the expression of Epas1 by immunoblotting. Gapdh was used as a loading control. N = 4. Representative immunoblots are shown. C, fibroblasts were isolated from adult cardiac tissue. After one passage, cells were fixed and stained with an antibody targeting Epas1 (red). Cells were counterstained with phalloidin to mark the cytoskeleton (green) and DAPI to mark nuclei (blue). Representative image shown from three experiments. D, fibroblasts were isolated from neonatal cardiac tissue. RNA was isolated after one, two or three passages. Following extraction, RNA was analyzed for the expression of Epas1 by qPCR. Epas1 expression values were normalized to the housekeeping gene Gapdh. N = 4. One-way ANOVA, F (2, 9) = 8.897 p = 0.0074, with Bonferroni post hoc tests were used to determine significance. Comparisons are shown to the P1 group: ns-not significant, ∗p < 0.05. E, protein was also isolated after one, two, or three passages of neonatal cardiac fibroblasts and analyzed for Epas1 protein levels. Gapdh was used as the loading control. A representative image is shown from three separate experiments. F, fibroblasts were isolated from neonatal cardiac tissue. RNA was isolated after one, two, or three passages. Following extraction, RNA was analyzed for the expression of the cellular aging markers Cdkn1a and Cdkn2a by qPCR. Expression values were normalized to the housekeeping gene Gapdh. N = 4. One-way ANOVA, Cdkn1a: F (2, 9) = 81.97 p < 0.0001; Cdkn2a: F (2, 9) = 43.51 p < 0.0001, with Bonferroni post-hoc tests were used to determine significance. Comparisons are shown to the P1 group: ∗∗p < 0.01, ∗∗∗p < 0.001. G, fibroblasts were isolated from neonatal cardiac tissue. After one passage, cells were transfected with miRNAs (negmiR or miR combo) and plasmid DNA (empty plasmid or a plasmid containing an Epas1 expression cassette). Protein was extracted 3 days after transfection and immunoblotted. Immunoblots were probed with antibodies targeting Epas1 and Gapdh (loading control). Representative image shown from three independent experiments. H, fibroblasts were isolated from neonatal cardiac tissue. After one passage, cells were transfected with miRNAs (negmiR or miR combo) and plasmid DNA (empty plasmid or a plasmid containing an Epas1 expression cassette). After 14 days, expression of the indicated cardiomyocyte-specific markers was determined by qPCR. Expression values were normalized to the housekeeping gene Gapdh. N = 4. One-way ANOVA ,Actn2: F (2, 9) = 8.015 p = 0.0100; Myh6: F (2, 9) = 12.46 p = 0.0026, Tnni3: F (2, 9) = 13.79 p = 0.0018, with Bonferroni post-hoc tests were used to determine significance. Comparisons are shown to the control group: ns-not significant, ∗p < 0.05, ∗∗p < 0.01.

Figure 4

Epas1 promotes fibroblasts to develop an angiogenic phenotype.A, neonatal and adult cardiac fibroblasts (passage 1) were analyzed for the expression of the indicated genes by qPCR. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows the individual expression values (as Z-scores) for three independent experiments. A two-way independent t test was used to determine significance between groups. Genes highlighted in bold were significantly different (p < 0.05) between neonatal and adult cardiac fibroblasts. Raw expression data can be found in Table S1. B, adult cardiac fibroblasts (passage 1) were transfected with either a control non-targeting siRNA or a siRNA targeting Epas1. After 3 days, the cells were analyzed for the expression of the indicated genes by qPCR. Expression values were normalized to the housekeeping gene Gapdh. The heatmap shows the individual expression values (as Z-scores) for three independent experiments. A two-way independent t test was used to determine significance between groups. Genes highlighted in bold were significantly different (p < 0.05) between the control and Epas1 siRNA groups. Raw expression data can be found in Table S1. C, fibroblasts were isolated from neonatal and adult cardiac tissue. After one passage, the cells were cultured under defined conditions to promote tube formation (angiogenesis). After 3 days, tube formation was visualized. Representative images are shown from 11 independent experiments. For each experiment, three fields were imaged and the tube number averaged. Scale bar 100 microns. The surface area for each imaged field was 0.40 mm². Arrows denote tubes. A two-way independent t test was used to determine the significance between the two groups (∗∗∗p < 0.001). D, fibroblasts were isolated from adult cardiac tissue. After one passage, the cells were transfected with a siRNA (non-targeting control, Epas1 targeting) and two plasmids (Gluc and Seap). The Gluc plasmid contains a promoter (Vegf, Gata4, or Hand2) coupled to firefly luciferase. The Seap plasmid contains renilla luciferase and is used to normalize for transfection efficiency. Three days after transfection, luciferase measurements were made and are expressed as the Gluc/Seap ratio. N = 4. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (two-way independent t test).

Figure 5

Epas1 knockdown improves the efficacy of fibroblast to cardiomyocyte reprogramming in vivo. Fibroblast lineage tracing Fsp1Cre:tdTomato mice were subjected to either a sham operation or myocardial infarction. Immediately after injury, fibroblast-targeting exosomes containing a combination of miRNAs and siRNAs were injected into the border zone. Mice received one of three combinations: (1) negmiR plus a non-targeting control siRNA; (2) miR combo plus a non-targeting control siRNA; or (3) miR combo plus an Epas1 targeting siRNA. A, reprogramming events 2 months post-injury. Cardiac tissue sections were incubated with antibodies targeting tdTomato or Tnni3 (cardiomyocyte marker). Sections were also counterstained with DAPI to visualize nuclei. Representative images from 5 to 6 mice per group are shown. The number of cardiomyocytes derived from fibroblasts (tdTomato+ Tnni3+) is expressed as a percentage of the total cardiomyocyte population (Tnni3+). N = 5 to 6 per group. One-way ANOVA, F (2, 14) = 44.22 p < 0.0001, and Bonferroni post hoc tests were used to determine significance. ∗Comparisons made to the negmiR + control siRNA group: ∗p < 0.05, ∗∗∗p < 0.001. #Comparisons made between the two miR combo groups: ###p < 0.001. Scale bar 50 microns. B, two months post-injury, cardiac tissue sections were stained with Masson’s Trichrome to quantify fibrosis. Representative serial sections are shown for each group. To determine the amount of fibrosis, the percentage of fibrosis over five serial sections was calculated and averaged. N = 5 to 6 per group. One-way ANOVA, F (2, 14) = 316.4 p < 0.0001, and Bonferroni post hoc tests were used to determine significance. ∗Comparisons made to the negmiR + control siRNA group: ∗∗∗p < 0.001. #Comparisons made between the two miR combo groups: #p < 0.05. Representative serial sections are shown. Scale bar 500 microns. C, cardiac function was determined 2 months post-injury. Functional parameters measured: FS (Fractional Shortening), EF (Ejection Fraction), mVcFc (mean velocity of circumferential fiber shortening), LVDd (Left ventricle diastolic diameter), Vol;d (Left ventricle volume in diastole), LVDs (Left ventricle systolic diameter), and Vol;s (Left ventricle volume in systole). One-way ANOVA and Bonferroni post hoc tests used to determine significance. Results of statistical testing can be found in Table S2. N = 3 to 6 per group. Significance to the control group (MI, negmiR + control siRNA) is shown: ns-not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.