Serum circulating proteins from pediatric patients with dilated cardiomyopathy cause pathologic remodeling and cardiomyocyte stiffness

Figures

• 

  Figure 1. DCM serum elicits significant transcriptional changes in NRVMs. (A) Volcano plot representation of RNA-Seq transcripts significantly differentially expressed (4079 genes total) in NRVMs treated with NF (n = 6) and DCM (n = 6) serum from 3 independent NRVM preps. Log2 fold change is represented on the x axis and –log10 q value is represented on the y axis. Horizontal dashed line represents significance cutoff, Welch’s 2-tailed t test, assuming equal variance q < 0.10. (B) Heatmap representing unsupervised clustering of NF serum–treated (n = 6) and DCM serum–treated (n = 6) NRVMs with 629 significantly expressed transcripts. DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes; NF, nonfailing, FC, fold change.

• 

  Figure 2. Fetal gene program expression changes in DCM serum– and NF serum–treated NRVMs with or without DNase and RNase. (A) RT-qPCR of FGP expression in NRVMs transfected with serum with or without DNase. Expression of BNP, ANF, and α-MyHC to β-MyHC ratios. Gene expression was normalized to 18S, and data are presented as a relative fold change to NF controls. DCM n = 8, NF n = 4 from 7 independent NRVM preps. (B) RT-qPCR of FGP expression analysis of serum-treated NVRMs with or without RNase. Expression of BNP, ANF, and α-MyHC/β-MyHC was evaluated. Gene expression was normalized to 18S, and data are presented as a relative fold change to NF controls. DCM n = 9, NF n = 5 from 9 independent NRVM preps. All groups are log2 transformed and error bar denotes mean ± SEM. P values are notated in the figure. Fitting a mixed model, Sidak’s multiple comparisons test was used for all data sets. DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes; NF, nonfailing; FGP, fetal gene program; BNP, natriuretic peptide B; ANF, atrial natriuretic factor; α-MyHC, α-myosin heavy chain; β-MyHC, β-myosin heavy chain.

• 

  Figure 3. Circulating miRNAs are dysregulated in serum from pediatric patients with DCM. (A) Volcano plot representation of the 123 differentially expressed miRNAs between NF (n = 12) and DCM (n = 32) patients. Log2 fold change is represented on the x axis and –log10 q value differentially expressed on the y axis. Horizontal line indicates significance; Wilcoxon’s test q < 0.10. (B) Heatmap of the 84 significantly differentially expressed circulating miRNAs. Unsupervised hierarchical clustering separated NF (n = 12) and DCM (n = 32) patients; Wilcoxon’s test q < 0.10. DCM, dilated cardiomyopathy; NF, nonfailing; FC, fold change.

• 

  Figure 4. DCM serum activation of the FGP is mediated by circulating proteins. RT-qPCR of FGP expression analysis of serum-treated NVRMs with or without PTNK. Expression of BNP, ANF, and α-MyHC/β-MyHC was evaluated. Gene expression was normalized to 18S. DCM n = 4, NF n = 4 from 4 different NRVM preps. Data are presented as a relative fold change to NF controls; all groups are log2 transformed and error bar denotes mean ± SEM. P values are notated in the figure. Fitting a mixed model, Sidak’s multiple comparisons test was used for all data sets. DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes; NF, nonfailing; FGP, fetal gene program; BNP, natriuretic peptide B; ANF, atrial natriuretic factor; α-MyHC, α-myosin heavy chain; β-MyHC, β-myosin heavy chain; PTNK, proteinase K.

• 

  Figure 5. Circulating proteins are dysregulated in plasma of pediatric patients with DCM. (A) Volcano plot representation of the 1141 differentially expressed proteins between NF (n = 4) and DCM (n = 8) patient plasma. Log2 fold change is represented on the x axis and –log10 q value differentially expressed on the y axis. Horizontal dashed line represents significance cutoff; Welch’s t test q < 0.10. (B) Heatmap of the 102 significantly differentially expressed proteins. Unsupervised hierarchical clustering separated NF (n = 4) and DCM (n = 8) patients; Welch’s t test q < 0.10. DCM, dilated cardiomyopathy; NF, nonfailing; FC, fold change.

• 

  Figure 6. Circulating MDK is upregulated in patients with DCM and increases expression of ANF and BNP. (A) Levels of circulating MDK in DCM (n = 8) compared with NF plasma (n = 4). P < 0.0001 by unpaired 2-tailed t test. (B) Levels of tissue MDK in DCM (n = 7) compared with NF (n = 7). P = 0.019. (C) BNP, ANF, and α-MyHC/β-MyHC expression in NRVMs treated with MDK compared with control. n = 8 independent experiments. (D) DCM serum treatment with MDK antibody blunts the increase in BNP or ANF. n = 4 independent experiments. (C and D) Data are presented as a relative fold change to NF controls; all groups are log2 transformed and error bar denotes mean ± SEM. P values are notated in the figure. A comparison between 2 groups was done using unpaired 2-tailed t test, whereas a 4-group analysis was done using Sidak’s multiple comparisons test. MDK, midkine; DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes; NF, nonfailing; BNP, natriuretic peptide B; ANF, atrial natriuretic factor; α-MyHC, α-myosin heavy chain; β-MyHC, β-myosin heavy chain.

• 

  Figure 7. Comparison of upregulated and downregulated from serum-treated NRVM transcripts and circulating proteins found in patient plasma. (A) Venn diagram representation of overlapping pathways found to be upregulated based on dysregulated transcripts from DCM serum–treated NRVMs and protein plasma from patients with DCM (369 unique transcriptome pathways, 660 unique protein pathways, and 243 overlapping pathways). (B) Venn diagram representation of overlapping pathways found to be downregulated based on dysregulated transcripts from serum-treated NRVMs and protein plasma from patients with DCM (362 unique transcriptome pathways, 440 unique protein pathways, and 11 overlapping pathways). DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes.

• 

  Figure 8. Comparison of pathways commonly altered in the serum-treated NRVM transcriptome and pediatric DCM transcriptome. Venn diagram representation of overlapping pathways altered in pediatric DCM left ventricle (LV) tissue and in serum-treated NRVMs. DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes.

• 

  Figure 9. Human recombinant protein sFRP1 increases stiffness in NRVMs. (A). Elasticity/stiffness (Young’s modulus) in NRVMs treated with NF and DCM pediatric patient serum. Mann-Whitney U test, P = 0.0009 (serum). Each dot represents an individual measurement. (B) Circulating levels of sFRP1 in DCM (n = 8) compared with NF (n = 4) controls. P = 0.0139. (C) Levels of sFRP1 mRNA in pediatric DCM tissue (n = 7) compared with NF controls (n = 7). P = 0.0002. (D) NRVMs treated with 1 μg/mL human recombinant MDK for 72 hours. (E) NRVMs treated with 1 μg/mL human recombinant (hsFRP1) for 72 hours. P < 0.0001. Error bar denotes mean ± SEM. Mann-Whitney U t test (AFM) or unpaired 2-tailed t test (log2 transformed circulating and tissue sFRP1) were used and P values are notated in the figure. Mann-Whitney was used in A, D, and E and t test in B and C. DCM, dilated cardiomyopathy; NRVMs, neonatal rat ventricular myocytes; NF, nonfailing; MDK, midkine; AFM, atomic force microscopy; sFRP1, secreted frizzled-related protein 1.