A Latent Cardiomyocyte Regeneration Potential in Human Heart Disease

Abstract

Background: Cardiomyocytes in the adult human heart show a regenerative capacity, with an annual renewal rate of ≈0.5%. Whether this regenerative capacity of human cardiomyocytes is employed in heart failure has been controversial.

Methods: We determined cardiomyocyte renewal in 52 patients with advanced heart failure, 28 of whom received left ventricular assist device support. We measured the concentration of nuclear bomb test-derived ¹⁴C in cardiomyocyte genomic DNA and performed mathematical modeling to establish cardiomyocyte renewal in heart failure with and without LVAD unloading.

Results: We show that cardiomyocyte generation is minimal in end-stage heart failure patients at rates 18 to 50× lower compared with the healthy heart. However, patients receiving left ventricle support device therapy, who showed significant functional and structural cardiac improvement, had a >6-fold increase in cardiomyocyte renewal relative to the healthy heart.

Conclusions: Our findings reveal a substantial cardiomyocyte regeneration potential in human heart disease, which could be exploited therapeutically.

Keywords: heart failure; heart-assist device ◼ polyploidy ◼ regeneration; myocytes, cardiac.

Figure 1.

Isolation of human cardiomyocytes and ploidy assessment. A, Flow cytometry–based sorting enables identification and isolation of cardiomyocyte nuclei with antibodies against PCM1 (inset). B, Flow cytometry analysis of cardiomyocyte nuclei DNA content reveals their ploidy profile (shown is a representative NICM sample). C, The distribution of cardiomyocyte nuclear ploidy populations shows a shift to higher ploidy levels in NICM (n=23) and ICM patients (n=11) compared with healthy adults (n=11). D, Cardiomyocyte nuclear ploidy level determined by flow cytometry is greater in both NICM (n=23) and ICM patients (n=11) than in healthy adults ([n=11] eg, 100% corresponds to only diploid nuclei, 200% to only tetraploid nuclei; Kruskal–Wallis 1-way ANOVA on ranks, H=30.4; P<0.0001, post hoc Dunn **P=0.0063, ****P<0.0001), lines show median with interquartile range. E, Ploidy increase in cardiomyopathy is not related to the age of the adult patient (NICM, R=0.03, P=0.89; ICM, R=0.35, P=0.28). F, Image cytometry identifies cardiomyocyte ploidy classes. Cardiomyocytes were digested from tissue samples and stained for α-actinin (α-act) and connexin-43 (Cx43). The integrated intensity of the DNA dye was used to designate the nuclear ploidy class of each imaged nucleus, as shown by the multicolor overlay. Cx43 enabled us to determine whether cardiomyocytes were intactly dissociated at the intercalated discs. The figure shows a compilation of several cardiomyocytes of different ploidy and nucleation levels. Scale bar=100 µm. G, The percentage of binucleated cardiomyocytes determined by image cytometry is greater among NICM (n=26) and ICM hearts (n=8) compared with healthy hearts ([n=11] ordinary 1-way ANOVA, F=43.5; P<0.00001, post hoc Tukey; ****P<0.0001). H, Both binucleation and nuclear ploidy levels were determined in isolated cardiomyocytes from healthy (n=7) and pathologic hearts in samples from individual patients (NICM n=8; ICM n=3) with image cytometry. I, The percentage of mononucleated diploid cardiomyocytes is lesser among NICM (10.4%) and ICM (7.8%) compared with healthy hearts ([24.0%] Kruskal–Wallis 1-way ANOVA on ranks; H=7.7; P=0.01, *post hoc Dunn P=0.02). 2n indicates diploid; 4n, tetraploid; 8n, octaploid; 16n, hexadecaploid; ICM, ischemic cardiomyopathy; NICM, nonischemic cardiomyopathy; PCM1, pericentriolar material 1; and SSC, side scatter.

Figure 2.

¹⁴C levels indicate increased DNA synthesis in cardiomyocytes of failing hearts. A, Presentation of ¹⁴C data. The black curve indicates the historic atmospheric ¹⁴C concentrations measured in the northern hemisphere.^{– 14}C measurements (delta ¹⁴C permille (‰) in relation to a universal ¹⁴C standard and corrected for radioactive decay, from heart samples are plotted as colored dots on the date of subject birth. Genomic ¹⁴C values of cardiomyocytes in healthy hearts ([blue dots] n=18; data from Bergmann et al, 2015) are close to atmospheric values at the time of birth, indicating a limited postnatal and adult renewal of cells. The deviation of ¹⁴C values of cardiomyocytes in nonischemic cardiomyopathy ([NICM] red dots; n=16) and ischemic cardiomyopathy (ICM) hearts (orange dots; n=8) from the atmospheric ¹⁴C curve suggest genomic DNA turnover. B, The estimated genomic ¹⁴C age of cardiomyocytes was calculated from subjects with postbomb birth dates and plotted at participant age. Black line indicates no turnover. Subjects born before the bomb spike (1963) were not included in this analysis because the genomic ¹⁴C age cannot be definitively determined, which related to shape of the atmospheric ¹⁴C curve.^– C, Both NICM (n=8) and ICM samples (n=4) show a greater deviation of the estimated genomic ¹⁴C age from the participant’s age compared with healthy hearts (n=12), suggesting increased DNA synthesis (Kruskal–Wallis 1-way-ANOVA on ranks, H=17.4; P=0.00017, post hoc Dunn multiple comparison, ***P=0.0004, *P=0.017).

Figure 3.

Mathematic modelling of cardiomyocyte renewal dynamics in failing hearts. A, Violin plot of the probability distribution of the true parameter of the turnover rate before onset and after onset for nonischemic cardiomyopathy (NICM) and ischemic cardiomyopathy ([ICM] Monte Carlo samples, n=200 000; Supplemental Material). Dashed line indicates the median turnover rate; dotted lines indicate the range of 50% credibility interval (ie, true value is in this range with 50% probability). In the disease phase, the percentage of newly formed cardiomyocytes drops from 0.6% ([0.5–0.6%] median [interquartile range]) to 0.03% (0.002%–0.5%) in NICM, and 0.01% (0.001%–0.1%) in ICM. B, The annual amount of DNA synthesis predicted by our model using the fitted renewal rate and the case-specific data such as age, date of diagnosis, and polyploidization. It shows an increase in both NICM and ICM compared with healthy hearts (Kruskal–Wallis 1-way ANOVA on ranks, H=41.6; P<0.00001, post hoc Dunn multiple comparison; ****P<0.0001). C, Origin of the annual amount of synthesized DNA attributed to cell renewal, binucleation and nuclear polyploidization estimated in diseased hearts predicted by our model (Supplemental Methods). After disease onset in NICM and ICM patients <0.3% of DNA synthesis can be attributed to proliferation-based cell renewal. The compositional data were analyzed as described in the statistics section. ANOVA revealed significant effects of condition on the source of newly formed DNA (F(2, 53)=1424, P<2e-16), post hoc Tukey honestly significant difference test; ****P<0.00001).

Figure 4.

¹⁴C levels in cardiomyocytes of LVAD-supported hearts. A, Presentation of ¹⁴C data from unloaded hearts. ¹⁴C measurements from heart samples are plotted as colored dots on the participant’s date of birth for diseased non-LVAD hearts ([red dots; n=24] these correspond to nonischemic cardiomyopathy [NICM] and ischemic cardiomyopathy [ICM] samples as shown in Figure 2A), LVAD nonresponders (n=13) and LVAD responders (n=15). The location of the dots from responder subjects suggests an increased deviation from the atmospheric ¹⁴C curve compared with non-LVAD subjects. B, The estimated genomic ¹⁴C age of cardiomyocytes was calculated from subjects with post-bomb birth dates and plotted at the person’s age. The black line indicates no turnover. C, Responders (n=8) show genomic ¹⁴C age, which is 19.3 years younger than the participant on average, compared with 13.9 years for LVAD nonresponders (n=3) and 12.6 years for non-LVAD patients (n=12). Ordinary 1-way ANOVA, F=4.2; P=0.03, post hoc Tukey, *P=0.025. LVAD indicates left ventricular assist device.

Figure 5.

Mechanical unloading increases cardiomyocyte renewal in LVAD responders. A, Cardiomyocyte renewal rate as determined using the 2-phase renewal model. In cardiomyopathy, the percentage of newly formed cardiomyocytes per year is 3.1% (2.2–3.9; median [interquartile range]) in LVAD responders, compared with 0.03% (0.002%–0.4%) in the non-LVAD group and 0.02% (0.001%–0.2%) in LVAD nonresponders. B, Percentage of DNA synthesis attributed to cell renewal, binucleation, and nuclear polyploidization determined at the time of sample collection. The compositional data were analyzed as described in the statistics section. The compositional data were analyzed as described in the statistics section. ANOVA revealed significant effects of condition on the source of newly formed DNA (F[2, 49]=21.05; P<3e^-07); post hoc Tukey honestly significant difference test, *****P<1e^-7, **P<0.001. LVAD indicates left ventricular assist device.