A simple and efficient method for purification of human pluripotent stem cell-derived cardiomyocytes via activating p53

Dear Editor,

Human pluripotent stem cell-derived cardiomyocytes (hPSC-CMs) are of great values for developmental studies, disease modeling, drug screening, cell therapy, and beyond [1]. Achieving high-purity of hPSC-CMs is essential for promoting their application. Purification of hPSC-CMs by switching the culture medium to lactate-rich, glucose-free, and nutrient-poor purification medium (lactate-method) has been widely-used [2], whereas it would induce “ischemic-like” phenotypes with spontaneous arrhythmic activity, abnormal calcium handling, and shortened sarcomeres [3]. Moreover, the majority of applications require more maturity of hPSC-CMs [4], which needs to extend cultivation time. However, prolonged cultivation results in the expansion of non-CMs (Fig. 1a, i), leading to a decreased purity of hPSC-CMs (Fig. 1a, ii). Furthermore, purification efficiency of the lactate-method for hPSC-CMs declined beyond differentiation day 20 (Fig. 1a, iii). Therefore, an alternative purification approach with less restrictions is required.

Fig. 1. A simple and efficient method for the purification of hPSC-CMs by activating p53.

a Growth of non-CMs (i) and the flow cytometry analysis of the purity of H9-CMs after lactate purification (ii). (iii) H9-CMs purified with the lactate method from differentiated day 12 or day 21. Scale bar, 100 µm. b Cell death analysis by flow cytometry of H9-CMs treated with 20 µmol/L of nutlin-3a for 4 days. c Purity of H9-CMs at day 4 post-recovery from lactate- or nutlin-treatment (20 µmol/L). The purification was initiated at differentiation day 12 and sustained for 2 days (i); purity (ii), and immunofluorescent staining (iii) of H9-CMs treated from differentiation day 21 for 2 or 4 days. (iv) The subtypes of H9-CMs and the action potential properties of ventricular-like H9-CMs at day 35. BR: beating rate, beats/min; MDP: maximum diastolic potential, mV; APA: action potential amplitude, mV; APD90: AP duration at 90% repolarization, ms; APD50: AP duration at 50% repolarization, ms. Twenty-one cells in control group and 20 cells in nutlin-3a groups were recorded. All 13 ventricular (V)-like cells in the control group and 14 in the nutlin-3a group were summarized. Scale bar, 50 µm. d Images (i) and cell death analysis by flow cytometry (ii) of H9 ESCs treated with 20 μmol/L of nutlin-3a up to 24 h. (iii) ALP staining of mixed H9 ESCs and H9-CMs after a 5-day recovery following the 2-day treatment with the lactate- or nutlin-method. (iv) Analysis of teratoma formation at 6 weeks after the injection of the mixed cells with or without nutlin-3a treatment (20 μmol/L). Scale bar, 50 µm in i; 100 µm in iii; 1 cm in iv. e Western blot of p53 with GAPDH as an internal control (i); images (ii) and cell death analysis (iii) of H1-wt and H1-p53^−/− hESCs treated with nutlin-3a (20 μmol/L) for 24 h; purity of H1- and H1-p53^-/--CMs treated with nutlin-3a (20 μmol/L) at differentiation day 21 for 2 days followed 4 days recovery (iv). Scale bar, 50 µm. f The purity of various hPSC-CMs purified with the lactate- or nutlin-method. The cells at differentiation day 21 were purified for 2 days and harvested after 4 days of recovery. n = 5 in aii, aiii, b, ci, cii, dii, diii, eiii, eiv. n = 3 in (f). All summary data are presented as mean ± SEM. Statistical significance was assessed by one-way ANOVA followed by Tukey’s test for aii, ci, cii, f; Student’s t-test in b, dii, diii; and two-way ANOVA analysis followed by Sidak’s test for aiii, eiii, eiv. ***P < 0.001. n.s. no significant difference.

To address this, we developed a purification method for hPSC-CMs based on the different sensitivity of cells to death-regulatory signaling. p53 is robustly activated during CM maturation [5], indicating the apoptotic-resistance of CMs after p53 activation. Indeed, no additional cell death was observed in H9 human embryonic stem cell (hESC)-derived CMs (H9-CMs) after treatment with 20 μmol/L of nutlin-3a, a p53 activator, for 4 days (Fig. 1b). We thus tested whether activating p53 can selectively purify hPSC-CMs by killing non-CM cells. Various concentrations of nutlin-3a were tested for purification of H9-CMs from differentiation day 21 for a 2-day treatment followed by a 4-day recovery. As shown in Fig. S1, flow cytometry analysis showed that the CMs represented by CTNT positive cells were 60.84% ± 2.35%, 65.4% ± 0.19%, 80.9% ± 1.21%, 90.74% ± 0.56%, 93.46% ± 0.80%, and 96.12% ± 0.55% in the cells treated with nutlin-3a at 0, 2.5, 5, 10, 20, and 40 μmol/L, respectively. No statistical significance in the purity of CMs was observed between the H9-CMs treated with 20 and 40 μmol/L of nutlin-3a (Fig. S1b).

Then, the cytotoxicity of nutlin-3a in H9-CMs was examined after the cells were treated up to 160 μmol/L from differentiation day 21 for a 4-day followed a 4-day recovery, which is a more rigorous condition than a 2-day treatment of nutlin-3a for purification. To limit the influence of cytotoxicity of nutlin-3a to non-CMs, the H9-CMs with high purity at differentiation day 21 were used for the test (Fig. S2a). The H9-CMs treated with same amount of DMSO were used as the control. No significant difference in the morphology of H9-CMs was observed up to 20 μmol/L of the nutlin-3a treatment, while the cell number was reduced in group treated with 40 μmol/L nutlin-3a, and almost all cells died in 80 and 160 μmol/L of nutlin-3a-treated groups (Fig. S2b). Consistently, the cell viability evaluated by CCK8 showed a significant reduction by 40 μmol/L of nutlin-3a and reached the bottom in the H9-CMs treated with 80 μmol/L and 160 μmol/L nutlin-3a, respectively (Fig. S2c). The 50% of lethal concentration of nutlin-3a in H9-CMs at differentiation day 21 was 29.27 μmol/L. It is notably that the CCK8 value was reduced to 77.17% in the H9-CMs treated with 20 μmol/L nutlin-3a (Fig. S2c), which may be caused by the inhibition of nutlin-3a in proliferation of tested cells as no additional cell death in H9-CMs treated with nutlin-3a at 20 μmol/L was detected compared with the cells without the nutlin-3a treatment (Fig. 1b and Fig. S2d). However, robust dead cells were detected in the H9-CMs treated with 40 and 80 μmol/L nutlin-3a (Fig. S2d), which is consistent with the morphology of cells (Fig. S2b) and CCK8 assay (Fig. S2b). These data demonstrate hPSC-CMs are resistant to nutlin-3a-induced cell death at the concentration of 20 μmol/L and lower. Taken together with the purification efficacy of various concentrations of nutlin-3a (Fig. S1), our data indicate that the proper concentration of nutlin-3a for the purification of hPSC-CMs lies within the range from 10 to 20 μmol/L and higher concentrations of nutlin-3a can cause significant death of the hPSC-CMs.

Nest, we compared the purification efficacy of the lactate-method and nutlin-3a-based method. For hESC-CMs at differentiation day 12, the purity of CMs was elevated from 69.82% ± 3.33% to 92.98% ± 1.38% in lactate-method group and 94.38% ± 1.64% in nutlin-3a group after purification for 2 days (Fig. 1c, i). However, for hESC-CMs purified from differentiation day 21, the nutlin-3a method maintained a high purification efficiency, whereas the lactate-method failed to remove the non-CM cells, even the purification period was prolonged to 4 days (Fig. 1c, ii). After recovery from the purification procedure for 6 days, the sarcomeric structure and shape were comparable between each group (Fig. 1c, iii). Moreover, similar proportions of nodal-, atrial-, and ventricular-like CMs were detected in both control and nutlin-3a-purified H9-CMs, and the comparable properties of action potentials (APs) of ventricular-like CMs were detected between the control and nutlin-3a-purified groups (purification starts from differentiation day 21 for 2 days) (Fig. 1c, iv).

The low percentage of undifferentiated hPSCs remains in hPSC-CM population [2], which is one of the critical safety concerns for the application of hPSC-CMs in cell therapy. hPSCs are hyper sensitive to p53 signaling and the p53 activation can induce a rapid cell death [6]. Thus, we tested the efficacy of the nutlin-3a-method in removing undifferentiated hPSCs. After changing culture medium from mTeSR1 to mTeSR1 + 20 μmol/L nutlin-3a, the H9 cells died rapidly within 8 h, and all cells died after treatment for 24 h (Fig. 1d, i and ii). We then tested the removement of undifferentiated hESCs in the mixture of H9 hESCs and H9-CMs. The H9-CMs at day 17 of differentiation were mixed with H9 hESCs (ratio of cell numbers, 100:1) and cultured as mixture for 3 days in mTeSR1, followed by a 2-day application of lactate- or nutlin-method. After the recovery for 5 days, small clones with high alkaline phosphatase activity, indicating undifferentiating hESCs, were observed in the lactate-treated cells but not in the nutlin-3a-treated ones (Fig. 1d, iii). The data suggest that nutlin-3a is more efficient in removing the undifferentiated hESCs. Consistently, no teratoma was observed after the growth for 6 weeks  in all immunodeficient NOD/ShiLtJGpt-Prkdc^em26Cd52Il2rg^em26Cd22/Gpt (NCG) mice injected with 1 million of mixed cells treated with nutlion-3a (20 μmol/L) for 2 days. However, teratomas with three germ layers were formed in all injected sites with the nutlin-3a-untreated cells (Fig. 1d, iv and Fig. S2).

Further, we examined whether removement of non-CMs (including the undifferentiated hPSCs) is p53 dependent. The TP53 gene was mutated by CRISPR after translation starting site, which resulted early termination of the protein translation in H1 hESCs (H1-p53^-/-). The deficiency of p53 protein was validated by Western blot (Fig. 1e, i). The cell death induced by nutlin-3a was totally blocked in H1-p53^-/- hESCs (Fig. 1e, ii, iii). In addition, the purification effect of nutlin-3a in H1-p53^-/--CMs was totally vanished (Fig. 1e, iv). These data indicate that the depletion of non-CMs is p53-dependent.

Further, we examined the purification efficacy of nutlin-3a (20 μmol/L) in the CMs derived from various hPSC lines by treating the hPSC-CMs at differentiation day 21 for 2 days following by a 4-day recovery. For all hPSC-CMs tested, nutlin-3a showed robust purification efficacies, whereas the lactate method could not improve the purity of hPSC-CMs (Fig. 1f).

In summary, we have developed a novel method for purification of hPSC-CMs with several advantages: (i) it is a simple, universal, and straightforward approach with high efficacy of purification for hPSC-CMs across early and later (over 20 days) stages; (ii) it effectively removes undifferentiated hPSCs, thereby enhancing the safety of applying hPSC-CMs in cell therapy; and (iii) the purified cells exhibit normal morphology and electrical properties. Thus, this alternative method for purifying hPSC-CMs could be employed under conditions that require extended culture times, thereby reducing assay variations in applications, such as disease modeling and drug screening. Additionally, it can provide seed cells with higher safety for the use in cell therapy.

Author contributions

PZ and HTY conceived the study. PZ and HTY designed experiments; PZ, SLR, XCW, PL, JZL, and WWJ performed the experiments. PZ and HTY analyzed the data and wrote the manuscript.

Competing interests

The authors have filed a patent (202410535584.6) for the application of purification of hPSC-CMs via activating p53.

Ethical approval

The animals were cared for in accordance with the Guidelines for Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication, 8th Edition, 2011). All procedures involving experimental animals were approved by the Institutional Animal Care and Use Committee of Shanghai Institute of Nutrition and Health (SIBS-2022-YHT-1).

Supplementary information

The online version contains supplementary material available at 10.1038/s41401-025-01543-5.