FOXO3-engineered human mesenchymal progenitor cells efficiently promote cardiac repair after myocardial infarction

Myocardial infarction (MI) is the irreversible cardiomyocyte death resulting from prolonged oxygen deprivation due to obstructed blood supply (ischemia), leading to contractile dysfunction and cardiac remodeling. In recent decades, stem cell transplantation has been extensively investigated for the repair of injured heart in animal studies and clinical trials (Kanelidis et al., 2017; Gyongyosi et al., 2018). Among cell-based therapies in clinical development, mesenchymal progenitor cells (MPCs) are attractive candidates due to their multi-lineage potential and immunomodulatory properties (Bagno et al., 2018). However, low quality (e.g., reduced proliferative ability and increased cellular senescence at late passages) and heterogeneous cell sources, as well as poor retention and survival rate of transplanted MPCs in an in vivo niche present obstacles towards broader clinical applications (Nguyen et al., 2016; Li et al., 2020). Forkhead box O3 (FOXO3), one of the most prominent genes related to human longevity, functions in diverse biological processes including DNA repair, oxidative stress response, cell proliferation and cellular senescence (Liu et al., 2018). We have previously reported that FOXO3 loss drives primate arterial aging and that constitutive activation of FOXO3 in human embryonic stem cell (hESC)-derived MPCs enhances their stress resistance and attenuates cellular senescence (Yan et al., 2019; Zhang et al., 2020). Here, we evaluated the cardiac repair after MI in immunodeficient mice intramyocardially transplanted with FOXO3-geneticallyenhanced MPCs (FOXO3-GE-MPCs). FOXO3-GE-MPCs were generated by directed differentiation of hESCs in which two FOXO3 phosphorylation sites were replaced with alanine (S253A, S315A) using targeted gene editing (Fig. S1A and S1B). The engineered FOXO3 could not be phosphorylated by AKT at S253 or S315 and was therefore constitutively active in the nucleus (Yan et al., 2019). Consistent with previous observations (Yan et al., 2019), FOXO3-GE-MPCs exhibited increased proliferation and decreased senescence-associated (SA)-β-gal activity relative to wildtype MPCs (WT-MPCs) (Fig. S1C–E). We next investigated whether FOXO3-GE-MPCs would be retained longer in the heart than WT-MPCs when intramyocardially delivered at the initiation of myocardial ischemia. In vivo imaging of luciferase-labelled MPCs revealed that transplanted WT-MPCs diminished within five days whereas FOXO3-GE-MPCs remained detectable until day 11 (Fig. 1A). Due to the limited resolution and sensitivity of in vivo imaging, we performed immunofluorescence staining of the human Golgi marker hTGN46 and RT-PCR of human GAPDH to further detect the transplanted cells in ischemic hearts and found that FOXO3 enhancement prolonged MPC retention up to 4 weeks after MI (Fig. 1B and 1C). Next, we used transthoracic echocardiography to explore whether FOXO3-GE-MPCs could ameliorate cardiac dysfunction and left ventricular (LV) remodeling after MI. In control mice (MI + vehicle group), we observed decreased cardiac contractility along with enlarged LV chamber at 4 weeks after MI. These effects were partially reversed by the transplantation with FOXO3-GE-MPCs, but not WT-MPCs (Fig. 1D). Similarly, an MI-induced increase in heart to body weight ratio and a decrease in running distance (Fig. 1E), along with cardiac fibrosis, compensatory hypertrophy and cardiomyocyte apoptosis, were all ameliorated only by FOXO3-GE-MPC transplantation (Figs. 1F–H and S1F). Collectively, our data indicate that FOXO3 enhancement promotes cardiac repair by MPCs after MI, suggesting that FOXO3-GE-MPCs may provide effective biomaterials for stem cell-based therapy against ischemic heart diseases. To dissect the underlying mechanisms of cardiac repair by FOXO3-GE-MPCs, we performed RNA-seq analysis of heart tissues of the infarct border zone (Figs. 2A, 2B and S1G–J) and identified a panel of MI-upregulated genes that were reversed by the transplantation of FOXO3-GE-MPCs, including those involved in inflammatory response (Fig. 2B). In addition, RelA (p65) that is a subunit of the NF-κB transcription complex (Wang et al., 2018) was upregulated in ischemic hearts and its upregulation was attenuated only by the transplantation with FOXO3-GE-MPCs (Figs. 2C and S1K). Some of the MI-upregulated genes rescued upon FOXO3-GE-MPC transplantation were enriched in NF-κB pathway (Fig. S1L). Furthermore, several NF-κB target genes including Cxcl13, Mmp9, Itgam, Tlr2 and Hmox1 were

divided into left and right ventricles, snap-frozen in liquid nitrogen and stored at -80℃ until use. 5 additional hearts for each group were fixed in 4% polyformaldehyde overnight and dehydrated in 30% sucrose, after which transversal frozen sections (10 μm) were prepared for histological assessment.

Immunofluorescence staining
Frozen sections were washed twice with PBS, permeabilized in 0.4% Triton X-100 for 10 min and blocked in 5% BSA in PBS for 30 min at room temperature. The sections were then incubated with primary antibodies diluted in blocking buffer (5% BSA in PBS) overnight at 4℃, followed by the incubation with fluorescent secondary antibodies and nuclear counterstain with Hoechst 33258. Immunofluorescence images were captured by laser scanning confocal microscopy and quantified using ImageJ. Antibodies used in this study are listed in Table S1.

TUNEL staining
Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was performed to evaluate cardiomyocyte apoptosis by using a one-step TUNEL apoptosis assay kit (C1088, Beyotime). The sections were subsequently stained with anti-cardiac troponin T antibody and Hoechst 33258. The images were captured by laser scanning confocal microscopy and the apoptotic rate of cardiomyocytes was quantified using ImageJ.

Masson's trichrome staining
After washing with PBS, myocardial sections were stained according to the protocol of Masson's Trichrome stain kit (G1340, Solarbio). All images were captured using a digital pathology slide scanner (Aperio CS2, Leica). Infarcted scar size in each heart was calculated as the ratio of total scar area to left ventricular area.

In vitro tube formation assay
HAECs were plated at a density of 6 × 10 3 cells per well onto Matrigel-coated 96-well plates and incubated at 37℃. Cell images were taken after 8 hours of incubation and the cumulative tube length was measured by ImageJ.

In vitro wound scratch assay
The HAECs were plated at a density of 1.5 × 10 4 cells per well onto collagen-coated 96-well plates and incubated at 37℃ until 100% confluence was reached. Cells were scratched once with a wound maker and the medium was changed. Subsequently, the wounding area was photographed once every hour for 16 hours using IncuCyte® Live-Cell analysis system (Essen BioScience, Hertfordshire, UK). The cell-free area was calculated using ImageJ. Migration ability was determined as the percentage of confluence relative to the initial size.

In vitro clonal expansion assay
Clonal expansion assay was performed as previously described with some modifications (Cheng et al., 2019). Briefly, HAECs were seeded at a density of 3 × 10 3 cells per well in collagen-coated 12-well plates and randomly divided into three experimental groups, FM group (EGM2 medium combined with fresh MPC medium at a ratio of 1:1), WT-MPC CM group (EGM2 medium combined with WT-MPC conditioned medium at a ratio of 1:1) and FOXO3-GE-MPC CM group (EGM2 medium combined with FOXO3-GE-MPC conditioned medium at a ratio of 1:1). Relative crystal violet-stained area was measured by ImageJ.

Reverse transcription PCR (RT)-PCR and quantitative reverse transcription PCR (RT-qPCR)
For RT-PCR, total RNA was extracted using TRIzol Reagent from mouse hearts. 500 ng of total RNA was reverse-transcribed to cDNA by using the GoScript Reverse Transcription System and oligo (dT) primer. PCR to detect human GAPDH in the mouse heart was carried out using Taq DNA Polymerase. Mouse Gapdh was used as an internal control. For RT-qPCR, total RNA was extracted using TRIzol Reagent from cultured MPCs. 2 μg of total RNA was reverse transcribed to cDNA as a template for real-time quantitative PCR analysis subsequently performed using a CFX384 Real-Time PCR system with iTaq Universal SYBR Green Super mix. GAPDH was used as an internal control. Primers used in this study are listed in Table  S2.

Measurement of serum inflammatory factors
Blood samples were collected through the abdominal aorta after the mice were anesthetized using a pro-coagulation tube. After centrifugation at 1,000 g for 10 min, serum was collected and used for measurement of inflammatory factors IFN-γ, IL-1β, and TNF-α via radioimmunoassay in collaboration with Beijing North Institute of Biotechnology Co., Ltd.

RNA-seq library construction and sequencing
Total RNA was extracted from the heart tissues from the infarct border zone after MI using the TRIzol reagent according to the manufacturer's protocol. Library construction and sequencing were carried out as previously described (Hu et al., 2020) mixed for each group and divided into 3 technical replicates. 2 μg of total RNA was used to construct sequencing libraries for Illumina by using a NEBNext® Ultra™ RNA Library Prep Kit. The libraries were sequenced on an Illumina HiSeq X-Ten platform.

RNA-seq data analysis
For human MPCs, reads from the previous data (accession number: GSE116277) were aligned to the human genome reference (hg19) with the hisat2 (version 2.0.4) and counted by HTSeq (version 0.11.0). Differentially expressed genes (DEGs) were calculated using R package DESeq2 (version 1.22.2) with a threshold of Benjamini-Hochberg adjusted P value < 0.05 and |log2 (fold change)| > 1. The DEGs were listed in Table S3.
For mouse hearts, low-quality reads were trimmed by TrimGalore (version 0.4.5) and clean reads were mapped to mouse mm10 genome using hisat2 and counted by HTSeq. DEGs were calculated using DESeq2 with a threshold of Benjamini-Hochberg adjusted P value < 0.05 and |log2 (Fold change)| > 0.58 and listed in Table S4. FPKM (fragments per kilobase per million) for each gene was calculated by StringTie (version 1.2.3). Gene ontology (GO) analysis (Biological Process) was conducted by using ToppGene (https://toppgene.cchmc.org/) by default parameters.

Statistical analysis
Two-tailed Student's t test or one-way ANOVA followed by Dunnett's test was performed as appropriate by GraphPad Prism 8.0 to assess the difference between groups. Data are presented as the mean ± SEMs. HAECs wound scratch assay was analyzed by two-way ANOVA followed by Dunnett's test. P < 0.05 is considered statistically significant.

Data Availability
RNA-seq data have been deposited in the NCBI Gene Expression Omnibus (GEO) under the accession numbers GSE149594 for mouse heart tissues and GSE116277 for human MPCs. Left, representative images of crystal violet staining. Right, the relative area of crystal violet-positive cells calculated by ImageJ and shown as the mean ± SEMs, n = 3. **, P < 0.01 (two-tailed t test). (E) SA-β-gal staining of WT-MPCs and FOXO3-GE-MPCs (P7). Scale bar, 50 μm.
Data are shown as the mean ± SEMs, n = 3. **, P < 0.01 (two-tailed t test). (F) Evaluation of apoptotic cardiomyocytes in mouse hearts immunostained with antibodies against α-actinin (green) and cleaved-caspase 3 (red). Scale bar, 25 μm. Quantitative data are shown as the mean ± SEMs. n = 3 for sham group; n = 5 for the other groups. ***, P < 0.001; ns, not significant (one-way ANOVA followed by Dunnett's test).   Table Legends  Table S1. Antibodies used in this study. Table S2. Primers used for RT-PCR and RT-qPCR. Table S3. DEGs identified in FOXO3-GE-MPCs and WT-MPCs, related to Figure 2. Table S4. DEGs identified in infarcted mouse hearts transplanted with FOXO3-GE-MPCs or vehicle and infarcted mouse hearts with MI or sham, related to Figure 2 and Figure S1.