Human decidual samples were collected from 20- to 30-year-old healthy women undergoing early pregnancy abortion in the clinic. Each sample was rinsed with saline, and then stored in a sterile bottle of a fully medium with 0.1% Penicillin and streptomycin until it was processed in the laboratory (<2 h later). All samples were rinsed for three times with sterile PBS, and then cut into 1–2 mm3 pieces, which were treated by 0.2% IV-type collaborate (Yeasen, Shanghai, China) and 0.25% trypsin. Ethylenediamine tetracetate (Trypsin EDTA; Gibco, Gaithersburg, MD, USA) covered the tissue, water bath (37°C/150 rpm) shaked it for 1 h subsequently. Cell digestion was filtered by a 200 μm nylon mesh (BD Falcon, San Jose, Ca, USA), which was cultured in DMEM/F12 (Hyclone, Southlogan, Utah, USA) medium containing 10% FBS (FBS; Sciencell, Sandiego, CA, USA) to terminate the enzymatic reaction. After being separated by centrifugation (1,000 rpm/5 min), cells were resuspended in 4 ml DMEM/F12 medium containing 10% FBS, and cultured in an incubator of 37°C, 5% CO2. On the second day, DMSCs were observed in the culture bottle (10). Similarly, BMSCs were also obtained from female patients (20–30 years old) who needed median thoracotomy in cardiac surgery, stored into a bottle pre-filled with heparin later, quickly transferred to the laboratory for treatment. An equal volume of lymphocyte separation solution was mixed with bone marrow by centrifugation at 2,500 × rpm for 20 min. BMSCs, the mononuclear cells layer, were obtained and washed twice. DMSCs and BMSCs were seeded in DMEM/F12 (HyClone) supplemented with 10% FBS (ScienCell) and cultured at 37°C in an atmosphere containing 5% CO₂ (10). Third-generation DMSCs/BMSCs were used for the following experiments.

All participants provided informed consent, and all procedures in the study were conducted in accordance with the ethical standards of the First/Second Affiliated Hospital of Harbin Medical University and the principles of the Helsinki Declaration.

The morphology of both MSCs was observed under a light microscope (Leica, Germany). As for cell identification, 2 × 10⁶ BMSC and DMSC cell suspensions were fully mixed with CD73-PerCP, CD90-FITC, CD105-APC, CD34-FITC and CD45-PerCP antibodies for 30 min followed by detection via flow cytometry (BD, USA).

DMSCs and BMSCs at an initial density of 1 × 10³ were cultured for 12, 24, 48, and 72 h in 96-well plates. Cell Counting Kit-8 (CCK-8; Dojindo, China) was utilized to detect proliferation ability according to the manufacturer's instructions (11).

DMSCs and BMSCs with initial densities of 300, 600, and 900/well were cultivated in 6-well plates and cultured for 2 weeks. After being fixed with 4% paraformaldehyde, the cells were covered with crystal violet staining solution for 10 min. A colony containing ≥20 cells was considered as a clonal colony.

The supernatants of DMSCs and BMSCs were deproteinized with 100 μL acetonitrile in centrifuge tubes, vortexed for 2–3 min, placed on ice water for 20 min, and then centrifuged at 12,000 rpm, 4°C for 15 min, which were transferred into another tube and centrifuged again with the mentioned conditions. At last, 200 μL supernatants were transferred into the sample bottle for LC-MS analysis. A 10 μL aliquot of the sample was injected into the ZORBAXSB-C18 column (AgilentTechnologies, INC) (12). Mass spectrometry analysis was assessed by using an Agilent Agilent 6,530-QTOF/MS instrument (Agilent Technologies) in both ESI + and ESI– modes. Finally, metabolic pathway enrichment was performed by using MetaboAnalyst.

Plasmid containing ODC shRNA or ODC sequences was transfected to inhibit or up-regulate ODC in DMSCs and BMSCs, respectively. Cells transfected with empty plasmids were regarded as the control group. Plasmid transfection was conducted via Lipofectamine (Invitrogen, USA) following the manufacturer's instructions (13). The ratio of transfection reagents to plasmid was 2 μl:1 μg. Western blot analysis was used to assess ODC protein levels as is previously reported (14). After 72 h culture, Transfected cells with GFP expression were visualized with fluorescence microscope (Leica, Germany).which was used to identify transfected cells and non-transfected cells. Then, transfection efficiency was assayed by fluorescence-activated cell sorting (FACS; BD, USA).

Using a transwell chamber coculture experiment system, the coculture experiment was performed in the chambers of 24-well plates (Corning, USA). DMSCs (10 × 10⁴) and BMSCs were inoculated in the upper chamber, and HUVECs (1 × 10⁴) were cocultured in the lower chamber for 24 h. HUVECs without hMSCs served as a negative control. VEGF and bFGF levels in the medium of the Transwell chamber were detected via ELISA kits (Cusabio, China) according to the manufacturer's instructions (15).

Human umbilical vein endothelial cells (HUVECs) were cocultivated with DMSCs or BMSCs for 24 h using a Transwell chamber coculture experiment system. The effects of DMSCs and BMSCs on HUVEC proliferation were examined by using a commercial EdU Kit (UE, China) according to the manufacturer's protocol. Images were obtained by using a fluorescence microscope (Leika, Germany) and analyzed with ImageJ (11).

Matrigel (100 μl/well; Corning, Standard, USA) was added to a 24-well plate (8 μm pore size; Corning, USA) and incubated for 30 min at 37°C. HUVECs (6 × 10⁴/well) were plated in Matrigel and inoculated in different media harvested from DMSC and BMSC culture media for 2 h. Images were obtained by light microscopy and analyzed by ImageJ.

A total of 50 adult male SD rats (200–220 g) were purchased from the animal experiment center of the Second Affiliated Hospital of Harbin Medical University and were divided into five groups: control, BMSCs, DMSCs,DMSCs + ODCshRNA and BMSCs + ODC. All procedures for animal experiments were performed in accordance with the guidelines reported in “Experimental Animal Care and Guide.” As is described in our recent study, rats were intraperitoneally injected with cyclosporin A (5 mg/kg; Novartis, Basel, Switzerland) per day from 3 days before surgery to the end of the experiment. Briefly, the rat was placed in a supine position to carry out the tracheal intubation (Artery puncture needle, 16G), the small animal inhalation anesthesia machine (Harvard Apparatus, Holliston, MA, USA) was used for anesthesia and maintaining breath, and inhaled gas was isoflurane. Through the left lateral thoracic incision, the left anterior descending coronary artery was ligated 1.5–2 mm below the left atrial appendage with a 5-0 Prolene needle. Myocardial infarction was confirmed by the whitening of the left ventricular anterior wall and apex. Ten min after ligation of the left anterior descending artery, 2.0 × 10⁶ cells from different groups were injected into the center and boundary of the infarct area.

Cardiac function was assessed based on the left ventricular ejection fraction (LVEF) and left ventricular shortening fraction (LVFS) via echocardiography. M-mode ultrasound images were collected in the section of the left ventricular papillary muscle via an S12–4 probe (EPIQ 5; Philips). All the data were collected based on at least four consecutive cardiac cycles.

The infarct size of MI rats was measured as is previously reported (10). The ratio of the scar length and the circumference were measured by Masson's trichrome staining.

One or 4 weeks after the operation, an immunofluorescence staining assay was performed to evaluate angiogenesis by detecting α-smooth muscle actin (α-SMA) (AF1032, Affinity) and vWF (ab6994, Abcam) expression. Moreover, transplanted cell survival was also detected by anti-cardiac troponin T (ab45932, Abcam) and anti-mitochondria (ab92824, Abcam) antibodies, according to the previous reports (15). Images were collected under an inverted fluorescence microscope (Leica, Germany).

Data were analyzed by using GraphPad Prism 5.0 and are presented as the mean ± SD (n = 5). The two groups were analyzed by t-test, and the multigroup comparison was analyzed with one-way ANOVA followed by Tukey's multiple comparisons post-hoc test. P < 0.05 was considered statistically significant. Each test was repeated at least thrice.

Both MSCs expressed characteristic surface markers CD73, CD90 and CD105 but not CD34 and CD45 (Figure 1A). Morphologically, DMSCs exhibited fibroblast-like wall growth and a large shuttle splicing hammer shape, which is similar to the morphology of BMSCs (Figure 1B). According to the results of CCK-8 and colony experiments, the proliferation rate of DMSCs was approximately 2 times higher than that of BMSCs (P_72h < 0.001 for CCK-8, P₉₀₀ < 0.001 for colony experiments; Figures 1C–E). The above results suggested that DMSCs exhibited better proliferation capability than BMSCs.

In addition, compared with BMSCs, DMSCs significantly increased HUVEC (Control) proliferation in vitro (P_{Control−BMSC} < 0.001, P_{Control−DMSC} < 0.001; Figures 1F,G). DMSCs effectively promoted the vascularization performance of HUVECs compared with BMSCs (P_BMSC−DMSC < 0.001; Figures 1H,I). Moreover, DMSCs stimulated HUVECs to produce more VEGF and bFGF, as is shown by ELISA (P_VEGF < 0.001, P_bFGF < 0.001; Figures 1J,K). These results indicated that DMSCs exhibit a more favorable capacity to boost the proliferation and angiogenesis of HUVECs than that of BMSCs.

Unsupervised clustering showed a significant difference in metabolic features between DMSCs and BMSCs (Figures 2A,B). In addition, partial least squares-discriminant analysis (PLS-DA) showed significant differences in metabolic profiles between BMSCs and DMSCs. The displacement test showed that the Q2 cum values were lower than the original values in almost all cases (Figures 2C,D), indicating favorable stability and effectiveness of the discriminated model. According to the KEGG enrichment results for all differential metabolites, arginine and proline metabolism were the most important metabolic pathways (Figure 2E). Compared with the supernatant of BMSCs, 4-hydroxyproline and niacinamide were decreased in DMSC supernatant, whereas only ornithine and spermine were increased significantly in DMSC supernatant (Figures 2F–I). The above results suggest that the increased ornithine and spermine levels in DMSC supernatant might represent the major factor of angiogenesis promoted by DMSCs more effectively compared with BMSCs.

The transfection of GFP plasmid could be observed under fluorescence inverted microscope, FC data showed that transfection efficiency was 58% in DMSCs and 31.5% in BMSCs (Figure 3A). After transfection of the ODC shRNA plasmid in DMSCs or ODC plasmid in BMSCs for 72 h, the protein levels of ODC were significantly down- and up-regulated in DMSCs and BMSCs, respectively (P_ODCshRNA < 0.01, P _ODC < 0.001; Figures 3B,C). As is expected, ODC inhibition in DMSCs obviously limited the proliferation of HUVEC, while ODC up-regulation in BMSCs promoted the proliferation of endothelial cells under co-cultivation conditions (P_{ODCshRNA−DMSC} < 0.01, P_ODC−BMSC < 0.001; Figures 3D,E). Further, whether ODC is involved in the regulation of endothelial function or not was assessed. Compared with the DMSC group, ODC inhibition for DMSCs resulted in the decrease capability of endothelial tubule formation (P_{ODCshRNA−DMSC} < 0.001; Figures 3F,G), and VEGF and bFGF secretion in HUVEC(P_VEGF < 0.001, P_bFGF < 0.001; Figures 3H,I). Conversely, ODC overexpression in BMSCs significantly promoted endothelial tubule formation (P_BMSC−ODC < 0.001; Figures 3F,G), and the secretion of VEGF and bFGF (P_VEGF < 0.001, P_bFGF < 0.001; Figures 3H,I). These results suggested that ODC is a critical target that promotes stem cell-mediated improvement of HUVECs proliferation and endothelial function in vitro.

In view of the survival rate after cell transplantation, anti-mitochondrial staining results suggested that DMSCs exhibited better survival rate than that of BMSCs (P_BMSC−DMSC < 0.001), whereas ODC inhibition significantly reduced the survival rate of DMSCs (P_{ODCshRNA−DMSC} < 0.001), and ODC overexpression improved the BMSCs survival post transplantation in rats heart with AMI (P_BMSC−ODC < 0.001) (Figures 4A,B). Microvessels and arterioles in the infarct area were visualized by immunofluorescence staining for vWF and a-SMA (Figures 4C,E). The number of microvessels and arterioles in the DMSC group was significantly more than that of BMSCs groups (P_BMSC−DMSC < 0.001; Figures 4D,F). ODC inhibition in DMSCs reduced the density of neovascularization compared with DMSC treatment (P_{ODCshRNA−DMSC} < 0.001), while ODC up-regulation in BMSCs significantly improved the angiogenesis capability compared with wile type BMSCs transplantation (P_BMSC−ODC < 0.001). The results suggested that DMSCs exhibited a greater survival rate and better angiogenesis in the infarcted area of MI partly by activating ODC pathway.

We further explored whether ODC mediated the effect of DMSCs/BMSCs transplantation on heart function after MI. Masson staining suggested that DMSC transplantation significantly inhibited myocardial fibrosis post-MI compared with BMSCs transplantation (Figures 5A,B). However, ODC inhibition masked the benefits of DMSCs transplantation, while ODC overexpression significantly improved the therapeutic effect of BMSCs to reduce the infarcted area (P_{ODCshRNA−DMSC} < 0.001, P_BMSC−ODC < 0.001; Figures 5A,B). Compared with the DMSC groups, DMSCs with ODC inhibition had lower LVEF and LVFS (P_{ODCshRNA−DMSC} < 0.001, P_BMSC−ODC < 0.01; Figures 5D,E), as well as thinner left ventricular wall, while ODC overexpression in BMSCs significantly improved the cardiac function (Figures 5C,E). The results showed that DMSCs could delay ventricular remodeling post-MI through ODC-mediated ornithine metabolism. The long-term LVEF and LVFS after DMSC and BMSC transplantation were significantly increased compared with the control group. However, LVEF and LVFS in the BMSC + ODC group were significantly increased compared with the BMSC groups (Figures 5C–E). These results suggest that DMSC transplantation can significantly improve cardiac function post-MI partly via ODC-mediated ornithine metabolism. There was no significant between-group difference in LVESd and LVEDd (Supplementary Figure 1).