All procedures with human material were approved by the ethical committee of Meshalkin National Medical Research Centre (permit No. 45 by 26.12.2014), and all donors signed an informed consent form. The animal tests conducted in an SPF vivarium were approved by the Institute of Cytology and Genetics ethical board (permit No. 22.4 by 30.05.2014).

Synthetic biodegradable scaffolds from polycaprolactone (PCL) (Sigma, #44074-5G) with gelatin (Sigma, #G2500-100G) and a low-permeable inner layer were prepared by electrospinning according to a previously published work (Popova et al., 2015). Synthetic scaffolds with a low-permeable layer were made by sequentially applying 3 layers of a matrix from polymer solutions of a certain composition using the electrospinning method, as described in the patent RU2572333. The thickness of the scaffold was ∼80 µm. The first inner matrix layer and the next low-permeable layer have about 10 µm each. Low-permeable layer was produced from 1.5% w/v PCL with 3.5% w/v gelatin in hexafluoroisopropanol, while remaining scaffold from 4.5% w/v PCL and 0.5% w/v gelatin in the same solvent. The electrospinning conditions include the solution feed rate of 1.2 ml/h, the voltage of 20–25 kV, the cylindrical collector rotation speed of 300 rpm, and the distance between the capillary and the collector of 20 cm. After fabrication, the scaffolds were removed from the collector and dried in a vacuum. After preparation, the matrix was incubated in phosphate buffer for at least an hour and then cross-linked with glutaraldehyde, as described previously (Chernonosova et al., 2018).

The scaffold microstructure was examined by scanning electron microscopy (SEM) as described earlier (Chernonosova et al., 2017). The fiber diameter was assessed from the SEM images according to ISO 7198:1998.

The morphology of the resulting scaffolds was studied using scanning electron microscopy. The image was obtained on a JSM-6460 LV scanning electron microscope (Jeol, Japan). This part of the work was carried out at the Department of Physicochemical Methods of the Institute of Catalysis, SB RAS. Electron microscopy images of the scaffold are shown in Figure 1.

According to SEM images, after exposure to phosphate buffer for an hour, the scaffolds have a fiber diameter of 0.5 ± 0.1 μm and a denser layer with low permeability. It was shown the low-permeable inner layer does not affect the strength of the tubes, displaying a tensile strength of 2.2÷2.5 MPa, and a thread breaking strength (suture retention strength) of 160 ± 20 g of force. A low-permeable inner layer reduces the material permeability for water from 5 to 10 ml to 0.1 ÷ 0.2 ml.

Endothelial and smooth muscle cells were isolated and cultivated as described in our previous work (Zakharova et al., 2017). For experiments with MMC, cells growing on plastics were detached with TrypLE, washed with PBS, centrifuged at 300 g for 5 min followed by dissolving the pellet in a growth medium containing 10 μg/ml MMC (Sigma, #M4287) (1 ml per 5 × 10⁵ cells) and incubated for 2 h at 37°C in 5% CO₂, vortexing occasionally. The cells were then washed 3 times with PBS (Biolot, #2.1.1) and seeded in a plastic tissue culture flask with a growth medium.

To prepare a tissue-engineered construct a PCL patch of 2 cm² was seeded with 3 × 10⁵ per cm² of cells. First, on one side of the patch, we plated smooth muscle cells in SmGM-2 medium (Lonza, #CC-3182). As a control, PCL patches without cells were placed into the same medium. After a day, the patch was turned over, and the other side was seeded with endothelial cells in a mixture of EGM-2 (Lonza, #CC-3162) + SmGM-2 media. Control PCL patches with no cells were placed in the same medium. The patches seeded with cells and unseeded control patches were cultured for another 5 days. The cells on the patches and unseeded control patches were incubated at 37°C in 5% CO₂.

We plated the cells on a scaffold for 6 days to allow them to form an extracellular matrix (ECM), which creates the physiological cellular microenvironment necessary for proper integration of the patch into the surrounding tissues and adequate colonization of the recipient cells. The ECM contains internal biochemical and mechanical signals that contribute to the formation of the correct hierarchy of recipient cells, populating the patch after engraftment (Hussey et al., 2018; Niklason, 2018; Zhu et al., 2019).

The animals were kept in single-sex family groups of 2–5 mice in individually ventilated cages using the Optimice system (Animal Care Systems, Centennial, CO, United States). The mice were maintained under controlled conditions: temperature, 22–26°C; relative humidity, 30–45%; and 12/12 light/dark periods with dawn at 02:00. Standard V1534-300 ssniff^® diet (ssniff Spezialdiäten GmbH, Soest, Germany) and reverse osmosis water, enriched with a mineral mixture Severyanka (St-Peterburg), were provided to the animals ad libitum.

We implanted 3D patches with and without cells in the abdominal aorta of 35 female SCID mice, 6–7 weeks old, average body weight 24 ± 4 g. The mice were divided into 2 groups. The experimental group included 23 mice implanted with patches pre-populated with cells. The control group consisted of 12 mice with implanted patches without cells. SCID mice were anesthetized with isoflurane and subsequently opened under standard sterile conditions. The abdominal aorta below the renal arteries up to the aortic bifurcation was isolated by a midline laparotomy approach. The aorta below the renal arteries and above the bifurcation was taken on holders (8/0 Prolene sutures). Next, the aorta was clamped with holders, and a longitudinal aortotomy of the abdominal aorta up to 0.5 mm in length was performed. Mitomycin C was not used during the aortotomy or patch implantation. Then a 0.5 × 0.5 mm patch was sewn into the aortotomy defect with 4 interrupted sutures using Premilene 10/0 suture. The patch implantation stages are shown in Supplementary Figure S1A. After confirmation of blood flow and hemostasis following removal of the clamp, the wound area was closed by layer-by-layer suturing of the anterior abdominal wall and skin with 6\0 threads. Prior, during, and after surgery, mice did not receive anticoagulants or antiplatelet agents.

Wall thickness of vascular patches before and 24 weeks after implantation was measured by electronic outside micrometer (0–25 mm, 0.001), Schut Geometrical Metrology, 908.750.

Abdominal aortic patency was assessed at 2, 4, 12, and 24 weeks after surgery. The study was carried out using magnetic resonance imaging (MRI) on an ultra-high-field tomograph BioSpec 117/16 USR (Bruker, Germany)—11.7 Tesla. F T1-weighted images were obtained by the FLASH method (Fast Low Angele SHot) with pulse sequence parameters (TE = 2.5 ms, TR = 200 ms), and an image (size 3 × 3 cm; on a matrix of 128 × 128 points; slice thickness 0.31 mm and distance between slices 0.33 mm).

Ultrasound examination was carried out at the same time. Under anesthesia with tiletamine (Zoletil 50, 5 mg/kg) and xylazine (XylaVet, 3 mg/kg) we assessed the permeability of the aorta, blood flow velocity in the area of the implanted patch after 2, 4, 12, and 24 weeks using a Vivid 4 apparatus with a linear vascular sensor (GE Medical Systems Israel Ltd., Ultrasound).

MMC-treated ECs and SMCs on plastic dishes were fixed with 4% FA (formaldehyde) (Sigma, #F8775-500 ML) for 10 min, permeabilized with 0.05% Triton X-100 (Sigma, #T8787-250 ML) for 10 min, and blocked with 1% BSA (bovine serum albumin) (VWR Life Science AMRESCO, #0332-100G).

Patch material at the 2nd, 4th, 12th, 24th weeks was collected by the mouse aorta transverse excision proximal and distal to the patch. After longitudinal dissection of the aorta posterior wall, the collected material was washed twice in PBS, fixed in 4% FA for 1 h, permeabilized with 1% Triton X-100 for 30 min, and blocked with 1% BSA.

The cells were stained with primary antibodies overnight at 4°C, washed with PBS, and incubated with secondary antibodies for 1 h at room temperature. The stained cells were analyzed with an inverted fluorescence microscope (Nikon Ti-E) using Nikon AR software. The cells in the tissue-engineered construct were analyzed using an LSM 780 confocal microscope (Zeiss) at the Common Facilities of Microscopic Analysis of Biological Objects, ICG SB RAS.

The following primary antibodies were used:

anti-human CD31 (DAKO, #M0823, 1:50), mouse IgG1; anti-mouse CD31 (Biolegend, #102502), rat IgG2a; anti-α-SMA (DAKO, #M0851, 1:50), mouse IgG2a; anti-smooth muscle myosin heavy chain 11 (Abcam, #ab82541, 1:500), rabbit polyclonal; anti-Von Willebrand factor (Abcam, #ab6994, 1:200), rabbit polyclonal; anti-fibronectin (Abcam, #ab6328, 1:200), mouse IgG1; anti-elastin (Abcam, #ab21610, 1:200), rabbit polyclonal; anti-collagen IV (Life Span, #LS-C79603, 1:200) mouse IgG1.

The following secondary antibodies were used:

Alexa Fluor 488 goat anti-mouse IgG1 (Life Technologies, #A21121, 1:400), Alexa Fluor 488 goat anti-rabbit IgG H + L (Life Technologies, #A11008, 1:400), Alexa Fluor 488 goat anti-rat IgG H + L (Life Technologies, #A11006, 1:400), Alexa Fluor 568 goat anti-mouse IgG1 (Life Technologies, #A21124, 1:400), Alexa Fluor 568 goat anti-mouse IgG2a (Life Technologies, #A21134, 1:400), Alexa Fluor 568 goat anti-rabbit IgG H + L (Life Technologies, #A11011, 1:400).

Representative images demonstrating negative control for autofluorescence with no primary antibody are shown in Supplementary Figure S2.

The patch material, explanted and fixed as for Immunofluorescent staining, was washed from the fixative at room temperature on a shaker in 125 mM glycine (Helicon, #Am-O167-0.5)/x1 PBS solution for 30 min, and then two times in PBS for 30 min. To restore the tissue structure, samples were incubated for 1 h in 20% sucrose at room temperature. They were then incubated overnight in 30% sucrose at + 4°C. The samples were embedded in O.C.T.-compound (Sakura, Tissue Tek, #4583), cooled to -22°C, 10 μm cryosections were made using a Cryostat HM 550 instrument on Superfrost Plus, Menzel-Glasses (Thermo Scientific, #J1800AMNZ). Histological preparations were stained with hematoxylin and eosin. Stained sections were examined by light microscopy using an AXIO Lab. A1 ZEISS microscope. The obtained microscopic images were used to assess the structure of the endothelial layer, the presence of cells in construct, patch fiber integrity, and the composition of the outer fibrous capsule.

A set of Mitomycin C (MMC) calibration solutions at various concentrations (1, 2, 5, 10, 50, 100, 250, and 500 ng/ml) was prepared from a standard stock solution of 2 mg/ml MMC by dilution with deionized water. The calibration solutions were analyzed via HPLC-MS/MS (MRM) to generate a standard curve. Various metabolites including MMC were extracted in H₂O: methanol (1: 4) from 5 EC and 5 SMC samples containing 1.3 to 2.5 million cells each. At first, 300 µl of H₂O was added to each tube with the cell sample followed by two cycles of freezing in liquid nitrogen and thawing in cold water. All tubes were vortexed between freeze-thaw cycles. Then 1.2 ml of cold (−20°C) methanol was added to each tube. The mixture was shaken vigorously for 10 min and left at −20°C for 30 min. Then the mixture was centrifuged at 16,100 g, + 4°C for 30 min to obtain a protein pellet and a methanol-H₂O layer. This upper aqueous layer was collected and lyophilized using an RVC 2-25 CD rotary vacuum concentrator (Martin Christ, Germany). The resulting protein-free extracts were used for LC-MS measurements.

The extracts for LC-MS analysis were re-dissolved in 100 µl of deionized water. The LC separation was performed on an Agilent 1,200 Series chromatograph (Agilent Technologies, United States) using Acclaim Polar Advantage II column (3 × 150 mm, 3 µm) with an Acclaim guard cartridge (Dionex, Germany). Solvent A consisted of 0.1% formic acid solution in H₂O, solvent B consisted of 0.1% formic acid solution in acetonitrile. The solvent B gradient was 10–55% (0–12 min), 55–95% (12–13 min), 95% (13–16 min), 95–10% (16–16.1 min), 10% (16.1–24 min); at the flow rate of 0.4 ml/min and the sample injection volume of 25 µl. The mass spectra were recorded at Agilent 6,410 Triple Quadrupole (Agilent Technologies, United States) in a positive mode. The instrument parameters were: electrospray voltage 3,500 V; nebulizer pressure 35 psi; dry gas flow 8 L/min; dry gas temperature 325°C, fragmentor voltage 100 V; collision gas nitrogen; collision energy 15 V, dwell time 500 ms, Delta EMV 400 V, mass transition m/z 335 to 242 for mitomycin C.

The proliferative activity of endothelial and smooth muscle cells non-treated and treated with MMC was assessed as previously described (Zakharova et al., 2017) by an XTT test with 2,3-bis- (2-methoxy-4-nitro-5- sulfophenyl) -2H-tetrazolium-5-carboxanilide (Roche, 11465015001, United States) according to the manufacturer’s instructions at 22 h after the reagents were added. The experiment was performed with exactly the same number of cells, which is 10⁴ per well in a 96-well plate. Before planting, cells were counted using Countess 2 Cell Counter (Thermo). For each point, the experiments were carried out in three replicates, with three wells of cells in each replication. Thus, nine values were analyzed for each point. XTT test values were recorded once a day within 8 days after cells were plated.

The viability of cells was assessed by staining with the Propidium Iodide/Annexin V*FITC kit (BioLegend, #640914) according to the manufacturer’s protocol and subsequent flow cytometry. The day before the analysis, cells were treated with MMC. Untreated cells were used as a control. The percentage of living, apoptotic, and necrotic cells was determined by flow cytometry for 10⁴ events using a FACS Canto II device (software—FACS Diva 7.0).

The viability of MMS-treated and untreated cells seeded on a PCL patch was assessed by staining with TMRM (tetramethylrhodamine, Thermo, #T668) and Annexin V*FITC (BioLegend, #640914) followed by life-imaging. TMRM is sequestered by active mitochondria and stains only living cells, while Annexin V*FITC identifies apoptotic cells. For the convenience of subsequent counting, cells were seeded on a patch at a low concentration (10⁴ cells per 1 cm² of the patch). In 5 days, cells were stained with TMRM and Annexin V*FITC according to the manufacturer’s protocols with modifications. First, the nuclei were stained with a NucBlue™ Live ReadyProbes™ Reagent (Thermo, #R37605) for 1 h under normal cell growth conditions in a CO₂ incubator.

After that, the patch was washed 1 time with PBS and placed in a 48-well with 500 μl Annexin V Binding Buffer (BioLegend, #640914), 1 μl of 50 μM TMRM, and 25 μl of Annexin V*FITC for 30 min at room temperature in the dark. Stained cells were visualized in the Annexin V Binding Buffer, using an LSM 780 NLO confocal microscope (Zeiss) at the Common Facilities of Microscopic Analysis of Biological Objects, ICG SB RAS. For each sample, the percentage of TMRM- and Annexin V-positive cells containing nuclei in 10 fields of view was calculated.

G-banding karyotyping was performed at passage 8 of EC and SMC as previously described (Malakhova et al., 2020). We analyzed 50 metaphase spreads using Nikon microscope TiE with Lucia Cytogenetics software.

To study the karyotype stability of MMC-treated cells, molecular karyotyping was carried out which is a study of copy number variations in genomic DNA by sequencing. Molecular karyotyping was performed by Genoanalytica (https://www.genoanalytica.ru). The analysis allows identifying microinsertions and microdeletions of 50,000 base pairs or longer. The karyotype was determined in 4 groups of samples, namely 1) endothelial cells treated with MMС (n = 3), 2) untreated endothelial cells (n = 3), 3) smooth muscle cells treated with MMС (n = 3), 4) untreated smooth muscle cells (n = 2). Genomic DNA was isolated by cell lysis followed by purification on glass fiber filters (QIAamp DNA Mini Kit reagents, Qiagen), then fragmented to 200 bp and used to prepare genomic libraries with NEB II Ultra reagents (New England Biolabs) for subsequent PCR amplification during 8 cycles. The nucleotide sequences were determined at the HiSeq 1,500 instrument (Illumina) with HiSeq Rapid SBS Kit v2 reagents (Illumina). The resulting 1,866,1932 reads are mapped to the reference human genome GRCh37/hg19 version. Rearrangements were identified as genome regions 5,000 kb or longer with abnormal read coverage determined by biased up or down coverage regression function at 1.5 times or more. Molecular karyotyping raw data is available at https://doi.org/10.6084/m9.figshare.17702513.

Statistical analysis of the number of detected CNVs was carried out taking into account the statistical weights of rearrangements in each group.

An in vivo Matrigel angiogenesis assay was conducted as previously described (Zakharova et al., 2017). Female immunodeficient (severe combined immunodeficiency, SCID) mice (6–8 weeks, 22–28 g) were used for experiments. Animals were kept in the SPF-vivarium of the ICG SB RAS (Novosibirsk, Russia). In the experiment, 17 mice were divided into groups as follows:

group 1. MMC treated ECs + SMCs, 5 mice; group 2. Untreated MMC ECs + SMCs, 5 mice; group 3. Untreated MMC ECs + SMCs stained with the intravital mitochondrial dye MitoTracker Deep Red FM (Invitrogen, M22426), 2 mice; group 4. Control—matrigel without cells, 5 mice.

Endothelial and smooth muscle cells (5 × 10⁵ each) in 50 μl of growth medium were mixed with 50 μl of Matrigel (Corning, #354234) immediately before injection. The total mixture in a volume of 100 μl was injected to mice subcutaneously into the abdominal region for 14 days. Isolectin B4 conjugated with biotin (Invitrogen, I21414) was injected into the Matrigel implant area 10 min before mouse sacrifice.

A Kodak In-Vivo Multispectral Imaging System was used to visualize cells stained with an in vivo mitochondrial dye (MitoTracker Deep Red FM).

Matrigel implants were removed, fixed in 4% PFA, frozen in Tissue-Tek O.C.T. Compound. Then, 10 μm cryosections were obtained from a vascularized implant on a Leica HM550 cryotome (Leica, Germany). The sections were incubated with a streptavidin-Cy3 fluorochrome conjugate (Sigma-Aldrich, S6402-1ML). Isolectin-positive vessels were imaged by a Nikon Ti-E microscope in 10 fields of view (FOV). Comparison between groups was carried out according to three parameters: the total vessels length, the average vessel length, and the total number of junctions in vasculature.

Imaged vasculature was assessed by the Angio Tool software (Zudaire et al., 2011), the results were compared by the Wilcoxon test with Bonferroni’s correction in R software version 3.5.1 (R Core Team, 2021). Representative images were taken using an LSM 780 NLO confocal microscope (Zeiss) at the Common Facilities of Microscopic Analysis of Biological Objects, ICG SB RAS.

The experiment was carried out at the Common Facilities Centre “SPF-vivarium of the ICG SB RAS” using female homozygous immunodeficient SCID mice at the age of 6–8 weeks. HEK293FT cells were used as a positive tumorigenic control (Stepanenko and Dmitrenko, 2015). Experimental mice groups are described above in Results.

We mixed 100 μl of Matrigel (BD Biosciences, United States) with 1 × 10⁶ cells in 100 μl of growth medium immediately before injection. We injected 200 μl of this mixture into mice subcutaneously into the abdominal region and left for 3 months. The positive control group was sacrificed earlier than the indicated period because tumor formation tends to necrotic changes.

Tissue samples taken at the site of injection were fixed in neutral (buffered, pH 7.6-7.8) 10% formalin and embedded in paraffin using standard techniques (Vaskova et al., 2015). Skin samples were oriented orthogonally to the epidermis. Series of 7–10 µm sections were mounted on Superfrost Plus Menzel-Glasses (Thermo Scientific, #J1800AMNZ) and stained as previously described (Vaskova et al., 2015).

Paraffin-embedded tissue sections (7–10 mm) were prepared and stained according to the previously described protocols (Vaskova et al., 2015).

Sample preparation, image capturing, and processing was performed at the Common Facilities Center of Microscopic Analysis of Biological Objects, ICG SB RAS (https://ckp.icgen.ru/ckpmabo/).

Statistical significance in value differences among experimental groups, we calculated by univariate analysis of variance (ANOVA) in Microsoft Excel (Microsoft Office 2013) software and Wilcoxon’s T-test in R version 3.5.1 (R Core Team, 2021). Bonferroni correction was applied for multiple comparisons. In the case of a binomial distribution for 10⁴ events, the significance of differences between classes was calculated using the chi-square method. Boxplots were built by RStudio version 3.5.1 (R Core Team, 2021).

We tested tissue-engineered constructs with EC and SMC from cardiac explants for the repair of vascular defects. We pre-seeded in vitro PCL patches with endothelial cells on one side and with smooth muscle cells on the other and then implanted them in 23 SCID mice of the experimental group with induced abdominal aortic defect (Supplementary Figure S1A). The control group consisted of 12 mice with the same defect, which was implanted with PCL patches without cells. Patch implantation details are presented in the video (Supplementary Video S1). The patency of the abdominal aorta was assessed at the 4th, 12th, and 24th weeks after surgery using ultrasound and MRI.

Doppler ultrasound scanning detected no significant differences in the linear blood flow velocity (BFV) between the experimental and control groups in the patch implantation area at 4th week of observation (Figure 2A). Similarly, we have not found any statistically significant differences in BFV after 24 weeks (Figure 2B).

Ultrasound triplex study showed that in all mice of the control and experimental groups at all control points the abdominal aorta patency was preserved, the blood flow rate was within the normal range (Figures 2A,B). Maintaining normal blood flow without increasing the BFV in the patch implantation area, allow us to exclude hemodynamically significant narrowing and/or occlusion of the abdominal aorta in mice.

No significant difference in the wall thickness was identified between the groups before implantation and 24 weeks after implantation (Figure 2C).

The abdominal aorta patency 4 weeks after surgery was confirmed by abdominal cavity MRI. In both the experimental and control groups, the abdominal aorta and iliac arteries were visualized without filling defects, which excluded hemodynamically significant stenosis or occlusions of the abdominal aorta. The absence of both aneurysms and ruptured patches indicates their mechanical strength (Figure 3).

Thus, at the hemodynamic examination level over 24 weeks, we did not find any advantages between PCL patches with and without cells.

The patch was implanted in such a way that the one side, covered with endothelial cells, was in contact with the bloodstream, while the other side, seeded with smooth muscle cells, was in contact with the surrounding tissue (Supplementary Figure S1B). After the patch was explanted and dissected from the aorta, a qualitative visual observation evidenced that about 80% of the patch from the inside was in contact with blood, while ∼20% was in contact with the arterial tissue.

In order for a tissue-engineered vascular patch to be able to integrate properly and perform its functions correctly, it requires to have a monolayer of endothelial cells and layers of smooth muscle cells. In addition, adequate patch functioning requires the cells within it to produce an appropriate extracellular matrix. We analyzed how these conditions are met for PCL patches explanted from the mouse abdominal aorta at weeks 2, 4, 12, and 24 of the experiment.

We find out that already 2 weeks after engraftment pre-seeded PLC patches on the inner surface contain not only human endothelial cells (hCD31), but also mouse endothelial cells (mCD31) (Figure 4). Human CD31-positive cells are detected at all observation points, while from the 12th to 24th week, their number drastically decreases. The number of murine CD31-positive cells increases by 12 weeks. At week 24, mouse CD31-positive cells almost completely cover the inner surface of the pre-seeded PLC patch. At all observation points on the outside surface of pre-seeded PLC patches, we also identified smooth muscle cells positive for the smooth muscle myosin heavy chain (SMMHC) marker (Figure 4).

During implantation even after 24 weeks, non-seeded PCL patches, in contrast to pre-seeded ones, lack a uniform monolayer of mouse endothelial and smooth muscle cells (Figure 5). In the absence of human EC and SMC, the density of mouse cells on the patches was higher near the edges adjacent to the recipient tissues. Usually, on the inner patch surfaces, along with or instead of endothelial cells, we detected either SMC or cells negative for vascular markers.

Extracellular matrix components as fibronectin and elastin attributable for ECs and SMCs respectively were detected homogeneously in the pre-seeded PLC patches at the 4 and 24 weeks (Figure 6; Supplementary Figure S8). By the 24th week, in the non-seeded PLC patches, extracellular matrix attributable to vessels are accumulated weakly, only in some areas. (Figure 6). Probably, not all cells that colonized to unseeded patch are capable of producing fibronectin and elastin. We propose that a weak production of extracellular matrix on the unseeded patches may be associated with uneven colonization of the patch with both EC and SMC, leading to the absence of a continuous monolayer. The unseeded patch contains no pre-formed niches for attracting the appropriate cell types. Colonization after implantation can also occur with nonspecific cells that do not produce the matrix in such an amount as this occurs on a patch pre-seeded with cells.

Histological staining with hematoxylin and eosin demonstrated that patches of the experimental group contained cell nuclei in the thickness of the PLC scaffold throughout the whole observation period at all control points (Figure 7). A monolayer of endotheliocyte-like cells was visualized throughout the experiment on the inner side of the patches that contact the bloodstream. From the second week, on the outer surface of the cell-seeded patches, we identified a fibrous capsule gradually thickening over time. At all stages of the experiment, the new capsule consists of dense fibrous tissue including a high content of smooth muscle and fibroblastic cells with microvessel formation areas.

PCL patches engrafted without cells, after explantation and staining with hematoxylin and eosin, also visualized cell nuclei in their thickness after 2 and 4 weeks. This indicates that the scaffold material does not interfere with the migration of cells inward (Figure 7). However, further histological examination showed a decrease in the number of nuclei from the 12th to the 24th week (Figure 7). Colonization of acellular vascular grafts with cells from circulating blood was described early (De Visscher et al., 2012; Sánchez et al., 2018; Smith et al., 2020).

Presumably, patches from the control group are colonized in vivo with blood cells, which are subsequently eliminated by apoptosis. In some non-seeded PCL patches, we noted areas with loss or even absence of polycaprolactone scaffold fibers, where both individual collagen fibers and fibrosis foci often occur. Our results suggest that in the absence of functional cells within the tissue-engineered construct, it tends to more rapid degradation, leading to a partial loss of the patch structural integrity. At all stages of the study, no intact monolayer of endotheliocyte-like cells was observed on the inner surface of non-seeded patches. In contrast to the experimental group, the external fibrous capsule in the control group is characterized by a thin layer of dense fibrous tissue with an insignificant content of fibroblast-like cells and the absence of a microvasculature (Figure 7).

The patches transplanted with pre-seeded EC and SMC after isolation from the aorta of experimental mice showed external surfaces with more pronounced local neovascularization, in contrast to the control patches, where new microvessels were less developed (Supplementary Figure S1B).

Thus, the patch pre-seeding with ECs and SMCs contributes to its integration into the recipient tissue, the formation of microvasculature, as well as the scaffold material integrity. After engraftment, cells in construct continue to produce an extracellular matrix, forming a highly specialized microenvironment that attracts corresponding recipient cells, which over time replace the transplanted cells.

In order to prevent the potential risk of malignancy from proliferating cells included in the tissue-engineered construct, endothelial and smooth muscle cells were treated with mitomycin C (MMC), which arrests cell division. We performed a series of tests to confirm that cells treated with MMC remain viable and functional despite the inability to divide. In particular, we showed that MMC-treated ECs and SMCs demonstrate normal viability >90%, retain specific markers, and ability to produce extracellular matrix as well as demonstrate an angiogenic potential in vivo.

Endothelial and smooth muscle cells untreated with MMS have a normal female karyotype and show no visible aberrations after examination of routine G-banding (Supplementary Figure S6A,B). Molecular karyotyping revealed 18 CNVs (microdeletions and microduplications) ranging from 50 to 230 kb (Supplementary Table S1), occurring from one to 11 times among 11 analyzed samples (EC MMC +, n = 3; EC MMC-, n = 3; SMC MMC +, n = 3; SMC MMC-, n = 2). We found 7 out of 18 CNVs in only MMC-untreated cells, 5 CNVs in only MMC-treated cells, and 6 CNVs in both MMC-untreated and treated cells (Supplementary Figure S6C). We showed that the number of CNVs did not significantly differ in the groups of treated and untreated cells (Supplementary Figure S6D). We also found no differences when comparing samples within endothelial or smooth muscle cells.

Using immunofluorescence staining, we assessed cell-specific properties in endothelial and smooth muscle cells treated with MMC. Endothelial cells treated with MMC demonstrate endothelial markers CD31 and von Willebrand factor, as well as produce extracellular matrix components type IV collagen and fibronectin (Figure 9A), retain the specific ability to form capillary-like 3D structures within matrigel layer (Figure 9B).

Smooth muscle cells treated with MMC showed properties attributable to normal SMCs as α-smooth muscle actin, SMMHC, and extracellular elastin production (Figure 10).

Thus, endothelial and smooth muscle cells treated with MMC do not lose specific markers and are able to accumulate extracellular matrix. Mitotically inactivated endothelial cells retain the ability to form capillary-like structures in vitro.

An important step in assessing the morphofunctional properties of endothelial and smooth muscle cells is to test their ability to induce angiogenesis in vivo. The standard test for angiogenic potential includes injecting cells within matrigel into the abdominal region of immunodeficient SCID mice. The angiogenic effect of the tested cells is due to the paracrine effect, i.e. angiogenic cytokines and growth factors secreted by cells and attracting the recipient’s cells (Yamahara et al., 2008; Yamahara and Itoh, 2009; Rufaihah et al., 2013). Angiogenic potential is assessed based on total length, branching, and other vasculature parameters in cryosections 14 days after injection. In our study, we compared the ability to stimulate neovascularization between actively dividing and mitotically inactivated cells in order to prove the functionality and potential suitability of ECs and SMCs treated with MMC for regenerative purposes. We injected a 1:1 mix of endothelial and smooth muscle cells since it is known that together they provide more pronounced angiogenesis when quantifying the total length and branching of capillaries (Yamahara and Itoh, 2009).

To show that throughout the experiment, the injected cells reside in the implantation area, we injected two mice with a mixture of endothelial and smooth muscle cells stained by the intravital mitochondrial dye MitoTracker Deep Red FM. We visualized viable fluorescently labeled cells in the injection area on the 14th day of the experiment (Figure 11A). Matrigel without cells was injected as a control.

Quantitative analysis of vasculature in cryosections 14 days after injection showed that the total vessel length was significantly higher in implants including matrigel with cells compared to matrigel only. At the same time, no significant difference was found between the groups that received the cells treated and untreated with MMC (Figure 11B).

Thus, ECs and SMCs after mitotic inactivation do not decrease their angiogenic potential in vivo. The combined injection of endothelial and smooth muscle cells increases the number of vascular structures by 2–3 times compared to the injection of matrigel without cells.

The tumorigenic effect from the mitotically inactivated endothelial and smooth muscle cells was evaluated on immunodeficient SCID mice. The experiment was carried out on 6 groups of mice injected with:

group 1. Endothelial cells untreated with MMC, 5 mice; group 2. Smooth muscle cells untreated with MMC, 4mice; group 3. Endothelial cells treated with MMC, 5 mice; group 4. MMC-treated smooth muscle cells, 4 mice; group 5. Control—matrigel without cells, 5 mice; group 6. Highly tumorigenic control—HEK 293FT cells, 5 mice.

The duration of the experiment in groups 1–5 was 12 weeks, in group 6 it was 19–21 days. The shortened experiment period in the highly tumorigenic control group is explained by the rapid formation of a significant tumor (Supplementary Figure S7).

Analysis of histological sections from the cell injection area showed the absence of any malignization in the groups of endothelial and smooth muscle cells treated and untreated with MMC, as well as in the control group of Matrigel without cells (Figure 12). In the group injected with HEK 293FT cells, tumorigenic foci were found at the injection site (Figure 12). Thus, EC and SMCs from cardiac explants appeared to be low- or non-tumorigenic per se, while mitotic inactivation seems to take away the remaining chance of malignization.