Mesenchymal Stromal Cells Improve Postoperative Cardiac Dysmaturation in a Juvenile Porcine Model

Abstract

Background

Mesenchymal stromal cells (MSCs) are known to protect tissues from inflammation and oxidative stress. This study aimed to evaluate the effects of cardiopulmonary bypass (CPB) and MSC treatment on cardiac maturation in a juvenile porcine model.

Methods

Two‐week‐old Yorkshire pigs were randomly assigned to 3 groups: (1) control (n=10), (2) CPB (n=10), and (3) CPB + MSC (n=15). MSCs (1 × 10⁷ cells/kg) were administered via CPB. Acute and long‐term responses were evaluated at 3 hours and 4 weeks post‐CPB. In addition to histological analysis, diffusion tensor imaging was used to assess cardiac microstructures.

Results

Oxidative stress marker 8‐hydroxy‐2ʹ‐deoxyguanosine (8‐OHdG) and Iba1⁺ (ionized calcium‐binding adapter molecule 1) monocytes/macrophages increased at 3 hours post‐CPB compared with controls. The increases in cardiac 8‐OHdG and monocytes/macrophages were still observed at 4 weeks, indicating prolonged oxidative/inflammatory stresses resulting from CPB. MSC delivery normalized both acute and prolonged oxidative/inflammatory insults. At 4 weeks, CPB caused reduced connexin 43 expression and increased cross‐sectional area of cardiomyocytes. Consistent with cellular changes, alterations in fractional anisotropy and helix and propagation angles were identified using diffusion tensor imaging 4 weeks post‐CPB. Additionally, differential tractography revealed impairments in fiber organization, demonstrating deleterious impacts of CPB on cellular and microstructural maturations of the developing heart. MSC delivery not only mitigated CPB‐induced oxidative/inflammatory stresses but also improved cellular and structural impairments following pediatric cardiac surgery.

Conclusions

CPB induces oxidative and inflammatory stresses, thereby causing cellular and structural changes in the developing heart. MSC treatment showed promise in improving CPB‐induced impairments of cardiac maturation in neonates and infants undergoing cardiac surgery.

Nonstandard Abbreviations and Acronyms

CPB

cardiopulmonary bypass

DTI

diffusion tensor imaging

FA

fractional anisotropy

HA

helix angle

MSC

mesenchymal stromal cell

PA

propagation angle

RD

radial diffusivity

Research Perspective.

What Is New?

• Cardiopulmonary bypass causes prolonged inflammation and oxidative stress, causing myocardial injury and disorganization in the developing heart.

• Mesenchymal stromal cell delivery via cardiopulmonary bypass mitigates injury to the post‐cardiopulmonary bypass myocardium and restores myocardial microstructures determined by cardiac diffusion tensor imaging.

What Question Should Be Addressed Next?

• Given that mesenchymal stromal cell delivery via cardiopulmonary bypass offers a clinically translatable approach to improve cardiac development, the safety and efficacy of this cell‐based treatment should be addressed in clinical trials involving infants with congenital heart disease who undergo cardiac surgery.

Cardiopulmonary bypass (CPB) is indispensable in cardiovascular surgery but is associated with the generation of reactive oxygen species, impaired antioxidant defense, and hemodynamic stress caused by blood exposure to nonendothelial surfaces. ¹ These factors contribute to systemic inflammatory response syndrome and severe oxidative stress, which are key drivers of postoperative cardiac complications such as arrhythmia and cardiac dysfunction.

Human cardiomyocytes transition to hypertrophic growth over several months to 2 years. ² Similar to humans, cardiac development in pigs is characterized by multinucleation and cardiomyocyte proliferation after postnatal day 15, followed by hypertrophic growth between 2 and 6 months. ² , ³ However, detailed cellular and structural development in neonatal hearts of the animal model is still largely unknown.

A mesenchymal stromal cell (MSC), which is a multipotent nonhematopoietic cell, possesses immunomodulatory and regenerative properties, including regulation of inflammation, antioxidant activity, antiremodeling properties, and activation of cell differentiation/cell cycle. ⁴ , ⁵ Our previous studies have demonstrated that MSCs via CPB provided neuroprotective effects. ⁶ Because MSCs migrate to the heart after the delivery through CPB, ⁶ we hypothesize that MSCs have the potential to alleviate CPB‐induced inflammation and oxidative stress in the developing cardiac tissues.

In the present study, we first examined the maturation of the porcine myocardium during the neonatal and infant periods by analyzing cell size, connexin 43 expression, proliferation rate, and the expressions of cardiac progenitor and metabolism markers. Diffusion tensor imaging (DTI) is an advanced magnetic resonance imaging (MRI) technique that provides the microstructural properties of the myocardium. ⁷ , ⁸ , ⁹ In addition to the cellular analyses, DTI was applied to determine microstructural maturation of the left ventricle (LV). We then evaluated the impact of cardiac surgery on the developing piglet heart, which closely resembles the human heart in anatomy and developmental profile. ³ Finally, the therapeutic effects of MSCs were assessed in the unique piglet model of CPB in vivo with both histological and MRI‐based approaches.

Methods

The data that support the findings of this study have been made publicly available at the Zenodo repository and can be accessed at 10.5281/zenodo.18010861.

Experimental Model

This study involved a total of 60 Yorkshire pigs (35 experimental piglets and 25 blood donor pigs) including 34 males and 26 females (Archer Farms, Darlington, MD). Two‐week‐old piglets were randomly assigned to 3 groups: control (no surgery, n=10), CPB (CPB for 150 minutes, n=10), and MSC (CPB with MSC administration during the rewarming, n=15). Controls were used to compare the effect of MSC treatment to the overall impact of cardiac surgery with CPB. CPB can provoke an immediate systemic inflammatory response with rapid elevations of circulating cytokines within 1to 3 hours of the completion. ¹⁰ Thus, the acute reactions of the developing heart were assessed at 3 hours post‐CPB (male, n=9; female, n=10), whereas the long‐term consequences were evaluated at 4 weeks following surgery (female, n=16).

CPB was established via ascending aortic perfusion and right atrial drainage. MSCs were manufactured from human bone marrow using the methods used for clinical trials at Children's National Hospital (Pro00011914; MeDCaP [Mesenchymal Stromal Cells Delivery Through Cardiopulmonary Bypass in Pediatric Cardiac Surgery]). Either 10 mL of phosphate‐buffered saline or 10 mL of phosphate‐buffered saline with MSCs (1 × 10⁷ cells/kg) was delivered through CPB (Figure 1A).

Figure 1. Study design and protocol.

A, CPB is performed for 150 minutes, and MSCs are administered at the rewarming phase. Assessments are conducted at 3 hours and 4 weeks post‐CPB. B, All DTI data sets are preprocessed by correcting motion‐related misalignment, generating Trk and NII files for further assessments. C and D, Differential tractography is generated to assess myocardial fiber connectivity. E, All DTI parameters are measured at 3 distinct areas on the anterior LV. F, Anterior LV is divided into 3 layers, and angles are measured at 3 distinct areas within each layer. G, Cellular analyses are performed using axial sections from the anterior LV, which are classified into 3 layers: endocardium, midmyocardium, and epicardium. 4W indicates 4 weeks; AD, axial diffusivity; CPB, cardiopulmonary bypass; DTI, diffusion tensor imaging; FA, fractional anisotropy; FF, full flow; HA, helix angle; Ht, hematocrit; L, left; LV, left ventricle; MD, mean diffusivity; MSC, mesenchymal stromal cell; NT, normal temperature; PA, propagation angle; R, right; RD, radial diffusivity; and RW, rewarming.

Cardiac positron emission tomography (PET) and MRI experiments were conducted in 3 and 17 piglets, respectively. We have previously found that intra‐arterial delivery via CPB uniforms MSCs distribution across multiple organs, including the brain, heart, and kidney using PET. ⁶ In the present study, PET was used to detect the MSCs labeled with fluorodeoxyglucose‐18 in the developing heart at 1 hour postdelivery (n=3). We have previously found that MSCs were distributed to the entire brain at 3 hours post‐CPB but disappeared in the T2‐weighted images at postoperative 4 weeks. ⁶ To validate these findings, we initially applied the same technique in 2 piglets at different time points (3 hours and 4 weeks post‐MSC delivery) to confirm the distribution pattern in the heart. Subsequently, ex vivo DTI (n=15 total) was performed on the 2‐week‐old control pre‐CPB hearts (n=3) and the hearts at post‐CPB week 4 (control, n=4; CPB, n=4; MSC, n=4) to evaluate microstructural alterations.

All DTI data sets were preprocessed using the Tolerably Obsessive Registration and Tensor Optimization Indolent Software Ensemble software package (https://tortoisedti.nichd.nih.gov/) and were corrected for motion‐related misalignment caused by frequency drifts, with appropriate rotations to the b‐matrix (Figure 1B). Trk and NII files were generated after Tolerably Obsessive Registration and Tensor Optimization Indolent Software Ensemble correction. Trk files were processed using Trackvis software (https://trackvis.org/) to quantify the tractography (Figure 1C). Differential tractography was generated by placing a 1.5‐μm spherical region of interest at 3 distinct locations within the anterior free wall of the LV to assess myocardial fiber connectivity in the anterior LV (Figure 1D). Fiber length, volume, number, and density were quantified for each region of interest, and the mean values were used for comparison.

All DTI parameters were measured at 3 distinct points on the anterior free wall of the LV, and the mean values were used for comparison (Figure 1E). Angle parameters, including helix angle (HA) and propagation angle (PA), were derived from NII files using DSI Studio (https://dsi‐studio.labsolver.org/) and custom Python code (Data S1). HA is defined as an angle of myocardial fibers relative to the circumferential axis in a long‐axis plane (Figure S1A), indicating helix rotation. ¹¹ PA is defined as an angle between 2 adjacent segments (primary eigenvectors) in a cardiomyocyte orientation streamline tractography (Figure S1B), reflecting fiber curvature. ⁸ For this analysis, the anterior LV was divided into 3 layers: endocardium (subendocardium), midmyocardium, and epicardium (subepicardium), and angles were measured at 3 distinct points within each layer (Figure 1F). The mean values of these measurements were used for comparison.

Cellular analyses using immunohistochemical methods were performed using axial sections from the anterior LV, which were classified into 3 layers (Figure 1G): endocardium (subendocardium), midmyocardium, and epicardium (subepicardium). Iba1 (ionized calcium‐binding adapter molecule 1) is a well‐established marker for residential microglia and infiltrating monocytes in the brain. Given that this marker can detect macrophages in various organs including cardiac macrophages, ¹² the present study used Iba1 to identify residential microglia and infiltrating monocytes in the developing porcine heart. Three areas in each layer were imaged using a confocal microscope at ×40 or ×80 magnification with 5‐μm Z levels, and the average numbers were used for quantification.

All experiments were performed in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study was approved by the Animal Care and Use Committee of the Children's National Hospital. Detailed methods are described in Data S1. Major resources are shown in Table S1.

Statistical Analysis

With small samples, choosing the form of analysis based on a preliminary Shapiro‐Wilk test can impair their calibration. ¹³ Thus, the present study used permutation versions of a t test and 1‐way ANOVA, followed by Bonferroni post hoc testing for 2‐group and 3‐group comparisons, respectively. Two‐way analyses were performed to test whether CPB‐induced insults or treatment effects differed across left ventricular regions. Given that the 2 time points represent biologically distinct phases following surgery, time was also included as a factor in the present study to assess the effect of treatment. Because exact permutation‐based 2‐way analysis is not readily implemented, the difference between the analytic/approximate approach and the permutation‐based approach was first assessed using 1‐way analysis of the same data. Because no substantial differences were observed (Table S2), a 2‐way repeated‐measures ANOVA with Bonferroni post hoc testing was used for the 2‐way analyses with region or time as a factor. Data distributions are presented as mean±SD or as bar graphs showing the mean with SD error bars and individual data points overlaid as dots. A P value of <0.05 was considered statistically significant. Permutation analyses were conducted using the coin package in R 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria). Other analyses were performed with the PRISM 10 software package (GraphPad Software, La Jolla, CA) and JMP Pro version 16 software (SAS Institute, Cary, NC).

Results

Pig Hearts Exhibit Cellular and Microstructural Maturation Between 2 and 6 Weeks Old

To understand developmental processes of cardiac maturation in the juvenile piglet model, we first characterized cellular and structural changes of the LV at 2 and 6 weeks old. When the size of myocytes was assessed using wheat germ agglutinin membrane staining, the cross‐sectional area of cardiomyocytes was significantly increased across all layers (Figure 2A and 2B). The ratio of connexin 43 expression to myosin heavy chain area remained unchanged in all 3 layers during this developmental period (Figure 2C and 2D). The number of NK2 homeobox5 (Nkx2.5)⁺ committed cardiomyocyte progenitors and GATA binding protein 4 (GATA4)⁺ cardiac progenitor cells significantly decreased (Figure 2E through 2H) as observed in the neonatal heart in rodents. ¹⁴ During this period, Antigen Kiel 67 (Ki67)⁺ cells accounted for 5% to 10% of the cell population in the piglet heart, whereas approximately 1% of cells were positive for Phospho‐histone H3 (PHH3) (Figure 2I through 2L).

Figure 2. Cellular and structural development of the postnatal porcine heart.

Representative immunofluorescence images and corresponding quantification of myocyte size (A and B), Cx43 (C and D), Nkx2.5 (E and F), GATA4 (G and H), Ki67 (I and J), PHH3 (K and L), and ATP5A (M and N) in the epicardium, midmyocardium, and endocardium at 2 and 6 weeks postnatally. DTI of the developing myocardium, including FA (O and P), MD (Q), RD (O and R), and AD (S). Immunostaining data are presented as mean±SD (n=5 per group) and analyzed by 2‐way ANOVA with Bonferroni post hoc tests. DTI data are shown as mean±SD (n=3 at 2 weeks, n=4 at 6 weeks) and analyzed by a permutation version of the Student t test. Scale bars=20 μm for MYH, Cx43, Nkx2.5, GATA4, and ATP5A, and 50 μm for Ki67 and PHH3. 2W indicates 2 weeks; 6W, 6 weeks; AD, axial diffusivity; ATP5A, alpha subunit of adenosine triphosphate synthase; cTnT, cardiac troponin T; Cx, connexin; DTI, diffusion tensor imaging; Endo, endocardium; Epi, epicardium; FA, fractional anisotropy; GATA4, GATA binding protein 4; Ki67, Antigen Kiel 67; MD, mean diffusivity; Mid, midmyocardium; MYH, myosin heavy chain; Nkx2.5, NK2 homeobox 5; PHH3, phospho‐histone H3; RD, radial diffusivity; and WGA, wheat germ agglutinin.

Alpha subunit of adenosine triphosphate (ATP) synthase (ATP5A) is a core component of mitochondrial ATP synthase and responsible for ATP production via oxidative phosphorylation. ¹⁵ When the ATP5A levels were assessed between the 2 age groups, we found that the expression was significantly increased during this developmental period (Figure 2M and 2N). The results indicate a postnatal shift toward increased mitochondrial oxidative metabolism, which is consistent with findings in postnatal cardiac maturation in the human infant. ¹⁶

When we applied DTI, a significant increase in fractional anisotropy (FA) and a decrease in radial diffusivity (RD) were revealed between 2 and 6 weeks old (Figure 2O through 2S). FA reflects directional coherence of water diffusion, whereas RD represents diffusion perpendicular to the fiber axis. ⁷ Thus, FA increases together with the reduction in RD observed in our analyses indicate restriction of transverse diffusion in the cardiac tissues (ie, more organized microstructure along 1 direction), demonstrating that cardiomyocytes become more aligned during this developmental period.

CPB Causes Oxidative Stress and Inflammation Responses Across All Layers of the Developing Heart

To assess the acute impact of cardiac surgery on the developing heart, we next compared cellular events between the control and CPB groups 3 hours post‐CPB. Our assessments included the accumulation of oxidative stress and inflammatory cells, alterations in cardiomyocyte size and gap junction expressions, and the presence of cardiomyocytes expressing cardiac progenitor markers.

When we assessed the oxidative and inflammatory responses, the 8‐hydroxy‐2ʹ‐deoxyguanosine (8‐OHdG)⁺ cells were significantly increased in the CPB group compared with the control group (Figure 3A and 3B). To understand the observed high 8‐OHdG accumulation following CPB, we also distinguish the acute oxidative stresses between DNA‐ and RNA‐derived damages. In our assessment, the number of 8‐OHdG⁺ cells on the DNase‐treated tissues was significantly lower compared with RNase‐treated hearts (Figure 3C and 3D), suggesting that the acute oxidative stresses resulting from CPB is primarily due to DNA damage. In the present study, Iba1 was used to identify monocyte–macrophage lineage cells. ¹² , ¹⁷ At 3 hours post‐CPB, the number of Iba1⁺ macrophages/monocytes was significantly increased, particularly in the epicardium (Figure 3E and 3F). When CD68⁺ (cluster of differentiation 68) cells, which are well‐known markers for human cardiac macrophages, ¹⁸ were assessed, we observed an overall increase in the distribution of Iba1⁺CD68⁺ cells relative to the total Iba1⁺ population (Figure 3G and 3H). The distribution pattern in the epicardium, however, remained similar during the acute stage following CPB‐induced injury. Together, these findings highlight that CPB induces acute oxidative and inflammatory stress in the developing heart.

Figure 3. Acute effects of CPB on the developing porcine heart.

Representative immunofluorescence images and corresponding quantification of 8‐OHdG (A and B), DNase‐ and RNase‐treated 8‐OHdG⁺ DAPI (C and D), total Iba1⁺ macrophages/monocytes (E and F), Iba1⁺CD68⁺ macrophages (G and H), and myocyte area delineated by WGA staining (I and J) in the epicardium, midmyocardium, and endocardium at 3 hours post‐CPB. Data are presented as mean±SD (n=5 per group) and analyzed using 2‐way ANOVA with Bonferroni post hoc tests. Scale bars=20 μm. CD68 indicates cluster of differentiation 68; CPB, cardiopulmonary bypass; Cont, control; Endo, endocardium; 8‐OHdG, 8‐hydroxy‐2ʹ‐deoxyguanosine; Epi, epicardium; Iba1, ionized calcium‐binding adapter molecule 1; Mid, midmyocardium; MYH, myosin heavy chain; and WGA, wheat germ agglutinin.

To assess the consequences of CPB‐induced oxidative and inflammatory stress, we examined caspase‐3 activation in cardiomyocytes. Consistent with our previous observations, caspase‐3⁺ cells were detected in the brain at 3 hours post‐CPB; however, no caspase‐3 activation was observed in the heart (Figure S2A). In this acute period, there were no changes in connexin 43⁺ expression (Figure S2B and S2C). In contrast, measurement of cell size revealed a significant increase in the cardiomyocyte cross‐sectional area across all 3 layers in the CPB group (Figure 3I and 3J). Given the acute timeline, it is likely that edematous changes contributed to the increased cardiomyocyte size observed 3 hours after CPB.

When we assessed the proliferation using markers for Ki67, there were no changes in the number of proliferating cardiomyocytes (Figure S2D and S2E). The number of Nkx2.5⁺ progenitor cells was also similar across all layers between the 2 groups (Figure S2F and S2G). On the other hand, the proportion of GATA4⁺ progenitor cells varied across the myocardial layer with a marked increase observed in the endocardium after CPB (Figure S2H and S2I), suggesting the possible acute endocardium damage as previously reported in other cardiac stress models. ¹⁴ , ¹⁹ , ²⁰

Prolonged Oxidative Stress and Inflammatory Response Induced by CPB Alter Developmental Processes of the Neonatal Hearts

We next assessed the developing porcine heart at 4 weeks post‐CPB. Notably, 8‐OHdG⁺ cells remained elevated across 3 myocardium layers in the CPB group (Figure 4A and 4B). Although DNA‐derived oxidation was more prominent at 3 hours post‐CPB (Figure 3C and 3D), the 8‐OHdG accumulation between DNase‐ and RNase‐treated myocytes became similar at 4 weeks post‐CPB (Figure 4C and 4D). These findings suggest that CPB‐induced oxidative stress to the developing cardiomyocyte may shift from DNA oxidation to RNA oxidation over time. Consistent with the prolonged oxidative stress, the number of Iba1⁺ macrophages/monocytes remained higher in the heart exposed to CPB (Figure 4E and 4F). At post‐CPB week 4, we observed a similar distribution between the Iba1⁺CD68⁻ and Iba1⁺CD68⁺ populations, suggesting that both infiltrating monocytes and tissue‐resident macrophages contributed to the observed prolonged increases of the monocyte/macrophage lineage cells in the developing heart.

Figure 4. Cellular changes at 4 weeks post‐CPB.

Representative immunofluorescence images and corresponding quantification of 8‐OHdG (A and B), DNase‐ and RNase‐treated 8‐OHdG⁺ cells (C and D), total Iba1⁺ macrophages/monocytes (E and F), Iba1⁺CD68⁺ macrophages (G and H), myocyte area delineated by WGA (I and J), Cx43 (K and L), proliferating cell markers Ki67⁺ (M and N) and PHH3⁺ (O and P), and cardiac progenitor markers Nkx2.5 (Q and R) and GATA4 (S and T) in the epicardium, midmyocardium, and endocardium at 4 weeks post‐CPB. Data are presented as mean±SD (n=5 per group) and analyzed using 2‐way ANOVA with Bonferroni post hoc tests. Scale bars=20 μm for 8‐OHdG, Iba1, MYH, Cx43, Nkx2.5, and GATA4, and 50 μm for Ki67 and PHH3. CPB indicates cardiopulmonary bypass; cTnT, cardiac troponin T; Cont, control; Cx, connexin; 8‐OHdG, 8‐hydroxy‐2ʹ‐deoxyguanosine; Endo, endocardium; Epi, epicardium; GATA4, GATA binding protein 4; Iba1, ionized calcium‐binding adapter molecule 1; Ki67, Antigen Kiel 67; Mid, midmyocardium; MYH, myosin heavy chain; Nkx2.5, NK2 homeobox 5; PHH3, phospho‐histone H3; and WGA, wheat germ agglutinin.

We observed a developmental increase of the cardiomyocyte size in the porcine heart (Figure 2A and 2B). The cross‐sectional area of myocytes exposed to CPB was further increased in all layers compared with controls (Figure 4I and 4J). Although the connexin 43 expression was similar between the 2 groups at 3 hours post‐CPB (Figure S2B and S2C), CPB impaired gap junction expression by postoperative 4 weeks (Figure 4K and 4L). In addition, the numbers of both Ki67⁺ and PHH3 proliferating cells significantly decreased across the layers after CPB (Figure 4M through 4P). Consistent with the findings, the number of Nkx2.5⁺ committed cardiomyocyte progenitors was lower in the CPB group compared with the controls (Figure 4Q and 4R). Acute increase of GATA4⁺ cells was displayed in the endocardium at 3 hours after CPB (Figure S2H and S2I). We also observed the same cellular event at postoperative week 4 (Figure 4S and 4T). Given the significant impact of oxidative injury and inflammation on cardiac development, ²¹ our data indicate that prolonged oxidative and inflammatory stresses induced by CPB alter the size and gap junction of the developing heart and shift the proliferation and maturation status.

CPB Induces Significant Alterations in Cardiac Fiber Architecture and Orientation and Disrupts Structural Connectivity With the Developing LV

To understand the effect of CPB‐induced damage on the structural organization within the heart, DTI was used. During normal cardiac development, an increase of FA and a decrease of RD were observed (Figure 2K). Although there were no differences in mean diffusivity and axial diffusivity, we observed a further increase of FA and a decrease of RD after CPB (Figure 5A through 5F), indicating more restricted tissue alignment in the transverse direction. The structural changes determined by DTI are consistent with hypertrophic remodeling but not tissue edema. ²²

Figure 5. Microstructural changes at 4 weeks post‐CPB.

A through E, DTI data, including FA, MD, RD, and AD. F and G, HA. H and I, PA. J, Tractography. K, Zoomed DTI in the anterior LV. Data are shown as mean±SD (n=4 each). P values were determined by a permutation version of the Student t test or 2‐way ANOVA with Bonferroni corrections. AD indicates axial diffusivity; Cont, control; CPB, cardiopulmonary bypass; DTI, diffusion tensor imaging; Endo, endocardium; Epi, epicardium; FA, fractional anisotropy; HA, helix angle; LV, left ventricle; MD, mean diffusivity; Mid, midmyocardium; PA, propagation angle; and RD, radial diffusivity.

To assess the myofiber architecture and microstructural orientation, HA and PA were used. In the control group, HA demonstrated the typical transmural gradient, including a left‐handed (counterclockwise) helical pattern (negative values) in the epicardium, a right‐handed (clockwise) pattern (positive values) in the endocardium, and intermediate values in the midmyocardium. In the CPB group, HA values were significantly increased across all myocardial layers (Figure 5G and 5H), suggesting a loss of normal helix rotation after CPB. PA in all layers was also increased in the CPB group (Figure 5I and 5J), indicative of CPB‐induced myofiber curvature.

To complement the DTI analyses, we further performed tractography‐based analyses. We first generated representative heart tractography to visualize the fiber direction and connectivity (Figure 5K). Overall, fiber alignment was disorganized in the CPB group compared with the control group (Figure 5K, upper panel). In the coronal views, we observed an alteration of the fiber direction after CPB, as evidenced by an increase in the red signals (Figure 5K). A detailed analysis of the anterior LV further revealed that CPB disrupted the smooth and continuous connectivity in the normal ventricle microstructures (Figure 5L). When we performed differential tractography to assess fibers connected to those in the anterior LV (Figure 5M), significant reductions in the number, volume, and density of myocardial fibers were observed (Figure 5N through 5Q), indicating that CPB impairs the structural connectivity with the anterior LV wall. Altogether, CPB not only altered myofiber architecture and microstructural orientation but also impaired the structural connectivity within the LV.

Intra‐Arterially Administered MSCs Are Initially Distributed to the Whole Heart but Are Not Retained Beyond 4 Weeks Post‐CPB

We observed CPB‐induced cellular and structural impairments in the developing heart. To assess the therapeutic effect of MSCs, we first quantified the distribution of MSCs within the heart using PET 1 hour post‐MSC delivery. Results from the PET study indicated that intra‐arterial delivery via CPB distributed MSCs equally to the 5 heart regions (Figure S3A and S3B). To further determine the distribution of MSCs, we assessed the superparamagnetic iron oxide particles on MRI scans by comparing them at 3 hours and 4 weeks post‐CPB. At 3 hours post‐CPB, T2‐weighted imaging showed a diffuse distribution of hypointense voxels (Figure S3C), indicative of superparamagnetic iron oxide‐labeled MSCs across the entire heart. By 4 weeks post‐CPB, these hypointense voxels were decreased (Figure S3C). These findings suggest that although intra‐arterially administered MSCs are initially distributed to the whole heart, there are little long‐term residual MSCs beyond 4 weeks post‐CPB, which is consistent with our previous findings observed in the brain. ⁶

MSCs Delivery Through CPB Inhibits the Disruption of Cardiac Maturation After Pediatric CPB Surgery

There were no differences in preoperative conditions and major intraoperative parameters between the CPB and MSC groups (Tables S3 and S4). Although there were differences in eosinophil numbers and creatine and calcium levels 3 hours post‐CPB between the 2 groups (Table S5), no differences were found 4 weeks after CPB in blood and urine examinations compared with the control group (Table S6).

CPB caused oxidative stress and inflammation persisted for up to 4 weeks in the developing heart (Figure 4A through 4H). Notably, MSC treatment inhibited 8‐OHdG expression (Figure 6A through 6C; Figure S4A and S4B) and reduced the number of Iba1⁺ macrophages/monocytes across all layers of myocardium (Figure 6D through 6F; Figure S4C and S4D) at both 3 hours and 4 weeks post‐CPB. Consistent with findings, the number of Iba1⁺CD68⁺ cells in the MSC groups was significantly lower compared with the CPB group at both time points (Figure S4E through S4I). Notably, MSC delivery via CPB normalized 8‐OHdG⁺ cell numbers and the number of both Iba⁺ and Iba⁺CD68⁺ cells across 3 layers (Figure 6A through 6F; Figure S4A through S4I), indicating MSC's protective effects against oxidative and inflammatory stresses.

Figure 6. Effect of MSCs on cellular development of the juvenile porcine heart.

Representative immunofluorescence images and corresponding quantification of 8‐OHdG (A–C), total Iba1⁺ macrophages/monocytes (D–F), myocyte area delineated by WGA (G–I), Cx43 (J and K), Nkx2.5⁺ progenitors (L and M), GATA4⁺ progenitors (N–P), Ki67⁺ proliferating cells (Q and R), and PHH3⁺ cells (S and T). Data are mean±SD (n=5 per group), analyzed by 2‐way ANOVA with Bonferroni post hoc tests. Scale bars=20 μm for 8‐OHdG, Iba1, WGA, Cx43, Nkx2.5, and GATA4, and 50 μm for Ki67 and PHH3. 3h indicates 3 hours; 4W, 4 weeks; Cont, control; CPB, cardiopulmonary bypass; cTnT, cardiac troponin T; Cx, connexin; 8‐OHdG, 8‐hydroxy‐2ʹ‐deoxyguanosine; Endo, endocardium; Epi, epicardium; GATA4, GATA binding protein 4; Iba1, ionized calcium‐binding adapter molecule 1; Ki67, Antigen Kiel 67; Mid, midmyocardium; MSC, mesenchymal stromal cell; MYH, myosin heavy chain; Nkx2.5, NK2 homeobox 5; PHH3, phospho‐histone H3; and WGA, wheat germ agglutinin.

Consistent with the reduction in CPB‐induced oxidative and inflammatory stresses, the cross‐sectional area of cardiomyocytes was normalized at both acute and chronic time points following MSC delivery (Figure 6G through 6I; Figure S4J and S4K). The increases of cell size observed in normal developmental conditions were also preserved in the MSC group (Figure 6G through 6I; Figure S4J and S4K). These results demonstrate that MSC treatment inhibits CPB‐induced acute edema and prevents pathological hypertrophy in developing cardiomyocytes. CPB caused the reduction of the expression of connexin 43 at 4 weeks postoperative (Figure 4K and 4L). Connexin 43 expression was significantly increased across the myocardium after MSC treatment (Figure 6J and 6K). These findings support the role of MSCs in preserving myocardial development in the post‐CPB heart.

CPB resulted in a significant reduction of Nkx2.5⁺ committed progenitors across the myocardium (Figure 4Q and 4R) and unique activation of GATA4 in the endocardium (Figure 4M and 4N). Our studies observed higher Nkx2.5⁺ progenitor numbers after MSC delivery compared with the CPB group (Figure 6L and 6M). Consistent with these results, the number of GATA4⁺ cells was normalized at both time points after MSC treatment (Figure 6N through 6P; Figure S4L and S4M). On the other hand, the number of Ki67⁺ and PHH3⁺ cells remained lower in both CPB and MSC groups compared with the controls (Figure 6Q through 6T). Altogether, the results demonstrate that although they show the limited capability to rescue cell proliferation, MSCs delivered through CPB into the developing heart inhibit the disruption of cardiac maturation after pediatric cardiac surgery.

MSC Administration Restores CPB‐Induced Microstructural Alterations of the Developing Heart

Finally, we assessed the effects of MSC treatment on significant microstructural alterations seen after CPB. CPB induced FA increase together with a reduction in RD (Figure 5A through D). Both DTI parameters were normalized after MSC treatment (Figure 7A through 7E; Figure S5A and S5B). The left‐handed helical pattern (negative values) in the normal epicardium determined by HA was altered after CPB (Figure 5G and 5H). MSCs restored the loss of helix rotation in the epicardium induced by CPB (Figure 7F and 7G; Figure S5C). Similarly, CPB‐induced fiber curvatures assessed by PA were inhibited across the myocardium after MSC delivery (Figure 7H and 7I and Figure S5D). In the representative heart tractography, disrupted fiber alignments after CPB were still seen in the heart in the MSC group (Figure S5E). On the other hand, the color code shown in axial and coronal views revealed that MSCs preserved the fiber orientation of the LV, similar to that of the normal heart. Furthermore, fiber alignment inside the heart demonstrated smooth and continuous connectivity in the MSC group as seen in the control group (Figure S5F). In addition, quantification in differential tractography revealed that CPB‐induced reductions in the number, density, and volume of myocardial fibers were improved in the MSC group (Figure 7J through 7N), indicating that MSC delivery via CPB preserves impairments of myocardial connectivity after pediatric CPB surgery. Finally, we did not observe any signs of stroke or ischemic damage in our DTI analysis.

Figure 7. Effect of MSCs on microstructural development of the juvenile porcine heart.

A through E, DTI data included FA, MD, RD, and AD. F and G, HA in the anterior LV. H and I, PA in the anterior LV. J through N, Differential tractography and fiber quantification. Data are shown as mean±SD (n=4 each). P values were determined by a permutation version of 1‐way ANOVA for DTI and 2‐way ANOVA for HA/PA with Bonferroni comparisons. AD indicates axial diffusivity; CPB, cardiopulmonary bypass; DTI, diffusion tensor imaging; Endo, endocardium; Epi, epicardium; FA, fractional anisotropy; HA, helix angle; LV, left ventricle; MD, mean diffusivity; Mid, midmyocardium; MSC, mesenchymal stromal cell; PA, propagation angle; and RD, radial diffusivity.

Discussion

The present study using the unique piglet model revealed that CPB induces persistent oxidative stress and cardiac inflammation and leads to cellular and microstructural alterations in the developing heart. Notably, MSC delivery via CPB normalized both acute and prolonged oxidative and inflammatory insults and inhibited CPB‐induced cellular changes. Consistent with the cellular changes, structural alterations due to CPB were improved after MSC administration.

Piglets Are a Clinically Relevant Model for Studying Postnatal Cardiac Development

Species‐specific differences in physiological and developmental timelines can significantly influence postnatal cardiac growth and the transition to hypertrophy. ² , ³ In humans, cardiomyocyte nucleation continues after birth, and hypertrophic growth begins a few months after birth. ²³ , ²⁴ In our study, we observed comparable developmental dynamics in piglets between 2 and 6 weeks old, including a physiological increase in cardiomyocyte cross‐sectional area and a postnatal shift toward increased mitochondrial oxidative metabolism. In addition, results from our cardiac DTI demonstrate for the first time an increase in FA and a decrease in RD of the developing porcine heart, consistent with findings from human neonatal studies. ²⁵ Taken together, these data indicate that piglets provide a translationally valuable platform for understanding human cardiac development and injury in neonates and young infants.

CPB Induced Prolonged Oxidative Stress and Inflammation, Leading to Cellular and Microstructural Changes

CPB imposes a broad range of unique stressors on developing organs including dynamic changes in temperatures, hemodilution, nonpulsatile blood flow, altered oxygenation, and inflammatory reactions. ²⁶ In pediatric patients undergoing CPB, oxidative stress is initiated almost immediately upon the establishment of CPB with rapid elevations of circulating cytokines. ¹⁰ This sequence was mirrored in our study by an increase in 8‐OHdG expression and the number of Iba1⁺ macrophages/monocytes as early as 3 hours post‐CPB, although classical monocyte infiltration following focal ischemic cardiac injury typically peaks days after the insult. ²⁷ The broad range of specific and unique pathological stimuli associated with CPB likely contributes to the accelerated monocyte recruitment observed at this early stage after cardiac surgery.

Notably, we found that the increases in 8‐OHdG expression and macrophage/monocyte numbers persist up to 4 weeks after surgery. The cytoplasmic localization of 8‐OHdG observed in our studies suggests mitochondrial damage following CPB, as seen in other disease models. ²⁸ , ²⁹ Our results further suggest that CPB‐induced oxidative stress to the developing cardiomyocyte may shift from DNA oxidation to RNA oxidation over time. In the present study, we observed CPB‐induced pathological cardiomyocyte hypertrophy, as evidenced by cardiac DTI. Given that hypertrophy has been previously linked to oxidative stress and inflammation including the activation of macrophage/monocytes, ³⁰ , ³¹ it is plausible that the sustained oxidative stress and inflammation resulting from cardiac surgery contribute to cardiomyocyte hypertrophy in the developing heart.

HA and PA are sensitive markers for microstructural changes and remodeling. ⁷ , ³² , ³³ In our study, both parameters were significantly reduced following CPB, indicating a loss of helical organization and impaired myocardial coordination. Given that these parameters are key determinants of effective cardiac contraction, such as fiber orientation and torsional dynamics, ⁷ , ³² , ³³ these microstructural alterations may reflect the cardiac dysfunction after pediatric cardiac surgery. It has been well documented that altered cardiac maturation in the early developmental periods significantly contributes to the development of heart failure in later life. ³⁴ , ³⁵ Further studies are required to understand the mechanism underlying sustained oxidative stress and inflammation following pediatric cardiac surgery and to elucidate the long‐term impact of the pathological cardiac remodeling observed in the present study.

Therapeutic Effects of MSCs on Cellular and Microstructural Alterations Post‐CPB

MSCs possess immunomodulatory and regenerative properties, making them promising candidates for therapeutic interventions targeting inflammation, oxidative stress, and pathological remodeling. ⁴ , ⁵ , ³⁶ In our study, MSCs exhibited both anti‐inflammatory and antioxidant effects on injured cardiomyocytes at both early and late phases following CPB. In addition to a reduction in the proinflammatory cytokine interferon‐gamma and an increase of the anti‐inflammatory cytokine IL (interleukin)‐10, our previous study found that MSC delivery via CPB increased plasma IL‐1b and IL‐8 levels. ³⁷ This cytokine profile suggests possible immunosuppressive reactions occurring within an inflammatory microenvironment. Furthermore, the correlation between the number of cerebral microglia and plasma cytokine levels during CPB observed in our previous work ³⁷ suggests an interaction between systemic inflammation and tissue macrophage responses. Given the rapid elevations of circulating cytokines post‐CPB, ¹ the timing of MSC administration in our protocol was aligned with the inflammatory cascade, allowing therapeutic action during the critical window.

MSCs also attenuated myocyte hypertrophy and preserved gap junctions, while enhancing the expression of key cardiac transcription factors. These results support the role of MSCs in remodeling by inhibiting inflammatory and oxidative signaling pathway, ³⁶ preserving gap junctions, ³⁸ and enhancing cardiomyogenic differentiation. ²⁰ In addition, our study showed that MSC treatment restored the CPB‐induced microstructural alterations determined by DTI. The normalization of helical pattern and fiber curvatures of ventricular tissue after MSC treatment is likely to inhibit pathological remodeling and promote functional recovery of the post‐CPB heart. ³² , ³⁹ , ⁴⁰ Together, these findings suggest that MSCs not only mitigate cellular stress responses but also preserve and restore the cellular and structural impairments, likely enhancing long‐term cardiac maturation after surgical repair of their cardiac anomalies.

We previously demonstrated neuroprotective effects of MSCs following CPB in the developing brain. ⁶ Our transcriptomic analyses suggested that microRNAs derived from MSC exosomes are key drivers of these observed neuroprotective functions ⁶ Although the regenerative capacity of MSCs was overstated by previous claims that they behave as pluripotent stem cells capable of replacing lost cardiomyocytes in vivo, their unique trophic and paracrine properties have been well recognized. ⁴ , ⁵ In the present study, we confirmed that intra‐arterially administered MSCs are initially distributed to the whole heart but are not retained beyond 4 weeks post‐CPB. These results support paracrine functions playing a major role in MSC‐mediated cardioprotection. Chemokine‐guided extravasation and migration of MSCs toward injury sites have been well documented. ⁴¹ Given the protective capacities of MSCs observed in both the developing brain and heart, it is possible that MSC delivery during cardiac surgery could inhibit deleterious impacts of CPB across multiple organ systems in neonates and young infants.

Limitations

This study has several limitations. Although Yorkshire pigs provide a valuable translational model due to their anatomical and physiological similarity to humans, species‐specific differences remain. Thus, results should be carefully interpreted. Some analyses, particularly assessments using cardiac PET and superparamagnetic iron oxide, were performed with small sample sizes, limiting statistical power. Similarly, a sham surgery group was not added in this study using piglets. Thus, part of the observed changes might be attributable to anesthesia or surgical incision. Our immunohistology and MRI analyses were focused on the anterior wall of the LV due to its importance in overall cardiac function. Additional regional assessments may be necessary to fully capture the regional heterogeneity of the developing heart. In our study, human bone marrow‐derived MSCs, which have been used for our clinical trial, were delivered in our porcine model. Although the absent to low major histocompatibility class II antigen expression in MSCs contributes to their hypoimmunogenicity after xenotransplantation, ⁴² , ⁴³ potential immunologic responses and interspecies compatibility may influence the results obtained from the present study. Additionally, MSCs were administered as a single dose via CPB. The effects of repeated or delayed administration have not been investigated yet. Although our study demonstrated well‐characterized structural alterations using DTI, it lacked functional assessments of myocardial performance. Therefore, further studies incorporating functional assays such as echocardiographic measurements are necessary to determine the impact on myocardial performance. Iba1 protein is evenly distributed in the cytoplasm of cell bodies and processes ¹⁰ , ¹⁷ Therefore, as demonstrated in studies of cerebral microglia, 3‐dimensional reconstructions and dynamic morphological assessments using Iba1 will likely enhance our understanding of how the monocyte–macrophage lineage contributes to inflammatory reactions in the cardiac system. However, the immune toolkit for the porcine model is not as robust as those available for rodents or humans. Consequently, although we used CD68, a well‐established marker for human cardiac macrophages, ¹⁸ our staining strategy may not have differentiated between infiltrating monocytes and tissue‐resident macrophages in the porcine heart. Furthermore, due to limited tools (Table S1), we were unable to assess neutrophil infiltration in the cardiac tissues following CPB. Given the critical influence of immune cells on cardiac repair and arrhythmia, ⁴⁴ , ⁴⁵ improving immune cell identification strategy in the porcine model will be essential to elucidating how immune responses to cardiac surgery impact cardiac maturation in children with congenital heart disease.

Conclusions

The present study using a translational piglet model of CPB identified prolonged oxidative stress and inflammation in the developing heart, alongside significant cellular and structural alterations following pediatric cardiac surgery. MSC delivery via CPB shows promise as the protective intervention that potentially facilitates cardiac maturation in children with congenital heart disease.

Sources of Funding

This work was supported by National Institutes of Health grants R01HL139712 (N.I.), R01HL146670 (N.I.), R01NS123442 (T.‐W.T.), and by the Office of the Assistant Secretary of Defense for Health Affairs through the Peer Reviewed Medical Research Program under award number W81XWH2010199 (N.I.) and HT94242510760 (N.I.). The Howard University Research Center for Minority Health Disparities imaging core was supported by National Institutes of Health grant 2U54MD007597.

Disclosures

None.

Supporting information

Tables S1–S6

Figures S1–S5

Supplemental Methods

References 46–50

Author_Checklist