Noninvasive Assessment of an Engineered Bioactive Graft in Myocardial Infarction: Impact on Cardiac Function and Scar Healing

Abstract Cardiac tissue engineering, which combines cells and biomaterials, is promising for limiting the sequelae of myocardial infarction (MI). We assessed myocardial function and scar evolution after implanting an engineered bioactive impedance graft (EBIG) in a swine MI model. The EBIG comprises a scaffold of decellularized human pericardium, green fluorescent protein‐labeled porcine adipose tissue‐derived progenitor cells (pATPCs), and a customized‐design electrical impedance spectroscopy (EIS) monitoring system. Cardiac function was evaluated noninvasively by using magnetic resonance imaging (MRI). Scar healing was evaluated by using the EIS system within the implanted graft. Additionally, infarct size, fibrosis, and inflammation were explored by histopathology. Upon sacrifice 1 month after the intervention, MRI detected a significant improvement in left ventricular ejection fraction (7.5% ± 4.9% vs. 1.4% ± 3.7%; p = .038) and stroke volume (11.5 ± 5.9 ml vs. 3 ± 4.5 ml; p = .019) in EBIG‐treated animals. Noninvasive EIS data analysis showed differences in both impedance magnitude ratio (−0.02 ± 0.04 per day vs. −0.48 ± 0.07 per day; p = .002) and phase angle slope (−0.18° ± 0.24° per day vs. −3.52° ± 0.84° per day; p = .004) in EBIG compared with control animals. Moreover, in EBIG‐treated animals, the infarct size was 48% smaller (3.4% ± 0.6% vs. 6.5% ± 1%; p = .015), less inflammation was found by means of CD25+ lymphocytes (0.65 ± 0.12 vs. 1.26 ± 0.2; p = .006), and a lower collagen I/III ratio was detected (0.49 ± 0.06 vs. 1.66 ± 0.5; p = .019). An EBIG composed of acellular pericardium refilled with pATPCs significantly reduced infarct size and improved cardiac function in a preclinical model of MI. Noninvasive EIS monitoring was useful for tracking differential scar healing in EBIG‐treated animals, which was confirmed by less inflammation and altered collagen deposit. Stem Cells Translational Medicine 2017;6:647–655


INTRODUCTION
Cardiac tissue engineering, which combines cells and biomaterials, is a promising therapy after myocardial infarction (MI) [1]. The main goal of this new approach includes both the restoration of damaged tissue and the recovery of cardiac function to limit or prevent adverse ventricular remodeling and end-stage heart failure. To this end, different biomaterials are being tested (i.e., collagen, alginate, hydrogel, and extracellular matrices) to provide a stable support for cell delivery in close contact with the damaged tissue [2]. Microscopic and ultrastructural scaffold topography is key for cellular homing and migration to the target tissue [3,4]. Decellularized tissues offer a natural microenvironment, driving cellular attachment, survival, migration, proliferation, and differentiation [ Prat-Vidal et al. previously reported preliminary data on a novel engineered bioactive impedance graft (EBIG) comprising a scaffold of decellularized human pericardium, green fluorescent protein (GFP)-labeled porcine adipose tissue-derived progenitor cells (pATPCs), and an electrical impedance spectroscopy (EIS) monitoring system [7]. The functional impact of EBIG in a preclinical model of MI and scar maturation characteristics remains unknown. Accordingly, cardiac function was evaluated in a noninvasive manner by using magnetic resonance imaging (MRI). Additionally, scar healing was evaluated by using a custom-designed EIS system incorporated within the implanted graft in the acute MI swine model.

Experimental Design
After a left lateral thoracotomy, an MI was induced via doubleligation of the first marginal branch of the circumflex artery, 1.5 cm distally from the atrioventricular groove (Prolene 5/0 W-8556 12-S, Ethicon, Somerville, NJ, http://www.ethicon.com) [8]. After 30 minutes, the EBIG was attached over the infarcted tissue with 0.1-0.2 ml of surgical glue (Glubran 2, CardioLink, Barcelona, Spain, http://www.cardiolink.es). Finally, the animals were allowed to recover and housed for 1 month before sacrifice. In order to ensure that all animals had a similar MI, we analyzed circulating troponin I (cTnI) and creatine kinase in serum samples collected by jugular venipuncture at baseline and 2 hours after MI induction. Cardiac biomarkers were measured in a fluorometric immunoassay analyzer (AQT90 FLEX, Radiometer Medical ApS, Brønshøj, Denmark, http://www.radiometer.com).
Twenty-six cross-bred Landrace 3 large white swine (30.2 6 3.6 kg) were randomly distributed into three groups ( Fig. 1): (a) control arm (n = 10), MI induction treated with apposition of a cell-free pericardial scaffold connected to the EIS system; (b) EBIG-treated arm (n = 12), MI induction treated with the EBIG; and (c) sham arm (n = 4), no MI, but the EBIG was implanted on top of healthy myocardium.
All animal studies were approved by the local Animal Experimentation Unit Ethical Committee (no. ES 100370001499) and complied with all guidelines concerning the use of animals in research and teaching as defined by the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 80-23, revised 1996). Human pericardial samples were obtained after written informed consent from patients undergoing cardiac surgery. The local ethics committee approved this study, and the protocol conformed to the principles outlined in the Declaration of Helsinki.

Engineered Bioactive Impedance Graft
Human pericardium obtained from patients undergoing cardiac surgery was used as a biological scaffold ( Fig. 2A). Pericardial decellularization was performed as previously described [7]. Briefly, the tissue was immersed in 1% SDS for 72 hours and then in 1% Triton-X for an additional 48 hours. Afterward, the membranes were lyophilized with a vacuum pump to extract all fluids, sterilized with g rays, and stored until use (Fig. 2B). Thirty minutes after MI induction, the metallic electrodes were separately anchored 1 cm apart in the thickness of the membrane (Fig. 2C,  2D). Then, the decellularized pericardium was rehydrated with 175 ml of Puramatrix hydrogel (BD Biosciences, San Jose, CA, http://www.bdbiosciences.com), and in treated animals, the EBIG included 2 3 10 6 GFP-pATPCs embedded in 175 ml of 10% sucrose. Finally, to ensure jellification, 350 ml of a-minimal essential medium was applied (Fig. 2E). Then, the bioimpedance devices, previously coated with biocompatible polydimethylsiloxane silicone and sterilized (Fig. 2F), were located in a subcutaneous pocket in the left supraescapular zone (Fig. 2G). Eighteen EIS systems were implanted (Fig. 2H) (seven animals in each experimental group and four sham animals) and connected to the EIS system (Fig. 2I, 2J) for active monitoring of MI scar evolution. Finally, animals were individually housed with an ad-hoc-developed antenna for noninvasive online bioimpedance signal transmission (Fig. 2K).

Noninvasive Scar Maturation Assessment
Impedance (magnitude and phase angle) was acquired every 5 minutes at 14 frequencies logarithmically spaced between 100 and 200 kHz. System temperature and battery voltage were also acquired. The contents of the EIS systems were downloaded after explantation through the Zigbee radio link controlled by a custom application (LabView, National Instruments, Austin, TX, http:// www.ni.com). The impedance measurements were processed with a median filter (n = 25, equivalent to 2 hours) to exclude sudden artifacts. Subsequently, a moving average filter (n = 150, equivalent to 12.5 hours) was used to smooth the time series by using the zero-shift double pass filter filtfilt (Matlab, Math-Works, Natick, MA, http://www.mathworks.com). In order to correct for long artifacts, the magnitude at the highest frequency (200 kHz) was subtracted from the magnitude at the other frequencies. Two estimators were chosen to display the integrity of the tissue in the monitored area: the slope of the time course of the impedance magnitude ratio between low frequency (LF; 1 kHz) and high frequency (HF; 100 kHz), and the slope of the time course of the phase angle difference between the LF and HF. The rationale for these estimators is presented in Discussion.

Histopathology Examination
Sacrifices were performed an average of 30.6 6 3 days after MI with an overdose of anesthesia. After lateral thoracotomy, the hearts were excised. Left ventricle (LV) infarct size was measured by examination of sections obtained 1.5 cm distally to the artery ligation by using the following equation: infarct size (%) = [(LV infarct area)/(LV total area)] 3 100. Quantitative morphometric and histological measurements were completed with Image-Pro Plus software (version 6.2.1; Media Cybernetics, Rockville, MD, http://www.mediacy.com).
On 4-mm paraffin slices, modified Gallego's and Masson's trichrome and Picrosirius Red staining were performed to analyze both pathological and histological changes and collagen deposition (type I red/yellow and type III green) under a computerassociated Leica DMI 6000B (Leica, Wetzlar, Germany, http:// www2.leicabiosystems.com) microscope with a polarized filter.

Statistical Analysis
Data are represented as the mean 6 SEM. Statistical analyses were performed by using the Student t test, one-way analysis of variance (ANOVA) with the Tukey's procedure for multiple comparisons, and the Mann-Whitney test for nonparametric data, using SPSS 19.0.1 (IBM, Armonk, NY, http://www.ibm.com). MRI data were analyzed as repeated measures by using ANOVA with the Greenhouse-Geisser correction. Bioimpedance data analyses were performed with the paired samples t test using SigmaStat software (Systat Software, Inc., San Jose, CA, https://systatsoftware.com). Values of p , .05 were considered significant.

RESULTS
Two animals died during MI induction due to ventricular fibrillation, and three were excluded from the study after postoperative infections. Therefore, 21 animals were included in the experimental protocol in the control (n = 8), EBIG-treated (n = 10), and sham (n = 3) groups (Fig. 1). Swine were sacrificed at predefined times (p = .5 between groups). In all animals, the graft was seen covering the infarct area; absence of deleterious graft-driven effects was detected in sham animals (supplemental online Fig. 1). Baseline infarct size between EBIG-treated and control animals was not significantly different as assessed by the necrosis biomarkers cTnI and creatine kinase-MB (0.39 6 0.1 vs. 0.32 6 0.1 mg/l, p = .63; and 0.59 6 0.09 vs. 0.54 6 0.14 mg/l, p = .75, respectively).

Cardiac Function Assessment
Baseline cardiac function did not differ between the EBIG-treated and control groups in LVEF (55.8% 6 2.2% vs. 56.5%   Table 1).
To understand the MRI functional benefit in the EBIGtreated arm, the infarct size and cellular differentiation were assessed. After digital morphometric analysis, LV infarct size was 48% smaller in EBIG-treated than control animals (3.4% 6 0.6% vs. 6.5% 6 1%, respectively; p = .015) (Fig. 3D-3F). Proper adhesion of the implanted graft with subjacent myocardium was observed in all control and EBIG-treated animals (supplemental online Fig. 2A-2D).

Cell Proliferation and GFP + -pATPC Endothelial and Cardiac Differentiation
As supplemental online Figure 3 depicts, there were cardiomyocytes expressing the cell proliferative marker Ki67 in the border zone of treated animals. However, Ki67-positive cells were not detected in the infarct core. Additionally, upon sacrifice, grafts from EBIG-treated and control animals showed vessel formation, some of which were positive for SMA, which is indicative of the existence of a vascular media layer (supplemental online Fig. 2E, 2F). Moreover, GFP + -pATPCs de novo expressed the endothelial markers IsoB4, SMA, vWF, and CD31 + within the EBIG and at the infarct zone, suggesting graft-driven neovascularization (Fig. 4A). Relative to cardiac differentiation analysis, we found that both GFP + -pATPCs within the graft and those that migrated to the infarct area were positive for NKX2.5, cKit, MEF2, cTnI, and cTnT. (Fig. 4B).

Scar Evolution Monitoring
Out of the 18 systems implanted (7 per experimental group and 4 in the sham group), 11 provided useful data (4 controls, 4 EBIGtreated, and 3 sham). Four of the remaining EIS systems were in deceased or excluded animals, and the remaining three stopped measuring after 1-3 days for unknown reasons.
Within 8 days, significant differences were observed between EBIG-treated and control animals. Figure 5 shows the time course of the impedance magnitude ratio (Fig. 5A) and phase angle difference (Fig. 5B) in the three groups. The magnitude ratio slope estimator was 20.48 6 0.07 per day for controls and 20.02 6 0.04 per day for EBIG-treated (p = .002). There was also a significant difference between the control and sham groups (20.15 6 0.07 per day; p = .013), whereas no differences were found between the EBIG-treated and sham groups (p = .39).
Regarding the phase angle difference slope estimator, the results were 23.52°6 0.84°per day for controls and 20.18°6 0.24°per day for EBIG-treated (p = .004). There was also a significant difference between the control and sham groups (21.08°6 0.17°per day; p = .03). The slopes were calculated as the firstorder coefficient of the linear regression for every magnitude ratio and phase difference time courses without any normalization. Figure 5 shows box plots of the magnitude ratio slope estimator (Fig. 5A) and the phase angle difference slope estimator (Fig. 5B).
To correlate the EIS findings with tissue differences between groups, collagen content and inflammation were examined by using histopathological techniques. Collagen content was measured At the bottom appears the box plot of the phase angle difference slope estimator, which also displayed a significant difference between controls and the other two groups. ‡, p = .004; x, p = .03. The means and confidence intervals can be found in Results. Abbreviations: EBIG, engineered bioactive impedance graft; HF, high frequency; LF, low frequency.

DISCUSSION
In the present study, we focused on a noninvasive assessment of cardiac function and scar healing using cardiac MRI and EIS to observe changes induced by a novel EBIG applied in a preclinical model of MI in swine. Two main conclusions emerge from our results: First, the increase in cardiac contractility (assessed by LVEF) in the EBIG-treated group was more than fivefold higher than that of controls; second, the EBIG positively impacted the healing process, providing impedance values similar to those of shamoperated animals, and altered scar collagen content.
Indeed, the remarkable cardiac function benefits with the EBIG in terms of LVEF and SV improvement by MRI analysis are in agreement with the 48% reduction in infarct size in EBIGtreated animals. Similar, but lower-magnitude, benefits were previously described when adipose-derived progenitor cells were implanted using different routes of administration (intramyocardial injection [9][10][11], intracoronary infusion [11], or peripheral intravenous delivery [12]) in murine and swine MI models.
Adipose-derived progenitor cells tested in cardiac regeneration therapies may be of subcutaneous [10][11][12] or cardiac origin [9,13]. Bayes-Genis et al. previously demonstrated that cardiac adipose-derived progenitors were more committed toward a cardiac-like phenotype than those of subcutaneous origin [9]. Moreover, when seeded into the scaffold, these cells showed repopulation and viability levels suitable for in vivo implantation [7]. In this study, upon sacrifice, we were able to find GFP + cells within both the EBIG and in underlying myocardium de novo expressing cTnI and NKX2.5. cTnI plays a key role in mature sarcomere organization, and NKX2.5 is central in activating the cascade of cardiomyogenic genes. Thus, it is likely that pATPCs are partly responsible for the beneficial effects of the EBIG because many studies have reported that ATPCs secrete factors such as adipokines, growth factors, and proteins from the extracellular matrix that exert an important role in cardiac regeneration [14]. Evidence is also accumulating showing that adipose-derived hormones can offer cardioprotective effects, including attenuation of cardiomyocyte apoptosis and reduction of infarct size [8]. Finally, neovascularization of the graft and underlying myocardium has also been suggested as a putative mechanism to limit infarct size [6]. In line with this, despite no vessel density differences in infarct core, within the graft, the scar, and the remote myocardium we found vessels positive for GFP, IsoB4, SMA, vWF, and CD31, but not in remote myocardium. In point of fact, GFP + pAPTCs within the scaffold also expressed SMA. Regarding cardiomyocyte proliferation, and in agreement with previous studies [15], the limited number of newly formed cardiomyocytes identified in EBIG-treated animals may only have a low contribution to the significant improvement observed.
Interestingly, to noninvasively monitor online myocardial scar evolution over 30 days, the grafts of sham, control, and EBIGtreated animals were connected to a customized EIS. Previous measurements of in vitro and in vivo myocardium in both normal and healed infarction states have been reported, displaying a frequency-dependent higher impedance for normal tissue and a resistive behavior for scar tissue [16,17]. Consequently, a gradual evolution between both states, with time-decreasing impedance at low frequencies, could be expected in the transition to healed scar tissue in control animals, whereas impedances at all frequencies should remain constant in EBIG-treated animals. Our initial aim was to fit all measurements to the Cole model for impedance and to study the time course of the model parameters to obtain information about the structural changes undergone by the tissue. The difficulty in performing suitable calibration because of differences in electrode impedances and in capacitive coupling between the cables and the animal's body forced us to abandon this goal and to define estimators based on low-and high-frequency measurements. The variability between subjects, as well as possible differences in the cell constant defined by interelectrode distance, made it necessary to define relative estimators like the magnitude ratio and phase angle difference between low-and high-frequency measurements. Finally, the slope of the time course of both the magnitude ratio and phase difference showed significant differences between the control and the EBIG-treated/sham animals. According to EIS results, the main healing changes occurred before the 8th day after the infarct, as shown by ample histopathology evidence [18]. Both the magnitude ratio and phase angle difference showed a clearly decreasing slope for the control group, whereas it remained almost constant for the EBIG-treated and sham groups. These results are consistent with the transition to scar tissue in the control group and the preservation of myocardial tissue (at least from an impedance perspective) in EBIG-treated animals. In accordance, analysis of myocardial fibrosis on the infarct core showed a lower collagen I/III ratio in EBIG-treated animals due to less collagen I deposition and higher collagen III, which is eventually replaced by healthy tissue.
Because inflammation is also a key step in myocardial healing after MI injury, we explored the impact of EBIG as an inflammatory modulator. Indeed, fewer activated lymphocytes accumulated in EBIGtreated animals, a phenomenon most likely caused by the adipose progenitors contained within the graft. Perea-Gil et al. have reported the immunomodulatory potential of cardiac adipose tissue progenitors abrogating T-cell alloproliferation in vitro [19]. A key event in graft rejection is the activation and proliferation of the recipient's lymphocytes, particularly T cells. Inhibition of T-cell activation (CD25 + ) is observed in patients under immunosuppressive treatment [20].

CONCLUSION
An advanced EBIG composed of acellular human pericardium refilled with pATPCs significantly reduced infarct size and improved cardiac function in a preclinical model of MI. In this context, the integration of a sophisticated noninvasive EIS monitoring system was useful to analyze the evolution of myocardial healing in EBIG-treated animals. Collectively, from an EIS perspective, myocardial tissue was preserved in EBIG-treated animals despite MI induction, which was confirmed by histopathological measurements of less inflammation and altered collagen I/III ratio.