Decellularized amniotic membrane attenuates postinfarct left ventricular remodeling

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.JournalofSurgicalResearch.com

Decellularized amniotic membrane attenuates postinfarct left ventricular remodeling Rajika Roy, PhD,a,1 Tobias Haase, PhD,a,1 Nan Ma, PhD,a Andreas Bader, Dipl Oec Troph,a Matthias Becker, MSc,a Martina Seifert, PhD,a Yeong-Hoon Choi, MD,b Volkmar Falk, MD,c and Christof Stamm, MDa,c,* a

Berlin-Brandenburg Center for Regenerative Therapies, Berlin, Germany Heart Center, University of Cologne, Cologne, Germany c Deutsches Herzzentrum Berlin, Berlin, Germany b

article info

abstract

Article history:

Background: Placenta and amnion have been suggested as sources of juvenile cells and

Received 18 May 2015

tissues for use in surgical regenerative medicine. We previously determined the impact of

Received in revised form

amniotic epithelial cells induced to undergo epithelial-to-mesenchymal transition (EMT)

24 July 2015

on myocardial remodeling processes and now evaluated the effects of naı¨ve and processed

Accepted 14 August 2015

amniotic membrane (AM) on postischemic left ventricular (LV) geometry and function.

Available online 22 August 2015

Methods: Human AM was used in unmodified form (AM), after EMT induction by transforming growth factor b (EMT-AM), and after decellularization (Decell-AM). After charac-

Keywords:

terization by histology, electron microscopy, splenocyte proliferation assay, and cytokine

Myocardial infarction

release, myocardial infarction was induced in 6e8-week old male BALB/c mice by perma-

Cardiac surgery

nent left anterior descending coronary occlusion, and AM patches were sutured to the

Left ventricular remodeling

anterior LV surface (n ¼ 10 per group). Infarcted hearts without AM or sham-operated mice

Amniotic membrane

were used as controls (n ¼ 10 each). After 4 weeks, LV pressureevolume curves were

Placenta

recorded using a conductance catheter before the animals were sacrificed and the hearts analyzed by histology. Results: TGF-ß treatment induced EMT-like changes in amniotic epithelial cells but increased AM xenoreactivity in vitro (splenocyte proliferation) and in vivo (CD4þ cell invasion). Moreover, in vitro interleukin-6 release from AM and from cardiac fibroblasts coincubated with AM was 300- or 100-fold higher than that of interleukin-10, whereas Decell-AM did not release any cytokines. AM- and Decell-AM-treated hearts had smaller infarct size and greater infarct scar thickness than infarct control hearts, but there was no difference in myocardial capillary density or the number of TUNEL positive apoptotic cells. LV contractile function was better in the AM and EMT-AM groups than in infarcted control hearts, but dP/dt max, dP/dt min, stroke work, and cardiac output were best preserved in mice treated with Decell-AM. Volume-based parameters (LV end-systolic and end-diastolic volume as well as LV ejection fraction) did not differ between AM and Decell-AM.

* Corresponding author. Deutsches Herzzentrum Berlin, Augustenburger Platz 1, 13353 Berlin, Germany. Tel.: þ49 30 4593 2109; fax: þ49 30 4593 2100. E-mail address: [email protected] (C. Stamm). 1 R.R. and T.H. contributed equally to this work. 0022-4804/$ e see front matter ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2015.08.022

410

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

Conclusions: Decellularized AM supports postinfarct ventricular dynamics independent of the actual regeneration processes. As a cell-free approach to support the infarcted heart, this concept warrants further investigation. ª 2016 Elsevier Inc. All rights reserved.

1.

Introduction

The adult mammalian myocardium has very limited capacity to regenerate, and myocardial infarction without immediate revascularization initiates remodeling processes that ultimately lead to congestive heart failure. Strategies to enhance myocardial regeneration by cell therapy or tissue engineering yielded promising results in experimental models but have shown little success in clinical trials so far. Moreover, the current regulatory framework greatly complicates the clinical translation of advanced therapy medicinal products involving viable cells. Recently, the impact of extracellular matrix (ECM) on tissue regeneration processes has been acknowledged, and the therapeutic potential of tissue-specific ECM products is being tested. Although cardiac ECM produced by decellularization was shown to positively influence the behavior of cardiac cells and stem cells, its availability for clinical application may be limited and production processes complex and costly [1,2]. Moreover, decellularized cardiac ECM has a gellike consistency so that it is difficult to handle surgically without additional supporting materials, whereas nonspecific ECM products are much easier to create and to apply. For instance, porcine small intestinal submucosa has been used in preclinical and clinical surgical settings, and decellularized porcine urinary bladder ECM was applied as a patch material for experimental reconstruction of the left ventricle and shown to facilitate skeletal muscle regeneration in humans [3e5]. Human amniotic membrane (AM), readily available from the placenta without ethical concerns or complex harvesting procedures, has long been used in ophthalmologic surgery and for burn wound healing [6e9], and was shown to improve heart function in experimental models of myocardial infarction [10]. Of note, Amensag and McFetridge [11] reported the biological and mechanical properties of the AM being enhanced in the rolled multilayered format of the membrane compared with a single layer. The biologic AM effects have mainly been attributed to resident amniotic epithelial cells (AECs), which are believed to retain pluripotent stem cell characteristics [12]; but, we have recently shown that those require induction of epithelial-to-mesenchymal transition (EMT) to augment their cardioprotective capacity [13]. We now tested the efficacy of a simplified, cell-free approach using decellularized AM in a surgically induced model of myocardial infarction in an immunologic mismatch model.

2.

Materials and methods

2.1.

Preparation of AM heart patches

Human full term placentas were obtained under sterile conditions immediately after cesarean delivery from women who

had given written informed consent for use of the placenta for research purposes, according to the declaration of Helsinki (7th revision, 2013, World Medical Association) and with approval of the local ethics committee. The placenta was placed on absorbent paper, and the umbilical cord was trimmed. Excess blood was removed by washing with Hanks balanced salt solution (HBSS; Life Technologies, Carlsbad, CA). The AM was mechanically peeled from the underlying chorion and washed several times in HBSS to remove blood. Then, the membrane was cut into 1cm2 pieces and placed in six-well plates (Greiner Bio One, Frickenhausen, Germany). In the cellular AM group, unmodified membrane was used. For the EMT-induced group (AMEMT), the membrane was incubated with transforming growth factor b1 (TGF-b1) in Dulbecco’s modified Eagle’s medium with high glucose supplemented with 10% fetal bovine serum, 100 mM of nonessential amino acids, 55 mM of 2-mercaptoethanol, 200 mM of L-glutamine, penicillinstreptomycin, and 10 ng/mL of epidermal growth factor (all purchased from Life Technologies, Carlsbad, CA) for a period of 7 days. For the decellularized membrane group (AM-Decell), AM patches were washed with phosphate buffered saline without calcium and magnesium (PBS) and placed in lysis buffer (10 mM TRIS and 0.1% EDTA) for 1 h at room temperature, followed by incubation in 0.5% sodium dodecyl sulfate (Carl Roth, Karlsruhe, Germany) solution for 4 h at room temperature. Next, the membrane patches were washed in PBS overnight under constant agitation.

2.2.

TGF-ß effect on AEC viability

For induction of EMT in non-decellularized patches, AM was treated for 7 days with increasing concentrations of TGF-b1 (25e100 ng/mL). To assess viability/metabolic activity under EMT, cell culture medium was carefully removed, and 20 mL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2(4-sulfophenyl)-2H-tetrazolium (MTS)/phenazine methosulfate (Promega, Madison, WI) mixture solution was added to 100 mL of cell culture and incubated for 4 h at 37 C. Absorbance at 490 nm was recorded using a standard plate reader (Molecular Devices, Sunnyvale, CA).

2.3.

Scanning electron microscopy

AM patches were fixed in 2.5% glutaraldehyde (Carl Roth, Karlsruhe, Germany) for 30 min at room temperature under constant agitation and washed three times in PBSþþ. The membrane was then dehydrated by sequential acetone (Carl Roth, Karlsruhe, Germany) treatment (30% acetone for 5 min, 50% for 5 min, 75% for 10 min, and 90% for 10 min, 100% for storage) before the samples were subjected to critical point drying. Membrane samples were mounted on a carbon-coated aluminum holder and coated with gold before visualization in a Hitachi S-2700 scanning electron microscope.

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

2.4.

Splenocyte proliferation assay

Splenocytes isolated from adult C57bl/6 mice (iQ Biosciences) were seeded in 96-well microtiter plates (3  10E5/well, Nunc; Thermo Fisher Scientific, Australia) in Roswell Park Memorial Institute (RPMI) medium, and T-cells were stimulated with 5 mL/mL concanavalin A (ConA; SigmaeAldrich, St. Louis, MO) or an equivalent amount of vehicle. Cellular AM pieces of a size containing approximately 0.1e1  10E5 AEC (AEC-tosplenocyte ratio 1:3e1:30) were minced and added to the respective wells, and equivalent pieces of decellularized AM were prepared and added to separate wells. The plates were then incubated at 37 C for 48 h, before 1 mCi/well [3H]thymidine (PerkinElmer, Waltham, MA) was added, followed by incubation for 18 h. AM fragments were then removed by centrifugation, splenocytes were collected on filter paper, and radioactivity was quantified using a Top Count Harvester (Packard Biosciences, Meriden, CT).

2.5.

Cytokine release

To study whether cytokines are released by Decell-AM and cellular AM, 8-mm membrane samples were placed with noncellular side down into 48-well plates and held in place by silicone rings cut from silicon tubing (Ismatec; MF0040). To assess whether AM may induce cytokine release from epicardial cells, separate experiments were done in the presence of human cardiac fibroblasts (generously donated by S. van Linthout, BCRT, Berlin). After incubation for 2 d in serumfree medium, the supernatant was collected, and interleukin (IL)-6 and IL-10 concentrations were determined by enzymelinked immunosorbent assay (BioLegend, San Diego, CA).

2.6.

Induction of myocardial infarction

The experimental protocol was approved by the Landesamt fu¨r Gesundheit und Soziales Berlin (State Office of Health and Social Affairs Berlin, G0456/08) and conforms to the Directive 2010/63/EU of the European Union. Male BALB/c mice aged 8 to 10 wks old (Charles River, Sulzfeld, Germany) were anesthetized by intraperitoneal injection of midazolam (5.0 mg/ kg), fentanyl (0.05 mg/kg), and medetomidine (0.5 mg/kg). The animals were intubated and ventilated with a tidal volume of 0.2 mL at a rate of 200/min (MiniVent, Hugo Sachs, MarchHugstetten, Germany). Anesthesia was maintained with 1.5% isoflurane at FiO2 ¼ 1.0. A left anterolateral thoracotomy was performed in the left third intercostal space, the pericardium was opened and the heart exposed. Then the left anterior descending coronary (LAD) was ligated with 7e0 prolene sutures (Ethicon, Norderstedt, Germany) 1 mm beneath the tip of the left atrial appendage. Myocardial ischemia distal to the ligation was evident by the discoloration of the myocardium and was confirmed by ECG. The AM patch was then placed on the anterior surface of the heart and was held in place using the LAD ligation suture ends. Care was taken to evenly spread the membrane over the myocardium and cover it with pericardium. The ribs were then approximated and the wounds closed with running 6e0 vicryl sutures. Animals were allowed to recover, extubated, and allowed to survive for 4 wk. Flunixin meglumine (2.5 mg/kg)

411

was administered subcutaneously for 72 h for pain relief. In the sham group, a thoracotomy was performed, but the heart was not manipulated.

2.7.

Hemodynamic measurements

Heart function was analyzed by recording ventricular pressuredvolume loops 28 d after surgery. Mice were anesthetized, intubated, and ventilated (see previous section), and an 1.4 F polyimide pressure-conductance catheter (Millar Instruments, Houston, TX) was inserted through the right carotid artery and guided into the left ventricle. Conversion of raw conductance data to calibrated volumes was performed by determination of parallel conductance (Vp) using a hypertonic saline dilution method, with 10 mL of 15% hypertonic saline injected via the left jugular vein [14]. To obtain absolute volume measurements, the catheter was calibrated with known volume of mouse blood. Measurements and data analysis were performed by a blinded investigator using LabChart software (AD Instruments, Dunedin, New Zealand). After catheterization, mice were sacrificed by cervical dislocation, and hearts were excised and snap frozen in Tissue-Tek O.C.T. embedding medium.

2.8.

Histology and immunofluorescence

For analysis of in vitro AM, naı¨ve and decellularized patches were fixed with 4% paraformaldehyde (Carl Roth, Karlsruhe, Germany) for 30 min at room temperature with constant shaking, or frozen in Tissue-Tek O.C.T. embedding medium (A. Hartenstein, Wu¨rzburg, Germany). Five-micrometer sections were mounted on Superfrost Plus microscope glass slides (Thermo Scientific, Waltham, MA, USA) and were stained with Masson trichrome stain (SigmaeAldrich, St. Louis, MO) according to the manufacturer’s instructions. For analysis of explanted hearts, transverse 7-mm sections were cut from apex to base and mounted on glass slides. Sections were fixed with 4% PFA and stained with Masson trichrome stain according to the manufacturer’s protocol. The areas of infarcted and non-infarcted myocardium were quantified using ImageJ (NIH, Bethesda, MD). The percentage of infarcted myocardium was calculated with respect to the total area of the myocardium, and the thickness of the infarct scar was measured at three points within the transmural infract area. Cardiac vascularization was evaluated by staining with isolectin IB4 conjugated to Alexa 647 (Molecular Probes, Eugene, OR). Apoptotic cells in the transplanted membrane and the peri-infarct area were detected by labeling fragmented DNA using the In Situ Cell Death Detection System (Roche Applied Science, Penzberg, Germany) according to the manufacturer’s instructions. Inflammatory cells were identified by staining with CD4 antibody (BioLegend, San Diego, CA) followed by secondary anti-goat Alexa 555 conjugated antibody (Molecular Probes, Eugene, OR). Nuclei were counterstained with 40 , 6diamidino-2-phenylindole (Life Technologies, Carlsbad, CA). Images were acquired with a Zeiss Axioskop Microscope (Carl Zeiss, Oberkochen, Germany). The number of positively stained cells was counted in the indicated region and expressed as positive cells per 100 mm2.

412 2.9.

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

Statistics

Data were analyzed using IBM SPSS statistics 20. Continuous data are expressed as mean  standard error of the mean. One-way analysis of variance with Bonferroni post hoc test was used to compare data sets of more than two experimental groups. If indicated, differences between individual data sets were tested by two-tailed, unpaired Students’ t-test. A value of P < 0.05 was considered statistically significant.

3.

Results

3.1.

AM patch characteristics

Masson trichrome staining of naı¨ve AM cross sections revealed a single layer of epithelial cells, scattered fibroblastoid mesenchymal stromal cells, and an intact basement

membrane consisting mainly of collagen fibrils as indicated by aniline blue staining (Fig. 1). The epithelial surface of the naı¨ve AM displayed typical polygonal, cobblestone morphology, consistent with that of epithelial cells in 2D cultures. Scanning electron microscopy images confirmed the compact structure of the cellular monolayer with tight intercellular junctions essential for sealing the amniotic cavity. The presence of microvilli on the surface of epithelial cells was also evident. Induction of EMT by TGF-b1 led to a loss of the cobblestone morphology of the epithelial layer (Fig. 1B), with irregularities in cell size and distribution. By electron microscopy, intercellular gaps became apparent, whereas the largely still villibearing spheroid cells remained attached to the basement membrane. Because TGF-b1 is also known to induce cellular senescence, we tested the effect of increasing TGF-b1 concentrations on amniotic cell viability (Fig. 1K). Indeed, increasing TGF-ß1 concentration dose-dependently impaired AEC viability, and we therefore chose a concentration of 25 ng/

Fig. 1 e Morphology of naı¨ve and modified amniotic membrane (AM). (AeC) Masson trichrome staining of unmodified AM (A), EMT-AM (B), and Decell-AM in cross section. Original magnification 603 (scale bar [ 50 mm). (D, E), Crystal violet staining of the AEC layer demonstrating a typical cobblestone pattern in naive AM (D), and irregular morphology in EMT-AM (E). Original magnification 403 (scale bar [ 50 mm). (FeH), Scanning electron microscopy of unmodified AM (F), EMT-AM (G), and Decell-AM (H). Original magnification 10003 (scale bar [ 20 mm). (I, J), Detail of intact and lost intercellular connections in unmodified AM (I) and EMT-AM (J). Original magnification 40003 (scale bar [ 6 mm). (K), AEC viability as assessed by MTS assay in response to increasing concentrations of TGF-ß1 for EM induction. *P < 0.05.

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

413

mL for in situ induction of EMT before AM transplantation. The decellularization protocol completely removed not only the epithelial cell surface layer but also the amniotic mesenchymal cells embedded in the stroma (Fig. 1C and H). Morphologically, ECM structure was preserved, including a distinct basement membrane.

3.2.

Fig. 2 e Intraoperative view of the mouse heart via a left lateral thoracotomy. The LAD has been ligated with a 7e0 prolene suture (*), which has also been used to attach the AM patch.

Postinfarct left ventricular function

After LAD ligation, naı¨ve (AM, n ¼ 10), TGF-b1 treated (EMTAM, n ¼ 10), or decellularized AM patches (Decell-AM, n ¼ 10) were transplanted with the epithelial layer directed to the surface of the infarcted left ventricular myocardium (Fig. 2). Infarct control mice received no AM (n ¼ 10). There were no surgical complications related to the AM transplantation. Within the 28-d observation period, one animal in the AM group and one animal in the EMT-AM group died (intergroup P ¼ 0.8). This infarct-related mortality is comparable to that observed in previous studies. In comparison to sham operation, LAD ligation alone led to severely impaired left ventricular (LV) function as measured by pressure/conductance catheter on day 28. As expected, LV ejection fraction decreased, whereas end-systolic and end-diastolic LV volume increased in infarcted compared to sham-operated hearts (Fig. 3). The most sensitive parameters of systolic and diastolic function, dP/dt max and dP/dt min, were both significantly improved in Decell-AM hearts but not in the AM or EMT-AM

Fig. 3 e (ALI), Heart function 4 wk postoperatively measured in vivo by cardiac catheterization and recoding of LV pressureevolume curves. CO [ cardiac output; EF [ ejection fraction; HR [ heart rate; SV [ stroke volume; SW [ stroke work; Ved [ end-diastolic volume; Ves [ end-systolic volume. *P < 0.05.

414

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

Fig. 4 e Cardiac histomorphology. (AeK), Cross-sectional photomicrographs of infarcted hearts w/wo epicardial AM application, Masson trichrome staining. (A, B) Control infarct group without AM transplantation. (CeE) unmodified AM; (FeH) EMT-AM, (IeK), Decell-AM. AM [ amniotic membrane. Left panels, middle panel original magnification 23. Right panel original magnification 203 (scale bar [ 50 mm). (L) Heart weight-to-body weight ratio. (M) Infarct size quantified by planimetry in Masson trichrome stained sections. (N) Average wall thickness in the infracted area of interest. *P < 0.05.

group (Fig. 3A and D). In line with this finding, LV stroke work, stroke volume, and cardiac output were also best preserved in hearts treated with Decell-AM (Fig. 3EeG). LV end-systolic volume tended to be smaller in the Decell-AM group, but this trend did not reach statistical significance and, despite the smaller infarct size in AM and Decell-AM hearts (see the

following section), there was no significant difference in enddiastolic volume (Fig. 3H and I). Taken together, the hemodynamic measurements indicate that postinfarct LV function was best preserved in hearts that underwent transplantation of Decell-AM, whereas unmodified AM or EMT-AM had no relevant impact on contractility or relaxation.

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

415

Fig. 5 e (A, C, E) Blood vessel density in the peri-infarct left ventricular myocardium visualized by isolectin IB4 staining. Original magnification 203 (scale bar [ 50 mm) (G) Average microvessel density in the peri-infarct area/100 mm2. (B, D, F) Apoptotic cells indicated by TUNEL staining in the peri-infarct area. Original magnification 203 (scale bar [ 50 mm). (H) Average number of TUNEL positive cells in the peri-infarct area/100 mm2. *P < 0.05.

3.3.

Myocardial structure

LAD ligation resulted in transmural myocardial infarction with typical histologic changes including thinning of the left

ventricular wall, fibrosis with extensive collagen deposition, and progressive ventricular chamber dilation (Fig. 4). The infarct scar tissue extended from the level of the LAD occlusion toward the apex of the heart. There were no

416

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

Fig. 6 e Immunohistology for visualization of CD4D cells in the infarct area treated with (A) unmodified AM (B) EMT-AM (C) Decell-AM. Original magnification 203 (scale bar [ 50 mm). (D) Average number of CD4D cells in the AM and adjacent myocardium 100 mm2. *P < 0.05.

significant differences in infarct size or the ratio of heart weight-to-body weight among the myocardial infarction groups. The AM patch was readily identifiable in all animals and was firmly attached to the myocardium. Compared to control hearts, infarct size was smaller in AM and Decell-AM treated hearts, but not in hearts that received EMT-AM (Fig. 4M). Although blood vessels and cells infiltrated the patch material, the epicardium was still visible as a distinct layer between the patch and the myocardium. The infarct scar was markedly thinner than normal myocardium, but wall thickness was best preserved in hearts treated with Decell-AM (Fig. 4N). In the ischemic myocardium adjacent to the patch and in the densely vascularized peri-infarct area, there was no difference in blood vessel density between the infarct groups as seen by isolectin IB4 staining (Fig. 5). There were a considerable number of apoptotic cells in the periinfarct area in all groups. Furthermore, apoptotic cells were visible in the transplanted naı¨ve (AM) and TGF-b1treated EMT-AM. However there was no significant difference in the number of TUNEL positive cells in the periinfarct area among the infarct groups (Fig. 5).

3.4.

Inflammatory response

CD4þ inflammatory cell invasion was detected in both AM and EMT-AM treated hearts, but was not higher than that in the control infarct group when decellularized AM was implanted (Fig. 6). We therefore also measured the splenocyte activation/ proliferation rate in vitro and found that EMT-induction with

TGF-ß1 abrogated the proliferation-inhibiting effect that AEC exerted on concanavalin-stimulated splenocytes, similar to complete decellularization (Fig. 7A). In addition, the proliferation of otherwise unstimulated splenocytes was enhanced only in the presence of EMT-AM (Fig. 7B). Decell-AM has no apparent impact on splenocyte proliferation. During 2d incubation in vitro, Decell-AM released neither IL-6 nor IL10, presumably because all soluble proteins have been removed during the decellularization process (Fig. 8A and B). Cellular AM, however, released both cytokines, but the amount of pro-inflammatory IL-6 was 300-fold higher than that of anti-inflammatory IL-10 (Fig. 8C). Also, only cellular AM but not Decell-AM induced the release of IL-6 and IL-10 from human cardiac fibroblasts, and again, there was a nearly 100fold dominance of IL-6.

4.

Discussion

Various strategies to counteract postinfarct myocardial remodeling processes have been proposed, including the intramyocardial implantation of cells alone, epicardial application of cellularized patch materials [15e17], intramural injection of polymer material [18], and ventricular stabilization with transventricular splints [19]. The primarily mechanical approaches have largely failed in clinical testing, and those involving viable somatic cells have also produced inconsistent or negative results. Although the therapeutic efficacy of autologous somatic cells is now considered insufficient, more

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

417

Fig. 7 e Splenocyte proliferation assay (C57bl/6 mouse cells). (A), splenocyte proliferation after T-cell stimulation with concanavalin in AEC-to-splenocyte ratio 1:3e1:30, indicating an immunosuppressive effect of unmodified AM. (B), baseline splenocyte proliferation rate indicating a xenogeneic pro-inflammatory stimulus of EMT-AM. CPM [ counts per min. *P < 0.05.

potent cell types created by means of genetic reprogramming are still far from being clinically usable. In this context, AM may possess unique characteristics in that it is believed to contain fetal cells with multilineage differentiation potential and mechanical properties that facilitate its use in surgery without the need for further manipulation. AM has been extensively used for ocular surface reconstructions since the 1940s [20,21], as a bioactive wound cover [6,7,22,23] and a pericardial substitute [24]. Cells from the AM were shown to have a beneficial effect on the myocardium in mouse and rat models of myocardial infarction [10,25], but transplantation of the entire membrane would reduce the time and costs associated with isolation, purification, and expansion of the cells. A stem cellelike character of AM epithelial cells (AEC) has been postulated because they display several pluripotency associated markers and were found to express selected targetcell genes/proteins on specific stimulation in vitro [12,26]. In vivo, we previously found that AECs exert cardioprotective effects in the ischemic heart, probably via paracrine mechanisms rather than by cellular transdifferentiation [13]. In those experiments, cardioprotection was much enhanced when AECs were induced to undergo EMT before transplantation. This observation then prompted us to study the effects of AEC when delivered to the epicardium while still attached to AM, which, in contrast to the intramyocardial

injection of cells in suspension, should act as a long-lasting cell deposit and provide mechanical stability. AM is considered to be immunoprivileged and even immunosuppressive in the allogeneic setting, as shown by early observations in humans and later experimental studies [27e29] and explainable by its function as part of the feto-maternal interface. In our xenogeneic experimental setting, however, we detected the invasion of CD4þ cells in both groups treated with cellular AM, which was not present when decellularized AM was used. CD4 is a glycoprotein found on the surface of mature T helper cells [30], macrophages, and dendritic cells [31]. Interestingly, the in vitro immune assays showed that only EMT-AM abrogates the immunosuppressive effect that naı¨ve AM exerts on stimulated splenocytes and also induces their activation above the baseline level. The interplay between EMT and the immune system has been studied mainly in the context of oncology research, and it has recently been described that EMT induced in tumor cells elicits an immune response mediated by the NKG2D receptor on natural killer cells and Tcells [32,33]. TGF-b itself has been shown to mediate immune tolerance [34] and aid tissue repair [35], but several studies have also shown that it plays a role in allograft rejection [36e38]. Because we previously observed a xenogeneic immune response to intramyocardial AEC injection, confirmed for epicardial AM application in the present study, we decided

Fig. 8 e Release of IL-6 and IL-10 from Decell-AM and naı¨ve AM, as well as from human cardiac fibroblasts co-incubated with AM for 2 d. (A) IL-6 concentration in the supernatant, (B) IL-10 concentration in the supernatant, (C) IL-67IL-10 ratio indicating the dominance of pro-inflammatory IL-6. CF [ cardiac fibroblasts. *P < 0.05.

418

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

to include decellularized AM in this set of experiments. Indeed, Decell-AM did not induce a visible cellular immune response, neither in vivo nor in vitro, and LV function and structure were better preserved in hearts treated with DecellAM, indicating that predominantly, the AM ECM components rather than AECs exert beneficial effects on the infarcted heart, whereas the cellular immune response to AECs is detrimental. In addition to the effects of AM immune cells, we also found that AM releases predominantly pro-inflammatory IL-6 in vitro, and also appears to induce the secretion of 100fold more IL-6 than IL-10 from human cardiac fibroblasts. Decell-AM, however, does not appear to contain any residual cytokines, nor does it stimulate cardiac fibroblasts to secrete those. This combination of allogeneic in vitro and xenogeneic in vivo data does not allow to draw definitive conclusions regarding the local immunologic effects of AM on the infarcted heart. However, it does indicate that cellular AM is not immunologically inert, but that the decellularization process effectively eliminates any immune activity, with beneficial consequences for the heart. Although the mechanism by which Decell-AM influences LV remodeling processes remains elusive, both mechanical and biological effects are likely to play a role. Decellularized AM has been shown to possess the same elasticity and tensile strength as the cellular AM and is also biocompatible in vitro [39]. Localized support of the dilating left ventricle may inhibit remodeling processes, and Decell-AM appears to have the appropriate characteristics to prevent LV dilatation without impairing diastolic relaxation. Although the membrane was sutured in place only with the LAD ligature, it adhered firmly to the epicardium, initially via dispersive adhesion forces and later by the formation of solid adhesions. However, the notion of a mechanical influence on LV physiology is impossible to prove in a mouse model and would be expected to lead to a relative decrease in LV volumes, which we did not observe. ECM has also been shown to influence the behavior of mature and progenitor cells by inhibiting apoptosis, supporting proliferation, and directing stem cell differentiation [1]. Various decellularized membranes have been suggested to aid in tissue repair by signaling mechanisms that involve not only the intact membrane but also the components of the degraded membrane [3,40]. These phenomena may have contributed to the observed effects but are difficult to demonstrate in an in vivo model such as the one we used. Also, several groups reported on the beneficial effects of multipotent stromal cells derived from AM [41e44] and expanded in vitro. However, AM-MSC are likely to play a lesser role in our experiments because they would have to exit the AM stroma and traverse the basement membrane and AEC layer to come into contact with the epicardium, and their number in a naı¨ve AM is limited as compared to the densely packed AEC. Taken together, we found that, although gross infarct size remained unchanged, decellularized AM attached to the surface of the acutely infarcted left ventricle helps prevent pathologic postinfarct remodeling processes, improving systolic and diastolic function. In this xenogeneic model, cellular AM, with or without EMT induction, elicited a local immune response that prevented a relevant improvement in heart function. Decellularized AM may therefore prove useful for

the development of epicardial approaches to counteract postischemic dysfunction.

Acknowledgment This project was supported by the German Bundesministerium fu¨r Bildung und Forschung (BMBF; FKZ 01GN0948 to C.S.) and by Deutsche Forschungsgemeinschaft (DFG, FKZ GSC 203) through the BerlindBrandenburg School for Regenerative Therapies. BCRT received funding from Charite´ Universita¨tsmedizin (FKZ 1315848A), Helmholtz-Zentrum Geesthacht (FKZ 1315848B), and BMBF (FKZ 0315848A). The authors thank Anne Gale for editorial assistance. Authors’ contributions: R.R., T.H., and M.B. conducted experiments, analyzed data, and prepared article. N.M. and A.B. analyzed data and revised article. Y.-H.C., V.F., and C.S., designed experiments and revised article.

Disclosure None of the authors have any personal or financial relationships to disclose that may have biased their work or conclusions.

references

[1] Oberwallner B, Brodarac A, Anic P, et al. Human cardiac extracellular matrix supports myocardial lineage commitment of pluripotent stem cells. Eur J Cardiothorac Surg 2015;47:416. [2] Oberwallner B, Brodarac A, Choi YH, et al. Preparation of cardiac extracellular matrix scaffolds by decellularization of human myocardium. J Biomed Mater Res A 2014;102:3263. [3] Badylak SF, Kochupura PV, Cohen IS, et al. The use of extracellular matrix as an inductive scaffold for the partial replacement of functional myocardium. Cell Transplant 2006;15(Suppl 1):S29. [4] Tan MY, Zhi W, Wei RQ, et al. Repair of infarcted myocardium using mesenchymal stem cell seeded small intestinal submucosa in rabbits. Biomaterials 2009;30:3234. [5] Sicari BM, Rubin JP, Dearth CL, et al. An acellular biologic scaffold promotes skeletal muscle formation in mice and humans with volumetric muscle loss. Sci Transl Med 2014;6: 234ra58. [6] Bennett JP, Matthews R, Faulk WP. Treatment of chronic ulceration of the legs with human amnion. Lancet 1980;1:1153. [7] Bose B. Burn wound dressing with human amniotic membrane. Ann R Coll Surg Engl 1979;61:444. [8] Clare G, Suleman H, Bunce C, Dua H. Amniotic membrane transplantation for acute ocular burns. Cochrane Database Syst Rev 2012;9:CD009379. [9] Dua HS, Gomes JA, King AJ, Maharajan VS. The amniotic membrane in ophthalmology. Surv Ophthalmol 2004;49:51. [10] Cargnoni A, Di Marcello M, Campagnol M, Nassuato C, Albertini A, Parolini O. Amniotic membrane patching promotes ischemic rat heart repair. Cell Transplant 2009;18: 1147. [11] Amensag S, McFetridge PS. Tuning scaffold mechanics by laminating native extracellular matrix membranes and effects on early cellular remodeling. J Biomed Mater Res A 2014;102:1325.

j o u r n a l o f s u r g i c a l r e s e a r c h 2 0 0 ( 2 0 1 6 ) 4 0 9 e4 1 9

[12] Miki T, Lehmann T, Cai H, Stolz DB, Strom SC. Stem cell characteristics of amniotic epithelial cells. Stem Cells 2005; 23:1549. [13] Roy R, Kukucka M, Messroghli D, et al. Epithelial to mesenchymal transition enhances the cardioprotective capacity of human amniotic epithelial cells. Cell Transplant 2015;24:985. [14] Steendijk P, Staal E, Jukema JW, Baan J. Hypertonic saline method accurately determines parallel conductance for dual-field conductance catheter. Am J Physiol Heart Circ Physiol 2001;281:H755. [15] Chang Y, Lai PH, Wei HJ, et al. Tissue regeneration observed in a basic fibroblast growth factor-loaded porous acellular bovine pericardium populated with mesenchymal stem cells. J Thorac Cardiovasc Surg 2007;134:65e73.e1-4. [16] Arana M, Gavira JJ, Pena E, et al. Epicardial delivery of collagen patches with adipose-derived stem cells in rat and minipig models of chronic myocardial infarction. Biomaterials 2014;35:143. [17] Godier-Furnemont AF, Martens TP, Koeckert MS, et al. Composite scaffold provides a cell delivery platform for cardiovascular repair. Proc Natl Acad Sci U S A 2011;108: 7974. [18] Sabbah HN, Wang M, Gupta RC, et al. Augmentation of left ventricular wall thickness with alginate hydrogel implants improves left ventricular function and prevents progressive remodeling in dogs with chronic heart failure. JACC Heart Fail 2013;1:252. [19] Hamner C, Ruth G, Raffe M, Schoen FJ, Schaff H. Safety and biocompatibility of the Myosplint systemda passive implantable device that alters ventricular geometry for the treatment of heart failure. Asaio J 2004;50:438. [20] Sorsby A, Haythorne J, Reed H. Further experience with amniotic membrane grafts in caustic burns of the eye. Br J Ophthalmol 1947;31:409. [21] Sorsby A, Symons HM. Amniotic membrane grafts in caustic burns of the eye (burns of the second degree). Br J Ophthalmol 1946;30:337. [22] Gruss JS, Jirsch DW. Human amniotic membrane: a versatile wound dressing. Can Med Assoc J 1978;118:1237. [23] Silverton JS, Trelford JD, Roussere JT, Wolfe BM, Conti S. The use of amniotic membrane in acute massive full-thickness loss of the abdominal wall from clostridial myonecrosis. Ann Plast Surg 1979;3:558. [24] Muralidharan S, Gu J, Laub GW, Cichon R, Daloisio C, McGrath LB. A new biological membrane for pericardial closure. J Biomed Mater Res 1991;25:1201. [25] Fang CH, Jin J, Joe JH, et al. In vivo differentiation of human amniotic epithelial cells into cardiomyocyte-like cells and cell transplantation effect on myocardial infarction in rats: comparison with cord blood and adipose tissue-derived mesenchymal stem cells. Cell Transplant 2012;21:1687. [26] Miki T, Strom SC. Amnion-derived pluripotent/multipotent stem cells. Stem Cell Rev 2006;2:133. [27] Dua HS, Azuara-Blanco A. Amniotic membrane transplantation. Br J Ophthalmol 1999;83:748.

419

[28] Akle CA, Adinolfi M, Welsh KI, Leibowitz S, McColl I. Immunogenicity of human amniotic epithelial cells after transplantation into volunteers. Lancet 1981;2:1003. [29] Ueta M, Kweon MN, Sano Y, et al. Immunosuppressive properties of human amniotic membrane for mixed lymphocyte reaction. Clin Exp Immunol 2002;129:464. [30] Killeen N, Sawada S, Littman DR. Regulated expression of human CD4 rescues helper T cell development in mice lacking expression of endogenous CD4. EMBO J 1993;12:1547. [31] Gibbings D, Befus AD. CD4 and CD8: an inside-out coreceptor model for innate immune cells. J Leukoc Biol 2009;86:251. [32] Lopez-Soto A, Zapico LH, Acebes-Huerta A, Rodrigo L, Gonzalez S. Regulation of NKG2D signaling during the epithelial-to-mesenchymal transition. Oncoimmunology 2013;2:e25820. [33] Lopez-Soto A, Huergo-Zapico L, Galvan JA, et al. Epithelialmesenchymal transition induces an antitumor immune response mediated by NKG2D receptor. J Immunol 2013;190: 4408. [34] Li MO, Wan YY, Sanjabi S, Robertson AK, Flavell RA. Transforming growth factor-beta regulation of immune responses. Annu Rev Immunol 2006;24:99. [35] Schmidt-Weber CB, Blaser K. Regulation and role of transforming growth factor-beta in immune tolerance induction and inflammation. Curr Opin Immunol 2004;16:709. [36] Little DM, Haynes LD, Alam T, Geraghty JG, Sollinger HW, Hullett DA. Does transforming growth factor beta 1 play a role in the pathogenesis of chronic allograft rejection? Transpl Int 1999;12:393. [37] Shihab FS, Tanner AM, Shao Y, Weffer MI. Expression of TGFbeta 1 and matrix proteins is elevated in rats with chronic rejection. Kidney Int 1996;50:1904. [38] Cohen AH, Nast CC. TGF-beta in renal allograft rejection. Miner Electrolyte Metab 1998;24:197. [39] Wilshaw SP, Kearney JN, Fisher J, Ingham E. Production of an acellular amniotic membrane matrix for use in tissue engineering. Tissue Eng 2006;12:2117. [40] Badylak SF, Freytes DO, Gilbert TW. Extracellular matrix as a biological scaffold material: structure and function. Acta Biomater 2009;5:1. [41] Danieli P, Malpasso G, Ciuffreda MC, et al. Conditioned medium from human amniotic mesenchymal stromal cells limits infarct size and enhances angiogenesis. Stem Cells Transl Med 2015;4:448. [42] Kim SW, Zhang HZ, Kim CE, Kim JM, Kim MH. Amniotic mesenchymal stem cells with robust chemotactic properties are effective in the treatment of a myocardial infarction model. Int J Cardiol 2013;168:1062. [43] Tsuji H, Miyoshi S, Ikegami Y, et al. Xenografted human amniotic membrane-derived mesenchymal stem cells are immunologically tolerated and transdifferentiated into cardiomyocytes. Circ Res 2010;106:1613. [44] Martinez EC, Vu DT, Wang J, et al. Grafts enriched with subamnion-cord-lining mesenchymal stem cell angiogenic spheroids induce post-ischemic myocardial revascularization and preserve cardiac function in failing rat hearts. Stem Cells Dev 2013;22:3087.