Fibrin Glue-aided, Instant Epicardial Placement Enhances the Efficacy of Mesenchymal Stromal Cell-Based Therapy for Heart Failure

Transplantation of mesenchymal stromal cells (MSCs) is a promising new therapy for heart failure. However, the current cell delivery routes result in poor donor cell engraftment. We therefore explored the role of fibrin glue (FG)-aided, instant epicardial placement to enhance the efficacy of MSC-based therapy in a rat ischemic cardiomyopathy model. We identified a feasible and reproducible method to instantly produce a FG-MSC complex directly on the heart surface. This complex exhibited prompt, firm adhesion to the heart, markedly improving initial retention of donor MSCs compared to intramyocardial injection. In addition, maintenance of retained MSCs was enhanced using this method, together contributing the increased donor cell presence. Such increased donor cell quantity using the FG-aided technique led to further improved cardiac function in association with augmented histological myocardial repair, which correlated with upregulation of tissue repair-related genes. We identified that the epicardial layer was eliminated shortly after FG-aided epicardial placement of MSCs, facilitating permeation of the donor MSC’s secretome into the myocardium enabling myocardial repair. These data indicate that FG-aided, on-site, instant epicardial placement enhances MSC engraftment, promoting the efficacy of MSC-based therapy for heart failure. Further development of this accessible, advanced MSC-therapy is justified.


FG-aided, instant epicardial placement markedly improved both initial retention and survival of donor MSCs compared to IM injection.
Female Lewis rats with post-MI ICM were transplanted with 4 × 10 6 MSCs, isolated from bone marrow of syngeneic male Lewis rats, using either FG-aided instant epicardial placement (FG-MSC group) or IM injection of MSC suspension (IM-MSC group). For epicardial placement, 40 μl of fibrinogen/MSC suspension and 40 μl of thrombin solution was simultaneously added on to the heart surface in a dropwise manner. This resulted in prompt generation of FG incorporating MSCs, which covered the infarct and peri-infarct myocardial areas as we targeted. This complex quickly and firmly adhered to the heart surface and remained in situ without dripping off. Real time PCR analysis of the male specific Sry gene demonstrated that the initial (1 hour after transplantation) retention of donor cells was markedly increased in the FG-MSC group compared to the IM-MSC group (Fig. 1a). Thereafter, donor cell presence was reduced in both groups; however, throughout the time course up to day 28, greater donor MSC presence was consistently observed in the FG-MSC group. Of note, the day 3 donor cell presence in the FG-MSC group was equivalent to 1-hour MSC presence in the IM-MSC group. Furthermore, calculated donor cell loss ratios between day 3 and 7 post-transplantation and between day 7 and 28, were both attenuated in the FG-MSC group (Fig. 1b), suggesting enhanced maintenance of retained MSCs after FG-aided epicardial placement.

MSCs engrafted using FG-aided epicardial placement remained on the heart surface.
Immunohistostaining revealed that the majority of donor MSCs in the FG-MSC group were retained on the epicardial surface until day 28 ( Fig. 1c). At 1 hour after placement, the FG-MSC complex covered all targeted areas of the infarct and surrounding border areas evenly. The FG-MSC layers were generally compacted at 1 hour, but became more spread by day 3 as donor MSCs became more elongated (Fig. 1d). However, this layer decreased in size by day 28. One hour after epicardial placement, FG was observed surrounding, round-shaped donor MSCs within the FG-MSC complex (Fig. 1d). At day 3, FG was still observed occasionally between MSCs, which showed an elongated morphology. At day 28, MSCs were more densely accumulated, and FG was hardly observed, suggesting degradation and/or absorption of FG by this time point after epicardial placement. In contrast, transplanted MSCs in the IM-MSC group formed isolated cell-clusters in the myocardium, while the size of these clusters decreased gradually over time (Fig. 1e). Apoptotic death of transplanted MSCs, assessed with cleaved caspase 3 staining, was reduced in the FG-MSC group compared to the IM-MSC group (Fig. 2a). On the other hand, proliferation of donor MSCs, as measured using immunostaining for Ki67, was not different between the groups; Ki67 + MSCs were rarely detected in either IM-MSC and FG-MSC group (Fig. 2b). We did not observe any cardiac Troponin T + cardiomyocytes derived from CM-DiI-labelled donor MSCs indicating a lack of cardiomyogenic differentiation of donor MSCs. In addition, there were no Oil-red O + adipocytes or Alizarin red + osteocytes originated from DiI + donor MSCs at day 28 (Supplemental Fig. S1), indicating safety of this treatment.
The epicardial cell layer was eliminated after FG-aided placement of MSCs. It is generally agreed that transplanted MSCs improve cardiac function via the paracrine effect based on their secretome 18 . Given that the majority of donor MSCs were retained on the heart surface after FG-aided epicardial placement (Fig. 1), we cautiously considered that the presence of the epicardial cell layer might pose a barrier to epicardially-placed MSCs in terms of communicating with the host myocardium to establish a paracrine effect. In this regard, we found that the ICAM-1 + epicardial cell layer was present in the normal heart and the sham-treated ICM hearts (Sham group), while these epicardial cells disappeared by day 1 after FG-aided epicardial placement of MSCs (Fig. 2c). In response to this finding, we found vascular formation in the FG-MSC complex, which included host-derived (DiI-negative) PECAM1 + endothelial cells (Fig. 2d). These data suggest that the epicardium would not block cellular and molecular interactions between the FG-MSC complex and the myocardium, including  FG-aided epicardial placement of MSCs upregulated reparative genes more extensively. We then investigated whether FG-aided, instant epicardial placement indeed increased upregulation of reparative factors relevant to myocardial repair. Quantitative RT-PCR analysis demonstrated that myocardial expression of a range of reparative factors, including Il10, Timp1, Vcam1, Cxcl12, Igf1, Hif1a and Tgfb, was upregulated in the IM-MSC group at day 3 compared with the Sham group (Fig. 3). This upregulation was more evident in the FG-MSC group at day 3. Furthermore, the majority of these genes remained upregulated by day 28 in the FG-MSC group, while those in the IM-MSC group returned to the baseline values equivalent to the Sham group. The FG-Cont group (epicardial placement of FG only in the ICM rat) induced upregulation of a subset of these genes, including Vcam1, Cxcl12, Igf1, Hif1a, Tmsb4, and Tgfb compared to the Sham group. However, this upregulation was much less extensive and not sustained compared to that of the FG-MSC group.

FG-aided epicardial placement of MSCs augmented myocardial tissue repair.
Corresponding to the upregulation of reparative genes as described above, histological analysis identified enhancement of tissue repair of the failing myocardium in the ICM rat hearts after FG-aided, instant epicardial placement of MSCs. At day 28 after treatment, Isolectin B4 staining indicated an increased capillary density in the peri-infarct area in the FG-MSC group compared to all other groups (Fig. 4a). In support of this observation, the percebtage of Ki67 + PECAM1 + proliferating endothelial cells was increased in the FG-MSC group compared to all other groups (Fig. 4b). Post-MI cardiomyocyte hypertrophy and interstitial collagen deposition in the peri-infarct areas were both attenuated in the FG-MSC group, compared to all other groups (Fig. 5a,b). Recently, it was suggested that alternative activated macrophages play an important role in the MSC-mediated paracrine effects as a secondary effector 19 . Consistent with this, we observed an accumulation of CD163 + alternatively activated macrophages in the peri-infarct area most notably in the FG-MSC group (Fig. 5c). The IM-MSC group also exhibited a reduction in cardiomyocyte hypertrophy, decrease in collagen deposition, and increase in CD163 + cell accumulation compared to the Sham group. However, the degrees of these effects in the IM-MSC group were less extensive when compared to the FG-MSC group.

FG-aided epicardial placement of MSCs further enhanced cardiac function. Myocardial repair by
FG-aided, instant placement of MSCs resulted in improved cardiac function recovery in the rat ICM model, compared to IM injection of MSCs (Table 1). At day 28 after MSC transplantation, echocardiography demonstrated that left ventricular ejection fraction (LVEF) and left ventricular fractional shortening (LVFS) in the FG-MSC and IM-MSC groups were increased as compared to the Sham group, with a greater increase measured for the FG-MSC group. In addition, the FG-MSC group showed reduced LV diameters compared to the Sham group. There was no significant difference in LVEF, LVFS, or LV dimensions between the Sham and FG-Cont groups. Heart rates were comparable among groups for echocardiography and catheterization. Baseline values prior to MSC transplantation were equivalent between groups. Consistently, cardiac catheterization showed the highest developed pressure, max dP/dt and min dP/dt, in the FG-MSC group among the group studied (Table 1). Left ventricular end-diastolic pressure was also reduced in the FG-MSC group, compared to other groups.

Discussion
This study demonstrated feasibility and therapeutic efficacy of FG-aided, instant epicardial placement of MSCs in a rat model of ICM. It was feasible and reproducible to instantly produce the FG incorporating MSC complex directly on the heart surface. This mixture promptly underwent gelation to form the FG-MSC complex, which firmly adhered to the targeted epicardial surface, enabling enhanced initial retention of MSCs compared to IM injection. No suture, additional glue or pericardial patch was required to fix the complex onto the heart, unlike epicardial placement of prefabricated FG constructs 15,16 . In addition, subsequent loss of retained MSCs was attenuated after FG-aided epicardial placement. These increased initial retention and improved maintenance of retained MSCs together amplified the quantity of engrafted MSCs, which corresponded to augmented myocardial tissue repair and improved cardiac function recovery. This enhanced therapeutic effect correlated with amplified myocardial upregulation of tissue repair-related genes. Freshly mixed suspension composed of fibrinogen, thrombin and MSCs is malleable enough to tightly fill the uneven heart surface. Such close-fitting adhesion would be beneficial for allowing permeation of biological factors secreted from donor MSCs into the myocardium. Our histological studies detected that the epicardial layer, which might block communications between MSCs and the myocardium, disappeared shortly after epicardial placement of the FG-MSC complex. This would allow permeation of the donor MSC secretome into the damaged myocardium. Collectively, FG-aided, on-site, instant epicardial placement of MSCs augmented initial retention and subsequent engraftment of MSCs, offering an enhanced therapeutic efficacy for ICM.
The enhanced donor cell quantity by FG-based cell transplantation has been reported previously 13,14 , and our results uncovered that this was a combined result from both increased initial retention and reduced subsequent densely aggregated. Scale bar = 100 µm. (d) Immunohistostaining analysis detected PECAM1 + (green) and CM-DiI + (yellow) vessels, suggesting trans-differentiation of donor MSCs to endothelial cells in both FG-MSC and IM-MSC groups (white arrows). As explained above, CM-DiI-labeled MSCs exhibited yellow color, but these yellow cells were distinguishable from the PECAM1 + CM-DiI + cells, which showed a different yellow color. In addition, PECAM1 + (green) but CM-DiI − vessels were observed, indicating host-derived endothelial cells (blue arrows). Scale bar = 100 µm.
loss of donor MSCs. Increased initial retention was achieved mainly by the unique nature of fibrin, which allowed the FG-MSC complex to firmly adhere to the heart, while securely retaining MSCs within. This feature attenuated the initial leakage of donor cells after transplantation, which is a major cause of the poor donor cell engraftment following IM and intracoronary injection 10,20 . Our in vivo results revealed that almost all transplanted MSCs were retained at 1 hour after FG-aided epicardial placement, while in contrast less than a half of donor cells were retained after IM injection of the MSC suspension. In addition, we uncovered attenuation of subsequent loss of donor MSCs both between day 3 and 7 and between day 7 and 28 after FG-aided epicardial placement compared to IM injection, additionally contributing to the increased quantity of engrafted donor MSCs. Underpinning this effect, apoptosis of transplanted MSCs was reduced after FG-aided epicardial placement compared to IM injection, while proliferation of transplanted MSCs rarely occurred in either cell-delivery method. Also, it is unlikely that engrafted, surviving MSCs are flushed out or migrate out from the heart at these later phases post-transplantation 20 . Collectively, the attenuation of subsequent loss of donor MSCs by FG-aided epicardial placement can be explained by reduced donor cell death. We speculate that increased survival of donor MSCs may be achieved by multiple FG-mediated mechanisms. Firstly, FG is known to provide a suitable three-dimensional scaffold to MSCs 21,22 , maintaining the high viability of incorporated MSCs. Secondly, we could expect FG to protect incorporated MSCs from external hazards, i.e. free radicals, by acting as a physical barrier. Thirdly, as reported in previous papers 22,23 , fibrin-based culture amplifies expression of anti-apoptotic factors in MSCs, including HIF1α and HGF, which would contribute to protection from insult in an autocrine manner 24 . Finally, while IM injection is known to induce myocardial inflammation that results in donor cell death 7,8 , epicardial placement does not cause such harmful inflammation.
Another possible mechanism by which FG-aided epicardial placement of MSCs achieved the augmented therapeutic effect might be enhancement of donor cell function via interactions with FG. As stated above, it has been reported that culture with fibrin increases the secretion of reparative factors, at least including HIF1α and HGF, from MSCs 22,23 . These upregulated factors could also play a role in the MSC-derived paracrine effect for myocardial repair [25][26][27] . Although the molecular mechanism underpinning the interaction between FG and MSCs that leads to the upregulation of reparative factors in MSCs is not certain, this may be explained by a characteristic component of the fibrin matrix, arginine-glycine-asparagine (RGD). RGD domain is known to bind integrin families expressed on the cell membrane of MSCs to mediate their intracellular signal cascades, resulting in amplified expression of reparative genes [28][29][30][31] . Further focused studies are needed to clarify the mechanism of FG-mediated enhancement of donor MSC function for myocardial repair.
Considering the clinical translation, FG-aided, instant epicardial placement of MSCs may be convenient and effective if added to coronary artery bypass grafting (CABG). The exposed heart during CABG offers an ideal opportunity to perform this MSC therapy in the safest and most reproducible manner. Although CABG is routinely performed to treat ICM patients, its outcome is often not satisfactory 32 . Of note, an addition of adult stem cell therapy is proven to augment the efficacy of CABG 33,34 . In the clinical scenario, an FG kit (containing fibrinogen and thrombin solutions) and ready-to-use allogeneic MSCs may be delivered from the manufacturer and/ or hub stem cell center, obviating the necessity of a clinical-grade, cell processing facility at each hospital where the treatment is undertaken. At the end of bypass anastomosis, FG-aided, instant epicardial placement of MSCs can be easily completed in a short time. Thus, this highly accessible and effective treatment (CABG + FG-aided, instant epicardial MSC placement) has a great potential to be a widely-adopted standard treatment for ICM. In addition, other applications of FG-aided, instant epicardial MSC placement, including sole application as well as combination with other heart surgery, i.e. left ventricular assist device implantation and valve repair/replacement, for the treatment of ischemic and non-ischemic heart failure, are also promising.
One possible limitation in our study may be the use of the fluorescent dye (CM-DiI) for histological donor cell trafficking. Although utility of this labeling method for over 4 weeks has been repeatedly reported 2-4,7,8 , we cannot exclude the possibility that CM-DiI might be released from dead/damaged donor MSCs and be subsequently taken up by host cells including macrophages. As such, to provide a reliable and quantitative assessment of donor cell retention/survival, we have carefully performed qPCR for Sry. Actually, the results of donor cell presence were consistent between the CM-DiI-based histological and PCR assessments. Other concerns associated with epicardial placement of the FG-MSC complex may include a risk of altered hemostasis, cardiac tamponade or adverse ventricular contractile effect. However, we did not observe bleeding, clot formation or pericardial fluid collection in either FG-MSC or FG-Cont groups at 1 hour, 3 days or 28 days. Furthermore, we have confirmed that LVEDP was the lowest, and min dP/dt was the most improved in the FG-MSC group among all groups studied at Day 28. This indicates that epicardial placement of the FG-MSC complex will not cause significant adverse constrictive force to the heart. Given our result showing that fibrin glues were dissolved and absorbed by day 28 after placement, with loss of the majority of donor cells (Fig. 1c), it is unlikely that any adverse constrictive effect would occur after this period. Finally, it will be interesting to investigate whether the effect of FG to improve therapeutic effects of cell therapy is limited to bone marrow-derived MSCs or widely applicable to other donor cell types, i.e. adipose tissue-derived MSCs and cardiac resident progenitor cells.
In summary, the results of this study represent the proof-of-concept data indicating the feasibility and efficacy of FG-aided, instant epicardial placement of MSCs for the treatment of ICM. The use of FG offers multiple benefits to this approach, including not only improvement of initial retention but also maintenance of retained cells. This collectively results in enhanced therapeutic effects of this advanced MSC-based therapy compared to the current cell-delivery method. Further pre-clinical and clinical development of this user-friendly, highly-effective MSC-based therapy for ICM is warranted.

Materials and Methods
All studies were performed with the approval of the institutional ethics committee at Queen Mary University of London and the Home Office, UK. The investigation conforms to the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals (US National Institutes of Health Publication, 1996). Assessments were carried out in a blinded manner wherever possible. Sample numbers are stated in Figure Legends or Table. Isolation, culture and labeling of rat bone marrow-derived MSCs. MSCs were isolated from the bone marrow of the tibias and femurs of male Lewis rats (100-150 g, Charles River UK, Margate, UK) and expanded as described previously [2][3][4]35 . Isolated cells were cultured in αMEM with 20% inactivated fetal bovine serum containing L-glutamine, penicillin and streptomycin, at 37 °C in a humidified atmosphere containing 95% air and 5% CO 2 . (incubator: Binder, Germany). Culture medium was aspirated and replaced every 2-3 hours without additional washing. When cell confluency reached 80-90%, cells were passaged by detachment using 0.25% Trypsin in 0.2% EDTA (Sigma). Plating concentrations for subsequent passages were approximately 1 × 10 4 cells/cm 2 . MSCs at passage 3 or 4 were used for all studies.  For tracking of donor cells, MSCs were labeled with a fluorescent dye, CM-DiI (Molecular Probes, Paisley, UK) as previously described [2][3][4] . We confirmed that >95% of the cells were positively stained with <2% cell death prior to transplantation (data not shown).
Induction of ICM rat model and MSC transplantation. MI was induced in female Lewis rats (150-200 g; Charles River UK) by ligating the left coronary artery through left thoracotomy under isoflurane anesthesia and mechanical ventilation as described previously 3,4,8 . At 4 weeks, rats were subjected to echocardiography, and those showing inappropriate left ventricular ejection fraction (LVEF > 45% or < 25%) were excluded from the study 3,4,8 . Included animals were randomly allocated into 4 treatment groups. For the FG-MSC group, the heart was exposed through re-left thoracotomy under mechanical ventilation. Then, 40 μl of phosphate-buffered saline (PBS) containing 174 mg/ml human fibrinogen, 60 U/ml human coagulation factor XIII, and 1000 KIU/ ml bovine aprotinin, in which 4 × 10 6 MSCs from male Lewis rat bone marrow were suspended, was added onto the epicardial heart surface, together with 40 μl of 5.88 mg/ml calcium chloride solution containing 500 IU/ml human thrombin, in a dropwise manner using two micropipettes, aiming to cover the infarct and surrounding peri-infarct myocardial areas. These solutions were from a Beriplast P Combi-set (CSL Behring). The volume of each solution was chosen as a result of a pilot study. For the IM-MSC group, 4 × 10 6 of MSCs were suspended in 200 μl PBS and intramyocardially injected into two sites (100 μl each), targeting the border and infract areas as we have previously described 2 , The Sham group received an open/close chest procedure at 4 weeks after coronary artery ligation, whilst the FG-Cont group consisted of FG (no MSC) placed on the heart surface to cover the infarct and surrounding peri-infarct areas in the same manner as in the FG-MSC group.
Cardiac function measurement. Transthoracic echocardiography was performed at 4 weeks after left coronary artery ligation (for pre-treatment data) and at day 28 post-treatment using Vevo-770 echocardiography machine (VisualSonics, Amsterdam, Netherlands) under 1.5% isoflurane inhalation via a nose cone 3,4 . LVEF and LVFS were calculated from the data obtained with 2-dimensional tracing. All data were collected from at least 3-5 different measurements in a blinded manner. In addition, hemodynamic parameters were measured by using cardiac catheterization (SPR-320 and PVAN3.2; Millar Instruments, Houston, TX) as previously described 3,4 . Briefly, under general anesthesia using 1.5% isoflurane inhalation and mechanical ventilation, the catheter was inserted into the left ventricular cavity through the right common carotid artery. Intra-LV and intra-aortic pressure signals were measured with a transducer and conditioner (MPVS-300; Millar Instruments) and digitally recorded with a data acquisition system (PowerLab 8/30; ADInstruments, Oxford, UK). All data were collected from at least 5 different measurements per animal in a blinded manner.
Quantitative assessment of donor cell presence. To quantify the presence of engrafted male rat MSCs in the syngeneic female rat heart, presence of the Y chromosome-specific Sry gene was quantitatively assessed by TaqMan real-time PCR (Prism 7900HT; Applied Biosystems) [2][3][4]7 . At the chosen time after treatment, the left ventricular myocardium was collected (with removal of atriums and right ventricular free walls), genomic DNA extracted using the DNeasy Blood&Tissue kit (Qiagen), and Sry analysis performed in technical duplicate. The signal in each sample was normalized to the amount of DNA by measuring the autosomal single-copy gene Spp1 (osteopontin) as an internal standard. To generate a standard curve, left ventricular myocardium from a female rat at day 28 after left coronary artery ligation were mixed with either 1 × 10 7 , 1 × 10 6 , 1 × 10 5 or 1 × 10 4 of male rat MSCs, and processed for Sry analysis (n = 3). Donor cell loss between day A and B was calculated using obtained donor cell presence as follows: % loss of donor cells between day A and day B = (donor cell presence at day A − donor cell presence at day B) ÷ donor cell death at day A × 100.