Formation of Cell-To-Cell Connection between Bone Marrow Cells and Isolated Rat Cardiomyocytes in a Cocultivation Model

Aims: Limited regenerative potential of cardiomyocytes (CMs) causes irreversible changes in heart tissue during pathological processes. However bone marrow mononuclear cells (BM-MNCs) can migrate to this tissue, incorporate to the area of dead or missing myocytes, and improve the global heart function. The mechanism of BMMSCs’ incorporation and interaction with CMs is not clear. Our aim was to create an in vitro model which would enable to study the interaction of BM-MNCs with CMs and to make a microscopy description of these interactions. Methods and Results: CMs were isolated from adult and newborn rats. BM-MNCs were isolated from bone marrow. BM-MNCs were added to the myocyte culture. Cell-to-cell adherence and Cx43 expression were evaluated by fluorescence microscopy, Ca2+ transients were evaluated in cardiomyocyte-BMC communication under electrical stimulation by fluo-4 fluorescence measurement. Analysis of calcein transport from BM-MNCs to CMs was performed using fluorescence microscopy. Conclusions: The adherence of BM-MNCs to CMs occurred quickly and was stable. Cx43 was detected in contact zones between BM-MNCs and CMs; pairs which displayed Cx43 positivity represented less than 1% from all BM-MNC-cardiomyocyte pairs in the coculture. Conductive structures between CMs and BM-MNCs were formed and verified by imaging calcein transfer and synchronous Ca2+ transients.


Introduction
Myocardial infarction is the leading cause of heart failure and death in developed countries. It is caused by restriction of blood supply to a part of the heart muscle. The consequent ischemia may lead to necrosis of a significant part of the heart [1,2]. Adult cardiomyocytes have a very limited potential to regenerate, and the necrotic tissue cannot be reconstituted into a functional syncytium of new cardiomyocytes [3,4]. Rather, necrotic tissue is gradually replaced by a highly organized scar containing mainly fibroblasts [5]. In the majority of cases, necrosis results in heart failure, for which today there is only one known causal therapy: heart transplantation. There are wellknown problems associated with allogeneic grafts, and insufficient numbers of organs are available for transplantation. Therefore, the potential for delivering stem/progenitor cells into the infarcted area is being intensely investigated. While different cell delivery methods have been used successfully in many cases [6], engraftment, subsequent regenerative effects, and the efficiencies of these processes require further study. Continued investigation is needed to develop a safe and efficient means of cell therapy.
Clinical studies and studies on experimental animal models have demonstrated the ability of delivered bone marrow mononuclear cells (BM-MNCs) or endothelial progenitor cells (EPCs) to improve the function of an ischemic heart. Nevertheless, the precise mechanism of cell homing in the affected heart and the cells' interaction with the residual functional cardiomyocytes (CMs) in the infarcted area remain unclear. While basic studies have indicated that BM-MNCs may transdifferentiate into CMs [7,8], other investigators suggest that BM-MNCs or EPCs play only a mechanical support role or act in vivo to restore the ion-conductive connection among the separated areas of heart tissue [9]. Another hypothesis states that BM-MNCs or EPCs are merely a source of paracrine factors that help CMs to function and to survive, or that eliminate pathophysiological fibroblast growth and left ventricular remodeling [10]. This divergence in published theories highlights the need for further investigation in this field, including developing in vitro experimental models. The aim of this study was to create conditions that would facilitate structural and functional connection in a model of BM-MNCs co-cultivated with isolated rat CMs and to perform its analysis. Our study used a traditional cocultivation setup which has been reported in several published works [11][12][13]. Genetically unmodified BM-MNCs were arranged in a microchamber with a rational concentration of neonatal or adult MCs. This arrangement permits continual microscopic analysis of tropism by BM-MNCs to the free ends of CMs (a model for interrupted CM strands in infarcted heart), quantitative analysis of CM-BM-MNC coupling (analogous with hypothetical mechanical and conductive coupling which initiate the regenerative process in vivo), and analysis of cell survival/differentiation of all cells in the coculture. Moreover, our setup performs variable confocal scanning of cells with high spatial resolution and recording of cell activity with high temporal resolution (using an ultra-high-speed camera). An important original aspect of our analysis was quantitative comparison of the interactions with myocytes of fresh and cultured BM-MNCs. A second original aspect of our work is its parallel use and comparison of neonatal and adult MCs as a substrate in cocultivation experiments. Although neonatal MCs are standardly used in almost all published cocultivations, their use as a model of in vivo adult heart tissue is still controversial and exact comparative analyses with adult MCs are both scarce and controversial in the literature [12].

Isolation of cardiomyocytes
Adult rat cardiomyocytes were obtained from young adult Wistar rats (250 ± 50 g). Rats were anesthetized by intramuscular administration of a mixture of 1% ketamine (Narkamon inj., Spofa, Czech Republic) (80 mg/kg) and 2% xylazine (Rometar inj., Spofa, Czech Republic) (12 mg/kg) into the right femoral muscle. After the loss of reflex, the sternum was quickly cut, the ribcage opened, and the beating heart cut from circulation and extracted. We proceeded using the method of enzymatic dissociation [14]. Explanted hearts were cannulated via the aorta and perfused using a peristaltic pump for 30 s with Tyrode's solution (0.9 mmol/l CaCl 2 ) at a constant flow rate of 5 ml/min. The heart was perfused with Ca 2+ free Tyrode's solution containing collagenase (Collagenase type II, Yakult, Japan) and protease (Protease type XIV, Sigma-Aldrich, Germany). The heart was then detached from the cannula and the ventricles were separated and dissociated into small fragments. The fragments were collected in Tyrode's-collagenase solution and gently mixed. After 30 s, the medium was replaced with standard Tyrode's solution (0.09 mmol/l CaCl 2 ) and mixed again for 10 s. The supernatant (more than 50% rod-shaped cells) was collected, filtered, and exposed to 0.9 mmol/l CaCl 2 . Morphology of the cells was recorded for 48 h ( Figure 1B).
Primary neonatal rat cardiomyocytes were isolated from Wistar rats 2-5 days old, as described previously [15]. Neonatal rats were anesthetized using diethyl ether and decapitated. Hearts were removed and dissociated in a balanced salt solution containing 20 mM HEPES, 120 mM NaCl, 1 mM NaH 2 PO 4 , 5.5 mM glucose, 5.4 mM KCl, and 0.8 mM MgSO 4 (pH 7.3-7.4). Cardiomyocytes were isolated by trypsin digestion (0.2% w/v), after which the cells were resuspended in a medium containing Iscove's modified Dulbecco's medium (IMDM) (Sigma-Aldrich, Germany) and Medium 199 (Sigma-Aldrich, Germany) in a 4:1 ratio and supplemented with horse serum (10%), fetal calf serum (5%), penicillin (100 U/mL), and streptomycin (100 μg/mL). The cells were plated for 1.5-2.0 h to allow for attachment and/or separation of non-myocardial cells (heart fibroblasts). The non-adhesive cells (cardiomyocytes) were transferred into a centrifugation tube, washed, then centrifuged at 150 × g rpm for 10 min. Suspension enriched in non-adhesive CMs was transferred to culture dishes coated with collagen I at density 5 × 10 4 cells/cm 2 . The cells were incubated in conditions of 21% O 2 and 5% CO 2 at 37°C. Cultivation medium was removed after 72 h and replaced by a medium containing IMDM and Medium 199 (4:1) with penicillin (100 U/ml) and streptomycin (100 mg/ml). Viability and function of the neonatal cardiomyocytes were verified by various means, and viability always exceeded 95%. Viable CMs displayed contractions at a rate of approximately 100 beats per minute. Maintenance of metabolically active mitochondria was verified using JC-1 fluorescent indicator of mitochondrial membrane potential where responsiveness to uncouplers was mandatory (Figure in Supplement-1).

Isolation of Bone Marrow Mononuclear Cells
Rat BM-MNCs were harvested and cultured, as previously described [12]. Briefly, cells were isolated from the bone marrow of femurs by flushing medium into the bone shaft (method of anesthesia of the rat based on ketamine and xylazine was the same as for the heart isolation protocol described above). Cell suspension was filtered through a 40 μm nylon cell strainer (BD Falcon, USA) and layered on Histopaque at 500×g for 30 min, mononuclear cell layer was plated in chambers of 24-well plates. Cells were grown in IMDM with 10% fetal bovine serum (FBS) at 37°C and 5% CO 2 .
The first group of BM-MNCs (named fresh BM-MNCs) was stained and used for cocultivation immediately (1h) after isolation. The second group of BM-MNCs (named cultivated BM-MNCs) was incubated under the described conditions for 3 day, the medium was replaced with fresh medium (with wash-out of non-adherent cells), and the adherent cells were grown to 80% confluence before passage was made. Four passages were performed during the following 4 weeks. Cells from the second to fourth passages (morphology at Figure  1A) were used for cocultivation experiments (cocultivation of neonatal and adult CMs was performed in parallel, as detailed in the next paragraph).

Staining of BM-MNCs and cocultivation
Before coculture experiments, fresh BM-MNCs and cultured BM-MNCs after 2-6 passages were labeled with 10 μM fluorescent dye PKH26 (Sigma-Aldrich, Germany) for 5 min at room temperature in phosphate-buffered solution. The PKH26-labeled BM-MNCs were plated at density 3 × 10 4 cells/cm 2 (fresh) or 1.5 ×10 4 cells/ cm 2 (cultured) in one 24-well dish containing 4 × 10 4 neonatal CMs/cm 2 or 1 × 104 adult CMs. Table 1 provides an overview of cocultivation experimental variants. M1 medium was used for cocultivation with neonatal or adult CMs. (Media were based on Tyrode's solution and IMDM media produced by Sigma-Aldrich, Germany. The complete composition of the M1 medium is available in Supplement 2. A comparison of the viability of CMs in IMDM [with addition of 10% FBS] and M1 medium was performed, as detailed in Results) All culture solutions were supplemented with penicillin (100 U/ml) and streptomycin (100 mg/ml). All four sets of cocultivation variants (with exceptions named below) underwent basic measurement of cell-to-cell adhesion and three different types of cellto-cell connection analysis 24 h after initialization of cocultivation, as described immediately below.

Basic detection of BM-MNCs adherence to CMs
The liquid inside the each cultivation chamber was gently resuspended by 100 uL automatic pipette (decoupling of cells in pairs which did not create real mechanical adherence but only short time adhesion), minimally 10 wide fields contained minimally 40 CMs was immediately identified in each cultivation chamber and basic visual analysis was used for quantification of the BM-MNCs with strong adherence to CMs. The percentage of adhered BM-MNC was quantified in 160 wide fields for each of 8 independent adult animal experiments and in 160 wide fields for each of 6 independent neonatal animal experiments.

Detection of Cx43 intracellular connection
BM-MNCs and CMs grown on cover slides on the bottom of dishes were fixed for 30 min in 4% paraformaldehyde. After three wash-outs with phosphate buffered saline, BM-MNCs and CMs were incubated 20 min with 0.3% Triton and stained with the primary IgG antibody anti-connexin43 for 2 h (1:400, C8093, Sigma-Aldrich, Germany). BM-MNCs and CMs were then washed twice with phosphate buffered saline and incubated for 1 h at room temperature with the secondary antibodies (1:100, F 9006 Sigma-Aldrich, Germany). To detect Cx43, fluorescence imaging was performed (confocal microscopy).

Detection of CM -BM-MNC coupling based on measurement of Ca 2+ transients
After 24 h, the coculture was washed with Tyrode's solution and incubated at 37°C and 5% CO 2 with fluo 4-AM (5 μM, Molecular Probes, Eugene, OR, USA) for 30 min. Cells were then transferred to the microscope chamber and overflowed with Tyrode's solution (pH 7.4, at 37°C). Two silver electrodes connected to a stimulator (type AT50, Chirana, Czech Republic) were used for electrical stimulation (1 Hz, 10 mA/cm 2 ). Images of beating CMs and surrounding BM-MNCs were obtained using an FITC filter set (excitation 488 nm, emission 512 nm) together with Moticam 2000 (Motic, China) and i-Speed 3 (Olympus, Japan) CCD cameras to obtain a set of images with 50 ms time crop. Synchronous fluo-4 fluorescence signals in CMs and neighboring BM-MNCs were recorded. Periodic increasing and decreasing intensity of fluorescence signals with the same time periods for both cells were considered as synchronic flashes. Sets of frames with 50 ms time crop were converted to pseudocolor images (transformation of intensity to rainbow color scale, using Olympus i-Speed lab software, Olympus, Japan).

Detection of CM -BM-MNC coupling based on measurement of calcein diffusion
BM-MNCs were incubated with 2.5 μM calcein-AM for 60 min at 37°C instead of PKH26 before cocultivation. BM-MNCs were washed out and used for cocultivation (the same setup as described in Table  1). Detection of calcein diffusion from BM-MNC to CM was performed using fluorescence microscopy.

Analysis of basic markers of cardiomyocyte transdifferentiation -recording of cell contractility and development of alpha-actinin sarcomeric structure
Occurrence of spontaneous active contraction of BM-MNCs was evaluated using a CCD camera on a fluorescence microscope to capture a set of images with 50 ms time crop. Contractions were also recorded during and after application of electrical stimulation. Antiactinin antibodies (primary antibodies A7732, Sigma-Aldrich, Germany) were used for staining the BM-MNCs. Anti-mouse IgG1 conjugated with FITC was used as a secondary antibody (1:100, F 9006 Sigma-Aldrich, Germany). FITC-positive intracellular structures of BM-MNCs were compared with positive controls for sarcomeric actin structure in freshly isolated adult and neonatal CMs.

Analysis of BM-MNCs adherence to CMs and CM -BM-MNC coupling (detection of Cx43 created between BM-MNCs and CMs)
Fluorescence microscopy analysis of the CMs cocultures with and BM-MNCs showed strong adherence of some BM-MNCs to CMs ( Figure 1C and 1D). The proportions of fresh and cultured BM-MNCs that strongly adhered to adult or neonatal CMs are shown in Table 2.

Detection of CM -BM-MNC coupling based on measurement of calcein molecules diffusion
The analysis of CMs which had contact with calcein-positive BM-MNCs showed that some CMs became calcein positive during the time of cocultivation. An illustrative image from fluorescence microscopy is shown in Figure 4. This suggests a possibility of calcein transfer through gap junctions or other connection structures between CMs and BM-MNCs. We found that percentage of adult BM-MNC -CM pairs that display this calcein transfer positivity was 3.

Discussion
Clinical studies and studies on experimental animal models have demonstrated the ability of delivered bone marrow mononuclear cells (BM-MNCs) or endothelial progenitor cells (EPCs) to settle down into the affected area of myocardium and to some extent improve the function of an ischemic heart [16]. The commonly acknowledged scheme of BM-MNC homing consists of three steps: 1) chemo-attractant factors are released by the infarcted area, 2) BM-MNCs migrate from the vessel through the endothelium, and 3) BM-MNCs adhere to CMs and can make conductive junctions with the CMs or eventually start transdifferentiation processes. Nevertheless, the precise mechanisms of the BM-MNCs homing in on the affected myocardium -and especially their interaction with the residual CMs in the infarcted area -remain unclear.
Aspects of BM-MNCs' interaction with the infarcted tissue (changes in morphology of the BM-MNCs, quantification of BM-MNCs' integration into the muscle, adhesion to endothelial cells and to the extracellular matrix) have been analyzed using electron and fluorescence microscopy on heart samples extracted and fixed postmortem [17][18][19]. The junctions created between BM-MNCs and CMs have also been analyzed [12] (application of cells to an extracorporeal rat beating heart model and subsequent analysis of fixed slices of the heart tissue). Direct analysis of BM-MNC -CM interaction in situ would be ideal, but this has not yet been achieved. The first intravital-microscopy setups for MSC tracking have been developed [20,21], but the problems with resolution and long-time invasiveness have not yet been settled. Until intra-vital microcopy will be perfected, research must rely on other techniques that have been developed to mimic the microstructure of infarcted heart in vivo and to allow the simulation and systematic observation of interaction between uncoupled myocytes and progenitor/stem cells. This type of experiment is known as a cocultivation experiment, and it is used also in this work.
The number of similar cocultivation experiments which simulate interaction of cardiac cells with BM-MNCs is very limited in the literature. We found no more than 15 original publications from the past 8 years, as well as several subsequent upgrades which were based on heterotypic cell-pair interactions in microchambers and a cell strands coculture setup on opto-electronical microchips (most of which are from the past 4 years; [22][23][24]). There also have been reports of state-of-the-art upgrades, including tissue bundles examination using external stimulatory forces [25] and micro-bioreactors with large ventricular slice [26,27]. To the best of our knowledge, however, almost all of these state-of-the-art cultivation settings have utilized modified MSCs (typically transfected for upgrade of ion channel) or another atypical type of stem/progenitor cells (iPS, ESC, genetically modified fibroblast) and these experiments principally did not analyze the simple interaction of fresh or cultured BM-MNCs.
Moreover, if we consider especially cocultivation experiments with unmodified BM-MNCs or MSCs, it can be seen that most of the published cocultivation experiments use only neonatal CMs as a substrate and the interactions are not well compared (or comparable) with those of adult CMs as these actually occur in infarcted heart. Our work is among the first in which neonatal and adult CMs are used in parallel for cocultivation experiments and whereby both myocyte types are handled under the same conditions (growth medium, concentration of the cell mass in the coculture related to probability of the cell interaction, experiment timeframe).

Cell survival in coculture and BM-MNC -CM adherence
The cultivation media influence the viability and reaction of the cells. This study shows that the M1 medium (details in Materials and Methods section) is better for cultivation of adult CMs than IMDM with 20% FBS (routinely used for the cultivation of BM-MNCs). M1 medium has ingredients similar to those of Tyrode's solution, which is routinely used in short-time experiments with freshly isolated CMs [14,28], but FBS contents and some ion concentrations are different. Despite relatively high viability, changes in shape from rectangular to elliptic were observed in 40-60% of adult viable CMs during the first 24 h. These changes had been described previously [29] and they remain a main reason why we limited the time frame for our analyses to 24 h. These changes of shape are similar to those seen in FBS-free Tyrode's solution.
Our results show that 2-5% of fresh BM-MNCs and 3-6% of cultured BM-MNCs strongly adhered to the CMs after 24 h. The concentrations of BM-MNCs and CMs in coculture used in this study produce very similar results in terms of the ability for and amounts of BM-MNC adherence as reported in previously published papers [12,13,30]. Decrease of BM-MNCs' concentration in coculture leads to a decreasing rate of adhered BM-MNCs, as we found during detailed analysis (data not shown). We hypothesize that the observed strong adherence of cells may play a role in vivo in a mechanical stabilization of tissue in the infarcted zone.

Communication between cells
Another aim was to evaluate the conductive function of the Cx43positive structures between BM-MNCs and CMs. Microphotographs from fluorescence microscopy show that the fluorescence probe calcein was transported from BM-MNCs to CMs. The results also proved synchronous Ca 2+ transients between BM-MNCs and CMs. BM-MNCs displayed periodic decrease and increase of cytosolic Ca 2+ concentration, and the Ca 2+ transients were synchronized with Ca 2+ transients in the neighboring CM.
Conducting structures were formed between CMs and cultured BM-MNCs, and calcein transfer occurred through these, where the fraction of positive cells was about 3% in BM-MNC -adult CM coculture and 5% in BM-MNC -neonatal CM coculture. Some other authors have observed more than 50% of cocultured cells displaying calcein-conducting positivity based on fluorescence-activated cell sorting [11,31]. Nevertheless, nonautomated precise optical fluorescence microscopy analysis did not show a similar 50% positivity, because flow cytometry may in this case suffer from overestimation and artificial results for this type of experiment [3]. Ion-conductive structures between BM-MNCs and CMs have been closely observed on neonatal CMs and cultured BM-MNCs in several published studies [30,32]. Results have shown that 24 h are sufficient for establishing the ion-conductive structure between BM-MNCs and neighboring CMs [30,33,34]. Analyses of adult rat CMs (as a cocultivation anchor) combined with fresh and cultured BM-MNCs (as fraction progenitor injected into the coculture) from the same species have not yet been published (only [11] described one principally similar experiment), and our results may thus point to a feasible path for further research.

Evaluation of transdifferentiation
No spontaneous contractility of BM-MNCs or contraction of BM-MNCs upon electrical stimulation by electrodes has been observed in cases of 24 h of cocultivation of BM-MNCs with neonatal or adult myocytes. In addition, no BM-MNCs displayed typical sarcomeric alpha-actinin structures. Similar conclusions were presented in a case of contact with adult CMs [12]. Transdifferentiation of BM-MNCs after coculture with neonatal CMs remains controversial. Our shortterm experiments produced no direct evidence of such transdifferentiation.

Limitations of the study
A basic limitation of our in vitro experimental setup consists in the simplified access of BM-MNCs to CMs. In vivo, the path for BM-MNCs to CMs consists of three steps: 1) chemo-attractant factors are released by the infarcted area, 2) BM-MNCs migrate from the blood vessel through the endothelium, and 3) BM-MNCs adhere to the CMs in tissue [35]. In our in vitro model, there are no vessels and no transendothelial border. BM-MNCs are present in culture directly next to the cardiomyocytes.
Another limitation of our experimental results derives from the contractile properties of adult CMs. Adult isolated CMs with nonaffected ion balance in their cytoplasm do not exhibit spontaneous contractility (only about 5% of CMs with affected ion balance in the cytoplasm exhibited periodic or non-periodic spontaneous contractions). On the other hand, neonatal CMs display spontaneous contractions with a high rate, approximately 100 beats per minute. Permanent contractions of CMs in vivo (or at least permanent periodic polarization of CMs) can be a necessary positive stimulus supporting adherence of BM-MNCs to CMs and formation of ionconductive structures between BM-MNCs and CMs, as well as possible transdifferentiation of BM-MNCs [36].

Conclusions
Fresh BM-MNCs and cultivated BM-MNCs are able to survive under specific conditions for 24 h in coculture with neonatal or adult CMs. About 5% of fresh BM-MNCs and cultured BM-MNCs in our in vitro setup strongly adhered to neonatal cardiomyocytes, and about 1.5% or 3% of fresh BM-MNCs or cultured BM-MNCs in our in vitro setup adhered to adult CMs. Our results show this relatively strong mechanical connection between BM-MNCs and CMs and very rapid formation in the first minutes and hours after first accession of the cells. This should be an important mechanical premise for following electrophysiological coupling of the cells.
Contact regions between cultured BM-MNCs and CMs (both adult and neonatal) displayed structural connection (Cx43 positivity) and functional conduction structure (vital dye intercellular transport and ion intercellular transport) in our coculture experiments. Nevertheless, the relative number of BM-MNC-CM pairs with functional conductive structure was limited. No transdifferentiation of BM-MNC was observed and no fresh BM-MNCs or cultured BM-MNCs developed cardiomyocyte-like differentiation during the 24 h coculture experiment.