Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

Contents lists available at ScienceDirect

Journal of Molecular and Cellular Cardiology

j ourna l homepage: www.e lsev ie r.com/ locate /y jmcc

Original article

Mesenchymal stromal cells affect cardiomyocyte growth through juxtacrine Notch-1/Jagged-1 signaling and paracrine mechanisms: Clues for cardiac regeneration

Chiara Sassoli a,1, Alessandro Pini a,1, Benedetta Mazzanti b, Franco Quercioli c, Silvia Nistri a,Riccardo Saccardi b, Sandra Zecchi- Orlandini a, Daniele Bani a, Lucia Formigli a,⁎a Dept. of Human Anatomy, Histology & Forensic Medicine, University of Florence, Viale G.B. Morgagni, 85, 50134 Florence, Italyb Dept. of Haematology, Placental Blood Bank, Careggi Hospital, University of Florence, Viale G.B. Morgagni, 85, 50134 Florence, Italyc National Institute of Optics, Sesto Fiorentino, Florence, Italy

Abbreviations: bFGF, basic fibroblast growth factor;differential interference contrast; Dil, dialkylcarbocyaninEagle's medium; FBS/FCS, fetal bovine/calf serum; HBSS,HEPES, 4-2-hydroxyethyl-1-piperazinyl-ethansolfonicinterleukin; LIF, leukemia inhibiting factor; M-CSF, mfactor; MIG, interferon-γ-inducedmonokine; MIP2, mac2; MSCs, mesenchymal stromal cells; PCR, polymeraseinterest; RT-PCR, reverse transcription-PCR; siRNA, shovascular endothelial growth factor 1.⁎ Corresponding author. Tel.: +39 055 4271 809; fax

E-mail addresses: csassoli@unifi.it (C. Sassoli), alessabenedetta.mazzanti@unifi.it (B. Mazzanti), franco.quercsilvia.nistri@unifi.it (S. Nistri), riccardo.saccardi@unifi.it(S.Z.- Orlandini), bani@unifi.it (D. Bani), formigli@unifi.

1 C. Sassoli and A. Pini contributed equally to this ma

0022-2828/$ – see front matter © 2011 Elsevier Ltd. Aldoi:10.1016/j.yjmcc.2011.06.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 March 2011Received in revised form 19 May 2011Accepted 1 June 2011Available online 13 June 2011

Keywords:Neonatal mouse cardiomyocytesMesenchymal stromal cellsNotch-1Jagged-1Cell growth

The possibility to induce myocardial regeneration by the activation of resident cardiac stem cells (CSCs) hasraised great interest. However, to propose endogenous CSCs as therapeutic options, a better understanding ofthe complex mechanisms controlling heart morphogenesis is needed, including the cellular and molecularinteractions that cardiomyocyte precursors establish with cells of the stromal compartment. In the presentstudy, we co-cultured immature cardiomyocytes from neonatal mouse hearts with mouse bone marrow-derived mesenchymal stromal cells (MSCs) to investigate whether these cells could influence cardiomyocytegrowth in vitro. We found that cardiomyocyte proliferation was enhanced by direct co-culture with MSCscompared with the single cultures. We also showed that the proliferative response of the neonatalcardiomyocytes involved the activation of Notch-1 receptor by its ligand Jagged-1 expressed by the adjacentMSCs. In fact, the cardiomyocytes in contact withMSCs revealed a stronger immunoreactivity for the activatedNotch-intracellular domain (Notch-ICD) as compared with those cultured alone and this response wassignificantly attenuated when MSCs were silenced for Jagged-1. The presence of various cardiotropiccytokines and growth factors in the conditioned medium of MSCs underscored the contribution of paracrinemechanisms to Notch-1 up-regulation by the cardiomyocytes. In conclusions these findings unveil apreviously unrecognized function of MSCs in regulating cardiomyocyte proliferation through Notch-1/Jagged-1 pathway and suggest that stromal-myocardial cell juxtacrine and paracrine interactions may contribute tothe development of new and more efficient cell-based myocardial repair strategies.

CSCs, cardiac stem cells; DIC,e; DMEM, Dulbecco's modifiedHank's Balanced Salts Solution;acid; HS, horse serum; IL,

acrophage colony stimulatingrophage inflammation protein-chain reaction; ROI, region ofrt interference RNA; VEGF-1,

: +39 055 4379500.ndro.pini@unifi.it (A. Pini),ioli@ino.it (F. Quercioli),(R. Saccardi), zecchi@unifi.itit (L. Formigli).nuscript.

l rights reserved.

© 2011 Elsevier Ltd. All rights reserved.

1. Introduction

There is growing support for the hypothesis that adult heartharbors a pool of resident cardiac stem/progenitor cells (CSCs),located mainly into epicardial cardiogenic niches [1,2] capable ofmaintaining tissue homeostasis and trigger heart repair [3–5].

Although these cells can regenerate lost cardiac muscle in lowervertebrates [6], they are unable to reconstruct heart tissue afterextensive injury in mammals, including humans [7–9]. Nevertheless,the discovery of stem cells within the mammalian heart has renewedthe interest for cardiac regeneration, based on the idea that recruitingresident cells inherently programmed to reconstitute the damagedmyocardium may achieve better results than forcing extra-cardiacstem cells to differentiate into contractile myocytes [10,11]. Indeed,experimental models and clinical trials performed using differentexogenous stem cell lineages transplanted into the post-ischemicheart have only demonstrated a modest cell engraftment andmyocardial regenerative capacities, being the observed beneficialeffects mainly attributable to the paracrine improvement of tissueremodeling and preservation of the residual myocardium by theengrafted cells [12–16].

It is becoming clear that, to propose endogenous CSCs as a target oftherapeutic options for cardiac regeneration, a better understandingof the mechanisms that regulate heart morphogenesis is required,with special reference to the cellular interactions that immaturecardiomyocytes may establish with other cells of the stromal

400 C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

compartment. Myocardial morphogenesis is, in fact, an intricateprocess in which cells derived frommany embryonic sources coalesceand interact to insure that the heart attains the appropriate size,shape, structure and function. In this context, stromal cells have beenshown to play an important and unique task, consisting in integratingcells into functional assemblies with well-defined three-dimensionalstructures [17]. Recent data from our group have highlighted thatpeculiar stromal cells, called telocytes, mediate myocardial compac-tion in vitro and regulate the ventricular wall organization duringmouse heart development [18]. Further support to the sculpturingrole of stromal cells in heart development comes from the in vitrodemonstration that embryonic, but not adult, fibroblasts are able topromote cardiomyocyte engraftment in a three-dimensional collagenmatrix [19], selectively express the GATA-4 cardiac transcriptionalfactor [20], and activate the dedifferentiation and cell cycle re-entry ofadult cardiomyocytes [21].

All these observations appear of particular interest taking intoconsideration that CSCs and stromal cells also establishmutual contactsin the adult heart within the cardiogenic niches [1,18]. The physicalinteraction between the two cell types are thought to control thegrowth,migration and commitment of CSCs, thereby contributing to thephysiological turnover of themyocardium [22,23]. Fromall these data, itappears that a better knowledge of the molecular mechanismscontrolling stromal/CSC interaction may be crucial for exploiting thisinformation to cardiac regeneration strategies. In theseperspectives, therecent finding of cardiomyocytes undergoingmitosis after the injectionof bone-marrow-derived mesenchymal stromal cells (MSCs) in thepost-ischemic porcine heart [24,25] suggests that these cells, acting asstromal/supporting cells, may stimulate the endogenous regenerativepotential of the myocardium [26].

In the present study we further explored the novel and attractiveconcept that MSCs may influence cardiac regeneration. Using a co-culture system, we have provided evidence that mouse MSCscollaboratively interact and regulate neonatal mouse cardiomyocyteproliferation and maturation through Notch-1/Jagged-1 signaling.

2. Materials and methods

A detailed version of Materials and methods is reported in theSupplemental file 1

2.1. Cardiomyocyte isolation and characterization

Primary cultures of neonatal mouse ventricular cardiomyocyteswere prepared and cultured as described previously [27]. Briefly, wholeheartswere excised fromhearts of 1–2-dayold newborn Swissmice andthe ventricles were minced and digested in 0.1% collagenase type II(Sigma-Aldrich, Milan, Italy) in Hank's solution for 30 min at 37 °C. Toremove fibroblasts, the cells were pre-plated twice on uncoated culturedishes for 40 min at 37 °C andmyocardial cells in the supernatant werecentrifuged and cultured on collagen-coated dishes or coverslips inDulbecco's modified Eagle's medium (DMEM) added with 10% horseserum (HS), 5% fetal calf serum (FCS) and 1% penicillin/streptomycin(Sigma). Cells were characterized for the expression of typicalcardiomyocyte markers and by the appearance of spontaneous beatingand cardiac electrophysiological properties [28].

2.2. Mesenchymal stromal cell (MSC) isolation and characterization

Mouse bone marrow-derived mesenchymal stromal cells (MSCs)were isolated from femura and tibiae of male C2F1 mice, as describedpreviously [29]. Bone marrow pellets were suspended in 10 ml ofHank's Balanced Salts Solution (HBSS, Euroclone, Milan, Italy)+1%Fetal Bovine Serum (FBS, HyClone, South Logan, UH), centrifuged(300 g, 7 min) and resuspended in low glucose DMEM supplementedwith L-glutamine, 25 mM HEPES, pyruvate, and 20% FBS. At sub-

confluence, the adherent cells were detached and sub-cultured toremove hematopoietic cells. After the 5th passage, the highly-enriched MSC cultures were used for the experiments. MSCphenotype was assessed by demonstrating their osteogenic andadipogenic potential and their specific antigenic profile [30].

2.3. Cell cultures

Soon after isolation (T0), cardiomyocytes were cultured alone or inco-culture with 5×104 MSCs at an approximately 3:1 ratio and grownfor 6, 12, 24 and 48 h. MSCs in the co-cultures were identified bylabeling with Vybrant™ Dil Cell-Labeling solutions (Molecular Probes,Eugene, OR) for 30 min at 37 °C according to the manufacturer'sprotocol before seeding. Parallel experiments were performed withMSCs silenced for Jagged-1 expression by specific si-RNA before beingco-cultured with cardiomyocytes. MSCs were cultured in the samemedium used for cardiomyocytes for the above reported times. Toreveal paracrine interaction between MSCs and cardiomyocytes, insome experiments, cardiomyocytes were grown in MSC-derivedconditioned medium for 24 h.

2.4. Time-lapse videomicroscopy

The dynamic features of the single cultures and the co-cultureswereanalyzed for 48 h by time-lapse videomicroscopy (1 frame/min,exposure time 0.5 s) using an inverted phase-contrast and fluorescencemicroscope (Nikon, Tokyo, Japan) equipped with a 10× objective and acooled video camera equipped with a motorized filter wheel and itsdedicated digital recording software (Chroma CX3, DTA, Cascina, Italy).This assaywas alsoused to evaluate the growthof the cell clusterswhichwas quantified by direct counting of cardiomyocytes in the beatingclusters at 24 h; the average number of cardiomyocytes was calculatedand reported as mean±SEM. An average of 50 clusters for eachexperimental conditionsperformed in triplicatewas assayed.Moreover,the number of cardiomyocytes and Dil-stained MSCs undergoingmitosis within the initial 24 h was calculated to evaluate the individualcontribution of each cell type to the clusters' growth. Single cultures ofcardiomyocytes were used as controls.

2.5. EdU (5-ethynyl-2′-deoxyuridine) incorporation assay

Cardiomyocyte proliferation was evaluated by the incorporation ofthe modified nucleoside, EdU (5-ethynyl-2′-deoxyuridine) using thefluorescent Click-iT® EdU Cell Proliferation Assays (Invitrogen LifeTechnologies, Grand Island, NY, USA) according to manufacturer'sinstructions. The cells were grown on glass coverslips and incubatedin the presence of 10 μM EdU; after 24 h incubation, the cells werefixed, permeabilized with 0.5% Triton X-100 and incubated with theAlexa Fluor 488 EdU detection solution. Observations were performedunder a confocal Leica TCS SP5 microscope (Leica Microsystems,Mannheim, Germany) equipped with a HeNe/Ar laser source forfluorescence measurements and with differential interference con-trast (DIC) optics, as reported previously [13].

The number of EdU+ nuclei was evaluated in 10 randommicroscopic fields (60× ocular) in each cell preparation and expressedas percentage of the total cell number. The experiments wereperformed in triplicate.

2.6. Transmission electron microscopy

Cardiomyocytes in single- and co-cultures were grown ontocollagen-coated cellulose membranes for 24 h, fixed in 4% glutaralde-hyde and 1% osmium tetroxide and embedded in Epon 812. Ultrathinsections were stained with uranyl acetate and alkaline bismuthsubnitrate and examined under a JEM 1010 electron microscope (Jeol,Tokyo, Japan) at 80 kV.

401C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

2.7. TUNEL assay

The DNA fragmentation was analyzed by TUNEL (terminaldeoxyribonucleotidyl transferase mediated dUTP nick end labeling)reaction using the In Situ Cell Death Detection Kit (Roche Diagnostics,Indianapolis, IN, USA) as reported in the manufacturer's instructions.Briefly, the cells were grown on glass coverslips, incubated withTUNEL reaction mixture, containing FITC-labeled nucleotides, coun-terstained with Hoechst (5 μg/ml; Molecular Probes) and finallyobserved under a Two-PhotonMicroscope [31]. TUNEL-positive nucleiwere counted in five microscopic fields (60×) for each cellpreparation. Three cell preparations for each experiment performedin duplicate were analyzed. TUNEL apoptotic index was thenexpressed as relative percentage of TUNEL positive nuclei on thetotal number of Hoechst-stained nuclei set at 100.

2.8. Total RNA extraction and RT-PCR

Expression levels of mRNA for Notch-1 and Jagged-1 were assayedby reverse transcription-PCR (RT-PCR) in cardiomyocytes and MSCs.Total RNA was extracted with Trizol Reagent (Invitrogen), accordingto the manufacturer's instructions. Briefly, 800 ng of total RNA wasreverse transcribed and amplified with SuperScript One-Step RT-PCRSystem (Invitrogen) for Notch-1, Jagged-1 and β-actin. PCR productswere separated by electrophoresis on 2% agarose and the ethidiumbromide-stained bands were quantified by densitometric analysisusing the Scion Image Beta 4.0.2 image analysis program (Scion Corp.,Frederick, MD).

2.9. Western blotting

Expression levels of Notch-1 and Jagged-1 proteins were assayedby Western blot in cardiomyocytes and MSCs. Forty μg of the totalproteins from cell lysates were electrophoresed by SDS–PAGE anddetected by the incubation with rabbit monoclonal anti-Notch-1antibody (1:2000; Abcam, Cambridge, UK) or goat polyclonal anti-Jagged-1 antibody (1:1000; Santa Cruz Biotechnology, Santa Cruz,CA). As control, the expression of β-actin was evaluated. Specificbands were detected using peroxidase-labeled secondary antibodies(1:15.000; Vector, Burlingame, CA) and ECL chemiluminescentsubstrate (Amersham Cologno Monzese, Italy). Densitometric analy-sis of the bands was performed using Scion Image Beta 4.0.2 imageanalysis software (Scion Corp.).

2.10. Confocal immunofluorescence microscopy

Cardiomyocytes in monoculture and in co-culture with Dil-labeledMSCs grown on glass coverslips were fixed and immunostained witheither rabbit polyclonal anti-cyclin A (1:100; Santa Cruz), or anti-Notch-1 antiserum (1:200; Abcam), or goat polyclonal anti-Jagged-1 antibody(1:100, SantaCruz) followedbygoat anti-rabbit or rabbit anti goatAlexaFluor 488-conjugated IgG (1:200; Molecular Probes). Negative controlswere performed with non immune rabbit serum substituted for theprimary antibodies. The cells were then viewed and analyzed under aconfocal Leica microscope. The number of cyclin A positive cardiomyo-cytes was calculated on digitized images of cardiomyocyte in single-and co-cultures and expressed as percentage of the total cell number.To quantify Notch-1 expression, densitometric analysis of theintensity of fluorescence signal was performed on digitized imagesusing ImageJ software (http://rsbweb.nih.gov/ij). Twenty regions ofinterest (ROI) of 400 μm2 were evaluated in each confocal stack.At least 10 images for each experimental condition performed intriplicate were assayed.

2.11. Silencing of Jagged-1 by siRNA

To inhibit the expression of Jagged-1, short interference RNAduplexes (siRNA) (Santa Cruz) were used. A non-specific scrambled(SCR) siRNA (Santa Cruz) was used as control. MSCs at 80% confluencewere transfected using siRNA trasfection medium (Santa Cruz) withthe mixed combination of Jagged-1–siRNA duplexes or with SCR-siRNA (20 nM). After 5 h, transfected cells were shifted in freshmedium for additional 24 h (T0), and then labeled with Dil and co-cultured with cardiomyocytes. Specific knock-down of Jagged-1 wasevaluated by confocal immunofluorescence (not shown) andWesternblotting. The efficiency of transfection was estimated to be approx-imately 70%.

2.12. Cytokine/growth factor assay

Interleukin (IL)-15, IL-18, basic fibroblast growth factor (bFGF),leukemia inhibiting factor (LIF), macrophage colony stimulatingfactor (M-CFS), interferon-γ-induced monokine (MIG), macrophageinflammation protein (MIP)-2 and vascular endothelial growth factor(VEGF)-1 levels were measured in 24 h culture medium of MSCs, ofcardiomyocytes grown in proliferation medium or in MSC-derivedconditioned medium and of co-cultures, by using Bio-Plex Pro MouseCytokine assay (Bio Rad Laboratories Inc., Hercules, CA, USA)following the manufacturer's instructions.

2.13. Statistical analysis

Data were reported as mean±SEM unless otherwise stated,statistical significance was determined by one-way ANOVA andNewman–Keuls multiple comparison test or Student's t test. Ap value≤0.05 was considered significant. Calculations were per-formed using GraphPad Prism software (GraphPad, San Diego, CA).

3. Results

3.1. MSCs promote neonatal cardiomyocyte proliferation

Time-lapse phase-contrast imaging showed that neonatal cardio-myocytes grew in the formof beating clusterswhichbecame largerwithtime (Figs. 1A, E). Of interest, the addition of MSCs to the culturessignificantly increased the amount of cardiomyocytes per clusterstarting from 12 h of co-culture (Figs. 1B, E). In order to analyze thespatial relationships between the two cell types and the individualgrowth rate, we next labeledMSCswith the fluorescent vital dyeDil andexamined the co-cultures by time-lapse fluorescence imaging (Figs. 1C,D). We found that Dil-labeledMSCswere located around and inside thecardiomyocyte clusters, in a mutual proportion of approximately 1MSCs/10 cardiomyocytes. Detection of dividing cells in the clustersduring the first 24 h showed that cardiomyocytes in the co-culturesgrew at a substantially higher rate compared to those in single cultures(percent increase in cell number at the 24 h-end point: 100±2% vs.30±1% respectively; pb0.05 by Student's t test; see also Supplementalfiles 2 and 3). MSCs also proliferated in the co-cultures, although at amarkedly slower rate than the adjacent cardiomyocytes. These findingssuggested that MSCs collaboratively interact with cardiomyocytes tostimulate and sustain their growth. To further confirm the aboveobservations, we then performed EdU incorporation experiments andanalyzed the expression of cyclin A by cardiomyocytes in the differentexperimental conditions (Figs. 1F–H; Fig. 2). We found that thepercentage of EdU+cardiomyocytes over the total cardiomyocyteswas higher in the co-cultures than in the single cultures after 24 h (30%±1% vs 5%±0.5% respectively; pb0.05 by Student's t test). Moreover,approximately 10% of the cardiomyocytes cultured alonewere cyclin A-positive; in these cells, a weak cytoplasmic and sparse nuclear cyclin Aimmunoreactivitywas observed at 24 h of culture, indicating that only a

Fig. 1. (A–D) Time lapse videomicroscopy ofmouse neonatal cardiomyocytes in single culture (A, C) and co-culture with unlabeledMSCs (B, asterisks) or with Dil-labeled (red)MSCs(D). (A, B) Phase contrast imaging; the presence of MSCs greatly increases the size of the cardiomyocyte clusters. (C, D) Fluorescence imaging; cardiomyocytes undergoing divisionappear as round-shaped and phase-bright cells and are more evident in the presence of MSCs (red). (E) Morphometric evaluation of the number of cardiomyocytes (CM) per cluster,performed on phase contrast images in the indicated experimental conditions: note that cardiomyocyte growth is significantly reduced when cells are co-cultured with MSCssilenced for Jagged-1 expression (Jagged-1–siRNA; compare with MSCs transfected with scrambled-siRNA, SCR-siRNA). Data are representative of at least three independentexperiments. Significance of differences (one-way ANOVA and Newman–Keuls post hoc test): *pb0.05 vs. single cultures (CM); °pb0.05 vs. co-cultures. (F–H) Superimposedfluorescence and DIC images of cardiomyocytes in (F) single and (G, H) co-culture with Dil-labeled MSCs showing EdU incorporation (green) in the proliferating cells.

402 C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

fewcardiomyocyteswere recruited to cell cycle (Figs. 2A, E). By contrast,when co-cultured with MSCs, the percentage of cardiomyocytesexpressing cyclin A increased significantly (about 25% of the cellsresulted positive): the protein immunolocalized intensely within thecytoplasm and distinctly in the nucleus of cardiomyocytes adjacent toMSCs, strongly suggesting that the physical contact with MSCs cells

could promote cardiomyocyte proliferation (Fig. 2B, E). The existence ofclose cell–cell interactions between cardiomyocytes and MSCs was alsorevealed by the ultrastructural examination (Fig. 3A, B), which showedpresumptive MSCs closely apposed to the immature cardiomyocytes,recognized by the presence of leptomeres, peculiar aggregates of cross-banded filaments preceding the formation of myofibrils. Of note, some

Fig. 2. (A–D) Representative superimposed confocal fluorescence and DIC images of cyclin A expression (green) in mouse neonatal cardiomyocytes in (A) single culture and 24 h co-culture with: (B) Dil-labeled (red) MSCs; (C) Dil-labeled MSCs trasfected with scrambled (SCR) si-RNA; (D) Dil-labeled MSCs silenced for Jagged-1 expression (Jagged-1-siRNA).Silenced MSCs were added to cardiomyocytes 24 h after siRNA transfection (T0) and co-cultured for further 24 h. (E) Quantitative analysis of the number of cyclin A-positivecardiomyocytes (CM) in the different experimental conditions. Data are representative of at least three independent experiments. Significance of differences (one-way ANOVA andNewman–Keuls post hoc test): *pb0.05, vs. single culture (CM); °pb0.05 vs. co-cultures.

403C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

cardiomyocytes adjacent to the MSCs also displayed mitotic figures(Fig. 3C). The induction of cardiomyocyte proliferation byMSCswas notassociated with significant changes of cell death (approximately 5% ofcardiomyocyte nuclei were TUNEL+ in both the single and co-cultures,Figs. 3D, E) suggesting that, under the noted co-culture conditions,MSCsdid not influence cardiomyocyte apoptosis.

3.2. MSCs promote neonatal cardiomyocyte proliferation throughjuxtacrine Notch-1-dependent mechanisms

Several lines of evidence indicate that Notch pathway is a crucialcell–cell signaling system for cardiac development, controllingmyocyte proliferation and differentiation [32,33]. This pathway ismediated by the physical interaction of Notch receptor and its ligandexpressed at the membrane of adjacent cells [34]. In search for a roleof this pathway in cardiomyocyte/MSC interaction, we then sought toexamine the expression of Notch-1 and its major ligand Jagged-1 inthe cardiomyocytes and MSCs. By combined RT-PCR and Westernblotting analyses, cardiomyocytes in single culture expressed highlevels of Notch-1 transcript and protein since the early culture times(12, 24 h); Notch-1 levels then decreased markedly at 48 h,conceivably due to cell differentiation. Moreover, the levels ofJagged-1 transcript and protein in the cardiomyocytes progressivelyincreased with time (Figs. 4A, B). On the other hand, in the MSCscultured alone, the levels of Notch-1mRNA and protein peaked at 24 hand suddenly decreased thereafter, whereas those of Jagged-1 peakedat 12 h and slowly decreased at the later time points, being virtuallyundetectable at 48 h (Figs. 4A, B). By confocal immunofluorescenceusing a specific antiserum recognizing both Notch-1 receptor and itsactivated form, Notch-1 intracellular domain (Notch-ICD), we found

that the levels of activated Notch-ICD in the cytoplasm and nucleus ofthe cardiomyocytes in co-culture with Dil-labeled MSCs weremarkedly increased compared to those in single cultures (Figs. 5A,B, E). The immunostaining was particularly evident in the cardio-myocytes located in close contact with MSCs, stressing the assump-tion that direct cellular contacts with the mesenchymal cells werecapable of potentiating Notch-mediated cardiomyocyte growth. Thisjuxtacrine paradigm was further supported by the finding that theexpression of Notch-ICD and cyclin A as well as the number ofcardiomyocytes per clusters was significantly reduced in thecardiomyocytes co-cultured with MSCs silenced for the expressionof Jagged-1 as compared with the levels of co-cultures with wild-typeor SCR-siRNA transfected MSCs (Figs. 1E; 2C–E; 4C; 5C–E).

3.3. Paracrine-mediated mechanisms

Although Jagged-1 silencing in MSCs yielded approximately 80%down-regulation (Fig. 4C), the expression of both Notch-ICD andcyclin A by cardiomyocytes in co-culture with silenced MSCs did notrevert to the levels found in the single cultures; this finding suggestedthe existence of additional paracrine mechanisms underlying thestimulation of cardiomyocyte growth by MSCs, as also indicated bythe recent literature [35]. To address this issue, we finally tested for apanel of cytokines and growth factors in the conditioned medium ofMSCs and cardiomyocytes in single and co-cultures. Of interest, MSCsreleased significant levels of several growth factors (Fig. 6), includingVEGF-1 and FGF, whose role in the regulation of cardiomyocyteproliferation and Notch-1 upregulation has been recently reported[36,37]. Cardiomyocytes also released detectable amount of VEGF-1and FGF, albeit at a substantially lower level; the release was greatly

Fig. 3. (A–C) Representative electron micrographs of mouse neonatal cardiomyocytes in single and co-cultures with MSCs grown on cellulose supports for 24 h. (A) Cardiomyocytesin single culture grow in clusters (g: electron-lucent glycogen areas). (B, C). Cardiomyocytes (cm) in co-culture are recognized by the presence of cross-banded leptomeres(arrowheads), whereas MSCs display a lighter chromatin pattern and an abundant rough endoplasmic reticulum. A cardiomyocyte undergoing mitosis is seen (asterisk).Juxtaposition between the two cell types can be clearly seen. (D, E) Fluorescence image of DNA fragmentation in cardiomyocytes cultured (D) alone and (E) in co-culture with Dil-labeled MSCs (red); TUNEL-positive nuclei are visualized in green, while the nuclei of viable cells are stained blue. Data are representative of at least three independent experiments.

404 C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

potentiated when the cells were exposed to MSC-derived conditionedmedium, indicating the ability of MSCs to stimulate growth factorproduction by cardiomyocytes through paracrine mechanisms. Nota-bly, high concentrations of VEGF-1 and FGF were also observed in theconditionedmedium of the co-cultures (Fig. 6), although these did notreach the levels induced by MSC-derived medium, suggesting thatcomplex cell–cell interactions mediated the functional cross-talkbetween the two cell types. Finally, the findings that Notch-1expression, at the mRNA and protein level, as well as cell growthand EdU incorporation (with approximately 20% of EdU+ cardio-myocytes) resulted significantly increased in the cardiomyocytesgrown in the presence of MSC-derived conditioned medium ascompared to the control cultures (Figs. 7A–F), were in perfectagreement with the assumption that paracrine mechanisms couldinfluence the proliferative response of cardiomyocytes to MSCs.

4. Discussion

Bone marrow MSCs are among the most used adult stem cells inregenerative medicine for cardiac repair. These cells, in fact, possess anumber of desirable attributes including ease of isolation and in vitroexpansion, clonogenicity, differentiation ability into mesodermal/mesenchymal lineages and immunosuppressive and anti-inflammatoryfunctions, potentially allowing allogeneic cell therapy [38]. Preclinicalstudies on stem cell therapy for cardiac regeneration have reportedsignificant improvements of ventricular pump function, ventricularremodeling, and myocardial perfusion after MSC transplantation in thepost-ischemic heart [39,40]. Several lines of evidence suggest that thetherapeutic potential of these cells in myocardial repair is related to

their ability to improve myocardial function establishing paracrineinteractions with the resident cardiac cells [14,35]. Once at the site ofinjury, in fact, the implanted cells secrete bioactive factors that promoteneo-angiogenesis, inhibit apoptosis and decrease the stiffness of thescarredventricularwall, thus actingas anactive graftwhichdynamicallycontributes to improve the myocardial performance [41,42]. Interest-ingly, the possibility has also been suggested that implanted MSCs mayexert paracrine effects on resident cardiac stem cells, activatingendogenous myocardial regeneration mechanisms [24,35,38,43].

In the present study, we provide new clues to extend the list ofMSC' effects on the host myocardium, showing that that these cellscan establish juxtacrine and paracrine interactions with neonatalimmature myocardial cells and enhance their proliferative potential.Indeed, we found that cardiomyocyte growth was significantlyincreased by co-culture with MSCs and that direct cell–cell in-teractions and release of cardiotropic soluble factors were cruciallyinvolved in the up-regulation of Notch-1 signaling by cardiomyocytes.These findings may contribute to explain the previous observationthat, in an experimental model of cell-based therapy for cardiacregeneration, new myocardial tissue is formed in the regions whereMSCs had been injected [24,25].

Notch signaling is known to play a critical role in the control ofcardiac development and in the self-renewal of myocardial pre-cursors, as well as in the maintenance of adult heart tissue integrity[32,33,43]. Notch proteins (Notch-1 to Notch-4 in mammals) aresingle-pass trans-membrane receptors with an extracellular ligand-binding domain, whose activation requires the interaction withspecific ligands, including Jagged-1 and -2, and Delta-like-1 to 4,which are membrane-anchored proteins expressed by adjacent cells

Fig. 4. (A) RT-PCR and (B)Western blotting time-course analyses of Notch-1 and Jagged-1 expressions in cardiomyocytes (CM) and MSCs. (C) Western blotting of MSCs silenced forJagged-1 expression (Jagged-1–siRNA) or transfected with scrambled siRNA (SCR-siRNA) at T0 and 24 h (24 h and 48 h after transfection respectively). The densitometric analyses ofthe band were normalized to β-actin. Data are representative of at least three independent experiments. Significance of differences (one-way ANOVA and Newman–Keuls post hoctest): *pb0.05 vs. T0 (A,B) or the corresponding SCR-siRNA controls (C).

405C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

Fig. 5. (A–D) Superimposed confocal fluorescence and DIC images of Notch-1 expression (green) in neonatal cardiomyocytes in (A) single culture and 24 h co-culture with: (B) Dil-labeled (red) MSCs; (C) Dil-labeled MSCs transfected with scrambled (SCR) si-RNA; (D) Dil-labeled MSCs silenced for Jagged-1 expression (Jagged-1-siRNA). The cells were stainedwith a specific antibody recognizing both the membrane Notch-1 receptor and its activated intracellular form (Notch-ICD). (E) Quantitative analysis of Notch-ICD expression incardiomyocytes (CM) in the different experimental conditions. Data are representative of at least three independent experiments. Significance of differences (one-way ANOVA andNewman–Keuls post hoc test): *pb0.05 vs. single culture (CM), °pb0.05 vs. co-cultures.

406 C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

and act as juxtacrine factors [34]. Once bound to its ligand, Notchundergoes proteolytic cleavage by metalloprotease and γ-secretase.This generates the release of Notch-ICD, which then translocates tothe nucleus and stimulates the transcription of target genescontrolling myocyte proliferation and maturation [44]. On thisbackground, the present findings allow us to suggest that MSCsexpressing the Jagged-1 ligand can behave as potent signal-sendingcells promoting the activation of Notch-1 in the adjacent cardiomyo-cytes and consequently increasing their intrinsic proliferative atti-tude. This is also supported by our observation that thedownregulation of Jagged-1 expression in MSCs through specificsiRNAs was able to reduce the pool of Notch-ICD-positive cardiomyo-cytes and the overall amount of proliferating cells. Moreover, we alsofound that Notch-1 could be up-regulated by the soluble factorscontained in the MSC's conditioned medium. Taken together, the

Fig. 6. Cytokine and growth factor secretion profiles byMSCs grown in cardiomyocyte prolifeconditioned medium (CM–MSC medium) and CM and MSCs in co-culture (CM–MSCs) for 2differences (one-way ANOVA and Newman-Keuls post hoc test): *pb0.05 vs. CM cultured a

reported findings suggest a model of cardiomyocyte–MSC interactionin which MSCs can exert growth-promoting effects on cardiomyo-cytes through a combination of juxtacrine and paracrine signalscapable of activating Notch-1 signaling. It cannot be ruled out thatother signaling systems, including the transfer of Notch-1 mRNAsfromMSCs to cardiomyocytes through microvesicle carriers, may alsobe operating, as previously reported in developing endothelial cells[45]. Our experimental model was not suited to confirm thepreviously reported protective effect of MSCs against cardiomyocyteapoptosis [46], possibly due to the absence of relevant pro-apoptoticstimuli in the co-cultures.

Based on the above data and considerations, it can be postulatedthat MSCs may behave as cardiac supporting stromal cells, whose rolein regulating proliferation of cardiac progenitor cells in the embryonicand adult heart has been recently underscored [18,23,32,35,44,47]. On

ration medium, cardiomyocytes grown in proliferation medium (CM) or inMSC-derived4 h. Data are representative of at least three independent experiments. Significance oflone.

Fig. 7. Notch-1 expression, detected by (A) RT-PCR and (B) confocal immunofluorescence in cardiomyocytes exposed for 24 h to MSC-derived conditioned medium (CM–MSCmedium). (C) Quantification of the fluorescence signal in the indicated experimental conditions. (D) Time lapse phase contrast imaging of cardiomyocytes exposed for 24 h to MSC-derived conditionedmedium. (E) Morphometric evaluation of the number of cardiomyocytes (CM) per cluster in the indicated experimental conditions. Data are representative of atleast three independent experiments. Significance of differences (A, Student's t test; C, E, one-way ANOVA and Newman–Keuls post hoc test): *pb0.05 vs. single culture (CM),°pb0.05 vs. co-cultures. (F) Superimposed fluorescence and DIC images of cardiomyocytes cultured in MSC-derived conditioned medium for 24 h showing EdU (green)incorporation in the proliferating cells.

407C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

the other hand, our findings do not support the concept of acardiomyogenic potential of MSCs, as we never observed MSCsexpressing myocardial-specific markers in the co-cultures (notshown). However, it seems feasible that MSCs progressively reducetheir proliferative attitude, as inferred by the downregulation ofNotch-1 and Jagged-1 expressions in the longer culture times.

The interaction between MSCs and cardiomyocytes raises theintriguing hypothesis that the endogenous repair capacity of theheart – which is inadequate per se to compensate for severe loss ofmuscle tissue upon myocardial infarction –may be potentiated by theinteraction between transplanted MSCs and resident CSCs. However,this strategy may be fraught with many technical impedimentsrelated to the hostile microenvironment of the post-infarcted scar andto the possibility that MSCs my establish cell interactions with othercell types (adult cardiomyocytes, fibroblasts, inflammatory cells).However, bioengineered three-dimensional matrices have recentlybeen proposed as appropriate devices to preserve the survival of thegrafted cells and assist their migration within the damaged myocar-dium [48–50]. Along this line of thought, experiments are in progress

in our laboratory consisting in the application of a biocompatiblescaffold seeded with MSCs over the post-ischemic heart, aimed atstrategically mending the epicardial layer and recruiting the localendogenous population of CSCs located in epicardial niches.

In conclusion, the present study provides novel insight into themechanisms underlying mesenchymal–cardiomyocyte interactions.We suggest that the attitude of MSCs to support proliferation ofmyocardial precursor cells may contribute to the development of newand more efficient cell-based myocardial repair strategies directed tostimulate the endogenous regenerative mechanisms of the heart.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.yjmcc.2011.06.004.

Sources of funding

This work was supported by research grants from the University ofFlorence and the Italian Ministry of University and Research (MIUR-PRIN 2008).

408 C. Sassoli et al. / Journal of Molecular and Cellular Cardiology 51 (2011) 399–408

Disclosure statement

None.

Acknowledgments

We gratefully acknowledge Dr. Marie Pierre Piccinni, Dept. ofInternal Medicine, University of Florence, for kind help in cytokineassay, Dr. Daniele Nosi, Dept of Anatomy, Histology and ForensicMedicine, University of Florence, for assistance in preparing theconfocal images, and Dr. Matteo Lulli, Dept. Experimental Pathology &Oncology, University of Florence, for kind help in time lapse phasecontrast videomicroscopy.

References

[1] Gherghiceanu M, Popescu LM. Cardiomyocyte precursors and telocytes in epicardialstem cell niche: electron microscope images. J Cell Mol Med 2010;14:871–7.

[2] Martínez-Estrada OM, Lettice LA, Essafi A, Guadix JA, Slight J, Velecela V, et al. Wt1is required for cardiovascular progenitor cell formation through transcriptionalcontrol of Snail and E-cadherin. Nat Genet 2010;42:89–93.

[3] Anversa P, Kajstura J, Leri A, Bolli R. Life and death of cardiac stem cells: a paradigmshift in cardiac biology. Circulation 2006;113:1451–63.

[4] Altarche-Xifró W, Curato C, Kaschina E, Grzesiak A, Slavic S, Dong J, et al. Cardiac c-kit+AT2+ cell population is increased in response to ischemic injury and supportscardiomyocyte performance. Stem Cells 2009;27:2488–97.

[5] Hsieh PC, Segers VF, Davis ME, MacGillivray C, Gannon J, Molkentin JD, et al.Evidence from a genetic fate-mapping study that stem cells refresh adultmammalian cardiomyocytes after injury. Nat Med 2007;13:970–4.

[6] Borchardt T, Braun T. Cardiovascular regeneration in non-mammalian modelsystems: what are the differences between newts and man? Thromb Haemost2007;98:311–8.

[7] van Amerongen MJ, Engel FB. Features of cardiomyocyte proliferation and itspotential for cardiac regeneration. J Cell Mol Med 2008;12:2233–44.

[8] Noort WA, Sluijter JP, Goumans MJ, Chamuleau SA, Doevendans PA. Stem cellsfrom in- or outside of the heart: isolation, characterization, and potential formyocardial tissue regeneration. Pediatr Cardiol 2009;30:699–709.

[9] Padin-Iruegas ME, Misao Y, Davis ME, Segers VF, Esposito G, Tokunou T, et al.Cardiac progenitor cells and biotinylated insulin-like growth factor-1 nanofibersimprove endogenous and exogenous myocardial regeneration after infarction.Circulation 2009;120:876–87.

[10] Ott HC, Taylor DA. From cardiac repair to cardiac regeneration—ready to translate?Expert Opin Biol Ther 2006;6:867–78.

[11] Rota M, Padin-Iruegas ME, Misao Y, De Angelis A, Maestroni S, Ferreira-Martins J,et al. Local activation or implantation of cardiac progenitor cells rescues scarredinfarcted myocardium improving cardiac function. Circ Res 2008;103:107–16.

[12] Formigli L, Zecchi-Orlandini S, Meacci E, Bani D. Skeletal myoblasts for heartregeneration and repair: state of the art and perspectives on the mechanisms forfunctional cardiac benefits. Curr Pharm Des 2010;16:915–28.

[13] Formigli L, Perna AM, Meacci E, Cinci L, Margheri M, Nistri S, et al. Paracrine effectsof transplanted myoblasts and relaxin on post-infarction heart remodelling. J CellMol Med 2007;11:1087–100.

[14] LaPar DJ, Kron IL, Yang Z. Stem cell therapy for ischemic heart disease: where arewe? Curr Opin Organ Transplant 2009;14:79–84.

[15] Chimenti I, Smith RR, Li TS, Gerstenblith G, Messina E, Giacomello A, et al. Relativeroles of direct regeneration versus paracrine effects of human cardiosphere-derived cells transplanted into infarcted mice. Circ Res 2010;106:971–80.

[16] Wollert KC, Drexler H. Cell therapy for the treatment of coronary heart disease: acritical appraisal. Nat Rev Cardiol 2010;7:204–15.

[17] Doljanski F. The sculpturing role of fibroblast-like cells inmorphogenesis. PerspectBiol Med 2004;47:339–56.

[18] Bani D, Formigli L, Gherghiceanu M, Faussone-Pellegrini MS. Telocytes assupporting cells for myocardial tissue organization in developing and adultheart. J Cell Mol Med 2010;14:2531–8.

[19] Pfannkuche K, Neuss S, Pillekamp F, Frenzel LP, AttiaW, Hannes T, et al. Fibroblastsfacilitate the engraftment of embryonic stem cell-derived cardiomyocytes onthree-dimensional collagenmatrices and aggregation in hanging drops. Stem CellsDev 2010;19:1589–99.

[20] Ieda M, Tsuchihashi T, Ivey KN, Ross RS, Hong TT, Shaw RM, et al. Cardiacfibroblasts regulate myocardial proliferation through beta1 integrin signaling. DevCell 2009;16:233–44.

[21] Zaglia T, Dedja A, Candiotto C, Cozzi E, Schiaffino S, Ausoni S. Cardiac interstitialcells express GATA4 and control dedifferentiation and cell cycle re-entry of adultcardiomyocytes. J Mol Cell Cardiol 2009;46:653–62.

[22] Leri A, Kajstura J, Anversa P. Cardiac stem cells and mechanisms of myocardialregeneration. Physiol Rev 2005;85:1373–416.

All in-text references underlined in blue are linked to publications on Re

[23] Ausoni S, Sartore S. The cardiovascular unit as a dynamic player in disease andregeneration. Trends Mol Med 2009;15:543–52.

[24] Mazhari R, Hare JM. Mechanisms of action of mesenchymal stem cells in cardiacrepair: potential influences on the cardiac stem cell niche. Nat Clin PractCardiovasc Med 2007;4:S21–6.

[25] Hatzistergos KE, Quevedo H, Oskouei BN, Hu Q, Feigenbaum GS, Margitich IS, et al.Bone marrow mesenchymal stem cells stimulate cardiac stem cell proliferationand differentiation. Circ Res 2010;107:913–22.

[26] Rose RA, Jiang H, Wang X, Helke S, Tsoporis JN, Gong N, et al. Bone marrow-derived mesenchymal stromal cells express cardiac-specific markers, retain thestromal phenotype, and do not become functional cardiomyocytes in vitro. StemCells 2008;26:2884–92.

[27] Formigli L, Francini F, Nistri S, Margheri M, Luciani G, Naro F, et al. Skeletalmyoblasts overexpressing relaxin improve differentiation and communication ofprimary murine cardiomyocyte cell cultures. J Mol Cell Cardiol 2009;47:335–45.

[28] Nistri S, Pini A, Sassoli C, Squecco R, Francini F, Formigli L, et al. Relaxin promotesgrowth andmaturation ofmouse neonatal cardiomyocytes in vitro: clues for cardiacregeneration. J Cell Mol Med 2011, doi:10.1111/j.1582-4934.2011.01328.x.

[29] Dobson KR, Reading L, Haberey M, Marine X, Scutt A. Centrifugal isolation of bonemarrow from bone: an improved method for the recovery and quantitation ofbone marrow osteoprogenitor cells from rat tibiae and femurae. Calcif Tissue Int1999;65:411–3.

[30] Peister A, Jason AM, Larson BL, Hall BM, Gibson LF, Prockop DJ. Adult stem cellsfrom bone marrow (MSCs) isolated from different strains of inbred mice vary insurface epitopes, rates of proliferation, and differentiation potential. Blood2004;103:1662–8.

[31] Mercatelli R, Soria S, Molesini G, Bianco F, Righini G, Quercioli F. Supercontinuumsource tuned by an on-axis monochromator for fluorescence lifetime imaging. OptExpress 2010;18:20505–11.

[32] Collesi C, Zentilin L, Sinagra G, Giacca M. Notch1 signaling stimulates proliferationof immature cardiomyocytes. J Cell Biol 2008;183:117–28.

[33] Nemir M, Pedrazzini T. Functional role of Notch signaling in the developing andpostnatal heart. J Mol Cell Cardiol 2008;45:495–504.

[34] Brou C. Intracellular trafficking of Notch receptors and ligands. Exp Cell Res2009;315:1549–55.

[35] Psaltis PJ, Paton S, See F, Arthur A, Martin S, Itescu S, et al. Enrichment for STRO-1expression enhances the cardiovascular paracrine activity of human bonemarrow-derived mesenchymal cell populations. J Cell Physiol 2010;223:530–40.

[36] Cohen ED, Wang Z, Lepore JJ, Lu MM, Taketo MM, Epstein DJ, et al. Wnt/beta-catenin signaling promotes expansion of Isl-1-positive cardiac progenitor cellsthrough regulation of FGF signaling. J Clin Invest 2007;117:1794–804.

[37] Kiec-Wilk B, Grzybowska-Galuszka J, Polus A, Pryjma J, Knapp A, Kristiansen K. TheMAPK-dependent regulation of the Jagged/Notch gene expression by VEGF, bFGFor PPAR gamma mediated angiogenesis in HUVEC. J Physiol Pharmacol 2010;61:217–25.

[38] English K, French A,Wood KJ. Mesenchymal stromal cells: facilitators of successfultransplantation? Cell Stem Cell 2010;7:431–42.

[39] Psaltis PJ, Zannettino AC, Worthley SG, Gronthos S. Concise review: mesenchy-mal stromal cells: potential for cardiovascular repair. Stem Cells 2008;26:2201–10.

[40] Nagaya N, Kangawa K, Itoh T, Iwase T, Murakami S, Miyahara Y, et al.Transplantation of mesenchymal stem cells improves cardiac function in a ratmodel of dilated cardiomyopathy. Circulation 2005;112:1128–35.

[41] Caplan AI, Dennis JE. Mesenchymal stem cells as trophic mediators. J Cell Biochem2006;98:1076–84.

[42] Gnecchi M, Zhang Z, Ni A, Dzau VJ. Paracrine mechanisms in adult stem cellsignaling and therapy. Circ Res 2008;10312:1204–19.

[43] Campa VM, Gutiérrez-Lanza R, Cerignoli F, Díaz-Trelles R, Nelson B, Tsuji T, et al.Notch activates cell cycle reentry and progression in quiescent cardiomyocytes. JCell Biol 2008;183:129–41.

[44] Boni A, Urbanek K, Nascimbene A, Hosoda T, Zheng H, Delucchi F, et al. Notch1regulates the fate of cardiac progenitor cells. Proc Natl Acad Sci USA 2008;105:15529–34.

[45] Deregibus MC, Cantaluppi V, Calogero R, Lo Iacono M, Tetta C, Biancone L, et al.Endothelial progenitor cell derived microvesicles activate an angiogenic programin endothelial cells by a horizontal transfer of mRNA. Blood 2007;110:2440–8.

[46] Yu XY, Geng YJ, Li XH, Lin QX, Shan ZX, Lin SG, et al. The effects of mesenchymalstem cells on c-kit up-regulation and cell-cycle re-entry of neonatal cardiomyo-cytes are mediated by activation of insulin-like growth factor 1 receptor. Mol CellBiochem 2009;332:25–32.

[47] Noseda M, Schneider MD. Fibroblasts inform the heart: control of cardiomyocytecycling and size by age-dependent paracrine signals. Dev Cell 2009;16:161–2.

[48] Chen CH, Wei HJ, Lin WW, Chiu I, Hwang SM, Wang CC, et al. Porous tissue graftssandwiched with multilayered mesenchymal stromal cell sheets induce tissueregeneration for cardiac repair. Cardiovasc Res 2008;80:88–95.

[49] Khait L, Hecker L, Blan NR, Coyan G,Migneco F, Huang YC, et al. Getting to the heartof tissue engineering. J Cardiovasc Transl Res 2008;1:71–84.

[50] Di Meglio F, Castaldo C, Nurzynska D, Romano V, Miraglia R, Bancone C, et al.Epithelial–mesenchymal transition of epicardial mesothelium is a source ofcardiac CD117-positive stem cells in adult human heart. J Mol Cell Cardiol2010;49:719–27.

searchGate, letting you access and read them immediately.