Extracellular matrix protein laminin enhances mesenchymal stem cell (MSC) paracrine function through αvβ3/CD61 integrin to reduce cardiomyocyte apoptosis

Abstract Myocardial ischaemia (MI) results in extensive cardiomyocyte death and reactive oxygen species (ROS)‐induced damage in an organ with little or no regenerative capacity. Although the use of adult bone marrow mesenchymal stem cells (BMMSCs) has been proposed as a treatment option, the high cell numbers required for clinical use are difficult to achieve with this source of MSCs, and animal studies have produced inconsistent data. We recently demonstrated in small and large animal models of acute MI that the application of human term placenta‐derived multipotent cells (PDMCs), a foetal‐stage MSC, resulted in reversal of cardiac injury with therapeutic efficacy. However, the mechanisms involved are unclear, making it difficult to strategize for therapeutic improvements. We found that PDMCs significantly reduced cardiomyocyte apoptosis and ROS production through the paracrine factors GRO‐α, HGF and IL‐8. Moreover, culturing PDMCs on plates coated with laminin, an extracellular matrix (ECM) protein, resulted in significantly enhanced secretion of all three paracrine factors, which further reduced cardiomyocyte apoptosis. The enhancement of PDMC paracrine function by laminin was mediated through αvβ3 integrin, with involvement of the signalling pathways of JNK, for GRO‐α and IL‐8 secretion, and PI3K/AKT, for HGF secretion. Our results demonstrated the utility of PDMC therapy to reduce cardiomyocyte apoptosis through modulation of ECM proteins in in vitro culture systems as a strategy to enhance the therapeutic functions of stem cells.


Introduction
Ischaemic heart disease (IHD) is a leading cause of death worldwide [1]. Atherosclerosis, the thickening and hardening of the coronary arteries resulting from the deposition of lipids in vessel walls, is the most common cause of MI, a type of IHD. When the blood supply is not quickly restored following coronary artery occlusion, the surrounding cardiac tissue undergoes necrosis and apoptosis [2,3]. Current treatment options consist of surgical repair to increase the blood supply, but once the myocardium is damaged, it is replaced by fibrosis and cannot be regenerated. This can result in permanent cardiac dysfunction and lead to death. Stem cell therapy performed with bone marrow (BM) mononuclear cells and stem cells -including mesenchymal stem cells (MSCs) -has been proposed for the treatment of IHD, but clinical results have not been as robust as those seen in animal studies [4,5]. This may be due to a number of factors, including cell volume used, method of cell delivery and post-delivery cell survival [6]. Moreover, these adult progenitors are quite rare, and cell numbers further decrease with increasing age [7]. Because IHD incidence increases with age, the use of any type of autologous BM stem cell would lose efficacy as the patient ages.
We have recently shown that human term PDMCs -a type of MSC with a foetal origin [8] -can modulate cardiac injury after MI induction in both small and large animal models [9]. Interestingly, the transplantation of PDMCs led to more significant improvements in cardiac function in the large animal model (pig) than in the small animal model (mouse). In addition to enhancing angiogenesis, as was seen in the mouse model, we found that PDMCs also reduced cardiomyocyte apoptosis in the pig model. Because cardiomyocyte death is the key pathologic reason for compromised cardiac function in IHD, we are interested in exploring the mechanisms involved in the anti-apoptotic effects of PDMCs. In this study, we established an in vitro system to mimic key events in MI-induced cardiomyocyte apoptosis using tumour necrosis factor-a (TNF-a), a key molecule involved in MI-induced cardiac injury. TNF-a is a multifunctional cytokine that promotes cell survival or death depending on its dose and context and is found at levels eightfold to 10-fold above normal in the infarct and border zones of the post-MI heart [10]. High levels of TNF-a result in not only cardiomyocyte apoptosis but also production of ROS, which lead to further cell injury and death [11,12]. To enhance the effects of PDMCs, we have modulated the in vitro cell culture conditions with the use of ECM proteins, which are known to be an important part of the stem cell microenvironment/niche and to regulate stem cell differentiation, proliferation, migration and function [13]. The in vitro use of specific ECM proteins has been shown to modulate stem cell differentiation capacity and lineage commitment [14][15][16]. However, it has not been reported whether ECM proteins modulate stem cell paracrine factor production in in vitro cell culture.
Our data in this study demonstrated that the paracrine function of PDMCs could be enhanced by the ECM protein laminin through avb3 integrin/CD61, an integrin that is highly expressed on MSCs [16,17] and to which laminin has high affinity [18]. Our findings thus can provide strategies to improve the therapeutic use of MSCs for the treatment of IHD.

Cell culture and related experiments
Term human placenta tissue samples (38-to 40-week gestation) from healthy donor mothers were obtained with informed consent and approval according to the procedures of the institutional review board. PDMCs were isolated and expanded as previously described [8]. Briefly, placental tissue was mechanically and enzymatically dissected and cultured in complete medium consisting of Dulbecco's modified Eagle's medium (DMEM; Gibco-Invitrogen, Carlsbad, CA, USA) supplemented with 10% foetal bovine serum (FBS; selected lots, Hyclone, Thermo Scientific, Waltham, MA, USA). When cells were at 80-90% confluence, they were subcultured at a dilution of 1:3. WS1, human foetal skin fibroblasts, were purchased from Bioresource Collection and Research Center (BCRC, Hsinchu, Taiwan) and were cultured as recommended in minimum essential medium (MEM; Gibco-Invitrogen) with 10% FBS (Gibco-Invitrogen). Mouse cardiomyocytes from 8-week-old C57BL/6 mice were isolated as previously described [19], cultured in DMEM (Gibco-Invitrogen) with 10% FBS (Gibco-Invitrogen), and characterized for expression of the cardiomyocytic markers cardiac troponin-T, Nkx2.5 and connexin43 [20]. For experiments with recombinant proteins, recombinant human proteins (all obtained from Peprotech, Rocky Hill, NJ, USA) were added to cell cultures at the indicated doses. For experiments involving coating with various ECM proteins, six-well plates were either not coated [indicated by the addition of phosphate-buffered saline (PBS)] or coated with ECM gel (50 lg/ml; Sigma-Aldrich, St Louis, MO, USA), gelatin (0.4%, Sigma-Aldrich), laminin (50 lg/ml; Invitrogen, Waltham, MA, USA), collagen type I (50 lg/ml; BD Biosciences, Sparks, MD, USA) or fibronectin (50 lg/ml; Sigma-Aldrich). Per the manufacturer's instructions, ECM gel was isolated from Engelbreth-Holm-Swarm (EHS) sarcoma generated in mice, and it contained laminin, collagen type IV, heparan sulphate proteoglycan, entactin and other minor components. All culture plate/well coating was performed by applying a solution containing the indicated protein/gel to culture plates/wells and incubating at 37°C for 2 hrs, then aspirating the solution from the plates/wells, and allowing them to air dry.

Enzyme-linked Immunosorbent Assay (ELISA)
CM from cells cultured in various conditions (i.e. on different ECM protein-coated plates, after RNAi knockdown) was collected, and the levels of human GRO-a, HGF, IL-6 and IL-8 were measured by enzyme-linked immunosorbent assay (ELISA). ELISA kits for GRO-a, HGF and IL-8 were obtained from R&D systems (Minneapols, MN, USA), and that for IL-6 was obtained from eBioscience. The experimental procedures were executed according to the manufacturer's instructions.

Western blotting
Western blotting was performed as previously reported [16]. Briefly, mouse cardiomyocytes were lysed in protein extraction reagent (Thermo Scientific). Total proteins were separated by electrophoresis on a 15% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The membranes were blotted with caspase-3 antibody (Cell Signaling Technology). After washing with PBS Tween-20, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 1 hr at room temperature. Signals were detected with chemiluminescent HRP substrate (Millipore, Billerica, MA, USA).

RNA interference
Small interfering RNA (siRNA) specific for avb3 integrin or the nontarget control (Gibco-Invitrogen) was transfected into PDMCs according to the manufacturer's recommendations and as previously reported [16]. Cells were seeded on six-well plates (1.5 9 10 5 cells/ well) and then treated with transfection reagent and either siRNA specific for avb3 integrin or non-target RNA. After 48 hrs, the expression levels of avb3 integrin in transfected cells were confirmed by RT-PCR.

Statistical analysis
All experiments were performed at least in triplicate, and the data are presented as the mean AE S.E.M. Both parametric analyses, Student's t-test for comparisons of two variables and ANOVA for comparisons of more than two variables, and nonparametric analyses, the Mann-Whitney U-test for comparisons of two variables and Kruskal-Wallis test for comparisons of more than two variables, were performed. A value of P < 0.05 was defined as statistically significant.

PDMC-CM reduces the number of apoptotic cardiomyocytes
Our previous work showed that PDMCs could ameliorate MI-induced cardiomyocyte apoptosis, likely through secreted factors. To ascertain whether these anti-apoptotic effects of PDMCs are mediated by secreted factors, we added PDMC-CM to TNF-a-treated cardiomyocytes (1:1 mixed with cultured medium). After 48 hrs, TNF-a induced apoptosis in approximately 18.1% of treated cells, as demonstrated by positive staining for annexin V, a marker for apoptosis (Fig. 1A). The addition of PDMC-CM to TNF-a-treated cardiomyocytes reduced the percentage of apoptotic cells down to 6.5%, whereas the addition of CM from mouse cardiomyocytes (mCardio-CM) only reduced the percentage to 13.5%. The quantitative results showed PDMC-CM but not mCardio-CM significantly reduced the percentage of apoptotic cells from an average of 16.5% down to 8.7% (Fig. 1B). Our results indicate that PDMC paracrine factors suppress TNF-a-induced cardiomyocyte apoptosis.
High levels of GRO-a, HGF and IL-8 but not IL-6 secreted by PDMCs are involved in reducing TNF-a-induced apoptosis in cardiomyocytes We have previously shown using qualitative cytokine arrays that PDMCs secrete high levels of GRO-a, HGF and IL-8, which are involved in modulating ischaemic cardiac injury [9]. To quantify the secreted levels of these three factors by PDMCs, we performed ELISA and found that PDMCs secrete higher levels of all three factors than WS1 human fibroblasts (Fig. 1C). To test whether these three PDMCsecreted factors are directly involved in anti-apoptotic effects, we induced cardiomyocyte apoptosis with TNF-a and treated them with GRO-a, HGF and IL-8 recombinant proteins alone or in combination, and we found that all three of these proteins significantly reduced the percentage of apoptotic cells, with IL-8 showing the strongest effect from 17.2% to 9.1% ( Fig. 2A and B). Furthermore, treatment with the combination of these factors appeared to have more potent anti-apoptotic effects than treatment with any factor alone, with HGF/IL-8 reducing apoptotic levels more effectively than HGF/GRO-a or GRO-a/ IL-8 (Fig. 2C). These effects are specific to these secreted factors because the addition of recombinant TIMP-1 and TIMP-2, two factors that are highly secreted by both PDMCs and BMMSCs [9], did not reduce TNF-a-induced apoptosis (data not shown). To ascertain whether GRO-a, HGF and IL-8 reduced the number of apoptotic cells through caspase-3, a critical mediator of apoptosis, we detected the activation of caspase-3 in TNF-a-induced apoptotic cardiomyocytes treated with GRO-a, HGF or IL-8 recombinant proteins. Interestingly, GRO-a, HGF and IL-8 reduced the formation of the active form of caspase-3 induced by TNF-a in cardiomyocytes (Fig. 2D). To assess the specificity of these effects, we administered IL-6 (10, 30 and 60 ng/ ml) -which is known to be highly secreted by MSCs and other stromal cells -to TNF-a-induced apoptotic cardiomyocytes and found no suppression of caspase-3 activation (Fig. 2E). Thus, our results indicate that GRO-a, HGF and IL-8 suppress cardiomyocyte apoptosis, with IL-8 having stronger effects than GRO-a and HGF.

GRO-a, HGF and IL-8 reduce the level of TNFa-induced ROS
ROS are known to be involved in cardiomyocyte apoptosis, so we assessed whether GRO-a, HGF and IL-8 affect ROS production in TNF-a-stimulated cardiomyocytes. Cardiomyocytes were treated with these three proteins alone or in combination, and then, intracellular ROS levels were measured performed with H 2 DCFDA staining. We found that TNF-a induced ROS production in cardiomyocytes, which could be reduced by any of the three factors, with the strongest effects seen when either GRO-a or HGF were added or when the combination of HGF/IL-8 was used (Fig. 3A and B). To further elucidate the suppressive effects of these paracrine factors on ROS production, we assayed for the expression of the antioxidant-associated genes Catalase and CuZnSOD and found that the addition of GRO-a, HGF and IL-8 increased the expression of Catalase and CuZnSOD in TNFa-stimulated cardiomyocytes (Fig. 3C). Our results demonstrate that PDMC-secreted factors, particularly HGF, enhance the expression of antioxidant enzymes and modulate TNF-a-induced ROS production in cardiomyocytes.

Laminin enhances PDMC secretion of GRO-a, HGF and IL-8
To better mimic the in vivo setting and assess whether paracrine functions can be modulated, we cultured PDMCs on plates that were coated with ECM proteins, including mixtures such as ECM gel and gelatin, or specific proteins such as laminin, collagen type I and fibronectin. We found that the ECM protein laminin promoted the secretion of all three factors. Culturing on laminin on average enhanced PDMC secretion of GRO-a from 25.0 to 29.2 ng/ml, HGF from 26.8 to 35.0 ng/ml and IL-8 from 27.8 to 39.4 ng/ml (Fig. 4). Culturing on collagen type I and fibronectin also increased the production of GROa and HGF, but not IL-8, significantly by PDMCs. Thus, culturing on laminin-coated plates enhanced the paracrine functions of PDMCs the most.

Laminin enhances the anti-apoptotic effects of PDMC-CM on cardiomyocytes
To verify whether the laminin-enhanced paracrine functions of PDMCs were functionally relevant, we collected the CM from PDMCs cultured as usual (uncoated; PDMC-CM) or cultured on various ECM protein-coated plates and treated TNF-a-stimulated cardiomyocytes with the CM. We found that CM collected from PDMCs cultured on laminin-coated plates (PDMC-lam-CM) was more effective at reducing TNF-a-induced apoptosis in cardiomyocytes than CM from PDMCs cultured on plates coated with the other ECM proteins (Fig. 5A and B). Moreover, PDMClam-CM also inhibited the formation of the activated form of caspase-3 (Fig. 5C). These data indicated that introducing laminin into the in vitro culture environment modulates the paracrine functions of PDMCs in a manner that is functionally relevant.

PDMC secretion of GRO-a, HGF and IL-8 but not IL-6 can be modulated by laminin through avb3 integrin
To investigate whether laminin can modulate PDMC paracrine functions through avb3 integrin/CD61, an integrin highly expressed on MSCs and known to have high affinity for laminin, we used siRNA to reduce avb3 integrin expression in PDMCs (Fig. 6A) and then assayed for various secreted factors. When the expression of avb3 integrin was silenced, it significantly reduced the protein secretion of GRO-a from 26.5 to 10.5 ng/ml, HGF from 27.4 to 17.9 ng/ml and IL-8 from 28.8 to 11.9 ng/ml even after culturing PDMCs on laminincoated plates, but IL-6 levels were relatively unchanged (Fig. 6B). Moreover, CM collected from PDMCs cultured on laminin with knockdown of avb3 integrin (siRNA-2-lam-CM) was not able to rescue cardiomyocytes from TNF-a-induced apoptosis, in contrast with CM collected from PDMCs with non-target (negative control) gene knockdown (siC-lam-CM; Fig. 6C).
Laminin promotes PDMC secretion of multiple factors through JNK, for GRO-a and IL-8 secretion, and PI3K/AKT, for HGF secretion Our finding of a single integrin being involved in the up-regulation of the secretion of multiple distinct paracrine factors was unexpected. To elucidate how this is mediated, we determined the signalling pathways that are involved in this process. We therefore treated laminincultured PDMCs with inhibitors of major signalling pathways (Fig. 7A) and found that JNK inhibition (with SP600125) led to the most significant decreases in GRO-a and IL-8 secretion, from 20.9 to 10.8 ng/ml and 18.0 to 12.4 ng/ml, respectively. Inhibition of the PI3K/AKT pathways (with MK2206 and LY294002, respectively), the MAP kinase pathways of MEK1/2 or MEK-1 (with U0126 and PD98059, respectively), or STAT3 (with Cpd188) had little or no effect. On the other hand, PDMC secretion of HGF was strongly affected by inhibition of either the PI3K (with MK2206) or AKT pathway (with LY294002), from 25.0 to 5.8 and 4.9 ng/ml, respectively, but was minimally affected by inhibition of other pathways. Collectively, these results demonstrated that laminin promoted the PDMC secretion of multiple factors mainly through JNK, for GRO-a and IL-8 secretion, and PI3K/ AKT, for HGF secretion. Thus, our findings demonstrate that the laminin/avb3 integrin axis promotes the PDMC secretion of GRO-a, IL-8 and HGF through multiple but specific signalling pathways (Fig. 7B).

Discussion
Our findings demonstrate that factors secreted by PDMCs suppress apoptosis and reduce ROS production in TNF-a-induced cardiomyocyte death. Moreover, these beneficial paracrine effects can be augmented in an in vitro setting by culturing PDMCs on laminin, which engages avb3 integrin expressed on these foetal-stage MSCs. Coupled with our recent results on the angiogenic and possible transdifferentiation of PDMCs in murine and porcine models of acute MI [9], our current data indicate that these foetal-stage MSCs may be good candidates for the treatment of IHD.
In this study, we found that PDMCs secrete high levels of GRO-a, HGF and IL-8, which have roles in reducing apoptosis and ROS production; the levels of these paracrine factors can be further up-regulated through modulation of the laminin/avb3 integrin axis. GRO-a and IL-8 are members of the CXC chemokine family, and they share a common receptor: CXCR2 [22]. Our data show that IL-8 is more potent at suppressing apoptosis than GRO-a; this may be because IL-8 is a ligand for two receptors, CXCR1 and CXCR2, whereas GROa is recognized only by CXCR2. Both CXCR1 and CXCR2 have been reported to regulate cell death by reducing apoptosis and necrosis [23]. CXCR2 is widely expressed on normal and tumour cells, and although we did not investigate the involved downstream signalling pathways, there are a number of reports in the literature demonstrating that the engagement of this receptor promotes cell survival and regulates anti-apoptotic genes [22,24], including suppression of TNF-a-induced apoptosis and ROS production [25][26][27]. Interestingly, our data also demonstrate that the laminin/avb3 integrin axismediated up-regulation of these two chemokines is mediated mainly through JNK, whereas the up-regulation of HGF is mainly mediated through PI3K/AKT (Fig. 7A). HGF, which is highly secreted by many types of stromal cells including MSCs, is known to be involved in multiple cellular processes, including proliferation and anti-inflammation [28][29][30]. Engagement of its only receptor, c-MET, leads to the activation of four major signalling pathways, and the PI3K/AKT pathway is commonly involved in the ligand's proliferative and anti-apoptotic functions in both normal and neoplastic cells [31]. In contrast, IL-6 -a cytokine also highly secreted by PDMCs and MSCs [32][33][34] did not reduce TNF-a-induced apoptosis and was not further up-regulated through the laminin/avb3 integrin axis. These results are in line with previous findings that the IL-6 inhibition of caspase-3 is mediated through TGF-b but not TNF-a [35]. IL-6, a pleomorphic cytokine, is highly linked with inflammation and disease, especially in the cardiac system, and involves ERK signalling [36,37]. The lack of an IL-6 response in PDMC-mediated cell therapy clearly can be beneficial in the setting of IHD and supports the specificity of laminin/avb3 integrin modulation and lack of ERK activation of PDMC paracrine functions. However, such speculation would require further research for substantiation.
In the endogenous environment of the stem cell niche, a number of ECM proteins have been found to regulate crucial stem cell functions, including proliferation, migration and lineage commitment [13,15]. One of the most active areas of research in stem cell therapy is how to mimic the in vivo niche for more efficient induction of lineage commitment ex vivo for the ultimate goal of improving therapeutic efficacy. Our data demonstrate that through the in vitro use of laminin, the paracrine effects of PDMCs can be enhanced with functional relevance. The use of ECM gel, which is more complex -containing laminin, collagens, heparan sulphate proteoglycan and other minor components -and presumed to be more reflective of in vivo environments, surprisingly was not as effective as using laminin only, a single type of ECM protein. It may be that in order to enhance the secretion of GRO-a, HGF and IL-8 by PDMCs over constitutive levels, a very potent agent is needed; thus, purified, single-agent proteins rather than a mixed compound -in which one single component would be at a lower concentration -may be more effective. Our results showing that knockdown of a specific integrin can reverse the effects of laminin may offer some indirect support. Moreover, these results also demonstrate that in MSCs/PDMCs, laminin, which binds to multiple integrin receptors [38,39], does bind with high affinity to the avb3 integrin [16,18]. Our findings on the multiple signalling pathways involved downstream in the laminin induction of PDMC paracrine function are in line with emerging reports on ECM/integrin interactions in cell 'sensing' of the microenvironment to modulate intracellular signalling pathways and coordinate cellular responses [40]. Overall, our findings on the potent and specific effects of the laminin/avb3 integrin axis further support the importance of the microenvironment in modulating stem cell functions and highlight the capacity for the manipulation of in vitro culture conditions to affect stem cell function.