Rap1-mediated nuclear factor-kappaB (NF-κB) activity regulates the paracrine capacity of mesenchymal stem cells in heart repair following infarction

Paracrine effect is the major mechanism that underlies mesenchymal stem cells (MSC)-based therapy. This study aimed to examine how Rap1, telomeric repeat-binding factor 2-interacting protein 1 (Terf2IP), which is a novel modulator involved in the nuclear factor-kappaB (NF-κB) pathway, regulates the paracrine effects of MSC-mediated heart repair following infarction. NF-κB activity of stromal cells was increased by Rap1 as measured by pNF-κB-luciferase reporter activity, and this was abolished by IkB-dominant-negative protein. Knockdown of Rap1 with shRap1 resulted in diminished translocation of p65-NF-κB from the cytoplasm to nuclei in response to tumor necrosis factor-α (TNF-α) stimulation. Compared with BM-MSCs, Rap1−/−-BM-MSCs displayed a significantly reduced ratio of phosphorylated NF-κB to NF-κB-p65 and of Bax to Bcl-2, and increased resistance to hypoxia-induced apoptosis by the terminal deoxynucleotidal transferase-mediated dUTP nick end labeling (TUNEL) assay. In contrast, re-expression of Rap1 in Rap1−/−-BM-MSCs resulted in loss of resistance to apoptosis in the presence of hypoxia. Moreover, absence of Rap1 in BM-MSCs led to downregulation of NF-κB activity accompanied by reduced pro-inflammatory paracrine cytokines TNF-α, IL (interleukin)-6 and monocyte chemotactic protein-1 in Rap1−/−-BM-MSCs compared with BM-MSCs. The apoptosis of neonatal cardiomyocytes (NCMCs) induced by hypoxia was significantly reduced when cocultured with Rap1−/−-BM-MSC hypoxic-conditioned medium (CdM). The increased cardioprotective effects of Rap1−/−-BM-MSCs were reduced when Rap1−/−-BM-MSCs were reconstituted with Rap1 re-expression. Furthermore, in vivo study showed that transplantation of Rap1−/−-BM-MSCs significantly improved heart function, decreased infarct size, prevented cardiomyocyte apoptosis and inhibited inflammation compared with controls and BM-MSCs (P<0.01). This study reveals that Rap1 has a critical role in the regulation of MSC paracrine actions. Compared with BM-MSCs, Rap1−/−-BM-MSCs decreased NF-κB sensitivity to stress-induced pro-inflammatory cytokine production and reduced apoptosis. Selective inhibition of Rap1 in BM-MSCs may be a novel strategy to enhance MSC-based therapeutic efficacy in myocardial infarction.


INTRODUCTION
Recent developments in stem cell biology to prevent or treat heart failure have moved from experimental research to clinical trials using different types of adult stem cells, such as adult bone marrow stem cells or mesenchymal stem cells (MSCs). 1,2 Despite the promising results observed in preclinical and clinical studies, the potential mechanisms underlying stem cell-based therapy have not been fully classified. Accumulating evidence demonstrates that apart from transdifferentiation, it is the paracrine effects of MSCs that are predominately responsible for cardiac repair. 3 Indeed, the exact extent to which these cells transdifferentiate into new cardiomyocytes to improve heart function remains highly controversial. 4,5 There is increasing recognition that MSCs produce a variety of cytokines that can directly promote cell survival and regulate inflammation following myocardial infarction (MI). 3,6 Various cytokines and growth factors have been investigated in different tissue-derived MSCs. 7,8 Some cytokines and growth factors have been shown to be critical for cardiac protection, for example, basic fibroblast growth factor, vascular endothelial growth factor, stromal-derived factor, 9,10 secreted frizzled related protein, 11 interleukin-10 (IL-10) 12 and metalloproteinase-9. 13 In contrast, some cytokines and factors produced from MSCs, including tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6), 14 are harmful for heart recovery. As a result, injection of MSCs or total MSC paracrine factors without optimization may limit their therapeutic efficacy. Therefore, optimization of MSCs before transplantation to maximize cell survival and beneficial paracrine factors is important. Nevertheless, the potential mechanism underlying regulation of MSC secretion is poorly understood and identification of the key molecules that govern MSC secretion to protect against heart ischemia injury remains urgently to be addressed. Rap1 (Trf2IP), a telomeric repeat-binding factor 2-interacting protein 1 (Terf2IP), is a novel modulator involved in the Nuclear factor-kappaB (NF-κB) pathway. 15,16 Many telomeric proteins apart from Rap1 including telomerase are now known to regulate inflammation through NF-κB signaling. 17,18 It has been documented that NF-κB family factors have important roles in the regulation of mitochondrial ROS/bioenergy, DNA replication, cell survival and inflammation in many cell types, [19][20][21][22] including MSCs. 23 Prior studies have also demonstrated that NF-κB activation mediates cytokine/growth factor secretion by MSCs. 24 Given the multifaceted effects of NF-κB activity on MSCs, efforts to identify important regulators (s) that modulate specificity in the functioning of the NF-κB signal pathway is a major challenge. Use of a genetic engineering will provide a greater understanding of the roles and mechanisms of specific factors that modulate NF-κB activation in MSC-mediated cardioprotection. It also has been reported that MSCs can secrete a variety of cytokines to regulate inflammation and enable cardiac repair posttransplantation. 25 Hence, to determine whether Rap1 can regulate MSC paracrine factors, which thereby impinges on cardiac repair, we designed this study. In this study, we proposed that inhibition of Rap1 can decrease pro-inflammatory factors secreted by MSCs, enhance cell survival and thus improve its therapeutic effects in MI. In this study, we showed that Rap1 − / − -BM-MSCs have a better therapeutic efficacy than wild-type (WT) BM-MSCs for cardiac repair post MI in mice. The greater therapeutic potential of Rap1 − / − -BM-MSCs for cardiac repair is not only attributed to the higher cell survival post-transplantation, but also to a reduced secretion of pro-inflammatory cytokines. Our results may prompt the development of new therapeutic strategies to enhance MSCbased therapy in MI.
Activation of NF-κB transcriptional activity by Rap1 In order to address the direct relationship between NF-κB and Rap1, the activation of NF-κB transcriptional activity by Rap1 was examined. A reporter gene construct carrying NF-κB-Luc was used to report NF-κB activity. This construct contains an NF-κB enhancer element located upstream of the secreted luciferase gene. Binding of transcription factors to the NF-κB enhancer element allows luciferase (tLuc) to be expressed and secreted into the surrounding medium (Catalog No. 631743; Clontech Laboratories, Mountain View, CA, USA). As shown in Figure 2a, HeLa cells were transfected with indicated plasmids. Groups i and iii received WT reporter constructs p NF-κB-Luc, whereas group ii received delta NF-κB-Luc. All groups received pSV-RLuc. The amount of expression plasmid for Rap1 (i and ii) and IkB-dominant-negative (iii) was progressively increased as indicated. Relative luciferase activity represents firefly luciferase activity recovered from pNF-κB-Luc or pdelta κB-Luc normalized to Renilla luciferase activity recovered from pSV-RLuc (Figure 2a). The results presented showed that the activity of NF-κB transcriptional activity was activated by Rap1.
Next, the relationship of IKKγ and Rap1 was examined. As shown in Figure 2b, HEK293T cells were transfected with expression plasmids for Flag-tagged Rap1 (lanes 3 and 4), HA-tagged IKKγ (lanes 2 and 4). All groups received His-tagged ubiquitin expression plasmid. His-tagged ubiquitin conjugates were pulled down. The ubiquitinated Rap1 and IKKγ were detected using western blot analysis with anti-Flag ( Figure 2bi) and anti-HA (Figure 2biii), respectively. The input lysates were also probed for anti-Flag and anti-HA to detect Rap1 and IKKγ, respectively (Figure 2bii and iv). These results show the IKKγmediated ubiquitination of Rap1.
Immunostaining results demonstrated that NF-κB translocated from the cytoplasm to nuclei when HeLa cells were stimulated with TNF-α (Figure 2ci and ii). Nonetheless, translocation of NF-κB-p65 from the cytoplasm to nuclei responding to TNF-a stimulation was diminished by knocking down Rap1 with shRap1 (Figure 2ciii and iv). These results show that the absence of Rap1 negatively regulates activation of NF-κB.    WT reporter constructs p NF-κB-Luc, whereas group B received delta NF-κB-Luc. All groups received pSV-RLuc. The amounts of expression plasmid for Rap1 (i and ii) and IkB-dominant-negative (iii) were progressively increased as indicated. Relative luciferase activity (RLA) represents firefly luciferase activity recovered from pNF-κB-Luc or p delta κB-Luc normalized to Renilla luciferase activity recovered from pSV-RLuc. (b) HEK293T cells were transfected with expression plasmids for Flag-tagged Rap1 (lanes 3 and 4) and HA-tagged IKKγ (lanes 2 and 4). All groups received His-tagged ubiquitin expression plasmid. His-tagged ubiquitin conjugates were pulled down. The ubiquitinated Rap1 and IKKγ was detected using western blot analysis with anti-Flag (i) and anti-HA (ii), respectively. The input lysates were also probed for anti-Flag and anti-HA to detect Rap1 and IKKγ, respectively (ii and iv). (c) Representative photographs show the location of NF-κB-p65 in HeLa cells (i) and shRap1-transfected HeLa cells (iii). NF-κB-p65 was translocated from the cytoplasm to nuclei responding to TNF stimulation (ii). Knocking down Rap1 with shRap1 resulted in diminished translocation of NF-κB-p65 from the cytoplasm to nuclei responding to TNF stimulation (iv).   Figures 6a and b, compared with the control group, apoptosis of cardiomyocytes was significantly increased in the MI group (Figures 6a and b, 23.3 ± 3.3% versus 1.5 ± 0.4%; Po 0.01). Transplantation of BM-MSCs or Rap1 − / − -BM-MSCs significantly decreased the cardiomyocyte apoptotic rate (Figures 6a and b; P o0.01). In addition, the apoptotic rate in the Rap1 − / − -BM-MSC group was much lower than in the  BM-MSC group (Figures 6a and b, 15.3 ± 2.1% versus 8.8 ± 1.5%; P o0.01).
The capillary density of the MI area was evaluated by PECAM (CD31) staining. Compared with the control group, the capillary density was dramatically reduced in the MI group (Figure 6ci and  Rap1-mediated MSC therapy for heart repair Y Zhang et al P o0.01), and the capillary density in the Rap1 − / − -BM-MSC group was much higher than in the BM-MSC group (Figure 6ci and ii, 51 ± 6.6/20 × versus 36.7 ± 5.8%; P o 0.01). The blood vessels of the MI area were evaluated by smooth muscle actin (SMA) staining; blood vessels were greatly increased in the BM-MSC and Rap1 − / − -BM-MSC groups compared with the MI group (Figure 6ciii and iv; P o0.01). Moreover, blood vessel density was much higher in the Rap1 − / − -BM-MSC group than in the BM-MSC group (Figure 6ciii and iv, 6.4 ± 1.1/20 × versus 3.8 ± 0.8/20 × ; P o 0.01).
One week post-cell transplantation, inflammation in heart tissue among the different experimental groups was evaluated by CD45 staining and ELISA. As shown in Figure 6di and ii, inflammation in the MI group was dramatically increased compared with the control group (Figure 6di  NF-κB has been documented as a pro-or anti-apoptotic gene that depends on the cellular context [26][27][28] and which upstream signaling it is activated by. NF-κB is also essential for the survival of many cells and functional response to stress. 23 In endothelial cells, inhibition of NF-κB activity significantly attenuates apoptosis induced by hypoxia. 29 In this study, under hypoxic conditions, apoptosis of Rap1 − / − -BM-MSCs was much lower than that of BM-MSCs because of inhibition of NF-κB activity and Bax/Bcl-2, and was confirmed using the western blot analysis. The Rap1 gene is a novel adaptor that regulates NF-κB activity by serving as the adaptor of the IKK complex, having a critical role in regulating cell apoptosis 15 and metabolism. 30 In this study, we showed that absence of Rap1 negatively regulated the activation of NF-κB in MSC. Our data also demonstrated that deletion of Rap1 in MSCs inhibits the activation of NF-κB, reduces the ratio of Bax/Bcl-2 and then directly attenuates the apoptosis induced by hypoxia, suggesting that inhibition of Rap1 enhances MSC survival capacity under hypoxia. These results prompt us to consider Rap1 − / − -BM-MSC transplantation instead of BM-MSCs for MI in that Rap1 − / − -BM-MSCs have a stronger ability to survive in the hostile environment of the injured heart. In our in vivo study, we confirmed that 4 weeks after transplantation, more Rap1 − / − -BM-MSCs than BM-MSCs survived and contributed more to the recovery of heart function.
NF-κB activation enhances expression of pro-inflammatory cytokines such as IL-6, IL-8 and TNF-α, and then provokes an excessive inflammatory response, with detrimental consequences. 31,32 Total inhibition of NF-κB leads to cell functional defects. 33 Selective inhibition of certain signals to reduce NF-κB activity has been proposed as a novel therapeutic strategy in many inflammatory diseases and cancer treatment. 34,35 Previous studies have revealed that MSCs have contradictive effects with proand anti-inflammatory properties. 36 Enhancement of the antiinflammatory effects of MSCs would be important to increase the efficacy of cardiovascular repair. 37,38 Thus, deletion of Rap1 can regulate NF-κB activation, thereby mediating pro-and antiinflammatory cytokine release. Our data demonstrated that, under hypoxia, the expression of pro-inflammatory cytokines IL-6, TNF-α and MCP-1 in the Rap1 − / − -BM-MSCs was much lower than in BM-MSC CdM. This suggests that the absence of Rap1 reduces NF-κB activation and decreases the pro-inflammatory cytokine production by MSCs. Furthermore, hypoxia-CdM from Rap1 − / − -BM-MSCs demonstrated a superior protective effect to BM-MSCs against myocardial death under hypoxia challenge, indicating that the beneficial paracrine function is enhanced in Rap1 − / − -BM-MSCs compared with WT MSCs. To further confirm the key role of Rap1 in regulating pro-inflammatory cytokine release by MSCs, we transfected Rap1 − / − -BM-MSCs with Rap1; when Rap1 − / − -BM-MSCs overexpressed Rap1, the expression of IL-6, TNF-α and MCP-1 was elevated under hypoxic conditions, suggesting that deletion of Rap1 has a strong capacity to regulate the paracrine action of MSCs.
There are some aspects of this study that need further investigation. First, the secretion of MSCs in the ischemic heart may differ under in vitro hypoxia. Second, apart from the antiinflammatory reaction of MSCs in this study, many other potential mechanisms of MSCs need to be further investigated. Third, whether deletion of Rap1 can influence other potential signaling pathways to mediate MSC-based therapy for cardiovascular disease also needs further investigation.
In summary, in this study we established that a novel NF-κB adaptor, Rap1, has an important role in the regulation of MSC paracrine components and MSC capacity for survival in a stressful environment. Compared with WT BM-MSCs, transplantation of Rap1 − / − -BM-MSCs achieved superior therapeutic efficacy in a mouse model of MI that may be attributed to Rap1-modulated cytokine bias and cell survival potential.

MATERIALS AND METHODS Mice
The WT and Rap1 − / − mice, provided by the National University of Singapore, were bred and kept in the Laboratory Animal Unit at the University of Hong Kong. All animal experiments in this study were approved by the Committee on the Use of Live Animals in Teaching and Research at The University of Hong Kong.
Isolation, culture and characterization of Rap1 − / − -BM-MSCs and BM-MSCs Rap1 − / − -BM-MSCs and BM-MSCs were isolated from 6~8-week-old Rap1 − / − and WT mice, respectively. Briefly, following killing, femurs and tibiae were quickly removed from mice and bone marrow was harvested by strong flushing with DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) and P/S on ice. The resultant cell suspension was filtered using a 70-mm filter and bone marrow cells were planted in culture dishes with complete DMEM medium and incubated at 37°C with 5% CO 2 in a humidified chamber. Non-adherent cells were removed 3 h later by replacing the culture medium. Thereafter, the culture medium was replaced with fresh medium and the process repeated every 8 h until 72 h of initial culture was reached. MSCs were characterized using flow cytometry. Oil red staining was performed to identify adipocytes, alcian blue staining for chondrocytes and alizarin red staining for osteocytes. In order to track MSCs posttransplantation, MSCs were labeled with the cell tracker dye Qtracker 655 Cell Labeling Kit (Q25021MP, Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instructions. The labeling efficiency was verified using flow cytometry. Rap1-mediated MSC therapy for heart repair Y Zhang et al

Preparation of CdM
The CdM of cultured Rap1 − / − -BM-MSCs and BM-MSCs was collected as previously reported. 39 Briefly, MSCs were rypsinized and plated on a 10-cm plate. After 24 h, the completed culture medium was replaced with serumand antibiotic-free DMEM. Twenty-four hours later, serum-free cell culture supernatants were collected, filtered by a 0.22-μm filter and centrifuged at 4°C, 4000 × g for 30 min using Amicon Ultra-4 Centrifugal Filter Devices (Millipore Billerica, MA, USA) to produce CdM. The final concentration was adjusted to 20 × of collected CdM.

NCMC isolation and culture
The NCMCs were isolated and cultured as described previously. 40 Briefly, following killing of neonatal Wistar rats (0-to 1-day-old), the hearts were promptly removed, rinsed four times with modified Hank's solution and cut into small pieces on ice. The tissue fragments were transferred to a 50-ml tube and warmed in a water bath with a magnetic bar for 10 min at 37°C. After discarding the supernatant, the minced myocardium was digested with fresh pre-warmed 0.25% trypsin for 5 min at 37°C, and the supernatant collected gently and transferred to a 50-ml tube on ice containing 7 ml FBS. These two steps were repeated to collect supernatant after which all supernatants were centrifuged at 156.8 × g for 5 min to collect the cells. Cells were re-suspended in NRVM culture medium to reduce fibroblast contamination. Finally, the supernatant was aspirated gently, and the cells were plated in MEA dishes at a density of 6 × 0 5 cells/ml. Culture medium was changed every day.

TUNEL assay
To directly visualize apoptosis of MSCs exposed to hypoxia and the NCMCs cocultured with MSC-CdM under hypoxic conditions, TUNEL assay was performed according to the manufacturer's instructions. Briefly, after washing with PBS, the cells on the cover slide were incubated with 1 μg/ml Proteinase K/10 mM Tris solution at room temperature for 15 min and then washed twice in PBS and finally incubated with 50 μl TUNEL reaction mixture in a humid chamber for 1 h at room temperature. Cover slides were washed twice with PBS, mounted with 4, 6-diamidino-2-phenylindole (DAPI), observed under a fluorescent microscope and finally photographed.

Western blot analysis
Total proteins were extracted from cells using RIPA buffer and the concentration measured. Protein (20 μg) was separated in a 10% SDS-PAGE gel and transferred on a PVDF membrane (Millipore). The membrane was washed three times in Tris-buffered saline (TBS) with 0.1% Tween-20 and blocked with 5% fat-free milk in TBS for 1 h at room temperature. Subsequently, the membrane was incubated with anti-Rap1, Bcl-2, Bax, p-NF-κB-p65 and NF-κB-p65 (Santa Cruz, Dallas, TX, USA; 1 : 2000) overnight at 4°C followed by incubation with horseradish peroxidaseconjugated secondary antibodies (1 : 10 000 dilution; Santa Cruz). Finally, the signal on the membranes was visualized using ECL and exposed to medical X-ray films for seconds or minutes.

Lentiviral construct packaging and infection
The lentivirus was packaged by co-transfecting 293T cells using the lentiviral packaging system comprising the recombinant lentiviral transfer Rap1 plasmid or shRap1 plasmid, packaging (GAG/Pol and REV) plasmids and envelope (VSV-G) plasmid. Following 48-h culture, the supernatant of transfecting 293T cells was collected, concentrated and tittered. Subsequently, the virus was used to infect Rap1 − / − -BM-MSCs or HeLa cells.

MI model and cell transplantation
A MI model was induced in ICR mice (6~8-week-old). Mice were anesthetized with an intraperitoneal injection of ketamine (100 mg/kg) and xylazine (20 mg/kg) and connected to a ventilator via tracheal intubation. The heart was exposed via a left side limited thoracotomy and the middle of the left anterior descending artery (LAD) ligated with an 8-0 silk suture. At 60 min after induction of MI, mice were randomized to receive intramyocardial injection of (1) PBS (MI group, n = 12); (2) 1.0 × 10 6 BM-MSCs (BM-MSC group, n = 13); or (3) 1.0 × 10 6 Rap1 − / − -BM-MSCs (Rap1 − / − -MSC group, n = 12) at four sites in the border of the infarct area. Another group of mice (n = 10) underwent thoracotomy without LAD ligation and served as controls (Control group).
Hemodynamic assessment At 4 weeks post-cell transplantation, mice were anesthetized and mechanically ventilated as described above. A 1.2-F pressure-volume conductance catheter connected to an ADVantage PV-Loop system (Scisence Inc., Ontario, Canada) for data acquisition was inserted into the LV cavity through the right carotid artery. Hemodynamic parameters, including LVESP and dp/dtmax, were recorded and analyzed using the LabScribe software (Scisence Inc.). The slope of the ESPVR was measured to evaluate heart function.

Assessment of fibrosis and apoptosis
Following hemodynamic assessment, mice were killed for histological and immunohistochemical study. Hearts were quickly taken out, washed with cold PBS and fixed in formalin for 24 h. Hearts were then embedded in paraffin and cut into 5-μm sections from the apex, mid-LV. Fibrosis in the different experimental groups was analyzed using a Masson's Trichrome Stain Kit (HT15, Sigma, St. Louis, MO, USA). The percentage infarct size was calculated as (fibrosis area/total LVA) × 100%. Apoptosis was evaluated with TUNEL staining as previously described.

Immunohistochemical staining
Immunohistochemical staining was performed according to the standard protocol. Following incubation with 5% bovine serum albumin for 30 min, heart sections or cells were stained with primary antibody and incubated overnight at 4°C with a 1 : 100 dilution. The antibodies were anti-Troponin (SC-8121, Santa Cruz), anti-α-SMA (SC-53142, Santa Cruz), anti-CD31 (SC-31054, Santa Cruz) and anti-CD45 (SC-25590, Santa Cruz). Sections were incubated with PBS instead of the primary antibody to serve as a negative control. Then, the second antibody with FITC-conjugated antimouse IgG (1 : 1000), anti-rat IgG (1 : 1000) or anti-goat IgG (1 : 1000) was added and incubated for 1 h at room temperature. Finally, heart sections were washed with PBS twice and mounted with 4, 6-diamidino-2phenylindole (DAPI) to stain the nucleus. Six mice from each group were analyzed; five sections were randomly collected from each mouse and then analyzed with a deconvoluted fluorescent microscope and Image J software (National Institutes of Health, Bethesda, MD, USA).