The present study examined the role of microRNA-125b (miR-125b) in myocardial ischaemia/reperfusion (I/R) injury. We constructed lentivirus-expressing miR-125b (LmiR-125b) and developed transgenic mice with overexpression of miR-125b.
LmiR-125b was transfected into mouse hearts through the right common carotid artery. Lentivirus vector (LmiR-Con) served as vector control. Untreated mice served as I/R control. Sham operation served as sham control. Seven days after transfection, the hearts were subjected to ischaemia (45 min) followed by reperfusion (4 h). Myocardial infarct size was analysed by 2,3,5-triphenyltetrazolium chloride staining. In separate experiments, hearts were subjected to ischaemia (45 min) followed by reperfusion for up to 7 days. Cardiac function was measured by echocardiography before, as well as 3 and 7 days after myocardial I/R. Increased expression of miR-125b significantly decreased I/R-induced myocardial infarct size by 60% and prevented I/R-induced decreases in ejection fraction (EF%) and fractional shortening (%FS). Transgenic mice with overexpression of miR-125b also showed the protection against myocardial I/R injury. Increased expression of miR-125b attenuated I/R-induced myocardial apoptosis and caspase-3/7 and -8 activities. Western blot showed that increased expression of miR-125b suppresses p53 and Bak1 expression in the myocardium. In addition, transfection of LmiR-125b decreased the levels of TNF receptor-associated factor 6 (TRAF6) and prevented I/R-induced NF-κB activation.
miR-125 protects the myocardium from I/R injury by preventing p53-mediated apoptotic signalling and suppressing TRAF6-mediated NF-κB activation.
It has well been documented that activation of NF-κB mediated by Toll-like receptor/interleukin-1 receptor (TLR/IL-1R) contributes to myocardial ischaemia/reperfusion (I/R) injury.1–4 Inhibition of NF-κB-binding activity has been shown to protect against myocardial I/R injury.2–6 The protective effects involve inhibition of innate immune and inflammatory responses and attenuated I/R-induced cardiac myocyte apoptosis.2–6 However, the mechanisms by which inhibition of NF-κB-binding activity decreases innate immune and inflammatory responses and attenuates cardiac myocyte apoptosis during myocardial I/R are still unclear.
Recent studies have demonstrated that activation of NF-κB regulates microRNA expression which, in turn, negatively regulates NF-κB-binding activity,7,8 thereby decreasing innate immune and inflammatory responses.9–11 MicroRNAs (miRs) are 21–23 nucleotide non-coding RNA molecules and have been identified as novel regulators of gene expression at the post-transcriptional level by binding to target messenger RNAs.9–14 miRs have been demonstrated to play a critical role in the negative regulation of innate immune and inflammatory responses by regulation of NF-κB-binding activity.9–11 Importantly, recent studies have shown that NF-κB activation regulates the expression of miRs,7,8 including miR-146, miR-155, and miR-21, etc., while these miRs, in turn, down-regulate NF-κB-binding activity. miR-21 has been demonstrated to play a protective role in myocardial I/R injury.15 We have reported that increased expression of miR-146a significantly decreases myocardial infarct size and attenuates I/R-induced cardiac dysfunction via down-regulation of NF-κB activation by targeting Interleukin-1 receptor associated kinase 1 and TNF receptor-associated factor 6 (TRAF6).16 Collectively, the data suggest that the miRs that regulate TLR/IL-R1-mediated NF-κB activation may be a new approach for the management and treatment of myocardial I/R injury.
MicroRNA-125b (miR-125b) is a homologue of lin-4, which is the first miR discovered and an important regulator of developmental timing in Caenorhabditis elegans.17 Recent studies have shown that activation of NF-κB decreases the expression of miR-125b.18,19 Tili et al. reported that the treatment of Raw 264.7 cell with lipopolysaccharide (LPS), a TLR4 ligand, suppresses the expression of miR-125b,18 while miR-125b suppresses TNF-α expression by targeting the 3′-untranslated region of TNF-α mRNA.18,19 miR-125b has been reported to play a role in down-regulation of apoptosis by repressing p53 and Bak-1.20,21 p53 is a tumour suppressor protein that plays a critical role in regulating cell cycle and apoptosis in response to hypoxia and ischaemic stress.22,23 Inhibition of p53-mediated apoptotic signalling significantly reduces I/R-induced myocardial injury.24 We have reported that increased expression of miR-125b in macrophages attenuates hypoxia/reoxygenation (H/R)-induced cell injury.25 However, whether miR-125b serves a protective role in myocardial I/R injury in vivo has not been investigated. miR-125b has been shown to target TNF-α26 and inhibit p53-mediated apoptotic signalling,27 therefore, it is possible that miR-125b serves a protective role in myocardial I/R injury.
In the present study, we examined the role of miR-125b in myocardial I/R injury. We observed that increased expression of miR-125b in the myocardium significantly decreases myocardial infarct size and prevents I/R-induced cardiac dysfunction. The mechanisms involve the inhibition of I/R-induced activation of NF-κB and the prevention of I/R-activated p53-mediated apoptotic signalling in the myocardium.
Male wild-type (WT) C57BL/6J mice were obtained from Jackson Laboratory. The experiments outlined in this manuscript conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication, 8th Edition, 2011). The animal care and experimental protocols were approved by the ETSU Committee on Animal Care.
qPCR assay of miRs
Construction of miR-125b into lentivirus-expressing system
miR-125b, mature sequence mmu-miR-125b-5p (MIMAT0000136), was constructed into lentivirus expression vector using a lentivirus-expressing system (Invitrogen Corporation) as described previously16,25 (see Supplementary material online, Methods).
Transgenic (Tg) mice with overexpression of miR-125b were developed with C57BL/6J background (see Supplementary material online, Methods).
In vitro experiments
The H9C2 rat cardiomyoblasts were obtained from the American Type Culture Collection (Rockville, MD, USA) and were maintained in Dulbecco's Modified Eagle's Medium (DMEM) supplemented 10% foetal bovine serum under 5% CO2 at 37°C.28 The cells were plated in six well plates at 1 × 105 cells/well. The cells were transfected with lentivirus-expressing miR-125b (LmiR-125b) or lentivirus-expressing vector that served as control (LmiR-con). The lentivirus-expressing vector contains a non-sense miR sequence that allows formation of a pre-miRNA hairpin predicated not to target any known vertebrate gene (Invitrogen Corporation). Stably transfected cells were selected using a blasticidin-resistant marker. The cells were subjected to hypoxia for 2 h followed by reoxygenation (H/R)25 for 24 h. The cells that were not subjected to H/R served as control (normoxia). There were three independent experiments in each group. The cells were harvested at 24 h for isolation of cellular protein.
In separate experiments, adult cardiac myocytes were isolated from nine male mice, as described previously.29 The cells were transfected with miR-125b, miR-scrambled control (miR-con), or anti-miR-125b, respectively, carried by exosomes that were isolated from bone marrow stromal cells (BMSCs)30 (see Supplementary material online, Methods). The cardiac myocytes were subjected to hypoxia (2 h) followed by reoxygenation for 24 h. Cardiac myocytes were harvested for analysis of the effect of miR-125b on H/R-induced cardiac myocyte injury.
Cardiac myocytes were isolated from three adult mice29 and subjected to hypoxia (2 h) followed by reoxygenation (24 h) (see Supplementary material online). The cells were harvested and total RNA was isolated for miR microarray analysis (see Supplementary material online, Methods).
In vivo transfection of LmiR-125b into mouse hearts
Induction of myocardial I/R injury
In situ apoptosis assay
Myocardial apoptosis was examined, as described previously,2,3,28,32 using the in situ cell death detection kit (Roche, USA). Three slides from each block were evaluated for the percentage of apoptotic cells and four fields on each slide were examined at the border areas using a defined rectangular field area with 20× magnification. A total of 100 nuclei were accounted. Numbers of apoptotic cardiac myocytes are presented as the percentage of total cells counted.
Measurement of cell viability and LDDH activity
Cell viability was assessed by measuring mitochondrial dehydrogenase activity using the MTT assay kit (Sigma). Cell injury was assessed by measurement of lactate dehydrogenase (LDH) activity in culture medium using a commercial kit (Cytotoxicity Detection Kit, Sigma).
Western blot was performed as described previously.2,3,28 The primary antibodies (anti-Fas, anti-p-53, anti-Bax, Bak-1, and TRAF6) and peroxidase-conjugated secondary antibody were purchased from Cell Signaling Technology, Inc. The signals were quantified using the G:Box gel imaging system by Syngene (Syngene, Fredrick, MD, USA).
Electrophoretic mobility shift assay (EMSA)
Nuclear proteins were isolated from heart samples as previously described.2,3,28 NF-κB-binding activity was measured using a LightShift Chemiluminescent EMSA kit (Thermo Fisher Scientific, Waltham, MA, USA).
Caspase 3/7 and -8 activities in heart tissues were measured as described previously33 using a Caspase-Glo assay kit (Promega).
ELISA for cytokine assay
The levels of cytokines (TNF-α and IL-1β) were measured by ELISA using OptEIA cytokine kits according to instructions provided by the manufacture (BD Biosciences).
Infiltration of neutrophils into the myocardium
Neutrophil accumulation in heart tissues was examined by staining with an anti-neutrophil marker antibody (NIMP-R14, Santa Cruz Biotechnology) as described previously34 (see Supplementary material online, Methods).
The data are expressed as mean ± SD. Comparisons of data between groups were made using one-way analysis of variance (ANOVA), and Tukey's procedure for multiple-range tests was performed. A P-value <0.05 was considered to be significant.
H/R decreased miR-125b expression via NF-κB activation in H9C2 cardiomyoblasts
We examined the effect of H/R on the expression of miR-125b in H9C2 cardiomyoblasts. As shown in Figure 1A, hypoxia (2 h) followed by reoxygenation (24 h) resulted in decreases in the levels of miR-125b by 33% compared with the normoxic control. To investigate whether miR-125b expression is regulated by NF-κB activation, we measured NF-κB-binding activity following H/R. Figure 1B shows that H/R induced increases in NF-κB-binding activity by 27% compared with the normoxic control. The treatment of the cells with an antioxidant, pyrrolidine dithiocarbamate (PDTC), which has been shown to inhibit NF-κB activation,35 significantly increases the expression of miR-125b in normoxic control and in H/R cells, when compared with the untreated normoxia and H/R groups, respectively. Administration of PDTC also significantly decreased NF-κB-binding activity following H/R.
LPS is a strong stimulator of NF-κB activation.36Figure 1C shows that LPS stimulation significantly induced decreases in miR-125b expression in H9C2 cells. However, treatment of the cells with antioxidants, PDTC or N-acetyl cysteine (NAC), prevented LPS-mediated suppression of miR-125b expression. Collectively, the data suggest that miR-125b expression during H/R is regulated by NF-κB activation.
We also analysed the effect of H/R on miR expression in adult cardiac myocytes. miR array showed that 43 miRNAs were differentially expressed in the cardiac myocytes after H/R at a false discovery rate of 0.05, when compared with non-H/R cells (normoxia) (see Supplementary material online, Tables S1 and S2).
Increased expression of miR-125b attenuated H/R-induced cell injury in H9C2 cardiomyoblasts
To determine whether miR-125b plays a role in the protection against H/R-induced cell injury, we generated stably transfected H9C2 cells with LmiR-125b or LmiR-control (LmiR-Con). The stably transfected cells were subjected to hypoxia (2 h) followed by reoxygenation (24 h). Untransfected H9C2 cells served as control. As shown in Figure 2A, transfection of H9C2 cells with LmiR-125b significantly increased the levels of miR-125b in the cells. Figure 2B shows that H/R induced increases in LDH activity by 5.0-fold compared with non-H/R (normoxia) groups. However, H/R-induced LDH activity in H9C2 cells was significantly attenuated by LmiR-125b transfection. LmiR-Con transfection did not affect H/R-induced LDH activity in H9C2 cells. Figure 2C shows that H/R markedly decreased cell viability (51%), compared with untreated normoxic cells. However, H/R-induced decrease in cell viability was significantly attenuated by LmiR-125b transfection, when compared with untreated H/R cells. Transfection of LmiR-con did not alter H/R-decreased cell viability.
Similarly, increased expression of miR-125b in isolated adult cardiac myocytes significantly attenuated H/R-induced cell injury (Figure 2D–F). In addition, inhibition of miR-125b expression in adult cardiac myocytes increased the susceptible to H/R-induced cell injury (Figure 2D–F).
We also examined the effect of LmiR-125b transfection on H/R-induced caspase-3 and -8 activities in H9C2 cells. Figure 2G shows that H/R induced caspase-3/7 activity by 79.8% and caspase-8 by 44.2%, when compared with normoxic cells. However, increased expression of miR-125b markedly attenuated H/R-increased caspase-3/7 and -8 activities. There was no significant difference in caspase-3/7 and -8 activities between LmiR-Con H/R cells and untreated H/R cells. The data suggest that increased expression of miR-125b plays a protective role in H/R-induced cellular injury.
miR-125b suppresses p53 and Bak-1 expression in H9C2 cardiomyoblasts
To understand the mechanisms by which overexpression of miR-125b attenuated H/R-induced caspase-3/7 and -8 activities, we examined the effect of miR-125b on p53 and Bak-1 expression in H9C2 cardiomyoblasts in the presence and absence of H/R. Figure 3 shows that the levels of p53 (A) and Bak-1 (B) were markedly lower in LmiR-125b-transfected normoxic cells than in normoxic control cells. H/R increased the levels of p53 (64.7%) and Bak-1 (91.3%) compared with normoxic control cells. The levels of p53 and Bak-1 in LmiR-125b-transfected cells were also increased following H/R stimulation. However, the levels of p53 and Bak-1 in LmiR-125b H/R cells were significantly lower by 49.3 and 42.5% compared with untransfected H/R cells and also were comparable with the normoxic control group. LmiR-Con transfection did not alter H/R-induced increases in p53 and Bak-1 expression.
Increased expression of miR-125b in isolated adult cardiac myocytes inhibited H/R-increased expression of p53 and Bak1. In contrast, inhibition of miR-125b expression in cardiac myocytes increased expression of p53 and Bak1 (Figure 3C and D).
H/R also induced increases in TNF-α production in the cardiomyoblasts compared with normoxic control (Figure 3E). TNF-α interacts with TNF receptors (TNFRs), resulting in activation of extrinsic apoptotic signalling.37 However, increased expression of miR-125b prevents H/R-induced increases in TNF-α production (Figure 3E).
In vivo increased expression of miR-125b decreased infarct size and improved cardiac function following myocardial I/R
To examine whether increased expression of miR-125b will induce protection against myocardial I/R injury, we transfected mouse hearts with LmiR-125b or LmiR-Con before the hearts were subjected to I/R. Figure 4A shows that the green fluorescent protein (GFP) that is carried by LmiR-125b or LmiR-Con is mainly expressed in the myocardium. qPCR data showed that the levels of miR-125b in LmiR-125b-transfected hearts were significantly increased by 8.2-fold compared with LmiR-Con-transfected hearts (Figure 4B). Figure 4C shows that I/R induced significant injury as denoted by infarct size in untreated hearts. In contrast, infarct size was significantly reduced (60%) following I/R in LmiR-125b-transfected mice compared with the untreated I/R group. LmiR-Con transfection did not alter I/R-induced myocardial infarct size.
We also examined the effect of increased expression of miR-125b on cardiac function following myocardial I/R. As shown in Figure 4D, ejection fraction (EF%) and fractional shortening (%FS) in untreated I/R hearts were significantly reduced by 39.6 and 44.6% on Day 3 and by 26.8 and 32.4% on Day 7 after myocardial I/R compared with baseline. However, I/R-induced cardiac dysfunction was prevented by LmiR-125b transfection. EF% and %FS values in LmiR-125b-transfected mice were not significantly decreased at 3 and 7 days after myocardial I/R compared with LmiR-125b baseline. Transfection of miR-Con did not alter I/R-induced cardiac dysfunction.
I/R-induced myocardial apoptosis was attenuated by LmiR-125b transfection
Cardiac myocyte apoptosis contributes to myocardial I/R injury.38 We examined the effect of increased expression of miR-125b on I/R-induced myocardial apoptosis. Figure 5A shows that I/R markedly induced myocardial apoptosis compared with sham control. In LmiR-125b-transfected mice, I/R-induced myocardial apoptosis was significantly attenuated, when compared with the untreated I/R group (12.0 ± 2.87 vs. 28.5 ± 1.72%). LmiR-Con transfection did not affect I/R-induced myocardial apoptosis.
Figure 5B shows that I/R increased caspase-3/7 (31.4%) and caspase-8 (40.5%) activities in the myocardium compared with sham control. However, increased expression of miR-125b prevented I/R-induced caspase-3/7 and -8 activities compared with the I/R group. There was no significant difference in caspase-3/7 and -8 activities between LmiR-Con I/R mice and the untreated I/R group.
Transfection of LmiR-125b prevented the increase in p53, Bak-1, Bax, and Fas levels in the myocardium following I/R
To determine the mechanisms by which increased expression of miR-125b-attenuated I/R-induced myocardial apoptosis, we examined the levels of pro-apoptotic effectors including p53, Bak-1, Bax, and Fas in the myocardium. As shown in Figure 5C, I/R increased the levels of p53 by 94%, Bak-1 by 72%, Bax by 90%, and Fas by 64.5%, respectively, compared with sham control. LmiR-Con transfection did not alter the levels of pro-apoptotic effectors in the myocardium. However, increased expression of miR-125b prevented the increases in p53, Bak-1, Bax, and Fas levels in the myocardium following I/R. Both p53 and Bak-1 levels in LmiR-125b-transfected hearts were markedly lower than in the untreated sham control.
Increased expression of miR-125b decreased TRAF6 expression and attenuated neutrophil infiltration in the myocardium
Activation of TLR-mediated NF-κB signalling contributes to myocardial I/R injury by promoting the inflammatory responses.2–4 We examined the effect of LmiR-125b transfection on NF-κB activation during myocardial I/R. TRAF6 is an important effector in the TLR-mediated NF-κB activation pathway.36Figure 6A shows that the levels of TRAF6 in LmiR-125b-transfected sham and I/R hearts were significantly lower than in untransfected sham and I/R groups. Transfection of LmiR-Con did not affect myocardial TRAF6 levels in the presence and absence of I/R. I/R significantly induced NF-κB-binding activity compared with the sham control (Figure 6B). In contrast, increased expression of miR-125b prevented I/R-induced NF-κB-binding activity. There was no significant difference in NF-κB-binding activity between the untransfected I/R group and LmiR-Con-transfected I/R mice. In addition, we have observed that increased expression of miR-125b significantly attenuated I/R-induced neutrophil infiltration into the myocardium (see Supplementary material online, Figure S1).
Transgenic mice with overexpression of miR-125b protect against myocardial I/R injury
To confirm our observation, we developed Tg mice with overexpression of miR-125b. Figure 7A shows increased expression of miR-125b in the myocardium of Tg mice. Tg and WT mice were subjected to I/R. Figure 7B shows that I/R markedly induced myocardial infarct size in WT mice. However, infarct size in Tg mice was significantly reduced by 50% compared with WT I/R mice. In contrast, inhibition of miR-125b by transfection of anti-miR-125b into the myocardium of WT mice resulted in susceptible to I/R-induced injury. Infarct size was markedly greater than in the untreated I/R group (Figure 7B). Tg mice also showed the prevention of cardiac dysfunction following I/R. As shown in Figure 7C, both EF% and %FS values were significantly decreased by 42.6 and 48.8% in WT mice after myocardial I/R. However, I/R-induced cardiac dysfunction was prevented in Tg I/R mice.
The present study demonstrates that increased expression of miR-125b in the myocardium significantly decreased myocardial infarct size and prevented I/R-induced cardiac dysfunction. To the best of our knowledge, this is the first report that miR-125b exerts a protective role in myocardial I/R injury. Inhibition of miR-125b resulted in significant susceptibility to I/R-induced myocardial injury. The mechanisms by which miR-125b protects against myocardial I/R injury involve the prevention of I/R-induced NF-κB activation and p53-mediated apoptotic signalling. Our data suggest that modulation of miR-125b may be a useful strategy for the induction of cardioprotection.
We have previously reported that scavenger receptor type A (SR-A) deficiency attenuates myocardial I/R injury.25 Interestingly, we have observed that the levels of miR-125b in SR-A deficient mice are significantly greater than in WT mice.25 Transfection of macrophages with miR-125b mimics attenuated H/R-induced cell injury,25 indicating that miR-125b may play a protective role in myocardial I/R. Indeed, we demonstrate in the present study that increased expression of miR-125b attenuates myocardial I/R injury and prevents I/R-induced cardiac dysfunction. In contrast, inhibition of miR-125b expression resulted in more injury of the myocardium following I/R. miR-125b is highly conserved among mammals, vertebrates, and nematodes.8 miR-125b is expressed in several organs, including brain, heart, lung, spleen, and skeletal muscle.39 Recent studies have reported that miR-125b expression is regulated by NF-κB activation,40 while miR-125b acts as a negative regulator of the NF-κB pathway by reducing the levels of tumour necrosis factor18,41 and by enhancing the stability of the NF-κB inhibitor NKIRAS2 (KBRAS2).41 We have observed that hypoxia followed by reoxygenation decreased the expression of miR-125b and increased NF-κB-binding activity in H9C2 cells. Inhibition of NF-κB-binding activity by anti-oxidant, PDTC, which has been reported to inhibit NF-κB activation,35 prevents H/R-induced decreases in miR-125b expression. The data indicate that H/R-induced decreases in the expression of miR-125b are mediated, in part, by NF-κB activation.
However, our in vivo data show that increased expression of miR-125b in the myocardium significantly prevents I/R-induced myocardial NF-κB-binding activity. Importantly, we have observed that the expression of TRAF6 was suppressed by transfection of LmiR-125b. TRAF6 plays a crucial role in the induction of inflammatory responses via activation of IκB kinases, leading to NF-κB nuclear translocation and activation.36,42 NF-κB activation regulates inflammatory cytokine expression.36,42 We have observed that transfection of H9C2 cells with LmiR-125b prevents H/R-induced increases in TNF-α production. Our observation is consistent with previous reports, showing that TNF-α mRNA is the target for miR-125b.18,19 Androulidaki et al.19 and Tili et al.18 reported that LPS, a TLR4 ligand, suppresses macrophage expression of miR-125b, while miR-125b negatively regulates TNF-α expression.18,19 Collectively, our data support the concept that NF-κB activation regulates miR-125b expression during H/R, while increased expression of miR-125b negatively regulates NF-κB activation. Therefore, inhibition of NF-κB activation by targeting TRAF6 in the myocardium could be an important mechanism for miR-125b protection against myocardial I/R injury. In addition, increased expression of miR-125b also down-regulated systemic inflammatory responses following myocardial I/R. We also observed that transfection of LmiR-125b significantly attenuated I/R-induced infiltration of neutrophils into the myocardium.
Myocardial apoptosis contributes to myocardial I/R injury.38 We have observed that increased expression of miR-125b significantly attenuated I/R-induced myocardial apoptosis. The mechanisms by which miR-125b attenuated I/R-induced myocardial apoptosis involve suppression of p53-mediated apoptotic signalling in the myocardium following myocardial I/R. p53 is a tumour suppressor protein that regulates and interacts with the apoptotic protein Bax. Bax acts as an antagonist against anti-apoptotic Bcl2, resulting in increases in mitochondrial membrane permeability and the release of cytochrome c.22,23 In addition, apoptotic lipid products serve as chemokines that promote infiltration of inflammatory cells into the myocardium during myocardial I/R.22,23 Therefore, p53 is a critical pro-apoptotic effector for myocardial apoptosis during myocardial I/R injury.24 Inhibition of p53 expression is an important approach for attenuation of myocardial I/R injury.24 Our in vitro data showed that increased expression of miR-125b in H9C2 cells and adult cardiac myocytes suppressed the expression of p53 and Bak-1 in both non-H/R and H/R cells, suggesting that miR-125b targets both p53 and Bak-1 in cardiomyoblasts. In vivo data demonstrated that transfection of the myocardium with LmiR-125b prevents I/R-induced increases in the expression of p53 and Bak-1 in the myocardium. In addition, increased expression of miR-125b also prevents I/R-induced increases in Fas levels and caspase-3/7 and -8 activities in the myocardium. The data indicate that anti-apoptotic properties of miR-125b include inhibition of both extrinsic and intrinsic apoptotic signalling pathways during myocardial I/R.
In summary, we demonstrated in the present study that miR-125b plays a significant role in the protection against myocardial I/R injury. The mechanisms involve the inhibition of NF-κB activation as well as TNF-α production and the prevention of p53-mediated apoptotic signalling following myocardial I/R. Our data suggest that miR-125b is a target for the induction of protection against myocardial I/R injury.
Conflict of interest: none declared.
This work was supported by NIH (HL071837 to C.L., GM083016 to C.L. and D.L.W., GM53522 to D.L.W.).