Monocyte-derived extracellular Nampt-dependent biosynthesis of NAD+ protects the heart against pressure overload

Nicotinamide phosphoribosyltransferase (Nampt) catalyzes the rate-limiting step in the salvage pathway for nicotinamide adenine dinucleotide (NAD+) biosynthesis, and thereby regulates the deacetylase activity of sirtuins. Here we show accommodative regulation of myocardial NAD+ by monocyte-derived extracellular Nampt (eNampt), which is essential for hemodynamic compensation to pressure overload. Although intracellular Nampt (iNampt) expression was decreased in pressure-overloaded hearts, myocardial NAD+ concentration and Sirt1 activity were preserved. In contrast, iNampt was up-regulated in spleen and monocytes, and circulating eNampt protein and nicotinamide mononucleotide (NMN), a key precursor of NAD+, were significantly increased. Pharmacological inhibition of Nampt by FK866 or depletion of monocytes/macrophages by clodronate liposomes disrupted the homeostatic mechanism of myocardial NAD+ levels and NAD+-dependent Sirt1 activity, leading to susceptibility to cardiomyocyte apoptosis and cardiac decompensation in pressure-overloaded mice. These biochemical and hemodynamic defects were prevented by systemic administration of NMN. Our studies uncover a crucial role of monocyte-derived eNampt in myocardial adaptation to pressure overload, and highlight a potential intervention controlling myocardial NAD+ against heart failure.

synthesized by the salvage pathway from its precursors nicotinic acid, nicotinamide (NAM), or NAM ribose 5 . In mammalian cells, conversion of NAM to NAM mononucleotide (NMN) by NAM phosphoribosyltransferase (Nampt) is the rate-limiting step of this pathway 6 . Intriguingly, Nampt exists not only as an intracellular form (iNampt) but also as an extracellular form (eNampt) in an enzymatically active dimer 7 , and eNampt, previously named pre-B cell colony-enhancing factor or visfatin, is produced and secreted by adipocytes, mononuclear cells, hepatocytes, and cardiomyocytes 7−9 . A high concentration of NMN is present in mouse plasma, and eNampt-mediated extracellular production of NMN regulates intracellular NAD + biosynthesis in metabolic tissues such as liver and white adipose tissue in high-fat diet-induced diabetic mice 10 .
In the heart, the expression level of Nampt protein was down-regulated under pathological conditions such as ischemia, ischemia/reperfusion, and pressure overload 11 . However, controversy exists over whether Nampt is beneficial or detrimental to cardiac pathophysiology 9,11,12 , and it remains unclear how myocardial NAD + biosynthesis is regulated by iNampt and eNampt under pathological conditions. Here we demonstrate that myocardial NAD + concentration is preserved in spite of decreased iNampt expression in mice after transverse aortic constriction (TAC). Mechanistically, monocyte-derived eNampt contributes to preservation of myocardial NAD + levels that is sufficient for the functional compensation to pressure overload. Our studies provide mechanistic insights into the inter-tissue regulation of cardiac homeostasis involving bone marrow-derived monocytes, and point toward a therapeutic strategy of manipulating this pathway for treatment of heart failure.

Results
Nampt-dependent biosynthesis of NAD + protects rat neonatal cardiomyocytes from doxorubicin-induced cardiotoxicity. We first examined the impact of Nampt-dependent NAD + biosynthesis on stress response to injury in cultured cardiomyocytes of neonatal rats. Doxorubicin (DOX), an anthracycline anticancer drug, has cardiotoxicity through multiple mechanisms including an increase in oxidative stress, myofibrillar deterioration, and intracellular Ca 2+ dysregulation 13 . Stimulation with DOX significantly decreased NAD + concentration and enzymatic activity of Sirt1, and conversely increased acetylation of nuclear proteins in cardiomyocytes, which was further exaggerated by concomitant treatment with FK866, a selective Nampt inhibitor 14 (Fig. 1a-c). It has been reported that Sirt1 regulates mitochondrial function and energy utilization, and protects cardiomyocytes from cell death and degeneration 15,16 . Consistently, DOX-induced decrease in cell viability and in expression levels of genes relevant to mitochondrial function, such as Sod2, Ppargc1a, Pparg, and mt-Co3, were further exaggerated by treatment with FK866 (Fig. 1d,e). Conversely, treatment of DOX-stimulated cardiomyocytes with NMN significantly increased NAD + concentration, enzymatic activity of Sirt1, decreased acetylation of nuclear proteins, and restored cell viability and the expression levels of mitochondrial genes ( Fig. 1f-j). These results suggest that Nampt-dependent biosynthesis of NAD + is protective against DOX-induced cardiotoxicity in cultured cells.
Myocardial NAD + concentration and Sirt1 activity are preserved in spite of decreased iNampt expression in TAC-operated mouse hearts. Next, to examine the pathophysiological connection between Nampt-dependent NAD + biosynthesis and heart failure, we imposed pressure overload on mice by producing TAC. In this model, we were able to induce adaptive cardiac hypertrophy with preserved systolic function at 2 w, and maladaptive and heart failure at 8 w after operation (Supplementary Table 1). At 8 w after operation, mRNA and protein levels of Nampt were significantly decreased in TAC-operated hearts compared with sham-operated hearts (Fig. 2a,b). But surprisingly, enzymatic activity as well as mRNA and protein levels of Sirt1, and acetylation level of forkhead box protein O1 (FoxO1) were unchanged between TAC-and sham-operated hearts ( Fig. 2c-f). In spite of decrease in iNampt protein levels, NAD + concentration was preserved in TAC-operated hearts, as revealed by direct measurement using an HPLC system (Fig. 2g). We also found that NMN concentration was significantly higher in TAC-operated hearts than that in sham-operated hearts (Fig. 2h). These results suggest that an increase in myocardial supply of NMN might compensate for a decrease in iNampt-dependent synthesis of NMN to achieve the steady concentration of NAD + in TAC-operated hearts. Nampt protein expression is increased in spleen and circulating CD11b + monocytes in TAC-operated mice. In TAC-operated mice, both plasma levels of eNampt protein and NMN were significantly increased, as compared with sham-operated mice (Fig. 3a,b), suggesting that the increase in eNampt-dependent biosynthesis of plasma NMN may contribute to the increase in myocardial NMN levels in TAC-operated mice. In addition, treatment of rat neonatal cardiomyocytes with TAC-operated mouse serum induced significantly higher deacetylase activity of Sirt1 than that with sham-operated mouse serum (Fig. 3c), suggesting that increase in eNampt-dependent biosynthesis of plasma NMN was sufficient for activation of Sirt1 in myocardium.
subpopulation potentially contributes to the increased Nampt expression in peripheral mononuclear cells, we separated cells from peripheral mononuclear cells of unoperated mice using myelomonocytic lineage marker CD11b (Supplementary Figure 2a), and found that Nampt mRNA expressions were significantly higher in CD11b + cells than in CD11b -cells (Fig. 3e). Furthermore, Nampt mRNA expressions in CD11b + cells from TAC-operated mice were significantly higher than those from sham-operated mice (Fig. 3f). CD11b + cells isolated from peripheral mononuclear cells showed high expression levels of Emr1 mRNA (Supplementary Figure 2b), and thus were considered to be CD11b+ , F4/80+ monocytes. Since spleen is a reservoir for extramedullary monocytes that are mobilized in response to myocardial ischemia 17 , we speculated that monocytes might be the possible source of increased plasma eNampt protein.
Nampt-dependent biosynthesis of myocardial NAD + is essential for functional compensation to pressure overload in mice. To examine functional significance of Nampt in hemodynamic compensation to pressure overload, we intraperitoneally administered FK866 or mock to TAC-or sham-operated mice. Since a decrease in cardiac iNampt expression and an increase in plasma eNampt expression were observed as early as at 1 w after TAC operation (Supplementary Figure 3), we started administration of FK866 on the day of operation. Echocardiographic examination revealed that administration of FK866 for 1 w induced a significant increase in left ventricular end-diastolic dimension (LVEDD) and a significant decrease in fractional shortening (FS) in TAC-operated mice, while these parameters were unchanged in sham-operated mice (Supplementary Table 2). As a consequence, approximately 65% of TAC-operated mice died within 1 w after treatment with FK866 (Fig. 4a). In TAC-operated mice, FK866 treatment induced a significant decrease in myocardial concentration of NAD + (Fig. 4b) and deacetylase activity of Sirt1 (Fig. 4c), and an increase in acetylation level of FoxO1 (Fig. 4d), although these changes were not observed in sham-operated mice ( Fig. 4b-d). As a consequence, FK866-induced reduction of NAD + led to a significant decrease in expression levels of genes relevant to mitochondrial function, such as Tfam, Pparg, and Esrra (Fig. 4e), and a significant increase in the number of TUNEL-positive cardiomyocytes (Fig. 4f,g). These results suggest that functional inhibition of Nampt disrupted the homeostatic mechanism of myocardial NAD + levels and NAD + -dependent Sirt1 activity, and thereby induced cardiac decompensation in pressure-overloaded mice.
To further assess whether the decrease in myocardial NMN and NAD + levels caused cardiac decompensation in FK866-treated TAC-operated mice, we intraperitoneally administered NMN or mock to a.  Table 3), and prevented premature death in FK866-treated TAC-operated mice (Fig. 5a). NMN treatment restored myocardial concentration of NAD + (Fig. 5b), deacetylase activity of Sirt1 (Fig. 5c), acetylation level of FoxO1 (Fig. 5d), expression levels of genes relevant to mitochondrial function, such as Sod2, Ppargc1a, Tfam, Pparg, Esrra, and mt-Co3 (Fig. 5e), and the number of TUNEL-positive cardiomyocytes ( Fig. 5f,g). Collectively, these results suggest that Nampt-dependent biosynthesis of NAD + is essential for myocardial Sirt1 deacetylase activity and functional compensation to pressure overload in mice.
Monocyte-derived eNampt is crucial for myocardial synthesis of NAD + sufficient for functional compensation to pressure overload in mice. The increase in Nampt expression in monocytes and plasma eNampt concentration suggested a causal link between monocyte-derived eNampt and Nampt-dependent functional compensation to pressure overload. To test the importance of monocyte-derived eNampt in this process, we depleted monocytes/macrophages by treatment with clodronate liposomes (CloLip) in mice 18 . We confirmed by flow cytometric analysis that, in unoperated mice, CloLip induced approximately 3.2-fold reduction in the percentage of circulating CD11b+ , F4/80 + monocytes at 1 d after the initiation of treatment (Fig. 6a, Supplementary Figure 4a). The percentage of CD11b+ , F4/80 + monocytes was increased approximately 3.1-fold at 5 d after TAC operation, but it was decreased approximately 1.6-fold by concomitant CloLip treatment (Fig. 6a, Supplementary Figure 4b).
In contrast, the populations of CD11b+ , F4/80-cells and CD11b -, F4/80-cells were not significantly changed after CloLip treatment (Supplementary Figure 4). CloLip-treated mice became lethargic and approximately 75% of CloLip-treated mice died within 5 d after TAC operation, although control liposomes (CntrlLip)-treated mice appeared normal (Fig. 6b). Echocardiographic examination revealed a significant increase in LVEDD and a significant decrease in %FS in CloLip-treated mice, while these parameters were unchanged in CntrlLip-treated mice (Supplementary Table 4). CloLip treatment induced a significant decrease in plasma eNampt protein (Fig. 6c) without a significant change in cardiac iNampt protein (Fig. 6d), leading to a significant decrease in NAD + concentration (Fig. 6e) and deacetylase activity of Sirt1 (Fig. 6f), and an increase in acetylation level of FoxO1 (Fig. 6g). CloLip-induced reduction of NAD + led to a significant increase in the number of TUNEL-positive cardiomyocytes (Fig. 6h,i). These results suggest that clodronate liposome-mediated reduction of CD11b + cells decreased plasma eNampt protein, and thereby induced cardiac decompensation in pressure-overloaded mice. We further assessed whether NMN treatment prevented CloLip-induced cardiac decompensation in pressure-overloaded mice. NMN treatment significantly improved echocardiographic parameters such as LVEDD and %FS (Supplementaty Table 4), and prevented premature death in CloLip-treated TAC-operated mice (Fig. 7a). NMN treatment restored myocardial concentration of NAD + (Fig. 7b), deacetylase activity of Sirt1 (Fig. 7c), acetylation level of FoxO1 (Fig. 7d), and the number of TUNEL-positive cardiomyocytes (Fig. 7e,f). Collectively, these results suggest that monocyte-derived eNampt is crucial for myocardial synthesis of NAD + sufficient for functional compensation to pressure overload in mice.

Discussion
Myocardial compensation to increased workload is achieved through complex and integrative processes involving not only a number of intracellular signaling pathways but also cross-talking networks of intercellular and inter-tissue communications. Our present study demonstrates a hitherto unknown mechanism of inter-tissue regulation of cardiac homeostasis by circulating monocytes. We provide proof  that an increase in circulating NMN mediated by up-regulation of monocyte-derived eNampt compensates for a decrease in myocardial iNampt expression to maintain the myocardial NAD + level in pressure-overloaded mice (Fig. 8).
It has been reported that down-regulation of Nampt increased apoptotic cell death in cultured neonatal rat cardiomyocytes 11 , and we also observed that Nampt-dependent biosynthesis of NAD + was protective against DOX-induced cardiotoxicity. Cardiac-specific overexpression of Nampt or administration of NMN increased myocardial NAD + levels, and protected the heart from ischemia or ischemia/reperfusion injury in mice, indicating that Nampt is crucial for myocardial NAD + synthesis and cardioprotection in stressed hearts 11,12 . On the other hand, Pillai et al. reported that cardiac-specific overexpression of Nampt induced cardiac hypertrophy and interstitial fibrosis spontaneously in mice, and that heterozygous knockout of Nampt attenuated isoproterenol-and angiotensin II-induced cardiac hypertrophy, indicating Nampt is a positive regulator of adverse cardiac remodeling 9 . Considering the much greater expression levels of myocardial Nampt in mice reported by Pillai et al., excessive production of Nampt might be detrimental to cardiac pathophysiology. Our present data clearly advance our understanding of hitherto controversial role of Nampt in stress response to hemodynamic overload, demonstrating that inhibition of Nampt induced cardiac decompensation in the early phase of pressure overload.
Recently, eNampt, like cytokines and growth factors, has been reported to act directly on cells, but the actions of eNampt on cardiomyocytes are enigmatic. Treatment with eNampt attenuated H 2 O 2 -induced a.  21 . It has remained unclear whether monocyte-derived eNampt has biological effects on cardiomyocytes, but we found that plasma NMN was significantly increased in TAC-operated mice, and that treatment of rat neonatal cardiomyocytes with TAC-operated mouse serum induced significantly higher deacetylase activity of Sirt1. Therefore, NMN might be extracellularly converted from NAM by eNampt at a systemic level, and used as a substrate for NAD + biosynthesis after transport flux into the cells via an unidentified NMN transporter in the hearts. Further studies using elaborate genetic models to delete Nampt specifically in cardiomyocytes or monocytes will be required to dissect the cooperative regulation of myocardial NAD + by iNampt and eNampt in the hearts subjected to pressure overload and ischemia/reperfusion. Nampt-dependent NAD + biosynthesis is a key determinant of the deacetylase activity of sirtuins 6 . Our analysis of Sirt1 deacetylase activity revealed that it was invariably correlated with myocardial NAD + concentration both in vitro and in vivo. In a mouse model of ischemia/reperfusion, cardiac-specific overexpression of Sirt1 reduced infarct size through activation of FoxO1 and manganese superoxide dismutase, while cardiac-specific knockout of Sirt1 exacerbated myocardial injury 22 . However, contradictory phenotypes attributable to gene dosages of the transgene were observed in the development of cardiac hypertrophy in mice with cardiac overexpression of Sirt1. Low to moderate levels of Sirt1 overexpression attenuated aging-dependent progression of cardiac remodeling and dysfunction, but high levels of Sirt1 overexpression increased oxidative stress and apoptosis, leading to cardiac dysfunction 23,24 . Beneficial effects of Sirt1 are mediated by a combination of multiple mechanisms such as deacetylation of peroxisome proliferator-activated receptor-γ coactivator-1α for modulating mitochondrial function 16,25 , up-regulation of manganese superoxide dismutase for scavenging ROS 22,26 , deacetylation of p53 for inhibiting apoptotic cell death 27 , and deacetylation of FoxO1 for inducing autophagy 28 . Conversely, Oka et al. reported that overexpressed Sirt1, by forming a complex with peroxisome proliferator-activated receptor-α , suppressed the estrogen-related receptors pathways, leading to mitochondrial dysfunction and progression of heart failure 29 . In TAC-operated mice, the expression levels of genes relevant to mitochondrial function were correlated with Sirt1 deacetylase activity, either when it was decreased by treatment with FK866 or when it was restored by treatment with NMN, supporting the possibility that Nampt-dependent biosynthesis of myocardial NAD + is implicated in the regulation of mitochondrial function. Insomuch as Sirt3 and Sirt7 as well as Sirt1 have been reported to have beneficial roles in the a. heart 15 , further studies will be required to understand the whole picture of compensatory mechanisms facilitated by eNampt-dependent biosynthesis of NAD + and NAD + -dependent activation of Sirt1 and other sirtuins in pressure-overloaded hearts.
In brain and pancreas, basal expression levels of iNampt are extremely low, and circulating NMN maintained by plasma eNampt functions as an essential substrate for intracellular NAD + biosynthesis 7 . Recent studies have demonstrated the effectiveness of NMN supplementation for restoration of NAD + in these tissues 7,10,30,31 . Under the pathological conditions in which circulating NMN levels decline, NMN supplementation may provide therapeutic benefit. It has been reported that the levels of Nampt protein as well as NAD + decline with aging in multiple tissues 10,32 . We observed that aged mice showed a significant decrease in the levels of myocardial NAD + and plasma eNampt protein, as compared with young mice, although the levels of myocardial iNampt protein were unchanged (Supplemental Figure S5). It is well known that the prevalence of heart failure increases with age, and that the elderly patients with heart failure have a worse prognosis than the younger ones 33 . It makes sense to hypothesize that age-dependent decline in eNampt-mediated supply of NMN might cause a decrease in myocardial NAD + , leading to increased susceptibility to heart failure with higher morbidity and mortality in the elderly. In addition, a recent study demonstrated that high levels of plasma eNampt are associated with favorable clinical outcome in patients with dilated cardiomyopathy 34 , and our study may provide a likely mechanistic explanation for the correlation between plasma eNampt and heart failure phenotypes. Our study highlighted a novel mechanism of circulating eNampt-mediated regulation of myocardial NAD + under stressed conditions, and opened up an opportunity to explore whether restoration of myocardial NAD + has a therapeutic potential for heart failure.

Mice, operation, monocyte/macrophage depletion, and transthoracic echocardiography.
All of the experiments were approved by the Institutional Animal Care and Use Committee of Osaka University, and carried out in accordance with the guidelines of Osaka University. C57BL/6J mice were purchased from CLEA Japan, Inc. For TAC operation, we anesthetized 8 to 9-week-old male mice by intraperitoneal injection of medetomidine hydrochloride (0.3 mg/kg), midazolam (4 mg/kg), and butorphanol (5 mg/kg) 35 , and anesthesia was monitored by pinching the toe. Respiration was artificially controlled with a tidal volume of 0.2 ml and a respiratory rate of 110 breaths/min. After median sternotomy, transverse aorta between the branches of right brachiocephalic and left common carotid arteries was constricted with 7-0 silk strings by ligating the aorta with splinting a blunted 27 gauge needle, which was removed after the ligation. The chest was then closed, and mice were allowed to recover from anesthesia while their body temperature was kept at 37 °C. Post-operative analgesia (meloxicam, 5 mg/kg/24 h) was administered subcutaneously for 48 h. The surgeon had no information about the mice used in this study. To inhibit Nampt-dependent biosynthesis of NAD + in vivo, we administered FK866 (10 mg/kg/d) or mock (corn oil) by intraperitoneal injection to mice for the duration of the experiment 36 . For NMN treatment, we administered NMN (500 mg/kg) or mock (normal saline) by intraperitoneal injection twice a day for the duration of the experiment 10,30 . Clodronate and control liposomes (Clophosome, Combo Kit) were purchased from FormuMax Scientific, Inc. To induce depletion of monocytes and macrophages, a dose of 25 μ l of clodronate or control liposomes was administered intraperitoneally. For evaluation of cardiac dimensions and contractility, transthoracic echocardiography was performed on conscious mice with a Vevo 770 Imaging System using a 25 MHz linear probe (Visual Sonics Inc.). Mice were sacrificed to collect tissue samples after 24 h starvation.
Cell culture and cell viability assay. The experiments for primary cultures of cardiomyocytes was approved by the Animal Study Committee of Osaka University, and were carried out in accordance with the guidelines of Osaka University. Briefly, cervical dislocation euthanasia was performed by a trained personal prior to harvesting the cardiac tissue of 1-day-old Wistar rats according to the American Veterinary Medical Association guidelines for the euthanasia of animals. Ventricles were minced and digested in Ca 2+ -free Hank's balanced salt solution (HBSS) (Hyclone) containing 0.1% trypsin (Hyclone) and 70 u/ml collagenase (Worthington Biochemical Corp.). Cardiomyocytes were enriched after pre-plating, plated at a field density of 1 × 10 5 cells/cm 2 , and cultured in DMEM supplemented with 10% fetal bovine serum (BGS) for 24 h. After serum-starvation (1% BGS) for 24 h, cardiomyocytes were stimulated with DOX (1 μ M for 3 h) and allosteric Nampt inhibitor FK866 (500 nM for 24 h) or NMN (500 μ M for 24 h) 10 . DOX was purchased from Sigma-Aldrich, and FK866 and NMN were purchased from Cayman Chemical and Sigma-Aldrich, respectively. Cell viability was determined using CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega) that is based on metabolic conversion of a tetrazolium compound MTS to a colored product by living cells.
Western blot analysis. Cells were lysed and tissues were homogenized in buffer (20 mM Tris, pH 7.4, Measurement of NAD + and NMN levels. NAD + and NMN levels were determined using an HPLC system (Shiseido Co., Ltd.) with a CAPCELL PAK C18 MGIII S5 column (15 cm × 2.0 mm; Shiseido) and a YMC-Pack Pro C18 RS column (15 cm × 4.6 mm; YMC Co., Ltd.), respectively, as described previously 10 . Frozen tissues or freshly collected plasma of mice, or cultured cardiomyocytes were extracted in 1 M perchloric acid (200 μ l/10 mg tissue, 100 μ l/10 μ l plasma, 500 μ l/6 cm dish). The extracts were centrifuged at 15,000 rpm for 10 min at 4 °C, and the resulting supernatants were neutralized in 3 M K 2 CO 3 on ice for 10 minutes. After clearing extracts, aliquots of 100 μ l were mixed with 50 μ l of buffer A (50 mM K 2 HPO 4 /KH 2 PO 4 , pH 7.0) and 50 μ l of water. For NAD + measurement, the HPLC was run at a flow rate of 200 μ l/min with 100% buffer A from 0-5 min, a linear gradient to 95% buffer A/5% buffer B (100% methanol) from 5-6 min, 95% buffer A/5% buffer B from 6-11 min, a linear gradient to 85% buffer A/15% buffer B from 11-13 min, 85% buffer A/15% buffer B from 13-23 min, and a linear gradient to 100% buffer A from 23-24 min. For NMN measurement, the HPLC was run at a flow rate of 700 μ l/min with 100% buffer A from 0-5 min, a linear gradient to 60% buffer A/40% buffer B from 5-10 min, 60% buffer A/40% buffer B from 10-32.5 min, a linear gradient to 100% buffer A from 32.5-35 min. NAD + and NMN were usually eluted at 14 min and 25 min, respectively. NAD + and NMN levels were quantitated based on the peak area compared to a standard curve and normalized to the weight of frozen tissues, the volume of plasma, and the number of cultured cardiomyocytes.