We have determined a novel new mechanism by which NAMPT regulates signaling pathways that influence major cellular fate decisions such as survival and programmed cell death. Our data clearly indicate that NAMPT-mediated effects on EC apoptosis are not triggered by iNAMPT-dependent NMN synthesis but rather involve eNAMPT secretion and subsequent engagement of TLR4 to produce reductions in EC apoptosis. This conclusion is supported by the effects of recombinant human NAMPT on TNF-α-induced EC apoptosis (caspase-3 cleavage) and by the inability of pharmacological inhibition of iNAMPT phosphoribosyltransferase activity (FK866) to alter levels of TNF-α-induced EC apoptosis. Furthermore, supplementation with NAMPT-containing condition media from control EC, but not NAMPT-depleted media from NAMPT-silenced EC, resulted in attenuation of EC susceptibility to TNF-α-induced EC apoptosis. This protection observed with NAMPT-containing condition media was abolished by the addition of an eNAMPT neutralizing antibody as well as by TLR4 inhibition.
Reductions in NMN levels via FK866-mediated inhibition of NAMPT enzymatic activity, induces a diverse range of biologic effects including effects on autophagy (38), oncosis (39) and ER-stress (40). In addition to iNAMPT regulation of intracellular NMN synthesis, iNAMPT may exert homeostatic effects that are dependent on protein-protein interactions involved in the regulation of oxidative stress and inflammation such as NADH dehydrogenase subunit 1 (ND1), interferon-induced transmembrane protein 3 (IFITM3), γ-glutamyl-transferase, ubiquitin conjugating enzyme E2L6 (UCE2L6) and ferritin light chain that were detected by our prior yeast 2 hybrid studies (41). While the precise impact of these potential interactions on cellular processes, including regulation of pro-survival mechanisms independently of NMN synthesis, remains poorly understood, reductions in intracellular NAMPT protein levels would be predicted to alter the magnitude of these interactions and thus alter EC susceptibility to cell death.
Our studies indicate very strong linkages between reductions in NAMPT expression and reduced eNAMPT secretion and, consequently, reduced autocrine/paracrine effects on cytokine- mediated apoptosis. NAMPT overexpression leads to increased eNAMPT secretion and enhanced autocrine/paracrine effects shown to be relevant to IL-1β-induced articular chondrocytes secretion of pro-inflammatory cytokines (IL-6, TNF-α) (16), by hepatocytes with increased NAMPT gene expression (14, 42, 43) and by transgenic mice overexpressing NAMPT with observed reductions in cardiac ischemia/reperfusion injury (44). Our proposed model adds clarity to observed increases in IL-8 secretion in TNF-α-challenged alveolar epithelial cell overexpressing NAMPT (45), reduced permeability in IL-1β-stimulated alveolar epithelium with reduced NAMPT expression (46) and reduced inflammation in human pulmonary microvascular EC infected with H1N1 influenza strain (31).
We also observed that inhibition of NMN synthesis failed to alter EC caspase-3 activation whereas NAMPT silencing in TNF-α-challenged ECs increased caspase-3 activation and PARP-1 cleavage compared to controls. FK866 is widely used as an anti-tumor agent (5) as NAD is a cofactor/substrate in PARP1-dependent DNA repair, and NAD requirements are increased in cancer cells repairing unstable genomic DNA. The multipotent mouse fibroblast cell line, C3H10T1/2, and the omnipotent pre-osteoblast cell line, MC3T3-E1, express higher levels of iNAMPT during osteogenic differentiation (47) and reduced NAMPT expression in C3H10T1/2 cells enhances adipogenesis and reduces osteogenesis (48). In cultured EC progenitor cells (EPCs), exogenous NAMPT significantly increased the translocation of NFκB in nuclei with subsequent impairment of migration and adhesion capacity (49).
Our ELISA-derived data validated that eNAMPT secretion into EC cultured medium is significantly reduced in NAMPT-silenced ECs and increased in NAMPT-overexpressing ECs. We have shown previously that the eNAMPT receptor is TLR4 with robust downstream NFκB signaling (30) resulting in altered expression of multiple targets involved in inflammation, oxidative stress, regulation of cell death and survival pathways (50). Secreted eNAMPT, therefore, may exert an overall protective survival effect in cells exposed to pro-apoptotic stimuli such as serum withdrawal, cytokines (TNF-α), LPS, hypoxia or excessive mechanical stress (25). Reductions in secreted levels of eNAMPT (after NAMPT silencing) interferes with such autocrine or paracrine effects, reducing cellular resistance to subsequent cellular stress. The capacity of conditioned medium from control EC (but not from NAMPT-silenced EC), supplemented to silenced EC, to restore the cellular defenses against cytokine challenge supports this underlying hypothesis. While it is possible that reduced NAMPT expression (due to silencing) indirectly reduces expression and/or secretion of cytokines that impact cell survival due to lowered levels of NMN synthesis, our studies, employing a eNAMPT neutralizing antibody to reduce TLR4 ligation, verified that secreted eNAMPT is responsible for reduced EC apoptotic responses to TNF-α.
In summary, we have identified a distinct and novel role for eNAMPT in regulating EC survival and apoptosis with intracellular NAMPT-mediated Sublingual NMN synthesis uninvolved in regulation of TNF-α-induced EC apoptosis. Secretion of eNAMPT and subsequent ligation of the NAMPT receptor, TLR4, provides a novel mechanism for NAMPT in activating pro-survival mechanisms in endothelium as a stress-response mechanism during inflammatory lung injury.
Figure 1. NAMPT expression modulates EC susceptibility to TNF-α-induced apoptosis. (A) EC transfection with NAMPT gene siRNA (72 hrs) reduced NAMPT expression and resulted in marked increases in full length PARP-1 expression (compared to transfection with scrambled siRNAs) in both vehicle- and TNF-α-challenged ECs (30 ng/ml, 5 hrs). TNF-α challenge of NAMPT-silenced ECs markedly increased PARP-1 cleavage proteolysis, whereas scrambled siRNA-treated EC exhibited minimal increases in PARP-1 cleavage. *, p<0.05. (B) ECs were pre- treated (6 hrs) with either vehicle (DMSO) or with a caspase-3 inhibitor (z-DEVD-fmk, 20 μM) prior to 24 hr challenge with TNF-α (30 ng/ml), FasL (100 ng/ml) or tBHP (5 mM). Inhibition of caspase-3 activity significantly reduced subsequent PARP-1 cleavage induced by TNF-α. (C) ECs were transfected with myc-tagged NAMPT plasmid resulting in marked NAMPT overexpression. Compared with EC transfected with myc-tag only, EC overexpressing NAMPT exhibited virtual abolishment of PARP cleavage after TNF-α challenge (30 ng/ml or 100 ng/ml, 24 hrs). Error bars indicate SEM of three independent experiments.
Figure 2. Pharmacological inhibition of EC NAMPT and reduced NMN synthesis fails to alter TNF-α- induced PARP cleavage. (A) ECs were pre-treated (24 hrs) with FK866 (10 μM) or vehicle (DMSO), followed by TNF-α challenge (100 ng/ml, 24 hrs). FK866 failed to produce significant differences in the levels of TNF-α-induced PARP-1 cleavage. (B) Comparisons of TNF- α-induced caspase-3 activity (colorimetric assay) in ECs silenced for NAMPT (72 hrs) or treated with FK866 (10 μM, 24 hrs). Only NAMPT silencing (siRNA) significantly increased TNF-α- induced caspase-3 activity compared to controls, with FK866 failing to alter caspase-3 activation upon subsequent TNF-α challenge compared to DMSO. *, p<0.05. Error bars indicate SEM of three independent experiments.
Figure 3. ECs secrete eNAMPT with NAMPT gene silencing reducing subsequent levels of eNAMPT secretion. (A) ECs were silenced for NAMPT (72 hrs) followed by sampling of supernatant (culture media). Secreted eNAMPT was enriched by IP and detected by western blot. eNAMPT levels in NAMPT-silenced media were reduced compared to silenced-controls and untransfected cells. (B) ELISA quantitative measurement of NAMPT levels in EC supernatants. Shown is the reduction in NAMPT in conditioned media (EC supernatant) derived from NAMPT- silenced EC compared to the unchallenged EC and control-silenced EC. *, p<0.05. Error bars indicate SEM of three independent experiments.
Figure 4. eNAMPT is responsible for TNF-α-induced cell death. (A) Schematic representation of the conditioned medium (CM) strategy. Donor CM was obtained from either control-silenced EC or EC silenced for NAMPT (24 hrs) (donor group), collected after 48 hrs. Figure 3B depicts the reduced NAMPT levels in CM from NAMPT silenced EC. CM was prepared by 50% conditioned medium (donor group) and 50% fresh EGM-2 medium to treat another set of ECs (acceptor group), which have already silenced for NAMPT expression (24 hrs). After 48 hrs of continuous incubation with CM, the acceptor group ECs were further treated with TNF-α for another 24 hrs. (B) Supplementation of the CM from NAMPT-silenced ECs significantly blunted the protection offered by conditioned media against TNF-α in NAMPT-silenced EC (100 ng/ml, 24 hr)-activated caspase- 3 activity compared to the control conditioned media. *, p<0.05. (C) Protection against TNF-α- activated caspase-3 activity in NAMPT-silenced EC induced by the conditioned media from control cells was attenuated by a goat polyclonal anti-NAMPT (1 μg/ml), but not by co-incubation with normal, control goat IgGs. (D) Incubation with conditioned media from ECs transfected with NAMPT-HA reduced the activation of caspase-3 by TNF-α (100 ng/ml, 24 hr) and the effect was significantly impaired by pre-incubation with LPS-RS (1 μg/ml). Co-incubation of TNF-α with a monoclonal anti-TNF-α antibody with neutralizing ability served as a positive control. *, p<0.05. Error bars indicate SEM of three independent experiments.
Figure 5. rhNAMPT attenuates TNF-α induced apoptosis markers. (A). ECs were pre- treated with rhNAMPT (100 ng/ml, 24 hrs) then challenged TNF-α (100 ng/ml, 24 hrs). ECs were next fixed and stained for the cleaved fragment of PARP-1 (89 kDa). EC pre-treated with rhNAMPT and TNF-α-challenged showed reduced fluorescent signal (AF488) compared to EC without rhNAMPT pre-treatment. (B) ECs were pre-treated with rhNAMPT (100 ng/ml, 24 hrs) followed by exposure to TNF-α (100 ng/ml, 24 hrs). Cells were assayed by TUNEL method and showed a reduction in the fluorescein signal in the samples pre-treated with rhNAMPT followed by TNF challenge. There were no differences between control and rhNAMPT-treated EC. *, p<0.05. Error bars indicate SEM of three independent experiments.