Structure-functional changes in eNAMPT at high concentrations mediate mouse and human beta cell dysfunction in type 2 diabetes
Progressive decline in functional beta cell mass is central to the development of type 2 diabetes. Elevated serum levels of extracellular nicotinamide phosphoribosyltransferase (eNAMPT) are associated with beta cell failure in type 2 diabetes and eNAMPT immuno-neutralisation improves glucose tolerance in mouse models of diabetes. Despite this, the effects of eNAMPT on functional beta cell mass are poorly elucidated, with some studies having separately reported beta cell-protective effects of eNAMPT. eNAMPT exists in structurally and functionally distinct monomeric and dimeric forms. Dimerisation is essential for the NAD-biosynthetic capacity of NAMPT. Monomeric eNAMPT does not possess NAD-biosynthetic capacity and may exert distinct NAD-independent effects. This study aimed to fully characterise the structure-functional effects of eNAMPT on pancreatic beta cell functional mass and to relate these to beta cell failure in type 2 diabetes.
CD-1 mice and serum from obese humans who were without diabetes, with impaired fasting glucose (IFG) or with type 2 diabetes (from the Body Fat, Surgery and Hormone [BodyFatS&H] study) or with or at risk of developing type 2 diabetes (from the VaSera trial) were used in this study. We generated recombinant wild-type and monomeric eNAMPT to explore the effects of eNAMPT on functional beta cell mass in isolated mouse and human islets. Beta cell function was determined by static and dynamic insulin secretion and intracellular calcium microfluorimetry. NAD-biosynthetic capacity of eNAMPT was assessed by colorimetric and fluorescent assays and by native mass spectrometry. Islet cell number was determined by immunohistochemical staining for insulin, glucagon and somatostatin, with islet apoptosis determined by caspase 3/7 activity. Markers of inflammation and beta cell identity were determined by quantitative reverse transcription PCR. Total, monomeric and dimeric eNAMPT and nicotinamide mononucleotide (NMN) were evaluated by ELISA, western blot and fluorometric assay using serum from non-diabetic, glucose intolerant and type 2 diabetic individuals.
eNAMPT exerts bimodal and concentration- and structure-functional-dependent effects on beta cell functional mass. At low physiological concentrations (~1 ng/ml), as seen in serum from humans without diabetes, eNAMPT enhances beta cell function through NAD-dependent mechanisms, consistent with eNAMPT being present as a dimer. However, as eNAMPT concentrations rise to ~5 ng/ml, as in type 2 diabetes, eNAMPT begins to adopt a monomeric form and mediates beta cell dysfunction, reduced beta cell identity and number, increased alpha cell number and increased apoptosis, through NAD-independent proinflammatory mechanisms.
We have characterised a novel mechanism of beta cell dysfunction in type 2 diabetes. At low physiological levels, eNAMPT exists in dimer form and maintains beta cell function and identity through NAD-dependent mechanisms. However, as eNAMPT levels rise, as in type 2 diabetes, structure-functional changes occur resulting in marked elevation of monomeric eNAMPT, which induces a diabetic phenotype in pancreatic islets. Strategies to selectively target monomeric eNAMPT could represent promising therapeutic strategies for the treatment of type 2 diabetes.
KeywordsBeta cell eNAMPT Extracellularnicotinamide phosphoribosyltransferase Inflammation Insulin secretion NAD Type 2 diabetes
Extracellular nicotinamide phosphoribosyltransferase
Recombinant wild-type extracellular nicotinamide phosphoribosyltransferase
Fasting plasma glucose
Glucose-stimulated insulin secretion
Impaired fasting glucose
Intracellular nicotinamide phosphoribosyltransferase
c-Jun N-terminal kinase
Mitogen-activated protein kinase
Signal transducer and activator of transcription 3
Elucidating the underlying mechanisms responsible for progressive decline in functional beta cell mass is essential for the design of novel treatments for type 2 diabetes . The protein nicotinamide phosphoribosyltransferase (NAMPT) exists in intracellular (iNAMPT) and extracellular (eNAMPT) forms. eNAMPT (also known as visfatin/pre-B cell enhancing factor [PBEF]) is a circulating protein with several reported functions [2, 3, 4]. Similar to iNAMPT, eNAMPT may exert NAD-biosynthetic effects, whereby eNAMPT catalyses the conversion of nicotinamide to nicotinamide mononucleotide (NMN) in serum [5, 6]. NMN may subsequently be transported into the cell (directly or via conversion into nicotinamide riboside), where it is converted into NAD by nicotinamide mononucleotide adenylyltransferases (NMNATs) 1–3. However, this mechanism has been disputed and NAD-independent functions, including eNAMPT-mediated proinflammatory effects, are also reported [4, 7, 8, 9].
Serum eNAMPT levels are elevated in type 2 diabetes, in association with declining beta cell function. Our in vivo studies indicate that eNAMPT may mediate type 2 diabetes pathophysiology [10, 11, 12]. Contrasting studies report that eNAMPT and its reaction product NMN exert beta cell-protective effects by boosting cellular NAD levels [6, 13, 14, 15, 16, 17, 18, 19].
A key, but frequently overlooked factor governing eNAMPT function is the presence of structurally and functionally distinct monomeric (~52 KDa) and dimeric (~104 KDa) eNAMPT. Dimerisation is essential for NAD-biosynthetic capacity of eNAMPT [6, 20] and the eNAMPT dimer predominates in normal physiology [12, 15]. The eNAMPT monomer does not possess NAD-biosynthetic capacity and may mediate NAD-independent proinflammatory effects. Our studies suggested a specific pathophysiological role for monomeric eNAMPT in mouse models of diabetes . We hypothesised that, in type 2 diabetes, eNAMPT monomer is selectively elevated, which exerts NAD-independent proinflammatory effects, and that this, combined with the loss of eNAMPT’s NAD-biosynthetic capacity, plays a key role in disease pathophysiology.
However, whether eNAMPT monomer is elevated in type 2 diabetes and to what extent the structurally distinct forms of eNAMPT directly affect the pancreatic beta cell are unknown. Previous studies, including our own, have examined NAD-boosting effects of NMN on beta cell function [6, 13, 14], or have examined supraphysiological concentrations or acute effects of eNAMPT [21, 22], neither of which accurately mimic in vivo (patho)physiology. Of perhaps greater importance, the selective effects of monomeric and dimeric eNAMPT on beta cell health have not been examined.
We used isolated mouse and human islets, and serum from humans with type 2 diabetes, to specifically characterise the effects of monomeric and dimeric eNAMPT on pancreatic beta cell functional mass.
Eight-week-old male CD-1 mice (28–33 g; Envigo, Blackthorn, UK) were housed in 12 h light/dark cycle, temperature-controlled conditions with ad libitum access to standard mouse chow (Envigo) and water. Procedures were performed in accordance with UK Home Office regulations (Animal Scientific Procedures Act, 1986).
Human serum was obtained from obese individuals without diabetes, (BMI >30 kg/m2; fasting plasma glucose [FPG] <5.6 mmol/l; n = 13 individuals), with impaired fasting glucose (IFG; BMI >30 kg/m2; FPG 5.6–6.9 mmol/l; n = 15 individuals)) or with type 2 diabetes (BMI >30 kg/m2; FPG >7.0 mmol/l; n = 27 individuals), as part of the Body Fat, Surgery and Hormone (BodyFatS&H) study (University of Auckland, Auckland, New Zealand) and from obese individuals with or at risk of developing type 2 diabetes as part of the VaSera trial (St Thomas’ Hospital, London) [23, 24]. Study participants gave informed consent. Investigations were approved by the Northern Regional Ethics Committee (NREC), Auckland, New Zealand (BodyFatS&H) and the National Research Ethics Service (NRES) Committee, London Central (VaSera). See electronic supplementary material (ESM) Methods for further details.
Native mass spectrometry
Native mass spectrometry was conducted using the Synapt G2S high definition mass spectrometer (HDMS; Waters, Manchester, UK) (see ESM Methods).
eNAMPT protein generation
Recombinant wild-type (WT) eNAMPT (eNAMPT-WT) and mutant SS199/200DD (eNAMPT monomer) were produced in Escherichia coli. See ESM Methods for details.
Pancreatic islet isolation
Mouse islets were isolated between 9:00 and 10:00 hours . Human islets were isolated from heart-beating non-diabetic donors, with ethical approval, at the King’s College Hospital Human Islet Isolation Unit (London, UK)  (see ESM Human Islets Checklist for further details). All isolated islets were incubated overnight (37°C, 5% CO2) prior to treatments (ESM Methods). RPMI media was used for all incubations. Islets were size-matched prior to treatments. Mouse and human islets were treated with eNAMPT (WT or monomer; 0.1–5 ng/ml) or NMN (100 μmol/l) for 24, 48 or 72 h. For inflammatory pathway inhibition, mouse islets were incubated with 1 ng/ml eNAMPT monomer or 5 ng/ml eNAMPT-WT, in combination with inhibitors of signal transducer and activator of transcription 3 (STAT3; using NCS74859; 50 μmol/l), P38-mitogen-activated protein kinase (MAPK; using SB203580; 1 μmol/l), NF-κB (using BAY 11-7082; 1 μmol/l) or c-Jun N-terminal kinase (JNK; using SP600125; 50 μmol/l), (Tocris Biosciences, Abingdon, UK). For CD73 inhibition, mouse islets were incubated with 0.5–5 ng/ml eNAMPT-WT for 48 h in combination with adenosine 5′-(α,β-methylene)diphosphate (AMP-CP; 1 μmol/l; Sigma-Aldrich, Poole, UK)
Static and dynamic glucose-stimulated insulin secretion
Static or dynamic insulin secretion was assessed in response to 2 mmol/l or 20 mmol/l glucose exposure, as previously described . Secreted insulin was measured by in-house 125I radioimmunoassay  (see ESM Methods).
Endotoxin concentrations were measured in recombinant eNAMPT preparations using the Pierce LAL Chromogenic Endotoxin Quantitation Kit (ThermoFisher Scientific, Altrincham, UK).
In vitro islet immunostaining
Mouse islets were pelleted and fixed in 4% (vol./vol.) buffered formalin. Sections (5 μmol/l) were incubated with anti-guinea pig insulin (Dako, Stockport, UK; catalogue no. A0564), anti-rabbit glucagon (Abcam, Cambridge, UK; ab92517) or anti-rat somatostatin (Abcam; ab30788) (1:100 for all; K) and DAPI (ThermoFisher). Secondary antibodies used were AlexaFluor 488-labelled donkey anti-rat (712-545-1250), AlexaFluor 594-labelled donkey anti-guinea pig (06-585-148) and AlexaFluor 488-labelled donkey anti-rabbit (711-545-152) (1:200 for all; Jackson ImmunoResearch, Cambridge UK). Sections were mounted on glass slides and captured on a Nikon TE2000 fluorescent microscope (Nikon Instruments, Melville, NY, USA) and quantified using ImageJ (https://imagej.nih.gov/ij/download.html; accessed December 2017–June 2018).
Mouse islets were treated with eNAMPT with or without cytokine cocktail (0.05 U/μl IL-1β, 1 U/μl TNF-α and 1 U/μl IFNγ; 48 h). Apoptosis was determined by the Caspase-Glo 3/7 luminescent assay (Promega, Southampton, UK).
Quantitative reverse transcription PCR
Gene expression was determined by SYBR Green quantitative reverse transcription PCR using ΔΔCt methodology and normalised against glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (QuantiTect; Qiagen, Manchester, UK). See ESM Methods and ESM Table 1.
Whole mouse islets, incubated in HEPES bicarbonate buffer (containing 2–20 mmol/l glucose), were loaded with Fura-2 AM and imaged using a Nikon Ti-E microscope equipped with FuraLEDs (Cairn Research, Faversham, UK) (excitation, 340/385 nm; emission, 470–550 nm).
MIN6 cell culture
MIN6 cells (mycoplasma free; obtained from J. I. Miyazaki, Osaka University, Osaka, Japan [27, 28]) were cultured at 37°C in DMEM (25 mmol/l glucose; 2 mmol/l glutamine, 10% FBS (vol./vol.), 100 U/ml penicillin, 100 μg/ml streptomycin) and treated with eNAMPT (1–5 ng/ml WT; 1 ng/ml monomer) or NMN (100 μmol/l) (48 h; Sigma-Aldrich), prior to NAD and NMN measurements.
eNAMPT, NAD and NMN measurements
Total serum eNAMPT was measured by ELISA (Adipogen, Seoul, South Korea). MIN6 intracellular NAD and NMN levels were measured by an NAD/NADH Quantification Kit, (Sigma-Aldrich) and an in-house fluorometric assay, respectively [29, 30] (see ESM Methods).
Serum eNAMPT monomer and dimer were measured by immunoblotting, under non-reducing conditions, as previously described . To ensure equivalent loading of each lane, serum protein was normalised to 10 μg and an equal volume (30 μl) was added to each well (see ESM Methods).
Data are expressed as mean ± SEM. Significance was tested using one- or two-way ANOVA with Tukey’s or Sidak’s post hoc test, using GraphPad PRISM 7 software (San Diego, CA, USA). Experimenters were not blind to group assignment and outcome assessment.
Serum eNAMPT monomer concentrations are elevated in type 2 diabetes and are positively correlated with increased HbA1c
eNAMPT exerts bimodal effects on beta cell function in mouse and human pancreatic islets
Increased HbA1c is indicative of poor glycaemic control and deteriorating beta cell function. To assess the (patho)physiological effects of eNAMPT on beta cells, we used serum eNAMPT concentrations (as detected in Fig. 1a and ) and exposed isolated islets to physiological (1 ng/ml) and pathophysiological (5 ng/ml) eNAMPT. To account for batch-to-batch differences between commercially available eNAMPT , we generated recombinant wild-type eNAMPT (eNAMPT-WT). To specifically examine eNAMPT monomer, we mutated Ser199/200 (key residues in the NAMPT binding interface [6, 20]) to aspartic acid. This generated a 52 KDa protein (eNAMPT monomer), which neither dimerised correctly, as demonstrated by size-exclusion chromatography and native MS, nor synthesised cellular NMN and NAD (ESM Fig. 2a–e). Neither eNAMPT-WT nor eNAMPT monomer contained notable endotoxin concentrations (ESM Fig. 2f)
We also assessed effects of a combination of eNAMPT-WT and eNAMPT monomer, which is more representative of normal (patho)physiology. eNAMPT-WT (1 ng/ml) alone induced GSIS as expected; however, when eNAMPT-WT (1 ng/ml) was added in combination with eNAMPT monomer, the insulin stimulatory effects of eNAMPT-WT were lost (Fig. 2n).
Together this demonstrates that low, physiological levels of eNAMPT enhance beta cell function, but as eNAMPT rises to pathophysiological levels (~5 ng/ml), beneficial effects are lost and eNAMPT mediates beta cell dysfunction. The effects of high concentrations of eNAMPT-WT are mimicked by non-dimerising eNAMPT monomer, consistent with the notion that eNAMPT-WT adopts monomeric structure and function at higher concentrations.
Pathophysiological eNAMPT reduces beta cell number and identity and increases alpha cell number
eNAMPT impairs beta cell function via proinflammatory mechanisms
Therefore, pathophysiological concentrations of eNAMPT-WT and eNAMPT monomer induce beta cell dysfunction, in part through promotion of STAT3 and P38-mediated islet inflammation.
Structure-functional changes mediate bimodal effects of eNAMPT-WT
Since the effects of eNAMPT monomer were mimicked by 5 ng/ml, but not 1 ng/ml eNAMPT-WT, we hypothesised that high concentrations of eNAMPT-WT (5 ng/ml) would be associated with reduced NAD-biosynthetic capacity (compared with 1 ng/ml eNAMPT-WT), indicative of reduced eNAMPT dimer and increased eNAMPT monomer levels.
These data provide evidence that the bimodal effects of eNAMPT-WT are related to loss of NAD-biosynthetic capacity of eNAMPT at higher concentrations.
This study demonstrates that eNAMPT exerts bimodal, concentration- and structure-functional-dependent effects on pancreatic beta cell functional mass, and provides clarification of previous contradictory eNAMPT studies.
Low, physiological levels of eNAMPT (1 ng/ml) enhance beta cell health through NAD-dependent mechanisms. Since islets are continually exposed to low eNAMPT concentrations in healthy non-diabetic individuals, this indicates a role for low eNAMPT concentrations in maintaining beta cell health. Beta cell identity genes were increased at 1 ng/ml eNAMPT, highlighting the potential importance of physiological levels of eNAMPT in maintaining beta cells in the fully differentiated state. This is consistent with studies showing beta cell de-differentiation following 3-day culture in the absence of circulating factors, including eNAMPT/NMN . The effects of 1 ng/ml eNAMPT are mediated via NAD biosynthesis, suggesting eNAMPT exists in the dimeric form at lower concentrations. This is consistent with our observations in non-diabetic human and mouse serum samples, which show eNAMPT dimer as the predominant structural form of eNAMPT at low physiological concentrations . Whether eNAMPT can exert NAD biosynthesis within the extracellular space remains controversial. Whilst RPMI media contains sufficient nicotinamide to support eNAMPT enzymatic activity, we were unable to detect eNAMPT within spent islet media after eNAMPT incubation, suggesting that at least some eNAMPT dimer is taken up into the islet/beta cell, after which NMN synthesis occurs intracellularly.
Mechanistically, eNAMPT-mediated NAD synthesis may improve beta cell health through activation of NAD-dependent sirtuins or via actions of NAD metabolites cyclic ADP-ribose (cADPR) and nicotinic acid adenine dinucleotide phosphate (NAADP), which reportedly induce intracellular calcium mobilisation [13, 14, 16, 19].
Conversely, as eNAMPT concentrations rise to pathophysiological levels, as in type 2 diabetes, eNAMPT induces islet inflammation, and initiates beta cell dysfunction, beta cell death and reduced beta cell mass and identity. The mechanisms driving reduced beta cell number and identity are unclear. We have demonstrated that high concentrations of eNAMPT enhance beta cell apoptosis, however beta to alpha cell trans-differentiation or beta cell de-differentiation may also play a role, and further studies are required to identify the precise mechanisms.
eNAMPT exerts beta cell dysfunction through NAD-independent mechanisms, suggesting a switch toward monomeric eNAMPT at higher concentrations. This is consistent with observations that the eNAMPT monomer was present at 1–2 ng/ml in serum from individuals with type 2 diabetes, but almost absent in serum from individuals without diabetes. It is also consistent with our observations that increasing serum eNAMPT did not correlate with increasing serum NMN.
Whilst elevated eNAMPT strongly correlated with increased HbA1c, we did not observe notable correlations between eNAMPT and BMI, and only a weak correlation between eNAMPT and serum insulin. We hypothesise that, as glucose rises during glucose intolerance, secretion of eNAMPT dimer increases, which promotes islet compensation. However, as eNAMPT levels continue to rise in response to elevated glucose, eNAMPT dimer begins to break apart into constituent eNAMPT monomers, which then drive beta cell failure. Thus, increases in eNAMPT strongly correlate with increases in blood glucose and HbA1c and do not necessarily precede hyperglycaemia. Precisely why eNAMPT preferentially exists as a monomer at high levels remains unclear but may relate to ligand-induced dimerisation, whereby eNAMPT is stabilised and maintained in dimer form by the presence of an endogenous ligand. We hypothesise that insufficient concentrations of the putative ligand at higher eNAMPT concentrations lead to break up of the dimer into constitutive monomers.
Mechanistically, eNAMPT promoted islet inflammation and beta cell failure through P38-MAPK and STAT3 pathways. Similar eNAMPT effects are reported in macrophages, monocytes and islets of mice fed a high-fat diet (HFD) [7, 9, 12, 42]. Upstream, these effects may be receptor-mediated, with eNAMPT reported to function via Toll-like receptor 4 (TLR4) , C-C chemokine receptor type 5 (CCR5) , IGF and insulin receptor  signalling, although these remain unconfirmed. Separately, the deleterious effects of eNAMPT monomer may also reflect decreased eNAMPT dimer-mediated NAD synthesis, occurring independently from direct effects of eNAMPT monomer.
Together, our results characterise structure-functional relationships as being of crucial importance to the effects of eNAMPT on beta cell health. Moreover, we demonstrate a novel mechanism of beta cell dysfunction in type 2 diabetes. Strategies to block the actions of the eNAMPT monomer by promoting dimerisation or stabilising eNAMPT in dimer form could represent promising therapeutic approaches for the treatment of diabetes.
We thank J. I. Miyazaki (Osaka University, Osaka, Japan) for provision and consent to use MIN6 cells. Imaging was conducted in the Nikon Imaging Centre at King’s College London (London, UK).
SRS, GCH, PC, RLB, NHFF, KJP, PEB, SB, DJH, GGL and PWC designed and conducted the research, analysed the data and reviewed and approved the final manuscript. A-TM, MPS and SDP designed, conducted and analysed the Body Fat, Surgery and Hormone (BodyFatS&H) study and reviewed and approved the final manuscript; CEM and JKC designed, conducted and analysed the VaSera study and reviewed and approved the final manuscript. SRS and PWC wrote the paper. PWC is the guarantor of this work.
PWC and SRS were supported by Diabetes UK project grants (15/0005154; 18/0005865), King’s Health Partners Challenge Fund (funded via MRC Confidence-in-concept), Society for Endocrinology Early Career Grant and EFSD/Lilly Fellowship. DJH was supported by a Diabetes UK R.D. Lawrence (12/0004431) Fellowship, and MRC (MR/N00275X/1) and Diabetes UK (17/0005681) Project Grants. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Starting Grant 715884 to DJH), and New Zealand Health Research Council (HRC) programme grant (190/C; SDP and A-TM). MPS was supported by University of Auckland FRDF fellowship and HRC project grant (14/191). GGL was supported by a Wellcome Trust Senior Fellowship. The VaSera trial was supported by FUKUDA Denshi Japan. The study funders were not involved in the design of the study; the collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
- 24.Faconti L, Mills CE, Govoni V et al (2019) Cardiac effects of 6 months’ dietary nitrate and spironolactone in patients with hypertension and with/at risk of type 2 diabetes, in the factorial design, double-blind, randomized controlled VaSera trial. Br J Clin Pharmacol 85(1):169–180PubMedCrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.