Thioredoxin-interacting protein is required for endothelial NLRP3 inflammasome activation and cell death in a rat model of high-fat diet
- 2.4k Downloads
Obesity and hypertension, known pro-inflammatory states, are identified determinants for increased retinal microvascular abnormalities. However, the molecular link between inflammation and microvascular degeneration remains elusive. Thioredoxin-interacting protein (TXNIP) is recognised as an activator of the NOD-like receptor pyrin domain containing-3 (NLRP3) inflammasome. This study aims to examine TXNIP expression and elucidate its role in endothelial inflammasome activation and retinal lesions.
Spontaneously hypertensive (SHR) and control Wistar (W) rats were compared with groups fed a high-fat diet (HFD) (W+F and SHR+F) for 8–10 weeks.
Compared with W controls, HFD alone or in combination with hypertension significantly induced formation of acellular capillaries, a hallmark of retinal ischaemic lesions. These effects were accompanied by significant increases in lipid peroxidation, nitrotyrosine and expression of TXNIP, nuclear factor κB, TNF-α and IL-1β. HFD significantly increased interaction of TXNIP–NLRP3 and expression of cleaved caspase-1 and cleaved IL-1β. Immunolocalisation studies identified TXNIP expression within astrocytes and Müller cells surrounding retinal endothelial cells. To model HFD in vitro, human retinal endothelial (HRE) cells were stimulated with 400 μmol/l palmitate coupled to BSA (Pal-BSA). Pal-BSA triggered expression of TXNIP and its interaction with NLRP3, resulting in activation of caspase-1 and IL-1β in HRE cells. Silencing Txnip expression in HRE cells abolished Pal-BSA-mediated cleaved IL-1β release into medium and cell death, evident by decreases in cleaved caspase-3 expression and the proportion of live to dead cells.
These findings provide the first evidence for enhanced TXNIP expression in hypertension and HFD-induced retinal oxidative/inflammatory response and suggest that TXNIP is required for HFD-mediated activation of the NLRP3 inflammasome and the release of IL-1β in endothelial cells.
KeywordsCaspase-1 High-fat diet Hypertension IL-1β Inflammasome Inflammation NLRP3 Obesity Oxidative stress Retinal acellular capillaries TXNIP
Association for Research in Vision and Ophthalmology
Ganglion cell layer
Glial fibrillary acidic protein
Green fluorescent protein
Human retinal endothelial
Inner nuclear layer
Inner plexiform layer
Nuclear factor κB
NOD-like receptor CARD containing-4
NOD-like receptor pyrin domain containing-1
NOD-like receptor pyrin domain containing-3
Outer nuclear layer
Palmitate coupled to BSA
Relative optical densitometry
Reactive oxygen species
Spontaneously hypertensive rats
SHR fed HFD
Small interfering RNA
Wistar control group rats
Wistar control rats fed HFD
Obesity and hypertension, the hallmarks of metabolic syndrome characterised by insulin resistance, are identified as independent risk factors for the development of retinopathy [1, 2]. Population studies have established obesity and hypertension as determinants for increased incidence of retinopathy and retinal microvascular abnormalities in the general non-diabetic population [3, 4]. Increased energy intake and consumption of saturated/trans-fat and cholesterol are the most plausible reasons behind the alarming epidemic rates of obesity and metabolic syndrome . High levels of plasma NEFA, and the saturated fatty acid palmitate in particular, are the result of high fat intake [6, 7] and contribute to metabolic syndrome-associated inflammation and insulin resistance [8, 9]. Palmitate has been shown to induce the activation of the NOD-like receptor pyrin domain containing-3 (NLRP3) inflammasome [6, 10, 11], a well-established multi-protein complex responsible for instigating obesity-induced inflammation [12, 13]. Activated NLRP3 oligomerises with the ASC (apoptosis-associated specklike) adaptor protein, which recruits procaspase-1, allowing its autocleavage and activation. Activated caspase-1 (cleaved caspase-1) enzyme in turn cleaves upregulated premature pro-inflammatory cytokines IL-1β and IL-18 and causes their release .
Thioredoxin-interacting protein (TXNIP) is the endogenous inhibitor and regulator of thioredoxin, a major cellular antioxidant and anti-apoptotic system . TXNIP expression, a well-established mediator of insulin resistance, has been shown to be consistently elevated in the muscles of insulin-resistant and diabetic individuals [16, 17]. In THP-1 human macrophages, inflammasome activators induce the dissociation of TXNIP from thioredoxin in a reactive oxygen species (ROS)-sensitive manner, allowing it to bind and activate NLRP3 . We and others have demonstrated a pivotal role of enhanced retinal TXNIP expression in the induction of the expression of pro-inflammatory cytokines, including IL-1β, ICAM-1 (intercellular adhesion molecule-1) and TNF-α, in vivo and in isolated retinal cultures [19, 20, 21, 22]. Increased levels of IL-1β were shown to increase the number of retinal acellular capillaries in vivo . Nevertheless, there is a gap in knowledge concerning whether TXNIP can mediate high-fat diet (HFD)-induced retinal inflammasome activation and whether this can occur directly within retinal endothelial cells to trigger inflammation and microvascular degeneration. Here, we test the hypothesis that HFD, as an essential component of the metabolic syndrome, results in upregulation of retinal TXNIP expression, NLRP3-inflammasome activation and microvascular degeneration in vivo and in vitro.
All animal studies were in accordance with the Association for Research in Vision and Ophthalmology (ARVO) and the Georgia Reagents University Animal Care and Use Committee. Retinas were obtained from two separate animal studies that examined the effect of HFD, either alone or in combination with hypertension. In the first study, 8-week old male Wistar Kyoto control rats (W) and spontaneously hypertensive rats (SHR) (Charles River, Wilmington, MA, USA) were randomly assigned to four groups: W and SHR control groups, fed ad libitum with normal rat chow (7% fat) for 10 weeks; W+F and SHR+F groups, fed with HFD (36 g %, 251 kJ [60 kcal] % fat, F2685 Bioserv [Frenchtown, NJ, USA]) for 10 weeks. Rats were weighed weekly and metabolic variables, plasma insulin and blood glucose were measured and published previously . In the second study, retinas were obtained from sham control 5-week-old male Wistar control rats (W) (Harlan Laboratories, Indianapolis, IN, USA) fed ad libitum for 8 weeks with an isoenergetic control diet (10% fat), and compared with retinas from W group rats fed with HFD (24 g %, 188 kJ [45 kcal] % fat, D12451 Research Diets [New Brunswick, NJ, USA]) (W+F) for 8 weeks. Rats were weighed weekly and metabolic variables were examined and published recently . Of note, in both studies, HFD resulted in significant increases in total body weight as an indicator for obesity. HFD induced increases in plasma cholesterol levels in the first study  but not in the second study  and there was no effect on blood pressure, blood glucose levels or insulin resistance in either study.
Determination of retinal acellular (degenerated) capillaries
Retinal vasculatures were isolated as described previously . Transparent vasculatures were laid out on slides and stained with periodic acid–Schiff and haematoxylin. Acellular capillaries were identified as capillary-sized blood-vessel tubes having no nuclei anywhere along their length. The number of acellular capillaries was averaged from eight different fields of the mid-retinal area and calculated as the average number/mm2 of retinal area using AxioObserver.Z1 Microscope (Zeiss North America, Thornwood, NY, USA).
Immunoprecipitation and western blot analysis
Retinas were lysed in modified radioimmunoprecipitation assay (RIPA) buffer (Millipore, Billerica, MA, USA) and 30 μg of total protein were separated by SDS-PAGE. Antibodies used were: anti-TXNIP (Santa Cruz Biotechnology, Dallas, TX, USA and Invitrogen, Carlsbad, CA, USA), anti-NFκB p65 (Santa Cruz Biotechnology), Anti-TNF-α (Novus Biologicals, Littleton, CO, USA), Anti-caspase-1, Anti-NLRP3 (Enzo lifesciences, Farmingdale, NY, USA), Anti-cleaved caspase-3 (Cell Signaling, Danvers, MA, USA) and anti-IL-1β (Abcam, Cambridge, MA, USA). One-hundred microgrammes of total protein was immunoprecipitated with anti-TXNIP antibody (5 μg/ml) and incubated with A/G agarose beads overnight. Precipitated proteins were analysed by SDS-PAGE and blotted with primary antibodies (anti-TXNIP and anti-NLRP3). Band intensities were quantified using Alpa Innotech Fluorchem (Santa Clara, CA, USA) imaging and densitometry software and expressed as relative optical density (ROD).
Detection of lipid peroxides
Levels of lipid peroxides (malondialdehyde, MDA) were assayed using thiobarbituric acid reactive substances as described previously . The Bradford assay (Bio-Rad, Hercules, CA, USA) was performed to determine the protein concentration of retinal lysate. Lipid peroxides were expressed in μmol MDA/mg total protein.
Detection of retinal nitrative stress
Slot-blot analysis was used to measure nitrative stress marker, nitrotyrosine (NY) as described previously . Five microgrammes of retinal homogenate were immobilised onto a nitrocellulose membrane and reacted with polyclonal anti-NY (Calbiochem/EMD Bioscience, La Jolla, CA, USA) and ROD was calculated compared with controls.
Quantitative real-time PCR
The One-Step qRT-PCR kit (Invitrogen) was used to amplify 10 ng retinal mRNA as described previously . PCR primers (listed in electronic supplementary material [ESM] Table 1) were purchased from Integrated DNA Technologies (Coralville, IA, USA). Quantitative PCR was performed using a Realplex Master cycler (Eppendorf North America, Hauppauge, NY, USA). The expression of Txnip, Nlrp3, Nlrc4 and Casp1 was normalised to the 18S level and expressed relative to W control.
Optimal cutting temperature compound (OCT)-frozen sections of the eyes (10 μmol/l) were fixed using 2% paraformaldehyde and reacted with the primary antibody (1:200 dilution), including polyclonal anti-TXNIP (Santa Cruz Biotechnology), polyclonal anti-GFAP (Pierce Biotect, Rockford, IL, USA), monoclonal anti-GFAP, monoclonal anti-glutamine synthetase (Chemicon-Millipore, Billerica, MA, USA) or negative control at 4°C overnight, followed by Oregon-green-conjugated goat anti-rabbit antibody or Texas-red goat anti-mouse antibody (Invitrogen). Retinal vasculature was localised using isolectin-B4 (Invitrogen). Images were collected using an AxioObserver.Z1 Microscope (Zeiss North America).
Human retinal endothelial cell culture studies
All human retinal endothelial (HRE) cell studies were in accordance with the ARVO and the Charlie Norwood Veterans Affairs Medical Center, research and ethics committee. HRE cells and supplies were purchased from Cell Systems Corporations (Kirkland, WA, USA) and VEC Technology (Rensselaer, NY, USA) as described previously . Sodium palmitate (catalogue No. P9767; Sigma-Aldrich, St. Louis, MO, USA) was dissolved in 50% ethyl alcohol solution, then added drop-wise to pre-heated 10% endotoxin- and fatty acid-free BSA (catalogue No. A8806; Sigma) in M199 at 50°C to create an intermediate stock solution of palmitate coupled to BSA (Pal-BSA). Confluent cells were switched to serum-free medium for 6 h then treated for 12 h with Pal-BSA solutions in a ratio of 1:10 to produce final concentrations of 200, 400 and 800 μmol/l of Pal-BSA. Equal volumes of 50% ethyl alcohol solution without any palmitate dissolved in BSA served as a control (BSA alone). Peroxynitrite (PN) was purchased from Calbiochem and diluted in 100 mmol/l NaOH and added at a final concentration of 100 μmol/l.
Silencing of TXNIP expression
Transfection of HRE cells with 0.6 μmol/l Txnip small interfering RNA (siRNA) was performed using Amaxa nucleofector primary endothelial cells kit (Lonza, Germany) as described previously . In addition, a chemical-transfection kit was used according to the manufacturer’s protocol (Santa Cruz Biotechnology). HRE cells (80% confluent) were incubated in the conditioned transfection medium with 300 ng of FITC-labelled scrambled (SC) or Txnip siRNA for 6 h, then left to recover in complete medium for 24 h before experiments were performed. Transfection efficiency was 70–80% for both methods as indicated by the number of cells expressing green fluorescent protein (GFP) or FITC-labelled SC siRNA (data not shown). Silencing of TXNIP expression was verified by western blot analysis.
Determination of IL-1β release
Secretion of cleaved IL-1β into the HRE cell conditioned media was determined using IL-1β ELISA sensitive kit (R&D systems, Minneapolis, MN, USA). Briefly, equal volumes of conditioned media for each group were concentrated using Ambion10K concentration columns (Millipore, Temecula, CA, USA), then loaded into IL-1β capture antibody pre-coated wells and processed according to the manufacturer’s protocol. IL-1β concentrations were expressed as pg/ml of the cell conditioned media used.
Life and death cell viability assay
HRE cell viability was tested by a Live/Dead assay (Invitrogen) following the manufacturer’s protocol. A working assay solution of 4 μmol/l ethidium homodimer-1 and 2 μmol/l calcein-AM was prepared and was added to the top of each culture well for 15 min at 37°C in a humidified atmosphere (20% O2 with 5% CO2). The staining solution was removed and samples were then viewed under a Ziess fluorescence microscope with filters (494 nm, green, viable; 528 nm, red, non-viable) and the percentage was calculated.
Results were expressed as means ± SEM. Two-way ANOVA, followed by Bonferroni multiple comparison test were used for testing differences among all the multiple experimental groups and for testing the interaction between the types of diet (HFD vs normal diet) across the groups of rats (W control vs SHR group, for the animal experiments) and between the presence or absence of PN across palmitate-treated or control cell groups (for the in vitro studies). Two-sided Student’s t test was used for testing differences between two experimental groups (W and W+F). Significance was defined as p < 0.05.
HFD causes retinal microvascular degeneration and oxidative and inflammatory stress
HFD induces retinal TXNIP expression and NLRP3 inflammasome activation
HFD induces TXNIP–NLRP3 interaction associated with retinal inflammasome activation
HFD-induced TXNIP expression co-localises within retinal glia and microvasculature
Palmitate induces TXNIP and activates caspase-1/IL-1β in HRE cells
Palmitate is a commonly elevated saturated NEFA in plasma, with a well-established pro-inflammatory effect, compared with other unsaturated fatty acids [7, 30, 31]. In addition, palmitate is a major component (60%) of the saturated fatty acids both in the F2685-Bioserv and D12451-Research Diets. To test the hypothesis that HFD can directly activate the NLRP3 inflammasome in endothelial cells, we examined different levels of palmitate coupled to BSA (Pal-BSA 200 μmol/l, 400 μmol/l and 800 μmol/l) in HRE cells. Pal-BSA induced a bell-shaped dose–response curve (ESM Fig. 2), reaching a maximum response at 400 μmol/l, at which point TXNIP, cleaved caspase-1 and cleaved IL-1β expression was increased by 1.6-, 1.8- and 1.75-fold, respectively, compared with BSA controls. Of note, 800 μmol/l induced HRE cell toxicity; hence 400 μmol/l of Pal-BSA was used for the rest of the in vitro studies to probe the role of HFD in NLRP3 inflammasome activation in endothelial cells
TXNIP is required for palmitate-induced NLRP3 inflammasome activation in HRE cells
TXNIP is required for palmitate-induced IL-1β maturation in HRE cells
TXNIP is required for palmitate-induced apoptosis and cell death in HRE cells
Although clinical evidence implicates metabolic syndrome in increasing the risk for retinopathy [3, 4], data from experimental models are lacking. Therefore, we attempted to model and characterise the detrimental elements of the metabolic syndrome on the early development of retinal microvascular lesions. The current study produced several major findings: (1) HFD-induced obesity or hypertension, or their combination, resulted in early retinal microvascular lesions, significant increases in retinal TXNIP expression, oxidative stress and inflammation; (2) HFD selectively induces retinal TXNIP–NLRP3 interaction and inflammasome activation resulting in increased expression of cleaved caspase-1 and IL-1β; (3) palmitate-induced TXNIP expression is required for NLRP3 inflammasome activation in retinal endothelial cells.
Population studies have shown that individuals with various components of the metabolic syndrome, including obesity, dyslipidaemia and hypertension, are more likely to have retinal microvascular abnormalities such as focal and generalised retinal arteriolar narrowing and venular dilatation [33, 34]. In our study, HFD-induced obesity or hypertension alone produced a significant increase in the formation of retinal acellular capillaries and this was further exacerbated upon their combination as early as 10 weeks. Accelerated retinal capillary dropout, vascular tortuosity and vascular leakage were previously reported in obese SHR rats starting at 12 weeks .
The thioredoxin system is one of the major antioxidant and anti-apoptotic defence mechanisms, and is directly inhibited by TXNIP. Increased levels of TXNIP bind more thioredoxin, limiting its availability for scavenging cellular ROS . Indeed, exposure of rats to HFD or hypertension, or their combination, significantly triggered retinal TXNIP expression, retinal lipid peroxides and NY levels, the footprint of PN formation, implicating endothelial dysfunction compared with controls (Figs 1, 2). Increased lipid peroxidation and PN generation and endothelial dysfunction were also reported in patients with hypertension-related microvascular changes [36, 37] and in retinas from BBZ rats, an obese and noninsulin-dependent model of diabetes , as well as coronary endothelial cells in response to HFD [39, 40]. Nevertheless, this is the first report to demonstrate increases in retinal TXNIP expression in HFD rats, SHR, or their combination. TXNIP can directly activate the redox-sensitive NFκB and its downstream inflammatory cytokines and adhesion molecules. HFD alone, hypertension alone or their combination significantly enhanced levels of NFκB and downstream TNF-α. In line with our findings, rats fed with an HFD or a high sucrose and cholesterol diet had a higher degree of inflammation in isolated retinal vasculature and higher levels of retinal microaneurysms, respectively [41, 42].
Lipotoxicity or impaired tissue homeostasis occurs as a result of lipid-induced changes in intracellular signalling or increased lipid use [9, 43]. Previous reports have established that saturated fatty acids, mainly palmitate but not unsaturated fatty acids, were able to induce pro-inflammatory responses in human coronary endothelial cells [7, 31] and NLRP3 inflammasome activation in cultured macrophages . Accumulated literature provides evidence supporting TXNIP as an important mediator of NLRP3 inflammasome activation. However, the relationship between HFD-induced retinal and endothelial TXNIP expression and inflammasome activation in HFD-induced obesity has not been examined. Here we show that HFD alone selectively induced expression and interaction of TXNIP and NLRP3 resulting in cleaved caspase-1 and IL-1β expression independent of hypertension (Figs 3, 4). These results lend further support to previously documented TXNIP-induced NLRP3 inflammasome activation in lung endothelial cells, macrophages, adipose tissue and pancreatic beta cells in response to ROS, obesity/hyperglycaemia and endoplasmic reticulum stress [18, 32, 44, 45]. To model HFD in vitro, we examined the effects of the saturated fatty acid palmitate alone or in combination with exogenous PN in HRE cells. Indeed, silencing Txnip abrogated palmitate-induced NLRP3 inflammasome activation and its associated increases in pro-apoptotic caspase-3 expression and reduced cell viability of HRE cells (Figs 6, 7, 8). These findings establish TXNIP as an essential activator of the NRLP3 inflammasome in HRE cells, a process which can result in increased retinal endothelial death/apoptosis. In line with our findings, pharmacological inhibition of caspase-1 or deletion of the IL-1 receptor suppressed the IL-1β-dependent formation of retinal acellular capillaries in diabetic animals [23, 46]. These findings support the proposed link between inflammasome activation and accelerated retinal microvascular degeneration. While the role of TXNIP expression was examined in retinal endothelial cells, the contribution of non-vascular retinal cell types is acknowledged and warrants further characterisation. To the best of our knowledge, this is the first report of increased retinal TXNIP expression and activation of endothelial TXNIP–NLRP3 inflammasome in experimental models of HFD.
Our data suggest an inverse relationship between the accumulation of intracellular cleavage/maturation of IL-1β and its release. Pal-BSA alone was able to induce both intracellular maturation and release of IL-1β in HRE cells and this was mitigated by silencing Txnip (Fig. 7). On the other hand, exogenous PN alone resulted in a higher surge of IL-1β release, which was independent of TXNIP inhibition, although it was not able to induce significant changes in its intracellular cleavage/maturation, suggesting accelerated activation and trafficking of IL-1β. Furthermore, this effect was quenched when exogenous PN was combined with Pal-BSA in the presence of TXNIP, whereas TXNIP inhibition reversed this process and facilitated PN-induced IL-1β release despite its combination with Pal-BSA. This data indicate that, while TXNIP is required for palmitate-induced NLRP3 inflammasome activation and IL-1β maturation in HRE cells, it does not facilitate the maximum IL-1β release. Recent evidence suggests that caspase-1 activation/processing of pro-IL-1β by caspase-1 and the release of mature IL-1β from human monocytes are distinct and separable events . TXNIP has been shown to shuffle between different cellular compartments, including the nucleus, mitochondria  and plasma membrane . A possible explanation for this intriguing observation might be due the nature of TXNIP as a member of the alpha arrestin scaffolding proteins, which are believed to play an important role in intracelleular cargo trafficking and/or internalisation of different proteins [49, 50]. Hence, subcellular localisation of TXNIP in response to different insults might reflect an enhancement or inhibition of mature IL-1β release.
In summary, our findings, in conjunction with the fact that obesity has been upgraded from a mere risk factor to a disease state, highlight the detrimental effect of HFD-induced obesity on the vasculature in general and development of retinal microvascular lesions even before reaching a state of hyperglycaemia and frank diabetes. Characterising the early impact of HFD-induced obesity on development of retinopathy and developing TXNIP as a therapeutic target will help to identify innovative strategies for intervention in obesity and related vascular complications affecting millions of patients worldwide.
The authors are grateful for the technical expertise of S. Matragoon and B. A. Pillai (Program in Clinical and Experimental Therapeutics, University of Georgia, Augusta, GA, USA). A. B. El-Remessy and A. Ergul are research pharmacologists at the Charlie Norwood VA Medical Center, Augusta, GA, USA.
This work was supported in part by grants from EY-022408, JDRF (4-2008-149) and Vision Discovery Institute to ABE, HL59699 to JDI, VA Merit Award BX00347 and NS054688 to AE and an American Heart Association pre-doctoral fellowship (12PRE10820002) for INM.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
INM and ABE conceived the hypothesis, and wrote and edited the manuscript. INM, SSH, AF and ABE performed experiments and analysed data to generate final figures, and edited the manuscript. AE and JDI helped with acquisition and interpretation of data and critically edited the manuscript. All authors have read and approved the final version of the manuscript.
- 21.Devi TS, Lee I, Huttemann M, Kumar A, Nantwi KD, Singh LP (2012) TXNIP links innate host defense mechanisms to oxidative stress and inflammation in retinal Müller glia under chronic hyperglycemia: implications for diabetic retinopathy. Exp Diabetes Res 2012:438238PubMedCentralPubMedCrossRefGoogle Scholar