The identification of brown adipose deposits in adults has led to significant interest in targeting this metabolically active tissue for treatment of obesity and diabetes. Improved methods for the direct measurement of heat production as the signature function of brown adipocytes (BAs), particularly at the single cell level, would be of substantial benefit to these ongoing efforts. Here, we report the first application of a small molecule-type thermosensitive fluorescent dye, ERthermAC, to monitor thermogenesis in BAs derived from murine brown fat precursors and in human brown fat cells differentiated from human neck brown preadipocytes. ERthermAC accumulated in the endoplasmic reticulum of BAs and displayed a marked change in fluorescence intensity in response to adrenergic stimulation of cells, which corresponded to temperature change. ERthermAC fluorescence intensity profiles were congruent with mitochondrial depolarisation events visualised by the JC-1 probe. Moreover, the averaged fluorescence intensity changes across a population of cells correlated well with dynamic changes such as thermal power, oxygen consumption, and extracellular acidification rates. These findings suggest ERthermAC as a promising new tool for studying thermogenic function in brown adipocytes of both murine and human origins.
Warm-blooded animals (or endotherms) have developed a number of metabolic processes for thermogenesis to maintain an optimal body temperature, even under extreme cold conditions1. Skeletal muscle tissue generates heat via shivering (involuntary tremor of the muscle)2 or uncoupling of sarco/endoplasmic reticulum Ca2+-ATPase (SERCA)-mediated ATP hydrolysis from Ca2+ transport3. These processes can generate substantial amounts of heat; however, they are insufficient to maintain body temperature under specific conditions, such as in newborns or during sustained cold exposure. Thus, a parallel mechanism has evolved in brown adipose tissue (BAT) to protect hibernating and newborn animals. Non-shivering thermogenesis in brown adipocytes (BAs) occurs via activation of uncoupling protein 1 (UCP1), which is located in the inner mitochondrial membrane and uncouples oxidative phosphorylation (and consequently ATP production) from respiration, resulting in increased heat production4, 5. The recent finding that BAT is not only present in neonates and small infants but also in adults6,7,8,9,10,11, along with the enhanced capacity of BAT for substrate combustion during non-shivering thermogenesis12, has generated substantial interest in BAT as a potential target for treatment of obesity and diabetes.
A key step for the evaluation of pharmaceutical and nutraceutical modulation of BA activity is the ability to monitor changes in BAT function. While the direct measurement of heat production would seem the most appropriate physiological read-out, developing appropriate methodologies has proven surprisingly challenging. Instead, indirect methods, such as mitochondrial depolarisation and oxygen consumption, are typically employed to quantitate heat production-related phenomena in differentiated adipocytes13. A number of techniques for the measurement of BAT thermal power have been introduced recently, including infrared thermography of activated human BA cultures14,15,16, the use of a bimaterial microcantilever, which is able to monitor small temperature variations (0.2 K) in the vicinity of norepinephrine-stimulated primary mouse BAs17, and microcalorimetry18,19,20. However, these techniques are not yet widely applied in metabolic research, most likely as a result of the high cell requirements and the inability to measure individual isolated primary cells. A new multi-channel isothermal microcalorimeter (calScreener), capable of monitoring heat production in growing bacteria, tumour micro-tissues, parasitic worms21 and soil-transmitted helminths22 in a convenient well-plate format has been introduced recently; however, this has not yet been tested with BAs.
As a means of measuring temperature changes at a single-cell scale, fluorescent thermosensors, capable of monitoring intracellular temperature changes in various cell types23,24,25,26,27,28,29,30,31,32,33,34,35, have also been described. To date, only two thermosensors, a genetically encoded fluorescent protein31 and a cytoplasmic nanogel25 have been used for measuring BA thermogenesis. However, both methods require laborious procedures (adenovirus transduction and microinjection, respectively) for probe delivery into the cell. As such, a more user-friendly small molecule-type dye33,34,35 would offer a practical advantage.
Here, we present a novel BODIPY-based thermosensitive dye, ERthermAC, with promising biocompatibility features (small size, excellent cellular uptake, low cytotoxicity, and robust photostability) that is capable of imaging heat production in individual isoproterenol (ISO)-stimulated mouse BAs and in forskolin-stimulated human BAs by confocal microscopy. Data obtained using this dye is comparable with that obtained by conventional methods measuring mitochondrial depolarisation, oxygen consumption, and extracellular acidification rates, and also with thermal power measurements obtained using the calScreener. We believe our findings will facilitate research on the mechanisms regulating brown fat thermogenesis as well as providing new in vitro tools for testing potential modulators of thermogenesis in this unique tissue.
Characterisation of ERthermAC
A diversity-oriented fluorescent library (DOFL) was previously generated through combinatorial synthesis and by the modification of side chains of different fluorescent dye backbones36. Screening of this library for temperature sensitivity led to the recent discovery of ER thermo yellow33. To improve the photostability of the original dye, we removed the alpha-chlorine from ER thermo yellow forming the acetyl-derivative designated ERthermAC. For synthesis (Fig. S1A), absorption and emission spectra (Fig. S1B), NMR spectra (Fig. S2), HPLC chromatogram and mass spectrum (Fig. S3) of ERthermAC, see Supplemental Information. ERthermAC was found to be remarkably photostable, showing only negligible bleaching under harsh conditions (continuous irradiation with high laser power at 543 nm) in contrast to ER thermo yellow (Fig. S4).
To demonstrate that this chemical modification did not change organelle specificity, we co-stained cells with the commercially available endoplasmic reticulum dye ER-Tracker Green. Indeed, strong co-localisation of both dyes was observed in murine brown adipocytes derived from in vitro differentiation of the established WT-1 brown preadipocytes37 (Fig. 1A), confirming ER-specificity of ERthermAC (chemical structure shown in Fig. 1B) in adipocytes.
The relationship between temperature and fluorescence intensity of ERthermAC staining was assessed in formaldehyde-fixed WT-1 cells under temperature-controlled conditions between 18 °C and 43 °C (Fig. 1C and D). A linear reduction in fluorescence intensity was observed between 18.1 °C and 35.0 °C (−1.07%/°C). Fluorescence was more rapidly decreased between 35.7 °C and 42.8 °C, but this reduction also showed an approximately linear relationship (−4.76%/°C, which is comparable to of ER thermo yellow)33.
Characterisation of WT-1 adipocytes
To confirm the suitability of WT-137 cells as a model of BAT, we first assessed their adipogenic differentiation efficiency. Lipid production and storage was visualised using the neutral lipid-specific dye BODIPY 493/503, revealing dense intracellular lipid droplet deposition in WT-1 cells following differentiation protocols (Fig. S5A).
Adipogenesis was further assessed via quantitative real-time PCR measurement of leptin, fatty acid binding protein 4 (fabp4), and peroxisome proliferator-activated receptor-γ 2 (pparg2) mRNA. Significant increases in the expression of these pan-adipocyte markers were observed in induced WT-1 cells compared with the uninduced control group (Fig. S5B). To verify differences in the thermogenic potential of BAT cell model, we then examined the expression of brown fat-specific genes (Fig. S5C). UCP1 mRNA expression was increased ~745-fold in differentiated WT-1 brown adipocytes compared with undifferentiated cells. Gene expression levels of peroxisome proliferator-activated receptor-γ coactivator 1-α (pgc1a) and cell death-inducing DFFA-like effector A (cidea), additional markers of brown fat, were also elevated in induced WT-1 cells (~12-fold and ~3760-fold, respectively). These data suggest that WT-1 cells represent suitable in vitro models of BAT.
Mitochondrial respiration has been shown to correlate closely with heat production38. Indeed, UCP1 activation results in collapse of the electrochemical proton gradient, which in turn blocks oxidative phosphorylation. The cell attempts to restore this gradient by accelerating the electron transport chain, a process requiring increased oxygen uptake39. Increased mitochondrial respiration and glycolytic activity upon adrenergic stimulation were confirmed in WT-1 cells using the Seahorse Extracellular Flux Analyzer. A significant increase in oxygen consumption (OCR) (Fig. S6A) and extracellular acidification (ECAR) rates (Fig. S6B) were observed in ISO-stimulated WT-1 cells immediately after ISO addition.
Consistent with these changes in OCR and ECAR, we observed a remarkable heat-producing response in WT-1 cells upon ISO stimulation using the calScreener isothermal microcalorimeter. At 75 min post-stimulation, the average thermal power of ISO- and vehicle-stimulated WT-1 cell cultures was 22.3 μW and 15.2 μW, respectively. Over the subsequent five hours, the heat flow of both groups gradually decreased to 10.7 μW (ISO) and 9.0 μW (vehicle) (Fig. S6C). Over this period, ISO-stimulated WT-1 cells produced on average 36.5% more heat (271.7 mJ/well) compared with the vehicle group (199.1 mJ/well) (Fig. S6D).
Fluorescence intensity of ERthermAC drastically decreases upon isoproterenol stimulation in WT-1 cells
Increasing temperature due to uncoupling respiration opens up competing, non-radiative relaxation pathways by altering the structural dynamics of the ERthermAC molecule that, in turn, decreases its fluorescence intensity. Upon ISO simulation, the majority of WT-1 cells displayed a rapid reduction in ERthermAC intensity (Fig. 2A and Supplementary Video 1). Stimulation with the solvent vehicle did not alter fluorescence intensity (Figs 2B and S7). The sudden decline in fluorescence intensity occurred after a lag period and at different time points in individual cells (ranging from 20 min onwards) where it reached a minimum level within 10 minutes (Fig. S8 and Supplementary Video 2). Approximately 25% of cells (11 out of 41 cells) were unaffected, suggesting that these non-responders had not acquired thermogenic capacity upon ISO simulation (Fig. 2C). In contrast, incubation with the uncoupling ionophore FCCP led to a simultaneous reduction in intensity across all cells (n = 35 cells from 2 cultures) without an apparent lag period (Figs 3A and S9). The inverse relationship between fluorescence intensity of ERthermAC and temperature indicates heat production in these cells upon adrenergic stimulation. For statistical analysis, we calculated the change in relative intensity (ΔF/F0) for each cell as:
ΔF/F0 values of vehicle and ISO (or FCCP) stimulation were then compared by Student’s t-test with Welch’s correction. A significant reduction in fluorescent activity was observed in ISO and FCCP-stimulated WT-1 cells compared with vehicle controls (both p < 0.0001), and the average reduction in relative intensity was comparable under both stimulatory conditions (ISO: −0.427 ± 0.042; FCCP: −0.520 ± 0.029 compared with the respective control groups). We also observed a morphological change in WT-1 cells (apparent thinning of cells), which occurred in parallel with the reduction in ERthermAC intensity (Fig. S8 and Supplementary Video 2).
Isoproterenol stimulation results in mitochondrial depolarisation in WT-1
BAT thermogenesis requires mitochondrial depolarisation; therefore, as a measure of mitochondrial depolarisation, we determined the effect of isoproterenol stimulation on the fluorescent activity of JC-1, a highly sensitive metachromatic dye-based probe for the assessment of mitochondrial membrane potential40. When the mitochondrion is polarised, JC-1 forms aggregates, which are retained in the mitochondrial matrix and emit light in the orange/red region of the spectrum (at a maximum absorbance of 590 nm). In depolarised mitochondria, JC-1 monomers leak into the cytosol and display green fluorescence (at a maximum absorbance of 527 nm).
FCCP stimulation induced a rapid JC-1 colour change simultaneously in all WT-1 cells with no lag period (Fig. 3B), consistent with the effects seen with ERthermAC. Moreover, in accordance with the changes observed in ERthermAC intensity in WT-1 cells following ISO stimulation, we observed a rapid red-to-green transition of JC-1 after an initial lag phase (Fig. 4, Supplementary Video 3). As before, this change occurred at different time points in individual cells (Fig. 4B). We also observed the previously noted morphological change in ISO-stimulated cells, indicating that this was not a specific effect of ERthermAC (Fig. S10 and Supplementary Video 4). Thus, the depolarisation kinetics determined using JC-1 in ISO and FCCP-stimulated WT-1 cells were comparable to the changes in ERthermAC intensity described above.
Characterisation of human adipocytes
A new immortalized brown preadipocyte cell line was developed using methods described in Xue et al. 41. After adipogenic induction and differentiation, human brown adipocytes (hBAs) showed dense lipid droplet accumulation (Fig. S11A). Moreover, brown preadipocytes exposed to induction cocktail expressed significantly higher levels of mature adipocyte markers, such as fatty acid synthase (FAS), FABP4 and PPARG2, compared to undifferentiated precursor cells (Fig. S11B). UCP1-expression was significantly higher (59,055-fold) in mature hBAs than in undifferentiated preadipocytes, suggesting high thermogenic capacity. Furthermore, mRNA levels of type II iodothyronine deiodinase (DIO2) and PGC1A were also significantly elevated (45.8 and 8.7-fold, respectively) in hBAs compared to the undifferentiated control group (Fig. S11C).
In accordance with results obtained by WT-1 cells, forskolin stimulation enhanced respiration in mature hBAs. OCR increased to 240% of basal level and it did not decrease throughout the measurement (Fig. S12A). Moreover, forskolin stimulation resulted in increased extracellular acidification upon stimulation (Fig. S12B).
Intensity drop of ERthermAC suggests heat production in human brown adipocytes
The majority of stimulated hBAs displayed a rapid reduction in ERthermAC intensity suggesting heat production (Fig. 5A, Supplementary Video 5). Similarly to WT-1 cells, the sudden decline in fluorescence intensity occurred after a lag period and at different time points in individual hBAs. In contrast, vehicle stimulation did not cause significant alterations in ERthermAC intensity (Fig. 5B). After forskolin stimulation, two different groups of cells could be identified, corresponding to responding (89.7%) and non-responding cells (10.3%), suggesting that the differentiated human brown adipocytes represent a heterogeneous population of cells. The average amplitude of intensity drop was −0.504 ± 0.056 in stimulated hBAs, which is comparable to that of WT-1 cells. In contrast, forskolin stimulation did not result in ERthermAC intensity drop in undifferentiated stromal vascular fraction (SVF) cells from brown fat tissue (Fig. S13).
JC-1 colour change suggests mitochondrial depolarisation in human brown adipocytes
Similarly to WT-1 cells, we observed a rapid red-to-green transition of JC-1 in hBAs (Fig. 6A, Supplementary Video 6), which change occurred at different time points in individual cells (Fig. 6B). These findings suggest that forskolin stimulation results in mitochondrial depolarisation in human brown adipocytes.
Changes in endoplasmic reticulum morphology, Ca2+ concentration and pH do not underlie the changes in ERthermAC fluorescence intensity
We then considered whether ERthermAC intensity might be affected by other major and dynamic cellular factors, such as morphology, Ca2+ concentration, and pH. Distribution and intensity of ER-Tracker Green was unaltered following ISO stimulation in WT-1 cells (Fig. S14), indicating that changes in cell morphology were unlikely to contribute to the observed decrease in ERthermAC fluorescence intensity. The effect of Ca2+ on ERthermAC intensity was assessed by incubation in HEPES buffer containing a range of physiological Ca2+ concentrations (10–1000 μM)42 by spectrophotometry. No change in fluorescence intensity was observed across this range (Fig. S15).
To estimate pH fluctuations in the ER of ISO-stimulated WT-1 cells, we used a super ecliptic pHluorin43 tagged with an ER-targeting KDEL signal peptide (for ER-pHluorin probe generation, see Supplemental Information). We observed a biphasic pH profile in response to ISO stimulation (Fig. S16). After an initial gradual decrease from basal levels (pH 7.3 ± 0.1) to pH 7.0 ± 0.1, most likely a result of fatty acid release during lipolysis44, the intra-endoplasmic pH was rapidly and substantially increased to pH 8.1 ± 0.3. Notably, this increase coincided with both the previously noted morphological changes as well as the respective changes in intensity and colour of ERthermAC and JC-1. However, determination of the relationship between pH and ERthermAC fluorescence using the same method as for ER-pHluorin sensor calibration45 revealed that fluorescence intensity was only reduced by approximately 8.5% between pH 7.0 and 8.1 (Fig. S16E; for detailed calibration protocol, see Supplemental Information). For comparison, the average ERthermAC intensity reduction in WT-1 cells was 42.7% following ISO stimulation, suggesting that alkalisation was not a major contributing factor to the signal changes observed.
Thermogenesis is a crucial physiological task for endothermic animals, and brown adipose tissue, in particular, has the capacity to produce high levels of heat via uncoupled respiration. Originally, BAT was assumed to be present only in human newborns and small infants to provide shiver-free thermogenesis, but has now also been shown to be present in adults6,7,8,9,10,11. This discovery has been the subject of substantial interest in both academic and pharmaceutical research46 as a result of the high level of glucose and free fatty acid consumption by activated murine BAT – it is estimated that 50 g of BAT can burn 250 kcal daily47. Stimulation of BAT energy expenditure via sympathetic48 or non-sympathetic49, 50 pathways is considered a potential new drug target for obesity and diabetes treatments, and there is increasing interest in the discovery of novel mechanisms and substances for this purpose (reviewed in ref. 51). However, to facilitate the development of breakthrough therapies, it is essential that techniques to measure the signature function of BAT are readily available.
Intracellular thermosensors, including thermoresponsive polymers23,24,25,26, dye-doped polymeric nanoparticles27,28,29, fluorescent proteins30, 31, and small molecules32,33,34,35, which are capable of monitoring intracellular temperature changes at the single-cell level have been described previously. Here, we used a newly developed small molecule-type thermosensitive dye, ERthermAC, that is easily taken up by cells within 30 minutes and accumulates rapidly and specifically in the ER. Although activation of UCP1 takes place in the mitochondria, we positioned ERthermAC into the endoplasmic reticulum for several reasons. Firstly, the chemical and physical properties of mitochondria, such as membrane potential, can vary extensively under thermogenic stimulation, potentially impairing accuracy of measurements. In contrast, ER membrane potential remains minimally depolarised during Ca2+ release52. Physiological Ca2+ concentrations (10–1000 μM)42 were not found to influence ERthermAC intensity, and pH, which varies within the ER between 7.0 and 8.1 during adrenergic stimulation, had only a modest impact (−8.5%) on its fluorescence intensity. Secondly, mitochondria and ER are in close proximity in BAs, with lipid droplets occupying the majority of intracellular space. Recent evidence suggests that ER membranes are fused with the outer mitochondrial membrane in BAT53, which would not only provide opportunities for intensive communication between these organelles but also sufficiently close proximity for heat measurement. Thus, placing ERthermAC into the extensive endoplasmic network creates an extended contiguous thermometer in the immediate mitochondrial vicinity while reducing the potential for interference from radical environmental changes (Fig. 7).
We assessed the application of ERthermAC intensity as readout of thermogenesis using the adrenergic stimulus, isoproterenol or forskolin. ISO mimics the physiological activation of BAT via binding to beta-adrenergic receptors and the subsequent activation of adenylate cyclase54 and generation of second messenger cyclic adenosine monophosphate (cAMP), which in turn activates the cAMP-dependent protein kinase A (PKA). In contrast, forskolin activates adenylate cyclase directly resulting in elevated cAMP levels. PKA can phosphorylate lipolytic enzymes, such as hormone-sensitive lipase, resulting in the release of free fatty acids (FFA) from lipid droplets. In activated BAs, FFAs interact with and activate UCP1, which short-circuits the proton gradient by shunting protons away from ATP synthesis55. This results in a gradient discharge with heat generated by the dissipation of the proton-motive force and also indirectly by elevated respiratory chain activity in an attempt to sustain the electrochemical gradient and ATP production46. Adrenergic stimulation of individual BAs differentiated from immortalised mouse37 and human precursor cells resulted in a rapid decline in intracellular ERthermAC intensity after a lag phase of several minutes, indicating a steep increase of temperature within the immediate vicinity of the dye. In contrast, addition of the chemical uncoupler FCCP induced an immediate cellular response in all WT-1 cells. Classical uncoupling reagents, such as FCCP, CCCP and 2,4-dinitrophenol, are lipid-soluble weak acids that directly uncouple mitochondrial oxidative phosphorylation by diverting protons across the inner mitochondrial membrane. The parallel transport leading to ATP production is short-circuited, and the energy of proton motive force is dissipated as heat56, 57. Thus, we assume the lag phase observed with ISO stimulation may correspond to the time needed to complete the adrenergic signalling cascade. However, further work is required to determine the basis of the staggered thermogenic activation kinetics in individual cells.
Notably, the kinetics of mitochondrial membrane depolarisation, as visualised using JC-1, were comparable to those of ERthermAC staining intensity in both ISO- and FCCP-exposed WT-1 cells and in forskolin stimulated human brown adipocytes including the staggered activation of cells. Overlapping emission spectra of both fluorescent probes precluded co-staining, but the respective colour and intensity profiles strongly suggest that the same thermogenic events were captured. Moreover, bright-field image series taken in parallel in the same cell culture with either ERthermAC or JC-1 staining revealed a fast, shrinkage-like cellular movement in activated BA that occurred in concert with the intensity reduction of ERthermAC and the colour change of JC-1, strengthening the hypothesis that the same cellular event was captured with ERthermAC and JC-1.
To validate the detection of thermogenesis in our model, we further employed the current standard Seahorse XF Analyzer and the novel multi-channel calScreener microcalorimeter on WT-1 cells as alternative measures. The profiles of enhanced oxygen consumption (OCR) and extracellular acidification (ECAR) rates (Fig. S6) in ISO-stimulated WT-1 cells corresponded closely with the averaged intensity change of ERthermAC (Fig. 2) and colour change of JC-1 (Fig. 4). In addition, the multi-channel design of the calScreener allowed us to simultaneously monitor heat generation in 32 different samples of adherent cell monolayers. Thus, we were able to compare the metabolic responses of differentiated mouse BAs in real-time, similarly to Seahorse measurements. Compared to the vehicle group, isoproterenol significantly increased the thermal power of WT-1 cells by 47.1% 75 minutes after stimulation and this increase was sustained for at least 6 hours post-stimulation, although the difference between treatment groups decreased gradually over time. We determined the thermal power of WT-1 cells to be 0.5 nW per cell in a given monolayer at 75 min post-stimulation, assuming that the number of cells was not significantly changed between seeding and the calorimetric measurement. This is within the range of previous studies that measured large amounts of suspended cells in a batch microcalorimeter (0.82 nW per hamster BAs and 5.1 nW per rat BAs)18, 19 or using nanocalorimetric sensors (1.6 nW per mouse BAs)20. However, technical limitations with the current calScreener detection instrumentation, including the 3-step equilibrium procedure after administration of the stimulus and the manual closures of the detection vials, mean that the first 75 min after stimulation, and thus the initial peak of thermogenesis, are not captured, potentially leading to an underestimation of thermal power. This limitation could be overcome with the use of an injection system to allow instantaneous monitoring of heat generation.
The question of temperature heterogeneity and the interpretation of single cell thermometry data is the subject of on-going debate in the field of cellular thermosensing58,59,60,61, and a recent commentary proposed that temperature differentials measurable in cells should not exceed a μK range58. As such, the apparent temperature increase from 25 °C to 42 °C in WT-1 cells following isoproterenol stimulation as measured by ERthermAC staining would appear, at first sight, surprising (assuming the initial temperature in the ER is equivalent to that of the culture medium at 25 °C). However, it is important to note that this finding is specific to the ER and does not mean that the whole cell would reach this temperature. Site-specific thermometry close to the point of heating i.e. probably the mitochondria in the case of uncoupled respiration allows the monitoring of local temperature changes and would show smaller temperature differentials further away from the point of heating60. Indeed, Kiyonaka et al. identified a 5 °C temperature change when a recombinantly expressed thermosensitive protein targeted the mitochondria but smaller differences when the thermosensor was located in the cytosol31. Moreover, a recent study has reported a maximum change of 1.3 °C in the cytoplasm of stimulated brown adipocytes25. Our results obtained by microcalorimetry, applied for the first time in adherent monolayer cultures, along with previous findings in other cell systems18,19,20, strengthen the hypothesis that local intracellular temperatures within BAs can increase by several degrees. Furthermore, when we take into consideration that not every cell is stimulated (approximately 75% of WT-1 cells showed changes in ERthermAC intensity within 2 hours), an average thermal power of individual cells in the range of several hundred picowatts to nanowatts would provide additional support to explain our findings with ERthermAC.
It is also important to note that brown adipose tissue thermogenesis has a major role in hibernating animals during arousal when they are able to restore their body temperature from 4 °C, or even below this, to normal levels62, thus requiring a very large increase in heat generation. In UCP1-ablated mice, peak rewarming rates from induced torpor were reduced by 50% (UCP1+/+: 0.24 ± 0.08 °C min−1; UCP1−/−: 0.12 ± 0.04 °C min−1)63, strengthening the hypothesis that UCP1-facilitated non-shivering thermogenesis in BAT provides an extraordinarily high heat output to allow fast rewarming of the animal. A number of protective mechanisms exist in BAs to prevent cellular damage during these fast and relatively high temperature changes and ensure that hibernating animals endure little to no damage to their tissues during torpor arousal, such as antioxidants and heat shock proteins64. Moreover, it was recently demonstrated that membrane lipids, such as cardiolipin, stabilise UCP1, increasing its thermal stability65 and conferring a higher temperature tolerance for mitochondrial membrane proteins.
Although our results suggest that Ca2+ concentration or morphology have no significant effect on ERthermAC staining intensity, and the pH fluctuation in the ER is only responsible for a slight change, we cannot rule out the possibility that other unknown factors might contribute to the high intensity change detected. Further experiments using fluorescence lifetime imaging or the development of a ratiometric dye based on ERthermAC are needed to clarify the exact amount of temperature change. Nevertheless, we believe that ERthermAC provides an elegant and user-friendly method for the qualitative study of heat generation in BAs at a single-cell level.
In summary, we have successfully demonstrated that the novel BODIPY-based thermosensitive dye, ERthermAC, is quickly and easily taken up into the endoplasmic reticulum of adipocytes where it forms a contiguous intraorganellar thermometer for the optical visualisation of thermogenesis. Following adrenergic stimuli, heat production occurred at random timings in each cell, and the dynamics of thermogenesis was consistent with mitochondrial depolarisation observed using JC-1. Consequently, ERthermAC is a promising tool for examination of brown adipose tissue thermogenesis, which is compatible with both time-lapse studies and image-based high-content screening.
Materials and Methods
Cell culture and adipogenic induction of WT-1 brown preadipocytes
Immortalised brown preadipocytes (WT-1)37 were cultured in high glucose DMEM (HG DMEM) supplemented with GlutaMAX, 10% FBS (#41F5724, #41F1623K and 42F7254K; 10270, Gibco/Life Technologies, Carlsbad, CA, USA), and Penicillin/Streptomycin (P/S; Gibco/Life Technologies). Cells were maintained at 37 °C in a humidified atmosphere of 5% CO2. To prevent spontaneous differentiation, cells were maintained at subconfluent levels prior to being detached using TrypLE Express (Gibco/Life Technologies), passaged at 1:10, and cultured to generate subsequent passages. Directed differentiation was carried out with cells at passage 28 to 38. This cell line was used as a BAT model in our experiments.
For adipogenic induction, WT-1 cells were seeded at a density of 3 × 104 cells/cm2 in basal medium (HG DMEM supplemented with GlutaMAX (Gibco/Life Technologies), 2% FBS, and P/S). Cells were pre-treated for three days with 3.3 nM BMP7 (354-BP, R&D Systems, Minneapolis, MN, USA), 20 nM insulin (91077 C, Sigma, St-Louis, MO, USA), and 1 nM T3 (T5516, Sigma). Cells were then exposed to an induction cocktail for two days, which consisted of 0.5 mM 3-isobutyl-1-methylxanthine (IBMX), 0.125 mM indomethacin, 5 μM dexamethasone, 20 nM insulin (I5879, I7378, D4902 and 91077 C, respectively, Sigma), and 1 nM T3 in basal medium, followed by a three-day maintenance phase (basal medium containing 20 nM insulin and 1 nM T3). Cells were cultured in HG DMEM without FBS and hormonal supplements for 24 h prior to experimental analysis. Non-induced control cells for qPCR analyses were maintained with basal medium alone on the same schedule.
Generation of immortalized human brown preadipocytes
This study was carried out in accordance with the institutional guidelines of and was approved by the Human Studies Institutional Review Boards of Beth Israel Deaconess Medical Center and Joslin Diabetes Center. Details on procedures of human subject collection were described previously41, 66. For characterisation and confocal microscopy experiments, neck fat from one subject was studied. The subject gave written informed consent before taking part in the study.
Isolation and immortalization procedure of primary stromal vascular fraction (SVF) from human neck fat was described previously66, 67. Briefly, deep neck fat was collected from a female subject (age 56, BMI 30.8). SVF cells were isolated and expanded in culture and split a few times before immortalization. To immortalize, hBAT-SVF cells were infected with retroviral particles encoding the plasmid pBABE-Hygro-hTERT (Addgene Plasmid #1773, Cambridge, MA, USA). Following retrovirus infection, cells were selected with 200 ug/ml Hygromycin (hBAT) for two weeks. Once drug selection was finished, immortalized cells were allowed to grow in HG DMEM medium containing 10% FBS.
Cell culture and adipogenic induction of human brown adipocytes
For adipogenic induction , cells were seeded in an initial density of 10.5 k cells/cm2. When reached confluence, the cells were pre-treated with 3.3 nM BMP7, 0.5 μM insulin and 2 nM T3 in high glucose DMEM,10% FBS and 1% P/S for 6 days. Then, the cells were treated with adipogenic induction cocktail for 18 days (medium was replaced every 3 days). The induction cocktail consisted of basal media (high glucose DMEM, 10% FBS and 1% P/S), 0.5 μM insulin, 0.1 μM dexamethasone, 30 μM indomethacine, 0.5 mM IBMX, 2 nM T3, 1 μM rosiglitazone, 33 μM biotin and 17 μM pantothenate. Non-induced control cells for qPCR analyses were maintained with basal medium alone on the same schedule. Cells were cultured in HG DMEM without FBS and hormonal supplements for 24 h prior to experimental analysis.
Staining with different fluorescent probes
Cells were cultured and differentiated as described above in glass-bottom dishes. Mature cells were incubated with different fluorescent probes (alone or in combination) for 30 min at 37 °C in HG DMEM at the following concentrations: 250 nM ERthermAC, 500 nM ER-Tracker Green (E34241, Molecular Probes, Life Technologies), 5 μM JC-1 (T3168, Molecular Probes, Life Technologies). Medium was replaced with fresh HG DMEM prior to imaging.
For assessment of lipid droplet content, cell cultures were rinsed with PBS, fixed in 4% formaldehyde (28908, Pierce Biotechnology Inc., Rockford, IL, USA) for 10 min at room temperature, then co-stained for 30 min with 0.1 μg/ml BODIPY 493/503 (D3922, Molecular Probes, Life Technologies) for cytoplasmic lipid droplets, and 0.5 μg/ml of 4′,6-diamidino-2-phenylindole (DAPI) (D3571, Molecular Probes, Life Technologies) for nuclear DNA.
Stained WT-1 cells were imaged with an Olympus FV 1000 confocal microscope with a 60× objective (PLAPON60XO, NA 1.42, oil immersion), equipped with stage incubator and CO2 supply. Microscopic images of ERthermAC and the red channel of JC-1 were obtained using a 543 nm laser (1.71 μW), dichroic mirrors DM405/488/543 and SDM560, and the emission band set at 555–655 nm. To capture ER-Tracker Green, BODIPY 493/503, ER-pHluorin, and the green channel of JC-1 images, a 488 nm laser (4.97 μW) and a dichroic mirror DM405/488/543 were used with the emission band set at 500–530 nm. Laser power was measured at the specimen using a laser power meter (NOVA II, 7Z01550, Ophir Optronics Solutions, Israel) and sensor (PD300, 7Z02410, Ophir Optronics Solutions).
For time-lapse imaging experiments (ERthermAC and JC-1), medium was replaced with 900 μl fresh HG DMEM (25 °C) after staining and equilibrated at 25 °C for 15 min. Imaging was initiated and 100 μl isoproterenol (ISO, final concentration: 10 μM; I6504, Sigma), carbonyl cyanide p-(tri-fluromethoxy)phenyl-hydrazone (FCCP, final concentration: 10 μM: XF Cell Mito Stress Test Kit, 101706–100, Seahorse Bioscience), or vehicle (HG DMEM) was added after 5 min. Cells were recorded over 120 min (ISO) or 90 min (FCCP), and images were acquired every 20 sec (ERthermAC) or 1 min (JC-1). Imaging speed was 4 μs/pixel; image size was 512 by 512. The intensity of fluorescence was analysed using ImageJ software. Regions of interest (ROI) were selected manually and pixel intensities were spatially averaged. Gain was set individually for each sample to avoid saturation of the signal. Consequently, results are interpreted as relative intensity (intensity of all ROIs divided by intensity at time 0).
Stained human brown adipocytes were imaged with a Zeiss LSM710NLO confocal microscope in point scan mode with a 40× objective (LD C-Apochromat 40×/1.1 W Korr), equipped with stage incubator and CO2 supply. Microscopic images of ERthermAC and the red channel of JC-1 were obtained using a 561 nm diode-pumped solid-state laser and the emission band was set at 589–667 nm. To capture BODIPY 493/503 and the green channel of JC-1 images, a 488 nm Argon laser were used with the emission band set at 499–560 nm. For Hoechst 33342 nuclear staining, a Chameleon Vision (Coherent, Santa Clara, CA, USA) laser was applied with the emission band set at 425–475 nm. Time-lapse imaging experiments were performed similarly as above described with the following modifications: cells were stimulated with forskolin (final concentration: 10 μM; F6886, Sigma) or vehicle (HG DMEM with DMSO) and they were recorded over 120 min (forskolin and vehicle) and images were acquired every 20 sec (ERthermAC) or 1 min (JC-1). Pixel dwell: 1.58 μsec; image size was 512 by 512.
Unless otherwise stated, all results are reported as mean ± standard error (SEM). Student’s t-test with Welch’s correction was used for comparison between two groups of time-lapse confocal imaging results, and Student’s t-test was used for mouse qPCR, Seahorse (both species) and calScreener experiments. For multiple pairwise comparisons (human qPCR), one-way ANOVA statistical analysis was performed followed by Tukey Multiple Comparisons Test using Graphpad Prism software (GraphPad Software, Inc., La Jolla, CA, USA), n.s. not significant; (*)p < 0.05; (**)p < 0.01; (***)p < 0.001; (****)p < 0.0001.
Bicego, K. C., Barros, R. C. & Branco, L. G. Physiology of temperature regulation. Comp. Biochem. Physiol. A. Mol. Integr. Physiol 147, 616–639 (2007).
Jansky, L. Shivering in Physiology and Pathophysiology of Temperature Regulation (ed. Blatteis, C.) 47–59 (World Scientific, Singapore, 1998).
de Meis, L., Arruda, A. P. & Carvalho, D. P. Role of sarco/endoplasmic reticulum Ca(2+)-ATPase in thermogenesis. Biosci. Rep. 25, 181–190 (2005).
Himms-Hagen, J. Nonshivering thermogenesis. Brain Res. 12, 151–160 (1984).
Ricquier, D. Uncoupling protein 1 of brown adipocytes, the only uncoupler: a historical perspective. Front. Endocrinol. (Lausanne) 2, 85 (2012).
Nedergaard, J., Bengtsson, T. & Cannon, B. Unexpected evidence for active brown adipose tissue in adult humans. Am. J. Physiol. Endocrinol. Metab 298, E444–452 (2007).
Saito, M. et al. High incidence of metabolically active brown adipose tissue in healthy adult humans: effects of cold exposure and adiposity. Diabetes 58, 1526–1531 (2009).
van Marken Lichtenbelt, W. D. et al. Cold-activated brown adipose tissue in healthy men. N. Eng. J. Med. 360, 1500–1508 (2009).
Virtanen, K. A. et al. Functional brown adipose tissue in healthy adults. N. Eng. J. Med. 390, 1518–1525 (2009).
Cypess, A. M. et al. Identification and importance of brown adipose tissue in adult humans. N. Engl. J. Med. 360, 1509–1517 (2009).
Wu, J. et al. Beige adipocytes are a distinct type of thermogenic fat cell in mouse and human. Cell 150, 366–376 (2012).
Lidell, M. E. & Enerbäck, S. Brown adipose tissue – a new role in humans? Nat. Rev. Endocrinol 6, 319–325 (2010).
Wikstrom, J. D. et al. Hormone-induced mitochondrial fission is utilized by brown adypocytes as an amplification pathway for energy expenditure. EMBO J. 33, 418–436 (2014).
Lee, P. et al. Irisin and FGF21 are cold-induced endocrine activators of brown fat function in humans. Cell Metab. 19, 302–309 (2014).
Lee, P., Werner, C. D., Kebebew, E. & Celi, F. S. Functional thermogenic beige adipogenesis is inducible in human neck fat. Int. J. Obes. (Lond.) 38, 170–176 (2014).
Rossato, M. et al. Human white adipocytes express the cold receptor TRPM8 which activation induces UCP1 expression, mitochondrial activation and heat production. Mol. Cell. Endocrin 383, 137–146 (2014).
Sato, M. K. et al. Temperature changes in brown adipocytes detected with a bimaterial microcantilever. Biophys. J. 106, 2458–2464 (2014).
Nedergaard, J., Cannon, B. & Lindberg, O. Microcalorimetry of isolated mammalian cells. Nature 267, 518–520 (1977).
Clark, D. G., Brinkman, M. & Neville, S. D. Microcalorimetric measurements of heat production in brown adipocytes from control and cafeteria-fed rats. Biochem. J. 235, 337–242 (1986).
Johannessen, E. A., Weaver, J. M., Bourova, L., Svoboda, P., Cobbold, P. H. & Cooper, J. M. Micromachined nanocalorimetric sensors for ultra-low volume cell-based assays. Anal. Chem. 74, 2190–21907 (2002).
Braissant, O. et al. Isothermal microcalorimetry accurately detects bacteria, tumorous microtissues and parasitic worms in a label-free well-plate assay. Biotechnol. J 10, 460–468 (2015).
Flores, D., Panic, G., Braissant, O. & Keiser J. A novel isothermal microcalorimetry tool to assess drug effects on Ancylostoma ceylanicum and Necator americanus. Appl. Microbiol. Biotechnol. (2015).
Gota, C., Okabe, K., Funatsu, T., Harada, Y. & Uchiyama, S. Hydrophilic fluorescent nanogel thermometer for intracellular thermometry. J. Am. Chem. Soc. 131, 2766–2767 (2009).
Okabe, K., Inada, N., Gota, C., Harada, Y., Funatsu, T. & Uchiyama, S. Temperature mapping with fluorescent polymeric thermometer and fluorescence lifetime imaging microscopy. Nat. Commun. 3, 705 (2012).
Hattori, K. et al. ASK1 signalling regulates brown and beige adipocyte function. Nat. Commun. 7, 11158 (2016).
Hayashi, T., Fukuda, N., Uchiyama, S. & Inada, N. A cell-permeable fluorescent polymeric thermometer for intracellular temperature mapping in mammalian cell lines. PLoS One 10, e0117677 (2015).
Peng, H. S., Huang, S. H. & Wolfbeis, O. S. Ratiometric fluorescent nanoparticles for sensing temperature. J. Nanopart. Res. 12, 2729–2733 (2010).
Oyama, K. et al. Walking nanothermometers: Spatiotemporal temperature measurement of transported acidic organelles in single living cells. Lab. Chip 12, 1591–1593 (2012).
Takei, Y. et al. A nanoparticle-based ratiometric and self-calibrated fluorescent thermometer for single living cells. ACS Nano. 8, 198–206 (2014).
Donner, J. S., Thompson, S. A., Kreuzer, M. P., Baffou, G. & Quidant, R. Mapping intracellular temperature using green fluorescent protein. Nano. Lett. 12, 2107–211 (2012).
Kiyonaka, S. et al. Genetically encoded fluorescent thermosensors visualize subcellular thermoregulation in living cells. Nat. Methods 10, 1232–1238 (2013).
Chapman, C. F., Liu, Y., Sonek, G. J. & Tromberg, B. J. The use of exogenous fluorescent probes for temperature measurements in single living cells. Photochem. Photobiol. 62, 416–425 (1995).
Arai, S., Lee, S. C., Zhai, D., Suzuki, M. & Chang, Y. T. A molecular fluorescent probe for targeted visualization of temperature at the endoplasmic reticulum. Sci. Rep 4, 6701 (2014).
Arai, S. et al. Mitochondria-targeted fluorescent thermometer monitors intracellular temperature gradient. Chem. Commun. 51, 8044–8047 (2015).
Itoh, H. et al. Direct organelle thermometry with fluorescence lifetime imaging microscopy in single myotubes. Chem. Commun. 52, 4458–4461.
Vendrell, M., Zhai, D., Er, J. C. & Chang, Y. T. Combinatorial strategies in fluorescent probe development. Chem. Rev. 112, 4391–4420 (2012).
Tseng, Y. H. et al. New role of bone morphogenetic protein 7 in brown adipogenesis and energy expenditure. Nature 454, 1000–1004 (2008).
Ricquier, D., Gaillard, J. L. & Turc, J. M. Microcalorimetry of isolated mitochondria from brown adipose tissue. Effect of guanosine-di-phosphate. FEBS Lett 99, 203–206 (1979).
Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K. & Walter, P. 2007. Energy conversion: mitochondria and chloroplasts. In Molecular biology of the cell, 5 th ed . 813–878 (Garland Science, Taylor & Francis Group, 2007).
Salvioli, S., Ardizzoni, A., Franceschi, C. & Cossarizza, A. JC-1, but not DiOC6(3) or rhodamine 123, is a reliable fluorescent probe to assess delta psi changes in intact cells: implications for studies on mitochondrial functionality during apoptosis. FEBS Lett 411, 77–78 (1997).
Xue, R. et al. Clonal analyses and gene profiling identify genetic biomarkers of the thermogenic potential of human brown and white preadipocytes. Nat Med 21, 760–768 (2015).
Demaruex, N. & Frieden, M. Measurements of the free luminal ER Ca(2+) concentration with targeted cameleon fluorescent proteins. Cell. Calcium. 34, 109–119 (2003).
Sankaranarayanan, S., De Angelis, D., Rothman, J. E. & Ryan, T. A. The use of pHluorins for optical measurements of presynaptic activity. Biophys. J. 79, 2199–2208 (2000).
Civelek, V. N., Hamilton, J. A., Tornheim, K., Kelly, K. L. & Corkey, B. E. Intracellular pH in adipocytes: effects of free fatty acid diffusion across the plasma membrane, lipolytic agonists and insulin. Proc. Natl. Acad. Sci. USA 93, 10139–10144 (1996).
Kim, J. H., Lingwood, C. A., Williams, D. B., Furuya, W., Manolson, M. F. & Grinstein, S. Dynamic measurement of the pH of the Golgi complex in living cells using retrograde transport of the verotoxin receptor. J. Cell Biol. 134, 1387–1399 (1996).
Tseng, Y. H., Cypess, A. M. & Kahn, C. R. Cellular bioenergetics as a target for obesity therapy. Nat. Rev. Drug Discov. 9, 465–482 (2010).
Rothwell, N. J. & Stock, M. J. Luxuskonsumption, diet-induced thermogenesis and brown fat: the case in favour. Clin. Sci. (Lond.) 64, 19–23 (1983).
Cypess, A. M. et al. Activation of human brown adipose tissue by β3-adrenergic receptor agonist. Cell Metab. 21, 33–38 (2015).
Villarroya, F. & Vidal-Puig, A. Beyond the sympathetic tone: the new brown fat activators. Cell Metab. 17, 638–643 (2013).
Bonet, M. L., Oliver, P. & Palou, A. Pharmacological and nutritional agents promoting browning of white adipose tissue. Biochim. Biophys. Acta. 1832, 969–985 (2013).
Kajimura, S., Spiegelman, B. M. & Seale, P. 2015. Brown and beige fat: physiological roles beyond heat generation. Cell Metab. 22, 546–549 (2015).
Lam, A. K. & Galione, A. The endoplasmic reticulum and junctional membrane communication during calcium signaling. Biochim. Biophys. Acta. 1833, 2542–2559 (2013).
de Meis, L., Ketzer, L. A., da Costa, R. M., de Andrade, I. R. & Benchimol, M. Fusion of the endoplasmic reticulum and mitochondrial outer membrane in rats brown dipose tissue: activation of thermogenesis by Ca2+. PLoS One 5, e9439 (2010).
Granneman, J. G. Norephinephrine infusions increase adenylate cyclase responsiveness in brown adipose tissue. J. Pharmacol. Exp. Ther. 245, 1075–1080 (1988).
Cannon, B. & Nedergaard, J. Brown adipose tissue: functional and physiological significance. Physiol. Rev. 84, 277–359 (2004).
Ting, H. P., Wilson, D. F. & Chance, B. Effects of uncouplers of oxidative phosphorylation on the specific conductance of bimolecular lipid membranes. Arch. Biochem. Biophys. 141, 141–146 (1970).
Heytler, P. G. Uncouplers of oxidative phosphorylation. Methods Enzymol 55, 462–542 (1979).
Baffou, G., Rigneault, H., Marguet, D. & Jullien, L. A critique of methods for temperature imaging in single cells. Nat. Methods 11, 899–901 (2014).
Suzuki, M., Zeeb, V., Arai, S., Oyama, K. & Ishiwata, S. The 10(5) gap issue between calculation and measurement in single-cell thermometry. Nat. Methods 12, 802–803 (2015).
Kiyonaka, S., Sakaguchi, R., Hamachi, I., Morii, T., Yoshizaki, T. & Mori, Y. Validating subcellular thermal changes revealed by fluorescent thermosensors. Nat. Methods 12, 801–802 (2015).
Baffou, G., Rigneault, H., Marguet, D. & Jullien, L. Reply to: “Validating subcellular thermal changes revealed by fluorescent thermosensors” and “The 10(5) gap issue between calculation and measurement in single-cell thermometry”. Nat. Methods 12, 803 (2015).
Hampton, M., Melvin, R. G. & Andrews, M. T. Transcriptomic analysis of brown adipose tissue across the physiological extremes of natural hibernation. PLoS One 8, e85157 (2013).
Oelkrug, R., Heldmaier, G. & Meyer, C. W. Torpor patterns, arousal rates and temporal organization of torpor entry in wildtype and UCP1-ablated mice. J. Comp. Physiol. B 181, 137–145 (2011).
Rouble, A. N., Tessier, S. N. & Storey, K. B. Characterization of adipocyte stress response pathways during hibernation in thirteen-lined ground squirrel. Mol. Cell. Biochem. 393, 271–282 (2014).
Crichton, P. G. et al. Trends in thermostability provide information on the nature of substrate, inhibitor and lipid interactions with mitochondrial carriers. J. Biol. Chem. 290, 8206–8217 (2015).
Cypess, A. M. et al. Anatomical localization, gene expression profiling and functional characterization of adult human neck brown fat. Nat. Med 19, 635–639 (2013).
Tchkonia, T. et al. Fat depot–specific characteristics are retained in strains derived from single human preadipocytes. Diabetes 55, 2571–2578 (2006).
Shamsi, F. & Tseng, Y. H. Protocols for generation of immortalized human brown and white preadipocyte cell lines. Methods Mol. Biol. 1566, 77-85 (2017).
The authors are grateful to the NUS Tissue Engineering Program for their continued support. We would like to thank Kerry McLaughlin of Insight Editing London for critical review of the manuscript. This study was supported through an intramural grant (MetaTec) from the Life Sciences Institute of the National University of Singapore (to MR). This research was also supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number 15K05251, and by PRESTO, JST (to MS), and by the Biomedical Research Council of Singapore (HI, EBL). HI was supported by an A*STAR Research Attachment Programme award. This work was also supported in part by US National Institutes of Health (NIH) grants R01DK077097 and R01DK102898 (to Y.-H.T.), and P30DK036836 (to Joslin Diabetes Center’s Diabetes Research Center).
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Kriszt, R., Arai, S., Itoh, H. et al. Optical visualisation of thermogenesis in stimulated single-cell brown adipocytes. Sci Rep 7, 1383 (2017). https://doi.org/10.1038/s41598-017-00291-9
This article is cited by
Substantial impact of 3-iodothyronamine (T1AM) on the regulations of fluorescent thermoprobe-measured cellular temperature and natriuretic peptide expression in cardiomyocytes
Scientific Reports (2022)
Biophysical Reviews (2022)
Nature Metabolism (2022)
Communications Biology (2021)
Nature Methods (2020)