In this protocol, methods relevant for BAT-optimized arteriovenous metabolomics using GC-MS in a mouse model are outlined. These methods allow for the acquisition of valuable insights into BAT-mediated metabolite exchange at the organismal level.
Brown adipose tissue (BAT) plays a crucial role in regulating metabolic homeostasis through a unique energy expenditure process known as non-shivering thermogenesis. To achieve this, BAT utilizes a diverse menu of circulating nutrients to support its high metabolic demand. Additionally, BAT secretes metabolite-derived bioactive factors that can serve as either metabolic fuels or signaling molecules, facilitating BAT-mediated intratissue and/or intertissue communication. This suggests that BAT actively participates in systemic metabolite exchange, an interesting feature that is beginning to be explored. Here, we introduce a protocol for in vivo mouse-level optimized BAT arteriovenous metabolomics. The protocol focuses on relevant methods for thermogenic stimulations and an arteriovenous blood sampling technique using Sulzer’s vein, which selectively drains interscapular BAT-derived venous blood and systemic arterial blood. Next, a gas chromatography-based metabolomics protocol using those blood samples is demonstrated. The use of this technique should expand the understanding of BAT-regulated metabolite exchange at the inter-organ level by measuring the net uptake and release of metabolites by BAT.
Brown adipose tissue (BAT) possesses a unique energy expenditure property known as non-shivering thermogenesis (NST), which involves both mitochondrial uncoupling protein 1 (UCP1)-dependent and UCP1-independent mechanisms1,2,3,4,5. These distinctive characteristics implicate BAT in the regulation of systemic metabolism and the pathogenesis of metabolic diseases, including obesity, type 2 diabetes, cardiovascular disease, and cancer cachexia6,7,8. Recent retrospective studies have shown an inverse association between BAT mass and/or its metabolic activity with obesity, hyperglycemia, and cardiometabolic health in humans9,10,11.
Recently, BAT has been proposed as a metabolic sink responsible for maintaining NST, as it requires substantial amounts of circulating nutrients as thermogenic fuel6,7. Furthermore, BAT can generate and release bioactive factors, referred to as brown adipokines or BATokines, which act as endocrine and/or paracrine signals, indicating its active involvement in systems-level metabolic homeostasis12,13,14,15. Therefore, understanding BAT's nutrient metabolism should enhance our understanding of its pathophysiological significance in humans, beyond its conventional role as a thermoregulatory organ.
Metabolomic studies employing stable isotope tracers, in combination with classic nutrient uptake studies using non-metabolizable radiotracers, have significantly improved our understanding of which nutrients are preferentially taken up by BAT and how they are utilized16,17,18,19,20,21,22,23,24,25,26,27. For instance, radioactive tracer studies have demonstrated that cold-activated BAT takes up glucose, lipoprotein-bound fatty acids, and branched-chain amino acids16,17,18,19,20,21,22,23,27. Recent isotope tracing combined with metabolomic studies has allowed us to measure the metabolic fate and flux of these nutrients within tissues and cultured cells24,25,26,28,29,30. However, these analyses primarily focus on the individual utilization of nutrients, leaving us with limited knowledge of BAT's systems-level roles in organ metabolite exchange. Questions regarding the specific series of circulating nutrients consumed by BAT and their quantitative contributions in terms of carbon and nitrogen remain elusive. Additionally, the exploration of whether BAT can generate and release metabolite-derived BATokines (e.g., lipokines) using nutrients is just beginning12,13,14,15,31,32.
Arteriovenous blood analysis is a classic physiological approach used to assess the specific uptake or release of circulating molecules in organs/tissues. This technique has previously been applied to the interscapular BAT of rats to measure oxygen and several metabolites, thereby establishing BAT as the major site of adaptive thermogenesis with its catabolic potential33,34,35,36,37. Recently, an arteriovenous study using rat interscapular BAT was coupled with a trans-omics approach, leading to the identification of undiscovered BATokines released by thermogenically stimulated BAT38.
Recent advances in high-sensitivity gas chromatography- and liquid chromatography-mass spectrometry (GC-MS and LC-MS)-based metabolomics have reignited interest in arteriovenous studies for the quantitative analysis of organ-specific metabolite exchange39,40,41. These techniques, with their high resolving power and mass accuracy, enable the comprehensive analysis of a wide range of metabolites using small sample quantities.
In alignment with these advancements, a recent study successfully adapted arteriovenous metabolomics for studying BAT at the mouse level, enabling the quantitative analysis of metabolite exchange activities in BAT under different conditions42. This article presents a BAT-targeted arteriovenous metabolomics protocol using GC-MS in a C57BL/6J mouse model.
All experiments were conducted with the approval of the Sungkyunkwan University Institutional Animal Care and Use Committee (IACUC). Mice were housed in an IACUC-approved animal facility located in a clean room set at 22 °C and 45% humidity, following a daily 12 h light/dark cycle. They were kept in ventilated racks and had access to a standard chow diet ad libitum (comprising 60% carbohydrate, 16% protein, and 3% fat). Bedding and nesting materials were changed on a weekly basis. For this study, male C57BL/6J mice aged 12 weeks and weighing between 25 g and 30 g were utilized. These animals were sourced from a commercial supplier (see Table of Materials).
1. Modulation of metabolic activity of the brown adipose tissue through temperature acclimation and pharmacological stimulation
NOTE: Temperature acclimation over several days to weeks or pharmacological stimulation using β-adrenergic receptor agonists are commonly employed methods for modulating BAT activity1. Therefore, a concise overview of the method is provided below to enable readers to choose the appropriate approach as required. To obtain metabolically inactive (less thermogenic) BAT, a baseline warm temperature, referred to as thermoneutrality (28-30 °C), is selected for C57BL/6J mice. This range ensures that the mice do not need to expend extra energy to maintain a constant body temperature. To obtain metabolically modestly or highly active (thermogenic) BAT, mild cold (20-22 °C) or severe cold (6 °C) temperatures can be chosen, respectively. For the purposes of this experiment, mice were raised under standard housing conditions at 22 °C, which, although mildly cold for mice, did not involve any pharmacological stimulations.
2. Arteriovenous blood sampling
NOTE: Mice over 12-14 weeks are best recommended for arteriovenous blood sampling. Younger mice may not have sufficiently sized Sulzer's veins, a distinct blood vessel that specifically drains venous blood from the interscapular BAT46.
3. Metabolite extraction from serum and chemical derivatization
4. Metabolomics analysis using GC-MS
NOTE: Single quadruple GC-MS (see Table of Materials) was employed to measure the various serum metabolites including carbohydrates, amino acids, and TCA cycle intermediates in derivatized samples from the Sulzer's vein and the left ventricle. Other columns can alternatively be used, although the experimental settings including the temperature program may vary depending on the types of columns used.
Figure 1 illustrates the experimental scheme of BAT-optimized AV metabolomics. As mentioned in the Protocol section, to obtain differentially stimulated brown adipose tissues, mice undergo temperature acclimation using rodent incubators or receive pharmacological administration such as β-adrenergic receptor agonists. Subsequently, mice are anesthetized, and blood samples are collected for metabolomic analysis (Figure 1A). For blood sampling, venous blood specifically draining from interscapular BAT is collected via the Sulzer's vein, while arterial blood is directly collected from the left ventricle of the heart (Figure 1B).
Figure 2 represents the schematics of serum metabolomics using GC-MS (Figure 2). Briefly, the methanol solution is added to serum samples followed by centrifugation to remove serum proteins. The supernatant containing the metabolites was dried in a SpeedVac and the dried samples were subjected to two-step derivatization using methoxyamine (MOX) and MTBSTFA (N-tert-Butyldimethylsilyl-N-methyltrifluoroacetamide). The derivatization step aims for the substitution of polar functional groups of the metabolites with TBDMS ester, allowing for GC-MS detection of metabolites as TBDMS derivatives. The derivatized samples were analyzed using single quadruple GC-MS. Metabolites are separated in the column according to their physiochemical properties. Electron ionization (EI) is employed to break the metabolites down into unique fragment ions detected by a mass detector. Compound identification is carried out through the detection of compound-specific fragment ions. The m/z value and retention time of the compound-specific fragment ions in the experiment are shown in the table (Table 1). An extracted ion chromatogram (EIC) is used to integrate the signal of a compound-specific fragment ion in the peak area of the metabolite, which defines the abundance of metabolites.
To determine whether BAT releases or uptakes a series of metabolites, Log2 ratios of ion counts between the Sulzer's vein and left ventricle (Log2(SV/LV)) can be calculated (Figure 3). If the Log2(SV/LV) value for a particular metabolite is >0, it indicates that the metabolite is more abundant in the SV than LV, suggesting that interscapular BAT net releases the metabolite. Conversely, if the Log2(SV/LV) value for a certain metabolite is <0, it suggests that the metabolite is taken up by the interscapular BAT.
Using this approach, it was found that mild cold-stimulated BAT significantly consumes glucose, lactate, 3-hydroxybutyrate (3-HB), BCAAs, aspartate, lysine, and tryptophan, while BAT significantly releases succinate and palmitate (Figure 3A). Most of these phenomena were observed in our previous studies on chronic cold BAT42, suggesting that mild cold BAT metabolically resembles chronic cold BAT, albeit to a lesser extent. To provide a quick visual overview of the data, a heat map displaying Log2(SV/LV) values is depicted (Figure 3B). It is essential to interpret the data with caution, as the heatmap represents Z-scores within the group.
Figure 1: Experimental scheme of BAT-optimized arteriovenous (AV) metabolomics. (A) Diagram showing the workflow of AV metabolomics. (1) To induce different versions of metabolically stimulated brown adipose tissue (BAT), mice are acclimated to different temperatures or receive pharmacological drug injections. (2) Subsequently, the arterial and venous blood samples are collected from the left ventricle and the Sulzer's vein of the mice, respectively. (3) The obtained serum samples are subjected to metabolite extraction, followed by metabolomic analysis. (B) Representative images providing guidance for blood sampling procedures. Arterial blood is collected from the left ventricle, while venous blood is obtained from the Sulzer's vein as it specifically drains blood from BAT. Please click here to view a larger version of this figure.
Figure 2: GC/MS-based targeted metabolomics. Experimental scheme of GC/MS-based targeted metabolomics. TBDMS: tert-butyldimethylsilyl, EI: Electron ionization, EIC: Extract ion chromatogram. Please click here to view a larger version of this figure.
Figure 3: Representative AV metabolomics data. (A) The uptake and subsequent release of selected prevalent and highly active circulating fuel metabolites by BAT, classified according to their respective groups. The data show Log2 Sulzer's vein to left ventricle (SV/LV) ratios. Bars indicating metabolites with mean Log2 (SV/LV) <0 are colored red, and Log2 (SV/LV) >0 are colored blue. Individual data points represent each mouse; n = 11. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. BCAA; branched-chain amino acid; EAA; essential amino acid; NEAA; non-essential amino acid. (B) A heat map illustrating the differential abundance of metabolites between Sulzer's vein (SV) and left ventricle (LV) blood. Each column shows a Z-score of average values within the group; n = 11. Please click here to view a larger version of this figure.
Compound | m/z of Fragment ions | Retention time | Formular | TBDMS derivative formular |
Pyruvate | 174.1 | 8.72 | C3H4O3 | C15H33O3NSi2 |
Aspartate | 418.2 | 18.45 | C4H7NO4 | C22H49NO4Si3 |
α-Ketoglutarate | 346.2 | 17.27 | C5H6O5 | C18H37NO5Si2 |
C16:0(Palmitate) | 313.3 | 19.72 | C16H32O2 | C22H46O2Si |
Lactate | 261.2 | 11.66 | C3H6O3 | C15H34O3Si2 |
Leucine | 200.2 | 14 | C6H13NO2 | C18H41NO2Si2 |
Isoleucine | 200.2 | 14.34 | C6H13NO2 | C18H41NO2Si2 |
Valine | 186.2 | 13.53 | C5H11NO2 | C17H39NO2Si2 |
Alanine | 260.2 | 12.17 | C3H7NO2 | C15H35NO2Si2 |
Serine | 390.2 | 17.04 | C3H7NO3 | C21H49NO3Si3 |
Glycine | 218.1 | 12.43 | C2H5NO2 | C14H33NO2Si2 |
Lysine | 300.2 | 20.48 | C6H14N2O2 | C24H56N2O2Si3 |
Succinate | 289.1 | 14.65 | C4H6O4 | C16H34O4Si2 |
Proline | 258.2 | 14.73 | C5H9NO2 | C17H37NO2Si2 |
Phenylalanine | 336.2 | 18 | C9H11NO2 | C21H39NO2Si2 |
Methionine | 320.2 | 16.84 | C5H11NO2S | C17H39NO2SSi2 |
Cysteine | 406.2 | 19 | C3H7NO2S | C21H49NO2SSi3 |
Asparagine | 417.2 | 19.86 | C4H8N2O3 | C22H50N2O3Si3 |
3-Hydroxybutyrate(3HB) | 275.2 | 12.81 | C4H8O3 | C16H36O3Si2 |
Tryptophan | 375.2 | 22.97 | C11H12N2O2 | C23H40N2O2Si2 |
Malate | 419.2 | 18.19 | C4H6O5 | C22H48O5Si3 |
Glutamate | 432.3 | 19.59 | C5H9NO4 | C23H51NO4Si3 |
Glutamine | 431.3 | 20.85 | C5H10N2O3 | C23H52N2O3Si3 |
Citrate | 459.2 | 22.38 | C6H8O7 | C30H64O7Si4 |
Glucose | 476.3 | 22.7 | C6H12O6 | C30H68O6Si4 |
Table 1: GC/MS compound fragment ions used for integration. This table contains a list of 25 serum metabolites identified in the present method, with the retention time and fragment ions corresponding to each compound.
A critical step in understanding the metabolic potential of BAT in whole-body energy balance is to define which nutrients it consumes, how they are metabolically processed, and what metabolites are released into the circulation. This protocol introduces a specialized arteriovenous sampling technique that enables access to the venous vasculature of interscapular BAT and systemic arterial vasculature in C57BL/6J mice, which was recently developed and validated by Park et al42. Below are key points you should critically consider when following the protocol.
For thermogenic stimulation, a key consideration is choosing the most appropriate type of thermogenic stimulation for your experimental setting. For example, when using an experimental group (vs. control) and aiming to observe differences in the metabolic adaptation capacity (crucial for its maximal non-shivering thermogenic potential) of BAT, chronic cold acclimation, or β-adrenergic receptor stimulation (lasting more than 3 days to 4 weeks) is suitable1,45. When BAT is exposed to acute cold exposure (a few hours less than 24 h), rather than metabolically adapting, BAT metabolizes its intrinsic stored fuels and communicates metabolically with shivering muscles for optimal thermogenesis1. A β-adrenergic receptor agonist represents a neuronal response to trigger BAT thermogenesis1,14, and there are other endocrine factors that enhance the thermogenic program55,56,57,58,59,60. Whenever performing thermogenic stimulation as mentioned above, having a baseline control at thermoneutrality (30 °C) should be useful to confirm whether the stimulation worked appropriately.
As mentioned in the protocol section, obtaining arteriovenous blood of the highest quality is key to determining successful establishment of the experiment. To achieve this, it is essential to prevent hemolysis (see step 2.5) and collect a consistent amount of blood to avoid dilution of the distinct characteristics of venous blood from the Sulzer's vein (less than 40 µL from the Sulzer's vein). To this end, during the initial set-up, the authors strongly recommend performing a pilot experiment for metabolite measurement using not only blood from Sulzer's vein, but also blood from non-BAT draining venous blood (e.g. inferior vena cava). This will help confirm whether the serum of the Sulzer's vein reveals distinct metabolite profiling compared to the non-BAT venous vessels. Regarding arterial blood collection from the left ventricle, it is important to consistently puncture the same area on the left ventricle to minimize variations in cardiokine levels61,62. Such variations could yield confusing results in arteriovenous-metabolomics or -proteomics.
Chromatography coupled with mass spectrometry is one of the most powerful tools for metabolomics, one method being GC-MS63. Indeed, the measurement of biologically relevant metabolites using GC-MS was considered challenging due to the coverage issue; GC-MS application is usually limited to the analysis of metabolites with highly volatile and non-polar properties as opposed to the numerous polar metabolites in tissues and biological fluids. However, the development of various extraction and derivatization methods, together with the upsides of GC-MS including excellent peak resolution, reproducibility, and extensive mass spectral libraries allowing for handy peak identification, has led to such a platform as an attractive option for analyzing biologically relevant metabolites64,65.
In this protocol, the GC-MS platform is employed for the analysis of 25 metabolites in serum collected using the aforementioned technique. To this end, metabolite extraction was carried out using 80% methanol solution. Of note, during this step, using certain types of tubes including Eppendorf tubes is recommended to achieve the lowest levels of organic substance contamination in the samples. Careful separation of supernatant from pellets following extraction is also needed to remove protein debris from the serum, which is critical for the maintenance of MS-ion source performance.
The derivatization step can be diversified based on the types of derivatizations including aryl derivatives, silylation, and acylation66. The authors employed two-step derivatization with N-methyl-N-tert-butyldimethylsilylrifluoroacetamide (MTBSTFA) for metabolite sialylation which is a commonly used method with the advantage of stable detection of biologically relevant polar metabolites including sugars, but with a drawback regarding the occasional loss of the N-trimethylsilyl group of amines and amino acids during the analysis step64. Of note, samples should be completely dried prior to the derivatization step, due to the moisture sensitivity of the reagents and derivatives.
One of the main limitations of the GC analysis, in general, still comes down to the range of detectable metabolites despite the aforementioned derivatization steps which improve GC-MS capacity regarding polar metabolite detection, especially as compared to LC-MS-based analysis. Quantitative metabolomics using LC-MS can greatly help gain better insight into the metabolic spectrum of BATs with respect to systemic metabolite exchange24,42.
Although the calculation used in the protocol-measurement of metabolite concentration gradient between the arterial vs. venous blood; Log2(SV/LV)-quantifies fractional absorption and release change by BAT, caution is needed for data interpretation. In particular, for those metabolites whose fractional net exchange value is '0' (e.g. citrate, αKG, malate, methionine, alanine, glutamine), the result could be attributed to either no uptake/release or equally active catabolism and synthesis (where both uptake and release are high). To verify which of the two possibilities is correct, additional metabolic flux experiments using isotope tracing in vivo with the taken-up metabolite and/or its substrate are necessary39.
The limitation of the fractional calculation is that it does not provide the quantitative landscape of the uptake and release for those metabolites42,67. For instance, although the Log2(SV/LV) value indicates a comparable relative net uptake between glucose and 3-hydroxybutyrate (3HB), the actual contribution of glucose to the carbon sources should be much higher than 3-HB due to the fact that the blood concentration of glucose is much higher than that of 3HB, and because glucose contains four more carbons (glucose has six carbons; 3HB has two carbons). If required, a comprehensive analysis incorporating additional parameters, such as blood flow rate, actual blood concentrations of each metabolite, and their chemical equations should inform us of the quantitative contributions of the metabolites taken-up/released and their contributions in total carbon or nitrogen flux42,67.
A major advantage of this miniaturized methodology for arteriovenous blood sampling and metabolomic analysis at the mouse level is its synergistic combination with various genetic models to study brown adipose tissue metabolism and physiology (e.g. UCP1-KO, UCP1-Cre driven BAT-specific loss-of-function models, transgenic models). Recent metabolomic analysis which requires only trace amounts of serum for metabolomic profiling makes this approach more feasible with smaller organisms (see protocol 3)39. However, proteomic analysis, which can provide better insights when considered together with metabolomics, may require larger amounts of samples. In this regard, conventional arteriovenous sampling using rats may be more suitable. We refer the readers to the most recent protocol for arteriovenous blood sampling for BAT in rats, written by Mestres-Arenas and colleagues47.
Although the current protocol adopts the anesthesia procedure using isoflurane as described in Park et al.42, this approach may negatively impact the thermogenic metabolism of brown adipocytes and/or brown adipose tissue in vivo68,69,70,71. Therefore, future arteriovenous metabolomics utilizing pentobarbital warrant investigation.
In summary, this protocol provides a foundational methodology to measure BAT-specific net metabolic activity (consumption vs. production) upon various thermogenic stimulations. This should give us valuable insights into the role of BAT as a systemic nutrient sink as well as provider by quantitatively listing key metabolic fuels as well as secreted metabolites. Moreover, this can also be useful for identifying previously unstudied metabolite-derived brown adipokines, especially when operated with discovery-based metabolomic platforms.
The authors have nothing to disclose.
We thank all members of the Choi and Jung laboratories for methodological discussion. We thank C. Jang and D. Guertin for advice and feedback. We thank M.S. Choi for critical reading of the manuscript. This work was funded by NRF-2022R1C1C1012034 to S.M.J.; NRF-2022R1C1C1007023 to D.W.C; NRF-2022R1A4A3024551 to S.M.J. and D.W.C. This work was supported by Chungnam National University for W.T.K. Figure 1 and Figure 2 were created using BioRender (http://biorender.com/).
0.5-20 µL Filter Tips | Axygen | AX.TF-20-R-S | |
1 mL Syringe with attached needle – 26 G 5/8" | BD Biosciences | 309597 | |
Agilent 5977B GC/MSD (mass selective detector) | Agilent | G7077B | |
Agilent 7693A Autosampler | Agilent | G4513A | |
Agilent 8890 GC System | Agilent | G3542A | |
Agilent J&W GC column (Capilary column) HP-5MS UI | Agilent | 19091S-433UI | |
Agilent MassHunter Workstation software_MS Quantitative analysis(Quant-My-way) | Agilent | G3335-90240 | |
C57BL/6J mouse | DBL | C57BL/6JBomTac | |
CentriVap -50 °C Cold Trap (with Stainless steel Lid) | LABCONCO | 7811041 | |
DL-Norvaline | Sigma-Aldrich | N7502-25G | |
Eppendorf centrifuge 5430R | Eppendorf | 5428000210 | |
Eppendorf Safe-Lock Tubes 1.5 mL | Eppendorf | 30120086 | |
Glass insert 250 μL | Agilent | 5181-1270 | |
Methanol (LC-MS grade) | Sigma-Aldrich | Q34966-1L | |
Methoxyamine hydrochloride | Sigma-Aldrich | 226904-5G | |
Microvette 200 Serum, 200 µL, cap red, flat base | Sarstedt | 20.1290.100 | |
MTBSTFA | Sigma-Aldrich | 394882-100ML | |
Pyridine(anhydrous, 99.8%) | Sigma-Aldrich | 270970-100ML | |
Refrigerated CentriVap Complete Vaccum Concentrators | LABCONCO | 7310041 | |
Rodent diet | SAFE | SAFE R+40-10 | |
Rodent incubator | Power scientific | RIT33SD | |
Ultra-Fine Pen Needles – 29 G 1/2" | BD Biosciences | 328203 | |
Vial Cap 9 mm | Agilent | 5190-9067 | |
Vial, ambr scrw wrtn 2 mL | Agilent | 5190-9063 | |
Vial, ambr scrw wrtn 2 mL+A2:C40 | Axygen | PCR-02-C |