Here, we present an inexpensive quantitative method utilizing deuterium oxide and gas chromatography mass spectrometry (GCMS) for the analysis of total fatty acid de novo lipogenesis in brown adipose tissue in vivo.
Fatty acid synthesis is a complex and highly energy demanding metabolic pathway with important functional roles in the control of whole-body metabolic homeostasis and other physiological and pathological processes. Contrary to other key metabolic pathways, such as glucose disposal, fatty acid synthesis is not routinely functionally assessed, leading to incomplete interpretations of metabolic status. In addition, there is a lack of publicly available detailed protocols suitable for newcomers in the field. Here, we describe an inexpensive quantitative method utilizing deuterium oxide and gas chromatography mass spectrometry (GCMS) for the analysis of total fatty acid de novo synthesis in brown adipose tissue in vivo. This method measures the synthesis of the products of fatty acid synthase independently of a carbon source, and it is potentially useful for virtually any tissue, in any mouse model, and under any external perturbation. Details on the sample preparation for GCMS and downstream calculations are provided. We focus on the analysis of brown fat due to its high levels of de novo fatty acid synthesis and critical roles in maintaining metabolic homeostasis.
Obesity and associated metabolic diseases are a pandemic that endanger present and future generations1,2. Commonly simplified as the consequence of the imbalance between energy intake and expenditure, the metabolic dysregulation associated with obesity affects a large number of metabolic pathways controlled by environmental and endogenous factors3. However, only a few pathways are routinely tested in animal models of metabolic dysregulation.
As an example, glucose disposal is routinely measured by glucose and insulin tolerance tests, probably due to the simplicity of using portable glucose monitors4. Whole body glucose and lipid oxidation relative rates are also routinely estimated based on the respiratory exchange ratio from indirect calorimetry assays5,6. However, the majority of all other aspects of metabolism are not routinely functionally assessed. This leads to incomplete interpretations of the metabolic status and missed therapeutic options. One of the main such pathways is de novo lipogenesis.
De novo lipogenesis (DNL) is the process by which new fatty acids are generated from precursors. Glucose is considered to be the main precursor contributing to whole-body DNL7, however other precursors, such as acetate, fructose, lactate, and branched chain amino acids, have been shown to be relevant carbon sources in a spatial and condition dependent manner8,9,10,11,12. DNL is a key contributor to metabolic homeostasis and is essential for normal development13. Additionally, alterations in DNL have been associated with cancer14,15 and metabolic16,17,18 and cardiovascular diseases19,20.
The DNL pathway is composed of the core enzymatic components ATP citrate lyase (ACLY), acetyl-CoA carboxylase (ACC1/2), and fatty acid synthase (FAS) which primarily produce palmitate, a 16-carbon saturated fatty acid. However, odd chain and branched chain fatty acids can also be produced at lower rates9. Elongases and desaturases further modify these fatty acids, creating a diverse range of fatty acids species useful for a variety of functions (e.g., long-term energy storage and manipulation of membrane fluidity).
The expression of the DNL enzymatic machinery is controlled by a short number of transcription factors. The most well described to date include the sterol regulatory element binding protein (SREBP) family, carbohydrate response element binding protein (ChREBP), and liver X receptor (LXR)21,22,23,24,25,26. Despite an apparent overlap in their functions, individual regulations based on cell type dominancy and physiological or pathological conditions have been reported21,22,27,28.
Remarkably, a number of inhibitors for selected steps of the DNL pathway have been approved for clinical use or are in the preclinical or clinical stages of development for a number of diseases, including obesity, nonalcoholic fatty liver disease/nonalcoholic steatohepatitis (NAFLD/NASH), and cardiovascular disease29. These efforts highlight the relevance of DNL in health and disease.
In recent years, the employment of methods to quantitatively assess de novo fatty acid synthesis has increased30. The most common method for assessing this is the use of heavy labelled water (D2O), where the heavy labelled hydrogen gets incorporated into acyl chains during synthesis both directly and indirectly, via deuterium exchange with the hydrogens of the DNL substrates NAPDH, acetyl-CoA, and malonyl-CoA. Although this approach is gaining in popularity, there is a lack of publicly available detailed protocols suitable for newcomers in the field. Here, we outline a method for quantitatively assessing the de novo synthesis of products of FAS using D2O and gas chromatography mass spectrometry (GCMS), with calculations previously developed by Lee et al.31. This method measures de novo fatty acid synthesis independently of a carbon source, and it is potentially useful for virtually any tissue, in any mouse model, and under any external perturbation. Here, we focus on the analysis of brown adipose tissue (BAT) due to its high levels of DNL and critical roles in maintaining metabolic homeostasis.
All experiments were approved by the Institutional Animal Care and Use Committee at Cincinnati Children's Hospital Medical Center.
1. Preparation of D2O
NOTE: To avoid experimental variation, prepare sufficient solution/drinking water for all mice for the duration of the experiment.
2. Modulation of BAT activity by temperature acclimation
3. Administration of D2O
4. Plasma and tissue collection, processing, and storage
5. Lipid extraction from adipose tissue
6. Preparation of fatty acid methyl esters (FAMEs) and GCMS analysis
7. Deuterium acetone exchange of plasma samples to determine body water enrichment
8. In vivo de novo lipogenesis calculations
Based on the D2O dosing described in step 1, we typically find that body water is enriched in the range of 2.5% to 6%, and that a baseline level of deuterium enrichment in body water is rapidly achieved in 1 h and maintained for the duration of the study via 8% enriched drinking water (Figure 1). Continuous steady state body water enrichment is an assumption of the calculations used in step 6, and so we recommend experimental validation of the kinetics of body water enrichment in new experimental models.
We have found that the amount of de novo synthesized fatty acids is increased at room temperature, relative to those at thermoneutrality in BAT (Figure 2). The mass isotopologue distribution of palmitate in BAT from mice at thermoneutrality and room temperature after 3 days of D2O administration shows a higher M1 and M2 deuterium enrichment found at room temperature (Figure 2A). This colder temperature enrichment does not only happen in palmitate, but also in a broad range of fatty acids in BAT (Figure 2B). Total palmitate abundance is also increased in BAT of mice at room temperature, and combining the fractional synthesis rate with the total palmitate levels, we found that total palmitate synthesis was increased in mice at room temperature (Figure 2C,D). Notably, plasma total fatty acid enrichment does not follow the same trend as BAT, but instead, fatty acid enrichment is increased with thermoneutrality (Figure 2E). Deconvoluting potential uptake from endogenous synthesis is not possible with a D2O single time-point approach, as described here, but these opposing trends indicate that the fatty acid enrichment pattern in BAT is not being driven by fatty acid uptake.
Figure 1: Percentage of D2O enrichment in plasma of mice over multiple time points, following injection with 0.035 ml/g body weight 0.9% NaCl D2O and replacement of drinking water with 8% D2O enriched water. Bar graphs represent mean ± SD. n = 9. Please click here to view a larger version of this figure.
Figure 2: Representative results in mice brown adipose tissue after 3 days of D2O administration. (A) Mass isotopologue distribution of palmitate in BAT after 3 days of D2O administration at room temperature (RT) and thermoneutrality (TN). (B) BAT molar enrichment of a range of fatty acids after 3 days of D2O administration at room temperature and thermoneutrality. (C) Total abundance and the (D) de novo synthesized palmitate in brown adipose tissue after 3 days of D2O administration at room temperature and thermoneutrality. (E) Plasma molar enrichment of a range of fatty acids after 3 days of D2O administration at room temperature and thermoneutrality. Bar graphs represent mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. n = 6. Statistical analysis was determined by two-tailed Students t-test. Please click here to view a larger version of this figure.
Compound | Ions | Estimated elution time (minutes) | Formula for isotope correction |
C16:0 | 270-275 | 20 | C17H34O2 |
C14:0 | 242-247 | 16.5 | C15H30O2 |
C15:0 | 256-261 | 18 | C16H32O2 |
Iso-C16:0 | 270-275 | 18.9 | C17H34O2 |
Iso-C17:0 | 284-289 | 21 | C18H36O2 |
Anteiso-C17:0 | 284-289 | 21.5 | C18H36O2 |
C16 D31 | 301 | 19.2 | ~ |
Acetone | 58-59 | 1.5 | C3H6O |
Table 1: GCMS compound fragment ions to integrate. This table covers a range of fatty acids produced by mammalian fatty acid synthase, but C16:0 is the primary product. All retention times are for the GCMS method detailed in step 6 except acetone, which is detailed in step 7.
Supplementary File: Example spreadsheet calculation. Please click here to download this File.
Understanding the balance and interaction between complex metabolic pathways is an indispensable step toward understanding the biological basis of metabolic related diseases. Here, we show a noninvasive and inexpensive methodology to determine changes in de novo fatty acid synthesis. This method is adapted from previously published methods which developed calculations for estimating de novo synthesis flux from fatty acid deuterium enrichment31 and for using deuterium-acetone exchange for determining the relative percentage of D2O in body water39. In recent years, we have taken the method described here and applied it to uncover changes in the synthesis of diverse fatty acids in adipose depots, the brain, liver, and tumors9,22,40. There are a number of critical steps and modifiable aspects of this protocol. These are related to the overall experimental design as well as the analytical approach and calculations utilized in this method. Alternate approaches for calculating de novo fatty acid synthesis from stable isotope tracers are also widely utilized. These include the alternate mass isotopomer distribution analysis (MIDA) approach developed by Hellerstein et al.41,42, isotopomer spectral analysis (ISA) developed by Kelleher et al.43,44, and newer approaches that have been developed to model the synthesis of longer chain fatty acids such as fatty acid source analysis (FASA)45. The benefit of the method here is that it is easily implementable with freely available software and the use of a spreadsheet. However, it is limited by the need for body water enrichment (BWE) to be at a steady state, caveats associated with the utilization of the N parameter (discussed below), and that it is only applicable to direct products of FAS which is primarily palmitate.
Brown fat DNL and temperature acclimation
DNL has surged as a main nodule of whole-body metabolic regulation and is essential for normal development13,29. Tissue targeted deletion and metabolomic analysis have pinpointed several key enzymatic and transcriptional players that control DNL in adipose tissue, and how it controls whole body fat accretion and normal metabolic homeostasis21,28,46,47,48,49. Although less appreciated, BAT is a highly active tissue in DNL, with a higher expression of core body DNL enzyme machinery (acly, acaca, fasn) than most other tissues and higher lipogenic activities22,50,51,52,53,54,55. BAT is however unique, in the sense that acclimation to sub-thermoneutral temperatures simultaneously induces DNL and fatty acid oxidation activity in BAT22,54,56. Although the mechanisms associated are not fully clear, it is accepted that BAT is a metabolic sink for nutrients, including those destined to DNL for lipid synthesis, which are essential for normal BAT activity53,55. As such, it is relevant to be able to account for all metabolic incomes and outcomes of BAT to assess its activity levels. However, DNL is usually only measured based on gene expression and functionally assessed on counted occasions. The protocol proposed here allows for the functional assessment of DNL, allowing a more complete interpretation of metabolic status.
D2O dose, timing, and body water enrichment measurement
One of the primary critical steps in the design is the dose and timing of the D2O administration. In this protocol, we employ a dose and administration approach (bolus injection followed by drinking water enrichment) that we have found leads to rapid achievement and maintenance of steady state BWE. The calculations used to determine fatty acid synthesis are based on the assumption of steady state BWE; thus, it is crucial that this is validated in new experimental models. If steady state enrichment is not possible due to problems with using this combined dosing approach, the reader is referred to other publications where alterations to the calculations allow for this39. The degree of variation from the steady state that can be tolerated, when employing a dosing approach that aims for rapid achievement of this depends on many factors. If BWE is increased in sustained spikes above the value used for calculation, then by using the approach here, DNL will be overestimated. If BWE is significantly lower than the value used for calculation for sustained amounts of time, DNL will be underestimated. While some slight variation around the average is expected, the most critical factor is that BWE does not vary more in one experimental group than the other, which could lead to differences in the calculated DNL being driven by variation in BWE. If both experimental groups display some slight but similar variation over time, the tolerance of this will depend on the accuracy and precision with which the investigator requires the measurement to perform, as well as the expected differences between experimental groups. For example, in the case of brown adipose tissue exposed to thermoneutrality versus room temperature, which was provided as an example here, the difference in the amount of de novo synthesized fatty acid present is greater than 50%; thus, small variations in BWE are unlikely to obscure the results.
In addition to consideration of the steady state, another modifiable aspect of this protocol is the level of BWE. Increased BWE potentially allows for increased label transfer onto fatty acids, and thus, fatty acid pools with lower turnover may be more easily detectable in a shorter amount of time. This could be an advantage in situations where the researcher is interested in acute responses in DNL. However, increased D2O levels in body water may cause unwanted physiological effects, and thus caution should be taken57. Administration of lower levels of D2O to achieve BWE levels typically around 0.5% is commonly employed in human studies due to the cost associated with D2O administration. In this case, methods with increased sensitivity may be needed to quantify BWE and fatty acid enrichment. For BWE, this may be achieved via ion ratio mass spectrometry58 or a modified acetone exchange protocol59. For fatty acid analysis, the use of a high resolution GCMS instrument has recently been shown to increase the sensitivity of the measurement of deuterium in fatty acids, and thus allow quantitation of DNL following very short time periods and/or low BWE60. Alternative methods can also be used to increase the throughput with which BWE is measured, by utilizing headspace analyses of the acetone generated in step 7.1.5 instead of extracting it with CHCl3 and performing liquid injection. This decreases the method run time and avoids multiple transfer steps, thus reducing sample prep time. If access to a headspace injector is available, this method is highly recommended61.
Choice of fatty acid pool to quantify and analytical approach
In this protocol, the method is designed to measure the total fatty acid pool in the tissue or plasma, irrespective of lipid class. However, lipid classes may be separated prior to FAME generation, via methods such as thin layer chromatography or liquid chromatography. In addition, when analyzing serum or plasma, lipids from specific lipoprotein fractions, such as very low density lipoprotein (VLDL), may be isolated via ultracentrifugation58. This method is commonly employed in human studies in order to ascertain hepatic DNL, which is likely to be the primary source of de novo made fatty acids in VLDL. The lipid extraction and derivatization protocol may also be modified to enhance the isolation of specific classes62. Finally, deuterium enrichment of lipid hydrogens can also be measured using 2H NMR spectroscopy rather than mass spectrometry. Although this method is less sensitive than MS and so requires increased material, its advantage is that it gives positional enrichment information, which can be used to better estimate elongation and desaturation rates63.
Calculation
The calculations in this protocol are based on previous calculations that take into account steady state BWE, fatty acid enrichment, and the number of exchangeable hydrogens on the fatty acids31. As with any calculations, there are a number of limitations and assumptions that should be considered when applying and interpreting the results from this approach. One of the primary considerations is the value used for N in section 8 of the method. During fatty acid synthesis, 2H from D2O is incorporated into the acyl chain, as well as indirectly via exchange with hydrogens on NADPH, acetyl-CoA, and malonyl-CoA64,65. Previous studies have shown that this exchange is incomplete31,66; thus, the calculation incorporates a correction factor (N) to allow for this65. A N of 22 is commonly used for many studies, as it has previously been found that this is appropriate for a range of tissues using the equation outlined in step 631,66. However, this may vary based on the experimental perturbation and animal model65; thus, we urge investigators to take this into consideration. A lower N increases the total synthesis flux, and the value employed for this parameter should be considered when comparing measurements across studies or labs. A limitation of the approach used here is that it is only relevant for the analysis of direct products of FAS which is primarily palmitate, with much lower amounts of odd chain fatty acids and mono-methyl branched chain fatty acids9. Although deuterium enrichment of all fatty acids can be detected, the formula used here is not appropriate for determining the synthesis of longer chain fatty acids, such as C18:0, as it does not allow for the uptake of unlabelled fatty acid and the elongation of these67. Similarly, desaturation rates may also be underestimated.
General limitations
Although the use of D2O for measuring fatty acids has generated crucial insight into the importance of DNL in physiological homeostasis, there are a number of limitations associated with this methodology. First, based on the timing outlined here, it is not possible to be completely confident in the tissue of origin of the newly synthesized fatty acids. It is easy to imagine a scenario in which a certain tissue generates fatty acids from DNL, and then those are transported to other tissues. Fine time course experiments after D2O treatment could increase the resolution of the tissue of origin for newly synthesized fatty acids, decreasing cross-tissue contamination. The 3 day time point which was utilized for the representative results was designed to detect deuterium enrichment in a range of fatty acids at both thermoneutrality (when flux is lower) and room temperature. Shorter time points post D2O administration, where palmitate is the primary target, minimizes cross-tissue fatty acid transport and so may be advantageous. Much shorter time points (e.g., 12 h) would be ideal when considering mice in cold conditions and when thermoneutrality is not included in the experimental design. Second, this method does not provide information about the substrate of origin of the new fatty acids. Identifying the specific substrates used in particular circumstances can be an additional layer of information necessary for full understanding of the regulatory map of DNL. This can be done with substrates labelled with stable isotopes.
Benefits
A particular benefit of this methodology is that animals are unrestrained (except for the initial injection of D2O) and conscious during the procedure, promoting a low stress environment for animals to behave naturally, including desired ambulation and a feeding pattern and quantity allowing for specific diet treatment if necessary. D2O is also remarkably easy to administer. Contrarily to liquid solutions of other tracers (e.g., 13C-U-glucose), D2O does not have distinct palatable features likely to impact water consumption. Beyond stable isotope tracers, the other most commonly used method is radioisotopes. The potential advantage of radioisotopes is choice of equipment for detection. A mass spectrometer is required for stable isotopes, whereas radioisotopes can be detected using a scintillation counter, which may be more easily accessible. However, there are a number of disadvantages associated with radioisotopes due to safety and ethical issues. In addition, the radioisotope is generally on glucose and so does not solely reflect changes in DNL, but rather changes in glucose derived DNL. In addition, determination of the synthesis of individual fatty acids is not possible. D2O more rapidly equilibrates across all tissues compared to specific substrates with isotope labels68.
DNL is a key component of metabolic homeostasis controlled by a number of enzymes and transcription factors that each independently affect development, metabolism, and disease states. Thus, DNL is bound to be a major metabolic pathway, that will need to be interrogated more broadly in research and therapy development. This protocol may be used as a first step forward in incorporating DNL analysis routinely in metabolic phenotyping.
The authors have nothing to disclose.
We thank the Sanchez-Gurmaches and Wallace lab members for valuable discussions. This work was supported by grants from the American Heart Association (18CDA34080527 to JSG and 19POST34380545 to RM), the NIH (R21OD031907 to JSG), a CCHMC Trustee Award, a CCHMC Center for Pediatric Genomics Award, and a CCHMC Center for Mendelian Genomics & Therapeutics Award. This work was supported in part by NIH P30 DK078392 of the Digestive Diseases Research Core Center in Cincinnati. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. RT and MW were supported by a UCD Ad Astra Fellowship.
4 mL Glass Vials | Fisher Scientific | 14-955-334 | |
0.2 µm filter | Olympus Plastic | 25-244 | |
26G needeled syringes | BD | 309597 | |
Acetone | Merck | 34850 | |
Acetonitrile | Merck | 900667 | |
Blue GC screw cap with septa | Agilent | 5190-1599 | |
Centrifuge | Eppendorf | 5424R | |
Chloroform | Sigma | 366927 | |
Deuterium oxide | Sigma | 151882 | |
Di-tert-butyl-4-methylphenol (BHT) Select FAME Column |
Merck | B1378 | |
Di-tert-butyl-4-methylphenol (BHT) Select FAME Column |
Agilent | CP7419 | |
EDTA tube | Sarstedt | 411395105 | |
Ethanol | Merck | 51976 | |
Hexadecenoic-d31 Acid | Larodan | 71-1631 | |
Hexane | Merck | 34859 | |
Methanol | Merck | 34860 | |
Microcentrifuge tube | Olympus Plastic | 24-282 | |
Mouse environmental chamber | Caron | Caron 7001-33 | |
Potasium Chloride | Fisher Bioreagents | BP366-500 | |
Potasium Phosphate | MP Biomedicals | 194727 | |
SafeLock microcentrifuge tubes | Eppendorf | 30120086 | |
Screw top amber GC vial | Agilent | 5182-0716 | |
Sodium Chloride | Fisher Bioreagents | BP358-212 | |
Sodium Hydroxide | Merck | S5881 | |
Sodium Phosphate, dibasic | Fisher Bioreagents | BP332-500 | |
Sodium Sulfate | Merck | 239313 | |
Sulfuric Acid | Merck | 258105 | |
Vial insert | Agilent | 5183-2088 |