Triglyceride lipolysis in adipocytes is an important metabolic process resulting in the liberation of free fatty acids and glycerol. Here, we provide a detailed protocol to measure basal and stimulated lipolysis in adipocytes and ex vivo adipose tissue from mice.
Adipocytes store energy in the form of triglycerides in lipid droplets. This energy can be mobilized via lipolysis, where the fatty acid side chains are sequentially cleaved from the glycerol backbone, resulting in the release of free fatty acids and glycerol. Due to the low expression of glycerol kinase in white adipocytes, glycerol re-uptake rates are negligible, while fatty acid re-uptake is dictated by the fatty acid binding capacity of media components such as albumin. Both glycerol and fatty acid release into media can be quantified by colorimetric assays to determine the lipolytic rate. By measuring these factors at multiple time points, one can determine the linear rate of lipolysis with high confidence. Here, we provide a detailed protocol for the measurement of lipolysis in in vitro differentiated adipocytes and ex vivo adipose tissue from mice. This protocol may also be optimized for other preadipocyte cell lines or adipose tissue from other organisms; considerations and optimization parameters are discussed. This protocol is designed to be useful in determining and comparing the rate of adipocyte lipolysis between mouse models and treatments.
Excess nutrients are stored in white adipose tissue in the form of triglycerides in the neutral lipid core of lipid droplets. Triglyceride stores are mobilized via lipolysis, a process by which the fatty acid side chains are sequentially cleaved by adipose tissue triglyceride lipase (ATGL), hormone-sensitive lipase (HSL), and monoglyceride lipase (MGL), resulting in the release of free fatty acids (FFAs) and the glycerol backbone1,2. Lipolysis is activated by catecholamine signaling in the adipose tissue. Sympathetic nerve terminals locally release catecholamines, which bind to β-adrenergic receptors on the adipocyte plasma membrane. Upon ligand binding, these G-protein coupled receptors (GPCRs) activate adenylyl cyclase via Gαs. Subsequent activation of protein kinase A (PKA) by cAMP results in the upregulation of both ATGL and HSL. The phosphorylation of perilipin-1 by PKA causes the dissociation of ABHD5 (also known as CGI-58), which binds and coactivates ATGL3. PKA directly phosphorylates HSL, promoting its translocation from the cytosol to the lipid droplet, where interaction with phosphorylated perilipin-1 further promotes its lipase activity4,5,6,7. The third lipase involved in lipolysis, MGL, does not appear to be regulated by catecholamine signaling8. Importantly, triglyceride synthesis in adipocytes is mediated by the glycerol lipid synthesis pathway, which does not involve the formation of monoglycerides as an intermediate; instead, glycerol-3-phosphate acyl transferases catalyze the formation of lysophosphatidic acid, which is combined with another fatty acyl-CoA to form phosphatidic acid, and then isomerized to diglycerides before the final synthesis of triglycerides (Figure 1)9,10,11.
Figure 1: Lipolysis and glycerol lipid synthesis pathways. Top: Lipolytic pathway; enzymes shown in red: adipose tissue triglyceride lipase (ATGL), hormone sensitive lipase (HSL), and monoglyceride lipase (MGL). Bottom: glycerol lipid synthesis pathway; enzymes shown in green: diglyceride acyltransferase (DGAT), phosphatidic acid phosphatase (PAP), lysophosphatidic acid acyltransferase (LPAT, also known as LPAATs), and glycerol-3-phosphate acyltransferase (GPAT). Lipids: triglyceride (TG), diglyceride (DG), monoglyceride (MG), free fatty acid (FFA), fatty acyl-CoA (FA-CoA), lysophosphatidic acid (LPA), and phosphatidic acid (PA). Other metabolites: inorganic phosphate (Pi) and glycerol 3-phosphate (G3P). Please click here to view a larger version of this figure.
Extracellular adenosine is another important regulator of lipolysis, working through Gs– and Gi-coupled GPCRs to impact adenyl cyclase activity. The predominant adenosine receptor in adipocytes, ADORA1, inhibits adenylyl cyclase, and thus lipolysis through the activation of Gi12. Expressed at lower levels, and primarily in brown adipocytes, ADORA2A activates lipolysis via Gs signaling13. ADORA1 impacts both basal lipolysis and the response to adrenergic agonists. The effect of adenosine on lipolysis can be controlled by adding adenosine deaminase to neutralize adenosine, as well as the ADORA1-specific agonist phenylisopropyladenosine14,15. Hormonal activation of Gq-coupled GPCRs can also affect lipolysis via the activation of phospholipase C and protein kinase C16,17,18,19. Inflammatory signals also impact lipolytic rates. TLR4 activation by LPS (and other endotoxins) increases the lipolytic rate by activating ERK, which phosphorylates perilipin-1 and HSL20. TNF-α also activates lipolysis via ERK and NF-κB activation, as well as transcriptional downregulation of the phosphodiesterase PDE-3B and CIDEC21,22,23. IL-6 has also been associated with increased adipocyte lipolysis, especially in mesenteric adipose tissue, whose FFA release impacts hepatic steatosis and gluconeogenesis24,25,26.
Lipolysis is suppressed during the fed state by insulin. AKT phosphorylates and activates PDE-3B to suppress cAMP signaling and prevent PKA activation27. Insulin also transcriptionally downregulates ATGL28. Obesity promotes catecholamine resistance through a variety of mechanisms, including the downregulation of β-adrenergic receptors in adipocytes29,30,31,32,33. Adipocytes express all three β-adrenergic receptors (β-1, β-2, and β-3). While β-1 and β-2 adrenergic receptors are ubiquitously expressed, the β-3 adrenergic receptor is predominately expressed in adipocytes in mice34,35. Adrb3 expression is induced by C/EBPα during adipogenesis36. The β-3 adrenergic receptor is highly expressed in mature adipocytes. The activation of β-1 and β-2 adrenergic receptors is self-limiting due to feedback inhibition by β-arrestin37. Feedback inhibition of the β-3 adrenergic receptor is mediated by other signaling pathways, which reduce Adrb3 expression33,38,39.
Numerous compounds can be used to activate adipocyte lipolysis. Catecholamines are major physiological activators of lipolysis. Norepinephrine (or noradrenaline) and epinephrine (or adrenaline) activate all three β-adrenergic receptors40. Norepinephrine and epinephrine also effect lipolysis via activation of α-adrenergic receptor signaling41. Commonly used β-adrenergic receptor agonists include isoproterenol, which is a non-selective β-adrenergic receptor agonist, and the β-3 adrenergic receptor agonists CL-316,243 and mirabegron42. Given that adipocytes predominantly express the β-3 adrenergic receptor, we use CL-316,243 as an example here. Its specificity for the β-3 adrenergic receptor also makes it a relatively specific activator of adipocyte catecholamine signaling, that can also be safely used in vivo. Note that the commonly used concentration of 10 µM CL-316,243 in cell culture is orders of magnitude higher than the ~0.1 µM dose required to achieve a maximal response33. Forskolin bypasses the adrenergic receptor, directly activating adenylyl cyclase and downstream lipolytic signaling. There are many more activators, as well as suppressors of lipolysis. When selecting a compound to stimulate lipolysis, the receptor-specificity and downstream signaling pathways should be carefully considered within the experimental design.
The rate of lipolysis in white adipose tissue is an important metabolic factor impacting cold tolerance and nutrient availability during fasting or exercise43,44,45,46. The purpose of this protocol is to measure the rate of lipolysis in adipocytes and adipose tissue, which will facilitate the understanding of adipocyte metabolism and how it may impact the metabolic phenotype of various murine models. To quantify the lipolytic rate, we measure the appearance of lipolytic products in the media (i.e., FFAs and glycerol). The method relies on the release of lipolytic products from the adipocyte into the media. Since white adipocytes express low levels of glycerol kinase, glycerol reuptake rates are low47. Conversely, the production of FFAs and glycerol by metabolic pathways other than lipolysis should also be considered. Adipocytes appear to express a phosphatase with activity against glycerol-3 phosphate, enabling the production of glycerol from glycerol-3-phosphate derived from glucose48,49,50. Glycolysis is a source of glycerol-3-phosphate used for FFA re-esterification in white adipocytes. When glucose levels are limited, glyceroneogenesis requires other 3-carbon sources, such as lactate and pyruvate51. The channeling of FFAs released by lipolysis within the cell and their metabolic fate is poorly understood; FFAs released by lipolysis must be converted to fatty acyl-CoA, before being re-esterified or undergoing β-oxidation. It appears that FFAs released by lipolysis likely exit the cell before being taken back up and converted to fatty acyl-CoA52,53,54,55,56,57,58,59,60,61,62. FFAs can be sequestered outside of the cell by albumin. Importantly, long-chain FFAs are known to feedback-inhibit lipolysis if they are not sequestered by albumin63,64,65,66,67. Thus, optimizing the FFA buffering capacity of the media during the lipolysis assay is critical. The procedure described here is similar to previously published methods to measure the lipolytic rate in adipocytes and ex vivo adipose tissue from mice and humans15,68,69,70,71. This protocol differs through the use of serial sampling; by performing serial sampling, we can internally validate that lipolysis is being measured in the linear phase and utilize multiple measurements to calculate the rate of lipolysis, thereby reducing measurement error to increase confidence in the final calculated value. The drawback of serial sampling is that the assay requires more time and reagents; however, the longer timeframe reduces impact of measurement error on the standard error of the estimates of the rate. Additionally, this protocol measures both FFA and glycerol release, and considers the ratio of FFA:glycerol release with the goal of achieving a 3:1 ratio, as would be expected from complete lipolysis and release of lipolytic products into the media72.
The use of all animals was approved by the Institutional Animal Care and Use Committee (IACUC) at Weill Cornell Medical College of Cornell University.
1. Preparation of buffers and collection plates
2. Sample preparation
3. Lipolysis assay
4. FFA colorimetric assay
5. Glycerol colorimetric assay
6. Calculation of lipolytic rate
We measured the basal and stimulated lipolytic rate of in vitro differentiated adipocytes. Primary preadipocytes from inguinal white adipose tissue were differentiated into adipocytes by the treatment of confluent cells with 5 µM dexamethasone, 0.5 mM IBMX, 1 µg/mL insulin, and 1 µM troglitazone for 4 days, followed by an additional 3 day treatment with 1 µg/mL insulin. Cells were incubated in media without insulin for 24 h prior to the lipolysis assay. At time = 0h, the cells were washed once with PBS, then phenol red-free DMEM with 2% BSA, containing either 10 µM CL-316,243 or vehicle control, was added to each well of the 12-well plate. At time = 1 h, 2 h, 3 h, and 4 h, 50% of the media was collected and replaced with 2% BSA in phenol red-free DMEM. FFA and glycerol levels were measured in the collected media and the moles of FFA, and glycerol secreted into media was calculated at each time point (Figure 2A,B). FFA and glycerol production were linear for the 4 h assay, with R2 values of 0.98 and above. The 4 h FFA levels were slightly low, indicating that lipolytic rates may have been slowing; however, analysis with and without the 4 h time point did not have a significant impact on the calculated lipolytic rate. Stimulated lipolytic rates were significantly higher than basal rates (Figure 2C). Lipolytic stimulation resulted in the production of FFA and glycerol at a near 3:1 molar ratio in all collected samples, as would be expected from complete triglyceride lipolysis without significant reuptake or retention (Figure 2D). However, in the unstimulated cells, the ratio was closer to 1, suggesting a non-lipolytic source of glycerol48,49,50.
Figure 2: Lipolytic rate in in vitro differentiated primary adipocytes. Preadipocytes were differentiated in a 12-well plate. Insulin was removed from the media 24 h prior to the lipolytic assay. At time = 0 h, the cells were stimulated with 10 µM CL-316,243 (CL) or treated with vehicle (V) control media. The nmol of (A) FFA and (B) glycerol produced in each well by each time point were plotted over time, and a linear regression line was fitted for each individual well. (C) Rate of FFA production. (D) Rate of glycerol production. Data are represented as mean ± SEM. (E) Molar ratio of FFA:glycerol in the collected media at each time point. Statistical analysis in (C) and (D) was performed using a student's t-test; the effect of CL was significant at α = 0.05. Please click here to view a larger version of this figure.
In cell culture, the monolayer of cells is in direct contact with the media, while in tissues cultured ex vivo, cells in the center are not in contact with the media. Thus, in ex vivo cultures, FFAs are more likely to be retained within the tissue. Larger chunks of tissue, which have a lower surface area to volume ratio, exhibit higher rates of FFA retention, resulting in lower FFA:glycerol molar ratios (Figure 3A). With decreasing tissue chunk size, the molar ratio of FFA:glycerol released approaches 3:1 (Figure 3A). However, tissue chunks can also be too small. In addition to the challenge of working with tiny pieces of adipose tissue, 1-2 mg chunks of adipose tissue exhibit reduced lipolytic rates, suggesting a reduced viability and functionality of the tissue (Figure 3B,C). While it is tedious and time consuming to cut adipose tissue into chunks of consistent size and shape, it is required to obtain comparable and reliable results. We recommend 5 to 10 mg chunks of adipose tissue, but consistency is most important. Lipolytic rates can still be reproducibly measured in larger tissue chunks, however, it is important to consider the feedback inhibition of lipolysis by FFAs63,64,65,66,67.
Figure 3: Effect of tissue chunk size on FFA release and lipolytic rate. Gonadal white adipose tissue from a high fat diet fed female C57Bl/6J mouse was collected and cut into chunks of varying size. All wells contained 25-30 mg of adipose tissue in total; for the 25 mg wells, this was one chunk of tissue; 10 mg wells had three chunks of tissue each of ~ 10 mg, the 5 mg wells contained five or six chunks of tissue, and the 2 mg wells contained 13-16 chunks. At time = 0 h, lipolysis was stimulated with 0.5 µM CL-316,243 (CL), or the control wells were treated with vehicle (V). The calculated lipolytic rate from samples collected after 1 h, 2 h, and 3 h is plotted for (A) FFAs and (B) glycerol. (C) Molar ratio of FFA:glycerol production in each well. Data are represented as mean ± SEM. Statistical analysis in (A) and (B) was performed using a two-way ANOVA with a Holm-Sidak post-hoc analysis, α = 0.05. The effect of CL was significant in all samples. Within the CL treated samples, the rates of FFA and glycerol production were significantly different in all pair-wise comparisons, except the 10 mg versus 5 mg samples. Please click here to view a larger version of this figure.
The negative impact of FFA retention can be observed in the 25 mg tissue chunks, which exhibit lower rates of both FFA and glycerol production upon stimulation (Figure 3B,C). This feedback inhibition is also apparent when the FFA binding capacity of the media is too low (i.e., there is not enough BSA in the media). When the BSA in the media was reduced to 0.5%, FFA release was reduced more than fourfold, and glycerol release reduced nearly twofold (Figure 4A,B). As an easy rule of thumb, micromolar levels of FFA in the media should not approach the percentage of BSA in the media (i.e., FFA levels should remain below 5 mM FFA in 5% BSA media and 0.5 mM in 0.5% BSA media [at this level there is 20:3 molar ratio of FFA:BSA]). To test the maximal BSA buffering capacity of the preparation of BSA media, collect media from a 24 h stimulated sample, and measure the concentration of FFA (the concentration will be high, so a lower sample volume or sample dilution is recommended). The FFA accumulation in the 5% BSA media after 24 h stimulation was about 5 mM, while the FFA content of the 0.5% BSA media was only 0.6 mM (Figure 4C). Looking at the FFA concentration in the collected media, it can be seen that the 0.5% BSA media FFA binding capacity is overburdened (Figure 4D). While FFA levels in the 5% BSA media are much higher, they do not exceed the buffering capacity of the media (Figure 4D). Ideally, the FFA concentration in the media should remain below a 3:1 FFA:BSA molar ratio (i.e., 2.3 mM FFA in 5% [0.75 mM] BSA).
Figure 4: Insufficient BSA levels result in reduced apparent lipolytic rates. Gonadal white adipose tissue from a high fat diet fed female C57Bl/6J mouse was collected and cut into ~5 mg chunks. A total of 20-30 mg of adipose tissue was placed in each well. At time = 0 h, the lipolysis was stimulated with 0.5 µM CL-316,243 (CL), or the control wells were treated with vehicle (V) in either 5% BSA media or media containing only 0.5% BSA. The calculated lipolytic rate from the samples collected after 4 h is plotted for (A) FFAs and (B) glycerol. The effect of CL was significant in all samples. Within the CL-treated samples, the rates of both FFA and glycerol production were significantly different between the 5% and 0.5% BSA media conditions. (C) FFA levels in the media from the stimulated wells after 24 h. (D) FFA levels in the samples collected at 4 h. (E) FFA standard curves with and without various preparations of 5% BSA in DMEM. Optical density calculated as (Reading B: A550 – A660) – (Reading A: A550 – A660). (F) Glycerol standard curve with and without BSA. Optical density is absorbant at 540 nm (A540). OD values at 5.6 mM were not linear and thus not included in the standard curve. (G) Rate of FFA release from female gonadal adipose tissue using different types of BSA in the assay media. Data are represented as mean ± SEM. Statistical analysis in (A) and (B) was performed using a two-way ANOVA with a Holm-Sidak post hoc analysis, * p value < 0.05. Please click here to view a larger version of this figure.
It is important to point out that some preparations of BSA contain FFAs. Each lot of BSA should be tested to ensure that the FFA content is negligible (note that some BSA preparations are marketed as FFA free). While not specifically marketed as FFA-free, the BSA used here does not contain detectable FFAs. We tested the assay media containing FFA-free BSA (Equitech Bio Inc, BAH66), the BSA used in experiments here (Sigma, A9418), and a cruder (less expensive) fraction V BSA (Sigma, 810531). The FFA-free BSA and A9418 BSA both did not produce any background signal or change the standard curve slope or intercept; the standard curves were superimposable (Figure 4E). No impact of the BSA DMEM was observed on the glycerol standard curve either (Figure 4F). The fraction V BSA, on the other hand, did produce a background signal in the FFA assay, which equated to an FFA concentration of 0.3 mM (Figure 4E). This background shifted the standard curve up, but did not significantly impact the slope, indicating that background subtraction is sufficient. We also performed a stimulated lipolysis experiment using these three different types of BSA, and found that after background subtraction, the calculated rate of lipolysis did not differ between BSA formulations (Figure 4F). However, this is often not the case, especially with less pure BSA preparations which can easily contain insulin or other components that impact the rate of lipolysis. Each lot of BSA should be tested and validated, and used consistently (i.e., samples assayed with different lots of BSA are not comparable). A full standard curve with the addition of BSA and DMEM should be performed when validating each lot of BSA to ensure that it does not contain anything that interferes with the enzymes in the colorimetric assays.
Taking serial samples and replacing the volume with fresh media helps to lower the FFA load in the media throughout the experiment. We measured lipolytic rates in the gonadal adipose tissue from a female mouse fed a high fat diet for 4 months. In this experiment, 5-8 mg tissue chunks (total weight of 25-30 mg) were incubated in 200 µL of 5% BSA media, and 100 µL was collected and replaced at 1 h, 2 h, 3 h, and 4 h. FFA and glycerol levels in the collected media were measured. At 3 h, the FFA levels in the media had reached the danger zone of 2.3 mM (Figure 5A). Feedback inhibition of lipolysis was suggested by the lack of increase in media FFA and glycerol levels at 4 h (Figure 5A,B). After calculating the µmol of FFA and glycerol released at each time point and fitting a linear curve, it is apparent that the lipolytic rate starts to decline at 4 h in the stimulated samples, with R2 values as low as 0.976 for FFA and 0.983 for glycerol (Figure 5C,D). Looking at the residual plot, it is clear that the 4 h values are below the linear trend (i.e., the residuals are negative) (Figure 5E,F). Exclusion of the 4 h time point increased the R2 values to 0.997 and 0.998 and above for FFA and glycerol, respectively. Inclusion of the 4 h time point causes a small but significant decrease in the calculated FFA and glycerol release rates (Figure 5G,H).
Figure 5: Calculation of linear lipolytic rate in ex vivo adipose tissue. Gonadal white adipose tissue from a high fat diet fed female C57Bl/6J mouse was collected and cut into ~5 mg chunks. A total of 20-30 mg of adipose tissue was placed in each well. At time = 0 h, the lipolysis was stimulated with 0.5 µM CL-316,243 (CL), or the control wells were treated with vehicle (V). Media collections were performed at 1 h, 2 h, 3 h, and 4 h. (A) FFA and (B) glycerol levels in the media. (C) FFA and (D) glycerol production plotted over time. Residual plates for (E) FFA and (F) glycerol production over time. Rate of release of (G) FFA and (H) glycerol, calculated with and without the 4 h time point. (I) FFA release in female gonadal adipose tissue. Nearly 50% of media changed at 1 h, 2 h, and 3 h. Between hours, the 15 min time points collected 2.75% of the media without replacement. Data are represented as mean ± SEM. Statistical analysis in (G) and (H) was performed using a two-way ANOVA with a Holm-Sidak post-hoc analysis, * p value < 0.05. The effect of CL was significant in all samples (p value < 0.05). Within the CL treated samples, the rates of both FFA and glycerol production were significantly different between calculations with and without the 4 h time point. Please click here to view a larger version of this figure.
Unlike previous protocols, this protocol utilizes multiple time points to determine the lipolytic rate, thereby reducing measurement error and providing internal validation that the linear rate of lipolysis is being measured. To maintain lipolysis in the linear phase at each collection point, 50% of the media is collected and replaced. Since linearity is critical, we wanted to ensure that the addition of fresh media was not causing a burst of lipolysis at each addition. To validate the linearity between time points, we took a very small sample (2.75%) every 15 min between the hourly time points from 1 h to 3 h, without replacement. At 1 h, 2 h, and 3 h, the normal 50% sample collection was made and replaced. The rate of FFA release was linear between the usual time points, indicating that there was not a burst of lipolysis with each addition of fresh media (Figure 5I).
Here, we provide a basic protocol for measuring the rate of lipolysis in adipocytes and ex vivo adipose tissue. To quantify lipolysis, it is important to measure lipolytic rate in the linear phase. We use a serial sampling technique, where a large fraction of media is collected and replaced with fresh media at regular intervals. This semiconservative method allows for the addition of fresh BSA with FFA buffering capacity and delays feedback inhibition, extending the duration of linear lipolysis. This experimental design attempts to recapitulate the vascularization of adipose tissue in vivo, which delivers fresh albumin to bind released FFAs75. This protocol was optimized to measure lipolysis in murine ex vivo white adipose tissue and inguinal adipose tissue primary preadipocytes differentiated in vitro. The protocol can likely be optimized to work well in brown and beige adipose depots, as well as adipose tissue from other organisms and immortalized cell lines with high adipogenic potential. The protocol allows for flexibility in the age, sex, and diet of the mice prior to the collection of adipose tissue for ex vivo assays. Some manipulations, such as fasting, may impact the basal lipolytic rate, while other interventions such as diet-induced obesity also impact the rate of stimulated lipolysis.
The protocol must be optimized for each experimental system, depending on whether lipolytic rates are higher or lower. To validate linearity, we recommend calculating the R2 values and looking at the residuals plot. The R2 values should be very close to 1. When evaluating the residual plot, one should be on the lookout for negative residual values at later time points, as these indicate that lipolytic rates are declining, and linearity is being lost. Examples of such residual plots are shown in Figure 5E,F. Eliminating the 4 h time points in this case provides a more accurate calculation of the linear lipolytic rate and improves the R2 values. However, since the rate is calculated from multiple time points, it is relatively robust, and inclusion of such a time point has only a small impact on the R2 values and calculated rate of lipolysis. If a single time point is used to calculate the lipolytic rate, the data can be more easily skewed.
The way in which the data are normalized can also impact the relative results between samples. Here, we recommend normalizing the ex vivo tissue to the tissue weight and the primary preadipocytes differentiated in vitro per well. However, these normalization techniques may not be appropriate for all experimental systems. Importantly, if the proliferation rate or differentiation efficiency differs between groups, normalization per well is not appropriate. Alternative normalization techniques include protein, lipid, and DNA. Each method of normalization has its advantages and disadvantages. Tissue weight and well normalization is simple and straightforward. Differences in adipocyte size can mean that lean adipose tissue contains many more adipocytes per gram than obese adipose tissue, while differences in differentiation efficiency can result in differences in the number of adipocytes per well. Lipids can be extracted with organic solvents (e.g., 2:1 chloroform:methanol [v/v]) and triglyceride content measured using a colorimetric reagent. While normalization to lipid content is not a traditionally used method, the lipid droplets are, after all, the organelle undergoing lipolysis, and this normalization method is specific to adipocytes, which can be useful when differentiation rates differ between cells. After lipid extraction, 0.1-0.3 N NaOH can be used to extract protein, which can then be assayed by Bradford or BCA protein assay68. Alternatively, the samples can be homogenized in lysis buffer, and the lipid fraction removed by centrifugation70. Note that lipid contamination will interfere with the protein assays. If protein is to be used for normalization, the cells must be washed extensively (at least three times) to remove BSA from the assay media. Sometimes, differentiated adipocytes do not adhere well to the plate and do not tolerate the three washes required to remove excess BSA. Normalization to DNA allows for normalization to cell number, and can be performed with a commercial DNA extraction kit, followed by the quantification of total DNA by absorbance or real-time PCR analysis of gene copy number. Alternatively, in plated cells, DAPI staining followed by imaging can be used to count nuclei and normalize to cell number. When normalizing using cell number or protein, it is important to consider non-adipocyte cells in the sample. Adipose tissue also contains immune and endothelial cells; in obese adipose tissue, inflammation and hypertrophy can result in as few as 1 in 10 nuclei coming from mature adipocytes. However, adipocytes still contribute to the majority of the mass and lipid content of the tissue. For in vitro differentiated preadipocytes, the differentiation efficiency impacts the relative percentage of mature adipocytes; when this is the case, normalization is complicated. Ideally, the differentiation efficiency should be optimized before utilizing the cells for a lipolysis assay. If not possible, normalization to lipid content is recommended if the lipid droplet volume is similar between differentiated adipocytes. Comparison of lipolytic rates across adipocyte cultures with unequal differentiation can lead to erroneous conclusions about lipolysis when the driving factor is actually adipogenesis; this is a limitation of the protocol.
Long chain FFAs feedback-inhibit lipolysis63,64,65,66,67. FFAs accumulate when they are not sufficiently sequestered in the media. Much of the troubleshooting associated with optimizing the measurement of lipolysis is related to minimizing FFA retention. Calculating the molar ratio of FFA:glycerol release into the media helps to identify issues related to FFA retention. If the molar ratio is 3:1 for FFA:glycerol, then all the products of lipolysis are being released and captured in the media, while a ratio below 2:1 is reason for concern. An important consideration for ex vivo assays is the size of the adipose tissue chunk. Larger chunks of tissue have a lower surface area to volume ratio, and are thus more likely to retain FFAs and exhibit reduced lipolytic rates as a consequence (Figure 4A,B). Bearing this in mind, it is critical to cut the adipose tissue for each sample into chunks of a consistent size and shape. To facilitate FFA sequestration in the media, it is also important to ensure that the media contains sufficient BSA to bind the released FFA. The FFA buffering capacity of the media can be increased by increasing the concentration of BSA or increasing the volume of media. It is equally effective to increase the collection volume or frequency. If one observes high lipolytic rates, but FFA:glycerol ratios below 2:1, we recommend increasing the collection frequency to every 30 min and stopping the assay at 2.5 h.
The protocol provided here is optimized for mouse ex vivo white adipose tissue and primary preadipocytes differentiated in vitro, both of which are highly lipolytic. The assay needs to be optimized to measure lipolysis in other systems, or if differences in the relatively low basal rate of lipolysis are of particular interest. For example, 3T3-L1 adipocytes have lower lipolytic activity32. When the lipolytic rate is low, the incubation volume should be decreased (i.e., instead of 400 µL per well in a 24-well plate, 200 µL in a 24-well plate should be used, or 300 µL in a 12-well plate, and 100-150 µL collected per time point. The collection volume should not drop below 100 µL, as larger sample volumes are required to accurately measure low FFA and glycerol levels. To assay glycerol using 50 µL of media per sample, the free glycerol reagent should be dissolved in 31 mL of water, and 150 µL of reagent used with 50 µL of sample in each well. FFA levels can be measured with 25 µL of sample per well, perhaps more, so long as linearity is maintained. Time points can also be extended to increase signal. When assaying lipolytic rates from cells other than white adipocytes, differences in cellular metabolism should be considered. For example, brown and beige adipocytes express much higher levels of glycerol kinase, and are thus able to retain glycerol release by lipolysis76,77. These metabolically active cells may also be able to channel released fatty acids into catabolic or synthetic pathways without release into the media. The metabolic fate of the FFA and glycerol in the specific cell type being assayed must be considered when interpreting the results of a lipolysis assay, especially in a cell type other than white adipocytes. Lipolysis assays in other cell types need to be optimized and validated.
This protocol is designed to evaluate differences in lipolytic rate in mouse models. Primary preadipocytes differentiated in vitro are a useful tool for investigating cell autonomous effects of genetic manipulations or pharmacological treatments in adipocytes. Adipose tissue, on the other hand, contains other cell types, including immune cells. The impact of adipose tissue immune cells in regulating lipolysis is important to consider when assessing whole tissue lipolysis. Cultured adipocyte lipolysis models may circumvent the potential complicating influence of immune cells and enable the thorough investigation of the specific cell types, while adipose tissue explants enable a more robust comparison to the in vivo environment. Additionally, the ex vivo assay can be used to investigate lipolytic rates in various adipose depots. In these systems, lipolysis is stimulated by compounds such as β-adrenergic receptor agonists, thus, any changes in sympathetic tone that may impact the in vivo mouse model will not be observed. In vivo measurement of lipolysis is complicated by the dynamics of FFA and glycerol release and uptake in various tissues throughout the body. Serum FFA and glycerol levels at any given time point are the balance of secretion and uptake, and it should not be presumed that changes in serum FFA and glycerol levels are attributable solely to adipose tissue lipolysis. Fasting-induced lipolysis is impacted by sympathetic tone, but does not produce rapid and robust changes in serum FFA or glycerol levels, making it difficult to interpret changes in fasted serum levels of FFA and glycerol. In vivo lipolysis can be stimulated using a β-3 adrenergic receptor agonist, such as CL-316,243, and increased serum FFA and glycerol are detectable within 20 min of stimulation. In vivo lipolytic assays should include both baseline measurements of serum FFA and glycerol, as well as a vehicle control; this way, the stimulation-specific change in serum FFA and glycerol levels can be determined.
The authors have nothing to disclose.
This work was supported by the US National Institutes of Health grant R01DK126944 to S.M.R.
24-Well tissue culture treated plate | Corning Inc | 3527 | Must be tissue culture treated for adipocyte differntiation |
48-Well flat bottom plate with lid | Corning Inc | 353078 | Can be tissue culture treated |
6-Well flat bottom plate with lid | Corning Inc | 353046 | Can be tissue culture treated |
96-Well PCR Plate | USA sceintific | 1402-9100 | Any conical 0.2 mL PCR plate will be convenient |
Bovine Serum Albumin | Sigma Aldrich | A9418 | FFA free BSA such as A8806, is also commonly used. The BSA should not have detectable FFA, also lot to lot variations in BSA can impact the observed rate of lipolysis |
CL-316,243 | Sigma Aldrich | C5976 | CAS #: 138908-40-4 availaible from other suppliers |
CO2 incubator | PHCBI | MCO-170AICUVH | CO2 should ideally be set to 10% for adipose tissue, however 5% CO2 will also work |
DMEM, low glucose, no phenol red | Thermofischer | 11054020 | Any phenol red free media should work, DMEM/F12, RPMI, but should contain volatile buffering capacity, i.e. biocarbonate |
FFA-free Bovine serum albumin | Equitech-Bio, Inc, | BAH66 | |
Free Glycerol Reagent | Sigma Aldrich | F6428 | |
Glycerol Standard Solution | Sigma Aldrich | G7793 | This can also be made by diluting glycerol to the desired concentration |
HR Series NEFA Standard Solution | Fujifilm | 276-76491 | |
HR Series NEFA-HR (2) Color Reagent A | Fujifilm | 999-34691 | |
HR Series NEFA-HR (2) Color Reagent B | Fujifilm | 991-34891 | |
HR Series NEFA-HR (2) Solvent A | Fujifilm | 995-34791 | |
HR Series NEFA-HR (2) Solvent B | Fujifilm | 993-35191 | |
Microbiological Incubator | Fischer Scientific | S28668 | Any incubator at 37C can be used |
Nunc MicroWell 96-Well Plates | Thermo Scientific | 269620 | Any optically clear, flat bottom 96-well plate works |
Silicone Laboratory Benchtop Mat | VWR | 76045-300 | Glass plate can also be used. Absorbant surfaces are not recommended |
Spectrophotometer/Microplate Reader | Molecular devices | SpectraMax i3x | Any plate reader that can read at 540, 550 and 660 mm will work |
V Bovine serum albumin | Sigma-Aldrich | 810531 | |
WypAll X70 Wipers | Kimberly-Clark | 41200 | Any high quality paper towel will work |