Measuring Mitochondrial Respiration in Previously Frozen Biological Samples

Measuring oxygen consumption allows for the role of mitochondrial function in biological phenomena and mitochondrial diseases to be determined. Although respirometry has become a common approach in disease research, current methods are limited by the necessity to process and measure tissue samples within 1 hr of acquisition. Detailed by Acin‐Perez and colleagues, a new respirometry approach designed for previously frozen tissue samples eliminates these hurdles for mitochondrial study. This technique allows for the measurement of maximal respiratory capacity in samples frozen for long‐term storage before testing. This protocol article describes the optimal tissue isolation methods and the combination of substrates to define electron transport chain function at high resolution in previously frozen tissue samples. © 2020 The Authors.


INTRODUCTION
Impaired mitochondrial respiration plays a key role in metabolic, aging-related, and cardiovascular disease (Liesa, Palacín, & Zorzano, 2009;Wallace, 2011). Mitochondrial respiration results from the transfer of electrons between complexes I, II, III, Figure 1 Flowchart of the frozen tissue respirometry protocol detailing the most important steps. The sections of the protocol are outlined in the top right box, which will be referred to in subsequent figures. and IV, with complex IV reducing oxygen to water; thus, oxygen consumption integrates electron transport activity from complexes I/II through IV. However, mitochondrial respirometry analysis currently requires processing and measurement of the living tissue sample within an hour of being taken from the patient. This requirement is set by the need to preserve the integrity of the inner mitochondrial membrane, needed both for integrated electron transfer between the complexes as well as to preserve the coupling of electron transfer to ATP synthesis. Consequently, the need to use fresh tissue to assess integrated electron transport activity makes respirometry analysis largely unfeasible for standard clinical practice and clinical studies. This limitation has significantly stalled scientific progress and virtually excluded the possibility of translating some discoveries in mitochondrial physiology into improved patient care. Here we describe a methodology to assess mitochondrial respirometry capacity in previously frozen biological specimens. This procedure (see flowchart in Fig. 1) overcomes the fundamental limitation of conventional respirometry approaches that require freshly isolated tissue, and also allows for the use of less biological material. Furthermore, our approach also allows for the use of a combination of substrates that permits selective measurements of respiratory capacity driven by complex I, II, or IV alone, resembling previous methods for determining isolated electron transport chain complex activity.
Detailed in Acin- Perez et al. (2020), the maximal respiratory capacity measured from frozen tissues using this assay is comparable to that measured in fresh tissue ( Figure 2A). Freeze-thawing impairs mitochondrial respiration by disrupting and permeabilizing the mitochondrial inner membrane, which releases electron carriers that support the electron transport chain ( Figure 2C). However, the interaction of electron transport complexes (supercomplexes) remains intact under freeze-thawing conditions, allowing for integrated electron transport as seen in fresh tissues. This article, therefore, provides instructions to rescue the defects induced by freeze-thawing, as well as to control the entry of electrons into the electron transport system by providing different electron carriers.
Osto et al.

of 17
Figure 2 Respirometry measurements of complex I, II, and IV comparing fresh versus frozen tissue protocols. (A) The Seahorse traces and associated OCR (oxygen consumption rate) quantifications of mouse liver homogenates. hFresh represents liver homogenate which was isolated and tested without being frozen. TMPD is added to a subset of the hFresh samples to uncouple the mitochondria to survey complex activity separately from ATP-linked respiration. hFrozen represents liver homogenates which underwent the frozen respirometry protocol. While complex IV was surveyed using TMPD/ascorbate for all three assays, the top graph was first stimulated with pyruvate + malate, the middle with NADH, and the bottom with succinate + rotenone. (B) Succinate + rotenone and TMPD + ascorbate−dependent oxygen consumption rate measured in the presence of a titration of ATP concentrations. (C) Oxygen consumption rate in response to NADH injection for frozen liver homogenate treated with or without digitonin to control whether the mitochondria are fully broken and substrates can directly fuel the electron transport chain. (D) Complex I and complex II activity (measured as oxygen consumption rate). (E) Oxygen consumption rates of fresh and frozen mitochondria respiring under succinate+rotenone. Fresh mitochondria respond to ADP, oligo, FCCP, and AA, whereas frozen mitochondria, being uncoupled from the beginning, only respond to AA.
As our procedure allows for the specific measurement of oxygen consumed by complex IV, mitochondrial isolation is no longer necessary, allowing for the use of homogenates to measure mitochondrial function, significantly simplifying tissue preparation (described in Basic Protocol 1).
The methodology described in this article is broken up into three subsections: Basic Protocol 1 for sample collection, storage, and homogenization for respirometry on previously frozen tissue; Basic Protocol 2 for running a Seahorse respirometry assay using previously frozen tissue samples; and Basic Protocol 3 for normalization to mitochondrial content for respirometry on previously frozen tissue (optional). Figure 1 provides a graphical representation of this method from start to finish. Basic Protocol 1 describes the process of dissecting a sample, freezing a sample for storage, and homogenizing a sample for a more simplified sample preparation. Basic Protocol 2 details the loading of the XF96 Seahorse sample plate, describes the compounds to be loaded into the Seahorse cartridge injection ports, and lays out the Seahorse protocol steps to measure respirometry in previously frozen tissues. Basic Protocol 3 describes how to use MitoTracker Deep Red staining to normalize samples to mitochondrial content rather than whole protein Osto et al.

of 17
Current Protocols in Cell Biology content. Completion of this series of protocols will allow users to progress from removing a tissue of interest from a specimen, freeze the tissue for later study, and set up a Seahorse assay to obtain complex I, II, and IV functional readings from the previously frozen tissue sample.

SAMPLE COLLECTION, STORAGE, AND HOMOGENIZATION FOR PREVIOUSLY FROZEN TISSUE RESPIROMETRY
The following protocol details how tissue samples should be collected and stored prior to being analyzed using the previously frozen tissue respirometry assay. In this protocol, the tissue of interest will be isolated from the organism and frozen for long-term storage. Samples will then be manually homogenized and isolated via centrifugation, or, optionally mitochondria may be isolated from the homogenate. Finally, a BCA protein assay will be performed to normalize the samples either for immediate loading or for normalization to mitochondrial mass (described in Basic Protocol 3). If the protocol is followed accurately, the samples should progress from whole tissue to a homogenate or mitochondrial isolate that is ready for normalized loading on a Seahorse sample plate in Basic Protocol 2.
This sample preparation protocol is optimized for mouse liver isolation. See Troubleshooting (Table 3) for alterations to be made for the collection, preparation, and measurement of other tissue types using the frozen tissue respirometry method. 3. Snap freeze in liquid nitrogen. For long-term storage, store the frozen tissue at −80°C.

Mice
Liver samples have been shown to have little loss in respirometry results if stored at room temperature or 4°C for up to 3 hr prior to being snap frozen (Fig. 3A).
Though snap freezing is the most ideal method of tissue preparation, the tissue may be immediately stored at −80°C or at −20°C for up to 4 hr prior to being stored at −80°C without affecting the respirometry results (Fig. 3B).

Figure 3
Tissue collection and homogenization. (A) Complex I, complex II, and complex IV oxygen consumption rates of mouse liver homogenates. Liquid nitrogen (LN) samples were immediately snap-frozen after isolation: 4°C 0.5 hr and 4°C 3 hr were kept at 4°C for 0.5 and 3 hr, respectively after isolation prior to being snap-frozen. RT 0.5 hr and RT 3 hr were kept at room temperature for 0.5 and 3 hr, respectively after isolation prior to being snapfrozen. (B) Complex I, complex II, and complex IV oxygen consumption rates of mouse liver homogenates. LN homogenates were isolated then snap-frozen in liquid nitrogen prior to storage at −80°C. −80°C samples were isolated then stored at −80°C until homogenization. −20°C samples were isolated then stored at −20°C for 4 hr before being stored at −80°C until homogenization. (C) Mouse liver before and after 10 strokes of homogenization using a glass-Teflon homogenizer.
5. Homogenize using 250 μl to 500 μl per 10 mg of tissue of ice-cold MAS buffer using the appropriate homogenizer ( Fig. 3C illustrates this process). For liver, brown adipose tissue, brain, kidney, lung, and skeletal muscle, use ten to twenty strokes in a glass-Teflon Dounce homogenizer. For heart and white adipose tissue, use twenty strokes in a glass-glass Dounce homogenizer.
CRITICAL: For muscle tissue, incubate homogenized tissue in collagenase Type II (0.25 mg/ml final concentration) in MAS at 37°C for 30 min prior to centrifugation.
Our studies have shown that using smaller-volume homogenizers (e.g., using a 7-ml homogenizer rather than a 15-ml homogenizer when working with <5 ml of sample material) results in less sample retention in the homogenizer and produces better homogenization.
Osto et al.

of 17
Current Protocols in Cell Biology

Figure 4
Homogenate isolation from frozen tissues. After homogenization, the homogenate must be isolated from the cellular debris. This is performed using a 1000 × g centrifugation for 5 min at 4°C. Highlighted in this figure is the pellet resulting from this centrifugation (boxed) and the supernatant (not boxed). The supernatant is your tissue homogenate, which should be collected and may be used for normalization prior to respirometry, for mitochondrial isolation, or for nonrespirometry assays like Blue Native Gel Electrophoresis. The pellet should be discarded.
CRITICAL: If using swinging-bucket centrifuge, measure the precise distance between rotor center and the center of mass of the liquid homogenate in the tube in the horizontal swinging position in order to correctly calculate the g to rpm conversion: rcf (g) = 1.12 * radius (mm) * rpm 1,000 2 7. Collect the resulting supernatant as the homogenate sample. Discard the pellet (Fig. 4). The supernatant (tissue homogenate) may be stored at −80°C for long-term storage.
This sample may also be used for other assays like enzymatic assays, western blots, or Blue Native gel electrophoresis.
8. If you are running the sample as a tissue homogenate, perform a BCA protein assay to normalize your samples according to protein content prior to loading on the Seahorse XF96 sample plate ( Fig. 5B, Method 1). If you are normalizing according to mitochondrial content, perform a BCA protein assay and continue with the steps detailed in the optional Basic Protocol 3 (Fig. 5B, Method 2). If you are isolating mitochondria from the tissue homogenate, proceed with the following steps for mitochondrial isolation protocol.

Mitochondrial isolation from homogenate (optional)
9. If mitochondrial isolation is desired, centrifuge the rest of the homogenate preparation (the supernatant from the above centrifugation) 10 min at 10,000 × g, 4°C.
10. Optionally, wash mitochondrial pellets twice in 1× volume of MAS buffer, each time centrifuging as in step 9.
11. Re-suspend final mitochondrial pellet in ice-cold MAS buffer.
12. Perform a BCA protein assay using the mitochondrial homogenates. This may be used to normalize the sample according to protein content for loading, or optionally Osto et al.

of 17
Current Protocols in Cell Biology

Figure 5
Normalization and loading of tissue homogenates or mitochondrial isolates. Normalization may be performed in one of two ways: protein normalization using a BCA assay or mitochondrial mass normalization using MTDR dye. Method 1 describes the protein-based normalization using the BCA assay. Method 2a describes mitochondrial normalization performed prior to running the XF96 Seahorse plate. Method 2b describes mitochondrial normalization performed after running the XF96 Seahorse plate.
to prepare for mitochondrial mass normalization using MitoTracker Deep Red dye as described in Basic Protocol 2. Figures 3C and 4 depict the results one should obtain after sample collection. Figure 3C shows a mouse liver sample that has been homogenized using a glass-Teflon homogenizer. Figure 4 shows the same sample after centrifugation, highlighting the pellet formed (which should be discarded) and the supernatant (which should be saved). If this sample separation does not occur after centrifugation, you should repeat the centrifugation step a second time.

RUNNING A SEAHORSE RESPIROMETRY ASSAY USING PREVIOUSLY FROZEN TISSUE SAMPLES
This protocol will detail the process of loading normalized samples onto an XF96 Seahorse plate, loading the Seahorse cartridge with compound injections, and running samples on the Seahorse platform. Samples will be loaded onto a XF96 sample plate, then centrifuged and diluted before loading the plate onto the Seahorse. Additionally, this protocol describes the loading of the Seahorse cartridge with compound injections. If this protocol is followed accurately, an operator will be capable of loading samples onto an XF96 sample plate, loading compound injections into a Seahorse cartridge, and running the plates on a Seahorse XF96 system. While this protocol is optimized for use with the Seahorse platform, with optimization, it could be adjusted to work with other respirometry platforms like the Oroboros.

Materials
Tissue homogenates or mitochondrial isolates (Basic Protocol 1) 1× MAS buffer (see recipe) Osto et al. Optimal loading concentration will vary depending on the sample material, but for most samples a concentration between 2.5 and 20 μg of sample per well will be sufficient. See Table 3 in the Troubleshooting section of this document for approximate loading concentrations for different tissue types.

of 17
Coating the XF96 plate with substrates promoting cell adhesion like laminin, Cell-Tak, and poly-D-lysine has been shown to not improve performance.
CRITICAL: Turn OFF the centrifuge brake and let buckets slow down by themselves. Sudden deceleration due to braking will result in uneven distribution of materials on the bottom of the plate.
3. Carefully add 130 μl (or enough to achieve a final volume of 150 μl) per well of 1× MAS buffer containing cytochrome c (10 μg/ml, final concentration) for homogenates, or the same volume of MAS buffer alone for isolated mitochondria.
CRITICAL: For brain and lung homogenates, add alamethicin (10 μg/ml, final) to the MAS buffer containing cytochrome c to allow complete membrane permeabilization to substrates.
All compound injections should be diluted in 1× MAS buffer. When making the TMPD + ascorbate solution, add 0.5 mM TMPD into 1 mM ascorbate and adjust the solution to pH 7.2 using potassium hydroxide (KOH) and hydrochloric acid (HCl). NADH and TMPD/ascorbate solutions must be made freshly on the day of the assay, but all other compound injections can be made ahead of time and stored at −20°C until time of use.
5. Mix and measure times for the Seahorse run protocol are 0.5 and 4 min, respectively.
If a protocol includes an oligomycin injection, a 2-min wait time should be included between the injection of oligomycin and the 4-min measurement period (Fig. 6B).   Figure 6C depicts a representative Seahorse trace for a standard assay which measures complex I, II, and IV in mouse liver homogenates. Figure 6C-I and C-II highlight the complex I NADH peak and complex II succinate + rotenone peak, respectively. Figure C-IV highlights the complex IV TMPD + ascorbate peaks from the complex I and complex II samples. This Seahorse trace not only has distinct peaks for all substrate injections, but also has strong and distinct responses to the inhibitors (antimycin A and azide) used in the assay, making it an optimal-looking trace.

NORMALIZATION TO MITOCHONDRIAL CONTENT FOR PREVIOUSLY FROZEN TISSUE RESPIROMETRY (OPTIONAL)
The following protocol describes normalization of the results from respirometry samples according to mitochondrial content. Homogenates or mitochondrial isolates (made in Basic Protocol 1) will undergo MitoTracker Deep Red (MTDR) staining and measurement. Using this data, samples can be normalized based on their mitochondrial content rather than using a whole protein measurement.

of 17
Current Protocols in Cell Biology equivalent protein amounts in the XF96 Seahorse plate as described in Basic Protocol 2 (Fig. 5B, Method 1).Normalization for mitochondrial content using MitoTracker Deep Red (MTDR) may be performed before or after running the Seahorse sample plate.
While MTDR normalization can be performed at any step after BCA values have been generated, loading onto the Seahorse sample plate is exclusively done based on BCA values. Normalization to mitochondrial content should only be done after oxygen consumption rate (OCR) values are obtained.
Normalizing to mitochondrial content before running a Seahorse experiment (Fig. 5B 6a. Calculate relative mitochondrial content after controlling for background fluorescence using a blank well. Normalizing to mitochondrial content after running a Seahorse experiment (Fig. 5B, Method 2b) 1b. Remove the Seahorse sample plate from the Seahorse machine.
2b. Carefully, without agitation of the sample material at the bottom of the Seahorse sample plate, remove the MAS buffer from the plate until approximately 80 μl of liquid remains in each well.
3b. Add 20 μl of 5 μM MTDR in MAS buffer to each sample such that each sample now has 100 μl of liquid with a final MTDR concentration of 1 μM.
4b. Incubate the sample plus dye at 37°C for 10 min. Ensure that the incubation period is no more than 10 min; too long of an incubation period can result in high background MTDR fluorescence.
5b. Add 100 μl of MAS buffer without MTDR to each sample. Remove 100 μl of solution from each well. Perform this washing procedure up to three times.
6b. Measure fluorescence in a plate reader or high-throughput microscope.
MTDR is excited at 625 nm and emits at 670 nm.
7b. Calculate relative mitochondrial content after controlling for background fluorescence using a blank well. For absolute quantification, use an isolated mitochondria standard.
Osto et al.

of 17
Current Protocols in Cell Biology a This figure shows an example data set of MTDR readings from mouse liver homogenate samples. The user will be generating data to fill the "raw absorbance units" and "ug protein loaded via BCA" columns of the table. The "raw absorbance units" column will be filled with the absorbance readings from the plate reader or high-throughput microscope taking MTDR measurements. The "μg protein loaded" via BCA will be filled with the amount of protein loaded (calculated via BCA) into the well for the MTDR reading. Using these two values, the final column titled "(Raw AU -background)/(μg protein)" can be calculated by dividing the raw absorbance minus background and the μg protein loaded via BCA. Normalizing your samples based upon the values in this column will result in equivalent sample loading based upon mitochondrial content. Table 1 showcases data from a MTDR plate reader assay to normalize mouse liver by mitochondrial content before loading a Seahorse sample plate. The columns titled "raw absorbance units" and "μg of protein loaded via BCA" will be filled in by the user based on the plate reader results and the amount of protein added to the plate for the MTDR reading (determined via BCA). Dividing the raw absorbance minus the background absorbance and the μg of protein loaded results in the data in the farthest right column entitled "(Raw AU -background)/(μg protein)," which is now a reading of mitochondrial content per μg of protein in the sample. Normalization may then be performed such that the OCR values obtained are normalized to measurements of OCR per mitochondrial content.

Background Information
Mitochondrial respiration results in the transfer of electrons between electron transport chain complexes I, II, III, and IV, with complex IV reducing oxygen to water. With oxygen playing a primary and crucial role in the process, respirometry platforms like the Seahorse or Oroboros use the measurement Osto et al.

of 17
Current Protocols in Cell Biology of oxygen consumption rate as a surrogate marker for mitochondrial activity. Current mitochondrial respirometry techniques for in vivo samples require that the processing and measurement of the living tissue sample happen within an hour of isolation from the specimen. This requirement is set by the need to preserve the integrity of the inner mitochondrial membrane, needed both for integrated electron transfer between the complexes as well as to preserve the coupling of electron transfer to ATP synthesis. Although freeze-thawing disrupts and permeabilizes the mitochondrial inner membrane, our novel respirometry method overcomes this challenge to allow for previously frozen tissue samples to be measured. Mitochondrial supercomplexes remain intact despite freezethawing; therefore, we have developed a set of substrates that are capable of measuring complex I, II, and IV through measurement of these supercomplexes. While this assay cannot measure coupled mitochondrial respiration, as can be done in fresh tissue respirometry, freezing tissues allows for long-term tissue storage and simplified sample processing through use of tissue homogenates, thereby enabling respirometry experiments that are less timeintensive, require less tissue processing, and are less expensive, facilitating large-scale respirometry research studies when compared to current respirometry methods.

Critical Parameters
Multiple parameters within this methodology are critical to obtaining good results from your respirometry run. In Basic Protocol 1, the most critical parameter to success is the homogenization step. If there is too little homogenization, mitochondria will not be released from the cell, resulting in low yield; too much homogenization might impact the mitochondrial supercomplexes. We have provided guidelines for the types of homogenizers and the number of strokes required for different tissues, but this could be a variable to optimize if you are getting suboptimal mitochondrial yield from your tissue samples.
Basic Protocol 2 is likely the part of the procedure that is most sensitive to certain experimental parameters. The most important parameter in this basic protocol is that centrifugation steps be done with the centrifuge brake turned OFF. Braking during deceleration of the centrifuge can result in an uneven distribution of the sample across the bottom of the Seahorse sample plate, adversely effecting OCR results and reproducibility. A centrifuge with the capability of having the brake turned off is necessary for best performance in this assay. An additional variable to consider for this part of the assay is how much sample is being loaded onto the sample plate. Too little sample and OCR values will be very low, but having too much sample can result in oxygen depletion and inconsistent responses to the compound injections (diagnosing these issues is discussed in the Troubleshooting, below). We recommend titrating your sample material to find the optimal loading concentrations that provide the best Seahorse results.
Basic Protocol 3 has few parameters that are critical to results, but it must be noted that while normalization to mitochondrial content is optional, it provides an additional analysis technique that may explain differences in respiration due to samples having differing numbers of mitochondria. An example of where this may be impactful is if two samples have the same OCR reading, but one sample has fewer mitochondria that have higher respiratory capabilities and the other has many mitochondria that have low respiratory capabilities. In this scenario, normalizing the data to mitochondrial content could elucidate these differences.

Troubleshooting
This troubleshooting section is designed to account for issues specific to frozen tissue respirometry, although traditional Seahorse troubleshooting techniques will likely also apply to this technique. Figure 6C displays a representative Seahorse trace for a standard assay measuring complexes I, II, and IV in liver homogenates. Table 2 describes factors that may cause poor respirometry results. Additionally, this table catalogs how to diagnose these errors and the potential solutions to improve respirometry sample performance. This protocol was initially optimized for frozen liver respirometry, but many other tissue types have been successfully tested using this procedure, with some modifications. Table 3 compiles some of the potential optimizations in tissue collection and respirometry when using alternative tissue types with this protocol. Further examples of respirometry results using alternative tissue types including Seahorse traces can be found in Acin- Perez et al. (2020); also see Table 3.

Statistical Analysis
When the Seahorse run is complete, export the oxygen consumption rates (OCRs) from the Agilent Wave software and (if applicable) Osto et al.

of 17
Current Protocols in Cell Biology   Calculating the complex I to complex IV or complex II to complex IV ratio may serve as an alternative measurement of complex respiratory capacity. The top graph of A shows OCR, and the bottom graph shows oxygen levels. At 4 μg of loaded sample, oxygen levels deplete linearly for both measurements after injection 1. As more sample is loaded, oxygen depletion becomes less linear, and between the two measurements oxygen levels do not fully recover, both of which are signs of oxygen depletion affecting respiration results. (B) A white adipose tissue (WAT) sample exhibiting inconsistencies between measurement 1 and measurement 2 after compound addition. With 50 μg of WAT loaded, measurement 2 after TMPD/ascorbate addition begins high, but drops dramatically as the measurement cycle continues, ultimately affecting the overall respirometry trace. When the sample input is decreased, measurements normalize and the drop in measured OCR is not observed. (C) A white adipose tissue (WAT) sample measured at 20 μg as a homogenate and measured at 10 μg after concentrating the sample by pelleting the sample and resuspending in a lower volume. Figure 6C exhibits a representative Seahorse trace. The trace in red corresponds to a complex I peak ( Figure 6C-I) followed by a complex IV peak ( Figure 6C-IV) induced by NADH and TMPD/ascorbate, respectively. The trace in blue corresponds to a complex II peak ( Figure 6C-II) followed by an additional complex IV peak ( Figure 6C-IV) induced by succinate/rotenone and TMPD/ascorbate, respectively. This sample responds to all substrate and inhibitor injections and has relatively low standard deviations for each measurement across the three replicates tested. When analyzing the data, the primary data points of importance will be Figure 6C,D, the complex I respiration induced by NADH addition, Figure 6C,C-II, the complex II respiration induced by succinate addition, and Figure  6C,C-IV, the complex IV respiration induced by TMPD + ascorbate addition. The measurements after antimycin A injection and azide injection are used primarily as background controls where the antimycin A measurement represents any respiratory activity not produced by the non-mitochondrial respiration of the electron-transport chain complexes, and the azide measurement represents any non-mitochondrial respiratory activity still in the sample. Figure 7 showcases Seahorse traces which are not ideal. The Troubleshooting section, above, describes potential reasons for why certain traces may look like those in Figure 7.

Time Considerations
Basic Protocol 1 will take approximately 2 hr for 15 samples, not including dissection of the tissue of interest from the specimen (the time to do this will depend on the specimen and tissue being used). This time is largely based on manual homogenization of the samples; an automated system of tissue homogenization could dramatically decrease the sample processing time, although these methods would need to be optimized for the purpose of this assay. Basic Protocol 2 will take approximately 2 hr to set up for 15 samples, then the Seahorse run will take approximately 90 min, with no user input required. This process could be easily scaled up for more samples with multiple Seahorse instruments, where larger volumes of each compound injection could be made for multiple Seahorse cartridges. Basic Protocol 3 will take approximately 30-60 min for 15 samples. In total, 15 samples will take approximately 6 hr, although this time can be split between 1 day of sample processing and storage and 1 day of running the Seahorse assay.