Here, we describe in detail methods for extracting macrophages from the bone marrow, spleen, and infarcted heart, and subsequently assessing metabolic flux in live cells.

Metabolic reprogramming is a hallmark of monocyte/macrophage activation and polarization between pro- and anti-inflammatory states. For example, pro-inflammatory (i.e., M1-like) monocytes/macrophages display more reliance on anaerobic glycolysis and less reliance on mitochondrial oxidative phosphorylation, whereas anti-inflammatory (M2-like) macrophages display more reliance on glucose and fatty acid oxidation in the mitochondria. Here, we describe in-depth protocols for extracting macrophages from the two major monocyte/macrophage reservoirs in the body, the spleen and bone marrow, as well as injured tissues such as the heart following myocardial infarction.

Macrophages or monocytes are extracted by immunomagnetic sorting by using antibody-tagged microbeads, which easily bind to cells without compromising their phenotypes. The extracted cells are then cultured in 96-well plates, followed by extracellular flux analysis using a metabolic flux analyzer. Both glycolysis and mitochondrial oxidative phosphorylation can be measured simultaneously in small numbers of cells (as little as 2-3 × 10⁵ cells). This method can easily be performed in 1 day and produces reliable and repeatable results. Ultimately, these methods help to enhance our understanding of metabolic changes during immune and inflammatory responses to injury and disease, which could lead to the development of novel therapeutic targets for immunometabolic pathways.

Immunometabolism is a blossoming field that studies the role of metabolic reprogramming in immune cells across different pathological disease and injury states. Macrophages are a key part of the innate immune system that play critical roles in inflammation, response to infection, antigen presentation, and wound healing¹. Understanding how macrophages polarize between pro- and anti-inflammatory (M1-like and M2-like) subsets across different disease states is an area of ongoing and intense investigation. Recent studies have identified metabolic reprogramming as a key mechanism underlying macrophage polarization. The current paradigm is that, broadly speaking, M1-like macrophages (which are typically monocyte-like) rely more on glycolysis to fuel pro-inflammatory functions, while M2-like macrophages rely more on mitochondrial oxidative phosphorylation to quell pro-inflammatory functions and fuel anti-inflammatory processes². Understanding how macrophage metabolism is altered in different disease states can provide insight into potential therapies that could be used to target metabolic pathways.

As an example, our lab has extensively investigated the role of macrophage metabolic reprogramming during myocardial infarction (MI)^3,4,5. Macrophages play a key role in the inflammatory and wound healing response during MI and, as such, undergo polarization from M1-like towards M2-like phenotypes as the infarcted heart undergoes remodeling to form replacement scar tissue. Using the methods described herein, we have demonstrated that this polarization is characterized by unique changes in glucose and glutamine metabolism, and mitochondrial function. We describe methods for extracting cardiac macrophages, as well as splenic macrophages and bone marrow monocytes, which can be combined with extracellular flux analysis to assess ex vivo metabolic flux in a single day. We hope the methods described offer a standardized approach for assessing immune cell metabolic phenotypes to enhance reproducibility across labs studying this important topic.

The methods below describe protocols for extracting macrophages and performing ex vivo analysis of metabolic flux from the infarcted heart following MI, the spleen, and the bone marrow (Figure 1). For the MI heart and spleen, the extraction method used is identical. All protocols involving mice were approved by the University of Mississippi Medical Center Institutional Animal Care and Use Committee (Protocol #1371, Mouton).

1. Extraction of macrophages from the infarcted heart and spleen

NOTE: Extraction of tissue macrophages uses a negative selection strategy to first remove neutrophils, which are labeled with Ly6G-microbeads, and then a positive selection strategy to obtain macrophages with CD11b-microbeads^3,4,5,6. When performing this protocol, work quickly and keep the cells cold.

1. Euthanize the mouse in accordance with AALAC and institutional guidelines by isoflurane overdose followed by removal of the heart.
2. Remove the heart and spleen quickly and weigh the tissues.
3. Prior to the assay, prepare PEB buffer by dissolving EDTA (2 mM) and 0.5% bovine serum albumin (w/v) in phosphate-buffered saline (PBS). Keep cold.
4. Place the heart in cold Hank's balanced salt solution (HBSS). Mince with a sterile scalpel into small pieces.
5. Prepare a digestion solution containing 600 U/mL collagenase type II and 60 U/mL DNase type I. Mix the tissue in 10 mL of the digestion solution in a MACS cell dissociation tube.
   NOTE: Make enough of the digestion solution to use 10 mL for each tissue (room temperature).
6. Incubate the tissues using the cell dissociator according to the manufacturer's protocol. This will incubate the minced tissue at 37 °C with gentle spinning for ~1 h to generate a cell suspension.
7. Move the suspension to 15 mL conical tubes and centrifuge at 300 × g for 10 min at 4 °C.
8. Resuspend the pellet in 1 mL of PEB buffer. To generate a single-cell suspension, apply the suspension over a 30 µm filter.
9. To lyse red blood cells, add 10 mL of 1x red blood cell lysis solution. Incubate for 5-10 min in the dark at 4 °C.
10. Centrifuge the suspension for 300 × g for 10 min at 4 °C.
11. Resuspend the pellet in 1 mL of PEB and count the cells using a hemacytometer.
12. Centrifuge the suspension at 300 × g for 10 min at 4 °C.
13. To remove neutrophils, prepare microbeads by mixing 90 µL of PEB buffer with 10 µL of anti-Ly6G microbeads per 10⁷ cells. Mix the cell pellet with the microbead suspension and incubate for 10 min in the dark at 4 °C.
14. To wash out unbound microbeads, add 10 mL of PEB buffer and centrifuge at 300 × g for 10 min at 4 °C.
15. Prepare magnetic columns (MS or LS) on a magnetic stand. Wet the columns by adding 0.5-1.0 mL of PEB buffer.
    NOTE: For tissues with an abundance of immune cells (i.e., spleen), the magnetic column can easily become clogged and stop flowing. Thus, we recommend using LS columns for spleen.
16. Resuspend the cell pellet in 500 µL of PEB buffer. Add the cell suspension to the magnetic column and let the unlabeled cells (i.e., non-neutrophils) run through. Wash the column 2x with 500 µL of PEB buffer. Collect the unlabeled cells (i.e., macrophages and other cells) in a 15 mL conical tube.
17. Centrifuge the suspension at 300 × g for 10 min at 4 °C.
18. Prepare microbeads by mixing 90 µL of PEB buffer with 10 µL of anti-CD11b microbeads per 10⁷ cells. Mix the cell pellet with the microbead suspension and incubate for 15 min in the dark at 4 °C.
19. Repeat steps 1.13-1.15.
20. The labeled macrophages will be bound to the magnetic column. To collect them, detach the magnetic column from the magnetic stand and place it over a new 15 mL conical tube.
21. Add 1 mL of RPMI media supplemented with 0.1% fetal bovine serum (FBS) to the column. Use the plunger to extract the macrophages into the tube. Count the macrophages using a hemacytometer.
    NOTE: The macrophages are now in the tube and ready to use for further assays.

2. Extraction of monocytes from bone marrow

NOTE: Extraction of monocytes from the bone marrow uses a negative selection strategy in which non-monocytes, including lymphocytes, natural killer cells, dendritic cells, erythroid cells, and granulocytes (i.e., neutrophils), are immunomagnetically labeled and removed.

1. Euthanize the mouse in accordance with AALAC and institutional guidelines by isoflurane overdose followed by removal of the heart.
2. Remove the legs from the mouse. In cold HBSS, dissect the muscle away from the tibia and femur until the bones are clean.
3. Carefully cut off both ends of the bone. Using a 10 mL syringe with a 27 G needle, flush out the bone marrow into a 15 mL conical tube with PEB buffer. Centrifuge the suspension at 300 × g for 10 min at 4 °C.
4. Perform steps 1.8-1.12.
5. Add 175 µL of PEB buffer, 25 µL of FcR blocking reagent, and 50 µL of monocyte biotin-antibody cocktail per 5 × 10⁷ of total cells. Incubate for 5 min in the dark at 4 °C.
6. Wash by adding 10 mL of PEB buffer and centrifuge for at 300 × g for 10 min at 4 °C.
7. Collect the pellet and re-suspend the cells in 500 µL of PEB buffer and 100 µL of anti-Biotin microbeads. Incubate for 10 min in the dark at 4 °C.
8. Prepare magnetic columns (LS) on a magnetic stand. Wet the columns by adding 1.0 mL of PEB buffer.
9. Immediately apply the cell suspension to the column. Use a 15 mL tube to collect the flowthrough, which contains the monocytes. The non-monocytes are bound to the magnetic column.
10. Wash the column 2x with 1 mL of PEB buffer. Count the monocytes using a hemacytometer.
11. To prepare the monocytes for the next step, centrifuge at 300 × g for 10 min at 4 °C. For flux analysis, resuspend the cell pellet in 1 mL of RPMI + 0.1% FBS. For flow cytometry, keep the cells in cold PEB and proceed with staining.

3. Preparation of cells for metabolic flux analysis

1. Hydrate the sensor cartridge (the day before the assay)
   1. Remove the sensor cartridge from its package. Remove the sensor plate (green) from the wells and pipette 200 µL of calibrant fluid into each well. Place the sensor plate back on the top of the cartridge plate so that the sensors are submerged in the calibrant fluid.
   2. Place the cartridge into a non-CO₂ humidified incubator (37 °C) overnight.
2. Plating the cells
   1. Plate the cells in a 96-well culture plate in 200 µL of media for at least 1 h. Leave at least one well blank for background measurements (use the same volume of media for blank wells).
      NOTE: The optimal cell number should be determined prior to experimentation. See "Troubleshooting Tips" in the discussion for more details.
   2. When ready to perform the assay, prepare the metabolic flux media (i.e., basal RPMI or DMEM with no glucose, glutamine, pyruvate, etc). For the glycolysis stress test, ensure the basal media is supplemented with 2 mM glutamine. For the mitochondrial stress test, ensure the basal media is supplemented with 2 mM glutamine, 10 mM glucose, and 1 mM pyruvate.
   3. Carefully replace the old media with new basal media by slowly pipetting (be careful not to dislodge any cells). Check the cells under a microscope to make sure they are still present.
   4. Place the cell plate in a non-CO₂ humidified incubator (37 °C).
3. Loading the flux plate
   1. Retrieve the hydrated sensor cartridge/flux plate from the incubator.
   2. Prepare the test compounds for either glycolysis or mitochondrial stress test.
      1. Glycolysis: Using glycolysis basal media, create stock solutions by dissolving glucose in 3 mL, oligomycin in 270 µL (100 µM), and 2-deoxyglucose in 3 mL. Create a working oligomycin solution of 10 µM.
      2. Mitochondrial: Using mitochondrial basal media, create stock solutions by dissolving oligomycin in 630 µL (100 µM), carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) in 720 µL (100 µM), and rotenone/antimycin A in 540 µL (50 µM). Create working solutions of oligomycin (0.5-2.5 µM), FCCP (0.125-2.0 µM), and rotenone/antimycin A (0.5 µM).
         NOTE: We have found that 1.5 µM oligomycin and 2.0 µM FCCP produce the best results.
   3. Load test compounds into ports using the following volumes: port A-20 µL, port B-22 µL, port C-25 µL, port D-27 µL. Optional: Load drug or compound of choice into port A, and test compounds into port B, C, and D to assess acute effects on glycolysis or mitochondrial respiration. If not using additional drug or compound, leave port D empty.
4. Running the assay
   1. Open the software and select Mitochondrial Stress Test.
      NOTE: This program can be used for glycolysis stress test as well, as extracellular acidification rate (ECAR) is measured.
   2. Select the wells that will be used. To subtract the background, ensure that at least 1 well has no cells.
   3. Run the program. Load the flux plate with the lid removed so that the plate is calibrated (~20 min).
   4. After the plate is calibrated, the software will give a prompt to load the cell plate. Take the cell plate out of the incubator, remove the bottom of the flux plate, and place the cell plate in the holder. Press run. The assay will take ~2.5 h to complete.
   5. Analyze the results using metabolic flux software⁷ or other analytical software of choice.

Typical cell numbers obtained from the different tissues depend on the size, age, and sex of the animal. For an adult male mouse (i.e., 16 weeks old, ~30 g), the spleen may yield 3.0-4.0 × 10⁶ macrophages, while the bone marrow (two tibias and two femurs) typically yields 1.0-1.5 × 10⁶ monocytes (Figure 2A). The infarcted heart typically yields high numbers as well, depending on the day post MI^4,5,6. At day 3, the yield is typically ~1.5 × 10⁶ cells/heart. The healthy heart typically has very few macrophages (0.1-0.2 × 10⁵/heart)⁶; thus, if using macrophages from the healthy (non-infarcted heart) as a control, pooling multiple hearts may be necessary. Figure 2B illustrates representative flow cytometry plots from extracted macrophages or monocytes, showing populations that are highly positive for CD11b and negative for Ly6G. Cardiac and splenic macrophages show heterogeneity for the monocyte marker Ly6C, while bone marrow monocytes are highly positive for Ly6C. Representative flow cytometry plots for sorted neutrophils (heart and spleen) and non-sorted cells from the infarcted heart, spleen, and bone marrow (i.e., whole tissue) are displayed in Supplemental Figure S1, Supplemental Figure S2, and Supplemental Figure S3.

The results of the extracellular flux analysis are represented as changes in the ECAR (mPh/min) and oxygen consumption rate (OCR, pmol O₂/min). The results are typically presented in a line graph with ECAR or OCR on the y-axis, and time on the x-axis (Figure 3). Specific measurements are displayed as bar graphs.

From the raw data, one can calculate several metabolic parameters. One can also calculate the OCR/ECAR ratio from each well to give an idea of the reliance on glycolysis versus mitochondrial respiration, which can change under different conditions⁴. For example, we fasted mice (adult male C57BL/6J) overnight to assess changes in macrophage metabolism. A fasted phenotype was confirmed by decreased blood glucose levels and increased blood ketones (Supplemental Figure S4), which decreased circulating leukocytes and neutrophils (Hemavet 950FS Auto Blood Analyzer; Supplemental Figure S5). We isolated bone marrow monocytes and splenic macrophages from fed versus fasted mice and assessed metabolic changes by extracellular flux analysis (Figure 4). In bone marrow monocytes, fasting decreased glycolysis and increased the OCR/ECAR ratio in both the mitochondrial stress test and glycolysis stress test (Figure 4A). In spleen macrophages, fasting did not significantly affect glycolysis but increased spare capacity and ATP-linked respiration (Figure 4B). Figure 4C shows an example of metabolic flux in macrophages isolated from the infarcted heart 3 days post surgery.

Figure 1: Representative schematic of metabolic flux analysis in extracted macrophages/monocytes. Please click here to view a larger version of this figure.

Figure 2: Representative results of cardiac and splenic macrophages and bone marrow monocytes isolated from mice. (A) Representative images of isolated cardiac and splenic macrophages, and bone marrow monocytes (4x); average cell numbers per mouse. Scale bar = 1 mm. (B) Representative flow cytometry plots from cardiac (myocardial infarction Day 3) and spleen macrophages and bone marrow monocytes. CD45 was used as a pan-leukocyte marker, and CD11b was used as a myeloid cell marker. Please click here to view a larger version of this figure.

Figure 3: Representative results of tests. (A) Glycolysis and (B) mitochondrial stress tests. N = 3 independent measurements (splenic macrophages). Mean ± SEM. Abbreviations: ECAR = extracellular acidification rate; OCR = oxygen consumption rate; FCCP = carbonyl cyanide-p-trifluoromethoxyphenylhydrazone. Please click here to view a larger version of this figure.

Figure 4: Impact of overnight fasting on cell metabolism. (A) Bone marrow monocyte and (B) spleen macrophage metabolism. (C) Representative extracellular flux plots from cardiac macrophages 3 days after myocardial infarction. N = 3-4 each group. Abbreviation: FCCP = carbonyl cyanide-p-trifluoromethoxyphenylhydrazone. Please click here to view a larger version of this figure.

Supplemental File 1: Representative flow cytometry plots and blood glucose and ketone levels from fed and fasted mice. Please click here to download this File.

Our method details the rapid extraction of macrophages from bone marrow, spleen, and the infarcted heart, which can then be used to perform downstream analyses such as extracellular metabolic flux analysis. The combination of these two methods is a powerful tool that can quantify metabolic changes in macrophages under different disease or injury states, or metabolic states such as exercise. While our method focused on bone marrow and splenic macrophages, other macrophage reservoirs can be used, such as the peritoneal compartment⁸. Although we are not the first to use this method, we hope that this article will increase exposure and improve standardization across labs.

While the data obtained from these methods gives invaluable insight into the metabolic state of different macrophage compartments, it is highly recommended to use other methods to complement the metabolic flux data to gain a more complete picture of metabolism. For example, combining the metabolic flux data with metabolomics (i.e., quantification of intracellular metabolites), and gene/protein expression and/or activity of metabolic enzymes is often used to create a well-rounded assessment of metabolic states in macrophages^5,9. In many cases, cells from the same animal can be used for different assays.

In our experience, optimization of the protocol may involve troubleshooting at several different steps. One of the most important things is to work quickly and keep the cells cold, which will increase cell yield and viability. Determining the optimal cell number for extracellular flux is another key point, as too few cells will not give a detectable signal, while too many cells will deplete nutrients in the media and thus will not display a response to the injected compounds. For macrophages, we have found that 2.0 × 10⁵ cells from the MI heart and 3.0 × 10⁵ in a 96-well plate give the most reliable and reproducible results. However, this may vary amongst labs. Extra care should be taken to ensure that equal cell numbers between samples and groups are used. Different cell numbers have different responses to the drugs used in the assay and can introduce unwanted variability.

Optimal concentrations of injected drugs included in the assay (oligomycin, FCCP) should also be optimized prior to performing experiments. For example, we have found that 1.5 µM oligomycin and 2.0 µM FCCP produce the most robust responses. Another common error is not allowing the cells enough time following extraction to adhere to the culture dish, which results in the loss of cells when changing to the extracellular flux media. We suggest at least 1 h, and up to 2 h, to allow the cells to adhere.

Pipetting too aggressively can also result in loss of cells; it is recommended to pipette very carefully and check the cells frequently under a microscope. A final point is to work quickly when preparing the flux plate, particularly when performing the glycolysis stress test. Since these cells are starved of glucose prior to the assay, longer incubation times (longer than 1 h in the basal media) can alter the initial response to glucose. We suggest immediately preparing the flux plate after replacing the normal media with the basal media (~30 min) and then calibrating the plate (which takes ~20 min). Interassay variability should always be considered. When making direct comparisons between groups, effort should always be made to perform all groups within the same plate or experiment.

While we described two standard assays that are frequently used (glycolysis stress and mitochondrial stress tests), it should be noted that there are several other standardized assays that can be used to assess other metabolic pathways, such as the mitochondrial fuel flex test¹⁰, fatty acid/palmitate oxidation¹¹, and ATP rate assay¹². New studies have demonstrated the efficacy of the BAM15 uncoupler, which is supplied in the T cell Metabolic Profiling Kit. BAM15 appears to be a more reliable uncoupler than FCCP with less cytotoxic effects¹³. Future studies should explore the potential of this uncoupler in the metabolic assessment of macrophages.

There are some limitations to this method. The first is that due to the number of steps, it can be difficult for one person to perform in a timely manner. In particular, when using higher numbers of samples (up to 8), it is recommended to have two people working together to improve the quality of the results. We have found that it is difficult to process more than eight samples in a day, even with two people. Another limitation is that since the assay is performed ex vivo, there is always the possibility of the extraction procedure causing changes in the cellular phenotype⁶. While there is no obvious way to eliminate this limitation, in vivo imaging techniques such as hyperpolarized magnetic resonance imaging have been used to visualize macrophage lactate production after MI¹⁴ and will hopefully continue to evolve as a complement or replacement for the ex vivo method we describe. A limitation of the extracellular flux method is that while it can capture shifts between glycolysis and OXPHOS, it does not fully reflect changes in ATP production rate within these pathways¹⁵. Furthermore, changes in acidification of the media (ECAR) may not always reflect lactic acid production by glycolysis but can also be influenced by CO₂ generation by the TCA cycle, which is converted to bicarbonate¹⁵. Thus, it may be advantageous for users to calculate the ATP production rate from glycolysis and OXPHOS, which can assess contributions from these pathways to overall ATP production.

The authors have no conflicts of interest to declare.

We would like to acknowledge the funding that supported this work: NIH/NHBLI 166737, NIH R00 HL146888, NIH U54HL169191, NIH/NIGMS P20GM104357, and NIH/NIGMS P30GM149404 for this work.