A Neuron‐Glia Co‐culture System for Studying Intercellular Lipid Transport

Abstract Neurons and glia operate in a highly coordinated fashion in the brain. Although glial cells have long been known to supply lipids to neurons via lipoprotein particles, new evidence reveals that lipid transport between neurons and glia is bidirectional. Here, we describe a co‐culture system to study transfer of lipids and lipid‐associated proteins from neurons to glia. The assay entails culturing neurons and glia on separate coverslips, pulsing the neurons with fluorescently labeled fatty acids, and then incubating the coverslips together. As astrocytes internalize and store neuron‐derived fatty acids in lipid droplets, analyzing the number, size, and fluorescence intensity of lipid droplets containing the fluorescent fatty acids provides an easy and quantifiable measure of fatty acid transport. © 2019 The Authors.


INTRODUCTION
Coupling of lipid metabolism between neurons and glia is an important area of neuroscience research. Astrocytes and microglia have long been known to supply neurons with lipids and cholesterol, which neurons use to build and repair membranes, via lipoprotein particles. Recent work has revealed that under oxidative stress, neurons transfer lipids to lipid droplets in astrocytes via lipid particles (Ioannou et al., 2019). As lipid dysfunction is a common theme in neurodegenerative disease and injury such as stroke, simple assays to quantify lipid transfer from neurons to glia are critical for elucidating the mechanisms involved.
Here, we report step-by-step protocols for performing assays of lipid transfer from neurons to glia (Ioannou et al., 2019). We first describe culturing of primary neurons and glia from postnatal rat pups (Basic Protocol 1) on separate glass coverslips (Support Protocol 1). The neurons are loaded with commercially available fluorescently labeled fatty acids and then incubated with astrocytes in a "sandwich" co-culture system where neurons do not physically contact glia (Basic Protocol 2). Fatty acid transfer can be measured by imaging the appearance of fluorescently labeled fatty acids in the glial cells 1 of 21 ( Fig. 1). We include a detailed protocol for quantification of fatty acid transport using ImageJ (Support Protocol 4). Finally, we provide related protocols for labeling cells with cell type-specific markers or markers for related organelles, such as lipid droplets (Support Protocol 2), and for controlling for phagocytosis and cellular debris (Support Protocol 3). Although our protocol follows fatty acids delivered from neurons to glia, this assay in principle can be used to study lipid transport in the reverse direction (from glia to neurons) or between any two types of cells.
NOTE: All culture incubations are performed in a humidified 37°C, 5% CO 2 incubator unless otherwise specified.

Prepare reagents for culture
1. Sterilize dissection instruments by autoclaving or rinsing with 70% ethanol.
2. Warm neuron medium and plating/glia medium to 37°C before use. Chill dissection medium on ice.
Even if you are only growing neurons, you will still need to use glia medium, which contains serum, to inactivate the papain solution used to dissociate the tissue.
3. Add 2 ml pre-warmed neuron or glia medium (depending on which cell type is to be cultured) to each well of 6-well cell culture plates containing poly-D-lysine-coated coverslips. Incubate at 37°C before starting dissections to allow the temperature to equilibrate.
4. Make 2× papain solution by dissolving one vial of papain (ß120 U enzyme for 10 pups) in 5 ml ice-cold dissection medium. Keep on ice.
7. Hold each head from the sides. Starting at the base of the brain, using extra-fine Bonn scissors, cut skull along the center, toward the front of the head. Be careful not to cut brain itself. Make two small cuts through skull, away from the initial incision.
8. Using Dumont-style #3 forceps, carefully peel skull off and to the sides.
9. Using a stainless steel laboratory spoon, scoop out brain and place it in a 60-mm plate containing ice-cold dissection medium (see step 5). Repeat steps 6 to 9 to collect the desired number of brains. Collect two brains per 60-mm plate.
10. Replace dissection medium in each 60-mm plate with fresh ice-cold dissection medium.
This step helps wash away some of the blood and damaged tissue from the brains. The brains need to be submerged in medium at all times.
The #3 forceps and spoon are now considered "dirty" and should not be used for any of the following steps.
11. Prepare a fresh ice pack (see step 5) to be used during dissection: again, fill a 10-cm cell culture plate or petri dish with ice, but keep lid on this time. Place one 60-mm plate containing two brains on top of ice pack.
Keeping the lid off the ice pack will keep your samples colder but will cause the 60-mm plate to slide around, making the dissection more difficult.
12. Place ice-packed plate on the stage of a dissection microscope with a light source. Position first brain upright and hold it in place by gently gripping cerebellum using Dumont-style #3 forceps.
We find 0.5× magnification to be optimal for this step.
13. Gently run an iris spatula between two hemispheres of the cortex to separate them. Use convex side of the spatula to gently open one hemisphere so that the hippocampus faces upward.
The hippocampus is the C-shaped structure that runs parallel to the cortex.
14. With the hippocampus facing up, place spatula between the hippocampus and the cortex. Using the spatula, separate tissue where the hippocampus meets the cortex to remove the hippocampus.
15. Using a 1-ml pipet tip that has been cut with sterile scissors to make the opening wider, transfer hippocampus in a new 60-mm cell culture plate containing fresh ice-cold dissection medium on an open ice pack (see step 5).
16. Collect all remaining hippocampi before moving on to the next step. 18. Using #5 forceps and a disposable scalpel, cut hippocampus into four pieces. Using a cut 1-ml pipet tip (see step 15), transfer hippocampal pieces into a 15-ml conical tube containing fresh cold dissection medium.
Hippocampi from multiple animals can be pooled in one tube.
19. Warm 2× papain solution from step 4 in a 37°C water bath.
Ioannou et al.

of 21
20. Wash hippocampal pieces three times with 5 ml cold dissection medium. To do this, allow pieces to fall to the bottom of the tube and aspirate dissection medium with a 1-ml pipet. Dissociate and plate neurons and/or glia 22. Using a 1-ml pipet, slowly remove dissection medium containing papain. Be careful not to aspirate tissue, as it will be extremely sticky at this point. Wash two times with 5 ml plating/glia medium.
23. Resuspend tissue in 3 ml glia medium. Gently triturate tissue by pipetting up and down, first 10 times with a 10-ml pipet and then 10 times with a 1-ml pipet tip.
24. Assemble a 70-µm nylon cell strainer over a 50-ml conical tube. Pass dissociated cells through the cell strainer. Rinse 15-ml conical tube with an additional 5 ml glia medium to collect any remaining cells that stuck to the tube and pass that medium through cell strainer as well.
25. To count the cells and assess their viability, transfer 10 µl cells to a microcentrifuge tube and add 10 µl of 0.4% trypan blue solution. Count cells using a hemocytometer.

Viable cells are colorless, whereas dead and damaged cells are blue.
26. Split suspended cells from step 24 into two 15-ml conical tubes. Centrifuge 5 min at 150 × g.

27.
Resuspend cells in either neuron medium or glia medium, depending on the cells to be grown, and plate desired number of cells onto the prepared coverslips in 6-well plates (see step 3).
Generally, the desired number of cells per coverslip is 2 × 10 5 for neuron cultures or 1.5 × 10 5 for glial cultures. Cells should not be kept in suspension for too long, as the cells clump together and the viability can be compromised.
Neurons secrete growth factors that support the health of the culture. For this reason, plating neurons too sparsely may negatively affect the health of the culture. However, plating neurons too densely will result in a higher number of contaminating astrocytes in the neuron culture, even in the presence of AraC.
Glia will divide, and therefore, more or fewer cells may be plated depending on the timing of the experiment. One consideration is that when glia are plated more sparsely, a larger number of microglia will grow.
28. A total of 6 hr after plating the cells, replace approximately half of the medium by gently removing 1 ml medium from each well and adding 1 ml fresh pre-warmed medium to the side of the well.
The cells will be weakly attached at this point, so be careful not to detach the cells.
29. The next day, completely replace medium in each well with 2 ml neuron or glia medium.
30. To reduce the growth of glia in neuronal cultures, add 2 µM AraC to culture medium 2 days after plating (see step 27).

of 21
This is critical to study lipids originating in neurons. Both astrocytes and microglia secrete lipoprotein particles, and therefore, it is important to have neuron cultures that are as pure as possible. This may not be necessary if you wish to study protein transfer and will express a protein in neurons using a neuron-specific promoter.
31. Replace half of the medium every 3 to 4 days with neuron or glia medium, depending on type of cell being grown.
If AraC was added, this will dilute out over time.
Although mixed glial cultures primarily contain astrocytes, microglia and oligodendrocytes are also present. We recommend immunostaining with antibodies to identify the cell type involved in internalizing transferred lipid or protein (see Support Protocol 2). Pure glial cultures can be achieved by a variety of methods. Both shaking and mild trypsinization protocols take advantage of the varying adherence strengths of different cell types (Jana, Jana, Pal, & Pahan, 2007;Saura, Tusell, & Serratosa, 2003). These protocols are cost and time effective and easy to employ; however, they have a small degree of cell contamination. If purity is critical for the experiments, we recommend immunopanning protocols previously described (Foo et al., 2011).

PREPARATION OF POLY-D-LYSINE-COATED COVERSLIPS
It is important to etch and clean the coverslips before coating them. Etching, in this case with a strong base, allows the poly-D-lysine to adhere to the glass and, consequently, the neurons to attach. Cleaning removes dust and debris from the glass, which is optimal for imaging. Coverslips can be prepared in advance; however, we found that coating the coverslips with poly-D-lysine overnight immediately prior to dissection (Basic Protocol 1) yields the healthiest neuron cultures.

Etch and clean coverslips
1. Prepare 1 M KOH solution by dissolving 56 g KOH in 1 L distilled pure water in a 1-L glass bottle.
Unused KOH solution can be stored at room temperature and used later. 2. Place individual 25-mm round glass coverslips in ceramic coverslip holders using forceps and place coverslip holders in Nalgene jars.
3. Fill jars with enough 1 M KOH to fully submerge the coverslips (ß100 ml). Take care to pour KOH solution slowly so as to not dislodge the coverslips from the holders.

Place jars in an ultrasonic water bath, with the lids slightly open to allow airflow.
Fill bath with enough water to be level with the KOH in the jars. Sonicate for 1 hr on high.
Do not fill the bath past the jar lid, as this will risk contaminating the coverslips with nonsterile bath water.
5. Rinse coverslips three times with distilled pure water by removing the ceramic holders containing coverslips from the jars, emptying the jars, refilling the jars with fresh distilled pure water, and submerging the holders containing coverslips in the fresh water.
7. Place jars in the ultrasonic water bath, with lids slightly open to allow airflow, and sonicate for 1 hr on high.
8. Remove ceramic holders containing coverslips, pour out ethanol, and refill the jars with 20% ethanol in distilled pure water to fully submerge the coverslips. Close lids tightly.
9. Store coverslips in 20% ethanol solution at 4°C until ready for use. When ready, open jars in a sterile culture hood and remove coverslips with sterile forceps to prevent contamination of the coverslips.
11. Place wax spacers in a 60-mm or 10-cm cell culture plate and sterilize wax by submerging in 100% ethanol for 5 min. Aspirate ethanol in a sterile culture hood and let wax pieces air-dry completely.
Ioannou et al.

of 21
12. The day before culturing cells, using sterile forceps, place cleaned coverslips (see step 9) in the wells of a 6-well or 35-mm culture dish and let dry thoroughly. Aspirate any remaining ethanol to speed the drying process.
13. Using sterile forceps, place three wax spacers (see step 11) on dry coverslips in a triangular pattern. Press gently so that wax adheres to the glass (Fig. 2 15. Add 250 µl of 1× poly-D-lysine working solution to center of each coverslip. Allow poly-D-lysine to spread over the surface of the coverslip. Put lid back on the cell culture dish and incubate coverslips a 37°C cell culture incubator for ࣙ2 hr (preferably overnight).
Surface tension will keep the poly-D-lysine on the coverslips, without spreading to the plastic dish. Be careful not to bump the dish to avoid causing the poly-D-lysine to spill onto the plastic. If this happens, fewer cells will grow on the coverslips.
16. Wash coverslips three times with 2 ml DPBS and proceed to Basic Protocol 1 (add desired medium in step 3 and proceed with plating cells in step 27).

LIPID TRANSFER ASSAY
This protocol describes how to use "sandwich"-style co-cultures to assay lipid transfer from neurons to glia. This protocol specifically follows the trafficking itinerary of the fluorescently labeled, saturated fatty acid analog BODIPY 558/568 C12 (Red-C12) into lipid droplets labeled with BODIPY 493/503 (Fig. 3). However, it is possible that this protocol could be used to investigate the trafficking of other fluorescently labeled lipids.

of 21
Figure 3 Appearance of neuron-derived fatty acids (Red-C12) in glial lipid droplets. After the transfer assay, glia were fixed, stained with BODIPY 493/503 (BD-493) to label lipid droplets, and imaged using a Zeiss 880 confocal microscope with a 63× objective. Scale bars are 10 µm.

Sterile forceps Tinfoil
Microscope slide (Thermo Fisher Scientific, 112-550-123) Kimwipes Confocal or widefield microscope with 63× objective NOTE: All solutions and equipment coming into contact with cells must be sterile, and proper sterile technique should be used accordingly.
NOTE: All culture incubations are performed in a humidified 37°C, 5% CO 2 incubator unless otherwise specified.
Addition of calcium to a final concentration of 2 mM is required for neuronal activity.
If storage is necessary, store at 4°C.
If storage is necessary, store at −20°C and warm to room temperature before use.
Store and use at room temperature.
Transfer assay 5. One day prior to the assay, load neurons on glass coverslips with Red-C12 by replacing half of the medium with 1 ml fresh pre-warmed neuron medium (leaving 1 ml neuron-conditioned medium) containing 1 µl of 5 mM Red-C12 (see step 2; 2.5 µM final concentration).
6. Approximately 18 hr later, wash neurons two times with 2 ml pre-warmed DPBS and incubate neurons in 2 ml fresh neuron medium for 1 hr at 37°C.
7. Wash neurons and glia on glass coverslips two times with 2 ml DPBS.
Ioannou et al.

of 21
8. Add 1 ml pre-warmed HBSS plus calcium (see step 1) to glia. Use sterile forceps to lift coverslips carrying neurons and place them face-down onto coverslips carrying astrocytes (Fig. 2, right panel).
The dental wax spacers will keep the coverslips from directly touching.
If the neurons are to be imaged, remove the wax spacers from the coverslips prior to mounting.
11. To fix the glial cells, wash coverslips two times with 2 ml DPBS, fix cells for 10 min in 1 ml of 3% paraformaldehyde (see step 3; protect from light using tinfoil), and wash two times with DPBS.
If storage is necessary, store at −20°C and warm to room temperature before use.
14. Incubate fixed glial cells from step 11 in 1 ml of 5 µg/ml BODIPY 493/503 working solution for 10 min at room temperature (protect from light using tinfoil) and wash three times with 2 ml DPBS. 21. Acquire a Z-stack with 0.5-µm step size per field of view.
We typically acquire 10 images per coverslip. At least one image of the DAPI staining is necessary for quantification.

IMMUNOSTAINING WITH CELL TYPE-SPECIFIC ANTIBODIES
As primary cultures often contain multiple cell types (mixed glial cultures), it may be useful to stain with antibodies to identify the cell types involved in the transfer assay (Basic Protocol 2). Here, we describe a protocol for immunostaining with antibodies that recognize neurons, astrocytes, and microglia (Fig. 4). This allows assessment of transfer of Red-C12 into different cell types (Fig. 5). In this particular protocol, the secondary antibodies were selected in order to perform four-color imaging, including imaging of the Red-C12 from Basic Protocol 2. This protocol can be adapted to include BODIPY 493/503, as described above, or to stain for oligodendrocytes and organelle-specific markers with appropriate antibodies. Immunostaining of different cell types in culture. Neuronal cultures in the absence of AraC and mixed glial cultures were immunostained for the neuronal marker tubulin β3 (Tuj), astrocyte marker GFAP, or microglia marker Iba1. Cells were imaged using a Zeiss 880 confocal microscope with a 20× objective. Tiling of 2 × 2 was used, with 10% overlap. Scale bars are 100 µm.

Figure 5
Immunostaining of astrocytes and microglia after transfer assay. After the transfer assay, glia were fixed and immunostained with anti-GFAP and anti-Iba1 to label astrocytes and microglia, respectively. Cells were imaged using a Zeiss 880 confocal microscope with a 63× objective. Scale bars are 10 µm. 2. Block and permeabilize cells by incubating in 1 ml blocking buffer for 1 hr at room temperature while shaking gently on an orbital platform shaker.
4. Incubate cells with primary antibodies for 1 hr at room temperature or overnight at 4°C while shaking gently on an orbital shaker.

If incubating overnight, the orbital shaker can be placed in a cold room.
Ioannou et al.

of 21
5. Wash three times with 2 ml wash buffer, 15 min each time, at room temperature while shaking gently on an orbital shaker. 7. Incubate cells with secondary antibodies for 1 hr at room temperature while shaking gently on an orbital shaker.
8. Wash three times with 2 ml wash buffer, 15 min each time, at room temperature while shaking gently on an orbital shaker.

Mount and image coverslips
9. Place a drop of mounting medium on a microscope slide.
10. Using forceps, submerge a coverslip carrying fixed cells (see step 8) in water and touch side of the coverslip to a Kimwipe to remove excess water.
11. Place coverslip gently onto the drop of mounting medium (see step 9), with the cells facing down. Be careful to avoid trapping bubbles.
12. Gently aspirate excess mounting medium from sides of the coverslip. Be careful not to move or twist the coverslip as this can damage the cells. If present, leave excess mounting medium on top of the coverslip in place for now. Cover to protect from light and let dry overnight at room temperature. Store mounted cells at 4°C protected from light until ready to image.
13. Before imaging, gently clean surface of the coverslip using a Kimwipe and lens cleaner or 70% ethanol.

CONTROLLING FOR PHAGOCYTOSIS AND TRANSFER OF CELL DEBRIS
Phagocytosis is an important mechanism that astrocytes and microglia use to take up debris and dying cells following injury. Phagocytosis of cell debris containing membranes labeled with fluorescently tagged fatty acids can contribute to fatty acid transfer and lipid droplet formation. Therefore, it is important to distinguish between fatty acid transfer due to phagocytosis and transfer mediated by lipid particles. Here, we outline two control experiments to quantify the contribution of phagocytosis of cell debris in the fatty acid transfer assay (Basic Protocol 2): a cell filtration protocol and a centrifugation protocol. These strategies can be used in parallel to the experiments described in Basic Protocol 2. Generally, neurons are loaded with Red-C12, and then after washing, they are placed in HBSS plus calcium instead of in a sandwich culture. After an incubation period equivalent to that of the transfer assay (Basic Protocol 2), the HBSS plus calcium is removed from Ioannou et al.
the neurons. Glial cells are then incubated either with the conditioned HBSS plus calcium or with conditioned HBSS plus calcium that has been filtered or centrifuged; as cellular debris and apoptotic bodies shed from dying neurons are roughly 0.8 to 5 µm in diameter, they can be removed by either filtration or centrifugation (Crescitelli et al., 2013).

Additional Materials (also see Basic Protocol 2)
2-ml microcentrifuge tubes 0.20-µm filter (Corning, 431219) 3.5-ml polycarbonate ultracentrifugation tubes (Beckman Coulter, 362305) Beckman Coulter Optima MAX-TL centrifuge and TLA-110 fixed-angle rotor 37°C water bath NOTE: All solutions and equipment coming into contact with cells must be sterile, and proper sterile technique should be used accordingly.
NOTE: All culture incubations are performed in a humidified 37°C, 5% CO 2 incubator unless otherwise specified.

Prepare cells
1. Approximately 18 hr prior to the assay, load neurons on glass coverslips with Red-C12 by replacing half of the medium with 1 ml fresh neuron medium (leaving 1 ml neuron-conditioned medium) containing 1 µl of 5 mM Red-C12 (see Basic Protocol 2, step 2; 2.5 µM final concentration). For the filtration protocol, prepare at least two coverslips or for the centrifugation protocol, prepare three coverslips.
2. Wash neurons two times with 2 ml pre-warmed DPBS and incubate neurons with 2 ml fresh neuron medium for 1 hr at 37°C.
4. Add 1.5 ml HBSS plus calcium (see Basic Protocol 2, step 1) to each neuronal coverslip and incubate for 4 hr at 37°C.

Filtration protocol
6a. Collect neuron-conditioned HBSS from two coverslips into two separate 2-ml microcentrifuge tubes.
10a. Fix, mount, and image cells as described in Basic Protocol 2, step 11 and steps 15 to 21.
7b. Set one tube on ice, protected from light (control).

Ioannou et al.
14 of 21 8b. Centrifuge second tube for 20 min at 15,000 × g. Transfer supernatant to a clean ultracentrifugation tube and store on ice protected from light.
This low-g sample is depleted of apoptotic bodies and cell debris.
9b. Centrifuge third tube for 2.5 hr at 250,000 × g. Transfer supernatant to a clean ultracentrifugation tube and store on ice protected from light.

This high-g sample is depleted of lipid particles and extracellular vesicles in addition to apoptotic bodies and cell debris.
10b. Briefly warm control conditioned HBSS (see step 7b) and low-g (see step 8b) and high-g (see step 9b) supernatants in a 37°C water bath.
12b. Add control-conditioned HBSS or low-g or high-g supernatant to glial coverslips and incubate for 4 hr at 37°C.
13b. Fix, mount, and image cells as described in Basic Protocol 2, step 11 and steps 15 to 21.

ANALYSIS OF LIPID DROPLETS WITH ImageJ
Here, we describe a protocol for analyzing transfer of labeled fatty acids into lipid droplets (Basic Protocol 2) using ImageJ, an open-source image-processing program. These experiments can be performed with a range of imaging modalities. We typically use confocal imaging to analyze co-localization due to superior contrast and widefield imaging with an electron-multiplying CCD (EMCCD) camera to collect images for quantification due to greater sensitivity and speed. Background subtraction is especially important if images are collected using a widefield microscope; this allows more accurate thresholding for particle detection (Fig. 6).
Useful tip: You can automate your analysis in ImageJ without knowledge of programming. In the Plugins menu, click on Macros and then select Record before analyzing one image. After you have completed the analysis, hit the Create button to save the code written by ImageJ.

Figure 6
Red-C12 detection to quantify fatty acid transfer. Thresholding in the absence of background subtraction can give inaccurate particle counts. Lipid droplets clustered together in the perinuclear region may be counted as one large lipid droplet, and smaller peripheral lipid droplets are not detected at all. Subtracting a Gaussian blurred duplicate of the original image removes the background, allowing for more accurate particle detection. Depicted here are maximum intensity projections of Z-stacks with 0.5-µm step size acquired using a Nikon Eclipse TiE widefield microscope with a 60× objective. Scale bars are 10 µm.
9. Adjust threshold so that lipid droplets are highlighted in red, either manually by using the sliding bars or using the custom thresholding algorithms.
It is not necessary to press Apply.
The Otsu, Yen, and Intermodal algorithms work well for detecting lipid droplets. The most important consideration, however, is that the same thresholding method is used consistently when comparing images from different treatment groups.
10. Select Analyze > Particles. Set size to "2-infinity pixel units" to remove single pixels. Check Summarize box. Click OK.
This reduces the likelihood of noise being detected as a particle.

It can be helpful to show Masks to visualize what is being included/excluded in your analysis so that you may modify your parameters accordingly.
11. Save data directly from the Summary window or copy and paste data into an Excel file.
Image analysis results will be added sequentially to the Summary window without erasing the previous analysis.
Quantify number and size of nuclei 12. To set measurements, select Analyze > Set Measurements.
13. Select Area and then select OK. 14. Select DAPI image.
17. Adjust threshold so that nuclei are highlighted in red, either manually by using the sliding bars or using the custom thresholding algorithms.
It is acceptable to have some background pixels detected after thresholding.
18. Select Analyze > Particles. Set size to "25-infinity pixel units" to remove single pixels. Check Summarize box. Click OK.
This reduces the likelihood of noise being detected.

It can be helpful to show Masks to visualize what is being included/excluded in your analysis so that you may modify your parameters accordingly.
19. Save data directly from the Summary window or copy and paste data into an Excel file.
Image analysis results will be added sequentially to the Summary window without erasing the previous analysis.

Calculating the number of nuclei can be done manually if the number of cells per image (using a 63× objective) is small (2 to 4 cells).
If the experiment is modified such that the number of nuclei is larger, then automated nuclei counting using ImageJ can be more time efficient and accurate. This method is, however, required for evaluating the size of nuclei to determine whether a specific treatment affected cell health. If there is no difference between the samples, phagocytosis of cell debris can be excluded.

Ioannou et al.
22. If the centrifugation protocol was performed (see Support Protocol 3), compare number of Red-C12-positive particles in the control sample versus low-g and high-g supernatants.
If there is a reduction in the number of particles following low-g centrifugation, fatty acids are being transferred by cell debris. If there is a reduction in the number of particles following high-g centrifugation, small, dense carriers such as extracellular vesicles or lipoprotein particles are mediating fatty acid transfer. If there is no difference in the number of particles detected after either low-g or high-g centrifugation, free fatty acids are mediating transfer. See Fig. 7. 23. As a secondary measure of cell health, compare mean nuclei size and number between treatment groups.
If differences between groups are observed, then alterations in cell health may be influencing the results and should be considered when interpreting all results.

Glucose, 20%
Add 10 g D-glucose (Thermo Fisher Scientific, D16-500) to 50 ml ddH 2 O and mix until fully dissolved. It can be helpful to warm the solution and/or use a vortex. Filter solution using a sterile syringe with a Luer-Lok tip (BD Biosciences, 302832) and 0.20-µm filter (Corning, 431219). Store ࣘ6 months at 4°C.

Neuron medium
Use NbActiv4 medium (BrainBits, Nb4-500), which already contains the supplements needed to support neuronal growth. Optionally, add 5 ml 100× antibioticantimycotic (Thermo Fisher Scientific, 15240096; 1× final) to 500 ml NbActiv4 to protect cells from contamination. Store ࣘ2 months at 4°C or according to expiry date listed by the manufacturer.

Background Information
The ability of glia to supply neurons with lipids and cholesterol has been appreciated for many decades. Astrocytes and microglia synthesize and secrete lipoprotein particles composed of a cholesterol and neutral lipid core surrounded by a phospholipid monolayer and apolipoproteins. Neurons use the lipids and cholesterol from glia-derived lipoprotein particles in order to build membranes and form synapses (Holtzman et al., 1995;Mauch et al., 2001). There are many challenges associated with studying lipid transport in the brain or in neuron-glia co-cultures. Lipids cannot be targeted by the same genetic tools commonly used for studying proteins, such as tagging with fluorescent proteins in a cell type-specific manner. Although numerous fluorescently tagged lipids are commercially available, there are limitations to use of exogenously applied lipids when studying cells of the nervous system. For example, fatty acid transport has been demonstrated between fibroblasts by co-culturing donor cells (loaded with fluorescent fatty acids) with acceptor cells in the same dish (Rambold, Cohen, & Lippincott-Schwartz, 2015); neurons and glia, however, need to be grown in different types of medium to facilitate their differentiation and survival and require several days to mature. Phagocytosis also becomes a confounding issue, as primary cells have more cell death upon initial plating compared to immortalized cell lines. These factors preclude the ability to perform fatty acid transfer assays in mixed neuron-glial cultures as previously described. Factors secreted by astrocytes and detected by neurons can be probed using a "sandwich" coculture system (Jones, Cook, & Murai, 2012). Combining the "sandwich" co-culture system with the lipid transfer assay overcomes the difficulties associated with primary neuron cultures and allows investigation of lipid trafficking between neurons and glia. Using this lipid transfer assay, we recently discovered that neurons also transfer lipids to astrocytes and microglia (Ioannou et al., 2019). This transfer is dependent on neuronal expression of ApoE, but much remains to be discovered regarding the mechanisms of lipid transport from neurons to astrocytes. The described lipid transfer assay will allow these mechanisms to be dissected in order to gain a fuller understanding of how neurons and astrocytes couple their lipid metabolism.

Critical Parameters
As with all experiments involving primary neuron cultures, the health of the neurons is critical. If neurons are unhealthy, they may release lipids via apoptotic bodies or via debris from dead cells, which will interfere with specific experiments. Testing for the mode of transport will determine whether this is a factor (Support Protocol 3). Another important consideration is the purity of the neuronal culture. As glia secrete lipoprotein particles that contribute to lipid transfer, glial contamination needs to be minimal to avoid misinterpretation of results. We recommend always keeping the percentage of contaminating glia to <2% (Support Protocol 2).

Troubleshooting
Please refer to Table 1 for a troubleshooting guide.

Understanding Results
As described above, phagocytosis of apoptotic bodies and cell debris can contribute to fatty acid transfer and lipid droplet formation (Support Protocol 3). This means that if a desired treatment induces neuronal cell death, more fatty acid transfer will likely be observed. In order to study the mechanisms of lipid particle-mediated transfer of fatty acids, one must carefully control for and monitor cell death. In addition, we include control experiments (Support Protocol 3) to directly test the mechanism of transfer involved. By using low-g and high-g centrifugation, various components can be depleted from neuronconditioned medium. If fatty acid transfer is abolished by low-g centrifugation, then apoptotic bodies and cell debris are mediating the transfer. If fatty acid transfer is abolished only by high-speed centrifugation, then dense carriers such as lipoprotein particles or extracellular vesicles mediate the transfer. Finally, if neither low-g nor high-g centrifugation affects the rates of fatty acid transfer, then the transfer is likely mediated by free fatty acids bound to carrier molecules such as albumin.
Another consideration is that because fluorescently labeled fatty acids accumulate in glial lipid droplets, the presence of glial lipid droplets is essential for detecting and quantifying transfer. Some treatments can influence fatty acid transfer and/or lipid droplet numbers in different ways. For example, neuronal activation using the chemogenetic receptor system hM3D(Gq) increases fatty acid transfer to glia. However, glutamate released by neural activity promotes glial consumption of fatty acids stored in lipid droplets, resulting in a reduction in the number of lipid droplets. Therefore, if glutamate or N-methyl-D-aspartate (NMDA) is added into the transfer assay, one might mistakenly conclude that there is a reduction in fatty acid transfer, when instead fatty acids/lipid droplets are being consumed at a faster rate. Therefore, it is important to analyze both the number of glial lipid droplets containing transferred fatty acids and the total number of lipid droplets.

Time Considerations
Preparing primary cell cultures (Basic Protocol 1) should take ß3 hr, including setup. Once the cells are plated, they will require medium changes after 6 hr and 24 hr and then every 3 to 4 days for maintenance. Neurons are ideally used after between 1 and 2 weeks in culture. Astrocytes can be used by 1 week. If older astrocytes are desired, they can be Ioannou et al.

of 21
initially plated on culture dishes and split to avoid over-confluence. For Support Protocol 1, it will take ß3 hr to clean the coverslips, which can be stored for later use, and coating the coverslips requires an overnight incubation. For lipid transfer assays (Basic Protocol 2), cells are treated with fluorescently labeled fatty acids the night before the assay. On the day of the assay, the labeled neurons are incubated in fresh medium for 1 hr prior to the assay. The sandwich co-culture takes 4 hr. Following the assay, cells can be fixed for 10 min and mounted. If immunolabeling is performed (Support Protocol 2), this will take an additional 24 hr. Experiments controlling for phagocytosis (Support Protocol 3) take ß10 hr or ß14 hr for the filtration and centrifugation assays, respectively. Images can be analyzed (Support Protocol 4) in ß15 min per coverslip.