We present a flow cytometry method to identify simultaneously different cell types retrieved from mouse brain or spinal cord. This method could be exploited to isolate or characterize pure cell populations in neurodegenerative diseases or to quantify the extent of cell targeting upon in vivo administration of viral vectors or nanoparticles.
Recent advances in viral vector and nanomaterial sciences have opened the way for new cutting-edge approaches to investigate or manipulate the central nervous system (CNS). However, further optimization of these technologies would benefit from methods allowing rapid and streamline determination of the extent of CNS and cell-specific targeting upon administration of viral vectors or nanoparticles in the body. Here, we present a protocol that takes advantage of the high throughput and multiplexing capabilities of flow cytometry to allow a straightforward quantification of different cell subtypes isolated from mouse brain or spinal cord, namely microglia/macrophages, lymphocytes, astrocytes, oligodendrocytes, neurons and endothelial cells. We apply this approach to highlight critical differences between two tissue homogenization methods in terms of cell yield, viability and composition. This could instruct the user to choose the best method depending on the cell type(s) of interest and the specific application. This method is not suited for analysis of anatomical distribution, since the tissue is homogenized to generate a single-cell suspension. However, it allows to work with viable cells and it can be combined with cell-sorting, opening the way for several applications that could expand the repertoire of tools in the hands of the neuroscientist, ranging from establishment of primary cultures derived from pure cell populations, to gene-expression analyses and biochemical or functional assays on well-defined cell subtypes in the context of neurodegenerative diseases, upon pharmacological treatment or gene therapy.
Gene and drug delivery technologies (such as viral vectors and nanoparticles) have become a powerful tool that can be applied to gain better insights on specific molecular pathways altered in neurodegenerative diseases and for development of innovative therapeutic approaches1,2,3. Optimization of these tools relies on quantification of: (1) the extent of penetration in the CNS upon different routes of administration and (2) targeting of specific cell populations. Histological analyses are usually applied to visualize fluorescent reporter genes or fluorescently-tagged nanoparticles in different CNS areas and across different cell types, identified by immunostaining for specific cell markers4,5. Even though this approach provides valuable information on the biodistribution of the administered gene or drug-delivery tools, the technique can be time-consuming and labor-intense since it requires: (1) tissue fixation, cryopreservation or paraffin-embedding and slicing; (2) staining for specific cellular markers sometimes requiring antigen retrieval; (3) acquisition by fluorescence microscopy, which usually allows the analysis of a limited number of different markers within the same experiment; (4) image processing to allow proper quantification of the signal of interest.
Flow cytometry has become a widely used technique which takes advantage of very specific fluorescent markers to allow not only a rapid quantitative evaluation of different cell phenotypes in cell suspensions, based on expression of surface or intracellular antigens, but also functional measurements (e.g., rate of apoptosis, proliferation, cell cycle analysis, etc.). Physical isolation of cells through fluorescent activated cell sorting is also possible, allowing further downstream applications (e.g., cell culture, RNAseq, biochemical analyses etc.)6,7,8.
Tissue homogenization is a critical step necessary to obtain a single cell suspension to allow reliable and reproducible downstream flow cytometric evaluations. Different methods have been described for adult brain-tissue homogenization, mainly with the aim to isolate microglia cells9,10,11; they can be overall classified in two main categories: (1) mechanical dissociation, which uses grinding or shearing force through a Dounce homogenizer (DH) to rip apart cells from their niches and form a relatively homogenized single cell suspension, and (2) enzymatic digestion, which relies on incubation of minced tissue chunks at 37 °C in the presence of proteolytic enzymes, such as trypsin or papain, favoring the degradation of the extracellular matrix to create a fairly homogenized cell suspension12.
Regardless of which method is utilized, a purification step is recommended after tissue homogenization to remove myelin through centrifugation on a density gradient or by magnetic selection9,12, before moving to the downstream applications.
Here, we describe a tissue processing method based on papain digestion (PD) followed by purification on a density gradient, optimized to obtain viable heterogeneous cell suspensions from mouse brain or spinal cord in a time-sensitive manner and suitable for flow cytometry. Moreover, we describe a 9-color flow cytometry panel and the gating strategy we adopted in the laboratory to allow the simultaneous discrimination of different CNS populations, live/dead cells or positivity for fluorescent reporters such as green fluorescent protein or rhodamine dye. By applying this flow cytometric analysis, we can compare different methods of tissue processing, i.e., PD versus DH, in terms of preservation of cellular viability and yields of different cell types.
The details we provide herein can instruct decision on the homogenization protocol and the antibody combination to use in the flow cytometry panel, based on the specific cell type(s) of interest and the downstream analyses (e.g., temperature-sensitive applications, tracking of specific fluorescent markers, in vitro culture, functional analyses).
All methods described here have been approved by the Institutional Animal Care and Use Committee (IACUC) of Dana Farber Cancer Institute (protocol number 16-024).
1. Preparation of solutions needed for the experiment
2. Animal euthanasia by intracardiac perfusion and tissue dissection
NOTE: Eight-week-old C57BL/6J mice, either sex, were used in the experiments. Perfusion with PBS solution is performed to eliminate blood contamination from organs, before proceeding with tissue digestion.
3. Enzymatic digestion of brain and spinal cord
NOTE: Volumes described in this section are enough for digestion of one-half brain or spinal cord.
4. Mechanical homogenization of brain and spinal cord
NOTE: Volumes described in this section are enough for homogenization of one-half brain or spinal cord. The protocol described in this section can be used as a method alternative to the one described in section 3, depending on user need as discussed below.
5. Debris removal
NOTE: Removal of debris, composed mainly of undigested tissue and myelin sheaths, is a critical step to allow efficient staining of the tissue homogenate for subsequent flow cytometric analyses.
6. Staining for flow cytometric evaluation of multiple cell types
We compared two different homogenization methods (DH versus PD) applied to mouse brain and spinal cord, to test the efficiency in retrieving different viable cell types suitable for downstream applications. To do so, we exploited a 9-color flow cytometry panel designed to characterize, in the same sample, different CNS cell types including microglia, lymphocytes, neurons, astrocytes, oligodendrocytes and endothelium.
Brain and spinal cord tissues were retrieved from different mice (n ≥ 6), split in two halves longitudinally, weighed and processed in parallel by applying either mechanical disruption using the Dounce homogenizer (DH method) or gently minced and digested enzymatically using the commercial NTDK based on papain (PD method) (Figure 1A). After debris removal, cells from the brain or the spinal cord were diluted 1/10 or 1/2−1/5, respectively, in Trypan blue to determine cell yield and viability with a Neubauer chamber (Figure 1B,C). The DH method overall produced a higher cell yield from both brain and spinal cord. However, majority of the cells retrieved were dead, resulting in only 13.8% ± 3.3% of viable cells in the brain and 10.5% ± 1.5% in the spinal cord (Figure 1B). Many of the dead cells formed aggregates (Figure 1C); this phenomenon could be due to the presence of highly interconnected cell networks (like the endothelial and glial cells lining the CNS vasculature) that could not be disaggregated by the shearing force applied with the DH. These aggregates of death cells were likely not removed by the density gradient and ended up in the final cell pellet used for cytofluorimetric analysis. On the contrary, the PD method determined an overall better preservation of cellular viability (90.6% ± 0.6% in the brain and 85.2% ± 2.8% in the spinal cord). Papain is able to digest the extracellular matrix and cell-to-cell junctions efficiently, leading to a more uniform single cell suspension. Some of the cells that die during the mincing process could be further digested by papain leading to formation of cell debris that are more efficiently separated through the density gradient. Overall, this likely determined a better preservation of cell viability with PD method, despite a slightly lower cell yield as compared with the DH method.
An aliquot of 100 µL from the brain and spinal cord cell suspensions was stained with the antibody mix (Table 1) and analyzed by flow cytometry with a 9-color panel. Figure 2A shows the gating strategy used to identify the different cell types from the brain and spinal cord cell suspensions. Briefly, the first gate identifies the general population according to forward scatter (FSC) and side scatter (SSC), excluding small cell debris. Then live (7-AAD-) cells are identified. Within total live cell population, CD45+ and CD45- cells are highlighted. Within the CD45+ gate, CD45+CD11b+ microglia/macrophages and CD45+CD11b- lymphocytes are identified. Within CD45- gate, cells are discriminated according to positivity for ACSA2 (astrocytes) or O4 (oligodendrocytes). CD45-ACS2-O4- cells are further subdivided according to positivity for Thy1 (neurons) or CD31 (endothelium). Remaining Thy1-CD31- cells are classified as “other cell types”, not accounted by our antibody mix.
As shown in Figure 2B, with DH method about 38% of the viable cells retrieved from the brain and about 32% of the viable cells retrieved from the spinal cord were of hematopoietic origin (CD45+). On the other hand, PD method allowed to retrieve a significantly high yield of viable cells in both tissues, with a very large fraction represented by non-hematopoietic CD45- cells (about 82% in the brain and 92% in the spinal cord). Remarkably, CD45+CD11b+ microglia/macrophages represented the most abundant viable cell fraction with the DH method (Figure 2C). However, PD method produced a more heterogeneous representation of cell types, including ACSA+ astrocytes, O4+ oligodendrocytes, CD31+ endothelial cells and Thy1+ neurons (Figure 2C). Interestingly, viable neurons and endothelial cells were hardly detectable with the DH method.
The DH method relies on mechanical grinding of the tissue between the glass pestle and mortar of the Dounce homogenizer to obtain tissue homogenization. This could cause some shear stress that will likely damage and affect viability of large or very sensitive cells such as neurons or cells of the neurovasculature. We evaluated the cellular viability (percentage of 7-AAD- cells) within each cell subpopulation identified through the antibody panel (Figure 3). Hematopoietic cells (CD45+) isolated from brain and spinal cord, including microglia/macrophages (CD45+CD11b+) and other non-myeloid cells (CD45+Cd11b-), displayed very high viability independently from the homogenization method that was used (Figure 3A). On the contrary, the DH method determined a significant reduction of viability of CD45- populations (Figure 3B) whereas the PD method determined an extensive preservation of different CNS cell types. In detail, neurons and endothelial cells were the subpopulations most significantly affected by DH and preserved by the PD method.
A schematic presentation of the critical steps required for proper sample preparation and efficient debris removal are summarized in Figure 4.
Figure 1: Yield of cells retrieved from brain and spinal cord is affected by the homogenization method.
(A) Experimental outline. Mice were anesthetized and intracardiacally perfused with PBS to remove intra-vascular circulating blood cells. The brain and spinal cord were carefully dissected and split in two halves longitudinally. Tissues were homogenized using either Dounce homogenizer (DH) or papain digestion (PD) as detailed in the main text. Myelin and tissue debris were then removed by centrifugation in a 30% density gradient medium solution resulting in a heterogeneous cell suspension containing different cell types that could be analyzed by flow cytometry. (B) Histograms showing the yield of cells retrieved from the brain or the spinal cord upon tissue homogenization with the DH or PD method. The mean ± SEM of at least 6 animals per condition is represented. (C) Representative brightfield microscope photomicrographs of Trypan blue positive (dead) and negative (live) cells retrieved from brain or spinal cord by the two methods. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 2: Relative proportions of different cell types retrieved from the CNS are affected by the tissue homogenization method.
(A) Representative flow cytometry plots showing the gating strategy to identify different cell subpopulation within cell preparations obtained from brain or spinal cord: cell population is gated on FSC and SSC physical parameters, followed by selection for 7-AAD- live cells; then cells are discriminated according to positivity for CD45 marker; microglia/macrophages are identified as CD11b+ cells within the CD45+ fraction whereas lymphocytes are CD11b-. Astrocytes, oligodendrocytes, endothelial and neuronal cells are identified as ACSA2+, O4+, CD31+ or Thy1+ cells within CD45-, respectively. (B) Histograms showing the percentage of CD45+ and CD45- cells within total live or dead cell populations, in brain or spinal cord upon homogenization with the DH or PD method. The statistical analysis of the results shown in the graphs is reported in Table 2. (C) Pie charts showing the percentage of different viable cell subtypes within total cell population, in brain or spinal cord upon homogenization with the DH or PD method. The percentage of total dead cells is also reported. N ≥ 6. CD45+CD11b+ = microglia/macrophages; CD45+CD11b- = lymphocytes/non-myeloid cells; CD45-ACSA2+ = astrocytes; CD45-O4+ = oligodendrocytes; CD45-Thy1+ = neurons; CD45-CD31+ = endothelial cells; Other = cells negative for all above-mentioned markers. The statistical analysis of the results shown in the graphs is reported in Table 2. Please click here to view a larger version of this figure.
Figure 3: Cellular viability of different CNS cell types is affected by the homogenization method applied.
(A) Histograms showing the percentage of 7-AAD- live cells within CD45+ hematopoietic populations including CD11b+ microglia/macrophages and CD11b- non-myeloid cells. (B) Histograms showing the percentage of 7-AAD- live cells within CD45- non-hematopoietic populations including astrocytes, oligodendrocytes, neurons, endothelial and other cell types. * = p < 0.05, ** = p < 0.01, Mann-Whitney between DH and PD. Please click here to view a larger version of this figure.
Figure 4: Schematic representation of the critical steps required for proper tissue processing.
A list of the most critical steps required for proper tissue processing and efficient removal of debris is shown. It is important to identify properly the debris disk (black arrow) and the cell pellet (blue arrow) formed after centrifugation of the samples on the 30% density gradient. The debris disk, together with the rest of the supernatant, must be carefully removed by aspiration without dislodging the cell pellet to avoid sample loss. Please click here to view a larger version of this figure.
Antibody mix | Initial concentration (µg/mL) | Final concentration (µg/mL) | Dilution factor |
anti CD45/BV510 | 200 | 2 | 100 |
anti CD11b/APC.780 | 200 | 2 | 100 |
anti CD31/BV421 | 200 | 2 | 100 |
anti ACSA2/APC | 150 | 0.75 | 200 |
anti O4/biotin | na | na | 40 |
anti CD90.2/PE.Cy7 | 200 | 2 | 100 |
Streptavidin mix | Initial concentration (µg/mL) | Final concentration (µg/mL) | Dilution factor |
Streptavidin/Alexa 680 | 1000 | 1 | 1000 |
Table 1: Recipe for preparation of mixes for flow cytometry staining. The table describes the optimal concentrations of antibodies and streptavidin used to allow flow cytometric analyses of multiple cell types. Please refer to Table of Materials for details on catalogue numbers of each reagent mentioned in the table.
Statistics for Figure 2B | ||||||||||
BRAIN (% cells) | ||||||||||
CD45+ | CD45- | |||||||||
LIVE | DEAD | LIVE | DEAD | |||||||
Method | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | ||
DH | 15.20 | ± 2.32 | 1.90 | ± 0.30 | 24.78 | ± 4.045 | 51.58 | ± 6.033 | ||
PD | 15.20 | ± 2.65 | 2.33 | ± 1.10 | 68.53 | ± 3.618 | 13.93 | ± 2.180 | ||
Mann-Whitney | ns | ns | *** | ** | ||||||
p-value | 0.9989 | 0.738 | 0.0006 | 0.0015 | ||||||
SPINAL CORD (% cells) | ||||||||||
CD45+ | CD45- | |||||||||
LIVE | DEAD | LIVE | DEAD | |||||||
Method | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | ||
DH | 15.00 | ± 8.21 | 1.41 | ± 0.11 | 31.64 | ± 8.21 | 51.95 | ± 16.52 | ||
PD | 7.49 | ± 4.99 | 1.15 | ± 0.68 | 84.27 | ± 9.39 | 7.09 | ± 3.75 | ||
Mann-Whitney | ns | ns | * | ns | ||||||
p-value | 0.5548 | 0.7236 | 0.0438 | 0.1144 |
Statistics for Figure 2C | |||||||||||||||||
BRAIN (% cells) | |||||||||||||||||
CD45+ | CD45- | ||||||||||||||||
CD11b+ | CD11b- | ACSA2 | O4 | Thy1 | CD31 | Other | Dead | ||||||||||
Method | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | |
DH | 19.32 | ± 3.88 | 1.17 | ± 0.27 | 9.52 | ± 2.68 | 3.41 | ± 1.01 | 1.39 | ± 0.77 | 0.48 | ± 0.29 | 10.52 | ± 4.49 | 53.83 | ± 5.79 | |
PD | 10.88 | ± 2.03 | 1.65 | ± 0.48 | 8.17 | ± 2.66 | 6.54 | ± 0.76 | 6.37 | ± 1.76 | 8.27 | ± 1.25 | 33.28 | ± 6.34 | 23.72 | ± 5.31 | |
Mann-Whitney | ns | ns | ns | * | ** | *** | ** | ** | |||||||||
p-value | 0.1206 | 0.4819 | 0.5894 | 0.0264 | 0.0093 | 0.0003 | 0.0084 | 0.0022 | |||||||||
SPINAL CORD (% cells) | |||||||||||||||||
CD45+ | CD45- | ||||||||||||||||
CD11b+ | CD11b- | ACSA2 | O4 | Thy1 | CD31 | Other | Dead | ||||||||||
Method | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | mean | ± SEM | |
DH | 21.23 | ± 6.25 | 2.51 | ± 0.57 | 4.26 | ± 2.34 | 9.40 | ± 1.89 | 2.82 | ± 1.51 | 0.97 | ± 0.50 | 22.74 | ± 9.04 | 35.28 | ± 1.89 | |
PD | 9.63 | ± 1.67 | 2.77 | ± 0.48 | 4.23 | ± 1.59 | 28.62 | ± 3.57 | 1.26 | ± 0.49 | 6.94 | ± 2.14 | 26.39 | ± 8.17 | 19.09 | ± 4.76 | |
Mann-Whitney | ns | ns | ns | * | ns | * | ns | ns | |||||||||
p-value | 0.1905 | >0.9999 | 0.7302 | 0.0159 | 0.7302 | 0.0317 | 0.7302 | 0.1111 |
Table 2: Statistical analysis of different populations retrieved by applying the DH or PD method. The table describes the statistics for the graphs shown in Figure 2B and Figure 2C. The average and SEM of at least six independent samples is represented. The p value and details on statistical test applied for each comparison are also reported.
Herein we describe a protocol for the co-purification and concurrent flow cytometric analysis of some of the most relevant CNS cells from mouse brain and spinal cord. Traditionally, histological analyses have been applied to describe the distribution of nanoparticles or the transduction efficiency of viral vectors in the CNS5,13, or to provide insights on morphological and molecular changes occurring in specific cell types during a pathology or upon pharmacological treatment14. However, histology lacks processivity and it does not allow comprehensive examination of multiple features in the same histological samples, due to the limited number of markers that can be concurrently analyzed. Our approach can be complementary to traditional histologic analyses and it can be coupled with several downstream applications (sorting, primary culture, biochemical or next-generation-sequencing analyses) to expand the compilation of information that can be obtained from individual samples. However, some key factors listed below must be considered as they can critically impact the success of this approach:
In summary, the protocol here presented takes advantage of a gentle enzymatic digestion followed by a 9-color staining allowing efficient simultaneous flow cytometric evaluation of different cell types from mouse brain and spinal cord. The protocol could be exploited to monitor in a streamline and comprehensive manner the efficiency of cell targeting by nanoparticles or viral vectors administered in the CNS15. Moreover, the protocol could be easily adopted for very delicate downstream applications, such as cell sorting, ex vivo subculture, single cell RNAseq, resulting of utmost importance not only for preclinical assessment of cell-targeting by therapeutics but also for in-depth characterization of pathological processes in neurodegenerative diseases.
A fraction of the whole CNS cell population is not discriminated by this protocol (see “other” cell types in Figure 2); this can be explained by the presence of other cell subtypes that are present in the CNS but are not captured by the antibodies we used. In our preliminary analyses, about 14% of the “other cells” fraction is positive for CD73, a mesenchymal cell marker enriched in the neurovasculature and involved in several neuroinflammatory processes16,17. Moreover, we hypothesize that the “other cells” fraction could also comprise less differentiated cells, like progenitors at different stages of maturation, such as nestin+ neural stem cells, nestin+ vimentin+ radial glia progenitors, doublecortin+ neural progenitors, NG2+ oligodendrocyte precursor cells, among others. These cell sub-types could be easily investigated by applying our flow cytometry protocol, since we chose a configuration of fluorescent dyes that allows to accommodate up to two additional cell markers conjugated with either the fluorescein isothiocyanate (FITC) or phycoerythrin (PE) fluorophores.
Overall, our approach could provide a new tool for more comprehensive investigations in the context of the CNS (in health and disease) taking advantage of a well-consolidated technology allowing both qualitative and high-throughput quantitative assessments such as flow cytometry.
The authors have nothing to disclose.
This study was funded by Boston Children’s Hospital start-up funds to A.B., ALSA grant nr. 17-IIP-343 to M.P., and the Office of the Assistant Secretary of Defense for Health Affairs through the Amyotrophic Lateral Sclerosis Research Program under Award No. W81XWH-17-1-0036 to M.P. We acknowledge DFCI Flow Cytometry Core for technical support.
10X HBSS (Calcium, Magnesium chloride, and Magnesium Sulfate-free) | Gibco | 14185-052 | |
70 mm Cell Strainer | Corning | 431751 | |
ACSA/ACSA2 anti-mouse antibody | Miltenyi Biotec | 130-117-535 | APC conjugated |
Bovine Serum Albumin | Sigma Aldrich | A9647-1KG | |
CD11b rat anti-mouse antibody | Invitrogen | 47-0112-82 | APC-eFluor 780 conjugated |
CD31 rat anti-mouse antibody | BD Bioscience | 562939 | BV421 conjugated |
CD45 rat anti-mouse antibody | Biolegend | 103138 | Brilliant Violet 510 conjugated |
CD90.1/Thy1.1 rat anti-mouse antibody | Biolegend | 202518 | PE/Cy7 conjugated |
CD90.2/Thy1.2 rat anti-mouse antibody | Biolegend | 1005325 | PE/Cy7 conjugated |
Conical Tubes (15 mL) | CellTreat | 229411 | |
Conical Tubes (50 mL) | CellTreat | 229422 | |
Dounce Tissue Grinder set (Includes Mortar as well as Pestles A and B) | Sigma-Aldrich | D9063-1SET | |
Fc (CD16/CD32) Block rat anti-mouse antibody | BD Pharmingen | 553142 | |
Fetal Bovine Serum | Benchmark | 100-106 | |
Neural Tissue Dissociation Kit (P) | Miltenyi Biotec | 130-092-628 | |
O4 anti mouse/rat/human antibody | Miltenyi Biotec | 130-095-895 | Biotin conjugated |
Percoll | GE Healthcare | 10266569 | sold as not sterile reagent |
Percoll | Sigma | 65455529 | sterile reagent (to be used for applications requiring sterility) |
Percoll PLUS | Sigma | GE17-5445-01 | reagent containing very low traces of endotoxin |
Streptavidin | Invitrogen | S3258 | Alexa Fluor 680 conjugated |