Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Nov 10.
Published in final edited form as: Methods Mol Biol. 2021;2308:83–94. doi: 10.1007/978-1-0716-1425-9_7

Flow Cytometry-Based Analysis of the Mouse Bone Marrow Stromal and Perivascular Compartment

Yuki Matsushita 1, Wanida Ono 1, Noriaki Ono 1
PMCID: PMC8580286  NIHMSID: NIHMS1750255  PMID: 34057716

Abstract

Bone marrow stromal cells (BMSCs) account for an extremely small percentage of total bone marrow cells; therefore, it is technically challenging to harvest a good quantity of BMSCs with good viability using fluorescence-activated cell sorting (FACS). Here, we describe the methods to effectively isolate BMSCs for flow cytometry analyses and subsequent FACS. Use of transgenic reporter lines facilitates FACS-based isolation of BMSCs, aiding to uncover fundamental characteristics of these diverse cell populations.

Keywords: Bone marrow stromal cells (BMSCs), Skeletal stem cells (SSCs), Mesenchymal stem cells (MSCs), Fluorescence-activated cell sorting (FACS), In vivo lineage-tracing experiments, Single cell RNA-seq, C-X-C motif chemokine ligand 12 (CXCL12), Transgenic reporter mouse lines

1. Introduction

Bone marrow contains a diverse array of cells, including hematopoietic cells, vascular endothelial cells, and skeletal (mesenchymal) cells [1, 2]. Among them, skeletal cells account for an extremely small percentage of total bone marrow cells [3, 4]. Skeletal cells encompass bone-making osteoblasts, bone marrow stromal cells (BMSCs) and other skeletal progenitor and precursor cells. BMSCs are mainly located in contact with blood vessels [5], a small subset of which assumingly function as skeletal stem cells [6, 7]. A large majority of BMSCs express two important genes, C-X-C motif chemokine ligand 12 (CXCL12, also known as stromal cell-derived factor 1, SDF1) [8–10] and leptin receptor (LepR), a receptor for a fat-specific hormone leptin [11]. BMSCs can be marked by their reporter strains, including Cxcl12-GFP [12], Cxcl12-DsRed [13], Cxcl12-creER; R26RtdTomato [10], and Lepr-cre; R26RtdTomato [11] mice; as a result, these cells are also termed as CXCL12+ or CXCL12+LepR+ cells [14]. BMSCs can be isolated from these transgenic fluorescent reporter strains by flow cytometry and fluorescence-activated cell sorting (FACS) in a highly reproducible manner, without involving technique-sensitive steps of antibody staining. With this approach, BMSCs can be clearly identified by strong signals emanating from products of fluorescent reporter genes, the existence of which can be further confirmed by subsequent transcriptomic analyses. Moreover, if a cell type-specific tamoxifen-inducible creER line is utilized, for example in Cxcl12-creER; R26RtdTomato mice, cell fates of a specific subset of BMSCs can be interrogated through in vivo lineage-tracing experiments.

Here as an example, we describe the methods to isolate BMSCs effectively from the femurs of Cxcl12GFP/+; Cxcl12-creER; R26RtdTomato mice, which were treated with tamoxifen at postnatal day (P) 21 and dissected at P28 (Fig. 1). We further describe the methods to harvest BMSCs using FACS and subsequently apply these cells to a droplet-based single-cell RNA-sequencing analysis. We also describe the method to culture lineage-marked BMSCs ex vivo [10, 15] to circumvent the common problem associated with FACS-isolated BMSCs that do not survive well in cultured conditions [16].

Fig. 1.

Fig. 1

Open in a new tab

Histological images of Cxcl12GFP/+; Cxcl12-creER; R26RtdTomato femurs. Cxcl12-GFP+ and/or Cxcl12-creER;tdTomato+ cells, which are located in contact with Endomucin (Emcn)+ blood vessels

2. Materials

2.1. Cell Preparation

  1. Transgenic reporter mice (here, Cxcl12GFP/+; Cxcl12-creER; R26RtdTomato mice, as an example).

  2. Tamoxifen.

  3. Sunflower seed oil.

  4. Absolute ethanol.

  5. Isoflurane.

  6. Ice.

  7. Scissors (sharp and dull).

  8. Forceps (1 × 2 teeth).

  9. Razor (single edge blade).

  10. Paper wipe.

  11. 6-well plate

  12. Liberase™ TM (collagenase, Sigma/Roche).

  13. Molecular biology–grade water.

  14. Pronase (neutral protease, Sigma/Roche).

  15. Ca2+, Mg2+-free Hank’s balanced salt solution (HBSS).

  16. HBSS medium: HBSS supplemented with 1 mM ethylenedia-minetetraacetic acid (EDTA) solution and 2% fetal bovine serum.

  17. Mortar and pestle.

  18. Luer-lock syringe 1 ml.

  19. 18G Needle.

  20. 70 μm cell strainer.

  21. 50 ml conical tube.

  22. DPBS–1% BSA: Dulbecco’s phosphate-buffered saline (DPBS) supplemented with 1% bovine serum albumin (BSA).

  23. DMEM medium: Low-glucose DMEM with GlutaMAX™ supplement (DMEM).

    supplemented with 10% mesenchymal stem cell–qualified FBS and 1% penicillin–streptomycin.

  24. Flow cytometry staining buffer (eBioscience).

  25. Allophycocyanin (APC)-conjugated CD45 antibody (30F-11).

  26. 5 ml polystyrene round-bottom tube with cell-strainer cap.

  27. Shaking incubator.

  28. Centrifuge.

2.2. Cell Sorting and Single-Cell RNA-Seq Analysis

  1. BD FACS Aria III (Ex.407/488/561/640 nm).

  2. RNase zap.

  3. Dulbecco’s phosphate-buffered saline (DPBS).

  4. Bovine serum albumin (BSA).

  5. Protein LoBind 1.5 ml tube.

  6. Trypan blue solution.

  7. Centrifuge.

  8. Automated Cell Counter.

  9. Chromium Single Cell 3′ v3 microfluidics chip (10× Genomics Inc).

  10. Sequencer (for example NovaSeq 6000 from Illumina).

  11. R (ver 3.5) and Rstudio (ver 1.1).

  12. Seurat R package [17].

2.3. Colony-Forming Assay, Subsequent Subcloning and In Vitro Trilineage Differentiation

  1. Counting chambers (Neubauer-improved).

  2. 6-well plate.

  3. DMEM medium: DMEM supplemented with 10% mesenchymal stem cell-qualified FBS and 1% penicillin–streptomycin (mix #3, #4, and #5).

  4. 70% ethanol.

  5. 50% ethanol.

  6. Methylene blue.

  7. Cloning cylinders.

  8. Dow high vacuum grease.

  9. 0.5% trypsin–EDTA.

  10. Non-tissue culture-coated V-bottom 96-well plate.

  11. StemPro™ Chondrogenesis differentiation kit (Gibco).

  12. Alcian Blue Staining Solution.

  13. 48-well plate.

  14. αMEM medium: αMEM with GlutaMAX™ supplemented with 10% FBS and 1% penicillin–streptomycin.

  15. Osteogenic medium: αMEM medium with 1 μM dexamethasone, 10 mM β-glycerophosphate, and 50 μg/ml ascorbic acid.

  16. 4% paraformaldehyde solution.

  17. Alizarin Red S.

  18. IBMX (3-isobutyl-1-methylxanthine).

  19. Adipogenic medium: αMEM medium with 1 μM dexamethasone, 0.5 mM and 1 μg/ml insulin.

  20. LipidTOX Green (Invitrogen).

  21. Oil Red O.

  22. Vortex.

  23. Centrifuge.

  24. Automated inverted fluorescence microscope with a structured illumination system (for example the Zeiss Axio Observer Z1 with ApoTome.2 system).

  25. Zeiss Zen 2 software (blue edition).

3. Methods

3.1. Cell Preparation

  1. Label CXCL12+ BMSCs with green and red fluorescent proteins, by administering tamoxifen to Cxcl12GFP/+; Cxcl12-creER; R26RtdTomato mice: typically, a single dose of 1 mg tamoxifen is injected intraperitoneally (100 μl of 10 mg/ml tamoxifen-oil mix) for mice at 3–6 weeks of age (see Note 1).

  2. Sacrifice the mice after a short-chase (typically 2–7 days after tamoxifen injection) using carbon dioxide overdose or isoflurane overdose in a drop jar followed by induction of bilateral pneumothorax.

  3. Make an incision in the skin of the leg by scissors and peel it off. Dissect the femur by excising the ligament at the femur head and at the knee joint using sharp scissors. Remove soft tissues carefully from dissected femurs using scissors and paper wipes.

  4. Remove distal epiphyseal growth plates using dull scissors and cutting off proximal ends (see Note 2).

  5. Open up the marrow cavity and expose bone marrow. To achieve this, cut the cortex of the femurs longitudinally, roughly by a razor for 1–2 times, into several pieces. Do not cut the femur into small pieces.

  6. Transfer the roughly cut bone pieces into a 6-well plate containing 2 ml HBSS.

  7. Digest the bone pieces with enzymes, by adding 2 Wunsch units of Liberase™ TM with or without 1 mg of Pronase to 2 ml HBSS in the 6-well plate (see Notes 3 and 4).

  8. Incubate the solution at 37 °C, 300 rpm for 60 min on a shaking incubator (see Note 5).

  9. Aspirate and mix the supernatant containing dissociated cells using an 18-gauge needle with a 1 ml Luer-Lok syringe, and filter it through a 70 μm cell strainer into a 50 ml tube on ice.

  10. Add 2 ml fresh HBSS to the 6-well plate containing digested bone pieces. Aspirate and mix the supernatant rigorously using the needle with the syringe used above, and filter it through the 70 μm cell strainer into the 50 ml tube on ice.

  11. Transfer the digested bone pieces to a mortar containing 2 ml fresh HBSS medium, then mechanically triturate the bone pieces using the 18-gauge needle with the 1 ml Luer-Lok syringe used above, as well as using a pestle, to release any remaining cells on bone pieces. Filter dissociated cells through the 70 μm cell strainer into the 50 ml tube on ice to prepare single cell suspension. Repeat this step for five times and collect the supernatant into the same tube.

  12. Centrifuge, 280 × g, at 4 °C for 10 min, then discard supernatant by decanting (see Note 6).

  13. Resuspend the pellet in 2 ml ice-cold DPBS–1% BSA (for single-cell RNA-seq) or 10 ml DMEM medium (for cell culture, skip to Subheading 3.3 below).

  14. Filter cell suspension into a 5 ml polystyrene round-bottom tube with a cell-strainer cap, then keep it on ice until loading onto a cell sorter.

3.2. Cell Sorting and Single-Cell RNA-Seq Analysis of FACS-Isolated Cells

  1. Set up a four-laser BD FACS Aria III (Ex.407/488/561/640 nm) high-speed cell sorter with a 100 μm nozzle (see Note 7.

  2. Rinse the nozzle with RNase zap then backflush.

  3. Run no-fluorescence control, and GFP+ or tdTomato+ singlecolor controls to set up the gate and the compensation to sort the desired target cells. Single-color samples are separately collected from control mice carrying only single colors (see Note 8) (Fig. 2).

  4. Load the sample onto a cell sorter and collect the target cells (sorting speed should be less than 10,000 events/s).

  5. Sort GFP+ cells and/or tdTomato+ cells directly into 600 μl ice-cold DPBS–1% BSA in a Protein LoBind 1.5 ml tube (see Note 9.

  6. Pellet cells by centrifugation (300 × g, at 4 °C for 10 min) and resuspend the pellet in 10–30 μl DPBS–1% BSA to achieve a concentration of 1000 cells/μl (see Note 10).

  7. Mix 5 μl sample (or cell count control) and 5 μl trypan blue solution for cell counting.

  8. Count total and live cell numbers by automated Cell Counter.

  9. Load cells onto the Chromium Single Cell 3′ v3 microfluidics according to the manufacturer’s protocol.

  10. Sequence cDNA libraries by Illumina NovaSeq 6000.

  11. Preprocess the sequencing data using the 10× Genomics software Cell Ranger.

  12. Generate and use a custom genome fasta and index file by including the sequences of eGFP and tdTomato-WPRE to the mouse genome (mm10) for alignment purposes.

  13. Perform further downstream analyses using the Seurat R package.

  14. Filter out cells with less than 1000 genes per cell and with more than 15% mitochondrial read content.

  15. Perform downstream analyses, including normalization, identification of highly variable genes across the single cells, scaling based on number of UMI, dimensionality reduction (PCA, CCA, and UMAP), unsupervised clustering, and the discovery of differentially expressed cell-type specific markers (Fig. 3).

  16. Perform differential gene expression analyses to identify cell-type specific genes.

Fig. 2.

Fig. 2

Open in a new tab

Gating strategy for cell sorting of Cxcl12-GFP+Cxcl12-creER+ cells isolated from Cxcl12GFP/+; Cxcl12-creER; R26RtdTomato femurs. (Figures modified and adapted with permission from [10])

Fig. 3.

Fig. 3

Open in a new tab

Single cell RNA-seq analysis of Cxcl12-GFP+ (including Cxcl12-creER;tdTomato+) BMSCs harvested from Cxcl12 GFP/+; Cxcl12-creER; R26RtdTomato femur bone marrow at P28 (tamoxifen injected at P21). Left panel: UMAP-based visualization of major classes of FACS-sorted cells, center panel: feature plots of tdTomato expression, right center panels: violin plots, tdTomato expression in each cluster. (Figures modified and adapted with permission from [10])

3.3. Colony-Forming Assay, Subcloning, and In Vitro Trilineage Differentiation

  1. Voltex DMEM-suspended cells.

  2. Count cells manually using a hemocytometer.

  3. Plate cells into a 6-well plate at a density of 105 cells/cm2 with DMEM medium for 10–14 days (see Note 11).

  4. Locate tdTomato+ colonies in a 6-well plate: capture tdTomato epifluorescence from live cultured cells by tile-scanning the entire 6-well plates with an inverted fluorescence microscope equipped with a fully automated stage (Fig. 4). For subcloning, skip to step 7.

  5. Aspirate the supernatant, then fix cells by adding 70% ethanol, for 5 min.

  6. Discard the ethanol, then add 2% methylene blue–50% ethanol (methylene blue staining) (Fig. 4).

  7. To subclone individual tdTomato+ colonies, isolate each tdTomato+ colony using cloning cylinders (see Note 12).

  8. Add 50 μl of 0.25% trypsin-EDTA into the cylinder, then incubate it at 37 °C for 5 min.

  9. Add 50 μl of a basal medium, and transfer the cell suspension to a new 6-well plate.

  10. Culture a single cell-derived clone of tdTomato+ cells in a 6-well plate with a basal medium described above at 37 °C with 5% CO2 with exchanges into fresh media every 3–4 days.

  11. Passage each clone when the culture reaches subconfluency to confluency, transferring approximately 1/10 of an old 6-well plate to a new 6-well plate with a fresh basal medium.

  12. Continue to passage clones to analyze self-renewability, or use passage 4–7 clones for trilineage differentiation assays (see steps 13, 19, or 22).

  13. To induce chondrocyte differentiation, transfer cells (approximately 1/6 of a confluent 6-well plate) into a non-tissue culture-coated V-bottom 96-well plate.

  14. Centrifuge at 150 × g for 5 min at room temperature to pellet the cells, then aspirate the supernatant carefully.

  15. Add StemPro Chondrogenesis medium, and centrifuge the plate at 150 × g for 5 min at room temperature to pellet the cells.

  16. Culture the pellet in the differentiation medium with exchanges into fresh media every 3–4 days for up to 3 weeks, each time with centrifugation at 150 × g for 5 min at room temperature to repellet the cells.

  17. Fix cell pellets with 70% ethanol for 5 min.

  18. Stain for Alcian-Blue Staining Solution for 30 min.

  19. To induce osteoblast differentiation, plate cells (approximately 1/6 of a confluent 6-well plate) on a 48-well plate and cultured with αMEM medium until confluency.

  20. Change to the osteogenic medium with exchanges into fresh media every 3–4 days for up to 3 weeks.

  21. Fix cells with 4% paraformaldehyde for 5 min and stain for 2% Alizarin Red S for 30 min.

  22. To induce adipocyte differentiation, plate cells (approximately 1/6 of a confluent 6-well plate) on a 48-well plate and cultured with αMEM medium until confluency.

  23. Change to the adipogenic medium with exchanges into fresh media every 3–4 days for up to 2 weeks.

  24. Stain cells with LipidTOX Green (1:200 in basal medium) for 3 h at 37 °C, or fix cells with 4% paraformaldehyde for 5 min and stain for Oil Red O for 30 min.

Fig. 4.

Fig. 4

Open in a new tab

tdTomato+ cells and total CFU-Fs

Acknowledgments

This research was supported by grants from National Institute of Health (R01DE026666 to N.O., R01DE029181 to W.O.), and the Japan Society for the Promotion of Science KAKENHI Grant JP19K19236 to Y.M. We thank for K. Mizuhashi for inventing innovative approaches described in this chapter.

Footnotes

1.

To make 10 mg/ml tamoxifen solution, dissolve 100 mg tamoxifen in 2.5 ml absolute ethanol in a 50 ml conical tube. Voltex persistently until crystals become completely dissolved. Add 10 ml sunflower seed oil and voltex persistently again. Open the cap and incubate at 65 ° C overnight in a chemical fume hood until ethanol completely evaporates. Close the cap and keep the tamoxifen-oil mix at room temperature until use. A dose of 25–175 mg/kg body weight is typically approved by an IACUC protocol. Average body weights of male C57BL/6J are 10.6 ± 1.9 g at 3-week, 16.5 ± 2.6 g at 4-week, 21.9 ± 1.8 g at 6-week and 25.0 ± 1.4 g at 8-week, according to The Jackson Laboratory. Tamoxifen should be administered to mice under a certified biosafety cabinet, and the mice should be housed in a special containment space.

2.

Epiphyses of femurs can be easily popped out using dull scissors.

3.

Reconstitute and aliquot Liberase solution according to the manufacturer’s protocol. Add 10 ml molecular biology-grade water to the amber bottle of Liberase™ TM (50 mg), and swirl the bottle well. Keep it on ice for 30 min, while inverting the bottle for every 5 min (do not use vortex). Aliquot 500 μl enzyme solution in a 1.5 ml tube, and keep it at −30 °C until use.

4.

Adding Pronase in the digestion cocktail increases the yield of BMSCs, especially those located in the central marrow space, without compromising RNA integrity number. However, Pronase digestion degrades protease-sensitive antigens such as CD31 (encoded by Pecam1).

5.

Length of digestion needs to be optimized for the age of mice. A 60 min digestion is sufficient for mice at 3–8 weeks of age.

6.

For standard flow cytometry analysis, cells should be resuspended in a 500 μl flow cytometry staining buffer containing appropriately diluted antibodies (such as 1:500 CD45-APC), then incubate the cells at 4 °C for 30 min. Add 2 ml flow cytometry staining buffer to wash. Centrifuge, 280 × g, at 4 °C for 10 min, then discard supernatant by decanting. Resuspend the cells in flow cytometry staining buffer (500–1000 μl, then run on flow cytometer.

7.

RNA quality of sorted cells depends on the size of the nozzle. Wider nozzles, such as 100 μm or 130 μm nozzles, give much better results than narrower nozzles, such as 70 or 85 μm nozzles. The sheath pressure is 20 psi using a 100 μm nozzle, which is close to the atmospheric pressure (15 psi). A 130 μm nozzle (10 psi) is ideal to prevent damage to the cells; however, the sorting speed becomes substantially slow. It may be difficult to harvest a sufficient number of cells within a limited amount of sorting time. A 70 μm nozzle (70 psi) or an 85 μm nozzle (45 psi) may inflict damage to the cells and compromise the RNA quality.

8.

Exclude FSChigh/SSChigh events as they contain duplets or multiplets of cells. We often experience that Cxcl12-GFP+ BMSCs are physically coupled with hematopoietic cells in a way indistinguishable from purely single cells.

9.

Sort more than 50,000 cells ideally. If less than 50,000 cells are harvested, a separate control tube should be set up solely for the purpose of cell counting, to estimate the cell number in an experimental tube. For the cell counting control tube, sort at least 10,000 GFP-low or tdTomato-low cells and dilute it at the same ratio as the experimental tube.

10.

Generally, 30% of the cells counted on FACS are lost due to centrifugation, between the sorting step and the actual cell counting step. As the ideal cell concentration for a 10× microfluidics device is 1000 cells/μl, sorted cells should be diluted to approximately 1400 cells/μl.

11.

Cell cultures are maintained at 37 °C in a 5% CO2 incubator.

12.

Print tile-scanned tdTomato+ colony images as the actual size of a 6-well plate. For this purpose, make 1.42 × 1.42-in. circles in 10 × 7.5-in. PowerPoint slide, and print it to a letter size paper. Then insert the printed images under the 6-well plate. After aspirating the medium, attach a cloning cylinder using grease to the position of the plate corresponding to the position of tdTomato+ colonies.

Declaration of interests: The authors declare no competing interests.

References

  • 1.Mendelson A, Frenette PS (2014) Hematopoietic stem cell niche maintenance during homeostasis and regeneration. Nat Med 20 (8):833–846 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Morrison SJ, Scadden DT (2014) The bone marrow niche for haematopoietic stem cells. Nature 505(7483):327–334 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Pittenger MF, Mackay AM, Beck SC et al. (1999) Multilineage potential of adult human mesenchymal stem cells. Science 284 (5411):143–147 [DOI] [PubMed] [Google Scholar]
  • 4.Zhu H, Guo ZK, Jiang XX et al. (2010) A protocol for isolation and culture of mesenchymal stem cells from mouse compact bone. Nat Protoc 5(3):550–560 [DOI] [PubMed] [Google Scholar]
  • 5.Sacchetti B, Funari A, Michienzi S et al. (2007) Self-renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic microenvironment. Cell 131(2):324–336 [DOI] [PubMed] [Google Scholar]
  • 6.Bianco P, Cao X, Frenette PS et al. (2013) The meaning, the sense and the significance: translating the science of mesenchymal stem cells into medicine. Nat Med 19(1):35–42 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Matsushita Y, Ono W, Ono N (2020) Skeletal stem cells for bone development and repair: diversity matters. Curr Osteoporos Rep 18 (3):189–198 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Greenbaum A, Hsu YM, Day RB et al. (2013) CXCL12 in early mesenchymal progenitors is required for haematopoietic stem-cell maintenance. Nature 495(7440):227–230 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Asada N, Kunisaki Y, Pierce H et al. (2017) Differential cytokine contributions of perivascular haematopoietic stem cell niches. Nat Cell Biol 19(3):214–223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Matsushita Y, Nagata M, Kozloff KM et al. (2020) A Wnt-mediated transformation of the bone marrow stromal cell identity orchestrates skeletal regeneration. Nat Commun 11(1):332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Zhou BO, Yue R, Murphy MM et al. (2014) Leptin-receptor-expressing mesenchymal stromal cells represent the main source of bone formed by adult bone marrow. Cell Stem Cell 15(2):154–168 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Ara T, Tokoyoda K, Sugiyama T et al. (2003) Long-term hematopoietic stem cells require stromal cell-derived factor-1 for colonizing bone marrow during ontogeny. Immunity 19 (2):257–267 [DOI] [PubMed] [Google Scholar]
  • 13.Ding L, Morrison SJ (2013) Haematopoietic stem cells and early lymphoid progenitors occupy distinct bone marrow niches. Nature 495(7440):231–235 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Seike M, Omatsu Y, Watanabe H et al. (2018) Stem cell niche-specific Ebf3 maintains the bone marrow cavity. Genes Dev 32 (5–6):359–372 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Mizuhashi K, Ono W, Matsushita Y et al. (2018) Resting zone of the growth plate houses a unique class of skeletal stem cells. Nature 563(7730):254–258 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Omatsu Y, Sugiyama T, Kohara H et al. (2010) The essential functions of adipo-osteogenic progenitors as the hematopoietic stem and progenitor cell niche. Immunity 33(3):387–399 [DOI] [PubMed] [Google Scholar]
  • 17.Butler A, Hoffman P, Smibert P et al. (2018) Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36(5):411–420 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES