Isolation and characterization of CD73+CD39+CD146+ mesenchymal stem cell subset from bone marrow

Abstract

BACKGROUND

Our mission is to cure hematopoietic malignancies through cell therapy. Time to transplant is a key challenge resulting in mortality of patients needing a transplant. Previous studies reported CD146+ mesenchymal stem cells (MSCs) regulating hematopoiesis in bone marrow (BM). In 2013, the study reported the existence in the synovium of a MSC subset, co-expressing CD73 and CD39, with greater osteo-chondrogenic potency and ability to produce adenosine. This subset expressed CD146, known to be associated with pericytes.

AIM

To investigate the presence and characterization of the CD73+CD39+CD146+ MSC subset in BM. Furthermore, we explored the existence of this subset in mobilized blood.

METHODS

BM cells were culture expanded up to passage 4. Flow cytometry was used to verify expression of CD73, CD39, and CD146 markers. Cell sorting was performed via BDFACS AriaTM Fusion. The subset was assessed for defined MSC characteristics and perivascular localization in BM sections. Peripheral blood derived MSCs were obtained through apheresis performed at Gift of Life under Institutional Review Board donor consent.

RESULTS

Our findings demonstrated that the combination of CD73, CD39, and CD146 enabled the identification and purification of a subset of MSCs from culture-expanded BM, up to passage 4. This subset exhibited a CD45-CD73+CD39+CD146+ phenotype, along with self-renewal and multipotency abilities, and was located in perivascular areas of BM sections. Additionally, this subset was found in both single and dual-mobilized leukopaks.

CONCLUSION

The CD73+CD39+CD146+ cell subset showed self-renewal and multipotency abilities and was located in perivascular areas of BM. Such cell subset was also reported in single and dual-mobilized leukopaks.

Key Words: Hematopoietic stem cells; Bone marrow-derived mesenchymal stem cells; Mesenchymal stem cells; Mesenchymal stem cell-like pericytes; Hematopoietic stem cell transplants; Peripheral blood stem cells; Single and dual mobilized leukopaks

Core Tip: In this study, the co-expression of CD73, CD39, and CD146 enabled the identification and enrichment of a mesenchymal stem cell subset with pericyte-like properties from bone marrow cultures. This subset was also found in mobilized blood of adult donors, which serve as a source for allogenic stem cell transplants. Since pericytes are known to play a vital role in maintaining hematopoiesis in vivo, we anticipate that using this subset in a co-culture system with hematopoietic stem cells could potentially lead, with further investigation, to methods for expanding hematopoietic stem cells and creating an inventory of products that are readily available for transplants.

INTRODUCTION

One person in the United States is diagnosed with blood cancer every 3 minutes[1]. Hematopoietic stem cell (HSC) transplant is currently the only life-saving treatment for patients suffering from hematological disorders. Current hematological stem cell transplants occur at 1:1 donor to recipient ratio. Although allogeneic HSC transplant is widely used for treatment, a key limitation is the time to transplant. Delay in donor search, selection, qualification, and procurement process results in patient mortality[2]. The Gift of Life Marrow Registry is currently investigating strategies to expand HSCs derived from ‘super donors’ with frequent human leukocyte antigen phenotype (2000 in our registry). The long-term goal is to generate multiple transplantable doses from a single donation and build an off-the-shelf inventory immediately available for patients in need. It is well known that HSCs differentiate in vitro losing their stemness, thus it is imperative to develop optimal culture conditions for their in vitro expansion[3-5]. In 2010, Walenda et al[5] reported the impact of mesenchymal stem cells (MSCs) on the properties of HSCs. The study demonstrated that co-culture of MSCs and HSCs resulted in the expansion of primitive HSCs and maintenance of their stemness, pointing to the significance of a co-culture system that more closely mimics the bone marrow (BM) niche[5-7]. MSCs are known to be a population of multipotent stromal cells, able to differentiate into osteogenic, chondrogenic, and adipogenic lineages[8]. The established set of non-specific surface markers CD73, CD90 and CD105 have been used so far to characterize MSCs[9], yet heterogeneity of MSCs exists across tissue sources, affecting the potential outcome of clinical applications[9,10]. In 2013, Gullo and De Bari[11] identified a subpopulation of MSCs from the human synovium co-positive for CD73, ecto-5’ nucleotidase, and CD39, ecto-NTPDase, which exhibited a greater chondro-osteogenic potency. Unpublished data showed that this population also expressed CD146, consistently up to passage 12 (P12), and produced adenosine. CD146 is known to be associated with pericytes, mural cells that have been suggested to be progenitors of MSCs[12]. Sacchetti et al[7] reported a population of MSC-like pericytes from the BM which maintained the hematopoietic environment in a xeno-transplantation model. MSC-like pericytes are associated with localization of HSCs in the BM niche by releasing certain factors. Among these factors is CXC motif chemokine ligand 12 (CXCL12), which is implicated in maintaining survival and controlling the proliferative/differentiating properties of HSCs, thus maintaining their stemness. Angiopoietin-1 plays a pivotal role in the BM niche as it is key to hematopoiesis and directly contributes to HSC regulation by interacting with HSC-expressed Tie-2[7]. MSC-like pericytes may represent a promising asset for successful transplantation and engraftment of HSCs for treatment of hematopoietic diseases and disorders[6,7]. Our previous findings in the human synovium together with well-established literature prompted us to investigate the existence of the CD73+CD39+CD146+ subset first in BM and then in mobilized blood (MB), a readily available source in our center. Using these markers, we were able to purify an MSC cell subset with perivascular location in the BM. Of note, this cell subset was also present in MB samples.

MATERIALS AND METHODS

BM derived-MSC cultures

BM derived-MSCs (BMSCs) were purchased at P2 (Millipore Sigma, Darmstadt, Germany, Cat.# SCC034), thawed per vendor specifications, and culture expanded with Dulbecco’s modified eagle medium (DMEM, Gibco, NY, United States, Cat.# 12491015), 10% fetal bovine serum (FBS, Gibco, NY, United States, A31604-02), and 1% penicillin/streptomycin/amphotericin (Hyclone, UT, United States, SV30079.01) in culture flasks (Corning, NY, United States, 431080) for 14 days at 37 °C and 50 mL/L CO₂ in incubator (Thermo Scientific, NY, United States, Forma 3110). On the day of confluence cells were trypsinized with 0.05% Trysin-EDTA (Gibco, NY, United States, 25300062), neutralized with equal volume of complete DMEM media, and set for sorting (see enrichment strategy of target cell subset). After sorting, the cells were plated in T150 flasks (Falcon-Corning, AZ, United States, 355001) with DMEM, 20% FBS and 1% antibiotic/antifungal for 2 days and switched to 10% FBS complete DMEM medium until confluent. Experiments were performed with expanded cells between P3 and P5.

Flow cytometry phenotypic analysis

BM cells were stained with 5 μL of Anti-Hu FC Receptor Binding (Invitrogen, CA, United States, 14-9161-71) for 15 minutes prior to antibody staining. Cells were verified as MSCs using MSC standard markers; the panel of markers was as follows: CD45, CD73, CD105, and CD90[13]. To verify the target subset, the following fluorophores were used: CD45-APC-AlexaFluor 750 (Beckman Coulter, IN, United States, A71119), CD73-APC (Miltenyi, Germany, 130-111-909) CD39-FITC (Miltenyi, Germany, 130-125-094), CD146-PE (Beckman Coulter, IN, United States, A07483), and 7AAD (Beckman Coulter, IN, United States, A07704). Acquisition was conducted at 10000-100000 events for BMSCs. Non-mobilized mononuclear cells (MNC) and single/dual mobilized peripheral blood stem cell (PBSC) acquisition were conducted at 500000 events to obtain a large number of CD45 negative events. All acquisitions were conducted on Beckman Coulter CytoFLEX. Analysis was performed using FLOWJO software. t-Distribution Stochastic Neighbor Embedding (tNSE) plots were used in combination with FlowSOM population generator and cluster explorer to generate a cluster which expressed the desired markers for MNCs and PBSCs[14]. Gates were established by using unstained and Fluorescence MinusOne controls. Details on the gating strategy can be found in Supplementary Figure 1. A similar gating strategy was employed for MNCs.

Enrichment strategy of target cell subset (CD73+CD39+CD45-)

Cells were counted with Via1 cassettes (Chemometec, NY, United States, 941-0012) on Neucleocounter NC-200 (Chemometec, NY, United States, 900-0200). BMSCs were harvested following culture (see BMSC culture) and stained for cell sorting using the following: DAPI (Ebiosciences, CA, United States, 62248), CD45-APC Alexafluor 750, CD73-APC, CD39-FITC. Unstained sample and single stain compensation beads (VersaComp Antibody Capture Bead Kit, Beckman Coulter, IN, United States, B22804) were used as controls. BM cells were sorted on BDFACS Aria^TM Fusion at UF Scripps Core[15]. Following cell sorting, isolated BMSCs were culture expanded and used for further assays.

Colony-forming-unit-fibroblast-assay

For colony-forming unit (CFU) assays cells were plated in a six-well plate (Corning, NY, United States, 0720083) at a density of 100 cells per well and cultured for 14 days. On day 14, cells were washed with phosphate buffered saline and fixed in 4% paraformaldehyde (PFA) for 10 minutes. Cultures were stained with 0.5% crystal violet (Electron Microscopy Sciences, PA, United States, 26301-01) for 1 hour and rinsed in distilled water (Gibco, NY, United States, 15230-147). Aggregates of ≥ 50 cells were defined as CFU-fibroblasts. Colony numbers were determined by counting under microscope (Inverted Microscope CKX53SF, Olympus, Japan).

Differentiation assays

In vitro chondrogenesis assay was performed as previously described[11] using a chondrogenic differentiation medium (Stem Cell Technologies, Canada, 05455). Micromasses stained with 1% Alcian blue at pH 2.5 (Electron Microscopy Sciences, PA, United States, 26323-01). To quantify, the Alcian blue micromasses were extracted with 6M guanidine HCl (ThermoFisher, NY, United States, 60-047-420) at room temperature for 6 hours. The optical density was measured at 600 nm.

Osteogenic differentiation assays were performed by culturing cells in osteogenic differentiation medium (Stem Cell Technologies, Canada, 05465) for 3 weeks as previously described[11]. After 3 weeks, cells were fixed with 4% PFA (Electron Microscopy Sciences, PA, United States, 157-4-5G) and stained with Alizarian Red (ICCA, Netherlands, R5004). After rinsing with water (Gibco, NY, United States, 15230-147), Alizarian red was solubilized with 0.5N HCl SDS (ThermoFisher, NY, United States, 035642.K2) and the absorbance was detected in a spectrophotometer at 405 nm.

Adipogenic differentiation assays were performed by culturing cells in adipogenic differentiation medium (Stem Cell Technologies, Canada, 05412) for 4 weeks. On day 28, cells were fixed with 4% PFA washed with water (Gibco, NY, United States, 15230-147) and stained with Oil Red O (StatLabs, CA, United States, STORO100) for 10 minutes. After staining, wells were washed thoroughly with 70% ethanol (Contec, SC, United States, 19-033-765). Post image colony counting was conducted by staining cells with 0.5% crystal violet for 1 hour and rinsed with water. Only colonies containing vacuoles were counted as adipose colonies.

Adenosine fluorometric assay

Cells were lifted trypsinized with 0.05% Trysin-EDTA (Gibco, NY, United States, 25300062), neutralized with equal volume of complete DMEM media, and counted with Via1 cassettes (Chemometec, NY, United States, 941-0012) on Neucleocounter NC-200 (Chemometec, NY, United States, 900-0200). Following counts, 2 × 10⁵ cells were lysed with RIPA Lysis and Extraction Buffer (Thermo Scientific, NY, United States, 89900), and placed on ice for 15 minutes, shaking occasionally. After lysing, cells were prepared for adenosine assay per manufactures protocol [Ab211094, Adenosine Assay Kit (Fluorometric), version 6c, Abcam, United Kingdom] in opaque 96-well, flat bottom plates (ThermoFisher, NY, United States, 08-771-226). The fluorescence was detected at 580-640 nm emission/520 nm excitation, and its intensity was proportional to the amount of adenosine in the samples.

Immunofluorescent staining

BM core slides were purchased from Creative Bioarray (NY, United States). Slides were baked at 60 °C for 16 hours then deparaffinized and rehydrated. To ensure optimal antibody retrieval, heat-induced epitope retrieval using 1 × citrate buffer and pressure cooker was conducted. BloxALL (Vector, Germany, SP-6000-100) and Opal Antibody Diluent/Block (Akoya, MA, United States, ARD1001EA) were used as blocking buffers to minimize non-specific binding. Primary antibody of anti-CD45 (Abcam, United Kingdom, ab243869), anti-CD39 (Abcam, United Kingdom, 223842), anti-CD146 (Invitrogen, CA, United States, MA5-29413), and anti-CD73 (CST, MA, Unites States, 13160), were used alongside Opal 620 (Akoya, MA, United States, OP-001004), Opal 480 (Akoya, MA, United States, OP-001001), Opal 690 (Akoya, MA, United States, OP-001006), and Opal 780 (Akoya, MA, United States, OP-001008) for multiplexing. Antibody stripping was conducted with 1 × citrate buffer overnight in water bath. Slides were washed with Tris-buffered saline with Tween 20 [BP152-1, Tris Base (Fisher Bioreagents), MA, United States, sodium chloride (VWR, PA, United States, 0241-5 kg), polysorbate 20 (VWR, PA, United States, 97062-332)] for all washing steps. Slides were counter-stained with Hoechst 33342 (Thermo Scientific, NY, United States, 62249). Immunofluorescence assays were validated, conducted, and analyzed by Applied Pathology Systems.

Apheresis process

Allogenic stem cell donors underwent full health history, screening, and medical evaluation to confirm as viable candidates for research grade product. All donors were provided with informed consent forms that had been approved by the Institutional Review Board for the research study. After donor consent, PBSC donors were given 10 μg/kg of granulocyte colony stimulating factor (G-CSF), filgrastim (Neupogen^TM) subcutaneous injections per day for five days to mobilize stem cells into the bloodstream. On day five, collection was conducted, and stem cells were sorted through cell-separating apheresis machine (Spectra Optia, Terumo, Japan). Dual-mobilized PBSC donors followed the filgrastim protocol for 5 days. On day 5, donors were given 24 mg/1.2 vial plerixafor (Mozobil^®) and an apheresis collection was conducted. MNC donor samples were collected under the same protocol without filgrastim or plerixafor mobilization. Samples were placed in 2-8 °C for no more than 24 hours before flow cytometry staining and analysis.

Statistical analysis

Experiments were conducted in duplicate or triplicate using appropriate controls. Data are presented as mean ± SD. SD and P values were determined by one-way ANOVA analysis Prism GraphPad software.

RESULTS

Co-expression of CD73 and CD39 in BMSCs

We first analyzed dynamics of expression from BMSCs following culture expansion (P2 and P3). As expected, all BMSCs were negative for the pan-hematopoietic marker CD45 (Figure 1A). Based on previous findings from synovial MSCs[11], we analyzed the co-expression of CD73 and CD39 within the CD45 negative population to assess presence of the subset of interest in BM (Figure 1A). The bar graph (Figure 1B) showed that CD45 negative cells analyzed at P2 co-expressed CD73 and CD39 (80%; n = 3). At P3, the expression of CD73 and CD39 was significantly higher at 96% (n = 3) (Figure 1B).

Figure 1 Co-expression of CD73 and CD39 in bone marrow-derived mesenchymal stem cells. A: Representative flow cytometry analysis showed, as expected, 99.9% of isolated bone marrow mesenchymal stem cells were negative for CD45 (a1). Representative flow cytometry analysis showed 96.0% of CD45- population of bone marrow-derived mesenchymal stem cells were co-positive for CD73+ CD39+ (a2); B: Bar graph showing expression of CD73 and CD39 markers at passage 2 and passage 3 (n = 3). Results are mean ± SD.

Isolation of CD45-CD73+CD39+ target subset from BM

Cell sorting experiments were performed at P3 to purify the subset of interest from BM (CD45-CD73+CD39+, Figure 2). The whole cell population was defined on a basis of forward scatter area (FSC-A) and side scatter area (Figure 2A), and doublet exclusion was conducted based on forward scatter height and width (FSC-H and FSC-W) (Figure 2A and B). Following the doublet exclusion, dead cells were ruled out by DAPI staining (Figure 2A and B) and only the viable cell fraction negative for CD45 was gated (Figure 2A and B). Within the CD45 negative population, cells co-expressing CD73 and CD39 were selected for sorting (Figure 2A and B). Post-sorting flow cytometry phenotypic analysis of sorted samples showed > 90% co-expression of CD73 and CD39 (Figure 2C).

Figure 2 Sorting strategy and purity post-sort for bone marrow-derived CD45-CD73+CD39+ subset. A: Sorting strategy for bone marrow (BM)-derived mesenchymal stem cells (MSCs) for first sorting (a1 and a2). Cell population was defined based on cell size, granularity, and single cells. Cells were gated on the viable population (a3). The cell fraction negative for CD45 was gated (a4) and cells which co-expressed CD73 and CD39 were gated and isolated based on single stained controls (a5). Controls not shown; B: Sorting strategy for BM-derived MSCs for second sort (b1 and b2). Cell population was defined based on cell size, granularity, and single cells. Cells were gated on the viable population (b3). The cell fraction negative for CD45 was gated (b4) and cells which co-expressed CD73 and CD39 were gated and isolated based on single stained controls (b5). Controls not shown; C: Bar graphs showing percentage of the CD73+CD39+ cell subset in samples obtained post-sorting.

CD45-CD73+CD39+ cell subset displays MSC characteristics

In 2006, the International Society for Cell and Gene Therapy established characteristics to define MSCs. Included in this is plastic adherence, the ability to self-renew and differentiate into various cell types including osteoblasts, adipocytes, and chondrocytes[13]. Thus, we investigated whether our subset displayed the MSC properties established by International Society for Cell and Gene Therapy. First, we conducted a clonal assay to verify their ability to self-renew. Data showed that the CD45-CD73+CD39+ BMSC subset exhibited plastic adherence and clonogenicity (Figure 3A). Following a 14-day culture, the results showed an average of 14 colonies per well (n = 2) (Figure 3A).

Figure 3 CD45-CD73+CD39+ cell subset displays mesenchymal stem cell characteristics. A: Clonogenicity of CD45-CD73+CD39+ isolated mesenchymal stem cells (MSCs). Cells were plated at 100 cells/well in a 6 well plate. After 14 days cells were stained with crystal violet. Aggregates of ≥ 50 cells were defined as colonies (a1). The bar graph represents the average number of colonies. Results are mean ± SD (n = 2). Representative image of one colony at day 14 following fixation and crystal violet staining at 40 × and scale bar = 24 μm (a1-a3); B: Each bar graph represents treatment group mean ± SD (n = 3), P < 0.0001, MSC subset compared to negative control. CD45-CD73+CD39+ subset was cultured at 10000 cells/well and treated with osteogenic differentiation medium. On day 21, cells were fixed and stained with alizarin red. Untreated cells and osteoblasts were used as controls (b1). Representative images of untreated subset, osteoblast, treated subset at day 21 following fixative and stain at 40 × and scale bar = 24 μm (b2-b4). Following imaging, cells were measured for absorbance at 405 nm; C: Each bar graph represents treatment group mean ± SD (n = 3), P < 0.001, MSC subset compared to negative control. CD45-CD73+CD39+ subset was cultured at 100 cells/well and treated with adipogenic differentiation medium. On day 28, cells were fixed and stained with Oil Red O. Aggregates of ≥ 50 cells and vacuoles were defined as adipogenic colonies. Untreated cells and adipocytes were used as controls (c1). Representative images of untreated subset, adipocytes, treated subset at day 28 following fixative and stain at 40 × and scale bar = 24 μm (c2-c4); D: Each bar graph represents treatment group mean ± SD (n = 3), P < 0.001, MSC subset compared to negative control. Cells were plated in micromasses in a 6 well plate at density of 250000 cells/ well. After 7 days, cells were fixed and stained with Alcian blue (d1). Representative images of untreated subset, chondrocyte, treated subset at 7 days following fixative and stain at 40 × and scale bar = 24 μm (d2-d4). Cells were measured for absorbance at 600 nm.

To assess whether the subset demonstrated multipotency, we conducted differentiation assays in the respective multipotency lineages (Figure 3B-D). The results showed that the CD45-CD73+CD39+ subset differentiated into bone (P < 0.0001, compared to the negative control) (Figure 3B), adipose (P < 0.001, compared to the negative control) (Figure 3C) and cartilage (P < 0.001, compared to the negative control) (Figure 3D). The subset exhibited a slightly greater osteogenic potential compared to the osteoblast positive control (Figure 3B).

CD45-CD73+CD39+ subset produces adenosine

CD73 and CD39 are enzymes known as ectonucleotidases that play a significant role in converting ATP into adenosine, a molecule with relevant immunomodulatory effects[11]. Previous unpublished findings highlighted that CD45-CD73+CD39+ synovial-derived MSCs exhibited the ability to produce adenosine. In this study, we conducted an adenosine fluorescent assay and found that the BM-derived CD73+CD39+CD146+ cell subset produced adenosine at a concentration of 21.6 mol/μL (n = 3), while the unsorted BM population yielded a concentration of 16.4 pmol/μL (n = 3), P < 0.0001 (Supplementary Figure 2).

In vitro and in vivo CD146 expression in CD45-CD73+CD39+ subset

Unpublished data from Gullo and De Bari[11] showed that CD73+ CD39+ MSCs from synovium also expressed CD146, known to be associated with pericytes[6,7,11,16]. Thus, we investigated whether our BM-derived CD45-CD73+CD39+ subset also expressed CD146. Flow cytometry phenotypic analysis showed that the entire CD45-CD73+CD39+ subset was also positive for CD146 at P3 and P4 (n = 3) (Figure 4A and B). To confirm that the CD45-CD73+CD39+CD146+ subset was present in vivo, antibody staining of healthy human BM core paraffin slides were stained for immunofluorescent analysis. Results showed that the CD45-CD73+CD39+CD146+ subset was present in vivo at the perivascular locations in BM (Figure 4C).

Figure 4 In vitro and in vivo CD146 expression in CD45-CD73+CD39+ subset. A: Representative flow cytometry analysis showed the entire population of cells copositive for CD73 and CD39 were also positive for CD146 post-sorting and culture (a1-a3); B: Frequency of parent of bone marrow mesenchymal stem cells post sorting for cultured cells for passage 3 and passage 4 (n = 3) with phenotype CD45-CD73+CD39+ and CD45-CD73+CD39+CD146+ as bar graphs are mean ± SD. No significant difference was observed between passage 3 and passage 4; C: Representative immunofluorescence analysis showing colocalization of CD73, CD39, and CD146 in human bone marrow slides. Human bone marrow slide with isotype control was used to verify staining. Image was taken at 40 × (c1). Human tonsils were previously validated for expressing CD45, CD73, CD39, and CD146 and served as a positive control. Image was taken at 40 × (c2). Representative image of human bone marrow slide showing colocalization of subset at perivascular locations where the arrows show locations of CD73, CD39, and CD146 expression and represents the vessel. Image was taken at 40 × (c3).

CD45-CD73+CD39+CD146+ subset expression in MB

Since the primary source of cells used for hematopoietic progenitor cell transplants at our center is apheresis peripheral blood, we investigated whether the CD45-CD73+CD39+CD146+ MSC subset was detectible in mobilized (PBSCs) and non-mobilized leukopaks (MNCs). Flow cytometry analysis showed that the CD45-CD73+CD39+CD146+ subset was not present in non-mobilized leukopaks (Figure 5A). Mobilized leukopaks were also investigated to assess whether mobilization of stem cells with filgrastim would affect the percentage of the target subset in the blood stream. The data showed that, in samples of PBSCs mobilized with filgrastim alone (single-mobilized PBSCs) (n = 6) (Figure 5B and Supplementary Figure 3) and PBSCs mobilized with filgrastim and plerixafor (dual-mobilized PBSCs) (n = 3) (Figure 5C), the CD45-CD73+CD39+CD146+ subset represented less than 1% of the entire population (Figure 5D and E). A flow cytometry strategy based on t-Distribution Stochastic Neighbor Embedding plot in combination with flowSOM and cluster explorer was used to identify the cell expression of CD73, CD39, and CD146 within the CD45 negative fraction. Details regarding strategy of gating are reported in Supplementary Figure 4[14].

Figure 5 CD45-CD73+CD39+CD146+ subset expressed in mobilized peripheral blood. A: Representative flow cytometry analysis showing target subset in mononuclear cells. Total CD45 negative fraction was gated (a1). t-Distribution Stochastic Neighbor Embedding (tSNE) plot in combination with flowSOM and cluster explorer was used. 1.8% of CD45- cells expressed CD73 and CD39 (a2). Only cells from the CD45- fraction presented in the t-SNE plot were gated[14], showing co-expression of CD73 and CD39 at 99.9% (a3). CD45-CD73+CD39+CD146+ population was not expressed in non-mobilized leukopaks (mononuclear cells) (a4); B: Representative flow cytometry analysis showing subset in single mobilized peripheral blood stem cells (PBSCs). CD45- fraction of total population was gated (b1). tSNE plot in combination with flowSOM and cluster explorer was used. 1.5% of CD45- cells expressed CD73 and CD39 (b2)[14]. Only cells from the CD45- fraction presented in the tSNE plot were gated, showing co-expression of CD73 and CD39 at 96.8% (b3)[14]. From CD45-CD73+CD39+ cells, approximately 80% are positive for CD146+ (b4); C: Representative flow cytometry analysis of target subset in single mobilized PBSCs. CD45 negative fraction of total population was gated (c1). tSNE plot in combination with flowSOM and cluster explorer was used. 5.7% of CD45- cells expressed CD73 and CD39 (c2)[14]. Only cells from cluster of CD45- fraction presented in the tSNE plot were gated and showed co-expression of CD73 and CD39 at 99.8% (c3)[14]. Among the CD45-CD73+CD39+ cells, 25.1% were positive for CD146+ (c4). All gates were established using Fluorescence MinusOne and unstained controls (A-C); D: Percent of CD45-CD73+CD39+CD146+ in single-mobilized PBSCs (< 1%). Bar graph is a result of mean ± SD (n = 6); E: Percent of CD45-CD73+CD39+CD146+ in dual-mobilized PBSCs (< 1%). Bar graph is a result of mean ± SD (n = 3).

DISCUSSION

Heterogeneous populations of MSCs with distinct biological properties can be sourced from various locations. In prior research, our lab identified a specific subset of MSCs from the human synovium that co-expresses CD73 and CD39, demonstrating a higher potency for osteogenic and chondrogenic differentiation. Previous unpublished data showed that this subset also produced adenosine and expressed CD146, a marker associated with pericytes in BM as well as other tissues such as adipose tissue[6,7,12,16]. This study aimed to investigate whether the CD45-CD73+CD39+CD146+ pericyte-like subset was present in BM first, followed by peripheral blood. CD146 expressing pericytes have been implicated as regulators of hematopoiesis in the BM niche[6,7]. Pericytes express factors such as CXCL12, involved in HSC survival and differentiation, and angiopoietin-1 an integral player in primitive/Long term-HSCs maintenance[6,7,17]. Pericytes are also primarily responsible for homing HSCs in BM[7]. MSCs have been used as assets in mediating graft vs hosts disease post-HSC transplants[18,19]. Data indicated that this was primarily associated with mechanisms related to repair of the BM niche and the inhibition of HSC differentiation[16,19-21]. The success of HSC transplants depends on the preservation of key mechanisms that are crucial for cell engraftment[16,19-21]. Currently, nearly 90% of HSC transplants are sourced via allogenic donation from peripheral blood[22]. Mobilization practices can be taxing for donors, potentially life-threatening, and a donor may only donate up to twice in their lifetime[22-24]. Thus, the in vitro expansion of HSCs presents a promising solution to the issues of donor availability and extended procurement times affecting patients awaiting transplants[2,25,26]. In 2013, Corselli et al[16] showed that co-cultures of CD146+ MSCs (pericytes) and HSCs enhanced post-transplantation engraftment. Overall, these studies support our hypothesis that using MSC-like pericytes in an in vitro co-culture system with HSCs could be a promising approach. This method may lead to an optimal system for expanding HSC cultures in vitro, allowing us to generate multiple therapeutic doses from a single donor, thereby creating off-the-shelf transplant products. Our previous study demonstrated that the synovium-derived CD73+CD39+ subset was present in culture in varying percentages across multiple passages and donors up to 25% at P10[11]. We also reported the presence of this subset in fresh synovium digests. In our current study, we found that the subset of interest was present both in vitro with 80% at P2 and 99% at P3, and in vivo, as indicated by our immunofluorescence data. The higher percentage of our subset in BM may be related to its potential role as pericyte-like MSCs in that tissue[6,7,12]. Sacchetti et al[7] reported that 99% of clonal cultures of BMSCs expressed CD146, while non-clonal cultures expressed the marker at a rate of 30%. Consistent with this study, our data indicated that 99% of the CD45-CD73+CD39+ subset also expressed CD146. In vivo, CD146+ MSCs located in the perivascular niches play a role in remodeling the BM niche and facilitating the flow of cells from the BM to the peripheral blood[7]. Our immunofluorescence data supported this finding, showing the triple positive localized near the vessels. Plastic adherence, self-renewal, and multipotency were assessed to determine whether the subset exhibited MSC-like characteristics[13]. Our data indicated that the subset displayed all the properties outlined by Dominici et al[13]. Interestingly, this subset showed slightly higher osteogenic ability compared to the positive control of osteoblasts. Our previous work highlighted that synovium derived CD73+CD39+ cells exhibited greater osteo-chondrogenic potency[12]. The increased osteogenic potential of the BM-derived subset may result from heterogenicity within the isolated subset or could represent a population of cells involved in bone repair[10,11,27,28]. Additionally, research suggested that CXCL12 enhanced the osteogenic differentiation of MSCs[27]. In culture, MSCs treated with CXCL12 in osteogenic growth medium exhibited more osteogenic potential than those cultured in osteogenic medium alone[27,28]. Since CXCL12 is a key factor expressed by BM-derived CD146+ cells (pericytes)[7], our data suggests that the higher osteogenic potency of our BM-derived CD146+ subset may be linked to CXCL12 secretion. This notion will require further investigation. The significance of CD146+ MSCs isolated from perivascular niches extends beyond the BM; their presence has also been reported in adipose tissue and peripheral blood[12,16]. Interestingly, the target cell subset produced adenosine, which is known to play a role in immune modulation; however, this interesting and novel concept will require further investigation. Since Gift of Life's primary source of material for HSC transplant is MB rather than BM, we investigated whether the CD45-CD73+CD39+CD146+ subset could be detected in peripheral blood collected through apheresis. Our facility also collects non-MB samples, so we investigated whether the target subset could be detected from this source. There is evidence that MSCs exist in peripheral blood[29-31] at varying percentages. Zvaifler et al[29] reported 0.3%-0.7% fractions were MSCs while Jain et al[31] reported 0.001% of white blood cells from peripheral blood. In our study, we did not find our target MSC subset in non-mobilized peripheral blood (MNCs). However, consistent with Zvaifler et al’s work[29], we observed that < 1% of the analyzed population consisted of MSCs (data not shown). We also evaluated whether mobilization using G-CSF (filgrastim) affected the mobilization of our subset in peripheral blood. Tormin et al[6] reported that BMSCs expressing CD146 have a unique ability to maintain HSCs in the BM via the CXCL12 and C-X-C chemokine receptor type 4 (CXCR4) axis. Disruption of the CXCL12 and CXCR4 axis, promoted by filgrastim, leads to the mobilization of HSCs from the BM into peripheral blood[6,30]. Jain et al[31] reported that administering G-CSFs for 5 days increased the percentage of MSCs in both peripheral blood and apheresis product. Our data aligns with Jain et al[31] showing a slight increase in our subset in MB following a 5-day G-CSF administration. Plerixafor is an antagonist of the CXCR4 receptor, which interferes with the CXCL12 domains that retain HSCs in BM[6,32]. The dual-mobilization of G-CSF and plerixafor has been found to enhance mobilization of HSCs[32,33], and plerixafor is approved as an adjuvant for G-CSF HSC mobilization[33-35]. Based on these findings, we investigated whether dual mobilization using filgrastim and plerixafor affected the presence of CD45-CD73+CD39+CD146+ MSC subset in the apheresis product. Our data indicated that this subset was present in dual mobilized PBSCs, but the dual mobilization did not significantly impact the subset compared to single mobilized PBSCs. This lack of impact may be due to the binding affinity plerixafor has to the CXCR4 receptor, which facilitates the release of HSCs into peripheral blood while leaving MSCs retained in the BM[32]. Another study reported that dual mobilization increased the presence of MSCs in BM but not in PB[36]. In a research study involving donors with multiple myeloma and lymphoma, plerixafor was administered after G-CSF mobilization to rescue MSCs; however, the results showed no significant impact in either peripheral blood or the apheresis product[31]. Our findings are consistent with these studies, although further research is needed to understand the underlying mechanisms. Additionally, donor receptivity to plerixafor may influence our dual mobilization results. Gift of Life has identified differences among donors in three groups: Non-mobilized, single mobilized, and dual mobilized, which affect all blood components (data not shown). Although we have not yet pinpointed the traits causing these differences, age and sex are known to influence blood cell levels[37]. Studies indicate that younger donors and males often yield more CD34+ HSCs[37]. All donors involved in this study were healthy individuals aged 18 to 35. However, a notable limitation of our research is the small sample size of mobilized peripheral blood collected, which consists of only six samples (n = 6). Challenges in finding donors and managing the risks of stem cell mobilization[2,22-24] restricted our sample size. In the future, we will need a larger donor pool to achieve more reliable results.

Gift of Life is exploring ways to grow adult HSCs in the lab for off-the-shelf transplants. HSCs can lose their stem-like abilities in culture, so creating the right environment is essential[3,5,25]. We believe that CD45-CD73+CD39+CD146+ cells could mimic the BM environment in vitro. It is essential to isolate this group of cells along with HSCs to understand their role. Future work will focus on studying this subset from mobilized peripheral blood.

CONCLUSION

The co-expression of CD73 and CD39 allowed for the identification and enrichment of a specific subset of MSCs located perivascularly in BM cultures. This subset was also positive for CD146 and existed in MB. We believe that leveraging this subset in a co-culture system with HSCs could lead to groundbreaking methods for expanding HSCs, ultimately creating a readily available inventory of products for transplants. Further studies will be conducted to assess the functionality of the CD73+CD39+CD146+ MSC subset derived from PBSCs.

ACKNOWLEDGEMENTS

We thank Gift of Life’s founder and CEO, Jay A Feinberg and Gift of Life’s COO, Marti Freud (JD) for the funding, support, and time granted to perform this work. We are further grateful to our donors who participated in this study; without you, this work would not have been possible. We also thank Dr. Miriam and Sheldon G Adelson Gift of Life-NMDP Collection Center, the apheresis team, and Bruce A Lenes (MD), Yen Michael S Hsu (MD, PhD, MBA, CABP), and Richaele Nichiporeko (DNP, APRN, FNP-BC) for their part in collection and support on this project. We also would like to thank Farida Khan [MT (ASCP), SCYM] and Virginia Meyer [MS, MLS (ASCP)] and the entire Gift of Life CLIA team for their part in sample procurement. A big thank you to Andrea Pena and Briana McIntosh and the entire Center for Cell and Gene Therapy Team for their laboratory expertise and endless support. We also would like to send our gratitude to Applied Pathology Systems and Lydia Zhang for their continued support and efforts with the bone marrow analysis. Additionally, we thank Emily Fox and the UF Scripps Flow Cytometry core for their advice and use of their facilities. A big thank you Christian Aguilera-Sandoval (PhD) for his support, training, and incite. Kathryn Martin would also like to acknowledge Kimberly Susan for inspiring the creativity which drove this work.