Performance of Developed Mononuclear Cell Separator for Regenerative Therapy Is Comparable to Manual

Abstract

Regenerative therapy involving transplanted bone marrow mononuclear cells (BM-MNCs) and hematopoietic stem cells (HSCs) is markedly effective against many diseases. However, manual MNC separation requires skilled labor and cell-processing centers. Thus, efforts have been targeted toward fractionating MNCs using cell separators. A double-blind, placebo-controlled study of myocardial infarction using BM-MNCs separated by an existing device was conducted, and no therapeutic effects were found. The development of a cell separator to replace manual techniques would significantly contribute to the widespread application of BM-MNC therapy. Therefore, we developed a BM-MNC separation device that can reproduce manual separation. We changed the shape of the injection port on the centrifuge container and determined its circuit to improve the performance of HSC separation and remove degenerative red blood cells (RBCs). We assessed HSC recovery and degenerative RBC removal rates using fluorescence-activated cell sorting. Additionally, we evaluated the therapeutic effects of cells separated using our device in mouse models of stroke. The HSC recovery and degenerative RBC removal rates using the device were comparable to those obtained using manual separation. We also confirmed the therapeutic effects of BM-MNCs separated by the device in the models. The new automated device could replace manual cell separation and render cell-based therapy using BM-MNC feasible without laborious manual tasks at dedicated cell-processing centers.

Impact Statement

Transplanted mononuclear cells (MNCs), including hematopoietic stem cells, confer beneficial effects on ischemic diseases, but they are not widely applied clinically. This is partially due to the need for costly manual techniques at dedicated cell-processing centers. Our device automatically separates MNCs that are sufficiently effective for transplantation. We believe that many patients with ischemic disease will benefit from MNC transplantation.

Introduction

The therapeutic mechanism of hematopoietic stem cell (HSC) transplantation in cerebral infarction involves energy supplied via gap junctions between them and damaged vascular endothelial cells.¹ Gap junctions connect the cytoplasm of adjacent cells and allow the rapid intercellular movement of small water-soluble molecules according to their concentration gradient.² We previously found that HSCs contain significantly higher levels of various metabolites in the cytosol compared with those in endothelial cells.³ Water-soluble molecules directly transferred between transplanted bone marrow mononuclear cells (BM-MNCs), a rich cell fraction of HSC, to the cerebral endothelium via gap junctions, induce angiogenesis by activating hypoxia-inducible factor-1α (HIF1α) at the endothelium in murine stroke models.¹ Furthermore, integrin β2-induced HSC adhesion is important in this process.⁴

Although the therapeutic mechanism of BM-MNC transplantation continues to be investigated and many clinical studies are being conducted,^5–8 therapy for ischemic diseases using these cells is insufficient. The reasons for this might be associated with the need for expensive dedicated cell-processing centers (CPCs), as well as complex, advanced, manual separation techniques. Although an automated closed-system cell separation device is needed to help address this issue,⁹ devices that can separate therapeutically effective MNCs are not available. Therefore, we aimed to develop a BM-MNC separation device.

We focused on the removal of degenerative red blood cells (RBCs) and the shape of the injection port on the container. A previous large double-blind, placebo-controlled clinical trial of BM-MNCs fractionated by automated devices rather than manual cell separation did not produce therapeutic effects.¹⁰ We investigated the cause of this and found that the cells separated using the existing device included clotted degenerative RBCs. These cells caused macrophages to accumulate around them, inhibiting angiogenesis when transplanted.¹¹ Furthermore, RBCs suppress the therapeutic effectiveness of autologous BM-MNC transplantation.¹² Therefore, we focused on removing degenerated RBCs. We considered that the shape of the injection port might be problematic. Cells are separated by density gradient centrifugation based on the speed of sedimentation, which depends on specific cell density and the premise that all cells will pass through a density gradient solution. However, in the conventional product, the upward convex inlet container, some cells did not pass through the density gradient media because cells were pushed outward along the wall surface by centrifugal force. Additionally, contamination of the buffy coat with RBCs attached to the wall of a downward-facing shape during the postseparation recovery process may occur. Therefore, we aimed to develop a BM-MNC separation device that would address these issues. We compared the separation performance between our automated device and the manual method and verified the quality of the separated cells. We previously found a significant increase in the amount of vascular endothelial growth factor (VEGF) uptake by human umbilical vein endothelial cells (HUVECs) cocultured with manually separated MNCs, including HSCs.¹ This increase indicated the potential of cocultured HSCs to promote vascular endothelial proliferation. The ability of HSC therapy is substantially influenced by the expression of CD18 (integrin β2), which is an adhesion factor expressed on HSCs.⁴ Therefore, we designed and developed a separation device and investigated the therapeutic effects of BM-MNCs obtained using it in mouse models of cerebral infarction.

Methods

Our device separated MNCs from BM using density gradient centrifugation, fluid control, a closed-cell separation circuit for the BM (Fig. 1A, B), density gradient medium, and cell washing solution. The device was equipped with pinch valves for switching flow channels, motors for centrifugation, pneumatic circuits to drive plungers in the containers, and camera detection systems to monitor the liquid volumes in the containers. Density gradient medium was injected into a vertical centrifuge container, then BM was gently introduced while rotating (Fig. 1C, Supplementary Movie). The BM was then separated into the following layers from the center of the container using centrifugation: plasma, buffy coat (including MNCs), Ficoll density gradient medium (specific gravity: 1.077 g/mL), granulocytes, and RBCs (Fig. 1D). Thereafter, the plunger was pressed to recover liquid from the center, which allowed plasma removal and the collection of buffy coat, including target MNCs, by switching the flow path. Suspended MNCs were washed in another rotating container, and the final sample was recovered. This system was designed to have a separation ability similar to that of manual separation (Fig. 1E). All processing times were 3 h.

FIG. 1.

Device configuration and cell separation. (A) Device for density gradient centrifugation and fluid control (left) and closed-cell separation circuit (right) set in the device. (B) Diagram and photo of vertical centrifuge container in closed-cell separation circuit. In diagram, the unit of size is millimeters. (C) Appearance of vertical centrifuge container at BM layering. Bone marrow is injected into transparent Ficoll in rotating container without disturbing borders. (D) Schematic diagram of horizontal and vertical sections of container contents after complete separation. Plasma layer (yellow), buffy coat layer (MNCs, blue), Ficoll layer (pink), and the granulocytes and RBCs layer (red) are shown from inside. Layer thickness is not precisely shown for clarity. (E) Status of manually separated cells before and after centrifugation. Before separation by centrifugation, BM sample is layered on top of Ficoll without mixing. Plasma, buffy coat (MNCs), Ficoll, and granulocyte and RBC layers are clear from the top after separation. BM, bone marrow; MNCs, mononuclear cells; RBCs, red blood cells.

Experiment

Experimental design

Ethics declaration

Umbilical cord blood (UCB) was collected from babies naturally delivered at full term, after obtaining informed consent from all mothers. All experiments were approved by the Ethics Committees of the Foundation for Biomedical Research and Innovation at Kobe. We confirm that the study complied with the ethical principles enshrined in the Declaration of Helsinki (2013 amendment). The Animal Care and Use Committee of the Foundation for Biomedical Research and Innovation in Kobe, Japan, approved the animal experiments (No. 21-04), which complied with the Guide for the Care and Use of Animals published by the Ministry of Education, Culture, Sports, Science, and Technology in Japan. The experiments and results are described according to the ARRIVE guidelines.

Cell preparation

Three samples of human BM (Lonza, Basel, Switzerland) shipped with heparin and used approximately 4 days after collection were diluted in 2% human albumin (albumin 25% I.V. 5 g/20 mL-BENESIS; Japan Blood Products Organization, Tokyo, Japan) and 15% ACD-A solution (Terumo, Tokyo, Japan) in RPMI 1640 Medium, no phenol red (Thermo Fisher Scientific Inc., Waltham, MA, USA) (hereinafter, this solution containing RPMI as the main component is abbreviated as RPMI). The Umbilical Cord Blood Bank donated UCB preserved in a solution of glucose and citrate. The UCB was diluted with RPMI as described above for BM immediately before use.

Manual separation and stock of MNCs

Human BM diluted five-fold with RPMI was layered into 50-mL tubes containing Ficoll–Paque Premium density centrifugation medium (Cytiva, Marlborough, MA, USA) and separated by centrifugation at 400 g for 40 min. While carefully avoiding contaminating clots, we manually separated and washed the buffy coat suspended in RPMI twice by centrifugation at 200 and 150 g for 10 min each. The enriched MNC suspension was analyzed using flow cytometry. Some cells were frozen and stored in liquid nitrogen to study VEGF uptake. Frozen samples were rapidly thawed at 37°C in a water bath. The UCB-MNCs were separated as described above for human BM-MNCs.

Automated MNC separation

A control device and closed-cell separation circuit were used to automatically collect stem cells. Three-fold diluted BM in RPMI (150 mL), Ficoll (300 mL), and 1.1 L of RPMI were added to wash cells using a bag in the closed-cell separation circuit. After the setup was complete and the process was initiated, BM was layered onto the Ficoll solution, followed by density gradient centrifugation. The buffy coat was collected, and MNCs were washed. Thereafter, BM-MNCs were collected in a recovery bag within 3 h of cell processing and analyzed using flow cytometry. Some MNCs were frozen. The UCB-MNCs were separated in the same way as human BM-MNCs.

Flow cytometry

Fluorescence-activated cell sorting (FACS) was performed using the FACSLyric flow cytometry system (BD Biosciences, Franklin Lakes, NJ, USA). Table 1 shows the antibodies and markers. Absolute counts were determined using Trucount tubes (BD Biosciences). We identified, counted, and characterized RBCs using the antibodies in Table 1 (Tube 1). Supplementary Figure S1 shows CD45⁻, CD61⁻, and CD235a⁺ cells with a diameter ≥3 µm. We counted degenerated RBCs using Annexin V, which binds to the eat-me signal phosphatidylserine to verify therapeutic inhibitory constituents.¹³ Removal rates of RBCs or Annexin V⁺ RBCs were calculated as follows:

\frac{(number i n sample before separation - number i n final suspension)}{number i n sample before separation}

Table 1.

Antibodies Used in Flow Cytometry

Tube	Antibody	Fluorochrome	Manufacturer	Catalog No.	Cell type
1	CD235a	FITC	Beckman Coulter	IM2212U	RBCs
	CD45	PE	Beckman Coulter	A07783	WBCs
	CD61	APC	BD Bioscience	564174	Platelets
	7AAD	—	BioLegend	640926	Dead cells
	Annexin V	Pacific Blue	BioLegend	640926	Degenerated cells
2	CD45	FITC	BD Biosciences	344563	WBCs
	CD34	PE	BD Bioscience	344563	HSCs
	7-AAD	—	BD Bioscience	344563	Dead cells
	CD3	Krome Orange	Beckman Coulter	B00068	Lymphocytes
	CD19	Krome Orange	Beckman Coulter	A96418	Lymphocytes
	CD56	BV510	BD Bioscience	563041	Lymphocytes
	CD14	APC	Beckman Coulter	IM2580	Monocytes
	CD33	APC	Beckman Coulter	IM2471	Monocytes
	CD16	Pacific Blue	Beckman Coulter	B36292	Neutrophils
	CD294	BV421	BD Bioscience	562992	Eosinophils
	CD123	BV421	BD Bioscience	567279	Basophils
3	CD34	APC	BD Biosciences	555824	HSCs
	CD45	FITC	BD Biosciences	555482	WBCs
	7-AAD	—	BD Biosciences	559925	Dead cells
	CD18	PE	BD Biosciences	Customized	—

7-AAD, 7-aminoactinomycin D; APC, allophycocyanin; BM, bone marrow; BV421, Brilliant Violet 421; CD, cluster of differentiation; FITC, fluorescein isothiocyanate; HSCs, hematopoietic stem cells; PE, phycoerythrin; RBC, red blood cells.

We quantified CD34⁺ HSCs using the International Society of Hematotherapy and Graft Engineering (ISHAGE) single platform¹⁴ (Table 1, Tube 2). Granulocytes were defined as the sum of neutrophils, eosinophils, and basophils. MNCs were the sum of monocytes and lymphocytes. The recovery rates of HSCs or MNCs were calculated as the numbers in the final suspension relative to those in the sample before separation. The removal rates of granulocytes were calculated using the equation described above for RBCs and Annexin V⁺ RBCs. We quantified CD18 expression in three leukocyte fractions (lymphocytes, monocytes, and granulocytes) and HSCs as described⁴ (Table 1, Tube 3) using BD QuantiBRITE (BD Biosciences).¹⁵ Median CD18 values were determined using the FACSLyric, then the absolute numbers of CD18 molecules were calculated. Antigens were quantified using a customized phycoerythrin-labeled CD18 antibody with a fluorescein/protein at a ratio of 1:1 based on mouse antihuman CD18 (clone L130; BD Biosciences). HSCs were gated using the ISHAGE single platform.¹⁴ The three leukocyte fractions were gated using SSC-CD45 scatter plots.

Comparison of separation performance according to injection port shape

We used Sepax Kit (Cytiva, CS-430.1) with a convex upward injection port into a container as the existing product. We created another container with a concave upward injection port with the same radius, which was smaller than the containers in our device; therefore, we used 60 mL of diluted BM. We compared the performance of density gradient centrifugation using the unwashed buffy fraction of human BM. The HSCs in each fraction were counted using FACS, which compared and verified the proportion of HSCs in the buffy fraction to the total number of HSCs.

Validation of separated cell viability

The human BM was inappropriate for verifying the viability of the separated cells because approximately 4 days had passed since they were harvested. We therefore verified viability using pig BMs (Hamaguchi Lab Plus, Osaka, Japan) on the day of or the day after harvest. These BMs were immediately suspended in 10 units of sodium heparin (Mochida Pharmaceutical Co., Tokyo, Japan) and diluted in RPMI. The pig BMs were separated as described for the other specimens, and viability was measured using an ADAM MC Automated Cell Counter (NanoEnTek, Seoul, Korea).

Uptake of VEGF by HUVECs

We assessed VEGF uptake in cocultured HUVECs and MNCs as described to verify the activation of vascular endothelial cells by MNCs in vitro.¹ The HUVECs were cultured as described by the manufacturer and used at the sixth passage. Biotin-conjugated recombinant human (rh)VEGF (BT10543-025, R&D Systems, Minneapolis, MN, USA) was incubated with streptavidin-conjugated allophycocyanin (APC; SA1005; Thermo Fisher Scientific Inc.) at a molecular ratio of 4:1 for 10 min at 25°C. The numbers of thawed cryopreserved BM-MNCs were measured in advance and cocultured with HUVECs and APC-labeled rh VEGF at 37°C for 2 h. The reaction ratio of BM-MNCs to HUVECs was the same for manual and automated separations. The cells were stained with PE-conjugated CD31 antibody (555446), FITC-conjugated antihuman CD45 antibody (555482), and 7-AAD (559925; all from BD Biosciences). The intensity of APC fluorescence in CD31⁺, CD45⁻, and 7AAD⁻ HUVECs was assessed using flow cytometry with the FACSLyric. The controls were HUVECs incubated without MNCs, and median ratios to the control were calculated. These experiments were conducted in quadruplicate.

Stroke model and cell transplantation

We created reliable stroke models using 5-week-old male severe combined immunodeficiency (SCID) mice (CB-17/lcr-scid/scidJcl; Oriental Yeast, Tokyo, Japan) as described.¹⁶ Focal cerebral ischemia was induced by permanent ligation and disconnection of the distal portion of the left middle cerebral artery using bipolar forceps under isoflurane anesthesia. Subsequently, 1 × 10⁵ BM-MNCs in 100 μL of RPMI or RPMI were intravenously injected into the tails of mice at 48 h after stroke induction.

Immunohistochemistry

We anesthetized mice with sodium pentobarbital and then transcardially perfused them with saline, followed by 2% paraformaldehyde. Coronal sections (20 µm) were cut using a vibratome (Leica, Wetzlar, Germany) and immunostained with 1:50-diluted primary antibodies against F4/80 (MCA497R; Bio-Rad Laboratories Inc., Hercules, CA, USA) and 1:50-diluted CD31 (550274; BD). The sections were visualized using the 3,3'-diaminobenzidine method and counterstained with Mayer’s hematoxylin solution (Wako, Osaka, Japan). We counted cells in one coronal section, each located at the bregma, and 320 µm anterior and posterior to it. The numbers of macrophages/microglia (F4/80⁺ cells) in a randomly selected region in 0.25 mm² of the peri-stroke area were quantified. Supplementary Figure S2 shows the blood vessels (CD31⁺ cells) quantified by an investigator who was blinded to whether the mice were injected with or without BM-MNCs.

Open-field test

We assessed cortical function by testing the behavior of mice testing in an open-field task 11–13 days after treatment with BM-MNCs. The mice were free to search the interior of a square acrylic box (30 × 30 × 30 cm²) for 60 min. Lights were on for the first 30 min (light period) and then turned off for 30 min (dark period). We compared how many times the mice crossed the infrared beam (2 cm above the floor and spaced at 10 cm intervals on the X and Y axes [locomotion]) during each period. We compared locomotor activity between the phases and the ratios of how many times the mice crossed the beam during the light and dark periods.

Statistics

Between-group differences were compared using Student t-tests. All results are presented as means ± standard deviation (SD). Values with p < 0.05 were considered statistically significant.

Experimental Results

Comparison of separation performance between containers with different inlet shapes

In the upward convex inlet container, the wall surface at the top of the container has a downward-sloping shape. This causes BM cells to be pushed outward along the wall surface by centrifugal force, resulting in some BM cells being centrifuged without passing through the Ficoll (Fig. 2A). In contrast, in the newly developed upward concave inlet container, the wall surface at the top of the container has an upward-sloping shape. Consequently, BM cells cannot move along the container wall surface against gravity. Therefore, all BM cells are subjected to centrifugal force and passed through Ficoll (Fig. 2B). The proportion of HSCs in the buffy fraction was higher in the container with the concave than that in the convex inlet (Table 2).

FIG. 2.

Comparison of separation behavior according to shape of injection port. (A) Some cells injected into the upside convex inlet container moved outside along the wall surface by centrifugal force, resulting in the cells being centrifuged without passing through Ficoll. (B) All cells passed through Ficoll without moving along the wall surface against the gravity when injected into the upside concave inlet container. Image 1 is immediately after bone marrow injection, and image 2 is a little later. Schematic diagram shows cell movement after injection. Red circle, bone marrow cell; green arrow, direction of container rotation; blue arrow, centrifugal force acting on cells; black arrow, direction of cell movement. The container was filled with Ficoll in advance.

Table 2.

Separation Performance Differs According to Inlet Shape

Inlet shape	n	Ratio of HSCs in buffy fraction (%)
Convex upward	3	48.9
		66.2
		70.1
		Average: 61.7%
Concave upward	2	75.5
		89.9
		Average: 82.7%

Comparison of RBC and annexin V ⁺ RBC removal rates between manual and automated device

The rates of removed RBCs exceeded 99.9% in both methods. The rates of removed Annexin V⁺ RBCs exceeded 97% in both methods and were similar between the two methods (Table 3).

Table 3.

Comparison of Human Bone Marrow Performance Between Manual and Automatic Separation

(%)	Manual	Device	p
Removal
RBCs	99.96 ± 0.02	99.95 ± 0.04	0.64
Annexin V⁺ RBCs	97.31 ± 1.27	97.01 ± 0.46	0.75
Granulocytes	65.9 ± 22.1	78.3 ± 9.4	0.40
Recovery
HSCs	31.5 ± 11.3	23.3 ± 7.2	0.46
MNCs	20.4 ± 10.1	18.0 ± 6.5	0.80

Results are expressed as means ± standard deviation. n = 3.

MNCs, mononuclear cells.

Comparison of separation performance between manual and automated separation

Several parameters were analyzed using FACSLyric to determine the separation performance of the two methods (Table 3). The recovery rates of HSCs and MNCs in the BM separation experiment and the removal rates of granulocytes were similar between the two methods. The SDs of recovery or removal rates were smaller and tended to be less variable using the automated method than those with the manual method. The other UCB source for HSCs was separated using our automated device and the manual method. The recovery rates for HSCs and MNCs were similar in both methods, but the SDs were lower for our automated device (Fig. 3A).

FIG. 3.

Comparison of performance between automated and manual separation. (A) Similar recovery rates of HSCs and MNCs separated from UCB using automated or manual methods. X indicates average value. (B) Similar VEGF uptake by HUVECs coincubated with BM-MNCs isolated manually or automatically. (C) Expression of CD18 in granulocytes derived from BM-MNCs after separation is significantly lower than that of BM before using our device, and tended to be lower with manual separation. *p < 0.05. n = 5 (A), n = 4 (B), and n = 3 (C). HSCs, hematopoietic cells; HUVECs, human umbilical vein endothelial cells; MNCs, mononuclear cells; UCB, umbilical cord blood; VEGF, vascular endothelial growth factor.

Viability of cells separated by automated device

We assessed the effects of separation on the viability of two recently acquired porcine BM cell samples. We found that the automated device maintained the cell viability at the level that it was before separation on the day of or the day after harvest. These findings confirmed that separation using our automated device did not affect cell viability (Table 4).

Table 4.

Comparison of Viability Between Manual and Automated Separation

(%)	Day of harvest	Day after harvest
(%)	Sample 1	Sample 1	Sample 2
Before	98.4	98.9	98.9
Manual	98.9	98.1	98.7
Automated	99.1	99.1	97.7

Sample 1 was analyzed on harvest day and the day after. Sample 2 was measured only on the day after harvest.

Amount of VEGF uptaken by HUVECs cocultured with MNCs

We assessed VEGF uptake in HUVECs cocultured with MNCs separated by our automated device in the same manner as manual separation¹ (Fig. 3B). The MNCs separated by the device retained the ability to promote vascular endothelial cells.

Expression of CD18 in separated cells

We compared CD18 expression in leukocytes (HSCs, monocytes, lymphocytes, and granulocytes) in the buffy coat layer with that in BM before separation. The expression of CD18 in granulocytes was significantly lower after than before separation, particularly in our device (Fig. 3C). The amount of CD18 expression in the other cells did not considerably differ between before and after.

Microglia/macrophage accumulation and microangiogenesis in transplanted MNCs

We examined the histological effects of BM-MNCs separated using our automated device and transplanted into mouse models of stroke. We quantified cerebral F4/80 positive microglia/macrophages in the peri-infarct area and microvessels in the infarct core and peri-infarct area 24 h after BM-MNCs were transplanted. The number of F4/80 positive cells in the peri-infarct area of mice transplanted with automatically separated BM-MNCs was similar to that in control mice. Microglia/macrophages did not accumulate when BM-MNCs were automatically separated (Fig. 4A, D). However, the number of microvessels in the peri-infarct area in mice treated with BM-MNCs was significantly increased compared to that in the control, and numbers in the infarct core tended to be higher than those in controls (Fig. 4B, E). These results show that automatically separated MNCs exerted significant therapeutic effects on the mouse stroke model.

FIG. 4.

Histological analysis of therapeutic effects of automatically separated BM-MNCs. (A) Numbers of F4/80 positive microglia/macrophages in peri-infarct area (0.25 mm²) did not differ between untreated controls and mice treated with BM-MNCs separated using our device. (B) Treatment with BM-MNCs separated using our device significantly enhanced microvascular regeneration in peri-infarct area. (C) Gray infarct area in coronal brain sections. Red boxes indicate lower magnification. (D) Accumulation of F4/80 positive microglia/macrophages around infarct area 24 h after cell transplantation in similar between untreated control and treated mice. (E) CD31⁺ microvasculature in peri-infarct area and infarct core 24 h after cell transplantation is more abundant in treated mice than in controls. Scale bars indicate 500 µm (lower magnification) and 200 µm (higher magnification). Square shows area of higher magnification (D, E). Control: n = 5, device: n = 4 (A, B). *p < 0.05.

Neurological function of mouse stroke models after cell transplantation

The neuronal functional effects of the automatically separated BM-MNCs were examined in the mouse stroke models at 11–13 days after transplant. The behavior of the control mice was abnormal in the open-field tests, with decreased activity under darkness. In contrast, the behavior of the mice transplanted with automatically separated MNCs was normal, and the mice tended to be more active in the dark (Fig. 5A). The ratio of locomotor activity in dark versus light conditions was significantly higher in the transplanted mice than that in the control mice (Fig. 5B).

FIG. 5.

Therapeutic effects of automatically separated BM-MNCs determined using behavioral tests. Transplanted BM-MNCs isolated automatically improved neural function in open-field tests. (A) Number of locomotor events under light (white bars) and dark (black bars) phases. (B) Ratios of number of events in dark and light periods. *p < 0.05. n = 5.

Discussion

We developed a new automated device with the potential to replace manual cell separation. Our automated cell separator will be ideal for treating strokes with cells without the need for complex, laborious manual procedures.

One of the main goals for our automated device was to remove clot-derived degenerated RBCs that are inhibitory constituents of transplanted MNCs.¹¹ Here, we used the eat-me signal, phosphatidylserine, which is associated with hypercoagulation during exposure, as a marker for degenerated RBCs.^17,18 These were removed using our device in the same way as the manual method. The current rates of RBC removal from BM are >99%^19–21 using current automatic separation devices that require a specific density liquid. However, treatment is occasionally ineffective.^10,11 One reason for this might be that degenerated RBCs have been overlooked. We believe that the therapeutic effects of automatically separated MNCs on stroke in mouse models were a result of removing degenerated RBCs.

Another key characteristic of our device is the concave injection site that differs from the existing site (Fig. 1B). When injected into a rotating container via the concave site, all blood cells were exposed to centrifugal force as they passed through the Ficoll. This allowed them to remain in concentric layers appropriate to each specific density. The HSCs were separated at a specific density similar to the manual method.

The recovery rates of HSCs or MNCs from BM samples tended to be lower than those reported.^19–21 This might be due to the fact that we used BM that had been harvested and stored long ago. Importantly, the rates were comparable between our automated and manual separations because cell therapy using manually separated MNCs is effective.^6–8,22 Automated separation notably did not affect cell viability, indicating that our device is suitable for clinical applications.

Furthermore, the automated device offers an advantage in terms of recovery rates, having a lower SD and less variability than those with manual separation (Table 3, Fig. 3A). Some disadvantages of manual work include the need for skilled technical personnel and variability among operators. Automated cell separation guarantees the effectiveness of this technology.

This research aims to improve existing automated systems; however, the focus of the design was not simply on developing equipment to increase the recovery rate of HSCs but on improving the ability to separate cells with therapeutic effects in line with therapeutic mechanisms. For example, by following this design, processing time is spent on BM overlay and washing, but it does not affect BM collection and treatment on the same day.

One of the important points to be considered in an automated system is reproducibility. In the case of autologous products in which there is substantial intrinsic variation in starting material, there is precedent for developing a prediction algorithm to guide the manufacturing steps to achieve the desired product profile.⁹ In this study, we found that the device had less variability compared to that with manual separation. However, it has not yet been verified whether the MNCs necessary for therapeutic effects can be separated qualitatively and quantitatively in all autologous BMs. We believe that accumulating data from future clinical studies will lead to the development of predictive algorithms.

Generally, one of the goals of automation system is to increase processing capacity in one go. However, we believe flexible scalability is not necessary for this device, as it is intended for cell separation for autologous BM therapy. In this study, we verified separation from 150 mL of diluted BM. Considering the amount of BM collected from a patient, there is likely no need to scale it up any further. This device is also capable of separating cells from even smaller amounts of diluted BM, but considering the need to ensure the number of cells required for therapy, we believe that reducing the sample volume is not necessary. Nevertheless, verification of appropriate amount of BM harvested and the number of cells administered through clinical studies would be necessary in the future. Moreover, if a treatment method using MNCs separated from more nonautologous sources is developed in the future and larger-scale production of cell preparations becomes necessary, it may be necessary to consider scaling up the method.

Less CD18 was expressed in granulocytes in the buffy coat layer than in BM before separation, which we believe was due to the influence of Ficoll. However, the mechanism remains unknown and requires future verification. Ficoll affects granulocyte chemotaxis.^23,24 The low expression of CD18 in granulocytes might promote HSCs’ binding to vascular endothelial cells, considering competition among lymphocytes.²⁵ Clinical²⁶ and research²⁷ findings have shown that MNC therapy using density centrifugation solutions other than Ficoll is ineffective. Ficoll might be a positive factor for using this device.

Although our innovative device is comparable to the manual method, it adequately separates therapeutic MNCs from good samples that do not progress through the coagulation cascade. This demands the use of anticoagulants.^28,29 However, integrins, including CD18 in leukocytes, are inhibited by heparin.^30–32 Therefore, we believe that excessive application of anticoagulants should be avoided.

This study has some limitations. We did not use human BM immediately after collection. Therefore, more clinical data are needed, and safety and stability require verification in human clinical studies. As a methodological limitation, this device was developed with a focus on separating therapeutically effective MNCs; therefore, its performance may be insufficient for the separation of other cells.

MNC therapy has substantial potential in regenerative medicine, and there are many potential applications for this device in the future. BM-MNC transplantation is currently undergoing clinical research for a variety of diseases,^5–8 and nonclinical studies have reported its effectiveness in treating dementia,³³ a global social issue of major importance.³⁴ Research into the clinical use of MNCs derived from UCB is also progressing.^35,36 Furthermore, we have discovered the potential for peripheral blood MNCs to be used in new treatments and diagnostics.³⁷ Therefore, the future potential of this device is not limited to the field of cell separation technology, but will likely lead to the resolution of various medical issues and even societal issues.

Conclusion

We developed a completely closed and fully automated device to separate BM-MNCs. Cells separated using this device exerted therapeutic effects in mice with cerebral infarction. This device overcomes factors that prevent the widespread therapeutic application of MNC transplantation, such as cumbersome manual work and the need for CPCs. We believe that our device will substantially contribute to the future clinical application of MNCs.

Footnotes

Acknowledgments

The authors are deeply grateful to the Medical Center for Translational Research, Osaka University Hospital, for their generous advice and the Japanese Red Cross Kinki Cord Blood Bank (Osaka, Japan) and Chubu Umbilical Cord Blood Bank, Inc. (Aichi, Japan) for their generous support with UCB procurement and would like to thank Editage () for English language editing.

Authors’ Contributions

O.S.: Formal analysis, investigation, writing—original draft, writing—review and editing, and visualization. T.S.: Methodology, formal analysis, investigation, writing—original draft, writing—review and editing, and visualization. K.N.: Methodology, investigation, and writing—review and editing. Y.O.: Investigation and writing—review and editing. H.M.: Methodology, investigation, and visualization. S.A.: Methodology and investigation. Y.S.: Investigation. H.K.: Methodology, supervision, and funding acquisition. T.S.M.: Methodology, software, investigation, writing—original draft, writing—review and editing, and visualization. K.K.: Methodology, software, and investigation. K.M.: Methodology, software, supervision, and funding acquisition. A.T.: Conceptualization, writing—review and editing, project administration, and funding acquisition.

Author Disclosure Statement

O.S., T.S., K.N., H.M., S.A., H.K., T.S.M., K.K., K.M., and A.T. have applied for a domestic patent (Japanese Patent Application No. 2024-117408), and O.S., T.S.M., K.K., K.M., and A.T. have applied for an international patent under the Patent Cooperation Treaty (International Application No. PCT/JP2025/025125) to express a closed-cell separation circuit. The other authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding Information

This research was supported by the Japan Agency for Medical Research and Development (AMED) under Grant Number JP23ym0126038 and JP24zf0127010.

Supplementary Material

References

Kikuchi-Taura

, Okinaka

, Takeuchi

, et al. Bone marrow mononuclear cells activate angiogenesis via gap junction-mediated cell-cell interaction. Stroke, 2020; 51(4):1279–1289; doi: 10.1161/STROKEAHA.119.028072

Bayraktar

, Lopez-Pigozzi

, Bortolozzi

. Calcium regulation of connexin hemichannels. Int J Mol Sci, 2024; 25(12):6594; doi: 10.3390/ijms25126594

Ogawa

, Akamatsu

, Fuchizaki

, et al. Gap junction–mediated transport of metabolites between stem cells and vascular endothelial cells. Cell Transplant, 2022; 31:9636897221136151–9636897221136112; doi: 10.1177/09636897221136151

Saino

, Ogawa

, Yasui

, et al. Integrin β2 plays a significant role in therapeutic angiogenesis through hematopoietic stem cell transplantation. Life (Basel), 2025; 15(2):195; doi: 10.3390/life15020195

Savitz

, Misra

, Kasam

, et al. Intravenous autologous bone marrow mononuclear cells for ischemic stroke. Ann Neurol, 2011; 70(1):59–69; doi: 10.1002/ana.22458

Choudry

, Hamshere

, Saunders

, et al. A randomized double-blind control study of early intra-coronary autologous bone marrow cell infusion in acute myocardial infarction: The REGENERATE-AMI clinical trial. Eur Heart J, 2016; 37(3):256–263; doi: 10.1093/eurheartj/ehv493

, Liu

, Duan

. Effects of bone marrow mononuclear cells delivered through a graft vessel in patients with previous myocardial infarction and chronic heart failure: An echocardiographic study of left ventricular dyssynchrony. J Clin Ultrasound, 2018; 46(8):512–518; doi: 10.1002/jcu.22609

Assmus

, Alakmeh

, De Rosa

, et al. Improved outcome with repeated intracoronary injection of bone marrow-derived cells within a registry: Rationale for the randomized outcome trial REPEAT. Eur Heart J, 2016; 37(21):1659–1666; doi: 10.1093/eurheartj/ehv559

, Kusuma

, Driscoll

, et al. Advances in automated cell washing and concentration. Cytotherapy, 2021; 23(9):774–786; doi: 10.1016/j.jcyt.2021.04.003

10.

Traverse

, Henry

, Pepine

, et al. TIME trial: Effect of timing of stem cell delivery following ST-elevation myocardial infarction on the recovery of global and regional left ventricular function: Final 2-year analysis. Circ Res, 2018; 122(3):479–488; doi: 10.1161/CIRCRESAHA.117.311466

11.

Okinaka

, Kikuchi-Taura

, Takeuchi

, et al. Clot-derived contaminants in transplanted bone marrow mononuclear cells impair the therapeutic effect in stroke. Stroke, 2019; 50(10):2883–2891; doi: 10.1161/STROKEAHA.119.026669

12.

Assmus

, Tonn

, Seeger

, et al. Red blood cell contamination of the final cell product impairs the efficacy of autologous bone marrow mononuclear cell therapy. J Am Coll Cardiol, 2010; 55(13):1385–1394; doi: 10.1016/j.jacc.2009.10.059

13.

van Engeland

, Nieland

LJW

, Ramaekers

FCS

, et al. Annexin V-Affinity assay: A review on an apoptosis detection system based on phosphatidylserine exposure. Cytometry, 1998; 31(1):1–9; doi: 10.1002/(SICI)1097-0320(19980101)31:1<1::AID-CYTO1>3.0.CO;2-R

14.

Keeney

, Chin-Yee

, Weir

, et al. Single platform flow cytometric absolute CD34+ cell counts based on the ISHAGE guidelines. International society of hematotherapy and graft engineering. Cytometry, 1998; 34(2):61–70.

15.

Pannu

, Joe

, Iyer

. Performance evaluation of quantiBRITE phycoerythrin beads. Cytometry, 2001; 45(4):250–258; doi: 10.1002/1097-0320(20011201)45:4<250::AID-CYTO10021>3.0.CO;2-T

16.

Kasahara

, Ihara

, Nakagomi

, et al. A highly reproducible model of cerebral ischemia/reperfusion with extended survival in CB-17 mice. Neurosci Res, 2013; 76(3):163–168; doi: 10.1016/j.neures.2013.04.001

17.

Gao

, Xie

, Yu

, et al. Procoagulant activity of erythrocytes and platelets through phosphatidylserine exposure and microparticles release in patients with nephrotic syndrome. Thromb Haemost, 2012; 107(4):681–689; doi: 10.1160/th11-09-0673

18.

Atichartakarn

, Angchaisuksiri

, Aryurachai

, et al. Relationship between hypercoagulable state and erythrocyte phosphatidylserine exposure in splenectomized haemoglobin E/β-thalassaemic patients. Br J Haematol, 2002; 118(3):893–898; doi: 10.1046/j.1365-2141.2002.03711.x

19.

Sorg

, Poppe

, Bunos

, et al. Red blood cell depletion from bone marrow and peripheral blood buffy coat: A comparison of two new and three established technologies. Transfusion (Paris), 2015; 55(6):1275–1282; doi: 10.1111/trf.13001

20.

Mazzanti

, Urbani

, Dal Pozzo

, et al. Fully automated, clinical-grade bone marrow processing: A single-centre experience. Blood Transfus, 2017; 15(6):577–584; doi: 10.2450/2016.0057-16

21.

Fantin

, Olivieri

, Spirito-Daffara

, et al. A comparison of two protocols for optimal red blood cell depletion using Sepax-2 device for ABO-major incompatible transplantation in adults. Curr Res Transl Med, 2019; 67(3):107–111; doi: 10.1016/j.retram.2019.03.003

22.

Taguchi

, Sakai

, Soma

, Doi , et al. Intravenous autologous bone marrow mononuclear cell transplantation for stroke: Phase1/2a clinical trial in a homogeneous group of stroke patients. Stem Cells Dev, 2015; 24(19):2207–2218; doi: 10.1089/scd.2015.0160

23.

Mosca

, Forte

WCN

. Comparative efficiency and impact on the activity of blood neutrophils isolated by percoll, ficoll and spontaneous sedimentation methods. Immunol Invest, 2016; 45(1):29–37; doi: 10.3109/08820139.2015.1085393

24.

Venaille

, Misso

NLA

, Phillips

, et al. Effects of different density gradient separation techniques on neutrophil function. Scand J Clin Lab Invest, 1994; 54(5):385–391; doi: 10.3109/00365519409088438

25.

Lub

, Van Kooyk

, Figdor

. Competition between lymphocyte function-associated antigen 1 (CD11a/CD18) and Mac-1 (CD11b/CD18) for binding to intercellular adhesion molecule-1 (CD54). J Leukoc Biol, 1996; 59(5):648–655; doi: 10.1002/jlb.59.5.648

26.

Lunde

, Solheim

, Aakhus

, et al.; ASTAMI investigators. Autologous stem cell transplantation in acute myocardial infarction: The ASTAMI randomized controlled trial. Intracoronary transplantation of autologous mononuclear bone marrow cells, study design and safety aspects. Scand Cardiovasc J, 2005; 39(3):150–158; doi: 10.1080/14017430510009131

27.

Seeger

, Tonn

, Krzossok

, et al. Cell isolation procedures matter: A comparison of different isolation protocols of bone marrow mononuclear cells used for cell therapy in patients with acute myocardial infarction. Eur Heart J, 2007; 28(6):766–772; doi: 10.1093/eurheartj/ehl509

28.

Wang

H-X

, Li

Z-Y

, Guo

Z-K

, Guo

Z-K

. Easily-handled method to isolate mesenchymal stem cells from coagulated human bone marrow samples. World J Stem Cells, 2015; 7(8):1137–1144; doi: 10.4252/wjsc.v7.i8.1137

29.

Salamanna

, Contartese

, Nicoli Aldini

, et al. Bone marrow aspirate clot: A technical complication or a smart approach for musculoskeletal tissue regeneration? J Cell Physiol, 2018; 233(4):2723–2732; doi: 10.1002/jcp.26065

30.

Peter

, Schwarz

, Conradt

, et al. Heparin inhibits ligand binding to the leukocyte integrin Mac-1 (CD11b/CD18). Circulation, 1999; 100(14):1533–1539; doi: 10.1161/01.CIR.100.14.1533

31.

Diamond

, Alon

, Parkos

, et al. Heparin is an adhesive ligand for the leukocyte integrin Mac-1 (CD11b/CD18). J Cell Biol, 1995; 130(6):1473–1482; doi: 10.1083/jcb.130.6.1473

32.

Lever

, Hoult

JRS

, Page

. The effects of heparin and related molecules upon the adhesion of human polymorphonuclear leucocytes to vascular endothelium in vitro. Br J Pharmacol, 2000; 129(3):533–540; doi: 10.1038/sj.bjp.0703099

33.

Kanamaru

, Kamimura

, Yokota

, et al. Intravenous transplantation of bone marrow-derived mononuclear cells prevents memory impairment in transgenic mouse models of Alzheimer’s disease. Brain Res, 2015; 1605(1):49–58; doi: 10.1016/j.brainres.2015.02.011

34.

Twiss

, McPherson

, Weaver

. Global diseases deserve global solutions: Alzheimer’s disease. Neurol Int, 2025; 17(6):92; doi: 10.3390/neurolint17060092

35.

Zarrabi

, Akbari

, Amanat

, et al. The safety and efficacy of umbilical cord blood mononuclear cells in individuals with spastic cerebral palsy: A randomized double-blind sham-controlled clinical trial. BMC Neurol, 2022; 22(1):123; doi: 10.1186/s12883-022-02636-y

36.

Zhu

, Poon

, Liu

, et al. Phase I–II clinical trial assessing safety and efficacy of umbilical cord blood mononuclear cell transplant therapy of chronic complete spinal cord injury. Cell Transplant, 2016; 25(11):1925–1943; doi: 10.3727/096368916X691411

37.

Okinaka

, Shinagawa

, Claussen

, et al. RNA analysis of circulating leukocytes in patients with Alzheimer’s disease. J Alzheimers Dis, 2024; 97(4):1673–1683; doi: 10.3233/JAD-230874

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.46 MB

1.10 MB

Performance of Developed Mononuclear Cell Separator for Regenerative Therapy Is Comparable to Manual

Abstract

Impact Statement

Introduction

Methods

Experiment

Experimental design

Ethics declaration

Cell preparation

Manual separation and stock of MNCs

Automated MNC separation

Flow cytometry

Comparison of separation performance according to injection port shape

Validation of separated cell viability

Uptake of VEGF by HUVECs

Stroke model and cell transplantation

Immunohistochemistry

Open-field test

Statistics

Experimental Results

Comparison of separation performance between containers with different inlet shapes

Comparison of RBC and annexin V + RBC removal rates between manual and automated device

Comparison of separation performance between manual and automated separation

Viability of cells separated by automated device

Amount of VEGF uptaken by HUVECs cocultured with MNCs

Expression of CD18 in separated cells

Microglia/macrophage accumulation and microangiogenesis in transplanted MNCs

Neurological function of mouse stroke models after cell transplantation

Discussion

Conclusion

Footnotes

Acknowledgments

Authors’ Contributions

Author Disclosure Statement

Funding Information

Supplementary Material

References

Supplementary Material

Comparison of RBC and annexin V ⁺ RBC removal rates between manual and automated device