Abnormal production or loss of red blood cells (RBCs) can cause life-threatening anemia and thrombocytopenia. Although blood transfusion is an essential treatment for these patients, inadequate blood supplies and transfusion-related adverse reactions pose potential challenges to the global healthcare system [1, 2]. In vitro generation of RBCs has become an important focus to overcome these problems and numerous attempts have been made using different cell types; including human adult hematopoietic stem cells (HSCs), cord blood (CB), bone marrow (BM), peripheral blood (PB), and pluripotent stem cells. Recently, induced pluripotent stem cells (iPSCs) and erythroblast progenitor cell lines have attracted attention owing to their unlimited expansion potential [3, 4].

Since the initial era of 2D culture systems for producing blood cells on a small scale in the laboratory, the use of modified culture systems has been reported. The current consensus is a 3-step method for liquid culture of erythroid cells (Table 1) [5]. Briefly, HSCs are expanded using erythropoietin (EPO), stem cell factor (SCF), and interleukin-3 (IL-3) in a base medium; Iscove’s modified Dulbecco’s medium (IMDM) or StemSpan serum-free expansion medium (SFEM, Stem Cell Technologies, Vancouver, BC, Canada). Erythroblast expansion in the presence of SCF, EPO, and transferrin is followed by terminal differentiation, with or without EPO. While the expansion of erythroblasts requires supplementation with SCF and EPO, cells at the terminal differentiation stage gradually lose receptors for these cytokines [6]. Since the enucleated reticulocytes still have ferritin receptors; transferrin, which provides iron to erythroid cells, should be supplemented until the end of the final maturation.

Table 1 . Overview of sources and culture media for the production of red blood cells..

	HSC commitment
Cell cultures	Erythropoiesis
Cell sources	CB, BM, PB, hESC, hiPSC, immortalized erythroid progenitor cell lines
Media	StemSpan, IMDM, Cell-Quina)
Additives	EPO, TPO, SCF, IL-3, transferrin, holo-transferrin, IL-6, Flt-3, heparin, IGF-1, glucocorticoids, TGF-β agonist, PPAR-α agonist

^a)GMP-grade homemade media..

Abbreviations: BM, bone marrow; CB, cord blood; EPO, erythropoietin; Flt-3, feline Mcdonough sarcoma-like tyrosine kinase 3; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell; IGF-1, insulin-like growth factor-1; IMDM, Isocove’s modified Dulbecco’s medium; PB, peripheral blood; PPAR-α, peroxisome proliferator-activated receptor α; SCF, stem cell factor; IL, interleukin; TGF-β, transforming growth factor β; TPO, thrombopoietin..

Enucleation is a crucial factor because nucleated erythroblasts are less effective at transporting oxygen and are more prone to hemolysis as they pass through small capillaries. Enucleated cells also have the advantage of no DNA and the inability to divide, thus preventing the transfer of malignancy to the recipient [7].

Additional additives, such as thrombopoietin, IL-6, feline McDonough sarcoma-like tyrosine kinase 3, heparin, glucocorticoids, insulin-like growth factor-1, transforming growth factor β agonists, and peroxisome proliferator-activated receptor α agonists, can be added and further modified; with or without serum or feeder layers [8-10].

A major concern is the scale of production required to meet the standard adult therapeutic transfusion dose. Because more than 2×10¹² RBCs are required per unit, a conventional 2D culture system requires 1,000 L of medium with a culture density of 2×10⁶ cells/mL [11]. High-density cell culture is critical for overcoming the massive volume of media required [5]. Cells grown in suspension should be maintained at a density of >1×10⁸ cells/mL to produce the numbers required within a minimal medium volume [12]. In addition, greater than 90% enucleation is important for better final erythrocyte yield [10]. Final erythrocytes should have adult hemoglobin and express accurate blood group antigens, such as O RhD-negative, for the universal blood type.

In this review, we discuss recent advances in the manufacturing of RBCs using various cell sources, 3D manufacturing methods, and clinical applications.

The first studies on cultured RBCs (cRBCs) focused primarily on human adult HSCs derived from CB, BM, and mobilized PB [13-15]. However, because of their limited proliferation capacity and low yield, RBCs derived from primary stem cells are difficult to use in clinical settings. An emerging model source uses the high proliferative potential of embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs). However, these cells show very low degrees of enucleation and very low adult hemoglobin expression. Recent studies have focused on the development of immortalized cell lines at the adult erythroblast stage.

Erythropoiesis by mimicking 3D BM niches

Complex 3D culture systems have been studied to mimic the BM niche which regulates the generation of blood cells with a focus on topography, local stiffness of physical constraints, soluble factors, small molecules, and feeder layers.

Severn et al. [16] inoculated PB CD34⁺ cells into porous polyurethane scaffolds. These scaffolds exhibited significantly higher total cell egress of hematopoietic progenitor cells (HPCs) than human bone scaffolds (3.90×10⁶ vs. 2.45×10⁶ cells in 1.5 mL of medium). Furthermore, HPCs were released for up to 28 days, achieving a high cell density of 3.7×10⁹ cells/mL. These erythroid progenitors were then cultured with proliferative polymerized high internal phase emulsion (polyHIPE) scaffolds similar to the honeycomb-like structure of the human BM. Cells in this study proliferated 5.18×10³-fold from the starting material and differentiated into 2.59×10⁹ erythroid cells in a spinner flask filled with 1.6 L medium [17].

Lee et al. [18] cultured erythroid cells in a 400 µm diameter microporous microcarrier. This increased cell density to 1×10⁷ cells/mL and resulted in a yield of 80% enucleated RBCs. However, when cultured with a microcarrier with a pore size of 30 µm, which is too narrow for erythroid cells to contact each other, the cell viability was too low (8.5%).

Bioreactors for RBC manufacturing

It is unclear which type of bioreactor is the best for RBC production. Culturing at a high cell density is inevitable because of the large quantity of cells required in the final product for transfusion. Perfusion-based hollow-fiber bioreactors are known to increase the cell culture density. In a small scale study using a 2 mL vessel, CB-derived CD34⁺ cells inoculated at a density of 800,000 cells/mL expanded 100-fold and differentiated into erythroid cells over a 7 day period [19]. Additionally, Allenby et al. [20] developed a hollow fiber bioreactor and seeded a culture of CB mononuclear cells (MNCs) at a density of 2×10⁷ cells/mL. After 28 days of culture, erythroid cells reached BM-like cell concentrations of more than 2×10⁹ cells/mL. The bioreactor was later modified using a spatiotemporal microenvironmental model that characterized the spatiotemporal distribution of cells, metabolites, and growth factors to guide the design of an improved version. The modified reactor generated a 50-fold increase in the RBC production efficiency when using a 5-fold lower concentration of growth factors than the previous model [21]. However, despite high-density cell culture using a 3D model mimicking BM niches, the limitation of inefficient oxygen and nutrient distribution need to be overcome for scale-up [22].

Wave-like rocking motion bioreactors, fluidized bed bioreactors [23, 24], rotating wall vessel bioreactors, and stirred tank reactors (STRs) have been proposed for the mass transfer of nutrients and oxygen to achieve greater erythroblast expansion and enucleation.

Since the effects of agitation on in vitro erythropoiesis in 24-well plates were studied in 2010 [25], the first 1 L culture in a wave-type bioreactor was demonstrated [26]. When CB-derived CD34⁺ cells were cultured in a wave-type CultiBag RM bioreactor system (Sartorius, Göttingen, Germany) with 50% dissolved oxygen and a fully defined culture medium, a 2.25×10⁸-fold proliferation was obtained; although this was a calculated rather than actual number of cultured cells.

The transfer of static culture to a flask-based system has been studied to scale up culture conditions. Griffiths et al. [27] reported that PB CD34⁺ cells in static culture flasks at a density of 2–10×10⁵ cells/mL within 500 mL were transferred to a 2 L stirred vessel at a density of 1–6×10⁵ cells/mL and cultured at 15 rpm. In this system, overall cell expansion was ≥10⁴ -fold with enucleation rates ranging from 55–95%, corresponding to 5 mL of packed RBCs. The procedure was then scaled up to yield 10 mL of packed RBCs using a good manufacturing practice (GMP)-compliant procedure with CD34⁺ cells from both CB and PB. Furthermore, they demonstrated the feasibility of large-scale culture in 1.5 L spinner flasks using Bristol Erythroid Line Adult (BEL-A) cell lines [4, 28].

Sivalingam et al. [12] demonstrated the scalability of hiPSC microcarrier aggregates on 6-well plates in 500 mL spinner flasks. In a 50-mL culture, the cells were cultured at a high density of 1.7×10⁷ cells/mL for erythroid differentiation. Recently, the same group established O-negative hiPSCs in combination with a perfusion bioreactor system capable of high-density cultures of 3.47×10⁷ cells/mL of erythroid cells [29].

Zhang et al. [30] showed that 1×10⁶ CB CD34⁺ cells cultured in roller bottles produced up to 2.9×10¹¹ total cells with a 50% enucleation rate. The amplification of total cells, including unenucleated cells, reached a plateau of approximately 2×10⁸-fold by day 21. If all of these cells differentiated into enucleated erythrocytes, the yield of erythrocytes from 1 CB unit (5 million CD34⁺ cells) could theoretically be equivalent to 500 blood transfusion units; a single unit of transfusable blood contains 2×10¹² RBCs. Heshusius et al. [10] expanded and differentiated CD34⁺ PB MNC-derived erythroid cells in 1 L gas permeable G-Rex bioreactors with a GMP grade medium. In this system, 3×10⁷-fold erythroblast expansion was achieved, yielding 24 mL of packed RBCs and more than 90% enucleation; adult hemoglobin expression and correct blood group antigen expression were identified.

An STR is a cylindrical vessel with a stirrer that is widely used to culture biological agents, including enzymes, antibodies, and cells. It has been widely used in industry because of its relatively easy scale-up, uniform distribution of cells and nutrients, sufficient oxygen transfer, ease of control and monitoring of culture conditions, and compliance with current GMP requirements [31, 32].

The Thomas group [33, 34] used 15 mL Ambr bioreactors using a design of experiments (DOE) methodology. However, the final cell yield was as low as 16-fold after 25 days and 8-fold after 16 days. This group later refined the experimental design for bioreactor process development, gas transfer parameters, and media volume. However, proliferation in the STR was lower than that in static cultures, with 12-fold growth in the STR and 14-fold growth in static cultures being achieved in 25 culture days. Differentiation and enucleation, in addition to a reduction in culture time and media volume, were enhanced in STR compared to static conditions [35]. Lee et al. [36] cultured an erythroblast cell line (ImEry) expressing c-MYC and BCL-XL genes in a 500 mL bioreactor using DOE principles to develop optimized media formulations.

Han et al. [37] recently identified the optimal process parameters for differentiating CB-derived erythroid cells in 10–11 mL bioreactors using DOE. Initially, the cells were inoculated into flasks and transferred to the STR. Up to 18 vessels were simultaneously operated to provide reproducible results and the optimal cell seeding density (5×10⁶ cells/mL), seeding timing (cell diameter 12–13 μm), pH (7.5), temperature (37°C), agitation speed (500 rpm), medium feeding regimen, and dissolved oxygen were determined. Recently, Gallego-Murillo et al. [38] used a single-use STR for expansion and differentiation of PBMC-derived erythroid cells. They scaled up from static cultures to 0.5 L, followed by 3 L cultures with optimal stirring speeds, aeration strategies, and GMP-grade homemade medium (Cellquin). However, further studies are needed to obtain the larger number of cRBCs required for even a single transfusion unit (Table 2).

Table 2 . 3D scaffolds and bioreactors for the production of red blood cells..

Types	Cell source	Culture period	Fold increase in total cells	Enucleation rate	Maximal cell density	Reference
Porous scaffold (polyHIPEs)	PB CD34+ cells	14 days of expansion (in scaffolds), 12 days of differentiation (in spinner flasks)	5,180- fold [(2.59±0.212) ×109 total cells]	16.25% (post- leuko-filtration)	1.6×106 cells/mL (in spinner flasks)	Severn et al. [17] 2019
Microcarrier (cytoline I)	Late erythroblasts differentiated CB CD34+ cells	1–3 days after culturing in 2D plates for 13–15 days	2,200-fold in 2D culture	80%	1×107 cells/mL	Lee et al. [18] 2015
Hollow fiber (BR001)	CB CD34+ cells	7 days	100-fold (small-scale of 2 mL or 8 mL)	40%	8×107 cells/mL	Housler et al. [19] 2012
Hollow fiber (3DHFR)	CB MNCs	28 days	4.4-fold (550-fold from the stimulated erythroid progenitors)	23%	2×109 cells/mL	Allenby et al. [20] 2019
Hollow fiber (BR2)	CB MNCs	28 days	50-fold	50%	2.5×109 cells/mL	Allenby et al. [21] 2022
Wave-type (CultiBag RM bioreactor)	CB CD34+ cells	21 days	1.73×106-fold	Not reported	1×105 cells/mL	Timmins et al. [26] 2011
2 L glass vessels (stirred)	PB CD34+cells	18–24 days	104-fold	55–95%	1–6×105 cells/mL	Griffiths et al. [27] 2012
1.5 L flasks (stirred)	PB and CB CD34+ cells	21 days	105-fold	PB: ≤60%	1–4×106 cells/mL from day14	Kupzig et al. [28] 2017
				CB: ≤38%
500 mL spinner flasks	hiPSC	39 days	206–805-fold	18.1–59.3%	1.7×107 cells/mL	Sivalingam et al. [12] 2021
Stirred-tank Applikon BioSep perfusion bioreactor	O-negative hiPSC	29 days	1510.7-fold	≤30%	3.47×107 cells/mL	Yu et al. [29] 2022
Bottle turning device culture system	CB CD34+cells	21 days	2×108-fold	50%	2.42×106 cells/mL	Zhang et al. [30] 2017
G-Rex bioreactor	PB MNCs	25 days	3×107-fold	≥90%	5–10×106 cells/mL	Heshusius et al. [10] 2019
Stirred micro-bioreactor (Ambr)	CB CD34+ cells	25 days	12-fold	80%	1–5×106 cells/mL	Bayley et al. [35] 2018
Shake flask	PB MNCs	10 days	13.8-fold	Not reported	3.06×106 cells/mL	Lee et al. [36] 2018
Stirred bioreactor (Ambr)	CB CD34+ cells	21 days (cultured in 2D plates for 13 days)	2.25×104-fold	50%	1.5×107 cells/mL	Han et al. [37] 2021
Stirred bioreactor (Single wall or AppliFlex)	PB MNCs	22 days	750-fold (0.5 L), 196-fold (3 L)	30–35%	0.7–2×106 cells/mL	Gallego-Murillo et al. [38] 2022

The first human transfusion of cRBCs was reported in 2011. HSCs were isolated from autologous mobilized PB, and 61,500±7,600-fold expansion was achieved by co-culture with mesenchymal stem cells. Subsequently, 10¹⁰ cRBCs were transfused and the half-life of the infused cells was approximately 26 days, demonstrating the safety of the cRBCs [39]. In 2022, a randomized and controlled phase I crossover trial titled RESTORE was announced to recruit at least ten healthy volunteers. Approximately 5–10 mL of lab-grown RBCs from allogeneic donors were transfused into 2 healthy individuals and no side effects have been reported. The details of current clinical trials of cultured red blood cells are shown in Table 3; ongoing studies will test the lifespan of lab-grown cells in the body and the donor’s fresh blood.

Table 3 . Clinical trials of cultured red blood cells..

Types	cRBC
Cell origin	Autologous PB stem cells	Allogeneic PB stem cells
Infused cells	RBC	RBC
Recipient	Healthy volunteer	Healthy volunteers
Enrollment no.	1	2
Identifiers	NCT00929266	ISRCTN:42886452
		EudraCT:2017-002178-38
Location	France	UK

Significant progress has been made in the generation of in vitro RBCs derived from immortalized cell lines and iPSCs and the development of the manufacturing process. Although product safety should be addressed further, pioneering clinical trials on cRBCs could herald a new era of transfusion medicine.

No potential conflicts of interest relevant to this article were reported.