In 1816, John Henry Leacock, a medical student in Edinburgh from Barbados, carried out the first systematic experiments on intra-species blood transfusion. He demonstrated that exsanguinated dogs could be resuscitated with a transfusion from another dog but not from a cat, thus establishing the requirement for a proper understanding of the nature of the product, a clear medical rationale and compatibility between donor and recipient. Leacock did not further develop his research, but it was picked up by a fellow medical student, James Blundell, who went on to carry out the first human–human transfusions in Guys Hospital in London [1].

Today, around 92 million red cell transfusions are administered per annum worldwide for a wide variety of clinical conditions. However, significant limitations persist, particularly when looked at from a global perspective, including sufficiency of supply, immunological incompatibility and the risk of the transfusion of transmitted infection.

For many decades, researchers have sought pharmaceutical alternatives to red cell transfusion, including the development of synthetic oxygen-carrying molecules and hemoglobin-containing solutions, but with limited success. This is perhaps unsurprising in light of the complex way in which red cells interact with the microcirculation in the regulation of tissue oxygenation and acid–base balance. The hematological system was the field within which the existence of stem cells was first postulated by Papenheim [2], demonstrated experimentally by McCulloch and Till [3], and clinically applied as bone marrow transplantation by E Donnall Thomas et al. [4]. Hence the generation of red cells in vitro is an attractive proposition in light of our scientific understanding of hematopoiesis, experience of this type of product and well-established clinical indications. However, the quest to achieve scalable in vitro red cell production is anything but simple and exemplifies the challenges in stem cell and developmental biology, analytic and process development and bioengineering characteristic of the next generation of cellular therapeutics.

Hematopoietic ontogeny & the importance of microenvironment

In humans, primitive hematopoiesis arises in the extra-embryonic mesoderm in the yolk sac at around day 21 of gestation, resulting in the production of megaloblastic nucleated red cells expressing embryonic hemoglobins (Gower 1, Gower 2 and Portland), which have higher oxygen affinity than fetal or adult hemoglobins and display less co-operativity and pH sensitivity. Specialized embryonic hematopoietic progenitor cells (HPC) arise in the yolk sac shortly thereafter [5]. Definitive hematopoietic stem cells (HSCs) arise independently from the intra-embryonic mesoderm, specifically from the hemogenic endothelium in the ventral floor of the dorsal aorta (and other arterial vessels) at around day 27 and differentiate into normoblastic enucleated erythrocytes expressing fetal hemoglobin [6,7]. From there they migrate to the embryonic liver at approximately day 32 where they massively expand and become the main source of fetal definitive hematopoiesis from around 6 weeks onwards. From around 20 weeks, hematopoietic stem and progenitor cells (HSPCs) migrate to the bone marrow, which is the source of adult definitive hematopoiesis following birth. Thus hematopoietic ontogeny involves a complex temporal and spatial interplay between HSPCs of differing origin and potentiality, and a variety of different microenvironments. This complexity needs to be properly understood and managed in order to achieve successful in vitro differentiation of erythrocytes from stem cell sources [Figure 1].

Erythroid differentiation from hematopoietic stem & progenitor cells

In the adult human, HSCs are thought to comprise around 1 out of 105 of the nucleated cells in the bone marrow and are capable of self-renewal and differentiation into the multiple different mature cell types of the adult hematopoietic and immune systems. They reside adjacent to the endosteal surface of the bone and their progeny migrate centripetally during differentiation. Erythoid precursors from the proerythroblast onwards are found in association with macrophages in erythroblastic islands close to the central venules. During enucleation the macrophages phagocytose the extruded nuclei and organelles, whilst the resulting reticulocyte undergoes substantive membrane remodeling as it is released into the circulation.

In clinical practice, CD34 is used as a surrogate marker of HSC potential, though in reality more than 99% of this phenotypically defined population are mature HPCs and endothelial cells. HPCs can be cultured short term in vitro in semi-solid matrices within which each will develop into a discrete colony as a result of a limited number of cell divisions followed by terminal differentiation. This system is, however, a one-stage continuous culture with limited flexibility and scope for expansion.

These limitations can be circumvented in suspension culture systems, which can be adjusted to allow sequential support of proliferation and differentiation. Over the past 20 years, a number of groups have developed a variety of suspension culture systems that use CD34⁺ enriched populations derived from bone marrow, granulocyte colony-stimulating factor mobilized peripheral blood or umbilical cord blood as a starting material.

The most simple of these used single-phase liquid cultures and a variety of cytokines in various combinations including interleukin-3 (IL-3), granulocyte-monocyte colony stimulating factor (GM-CSF) and erythopoietin (EPO) [8]. After 18–21 days in culture, erythroblastic islands developed followed by enucleation of 10–40% of cells. The enucleated cells had the characteristics of reticulocytes and expressed gamma globins (cord blood) or beta globin (adult bone marrow and peripheral blood). A higher level of expansion was reported by Panzenbock et al. who cultured CD34⁺ cells in EPO, stem cell factor (SCF), insulin-like growth factor 1 (IGF-1), dexamethasone and estradiol, resulting in a 105-fold expansion in erythroid progenitors by days 15–18 [9]. Erythroid progenitors were recovered and cultured further in EPO and insulin in order to induce terminal differentiation to erythrocytes. Similarly, Freyssinier et al. cultured CD34⁺ cells for 7 days in serum-free conditions with SCF, IL-3 and IL-6, resulting in a 30-fold expansion in cell numbers and the appearance of high numbers of CD36⁺ cells [10]. The CD36⁺ cells were purified by immunomagnetic separation and further cultured in the same medium and cytokines with the addition of EPO for an additional 3 days, leading to a 100-fold cell expansion and maturation to glycophorin A-positive erythroblasts and a small proportion of enucleated red cells. Unlike in Malik et al’s [8] study, enucleation did not seem to be dependent on the presence of macrophages in this system. Similarly, Miharada et al. developed a method to produce enucleated erythrocytes using vascular endothelial growth factor (VEGF), IGF-II and mifepristone (a glucocorticoid antagonist) but did not make use of feeder cells or macrophages [11]. Maggakis-Kelemen et al. [12]also used a similar culture system to that of Malik et al. [8], but with the addition of dimethylsulphoxide (DMSO), ferrous citrate and transferrin. They found that the latter dramatically enhanced hemoglobin synthesis but that whilst precursor cells showed a high degree of deformability, the cultured reticulocytes and erythrocytes showed reduced shear modulus compared to controls. Fujimi et al. [13] showed that co-culture with macrophages significantly enhanced the production of red cells from CD34⁺ HSPCs.

The most extensive functional analysis of in vitro generated erythroid cells has been carried out by Douay and colleagues who developed a protocol for differentiation of cord blood CD34⁺ cells in defined media, using a 3-step protocol involving stimulation of CD34⁺ HSPCs using SCF, Flt3-ligand and thrombopoietin (TPO), followed by expansion of erythroid progenitors using SCF, EPO and IGF-1; and terminal erythroid differentiation using EPO and IGF-1 [14,15]. The system produced 2 x 10⁵-fold amplification with dominant erythroid differentiation. This group have subsequently modified their protocol for application to adult blood and bone marrow [16,17] substituting IL-3 for Flt3-ligand in the first phase, culture with EPO in the presence of a murine stromal cell line (MS5) or human mesenchymal cells for 3 days; followed by the stromal cells alone without growth factors for up to a further 10 days [18,19]. Commitment towards the erythroid lineage was morphologically apparent from day 8, with terminal differentiation to 65–80% reticulocytes by day 15 and maturation to erythrocytes from days 15 to 18 evidenced by loss of transferrin receptor (CD71) expression, 90–100% enucleation and typical morphological characteristics. These cells also showed normal glucose 6 phosphate dehydrogenase (G6PD) and pyruvate kinase (PK) enzyme levels, membrane deformability and oxygen dissociation characteristics and similar survival to native red cells following carboxyfluorescein diacetate succinimidyl ester (CFSE) labeling and intraperitoneal infusion into non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice. Douay and colleagues have used a Good Manufacturing Process (GMP) version of this protocol, omitting the co-culture step, to manufacture 10¹⁰ cultured erythroblasts from autologous peripheral blood CD34⁺ cells that were 51Cr labeled and reinfused into the human volunteer. The level of the cells in the circulation after 26 days was comparable to that seen with peripheral blood red cells, supporting their overall quality, safety and compatibility, and amounting to clinical proof of principle [20]

Anstee and his colleagues have developed a simpler, feeder-free two-stage culture system using enriched CD34⁺ cells cultured for 5 days in serum-free medium supplemented with SCF, IL-3 and EPO, lipoprotein and tacrolimus, and then transferred to medium supplemented with human AB serum, transferrin, EPO, insulin and T3 [21,22]. This system produces a 10⁴-fold expansion in cell numbers with 55–95% enucleation. Leucocyte filtration was used to remove normoblasts, free nuclei and R1 reticulocytes leaving approximately 3 x 10¹⁰ R2 reticulocytes and erythrocytes, which displayed normal morphology, protein expression and glycosylation, membrane deformability and oxygen dissociation and superior recovery and survival in immunodeficient mice. A larger scale clinical recovery and survival study using a GMP version of this protocol is due to commence in 2017.

Overall, although it is clear that bone marrow, cord blood and peripheral blood can be used as sources of HSPC to generate sufficient functional red cells for a small number of units, large-scale manufacture is not likely to be possible because of the limited replication capacity of HSCs and the remaining need for donation of this starting material.

Erythroid differentiation from pluripotent stem cells

Even though long-term repopulating HSCs have not yet been successfully generated from human pluripotent stem cells (PSCs), several groups have demonstrated differentiation of downstram hematopoietic cells [23,24] using two different approaches.

The first approach entails co-culturing PSCs with murine stromal lines, such as S17 or OP9 [25], or the yolk sac endothelial cell line C166 [26] to promote or support the commitment of pluripotent PSCs specifically to the hematopoietic lineage. Vodyanik et al. [25] showed that OP9 stroma supports the production of up to 10 times more HSPCs compared with other murine bone marrow stromal cell lines. This group demonstrated production of up to 20% CD34⁺ cells and isolated up to 10⁷ CD34⁺ cells from a similar number of hPSCs. In an attempt to recapitulate the microenvironment of developing HSCs in the embryo, Ledran et al. showed that co-culture of hPSC with monolayers of stromal cells derived from murine AGM, fetal liver or other developmental niches enhanced early differentiation of CD34⁺ HSPCs [27]. These HSPCs were capable of primary and secondary engraftment into immunodeficient mice at low levels and also generated both erythroid and myeloid cells in colony-forming assays. Interestingly, the AGM-derived AM201B4 line resulted in the best reconstituting capacity suggesting, perhaps logically, that PSCs may be better supported by stroma from the sites where definitive hematopoiesis first occurs, rather than sites such as the bone marrow, which is more active in the adult.

The Bouhassira group have shown that hESC culture on hTERT-transformed human fetal liver cell lines can generate 2–10% CD34⁺ cells, which can then be differentiated into an erythroid population using a 4-stage liquid culture protocol [28,29]. Using this system, they further reported that increasing the duration of culture on the FH-hTERT cell line from 14 to 21 or 35 days lead to a 5000-fold increase in cell number of erythroid cells with differentiation of the predominant population of megaloblastic nucleated erythroblasts to a small population of normoblastic enucleated red cells with an increase in alpha/zeta and beta/epsilon chain ratios recapitulating in vitro the shift from primitive to definitive hematopoiesis [29-31].

The second approach entails the removal of PSCs from feeder cells and culture in suspension, which leads to the formation of embryoid bodies (EBs). Chadwick et al. demonstrated that bone morphogenetic protein 4 (BMP4), a ventral mesoderm inducer, strongly promotes hematopoietic differentiation in the context of a combination of hematopoietic cytokines (SCF, Flt-3 ligand, IL-3, IL-6 and G-CSF) [32]. It has also been demonstrated that addition of VEGF-A165, which is normally produced in response to hypoxia, further promotes erythropoiesis, particularly when augmented by EPO [33]. Chang et al. [34] have described the generation of erythroid cells from PSCs by subsequent processing of cells present at either early or late stages of EB formation, whereas Zambidis et al. [35] have more closely defined the progress of differentiation in EB through primitive and definitive stages of hematopoiesis. Much of the seminal work characterizing hematopoiesis in EB has come from the Keller group who have demonstrated that at 3–4 days, EB differentiation give rise to cell colonies expressing both endothelial and hematopoietic markers (KDR [FLK1] and CD117) [36,37]. Other groups have also shown the appearance of CD34⁺ cells from 5 days of culture, peaking at day 12–15, with evidence of primitive erythropoiesis between 7 and 12 days and definitive erythropoiesis emerging from days 12– 20 [32,33,35].

One of the more clinically relevant published studies is that of Lu et al., who have demonstrated the differentiation of hESCs into functional erythrocytes on a large scale using a 4-step procedure [38]. The first step involved the formation of EBs by culturing PSCs in serum-free media, BMP4, VEGF and bFGF with the addition of SCF, TPO and FLT-3 ligand after 48 h. The second involved dissociation of the EBs with trypsin and culture of the single-cell suspension in blast colony growth medium, bFGF and recombinant tPTD-HOXB4 fusion protein for 10 d. The third step consisted of further culture with EPO for 5 days and thereafter with SCF, EPO and methylcellulose. Finally, the erythroid cells were enriched through a plastic adherence step. Starting with one 6-well plate of hESC, they generated 10¹¹–10¹² nucleated erythroid cells with comparable oxygen dissociation curves, response to pH (Bohr effect) and 2,3-diphosphoglycerate (DPG) activity to normal adult red cells, despite expressing mainly embryonic and fetal globins. However, after 28 days of culture only 16% of the cells expressed beta globin and 10–30% of the cells became enucleated. Also, in order to achieve this impressive expansion in cell numbers during differentiation, it was necessary to treat the cells with a recombinant HOXB4 protein.

These two methods of inducing hematopoietic differentiation have also been combined; Wang et al. demonstrated that the treatment of PSCs during EB formation with a combination of low-dose cytokines and human bone marrow stroma significantly increased hematopoietic differentiation [39]. In a very promising study, Ma et al. demonstrated that PSC co-cultured with murine fetal liver-derived stromal cells after an initial EB stage, produced large numbers of erythroid progenitors that underwent definitive differentiation with a shift from epsilon-globin expression to beta-globin expression and production of enucleated erythrocytes. Comparative analysis against cord blood-derived progenitors showed similar expression of glycophorin A, CD71 and CD81, G6PD activity and oxygen-carrying capacity [40]. However, they were only able to achieve a 100-fold expansion from PSCs to mature erythrocytes.

Most recently, Jaichandran and colleagues [41] have reported a serum-free chemically defined microcarrier-based suspension culture based on the differentiation protocol developed by Lu et al. resulting in a 6-fold improvement in HPC expansion and an 80-fold improvement in erythroid cell generation compared with their standard EB method. They report efficient terminal maturation and production of up to 50% mature enucleated erythrocytes using co-culture with mesenchymal stromal cells.

We feel that the key to improving the process and making it possible to efficiently achieve hematopoiesis and erythroid maturation at scale and in a manner that is current GMP-compliant, is to identify and replace the array of signals delivered by stroma with a defined animal- and feeder-free culture system. We have developed such a system using combinations of cytokines and small molecules [41]. To induce differentiation, hPSCs are cultured in BMP4, VEGF, Activin A, Wnt3A, and GSKIIIbeta inhibitor VIII or A-A014418 (Mix A). On day 2 the existing set of cytokines and small molecules is augmented and FGF, SCF and beta-estradiol are added. On day 3 the EBs are dissociated and re-suspended in medium containing BMP4, VEGF, FGFa, SCF, IGF, TPO, heparin, IBMX and beta-estradiol (Mix B). On day 5 this medium was refreshed and StemRegenin added. On day 10 the cells are replated in erythroid liquid culture conditions with hydrocortisone, SCF, Flt3-ligand, BMP4, IL3, IL11, IBMX and EPO (Mix C). On day 17 the cells are cultured in medium containing BSA, insulin, transferrin, lipid mixture, hydrocortisone, SCF, IGF, IL3, IL11 and EPO (Mix D). From day 24 the cells are suspended in medium containing EPO and from day 26 in medium alone [42]. This protocol supports 50,000–200,000-fold expansion in cell numbers by day 31 of which 99% are CD235a-positive with 90–95% fetal hemoglobin. The approach has been demonstrated to be reproducible across a number of embryonic and induced PSC lines. However, enucleation of these cells appears to be unstable giving rise to abnormal R1 reticulocytes, which fragment easily leaving only around 10% stable reticulocytes. The cause of this is unclear, but may represent the PSCs adopting a fate similar to that of the transient but definitive erythroid/myeloid precursor population reported by the Palis laboratory [5]

.

Erythroid cell lines

In light of the challenges in achieving definitive adult erythropoiesis and stable reticulocytes through PSC differentiation, several groups have explored the possibility of generating immortalized erythroid cell lines (ECL). Hirose et al. described the immortalization of erythroblasts from hiPSC, using doxycycline (DOX)-inducible overexpression of BCL-XL and c-MYC to establish a highly proliferative population of immortalized erythrocyte progenitor cells that underwent differentiation on removal of DOX [43]. Neither gene alone was sufficient to cause the immortalization, and survival and proliferation of the clonal imERYPC lines in the presence of DOX was dependent only on EPO. After removal of DOX, the maturing normoblasts were shown to have gene expression profiles broadly similar to those derived from CB CD34⁺ cells including predominant expression of HbF (~80%) and some HbA and HbE. Enucleation of the differentiated immortalized erythroid progenitors was low in this report – seemingly only occurring in 1–2% of cells, but injection of erythroblasts into clodronate-treated NOD-SCID mice resulted in extensive enucleation in vivo. This line offers a very good system for studying erythroid development but the potential to use such a line to generate therapeutic red blood cells is limited by the lack of in vitro enucleation and the expression of HbF rather than HbA. These important properties might be improved by the use of a similar immortalization strategy using adult CD34⁺ as the starting material rather than iPSC.

Kurita et al. transfected both human PSC and HSPC with lentiviral vectors containing human papilloma virus (HPV)-16 E6/E7 under a tetracycline–inducible promoter. CD34⁺ cells were cultured in serum-free media with SCF, TPO and Flt3-ligand for 1 day and the HPV-16 E6/E7 expression system was then introduced by lentiviral transduction [44]. They were further cultured for 4 days in SCF, EPO and dexamethasone. Non-adherent cells were collected and cultured in the presence of SCF, EPO, dexamethasone and doxycycline for a prolonged period with regular changes of medium. The cells proliferated continuously for more than 6 months (80 population doublings). Erythroid cell lines have an erythroblastic morphology and express Glycophorin A, CD71, CD36 and c-kit. In our hands, terminal differentiation can be induced through culture in EPO alone to produce 15–30% stable reticulocytes expressing fetal or adult hemoglobin dependent on the source of the CD34⁺ starting material.

Prospects & challenges for manufacture & clinical application

The manufacture of red cells for clinical application remains very challenging. Whilst the culture of CD34⁺ cells derived from bone marrow, peripheral blood or umbilical cord blood gives rise to essentially normal reticulocytes with high levels of efficiency, overall expansion of cell numbers is low and given that a unit of clinical red cell concentrate contains around 2.5 x 10¹² erythrocytes, the prospect of manufacturing more than a handful of such units from each donation of starting material appears to be low. PSCs offer the possibility of indefinite expansion of the starting material and considerable expansion during differentiation; however, the differentiation process remains incompletely understood and may require additional signaling to achieve stable enucleation. ECLs offer the prospect of simplifying the complexity of the differentiation process and produce stable reticulocytes, but the majority of the cell expansion would need to be provided by the cell line itself. Moreover, both PSCs and ECLs engender the risk of infection during long-term culture and of neoplasia due to the use of oncogenes during derivation, insertional mutagenesis or genetic instability within the cell line itself. Whilst it can be argued that the depletion of nucleated cells from the final product and/or the use of pathogen reduction technologies may reduce these risks, they remain of concern.

The scale and complexity of the bioprocess required to produce a unit of red cells is challenging. It is not simply a matter of scale up (or scale out), but also of intensifying the cell culture, achieving control over a multifactorial process during which the cells undergo substantial phenotypic and metabolic changes, and controlling the costs of manufacture. The intensification challenge can be evaluated with respect to the known limits on bioprocess in more established suspension cell cultures such as CHO lines for protein manufacture. In such systems, there is a mechanical stress tolerance of the cells that determines the mixing and gas delivery operation of the bioreactor, an oxygen demand of the cells that (in conjunction with dissolved oxygen input rate to the bioreactor) determines the absolute density limit at which cells can be cultured, and a tolerance to soluble factors such as nutrients and cytokines or detrimental metabolites that determines medium provision and exchange strategy. Such limitations have tended to restrict culture systems for common cell lines to absolute maximum densities in the order of 1 x 10⁷ cells/ml after substantial optimization and, in some cases, cell line engineering. These densities need to be surpassed without requiring uneconomic medium perfusion or exchange rates to produce 2 x 10¹² red cell doses in economically reasonable volumes.

It has been shown that erythroid progenitors through to enucleated reticulocytes can be cultured in various scalable platforms including rocked bags and mechanically stirred tanks. There is therefore no compatibility barrier between stresses imposed by standard manufacturing equipment and reticulocyte production. Further, oxygen demand of cells during erythroid commitment is low relative to cell lines for protein manufacture. Similarly, studies of common nutrient and metabolite demand have not identified primary limits of culture intensification. This is both a challenge and an opportunity; it suggests very high intensity production is plausible with respect to common mass transfer and metabolic constraints, but achieving it may require identifying and addressing novel inhibitory mechanisms.

Given that manufacturing platforms are not incompatible with the process, manufacturing challenges (beyond autocrine growth inhibition) fall into four key areas. A particular challenge of bioprocess for red cell culture, similar to other stem cell products, is that the engineering considerations are different for each phenotypic stage of development. Further, metabolic factors governed by nutrient supply, or autocrine exposure determined by media exchange strategy, can themselves alter the phenotypic trajectory of cells, either changing the nature of the end product or altering the total proliferative capacity of the culture. This creates a complex control challenge where the control window is tracking a moving target, and any deviations can have compounding effects that necessitate control changes throughout the rest of the process. Secondly, production is not only limited by the intensification of the system at any phenotypic point (cells produced per volume), but also by the total proliferative capacity of the system that determines end product yield (cells produced per input cell). Emerging cell-line-based approaches to expansion may overcome absolute proliferative limits, but the metabolic state and sensitivities of the cells responsible for the bulk of production will be key to controlling cost of goods; proliferation needs to be weighted in the part of the process with lower bioprocess demand and therefore greater intensification potential. Thirdly, phenotypic heterogeneity in culture will dictate the yield curve with time of final enucleated product; greater homogeneity will allow harvest of more homogenously aged product and reduce the requirement for bioprocess conditions that support both the latter stages of culture and enucleated cell stability. The homogeneity of maturation and stability of enucleated cells in culture will determine whether continual or multiple harvest strategies are required to deliver required efficiency and product quality. Finally, downstream processing of enucleated cells at scale will require new technologies; current leukofiltration is not necessarily optimized to deal with the balance of cell material from a bioprocess. Although specification of downstream process requirements won’t be fully definable until the harvest material is defined, it is likely that tangential flow filtration or similar strategies may be necessary to deliver success at manufacturing scale.

Some of these challenges require mathematical process models to address them. In particular, the intensification of the system at a given point of phenotypic development, and the impact of that intensification on the trajectory and proliferative capacity of the culture, require models that can describe cell culture dynamics and evolving consequences of autocrine exposure or nutrient depletion. The limitation to the application of such approaches are the skills barrier to entry and a poor understanding of the level of model complexity required to provide manufacturing utility in the field. The mechanistic understanding associated with such models is necessary to define current system potential and the requirement for new or modified bioprocess engineering solutions.

Realizing the long-term objective of clinical supply of manufactured red cells requires further work on understanding and abbreviating erythroid differentiation protocols, achieving more consistent and stable enucleation, and intensification of both upstream and downstream bioprocessing. Most important though is clear demonstration that the red cell products of the manufacturing process are fit for purpose both in vitro and in vivo. The task remains challenging, but is helping to inform generic approaches to the large-scale manufacture of cellular therapeutics.

Financial & competing interests disclosure

The authors are funded in part by the Wellcome Trust (Grants 087430/Z/08 and 102610) and the Scottish National Blood Transfusion Service are members of the Novosang consortium HYPERLINK “http://www.novosang.co.uk” www.novosang.co.uk. JCM is named as an inventor on Patent Pub No. WO/2014/013255; International Application No. PCT/GB2013/051917 – ‘Erythroid Production’ filed by The University of Glasgow. This patent application describes a method of red cell production included in the discussion herein. No writing assistance was utilized in the production of this manuscript.

References

1. Blundell J. Observations on transfusion of blood, with a description of his gravitator. Lancet 1828; 2, 321–24. Website

2. Pappenheim A. Abstammung und entstehung der rothen blutzelle. Virchows Archives 1898; 151, 89–158. CrossRef

3. McCulloch Till J. The radiation sensitivity of normal mouse bone marrow cells, determined by quantitative marrow transplantation into irradiated mice. Radiat. Res. 1960; 13(1), 115–25. CrossRef

4. Thomas ED, Lochte H, Lu W, Ferrebee J. Intravenous infusion of bone marrow in patients receiving radiation and chemotherapy. N. Engl. J. Med. 1957; 257, 491–6. CrossRef

5. McGrath KE, Frame JM, Fegan JH et al. distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep. 2015; 11(12), 1892–904. CrossRef

6. Medvinski A, Dzierzak E. Definitive hematopoiesis is autonomously initiated by the AGM region. Cell 1996; 86, 897–906. CrossRef

7. Taoudi S, Gonneau C, Moore K et al. Extensive hematopoietic stem cell generation in the AGM region via maturation of VE-cadherin+ CD45+ pre-definitive HSCs. Cell Stem Cell 2008; 3, 99–108. CrossRef

8. Malik P, Fisher T, Barsky L et al. An in vitro model of human red blood cell production from hematopoietic progenitor cells. Blood 1998; 91, 2664–71. Website

9. Panzenbock B, Bartunek P, Mapara M, Zenke M. Growth and differentiation of human stem cell factor/erythropoietic dependent erythroid progenitor cells in vitro. Blood 1998; 92, 3658–68. Website

10. Freyssinier J, Lecoq-Lafon C, Amsellen S et al. Purification, amplification and characterization of a population of human erythroid progenitors. Br. J. Haematol. 1999; 106, 912–22. CrossRef

11. Miharada K, Hiroyama T, Sudo K, Nagasawa T, Nakamura Y. Efficient enucleation of erythroblasts differentiated in vitro from hematopoietic stem and progenitor cells. Nature Biotechnol. 2006; 24, 1255–56. CrossRef

12. Maggakis-Kelemen C, Bork M, Kayser P, Biselli M, Artmann G. Biological and mechanical quality of red blood cells cultured from human umbilical cord blood stem cells. Med. Biol. Eng. Comput. 2003; 41, 350–56. CrossRef

13. Fujimi A, Matsunaga T, Kobune M et al. Ex vivo large-scale generation of human red blood cells from cord blood CD34+ cells by co-culturing with macrophages. Int. J. Hematol. 2008; 87, 339–50. CrossRef

14. Douay L. Experimental culture conditions are critical for ex vivo expansion of hematopoietic cells. J. Hematother. Stem Cell Res. 2001; 21, 341–46. CrossRef

15. Neildez-Nguyen T, Wajcman H, Marden M et al. Human erythroid cells produced ex vivo at large scale differentiate into red blood cells in vivo. Nature Biotechnol. 2002; 5, 467–72.CrossRef

16. Giarratana M, Kobari L, Lapillonne H et al. Ex vivo generation of fully mature human red blood cells from hematopoietic stem cells. Nature Biotechnol. 2005; 23, 69–74. CrossRef

17. Douay L, Andreu G. Ex vivo production of human red blood cells from hematopoietic stem cells: What is the future in transfusion? Transfus. Med. Rev. 2007; 2, 91–100. CrossRef

18. Douay L, Lapillone H, Turhan A. Stem cells – a source of adult red blood cells for transfusion purposes: present and future. Crit. Care Clin. 2009; 25, 383–98. CrossRef

19. Douay L, Giarratana M. Ex vivo generation of human red blood cells: a new advance in stem cell engineering. In: Stem Cells in Regenerative Medicine, Methods and Protocols. J Audent, Stanford WL (editors). 2009; 482, Humana Press, NY, USA, 127–40. CrossRef

20. Giarratana M, Rouard H, Dumont A et al. Proof of principle for transfusion of in vitro generated red blood cells. Blood 2011; 118 (19), 5071-79. CrossRef

21. Mankelow TJ, Griffiths RE, Trompeter S et al. Autophagic vesicles on mature human reticulocytes explain phosphatidylserine positive red cells in sickle cell disease. Blood 2015; 126, 1831–34. CrossRef

22. Griffiths R, Kupzig S, Cogan N et al. Maturing reticulocytes internalise plasma membrane in glycophorin A containing vesicles that fuse with autophargosomes before exocytosis. Blood 2012; 11, 6296–306. CrossRef

23. Kyba M. & Daley G. Hematopoiesis from embryonic stem cells: lessons to and from ontogeny. Exp. Hematol. 2003; 31, 994–1006. CrossRef

24. Lengerke C, Grauer M, Niebuhr N et al. Hematopoietic development from human induced pluripotent stem cells. Ann. NY Acad. Sci. 2009; 1176, 219–27. CrossRef

25. Vodyanik M, Bork J, Thomson J, Slukvin II. Human embryonic stem cell-derived CD34+ cells: efficient production in the coculture with OP9 stromal cells and analysis of lymphohematopoietic potential. Blood 2005; 105, 617–26. CrossRef

26. Kaufman D, Hanson E, Lewis R et al. Hematopoietic colony-forming cells derived from human embryonic stem cells. Proc. Natl Acad. Sci. USA 2001; 98, 10716–21. CrossRef

27. Ledran, M, Krassowska A, Armstrong L et al. Efficient hematopoietic differentiation of human embryonic stem cells on stromal cells derived from hematopoietic niches. Cell Stem Cell 2008; 3, 85–98. CrossRef

28. Qiu C, Hanson E, Olivier E et al. Differentiation of human embryonic stem cells into hematopoietic cells by coculture with human fetal liver cells recapitulates the globin switch that occurs early in development. Exp. Hematol. 2005; 33, 1450–8. CrossRef

29. Olivier E, Qiu C, Velho M, Hirsch R, Bouhassira E. Large-scale production of embryonic red blood cells from human embryonic stem cells. Exp. Hematol. 2006; 34, 1635–42. CrossRef

30. Olsen A, Stachura D, Weiss M. Designer blood: creating hematopoietic lineages from embryonic stem cells. Blood 2006; 107, 1265–75. CrossRef

31. Qiu C, Olivier E, Velho M, Bouhassira E. Globin switches in yolk sac-like primitive and fetal-like definitive red blood cells produced from human embryonic stem cells. Blood 2008; 111, 2400–08. CrossRef

32. Chadwick K, Wang L, Li L et al. Cytokines and BMP-4 promote hematopoietic differentiation of human embryonic stem cells. Blood 2003; 102, 906–15. CrossRef

33. Cerdan C, Rouleau A, Bhatia M. VEGF-A165 augments erythryopoietic development from human embryonic stem cells. Blood 2004; 103, 2504–12. CrossRef

34. Chang K, Nelson A, Cao H et al. Definitive-like erythroid cells derived from human embryonic stem cells coexpress high levels of embryonic and fetal globins with little or no adult globin. Blood 2006; 108, 1515–23. CrossRef

35. Zambidis E, Peault B, Park T, Bunz F, Civin C. Hematopoietic differentiation of human embryonic stem cells progresses through sequential hematoendothelial, primitive, and definitive stages resembling human yolk sac development. Blood 2005; 106, 860–70. CrossRef

36. Keller G, Kennedy M, Papayannopoulou T, Wiles M. Hematopoietic commitment during embryonic stem cell differentiation in culture. Mol. Cell. Biol. 1993; 13, 473–86. CrossRef

37. Kennedy M, D’ Souza S, Lynch-Kattman M et al. Development of the hemangioblast defines the onset of hematopoiesis in human ES cell differentiation culture. Blood 2007; 109, 2679–87. CrossRef

38. Lu S, Feng Q, Park J et al. Biological properties and enucleation of red blood cells from human embryonic stem cells. Blood 2008; 112, 4475–84. CrossRef

39. Wang J, Zhao H, Lin G et al. In vitro hematopoietic differentiation of human embryonic stem cells induced by co-culture with human bone marrow stromal cells and low dose cytokines. Cell Biol. Int. 2005; 29, 654–61. CrossRef

40. Ma F, Ebihara Y, Umeda K et al. Generation of functional erythrocytes from human embryonic stem cell-derived definitive hematopoiesis. Proc. Natl Acad. Sci. USA 2008; 105, 13087–92. CrossRef

41. Jaichandran S, Tin-Lun A, Yu C et al. Superior red blood cell generation from human pluripotent stem cells through a novel microcarrier based embryoid body platform. Tissue Eng. Part C 2016; 22(8), 765–80. CrossRef

42. Olivier E, Marenah L, McCahill A, Condie A, Cowan S, Mountford JC. High-efficiency serum-free feeder-free erythroid differentiation of human pluripotent stem cells using small molecules. Stem Cells Transl. Med. 2016; 5, 1–12. CrossRef

43. Hirose S, Takayama N, Nakamura S et al. Immortalisation of erythroblasts by c-MYC and BCL-XL enables large scale erythrocyte production from human pluripotent stem cells. Stem Cell Rep. 2013; 1, 499–508. CrossRef

44. Kurita R, Suda N, Sudo K et al. Establishment of immortalised human erythroid progenitor cell lines able to produce enucleated red blood cells. PLOS ONE 2013; 8, e59890. CrossRef

Affiliations

Joanne C Mountford^1,2, Robert J Thomas³ & Marc L Turner^1,4
1Scottish National Blood Transfusion Service, UK
²University of Glasgow, UK
³University of Loughborough, UK
⁴University of Edinburgh, UK

This work is licensed under a Creative Commons Attribution- NonCommercial – NoDerivatives 4.0 International License.