Advances in the characterization of in vitro–generated red blood cells: from biophysical properties to functional applications

Abstract

Background

Red blood cells (RBCs), essential for oxygen transport and carbon dioxide removal, are pivotal for maintaining systemic metabolic homeostasis. However, global blood shortages and limitations in current transfusion practices underscore the urgent need for alternative sources, such as in vitro–generated RBCs. Among these, induced pluripotent stem cell (iPSC)-derived RBCs have gained attention for their potential patient-specific, pathogen-free, and immunologically compatible solutions. Yet, conventional assays of RBC function provide only a partial view of the complex molecular programs that govern erythropoiesis and maturation. These gaps motivate the integration of multi-omics platforms to comprehensively profile the developmental and functional states of in vitro–generated RBCs.

Main body

iPSC-derived RBCs hold broad translational promise, yet recapitulating erythropoiesis in vitro remains difficult given the niche’s hypoxia, extracellular cues, and multilayered regulation. We review erythroid development from primitive to definitive programs, the roles of HSCs and erythroblastic islands, and key pathways (e.g. EPO, GATA1, KLF1, TGF-β), with emphasis on transcriptional networks and chromatin remodeling that drive maturation. Integrating multi-omics provides a quantitative view of erythropoietic state transitions, explicitly linking gene-regulatory programs with metabolic and proteomic remodeling. By clarifying the mechanisms behind incomplete enucleation and hemoglobin switching, this systems-level perspective guides optimization toward scalable iPSC-derived RBC manufacture. We outline how transcriptomic, epigenomic, proteomic, and metabolomic readouts together map the regulatory landscape, and we discuss translational applications—transfusion medicine, rare blood type management, disease modeling, and drug delivery—alongside safety, ethical, and regulatory considerations.

Conclusion

This review underscores both current advances and persistent challenges in characterizing in vitro–generated RBCs, and it proposes a translational framework that integrates multi-omics data into their development and evaluation. By explicitly linking molecular regulation with process control and quality metrics, this approach provides a roadmap to optimize functional performance and ensure clinical readiness of in vitro–generated RBCs.

Background

Erythrocytes play a vital physiological role in oxygen transport, supplying oxygen to all cells and facilitating carbon dioxide removal through the lungs. Globally, blood donations are insufficient to meet the transfusion demand [1, 2], posing grave challenges in emergency and surgical settings, as well as for individuals with chronic conditions and those requiring regular transfusions [3]. Infection risks and immune rejection also compromise the safety of transfusions [4].

However, erythropoiesis is a complex process to replicate in vitro. In vivo erythropoiesis occurs within the bone marrow, a specialized microenvironment comprising intricate cellular interactions, extracellular matrix components, and specific cytokines and growth factors [5]. Hematopoietic stem cells (HSCs) rely on this precisely regulated niche for their proliferation and differentiation. Recreating this microenvironment ex vivo is technically fraught, as each component must be accurately mimicked outside the body. Moreover, the hematopoietic niche within the bone marrow is maintained under hypoxic conditions, which preserves HSC characteristics [6, 7] and regulates self-renewal and differentiation. However, stable hypoxia cannot be accurately replicated in vitro, affecting HSC properties and differentiation [8].

Physiological erythropoiesis is a tightly regulated developmental process in which HSCs give rise to erythroid progenitors that progressively mature into functional erythrocytes under the coordinated control of multiple signaling pathways and transcription factors [9, 10]. In parallel, in vitro erythrocyte production has advanced through the use of cell sources such as HSCs and induced pluripotent stem cells (iPSCs) [2, 11–13] with the ultimate goal of generating cells that closely resemble native erythrocytes at the molecular and functional levels. In this review, we first outline the physiological and molecular mechanisms governing erythropoiesis and the current strategies used for in vitro erythrocyte generation. We then compare the morphological, biophysical, and functional properties of in vitro-generated erythrocytes with those of naturally derived erythrocytes, with a particular focus on insights obtained from multi-omics analyses that integrate gene expression, protein interaction, and epigenetic data. Finally, we discuss both the potential and the current limitations of artificial erythrocytes in recapitulating the properties of physiological erythrocytes, and how these advances may guide their further development and clinical translation.

Overview of physiological hematopoiesis

Hematopoiesis is broadly classified into primitive hematopoiesis and definitive hematopoiesis. Primitive hematopoiesis occurs in the yolk sac’s extra-embryonic mesoderm, where the first blood cells arise within blood islands [14]. These early cells originate from hemangioblasts [15, 16] and mainly produce primitive erythrocytes, macrophages, and megakaryocytes [15, 17]. Among these, primitive erythrocytes retain their nuclei, are relatively large (10 μm), and facilitate rapid oxygen transport during the early embryonic stages [18].

However, primitive hematopoiesis does not involve the production of long-term self-renewing HSCs, being subsequently replaced by definitive hematopoiesis [19, 20]. Definitive hematopoiesis generates multipotent, self-renewing HSCs that sustain lifelong blood production [21]. The first HSCs emerge from endothelial cells in the aorta-gonad-mesonephros (AGM) region via endothelial-to-hematopoietic transition (EHT) [22]. These cells migrate to the fetal liver for expansion and differentiation, then home to the bone marrow to establish a self-sustaining hematopoietic system [23]. Definitive hematopoiesis ensures the continuous and multipotent generation of blood cells, supporting hematopoiesis throughout adult life (Fig. 1).

Fig. 1.

Stepwise Evolution of Hematopoiesis During Development: Primitive to Definitive Transition

Erythropoiesis, the differentiation of HSCs into erythrocytes, progresses through defined stages. HSC fate toward the erythroid lineage is guided by external cues such as erythropoietin (EPO) and transcription factors including GATA1 [24–26]. HSCs first give rise to common myeloid progenitors (CMPs), which differentiate into megakaryocyte–erythroid progenitors (MEPs) regulated by transcription factors such as PU.1, SCL, and C/EBPα [27–29]. MEPs produce burst-forming unit erythroid (BFU-E) and colony-forming unit erythroid (CFU-E) progenitors [9], which express the EPO receptor (EPOR); EPO signaling promotes their survival and proliferation [30, 31]. CFU-E cells further differentiate into proerythroblasts and subsequent erythroblast stages (Fig. 2).

Fig. 2.

Terminal erythroid differentiation and enucleation process

Proerythroblasts are the earliest morphologically identifiable precursors, with large nuclei and basophilic cytoplasm [32, 33]. They rapidly divide and mature into basophilic erythroblasts, where ribosome activation initiates hemoglobin synthesis [34]. As hemoglobin accumulates, polychromatic erythroblasts display mixed cytoplasmic staining and undergo nuclear condensation [32, 35, 36]. In the orthochromatic erythroblast stage, the cytoplasm becomes eosinophilic due to hemoglobin predominance, and cell division ceases [32]. The nucleus is expelled to form a reticulocyte, which enters circulation and loses residual organelles to become a mature erythrocyte optimized for oxygen transport [32, 37, 38].

Adult human erythrocytes are produced at a rate of 2–3 million cells per second [39]. Erythropoiesis predominantly occurs in the bone marrow, generating RBCs that circulate in the bloodstream for around 120 days. During this period, erythrocytes maintain high flexibility, enabling them to pass through narrow capillaries. Eventually, their membranes become damaged, and metabolic functions deteriorate [37]. Senescent erythrocytes are primarily removed by macrophages in the spleen [39, 40]. During this process, hemoglobin in erythrocytes is degraded into heme and globin, with heme further catabolized into iron and bilirubin [41–43]. The iron is released into the bloodstream and recycled for new erythropoiesis and bilirubin is transported to the liver and excreted in bile [43].

Mechanistic insights into erythroid development

Primitive erythroblasts

Primitive erythroblasts originate primarily from KDR⁺ (a mesodermal marker associated with hemangiogenic potential) GYPA+ (CD235a; glycophorin A) mesoderm; their differentiation being regulated by activin-nodal signaling and inhibition of the Wnt pathway [44]. KDR⁺GYPA⁺ mesodermal cells transition through the hemangioblast and hemogenic endothelium stages to become primitive hematopoietic progenitors, which differentiate into primitive erythroblasts, macrophages, and megakaryocytes [15, 44, 45]. Primitive erythroblasts express GYPA, a key marker of early erythroid progenitors [44, 46]. They also express embryonic globin (HBE), which distinguishes them from definitive erythroid cells expressing fetal globin (HBG) or adult globin (HBA) [47]. During the hemogenic endothelium stage, transcription factors such as HOXA3, RUNX1, SOX17, and NOTCH1 are activated [44, 48–50]. They regulate EHT, promoting the conversion to primitive hematopoietic cells [51, 52], which is controlled by activin-nodal signaling [53]. In contrast, inhibition of the Wnt/β-catenin pathway facilitates primitive hematopoiesis while suppressing definitive hematopoiesis [44].

The differentiation of primitive erythroblasts relies on sequential regulation by essential transcription factors, including SCL/TAL1, GATA1, Krüppel-like factor 1 (KLF1), and LMO2 [54]. Deficiency in any of these factors leads to the failure of primitive erythropoiesis in mice, resulting in embryonic lethality. SCL/TAL1 regulates hematopoietic lineage differentiation and maturation, functioning as part of a multiprotein complex [55–57].

Endothelial-to-hematopoietic transition

Endothelial cells exhibit remarkable plasticity during early embryonic development, differentiating into various cell types with distinct characteristics, including arterial, venous, and lymphatic identities. Moreover, some endothelial cells can transition into blood cells or acquire mesodermal traits essential for heart development [58]. Primitive, undifferentiated endothelial cells derived from mesodermal progenitors in the yolk sac are regulated in their formation and differentiation by signaling pathways involving BMP4, FGF2, and VEGF-A [59, 60]. VEGF-A signals through VEGFR1 (Flt-1) and VEGFR2 (Flk-1) receptors, which regulates vasculogenesis [60].

Through EHT, vascular endothelial cells are converted into hemogenic endothelium, which generates multipotent hematopoietic stem and progenitor cells (HSPCs) [61]. Herein, retinoic acid signaling induces c-Kit expression, while Notch signaling regulates the cell cycle to determine the characteristics of hemogenic endothelium [62]. Through EHT, hemogenic endothelial cells in the AGM region yield intra-aortic hematopoietic clusters [63] comprising pre-HSCs that express a combination of endothelial markers (VE-cadherin, CD31, CD34) and hematopoietic markers (CD41, c-Kit, CD43, CD45) at varying levels [64, 65]. This implies intermediate cell states during the transition from endothelial cells to HSPCs.

Erythropoiesis in the fetal liver

The fetal liver is the primary hematopoietic organ during mid-gestation [66]. In the human fetus, HSCs migrate to the fetal liver between 7 and 17 weeks, where they undergo extensive proliferation and differentiation [67, 68]. These HSCs cohabit with yolk sac–derived erythro-myeloid progenitors (EMPs) that seed the liver earlier; accumulating evidence indicates that early hepatic erythropoiesis is predominantly EMP-driven, whereas HSC expansion in the liver primarily prepares for subsequent bone marrow colonization. The fetal liver remains the dominant hematopoietic site until the mid-second trimester; thereafter, hematopoiesis gradually transitions to the bone marrow [68]. HSCs in the fetal liver proliferate rapidly; however, this expansion primarily supports subsequent bone marrow seeding rather than serving as the main source of erythropoiesis, which largely derives from EMP-origin progenitors. Single-cell transcriptomic studies reveal that erythropoiesis in the fetal liver is highly active during early developmental stages, gradually giving way to a more diverse hematopoietic landscape as hematopoiesis transitions to the bone marrow [68].

Erythropoiesis is highly active in the fetal liver due to its unique hematopoietic niche, which supports the expansion of HSCs and sustains robust EMP-derived erythropoiesis. VCAM1⁺ macrophage-mediated EBIs facilitate the proliferation and maturation of erythroid progenitors in the fetal liver [56, 68, 69]. Erythropoiesis in the fetal liver is critically dependent on erythropoietin (EPO) signaling, as demonstrated in animal models [56]. EPO signaling regulates the proliferation and differentiation of colony forming unit-erythroid (CFU-E) cells, and its deficiency severely impairs erythropoiesis [70]. Excessive iron accumulation can alter hepatocyte metabolism, disrupting hematopoietic function [70]. Iron overload occurring in the fetal liver under EPO-deficient conditions causes metabolic stress. Furthermore, transcription factors such as RUNX1 and GATA2 are critical for maintaining HSCs and regulating erythroid progenitor differentiation, while Notch signaling is important in determining hematopoietic cell fate [68].

Erythropoiesis in the fetal liver differs from hematopoiesis in the adult bone marrow in several ways. HSCs in the fetal liver proliferate approximately 38 times faster than those in the adult bone marrow, a surge interpreted as expansion toward bone marrow seeding rather than direct support of bulk erythropoiesis. While HSCs in the adult bone marrow predominantly remain in a quiescent state, fetal liver HSCs maintain an actively proliferative state [71, 72]. The composition of hematopoietic lineages also shifts dynamically throughout development; in the early fetal liver, erythroid lineage cells dominate; later, myeloid and lymphoid lineage cells progressively increase [68]. This transition might facilitate the eventual shift of hematopoiesis from the fetal liver to the bone marrow in preparation for postnatal hematopoiesis.

Erythropoiesis in the adult bone marrow

Adult erythropoiesis in the bone marrow differs considerably in transcriptional regulatory mechanisms and gene expression patterns from erythropoiesis in the fetal liver. Erythropoiesis in the fetal liver is characterized by a high proliferative capacity and rapid cell cycle progression, predominantly producing fetal hemoglobin (HbF) containing γ-globin. However, erythropoiesis in the bone marrow occurs under more tightly regulated mechanisms, with increased β-globin expression facilitating the transition from HbF to HbA through globin switching [73].

Erythropoiesis in the bone marrow is distinguished from fetal erythropoiesis by the regulation of iron metabolism, chromatin structural modifications, and the action of specific transcription factors. BCL11A suppresses γ-globin gene expression, facilitating the transition from HbF to hemoglobin A (HbA) [74, 75]. The regulatory region of BCL11A is associated with histone modifications that alter chromatin accessibility; changes in DNA methylation and histone acetylation repress HbF and enhance HbA expression [76]. Another distinguishing feature of erythropoiesis in the bone marrow compared to the fetal liver is the activation of SOX6, which plays a crucial role in γ-globin repression and adult erythroid differentiation [77]. SOX6 cooperates with B-cell lymphoma/leukemia 11 A (BCL11A) and KLF1 to suppress γ-globin and promote β-globin expression, serving as a key regulator of the HbF-to-HbA switch, which predominates in the fetal liver and shifts toward HbA production in the bone marrow [77, 78]. Chromatin accessibility regulation is another critical mechanism in erythropoiesis within the bone marrow [57]. Erythroid progenitors in the bone marrow undergo more intricate chromatin remodeling than those in the fetal liver, leading to precise stage-specific gene expression regulation [76].

Erythropoiesis in the bone marrow is tightly regulated through close interactions with the hematopoietic niche. In adult hematopoietic tissues, erythropoiesis is modulated not only by EPO signaling but also by growth factors such as SCF and IL-3, which ensure the survival and proliferation of erythroid progenitor cells [79]. In adult bone marrow, the differentiation of HSCs into erythroid lineage cells is stringently regulated, ensuring that erythropoiesis is activated only when needed through a feedback mechanism [80, 81].

Erythroblastic Islands

Erythroblastic islands (EBIs), first described in 1958 by Marcel Bessis [82], primarily function to scavenge pyrenocytes during enucleation, while also regulating erythroblast maturation, iron delivery, nuclear extrusion, and intercellular signaling. EBIs play a vital supportive role throughout multiple stages of erythropoiesis, providing a specialized microenvironment that supports the differentiation of erythroblast precursors into reticulocytes [83, 84]. Central macrophages in EBIs supply iron to erythroblasts via ferritin and transferrin, promoting hemoglobin synthesis [85, 86]. Macrophage–erythroblast interactions are essential for supporting terminal erythrocyte development and preventing apoptosis [84, 87]. Moreover, macrophage–erythroblast communication is also regulated independently of growth factors, orchestrating erythroblast differentiation and proliferation [88].

The interaction between the erythroblasts and macrophages provides the niche necessary for erythropoiesis, supporting erythroblast survival, proliferation, and maturation. Central macrophages store iron through transferrin and ferritin (FTH1, FTL), releasing it into circulation via ferroportin (SLC40A1) [86, 89, 90]. Erythroblasts, in turn, express the transferrin receptor (TFRC, CD71) to acquire iron for hemoglobin synthesis [91]. Heme oxygenase (HMOX1) in central macrophages helps break down heme and release iron [92]. Erythroblasts express genes such as ALAS2 and AHSP, contributing to heme biosynthesis and α-globin chain stabilization [93–95]. Thus, the coordinated activity between erythroblasts and macrophages is indispensable for complete RBC maturation.

Because of its importance, co-culture of erythroblasts and macrophages was first reported in 1979 and has since been widely used following EBIs in vitro culture [88]. However, it is difficult to fully replicate the bone marrow environment [96]. Moreover, the relationship between in vitro and in vivo EBIs remains unclear, suggesting potential differences in gene expression and functional properties.

Molecular regulation of erythropoiesis: signaling pathways and transcriptional networks

The major signaling pathways and transcription factors involved in erythropoiesis regulate the proliferation, differentiation, and survival of erythroid progenitors, determining the efficiency and maturation of RBCs during their production (Fig. 3) Table 1.

Fig. 3.

Key signaling pathways regulating erythropoiesis

Table 1.

Expression patterns of representative erythroid genes across distinct hematopoietic stages and cell types

Gene	Primitive erythroblast (yolk sac)	Fetal HSC	Fetal erythroblast	Fetal erythrocyte	Adult HSC	Adult erythroblast	Adult erythrocyte	Ref
GATA1	*****	**	****	***	*	****	***	[25, 116–268]
GATA2	*****	*****	***	*	*****	**	-	[117, 269–271]
RUNX1	*	*****	***	*	***	*	-	[272–276]
ALAS1	*	***	-	-	***	*	-	[277–279]
ALAS2	***	*	*****	*****	-	*****	*****	[278–280]
AHSP	*	***	*****	***	**	*****	-	[68, 281–284]
EPOR	***	***	*****	***	***	*****	-	[79, 285–288]
CD71	*****	***	*****	***	**	*****	*	[9, 68, 285, 289, 290]
CD235a	*****	*	*****	*****	*	*****	*****	[68, 285], 290– [292]
KLF1	*****	**	*****	****	*	*****	***	[129–294]
BCL11A	***	****	*****	*****	****	*****	***	[68, 74, 291, 294, 295]
HBE	*****	-	-	-	-	-	-	[296, 297]
HBZ	*****	-	-	-	-	-	-	[296, 297]
HBG	***	*	*****	*****	-	**	*	[296]
HBA	***	*	*****	*****	*	*****	*****	[39, 296]
HBB	*	-	**	**	-	*****	*****	[39, 296]

EPO signaling pathway

EPO, a glycoprotein hormone primarily produced in the kidney, regulates erythropoiesis and promotes the survival and proliferation of erythroblasts. EPO binds to EPOR expressed on the surface of erythroblasts, initiating the JAK2/STAT5 signaling pathway to regulate erythroblast survival and differentiation [69]. EPO binding activates JAK2, leading to phosphorylation of the intracellular domain of EPOR, which recruits and activates STAT5 [69, 97]. Activated STAT5 translocates to the nucleus and induces the expression of survival-related genes, such as BCL-XL (BCL2L1) [98].

EPO signaling also activates the PI3K/AKT and MAPK/ERK pathways, regulating the survival and proliferation of erythroid progenitor cells [99]. The PI3K/AKT pathway is essential for the terminal stages of erythroid differentiation. When EPO binds to EPOR, the regulatory subunit (P85) of PI3K is activated, phosphorylating AKT [100, 101]. Activated AKT translocate to the nucleus, where it modulates transcription factors such as FOXO3 and GATA1, promoting erythroid development [102, 103]. Simultaneously, it regulates HIF1α and mTOR signaling, enhancing the erythropoietic response under hypoxic conditions [104, 105]. Meanwhile, the MAPK pathway plays a crucial role in the early stages of erythroid differentiation; however, its activation at later stages inhibits differentiation. This pathway involves ERK1/2, p38 MAPK, and JNK signaling and is regulated through RAS and RAF [106]. ERK1/2 cooperates with EPOR to regulate transcription factors such as c-Myc and ETS proteins, promoting erythroid progenitor proliferation [99, 107]. Meanwhile, p38 MAPK is implicated in stress erythropoiesis, and under certain conditions, it may suppress erythroid differentiation [108]. The EPO signaling network is highly interconnected, featuring cross-inhibition mechanism between the PI3K/AKT and MAPK pathways [109]. This regulatory interplay enables erythropoiesis to adapt to diverse physiological conditions.

During early human erythroid progenitor stages, SCF–c-Kit signaling cooperates with EPO to activate the PI3K/AKT–MAPK axis and enhance proliferation and survival, with particularly strong synergy observed at the level of ERK1/2 activation [110, 111]. SCF also augments EPO-induced STAT5 transcriptional activity, thereby potentiating EPO signaling [112]. In human BFU-E, colony formation requires the combination of SCF and EPO; by the subsequent CFU-E stage, colonies can form with EPO alone and the dependence on SCF declines markedly [113]. Consistently, as differentiation proceeds, c-KIT expression and function are down-regulated, and repression of KIT expression via the EPO–Lyn axis promotes the transition toward terminal maturation [114].

GATA1 transcription factor

GATA1 orchestrates the maturation and function of erythroid cells. It comprises two zinc finger domains and an N-terminal transactivation domain (N-TAD), which regulate the expression of genes involved in erythroid differentiation [115]. GATA expression and activity are tightly controlled; high levels of GATA1 protein must be maintained during the early stages of erythroid maturation, whereas its downregulation is essential for terminal differentiation [116].

GATA1 coordinates the transcriptional network governing erythropoiesis. By interacting with FOXO3, GATA2, and KLF1, it regulates erythroid-specific gene expression and modulates key transcriptional programs activated by EPO signaling [102, 117]. GATA1 also controls genes associated with mitochondrial autophagy and RNA degradation, thereby facilitating erythroid maturation [118]. GATA1 expression is initially low in HSCs and progenitor cells, increasing during erythroid commitment, and subsequently declining sharply at the final maturation stage [25, 119]. GATA1 is regulated at multiple levels—including transcription, mRNA translation, post-translational modification, and protein turnover—via mechanisms that modulate translation efficiency, histone acetylation, ubiquitination, and protein stability [120, 121].

GATA1 is the key transcriptional regulator of erythroid commitment and forms transcriptional activation and repression complexes, with FOG1 as a major interacting partner. Although FOG1 does not bind DNA directly, its association with GATA1 enhances the function of transcriptional regulatory complexes [55, 122]. The GATA1–FOG1 complex interacts with the nucleosome remodeling and deacetylating (NuRD) complex to exert dual regulatory roles: repressing specific gene expression while simultaneously activating genes essential for erythropoiesis [123, 124]. This cooperative interaction between GATA1 and FOG1 establishes a sophisticated regulatory mechanism that promotes erythroid-specific gene expression while suppressing alternative hematopoietic differentiation pathways, such as myeloid lineage differentiation [125].

KLF1 and BCL11A

KLF1 and BCL11A are important regulatory factors in erythroblast differentiation and globin switching. KLF1 is an erythroid-specific transcription factor governing the expression of genes essential for erythroblast maturation [126]. It directly activates β-globin gene (HBB), aminolevulinate synthase 2 (ALAS2), and alpha-hemoglobin stabilizing protein (AHSP)—critical for hemoglobin synthesis and erythrocyte function [127, 128]. BCL11A is a central regulator of globin switching. During fetal development, γ-globin is expressed, forming HbF; postnatally, its expression is downregulated and replaced by β-globin, forming HbA [129]. BCL11A facilitates this transition by repressing γ-globin expression and promoting β-globin expression [75], a process regulated upstream by KLF1, which induces BCL11A expression to ensure the precise timing of globin switching [75, 129]. Dysregulation of these factors can lead to hematologic disorders. KLF1 mutations are associated with anemia and erythrocyte structural defects, whereas BCL11A deficiency leads to persistent γ-globin expression and abnormal retention of HbF [130, 131].

TGF-β signaling and SMAD pathway

The SMAD and TGF-β signaling pathway exerts an inhibitory effect during the late stages of erythroid differentiation [132]. TGF-β signaling regulates hematopoietic cell development, influencing the differentiation of HSCs and progenitor cells. In erythropoiesis, ligands of the TGF-β family, such as activin and GDF11, bind to the TGF-β type II receptor, initiating the signaling cascade and activating the TGF-β type I receptor (ALK5) [133]. This activation leads to the phosphorylation of SMAD2 and SMAD3 (R-SMADs), which form a complex with SMAD4 and translocate to the nucleus, where they regulate the expression of genes inhibiting erythroid differentiation [134].

TGF-β signaling acts as a negative regulator of hematopoiesis by inducing cell-cycle arrest and apoptosis in late-stage erythroid differentiation [135]. Under normal physiological conditions, GDF11 expression decreases, while EPO signaling increases, thereby promoting erythropoiesis [135]. However, TGF-β pathway hyperactivation can cause hematopoietic disorders such as SMAD2/3-driven ineffective erythropoiesis. Therefore, TGF-β pathway inhibitors, such as ALK5 inhibitors and activin receptor ligand traps, are being investigated as potential therapeutic agents for hematopoietic disorders [136–138].

Additional regulatory factors

The hypoxia-inducible factor (HIF) pathway regulates EPO synthesis while modulating iron metabolism and the bone marrow microenvironment to promote erythroid progenitor proliferation and differentiation [139–141]. HIF-2α has been identified as the primary transcription factor regulating EPO production in adult kidneys and liver [141, 142]. While HIF-1α is crucial for embryonic erythropoiesis, HIF-2α becomes the predominant regulator in adults [56, 143, 144]. HIF-2α-deficient mouse models exhibit anemia and hypocellular bone marrow, suggesting a link between erythropoiesis and HIF-2-dependent renal EPO production. HIF is also closely associated with iron metabolism [145]. HIF-2α suppresses hepcidin, a key regulator of iron homeostasis, thereby enhancing iron absorption and optimizing iron availability for erythropoiesis [146, 147]. The proliferation and maturation of erythroid progenitor cells within the bone marrow are also regulated by HIF signaling. Hypoxic conditions have been shown to promote erythroid progenitor maturation and differentiation within the bone marrow, with HIF-2α playing a key role in shaping the erythroid niche [148]. Furthermore, HIF-2α deficiency has been shown to regulate the expression of vascular cell adhesion molecule-1 (VCAM-1), thereby impairing erythroid maturation [149].

Notch signaling regulates erythropoiesis by orchestrating apoptosis, HSC fate determination, and erythropoietic response under stress conditions [150, 151]. In Notch/RBPjκ-deficient mice, erythroid cells accumulate excessively in the yolk sac and bone marrow, suggesting that Notch signaling regulates erythroid differentiation and cell survival by suppressing GATA-1 activity through Hes1 [150, 152]. The activation of the Notch2–Hes1 axis promotes erythropoiesis, whereas its inhibition increases platelet production. Under stress-induced hematopoiesis, Notch2 expression is upregulated, facilitating erythroid progenitor activation, while Notch deficiency results in impaired erythroid recovery following radiation-induced damage [153]. At the HSC stage, Notch signaling maintains GATA-2 expression and suppresses GATA-1 transcriptional activity, thereby sustaining an undifferentiated state; conversely, the loss of Notch1 signaling promotes erythroid lineage differentiation [148].

The JAK2 and NF-κB signaling pathway is important for regulating cell survival, proliferation, and differentiation during erythropoiesis. In normal erythroid development, NF-κB transcription factors (e.g. p50, p52, p65) are highly expressed in early erythroid progenitors (BFU-E), declining gradually as differentiation progresses [154, 155]. NF-κB modulates c-myc and c-myb expression, thereby influencing erythroid progenitor proliferation and differentiation, with granulocyte-macrophage colony-stimulating factor stimulation promoting the nuclear translocation of p52 and p65 to activate these target genes [154]. Aberrant NF-κB activation, however, can impair the self-renewal capacity of HSPCs, leading to hematopoietic dysfunction [156, 157]. Persistent activation of the non-canonical NF-κB pathway via NF-κB-inducing kinase reduces HSPCs within the bone marrow and disrupts the bone marrow microenvironment, contributing to bone marrow failure [158]. Sustained activation of IKK2, a key regulator of the canonical NF-κB pathway, results in excessive NF-κB signaling, which impairs maintenance of stemness and aberrant erythroid differentiation [159]. NF-κB also plays a crucial role in hematopoietic stress responses by mediating inflammatory cytokine signaling [160]. Furthermore, NF-κB interacts with EPO signaling, where JAK2-mediated phosphorylation of IκB facilitates NF-κB nuclear translocation and regulates cell survival and signal transduction during erythropoiesis [161].

Applications of multi-omics analysis and recent examples

Functional characterization of in vitro–generated red blood cells (iRBCs) has long relied on morphological, biochemical, and biophysical assays [162–165]. Representative approaches include microscopic observation and flow-cytometric analysis to assess cellular morphology and enucleation efficiency, which indicate the degree of terminal maturation during erythroid differentiation [166, 167]. Measurement of the oxygen dissociation curve (P₅₀) provides a quantitative evaluation of the oxygen-binding and release capacities of iRBCs, and can serve as an informative parameter for functional comparison with native red blood cells when integrated with complementary assays such as deformability, metabolic, and proteomic analyses [168–170]. Osmotic fragility and hemolysis assays are routinely employed to verify membrane mechanical stability and permeability by assessing lysis under varying osmotic or oxidative conditions [171, 172]. In addition, ektacytometry or microfluidic deformability assays measure the elongation index of cells under defined shear stress to determine membrane elasticity and viscoelasticity, parameters closely related to microcirculatory transit [164]. Finally, surface antigen profiling using flow cytometry or antibody binding tests enables the detection of major blood-group antigens (e.g., ABO, Rh, Kell), thereby validating the immunohematologic safety of iRBCs [170, 173]. Collectively, these classical assays provide essential information for evaluating the morphological and biochemical properties of iRBCs and have established the fundamental framework for comparing laboratory-generated RBCs with their physiological counterparts.

While conventional approaches have provided valuable insights into iRBC morphology and function, they may have inherent constraints that limit their ability to fully capture the molecular dynamics underlying erythroid differentiation and maturation. Therefore, while classical single-parameter analyses remain indispensable for assessing iRBC phenotype and functionality, integrative and multi-dimensional approaches are increasingly recognized as essential for obtaining a more comprehensive understanding of iRBC maturation and functional integrity.

To overcome these constraints, multi-omics analysis has emerged as a powerful solution. Multi-omics refers to the integrated readout of molecular layers—typically, transcriptome and chromatin accessibility at single-cell resolution (scRNA-seq, scATAC-seq, or multiome), protein and surface phenotypes (CITE-seq, proteomics), metabolites and lipids (metabolomics), and spatial context [174]. This integration can align regulatory programs with phenotypes and functional endpoints [175, 176] providing a comprehensive framework for connecting molecular regulation to cell function.

The multi-omics-based paradigm thus complements and extends classical assays by offering molecular insights previously unattainable. Although comprehensive multi-omics studies on in vitro–generated red blood cells are still limited, integrating diverse molecular layers—such as transcriptomic, proteomic, and metabolomic profiles—holds strong potential to uncover the complex regulatory networks governing erythroid differentiation, maturation, and storage biology. Such approaches are expected to bridge molecular regulation with functional phenotypes, ultimately establishing new standards for iRBC quality assessment, functional optimization, and clinical translation in transfusion medicine.

Building on these classical foundations, the following sections explore how integrative multi-omics approaches may redefine functional assessment and translational readiness of iRBCs.

Multi-omics strategies: native human RBCs vs. PSC-derived RBCs

Comparing in vivo human erythroid biology with in vitro stem-cell models reveals system visibility differences. Across RBC omics, enucleation and minimal translation indicate scarce poly(A)-selected transcripts and weak RNA–protein or RNA–function coupling, and low molecule counts make clustering sensitive to stochastic dropouts [176, 177]. Besides, mRNA-centric assays underrepresent the small-RNA landscape of RBCs, and protein/metabolic states often carry more informative signals [175, 176]. Consequently, single-cell transcriptomics is most useful just before enucleation, whereas proteomics, metabolomics, and small-RNA profiling better anchor analyses of fully mature RBCs.

Bone marrow provides access to pre-enucleation stages where multi-omics is most revealing [178, 179]. Large-scale CITE-seq efforts have aligned surface phenotypes with transcriptomes across dozens of erythroid and non-erythroid states, bridging historical gating with single-cell atlases; complementary scATAC-seq and mass cytometry (CyTOF) further refine regulatory programs and protein phenotypes at these stages [179, 180]. Spatial multiplex imaging (e.g., CODEX) situates erythroid states within stromal neighborhoods, quantifying niche interactions, which in vitro monocultures cannot achieve [181]. Public multimodal resources pairing RNA + ATAC and CITE-seq serve as practical yardsticks for cross-study comparisons and for aligning in vivo marrow with in vitro trajectories [178, 180].

In pluripotent systems, integrative designs concentrate on fate-decision checkpoints. Time-resolved single-cell trajectories link stage transitions. Combining chromatin accessibility and histone-mark maps with bulk and single-cell transcriptomics has pinpointed regulators of hemogenic specification and EHT that RNA alone would miss [182, 183]. Downstream, CITE-seq in iPSC-derived hematopoietic cells helps standardize stage enrichment and enables stage-matched comparisons to marrow [184]. Single-cell comparisons consistently place iPSC-derived definitive erythroid cells near fetal-liver programs, and β-globin reporter studies show frequent co-expression of embryonic, fetal, and adult globins—indicating that “adult-like” RNA profiles must be tested against regulatory closure and protein/metabolic convergence [185]. Pairing RNA with ATAC or protein readouts offers the most robust bridge between pre-enucleation regulation and terminal RBC function [183].

Benchmarking erythroid maturation: in vivo marrow vs. PSC-derived cultures

Erythropoiesis in human bone marrow executes a tightly coupled program whereby hemoglobin expression switches from fetal to adult forms, chromatin is progressively compacted, organelles are cleared, and metabolism is rewired to support the hemoglobinized, enucleated end state. Fetal-to-adult hemoglobin switching is driven by lineage-specific regulators (e.g., BCL11A, ZBTB7A/LRF, KLF1) and coincides with epigenomic remodeling at the β-globin locus [186]. Single-cell and bulk studies show decreasing accessibility and consolidating chromatin states across late erythroblasts as nuclei are expelled [187, 188]. In parallel, mitophagy eliminates mitochondria, consolidating the shift toward glycolysis and redox programs typical of mature RBCs [189, 190]. These layers produce a coherent multi-omics signature of maturation that multi-modal assays (scRNA-seq, scATAC-seq, proteo- and metabolomics, and spatial protein imaging) capture with high fidelity in marrow [175, 178, 181, 191].

Many in vitro stem-cell differentiation systems capture this terminal phase only incompletely. iPSC/ESC-derived erythroid cultures frequently show lower and more variable enucleation efficiencies than marrow, continued γ-globin expression despite progression into late stages, and epigenomic “lag,” e.g., ATAC-seq signatures consistent with incomplete closure/condensation when chromatin-modifying steps are perturbed or suboptimal [192–194]. Metabolic features often reflect delayed organelle clearance and residual mitochondrial activity, which can impair the mechanics of nuclear extrusion and final maturation [195]. These discrepancies are readily detected by multi-omics: scRNA-seq reports globin ontogeny and stage composition, scATAC-seq and histone-mark maps benchmark regulatory closure, and proteomics and metabolomics resolve the functional endpoint [182, 196]. As protocols improve, meaningful comparisons will require in vivo marrow references to assess convergence across globin switching, enucleation, epigenetic closure, and metabolic rewiring simultaneously.

Trajectory fidelity and cellular heterogeneity

Human bone marrow erythropoiesis follows an ordered trajectory—from proerythroblasts, through basophilic and polychromatic stages, to orthochromatic erythroblasts—marked by coordinated transcriptional programs and progressive chromatin consolidation that culminate in enucleation [177, 188]. Single-cell and bulk time-series show that late stages compress into a narrow pseudotime with decreasing chromatin accessibility and consolidated regulatory states, providing a high-fidelity scaffold for stage assignment [57, 187, 197]. Meanwhile, pluripotent stem cell–derived erythroid cultures display broader, partially overlapping clusters and branchier trajectories, reflecting asynchronous differentiation and the coexistence of primitive and definitive ontogenies; even when late-stage markers are expressed, stage composition remains more heterogeneous than in marrow [178, 182, 196]. Multi-omics makes these discrepancies explicit: CITE-seq anchors transcript states to surface immunophenotypes and tightens stage calls, while scATAC-seq benchmarks the degree of regulatory “closure” against marrow references—often revealing incomplete compaction in vitro that explains blurred or detoured paths [178, 183, 184]. Together, these modalities transform trajectory fidelity from a heuristic into a measurable alignment problem against an in vivo reference manifold.

Multi-omics benchmarks for RBC manufacturing and gene editing

In vivo multi-omics of human marrow and blood provides mechanistic and cartographic references for translational work. Reference atlases linking surface immunophenotypes to transcriptomes within stromal neighborhoods pinpoint disease-relevant circuits, offering benchmarks for dysregulated erythropoiesis in anemia and thalassemia and niche remodeling [178, 181, 198–200]. In practice, we treat these in vivo references as multi-axis benchmarks to judge whether manufactured or gene-edited RBCs converge simultaneously on four maturation layers—globin composition, enucleation, epigenetic closure, and metabolic function [178]. In parallel, proteomic and metabolomic studies of red cells highlight functional endpoints that can serve as clinically meaningful biomarkers.

Pluripotent stem cell–derived systems enable intervention and iteration. Patient-specific iPSC lines allow correction of pathogenic HBB variants and re-testing of erythroid maturation in isogenic backgrounds, while genome editing strategies reactivating fetal programs have matured from concept to clinic [201–204]. In vitro erythropoiesis provides a controllable platform to evaluate how editing designs, differentiation conditions, or candidate small molecules affect globin switching, enucleation, and metabolic rewiring before advancing to trials or manufacturing runs. The recent clinical success of CRISPR-based exagamglogene autotemcel (BCL11A enhancer editing) in sickle cell disease and transfusion-dependent β-thalassemia underscores how regulatory insights translate into durable benefit, connecting atlas-guided hypotheses and therapy [205, 206].

Connecting in vitro readouts to in vivo fidelity requires multi-axis benchmarks. For manufactured or gene-edited RBCs, the most useful assessments align four layers against marrow references: (i) globin composition (adult vs. fetal at RNA and protein levels), (ii) enucleation (efficiency and morphology), (iii) epigenetic closure (RNA + ATAC/selected histone marks at β-globin locus and maturation loci), and (iv) metabolic/functional endpoints (ATP, redox resilience, deformability, antigen repertoire). Multi-omics makes this alignment quantitative: CITE-seq to anchor phenotype with state, scATAC-seq to benchmark regulatory closure, and proteo-/metabolomics to certify functional convergence. Thus, similarity to native RBCs becomes a measurable gate for preclinical decision-making and release criteria in advanced manufacturing [178, 181, 190, 207].

Translational value of in vitro-generated rbcs: the case for multi-omics

iRBCs offer clear clinical promise across multiple fronts, beginning with transfusion medicine, where they could not only alleviate shortages but also serve alloimmunized patients and those with rare blood phenotypes for whom matched units are scarce [173, 208]. Building on that clinical rationale, patient-specific iPSC lines extend the value of iRBCs to disease modeling, enabling isogenic pre-/post-correction comparisons in hemoglobinopathies and red-cell enzymopathies and providing tractable systems to probe host–pathogen interactions such as malaria [209–211]. This experimental tractability naturally feeds into drug discovery, where erythroid cultures function as controllable platforms for small molecule screening to enhance enucleation, globin switching, and membrane biomechanics, and for testing concepts in RBC-mediated therapeutic delivery [212]. Taken together, these converging use-cases identify the most compelling near-term clinical evaluations. These opportunities are concentrated in settings with significant unmet need and a favorable benefit–risk profiles, such as rare blood types, highly sensitized patients requiring antigen-matched units, and defined research or diagnostic applications [208].

Substantial advances in process and culture optimization—including greater granularity in stem cell differentiation protocols, adoption of xeno-free/animal-component–free media, and high-density bioreactor–based manufacturing with enhanced process control—have markedly improved the manufacturability of iRBCs [213, 214]. In HSPC-derived systems, near-complete erythroid differentiation has been demonstrated even under feeder- and animal-free conditions, underscoring the field’s progress [215]. Nevertheless, several structural hurdles continue to delimit clinical translation. First, achieving clinically relevant yields that reproducibly approach a therapeutic unit (~ 2 × 10^12 cells) while maintaining an acceptable cost of goods remains a central bottleneck, tightly coupled to scale-up/scale-out strategy and process economics [215, 216]. Second, complete enucleation and functional maturation—including adult hemoglobin switching with appropriate 2,3-DPG regulation, deformability and microvascular transit, and splenic filtration resilience—are not yet uniformly secured across production batches and thus contribute to inter-batch variability at the preclinical stage [162, 211]. Third, issues specific to iPSC-derived products—genetic/epigenetic stability and the residual undifferentiated cell–associated tumorigenicity risk—necessitate high-sensitivity quality control and well-defined release testing panels throughout the manufacturing chain [217]. Finally, an evidence gap at the clinical level remains salient. While the randomized cross-over RESTORE phase-1 trial in humans is evaluating survival and safety of transfused iRBCs produced from adult HSPCs, iPSC-derived RBCs remain focused on preclinical validation and manufacturing readiness [218]. Taken together, convergence on standardized, serum-free and scalable processes, tighter batch-to-batch uniformity in functional maturation, and a pluripotent stem cell-tailored safety and regulatory framework should bring stem cell–derived RBCs appreciably closer to clinical reality [214].

Multi-omics can serve as a quantitative bridge linking process variables, cellular states, and clinically relevant functions in iRBC manufacturing [219, 220]. For scale-up and xeno-free transition, targeted metabolomics and lipidomics—complemented by ¹³C-based flux analysis—enable active control over nutrient delivery, redox balance, and energy charge (ATP, 2,3-DPG), informing feed strategies and chemically defined media design [219]. Recent studies further show that storage and distribution parameters (e.g., supercooling) imprint discernible metabolic fingerprints, underscoring the feasibility of metabolism-anchored critical quality attributes (CQAs) that couple process to quality [221]. Together with integrative “storage-lesion” omics, these data support a two-tier approach: broad discovery-scale omics during development, compressed to pragmatic surrogate assays (e.g., ATP/2,3-DPG panels) for routine release testing [220, 222, 223].

Uniform functional maturation will benefit from single-cell, chromatin-resolved maps that reveal residual immature clusters along the erythroid trajectory and the regulators of fetal-to-adult hemoglobin switching [224, 225]. In parallel, quantitative proteomics of the membrane–cytoskeleton complex (Band 3–ankyrin–spectrin and partners) offers mechanism-proximal surrogates for deformability, microvascular transit, and splenic filtration resilience; glycomics/glycoproteomics can audit the completeness of the blood-group antigen repertoire [226, 227]. Mass-spectrometry–based antigen phenotyping is increasingly promising as a QC adjunct, although it does not yet fully replace clinical serology in all settings [228, 229].

A pluripotent-specific safety framework is best implemented as a layered strategy: whole-genome/exome and copy-number profiling at the bank level (MCB/WCB) to track genomic stability; targeted methylome/transcriptome panels in in-process intermediates to detect residual pluripotency signatures; and focused, low-cost qPCR and immunophenotyping in the final product [230, 231]. This development-to-release contraction aligns with current translational guidance and addresses tumorigenicity risks associated with residual undifferentiated cells and culture-acquired variants [232, 233].

Finally, time-series metabolomics/proteomics/lipidomics across pre-storage, storage, and post-storage stages can quantify how oxidation, carbonylation, Band-3 processing, and extracellular vesiculation relate to hemolysis, rheology, and immuno-hemostatic safety [219, 220]. The same datasets provide a principled way to benchmark iRBCs against donor RBCs and to distill robust release surrogates—again converging on ATP, 2,3-DPG, and a small panel of membrane and damage markers [219].

In summary, the combination of classical functional assays and integrated multi-omics strategies provides a comprehensive framework for evaluating iRBCs. Traditional morphological, biochemical, and biophysical measurements establish the foundation for assessing iRBC phenotype and performance, while multi-omics enables molecular-level insights into the regulatory programs governing erythroid differentiation, maturation, and functional fidelity. By linking molecular signals to key functional readouts—such as enucleation, hemoglobin switching, membrane stability, and oxygen transport capacity—multi-omics not only deepens mechanistic understanding but also facilitates the development of quantitative surrogate markers for process control and quality assurance. This integrated approach systematically supports the optimization of iRBC production, enhances batch-to-batch consistency, and strengthens translational readiness.

Clinical potential and remaining challenges

Clinical potential

Management of rare blood types with in vitro-generated red blood cells

Patients with rare blood types or complex alloimmunization profiles represent a key clinical niche for in vitro–generated RBCs. It remains particularly challenging for blood services to source compatible units for multiply alloimmunized patients with chronic transfusion-dependent anemia or rare blood types, even in healthcare systems with established rare donor registries and frozen inventories [214, 234, 235]. In this setting, ex vivo manufactured RBCs are envisioned not as a universal replacement for donor blood, but as a targeted option for patients for whom no alternative compatible product is available [213, 236, 237].

To address this unmet need, several groups have proposed banking pluripotent stem cell lines from carefully selected blood donors, including those with rare phenotypes. Park et al. established iPSC lines from peripheral blood mononuclear cells of donors with common O D-positive blood and two rare phenotypes, D − and Jr(a−), and differentiated them into erythroid cells using a 31-day serum-free, xeno-free protocol compatible with Good Manufacturing Practice (GMP) requirements [238]. The resulting cells progressed from hematopoietic progenitors to CD71⁺ early erythroid precursors and CD235a⁺ late erythroid cells while maintaining normal karyotypes and pluripotency marker expression, demonstrating the technical feasibility of building iPSC line banks from rare blood donors as a future resource for antigen-matched RBC production [12, 236, 238].

In the near term, a realistic clinical application of rare blood–derived iPSCs is the generation of diagnostic reagent RBCs rather than full transfusion units. Serologic evaluations in patients with rare blood group antigens are frequently constrained by the limited availability of appropriately typed panel cells. Several studies have shown that iPSC-derived erythroid cells expressing defined blood group antigens can be produced in relatively small numbers and used as customized reagent RBCs for antibody identification and compatibility testing [170, 173, 213]. Because the cell dose required for laboratory use is orders of magnitude lower than that required for therapeutic transfusion, this approach is compatible with current differentiation efficiencies and culture scales. Accordingly, iPSCs from rare blood group donors have been proposed as an essentially unlimited source of such reagent cells [173].

Beyond immediate diagnostic applications, iPSC-based erythroid systems are also being explored for broader transfusion uses. In the longer term, rare phenotypes and gene-edited lines are also being developed as templates for “universal” or multi-compatible RBC products. O-negative hiPSC lines have been selected and optimized for high-density erythroid differentiation and are explicitly pursued as a platform for universal donor RBC manufacture [239, 240]. In parallel, CRISPR/Cas9-mediated editing of major blood group genes, including ABO and Rh, has been used to generate iPSC derivatives with reduced antigen expression or altered blood group, with the goal of broadening transfusion compatibility [223, 241, 242].

Collectively, these advances have led recent reviews to highlight rare blood type management and multiply alloimmunized patients as priority indications in which the high manufacturing cost and complexity of in vitro–generated RBCs may still be clinically justifiable [214, 216, 237]. At the same time, key hurdles remain, including scalable production of fully enucleated, adult-type RBCs, detailed characterization of blood group antigen expression on cultured cells, stringent genomic and functional quality control of parental iPSC lines, and the establishment of robust GMP and regulatory frameworks—challenges consistently emphasized across contemporary erythroid differentiation and “blood pharming” literature [162, 213, 243, 244].

Disease modeling of inherited red blood cell disorders with in vitro–generated RBCs

iPSC–derived erythroid cells provide a versatile platform to model inherited RBC disorders in a human genetic context. By differentiating patient-specific iPSCs into erythroid progenitors and reticulocytes, disease-specific phenotypes can be reproduced in vitro and directly linked to defined mutations, while also allowing genetic correction and drug testing under controlled conditions [209, 245, 246].

Pyruvate kinase deficiency (PKD), congenital dyserythropoietic anemia (CDA) type IV, and sickle cell disease (SCD) collectively illustrate the range of inherited red blood cell disorders that can be modeled with in vitro–generated RBCs. In PKD, patient-derived iPSCs carrying PKLR mutations were used to generate erythroid cells that reproduced the characteristic energetic imbalance, while TALEN-mediated insertion of a codon-optimized R-type pyruvate kinase cDNA into the endogenous locus restored enzyme activity and ATP levels, demonstrating both disease recapitulation and functional rescue within the same platform [209]. In CDA type IV, iPSCs harboring the KLF1 E325K mutation differentiated into erythroid cells that showed hallmark features of dyserythropoiesis, and inducible expression of KLF1 E325K alone was sufficient to impose G1 arrest and broadly perturb expression of KLF1 target genes, clarifying how this single transcription factor mutation disrupts erythroid maturation [210]. More recently, an optimized three-stage, xeno-free differentiation protocol has enabled the generation of enucleated, β-globin–expressing RBCs from iPSCs derived from healthy donors and homozygous SCD patients; these induced RBCs display a GlyA⁺Band3⁺CD71^low phenotype, undergo hypoxia-induced sickling, and reveal dysregulated molecular pathways by transcriptome profiling, providing a functionally and molecularly informative model of SCD in vitro [211].

Together, these studies highlight the breadth of disease modeling that can be achieved with in vitro–generated RBCs: PKD models allow detailed interrogation and correction of metabolic defects, CDA type IV models dissect the consequences of erythroid transcription factor mutations on cell-cycle progression and membrane integrity, and SCD models reproduce hallmark functional phenotypes such as hypoxia-induced sickling in enucleated, adult hemoglobin–expressing cells [209–211]. At the same time, most iPSC-derived erythroid systems still display residual developmental immaturity and variable enucleation efficiency, underscoring the need for further optimization and standardization before they can serve as fully predictive platforms for therapeutic screening and translational studies [245, 246].

Drug discovery and screening with in vitro–generated red blood cells

In vitro–generated RBCs, especially those derived from human iPSCs, provide a renewable and genetically defined source of erythroid cells for pharmacologic studies. Patient-specific or gene-edited iPSC lines can be differentiated into hematopoietic progenitors and erythroid cells that recapitulate disease-relevant phenotypes, creating a platform on which small molecules can be tested in a human context [245, 247, 248]. Recent reviews on iPSC-based disease modeling consistently emphasize their potential for drug discovery and precision medicine in hematologic diseases, including disorders of erythropoiesis [249–251].

The most compelling proof-of-concept for an iPSC-based drug screen targeting an inherited red cell disorder comes from Diamond–Blackfan anemia (DBA). Doulatov et al. generated iPSCs from DBA patients, produced hematopoietic progenitors that reproduced the erythroid differentiation block, and then performed an unbiased small-molecule screen using this system [212]. They identified SMER28 as a compound that enhances erythropoiesis in DBA models; mechanistic work showed that SMER28 acts via the autophagy factor ATG5 to stimulate erythroid expansion and up-regulate globin gene expression, thereby nominating autophagy as a therapeutic pathway in DBA and establishing an iPSC-based chemical screen that yields a concrete candidate therapy. Complementary work in DBA used zebrafish and mammalian models with ribosomal protein deficiency to identify calmodulin inhibitors, including the U.S. Food and Drug Administration–approved drug trifluoperazine, which rescued hematopoietic phenotypes in zebrafish and mouse models and improved erythroid differentiation in human CD34⁺ cells from healthy donors and a DBA patient, illustrating how red cell–focused screens can progress across model systems toward human cells [252, 253].

Collectively, these studies demonstrate that in vitro–generated erythroid cells—especially iPSC-derived systems—can actively enable both unbiased and pathway-guided small-molecule screens, revealing actionable targets and candidate therapeutics for rare red cell disorders. At the same time, current iPSC-derived erythroid cells often retain a fetal or neonatal globin expression pattern, show incomplete enucleation, and remain technically demanding to scale, which limits their immediate use in very large high-throughput campaigns [244, 247, 254]. As differentiation protocols and culture technologies continue to improve, drug discovery and screening for inherited anemias—exemplified by DBA—are likely to remain among the most feasible near-term translational applications of in vitro–generated RBCs, bridging the gap between mechanistic insight and candidate therapeutics in a human erythroid context [245, 247, 255].

Toward transfusion-grade production: remaining challenges for in vitro–generated RBCs

Scaling to clinically relevant volumes remains the central bottleneck for in vitro–generated RBCs. Most iPSC-erythropoiesis pipelines still struggle with limited volumetric productivities and high cost of goods. Also, significant batch-to-batch variability during mesoderm induction, erythroid amplification, and terminal maturation/enucleation are also observed [213, 256, 257]. Recent analyses outline integrated upstream–downstream strategies (high-density expansion, perfusion/bioreactor control of oxygen and shear, process analytics) for large-scale production of transfusion-ready RBCs, but also emphasize that yields and enucleation efficiencies must increase further before routine clinical manufacturing is realistic [256].

Downstream handling and storage constitute a second set of hurdles. Although standard cold storage at 1–6 °C is optimized for donor RBCs, it remains to be determined whether cultured counterparts display comparable storage lesion kinetics, oxidative fragility, and membrane remodeling profiles [258]. Recent work highlights innovation in short- and long-term preservation, including additive solution design and cryo/supercooling approaches, while supercooled storage at − 5 °C with antioxidant-supplemented solutions (e.g., Erythro-Sol 5 plus resveratrol/serotonin/melatonin/Trolox) has extended post-collection durability by mitigating time- and temperature-dependent oxidative injury [222, 258–260]. These advances define testable conditions for cultured RBCs but require head-to-head evaluation against donor standards and clinically relevant endpoints such as hemolysis, deformability, and post-transfusion survival.

A third challenge is biological variability tied to cell source. Epigenetic memory from the somatic cell of origin can persist in iPSCs, yet early comparative studies reported broadly similar hematopoietic/erythroid induction capacities across sources; more recent work systematically profiling hiPSC lines indicates that input cell types and hematopoietic progenitor sources still influence erythroid differentiation and maturation efficiencies [247], ]. For clinical translation, this implies the need for source selection criteria and release assays that capture functional potency, not just surface phenotype [244].

Variability also arises upstream at the reprogramming step. Reprogramming efficiency and resultant line quality depend on donor-specific factors (age, genetic background), the somatic cell type, and the reprogramming platform, introducing lot-to-lot differences that propagate into erythroid performance [247, 263]. Manufacturing-oriented reviews recommend mitigating strategies—line qualification gates, parallel clone selection, and platform standardization—to reduce variance before erythroid induction [263, 264].

Finally, sex-linked effects may modulate pluripotent and erythroid phenotypes. Human pluripotent stem cells display sex-dependent transcriptional programs (including SRY expression in male hPSCs and >200 autosomal differentials), with evidence that endocrine cues can bias lineage responses; whether and how these programs impact quantitative metrics of iPSC-erythropoiesis (e.g., proliferation, enucleation, globin switching) warrants targeted study in male/female isogenic systems [265].

Together, these constraints—industrial-scale yield and enucleation, preservation compatible with transfusion practice, source- and reprogramming-driven variability, and potential sex effects—define a tractable roadmap for de-risking iPSC-derived RBC manufacturing: qualify lines up front, standardize bioprocess inputs, stress-test storage formulations on cultured cells with donor-grade readouts, and incorporate sex as a biological variable in design and analysis. Addressing these points systematically should narrow the gap between current lab-scale protocols and robust, compliant production of transfusion-grade RBCs.

Conclusions

iRBCs represent a pragmatic alternative to donor-dependent transfusion systems, with particularly strong clinical impact in areas of substantial unmet need such as rare blood groups and highly sensitized patients. Despite clear gains in manufacturability over the past decade through process/culture optimization and scale-up, key hurdles remain: achieving therapeutic-unit yields at acceptable cost, securing batch-to-batch uniformity of functional maturation, and establishing a pluripotent stem cell–specific safety framework.

Here, the value of multi-omics lies not in replacement but in quantitative complementarity. With classical biochemical and biophysical assays maintained as the standard, multi-omics enables numerical benchmarking of iRBC maturation and functional fidelity and supports targeted control of process variables. In development, broad discovery-scale omics can nominate candidate readouts; at release, these can be compressed into a small surrogate panel—such as ATP, 2,3-DPG, and membrane/damage markers—constituting a practical two-tier strategy. This approach acts across the four critical axes of scale-up, functional uniformity, safety qualification, and storage/logistics, thereby sequentially relieving translational bottlenecks. Furthermore, by enabling the concurrent interrogation of multiple pathways, multi-omics approaches expand the experimental scope and provide integrative insights that are not attainable through single-layer analyses.

By sharpening product definition and process control through multi-omics, the value proposition of iRBCs clearly extends well beyond transfusion replacement. Scalable, closed-system bioreactors enable mass cultivation of pathogen-minimized products suited for safer transfusion and a reduced risk of infection, while standardized erythroid cultures support diagnostic applications and drug testing aimed at improving enucleation, globin switching, and membrane biomechanics. Strategic reserves produced in this way can reinforce backup blood supply for emergency and military settings, and patient-specific lines open paths to personalized therapy for individuals with rare blood types or complex alloimmunization (Fig. 4). Coupled with interdisciplinary alignment in regulatory science, ethics, and public health—and guided by multi-omics–justified surrogate metrics within GMP-compliant processes—these trajectories bring the clinical realization of iRBCs appreciably closer.

Fig. 4.

Clinical application potential of iPSC-derived red blood cells (iRBCs)

Collectively, these elements position iRBCs on a credible path toward clinical adoption, provided that safety, quality, and scalability continue to advance in concert.

Moreover, the value proposition of iRBCs reaches well beyond transfusion replacement: closed, scalable bioreactors enable pathogen-minimized products for safer transfusion, while standardized erythroid cultures empower diagnostics and drug testing aimed at improving enucleation, globin switching, and membrane biomechanics. Strategic reserves can reinforce emergency and military supply, and patient-specific lines open paths to personalized therapy for rare blood types or complex alloimmunization.

Codified within GMP workflows, this two-tier, multi-omics–enabled framework furnishes the operational backbone for translation. Coupled with interdisciplinary alignment in regulatory science, ethics, and public health—and guided by multi-omics–justified surrogate metrics within GMP-compliant processes—these trajectories bring the clinical realization of iRBCs appreciably closer. Collectively, these insights establish not only a synthesis of current progress but also a novel translational framework that operationalizes multi-omics as a quantitative bridge from discovery to clinical deployment.

Abbreviations

HSCs

Hematopoietic stem cells

iPSCs

Induced pluripotent stem cells

AGM

Aorta-gonad-mesonephros

EHT

Endothelial-to-hematopoietic transition

EPO

Erythropoietin; EPOR, EPO receptor

MEP

Megakaryocyte-erythroid progenitors

BFU-E

Burst-forming unit erythroid

CFU-E

Colony-forming unit erythroid

EBI

Erythroblastic island

iRBCs

RBCs generated from human iPSCs

Author contributions

Y.J. conceptualized the review, wrote the manuscript, and prepared the figures. Y.J.E. contributed to the writing of the main text. S.-H.G. provided critical revision of the manuscript for important intellectual content. Y.K. and J.H. contributed to revising and refining the manuscript by reviewing the draft and assisting in the identification and selection of key references. Y.N., Y.A.R., and J.H.J. provided conceptual advice and critical revision of the manuscript. J.H.J. and Y.A.R. supervised the overall direction of the work. All authors reviewed and approved the final version of the manuscript.

Funding

This work was supported by the Multi-Ministerial Research Project, Republic of Korea, under the grant for the Manufacturing Human Cell-based Artificial Blood and Platform Technology Development for Transfusion (grant number: RS-2023-KH142779). This research was supported by a grant of Korean ARPA-H Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health &Welfare Republic of Korea (grant number: RS-2024-00512348). This research was also supported by Basic Medical Science Facilitation Program through the Catholic Medical Center of the Catholic University of Korea funded by the Catholic Education Foundation. This research was supported by Basic Medical Science Facilitation Program through the Catholic Medical Center of the Catholic University of Korea funded by the Catholic Education Foundation. AID center.

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.