Establishment of an immortalized erythroid model for Hb Bart’s hydrops fetalis with homozygous α⁰-thalassemia Southeast Asian deletion

Abstract

Hemoglobin Bart’s hydrops fetalis syndrome (BHFS) is the most severe form of α-thalassemia and is highly prevalent in Southeast Asia. It is most commonly caused by homozygosity for the α⁰-thalassemia Southeast Asian deletion (--^SEA/--^SEA), which eliminates α-globin production and results in profound fetal anemia. Current management is limited to intrauterine transfusions with only transient benefit, and no pharmacological or gene-based therapies are available. Progress toward therapy has been hampered by the lack of disease-relevant in vitro models. Here, we established Chery-Bart’s, an immortalized erythroid progenitor cell line derived from umbilical cord blood CD34⁺ hematopoietic stem and progenitor cells of a fetus with --^SEA/--^SEA and β^{IVS I−1 (G> T)}/β using a tetracycline-inducible HPV16 E6/E7 system. Chery-Bart’s cells exhibit sustained proliferation beyond the Hayflick limit, stable recovery after cryopreservation, and enrichment in the basophilic erythroblast stage. In vitro erythropoiesis assays showed that the cell line retains key features of the parental cells, including the capacity for erythroid maturation, expression of fetal globin transcripts, and robust production of Hb Bart’s (γ₄), the hallmark pathological hemoglobin of BHFS. Although Chery-Bart’s carries a β-thalassemia mutation, this is unlikely to modify disease severity, as γ-globin predominates in this model. Collectively, Chery-Bart’s represents the first durable and disease-specific cellular model of BHFS with --^SEA/--^SEA genotype. This platform provides a scalable resource for mechanistic studies and preclinical evaluation of therapeutic approaches, including drug discovery, gene therapy, and α-like globin chain induction. Beyond BHFS, Chery-Bart’s may also serve as a model for investigating severe α-thalassemia syndromes associated with the --^SEA allele, such as deletional and non-deletional HbH disease.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-54128-5.

Introduction

α-Globin is a critical component of hemoglobin in red blood cells and is encoded by two functional genes, the α2-globin gene (HBA2) and the α1-globin gene (HBA1). Together, these constitute the αα genotype and reside within the α-globin gene cluster on chromosome 16. α-Thalassemia is among the most common inherited hematologic disorders affecting approximately 5% of the global population. It is most often caused by large deletions within the α-globin gene cluster, leading to the loss of one (-α) or both (--) α-globin genes^1–4. These genetic defects lead to a reduction or complete absence of α-globin chain production, classified as α⁺-thalassemia (α-thalassemia 2) and α⁰-thalassemia (α-thalassemia 1), respectively.

Southeast Asian (--^SEA) deletion is a well-characterized structural variant that removes both cis α-globin genes (GRCh38.p14; NC_000016.10:g.165397_184700del) and represents the most prevalent α⁰-thalassemia genotype worldwide. Its highest frequencies are observed in Southeast Asia, particularly in southern China and Thailand^5–9. Individuals heterozygous for --^SEA deletion (--^SEA/αα) are typically asymptomatic; however, compound heterozygosity with other α-thalassemia mutations can result in clinically significant and sometimes life-threatening disorders, including deletional HbH disease (--^SEA/-α) and non-deletional HbH disease (--^SEA/α^Tα). Homozygosity for --^SEA deletion (--^SEA/--^SEA) causes the most severe form of α-thalassemia, characterized by complete absence of α-globin chain production and leading to profound anemia with lethal complications due to the inability to synthesize functional hemoglobin containing α-globin chains^1,4.

Fetal hemoglobin (HbF; α₂γ₂) is the predominant hemoglobin in red blood cells during fetal development. In the complete absence of α-globin production, HbF levels decrease, leading to severe fetal anemia, while excess unpaired γ-globin chains form homotetramers known as hemoglobin Bart’s (Hb Bart’s; γ₄)^10,11. These tetramers possess abnormally high oxygen affinity and are unable to release oxygen effectively to tissues, resulting in profound hypoxia¹². The combination of fetal anemia and hypoxia triggers massive compensatory responses, including increased cardiac output and marked extramedullary erythropoiesis in the liver and other organs. Despite these adaptations, compensation ultimately fails, leading to multi-organ dysfunction and the development of hydrops fetalis characterized by generalized edema and fluid accumulation in the peritoneal, pleural, and pericardial cavities. This clinical presentation is known as Hb Bart’s hydrops fetalis syndrome (BHFS)^1,13. Without intervention, most affected fetuses die in utero or within hours after birth. BHFS is most often caused by --^SEA/--^SEA genotype and is estimated to affect several thousand pregnancies worldwide each year^3,8,13,14.

As with many inherited disorders, the gold standard for preventing and controlling severe α-thalassemia is genetic counseling in combination with prenatal diagnosis (PND). Preimplantation genetic diagnosis (PGD) can also be applied to avoid termination of affected pregnancies and related complications, offering advantages over PND. Nonetheless, new cases continue to be identified, particularly in regions where --^SEA allele is highly prevalent. Despite advances in prenatal diagnostics and the development of intrauterine interventions, BHFS remains a lethal condition with few therapeutic options. Intrauterine transfusion can markedly improve perinatal survival by temporarily correcting fetal anemia; however, its efficacy depends on early diagnosis and access to specialized care, and most survivors remain transfusion dependent^15,16. Postnatal management requires lifelong red blood cell transfusions with concurrent iron chelation therapy, which carries the risk of iron overload and related complications. At present, hematopoietic stem cell transplantation is the only curative treatment, but its use is limited by donor availability, transplant-related morbidity, and the inability to reverse organ damage resulting from early-life hypoxia^4,13.

Alternative strategies, including the reactivation of exogenous α-globin genes or the embryonic α-like ζ-globin gene (HBZ), have shown promise in preclinical studies as potential curative approaches by restoring functional hemoglobin production^17–20. These approaches remain confined to laboratory research and early clinical investigations (ClinicalTrials.gov ID NCT05851105 and NCT05757245), with no official results reported to date. Among currently available in vitro systems, the human umbilical cord blood-derived erythroid progenitor 2 (HUDEP2) cell line is widely used to study the molecular pathology of hemoglobinopathies and to evaluate emerging therapeutic strategies^17,20–22. HUDEP2 utilizes a doxycycline-inducible Tet-on system, enabling controlled transitions between continuous proliferation and erythroid differentiation²³. This cell line predominantly expresses adult globin chains, making it representative of adult hemoglobin profiles. HUDEP2 cells have been genetically engineered to harbor α-globin gene deletions that model various α-thalassemia genotypes and have been employed to investigate the therapeutic potential of exogenous α-globin and ζ-globin supplementation^17,20. Complementary to these in vitro systems, an adult mouse model of homozygous α⁰-thalassemia has been developed via conditional knockout of α-globin genes and used to evaluate α-globin gene addition therapies¹⁹. While these models have provided important mechanistic insights and informed therapeutic development, they rely on genetically modified wild-type cells or animals and may not fully replicate the native cellular environment or the complex pathophysiology observed in naturally occurring BHFS.

To overcome the limitations of existing models, we established a novel cellular model, Chery-Bart’s. This Chiang Mai immortalized erythroid progenitor cell line was derived from umbilical cord blood CD34⁺ cells of a fetus with BHFS carrying --^SEA/--^SEA genotype. The model was designed to recapitulate the biological properties of erythroid progenitors in BHFS. Comparative in vitro erythropoiesis assays demonstrated marked ineffective erythropoiesis, consistent with the phenotypic and developmental abnormalities observed in the parental BHFS-derived CD34⁺ cells from the same --^SEA/--^SEA fetus prior to immortalization. These results establish Chery-Bart’s as a physiologically relevant, disease-specific, and sustainable model for mechanistic studies of BHFS resulting from the --^SEA/--^SEA genotype. In addition, this platform may hold promise as a resource for high-throughput drug screening and the development of targeted therapeutic approaches for severe α-thalassemia diseases.

Results

Stable erythroid progenitor cell line derived from BHFS with --^SEA/--^SEA is generated and exhibits sustained proliferation

CD34⁺ hematopoietic stem/progenitor cells (HSPCs) were isolated from the umbilical cord blood of three stillborn fetuses diagnosed with BHFS due to --^SEA/--^SEA and transduced with a lentiviral vector encoding a tetracycline-inducible HPV16 E6/E7 expression cassette (Fig. 1A and Supplementary Figure S1). Of the three independent transduction attempts using unrelated donor samples, only the culture derived from a 31-week-gestation male fetus with --^SEA/--^SEA and β^{IVS I−1 (G>T)}/β achieved successful immortalization, leading to the establishment of a stable erythroid progenitor cell line. Following transduction, integration of the E6/E7 oncogene cassette was confirmed with a vector copy number (VCN) of 0.8 ± 0.1 (Fig. 1B), and transcripts were readily detected (Fig. 1C, D and Supplementary Figure S2). Robust proliferation became evident around day 40 and continued to accelerate beyond day 140 in expansion medium (Supplementary Figure S3A). Cell viability, initially approximately 50% during the early immortalization phase, increased to about 80% after day 140, correlating with reduced apoptosis in parallel assays (Supplementary Figures S3B, C). Cells were cryopreserved at regular intervals throughout the expansion phase (Supplementary Table S1). In contrast, cultures maintained without doxycycline exhibited marked proliferative impairment and complete cell death by day 37 (Supplementary Figures S3A, B). At the time of manuscript preparation, the cell line had undergone continuous proliferation for five months, exceeding 70 passages and surpassing the Hayflick limit²⁴, while maintaining consistent recovery after freeze-thaw cycles. The cell line, designated Chery-Bart’s (Chiang Mai immortalized erythroid progenitor cell line from a BHFS fetus with --^SEA/--^SEA genotype), displayed a mean doubling time of approximately 30 h after day 140 in continuous culture. Karyotype analysis revealed chromosomal abnormalities acquired during immortalization (Supplementary Figure S4); however, thalassemia genotypes were retained (Fig. 1E and Supplementary Figure S5,6), and erythroid lineage identity was preserved. Morphological analysis showed a predominance of basophilic erythroblasts, which persisted throughout the immortalization process (Fig. 1F). Flow cytometric profiling demonstrated stable expression of erythroid-specific markers CD36, CD71, and CD235a (Fig. 1G, H and Supplementary Figure S7), confirming their identity as erythroid progenitors. Collectively, these findings establish Chery-Bart’s as a disease-specific, HPV16 E6/E7-mediated immortalized erythroid progenitor cell line with sustained proliferation and preserved erythroid characteristics.

Fig. 1.

Generation of Chery-Bart’s cell line. (A) Schematic illustration of the immortalization workflow. (B) Vector copy number (VCN) per diploid cell in Chery-Bart’s cells at day 32 and day 145 compared with untransduced controls. Expression of HPV16 E6 (C) and E7 (D) mRNA in Chery-Bart’s cells at day 32 and day 145 compared with control cells at day 12 (transduced cells were cultured in expansion medium without doxycycline). (E) Multiplex gap-PCR analysis for four common α-thalassemia deletions (-α^3.7, -α^4.2, --^SEA, and --^THAI) in donor cord blood and Chery-Bart’s at day 141. M, GeneRuler 1 kb Plus DNA Ladder. (F) Morphological features of Chery-Bart’s cells during immortalization in expansion medium. Scale bars = 50 μm. (G) Representative flow cytometry histograms showing the percentage of positive cells for each marker. (H) Hematopoietic cell markers analyzed by flow cytometry in Chery-Bart’s cells at day 141. Bar graphs represent mean ± SD of triplicates. ***p < 0.001.

Chery-Bart’s cell line recapitulates pathological erythropoiesis of primary BHFS erythroblasts

To evaluate the erythroid differentiation characteristics of the Chery-Bart’s cell line in comparison with its parental BHFS-derived CD34⁺ cells obtained from the same subject prior to immortalization (hereafter referred to as parental cells), both were subjected to in vitro differentiation using established erythroid differentiation media. In parental cells, robust proliferation was observed during phase II of culture (days 7–11), followed by a decline in total cell numbers during phase III (day 11 onward) (Fig. 2A). Cell viability also decreased significantly during differentiation relative to day 0; however, overall viability remained above 80% (Fig. 2B), consistent with the modest increase in apoptotic cell death detected by the apoptosis assay (Supplementary Figure S8A). Under the three-phase erythroid culture conditions, enucleated erythroid cells typically appear by phase III of culture in non-thalassemic controls²⁵. In contrast, differentiated parental cells exhibited delayed maturation, as demonstrated by flow cytometry and morphological analyses (Fig. 2C, D), consistent with ineffective erythropoiesis described in various thalassemia genotypes²⁶ and indicative of impaired erythroid maturation in BHFS with the --^SEA/--^SEA genotype. Compared to the parental cells, Chery-Bart’s displayed an even more pronounced dyserythropoiesis. Upon induction with erythroid differentiation medium, Chery-Bart’s cells exhibited an initial proliferative phase during phase I of culture (days 0–4), followed by a gradual reduction in expansion (Fig. 2E). A marked decrease in cell viability was observed during differentiation, particularly from phase II onward (day 4 onward) (Fig. 2F), which corresponded with the pronounced increase in apoptotic cell death detected by apoptosis assay (Supplementary Figure S8B). Cell morphology during differentiation of Chery-Bart’s cells demonstrated clear erythroid maturation, as confirmed by both morphological assessment and flow cytometry (Fig. 2G, H). Notably, differentiated Chery-Bart’s cells at days 4–8 expressed levels of the erythroid markers CD71 and CD235a comparable to those observed in differentiated parental cells at days 11–16 (Fig. 2D, H). However, the morphology of differentiated Chery-Bart’s cells appeared more aberrant than that of the differentiated parental cells (Fig. 2C, G). This discrepancy may reflect differences in the initial developmental stage of the starting populations, as Chery-Bart’s cells were primarily in the basophilic erythroblast stage prior to induction of differentiation (Fig. 1F), together with the accumulation of chromosomal abnormalities during immortalization (Supplementary Figure S4). E6 and E7 expression progressively declined despite continuous doxycycline exposure (Supplementary Figure S9), likely reflecting nuclear inactivation associated with erythroid maturation, as supported by both morphological and flow cytometric analyses (Fig. 2G, H). Collectively, these kinetics support an 8-day differentiation window as optimal for downstream studies, since extended culture led to substantial cell death and accumulation of cellular debris, which could confound experimental outcomes. Although the heightened dyserythropoiesis in Chery-Bart’s may partially result from immortalization-associated changes, the cell line recapitulates key differentiation defects observed in the parental cells.

Fig. 2.

Erythroid differentiation of parental CD34⁺ HSPCs and Chery-Bart’s cell line. (A) Cell proliferation and (B) viability of parental CD34⁺ HSPCs during differentiation in three-phase erythroid culture medium. (C) Cell morphology and (D) flow cytometry analysis of parental CD34⁺ HSPCs during differentiation in three-phase erythroid culture medium. Scale bars = 50 μm. (E) Cell proliferation and (F) viability of Chery-Bart’s cells derived from day 141 during differentiation in three-phase erythroid culture medium. (G) Cell morphology and (H) flow cytometry analysis of Chery-Bart’s cells during differentiation. Scale bars = 50 μm. Line graphs represent mean ± SD of triplicates with data compared to day 0 (prior to induction of differentiation). *p < 0.05; **p < 0.005; ***p < 0.001.

Chery-Bart’s cell line undergoes hemoglobinization following erythroid differentiation

Hb Bart’s is the hallmark pathological hemoglobin predominantly found in the red blood cells of fetuses affected by BHFS. To assess the capacity of the Chery-Bart’s cell line to produce Hb Bart’s, cells were induced to undergo erythroid differentiation for eight days, followed by hemoglobin analysis. During the expansion phase, Chery-Bart’s cell pellets appeared white to pale pink (Fig. 3A). Upon induction, the pellets progressively acquired a red coloration, with a clear change by day 8 and the most intense hue observed at this time point, indicating ongoing hemoglobinization (Fig. 3A). Absolute mRNA expression levels of six globin genes, including ζ-globin (HBZ), α-globin (HBA), ε-globin (HBE), γ-globin (HBG), δ-globin (HBD), and β-globin (HBB), were quantified throughout Chery-Bart’s differentiation (Fig. 3B-G). While HBA expression was undetectable, HBG and HBZ were predominantly expressed, and HBB, HBE, and HBD were also detected during differentiation. This globin profile corresponds to a fetal-like pattern characteristic of BHFS caused by --^SEA/--^SEA genotype. To confirm hemoglobin production, differentiated cell pellets were analyzed by ion-exchange high-performance liquid chromatography (HPLC) and a monoclonal antibody-based assay specific for human Hb Bart’s. The chromatographic profile of differentiated Chery-Bart’s cells closely resembled that of donor-matched cord blood (Fig. 3H), and Hb Bart’s was clearly detected (Fig. 3I). These results demonstrate that the Chery-Bart’s cell line reproduces the pathological hemoglobinization of BHFS, supporting its utility as a robust and physiologically relevant in vitro model for disease investigation.

Fig. 3.

Globin and hemoglobin analysis during Chery-Bart’s differentiation. (A) Hemoglobinization of cell pellets during differentiation of Chery-Bart’s cells derived from day 141. Globin gene expression during differentiation of Chery-Bart’s cells derived from day 141, including (B) HBZ, (C) HBA, (D) HBE, (E) HBG, (F) HBD, and (G) HBB. Bar graphs represent mean ± SD of triplicates with data compared to day 0 (prior to induction of differentiation). u.d., undetectable. *p < 0.05. (H) Representative chromatograms of hemoglobin analysis from donor cord blood and differentiated Chery-Bart’s cells at day 8. (I) Immunochromatographic strip test for detection of Hb Bart’s in Chery-Bart’s cells before (day 0) and after (day 8) differentiation.

Discussion

To date, most curative strategies for α-thalassemia remain at the preclinical stage, primarily due to the absence of suitable disease models for evaluating drug efficacy and refining genetic modification platforms. The lack of sustainable, disease-relevant models continues to represent a major obstacle to therapeutic advancement in α-thalassemia. Primary cells are often regarded as the gold standard for in vitro erythroid studies because of their close resemblance to native cellular functions²⁷. However, their inherently limited lifespan remains a major obstacle to long-term experimentation and large-scale applications. In contrast, immortalized cell lines have become indispensable tools in hematopoietic research, offering scalable, reproducible, and experimentally tractable platforms for investigating cellular and molecular mechanisms. Among these, the HUDEP series has become a cornerstone in erythroid biology. HUDEP cells are generated from CD34⁺ HSPCs isolated from normal human umbilical cord blood and immortalized using a tetracycline-inducible HPV16 E6/E7 system²³. They express key erythroid lineage markers and provide several advantages over other hematopoietic cell lines, including sustained proliferative capacity, efficient enucleation upon induction of terminal differentiation, and robust globin gene expression. Currently two HUDEP cell lines are widely used. HUDEP1 predominantly expresses γ-globin and exhibits a fetal hemoglobin profile, while HUDEP2 primarily expresses β-globin consistent with adult hemoglobin production²³. These cell lines have greatly advanced the understanding of erythropoiesis and have facilitated translational research, particularly in thalassemia. Disease-mimicking HUDEP2 sublines have been generated through targeted genetic modifications^17,21,22,28; however, these engineered models may not fully capture disease-specific phenotypes, especially when evaluated across integrated multi-omics datasets. This underscores the need for physiologically relevant models that more reproduce the complex pathophysiology of hematologic disorders.

Inheritance of --^SEA gives rise to various α-thalassemia genotypes, including HbH disease and BHFS, the latter constituting the lethal phenotype^4,13. Previous attempts to develop cellular models of --^SEA/--^SEA genotype have been limited. One approach generated an induced pluripotent stem cell (iPSC) line from skin fibroblasts of a fetus carrying --^SEA/--^SEA genotype using a non-integrative reprogramming method²⁹. The iPSCs were differentiated into hemoglobin-producing cells through co-culture with OP9 stromal cells and were found to express predominantly fetal globin transcripts. However, their erythroid differentiation potential and hemoglobin composition were not further examined. Another model was created by CRISPR/Cas9-mediated deletion of all functional α-globin genes in HUDEP2 cells to mimic --^SEA/--^SEA genotype¹⁷. Although this engineered line replicates the genetic defect, it retains the adult globin expression profile of HUDEP2 and lacks detailed characterization of erythroid maturation and hemoglobin production. These shortcomings highlight the need for a physiologically relevant erythroid model that reproduces the molecular and cellular features of BHFS caused by --^SEA/--^SEA.

In this study, we established the Chery-Bart’s cell line by immortalizing CD34⁺ HSPCs from a fetus with BHFS resulting from --^SEA/--^SEA genotype. Chery-Bart’s cells display erythroid lineage characteristics, including expression of CD36, CD71, and CD235a, and are predominantly arrested at the basophilic erythroblast stage. These phenotypic traits have remained stable during continuous expansion for five months at the time of manuscript preparation. Importantly, the cell line retains the ability to undergo erythroid maturation in response to differentiation medium. Differentiated Chery-Bart’s cells reproduce key phenotypes of the parental cells, including ineffective erythropoiesis, a fetal globin expression profile, and production of Hb Bart’s, the pathognomonic hemoglobin in BHFS. In addition to --^SEA/--^SEA genotype, Chery-Bart’s cells also carry a β⁰-thalassemia mutation, IVS I-1 (G > T) [HBB:c.92 + 1G > T]. In general, β-globin is the predominant β-like globin synthesized in adult human erythroid cells, and its production is normally balanced with that of α-globin. In adults with α-thalassemia, reduced α-globin synthesis leads to an excess of unpaired β-globin chains, which tend to precipitate and cause erythroid cell damage. Therefore, adults with α-thalassemia who also harbor β-thalassemia mutations may theoretically present with milder clinical manifestations due to a relative reduction in β-globin excess. However, during the fetal stage, γ-globin rather than β-globin is the predominant β-like globin. Consequently, the co-inheritance of β-thalassemia is unlikely to significantly influence erythroid physiology in α-thalassemia during fetal life. This is consistent with our findings, as the major globin expressed in the Chery-Bart’s cell line, which was derived from fetal cord blood, is γ-globin. Based on these observations, we believe that the co-inherited β-thalassemia mutation does not affect the cellular phenotype of this model. Furthermore, the presence of the β-thalassemia mutation did not ameliorate the clinical severity of BHFS, as evidenced by the intrauterine death of the affected fetus.

Chery-Bart’s cells were established using a tetracycline-inducible system to drive HPV16 E6 and E7 expression, an established method for immortalizing erythroid progenitor cells used in other models^23,30,31. In this system, transcription of E6 and E7 is controlled by a tetracycline-responsive element (TRE), activated when the reverse tetracycline-controlled transactivator (rtTA) binds in the presence of tetracycline or its derivatives, such as doxycycline. The oncogenic activity involves E6-mediated degradation of the tumor suppressor p53, which inhibits apoptosis, and E7-driven inactivation of the retinoblastoma protein (pRb), thereby promoting cell cycle progression³². Following lentiviral transduction, HPV16 E6/E7 integration was detected within a month, with expression of both transcripts. Molecular characterization indicated that Chery-Bart’s cells harbor approximately one copy of the HPV16 E6/E7 cassette per genome and predominantly express the E6* isoform and E7. This profile mirrors that of HUDEP cell lines, suggesting that E6*, in combination with E7, is critical for immortalization. This is consistent with HPV-mediated carcinogenesis, where E6* and E7 confer a proliferative advantage and are frequently detected in cervical carcinoma cell lines and cervical intraepithelial neoplasia lesions³³. Despite this, the initial expansion and viability of the transduced cells remained limited. A marked increase in proliferative capacity and cell viability was observed after day 140, suggesting the acquisition of additional genetic or epigenetic alterations that enabled the cells to bypass the Hayflick limit and achieve immortalization. Another additional possible factor that would help the immortalization is having aneuploidy which is a common feature of human cancer presenting in many solid tumors and hematopoietic cancers. Consistent with findings in certain hematopoietic malignancies and other established hematopoietic cell lines^34–37, Chery-Bart’s cells exhibited an abnormal karyotype, indicating a degree of genomic instability. These findings support the notion that, while E6 and E7 expression is essential, it may not be solely sufficient to induce immortalization in this model. The contribution of additional events, such as chromosomal aberrations, may play a synergistic role in facilitating the immortalization process in Chery-Bart’s cell line.

Differentiation of Chery-Bart’s cells was induced by culturing them in erythroid differentiation media containing defined factors that support maturation. Our results indicate that doxycycline primarily promotes cellular proliferation with minimal impact on differentiation. Chery-Bart’s cells retained the ability to undergo erythroid maturation when doxycycline was maintained throughout both phase I and phase II of differentiation. Notably, E6 and E7 expression declined spontaneously during erythroid maturation, even in the continued presence of doxycycline. These findings suggest that the loss of mitotic activity in maturing erythroid cells, together with factors unique to the differentiation media rather than doxycycline alone, plays a more critical role in driving terminal erythroid maturation. During erythroid differentiation, both the parental cells and the Chery-Bart’s cell line exhibited delayed terminal maturation. In our experience, enucleated erythroid cells typically emerge by phase III of differentiation in non-thalassemic cells²⁵; however, this was markedly reduced in erythroid progenitors derived from BHFS caused by --^SEA/--^SEA genotype. This finding is consistent with the previous in vitro erythropoiesis study in both α- and β-thalassemia²⁶, supporting the presence of ineffective erythropoiesis likely driven by the imbalance in globin chain synthesis inherent to thalassemic cells. Notably, differentiated Chery-Bart’s cells displayed more pronounced ineffective erythropoiesis, particularly increased cell death, compared with their differentiated parental counterparts. This phenomenon may reflect the accumulation of aberrant cellular features acquired during the dynamic changes associated with immortalization. As observed in other established cell lines, atypical cellular properties particularly at high passage numbers may limit the physiological relevance of immortalized models^38,39. Therefore, comprehensive multi-omics analyses comparing genetically matched parental and immortalized cells are essential to delineate these differences. Another consideration is that Chery-Bart’s cells were established and characterized as a bulk population of immortalized cells, resulting in diverse karyotypic variants within the culture. Consequently, performing limiting dilution may be beneficial for isolating specific clones with defined karyotypes. Despite such limitations, established cell lines remain indispensable, offering robust and scalable platforms that continue to drive both mechanistic insights and translational advances in cellular and molecular hematology.

Beyond its relevance to BHFS, Chery-Bart’s cell line represents a valuable platform for investigating other α-thalassemia syndromes associated with --^SEA deletion, including severe HbH disease. Historically, HbH disease arising from --^SEA inheritance has been considered as relatively benign with most patients exhibiting moderate hypochromic microcytic anemia and mild clinical symptoms. However, growing clinical evidence and molecular data have revealed substantial heterogeneity in disease severity, with presentations ranging from mild anemia to life-threatening complications^40,41. Chery-Bart’s model offers a unique opportunity to explore therapeutic strategies, including genome editing approaches or interventions aimed at enhancing α-like globin expression to mitigate globin chain imbalance. Such approaches hold significant promise for improving clinical outcomes in severe α-thalassemia diseases.

In conclusion, we have established Chery-Bart’s cell line, the physiologically relevant and disease-specific human model of BHFS caused by --^SEA/--^SEA genotype. This stable erythroid progenitor line recapitulates the molecular and cellular hallmarks of the disease, overcoming the major limitations of existing engineered or finite-lifespan primary cell models. Moreover, Chery-Bart’s may provide a robust and scalable platform for high-throughput drug screening, mechanistic dissection of α-thalassemia pathophysiology, and precise genome editing. These applications have the potential to accelerate translation and offer significant promise for developing novel therapeutic strategies for α-thalassemia disorders associated with --^SEA allele, including BHFS and severe HbH disease.

Materials and methods

Thalassemia diagnosis

Genomic DNA was extracted from whole blood (300 µL) or cultured cells (2–5 × 10⁵ cells) using the Blood/Cell DNA Mini Kit (Geneaid, New Taipei City, Taiwan) according to the manufacturer’s instructions. DNA concentration was determined using the Qubit dsDNA BR Assay Kit and Qubit 4 Fluorometer (Invitrogen, Eugene, OR). Four common deletional α-thalassemia genotypes including α⁺-thalassemia 3.7 kb deletion (-α^3.7), α⁺-thalassemia 4.2 kb deletion (-α^4.2), α⁰-thalassemia Southeast Asian deletion (--^SEA), and α⁰-thalassemia Thai deletion (--^THAI) were detected by multiplex gap-PCR with specific primers (Supplementary Table S2) as previously described⁴². Breakpoint for the --^SEA deletion was confirmed by Sanger sequencing (1st BASE, Selangor, Malaysia) using a specific primer (Supplementary Table S2). The β-globin gene, including its promoter region, was amplified and sequenced by Sanger sequencing using specific primers (Supplementary Table S2) to identify β-thalassemia mutations.

Lentivirus production

HEK293T cells (5 × 10⁶) were transfected with 1.64 pmol of CSIV-TRE-HPV-E6/E7-UbC-KT plasmid (provided by the RIKEN BRC through the National BioResource Project of the MEXT, Japan; #RDB17719), 0.72 pmol of pMD2.G (Addgene #12259), and 1.30 pmol of psPAX2 (Addgene #12260) using polyethylenimine “Max” (PEI MAX; Polysciences, Warrington, PA) at a DNA-to-PEI ratio of 1:4. Viral supernatants were harvested at 48 and 72 h post-transfection, filtered through a 0.45 μm syringe filter, and concentrated overnight at 4 °C with constant rocking (60 rpm) in 10% (w/v) PEG-8000 and 0.3 M NaCl (pH 7.2). Lentiviral particles were pelleted by centrifugation at 1,500 × g for 1 h at 4 °C, resuspended in 1× Dulbecco’s phosphate-buffered saline (DPBS; Stemcell Technologies, Vancouver, Canada), snap-frozen in liquid nitrogen, and stored at -80 °C until use.

CD34⁺ HSPC purification

Heparinized umbilical cord blood (15 mL) was obtained from three independent fetuses diagnosed with BHFS due to --^SEA/--^SEA genotype, each with an Apgar score of 0. Umbilical cord blood mononuclear cells (UBMCs) were isolated by density gradient centrifugation using Lymphoprep (Stemcell Technologies). CD34⁺ HSPCs were subsequently purified from UBMCs using the CD34 MicroBead Kit and LS Columns (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s protocol. Purified CD34⁺ HSPCs were immediately processed for downstream applications.

Erythroid immortalization

Immortalization was performed as previously described²³ with minor modifications (Supplementary Figure S1). Briefly, freshly isolated CD34⁺ HSPCs were cultured for 48 h in prestimulation medium composed of X-VIVO 15 (04-418Q; Lonza, Walkersville, MD) supplemented with 100 ng/mL recombinant human stem cell factor (SCF; Peprotech, Rehovot, Israel), 100 ng/mL recombinant human Flt3 ligand (Flt3-L; Peprotech), 100 ng/mL recombinant human thrombopoietin (TPO; Peprotech), 2 mM GlutaMAX (Gibco, Grand Island, NY), and 100 U/mL penicillin with 100 µg/mL streptomycin (1× P/S; Gibco). Following prestimulation, 5 × 10⁵ cells were transduced with the lentivirus in the same prestimulation medium supplemented with 8 µg/mL polybrene (hexadimethrine bromide; MedChemExpress, Monmouth Junction, NJ) and cultured for an additional four days. Cells were then transferred to expansion medium consisting of StemSpan SFEM (#09650; Stemcell Technologies) supplemented with 2 mM GlutaMAX, 1× P/S, 50 ng/mL SCF, 3 U/mL Epoetin alfa (Eprex; Cilag AG, Schaffhausen, Switzerland), 1 µM dexamethasone (Dex; Sigma-Aldrich), and 1 µg/mL doxycycline hyclate (Dox; Sigma-Aldrich). Fresh expansion medium was replenished every 48–72 h, maintaining a seeding density of 2–3 × 10⁵ cells/mL and preventing the density from exceeding 1 × 10⁶ cells/mL throughout the immortalization process. Cells were periodically harvested and cryopreserved in freezing medium containing 10% (v/v) dimethyl sulfoxide (DMSO; Sigma-Aldrich) in fetal bovine serum (FBS) (SV3016003; Hyclone, Logan, UT) for downstream applications.

Karyotyping

At day 141 of expansion, chromosomal analysis of the Chery-Bart’s cell line was performed at the Human Genetics Laboratory, Department of Anatomy, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand. Thirty metaphase spreads were examined using standard G-banding.

Erythroid differentiation

CD34⁺ HSPCs were differentiated using a three-phase erythroid differentiation medium (EDM) as previously described⁴³ with minor modifications (Supplementary Figure S1). Briefly, EDM was prepared by supplementing Iscove’s Modified Dulbecco’s Medium (IMDM) (SH30259.02; HyClone) with 5% (v/v) male AB serum, 330 µg/mL human holo-transferrin (ProSpec-Tany, Ness Ziona, Israel), 10 µg/mL recombinant human insulin (ProSpec-Tany), 2 U/mL heparin (Leo, Ballerup, Denmark), 1× P/S, 2 mM GlutaMAX, and 3 U/mL Eprex. From days 0–7, cells were cultured in EDM supplemented with 50 ng/mL SCF, 5 ng/mL interleukin-3 (IL-3; Peprotech), and 1 µM Dex. From days 7–11, cultures were maintained in EDM with 50 ng/mL SCF. From days 11–16, cells were grown in EDM without additional supplements. Differentiation of immortalized erythroid cells was performed as previously described⁴⁴ with modifications (Supplementary Figure S1). From days 0–4, cells were cultured in EDM supplemented with 50 ng/mL SCF and 1 µg/mL Dox. From days 4–8, cells were maintained in EDM containing 1 µg/mL Dox, and from days 8–12, cultures were continued in EDM without supplementation. Throughout the differentiation period, cell density was maintained at 2–3 × 10⁵ cells/mL and kept below 1 × 10⁶ cells/mL to ensure optimal erythroid maturation.

VCN analysis

Standard curves were generated using Precision Melt Supermix (Bio-Rad) and 10-fold serial dilutions of either the CSIV-TRE-HPV-E6/E7-UbC-KT plasmid or a β-actin (ACTB)-containing plasmid, each amplified with specific primers (Supplementary Table S2). Absolute quantitative real-time PCR was then performed to determine the copy number of the rtTA sequence within the lentiviral expression cassette. Lentiviral vector integration was quantified relative to ACTB gene copy number, and results were expressed as VCN per diploid genome.

Cell proliferation and viability assessment by trypan blue exclusion

Cell suspensions were harvested and mixed with 0.4% trypan blue solution (Gibco) at the appropriate dilution. Cell counts were performed using a hemocytometer, with viable and nonviable cells distinguished by trypan blue exclusion. Cell proliferation was expressed as the fold increase in total cell number relative to the initial count on day 0 of culture, and viability was reported as the percentage of viable cells out of the total population.

Cell apoptosis assay

Apoptosis was assessed using the FITC Annexin V Apoptosis Detection Kit with propidium iodide (PI) (BioLegend, San Diego, CA), following a modified protocol at half the manufacturer’s recommended reaction volume. In brief, 1 × 10⁵ cells were pelleted, washed once with 1× DPBS, and resuspended in 100 µL of Annexin V Binding Buffer. FITC-conjugated Annexin V and PI were added (5 µL each), and the suspension was incubated for 15 min at room temperature in the dark. Samples were analyzed on a BD Accuri C6 Plus flow cytometer (BD Biosciences, San Jose, CA), and data were processed using FlowJo software version 11.0.0 (BD Life Sciences, Ashland, OR).

Immunophenotyping

The expression of hematopoietic surface markers CD34, CD36, CD45, CD71, and CD235a was analyzed by flow cytometry. Cells were stained with fluorochrome-conjugated anti-human antibodies including CD34 PerCP (clone 581; BioLegend), CD36 FITC (clone 5-271; BioLegend), CD45 PE (clone HI30; BioLegend), CD71 PE (clone CY1G4; BioLegend), and CD235a APC (clone HIR2; BioLegend). Briefly, 1 × 10⁵ cells were harvested, washed once with 1× DPBS, and resuspended in 100 µL of 1× DPBS. Antibodies were added according to the manufacturer’s instructions, and samples were incubated for 15 min at room temperature in the dark. Stained cells were analyzed using a BD Accuri C6 Plus flow cytometer, and data were processed using FlowJo software.

Morphological analysis

Cell suspensions (1 × 10⁵ cells) were harvested, washed once with 1× DPBS, and cytocentrifuged at 600 rpm for 4 min using a Cytospin 3 cytocentrifuge (Thermo Shandon, Cheshire, UK). Slides were air-dried at room temperature, fixed in absolute methanol, and stained with Wright-Giemsa. Erythroid morphology was assessed under a light microscope at 400× magnification (Leica DM750; Leica Microsystems, Heerbrugg, Switzerland), and representative images were captured for documentation.

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from 2 to 5 × 10⁵ cells using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. RNA concentration was determined using the Qubit RNA HS Assay Kit (Invitrogen) with a Qubit 4 Fluorometer. First-strand cDNA was synthesized from ≥ 200 ng of total RNA using the FastKing RT Kit with gDNase (Tiangen, Beijing, China) following the manufacturer’s protocol. qRT-PCR was performed on a CFX Opus 96 Real-Time PCR System (Bio-Rad, Hercules, CA) using the SensiFAST SYBR No-ROX Kit (Bioline, London, UK) and specific primers (Supplementary Table S2). Gene expression levels were normalized to GAPDH and quantified either relatively using the 2-^ΔΔCt method or absolutely using standard curves generated from plasmids containing the target gene sequence.

Hb analysis

Hb profiles were determined using the VARIANT II β-Thalassemia Short Program (Bio-Rad), which operates on the principle of ion-exchange high-performance liquid chromatography. Whole blood samples were analyzed by direct injection into the HPLC system following the manufacturer’s instructions. For cultured cells, at least 1 × 10⁶ cells were pelleted, resuspended in 500 µL of Wash/Diluent Solution (Bio-Rad), and thoroughly homogenized before injection into the system.

Hb Bart’s detection

Hb Bart’s (γ₄) was detected using a monoclonal anti-Hb Bart’s antibody in a lateral flow chromatographic immunoassay (i + LAB αTHAL; i + MED Laboratories, Bangkok, Thailand). Briefly, at least 1 × 10⁵ cells were harvested, washed with 1× DPBS, and resuspended in 100 µL of the supplied lysis buffer. The test strip was vertically immersed in the lysate for 5 minutes, and results were visually interpreted within 15 min.

Statistical analyses

Data are presented as mean ± standard deviation (SD) from three independent experiments. Statistical analyses were performed using an unpaired Student’s t-test in GraphPad Prism version 9.4.1 (GraphPad Software, Boston, MA). A p value < 0.05 was considered statistically significant.

Ethical approval

All protocols involving human subjects were approved by the Research Ethics Committee Panel 5, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand, in accordance with the Guideline for Good Clinical Practice (ICH GCP) and relevant international ethical guidelines, as well as applicable laws and regulations (Study code: PED-2566-0617). Samples were obtained after written informed consent was secured in accordance with ethical guidelines.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Thaw Naing Zin: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Data curation, Writing – Original draft preparation, Visualization, Funding acquisition. Phudit Jatavan: Conceptualization, Resources, Writing – Review and Editing. Nittaya Sakunpansa: Resources, Writing – Review and Editing. Pimlak Charoenkwan: Conceptualization, Resources, Writing – Review and Editing. Pinyaphat Khamphikham: Conceptualization, Methodology, Validation, Data curation, Writing – Original draft preparation, Visualization, Supervision, Project administration, Funding acquisition.

Funding

T.N.Z. was supported by CMU Presidential Scholarship (Ref. No. 8393(25)/1688), Master’s Degree Program in Medical Technology (International Program) at the Faculty of Associated Medical Sciences, Chiang Mai University (CMU), Chiang Mai, Thailand.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.