Evaluation of 2D and 3D Erythroid Differentiation Protocols Using Sickle Cell Disease and Healthy Donor Induced Pluripotent Stem Cells

Background: Sickle cell disease (SCD) is a highly prevalent genetic disease caused by a point mutation in the HBB gene, which can lead to chronic hemolytic anemia and vaso-occlusive events. Patient-derived induced pluripotent stem cells (iPSCs) hold promise for the development of novel predictive methods for screening drugs with anti-sickling activity. In this study, we evaluated and compared the efficiency of 2D and 3D erythroid differentiation protocols using a healthy control and SCD-iPSCs. Methods: iPSCs were subjected to hematopoietic progenitor cell (HSPC) induction, erythroid progenitor cell induction, and terminal erythroid maturation. Differentiation efficiency was confirmed by flow cytometry analysis, colony-forming unit (CFU) assay, morphological analyses, and qPCR-based gene expression analyses of HBB and HBG2. Results: Both 2D and 3D differentiation protocols led to the induction of CD34+/CD43+ HSPCs. The 3D protocol showed good efficiency (>50%) and high productivity (45-fold) for HSPC induction and increased the frequency of BFU-E, CFU-E, CFU-GM, and CFU-GEMM colonies. We also produced CD71+/CD235a+ cells (>65%) with a 630-fold cell expansion relative to that at the beginning of the 3D protocol. After erythroid maturation, we observed 95% CD235a+/DRAQ5- enucleated cells, orthochromatic erythroblasts, and increased expression of fetal HBG2 compared to adult HBB. Conclusion: A robust 3D protocol for erythroid differentiation was identified using SCD-iPSCs and comparative analyses; however, the maturation step remains challenging and requires further development.


Introduction
Sickle cell disease (SCD) is one of the most common hereditary hematological diseases worldwide, particularly affecting African descendants [1,2]. However, it remains a neglected disease with high morbidity rates, and until recently, only one disease-modifying drug, hydroxyurea, has been approved for treatment [3]. Despite being a monogenic disease caused by a missense mutation in the HBB gene, significant variability is observed in the degree of clinical manifestations among SCD patients due to the influence of interactions with other polymorphisms present in the individual genetic background [4]. Therefore, the field would benefit from the establishment of novel methods and models capable of providing a better understanding of pathophysiological mechanisms and drug discovery.
Considering that most preclinical studies on SCD drug discovery have used immortalized CD34 + cells or animal models, the standardization of efficient hematopoietic differentiation protocols from patient-derived iPSCs would be highly beneficial [21].
In principle, iPSC hematopoietic differentiation protocols attempt to recapitulate the steps of embryonic development. Many protocols have been proposed to produce hematopoietic stem/progenitor cells (HSPCs), erythroid progenitor cells, and mature erythroid cells [22], such as by employing the use of co-culture with hematopoietic stimulating cells [23], feeder-and serum-free cultivation [16,24], or genetic modifications for efficient maturation [25,26]. Such works provide valuable directions to achieve a robust and reliable erythroid differentiation protocol, but results are often discrepant among them. The perspective of altered signaling pathways in patient cells implies that these protocols should be validated using both control and patient iPSCs. Indeed, the role of altered inflammatory pathways is now acknowledged as critical to the pathophysiology of SCD [26][27][28][29]. In the present study, we evaluated the efficiency of different protocols using 2D or 3D culture methods to establish a method of differentiating iPSCs towards the erythroid lineage using control and SCD-iPSC lines.

Generation and Characterization of iPSCs
This study was approved by the local Ethics Committee at São Rafael Hospital (#30875820.8.0000.0048). The iPSC lines used in this project were previously generated by reprogramming erythroblasts expanded from peripheral blood mononuclear cells (PBMCs) from three SCD patients and three healthy donors. The donors were included in the study after confirming the SCD diagnosis by hemoglobin electrophoresis and obtaining written informed consent. PBMCs were cultured in erythroblast expansion medium and transfected with episomal vectors to generate iPSCs, as previously described [9,10]. Extensive characterization of the cell lines was previously performed, including pluripotency markers, karyotype, STR analysis, mutation detection by Sanger sequencing, and β S globin gene haplotype determination by polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) [9,10,30].
(B.) Commercial medium: STEMdiff™ APEL™2 Medium (APEL; STEMCELL Technologies) was used for EB formation based on the protocol of Ng and collaborators [34]. The medium was supplemented with BMP4 (10 ng/mL), bFGF (10 ng/mL), SCF (50 ng/mL), VEGF (10 ng/mL), penicillin/streptomycin (1%), and Y-27632 (10 µM). After dissociation and counting, the iPSCs were resuspended in this medium and plated for aggregation on ULA or Aggw plates. In both cases, each EB was formed from 3000 starting iPSCs. On D6, 50 µL fresh medium was added to each ULA well. For Aggw, 750 µL of the medium was removed and the same volume of fresh medium was added. On D9, the EBs were transferred to non-adherent straight-bottomed 24-well plates at a proportion of 100 µL per transferred EB, and fresh medium was added. On D12 and D16, the cells in the supernatant were collected for analysis.

Differentiation Step 3-Terminal Maturation of Erythroid Cells
The cells in the culture supernatant were collected at the end of step 2 and incubated with the specific medium for final erythroid maturation, based on the protocol by Huang et al. [6], composed of Iscove's Modified Dulbecco's Medium (IMDM) (Thermo Fisher Scientific) supplemented with EPO (3 U/mL), holo-transferrin (500 µg/mL), heparin (4 U/mL), and insulin (10 µg/mL) (Sigma-Aldrich, St. Louis, MI, USA). The medium was changed every 2-3 days, and the cells were counted for volume adjustment and passaging, targeting a maximum cell density of 4 × 10 6 cells/mL. At the end of this stage (D33), cells were collected for analysis.

Colony Forming Unit (CFU) Assay
MethoCult (STEMCELL Technologies) was used to assess the functional capacity of HSPCs following the manufacturer's instructions. Briefly, 5000 cells were collected based on the percentage of CD34 + cells in the samples from D16 of Step 1 and then added to 3 mL of MethoCult. After homogenization, 1.1 mL was placed in a 35 mm culture plate in duplicate and incubated for 14 days. CFUs were registered using an inverted microscope (Nikon Eclipse Ti-U, Nikon Instruments Inc., Melville, NY, USA) for the identification of colony units and quantification.

Morphological Evaluation
Cells at different stages of differentiation were collected, washed with 1X PBS, and centrifuged (23× g for 5 min, acceleration medium, 21 • C) on a glass slide using a Shandon CytoSpin 4 Cytocentrifuge (Thermo Fisher Scientific). The slides were stained with Giemsa and analyzed under a microscope (Nikon Eclipse Ti-U) at 400× or 1000×.

Analysis of Hemoglobin Expression by RT-qPCR
Total RNA from differentiated cells was extracted using the TRIzol chloroform method, followed by precipitation with isopropanol. The samples were quantified using a Nan-oDrop 8000 spectrophotometer (Thermo Fisher Scientific). cDNA was produced from the reverse transcription of 0.5 µg RNA using a high-capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific), according to the manufacturer's recommendations. RT-qPCR reactions were performed in triplicate using the endogenous normalizing gene GAPDH (Hs99999905_m1), in addition to specific probes for beta globin (Hs00747223_g1 HBB) and gamma globin (Hs00361131_g1 HBG2). For the reactions, TaqMan Universal PCR MasterMix (Thermo Fisher Scientific) was used under the following cycling conditions: 50 • C 20 s, 95 • C 10 min, 95 • C 15 s, 60 • C 1 min (40×). The equipment used was ABI 7500 FAST (Thermo Fisher Scientific).

Relevant Reagents and Statistical Analysis
The relevant commercial basal media, growth factors, cytokines, isotypes, and specific primary antibodies are listed in Table S1.
For the statistical analysis, parametric data were analyzed using unpaired t-tests, and standard errors of the mean were plotted on a graph. The results were considered significant at p < 0.05. mRNA expression was calculated by the threshold method [38] using log 2 (fold change) values obtained through the calculation of 2 −∆∆Ct . Data were obtained in triplicate and evaluated by ANOVA using the Bonferroni post-test with the aid of Prism v7 software (GraphPad, Boston, MA, USA). The results were considered significant at p < 0.05 (unadjusted p-value).

Induction of iPSCs towards HSPCs (Step 1)
The iPSC lines (n = 4) included in the experiments and their respective genotypes are listed in Table 2. Initially, we tested different conditions using a 2D cultivation method ( Figure 1A) with the intention of developing a robust, cost-effective, and faster protocol. At the beginning of the protocol, the iPSCs on D0 showed compact colony formation ( Figure 1B). After incubation with either SFM or StemSpan supplemented with growth factors, we observed expansion and morphological alterations of the adhered cells in culture ( Figure 1C). Flow cytometry analysis of the cell suspension revealed a very low percentage of CD34 + /CD43 + cells ( Figure S1). listed in Table 2. Initially, we tested different conditions using a 2D cultivation method ( Figure 1A) with the intention of developing a robust, cost-effective, and faster protocol. At the beginning of the protocol, the iPSCs on D0 showed compact colony formation (Figure 1B). After incubation with either SFM or StemSpan supplemented with growth factors, we observed expansion and morphological alterations of the adhered cells in culture (Figure 1C). Flow cytometry analysis of the cell suspension revealed a very low percentage of CD34+/CD43+ cells ( Figure S1).  For subsequent assays, we compared different erythroid differentiation protocols using 3D culture techniques based on previously reported methods [6] (Figure 2A). We compared three cell aggregation methods (ULA, Hdrop, and Aggw) using HSPCim with factors and observed similar EB sizes and morphologies on D12 for all aggregation methods ( Figure 2B). This finding showed that these methods have the capacity to aggregate iPSCs and form 3D structures. Although all the methods were efficient, the Hdrop process was not performed in the following tests because it is a labor-intensive method. For the APEL media test assay, we used the ULA and AggW aggregation methods and observed that self- assembly occurred with lower numbers of non-aggregated dispersed cells surrounding the EB on D1 when compared to HSPCim ( Figure 2C). At D12, the ULA EBs from both cell lines were larger than those obtained using HSPC induction media. In contrast, the AggW EBs from both cell lines underwent spontaneous disaggregation ( Figure 2C). We also observed the presence of cells sprouting from the EB and expansion in the suspension ( Figure 2D). Based on these observations, new differentiation experiments were performed using ULA cell aggregates and APEL medium supplemented with the described growth factors.  To evaluate the presence of HSPCs in the samples, immunophenotyping was performed using flow cytometry. All cell lines successfully differentiated into CD34 + /CD43 + HSPCs. Our results showed a 34.7% ± 15.3% HSPC subpopulation on D12, and a significant percentage increase (p < 0.0031) was observed on D16, with the HSPC subpopulation representing 55.3% ± 16.4% ( Figure 3B). Although HSPC subpopulations increased in percentage, CD34+/CD43+ cell quantification showed a non-statistically significant increasing trend from D12 (4.6 ± 1.1 × 10 4 cells) to D16 (13.6 ± 4.6 ×10 4 cells) ( Figure 3C). Considering the initial number of cells on D0 and D16, we obtained an approximately 45-fold expansion of HSPCs.  The hematopoietic functional test of the HSPCs produced on D16 showed colonies that were identified as erythroid progenitor cells, including CFU-E, BFU-E, large BFU-E, granulocyte and/or macrophage progenitor cells, and CFU-GM (Figure 4). The percentage range of observed colonies was 46 ± 11% for BFU-E (regular and large), 21 ± 6.1% for CFU-E, 32 ± 6.7% for CFU-GM, and 0.95% ± 1.9% for CFU-GEMM.

Induction of HSPC towards EPCs (Step 2)
In Step 2, we compared the production of EPCs promoted by the SP34+4F and SP43+6F media ( Figure 5A). Samples from D10 of Step 2 were analyzed using CD71 and CD235a antibodies as EPC markers. The collected samples showed vivid red pellets after centrifugation ( Figure 5B). Immunophenotype analysis showed the presence of CD71+/CD235a+ cells in the samples ( Figure 5C), although significant differences were not observed in this subpopulation cultivated with SP34+4F media (65.7% ± 16%) and SP34+6F media (62.0% ± 15.6%) ( Figure 5D). Meanwhile, the quantification of CD71+/CD235a+ cells showed a significant increase (p < 0.0001) when using SP34+6F (8.81 ± 8.3 × 10 5 cells) compared to SP34+4F (2.02 ± 5.2 × 10 6 cells) ( Figure 5E). On D16 in Step 1 and D10 in Step 2, we obtained a 14-fold expansion of erythroid precursor cells. We also analyzed globin gene expression in total cells and observed a high expression of HBG2 compared to that in peripheral blood (PB) (Figure 5F), showing that the produced cells present a fetal phenotype. In contrast, the samples showed low HBB gene expression (Figure 5G).

Induction of HSPC towards EPCs (Step 2)
In Step 2, we compared the production of EPCs promoted by the SP34+4F and SP43+6F media ( Figure 5A). Samples from D10 of Step 2 were analyzed using CD71 and CD235a antibodies as EPC markers. The collected samples showed vivid red pellets after centrifugation ( Figure 5B). Immunophenotype analysis showed the presence of CD71 + /CD235a + cells in the samples ( Figure 5C), although significant differences were not observed in this subpopulation cultivated with SP34+4F media (65.7% ± 16%) and SP34+6F media (62.0% ± 15.6%) ( Figure 5D). Meanwhile, the quantification of CD71 + /CD235a + cells showed a significant increase (p < 0.0001) when using SP34+6F (8.81 ± 8.3 × 10 5 cells) compared to SP34+4F (2.02 ± 5.2 × 10 6 cells) ( Figure 5E). On D16 in Step 1 and D10 in Step 2, we obtained a 14-fold expansion of erythroid precursor cells. We also analyzed globin gene expression in total cells and observed a high expression of HBG2 compared to that in peripheral blood (PB) (Figure 5F), showing that the produced cells present a fetal phenotype. In contrast, the samples showed low HBB gene expression ( Figure 5G). Cells 2023, 12, x FOR PEER REVIEW 11 of 19

Maturation of Erythroid Cells (Step 3)
Step 3 was performed by culturing the cells in suspension in maturation medium for another 7 days to generate mature erythroid cells ( Figure 6A). Flow cytometry anal-ysis ( Figure 6B) showed a CD235a subpopulation (33% ± 17%), in which DRAQ5-cells accounted for 96% ± 4.9% ( Figure 6C), suggesting that CD235a + erythroid cells underwent enucleation, indicating erythroid maturation. Morphological evaluation of the stained cells allowed us to identify orthochromatic erythroblasts at the beginning of the maturation step and reticulocyte-like cells at the end of the assay ( Figure 6D).

Discussion
Several studies have demonstrated that iPSCs can be used to obtain erythroid cells based on either EBs (3D culture) [35,36,39] or monolayers (2D culture) [24,40,41]. In the present study, we present a comparative analysis of different protocols to generate hematopoietic progenitor cells and erythroid cells from iPSCs obtained from healthy donors and SCD-iPSCs. Due to the small sample size, our data does not allow for comparison between healthy cells and SCD-iPSCs, but it contributes to the evaluation of the reproducibility of published reports, and it helps define the most appropriate methodology for obtaining erythroid cells from iPSCs, which is necessary before these models can be successfully applied in drug screening and disease-modeling studies. Indeed, the importance of validation studies is demonstrated by differences in the efficiency of human iPSC differentiation protocols, which may be influenced by several factors, including inter-laboratory and cell line variability.
The first step in differentiation is to obtain HSPCs, which are defined by the coexpression of CD34 and CD43 [23] that characterize primitive hematopoietic progenitors [42]. Although we did not achieve satisfactory results using the 2D culture methodology, other authors have achieved in vitro erythropoiesis through monolayer culture. Nonetheless, most of the reported 2D protocols involve coculture with OP9 cells [42,43], which were avoided here because of the goal of establishing xeno-free protocols.
Cellular interactions in the microenvironment are critical for proper hematopoietic differentiation [44][45][46][47]. Therefore, by enhancing cell interactions, 3D models based on EB differentiation can provide a theoretical advantage over 2D monolayer models. Data from different studies [48] show that within the formed 3D structure, the molecular and morphological events resemble the beginning of human embryonic development [49] and favor the induction of hematopoietic lineages [6,35,50]. Our data support the more efficient generation of HSPCs and increased cell expansion using 3D methods, and these results could be further scaled up. Among the different methods used for EB formation, the hanging drop method has been widely applied in research on mouse pluripotent stem cells, with few successful reports in human cell studies. As a labor-intensive method, other aggregation methods have been shown to be more efficient, corroborating the data described herein [51][52][53]. We also observed that Aggw plate and APEL medium presented EB deterioration along the hematopoietic differentiation. In our experiments, we utilized the same media and cell numbers for EB formation for the Aggw and ULA methods, and although both methods were able to initially form EBs with similar size and characteristics, an important difference was observed in the steps following aggregation. It was not in our scope to evaluate the mechanisms involved in that event, but this observation ended up favoring the ULA plates aggregation method. It is known that large EBs may present areas of core necrosis due to low oxygen/nutrient diffusion, and EBs that are too small may not survive or perform inefficiently in differentiation protocols [54]. EB viability and the yield in terminal differentiation may vary in a size-dependent manner. Our results highlight the importance of choosing and validating an aggregation method that better suits the desired applications. For instance, different results in terms of EB survival and differentiation performance have been reported for EB formed even for V and U bottom plates [55].
After differentiation Step 1, we achieved an average induction efficiency of 55% for iPSCs towards CD34 + /CD43 + HSPCs. Kessel et al. [56] obtained an induction efficiency of approximately 10% for EBs towards CD34 + /CD43 + HSPCs using a similar protocol. In addition, we generated HSPCs in high quantities when using the APEL medium, which allowed for the continuation of differentiation Steps 2 and 3. These data corroborate the study published by Reis et al. [11] concerning the use of SCD-derived iPSCs to perform hematopoietic differentiation with EB formation, including the use of APEL medium and the emergence of surrounding cells from EB, as previously mentioned. In contrast, our work showed an increase in the percentage of CD34 + /CD43 + cells using the methodology described herein, while Reis et al. observed a reduction or loss in the percentage of these hematopoietic markers. In addition to the already known epigenetic memory effect over iPSCs [57,58], the observed discrepancy may be explained by the intrinsic cell line variability, such as alterations in the expression levels of hematopoietic-related genes and/or other gene variants that can alter cell behavior to the same stimuli [59].
Our Step 1 cell quantification comparison of D12 and D16 suggests that increasing the time of induction delivers a higher number of HSPCs, thus allowing robust HSPC expansion, which is useful for detailed analysis of hematopoietic development and disease modeling. Concerning the number of erythroid progenitor cells at D10 of Step 2, we observed a 14-fold expansion of CD71 + /CD235a + from HSPCs, which resulted in a 630-fold cell expansion related to that at the beginning of the 3D protocol. Other authors have shown an expansion of approximately 10-fold in erythroblasts derived from different sources of iPSCs using a co-culture with the OP9 cell line [60]. Deng et al. obtained a 530-fold expansion at the end of the erythroblast phase using platelet lysate in a medium [61]. Although these differences can be associated with the cultivation method, they serve as references to demonstrate the enforceability of our protocol. Another study by the above-cited group [16] also evaluated the functional activity of HSPCs by CFU assay and identified the formation of colonies of macrophages, granulocytes, and erythrocytes. Our work corroborates these findings, which demonstrate that the functional abilities of healthy donors and sickle cell iPSC-derived HPSCs are similar and able to generate hematopoietic colonies.
After cultivation with specific factors for erythroid differentiation and expansion of erythroblasts, in Step 2, we obtained approximately 65% CD71 + /CD235a cells. The yield of cells obtained at the end of stage II was 34-and 42-times the number of the initial cells depending on the analyzed strain, and these values are close to those obtained by Huang and collaborators for the same stage [6]. However, other studies that used EB formation to obtain erythroid cells reported the presence of more than 80% of cells labeled for either CD71 or CD235a [36,62]. The red cell pellet suggests that the induced cells produced hemoglobin. The gene expression of HBG2 and HBB in iPSC-derived erythroid progenitor cells suggests that differentiated cells have a fetal phenotype and are not capable of switching to a mature phase. Other studies have reported the same profile [23,42,62]. Maturation and hemoglobin switching are critical steps for hematopoietic disease modeling, especially in SCD cell lines, and these steps should be optimized in future studies so that a disease phenotype (i.e., sickling) can be reproduced in vitro. In this sense, alternative protocols that show highly efficient production of hemoglobin β using different approaches, such as immortalization of erythroid precursor cells and modulation of BCL11A expression [25,26,63], in vivo maturation in NOD/SCID mice [64], or using cells treated with human platelet lysate [63], should be further assessed. Recently, Chen et al. reported that erythroid terminal differentiation was improved in a culture of HSPCs derived from human embryonic stem cells by adding an aryl hydrocarbon receptor antagonist [65], and this represents a less laborious approach that is compatible with SCD disease models.

Conclusions
In this study, we describe the application of iPSC lines derived from healthy donors or SCD individuals to obtain erythroid cells. By optimizing protocols described in the literature and different methodologies, a protocol was developed for obtaining erythroid cells from iPSCs that express specific markers and have typical morphological characteristics; however, the cells present a fetal phenotype, requiring further efforts to improve maturation. The results of this study will serve as a basis for the development of 3D protocols and may contribute to establishing a platform for screening drugs for sickle cell anemia, as well as studying the mechanisms involved in the disease and developing new therapies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells12081121/s1, Table S1: List of relevant commercial reagents; Figure S1: Flow cytometry analysis of 2D erythroid differentiation method. Institutional Review Board Statement: The study was conducted under the Declaration of Helsinki, following a protocol approved by the local ethics IRB at São Rafael Hospital) CAAE: 30875820.8.0000.0048). All patients gave written informed consent to participate in the study.

Informed Consent Statement:
The ethical approvals for using the iPSC lines, as stated in the IRB statement, include written consent from the donors.

Data Availability Statement:
All relevant data is available from the authors upon reasonable request.