Applications of the Human Bone Marrow Organoid System to Study Hematopoiesis and Anemia Stress

Ribonucleoprotein nucleofection

Gene-edited CD34⁺ HSPCs were purchased from the HSPC Gene Editing core at the Fred Hutchinson Cancer Research Center. Cells were thawed 2–4 h before nucleofection. Gene editing was conducted according to previously defined protocols for optimized nucleofection of single guide RNA (sgRNA):Cas9 ribonucleoprotein (RNP) complexes. ³⁸ Briefly, two lyophilized single guide RNA oligos per gene (Synthego) were dissolved in Tris-EDTA (pH 8.0) and mixed at a concentration of 100 pmol/µL (50 pmol/µL per guide). RNP complexes were assembled by mixing 1 µL of the sgRNAs with 2 µL of sNLS-SpCas9-sNLS nuclease (25 pmol/µL; Aldevron) and 17 µL of Complete P3 solution (P3 4D-Nucleofector Kit) to a total volume of 20 µL. 2–4 ×10⁵ CD34⁺ cells (resuspended in P3 + RNP complexes) were nucleofected using the DS-150 program on the 4D Nucleofector (Lonza). Freshly nucleofected cells were transferred to CD34 expansion media (StemSpan SFEM II (Stemcell Tech., 09655) with 100 ng/mL TPO, SCF, interleukin-6 [IL-6], fms related receptor tyrosine kinase 3 [FLT-3]) for 4 h and then cryopreserved in CryoStor CS10 (Sigma).

CRISPR editing analysis

Quantitation of Adeno-Associated Virus Integration Site 1 (AAVS1)-CTRL editing efficiencies was determined using the Interference of CRISPR Editing (ICE) PCR assay, while SAMD1 knockout was partially confirmed by SYBRgreen qPCR. Briefly, genomic DNA (MicroElute Genomic DNA Kit, Omega Bio-Tek) was isolated from cells 48 h postnucleofection. PCR amplification of the region surrounding the AAVS1 cut sites was performed using 50 ng DNA, Phusion High-Fidelity DNA Polymerase with GC Buffer, AAVS-ICE-F/R primers, and 1M Betaine. Size-verified PCR products were column purified (Monarch PCR and DNA Clean-up, New England Biolabs) and submitted for Sanger sequencing with the AAVS-ICE-F primer (Genewiz). Trace files were used to quantify editing using the EditCo tool (https://ice.editco.bio). Due to issues with generating high-quality Sanger sequencing data at the SAMD1 edit site, qPCR assays were used to partially quantify editing efficiencies based on a standard curve generated for absolute quantification. One assay quantified the deletion frequency of the region between the two sgRNAs (primers SAMD1-F1/R1), while a second partially quantified sgSAMD1-166 editing as well as deletion between the sgRNAs (Supplementary Fig. S1A-C; primers SAMD1-F/R).

Generation of human BMOs

Human bone marrow organoids were differentiated from human iPSCs (CVCL_RM92) as previously described.^33,34 Briefly, the generation of organoids is performed in five phases. Phase 0 iPSCs are detached and aggregated in StemFlex and Revitacell for 1 day. Phase I is 3 days of mesodermal commitment in APEL2 medium containing 50 ng/mL of BMP4, Fibroblast Growth Factor-2 (FGF2), and Vascular Endothelial Growth Factor A (VEGFA) and 3 μM of the CHIR99021 GSK3 inhibitor performed in 2% O₂, 5% CO₂, 93% N₂. Vascular and hematopoietic commitment is induced for 2 days in Phase II in APEL2 medium containing 50 ng/mL of BMP4, FGF2, and VEGFA and 25 ng/mL of hSCF and FLT-3. At phase III, iPSC aggregates were embedded in mixed collagen hydrogels and cultured in APEL2 medium containing 5 U/mL of heparin, 5% KO serum (ThermoFisher Scientific), and 100 ng/mL of VEGFA, 50 ng/mL of Vascular Endothelial Growth Factor C (VEGFC), FGF2, TPO, erythropoieting (EPO), hSCF, BMP4, and Granulocyte-Colony Stimulating Factor (G-CSF), and 20 ng/mL of IL-3 and IL-6 for 5 days. After 2 days, 1 mL of medium was added to each well containing half the concentration of the cytokines and growth factors. Day 10 hBMOs were distributed into individual wells of U-bottomed 96-well plates and cultured for an additional 2 days in Phase IV medium before engraftment or stress experiments.

Engraftment of gene-edited CD34⁺ cells

Cryopreserved, nucleofected CD34⁺ HSPCs were thawed one day prior to engraftment and cultured at 1 × 10⁵ cells/mL in StemSpan serum-free expansion media containing recombinant human FLT-3, SCF, IL-3, and IL-6 (Stemcell Technologies). Alternatively, CD34⁺ cells can be freshly nucleofected one day prior to engraftment. To label engrafted cells, nucleofected CD34⁺ HSPCs were washed and incubated with 5 μm of carboxyfluorescein succinimidyl ester (CFSE; Biolegend). Cells were resuspended in Phase IV medium composed of StemPro-34 (Gibco) containing 5 U/mL of heparin, 5% KO serum (Thermo Fisher Scientific), 25 ng/mL of VEGFA, VEGFC, FGF2, hSCF, and FLT-3, and 10 ng/mL of TPO, EPO, IL3, and IL6. 5000 CD34⁺ cells in 100 μL of Phase IV medium were added to organoid cultures for engraftment. Engrafted organoids were cultured for 7 days. Every 48–72 h, 100 μL of Phase IV media was removed and replaced with fresh Phase IV media.

Induction of “stress” erythropoiesis

Day 12 hBMOs were grown in 100 μL of Phase IV media containing 15 ng/mL of hBMP4 (R&D Systems). Organoids were incubated in either normoxic (20% O₂) or hypoxic (2% O₂) conditions for 6 days with media changes every 48–72 h.

Dissociation of organoids and flow cytometry

To dissociate organoids into a single-cell suspension, 5–10 organoids per replicate were collected in a 15 mL conical tube using a P1000 trimmed tip. To remove nonengrafted cells, 5 mL of PBS was added and incubated at room temperature for 4 min, then aspirated. Organoids were washed in 10 mL of PBS at 400 g for 5 min. PBS was aspirated, and organoids were resuspended in 500 μL of collagenase (10 mg/mL) in Hank’s Balanced Salt Solution (Corning). Organoids were incubated at 37°C for 6 min and then mixed with a P1000 followed by another 6-min incubation at 37°C. After mixing with a P200, the remaining visible organoids were incubated in 3-min increments at 37°C, followed by mixing with a P1000 until completely dissociated by eye. Immediately after dissociation, 3 mL of 10% FBS in PBS was added to stop enzyme activity. Cells were washed at 200 g for 5 min and resuspended in PBS.

Engrafted and organoid cells were stained with preconjugated antibodies to detect the surface proteins CD34 (clone 561), CD38 (clone HIT2), CD45RA (clone HI100), CD49f (clone GoH3), CD90 (clone 5E10), CD71 (clone CY1G4), CD235a (clone HI264), CD45 (clone 2D1), CD11b (clone M1/70), fixable Live/Dead Blue stain (Thermo Fisher Scientific), and a biotin-conjugated lineage negative cocktail: CD3 (clone OKT3), CD19 (clone HIB19), CD14 (clone M5E2), CD56 (clone HCD56), and CD20 (clone 2H7). Data were collected on the LSR Fortessa and analyzed in FlowJo v10.10.0.

Statistical analysis

Statistics were performed as stated in each figure and were performed in GraphPad Prism v10.4.2. Quantitative data are expressed as +/− standard error of the mean. Significance is defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

Efficient engraftment of gene-edited CD34⁺ cells into human bone marrow organoids

We generated organoids using a 10-day protocol outlined in Figure 1A, according to Khan et al. ³³ Between days 10 and 12, we developed new protocols to evaluate engraftment of gene-edited cells and the role of low oxygen/BMP-4 treatment on organoid erythropoiesis (Fig. 1B). On day 12, organoids developed characteristic features, including vessel formation and cells with erythroid morphology within vessels (Fig. 1C). To establish hBMOs as a platform for functional genetic interrogation of human hematopoiesis, we first evaluated whether CRISPR/Cas9-edited CD34⁺ HSPCs could be efficiently engrafted and maintained within the organoid microenvironment. As proof of principle, CD34⁺ cells from healthy human donors were edited by RNP nucleofection using two single-guide RNAs targeting the Sterile Alpha Motif Domain-1 (SAMD1) gene, a transcription factor that plays a role in hematopoiesis, erythropoiesis, and megakaryopoiesis, ³⁹ or using guides targeting the AAVS1 locus as a control. AAVS1-CTRL editing efficiencies were quantified by ICE analysis of Sanger sequencing (100% indel frequency; Supplementary Fig. S1). Due to technical issues with Sanger sequencing data covering the SAMD1 edited region, we developed a new method to test editing efficiencies using absolute quantitation of SYBRgreen qPCR data (Supplementary Fig. S1). The deletion frequency for the two sgRNAs targeting SAMD1 were 44.9% and 8.8%, respectively. To confirm loss of SAMD1 protein, edited cells were expanded in StemSpan liquid culture containing cytokines (see methods) for 6 days (Fig. 2A).

OPEN IN VIEWER

FIG. 1.

Generation of vascularized organoids for gene loss-of-function experiments and erythropoietic stress. (A) Schematic for generating organoids from iPSCs. (B) Schematic for conducting loss-of-function and erythropoietic stress experiments. (C) Representative Hematoxylin and eosin staining of hBMO cross-sections (20× and 40× magnifications). hBMOs, human bone marrow organoids; iPSCs, induced pluripotent stem cells.

OPEN IN VIEWER

FIG. 2.

Gene-edited CD34⁺ HSPCs engraft into hBMOs. (A) Western blotting for Sterile Alpha Motif Domain 1 (SAMD1) and β-actin of protein lysates extracted from CD34⁺ progenitors nucleofected with RNP complexes containing sgRNAs targeting control genomic regions (sgAAVS1) or targeting the SAMD1 gene (sgSAMD1). (B) Images of CFSE-positive cells taken in the green fluorescence channel 7 days after organoid engraftment of CD34⁺ HSPCs. (C) Representative flow cytometry plot of CFSE⁺ cells 7 days after engraftment. (D) Quantitation of percent engraftment (CFSE⁺) in total live hBMOs after 7 days (n = 9). CFSE, carboxyfluorescein succinimidyl ester; HSPCs, hematopoietic stem and progenitor cells; RNP, ribonucleoprotein.

Control and gene-edited (sgSAMD1) CD34⁺ HSPCs were labeled with CFSE and engrafted into pre-established hBMOs 24 hours after thawing. After 7 days, fluorescence imaging confirmed the presence of CFSE⁺ cells in the organoid (Fig. 2B). The CFSE⁺ cells appeared to be spread evenly throughout the organoids and not clustered at one engraftment location. To examine the percentage of engrafted cells in the organoid, hBMOs were dissociated for flow cytometry. For this, 6.5% of the organoid was engrafted cells, and no difference was observed between the cells nucleofected with nontargeting and gene-targeting guides (Fig. 2C and D). These results demonstrate that hBMOs support the engraftment of gene-edited human HSPCs.

Gene-edited CD34⁺ cells are maintained and undergo multilineage differentiation in hBMOs

To determine whether gene-edited CD34 cells’ retain immunophenotypic HSPC properties in hBMOs, organoids were dissociated after 7 days using collagenase II and manual pipetting for immunophenotypic analysis. Flow cytometric profiling revealed that 10.6% of the engrafted CFSE⁺ cells were CD34⁺Lin⁻, indicating that engrafted, edited immature progenitor populations remained in organoids. The CFSE⁺CD34⁺ Lin⁻ population consisted predominantly of common myeloid progenitors (CMPs; CD38⁺ CD45RA⁻), with smaller fractions of megakaryocyte/erythroid progenitors (MEPs; CD38^low CD45RA⁻), granulocyte/monocyte progenitors (GMPs; CD38⁺ CD45RA⁺), and a detectable populations of HSCs (CD38⁻ CD45RA⁻, CD90⁺, and CD49f⁺), and multipotent progenitors (MPPs; CD38⁻ CD45RA⁻ and CD90⁻ CD49f⁻; Fig. 3). ⁴⁰ Common lymphoid progenitors (CLPs) were not observed in the engrafted CD34 population. The lack of CLPs is expected as hBMOs themselves do not support lymphoid differentiation. ³³

OPEN IN VIEWER

FIG. 3.

hBMOs support engrafted CD34⁺ immunophenotypically defined progenitor populations. Representative gating strategy of HSPC cell types in control CD34⁺ cells nucleofected with sgRNA targeting the AAVS1 locus at day 6: hematopoietic stem/progenitor cells (CD34⁺ Lin⁻), CMPs (CD38⁺ CD45RA⁻), MEPs (CD38^low CD45RA⁻), and GMPs (CD38⁺ CD45RA⁺). CD34⁺ Lin⁻ shown as percent of CFSE⁺. CMP, GMP, MEP, CD38⁻ CD45RA⁻, HSC, and MPP are shown as a percentage of CD34⁺ Lin⁻. CMP, common myeloid progenitor; HSC, hematopoietic stem cell; GMP, granulocyte/monocyte progenitors; MEP, megakaryocyte/erythroid progenitors; MMP, multipotent progenitor.

Next, we evaluated whether edited CD34 cells are capable of multilineage differentiation. Analysis of CFSE⁺ CD45⁺ cells revealed myeloid differentiation, based on the acquisition of a CD34⁻CD11b⁺ population (Fig. 4A). CD71, also known as transferrin receptor, and CD235a can be used to evaluate stages of erythroid maturation. We observed that engrafted CFSE⁺ cells differentiated to early erythroid progenitors (CD71⁺ CD235a⁻) and late erythroid progenitors/mature erythrocytes (CD235a⁺), making up approximately 20.5% and 27% of CFSE⁺ cells, respectively. A representative flow panel from three replicates of pooled organoids is shown in Figure 4B. These data demonstrate that hBMOs maintain the HSPC populations of engrafted CD34⁺ cells after 7 days in culture and also support multilineage differentiation.

OPEN IN VIEWER

FIG. 4.

hBMO-engrafted CD34⁺ HSPCs differentiate to the myelomonocytic and erythroid lineages. (A) Representative gating strategy of myelomonocytic cells 7 days after CD34⁺ cell engraftment in the CFSE⁺ fraction. CD45⁺ cells are shown as a percent of CFSE⁺. CD34⁻ cells and myelomonocytic cells (CD11b⁺) are shown as a percent of CD45⁺ cells. (B) Representative flow cytometry plot of CD71⁺ and CD235a⁺ cells 7 days after CD34⁺ cell engraftment in the CFSE⁺ fraction. Shown as percent of CFSE⁺.

Low O₂ and the addition of BMP4 alters hematopoiesis in hBMOs

Stress erythropoiesis occurs during acute anemia in mice and humans to replenish lost erythrocytes, and foundational work showed that BMP4 and O₂ levels can mimic this process in liquid cultures ex vivo.^20,21 In this light, we tested whether hematopoietic and erythroid progenitor cells in hBMOs can respond to similar environmental signals to increase erythropoietic output. We shifted day 12 organoids to a culture media containing an additional 15 ng/mL BMP4 (or PBS control) and/or low oxygen for 6 days. Pooled organoids were dissociated for comparative flow cytometry analysis. The percentage of Lin⁻CD34⁺ cells in hBMOs increased from 4.8% in normoxia to 12.5% in hypoxia (N = 3; Fig. 5A). Hypoxia increased CD34⁺ cells compared to normoxia in both control media and BMP-4-containing media (Fig. 5B). Next, we analyzed MPP, HSC, CMP, and MEP populations (Fig. 5C). No significant changes were observed in the percentages of progenitor populations within the CD34⁺ fractions (Fig. 5D). We tested whether the erythroid populations changed in response to hypoxia and/or BMP-4 (Fig. 6A). In liquid cultures, hypoxia and BMP-4 act combinatorially to induce stress erythropoiesis, a process involved in recovery from acute anemia. Growing hBMOs in low O₂ environments combined with the addition of BMP-4 to media increased the percentage of erythroid cells within organoids by 2.9-fold (Fig. 6B). The changes in erythroid output caused by modifications to these two variables is consistent with a “stress erythropoiesis” phenotype induced by BMP-4 and hypoxia in liquid culture systems.^19,20 However, functional testing is needed to confirm this. Evaluating stress erythropoiesis in organoids is a more physiologically relevant human model system which retains the original stem cell populations and is therefore capable of capturing both short- and long-term mechanistic actions of erythroid progenitors during specification, expansion, differentiation, and resolution stages. Combined with gene editing approaches, we can use these modified culture systems to investigate loss-of-function phenotypes, the influence of additional growth factors, the incorporation of other cell types, or how drug treatments may influence human stress erythropoiesis.

OPEN IN VIEWER

FIG. 5.

Low O₂ and BMP4 alter the HSPC compartment in hBMOs. (A) Representative flow cytometry plots after 6 days of low O₂ and/or BMP4 addition. (B) Quantitation of CD34⁺ cells determined by flow cytometry (n = 3). Numbers in boxes represent a percentage of live cells. (C) Representative flow cytometry plots of lineage-negative, CFSE⁺CD34⁺ cells: HSCs (CD90⁺ITGA6⁺), CMPs (CD38⁺CD45RA⁻), MEPs (CD38^lowCD45RA⁻), GMPs (CD38⁺CD45RA+). (D) Quantitation of HSCs, CMPs, MEPs, and GMPs after 6 days of low O₂ and/or BMP4 addition as determined by flow cytometry (n = 3). **p < 0.01 (two-tailed unpaired Student’s t-test).

OPEN IN VIEWER

FIG. 6.

Low O₂ and BMP4 increase erythroid output in hBMOs. (A) Representative flow cytometry plots of anti-CD71 PE and anti-CD235a after 6 days of low O₂ and/or BMP4 addition. Numbers in boxes represent a percent of live cells. (B) Quantitation of CD71⁺ CD235a⁺ (erythroid) cell percentages in organoids after 6 days of low O₂ and/or BMP4 addition as determined by flow cytometry (n = 6). **p < 0.01 (two-tailed unpaired Student’s t-test).