Directed differentiation of hematopoietic precursors and functional osteoclasts from human ES and iPS cells

Abstract

The directed differentiation of human pluripotent stem cells offers the unique opportunity to generate a broad spectrum of human cell types and tissues for transplantation, drug discovery, and studying disease mechanisms. Here, we report the stepwise generation of bone-resorbing osteoclasts from human embryonic and induced pluripotent stem cells. Generation of a primitive streak-like population in embryoid bodies, followed by specification to hematopoiesis and myelopoiesis by vascular endothelial growth factor and hematopoietic cytokines in serum-free media, yielded a precursor population enriched for cells expressing the monocyte-macrophage lineage markers CD14, CD18, CD11b, and CD115. When plated in monolayer culture in the presence of macrophage colony-stimulating factor and receptor activator of nuclear factor-kappaB ligand (RANKL), these precursors formed large, multinucleated osteoclasts that expressed tartrate-resistant acid phosphatase and were capable of resorption. No tartrate-resistant acid phosphatase-positive multinucleated cells or resorption pits were observed in the absence of RANKL. Molecular analyses confirmed the expression of the osteoclast marker genes NFATc1, cathepsin K, and calcitonin receptor in a RANKL-dependent manner, and confocal microscopy demonstrated the coexpression of the alphavbeta3 integrin, cathepsin K and F-actin rings characteristic of active osteoclasts. Generating hematopoietic and osteoclast populations from human embryonic and induced pluripotent stem cells will be invaluable for understanding embryonic bone development and postnatal bone disease.

Figure 1

Specification of hematopoietic and osteoclast lineages from human hESCs and hiPSCs. A schematic of the stepwise protocol used for the differentiation of H1, HES2, and MSC-iPS cells to the osteoclast lineage (see “Commitment of hESCs and hiPSCs to the monocyte-macrophage lineage” for details).

Figure 2

Colony-forming potential of hESCs and hiPSCs. EBs from hESCs (H1, HES2) and hiPSCs (MSC-iPS1) were harvested at the indicated times during hematopoietic cell commitment and myeloid cell expansion and colony forming potential was measured in methylcellulose cultures. (A) Quantification of the number of erythroid (Ery), macrophage (Mac), and myeloid colonies. Erythroid includes both pure erythroid colonies (> 90%) and mixed erythroid/myeloid colonies (< 10%). The majority of the myeloid colonies are mast cell colonies. Data represent the mean ± SEM of 6 to 8 cultures from 3 independent experiments. Values are presented as a percentage of total colonies, to demonstrate the emergence of macrophage progenitors with time during EB differentiation. We typically observe between 320 and 850 total colonies per 5 × 10⁴ cells at 2 weeks for all 3 cell lines, which increases to 490 to 1475 colonies per 5 × 10⁴ cells by 3 weeks. (B) Representative photographs of typical H1-derived colonies and cells within colonies after cytospin preparations. Images were acquired with a Zeiss Axioskop2 plus microscope using plan-Neofluar objectives and an Axiocam camera.

Figure 3

Monocyte-macrophage surface marker expression on developing EBs. Flow cytometric analysis of EBs from hESCs (H1, HES2) and hiPSCs (MSC-iPS1) harvested at the indicated times during hematopoietic differentiation, showing expression patterns of CD18, CD11b, CD115, and CD14, expressed as a percent of CD45⁺ hematopoietic progenitors. Open histograms represent populations stained with the indicated antibodies and shaded histograms represent unstained control samples.

Figure 4

Differentiation of functional osteoclasts from hESC and hiPSCs. Representative micrographs of multinucleated osteoclasts derived from H1-derived (A-B) and MSC-iPS1–derived (C-D) hematopoietic precursors. Cells from day 20 cultures were seeded onto dentine slices in the presence of M-CSF and RANKL, and osteoclasts were stained 2 weeks later for TRAP activity (↑). Clear multinucleation is evident (B,D) and prominent resorption trails (*) are associated with all osteoclasts as viewed under reflected light. Resorption lacunae created by H1-derived (E,F,I) and MSC-iPS1–derived (G-H) osteoclasts were identified after removal of cells and staining with toluidine blue or by scanning electron microscopy (I). Cells cultured in the absence of RANKL did not differentiate into TRAP-positive cells, and resorption pits were never observed (J). Images were acquired with a Leica MZ FLIII stereomicroscope using a DFC300 FX camera. Auto adjustments were performed with Adobe Photoshop software. Original magnifications: ×4 (E); ×10 (A,C,G,J); ×20 (F,H); and ×40 (B,D).

Figure 5

RANKL-dependent differentiation of osteoclasts. Quantification of the number of multinucleated osteoclasts and their resorptive activity, generated by H1- and MSC-iPS1–derived osteoclast precursors. Cells from day 21 cultures were seeded at 10⁵ cells/96-mm well on dentine slices in the absence or presence of RANKL as indicated and harvested 14 days later for TRAP staining (A) and resorption analysis (B). The data represent the mean ± SD of 4 to 6 wells. *P < .05.

Figure 6

Expression of mature osteoclast markers. Mature osteoclasts derived from hESCs express F-actin rings (A-C) and VNR (E-G). H1-derived hematopoietic precursors from day 21 cultures were seeded at 10⁵ cells/96-mm well onto dentine slices in the presence of M-CSF and RANKL and stained 14 days later with TRITC-phalloidin (A-C) or by immunohistochemistry for VNR with the 23C6 antibody (E-G). F-actin rings (white arrows, panels A-C), ruffled borders (white arrowheads, panels B-C), and strong VNR expression (black arrows, panels E-G) are present in actively resorbing, multinucleated osteoclasts. Resorption trails are clearly visible in panels E-G (*). Images were acquired with a Zeiss Axioskop2 Plus microscope equipped with epifluorescence using plan-Neofluar objectives and an Axiocam camera. Auto adjustments were performed with Adobe Photoshop software. Laser confocal microscopy demonstrates the coexpression of β3 integrin (green), F-actin rings (red), and cathepsin K (blue) in H1-dervied (H) and MSC-iPS1–derived (I) osteoclasts. F-actin ring–positive cells were never detected in cultures lacking RANKL (D). Original magnifications: ×10 (A,D); ×20 (B,C,E,F); and ×40 (G).

Figure 7

Quantitative PCR analysis of osteoclast marker genes. RANKL-dependent expression of osteoclast marker genes in cultures of H1-derived (A) and MSC-iPS1–derived (B) precursors. Cells from day 21 EB cultures were seeded at 10⁵ cells/96-mm well on dentine slices in the absence or presence of RANKL as indicated and harvested 14 days later for RNA isolation and analysis of the osteoclast marker genes cathepsin K (CATK), calcitonin receptor (CTR), and NFATc1 by quantitative reverse-transcription PCR. The data represent the mean ± SD of 3 to 6 wells. *P < .05.