Induction of hematopoietic differentiation of mouse embryonic stem cells by an AGM-derived stromal cell line is not further enhanced by overexpression of HOXB4

Abstract

Hematopoietic differentiation of embryonic stem (ES) cells can be enhanced by co-culture with stromal cells derived from hematopoietic tissues and by overexpression of the transcription factor HOXB4. In this study, we compare the hematopoietic inductive effects of stromal cell lines derived from different subregions of the embryonic aorta-gonad-mesonephros tissue with the commonly used OP9 stromal cell line and with HOXB4 activation. We show that stromal cell lines derived from the aorta and surrounding mesenchyme (AM) act at an earlier stage of the differentiation process compared with the commonly used OP9 stromal cells. AM stromal cells were able to promote the further differentiation of isolated brachyury-GFP(+) mesodermal cells into hematopoietic progenitors, whereas the OP9 stromal cells could not support the differentiation of these cells. Co-culture and analyses of individual embryoid bodies support the hypothesis that the AM stromal cell lines could enhance the de novo production of hematopoietic progenitors, lending support to the idea that AM stromal cells might act on prehematopoietic mesoderm. The induction level observed for AM stromal cells was comparable to HOXB4 activation, but no additive effect was observed when these 2 inductive strategies were combined. Addition of a γ-secretase inhibitor reduced the inductive effects of both the stromal cell line and HOXB4, providing clues to possible shared molecular mechanisms.

FIG. 1.

Induction of hematopoietic differentiation by co-culture with stromal cell lines. The total number of hematopoietic colonies (A) and the number of specific colony types including CFU-Mix (B), CFU-GM (C), CFU-M (D), and Ery/Mac (E) per 3 × 10⁵ embryonic stem (ES)-derived cells (7a-GFP ESC) in control gelatin cultures or after co-culture on stromal cell lines to 6 days differentiation. Data in A represent the mean of between 3 and 11 independent experiments, whereas data in B–D represent the mean (±SD) of 3 representative experiments, *P < 0.05 compared to gelatin controls. To ensure the entire hematopoietic output was represented in A, remaining colonies such as definitive erythroid and CFU-mast were categorized as “other.”

FIG. 2.

Stromal cell lines mediate their effects after brachyury expression in embryoid bodies (EBs). (A) Percentage of brachyury-eGFP expressing cells in Bry-201 EBs (Days 2–8) differentiated under the different co-culture conditions, P > 0.05 between conditions (1 representative experiment of 3 is shown). (B) Proportion of brachyury-eGFP expressing embryonic stem (ES) cells in co-cultures at Day 5 (P > 0.2 when comparing each stromal co-culture to gelatin controls, n = 3 experiments). (C) Experimental strategy: brachyury-eGFP positive and negative cells were isolated by fluorescence-activated cell sorter (FACS) from 4-day-old suspension EBs and then co-cultured for a further 6 days before assessing hematopoietic activity. (D) Representative example of FACS profile of unsorted and sorted populations. (E) Number of hematopoietic colonies produced from brachyury-eGFP positive and negative populations after co-culture on gelatin, AM14.1C4 or OP9 stromal layers for 6 days (*P < 0.02 compared to corresponding gelatin and OP9 co-cultures).

FIG. 3.

No additive effect when the AM stroma and HOXB4 induction strategies are combined. Cos7 cells, transiently transfected with the pCAGHOXB4ER^T2IP construct, cultured in the absence (A) or the presence (B) of tamoxifen followed by immunocytochemical analysis using an anti-HOXB4 antibody (green) and counterstaining of the nucleus with DAPI (blue). (C) Number of hematopoietic colonies observed when control CGR8 embryoid bodies (EBs) and CGR8-HOXB4-ER^T2 inducible EBs were cultured in the presence (+) or absence (−) of tamoxifen (added on Days 1, 3, and 5 of differentiation) and/or co-cultured with AM14.1C4 or OP9 stromal cell layers. CFU assays were setup at 6 days differentiation.

FIG. 4.

γ-Secretase inhibition of embryoid bodies (EBs) differentiated using the 2 induction strategies. (A) Frequency of multipotent hematopoietic colonies (CFU-Mix, CFU-GM, Ery/Mac) per 1 × 10⁵ input embryonic stem (ES)-derived cells differentiated to 6 days on gelatin, AM14.1C4 or OP9 in the presence (+) or absence (−) of γ-secretase inhibitor (GSI), added between Days 4 and 6. One representative experiment of 3 is shown. (B) Quantitative reverse transcription-polymerase chain reaction (RT-PCR) showed a significant reduction in Hey1 gene transcript in EBs in the presence of γ-secretase inhibitor (GSI). (C) γ-Secretase inhibition of HOXB4-ER^T2 EBs differentiated on gelatin. Cells were either noninduced or induced with tamoxifen in the presence (+) or absence (−) of GSI, added between Days 4 and 6. The fold induction in CFU-Mix, CFU-GM, and Ery/Mac colony numbers over noninduced control HOXB4-ER^T2 EBs is shown (±SD between experiments). (D) Quantitative RT-PCR showing an increase in Hey1 gene transcript after HOXB4 activation which was inhibited in the presence of the γ-secretase inhibitor (GSI). (E) Quantitative RT-PCR showing the expression pattern of Jagged1 in control CGR8 EBs or HOXB4-ER^T2 EBs differentiated in the presence (+) or absence (−) of tamoxifen, added from Days 2 to 4.