Genetic Engineering of Human Embryonic Stem Cells for Precise Cell Fate Tracing during Human Lineage Development

Abstract

It is highly desirable to specify human developmental principles in an appropriate human model with advanced genetic tools. However, genetically engineering human cells with lineage-tracing systems has not been achieved. Here we introduce strategies to construct lineage-tracing systems in human embryonic stem cells (hESCs). The AAVS1 locus was suitable for the integration of the conditional reporter. The Cre-LoxP and Flp-FRT systems were highly sensitive, which may cause inaccurate lineage labeling in human cells. The recombination sensitivity and tracing fidelity could be finely tuned by modification of the LoxP recombination site. Moreover, tamoxifen-controllable Cre^ERT2-LoxP and Flp^ERT2-FRT systems showed compelling advantages in tightly tracing human lineages temporally. In proof-of-principle experiments, we traced human PAX6⁺ neuroectoderm cells and revealed their full neural lineage differentiation potency both in vitro and in vivo. Devising and optimizing of lineage-tracing systems in hESCs will thus set up a solid foundation for human developmental studies.

Keywords: FOXA2; PAX6; development; genetic engineering; human embryonic stem cells; lineage-tracing.

Graphical abstract

Figure 1

Conditional Tracer Integrated in the Human AAVS1 Locus Is Recombinase Controllable (A) Schematic diagram for constructing of AAVS1-LSL-GFP hPSCs lines through TALEN-mediated gene targeting. Genotyping PCR primer sets are labeled with arrows. Location of restriction enzyme cutting sites for Southern blot analysis are marked with black bars. The 5′ arm and GFP probes for Southern blot analysis are marked with blue bars at indicated genetic regions after homologous recombination (HR). (B and C) Genomic DNA PCR results identify monoallelic (B) and biallelic (C) HR colonies after genetic engineering. (D and E) Southern blot analyses of wild-type (WT), monoallelic, and biallelic HR colonies with 5′ arm probe (D) and GFP probe (E) show precise gene targeting. (F) Bright-field (BF) and fluorescent GFP images of AAVS1-LSL-GFP hPSCs line infected with or without lentivirus expressing Cre. Scale bars, 100 μm.

Figure 2

Tight Control and Efficient Cleavage of PAX6-P2A-Cre or FOXA2-P2A-Cre Fusion Proteins (A) Schematic diagram for constructing PAX6-P2A-Cre and FOXA2-P2A-Cre hPSCs lines through a gRNA-guided CRISPR/Cas9 system. Genotyping PCR primer sets are labeled with arrows. (B) Genomic DNA PCR results identify five monoallelic HR colonies after genetic engineering. (C) Immunolabeling of FOXA2 and Cre in FOXA2-P2A-Cre lines differentiated toward a cortical NP or FP fate for 6 or 12 days. Nuclei are counter stained with Hoechst. Scale bar, 100 μm. (D) Western blot analysis of FOXA2-P2A-Cre lines differentiated toward a cortical NP or FP fate for 6, 12, or 20 days. (E) Genomic DNA PCR results identify two monoallelically recombined PAX6-P2A-Cre colonies after genetic engineering. (F) Immunolabeling of PAX6 and Cre in PAX6-P2A-Cre lines at hESC stage, day 10 NE, day 17 cortical NPs, or day 17 ventral NPs. Nuclei are counter stained with Hoechst. Scale bar, 100 μm. (G) Western blot analysis of PAX6-P2A-Cre lines along with neural differentiation. See also Figure S1.

Figure 3

Tracing PAX6 Expressing NE In Vitro and In Vivo (A) AAVS1 left and right TALENs and AAVS1-LSL-GFP donor plasmids are electroporated into PAX6-P2A-Cre H9 hESCs. Genomic DNA PCR analysis identifies monoallelically and biallelically targeted colonies. (B) Southern blot analysis of no. 4 biallelic HR colony shows precise engineering without off-target recombination. (C) Immunostaining of GFP, PAX6, Cre, and NKX2.1 in PAX6-P2A-Cre/AAVS1-LSL-GFP H9 hESC line at day 17 dorsal or ventral NPs reveals that both regional NPs are developed from PAX6 expression NE. Scale bar, 100 μm. (D) H&E and immunolabeling results of adjacent sections from PAX6-P2A-Cre/AAVS1-LSL-GFP H9 hESC line-derived teratomas. GFP labels all existing neural tubes (NT), but it is not present in cartilages (CT) or intestine tissues (IN). Scale bar, 200 μm. (E and F) H&E and immunolabeling results of adjacent sections from PAX6-P2A-Cre/AAVS1-LSL-GFP H9 hESC line-derived teratomas. Both PAX6⁺ dorsal NPs and NKX2.1⁺ ventral NPs are labeled with GFP. Scale bar, 100 μm. See also Figure S2.

Figure 4

The Cre-LoxP Recombination System Is Hypersensitive in Human Cells (A) Genomic DNA PCR results of retrieved colonies after recombination of AAVS1-LSL-GFP into FOXA2-P2A-Cre seed H9 hESC line. (B) Southern blot analysis of no. 3 FOXA2-P2A-Cre/AAVS1-LSL-GFP colony with GFP probe. (C) Immunolabeling with GFP, OCT4, Cre, and FOXA2 of no. 3 FOXA2-P2A-Cre/AAVS1-LSL-GFP H9 hESC line maintained in a pluripotent stage or 2 days post differentiation toward an FP fate. Scale bar, 100 μm. (D) Bright-field (BF) and fluorescent images of FOXA2-P2A-Cre/AAVS1-LSL-GFP hESC line at different passages indicated. Scale bar, 100 μm. See also Figure S3.

Figure 5

A Series of LSL Variants with Distinct Recombination Sensitivities (A) Sequences of LSL and LSL variants. Mutated nucleotides are labeled in red. (B) HEK29 cells are transfected with pLenti-Cre and AAVS1-LSL-GFP or AAVS1-LSLm1-4-GFP plasmids. GFP⁺ cells are calculated to evaluate the sensitivity of Cre-mediated recombination of each LSL variant. Data are presented as mean ± SEM of three independent experiments. (C) Genomic DNA PCR results of retrieved colonies after recombination of AAVS1-LSLm1-GFP or AAVS1-LSLm2-GFP into FOXA2-P2A-Cre seed H9 hESC line. (D) Bright-field (BF) and fluorescent images of FOXA2-P2A-Cre/AAVS1-LSLm2-GFP hESC line at passages 35 and 49. Scale bar, 100 μm. (E and F) Immunolabeling of GFP, FOXA2, and Cre in FP differentiation derivatives of FOXA2-P2A-Cre/AAVS1-LSLm2-GFP (E) or FOXA2-P2A-Cre/AAVS1-LSLm1-GFP (F) lines at different days indicated. Scale bars, 100 μm. See also Figure S4.

Figure 6

Improvement of the Tracing Fidelity by Introducing Controllable Recombinases (A) Schematic diagram for generating FOXA2-Cre^ERT2 lines. Primer sets for genomic DNA PCR are labeled with arrows. (B and C) Genomic DNA PCR results showing recombination of Cre^ERT2 at the ATG region of FOXA2 (B) and CAG-LSL-GFP (C) cassettes within the AAVS1 intron. (D) Southern blot analysis of no. 4 FOXA2-Cre^ERT2/AAVS1-LSL-GFP colony with GFP probe. (E and F) Immunolabeling of GFP, FOXA2, Cre, and PAX6 in dorsal NPs and FP cells with (E) or without (F) 4-OHT treatment at days 6 or 12 post differentiation. 4-OHT was administered 2 days before staining. Scale bars, 100 μm. See also Figure S5.

Figure 7

Temporal Tracing of PAX6 Expression Cells Identifies their Sequential NE and Dorsal NP Identities (A) Schematic diagram for generating PAX6-Cre^ERT2 lines. Primer sets for genomic DNA PCR are labeled with arrows. (B and C) Genomic DNA PCR results showing recombination of Cre^ERT2 at the ATG region of PAX6 (B) and CAG-LSL-GFP (C) cassettes within the AAVS1 intron. (D) Immunolabeling of GFP, PAX6, Cre, and NKX2.1 in dorsal and ventral NPs at day 17 in PAX6-Cre^ERT2/AAVS1-LSL-GFP H9 hESC line treated with 4-OHT through days 8–10. Scale bar, 100 μm. (E) Immunolabeling of GFP, PAX6, Cre, and NKX2.1 in dorsal and ventral NPs at day 17 in PAX6-Cre^ERT2/AAVS1-LSL-GFP H9 hESC line treated with or without 4-OHT through days 15–17. Scale bar, 100 μm. See also Figure S6.