Bi-directional gene activation and repression promote ASC differentiation and enhance bone healing in osteoporotic rats

Abstract

Calvarial bone healing is challenging, especially for individuals with osteoporosis because stem cells from osteoporotic patients are highly prone to adipogenic differentiation. Based on previous findings that chondrogenic induction of adipose-derived stem cells (ASCs) can augment calvarial bone healing, we hypothesized that activating chondroinductive Sox Trio genes (Sox5, Sox6, Sox9) and repressing adipoinductive genes (C/ebp-α, Ppar-γ) in osteoporotic ASCs can reprogram cell differentiation and improve calvarial bone healing after implantation. However, simultaneous gene activation and repression in ASCs is difficult. To tackle this problem, we built a CRISPR-BiD system for bi-directional gene regulation. Specifically, we built a CRISPR-AceTran system that exploited both histone acetylation and transcription activation for synergistic Sox Trio activation. We also developed a CRISPR interference (CRISPRi) system that exploited DNA methylation for repression of adipoinductive genes. We combined CRISPR-AceTran and CRISPRi to form the CRISPR-BiD system, which harnessed three mechanisms (transcription activation, histone acetylation, and DNA methylation). After delivery into osteoporotic rat ASCs, CRISPR-BiD significantly enhanced chondrogenesis and in vitro cartilage formation. Implantation of the engineered osteoporotic ASCs into critical-sized calvarial bone defects significantly improved bone healing in osteoporotic rats. These results implicated the potential of the CRISPR-BiD system for bi-directional regulation of cell fate and regenerative medicine.

Keywords: CRISPR-AceTran; CRISPR-BiD; CRISPRi; bi-direction gene regulation; regenerative medicine.

Graphical abstract

Figure 1

Development of CRISPR-AceTran to potentiate Sox Trio activation and chondrogenesis (A) Illustration of the hybrid baculovirus vectors Bac-p300-VPR (harboring the CRISPR-AceTran system) and Bac-Cre expressing Cre recombinase, as well as the genomic loci of the eight gRNAs targeted. The CRISPR-AceTran system consisted of two parts: one cassette expressing dSpCas9-p300 and MCP-VPR (separated by the self-cleavage P2A peptide) under the control of the rat elongation factor 1α (rEF-1α) promoter; the other part was an array of eight human U6 (hU6) promoter-driven gRNA expression cassettes targeting the Sox Trio (Sox5, Sox6, and Sox9) promoter regions. Bac-p300-VPR transduction of cells would enable co-expression of dSpCas9-p300, MCP-VPR, and the eight scaffold gRNAs. (B) Mode of action of CRISPR-AceTran. dSpCas9-p300 would coordinate with gRNA to bind the target genes for histone acetylation and chromatin relaxation. The MS2 aptamers on the gRNA recruit MCP-VPR for transcription activation. (C–E) Sox5, Sox6, and Sox9 expression levels as measured by qRT-PCR at 3 dpt. (F and G) Expression levels of chondrogenic markers Acan and Col2a1 at 7 and 14 dpt. (H) Safranin O staining of cells cultured in chondroinduction medium at 21 dpt. (I) Spectrophotometric analysis of Safranin O stain. ASCs and OVXASCs were mock transduced (mock group), transduced with Bac-p300-VPR at a multiplicity of infection (MOI) of 200 (C–E) or co-transduced with Bac-p300-VPR and Bac-Cre at an MOI of 200 for each (F–I). Imaging data are representative of three independent culture experiments. qRT-PCT data represent means ± SD of three independent culture experiments. The data of experimental groups were normalized to those of the mock groups. p values were calculated by an unpaired two-tailed Student’s t tests (C–E) or two-way ANOVA with a Holm-Sidak post hoc test (F, G, and I). WPRE, Woodchuck hepatitis virus posttranscriptional regulatory element; SV40 pA, SV40 virus polyadenylation signal; loxP, loxP recognition sequences; NLS, nuclear localization signal. ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, ∗∗∗∗p ≤ 0.0001; ns, not significant.

Figure 2

CRISPRi using dSaCas9-DNMT3A for repression of adipogenesis in ASCs and OVXASCs (A) Illustration of Bac-D3A that harbored the CRISPRi system. The CRISPRi system included the dSaCas9-DNMT3A fusion gene under the control of the rEF-1α promoter and the gRNA array consisting of six hU6-driven gRNA expression cassettes targeting C/ebp-α and Ppar-γ. (B and C) Expression levels of C/ebp-α and Ppar-γ as measured by qRT-PCR at 3 dpt. (D) Expression levels of adipogenic marker Fabp4 at 7 and 14 dpt. (E and F) Western blot and densitometry analyses of FABP4 protein level at 7 and 14 dpt. (G) Oil red O staining at 21 dpt. (H) Spectrophotometric analysis of oil red O stain at 21 dpt. ASCs and OVXASCs were mock transduced (mock group), transduced with Bac-D3A (D3A group) at an MOI of 200 (B and C), or co-transduced with Bac-D3A and Bac-Cre (D3A/Cre group) at an MOI of 200 for each (D–H). The cells were induced toward the adipogenic pathway by culturing in adipoinduction medium. Imaging data are representative of three independent culture experiments. qRT-PCT and western blot data represent means ± SD of three independent culture experiments. Quantitative data of the experimental groups were normalized to those of the mock groups. p values were calculated by an unpaired two-tailed Student’s t test (B and C) or a two-way ANOVA with a Holm-Sidak post hoc test (D, F, and H). ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, ∗∗∗∗p ≤ 0.0001.

Figure 3

CRISPR-BiD promoted OVXASC chondrogenesis and cartilage formation (A) Safranin O staining. (B) Quantitative analyses of Safranin O staining. (C) Gross appearance of engineered cartilage. (D) H&E staining. (E) Alcian blue staining. (F) Quantitative analysis of GAG. (G) Quantitative analysis of Col II. OVXASCs were mock transduced (mock group) or co-transduced with Bac-p300-VPR/Bac-D3A/Bac-Cre (CRISPR-BiD group) in 15-cm dishes and cultured in chondrogenic induction medium. After overnight culture, the cells continued to be cultured in chondroinductive medium for 21 days or were seeded into porous gelatin discs (diameter, 6 mm; thickness, 1 mm; 5 × 10⁶ cells/disc; n = 9). The OVXASC/gelatin constructs were cultured in chondroinductive medium for 21 days until analyses. Images are representative of three independent culture experiments. Quantitative data represent means ± SD of three independent culture experiments. p values were calculated by an unpaired two-tailed Student’s t test. ∗∗∗∗p ≤ 0.0001; nd, not detectable.

Figure 4

μCT evaluation of in vivo bone regeneration (A) Transverse view. (B) 3D projection images. (C–E) Quantitative analysis of bone area, bone volume, and bone density. OVXASCs were mock transduced (mock group) or co-transduced with Bac-p300-VPR/Bac-D3A/Bac-Cre (CRISPR-BiD group) and seeded into gelatin discs (n = 6 each) as in Figure 3. The OVXASC/scaffold constructs were implanted into the critical-size defects (diameter, 6 mm) at the calvaria of osteoporotic (OVX) rats. Bone healing was assessed by μCT imaging at weeks 4 and 8 after implantation, and the quantitative data were normalized to those of the mock group at week 4. p values were calculated by two-way ANOVA with a Holm-Sidak post hoc test. ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, ∗∗∗∗p ≤ 0.0001.

Figure 5

Histological evaluation of regenerated bone (A and B) H&E staining. (C–H) Immunostaining for BSP (C and D), OPN (E and F), and OCN (G and H). (I–K) Quantitative analysis of BSP, OPN, and OCN intensity. The animal experiments were performed as in Figure 4, and the animals were sacrificed at week 8. The data are representative of six bone sections. Average intensity of each bone section was derived from ImageJ processing of five images covering the entire defect. FT, fibrous tissue; NaB, native bone; NB, new bone; OB, osteoblast; OS, osteocyte. Filled arrowheads indicate OB. Vacant arrowheads indicate OS. Arrows indicate BSP, OCN, or OPN. p values were calculated by an unpaired two-tailed Student’s t test. ∗∗∗p ≤ 0.001.