The Combination of CRISPR/Cas9 and iPSC Technologies in the Gene Therapy of Human β-thalassemia in Mice

β-thalassemia results from point mutations or small deletions in the β-globin (HBB) gene that ultimately cause anemia. The generation of induced pluripotent stem cells (iPSCs) from the somatic cells of patients in combination with subsequent homologous recombination-based gene correction provides new approaches to cure this disease. CRISPR/Cas9 is a genome editing tool that is creating a buzz in the scientific community for treating human diseases, especially genetic disorders. Here, we reported that correction of β-thalassemia mutations in patient-specific iPSCs using the CRISPR/Cas9 tool promotes hematopoietic differentiation in vivo. CRISPR/Cas9-corrected iPSC-derived hematopoietic stem cells (HSCs) were injected into sublethally-irradiated NOD-scid-IL2Rg−/− (NSI) mice. HBB expression was observed in these HSCs after hematopoietic differentiation in the NSI mice. Importantly, no tumor was found in the livers, lungs, kidneys, or bone marrow at 10 weeks in the NSI mice after implantation with these HSCs. Collectively, our findings demonstrated that CRISPR/Cas9 successfully corrects β-thalassemia mutations in patient-specific iPSCs. These CRISPR/Cas9-corrected iPSC-derived HSCs express normal HBB in mice without tumorigenic potential, suggesting a safe strategy for personalized treatment of β-thalassemia.


Results
Gene-correction of HBB mutation in β-thalassemia iPSCs by CRISPR/Cas9. In order to correct the mutation of the HBB gene in the piPSCs, CRISPR/Cas9-gRNA was used to cleave the iPSCs genome, whereas the polymerase chain reaction (PCR) products with the normal gene information surrounding the mutation site were used as the donor (Fig. 1A). To confirm Cas9-gRNAs cleavage of the genome of β-iPS-41/42 cells, the Cas9-gRNA vector was transfected into the piPSCs, and then the genomic DNA was extracted following PCR amplification of the surrounding sequence. T7E1 assay confirmed the activity of gRNA (Fig. 1B). To further obtain the gene correction clones, Cas9-gRNA vector plus PCR products were cotransfected into β-iPS-41/42 cells and then selected with puromycin for two days. One week later, the clones were identified by sequencing to observe whether the mutation in the piPSCs was corrected. Indeed, β-thalassemia mutations in piPSCs were corrected by the CRISPR/ Cas9 system after sequencing (Fig. 1C). These CRISPR/Cas9-corrected iPSCs (ciPSCs) showed a tightly compacted morphology with high nuclear/cytoplasm ratio clonal (Fig. 1D) and normal karyotypes (46, XY) (Fig. 1E). Short tandem repeat (STR) analyses with 16 loci further revealed the same origin with identical signal peaks in both piPSCs and ciPSCs, suggesting that the ciPSCs were derived from the piPSCs (Fig. 1F). Collectively, these findings indicate that the HBB gene mutation in piPSCs is corrected by CRISPR/Cas9.
To further confirm these findings in vivo, cells were subcutaneously injected into immunodeficient (SCID) mice. A significant change to teratoma formation in SCID mice was observed at eight weeks after injection of ciPSCs. Histological examinations further confirmed that the teratomas contained various tissues comprising all three germ layers (endoderm, mesoderm, and ectoderm) (Fig. 2C). As a positive control, human embryonic stem cells (hESCs) exhibited a similar phenotype as ciPSCs in pluripotency experiments. There were no differences between these iPSCs and hESCs in their ability to either self-renew or differentiate.
Hematopoietic differentiation in vitro. Stem cell co-culture with stromal cell lines such as S17, OP9, and fetal liver-derived stromal cells can result in spontaneous generation of hematopoietic cells 33 . To assay the hematopoietic differentiation ability of piPSCs and ciPSCs, all iPSCs and hESCs were co-cultured with OP9 stromal cells, separately. These cells exhibited dynamic similarity in morphology (Fig. 2D). Notably, both uncorrected and corrected ciPSCs had similar abilities to differentiate and produce HSCs after co-culture with stromal cells. Flow cytometry analysis showed that the proportion of CD34 + cells derived from the piPSCs and ciPSCs at day 10 was 2.8% and 3.1%, respectively (Fig. 2E). These results demonstrated that both piPSCs and ciPSCs undergo successful hematopoietic differentiation in vitro. Hematopoietic differentiation in vivo. NSI mice lack mature T cells, B cells, and natural killer cells and therefore permit the engraftment of a wide range of primary human cells. We therefore used NSI mice to observe hematopoietic differentiation of ciPSCs in vivo. Female NSI mice with similar weights were sub-lethally irradiated with 1.0 Gy/min X-ray for 1 min and then randomly divided into six groups: the α-MEM group, piPSC group, ciPSC group, niPSC (iPSCs derived from normal human blood) group, UCB (Umbilical Cord Blood) group, and the hESC group. Then, 5 × 10 5 CD34 + HSCs derived from these cells were enriched by magnetic-activated cell sorting (MACS) and transplanted into the right femurs of the irradiated NSI mice. Routines blood tests were performed every week for four weeks after transplantation. Both HB and red blood cells (RBCs) in the peripheral blood dropped after irradiation (P < 0.05). HB levels in the ciPSC group, niPSC group, UCB group, and hESC group were higher than in the α-MEM group and piPSC group at two weeks after transplantation (Table 1). However, there were no differences among these groups for all weeks and indexes (P > 0.05).
The cluster of differentiation (often abbreviated as CD) is widely used for the identification and immunophenotyping of cells. We next analyzed CD marks in the blood using flow cytometry analysis. Anti-human CD3 (present on 70-80% of normal human peripheral blood lymphocytes and 60-85% of thymocytes), anti-human CD45 (present on all human leukocytes including lymphocytes, monocytes, granulocytes, eosinophils, and thymocytes), anti-human CD71 (expressed on erythroid progenitors), and anti-human CD235a (expressed on human erythrocytes and erythroid precursor cells) were performed every week, whereas anti-human CD8 (expressed by the majority of thymocytes), anti-human CD31 (expressed on platelets, monocytes, granulocytes, and in high amounts on endothelial cells), and anti-human CD43 (expressed on T cells, pre-B cells and activated B cells, NK cells and granulocytes) were detected every other week after transplantation. The levels of these CD markers in the piPSC, ciPSC, niPSC, UCB, and hESC groups were higher than those in the α-MEM group (Fig. 3A,B). Analysis of the α-MEM, piPSC, ciPSC, and niPSC groups are also shown in Fig. 3C.
All of the mice were sacrificed and the bone marrow (BM) cells from the transplantation (right femur) and the non-transplantation sides (left femur) were collected and analyzed with flow cytometry using specific anti-human CD3, CD45, CD71, CD235a, CD34 (expressed on hematopoietic progenitor cells), CD31, and CD43 at 10 weeks after transplantation. The levels of these CD markers in the piPSC, ciPSC, niPSC, UCB, and hESC groups were higher than those in the α-MEM group. Analysis of the α-MEM, piPSC, ciPSC, and niPSC groups is shown in Fig. 4. Analyses of these CD markers in the peripheral blood and BM demonstrated that hematopoiesis of the transplanted human cells occurred in the NSI mice.
The sex-determining region Y (SRY) is a gene found on Y chromosomes that leads to the development of male phenotypes. RT-PCR was further used to detect the human SRY gene in the blood and BM cells from the transplantation and the non-transplantation sides of the female NSI mice. If the ciPSC transplant fails, no human SRY gene can be tested in female NSI mice. The human SRY gene was detected in the ciPSC group (Fig. 5), further confirming that hematopoiesis of the transplanted human cells had occurred in the NSI mice.
To address whether ciPSC-derived HSCs can produce HBB proteins in vivo, we further tested HBB proteins levels using two different testing methods. First, 2 μl of mouse peripheral blood cells underwent flow cytometry analysis with specific anti-human CD235a and anti-human HBB. Indeed, CD235a and HBB positive cells were found in both the piPSC and ciPSC groups, but the ratio of HBB positive cells in the ciPSC group was higher than that in the piPSC group (P < 0.01) (Fig. 6A,B). Secondly, 1 ml of peripheral blood cells of the mice was sorted with specific anti-human CD235a by flow cytometry. These CD235a positive cells were further used to detect HBB protein levels by western blot analysis. The expression of HBB protein in the ciPSC group was higher than that in the piPSC group (P < 0.05) ( Fig. 6C-E). These findings indicate that ciPSC-derived HSCs can produce normal HBB proteins in vivo.
Pathological results. The potential tumorigenicity of iPSCs is a major concern in the clinical development of human iPSCs for therapy. Key factors that influence the tumorigenicity of iPSCs include the reprogramming method, the selection of clonal lines, and how the iPSCs are treated thereafter. We next evaluated the tumorigenic potential of ciPSCs in tissues from mice by histopathology analysis. Indeed, no tumors were observed in the livers, lungs, kidneys, or BM 10 weeks after injection (Fig. 7). Moreover, no unexplained premature deaths were observed after these mice received transplants. Thus, our studies indicate that ciPSC-derived HSCs express normal HBB in mice without tumorigenic potential.

Discussion
Advances in the development of genome engineering tools are allowing us to cure human diseases caused by genetic mutations. β-thalassemia syndromes are a group of hereditary disorders characterized by a genetic deficiency in the synthesis of HBB. We and others have shown that the generation of HSCs from the iPSCs of patients with subsequent gene correction could recover the production of HBB 17,[25][26][27][28][29] . Notably, the CRISPR/Cas9 system is the most popular homologous recombination-based gene correction method, which has proven to be a powerful tool for targeted genetic modification 17,29,34,35 . In our study, the genetically corrected homozygous 41/42 deletion of patient-specific iPSCs (one allele of the mutated HBB gene corrected) also maintained normal karyotypes (46, XY), retained their pluripotency, and differentiated into HSCs. However, before the clinical application of iPSCs, a number of hurdles must be overcome. Mysteries remain, such as the ability of HSCs from the gene correction of patient iPSCs in vivo to survive and differentiate; HSC recovery of the production of HBB in vivo; and whether tumors form after transferring HSCs. Therefore, we focused on HSCs differentiated from CRISPR/ Cas9-corrected iPSC in vivo.
In this study, we used NSI mice, which with a more severely impaired immune system, attained a higher tumor engraftment index score and were more suitable for xenograft and allograft experiments 36 . Typical routine blood test results showed that the levels of both HB and RBCs in the peripheral blood dropped after irradiation. HB levels of transplanted mice in the ciPSC, niPSC, UCB, and hESC groups were more highly elevated than those in Scientific RepoRts | 6:32463 | DOI: 10.1038/srep32463 Weight (g) RBC (×10 12 /l) HB (g/l) RE 10 9 /L RET (%) IFR (%) Week 0 the α-MEM group and piPSC group at two weeks after transplantation, but there was no statistical significance. Unlike our study, another study indicated that genetically corrected iPSC-derived HSCs could positively stimulate hematopoiesis in SCID mice after irradiation 26 . Some of the possible reasons for the difference would be the change in numbers, strains of animals, and quality of the HSCs.
In order to confirm that HSCs from the iPSCs of patients could survive and differentiate in vivo, flow cytometry analysis of human CD markers were performed. Anti-human CD3, CD45, CD71, CD31, CD34, CD43, and CD235a were detected in BM cells from the transplantation and the non-transplantation sides of the mice. The presence of the human SRY gene was also detected in genomic DNA extracted from the peripheral blood and BM cells from both sides of the cell-transplanted female NSI mice. These findings confirmed that the transplanted human cells underwent hematopoiesis in the NSI mice.
Furthermore, flow cytometry analysis and western blot were used to detect human HBB proteins in the peripheral blood. First, flow cytometry analysis with specific anti-human CD235a and anti-human HBB was performed. Secondly, flow cytometry was sorted with specific anti-human CD235a, and then the CD235a cells were used to detect the HBB proteins with western blot. Both analyses suggested that the genetically corrected homozygous mutation of β-thalassemia iPSCs could produce HBB after hematopoietic differentiation in vivo. These results were consistent with findings from a previous study using quantitative PCR analysis, high-performance liquid chromatography, and western blot analysis 26 .
In summary, we demonstrated that HSCs from the correction of homozygous mutation of β-thalassemia iPSCs could survive, differentiate into RBCs, and produce β-globin after hematopoietic differentiation in vivo. There were no observations of tumor formation in NSI mice after ciPSC transplantation. Our results highlight a safe gene therapy strategy of combining iPSCs and CRISPR/Cas9 technology to treat β-thalassemia. Further studies are needed to observe the innate immune response to this bacterial original CRISPR/Cas9 genome editing tool.   gRNA assembly and PCR donor purification. Sequences used for cloning the gRNA into the PX459 vector were 5′-CACCGCCCCAAAGGACTCAACCTC-3′ and 5′-AAACTTGGACCCAGAGGTTGAGTCC-3′. To prepare the PCR donor, we used the primers 5′-GAGAAGACTCTTGGGTTTCTGATAG-3′ and 5′-CAGCTCAC TCAGTGTGGCA-3′ to amplify, and used the PCR products as donor. 5′-GACAGAGAAGACTCTT GGGTTTCTGATA-3′ and 5′-CAGCTCACTCAGTGTGGC A-3′ were used to perform the T7E1 assay.

MACS cell sorting for CD34 + cells.
Cells from hESCs/OP9, iPSCs/OP9 co-cultures, and human umbilical cord blood were collected, washed with MACS buffer (PBS with 2% FBS), and incubated with CD34 MicroBeads (Miltenyi Biotec) for magnetic labeling. Then cells were loaded onto the column. Magnetic separation was performed in the magnetic field according to instructions.
Bone marrow transplantation. Twelve to 24 hours before transplantation, NSI (NOD-scid-IL2Rg−/−, provided from Dr. Peng Li, which with a more severely impaired immune system, attained a higher tumor engraftment index score and were more suitable for xenograft and allograft experiments 36 ) mice at the age of six to eight weeks old were sublethally irradiated with 1.0 Gy/min for 1 min (RS2000 X-ray). Mice were randomly divided into six groups: the α-MEM group (n = 4), piPSC group (n = 5), ciPSC group (n = 5), niPSC group (n = 4), UCB group (n = 3), and the hESC group (n = 4). For transplantation, 5 × 10 5 MACS sorted CD34 + HSCs were collected and resuspended in 20 μl α-MEM medium. After anesthesia, the cell suspension in the 28 G insulin syringe was gently injected into the right femurs of the mice.  Cells were then stained with the anti-human FITC-labeled HB β monoclonal antibody (Santa Cruz Biotech). For sorting of the human CD235a + cells, the peripheral blood cells of mice were collected and stained with anti-human CD235-APC monoclonal antibody (BD) and then sorted according to the instructions. Flow cytometry was performed using FACSAria ™ III flow cytometer (BD) and analyzed with FACSDiva ™ software (BD).

Western blot analysis.
After sorting, the CD235 + cells were washed twice with cold PBS and then extracted into RIPA lysis buffer with 1 mM PMSF (Beyotime, China) on ice as previously described 38,39 . After centrifugation  at 15,000 × g for 5 min, the protein content was determined using a BCA protein assay kit (Thermo Fisher) according to the manufacturer's instructions. Equivalent amounts of protein were separated on 15% sodium dodecyl sulfate (SDS)-polyacrylamide gels and blotted onto nitrocellulose membranes. After being blocked at room temperature for 2 h with 5% nonfat milk in TBS with 0.1% Tween-20, the membranes were probed with anti-rabbit human HB beta polyclonal antibody (1:1000; Abgent), anti-GAPDH (1:5000; Cell Signaling Technology), and horseradish peroxidase-conjugated IgG antibodies (1:10000; Cell Signaling Technology). The protein bands were then visualized by exposure to film (Fuji). The volumes of the bands were determined by standard scanning densitometry with normalization of densitometry measures to the expression of GAPDH.
Statistical analysis. Data are expressed as the mean ± standard error of the mean (SEM) and were compared by one-way analysis of variance (ANOVA). When ANOVA results were significant, differences between groups were assessed with post-hoc testing using LSD tests (SPSS version 17.0 for Windows). Differences with P < 0.05 were considered statistically significant.