目的：CRISPR/Cas9作为一种常用的基因编辑技术，已广泛用于哺乳动物细胞的基因编辑研究。然而小鼠原代造血干细胞(HSC)由于数量稀少，难以被携带大片段的外源基因感染，因此在原代HSC水平上进行基因编辑仍然比较困难。本课题主要利用CRISPR/Cas9技术建立了在小鼠原代HSC水平进行基因敲除的方法。

方法：首先利用Cas9-EGFP基因敲入小鼠与Cre-ER小鼠杂交产生可诱导表达Cas9蛋白的Cas9-EGFP/Cre-ER转基因小鼠模型；利用CHOPCHOP网站设计靶向小鼠Pax5基因的sgRNA序列，以sgScramble作为对照，并构建包含该序列的慢病毒载体并包装病毒；利用流式细胞术分选该小鼠骨髓原代HSC(HSC1: Lin^-c-Kit⁺Sca-1⁺CD150⁺CD48^-CD201⁺; HSC2: Lin^-c-Kit⁺Sca-1⁺CD150^-CD48^-CD201⁺)细胞，使用sgPax5-Crimson慢病毒进行体外感染以实现对Pax5基因的敲除，以sgScramble-Crimson作为对照，并对感染后细胞进行流式分析检测感染效率；同时提取感染后细胞的基因组DNA进行Pax5基因的靶向测序以检测基因敲除效率；将感染后细胞移植进入受致死剂量照射的受体小鼠，移植后监测在HSC水平敲除Pax5基因后对造血重建及谱系分化的影响。

结果：

1. Cas9-EGFP基因敲入小鼠(B6;129)首先与C57BL/6J(CD45.2)回交十代，以获得纯合的CD45.2遗传学背景，之后与Cre-ER^T2(CD45.2)小鼠进行交配，根据基因型鉴定结果从子代中筛选获得Cas9^fl/-Cre^+/-双阳性子代小鼠。

2. 给予Cas9^fl/-Cre^+/-小鼠连续五天腹腔注射他莫昔芬诱导Cas9-EGFP的表达：两周后外周血中只有髓系细胞(Mac-1/Gr-1⁺)能够表达EGFP(20%)；四周后外周血中除成熟红细胞(Ter119⁺)外，淋系细胞(B: B220⁺; T: CD3/4/8⁺)、髓系细胞以及血小板(CD41⁺)都能够表达一定比例的EGFP(10%-20%)；八周后外周血各谱系细胞EGFP表达稳定，其比例与诱导四周后EGFP比例无统计学差异。

3. 基于造血干祖细胞单细胞RNA测序及移植结果发现：HSC2(Lin^-c-Kit⁺Sca-1⁺CD150^-CD48^-CD201⁺)为一群具有多谱系分化能力但偏向淋系分化的HSC，在淋系细胞的成熟发育过程中发挥着重要的作用。

4. 使用sgPax5/sgScramble-Crimson慢病毒对HSC(HSC1以及HSC2进行体外感染72h后检测其感染效率：HSC1的Pax5敲除组和对照组的感染效率分别为69.0 %± 21.0 vs.74.8 %± 13.6；HSC2的Pax5敲除组和对照组的感染效率分别为51.9 %± 33.2 vs. 61.4% ± 17.7。

5. 使用sgPax5/sgScramble -Crimson慢病毒对HSC进行体外感染48h后，再与OP-9小鼠基质细胞共培养7天，分选阳性感染细胞并提取全基因组DNA进行Pax5靶向测序，结果显示CRISPR/Cas9的编辑效率为92%。

6. Pax5敲除组外周血中供体来源的细胞总重建率无明显差异：其中髓系细胞与T细胞重建正常，而敲除组受体小鼠外周血中缺失成熟B细胞。HSC1与HSC2组表型相似，无明显差异。

7. 与对照组相比，Pax5敲除组小鼠骨髓中供体来源的B细胞比例明显升高，其中以pro-B为主(45.3%± 15.7 vs .7.0% ± 12.6, P＜0.01)；T淋巴细胞以及髓系细胞比例相对于对照组未有显著变化；造血祖细胞MPP3比例的明显增高(0.49% ± 8.2 vs. 2.8% ± 25.7, P＜0.01)；而HSC比例无显著改变。HSC1与HSC2组无明显差异。

结论：

1. 本方法通过优化HSCs体外培养体系及简化外源性质粒大小，建立了在少量高度纯化的HSCs水平的一种高效且便捷的基因敲除方法。

2. 基于此方法，本研究成功在小鼠骨髓原代HSC水平敲除Pax5，导致骨髓中出现了Pro-B的阻滞现象，外周血中成熟B细胞缺失现象，与既往研究结果一致。

Objective: CRISPR/Cas9 is a commonly used gene editing technology that has been widely used in the study of gene editing in mouse and human cells. Primary hematopoietic stem cells (HSCs) are difficult to be infected by exogenous genes carrying large fragments due to their scarce cell number, so gene editing in primary HSCs is very difficult. In this topic, we mainly use CRISPR/Cas9 technology to establish the methodology of gene editing at the HSC level.

Methods: Cas9-EGFP/Cre-ER transgenic mouse models with inducible expression of Cas9 protein were first generated by crossing Cas9-EGFP knockout mice with Cre-ER mice. Primary HSC1 as well as HSC2 cell cells from this mouse were sorted using flow cytometry and infected using sgPax5-Crimson lentivirus to knock out the Pax5 gene, and the results showed that the infection efficiency of the lentivirus was able to reach 95%. The infected cellular DNA was subjected to targeted sanger sequencing, which showed a knockout efficiency of 97.85%. The infected cells were then transplanted into lethally irradiated recipients to examine the effect of HSC knockout of the Pax5 gene on reconstitution and lineage differentiation

Results:

1) Cas9-EGFP knock-in mice were first backcrossed with C57BL/6J (CD45.2) for ten generations to ensure a pure CD45.2 genetic background, and then crossed with Cre-ERT2 (CD45.2) mice to finally obtain Cas9^fl/-Cre ^+/− positive offspring mice.
2) Two weeks after Tamoxifen injection, only myeloid cells in peripheral blood could express EGFP; at four weeks after induction, lymphoid cells, myeloid cells, and platelets in peripheral blood could express a certain proportion of EGFP in addition to mature erythrocytes; at the 8th week after induction, the expression of EGFP in peripheral blood lineage cells was in a stable state.
3) According to the results of single-cell RNA sequencing and single-cell transplantation, it was found that HSC2 is a group of hematopoietic stem cells with multi-lineage differentiation ability but lymphoid-bias, which plays an important role in the generation and development of lymphoid cells.
4) HSCs (HSC1 and HSC2) were infected in vitro with sgPax5/sgScramble-Crimson lentivirus for 72 h. The infection efficiency of HSC1 Pax5 knockout and control group was 69.0% ± 21.0 vs. 74.8% ± 13.6, respectively; the infection efficiency of Pax5 knockout control group of HSC2 was 51.9% ± 33.2 vs. 61.4% ± 17.7, respectively.
5) After 48 hours of in vitro infection of HSCs with sgPax5/sgScramble-Crimson lentivirus, they were transferred to co-cultured with OP-9 mouse stromal cells for 7 days. Whole genomic DNA of infection-positive cells was extracted, and Pax5-targeted sequencing was performed. The results showed that the editing efficiency of CRISPR/Cas9 was about 92%.
6) The total reconstitution rate in the peripheral blood of Pax5 knockout recipient mice was lower than that of the control group, in which myelocytes and T cells were able to reconstitute normally, while mature B cells were lacked in the peripheral blood. There is no big difference between HSC1 and HSC2 group.
7) After Pax5 knockout, the proportion of B cells in the bone marrow of recipient mice was significantly increased, of which pro-B(45.3%± 15.7 vs .7.0% ± 12.6, P＜0.01) was predominant; the proportion of T lymphocytes and myeloid cells did not change greatly compared with the control group; in the level of hematopoietic progenitor cells, a significant increase in the proportion of MPP3(0.49% ± 8.2 vs. 2.8% ± 25.7, P＜0.01) occurred; there was no significant effect at the level of hematopoietic stem cells. There is no significant difference between HSC1 and HSC2 group.

Conclusions:

1)    An efficient and convenient gene knockout method at the level of a small number of highly purified HSCs was established by optimizing the in vitro culture system of HSCs and simplifying the exogenous plasmid size.

2)    Based on this approach, this study successfully knocked out Pax5 at the level of mouse bone marrow primary HSCs, resulting in the arrest of Pro-B in the bone marrow and the absence of mature B cells in the peripheral blood, consistent with the results of previous studies.