Nucleic Acids Research: Questioning the safety of in vivo gene editing

In recent years, a new generation of gene editing tools such as CRISPR-Cas has shown great application prospects
in the process of tumor immunotherapy and gene in vivo therapy.
The optimization of gene editing tools has gone through three periods: first, the improvement and expansion of editing efficiency and application scope; the second is the inhibition of off-target activity; The third is to focus on the genomic toxicity of gene editing, which mainly refers to chromatin structural variation
.
Chromosomal structural variation includes chromosomal translocations and large fragments of DNA loss1
.
Chromosome structural variation seriously threatens the stability of the genome and interferes with the normal life activities of cells, thereby promoting cell death, malignant proliferation and cancer, such as a variety of lymphomas and leukemias caused by
chromosomal translocation.
In 2021, the NIH Somatic Cell Editing Project wrote an article in Nature emphasizing the importance and necessity of paying attention to chromosomal structural variation caused by gene editing2
.
In the same year, due to the discovery of chromosomal translocations caused by gene editing in patients infused with universal CAR-T cells, the US FDA urgently suspended clinical research of CAR-T cell therapy developed by Allogene Therapeutics3
.

Previous studies by Hu Jiazhi's group found that in general-purpose CAR-T cells for tumor treatment, Cas9 editing multiple sites at the same time will result in chromosomal translocations of up to about 2% between target sites and between target sites and other locations in the genome, and can be continuously detected during in vitro culture for 2 weeks4
.
However, the fate of chromosomal structural abnormalities of T cells caused by gene editing after infusion into the body is still blank, which hinders the large-scale application of gene editing in clinical practice and becomes the sword
of Dalmose that continues to hang over the heads of related fields.

On October 16, 2022, the research group of Hu Jiazhi, a researcher from the School of Life Sciences, Peking University, and Xu Mo's research group from the Beijing Institute of Biological Sciences/Tsinghua University Institute of Biomedical Sciences published a paper entitled "CRISPR/Cas9-induced structural variations expand in Lymphocytes in vivo" online at Nucleic Acids Research.
The research paper comprehensively, quantitatively and long-term tracked and analyzed the dynamic changes
of T cell chromosomal structural abnormalities caused by gene editing after infusion into mice.
The study found that chromosomal translocations, large-scale chromosome loss, and viral DNA insertion caused by gene editing did not disappear over time within 2 months of adoptive mice, but continued to maintain a high level and showed a clear random clonal amplification phenomenon
.
It is particularly noteworthy that some structural variants can even expand violently to tens of thousands of copies in the body, showing signs of malignant amplification
.

The authors used the mouse enteritis model previously established by Xu Mo's group5 to study the fate
of chromosomal structural mutations caused by gene editing in adoptive T cells in vivo.
In this system, T cells expressing only TCRs that specifically recognize the Helicobacter hepaticus bacterial antigen are first activated in vitro and edited
by Cas9.
Successfully edited cells will be infused into immunodeficient mice that have colonized Helicobacter hepaticus, thereby inducing T-cell-dependent enteritis (Figure 1), an inflammatory process that closely resembles
TCR-T or CAR-T cell-mediated tumor immunotherapy 。 Subsequently, the authors used the highly sensitive sequencing method PEM-seq1,6 (Nucleic Acids Research |) to develop a highly sensitive sequencing method by Hu Jiazhi's research group Hu Jiazhi's research group developed a new method for comprehensive evaluation of gene editing products, the detailed protocol has been published in STAR protocols7 in 2022), which traces and quantifies the proportion and distribution characteristics
of chromosomal structural abnormalities in Cas9-edited T cells before infusion, 3 weeks and 2 months after infusion.

Figure I.
The HH7-2tg mouse model mimics the TCR-T cell editing and infusion treatment process

Taking the most harmful chromosomal translocation as an example, the authors found that the edited T cells carried about 1% of the total editing events before infusion (Figure 2A), which is comparable to the previous results in human CAR-T cells4 and means that at least millions of T cells carrying chromosomal translocations will be infused into the patient
in a single CAR-T cell treatment 。 After 3 weeks and 2 months of infusion into mice, the chromosomal translocation level was maintained at 0.
2~3.
36% (3.
36% for only one mouse, and 0.
79% for the other 6 mice) and 0.
17~0.
59% (4 mice) (Fig.
2A), which corrected the finding made by PCR with limited sensitivity and accuracy in the field in the early stage - structural abnormalities such as chromosomal translocation will eventually disappear
with cell growth 。 In addition, chromosomal translocations in Cas9-edited T cells were relatively evenly distributed on the genome before infusion; The chromosomal translocations in T cells infused into mice for 3 weeks or 2 months were mainly composed of several to dozens of chromosomal translocation events, indicating that these chromosomal translocations occurred clonal expansion with the proliferation of T cells (a total of 128 clonal amplified chromosomal translocations were found in this study) (Figure 2B).

。 Of particular interest, the authors found that the edit-induced chromosomal translocation produced 20,615 clones (72.
2% of all chromosomal translocation events) in one mouse, resulting in a 2.
4-fold increase in chromosomal translocation levels compared to pre-infusion (Figure 2C), suggesting that the chromosomal translocation may have contributed to the cells' growth advantage (Figure 2D).

Figure II.
PEM-seq detects the proportion and high-frequency sites of chromosomal translocation in inflammatory T cells

Similar to chromosomal translocations, the authors found that gene editing resulted in large DNA loss at target sites and viral DNA insertion at similar levels before and after infusion into mice, and exhibited random clonal amplification
.
Similarly, the authors found that a single viral insertion event produced 30,401 clones in a mouse and resulted in a nearly two-fold increase in viral DNA insertion levels compared to pre-infusion (Figure 3).

Figure III.
PEM-seq identified viral DNA integration and the proportion of individual viral DNA sites in inflammatory T cells

In summary, the study found that the by-products of chromosomal translocation, large DNA deletion at target sites and viral DNA insertion, which seriously endanger genome stability caused by gene editing, not only did not disappear within 2 months after being infusion into mice, but also drastically random clonal amplification
of some chromosome structural abnormalities in each infusion mouse (a total of 16).
Moreover, the study detected a higher level of chromosomal translocation in four (25%) of the infusion-infused mice than before the infusion, suggesting that abnormal proliferation of T cells carrying structural chromosomal abnormalities may occur
frequently 。 Therefore, this study not only provides enlightenment for the development of oncogenic chromosomal translocations in vivo in lymphoma or leukemia, but also suggests the necessity of
using methods such as PEM-seq to continuously track the dynamic changes and fate of gene editing products in vivo in adoptive cell therapies (such as general-purpose CAR- or TCR-T, gene-edited HSCs or islet cells, etc.
) or in vivo gene therapy (macular degeneration, etc.
).
。 In addition, the study also sheds light on the need for safer gene-editing tools, such as Cas9TX, which can reduce chromosome structural abnormalities to background levels4 (Nature Communications |).
Hu Jiazhi's research group has developed the most secure Cas9 gene editing tool variant Cas9TX).


Dr.
Liu Yang, researcher and research group of Peking University, and Xu Mo, researcher of Beijing Institute of Biological Sciences/Tsinghua University Institute of Biomedical Sciences, are co-corresponding authors
of the paper.
Wu Jinchun, doctoral students in Hu Jiazhi's laboratory, Zou Ziye, doctoral student in Xu Mo's laboratory, and Dr.
Liu Yang of Hu Jiazhi's research laboratory are the co-first authors of the paper, and Liu Xuhao and Zhang Dingzhengrong, doctoral students in Hu Jiazhi's research laboratory, have also made important contributions
to this work 。 This work has been strongly supported
by the Ministry of Agriculture and Rural Affairs of China, the Peking University-Tsinghua Joint Center for Life Sciences, the National Key R&D Program of the Ministry of Science and Technology, the National Natural Science Foundation of China, the Key Laboratory of Cell Proliferation and Differentiation of the Ministry of Education, the Instrument Center (imaging platform and flow cytometry platform) of the School of Life Sciences, Peking University, the Beijing Institute of Biological Sciences/Tsinghua University Institute of Biomedical Sciences, and the Beijing Municipal Science and Technology Commission.

References

1.
Liu, M.
et al.
Global detection of DNA repair outcomes induced by CRISPR–Cas9.
Nucleic Acids Research 49, 8732-8742, doi:10.
1093/nar/gkab686 (2021).

2.
Saha, K.
et al.
The NIH Somatic Cell Genome Editing program.
Nature 592, 195-204, doi:10.
1038/s41586-021-03191-1 (2021).

3.
Sheridan, C.
Off-the-shelf, gene-edited CAR-T cells forge ahead, despite safety scare.
Nat Biotechnol 40, 5-8, doi:10.
1038/d41587-021-00027-1 (2022).

4.
Yin, J.
et al.
Cas9 exo-endonuclease eliminates chromosomal translocations during genome editing.
Nat Commun 13, 1204, doi:10.
1038/s41467-022-28900-w (2022).

5.
Xu, M.
et al.
c-MAF-dependent regulatory T cells mediate immunological tolerance to a gut pathobiont.
Nature 554, 373-377, doi:10.
1038/nature25500 (2018).

6.
Yin, J.
et al.
Optimizing genome editing strategy by primer-extension-mediated sequencing.
Cell discovery 5, 1-11, doi: 10.
1038/s41421-019-0088-8 (2019).

7.
Liu, Y.
et al.
PEM-seq comprehensively quantifies DNA repair outcomes during gene-editing and DSB repair.
STAR Protocols 3, 101088, doi:https://doi.
org/10.
1016/j.
xpro.
2021.
101088 (2022).