NIST CD19 CAR-T VCN cell lines

Chimeric antigen receptor (CAR) T cell therapy is a revolutionary novel cancer therapy and has shown promise in both leukemia and solid tumor. Lentivirus vector (LVV) transduction is one of the major methods to generate engineered CAR T cells. The LVV integration is uncontrolled. High copies of LVV integration may cause insertional mutagenesis, and/or oncogenicity. The US FDA suggested that integrated vector copy number (VCN) shall be less than 5 copies per cell. Frequently used measurement methods include qPCR, ddPCR, and next generation sequencing (NGS). These methods have their limitations and a WHO inter-laboratory study highlighted the challenges of VCN measurement especially at high VCN. LVV integration caused secondary cancers suggest the importance of determination of LVV integration sites. The need and benefit of reference standards in the field for assay validation and benchmarking is increasingly recognized.In this project, we developed a set of clonal cell lines with defined 0-5 copies of integrated CD19 CAR LVV. CD19 CAR lentivirus was used to transduce Jurkat cells at MOI 1 or 5. GFP positive cells were selected after selected bleomycin resistant single cells were printed into 96-well plates. The clonal cell lines were further screened by ddPCR to determine the integrated VCN. The majority of the clonal cell lines at this stage had one copy of insertion (VCN=1). Therefore, we used one of the clonal cell lines to generate higher VCN cell lines by transducing VCN1 cells with CD19 CAR lentivirus. The transduced single cells were printed into each well of 96-well plates, and the VCN of each clonal cell line was determined by ddPCR. We generated VCN2, 3, and 4 clonal cell lines. VCN4 clonal cell line to generate VCN5 cell line with the same method.The integrated VCN was analyzed by ddPCR and NGS assays. NGS data was analyzed at NIST. Six LVV components were analyzed to determine the VCNs: LTR, Psi, RRE, CD19 CAR, eGFP and WPRE. VCN0, the parental Jurkat cell line, was not detected with any targets. VCN1 and VCN2 clonal cell lines contained one and two copies of each element per cell respectively. Three copies of most elements, and an extra copy of WPRE were detected in VCN3 cells. VCN4 cells had four copies of LTR, Psi, RRE, CD19 CAR, and WPRE, and five copies of eGFP. Since VCN5 cells were derived from VCN4, we detected five copies of LTR, Psi, RRE, CD19 CAR, and WPRE, and six copies of eGFP.NGS analysis revealed that the proviral DNA integrated in the Jurkat genome shows a variety of SNPs, deletions, repeats, and other unintended insertions. VCN1 showed one copy of proviral DNA inserted on chromosome (chr) 17 (78,730,345, minus strand). Further, there was a 724 bp deletion in the EF1a promoter region. Since VCN2-5 were all derived from VCN1, all cell lines contained the 724 bp deletion on chr 17. VCN2 cells had an additional insertion on chr19 (21,610,872, minus strand) containing 4 SNPs. VCN3 had an additional insertion on chrX (103,930,456, minus strand), one copy inherited from VCN1 cells and another copy inserted on chr1 (23,307,863, minus strand) with a 789 bp repeat (part of the WPRE sequence) at the end of 3 LTR sequence, and 1 SNP. NGS also revealed that there was a 260 bp random insertion on chr8 (99,830,407, minus strand), composed mostly from LTR sequence. VCN4 contained an additional insertion with expected sequence on chr6 (24,825,065, plus strand) and an insertion on chr3 (18,115,277, minus strand) with a single SNP. An additional insertion was found in chr1 (243,273,502 plus strand) with one SNP, however, at the end of 3 LTR sequence there was a 41 bp sequence of unknown origin. Additionally, a 2046 bp 3 portion of the proviral DNA, containing the eGFP gene, was found inserted in chr6 (15,601,063, minus strand). The VCN5 cells were derived from VCN4 cells, therefore VCN5 cells inherited insertions found in VCN4. Additionally, VCN5 had an as expected insertion in chr5 (39,044,694, minus strand).

嵌合抗原受体（Chimeric antigen receptor, CAR）T细胞疗法是一种革命性的新型癌症治疗手段，在白血病与实体瘤治疗中均展现出应用潜力。慢病毒载体（Lentivirus vector, LVV）转导是制备工程化CAR-T细胞的主要方法之一，但LVV的整合过程不受调控，高拷贝数的LVV整合可能引发插入突变或致癌性。美国食品药品监督管理局（US FDA）建议，整合载体拷贝数（vector copy number, VCN）应低于每细胞5拷贝。当前常用的检测方法包括定量聚合酶链反应（qPCR）、液滴数字聚合酶链反应（ddPCR）与下一代测序（next generation sequencing, NGS），但这些方法均存在局限性，世界卫生组织（WHO）的室间比对研究也强调了VCN检测的挑战，尤其是在高VCN场景下。LVV整合引发的继发性癌症凸显了LVV整合位点检测的重要性，该领域用于检测方法验证与基准比对的参考标准的需求与价值正日益获得认可。本研究构建了一系列整合CD19 CAR LVV拷贝数明确为0~5的克隆细胞系：以感染复数（Multiplicity of Infection, MOI）1或5的CD19 CAR慢病毒转导Jurkat细胞，经博来霉素抗性筛选后，将单克隆细胞接种至96孔板，随后分选绿色荧光蛋白（GFP）阳性细胞；通过ddPCR对克隆细胞系进行进一步筛选以确定其整合VCN，此阶段多数克隆细胞系仅携带1个拷贝的整合位点（VCN=1）。因此，研究选取其中一株VCN=1的细胞系，通过再次转导CD19 CAR慢病毒以构建更高VCN的细胞系：将转导后的单细胞接种至96孔板的每一个孔中，通过ddPCR测定每株克隆细胞系的VCN，最终获得VCN为2、3、4的克隆细胞系，并以相同方法由VCN4细胞系构建VCN5细胞系。通过ddPCR与NGS对整合VCN进行分析，NGS数据由美国国家标准与技术研究院（NIST）完成分析，针对6种LVV组分进行VCN检测：长末端重复序列（long terminal repeat, LTR）、包装信号（packaging signal, Psi）、Rev应答元件（Rev response element, RRE）、CD19 CAR、eGFP与土拨鼠肝炎病毒转录后调控元件（Woodchuck hepatitis virus posttranscriptional regulatory element, WPRE）。VCN0即亲本Jurkat细胞系，未检测到任何靶标序列；VCN1与VCN2细胞系分别携带每个靶标元件1、2个拷贝；VCN3细胞系中多数靶标元件为3个拷贝，且额外携带1个WPRE拷贝；VCN4细胞系携带LTR、Psi、RRE、CD19 CAR与WPRE各4个拷贝，eGFP为5个拷贝；由于VCN5细胞系由VCN4细胞系衍生而来，其携带LTR、Psi、RRE、CD19 CAR与WPRE各5个拷贝，eGFP为6个拷贝。NGS分析显示，整合至Jurkat基因组的前病毒DNA存在多种单核苷酸多态性（SNPs）、缺失、重复及其他非预期插入事件：VCN1细胞系的前病毒DNA整合至17号染色体（负链，78,730,345位点），且EF1α启动子区域存在724 bp的缺失；由于VCN2~5细胞系均由VCN1细胞系衍生而来，所有细胞系均携带17号染色体上的724 bp缺失。VCN2细胞系额外在19号染色体（负链，21,610,872位点）存在一处插入，包含4个SNPs；VCN3细胞系额外存在两处插入：一处为X染色体（负链，103,930,456位点）的插入（拷贝数继承自VCN1细胞系），另一处为1号染色体（负链，23,307,863位点）的插入，其3'LTR序列末端存在789 bp的重复（部分WPRE序列），且带有1个SNP。NGS还显示，8号染色体（负链，99,830,407位点）存在一处260 bp的随机插入，序列主要由LTR组成；VCN4细胞系额外存在三处插入：一处为6号染色体（正链，24,825,065位点）的预期序列插入，一处为3号染色体（负链，18,115,277位点）的插入（带有1个SNP），还有一处为1号染色体（正链，243,273,502位点）的插入（带有1个SNP），但其3'LTR序列末端存在一段41 bp的未知来源序列；此外，一段包含eGFP基因的2046 bp前病毒DNA 3'端片段被插入至6号染色体（负链，15,601,063位点）。VCN5细胞系由VCN4细胞系衍生而来，因此继承了VCN4的所有插入事件，此外还在5号染色体（负链，39,044,694位点）存在一处预期的插入事件。