AUTHORS
zhixiang yan, Yanan Yue2, Heng Mei, MD, PhD3, yiwen zhou4, xiaobin song5, xiaoling wang6, Jia Xu7, lihong shen8, Lin Liu9, jie zhou10, Yichi Zhang11, tianying chen12, Yingyong Xu, MD, PhD, MBA13, Luhan Yang, PhD14
AFFILIATIONS
1 Qihan Biotechnology, Hangzhou, China
2 Qihan Biotechnology, Hangzhou, China
3 Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, wuhan, China
4 Qihan Biotechnology, Hangzhou, China
5 Qihan Biotechnology, Hangzhou, China
6 Qihan Biotechnology, Hangzhou, China
7 Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, wuhan, China
8 Qihan Biotechnology, Hangzhou, China
9 Institute of Hematology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, wuhan, China
10 Qihan Biotechnology, Hangzhou, China
11 Qihan Biotechnology, Hangzhou, China
12 Qihan Biotechnology, Hangzhou, China
13 Qihan Biotechnology, Hangzhou, China
14 Qihan Biotechnology, Hangzhou, China
ABSTRACT
Background In vivo CAR-T therapy generates autologous CAR-T cells directly in patients and may overcome limitations of ex vivo manufacturing and access. Key open questions are whether in vivo–generated CAR-T cells achieve deep pharmacodynamic target clearance, whether lentiviral transduction perturbs T-cell differentiation or drives exhaustion, and how many T cells are actually transduced in vivo. QI-019A is a T cell–targeted lentiviral vector encoding a dual CD19/BCMA CAR for relapsed/refractory multiple myeloma (RRMM). We applied longitudinal single-cell RNA sequencing (scRNA-seq) with paired single-cell TCR sequencing (scTCR-seq) to address these questions in a patient treated with a single intravenous dose of QI-019A.
Methods Peripheral blood (PB) was collected at baseline (BL), peak CAR-T expansion (Cmax), and Day 28; bone marrow (BM) was collected at BL and Day 14.
At Cmax, PBMCs were sorted into CAR-positive and CAR-negative fractions by FACS, then profiled by scRNA-seq, scTCR-seq, and CAR-transcript detection. CAR status was determined using concordant FACS and CAR-transcript measurements: only cells with concordant FACS and CAR-transcript calls were retained as high-confidence CAR-positive or CAR-negative cells; discordant cells were excluded. High-confidence populations were compared for differentiation state, activation, cytotoxicity, proliferation, metabolic fitness, and exhaustion programs.
Productive paired TCRαβ sequences defined clonotypes; a Heaps-law model fitted to rarefaction data, together with the estimated absolute CAR-positive T-cell count, was used to extrapolate total clonotype richness and derive a lower-bound estimate of independent in vivo founder transduction events.
Results:
Deep clearance of malignant plasma cells.QI-019A induced a marked reduction of malignant plasma cells in BM, accompanied by depletion of circulating CD19-positive B cells and BCMA-positive plasma cells. These changes tracked CAR-T expansion and supported on-target activity of dual CD19/BCMA targeting.
Effector activation with preserved functional competence. High-confidence CAR-positive cells comprised both CD4 and CD8 T cells and were enriched for effector-memory (Tem) and effector (Teff) states, with smaller progenitor-exhausted-like, terminal-exhaustion-like, and TEMRA-like subsets. Compared with CAR-negative T cells at Cmax, CAR-positive cells showed distinct subset composition and enhanced activation, cytotoxicity, and proliferation, while metabolic programs showed selective remodeling without uniform enhancement. Although exhaustion-associated features were increased in subsets, terminally exhausted cells remained a minority and coexisted with robust effector and proliferative activity. Overall, the CAR-positive population demonstrated strong antigen-driven activation with preserved functional competence rather than widespread terminal exhaustion.
Polyclonal origin and estimated transduction scale. Despite expansion of several clonotypes, the CAR-positive compartment retained substantial TCR diversity, comprising 1,040 clonotypes among 4,236 cells. Extrapolation to an estimated 4 × 108 circulating CAR-positive T cells at Cmax suggested TCR clonotype richness on the order of 106. Because each distinct CAR-positive clonotype requires at least one independently transduced founder T cell, this provides a model-based lower bound of approximately 106 founder transduction events. The actual number may be higher because multiple cells with the same TCR clonotype may have been independently transduced.
Conclusions:
In one patient with RRMM, QI-019A generated dual CD19/BCMA CAR-T cells in vivo and produced marked depletion of malignant plasma cells and targeted B-lineage compartments. Paired scRNA-seq and scTCR-seq showed memory-enriched states, broad TCR diversity, and a model-based lower bound of approximately 106 founder transduction events. These findings support the pharmacodynamic activity, cellular fitness, and preliminary clonal safety of QI-019A.
DISCLOSURES
Luhan Yang, PhD: No relevant conflicts of interest to declare.