Miao Li

Projects Research + code

CAR-T vs. CAR-NKT in solid tumors

Following two cell therapies through tissues and time to learn why one of them does better in solid tumors — and which checkpoint each one needs.

When
2022 – 2026
My role
Co-first author; led the single-cell and transcriptomic analyses
Built with
scRNA-seq, scTCR-seq, Seurat, Scanpy, R, Python
Hand-drawn graphic: single-cell and TCR profiling of 274,878 cells and over 3,200 receptor-ligand pairs follows PBMC-derived CAR-T and stem cell-derived CAR-NKT cells into MSLN-targeted solid tumors. CAR-T cells are held back by TIGIT-CD112 and pair with anti-TIGIT; CAR-NKT cells home better, persist longer, are held back by CD96-CD155, and pair with anti-CD96. Each cell therapy gets its own checkpoint partner.
Cells to checkpoints: single-cell discovery on the left, the matched checkpoint partner for each cell product on the right.

Project introduction

An introduction to the CAR-T vs. CAR-NKT spatiotemporal profiling study.

The question

CAR-T cells transform some blood cancers but struggle in solid tumors, where they infiltrate poorly and meet a hostile microenvironment. Invariant NKT cells home to tissues naturally and can reshape that environment, which makes allogeneic, stem cell-derived CAR-NKT cells an attractive alternative. We wanted to know how the two actually behave inside the body.

What I did

I led the single-cell RNA and TCR sequencing analyses, profiling conventional CAR-T cells and IL-15-enhanced CAR-NKT cells across multiple tissues and time points in solid-tumor models: where each product goes, how long it stays, and which gene programs and checkpoint receptors it switches on inside tumors.

What we found

The two products had distinct pharmacokinetic, pharmacodynamic, and immunoregulatory profiles. CAR-NKT cells homed to, infiltrated, and localized within tumors better and persisted longer, with their own checkpoint-receptor landscape. That landscape pointed to different combinations: CAR-T cells worked synergistically with TIGIT blockade, while CAR-NKT cells were more sensitive to CD96 blockade.

Why it matters

Cell therapies are not interchangeable, and neither are their combination partners. Profiling them in space and time gives a rational way to pair each product with the checkpoint therapy most likely to help it.

Read the paperAll projects