A carcinoembryonic antigen-specific cell therapy selectively targets tumor cells with HLA loss of heterozygosity in vitro and in vivo.
Mark L SandbergXueyin WangAaron D MartinDaniel P NampeGrant B GabrelowChuck Z LiMichele E McElvainWen-Hua LeeSanam ShafaattalabSara MartireFernando A FisherYuta AndoEdwin LiuDavid JuLu Min WongHan XuAlexander KambPublished in: Science translational medicine (2022)
The CEACAM5 gene product [carcinoembryonic antigen (CEA)] is an attractive target for colorectal cancer because of its high expression in virtually all colorectal tumors and limited expression in most healthy adult tissues. However, highly active CEA-directed investigational therapeutics have been reported to be toxic, causing severe colitis because CEA is expressed on normal gut epithelial cells. Here, we developed a strategy to address this toxicity problem: the Tmod dual-signal integrator. CEA Tmod cells use two receptors: a chimeric antigen receptor (CAR) activated by CEA and a leukocyte Ig-like receptor 1 (LIR-1)-based inhibitory receptor triggered by human leukocyte antigen (HLA)-A*02. CEA Tmod cells exploit instances of HLA heterozygous gene loss in tumors to protect the patient from on-target, off-tumor toxicity. CEA Tmod cells potently killed CEA-expressing tumor cells in vitro and in vivo. But in contrast to a traditional CEA-specific T cell receptor transgenic T cell, Tmod cells were highly selective for tumor cells even when mixed with HLA-A*02-expressing cells. These data support further development of the CEA Tmod construct as a therapeutic candidate for colorectal cancer.
Keyphrases
- induced apoptosis
- cell cycle arrest
- cell therapy
- oxidative stress
- endoplasmic reticulum stress
- gene expression
- clinical trial
- magnetic resonance
- magnetic resonance imaging
- binding protein
- signaling pathway
- early onset
- randomized controlled trial
- machine learning
- big data
- pi k akt
- transcription factor
- electronic health record
- peripheral blood
- copy number
- artificial intelligence