AUTHOR=Pospiech Mateusz , Tamizharasan Mukund , Wei Yu-Chun , Kumar Advaith Maya Sanjeev , Lou Mimi , Milstein Joshua , Alachkar Houda TITLE=Features of the TCR repertoire associate with patients' clinical and molecular characteristics in acute myeloid leukemia JOURNAL=Frontiers in Immunology VOLUME=14 YEAR=2023 URL=https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2023.1236514 DOI=10.3389/fimmu.2023.1236514 ISSN=1664-3224 ABSTRACT=Background

Allogeneic hematopoietic stem cell transplant remains the most effective strategy for patients with high-risk acute myeloid leukemia (AML). Leukemia-specific neoantigens presented by the major histocompatibility complexes (MHCs) are recognized by the T cell receptors (TCR) triggering the graft-versus-leukemia effect. A unique TCR signature is generated by a complex V(D)J rearrangement process to form TCR capable of binding to the peptide-MHC. The generated TCR repertoire undergoes dynamic changes with disease progression and treatment.

Method

Here we applied two different computational tools (TRUST4 and MIXCR) to extract the TCR sequences from RNA-seq data from The Cancer Genome Atlas (TCGA) and examine the association between features of the TCR repertoire in adult patients with AML and their clinical and molecular characteristics.

Results

We found that only ~30% of identified TCR CDR3s were shared by the two computational tools. Yet, patterns of TCR associations with patients’ clinical and molecular characteristics based on data obtained from either tool were similar. The numbers of unique TCR clones were highly correlated with patients’ white blood cell counts, bone marrow blast percentage, and peripheral blood blast percentage. Multivariable regressions of TCRA and TCRB median normalized number of unique clones with mutational status of AML patients using TRUST4 showed significant association of TCRA or TCRB with WT1 mutations, WBC count, %BM blast, and sex (adjusted in TCRB model). We observed a correlation between TCRA/B number of unique clones and the expression of T cells inhibitory signal genes (TIGIT, LAG3, CTLA-4) and foxp3, but not IL2RA, CD69 and TNFRSF9 suggestive of exhausted T cell phenotypes in AML.

Conclusion

Benchmarking of computational tools is needed to increase the accuracy of the identified clones. The utilization of RNA-seq data enables identification of highly abundant TCRs and correlating these clones with patients’ clinical and molecular characteristics. This study further supports the value of high-resolution TCR-Seq analyses to characterize the TCR repertoire in patients.