The Immune Editing Process Reshapes the AML1-ETO Fusion Protein-Induced Pre-Leukemia Microenvironment

Objective: Currently, the process of common chromosome translocation, such as AML1-ETO, recruiting chromatin-modifying enzymes that drive the malignant phenotype is an object of intense investigation. However, there exists an indeterminate understanding of how this chromosome translocation reshapes the immune microenvironment, and whether immune editing occurs during the pre-leukemia stage to facilitate their own progression. In this study, we evaluate how the AML1-ETO chromosome translocation reshapes the immune microenvironment by employing the Aml1 ^Eto/+; Mx1-Cre mouse model to identify potential pivotal immunotherapy targets.

Methods: Four weeks after PIPC induction (5mg/kg), bone marrow-derived T cells without the AML1-ETO (AE) fusion gene and Lin ^- c-Kit ⁺ Sca-1 ⁺ (LSK) cells were sorted from both Aml1 ^Eto/+; Mx1-Cre mice (experimental group) and Aml1 ^Eto/+; w/o Mx1-Cre mice (control group), respectively, and subsequently mixed at 8:2 ratio for single-cell RNA-Seq(scRNA) and single-cell TCR-seq. The results are further validated by real-time PCR and flow cytometry.

Results:

In our preliminary cell-sorting-based dissection research, we found that despite antigen presentation existing in clonally expanded AE precursor cells and the differentiated T cells like effector T cells expanded, the established AE-expressed precursor cells are not eradicated by the host immune system as expected.

To describe the mechanism underlying it, the first step is to find whether the potentially tumor-reactive T cells (pTRTs) existed. We examined the single-cell data based on the following four dimensions: the tumor-enrichment index, the proliferation index, the clonal expansion index, and the TCR signaling pathway activation index. Two clusters of Cd8 ⁺ exhausted T cells and four clusters of Cd4 ⁺Runx3 ⁺ cytotoxic T cells were identified as pTRTs. While Cd8 ⁺ pTRTs demonstrated high expression levels of genes like Gzmk, Ifng, and Prf1, they also exhibited high expression levels of exhaustion markers such as Pdcd1, Tox, and Ctla4. More importantly, compared with virus-specific T cells recognized by machine learning trained from validated TCR databases, the Cd8 ⁺ pTRTs showed the feature of metabolic insufficiency via the Compass analysis. Cell-cell interaction analysis by CellphoneDB and NicheNet and following flow cytometry confirmation revealed that the highly expressed PVR and NECTIN2 on AE ⁺ precursor cells interacted with pTRTs via inhibitory receptor TIGIT, while they interacted with bystander T cells through stimulatory receptor CD226.

At the same time, Tregs played a vital role in the maintenance of the immune equilibrium state as the Treg cells increased in the AE microenvironment but not homogeneously. There were four different Treg clusters identified, including Cd4_Treg_Ccr7+, Cd4_Treg_Tnfrsf9-, Cd4_Treg_ Cdc25b+ with high apoptosis and proliferation score, and Cd4_Treg_Tnfrsf9+ with high suppression\exhaustion/activation/migration score which was recognized as effector Treg. It was observed that only the effector Treg exhibited an increase in proportion. Monocle analysis showed that Cd4_Treg_Ccr7+ was at the beginning of common origin while the Cd4_Treg_Cdc25b+ and Cd4_Treg_Tnfrsf9+ were at the endpoint of two distinct differentiation pathways and the AE microenvironment shifted to the latter one. The flow cytometry results clearly showed a notable rise in KI67 ^-TNFRSF9 ⁺ effector Tregs, along with a significant decline in KI67 ⁺TNFRSF9 ^- proliferative Tregs.

Conclusions: Our findings demonstrate that, during the pre-leukemia stage, the immune editing process has already occurred, leading to substantial alterations in the immune microenvironment. It was identified that PVR/NECTIN2-TIGIT interaction and TNFRSF9 ⁺Treg as targetable mechanisms to restore the function of pTRTs and break the immune equilibrium dilemma.