Single-Cell Analysis Reveals Altered Tumor Microenvironments of Relapse/ Refractory and Remission-Associated Acute Myeloid Leukemia Treated By ABC-14 Regimen

Xin Huang, Minming Li, Ping Wu, Chengxin Deng, Zesheng Lu, Yulian Wang, Suxia Geng, Chengwei Luo, Peilong Lai, Xin Du, Jianyu WengBlood 2024 · DOI

Abstract

The ABC-14 regimen (Azacitidine, Venetoclax, and Chidamide) has demonstrated clinical efficacy in newly diagnosed Acute Myeloid Leukemia (AML), particularly in overcoming resistance seen with the Azacitidine-Venetoclax (VA) combination and reducing the toxicity of traditional "3+7" chemotherapy. This study utilizes single-cell RNA sequencing (scRNA-seq) to characterize the bone marrow (BM) microenvironment in AML patients across various clinical states: diagnosis, remission, relapse, and refractory disease. Our analysis identifies distinct immune cell signatures—including exhausted TIGIT+ CD8+ T cells and specific monocyte clusters—that correlate with clinical outcomes. Furthermore, we observe significant downregulation of p53 signaling in refractory blasts and MHC Class II-related genes in relapse-associated T cells. These findings provide a high-resolution map of the AML microenvironment under ABC-14 therapy, offering insights for the development of targeted therapeutic strategies.

Introduction

Acute Myeloid Leukemia (AML) treatment has traditionally relied on intensive "3+7" induction chemotherapy. While effective, its high toxicity and the emergence of drug resistance necessitate more refined therapeutic approaches. The combination of Azacitidine and Venetoclax (VA) represented a significant advancement, yet resistance remains a challenge.

The "ABC-14" regimen incorporates Chidamide, a histone deacetylase inhibitor, to enhance the efficacy of the VA backbone. Preliminary data (2023 ASH Poster #1522) suggests ABC-14 is comparable to "3+7" in efficacy while offering a superior safety profile and better outcomes for specific subtypes like M4/M5. To optimize this regimen, it is critical to move beyond the malignant "seeds" (blasts) and understand the cellular and molecular dynamics of the bone marrow "soil" (microenvironment) that drive either continuous complete remission (CCR) or treatment failure.

Methodology

We performed scRNA-seq on bone marrow aspirates from 15 AML patients (6 at diagnosis, 3 in CCR, 2 at short-term relapse, and 4 refractory cases) and healthy individuals. Samples were obtained from the Guangdong Provincial People's Hospital under ethical guidelines with informed consent. The bioinformatics workflow included:

  • Cell Clustering and Annotation: Identification of blasts, monocytes, T cells, and macrophages.
  • CNV Estimation: Distinguishing malignant blasts from healthy hematopoietic cells via Copy Number Variation analysis.
  • Differential Expression Analysis (DEGs): Identifying genes uniquely expressed in CCR versus relapse/refractory groups.
  • Pathway Enrichment: Utilizing GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) databases to map altered biological processes.

Results

Immune Signatures of Remission and Relapse

Our analysis revealed that the cellular composition of the bone marrow microenvironment differs significantly based on patient response:

  • Remission (CCR) Group: Characterized by an enrichment of specific CD14+ monocyte clusters expressing IFIT1, S100A12, HLA-DPB1, and AQP9. These markers suggest a more robust and functional innate immune presence associated with successful treatment.
  • Relapse Group: Associated with an increase in CD16+ monocytes, C1QB+ macrophages, and exhausted TIGIT+ CD8+ T cells. The presence of TIGIT, a known immune checkpoint, indicates a state of T-cell exhaustion that may facilitate leukemic escape and disease recurrence.

Molecular Drivers of Resistance

We identified key molecular pathways that are downregulated in patients who fail to respond to ABC-14:

  • p53 Signaling: Primary refractory patients and those with post-therapy residual blasts showed a significant downregulation of genes involved in p53-mediated signal transduction. As p53 is essential for inducing apoptosis in response to DNA damage, its suppression likely contributes to primary drug resistance.
  • Antigen Presentation: In the relapse group, a specific T-cell cluster exhibited downregulation of MHC Class II genes and T-cell activation pathways. This suggests a mechanism of immune evasion where the leukemia cells effectively hide from the adaptive immune system by preventing the presentation of peptide antigens.

Discussion

This study provides a comprehensive single-cell landscape of AML under the ABC-14 regimen. The identification of exhausted T cells and immunosuppressive monocyte/macrophage populations in relapsed patients suggests that combining ABC-14 with immune checkpoint inhibitors (e.g., anti-TIGIT) could potentially improve clinical outcomes. Additionally, the downregulation of p53 and MHC II pathways highlights specific molecular vulnerabilities that could be targeted to overcome primary resistance and prevent relapse.

While limited by a small cohort of 15 cases, these results lay the groundwork for larger validation studies. Future work will focus on determining whether these cellular signatures are causative of relapse or secondary to it, ultimately facilitating the development of personalized therapeutic strategies for relapsed/refractory AML patients.

Summary updated

Abstract

The preliminary study (2023 ASH Poster #1522) demonstrated that ABC-14 regimen (Azacitidine, Venetoclax, and Chidamide) was comparable to “3+7” regimen in the induction therapy of newly diagnosed AML, while overcoming the limitations of drug resistance of VA regimen (Azacitidine and Venetoclax), even with better for the subtype of AML-M4/M5, and reducing the toxicity of the traditional “3+7” regimen. However, it is important to identify deeper characterization of the disease that can be used to better guide ABC-14 regimen therapy decisions and improve clinical outcomes. Here, we utilize single-cell RNA sequencing (scRNA-seq) to analyze AML bone marrow (BM) samples from diagnosis (6 cases), remission (3 cases), short time relapse (2 cases), and refractory (4 cases) timepoints, which provide insights into AML tumor and microenvironment. Bone marrow samples aspirates from AML patients and healthy individuals were obtained from the Guangdong Provincial People's Hospital under ethical guidelines with informed consent. scRNA-seq data analysis including cell clustering and cell-type annotation, CNV estimation in single cells, differentially expressed genes (DEGs) testing, GO/KEGG enrichment analysis. We focused on the distinct clusters of primary refractories, short time relapse, and continuous complete remission (CCR) associated AML blasts, monocytes, T cells, and macrophages with differential expression of genes. Relapse-associated group have more CD16+Mono, exhausted TIGIT+CD8+T cells and C1QB+macrophages, while CCR-associated group have more CD14+IFIT1+Mono, CD14+S100A12+Mono, CD14+HLA-DPB1+Mono, CD14+AQP9+Mono. Post-therapy residual blasts and primary refractory group exhibits downregulation of signal transduction by p53 class mediator genes. Also, a post-therapy T-cell cluster associated with relapse group exhibits downregulation of antigen processing and presentation of peptide antigen via MHC Class II and T-cell activation genes. Altogether, this study characterizes ABC-14 regiment treated AML relapsed /refractory and CCR associated groups to provide insights into the BM microenvironment landscape. Our results will facilitate a better understanding and the development of novel therapeutic strategies for relapsed/refractory AML patients.
BibTeX
@inproceedings{huang2024single,
  author = {Xin Huang and Minming Li and Ping Wu and Chengxin Deng and Zesheng Lu and Yulian Wang and Suxia Geng and Chengwei Luo and Peilong Lai and Xin Du and Jianyu Weng},
  title = {{Single-Cell Analysis Reveals Altered Tumor Microenvironments of Relapse/ Refractory and Remission-Associated Acute Myeloid Leukemia Treated By ABC-14 Regimen}},
  booktitle = {Blood},
  year = {2024},
  doi = {10.1182/blood-2024-204516},
}
Figure