Abstract
Despite recent advances in the treatment of acute myeloid leukemia (AML), most patients with AML relapse and succumb to the disease. Achieving complete remission (CR) does not equate to cure, especially in TP53 mutated (TP53mut) AML. Relapse after CR is caused by measurable/minimal residual disease (MRD), which is resistant to prior therapies due to cell-intrinsic resistance mechanisms or the protective bone marrow microenvironment (BME). TP53mut AMLs reside in an immune dysregulated BM niche, however, a detailed characterization of the MRD-supportive BME has not been undertaken. In this study, we used spatial proteomics analysis of paired AML BM biopsy specimens to dissect spatial characteristics of AML and the BME using COMET (Lunaphore) and Visiopharm (Visiopharm) analytics.
Methods: To investigate the spatial architecture of TP53mut and TP53 wild-type (TP53wt) AMLs, BM biopsy specimens from TP53mut AML, TP53wt AML patients and healthy donors (N = 16, 10 and 3, respectively) were subjected to sequential multiplex immunofluorescence staining with a 20-antibody panel. Seven AML biopsies were paired at diagnosis and remission. All TP53wt samples carried NPM1 mutations (NPM1c). The panel includes markers for immature blasts, hematopoietic cells, and structural elements including mesenchymal stromal cells (MSCs), endothelial cells, adipocytes, bones, and arterioles. Strong p53 positivity by immunohistochemistry indicates TP53 mutations (Blood 2022), and anti-p53 and anti-NPM1c antibodies were used to detect TP53mut and NPM1mut/TP53wt leukemia cells. Patients received either Aza/Ven- or Ara-C/Idarubicin-based regimens combined with magrolimab or gilteritinib.
Results: In total, 745,285 cells were analyzed across 29 whole BM images after QC steps, and 605,208 cells were successfully annotated, on average 25,699 cells (range 1,585 - 86,639) per sample. In TP53 mutant biopsies, we found 43.7% (16.0 - 59.2%) p53 positive (p53+) cells at diagnosis and 6.14% (0.167 – 17.7) in remission, respectively, comparable to flowcytometry-based MRD detection (FC-MRD) of 2.1% (0 – 3.7). In TP53wt samples, NPM1c positive cells at diagnosis and in remission accounted for 77.2% (62.0 - 93.6) and 0.031% (0 – 0.63), respectively. FC-MRD was 0% in these remission samples.
p53+ leukemia cells were subclassified into p53+, p53+/CD34+ immature, p53+/CD71+ erythroid, and p53+/CD14+ monocytic cells. We noticed that megakaryocytes (MgK) express high levels of p53, suggesting TP53 mutations in MgK. p53+ erythroid cells were prevalent at diagnosis (33%) and further enriched in remission (60%), suggesting persistence of p53+ erythroid MRD cells. Spatial neighborhood clustering revealed that lymphocytes, erythroblasts, and mature myeloid cells each form distinct clusters across all samples, indicating spatial segregation by cell type. Notably, p53+ leukemia cells are not uniformly distributed at diagnosis and in remission. p53+ immature cells form clusters with themselves or with p53+ erythroid cells and are in close proximity to MSCs and adipocytes. p53+ erythroid cells co-localize with erythroblasts and are also close to MSCs. p53+ monocytic cells are structurally associated with T-cell clusters. These findings suggest a heterogeneous niche specific for each of the p53+ subtypes. Lymphocyte clusters in TP53mut AML BMs encompass regulatory T-cells(Tregs, CD3+/FOXP3+) but not B-cells. In addition, p53+ erythroid and immature cells are devoid of T-cell clusters. In contrast, TP53wt leukemia cells in remission are colocalized with lymphocyte clusters lacking Tregs. Strikingly, TIGIT surge, a distance-dependent increase of TIGIT levels, was observed in T and NK cells near p53+ cells. These findings collectively support the presence of a localized immunosuppressive microenvironment in TP53mut AML.Conclusions: single-cell proteomics analysis of bone marrow biopsies provides a spatially resolved view of the TP53mut AML microenvironment, uncovering persistent p53+ MRD cells enriched for erythroid phenotypes. Their structural niches suggest MRD-protective roles of MSCs and adipocytes, and immunosuppressive features such as spatial separation between T-cell clusters and p53+ cells, involvement of Tregs within lymphocyte clusters and a distance-dependent TIGIT surge of T and NK-cells in proximity to p53+ cells. This comprehensive spatial analysis offers new insights into the mechanisms of immune escape and MRD persistence in TP53mut AML.
