Search bioRxiv⌕ Search

bioRxiv · 10.1101/2023.02.22.529592

Machine Learning Investigation Of Gene Expression Datasets Reveals TP53 Mutant-like AML With Wild Type TP53 And Poor Prognosis

Abstract

Acute myeloid leukemia (AML) with TP53 mutations (TP53Mut) has poor clinical outcomes with 1-year survival rates of less than 10%. We investigated whether this AML subtype harbors a distinct gene expression profiling (GEP), what this GEP reveals about TP53Mut AML pathophysiology, and whether this GEP is prognostic in TP53 wild type (TP53WT) AML. We applied a supervised machine-learning approach to assess whether a unique TP53Mut GEP could be detected. Using the BEAT-AML dataset, we randomly divided the samples into training and testing datasets, while the TCGA dataset was reserved as a validation dataset. We trained a ridge regression machine learning model to classify TP53Mut and TP53WT cases. This model was highly accurate in distinguishing TP53Mut versus TP53WT cases in both the test and validation data sets. Additionally, we noted a cohort of TP53WT samples with high ridge regression scores and poor overall survival, suggesting share clinical and GEP features with TP53Mut AML. We defined these TP53WT samples as TP53 mutant-like (TP53Mut-like) AMLs. We trained a second ridge regression model to specifically detect TP53Mut-like samples in the BEAT AML dataset and found that TCGA data also harbors TP53Mut-like samples. The TP53Mut-like samples in the TCGA also have a worse OS rate than TP53WT cases. Using drug sensitivity data from 122 small molecules in the BEAT AML dataset, we found TP53Mut-like AMLs have distinct drug sensitivity patterns compared to TP53WT. Finally, we identified a 25 gene signature that can identify TP53Mut-like cases. This signature could be used clinically to identify this novel subset of poor-prognosis AML.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Lee, Y., Baughn, L. B., Sachs, Z., Myers, C. L.. 2023-02-23. Machine Learning Investigation Of Gene Expression Datasets Reveals TP53 Mutant-like AML With Wild Type TP53 And Poor Prognosis. https://doi.org/10.1101/2023.02.22.529592

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

Minor intron splicing restricts E2F transcription factor activity in RB-deregulated prostate cancer

Castration-resistant prostate cancer with neuroendocrine differentiation (CRPC-NE) arises in RB1-mutated lineage plastic cells in response to androgen receptor (AR)-targeted therapies. RB1 loss leads to constitutive E2F activation and cylin-dependent kinase (CDK) inhibitor resistance, making it challenging to restrict E2F activity. A post-transcriptional node is that typical E2Fs contain a rare but highly conserved minor intron, processed by the minor spliceosome (MiS). In an in vivo CRISPR screen, we nominate the MiS as the strongest dependency, and U6atac correlates strongest with E2F transcriptional programs in a clinical CRPC cohort. U6atac depletion led to mis-splicing of typical E2Fs, cell cycle arrest, and translation of a minor intron-containing E2F1 isoform (E2F1-MIRI) as a pharmacodynamic biomarker. Combined E2F activator depletion, induced further neuroendocrine (NE) specification. U6atac depletion shifted NE identity along the ASCL1 axis without the secretory granule program, and induced cell-surface antigens including DLL3, nominating MiS inhibition as a combination partner for antigen-directed agents.

cancer biology↗

An expanded, phenotypically distinct tumor cell subset with low HPV mRNA increases recurrence risk

Human papillomavirus-related (HPV+) cancers harbor a subset of tumor cells expressing certain senescence-related transcripts and reduced levels of HPV mRNAs. We aimed to define this HPV-low population's biologic traits and clinical significance via integrated analysis of recurrence-prone human HPV+ oropharyngeal cancers (OPCs) and patient-derived xenografts (PDXs) of variable therapy sensitivity. Single-cell mRNA sequencing of PDXs and human tumors detected gene expression profiles in HPV-low cells indicating lack of senescence or terminal differentiation and intact tumorigenic capacity. HPV-low cells in cisplatin-refractory PDXs and particular human OPCs differentially upregulated multiple pathways linked to aggressive tumor biology, including RAS signaling, IL-6 signaling, and EMT. Predicted loss of cell-cell interactions among HPV-low cells reflected reduced epithelial identity, further supporting their more aggressive biology. Accordingly, the size of the pre-treatment HPV-low fraction had strong, independent association with recurrence risk. Genetic subclones within OPCs varied widely in content of HPV-low cells, which were enriched in particular lineages. Our findings uncover novel pro-tumor functions of HPV-low cells and support a central role for them in therapy resistance as well as potential utility as a clinical biomarker. Further elucidating mechanisms of cooperativity and plasticity between HPV-low and high states holds promise to reveal targets in treatment-refractory HPV+ malignancies.

cancer biology↗

Discovery of molecular glues for modulating MHC presentation of cancer neoantigens

Antigen presentation by MHC class I molecules is essential for cancer immunosurveillance, but many cancer-specific antigens fail to form complexes with Class I MHC and are poorly presented on the cell surface. In this study, we present a chemical strategy to overcome this limitation by identifying peptide MHC glues (PMGs) that promote the formation of peptide-MHC complexes with otherwise unpresentable cancer antigens. Crystal structures of these complexes show that our compounds stabilize antigen-MHC binding by making simultaneous interactions with the MHC protein and the antigen peptides. These findings reveal a new approach for chemically rewiring Class I MHC antigen presentation and accessing rare cancer antigens for therapeutic targeting.

cancer biology↗