Ebinger, M. | bioRxiv Science

bioRxiv2025

A platform of robust patient-derived leukemia models covering subgroups for which no cell lines exist

Preclinical cancer research requires robust model systems, especially for poor prognosis entities like acute myeloid leukemia (AML), a highly aggressive blood cancer. Here, primary tumor cells from 137 AML patients of all age groups were transplanted into immune compromised mice to generate patient-derived xenografts (PDX). From these, 23 models enable robust, virtually endless serial re-transplantation and are amenable to lentiviral genetic engineering (*PDX AML models). These models primarily originate from patients with highly aggressive, relapsed disease. Comprehensive genomic, transcriptomic, and epigenomic analyses confirmed that they replicate primary AML biology more faithfully than conventional cell lines. Notably, *PDX AML models include AML subgroups that are underrepresented or absent in existing model systems, such as cytogenetically normal or IDH1/2-mutant AML. They withstand freeze-thaw cycles, making them suitable for broad distribution and reproducibility across research institutions. Luciferase-based in vivo imaging enables real-time monitoring of tumor progression and treatment responses in preclinical trials. Surprisingly, long-term treatment, including repeated cytarabine therapy over a period of one year, showed a gradual reduction in leukemia cell proliferation, which decreased continuously after each treatment block. Collectively, our *PDX models represent a robust, versatile, and relevant platform that holds great promise to accelerate translational research for the benefit of cancer patients. Visual Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=123 SRC="FIGDIR/small/677299v1_ufig1.gif" ALT="Figure 1"> View larger version (23K): org.highwire.dtl.DTLVardef@1ead31eorg.highwire.dtl.DTLVardef@1ca702corg.highwire.dtl.DTLVardef@18a7137org.highwire.dtl.DTLVardef@edbe6_HPS_FORMAT_FIGEXP M_FIG C_FIG Key PointsO_LIWe present new robust AML PDX models covering subgroups for which no cell lines exist for use in various ex vivo and in vivo applications. C_LIO_LI*PDX models enable serial transplantation, genetic engineering and better representation of primary AML biology than cell lines. C_LIO_LIOne-year in vivo trials mimicking clinical chemotherapy showed surprising gradual decline in leukemia growth after each treatment block. C_LI

cancer biology↗

bioRxiv2024

Super-enhancer-driven CACNA2D2 is an EWSR1::WT1 signature gene encoding a diagnostic marker for desmoplastic small round cell tumor (DSRCT)

Desmoplastic small round cell tumor (DSRCT) is a highly aggressive cancer predominantly occurring in male adolescents and young adults. The lack of a comprehensive understanding on the biology of the disease is paralleled by its dismal survival rates (5-20%). To overcome this challenge, we first identified and prioritized urgently needed resources for clinicians and researchers. Thus, we established genome-wide single-cell RNA-sequencing and bulk proteomic data of in vitro and in vivo-generated knockdown models of the pathognomonic DSRCT fusion oncoprotein (EWSR1::WT1) and combined them with an original systems-biology-based pipeline including patient data and the largest histology collection of DSRCTs and morphological mimics available to date. These novel tools were enriched with curated public datasets including patient- and cell line-derived ChIP-seq, bulk and single-cell RNA-seq studies resulting in a multi-model and multi-omic toolbox for discovery analyses. As a proof of concept, our approach revealed the alpha-2/delta subunit of the voltage-dependent calcium channel complex, CACNA2D2, as a highly overexpressed, super-enhancer driven, direct target of EWSR1::WT1. Single-cell and bulk-level analyses of patient samples and xenografted cell lines highlighted CACNA2D2 as a critical component of our newly established EWSR1::WT1 oncogenic signature, that can be employed to robustly identify DSRCT in reference sets. Finally, we show that CACNA2D2 is a highly sensitive and specific single biomarker for fast, simple, and cost-efficient diagnosis of DSRCT. Collectively, we establish a large-scale multi-omics dataset for this devastating disease and provide a blueprint of how such toolbox can be used to identify new and clinically relevant diagnostic markers, which may significantly reduce misdiagnoses, and thus improve patient care.

cancer biology↗