Biology subjects

Bullinger, L.

Publications and source records attributed to Bullinger, L..

9 recordsLinked to original sources

bioRxiv2026

Multiscale biological interactions define clinical trajectories in acute myeloid leukemia

Cancer is characterized by complex interactions across genetic, cellular, and microenvironmental scales. However, a quantitative understanding of how these interactions shape clinical trajectories remains limited. Here, we present a multi-scale single-cell dataset from 184 treatment-naive acute myeloid leukemia (AML) patients spanning all major genetic subtypes, together with an analytical framework to dissect interactions across biological scales. We show that distinct clinical outcomes are encoded by specific cross-scale, cross-compartment interactions present at diagnosis: response to induction therapy is governed by interactions between genetic alterations and leukemic differentiation state; relapse following chemotherapy is associated with non-genetic programs linked to metabolism; and relapse after allogeneic stem cell transplantation is driven by interactions between the immune microenvironment and residual healthy hematopoiesis. Together, our study provides a framework to resolve intra- and inter-patient heterogeneity in cancer and supports a model in which clinical trajectories in AML emerge from defined interactions across biological scales.

cancer biology↗

bioRxiv2026

Hierarchical classification of hematologic malignancies using epigenetic and genetic information

Molecular testing in hematology requires different assays for disease subgroup identification, risk stratification and selection of appropriate treatment regimens. Yet, molecular tests are not necessarily standardized between diagnostic laboratories, resulting in varying turnaround times and potentially divergent results. To resolve this issue and enable single-assay molecular testing, we have developed a hierarchical classification framework that combines epigenetic and genetic data from whole genome nanopore sequencing (WGNS) with machine learning to determine disease entities, epigenetic subgroups (epitypes) and genetic aberrations in hematopoietic neoplasms. We curated DNA methylation data from 5,420 samples and trained a classifier allowing entity-level diagnostics featuring 21 conditions, including healthy controls, acute and chronic myeloid and lymphoid neoplasms. This classifier was subsequently combined with entity-specific epitype classifiers predicting 44 therapeutically or prognostically relevant states, followed by integration of genetic data. Benchmarking of the combined (epi-)genetic testing strategy using WGNS confirmed high accuracy in the detection of diagnostic groups and risk stratification, and identified diagnosis-defining molecular alterations that were not reported by standard-of-care work-up.

cancer biology↗

bioRxiv2025

CCAAT-enhancer binding protein delta functions as a tumor suppressor gene in acute myeloid leukemia

There is a continued need for identification of novel disease drivers of acute myeloid leukemia as many patients experience relapse and have poor clinical outcomes. Analyses from our study and publicly available datasets predicted CEBPD as a novel tumor suppressor gene in acute myeloid leukemia. Consistent with the analyses, CEBPD knockdown experiments showed activation of MAPK signaling with concomitant increase in cell growth rate, while upregulation experiments suggested induction of myeloid differentiation marker CD14 expression in AML cell lines OCI-AML2 and OCI-AML5. Consistent with a previous report, our genomics analyses and azacytidine treatment experiments suggested a role for DNA methylation in downregulation of CEBPD expression during AML pathogenesis. Altogether, our results provide experimental evidence for a tumor suppressor function of CEBPD in AML.

cancer biology↗

bioRxiv2024

Pyjacker identifies enhancer hijacking events in acute myeloid leukemia including MNX1 activation via deletion 7q

Acute myeloid leukemia with complex karyotype (ckAML) is characterized by high genomic complexity, including frequent TP53 mutations and chromothripsis. We hypothesized that the numerous genomic rearrangements could reposition active enhancers near proto-oncogenes, leading to their aberrant expression. We developed pyjacker, a computational tool for the detection of enhancer hijacking events, and applied it to a cohort of 39 ckAML samples. Pyjacker identified motor neuron and pancreas homeobox 1 (MNX1), a gene aberrantly expressed in 1.4% of AML patients, often as a result of del(7)(q22q36) associated with hijacking of a CDK6 enhancer. MNX1-activated cases show significant co-occurrence with BCOR mutations and a gene signature shared with t(7;12)(q36;p13) pediatric AML. We demonstrated that MNX1 is a dependency gene, as its knockdown in a xenograft model reduces leukemia cell fitness. In conclusion, enhancer hijacking is a frequent mechanism for oncogene activation in AML. Statement of significanceThis study examines the consequences of structural alterations and demonstrates that proto-oncogene activation by enhancer hijacking is an overlooked pathomechanism in AML. MNX1 overexpression demonstrates that deletions on chromosome 7q can not only lead to haploinsufficiency, but also to activation of oncogenes by enhancer hijacking, providing a novel leukemogenic mechanism.

cancer biology↗

bioRxiv2023

Altered enhancer-promoter interaction leads to MNX1 expression in pediatric acute myeloid leukemia with t(7;12)(q36;p13)

Acute myeloid leukemia (AML) with the t(7;12)(q36;p13) translocation occurs only in very young children and has a poor clinical outcome. The expected oncofusion between breakpoint partners (MNX1 and ETV6) has only been reported in a subset of cases. However, a universal feature is the strong transcript and protein expression of MNX1, a homeobox transcription factor that is normally not expressed in hematopoietic cells. Here, we map the translocation breakpoints on chromosomes 7 and 12 in affected patients to a region proximal to MNX1 and either introns 1 or 2 of ETV6. The frequency of MNX1 overexpression in pediatric AML (n=1556, own and published data) is 2.4% and occurs predominantly in t(7;12)(q36;p13) AML. Chromatin interaction assays in a t(7;12)(q36;p13) iPSC cell line model unravel an enhancer-hijacking event that explains MNX1 overexpression in hematopoietic cells. Our data suggest that enhancer-hijacking is a more common and overlooked mechanism for structural rearrangement-mediated gene activation in AML. Key pointsO_LIExpression analysis of over 1500 pediatric AML samples demonstrates MNX1 expression as a universal feature of t(7;12)(q36;p13) AML as well as in rare cases without t(7;12)(q36;p13) C_LIO_LIMNX1 is activated by an enhancer-hijacking event in t(7;12)(q36;p13) AML and not, as previously postulated, by the creation of a MNX1::ETV6 oncofusion gene. C_LI

cancer biology↗

bioRxiv2023

Covering Hierarchical Dirichlet Mixture Models on binary data to enhance genomic stratifications in Onco-Hematology

Onco-hematological studies are increasingly adopting statistical mixture models to support the advancement of the genetically-driven classification systems for blood cancer. Targeting enhanced patients stratification based on the sole role of molecular biology attracted much interest and contributes to bring personalized medicine closer to reality. In particular, Dirichlet processes have become the preferred method to approach the fit of mixture models. Usually, the multinomial distribution is at the core of such models. However, despite their advanced statistical formalism, these processes are not to be considered black box techniques and a better understanding of their working mechanisms enables to improve their employment and explainability. Focused on genomic data in Acute Myeloid Leukemia, this work unfolds the driving factors and rationale of the Hierarchical Dirichlet Mixture Models of multinomials on binary data. In addition, we introduce a novel approach to perform accurate patients clustering via multinomials based on statistical considerations. The newly reported adoption of the Multivariate Fishers Non-Central Hypergeometric distributions reveals promising results and outperformed the multinomials in clustering both on simulated and real onco-hematological data. Author summaryExplainable models are particularly attractive nowadays since they have the advantage to convince clinicians and patients. In this work we show that a deeper understanding of the Hierarchical Dirichlet Mixture Model, a non-black box method, can lead to better data modelling. In onco-hematology Hierarchical Dirichlet Mixture Models typically help to cluster molecular alterations rather than patients. Here, an intuitive statistical approach is presented to tackle patient classification based on the Hierarchical Dirichlet Mixture Models outcome. Additionally, molecular alterations are usually modelled by Hierarchical Dirichlet Mixture Models as a mixture of multinomial distributions. This work highlights that the alternative Fishers Non-Central Hypergeometric distribution can provide even better results and can give a higher priority to rare molecular alterations for patient classification.

cancer biology↗

bioRxiv2023

Imetelstat-Mediated Alterations in Fatty Acid Metabolism To Induce Ferroptosis As Therapeutic Strategy for Acute Myeloid Leukemia

Telomerase enables replicative immortality in most cancers including acute myeloid leukemia (AML). Imetelstat is a first-in-class telomerase inhibitor with clinical efficacy in myelofibrosis and myelodysplastic syndromes. Here, we develop an AML patient-derived xenograft (PDX) resource, and perform integrated genomics, transcriptomics, and lipidomics analyses combined with functional genetics to identify key mediators of imetelstat efficacy. In a randomized Phase II-like preclinical trial in PDX, imetelstat effectively diminishes AML burden, and preferentially targets subgroups containing mutant NRAS and oxidative stress-associated gene expression signatures. Unbiased, genome-wide CRISPR/Cas9 editing identifies ferroptosis regulators as key mediators of imetelstat efficacy. Imetelstat promotes the formation of polyunsaturated fatty acid-containing phospholipids, causing excessive levels of lipid peroxidation and oxidative stress. Pharmacological inhibition of ferroptosis diminishes imetelstat efficacy. We leverage these mechanistic insights to develop an optimized therapeutic strategy using oxidative stress-inducing chemotherapy to sensitize patient samples to imetelstat causing significant disease control in AML.

cancer biology↗

bioRxiv2022

Longitudinal single-cell transcriptomics reveals distinct patterns of recurrence in acute myeloid leukemia

The heterogeneity and evolution of AML blasts can render therapeutic interventions ineffective in a yet poorly understood patient-specific manner. To gain insight into the clonal heterogeneity of diagnosis (Dx) and relapse (Re) pairs, we employed whole-exome sequencing and single-cell RNA-seq to longitudinally profile two t(8;21) (AML1-ETO = RUNX1-RUNX1T1), and four FLT3-ITD AML cases. The single cell RNA data underpinned the tumor heterogeneity amongst patient blasts. The Dx-Re transcriptomes of high risk FLT3-ITD pairs formed a continuum from extensively changed in the absence of significantly mutational changes in AML-associated genes to rather similar Dx-Re pair of an intermediate risk FLT3-ITD. In one high risk FLT3-ITD pair, a pathway switched from an AP-1 regulated network in Dx to mTOR signaling in Re. The distinct AML1-ETO pairs comprise clusters that share genes related to hematopoietic stem cell maintenance and cell migration suggesting that the Re leukemic stem cell-like (LSC-like) cells probably evolved from the Dx LSC-like cells. In summary, our study revealed a continuum from drastic transcriptional changes to extensive similarities between respective Dx-Re pairs that are poorly explained by the well-established model of clonal evolution. Our results suggest alternative and currently unappreciated and unexplored mechanisms leading to therapeutic resistance and AML recurrence.

cancer biology↗

bioRxiv2016

Partitioned learning of deep Boltzmann machines for SNP data

Learning the joint distributions of measurements, and in particular identification of an appropriate low-dimensional manifold, has been found to be a powerful ingredient of deep leaning approaches. Yet, such approaches have hardly been applied to single nucleotide polymorphism (SNP) data, probably due to the high number of features typically exceeding the number of studied individuals. After a brief overview of how deep Boltzmann machines (DBMs), a deep learning approach, can be adapted to SNP data in principle, we specifically present a way to alleviate the dimensionality problem by partitioned learning. We propose a sparse regression approach to coarsely screen the joint distribution of SNPs, followed by training several DBMs on SNP partitions that were identified by the screening. Aggregate features representing SNP patterns and the corresponding SNPs are extracted from the DBMs by a combination of statistical tests and sparse regression. In simulated case-control data, we show how this can uncover complex SNP patterns and augment results from univariate approaches, while maintaining type 1 error control. Time-to-event endpoints are considered in an application with acute myeloid lymphoma patients, where SNP patterns are modeled after a pre-screening based on gene expression data. The proposed approach identified three SNPs that seem to jointly influence survival in a validation data set. This indicates the added value of jointly investigating SNPs compared to standard univariate analyses and makes partitioned learning of DBMs an interesting complementary approach when analyzing SNP data.

bioinformatics↗