Esophageal adenocarcinoma (EAC) is genetically heterogeneous with relatively few recurrent actionable drivers, which limits the development of targeted therapies. The disease typically arises from the premalignant condition Barrett's esophagus, yet mechanisms that drive progression from Barrett's esophagus to invasive EAC remain incompletely understood. The study summarized in this preprint applied a functional genomics strategy to link loss-of-function events to phenotypic consequences in vivo and to characterize resulting transcriptional states.
This work aimed to define functional tumor suppressors that promote transformation and to map how diverse genetic losses converge on common biological programs that could inform mechanistic studies and translational efforts.
The authors combined three complementary experimental modalities. First, pooled CRISPR-Cas9 loss-of-function screening was used to perturb candidate genes. Second, in vivo tumorigenicity assays assessed which perturbations promote progression to invasive EAC. Third, Perturb-seq transcriptional profiling characterized the gene expression programs associated with specific perturbations. Together these approaches enabled functional annotation of candidate tumor suppressors in a context that models tumor initiation and progression.
The study design links genotype perturbation to both in vivo phenotype and single-cell transcriptional readouts, allowing the authors to identify genes whose loss promotes tumor formation and to define the downstream molecular consequences of those losses.
Using the combined screening and assay strategy, the authors identified 37 genes whose loss-of-function promotes progression from Barrett's esophagus to esophageal adenocarcinoma. These genes are reported as functional tumor suppressors in the context of EAC initiation. The identification of 37 tumor suppressors provides a functional landscape of events that can drive tumorigenesis in this disease setting despite the underlying genetic heterogeneity observed in EAC.
The list of 37 genes is presented in the source article as the output of the integrated pipeline. Details on individual gene effects, effect sizes, or validation experiments beyond the high-level summary were not reported in the abstract.
Although the loss-of-function events were genetically diverse, the associated transcriptional responses clustered into four recurrent programs. These four convergent programs are:
Metabolic reprogramming — alterations in metabolic gene expression consistent with shifts in cellular metabolism during transformation.
Cell cycle progression — transcriptional changes favoring proliferation and cell-cycle entry.
RNA processing — perturbations affecting RNA metabolism and processing pathways.
Cellular motility — programs linked to cytoskeletal and motility-related gene expression that could contribute to invasive behavior.
By mapping individual tumor suppressor losses to these transcriptional programs using Perturb-seq, the study provides a framework for understanding how diverse tumor-suppressive hits can produce shared functional outcomes that promote malignant progression.
Among the tumor suppressors identified, loss of NIPBL, TGFBR2, and RPL22 was specifically reported as mediating resistance to platinum- and taxane-based chemotherapy. This observation links particular loss-of-function events to clinically relevant treatment resistance phenotypes in EAC.
The abstract does not report detailed experimental conditions, quantitative resistance measures, or clinical correlates; such details would be available in the full preprint and require careful review.
The principal implication of this work is that functional annotation can reconcile the genomic heterogeneity observed in EAC by revealing that many distinct tumor-suppressive losses converge on a smaller set of actionable transcriptional programs. This convergence may offer opportunities for therapeutic intervention that target downstream processes rather than individual upstream genetic lesions.
By identifying specific genes whose loss confers chemotherapy resistance, the study also suggests candidate biomarkers or mechanistic targets for overcoming resistance to standard agents such as platinum compounds and taxanes. The authors describe the dataset and results as a resource to guide subsequent mechanistic studies and translational efforts aimed at improving treatment strategies for this aggressive cancer.
This work is presented as a preprint on bioRxiv and was posted July 21, 2026. The preprint has not been certified by peer review. Funding information reported includes support from the National Health and Medical Research Council (grant 1182525). The abstract summarizes the main findings but does not provide full methodological details, quantitative results, or peer-reviewed validation; these limitations should be considered when interpreting the conclusions. Further validation and peer review will be required to confirm the reported tumor suppressors, the mapped transcriptional programs, and the chemotherapy resistance associations.
Collectively, the study introduces a functional genomics resource that links loss-of-function events to in vivo tumorigenicity and single-cell transcriptional phenotypes in EAC, highlighting convergent pathways and specific genes of potential translational interest.