Luca Ducoli

Insight into population-level gene-environment (GxE) interactions is a major goal in understanding polygenic diseases, such as breast cancer. We integrated massively parallel reporter assays (MPRA), chromatin profiling, and network analysis to identify estrogen-responsive breast cancer variants. Screening 1,604 GWAS-identified breast cancer variants identified 73 estrogen-modulated SNVs (emSNVs). Chromatin accessibility modeling validated allele-specific effects, identifying variants that disrupt pioneer factor binding to chromatin and others that modulate transcription factor (TF) recruitment to pre-accessible enhancers. emSNV targets converged on pathways including mitochondrial metabolism, NF-κB signaling, and chromatin regulation. Aggregating emSNVs into a polygenic risk score (PRSE2) revealed interactions with reproductive risk factors in 13,026 post-menopausal BRCA cases and 108,265 controls, including age at first birth (p=0.0052); a control PRS lacking estrogen-responsive variants showed no interactions. This framework bridges molecular and epidemiological GxE studies to uncover variants whose disease associations depend on environmental context, with implications for understanding polygenic disease risk.

RNA-dependent protein assemblies — including the spliceosome, ribosome and RNA-dependent membraneless organelles — have crucial roles in diverse cellular processes through RNA scaffolding and hierarchical assembly. Various empirical techniques and artificial intelligence algorithms have been developed to help understand the architecture, dynamics and functional implications of RNA–protein complexes, and their further development is underway to comprehensively integrate this information. This Review explores how combining these diverse technologies will enhance our understanding of the biological functions of RNA-dependent protein assemblies. We first explore methodological frontiers, contrasting traditional approaches with new platforms, which enable the identification and tracking of RNA–protein assembly dynamics on the same RNA molecules. We then present avenues for integrating these new experimental techniques with machine-learning methods to improve both predictive models of RNA–protein assembly and functional RNA design. We discuss how the synergy between experimental and digital biology can drive new insights into disease mechanisms and therapeutic strategies, including targeted modulation of pathogenic RNA–protein assemblies. Finally, we examine roadmaps for future research, emphasizing the potential of closed-loop systems that iteratively refine our understanding of RNA–protein assemblies through cycles of hypothesis generation, prediction, experimentation and validation.

Identifying noncoding single nucleotide variants (SNVs) in regulatory DNA linked to polygenic disease risk, the transcription factors (TFs) they bind, and the genes they dysregulate is a goal in polygenic disease research. Here, we use massively parallel reporter analysis of 3451 SNVs linked to risk for polygenic skin diseases with disrupted epidermal homeostasis to identify 355 differentially active SNVs (daSNVs). daSNV target gene analysis, combined with daSNV editing, underscored dysregulated epidermal differentiation as a shared pathomechanism. CRISPR knockout screens of 1772 human TFs revealed 123 TFs essential for epidermal homeostasis, highlighting ZNF217 and CXXC1. Population sampling CUT&RUN of 27 homeostatic TFs identified allele-specific DNA binding (ASB) differences at daSNVs enriched near epidermal homeostasis and monogenic skin disease genes, with notable representation of SP/KLF and AP-1/2 TFs. High TF-occupancy promoters were “buffered” against ASB. This resource implicates dysregulated binding of specific homeostatic TF families in risk for diverse polygenic skin diseases.

RNA-binding proteins (RBPs) control varied processes, including RNA splicing, stability, transport and translation^1,2,3. Dysfunctional RNA–RBP interactions contribute to the pathogenesis of human disease^1,4,5; however, characterizing the nature and dynamics of multiprotein assemblies on RNA has been challenging. Here, to address this, non-isotopic ligation-based ultraviolet-light-induced cross-linking and immunoprecipitation⁶ was combined with mass spectrometry (irCLIP-RNP) to identify RNA-dependent associated proteins (RDAPs) co-bound to RNA with any RBP of interest. irCLIP-RNP defined landscapes of multimeric protein assemblies on RNA, revealing patterns of RBP–RNA associations, including cell-type-selective combinatorial relationships between RDAPs and primary RBPs. irCLIP-RNP also defined dynamic RDAP remodelling in response to epidermal growth factor (EGF), revealing that EGF-induced recruitment of UPF1 adjacent to HNRNPC promotes splicing surveillance of cell proliferation mRNAs. To identify the RNAs simultaneously co-bound by multiple studied RBPs, a sequential immunoprecipitation irCLIP (Re-CLIP) method was also developed. Re-CLIP confirmed binding relationships observed in irCLIP-RNP and identified HNRNPC and UPF1 RBP co-binding on RND3 and DDX3X mRNAs. irCLIP-RNP and Re-CLIP provide a framework to identify and characterize dynamic RNA–protein assemblies in living cells.

Cutaneous squamous cell carcinoma (cSCC) accounts for 20% of all skin cancer deaths globally, making it the second-highest subtype of skin cancer. The prevalence of cSCC in humans, as well as the poor capacity for an efficient prognosis, highlights the need to uncover alternative actors and mechanisms at the foundation of skin cancer development. Significant advances have been made to better understand some key factors in cSCC progression. However, little is known about the role of noncoding RNAs, particularly of a specific category called long noncoding RNA (lncRNA). By performing pseudobulk analysis of single-cell sequencing data from normal and cSCC human skin tissues, we determined a global portfolio of lncRNAs specifically expressed in keratinocyte subpopulations. Integration of CRISPR interference screens in vitro and the xenograft model identified several lncRNAs impacting the growth of cSCC cancer lines both in vitro and in vivo. Among these, we further validated LINC00704 and LINC01116 as proliferation-regulating lncRNAs in cSCC lines and potential biomarkers of cSCC progression. Taken together, our study provides a comprehensive signature of lncRNAs with roles in regulating cSCC progression.

Tumor-draining lymph nodes (LNs), composed of lymphocytes, antigen-presenting cells, and stromal cells, are highly relevant for tumor immunity and the efficacy of immunotherapies. Lymphatic endothelial cells (LECs) represent an important stromal cell type within LNs, and several distinct subsets of LECs that interact with various immune cells and regulate immune responses have been identified. In this study, we used single-cell RNA sequencing (scRNA-seq) to characterize LECs from LNs draining B16F10 melanomas compared to non-tumor-draining LNs. Several upregulated genes with immune-regulatory potential, especially in LECs lining the subcapsular sinus floor (fLECs), were identified and validated. Interestingly, some of these genes, namely, podoplanin, CD200, and BST2, affected the adhesion of macrophages to LN LECs in vitro. Congruently, lymphatic-specific podoplanin deletion led to a decrease in medullary sinus macrophages in tumor-draining LNs in vivo. In summary, our data show that tumor-derived factors induce transcriptional changes in LECs of the draining LNs, especially the fLECs, and that these changes may affect tumor immunity. We also identified a new function of podoplanin, which is expressed on all LECs, in mediating macrophage adhesion to LECs and their correct localization in LN sinuses.