Biology subjects

Deplancke, B.

Publications and source records attributed to Deplancke, B..

At least 19 recordsLinked to original sources

bioRxiv2026

Growth factor-free synthetic matrices reveal intrinsic skeletal progenitor competence

Clinical translation of human skeletal progenitor cells has been limited by the inability to reliably predict their regenerative performance, as current characterization methods rely on phenotypic markers or instructive differentiation assays that obscure intrinsic cell function. Here we demonstrate that human skeletal progenitor competence can only be revealed within a precisely defined and growth factor-free synthetic microenvironment. By independently tuning matrix mechanics, degradability, adhesion, and cell density within a fully synthetic poly(ethylene glycol) hydrogel, we create a permissive niche that enables human bone marrow stromal cells (hBMSCs) to undergo spontaneous bone formation in vivo without osteoinductive supplementation. These exogenous growth factor-free conditions expose donor-specific differences in bone forming capacity that are otherwise masked. Strikingly, supplementation with even low concentrations of BMP-2 abolished inter-donor variability, directly demonstrating that osteoinductive factors override intrinsic progenitor competence. Single-cell transcriptomics, proteomic profiling, and longitudinal in vivo analyses converge to show that high performing donors are defined by chondrocyte-primed transcriptional states and coordinated extracellular matrix (ECM) remodeling programs that establish a pro-osteochondral niche and drive bone formation via endochondral ossification. In contrast, low performing donors fail to initiate this cascade despite comparable phenotypic profiles and proliferation rates. Together, these findings implicate matrix context as a critical regulator of stem cell competence. Within this permissive synthetic niche, intrinsic ECM remodeling capacity emerges as a defining feature of donor-specific osteogenic potential, one that remains invisible to conventional assays. These matrices thus provide a functional platform for prospective stratification of skeletal progenitor cells, with direct implications for donor selection in regenerative medicine. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=74 SRC="FIGDIR/small/731042v1_ufig1.gif" ALT="Figure 1"> View larger version (38K): org.highwire.dtl.DTLVardef@6fa9ceorg.highwire.dtl.DTLVardef@f51dddorg.highwire.dtl.DTLVardef@7bb040org.highwire.dtl.DTLVardef@706a36_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗

bioRxiv2026

Epithelial innate immune sensing of pneumococci is inherently restricted to a small cellular minority across species and infection niches

Streptococcus pneumoniae colonises the nasopharynx asymptomatically yet causes life-threatening invasive disease. How it navigates early epithelial immune surveillance to cause disease remains unclear. Conventional innate immune models predict coordinated, population-wide epithelial responses to bacterial infection. Using single-cell RNA sequencing, RNA fluorescence in situ hybridization and in vivo mouse and zebrafish models, pneumococcal infection is instead shown to activate innate immune genes including chemokine, NF-{kappa}B regulatory, and prostaglandin pathway genes, in only 1-4% of lung epithelial cells. This restriction is seemingly pneumococcal-specific as Escherichia coli triggers responses in over 40% of the same cells. Strikingly, nasopharyngeal epithelial cells show complete immune silence to pneumococci while responding robustly to E. coli and Staphylococcus aureus, suggesting niche-specific immune evasion. Additionally, pharmacological inhibition of COX-2 significantly increased mortality in a zebrafish meningitis model, identifying prostaglandin signalling as a protective host response during invasive disease. Competence-associated surface remodelling contributes modestly and incrementally to immune restriction, while the predominant dampening is competence-independent. These findings challenge canonical epithelial immunity models against bacterial infection and provide a cellular framework for understanding pneumococcal commensalism and pathogenesis. Significance statementClassical innate immune models predict that bacterial infection triggers coordinated, population-wide transcriptional responses across the epithelium. Using single-cell RNA sequencing, RNA fluorescence in situ hybridization, and in vivo zebrafish and mouse models, we show that Streptococcus pneumoniae, responsible for over one million deaths annually, activates innate immune genes in only 1-4% of lung epithelial cells. Escherichia coli triggers responses in over 40% of the same cells, demonstrating this restriction is pneumococcal-specific. Nasopharyngeal epithelial cells, the bacteriums primary colonization niche, show complete immune silence to pneumococci, suggesting niche-specific evolutionary adaptation. The prostaglandin pathway is identified as a protective host response during invasive disease. These findings challenge canonical models of epithelial immunity and provide a cellular framework for pneumococcal commensalism and pathogenesis.

microbiology↗

bioRxiv2026

scFAIR Consortium: a decentralized hub for single-cell RNA-Seq data standardization and unification

The rapid accumulation of single-cell RNA-Seq (scRNA-seq) data across multiple repositories presents major challenges for data accessibility, integration, and reproducibility. While primary repositories provide raw data, they rarely include structured cell-type annotations or descriptions of analytical workflows, limiting the ability to reuse and integrate datasets in a FAIR (Findable, Accessible, Interoperable, Reusable) manner. Here we present scFAIR, a consortium of single-cell data resources that has developed a unified metadata schema and common curation framework to improve the FAIRness of scRNA-seq data. Building on and extending the CZ CELLxGENE Discover metadata schema, the scFAIR consortium has been instrumental in driving key schema improvements, including the expansion of supported organisms, richer biological context, and structured reporting of computational workflows. To provide unified access to decentralized datasets, the consortium developed the sc-fair.org portal, which currently aggregates 2,346 datasets across partner resources through ontology-aware semantic search. We demonstrate the practical value of FAIR-compliant datasets through a cross-species validation between human and mouse Allen Brain Atlases, showing that standardized ontology annotations enable reliable annotation transfer across species, with 90% of neuronal clusters receiving an exact or equivalent label. Together, the scFAIR schema, validator, and portal constitute a community-driven framework that advances single-cell data standardization and lays the foundation for reproducible, large-scale integration of single-cell datasets.

bioinformatics↗

bioRxiv2026

Generative modeling reveals the connection between cellular morphology and gene expression

The understanding of how transcriptional programs give rise to cellular morphology, and how morphological features reflect and influence cell identity and function remains limited. This is due in part to the lack of large-scale datasets pairing the two modalities as well as the absence of computational frameworks capable of modeling their cross-modal structure. Here, we introduce COSMIC, a bidirectional generative framework that enables quantitative decomposition of transcriptional variance reflected in morphology and morphological variance explained by gene expression. COSMIC builds on a foundation model trained on over 21 million segmented nuclei and couples it with existing transcriptomic embeddings. To enable cross-modal learning, we leveraged a newly generated multimodal dataset acquired using IRIS, a technology that captures high-resolution images and transcriptomes from the same single cells at scale. COSMIC accurately modeled cell type identity, as well as continuous dynamics such as cell-cycle progression, establishing a quantitative link between morphological phenotypes and underlying gene expression. In prostate cancer cells, COSMIC identified morphological and transcriptomic differences between chemotherapy drug treatment-responsive and -resistant cells, and revealed morphology-associated genes linked to tumor state. Together, these results demonstrate that generative modeling powered by paired single-cell measurements can capture the bidirectional flow of information between cellular form and gene expression, opening new avenues for mechanistic discovery and predictive modeling in both basic and translational cell biology.

bioinformatics↗

bioRxiv2025

Single-cell phenomics through integrated imaging and molecular profiling

Single-cell technologies such as transcriptomics, microscopy, and flow cytometry have revolutionized the study of cellular identity and function. While each of these technologies is powerful on its own, their full potential lies in their integration, enabling multimodal profiling of the same cell and revealing how distinct modalities influence one another. Here, we introduce IRIS (Interconnected Robotic Imaging and Single cell transcriptomics), a deterministic single-cell platform technology that seamlessly couples high-resolution microscopy with droplet-based single-cell RNA sequencing. IRIS enables precise cell positioning, multimode imaging across brightfield and fluorescent channels, and subsequent molecular capture from the same cell, directly linking high-resolution morphological features to matched transcriptomes. We validate IRIS by recovering cell cycle progression states and transcriptional programmes associated with canonical morphologies and demonstrate its discovery power by molecularly resolving two nuclear-ER architectures within naive CD8+ T cells, each defined by distinct gene expression profiles and functional markers. IRIS establishes an integrative single-cell phenomics framework, opening new avenues for dissecting how cellular form relates to molecular state and function.

immunology↗

bioRxiv2025

Inferring binding specificities of human transcription factors with the wisdom of crowds

DNA motif discovery and, particularly, computational modeling of transcription factor binding motifs, has been a mecca of algorithmic bioinformatics for several decades. Here, we report the results of the largest open community challenge in Inferring BInding Specificities (IBIS), where participants all over the world were invited to construct binding specificity models from multi-assay experimental data for poorly studied human transcription factors. The submissions were rigorously tested against a rich held-out dataset. Benchmarking demonstrated a consistent advantage of properly designed deep learning models over traditional positional weight matrices and other machine learning methods. Yet, the positional weight matrices displayed a surprisingly strong performance out of the box, being only slightly behind the best deep learning models. A post-challenge assessment of a selection of other deep learning methods further solidified this finding. IBIS highlights the power of benchmarking in finding adequate DNA motif representations, emphasizes the pros and cons of various machine learning methods applied to DNA motif modeling, and establishes a rich dataset, benchmarking protocols, and computational framework for a fair cross-platform evaluation of future models of transcription factor binding motifs in DNA sequences. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=175 SRC="FIGDIR/small/688692v1_ufig1.gif" ALT="Figure 1"> View larger version (64K): org.highwire.dtl.DTLVardef@1c6677corg.highwire.dtl.DTLVardef@b4124aorg.highwire.dtl.DTLVardef@1ce2b1org.highwire.dtl.DTLVardef@66e917_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2025

The cellular basis of mitochondrial stress signaling in the brain

Neuronal vulnerability to stress is highly cell-type specific, but the underlying mechanisms are poorly understood. Here we show that the ability to activate endoplasmic reticulum (ER) stress signaling is encoded by the relative mitochondrial metabolic activity of neurons. Genetically inducing mitochondrial dysfunction in the Drosophila brain caused the emergence of a novel cluster of neurons, detected using single nuclear RNA-sequencing, that activated the ER unfolded protein response (UPR). UPR activation occurred only in neurons with high mitochondrial activity. Unexpectedly, mitochondria-ER contacts were also abundant in the neurons that activated the UPR but virtually undetectable in neurons with low mitochondrial activity. In the human brain, strikingly, excitatory neurogranin neurons had the highest mitochondrial activity and triggered the UPR. The selective activation of the UPR only in neurons with high mitochondrial activity is therefore conserved. This study reveals the remarkable dependence on the inherent oxidative metabolic activity of neurons to activate the UPR.

neuroscience↗

bioRxiv2024

A three-dimensional ex vivo model recapitulates in vivo features and unravels increased drug resistance in childhood acute lymphoblastic leukemia

Acute lymphoblastic leukemia (ALL) preferentially localizes in the bone marrow (BM) and displays recurrent patterns of medullary and extra-medullary involvement. Leukemic cells exploit their niche for propagation and survive selective pressure by chemotherapy in the BM microenvironment, suggesting the existence of protective mechanisms. Here, we established a three-dimensional (3D) BM mimic with human mesenchymal stromal cells and endothelial cells that resemble vasculature-like structures to explore the interdependence of leukemic cells with their microenvironment. This model recapitulates recurrent topologic differences between B-cell and T-cell precursor ALL, whereby B-ALL interacts more closely with the mesenchymal compartment. Migration versatility was found to be associated with subtype, consistent with increased motility observed in T-ALL in vivo. Single-cell RNA signatures revealed similarities to profiles from in vivo patient derived xenografts, suggesting relevant states ex vivo. Furthermore, enhanced migration, adherence and cell cycle heterogeneity was visualized in our co-culture model. Finally, drug response profiling experiments in this 3D system reproduced established response patterns and indicated that drug resistant leukemic subpopulations may be detected more faithfully compared to information from two-dimensional models.

cancer biology↗

bioRxiv2024

EXTRA-seq: a genome-integrated extended massively parallel reporter assay to quantify enhancer-promoter communication

Precise control of gene expression is essential for cellular function, but the mechanisms by which enhancers communicate with promoters to coordinate this process are not fully understood. While sequence-based deep learning models show promise in predicting enhancer-driven gene expression, experimental validation and human-interpretable mechanistic insights lag behind. Here, we present EXTRA-seq, a novel EXTended Reporter Assay followed by sequencing designed to quantify enhancer activity in endogenous contexts over kilobase-scale distances. We demonstrate that EXTRA-seq can be targeted to disease-relevant loci and captures expression changes at the resolution of individual transcription factor binding sites, enabling mechanistic discovery. Using engineered synthetic enhancer-promoter combinations, we reveal that the TATA-box acts as a dynamic range amplifier, modulating expression levels in function of enhancer strength. Importantly, we find that integrating state-of-the-art deep learning models with plasmid-based enhancer assays improves the prediction of gene expression as measured by EXTRA-seq. These findings open new avenues for predictive modeling and therapeutic applications. Overall, our work provides a powerful experimental platform to interrogate the complex interplay between enhancers and promoters, bridging the gap between in silico predictions and human-interpretable biological mechanisms.

genomics↗

bioRxiv2024

Cross-platform DNA motif discovery and benchmarking to explore binding specificities of poorly studied human transcription factors

A DNA sequence pattern, or "motif", is an essential representation of DNA-binding specificity of a transcription factor (TF). Any particular motif model has potential flaws due to shortcomings of the underlying experimental data and computational motif discovery algorithm. As a part of the Codebook/GRECO-BIT initiative, here we evaluated at large scale the cross-platform recognition performance of positional weight matrices (PWMs), which remain popular motif models in many practical applications. We applied ten different DNA motif discovery tools to generate PWMs from the "Codebook" data comprised of 4,237 experiments from five different platforms profiling the DNA-binding specificity of 394 human proteins, focusing on understudied transcription factors of different structural families. For many of the proteins, there was no prior knowledge of a genuine motif. By benchmarking-supported human curation, we constructed an approved subset of experiments comprising about 30% of all experiments and 50% of tested TFs which displayed consistent motifs across platforms and replicates. We present the Codebook Motif Explorer (https://mex.autosome.org), a detailed online catalog of DNA motifs, including the top-ranked PWMs, and the underlying source and benchmarking data. We demonstrate that in the case of high-quality experimental data, most of the popular motif discovery tools detect valid motifs and generate PWMs, which perform well both on genomic and synthetic data. Yet, for each of the algorithms, there were problematic combinations of proteins and platforms, and the basic motif properties such as nucleotide composition and information content offered little help in detecting such pitfalls. By combining multiple PMWs in decision trees, we demonstrate how our setup can be readily adapted to train and test binding specificity models more complex than PWMs. Overall, our study provides a rich motif catalog as a solid baseline for advanced models and highlights the power of the multi-platform multi-tool approach for reliable mapping of DNA binding specificities. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=141 SRC="FIGDIR/small/619379v2_ufig1.gif" ALT="Figure 1"> View larger version (61K): org.highwire.dtl.DTLVardef@79561forg.highwire.dtl.DTLVardef@54c0aorg.highwire.dtl.DTLVardef@1c33f34org.highwire.dtl.DTLVardef@16a93ba_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical AbstractC_FLOATNO C_FIG

bioinformatics↗

bioRxiv2024

Identification of methylation-sensitive human transcription factors using meSMiLE-seq

Transcription factors (TFs) are key players in eukaryotic gene regulation, but the DNA binding specificity of many TFs remains unknown. Here, we assayed 284 mostly poorly characterized, putative human TFs using selective microfluidics-based ligand enrichment followed by sequencing (SMiLE-seq), revealing 72 new DNA binding motifs. To investigate whether some of the 158 TFs for which we did not find motifs preferably bind epigenetically modified DNA (i.e. methylated CG dinucleotides), we developed methylation-sensitive SMiLE-seq (meSMiLE-seq). This microfluidic assay simultaneously probes the affinity of a protein to methylated and unmethylated DNA, augmenting the capabilities of the original method to infer methylation-aware binding sites. We assayed 114 TFs with meSMiLE-seq and identified DNA-binding models for 48 proteins, including the known methylation-sensitive binding modes for POU5F1 and RFX5. For 11 TFs, binding to methylated DNA was preferred or resulted in the discovery of alternative, methylation-dependent motifs (e.g. PRDM13), while aversion towards methylated sequences was found for 13 TFs (e.g. USF3). Finally, we uncovered a potential role for ZHX2 as a putative binder of Z-DNA, a left-handed helical DNA structure which is adopted more frequently upon CpG methylation. Altogether, our study significantly expands the human TF codebook by identifying DNA binding motifs for 98 TFs, while providing a versatile platform to quantitatively assay the impact of DNA modifications on TF binding.

genomics↗

bioRxiv2024

Perspectives on Codebook: sequence specificity of uncharacterized human transcription factors

Gene expression is regulated by transcription factors (TFs), which recognize specific DNA sequence motifs. Several hundred putative human TFs, identified mainly by an apparent DNA-binding domain, lack known binding motifs1, and even for well-characterized TFs, it remains controversial to what degree motifs accurately reflect binding sites in living cells2,3. Here, we describe a systematic effort ("Codebook") to determine the sequence specificity of 332 putative and poorly characterized human TFs. Over 4,000 independent experiments, encompassing multiple in vitro and in vivo assays, produced motifs for just over half (177, or 53%), of which most are unique to a single protein, thereby extending the vocabulary of sequence recognition encoded by human TFs by [~]100 distinct motifs. Moreover, binding motifs identified in vitro are strongly enriched within cellular binding sites. Collectively, the data reveal tens of thousands of previously unknown, conserved, and direct TF binding sites across the human genome. These sites are concentrated in promoter regions, and are predictive of gene expression, illustrating that this new data atlas provides an important step forward in decoding the human genome.

genomics↗

bioRxiv2024

Lack of Single Amino Acids Transcriptionally Remodels Sensory Systems to Enhance the Intake of Protein and Microbiota

Adequate intake of dietary essential amino acids (eAAs) is vital for protein synthesis and metabolism. Any single eAA deprivation is sufficient to increase protein intake in Drosophila melanogaster. How such nutritional "needs" are transformed into behavioral "wants" remains poorly understood. We derived transcriptomes from the heads of flies deprived of individual eAAs to identify mechanisms by which this is achieved. We found that, while specific eAA deprivations have unique effects on gene expression, a large set of changes are shared across deprivations. We show that Or92a upregulation upon eAA deprivation increases the exploitation of yeast, the main protein source of flies. Furthermore, Ir76a upregulation was crucial for feeding on Lactobacillus, a gut bacterium that ameliorates the fitness of eAA-deprived flies. Our work uncovers common and unique transcriptional changes induced by individual eAA deprivations in an animal and reveals novel mechanisms underlying the organisms behavioral and physiological response to eAA challenges.

neuroscience↗

bioRxiv2024

Vestigial like 4 regulates the adipogenesis of classical brown adipose tissue

Brown adipose tissue (BAT) is mammals primary non-shivering thermogenesis organ, and the molecular mechanisms regulating BAT growth and adipogenesis are largely unknown. The Hippo-YAP pathway has been well-known for controlling organ size, and Vestigial like 4 (VGLL4) is a transcriptional regulator that modulates the Hippo-YAP pathway by competing against YAP for binding to TEAD proteins. In this study, we dissected the function of VGLL4 in regulating BAT development. We generated a conventional Vgll4 mutant mouse line, in which the two Tondu (TDU) domains of VGLL4 were disrupted. We found that deletion of the TDU domains of VGLL4 resulted in perinatal lethality and paucity of the interscapular BAT. Histological and magnetic resonance imaging studies confirmed that the adipogenesis of BAT was impaired in Vgll4 mutants. Adeno-associated virus (AAV) mediated, brown adipocyte-specific overexpression of VGLL4 increased BAT volume and protected the adult male mice from acute cold stress. Genomic studies suggest that VGLL4/TEAD1 complex directly regulates the myogenic and adipogenic gene expression programs of BAT. In conclusion, our data identify VGLL4 as a previously unrecognized adipogenesis factor that regulates classical BAT development.

developmental biology↗

bioRxiv2024

Food texture preference reveals multisensory contributions of gustatory organs in behaviour and physiology

Food presents a multisensory experience, with visual, taste, and olfactory cues being important in allowing an animal to determine the safety and nutritional value of a given substance1. Texture, however, remains a surprisingly unexplored aspect, despite providing key information about the state of the food through properties such as hardness, liquidity, and granularity. Food perception is achieved by specialised sensory neurons, which themselves are defined by the receptor genes they express. While it was assumed that sensory neurons respond to one or few closely-related stimuli, more recent findings challenge this notion and support evidence that certain sensory neurons are more broadly tuned. In the Drosophila taste system, gustatory neurons respond to cues of opposing hedonic valence or to olfactory cues. Here, we identified that larvae ingest and navigate towards specific food substrate hardnesses, and probed the role of gustatory organs in this behaviour. By developing a genetic tool targeting specifically gustatory organs, we show that these organs are major contributors for evaluation of food texture and ingestion decision-making. We find that ablation of gustatory organs not only results in loss of chemosensation, but also navigation and ingestion preference to varied substrate textures. Furthermore, we show that certain neurons in the primary taste organ exhibit varied and concurrent physiological responses to mechanical and multimodal stimulation. We show that individual neurons house independent mechanisms for multiple sensory modalities, challenging assumptions about capabilities of sensory neurons. We propose that further investigations, across the animal kingdom, may reveal higher sensory complexity than currently anticipated.

neuroscience↗

bioRxiv2024

A humoral immune response to parasitoid wasps in Drosophila is regulated by JAK/STAT, NF-κB and GATA

The two arms of innate immunity consist of the cell-mediated cellular defenses and the systemic humoral immune responses. Drosophila humoral immune defenses in the context of antimicrobial immunity, particularly the regulation and activation of antimicrobial peptide secretion from the fat body, have been studied extensively. How Drosophila regulates humoral immunity against another major natural enemy, the parasitoid wasp, is less well-characterized. In this study, we focused on a gene crucial in anti-parasitoid immunity, lectin-24A, which is specifically induced following parasitization. We found that a fluorescent reporter driven by the region upstream of lectin-24A showed localized posterior expression in the larval fat body, the Drosophila tissue mediating humoral immunity. Furthermore, with RNA sequencing of the anterior and posterior fat body sections, we found that components of JAK/STAT, GATA, and Toll pathways were regulated differentially in the anterior-posterior axis of the fat body and/or by infection. Predicted binding motifs for transcription factors in all three of these pathways were identified in the 444bp upstream region of the lectin-24A gene, where scrambling these motifs leads to reduced basal or induced expression of the fluorescent reporter. Investigating each of these pathways, we found that JAK/STAT, the GATA factor Pannier, and the NF-{kappa}B factor dorsal all modulate the expression of lectin-24A. The binding motifs associated with these transcription factors were also enriched in the upstream sequences of parasitism-induced genes in the fat body. Taken together, these results indicate that JAK/STAT, Pannier, and NF-{kappa}B signaling are involved in the regulation of lectin-24A and, more generally, Drosophila humoral anti-parasitoid immunity after infection.

immunology↗

bioRxiv2024

Dissecting reprogramming heterogeneity at single-cell resolution using scTF-seq

Reprogramming approaches often produce heterogeneous cell fates and the mechanisms behind this heterogeneity are not well-understood. To address this gap, we developed scTF-seq, a technique inducing single-cell barcoded and doxycycline-inducible TF overexpression while quantifying TF dose-dependent transcriptomic changes. Applied to mouse embryonic multipotent stromal cells (MSCs), scTF-seq produced a gain-of-function atlas for 384 murine TFs. This atlas offers a valuable resource for gene regulation and reprogramming research, identifying key TFs governing MSC lineage differentiation, cell cycle control, and their interplay. Leveraging the single-cell resolution, we dissected reprogramming heterogeneity along dose and pseudotime. We thereby revealed TF dose-dependent and stochastic cell fate branching, unveiling gene expression signatures that enhance our understanding and prediction of reprogramming efficiency. scTF-seq also allowed us to classify TFs into four sensitivity classes based on dose response and determining features. Finally, in combinatorial scTF-seq, we observed that the same TF can exhibit both synergistic and antagonistic effects on another TF depending on its dose. In summary, scTF-seq provides a powerful tool for gaining mechanistic insights into how TFs determine cell states, while offering novel perspectives for cellular engineering strategies.

cell biology↗

bioRxiv2023

Dissection and reconstruction of the colorectal cancer tumor microenvironment

Patient-derived organoids (PDOs) are the reference in vitro human disease models. However, the utility of colorectal cancer (CRC) PDOs is hindered by the lack of a tumor microenvironment (TME). To address this limitation, we built a living biobank of CRC PDOs with autologous stromal and immune TME. We characterized the original tumors and traditional monocultures using single-cell RNA-seq (scRNA-seq) and whole exome sequencing (WES) to obtain insights into cell type selection and phenotypic drift in culture. Subsequently, we developed culture conditions supporting all cell types to recapitulate the CRC-TME around PDOs. From the transcriptomes of >180k cells obtained from 260 such co-cultures, we illuminated the mutual influence of cells within CRC tumors. Based on original tumor data, atlases of predicted interactions and transcriptional networks elucidated why monocultures were altered and suggested that TME reconstruction more accurately reflected original tumor behavior. We found that inflammatory signals were absent in vitro and recovered upon co-culture with tumor-infiltrating lymphocytes (TILs). We also functionally confirmed that stromal, not cancer cells, mediated immune evasion. Additionally, stroma induced an invasive phenotype in cancer cells. From this deep dive into CRC-TME interactions, we built the human CRC-TME atlas (https://crc-tme.com/), an online portal for interactive exploration of gene expression data, prediction of cell-cell interactions at the pathway and receptor/ligand levels, transcriptional networks, and more. We anticipate PDO cultures with reconstructed TMEs will be valuable for discovery efforts, preclinical studies, and personalized medicine, with the atlas as a framework and inspiration for future CRC-TME studies.

cancer biology↗