Cancer-associated fibroblasts drive metabolic heterogeneity in colorectal cancer cells: predictions from metabolic modeling

npj Systems Biology and Applications · 2026-03-04

KRAS-mutant colorectal cancer (CRC) undergoes metabolic reprogramming that promotes tumor progression and drug resistance. Cancer-associated fibroblasts (CAFs), a major component of the tumor microenvironment (TME), play a pivotal role in modulating these metabolic adaptations in CRC. This study applies flux sampling combined with representation learning and hierarchical clustering to a computational model of central carbon metabolism to understand how CAFs influence KRAS-mutant CRC metabolic reprogramming following targeted enzyme knockdowns. Focusing on 12 key enzymes involved in glycolysis and the pentose phosphate pathway, knockdowns were simulated under normal CRC media and CAF-conditioned media (CCM) conditions. Analysis revealed CCM induces greater metabolic heterogeneity, with knockdown models exhibiting more variable and distinct metabolic states compared to those cultured in normal CRC media, indicating CAF-derived factors diversify the metabolic responses of CRC cells to enzyme perturbations. Pathway-level flux analysis demonstrated media-specific shifts in central carbon metabolism. Predicted biomass flux showed enzyme knockdowns reduced growth across both conditions, but CCM models indicated a protective effect against perturbation. Overall, simulations illustrated CCM enhances the metabolic adaptability of KRAS-mutant CRC cells to perturbations, emphasizing the importance of including TME components in metabolic modeling and therapeutic development and suggesting that targeting tumor-CAF metabolic interactions may improve treatment strategies.

Merging metabolic modeling and imaging for screening therapeutic targets in colorectal cancer

npj Systems Biology and Applications · 2025-01-28 · 11 citations

Cancer-associated fibroblasts (CAFs) play a key role in metabolic reprogramming and are well-established contributors to drug resistance in colorectal cancer (CRC). To exploit this metabolic crosstalk, we integrated a systems biology approach that identified key metabolic targets in a data-driven method and validated them experimentally. This process involved a novel machine learning-based method to computationally screen, in a high-throughput manner, the effects of enzyme perturbations predicted by a computational model of CRC metabolism. This approach reveals the network-wide effects of metabolic perturbations. Our results highlighted hexokinase (HK) as a crucial target, which subsequently became our focus for experimental validation using patient-derived tumor organoids (PDTOs). Through metabolic imaging and viability assays, we found that PDTOs cultured in CAF-conditioned media exhibited increased sensitivity to HK inhibition, confirming the model predictions. Our approach emphasizes the critical role of integrating computational and experimental techniques in exploring and exploiting CRC-CAF crosstalk.

Prolactin receptor localization and dynamics: Insights from quantitative imaging and mathematical modeling

bioRxiv (Cold Spring Harbor Laboratory) · 2025-10-14

Abstract Signal transduction through the prolactin receptor (PRLR) is crucial in pancreatic β-cell pro-liferation, impacting pancreatic homeostasis. PRLR-induced JAK/STAT signaling is dynamic, involving changes in spatial organization of signaling molecules. Thus, the spatial organization of PRLR could have strong implications on signaling output. Internalization has been shown and modeled in other signaling pathways but has not been considered in a mathematical model of PRLR signaling. Here, we use live-cell fluorescence imaging, reconstitution approaches, and fluorescence correlation spectroscopy (FCS) to inform a mathematical model of PRLR signaling. Internal PRLR localization is observed in primary pancreatic tissue and in an engineered PRLR expression system. Our imaging data indicate the presence of intracellular and plasma membrane-bound receptor populations. We use FCS to resolve the membrane-bound PRLR population. Based on our data, we include internalization dynamics within an ordinary differential equation (ODE) model of PRLR signaling. We employ the model to explore how the spatial heterogeneity of PRLR affects downstream signaling. We show that the model is more sensitive to PRLR trafficking rates and ability to promote signaling than to its initial spatial distribution. Our data underscore the versatility of a modeling-imaging framework to quantitatively understand signal transduction in and beyond β-cells. Significance Statement Prolactin receptor (PRLR) signal transduction impacts the growth and survival of insulin-secreting cells, making this pathway a target for building our understanding of pancreatic homeostasis and exploring potential diabetes therapeutics. Live fluorescence imaging techniques applied within an engineered PRLR expression platform indicate PRLR localization patterns consistent with primary pancreatic tissue and the presence of two spatially distinct PRLR populations. These observations inform a predictive mathematical model of PRLR signaling. Integrating experimental data tailored to computational approaches shapes our understanding of complex, multiscale systems such as signal transduction. A generalizable modeling-imaging framework enables the study of molecular dynamics beyond β-cells.

Cancer-associated fibroblasts drive metabolic heterogeneity in KRAS-mutant colorectal cancer cells

bioRxiv (Cold Spring Harbor Laboratory) · 2025-10-01

Abstract KRAS-mutant colorectal cancer (CRC) is characterized by metabolic reprogramming that can lead to tumor progression and drug resistance. The tumor microenvironment (TME) plays a pivotal role in modulating these metabolic adaptations. In particular, cancer-associated fibroblasts (CAFs), which make up a large portion of the TME, have been shown to strongly contribute to metabolic reprogramming in CRC. This study applies flux sampling, a computational method that explores the full range of feasible metabolic states, combined with representation learning and hierarchical clustering, to a computational model of central carbon metabolism to understand how CAFs influence metabolic adaptations of KRAS-mutant CRC cells following targeted enzyme knockdowns. Focusing on twelve key enzymes involved in glycolysis and the pentose phosphate pathway, knockdowns were simulated under both normal CRC media and CAF-conditioned media (CCM) conditions. Analysis revealed that CCM induces greater metabolic heterogeneity, with knockdown models exhibiting more variable and distinct metabolic states compared to those cultured in normal CRC media. While some enzyme knockdowns produced similar metabolic states, this overlap was less frequent in CCM, indicating that CAF-derived factors diversify the metabolic responses of CRC cells to enzyme perturbations. Pathway-level flux analysis demonstrated media-specific shifts in central carbon metabolism pathways. Importantly, the predicted biomass flux showed that enzyme knockdowns reduced growth across both conditions, but models in the CCM condition indicated CAFs could offer a protective effect against metabolic perturbation. Overall, this study reveals that CCM significantly influences the metabolic state and adaptability of KRAS-mutant CRC cells to enzyme perturbations, emphasizing the importance of including TME components in metabolic modeling and therapeutic development. These findings provide valuable insights into the metabolic adaptability of CRC and suggest that targeting tumor-CAF metabolic interactions may improve treatment strategies. Graphical Abstract Overview of computational workflow Models of interest represent simulated enzyme knockdowns in central carbon metabolism. Flux sampling searches the entire metabolic solution space and results in a distribution of flux values for each reaction within each model. Samples can be organized by knockdown and condition into matrices for input into representation learning. Representation learning is applied to sampling data to identify shared and independent metabolic states. Metabolic states indicate a heterogeneous response to enzyme knockdowns. Overlap of dark and light blue flux distributions, sampling clusters, and metabolic responses exemplify a shared metabolic state separate from to the gray unperturbed state. This workflow provides a low-dimensional representation of metabolic state that captures both the pathway- and reaction-level differences that describe each simulated knockdown.

Impact of expression of metabolic genes on patient (pts) outcomes in metastatic colorectal cancer (mCRC): Data from CALGB/SWOG 80405 (Alliance).

Journal of Clinical Oncology · 2025-01-27

224 Background: Metabolic pathways are reprogrammed during CRC development, leading to increased glycolysis, glutaminolysis, and fatty acid synthesis. Understanding the impact of metabolic reprogramming on treatment outcomes is paramount. Hence, we investigated whether the tumor gene expression of 4 selected major metabolic genes ( PKM, SLC2A1 , SLC16A1 , and CAV1 ) could affect treatment response in pts enrolled in the CALGB/SWOG 80405 trial. Methods: 433 mCRC pts treated with either bevacizumab (bev, n = 226) or cetuximab (cet, n = 207) in combination with first-line chemotherapy were analyzed. RNA was isolated from FFPE tumor samples and sequenced on the HiSeq 2500 (Illumina). Overall survival (OS) and progression-free survival (PFS) were compared between groups of pts categorized by tertiles of gene expression (high [H], medium [M] and low [L]). Logrank P -values provide a non-parametric unadjusted assessment of differences. Likelihood ratio tests (LRT) were computed from multivariate Cox proportional hazards models, adjusting for age, sex, ECOG, tumor side, number of metastatic sites, KRAS , CMS, and treatment. Results: PKM gene expression was associated with cet treatment outcomes with PKM- low tumors showing significantly longer PFS and OS (median PFS: 14.3 vs 9.8 vs 8.1 months, L vs M vs H, P < 0.0001, LRT P = 0.011; median OS: 45.9 vs 31.2 vs 20.9 months, P < 0.0001, LRT P = 0.12). Low SLC2A1 was also associated with longer OS in cet-treated pts (32.4 vs 35.8 vs 25.2 months, P = 0.02, LRT P < 0.0001). Similar results were observed in bev-treated pts where low SLC2A1 expressing tumors had longer OS (37.4 vs 26.1 vs 29 months, P = 0.037, LRT P = 0.069) and a non-significant trend for longer PFS (13.1 vs 11 vs 9.5 months, P = 0.076). CAV1 low gene expression was associated with longer OS in cet-treated pts (37.4 vs 34 vs 21.5 months, P = 0.0048, LRT P = 0.013), while no significant associations were found in bev-treated (interaction LRT P = 0.0009). No significant results were observed for SLC16A1 . Conclusions: The intratumoral expression of metabolic genes was prognostic and predictive in mCRC pts treated with first-line therapy. GLUT1 (encoded by SLC2A1 ) and PKM2 (encoded by PKM ) promote aerobic glycolysis of cancer cells, known as Warburg effect, supporting rapid cell proliferation and survival. CAV1 has been shown to be involved in mitochondrial bioenergetics and fatty acid metabolism and to play either a tumor suppressor or oncogenic role depending on the cancer type and stage. Lower expression of SLC2A1 and PKM was associated with improved survival and increased benefit from cet treatment in our cohort. Low CAV1 expression was also associated with increase survival suggesting a cancer-promoting role in mCRC. Our results suggest that targeting cancer cell metabolism through novel inhibitors of the aforementioned pathways may be a promising therapeutic strategy.

Metabolic Reprogramming and Therapeutic Vulnerabilities in the Tumor Microenvironment Revealed by Multi-scale Network Geometry

bioRxiv (Cold Spring Harbor Laboratory) · 2025-11-17

Abstract The tumor microenvironment comprises diverse cell populations that coordinate metabolic activities to sustain malignant growth, yet the systems-level organization of these interactions remains poorly understood. Here, we present an integrated framework combining single-cell transcriptomics, genome-scale metabolic modeling, and multi-scale network geometry to decode metabolic coordination in colorectal cancer. We demonstrate that FAP + cancer-associated fibroblasts and MARCO + tumor-associated macrophages undergo extensive reprogramming, establishing metabolic division of labor: fibroblasts specialize in amino acid and fatty acid metabolism while macrophages adopt cancer-like nucleotide biosynthesis programs. Systematic knockout analysis identified 19 tumor-selective vulnerabilities in branched-chain amino acid catabolism, with MAOB validated as a prognostic marker through patient survival analysis. To reveal architectural organization, we applied multifractal geometric characterization and Ollivier-Ricci curvature analysis for the first time to flux-weighted metabolic networks derived from context-specific genome-scale models. While conventional network metrics failed to distinguish tumor from normal phenotypes, multifractal analysis successfully separated tissue states through coordinated architectural changes across hierarchical scales. Role transition analysis revealed that 20–25% of metabolites undergo functional reorganization, with prostaglandin and bile acid derivatives emerging as critical communication hubs between stromal populations. Curvature analysis identified pathway-specific geometric remodeling in fatty acid metabolism (fibroblasts) and leukotriene metabolism (macrophages). Our findings establish that metabolic adaptation represents ecosystem-level network reorganization rather than isolated pathway changes, providing a generalizable framework for identifying therapeutic strategies targeting cooperative metabolic networks.

Multiscale Probabilistic Modeling: A Bayesian Approach to Augment Mechanistic Models of Cell Signaling with Machine-Learning Predictions of Binding Affinity

bioRxiv (Cold Spring Harbor Laboratory) · 2025-05-23

Abstract Computational models in systems biology are often underdetermined—that is, there is little data relative to the complexity and size of the model. The lack of data is primarily due to limits in our ability to observe specific biological systems and restricts the utility of computational models. However, there are a growing number of experimental databases in biology. While these databases provide more observations, they often do not have observations that match the system of interest exactly. For example, database measurements might be collected at different experimental conditions or on a different scale compared to the system of interest. Here, we investigate what information can be gleaned from generalizing databases across these differences in the context of modeling a specific system – cell signaling. Ultimately, our goal is to better determine models of specific systems, thereby increasing their utility. To do this, we propose a novel, multiscale, probabilistic framework. We use this framework to integrate measurements of protein structure from the Protein Data Bank and measurements of amino acid sequence from the Universal Protein Resource into the parameter inference of cell signaling models. Then, we quantify exactly what information is gained from these measurements when modeling cell signaling. We choose to investigate the utility of these databases in the context of dynamic cell signaling models because experimental measurements of the variables of interest, protein dynamics, are still quite limited. We find that we can successfully integrate measurements from these databases to significantly improve parameter estimation of signaling models. The impact of sequence and structure measurements on model predictions depends on the sensitivity of the prediction to perturbations in the parameter values. Overall, this study demonstrates that measurements of protein structure and amino acid sequence can be leveraged to better inform parameters in models of cell signaling. Author Summary Computational models of cell signaling have provided mechanistic insights into complex biological systems, including in physiological and disease settings. Accurate and predictive modeling critically depends on the precise estimation of model parameters, which is often hindered by the limited availability of experimental data. In this study, we present a novel multiscale probabilistic inference framework that broadens the scope of data types that can be leveraged for parameter estimation for models of cell signaling. The framework integrates a machine learning pipeline with a generalizable parameter inference approach, enabling the use of experimental data across scales. Specifically, we demonstrate that incorporating protein amino acid sequence and 3D structural data enhances parameter estimation compared to traditional measurements such as protein concentrations over time. Improving parameter estimation increases the robustness and applicability of cell signaling models. Ultimately, our framework facilitates use of a broader range of data and supports the development of predictive computational models that increase our understanding of cell signaling.

BPS2025 - A tale of trafficking: On prolactin receptor localization in pancreatic β-cells

Biophysical Journal · 2025-02-01

Clinical and molecular characterization of <i>FAP</i> and <i>SPP1</i> in colorectal cancer (CRC), CALGB (Alliance)/SWOG 80405 and real-world data.

Journal of Clinical Oncology · 2025-01-27

276 Background: FAP and SPP1 contribute to immune modulation in the CRC tumor microenvironment (TME). FAP is expressed by cancer associated fibroblasts, aiding in tissue remodeling and tumor invasion. SPP1 is an integrin-binding protein expressed by tumor associated macrophages that promotes tumor growth, adhesion, and metastasis. We present a clinical and molecular characterization of FAP and SPP1 in CRC. Methods: We analyzed 24,257 CRC samples tested at Caris Life Sciences (Phoenix, AZ) with WTS (Illumina NovaSeq), NextGen DNA sequencing (NextSeq), and PD-L1 expression (SP142, positive ≥ 2+, 5%). RNA deconvolution analysis estimated cell infiltration in the TME. Data from the phase 3 CALGB/SWOG 80405 trial (NCT00265850) on 433 metastatic CRC patients treated with bevacizumab (Bev, n = 226) or cetuximab (Cet, n = 207) in combination with first-line chemotherapy were also evaluated. RNA isolated from FFPE tumor samples were sequenced with HiSeq 2500 (Illumina). Overall survival (OS) and progression-free survival (PFS) were compared using Cox regression in categorical gene expression tertiles (high (T3), medium (T2), and low (T1)) for FAP and SPP1 . Results: FAP -T3 and SPP1 -T3 had increased PD-L1 positivity (q&lt;0.05), while SPP1 -T3 also demonstrated increased MSI-H (8.4% vs 5.3%) and TMB-H(&gt;10 mt/mb, 15.0% vs 10.1%) status relative SPP1 -T1 (all q&lt;0.001). T3 of both genes of correlated with increased M1/M2 macrophages, NK cells and T cell inflamed score (q&lt;0.001); dendritic cells and neutrophils were increased in SPP1 -T3 and FAP -T1 tumors (q&lt;0.05). FAP -T3 and SPP1 -T3 had increased pathway activation of epithelial-mesenchymal transition (EMT), inflammatory response, TNF-a signaling, angiogenesis and KRAS signaling (all q &lt; .005). In Caris cohorts, FAP -T3 demonstrated worse OS in Cet/panitumumab treated CRC (T3: 23.6 vs T1: 21.0 months [mo], P = .005; HR 0.85, 95% CI [0.77-0.95]), but SPP1 did not correlate with OS. In 80405, FAP -T3 showed shorter PFS (T3: 9.5 vs T2: 11.5 vs T1: 12.6 mo, T3 vs T1 (reference) adjusted HR 1.27 [1.07-1.51]) and OS (25.2 vs 29.4 vs 35.5 mo, adjusted HR 1.31 [1.10-1.57]). Similarly, SPP1 -T3 demonstrated worse PFS (9.0 vs 12.7 vs 14.0 mo, adjusted HR 1.29 [1.12-1.48]) and OS (20.9 vs 34.0 vs 36.3 mo, HR 1.24 [1.07-1.44], P &lt; 0.001). Treatment interaction tests noted FAP -T1 CRC benefited from Cet over Bev with respect to PFS ( P = 0.003) and OS ( P = 0.044), while SPP1-T1 tumors demonstrated a PFS benefit with Cet ( P = 0.009). Conclusions: Our results indicate that increased FAP and SPP1 expression is associated with immune cell infiltration, EMT, and inflammatory signaling. Additionally, their expression may be prognostic and predictive of targeted therapy. These data support the evaluation of FAP and SPP1 as predictive markers and therapeutic targets in CRC.