Acquired resistance to BRAF inhibitory treatments requires tumor tissue remodeling and reveals targetable vulnerabilities in colorectal cancer

Acquired resistance to BRAF inhibitory treatments requires tumor tissue remodeling and reveals targetable vulnerabilities in colorectal cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Acquired resistance to BRAF inhibitory treatments requires tumor tissue remodeling and reveals targetable vulnerabilities in colorectal cancer Hector Palmer, Lorena Ramírez, Javier Ros, Oriol Arqués, Jordi Martinez-Quintanilla, and 25 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6770441/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The combination of BRAF V600E and EGFR inhibitors benefits advanced colorectal cancer (CRC) patients, but resistance can develop through non-genetic mechanisms not fully understood. This study reveals that BRAF mutant CRC tumors undergo a tissue remodeling to resist BRAF V600E inhibitors, characterized by tumor cell mucinous differentiation and extracellular matrix transformation, allowing an increased infiltration of activated fibroblasts, immune cells, and vasculature. Some of these changes are essential for acquiring resistance. In fact, we demonstrate that blocking new vascularization with the anti-angiogenic antibody bevacizumab against VEGFA extends the benefit of BRAF inhibitory therapies in CRC models. These findings are based on patient-derived xenograft (PDX) models and validated in patient samples, offering deeper insights into resistance mechanisms and suggesting rational combinations to prolong therapy effectiveness. A clinical trial is being initiated to test the combination of bevacizumab and BRAF inhibitory therapy for improving CRC patients’ treatment. Health sciences/Oncology/Cancer/Cancer therapy/Cancer therapeutic resistance Health sciences/Oncology BRAFV600E acquired drug resistance encorafenib tissue remodeling bevacizumab colorectal cancer Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Main Colorectal cancer (CRC) is a significant global health concern, with approximately 1.9 million new cases and 935,173 deaths in 2020( 1 ). About 10% of CRC patients harbor the BRAF V600E mutation( 2 , 3 ), linked to poor prognosis and limited treatment options( 2 , 4 – 7 ). This mutation, resulting from a thymine-to-adenine change( 2 , 4 , 5 ), leads to constitutive BRAF activation and downstream MAPK signaling, promoting tumor proliferation and survival( 8 ). While BRAFi inhibitors (BRAFi) monotherapies show efficacy in melanoma and other tumors( 9 ), their effectiveness in CRC is hindered by rapid feedback activation of EGFR( 10 , 11 ). Clinical trials, including BEACON, demonstrated that combining EGFR inhibitors (EGFRi) with BRAFi with or without MEK inhibitors (MEKi) improves outcomes in metastatic CRC (mCRC)( 9 , 12 ). However, resistance to these combinations often develops, necessitating an understanding of its mechanisms. Known resistance mechanisms can involve genetic alterations that reactivate the MAPK pathway, including KRAS and MAP2K1 mutations, BRAF and MET amplifications and others( 13 – 21 ). However, approximately 40% of resistant cases lack these oncogenetic changes( 14 , 22 ), indicating the relevance of non-genetic mechanisms of resistance. Our study utilizing patient-derived xenografts (PDX) reveals that resistance to combined BRAFi therapies can arise from non-genetic factors linked to acute tissue remodeling, such as mucinous differentiation, extracellular matrix (ECM) remodeling, increased immune stroma( 23 ) and vasculature. We observed that blocking angiogenesis with bevacizumab in PDX models extended combined BRAFi therapy efficacy, delaying resistance and enhancing progression-free survival (PFS). This underscores the importance of understanding resistance mechanisms, as patients with limited options after BRAF-targeted therapies need new strategies. Our findings provide insights into the processes underlying resistance to combined BRAFi therapies in mCRC and support the initiation of BRAVE, a phase I/II clinical trial (NCT06411600) to evaluate bevacizumab in combination with encorafenib (BRAFi) and cetuximab (EGFRi). In summary, we reveal that BRAF V600E mCRC can develop resistance through cell plasticity and tissue remodeling, suggesting new rational combinations for improving treatment outcomes in advanced CRC patients. Results PDX models effectively mimic the development of resistance to combined BRAFi therapy in BRAF V600E mCRC All treatments with a BRAFi in combination with other drugs such as EGFRi, MEKi or others were indicated as BRAFi therapies for simplification purposes, since all have a common mutant BRAF inhibition strategy. We established 32 PDX models from 22 patients, including 17 derived from patients naive (PDX-N) and 15 resistant (PDX-R) upon progression to combined BRAFi therapies (three of them were derived from paired biopsies) (Fig. 1 A and Supplementary Table S1 ). Whole Exome Sequencing (WES) revealed that two resistant patients exhibited oncogenic subclonal KRAS mutations, a known resistance mechanism( 13 – 20 ) (Supplementary Table S1 ). The mutational landscape of our PDX models was consistent with TCGA BRAF V600E CRC tumors, featuring mutations in APC , TTN , and TP53 (Supplementary Fig. S1 A and Supplementary Table S2), but lacked exclusive oncogenic mutations distinguishing resistant from naive models. We only derived one PDX model with microsatellite instability (MSI) from a patient who progressed to combined BRAFi therapy because they were preferably enrolled in immunotherapy treatments (Supplementary Table S1 ). The analysis of single base substitution (SBS) COSMIC signatures also indicated similar profiles between naive and resistant models and a low incidence of mismatch repair deficiency (MMRd) signatures and MSI due to priority selection for immunotherapy treatment (Supplementary Fig. S1 B). Our sequencing results indicate that our models recapitulate human CRC genetics (TCGA-COREAD) and suggest that resistance to BRAF inhibitors could be driven by non-genetic mechanisms rather than genetic alterations. To explore this, we selected naive (CTAX2-N, M64-N) and resistant (CTAX90-R, CTAX91-R) PDX models. Aiming to use a balanced set of models we considered that all were derived from patients treated with encorafenib + cetuximab backbone and benefited more than 120 days (Supplementary Table 1). We also balanced the model selection by using one MSI and one with microsatellite stability (MSS) for each naïve or resistant. For our initial discovery phase, we combined encorafenib (BRAFi), cetuximab (EGFRi) and binimetinib (MEKi) in most of our experiments which is indicated as triplet treatment (Fig. 1 B). While the standard treatment for patients in Europe and USA is a BRAFi + EGFRi doublet, our results with PDX indicated that only a triple combination adding a MEKi produced an initial prolonged response phase followed by a progression whereas xenografts treated with the doublet although slowed proliferation, progressed early as non-treated control animals (Kapplan Meier plots) (Supplementary Fig. S2A). In fact, the triplet has been adopted as the Standard-of-Care (SoC) in Japan because of its higher overall response rate in BRAF V600E CRC patients( 9 , 20 , 24 ) (Supplementary Fig. S2B). Importantly, the most significant discoveries in our study using the triplet were also observed using doublet treatments in PDX and patients. Both naive models initially showed robust responses but subsequently acquired resistance (Fig. 1 B). Importantly, in both naive models, the treatment blocked tumor growth for the initial 20–35 days. In contrast, treatment delayed the growth rate of resistant PDXs, but all tumors progressed soon after treatment initiation without showing a stabilized disease phase. Interestingly, the triplet treatment equally blocked the MAPK signaling in tumors of both naive and resistant PDX models, as observed by the inhibition of pathway target genes DUSP6 and SPRY4 expression (Supplementary Fig. S2C). We observed that both naive PDX tumors in different mice treated with the triplet acquired resistance almost simultaneously after a response phase (Supplementary Fig. S2D) and WES did not show the appearance of any oncogenic mutation that could explain such resistance (Supplementary Fig. S2E and Supplementary Table S3). In summary our results suggested that our PDX models recapitulate patients´s genomics and their response to treatment and that non-genetic mechanisms could drive the acquisition of resistance to BRAFi therapies in mCRC instead of the outgrowth of a selected genetic clone. MAPK oncogenic signal blockade induced significant molecular reprogramming in BRAF V600E CRC tumor cells and surrounding stroma We analyzed whole gene expression in naive and resistant PDX models treated with a triplet therapy at various time points (Fig. 1 C and Supplementary Fig. S3A, Supplementary Table S4). Treatment inhibited proliferation and triggered distinct responses: naive models exhibited rapid interferon-based immune activation( 23 ), while resistant models engaged in drug detoxification (Supplementary Fig. S3A and S3B). Gene set enrichment analysis (GSEA) indicated positive enrichment of hypoxia, vasculature, mucinous differentiation and tissue remodeling pathways in naive models, while proliferation-related pathways were negatively impacted in both (Fig. 1 D, 1 E, Supplementary Fig. S3B, Supplementary Table S5). Interestingly, we observed that this switch was deeper in naive than in resistant models when acquired resistance (end points), as observed by the appearance of a larger variety of biological processes (gene sets) positively enriched (Fig. 1 D). Aiming to mine deeper in understanding this non-genetic process of tissue plasticity we selected naïve M64-N model as the most sensitive to triplet treatment for further detailed investigations. Subsequent proteomic analysis of M64-N confirmed the induction of proteins related to mucinous differentiation, immune response, vasculature, RNA processing, tissue remodeling or EGFR signaling blockade, alongside repression of protein translation and proliferation-associated proteins (Fig. 2 A-D and Supplementary Table S6 and S7). BRAF V600E oncogenic signaling was previously described to control gene expression by modulating genome methylation( 25 ). Methylome analysis showed progressive reprogramming of DNA methylation upon triplet treatment in all models, with hypomethylation occurring earlier in naive than resistant models, particularly in regions controlling gene expression (Fig. 2 E-G, Supplementary Fig. S4A-D and Supplementary Table S8). This hypomethylation correlated with increased gene expression prior to resistance acquisition (Fig. 2 G-I and Supplementary Fig. S4C, S4D and Supplementary Table S9). All our samples presented a general methylome profile distinctive of TCGA-COREAD (Supplementary Fig. S4A). MUC20 gene is a clear example, where specific CpG sites in its promoter were hypomethylated and showed increased expression upon triplet treatment (Fig. 2 J). BRAF inhibition in CRC can promote a cancer cell intrinsic biological change that is heterogenous among cells within the same tumor( 26 ). For reveling this potential complexity, we performed single-cell RNA sequencing (scRNAseq) in the most triplet sensitive naive model M64-N at endpoint treatment. We revealed marked changes in gene expression profiles among tumor and stroma cells following triplet treatment involving an increased infiltration of immune cells, fibroblasts and endothelial cells confirming significant tissue remodeling (Fig. 3 A-C). We observed the modulation of specific genes by the triplet in tumor cells observed by bulk analyses, such as the repression of DUSP6 as MAPK target gene or induction of metalloprotease MMP7 involved in tissue remodeling and MUC4 in mucinous differentiation (Fig. 3 D). Moreover, treatment led to an increase of stem cell marker PROM1 /CD133 which has been described to stabilize the angiogenic ligand VEGFA( 27 ) (Fig. 3 D and 3 E). We observed increase in the PROM1 /CD133 + cancer stem cell population( 28 , 29 ) and in mucinous differentiated cells (MUC4) (Fig. 3 D-G), indicating a complex interplay between progenitor and differentiated tumor cells during resistance development. An independent laboratory performed spatial transcriptomics in two extra BRAF V600E CRC PDX models treated with dabrafenib (BRAFi), cetuximab (EGFRi) and trametinib (MEKi). We observed a nice separation of treated versus not treated areas by UMAP analyses and a clear enrichment of infiltrated mouse stroma upon triplet treatment (Data not shown). Interestingly, markers of mucinous differentiation ( MUC4 , CLCA ), CAF infiltration ( Acta2 ), tissue remodeling ( COL1A1 ) and others relevant for the drug-induced phenotype such as PROM1 were increased in the tissue upon triplet treatment. Again, VEGFA was not regulated at mRNA level, whereas others such as OLFM4 was again induced by the triplet. In conclusion, our results suggested that MAPK blockade regardless of the specific drugs (dabrafenib vs encorafenib, binimetinib vs trametinib), promoted a multifaceted molecular shift in BRAF V600E CRC, fostering an environment conducive to resistance through epigenetic reprogramming, tumor cell mucinous differentiation and stroma remodeling. BRAF inhibitory therapy promotes a mucinous differentiation of tumor cells that is required for acquiring resistance to treatment This was confirmed in PDX models, where a significant increase in mucus production was observed macroscopically after treatment, and driven by transcription factor ATOH1 and mucins upregulation ( MUC1 , MUC2 , MUC4 , MUC13 ) (Fig. 4 A-C, Supplementary Fig. S5A, S5B and Supplementary Table S1 0). ATOH1 is a transcription factor key for directing intestinal stem cells towards goblet cell terminal differentiation (Supplementary Fig. S5C). This differentiation led to the accumulation of goblet-like cells that secreted mucins (alcian blue), measured as MUC2-positive cells (Fig. 4 D-G and Supplementary Fig. S6A), being MUC2 a direct ATOH1 target gene( 30 – 32 ). In collaboration, another independent team observed in parallel the same progressive mucinous differentiation phenotype in seven BRAF V600E PDX models (Supplementary Fig. S6B-D). Interestingly, this blind validation involved different MAPK inhibitory drugs: dabrafenib (BRAFi), trametinib (MEKi), and cetuximab (EGFRi). Futhermore, increased mucin levels were detected in five out of ten patients when comparing baseline biopsies to those taken upon disease progression under BRAFi-combined therapy (Fig. 4 H, 4 I and Supplementary Table S1 ). While cetuximab resistance has been associated with Paneth cell differentiation in CRC( 33 ), gene expression profiling ruled out this lineage differentiation in triplet-treated PDX models (Supplementary Fig. S6E and Supplementary Table S1 1). Knockout experiments targeting mucin maturation enzymes ( GCNT1/GCNT3 ) in HT29 CRC cell line (Supplementary Fig. S6F), revealed that mucin maturation driven by triplet treatment, was required for resistance acquisition (Fig. 4 J). We observed that double-KO cells grew subcutaneously in NOD-SCID immunodeficient mice as rapidly as their wild type (WT) counterparts. However, any double-KO tumor was capable to progress and acquire resistance to triplet treatment as WT did (Fig. 4 J and Supplementary Fig. S6G and S6H). Double-KO tumors still commit to mucinous differentiation and produce immature mucins (alcian blue) (Fig. 4 K), however a reduction of proliferation upon triplet treatment was observed in KO but not in WT tumors (Fig. 4 K and Supplementary Fig. S6I). In conclusion, we discovered a drastic increase in mucin production upon triplet treatment and mucin maturation as a key factor for acquiring resistance to BRAFi therapies. MAPK inhibition induces extracellular matrix (ECM) remodeling We observed an enrichment of tissue remodeling genes, such as metalloproteases ( MMP7 ), cathepsins ( CTSS ), and complement components ( CFB , C1S , SERPING1 ), in naive, but less in resistant PDX models that was more acute at long term treatment before acquisition of resistance (end point) (Fig. 5 A, 5 B and Supplementary Table S12 and S13). MMP7 expression was notably elevated in naive models by triplet treatment (Fig. 5 B-E). MMP7 mRNA and protein were particularly induced by triplet in naive but not in resistant PDX models, which already presented higher basal levels (Fig. 5 C-E and Supplementary Fig. S7A and S7B). This molecular shift also increased collagen deposition (Fig. 5 C-J) and cancer associated fibroblasts (CAFs) infiltration, contributing to a distinctive ECM phenotype (Supplementary Fig. S7C and S7D). We could conclude that triplet treatment promoted an early ECM remodeling that was consolidated on persistent small tumors before restart tumor growth. Since this is a complex process involving a coordinated network of factors, we could not perform an intervention experiment to prove its requirement for resistance acquisition. Indeed, we knocked down MMP7 in HT29 cells and it was not enough to block the acquisition of resistance to triplet treatment (data not shown). Further analyses were preformed to understand tumor cell signaling consequences upon MAPK signaling blockaded. We confirmed that triplet therapy blocked MAPK signaling reducing phospho-ERK. We also observed a decrease of GSK-3β phosphorylation at Tyr216, a target site of MEK1 kinase, required for its activity ( 34 ) (Supplementary Fig. S8A). This reduction of active phosphorylated GSK-3β decreased in turn β-catenin phosphorylation and degradation, leading to its nuclear accumulation in cells and tumors, and activation of Wnt pathway target genes, including MMP7 and OLFM4 (Supplementary Fig. S8A-D). β-catenin is the major effector of the Wnt pathway, a central force in CRC tumorigenesis 34–37 . Tankyrase (TNK) inhibition reduced Wnt/β-catenin signaling( 37 ), but genes like DUSP6 , directly regulated by MAPK signaling, were unaffected by Wnt inhibition. Similarly, PROM1 was induced by triplet but unaffected by TNKi treatment. This suggests that BRAFi therapy triggers a signaling shift that suppresses MAPK while enhancing Wnt/β-catenin signaling, orchestrating a precise gene expression profile that determines tissue architecture remodeling during the acquisition of resistance in BRAF V600E CRC tumors. Combined BRAF inhibitory therapy increased tumor vascularization, which can be effectively blocked by bevacizumab to prolong progression-free survival (PFS) At this point of the study, despite our significant advances in understanding the resistance of mechanisms to BRAFi therapies in colorectal cancer (CRC), a strategy to improve patient outcomes remained elusive. The complex tumor tissue remodeling observed under BRAFi treatment posed challenges for direct pharmacological targeting. We thus shifted our focus to tumor vascularization, a process which genes were significantly enriched in triplet-treated naive PDX models (Fig. 1 D, 3 B-D). Indeed, gene sets associated with vascular development were particularly upregulated in these models following prolonged triplet treatment (Fig. 6 A and 6 B, Supplementary Table S14, S15). These included genes such as MSLN , PROM1 , TSPAN12 , ST6GAL1 , IL6 , SEMA5A , and PDGFA ; all involved in tumor angiogenesis via various mechanisms. Notably, PROM1 , encoding the stem cell marker CD133, was upregulated by triplet treatment (Fig. 6 B-D). CD133 has been implicated in stabilizing VEGFA( 27 ), a key driver of angiogenesis targeted by the monoclonal antibody bevacizumab, a standard treatment in CRC( 38 ). Although triplet therapy did not consistently increase total VEGFA mRNA or protein levels (Supplementary Fig. S9A-D), it did promote extracellular VEGFA accumulation in BRAF V600E CRC cell line organoid models (Fig. 6 E-F, Supplementary Fig. S9E-F). Further in vivo analyses showed that triplet treatment significantly elevated microvascular density (MVD) in naive PDX models, with resistant models already presenting higher baseline MVD (Fig. 7 A-D). Importantly, six out of ten CRC patients exhibited increased MVD after treatment with BRAFi therapies (Fig. 7 E and Supplementary Table S1 ). These findings highlighted tumor vascularization as a potential vulnerability that could be exploited therapeutically. Indeed, the combination of triplet therapy with bevacizumab effectively prolonged PFS in HT29 xenografts and naive PDX models (M64-N) (Fig. 7 F and 7 G). In the M64-N model, nearly all tumors treated with triplet and bevacizumab showed no progression after 50 days, in contrast to the lower efficacy of triplet alone (Fig. 7 G). Bevacizumab also demonstrated efficacy in reducing tumor growth after acquiring resistance to triplet. Moreover, the combination of triplet and bevacizumab modestly slowed the growth of resistant CTAX91-R xenografts (Fig. 7 G). In conclusion, increased MVD during tissue remodeling represents a targetable vulnerability in BRAFi-resistant CRC. These preclinical findings us led to the initiation of the BRAVE phase I/II clinical trial (NCT06411600), in which we are enrolling 94 BRAF V600E mCRC patients (BRAFi-naive) treated with cetuximab (EGFRi), encorafenib (BRAFi), and bevacizumab (VEGFAi). Considering our preclinical results with PDX models, the trial’s primary objective is PFS, with secondary endpoints including response rate, time to response, duration of response, and overall survival. 15 patients have been already enrolled and results are encouraging. DISCUSSION In the phase III BEACON trial, we observed that the combination of encorafenib (BRAFi), cetuximab (EGFRi), and/or binimetinib (MEKi) significantly improved overall survival and response rates in mCRC patients with the BRAF V600E mutation, compared to standard chemotherapy( 9 ). This study was the base for FDA and EMA approval of encorafenib and cetuximab combination as the new SoC for mCRC patients. However, resistance to these combinations occurred in almost all patients. Therefore, understanding the mechanisms ruling this acquisition of resistance remained a clinical challenge. In response our study has explored the non-genetic mechanisms of resistance to MAPK oncogenic signal inhibition by combining encorafenib, cetuximab, and binimetinib in mCRC. We observed tumor cell intrinsic reprogramming, including the previously described immune response, but also an unanticipated mucinous differentiation and massive mucin secretion. Importantly, it was the maturation of biological active mucins a critical determinant for resistance acquisition. Blocking this process could be a potential therapeutic strategy that was tested in a clinical trial using Bromelain as an unspecific mucolytic against advanced solid tumors (NCT02340845). Future drugs with more specific mode of action against mucin secretion or function could extend the benefit of advanced CRC patients progressing to BRAFi-based therapies. Some mucins can activate the oncogenic Wnt/b-catenin signaling that is a driving force of CRC malignancy. For instance, the MUC1 cytoplasmic domain (CD) has been shown to bind β-catenin( 39 ) blocking its phosphorylation by GSK3b, its proteasome degradation and final accumulation. We observed a coordinated increase of mucins and nuclear b-catenin upon GSK3b blockade. In addition, b-catenin promotes the expression of MUC4 ( 40 ) and other tissue remodeling genes such as MMP7 acting as a transcriptional coactivator. In fact, Wnt/b-catenin has also been considered responsible for the resistance of different cancer types to a variety of drugs( 41 , 42 ). In summary, our results indicate that a Mucin/Wnt/b-catenin axis loop could play a relevant role in the acquisition of resistance to BRAFi-based therapies in CRC and is an attractive therapeutic target. Additionally, detecting mucins in blood could improve monitoring response and early progression of CRC patients treated with combined BRAFi therapies as previously describe in mucinous ovarian or breast carcinomas( 43 , 44 ). The triplet therapy unexpectedly enhanced mucinous differentiation in tumor cells, typically associated with low proliferation. Nevertheless, mucinous tumors continued to grow and progress under triplet. Additionally, there was a concurrent increase in PROM1 /CD133 positive cells, indicating that this cancer stem cell-like population may feed tumor growth and progression( 28 , 29 ). ATOH1 upregulation could promote a rapid and constant transit of proliferating stem cell-like cells to mucinous goblet-like differentiated cells and sustain these populations balanced. Indeed, hypersecretory cancer cells producing mucins and collagens could be essential for building a biochemical fortress to protect minimal residual disease in tumors such as CRC and pancreas. Similar phenomenon has been described in lung cancer persistent tumors where minimal residual disease has shown heterogeneous cell profiles including alveolar differentiation and progenitor cells downstream of an activated Wnt/b-catenin pathway ( 45 , 46 ). BRAFi treatments also promoted early stroma remodeling, including deposition of a distinct ECM with specific collagens and metalloproteases, changing tumor tissue stiffness and may creating low proliferation niches ( 47 ). These ECM changes could also facilitate infiltration and activation of cancer-associated fibroblasts (CAFs), known promoters of tumor progression ( 48 ). Key factors essential for these processes that are required for resistance, could also serve as valuable targets to prolong the effectiveness of BRAFi therapies. Further investigations are needed to identify key drug targets relevant in these ECM remodeling. BRAFi treatments in BRAF V600E mCRC induced a tumor cell persistence based in cancer cell intrinsic rewiring/plasticity and complex stroma remodeling, involving immune responses, ECM alterations, and increased vasculature. These processes revealed unexpected therapeutic targets critical for stroma function. Further investigation of this resistance process may reveal common mechanism across inhibitors like Tyrosine Kinase Receptor inhibitors, BRAFi, KRASi, and MEKi, facilitating the design of new drug combinations to benefit a wider range of patients progressing to these MAPK pathway-targeted therapies. Recent perspectives suggest aggressive therapies such as chemotherapy, radiotherapy or some target-directed drugs that block cancer cell proliferation may promote the Darwinian selection of resistant genetic clones. Instead, combining these therapies with drugs that modulate other aspects of tumor biology, such as cancer cell invasion, plasticity, persistence, dormancy or stroma environment, may prevent the selection of more aggressive resistant genetic clones and prolong patient benefit ( 49 , 50 ). A good example of this strategy is blocking the MAPK pathway in mCRC with dabrafenib (BRAFi) and trametinib (MEKi) eliciting an immune response that, when combined with anti-PD-1 immunotherapy (spartalizumab), prolonged therapeutic benefit by modulating the immune stroma ( 23 ). Following an equivalent rationale, we focused on blocking stromal tumor vasculature induced by BRAFi therapy with bevacizumab to prevent acquisition of resistance. We observed increased expression of vasculature-related genes through not classic angiogenesis regulators, but not VEGFA, which instead showed an increase in protein secretion. Interestingly, MUC1 has also being described to promote the expression and secretion of VEGFA ( 51 ), completing a Mucin/Wnt-b-catenin/VEGFA axis behind the acquisition of resistance to BRAFi therapies. In addition, we observed an increased expression of PROM1 /CD133, which has been described to stabilize secreted VEGFA( 27 ). Indeed, we demonstrated that blocking VEGFA with bevacizumab extended the benefit of BRAFi therapy, whether used upfront or after progression, demonstrating that the drug-enhanced vasculature in BRAF mutant mCRC tumors is a vulnerability that can be exploited early in resistance acquisition. Consequently, our findings led to the initiation of the BRAVE clinical trial (NCT06411600), where we are testing the combination of encorafenib, cetuximab, and bevacizumab in mCRC patients refractory to previous lines of chemotherapy and naive for BRAFi. This trial aims to extend PFS, leveraging our research insights into stroma and vasculature remodeling during acquisition of resistance to BRAFi therapies, for improving patients’ outcome. We have currently enrolled 15 patients showing promising results. Methods Cell lines Human COLO 205 colorectal cancer (CRC) cells were generously provided by Dr. Diego Arango. Other human CRC cell lines used were HT29 and SW620, both procured from the American Type Culture Collection (ATCC). These cell lines were maintained in Dulbecco’s modified eagle medium (DMEM) supplemented with 10% Fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) and maintained at 37°C within a humidified incubator with 5% CO 2 . Generation of GCNT1/3 knockout cell lines The primer sequences used to generate GCNT1 and GCNT3 knockout cell lines were: CRISPR sgRNA GCNT1 forward:5-’CACCGAAGCGGTATGAGGTCGTTAA -3’. CRISPR sgRNA GCNT1 reverse:5’-AAACTTAACGACCTCATACCGCTTC-3’. CRISPR sgRNA GCNT3 forward:5’-CACCGGCGGGCTATCTGCGTTTATG-3’. CRISPR sgRNA GCNT3 reverse:5-’AAACCATAAACGCAGATAGCCCGCC-3’. CRISPR sgRNA Scramble forward:5’-CACCGGCACTACCAGAGCTAACTCA-3’. CRISPR sgRNA Scramble reverse:5’-AAACTGAGTTAGCTCTGGTAGTGCC-3’. GCNT1 and GCNT3 knockout (GCNT1/3 KO) and control (Scramble) cell lines were generated using CRISPR-Cas9 system. HT29 cells were transfected with pSpCas9-sgRNAGCNT1guide1-2A-Puro and pSpCas9-sgRNA-GCNT3guide1-2A-Puro constructs using linear polyethylenimine (PEI 25000, Polysciences). Following a 48 hours post-transfection period, the cells that had incorporated the plasmid were selected using puromycin at a concentration of 1µg/ml for 72 hours. Subsequently, individual clones were expanded. Validation of the HT29 GCNT1/3 KO models was carried out using a PCR-based screening approach (52). Validation primers: GCNT1 forward: 5’-GCAATGAGTGCAAACTGGAA-3’. GCNT1 reverse: 5’-TTCAGGAATCCTTTGGATGG-3’. GCNT3 forward: 5’-CAGGGAATGCGTACATTGTG-3’. GCNT3 reverse: 5’- CACAGGTAGCAACGCTCTCA-3’. Two-dimensional (2D) and Three-dimensional (3D) cell culture For the 2D culture of human colorectal cancer cell lines (COLO 205, HT29, SW620), cells were sustained in Dulbecco’s modified eagle medium (DMEM); 10% Fetal Bovine Serum (FBS) and 1% Penicillin-Streptomycin (P/S). For the 3D culture of the HT29 cell line, 2x10 5 cells were suspended in 200 µl of medium combined with Matrigel (1:3 v/v ratio) with Human Intestinal Stem Cell Media (HISC)(53) when single cell es required for the experiment the HISC media is complemented with Rock inhibitor (10µM) and GSK3 inhibitor (5µM) In both cases, 2D and 3D cultures, after 72 hours, cells were maintained under serum-limiting condition (0.5% FBS). Following an overnight incubation, the cells were exposed to encorafenib (1µM), a combination therapy of cetuximab (100µg/ml), encorafenib (1µM) and binimetinib (100nM) – collectively referred to as the 'triplet' – or vehicle in DMEM with 2.5% FBS for 48 hours. All cell lines used in this study were cultured in a humidified incubator at 37°C with 5% CO 2 . Patient samples Samples were collected from patients at Vall d'Hebron University Hospital and Hospital Sant Joan Despí Moisès Broggi. Novartis Biomedical Research procured patient tumor specimens from the National Disease Research Interchange (Philadelphia, Pennsylvania, USA), Cooperative Human Tissue Network which is funded by the National Cancer Institute (Rockville, Maryland, USA) and Maine Medical Center (Portland, Maine, USA). All patients provided informed consent for the use of their samples. Mice The experiments were conducted in accordance with the animal care directive of the European Union (86/609/CEE) and were approved by the Ethical Committee of Animal Experimentation of Vall d'Hebron Institute of Research (VHIR)/Vall d'Hebron Institute of Oncology (VHIO); ID: 12/18 CEEA). Female NOD.CB17- Prkdc scid /Rj mice, approximately 8 weeks old, were purchased from Janvier Labs. The mice were housed in groups of 5 animals per ventilated cage, with ad libitum access to food and water. Environmental conditions encompassed a 12-hour light-dark cycle, a temperature of 22 ± 2°C, and a relative humidity of 45-65%. The randomization of mice was performed based on tumor size. Continuous monitoring of the mice was conducted by the authors, facility technicians, and a veterinary scientist responsible for overseeing animal welfare. Patient-derived xenograft (PDX) establishment Human colorectal carcinoma tissues were obtained upon surgery in accordance with the ethical standards of the institutional committee on human experimentation. Histological diagnosis was based on microscopic features of carcinoma cells to determine both histological type and grade. The excised cancer tissues were washed 3 times in cold PBS solution and incubated overnight in DMEM/F-12 supplemented with an antibiotic and antifungal cocktail (penicillin at 250 U/ml, streptomycin at 250 μg/mL, fungizone at 10 μg/mL, kanamycin at 10 μg/mL, gentamycin at 50 μg/mL, and nystatin at 5 μg/mL). Enzymatic digestion was performed using collagenase (1.5 mg/mL) and DNase I (20 μg/mL) in a medium containing the same antibiotic and antifungal cocktail during 1 h at 37°C with intermittent pipetting every 15 minutes to disperse cells. After digestion, the dissociated sample was filtered through a 100 μm pore size filter and then washed with fresh medium. To remove red blood cells, the sample was subjected to a 10 minutes ammonium chloride treatment, followed by another wash. Finally, a suspension of 1x10 5 cells in 50 μL of PBS was mixed with an equal volume of Matrigel. This cell-Matrigel mixture was then subcutaneously injected into NOD-SCID flank mice to establish subcutaneous tumors for subsequent histological, protein, DNA, and RNA analyses. Novartis Biomedical Research generated PDX models as previously described (54). In vivo studies Human colon carcinoma models were used and grown as previously described (55). 10 5 cells for PDX or 10 6 cells for cell line, suspended in PBS, were mixed with Matrigel (1:1 v/v ratio) and subcutaneously injected into both flanks of NOD-SCID mice. For the experiment involving tumor pieces, 3x3 mm tumor pieces obtained from a resistant mouse PDX to BRAFi tumors were immediately implanted subcutaneously in both flanks. Once the tumors reached a volume of 60 mm 3 , mice were randomly assigned into distinct experimental groups, including control (vehicle), encorafenib, bevacizumab, combination therapy involving cetuximab, encorafenib, and binimetinib (triplet), as well a combination therapy consisting of the triplet and bevacizumab. Cetuximab was administered intraperitoneally at a dosage of 20 mg/kg twice a week. Encorafenib was orally administered daily at 20 mg/kg, binimetinib was given orally twice a day at 6 mg/kg, and bevacizumab was intraperitoneally administered twice a week at a dose of 5 mg/kg. When the tumors reached the end-point criteria, the mice were euthanized, and complete necropsies were performed. The subcutaneous tumors were fixed in formalin or frozen in liquid nitrogen immediately after the extraction to perform the histological, RNA, DNA and protein analysis. Tumors were measured 3 times per week, and their volume was estimated using the formula: V = (length×width 2 )/2, where length represents the largest tumor diameter and width represents the perpendicular tumor diameter. The number of mice for each experiment is reported in figure legends. PFS and survival curves Progression-free survival (PFS) was determined in the subcutaneous tumor growth experiments using the response evaluation criteria in solid tumors (RECIST) (56). Disease progression was identified by a percentage change of approximately 25% in tumor size from the baseline (time of treatment initiation). To calculate the time elapsed between the baseline tumor size and a 25% increase in tumor volume, a second-order polynomial (quadratic) equation was applied through the GraphPad Prism software. Statistical significance was evaluated utilizing the Kaplan-Meier survival curve along with the log-rank (Mantel-Cox) test, with significance considered at p < 0.05. Best response analysis An analysis of the best treatment response was performed, and this has been depicted using a waterfall plot. Using GraphPad Prism software, we generated a second-degree polynomial equation to calculate the elapsed time between the baseline tumor volume of the control group and a 50% increase. Once this day was calculated, the graph was extrapolated to indicate the value corresponding to the best response of each tumor. Sample preparation for Western blot analyses Tumors Protein extracts from tumors were processed in SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM dithiothreitol, DTT) supplemented with protease and phosphatase inhibitor cocktails. The samples were homogenized in the lysis buffer using a pestle. Then, samples were heated at 95ºC using a dry bath for 5 minutes and sonicated. Once the samples were thoroughly homogenized, they were centrifuged at 15000g for 10 minutes. Finally, the supernatant was collected and quantified using the Pierce BCA Protein Assay Kit. Two-dimensional (2D) cell culture Total cell extracts from cells were homogenized in SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) containing protease and phosphatase inhibitor cocktails. Then, samples were boiled at 95ºC using a dry bath for 5 minutes, followed by sonication and centrifugation at 15000 g for 10 minutes. The resulting supernatant was collected and subsequently quantified using the Pierce BCA Protein Assay Kit. Three-dimensional (3D) cell culture Prior to initiating the protein extraction from the spheroids, the supernatant was removed and set aside for later protein extraction. Cell recovery solution was added to the Matrigel embedded spheroids and incubated at 4ºC on a shaker at 30 rpm for 30 minutes. This solution dissolved the Matrigel, enabling the intact retrieval of the spheroids. After the 30-minute incubation, the mixture was centrifuged at 1400 rpm for 5 minutes at 4ºC. The supernatant was then collected and reserved for subsequent protein extraction that might have been released to the Matrigel. The resulting pellet (spheroids) was washed with cold PBS and lysed using an SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) supplemented with a cocktail of protease and phosphatase inhibitors. It was heated at 95ºC using a dry bath for 5 minutes, sonicated, and then centrifuged at 15000g for 10 minutes. The supernatant was collected and stored for subsequent quantification. The proteins present in the Matrigel and in the supernatant were precipitated using trichloroacetic acid (TCA). For this, 1 volume of TCA was added to 4 volumes of the sample in both the supernatant and the Matrigel, followed by incubation for 15 minutes on ice. After this period, the samples were centrifuged at 4ºC at 16000g for 15 minutes. The supernatant was discarded, and the pellet was resuspended with 600 µl of cold acetone. The mixture was then centrifuged again at 4ºC at 16000g for 15 minutes, followed by an additional acetone wash. The supernatant was discarded, and the pellet was air-dried for 5 minutes at 95ºC. Subsequently, the pellet was resuspended in an SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) supplemented with a cocktail of protease and phosphatase inhibitors. The samples were heated at 95ºC for 5 minutes and then sonicated. The protein extract was stored for subsequent quantification. Western blot Western blots were conducted in accordance with established protocols. Briefly, samples were boiled at 95°C for 5 minutes. Lysates were then separated by SDS-PAGE and proteins were transferred onto a nitrocellulose membrane (BioRad). Membranes were blocked with TBS, 0.1% Tween 20, and either 5% non-fat milk or BSA. They were then incubated overnight at 4°C with the appropriate primary antibodies (EGFR 1:1000 (#SC-03), pEGFR 1:1000 (#2234), ERK 1:3000 (#9102), pERK 1:3000 (#9101), MMP7 1:2000 (#ab207299), VEGFA 1:1000-1/5000 (#ab183100, #ab155944) , Vinculin 1:5000 (#V9131), GAPDH 1:10000 (#ab128915)). Subsequently, the membranes were exposed to HRP-conjugated secondary antibodies diluted in the blocking solution for 1 hour at room temperature, after that, the membranes were revealed using SuperSignal West Pico PLUS Chemiluminescent Substrate and were visualized using the Amersham Imager 600 (GE Healthcare). Sample preparation for proteomics analyses Protein extraction PDX tumor samples, around 50 mg wet weight, where lysed with 150 µL of lysis buffer (7M urea, 2M thiourea, 4% CHAPS, 30mm Tris.HCl pH=8.5), supplemented with 1M NaF and 0.1M Na 3 VO 4 as phosphatase inhibitors. Samples were first mechanically disrupted using a micro pestle, and then sonicated using a probe sonicator ( VCX 150; Sonics & Materials Inc. USA ) with 10 cycles of 15 seconds of ultrasound bursts, followed by 10 seconds of cooling intervals, while keeping the tube ice-cooled. Then, lysed samples were centrifuged for 5 min at 16000 g and the supernatants were collected. Protein extracts were further purified by a modified TCA-acetone precipitation (2D-CleanUp Kit, GE Healthcare) and, finally, resuspended in 100 µL of 8M urea, 50mM ammonium bicarbonate, plus phosphatase inhibitors. Protein concentration was determined using the Bio-Rad RCDC Protein Assay (Bio-Rad, UK). Trypsin digestion for total proteome analysis 20 µg of total protein from each sample were digested for total proteome analysis. Proteins were first reduced with DTT by addition of freshly prepared 700 mM DTT solution to a final concentration of 10 mM, for 1h at RT. Next, they were carbamidomethylated with iodoacetamide (IAA), by addition of the required volume of freshly prepared 700 mM IAA to obtain a final concentration of 30 mM in the sample. Alkylation was allowed to proceed for 30 minutes at RT in the dark, and then the reaction was quenched by addition of N-acetyl-L-cysteine to a final concentration of 35 mM, followed by incubation for 15 minutes at RT in the dark. Samples were then diluted with 50 mM ammonium bicarbonate to a final concentration of 1M Urea, and then modified porcine trypsin (Promega Gold) was added in a ratio of 1:20 (w/w), and the mixture was incubated overnight at 37 °C. The reaction was stopped with formic acid (FA) to a final concentration of 0.5%, and the digest was kept at -20ºC until further analysis. Liquid chromatography-Mass spectrometry analysis (LC-MS) For LC-MS/MS analysis peptide mixtures were diluted in 3% ACN, 1% FA and the sample was loaded to a 300 μm × 5 mm Pep-Map C18 (Thermo Scientific) at a flow rate of 15 μl/min using a Thermo Scientific Dionex Ultimate 3000 chromatographic system (Thermo Scientific). Peptides were separated using a C18 analytical column (NanoEase MZ HSS T3 column, 75 μm × 250 mm, 1.8 μm, 100Å, Waters) with a 210 minutes run for total proteome samples, comprising four consecutive steps, first 3 minutes of isocratic flow at 3% B, then gradient flows from 3 to 35% B in 180 minutes, from 35 to 50% B in 5 minutes, from 50 to 85% B in 1 minute, followed by isocratic elution at 85 % B in 5 minutes and stabilization to initial conditions (A= 0.1% FA in water, B= 0.1% FA in ACN), comprising four consecutive steps, first 3 minutes of isocratic gradient at 3%B, from 3 to 35% B in 90 minutes, from 35 to 50% B in 5 minutes, from 50 to 85% B in 1 minute, followed by isocratic elution at 85% B in 5 minutes and stabilization to initial conditions (A= 0.1% FA in water, B= 0.1% FA in ACN). Flow rate was 250 nL/min and the column was kept at 40 ºC. The column outlet was directly connected to an Advion TriVersa NanoMate (Advion) fitted on an Orbitrap Fusion Lumos™ Tribrid (Thermo Scientific). The mass spectrometer was operated in a data-dependent acquisition (DDA) mode. Survey MS scans were acquired in the orbitrap with the resolution (defined at 200 m/z) set to 120,000. The lock mass was user-defined at 445.12 m/z in each Orbitrap scan. The top speed (most intense) ions per scan were fragmented in the HCD cell and detected in the orbitrap at 30000 resolution. Quadrupole isolation was employed to selectively isolate peptides of 350-1700 m/z. The predictive automatic gain control (pAGC) target was set to 4e5. The maximum injection time was set to 50ms for MS1 and 70ms for MS2 scan. Included charged states were 2 to 7. Target ions already selected for MS/MS were dynamically excluded for 15 s. The mass tolerance of this dynamic exclusion was set to ±2.5 ppm from the calculated monoisotopic mass. Spray voltage in the NanoMate source was set to 1.7 kV. RF Lens were tuned to 30%. Minimal signal required to trigger MS to MS/MS switch was set to 5000 and activation Q was 0.250. The spectrometer was working in positive polarity mode and singly charge state precursors were rejected for fragmentation. Protein identification, quantitative differential and functional analysis. Progenesis ® QI for proteomics software v3.0 (Nonlinear dynamics, UK) was used for MS data analysis. The LC-MS runs were aligned to an automatically selected reference sample Alignments were then manually supervised. Only features within the 400 to 1,600 m/z range, 45 to 190 minutes retention time, for total proteome, and with positive charges between 2 to 4 were considered for identification and quantification. Peak lists (mgf files) were generated using Progenesis and loaded to Proteome Discoverer v2.1 (Thermo Fisher Scientific) for protein identification. Proteins were identified using Mascot v2.5 (Matrix Science, London UK) to search the SwissProt database. Two different searches were performed, restricting taxonomy to human or mouse proteins, respectively. MS/MS spectra were searched with a precursor mass tolerance of 10 ppm, fragment tolerance of 0.02 Da, trypsin specificity with a maximum of 2 missed cleavages, cysteine carbamidomethylation set as fixed modification and methionine oxidation as variable modification for total proteome analysis. Significance threshold for the identifications was set to p<0.05, minimum ions score of 20. Protein quantification of human, mouse or ambiguous proteins was performed by addition of the corresponding integrated peptide MS signals derived from the Progenesis analysis. On the basis of the two searches performed, MS features were classified in those attributable to mouse or human exclusive peptides and those corresponding to shared sequences. For those proteins presenting the three types of peptide features, the abundances of human and mouse proteins was estimated by deconvolution of the shared feature signals, as described in Saltzman et al. 2018 (57). The algorithm was implemented as an Excel macro. Statistical analysis was performed using the Perseus software platform (58). Proteins presenting Fold change|>2 and adjusted p-value < 0.05 (T-test on log2 transformed abundance values) were considered as differential between the control and treated groups. After the proteome differential expression analysis, functional analysis was performed using the Pre-Ranked Gene Set Enrichment Analysis (GSEA) method (59,60). This analysis was implemented with the clusterProfiler (61) R package v. 4.8.2 . Enrichment analysis was conducted on the following gene set collections from the Molecular Signatures Database (MSigDB), v. 2023.2: Hallmark gene sets, C2, C5 gene ontology (GO) gene sets and C6 oncogenic signatures. Immunohistochemical (IHC) staining Formalin-fixed paraffin-embedded (FFPE) tissues underwent standard deparaffination and rehydration processes. For antigen retrieval, the slides were incubated with a 10 mM sodium citrate pH 6 buffer. Following the inhibition of endogenous peroxidase activity using a 3% H 2 O 2 solution, the slides were permeabilized using a 1% Tween 20 solution in PBS for 15 minutes. Then, the tissue specimens were blocked with a 3% BSA solution in PBS for 1 hour and then incubated with the corresponding primary antibodies (CD31 1:20 (#DIA-310), Ki67 1:100 (#M7240), Muc2 1:100 (#BD555926), α-SMA 1:1500(#A2547)) diluted in blocking solution at 4°C overnight. After washing, sections were incubated with corresponding HRP-conjugated secondary antibodies for 1 hour at room temperature (RT). For chromogenic detection, the Dako Liquid DAB+ Substrate Chromogen System was added onto the slides and incubated up to 10 minutes. Finally, the slides were counterstained with haematoxylin, dehydrated, and mounted. The NanoZoomer 2.0-HT Digital slide scanner C9600 was used to acquire a high-resolution whole slide scanner of the immunostainings. Alcian blue staining Alcian blue staining was carried out on FFPE tumor xenograft sections and patient sections to identify mucinous differentiation. This staining procedure was conducted by the Translational Molecular Pathology service at VHIR (Vall d’Hebron Institute of Research). In brief, following the process of deparaffinization, FFPE sections were hydrated and subjected to a 3-minute treatment in absolute ethanol, followed by 5 minutes in acetic acid, 30 minutes in 1% Alcian blue solution, and 10 minutes in nuclear fast red. After each step, the sections were rinsed three times for 5 minutes each using PBS washes. Subsequently, the samples were dehydrated and coverslipped using mounting media. The NanoZoomer 2.0-HT Digital slide scanner C9600 was employed to visualize and evaluate the Alcian blue staining. For image quantification, the open-source software ImageJ was used. The analysis of Alcian blue staining involved quantifying the percentage of the Alcian blue-positive area in each tumor section. Picro-sirius red staining Sirius red staining was performed on FFPE tumor xenograft sections and patient sections to identify collagen fibers. This staining procedure was conducted by the Translational Molecular Pathology service at VHIR (Vall d'Hebron Institute of Research). In brief, after the deparaffinization process, FFPE sections were hydrated and then incubated in the picro-Sirius red solution for 1 hour. This was followed by a 5 second immersion in a 0.1% hydrochloric acid solution. After this step, the slides were rinsed in tap water for 10 minutes and then washed in distilled water. Finally, the slides were counterstained with haematoxylin, dehydrated, and mounted. The NanoZoomer 2.0-HT Digital slide scanner C9600 was utilized to visualize and assess the Sirius red staining. The picro-Sirius red staining was analyzed using the Twombli (62), a Fiji macro designed for quantifying patterns in the extracellular matrix. RNA extraction For RNA extraction of 2D culture experiments, the plates were washed 3 times with cold PBS, and the extraction protocol was initiated. For in vivo experiments, snap-frozen samples were homogenized using a pestle. For both scenarios, the RNA extraction was performed using the RNeasy Mini Kit (Qiagen), following the manufacturer's recommendations. The obtained RNA was quantified using a NanoDrop spectrophotometer (ThermoFisher Scientific). Gene expression analyses Samples were processed according to the following Affymetrix protocols: GeneChip WT PLUS Reagent kit (P/N 703174 2017) and Expression Wash, Stain and Scan User Manual (P/N 702731 2017) (Affymetrix Inc., Santa Clara, CA, USA) and hybridized to Clariom D Human expression array (ThermoFisher Scientific). Data were normalized using the robust multi-array average (RMA) method from the oligo R package (63) and differential gene expression was calculated using the limma R package (64). P-values were adjusted for multiple testing using the Benjamini and Hochberg (BH)correction. Genes differentially expressed (|Fold change|>2 and adjusted p-value < 0.05) in any of the time points for either Naive or Resistant models were represented in a heatmap. To evaluate the gene response to the treatment over time, z-scores for each gene were calculated using the scale function in R and the median of the different replicates within each time point for each group was calculated. Over representation analysis (ORA) and gene set enrichment analysis (GSEA) were performed with the clusterProfiler R package (61,65). Gene set collections (Gene Ontology, Reactome, KEGG, Chemical and Genetic Perturbations and Hallmark) were obtained from the Molecular Signature Database (MSigDb 2022). To construct heatmaps both pheatmap and ComplexHeatmap R packages were used (66,67) while for the rest of the graphs ggplot2 was used. Gene sets were classified as Naive specific, Resistant specific or commonly regulated depending on which condition they were significantly enriched (adjusted p-value < 0.01). Then, they were clustered by the similarity of the genes in their core enrichment using the GSEAmining package (68). Gene set clusters were then manually classified to specific biological categories (Figure 1D) and represented using alluvial plots with the ggalluvial R package (69). GSEA plots were performed using the gseaplot2 function from the enrichplot package (70). Goblet cell markers were obtained from (71,72).Those that were up-regulated in each of the time points (|Fold change|>1.2 and adjusted p-value < 0.05) were represented in a heatmap. Normalized enrichment score (NES) from gene sets in the alluvial plot annotated in tissue remodelling and vasculature processes positively enriched in naive models were represented in bubble plots. From those selected gene sets, genes that were differentially expressed in each time point present in the core enrichment genes (from leading edge analysis) of the gene sets were represented in a heatmap. Quantitative RT-PCR analyses To analyze the expression of selected genes, RNA retrotranscription was performed using the iScript™ Advanced cDNA Synthesis Kit following the manufacturer’s instructions. Analyses were carried out in triplicate with 10 ng of cDNA using using PerfeCTa SYBR Green FastMix on a Quant Studio 6 Flex cycler. Specific pairs of primers were designed (https://www.ncbi.nlm.nih.gov/gene/) and used to detect the indicated transcripts. Relative gene expression was determined using the comparative CT method. TATA-binding protein (TBP) and Hypoxanthine Phosphoribosyltransferase 1 (HPRT) were used as housekeeping genes. ATOH1 forward primer: 5’-AGCTTCTTGTCGTTGTTG-3’ ATOH1 reverse primer: 5’-AGGTGAATGGGGTGCAGA-3’. CLCA1 forward primer: 5’-TCGTTGCAATCGACCCCAAT-3’. CLCA1 reverse primer: 5’-CCTGGGTCACCATGTCCTTT-3’. COL1A1 forward primer: 5’-GATTCCCTGGACCTAAAGGTGC-3’. COLA1 reverse primer: 5’-AGCCTCTCCATCTTTGCCAGCA-3’. MMP7 forward primer: 5’-TCGGAGGAGATGCTCACTTCGA-3’. MMP7 reverse primer: 5’-GGATCAGAGGAATGTCCCATACC-3’. MUC4 forward primer: 5’-AACACAGCCTGCTAGTCCAGCA-3’. MUC4 reverse primer: 5’-TGGAGAGGATGGCTTGGTAGGT-3’. VEGFA forward primer: 5’-AGGGCAGAATCATCACGAAGT-3’. VEGFA reverse primer: 5’-AGGGTCTCGATTGGATGGCA-3’. TBP forward primer: 5’- CGGCTGTTTAACTTCGCTTC -3’ TBP reverse primer: 5’-CACACGCCAAGAAACAGTGA -3’. HPRT forward primer: 5’-CTGGCGTCGTGATTAGTGAT -3’. HPRT reverse primer: 5’-GGCTACAATGTGATGGCCT-3’. Single Cell Analysis Analysis of the Single Cell data was performed in R using the Seurat package (version 5.0.3). Cells that expressed less than 200 genes were filtered out, together with genes that were only expressed in less than 3 cells. Additional filtering was done to remove droplets containing more than one cell, by filtering for cells with more than 1000 and less than 8000 features. Dead cells were removed by filtering for less than 25% mitochondrial DNA. The filtered gene expression data was normalized using the 'LogNormalize' method with a scale factor of 10.000, implemented through the ‘NormalizeData’ function in the Seurat package. Next, the 2000 most variable feature were identified, and the data was scaled for all genes. Using the ‘FindNeighbours’ and the ‘FindCluster’ function from Seurat, cells were clustered, with a resolution of 0.5 and non-linear dimensionality reduction, with the ‘RunUMAP’ function, was done using the first 20 principal components, to visualize the high-dimensional data in a low-dimensional space. A total of 14 clusters were found and annotated using a similar set of features as described in the single-cell transcriptomic analysis paper by Sebastian et al (73). DNA extraction For in vivo experiments, DNA extraction from snap-frozen samples was performed using the QIAGEN QIAamp DNA Mini and Blood Mini Kit, following the manufacturer's recommendations. Whole-Exome Sequencing (WES) The libraries for Whole Exome Sequencing (WES) were prepared from 500 ng of DNA per sample using the SureSelect XT Target Enrichment kit for Illumina platforms. For enrichment, the SureSelect XT Human All Exon v5 capture set (Agilent) was used, and sequencing was performed on the HiSeq 2500 sequencer (chemistry v3, high-output mode). This achieved a coverage of approximately 150x with a read length of 2x100 bp. Data processing and analysis The sequencing raw data from PDX were pre-processed using BBSplit (BBTools) (74) to separate human-origin reads and remove mouse contamination. We then utilized the sarek/nf-core 3.2.3 (75), with default settings, to extract somatic variants from paired samples. Quality checks for the Fastq files were performed by FastQC v0.11.4 (76), followed by trimming low-quality regions using fastP v0.23.4.(77). Alignment to the human reference genome GRCh38 was successfully done using Burrows-Wheeler Alignment with mem algorithm (BWA-MEM) v.0.7.17-r1188 (78). After error correction based on base recalibration using GATK v.4.4.0 (79,80) and duplicate removal by Picard v2.9.2 (81), somatic variant calling was done with Mutect2 (82) in a control versus tumor setup to identify single nucleotide variants (SNVs) and short insertions and deletions (indels), which were then converted to Variant Call Format (VCF). Publicly available resources such as normal sample panels of GATK and databases of already known mutations like Mills (83) for short indels and dbSNP (84) for single nucleotide polymorphisms were used in order to avoid artifacts and perform base recalibration effectively. Variants failing to meet quality criteria of GATK Best Practices were filtered out, and those with variant allele frequencies below 5% or supported by fewer than 10 reads were excluded. Variant annotation was performed using SNPeff v.5.1d (85), and conversion into Mutation Annotation Format (MAF) was conducted using the maftools package (86). Heatmaps and oncoprints were generated using the ComplexHeatmap R-package (66), while tumor mutational burden (TMB) was calculated with the maftools package. Additional analyses were carried out using Python v3.7 and R v4.2.2. We used data from Colorectal Adenocarcinoma cohort of TCGA (COADREAD, PanCancer Atlas). Data was downloaded from cBioportal (87) selecting only BRAF-V600E mutated patients. Tumor mutational signatures were obtained from variant calling format (VCFs) data extracted in earlier stages and analyzed using SigProfilerAssignment v.0.1.4 (88). This software assigns specific established signatures to individual samples based on the type and frequency of variants, with a focus on single nucleotide polymorphisms (SNPs) for single base substitution signatures (SBS). Signature data were sourced from the latest release of the Catalogue of Somatic Mutations in Cancer (COSMIC), version 3.4 (89). Contribution of each signature was determined by evaluating the number of mutations attributed to it relative to the total mutation burden. To extract TCGA signatures, MAF data from the same previous cohort COADREAD TCGA Pan Cancer Atlas were downloaded in GRCh37 format from cBioPortal. Only samples harboring BRAFV600E mutations were selected for further analysis. The MAF data were converted to VCF using maf2vcf (18) and adjusted to GRCh38 using the Liftover function of Picard v.2.9.2. The resulting VCF files were successfully processed using SigprofilerAssignment (90). DNA methylation analyses Sample preparation FFPE DNA extraction. DNA isolation from FFPE embedded samples from PDX tissues was performed using the ReliaPrep FFPE gDNA Miniprep System Kit (Promega, USA). All DNA samples were treated with RNase A for 1 h at 45ºC, quantified by the fluorometric method (Quant-iT PicoGreen dsDNA Assay, Life Technologies, CA, USA), and assessed for purity by NanoDrop (Thermo Scientific, MA, USA) 260/280 and 260/230 ratio measurements . Quality check of FFPE DNA. Following Illumina’s recommendations, the DNA quality of the FFPE samples was checked by performing a quantitative PCR with 2 ng of FFPE DNA. The average value of a standard provided by Illumina was subtracted in order to calculate the ΔCq. FFPE DNAs with ΔCq <5 indicate suitability for FFPE restoration. Bisulphite conversion and restoration. A minimum of 300 ng of DNA was bisulphite-converted using the EZ-96 DNA Methylation kit (Zymo Research Corp., CA, USA), following manufacturer’s instructions for Infinium assays. Next, bisulfite-converted DNA (bs-DNA) from FFPE samples was restored as previously described (91). In brief, after DNA denaturation (NaOH 0.1N, 10 min at room temperature), PPR (Primer Pre Restore) and AMR (Amplification Mix Restore) reagents were added, and samples were incubated for 1 h at 37ºC. DNA was cleaned with ZR-96 DNA Clean & Concentrator-5 kit (Zymo Research Corp.) and eluted in 13 μl of ERB (Elution Restore Buffer Reagent). Cleaned DNA was then denatured for 2 min at 95°C, followed by ligation incubation at 37°C for 1 h with ER (Elution Restore) and CMM (Convert Master Mix) reagents. After a second cleaning step with ZR-96 DNA Clean & Concentrator-5 kit (Zymo Research Corp.), DNA was finally eluted in 10 μl of DiH20. Array hybridization. Eight microliters of restored FFPE bs-DNA were used as input for the Illumina Infinium HD Methylation Assay Protocol, as previously described [Moran et al., 2016]. Briefly, the process includes a whole genome amplification step followed by enzymatic end-point fragmentation, precipitation, and resuspension. The resuspended samples were hybridized on the Illumina Infinium MethylationEPIC BeadChip at 48°C for 16 h. Then unhybridized and non-specifically hybridized DNA were washed away, followed by a single nucleotide extension using the hybridized bisulfite-treated DNA as a template. The nucleotides incorporated were labeled with biotin (ddCTP and ddGTP) and 2,4-dinitrophenol (DNP) (ddATP and ddTTP). After the single base extension, repeated rounds of staining were performed with a combination of antibodies that differentiated DNP and biotin by fixing them different fluorophores. Finally, the BeadChip was washed and scanned using the Illumina iScan with Autoloader system. Methylation Analysis Illumina EPIC Methylation array data in the form of IDAT files were imported into R using the minfi package (version 1.48.0). Stratified quantile normalization was conducted and subsequently probes with p-values exceeding 0.01 were filtered out, using the minfi function detectionP. Additional probe filtration was done removing those containing SNPs or being cross-reactive, for the latter utilizing data from Pidsley et al. (92). Probes were annotated using the IlluminaHumanMethylationEPICanno.ilm10b4.hg19 package (version 0.6.0). Differential methylation analysis was conducted on the M-values for each site using the limma package to identify differentially methylated probes (DMP), or CpGs, across the same comparisons employed in the gene expression analysis (Fig. 1c). P-values were adjusted for multiple testing using the BH method, and DMPs were selected based on a |Fold change|>1.5 and adjusted p-value < 0.05. To visualize the overlap of differentially methylated CpGs, Venn diagrams were constructed using the VennDiagram package (version 1.7.3). Stacked bar plots were generated to depict the distribution of differentially methylated CpGs in the genome, categorizing them based on CpG island status and location within promoters, gene bodies, or other functional regions. To compare gene expression and methylation on a gene-basis, the median methylation value of probes located in the TSS200 region and gene body, separately, was taken. Limma analysis was performed on these summarized methylation values to identify genes exhibiting differential methylation in their promoter or body regions. To explore the association between methylation and gene expression, scatterplots were created, plotting the log 2 Fold change of methylation against expression. Quantification and statistical analysis The statistical analyses of the "wet lab/mice" experiments were conducted using GraphPad Prism 8.1 and R v4.0.4 (93). Most significant differences were assessed using the Student's t-test for independent samples. Data were presented as means with SEM unless otherwise specified. For comparisons involving two or more groups that encompass continuous variables (such as gene expression, methylation), Mann-Whitney U and Kruskal-Wallis tests were employed. Statistical analyses were adjusted using the False Discovery Rate (FDR) or Bonferroni correction. Survival was measured using the Kaplan-Meier method. Asterisks codes were employed to denote different levels of statistical significance: * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001. Declarations The study including patients’ samples and data was approved (ID: PR(AG)113_2015) by the Vall d’Hebron University Hospital institutional ethical review board according to the guidelines of the European Network of Research Ethics Committees, following European, national and local laws. Written informed consent was signed by all patients. Acknowledgements We thank the IJC Genomics Unit. VHIO would like to acknowledge the Cellex Foundation for providing research facilities and equipment, the FERO Foundation for their funding support, and the Centro de Investigación Biomédica en Red de Cáncer (CIBERONC). We thank the Institute of Health Carlos III (ISCIII-FEDER) PI20/00968, Fundación AECC (CLSEN19001ELEZ) (ACRCelerate), Department of Health (Generalitat de Catalunya, 2021SGR01567), Mutual Médica Award, GRUPO DE TRATAMIENTO DE LOS TUMORES DIGESTIVOS (GTTD) and Fundación CRIS Contra el Cáncer. This research was partially funded by the SCITRON program from Novartis. We thank CERCA Programme / Generalitat de Catalunya for institutional support. Data availability The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium ( http://proteomecentral.proteomexchange.org ) via the PRIDE partner repository ( 57 ) with the dataset identifier PXD54589. Expression and methylathion data that support the findings of this study have been deposited in the Gene Expression Omnibus (GEO) under accession code GSE280600. ScRNA-seq and Exome Sequencing data have been deposited in the European Genome-phenome Archive (EGA) under the accession codes EGAD50000000939 and EGAD50000000940. References Sung, H. et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA. Cancer J. Clin. 71 , (2021). Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417 , 949–954 (2002). Dhomen, N. & Marais, R. New insight into BRAF mutations in cancer. Curr. Opin. Genet. Dev. 17 , 31–39 (2007). Taieb, J., Lapeyre-Prost, A., Laurent Puig, P. & Zaanan, A. Exploring the best treatment options for BRAF-mutant metastatic colon cancer. British Journal of Cancer vol. 121 434–442 (2019). Bond, C. E. & Whitehall, V. L. J. How the BRAF V600E mutation defines a distinct subgroup of colorectal cancer: Molecular and clinical implications. Gastroenterol. Res. Pract. (2018) doi:10.1155/2018/9250757. Roth, A. D. et al. Prognostic role of KRAS and BRAF in stage II and III resected colon cancer: Results of the translational study on the PETACC-3, EORTC 40993, SAKK 60-00 trial. J. Clin. Oncol. 28 , 466–474 (2010). Cantwell-Dorris, E. R., O’Leary, J. J. & Sheils, O. M. BRAFV600E: Implications for carcinogenesis and molecular therapy. Molecular Cancer Therapeutics vol. 10 385–394 (2011). Pratilas, C. A. et al. V600EBRAF is associated with disabled feedback inhibition of RAF-MEK signaling and elevated transcriptional output of the pathway. Proc. Natl. Acad. Sci. U. S. A. 106 , 4519–4524 (2009). Kopetz, S. et al. Encorafenib, Binimetinib, and Cetuximab in BRAF V600E–Mutated Colorectal Cancer . N. Engl. J. Med. 381 , 1632–1643 (2019). Prahallad, A. et al. Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature 483 , 100–104 (2012). Corcoran, R. B. et al. EGFR-mediated reactivation of MAPK signaling contributes to insensitivity of BRAF-mutant colorectal cancers to RAF inhibition with vemurafenib. Cancer Discov. 2 , (2012). Yaeger, R. D. et al. Pilot study of vemurafenib and panitumumab combination therapy in patients with BRAF V600E mutated metastatic colorectal cancer. J. Clin. Oncol. (2015) doi:10.1200/jco.2015.33.3_suppl.611. Van Geel, R. M. J. M. et al. A phase Ib dose-escalation study of encorafenib and cetuximab with or without alpelisib in metastatic BRAF-mutant colorectal cancer. Cancer Discov. 7 , 610–619 (2017). Xu, T. et al. Molecular mechanisms underlying the resistance of BRAF V600E-mutant metastatic colorectal cancer to EGFR/BRAF inhibitors. Ther. Adv. Med. Oncol. 14 , 175883592211050 (2022). Ahronian, L. G. et al. Clinical acquired resistance to RAF inhibitor combinations in BRAF-mutant colorectal cancer through MAPK pathway alterations. Cancer Discov. 5 , 358–367 (2015). Yaeger, R. et al. Mechanisms of acquired resistance to BRAF V600E inhibition in colon cancers converge on RAF dimerization and are sensitive to its inhibition. Cancer Res. 77 , 6513–6523 (2017). Corcoran, R. B. et al. Combined BRAF, EGFR, and MEK inhibition in patients with BRAF V600E -mutant colorectal cancer. Cancer Discov. 8 , 428–443 (2018). Hong, D. S. et al. Phase 1B study of vemurafenib in combination with irinotecan and cetuximab in patients with BRAF -mutated advanced cancers and metastatic colorectal cancer. J. Clin. Oncol. 32 , 3516–3516 (2014). Pietrantonio, F. et al. MET-driven resistance to dual EGFR and BRAF blockade may be overcome by switching from EGFR to MET inhibition in BRAF-mutated colorectal cancer. Cancer Discov. 6 , 963–971 (2016). Ros, J. et al. Encorafenib plus cetuximab for the treatment of BRAF-V600E-mutated metastatic colorectal cancer. Therap. Adv. Gastroenterol. 15 , 175628482211106 (2022). Kopetz, S. et al. Molecular profiling of BRAF-V600E-mutant metastatic colorectal cancer in the phase 3 BEACON CRC trial. Nat. Med. (2024) doi:10.1038/s41591-024-03235-9. Kopetz, S. et al. 316O Genomic mechanisms of acquired resistance of patients (pts) with BRAF V600E-mutant (mt) metastatic colorectal cancer (mCRC) treated in the BEACON study. Ann. Oncol. 33 , (2022). Tian, J. et al. Combined PD-1, BRAF and MEK inhibition in BRAFV600E colorectal cancer: a phase 2 trial. Nat. Med. 29 , 458–466 (2023). Corcoran, R. B. et al. Combined BRAF, EGFR, and MEK Inhibition in Patients With BRAFV600E-Mutant Colorectal Cancer. Cancer Discov. 8 , 428 (2018). Weisenberger, D. J. et al. CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat. Genet. 38 , 787–793 (2006). Hazar-Rethinam, M. et al. Convergent therapeutic strategies to overcome the heterogeneity of acquired resistance in BRAFv600e colorectal cancer. Cancer Discov. 8 , 417–427 (2018). Adini, A. et al. The stem cell marker prominin-1/CD133 interacts with vascular endothelial growth factor and potentiates its action. Angiogenesis 16 , 405–416 (2013). Ricci-Vitiani, L. et al. Identification and expansion of human colon-cancer-initiating cells. Nature 445 , 111–115 (2007). Ricci-Vitiani, L., Fabrizi, E., Palio, E. & De Maria, R. Colon cancer stem cells. J. Mol. Med. 87 , 1097–1104 (2009). Kim, Y. S. & Ho, S. B. Intestinal goblet cells and mucins in health and disease: Recent insights and progress. Current Gastroenterology Reports vol. 12 319–330 (2010). VanDussen, K. L. & Samuelson, L. C. Mouse atonal homolog 1 directs intestinal progenitors to secretory cell rather than absorptive cell fate. Dev. Biol. 346 , 215–223 (2010). Yang, Q., Bermingham, N. A., Finegold, M. J. & Zoghbi, H. Y. Requirement of Math1 for secretory cell lineage commitment in the mouse intestine. Science (80-. ). 294 , 2155–2158 (2001). Lupo, B. et al. Colorectal cancer residual disease at maximal response to EGFR blockade displays a druggable Paneth cell–like phenotype. Sci. Transl. Med. 12 , (2020). Takahashi-Yanaga, F. et al. Glycogen synthase kinase-3β is tyrosine-phosphorylated by MEK1 in human skin fibroblasts. Biochem. Biophys. Res. Commun. 316 , 411–415 (2004). Clevers, H. Wnt/β-Catenin Signaling in Development and Disease. Cell vol. 127 469–480 (2006). Clevers, H. & Nusse, R. Wnt/β-catenin signaling and disease. Cell vol. 149 1192–1205 (2012). Lau, T. et al. A novel tankyrase small-molecule inhibitor suppresses APC mutation-driven colorectal tumor growth. Cancer Res. 73 , 3132–3144 (2013). Hurwitz, H. et al. Bevacizumab plus Irinotecan, Fluorouracil, and Leucovorin for Metastatic Colorectal Cancer. N. Engl. J. Med. 350 , 2335–2342 (2004). Yamamoto, M., Bharti, A., Li, Y. & Kufe, D. Interaction of the DF3/MUC1 breast carcinoma-associated antigen and β- catenin in cell adhesion. J. Biol. Chem. 272 , 12492–12494 (1997). Pai, P. et al. The canonical Wnt pathway regulates the metastasis-promoting mucin MUC4 in pancreatic ductal adenocarcinoma. Mol. Oncol. 10 , 224–239 (2016). Zhan, T., Rindtorff, N. & Boutros, M. Wnt signaling in cancer. Oncogene vol. 36 1461–1473 (2017). Tenbaum, S. P. et al. β-Catenin confers resistance to PI3K and AKT inhibitors and subverts FOXO3a to promote metastasis in colon cancer. Nat. Med. 18 , 892–901 (2012). Bademler, S. et al. Clinical Significance of Serum Membrane-Bound Mucin-2 Levels in Breast Cancer. Biomolecules 9 , 40 (2019). Yin, B. W. T., Dnistrian, A. & Lloyd, K. O. Ovarian cancer antigen CA125 is encoded by the MUC16 mucin gene. Int. J. Cancer 98 , 737–740 (2002). Nabhan, A.N.; Brownfield, D.G.; Harbury, P.B.; Krasnow, M.A.; Desai, T.J. Single-cell Wnt signaling niches maintain stemness of alveolar type 2 cells. Science 2018 , 359 , 1118–1123. Desai, T.J.; Brownfield, D.G.; Krasnow, M.A. Alveolar progenitor and stem cells in lung development, renewal and cancer. Nature 2014 , 507 , 190–194. Di Martino, J. S. et al. A tumor-derived type III collagen-rich ECM niche regulates tumor cell dormancy. Nat. Cancer 3 , 90–107 (2022). Liotta, L. A. & Kohn, E. C. The microenvironment of the tumour - Host interface. Nature vol. 411 375–379 (2001). Shen, S., Vagner, S. & Robert, C. Persistent Cancer Cells: The Deadly Survivors. Cell vol. 183 860–874 (2020). Škarková, A. et al. Educate, not kill: treating cancer without triggering its defenses. Trends Mol. Med. (2024) doi:10.1016/J.MOLMED.2024.04.003. Woo, J. K. et al. Mucin 1 enhances the tumor angiogenic response by activation of the AKT signaling pathway. Oncogene 31 , 2187–2198 (2012). Giuliano, C. J., Lin, A., Girish, V. & Sheltzer, J. M. Generating Single Cell–Derived Knockout Clones in Mammalian Cells with CRISPR/Cas9. Curr. Protoc. Mol. Biol. 128 , e100 (2019). Sato, T. et al. Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett’s epithelium. Gastroenterology 141 , 1762–1772 (2011). Gao, H. et al. High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response. Nat. Med. 21 , 1318–1325 (2015). Puig, I. et al. A personalized preclinical model to evaluate the metastatic potential of patient-derived colon cancer initiating cells. Clin. Cancer Res. 19 , 6787–6801 (2013). Eisenhauer, E. A. et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur. J. Cancer 45 , 228–247 (2009). Saltzman, A. B. et al. GpGrouper: A peptide grouping algorithm for gene-centric inference and quantitation of bottom-up proteomics data. Mol. Cell. Proteomics 17 , 2270–2283 (2018). Tyanova, S. et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat. Methods 2016 139 13 , 731–740 (2016). Perez-Riverol, Y. et al. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic Acids Res. 50 , D543 (2022). Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 102 , 15545–15550 (2005). Yu, G., Wang, L. G., Han, Y. & He, Q. Y. ClusterProfiler: An R package for comparing biological themes among gene clusters. Omi. A J. Integr. Biol. 16 , 284–287 (2012). Wershof, E. et al. A FIJI macro for quantifying pattern in extracellular matrix. Life Sci. Alliance 4 , (2021). Gautier, L., Cope, L., Bolstad, B. M. & Irizarry, R. A. Affy - Analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20 , 307–315 (2004). Ritchie, M. E. et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43 , e47 (2015). Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation 2 , (2021). Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32 , 2847–2849 (2016). Gu, Z. Complex heatmap visualization. iMeta 1 , e43 (2022). Arqués, O. GSEAmining: Make Biological Sense of Gene Set Enrichment Analysis Outputs. Bioconductor open source Softw. Bioinforma. (2023). Brunson, Jason Cory and Read, Q. D. Alluvial Plots in ggplot2. Rpackage version 0.12.5. The Comprehensive R Archive Network https://corybrunson.github.io/ggalluvial/articles/ggalluvial.html (2023). Yu G. enrichplot: Visualization of Functional Enrichment Result. R package version 1.20.1. (2023). Burclaff, J. et al. A Proximal-to-Distal Survey of Healthy Adult Human Small Intestine and Colon Epithelium by Single-Cell Transcriptomics. Cmgh 13 , 1554–1589 (2022). Franzén, O., Gan, L. M. & Björkegren, J. L. M. PanglaoDB: A web server for exploration of mouse and human single-cell RNA sequencing data. Database 2019 , 1–9 (2019). Sebastian, A. et al. Single-cell transcriptomic analysis of tumor- derived fibroblasts and normal tissue-resident fibroblasts reveals fibroblast heterogeneity in breast cancer. Cancers (Basel). 12 , (2020). Bushnell, B. BBMap: A Fast, Accurate, Splice-Aware Aligner (No. LBNL-7065E). in Berkeley National Lab. (LBNL), Berkeley, CA (United States). (Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)., 2014). Nystedt, B. et al. Sarek: A portable workflow for whole-genome sequencing analysis of germline and somatic variants. F1000Research 9 , (2020). Andrews, S. & others. FastQC: a quality control tool for high throughput sequence data. 2010. Https://Www.Bioinformatics.Babraham.Ac.Uk/Projects/Fastqc/ 5 , http://www.bioinformatics.babraham.ac.uk/projects/ (2010). Chen, S., Zhou, Y., Chen, Y. & Gu, J. Fastp: An ultra-fast all-in-one FASTQ preprocessor. in Bioinformatics vol. 34 i884–i890 (Oxford Academic, 2018). Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. (2013). McKenna, A. et al. The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20 , 1297–1303 (2010). Van der Auwera, G. A. et al. From fastQ data to high-confidence variant calls: The genome analysis toolkit best practices pipeline. Curr. Protoc. Bioinforma. 11 , 11.10.1 (2013). Picard Tools - By Broad Institute. https://broadinstitute.github.io/picard/. Cibulskis, K. et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat. Biotechnol. 31 , 213–219 (2013). Mills, R. E. et al. An initial map of insertion and deletion (INDEL) variation in the human genome. Genome Res. 16 , 1182–1190 (2006). Sherry, S. T. et al. DbSNP: The NCBI database of genetic variation. Nucleic Acids Res. 29 , 308–311 (2001). Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin). 6 , 80–92 (2012). Mayakonda, A., Lin, D. C., Assenov, Y., Plass, C. & Koeffler, H. P. Maftools: Efficient and comprehensive analysis of somatic variants in cancer. Genome Res. 28 , 1747–1756 (2018). Cerami, E. et al. The cBio Cancer Genomics Portal: An open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2 , 401–404 (2012). Díaz-Gay, M. et al. Assigning mutational signatures to individual samples and individual somatic mutations with SigProfilerAssignment. Bioinformatics 39 , (2023). Tate, J. G. et al. COSMIC: The Catalogue Of Somatic Mutations In Cancer. Nucleic Acids Res. 47 , D941–D947 (2019). Cyriac Kandoth. mskcc/vcf2maf: vcf2maf v1.6.19. (2020) doi:10.5281/ZENODO.593251. Moran, S., Arribas, C. & Esteller, M. Validation of a DNA methylation microarray for 850,000 CpG sites of the human genome enriched in enhancer sequences. Epigenomics 8 , 389–399 (2016). Pidsley, R. et al. Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. Genome Biol. 17 , (2016). R Core Team. R Core Team 2021 R: A language and environment for statistical computing. R foundation for statistical computing. https://www.R-project.org/. R Found. Stat. Comput. 2 , 2019 (2022). Additional Declarations Yes there is potential Competing Interest. LR, OA, JMQ, DC, AME, IC, IP, GF, RF and HGP report financial support from Cyclacel Pharmaceuticals, Boundless Bio, Grifols, Roche, AstraZeneca and Vivan Therapeutics outside the submitted work. I.P. is cofounder of Oniria Therapeutics S.L. and reports financial support from this company outside the submitted work. HGP is a CSO, co-Founder & Board Member from Oniria Therapeutics S.L and reports financial support from this company outside the submitted work. JR has received personal speaker honoraria from Sanofi and Amgen, and accommodation expenses from Pierre-Fabre, Servier, Amgen and Merck. PN discloses personal financial interests receiving honoraria or consultation fees from Novartis, Bayer, Targos Molecular Pathology GmbH and MSD Oncology; and receiving travel, accommodation paid or reimbursed by Novartis.EE discloses personal financial interests receiving honoraria for advisory role, travel grants and research grants from Amgen, Bayer, Hoffman-La Roche, Merck Serono, Sanofi, Pierre Fabre, MSD, Organon, Novartis and Servier. JT reports personal financial interest in form of scientific consultancy role for Array Biopharma, AstraZeneca, Bayer, Boehringer Ingelheim, Cardiff Oncology, Chugai, Daiichi Sankyo, F. Hoffmann-La Roche Ltd, Genentech Inc, HalioDX SAS, Hutchison MediPharma International, Ikena Oncology, Inspirna Inc, IQVIA, Lilly, Menarini, Merck Serono, Merus, MSD, Mirati, Neophore, Novartis, Ona Therapeutics, Orion Biotechnology, Peptomyc, Pfizer, Pierre Fabre, Samsung Bioepis, Sanofi, Scandion Oncology, Scorpion Therapeutics, Seattle Genetics, Servier, Sotio Biotech, Taiho, Tessa Therapeutics, TheraMyc and Tolremo Therapeutics. Stocks: Oniria Therapeutics and also educational collaboration with Imedex/HMP, Medscape Education, MJH Life Sciences, PeerView Institute for Medical Education and Physicians Education Resource (PER). The other authors declare no competing interests. Supplementary Files SUPPLEMENTARYFIGURELEGENDS.docx Supplementary Figure Legends SUPPLEMENTARYFIGURE2.pdf SUPPLEMENTARY FIGURE 2 SUPPLEMENTARYFIGURE1.pdf SUPPLEMENTARY FIGURE 1 SUPPLEMENTARYFIGURE4.pdf SUPPLEMENTARY FIGURE 4 SUPPLEMENTARYFIGURE5.pdf SUPPLEMENTARY FIGURE 5 SUPPLEMENTARYFIGURE6.pdf SUPPLEMENTARY FIGURE 6 SUPPLEMENTARYFIGURE9.pdf SUPPLEMENTARY FIGURE 9 SUPPLEMENTARYFIGURE7.pdf SUPPLEMENTARY FIGURE 7 SUPPLEMENTARYFIGURE8.pdf SUPPLEMENTARY FIGURE 8 SUPPLEMENTARYFIGURE3.pdf SUPPLEMENTARY FIGURE 3 SUPPLEMENTARYTABLESS115.xlsx SUPPLEMENTARY TABLES S1-15 Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6770441","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":469383503,"identity":"fe0409ad-3cbe-418c-ae36-d4a5ccc2db42","order_by":0,"name":"Hector Palmer","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYBACAxDxAIj5GXgYDhCvJQGIJRugWniI1mJwAKqWoBZz/rUHPyTm2OQb38g9eOADwx05e0JaLGe8S5ZI3JZmue1GXsLBGQzPjAk77MYZA6CWwwZmZ84YHOZhOJzYQ4QW4x+J2/4bGPdAtNQT1nK+xwxoywEDA/YesJYEIhzGl2aRuC3ZQOJ4j8HBGQaHDXsOELTl7OEbH7fZGfA38xh/+FBxWJ69gZA1EjkoJhBSDgL8Z4hRNQpGwSgYBSMaAABR5UT9bJHB3wAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-9823-5638","institution":"Vall d´Hebron Institute of Oncology","correspondingAuthor":true,"prefix":"","firstName":"Hector","middleName":"","lastName":"Palmer","suffix":""},{"id":469383504,"identity":"d78fe827-ada8-468c-8bbd-bf55436fc3de","order_by":1,"name":"Lorena Ramírez","email":"","orcid":"https://orcid.org/0000-0003-1792-7390","institution":"Vall d'Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Lorena","middleName":"","lastName":"Ramírez","suffix":""},{"id":469383505,"identity":"2ce52d3e-00b9-4f9f-bc26-b14d92daf717","order_by":2,"name":"Javier Ros","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Javier","middleName":"","lastName":"Ros","suffix":""},{"id":469383506,"identity":"0df172b0-d5ab-4f19-ab50-7bf82e761606","order_by":3,"name":"Oriol Arqués","email":"","orcid":"","institution":"Vall d´Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Oriol","middleName":"","lastName":"Arqués","suffix":""},{"id":469383507,"identity":"eddf9e71-700b-425d-9ac5-a20a1a3f2f42","order_by":4,"name":"Jordi Martinez-Quintanilla","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Jordi","middleName":"","lastName":"Martinez-Quintanilla","suffix":""},{"id":469383508,"identity":"1dc9ce61-1da0-4ccd-a7aa-b0626096b766","order_by":5,"name":"Débora Cabot","email":"","orcid":"","institution":"Vall d´Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Débora","middleName":"","lastName":"Cabot","suffix":""},{"id":469383509,"identity":"0b69520f-5858-49e5-9cdd-bbc769b864f3","order_by":6,"name":"Alex Mur-Espinosa","email":"","orcid":"https://orcid.org/0000-0002-2825-2475","institution":"Vall d´Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Alex","middleName":"","lastName":"Mur-Espinosa","suffix":""},{"id":469383510,"identity":"8423afb1-e87a-4e5f-b456-16b6edd85de4","order_by":7,"name":"Irene Chicote","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Irene","middleName":"","lastName":"Chicote","suffix":""},{"id":469383511,"identity":"c7ec5fdb-ab99-4137-be03-ed294fd2a912","order_by":8,"name":"Isabel Puig","email":"","orcid":"","institution":"Vall d´Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Isabel","middleName":"","lastName":"Puig","suffix":""},{"id":469383512,"identity":"e2dea25e-cbb3-42ee-b882-4c3977da6b3f","order_by":9,"name":"Gilles Flamen","email":"","orcid":"","institution":"Vall d'Hebron Institut d'Oncologia","correspondingAuthor":false,"prefix":"","firstName":"Gilles","middleName":"","lastName":"Flamen","suffix":""},{"id":469383513,"identity":"831db15c-7120-4cc9-aed6-e767a9507269","order_by":10,"name":"Ángel de la Torre","email":"","orcid":"","institution":"Vall d'Hebron Institut d'Oncologia","correspondingAuthor":false,"prefix":"","firstName":"Ángel","middleName":"de la","lastName":"Torre","suffix":""},{"id":469383514,"identity":"ff2935c0-7325-4612-8a9a-2d133bcea44a","order_by":11,"name":"Ania Alay","email":"","orcid":"","institution":"Institut Català d’Oncologia","correspondingAuthor":false,"prefix":"","firstName":"Ania","middleName":"","lastName":"Alay","suffix":""},{"id":469383515,"identity":"4a1d2064-d4ab-4e4d-9cd6-faac3d59e627","order_by":12,"name":"Luna Martín","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Luna","middleName":"","lastName":"Martín","suffix":""},{"id":469383516,"identity":"94c63511-a123-4589-8970-0bdc320b6f06","order_by":13,"name":"Raquel Flores","email":"","orcid":"","institution":"Vall d'Hebron Institut d'Oncologia","correspondingAuthor":false,"prefix":"","firstName":"Raquel","middleName":"","lastName":"Flores","suffix":""},{"id":469383517,"identity":"b0e565ac-2857-4265-b687-6708cd19fd0e","order_by":14,"name":"Ariadna García","email":"","orcid":"","institution":"Hospital Universitari Vall d'Hebron","correspondingAuthor":false,"prefix":"","firstName":"Ariadna","middleName":"","lastName":"García","suffix":""},{"id":469383518,"identity":"33d95bab-4d08-4c33-8a0c-c2fbb723c924","order_by":15,"name":"Jose Jimenez","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Jose","middleName":"","lastName":"Jimenez","suffix":""},{"id":469383519,"identity":"73a61afb-eb96-44cd-b53d-4a8c0fe8b2c1","order_by":16,"name":"Veronica Davalos","email":"","orcid":"https://orcid.org/0000-0003-4077-5137","institution":"Josep Carreras Leukemia Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Veronica","middleName":"","lastName":"Davalos","suffix":""},{"id":469383520,"identity":"eb263a49-f26d-4c2c-89ea-5092c734a3cf","order_by":17,"name":"Lara Nonell","email":"","orcid":"","institution":"Vall d'Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Lara","middleName":"","lastName":"Nonell","suffix":""},{"id":469383521,"identity":"7204b71a-90d9-489e-b037-ca987b3ea347","order_by":18,"name":"Xavier Sole","email":"","orcid":"https://orcid.org/0000-0002-2197-3325","institution":"Center for Biomedical Diagnostics (CDB), Hospital Cl\u0026#x00ED","correspondingAuthor":false,"prefix":"","firstName":"Xavier","middleName":"","lastName":"Sole","suffix":""},{"id":469383522,"identity":"7be299ec-07f1-4535-9bda-06153a15b3af","order_by":19,"name":"David Kodack","email":"","orcid":"","institution":"Novartis Biomedical Research","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"Kodack","suffix":""},{"id":469383523,"identity":"2bec8399-8ba3-464e-9c15-897ae834a7d3","order_by":20,"name":"Patrizia Barzaghi-Rinaudo","email":"","orcid":"","institution":"Novartis Biomedical Research","correspondingAuthor":false,"prefix":"","firstName":"Patrizia","middleName":"","lastName":"Barzaghi-Rinaudo","suffix":""},{"id":469383524,"identity":"bb4e71d5-2b42-48f9-a52a-6444f2561c22","order_by":21,"name":"Lisa Kattenhorn","email":"","orcid":"https://orcid.org/0000-0002-9360-8677","institution":"Novartis Biomedical Research","correspondingAuthor":false,"prefix":"","firstName":"Lisa","middleName":"","lastName":"Kattenhorn","suffix":""},{"id":469383525,"identity":"57f78a3c-ed38-4e53-ad20-09bae320f86a","order_by":22,"name":"Slavica Dimitrieva","email":"","orcid":"","institution":"Disease area Oncology, Novartis Institute for Biomedical Research","correspondingAuthor":false,"prefix":"","firstName":"Slavica","middleName":"","lastName":"Dimitrieva","suffix":""},{"id":469383526,"identity":"5944c425-4dcd-4107-b288-9c410895abd2","order_by":23,"name":"Paolo Nuciforo","email":"","orcid":"https://orcid.org/0000-0003-1380-0990","institution":"Vall d’Hebron Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Paolo","middleName":"","lastName":"Nuciforo","suffix":""},{"id":469383527,"identity":"92db0ce3-06a4-43e6-b820-55bbe58de547","order_by":24,"name":"Manel Esteller","email":"","orcid":"https://orcid.org/0000-0003-4490-6093","institution":"Josep Carreras Leukaemia Research Institute (IJC)","correspondingAuthor":false,"prefix":"","firstName":"Manel","middleName":"","lastName":"Esteller","suffix":""},{"id":469383528,"identity":"4e9dfcbf-be4d-4c3c-81dc-7a1d101b669f","order_by":25,"name":"Francesc Canals","email":"","orcid":"","institution":"VHIO","correspondingAuthor":false,"prefix":"","firstName":"Francesc","middleName":"","lastName":"Canals","suffix":""},{"id":469383529,"identity":"70757783-1018-4b96-891a-997938fc8609","order_by":26,"name":"Ana Vivancos","email":"","orcid":"https://orcid.org/0000-0003-2888-6512","institution":"Vall d'Hebrón Institut d´Oncología (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Ana","middleName":"","lastName":"Vivancos","suffix":""},{"id":469383530,"identity":"4b0ea165-570f-4a3c-90de-ebfa615ca398","order_by":27,"name":"Diana Graus Porta","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Diana","middleName":"Graus","lastName":"Porta","suffix":""},{"id":469383531,"identity":"caf716d3-d1b8-4018-8518-7ed797607346","order_by":28,"name":"Josep Tabernero","email":"","orcid":"https://orcid.org/0000-0002-2495-8139","institution":"Vall d'Hebron University Hospital and Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Josep","middleName":"","lastName":"Tabernero","suffix":""},{"id":469383532,"identity":"39205b50-e3aa-4755-918e-3aa311a3c2e0","order_by":29,"name":"Elena Elez","email":"","orcid":"https://orcid.org/0000-0002-4653-6324","institution":"Vall d´Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Elena","middleName":"","lastName":"Elez","suffix":""}],"badges":[],"createdAt":"2025-05-28 18:15:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6770441/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6770441/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85358011,"identity":"d3312fbc-62e7-43f6-abf7-946c27a4d245","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":438768,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePDX \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eBRAF\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cstrong\u003eV600E \u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003emodels recapitulate patients’ response to treatment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e Schematic representation of the total number of patients’ samples analyzed and PDX derived from patients naive (N) or resistant (R) to a BRAFi-based treatment, as well as the number of paired samples from the same patient. \u003cstrong\u003eB,\u003c/strong\u003e On the left side, tumor growth curves of the indicated subcutaneous xenografts from control or triplet treatments, where the red line indicates the period of stabilized disease (SD), n= 7-46 xenografts per condition. Values are presented as mean ± SEM. On the right side, progression-free survival curves (PFS), with the median survival indicated for each group in the same graph. \u003cstrong\u003eC, \u003c/strong\u003eSchematic graph illustrating the experimental design of the study, indicating the specific time points at which the samples were collected: cST (control short time), cEP (control end point), tST (triplet short time), tLT (triplet long time), tEP (triplet end point). Additionally, the different comparisons conducted between the various groups are detailed. \u003cstrong\u003eD,\u003c/strong\u003e Alluvial plot representing the proportion of gene sets differentially enriched in each treatment time point. \u003cstrong\u003eE,\u003c/strong\u003e Gene set enrichment analysis (GSEA) results for representative gene sets in selected categories at LT, except mucinous differentiation that is at EP. T-Test analyses were performed to compare the area under the curve in left panel \u003cstrong\u003eB\u003c/strong\u003e. The log-rank (Mantel–Cox) test was used to compare survival curves among groups in right \u003cstrong\u003eB\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/ee97b82da20954760661485b.png"},{"id":85359096,"identity":"3f43f793-3cb9-41b4-b029-1eb4a9367832","added_by":"auto","created_at":"2025-06-25 05:40:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":392939,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMultiomic integration and methylation status upon triplet treatment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e Volcano plot representing the proteins downregulated (in blue) or upregulated (in red) in the M64-N (naive) model (n=6). \u003cstrong\u003eB,\u003c/strong\u003e Venn diagram intersection comparing proteins (violet circle) and genes (green cercle) up or downregulated by the triplet in M64-N at EP (n=6). \u003cstrong\u003eC\u003c/strong\u003e, proteins and genes commonly up or downregulated by the triplet. \u003cstrong\u003eD,\u003c/strong\u003e Buble plot depicting the top 50 gene sets obtained by the pre-ranked Gene Set Enrichment Analysis (GSEA) using the hallmark, C2, C5 and C6 gene set collections from the molecular signatures database (n=6). \u003cstrong\u003eE,\u003c/strong\u003ePrincipal component analysis (PCA) of methylation patterns in the M64-N model control or treated at different time points. \u003cstrong\u003eF,\u003c/strong\u003e Bar plot depicting the relative frequency of the different genomic regions showing methylation changes over time after treatment in the M64-N model. cST: Control short time, cEP: Control end point, tST: Triplet short time, tLT: Triplet long time, tEP: Triplet end point. \u003cstrong\u003eG,\u003c/strong\u003e Scatter plot showing the correlation between gene expression and methylation levels of the corresponding promoters (TSS200) at EP in the M64-N model.\u003cstrong\u003e \u003c/strong\u003eRed squares highlight demethylated and higher expressed genes upon triplet treatment. Some genes following this patter are indicated in red. \u003cstrong\u003eH,\u003c/strong\u003e Venn diagram representing the overlap of the genes which promoters (TSS200) are hypomethylated and overexpressed after the treatment in all models at EP. \u003cstrong\u003eI,\u003c/strong\u003e Bar plot showing the evolution of methylation status in naive and resistant models for genes that are exclusively hypomethylated and overexpressed in the naive models at EP (n = 24 genes). \u003cstrong\u003eJ, \u003c/strong\u003eLine plot showing the methylation levels in the different probes for the \u003cem\u003eMUC20\u003c/em\u003e gene in the indicated PDX treated and non-treated. Significant differentially methylated position (DMP) (adjusted \u003cem\u003ep value\u003c/em\u003e \u0026lt; 0.05) is indicated with a red line and box.\u003c/p\u003e","description":"","filename":"fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/80c1c4b930e568cb5716a89c.png"},{"id":85359615,"identity":"49862e57-5d62-4d0c-a751-b7819d3b4f15","added_by":"auto","created_at":"2025-06-25 05:48:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":653874,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMAPK pathway inhibition promotes tissue cell remodeling in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eBRAF\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cstrong\u003eV600E\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e CRC tumors.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e UMAP showing the distribution of treated and non-treated cells based on single-cell RNA-seq (scRNAseq). \u003cstrong\u003eB,\u003c/strong\u003e UMAP showing the detailed distribution of treated and non-treated endothelial cells based on scRNAseq results. \u003cstrong\u003eC,\u003c/strong\u003e Bar plot showing the percentage of different cell types present in treated and non-treated tumors based on scRNAseq results. \u003cstrong\u003eD,\u003c/strong\u003e UMAP showing the expression of the indicated genes in tumor cells based on scRNAseq results. \u003cstrong\u003eE,\u003c/strong\u003e Violin plot showing the expression levels of \u003cem\u003ePROM1\u003c/em\u003e in treated and non-treated tumor cells based on scRNAseq results, ****\u003cem\u003ep\u003c/em\u003e ≤ 0.0001.\u003cstrong\u003e G,\u003c/strong\u003e UMAP showing treated and non-treated human tumor cells expressing \u003cem\u003eMUC4\u003c/em\u003e (in red), \u003cem\u003ePROM1\u003c/em\u003e (in blue), and co-expression (in green) based on scRNAseq results. \u003cstrong\u003eG,\u003c/strong\u003e Bar plot showing the absolute number of treated and non-treated human tumor cells expressing \u003cem\u003eMUC4\u003c/em\u003e (in red), \u003cem\u003ePROM1 \u003c/em\u003e(in blue), and co-expression (in green) based on scRNAseq results. C: Control; T: Triplet for panels \u003cstrong\u003eC\u003c/strong\u003e and \u003cstrong\u003eG\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/902a706f69f307a316facf34.png"},{"id":85358016,"identity":"4efc38ad-ef7b-46c2-838b-9e674200d560","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":3775946,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMAPK pathway inhibition promotes a mucinous differentiation of \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eBRAF\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cstrong\u003eV600E\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e CRC tumor cells required for acquisition of resistance to treatment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e Hierarchical cluster showing the expression of genes related to mucinous differentiation in indicated subcutaneous xenografts\u003cstrong\u003e. B,\u003c/strong\u003e Basal expression of \u003cem\u003eATOH1\u003c/em\u003e in control animals at short time (cST) in the indicated subcutaneous xenografts, evaluated by qRT-PCR (n= 3). Mean ± SEM.\u003cstrong\u003e C, \u003c/strong\u003eExpression of \u003cem\u003eATOH1\u003c/em\u003e of the indicated subcutaneous xenografts at short time (ST), long time (LT) and end point (EP), evaluated by qRT-PCR (n= 3). Mean ± SEM. \u003cstrong\u003eD,\u003c/strong\u003e Representative pictures of MUC2 immunohistochemical staining in subcutaneous xenografts. Scale bar, 100 μm. \u003cstrong\u003eE,\u003c/strong\u003e Representative pictures of alcian blue staining in subcutaneous xenografts. Scale bar, 100 μm. Mucin (M). Arrowhead point lining goblet-like cells. \u003cstrong\u003eF,\u003c/strong\u003e Quantification of the mucinous differentiation at cST of the indicated subcutaneous xenografts, evaluated by alcian blue staining (n= 6 per PDX). Mean ± SEM. \u003cstrong\u003eG,\u003c/strong\u003e Quantification of mucinous differentiation of the indicated subcutaneous xenografts at ST, LT and EP, evaluated by alcian blue staining (n= 3-6). Mean ± SEM. \u003cstrong\u003eH,\u003c/strong\u003e Representative images of mucinous differentiation evaluated by alcian blue staining in patient #9 under basal conditions and at progression to BRAF inhibitors. Scale bar, 100 μm. \u003cstrong\u003eI,\u003c/strong\u003e Quantification of mucinous differentiation evaluated by alcian blue staining in human samples prior, during or at progression to BRAFi treatment. \u003cstrong\u003eJ, \u003c/strong\u003eTumor growth curves of the indicated subcutaneous xenografts from control or triplet treatments condition. Values are presented as mean ± SEM (n=10). \u003cstrong\u003eK,\u003c/strong\u003e Representative images illustrating proliferation evaluated by immunohistochemical Ki67 and mucinous differentiation examined through alcian blue staining in subcutaneous xenografts. Scale bar, 100 μm. r.u.:\u003cstrong\u003e \u003c/strong\u003erelative units for panels \u003cstrong\u003eB\u003c/strong\u003e, \u003cstrong\u003eC, F\u003c/strong\u003e and \u003cstrong\u003eG\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/b19cad4a9fc1c3f1e54754b7.png"},{"id":85359098,"identity":"62b98e50-db7f-436f-b44c-1241be0f687e","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1769725,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMAPK pathway inhibition promotes stroma remodeling in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eBRAF\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cstrong\u003eV600E \u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eCRC tumors.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e Bubble plot representing gene sets classified as tissue remodeling clusters for naive and resistant PDX models at short time (ST), long time (LT) and end point (EP). NES: Normalized Enrichment Score. \u003cstrong\u003eB,\u003c/strong\u003e Genes from the leading edge of the gene sets indicated in panel \u003cstrong\u003eA\u003c/strong\u003e are represented as a heatmap, again from PDX treated at ST, LT and EP. \u003cstrong\u003eC,\u003c/strong\u003e Basal expression of \u003cem\u003eMMP7\u003c/em\u003e (top) and \u003cem\u003eCOL1A1\u003c/em\u003e (bottom)\u003cem\u003e \u003c/em\u003eat short time (cST) of the indicated subcutaneous xenografts, evaluated by qRT-PCR (n= 3). Mean ± SEM. \u003cstrong\u003eD,\u003c/strong\u003e Expression of \u003cem\u003eMMP7\u003c/em\u003e (top) and \u003cem\u003eCOL1A1\u003c/em\u003e (bottom)\u003cem\u003e \u003c/em\u003eof the indicated subcutaneous xenografts at ST, LT and EP, evaluated by qRT-PCR (n= 3). Mean ± SEM. \u003cstrong\u003eE,\u003c/strong\u003e Western blot of MMP7\u003cem\u003e \u003c/em\u003ein the indicated subcutaneous xenografts at ST, and EP. GAPDH was used as a loading control. \u003cstrong\u003eF, \u003c/strong\u003eRepresentative pictures of collagen fibers stained with Sirius red at EP in subcutaneous xenografts. Scale bar, 100 μm. \u003cstrong\u003eG,\u003c/strong\u003e Quantification of the Sirius red area at cST of the indicated subcutaneous xenografts (n= 6 per PDX). Mean ± SEM. \u003cstrong\u003eH,\u003c/strong\u003e Quantification of the Sirius red area of the indicated subcutaneous xenografts at ST, LT and EP (n= 3-6). Mean ± SEM. \u003cstrong\u003eI, \u003c/strong\u003eBasal quantification of the percentage of high-density matrix (HDM) at cST of the indicated subcutaneous xenografts, evaluated by Sirius red staining (n= 6 per PDX). Mean ± SEM. \u003cstrong\u003eJ,\u003c/strong\u003e Quantification of the percentage of HDM of the indicated subcutaneous xenografts at ST, LT and EP, evaluated by Sirius red staining (n= 3-6). Mean ± SEM. r.u.: relative units for panels \u003cstrong\u003eC\u003c/strong\u003e and \u003cstrong\u003eD\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/e9062cf353c70bf1d3980355.png"},{"id":85359099,"identity":"1a50f38a-f384-4f88-bed0-dcc449a220c0","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1021893,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMAPK pathway inhibition promotes the expression of vasculature-related genes and the extracellular accumulation of VEGFA.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA, \u003c/strong\u003eBubble plot representing NES and adjusted \u003cem\u003ep value\u003c/em\u003e at different time points for naive and resistant models for gene sets classified as vasculature clusters.\u003cstrong\u003e B, \u003c/strong\u003eAmong the genes present in the leading-edge analysis of gene sets in panel \u003cstrong\u003ea\u003c/strong\u003e, normalized expression levels of those differentially expressed in naive models are shown in a heatmap.\u003cstrong\u003e C, \u003c/strong\u003eBasal expression of \u003cem\u003ePROM1 \u003c/em\u003eat short time (cST) of the indicated subcutaneous xenografts, evaluated by qRT-PCR (n= 3). Mean ± SEM.\u003cstrong\u003e D, \u003c/strong\u003eExpression of \u003cem\u003ePROM1 \u003c/em\u003eof the indicated subcutaneous xenografts at short time (ST), long time (LT) and end point (EP), evaluated by qRT-PCR (n= 3). Mean ± SEM.\u003cstrong\u003e E, \u003c/strong\u003eWestern blot detecting VEGFA in the supernatant (SN) and Matrigel (M) fraction of HT29 cell line cultured in 3D embedded in Matrigel. Ponceau was used as a loading control. Enco: encorafenib.\u003cstrong\u003e F \u003c/strong\u003eBar plots representing VEGFA protein quantification from the western blot in panel \u003cstrong\u003eE\u003c/strong\u003e. r.u.: relative units in panels \u003cstrong\u003eC\u003c/strong\u003e, \u003cstrong\u003eD\u003c/strong\u003e and \u003cstrong\u003eF\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/9c5b64374c024d70f79e0119.png"},{"id":85359103,"identity":"7cc33b03-2471-4cd9-b70a-c289057448fa","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1671752,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVEGFA blockade with bevacizumab improves BRAFi therapies in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eBRAF\u003c/strong\u003e\u003c/em\u003e\u003csup\u003e\u003cstrong\u003eV600E \u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eCRC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA,\u003c/strong\u003e Representative images of Microvessel Density (MVD) quantification in CRC \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e PDX at cEP. Scale bar, 100 μm. \u003cstrong\u003eB,\u003c/strong\u003e Exemplifying pictures of CD31 immunostaining at end point (EP) in representative subcutaneous xenografts. Scale bar, 100 μm. Arrowheads point CD31-positive endothelial cells. \u003cstrong\u003eC,\u003c/strong\u003e MVD quantification at basal levels at short time (cST) of the indicated subcutaneous xenografts (n= 6). Mean ± SEM. \u003cstrong\u003eD,\u003c/strong\u003e MVD of the indicated subcutaneous xenografts at short time ST, long time (LT) and end point EP (n= 3-6). Mean ± SEM. \u003cstrong\u003eE,\u003c/strong\u003e Quantification of MVD evaluated in human biopsy samples prior, during or at progression to treatment with BRAFi-based therapies. \u003cstrong\u003eF,\u003c/strong\u003e On the left side, tumor growth curves of HT29 subcutaneous xenografts (n= 8-32). Presented as Mean ± SEM. In the middle, progression-free survival curves (PFS) of the indicated groups, with the median survival indicated for each group within the same graph. On the right side, results of the statistical analysis for the PFS. \u003cstrong\u003eG,\u003c/strong\u003e On the left side, tumor growth curves of the indicated subcutaneous xenografts (n= 7-56). Presented as Mean ± SEM. In the middle, PFS curves of the indicated groups, with the median survival indicated for each group within the same graph. On the right side, results of the statistical analysis for the PFS.\u003c/p\u003e","description":"","filename":"fig7.png","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/6217973566b010d45a96de44.png"},{"id":85360396,"identity":"9528f441-92b8-4733-9454-827224fbc5bd","added_by":"auto","created_at":"2025-06-25 05:56:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":11811242,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/43cd28cf-14c0-4f06-b198-b6dc9069bd23.pdf"},{"id":85358010,"identity":"291bac25-46be-4a63-ba9a-e9b88df2c12d","added_by":"auto","created_at":"2025-06-25 05:32:03","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":23670,"visible":true,"origin":"","legend":"Supplementary Figure Legends","description":"","filename":"SUPPLEMENTARYFIGURELEGENDS.docx","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/c8616239744d74162bafd772.docx"},{"id":85358013,"identity":"35532e8e-2c76-446e-a003-726c8e78a8fd","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":318683,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 2","description":"","filename":"SUPPLEMENTARYFIGURE2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/c1121bfc9aea9dbaee7733fa.pdf"},{"id":85359101,"identity":"235d8b81-0a86-44bc-81ef-b4c21e9b2d15","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":273816,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 1","description":"","filename":"SUPPLEMENTARYFIGURE1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/7954f32c3f7800d592a37a11.pdf"},{"id":85359102,"identity":"a8053b63-2101-492f-8c93-b548503711ee","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":595109,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 4","description":"","filename":"SUPPLEMENTARYFIGURE4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/dfe8ca89d3458d1a214dd25a.pdf"},{"id":85359104,"identity":"2b764956-b0e4-420b-86f7-05aaf853d3d8","added_by":"auto","created_at":"2025-06-25 05:40:04","extension":"pdf","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":180346,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 5","description":"","filename":"SUPPLEMENTARYFIGURE5.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/f565af98d074d5e21b912351.pdf"},{"id":85358025,"identity":"89523b64-c354-40a9-8060-8de3230737d8","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"pdf","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":401403,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 6","description":"","filename":"SUPPLEMENTARYFIGURE6.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/7667d2341fa32b9ed6acab7c.pdf"},{"id":85359616,"identity":"9df5830f-0adf-4e56-b563-eff19dca6d72","added_by":"auto","created_at":"2025-06-25 05:48:04","extension":"pdf","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":226554,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 9","description":"","filename":"SUPPLEMENTARYFIGURE9.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/86571c06fe1d2280edfda56f.pdf"},{"id":85358032,"identity":"32dcef57-81ff-4761-9813-a87b51fc125e","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"pdf","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":259513,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 7","description":"","filename":"SUPPLEMENTARYFIGURE7.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/178de23fc5643ff92fc1f068.pdf"},{"id":85358020,"identity":"09240c7b-29ee-4622-b3e2-82067b364fe1","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"pdf","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":299797,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 8","description":"","filename":"SUPPLEMENTARYFIGURE8.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/f4ca04a4c42b4e3f9b5df100.pdf"},{"id":85358021,"identity":"460ca7ba-1a38-45b5-85bc-4dc326118979","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"pdf","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":355230,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURE 3","description":"","filename":"SUPPLEMENTARYFIGURE3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/1a8c76e19c55543773f9f8fe.pdf"},{"id":85358026,"identity":"1ef082df-1a97-467b-b610-18e890ff1137","added_by":"auto","created_at":"2025-06-25 05:32:04","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":3955428,"visible":true,"origin":"","legend":"SUPPLEMENTARY TABLES S1-15","description":"","filename":"SUPPLEMENTARYTABLESS115.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6770441/v1/805bd2870ae3f6ada0ebf4c5.xlsx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nLR, OA, JMQ, DC, AME, IC, IP, GF, RF and HGP report financial support from Cyclacel Pharmaceuticals, Boundless Bio, Grifols, Roche, AstraZeneca and Vivan Therapeutics outside the submitted work. I.P. is cofounder of Oniria Therapeutics S.L. and reports financial support from this company outside the submitted work. HGP is a CSO, co-Founder \u0026 Board Member from Oniria Therapeutics S.L and reports financial support from this company outside the submitted work. JR has received personal speaker honoraria from Sanofi and Amgen, and accommodation expenses from Pierre-Fabre, Servier, Amgen and Merck. PN discloses personal financial interests receiving honoraria or consultation fees from Novartis, Bayer, Targos Molecular Pathology GmbH and MSD Oncology; and receiving travel, accommodation paid or reimbursed by Novartis.EE discloses personal financial interests receiving honoraria for advisory role, travel grants and research grants from Amgen, Bayer, Hoffman-La Roche, Merck Serono, Sanofi, Pierre Fabre, MSD, Organon, Novartis and Servier. JT reports personal financial interest in form of scientific consultancy role for Array Biopharma, AstraZeneca, Bayer, Boehringer Ingelheim, Cardiff Oncology, Chugai, Daiichi Sankyo, F. Hoffmann-La Roche Ltd, Genentech Inc, HalioDX SAS, Hutchison MediPharma International, Ikena Oncology, Inspirna Inc, IQVIA, Lilly, Menarini, Merck Serono, Merus, MSD, Mirati, Neophore, Novartis, Ona Therapeutics, Orion Biotechnology, Peptomyc, Pfizer, Pierre Fabre, Samsung Bioepis, Sanofi, Scandion Oncology, Scorpion Therapeutics, Seattle Genetics, Servier, Sotio Biotech, Taiho, Tessa Therapeutics, TheraMyc and Tolremo Therapeutics. Stocks: Oniria Therapeutics and also educational collaboration with Imedex/HMP, Medscape Education, MJH Life Sciences, PeerView Institute for Medical Education and Physicians Education Resource (PER). The other authors declare no competing interests.","formattedTitle":"Acquired resistance to BRAF inhibitory treatments requires tumor tissue remodeling and reveals targetable vulnerabilities in colorectal cancer","fulltext":[{"header":"Main","content":"\u003cp\u003eColorectal cancer (CRC) is a significant global health concern, with approximately 1.9\u0026nbsp;million new cases and 935,173 deaths in 2020(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). About 10% of CRC patients harbor the \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e mutation(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e), linked to poor prognosis and limited treatment options(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan additionalcitationids=\"CR5 CR6\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). This mutation, resulting from a thymine-to-adenine change(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e), leads to constitutive BRAF activation and downstream MAPK signaling, promoting tumor proliferation and survival(\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWhile BRAFi inhibitors (BRAFi) monotherapies show efficacy in melanoma and other tumors(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e), their effectiveness in CRC is hindered by rapid feedback activation of EGFR(\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Clinical trials, including BEACON, demonstrated that combining EGFR inhibitors (EGFRi) with BRAFi with or without MEK inhibitors (MEKi) improves outcomes in metastatic CRC (mCRC)(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). However, resistance to these combinations often develops, necessitating an understanding of its mechanisms.\u003c/p\u003e \u003cp\u003eKnown resistance mechanisms can involve genetic alterations that reactivate the MAPK pathway, including \u003cem\u003eKRAS\u003c/em\u003e and \u003cem\u003eMAP2K1\u003c/em\u003e mutations, \u003cem\u003eBRAF\u003c/em\u003e and \u003cem\u003eMET\u003c/em\u003e amplifications and others(\u003cspan additionalcitationids=\"CR14 CR15 CR16 CR17 CR18 CR19 CR20\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). However, approximately 40% of resistant cases lack these oncogenetic changes(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e), indicating the relevance of non-genetic mechanisms of resistance. Our study utilizing patient-derived xenografts (PDX) reveals that resistance to combined BRAFi therapies can arise from non-genetic factors linked to acute tissue remodeling, such as mucinous differentiation, extracellular matrix (ECM) remodeling, increased immune stroma(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e) and vasculature.\u003c/p\u003e \u003cp\u003eWe observed that blocking angiogenesis with bevacizumab in PDX models extended combined BRAFi therapy efficacy, delaying resistance and enhancing progression-free survival (PFS). This underscores the importance of understanding resistance mechanisms, as patients with limited options after BRAF-targeted therapies need new strategies.\u003c/p\u003e \u003cp\u003eOur findings provide insights into the processes underlying resistance to combined BRAFi therapies in mCRC and support the initiation of BRAVE, a phase I/II clinical trial (NCT06411600) to evaluate bevacizumab in combination with encorafenib (BRAFi) and cetuximab (EGFRi). In summary, we reveal that \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e mCRC can develop resistance through cell plasticity and tissue remodeling, suggesting new rational combinations for improving treatment outcomes in advanced CRC patients.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003ePDX models effectively mimic the development of resistance to combined BRAFi therapy in\u003c/strong\u003e \u003cstrong\u003eBRAF\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003eV600E\u003c/strong\u003e\u003c/sup\u003e \u003cstrong\u003emCRC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll treatments with a BRAFi in combination with other drugs such as EGFRi, MEKi or others were indicated as BRAFi therapies for simplification purposes, since all have a common mutant BRAF inhibition strategy.\u003c/p\u003e\n\u003cp\u003eWe established 32 PDX models from 22 patients, including 17 derived from patients naive (PDX-N) and 15 resistant (PDX-R) upon progression to combined BRAFi therapies (three of them were derived from paired biopsies) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eA and Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e). Whole Exome Sequencing (WES) revealed that two resistant patients exhibited oncogenic subclonal \u003cem\u003eKRAS\u003c/em\u003e mutations, a known resistance mechanism(\u003cspan class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e) (Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e). The mutational landscape of our PDX models was consistent with TCGA \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003e\u003cem\u003eV600E\u003c/em\u003e\u003c/sup\u003e CRC tumors, featuring mutations in \u003cem\u003eAPC\u003c/em\u003e, \u003cem\u003eTTN\u003c/em\u003e, and \u003cem\u003eTP53\u003c/em\u003e (Supplementary Fig. \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003eA and Supplementary Table S2), but lacked exclusive oncogenic mutations distinguishing resistant from naive models. We only derived one PDX model with microsatellite instability (MSI) from a patient who progressed to combined BRAFi therapy because they were preferably enrolled in immunotherapy treatments (Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eThe analysis of single base substitution (SBS) COSMIC signatures also indicated similar profiles between naive and resistant models and a low incidence of mismatch repair deficiency (MMRd) signatures and MSI due to priority selection for immunotherapy treatment (Supplementary Fig. \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003eB). Our sequencing results indicate that our models recapitulate human CRC genetics (TCGA-COREAD) and suggest that resistance to BRAF inhibitors could be driven by non-genetic mechanisms rather than genetic alterations.\u003c/p\u003e\n\u003cp\u003eTo explore this, we selected naive (CTAX2-N, M64-N) and resistant (CTAX90-R, CTAX91-R) PDX models. Aiming to use a balanced set of models we considered that all were derived from patients treated with encorafenib\u0026thinsp;+\u0026thinsp;cetuximab backbone and benefited more than 120 days (Supplementary Table\u0026nbsp;1). We also balanced the model selection by using one MSI and one with microsatellite stability (MSS) for each na\u0026iuml;ve or resistant.\u003c/p\u003e\n\u003cp\u003eFor our initial discovery phase, we combined encorafenib (BRAFi), cetuximab (EGFRi) and binimetinib (MEKi) in most of our experiments which is indicated as triplet treatment (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eB). While the standard treatment for patients in Europe and USA is a BRAFi\u0026thinsp;+\u0026thinsp;EGFRi doublet, our results with PDX indicated that only a triple combination adding a MEKi produced an initial prolonged response phase followed by a progression whereas xenografts treated with the doublet although slowed proliferation, progressed early as non-treated control animals (Kapplan Meier plots) (Supplementary Fig. S2A). In fact, the triplet has been adopted as the Standard-of-Care (SoC) in Japan because of its higher overall response rate in \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003e\u003cem\u003eV600E\u003c/em\u003e\u003c/sup\u003e CRC patients(\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e) (Supplementary Fig. S2B). Importantly, the most significant discoveries in our study using the triplet were also observed using doublet treatments in PDX and patients.\u003c/p\u003e\n\u003cp\u003eBoth naive models initially showed robust responses but subsequently acquired resistance (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eB). Importantly, in both naive models, the treatment blocked tumor growth for the initial 20\u0026ndash;35 days. In contrast, treatment delayed the growth rate of resistant PDXs, but all tumors progressed soon after treatment initiation without showing a stabilized disease phase. Interestingly, the triplet treatment equally blocked the MAPK signaling in tumors of both naive and resistant PDX models, as observed by the inhibition of pathway target genes \u003cem\u003eDUSP6\u003c/em\u003e and \u003cem\u003eSPRY4\u003c/em\u003e expression (Supplementary Fig. S2C).\u003c/p\u003e\n\u003cp\u003eWe observed that both naive PDX tumors in different mice treated with the triplet acquired resistance almost simultaneously after a response phase (Supplementary Fig. S2D) and WES did not show the appearance of any oncogenic mutation that could explain such resistance (Supplementary Fig. S2E and Supplementary Table S3).\u003c/p\u003e\n\u003cp\u003eIn summary our results suggested that our PDX models recapitulate patients\u0026acute;s genomics and their response to treatment and that non-genetic mechanisms could drive the acquisition of resistance to BRAFi therapies in mCRC instead of the outgrowth of a selected genetic clone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMAPK oncogenic signal blockade induced significant molecular reprogramming in\u003c/strong\u003e \u003cstrong\u003eBRAF\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003eV600E\u003c/strong\u003e\u003c/sup\u003e \u003cstrong\u003eCRC tumor cells and surrounding stroma\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe analyzed whole gene expression in naive and resistant PDX models treated with a triplet therapy at various time points (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eC and Supplementary Fig. S3A, Supplementary Table S4). Treatment inhibited proliferation and triggered distinct responses: naive models exhibited rapid interferon-based immune activation(\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e), while resistant models engaged in drug detoxification (Supplementary Fig. S3A and S3B). Gene set enrichment analysis (GSEA) indicated positive enrichment of hypoxia, vasculature, mucinous differentiation and tissue remodeling pathways in naive models, while proliferation-related pathways were negatively impacted in both (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eD, \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eE, Supplementary Fig. S3B, Supplementary Table S5). Interestingly, we observed that this switch was deeper in naive than in resistant models when acquired resistance (end points), as observed by the appearance of a larger variety of biological processes (gene sets) positively enriched (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eD).\u003c/p\u003e\n\u003cp\u003eAiming to mine deeper in understanding this non-genetic process of tissue plasticity we selected na\u0026iuml;ve M64-N model as the most sensitive to triplet treatment for further detailed investigations. Subsequent proteomic analysis of M64-N confirmed the induction of proteins related to mucinous differentiation, immune response, vasculature, RNA processing, tissue remodeling or EGFR signaling blockade, alongside repression of protein translation and proliferation-associated proteins (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eA-D and Supplementary Table S6 and S7).\u003c/p\u003e\n\u003cp\u003eBRAF\u003csup\u003eV600E\u003c/sup\u003e oncogenic signaling was previously described to control gene expression by modulating genome methylation(\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e). Methylome analysis showed progressive reprogramming of DNA methylation upon triplet treatment in all models, with hypomethylation occurring earlier in naive than resistant models, particularly in regions controlling gene expression (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eE-G, Supplementary Fig. S4A-D and Supplementary Table S8). This hypomethylation correlated with increased gene expression prior to resistance acquisition (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eG-I and Supplementary Fig. S4C, S4D and Supplementary Table S9). All our samples presented a general methylome profile distinctive of TCGA-COREAD (Supplementary Fig. S4A). \u003cem\u003eMUC20\u003c/em\u003e gene is a clear example, where specific CpG sites in its promoter were hypomethylated and showed increased expression upon triplet treatment (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eJ).\u003c/p\u003e\n\u003cp\u003eBRAF inhibition in CRC can promote a cancer cell intrinsic biological change that is heterogenous among cells within the same tumor(\u003cspan class=\"CitationRef\"\u003e26\u003c/span\u003e). For reveling this potential complexity, we performed single-cell RNA sequencing (scRNAseq) in the most triplet sensitive naive model M64-N at endpoint treatment. We revealed marked changes in gene expression profiles among tumor and stroma cells following triplet treatment involving an increased infiltration of immune cells, fibroblasts and endothelial cells confirming significant tissue remodeling (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eA-C). We observed the modulation of specific genes by the triplet in tumor cells observed by bulk analyses, such as the repression of \u003cem\u003eDUSP6\u003c/em\u003e as MAPK target gene or induction of metalloprotease \u003cem\u003eMMP7\u003c/em\u003e involved in tissue remodeling and \u003cem\u003eMUC4\u003c/em\u003e in mucinous differentiation (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD). Moreover, treatment led to an increase of stem cell marker \u003cem\u003ePROM1\u003c/em\u003e/CD133 which has been described to stabilize the angiogenic ligand VEGFA(\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD and \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eE). We observed increase in the \u003cem\u003ePROM1\u003c/em\u003e/CD133\u0026thinsp;+\u0026thinsp;cancer stem cell population(\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e29\u003c/span\u003e) and in mucinous differentiated cells (MUC4) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD-G), indicating a complex interplay between progenitor and differentiated tumor cells during resistance development.\u003c/p\u003e\n\u003cp\u003eAn independent laboratory performed spatial transcriptomics in two extra \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003e\u003cem\u003eV600E\u003c/em\u003e\u003c/sup\u003e CRC PDX models treated with dabrafenib (BRAFi), cetuximab (EGFRi) and trametinib (MEKi). We observed a nice separation of treated versus not treated areas by UMAP analyses and a clear enrichment of infiltrated mouse stroma upon triplet treatment (Data not shown). Interestingly, markers of mucinous differentiation (\u003cem\u003eMUC4\u003c/em\u003e, \u003cem\u003eCLCA\u003c/em\u003e), CAF infiltration (\u003cem\u003eActa2\u003c/em\u003e), tissue remodeling (\u003cem\u003eCOL1A1\u003c/em\u003e) and others relevant for the drug-induced phenotype such as \u003cem\u003ePROM1\u003c/em\u003e were increased in the tissue upon triplet treatment. Again, \u003cem\u003eVEGFA\u003c/em\u003e was not regulated at mRNA level, whereas others such as \u003cem\u003eOLFM4\u003c/em\u003e was again induced by the triplet.\u003c/p\u003e\n\u003cp\u003eIn conclusion, our results suggested that MAPK blockade regardless of the specific drugs (dabrafenib vs encorafenib, binimetinib vs trametinib), promoted a multifaceted molecular shift in \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e CRC, fostering an environment conducive to resistance through epigenetic reprogramming, tumor cell mucinous differentiation and stroma remodeling.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBRAF inhibitory therapy promotes a mucinous differentiation of tumor cells that is required for acquiring resistance to treatment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis was confirmed in PDX models, where a significant increase in mucus production was observed macroscopically after treatment, and driven by transcription factor ATOH1 and mucins upregulation (\u003cem\u003eMUC1\u003c/em\u003e, \u003cem\u003eMUC2\u003c/em\u003e, \u003cem\u003eMUC4\u003c/em\u003e, \u003cem\u003eMUC13\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eA-C, Supplementary Fig. S5A, S5B and Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e0). ATOH1 is a transcription factor key for directing intestinal stem cells towards goblet cell terminal differentiation (Supplementary Fig. S5C). This differentiation led to the accumulation of goblet-like cells that secreted mucins (alcian blue), measured as MUC2-positive cells (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eD-G and Supplementary Fig. S6A), being \u003cem\u003eMUC2\u003c/em\u003e a direct ATOH1 target gene(\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e32\u003c/span\u003e). In collaboration, another independent team observed in parallel the same progressive mucinous differentiation phenotype in seven \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e PDX models (Supplementary Fig. S6B-D). Interestingly, this blind validation involved different MAPK inhibitory drugs: dabrafenib (BRAFi), trametinib (MEKi), and cetuximab (EGFRi). Futhermore, increased mucin levels were detected in five out of ten patients when comparing baseline biopsies to those taken upon disease progression under BRAFi-combined therapy (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eH, \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eI and Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e). While cetuximab resistance has been associated with Paneth cell differentiation in CRC(\u003cspan class=\"CitationRef\"\u003e33\u003c/span\u003e), gene expression profiling ruled out this lineage differentiation in triplet-treated PDX models (Supplementary Fig. S6E and Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e1).\u003c/p\u003e\n\u003cp\u003eKnockout experiments targeting mucin maturation enzymes (\u003cem\u003eGCNT1/GCNT3\u003c/em\u003e) in HT29 CRC cell line (Supplementary Fig. S6F), revealed that mucin maturation driven by triplet treatment, was required for resistance acquisition (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eJ). We observed that double-KO cells grew subcutaneously in NOD-SCID immunodeficient mice as rapidly as their wild type (WT) counterparts. However, any double-KO tumor was capable to progress and acquire resistance to triplet treatment as WT did (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eJ and Supplementary Fig. S6G and S6H). Double-KO tumors still commit to mucinous differentiation and produce immature mucins (alcian blue) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eK), however a reduction of proliferation upon triplet treatment was observed in KO but not in WT tumors (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eK and Supplementary Fig. S6I).\u003c/p\u003e\n\u003cp\u003eIn conclusion, we discovered a drastic increase in mucin production upon triplet treatment and mucin maturation as a key factor for acquiring resistance to BRAFi therapies.\u003c/p\u003e\n\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003cp\u003e\u003cstrong\u003eMAPK inhibition induces extracellular matrix (ECM) remodeling\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eWe observed an enrichment of tissue remodeling genes, such as metalloproteases (\u003cem\u003eMMP7\u003c/em\u003e), cathepsins (\u003cem\u003eCTSS\u003c/em\u003e), and complement components (\u003cem\u003eCFB\u003c/em\u003e, \u003cem\u003eC1S\u003c/em\u003e, \u003cem\u003eSERPING1\u003c/em\u003e), in naive, but less in resistant PDX models that was more acute at long term treatment before acquisition of resistance (end point) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eA, \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eB and Supplementary Table S12 and S13). MMP7 expression was notably elevated in naive models by triplet treatment (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eB-E). \u003cem\u003eMMP7\u003c/em\u003e mRNA and protein were particularly induced by triplet in naive but not in resistant PDX models, which already presented higher basal levels (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eC-E and Supplementary Fig. S7A and S7B). This molecular shift also increased collagen deposition (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eC-J) and cancer associated fibroblasts (CAFs) infiltration, contributing to a distinctive ECM phenotype (Supplementary Fig. S7C and S7D).\u003c/p\u003e\n \u003cp\u003eWe could conclude that triplet treatment promoted an early ECM remodeling that was consolidated on persistent small tumors before restart tumor growth. Since this is a complex process involving a coordinated network of factors, we could not perform an intervention experiment to prove its requirement for resistance acquisition. Indeed, we knocked down \u003cem\u003eMMP7\u003c/em\u003e in HT29 cells and it was not enough to block the acquisition of resistance to triplet treatment (data not shown).\u003c/p\u003e\n \u003cp\u003eFurther analyses were preformed to understand tumor cell signaling consequences upon MAPK signaling blockaded. We confirmed that triplet therapy blocked MAPK signaling reducing phospho-ERK. We also observed a decrease of GSK-3\u0026beta; phosphorylation at Tyr216, a target site of MEK1 kinase, required for its activity (\u003cspan class=\"CitationRef\"\u003e34\u003c/span\u003e) (Supplementary Fig. S8A). This reduction of active phosphorylated GSK-3\u0026beta; decreased in turn \u0026beta;-catenin phosphorylation and degradation, leading to its nuclear accumulation in cells and tumors, and activation of Wnt pathway target genes, including \u003cem\u003eMMP7\u003c/em\u003e and \u003cem\u003eOLFM4\u003c/em\u003e (Supplementary Fig. S8A-D). \u0026beta;-catenin is the major effector of the Wnt pathway, a central force in CRC tumorigenesis\u003csup\u003e34\u0026ndash;37\u003c/sup\u003e. Tankyrase (TNK) inhibition reduced Wnt/\u0026beta;-catenin signaling(\u003cspan class=\"CitationRef\"\u003e37\u003c/span\u003e), but genes like \u003cem\u003eDUSP6\u003c/em\u003e, directly regulated by MAPK signaling, were unaffected by Wnt inhibition. Similarly, \u003cem\u003ePROM1\u003c/em\u003e was induced by triplet but unaffected by TNKi treatment. This suggests that BRAFi therapy triggers a signaling shift that suppresses MAPK while enhancing Wnt/\u0026beta;-catenin signaling, orchestrating a precise gene expression profile that determines tissue architecture remodeling during the acquisition of resistance in \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e CRC tumors.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eCombined BRAF inhibitory therapy increased tumor vascularization, which can be effectively blocked by bevacizumab to prolong progression-free survival (PFS)\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eAt this point of the study, despite our significant advances in understanding the resistance of mechanisms to BRAFi therapies in colorectal cancer (CRC), a strategy to improve patient outcomes remained elusive. The complex tumor tissue remodeling observed under BRAFi treatment posed challenges for direct pharmacological targeting. We thus shifted our focus to tumor vascularization, a process which genes were significantly enriched in triplet-treated naive PDX models (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eD, \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eB-D). Indeed, gene sets associated with vascular development were particularly upregulated in these models following prolonged triplet treatment (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eA and \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eB, Supplementary Table S14, S15). These included genes such as \u003cem\u003eMSLN\u003c/em\u003e, \u003cem\u003ePROM1\u003c/em\u003e, \u003cem\u003eTSPAN12\u003c/em\u003e, \u003cem\u003eST6GAL1\u003c/em\u003e, \u003cem\u003eIL6\u003c/em\u003e, \u003cem\u003eSEMA5A\u003c/em\u003e, and \u003cem\u003ePDGFA\u003c/em\u003e; all involved in tumor angiogenesis via various mechanisms.\u003c/p\u003e\n \u003cp\u003eNotably, \u003cem\u003ePROM1\u003c/em\u003e, encoding the stem cell marker CD133, was upregulated by triplet treatment (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eB-D). CD133 has been implicated in stabilizing VEGFA(\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e), a key driver of angiogenesis targeted by the monoclonal antibody bevacizumab, a standard treatment in CRC(\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e). Although triplet therapy did not consistently increase total VEGFA mRNA or protein levels (Supplementary Fig. S9A-D), it did promote extracellular VEGFA accumulation in \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e CRC cell line organoid models (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eE-F, Supplementary Fig. S9E-F). Further \u003cem\u003ein vivo\u003c/em\u003e analyses showed that triplet treatment significantly elevated microvascular density (MVD) in naive PDX models, with resistant models already presenting higher baseline MVD (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eA-D). Importantly, six out of ten CRC patients exhibited increased MVD after treatment with BRAFi therapies (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eE and Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThese findings highlighted tumor vascularization as a potential vulnerability that could be exploited therapeutically. Indeed, the combination of triplet therapy with bevacizumab effectively prolonged PFS in HT29 xenografts and naive PDX models (M64-N) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eF and \u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eG). In the M64-N model, nearly all tumors treated with triplet and bevacizumab showed no progression after 50 days, in contrast to the lower efficacy of triplet alone (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eG). Bevacizumab also demonstrated efficacy in reducing tumor growth after acquiring resistance to triplet. Moreover, the combination of triplet and bevacizumab modestly slowed the growth of resistant CTAX91-R xenografts (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003eG).\u003c/p\u003e\n \u003cp\u003eIn conclusion, increased MVD during tissue remodeling represents a targetable vulnerability in BRAFi-resistant CRC. These preclinical findings us led to the initiation of the BRAVE phase I/II clinical trial (NCT06411600), in which we are enrolling 94 \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e mCRC patients (BRAFi-naive) treated with cetuximab (EGFRi), encorafenib (BRAFi), and bevacizumab (VEGFAi). Considering our preclinical results with PDX models, the trial\u0026rsquo;s primary objective is PFS, with secondary endpoints including response rate, time to response, duration of response, and overall survival. 15 patients have been already enrolled and results are encouraging.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn the phase III BEACON trial, we observed that the combination of encorafenib (BRAFi), cetuximab (EGFRi), and/or binimetinib (MEKi) significantly improved overall survival and response rates in mCRC patients with the \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003e\u003cem\u003eV600E\u003c/em\u003e\u003c/sup\u003e mutation, compared to standard chemotherapy(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). This study was the base for FDA and EMA approval of encorafenib and cetuximab combination as the new SoC for mCRC patients. However, resistance to these combinations occurred in almost all patients. Therefore, understanding the mechanisms ruling this acquisition of resistance remained a clinical challenge.\u003c/p\u003e \u003cp\u003eIn response our study has explored the non-genetic mechanisms of resistance to MAPK oncogenic signal inhibition by combining encorafenib, cetuximab, and binimetinib in mCRC. We observed tumor cell intrinsic reprogramming, including the previously described immune response, but also an unanticipated mucinous differentiation and massive mucin secretion. Importantly, it was the maturation of biological active mucins a critical determinant for resistance acquisition. Blocking this process could be a potential therapeutic strategy that was tested in a clinical trial using Bromelain as an unspecific mucolytic against advanced solid tumors (NCT02340845). Future drugs with more specific mode of action against mucin secretion or function could extend the benefit of advanced CRC patients progressing to BRAFi-based therapies.\u003c/p\u003e \u003cp\u003eSome mucins can activate the oncogenic Wnt/b-catenin signaling that is a driving force of CRC malignancy. For instance, the MUC1 cytoplasmic domain (CD) has been shown to bind β-catenin(\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e) blocking its phosphorylation by GSK3b, its proteasome degradation and final accumulation. We observed a coordinated increase of mucins and nuclear b-catenin upon GSK3b blockade. In addition, b-catenin promotes the expression of \u003cem\u003eMUC4\u003c/em\u003e(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e) and other tissue remodeling genes such as \u003cem\u003eMMP7\u003c/em\u003e acting as a transcriptional coactivator. In fact, Wnt/b-catenin has also been considered responsible for the resistance of different cancer types to a variety of drugs(\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). In summary, our results indicate that a Mucin/Wnt/b-catenin axis loop could play a relevant role in the acquisition of resistance to BRAFi-based therapies in CRC and is an attractive therapeutic target. Additionally, detecting mucins in blood could improve monitoring response and early progression of CRC patients treated with combined BRAFi therapies as previously describe in mucinous ovarian or breast carcinomas(\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe triplet therapy unexpectedly enhanced mucinous differentiation in tumor cells, typically associated with low proliferation. Nevertheless, mucinous tumors continued to grow and progress under triplet. Additionally, there was a concurrent increase in \u003cem\u003ePROM1\u003c/em\u003e/CD133 positive cells, indicating that this cancer stem cell-like population may feed tumor growth and progression(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). \u003cem\u003eATOH1\u003c/em\u003e upregulation could promote a rapid and constant transit of proliferating stem cell-like cells to mucinous goblet-like differentiated cells and sustain these populations balanced. Indeed, hypersecretory cancer cells producing mucins and collagens could be essential for building a biochemical fortress to protect minimal residual disease in tumors such as CRC and pancreas. Similar phenomenon has been described in lung cancer persistent tumors where minimal residual disease has shown heterogeneous cell profiles including alveolar differentiation and progenitor cells downstream of an activated Wnt/b-catenin pathway (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBRAFi treatments also promoted early stroma remodeling, including deposition of a distinct ECM with specific collagens and metalloproteases, changing tumor tissue stiffness and may creating low proliferation niches (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). These ECM changes could also facilitate infiltration and activation of cancer-associated fibroblasts (CAFs), known promoters of tumor progression (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). Key factors essential for these processes that are required for resistance, could also serve as valuable targets to prolong the effectiveness of BRAFi therapies. Further investigations are needed to identify key drug targets relevant in these ECM remodeling.\u003c/p\u003e \u003cp\u003eBRAFi treatments in \u003cem\u003eBRAF\u003c/em\u003e\u003csup\u003eV600E\u003c/sup\u003e mCRC induced a tumor cell persistence based in cancer cell intrinsic rewiring/plasticity and complex stroma remodeling, involving immune responses, ECM alterations, and increased vasculature. These processes revealed unexpected therapeutic targets critical for stroma function. Further investigation of this resistance process may reveal common mechanism across inhibitors like Tyrosine Kinase Receptor inhibitors, BRAFi, KRASi, and MEKi, facilitating the design of new drug combinations to benefit a wider range of patients progressing to these MAPK pathway-targeted therapies.\u003c/p\u003e \u003cp\u003eRecent perspectives suggest aggressive therapies such as chemotherapy, radiotherapy or some target-directed drugs that block cancer cell proliferation may promote the Darwinian selection of resistant genetic clones. Instead, combining these therapies with drugs that modulate other aspects of tumor biology, such as cancer cell invasion, plasticity, persistence, dormancy or stroma environment, may prevent the selection of more aggressive resistant genetic clones and prolong patient benefit (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). A good example of this strategy is blocking the MAPK pathway in mCRC with dabrafenib (BRAFi) and trametinib (MEKi) eliciting an immune response that, when combined with anti-PD-1 immunotherapy (spartalizumab), prolonged therapeutic benefit by modulating the immune stroma (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFollowing an equivalent rationale, we focused on blocking stromal tumor vasculature induced by BRAFi therapy with bevacizumab to prevent acquisition of resistance. We observed increased expression of vasculature-related genes through not classic angiogenesis regulators, but not VEGFA, which instead showed an increase in protein secretion. Interestingly, MUC1 has also being described to promote the expression and secretion of VEGFA (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e), completing a Mucin/Wnt-b-catenin/VEGFA axis behind the acquisition of resistance to BRAFi therapies. In addition, we observed an increased expression of \u003cem\u003ePROM1\u003c/em\u003e/CD133, which has been described to stabilize secreted VEGFA(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). Indeed, we demonstrated that blocking VEGFA with bevacizumab extended the benefit of BRAFi therapy, whether used upfront or after progression, demonstrating that the drug-enhanced vasculature in BRAF mutant mCRC tumors is a vulnerability that can be exploited early in resistance acquisition. Consequently, our findings led to the initiation of the BRAVE clinical trial (NCT06411600), where we are testing the combination of encorafenib, cetuximab, and bevacizumab in mCRC patients refractory to previous lines of chemotherapy and naive for BRAFi. This trial aims to extend PFS, leveraging our research insights into stroma and vasculature remodeling during acquisition of resistance to BRAFi therapies, for improving patients\u0026rsquo; outcome. We have currently enrolled 15 patients showing promising results.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003e\u003cu\u003eCell lines\u0026nbsp;\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHuman COLO 205 colorectal cancer (CRC) cells were generously provided by Dr. Diego Arango. Other human CRC cell lines used were HT29 and SW620, both procured from the American Type Culture Collection (ATCC). These cell lines were maintained in Dulbecco\u0026rsquo;s modified eagle medium (DMEM) supplemented with 10% Fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) and maintained at 37\u0026deg;C within a humidified incubator with 5% CO\u003csub\u003e2\u003c/sub\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGeneration of\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eGCNT1/3\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;knockout cell lines\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe primer sequences used to generate \u003cem\u003eGCNT1\u003c/em\u003e and \u003cem\u003eGCNT3\u003c/em\u003e knockout cell lines were:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eCRISPR sgRNA\u003cem\u003e\u0026nbsp;GCNT1\u003c/em\u003e forward:5-\u0026rsquo;CACCGAAGCGGTATGAGGTCGTTAA\u0026nbsp;-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003eCRISPR sgRNA \u003cem\u003eGCNT1\u0026nbsp;\u003c/em\u003ereverse:5\u0026rsquo;-AAACTTAACGACCTCATACCGCTTC-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003eCRISPR sgRNA \u003cem\u003eGCNT3\u003c/em\u003e forward:5\u0026rsquo;-CACCGGCGGGCTATCTGCGTTTATG-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003eCRISPR sgRNA \u003cem\u003eGCNT3\u0026nbsp;\u003c/em\u003ereverse:5-\u0026rsquo;AAACCATAAACGCAGATAGCCCGCC-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003eCRISPR sgRNA Scramble forward:5\u0026rsquo;-CACCGGCACTACCAGAGCTAACTCA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003eCRISPR sgRNA Scramble reverse:5\u0026rsquo;-AAACTGAGTTAGCTCTGGTAGTGCC-3\u0026rsquo;.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cem\u003eGCNT1\u003c/em\u003e and \u003cem\u003eGCNT3\u003c/em\u003e \u003cem\u003eknockout\u003c/em\u003e (GCNT1/3 KO) and control (Scramble) cell lines were generated using CRISPR-Cas9 system. HT29 cells were transfected with pSpCas9-sgRNAGCNT1guide1-2A-Puro and pSpCas9-sgRNA-GCNT3guide1-2A-Puro constructs using linear polyethylenimine (PEI 25000, Polysciences). Following a 48 hours post-transfection period, the cells that had incorporated the plasmid were selected using puromycin at a concentration of 1\u0026micro;g/ml for 72 hours.\u0026nbsp;Subsequently, individual clones were expanded.\u003c/p\u003e\n\u003cp\u003eValidation of the HT29 GCNT1/3 KO models was carried out using a PCR-based screening approach (52). Validation primers:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cem\u003eGCNT1\u003c/em\u003e forward: 5\u0026rsquo;-GCAATGAGTGCAAACTGGAA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eGCNT1\u003c/em\u003e reverse: 5\u0026rsquo;-TTCAGGAATCCTTTGGATGG-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eGCNT3\u0026nbsp;\u003c/em\u003eforward: 5\u0026rsquo;-CAGGGAATGCGTACATTGTG-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eGCNT3\u0026nbsp;\u003c/em\u003ereverse: 5\u0026rsquo;- CACAGGTAGCAACGCTCTCA-3\u0026rsquo;.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eTwo-dimensional (2D) and Three-dimensional (3D) cell culture\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the 2D culture of human colorectal cancer cell lines (COLO 205, HT29, SW620), cells were sustained in Dulbecco\u0026rsquo;s modified eagle medium (DMEM); 10% Fetal Bovine Serum (FBS) and 1% Penicillin-Streptomycin (P/S).\u003c/p\u003e\n\u003cp\u003eFor the 3D culture of the HT29 cell line, 2x10\u003csup\u003e5\u0026nbsp;\u003c/sup\u003ecells were suspended in 200 \u0026micro;l of medium combined with Matrigel (1:3 v/v ratio) with Human Intestinal Stem Cell Media (HISC)(53) when single cell es required for the experiment the HISC media is complemented with Rock inhibitor (10\u0026micro;M) and GSK3 inhibitor (5\u0026micro;M)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn both cases, 2D and 3D cultures, after 72 hours, cells were maintained under serum-limiting condition (0.5% FBS). Following an overnight incubation, the cells were exposed to encorafenib (1\u0026micro;M), a combination therapy of cetuximab (100\u0026micro;g/ml), encorafenib (1\u0026micro;M) and binimetinib (100nM) \u0026ndash; collectively referred to as the \u0026apos;triplet\u0026apos; \u0026ndash; or vehicle in DMEM with 2.5% FBS for 48 hours.\u003c/p\u003e\n\u003cp\u003eAll cell lines used in this study were cultured in a humidified incubator at 37\u0026deg;C with 5% CO\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003ePatient samples\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSamples were collected from patients at Vall d\u0026apos;Hebron University Hospital and Hospital Sant Joan Desp\u0026iacute; Mois\u0026egrave;s Broggi. Novartis Biomedical Research procured patient tumor specimens from the National Disease Research Interchange (Philadelphia, Pennsylvania, USA), Cooperative Human Tissue Network which is funded by the National Cancer Institute (Rockville, Maryland, USA) and Maine Medical Center (Portland, Maine, USA). All patients provided informed consent for the use of their samples.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eMice\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe experiments were conducted in accordance with the animal care directive of the European Union (86/609/CEE) and were approved by the Ethical Committee of Animal Experimentation of Vall d\u0026apos;Hebron Institute of Research (VHIR)/Vall d\u0026apos;Hebron Institute of Oncology (VHIO); ID: 12/18 CEEA). Female NOD.CB17-\u003cem\u003ePrkdc\u003csup\u003escid\u003c/sup\u003e\u003c/em\u003e/Rj mice, approximately 8 weeks old, were purchased from Janvier Labs. The mice were housed in groups of 5 animals per ventilated cage, with ad libitum access to food and water. Environmental conditions encompassed a 12-hour light-dark cycle, a temperature of 22 \u0026plusmn; 2\u0026deg;C, and a relative humidity of 45-65%.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe randomization of mice was performed based on tumor size. Continuous monitoring of the mice was conducted by the authors, facility technicians, and a veterinary scientist responsible for overseeing animal welfare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient-derived xenograft (PDX) establishment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHuman colorectal carcinoma tissues were obtained upon surgery in accordance with the ethical standards of the institutional committee on human experimentation. Histological diagnosis was based on microscopic features of carcinoma cells to determine both histological type and grade. The excised cancer tissues were washed 3 times in cold PBS solution and incubated overnight in DMEM/F-12 supplemented with an antibiotic and antifungal cocktail (penicillin at 250 U/ml, streptomycin at 250 \u0026mu;g/mL, fungizone at 10 \u0026mu;g/mL, kanamycin at 10 \u0026mu;g/mL, gentamycin at 50 \u0026mu;g/mL, and nystatin at 5 \u0026mu;g/mL).\u003c/p\u003e\n\u003cp\u003eEnzymatic digestion was performed using collagenase (1.5 mg/mL) and DNase I (20 \u0026mu;g/mL) in a medium containing the same antibiotic and antifungal cocktail during 1 h at 37\u0026deg;C with intermittent pipetting every 15 minutes to disperse cells. After digestion, the dissociated sample was filtered through a 100 \u0026mu;m pore size filter and then washed with fresh medium. To remove red blood cells, the sample was subjected to a 10 minutes ammonium chloride treatment, followed by another wash.\u003c/p\u003e\n\u003cp\u003eFinally, a suspension of 1x10\u003csup\u003e5\u003c/sup\u003e cells in 50 \u0026mu;L of PBS was mixed with an equal volume of Matrigel. This cell-Matrigel mixture was then subcutaneously injected into NOD-SCID flank mice to establish subcutaneous tumors for subsequent histological, protein, DNA, and RNA analyses.\u003c/p\u003e\n\u003cp\u003eNovartis Biomedical Research generated PDX models as previously described (54).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u003cu\u003eIn vivo\u003c/u\u003e\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cu\u003e\u0026nbsp;studies\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHuman colon carcinoma models were used and grown as previously described (55). 10\u003csup\u003e5\u003c/sup\u003e cells for PDX or 10\u003csup\u003e6\u003c/sup\u003e cells for cell line, suspended in PBS, were mixed with Matrigel (1:1 v/v ratio) and subcutaneously injected into both flanks of NOD-SCID mice. For the experiment involving tumor pieces, 3x3 mm tumor pieces obtained from a resistant mouse PDX to BRAFi tumors were immediately implanted subcutaneously in both flanks. Once the tumors reached a volume of 60 mm\u003csup\u003e3\u003c/sup\u003e, mice were randomly assigned into distinct experimental groups, including control (vehicle), encorafenib, bevacizumab, combination therapy involving cetuximab, encorafenib, and binimetinib (triplet), as well a combination therapy consisting of the triplet and bevacizumab.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCetuximab was administered intraperitoneally at a dosage of 20 mg/kg twice a week. Encorafenib was orally administered daily at 20 mg/kg, binimetinib was given orally twice a day at 6 mg/kg, and bevacizumab was intraperitoneally administered twice a week at a dose of 5 mg/kg.\u003c/p\u003e\n\u003cp\u003eWhen the tumors reached the end-point criteria, the mice were euthanized, and complete necropsies were performed. The subcutaneous tumors were fixed in formalin or frozen in liquid nitrogen immediately after the extraction to perform the histological, RNA, DNA and protein analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTumors were measured 3 times per week, and their volume was estimated using the formula: V = (length\u0026times;width\u003csup\u003e2\u003c/sup\u003e)/2, where length represents the largest tumor diameter and width represents the perpendicular tumor diameter.\u003c/p\u003e\n\u003cp\u003eThe number of mice for each experiment is reported in figure legends.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePFS and survival curves\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProgression-free survival (PFS) was determined in the subcutaneous tumor growth experiments using the response evaluation criteria in solid tumors (RECIST)\u0026nbsp;(56). Disease progression was identified by a percentage change of approximately 25% in tumor size from the baseline (time of treatment initiation). To calculate the time elapsed between the baseline tumor size and a 25% increase in tumor volume, a second-order polynomial (quadratic) equation was applied through the GraphPad Prism software.\u003c/p\u003e\n\u003cp\u003eStatistical significance was evaluated utilizing the Kaplan-Meier survival curve along with the log-rank (Mantel-Cox) test, with significance considered at \u003cem\u003ep\u0026nbsp;\u003c/em\u003e\u0026lt; 0.05.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBest response analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAn analysis of the best treatment response was performed, and this has been depicted using a waterfall plot. Using GraphPad Prism software, we generated a second-degree polynomial equation to calculate the elapsed time between the baseline tumor volume of the control group and a 50% increase. Once this day was calculated, the graph was extrapolated to indicate the value corresponding to the best response of each tumor.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eSample preparation\u0026nbsp;\u003c/u\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cu\u003efor Western blot analyses\u0026nbsp;\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eTumors\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eProtein extracts from tumors were processed in SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM dithiothreitol, DTT) supplemented with protease and phosphatase inhibitor cocktails. The samples were homogenized in the lysis buffer using a pestle. Then, samples were heated at 95\u0026ordm;C using a dry bath for 5 minutes and sonicated. Once the samples were thoroughly homogenized, they were centrifuged at 15000g for 10 minutes. Finally, the supernatant was collected and quantified using the Pierce BCA Protein Assay Kit.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eTwo-dimensional (2D) cell culture\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eTotal cell extracts from cells were homogenized in SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) containing protease and phosphatase inhibitor cocktails. Then, samples were boiled at 95\u0026ordm;C using a dry bath for 5 minutes, followed by sonication and centrifugation at 15000 g for 10 minutes. The resulting supernatant was collected and subsequently quantified using the Pierce BCA Protein Assay Kit.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eThree-dimensional (3D) cell culture\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePrior to initiating the protein extraction from the spheroids, the supernatant was removed and set aside for later protein extraction. Cell recovery solution was added to the Matrigel embedded spheroids and incubated at 4\u0026ordm;C on a shaker at 30 rpm for 30 minutes. This solution dissolved the Matrigel, enabling the intact retrieval of the spheroids.\u003c/p\u003e\n\u003cp\u003eAfter the 30-minute incubation, the mixture was centrifuged at 1400 rpm for 5 minutes at 4\u0026ordm;C. The supernatant was then collected and reserved for subsequent protein extraction that might have been released to the Matrigel.\u003c/p\u003e\n\u003cp\u003eThe resulting pellet (spheroids) was washed with cold PBS and lysed using an SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) supplemented with a cocktail of protease and phosphatase inhibitors. It was heated at 95\u0026ordm;C using a dry bath for 5 minutes, sonicated, and then centrifuged at 15000g for 10 minutes. The supernatant was collected and stored for subsequent quantification.\u003c/p\u003e\n\u003cp\u003eThe proteins present in the Matrigel and in the supernatant were precipitated using trichloroacetic acid (TCA). For this, 1 volume of TCA was added to 4 volumes of the sample in both the supernatant and the Matrigel, followed by incubation for 15 minutes on ice. After this period, the samples were centrifuged at 4\u0026ordm;C at 16000g for 15 minutes. The supernatant was discarded, and the pellet was resuspended with 600 \u0026micro;l of cold acetone. The mixture was then centrifuged again at 4\u0026ordm;C at 16000g for 15 minutes, followed by an additional acetone wash. The supernatant was discarded, and the pellet was air-dried for 5 minutes at 95\u0026ordm;C.\u003c/p\u003e\n\u003cp\u003eSubsequently, the pellet was resuspended in an SDS lysis buffer (1% SDS, 10 mM EDTA pH 8, and 1 mM DTT) supplemented with a cocktail of protease and phosphatase inhibitors. The samples were heated at 95\u0026ordm;C for 5 minutes and then sonicated. The protein extract was stored for subsequent quantification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWestern blot\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWestern blots were conducted in accordance with established protocols. Briefly, samples were boiled at 95\u0026deg;C for 5 minutes. Lysates were then separated by SDS-PAGE and proteins were transferred onto a nitrocellulose membrane (BioRad). Membranes were blocked with TBS, 0.1% Tween 20, and either 5% non-fat milk or BSA. They were then incubated overnight at 4\u0026deg;C with the appropriate primary antibodies (EGFR 1:1000 (#SC-03), pEGFR 1:1000 (#2234), ERK 1:3000 (#9102), pERK 1:3000 (#9101), MMP7 1:2000 (#ab207299), VEGFA 1:1000-1/5000 (#ab183100, #ab155944) , Vinculin 1:5000 (#V9131), GAPDH 1:10000 (#ab128915)). Subsequently, the membranes were exposed to HRP-conjugated secondary antibodies diluted in the blocking solution for 1 hour at room temperature, after that, the membranes were revealed using SuperSignal West Pico PLUS Chemiluminescent Substrate and were visualized using the Amersham Imager 600 (GE Healthcare).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eSample preparation for proteomics analyses\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eProtein extraction\u0026nbsp;\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePDX tumor samples, around 50 mg wet weight, where lysed with 150 \u0026micro;L of lysis buffer (7M urea, 2M thiourea, 4% CHAPS, 30mm Tris.HCl pH=8.5), supplemented with 1M NaF and 0.1M Na\u003csub\u003e3\u003c/sub\u003eVO\u003csub\u003e4\u003c/sub\u003e as phosphatase inhibitors. Samples were first mechanically disrupted using a micro pestle, and then sonicated using a probe sonicator (\u003cem\u003eVCX 150; Sonics \u0026amp; Materials Inc. USA\u003c/em\u003e) with 10 cycles of 15 seconds of ultrasound bursts, followed by 10 seconds of cooling intervals, while keeping the tube ice-cooled. Then, lysed samples were centrifuged for 5 min at 16000 g and the supernatants were collected. Protein extracts were further purified by a modified TCA-acetone precipitation (2D-CleanUp Kit, GE Healthcare) and, finally, resuspended in 100 \u0026micro;L of 8M urea, 50mM ammonium bicarbonate, plus phosphatase inhibitors. Protein concentration was determined using the Bio-Rad RCDC Protein Assay (Bio-Rad, UK).\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eTrypsin digestion for total proteome analysis\u003c/u\u003e\u003cu\u003e\u0026nbsp;\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003e20 \u0026micro;g of total protein from each sample were digested for total proteome analysis. Proteins were first reduced with DTT by addition of freshly prepared 700 mM DTT solution to a final concentration of 10 mM, for 1h at RT. Next, they were carbamidomethylated with iodoacetamide (IAA), by addition of the required volume of freshly prepared 700 mM IAA to obtain a final concentration of 30 mM in the sample. Alkylation was allowed to proceed for 30 minutes at RT in the dark, and then the reaction was quenched by addition of N-acetyl-L-cysteine to a final concentration of 35 mM, followed by incubation for 15 minutes at RT in the dark. Samples were then diluted with 50 mM ammonium bicarbonate to a final concentration of 1M Urea, and then modified porcine trypsin (Promega Gold)\u0026nbsp;was added in a ratio of 1:20 (w/w), and the mixture was incubated overnight at 37 \u0026deg;C. The reaction was stopped with formic acid (FA) to a final concentration of 0.5%, and the digest was kept at -20\u0026ordm;C until further analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLiquid chromatography-Mass spectrometry analysis (LC-MS)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor LC-MS/MS analysis peptide mixtures were diluted in 3% ACN, 1% FA and the sample was loaded to a 300 \u0026mu;m \u0026times; 5 mm Pep-Map C18 (Thermo Scientific) at a flow rate of 15 \u0026mu;l/min using a Thermo Scientific Dionex Ultimate 3000 chromatographic system (Thermo Scientific). Peptides were separated using a C18 analytical column (NanoEase MZ HSS T3 column, 75 \u0026mu;m \u0026times; 250 mm, 1.8 \u0026mu;m, 100\u0026Aring;, Waters) with a 210 minutes run for total proteome samples, comprising four consecutive steps, first 3 minutes of isocratic flow at 3% B, then gradient flows from 3 to 35% B in 180 minutes, from 35 to 50% B in 5 minutes, from 50 to 85% B in 1 minute, followed by isocratic elution at 85 % B in 5 minutes and stabilization to initial conditions (A= 0.1% FA in water, B= 0.1% FA in ACN), comprising four consecutive steps, first 3 minutes of isocratic gradient at 3%B, from 3 to 35% B in 90 minutes, from 35 to 50% B in 5 minutes, from 50 to 85% B in 1 minute, followed by isocratic elution at 85% B in 5 minutes and stabilization to initial conditions (A= 0.1% FA in water, B= 0.1% FA in ACN). Flow rate was 250 nL/min and the column was kept at 40 \u0026ordm;C. The column outlet was directly connected to an Advion TriVersa NanoMate (Advion) fitted on an Orbitrap Fusion Lumos\u0026trade; Tribrid (Thermo Scientific). The mass spectrometer was operated in a data-dependent acquisition (DDA) mode. Survey MS scans were acquired in the orbitrap with the resolution (defined at 200 m/z) set to 120,000. The lock mass was user-defined at 445.12 m/z in each Orbitrap scan. The top speed (most intense) ions per scan were fragmented in the HCD cell and detected in the orbitrap at 30000 resolution. Quadrupole isolation was employed to selectively isolate peptides of 350-1700 m/z. The predictive automatic gain control (pAGC) target was set to 4e5. The maximum injection time was set to 50ms for MS1 and 70ms for MS2 scan. Included charged states were 2 to 7. Target ions already selected for MS/MS were dynamically excluded for 15 s. The mass tolerance of this dynamic exclusion was set to \u0026plusmn;2.5 ppm from the calculated monoisotopic mass. Spray voltage in the NanoMate source was set to 1.7 kV. RF Lens were tuned to 30%. Minimal signal required to trigger MS to MS/MS switch was set to 5000 and activation Q was 0.250. The spectrometer was working in positive polarity mode and singly charge state precursors were rejected for fragmentation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProtein identification, quantitative differential and functional analysis.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProgenesis \u0026reg; QI for proteomics software v3.0 (Nonlinear dynamics, UK) was used for MS data analysis. The LC-MS runs were aligned to an automatically selected reference sample Alignments were then manually supervised. Only features within the 400 to 1,600 m/z range, 45 to 190 minutes retention time, for total proteome, and with positive charges between 2 to 4 were considered for identification and quantification. Peak lists (mgf files) were generated using Progenesis and loaded to Proteome Discoverer v2.1 (Thermo Fisher Scientific) for protein identification. Proteins were identified using Mascot v2.5 (Matrix Science, London UK) to search the SwissProt database. Two different searches were performed, restricting taxonomy to human or mouse proteins, respectively. MS/MS spectra were searched with a precursor mass tolerance of 10 ppm, fragment tolerance of 0.02 Da, trypsin specificity with a maximum of 2 missed cleavages, cysteine carbamidomethylation set as fixed modification and methionine oxidation as variable modification for total proteome analysis. Significance threshold for the identifications was set to p\u0026lt;0.05, minimum ions score of 20.\u003c/p\u003e\n\u003cp\u003eProtein quantification of human, mouse or ambiguous proteins was performed by addition of the corresponding integrated peptide MS signals derived from the Progenesis analysis. On the basis of the two searches performed, MS features were classified in those attributable to mouse or human exclusive peptides and those corresponding to shared sequences. For those proteins presenting the three types of peptide features, the abundances of human and mouse proteins was estimated by deconvolution of the shared feature signals, as described in Saltzman et al. 2018 (57). The algorithm was implemented as an Excel macro.\u003c/p\u003e\n\u003cp\u003eStatistical analysis was performed using the Perseus software platform (58). Proteins presenting Fold change|\u0026gt;2 and adjusted p-value \u0026lt; 0.05 (T-test on log2 transformed abundance values) were considered as differential between the control and treated groups.\u003c/p\u003e\n\u003cp\u003eAfter the proteome differential expression analysis, functional analysis was performed using the Pre-Ranked Gene Set Enrichment Analysis (GSEA) method (59,60). This analysis was implemented with the clusterProfiler (61) R package v. 4.8.2 . \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEnrichment analysis was conducted on the following gene set collections from the Molecular Signatures Database (MSigDB), v. 2023.2: Hallmark gene sets, C2, C5 gene ontology (GO) gene sets and C6 oncogenic signatures.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eImmunohistochemical (IHC) staining\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFormalin-fixed paraffin-embedded (FFPE) tissues underwent standard deparaffination and rehydration processes. For antigen retrieval, the slides were incubated with a 10 mM sodium citrate pH 6 buffer. Following the inhibition of endogenous peroxidase activity using a 3% H\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e solution, the slides were permeabilized using a 1% Tween 20 solution in PBS for 15 minutes. Then, the tissue specimens were blocked with a 3% BSA solution in PBS for 1 hour and then incubated with the corresponding primary antibodies (CD31 1:20 (#DIA-310), Ki67 1:100 (#M7240), Muc2 1:100 (#BD555926), \u0026alpha;-SMA 1:1500(#A2547)) diluted in blocking solution at 4\u0026deg;C overnight. After washing, sections were incubated with corresponding HRP-conjugated secondary antibodies for 1 hour at room temperature (RT). For chromogenic detection, the Dako Liquid DAB+ Substrate Chromogen System was added onto the slides and incubated up to 10 minutes. Finally, the slides were counterstained with haematoxylin, dehydrated, and mounted. The NanoZoomer 2.0-HT Digital slide scanner C9600 was used to acquire a high-resolution whole slide scanner of the immunostainings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eAlcian blue staining\u0026nbsp;\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlcian blue staining was carried out on FFPE tumor xenograft sections and patient sections to identify mucinous differentiation. This staining procedure was conducted by the Translational Molecular Pathology service at VHIR (Vall d\u0026rsquo;Hebron Institute of Research). In brief, following the process of deparaffinization, FFPE sections were hydrated and subjected to a 3-minute treatment in absolute ethanol, followed by 5 minutes in acetic acid, 30 minutes in 1% Alcian blue solution, and 10 minutes in nuclear fast red. After each step, the sections were rinsed three times for 5 minutes each using PBS washes. Subsequently, the samples were dehydrated and coverslipped using mounting media.\u003c/p\u003e\n\u003cp\u003eThe NanoZoomer 2.0-HT Digital slide scanner C9600 was employed to visualize and evaluate the Alcian blue staining. For image quantification, the open-source software ImageJ was used. The analysis of Alcian blue staining involved quantifying the percentage of the Alcian blue-positive area in each tumor section.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003ePicro-sirius red staining\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSirius red staining was performed on FFPE tumor xenograft sections and patient sections to identify collagen fibers. This staining procedure was conducted by the Translational Molecular Pathology service at VHIR (Vall d\u0026apos;Hebron Institute of Research). In brief, after the deparaffinization process, FFPE sections were hydrated and then incubated in the picro-Sirius red solution for 1 hour. This was followed by a 5 second immersion in a 0.1% hydrochloric acid solution. After this step, the slides were rinsed in tap water for 10 minutes and then washed in distilled water. Finally, the slides were counterstained with haematoxylin, dehydrated, and mounted.\u003c/p\u003e\n\u003cp\u003eThe NanoZoomer 2.0-HT Digital slide scanner C9600 was utilized to visualize and assess the Sirius red staining. The picro-Sirius red staining was analyzed using the Twombli\u0026nbsp;(62), a Fiji macro designed for quantifying patterns in the extracellular matrix.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eRNA extraction\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor RNA extraction of 2D culture experiments, the plates were washed 3 times with cold PBS, and the extraction protocol was initiated. For in vivo experiments, snap-frozen samples were homogenized using a pestle. For both scenarios, the RNA extraction was performed using the RNeasy Mini Kit (Qiagen), following the manufacturer\u0026apos;s recommendations. The obtained RNA was quantified using a NanoDrop spectrophotometer (ThermoFisher Scientific).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGene expression analyses\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSamples were processed according to the following Affymetrix protocols: GeneChip WT PLUS Reagent kit (P/N 703174 2017) and Expression Wash, Stain and Scan User Manual (P/N 702731 2017) (Affymetrix Inc., Santa Clara, CA, USA) and hybridized to Clariom D Human expression array (ThermoFisher Scientific).\u003c/p\u003e\n\u003cp\u003eData were normalized using the robust multi-array average (RMA) method from the oligo R package (63) and differential gene expression was calculated using the limma R package (64). P-values were adjusted for multiple testing using the Benjamini and Hochberg (BH)correction. Genes differentially expressed (|Fold change|\u0026gt;2 and adjusted p-value \u0026lt; 0.05) in any of the time points for either Naive or Resistant models were represented in a heatmap. To evaluate the gene response to the treatment over time, z-scores for each gene were calculated using the scale function in R and the median of the different replicates within each time point for each group was calculated. Over representation analysis (ORA) and gene set enrichment analysis (GSEA) were performed with the clusterProfiler R package (61,65). Gene set collections (Gene Ontology, Reactome, KEGG, Chemical and Genetic Perturbations and Hallmark) were obtained from the Molecular Signature Database (MSigDb 2022). To construct heatmaps both pheatmap and ComplexHeatmap R packages were used (66,67) while for the rest of the graphs ggplot2 was used.\u003c/p\u003e\n\u003cp\u003eGene sets were classified as Naive specific, Resistant specific or commonly regulated depending on which condition they were significantly enriched (adjusted p-value \u0026lt; 0.01). Then, they were clustered by the similarity of the genes in their core enrichment using the GSEAmining package (68). Gene set clusters were then manually classified to specific biological categories (Figure 1D) and represented using alluvial plots with the ggalluvial R package (69). GSEA plots were performed using the gseaplot2 function from the enrichplot package (70).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGoblet cell markers were obtained from (71,72).Those that were up-regulated in each of the time points (|Fold change|\u0026gt;1.2 and adjusted p-value \u0026lt; 0.05) were represented in a heatmap. Normalized enrichment score (NES) from gene sets in the alluvial plot annotated in tissue remodelling and vasculature processes positively enriched in naive models were represented in bubble plots. From those selected gene sets, genes that were differentially expressed in each time point present in the core enrichment genes (from leading edge analysis) of the gene sets were represented in a heatmap.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eQuantitative RT-PCR analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo analyze the expression of selected genes, RNA retrotranscription was performed using the iScript\u0026trade; Advanced cDNA Synthesis Kit following the manufacturer\u0026rsquo;s instructions. Analyses were carried out in triplicate with 10 ng of cDNA using using PerfeCTa SYBR Green FastMix on a Quant Studio 6 Flex cycler. Specific pairs of primers were designed (https://www.ncbi.nlm.nih.gov/gene/) and used to detect the indicated transcripts. Relative gene expression was determined using the comparative CT method. TATA-binding protein (TBP) and Hypoxanthine Phosphoribosyltransferase 1 (HPRT) were used as \u0026nbsp;housekeeping genes.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cem\u003eATOH1\u003c/em\u003e forward primer: 5\u0026rsquo;-AGCTTCTTGTCGTTGTTG-3\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eATOH1\u003c/em\u003e reverse primer: 5\u0026rsquo;-AGGTGAATGGGGTGCAGA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eCLCA1\u003c/em\u003e forward primer: 5\u0026rsquo;-TCGTTGCAATCGACCCCAAT-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eCLCA1\u003c/em\u003e reverse primer: 5\u0026rsquo;-CCTGGGTCACCATGTCCTTT-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eCOL1A1\u003c/em\u003e forward primer: 5\u0026rsquo;-GATTCCCTGGACCTAAAGGTGC-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eCOLA1\u003c/em\u003e reverse primer: 5\u0026rsquo;-AGCCTCTCCATCTTTGCCAGCA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eMMP7\u003c/em\u003e forward primer: 5\u0026rsquo;-TCGGAGGAGATGCTCACTTCGA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eMMP7\u003c/em\u003e reverse primer: 5\u0026rsquo;-GGATCAGAGGAATGTCCCATACC-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eMUC4\u003c/em\u003e forward primer: 5\u0026rsquo;-AACACAGCCTGCTAGTCCAGCA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eMUC4\u003c/em\u003e reverse primer: 5\u0026rsquo;-TGGAGAGGATGGCTTGGTAGGT-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eVEGFA\u003c/em\u003e forward primer: 5\u0026rsquo;-AGGGCAGAATCATCACGAAGT-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eVEGFA\u003c/em\u003e reverse primer: 5\u0026rsquo;-AGGGTCTCGATTGGATGGCA-3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eTBP\u003c/em\u003e forward primer: 5\u0026rsquo;- CGGCTGTTTAACTTCGCTTC -3\u0026rsquo;\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eTBP\u003c/em\u003e reverse primer: 5\u0026rsquo;-CACACGCCAAGAAACAGTGA -3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eHPRT\u003c/em\u003e forward primer: 5\u0026rsquo;-CTGGCGTCGTGATTAGTGAT -3\u0026rsquo;.\u003c/li\u003e\n \u003cli\u003e\u003cem\u003eHPRT\u003c/em\u003e reverse primer: 5\u0026rsquo;-GGCTACAATGTGATGGCCT-3\u0026rsquo;.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eSingle Cell Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnalysis of the Single Cell data was performed in R using the Seurat package (version 5.0.3). Cells that expressed less than 200 genes were filtered out, together with genes that were only expressed in less than 3 cells. Additional filtering was done to remove droplets containing more than one cell, by filtering for cells with more than 1000 and less than 8000 features. Dead cells were removed by filtering for less than 25% mitochondrial DNA. The filtered gene expression data was normalized using the \u0026apos;LogNormalize\u0026apos; method with a scale factor of 10.000, implemented through the \u0026lsquo;NormalizeData\u0026rsquo; function in the Seurat package. Next, the 2000 most variable feature were identified, and the data was scaled for all genes. Using the \u0026lsquo;FindNeighbours\u0026rsquo; and the \u0026lsquo;FindCluster\u0026rsquo; function from Seurat, cells were clustered, with a resolution of 0.5 and non-linear dimensionality reduction, with the \u0026lsquo;RunUMAP\u0026rsquo; function, was done using the first 20 principal components, to visualize the high-dimensional data in a low-dimensional space. A total of 14 clusters were found and\u0026nbsp;annotated using a similar set of features as described in the single-cell transcriptomic analysis paper by Sebastian et al (73).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eDNA extraction\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor in vivo experiments, DNA extraction from snap-frozen samples was performed using the QIAGEN QIAamp DNA Mini and Blood Mini Kit, following the manufacturer\u0026apos;s recommendations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWhole-Exome Sequencing (WES)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe libraries for Whole Exome Sequencing (WES) were prepared from 500 ng of DNA per sample using the SureSelect XT Target Enrichment kit for Illumina platforms. For enrichment, the SureSelect XT Human All Exon v5 capture set (Agilent) was used, and sequencing was performed on the HiSeq 2500 sequencer (chemistry v3, high-output mode). This achieved a coverage of approximately 150x with a read length of 2x100 bp.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData processing and analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe sequencing raw data from PDX were pre-processed using BBSplit (BBTools) (74) to separate human-origin reads and remove mouse contamination. We then utilized the sarek/nf-core 3.2.3 (75), with default settings, to extract somatic variants from paired samples. Quality checks for the Fastq files were performed by FastQC v0.11.4 (76), followed by trimming low-quality regions using fastP v0.23.4.(77). Alignment to the human reference genome GRCh38 was successfully done using Burrows-Wheeler Alignment with mem algorithm (BWA-MEM) v.0.7.17-r1188 (78). After error correction based on base recalibration using GATK v.4.4.0 (79,80) and duplicate removal by Picard v2.9.2 (81), somatic variant calling was done with Mutect2 (82) in a control versus tumor setup to identify single nucleotide variants (SNVs) and short insertions and deletions (indels), which were then converted to Variant Call Format (VCF). Publicly available resources such as normal sample panels of GATK and databases of already known mutations like Mills (83) for short indels and dbSNP (84) for single nucleotide polymorphisms were used in order to avoid artifacts and perform base recalibration effectively. Variants failing to meet quality criteria of GATK Best Practices were filtered out, and those with variant allele frequencies below 5% or supported by fewer than 10 reads were excluded. Variant annotation was performed using SNPeff v.5.1d (85), and conversion into Mutation Annotation Format (MAF) was conducted using the maftools package (86). Heatmaps and oncoprints were generated using the ComplexHeatmap R-package (66), while tumor mutational burden (TMB) was calculated with the maftools package. Additional analyses were carried out using Python v3.7 and R v4.2.2. We used data from Colorectal Adenocarcinoma cohort of TCGA (COADREAD, PanCancer Atlas). Data was downloaded from cBioportal (87) selecting only BRAF-V600E mutated patients.\u003c/p\u003e\n\u003cp\u003eTumor mutational signatures were obtained from variant calling format (VCFs) data extracted in earlier stages and analyzed using SigProfilerAssignment v.0.1.4 (88). This software assigns specific established signatures to individual samples based on the type and frequency of variants, with a focus on single nucleotide polymorphisms (SNPs) for single base substitution signatures (SBS). Signature data were sourced from the latest release of the Catalogue of Somatic Mutations in Cancer (COSMIC), version 3.4 (89). Contribution of each signature was determined by evaluating the number of mutations attributed to it relative to the total mutation burden.\u003c/p\u003e\n\u003cp\u003eTo extract TCGA signatures, MAF data from the same previous cohort COADREAD TCGA Pan Cancer Atlas were downloaded in GRCh37 format from cBioPortal. Only samples harboring BRAFV600E mutations were selected for further analysis. The MAF data were converted to VCF using maf2vcf (18) and adjusted to GRCh38 using the Liftover function of Picard v.2.9.2. The resulting VCF files were successfully processed using SigprofilerAssignment (90).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDNA methylation analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSample preparation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFFPE DNA extraction.\u0026nbsp;\u003c/strong\u003eDNA isolation from FFPE embedded samples from PDX tissues was performed using the ReliaPrep FFPE gDNA Miniprep System Kit (Promega, USA). All DNA samples were treated with RNase A for 1 h at 45\u0026ordm;C, quantified by the fluorometric method (Quant-iT PicoGreen dsDNA Assay, Life Technologies, CA, USA), and assessed for purity by NanoDrop (Thermo Scientific, MA, USA) 260/280 and 260/230 ratio measurements\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eQuality check of FFPE DNA.\u0026nbsp;\u003c/strong\u003eFollowing Illumina\u0026rsquo;s recommendations, the DNA quality of the FFPE samples was checked by performing a quantitative PCR with 2 ng of FFPE DNA. The average value of a standard provided by Illumina was subtracted in order to calculate the \u0026Delta;Cq. FFPE DNAs with \u0026Delta;Cq \u0026lt;5 indicate suitability for FFPE restoration.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBisulphite conversion and restoration.\u0026nbsp;\u003c/strong\u003eA minimum of 300 ng of DNA was bisulphite-converted using the EZ-96 DNA Methylation kit (Zymo Research Corp., CA, USA), following manufacturer\u0026rsquo;s instructions for Infinium assays. Next, bisulfite-converted DNA (bs-DNA) from FFPE samples was restored as previously described (91). In brief, after DNA denaturation (NaOH 0.1N, 10 min at room temperature), PPR (Primer Pre Restore) and AMR (Amplification Mix Restore) reagents were added, and samples were incubated for 1 h at 37\u0026ordm;C. DNA was cleaned with ZR-96 DNA Clean \u0026amp; Concentrator-5 kit (Zymo Research Corp.) and eluted in 13 \u0026mu;l of ERB (Elution Restore Buffer Reagent). Cleaned DNA was then denatured for 2 min at 95\u0026deg;C, followed by ligation incubation at 37\u0026deg;C for 1 h with ER (Elution Restore) and CMM (Convert Master Mix) reagents. After a second cleaning step with ZR-96 DNA Clean \u0026amp; Concentrator-5 kit (Zymo Research Corp.), DNA was finally eluted in 10 \u0026mu;l of DiH20.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eArray hybridization.\u0026nbsp;\u003c/strong\u003eEight microliters of restored FFPE bs-DNA were used as input for the Illumina Infinium HD Methylation Assay Protocol, as previously described [Moran et al., 2016]. Briefly, the process includes a whole genome amplification step followed by enzymatic end-point fragmentation, precipitation, and resuspension. The resuspended samples were hybridized on the Illumina Infinium MethylationEPIC BeadChip at 48\u0026deg;C for 16 h. Then unhybridized and non-specifically hybridized DNA were washed away, followed by a single nucleotide extension using the hybridized bisulfite-treated DNA as a template. The nucleotides incorporated were labeled with biotin (ddCTP and ddGTP) and 2,4-dinitrophenol (DNP) (ddATP and ddTTP). After the single base extension, repeated rounds of staining were performed with a combination of antibodies that differentiated DNP and biotin by fixing them different fluorophores. Finally, the BeadChip was washed and scanned using the Illumina iScan with Autoloader system.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethylation Analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIllumina EPIC Methylation array data in the form of IDAT files were imported into R using the minfi package (version 1.48.0). Stratified quantile normalization was conducted and subsequently probes with p-values exceeding 0.01 were filtered out, using the minfi function detectionP. Additional probe filtration was done removing those containing SNPs or being cross-reactive, for the latter utilizing data from Pidsley et al. (92).\u003c/p\u003e\n\u003cp\u003eProbes were annotated using the IlluminaHumanMethylationEPICanno.ilm10b4.hg19 package (version 0.6.0). Differential methylation analysis was conducted on the M-values for each site using the limma package to identify differentially methylated probes (DMP), or CpGs, across the same comparisons employed in the gene expression analysis (Fig. 1c). P-values were adjusted for multiple testing using the BH method, and DMPs were selected based on a |Fold change|\u0026gt;1.5 and adjusted p-value \u0026lt; 0.05. To visualize the overlap of differentially methylated CpGs, Venn diagrams were constructed using the VennDiagram package (version 1.7.3). Stacked bar plots were generated to depict the distribution of differentially methylated CpGs in the genome, categorizing them based on CpG island status and location within promoters, gene bodies, or other functional regions.\u003c/p\u003e\n\u003cp\u003eTo compare gene expression and methylation on a gene-basis, the median methylation value of probes located in the TSS200 region and gene body, separately, was taken. Limma analysis was performed on these summarized methylation values to identify genes exhibiting differential methylation in their promoter or body regions. To explore the association between methylation and gene expression, scatterplots were created, plotting the log\u003csub\u003e2\u003c/sub\u003e Fold change of methylation against expression.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eQuantification and statistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe statistical analyses of the \u0026quot;wet lab/mice\u0026quot; experiments were conducted using GraphPad Prism 8.1 and R v4.0.4 (93). Most significant differences were assessed using the Student\u0026apos;s t-test for independent samples. Data were presented as means with SEM unless otherwise specified. For comparisons involving two or more groups that encompass continuous variables (such as gene expression, methylation), Mann-Whitney U and Kruskal-Wallis tests were employed. Statistical analyses were adjusted using the False Discovery Rate (FDR) or Bonferroni correction. Survival was measured using the Kaplan-Meier method. Asterisks codes were employed to denote different levels of statistical significance: *\u003cem\u003ep\u003c/em\u003e \u0026le; 0.05, **\u003cem\u003ep\u003c/em\u003e \u0026le; 0.01, ***\u003cem\u003ep\u003c/em\u003e \u0026le; 0.001, ****\u003cem\u003ep\u003c/em\u003e \u0026le; 0.0001.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cspan\u003eThe study including patients\u0026rsquo; samples and data was approved (ID: PR(AG)113_2015) by the Vall d\u0026rsquo;Hebron University Hospital institutional ethical review board according to the guidelines of the European Network of Research Ethics Committees, following European, national and local laws. Written informed consent was signed by all patients.\u003c/span\u003e\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eWe thank the IJC Genomics Unit. VHIO would like to acknowledge the Cellex Foundation for providing research facilities and equipment, the FERO Foundation for their funding support, and the Centro de Investigaci\u0026oacute;n Biom\u0026eacute;dica en Red de C\u0026aacute;ncer (CIBERONC). We thank the Institute of Health Carlos III (ISCIII-FEDER) PI20/00968, Fundaci\u0026oacute;n AECC (CLSEN19001ELEZ) (ACRCelerate), Department of Health (Generalitat de Catalunya, 2021SGR01567), Mutual M\u0026eacute;dica Award, GRUPO DE TRATAMIENTO DE LOS TUMORES DIGESTIVOS (GTTD) and Fundaci\u0026oacute;n CRIS Contra el C\u0026aacute;ncer. This research was partially funded by the SCITRON program from Novartis. We thank CERCA Programme / Generalitat de Catalunya for institutional support.\u003c/p\u003e\n\u003ch2\u003eData availability\u003c/h2\u003e\n\u003cp\u003eThe mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://proteomecentral.proteomexchange.org\u003c/span\u003e\u003c/span\u003e) via the PRIDE partner repository (\u003cspan class=\"CitationRef\"\u003e57\u003c/span\u003e) with the dataset identifier PXD54589.\u003c/p\u003e\n\u003cp\u003eExpression and methylathion data that support the findings of this study have been deposited in the Gene Expression Omnibus (GEO) under accession code GSE280600.\u003c/p\u003e\n\u003cp\u003eScRNA-seq and Exome Sequencing data have been deposited in the European Genome-phenome Archive (EGA) under the accession codes EGAD50000000939 and EGAD50000000940.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSung, H. \u003cem\u003eet al.\u003c/em\u003e Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. \u003cem\u003eCA. Cancer J. Clin.\u003c/em\u003e \u003cstrong\u003e71\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eDavies, H. \u003cem\u003eet al.\u003c/em\u003e Mutations of the BRAF gene in human cancer. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e417\u003c/strong\u003e, 949\u0026ndash;954 (2002).\u003c/li\u003e\n\u003cli\u003eDhomen, N. \u0026amp; Marais, R. New insight into BRAF mutations in cancer. \u003cem\u003eCurr. Opin. Genet. Dev.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 31\u0026ndash;39 (2007).\u003c/li\u003e\n\u003cli\u003eTaieb, J., Lapeyre-Prost, A., Laurent Puig, P. \u0026amp; Zaanan, A. Exploring the best treatment options for BRAF-mutant metastatic colon cancer. \u003cem\u003eBritish Journal of Cancer\u003c/em\u003e vol. 121 434\u0026ndash;442 (2019).\u003c/li\u003e\n\u003cli\u003eBond, C. E. \u0026amp; Whitehall, V. L. J. How the BRAF V600E mutation defines a distinct subgroup of colorectal cancer: Molecular and clinical implications. \u003cem\u003eGastroenterol. Res. Pract.\u003c/em\u003e (2018) doi:10.1155/2018/9250757.\u003c/li\u003e\n\u003cli\u003eRoth, A. D. \u003cem\u003eet al.\u003c/em\u003e Prognostic role of KRAS and BRAF in stage II and III resected colon cancer: Results of the translational study on the PETACC-3, EORTC 40993, SAKK 60-00 trial. \u003cem\u003eJ. Clin. Oncol.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 466\u0026ndash;474 (2010).\u003c/li\u003e\n\u003cli\u003eCantwell-Dorris, E. R., O\u0026rsquo;Leary, J. J. \u0026amp; Sheils, O. M. BRAFV600E: Implications for carcinogenesis and molecular therapy. \u003cem\u003eMolecular Cancer Therapeutics\u003c/em\u003e vol. 10 385\u0026ndash;394 (2011).\u003c/li\u003e\n\u003cli\u003ePratilas, C. A. \u003cem\u003eet al.\u003c/em\u003e V600EBRAF is associated with disabled feedback inhibition of RAF-MEK signaling and elevated transcriptional output of the pathway. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 4519\u0026ndash;4524 (2009).\u003c/li\u003e\n\u003cli\u003eKopetz, S. \u003cem\u003eet al.\u003c/em\u003e Encorafenib, Binimetinib, and Cetuximab in BRAF V600E\u0026ndash;Mutated Colorectal Cancer . \u003cem\u003eN. Engl. J. Med.\u003c/em\u003e \u003cstrong\u003e381\u003c/strong\u003e, 1632\u0026ndash;1643 (2019).\u003c/li\u003e\n\u003cli\u003ePrahallad, A. \u003cem\u003eet al.\u003c/em\u003e Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e483\u003c/strong\u003e, 100\u0026ndash;104 (2012).\u003c/li\u003e\n\u003cli\u003eCorcoran, R. B. \u003cem\u003eet al.\u003c/em\u003e EGFR-mediated reactivation of MAPK signaling contributes to insensitivity of BRAF-mutant colorectal cancers to RAF inhibition with vemurafenib. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, (2012).\u003c/li\u003e\n\u003cli\u003eYaeger, R. D. \u003cem\u003eet al.\u003c/em\u003e Pilot study of vemurafenib and panitumumab combination therapy in patients with BRAF V600E mutated metastatic colorectal cancer. \u003cem\u003eJ. Clin. Oncol.\u003c/em\u003e (2015) doi:10.1200/jco.2015.33.3_suppl.611.\u003c/li\u003e\n\u003cli\u003eVan Geel, R. M. J. M. \u003cem\u003eet al.\u003c/em\u003e A phase Ib dose-escalation study of encorafenib and cetuximab with or without alpelisib in metastatic BRAF-mutant colorectal cancer. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 610\u0026ndash;619 (2017).\u003c/li\u003e\n\u003cli\u003eXu, T. \u003cem\u003eet al.\u003c/em\u003e Molecular mechanisms underlying the resistance of BRAF V600E-mutant metastatic colorectal cancer to EGFR/BRAF inhibitors. \u003cem\u003eTher. Adv. Med. Oncol.\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 175883592211050 (2022).\u003c/li\u003e\n\u003cli\u003eAhronian, L. G. \u003cem\u003eet al.\u003c/em\u003e Clinical acquired resistance to RAF inhibitor combinations in BRAF-mutant colorectal cancer through MAPK pathway alterations. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 358\u0026ndash;367 (2015).\u003c/li\u003e\n\u003cli\u003eYaeger, R. \u003cem\u003eet al.\u003c/em\u003e Mechanisms of acquired resistance to BRAF V600E inhibition in colon cancers converge on RAF dimerization and are sensitive to its inhibition. \u003cem\u003eCancer Res.\u003c/em\u003e \u003cstrong\u003e77\u003c/strong\u003e, 6513\u0026ndash;6523 (2017).\u003c/li\u003e\n\u003cli\u003eCorcoran, R. B. \u003cem\u003eet al.\u003c/em\u003e Combined BRAF, EGFR, and MEK inhibition in patients with BRAF V600E -mutant colorectal cancer. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 428\u0026ndash;443 (2018).\u003c/li\u003e\n\u003cli\u003eHong, D. S. \u003cem\u003eet al.\u003c/em\u003e Phase 1B study of vemurafenib in combination with irinotecan and cetuximab in patients with BRAF -mutated advanced cancers and metastatic colorectal cancer. \u003cem\u003eJ. Clin. Oncol.\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 3516\u0026ndash;3516 (2014).\u003c/li\u003e\n\u003cli\u003ePietrantonio, F. \u003cem\u003eet al.\u003c/em\u003e MET-driven resistance to dual EGFR and BRAF blockade may be overcome by switching from EGFR to MET inhibition in BRAF-mutated colorectal cancer. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 963\u0026ndash;971 (2016).\u003c/li\u003e\n\u003cli\u003eRos, J. \u003cem\u003eet al.\u003c/em\u003e Encorafenib plus cetuximab for the treatment of BRAF-V600E-mutated metastatic colorectal cancer. \u003cem\u003eTherap. Adv. Gastroenterol.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 175628482211106 (2022).\u003c/li\u003e\n\u003cli\u003eKopetz, S. \u003cem\u003eet al.\u003c/em\u003e Molecular profiling of BRAF-V600E-mutant metastatic colorectal cancer in the phase 3 BEACON CRC trial. \u003cem\u003eNat. Med.\u003c/em\u003e (2024) doi:10.1038/s41591-024-03235-9.\u003c/li\u003e\n\u003cli\u003eKopetz, S. \u003cem\u003eet al.\u003c/em\u003e 316O Genomic mechanisms of acquired resistance of patients (pts) with BRAF V600E-mutant (mt) metastatic colorectal cancer (mCRC) treated in the BEACON study. \u003cem\u003eAnn. Oncol.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eTian, J. \u003cem\u003eet al.\u003c/em\u003e Combined PD-1, BRAF and MEK inhibition in BRAFV600E colorectal cancer: a phase 2 trial. \u003cem\u003eNat. Med.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 458\u0026ndash;466 (2023).\u003c/li\u003e\n\u003cli\u003eCorcoran, R. B. \u003cem\u003eet al.\u003c/em\u003e Combined BRAF, EGFR, and MEK Inhibition in Patients With BRAFV600E-Mutant Colorectal Cancer. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 428 (2018).\u003c/li\u003e\n\u003cli\u003eWeisenberger, D. J. \u003cem\u003eet al.\u003c/em\u003e CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. \u003cem\u003eNat. Genet.\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 787\u0026ndash;793 (2006).\u003c/li\u003e\n\u003cli\u003eHazar-Rethinam, M. \u003cem\u003eet al.\u003c/em\u003e Convergent therapeutic strategies to overcome the heterogeneity of acquired resistance in BRAFv600e colorectal cancer. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 417\u0026ndash;427 (2018).\u003c/li\u003e\n\u003cli\u003eAdini, A. \u003cem\u003eet al.\u003c/em\u003e The stem cell marker prominin-1/CD133 interacts with vascular endothelial growth factor and potentiates its action. \u003cem\u003eAngiogenesis\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 405\u0026ndash;416 (2013).\u003c/li\u003e\n\u003cli\u003eRicci-Vitiani, L. \u003cem\u003eet al.\u003c/em\u003e Identification and expansion of human colon-cancer-initiating cells. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e445\u003c/strong\u003e, 111\u0026ndash;115 (2007).\u003c/li\u003e\n\u003cli\u003eRicci-Vitiani, L., Fabrizi, E., Palio, E. \u0026amp; De Maria, R. Colon cancer stem cells. \u003cem\u003eJ. Mol. Med.\u003c/em\u003e \u003cstrong\u003e87\u003c/strong\u003e, 1097\u0026ndash;1104 (2009).\u003c/li\u003e\n\u003cli\u003eKim, Y. S. \u0026amp; Ho, S. B. Intestinal goblet cells and mucins in health and disease: Recent insights and progress. \u003cem\u003eCurrent Gastroenterology Reports\u003c/em\u003e vol. 12 319\u0026ndash;330 (2010).\u003c/li\u003e\n\u003cli\u003eVanDussen, K. L. \u0026amp; Samuelson, L. C. Mouse atonal homolog 1 directs intestinal progenitors to secretory cell rather than absorptive cell fate. \u003cem\u003eDev. Biol.\u003c/em\u003e \u003cstrong\u003e346\u003c/strong\u003e, 215\u0026ndash;223 (2010).\u003c/li\u003e\n\u003cli\u003eYang, Q., Bermingham, N. A., Finegold, M. J. \u0026amp; Zoghbi, H. Y. Requirement of Math1 for secretory cell lineage commitment in the mouse intestine. \u003cem\u003eScience (80-. ).\u003c/em\u003e \u003cstrong\u003e294\u003c/strong\u003e, 2155\u0026ndash;2158 (2001).\u003c/li\u003e\n\u003cli\u003eLupo, B. \u003cem\u003eet al.\u003c/em\u003e Colorectal cancer residual disease at maximal response to EGFR blockade displays a druggable Paneth cell\u0026ndash;like phenotype. \u003cem\u003eSci. Transl. Med.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eTakahashi-Yanaga, F. \u003cem\u003eet al.\u003c/em\u003e Glycogen synthase kinase-3\u0026beta; is tyrosine-phosphorylated by MEK1 in human skin fibroblasts. \u003cem\u003eBiochem. Biophys. Res. Commun.\u003c/em\u003e \u003cstrong\u003e316\u003c/strong\u003e, 411\u0026ndash;415 (2004).\u003c/li\u003e\n\u003cli\u003eClevers, H. Wnt/\u0026beta;-Catenin Signaling in Development and Disease. \u003cem\u003eCell\u003c/em\u003e vol. 127 469\u0026ndash;480 (2006).\u003c/li\u003e\n\u003cli\u003eClevers, H. \u0026amp; Nusse, R. Wnt/\u0026beta;-catenin signaling and disease. \u003cem\u003eCell\u003c/em\u003e vol. 149 1192\u0026ndash;1205 (2012).\u003c/li\u003e\n\u003cli\u003eLau, T. \u003cem\u003eet al.\u003c/em\u003e A novel tankyrase small-molecule inhibitor suppresses APC mutation-driven colorectal tumor growth. \u003cem\u003eCancer Res.\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 3132\u0026ndash;3144 (2013).\u003c/li\u003e\n\u003cli\u003eHurwitz, H. \u003cem\u003eet al.\u003c/em\u003e Bevacizumab plus Irinotecan, Fluorouracil, and Leucovorin for Metastatic Colorectal Cancer. \u003cem\u003eN. Engl. J. Med.\u003c/em\u003e \u003cstrong\u003e350\u003c/strong\u003e, 2335\u0026ndash;2342 (2004).\u003c/li\u003e\n\u003cli\u003eYamamoto, M., Bharti, A., Li, Y. \u0026amp; Kufe, D. Interaction of the DF3/MUC1 breast carcinoma-associated antigen and \u0026beta;- catenin in cell adhesion. \u003cem\u003eJ. Biol. Chem.\u003c/em\u003e \u003cstrong\u003e272\u003c/strong\u003e, 12492\u0026ndash;12494 (1997).\u003c/li\u003e\n\u003cli\u003ePai, P. \u003cem\u003eet al.\u003c/em\u003e The canonical Wnt pathway regulates the metastasis-promoting mucin MUC4 in pancreatic ductal adenocarcinoma. \u003cem\u003eMol. Oncol.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 224\u0026ndash;239 (2016).\u003c/li\u003e\n\u003cli\u003eZhan, T., Rindtorff, N. \u0026amp; Boutros, M. Wnt signaling in cancer. \u003cem\u003eOncogene\u003c/em\u003e vol. 36 1461\u0026ndash;1473 (2017).\u003c/li\u003e\n\u003cli\u003eTenbaum, S. P. \u003cem\u003eet al.\u003c/em\u003e \u0026beta;-Catenin confers resistance to PI3K and AKT inhibitors and subverts FOXO3a to promote metastasis in colon cancer. \u003cem\u003eNat. Med.\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 892\u0026ndash;901 (2012).\u003c/li\u003e\n\u003cli\u003eBademler, S. \u003cem\u003eet al.\u003c/em\u003e Clinical Significance of Serum Membrane-Bound Mucin-2 Levels in Breast Cancer. \u003cem\u003eBiomolecules\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 40 (2019).\u003c/li\u003e\n\u003cli\u003eYin, B. W. T., Dnistrian, A. \u0026amp; Lloyd, K. O. Ovarian cancer antigen CA125 is encoded by the MUC16 mucin gene. \u003cem\u003eInt. J. Cancer\u003c/em\u003e \u003cstrong\u003e98\u003c/strong\u003e, 737\u0026ndash;740 (2002).\u003c/li\u003e\n\u003cli\u003eNabhan, A.N.; Brownfield, D.G.; Harbury, P.B.; Krasnow, M.A.; Desai, T.J. Single-cell Wnt signaling niches maintain stemness of alveolar type 2 cells. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e2018\u003c/strong\u003e, \u003cem\u003e359\u003c/em\u003e, 1118\u0026ndash;1123. \u003c/li\u003e\n\u003cli\u003eDesai, T.J.; Brownfield, D.G.; Krasnow, M.A. Alveolar progenitor and stem cells in lung development, renewal and cancer. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e2014\u003c/strong\u003e, \u003cem\u003e507\u003c/em\u003e, 190\u0026ndash;194. \u003c/li\u003e\n\u003cli\u003eDi Martino, J. S. \u003cem\u003eet al.\u003c/em\u003e A tumor-derived type III collagen-rich ECM niche regulates tumor cell dormancy. \u003cem\u003eNat. Cancer\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 90\u0026ndash;107 (2022).\u003c/li\u003e\n\u003cli\u003eLiotta, L. A. \u0026amp; Kohn, E. C. The microenvironment of the tumour - Host interface. \u003cem\u003eNature\u003c/em\u003e vol. 411 375\u0026ndash;379 (2001).\u003c/li\u003e\n\u003cli\u003eShen, S., Vagner, S. \u0026amp; Robert, C. Persistent Cancer Cells: The Deadly Survivors. \u003cem\u003eCell\u003c/em\u003e vol. 183 860\u0026ndash;874 (2020).\u003c/li\u003e\n\u003cli\u003e\u0026Scaron;karkov\u0026aacute;, A. \u003cem\u003eet al.\u003c/em\u003e Educate, not kill: treating cancer without triggering its defenses. \u003cem\u003eTrends Mol. Med.\u003c/em\u003e (2024) doi:10.1016/J.MOLMED.2024.04.003.\u003c/li\u003e\n\u003cli\u003eWoo, J. K. \u003cem\u003eet al.\u003c/em\u003e Mucin 1 enhances the tumor angiogenic response by activation of the AKT signaling pathway. \u003cem\u003eOncogene\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 2187\u0026ndash;2198 (2012).\u003c/li\u003e\n\u003cli\u003eGiuliano, C. J., Lin, A., Girish, V. \u0026amp; Sheltzer, J. M. Generating Single Cell\u0026ndash;Derived Knockout Clones in Mammalian Cells with CRISPR/Cas9. \u003cem\u003eCurr. Protoc. Mol. Biol.\u003c/em\u003e \u003cstrong\u003e128\u003c/strong\u003e, e100 (2019).\u003c/li\u003e\n\u003cli\u003eSato, T. \u003cem\u003eet al.\u003c/em\u003e Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett\u0026rsquo;s epithelium. \u003cem\u003eGastroenterology\u003c/em\u003e \u003cstrong\u003e141\u003c/strong\u003e, 1762\u0026ndash;1772 (2011).\u003c/li\u003e\n\u003cli\u003eGao, H. \u003cem\u003eet al.\u003c/em\u003e High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response. \u003cem\u003eNat. Med.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 1318\u0026ndash;1325 (2015).\u003c/li\u003e\n\u003cli\u003ePuig, I. \u003cem\u003eet al.\u003c/em\u003e A personalized preclinical model to evaluate the metastatic potential of patient-derived colon cancer initiating cells. \u003cem\u003eClin. Cancer Res.\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 6787\u0026ndash;6801 (2013).\u003c/li\u003e\n\u003cli\u003eEisenhauer, E. A. \u003cem\u003eet al.\u003c/em\u003e New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). \u003cem\u003eEur. J. Cancer\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 228\u0026ndash;247 (2009).\u003c/li\u003e\n\u003cli\u003eSaltzman, A. B. \u003cem\u003eet al.\u003c/em\u003e GpGrouper: A peptide grouping algorithm for gene-centric inference and quantitation of bottom-up proteomics data. \u003cem\u003eMol. Cell. Proteomics\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 2270\u0026ndash;2283 (2018).\u003c/li\u003e\n\u003cli\u003eTyanova, S. \u003cem\u003eet al.\u003c/em\u003e The Perseus computational platform for comprehensive analysis of (prote)omics data. \u003cem\u003eNat. Methods 2016 139\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 731\u0026ndash;740 (2016).\u003c/li\u003e\n\u003cli\u003ePerez-Riverol, Y. \u003cem\u003eet al.\u003c/em\u003e The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, D543 (2022).\u003c/li\u003e\n\u003cli\u003eSubramanian, A. \u003cem\u003eet al.\u003c/em\u003e Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e \u003cstrong\u003e102\u003c/strong\u003e, 15545\u0026ndash;15550 (2005).\u003c/li\u003e\n\u003cli\u003eYu, G., Wang, L. G., Han, Y. \u0026amp; He, Q. Y. ClusterProfiler: An R package for comparing biological themes among gene clusters. \u003cem\u003eOmi. A J. Integr. Biol.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 284\u0026ndash;287 (2012).\u003c/li\u003e\n\u003cli\u003eWershof, E. \u003cem\u003eet al.\u003c/em\u003e A FIJI macro for quantifying pattern in extracellular matrix. \u003cem\u003eLife Sci. Alliance\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eGautier, L., Cope, L., Bolstad, B. M. \u0026amp; Irizarry, R. A. Affy - Analysis of Affymetrix GeneChip data at the probe level. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 307\u0026ndash;315 (2004).\u003c/li\u003e\n\u003cli\u003eRitchie, M. E. \u003cem\u003eet al.\u003c/em\u003e Limma powers differential expression analyses for RNA-sequencing and microarray studies. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e, e47 (2015).\u003c/li\u003e\n\u003cli\u003eWu, T. \u003cem\u003eet al.\u003c/em\u003e clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. \u003cem\u003eInnovation\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eGu, Z., Eils, R. \u0026amp; Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 2847\u0026ndash;2849 (2016).\u003c/li\u003e\n\u003cli\u003eGu, Z. Complex heatmap visualization. \u003cem\u003eiMeta\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, e43 (2022).\u003c/li\u003e\n\u003cli\u003eArqu\u0026eacute;s, O. GSEAmining: Make Biological Sense of Gene Set Enrichment Analysis Outputs. \u003cem\u003eBioconductor open source Softw. Bioinforma.\u003c/em\u003e (2023).\u003c/li\u003e\n\u003cli\u003eBrunson, Jason Cory and Read, Q. D. Alluvial Plots in ggplot2. Rpackage version 0.12.5. \u003cem\u003eThe Comprehensive R Archive Network\u003c/em\u003e https://corybrunson.github.io/ggalluvial/articles/ggalluvial.html (2023).\u003c/li\u003e\n\u003cli\u003eYu G. enrichplot: Visualization of Functional Enrichment Result. R package version 1.20.1. (2023).\u003c/li\u003e\n\u003cli\u003eBurclaff, J. \u003cem\u003eet al.\u003c/em\u003e A Proximal-to-Distal Survey of Healthy Adult Human Small Intestine and Colon Epithelium by Single-Cell Transcriptomics. \u003cem\u003eCmgh\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1554\u0026ndash;1589 (2022).\u003c/li\u003e\n\u003cli\u003eFranz\u0026eacute;n, O., Gan, L. M. \u0026amp; Bj\u0026ouml;rkegren, J. L. M. PanglaoDB: A web server for exploration of mouse and human single-cell RNA sequencing data. \u003cem\u003eDatabase\u003c/em\u003e \u003cstrong\u003e2019\u003c/strong\u003e, 1\u0026ndash;9 (2019).\u003c/li\u003e\n\u003cli\u003eSebastian, A. \u003cem\u003eet al.\u003c/em\u003e Single-cell transcriptomic analysis of tumor- derived fibroblasts and normal tissue-resident fibroblasts reveals fibroblast heterogeneity in breast cancer. \u003cem\u003eCancers (Basel).\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eBushnell, B. BBMap: A Fast, Accurate, Splice-Aware Aligner (No. LBNL-7065E). in \u003cem\u003eBerkeley National Lab. (LBNL), Berkeley, CA (United States).\u003c/em\u003e (Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)., 2014).\u003c/li\u003e\n\u003cli\u003eNystedt, B. \u003cem\u003eet al.\u003c/em\u003e Sarek: A portable workflow for whole-genome sequencing analysis of germline and somatic variants. \u003cem\u003eF1000Research\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eAndrews, S. \u0026amp; others. FastQC: a quality control tool for high throughput sequence data. 2010. \u003cem\u003eHttps://Www.Bioinformatics.Babraham.Ac.Uk/Projects/Fastqc/\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, http://www.bioinformatics.babraham.ac.uk/projects/ (2010).\u003c/li\u003e\n\u003cli\u003eChen, S., Zhou, Y., Chen, Y. \u0026amp; Gu, J. Fastp: An ultra-fast all-in-one FASTQ preprocessor. in \u003cem\u003eBioinformatics\u003c/em\u003e vol. 34 i884\u0026ndash;i890 (Oxford Academic, 2018).\u003c/li\u003e\n\u003cli\u003eLi, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. (2013).\u003c/li\u003e\n\u003cli\u003eMcKenna, A. \u003cem\u003eet al.\u003c/em\u003e The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. \u003cem\u003eGenome Res.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 1297\u0026ndash;1303 (2010).\u003c/li\u003e\n\u003cli\u003eVan der Auwera, G. A. \u003cem\u003eet al.\u003c/em\u003e From fastQ data to high-confidence variant calls: The genome analysis toolkit best practices pipeline. \u003cem\u003eCurr. Protoc. Bioinforma.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 11.10.1 (2013).\u003c/li\u003e\n\u003cli\u003ePicard Tools - By Broad Institute. https://broadinstitute.github.io/picard/.\u003c/li\u003e\n\u003cli\u003eCibulskis, K. \u003cem\u003eet al.\u003c/em\u003e Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. \u003cem\u003eNat. Biotechnol.\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 213\u0026ndash;219 (2013).\u003c/li\u003e\n\u003cli\u003eMills, R. E. \u003cem\u003eet al.\u003c/em\u003e An initial map of insertion and deletion (INDEL) variation in the human genome. \u003cem\u003eGenome Res.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1182\u0026ndash;1190 (2006).\u003c/li\u003e\n\u003cli\u003eSherry, S. T. \u003cem\u003eet al.\u003c/em\u003e DbSNP: The NCBI database of genetic variation. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 308\u0026ndash;311 (2001).\u003c/li\u003e\n\u003cli\u003eCingolani, P. \u003cem\u003eet al.\u003c/em\u003e A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. \u003cem\u003eFly (Austin).\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 80\u0026ndash;92 (2012).\u003c/li\u003e\n\u003cli\u003eMayakonda, A., Lin, D. C., Assenov, Y., Plass, C. \u0026amp; Koeffler, H. P. Maftools: Efficient and comprehensive analysis of somatic variants in cancer. \u003cem\u003eGenome Res.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 1747\u0026ndash;1756 (2018).\u003c/li\u003e\n\u003cli\u003eCerami, E. \u003cem\u003eet al.\u003c/em\u003e The cBio Cancer Genomics Portal: An open platform for exploring multidimensional cancer genomics data. \u003cem\u003eCancer Discov.\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 401\u0026ndash;404 (2012).\u003c/li\u003e\n\u003cli\u003eD\u0026iacute;az-Gay, M. \u003cem\u003eet al.\u003c/em\u003e Assigning mutational signatures to individual samples and individual somatic mutations with SigProfilerAssignment. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e39\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eTate, J. G. \u003cem\u003eet al.\u003c/em\u003e COSMIC: The Catalogue Of Somatic Mutations In Cancer. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, D941\u0026ndash;D947 (2019).\u003c/li\u003e\n\u003cli\u003eCyriac Kandoth. mskcc/vcf2maf: vcf2maf v1.6.19. (2020) doi:10.5281/ZENODO.593251.\u003c/li\u003e\n\u003cli\u003eMoran, S., Arribas, C. \u0026amp; Esteller, M. Validation of a DNA methylation microarray for 850,000 CpG sites of the human genome enriched in enhancer sequences. \u003cem\u003eEpigenomics\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 389\u0026ndash;399 (2016).\u003c/li\u003e\n\u003cli\u003ePidsley, R. \u003cem\u003eet al.\u003c/em\u003e Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. \u003cem\u003eGenome Biol.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, (2016).\u003c/li\u003e\n\u003cli\u003eR Core Team. R Core Team 2021 R: A language and environment for statistical computing. R foundation for statistical computing. https://www.R-project.org/. \u003cem\u003eR Found. Stat. Comput.\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 2019 (2022).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"BRAFV600E, acquired drug resistance, encorafenib, tissue remodeling, bevacizumab, colorectal cancer","lastPublishedDoi":"10.21203/rs.3.rs-6770441/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6770441/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe combination of BRAF\u003csup\u003eV600E\u003c/sup\u003e and EGFR inhibitors benefits advanced colorectal cancer (CRC) patients, but resistance can develop through non-genetic mechanisms not fully understood. This study reveals that \u003cem\u003eBRAF\u003c/em\u003e mutant CRC tumors undergo a tissue remodeling to resist BRAF\u003csup\u003eV600E\u003c/sup\u003e inhibitors, characterized by tumor cell mucinous differentiation and extracellular matrix transformation, allowing an increased infiltration of activated fibroblasts, immune cells, and vasculature. Some of these changes are essential for acquiring resistance. In fact, we demonstrate that blocking new vascularization with the anti-angiogenic antibody bevacizumab against VEGFA extends the benefit of BRAF inhibitory therapies in CRC models. These findings are based on patient-derived xenograft (PDX) models and validated in patient samples, offering deeper insights into resistance mechanisms and suggesting rational combinations to prolong therapy effectiveness. A clinical trial is being initiated to test the combination of bevacizumab and BRAF inhibitory therapy for improving CRC patients\u0026rsquo; treatment.\u003c/p\u003e","manuscriptTitle":"Acquired resistance to BRAF inhibitory treatments requires tumor tissue remodeling and reveals targetable vulnerabilities in colorectal cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-25 05:31:59","doi":"10.21203/rs.3.rs-6770441/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e90e7a51-abd6-4f1d-ac4c-2daedf8cf555","owner":[],"postedDate":"June 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":49836945,"name":"Health sciences/Oncology/Cancer/Cancer therapy/Cancer therapeutic resistance"},{"id":49836946,"name":"Health sciences/Oncology"}],"tags":[],"updatedAt":"2025-08-22T15:14:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-25 05:31:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6770441","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6770441","identity":"rs-6770441","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}