Search

Rectal carcinoma, representing about a third of all newly diagnosed colorectal cancers, is one of the most common malignant tumors. The standard of care for locally advanced rectal cancer (LARC), consisting of neoadjuvant chemo/radiotherapy prior to surgical resection, is poorly effective, leading to complete tumor regression only in 10-30% of the cases. Approximately 40% of LARC harbor activating KRAS mutations, which have been extensively associated with primary resistance to targeted EGFR therapy and with radioresistance as well. Previous work showed that the MET receptor tyrosine kinase supports radioresistance and can be inhibited to radiosensitize tumor cell subpopulations retaining stem-like properties. Here, we show that LARC often express high levels of MET and can be successfully radiosensitized by MET inhibition. This was assessed in rectal stem-like cells isolated from human tumors (rectospheres), and transplanted in the mouse to regenerate tumors that faithfully reproduce the phenotype, the genotype and the therapeutic response of the original tumor. Mechanistically, we found that radioresistant KRAS-mutant rectospheres display significantly higher basal levels of RAD51, a master regulator of DNA homologous recombination repair, and increased RAD51 recruitment to irradiation-induced DNA double-strand breaks, as compared with KRAS-wild type rectospheres. Importantly, we showed that MET pharmacological inhibition by small-molecule kinase inhibitors, combined with radiotherapy, impairs RAD51 expression and function. This leads to DNA damage accumulation and results in effective radiosensitization of K-RAS-mutant rectal stem-like cells in vitro and in vivo, and significant inhibition of experimental tumors. Therefore, preclinical evidence is provided that MET can be exploited as a therapeutic target to radiosensitize KRAS-mutant rectal cancer at stem-like cell level. Citation Format: Antonio D'Ambrosio, Federica Verginelli, Francesca Orzan, Raffaella Albano, Elena Casanova, Paolo Luraghi, Francesca De Bacco, Andrea Bertotti, Livio Trusolino, Rebecca Senetta, Anna Sapino, Elena Del Mastro, Marco Gatti, Paolo M. Comoglio, Carla Boccaccio. MET inhibition radiosensitizes KRAS-mutant rectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1387.

In colorectal cancer, the genetic mechanisms underlying the metastatic switch are still poorly understood. Here we show that overexpression of miRNA-483-3p, encoded by a frequently amplified gene locus encompassing also insulin-like growth factor 2, confers an aggressive phenotype to stem-like cells derived from colorectal cancer metastases (m-colospheres). As result of miRNA-483-3p ectopic overexpression, m-colospheres displayed (i) increased proliferative response to exogenous EGFR family ligands EGF and NRG1; (ii) increased spontaneous and growth factor-induced in vitro invasiveness and epithelial-mesenchymal transition (EMT); (iii) enhanced stem-cell frequency and resistance to differentiation. By transcriptomic analyses and functional validation, we found that miRNA-483-3p targets NDRG1, a known ‘metastasis suppressor', which is responsible for degradation of EGFR family members, in particular HER3. As result, ectopic or native miRNA-483-3p overexpression was associated with hyper-activation of the signaling pathway triggered by HER3, including AKT and GSK3β, responsible for activation of EMT transcription factors. Consistently, treatment of miRNA-483-3p overexpressing m-colospheres with HER3 specific antibodies counteracted their proliferative and invasive phenotype. The pro-invasive role of miRNA-483-3p in patients was further confirmed by (i) analysis of colorectal tumors, where miRNA-483-3p expression levels directly correlated with expression of EMT transcription factors and poor-prognosis, and (ii) downregulation of naturally occurring miRNA-483-3p overexpression, which prevented invasion of tumors formed by m-colosphere transplantation. These results indicate that miRNA-483-3p can support colorectal cancer invasion through a signaling pathway amenable to targeting in human patients. Citation Format: Ermes Candiello, Gigliola Reato, Federica Verginelli, Antonio D'Ambrosio, Gennaro Gambardella, Raffaella Albano, Paolo Luraghi, Paolo M. Comoglio, Andrea Bertotti, Livio Trusolino, Carla Boccaccio. miRNA-483-3p overexpression unleashes invasiveness of metastatic colorectal cancer by NDRG1 targeting and upregulation of the HER3-AKT axis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2358.

Background: Moving stage III Colon Cancer (CC) into the precision medicine space is a priority in view of the lack of molecular markers driving adjuvant treatment. Retrospective studies have demonstrated the tremendous prognostic impact of circulating tumor DNA (ctDNA) analysis after curative intent surgery, and suggested that lack of conversion of ctDNA from detectable to undetectable after adjuvant chemotherapy reflects treatment failure. With these premises, we have designed the PEGASUS trial (EudraCT 2019-002074-32) to prove the feasibility of using liquid biopsy to guide the post-surgical and post-adjuvant clinical management of early colon cancer patients. Methods: PEGASUS is a prospective multicentric vanguard study designed to test the feasibility of performing serial interventional liquid biopsies (LB) for ctDNA determination in 140 microsatellite stable Stage-III and T4N0 Stage-II CC patients. The LUNAR1 (Guardant Health, Redwood City, CA, USA) will be used for ctDNA determination. For the efficacy analysis, the PEGASUS cohort will be compared with a cohort of 420 patients from the TOSCA trial (NCT00646607) population matched 3:1 for all known prognostic phenotypes. A post-surgical LB executed 2-4 weeks after surgery will guide a “Molecular Adjuvant” treatment as follows: i) ctDNA+ patients will receive CAPOX for 3 months and ii) ctDNA- patients will receive capecitabine (CAPE) for 6 months but will be retested after 1 cycle, and if found ctDNA+ will be switched to CAPOX treatment. Immediately after the end of “Molecular Adjuvant” treatment a further LB will be performed and instruct subsequent treatment. Positive patients (ctDNA+/+ and ctDNA-/+) will receive an up-scaled “Molecular Metastatic” systemic treatment for 6 months or until radiological progression or toxicity as follows: i) ctDNA+/+ patients will be treated with FOLFIRI; ii) ctDNA-/+ patients with CAPOX. These patients will be subjected to a LB after 3 months and at the end of treatment and in case of positivity will be switched to FOLFIRI. Patients experiencing ctDNA conversion to negative (ctDNA+/-) will receive a de-escalated treatment with CAPE for 3 months. 3 LB will be performed within 3 months and in case of positivity the patient will be switched to FOLFIRI. Patients with ctDNA-/- will be subjected to an interventional follow-up comprising 2 further LB and in case of positivity they will be switched to CAPOX treatment. PEGASUS is piggybacked to AlfaOmega (NCT04120935), a Master Observational Protocol that will follow patients from diagnosis to 5 years or recurrence/death (whichever comes first), collecting clinical data, radio images and biological samples. AlfaOmega provides a clinical and logistic ecosystem for the seamless integration of PEGASUS clinical results with the biological underpinning of colon cancer. Citation Format: Elena Elez, Filippo Pietrantonio, Andrea Sartore-Bianchi, Clara Montagut, Andres Cervantes, Stefania Sciallero, Maria Giulia Zampino, Stefania Mosconi, Valter Torri, Noelia Tarazona, Luca Lazzari, Paolo Luraghi, Susana Muñoz, Matteo Fassan, Enzo Medico, Livio Trusolino, Maria Rescigno, Salvatore Siena, Alberto Sobrero, Roberto Labianca, Josep Tabernero, Alberto Bardelli, Sara Lonardi, Silvia Marsoni. Post-surgical liquid biopsy-guided treatment of stage III and high-risk stage II colon cancer patients: The PEGASUS trial [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr CT263.

Background: Breast cancer (BC) is the most frequent tumor in women worldwide. BC lethality is caused by aggressive, therapy-resistant metastases (mBC). Preliminary data have shown that mBC lesions are invariably embedded into a densely packed network of fibrous extracellular matrix, making the metastatic microenvironment a potent inducer of mechanical inputs, ultimately leading to the activation of the transcription factors YAP/TAZ. Aberrant mechano-signaling could thus represent a vulnerability of metastasis, which can be exploited to develop new therapeutic strategies. To investigate how metastatic outgrowth is regulated by the physical properties of the microenvironment, and how the altered mechano-transduction of human BC metastasis can be exploited to define new unconventional therapies, we designed and activated METAMECH, a Master Observational Trial (MOT). MOTs are trial constructs hybridizing a canonical observational trial with a multiplex sample-collection platform, aimed to empower the bi-directional collaboration between pre-clinical and clinical research, an essential prerequisite to feed and implement precision oncology. Methods: METAMECH (IFOM-CPO007/2019/PO006) is a MOT that will follow a stage-mixed cohort of at least 500 BC patients throughout their course of treatments, until death or a minimum of 5 years. METAMECH has been designed to streamline the study of the co-evolutionary landscape between tumor and host cells across all BC subtypes. METAMECH is a resource for integrative clinical and imaging data and fresh and archival sample collection which will allow to auction mechanotransduction mechanisms supporting the outgrowth of human metastases, mine for new potentially actionable targets and permit the selection of appropriate patients for mechano-biology experimentally-driven trials. METAMECH is a multi-tiers research scaffold that will allow to integrate the clinical and basic aspects of the project, to optimize patients enrollment and the logistic of longitudinal collection of theirs data/samples. A customized data lake is being designed for data storage, analysis and retrieval. To facilitate patients access, METAMECH has been designed as a flexible infrastructure organized in TIERS, which allow to pursue different objectives: TIER0/1, Retrieving/Recording: to retrospectively and prospectively access to selected cohorts of clinically annotated BC samples to validate/discover mechanotransduction-linked biomarkers and create a mechano-classifier; TIER2, Modelling: to develop pertinent experimental models to study the aberrant mechanisms underlying the metastatic outgrowth and define mechanotransduction-targeting therapeutic strategies; TIER3, Linking: to access data and samples of patients enrolled in Proof-of-Concept trials to prove the efficacy and study/understand resistance mechanisms of mechanotransduction-targeting therapies. Citation Format: Paolo Luraghi, Luca Lazzari, Maria Vittoria Dieci, Serena Bellani, Chiara Soffientini, Rosario Rizzuto, Antonio Rosato, Vincenzo Costanzo, Giorgio Scita, Giannino Del Sal, Massimiliano Pagani, Marco Foiani, Claudio Tripodo, Silvio Bicciato, Giovanni Blandino, Alberto Zambelli, Valter Torri, Matteo Fassan, PierFranco Conte, Silvia Marsoni, Stefano Piccolo. METAMECH -A Master Observational Trial empowering mechanobiology translational research and mechanobased proof of concept trials in breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr CT261.

TPS4124 Background: Moving stage III Colon Cancer (CC) into the precision medicine space is a priority in view of the lack of molecular markers driving adjuvant treatment. Retrospective studies have demonstrated the tremendous prognostic impact of circulating tumor DNA (ctDNA) analysis after curative intent surgery, and suggested that lack of conversion of ctDNA from detectable to undetectable after adjuvant chemotherapy reflects treatment failure. With these premises, we have designed the PEGASUS trial (NCT04259944). Methods: PEGASUS is a prospective multicentric study designed to prove the feasibility of using liquid biopsy (LB) to guide the post-surgical and post-adjuvant clinical management in 140 microsatellite stable Stage-III and T4N0 Stage-II CC patients. The LUNAR1 test (Guardant Health, Redwood City, CA, USA) will be used for ctDNA determination. For the efficacy analysis, the PEGASUS cohort will be compared with a 3:1 matched cohort of 420 patients from the TOSCA trial (NCT00646607). A LB executed 2-4 weeks post-surgery will guide a “Molecular Adjuvant” treatment: i) ctDNA+ patients will receive CAPOX for 3 months and ii) ctDNA- patients will receive capecitabine (CAPE) for 6 months but will be retested after 1 cycle, and if found ctDNA+ will be switched to CAPOX treatment. At the end of the “Molecular Adjuvant” treatment a further LB will be performed and instruct subsequent treatment. Positive patients (ctDNA+/+ and ctDNA-/+) will receive an up-scaled “Molecular Metastatic” systemic treatment for 6 months or until radiological progression or toxicity: i) ctDNA+/+ patients will be treated with FOLFIRI; ii) ctDNA-/+ patients with CAPOX. These patients will be subjected to a LB after 3 months and at the end of treatment: in case of positivity will be switched to FOLFIRI. ctDNA+/- patientswill receive a de-escalated treatment with CAPE for 3 months. 3 LB will be performed within 3 months and in case of positivity the patient will be switched to FOLFIRI. Patients with ctDNA-/- will be subjected to an interventional follow-up comprising 2 further LB and in case of positivity they will be switched to CAPOX treatment. PEGASUS is piggybacked to AlfaOmega (NCT04120935), a Master Observational Protocol that will follow patients from diagnosis to 5 years or recurrence/death (whichever comes first), collecting clinical data, radio-images and biological samples. AlfaOmega provides a clinical and logistic ecosystem for the seamless integration of PEGASUS clinical results with the biological underpinning of colon cancer. Clinical trial information: NCT04259944 .

Background The overall survival of mCRC patients has been increased by the availability of new cytotoxic and targeted agents and today potentially by the advent of immunotherapies. However, the impact of these advances has been incremental rather than transformative, and a number of unmet medical needs still await rational solutions. In order to navigate the co-evolutionary pathways of CRC tumors during their natural history and under the Darwinian pressure of therapies, we need to feed the experimental laboratories with “the right sample, at the right time, for the right experiment”. This can be accomplished through the design of master protocols that represent a new subset of observational trials aimed to empower the bi-directional collaboration between pre-clinical and clinical research, an essential prerequisite to feed and implement precision oncology. Methods AlfaΩmega (protocol number IFOM-CPO003/2018/PO002) has been designed to streamline the study of the co-evolutionary landscape between tumor and host cells in a stage-mixed cohort of at least 500 CRC patients, with the aim of understanding how their outcomes can be significantly improved. This resource for integrative clinical data and sample collection will allow the molecular story-telling of CRC metastatic spread along time and space and the selection of appropriate patients for experimentally-driven trials. To achieve the required level of ‘experimental precision’, patients can enter AlfaΩmega at two different ‘therapeutic checkpoints’: i) prior to a surgical event or ii) prior to a systemic treatment. Moreover, to optimize the enrollment of patients, the longitudinal collection of data/samples and their logistic management, AlfaΩmega has been designed as a flexible infrastructure organized in TIERs, each of which represents a building block for the stepwise comprehension of the biological processes that drive tumor evolution and that will have an independent informed consensus process: § TIER1, Monitoring: the ability to follow CRC evolution under standard of care treatments and to define new evolution-linked biomarkers: access to clinical & imaging data, FFPE, plasma and PBMCs. TIER1 informed consent is mandatory for the enrolment. § TIER2, Modelling: the ability to develop pertinent experimental models to study evolutionary mechanisms and define evolution-targeting therapeutic strategies: access to fresh tissue, blood, stools, buccal swabs and other fluids. Sample collection in TIER2 is discretionary, i.e. may not be applied to all patients entering TIER1. § TIER3, Linking: the ability to access data and samples of patients enrolled in proof-of-concept trials to prove the efficacy and study/understand resistance mechanisms of evolution-targeting therapies: linking with AlfaΩmega logistic backbone and IT architecture. Citation Format: Luca Lazzari, Paolo Luraghi, Nabil Amirouchene-Angelozzi, Andrea Bertotti, Enzo Medico, Federica Di Nicolantonio, Fabrizio d'Adda di Fagagna, Sergio Abrignani, Daniele Regge, Anna Sapino, Emanuela Bonoldi, Angelo Vanzulli, Vittorina Zagonel, Filippo de Braud, Livio Trusolino, Alberto Bardelli, Salvatore Siena, Silvia Marsoni. AlfaOmega- a master protocol empowering precision research in colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr CT214.

TPS2659 Background: Metastatic colorectal cancer (CRC) harbouring genetic defects in the mismatch-repair pathway (MMRd) presents with a high tumor mutational burden (TMB), and is highly sensitive to anti–programmed cell death protein 1 (PD-1) immune checkpoint inhibitors. We recently showed in preclinical models that the pharmacological treatment with temozolomide (TMZ) can induce the inactivation of MMR genes, and consequently the increase of TMB and immunogenic neoantigens, thus suggesting that TMZ could be used to prime MMR proficient (MMRp) tumors for response to checkpoint inhibitors. Accordingly, mCRC patients recruited in previous clinical trials where TMZ was administered, acquired alterations of MMR genes upon treatment and showed remarkable increase in TMB at disease progression (PD). We thus designed the ARETHUSA clinical trial to test whether a priming course with TMZ in patients can sensitize mCRC to the anti-PD1 inhibitor pembrolizumab. Methods: Arethusa (NCT03519412) is a 2-cohorts, phase II trial consisting of three different phases. In the SCREENING, 348 mCRC RAS-mutated patients will be tested for MMR status. MMRd patients will proceed directly to TRIAL for immediate pembrolizumab treatment (expected 14). MMR-proficient (MMRp) patients will be further tested for expression of O6-methylguanine-DNA methyltransferase (MGMT) by immunohistochemistry and by promoter methylation analysis. IHC-negative, promoter methylation-positive MMRp patients (expected 67) will enter in the PRIMING phase and will be treated with TMZ until PD. TMB will then be assessed on tumor biopsies at resistance. Those patients that will have > 20 mutations/megabase will proceed to TRIAL (expected 20) and will be treated with pembrolizumab. Overall response rate (primary outcome), Progression Free, and Overall Survival, and treatment related toxicities (secondary outcomes) in MMRp pembrolizumab-treated patients will be estimated., while the MMRd cohort will be used for comparison. Tissue biopsies, longitudinal blood and stool collection will be used for discovery of predictive molecular biomarkers and assessment of tumor evolution. Clinical trial information: NCT03519412.

e15113 Background: Despite a reduction of Colorectal Cancer (CRC) incidence in western countries in the past decades, Early-Onset CRCs (EO-CRC, patients diagnosed with CRC ≤ 40 years old) incidence has increased. Although frequently occurring in the context of familial syndromes, EO-CRCs are mainly sporadic cases and phenotipically enriched for distal localisation and advanced stage at diagnosis. Whether EO-sporadic CRCs pathogenesis differs from that of normal-onset (NO) CRC and how this might impact incidence rates is currently unknown. This had prompted us to ask if, at the genetic level, “traces” of peculiar pathogenic processes could be identified in EO-specific (or -enriched) genetic signatures (GS). Methods: The mutational signatures of 424 TCGA CRC patient samples (19 EO and 405 ≥ 50 years at diagnosis) were analyzed and the similarity between each mutational profile and COSMIC GS was calculated using Bioconductor R package Mutational Patterns (doi: 10.1186/s13073-018-0539-0). Unsupervised hierarchical clustering of the samples according to similarity to GS was performed in single cohort and pooled analysis. Association between age and individual GS was assessed through grouped and linear correlation analysis. Results: EO-CRC patients were grouped in three main clusters: Cluster1 exposing a major similarity with GS10 (associated with defects in polymerase proofreading activity), Cluster2 showing stronger similarities with GS6 (associated with mismatch repair deficiency), and GS1 (associated with 5-MeC deamination) and Cluster3 presenting similarity with multiple GSs, such as 3 (associated with homologous recombination deficiency) and 5 (pathogenesis unknown). Overall this clustering was maintained when EO-CRC samples were pooled with NO CRC. Grouped and correlation analysis revealed no significant association between age and individual GSs, including GS1 (associated with age). Conclusions: These preliminary analyses show that the relative contribution of known GS is similar in EO and NO cohorts of patients. Possible enrichment for EO-CRC in specific signature clusters will be analysed on a wider sample series.

Background Angiogenesis inhibition is a promising approach for treating metastatic colorectal cancer (mCRC). Recent evidences support the seemingly counterintuitive ability of certain antiangiogenic drugs to promote normalization of residual tumor vessels with important clinical implications. Lenalidomide is an oral drug with immune-modulatory and anti-angiogenic activity against selected hematologic malignancies but as yet little is known regarding its effectiveness for solid tumors. The aim of this study was to determine whether lenalidomide can normalize colorectal cancer neo-vessels in vivo, thus reducing tumor hypoxia and improving the benefit of chemotherapy. Methods We set up a tumorgraft model with NOD/SCID mice implanted with a patient-derived colorectal cancer liver metastasis. The mice were treated with oral lenalidomide (50 mg/Kg/day for 28 days), intraperitoneal 5-fluorouracil (5FU) (20 mg/Kg twice weekly for 3 weeks), combination (combo) of lenalidomide and 5FU or irrelevant vehicle. We assessed tumor vessel density (CD146), pericyte coverage (NG2; alphaSMA), in vivo perfusion capability of residual vessels (lectin distribution essay), hypoxic areas (HP2-100 Hypoxyprobe) and antitumor activity in vivo and in vitro. Results Treatment with lenalidomide reduced tumor vessel density (p = 0.0001) and enhanced mature pericyte coverage of residual vessels (p = 0.002). Perfusion capability of tumor vessels was enhanced in mice treated with lenalidomide compared to controls (p = 0.004). Accordingly, lenalidomide reduced hypoxic tumor areas (p = 0.002) and enhanced the antitumor activity of 5FU in vivo. The combo treatment delayed tumor growth (p = 0.01) and significantly reduced the Ki67 index (p = 0.0002). Lenalidomide alone did not demonstrate antitumor activity compared to untreated controls in vivo or against 4 different mCRC cell lines in vitro. Conclusions We provide the first evidence of tumor vessel normalization and hypoxia reduction induced by lenalidomide in mCRC in vivo. This effect, seemingly counterintuitive for an antiangiogenic compound, translates into indirect antitumor activity thus enhancing the therapeutic index of chemotherapy. Our findings suggest that further research should be carried out on synergism between lenalidomide and conventional therapies for treating solid tumors that might benefit from tumor vasculature normalization.

Metastatic colorectal cancer remains largely incurable, although in a subset of patients, survival is prolonged by new targeting agents such as anti-EGF receptor (anti-EGFR) antibodies. This disease is believed to be supported by a subpopulation of stem-like cells termed colon cancer–initiating cell (CCIC), which may also confer therapeutic resistance. However, how CCICs respond to EGFR inhibition has not been fully characterized. To explore this question, we systematically generated CCICs through spheroid cultures of patient-derived xenografts of metastatic colorectal cancer. These cultures, termed “xenospheres,” were capable of long-term self-propagation in vitro and phenocopied the original patient tumors in vivo, thus operationally defining CCICs. Xenosphere CCICs retained the genetic determinants for EGFR therapeutic response in vitro and in xenografts; like the original tumors, xenospheres harboring a mutated KRAS gene were resistant to EGFR therapy, whereas those harboring wild-type RAS pathway genes (RASwt) were sensitive. Notably, the effects of EGFR inhibition in sensitive CCICs could be counteracted by cytokines secreted by cancer-associated fibroblasts. In particular, we found that the MET receptor ligand hepatocyte growth factor (HGF) was especially active in supporting in vitro CCIC proliferation and resistance to EGFR inhibition. Ectopic production of human HGF in CCIC xenografts rendered the xenografts susceptible to MET inhibition, which sensitized the response to EGFR therapy. By showing that RASwt CCICs rely on both EGFR and MET signaling, our results offer a strong preclinical proof-of-concept for concurrent targeting of these two pathways in the clinical setting. Cancer Res; 74(6); 1857–69. ©2014 AACR.

Although EGFR targeted therapy has improved the survival outcome of patients with metastatic colorectal cancer, lack of response or emergence of secondary resistance frequently occurs as result of genetic alterations that hyperactivate the RAS pathway or by activation of compensatory pathways by signals from the tumor microenvironmet that can also induce a stem-like phenotype. From patient-derived xenografts (“xenopatients”) of colorectal cancer liver metastases, we generated primary sphere cultures (“xenospheres”) that display the properties of cancer stem-like cells. Indeed, these cells can long-term self-propagate in vitro, and form phenocopies of original patient tumors in vivo (“spheropatients”), maintaining the same response to anti-EGFR therapy. We obtained a panel of xenospheres derived from more than 40 different xenopatients, all genetically characterized for the presence of mutations in genes that are predictive for response to EGFR targeted therapy. The panel included xenospheres harboring mutations in RAS genes (Rasmut), or recently identified rare predictors of resistance: ERBB2 amplification/mutation, MET amplification, EGFR mutation in the extracellular domain, and IGF2 overexpression. Moreover, we conducted an extensive exome-sequencing analysis of 10 xenospheres, for which the exome data of the matched xenopatients were available, thus allowing a complete and robust comparison. Xenospheres displayed a good genetic stability and a remarkable correspondence with xenopatients, in both the mutational and gene expression profiles. While RASmut xenospheres displayed self-sustained proliferative ability, those lacking mutations affecting the RAS pathway (RASwt) were strongly dependent of exogenous growth factors. We thus assessed the response of all RASwt xenospheres to ligands of the EGF family (TGFA, EREG, AREG, HB-EGF, NRG1), or bFGF, or HGF, testing both proliferation and response to EGFR inhibition (cetuximab). We found that TGFA and HB-EGF induced proliferation and resistance to cetuximab with almost the same potency of EGF, while EREG and AREG had a weak mitogenic activity and did not protect against cetuximab treatment, accordingly with observations in patients and xenopatients. The different proliferative activity (and protection against cetuximab) can be at least in part explained by the effect on receptor stabilization and exposure to the cell membrane exerted by different ligands. Interestingly, we found that, among all EGF family ligands, NRG1 (binding ERBB3/4) had the most potent mitogenic activity and protection against cetuximab treatment. NRG1 was able to replace EGF in sustaining xenosphere propagation, without altering the global gene expression profile, implying that it could sustain cancer-stem like cell properties similarly to EGF. To study the role of NRG1 in vivo we generated both an autocrine and a paracrine mode, by inducing its expression infecting either RASwt xenospheres or a cell line of murine fibroblasts. NRG1-expressing RASwt xenospheres became growth factor independent, resistant to cetuximab but sensitive to lapatinib (a dual EGFR/ERBB2 small molecule). Similar results were obtained using the conditioned medium of NRG1-expressing murine fibroblasts. By injecting parental and modified RASwt xenospheres into immunocompromised mice, we found that NRG1-expressing xenospheres had a strongly increased tumor onset, and that these tumors were resistant to cetuximab but sensitive to lapatinib treatment, suggesting its possible use in the clinical practice of RASwt patients. We conclude that xenospheres are a robust patient-derived in vitro model of colorectal cancer-stem like cells, amenable to study molecular mechanisms of response or resistance to targeted therapies. Citation Format: Paolo Luraghi, Viola Bigatto, Gigliola Reato, Elia Cipriano, Francesco Sassi, Claudio Isella, Andrea Bertotti, Livio Trusolino, Paolo Maria Comoglio, Carla Boccaccio. Xenospheres: a comprehensive patient-derived in vitro model to study response and resistance to targeted therapies in metastatic colorectal cancer. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr B17.