Final Report Summary - PERSMEDOMICS (Bioinformatics and Integrative Genomics for a Novel Personalized Cancer Therapy)
Summary description of project objectives:
Pancreatic Ductal Adenocarcinoma (PDAC) is essentially a metastatic disease with one of the worst overall survival rates. Patients with metastatic pancreatic cancer are currently managed with systemic chemotherapy with poor results. Indeed, the median survival of these patients with more aggressive regimens in the 10-11 months range. Thus, an important need is the development of newer and more effective treatment modalities for this disease that result in an improvement in survival. It’s known that massively parallel sequencing data analysis such as WES (whole-exome sequencing) results in the identification of hundreds (sometimes thousands) of mutations. Only a small fraction of these alterations might have a clinical impact and their identification has become one of the biggest challenges to succesfully translate cancer genomic to the clinic. The current understanding of PDAC genetics indicates that this disease is genetically complex, with a significant number of key genetic alterations per tumor and very heterogeneous with substantial inter-patient variability with regards to potentially targetable genetic alterations.
Identifying the most appropriate therapies from genomics data is a major challenge in personalized cancer medicine. Thus, linking genetic alterations with therapeutic interventions is a potential useful avenue to combat PDAC.
The project’s main objective is to gain a better understanding of the impact of cancer genomics on making clinical decisions by developing new computational pipelines that priorizes drug therapies with the wide range of actionable genomic alterations for each individual patient instead of the population average data. These computational approaches can help to guide the selection of personalized treatments as well as to discover new genomic biomarkers of drug response. The strength of the approach lies in the combination of clinical research, genomics and latest in silico (computational) analyses.
Description of work performed and major findings:
To achieve project’s objectives our main research activity over last four years it has been focused on the development of two novel computational tools publicily available for the scientific community for the selection and prioritization of antitumoral drugs from the results of genome-wide analyses of cancer patients. PanDrugs (http://www.pandrugs.org/; Piñeiro-Yañez et al. 2017 submitted) identifies actionable molecular alterations based on the analysis and integration of genomic data (mutations, copy number variations or gene expression levels), with functional data (protein essentiality) and pharmacological data (sensitivity or resistance to antitumor drugs). PanDrugs prioritizes drugs by calculating gene-drug scores that combines both biological and clinical relevance. This curated target-drug resource includes single gene-drug and pathway-drug association and an extensive clasification of these drugs based on their status (approved, clinical candidates or experimental probes) and drug/target family. We have also implemented SATIE (Sequential Antitumor Treatment Inference and Enrichment, http://satie.bioinfo.cnio.es/; Tejero et al. 2017 submitted), a web tool to predict sequential treatments in cancer using transcriptional signatures of drug effect. This tool exploits the possibility of inferring a sequential treatment with a first drug that might sensitize to the immediate response for a second drug in a way that they could act synergistically. SATIE generates ranked lists of candidate drugs for sensitization, first line treatment and second line therapies, combinations of drugs or even therapeutic interventions for acquired resistance to a given drug.
These tools have been extensively used to analyze PDAC patients sequencing data in order to propose and prioritize those may have therapeutic implications. In the context of clinical research projects and clinical trials, PanDrugs has been integrated as a new module in the sequencing analysis pipeline to allow categorize patient’s tumors and match them to effective drugs or treatments. So far, we have analyzed more than 500 patients and this new pipeline has been able to identify actionable mutations in nearly half of patients. For instance, in Garralda et al. study included 25 patients integrating NGS and patient-derived xenografts (PDXs) mouse models. Using PanDrugs, we were able to identify putatively actionable tumor-specific genomic alterations in most of the cases and experimental testing of candidate treatments in PDXs models helped to select empirical treatments in patients with no actionable mutations. In 2015, we started testing this strategy in a prospective randomized clinical trial (recently funded by ERC Advanced grant, ERC-ADG-2014) in ~150 patients with standard of care resistant metastatic pancreatic cancer aiming to test the hypothesis that an integrated personalized treatment approach improves survival compared to the conventional treatment strategy. PanDrugs is being used to select second line therapies for the experimental arm patients. SATIE approach has been experimentally validated in a recently published study (Rajeshkumar et al 2017 Clin Cancer Res) using Pancreatic Ductal Adenocarcinoma (PDAC) gene expression profiles from PDXs to test whether drugs response pattern could be correlated with genomic status as well as transcriptional signatures. We evaluated the activity of some metabolic inhibitors in pancreatic cancer PDXs. Using a phenformin-induced transcriptomic signature from the LINCS database, we found a correlation between baseline expressions of the signature with sensitivity to phenformin suggesting that this drug could be revived for clinical use for treatment of pancreatic cancer.
Impact and results dissemination:
Over last four years, Bioinformatics Unit at CNIO leaded by Dr. Al-Shahrour has organized several training activities and participated in conferences related to the project (15 invited talks and 10 posters in scientific conferences). Our research outcome includes 20 peer-reviewed published articles. We participate in national and international consortiums such as: WG2 in Eupancreas (http://eupancreas.com/) and EuroPDX (http://www.europdx.eu/). Since 2014, we participate in the Pan-Cancer Analysis of Whole Genomes (PCAWG) initiative in the working group: “Translating cancer genomes to the clinic”. In 2017, CNIO Bioinformatics Unit as a new node of INB/ELIXIR-ES (https://www.elixir-europe.org/) aims to provide the framework and expertise for the systematic analysis and interpretation of cancer genomes. We have established collaborations with hospitals from Spanish National Health System that provide an initial framework to promote the application of Bioinformatics resources into the clinic. Importantly, CNIO Bioinformatics Unit participates in the European project SUDOE ONCONET along with others partners. Through its participation in the SUDO ONCONET project, CNIO Bioinformatics Unit can promote the expertise and resources through ELIXIR initiatives. The actions will lead to the production of white papers, workshops and conferences, as well as the creation of educational tools aimed at health professionals, regulatory authorities and the general public.
Through our scientific interactions and international collaborations, we expect that results of this project will facilitate the development of genomic-based clinical tests to help categorize tumors and match them to effective drugs or treatments based on their activation/deregulation profiles.
Pancreatic Ductal Adenocarcinoma (PDAC) is essentially a metastatic disease with one of the worst overall survival rates. Patients with metastatic pancreatic cancer are currently managed with systemic chemotherapy with poor results. Indeed, the median survival of these patients with more aggressive regimens in the 10-11 months range. Thus, an important need is the development of newer and more effective treatment modalities for this disease that result in an improvement in survival. It’s known that massively parallel sequencing data analysis such as WES (whole-exome sequencing) results in the identification of hundreds (sometimes thousands) of mutations. Only a small fraction of these alterations might have a clinical impact and their identification has become one of the biggest challenges to succesfully translate cancer genomic to the clinic. The current understanding of PDAC genetics indicates that this disease is genetically complex, with a significant number of key genetic alterations per tumor and very heterogeneous with substantial inter-patient variability with regards to potentially targetable genetic alterations.
Identifying the most appropriate therapies from genomics data is a major challenge in personalized cancer medicine. Thus, linking genetic alterations with therapeutic interventions is a potential useful avenue to combat PDAC.
The project’s main objective is to gain a better understanding of the impact of cancer genomics on making clinical decisions by developing new computational pipelines that priorizes drug therapies with the wide range of actionable genomic alterations for each individual patient instead of the population average data. These computational approaches can help to guide the selection of personalized treatments as well as to discover new genomic biomarkers of drug response. The strength of the approach lies in the combination of clinical research, genomics and latest in silico (computational) analyses.
Description of work performed and major findings:
To achieve project’s objectives our main research activity over last four years it has been focused on the development of two novel computational tools publicily available for the scientific community for the selection and prioritization of antitumoral drugs from the results of genome-wide analyses of cancer patients. PanDrugs (http://www.pandrugs.org/; Piñeiro-Yañez et al. 2017 submitted) identifies actionable molecular alterations based on the analysis and integration of genomic data (mutations, copy number variations or gene expression levels), with functional data (protein essentiality) and pharmacological data (sensitivity or resistance to antitumor drugs). PanDrugs prioritizes drugs by calculating gene-drug scores that combines both biological and clinical relevance. This curated target-drug resource includes single gene-drug and pathway-drug association and an extensive clasification of these drugs based on their status (approved, clinical candidates or experimental probes) and drug/target family. We have also implemented SATIE (Sequential Antitumor Treatment Inference and Enrichment, http://satie.bioinfo.cnio.es/; Tejero et al. 2017 submitted), a web tool to predict sequential treatments in cancer using transcriptional signatures of drug effect. This tool exploits the possibility of inferring a sequential treatment with a first drug that might sensitize to the immediate response for a second drug in a way that they could act synergistically. SATIE generates ranked lists of candidate drugs for sensitization, first line treatment and second line therapies, combinations of drugs or even therapeutic interventions for acquired resistance to a given drug.
These tools have been extensively used to analyze PDAC patients sequencing data in order to propose and prioritize those may have therapeutic implications. In the context of clinical research projects and clinical trials, PanDrugs has been integrated as a new module in the sequencing analysis pipeline to allow categorize patient’s tumors and match them to effective drugs or treatments. So far, we have analyzed more than 500 patients and this new pipeline has been able to identify actionable mutations in nearly half of patients. For instance, in Garralda et al. study included 25 patients integrating NGS and patient-derived xenografts (PDXs) mouse models. Using PanDrugs, we were able to identify putatively actionable tumor-specific genomic alterations in most of the cases and experimental testing of candidate treatments in PDXs models helped to select empirical treatments in patients with no actionable mutations. In 2015, we started testing this strategy in a prospective randomized clinical trial (recently funded by ERC Advanced grant, ERC-ADG-2014) in ~150 patients with standard of care resistant metastatic pancreatic cancer aiming to test the hypothesis that an integrated personalized treatment approach improves survival compared to the conventional treatment strategy. PanDrugs is being used to select second line therapies for the experimental arm patients. SATIE approach has been experimentally validated in a recently published study (Rajeshkumar et al 2017 Clin Cancer Res) using Pancreatic Ductal Adenocarcinoma (PDAC) gene expression profiles from PDXs to test whether drugs response pattern could be correlated with genomic status as well as transcriptional signatures. We evaluated the activity of some metabolic inhibitors in pancreatic cancer PDXs. Using a phenformin-induced transcriptomic signature from the LINCS database, we found a correlation between baseline expressions of the signature with sensitivity to phenformin suggesting that this drug could be revived for clinical use for treatment of pancreatic cancer.
Impact and results dissemination:
Over last four years, Bioinformatics Unit at CNIO leaded by Dr. Al-Shahrour has organized several training activities and participated in conferences related to the project (15 invited talks and 10 posters in scientific conferences). Our research outcome includes 20 peer-reviewed published articles. We participate in national and international consortiums such as: WG2 in Eupancreas (http://eupancreas.com/) and EuroPDX (http://www.europdx.eu/). Since 2014, we participate in the Pan-Cancer Analysis of Whole Genomes (PCAWG) initiative in the working group: “Translating cancer genomes to the clinic”. In 2017, CNIO Bioinformatics Unit as a new node of INB/ELIXIR-ES (https://www.elixir-europe.org/) aims to provide the framework and expertise for the systematic analysis and interpretation of cancer genomes. We have established collaborations with hospitals from Spanish National Health System that provide an initial framework to promote the application of Bioinformatics resources into the clinic. Importantly, CNIO Bioinformatics Unit participates in the European project SUDOE ONCONET along with others partners. Through its participation in the SUDO ONCONET project, CNIO Bioinformatics Unit can promote the expertise and resources through ELIXIR initiatives. The actions will lead to the production of white papers, workshops and conferences, as well as the creation of educational tools aimed at health professionals, regulatory authorities and the general public.
Through our scientific interactions and international collaborations, we expect that results of this project will facilitate the development of genomic-based clinical tests to help categorize tumors and match them to effective drugs or treatments based on their activation/deregulation profiles.