Identification of the metabolic reprogramming underlying metastasis: use and development of PheNoMenal e-infraestructures for
analyzing metabolic phenotype data
Marta CascanteUniversity of Barcelona
EASYM, October 2016
Sustainingproliferative
signaling
Evading GrowthSupressors
Avoiding Inmune Destruction
EnablingReplicativeInmortality
Tumor-promoting
Inflammation
Invasion and metastasis
InducingAngiogenesis
Genomeinstability
Resisting Cell Death
Metabolicreprogramming
CANCER HALLMARKS: METABOLISM
Metabolic reprogramming underlyingmetastatic potential ?
Metastasis: the
leading cause of
death by cancer.
Cancer cells are perfect systems to invade and parasite other tissues
Robust metabolic profile
FRAGILITY
Exploitable Target against tumor progression
and metastasis?
unexpected perturbations
CANCER AS A METABOLIC ALTERATION
Tumor metabolism robustness counteracts single hits
Multiple hit strategies can avoid bypass of single inhibitions
Tumor metabolism response to multiple hits is unpredictable
Rational design of new therapeuticalcombinations is necessary
Highest capacity to predictphenotype:
Understanding metabolic reprogramming in metastatic cancer will permit to discover new drug targets
Metabolites are not only the “end point” also the “driving force”
Fluxomics (analyze the metabolic flux distribution in the cell metabolic network)
EXPERIMENTAL APPROACHES
Integration of data in computer models
Cytomics
Genomics
Proteomics
Metabolomics
“IN SILICO” SIMULATOR
• Analyze metabolic adaptations supporting metastatic potential
• Design metabolic Interventions in drug development
METABOLIC FLUX REPROGRAMMING FINGERPRINT:
Tracer based metabolic flux models and GSMM: two versatile, potent and informative tools in the study of the metabolic
fingerprint of cancer cells
13C-basedMETABOLIC FLUX
MODELS
GENOME-SCALE METABOLIC MODELS
(GSMM)
Fluxomics in a systems medicine approach
METABOLIC FLUX REPROGRAMMING FINGERPRINT:
Tracer based metabolic flux models and GSMM: two versatile, potent and informative tools in the study of the metabolic
fingerprint of cancer cells
13C-basedMETABOLIC FLUX
MODELS
GENOME-SCALE METABOLIC MODELS
(GSMM)
Fluxomics in a systems medicine approach
13C TRACER-BASED FLUXOMICS WORKFLOW
Incubation withtracer substrates
-[1,2-13C2]-glucose-[3-13C]-glutamine- [U-13C]-glutamine-etc…
Study of isotopomericdistribution
-Glucose-Lactate-Ribose-Fatty Acids-Krebs Cycleintermediates-Aminoacids…
Introduction of additional constraints
-Enzymatic activities- Metabolites uptake/intake
- Glucose- Glutamine- Lactate- Glutamate
- Targeted metabolomics-(e.g. BIOCRATES kits)
Generation of fluxomic model
Cell culture LC/MS, GC/MSLC/MS,
SpectrophotometryComputational
Metabolization of 13C-labelled substrates inside the cells
Incorporation of 13C in intracellular and excreted metabolites
Fluxome: “Total set of fluxes in the metabolicnetwork of a cell.”
…Qualitative: from MIDA (Mass Isotopomer
Distribution Analyisis)
…Quantitative: from software packages (i.e. ISODYN in-house)
13C-basedMETABOLIC FLUX
MODELS
GENOME-SCALE METABOLIC MODELS
(GSMM)
Tracer based metabolic flux models and GSMM: two versatile, potent and informative tools in the study of the metabolic
fingerprint of cancer cells
METABOLIC FLUX REPROGRAMMING FINGERPRINT:
Fluxomics in a systems medicine approach
Feasible solution space(non-condition specific)
Condition specific solution space
Unconstrained solution space
• Stoichiometric constraints• Thermodynamics constraints
GSMMs AND FLUX BALANCE ANALYSIS (FBA)
• Transcriptomic profile• Proteomic and Metabolic profile
Optimal Solution
• Calculate optimal steady-state flux
distributions in metastatic and non-
metastatic cancer cell
subpopulations
GSMMs of human metabolism flux map distribution: constraint based approach:
Challenge: Different tumor subpopulations will have different
vulnerabilities?
• Mathematical representation of the metabolic reaction encoded by an organism’sgenome.
WE KNOW THAT tumor metastatic potential:
● Resides in a minority of malignant cells (tumor heterogeneity), known as
tumor initiating cells (TICs) or cancer stem cells (CSCs).
● It uses at several critical steps the epithelial-mesenchymal transition (EMT).
WE AIM TO:
● Characterize TICs/CSCs and EMT metabolic patterns
● To unveil metabolic vulnerabilities of tumor invasion and metastasis.
IDENTIFICATION OF THE METABOLIC REPROGRAMMING UNDERLYING PROSTATE CANCER METASTASIS:
A dual cell model to study prostate cancer metastasis:
Two related cell subpopulations isolated from the PC-3 prostate
cancer cell line according to their metastatic and invasive potential:
PC-3 cell line
cells isolated from nude mice liver metastasesafter intrasplenic injection of PC-3 cells.
PC-3/M cells
Selection by limiting dilution: usingthe Matrigel invasiveness assay
PC-3/S cells
EMT
Celia-Terrassa T, et al. (2012) . J Clin Invest 122(5):1849-1868 .
-Can be transformed into each other (factors, ↓ or ↑ genes, etc.)
PC-3/M
-Cooperation to establish metastasis
Celia-Terrassa T, et al.(2012) . J Clin Invest122(5):1849-1868 .
Metabolic characterization to find metabolic targets in order to:
• Kill these cells• Promote the transformation of these distinct phenotypes
Introduction to the PC-3/M and PC-3/S cell model
1
5
9
13
0 24 48 72
Re
lati
ve c
ell
nu
mb
er (
Ntf
/ N
t=0
h)
Hours
Growth curve
PC-3/S
PC-3/M ↑ metastatic potential↓ invasive
TICs/CSCs (stem cellmarkers)
PC-3/S ↓ metastatic potential
↑ invasiveEMT
Here we propose an strategy to characterize the metabolism
of these cells in order to find specific targets, combining and
integrating experimental and computational methods:
• Detailed metabolic characterization of both PC-3/M and PC-3/S cell lines
• Novel genome scale metabolic network reconstructionanalysis that integrates the transcriptomic and metabolicanalyses on the two clonal cell lines.
Combining Experimental and Computational approaches
Here we propose an strategy to characterize the
metabolism of these cells in order to find specific targets,
combining and integrating experimental and
computational methods:
● Detailed metabolic characterization of both PC-3/M and
PC-3/S cell lines.
● Novel genome scale metabolic network reconstruction
analysis that integrates transcriptomic and metabolic
analyses on the two clonal cell lines.
Combining Experimental and Computational approaches
Aguilar, E Marín de Mas, I, Zodda E, Marin S., Morrish F; Selivanov V; Meca-Cortés O; Delowar H; Pons M; Izquierdo I., Celia-Terrassa T; de Atauri P; Centelles JJ; Hockenbey D; Thomson TM; Cascante M, Stem Cells, 2016 May;34(5):1163-76. doi: 10.1002/stem.2286.
PC-3M have enhanced glycolytic metabolism and fuel TCA cycle using glutamine
***
0,0
0,2
0,4
0,6
0,8
1,0
PC-3M PC-3S
mm
ol/h
· 1
06
cells
Glucose consumption
***
0,0
1,0
2,0
3,0
PC-3M PC-3S
mm
ol/h
· 1
06
cells
Lactate production
**
0,00
0,05
0,10
0,15
PC-3M PC-3S
mm
ol/h
· 1
06
cells
Glutamine consumption
***
*****
*** **
0,0
0,1
0,2
0,3
0,4
0,5
SIG
m
Label enrichment in TCA intermediates
**
***
0,0
0,1
0,2
0,3
0,4
0,5
0,6
Pyruvate Lactate Alanine
SIG
m
Label enrichment
PC-3/M
PC-3/S
100% 1,2-13C2-glucose
***
***
* ** **
0,0
0,2
0,3
0,5
0,6
0,8
0,9
SIG
m
Label enrichment in TCA intermediates
100% U-13C5-glutamine
- Glucose contributes to a greater extent to the labellingof several final glycolytic products in PC-3M
-Glutamine contributes to a greater extent tothe labeling of the TCA intermediates in PC-3M
Aguilar, E Marín de Mas, I, Zodda E, Marin S., Morrish F; Selivanov V; Meca-Cortés O; DelowarH; Pons M; Izquierdo I., Celia-Terrassa T; de Atauri P; Centelles JJ; Hockenbey D; Thomson TM; Cascante M, Stem Cells, 2016 May;34(5):1163-76. doi: 10.1002/stem.2286.
ISODYN MODEL PREDICTS:
• PC3-M (e-CSC) cells display enhanced glycolysis and fuel TCA cycle using glutamine
• increased TCA cycle relative to glucose uptake and mitochondrial respiration fluxes in PC-3S cells
• Increased one-carbon metabolism in PC3-M cells
PC-3S (stable EMT and non-eCSC) have enhanced TCA cycle and mitochondrial respiration
Integration of 13C-based metabolomics data in a flux model generated usingISODYN software package to estimate quantitative flux maps in PC-3S and PC-3M cells:
Aguilar, E Marín de Mas, I, Zodda E, Marin S., Morrish F; Selivanov V; Meca-Cortés O; Delowar H; Pons M; Izquierdo I., Celia-Terrassa T; de Atauri P; Centelles JJ; Hockenbey D; Thomson TM; Cascante M, Stem Cells, 2016 May;34(5):1163-76. doi: 10.1002/stem.2286.
Metastatic metabolic gene signature
• Metastatic metabolic gene signature (MMGS) : (genes over-expressed in PC-3/M compared with PC-3/S)
• MMGS is consistent with metabolic differences observed between PC-3/M and PC-3/S:
• Correlation between MMGS and prostate cancer progression by applying GSEA:
We found a significant correlation between MMGS and tumor progression in prostate cancerand in other 11 cancer types.
Aguilar, E Marín de Mas, I, Zodda E, Marin S., Morrish F; Selivanov V; Meca-Cortés O; Delowar H; Pons M; Izquierdo I., Celia-Terrassa T; de Atauri P; Centelles JJ; Hockenbey D; Thomson TM; Cascante M, Stem Cells, 2016 May;34(5):1163-76. doi: 10.1002/stem.2286.
Here we propose an strategy to characterize the metabolism ofthese cells in order to find specific targets, combining andintegrating experimental and computational methods
● Detailed metabolic characterization of both PC-3/M and PC-
3/S cell lines
● Novel genome scale metabolic network reconstruction
analysis that integrates the transcriptomic and metabolic
analyses on the two clonal cell lines. (work in progress)
Combining Experimental and Computational approaches
Computational approach based on CBM:
● Uses transcriptomic data of PC-3/M and PC-3/S to infer the activity
state of their metabolic networks
Cancer-specificMetabolic model
Infer the activity stateof metabolic networks
Non-specific genome-scale
metabolic network reconstruction.
transcriptomicdata of PC-3/M & PC-3/S
Constraint-based method on genome-scale models of human metabolism (GSMMs)
This approach uses an objective function maximizes the similarity
between gene expression and reaction activity state:
● Maximizes the number
of active reactions
associated to
highly expressed genes
● Minimizes the number
of active reactions
associated to
lowly expressed genes
This analysis was performed by using the software fasimu
Shlomi et al; Nat Biotechnol. 2008 Sep;26(9):1003-10
Constraint-based method on genome-scale models of human metabolism (GSMMs)
• Use and development of e-infraestructures for analyzing metabolic phenotype data:
Making existing Metabolomics and Fluxomics toolsavailable to serve Systems Medicine:
• PhenoMeNal will support the data processing and analysis pipelines for molecular phenotype data generated by metabolomics applications.
• PhenoMeNal will provide services enabling computation and analysis to improve the understanding of the causes and mechanisms underlying health and diseases.
(Phenome and Metabolome aNalysis)
This project has received funding from theEuropean Union’s Horizon 2020 research andInnovation programme under grant agreementNo 65424121Consortium Coordinator: Christoph Steinbeck
GC/MS(raw) Isotopologue
distributions
Additional Inputs:-SBML model-Flux
boundaries
Etc..
Patient 4
Midcor
Iso2FluxPatient specific
flux distribution
GC/MS(raw) Isotopologue
distributions
Additional Inputs:-SBML model-Flux
boundaries
Etc..
Patient 3
Midcor
Iso2FluxPatient specific
flux distribution
GC/MS(raw) Isotopologue
distributions
Additional Inputs:-SBML model-Flux
boundaries
Etc..
Patient 2
Midcor
Iso2FluxPatient specific
flux distribution
ISODYN
GC/MSLC/MS
(raw)Metabolights
Isotopologuedistributions
Additional Inputs:
-SBML model-Flux
boundariesEtc..
Patient 1
Midcor
Iso2Flux
Patient specific
flux distribution
Automatic workflows for 13C- fluxomics in the framework of PhenoMeNal in development
Making existing Metabolomics and Fluxomics toolsavailable to serve Systems Medicine:
EC- H2020
Automatic Data Curation
CONCLUSIONS
• Metabolic programs underlying Cancer stem cell (CSC) and epithelial-mesenchymal transition (EMT) programs are mutually uncoupled and associated with preferred metabolic dependencies required for optimal cell proliferation and sustaining specific cell phenotypes.
• The complementary 13C-based fluxomics and genome scale metabolic models applied here have identified tumor subpopulation specific metabolic targets with potential impact on the design of new treatments for metastatic cancer.
• 13C-based fluxomics tools and GSMMs in conjunction with simulation techniques such as FBA need to be integrated.
• Automatic workflows for 13C-fluxomics can be developed to serve systems medicine (PhenoMeNal e-infrastructure in development ).
Spanish Systems Medicine Network
Objectives
Update the pathophysiology of the massive
data provided by omics and quantitative
biology and processing using
computational tools
Increase Spanish science in the medicine
field by enhancing cooperation and
synergy in biomedical research and clinical
field industry at national and international
level
Groups
University of Barcelona
PR: Marta Cascante
Institute of Biomedical Research
(IIBM-CSIC)
PR: Lisardo Boscá Gomar
Pompeu Fabra University
PR: Pura Muñoz Canoves
Institute of Biomedicine
of Valencia (IBV-CSIC)
PR: Marta Casado Pinna
Institute of Biology and
Molecular Genetics (CSIC)
PR: Mariano Sanchez
Crespo
Institute of Parasitology and
Biomedicine López Crespo (CSIC)
PR: Dolores González Pacanowsca
Institute of Biomedical
Research (IIBM-CSIC)
PR: Paloma Martin-Sanz
University of Navarra
PR: Fernando J. Corrales
Institute Consortium of
Biomedical Research August Pi
and Sunyer (IDIBAPS)
PR: Josep Roca Torrent
Contact:Network Coordinator: Marta CascateEmail: [email protected] Assistant: Pablo M. BlancoEmail: [email protected]
Acknowledgements
THANK YOU FOR YOUR ATTENTION
Fondos Feder “Una manera de hacer Europa”
Group of Integrative Systems Biology, Metabolomics and Cancer
Department of Biochemistry and Molecular Biomedicine, University of Barcelona
Contact: Marta Cascante, [email protected]
Collaborators:
• Team of Dr Timothy Thomson, IBMB-CSIC, Barcelona
• Team of Dr Balazs Papp BiologicalResearch Center Szeged, Hungary
• Team of Dr David Hockenbery and DrFionnuala Morrish, Fred Hutchinson Cancer Research Center, Seattle
EC- H2020
EC-H2020
Pedro de Atauri
Vitaly Selivanov
Carles Foguet
Esther Aguilar
Erika Zodda
Josep Centelles
Silvia Marín
Igor Marín de Mas