Skip to main content
European Commission logo print header

Characterization and modelling of dietary effects mediated by gut microbiota on lipid metabolism

Final Report Summary - ETHERPATHS (Characterization and modelling of dietary effects mediated by gut microbiota on lipid metabolism)

Executive Summary:

The general objective of ETHERPATHS was to build a suite of tools that will afford studies of foods modulating lipid metabolism, from systemic to cellular levels. The theme of the project was the control of systemic lipid homeostasis and the role that nutrition may play in its maintenance, improvement or dysregulation. In particular, we studied the relevance of plasmalogens, which are enriched in n-3 fatty acids and are already known to play an important role in human health as a factor involved in aging, obesity, diabetes, and diseases of the central nervous system.

As part of the ETHERPATHS platform for nutritional systems biology, in vitro studies were supporting the studies in vivo as well as the clinical intervention study. For example, in vitro colon model was applied to ten substrates used in the diets of human intervention conducted in the project, and predicted the compliance markers from polyphenol-rich diet. These markers were later validated in the nutritional intervention study. The metabolite profiles of samples obtained from the colon model were analysed. The purified faecal water extracts were prepared from these studies and applied in hepatocyte and adipocyte interventions. The in silico strategies to complement the in vitro studies included (1) modifications of the IsoDyn model, a model of central carbon metabolism, and (2) development of a model of palmitate conversion into VLDL in hepatocytes.

The main focus of in vivo studies was on generation and characterization of the plasmalogen deficient mice (Gnpat KO mouse). A model of Gnpat depletion and phenotyping showed that Gnpat depletion has an impact in the overall energy homeostasis. Studies were also conducted to investigate how the gut microbiota affects host lipid metabolism by comparing serum, white adipose tissue, and liver lipidomes in germ free and conventionally raised mice under different diets.

The four diets were developed differing for n-3 fatty acid and polyphenol content, which were used in the ETHERPATHS nutritional intervention study in humans: (A) Control diet low in n-3 fatty acids and polyphenols; (B) Diet rich in n-3 fatty acids and low in polyphenols; (C) Diet rich in polyphenols and low in n-3 fatty acids; (D) diet rich in n-3 fatty acids and polyphenols. The schedule for type, time and modalities of sampling, storing and shipping for the blood, urine, faeces and adipose tissue samples to be analysed at the end of the intervention periods has been defined. Dietary compliance was optimal. The diets were found to have distinct effects on the lipid profiles as well as on gut microbial composition.

New computational and analytical tools were developed to support the nutritional systems biology platform. Metabolomics data processing software MZmine 2 was released and optimized to detect stable isotope tracer signals. A method for use of heavy water as tracer for de novo lipogenesis has been implemented. The protocol for new kinetic studies has been developed. New efficient methodologies for the characterization of lipid and small polar metabolite profiles in biological samples were set up. For the metabolite identification, mass spectra libraries utilising the accurate mass and high resolution mass spectrometry were constructed for different sample matrices, including serum and tissue. New data analysis methods for the treatment of the metabolomics data were developed and released.

The Web platform was set up which integrates tools for experimental data management and a Systems Biology Markup Language (SBML) model designer and manager. Also, the Meta-Model of lipid spillover in whole organism was developed and integrated with the Web platform. Kinetic model of fatty acid metabolism in hepatocyte was added to Meta-Model. Its adequacy was demonstrated through virtual experiments as well as by generating and validating the predictions based on experimental data generated in the project.

During the last period of the project, also the dissemination and exploitation activities have been intensifying. The ETHERPATHS platform has been presented at multiple conferences (e.g. special ETHERPATHS session at NuGOweek 2012). Also a book has been released ‘A Systems Biology Approach to Study Metabolic Syndrome’ (Springer Verlag, ISBN 978-3-319-01007-6, eds. Matej Orešič and Antonio Vidal-Puig), which includes among others 10 contributions from the ETHERPATHS participants.

Project Context and Objectives:

The collaborative project ETHERPATHS specifically addressed the call FP7 Cooperation Work Programme: KBBE-2007-2-2-08: Systems Biology and bioanalytical tools for nutrition research.

Understanding of living organisms in the context of coordinated gene and molecular function and translation of this knowledge to improve human health is a great challenge and certainly one of the central aims of medical systems biology. The interface between biological systems and nutritional factors represents the next level of complexity involving interactions between nutrients, host metabolism and gut microbiota that still have to be understood. Nutritional systems biology thus requires bridging across multiple levels and concepts, e.g. cellular–organismal, host–microbial, short-term–long-term effects. To face such a complex task, it is essential that the systems biology tools are built by combining multiple experimental and theoretical approaches, based on concrete problems and clearly defined questions.

The theme of this collaborative project was the control of systemic lipid homeostasis and the role that nutrition may play in its maintenance, improvement or dysregulation. In particular, we will study the relevance of plasmalogens, which are enriched in n-3 fatty acids and are already known to play an important role in human health as a factor involved in aging, obesity, diabetes, and diseases of the central nervous system. The concept is that the study of the regulation of systemic lipid homeostasis in the context of nutrition is an excellent case study, with the sort of clearly definable questions needed to build the tools for nutritional systems biology.

The overall objective of the project was to build a suite of tools that will afford studies of foods modulating lipid metabolism, from cellular to systemic levels.

The expected impact is the development of new informatics and analytical tools to bridge data across in vitro, in vivo, and clinical levels by using cellular, tissue level, and physiological models; and to gain new knowledge on how various dietary components, such as n-3 fatty acids and dietary polyphenols, as well as gut microbial composition modulate lipid homeostasis beneficially for human health.

The main objective of ETHERPATHS was met by combinations of wet-lab and dry-lab strategies at in vitro, in vivo, and clinical levels, divided into six specific objectives:

O1. Provide in vitro and in silico platforms for studies of gut microbiota in the context of nutritional interventions.

ETHERPATHS will focus on linking metabolism of fatty acid supplements with those containing polyphenols, which are converted by gut microbiota and circulated in entero-hepatic-circulation and plasma.

O2. Generate and characterize a plasmalogen-deficient mouse model in the context of nutritional interventions, specific lipid metabolism modulation, and gut microbiota variation.

ETHERPATHS will identify the mechanistic links between nutritional interventions involving fatty acid supplements and polyphenols, and systemic plasmalogen status. New pathway reconstruction tools will be developed for that purpose, which will complement and link to the modelling tools and results from in vitro studies.

O3. Provide bioanalytical and dynamic modelling tools for studies of systemic lipid metabolism in the context of nutritional interventions.

ETHERPATHS will use samples and data from an ongoing clinical trial, as well as conduct a new trial involving specific nutritional interventions. New physiological models of lipid metabolism will be developed based on the data, which will include modelling of lipids in very low–density lipoproteins (VLDL) as well as in chylomicron particles, thus establishing a link with the intestinal lipid metabolism.

O4. Provide software and modelling tools for integration of computational models and data across multiple organism levels.

ETHARPATHS will combine modelling tools used to reach objectives 1-3, in a single Systems Biology Markup Language (SBML)-compliant software platform. Using advanced and emerging integrative bioinformatics strategies such as conceptual spaces, the data and model entities will be linked to established knowledge repositories and databases. New theoretical models will be developed aiming to investigate control and dependencies between different levels, e.g. gut microbiota and its metabolites, tissues, and systemic lipid metabolism.

O5. Design and perform nutritional interventions, including food components that are metabolized in the colon, to define their specific alterations in lipid homeostasis.

Evaluate the effects of diets rich in -3 fatty acids and polyphenols on lipid metabolism. Investigate possible mechanisms of their effects.

O6. Develop a high-throughput mass spectrometry-based platform for sensitive profiling of lipids.

ETHERPATHS will build on most recent developments in nano-electrospray ionization (nanoESI) microfluidistic chip and electrocapture technologies to develop highly sensitive and quantitative platforms for global profiling of lipids and hydrophilic metabolites.

Project Results:

The research progressed in all six workpackages (WPs, shown in Figure 1 below), corresponding to six specific objectives of the project.

WP1

In WP1, during the first period (months 1-18), in vitro colon model was applied to eight substrates used in the diets of ETHERPATHS human intervention study, which was performed in Naples. Purification techniques to isolate faecal microbiota from faecal water were tested and hydrophobic PTFE membrane was selected. Metabolite profiles were analysed. Statistical analysis methods for quantitative targeted analysis and semi-quantitative metabolomics were established and applied to all data and to red wine, respectively. In the 2nd period (months 19-36), semi-quantitatve metabolomics was applied to those selected foods, which showed formation of known microbial metabolites of phenolic compounds, namely green tea, orange, rocket salad. Furthermore during the 2nd period additional substrates (spinach, artichoke and coffee) were subjected to the colon model for identification of their food-related intestinal metabolome. Exctracts fro cell assays were prepared, including converted green tea (with microbiota), non-converted green tea (in buffer, no microbiota) and microbial control (faecal microbiota, no green tea) after oincubation of 6 hours. Experimental data from VTT (production of polyphenols and other metabolites by colonic microbiota) were processed at ISBSPb and kinetics parameters for meta-model were estimated. Using developed version of meta-model the preliminary predictions of blood dynamics of different metabolites were made (Figure 2).

During the 3rd and last reporting period (months 37-54), the prediction of average human urinary excretion of quantitative metabolite profiles was calculated from the colon model data on the basis of metabolite profiles at 6 hour time point, the fresh weights and portions and frequency of the intake of polyphenol rich foods. The data showed the same significant metabolites as the most ubundant hydroxylated phenylpropionic, -acetic and benzoic acids, indicated as major metabolites in the colon model. The compliance was also indicated by 7-day food records and the results were in agreement. Further conversion of microbial metabolites from green tea in hepatocytes also confirmed the role of liver in the formation of the metabolic pool and filled the gap between colon model and human urinary excretion of polyphenol metabolites.

During the 1st period, the experiments related to hepatocytes and adipocytes were initiated by protocol modifications for hepatocytes and adipocytes. Cultivation of hepatocytes has started and the first experiments with pre-adipocytes and adipocytes have been performed to design the future incubation conditions. Purified faecal water extracts were prepared from selected colon model samples and corresponding controls for testing their purity in the hepatocyte interventions. Purification of microbial extract was adjusted. The generation of an in vitro model of plasmalogen depletion was conducted to investigate the effects of plasmalogens and dietary components in adipose tissue using loss of function strategies, i.e. establishment of Gnpat knockdown 3T3L1 white preadipocyte cell lines.

During the 2nd period we found that benzoic acid derivatives were the main metabolites from incubation of hepatocytes with colonic phenylpropionic acid-rich extracts and that it is feasible to study oleic acid intervention to show lipoprotein secretion from human primary hepatocytes. In the 3rd reporting period the hepatic metabolism of green tea and its microbial metabolites were confirmed using human primary hepatocytes. 3,4-Dihydroxypropionic acid, 3-hydroxybenzoic acid, ferulic acid and hippuric acid were confirmed the major hepatic metabolites from converted green tea. We also found that McA-RH777 hepatocyte cells produced glucose, lactate and glutamine without an apparent consumption of amino acids. It would suggest that this cell line has another source of carbons that is able to contribute to the central carbon metabolism. The isotopologue distributions obtained after the 8 hours of incubation with the tracer suggest that these cells use glucose from the medium to recycle glycogen while they produce new glucose molecules from other carbon sources. However, other metabolites such as lactate, amino acids (Figure 3) and TCA cycle intermediates seem to be mainly produced by other carbon source/s.

The three tested colonic extracts have different effects in the McA-RH7777 central carbon metabolism. Despite the lack of modelling of the data, a different distribution in some of the central carbon metabolic pathways is expected, especially in those involving pentose phosphate pathway and TCA cycle. The highest metabolic effects were induced by the colonic extracts containing fermented and non-fermented green tea polyphenols. Moreover, their effects in McA-RH7777 cells were different, confirming that the fermentation is important for the metabolic effects.

During the 2nd period, we found:

1. PUFA intervention. The metabolic disturbances associated to the lack of plasmalogens (eg. lipid accumulation, markers of adipogenesis and lipogenesis) are directly related to the depletion of that specific subset of lipids or some products of their metabolization. No dietary or pharmacological interventions other than metabolites aimed to rescue plasmalogens levels “per se” totally recovered the phenotype but only improved some partial features. Globally considered our data indicate that plasmalogens (ether lipids) play a unique role in the differentiation programme of the adipocytes.

2. Effects of extracts derived from green tea pre- and post-colonic fermentation on cell cultured adipocytes. Globally considered, our results suggest that all three green tea experimental groups (Green tea with microbiota, Green tea in buffer and Faecal control) tested impacted at different degrees the metabolism of the 3t3l1 line after acute exposition. Indeed, increased lipogenic capacity as well as increased catabolism of lipids (lipolysis and fatty acid oxidation programme) and possibly, increase in glucose uptake is a common hallmark in response to these treatments.

During the 3rd period, we implemented the data regarding the metabolic impact of non-converted or converted green tea of the microbial control (no green tea) in adipocytes. These results demonstrated that those compounds affect glycolysis, lipolysis and mitochondrial respiration as well as insulin signaling cascade. The identification of the single molecules or combination of molecules from the different extracts responsible for such effects would be essential to explore and exploit any potential therapeutic use of those natural compounds in future. The other approach is that different extracts from the microbial conversion of different foods would be tested using varying time lengths and concentrations. The most important difference is to detect whether the pre- or post-colonic extracts are active and also take into account the hepatic metabolism of the non-converted food extracts.

During the 2nd period preliminary results showed little disturbance of the main adipocyte metabolism, suggesting a subtle effect of the colonic extracts studied at the concentrations used. During the last reporting period, converted Green tea extracts caused a state in the 3T3-L1 adipocytes, in which the cells required less reductive power. The changes were subtle, which also indicated that the adipocytes were not affected harmfully.

During the 1st period, the IsoDyn model for metabolic modelling based on 13C tracer data was also modified. The draft model of palmitate conversion into VLDL in hepatocyte was constructed as a link between central carbon metabolism model and the model of systemic lipid metabolism. Draft version of the model was updated to describe three fatty acids (palmitate, stearate, alpha-linolenic acid) conversion into VLDL particle.

During the 2nd period, methodology was implemented in IsoDyn software which can reveal compartmental structure and metabolic flux distribution from the distribution of 13C isotopomers measured in the products of cells incubated with 13C labeled substrates. The application of this methodology to the analysis of 13C isotopomer distributions measured in metabolites of isolated liver cells revealed a separate compartment of hexose phosphates related with substrate channeling in glycogen metabolism. This analysis provided the distribution of metabolic fluxes in central carbohydrate metabolism of hepatocytes incubated with 13C labeled glucose, and revealed the changes of fluxes in hepatocytes that were induced by addition of other sources of carbon in the incubation media.

A metabolic model for the analysis of flux distribution in adipocytes has been constructed. Simulation of 13C isotopomer distribution with Isodyn revealed greatly increased fluxes from Krebs cycle and biosynthetic pathways in adipocytes compared to their undifferentiated precursors (Figure 4).

All these analyses indicate that the tested polyphenol- or their microbial metabolite-enriched extracts do not affect hepatocyte metabolism, while the converted and non-converted polyphenols of green tea affect glycogen and glycolysis futile recycling in adipocytes.

During the 1st period, NorayBio started to prepare the design the DDBB (database) and the software structure during the 1st period and data of the first food related metabolite profiles were sent to NorayBio to facilitate the construction of the database. ISBSPb obtained experimental data from VTT (production of polyphenols and other metabolites by colonic microbiota) and the data were processed and kinetics parameters for meta-model were estimated. Using developed version of meta-model the preliminary predictions of blood dynamics of different metabolites were made.

During the second period NorayBio has coordinated the data collection among WP1 partners that was then used to feed the Meta-Model in WP4 (metabolite profiles, flux analysis data). Individual models were requested too to build the Meta-model. The requirements for the kind of models and data needed were studied and this information were sent to the partners.

Data files and format examples were used to test the performance of the Web platform.

During the 3rd reporting period NorayBio and ISBSPB applied their software and modelling tools to model the effects of microbial metabolome on hepatocytes and adipocytes optimized and validated them with the experimental data obtained in the project.

WP2

A pilot nutritional intervention was performed in ob/ob mouse models that suggests alkylglycerol (precursor of plasmalogens) treatment has a beneficial impact on serum triglyceride, free fatty acid and ketone body levels. This is accompanied by mild increases in genes involved in lipid oxidation in highly oxidative tissues such as heart and liver. In addition, lipidomic analysis also revealed remodelling in the lipid profile of key organs, including an increase in ether-lipid species. This supports the use of alkylglycerols as strategy to replenish ether-lipid pool.

University of Cambridge (UCAM) team also obtained a model of Gnpat (key gene in ether lipid synthesis) depletion (a hypomorphic) and started the metabolic phenotyping of the mice fed chow diet. Our data up to date show Gnpat hypomorphic mice are more glucose tolerant than the wildtype counterparts. Blood biochemistry revealed a substantial increase in ketone bodies. This latter result is in line with the gene profiling in liver increased expression of genes involved in fatty acid oxidation and ketone bodies production. Interestingly, skeletal muscle and brown adipose tissue also showed and increase in fatty acid uptake and oxidation as well as increase in markers related to mitochondrial biogenesis.

VTT performed lipidomic analysis of different tissues of the gnpat C57BL6 ko mouse vs. wt littermates and confirmed the depletion of etherlipids was globally exacerbated in comparison to the mix background 129Ola/C57BL6. A new pathway reconstruction methodology was developed for understanding cellular lipid metabolism that can be applied to the in vivo profiling. Methodology was demonstrated in lipidomics data from liver tissue in plasmalogen-depleted mouse models but the methodology is flexible for any tissue data.

All in all, the data in this WP suggest that Gnpat depletion has an impact in the overall energy homeostasis.

University of Gothenburg team (UGOT; Fredrik Bäckhed) has investigated how the gut microbiota affects host lipid metabolism by comparing serum, white adipose tissue, and liver lipidomes in CONV-R and GF mice. They demonstrated that the microbiota modified a number of lipid classes in the serum, adipose tissue and liver, with its most dramatic effect on triglyceride and phosphatidylcholine species (Figure 5).

WP3

UGOT (Jan Boren) has developed a new protocol for kinetic studies (Figure 6) and the outline of the model has been set.

In the human kinetic studies, main core and surface lipids of VLDL1 and VLDL2 lipoproteins showed expected results with core lipids (TG and DAG) being 8-fold higher in VLDL1 than VLDL2 and surface lipids being about 2-fold higher. In contrast the surface lipid PE was markedly enriched (i.e more than two fold) in VLDL1. This may reflect a different site of maturation for VLDL1 than for VLDL2. This is in line with our results showing differential regulation of VLDL1 and VLDL2 secretion both in vitro and in vivo

The nutritional intervention showed altered levels in glycerophospholipids, mainly PC and PE, due to the high n-3 fatty acid diets (with or without a high content in polyphenols), while exclusive polyphenol modification caused only minor changes in the HDL fraction profile.

In the mouse lipoproteins analysis, lipidomic profiling of plasma HDL fractions showed that the depletion of ether phospholipids in Gnpat KO mice could be modulated by diet. Ether-lipids could partly be replenished with dietary n-3 fatty acids.

NorayBio has coordinated the data collection among WP3 partners that was then used to build the Meta-Model. The requirements for the kind of models and data needed were studied and this information was sent to the partners. Data files and format examples were used to test the performance of the Web platform. The physiological module of the Meta-Model (WP4) was updated using experimental data from WP3 and with published data. This was implemented in the Circulator software package (ISBSPb).

WP4

First 18 months of the project has been a very intense period in WP4. Outstanding results have been obtained at both main tasks of the WP: “Development of the software platform”, and “Development and implementation of the systemic Meta-Model”. Regarding the first one, a first version of the Web platform has been developed, which integrated tools for experimental data management and a SBML model designer and manager. The platform will also gives access to the Meta-Model server (MMS). Interface and communication to the MMS has been also defined. Regarding the second objective, the draft version of the Meta-Model of lipid spillover in whole organism was developed. Kinetic model of FAs metabolism in hepatocyte was added to Meta-Model. Its adequacy has been initially demonstrated through virtual experiments.

In the 2nd period, the development focus was on 1) new informatics module that gives access to the Metamodel Server as well as improvement and validation of the existing ones and 2) Systemic Meta-Model. Regarding the first one, a new version of the platform was released. Improvements and validation works were carried out for Experiments manager and SBML Models Designer modules. Meta-Model Server module was implemented based on the previous requirements analysis. Regarding the second one, inputs and outputs of the mathematical models have been interconnected with corresponding outputs and inputs of meta-model of lipid spillover in whole organism. The Meta-Model of lipid spillover in whole organism was updated too.

In the 3rd reporting period, NorayBio finished the web platform to allow data interchange and to make and manage simulations against the Meta-Model Server and their results (Figure 7). It has been programmed in such a way that will be easily updated in case new processes are implemented in the future in the Meta-Model Server.

The University of Barcelona team (UB) released two models: model of adipocyte central carbon metabolism and model of hepatocyte central carbon metabolism, both implemented in SBML and adjusted with IsoDyn software. ISBSPb implemented the final version of meta-model that integrates literature available experimental data and reproduces some of experimental data obtained by other participants of project was developed. This version was used for optimization and validation of meta-model server and GUI.

Unlike other software platforms, the software platform released in WP4 allows the joint study of digestive, lymphatic and circulatory systems in humans and aims to become a new reference tool for the research on lipids nutrition.

WP5

The four diets have been developed differing for -3 fatty acids and polyphenols content to be used in the ETHERPATHS nutritional intervention study in humans: (A) Control diet low in n-3 fatty acids and polyphenols; (B) Diet rich in n-3 fatty acids and low in polyphenols; (C) Diet rich in polyphenols and low in n-3 fatty acids; (D) diet rich in n-3 fatty acids and polyphenols.

Beside the different content in n-3 fatty acids and polyphenols, the four diets are similar for all the other characteristics, including macronutrient composition and content of micronutrient that could possibly affect outcomes, especially for their antioxidant properties.

Polyphenol-rich foods were subjected to the colon model for identification of the microbial metabolites of polyphenols in WP1. Acceptability of the foods identified for the different dietary approaches has been evaluated in a group of healthy subjects.

The screening of the subjects for the intervention study started in September 2009 and continued until month 38 of the project.

The preparation of the foods/meals to be given to the participants to the intervention study started in the 1st project period. The schedule for type, time and modalities of sampling, storing and shipping for the blood, urine, faeces and adipose tissue samples to be analysed at the end of the intervention periods was defined after an intensive email communication among the WP5 participants.

Screening, enrolment, and completion of the study during the 1st period: Out of the 50 subjects fulfilling inclusion criteria, 18 was enrolled and randomized to the dietary intervention. Of these, 13 subjects completed the study by month 18. Five subjects were enrolled and completed in month 19. There have been 2 dropouts, one because of moving to another town, one because of serious family health problems. During months 19-54, 65 subjects were enrolled and randomized to the dietary intervention. Out of these, 62 subjects completed the study and 3 were dropouts, because of moving to another town, or serious family health problems. All planned subjects completed the dietary intervention by month 42..

The main significant changes were observed for the groups with high intake of polyphenols. There was number of significantly increased microbial metabolites in the 24h-urinary excretion, indicating the microbial metabolites as compliance biomarkers of polyphenol intake.

Dietary compliance was optimal shown as 7-day food record and as for polyphenols according to the microbial metabolites profile in the urine. Urinary phenolic metabolites increased with high polyphenols diet. Differences with in vitro colon model results likely also reflected polyphenols hepatic metabolism.

The results show trends to differential effects of dietary interventions on postprandial lipoproteins. In particular the diet rich in polyphenols (Figure 8)

- Reduced triglyceride-rich lipoproteins both at fasting and in the postprandial period
- Reduced HDL cholesterol and triglycerides both at fasting and in the postprandial period
- Had no significant effect on fasting and postprandial glucose and insulin levels
- Reduced urinary isoprostanes

VTT evaluated whether different storage conditions and/or the mostly used DNA-extraction kits affect the Firmicutes and Bacteroides results derived from human faecal samples. The samples were analyzed as fresh and after three different storage temperatures. DNA was extracted using two different, mostly used, commercial kits with 8 different protocols. One kit was based on enzymatic lysis (Qiagen) and the other on mechanical lysis (QBiogene).

The results showed that the mechanical DNA-extraction method to be used for the analysis of the faecal samples of the human nutritional intervention trial to quantify and characterize both Gram-positive and Gram-negative bacteria is as efficient as possible. In addition, the chosen storage temperature (-70°) is optimal. The resulst have been published.

Different diets had different impacts on the faecal bacterial populations. In particular

- The predominant bacterial profiles (Univ-DGGE) were most stable after high consumption of n-3 fatty-acids
- Bifidobacterial profiles (Bif-DGGE) were most unstable after high consumption of polyphenols, The most inter-individual similarity was observed for the Eubacterium rectale group.
- With the diet rich in n-3 fatty acids and polyphenols the most diverse was Bacteroides spp. and the bacterial numbers of the studied bacterial groups increased in over half of the subjects.
- Thus, in the case of fecal microbiota the response to the combination of nutritional substances does not necessarily equal to the cumulative response of singular nutritional substances.

WP6

The initial focus of WP6 was on setting up the infrastructure for the planned analytical developments. The novel time-of-flight mass spectrometric system was installed and testing of the system was performed. In addition the development of new, more efficient methodologies for the characterization of lipid profiles in biological samples was performed. The development was divided into four parts, namely improvement of the global screening method for molecular lipids, development of tools for more accurate identification of unknown molecular species, development of fractionation procedure and development of methodology for the determination of hydrophilic metabolites. During the 1st reporting period, new analytical method was developed for the global screening of the lipid profile and it was utilized in WPs 1-3. For the identification, mass spectra libraries utilising the accurate mass and MSn fragmentation obtained by the UPLC-Nanomate_LTQ-Orbitrap instrumentation were collected for different sample matrices, including serum and tissue.

During the 2nd reporting period, the fractionation procedure for separation of neutral lipids, glycolipids and phospholipids into individual classes was developed. A UPLC-Nanomate-LTQ-Orbitrap MS instrument was used for structural elucidation of molecular lipids, especially for the characterisation of alkyl-ether and plasmalogen lipids. In addition, the new UPLC-QQQMS instrument was utilised in sensitive analysis of hydrophilic metabolites.

During the 3rd period, the method development for the analysis of hydrophilic metabolites was continued to cover additional classes of endogenous molecules. The improved lipidomics method together with the new lipid library was used for fasted plasma samples of WP 5 human intervention.


New data analysis methods for the treatment of the metabolomics data were developed. For the GC×GC-TOFMS data, an updated version of the software has been developed, allowing for more reliable alignment and normalisation of the data, and as new features, functional group-type identification based on mass spectral fragmentation patterns, as well as several practical features such as search of PubChem and KEGG indexes, features for statistical analyses (PCA, ANOVA, t-test, fold change). The main features of the software are summarised in Figure 9.

In summary, developments in different research lines as divided into different work packages, were increasingly intertwined as the project progressed. All the main objectives of the project were so met, and both the computational/informatics, analytical tools and protocols were developed which together comprise the nutritional systems biology platform.

Potential Impact:

Impact listed in the call
ETHERPATHS aimed at achieving three main impacts, based on expected impacts listed in the call:

(1) To improve and strengthen the collaboration between theoretical (dry-lab) and bioanalytical (wet-lab) research.

ETHERPATHS involved close cooperation of relevant theoretical and bioanalytical research at multiple levels: (A) the in vitro studies, focusing on gut microbiota as well as the effect of their metabolic products on host cells, were complemented with modelling of host (adipocyte and hepatocyte) cells, (B) animal model studies were complemented by computational and bioinformatic studies aiming to reconstruct the tissue specific pathways regulated by specific nutritional interventions, and (C) clinical studies with nutritional interventions were complemented with physiological modelling, specifically dynamic models of systemic lipid metabolism. In order to facilitate such dry-lab/wet-lab cooperation in practice, the organization of S/T work packages reflected such a scientific strategy.

In addition, in silico modelling was being used to translate the wet-lab research across multiple levels. Specifically, models of systemic lipid metabolism, obtained from human kinetic studies, along with detailed molecular characterization of lipoprotein composition, helped to translate the knowledge obtained from nutritional intervention trial into the human physiological context. As a result, tissue-specific (dys)regulated pathways obtained by pathway reconstruction in specific in vivo models were interpreted in the context of human physiology. Furthermore, the cellular models of adipocyte and hepatocyte metabolism were investigated in the context of these pathways. This was made possible by application of in vitro colon model and identification and isolation of metabolite pools obtained from nutritional interventions, that was applied to the host cells, thus mimicking the physiological setting. The meta-model developed by ETHERPATHS provides a framework for combining models and the data obtained from the models as well as from the bioanalytical platforms across in vitro, in vivo, and clinical levels.

(2) To increase the knowledge and give new impetus for health-oriented nutrition research.

The ETHERPATHS focus on lipid homeostasis and its dietary modulation is of central importance to human health. By using powerful systems biology and bioanalytical tools available to and developed by the ETHERPATHS consortium, new knowledge was generated and disseminated on cellular and systemic metabolism of plasmalogens, on dietary components altering their homeostasis, as well as investigation of potential role of gut microbiota and its metabolites in mediation of plasmalogen metabolism. Such knowledge will be of high importance for health-oriented nutrition research, and will open new possibilities for development of health-oriented nutritional products aiming to maintain or improve lipid homeostasis.

(3) Computer based tools describing mathematical models of metabolite and energy control is a basic demand for the further development of nutrigenomics and personalized nutrition concepts.

ETHERPATHS was heavily focusing both on developments of computational models at multiple levels, as well as bridging them with bioanalytical tools providing data for model building as well as for validation. In ETHERPATHS, complex mathematical models at cellular level, focusing on host mammalian cells and enforced by a module for stable isotope tracer data analysis, combined with experimental metabolomic data from the in vitro colon model, were complemented by tissue-specific pathway reconstruction methods as part of the in vivo studies, and with the dynamic physiological models, focusing on systemic lipid metabolism, i.e. the systemic level of energy control. Unique strength of ETHERPATHS was that these multiple levels were combined both by physiologically motivated models, as well as by practical implementation using powerful software engineering and informatics tools.

Participating SMEs

The project contributed to developments of innovative products and offerings of all three participating SMEs. Noray Bioinformatics has developed a database system, software and web interface which facilitates construction and use of the mathematical models for nutrition research. Institute for Systems Biology SPb has developed new models for systemic multi-level and multi-compartmental studies of metabolism. The modelling and bioinformatics tools developed by these two SMEs will be incorporated as part of their offering and thus increase their competitiveness. The two participating analytical SMEs (Advion Biosciences, BioMotif) have benefited by further developments of analytical protocols on their platforms, which will lead to new application possibilities and thus potentially to opening new markets for the SMEs.

Wider socio-economic impacts

The development of health claims for food products has been a great challenge for companies in the nutrition sector. One of the difficulties has been assessment of the scientific substantiation of the claims. Ideally, the scientific evidence for any specific claim would include strong clinical evidence for specific health-related (but not medical) outcomes of interest, such as e.g. amount of liver fat, as well as provide a mechanism how a specific food product affects the outcomes of interest, i.e. establish a direct mechanistic link between the food product consumption and the clinical outcome. Such mechanistic insights can only be acquired if one is also studying the physiological effect of foods at different levels, including in silico, in vitro and in vivo. ETHERPATHS has developed a platform for such a purpose and demonstrated it by studying the effects of specific foods on lipid metabolism. However, the platform as such is generic and with problem-specific modification can be applied to study different food products, related to different health-related outcomes.

The model how ETHERPATHS could be applied in nutrition research is shown in Figure 10 below. In short, the health-related outcomes such as specific markers (e.g. the amount of liver fat, which is a known risk factor of diabetes and cardiovascular disease) would be derived from clinical research. A specific food product aimed at affecting the liver fat would then be studied at multiple levels using the ETHERPATHS platform (e.g. by pursuing studies with in vitro colon model, hepatocyte and adipocyte cell cultures, suitable animal models, and ultimately a nutritional intervention study in humans – all connected with a suite of tools comprising the ETHERPATHS platform).

Together, the ETHERPATHS has a potential for economic impact by providing an improved pipeline for health substation of food products, a possibility which will be exploited in the future. Furthermore, use of the ETHERPATHS platform in this setting may ultimately lead to healthier and safer foods, thus leading to public health benefit.

Dissemination

The project participants have been active in publishing scientific papers during the whole project, and as expected these activities have intensified during the last period. Most of the key outcomes of ETHERPATHS are yet to be published. So far, ETHERPATHS papers have occurred in many leading journals in specific domains, including Cell Metabolism, PLoS Biology, PLoS Computational Biology, Gut, Analytical Chemistry, Journal of Nutrition, Bioinformatics, Diabetes, and Diabetes Care, to name a few.


During the last period of the project, the dissemination and exploitation activities have been intensifying. The ETHERPATHS platform has been presented at multiple conferences. For example, there was a special ETHERPATHS session at NuGOweek 2012.

Also a book is released in September 2013 ‘A Systems Biology Approach to Study Metabolic Syndrome’ (Springer Verlag, ISBN 978-3-319-01007-6, eds. Matej Orešič and Antonio Vidal-Puig, ETHERPATHS coordinator and WP2 leader, respectively). The edited volume includes among others also 10 contributions from ETHERPATHS participants, introducing some of the key methodologies and tools comprising the ETHERPATHS platform. The project is specifically acknowledged in the Preface of the book.

List of Websites:

Contacts

Coordinator: Professor Matej Orešič; email: matej.oresic@vtt.fi
Financial Officer: Sinikka Soirinsuo; email: sinikka.soirinsuo@vtt.fi
Scientific Officer. Anna-Marja Aura; email: anna-marja.aura@vtt.fi
The address of the project public website: www.etherpaths.org