Final Report Summary - FIBEBIOTICS (Dietary Fibers supporting Gut and Immune Function - From polysaccharide compound to health claim)
Executive Summary:
The goal of the FP7 KBBE EU FibeBiotics project was to support the development of functional food ingredients and products that are beneficial for the human gut and immune system and therefore of crucial importance for quality of life. The project studied the effects of dietary fibres, more specifically non-digestible polysaccharides (NPS) with a sugar backbone of 10 units and more. Previous research, mainly based on animal trials, has shown potential health effects of the NPSs towards the gut and immune system. In this project the health effects of the NPSs have been focused around enhancing immune defence against pathogens, increasing the gut barrier function and the reduction of infectious diseases such as common cold and influenza of elderly. The project wanted to make use of European Food Safety Authority (EFSA) supported biomarkers that would support immune function claims and underpin the mechanism involved. The studied mechanisms involved in the health effect of NPSs were the innate and adaptive immune system and the possible involvement of the microbiota and microbiota-mediated products such as short chain fatty acids which also have shown to have immune modulating properties. To achieve this goal, new and existing NPSs have been studied for their health effects in a systematic way by developing a toolbox of dedicated assays and models that can now be used by research entities and industry to study bioactivity of food ingredients with a similar health focus. The project has:
1) performed biochemical analyses to characterise the NPSs and study effects of fermentation in the gut and the fate of NPS when formulating final food products,
2) developed standardized in vitro screening methods to compare different NPS for their bioactivity towards gut and immune cells and their effects on microbiota and microbiota derived compounds,
3) performed dedicated in vivo and ex vivo analyses to study the effect of NPSs on gut barrier function,
4) studied the effect of a 5 weeks intervention of NPS in humans based on a pilot trial (6 arms, with 5 different NPSs and a placebo) and of a 3 months intervention in a pivotal trial (2 arms, 1 NPS and 1 placebo) with vaccination and common cold as outcome parameters,
5) developed a database and explored different mathematical methods for data analyses,
6) analysed the consumer opinion on fibres and immune support and also kept track of EFSA guidelines.
Dissemination has mainly been realised via an established Industrial Platform, to which 42 large enterprises and SMEs subscribed and that attended the meetings organised by the EU FibeBiotics consortium. Several scientific papers have been realised but still many are in the pipeline.
Project Context and Objectives:
Project Context
Dietary fibres are, by all food-related-authorities, recognised for their contribution to human health. This positive effect of fibres has mainly been postulated on the basis of epidemiological studies and, to a lesser extent, intervention studies. This has led to the recommendation of 30-40g fibre intake per day depending on the countries responsible for the advice. The general consumer, however, hardly reaches an intake of 20g dietary fibre per day. So, the awareness among consumers is not high or the willingness to act according to food council advices is not high or people do not have knowledge about which food products contribute to a high fibre intake.
Advising on the intake of dietary fibre also brings some uncertainties, as there are thousands of different dietary fibres in our food. This warrants the question whether they all contribute to a similar extend to maintenance of health. Controlled intervention trials have shown that some dietary fibres can reduce cholesterol levels while other fibres failed to show these effects. Apparently, the type of fibre can have different bioactive effects just like these fibres show differences techno-functional properties, such as water binding capacity, or conferring solubility and stability etc. to food products.
Dietary fibres include non-starch polysaccharides, resistant starch, resistant oligosaccharides (with three or more monomeric units), and other non-digestible polysaccharides. The terms “soluble” and “insoluble” have been used to classify dietary fibres in an attempt to link different physical-chemical properties of fibres to different physiological effects. As a consequence of the complexity of the large group of dietary fibre, public food databases (e.g. USDA food database, Euro FIR) can only describe the fibre composition and cannot relate it to health effect. In FibeBiotics, we focused on a specific set of compounds among the dietary fibres, the high molecular weight non-digestible polysaccharides (excluding e.g. the oligosaccharides), and the term NPS (Non-digestible PolysaccharideS) has been chosen to specify this.
This proposal aimed at developing foods ingredients, based on non-digestible polysaccharides, which were targeted towards a maintained and improved host’s immune system. The immune system is of crucial importance for human quality of life. A dysfunctional immune status will result in numerous health problems including infection diseases, autoimmune diseases, allergies, inflammation, atherosclerosis, cancer etc. Given its important role and the fact that the majority of the immune system is located in the gut, the gastrointestinal tract has become one of the most important target sites to maintain or improve the function of the immune system even in apparently healthy individuals.
Nutrition has been shown to be an important modulator of gut and immune functions. It is, therefore, not surprising that the majority of the functional food market is directed towards the gut and immune system, e.g. probiotics and prebiotics (mostly based on oligosaccharides) and fibres or non-digestible polysaccharides. In the FibeBiotics project, we included several NPSs which have been hypothesized, based on preliminary research or published data, that they can have beneficial effects for gut function, mainly in relation to bowel habit, cholesterol lowering effects and reduced risk of cancer.
Less known were the direct immune supportive effects of NPS that have been demonstrated in in vitro and in vivo animal models and a small number of human trials. In particular, cancer related research has shown the potency of different NPS and their contribution to the immune system, a crucial pillar in the combat towards tumours. These studies have also indicated the variation in bioactivity among these NPSs and the gaps in knowledge such as structure-function relationships and mechanisms of health effects that need to be fully clarified in order to apply fibres efficiently. Unravelling the mechanism and developing good models that can predict health effects in humans, to select the NPSs with most activity were therefore main goal of the EU FibeBiotics project.
The European Food and Safety Authority (EFSA) is responsible for the evaluation of the scientific basis for the health effects of food compounds and products. Based on scientific panels, they are evaluating biomarkers, trial designs and study results and come with guidelines how to substantiate the health effects in specific areas. The legal limitation to approach food not in relation to disease threats but rather from the perspective to maintain health of healthy individuals, limited the possibilities to substantiate the effects. This requires the need for validated and broadly accepted (early) risk factors that warns for derailing of homeostasis or indicates a deficiency for a certain (cellular) function.
Based on the evaluation of immune related biomarkers, EFSA (draft guidance of the EFSA, 2010 and repeated in updated guidelines) stated that vaccination efficiency as objective biomarker could suffice to demonstrate positive effects of interventions towards the immune system. This could lead to health claims that would go broader then the used vaccination model but would indicate a more general enhancement of the immune system. These analyses would, preferably, also be supported by endpoint measurement like showing an effect of the intervention on reduced symptoms in the population.
This vaccination efficacy was used as one of the pillars of the FibeBiotics project. In order to have a broad impact of the intervention (specific NPSs that support the immune system), we wanted to target a large population that could benefit from this approach, the elderly. Elderly have been chosen since clinical efficacy of the flu vaccines in elderly subjects is around 60-90%, indicating that their immune system might be prone to be boosted to reach the same level of protection as young aged persons. Flu and post-flu pneumonia are important causes of mortality in elderly subjects. That is why, we also chose to use the common season flu vaccination as a model in order to analyse whether an intervention with specific NPSs could increase the antibody titres and number of individuals that are protected against the flu. Dietary NPSs may provide a valuable solution to this health problem which may be applied in a broad, population-wide approach both to support other vaccination models with low efficacy, use in other populations with a suppressed immune system (or temporarily due to extreme high challenges like sports or high risk jobs) or support populations that have high exposures to flu and colds (e.g. day-care workers). So, when these beneficial effects could have been proven by scientific sound evidence, the benefits for the population are enormous and costs for loss of working productivity and costs for health care will decline.
The NPSs are non-digestable dietary fibres, meaning that the human digestive system cannot degrade them like it can for instead degrade starch into glucose. However, when fibres enter the colon, they can be the food for the bacteria that are living there. The microbiota is important for human health as they produce and release vitamins, minerals, fatty acids etc. which still can be taken up from the colon by the host. These microbiota gain more and more interest as they can help to reduce the growth of pathogens that are latently present in the gut or can reduce the impact of new intestinal pathogens that enter our system orally. Correlation analysis between the presence of specific species in the intestine and health and disease have led to the identification of potential risky and beneficial gut bacteria. So are the colony numbers for Bifidobacteria and Lactobacilli often found to be high in healthy subjects. Research is still ongoing to unravel these correlations and whether they are cause or consequence of changes in health. At this moment, no general consensus has been reached within the scientific community and that is why the microbiota profile cannot be used to substantiate a risk-benefit analysis of a certain intervention. Only the more commonly known gut pathogens can be seen as risk factors. Demonstrating a significant reduction in the presence of these in lumen samples could help to substantiate a health effect. NPSs, as they can be selective food for some bacteria that can use them for their own resources, may also have a positive contribution to reduction of pathogens. Research, however, is difficult as natural infection rates among people are low and working with infected people difficult, for ethical reasons. Therefore in the FibeBiotics project, we included research to understand the impact of NPS on microbiota but did not make it the focus of the proposal.
Establishing a health effect of food products or compounds should be performed by controlled intervention trials. These trials, however, are costly, laborious and impact the included persons. Models that can predict the outcome of a human trial with high certainty would be very valuable as it can be used to: 1) support the mechanism of action of the bioactivity involved; 2) can be used to select and compare food ingredients to select the best bioactive ingredient; 3) can support product development with sustained bioactivity (knowing that matrix or other processing unit operations could breakdown or counteract bioactivities); 4) can be used to check products with a certain health claim and expected bioactive content for its reliability. As very often biological in vivo activity cannot be simulated by one model and as the bioactivity of these NPSs can be hypothesized to be brought about by different mechanisms we have set up, in the EU FibeBiotics project, several different in vitro models. These models could be used to systematically compare the different NPSs for their bioactivity. This type of systemic comparisons is hardly performed within the scientific community where a large diversity of protocols is used and focus is only on very limited numbers of compounds of interest. The adoption of these standardised models is important as they will make studies between labs more comparable, leading to better understanding.
As discussed at the start, many people know about the health effects of dietary fibres but on average do not reach the daily recommended doses. That it is why, it is important to comprehend consumer behaviour, both assessing the current knowledge level on fibres, the associations that are made, what kind of product they would take at what moment to increase fiber intake and what food products they would find acceptable to be enriched with bioactive fibres. Therefore, FibeBiotics also included research towards European consumers. Knowing that much of the Business to Business (BtB) or Business to Consumers (BtC) approaches start with presenting correct, credible and attractive information on the product, we also evaluated current strategies applied in this field and how companies associated to the FibeBiotics project could position their own products.
The EU call from which this project was selected requested a clear involvement and commitment of SMEs. The SME partners within the FibeBiotics consortium were in the field of technology and service provision in the area of food and health research and those that had their main business as supplier of ingredients for which they would like to be informed on the health effects of the NPSs within their product portfolio.
The objectives of the project
The overall objective of the FibeBiotics project was the development of functional food products, based on bioactive NPSs that support gut and immune function to maintain a healthy homeostasis in the elderly population.
Our objectives were to support the entire food industry (and not only the associated partners) by developing standardized methods, tools, pipeline and systematic analyses of compounds for their bioactivity. These technologies should be available to other partners in the future to develop more products that will meet the scientific requirements to allow claims on gut and immune function. These tools have been developed in the project by focusing on three main NPS compounds delivered by three of the SME partners but ensured at the same time, the translation of knowledge to other industrial partners that have the same or similar aim in developing products that support the gut and the immune system.
The objective to develop a NPS based functional food for promoting immune functioning, had been divided into five main objectives:
• Define methods to detect and characterize purified and processed NPS. Determine biochemical and bioactive alterations when designing functional food products in order to make products with remained activity.
• Develop highly standardized methods to classify purified polysaccharides for their bioactive effects on immune cells and microbiota and microbiota-mediated metabolites using in vitro and ex vivo assays.
• Improve the knowledge on the mechanism of action of NPS based on in vivo and ex vivo analysis as a result of small scale interventions.
• Combine all knowledge and show health promoting effects in a large pivotal human clinical trial.
• Perform all necessary translational activities for implementing this knowledge to support the claim substantiation of the studied NPS but also provide the tools, knowledge and part of the data to enable functional food research with a similar health target.
These objectives have resulted in many sub-objectives that will be addressed when discussing approaches and results.
Project Results:
Before going into details on the main S&T results/foregrounds it has to be said that the project had two phases. The first one was on developing, testing and establishing novel tools and strategies open to apply for everybody in the field to study health effects of dietary fibres and similar compounds. The second was to apply these tools to a selected set of NPSs, supplied by some of the FibeBiotics partners in order to build on a portfolio that might be used in the future in applying for an EFSA health claim when results turned out positive. Especially, the second goal of the project resulted in a careful evaluation of all the results before making them public, not disclosing results that could jeopardise potential IP that could be filed and not damaging business of involved identities. That is why all reports remained restricted within the consortium and also now we will talk in general terms on the results making it impossible to relate effects to specific NPSs studied in the consortium.
The NPSs within the project
Research could not have been done without NPSs as method development requires these NPSs, specially as we are exploring a new field on gut and immune support for which hardly positive control products, nor medicines, are available.
Within the consortium we had six partners that were responsible for preparing and providing these NPSs to the other consortium members for their research (see figure 1). Besides these main NPSs, in specific experiments other NPSs were taken along as reference when in the literature certain bioactive effects had been described which we wanted to check or compare with the newly included NPSs in the same experimental set up.
Biochemical analysis of NPSs
NPSs are derived from biological material which can vary due to genetic differences of the starting material, differences in growth conditions, harvest time, post-harvest treatments and differences in purification method. Therefore the project mostly starts with biochemical characterisation of the NPSs to determine how much of a specific NPS is present in the sample, whether it is one specific NPS or a mixture of NPSs and whether other compounds are in the same sample. This is important to compare different NPSs (based on a similar concentration) but also to learn about the variation in different batches of the same NPS. As long as we do not know what exactly determines the bioactivity of an NPS we also do not know whether these variations in samples are important. Therefore, at this stage, we liked to know as much as possible on the biochemical structure and conformation of the NPS and the other constituents in the sample that might have an impact on the bioactive properties subjected to our study.
Within the EU FibeBiotics project, we had two groups with expertise on biochemical characterisation of polysaccharides like the NPS studied here. As these NPSs are very complex and it requires several biochemical methods to fully characterise samples, these experts developed a schematic approach how to analyse these samples step-wise (see figure 2).
These methods and strategies have also been included in deliverable reports. They can support other groups in this field to analyse samples in the hope to reach more consensus on how to describe samples, as a first step to a more systematic comparison of NPSs and their bioactivity.
An example of the results of such a biochemical characterised can be found in Table 1. These results have been used in one of the publications where the effects of these NPSs preparations have been studied towards the small intestinal epithelium (see later). As part of sample characterisation we also analysed the presence of potential contamination (like endotoxins) in the sample. These small amounts of contaminants are no problem for most of the research to be done, but could form a problem (i.e. false positives) when studying the direct immune effect of NPSs on immune cells as these cells are very sensitive to these endotoxins. In scientific literature, sometimes proper controls for contaminants are not presented which, to our opinion, hampers interpretation of several publications worldwide.
More than expected at the start of the project we indeed encountered the presence of too high concentrations of endotoxins in the samples to be able to use them directly in some of the immune assays. Therefore extra efforts have been invested to further purify some of the samples for immune cell related research. As purification might again change biochemical structure of the NPS, the development of this purification protocol was supported with biochemical analysis. Figure 3 shows the effect of the purification strategy developed, in order to reduce endotoxin to acceptable levels, for one of the NPSs. It is crucial to find a method with the right balance between ‘not too harsh to change structure like the molecular weight’ of the starting NPS material (Figure 3 - blue line) but ‘harsh enough to remove the endotoxin of the sample’. The research resulted in a paper accepted for publication and was the key to continue the research towards analysing the immune supportive effect of NPS. The developed method can also be applied in similar studies to remove endotoxines from other samples for which immune modulating effects are studied.
Product formulation with maintained NPSs
In parallel to the research to establish the effect of NPS towards the gut and the immune system it is important to know whether you can make products that still contain the same NPSs as we used in the bioactivity studies (the purified compound without the processing steps of a real food formulation). That is why within the FibeBiotics project, several products have been prepared including cheese that contain EPS, smoothies that contain beta-glucan from oat and yeast and crisp bread that contain different NPSs supplied by the consortium (see figure 4).
Biochemical methods were subsequently used to study the fate of the NPSs in these food matrix in order to understand potential risk when preparing products enriched with these NPSs. In addition, these biochemical methods will be important for food authorities to check compliance of food producers that the claim they might make on the package (like ‘yeast beta glucan inside’) is indeed reliable. These products have subsequently been used to validate strategies to analyse these specific NPSs in these food matrices. This is not an easy task as food matrices are complex and can have very similar compounds present (e.g. how to distinguish yeast used for making the dough from those derived from the yeast NPS specifically added to the product because of health effect). But we can say that major steps have been achieved and have resulted in methods, protocols and strategies that can be used or further developed (see figure 5). One paper has been published on NPS detection methods and another, describing the methods to study fungal beta glucans in different food matrices making use of alanine blue, will be prepared and submitted.
Fermentation of NPSs in the gut
To be able to demonstrate gut and immune supportive effects it must be clear that the NPSs will not be hydrolyzed into small mono sugars units in the mouth, stomach or small intestine. As defined by the term NPS (non-digestible Polysaccharide) it is expected that these polysaccharides resist the passage over the small intestine (or at least most of it) until the colon were they then might be degraded by microbes. The subsequent degradation of the NPSs in the colon will mostly depend on the bacterial species present and the enzymes they produce. However there is also a lot of redundancy between the species present which can take over the same capacity for fermentation in the gut ecosystem. To check whether these NPSs indeed are not broken down too rapidly in the gut, several models can be applied. The most simple method is based on single tubes in which human enzymes (like amylase and brush border enzymes), pH changes and bile acid etc. are added stepwise to simulate the mouth, stomach and small intestine. Indeed the NPSs of interest in this consortium survived these upper gastric gut simulations. The next step was to culture these samples in the presence of a human feacal sample containing a high complexity of microbiota. This research led to the conclusion that some of the NPS are more resistant to degradation by bacteria than others. An example is given in figure 6 for the NPS AX indicating that after an incubation of 24 h some parts of the molecule are degraded (bacteria needed to switch on the right enzymes and have to grow in number) while other parts of the molecule are to difficult to digest and remain in the sample until the end of the experiment (48h). These type of analyses made clear that NPSs can differ in the resistance towards fermentation in the colon.
Fermentation of NPSs can also be associated by gas production which is in generally less preferred as these can cause bloating effects in the gut.
In vitro analysis of NPS towards the microbiota and related metabolites
These relative simple experiments as described above were given a follow up in more advanced research models. One is the SHIME (Simulation of Human Intestinal Microbial Ecosystem) that can simulate the trajectory of a product from mouth to colon in a more automated way and that also includes a separation of the colon in three parts. This type of analysis can support further understanding of the fate of NPSs in the human intestinal system (see figure 7).
This model has been further optimised within this FibeBiotics project and subsequently has been applied to study the fate of some of the NPSs in more detail. Not only the breakdown of the NPSs is important knowledge, also the formation of fermentation products like short chain fatty acids (SCFA) that can support in local and systemic metabolic regulation and can signal towards the immune system are important to measure. This fermentation of fibres in SCFA is in general considered as a generic positive health effect of dietary fibre as almost all dietary fibres are converted into SCFA. However, the type of SCFA (e.g. acetic acid, propionic acid, butryric acid that are the major ones) produced during the fermentation can be different, the ratio among these can vary and the total concentration of SCFA can slightly vary between the different dietary fibres. Besides, depending on the bacterial species that are involved in this fermentation, fermentation of the same NPS can therefore depend on the feacal sample used for the fermentation study. Within a SHIME these SCFA can be determined at different places in the intestine and give more accurate measurements than analysis done on humans for which on a routine base only faecal sample are available.
However, these faecal samples indicate produced SCFA that have not been absorbed by the host and therefore hardly will indicate the real benefit to the host. Part of the FibeBiotics project was therefore dedicated towards the development of an on line SCFA sensor that might be applied in vivo in the intestine when possible or applied in a model like SHIME to get more accurate measurements. One of the deliverables describes the development and the test phases in more detail (figure 8).
To cut a long story short, it has to be concluded that another type of sensor than the current MOS (metal oxide sensor) type sensors might be more successful (due to stability, energy supply, working in a 100% humid environment and at low oxygen levels as in the gut) therefore more developments will be needed.
Besides the SCFAs, the SHIME also enables the simulation of the effect on pH which might be partly linked to SCFA production. The hypothesis is that lowering the pH in the colon may help to boost the most positive know SCFA butyrate and might also enhance butyrate producing bacteria. Indeed, using this model also effects on pH have been measured.
Like stated earlier, the microbiota in the intestine can bring both risk and benefit to the host. Risk factors are endotoxins, exotoxins and harmful metabolites that can be produced by the bacteria. On the other side vitamins, SCFA, and pathogen-associated molecular patterns (PAMPs) that make the immune system alert but without affecting the host negatively can have immune modulating activity (and hence represent a potential indirect immunomodulatory effect of NPSs). We studied, making use of the SHIME model, changes to gut microbiota and metabolites when replacing starch, given as food intake in the SHIME model, by different types of NPSs. Figure 9 gives an example of such an experiment indicating that, when starch is replaced by beta glucan from oat, more Lactobacilli, a group of bacteria associated with health benefits, will grow in the colon with higher (relative) abundance.
Supportive tool to study microbiota
Some human diseases have shown correlation with the present of absence or abnormal ratios of certain microbiota in faecal samples. There are several technologies to study the microbiota community composition ranging from specific quantification of only a very limited number of species to full sequencing of all DNA, RNA or rRNAs or certain areas of the genome (mainly 16S) which allows identification and quantification. These full sequencing analyses decrease in cost with the Next Generation Sequencing technologies that are being developed but still are costly when many samples from a large cohort or intervention trial need to be analysed or when it would be used for routine screening of a population. Therefore, the FibeBiotics project also gave attention to new cost-efficient microbiota profiling technologies that could be beneficial for some specific application.
The GA-map™ analysis consists of various steps starting from DNA isolation of faecal samples to analysis of the microbial profile (figure 10). This profile is based on 48 unique bacteria probes able to be detected simultaneously in one single tube. In the FibeBiotics context the technology was further optimized and adapted in order to study the 48 different bacterial groups that can profile the microbial community of elderly European subjects. The developed probe-set has been tested in selected samples which are representative of the elderly population, and compared to the 454 pyrosequencing data to confirm correlation. A very good correlation has been observed. Several optimization steps have been included to increase robustness and reproducibility. The University of Bologna has been included to for data processing and statistical analysis. Subsequently the technology was used in the human intervention trials to study the effect of NPS intervention towards changes in the microbiota composition.
NPS bioavailability
In pharmacolog,y it is very common to perform pharmakinetic research in order to determine the bioavailability, organ targeting and metabolisation of the compound of interest. For food compounds these data are less required, especially as the food ingredient already has a novel food status. However, the place of action of NPSs in the body is a recurrent question; will these often high molecular weight molecules be restricted to the intestinal lumen or can they pass the intestinal epithelium and enter the blood stream to interact fully with all kind of immune cells and other organs than the intestine.
Previous research had shown that a labeled, high molecular weight NPS, could be traced back in the blood after oral intake, although at low percentage. We wanted to study whether this could also be the case for other NPSs but then not making use of labelled NPSs (expensive) but by sensitive biochemical methods to make it more generic. When succesful, these type of analysis might be added to EFSA dossier in order to support a specific hypothesis. Spiking experiments in which NPS were added to blood and then analysed gave some results but also a lot of background from the very complex blood matrix, even after applying several purification methods. It had to be concluded that it will be impossible to detect the NPSs in blood after oral intake as concentrations will be too low in relation to disturbing factors in the blood. When EFSA would require such information in dossiers, research should be performed with expensive labeling experiments.
In vitro analysis of the effects of NPS towards gut and immune system
As explained before we would like to be able to characterize and select NPSs based on their bioactivity in humans. However, still several hypotheses for this function in humans are possible. Therefore we studied this bioactivity towards different cell types and with different response read-out in order to study whether there are differences in bioactivity between the NPSs of interest. Besides, these analysis hint towards the potential involved mechanism underpinning the effects found in vivo. The in vitro assay will then also be able to support the development of products with the best possible effect resulting in a highly potential positive outcome in human intervention trials. Besides, these in vitro assay can support quality control of the product when something changes in the production and processing of the product and can be used to check products for their claimed bioactivity.
Several assays have been developed and used to study the effect of NPSs. The first is based on human small intestinal epithelial cells, the first cells that are exposed to NPS when orally provided. These cells are cultured in the lab and can be exposed to NPSs. Based on gene expression and analysis of compounds in the excreted medium it was concluded that some of the NPSs could activate immune related genes like NFkB and as a consequence could secrete chemokines basolaterally to attract other immune cells to the place of action. This could support the hypothesis that some NPSs are recognised by the intestinal cells as ‘non self’ or ‘strange’ (as some have similarities with cell walls of pathogens) and subsequently alert the immune system. A publication on the obtained results has been submitted.
The second assay that was developed and used is based on human macrophages. Macrophages are shown to reside in the gut epithelium and are efficient in sampling compounds from the lumen and present it to other immune cells (antigen presenting). Research has been performed with reporter cell line of macrophages in which specifically the NFkB pathway was used as a read-out. Besides, research was performed based on primary cell, isolated from blood of healthy donors. Especially based on these primary macrophages interesting results were obtained in which it was shown that these macrophages that classically are divided in tolerogenic (M2) or pro-inflammatory (M1) type of macrophages now, after NPS exposure, show a new phenotype (see figure 11). Follow-up experiments demonstrated the presence of interesting cocktails of chemokines produced by the cells and excreted in the medium. In real life this could indicate that macrophages in the intestinal mucosa that are exposed to specific NPSs could start to produce chemokines to attract other immune cells that can support in direct killing of unwanted pathogens. This is in line with the observations on intestinal epithelial cells, but in a more pronounced way and again indicates that NPSs are seen as a kind of warning signals to make the immune system more alert. The research will be materialised in several articles and a PhD-thesis for the PhD Y. Tang that participated in this FibeBiotics project.
Besides these assays several other in vitro assays have been developed and used to characterise the potential bioactivity of NPSs. A complete overview on the in vitro tests and results from the NPS exposure experiments can be found in the (restrictively disseminated) deliverable reports but some of the results will be included in scientific papers.
Effect of NPS towards the gut barrier function ex vivo
Experiments with human epithelial cells grown in vitro in the lab showed that these cells can respond to the NPSs. However these cells cultured in the lab are only one type of cell (enterocytes) while the intestinal epithelial is a mixture of different cross-talking cell types with different functions. Therefore research was started to expose NPS to biopsies taken from healthy subjects from different location in the intestine to study the effect of NPS exposure on gut barrier function. As these biopsies normally have a high integrity and not allow compounds passing the epithelial layer, the exposure of NPSs was combined with a stressor that temporarily reduces the integrety. Subsequently the effect was studied by fluorescent molecules in order to study the passage through the intestinal biopsy and how well the cells could resist the stress. This research was performed under an approved Ethical protocol. The project encountered several drawbacks due to variable electric supply, breakdown of essential parts of the Ussing Chamber equipment, unrepeatable effect of a stressor which forced to change towards another stressor (to which some of the volunteers did not show any response) and bruising of biopsy material upon transport. The biopsies were also found not to survive exposures more then 1.5-2 hours, which hampered reliable gene expression analysis of the tissue and analysis of compounds excreted in the media that were all very low due to this short incubation time relative to the large volume of surrounded media.
However, after selection of subjects that responded to the stressor, the effect of NPS on biopsy barrier function could be studied. Both young and healthy subjects were included in the study as well as senior participants with gastrointestinal symptoms of constipation and/or diarrhoea. This resulted in the observation that one of the NPSs studied gave a statistically significant reduction in paracellular permeability and transcellular permeability compared to intervention with only the stressor. These results indicate that some of the NPSs that were tested might contribute to maintain homeostasis of the gut barrier function (figure 12).
Effect on NPS towards the gut barrier function in human volunteers
The human intestine has a complex and crucial function as it needs to take up nutrients from our food and at the same form an efficient barrier to the commensal bacteria that we host in our intestinal system and the potential pathogenic and immune potentiating factors they produce. Several researchers already launched the ‘leaky gut’ hypothesis indicating that the barrier function of the gut is reduced due to life style habits, including western diet. Due to this reduced barrier function our immune system would be chronically overloaded with risk factors to it is forced to respond, finally leading to an exhausted immune system that fails to work properly. This would result in increased sensitivity to pathogens and stresses. The validity of this hypothesis is still in debate and supportive evidence is still the research subject of several groups. One of the proposed strategies to study the differences in gut barrier function in volunteers is passage of sugar probes. Different sugars are taken up differently by the gut epithelium and at different regions of the intestinal system. This resulted in a 5 sugar test that can be administered orally to subjects after which the urine is analysed 5 and 24 hours after the intake. The amount of the sugars found in these urine samples can be seen as an indication of the barrier and uptake function of the intestine in different regions (duodenum, small intestine, colon). However, under non stress condition these barrier and uptake functions between healthy individuals are hardly different. This sugar test, can be combined with a stressor (like used for the biopsies studied reported earlier) that can reduce the barrier function of the intestine. In this format, the test can differentiate between persons on the basis how they cope with the stress provided. Several trials have been performed in which the reproducibility of a stressor was tested resulting in the selection of indomethacin as stressor.
The FibeBiotics in vivo barrier trial, a double-blinded placebo-controlled randomised parallel study, aimed to investigate whether 6 weeks of oral intake of two different NPSs (compared to a placebo) could improve intestinal barrier function against this indomethacin-induced barrier disruption. 60 healthy individuals representing a general population of elderly people (65 years old and above) were recruited and randomly divided over the three groups. The results from the 6-week intervention study with daily oral intake of the NPSs did not show any improvement on indomethacin-induced hyperpermeability nor on baseline permeability. Although indomethacin significantly increases small intestinal and colonic permeability at baseline, the interventions did not show any statistically significant improvement (figure 13). All the results have to be studied in more detail and will be used to write one or more publications which will also make part of the thesis of the involved PhD student at Orebro University, Sweden.
Effect of NPS towards vaccination efficacy based on human pilot trial
The heart of the FibeBiotics project was the human pilot trial in which the effect of NPS intervention was evaluated making use of vaccination efficacy, an EFSA supported biomarker for the immune system. The trial was designed as a pilot trial as at that stage we wanted to explore more biomarkers and wanted to see how other immune related biomarkers might correlate with the primary biomarker related to the vaccination. At that stage, we also wanted to compare the different selected NPSs within this project to be able to compare results obtained in vivo with those that are obtained in the different in vitro models in order to find potential underlying mechanisms and strategies to predict the potential outcome of an in vivo trial based on in vitro selection and product design. An ethical approval was requested and provided by the responsible authorities.
The trial was performed with healthy volunteers, in the range of 50 to 78,5 years and with a mean age of 66 +/- 6, gender balanced that received a standard seasonal flu vaccination. The trial was highly complicated in which many persons had to be recruited, health status had to be recorded, samples have been taken (of which some had to be processed directly) and biomarkers had to be measured. An schematic outline of the pilot trial can be found in figure 14. The trial was very well organised by Clinical Research Centre Kiel, Germany and resulted in large data sets on which extensive analyses have been performed, and still can be performed when new hypotheses on certain correlation of biomarkers are hypothesized. Unfortunately, publications have not been completed yet.
The trial can actually be seen as three trials. The first part includes the analyses that were performed before the start of the intervention and collected at visit 0 and visit 1. This part of the trial is called the association study as we can relate the many measurements performed and biomarkers determined to each other and to the individual parameters like age, gender, BMI, waist circumference etc. As the general population showed BMI of 27.2 which have been classified as over-weighted (although other see it as the current normal/standard) it has been possible to find variations in the parameters for Metabolic Syndrome, liver sclerosis, endothelial function etc. and how they correlate with other biomarkers determined. These are currently incorporated in a manuscript that presents new knowledge on health parameters and how these can be related.
The second part of the study can be regarded as the main study in which the effect of the NPS intervention is studied on the primary parameter, vaccination efficacy. The standard seasonal flu vaccine is a combined vaccination of three different vaccines for which antibody titres have been determined separately. This resulted in HI titres at four different time points and many type of analysis as results can be approached as HI titter (e.g. V3 and V4), delta or fold difference between (V1 or V2 verses V3 or V4), geometric mean fold increase, (increase in) protection rate and seroconversion rate (subject with a four-fold increase in HI titres). Besides, calculated for the individual vaccines, they can also be calculated for the combined titres. To make it even more complex, the data can or should be analysed by different statistical tests depending on the distribution of data points. This resulted in a large summarising table, from which it was concluded that none of the NPSs showed a significant effect on these parameters (as also was expected based on power calculation) and that one or more showed promising trend in increasing the HI titres that might be caused by the intervention (see as an example of these analysis figure 15).
In addition, microneutralizing titres have been determined (another technique used by the to determine that people have serum cross-reactive antibodies to the influenza viruses used in the vaccine which is more specific than HI titres) and common cold/flu-like symptoms were scored. Some of the NPSs performed promisingly in reducing these symptoms although the incidences were very low (as can be imagined for 40 persons in a 5 weeks period), making it difficult to draw biologically relevant conclusions. Based on these results it had to be concluded that: 1) some of the NPS showed positive trends, power calculations indicating that when the arms in the trial will raise from 40 to ~170 results could become significant; 2) that, although a trend is seen for some of the NPSs, it will be hard to correlate the effects in vivo to in vitro, except that the one that looks most promising would be expected also to show most promising effect in the in vitro tests; 3) that the flu vaccination model might not be the best model to analyse these effects as people have built up memory (as often same vaccines have been given previously) making it uncertain whether primary vaccination responses are analysed or whether also secondary memory-driven responses are included; 4) the results in reducing common cold like symptoms or disease episodes were very interesting and potentially also more appealing as supportive evidence to consumers than vaccination efficacy which will be harder to understand.
The third part of the study was dedicated to identify potential other effects that might have a link to the observed effect in vaccination efficacy of the NPS interventions. Cellular parameters were determined by analysing the blood of the volunteers for cytokines (small molecules in the blood that help to communicate between cells and therefore important for metabolic function). These cytokines were analysed in unstimulated blood but also after stimulation with LPS, conA and vaccine in order to analyse how people respond to these challenges and whether people within a specific NPS intervention group might respond differently. Despite the extensive data set, no major effects have been discovered yet. The bioinformatics experts will probably re-analyse these data when new hypothesis are generated and tested based on this dataset.
Within this third part of the study also the microbiota analysis can be grouped. Effect of NPS on microbiota can be expected for some of the NPS that were given in high amounts in the intervention. These expectations were also fed by the in vitro analysis which shown effect towards increased Bifidobacterium or Lactobacilli in some groups. Some of the samples were analysed by 16S rRNA seq analysis and all samples analysed by the GA-map technology as explained above. The results obtained by both technologies have been compared and the effect of the NPS intervention studied, also in connection to SCFA that were analysed. These results will also be linked together in a publication.
A pivotal trial to study effect towards episodes of common cold
Based on these results, we had to design the pivotal trial in which we hope to establish for one of the NPSs that, by increasing the power by including more people in the arms of a trial, results will be more (statistical) significant and later can be used as part of an EFSA claim submission.
After a long discussion and consulting the Scientific Advisory Board including a person with experience in EFSA, setting up the guidelines and involved in dossier evaluation, it was decided to change the strategy from vaccination efficacy to common cold/upper respiratory tract infection symptoms scored as primary parameter to study the effect of one of most promising NPSs as concluded from the pilot trial.
The pivotal study was designed, ethical approval obtained, NPS and placebo were produced and packed and recruitment was started. As we decided to focus on common-cold/upper respiratory tract infection related symptoms we had to perform the trial during the winter season. In order to have a high chance that people would experience these symptoms, the intervention period was extended from the 5 weeks in the pilot trial to 12 weeks in this pivotal trial. To work according to the guidelines of EFSA, disease symptoms scored by the participants were also checked by a physician. A full overview of the study design can be found in figure 16.
Inclusion criteria of the trial were men and postmenopausal women aged ≥ 50 years that were willing to abstain from food and supplements containing probiotics, prebiotic supplements and other fermented products. Besides, they should be willing to sustain a low dietary fibre diet for 14 weeks and willing to obtain the influenza vaccination season 2014/2015 (but did not receive it yet). Like all studies, we also had a list of exclusion criteria.
The symptoms scores were recorded (=number of subjects affected by at least one episode) according to self-reporting (questionnaire based on Jackson et al. 1958) and physicians diagnosis and then calculated according to the best validated and blinded method (so people that do the statistical analysis only know whether people had belong to group A or B and not whether they are in the NPS or placebo group).
Many different analyses have been performed on the dataset, like calculations based on occurrence, incidences, severity of the symptoms, duration of the episodes and for type of symptoms separately and combined. An example of such analysis can be found in figure 17 which illustrates that both groups shows some differences on some of the scorings. However, when statistical analysis were performed on these data, none of the parameters showed to be significantly different. At this stage the trial is still blinded until the pilot trial is published. Upon this publication, this pivotal trial will be unblinded and meta-analysis performed to see when both the pilot trial and the pivotal trial are combined, for this NPS and placebo control arm, whether more promising effects can be concluded.
It can be concluded that the NPS tested in both the pilot and the pivotal trial showed: 1) an increase in GMFI of HI titres against virus A H1N1 and combined virus strains; 2) an increase in Δ seroprotection rate against A H1N1 (23%), B (17%), H3N2 (6%) and combined virus strains; 3) an increase in seroconversion rate against A H1N1; 4) a higher local vaccination response; 5) a lower incidence/occurrence of common cold (1 vs. 5 according to Predy; 1 vs. 8 according to Jackson) and 6) a higher IFNγ (unstimulated) ex vivo but that these effects are not statistically significant. For the pivotal trial, no conclusions can be drawn yet, as the experiment is still blinded.
Database and warehousing for now and the future
Data management and data analysis get more and more important, also in research. Many researchers but also industrial partners still work with separate files, stored in a hardly organised way in project folders which can get lost or are very hard to interpret by others as identifiers for the different columns and parameters are missing. The idea of the FibeBiotics project was to select/develop a user-friendly software system, highly flexible for the data sets that are developed and implemented in this kind of research project. This could then have been a good and easy experience for all partners, and the industrial partners that are regularly updated about such a tool to encourage future groups to work which such databases. An appealing property of databases (when working properly) is that data are well organised with links to the protocols used, processing steps of data are documented and data can be combined quickly to perform new types of correlation studies that else would take a lot of time to subtract from individual excel files with different formats.
This part of the project was a major challenge as the research performed in the whole FibeBiotics project was very diverse and so were the data generated: biochemical analysis, NPS description and samples from different batches and different purity, different cell lines and read out (microarray analysis, cytokine analysis, NO, qPCR etc.) and human trials. These human trials are very complicated, with general information (which should be coded to protect privacy), food intake data, questionnaire data, and the analyses done on all samples. To give an example, one blood sample was used for HI titres (3 individual vaccines), MN titres, cytokine (each sample had 4 different treatments and then 7 different cytokines were measured) but also many other health related blood values like HDL, triglycerides, insulin, liver function markers, LBP etc. Some of the analysis are performed on all samples, some only on a limited number of visits or measured only at the start. So besides this being a logistic challenge to do all these analyses, it is also a data management challenge to combine them all together in a database in such a way that other researchers that might want to work with the dataset in the future still understand the data and how they are connected.
A data warehousing system, developed and used in the large EU framework NUGO was selected and installed by the partner responsible for this task and made available for the partners of the FibeBiotics consortium. Due to misconnection between software system and regular updates of programmes and links used to internet pages, many maintenance actions were needed to make it work and to make sure that it will keep working properly in the future. As in this project different partners will have different interest and responsibilities, especially as it also involved data from a human trial, for which the possibilities for data sharing are restricted by law, specially strategies were implemented concerning ownership of data and disclaimers on how to ask permission for use of data from the database and how to acknowledge persons and groups which made the data available.
Then another challenge was on ontologies, how to label data, how to upload them etc. A first attempt to have all partners upload data themselves somewhat failed due to uncertainties about field names, data sets that differed from the standard formatting and the unfriendly-user interface and the inflexibility of the software. An instruction movie prepared by the responsible partner helped to instruct partners better but still some confusion remained.
Starting from a new installed version of the database, it was then decided to re-upload all data by the ICT partner in a much more organised way. After a short explanation of the dataset, by the person who generated the data, and how it was provided and what each field in the file meant, the responsible partner, knowing the programme best, uploaded the data in the database (see Figure 18, Start screen of the FibeBiotics database).
An overview of the data currently available in the database is in one of the project's deliverables. In addition, at the end of the project we have secured continuation of the database for at least another two years but also the data will be collected in the original files and stored at the coordinator for further use.
Mathematical strategies, can it support to explore the data better
As data become more and more complex also data analysis, statistical tools and mathematical models increase in complexity but also present opportunities. In most human clinical research, the number and type of statistical tests are often dictated by the type of data available and the type of comparison one wishes to make. However in research, when you want to build new hypotheses on correlation, complex interactions between parameters and classifying data to increase understanding, many new tools, models and strategies can be explored to make more out of your data. The FibeBiotics partner, University of Bologna, is strong in statistics and mathematical modelling of the data and has supported the data analysis that we generated and that was provided by other groups. Some of this knowledge was more theoretic or tested on other dataset than based on the FibeBiotics project, but which helped to generate new strategies. This led to several publications of the group of Bologna in which FibeBiotics was acknowledged. Presentations were given by the Bologna group during the project progress meetings and the Industrial Platform meetings to make people more aware of the possibilities.
One of the research examples that was based on FibeBiotics data is the research performed on microbiota analysis. In the intestinal system, more than 1000 bacterial specifies can be present and distinguished by 16S rRNA sequence. Each person will have another composition of the microbiota. This makes the analysis of an intervention of a whole group of persons very challenging, as it can depend very much on your starting situation. Based on data generated on the faecal samples from the human pilot trial and the 16S rRNA sequences analysed by Teagasc the Bologna group thought about a new strategy to analyse these data. They modelled humans as a steady-state process in which you have an inlet of new bacteria (entering orally), reproduction of bacterial species due to available food, natural death of bacteria and an outlet of bacteria as can be found in the faeces: the samples that were analysed. This type of analyses could generate new ideas on what is really happing in the gut as a result of changing dietary interventions. This has led to an interesting publication (Sala C et al (2016) BMC Bioinformatics. 20;17 Suppl 2:16) from which figure 19 is taken.
The EFSA guidelines and practise
The food industry is confronted with the EFSA based legislation about what is allowed to declare on the package of a food product, how much can/should be in products of a certain bioactive ingredient, what can be communicated to consumers on content, nutritional value and what on health effects. As this type of EFSA regulations have not been settled yet, it is of interest to the food industry what EFSA requires in order to authorise a certain content or health claim for a product. Therefore, the FibeBiotics kept track on EFSA guidelines, EFSA opinions, submitted dossiers in the area of fibres and in the area of gut and immune function. These evaluations have been reported to the FibeBiotics consortium and to the Industrial Platform as they are important to keep in mind to perform research when the aim is to use the data to submit a dossier to substantiate a health claim.
In the domain of immune support, the food compounds that have authorised claims are mainly established vitamins and minerals which have shown effect on the immune system mainly (or almost only) when deficiencies and diseases occur in people due to malnutrition. So results from a vaccination trial like used in this FibeBiotics project (as one of the only markers that is currently broad accepted as an indicator for the immune system) were not the basis for those health claims (although some show effect on vaccination efficacy based on populations in developing countries).
Dietary fibres like the NPSs have been recognised for other health effects than immune effects, notably cholesterol lowering, reduction of post-prandial glycaemic response, increasing faecal bulk, reducing transit time, maintain normal bowel function and reduction of body weight (see table 2). These parameters are easier to study than immune related responses as for cholesterol a widely accepted biomarker is available (HDL/LDL) and the others are relatively easy to quantify. The BGO as used in FibeBiotics have already been authorised with several of these health claims. Other of these health claims can be seen as ‘low hanging fruits’. As an example, it can be expected that all non-digestive polysaccharide will contribute to a reduction in post-prandial glycaemic response and that due to water binding capacity, as seen for many of the fibres, they also will have a positive contribution of increase of faecal bulk, stool frequency and stool consistence.
Due to the lack of accepted biomarkers other than vaccination and the complexity of the vaccination response due to memory reaction and the variation between individuals, we expect that it will still take some time before the first product with an immune claim is accepted. In 2012, at the start of the FibeBiotics project, we had expected that certain probiotic strains that have shown effects on the immune system would also have a dossier on immune health support authorised by now; which is not the case. This is neither the case for probiotics: although some studies show clear effects on the duration of common cold episodes, no health claim has been approved yet.
In the new general guidelines and those specifically for gut and immune health, we see that the EFSA allows some room to base research not only on healthy people (for which it is very difficult to demonstrate health effects when no risk factors are general accepted) but to also include those who have mild forms of a disease when it can be made acceptable that the product and established effect can also be active in healthy persons to maintain health. We have to await how this field will develop for NPS and immune support in the future.
Consumers, the unpredictable factor
The final goal of the FibeBiotics is that healthy food products are developed, based on scientific evidence and that consumers will buy and consume them in order to support their health. The food ingredient, product, information provided and health effect should, therefore, be of interest to consumers. In order to investigate the knowledge level on fibres and current opinion of consumers, we developed a questionnaire which was used as market study among consumers in three European countries. The full result of the market study can be found as one of the deliverables. Here, we only will summarize some of the main results.
The consumer survey was performed in men and women in three countries: Germany (N=1263), France (N=1228) and The Netherlands (N=1234). All three consumer samples were representative on the variables gender, age, education level, and social economic status. The questionnaire was translated into the three applicable languages. The consumer survey was performed in the first months of 2014, since it was decided that it would be preferred to execute the survey in the ‘flu/common cold period’ due to the specific questions on potential health effects of fibres.
Health impact of food becomes increasingly important with age. Most respondents perceived eating fresh and having a well-balanced diet as eating healthy. In general, healthy foods were considered natural, tasty, enjoyable and easy accessible. About half of the consumers did not use non-prescribed health improving foods. However, if one wanted to improve health, vitamins were mentioned as the most consumed, followed by calcium and fibres.
In general, most important aspects for consumers when buying food were price and taste. But for older consumers (55+) health impact was the most important factor. Some differences among countries were found: taste was found more important in Germany, whereas health impact was a more important aspect in France.
Overall, consumers were (very much) interested in consuming fibres. Consumers perceived bread and rice as natural carriers of fibre. Enriched foods were perceived as accessible, enjoyable and good tasting, but also as expensive and unnatural. Almost half of all respondents indicated not to have a strong opinion. This might indicate little knowledge and experience regarding enriched foods.
Although consumers were generally interested in products which help them to lose weight or enhance well-being and slow down the ageing process, health benefits by fibre enriched products, like improvement of the immune system and reduction of intestinal problems, were also appealing health benefits (figure 20). Fibres and whole grain products are most likely to be consumed in the case of stomach and digestion problems.
German (69%) and French (61%) consumers seemed to be more interested than the Dutch (50%) in consuming products that benefit the immune system and will reduce intestinal problems. In addition, middle aged adults (30-55 years) were most interested in enriched food products.
Consumers from all three countries preferred fibres in cereal based products, like breakfast cereals and a bar. Baked goods are also a preferred option in Germany and France, however surprisingly not in The Netherlands (figure 21). Functional (healthy) foods are preferably bought in drugstores/health stores (Germany & The Netherlands) or in pharmacies (France).
Most consumers paid little attention to nutritive information and hardly do any research on their own in order to find additional information regarding nutritional values. Consumers in general are unfamiliar with functional foods. Thus, providing more knowledge about and giving the opportunity to get experience with enriched foods is needed.
Consumers based their decision whether or not to believe information about food and health mostly on their own judgment. Medical sources and scientific research were considered less convincing (figure 22).
Especially, this last result is an eye opener for scientists such as within the FibeBiotics project, who have another opinion; research should be leading in how we are able to judge the potential health effect of our food. As this counts for the whole food and health domain worldwide, all scientists should try to (re)gain the trust of consumers. In food and health, also many scientists combat each other with different opinions, and the media are as well involved in the controversy as they like to overstate the research results to seek the attention of people. A more balanced discussion and reliable performance of scientists might help to improve this.
Conclusions from the Consumer Survey
Overall, consumers are interested in consuming fibre (enriched) products. Differences among countries were found. In order to position functional foods in a rapidly growing market, it is crucial to use an effective strategy. To position products successfully into the consumer market the following recommendations are made:
• Taste and price are the most important aspects to consumers and, therefore, must meet consumers’ expectations.
• The enrichment of a product needs to be in line with natural characteristics, e.g. fibres in breakfast cereals.
• Cereal-based products are preferred as fibre carrier.
• Drugstores/health stores or pharmacies are a right product outlet for functional foods.
• The product could be best targeted towards middle aged to older consumer segment. Older adults are more focused on healthy products and healthy ageing.
• Perceived health benefits are crucial. Though fibres are mostly associated with digestive health, consumers would buy fibre products if they have immune benefits.
In line with literature, this study also has shown that taste is the most important aspect of food products. Therefore, sensory research before positioning in the consumer market is recommended.
Potential Impact:
The original objective of the FP7 Work Programme 2011 was to improve health and well-being of European consumers by increased consumption of health products and at the same time to increase the strength of European SMEs. Moreover, we have tried to support the food industry and further help to study the EFSA approach on biomarkers to study immune support.
The impact for the society
Cold viruses are the most common infectious agents humans encounter. Cold viruses result in significant costs to the economy, considering lost workdays and decreased school attendance. Cold episodes range from 2-10 colds per year, depending on age and exposure. Common colds resulted in more than 100 million physician visits annually, at a total cost of $7.7 billion. At least one-third of these patients received an antibiotic, even though antibiotics have no effect on viral infections, not only adding to the cost but also contributing to the threat of antibiotic resistance. Research that integrated all costs (including absence of employees with sick children at home) showed that common colds and flu are responsible for a cost of $ 40 billion annually. Finding alternative strategies like the NPSs studied here, could reduce these incidences and episodes and therefore would have a high impact on the community.
Influenza is an acute respiratory illness caused primarily by the influenza virus (serotypes A and B). It occurs worldwide and is responsible for considerable morbidity and mortality. Serious illness and death rates are highest among the elderly, young children under the age of two, and those with medical conditions, placing them at increased risk for influenza complications. Because of the potential severity and epidemic/pandemic threat in most countries it is practiced to vaccinate elderly persons and people at high risk like health care workers. The idea is that immunisations will prevent hospitalisation and/or death; reduce influenza-related respiratory illnesses; and decrease physician visits among all age groups. However, it is generally known that the effectiveness of these vaccines declines with increasing age which explains why the preventive effect is only observed in persons below 80 years of age. Some studies even show marginal significant effects between vaccinated and non-vaccinated persons. At the same time, it was concluded that the burden of influenza in the elderly is substantial and even a modest protection of a 30% reduction in influenza-related hospital admissions would be beneficial. The pilot trial showed that NPS intervention group 4 shown a delta seroprotection rate increase of 26% for influence A H1N1 and also for the other viral strain strains. Although this difference was not significant based on a group of 40 persons, when this line would be true for the total vaccinated population, this NPS would help to save many lifes and would reduce costs for medicines and health care drastically. As no vaccination analyses have been performed in the pivotal trial we will never know whether the results found in the pilot trial could be repeated and would indeed point to an significant increase in the influenza protection level of a population. Therefore, more follow-up studies are needed to study the reduction of cold incidences and reduced duration of the illness by NPS. Based on these outcomes the full impact on the quality of life, life expectation and health care costs, can then be calculated.
It is important to mention that the immune system, which is the target of the proposal, has a much higher impact via many other immune-related diseases. As these other diseases cannot be studied within the course of a short intervention study (as will be performed in the present EU project) these other diseases have not been mentioned before. However an overload of publications has shown that an overall improved immune status can also lead to a reduced risk on cancer, Alzheimer, Crohn’s disease, arthritis, diabetes etc. Based on in vitro research, new hypotheses have been generated on how NPS can contribute to certain immune related mechanisms. We therefore hope to study these NPS in order to support in reducing incidences of immune related diseases or reduce recurrence of diseases when treated. We think that these can be efficiently combined with current treatments and make this more efficient. Time will tell, and future research will learn.
The impact for the research partners
The goal of the FP7 framework, and particular also this call in the KBBE domain, aimed to boost also SMEs as accelerators of our economic growth. In our project, we included SMEs, both food ingredient developers/supplies and those more towards technology services in the food and health area. Both had much benefit from this project.
The food ingredient suppliers (BioActor, Swedish Oat Fibre, Immitec) have gathered knowledge on the biochemical composition, how the NPSs behave when applied in food products, have gained knowledge on bioactive effects towards the immune system and towards gut fermentation and effects on microbiota, obtained in vivo data for which some of the NPS showed somewhat more potent in inducing effects than others etc. But also when no promising effects are found, these results are still of interest indicating that either the product formulation chosen was not bioactive or that immune health is not the most likely direction to continue and the health effects could better be studied towards other health aspects like bowel function or reduction of pathogens. All these research directions could not be addressed in this project. The research network as a whole and the interaction with the industry organised by the FibeBiotics consortium was also seen as important for the SMEs and have improved their position. The same counts for Winclove, the company mainly developing probiotics, which has been able to study the interaction of probiotics with NPSs in an early stage and could scan this area for interesting new developments that could help to strengthen their portfolio either by using the NPSs studied or by applying the research tools developed to use for substantiation of probiotics which are also studied for their immune support.
The technology and service providing SMEs have also greatly benefited from the project. CRC Kiel performed and organised the human trial perfectly and when results are published can help to establish a good market position in order to gain new assignments. Genetic Analysis has been able to further optimise the GA-map technology and made it dedicated towards European elderly which might help also in the future to develop diagnostic test to diagnose abnormal microbiota composition in relation to diseases. This project generated many samples which now can be used to link microbiota and, therefore, the GA-map technology to health status both at baseline and as a result of the intervention. ProDigest has already had a well-established position in the area of gut intestinal simulation as the SHIME model is, together with the TIM model at TNO, one of the best in full simulation of the intestinal tract. This model has been further developed in this project and experiments formed have shown good correlation with effects found in vivo, further supporting the potential of this model to replace animal and human trials. The aim to develop this SHIME model further by including in line exposures to human cells did not work out yet and Prodigest will invest in that further. Besides SHIME experiments, the group has been able to gain more knowledge on cell culture research that could be performed off line with samples derived from the SHIME. R-consultancy was involved in building the data base and accompanied software to simplify uploading, warehousing and subtraction of data. Due to problems with compatibility of systems and limited flexibility of the system, the database was less optimal than expected and also none of the partners has implemented the database structure for their own in-company warehousing. This might be due to the small size of some of the companies or the availability to other commercial software system developers that jump into this niche.
All research partners, both universities and institutes have benefited from this project to strengthen their knowledge and develop new research tools and strategies in the area of gut immune effects.
The results obtained for the different NPS compounds within this consortium can be used as a reference to compare and characterise effect of other NPSs in order to develop other functional food products directed towards gut and immune function. The appeared and to-appear publications will help them in exposing their knowledge and to attract new research capacity to continue their research.
Dissemination towards the food industry and the FibeBiotics Industrial Platform
EU funding should support not only a limited number of partners involved in a project but to our opinion result in useful tools, knowledge and strategies which can be applied broadly by the community. In this food and health related research, mainly the food industry should be the target for dissemination. As not all food companies could become project partner, we established an Industrial Platform (InP) for which we invited all industrial stakeholders interested to be informed on the progress of this project. This would generate short links with groups which can answer questions from the industry directly, will support networking with other companies with an interest in fibres and connect ingredient suppliers with end users.
The FibeBiotics Industrial Platform was established to ensure that knowledge and new developed technologies/research tool, once proven to be suitable for application, are indeed used to the benefit of industry. The Platform was continuously open for new interested members. At the end of the project, the Industrial Platform included 42 members which all signed a participation document and indicated their specific interests.
For the Industrial Plaform members, we established a members-only website, which contained among others the information on the project’s progress and (related) publications. The Platform members also received Newsletters on a regular basis and the FibeBiotics consoritum organised annual Industrial Platform meetings with presentations from the FibeBiotics partners and also from the industry themselves to exchange ideas on in-house research performed by the companies, their products portfolio and their market strategies. The InP meetings have been organised in:
• Ghent, Belgium (2012), including a visit to the facilities of partner ProDigest;
• Bologna, Italy (2013) alongside the EFFoST Annual Meeting: Bio-based technologies in the context of European food innovation;
• Cork, Ireland (2014), including a visit to cheese factory of Teagasc;
• Paris, France (2015), including a visit to the lab facilities of Institute Pasteur;
• Wageningen, The Netherlands (2016), including a visit to the lab facilities of Wageningen University & Research.
These events were very well visited and appreciated by the InP members.
The impact for the food industry could be summarised as that FibeBiotics has brought them in contact with new strategies and research tools, tested human trial designs supported by EFSA which probably has helped them to consider whether to pursue this direction, or learn from this and change future plans. They could reflect and improve in-house research with the new knowledge generated from the FibeBiotics project at a very early stage (as many major results still need to be published but they already have been shared with the industry that now can benefit from this new insight knowledge).
For some a pity and for some a relief, but we have to realise that we have not been able to come closer towards an immune-related health claim for one or more NPSs. If we had results that would make a proper EFSA dossier, this would have been authorised and could have been a major breakthrough. Such an authorised claim and the way the dossier content is generated, would have made it potentially easier - also for other NPSs - to obtain an EFSA-approved health claim, in case of similar results by similar human trials. The pilot trial and the pivotal trial could be used as supportive evidence in the future but, as effects are not statistically significant, it will not be sufficient. We also have not been able to identify other immune related biomarkers which might become useful to substantiate effect in the future. However, the biomarker for barrier function might still be a valid approach, but we are not sure whether food compounds really can counteract the effect of a stressor and how biologically relevant this is for intestinal function.
However, many research tools, knowledge (such as those related to microbiota and microbiota related metabolites) can be expected to become very useful in the future when more relations to health impact are known.
Impact on the European policy on immune claims
Digging into the area of vaccination efficacy as one of the only, by EFSA, accepted models to study potential immune supportive effect led to the realisation that very little evidence is available on adjuvants in a healthy population. Almost all of the authorised health claims in this area are based on deficiencies of a certain nutrient in populations (in developing countries) that really suffered from a deficiency. That is how some vitamins and minerals reached their immune claim; these were not achieved by showing adjuvant effects on vaccination efficacy as the result of extra intake in a healthy (elderly) population. And even when studies were performed on populations with malnutrition, then also the supplementation of the nutrient (e.g. zinc) did not always show this increase in vaccination efficacy.
Some studies have shown positive effects of a food intervention on vaccination efficacy but these have been hardly repeated in order to analyse how reliably the effects can be monitored. One possibility might be that the type of vaccination model used is of main importance. We used a flu vaccination model for which effect will be influenced by vaccination history, encountered flu infections, and health status. So we concluded that it would be wise to use vaccination models that a person did not have before (nor the vaccination nor the disease which also can boost memory) or use suboptimal doses of a vaccination. In order to detect a positive effect of an intervention you need a window to quantify the improvement of antibody titers. When using very optimised vaccinations the vaccine treatment itself might induce often towards its maximum and does not leave room for improvement by an intervention. However it is currently not known how the EFSA will judge studies in which suboptimal vaccination models are used> The hopes are that this will be accepted and evaluated in next guidelines.
With the new updated guidelines, the EFSA opened up new potential valid strategies as they concluded: ”studies conducted in subjects with a disease may be used to substantiate function claims for the general population or subgroups thereof (without the disease) as long as the effect of the food/constituent on the body function which is named in the claim is expected to occur in subjects without the disease and a rationale is given for such expectations”. This might be a strategy to follow in the future to study gut and immune supportive effects, although the interpretation of ‘expected to occur in subjects without the disease’ can still be explained in different ways and therefore forms a narrow basis for companies to target their human trial designs. When this rule has been applied and implemented in real evaluations it will become more clear what the possibilities are to substantiate health effect making use of a somewhat diseased or medically compromised population.
All current trial designs and statistical method in the area of health claims are still completely leaning on the strategies used for medicines based on diseased subjects. However, effects of pharmaceuticals can be very individually determined, making that within the population there are responders and non-responders. Not without a purpose, medical research is focussing a lot on strategies for personalised medicatio. When treatments have small effects (low specific activity), as can be the case for food that might subtly correct homeostasis, effects based on the average values will require very large sample sizes. This will require expensive trials that might be possible for drug research (due to recovery of investment) but this will be much more difficult in the food area. When it is generally accepted that a randomly choosen population included in a trial will consist of responders and non-responders new concepts are needed to evaluate the results and the conclusions on benefit to the general population (or to responders within the population). We hope that EFSA guidelines in the future will take these considerations into account so that food intervention studies that show positive effect for specific individuals in the population but does not harm the rest of the population still can be used to substantiate (personal) health effect.
List of Websites:
www.fibebiotics.eu
Dr. Jurriaan Mes, project coordinator
Email: jurriaan.mes@wur.nl
Tel: +31 317 48 11 74
The goal of the FP7 KBBE EU FibeBiotics project was to support the development of functional food ingredients and products that are beneficial for the human gut and immune system and therefore of crucial importance for quality of life. The project studied the effects of dietary fibres, more specifically non-digestible polysaccharides (NPS) with a sugar backbone of 10 units and more. Previous research, mainly based on animal trials, has shown potential health effects of the NPSs towards the gut and immune system. In this project the health effects of the NPSs have been focused around enhancing immune defence against pathogens, increasing the gut barrier function and the reduction of infectious diseases such as common cold and influenza of elderly. The project wanted to make use of European Food Safety Authority (EFSA) supported biomarkers that would support immune function claims and underpin the mechanism involved. The studied mechanisms involved in the health effect of NPSs were the innate and adaptive immune system and the possible involvement of the microbiota and microbiota-mediated products such as short chain fatty acids which also have shown to have immune modulating properties. To achieve this goal, new and existing NPSs have been studied for their health effects in a systematic way by developing a toolbox of dedicated assays and models that can now be used by research entities and industry to study bioactivity of food ingredients with a similar health focus. The project has:
1) performed biochemical analyses to characterise the NPSs and study effects of fermentation in the gut and the fate of NPS when formulating final food products,
2) developed standardized in vitro screening methods to compare different NPS for their bioactivity towards gut and immune cells and their effects on microbiota and microbiota derived compounds,
3) performed dedicated in vivo and ex vivo analyses to study the effect of NPSs on gut barrier function,
4) studied the effect of a 5 weeks intervention of NPS in humans based on a pilot trial (6 arms, with 5 different NPSs and a placebo) and of a 3 months intervention in a pivotal trial (2 arms, 1 NPS and 1 placebo) with vaccination and common cold as outcome parameters,
5) developed a database and explored different mathematical methods for data analyses,
6) analysed the consumer opinion on fibres and immune support and also kept track of EFSA guidelines.
Dissemination has mainly been realised via an established Industrial Platform, to which 42 large enterprises and SMEs subscribed and that attended the meetings organised by the EU FibeBiotics consortium. Several scientific papers have been realised but still many are in the pipeline.
Project Context and Objectives:
Project Context
Dietary fibres are, by all food-related-authorities, recognised for their contribution to human health. This positive effect of fibres has mainly been postulated on the basis of epidemiological studies and, to a lesser extent, intervention studies. This has led to the recommendation of 30-40g fibre intake per day depending on the countries responsible for the advice. The general consumer, however, hardly reaches an intake of 20g dietary fibre per day. So, the awareness among consumers is not high or the willingness to act according to food council advices is not high or people do not have knowledge about which food products contribute to a high fibre intake.
Advising on the intake of dietary fibre also brings some uncertainties, as there are thousands of different dietary fibres in our food. This warrants the question whether they all contribute to a similar extend to maintenance of health. Controlled intervention trials have shown that some dietary fibres can reduce cholesterol levels while other fibres failed to show these effects. Apparently, the type of fibre can have different bioactive effects just like these fibres show differences techno-functional properties, such as water binding capacity, or conferring solubility and stability etc. to food products.
Dietary fibres include non-starch polysaccharides, resistant starch, resistant oligosaccharides (with three or more monomeric units), and other non-digestible polysaccharides. The terms “soluble” and “insoluble” have been used to classify dietary fibres in an attempt to link different physical-chemical properties of fibres to different physiological effects. As a consequence of the complexity of the large group of dietary fibre, public food databases (e.g. USDA food database, Euro FIR) can only describe the fibre composition and cannot relate it to health effect. In FibeBiotics, we focused on a specific set of compounds among the dietary fibres, the high molecular weight non-digestible polysaccharides (excluding e.g. the oligosaccharides), and the term NPS (Non-digestible PolysaccharideS) has been chosen to specify this.
This proposal aimed at developing foods ingredients, based on non-digestible polysaccharides, which were targeted towards a maintained and improved host’s immune system. The immune system is of crucial importance for human quality of life. A dysfunctional immune status will result in numerous health problems including infection diseases, autoimmune diseases, allergies, inflammation, atherosclerosis, cancer etc. Given its important role and the fact that the majority of the immune system is located in the gut, the gastrointestinal tract has become one of the most important target sites to maintain or improve the function of the immune system even in apparently healthy individuals.
Nutrition has been shown to be an important modulator of gut and immune functions. It is, therefore, not surprising that the majority of the functional food market is directed towards the gut and immune system, e.g. probiotics and prebiotics (mostly based on oligosaccharides) and fibres or non-digestible polysaccharides. In the FibeBiotics project, we included several NPSs which have been hypothesized, based on preliminary research or published data, that they can have beneficial effects for gut function, mainly in relation to bowel habit, cholesterol lowering effects and reduced risk of cancer.
Less known were the direct immune supportive effects of NPS that have been demonstrated in in vitro and in vivo animal models and a small number of human trials. In particular, cancer related research has shown the potency of different NPS and their contribution to the immune system, a crucial pillar in the combat towards tumours. These studies have also indicated the variation in bioactivity among these NPSs and the gaps in knowledge such as structure-function relationships and mechanisms of health effects that need to be fully clarified in order to apply fibres efficiently. Unravelling the mechanism and developing good models that can predict health effects in humans, to select the NPSs with most activity were therefore main goal of the EU FibeBiotics project.
The European Food and Safety Authority (EFSA) is responsible for the evaluation of the scientific basis for the health effects of food compounds and products. Based on scientific panels, they are evaluating biomarkers, trial designs and study results and come with guidelines how to substantiate the health effects in specific areas. The legal limitation to approach food not in relation to disease threats but rather from the perspective to maintain health of healthy individuals, limited the possibilities to substantiate the effects. This requires the need for validated and broadly accepted (early) risk factors that warns for derailing of homeostasis or indicates a deficiency for a certain (cellular) function.
Based on the evaluation of immune related biomarkers, EFSA (draft guidance of the EFSA, 2010 and repeated in updated guidelines) stated that vaccination efficiency as objective biomarker could suffice to demonstrate positive effects of interventions towards the immune system. This could lead to health claims that would go broader then the used vaccination model but would indicate a more general enhancement of the immune system. These analyses would, preferably, also be supported by endpoint measurement like showing an effect of the intervention on reduced symptoms in the population.
This vaccination efficacy was used as one of the pillars of the FibeBiotics project. In order to have a broad impact of the intervention (specific NPSs that support the immune system), we wanted to target a large population that could benefit from this approach, the elderly. Elderly have been chosen since clinical efficacy of the flu vaccines in elderly subjects is around 60-90%, indicating that their immune system might be prone to be boosted to reach the same level of protection as young aged persons. Flu and post-flu pneumonia are important causes of mortality in elderly subjects. That is why, we also chose to use the common season flu vaccination as a model in order to analyse whether an intervention with specific NPSs could increase the antibody titres and number of individuals that are protected against the flu. Dietary NPSs may provide a valuable solution to this health problem which may be applied in a broad, population-wide approach both to support other vaccination models with low efficacy, use in other populations with a suppressed immune system (or temporarily due to extreme high challenges like sports or high risk jobs) or support populations that have high exposures to flu and colds (e.g. day-care workers). So, when these beneficial effects could have been proven by scientific sound evidence, the benefits for the population are enormous and costs for loss of working productivity and costs for health care will decline.
The NPSs are non-digestable dietary fibres, meaning that the human digestive system cannot degrade them like it can for instead degrade starch into glucose. However, when fibres enter the colon, they can be the food for the bacteria that are living there. The microbiota is important for human health as they produce and release vitamins, minerals, fatty acids etc. which still can be taken up from the colon by the host. These microbiota gain more and more interest as they can help to reduce the growth of pathogens that are latently present in the gut or can reduce the impact of new intestinal pathogens that enter our system orally. Correlation analysis between the presence of specific species in the intestine and health and disease have led to the identification of potential risky and beneficial gut bacteria. So are the colony numbers for Bifidobacteria and Lactobacilli often found to be high in healthy subjects. Research is still ongoing to unravel these correlations and whether they are cause or consequence of changes in health. At this moment, no general consensus has been reached within the scientific community and that is why the microbiota profile cannot be used to substantiate a risk-benefit analysis of a certain intervention. Only the more commonly known gut pathogens can be seen as risk factors. Demonstrating a significant reduction in the presence of these in lumen samples could help to substantiate a health effect. NPSs, as they can be selective food for some bacteria that can use them for their own resources, may also have a positive contribution to reduction of pathogens. Research, however, is difficult as natural infection rates among people are low and working with infected people difficult, for ethical reasons. Therefore in the FibeBiotics project, we included research to understand the impact of NPS on microbiota but did not make it the focus of the proposal.
Establishing a health effect of food products or compounds should be performed by controlled intervention trials. These trials, however, are costly, laborious and impact the included persons. Models that can predict the outcome of a human trial with high certainty would be very valuable as it can be used to: 1) support the mechanism of action of the bioactivity involved; 2) can be used to select and compare food ingredients to select the best bioactive ingredient; 3) can support product development with sustained bioactivity (knowing that matrix or other processing unit operations could breakdown or counteract bioactivities); 4) can be used to check products with a certain health claim and expected bioactive content for its reliability. As very often biological in vivo activity cannot be simulated by one model and as the bioactivity of these NPSs can be hypothesized to be brought about by different mechanisms we have set up, in the EU FibeBiotics project, several different in vitro models. These models could be used to systematically compare the different NPSs for their bioactivity. This type of systemic comparisons is hardly performed within the scientific community where a large diversity of protocols is used and focus is only on very limited numbers of compounds of interest. The adoption of these standardised models is important as they will make studies between labs more comparable, leading to better understanding.
As discussed at the start, many people know about the health effects of dietary fibres but on average do not reach the daily recommended doses. That it is why, it is important to comprehend consumer behaviour, both assessing the current knowledge level on fibres, the associations that are made, what kind of product they would take at what moment to increase fiber intake and what food products they would find acceptable to be enriched with bioactive fibres. Therefore, FibeBiotics also included research towards European consumers. Knowing that much of the Business to Business (BtB) or Business to Consumers (BtC) approaches start with presenting correct, credible and attractive information on the product, we also evaluated current strategies applied in this field and how companies associated to the FibeBiotics project could position their own products.
The EU call from which this project was selected requested a clear involvement and commitment of SMEs. The SME partners within the FibeBiotics consortium were in the field of technology and service provision in the area of food and health research and those that had their main business as supplier of ingredients for which they would like to be informed on the health effects of the NPSs within their product portfolio.
The objectives of the project
The overall objective of the FibeBiotics project was the development of functional food products, based on bioactive NPSs that support gut and immune function to maintain a healthy homeostasis in the elderly population.
Our objectives were to support the entire food industry (and not only the associated partners) by developing standardized methods, tools, pipeline and systematic analyses of compounds for their bioactivity. These technologies should be available to other partners in the future to develop more products that will meet the scientific requirements to allow claims on gut and immune function. These tools have been developed in the project by focusing on three main NPS compounds delivered by three of the SME partners but ensured at the same time, the translation of knowledge to other industrial partners that have the same or similar aim in developing products that support the gut and the immune system.
The objective to develop a NPS based functional food for promoting immune functioning, had been divided into five main objectives:
• Define methods to detect and characterize purified and processed NPS. Determine biochemical and bioactive alterations when designing functional food products in order to make products with remained activity.
• Develop highly standardized methods to classify purified polysaccharides for their bioactive effects on immune cells and microbiota and microbiota-mediated metabolites using in vitro and ex vivo assays.
• Improve the knowledge on the mechanism of action of NPS based on in vivo and ex vivo analysis as a result of small scale interventions.
• Combine all knowledge and show health promoting effects in a large pivotal human clinical trial.
• Perform all necessary translational activities for implementing this knowledge to support the claim substantiation of the studied NPS but also provide the tools, knowledge and part of the data to enable functional food research with a similar health target.
These objectives have resulted in many sub-objectives that will be addressed when discussing approaches and results.
Project Results:
Before going into details on the main S&T results/foregrounds it has to be said that the project had two phases. The first one was on developing, testing and establishing novel tools and strategies open to apply for everybody in the field to study health effects of dietary fibres and similar compounds. The second was to apply these tools to a selected set of NPSs, supplied by some of the FibeBiotics partners in order to build on a portfolio that might be used in the future in applying for an EFSA health claim when results turned out positive. Especially, the second goal of the project resulted in a careful evaluation of all the results before making them public, not disclosing results that could jeopardise potential IP that could be filed and not damaging business of involved identities. That is why all reports remained restricted within the consortium and also now we will talk in general terms on the results making it impossible to relate effects to specific NPSs studied in the consortium.
The NPSs within the project
Research could not have been done without NPSs as method development requires these NPSs, specially as we are exploring a new field on gut and immune support for which hardly positive control products, nor medicines, are available.
Within the consortium we had six partners that were responsible for preparing and providing these NPSs to the other consortium members for their research (see figure 1). Besides these main NPSs, in specific experiments other NPSs were taken along as reference when in the literature certain bioactive effects had been described which we wanted to check or compare with the newly included NPSs in the same experimental set up.
Biochemical analysis of NPSs
NPSs are derived from biological material which can vary due to genetic differences of the starting material, differences in growth conditions, harvest time, post-harvest treatments and differences in purification method. Therefore the project mostly starts with biochemical characterisation of the NPSs to determine how much of a specific NPS is present in the sample, whether it is one specific NPS or a mixture of NPSs and whether other compounds are in the same sample. This is important to compare different NPSs (based on a similar concentration) but also to learn about the variation in different batches of the same NPS. As long as we do not know what exactly determines the bioactivity of an NPS we also do not know whether these variations in samples are important. Therefore, at this stage, we liked to know as much as possible on the biochemical structure and conformation of the NPS and the other constituents in the sample that might have an impact on the bioactive properties subjected to our study.
Within the EU FibeBiotics project, we had two groups with expertise on biochemical characterisation of polysaccharides like the NPS studied here. As these NPSs are very complex and it requires several biochemical methods to fully characterise samples, these experts developed a schematic approach how to analyse these samples step-wise (see figure 2).
These methods and strategies have also been included in deliverable reports. They can support other groups in this field to analyse samples in the hope to reach more consensus on how to describe samples, as a first step to a more systematic comparison of NPSs and their bioactivity.
An example of the results of such a biochemical characterised can be found in Table 1. These results have been used in one of the publications where the effects of these NPSs preparations have been studied towards the small intestinal epithelium (see later). As part of sample characterisation we also analysed the presence of potential contamination (like endotoxins) in the sample. These small amounts of contaminants are no problem for most of the research to be done, but could form a problem (i.e. false positives) when studying the direct immune effect of NPSs on immune cells as these cells are very sensitive to these endotoxins. In scientific literature, sometimes proper controls for contaminants are not presented which, to our opinion, hampers interpretation of several publications worldwide.
More than expected at the start of the project we indeed encountered the presence of too high concentrations of endotoxins in the samples to be able to use them directly in some of the immune assays. Therefore extra efforts have been invested to further purify some of the samples for immune cell related research. As purification might again change biochemical structure of the NPS, the development of this purification protocol was supported with biochemical analysis. Figure 3 shows the effect of the purification strategy developed, in order to reduce endotoxin to acceptable levels, for one of the NPSs. It is crucial to find a method with the right balance between ‘not too harsh to change structure like the molecular weight’ of the starting NPS material (Figure 3 - blue line) but ‘harsh enough to remove the endotoxin of the sample’. The research resulted in a paper accepted for publication and was the key to continue the research towards analysing the immune supportive effect of NPS. The developed method can also be applied in similar studies to remove endotoxines from other samples for which immune modulating effects are studied.
Product formulation with maintained NPSs
In parallel to the research to establish the effect of NPS towards the gut and the immune system it is important to know whether you can make products that still contain the same NPSs as we used in the bioactivity studies (the purified compound without the processing steps of a real food formulation). That is why within the FibeBiotics project, several products have been prepared including cheese that contain EPS, smoothies that contain beta-glucan from oat and yeast and crisp bread that contain different NPSs supplied by the consortium (see figure 4).
Biochemical methods were subsequently used to study the fate of the NPSs in these food matrix in order to understand potential risk when preparing products enriched with these NPSs. In addition, these biochemical methods will be important for food authorities to check compliance of food producers that the claim they might make on the package (like ‘yeast beta glucan inside’) is indeed reliable. These products have subsequently been used to validate strategies to analyse these specific NPSs in these food matrices. This is not an easy task as food matrices are complex and can have very similar compounds present (e.g. how to distinguish yeast used for making the dough from those derived from the yeast NPS specifically added to the product because of health effect). But we can say that major steps have been achieved and have resulted in methods, protocols and strategies that can be used or further developed (see figure 5). One paper has been published on NPS detection methods and another, describing the methods to study fungal beta glucans in different food matrices making use of alanine blue, will be prepared and submitted.
Fermentation of NPSs in the gut
To be able to demonstrate gut and immune supportive effects it must be clear that the NPSs will not be hydrolyzed into small mono sugars units in the mouth, stomach or small intestine. As defined by the term NPS (non-digestible Polysaccharide) it is expected that these polysaccharides resist the passage over the small intestine (or at least most of it) until the colon were they then might be degraded by microbes. The subsequent degradation of the NPSs in the colon will mostly depend on the bacterial species present and the enzymes they produce. However there is also a lot of redundancy between the species present which can take over the same capacity for fermentation in the gut ecosystem. To check whether these NPSs indeed are not broken down too rapidly in the gut, several models can be applied. The most simple method is based on single tubes in which human enzymes (like amylase and brush border enzymes), pH changes and bile acid etc. are added stepwise to simulate the mouth, stomach and small intestine. Indeed the NPSs of interest in this consortium survived these upper gastric gut simulations. The next step was to culture these samples in the presence of a human feacal sample containing a high complexity of microbiota. This research led to the conclusion that some of the NPS are more resistant to degradation by bacteria than others. An example is given in figure 6 for the NPS AX indicating that after an incubation of 24 h some parts of the molecule are degraded (bacteria needed to switch on the right enzymes and have to grow in number) while other parts of the molecule are to difficult to digest and remain in the sample until the end of the experiment (48h). These type of analyses made clear that NPSs can differ in the resistance towards fermentation in the colon.
Fermentation of NPSs can also be associated by gas production which is in generally less preferred as these can cause bloating effects in the gut.
In vitro analysis of NPS towards the microbiota and related metabolites
These relative simple experiments as described above were given a follow up in more advanced research models. One is the SHIME (Simulation of Human Intestinal Microbial Ecosystem) that can simulate the trajectory of a product from mouth to colon in a more automated way and that also includes a separation of the colon in three parts. This type of analysis can support further understanding of the fate of NPSs in the human intestinal system (see figure 7).
This model has been further optimised within this FibeBiotics project and subsequently has been applied to study the fate of some of the NPSs in more detail. Not only the breakdown of the NPSs is important knowledge, also the formation of fermentation products like short chain fatty acids (SCFA) that can support in local and systemic metabolic regulation and can signal towards the immune system are important to measure. This fermentation of fibres in SCFA is in general considered as a generic positive health effect of dietary fibre as almost all dietary fibres are converted into SCFA. However, the type of SCFA (e.g. acetic acid, propionic acid, butryric acid that are the major ones) produced during the fermentation can be different, the ratio among these can vary and the total concentration of SCFA can slightly vary between the different dietary fibres. Besides, depending on the bacterial species that are involved in this fermentation, fermentation of the same NPS can therefore depend on the feacal sample used for the fermentation study. Within a SHIME these SCFA can be determined at different places in the intestine and give more accurate measurements than analysis done on humans for which on a routine base only faecal sample are available.
However, these faecal samples indicate produced SCFA that have not been absorbed by the host and therefore hardly will indicate the real benefit to the host. Part of the FibeBiotics project was therefore dedicated towards the development of an on line SCFA sensor that might be applied in vivo in the intestine when possible or applied in a model like SHIME to get more accurate measurements. One of the deliverables describes the development and the test phases in more detail (figure 8).
To cut a long story short, it has to be concluded that another type of sensor than the current MOS (metal oxide sensor) type sensors might be more successful (due to stability, energy supply, working in a 100% humid environment and at low oxygen levels as in the gut) therefore more developments will be needed.
Besides the SCFAs, the SHIME also enables the simulation of the effect on pH which might be partly linked to SCFA production. The hypothesis is that lowering the pH in the colon may help to boost the most positive know SCFA butyrate and might also enhance butyrate producing bacteria. Indeed, using this model also effects on pH have been measured.
Like stated earlier, the microbiota in the intestine can bring both risk and benefit to the host. Risk factors are endotoxins, exotoxins and harmful metabolites that can be produced by the bacteria. On the other side vitamins, SCFA, and pathogen-associated molecular patterns (PAMPs) that make the immune system alert but without affecting the host negatively can have immune modulating activity (and hence represent a potential indirect immunomodulatory effect of NPSs). We studied, making use of the SHIME model, changes to gut microbiota and metabolites when replacing starch, given as food intake in the SHIME model, by different types of NPSs. Figure 9 gives an example of such an experiment indicating that, when starch is replaced by beta glucan from oat, more Lactobacilli, a group of bacteria associated with health benefits, will grow in the colon with higher (relative) abundance.
Supportive tool to study microbiota
Some human diseases have shown correlation with the present of absence or abnormal ratios of certain microbiota in faecal samples. There are several technologies to study the microbiota community composition ranging from specific quantification of only a very limited number of species to full sequencing of all DNA, RNA or rRNAs or certain areas of the genome (mainly 16S) which allows identification and quantification. These full sequencing analyses decrease in cost with the Next Generation Sequencing technologies that are being developed but still are costly when many samples from a large cohort or intervention trial need to be analysed or when it would be used for routine screening of a population. Therefore, the FibeBiotics project also gave attention to new cost-efficient microbiota profiling technologies that could be beneficial for some specific application.
The GA-map™ analysis consists of various steps starting from DNA isolation of faecal samples to analysis of the microbial profile (figure 10). This profile is based on 48 unique bacteria probes able to be detected simultaneously in one single tube. In the FibeBiotics context the technology was further optimized and adapted in order to study the 48 different bacterial groups that can profile the microbial community of elderly European subjects. The developed probe-set has been tested in selected samples which are representative of the elderly population, and compared to the 454 pyrosequencing data to confirm correlation. A very good correlation has been observed. Several optimization steps have been included to increase robustness and reproducibility. The University of Bologna has been included to for data processing and statistical analysis. Subsequently the technology was used in the human intervention trials to study the effect of NPS intervention towards changes in the microbiota composition.
NPS bioavailability
In pharmacolog,y it is very common to perform pharmakinetic research in order to determine the bioavailability, organ targeting and metabolisation of the compound of interest. For food compounds these data are less required, especially as the food ingredient already has a novel food status. However, the place of action of NPSs in the body is a recurrent question; will these often high molecular weight molecules be restricted to the intestinal lumen or can they pass the intestinal epithelium and enter the blood stream to interact fully with all kind of immune cells and other organs than the intestine.
Previous research had shown that a labeled, high molecular weight NPS, could be traced back in the blood after oral intake, although at low percentage. We wanted to study whether this could also be the case for other NPSs but then not making use of labelled NPSs (expensive) but by sensitive biochemical methods to make it more generic. When succesful, these type of analysis might be added to EFSA dossier in order to support a specific hypothesis. Spiking experiments in which NPS were added to blood and then analysed gave some results but also a lot of background from the very complex blood matrix, even after applying several purification methods. It had to be concluded that it will be impossible to detect the NPSs in blood after oral intake as concentrations will be too low in relation to disturbing factors in the blood. When EFSA would require such information in dossiers, research should be performed with expensive labeling experiments.
In vitro analysis of the effects of NPS towards gut and immune system
As explained before we would like to be able to characterize and select NPSs based on their bioactivity in humans. However, still several hypotheses for this function in humans are possible. Therefore we studied this bioactivity towards different cell types and with different response read-out in order to study whether there are differences in bioactivity between the NPSs of interest. Besides, these analysis hint towards the potential involved mechanism underpinning the effects found in vivo. The in vitro assay will then also be able to support the development of products with the best possible effect resulting in a highly potential positive outcome in human intervention trials. Besides, these in vitro assay can support quality control of the product when something changes in the production and processing of the product and can be used to check products for their claimed bioactivity.
Several assays have been developed and used to study the effect of NPSs. The first is based on human small intestinal epithelial cells, the first cells that are exposed to NPS when orally provided. These cells are cultured in the lab and can be exposed to NPSs. Based on gene expression and analysis of compounds in the excreted medium it was concluded that some of the NPSs could activate immune related genes like NFkB and as a consequence could secrete chemokines basolaterally to attract other immune cells to the place of action. This could support the hypothesis that some NPSs are recognised by the intestinal cells as ‘non self’ or ‘strange’ (as some have similarities with cell walls of pathogens) and subsequently alert the immune system. A publication on the obtained results has been submitted.
The second assay that was developed and used is based on human macrophages. Macrophages are shown to reside in the gut epithelium and are efficient in sampling compounds from the lumen and present it to other immune cells (antigen presenting). Research has been performed with reporter cell line of macrophages in which specifically the NFkB pathway was used as a read-out. Besides, research was performed based on primary cell, isolated from blood of healthy donors. Especially based on these primary macrophages interesting results were obtained in which it was shown that these macrophages that classically are divided in tolerogenic (M2) or pro-inflammatory (M1) type of macrophages now, after NPS exposure, show a new phenotype (see figure 11). Follow-up experiments demonstrated the presence of interesting cocktails of chemokines produced by the cells and excreted in the medium. In real life this could indicate that macrophages in the intestinal mucosa that are exposed to specific NPSs could start to produce chemokines to attract other immune cells that can support in direct killing of unwanted pathogens. This is in line with the observations on intestinal epithelial cells, but in a more pronounced way and again indicates that NPSs are seen as a kind of warning signals to make the immune system more alert. The research will be materialised in several articles and a PhD-thesis for the PhD Y. Tang that participated in this FibeBiotics project.
Besides these assays several other in vitro assays have been developed and used to characterise the potential bioactivity of NPSs. A complete overview on the in vitro tests and results from the NPS exposure experiments can be found in the (restrictively disseminated) deliverable reports but some of the results will be included in scientific papers.
Effect of NPS towards the gut barrier function ex vivo
Experiments with human epithelial cells grown in vitro in the lab showed that these cells can respond to the NPSs. However these cells cultured in the lab are only one type of cell (enterocytes) while the intestinal epithelial is a mixture of different cross-talking cell types with different functions. Therefore research was started to expose NPS to biopsies taken from healthy subjects from different location in the intestine to study the effect of NPS exposure on gut barrier function. As these biopsies normally have a high integrity and not allow compounds passing the epithelial layer, the exposure of NPSs was combined with a stressor that temporarily reduces the integrety. Subsequently the effect was studied by fluorescent molecules in order to study the passage through the intestinal biopsy and how well the cells could resist the stress. This research was performed under an approved Ethical protocol. The project encountered several drawbacks due to variable electric supply, breakdown of essential parts of the Ussing Chamber equipment, unrepeatable effect of a stressor which forced to change towards another stressor (to which some of the volunteers did not show any response) and bruising of biopsy material upon transport. The biopsies were also found not to survive exposures more then 1.5-2 hours, which hampered reliable gene expression analysis of the tissue and analysis of compounds excreted in the media that were all very low due to this short incubation time relative to the large volume of surrounded media.
However, after selection of subjects that responded to the stressor, the effect of NPS on biopsy barrier function could be studied. Both young and healthy subjects were included in the study as well as senior participants with gastrointestinal symptoms of constipation and/or diarrhoea. This resulted in the observation that one of the NPSs studied gave a statistically significant reduction in paracellular permeability and transcellular permeability compared to intervention with only the stressor. These results indicate that some of the NPSs that were tested might contribute to maintain homeostasis of the gut barrier function (figure 12).
Effect on NPS towards the gut barrier function in human volunteers
The human intestine has a complex and crucial function as it needs to take up nutrients from our food and at the same form an efficient barrier to the commensal bacteria that we host in our intestinal system and the potential pathogenic and immune potentiating factors they produce. Several researchers already launched the ‘leaky gut’ hypothesis indicating that the barrier function of the gut is reduced due to life style habits, including western diet. Due to this reduced barrier function our immune system would be chronically overloaded with risk factors to it is forced to respond, finally leading to an exhausted immune system that fails to work properly. This would result in increased sensitivity to pathogens and stresses. The validity of this hypothesis is still in debate and supportive evidence is still the research subject of several groups. One of the proposed strategies to study the differences in gut barrier function in volunteers is passage of sugar probes. Different sugars are taken up differently by the gut epithelium and at different regions of the intestinal system. This resulted in a 5 sugar test that can be administered orally to subjects after which the urine is analysed 5 and 24 hours after the intake. The amount of the sugars found in these urine samples can be seen as an indication of the barrier and uptake function of the intestine in different regions (duodenum, small intestine, colon). However, under non stress condition these barrier and uptake functions between healthy individuals are hardly different. This sugar test, can be combined with a stressor (like used for the biopsies studied reported earlier) that can reduce the barrier function of the intestine. In this format, the test can differentiate between persons on the basis how they cope with the stress provided. Several trials have been performed in which the reproducibility of a stressor was tested resulting in the selection of indomethacin as stressor.
The FibeBiotics in vivo barrier trial, a double-blinded placebo-controlled randomised parallel study, aimed to investigate whether 6 weeks of oral intake of two different NPSs (compared to a placebo) could improve intestinal barrier function against this indomethacin-induced barrier disruption. 60 healthy individuals representing a general population of elderly people (65 years old and above) were recruited and randomly divided over the three groups. The results from the 6-week intervention study with daily oral intake of the NPSs did not show any improvement on indomethacin-induced hyperpermeability nor on baseline permeability. Although indomethacin significantly increases small intestinal and colonic permeability at baseline, the interventions did not show any statistically significant improvement (figure 13). All the results have to be studied in more detail and will be used to write one or more publications which will also make part of the thesis of the involved PhD student at Orebro University, Sweden.
Effect of NPS towards vaccination efficacy based on human pilot trial
The heart of the FibeBiotics project was the human pilot trial in which the effect of NPS intervention was evaluated making use of vaccination efficacy, an EFSA supported biomarker for the immune system. The trial was designed as a pilot trial as at that stage we wanted to explore more biomarkers and wanted to see how other immune related biomarkers might correlate with the primary biomarker related to the vaccination. At that stage, we also wanted to compare the different selected NPSs within this project to be able to compare results obtained in vivo with those that are obtained in the different in vitro models in order to find potential underlying mechanisms and strategies to predict the potential outcome of an in vivo trial based on in vitro selection and product design. An ethical approval was requested and provided by the responsible authorities.
The trial was performed with healthy volunteers, in the range of 50 to 78,5 years and with a mean age of 66 +/- 6, gender balanced that received a standard seasonal flu vaccination. The trial was highly complicated in which many persons had to be recruited, health status had to be recorded, samples have been taken (of which some had to be processed directly) and biomarkers had to be measured. An schematic outline of the pilot trial can be found in figure 14. The trial was very well organised by Clinical Research Centre Kiel, Germany and resulted in large data sets on which extensive analyses have been performed, and still can be performed when new hypotheses on certain correlation of biomarkers are hypothesized. Unfortunately, publications have not been completed yet.
The trial can actually be seen as three trials. The first part includes the analyses that were performed before the start of the intervention and collected at visit 0 and visit 1. This part of the trial is called the association study as we can relate the many measurements performed and biomarkers determined to each other and to the individual parameters like age, gender, BMI, waist circumference etc. As the general population showed BMI of 27.2 which have been classified as over-weighted (although other see it as the current normal/standard) it has been possible to find variations in the parameters for Metabolic Syndrome, liver sclerosis, endothelial function etc. and how they correlate with other biomarkers determined. These are currently incorporated in a manuscript that presents new knowledge on health parameters and how these can be related.
The second part of the study can be regarded as the main study in which the effect of the NPS intervention is studied on the primary parameter, vaccination efficacy. The standard seasonal flu vaccine is a combined vaccination of three different vaccines for which antibody titres have been determined separately. This resulted in HI titres at four different time points and many type of analysis as results can be approached as HI titter (e.g. V3 and V4), delta or fold difference between (V1 or V2 verses V3 or V4), geometric mean fold increase, (increase in) protection rate and seroconversion rate (subject with a four-fold increase in HI titres). Besides, calculated for the individual vaccines, they can also be calculated for the combined titres. To make it even more complex, the data can or should be analysed by different statistical tests depending on the distribution of data points. This resulted in a large summarising table, from which it was concluded that none of the NPSs showed a significant effect on these parameters (as also was expected based on power calculation) and that one or more showed promising trend in increasing the HI titres that might be caused by the intervention (see as an example of these analysis figure 15).
In addition, microneutralizing titres have been determined (another technique used by the to determine that people have serum cross-reactive antibodies to the influenza viruses used in the vaccine which is more specific than HI titres) and common cold/flu-like symptoms were scored. Some of the NPSs performed promisingly in reducing these symptoms although the incidences were very low (as can be imagined for 40 persons in a 5 weeks period), making it difficult to draw biologically relevant conclusions. Based on these results it had to be concluded that: 1) some of the NPS showed positive trends, power calculations indicating that when the arms in the trial will raise from 40 to ~170 results could become significant; 2) that, although a trend is seen for some of the NPSs, it will be hard to correlate the effects in vivo to in vitro, except that the one that looks most promising would be expected also to show most promising effect in the in vitro tests; 3) that the flu vaccination model might not be the best model to analyse these effects as people have built up memory (as often same vaccines have been given previously) making it uncertain whether primary vaccination responses are analysed or whether also secondary memory-driven responses are included; 4) the results in reducing common cold like symptoms or disease episodes were very interesting and potentially also more appealing as supportive evidence to consumers than vaccination efficacy which will be harder to understand.
The third part of the study was dedicated to identify potential other effects that might have a link to the observed effect in vaccination efficacy of the NPS interventions. Cellular parameters were determined by analysing the blood of the volunteers for cytokines (small molecules in the blood that help to communicate between cells and therefore important for metabolic function). These cytokines were analysed in unstimulated blood but also after stimulation with LPS, conA and vaccine in order to analyse how people respond to these challenges and whether people within a specific NPS intervention group might respond differently. Despite the extensive data set, no major effects have been discovered yet. The bioinformatics experts will probably re-analyse these data when new hypothesis are generated and tested based on this dataset.
Within this third part of the study also the microbiota analysis can be grouped. Effect of NPS on microbiota can be expected for some of the NPS that were given in high amounts in the intervention. These expectations were also fed by the in vitro analysis which shown effect towards increased Bifidobacterium or Lactobacilli in some groups. Some of the samples were analysed by 16S rRNA seq analysis and all samples analysed by the GA-map technology as explained above. The results obtained by both technologies have been compared and the effect of the NPS intervention studied, also in connection to SCFA that were analysed. These results will also be linked together in a publication.
A pivotal trial to study effect towards episodes of common cold
Based on these results, we had to design the pivotal trial in which we hope to establish for one of the NPSs that, by increasing the power by including more people in the arms of a trial, results will be more (statistical) significant and later can be used as part of an EFSA claim submission.
After a long discussion and consulting the Scientific Advisory Board including a person with experience in EFSA, setting up the guidelines and involved in dossier evaluation, it was decided to change the strategy from vaccination efficacy to common cold/upper respiratory tract infection symptoms scored as primary parameter to study the effect of one of most promising NPSs as concluded from the pilot trial.
The pivotal study was designed, ethical approval obtained, NPS and placebo were produced and packed and recruitment was started. As we decided to focus on common-cold/upper respiratory tract infection related symptoms we had to perform the trial during the winter season. In order to have a high chance that people would experience these symptoms, the intervention period was extended from the 5 weeks in the pilot trial to 12 weeks in this pivotal trial. To work according to the guidelines of EFSA, disease symptoms scored by the participants were also checked by a physician. A full overview of the study design can be found in figure 16.
Inclusion criteria of the trial were men and postmenopausal women aged ≥ 50 years that were willing to abstain from food and supplements containing probiotics, prebiotic supplements and other fermented products. Besides, they should be willing to sustain a low dietary fibre diet for 14 weeks and willing to obtain the influenza vaccination season 2014/2015 (but did not receive it yet). Like all studies, we also had a list of exclusion criteria.
The symptoms scores were recorded (=number of subjects affected by at least one episode) according to self-reporting (questionnaire based on Jackson et al. 1958) and physicians diagnosis and then calculated according to the best validated and blinded method (so people that do the statistical analysis only know whether people had belong to group A or B and not whether they are in the NPS or placebo group).
Many different analyses have been performed on the dataset, like calculations based on occurrence, incidences, severity of the symptoms, duration of the episodes and for type of symptoms separately and combined. An example of such analysis can be found in figure 17 which illustrates that both groups shows some differences on some of the scorings. However, when statistical analysis were performed on these data, none of the parameters showed to be significantly different. At this stage the trial is still blinded until the pilot trial is published. Upon this publication, this pivotal trial will be unblinded and meta-analysis performed to see when both the pilot trial and the pivotal trial are combined, for this NPS and placebo control arm, whether more promising effects can be concluded.
It can be concluded that the NPS tested in both the pilot and the pivotal trial showed: 1) an increase in GMFI of HI titres against virus A H1N1 and combined virus strains; 2) an increase in Δ seroprotection rate against A H1N1 (23%), B (17%), H3N2 (6%) and combined virus strains; 3) an increase in seroconversion rate against A H1N1; 4) a higher local vaccination response; 5) a lower incidence/occurrence of common cold (1 vs. 5 according to Predy; 1 vs. 8 according to Jackson) and 6) a higher IFNγ (unstimulated) ex vivo but that these effects are not statistically significant. For the pivotal trial, no conclusions can be drawn yet, as the experiment is still blinded.
Database and warehousing for now and the future
Data management and data analysis get more and more important, also in research. Many researchers but also industrial partners still work with separate files, stored in a hardly organised way in project folders which can get lost or are very hard to interpret by others as identifiers for the different columns and parameters are missing. The idea of the FibeBiotics project was to select/develop a user-friendly software system, highly flexible for the data sets that are developed and implemented in this kind of research project. This could then have been a good and easy experience for all partners, and the industrial partners that are regularly updated about such a tool to encourage future groups to work which such databases. An appealing property of databases (when working properly) is that data are well organised with links to the protocols used, processing steps of data are documented and data can be combined quickly to perform new types of correlation studies that else would take a lot of time to subtract from individual excel files with different formats.
This part of the project was a major challenge as the research performed in the whole FibeBiotics project was very diverse and so were the data generated: biochemical analysis, NPS description and samples from different batches and different purity, different cell lines and read out (microarray analysis, cytokine analysis, NO, qPCR etc.) and human trials. These human trials are very complicated, with general information (which should be coded to protect privacy), food intake data, questionnaire data, and the analyses done on all samples. To give an example, one blood sample was used for HI titres (3 individual vaccines), MN titres, cytokine (each sample had 4 different treatments and then 7 different cytokines were measured) but also many other health related blood values like HDL, triglycerides, insulin, liver function markers, LBP etc. Some of the analysis are performed on all samples, some only on a limited number of visits or measured only at the start. So besides this being a logistic challenge to do all these analyses, it is also a data management challenge to combine them all together in a database in such a way that other researchers that might want to work with the dataset in the future still understand the data and how they are connected.
A data warehousing system, developed and used in the large EU framework NUGO was selected and installed by the partner responsible for this task and made available for the partners of the FibeBiotics consortium. Due to misconnection between software system and regular updates of programmes and links used to internet pages, many maintenance actions were needed to make it work and to make sure that it will keep working properly in the future. As in this project different partners will have different interest and responsibilities, especially as it also involved data from a human trial, for which the possibilities for data sharing are restricted by law, specially strategies were implemented concerning ownership of data and disclaimers on how to ask permission for use of data from the database and how to acknowledge persons and groups which made the data available.
Then another challenge was on ontologies, how to label data, how to upload them etc. A first attempt to have all partners upload data themselves somewhat failed due to uncertainties about field names, data sets that differed from the standard formatting and the unfriendly-user interface and the inflexibility of the software. An instruction movie prepared by the responsible partner helped to instruct partners better but still some confusion remained.
Starting from a new installed version of the database, it was then decided to re-upload all data by the ICT partner in a much more organised way. After a short explanation of the dataset, by the person who generated the data, and how it was provided and what each field in the file meant, the responsible partner, knowing the programme best, uploaded the data in the database (see Figure 18, Start screen of the FibeBiotics database).
An overview of the data currently available in the database is in one of the project's deliverables. In addition, at the end of the project we have secured continuation of the database for at least another two years but also the data will be collected in the original files and stored at the coordinator for further use.
Mathematical strategies, can it support to explore the data better
As data become more and more complex also data analysis, statistical tools and mathematical models increase in complexity but also present opportunities. In most human clinical research, the number and type of statistical tests are often dictated by the type of data available and the type of comparison one wishes to make. However in research, when you want to build new hypotheses on correlation, complex interactions between parameters and classifying data to increase understanding, many new tools, models and strategies can be explored to make more out of your data. The FibeBiotics partner, University of Bologna, is strong in statistics and mathematical modelling of the data and has supported the data analysis that we generated and that was provided by other groups. Some of this knowledge was more theoretic or tested on other dataset than based on the FibeBiotics project, but which helped to generate new strategies. This led to several publications of the group of Bologna in which FibeBiotics was acknowledged. Presentations were given by the Bologna group during the project progress meetings and the Industrial Platform meetings to make people more aware of the possibilities.
One of the research examples that was based on FibeBiotics data is the research performed on microbiota analysis. In the intestinal system, more than 1000 bacterial specifies can be present and distinguished by 16S rRNA sequence. Each person will have another composition of the microbiota. This makes the analysis of an intervention of a whole group of persons very challenging, as it can depend very much on your starting situation. Based on data generated on the faecal samples from the human pilot trial and the 16S rRNA sequences analysed by Teagasc the Bologna group thought about a new strategy to analyse these data. They modelled humans as a steady-state process in which you have an inlet of new bacteria (entering orally), reproduction of bacterial species due to available food, natural death of bacteria and an outlet of bacteria as can be found in the faeces: the samples that were analysed. This type of analyses could generate new ideas on what is really happing in the gut as a result of changing dietary interventions. This has led to an interesting publication (Sala C et al (2016) BMC Bioinformatics. 20;17 Suppl 2:16) from which figure 19 is taken.
The EFSA guidelines and practise
The food industry is confronted with the EFSA based legislation about what is allowed to declare on the package of a food product, how much can/should be in products of a certain bioactive ingredient, what can be communicated to consumers on content, nutritional value and what on health effects. As this type of EFSA regulations have not been settled yet, it is of interest to the food industry what EFSA requires in order to authorise a certain content or health claim for a product. Therefore, the FibeBiotics kept track on EFSA guidelines, EFSA opinions, submitted dossiers in the area of fibres and in the area of gut and immune function. These evaluations have been reported to the FibeBiotics consortium and to the Industrial Platform as they are important to keep in mind to perform research when the aim is to use the data to submit a dossier to substantiate a health claim.
In the domain of immune support, the food compounds that have authorised claims are mainly established vitamins and minerals which have shown effect on the immune system mainly (or almost only) when deficiencies and diseases occur in people due to malnutrition. So results from a vaccination trial like used in this FibeBiotics project (as one of the only markers that is currently broad accepted as an indicator for the immune system) were not the basis for those health claims (although some show effect on vaccination efficacy based on populations in developing countries).
Dietary fibres like the NPSs have been recognised for other health effects than immune effects, notably cholesterol lowering, reduction of post-prandial glycaemic response, increasing faecal bulk, reducing transit time, maintain normal bowel function and reduction of body weight (see table 2). These parameters are easier to study than immune related responses as for cholesterol a widely accepted biomarker is available (HDL/LDL) and the others are relatively easy to quantify. The BGO as used in FibeBiotics have already been authorised with several of these health claims. Other of these health claims can be seen as ‘low hanging fruits’. As an example, it can be expected that all non-digestive polysaccharide will contribute to a reduction in post-prandial glycaemic response and that due to water binding capacity, as seen for many of the fibres, they also will have a positive contribution of increase of faecal bulk, stool frequency and stool consistence.
Due to the lack of accepted biomarkers other than vaccination and the complexity of the vaccination response due to memory reaction and the variation between individuals, we expect that it will still take some time before the first product with an immune claim is accepted. In 2012, at the start of the FibeBiotics project, we had expected that certain probiotic strains that have shown effects on the immune system would also have a dossier on immune health support authorised by now; which is not the case. This is neither the case for probiotics: although some studies show clear effects on the duration of common cold episodes, no health claim has been approved yet.
In the new general guidelines and those specifically for gut and immune health, we see that the EFSA allows some room to base research not only on healthy people (for which it is very difficult to demonstrate health effects when no risk factors are general accepted) but to also include those who have mild forms of a disease when it can be made acceptable that the product and established effect can also be active in healthy persons to maintain health. We have to await how this field will develop for NPS and immune support in the future.
Consumers, the unpredictable factor
The final goal of the FibeBiotics is that healthy food products are developed, based on scientific evidence and that consumers will buy and consume them in order to support their health. The food ingredient, product, information provided and health effect should, therefore, be of interest to consumers. In order to investigate the knowledge level on fibres and current opinion of consumers, we developed a questionnaire which was used as market study among consumers in three European countries. The full result of the market study can be found as one of the deliverables. Here, we only will summarize some of the main results.
The consumer survey was performed in men and women in three countries: Germany (N=1263), France (N=1228) and The Netherlands (N=1234). All three consumer samples were representative on the variables gender, age, education level, and social economic status. The questionnaire was translated into the three applicable languages. The consumer survey was performed in the first months of 2014, since it was decided that it would be preferred to execute the survey in the ‘flu/common cold period’ due to the specific questions on potential health effects of fibres.
Health impact of food becomes increasingly important with age. Most respondents perceived eating fresh and having a well-balanced diet as eating healthy. In general, healthy foods were considered natural, tasty, enjoyable and easy accessible. About half of the consumers did not use non-prescribed health improving foods. However, if one wanted to improve health, vitamins were mentioned as the most consumed, followed by calcium and fibres.
In general, most important aspects for consumers when buying food were price and taste. But for older consumers (55+) health impact was the most important factor. Some differences among countries were found: taste was found more important in Germany, whereas health impact was a more important aspect in France.
Overall, consumers were (very much) interested in consuming fibres. Consumers perceived bread and rice as natural carriers of fibre. Enriched foods were perceived as accessible, enjoyable and good tasting, but also as expensive and unnatural. Almost half of all respondents indicated not to have a strong opinion. This might indicate little knowledge and experience regarding enriched foods.
Although consumers were generally interested in products which help them to lose weight or enhance well-being and slow down the ageing process, health benefits by fibre enriched products, like improvement of the immune system and reduction of intestinal problems, were also appealing health benefits (figure 20). Fibres and whole grain products are most likely to be consumed in the case of stomach and digestion problems.
German (69%) and French (61%) consumers seemed to be more interested than the Dutch (50%) in consuming products that benefit the immune system and will reduce intestinal problems. In addition, middle aged adults (30-55 years) were most interested in enriched food products.
Consumers from all three countries preferred fibres in cereal based products, like breakfast cereals and a bar. Baked goods are also a preferred option in Germany and France, however surprisingly not in The Netherlands (figure 21). Functional (healthy) foods are preferably bought in drugstores/health stores (Germany & The Netherlands) or in pharmacies (France).
Most consumers paid little attention to nutritive information and hardly do any research on their own in order to find additional information regarding nutritional values. Consumers in general are unfamiliar with functional foods. Thus, providing more knowledge about and giving the opportunity to get experience with enriched foods is needed.
Consumers based their decision whether or not to believe information about food and health mostly on their own judgment. Medical sources and scientific research were considered less convincing (figure 22).
Especially, this last result is an eye opener for scientists such as within the FibeBiotics project, who have another opinion; research should be leading in how we are able to judge the potential health effect of our food. As this counts for the whole food and health domain worldwide, all scientists should try to (re)gain the trust of consumers. In food and health, also many scientists combat each other with different opinions, and the media are as well involved in the controversy as they like to overstate the research results to seek the attention of people. A more balanced discussion and reliable performance of scientists might help to improve this.
Conclusions from the Consumer Survey
Overall, consumers are interested in consuming fibre (enriched) products. Differences among countries were found. In order to position functional foods in a rapidly growing market, it is crucial to use an effective strategy. To position products successfully into the consumer market the following recommendations are made:
• Taste and price are the most important aspects to consumers and, therefore, must meet consumers’ expectations.
• The enrichment of a product needs to be in line with natural characteristics, e.g. fibres in breakfast cereals.
• Cereal-based products are preferred as fibre carrier.
• Drugstores/health stores or pharmacies are a right product outlet for functional foods.
• The product could be best targeted towards middle aged to older consumer segment. Older adults are more focused on healthy products and healthy ageing.
• Perceived health benefits are crucial. Though fibres are mostly associated with digestive health, consumers would buy fibre products if they have immune benefits.
In line with literature, this study also has shown that taste is the most important aspect of food products. Therefore, sensory research before positioning in the consumer market is recommended.
Potential Impact:
The original objective of the FP7 Work Programme 2011 was to improve health and well-being of European consumers by increased consumption of health products and at the same time to increase the strength of European SMEs. Moreover, we have tried to support the food industry and further help to study the EFSA approach on biomarkers to study immune support.
The impact for the society
Cold viruses are the most common infectious agents humans encounter. Cold viruses result in significant costs to the economy, considering lost workdays and decreased school attendance. Cold episodes range from 2-10 colds per year, depending on age and exposure. Common colds resulted in more than 100 million physician visits annually, at a total cost of $7.7 billion. At least one-third of these patients received an antibiotic, even though antibiotics have no effect on viral infections, not only adding to the cost but also contributing to the threat of antibiotic resistance. Research that integrated all costs (including absence of employees with sick children at home) showed that common colds and flu are responsible for a cost of $ 40 billion annually. Finding alternative strategies like the NPSs studied here, could reduce these incidences and episodes and therefore would have a high impact on the community.
Influenza is an acute respiratory illness caused primarily by the influenza virus (serotypes A and B). It occurs worldwide and is responsible for considerable morbidity and mortality. Serious illness and death rates are highest among the elderly, young children under the age of two, and those with medical conditions, placing them at increased risk for influenza complications. Because of the potential severity and epidemic/pandemic threat in most countries it is practiced to vaccinate elderly persons and people at high risk like health care workers. The idea is that immunisations will prevent hospitalisation and/or death; reduce influenza-related respiratory illnesses; and decrease physician visits among all age groups. However, it is generally known that the effectiveness of these vaccines declines with increasing age which explains why the preventive effect is only observed in persons below 80 years of age. Some studies even show marginal significant effects between vaccinated and non-vaccinated persons. At the same time, it was concluded that the burden of influenza in the elderly is substantial and even a modest protection of a 30% reduction in influenza-related hospital admissions would be beneficial. The pilot trial showed that NPS intervention group 4 shown a delta seroprotection rate increase of 26% for influence A H1N1 and also for the other viral strain strains. Although this difference was not significant based on a group of 40 persons, when this line would be true for the total vaccinated population, this NPS would help to save many lifes and would reduce costs for medicines and health care drastically. As no vaccination analyses have been performed in the pivotal trial we will never know whether the results found in the pilot trial could be repeated and would indeed point to an significant increase in the influenza protection level of a population. Therefore, more follow-up studies are needed to study the reduction of cold incidences and reduced duration of the illness by NPS. Based on these outcomes the full impact on the quality of life, life expectation and health care costs, can then be calculated.
It is important to mention that the immune system, which is the target of the proposal, has a much higher impact via many other immune-related diseases. As these other diseases cannot be studied within the course of a short intervention study (as will be performed in the present EU project) these other diseases have not been mentioned before. However an overload of publications has shown that an overall improved immune status can also lead to a reduced risk on cancer, Alzheimer, Crohn’s disease, arthritis, diabetes etc. Based on in vitro research, new hypotheses have been generated on how NPS can contribute to certain immune related mechanisms. We therefore hope to study these NPS in order to support in reducing incidences of immune related diseases or reduce recurrence of diseases when treated. We think that these can be efficiently combined with current treatments and make this more efficient. Time will tell, and future research will learn.
The impact for the research partners
The goal of the FP7 framework, and particular also this call in the KBBE domain, aimed to boost also SMEs as accelerators of our economic growth. In our project, we included SMEs, both food ingredient developers/supplies and those more towards technology services in the food and health area. Both had much benefit from this project.
The food ingredient suppliers (BioActor, Swedish Oat Fibre, Immitec) have gathered knowledge on the biochemical composition, how the NPSs behave when applied in food products, have gained knowledge on bioactive effects towards the immune system and towards gut fermentation and effects on microbiota, obtained in vivo data for which some of the NPS showed somewhat more potent in inducing effects than others etc. But also when no promising effects are found, these results are still of interest indicating that either the product formulation chosen was not bioactive or that immune health is not the most likely direction to continue and the health effects could better be studied towards other health aspects like bowel function or reduction of pathogens. All these research directions could not be addressed in this project. The research network as a whole and the interaction with the industry organised by the FibeBiotics consortium was also seen as important for the SMEs and have improved their position. The same counts for Winclove, the company mainly developing probiotics, which has been able to study the interaction of probiotics with NPSs in an early stage and could scan this area for interesting new developments that could help to strengthen their portfolio either by using the NPSs studied or by applying the research tools developed to use for substantiation of probiotics which are also studied for their immune support.
The technology and service providing SMEs have also greatly benefited from the project. CRC Kiel performed and organised the human trial perfectly and when results are published can help to establish a good market position in order to gain new assignments. Genetic Analysis has been able to further optimise the GA-map technology and made it dedicated towards European elderly which might help also in the future to develop diagnostic test to diagnose abnormal microbiota composition in relation to diseases. This project generated many samples which now can be used to link microbiota and, therefore, the GA-map technology to health status both at baseline and as a result of the intervention. ProDigest has already had a well-established position in the area of gut intestinal simulation as the SHIME model is, together with the TIM model at TNO, one of the best in full simulation of the intestinal tract. This model has been further developed in this project and experiments formed have shown good correlation with effects found in vivo, further supporting the potential of this model to replace animal and human trials. The aim to develop this SHIME model further by including in line exposures to human cells did not work out yet and Prodigest will invest in that further. Besides SHIME experiments, the group has been able to gain more knowledge on cell culture research that could be performed off line with samples derived from the SHIME. R-consultancy was involved in building the data base and accompanied software to simplify uploading, warehousing and subtraction of data. Due to problems with compatibility of systems and limited flexibility of the system, the database was less optimal than expected and also none of the partners has implemented the database structure for their own in-company warehousing. This might be due to the small size of some of the companies or the availability to other commercial software system developers that jump into this niche.
All research partners, both universities and institutes have benefited from this project to strengthen their knowledge and develop new research tools and strategies in the area of gut immune effects.
The results obtained for the different NPS compounds within this consortium can be used as a reference to compare and characterise effect of other NPSs in order to develop other functional food products directed towards gut and immune function. The appeared and to-appear publications will help them in exposing their knowledge and to attract new research capacity to continue their research.
Dissemination towards the food industry and the FibeBiotics Industrial Platform
EU funding should support not only a limited number of partners involved in a project but to our opinion result in useful tools, knowledge and strategies which can be applied broadly by the community. In this food and health related research, mainly the food industry should be the target for dissemination. As not all food companies could become project partner, we established an Industrial Platform (InP) for which we invited all industrial stakeholders interested to be informed on the progress of this project. This would generate short links with groups which can answer questions from the industry directly, will support networking with other companies with an interest in fibres and connect ingredient suppliers with end users.
The FibeBiotics Industrial Platform was established to ensure that knowledge and new developed technologies/research tool, once proven to be suitable for application, are indeed used to the benefit of industry. The Platform was continuously open for new interested members. At the end of the project, the Industrial Platform included 42 members which all signed a participation document and indicated their specific interests.
For the Industrial Plaform members, we established a members-only website, which contained among others the information on the project’s progress and (related) publications. The Platform members also received Newsletters on a regular basis and the FibeBiotics consoritum organised annual Industrial Platform meetings with presentations from the FibeBiotics partners and also from the industry themselves to exchange ideas on in-house research performed by the companies, their products portfolio and their market strategies. The InP meetings have been organised in:
• Ghent, Belgium (2012), including a visit to the facilities of partner ProDigest;
• Bologna, Italy (2013) alongside the EFFoST Annual Meeting: Bio-based technologies in the context of European food innovation;
• Cork, Ireland (2014), including a visit to cheese factory of Teagasc;
• Paris, France (2015), including a visit to the lab facilities of Institute Pasteur;
• Wageningen, The Netherlands (2016), including a visit to the lab facilities of Wageningen University & Research.
These events were very well visited and appreciated by the InP members.
The impact for the food industry could be summarised as that FibeBiotics has brought them in contact with new strategies and research tools, tested human trial designs supported by EFSA which probably has helped them to consider whether to pursue this direction, or learn from this and change future plans. They could reflect and improve in-house research with the new knowledge generated from the FibeBiotics project at a very early stage (as many major results still need to be published but they already have been shared with the industry that now can benefit from this new insight knowledge).
For some a pity and for some a relief, but we have to realise that we have not been able to come closer towards an immune-related health claim for one or more NPSs. If we had results that would make a proper EFSA dossier, this would have been authorised and could have been a major breakthrough. Such an authorised claim and the way the dossier content is generated, would have made it potentially easier - also for other NPSs - to obtain an EFSA-approved health claim, in case of similar results by similar human trials. The pilot trial and the pivotal trial could be used as supportive evidence in the future but, as effects are not statistically significant, it will not be sufficient. We also have not been able to identify other immune related biomarkers which might become useful to substantiate effect in the future. However, the biomarker for barrier function might still be a valid approach, but we are not sure whether food compounds really can counteract the effect of a stressor and how biologically relevant this is for intestinal function.
However, many research tools, knowledge (such as those related to microbiota and microbiota related metabolites) can be expected to become very useful in the future when more relations to health impact are known.
Impact on the European policy on immune claims
Digging into the area of vaccination efficacy as one of the only, by EFSA, accepted models to study potential immune supportive effect led to the realisation that very little evidence is available on adjuvants in a healthy population. Almost all of the authorised health claims in this area are based on deficiencies of a certain nutrient in populations (in developing countries) that really suffered from a deficiency. That is how some vitamins and minerals reached their immune claim; these were not achieved by showing adjuvant effects on vaccination efficacy as the result of extra intake in a healthy (elderly) population. And even when studies were performed on populations with malnutrition, then also the supplementation of the nutrient (e.g. zinc) did not always show this increase in vaccination efficacy.
Some studies have shown positive effects of a food intervention on vaccination efficacy but these have been hardly repeated in order to analyse how reliably the effects can be monitored. One possibility might be that the type of vaccination model used is of main importance. We used a flu vaccination model for which effect will be influenced by vaccination history, encountered flu infections, and health status. So we concluded that it would be wise to use vaccination models that a person did not have before (nor the vaccination nor the disease which also can boost memory) or use suboptimal doses of a vaccination. In order to detect a positive effect of an intervention you need a window to quantify the improvement of antibody titers. When using very optimised vaccinations the vaccine treatment itself might induce often towards its maximum and does not leave room for improvement by an intervention. However it is currently not known how the EFSA will judge studies in which suboptimal vaccination models are used> The hopes are that this will be accepted and evaluated in next guidelines.
With the new updated guidelines, the EFSA opened up new potential valid strategies as they concluded: ”studies conducted in subjects with a disease may be used to substantiate function claims for the general population or subgroups thereof (without the disease) as long as the effect of the food/constituent on the body function which is named in the claim is expected to occur in subjects without the disease and a rationale is given for such expectations”. This might be a strategy to follow in the future to study gut and immune supportive effects, although the interpretation of ‘expected to occur in subjects without the disease’ can still be explained in different ways and therefore forms a narrow basis for companies to target their human trial designs. When this rule has been applied and implemented in real evaluations it will become more clear what the possibilities are to substantiate health effect making use of a somewhat diseased or medically compromised population.
All current trial designs and statistical method in the area of health claims are still completely leaning on the strategies used for medicines based on diseased subjects. However, effects of pharmaceuticals can be very individually determined, making that within the population there are responders and non-responders. Not without a purpose, medical research is focussing a lot on strategies for personalised medicatio. When treatments have small effects (low specific activity), as can be the case for food that might subtly correct homeostasis, effects based on the average values will require very large sample sizes. This will require expensive trials that might be possible for drug research (due to recovery of investment) but this will be much more difficult in the food area. When it is generally accepted that a randomly choosen population included in a trial will consist of responders and non-responders new concepts are needed to evaluate the results and the conclusions on benefit to the general population (or to responders within the population). We hope that EFSA guidelines in the future will take these considerations into account so that food intervention studies that show positive effect for specific individuals in the population but does not harm the rest of the population still can be used to substantiate (personal) health effect.
List of Websites:
www.fibebiotics.eu
Dr. Jurriaan Mes, project coordinator
Email: jurriaan.mes@wur.nl
Tel: +31 317 48 11 74
