Professor Konstantinos Gerasimidis, University of Glasgow, UK

A wealth of scientific knowledge about diet and the gut microbiome (GMB) in health and disease has emerged as a result of a better understanding of the etiopathogenesis of non-communicable diseases (NCDs) and the development of molecular and computational tools to characterize the GMB. Nevertheless, the study of the GMB in health and disease remains complex because the human gut microbiota (GM) itself is an intricate system with as many bacteria as cells in the human body, and with a highly diverse composition, unique to each individual. Moreover, its pangenome contains 100 times more genes than that of the host, translating into an enormous functional capacity, able to carry out metabolic activities, some of which are no longer present in the host. Due to this complexity, there is no consensus definition of a “healthy microbiota” as it can vary depending on the context. A review by Van Hul comprehensively addresses this issue (Van Hul et al., Reference Van Hul, Cani, Petitfils, De Vos, Tilg and El-Omar2024). Other aspects that complicate this area of research include the fact that many of the studies that have shown associations between the GM and the aetiology of NCDs have been conducted in animal models. When studies focused on humans, they have mostly been carried out in people who already have the disease (making it difficult to distinguish whether the GM changes are the cause or the consequence of the disease). Lastly, there are many confounding factors when studying the (healthy) GM, the most important of which is diet.

Diet is a major determinant of GM composition and function in physiological and pathological contexts. During the first 14 months of life, the feeding mode (breast- or formula-feeding) determines infant GM development and is associated with metabolic capacity (Stewart et al., Reference Stewart, Ajami, O’Brien, Hutchinson, Smith, Wong, Ross, Lloyd, Doddapaneni, Metcalf, Muzny, Gibbs, Vatanen, Huttenhower, Xavier, Rewers, Hagopian, Toppari, Ziegler, She, Akolkar, Lernmark, Hyoty, Vehik, Krischer and Petrosino2018). Another example is how differences between the diets of children living in a rural environment in Burkina Faso and those living in an urban area in Italy result in vastly different GM richness, diversity, and metabolic capacity (De Filippo et al., Reference De Filippo, Cavalieri, Di Paola, Ramazzotti, Poullet, Massart, Collini, Pieraccini and Lionetti2010). A more recent study examined how different major diets and their main bioactive compounds influence GM composition and impact health (Ross et al., Reference Ross, Patangia, Grimaud, Lavelle, Dempsey, Ross and Stanton2024). In the context of a disease, it has been shown that children with intestinal failure, and whose nutrition depends on intravenous feeding, have changes in the composition and metabolic activity of their GM, compared to healthy controls (Neelis et al., Reference Neelis, de Koning, Hulst, Papadopoulou, Kerbiriou, Rings, Wijnen, Nichols and Gerasimidis2022).

Of all the bioactive dietary components that influence the GM, fibre has the most profound effect. Different dietary fibres are fermented differently or are not fermented at all (Gerasimidis et al., Reference Gerasimidis, Nichols, McGowan, Svolos, Papadopoulou, Kokkorou, Rebull, Bello Gonzalez, Hansen, Russell and Gaya2022b), having varying effects on gut and metabolic health. It is therefore important to consider what the goal of the dietary therapy is when advising patients to consume fibre. Fermented foods (FFs) are another category that has been shown to increase GM diversity and decrease inflammatory biomarkers (Wastyk et al., Reference Wastyk, Fragiadakis, Perelman, Dahan, Merrill, Yu, Topf, Gonzalez, Van Treuren, Han, Robinson, Elias, Sonnenburg, Gardner and Sonnenburg2021). As the effects of FFs and fibre on the GM differ, it may be important to include both in our diets. Diet also includes food additives that may adversely affect the human GM. In a study conducted in an in vitro gut model using a fibre mixture as substrate, the addition of emulsifiers altered short-chain fatty acid (SCFA) production and promoted shifts in GM diversity with the growth of potential pathobionts (Gerasimidis et al., Reference Gerasimidis, Bryden, Chen, Papachristou, Verney, Roig, Hansen, Nichols, Papadopoulou and Parrett2020). Dietary emulsifiers have also been shown to promote dysbiosis and increase inflammation and colitis in animals predisposed to this condition (Chassaing et al., Reference Chassaing, Koren, Goodrich, Poole, Srinivasan, Ley and Gewirtz2015), but there are currently no human studies linking them.

GM has been implicated in the pathogenesis of several NCDs through its interaction with diet. However, the relationship between GM, diet, and NCDs is extremely complex, and it is important to examine the direction of causality. A number of studies have shown that many of the pathogenic effects attributed to the GM in certain diseases are actually cases of reverse causation. Examples include celiac disease, autism spectrum disorder (ASD), and obesity. In celiac disease, one study showed that dysbiosis was absent in children with new-onset celiac disease, and it was the gluten-free diet that led to changes in GM (Zafeiropoulou et al., Reference Zafeiropoulou, Nichols, Mackinder, Biskou, Rizou, Karanikolou, Clark, Buchanan, Cardigan, Duncan, Wands, Russell, Hansen, Russell, McGrogan, Edwards, Ijaz and Gerasimidis2020). Similarly, in ASD, direct associations between the GMB and ASD were negligible, but the restricted and less diverse diets consumed by patients with ASD were strongly associated with changes in the GMB (Yap et al., Reference Yap, Henders, Alvares, Wood, Krause, Tyson, Restuadi, Wallace, McLaren, Hansell, Cleary, Grove, Hafekost, Harun, Holdsworth, Jellett, Khan, Lawson, Leslie, Frenk, Masi, Mathew, Muniandy, Nothard, Miller, Nunn, Holtmann, Strike, de Zubicaray, Thompson, McMahon, Wright, Visscher, Dawson, Dissanayake, Eapen, Heussler, McRae, Whitehouse, Wray and Gratten2021). In the case of obesity, it is widely accepted that the higher ratio of Bacillota (formerly Firmicutes) to Bacteroidota in the GM contributes to weight gain in mice and humans by increasing the capacity to extract energy from the diet (Turnbaugh et al., Reference Turnbaugh, Ley, Mahowald, Magrini, Mardis and Gordon2006). Professor Gerasimidis analysed the causal direction of this relationship by comparing the GM and faecal SCFAs in children with “bacteria-induced” obesity and those with genetic obesity (Prader–Willi syndrome). If the GM caused obesity, the causal direction would be different in both groups. The study showed that SCFA levels were higher in obese compared to non-obese participants, but there were no differences between the two types of obesity, concluding that the main reason obese people have different GMs is because they eat more and eat differently (unpublished). The above is not to dismiss the role of the GM and its interaction with diet in the pathogenesis of NCDs. For example, studies suggest that an increase in trimethylamine N-oxide (TMAO), an oxidized derivative of trimethylamine (TMA) produced by the GM metabolism of excess dietary phosphatidylcholine, contributes to the development of atherosclerosis (Heianza et al., Reference Heianza, Ma, Manson, Rexrode and Qi2017).

In the view of Professor Gerasimidis, the most important features of the GM are its utility as a disease-specific signature for diagnosing NCDs or predicting responses to drugs and foods, and its modifiability by pharmacological/dietary interventions to relieve patients’ symptoms. A study in children with Crohn’s disease (CD) showed that it is possible to predict the response to exclusive enteral nutrition with high sensitivity and specificity based on GM signatures (Nichols et al., Reference Nichols, Briola, Logan, Havlik, Mascellani, Gkikas, Milling, Ijaz, Quince, Svolos, Russell, Hansen and Gerasimidis2024). This information can be translated into clinical practice. Regarding symptom relief through dietary manipulation, it is certainly a plausible therapy for certain microbiome-associated diseases, but it is important to keep in mind that one size does not fit all (Gerasimidis et al., Reference Gerasimidis, Gkikas, Stewart, Neelis and Svolos2022a). For example, while the use of probiotics has improved certain symptoms in irritable bowel syndrome (IBS) (although different strains at different doses may have different effects) (Goodoory et al., Reference Goodoory, Khasawneh, Black, Quigley, Moayyedi and Ford2023), it has failed to improve disease outcomes in CD (Limketkai et al., Reference Limketkai, Akobeng, Gordon and Adepoju2020; Lorentz and Müller, Reference Lorentz and Müller2022).

Next, Professor Gerasimidis explained the evolution of nutritional therapy. The traditional approach is one-size-fits-all, and we are moving from that to a “stratified approach” where patients are grouped based on certain clinical characteristics (e.g., luminal vs. fistulizing CD), and towards a “precision therapy” where diseases are treated based on individual “omic” biology and immunophenotype, leading to precision nutrition therapy. The postprandial glycaemic response to carbohydrate consumption varies between individuals. One study found that personal glycaemic response can be predicted from GMB, and its data were used to develop and validate an algorithm that may be useful in future personalized management of prediabetic and diabetic persons (Zeevi et al., Reference Zeevi, Korem, Zmora, Israeli, Rothschild, Weinberger, Ben-Yacov, Lador, Avnit-Sagi, Lotan-Pompan, Suez, Mahdi, Matot, Malka, Kosower, Rein, Zilberman-Schapira, Dohnalová, Pevsner-Fischer, Bikovsky, Halpern, Elinav and Segal2015; Korem et al., Reference Korem, Zeevi, Zmora, Weissbrod, Bar, Lotan-Pompan, Avnit-Sagi, Kosower, Malka, Rein, Suez, Goldberg, Weinberger, Levy, Elinav and Segal2017). Another example is a study of the response to a diet low in fermentable oligosaccharides, disaccharides, monosaccharides, and polyols (FODMAP) in IBS patients. The GM of the patients before starting the low-FODMAP diet differed from that of healthy controls, and the more divergent, the greater therapeutic benefit they received from the diet (Vervier et al., Reference Vervier, Moss, Kumar, Adoum, Barne, Browne, Kaser, Kiely, Neville, Powell, Raine, Stares, Zhu, De La Revilla Negro, Lawley and Parkes2022). Thus, it is important to analyse the GM before prescribing a diet with therapeutic goals, but there is currently no causal evidence to recommend microbiome-based precision medicine for the management of NCDs.

Finally, Professor Gerasimidis is currently conducting the UC-TREAT trial, which builds on studies showing that certain GMB signatures of ulcerative colitis (UC) patients can identify responders/non-responders before biologics treatment (Ananthakrishnan et al., Reference Ananthakrishnan, Luo, Yajnik, Khalili, Garber, Stevens, Cleland and Xavier2017; Lee et al., Reference Lee, Plichta, Hogstrom, Borren, Lau, Gregory, Tan, Khalili, Clish, Vlamakis, Xavier and Ananthakrishnan2021). The study aims to improve the response of UC patients to anti-cytokine or anti-integrin therapy by feeding them a diet high in fibre, FF, and polyphenols, and expects to increase the response rate from 50% to 70–80%. Pre-conditioning a patient’s GMB to make it more responsive to a drug through specific microbiome-modifying dietary regimes is another strategy with high potential for future application.

Seeding and colonization of the early-life microbiome: consensus and debate?

While there exists currently a consensus on the importance of the development of a functional microbiome during the first 1,000 days, for example, driving the development of a functional immune system or the acquisition of microorganisms that help ferment (complex) carbohydrates to health-supporting SCFAs, there is no real consensus as to the optimal source of the microbes that will result in the development of a healthy holobiont. During the symposium, two possible research axes were highlighted, focusing on the importance of the environment versus human breast milk (HBM) as an important source for the colonization of the gut.

The role of human breast milk microbiota

According to Professor Flavia Indrio (University of Salento, Italy), human breast milk (HBM) is beneficial per se, because it is a unique food that contains all the macro- and micronutrients the infant needs, as well as bioactive compounds such as human milk oligosaccharides (which shape the infant’s GM through their bifidogenic effect), lactoferrin, immunomodulators, hormones, growth factors, nucleotides, stem cells, and microbes (Goldman, Reference Goldman2000). As it contains probiotics, prebiotics, and postbiotics, HBM can be called a “panbiotic.”

HBM composition constantly changes according to the development of each infant (personalized nutrition), for whom breastfeeding reduces the risk of infectious diseases, allergies, autoimmune diseases, malignancies, necrotizing enterocolitis, sudden infant death syndrome, and mortality, while promoting cognitive and immune development (Victora et al., Reference Victora, Bahl, Barros, França, Horton, Krasevec, Murch, Sankar, Walker and Rollins2016). Breastfeeding also benefits the mothers by reducing the risk of type 2 diabetes (T2D), breast cancer, and ovarian cancer. Breastfeeding also benefits society in many ways, for example, by reducing mortality and preventing disease in both mothers and children, reducing economic losses (UNICEF, 2018; World Health Organization, n.d.-a).

Many studies have shown that the HBM, with 10⁵–10⁸ colony-forming units (CFUs)/ml, is a source of microbes that help colonize the infant’s gut (Heikkilä and Saris, Reference Heikkilä and Saris2003; Martín et al., Reference Martín, Langa, Reviriego, Jimínez, Marín, Xaus, Fernández and Rodríguez2003, Reference Martín, Jiménez, Olivares, Marín, Fernández, Xaus and Rodríguez2006, Reference Martín, Heilig, Zoetendal, Smidt and Rodríguez2007a, Reference Martín, Heilig, Zoetendal, Jiménez, Fernández, Smidt and Rodríguez2007b, Reference Martín, Jiménez, Heilig, Fernández, Marín, Zoetendal and Rodríguez2009, Reference Martín, Maldonado-Barragán, Moles, Rodriguez-Baños, Campo, Fernández, Rodríguez and Jiménez2012; Jost et al., Reference Jost, Lacroix, Braegger and Chassard2013, Reference Jost, Lacroix, Braegger and Chassard2015; Stewart et al., Reference Stewart, Ajami, O’Brien, Hutchinson, Smith, Wong, Ross, Lloyd, Doddapaneni, Metcalf, Muzny, Gibbs, Vatanen, Huttenhower, Xavier, Rewers, Hagopian, Toppari, Ziegler, She, Akolkar, Lernmark, Hyoty, Vehik, Krischer and Petrosino2018). Their origin remains unclear, but research suggests that they come from the mother’s skin, breast tissue, gut, and the infant’s oral cavity (Le Doare et al., Reference Le Doare, Holder, Bassett and Pannaraj2018). The most common bacteria found in HBM are species from the genera Streptococcus, Staphylococcus, and Propionibacterium, followed by lactobacilli and Bifidobacterium spp. Strictly anaerobic microbes and butyrate-producing bacteria such as Faecalibacterium have also been isolated (Hunt et al., Reference Hunt, Foster, Forney, Schütte, Beck, Abdo, Fox, Williams, McGuire and McGuire2011). Of these, Bifidobacterium, the most abundant taxon in the GM of exclusively breastfed infants (O’Callaghan and van Sinderen, Reference O’Callaghan and van Sinderen2016), is of particular interest for infant health, given its role in immune cell education and butyrate production (Stuivenberg et al., Reference Stuivenberg, Burton, Bron and Reid2022). Regarding other microorganisms, the use of advanced sequencing techniques has allowed a more detailed analysis of the HBM in more detail, identifying Archaea, fungi, and viruses in addition to bacteria (Boix-Amorós et al., Reference Boix-Amorós, Martinez-Costa, Querol, Collado and Mira2017; Pannaraj et al., Reference Pannaraj, Li, Cerini, Bender, Yang, Rollie, Adisetiyo, Zabih, Lincez, Bittinger, Bailey, Bushman, Sleasman and Aldrovandi2017; Moossavi et al., Reference Moossavi, Sepehri, Robertson, Bode, Goruk, Field, Lix, de Souza, Becker, Mandhane, Turvey, Subbarao, Moraes, Lefebvre, Sears, Khafipour and Azad2019; Dinleyici et al., Reference Dinleyici, Pérez-Brocal, Arslanoglu, Aydemir, Ozumut, Tekin, Vandenplas, Moya and Dinleyici2020, Reference Dinleyici, Pérez-Brocal, Arslanoglu, Aydemir, Sevuk Ozumut, Tekin, Vandenplas, Moya and Dinleyici2021). Although information on the HBM mycobiota is limited, three studies analysing the fungi at the genus level found that the most abundant taxa were Malassezia and Saccharomyces, which varied with HBM stage, delivery mode, gestational age, and birth weight (Boix-Amorós et al., Reference Boix-Amorós, Martinez-Costa, Querol, Collado and Mira2017, Reference Boix-Amorós, Puente-Sánchez, du Toit, Linderborg, Zhang, Yang, Salminen, Isolauri, Tamames, Mira and Collado2019; Dinleyici et al., Reference Dinleyici, Pérez-Brocal, Arslanoglu, Aydemir, Ozumut, Tekin, Vandenplas, Moya and Dinleyici2020). Similarly, information on the HBM virome is limited, but a study examining the transient human milk (4–10 days postpartum) and infant stool viromes of mother–infant pairs found that there were some viruses shared by both samples, presumably as a result of the infant’s ingestion of HBM (Pannaraj et al., Reference Pannaraj, Ly, Cerini, Saavedra, Aldrovandi, Saboory, Johnson and Pride2018). Most viruses identified were bacteriophages, which are likely to help shape the infant’s GM, while a minority were eukaryotic viruses (Pannaraj et al., Reference Pannaraj, Ly, Cerini, Saavedra, Aldrovandi, Saboory, Johnson and Pride2018; Mohandas and Pannaraj, Reference Mohandas and Pannaraj2020). The importance of the gut immune system is that it drives the immunity of the whole body; therefore, its proper training is essential.

HBM participates in this process because the neonatal immune system must be trained through complex interactions with the early GM, which is allowed to grow by the presence of specific HBM-derived substrates.

As mentioned above, according to Professor Edwards, there are indeed many theories about the origin of the infant’s GM. Besides the exposure to maternal vaginal-, skin-, and faecal microbiota at birth, prenatal exposure to placental microbiota (Collado et al., Reference Collado, Rautava, Aakko, Isolauri and Salminen2016) and HBM microbiota is often mentioned. Related to the contribution of the placental microbiome, Professor Edwards and others conducted a systematic review that closely examined the evidence for a role for the placental microbiome of healthy mothers in colonizing the infant’s gut. Their findings found that while this origin had previously seemed plausible, the reassessment found that most studies had methodological problems and were highly heterogeneous (Gil et al., Reference Gil, Rueda, Ozanne, van der Beek, van Loo-Bouwman, Schoemaker, Marinello, Venema, Stanton, Schelkle, Flourakis and Edwards2020). Other studies suggested that the detection of bacteria in the placenta was due to contamination during sample collection and processing (de Goffau et al., Reference de Goffau, Lager, Sovio, Gaccioli, Cook, Peacock, Parkhill, Charnock-Jones and Smith2019; Sterpu et al., Reference Sterpu, Fransson, Hugerth, Du, Pereira, Cheng, Radu, Calderón-Pérez, Zha, Angelidou, Pennhag, Boulund, Scheynius, Engstrand, Wiberg-Itzel and Schuppe-Koistinen2021), while a systematic review in 2022 supported the existence of a low-biomass placental microbiota in healthy pregnancies, with some of the microbial taxa in the placenta might have an oral origin (Zakis et al., Reference Zakis, Paulissen, Kornete, Kaan, Nicu and Zaura2022). The high risk of quality bias for the majority of the included studies indicates that the results of individual papers should be interpreted with caution.

Microbial transfer through HBM, as discussed by Professor Indrio, was also questioned by Professor Edwards, as some studies found strains of bacteria that were common between the HBM and the infant’s stool, but it was not certain whether they were transferred from the HBM to the infant and whether they actually colonized the infant’s gut. A systematic review conducted by Professor Edwards found that some of these studies (Grönlund et al., Reference Grönlund, Gueimonde, Laitinen, Kociubinski, Grönroos, Salminen and Isolauri2007; Martín et al., Reference Martín, Jiménez, Heilig, Fernández, Marín, Zoetendal and Rodríguez2009, Reference Martín, Maldonado-Barragán, Moles, Rodriguez-Baños, Campo, Fernández, Rodríguez and Jiménez2012; Hunt et al., Reference Hunt, Foster, Forney, Schütte, Beck, Abdo, Fox, Williams, McGuire and McGuire2011; Cabrera-Rubio et al., Reference Cabrera-Rubio, Collado, Laitinen, Salminen, Isolauri and Mira2012; González et al., Reference González, Maldonado, Martín, Mandomando, Fumadó, Metzner, Sacoor, Fernández, Macete, Alonso, Rodríguez and Menendez2013) had not monitored true transfer or had methodological problems (Edwards et al., Reference Edwards, Van Loo-Bouwman, Van Diepen, Schoemaker, Ozanne, Venema, Stanton, Marinello, Rueda, Flourakis, Gil and Van der Beek2022).

After excluding potentially flawed papers and keeping those that really showed “transfer,” which needed identification of strains or specific markers, it was concluded that there is evidence for the presence of Staphylococcus, Streptococcus, Lactococcus, Pseudomonas, lactobacilli, bifidobacteria, and Archaea in HBM (Togo et al., Reference Togo, Grine, Khelaifia, Robert, Brevaut, Caputo, Baptiste, Bonnet, Levasseur, Drancourt, Million and Raoult2019; Lugli et al., Reference Lugli, Duranti, Milani, Mancabelli, Turroni, Alessandri, Longhi, Anzalone, Viappinai, Tarracchini, Bernasconi, Yonemitsu, Bode, Goran, Ossiprandi, van Sinderen and Ventura2020) but little evidence of colonization of the infant gut. Two studies clearly identified strains shared by HBM and infant stools (Benito et al., Reference Benito, Lozano, Jiménez, Albújar, Gómez, Rodríguez and Torres2015; Kordy et al., Reference Kordy, Gaufin, Mwangi, Li, Cerini, Lee, Adisetiyo, Woodward, Pannaraj, Tobin and Aldrovandi2020). One study examining antibiotic-resistant genes (ARGs) in newborn stool showed shared resistomes with maternal gut and HBM microbiota (Pärnänen et al., Reference Pärnänen, Karkman, Hultman, Lyra, Bengtsson-Palme, Larsson, Rautava, Isolauri, Salminen, Kumar, Satokari and Virta2018), but no evidence of colonization of the infant gut. In fact, the opposite has been found. Makino et al. found strains of Bifidobacterium longum shared by mother–infant pairs (Makino et al., Reference Makino, Kushiro, Ishikawa, Muylaert, Kubota, Sakai, Oishi, Martin, Ben Amor, Oozeer, Knol and Tanaka2011), but later reported that they appeared in the infant stool before the HBM intake (Makino et al., Reference Makino, Martin, Ishikawa, Gawad, Kubota, Sakai, Oishi, Tanaka, Ben-Amor, Knol and Kushiro2015). Another study found that probiotic bacteria given to infants appeared later in the HBM (Treven et al., Reference Treven, Mahnič, Rupnik, Golob, Pirš, Matijašić and Lorbeg2019). Infant oral microbiota colonization of the HBM is plausible because infants pull the breast into the back of their mouths and the suction creates a vacuum in the mammary ducts (Edwards et al., Reference Edwards, Van Loo-Bouwman, Van Diepen, Schoemaker, Ozanne, Venema, Stanton, Marinello, Rueda, Flourakis, Gil and Van der Beek2022). On the other hand, it is reasonable to assume that at least some of the bacteria in the infant’s mouth come from the HBM. One study found the same bacterial species in the infant’s mouth and the pre-colostrum before contact with the newborn in mother–infant pairs (Ruiz et al., Reference Ruiz, Bacigalupe, García-Carral, Boix-Amoros, Argüello, Silva, de Los Angeles Checa, Mira and Rodríguez2019). Taken together, this suggests that there is a bidirectional exchange of bacterial material between the HBM microbiota and the infant’s oral microbiota, and between the HBM microbiota and the infant’s faecal microbiota (Edwards et al., Reference Edwards, Van Loo-Bouwman, Van Diepen, Schoemaker, Ozanne, Venema, Stanton, Marinello, Rueda, Flourakis, Gil and Van der Beek2022).

The possible long-term effect on health of the early microbiota development

As pointed out by Professor Indrio, during the first 1,000 days of life, the maturation of the GM and the immune system work in concert to set the stage for future immune competence (Wopereis et al., Reference Wopereis, Oozeer, Knipping, Belzer and Knol2014). Thus, factors such as maternal health, diet, lifestyle, mode of delivery, gestational age, and geographic location that have the potential to cumulatively alter the microbial composition of the GM, as well as factors that cause dysregulation of the microbiota-immunity interplay during infancy, with the potential to lead to immune dysregulation and to disease, such as autoimmune diseases and those related to gut barrier integrity. Finally, the HBM is involved in nutritional programming, the ability of compounds normally present in the diet or molecules synthesized de novo in our organism, to modulate and support early childhood development (Gialeli et al., Reference Gialeli, Panagopoulou, Liosis and Siahanidou2023). Exosomes and stem cells contained in the HBM can alter the expression of our genes, and thus determine the short- and long-term health or disease status. Exosomes carrying microRNAs (miRNAs) are produced by mammary gland cells and released into the milk. They are taken up by the infant’s intestinal cells and, once in the systemic circulation, can cross the blood–brain barrier (BBB). In the brain, exosomes release miRNAs that bind to mRNA and modulate neuronal gene expression. Stem cells can also cross the BBB and differentiate into neural cells or release bioactive molecules that have epigenetic effects. Interestingly, the infant’s GM may also induce epigenetic changes or regulate key molecules involved in brain development. Together, these influence the development and plasticity of the infant’s brain (Gialeli et al., Reference Gialeli, Panagopoulou, Liosis and Siahanidou2023).

Professor Indrio concluded her presentation by paraphrasing a quote from the economist James J. Heckman: “The highest rate of return for any society is to invest in the first 1,000 days of an infant’s life; if we lay the foundation for a healthy infant, we lay the foundation for a strong society,” to emphasize that breastfeeding is the cornerstone for preventing many short- and long-term health risks for infants and mothers, impacting society as a whole. Prevention requires large investments, but the returns to society and the economy are also large.

Professor Edwards concluded by describing how information on the development of the infant GM and the factors that alter it was already used by some researchers to create a well-being index that can predict the overall health of the children over the first 5 years. Five distinct microbiota communities with distinct health trajectories associated with infant exposures were defined using clustering and trajectory modelling based on GM and health data for the first 2 years of life of 1,000 term-born infants (Hickman et al., Reference Hickman, Salonen, Ponsero, Jokela, Kolho, de Vos and Korpela2024). This index could be valuable for improving the prediction of disease risk and interventions to improve health, and could steer future research and communication towards future mothers.

Associate Professor Silvia Turroni, University of Bologna, Italy

The exposome encompasses the totality of exposures (specific or general external factors and internal biological responses) experienced by an individual from conception to death (Wild, Reference Wild2005; Vrijheid, Reference Vrijheid2014; Zhang et al., Reference Zhang, Arora, Chaleckis, Isobe, Jain, Meister, Melén, Perzanowski, Torta, Wenk and Wheelock2019) and influences human, animal, plant, and environmental health, all interconnected and integrated under the concept of One Health (World Health Organization, n.d.-b).

The exposome drives variation in the GMB (Levin et al., Reference Levin, Sitarik, Havstad, Fujimura, Wegienka, Cassidy-Bushrow, Kim, Zoratti, Lukacs, Boushey, Ownby, Lynch and Johnson2016; Gacesa et al., Reference Gacesa, Kurilshikov, Vich Vila, Sinha, Klaassen, Bolte, Andreu-Sánchez, Chen, Collij, Hu, Dekens, Lenters, Björk, Swarte, Swertz, Jansen, Gelderloos-Arends, Jankipersadsing, Hofker, Vermeulen, Sanna, Harmsen, Wijmenga, Fu, Zhernakova and Weersma2022; Oesterle et al., Reference Oesterle, Ayeni, Ezekiel, Berry, Rompel and Warth2024). Many studies have compared the microbiome–exposome interaction of traditional (hunter–gatherer and rural) and urban communities to understand the extent to which GM and host co-evolution are responsible for our physiological flexibility and environmental adaptation (De Filippo et al., Reference De Filippo, Cavalieri, Di Paola, Ramazzotti, Poullet, Massart, Collini, Pieraccini and Lionetti2010; Yatsunenko et al., Reference Yatsunenko, Rey, Manary, Trehan, Dominguez-Bello, Contreras, Magris, Hidalgo, Baldassano, Anokhin, Heath, Warner, Reeder, Kuczynski, Caporaso, Lozupone, Lauber, Clemente, Knights, Knight and Gordon2012; Tyakht et al., Reference Tyakht, Kostryukova, Popenko, Belenikin, Pavlenko, Larin, Karpova, Selezneva, Semashko, Ospanova, Babenko, Maev, Cheremushkin, Kucheryavyy, Shcherbakov, Grinevich, Efimov, Sas, Abdulkhakov, Abdulkhakov, Lyalyukova, Livzan, Vlassov, Sagdeev, Tsukanov, Osipenko, Kozlova, Tkachev, Sergienko, Alexeev and Govorun2013; Schnorr et al., Reference Schnorr, Candela, Rampelli, Centanni, Consolandi, Basaglia, Turroni, Biagi, Peano, Severgnini, Fiori, Gotti, De Bellis, Luiselli, Brigidi, Mabulla, Marlowe, Henry and Crittenden2014; Martínez et al., Reference Martínez, Stegen, Maldonado-Gómez, Eren, Siba, Greenhill and Walter2015; Obregon-Tito et al., Reference Obregon-Tito, Tito, Metcalf, Sankaranarayanan, Clemente, Ursell, Zech Xu, Van Treuren, Knight, Gaffney, Spicer, Lawson, Marin-Reyes, Trujillo-Villarroel, Foster, Guija-Poma, Troncoso-Corzo, Warinner, Ozga and Lewis2015; Zhang et al., Reference Zhang, Guo, Xue, Sun, Zhang, Wang, Wang, Wang, Xu, Cao, Xu, Lv, Zhong, Chen, Qimuge, Menghe, Zheng, Zhao, Chen and Zhang2015; Gomez et al., Reference Gomez, Petrzelkova, Burns, Yeoman, Amato, Vlckova, Modry, Todd, Jost Robinson, Remis, Torralba, Morton, Umaña, Carbonero, Gaskins, Nelson, Wilson, Stumpf, White, Leigh and Blekhman2016). A comparative study on the Hadza hunter–gatherer community in Tanzania and urban Italians showed that the Hadza GM was highly diverse, enriched in pathways involved in complex polysaccharide degradation, had high numbers of fibrolytic and opportunistic bacteria, lacked bifidobacteria, and had a resistome similar to that of soil. In contrast, the Italian GM was enriched in pathways involved in xenobiotic metabolism, had a higher number of mobile genetic elements, and its resistome showed a food-derived (“farm to fork”) origin (Rampelli et al., Reference Rampelli, Schnorr, Consolandi, Turroni, Severgnini, Peano, Brigidi, Crittenden, Henry and Candela2015). This was also confirmed by a study conducted in Italy as part of the Controlling Microbiomes Circulation for Better Food Systems (CIRCLES) project (https://circlesproject.eu/), which found that microorganisms and ARGs spread during the production process in pig and poultry farming systems, and that a relevant fraction of ARGs associated with mobile genetic elements favoured their spread to human-associated microorganisms in the population (Scicchitano et al., Reference Scicchitano, Babbi, Palladino, Turroni, Mekonnen, Laczny, Wilmes, Leekitcharoenphon, Castagnetti, D’Amico, Brigidi, Savojardo, Manfreda, Martelli, De Cesare, Aarestrup, Candela and Rampelli2024a, Reference Scicchitano, Leuzzi, Babbi, Palladino, Turroni, Laczny, Wilmes, Correa, Leekitcharoenphon, Savojardo, Luise, Martelli, Trevisi, Aarestrup, Candela and Rampelli2024c).

Another study compared the Bassa, a rural population from Nigeria, with the local urban population and found that ancient microbiome and metabolome traits were progressively lost with urbanization (Ayeni et al., Reference Ayeni, Biagi, Rampelli, Fiori, Soverini, Audu, Cristino, Caporali, Schnorr, Carelli, Brigidi, Candela and Turroni2018).

Unexpectedly, the GM of the Bassa infants was overall adult-like, highly diverse, and essentially devoid of bifidobacteria, which can be explained by the extensive sharing of all aspects of daily life, resulting in a high degree of microbial dispersal that allows the human microbiome to behave as a meta-community, thus eliminating the differences between infant and adult microbiomes. Thus, assumptions about a healthy microbiome should not blindly follow studies in Western populations but be based on the specific human ecological context.

Another example of GM differences resulting from lifestyle adaptation relates to xenobiotics: shotgun metagenomic analysis of faecal samples from individuals aged 22 to over 100 years revealed an age-dependent progressive increase in the number of genes involved in xenobiotic biodegradation and metabolism, with a concomitant decrease in genes involved in carbohydrate metabolism. Most strikingly, the gut of centenarians and semi-supercentenarians (i.e., individuals living to 105 years and older) contained an increased number of genes involved in the metabolism of chemicals derived from industrial production, municipal waste, and indoor consumer products (Rampelli et al., Reference Rampelli, Soverini, D’Amico, Barone, Tavella, Monti, Capri, Astolfi, Brigidi, Biagi, Franceschi, Turroni and Candela2020), likely reflecting a new equilibrium with their environment.

But how is the microbiome “modernizing” and what are the underlying reasons? The Multiple-Hit Hypothesis (Sonnenburg and Sonnenburg, Reference Sonnenburg and Sonnenburg2014) postulates that the development of agriculture, the industrialization of food production, and the introduction of processed sanitized foods have cumulatively reduced the diversity and amount of dietary microbiota-accessible carbohydrates during human evolution. This has led to a loss of microbial diversity and members of the ancestral microbiota, including the so-called “old friends,” and is now tipping towards dysbiosis, predisposing us to NCDs. The “Old Friend Hypothesis” (Rook, Reference Rook2010) refers to exposure to microbes that were present during the evolution of the human immune system in hunter–gatherer times.

More recently, the term “microbiota insufficiency syndrome” has been introduced to describe the loss of co-evolved microbial community members, such as “volatile and/or associated negatively with industrialized societies of humans” (VANISH) taxa, counterbalanced by the gain of the “bloom or selected in societies of urbanization/modernization” (BloSSUM) taxa, many of which are mucolytic and can provoke maladaptive host responses under certain conditions, explaining many of the associated NCDs (Sonnenburg and Sonnenburg, Reference Sonnenburg and Sonnenburg2019).

The diversity of microbial exposures, especially during infancy, is key as it interacts with the regulatory systems that keep the immune system in balance (Bloomfield et al., Reference Bloomfield, Rook, Scott, Shanahan, Stanwell-Smith and Turner2016). Recent studies suggest that it is possible to restore our “old friends” and microbiota diversity by promoting interactions with the environmental microbiota through ecological restoration and holistic public health interventions that mitigate anthropogenic pressures, ecosystem degradation, dysbiotic drift, and the degeneration of the “Lovebug Effect” (Robinson and Breed, Reference Robinson and Breed2020). Evidence for this was provided by exposome-based ecological restoration interventions that studied children attending daycare centres with standard or biodiversity yards (intervention) and compared them with control daycares that made daily visits to a nearby forest. After 28 days, the biodiversity intervention improved immune regulation and health-related commensal skin and gut microbiota (Roslund et al., Reference Roslund, Puhakka, Grönroos, Nurminen, Oikarinen, Gazali, Cinek, Kramná, Siter, Vari, Soininen, Parajuli, Rajaniemi, Kinnunen, Laitinen, Hyöty and Sinkkonen2020). A replication of the study, increasing the sample size and extending the duration to 2 years, confirmed the previous findings and concluded that the implementation of biodiverse, nature-based exposomic solutions in urban children’s environments can improve the commensal microbiota and may be a feasible approach to reduce immune-mediated diseases (Roslund et al., Reference Roslund, Puhakka, Nurminen, Oikarinen, Siter, Grönroos, Cinek, Kramná, Jumpponen, Laitinen, Rajaniemi, Hyöty and Sinkkonen2021). Professor Turroni’s group also conducted a pilot biodiversity intervention study in which urban children interacted daily with horses at a farm during a summer camp. After 15 days, the intervention resulted in increased diversity and a health-promoting GMB, as well as the recovery of anti-inflammatory metabolites and bacteria capable of bioconverting plant polyphenols to anti-inflammatory metabolites. Importantly, there was no evidence of transmission of virulent and/or toxin-producing strains (Scicchitano et al., Reference Scicchitano, Foresto, Laczny, Cinti, Vitagliano, Halder, Morri, Turroni, D’Amico, Palladino, Fiori, Wilmes, Rampelli and Candela2024b).

In conclusion, it is possible to rewild the modernized GM, but this requires broad exposomic interventions in line with the holistic One Health vision, that is, promoting the integrated health of ourselves – humans, plants, animals, and the environment.

Professor Max Nieuwdorp, Amsterdam University Medical Center, The Netherlands

The endocrine system resembles an orchestra with the pituitary gland as conductor, regulating the secretion of hormones by other glands while constantly monitoring their levels. Each gland secretes specific hormones, which are messenger molecules that carry signals through the blood and other bodily fluids to coordinate various functions in the body. This is done by hormones binding to specific receptors on target cells. In addition to the glands, hormones and hormone-like metabolites are also produced by the GI tract and the GM. Finally, another group of chemical messengers is the pheromones, which are hormone-like molecules that are secreted outside the body (e.g., saliva, sweat, and faeces) that trigger a behaviour in individuals of the same species (e.g., attraction, distancing, and aggression). Animals sense pheromones through the vomeronasal organ, a structure that was thought to be non-functional in humans but given that in 25% of human adults it is still present inside the nose. Of note, some studies suggest that the microbiota alter pheromone levels, which may be an area of research worth exploring.

History is full of examples of hormonal disorders that were not recognized as such at the time or were hidden for social or political reasons. Some, like Charles Byrne, the “Irish Giant,” who suffered from acromegaly secondary to a pituitary tumour, were considered curiosities. And world leaders like US President John F. Kennedy, who had Addison’s disease, kept it a secret to present himself as a healthy, energetic persona. Fortunately, our understanding of hormones has come a long way in the last 100 years, and therapies have been developed, but much remains to be discovered. What we do know is that hormones play a vital role in our daily lives, and that external factors can affect hormone function.

As stated earlier, the GM can also produce hormones, and bidirectional interactions exist between them. The following examples are evidence of this. In the context of weight maintenance, there is a close interaction between diet, GM, and hormones insulin and glucagon-like peptide-1 (GLP-1). The secretion of GLP-1 produced by L cells in the (small) intestine as a counterpart of insulin can be increased by butyrate produced by gut bacteria (Yadav et al., Reference Yadav, Lee, Lloyd, Walter and Rane2013). In patients with T2D, the diversity of the GM is altered, and the variation of certain bacterial species and the presence of dysbiosis is associated with altered insulin levels (Karlsson et al., Reference Karlsson, Tremaroli, Nookaew, Bergström, Behre, Fagerberg, Nielsen and Bäckhed2013). Finally, faecal microbiota transplantation (FMT) from lean donors to patients with metabolic syndrome temporarily improved their insulin sensitivity (Vrieze et al., Reference Vrieze, Van Nood, Holleman, Salojärvi, Kootte, Bartelsman, Dallinga-Thie, Ackermans, Serlie, Oozeer, Derrien, Druesne, Van Hylckama Vlieg, Bloks, Groen, Heilig, Zoetendal, Stroes, de Vos, Hoekstra and Nieuwdorp2012).

As we age, the production of sex hormones declines in both sexes, but the dynamics of hormonal change are entirely different in women and men. In women, oestrogen and progesterone decline sharply in middle age (menopause), while in men, testosterone starts to decline slowly after 30 years of age and continues until death (andropause). Not surprisingly, these two different trajectories of sex hormones have different effects on the body.

During menopause, decreasing oestrogen increases the risk of cardiovascular disease (CVD) (Bittner, Reference Bittner2009) and affects oestrogen receptors in the gut, leading to more frequent GI symptoms. Menopause is also associated with reduced GM diversity (Peters et al., Reference Peters, Santoro, Kaplan and Qi2022) and increased incidence of non-alcoholic fatty liver disease (NAFLD).

With regard to the latter, it is observed that a decline in oestrogen leads to dysbiosis, and the altered microbial fermentation produces substances (e.g., ethanol) that contribute to the development of the disease.

NAFLD is associated with auto-brewery syndrome, a condition in which high levels of ethanol are produced by endogenous fermentation in the small intestine and colon, and patients experience symptoms of alcohol intoxication after consuming sugar and carbohydrates despite not having consumed exogenous alcohol (Cope et al., Reference Cope, Risby and Diehl2000). To investigate the pathogenesis of this condition, Professor Nieuwdorp conducted a study in which GM-produced ethanol was measured in portal and peripheral blood in subjects with and without NAFLD under fasting and postprandial (high carbohydrate diet) conditions. Peripheral blood ethanol concentrations increased with the mixed meal (high carbohydrate) in all participants, but the increase was greater in patients with NAFLD. Portal blood ethanol concentrations were low in subjects without steatosis and ~25-fold higher in those with NAFLD. Focusing on the subjects with NAFLD, ethanol concentration in the peripheral blood was 125 times lower than in portal blood, indicating that the liver continuously filters out most of the ethanol produced in the gut. When subjects were given an alcohol dehydrogenase inhibitor before the test meal, ethanol concentrations increased 10- to 15-fold in NAFLD patients, but this was reversed after oral antibiotic treatment that temporarily reduced GM (Meijnikman et al., Reference Meijnikman, Davids, Herrema, Aydin, Tremaroli, Rios-Morales, Levels, Bruin, de Brauw, Verheij, Kemper, Holleboom, Tushuizen, Schwartz, Nielsen, Brandjes, Dirinck, Weyler, Verrijken, de Block, Vonghia, Francque, Beuers, Gerdes, Bäckhed, Groen and Nieuwdorp2022).

It is thought that the ethanol produced by excessive fermentation by a dysbiotic GM via the portal vein reaches the liver, where it is metabolized and stored as hepatic fat in NAFLD patients, which may also be the case in postmenopausal women. And, when liver metabolism is saturated, ethanol circulates throughout the body and causes symptoms of intoxication simply by consuming a sugary drink, providing evidence that auto-brewery syndrome is present in these patients.

Regarding GM and longevity, studies have shown that the microbiota of centenarians is associated with youth-associated bacterial signatures such as stability of Bacteroides and increased species evenness (Pang et al., Reference Pang, Chen, Lu, Meng, Huang, Yu, Huang, Ye, Chen, Liang, Peng, Luo and Wang2023). Since there are “immortal” creatures in nature, such as the jellyfish Turritopsis dohrnii, that can revert from a mature to an immature state in response to physical damage and ageing, some researchers working on anti-ageing technology intended to extend life believe that in the future, old age could be thought of as a hormonal disorder that can be treated. Professor Nieuwdorp believes that there is some truth to this, and he ended the presentation with food for thought: given the sociological and psychological implications, do we want to live past 100 years?

Professor Sarah Lebeer, University of Antwerp, Belgium

The vaginal microbiota (VM) is as important for human health as the GM and is characterized by low diversity with a dominance of lactobacilli (Ravel et al., Reference Ravel, Gajer, Abdo, Schneider, Koenig, McCulle, Karlebach, Gorle, Russell, Tacket, Brotman, Davis, Ault, Peralta and Forney2011; Lebeer et al., Reference Lebeer, Ahannach, Gehrmann, Wittouck, Eilers, Oerlemans, Condori, Dillen, Spacova, Vander Donck, Masquillier, Allonsius, Bron, Van Beeck, De Backer, Donders and Verhoeven2023), whose lactic acid production prevents pathogen overgrowth by keeping the vaginal pH acidic (Alakomi et al., Reference Alakomi, Skyttä, Saarela, Mattila-Sandholm, Latva-Kala and Helander2000; Lebeer et al., Reference Lebeer, Vanderleyden and De Keersmaecker2008; Petrova et al., Reference Petrova, van den Broek, Balzarini, Vanderleyden and Lebeer2013, Reference Petrova, Lievens, Malik, Imholz and Lebeer2015; Allonsius et al., Reference Allonsius, van den Broek, De Boeck, Kiekens, Oerlemans, Kiekens, Foubert, Vandenheuvel, Cos, Delputte and Lebeer2017). In fact, high diversity in the VM is mostly associated with disease (Ravel et al., Reference Ravel, Gajer, Abdo, Schneider, Koenig, McCulle, Karlebach, Gorle, Russell, Tacket, Brotman, Davis, Ault, Peralta and Forney2011; France et al., Reference France, Ma, Gajer, Brown, Humphrys, Holm, Waetjen, Brotman and Ravel2020; Oerlemans et al., Reference Oerlemans, Bellen, Claes, Henkens, Allonsius, Wittouck, van den Broek, Wuyts, Kiekens, Donders and Lebeer2020a, Reference Oerlemans, Wuyts, Bellen, Wittouck, De Boeck, Ruban, Allonsius, van den Broek, Donders and Lebeer2020b). Humans have a unique symbiosis with lactobacilli, which begin to dominate the VM at puberty, increase during pregnancy, and decrease dramatically at menopause (Ravel et al., Reference Ravel, Gajer, Abdo, Schneider, Koenig, McCulle, Karlebach, Gorle, Russell, Tacket, Brotman, Davis, Ault, Peralta and Forney2011; Hickey et al., Reference Hickey, Zhou, Settles, Erb, Malone, Hansmann, Shew, Van Der Pol, Fortenberry and Forney2015; Lebeer et al., Reference Lebeer, Ahannach, Gehrmann, Wittouck, Eilers, Oerlemans, Condori, Dillen, Spacova, Vander Donck, Masquillier, Allonsius, Bron, Van Beeck, De Backer, Donders and Verhoeven2023). Most women experience at least one vaginal infection in their lifetime, and VM dysbiosis has been linked to fertility problems, pregnancy complications, cancer, and mental health. In infants, maternal VM is important for early gut colonization, and its alteration is associated with the risk of preterm birth and neonatal infections. And in men, partner VM has implications for reproductive health, urinary tract infections, sexually transmitted infections, and mental health (Lewis et al., Reference Lewis, Bernstein and Aral2017; Toh et al., Reference Toh, Xing, Gao, Jordan, Batteiger, Batteiger, Van Der Pol, Muzny, Gebregziabher, Williams, Fortenberry, Fortenberry, Dong and Nelson2023).

Current knowledge of the VM is largely descriptive and epidemiological, and while the unique nature of VM requires it to be studied in humans due to its uniqueness, large numbers of samples are needed to understand the role of vaginal lactobacilli in health. This raises the issue of reaching a large group of healthy women and the potential reluctance to provide intimate samples and data. To address these challenges, Professor Lebeer’s group used citizen science (Lebeer et al., Reference Lebeer, Ahannach, Gehrmann, Wittouck, Eilers, Oerlemans, Condori, Dillen, Spacova, Vander Donck, Masquillier, Allonsius, Bron, Van Beeck, De Backer, Donders and Verhoeven2023; Muliyil, Reference Muliyil2023) to ethically collect samples and data from participants and to involve them in the planning of the project. This was combined with a digital platform to communicate about VM and women’s health and to raise awareness about the importance of conducting a scientific project on the VM. They named the project ISALA after Dr. Isala Van Diest, the first female physician in Belgium (https://isala.be).

Within a few days after the call for participation in ISALA was launched, over 6,000 women had registered, with a total of 4,682 women completing a long questionnaire that included intimate questions and 3,345 sending in the required two swabs. The results of the analysis of the VM composition of each participant were first returned personally, as they are the main stakeholders of the project. Further analyses were then performed to address the scientific questions of whether lactobacilli are still present in sufficient numbers in Belgian women, to analyse the impact of lifestyle and life course events, and to isolate vaginal lactobacilli model organisms for basic genetic and functional research.

The analysis showed that lactobacilli were still abundant in the VM of Belgian women, but there was a hidden diversity even in a community dominated by lactobacilli. Professor Lebeer used the datasets of the Vaginal Community State Type Nearest Centroid Classifier (VALENCIA) community state types (CST) (France et al., Reference France, Ma, Gajer, Brown, Humphrys, Holm, Waetjen, Brotman and Ravel2020), the Human Microbiome Project, and the ISALA project to explore patterns of bacteria co-occurring with lactobacilli for metabolic and physical interaction. Although the current concept of VM composition is that the microbes can be grouped into five CSTs (Ravel et al., Reference Ravel, Gajer, Abdo, Schneider, Koenig, McCulle, Karlebach, Gorle, Russell, Tacket, Brotman, Davis, Ault, Peralta and Forney2011), the results indicated that the composition of the healthy VM is rather a continuum with abundant taxa co-occurring (Lebeer et al., Reference Lebeer, Ahannach, Gehrmann, Wittouck, Eilers, Oerlemans, Condori, Dillen, Spacova, Vander Donck, Masquillier, Allonsius, Bron, Van Beeck, De Backer, Donders and Verhoeven2023). These were often Lactobacillus species, mostly L. crispatus, associated with others like Lactobacillus johnsonii and Limosilactobacillus reuteri. The nature and function of these interactions are being investigated and may open new perspectives for the development of vaginal probiotics.

Other findings were that the VM is dynamic and is influenced by age, menstrual cycle, childbirth, hormonal contraceptives, sexual activity, menstrual hygiene products, and diet. While the most important factor was age, menopause, as such, was much less of a driver of the vaginal community. Childbearing and parity were almost as important as age. Beta diversity was higher in the luteal phase of the menstrual cycle, strongly associated with L. crispatus, as was the use of the traditional birth control pill. Regarding menstrual hygiene products, the menstrual cup was positively associated, and the traditional menstrual pad was negatively associated with L. crispatus. In terms of diet, the most important factor affecting VM was the monthly consumption of sugary drinks, which was negatively associated with L. crispatus, while monthly alcohol consumption was positively associated with L. crispatus (Lebeer et al., Reference Lebeer, Ahannach, Gehrmann, Wittouck, Eilers, Oerlemans, Condori, Dillen, Spacova, Vander Donck, Masquillier, Allonsius, Bron, Van Beeck, De Backer, Donders and Verhoeven2023). These results are associations that further require intervention studies to establish causality.

Due to the interest and importance of menstrual hygiene products to the participants, the first intervention study is currently being conducted in healthy women using five different menstrual hygiene products to evaluate their effect on VM. The second intervention study, currently in preparation, is a placebo-controlled RCT in healthy women orally taking Li. reuteri AMBV339 (Isala bacterium) and/or vitamin B2 to investigate their effect on the GM and VM. Li. reuteri AMBV339 is a member of the community module identified in the participants. The strain can produce vitamin B2 and is thought to be involved in the stability of the vaginal ecosystem. Although the strain is administered orally, the researchers expect that some of the lactobacilli will reach the vagina (gut–vagina axis). These two studies are examples of how data obtained by mapping the microbiota can be used to design clinical trials.

By communicating the study through social media, the study has been taken up around the world, and the ISALA team has been contacted by international researchers to run similar projects in their own countries. With 20 sister projects around the world, the ISALA citizen science project is certainly making history.

Associate Professor Ludivine Doridot, Université Paris Cité, Institut Cochin, France

Endometriosis (EM) is a heterogeneous gynaecological disease in which endometrial-like tissue grows outside the uterus, mainly in the pelvic cavity. Like normal uterine lining, EM lesions thicken, break down, and bleed with each menstrual cycle, causing chronic inflammation in the pelvic cavity and a variety of symptoms. The main symptom is pelvic pain, but it can also cause fatigue, dysuria, and digestive symptoms. In addition, ~40% of the patients suffer from infertility and have an elevated risk of miscarriage if they do become pregnant (Santulli et al., Reference Santulli, Marcellin, Menard, Thubert, Khoshnood, Gayet, Goffinet, Ancel and Chapron2016; Zullo et al., Reference Zullo, Spagnolo, Saccone, Acunzo, Xodo, Ceccaroni and Berghella2017). EM affects 10% of women, mainly in their reproductive years, but for unknown reasons the age of onset is very heterogeneous. Together, these factors make EM detrimental to quality of life (QoL) and a huge financial burden for patients and society, so that EM is now recognized as a major public health problem in many countries.

Factors that increase the risk of EM include a family history of EM, short monthly cycles, long heavy periods, and exposure to an adverse perinatal environment. Conversely, factors that decrease the risk include late menarche, higher parity, breastfeeding, and consumption of HBM (Youseflu et al., Reference Youseflu, Savabi-Esfahani, Asghari-Jafarabadi and Maleki2022). Associations have been described with increased risk of ovarian cancer and a decreased risk of cervical cancer (Zullo et al., Reference Zullo, Spagnolo, Saccone, Acunzo, Xodo, Ceccaroni and Berghella2017; Kvaskoff et al., Reference Kvaskoff, Mahamat-Saleh, Farland, Shigesi, Terry, Harris, Roman, Becker, As-Sanie, Zondervan, Horne and Missmer2021), as well as inflammatory and autoimmune diseases, including celiac disease and IBD (Shigesi et al., Reference Shigesi, Kvaskoff, Kirtley, Feng, Fang, Knight, Missmer, Rahmioglu, Zondervan and Becker2019). The relationship between EM and these GI diseases is still unknown, raising the question of the involvement of the microbiome in their pathophysiology.

EM is diagnosed by medical history, pelvic examination, and imaging. A common problem is the long delay in diagnosis (currently around 8 years), explained by both patient and/or clinician factors (Fryer et al., Reference Fryer, Mason-Jones and Woodward2025). Another reason may be that a definitive diagnosis requires laparoscopy, a surgical procedure that clinicians delay until it is absolutely necessary. There is no cure for EM, but there are non-specific treatments to manage it. The first line of treatment is hormonal therapy, combined with analgesics, but if this is ineffective or there are side effects, surgery is an alternative. Surgery, however, is associated with high risks and the recurrence rate is high. In the case of infertility, standard assisted reproduction techniques are used. The cause of EM is unclear. The main hypothesis currently proposed for its etiopathogenesis is retrograde menstruation.

However, while this occurs in most women, only 10% develop EM, suggesting that other factors may be involved. One such factor is a dysfunctional immune system in the peritoneal cavity. Indeed, macrophages are frequently found in the peritoneal fluid and are involved in inflammation, angiogenesis, and cell growth (Symons et al., Reference Symons, Miller, Kay, Marks, Liblik, Koti and Tayade2018). A recent study confirmed decreased macrophage phagocytosis and NK cell cytotoxic activity, as well as increased pro-inflammatory and chemotactic signals in the peritoneal fluid (Zou et al., Reference Zou, Wang, Xu, Xu, Zhu, Yu, Peng, Guo, Li and Zhang2021).

Based on “trained immunity,” Associate Professor Doridot’s group studied the effect of immunomodulation on the development of EM in a mouse model of surgically induced EM (Jeljeli et al., Reference Jeljeli, Riccio, Chouzenoux, Moresi, Toullec, Doridot, Nicco, Bourdon, Marcellin, Santulli, Abrão, Chapron and Batteux2020). Exposure (training) of macrophages or mice to the BCG vaccine induces a pro-inflammatory response to subsequent triggers, whereas exposure to repeated low doses of lipopolysaccharide (ld-LPS, a component of Gram-negative bacteria) mimics post-sepsis immunoparalysis. Macrophages were trained in vivo (intraperitoneal injection of BCG or ld-LPS) or in vitro (LPS in culture) and injected into mice.

In both experiments, lesion development was promoted by BCG, whereas the size of the EM lesions was reduced by LPS. Co-culture of human endometrial cells with BCG- or LPS-trained macrophages showed that LPS-trained macrophages reduced the inflammatory, fibrotic, and adhesion profiles in the culture. Furthermore, a study done in a cohort of EM patients found that women with a history of a Gram-positive infection had a more severe phenotype, with more pelvic pain, whereas those with a history of Gram-negative infection had fewer EM lesions, less pain, and a shorter duration of infertility (Jeljeli et al., Reference Jeljeli, Riccio, Chouzenoux, Moresi, Toullec, Doridot, Nicco, Bourdon, Marcellin, Santulli, Abrão, Chapron and Batteux2020). These associations need to be confirmed in a larger study but have prompted further investigation into the potential association between EM and certain infections. In fact, a Japanese group identified Fusobacterium, a Gram-negative bacterium, in the EM lesions and demonstrated its potential pathogenic role in EM (Muraoka et al., Reference Muraoka, Suzuki, Hamaguchi, Watanabe, Iijima, Murofushi, Shinjo, Osuka, Hariyama, Ito, Ohno, Kiyono, Kyo, Iwase, Kikkawa, Kajiyama and Kondo2023).

Fusobacterium infection causes macrophage infiltration and TGB-β production, which induces the transformation of stromal fibroblasts into myofibroblasts with high mobility and adhesion. In a mouse model of EM, transplantation of Fusobacterium-infected EM tissue increased the number of myofibroblasts and led to lesion growth, while tissue from uninfected mice did not. EM tissue infected with Lactobacillus iners (normal vaginal inhabitant) did not grow lesions. In addition, antibiotic-treated Fusobacterium-infected tissue grew fewer lesions. These data suggest that Fusobacterium may indeed play a role in the pathogenesis of EM, and that antibiotics may be a potential treatment for EM, but this requires further investigation. Regarding the involvement of vaginal microbiota or changes in GM diversity, studies have failed to find associations with EM (Colonetti et al., Reference Colonetti, Saggioratto, Grande, Colonetti, Junior, Ceretta, Roever, Silva and da Rosa2023; Pérez-Prieto et al., Reference Pérez-Prieto, Vargas, Salas-Espejo, Lüll, Canha-Gouveia, Pérez, Fontes, Salumets, Andreson, Aasmets, team, Whiteson, Org and Altmäe2024).

However, a study of the association between EM and dietary patterns found a lower risk of EM with a healthy diet, while a Western diet was associated with a higher risk (Dougan et al., Reference Dougan, Fest, Cushing-Haugen, Farland, Chavarro, Harris and Missmer2024). Although this study did not analyse the microbiota, it hints at an involvement of the microbiome in the development of EM.

Associate Professor Doridot’s current research focuses on the relationship between EM and infection. To this end, her team is developing EM epithelial-stromal cell organoids that will be co-cultured with microbial derivatives in future experiments to understand the mechanisms involved.

Associate Professor Alejandro Arias-Vásquez, Radboud University Medical Center, The Netherlands

The microbiota–gut–brain axis (MGBA) is a bidirectional pathway between the gut and the brain that communicates via the immune system, the vagus nerve, and the neuroendocrine system. Associate Professor Arias-Vásquez is investigating the role of the MGBA in modulating brain function and behaviour, and its involvement in neurodevelopmental disorders (NDDs) and psychiatric disorders (PsyDs), focusing on the neuroendocrine pathway.

There are two critical windows of vulnerability for mental health during development: early childhood for NDDs, and adolescence for PsyDs. Like all the processes that occur in our organism throughout life, neuropsychiatric processes are modulated by the exposome and several behavioural traits associated with genetic information shared by NDDs and PsyDs can be influenced by the environment, including the GM.

Bacteria produce metabolites that are structurally similar to neurotransmitters or their precursors, which enter the bloodstream, cross the BBB, and alter the kinetics of neurotransmission in the brain, thereby compensating for the effects of genetic or other environmental factors (Aarts et al., Reference Aarts, Ederveen, Naaijen, Zwiers, Boekhorst, Timmerman, Smeekens, Netea, Buitelaar, Franke, van Hijum and Arias Vasquez2017). It is therefore plausible that modulation of the GM influences the risk of developing NDDs and PsyDs.

The study of the GM in relation to phenotypes and outcomes of mental health disorders defined by classical clinical diagnosis has become popular. This has led to a wealth of data on various conditions, but research in this area suffers from methodological, technical/logistical, and analytical problems. Of all these problems, the irreproducibility of results is the one that concerns Associate Professor Arias-Vásquez the most. Therefore, his research consists of a wide range of preclinical and clinical research methodologies, using combinations of statistical approaches for data analysis. Moreover, he incorporates a Research Domain Criteria (RDoc)-based classification, which uses more quantitative traits than classical diagnosis, to replicate or find converging association signals that are robust enough to be replicated by independent groups.

For example, in a meta-analysis of the GM in adult attention-deficit/hyperactivity disorder (ADHD), his team was confronted with small sample sizes and studies that were heterogeneous in methods used and outcomes obtained. However, by systematically harmonizing bioinformatic data and applying various meta-analytic methods, they succeeded in finding converging results on bacterial genera associated with ADHD (Jakobi et al., Reference Jakobi, Vlaming, Mulder, Ribases, Richarte, Ramos-Quiroga, Tendolkar, van Eijndhoven, Vrijsen, Buitelaar, Franke, Hoogman, Bloemendaal and Arias-Vasquez2024).

A characteristic of mental disorders that needs to be considered in research is that pure forms of mental disorders are rare and they usually come with high levels of comorbidity. This is addressed using RDoc functional domains, a transdiagnostic approach that better reflects the complexity and comorbidity of mental health, and views clinical diagnosis through its components. For example, in a study looking for associations between the GM composition and various psychiatric disorders, rather than studying people with a specific mental disorder, the researchers looked at responses to questionnaires reflecting diverse and complex behaviours, but which were grouped into four RDoc functional domains. Examining the relationships between these and the GM, using two statistical methods, identified several associations between the gut bacteria and the RDoC functional domains across stress-related and NNDs (Mulder et al., Reference Mulder, Jakobi, Shi, Mulders, Kist, Collard, Vrijsen, van Eijndhoven, Tendolkar, Bloemendaal and Arias Vasquez2024).

The lack of objective biomarkers in psychiatry makes it difficult to measure treatment response in patients. Since there is evidence that the GM is involved in treatment response in mental disorders, differential microbial associations have the potential to be used as treatment response indicators. In a (nested) case–control study using data from the Marburg–Munster Affective Cohort Study (Winter et al., Reference Winter, Leenings, Ernsting, Sarink, Fisch, Emden, Blanke, Goltermann, Opel, Barkhau, Meinert, Dohm, Repple, Mauritz, Gruber, Leehr, Grotegerd, Redlich, Jansen, Nenadic, Nöthen, Forstner, Rietschel, Groß, Bauer, Heindel, Andlauer, Eickhoff, Kircher, Dannlowski and Hahn2022), Associate Professor Arias-Vásquez examined the bacteria that were differentially associated with cases and controls in order to determine whether serotonergic antidepressants could induce GM differences in people with major depressive disorder (MDD). Nine genera were differentially abundant between MDD cases and controls, linked to the treatment (Borgiani et al., Reference Borgiani, Possidente, Fabbri, Oliva, Bloemendaal, Arias Vasquez, Dinan, Vieta, Menchetti, De Ronchi, Serretti and Fanelli2025). An analysis looking for differences in the MDD group between responders and non-responders is ongoing. Similarly, in a case–control study of multiple sclerosis, 33 genera were differentially abundant between groups, but no differences were found between users and non-users of disease-modifying therapies.

In the area of ADHD research, Associate Professor Arias-Vásquez’s team has conducted a study in which faecal microbiota from ADHD patients or controls were transplanted into mice, showing that stools from ADHD patients with altered composition led to changes in brain structure, function, and behaviour in the mice (Tengeler et al., Reference Tengeler, Dam, Wiesmann, Naaijen, van Bodegom, Belzer, Dederen, Verweij, Franke, Kozicz, Arias Vasquez and Kiliaan2020).

Regarding dietary interventions, they conducted a randomized, controlled trial (RCT) with follow-up, comparing the effects of an elimination diet and a healthy diet on reducing symptoms in children with ADHD. Both studies favoured the healthy diet (Huberts-Bosch et al., Reference Huberts-Bosch, Bierens, Ly, van der Velde, de Boer, van Beek, Appelman, Visser, Bos, Reijmers, van der Meer, Kamphuis, Draaisma, Donders, van de Loo-Neus, Hoekstra, Bottelier, Arias-Vasquez, Klip, Buitelaar, van den Berg and Rommelse2024, Reference Huberts-Bosch, Bierens, Rucklidge, Ly, Donders, van de Loo-Neus, Arias-Vasquez, Klip, Buitelaar, van den Berg and Rommelse2025), and GM analysis is currently underway. Finally, a multicentre RCT investigated the effect of a symbiotic product in adults with ADHD and/or borderline personality disorder, and showed that adding the symbiotic to their treatment was associated with an improvement of irritability in a subgroup of patients (Arteaga-Henríquez et al., Reference Arteaga-Henríquez, Ramos-Sayalero, Ibañez-Jimenez, Karina Rosales-Ortiz, Kilencz, Schiweck, Schnorr, Siegl, Arias-Vasquez, Bitter, Fadeuilhe, Ferrer, Lavebratt, Matura, Reif, Réthelyi, Richarte, Rommelse and Antoni Ramos-Quiroga2024). For now, the evidence suggests that modulation of the GM through dietary interventions may have a positive impact on mental health, but more and better research is needed on its proposed role in behavioural and psychiatric outcomes.

Taken together, there is growing evidence that the GM is involved in mental health, and that modulating it is a promising avenue for improving mental health and QoL. At present, however, the lack of replication of the findings makes it difficult to establish associations with confidence, and only harmonization and collaboration will overcome this.

Professor Yvonne Nolan, University College Cork, Ireland

In our ageing global population, the incidence of cognitive decline associated with dementia is increasing. It is predicted that by 2050 the number of people over 60 years will have doubled (United Nations Population Division, 2019), and the prevalence of dementia is set to triple (Nichols et al., Reference Nichols, Steinmetz, Vollset, Fukutaki, Chalek, Abd-Allah, Abdoli, Abualhasan, Abu-Gharbieh, Akram, Al Hamad, Alahdab, Alanezi, Alipour, Almustanyir, Amu, Ansari, Arabloo, Ashraf, Astell-Burt, Ayano, Ayuso-Mateos, Baig, Barnett, Barrow, Baune, Béjot, Bezabhe, Bezabih, Bhagavathula, Bhaskar, Bhattacharyya, Bijani, Biswas, Bolla, Boloor, Brayne, Brenner, Burkart, Burns, Cámera, Cao, Carvalho, Castro-de-Araujo, Catalá-López, Cerin, Chavan, Cherbuin, Chu, Costa, Couto, Dadras, Dai, Dandona, Dandona, De la Cruz-Góngora, Dhamnetiya, Dias da Silva, Diaz, Douiri, Edvardsson, Ekholuenetale, El Sayed, El-Jaafary, Eskandari, Eskandarieh, Esmaeilnejad, Fares, Faro, Farooque, Feigin, Feng, Fereshtehnejad, Fernandes, Ferrara, Filip, Fillit, Fischer, Gaidhane, Galluzzo, Ghashghaee, Ghith, Gialluisi, Gilani, Glavan, Gnedovskaya, Golechha, Gupta, Gupta, Gupta, Haider, Hall, Hamidi, Hanif, Hankey, Haque, Hartono, Hasaballah, Tasdik Hasan, Hassan, Hay, Hayat, Hegazy, Heidari, Heidari-Soureshjani, Herteliu, Househ, Hussain, Hwang, Iacoviello, Iavicoli, Ilesanmi, Ilic, Ilic, Irvani, Iso, Iwagami, Jabbarinejad, Jacob, Jain, Jayapal, Jayawardena, Prakash Jha, Jonas, Joseph, Kalani, Kandel, Kandel, Karch, Kasa, Kassie, Keshavarz, Khan, Khatib, Khoja, Khubchandani, Kim, Kim, Kisa, Kisa, Kivimäki, Koroshetz, Koyanagi, Anil Kumar, Kumar, Lak, Leonardi, Li, Lim, Liu, Liu, Logroscino, Lorkowski, Lucchetti, Saute, Magnani, Malik, Massano, Mehndiratta, Menezes, Meretoja, Mohajer, Ibrahim, Mohammad, Mohammed, Mokdad, Mondello, Moni, Moniruzzaman, Mossie, Nagel, Naveed, Nayak, Kandel, Nguyen, Oancea, Otstavnov, Otstavnov, Owolabi, Panda-Jonas, Pashazadeh Kan, Pasovic, Patel, Pathak, Peres, Perianayagam, Peterson, Phillips, Pinheiro, Piradov, Pond, Potashman, Pottoo, Prada, Radfar, Raggi, Rahim, Rahman, Ram, Ranasinghe, Rawaf, Rawaf, Rezaei, Rezapour, Robinson, Romoli, Roshandel, Sahathevan, Sahebkar, Sahraian, Sathian, Sattin, Sawhney, Saylan, Schiavolin, Seylani, Sha, Shaikh, Shaji, Shannawaz, Shetty, Shigematsu, Shin, Shiri, Santos Silva, Pedro Silva, Silva, Singh, Skryabin, Skryabina, Smith, Soshnikov, Spurlock, Stein, Sun, Tabarés-Seisdedos, Thakur, Timalsina, Tovani-Palone, Tran, Tsegaye, Tahbaz, Valdez, Venketasubramanian, Vlassov, Thu Vu, Gia Vu, Wang, Wimo, Winkler, Yadav, Jabbari, Yamagishi, Yang, Yano, Yonemoto, Yu, Yunusa, Zadey, Zastrozhin, Zastrozhina, Zhang, Murray and Vos2022). Alzheimer’s disease (AD) is the most common form of dementia, and as there is no effective treatment for it, research into how we can delay or prevent it is essential. Throughout the lifespan, exposure to positive or negative environmental and lifestyle factors affects brain plasticity and cognitive function. Professor Nolan’s research focuses on how our environment and lifestyle alter our GM and how this affects brain aging and changes in brain plasticity and memory behaviours associated with AD.

Neurogenesis in the hippocampus region of the brain is a unique form of brain plasticity, which is necessary for certain forms of memory and emotional regulation. Neurogenesis is the production of new neurons and their development into brain circuits (Kozareva et al., Reference Kozareva, Cryan and Nolan2019), a process that declines with age but persists in neurologically healthy people as they age. From the earliest stages of AD, neurogenesis is reduced, neurons are damaged and die, and synapses degenerate, all of which leads to reduced brain plasticity and consequently cognitive decline (Drew, Reference Drew2018; Moreno-Jiménez et al., Reference Moreno-Jiménez, Flor-García, Terreros-Roncal, Rábano, Cafini, Pallas-Bazarra, Ávila and Llorens-Martín2019; Salta et al., Reference Salta, Lazarov, Fitzsimons, Tanzi, Lucassen and Choi2023).

Recently, the MGBA has attracted attention in the field of brain aging and AD (Capocchi et al., Reference Capocchi, Figueroa-Romero, Dunham, Faraci, Rothman, Whiteson, Seo, Holtzman, Grabrucker, Nolan, Kaddurah-Daouk and Jett2024). The GM, which is responsive to environmental and lifestyle factors, may be a peripheral route affecting the hippocampus. One study showed that neurogenesis is altered in the brain of germ-free mice (Ogbonnaya et al., Reference Ogbonnaya, Clarke, Shanahan, Dinan, Cryan and O’Leary2015). Similarly, Professor Nolan’s group has shown that neurogenesis and spatial pattern separation (a hippocampus-dependent memory process that relies on neurogenesis) are impaired in rats following a 10-week course of an antibiotic cocktail, given to deplete their microbiota (Nicolas et al., Reference Nicolas, Dohm-Hansen, Lavelle, Bastiaanssen, English, Cryan and Nolan2024).

A recent meta-analysis of GM composition in AD and mild cognitive impairment found that diversity is altered in AD, with a decrease in Firmicutes and an increase in Bacteroidetes (G. Chen et al., Reference Chen, Zhou, Zhu, Shi and Kong2023). Similar changes were observed in the microbiota of AD patients in a study conducted by Professor Nolan’s group (Grabrucker et al., Reference Grabrucker, Marizzoni, Silajdžić, Lopizzo, Mombelli, Nicolas, Dohm-Hansen, Scassellati, Moretti, Rosa, Hoffmann, Cryan, O’Leary, English, Lavelle, O’Neill, Thuret, Cattaneo and Nolan2023), who also found that faecal Bacteroidota (a rather pro-inflammatory phylum) correlated with pro-inflammatory cytokines in the blood of AD patients, as previously reported (Cattaneo et al., Reference Cattaneo, Cattane, Galluzzi, Provasi, Lopizzo, Festari, Ferrari, Guerra, Paghera, Muscio, Bianchetti, Volta, Turla, Cotelli, Gennuso, Prelle, Zanetti, Lussignoli, Mirabile, Bellandi, Gentile, Belotti, Villani, Harach, Bolmont, Padovani, Boccardi and Frisoni2017). At the genus level, Coprococcus (butyrate producers) was decreased, and Desulfovibrio (hydrogen sulphide and lipopolysaccharide producers) was increased in AD patients. When the correlation of Coprococcus and Desulfovibrio was analysed, a positive correlation was found between the Mini-Mental State Examination (MMSE; cognitive performance test) and Coprococcus and an inverse one between MMSE and Desulfovibrio. To investigate the causality of these GM changes at the behavioural and neuronal level in relation to AD, the authors inoculated faecal microbiota from AD patients (AD-FMT) or age-matched healthy controls to healthy young rats with an antibiotic-depleted microbiota.

Forty percent of the human donor taxa engrafted into the cecum of the recipient rats, and the composition reflected that of the AD donor patients showed no changes in AD-FMT rats. Rats that received FMT from AD patients showed impaired performance in the Morris water maze (spatial memory), novel object recognition (recognition memory), and spatial pattern separation tasks, but no change in the open field (locomotion), elevated plus maze (anxiety), and forced swim (depression) tests. Histological analysis of the hippocampus of the animals showed reduced new neuron survival, cell proliferation, dendrite growth, and development, as well as reduced dendrite length and complexity in AD-FMT rats compared to controls. Finally, to investigate the pathway through which microbiota affects neurogenesis, an untargeted metabolomic analysis of the caecal content and hippocampal tissue was performed. Thirteen metabolites in the caecal content and three in the hippocampus were differentially abundant in AD-FMT rats compared to controls. Hydroxybutyrate (reported to promote neurogenesis and cognitive function) was reduced in the caecum.

The reduction in 4-guanidinobutyric acid, homo-citrulline, and taurine (an amino acid abundant in seafood and known to stimulate neurogenesis (Gebara et al., Reference Gebara, Udry, Sultan and Toni2015), observed in the hippocampus, reflects a previous report in an AD mouse model (Zhao et al., Reference Zhao, Chen, Iqbal, Liu, Zhang, Xiao, Jin, Yao and Shen2021). Taken together, this study (Grabrucker et al., Reference Grabrucker, Marizzoni, Silajdžić, Lopizzo, Mombelli, Nicolas, Dohm-Hansen, Scassellati, Moretti, Rosa, Hoffmann, Cryan, O’Leary, English, Lavelle, O’Neill, Thuret, Cattaneo and Nolan2023) confirmed the causal role of the GM in AD and identified hippocampal neurogenesis as a central process that mediates changes in the gut microbiota in AD (Grabrucker et al., Reference Grabrucker, Marizzoni, Silajdžić, Lopizzo, Mombelli, Nicolas, Dohm-Hansen, Scassellati, Moretti, Rosa, Hoffmann, Cryan, O’Leary, English, Lavelle, O’Neill, Thuret, Cattaneo and Nolan2023).

The “Lancet Commission on Dementia Prevention, Intervention and Care” (Livingston et al., Reference Livingston, Huntley, Liu, Costafreda, Selbæk, Alladi, Ames, Banerjee, Burns, Brayne, Fox, Ferri, Gitlin, Howard, Kales, Kivimäki, Larson, Nakasujja, Rockwood, Samus, Shirai, Singh-Manoux, Schneider, Walsh, Yao, Sommerlad and Mukadam2024) has identified that nearly half of all dementia cases could be delayed or prevented by addressing 14 modifiable risk factors, many of which are associated with or susceptible to changes in GM. The majority of these risk factors are most prevalent during midlife. This could therefore be a time of the lifespan prodromal to dementia but also a window of opportunity to improve gut and brain health through lifestyle changes. Ultimately, while we should promote positive lifestyle choices to achieve a healthy brain and quality of life, we should take into account the complexity of individuals, societies, and situations, which may influence choices about personal lifestyles.

Mr. Kazunori Suda, Yakult Honsha European Research Center for Microbiology, Belgium

Stress and sleep disorders are associated with presenteeism (being at work but underperforming due to poor physical and mental health) among office workers (Furuichi et al., Reference Furuichi, Shimura, Miyama, Seki, Ono, Masuya and Inoue2020), with the former causing greater economic losses. As shown earlier in this session, there is abundant evidence for the role of the GM in gut–brain communication during physiological and pathological states. At the same time that the MGBA modulate our responses to stress (Foster et al., Reference Foster, Rinaman and Cryan2017), stress affects the GM, resulting in a “downward spiral.” Interventions that modulate the GM could balance or reverse this stress-induced spiral.

The effect of Lacticaseibacillus paracasei strain Shirota (LcS, formerly called Lactobacillus casei Shirota) on the GBA under stressful conditions has been extensively investigated using an academic examination stress model in healthy medical students (Takada et al., Reference Takada, Nishida, Kataoka-Kato, Gondo, Ishikawa, Suda, Kawai, Hoshi, Watanabe, Igarashi, Kuwano, Miyazaki and Rokutan2016). This is a model in which psychological stress gradually accumulates in the individual as a scheduled stressful event approaches (in this case, the national medical licensing examination) (Kurokawa et al., Reference Kurokawa, Tanahashi, Murata, Akaike, Katsuura, Nishida, Masuda, Kuwano, Kawai and Rokutan2011). Several double-blind, placebo-controlled studies using this model have shown that LcS prevents the onset of GI discomfort and cold-like symptoms (Kato-Kataoka et al., Reference Kato-Kataoka, Nishida, Takada, Suda, Kawai, Shimizu, Kushiro, Hoshi, Watanabe, Igarashi, Miyazaki, Kuwano and Rokutan2016), attenuates the increase in salivary cortisol (Takada et al., Reference Takada, Nishida, Kataoka-Kato, Gondo, Ishikawa, Suda, Kawai, Hoshi, Watanabe, Igarashi, Kuwano, Miyazaki and Rokutan2016), and preserves sleep quality, particularly sleep latency and deep sleep (Takada et al., Reference Takada, Nishida, Gondo, Kikuchi-Hayakawa, Ishikawa, Suda, Kawai, Hoshi, Kuwano, Miyazaki and Rokutan2017).

Mr. Suda studied whether LcS could improve the daytime performance of 12 stressed office workers with sleep complaints by conducting a double-blind, randomized, crossover, placebo-controlled trial. They were divided into two groups that alternately consumed placebo or LcS-fermented milk (4 weeks each, separated by a 4-week washout period), and measured physiological indicators of performance (EEG and heart rate variability [HRV]) and mood in the morning and afternoon during the last week of each intervention period (Kikuchi-Hayakawa et al., Reference Kikuchi-Hayakawa, Ishikawa, Suda, Gondo, Hirasawa, Nakamura, Takada, Kawai and Matsuda2023). Performance was assessed using the auditory oddball task. This consists of a sequence of repetitive tones (1,000 Hz, non-target) randomly interrupted by atypical tones (2,000 Hz, target). Participants listen for 5 min and react by pushing a button when recognizing the target tones. The perception and recognition of the target generate a characteristic wave (P300) in the brain that is recorded in the EEG. When the P300 latency time, button press reaction time, and button press accuracy rate of P300 were analysed, latency was significantly shorter in the morning, and reaction times tended to be shorter in the afternoon during the LcS period compared to the placebo period.

The accuracy rate was high during both intake periods and did not differ between the two periods. The EEG was also subjected to power spectrum analysis (decomposing the complex EEG wave graph into alpha, beta, delta, and theta waves, based on frequency). Only waves related to brain activity in the awake state during the oddball task and the non-task state were analysed. Alpha and beta waves (relaxed or eyes closed, and concentrated or tense state, respectively) did not differ between placebo and LcS periods. Theta power (sleepiness) was significantly lower in the LcS period versus the placebo period for both the oddball task and the resting state. HRV provides information about the activity of the autonomic nervous system: sympathetic and parasympathetic activities are reflected in the low frequency (LF) and high frequency (HF) components, respectively. The LF/HF ratio was significantly lower during the oddball task in the afternoon when LcS was taken (compared to the placebo period). Daytime mood was assessed using a visual analogue scale (VAS) and the Karolinska sleepiness scale questionnaires (KSS). Results from VAS showed that attention was significantly improved by taking LcS in the afternoon, but fatigue, motivation, optimism, and sleepiness, although better in the LcS period regardless of the time of day, did not show significant differences between the treatments. Sleepiness, according to KSS, was higher in the afternoon than in the morning, but there was no difference between the treatments. Finally, the Spearman’s rank correlation coefficient analysis between perceived daytime mood and physiological indices showed a negative correlation between attention and theta power (sleepiness). Summarizing these results, the consumption of LcS increases daytime wakefulness according to both objective and subjective indices. It improves brain performance in the morning, while reducing sleepiness, modifying the activity of the autonomic nervous system, and improving attention in the afternoon.

Can LcS improve the daytime performance of office workers? Mr. Suda confidently answered “yes” but explained that the relationship between improving stress and sleep and enhancing brain response and attention, as well as the underlying mechanisms, are not yet clear and further research continues.

Professor Paul Cotter, Teagasc Food Research Centre and APC Microbiome Ireland, Ireland

Fermented foods (FFs) are “foods made through desired microbial growth and enzymatic conversions of food components” (Marco et al., Reference Marco, Sanders, Gänzle, Arrieta, Cotter, De Vuyst, Hill, Holzapfel, Lebeer, Merenstein, Reid, Wolfe and Hutkins2021) and have a long history of safe use when made properly. In this definition, the word “desired” refers to the effect of conferring valuable properties to the food (e.g., extending shelf life) and is the key to distinguishing FFs from spoilage.

In recent years, omics-based approaches have improved our understanding of the microbial communities in FFs and the gut, and these insights are helping us to understand how FFs can benefit health. The non-microbial components of FFs (minerals, enzymes, amino acids and derivatives, and bioactive peptides) can contribute to health in a variety of ways, including effects on the gut epithelium, on nutrient bioavailability and absorption, and as neurotransmitters and antimicrobials. The microorganisms and their metabolites (organic acids, SCFAs, polyphenols, and exopolysaccharides) support the GM balance, intestinal epithelial integrity, and health, and regulate the immune system. In addition to their effects on the host, they also remove undesirable components (e.g., mycotoxins, lactose, and FODMAPs) (Mukherjee et al., Reference Mukherjee, Breselge, Dimidi, Marco and Cotter2024). Modern lifestyles, along with industrialized and sanitized foods, have reduced exposure to the right microbes needed to train our immune system and thereby contributed to the development of various diseases (Sonnenburg and Sonnenburg, Reference Sonnenburg and Sonnenburg2019). In light of this, FFs could be used to increase the number of live microbes in our diets and result in significant health benefits.

Live dietary microbes (LDMs) are a very heterogeneous group of organisms, and organizations such as the Institute for the Advancement of Food and Nutrition Sciences (IAFNS) and the International Scientific Association for Probiotics and Prebiotics (ISAPP) are interested in determining if there are health benefits to consuming LDM and whether it is worthy of further research. Interestingly, these organizations have taken different approaches to answer this question. IAFNS conducted a scoping review to determine the breadth of evidence available to support the existence of health benefits from the consumption of LDM and reached two conclusions: “a dose of ≥2 × 10⁹ CFU/day was associated with non-negative outcomes” and “older populations with a median age of 39 years were associated with positive outcomes” (Iyer et al., Reference Iyer, Mukherjee, Gómez-Sala, O’Connor, Kenny and Cotter2024).

ISAPP, meanwhile, used the data from the 2001–2018 National Health and Nutrition Examination Survey (NHANES) on the estimated number of live microorganisms present in foods to extrapolate daily adult microbial intakes and analyse associations with specific health parameters. The results suggested that “foods with high microbial concentrations (>10⁷ live CFU/g) are associated with modest health improvements across a range of outcomes” (Hill et al., Reference Hill, Tancredi, Cifelli, Slavin, Gahche, Marco, Hutkins, Fulgoni, Merenstein and Sanders2023). The comparable results of both studies demonstrated that there is sufficient evidence to warrant further research in this area.

One of the most important ways in which FFs may impact health is through possible interaction with the small intestinal microbiota (SIM). Until recently, the study of the SIM was limited by the technical challenges of sampling from this anatomically difficult-to-reach site, while samples from biopsies or ileostomy bags did not provide accurate information (Kastl et al., Reference Kastl, Terry, Wu and Albenberg2020). However, with the development of innovative technologies, such as orally ingestible capsules that collect fluids from the lumen of the small intestine as they pass through, more and more accurate information has become available. The SIM is a dynamic, resilient, and easily modulated ecosystem, dominated by lactic acid bacteria and microbes that are closely related to those found in FFs, especially dairy FFs. One study showed that after acute consumption of fermented milk, its microbes (10^10.9 CFU) temporarily constituted 90% of the SIM, and 10% of them were alive (Takada et al., Reference Takada, Chinda, Mikami, Shimizu, Oana, Hayamizu, Miyazawa, Arai, Katto, Nagara, Makino, Kushiro, Oishi and Fukuda2020). All these characteristics make the small intestine a perfect environment for a FF microbe to exert its influence, and emerging data suggest that the SIM is indeed important for health.

FFs are multi/panbiotics, that is, they may contain more than just one “biotic,” and although a health benefit is not required for FFs, these “biotic” components may provide health benefits (Vinderola et al., Reference Vinderola, Cotter, Freitas, Gueimonde, Holscher, Ruas-Madiedo, Salminen, Swanson, Sanders and Cifelli2023). As such, dairy fermentates are currently the subject of intense research. Fermentates are powdered multibiotic materials produced by the fermentation of a dairy substrate. They are of interest to the food industry as commercial ingredients due to their stability under different conditions, the ease of storage and transportation, in addition to their bioactivity and safety. Professor Cotter’s group is conducting a human trial of a fermentate selected from 72 microorganisms potentially associated with health benefits, safe for use in food, and thoroughly evaluated in in vitro and ex vivo tests (Finnegan et al., Reference Finnegan, Connolly, Mechoud, FitzGerald, Beresford, Mathur, Brennan, Cotter and Loscher2024a, Reference Finnegan, Mechoud, FitzGerald, Beresford, Mathur, Cotter and Loscher2024b). An important consideration for industrial application of fermentates is the need to grow the microorganisms on a large scale and produce safe and quality-controlled batches. This can be challenging as microorganisms do not always grow and do not necessarily produce the same fermentates when growth conditions change. Further research and development of appropriate protocols to facilitate the scaling process might be critical for their future development.

As for health claims, while there is a substantial amount of knowledge and research on the potential health benefits of FFs, the reality is that regulatory approval of health claims is a challenge that will need to be addressed in the coming years. In this regard, yogurt, a forerunner in health claims related to FFs, received approval from the European Food Safety Authority (EFSA) in 2006 for a health claim related to improved lactose digestion (EFSA, 2010) and was granted a Food and Drug Administration (FDA)-qualified health claim related to reduced risk of T2D in early 2024 (FDA, 2024), paving the way for the formal recognition of health benefits for other FFs in the future. Professor Cotter is addressing this issue by participating in a multinational project called “Promoting Innovation of ferMENted fOods” (PIMENTO, https://fermentedfoods.eu/ds), which facilitates the building of networks of communities working on FFs to promote scientific collaboration, foster innovation, and tackle specific research challenges.

One of the PIMENTO working groups reviewed the available evidence on 16 functional areas and divided them into E-projects, covering clinical and biological endpoints associated with major health indicators as defined by EFSA guidelines, and S-projects, addressing some broader biological questions (Todorovic et al., Reference Todorovic, Akpinar, Assunção, Bär, Bavaro, Berkel Kasikci, Domínguez-Soberanes, Capozzi, Cotter, Doo, Gündüz Ergün, Guzel, Harsa, Hastaoglu, Humblot, Hyseni, Hosoglu, Issa, Karakaş-Budak, Karakaya, Kesenkas, Keyvan, Künili, Kütt, Laranjo, Louis, Mantzouridou, Matalas, Mayo, Mojsova, Mukherjee, Nikolaou, Ortakci, Paveljšek, Perrone, Pertziger, Santa, Sar, Savary-Auzeloux, Schwab, Starowicz, Stojanović, Syrpas, Tamang, Yerlikaya, Yilmaz, Malagon-Rojas, Salminen, Frias, Chassard and Vergères2024). Professor Cotter is participating in an E-project, reviewing the evidence for the beneficial effects of FFs on GI health. The aim is to develop new health claim dossiers, or at least to identify the gaps that need to be filled through studies that meet the requirements common to all RCTs. At the same time, the proposed studies should consider the multi-biotic nature of FFs: pay particular attention to the precise definition of the intervention (e.g., studies with one component, with one FF, and with a group of FF), consider an in-depth characterization of the intervention as to provide metadata for subsequent investigations, use an adequate control component or food, recognize and accommodate for the many potential confounding factors, and finally, caution in extrapolation and generalization of results to avoid over-interpretation.

Future perspectives for the study and application of FFs depend on the provision of accessible metagenomic data on the microbiomes present in FFs. Researchers at Microbiome Applications for Sustainable Food Systems through Technologies and EnteRprise (MASTER, https://www.master-h2020.eu) created a compendium of 2,500 food-associated metagenomes by integrating newly sequenced and publicly available food metagenomes. From this dataset, 10,000 prokaryotic food metagenome-assembled genomes (MAGs, a [pseudo]genome reconstructed from food samples without culturing the microbe) were reconstructed (Carlino et al., Reference Carlino, Blanco-Míguez, Punčochář, Mengoni, Pinto, Tatti, Manghi, Armanini, Avagliano, Barcenilla, Breselge, Cabrera-Rubio, Calvete-Torre, Coakley, Cobo-Díaz, De Filippis, Dey, Leech, Klaassens, Knobloch, O’Neil, Quijada, Sabater, Skirnisdottir, Valentino, Walsh, MASTER EU, Alvarez-Ordonez, Asnicar, Fackelmann, Heidrich, Margolles, Marteinsson, Rota Stabelli, Wagner, Ercolini, Cotter, Segata and Pasolli2024). The MAGs were then grouped into three species-level genome bins (SGBs, clusters of DNA sequences thought to be from the same species): 613 known, 211 unknown (previously undescribed taxa), and, most remarkably, 320 unknown but related food-specific genomes (never before found in food), representing potentially new species that are targets for future studies.

Comparing food metagenomes (more than half of the food categories were from fermented foods) and human metagenomes (from mouth and stool), many overlapping microbial species were consistently detected in both (Carlino et al., Reference Carlino, Blanco-Míguez, Punčochář, Mengoni, Pinto, Tatti, Manghi, Armanini, Avagliano, Barcenilla, Breselge, Cabrera-Rubio, Calvete-Torre, Coakley, Cobo-Díaz, De Filippis, Dey, Leech, Klaassens, Knobloch, O’Neil, Quijada, Sabater, Skirnisdottir, Valentino, Walsh, MASTER EU, Alvarez-Ordonez, Asnicar, Fackelmann, Heidrich, Margolles, Marteinsson, Rota Stabelli, Wagner, Ercolini, Cotter, Segata and Pasolli2024). This does not mean that FFs are responsible for the microbes that colonize the gut, but strain-level metagenomic analysis showed that some branches of particular taxa of the phylogenetic trees contained microbial strains from food and humans, implying that there is potential transfer of strains from the food to the gut. If a comparable amount of metagenomic data were available from the small intestine, the overlap would be expected to be greater than currently observed in stool samples. The food metagenomic data from MASTER can also be used to study specific foods in depth, such as kefir (Barcenilla et al., Reference Barcenilla, Cobo-Díaz, De Filippis, Valentino, Cabrera Rubio, O’Neil, Mahler de Sanchez, Armanini, Carlino, Blanco-Míguez, Pinto, Calvete-Torre, Sabater, Delgado, Ruas-Madiedo, Quijada, Dzieciol, Skírnisdóttir, Knobloch, Puente, López, Prieto, Marteinsson, Wagner, Margolles, Segata, Cotter, Ercolini and Alvarez-Ordóñez2024; Walsh et al., Reference Walsh, Macori, Kilcawley and Cotter2020, Reference Walsh, Coakley, Walsh, Crispie, O’Toole and Cotter2023, Reference Walsh, Breselge, Martin, Coakley, Ferguson, Stapleton, Crispie, O’Toole and Cotter2024). Professor Cotter’s team generated a pangenome from 69 water kefirs, and multi-omics analysis revealed a total of 96 distinct bacterial and yeast species, of which 76 had enough data to be assembled into MAGs. Their grouping into SGBs suggested the presence of 2 fungal and 18 bacterial novel species, including 2 putative novel bifidobacteria (Breselge et al., Reference Breselge, Skibinska, Yin, Brennan, Kilcawley and Cotter2025). The isolation, characterization (including novelty), and selection of novel strains to be introduced into food are aspects that are being studied.

One of Professor Cotter’s current interests is to combine the benefits of artisanal and industrial production of FFs with a focus on novel species. Both types of production have advantages and disadvantages, and the benefits sought from artisanal products are complex microbiota, health-promoting potential, and flavour richness. Those sought from industrial production are large-scale culture, assured stability, quality, and safety. Both benefits can be combined by making new and improved “pitched” FFs. Pitched FFs are produced with a small number of specific strains to recapitulate the health benefits of a traditional FF, which requires growing and harnessing novel species and recreating the microbial communities. As an example, the two novel bifidobacterial strains found in water kefir were taxonomically classified and cultured (Breselge et al., Reference Breselge, Bellassi, Barcenilla, Álvarez-Ordóñez, Morelli and Cotter2024) in the production of a pitched water kefir. This approach has previously been applied to milk kefir, successfully producing a pitched version that, compared to commercial milk kefir, improved cholesterol metabolism and inflammatory markers in animals, confirmed in a pilot study in humans (Bourrie et al., Reference Bourrie, Ju, Fouhse, Forgie, Sergi, Cotter and Willing2021, Reference Bourrie, Forgie, Makarowski, Cotter, Richard and Willing2023). To corroborate these findings, a full-scale trial of pitched kefir consumption over a 6-month period is being conducted as part of the Domino Project, an initiative that aims to attribute health benefits to traditional FFs and to develop novel FFs (https://www.domino-euproject.eu).

The last concept related to research into new horizons concerns the neuroactive potential of FFs and their components. Studies over the last few years have shown that they can modulate the MGBA in multiple ways (van de Wouw et al., Reference van de Wouw, Walsh, Crispie, van Leuven, Lyte, Boehme, Clarke, Dinan, Cotter and Cryan2020, Reference van de Wouw, Walsh, Vigano, Lyte, Boehme, Gual-Grau, Crispie, Walsh, Clarke, Dinan, Cotter and Cryan2021; Balasubramanian et al., Reference Balasubramanian, Schneider, Gunnigle, Cotter and Cryan2024).

Based on this, analysis of the vast metagenomic data on FFs is underway to identify those enriched in pathways leading to neurotransmitter production. The effects of some of the promising foods are currently being investigated using ex vivo studies and are progressing towards in vivo studies.

In conclusion, the knowledge on FFs may be on the verge of a breakthrough, facilitating the development of new and improved foods with enhanced health benefits. Professor Cotter’s work has been a catalyst for this progress, and more is expected to come.

Dr. Yukitoshi Aoyagi, Tokyo Metropolitan Institute of Gerontology, Japan

Chronic diseases and their complications are highly prevalent among older people worldwide. This represents an increasing economic burden at both individual and national levels. Concerned about improving the health and quality-adjusted life expectancy of Japan’s ageing population, Dr. Aoyagi began an observational epidemiological study in the year 2000 on habitual physical activity and diet and their relationship with mental and physical health in 65+-year-old residents in Nakanojo Town (Gunma, Japan) (Aoyagi, Reference Aoyagi2007, Reference Aoyagi2009). The study included medical examinations, blood tests, self-administered questionnaires, and, for some subgroups, genetic and microbiota analyses. But the core of the Nakanojo Study is the measurement of average habitual physical activity. This is measured using a pedometer/accelerometer, worn continuously by participants for 24 h/day, which calculates the number of steps taken daily and the amount of time spent in moderate-intensity activity (defined as three to six times the resting metabolic rate, e.g., walking at a moderate pace) (Aoyagi, Reference Aoyagi2009; Aoyagi and Shephard, Reference Aoyagi and Shephard2010). He has conducted many studies within this framework and presented those focusing on hypertension and infrequent bowel movements/constipation.

Dr. Aoyagi started by determining the average daily habitual physical activity required to prevent diseases common in the elderly population. Based on the results of 500 participants (Aoyagi, Reference Aoyagi2009; Aoyagi et al., Reference Aoyagi, Park, Park and Shephard2010), he identified the thresholds of habitual physical activity to, for example, be 8,000 steps/day and/or 20 min/day of moderate-intensity activity for the prevention of hypertension and diabetes, and 7,000 steps/day and/or 15 min/day for atherosclerosis, cancers of the GI tract and accessory organs, lung, and breast (Aoyagi, Reference Aoyagi2009), concluding that engaging in any amount of moderate-intensity activity is associated with better health in older adults.

One of the topics he focused on was the relationship between the intake frequency of LcS-containing products (LcS-p) over 5 years and the development of hypertension (Aoyagi et al., Reference Aoyagi, Park, Matsubara, Honda, Amamoto, Kushiro, Miyazaki and Shephard2017). The study included 352 normotensive participants divided into two groups based on the frequency of LcS-p intake per week: <3 and ≥3 days/week.

The incidence of hypertension over 5 years was significantly lower in the ≥3 days/week group than in the <3 days/week group (6% vs. 14%), and logistic regression analysis after controlling for confounders (age, sex, BMI, smoking, and alcohol consumption) showed that the relative risk of hypertension among participants who consumed LcS-p ≥ 3 days/week was 40%, compared with those who consumed the LcS-p < 3 days/week. This suggested that long-term consumption of LcS-p at least 3 days/week may contribute to a reduced risk of hypertension (Aoyagi et al., Reference Aoyagi, Park, Matsubara, Honda, Amamoto, Kushiro, Miyazaki and Shephard2017). Based on a previous report (Furushiro et al., Reference Furushiro, Sawada, Kouichi, Motoike, Sansawa, Kobayashi, Watanuki and Yokokura1990), Dr. Aoyagi attributed the antihypertensive effect of LcS to increased biosynthesis of prostaglandin I₂, which reduces peripheral vascular resistance.

Another aspect that Dr. Aoyagi investigated is the relationship between the independent and interactive effects of the frequency of LcS-p intake and habitual physical activity and constipation (≤3 days evacuations/week). The study included 338 participants who were divided into 3 groups according to the frequency of LcS-p intake in the previous month: 0–2, 3–5, and 6–7 days/week (n = 204, 54, and 80, respectively), and 2 groups based on the amount of physical activity in the past month: <7,000 and ≥7,000 steps/day (n = 187 and 151, respectively). Defecation frequency over the past week was estimated by a self-administered questionnaire, and faecal microbiota was analysed qualitatively and quantitatively. After adjustment for confounders, the odds of developing constipation were lower (38%) in the participants who consumed LcS-p 6–7 days/week, compared with those who consumed 0–2 days/week; and lower (44%) in participants who walked ≥7,000 steps/day compared with those who walked <7,000 steps/day. The analysis of the interactive effect showed that the combination of 6–7 days/week and physical activity >7,000 steps was most effective in reducing the odds of developing constipation compared with any combination. The total number of lactobacilli and LcS tended to be higher in the group that consumed LcS-p more frequently, while the composition of the faecal microbiota did not differ between the two physical activity groups. This suggests that two mechanisms are associated with the stimulation of intestinal motility and the subsequent reduction in the risk of constipation. One is the effect of the increased number of lactobacilli and their metabolites (e.g., lactic acid) brought about by the consumption of LcS-p, and the other is the mechanical stimulation by physical activity (Aoyagi et al., Reference Aoyagi, Amamoto, Park, Honda, Shimamoto, Kushiro, Tsuji, Matsumoto, Shimizu, Miyazaki, Matsubara and Shephard2019).

In conclusion, the risk of developing hypertension is inversely associated with the frequent consumption of LcS-p, and the risk of constipation is inversely associated with the frequency of consumption of LcS-p and the amount of physical activity in older adults. These studies did not include a group that functions as a control. Dr. Aoyagi hinted that an intervention study will follow in the near future as part of his research on the role of lactic acid bacteria in promoting a healthy and long life.

Mr. Masatoshi Morikawa, Yakult Honsha European Research Center for Microbiology, Belgium

Numerous studies have shown that LcS has beneficial health effects. In terms of immune modulation, human studies have demonstrated that LcS reduces the risk of GI tract, bladder, and breast cancers by enhancing humoral and cellular responses, particularly NK cell activity, while it also shortens the duration of fever and relieves symptoms of upper respiratory tract infections (Ishikawa et al., Reference Ishikawa, Akedo, Otani, Suzuki, Nakamura, Takeyama, Ishiguro, Miyaoka, Sobue and Kakizoe2005; Gleeson et al., Reference Gleeson, Bishop, Stensel, Lindley, Mastana and Nimmo2011; Nagata et al., Reference Nagata, Asahara, Wang, Suyama, Chonan, Takano, Daibou, Takahashi, Nomoto and Yamashiro2016; Shida et al., Reference Shida, Sato, Iizuka, Hoshi, Watanabe, Igarashi, Miyazaki, Nanno and Ishikawa2017). But where and how does LcS interact with the immune system? Mr/ Morikawa investigated this through a series of experiments focusing on the phagocytes of the small intestine and the cell wall polysaccharides of LcS.

It is well known that mononuclear phagocytes (monocytes, macrophages, and dendritic cells [DCs]) function as initiators and commanders of immune responses, while lymphocytes (B-cells, T-cells, and NK cells) function as effectors. Phagocytes come into contact with bacteria in the small intestine, where the mucus that covers the epithelium is single-layered and thin (Luis and Hansson, Reference Luis and Hansson2023). Additionally, the small intestine harbours a subset of DCs that extend their processes through the epithelial tight junctions to capture bacteria in the lumen (Rescigno et al., Reference Rescigno, Urbano, Valzasina, Francolini, Rotta, Bonasio, Granucci, Kraehenbuhl and Ricciardi-Castagnoli2001; Luis and Hansson, Reference Luis and Hansson2023). A human study of the microbial composition of the ileal content after a single intake of LcS-fermented milk (LcS-FM) showed that LcS reaches the ileum 2–4 hours after ingestion and remains there for several hours, with its relative abundance exceeding 90% during this time (Takada et al., Reference Takada, Chinda, Mikami, Shimizu, Oana, Hayamizu, Miyazawa, Arai, Katto, Nagara, Makino, Kushiro, Oishi and Fukuda2020). This suggested that LcS was likely to encounter phagocytes in the small intestine, but it was not known whether they interact with bacteria in the steady state. Using phagocytes isolated from the lamina propria of the mouse small intestine, Mr. Morikawa showed by histological staining and fluorescence in situ hybridization that bacteria are engulfed by phagocytes, even in the steady state (Morikawa et al., Reference Morikawa, Tsujibe, Kiyoshima-Shibata, Watanabe, Kato-Nagaoka, Shida and Matsumoto2016). To understand how this interaction activates systemic immunity, Mr. Morikawa assumed that LcS is also phagocytosed in the small intestine, and a double-blind RCT was set up in which healthy office workers consumed either an LcS-FM or a control milk drink daily for 6 weeks. The expression of CD40 (a trigger of cytokine production, such as interleukin 12 [IL-12]) on the surface of peripheral blood phagocytes was measured on days 3, 10, 28, and 42. CD40 was significantly increased in the phagocytes of the LcS-FM group on day 3 (Naito et al., Reference Naito, Morikawa, Yamamoto-Fujimura, Iwata, Maki, Kato-Nagaoka, Oana, Kiyoshima-Shibata, Matsuura, Kaji, Watanabe, Shida, Matsumoto and Hori2023). These studies suggest that LcS may influence systemic immunity through signals originating in small intestinal phagocytes and directed to peripheral immune effectors.

Next, to elucidate what happens after LcS is phagocytosed, the resistance of LcS to digestion was evaluated through stimulation of macrophages or exposure to N-acetyl-muramidase. Heat-killed LcS (HK-LcS) engulfed by macrophages remained undigested even after 24 hours and resisted enzymatic digestion after 2 hours. In contrast, control lactobacilli (L. plantarum and L. johnsonii) were digested within 1 hour (Shida et al., Reference Shida, Kiyoshima-Shibata, Nagaoka, Watanabe and Nanno2006, Reference Shida, Kiyoshima-Shibata, Kaji, Nagaoka and Nanno2009). The components that confer LcS this resistance were identified by enzymatic fractionation of HK-LcS into intact cell wall fraction (ICW), protoplast fraction (PP), proteoglycan fraction (PG), and polysaccharides (PS). When ICW and PG were incubated with macrophages or exposed to N-acetyl-muramidase, ICW remained undigested, whereas PG was mostly digested within 1 hour. As they differ only in the presence of polysaccharides, it was concluded that the polysaccharides in ICW could make LcS resistant to digestion (Shida et al., Reference Shida, Kiyoshima-Shibata, Kaji, Nagaoka and Nanno2009). Electron microscopy of a sample treated with a monoclonal antibody against its polysaccharides showed that LcS is completely covered by polysaccharides, identified as LCPS-1 (100 kDa, linear) and LCPS2 (32 kDa, branched) (Nagaoka et al., Reference Nagaoka, Muto, Nomoto, Matuzaki, Watanabe and Yokokura1990; Mizukoshi et al., Reference Mizukoshi, Kimura, Ikemura, Mori and Nagaoka2022).

The function of the LcS cell wall is not only to provide resistance to digestion, but it also plays a role in enhancing NK cell activity via IL-12, as was demonstrated in a series of in vitro studies. First, stimulation of mouse peritoneal macrophages with HK-LcS induced IL-12 production in a dose-dependent manner (Kaji et al., Reference Kaji, Kiyoshima-Shibata, Nagaoka, Nanno and Shida2010). Second, when the capacity of HK-LcS to activate NK cells was tested in human PBMCs fractionated into monocytes and other cells (which included NK cells), HK-LcS increased NK cell activity only in the presence of monocytes. And finally, the addition of anti-IL-12 antibody to an incubation of LcS and PBMCs reduced LcS-stimulated NK cell activity by ~50% (Takeda et al., Reference Takeda, Suzuki, Shimada, Shida, Nanno and Okumura2006). Next, the LcS structure responsible for IL-12 induction was identified by exposing macrophages to HK-LcS or fractionated cell wall components. ICW induced IL-12 production at levels comparable to HK-LcS, whereas PG induced little production, suggesting that the polysaccharide-bound structure of the cell wall is essential for IL-12 induction (Shida et al., Reference Shida, Kiyoshima-Shibata, Nagaoka, Watanabe and Nanno2006). This finding was further assessed using LcS polysaccharide-peptidoglycan complexes (PS-PG)-bound microbeads or soluble PS-PG. When added to macrophages, only the PS-PG-bound microbeads increased IL-12 in a concentration-dependent manner (Nagahama et al., Reference Nagahama, Kumano, Nakata, Tsuji, Moriyama, Shida, Nomoto, Chiba, Koumoto and Matsui2015), confirming that the undigested structure bound to PS-PG is important for IL-12 induction.

Mr. Morikawa concluded his presentation by emphasizing the key role played by phagocytes and polysaccharides (the two “P”‘s) in the immune system-strengthening effect of LcS.

Dr. Bastiaan Haak, Wellcome Sanger Institute, UK & Amsterdam UMC, The Netherlands

Sepsis is “a life-threatening organ dysfunction caused by a dysregulated host response to infection” (Singer et al., Reference Singer, Deutschman, Seymour, Shankar-Hari, Annane, Bauer, Bellomo, Bernard, Chiche, Coopersmith, Hotchkiss, Levy, Marshall, Martin, Opal, Rubenfeld, van der Poll, Vincent and Angus2016) and is a major public health problem with high costs. Yet we still lack adequate diagnosis and treatment.

The GM of patients with sepsis often exhibits a loss of commensal obligate anaerobes and an overgrowth of pathogenic bacteria. Evidence suggests that these perturbations are associated with increased mortality. The antibiotics commonly used to treat sepsis are vancomycin and piperacillin-tazobactam, which are effective against harmful bacteria but also affect commensal anaerobes. In 2015, the use of cefepime, an anaerobe-sparing antibiotic, replaced piperacillin-tazobactam due to a shortage of the latter, and a retrospective cohort study of sepsis hospitalizations between 2014 and 2018 found that the 90-day mortality was higher in those who received piperacillin-tazobactam compared to those who received cefepime (Chanderraj et al., Reference Chanderraj, Admon, He, Nuppnau, Albin, Prescott, Dickson and Sjoding2024). This study highlighted the importance of sparing obligate anaerobes in the treatment of sepsis.

The concept of the gut–lung axis emerged a decade ago from observations in a study in which either mice depleted of their microbiota by broad-spectrum antibiotics, or non-treated controls, were infected intranasally with Streptococcus pneumoniae. Microbiota-depleted mice had increased bacterial dissemination, inflammation, organ damage, and mortality after pneumonia compared to controls. FMT with faeces collected before antibiotic administration normalized their response to infection (Schuijt et al., Reference Schuijt, Lankelma, Scicluna, de Sousa e Melo, Roelofs, de Boer, Hoogendijk, de Beer, de Vos, Belzer, de Vos, van der Poll and Wiersinga2016). Since then, many animal studies have solidified the concept of the gut–lung axis, but the role of anaerobes is still not fully understood. Central to this axis are the metabolites produced by gut anaerobes, particularly butyrate. In addition to its functions in the gut, butyrate acts as an immunoregulator in the lungs by reaching the organ via the blood or lymphatic circulation and by increasing the repertoire of regulatory T cells involved in limiting inflammation (Haak and Wiersinga, Reference Haak and Wiersinga2017; Haak et al., Reference Haak, Prescott and Wiersinga2018b; Schulthess et al., Reference Schulthess, Pandey, Capitani, Rue-Albrecht, Arnold, Franchini, Chomka, Ilott, Johnston, Pires, McCullagh, Sansom, Arancibia-Cárcamo, Uhlig and Powrie2019). Based on these observations, Professor Haak studied the relationship between the abundance of butyrate-producing gut bacteria (BPGB) and the development of infections in three groups of high-risk patients: stem cell transplant patients, acute stroke patients, and kidney transplant patients. The studies measured clinical and microbiota risk factors for lower respiratory tract (LRT) or other infections, and all confirmed an inverse association between BPGB abundance and the risk of developing LRT infections, independent of other predictive factors (Haak et al., Reference Haak, Littmann, Chaubard, Pickard, Fontana, Adhi, Gyaltshen, Ling, Morjaria, Peled, van den Brink, Geyer, Cross, Pamer and Taur2018a; Lee et al., Reference Lee, Huang, Magruder, Zhang, Gong, Sholi, Albakry, Edusei, Muthukumar, Lubetzky, Dadhania, Taur, Pamer and Suthanthiran2019; Haak et al., Reference Haak, Westendorp, van Engelen, Brands, Brouwer, Vermeij, Hugenholtz, Verhoeven, Derks, Giera, Nederkoorn, de Vos, van de Beek and Wiersinga2021). These are associative data, but the replication of the results in three patient cohorts provides confidence that this signal warrants further investigation.

Obligate anaerobic bacteria can also be used as diagnostic or predictive indicators of susceptibility to infectious diseases in both at-risk and general populations. Two cohorts of voluntarily participating adults, namely the “Healthy Life in an Urban Setting” (HELIUS) cohort (Snijder et al., Reference Snijder, Galenkamp, Prins, Derks, Peters, Zwinderman and Stronks2017) and the FINRISK cohort (Borodulin et al., Reference Borodulin, Tolonen, Jousilahti, Jula, Juolevi, Koskinen, Kuulasmaa, Laatikainen, Männistö, Peltonen, Perola, Puska, Salomaa, Sundvall, Virtanen and Vartiainen2018), provided an opportunity to analyse their GM (Deschasaux et al., Reference Deschasaux, Bouter, Prodan, Levin, Groen, Herrema, Tremaroli, Bakker, Attaye, Pinto-Sietsma, van Raalte, Snijder, Nicolaou, Peters, Zwinderman, Bäckhed and Nieuwdorp2018; Salosensaari et al., Reference Salosensaari, Laitinen, Havulinna, Meric, Cheng, Perola, Valsta, Alfthan, Inouye, Watrous, Long, Salido, Sanders, Brennan, Humphrey, Sanders, Jain, Jousilahti, Salomaa, Knight, Lahti and Niiranen2021). A 6-year follow-up of these two cohorts showed that a lower number of BPGB was associated with a higher incidence of severe infections, suggesting that the baseline GM data can predict hospitalization or mortality due to severe infections independently of other predictive factors (Kullberg et al., Reference Kullberg, Wikki, Haak, Kauko, Galenkamp, Peters-Sengers, Butler, Havulinna, Palmu, McDonald, Benchraka, Abdel-Aziz, Prins, Maitland van der Zee, van den Born, Jousilahti, de Vos, Salomaa, Knight, Lahti, Nieuwdorp, Niiranen and Wiersinga2024). These are again associations, and further studies are needed to establish causality. If confirmed, however, BPGB are ideal GM members that could be the basis for interventions to modulate the risk of developing infections or reduce the severity of infection outcomes. Of these, Faecalibacterium prausnitzii, one of the most abundant BPGB in the healthy gut, has already been used safely in animal models. In one experiment, mice were given F. prausnitzii or vehicle and infected with Klebsiella pneumoniae after 4 weeks. The severity of pneumonia, organ damage, and sepsis was significantly lower in mice colonized with F. prausnitzii than in control mice, while their levels of CXCL1 and IL-6 were lower, indicating a downregulated immune response. A review based on this observation in a cohort of patients with community-acquired pneumonia found that patients with higher stool levels of F. prausnitzii had a shorter hospital stays than those who lost it (Kullberg et al., in preparation). Thus, the risk or severity of infection may be a treatable feature that warrants further investigation.

The future holds two possibilities. One is new treatment options to prevent and treat severe infections and prevent hospital readmissions by administering the right GM-modulating intervention at the right time (preventive, in-hospital, or immediately after hospital discharge) (Haak et al., Reference Haak, Prescott and Wiersinga2018b). However, while there is evidence to support this treatment option, physicians are still cautious about translating these findings into clinical practice due to safety concerns. The other possibility is based on the awareness of the importance of the gut anaerobic microbiota in severe infections. Data from critically ill patients, for example, admitted to the emergency department and patients with sepsis, show that patients unnecessarily treated with anti-anaerobic antibiotics have a poor prognosis and a higher risk of mortality (Kullberg et al., Reference Kullberg, Haak, Chanderraj, Prescott, Dickson and Wiersinga2025). Physicians are responsible for careful antibiotic selection and should not administer an anti-anaerobic antimicrobial unless indicated. In this regard, the development of lolamicin, a new antibiotic that selectively targets harmful Gram-negative bacteria while sparing the anaerobic microbiota, is one of the most important recent advances in this field (Muñoz et al., Reference Muñoz, Ulrich, Vasan, Sinclair, Wen, Holmes, Lee, Hung, Fields, Tajkhorshid, Lau and Hergenrother2024).

The same can be said for the development of consortia of “next-generation probiotics” such as the one composed of the commensal bacteria Blautia producta (a lantibiotic producer) and its cross-feeders, Bacteroides sartorii, Clostridium bolteae, and Parabacteroides distasonis (Kim et al., Reference Kim, Becattini, Moody, Shliaha, Littmann, Seok, Gjonbalaj, Eaton, Fontana, Amoretti, Wright, Caballero, Wang, Jung, Morjaria, Leiner, Qin, Ramos, Cross, Narushima, Honda, Peled, Hendrickson, Taur, van den Brink and Pamer2019).

In preclinical studies, this mixture demonstrated the ability to reduce the growth of vancomycin-resistant Enterococcus faecium, a major problem in stem cell transplantation (Caballero et al., Reference Caballero, Kim, Carter, Leiner, Sušac, Miller, Kim, Ling and Pamer2017). By understanding its principle and mechanism, the development of consortia with, for example, F. prausnitzii, is within reach in the near future. Professor Haak concluded by calling for collaboration to further explore and exploit the potential of microbiota-targeted therapies to prevent or treat severe systemic infections.

Professor Egija Zaura, Vrije Universiteit Amsterdam, The Netherlands

The mouth reflects systemic health and vice versa. Professor Zaura explained how the mouth affects the body, the benefits we receive as hosts when we are in balance with the oral microbiota (OM), and what happens when the balance is lost.

The OM changes at various stages of life. Teething causes a dramatic change as the teeth and the newly formed gingival crevice become a new microbial niche. Later, the OM changes as we transition from deciduous to permanent teeth, get braces or implants, lose some or all our teeth, or use removable dentures (Kaan et al., Reference Kaan, Kahharova and Zaura2021). Other characteristics of the OM are its complexity and resilience. The oral microbiome, particularly the tongue microbiome, has been shown in several studies to be the most stable in the human body (Liu et al., Reference Liu, Tong, Zhu, Tian, Jie, Zou, Lin, Liang, Li, Ju, Qin, Zou, Lu, Zhu, Jin, Xu, Yang, Wang, Zong, Liu, Hou, Jia and Zhang2021). Resilience is essential because the mouth is under constant stress from a variety of factors, one of which is carbohydrates in the diet. Studies of dental tartar from ancient dentitions have shown that the introduction of carbohydrates into the diet during the Neolithic period and flour and sugar during the Industrial Revolution caused drastic changes in its composition (Adler et al., Reference Adler, Dobney, Weyrich, Kaidonis, Walker, Haak, Bradshaw, Townsend, Sołtysiak, Alt, Parkhill and Cooper2013). Interestingly, the mouth condition of people who lived as if they were in the Stone Age (foraging for food and without oral hygiene) for 4–6 weeks in a reality TV competition had more plaque, but it contained more health-related bacteria, and their gums were not inflamed (Baumgartner et al., Reference Baumgartner, Imfeld, Schicht, Rath, Persson and Persson2009). This suggests that plaque itself is not inherently pathogenic, but dietary sugars promote the growth of acidogenic and pro-inflammatory taxa within plaque. Sugar reduces microbial diversity and leads to the dominance of bacteria that produce acids that dissolve tooth enamel. It also causes inflammation and alters the immune system so that it no longer responds properly to the microbes that accumulate in the mouth.

We still do not know exactly what constitutes a healthy OM, but it can be thought of as a balanced ecosystem from which the host derives numerous benefits. A study of pacifier cleaning practices showed that infants of parents who cleaned pacifiers by sucking on them were less likely to develop allergies than infants whose parents used other thorough cleaning methods (Hesselmar et al., Reference Hesselmar, Sjöberg, Saalman, Åberg, Adlerberth and Wold2013). At 4 months of age, the composition of the salivary microbiota was different between the two groups, suggesting that the reduction in allergies was due to immune stimulation by microbes transferred from parents via saliva. Another example of OM’s beneficial contributions to our health is its role in nitric oxide (NO) metabolism. The precursor of NO is dietary nitrate, and it must be reduced to nitrite before it reaches the stomach to be converted to NO, a vasodilator that helps lower blood pressure. As we do not have the enzyme to carry out this reaction, the OM does it for us (Hezel and Weitzberg, Reference Hezel and Weitzberg2015). From this point of view, broad-spectrum antiseptics (e.g., chlorhexidine) containing mouthwashes should be used with caution because they may disrupt the OM, leading to a reduction in nitrite bioavailability.

A disease in the mouth is comparable to an ecological disaster, but the small size of the oral cavity often leads to an underestimation of its severity and impact on systemic health. For example, in periodontal disease, where the jawbone in the mouth is reduced and the gums are severely inflamed, microbes and their toxins can enter the bloodstream through the open vessels of the ulcerated gums and affect overall health. In a meta-analysis of C-reactive protein (CRP) in relation to periodontitis, most of the studies showed that CRP levels were higher in patients with periodontitis than in controls (Paraskevas et al., Reference Paraskevas, Huizinga and Loos2008). The mechanism by which the dysbiotic OM creates a pro-inflammatory state that affects systemic health has been reviewed by Hajishengallis (Reference Hajishengallis2015). Another example of the association between periodontitis and systemic health is diabetes (Genco and Sanz, Reference Genco and Sanz2020). In clinical practice, it is possible to identify undiagnosed pre-diabetic or diabetic individuals among patients with periodontal disease. Conversely, people with diabetes or pre-diabetes are more likely to have periodontal disease. Proper treatment of periodontal disease improves CRP levels and restores glycaemic control in people with diabetes.

Regarding the relationship between the OM and GM, a recent study found no evidence of oral bacterial colonization of the gut in healthy adult samples from a shotgun metagenomic database (Rashidi et al., Reference Rashidi, Gem, McLean, Kerns, Dean, Dey and Minot2024). However, OM can migrate and exacerbate a disease by producing virulence factors when there is disease-induced dysbiosis in the gut. A recent review describes the underlying mechanisms (Kunath et al., Reference Kunath, De Rudder, Laczny, Letellier and Wilmes2024). It is also possible that oral bacteria associated with periodontitis contribute to gut dysbiosis. One study compared the composition of the OM and GM in periodontitis patients with controls and evaluated the effect of periodontitis treatment. Both oral and gut microbiota of patients differed from controls at baseline, and periodontitis treatment was associated with changes in both, returning to profiles similar to controls (Baima et al., Reference Baima, Ferrocino, Del Lupo, Colonna, Thumbigere-Math, Caviglia, Franciosa, Mariani, Romandini, Ribaldone, Romano and Aimetti2024).

Professor Zaura concluded by expressing her concern that studies using OM to predict outcomes in a variety of diseases often go wrong because they attempt to conclude from highly heterogeneous data. Her team experienced this when studying whether the OM of people with AD differed from that of healthy people (Pruntel et al., Reference Pruntel, Leusenkamp, Zaura, Vissink and Visser2024). She emphasized that when studying the OM, data on oral health, lifestyle, general health, demographics, and biology, as well as technical factors, are the minimum information needed to draw sound conclusions (Zaura et al., Reference Zaura, Pappalardo, Buijs, Volgenant and Brandt2021).

Professor Marie-Christine Simon, University of Bonn, Germany

In the MGBA, the gut is not only home to the microbiota but also has a neuroendocrine role that is involved in the regulation of systemic metabolism and energy regulation. Dr Simon’s work has focused on the role of gut peptides and hormones (Huber et al., Reference Huber, Sproten and Simon2021; Plassmann et al., Reference Plassmann, Schelski, Simon and Koban2022) in the bidirectional communication pathways of the MGBA, and she explains the results of her nutritional interventions intended to modulate the microbiota and gut peptides to influence metabolism and neurological function.

GLP-1 is a peptide hormone secreted by intestinal neuroendocrine cells in response to food intake that acts in the central nervous system to mediate satiety. This was first reported by Turton, who injected GLP-1 into the brain ventricles of fasting rats and observed a dose-dependent inhibition of feeding (Turton et al., Reference Turton, O’Shea, Gunn, Beak, Edwards, Meeran, Choi, Taylor, Heath, Lambert, Wilding, Smith, Ghatei, Herbert and Bloom1996). Years later, mice were shown to express GLP-1 receptors in brain regions relevant to satiety and appetite regulation (Jensen et al., Reference Jensen, Pyke, Rasch, Dahl, Knudsen and Secher2018). It was only in 2022 that vascular smooth muscle cells and, importantly, astrocytes of the hippocampus of the human brain were found to express GLP-1 receptors (Yang et al., Reference Yang, Vest, Kern, Lee, Agam, Maat, Losada, Chen, Schaum, Khoury, Toland, Calcuttawala, Shin, Pálovics, Shin, Wang, Luo, Gate, Schulz-Schaeffer, Chu, Siegenthaler, McNerney, Keller and Wyss-Coray2022). The hippocampus, the primary structure affected in AD, modulates memory aspects of feeding and is involved in appetite regulation along with the hypothalamus.

The first clinical application of GLP-1 analogues (GLP1As) more than 10 years ago was in the treatment of T2D, based on their ability to increase insulin secretion and modulate glucagon release from the pancreas. Since their recent approval for the treatment of obesity, their use in weight management has skyrocketed worldwide (Pi-Sunyer et al., Reference Pi-Sunyer, Astrup, Fujioka, Greenway, Halpern, Krempf, Lau, le Roux, Violante Ortiz, Jensen and Wilding2015; Kelly et al., Reference Kelly, Auerbach, Barrientos-Perez, Gies, Hale, Marcus, Mastrandrea, Prabhu and Arslanian2020; Jastreboff et al., Reference Jastreboff, Aronne, Ahmad, Wharton, Connery, Alves, Kiyosue, Zhang, Liu, Bunck and Stefanski2022). Numerous GLP1As have since been developed, some of which have also been shown to have cardioprotective and neuroprotective effects (Muscogiuri et al., Reference Muscogiuri, Cignarelli, Giorgino, Prodam, Santi, Tirabassi, Balercia, Modica, Faggiano and Colao2014; Tai et al., Reference Tai, Liu, Li, Li and Hölscher2018; Siddeeque et al., Reference Siddeeque, Hussein, Abdelmaksoud, Bishop, Attia, Elshazli, Fawzy and Toraih2024), and have even been used off-label in the treatment of AD.

The mechanism behind the neuroprotective effect of GLP1A has been studied in animal models and has shown that it reduces neuroinflammation and oxidative stress in the cortex and the hippocampus, which reduces neuronal apoptosis and neurotoxicity, and improves synaptic plasticity and insulin signalling (Yaribeygi et al., Reference Yaribeygi, Rashidy-Pour, Atkin, Jamialahmadi and Sahebkar2021). However, the role of GLP-1 in cognition is an area that requires further research.

Bringing the focus to the GBA and how GLP-1 secretion in the gut can be modulated by dietary interventions, Professor Simon presented the results of a double-blind RCT, in which glucose-tolerant participants consumed Li. reuteri or placebo for 4 weeks. Faecal microbiota, GLP-1 secretion, and insulin response (via the incretin effect) were measured at baseline and at the end of the intervention. GM composition did not change except for the presence of the intervention strain, and Li. reuteri increased the glucose-stimulated GLP-1 release along with higher insulin secretion compared to placebo (Simon, Reference Simon2012; Simon et al., Reference Simon, Strassburger, Nowotny, Kolb, Nowotny, Burkart, Zivehe, Hwang, Stehle, Pacini, Hartmann, Holst, MacKenzie, Bindels, Martinez, Walter, Henrich, Schloot and Roden2015). A study in mice showing that microbiota regulates basal levels of GLP-1 has also been published (Wichmann et al., Reference Wichmann, Allahyar, Greiner, Plovier, Lundén, Larsson, Drucker, Delzenne, Cani and Bäckhed2013). In terms of mechanisms, in vitro and animal studies have shown that SCFAs regulate glucose homeostasis and appetite by stimulating GLP-1 secretion through interaction with host G-protein coupled receptors GPR41 and GPR43. Given the central role of the microbiota in the production of SCFAs in the gut, it is plausible that changes in the GM alter the secretion of GLP1 and other effectors that influence satiety and postprandial glycemia (Simon and Müssig, Reference Simon and Müssig2015).

Professor Simon’s group conducted a three-arm nutritional intervention study in which cognitively healthy obese adults consumed a Nordic diet (ND), a vegetarian diet (VD), or a habitual diet (control) for 6 weeks to investigate their effects on cardiometabolic risk factors, microbiome, plasma markers associated with neurodegenerative diseases (glial fibrillary acidic protein; neurofilament-light amyloid-beta, Aβ-42/40; phosphorylated tau, p-tau 181 and 231; and total tau [Teunissen et al., Reference Teunissen, Verberk, Thijssen, Vermunt, Hansson, Zetterberg, van der Flier, Mielke and Del Campo2022]) and metabolic markers along the GBA (GLP-1 and incretin effect). The kinetics of markers of AD and metabolism were studied during 180 min in two states: fasting (basal vs. 6 weeks) and postprandial (after ingestion of a mixed meal, basal vs. 6 weeks).

Anthropometric measurements did not change between the groups during the study. Both the ND and VD induced taxonomic shifts in the GM. The ND resulted in significant improvements in fasting lipid metabolism and liver function. ND-induced improvements in lipid metabolism were mediated by the genetic makeup and associated with the basal microbial composition. However, no diet-induced changes were observed in either GLP-1 or plasma biomarkers of neurodegeneration in fasting or postprandial states. A finding that was not expected was the sharp increase in plasma AD biomarkers after 120 min in the postprandial time course. To investigate these kinetic changes, they conducted an intervention study examining the effect of acute food intake on the plasma AD markers in cognitively healthy obese adults and compared them to fasting controls. The result confirmed that the kinetics of plasma AD biomarkers changed significantly in the postprandial state compared to fasting controls: some markers increased (e.g., Aβ-40 and 42) and others decreased (e.g., neurofilament light and β-tau) (Huber et al., Reference Huber, Ashton, Schieren, Montoliu-Gaya, Molfetta, Brum, Lantero-Rodriguez, Grötschel, Stoffel-Wagner, Coenen, Weinhold, Schmid, Blennow, Stehle, Zetterberg and Simon2023). Professor Simon’s group is currently investigating what these findings mean in terms of cognition and is conducting a study of GLP-1 in patients with AD or mild cognitive impairment.

There is still much to learn about GLP-1. One of the missing pieces of basic information is the normal fasting level of GLP-1, which is not known even in people with diabetes or obesity. Professor Simon and her team are working on this and on mechanistic insights for GLP-1 that need further study in the context of metabolic and neurodegenerative disorders, and we look forward to hearing about them in the future.

Professor Jingyuan Fu, University Medical Centre Groningen, The Netherlands

In the last two decades, many gut–organ axes have been described, confirming the systemic impact of the gut and GMB on our health (Afzaal et al., Reference Afzaal, Saeed, Shah, Hussain, Rabail, Socol, Hassoun, Pateiro, Lorenzo, Rusu and Aadil2022). Among the many mechanisms by which the GM interacts with the human body, Professor Fu’s research, as part of the LifeLines (LL) Cohort Study, focuses on how microbes interact with diet and the host to influence metabolism.

The LL Cohort Study is a prospective, population-based, longitudinal study investigating the links between genetics, environment, lifestyle, and diet in the development of disease (https://www.lifelines.nl). It includes the LL Deep sub-cohort that has undergone in-depth data collection and analysis of the determinants (among them the GMB) and mechanisms underlying diseases, the LL Dutch Microbiome Project, which collected samples and analysed the microbiota from the mouth, tongue, and nose, in addition to stool samples, and the LL NEXT, a cohort that is currently following pregnant women and their newborns. The extensive metadata collected (2,000 factors/participant), the deep “omics” analysis of the samples, and the longitudinal follow-up (at least 30 years) allow researchers to systematically dissect which factors determine human health and disease development.

Many of the plasma metabolites are products of host–microbe co-metabolism of dietary and other xenobiotic compounds. The following are examples of how these interactions are being unravelled. The first is TMAO, a metabolite associated with an increased risk of CVD. TMAO has precursors such as phosphatidylcholine, betaine, and L-carnitine that are rich in different diets, which are converted to TMA by the GM, absorbed and transported to the liver, where it is oxidized to TMAO. Analysis of the relationship between plasma levels of TMAO and its dietary precursors, GM, and host genetics, analysed in five Dutch cohorts, including LL, revealed that the GM was positively associated with the TMAO: precursors ratio, that diet was associated with concentrations of TMAO and its precursors (although could not explain the variation in plasma TMAO), and that host genetics was not associated with TMAO. In addition, TMAO was found to be associated with all-cause mortality in men, but not with CVD risk (Andreu-Sánchez et al., Reference Andreu-Sánchez, Ahmad, Kurilshikov, Beekman, Ghanbari, van Faassen, van den Munckhof, Steur, Harms, Hankemeier, Ikram, Kavousi, Voortman, Kraaij, Netea, Rutten, Riksen, Zhernakova, Kuipers, Slagboom, van Duijn, Fu and Vojinovic2024). The second example is bile acids (BAs), which have been implicated in obesity-related diseases such as metabolic dysfunction-associated fatty liver disease (MAFLD). BAs in plasma and faeces are subject to microbial and host metabolism. Primary BAs are synthesized from cholesterol in the liver, stored in the gallbladder, and secreted into the gut, where most are reabsorbed and returned to the liver (enterohepatic circulation). Some, however, are converted by the GM into secondary bile acids; 95% of bile acids recirculate between the intestine and the liver, while the remainder are excreted in the faeces or enter the bloodstream. A study measuring plasma and faecal BAs in individuals with obesity and analysing the effect of host genetics and GMB confirmed associations of plasma and faecal BAs with genetics and revealed correlations with GMB (Chen et al., Reference Chen, van den Munckhof, Schraa, ter Horst, Koehorst, van Faassen, van der Ley, Doestzada, Zhernakova, Kurilshikov, Bloks, Groen, Riksen, Rutten, Joosten, Wijmenga, Zhernakova, Netea, Fu and Kuipers2020). A follow-up study investigated how the latter occurs. Structural variants (SVs) of bacterial species in the GMB of two Dutch cohorts were characterized and systematically analysed against plasma BA parameters (Wang et al., Reference Wang, Doestzada, Chen, Andreu-Sánchez, van den Munckhof, Augustijn, Koehorst, Ruiz-Moreno, Bloks, Riksen, Rutten, Joosten, Netea, Wijmenga, Zhernakova, Kuipers and Fu2021). SVs based on deletions, insertions, or rearrangements of large genomic regions are common in bacterial genomes and are useful for identifying regions containing metabolically relevant genes. Indeed, the study found more than 800 associations. For example, the abundance of a Coprococcus comes variant correlated positively with the ratio of secondary: primary BAs, and the SV region was found to contain genes encoding bile salt hydrolases that deconjugate conjugated primary BAs, which can subsequently form secondary BAs.

Bacteria can also take up compounds synthesized in the human body as substrates for their metabolism. A study investigating how host genetics drives the genetic diversity of GMB using the data from four Dutch cohorts identified an association of the human FUT2 gene and the ABO locus with a particular strain of F. prausnitzii.

Blood group antigens determined by ABO can be secreted into the intestinal mucosa by the action of fucosyltransferase 2 encoded by the FUT2 gene, and the terminal oligosaccharide of blood group A antigen is N-acetyl-galactosamine (GalNAc), which is used as a carbon source by the F. prausnitzii variant of interest. Thus, this F. prausnitzii variant has a better chance of colonizing the gut and influencing health in a person with blood type A and the FUT2 secretory phenotype (Zhernakova et al., Reference Zhernakova, Wang, Liu, Andreu-Sánchez, Zhang, Ruiz-Moreno, Peng, Plomp, Del Castillo-Izquierdo, Gacesa, Lopera-Maya, Temba, Kullaya, van Leeuwen, Study, Xavier, de Mast, Joosten, Riksen, Rutten, Netea, Sanna, Wijmenga, Weersma, Zhernakova, Harmsen and Fu2024). This suggests that a blood type-based probiotic could be developed.

Some bacteria of the GM show temporal variation, and these changes may explain metabolic changes. Examination of faecal samples collected from a subset of the LL Deep cohort at baseline and 4 years later revealed that microbial genetic stability varied over time across species (Chen et al., Reference Chen, Wang, Garmaeva, Kurilshikov, Vich Vila, Gacesa, Sinha, Segal, Weersma, Wijmenga, Zhernakova and Fu2021). F. prausnitzii is one of the unstable species, with two distinct strain clusters associated with 15 metabolites, of which p-cresol sulphate, a microbiota-derived metabolite that is a uremic toxin, was negatively associated with F. prausnitzii strain 2. While most of the participants maintained either strain over the 4 years, some switched strains from strain 1 to 2 or from strain 2 to 1. In the former direction, plasma L-cresol sulphate decreased, whereas in the latter direction it increased, suggesting a causal relationship between them (Chen et al., Reference Chen, Wang, Garmaeva, Kurilshikov, Vich Vila, Gacesa, Sinha, Segal, Weersma, Wijmenga, Zhernakova and Fu2021).

Finally, Professor Fu described an in vitro model of the microbiota–gut–liver axis using organ-on-a-chip technology. Human induced pluripotent stem cells (hiPSCs) can be generated from somatic cells and further differentiated into intestinal and liver organoids. The chips (with an upper chamber mimicking the gut or liver microenvironment and a lower chamber mimicking the bloodstream) have been validated and are commercially available. In collaboration with a group that has successfully constructed a gut-on-a-chip from hiPSCs (Palasantzas et al., Reference Palasantzas, Tamargo-Rubio, Le, Slager, Wijmenga, Jonkers, Kumar, Fu and Withoff2023; Moerkens et al., Reference Moerkens, Mooiweer, Ramírez-Sánchez, Oelen, Franke, Wijmenga, Barrett, Jonkers and Withoff2024), her team is working with hiPSC-derived liver-on-a-chip and has demonstrated expression of cytochrome P450 and hepatocyte nuclear factor 4 alpha and production of albumin (Fu et al., Reference Fu, Qiu and Tan2023). Professor Fu concluded by outlining future research to integrate currently separate systems into one by co-culturing microbes and gut organoids on a single chip and connecting it to the liver-on-a-chip. As LL Deep has collected live GMB and biological samples with cells from each participant, it may be possible in the not-too-distant future to recreate the microbiota–gut–liver axis in an individual-specific way.

ROUNDTABLE: clinical trials in the nutrition field: a solution or a nightmare?

During the symposium, a roundtable discussion on the current state of clinical trials in nutrition and the scarcity of approved health claims for “biotics” and live dietary microbes (LDM) was held to analyse the key problems facing the field and consider practical solutions from different perspectives. Panellists included Dr. Gwen Falony (Host-Microbe Interactomics, Wageningen University, The Netherlands), Professor Wendy Hall (Nutritional Sciences, King’s College London, UK), Professor Flavia Indrio (Experimental Medicine, University of Salento, Italy), Professor Hans Verhagen (Food Safety & Nutrition Consultancy, The Netherlands), Professor Bruno Pot (Science Director, Yakult Europe, The Netherlands), and Dr. Emily Prpa (Science Manager, Yakult UK), who moderated the discussion.

Difficulties in setting up scientifically controlled human nutritional studies: the conundrum of how to apply drug clinical trial design to foods

Clinical trials in nutrition have come a long way in the past 25 years. Requirements such as pre-trial registration of interventional trials in a public registry and reporting of results according to guidelines such as the Consolidated Standards of Reporting Trials (CONSORT) have improved the ethical and scientific quality (e.g., transparency, completeness, consistency, and rigor) of trials and have facilitated interpretation, comparison, and pooling of data for meta-analysis.

However, while there is agreement on the need to support food health claims with sound science, and most participants agreed that clinical trials provide the highest level of evidence, many panellists acknowledged that it is challenging to apply clinical trial designs developed for testing drugs to foods. This is because there is a disconnect between the expectation and the application. Drugs and foods are fundamentally different. A drug interacts with specific biological targets to produce a specific response to prevent, treat, or cure medical conditions, whereas food has a different mode of action, providing nutrients to maintain the functioning of our organism and overall health. Thus, the objectives and target populations of interventional studies are inherently different. This, combined with the complexity of food consumption itself and the lack of consensus in the field of nutrition in general (e.g., different recommended dietary allowances for nutrients in different countries), makes it unsurprising that nutrition clinical trials face the challenges we see today and that a re-evaluation of the concepts and criteria in nutrition in general may be needed. The definitions of “clinical trial” by the European Union and the US Food and Drug Administration do not include food as an intended target of testing (European Union, 2001; Food and Drug Administration, 2020). The question is: are there scientifically sound methods that allow us to estimate the long-term benefits of foods in a large population without conducting clinical trials? Many study designs, including mechanistic, observational, and experimental (e.g., a collection of individual case studies, studies conducted only in responders, and a focus on subgroups), were proposed, but these would not be accepted as substantiating evidence by regulatory agencies for various reasons, including bias, lack of generalizability, and the low ranking of these study designs in the hierarchy of scientific evidence. The conclusion was that, despite the discrepancy, a human intervention study in the form of a well-conducted RCT is needed to improve the acceptability of the outcomes of nutrition studies.

The regulatory viewpoint

In the context of a health claim application, knowing the answer to the question “Does it (food or ingredient) work?” is a pre-requisite for starting to prepare a dossier. Since EFSA focuses meticulously on the applicant asking the right questions, the most important element when planning a clinical trial is to define the question(s) to be answered, then to design the right experiments (in accordance with established standards for the design and conduct of clinical trials) and to collect all the resulting data and surrounding information in the dossier. Mechanistic studies are helpful to illustrate that the effect exists but are not essential.

On the other hand, it may not be so problematic to show effects in nutritional studies, but compiling a consistent dossier may be the hurdle. If the studies included in a dossier show a beneficial effect, but have methodological inconsistencies (e.g., different primary endpoints between studies), that makes it difficult to conclude that a cause-and-effect relationship has been established and the health claim will not be approved. Including all relevant material in the dossier is also essential. EFSA will evaluate what is submitted and will not dive into the literature to find further information (except in specific cases where omission of negative data is suspected). Therefore, it is not only the clinical trials that are important, but also the careful preparation of the dossier.

Challenges and potential solutions in nutrition clinical trials involving the gut microbiota

Recognizing that clinical trials are the best study design to demonstrate the cause-and-effect relationship between the food product/component and the outcomes, the puzzle pieces that go into nutrition clinical trials in the GM field were discussed.

The challenges identified can be grouped into the following categories:

• • Food/intervention related (heterogeneity of biotics and LDM, differences in composition, and number of microorganisms, dosage, etc.)

• • Population related (exposome, age, genetics, metabolome, inter-, and intra-individual variation)

• • GM inherent (composition, diversity, microbiome, and metabolome)

• • Diet, lifestyle, and environmental factors

These are confounding factors that, when possible, must be controlled for in the studies. Conducting the trial itself presents challenges, such as the lack of good biomarkers, in addition to meeting the universal requirements of any clinical trial, such as adequate sample size, consistent methodology and analytical methods, and proper reporting.

The proposed solutions for modifying current study settings to overcome barriers and improve the acceptability of the outcomes of nutrition clinical trials can be grouped as follows:

• • Collaborative efforts as a community of nutritional scientists, microbiologists, clinicians, and industry to:

  1. a) Create standards for:

     1. i) conducting nutrition trials that are specific for each component

     2. ii) studying the microbiome (creation of a checklist)

  2. b) Define biomarkers of:

     1. i) each food or food component

     2. ii) unhealthy microbiota (a healthy microbiota cannot be defined, but an unhealthy one can. The approach of “preventing the occurrence of an unhealthy microbiota” is noticeably clear)

  3. c) Create or expand databases and make them available to all researchers on:

     1. i) microbiota patterns

     2. ii) microbiome composition

     3. iii) multitude of additives and other non-nutrient bioactives that may affect the GM/GMB (such as the “Phenol-Explorer”)

• • Individual efforts of research teams when designing and conducting RCTs:

  1. a) adhere to the principles of rigor and transparency

  2. b) identify the research question and formulate a hypothesis rather than following the recent trend of data-driven science

  3. c) define the study population (it must be representative of the target population of the health claim; at-risk populations can be used but the results of such studies may be difficult to extrapolate to the general population)

  4. d) ensure that the sample size is properly calculated (i.e., sufficiently powered)

  5. e) define simple and concrete outcomes, quantitative in nature

  6. f) use rigorously selected standardized dietary controls (or placebo), given the multiple confounding factors

  7. g) address inter- and intra-individual variation (use an appropriate study design, including randomization and stratification), and accepted statistical methods to account for inter-individual variation. Intra-individual variation determines statistical power more than inter-individual variation (Larsen et al., Reference Larsen, Claassen and Brummer2020), and to account for it, the study population must adhere to a strict regimen to maintain their lifestyle, including dietary habits, throughout the study)

  8. h) identify and address confounding factors (randomise the participants, analyse the data using stratification, regression, and other statistical tools. From a microbiological point of view, the most important confounding factor when studying GM is transit time. Therefore, options to measure it [e.g., orally ingestible capsules, non-absorbable dye] should be considered. If none are possible, a proxy like stool moisture could be measured)

  9. i) follow-up for longer time

  10. j) consider ethical and safety aspects

Clinical trials in nutrition are often criticized for a variety of these differences, even though trials can be of excellent quality and design. The question that was raised, though not answered, was whether the different standards that dietary trials are being held to by EFSA or other authorities need to be challenged.

The role of microbiology in nutrition clinical trials

The primary endpoint of nutrition clinical trials is a health outcome related to the host, not the GM itself. The interest in the GM relates to its assumed or accepted influence on the primary endpoint. It is now known that dysbiosis (previously referred to as the microbiota carried by a patient) is not a state exclusive to patients but can also be present in healthy individuals. As dysbiosis is a relatively stable state, it is not only possible but advisable for basic researchers to support clinical trials in nutrition by looking at the GM profile of the participants before starting a dietary intervention study and to randomize them additionally according to their microbiota, based on established expertise.

Regarding GM testing in the general population, a citizen science-type approach using GM test kits to collect samples at home has proven to be possible. In fact, this is already happening in Europe, for example, the LifeLines Deep Cohort Study (The Netherlands), the Flemish Gut Flora Project (Belgium), and the Le French Gut Project (France). It can be said that sufficient samples are available to study GM/GMB diversity, and data are accessible to all researchers by submitting an application. Essential to this approach is that necessary metadata is simultaneously collected, to allow the use of those samples in a proper randomization process.

Other interesting points

An audience member commented on the discrepancy between what regulatory agencies and academia expect from research. Whereas regulatory agencies look for replication of results, academic science is under pressure to publish in high-impact journals that look for novelty when reviewing submissions. This is inherently contradictory, but in addition, the funding for replication studies would come from industry, leading to scepticism about the results, even if the studies are conducted to rigorous standards. Another comment was made about the proliferation of food products that claim to provide health benefits by adding an ingredient whose health claim has been approved by regulatory agencies (e.g., vitamin C to claim immune system effects). This is not what the food safety agencies were seeking with the implementation of the health claims legislation. These are issues that have implications for nutritional intervention studies at another level that also must be resolved.

Many challenges remain in improving the acceptability of the outcomes of nutrition clinical trials, but they must be seen as windows of opportunity, and solutions should be sought through collaborative efforts to advance the field and ultimately protect consumers and improve their health.

Concurrent with the discussion, a poll was conducted in the audience, and a word cloud of the responses was projected on a screen. The four poll questions and the audience responses are summarized in Table 1.

Table 1.

The result of a 4 question poll conducted in the audience; question and audience responses

Table 1. Long description

The table consists of four rows of data under the headers Question and Top words.

* Row 1: For the question ‘What are the confounding factors you consider most important to control?’, the top words are Diet, lifestyle, age, health status, exercise/physical activity, nutrition, sex, and stress.

* Row 2: For the question ‘Which sub-healthy or at-risk population do you think are suitable for nutritional studies?’, the top words are I B S, obese, metabolic syndrome, older adults, menopausal women, stressed, eczema, athletes, A D H D, and prediabetes.

* Row 3: For the question ‘Are validated biomarkers an absolute must for nutritional studies (versus survey-like outcomes)?’, the results are Yes (32) versus No (12).

* Row 4: For the question ‘Which alternative (to) clinical settings would you prefer (more than one answer possible)? You can propose your own approach!’, the top words are overwhelmingly citizen science, followed by pop-up labs, total evidence load, user trials, and 3 D gut models.

Critical perspectives and forward-looking analysis

The realization that the microbiota-containing organs and their interactions with each other and the host are essential to the physiology and homeostasis of our organism was evident at this symposium. It was also clear that the exposome, including perinatal factors, hormones, urban/rural living, and especially diet, influences the microbiota.

Through a series of enlightening presentations, new applications for GMB profiles, such as the possibility of predicting responders from non-responders to a treatment, or personal glycaemic responses useful in the management of diabetes, were explored, next to practical applications arising from the discovery of novel microbial species that can help produce “pitched” foods or offer the prospect of producing “next-generation probiotic consortia.” All these ideas, together with the introduction of tools such as ingestible robotic capsules to study the GI tract and the organ-on-a-chip approach, allowed us to envision an open horizon for future research and applications.

Recurrent themes, repeatedly mentioned, that will need to be addressed in the future to improve the acceptance and reliability of microbiome-based interventions are the design of clinical trials, especially to take into account the individual microbiota variability (consider studying subpopulations) and the difficult-to-control confounding factors (diet, stress, age, environment,…) in order to further improve the study reproducibility. Besides the difficulties for true placebo-controlled trials in food studies, the need for validated biomarkers not only related to disease, but also to health, has also been brought up on several occasions. The role of the gut microbiome, as the conductor of health, occupies an important, central role in maintaining health. Microbiota-based parameters have even been suggested as the basis for substantiating a health index. It is therefore extremely important to keep a finger on the microbiome pulse in order to detect early shifts in (bio)markers that are known to be related to health risks.

The future of microbiota research seems to cover a broad variety of research fields, from further developments in food-oriented clinical trial types, improved technologies to study and understand the complicated metabolism of the different ecosystems in, for example, the small bowel versus the large bowel (Jensen et al., Reference Jensen, Heyndrickx, Jonkers, Mackie, Millet, Naghibi, Pærregaard, Pot, Saulnier, Sina, Sterkman, Van den Abbeele, Venlet, Zoetendal and Ouwehand2023), the importance of cross-feeding in the design of pro-, pre-, or postbiotics, to tools that will allow us to better understand the interaction with the host (immunological, physiological, and endocrinological). Fundamental, technological, and applied research will need to march hand-in-hand to answer the remaining questions, certainly raising new questions but without any doubt opening new doors for new discoveries, new diagnostics, and new products to prevent or even cure disease.

During this symposium, we were also made aware of the growing awareness of the public and their participation in cohort studies, allowing us to collect large amounts of data with long follow-up times, essential to the above-mentioned research. The interest of the public might also help to drive the regulation towards a more homogeneous approach in guaranteeing safe and efficient solutions.

The same urgency for the accumulation of massive datasets will also stimulate the development of international research consortia, in charge of multinational research projects that will help to establish the databases necessary to encompass and cover the huge individual variety and the large number of confounders that need to be considered in the further exploration of the conductor role of the gut microbiota. There is no doubt that we can expect a fast increase in knowledge, leading to solutions to current challenges, and guaranteeing continuous innovation in the field.