Summary Experts Meeting on Microbiome & Human Health

In the development of new products for human health, there are both opportunities and challenges. Infant Health is an important focus because microbiome dysbiosis results in poorer health outcomes later in life. How can genomics help us shape the infant microbiota to ensure all infants have a healthy outlook?

Probiotics are widely used for general health and also to prevent conditions such as antibiotic-associated diarrhoea. What are the coming trends in probiotics, and how can we use them to improve human health? Are there new application areas and what do clinical trials show?

To answer these questions, BaseClear held an Experts Meeting on the topic “Microbiome and Human Health” on Thursday December 2, 2021, chaired by Dr. Derek Butler and Dr. Radhika Bongoni. Seven microbiome scientists presented on varied topics within human health, and each session was followed by a panel discussion. Over 80 people attended the online presentations.

Session 1: Early Life Nutrition and Infant Health

Presentation 1: The development of early life gut microbiota and its implication in later life
Dr. Himanshu Kumar, Senior Scientist – Danone

The composition of the gut microbiota changes over the life cycle. In infants, the gut microbiota is dominated by Bifidobacteria, especially in breastfed infants. Once infants start to receive solid food, their microbiome changes profoundly. Factors such as malnutrition and antibiotic use can also have a significant impact on the microbiome at any age.

In infants, certain eco-physiological factors drive maternal transmission, colonization and succession of the gut microbiome. These include mode of delivery, the composition of infant feed whether it is formula or breastmilk, and cross-feeding between microbiome species that affects short chain fatty acid production and pH, thereby influencing the microbiome as a whole. There is a trend towards increasing species diversity in the microbiome through the lifecycle that stabilizes in childhood. Certain events that happen during from birth can negatively affect both health and the microbiome such as disease. For example, infants born by Caesarean section or who receive antibiotics early in life have a greater risk of allergies, eczema and asthma, and there are differences in their microbiome compared to vaginal births. Are there strategies to modify the microbiome of infants born by Caesarean section to resemble that of vaginally-delivered infants, and can that affect allergy symptoms?

The JULIUS study (main study report) was conducted to see whether prebiotics and a symbiotic treatment could reduce the effect of Caesarean section on the microbiome. The study randomized 153 elective Caesarean section infants to a control group, prebiotic treatment (scGOS/lcFOS) or synbiotic treatment (Bifidobacterium breve M-16V plus scGOS/lcFOS). The intervention time was 16 weeks and follow-up 6 weeks. A reference group of 30 vaginal delivery infants was also included. The infant gut microbiome, faecal pH and short chain fatty acids, and safety and tolerance of the intervention were measured.

By comparing the reference and control groups, delayed colonization with Bifidobacterium over the first 4-8 weeks was shown in the Caesarean section group. The infants given the symbiotic from birth were colonized with Bifidobacterium in a similar way to the reference group from the first few days of birth, thus reversing the dysbiosis found in the control group. The faecal pH was also significantly reduced compared to the control group, indicating a more favourable environment for a normal infant microbiome. Infants in the synbiotic group also had a reduction in skin-related disorders and eczema, indicating that the treatment both modified the microbiome and had beneficial effects on health for the infants.

The synergistic effect of the combination of probiotics and prebiotics was beneficial for microbiome modulation and health effects. The delayed colonisation with Bifidobacterium found in Caesarean section infants was prevented and there was a possible preventive effect on eczema.

Antibiotics are an important part of health care and are used to treat many diseases in infants. For example, neonatal sepsis and pneumonia can be life threatening. However, there are some negative effects associated with antibiotic usage later in life. For example, antibiotic usage is associated with obesity. Antibiotics also negatively affect the microbiota: use of antibiotics had a clear effect on the development of the gut microbiome of infant boys in an observational study from Kumar et al. Studies performed using faecal transplant in mice show that the microbiome of antibiotics-exposed animals impaired growth when transferred to germ-free male mice. Growth impairment associated with antibiotic use may carry long term physiological consequences.

In another area of neonatal health, the effect of mild therapeutic hypothermia for neonatal hypoxic ischemic encephalopathy (NHIC) on the microbiome has been researched. NHIC is a brain dysfunction caused by a lack of oxygen or blood flow to the brain, which can occur during pregnancy, labour, delivery or the postnatal period. There can be severe and permanent disability that results, including cerebral palsy or epilepsy. Cooling reduces the oxygen demand of organs and can prevent long term effects. The results of an exploratory study show that cooled babies had a reduced alpha diversity, higher abundance of Staphylococcus and lower abundance of Bacteroides. Is it possible to develop an intervention to revert the microbiome of cooled babies to one similar to healthy controls? Could that have a beneficial health effect?

There are many applications of microbiome-modulating therapies in neonatal health, and they could be very important due to follow-on effects later in the lifecycle.

Presentation 2: Advanced biostatistical analysis to enhance microbiome data mining for health discoveries
Dr. Eline Klaassens, Product Manager Human Health – BaseClear

There are many genomic tools available that are instrumental in developing microbiome-based human health solutions. Attention to the different parts of the research process can yield the best results and most value from the data.

  1. The right approach to sampling and logistics ensures that samples are collected in a reliable way and can be analysed successfully. Care here must be taken in ensuring that adequate biomass is collected.
  2. DNA/RNA extraction is an important part of the process, especially for (meta) transcriptomics.
  3. During microbial profiling, the technique used is important. Do you need specific or non-specific metabolomics? Is strain-specific qPCR required? Do assays need to be developed or validated?
  4. Bioinformatics and visualisation are an important step. Dedicated databases can improve the robustness of the results; these can also be customised with proprietary strains.
  5. Meta-data collection is becoming increasingly important, especially for key microbial biomarkers.
  6. Reporting and interpretation of large datasets requires expert knowledge.

A dedicated database is one that contains microorganisms that can be found in the ecosystem that has been sampled. Using a dedicated database has a number of important advantages. It enables the use of shallow shotgun data for taxonomic classification and can sample all domains including archaea and viruses. It improves the proportion of classified reads and allows classification to strain level. Strains of interest can be added, and the database can be validated in several different ways. When using a dedicated database, there is no need to filter out species with a low abundance, thus a more accurate picture of the microbiome can be obtained. This enables improved detection of key microbial biomarkers. Dedicated databases include:

  • Infant (recently developed)
  • Skin (includes microbes found in different skin niches)
  • Pathogen (focus on the food industry and low-abundance pathogens)
  • Animal
  • Phage
  • Human gut

The taxonomic profiling pipeline takes paired-end reads, optionally with subsampling and host-filtering, and assigns taxonomic labels to them via the Kraken 2 database. Using Bracken, abundance at different taxonomic levels is calculated.

The Kraken 2 tool for metagenomic analysis has several advantages over the Kraken 1 tool, reducing memory usage substantially while maintaining high accuracy and speed (see Wood et al., 2019). The Kraken 2 approach is based on minimizers rather that then k-mers used in Kraken 1. Kraken 2 uses a probabilistic, compact hash table to map minimizers to LCAs. The minimizer itself is used as a sub-string for comparison to the reference set, reducing the memory requirement and allowing a faster processing time. Other advantages include better determination of false and true positives based on the number of and the ratio of distinct minimizers.

Key Microbial Biomarker selection is an important tool in human health applications. It allows the development of diagnostic tools for early disease detection, there are applications in personalized medicine, and patients can be stratified on relevant biomarkers in clinical studies. Overall, it allows an improved understanding of results in pre-clinial and clinical trials, particularly in the development of live biotherapeutics. Microbial biomarkers or microbial signatures can be used. Machine learning tools can also be applied, and have been applied successfully in inflammatory bowel disease and colon cancer studies.

For Key Microbial Biomarker selection, an ensemble method is used that combines the results of four different models to overcome weaknesses of individual methods. When the results using different methods agree, this lends strength to the overall conclusions. The ensemble methods are:

  1. Feature importance evaluation using Random Forest
  2. LASSO logistic regression
  3. LEFSe biomarker validation
  4. Differential abundance analysis

The results of the methods can be plotted to check the performance of the models. This method has been validated with data from colorectal cancer and with gut microbiome analyses. A custom report is produced after the analysis.

For human health and infant health applications, there are a number of microbial genomics tools available to enhance research. They allow species- and strain-level identification of microbes, use dedicated databases to extract more from the data, and allow the detection of false positives. Machine learning models are available to select Key Microbial Biomarkers and Microbial Signatures. The main applications are for the stratification of patients in clinical trials, the development of diagnostics, and selection of biotherapeutics.

Presentation 3: From Mum to Bum, how human milk oligosaccharides shape the microbiome
Dr. Clara Belzer, Associate Professor, Wageningen UR

The microbiota of infants starts to develop at birth with transplantation of many species from the mother. The pioneer species received from the mother are important, and they derive from the maternal vaginal flora, lower intestinal tract and breast milk. The pioneers reduce oxygen tension, thus favouring the succession of strict anaerobes. Common infant gut microbiome inhabitant Bifidobacterium can help with the digestions of human milk.

Microbiome acquisition is affected by many different factors. Those that have been investigated include genotype, diet, health status, antibiotic use in the mother, mode of delivery and feed type, and infant genotype. There are a number of relatively unexplored areas of research such as cross-domain interactions, the oral microbiota, changes in the maternal microbiome and medical-based feeding strategies.

Infants born via Caesarean section are different to infants born vaginally. Recently a proof-of-concept study was performed to see whether maternal faecal microbiota transplantation could restore Caesarean-section born infants (see Korpela et al., 2020). Faecal microbiota samples taken from the mother three weeks before planned Caesarean section was used to inoculate infants after birth. The infants were successfully colonised and had a microbiome that resembled that of vaginal delivery infants after the intervention.

Human breastmilk contains a wide range of oligosaccharides, and various immune molecules. It is involved with the maturation of the infant immune system. Bifidobacteria are specialised in using Human Milk Oligosaccharides (HMOs), explaining their presence particularly in the gut microbiome of human infants. The diversity of HMOs offers many ecological niches for different strains of Bifidobacterium, that not only degrade HMOs, but also acidify the gut, degrade sugars, produce vitamins and may be involved in microbial crosstalk or signalling within the gut. Even though the abundance of Bifidobacterium are similar between breast- and formula-fed infants by one month of age, the specific composition of different strains differs. Infant formula now contains the HMO 2-fucosyllactose to more closely resemble breast milk.

The discovery that Bifidobacteria can use urea (found in breastmilk) as a source of nitrogen has added an extra dimension to research into human milk. Growth was stimulated by the addition of urea in in vitro experiments. Some Bifidobacteria appear to be involved with cross-feeding and therefore stimulate a more diverse microbiome.

The use of antibiotics reduces or delays colonisation by Bifidobacterium. The “Toddlers receiving synbiotics after antibiotics” or TOBBI study was designed to assess whether an intervention can restore a normal gut microbiome (see the website: https://www.wur.nl/en/project/tobbi-trial.htm).

In summary, the transmission of microbes from mother to child depends on many factors. Mode of feeding and nutrients stimulate the development of the microbiome. The growth of Bifidobacteria is stimulated by HMO and urea, so their growth can be modulated via prebiotics to produce a beneficial role in the infant gut.

For readers interested in more in-depth knowledge about the microbiome in human health, there is a MOOC available on the subject via this link: https://www.edx.org/course/nutrition-and-health-human-microbiome

Presentation 4: In vitro and in silico Human Milk Oligosaccharide degradation capacity by bifidobacteria
Dr. Gerben Hermes, Senior Scientist – Chr. Hansen

Human milk is a complex substrate, and it is unique in the animal kingdom due to the wide range of Human Milk Oligosaccharides (HMOs) present. See Laursen et al. for a summary of normal infant microbiome development. The HMO content of breastmilk is responsible for high numbers of Bifidobacterium. In fact, infants do not produce the glycosidases that digest HMOs, but Bifidobacteria do.

Recently, an infant formula containing the five most common HMOs was tested for safety and tolerability (Parschat et al., 2021). The formula appeared safe and well-tolerated, and there were several changes in stool consistency such that it was more similar to breastfed infants than formula fed.

Several challenges still remain regarding how Bifidobacteria degrade HMOs as it pertains to infant feeding. Understanding of consumption pathways is incomplete, and there is a lack of knowledge of the effect of how different species and strains affect assimilation of HMOs.

In vitro tests were done to test the growth of different strains alone and in combination using HMOs. It seems that HMO utilisation is strain-specific, and only a few strains could grow on some HMOs. Clusters of HMO degraders were species-specific. Gene prediction on glycosyl hydrolase genes was used to predict behaviour in the gut. HMO utilisation strategies are species-specific via gene clusters, and there were large differences in gene content at the strain level. The in silico findings need to be confirmed with experimental data.

Sialidases cleave the terminal sialic acids from complex carbohydrates including HMOs, freeing them for use as an energy source or for brain growth and development. Different Bifidobacterium strains differ in their specificity for sialidases on sialylated HMOs. This could be a differentiator in the infant microbiome.

Session 2: Probiotics and their health benefits

Presentation 5: Microbiota management: a one health perspective
Dr. Olaf Larsen, Senior Manager Science – Yakult

Since the 1950s, the incidence of many common infectious diseases has dropped due to health care interventions such as improved hygiene and vaccination. Over the same period of time, non-communicable disease (NCD) incidence has increased, including immune-related diseases such as type I diabetes and asthma. The hygiene hypothesis suggests that microbiome changes due to better hygiene have caused the increase in NCDs.

The gut microbiome is also associated with the response to SARS-CoV-2 infection, and the risk of serious illness increases with several NCDs as well as microbial dysbiosis. A holistic approach looking at inter-kingdom cross-talk could offer novel insights. Interestingly, prior infection with a helminth is associated with a reduction in inflammation-related co-morbidities (Cepon-Robins and Gilmore, 2020).

After reaching adulthood, we tend to lose microbial diversity in the gut. Due to improved hygiene, it is now more difficult now to replenish lost microbes. What are the key determinants of ecosystem structure and stability, and their associated functions? Are there inter-ecosystem determinants?

Some interventions have been able to increase biodiversity in humans, with beneficial effects on the immune system. For example, increasing plant biomass in highly urbanised environments increases microbial diversity in humans, and one study conducted in day care children showed beneficial changes in immune regulation (Roslund et al., 2021).

One way to look at the health of microbial ecosystems is by considering “keystone taxa” – these are specific microbes that interact with others and are key to maintain the structure of the entire ecosystem. Keystone taxa are indispensable for the recovery of the ecosystem from a shock such as antibiotic use, and can predict the collapse of the ecosystem into dysbiosis. A metric has been developed, called a “K core” used to specify the number of keystone taxa (see Morone, Del Ferraro and Makse, 2019). A similar system using microbial “guilds” has also been developed by Wu and colleagues (2021). Researchers can identify guilds to work out which are important for the robustness and functionality of the microbiome. The number of guilds needed depends on the entire structure of the ecosystem, but generally a higher number is considered to be more stable. A practical application is that by identifying key species required, potentially risky procedures such as faecal transplantation can be avoided as only several known species are used.

Presentation 6: Microbiome and probiotics in metabolic health
Dennis Zeilstra – Senior Science Analyst – Winclove Probiotics

Metabolic health is defined by having normal blood sugar, triglycerides, cholesterol, blood pressure and waist circumference without the need for medication. Many chronic diseases are connected to metabolic health. The two hallmarks of metabolic disease include systemic low-grade inflammation and insulin resistance, which occurs when cells have a decreased sensitivity to insulin. It is a normal process, however when present chronically it is associated with health issues.

The microbiota is linked to insulin resistance and systemic low-grade inflammation. There are three proposed mechanisms: (1) the gut microbiota regulates immune development and activity; (2) the gut senses nutrients; (3) nerve cells in the gut regulate the glycaemic response, and are in contact with microbes. The immune response is advantageous but requires a lot of energy to maintain. Metabolism and immunity co-evolved, thus humans and their commensals can be considered holobionts.

There is potential for metabolic health to be modulated with probiotics. Strain selection is important. A probiotic mixture (Ecologic® Barrier from Winclove Probiotics) consisting of various Bifidobacteria, Lactobacillus and Lactococcus species has been tested for its effect on inflammation, the immune system and barrier function. A recent clinical trial investigated this probiotic in 96 treatment-naïve type 2 diabetic patients. The intervention, given daily for 6 months, reduced HOMA and systemic inflammation compared to control. Another clinical trial conducted in 81 obese postmenopausal women gave the same probiotic mixture for 12 weeks as a low-dose and high-dose treatment. This second trial showed an improvement in insulin resistance in the two probiotics groups. These studies show that probiotics can have a clinical benefit in metabolic disorders.

Presentation 7: The human vulvar microbiome
Lisa Pagan, Research Physician – Centre for Human Drug Research

There is a lack of knowledge about the vulval microbiome, what is considered healthy, and how it contributes to vulval disease. For example, vulval cancer is the “forgotten cancer” in women, making up 5% of gynaecological cancers. Some cancers are related to high-grade squamous intraepithelial lesions (HSIL) caused by the human papillomavirus (HPV), while others are not. The condition lichen sclerosus, which causes blotchy, red, itchy patches on the genitals, increases risk of vulval cancer. It is also uncomfortable with peak incidence in young girls and postmenopausal women, and causes considerably morbidity. The corticosteroid treatment is messy and must be used lifelong, and there are no new treatments on the horizon. The microbiome may be implicated in vulvar disease, and further research is needed to identify mechanisms, biomarkers of disease, or drug development targets.

First step is to identify the healthy vulvar microbiome composition. Second step is profiling the vulvar microbiome in disease. In the literature, there are only a handful of articles that describe the healthy vulval microbiome, and a weakness is the lack of diversity in subjects, and no shotgun metagenomics data. The vulva is an unique niche with a different microbiome composition in different locations. It is a collection of vaginal, cutaneous and intestinal species.

A concern with vulval microbiome studies is whether there is sufficient biomass yield for metagenomic sequencing. A feasibility study was conducted and most samples were sufficient for shotgun analysis, and a fungal fraction was identified, with results similar to in the literature.

In an observational study at the Center for Human Drug Research in Leiden, the aim was to identify ideally non-invasive biomarkers to differentiate healthy vulvar tissue from that affected by disease (lichen sclerosis, HSIL or vulval cancer). The aim is to combine many facets of vulval health to identify key factors, including imaging, immunohistochemistry, micro-biomics, histology, biophysical and patient-related. Genomics technology has potential here to develop a novel application to improve quality of life for women affected by vulval disease.

Get in touch