Ageing and unravelling complex interactions with diet and microbiome

Ageing is the largest risk factor for disease^{(Reference Niccoli and Partridge1)}, and thus, slowing ageing could have a huge impact on health and well-being as well as reducing the burden of an ageing population on health and other public services, while increasing productivity by allowing people to work for longer^{(Reference Scott2)}. It was recently estimated that increasing life expectancy by 1 year would be worth US$38 trillion in cumulative value to individuals^{(Reference Scott, Ellison and Sinclair3)}. Dietary interventions can help maintain health with age, but there are many unknowns about the underlying mechanisms and, therefore, the best diet to follow. It is known that the gut microbiome has several interactions with our diet and human health that could in principle influence ageing^{(Reference Aggarwal, Kitano and Puah4)}, but the complexity of the interactions means that specific mechanisms are difficult to uncover, especially within the timeframe of human ageing^{(Reference Maynard and Weinkove5)}. Experiments in rodent animal models can be very informative but are also time-consuming and expensive and have ethical issues. The nematode Caenorhabditis elegans in which the bacterial diet and the nutrient medium can be precisely controlled provides a model to rapidly test interventions for their ability to slow ageing^{(Reference Maynard and Weinkove5)}. Furthermore, the molecular mechanisms of action can be uncovered using the established genetics and other tools developed over decades of research using this model. This review will discuss how research in C. elegans has revealed an impact of folate synthesis within bacteria on host ageing and how that might have implications for how folic acid supplements interact with the host microbiome and affect health.

The key features of the C. elegans model

C. elegans is a well-established model organism for biomedical research, having revealed some key findings such as the genes for apoptosis, the mechanism of RNA interference and the identity of many genes involved in neurobiology^{(Reference Markaki and Tavernarakis6)}. It has recently been shown that research in C. elegans can be effective in drug discovery^{(Reference Weinkove and Zavagno7,Reference Iyer, Sam and DiPrimio8)} . However, more case studies are needed to demonstrate the material contribution of C. elegans research to the development of approved drugs or dietary interventions. C. elegans is well established in ageing research and benefits from the fact that it has a lifespan of 2–3 weeks and shows signs of ageing within a week, making experiments fast to perform^{(Reference Kenyon9)}. Much of what is known about molecular pathways that influence ageing comes from research using C. elegans ^{(Reference Kenyon9)}. For example, mutants that disrupt the insulin/insulin-like growth factor signalling pathway were first discovered to influence ageing in C. elegans ^{(Reference Kenyon9)}. Disruption of this pathway using genetics has been also shown to increase lifespan in the fruit fly Drosophila melanogaster and in mice^{(Reference Kenyon9)} and is linked to why large dogs with increased Insulin-like growth factor signalling live shorter than small dogs^{(Reference Sándor and Kubinyi10)}. The pathway ends in the regulation of the DAF-16/FOXO transcription factor. Polymorphisms in human FOXO3A are associated with increased human lifespan, further demonstrating the relevance of this pathway to human ageing^{(Reference Revelas, Thalamuthu and Oldmeadow11)}.

In the wild, C. elegans live in rotting vegetation where they feed on multiple strains of bacteria growing within that vegetation^{(Reference Felix and Duveau12)}. In the lab, the worms are fed with a single strain of E. coli, which is grown as a bacterial lawn on a Petri dish with nutrient agar. Using a benign lab strain of E. coli as food is the universal method used to culture C. elegans in the lab. There are no other microbes present as the worms can easily be made to be germ free. This system can be manipulated by changing the constituents of the agar medium, the genetics of the E. coli and the genetics of the worm itself^{(Reference Weinkove13)}. Compounds that impact the bacteria or the animal can be added to the agar. In some ways, the components system represents nutrients (the agar), the microbiome (E. coli or other bacteria) and the host (C. elegans). In the wild, C. elegans extracts its nutrients from vegetation such as an apple by eating the microbes the apple supports when it rots. Humans also use microbes to extract nutrients from apples, but the microbes are contained within the human gastrointestinal tract.

Inhibiting folate synthesis in E. coli extends C. elegans lifespan

In a serendipitous discovery, it was found that a particular strain of E. coli with a spontaneous mutation led to an increased lifespan of C. elegans when the worms were cultured on it^{(Reference Virk, Correia and Dixon14)}. To find the identity of the mutations, researchers used the observation that the mutated strain had a growth defect on minimal media and conducted a screen to find DNA from the wild-type E. coli to find a gene that would restore growth. The mutation was found to be in the gene aroD, which encodes an enzyme in the shikimic acid pathway used to make aromatic compounds, such as aromatic amino acids and a few other metabolites. By adding back aromatic compounds to the medium to test which one influenced lifespan, it was discovered that the addition of para-aminobenzoic acid (PABA), the precursor of folate, prevented the lifespan increase in the mutant E. coli strain^{(Reference Virk, Correia and Dixon14)}. Thus, while the aroD mutation prevented the synthesis of several aromatic compounds, the bacteria could get enough of these compounds from the media to support growth without requiring de novo synthesis. However, the prevention of folate synthesis in the bacterial mutant led to the increased lifespan of the animal. To test this hypothesis further, folate synthesis in the wild-type bacteria was inhibited by the sulphonamide drug, sulfamethoxazole (SMX). Treatment with this drug resulted in a dose-dependent increase in C. elegans lifespan and thus confirmed that inhibition of folate synthesis in the bacteria can slow ageing in the worm as measured both in lifespan and healthspan^{(Reference Virk, Correia and Dixon14–Reference Zavagno, Raimundo and Kirby16)}.

Nutritional requirements for folate

All cells require folate for one-carbon metabolism, a process that makes the building blocks of biosynthesis^{(Reference Fox and Stover17)}. E. coli, like many but not all bacterial species, makes its own folate^{(Reference Magnúsdóttir, Ravcheev and de Crécy-Lagard18)}, whereas animals, including C. elegans and humans, obtain folate from their diet or associated microbes. Under the conditions in which C. elegans is cultured in the lab, inhibiting E. coli folate synthesis to the levels in which lifespan was extended did not prevent growth of the E. coli or C. elegans, showing that sufficient folate remained^{(Reference Virk, Correia and Dixon14)}. Thus, under standard lab conditions, E. coli makes much more folate than they need for growth, and likewise, C. elegans absorbs more folate from the bacteria than it needs for growth and reproduction.

A folate deficiency model in C. elegans

Folate deficiency in humans leads to neural tube defects during embryonic development. In order to test what happens when folate uptake is restricted in C. elegans, use was made of a mutant in a gene that encodes a homologue of the enzyme glutamate decarboxypeptidase GCP2. This enzyme assists folate uptake in the human gut by removing glutamates from polyglutamated folates and thus making the folates more able to be transported into intestinal cells^{(Reference Halsted, Ling and Luthi-Carter19)}. The mutant (gcp-2.1) that lacks this enzyme developed and grew normally, but worms developed abnormally on bacteria made to be low in folate from SMX treatment^{(Reference Virk, Correia and Dixon14)}. This developmental defect was not found when folinic acid or folic acid was added to the media, demonstrating that the developmental defect was caused by a lack of folate, and thus, the combination of the gcp-2.1 mutant and treatment with SMX constituted a folate deficiency model in C. elegans. This deficiency model can be used to titrate folate supplementation because the level of folate supplementation that rescues the developmental defect can be deemed to be at a sufficient level.

Low folate acts in the bacteria, not in the animal, to extend lifespan

A key question was whether the worms live longer because there is less folate in the bacteria or because there is less folate in the animal. Using a dose of folinic acid that prevented folate deficiency in the developmental model, the worms were supplemented at the same time as inhibiting E. coli folate synthesis using the folate synthesis inhibitor SMX^{(Reference Virk, Jia and Maynard15)}. Under these conditions, the worms lived longer, suggesting that lowered folate in the bacteria increased lifespan, not because it reduced dietary folate intake (and thereby lowered folates in the worms) but because it prevented the bacteria accelerating ageing^{(Reference Virk, Jia and Maynard15)}. Bacteria might enhance ageing in a folate-dependent way either by producing specific metabolites or some kind of pathogenic physical interaction with the host. We still do not know how this age-accelerating activity works.

Translation to humans

Looking back through the literature, it had been shown in 1958 that a folate synthesis-inhibiting sulphonamide very similar to SMX increased the lifespan of mice and rats and even dogs^{(Reference Hackmann20)}. Again, as we find in C. elegans, the lifespan extension was reversed by adding PABA to the food, most likely rescuing folate synthesis in the gut microbes in the animals, though that explanation was not considered in the paper. The study was small and needs to be repeated, but it points to the mechanism of slowing ageing being conserved in mammals. In the gut microbiome, some microbes make folate and others do not have the enzymes to do so, relying on folate from other microbes^{(Reference Magnúsdóttir, Ravcheev and de Crécy-Lagard18)}. The proteobacteria, which include E. coli as well as many pathogens, are folate producers, and proteobacteria are associated with conditions like obesity and small intestine bacterial overgrowth (SIBO), in which bacteria colonise the small intestine^{(Reference Miazga, Osiński and Cichy21)}. SIBO is clinically associated with high plasma folate levels, and this folate may derive from the invading bacteria^{(Reference Camilo, Zimmerman and Mason22)}. Perhaps inhibition of bacterial folate synthesis could help treat the condition, not by killing the bacteria but by making them less pathogenic. It is known that disrupting folate synthesis makes many microbes less pathogenic^{(Reference Stocke and Lamont23)}. Our findings in C. elegans have raised a number of possibilities for treating human disease and ageing that need to be investigated by studies in mammalian models and humans.

Folic acid – a breakdown product that feeds bacteria

In the experiments with the folate-deficient C. elegans model, we found that folic acid could not rescue the development defect anywhere near as effectively as folinic acid^{(Reference Virk, Jia and Maynard15,Reference Maynard, Cummins and Green24)} . See Table 1 for the differences between various folates. Folic acid is an oxidised form of folate, which is not found in nature but is more stable than reduced folates, and thus used in supplements and in food fortification. We found that folic acid rescues the folate deficiency, not by directly increasing animal folates but by an indirect pathway whereby a breakdown product of folic acid, PABA-glu, is taken up by E. coli via the AbgT transporter and broken down further to PABA. This PABA is then used to synthesise new bacterial folate, which can be taken up by the worm when it eats the bacteria^{(Reference Maynard, Cummins and Green24)}. Interestingly, the PABA-glu breakdown product was found in all sources of folic acid that we examined, including a consumer supplement in which it was present at 3–4% of the intact folic acid^{(Reference Maynard, Cummins and Green24)}. Thus, when people take folic acid supplements or fortified food, they are also ingesting PABA-glu, which cannot be taken up by humans but can be used by many folate-synthesising bacteria.

Questions still to be answered

We still don’t know how inhibiting folate synthesis prevents E. coli from accelerating ageing in C. elegans. One approach is a genetic screening of single gene mutants of E. coli to see which extends lifespan. The hypothesis is that if a gene is responsible for accelerating ageing in an age-dependent manner, the deletion of that gene would make C. elegans live longer. We screened over 1000 single gene mutants in E. coli and found 9 mutants that did extend lifespan^{(Reference Virk, Jia and Maynard15)}. Two were involved in folate synthesis, and the other seven were in genes with diverse functions. While interesting areas of research, none of them clearly revealed the mechanism by which inhibition of bacterial folate synthesis increased lifespan. Another group screened almost 4000 E. coli mutants for those that increased lifespan and found 29 mutants including aroD and others involved in folate synthesis, but again, none of the mutants clearly revealed a relevant mechanism^{(Reference Han, Sivaramakrishnan and Lin25)}. Further work with omics analysis is required to uncover potential metabolites or peptides that mediate the age-promoting effects of E. coli. Once these have been uncovered in the E. coli: C. elegans system they can be investigated in the human microbiome and tested for correlation with disease.

We do not know what the effect of PABA-glu is on the human microbiome and whether the amounts found in folic acid supplements have a negative effect on health. Clinical studies would allow some of these questions to be investigated.