You have a stomach. I have a stomach. It is one of our few universals. Humans, mate, sing, talk, and raise their children in many different ways, but we’ve all got stomachs. The question is why.

Stomachs help to digest food; they get the process rolling, boiling and grinding by coating our food in slime, enzymes and acid. This is the textbook explanation and no one is saying it is wrong, but in one of my treasured meanders through the library, I recently stumbled upon a paper suggesting this explanation is incomplete, perhaps woefully so. Just as important to our survival may be the stomach’s role in separating, sieving one might say, bacteria that are good for our guts from those that are bad. The study I found was led by Dr. Orla-Jensen, a retired professor from the Royal Danish Technical College. Orla-Jensen tested this new idea about the stomach by comparing the gut bacteria of young people, healthy older people and older people suffering from dementia. What Orla-Jensen found is potentially a major piece in the puzzle of the ecology of our bodies.

Image 1. A diagram of the human stomach. The stomach may act as a sieve, allowing only some kinds of microbes through to the small intestines.

Orla-Jensen and colleagues began by positing, or perhaps assuming is the better word, that a key function of the stomach is to kill bad bacteria with acid. The acid, they argue, serves as a sieve. It stops bad bacteria, particularly the most opportunistic of pathogens, but it does not stop all bacteria. It lets those beneficial bacteria that have adaptations for dealing with stomach acid--adaptations honed over many thousands of generations--on down the gastrointestinal road. In their model, if the stomach fails to kill bad bacteria, pathogens dominate the intestines. They do so in place of the beneficial microbes that help our bodies to digest food and produce nutrients. And when they do... death or at least the failure to thrive is nearly inevitable.

Orla-Jensen and colleagues knew from earlier work that the pH of the human stomach increases with age; the stomach becomes less acidic. This effect is most acute in individuals over seventy years of age. In these individuals Orla-Jensen predicted that the stomach's effectiveness as a killer of bad microbes might be compromised. In turn, the intestines, recipients of everything that leaves the stomach, living or dead, might become dominated by pathogenic species such as the weedy and deadly Clostridium dificile or by oral species, that while beneficial in the mouth can become a pathogen in the gut. It was a simple enough prediction, but perhaps too simple. The biota of the gut is complex. It can contain thousands of species and is influenced by many, many factors which have proven in many ways intractable. Could the stomach’s pH really matter enough to make a measurable difference? As I read Orla-Jensen’s paper, I was skeptical, but I was curious enough to read through to the results. I sat down on the floor in the library and prepared to stay a while.

Image 2. Micrograph of Clostridium dificile. Image courtesy of CDC/ Lois S. Wiggs (PHIL #6260), 2004.

To test their hypothesis, Orla-Jensen and colleagues cultured bacteria they had collected from fecal samples of ninety human participants, one third of whom were between 30 and 40 years old and two thirds of whom were over seventy. They then compared the microbes found in the samples from these different age groups. Again, they would expect that in the older individuals that the bad bacteria and oral bacteria should be more common and, in their abundance, displace the good necessary bacteria, such as Bifidobacterium.

Remarkably, the authors’ predictions from the sieve hypotheses held up. I have reproduced and slightly modified their main table below. Nine percent of the individuals over seventy had more than a million cells of the bad news Clostridum bacteria per gram of feces; none of the thirty to forty-year-olds did. What was more, a third of the individuals over seventy had more than a billion cells per gram of feces of the oral bacteria, Streptococcus salivarius. Again, none of the thirty to forty-year-olds did. But were these pathogenic and oral bacteria doing well enough to actually compromise the success of good bacteria in the gut? Yes. While all of the thirty to forty year olds had at least a million cells of the good gut bacteria Bifidobacterium per gram of sample, less than half of the individuals over seventy did.

Interestingly, the guts of those individuals over seventy years of age who had dementia were in the worst shape, by far. Nearly each and every one of their guts was dominated by Clostridium and oral bacteria. Other studies seem to lend support to these general findings, albeit from different angles. A study comparing healthy individuals and individuals with low stomach acidity found that those with low stomach acidity were less likely to have Bifidobacterium even though their total density of intestinal bacteria, particularly the pathogens, increased. Another study found that individuals with low stomach acidity tend to be more likely to suffer from diarrhea, as would be expected if their guts were being taken over by pathogens.

The differences seen here as a function of age are much more pronounced than those seen in another study, recently published in the journal Nature. The Nature article compares the gut microbes of more than five hundred individuals of different ages and ethnicities. In the Nature study the authors found little effect of age on gut microbes after the first few years of life (during which there was a large effect as newborns slowly acquired adult microbes). However, the Nature study only considered four individuals over seventy years of age (they also did not specifically look for shifts in beneficial versus problematic species, perhaps they will in the future). Orla-Jensen's work suggests that it is precisely the very old individuals in whom the differences begin to be pronounced. Sometimes it takes the perspective of many studies and time to see the full picture. This is probably where I should point out that the Orla-Jensen study I’m discussing was published in 1948. Interesting ideas can get lost in unread scientific articles; many, perhaps most, are. Orla-Jensen’s paper has only rarely been cited and never in the context of the discussion of the function of the stomach or even in the context of aging and the microbial wilderness of our bodies.

Table 1. Reproduced (with updates) from Orla-Jensen et al., 1948. Sample size for each group = 30 individuals. The author of this paper, Prof. Orla-Jensen was 77 at the time of the publication of this paper in 1948 and so had a personal interest in these results. One wonders if he sampled himself.

Percent of individuals with > than 1 billion cells of each bacteria per gram of feces, or, in parentheses, percent of individuals with > 1 million cells per gram.
VolunteersMutualist BifidobacteriumPathogen ClostridiumOral bacteria, Streptococcus salivarius
Aged 30-40 (Healthy)57 (100)00
> 70 years (Healthy)25 (44)931
> 70 years (w/ Dementia)7 (9)4835

More than sixty-five years later it is now up to us to figure out what other predictions the sieve hypothesis might make <sup>2/<sup>. Perhaps the most obvious prediction is that as one travels the body, from the skin to the mouth to the stomach and on into the intestines, that one should encounter, at each step, diminishing subsets of microbial lineages. Is this true? It seems hard to believe. After all, a huge number of studies have proudly announced the great diversity of microbes in the gut, a terrible diversity. Let’s look.

The best study I know of included samples from mouth and gut, and considered which taxa of microbes were found in the different habitats. The diversity of major lineages drops by half as you go from the mouth to the stomach AND the lineages present in the gut, particularly the colon, are a subset of those in the stomach which are a subset of those in the mouth (see Figure 2). Comparing the results of this studies with those of others suggests the mouth itself also serves as a kind of filter, winnowing the species that land on the skin and lips or in the mouth to the subset that are most beneficial. From this subset, the stomach further cleaves.

If the sieve hypothesis holds, there must be additional predictions. I have not thought this through terribly well, but I think I would probably expect differences in the stomachs of animals eating different foods. Animals that eat foods that are more likely to include pathogens ought to have filters that are more finely tuned to weeding out bad microbes; they ought, I think, to err on the side of killing too many. This does appear to be the case for some vultures. The stomach of the white-backed vulture has a pH of 1! Conversely it seems plausible to predict that animals that eat diets less likely to lead them to pathogens, fruit eaters for example, should be expected to relax the sieve, open it up a little to make sure that many good microbes make it through. I don’t know that it has been tested. There must be more predictions for the differences one expects among species. A broad survey of the evolution of the stomach seems in order.

Image 3. White backed vultures feeding on a wildebeest. These vultures need to very actively fight the pathogens in the dead meat on which they indulge. One way they do so is by having very, very, acidic stomachs. Photo by Magnus Kjaergaard.

Modern living also presents us with another testable prediction about the stomach and its effects on microbes. Bariatric surgery is an ever more common medical intervention in which the size of a patient’s stomach is reduced so as to reduce the amount of food he or she can eat in one sitting. The surgery also has the consequence, however, of increasing the pH in the stomachs of those who have the surgery, making their stomachs less acidic. If the sieve hypothesis is right these individuals ought to have gut bacteria that look more like those of seventy years old than those of thirty year olds. They do. Recently a study has found that good Bifidobacterium species become more rare after bariatric surgery while oral bacteria (in this case Prevotella) and E. coli, which can be a pathogen, become more common. These results seem to be what the sieve hypothesis would predict.

I am sure there are more predictions. I’ll leave you to them. The good news is that if there are more predictions now is a great time to look, to test them. The study of the microbes of our body is now hip, as sexy as a field of study that often involves the word fecal can be (see Image 4 or check out your own sexy fecal bugs at American Gut). New data are published every day. If we can develop good predictions they can be tested. We might finally figure out what the stomach does, or rather the complex mix of its roles, its churning melange of duties. No one denies that the stomach helps to break down proteins, it just might not be its most important job.

Image 4. Microbiologist Jonathan Eisen wearing his microbiome. Image courtesy of Jonathan Eisen.

Meanwhile, there is an interesting coda to this story. In addition to considering the difference between old and young individuals, Orla-Jensen, as you might remember, considered the difference between healthy individuals over seventy and individuals over seventy with dementia. The individuals with dementia had even more pathogens and oral microbes in their guts than did the healthy seventy-year-olds. This is interesting, but what is the cause and what is the effect here? Could a poorly functioning stomach lead to a pathogen heavy microbe community in the gut and could that gut community in turn lead to dementia? Could our minds really fail because our stomachs do? A few recent studies have begun to explore the possibility that dementia might result from infection, but it is WAY too soon to say anything conclusive. One is left to imagine the mechanism behind such a decline. I have some ideas, but I’ll need to think them over some more. Meanwhile, you can offer your hypotheses too, and I'll go back to the library and see what other gems I can find, old studies that are as revolutionary as the new ones you read about in the press, studies that whether right or wrong confirm just how little we know and how slow and circular progress can be.

Footnotes (more to be added)

1- They did not sequence the genes of these microbes—now a common technique—and so their results represent just part of what was going on in the sampled guts, a few kinds of common trees in a diverse forest, and yet it was probably a reasonable measure of those trees.

2- Which, I will confess, I've named here. Orla-Jensen and colleagues thought the idea so obvious as to not even deserve a name.