All this told Bruce and his collaborators two things about breast milk. Firstly, since you can grow a healthy human for many years exclusively on breast milk, it will have everything that a baby needs. Secondly, there is not likely to be anything in it that the baby doesn’t need. From the moment the earliest mammals started producing milk, any mothers that wasted energy on putting unnecessary stuff into it would have quickly been plucked out of the gene pool. The solid components of milk that cost the mother most of the calories by order of amount are 1) fat; 2) sugar; 3) complex chains of sugar molecules called Human Milk Oligosaccharides or HMOs; and 4) protein. Each of these must be of absolutely vital importance to the infant. But here’s the bizarre thing. The third largest solid component of milk, those Human Milk Oligosaccharides, are totally indigestible by a human infant. More than bizarre, it seems absurd. In the words of Bruce German, ‘the mother is expending tremendous amounts of energy to produce these varied and complex molecules and yet they have no apparent nutritional value’.
Human Milk Oligosaccharides are chains of sugar molecules. To put that in context it may be useful to understand a little about different sugar molecules. Monosaccharides are single molecules, usually rings of carbon with a few hydrogen and oxygen atoms added on. Glucose is a familiar example. As a single molecule, it can be absorbed into cells and used for making energy without any breaking down in the gut. Disaccharides are made of two molecules. The white refined sugar in your kitchen is a disaccharide called sucrose, made of a glucose molecule joined to a fructose molecule. The chemical bond that joins the two molecules needs to be broken down by enzymes in your gut before you can use the individual sugar molecules for generating energy. Polysaccharides are long chains of 200–2,000 sugar molecules. They’re often indigestible, like cellulose. Oligosaccharides sit in the middle. The ones in breast milk are branched chains of between 3 and 22 sugar molecules with unhelpful names like di-sialyl-lacto-N-tetraose and lacto-N-fucopentaose V. There are around 200 unique and different types of oligosaccharide in human breast milk, each with different sugar molecules joined together in different chains. Crucially these all need different enzymes to digest them. And humans have none of them. We know that because, in the words of Bruce, if you feed a modern American child human breast milk, ‘the HMOs come out the other end’.
So why is there the same amount of these totally indigestible oligosaccharides as there is protein in human milk? To feed bacteria. In fact, to feed a single bacteria: Bifidobacterium infantis.
BUG FOOD
The idea that the HMOs might be present to feed bacteria rather than humans is an old one. Over a century ago paediatricians, microbiologists and chemists were already trying to understand the health benefits and constituents of breast milk. In the last part of the nineteenth century in Europe and America one in three children died before the age of 5, but it was clear that the chances of survival were higher for breast-fed infants. By 1900 Austrian doctors and scientists had detected differences in the bacteria found in the faeces of breast-fed compared with bottle-fed infants, a remarkable achievement considering the technology of the time. As early as 1888 sugars other than lactose were identified as being in milk, and by 1926 it was reported that there were factors in human milk to promote the growth of a genus of bacteria called Bifidobacterium, but the extraordinary details of the relationship between human mothers and these bacteria took almost another century to determine and required huge advances in genetics and microbiology.
THE HUMAN MICROBIOTA
Estimates for the total number of bacterial cells found in association with the human body have varied between 10 and 1.5 bacteria for each and every human cell. The total number of bacterial genes associated with the human microbiota could exceed the total number of human genes by a factor of 80 to 1. Conservative estimates suggest that an average 70 kg human being is composed of about 30 trillion human cells … and 40 trillion bacterial cells.
The human body provides a rich and varied environment for bacteria with different parts of the body hosting very different communities. In the right location the bacteria perform useful functions, but the concept of ‘good and bad bacteria’ is simplistic. The crucial thing is to have the right bacteria in the right place. Mouth bacteria on a heart valve are bad. Gut bacteria in the urinary tract are bad.
There are a wide range of Bifidobacteria species but they all have a bifurcating shape under a microscope. Aside from that, distinguishing them all is not easy. Starting with the hypothesis that Human Milk Oligosaccharides would nourish them, Bruce German, together with David Mills, a microbiologist, started testing different bacteria, including multiple Bifidobacteria species, to see if they could be cultured in the laboratory using Human Milk Oligosaccharides as their only source of food. Surprisingly initial tests showed very unenthusiastic growth by any of the species tested. They seemed to lack the enzymes necessary to digest the wide variety of sugars in breast milk. But then the team tested B. infantis, a bacteria first isolated from the stool of a breast-fed infant, and it flourished.
If digesting HMOs required a single enzyme, then this ability could be put down to coincidence. Perhaps B. infantis had evolved to digest similar molecules in other environments. But digesting HMOs requires a vast toolkit of genes. B. infantis, and only B. infantis, has them all. An analysis of the genome of the bacteria revealed over 700 unique genes compared to other Bifidobacteria. These include genes for grabbing the HMOs and taking them inside the cell, as well as a series of enzymes able to break down the full range of linkages between the sugar molecules. They are the only bacteria able to completely break down HMOs and there can be no doubt that they have evolved to do this. More importantly the evidence from other species shows a process of co-evolution. As the bacteria evolved to digest milk so the milk evolved to feed them. In the case of humans, the reason why we produce such a complex range of HMOs must be to specifically advantage B. infantis over other bacteria.
So what do we get from the deal? Why do we want a single bug flourishing in our infant gut? The primary reason is probably competition. Human infants are born with a gut that is ostensibly sterile and provides an amazing opportunity for bacteria. It is warm, wet and full of a steady supply of nutrition from food and milk. It is also relatively unprotected by the naive infant immune system. From the moment the mother’s waters break, the baby starts swallowing a range of bacteria. The vagina is full of a carefully controlled mix of bacteria. And what strikes anyone watching a normal vaginal delivery is that the baby is born, face down, into a pile of faeces. It was my job as a medical student to hold a little gauze over the stool, protecting the child and to some extent the dignity of the mother (today, with our advancing knowledge of the microbiome, I wonder if it would have been better not to). Part of the role of breast milk then is to encourage the growth of ‘good’ bugs. This is the most obvious way in which B. infantis protects us: by binding to the cells lining the baby’s large intestine, preventing other, more harmful bacteria doing so. And crucially it seems to bind more strongly when grown on breast milk.
It’s not just a matter of outcompeting the pathogens – B. infantis may keep them at bay with secretions. When grown on HMOs it produces short chain fatty acids (SCFAs). These are molecules that have become famous from their beneficial effects shown in adults eating a high-fibre diet, but they seem to be equally important in children. Some of them may directly kill harmful bacteria. Bruce describes this in terms of the concept of a ‘shelf stable baby’; preserved from the inside by the secretions of the friendly bacteria. Other SCFAs like acetate may feed the developing infant brain. This role in brain development may in part explain the huge complexity of the HMO in human milk compared even to our closest chimpanzee cousins.
AT THE END OF MY CONVERSATION WITH MARK I ASKED IF HE HIMSELF HAD BEEN BREASTFED. ‘NO,’ HE REPLIED. ‘PERHAPS IF I HAD BEEN BREASTFED I’D HAVE BEEN A SURGEON.’
There is a catalogue of other benefits, too. It’s been shown to be anti-inflammatory, specifically in infant gut tissue when compared to adults. Understanding