Diet, Childhood Nutrition and the Microbiome – Kathryn Dewey

Diet, Childhood Nutrition and the Microbiome – Kathryn Dewey


Kathryn Dewey:
Thank you very much. I’d like to thank the organizers for inviting me, and, actually,
I was a bit surprised and honored to be invited because I am a newbie in this area. I was
introduced to the microbiome world a couple of years ago when I became involved in a research
consortium led by Jeff Gordon at Washington University in St. Louis called the Breast
Milk Gut Microbiome and Immunity Project. And my part in that consortium is really because
of work that I’ve been involved in Africa with child malnutrition. And so I’ve had to
kind of do a crash course in the microbiome field to prepare for this presentation, so
I hope that my colleagues in the audience who know more about some of this than I do
will chip in during the discussion. Now I, myself, work at the far end of the
translational spectrum when it comes to my research. So I’m involved in human population
research in communities, and I hope that the discussion today and tomorrow can lead us
even further in the direction of linking the very exciting work that we’ve seen presented
today at the basic level to its translation into application in the field, and for many
different types of populations; we’ve just heard about the elderly, and I’m going to
talk about the other end of the age spectrum. Just one last thing to say about my background:
I was originally trained as an ecologist, and so for me, it’s really delightful to see
the human body now viewed as an ecosystem. And I only wish that more M.D.s would see
it that way. So when thinking about what to review for
this presentation, I think that the key questions in my mind were, how does diet in early life
influence the microbiome? How does the microbiome influence child nutritional status? And how
do differences in microbiomes structure and function affect nutritional outcomes during
the first two to three years of life, as well as long-term health and developmental outcomes?
And what I was able to find was really very scant and really almost nothing on the third
question listed there. So I’m going to start first with something
that you all know, we’ve heard a couple time previously, that the beginning stage for the
infant is, theoretically at least, a sterile environment in the infant intestine, and that
various factors then influence the individual’s gut macrobiotic composition thereafter, including
gestational age, delivery mode, maternity ward or neonatal clinic, feeding mode, others
foods and fluids, antibiotic use, paternal skin and the skin of other caregivers, and
the environment. And this study shown in this slide has also been referred to before, but
I want to mention it again for one specific reason. This was the cross-cultural study done by
Yatsunenko and colleagues published last year in which they characterize bacterial species
present in fecal samples from three very different populations around the world. And they were
able to show that the differences between the younger profiles and the adults narrowed
with age. So you can see that in all three of the populations the children were quite,
quite different from the adults, and then achieved, more or less, the adult type by
about three years of age. So the first three years of life are a period of very rapid transformation
of the microbiota, and an important period in life for us to focus on. In that same study, they were able to show
that this fundamental sort of change that occurs in the first couple of lifes — couple
of years of life — is in the diversity in the microbiota. And if you look at that blown
up over here, in all three of the populations, that was the trend between zero and three
years of age. Now, of course, the very first diet we hope
the infant will receive is breast milk. And this plays a fundamental role in influencing
the infant’s microbiome, and I’m going to talk about three aspects to that. First of
all, there are microbes in breast milk that have been hypothesized to seed the infant’s
GI tract. There are many, many constituents in human milk that are considered pre-biotic
that are very important at promoting the growth of certain types of bacteria. And there are
quite dramatic differences in the microbiome between breastfed and formula-fed infants. So starting with the first one, I’m just going
to show you this one study by Cabrera-Rubio et al in which they characterize the milk
microbial community for 18 lactating women within two days of childbirth, and then again
at one and six months post-partum. Now they found several hundred species of bacteria,
and — or that they suspect that there are several hundred species, and that these are
compositionally distinct from other human niches. And they conclude that they are not
simply contaminants from the skin. They found that colostrum has a higher microbial diversity
than transitional or mature milk. And in the colostrum samples, the dominant types of bacteria
are shown here. And that by one and six months, the typical inhabitants of the oral cavity
became more prevalent in the milk, and they speculated that this was due to bacteria from
the infant’s mouth colonizing the milk ducts and the areola. Interestingly, they found
that milk from obese mothers tended to contain a different and less diverse bacterial community
compared to milk from normal weight mothers. And they also found that milk samples from
mothers who went through elective, but not non-elective, C-section, contained a different
bacterial than the milk samples from mothers with the vaginal delivery. Because it was
only seen in the mothers with the elective C-section, they concluded that it was the
process of at least going through some labor, the physiological stress of that, that linked
the mothers with vaginal delivery and non-elective C-sections in having some commonality in their
microflora in the milk. Now I’m not going to spend a lot of time on
the prebiotics in human milk because you’re going to hear a wonderful lecture tomorrow
by my colleague David Mills. And I am showing just one slide from one of their papers recently
published characterizing the — some of the prebiotics in human milk, particularly the
human milk oligosaccharides and glycoconjugates that are quite diverse, many, many, many different
types of these that are very abundant in human milk and pretty unique to human milk, and
are known to be utilized especially by bifidobacteria. So this is a structural example of the human
milk oligosaccharide. These are some of the other glycans of different types, and over
here, a glycolipid. And David will explain to you how all of those are operating in terms
of their impact of the microflora. So as a result of some of these differences
in the experience between breastfed and non-breastfed infants, there are differences in the microbiome
by feeding mode. But the results from different studies are actually somewhat mixed. Just
showing two examples here, this is a study of a fairly large sample of more than 600
from European countries, which found that bifidobacteria dominated the microbiota of
the breastfed infants, whereas the formula-fed infants had higher proportions of bacteroides
and members of the clostridium coccoides and lactobacillus groups. In this study, which
is much smaller, they found that formula-fed infants had increased richness of species
with over-representation of clostridium difficile, but no difference in bifidobacteria compared
to breastfed infants. So there has been some mixture in findings,
and there was a review conducted a few years back that contrasted the studies done before
1980 and after 1980 with the purpose of making the point that methodological differences,
as well as differences perhaps in the definition of breastfeeding and how subjects were categorized,
may have led to differences in some of the findings. So, for example, here in the bifidobacteria
line, you can see before 1980, five out of six studies found those were increased in
breastfed infants, but after 1980, that was not as commonly demonstrated. On the other
hand, there are other types over here that are lower in the breastfed infants, and that’s
fairly consistently found including bacteroides and clostridia. So I think that we don’t fully
understand all these differences, but certainly there are differences in microbial composition
in the gut by feeding mode. Now to straighten out some of these — some
of this confusion, I thought it would be helpful to look at one study from an animal model,
in this case a primate, which is very close to humans. This was just recently published,
and in this case they looked at Rhesus monkey infants who were randomized to receive breast
milk from their mothers or formula exclusively from birth to three months of age. It’s a
small sample. When you’re working with monkeys, that’s the way it is. But the value of this
particular model is that nutritional needs are actually quite similar to those of human
infants. And they found that the formula-fed monkeys had more rapid growth and higher serum
insulin. That’s also true in human infants who are formula fed. And they had higher levels
of bacteria from the ruminococcus genus and lower levels from the lactobacillus genus.
And along with these differences they found elevated levels of many cytokines, chemokines,
and growth factors at four weeks of age. So in addition to characterizing the microbiome
in this study, they also did metabolomics work, and you can see here, I’m not going
to go into the details, but that there’s quite dramatic separation in both serum and urine
in the profiles that they were able to find. So this difference between receiving breast
milk or receiving formula has fundamental effects on the child’s metabolism. The next phase in terms of what the infant
would experience would be the introduction of solid foods, and this also plays a role
in the changes that occur in the microbiome. This was a study, I mentioned before, from
Europe in which they assess the fecal microbiota composition of infants from five different
countries. They sampled them at six weeks of age before any solid foods were introduced,
and again four weeks the introduction of solid foods. And in this particular set of infants,
59 percent were fully breastfed, 27 percent fully formula fed, and 14 percent were mixed
fed at baseline. You can see in this slide that after solid foods were introduced, there
was a reduction in bifidobacteria, as well as in some of the clostridia species, and
in the enterobacteria here. And there was an increase in a couple of other species of
clostridia. So that shift makes a difference. What we
don’t know is what happens is when you give different types, or different composition
of solids foods, and that’s been uninvestigated to the best of my knowledge. The other study I wanted to mention along
these lines is one from De Filippo et al, and this was mentioned in an earlier presentation
today. And in this case, they contrasted the gut microbiota of children in Burkina Faso
in West Africa and children in Europe. And they found that the composition of the microflora
diverged after weaning. Both populations, breastfeeding was fairly common in very early
life, and when they starting receiving other foods, they received very different types
of diets. So in the Burkinabe diet, this was low in fat and animal protein, rich in starch,
fiber, and plant polysaccharides, and predominantly vegetarian. The European diet being higher
in fat, animal protein, sugar, and starch, and lower in fiber. Now I wanted to mention
this is an agricultural population here, so I would not consider this to be any kind of
model diet, certainly not an ancestral diet, if you go back to more of the evolutionary
past for humans. It’s a cereal-based diet, and nutritionally inadequate in many different
ways. So I don’t want to give you the impression that this is a better diet than the European
one in any way. As I said, the differences in the microbiota
became evident after the period of predominant breast feeding, and at that point, they found
higher microbial richness and biodiversity in the Burkina Faso samples. They found a
greater representation of actinobacteria and bacteroidetes, and, in contrast, in European
children, a greater abundance of firmicutes and proteobacteria. And this shows, in color,
the very dramatic difference in the microbiota in these two populations. Over here, very
much dominated by prevotella and other bacteroidetes, and over here, you can see a lot more representation
of the ones in this area here. Now there were some other correlates to these
differences in the microbiota. In the Burkinabe children, they had greater total short-chain
fatty acids in fecal samples, and we just heard about the importance of short-chain
fatty acids that are produced during fermentation by intestinal microbiota based on plant polysaccharides.
These substances are very important precursors for many things that are — potentially have
health outcomes, and have theorized to prevent established potentially pathogenic intestinal
microbes, and have a role against gut inflammation. In this particular situation I’m not sure
we could argue that the Burkinabe children were protected against some of these pathogenic
microbes, because diarrhea rates and other illnesses are very, very common there. I want to move on to think about specific
nutrients because a lot of times, when people think about nutrition and the microbiota,
they’re thinking in very general terms about total protein, total fat, carbohydrates, maybe
fiber. But they’re really not thinking very often about micronutrients. And in nutrition
we think about the huge number of different nutrients that we think about every day. There’s
probably 40 or more nutrients that we might consider. There’s very, very little in the
literature on the specific nutrients and how that might interact with microbiota. I’m actually
just going to mention two that I could find studies related to, and one is iron, and the
other is fatty acids. In the case of iron, it’s an important one
to consider because it’s an essential nutrient for many gut microbes, but some of the beneficial
barrier bacteria, in particular lactobacilli, do not require iron. So when you introduced
iron into the diet in large amounts, you can shift the profile of bacteria in the gut.
And in particular, for some enteric gram-negative bacteria, and some of these are pathogens,
iron acquisition plays an essential role in the virulence and colonization of the pathogenic
strains. With regard to fatty acids, there’s just very little information, but — suggesting
that the omega-3 long-chain polyunsaturated fatty acids may modulate growth and adhesion
of lactobacilli, and may impair the growth of certain species of Bacteroides. So I wanted to show you one important study
that was published recently by Nancy Krebs and her group in Colorado looking at effects
of different complementary feeding regimens on iron status and enteric microbiota in breastfed
infants. In this case it was a randomized trial with four to five — 45 exclusively-breastfed
five-month-old infants randomized to one of three feeding groups, either commercially
available pureed meats, iron- and zinc-fortified infant cereals, or iron-only fortified infant
cereals. And they followed them until nine to 10 months of age. And in the last 14 infant
in the study, they were able to get samples to characterize the enteric microbiome. And
they got fecal samples monthly between five and nine months. One of the important aspects of this study
is that the cereal groups had much higher iron intake than the meat group, and that’s
shown here in this slide. Right here is the iron- and zinc-fortified cereal, this is the
iron-only fortified cereal, and this is the meat group. And this is dietary iron between
six and nine months of age. You can see a huge jump with both of the fortified cereals,
and a much smaller increase in dietary iron in the meat group. Concomitantly with that,
they found differences in the microflora in this relatively small sample of infants, but
only in the iron-only fortified group. So, in that case, they had a reduction here in
the actinobacteria, and they had increase in the bacteroidetes, a decrease over here
in the lactobacilli, and the only difference that the meat group showed was an increase
in clostridia group, a clade in this particular case. So these are, I would say, probably fairly
preliminary data showing that the type of complementary food, and in particular the
iron content of that food, could actually make quite a large difference in the microflora.
To illustrate that point further, I have one slide from an older age group; this was conducted
in school-age children in Cote d’Ivoire, in Africa, by Michael Zimmermann and his group.
This was the double-blind randomized control trial in which they received either an iron-fortified
or non-fortified biscuit for six months. And those biscuits contained 20 milligrams of
iron per day, given four times per week, as electrolytic iron. When they look at the gut
microflora, they did not find a significant difference in the bifidobacteria. This is
control group at baseline and at six months; iron at baseline and at six months. But they
found an increase in the iron group in enterobacteria, a decrease in lactobacilli, and, along with
those changes in the flora, they found quite a large increase in a marker of gut inflammation,
which is fecal calprotectin. So there are some concerns that providing iron may shift
the flora in such a way that it may be disadvantageous to the child. With that said, maybe I should back up for
just a second and say that we worry about iron a lot because iron deficiency is probably
the most common nutritional disorder in the world. And we estimate probably about 50 percent
of young children are iron-deficient, and similarly, a very large percentage of pregnant
women globally in low-income countries. And so that’s why fortified products are one of
the strategies being used to reduce iron deficiency and iron deficiency anemia. But we need to
be concerned about whether there are potential side-effects from those kinds of strategies. Now, as I mentioned, the other nutrient I
could find some information about were fatty acids, and this is just one study I found
on the effect of fish oil supplements on fecal microbiota from nine to 18 months. This was
conducted in Denmark with 132 healthy infants. And they were looking at molecular fingerprints
of bacterial DNA, and so they were characterizing the outcomes here in terms of sort of general
types. I don’t have genus, but these were — both of these particular categories were
presumed to be bacteroidetes. And they were modified by the breastfeeding status of the
child. So what they found is that in those who were not breast-fed at baseline, at nine
months, there was a greater increase in one of these types in the fish oil group, and
a greater increase in the other type in the sunflower oil group. It’s a little bit hard
to interpret what that means in terms of potential health consequences, but what is does mean
is that the fatty acid composition of the diet can cause shifts in the microflora in
the infant’s gut. Though switching to the second way of looking
at this, how does the microbiome influence the child’s nutritional status, I’m going
to mention a study that has also been mentioned by previous speakers, just published in Science
this year, in which the gut microbiomes of Malawian twin pairs, discordant for a form
of severe malnutrition called kwashiorkor, were compared. And in this study they compared,
or they examined, 317 twin pairs that were followed through the first three years of
life. And in this case, 50 percent of the twin pairs did not develop acute malnutrition.
Forty-three percent became discordant, which means one twin pair — twin of the pair did,
and the other did not. And 7 percent both developed acute malnutrition. Whenever one
or both of the members of the twin pair developed malnutrition, they were both treated with
therapeutic food, peanut-based, ready-to-use therapeutic food. And so in this study they assessed the microbiomes
of nine same gender twin pairs without malnutrition, and thirteen pairs discordant for kwashiorkor.
And in this slide they were able to show that the microbiomes of the twins with kwashiorkor
were less mature compared to their non-malnourished co-twins. So these are the co-twins who remained
non-malnourished. I don’t personally call them healthy, because in that environment
that’s the exception rather than the rule, but they didn’t meet the definition of severe
malnutrition. And you can see there’s an increase over time of the diversity of the microbiome.
And in the healthy co-twin of the pair, where one of them developed kwashiorkor, that same
sort of trend was exhibited, but in the twin with kwashiorkor, there was a transient increase
in maturity when they received the RUTF, but that was not sustained, so they ended up not
really having that same temporal maturation as their co-twin. And as you know, in this study, they were
able to take the fecal samples from these pairs of twins and implant them into gnotobiotic
mice, and I won’t go over again how that method works; I think most of you here are familiar
with it. But I do want to show a slide that’s been shown before just to reiterate that,
in fact, the results in the mice who were recipients of these microbial transplants
exhibited the phenotype of losing weight only when they were on the Malawian diet. And this
slide doesn’t show that when they were on a control mouse chow, they did not exhibit
a phenotypic difference. So there is a diet-microbiome interaction, and I think it’s very important
for all of us here to remember when thinking about consequences. So here is the group that got the transplant
from the twin with kwashiorkor, and then this from the co-twin that did not develop kwashiorkor.
And you can see that these mice lost weight during that period of time. They recovered
that during the period of feeding with RUTF, but then did not really sustain that weight
gain over time. And similar profile was seen in the fecal microbiota, very dramatic reconfiguration
during the period of re-feeding with RUTF, but then became more similar to the co-twin
thereafter. Now in terms of the types of species and taxa
that changed when the mouse received the microbiota from the kwashiorkor twin, there were increases
in certain types of bacteria and decreases in others. And some of the taxa also changed
in the mice that got the healthy co-twin transplant. And I’m not going to go into the details here,
but you can just see visually that this is the period during the Malawian diet, this
is on RUTF — now you see a lot of changes occurring — and this back on the Malawian
diet here. So where do we go from here? This is really
what drew me into the field, is the project that is attempting to bring together various
disciplines to look at all the relationships between the breast milk, gut microbiome, and
immunity using populations in — populations with a lot of malnutrition. And the goals
of this project are to identify and validate new pre- and pro-biotic interventions to improve
the health and development of infants and children in the developing world, and then
demonstrate a process by which new such interventions can be identified and validated in the future. And the point here is that, as several people
have mentioned before, it really requires lots of different disciplinary expertise to
put this all together, to take the information from the trials that have been done in humans,
to then look at what that means in animal models, to identify leads for pro — pre,
pro, and symbiotic. Then go back and test those in animals, and eventually test those
in human populations. So that’s where we are aiming. I was asked to identify gaps, so this is my
last slide. Just wanted to say that apart from the studies comparing the breastfed and
formula-fed infants, there’s very little information on how dietary composition or nutrient intake
affects the microbiome of children. So I would very strongly urge greater funding and research
in that area. And then in terms of the emerging link between malnutrition and the microbiome,
this is something that we need to understand much better in terms of what the causality
and the direction of the causality is, as well as the effect of different types of interventions.
And lastly, we really desperately need prospective studies that include long-term follow-up of
intervention trials to look at consequences later in life of these changes in early childhood. I’d just like to finish by thanking the BMMI
team that introduced me to this world, and to thank all of you for your attention. [applause] Female Speaker:
Thank you, Kathryn, that was excellent. I’m going to take a moderator’s privilege and
ask one quick thing. Relative to the comparing or doing research in formula-fed babies versus
breastfed babies, formulas that are now being enriched with probiotics at an increasing
rate in the United States, and prebiotics as well, and this has been a trend within
the last five years or so. So there needs to be that caution to control for that, and
it’s going to be difficult to compare to historic studies because of that. Kathryn Dewey:
Yeah, thank you for bringing that up. I think part of the reason there’s so much diversity
in the findings of studies doing this is that formulas have changed dramatically over the
last 30 years, not even just the last five or 10 years, with a lot of different compositional
changes, and so it’s very difficult to compare across time, as you say. Male Speaker:
Hi, you may have said this, it went by kind of fast, but for the case of bacteria in mother’s
milk and appearance of those lineages in babies, what’s the overview of studies on that? Reason
I ask is we did a study with grey cell dravondi [spelled phonetically] where we tracked OTUs
in mother’s milk and OTUs in baby poop, and we really didn’t see any linkage between the
two. Babies’ poop was no more similar to their own mother’s milk than somebody else’s mother’s
milk, and vice versa. And I was wondering what the line of studies is on that. Kathryn Dewey:
Well, I think you probably know better than I do. I mean, I don’t believe the study that
I was mentioning tried to look at how the microbes in the mother’s milk were related
to the infant’s microflora. And I think that’s a good point: If it is seeding the gut, is
that really contributing to the growth of bacteria in the infant over time? So I think
it’s an open question, unless somebody else wants to answer that. Male Speaker:
So, with regard to the microbes in mother’s milk, and, yeah, I mean, how is there sort
of a special organ that cultivates the microbes, or is it just sort of distributed throughout
the gland? I mean, it seems to me that this is some sort of a evolutionary — I mean,
some need to have these microbes in there, and yet nobody actually talks about what’s
maintaining them, and how they’re actually kept in the breast, and how they accumulate.
Where do they come from originally? Is it sort of spores that are there for all of life
and suddenly wake up later? I mean, any ideas on that? Kathryn Dewey:
Well, those are great questions. I mean, I can only speculate that there is a portal
to the outside world even before the mammary glands start to lactate. Male Speaker:
[inaudible] Kathryn Dewey:
Right, but once they get inside at least the ducts, then they have this wonderful food
to start flourishing on, which is not present on the skin or in other parts of the body.
So I’m just speculating here. If anybody wants to jump in, please go ahead, but those are
great questions. Female Speaker:
Thank you. Let’s give another round of applause to Kathryn, thank you. [applause] Female Speaker:
And it’s my pleasure to introduce the last speaker for this session, Dr. Johanna Lampe,
from the Fred Hutchinson Cancer Research Center, and she’s going to talk about Gut Microbial
Metabolism of Food Constituents: Modulating Human Dietary Exposures.