I have a piece today in the NYT Sunday Review on the microbiome in celiac disease. A lot of interesting material ended up on the cutting room floor. I’ll post some below. And I’ll try to preemptively answer questions that, if I were reading the piece for the first time, would occur to me.
When exactly am I supposed to introduce gluten to my baby?
The NYT fact-checker and I had a long discussion about this. The problem is, the American Academy of Pediatrics isn’t completely clear. Citing the following evidence, the academy emphasizes introducing gluten while still breastfeeding.
There is a reduction of 52% in the risk of developing celiac disease in infants who were breastfed at the time of gluten exposure. Overall, there is an association between increased duration of breastfeeding and reduced risk of celiac disease when measured as the presence of celiac antibodies. The critical protective factor appears to be not the timing of the gluten exposure but the overlap of breastfeeding at the time of the initial gluten ingestion. Thus, gluten-containing foods should be introduced while the infant is receiving only breast milk and not infant formula or other bovine milk products.
But it doesn’t say explicitly WHEN to introduce gluten. It does say, however, to begin complementary feeding around six months of age.
The AAP reaffirms its recommendation of exclusive breastfeeding for about 6 months, followed by continued breastfeeding as complementary foods are introduced, with continuation of breastfeeding for 1 year or longer as mutually desired by mother and infant.
I thought that gluten was probably included in the category “complementary foods.” The fact-checker (probably correctly) didn’t agree. So we left out any mention of a defined timeframe. Why would explicit instructions on timing be nice? Because introducing too late may be problematic, as may an overly early introduction. Some scientists talk about an optimal “window” of first exposoure to induce tolerance to gluten.
The AAP’s European counterpart, the ESPGHAN, is much clearer about timing. Introduce small amounts of gluten not before 4 months of age, but not after 7 months, while still breastfeeding. That’s the imagined window in Europe. But who knows if this recommendation will hold as new evidence emerges.
Anneli Ivarsson, the Swedish scientist I quote in the piece with regards to the Swedish celiac epidemic, just had a study published in Pediatrics suggesting that children exposed to gluten at 4 months (while breastfeeding) had a lower risk of celiac compare to those first exposed later. So maybe earlier is better. It’s just not clear yet. But as the AAP says, just make sure to do it while breast feeding.
Also, one study I cite in the article, by Alessio Fasano, introduced gluten to two groups at two different times — 6 months and 12 months — to see if it affected susceptibility to celiac. When he published the paper, it appeared that those who encountered the protein later developed celiac less often. But he tells me that as time went on — he’s at year 3.5 with the cohort — the timing of introduction didn’t really matter. That’s not published, of course, so…
C-sections, microbes and celiac
Children born by cesarean section who are deprived of their mother’s vaginal microflora at birth, have a different colonization pattern of microbes than children born vaginally. A number of studies now find that these children have a higher risk of allergic disease, although one mitigating factor seems to be if the parents have allergies. This study finds that children born by c-section to allergic parents have a greater risk of asthma than non-c-section kids, but the risk is less for c-section children born to non-asthmatic parents.
Other studies have observed a higher risk of type 1 diabetes (this meta-analysis yields an increased risk of 20 percent for these kids) and celiac disease. Causality is hard to prove. It could be some other factor. Maybe complications during pregnancy prompt obstetricians to perform more c-sections, and this primary factor really explains the correlation. That other factor — maternal infection, say — may increase the risk of celiac AND complications during pregnancy that lead to c-section.
This study attempted to control for that possibility by separating women who underwent elective c-sections, and the children still had an elevated risk of celiac.
I’m still not completely convinced, however, that this association results from altered microbial colonization patterns. Yes, c-section children are colonized by different bugs early on. But as this group points out, they’ve also circumvented what’s probably the greatest trauma of most new lives: passage through the birth canal.
The scientists note that children born by c-section have a different methylation pattern of genes in their white blood cells. That is, certain genes remain “off” compared to vaginally born counterparts. Presumably, the severe shock of the birthing process would have turned them on.
This phenomenon, I think, may also explain the correlation of type 1 diabetes, celiac and allergic diseases with c-section birth. The absence of a severely stressful event affects gene expression, and somehow that predisposes to immune-mediated disease.
Here’s the thing, though: these two ideas aren’t mutually exclusive. What colonizes you is partly determined by exposures to microbes (e.g. your mother’s vaginal microbiome), your gene variants, AND how you express the genes you’ve got — how you translate them into living flesh and blood. Think of yourself as an ecological niche. Altering gene expression will almost certainly affect who can move in, and who not. It will impact the bugs that predominate.
And that, in turn, may increase or decrease one’s vulnerability to autoimmune and allergic disease.
Mom’s immune system during pregnancy
Much evidence now suggests that our immune systems start learning while we’re still in the womb, taking cues from Mom. How does this apply to celiac? One study found that elevated markers of pro-inflammatory factors early in pregnancy predicted celiac disease in the children. You’d think this was pure correlation — that there was no causation. Maybe the same genes that increase the risk of celiac tend to inflame during pregnancy, for example.
The problem with that argument is the animal models very strongly suggest otherwise. If you inflame an ewe during pregnancy — the timing of organ development in sheep fetuses more resembles ours than in rodents — it changes the immunology of the fetal gut. It increases permeability, causing what’s sometimes called “leaky gut,” and decreases the cells that help us tolerate foreign proteins AND our own tissues — regulatory t-cells.
Both these types of changes are suspected not only in celiac disease, but inflammatory bowel disease specifically, and other autoimmune disorders generally. These changes will, in turn, affect the microbial community, possibly selecting for a less friendly one.
How genes affect our microbial communities
In the article, I mention two two studies — one by Alessio Fasano, another by Yolanda Sanz — showing that genes long associated with celiac seem to affect the composition of the microbiome in an unhealthy way. A shift in the microbial community, called dysbiosis, has been observed in these diseases for years now. And the predisposing genotypes have for some time been defined as well. Now the two observations seem to be converging.
A working hypothesis is that these genes may induce autoimmunity by increasing the risk of accruing an odd, disease-causing microbiome.
That’s the observation in inflammatory bowel disease regarding those infamous NOD2 variants. Whether carriers have IBD or not, people with these variants tend to have a microbiome that’s edging toward the IBD state. As if — microbiologically speaking — they’re halfway there already. This study suggests a similar pattern in relatives of people with celiac — they harbor a microbiome that’s intermediate between no-risk controls, and people with overt disease. Whether that’s due to shared genes or environment remains to be elucidated.
But, wait — that doesn’t prove that the microbiome causes disease! Well, as I blogged a few weeks ago, scientists have engineered mice to express those NOD2 variants that, in humans, predispose to Crohn’s disease. They find these “humanized” mice to also accrue a weird microbiome, which then prompts colitis and eventually cancer.
And when they transplant that nasty community into perfectly healthy mice without those NOD2 variants, the microbial community alone is enough to cause disease. That’s a compelling demonstration that, although your genes shape your microbiome, the microbes themselves cause the problems.
The evidence presented in the article is pretty slim that microbial shifts precede autoimmunity
That’s true. But science doesn’t occur in a vacuum. And the fact is that Alessio Fasano’s find — two kids showing weird microbial patterns before the onset of autoimmune disease — has many precedents. This study noticed something very similar in kids at-risk for type 1 diabetes. An impoverished microbiome to begin with, and then strange fluctuations preceding disease onset in those who developed disease.
This one observed microbial shifts early in rheumatoid arthritis, an autoimmune condition of the joints. And the pattern has been observed for over a decade in allergic disease, although in that case, the lesson seems to be that loss of microbial diversity precedes disease onset. Kids with a richer collection of microbes less often develop allergies.
The finds of Yolanda Sanz and Alessio Fasano are at odds with one another
On the surface, that seems to be true. Sanz thinks certain bacteroides are acting virulently and contributing to disease in her celiac cases. Fasano thinks that bacteroidetes protect against autoimmunity — in his study, they were relatively depleted in children genetically at-risk for celiac — and he points to a comparative study on Italian children and children in Burkina Faso to make his case. The African microbiome, surmised to protect against inflammation, was enriched in bacteroidetes.
I asked both about this, and they said several interesting things: First, they used different analyzing methods, so hard to compare; second, and more important, the two finds aren’t necessarily mutually exclusive. Sanz may have identified a virulent strain of what’s otherwise a healthy family of bugs. Fasano may be observing the absence of related bugs that are healthy, but not seeing any particular virulent strain.
Sanz also points out that Fasano’s study had no real control — which is true. He references other studies for his “normal and healthy” microbiome. But his study was a “proof-of-concept” — that is, more exploratory than anything else. And I get the feeling that more will come of it with time.
The other major point is that both studies observe relative microbial impoverishment in the diseased state. Loss of diversity is a theme in this science. And that’s probably because the rules that govern outer ecosystems govern our inner one. The more diverse a community, the more stable, and the less prone to wild swings that can lead to collapse and catastrophe.
About historical wheat consumptions in the US
Some argue that the increased rates of wheat consumption are driving the increase in celiac disease. The problem with that argument is that while we have consumed more wheat since the 1970s, actually, in the late 19th century, we consumed far more wheat than today. It’s true that we don’t really know the prevalence of celiac back then. But food for thought nonetheless.
From the USDA: