The HMO microbiome story begins before the microbiome does anything at all. HMOs first have to survive the host. What is interesting is that they mostly do. Human digestive enzymes are well equipped to process starch, sucrose, lactose and many dietary glycans, but we lack the specific glycosidases needed to cleave the diverse glycosidic linkages found in human milk oligosaccharides. So HMOs are not meaningfully broken down in the mouth, stomach or small intestine1,2. They travel on.
Here is why that matters: a compound cannot act as a selective microbial substrate if the host absorbs it first. The non-digestibility of HMOs is therefore central to HMO prebiotic selectivity. It places these structures in the category of prebiotics, substrates selectively used by host microorganisms to confer a health benefit. In mature human milk, HMOs are typically present at approximately 5 to 15 grams per litre, and much of that reaches the infant colon each day3. That is a serious substrate load, not a trace signal.
Once there, HMOs become available to HMO gut bacteria with the right enzymatic machinery. For broader context on their role in gut and immune development, see our HMO gut health guide. The basic principle is simple, but powerful: HMOs are selective because the host leaves them for microbes that can use them.
What is remarkable about the relationship between HMOs and the microbiome is not simply that HMOs are non-digestible, but that they are selectively utilisable. We have known for some time that bifidobacteria dominate the infant gut when HMOs are present at high concentrations, yet the mechanistic basis for this pattern is more nuanced than a simple growth advantage. It turns out to be a question of enzymatic capability, transport systems, and ecological competition, all acting together.
Here is the key bit: HMOs are structurally diverse glycans, and many gut bacteria cannot efficiently access or metabolise them in intact form. Bifidobacteria adapted to HMOs can, and so can a limited set of other gut bacteria, which usually access only some HMO structures. That asymmetry helps explain the selective prebiotic effect seen most clearly in the infant gut4.
The most studied example is Bifidobacterium longum subsp. infantis. The species’ HMO interaction remains one of the clearest examples of glycan-driven microbial selectivity. This organism encodes a highly specialised set of glycosidases, including fucosidases, sialidases, and beta-galactosidases5.6. Together, these enzymes act on terminal HMO moieties, progressively dismantling complex oligosaccharides into utilisable monosaccharides.
What this means is straightforward in biochemical terms. The organism can unlock energy from substrates that remain inaccessible to most other members of the gut microbiota.
The genes encoding this machinery are organised in what is often referred to as an HMO gene cluster. Sela et al. identified a specialised HMO utilisation cluster in B. longum subsp. infantis, including genes for HMO transport and glycoside hydrolysis4. Later comparative work supports the view that this machinery is most complete in infant-associated bifidobacteria5. Some other genera carry partial toolkits, often borrowed from pathways for breaking down mucus. Many organisms, including E. coli and numerous Salmonella species, may be metabolically versatile, but they lack the full HMO utilisation toolkit needed to access these substrates efficiently7,8.
The advantage of bifidobacteria is encoded in their genome. HMOs are not a private resource, but bifidobacteria are among the best equipped to use them. Without the right enzymatic toolkit, HMOs remain unused, reinforcing a strong selective advantage for species that carry these pathways.
It is worth noting that not all Bifidobacterium behave the same way. B. longum subsp. infantis is the most specialised HMO utiliser, often considered the gold standard for HMO adaptation. Bifidobacterium bifidum and Bifidobacterium breve also metabolise HMOs, but with a more limited or strain-dependent capacity. B. bifidum breaks HMOs down outside the cell and shares the released sugars with other bifidobacteria, whereas B. breve imports selected structures such as LNT and LNnT, and its use of fucosylated and sialylated HMOs varies strongly between strains9,10,11.
Here is the key distinction: B. infantis imports intact HMOs into the cell before breaking them down. This intracellular strategy is energetically efficient and reduces exposure to competing microbes. B. bifidum, by contracts, degrades HMOs externally, which is less efficient but still ecologically beneficial in mixed communities.
In practice, differences will also occur at the strain level9,12. As such, to be sure of the strain’s metabolic capabilities, always check the strain’s genetics at the DNA level.
Different HMO structures appear to favour different bifidobacterial species and strains. Fucosylated and sialylated HMOs are utilised differently depending on each organism’s transport systems and glycosidase repertoire. A mixed HMO profile may therefore give several bifidobacterial species something to work with, rather than pushing the system towards one dominant responder. This is supported by studies showing strain and species level differences in HMO consumption across Bifidobacterium, including differential use of fucosylated and sialylated structures9,12.
Bifidobacterium is the most studied HMO consumer, but not the only one. Bacteroides thetaiotaomicron and Bacteroides fragilis can grow on HMOs as their sole carbon source, by redeploying pathways they normally use to break down intestinal mucus13. Akkermansia muciniphila degrades 2’FL and LNT in vitro and releases simpler sugars that other bacteria can use14, although this ability differs between strains15. Butyrate producers such as Roseburia hominis and Eubacterium ramulus can use LNT and related HMO building blocks directly16, and some Ruminococcus gnavus strains have been reported to grow on fucosylated HMOs17. Most of this evidence comes from pure cultures and in vitro models. How much these organisms contribute in the human gut depends on age, diet, and the existing microbiome5. The resulting picture is a network rather than a single winner: specialist bifidobacteria, generalist glycan degraders and cross-feeders, each accessing different HMO structures or their breakdown products.
Once Bifidobacterium gain access to HMOs, they do not simply grow in isolation. They reshape the ecosystem. Rapid consumption of HMOs reduces substrate availability for opportunistic organisms, including potential pathogens that lack the enzymatic machinery for utilisation.
What follows is a shift in metabolic output. As Bifidobacterium produces more lactate and acetate, the lumen becomes more acidic. That shift in pH makes it harder for many opportunistic organisms to expand18, especially those less tolerant of acidic conditions. The ecology is fairly simple: the bacteria that can use HMOs early help make the space less favourable for competitors.
Walsh et al. showed that HMO-primed bifidobacteria had enhanced adhesion to intestinal epithelial cells in vitro, suggesting a potential colonisation advantage that should still be interpreted cautiously19. To put it simply, HMOs do not just feed beneficial bacteria. They help shift the local ecology in favour of HMO-adapted microbes. This is the essence of HMO prebiotic selectivity, and a central reason why the HMO microbiome axis continues to attract sustained scientific interest.
Once HMOs reach the colon and are taken up by specialist HMO gut bacteria, the story shifts from substrate access to metabolic output. So what happens next? The answer is partly about short-chain fatty acids, partly about pH, and partly about how microbial fermentation changes the ecological tone of the gut.
Fermentation of HMOs by B. longum subsp. infantis produces acetate as the dominant end product, with formate and lactate also generated through bifid shunt metabolism20. This point is worth being precise about. B. infantis is not a major butyrate producer. It produces acetate, and that acetate can then be used by other organisms.
Here is where it gets interesting: acetate can cross-feed butyrate-producing taxa, including species such as Roseburia 21. In that setting, HMO short-chain fatty acids are not simply primary fermentation products; they become part of a wider metabolic network. Butyrate, when produced downstream, acts as a key energy source for colonocytes, supports tight junction integrity, and appears to reduce inflammatory signalling within the mucosa22.
The nuance matters. The HMO microbiome effect is not that HMOs directly drive B. infantis to produce butyrate. Rather, HMOs enrich organisms that produce acetate, and acetate may support secondary butyrate production in mixed communities. That cross-feeding effect is biologically plausible and increasingly supported, although still more context dependent than acetate production itself 21.
Metabolism changes habitat. Once HMOs are fermented, the downstream effects extend beyond bacterial growth itself. SCFA production, mucin signalling and reduced proteolytic fermentation all contribute to a gut environment that is more oriented towards saccharolytic metabolism and barrier support.
What is fascinating is how these effects extend beyond pathogen suppression. SCFA signalling is associated with increased mucin production, which can thicken the mucus layer and improve barrier integrity23. There is also evidence that HMO fermentation can reduce branched-chain fatty acid formation, a marker of proteolytic fermentation, including in faecal microbiota from children with Crohn’s disease, although FOS had a similar effect in that study24. In practical terms, HMOs appear to shift microbial metabolism away from protein fermentation and towards saccharolytic fermentation. In an ex vivo study with microbiota from children and adults, HMOs also increased aromatic lactic acids, microbial metabolites linked to immune regulation25.
It is worth noting that the adult gut is not simply an older version of the infant gut. The ecological starting point is different. Infants, particularly during early milk feeding, often have a microbiome in which Bifidobacterium can become highly abundant. Adults usually begin from a lower Bifidobacterium baseline, with a more complex and competitive microbial community already in place. The HMO gut microbiome adults question is therefore less about whether HMOs work, and more about how strongly they shift an already complex microbial ecosystem.
That matters for interpretation. The evidence suggests that HMOs still act as selective prebiotics in adults, but the effect size is usually more modest than we see in infants. Elison et al. reported in 2016 that healthy adults consuming 2’FL, LNnT, or a 2’FL and LNnT mixture for two weeks showed measurable shifts in gut microbiota composition, with good tolerance at daily intakes up to 20 grams26. This is not a trivial finding. It shows that the HMO microbiome mechanism remains biologically relevant beyond infancy.
What we see, though, is context dependence. Bajic et al. used ex vivo SIFR technology to compare the effects of 2’FL, LNnT, 3’SL, 6’SL and HMO blends on microbiota from 6-year-old children and adults25. They found age-dependent bifidogenic responses, linked to differences in baseline Bifidobacterium species composition. Butyrate production, for example, increased only with adult microbiota25. In other words, HMOs can still feed the bacteria equipped to use them, but the starting microbiome helps determine which species respond. In adults, where bifidobacteria are usually less dominant, other HMO users may play a relatively larger role. One study reported that sialyllactose increased short-chain fatty acid production in adults through changes in the faecal microbiome that were not driven by bifidobacteria27.
The practical implication is fairly clear. HMOs remain selective substrates for gut bacteria in adults, and they help illustrate how HMOs feed beneficial bacteria with unusual precision. The magnitude of Bifidobacterium enrichment is likely smaller than in infants, but that does not diminish their value. It reflects the ecology we are working with. For broader context on HMO mechanisms across gut and immune health, see our full HMO gut health and immunity guide.
So how do HMOs stack up against better known prebiotics such as GOS, FOS and inulin? Here is the short version: all can increase carbohydrate fermentation, but they do not apply the same ecological pressure. Conventional prebiotics tend to be shared substrates. Many organisms can use them, directly or through cross feeding, which makes their microbiome effects useful but relatively broad.
HMOs are different because access is restricted. The selective prebiotic effect HMO researchers describe comes from the need for specific transporters and glycosidases, particularly in Bifidobacterium species adapted to milk oligosaccharides. Ex vivo comparisons suggest that HMOs can be more selectively bifidogenic than FOS or inulin24.25. It is not simply a question of feeding bacteria. It is a question of which bacteria have the molecular equipment to enter the system.
For formulators, the HMO vs GOS microbiome distinction is practical. GOS and inulin can still be valuable, especially where broader fermentation is desired. HMOs are better viewed as precision substrates, especially when the goal is to enrich Bifidobacterium with a tighter ecological signal.
Prebiotic | Selectivity | Main bacteria enriched |
HMOs, for example 2’FL, LNnT | More restricted, targets Bifidobacterium via specific enzymatic toolkit | Bifidobacterium (notably B. infantis, B. bifidum, B. breve, B. longum), plus some Bacteroides, Akkermansia and butyrate producers, depending on structure |
GOS | Moderate. Broadly fermented by multiple species | Bifidobacterium, Lactobacilli, mixed |
FOS and inulin | Moderate. Broadly fermented; bifidogenic, also used by many other taxa. | Bifidobacterium, and many other taxa. |
Lactulose | Broader access, very broad fermentation substrate | Mixed flora, non specific |
Explore HMO microbiome applications
For background on the broader evidence base, see our complete HMO guide and our HMO gut health and immunity guide. For researchers and formulators interested in exploring HMO applications, Inbiose offers human-identical HMO ingredients backed by rigorous science. Contact our team.
In short, HMOs are selective because only certain gut bacteria carry the right enzymatic and transport machinery to use them efficiently. HMO Bifidobacterium species can recognise, import, and cleave complex HMO structures that many competing organisms cannot access. That is the core of how HMOs feed beneficial bacteria: they create a substrate advantage for microbes with the correct glycosidases.
The quick answer is B. longum subsp. infantis. It appears to be the most specialised HMO utiliser, particularly because it can import intact HMOs before digesting them intracellularly. B. bifidum and B. breve also use HMOs, although generally with broader or less complete capacity. So the HMO microbiome effect is strongly (sub)species specific, even within Bifidobacterium.
No. Bifidobacteria, particularly B. infantis, are the most efficient HMO users, but other gut bacteria can use specific structures as well (such as certain Bacteroides species, Akkermansia muciniphila or some butyrate-producing bacteria). Their capacity is usually narrower and differs between strains, and most of the evidence comes from laboratory studies. This is one reason why the effect of an HMO depends on the microbiome it meets.
Not exactly. Here is what we know: adults usually start with lower Bifidobacterium abundance and a more diverse, competitive microbiome than infants. HMOs can still shift the adult gut microbiome, including measurable Bifidobacterium enrichment, but the magnitude is typically smaller. That does not mean the effect is weak; it means the ecological baseline is different.
HMO short-chain fatty acids are mainly generated through bifidobacterial fermentation. B. infantis primarily produces acetate, along with lactate and formate. It is worth being precise here: B. infantis does not produce much butyrate directly. But acetate and lactate can support cross feeding by butyrate producing bacteria, so butyrate may increase as a secondary community effect.
The timing depends on dose, baseline microbiome composition, and which HMO is used. In adult studies, measurable microbiome shifts have been observed within about two weeks, particularly with supplementation of 2’FL, LNnT or a combination of both. In infants, responses can be more pronounced because Bifidobacterium often has a stronger ecological foothold. The HMO microbiome response is real, but it is not identical across populations.
Generally, yes, although HMOs are not used by Bifidobacterium alone. The HMO vs GOS microbiome comparison is really a comparison of access. GOS and FOS are useful prebiotics, but they are fermented by a wider range of bacteria. HMOs are more selective because their utilisation requires specialised HMO related glycosidases and transporters. That gives certain Bifidobacterium species a clearer competitive advantage.
Likely, yes. Different Bifidobacterium species vary in which HMO structures they can use, so mixtures containing several HMOs may support broader bifidobacterial diversity than a single structure alone. This is one reason blended HMO approaches are scientifically interesting. The aim is still selectivity, but with a wider range of beneficial Bifidobacterium responders. Mixtures may also support HMO users beyond Bifidobacterium, since different genera target different structures.
While Inbiose strives to ensure the accuracy and timeliness of the information presented in this document, the company makes no guarantees or assurances regarding the precision, reliability, or completeness of the content. This information is intended solely for scientific and technical purposes in business-to-business communications. Local regulations and requirements should be reviewed when considering product labeling or advertising to end consumers. The information provided does not constitute medical or scientific advice, diagnosis, or treatment, and is offered without any warranties, express or implied. Inbiose shall not be held liable for any damages resulting from the use or interpretation of the material presented herein. The reader assumes full responsibility for their use of the information. No liability is accepted for any infringement of patents resulting from the use of the information provided. Content may be updated without prior notice. For further information, please contact your local Inbiose representative. All trademarks mentioned are owned by or licensed to Inbiose unless otherwise noted.
