The Human Milk Glycome as a Defense Against Infectious Diseases: Rationale, Challenges, and Opportunities

Abstract

Each year over 3 million people die from infectious diseases with most of these deaths being poor and young children who live in low- and middle-income countries. Infectious diseases emerge for a multitude of reasons. On the social front, reasons include a breakdown of public health standards, international travel, and immigration (for financial, civil, and social reasons). At the molecular level, the modern rise of infectious diseases is tied to the juxtaposition of drug-resistant pathogens and a lack of new antimicrobials. The consequence is the possibility that humankind will return to the preantibiotic era wherein millions of people will perish from what should be trivial illnesses. Given the stakes, it is imperative that the chemistry community take leadership in delivering new antibiotic leads for clinical development. We believe this can happen through innovation in two areas. First is the development of novel chemical scaffolds to treat infections caused by multidrug-resistant pathogens. The second area, which is not exclusive to the first, is the generation of antibiotics that do not cause collateral damage to the host or the host’s microbiome. Both can be enabled through advances in chemical synthesis. It is with this general philosophy in mind that we hypothesized human milk oligosaccharides (HMOs) could serve as novel chemical scaffolds for antibacterial development. We provide herein a personal account of our laboratory’s progress toward the goal of using HMOs as a defense against infectious diseases.

Antimicrobial resistance is among the most complex and concerning public health challenges facing humankind.^1,2 Warnings that antibiotics are losing effectiveness due to clinical misuse and overuse have been largely ignored. Common illnesses, such as pneumonia, as well as the world’s most prevalent infectious diseases (human immunodeficiency virus (HIV), malaria, and tuberculosis) are becoming increasingly difficult to treat due to drug resistance. Infections caused specifically by antibiotic-resistant bacteria continue to challenge physicians (Table 1). The World Health Organization (WHO) has shown that rates of infection attributable to methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococcus faecium (VRE), and fluoroquinolone-resistant Pseudomonas aeruginosa are increasing.³ In fact, more people in the United States die each year from MRSA infection than from HIV/acquired immunodeficiency syndrome (AIDS) and tuberculosis combined.⁴ Moreover, several antibiotic-resistant Gram-negative pathogens, including the Acinetobacter species, Escherichia coli, Klebsiella species, and multidrug-resistant (MDR) Pseudomonas aeruginosa, have emerged as significant players in human infection.⁵

Table 1.

WHO Priority Pathogens for New Antibiotic R&D⁷

priority level 1: critical		priority level 2: high		priority level 3: medium
pathogen	resistance	pathogen	resistance	pathogen	resistance
Acinetobacter baumannii	carbapenem	Enterococcus faecium	vancomycin	Streptococcus pneumoniae	penicillin
Pseudomonas aeruginosa	carbapenem	Staphylococcus aureus	methicillin, vancomycin	Haemophilus influenzae	ampicillin
Enterobacteriaceaea	carbapenem, cephalosporin	Helicobacter pylori	clarithromycin	Shigella	fluoroquinolone
		Campylobacter	fluoroquinolone
		Salmonella	fluoroquinolone
		Neisseria gonorrhoeae	cephalosporin fluoroquinolone

^aEnterobacteriaceae include: Enterobacter spp., Escherichia coli, Klebsiella pneumonia, Morganella spp., Proteus spp., Providencia spp., and Serratia spp.

Before initiating a discussion on what we view to be a frontier strategy in the treatment of infectious diseases, we believe it is beneficial to briefly inform the reader on the inspiration and risk parameters associated with the development of antimicrobial agents. To this end, we digress momentarily to recount various teachings from the fields of infectious disease and human milk glycobiology, both of which are relevant to our program mission.

BACKGROUND

As their taglines often read, pharmaceutical companies are the greatest source of therapeutics in the world. Perhaps paradoxically, however, most have limited investments in the development of new antibiotics.⁶ Financially speaking, the return on investment for antibiotic development is poor as bacteria quickly develop resistance to small molecules. This means companies have a limited window to recoup the investment. A second issue is that new targets are often species, and even strain, specific. In the clinic, new antimicrobials are expected to feature broad-spectrum activity. This is unfortunate as highly effective antibiotics like fidaxomicin, a narrow spectrum macrolide antibiotic for Clostridium difficile that does not affect the host microbiome, are not competitive in a market which features broadly functioning warheads like vancomycin. Broad-spectrum antimicrobials, however, predictably cause collateral damage to the host microbiome.

Refreshingly, there are exceptions to this trend. Kaleido Biosciences is a new biotech company with the primary goal of developing orally available compounds that modulate the human microbiome. Excitingly, Kaleido has an interest in studying complex oligosaccharides and glycoconjugates in human gut microbiome assays that are designed to reflect healthy and diseased human bowels.

At this stage, it is important to emphasize that, when it comes to clinical impact, quality is better than quantity. Going forward, it is imperative that new antimicrobials provide advances in treatment when compared to available therapies. Ideally, new compounds would function through new modes of action.^8–13 Upon analysis of the recent arsenal of drugs in the latest stages of development, it is evident that they do not advance our ability to treat infections of resistant pathogens, specifically Gram-negative species. Unfortunately, due to limited options, clinicians are repurposing once abandoned drugs.¹⁴ In our search for new scaffolds that may have new modes of action, we turned to human milk.

HUMAN MILK OLIGOSACCHARIDES

While breastfeeding has long had its critics, it is now widely accepted that nursing has a profoundly positive effect on the short- and long-term health of neonates.¹⁵ Under even the harshest of scenarios and even when the mother’s own nutrition is compromised, human milk provides all vitamins, nutrients, and macromolecules that are essential to the development of the child. Interestingly, the composition of macromolecules in human milk is dynamic and adapts itself to complement the neonate defense system.^16–18

The host defense mechanisms of a newborn can be classified as nonspecific (innate) or specific (acquired).^19–21 Nonspecific mechanisms are effective without prior exposure to a microorganism or its antigens. On the basis of the data, it appears as though human milk oligosaccharides (HMOs) are part of the nonspecific response.^22–28 Historically, HMOs have been shown to inhibit the growth of a number of viral and bacterial pathogens.^24–27,29 In the next two sections, we will provide brief overviews of the roles of HMOs as prebiotics and antimicrobial agents as both would be of primary interest to this community based on the WHO’s priority listing of bacterial pathogens.

IMPACT OF THE HUMAN MILK GLYCOBIOME ON SYMBIOTIC BACTERIA

Shortly after parturition, a newborn’s microbiome begins to develop as a succession of bacteria colonizes the neonatal gut. Additional factors, such as mode of delivery (cesarean vs vaginal birth) and antibiotic use, influence the gut flora.^30,31 While little is known about the mechanisms that connect the variables described here to microbiota composition, it is well established that human milk and HMOs are a primary driver of healthy microbiome maturation.

Human milk serves as a primary source of continued microbial inoculation as it can contain ca. 700 different species of bacteria.³² Moreover, because only ca. 1% of HMOs are absorbed into circulation, the majority reach the distal intestine where they can be metabolized by mutualistic symbiotic bacteria. Aerobic and facultative anaerobic species are favored early in life when the gut is oxygen replete. Anaerobic bacteria, such as Bifidobacteria and Bacteroides, are established as oxygen levels decrease. Over time, the gut microbiome achieves an adult-like composition.

HMOs are believed to be a specific growth factor capable of enriching the gut flora. Breastfed infants have a microbiota rich in Bacteroides and Bifidobacteria. It has been demonstrated that some species of Bacteriodes consume long-chain HMOs via mucin-utilization pathways.³³ Contrarily, a number of species of Bifidobacteria consume short-chain HMOs.^34,35 One can infer from this data that long-chain HMOs serve as mimics to mucins and thus promote the growth of symbiotes, like Bacteroides, which can metabolize these molecules. Shorter-chained HMOs, however, are structurally divergent from O-linked mucin-type glycans and glycoproteins. It is conceivable then that these molecules are used specifically by certain species of Bifidobacteria which do not metabolize mucins and thus would otherwise be outcompeted. In sum, HMOs select for the growth of both HMO-metabolizing Bifidobacterium species and mucin-metabolizing Bacteroides species (Table 2).^36–38

Table 2.

HMO-Promoted Growth of Symbiotic Bacteria

symbiote	action	reference
B. bifidum, B. longum	major strains found in breastfed infant feces can grow using HMOs as the sole carbon source metabolizes “small” oligosaccharides found in human milk	39
B. breve, B. adolescentis	major strains associated with adult gut flora do not grow efficiently on HMOs	39
B. fragilis, B. thetaiotaomicron	HMO use in B. fragilis and B. thetaiotaomicron coupled to up-regulation of mucin degradation pathways	33 and 40
B. ovatus, B. stercoris	do not exhibit growth in the presence of HMOs	33 and 40
L. plantarum, L. acidophilus	do not digest complex HMOs metabolize neutral HMOs ferment lactose, glucose, N-acetylglucosamine, and fucose	41 and 42
L. reuteri, L. fermentum, S. thermophilus	do not metabolize HMOs	41 and 42

HMO-MEDIATED PATHOGEN PROTECTION

It has been established that breastfed infants experience decreased instances of diarrhea, respiratory infection, urinary tract infection, ear infection, necrotizing enterocolitis (NEC), and sudden infant death syndrome (SIDS), compared to their formula-fed counterparts.^15,43,44 In agreement with these findings, breastfed neonates tend to be colonized to a lesser extent by infectious species such as E. coli, C. diff, and C. jejuni. Importantly, many of these protective properties have been attributed to the HMO component of milk.^45–50 For example, a study by the Donovan lab showed that HMO supplementation shorted the duration of rotavirus infection. Rotavirus is one of the leading causes of diarrhea in infants.⁵¹ Bovine milk, from which most formula is based, however, contains a negligible oligosaccharide component. Additionally, bovine milk oligosaccharides (BMOs) lack the structural complexity and diversity of HMOs.¹⁷ Consequently, formula-fed infants unfortunately do not obtain comparable oligosaccharide-fostered protections as those that are breastfed.

Broadly speaking, HMO-fostered protection can be broken down into two categories. The first is protection resulting from the selective metabolism of HMOs by symbiotic bacteria. Selective utilization by symbiotes, such as Bifidobacteria, affords these species a competitive edge over pathogens which cannot metabolize HMOs. In a study by the Miller lab, it was found that 0 of 10 Enterobacteriaceae strains tested, including several E. coli strains and one Shigella dysenteriae strain, were incapable of growing on the HMOs 2′-fucosyllactose (2′-FL), 6′-sialyllactose (6′-SL), and lacto-N-netotetraose (LNnT). Several of these strains were, however, able to grow on galactooligosaccharides (GOS) and mono- and disaccharide HMO components.⁵² As a result of this selective metabolism, symbiotes can grow and outcompete harmful pathogens. Moreover, metabolism of HMOs results in the production of short-chain fatty acids (SCFAs). SCFAs lower the pH of the gut which further stunts the growth of many pathogenic species.⁴⁶

The second protective mechanism arises from more direct interaction with pathogens. Namely, HMOs can act as antiadhesive antimicrobials by serving as soluble decoy receptors for pathogens or pathogenic virulence agents such as toxins. This ability is made possible by the resemblance of HMOs to various cell surface glycan receptors. As a result, pathogens bind to HMOs rather than to cell surface glycans thereby prohibiting the binding of pathogenic species to epithelial cells which is often the first step of infection. For instance, the potential for HMOs to protect against norovirus infection is hypothesized by Hansman and coworkers to be due to commonalities between HMO and histo-blood group antigen (HBGA) fucosylation patterns. These structural similarities would allow HMOs to act as natural decoys for the HBGA binding pocket of noroviruses.^53,54 Similar findings by the Newburg lab have been seen for C. jejuni wherein α-1,2 fucosylated HMOs were able to inhibit adherence to host cell receptors.^55,56 A summary of HMO-fostered protections against numerous bacterial pathogens is provided in Table 3.

Table 3.

HMO-Fostered Inhibition of Bacterial Pathogens

bacterial species	action	HMOs	reference
Acinetobacter baumannii	inhibition of growth	pooled HMOs	unpublished
Campylobacter jejuni	inhibition of adhesion to epithelial cells	2′-FL	55, 56, and 59
	inhibition of inflammatory signaling	other 2-linked fucosylated oligosaccharides
	decreased Campylobacter-attributable diarrhea
Candida albicans	inhibition of adhesion to epithelial cells	pooled HMOs	60
	interference with hyphal morphogenesis
Clostridium difficile	binding to exotoxins A (TcdA) and B (TcdB) (prevents interactions of toxin with cellular receptors)	fucosylated single-entity HMOs (e.g., LNFP I, LNFP III)	61 and 62
		acidic single-entity HMOs (e.g., LST b and c) LNT, LNnH
Enterococcus faecium	faster vancomycin-resistant E. faecium (VRE) colonization reduction compared to non-HMO treatment	mixtures of fucosylated HMOs	63
Escherichia coli	interference with intracellular signals used by UPEC to cause cell damage	acidic and neutral HMO mixtures	64–68
	inhibition of UPEC adhesion to epithelial cells	neutral and acidic single-entity HMOs (e.g., 2′-FL, 6′-SL, LNFP I and II)
	inhibition of EPEC adhesion to epithelial cells
	binding to heat-labile enterotoxin type 1 (HLT)
Haemophilus influenzae	inhibition of adhesion to epithelial cells	high molecular weight fraction of milk	69
Helicobacter pylori	inhibition of adhesion to epithelial cells	acidic HMOs (e.g., 3′-SL and 6′-SL)	70
Pseudomonas aeruginosa	inhibition of adhesion to epithelial cells	2′-FL and 3-FL	71 and 72
	reduction of adhesion to and internalization in pneumocytes	3′-SL and 6′-SL
Streptococcus agalactiae	bacteriostatic and antibiofilm (mechanism not yet determined)	neutral HMOs mixtures	57 and 58
		single-entity HMOs (e.g., LNT and LNFP I) pooled HMOs
Streptococcus pneumoniae	inhibition of adhesion to epithelial cells	low and high molecular weight milk fractions	69
		single-entity HMOs (e.g., LNT)
Shigella dysenteriae	binding to Shiga toxins Stx2 and Stx1B5	acidic and neutral single-entity HMOs (e.g., 2′-FL, 6′-SL, LNDFH I, LNFP III)	68
Salmonella fyris	inhibition of adhesion to epithelial cells	acidic and neutral low molecular weight HMOs (e.g., 3-FL and 6 ′-FL)	65
Staphylococcus aureus	promotes growth without HMO metabolism; proposed to act as growth stimulant	pooled HMOs	73
	inhibition of biofilm
Norovirus and Rotovirus	inhibition of binding to HBGAs [Norovirus]	sialylated HMOs (e.g., 3′-SL and 6′-FL)	53, 54, and 74
	inhibition of adhesion to epithelial cells [Rotavirus]

While inhibition of binding to cell surface glycans is a common mode of HMO-attributable protection, there nevertheless remain cases where the mechanism of inhibition has yet to be established. Work from our laboratory as well as work from the Bode laboratory has shown that HMOs have both bacteriostatic and antibiofilm activities against Streptococcus agalactiae (Group B Step, GBS). While the Bode laboratory has evidence which suggests that HMOs may serve as alternative substrates capable of impairing growth kinetics, a definitive mechanism of inhibition has yet to be determined.^57,58

The relationship between HMOs and S. aureus remains ambiguous. An early study by the McGuire lab showed that HMOs actually stimulated the growth of S. aureus.⁷³ The increased growth was not, however, attributable to the metabolism of HMOs. As a result, it was suggested that HMOs perhaps serve as growth stimulants. Recent work in our laboratory showed HMOs to act as antibiolfilm agents against S. aureus, although again, a mechanism for this activity has yet to be determined.

CHALLENGES AND FUTURE OPPORTUNITIES

It was first observed in 1900 that human milk governs the composition of the infant gut flora. A century later, the community still has a poor understanding of the mode of action for antimicrobial activity. Specifically, questions remain about exactly how HMOs select for the growth of symbiotic species while also providing the means to defeat a number of pathogens (Table 4).

Table 4.

Key Roadblocks to Leveraging the Human Milk Glycome To Fight Infectious Diseases

objective	stage
identify which pathogens are susceptible to HMOs	late
characterize and purify anti-infective HMOs	early
determine mechanism of antimicrobial or antivirulence activity	early
chemically/chemoenzymatically synthesize single-entity HMOs	early
use HMOs as dietary supplements	late

We believe the future of HMO research ultimately will involve orally administering synthetic HMOs as a new generation of antimicrobial agents or dietary supplements. In fact, a number of for-profit and nonprofit companies are pursuing this concept. For example, Sugarlogix is a biotech start-up looking to synthesize HMOs as a supplement for infant formula. Our personal view is that, depending on the pathogen, HMOs could serve as either therapeutics to treat current infections or prophylactics to prevent infection. For instance, HMO antimicrobial cocktails could be delivered to children at risk for infectious diseases.

The greatest barrier to research in the field of HMO glycobiology is the limited availability of HMOs. At a basic level, there is a limited supply of donor milk accessible in the United States. Rightfully so, this product is prioritized for sick neonates who are most likely to benefit from exclusive consumption of human milk. Given the need to prioritize milk distribution, the amount of breast milk available to researchers for preliminary studies remains small.

Reflective of this reality, a common theme in HMO glycobiology research remains that researchers resort to pooling together small donor milk samples in order to generate enough material for evaluation. While the use of pooled milk has helped elucidate several actions and benefits of HMOs, researchers nevertheless often cite this method as a shortcoming of their studies. This proclamation of limitation is more often than not followed by the authors’ lament about the lack of access to single-entity compounds and/or the prohibitively expensive nature of the few commercially available HMOs.

In general, our team is against pooling milk samples for evaluation as it removes the ability to observe differences in HMO activity based on maternal phenotype. Additionally, this approach can make identifying specific HMO structure activity relationships akin to finding a needle in a very large haystack.

A hallmark example of the limitations associated with using pooled milk samples can be seen in a recent study by the Bode lab which investigated the effects of HMOs on NEC in a neonatal rat model.⁷⁵ Initial studies found pooled HMOs to reduce NEC incidence. Further investigation, however, revealed that this protection was unique to a single HMO, disialyllacto-N-tetraose (DSLNT). Furthermore, it was found that both sialic acid residues were required to protect against NEC.

In addition to issues related to procuring HMOs, an obscure issue is that the composition of human milk changes on a regular basis. While this is certainly evolutionary genius, perhaps to defend against resistance, the scientific consequence is that, once a sample has been exhausted, one will never be able to replicate the biological results observed from the initial screening! Moreover, based on the average milk donation, approximately 4–5 bacterial strains can be assayed from a single milk sample. Thus, innovative solutions are necessary to determine which HMOs, or which combinations of HMOs, are most important for anti-infective activity. These results would in turn drive innovation in both chemical and chemoenyzmatic synthesis.

As carbohydrate synthesis continues to advance, more HMOs are becoming available.^76–78 These HMOs are, however, generally restricted to small (tri- and tetrasaccharides) and structurally simple compounds. More specifically, these structures are generally restricted to compounds featuring minimal decorations or elongations of lactose, which is the disaccharide found at the reducing end of all HMOs. Examples include monofucosylated or sialylated lactose and lactose elongated in a linear fashion with one lacto-N-biose or N-acetyllactosamine residue. Unfortunately, these HMOs represent only a small portion of the human milk glycome which is estimated to contain well-over 200 unique structures.⁷⁹ It is our hope that the high-level nature of the problem will continue to inspire novel strategies and tactics to produce complex HMOs.^77,78

There exist exciting opportunities for multidisciplinary teams to learn exactly how HMOs confer beneficial effects.⁸⁰ For example, the NIH has established a glycoscience common fund program to develop new methodologies and resources to study glycans. The goal of the program is to make carbohydrates, including HMOs, accessible to the broader biomedical research community. Recently, UC San Diego has initiated an amazing effort to study this topic. Prof. Lars Bode, an expert in HMO research, is leading an initiative which will endow a faculty chair in human milk research and support seed grants for collaborative projects. Hopefully, as time progresses, more initiatives will be established at UCSD and other universities to assist multidisciplinary teams in performing research on this frontier topic.

ABBREVIATIONS

HMO

human milk oligosaccharide

WHO

World Health Organization

MRSA

methicillin-resistant S. aureus

VRE

vancomycin-resistant E. faecium

HIV

human immunodeficiency virus

AIDS

acquired immunodeficiency syndrome

MDR

multidrug-resistant

spp

multiple species

NEC

necrotizing enterocolitis

SIDS

sudden infant death syndrome

SCFA

short-chain fatty acids

GBS

Group B Step

2′-FL

2′-fucosyllactose

3-FL

3-fucosyllactose

6′-SL

6′-sialyllactose

LNnT

lacto-N-netotetraose

GOS

galactooligosaccharides

LNFP I

lacto-N-fucopentaose I

LNFP III

lacto-N-fucopentaose III

LST b

sialyl-lacto-N-tetraose b

LST c

sialyl-lacto-N-tetraose c

LNT

lacto-N-tetraose

LNnH

lacto-N-neohexaose

UPEC

uropathogenic E. coli

EPEC

enteropathogenic E. coli

LNFP II

lacto-N-fucopentaose II

3′-SL

3′-sialyllactose

LNDFH I

lacto-N-difucohexaose I

HBGA

histo-blood group antigens

DSLNT

disialyllacto-N-tetraose