MALNUTRITION

The Human Gut Microbiota and Undernutrition Jefrey I. Gordon,1* Kathryn G. Dewey,2 David A. Mills,3,4 Ruslan M. Medzhitov5 Childhood malnutrition is a global health problem that cannot be attributed to food insecurity alone. The gut microbiota may contribute to this devastating health disorder. In this Perspective, we call for the application of tools and concepts emerging from studies of the human gut microbiota to better understand the nutritional needs of infants and children and the role of the microbiota in the pathogenesis and treatment of undernutrition. This efort will require elucidation of the interrelationships between breast milk composition and the development of the microbiota and immune system in the context of the maternal-infant dyad.

WHAT DO WE MEAN BY MALNUTRITION? Malnutrition is a general term that encompasses many diferent manifestations of inadequate nutrition, including delayed growth of children as well as signs and symptoms of defciencies in vitamins, minerals, essential fatty acids, and protein. Malnutrition includes both under- and overnutrition, but here we will focus on undernutrition. Child growth has been used as an indicator of undernutrition for decades, but the most commonly measured outcome— low weight-for-age—is not the optimal indicator because it can refect either acute or chronic undernutrition or both. Characterizing growth in terms of both heightfor-age (refective of chronic undernutrition) and weight-for-height (refective of acute undernutrition) is more useful. Low height-for-age, typically defned as less than two standard deviations (SD) below the median based on the World Health Organization (WHO) Child Growth Standards, is termed “stunting.” Low weight-for-height is termed “wasting”; wasting is considered “moderate” if between –2 and –3 SD below the median and “severe” if less than –3 SD below the median. Stunting and certain specifc nutritional defciencies are key risk factors for diminished survival, poor child 1

Center for Genome Sciences and Systems Biology, Washington University School of Medicine, St. Louis, MO 63108, USA. 2Department of Nutrition, University of California, Davis, Davis, CA 95616, USA. 3Departments of Viticulture & Enology and Food Sciences & Technology, University of California, Davis, Davis, CA 95616, USA. 4Foods for Health Institute, University of California, Davis, Davis, CA 95616, USA. 5Department of Immunobiology and Howard Hughes Medical Institute, Yale University School of Medicine, New Haven, CT 06520, USA. *Corresponding author. E-mail: [email protected]

and adult health, decreased learning capacity, and lost future productivity (1, 2). Challenges to reducing stunting. Worldwide, stunting afects 27% of children under 5 years of age in low- and middleincome countries, or a total of 170 million children (3). Stunting ofen goes unrecognized by families who live in communities where short stature is so common that it is construed as normal. Even among health workers, stunting generally does not receive the same attention as being underweight or wasting, especially if height is not routinely measured as part of community health programs. In low- and middle-income countries, stunting is more prevalent than underweight (low weight-for-age, 16%) or wasting (9%) (3). Tis is probably because growth in height is highly sensitive to poor dietary quality, not just quantity (4). Intrauterine growth retardation, inadequate nutrition to support the rapid growth and development of infants and young children, and frequent infections during early life are the factors most frequently invoked as causes of stunting (5, 6). Although a child may not be classifed as “stunted” until 2 to 3 years of age, the process of becoming stunted typically begins in utero. Tus, stunting usually refects the cumulative efects of poor nutrition and other insults that ofen span several generations (2). Undernutrition during critical periods of development has been linked to insulin-like growth factor-I (IGF-I) and growth hormone signaling pathways, which appear to modulate height in response to an unfavorable environment, whether in utero or postnatally (7, 8). Tis may “program” the child’s future growth, which may explain why reductions in stunting in response to nutritional interventions

are typically rather modest. It ofen takes several generations to see substantial declines in stunting rates (9). Te archaeological record indicates that the average height of preagricultural humans was similar to that of well-nourished modern populations and that it declined in many prehistoric societies afer the introduction of agriculture ~10,000 years ago (10). Average height has increased over the past ~100 years in most industrialized countries; this has been attributed to improved nutrition (such as increased consumption of animal-source foods) and a reduction in infectious diseases. However, in lowincome populations that obtain most of their calories from cereal-based diets, stunting remains widespread. Preagricultural humans consumed a wide variety of animalsource foods, including wild game, fsh, shellfsh, and insects. It is estimated that human hunter-gatherers obtained 45 to 65% of their energy from animal-source foods (11). In combination with a wide variety of wild plant foods such as leaves, fowers, roots and tubers, nuts, seeds, berries, and other fruits (but virtually no cereal grains), the nutritional quality of the prehistoric diet is thought to have been very high compared to modern diets (12). Te decline in dietary quality that accompanied the agricultural revolution is likely to be a major reason for the high rates of stunting that persist today. Infection and subclinical conditions. Epidemiological studies indicate that food insecurity alone cannot explain the incidence of moderate to severe forms of malnutrition in various human populations (6). Infections are very common in early life, with children under 2 experiencing an average of three to fve episodes of diarrhea per year in developing countries (13). Te incidence of diarrhea peaks at 6 to 11 months of age, as infants eat increasing amounts of complementary foods that may be contaminated, and begin to crawl and explore their environment, thus putting them in direct contact with multiple sources of pathogens (14). During an infection, the immune system requires a broad range of nutrients to mount a defense against the invading organism. Subclinical conditions may also have a strong, perhaps cumulative efect on gut function, metabolism, and growth. One condition likely to be prevalent in developing countries is environmental enteropathy (also known as tropical enteropathy). Tis condition ofen has no outward manifestation, but it can cause nutrient malabsorp-

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

1

Downloaded from stm.sciencemag.org on June 23, 2014

PERSPECTIVE

tion by changing gut structure and function and may decrease the efcacy of nutritional interventions (14). A fundamental challenge for understanding environmental enteropathy is the absence of noninvasive assays for its diagnosis, classifcation and assessment of therapeutic responses. Biopsy of the small intestine has been the most defnitive test. Changes associated with enteropathy include reductions in the height of small intestinal villi and increases in the number of intraepithelial lymphocytes and lamina propria T cells with a predominant T helper 1 (T1) phenotype. Functional tests have shown impaired mucosal barrier integrity. No single pathogen or constellation of pathogens has been identifed in the gut microbiota of a majority of individuals diagnosed with environmental enteropathy. Tis has led to the notion that this disorder could be the result of colonization with microbial consortia that induce infammation in the context of a susceptible host. What constitutes a “normal” immune repertoire in a healthy gut may vary considerably depending upon a person’s environmental exposures and microbiota confguration (15). Tus, linking the term “environmental” together with “enteropathy” is sensible because it underscores the importance of placing a host’s immune system and gut microbial community in the context of exposures to various environmental factors. The frst 1000 days. Tere is a growing consensus that pregnancy and the frst 2 years of life are the most critical time period for interventions to improve child growth and development. Tis is the age of greatest vulnerability to undernutrition and infection. Potentially irreversible long-term physical and mental damage due to undernutrition occurs during this period; for example, it is difcult to reverse stunting if intervention begins afer age 2. In addition, the gut microbial community is established within the frst 3 years of life, and repair of defects in the development of this metabolic “organ” during this time period may promote healthy growth. Although a variety of environmental and genetic factors have been postulated to infuence the development of moderate to severe forms of undernutrition, the underlying mechanisms remain poorly defned. For example, in Malawi, discordance for moderate to severe malnutrition between twins within the same household and fed similar diets is only approximately 40% (16). Tis

fnding raises a number of questions. Do diferent gut community confgurations predispose one co-twin to malnutrition? Is the incidence of discordance for moderate or severe forms of malnutrition diferent in monozygotic compared to dizygotic twin pairs? Both the microbiota and breast milk are “transmitted” from mother to infant, yet we have much to learn about how breast milk and the infant microbiota codevelop during the suckling period when a mother is healthy or when she is malnourished. Important feedback mechanisms may be revealed. Undernutrition may delay the maturation of the infant’s gut microbiota or skew it toward a diferent and persistent confguration that either lacks necessary functions for healthy growth or increases the risk for disease, including immuno-infammatory disorders. If this is the case, would nutrient repletion return the gut microbiota to a “normal” pre-defcient state, or are there persistent structural and functional perturbations that require continued nutritional supplementation to correct? BREAST MILK, THE GUT MICROBIOTA, AND UNDERNUTRITION. Human milk, the ultimate “functional food,” helps to orchestrate development of the infant. In addition to its nutritional components, human milk contains a constellation of bioactive and immunological factors that drive maturation of the infant intestine and companion microbial community. Various studies have shown that breast-feeding protects infants in both economically developed and less developed countries (17–19). Protection is aforded by an array of biomolecules, including secretory IgA (sIgA), lactoferrin, lysozyme, α-lactalbumin, complex lipids, as well as free oligosaccharides and other glycoconjugates. Secretory IgA, which is present at high concentrations in peripartum breast milk (colostrum) (19), complements the lack of sIgA production in newborns, thus providing the neonate time to develop its capacity for immunoglobulin production. Free oligosaccharides, which serve as decoy receptors for epithelial binding sites, are also found at higher concentrations in colostrum than in mature milk (20, 21), thereby reducing colonization by pathogens while simultaneously promoting establishment of benefcial members of the microbiota, including Bifdobacterium spp. (through mechanisms that remain to be defned) (22). Colostrum contains higher concentrations of total protein than

mature milk, and there are changes in overall protein composition over the course of lactation (23). Milk proteins also undergo dramatic changes in glycosylation during lactation (24). Changes in milk protein glycosylation are known to impact protein function, stability, and structure (25) and to infuence interactions with pathogens (26). Other factors that modulate the gut microbiota are activated during transit of milk components through the infant gastrointestinal tract. Proteolytic processing of the glycoprotein κ-casein releases glycomacropeptide, which likely acts as a receptor analog preventing colonization of the gut by pathogens (27). Similarly, lactoferrin is converted by proteolysis to lactoferricin, a potent antimicrobial (28). Milk provides an array of lipids to the growing infant. Tis component includes complex lipid structures: eicosanoids (29), phospholipids, glycosphingolipids, free cholesterol, cholesterol esters, and cholesterol intermediates. Te fatty acid composition of maternal milk is refective of maternal stores, diet, and synthesis in the mammary gland (30, 31). Long chain polyunsaturated fatty acids such as docosahexaenoic and arachidonic acids are necessary for growth and neurological development, yet concentrations vary globally among low- to high- income countries (32). Fatty acid composition changes over the course of lactation, with colostrum generally possessing less linoleic and alpha-linolenic acids and a higher percentage of long chain polyunsaturated fatty acids (33). Mean triglyceride content increases during the frst week postpartum and then remains constant throughout lactation. In contrast, cholesterol decreases over the course of lactation (34), whereas the phospholipid concentration remains stable (35). Human milk also contains a number of other factors, including cytokines and chemokines, whose production is delayed in the neonate (36). Human epidermal growth factor, which produces a variety of biological responses, including repair of damaged mucosa, is found at the highest concentration in breast milk produced during the frst few days afer parturition and then declines over the frst month of lactation (37). Together, these programmed modulations attest to the amazingly dynamic (and little studied) nature of the assemblage of milk factors delivered to the developing infant. The gut microbiota and postnatal development. Studies of infants, children,

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

2

Downloaded from stm.sciencemag.org on June 23, 2014

PERSPECTIVE

teenagers, and adults living in countries with diferent life-styles and cultural traditions are beginning to reveal shared as well as distinct features of the gut microbiota during postnatal development. In a recent survey of 531 individuals, fecal samples were obtained from families of Guahibo Amerindians residing in two villages located near Puerto Ayacucho in the Amazonas State of Venezuela, from families living in four rural communities in Malawi, and from families distributed across the United States (38) (Fig. 1). In each population of breast-fed healthy children sampled, the gut microbiota did not reach an adultlike confguration until ~3 years of age. Te fecal bacterial species assemblages in Malawian and Amerindian gut microbiotas were more diverse and clustered together compared to Western gut communities. Tis diference in diversity was apparent in children at 3 years of age. Shared features of the functional maturation of gut microbiomes across the three populations included higher representation of genes involved in folate biosynthesis in infants with a later shif to folate metabolism and reduced representation of genes involved in (i) synthesis of cobalamin (vitamin B12), biotin (vitamin B7), and thiamine (vitamin B1), (ii) fermentation and methanogenesis, and (iii) metabolism of certain amino acids (arginine, lysine, glutamate, aspartate) in children compared to adults. However, diferences in Amerindian and Malawian children compared to children in the United States are also evident. Amerindian and Malawian children have gut microbiomes with a greater proportional representation of genes that produce ribofavin (vitamin B2). Te same is true for the genes encoding glycoside hydrolases that process glycans in breast milk and in the gut mucosa. Te microbiomes of children living in non-Western environments also exhibit more prominent representation of genes encoding urease, which breaks down urea, an important source of nitrogen in breast milk, into ammonia. Microbes can use ammonia for production of essential and nonessential amino acids. Urease production by the microbiome would be benefcial when protein is not very abundant in the diet. Diferences in the representation of genes in the microbiomes of adults in the United States compared to Malawian and Amerindian adult microbiomes follow diferences in diet (38). Although more studies of this type are needed, the results suggest that there is an identifable program of functional matura-

Fig. 1. Tackling the scourge of undernutrition. Characterizing the interrelationships between the gut microbiota and nutritional status involves surveys of healthy and malnourished individuals representing different geographical locations, cultural traditions, and life-styles. These surveys require a comprehensive biospecimen collection effort including sampling of breast milk from mothers during the course of lactation, together with assessments of maternal and infant nutritional and metabolic status and microbial ecology. The images shown were taken at sites where sampling of infants, children, and their mothers are being conducted by the authors and their colleagues. (A to F) Malawi, (G) Amazonas state of Venezuela, (H and I) Bangladesh.

tion of the human gut microbiota, driven in part by breast milk, and that this program varies between individuals living in Western compared to non-Western countries. Tese data also indicate that when defning the nutritional requirements of humans at diferent ages, we need to consider the gut microbiota. INTERACTIONS OF THE IMMUNE SYSTEM AND GUT MICROBIOTA Host-microbial relations have been traditionally considered in the context of antagonistic interactions between pathogens and the immune system. Accordingly, the best-understood function of the immune system is to detect invasive pathogens and to promote their destruction and elimina-

tion. Furthermore, many view the immune system as having been specifcally designed to attack pathogenic microbes while sparing, ignoring, or tolerating nonpathogenic (benefcial) symbionts. However, the immune system clearly detects and responds to myriad components of the microbiota: Antimicrobial defensins and IgA are just two examples of immune defenses directed at gut microbiota constituents [see Review by Hooper et al. (39)]. Te current framework of host-microbial interactions does not fully explain the decision-making process of the immune system when it comes to these resident microbes. Tere is increasing appreciation for the fact that diferent resident members of the microbiota can have a positive or negative

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

3

Downloaded from stm.sciencemag.org on June 23, 2014

CREDIT: PHOTOS COURTESY OF TANYA YATSENUNKO (A-E), INDI TREHAN (F), MARIA GLORIA DOMINGUEZ-BELLO (G), AND AMBER BENEZRA AND JOSEPH DESTEFANO (H, I).

PERSPECTIVE

impact, depending on their relative abundance, community structure, and status of the host. When host defenses are compromised, normal microbial community structure is disrupted, resulting in an outgrowth of detrimental microorganisms or the conversion of normally innocuous microorganisms into pathogens (pathobionts) if their “virulence” depends on their abundance. Tis suggests that the immune system can somehow suppress the expansion of detrimental microorganisms yet maintain the abundance of benefcial microbes. However, it is not clear how the immune system maintains the appropriate density, composition, and overall confguration of the gut microbiota. It is likely that additional modes of interaction between microbes and the immune system exist that, on the one hand, allow the immune system to detect specifc confgurations of the gut microbiota and, on the other hand, promote confgurations that are benefcial to the host. For the purpose of this discussion, “microbiota confguration” can be loosely defned as microbial consortia (with their specifc characteristics) that are relevant to the ftness of the host. Some confgurations are better than others and, therefore, an optimal confguration or confgurations must exist for a given host in a given environment. It is conceivable that the host plays an important role in determining the microbiota confguration. Te most obvious example is the provision of human milk oligosaccharides (HMOs) to HMO-consuming Bi dobacterium spp. in the gut. Tus, hostderived HMOs promote the expansion and maintenance of a coevolved bacterial clade benefcial to the infant. Conversely, a variety of antimicrobial agents (including HMOs) in breast milk suppress bacteria that are detrimental to an infant’s ftness. Lactation in mammals and analogous secretions in birds (such as “crop milk”), and other processes such as premastication, regurgitation, and consumption of feces (coprophagy) across many types of animals, may facilitate optimal microbial colonization of the gut in the ofspring. Immune sensing of microbiota confgurations. Te immune system may not only protect from invasive pathogens but also promote the optimal confguration of the gut microbiota. Although such a function would have an obvious evolutionary and physiological rationale, it is not clear how immune system function could modulate microbial ecology. We suggest that the

immune system has a sensing mechanism that evaluates the confguration of the gut microbiota. Pattern recognition by the innate immune system cannot fulfll this role, as it does not have the requisite level of resolution to discriminate between diferent classes and characteristics of microbes. An additional mode of communication between the microbiota and the host immune system could be based on sensing metabolites [see Review by Holmes et al., this issue (40) and Review by Nicholson et al. (41)]. Tere are three groups of metabolites that could be involved. Te frst group of metabolites would be uniquely produced by diferent classes of microbes from precursors present in the diet (such as short-chain fatty acids). Tese metabolic signals would be sensed by the digestive system (enteroendocrine cells of the gut mucosa, cells of the pancreas and liver, and adipocytes), where they would report on the quality and quantity of incoming nutrients. Te immune system could sense these metabolites to evaluate the characteristics of the gut microbiota. Some of these small-molecule metabolites sensed by the immune system may not be unique to microbes but rather may be produced in amounts or locations characteristic of a microbial origin. Te metabolites may have anti- or proinfammatory activity depending on whether the microbial communities they “report on” are desirable or detrimental to the host. Te second group of metabolites would be unmodifed nutrients that the immune system could use to evaluate the presence of microbes that metabolize the nutrients into forms that can no longer be sensed. Te amounts of these nutrients could be an indicator of the relative densities of the microbes that metabolize them. Certain micronutrients, such as retinoic acid, may fall into this category. Tese micronutrients would be expected to have anti-infammatory activity and their concentrations would be a function of the abundance of “detrimental” microbes that metabolize them. To function in this manner, the concentrations of micronutrients with signaling capabilities should not be limited by the host’s dietary intake (they should be abundant in the “natural” diet). Tis feature could be one factor that discriminates Western from non-Western diets. Te third group of metabolites would be produced by the host but metabolized by specifc members of a microbial con-

sortium. Bile acids produced by the liver and secreted into the jejunum can be either reabsorbed in the ileum or metabolized into secondary bile acids (for example, deoxycholic and lithocholic acids) by 7α-dehydroxylation mediated by microbes in the colon. Both primary and secondary bile acids have signaling functions, including potentially anti-infammatory activities (42). Although mechanistic details still need to be elucidated, secondary bile acids may be an example of “molecular probes” that report to the host on the presence and abundance of specifc enzymatic activities of certain microorganisms, such as bacteria that produce the enzyme 7α-dehydroxylase. In all three cases, metabolite sensing could provide the immune system with information about the presence and abundance of specifc metabolic activities and confgurations of microbial consortia. Although the impact of microbial metabolites and some nutrients on the immune system has been described (15, 43), the logic of this mode of host-microbial communication has been unclear. We propose that the host immune system’s ability to sense microbial metabolites and dietary micronutrients may serve the function of reporting on the “quality” of a gut microbiota confguration. Relative abundance of metabolites with anti-infammatory or immunosuppressive activities (for example, short-chain fatty acids) should refect the desired microbiota confguration. If the gut microbiota is suboptimal, immune responsiveness would be tuned up (or specifc efector responses would be upregulated) to promote reconfguration to a more optimal state. Tis could be achieved through specifc efects of IgA, less specifc efects of innate efectors (such as defensins), or nonspecifc efects such as diarrhea, which the host uses as a last resort to eliminate undesirable microbial communities and to prepare niches for recolonization with more optimal microbial consortia. TESTING HYPOTHESES Studying how the gut microbiota and the immune system codevelop in healthy and undernourished infants and children, and the role of breast milk in this process, would allow testing of several hypotheses. Hypothesis 1. Te gut microbiota is a metabolic organ that is needed for healthy postnatal development. Malnutrition afects the assembly of metabolic functions encoded and expressed by the gut microbiome. Te assumption being tested is that the gut micro-

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

4

Downloaded from stm.sciencemag.org on June 23, 2014

PERSPECTIVE

biota is both a biomarker and mediator of key metabolic functions needed to promote the nutritional health of mothers and the healthy growth of their ofspring. A related assumption is that the maturation of breast milk during the course of lactation impacts the assembly of the infant gut microbial community, thereby shaping the microbiota’s metabolic potential and expressed metabolic activities. A corollary is that periods of malnutrition may have long-lasting effects on microbiota-based metabolic activities that extend beyond the period of overt nutrient defciency. Tis would require sustained nutrition support to ensure rescue of the normal development of microbial and host physiological functions. Tis hypothesis addresses the possibility that the efects of undernutrition on the developmental program of the gut microbiota could be modifed at key stages of postnatal life by administering live microbes that fll missing metabolic niches (probiotics) or prebiotics that boost metabolic functions of key microbes in the gut microbiota. A combination of prebiotics and probiotic consortia (synbiotics) could be required for restoration of a benefcial microbiota. Te hypothesis raises an important question about microbiota-directed therapeutics. Would therapeutic interventions have greater efcacy if implemented early during community assembly when the microbiota manifests greater instability and seemingly stochastic behavior, or later, when directed movement toward a particular functional confguration is more apparent (although by this time, attempted “repair” of the microbiota may be more difcult to achieve)? Hypothesis 2. Te functions of the normal developing innate and adaptive immune systems are modulated by the metabolic activities of the gut microbiota. Te constituents in breast milk have a major efect on these metabolic activities and thus on the immune system. Tis testable hypothesis is motivated by the view that innate and adaptive immune cells may have sensors that respond to breast milk constituents or to their breakdown products generated by microbial metabolism. Tese sensors and their associated signaling pathways may modulate immune cell subsets, including activities relevant to maturation of intestinal barrier function, resistance and responses to infection, and robust responses to enteric vaccines. Tis hypothesis emphasizes the need to delineate the efects of breast milk–dependent bioactive compounds—that is, breast milk

components presented afer processing by the host, by the microbiota, and by hostmicrobial cometabolism—on major components of the intestine’s immune system. Hypothesis 3. Immunopathological states associated with childhood undernutrition are caused by the co-occurrence of several interrelated factors, including nutrient availability, defects in gut microbiota maturation, abnormal nutrient processing by the microbiota, aberrant nutrient sensing by immune cells, and the presence of enteropathogens. If this hypothesis is true, the gut microbiota and its metabolic activities may be useful biomarkers for diagnosing and classifying immunopathological states associated with malnutrition. Administering gut microbiota constituents from healthy donors and breast milk components could reshape microbial community metabolism and immune function in ways that mitigate the risk for developing environmental enteropathy. THE VALUE OF HUMAN STUDIES Tere are several reasons we believe that twins, particularly twins discordant for moderate or severe malnutrition, will be informative for clinical studies of the role of the gut microbiota in disease, especially if both cotwins in a discordant pair are subjected to the same nutritional intervention (16). First, substantial interpersonal differences in microbial community confgurations normally exist between unrelated healthy individuals; intra- and interpersonal variations are particularly great during the early postnatal period (38). Second, there may be a relatively large number of microbiota confgurations associated with malnutrition in children, each shared by relatively few individuals. Tird, if given the same therapeutic food intervention, each sib in a discordant twin pair could serve as his or her own control, with the healthy cotwin in the discordant pair serving as another control. Fourth, comparing discordant monozygotic and dizygotic twin pairs allows initial broad assessment of the role of host genetics in defning disease. A logical question arising from such twin studies would be whether diferences are causally or casually related to malnutrition. Recent work has shown that it is possible to transplant a previously frozen fecal microbiota (or a collection of cultured microbes generated from the fecal sample), from a given human donor (representing an age, phenotype, and cultural tradition of interest) into multiple germ-free mice.

Tere is efcient capture of the genes and taxa of the donor’s gut microbial community in the gastrointestinal tracts of these recipient gnotobiotic mice (44, 45), and these communities can be transmitted from one generation to the next. In addition, recipient mice can be fed diets that are representative of those consumed by the human donor. Tese types of experiments provide an opportunity to determine how much of a donor’s phenotype can be transmitted to gnotobiotic animal recipients (and their ofspring) via the microbiota and to perform proof-of-concept therapeutic tests with prebiotics, probiotic consortia, or synbiotics. In principle, these approaches can be used to determine whether the efects of nutrient supplements and supplement-responsive taxa can be generalized to humans living in multiple geographic locations with distinct cultural traditions and diets. Such preclinical platforms could also allow identifcation and validation of new surrogate- and mechanism-based biomarkers of efcacy, as well as dosing schemes and safety tests that will inform clinical study design. Te advantage of these “personalized” gnotobiotic models is that the targeted human population of interest has to be defned at the very onset of model creation, thereby facilitating subsequent translation. Te discovery pipelines created by these preclinical models and the clinical studies that they could support raise a number of challenges for regulatory agencies, issues about ownership of microbes, and ethical challenges related to how microbiota-directed therapeutics could or should be applied to mothers and infants. Tey also represent formidable challenges and tremendous opportunities for the agricultural, food, and pharmaceutical industries; government agencies; and other elements of society. With the world’s human population predicted to top 9 billion by mid-century, new approaches and their implementation to improve nutritional health cannot come soon enough. REFERENCES AND NOTES 1. C. G. Victora, M. de Onis, P. C. Hallal, M. Blössner, R. Shrimpton, Worldwide timing of growth faltering: Revisiting implications for interventions. Pediatrics 125, e473–e480 (2010). 2. K. G. Dewey, K. Begum, Long-term consequences of stunting in early life. Matern. Child Nutr. 7 (suppl. 3), 5–18 (2011). 3. C. K. Lutter, B. M. Daelmans, M. de Onis, M. T. Kothari, M. T. Ruel, M. Arimond, M. Deitchler, K. G. Dewey, M. Blössner, E. Borghi, Undernutrition, poor feeding practices, and low coverage of key nutrition interventions. Pediatrics 128, e1418–e1427 (2011).

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

5

Downloaded from stm.sciencemag.org on June 23, 2014

PERSPECTIVE

4. M. H. Golden, Proposed recommended nutrient densities for moderately malnourished children. Food Nutr. Bull. 30 (suppl.), S267–S342 (2009). 5. E. A. Frongillo Jr., Symposium: Causes and Etiology of Stunting. Introduction. J. Nutr. 129 (suppl.), 529S–530S (1999). 6. R. Martorell, M. F. Young, Patterns of stunting and wasting: Potential explanatory factors. Adv. Nutr. 3, 227–233 (2012). 7. A. Larnkjær, C. Mølgaard, K. F. Michaelsen, Early nutrition impact on the insulin-like growth factor axis and later health consequences. Curr. Opin. Clin. Nutr. Metab. Care 15, 285–292 (2012). 8. T. D. Walters, A. M. Griffiths, Mechanisms of growth impairment in pediatric Crohn’s disease. Nat. Rev. Gastroenterol. Hepatol. 6, 513–523 (2009). 9. M. de Onis, M. Blössner, E. Borghi, Prevalence and trends of stunting among pre-school children, 1990-2020. Public Health Nutr. 14, 1–7 (2011). 10. C. S. Larsen, Animal source foods and human health during evolution. J. Nutr. 133 (suppl. 2), 3893S–3897S (2003). 11. L. Cordain, J. B. Miller, S. B. Eaton, N. Mann, S. H. A. Holt, J. D. Speth, Plant-animal subsistence ratios and macronutrient energy estimations in worldwide hunter-gatherer diets. Am. J. Clin. Nutr. 71, 682–692 (2000). 12. S. B. Eaton, S. B. Eaton, 3rd, Paleolithic vs. modern diets— Selected pathophysiological implications. Eur. J. Nutr. 39, 67–70 (2000). 13. M. Kosek, C. Bern, R. L. Guerrant, The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bull. World Health Organ. 81, 197–204 (2003). 14. K. G. Dewey, D. R. Mayers, Early child growth: How do nutrition and infection interact? Matern. Child Nutr. 7 (suppl. 3), 129–142 (2011). 15. A. L. Kau, P. P. Ahern, N. W. Griffin, A. L. Goodman, J. I. Gordon, Human nutrition, the gut microbiome and the immune system. Nature 474, 327–336 (2011). 16. T. Yatsunenko, M. Manary, I. Trehan, R. Knight, J. I. Gordon, Human gut microbiome in children with severe malnutrition: a comparative metagenomic study of Malawian twins. Abstract in Malnutrition, Gut-Microbial Interactions and Mucosal Immunity to Vaccines. p 85. Part of the Keystone Symposia Global Health Series, New Deli, India, November 7-11, 2011 (Publisher: Keystone Symposia, Silverthorne, Colorado, 2011). 17. L. Duijts, M. K. Ramadhani, H. A. Moll, Breastfeeding protects against infectious diseases during infancy in industrialized countries. A systematic review. Matern. Child Nutr. 5, 199–210 (2009). 18. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality, Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: A pooled analysis. Lancet 355, 451–455 (2000). 19. G. M. Ruiz-Palacios, J. J. Calva, L. K. Pickering, Y. LopezVidal, P. Volkow, H. Pezzarossi, M. S. West, Protection of breast-fed infants against Campylobacter diarrhea by antibodies in human milk. J. Pediatr. 116, 707–713 (1990). 20. R. M. Erney, W. T. Malone, M. B. Skelding, A. A. Marcon, K. M. Kleman-Leyer, M. L. O’Ryan, G. Ruiz-Palacios, M. D. Hilty, L. K. Pickering, P. A. Prieto, Variability of human milk

21.

22.

23.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

neutral oligosaccharides in a diverse population. J. Pediatr. Gastroenterol. Nutr. 30, 181–192 (2000). S. Martín-Sosa, M. J. Martín, L. A. García-Pardo, P. Hueso, Sialyloligosaccharides in human and bovine milk and in infant formulas: Variations with the progression of lactation. J. Dairy Sci. 86, 52–59 (2003). A. M. Zivkovic, J. B. German, C. B. Lebrilla, D. A. Mills, Human milk glycobiome and its impact on the infant gastrointestinal microbiota. Proc. Natl. Acad. Sci. U.S.A. 108 (suppl. 1), 4653–4658 (2011). B. Lönnerdal, E. Forsum, L. Hambraeus, A longitudinal study of the protein, nitrogen, and lactose contents of human milk from Swedish well-nourished mothers. Am. J. Clin. Nutr. 29, 1127–1133 (1976). J. W. Froehlich, E. D. Dodds, M. Barboza, E. L. McJimpsey, R. R. Seipert, J. Francis, H. J. An, S. Freeman, J. B. German, C. B. Lebrilla, Glycoprotein expression in human milk during lactation. J. Agric. Food Chem. 58, 6440–6448 (2010). S. E. O’Connor, B. Imperiali, Modulation of protein structure and function by asparagine-linked glycosylation. Chem. Biol. 3, 803–812 (1996). M. Barboza, J. Pinzon, S. Wickramasinghe, J. W. Froehlich, I. Moeller, J. T. Smilowitz, L. R. Ruhaak, J. Huang, B. Lönnerdal, J. B. German, J. F. Medrano, B. C. Weimer, C. B. Lebrilla, Glycosylation of human milk lactoferrin exhibits dynamic changes during early lactation enhancing its role in pathogenic bacteria-host interactions. Mol. Cell. Proteomics, (published online 19 January 2012 10.1074/ mcp.M111.015248). E. P. Brody, Biological activities of bovine glycomacropeptide. Br. J. Nutr. 84 (suppl. 1), S39–S46 (2000). E. M. Jones, A. Smart, G. Bloomberg, L. Burgess, M. R. Millar, Lactoferricin, a new antimicrobial peptide. J. Appl. Bacteriol. 77, 208–214 (1994). J. S. Hawkes, D. L. Bryan, M. J. James, R. A. Gibson, Cytokines (IL-1beta, IL-6, TNF-alpha, TGF-beta1, and TGFbeta2) and prostaglandin E2 in human milk during the first three months postpartum. Pediatr. Res. 46, 194–199 (1999). N. Fidler, T. Sauerwald, A. Pohl, H. Demmelmair, B. Koletzko, Docosahexaenoic acid transfer into human milk after dietary supplementation: A randomized clinical trial. J. Lipid Res. 41, 1376–1383 (2000). M. Rodriguez-Cruz, A. R. Tovar, B. Palacios-González, M. Del Prado, N. Torres, Synthesis of long-chain polyunsaturated fatty acids in lactating mammary gland: Role of Delta5 and Delta6 desaturases, SREBP-1, PPARalpha, and PGC-1. J. Lipid Res. 47, 553–560 (2006). J. T. Brenna, B. Varamini, R. G. Jensen, D. A. DiersenSchade, J. A. Boettcher, L. M. Arterburn, Docosahexaenoic and arachidonic acid concentrations in human breast milk worldwide. Am. J. Clin. Nutr. 85, 1457–1464 (2007). R. A. Gibson, G. M. Kneebone, Fatty acid composition of human colostrum and mature breast milk. Am. J. Clin. Nutr. 34, 252–257 (1981). G. Harzer, M. Haug, I. Dieterich, P. R. Gentner, Changing patterns of human milk lipids in the course of the lactation and during the day. Am. J. Clin. Nutr. 37, 612–621 (1983). H. Shoji, T. Shimizu, N. Kaneko, K. Shinohara, S. Shiga, M. Saito, K. Oshida, T. Shimizu, M. Takase, Y. Yamashiro, Comparison of the phospholipid classes in human milk in Japanese mothers of term and preterm infants. Acta Paediatr. 95, 996–1000 (2006).

36. R. Garofalo, Cytokines in human milk. J. Pediatr. 156, (Suppl), S36–S40 (2010). 37. B. Dvorak, Milk epidermal growth factor and gut protection. J. Pediatr. 156 (suppl.), S31–S35 (2010). 38. T. Yatsunenko, F. E. Rey, M. J. Manary, I. Trehan, M. G. Dominguez-Bello, M. Contreras, M. Magris, G. Hidalgo, R. N. Baldassano, A. P. Anokhin, A. C. Heath, B. Warner, J. Reeder, J. Kuczynski, J. G. Caporaso, C. A. Lozupone, C. Lauber, J. C. Clemente, D. Knights, R. Knight, J. L. Gordon, Human gut microbiome viewed across age and geography. Nature (2012). 10.1038/nature11053 39. L. V. Hooper, D. R. Littman, A. J. Macpherson, Interactions between the microbiota and the immune system. Science, in press (2012). 40. E. Holmes, J. Kinross, G. G. Gibson, R. Burcelin, W. Jia, S. Pettersson, J. K. Nicholson, Therapeutic modulation of microbiota-host metabolic interactions. Sci. Transl. Med. 4, 137rv6 (2012). 41. J. K. Nicholson, E. Holmes, J. Kinross, R. Burcelin, G. Gibson, W. Jia, S. Pettersson, Host-gut microbiota metabolic interactions. Science 336, 1262 (2012). 42. P. Lefebvre, B. Cariou, F. Lien, F. Kuipers, B. Staels, Role of bile acids and bile acid receptors in metabolic regulation. Physiol. Rev. 89, 147–191 (2009). 43. K. M. Maslowski, C. R. Mackay, Diet, gut microbiota and immune responses. Nat. Immunol. 12, 5–9 (2011). 44. P. J. Turnbaugh, V. K. Ridaura, J. J. Faith, F. E. Rey, R. Knight, J. I. Gordon, The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci. Transl. Med. 1, 6ra14 (2009). 45. A. L. Goodman, G. Kallstrom, J. J. Faith, A. Reyes, A. Moore, G. Dantas, J. I. Gordon, Extensive personal human gut microbiota culture collections characterized and manipulated in gnotobiotic mice. Proc. Natl. Acad. Sci. U.S.A. 108, 6252–6257 (2011). Acknowledgments: We thank the Bill & Melinda Gates Foundation for challenging us to better understand the relationship between breast milk, gut microbiome, immunity, and the healthy growth of children worldwide. We are grateful to our colleagues who are members of the Breast Milk, gut Microbiome, and Immunity (BMMI) Project funded by the Bill & Melinda Gates Foundation: Kenneth Maleta, Per Ashorn, Tahmed Ahmed, Gagandeep Kang, Pascal Bessong, Aldo Lima, Margaret Kosek, Michael Gottlieb, Dennis Lang, Jeremy Nicholson, Elaine Holmes, Linda Saif, Rob Knight, Michael Barratt, Karen Seibert, David D’Argenio, and Andrew Serazin. JIG thanks Mark Manary, Indi Trehan, Michelle Smith, and Tanya Yatsunenko for their collaboration with twin studies of childhood malnutrition in Malawi. DAM thanks Jennifer Smilowitz for her contributions to the description of changes in breast milk following parturition. Funding: Authors are supported by grants from the Bill & Melinda Gates Foundation (J. I. G., K. G. D., D. A. M.), the NIH (J. I. G., D. A. M., R. M. M.), the Crohn’s and Colitis Foundation of America (J. I. G.), the California Dairy Research Foundation (D. A. M.), the University of California Discovery Program (D. A. M.), and the Howard Hughes Medical Institute (R. M. M.).

10.1126/scitranslmed.3004347 Citation: J. I. Gordon, K. G. Dewey, D. A. Mills, R. M. Medzhitov, The human gut microbiota and undernutrition. Sci. Transl. Med. 4, 137ps12 (2012).

www.ScienceTranslationalMedicine.org 6 June 2012 Vol 4 Issue 137 137ps12

6

Downloaded from stm.sciencemag.org on June 23, 2014

PERSPECTIVE