Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate Nestlé Nutrition Institute Workshop Series

Vol. 94 Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate

Editors

Pearay L. Ogra Buffalo, NY W. Allan Walker Boston, MA Bo Lönnerdal Davis, CA © 2020 Nestlé Nutrition Institute, Switzerland CH 1814 La Tour-de-Peilz S. Karger AG, P.O. Box, CH–4009 Basel (Switzerland) www.karger.com

Library of Congress Cataloging-in-Publication Data

Names: Nestle Nutrition Workshop (94th : 2019 : Lausanne, Switzerland), author. | Ogra, Pearay L., editor. | Walker, W. Allan, editor. | Lonnerdal, Bo, 1938- editor. | Nestle Nutrition Institute, issuing body. Title: Milk, mucosal immunity and the microbiome : impact on the neonate / editors, Pearay L. Ogra, W. Allan Walker, Bo Lonnerdal. Other titles: Nestle Nutrition Institute workshop series ; v. 94. 1664-2147 Description: Basel ; Hartford : Karger ; Switzerland : Nestle Nutrition Institute, [2020] | Series: Nestle Nutrition Institute workshop series, 1664-2147 ; vol. 94 | Includes bibliographical references and index. | Summary: “This publication covers the 94th Nestle Nutritional Institute Workshop, which was designed to provide a comprehensive overview on the latest human milk research and its role in modulating mucosal immunity, the microbiome, and its impact on the neonate. This publication should provide scientific support to anyone seeking a deeper understanding of human milk and its immunological properties, and enlarge the knowledge of those who specialize in human milk research”-- Provided by publisher.

The material contained in this volume was submitted as previously unpublished material, except in the instances in which credit has been given to the source from which some of the illustrative material was derived. Great care has been taken to maintain the accuracy of the information contained in the volume. However, neither Nestlé Nutrition Institute nor S. Karger AG can be held responsible for errors or for any consequences arising from the use of the information contained herein. © 2020 Nestlé Nutrition Institute (Switzerland) and S. Karger AG, Basel (Switzerland). All rights reserved. This book is protected by copyright. No part of it may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, or recording, or otherwise, without the written permission of the publisher.

Printed on acid-free and non-aging paper (ISO 9706) ISBN 978–3–318–06684–5 e-ISBN 978–3–318–06685–2 ISSN 1664–2147 e-ISSN 1664–2155

Basel · Freiburg · Hartford · Oxford · Bangkok · Dubai · Kuala Lumpur · Melbourne · Mexico City · Moscow · New Delhi · Paris · Shanghai · Tokyo Contents

VII Preface X Foreword XII Contributors XV Dedication

Immunology of Milk and Lactation

1 The Evolution of Lactation in Mammalian Species Oftedal, O.T. (USA) 11 Immunology of Human Milk and Lactation: Historical Overview Ogra, P.L. (USA) 27 The Mammary Gland as an Integral Component of the Common Mucosal Immune System Mestecky, J. (USA/Czech Republic) 38 Immunomodulatory Components of Human Colostrum and Milk Tlaskalová-Hogenová, H.; Kverka, M.; Hrdý, J. (Czech Republic) 48 Breastfeeding, a Personalized Medicine with Influence on Short- and Long-Term Immune Health Verhasselt, V. (Australia) 59 Summary on Immunology of Milk and Lactation Ogra, P.L. (USA)

Microbiology of Milk and Lactation: Influence on Gut Colonization

65 Milk Microbiome and Neonatal Colonization: Overview Rautava, S. (Finland) 75 Human Milk Microbiota: Origin and Potential Uses Fernández, L.; Rodríguez, J.M. (Spain) 86 Beyond the Bacterial Microbiome: Virome of Human Milk and Effects on the Developing Infant Mohandas, S.; Pannaraj, P.S. (USA)

V 94 Gut Microbiota, Host Gene Expression, and Cell Traffic via Milk Neu, J. (USA) 103 Breast Milk and Microbiota in the Premature Gut: A Method of Preventing Necrotizing Enterocolitis Walker, W.A.; Meng, D. (USA) 113 Summary on Microbiota of Milk and Lactation: Influence on Gut Colonization Walker, W.A. (USA)

Protective Factors in Human Milk

115 Human Milk Oligosaccharides: Structure and Functions Bode, L. (USA) 124 Oligosaccharides and Viral Infection: Human Milk Oligosaccharides versus Algal Fucan-Type Polysaccharides Hanisch, F.-G. (Germany); Aydogan, C. (Switzerland) 133 Milk Fat Globule Membranes: Effects on Microbiome, Metabolome, and Infections in Infants and Children Hernell, O. (Sweden); Lönnerdal, B. (USA); Timby, N. (Sweden) 141 Clinical Trials of Lactoferrin in the Newborn: Effects on Infection and the Gut Microbiome Embleton, N.D.; Berrington, J.E. (UK) 152 Effects of Milk Osteopontin on Intestine, Neurodevelopment, and Immunity Jiang, R.; Lönnerdal, B. (USA) 158 Effects of Milk Secretory Immunoglobulin A on the Commensal Microbiota Dunne-Castagna, V.P.; Mills, D.A.; Lönnerdal, B. (USA) 169 Summary on Protective Factors in Human Milk Lönnerdal, B. (USA)

172 Subject Index

For more information on related publications, please consult the NNI website: www.nestlenutrition-institute.org

VI Contents Published online: April 8, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp VII–IX (DOI:10.1159/000505372)

Preface

Lactation and the process of breastfeeding has been integral to the survival and long-term well-being of neonates and infants of most mammalian species. Mothers’ milk has been considered as a complete source of nutrition for the suckling infant from times immemorial, often with many magical healing powers. It is only in the past 2 centuries that significant scientific information has become available about the evolution of the mammary glands and the development of lactation and its impact on the suckling mammalian neonate. Since the observations of Paul Ehrlich over 120 years ago, it is now clear that mammalian breastfeeding is associated with significant reduction in infant mortality, protection against enteric, respiratory, and other mucosal and systemic infections, and protection against the development of allergic disorders. Recent observations have demonstrated that breastfeeding has a profound impact on the development and function of the neonatal immune system, mucosal microbiological homeostasis, and long-term protection against autoimmune and other inflammatory disease states and malignancy. The nutritional and immunobiological benefits attributed to breastfeeding are related to the diverse spectrum of specific cellular and soluble products present in the early colostrum and milk. It is now clear that the immunologic activity in the products of lactation represents the effector functional elements of the common mucosal immune system. The discovery of the secretory IgA (SIgA) immunoglobulin isotype in the milk, followed by the identification of antibacterial, antiviral, and antiparasitic activity in the milk associated with SIgA and demonstration of important elements of cellular immunity in the milk represent crucial milestones in the understanding of lactation as the single most important element of neonatal health in most mammalian species. The 94th Nestlé Nutrition Institute workshop is dedicated to Prof. Lars A. Hanson, who was the first investigator to identify SIgA in the colostrum and milk. He has been one of the most devoted scholars to the study of mammalian lactation and breastfeeding, and has also been instrumental in a global effort to foster breastfeeding in the developing world and for undernourished infants. He is rightfully considered as the “father of modern breastfeeding.” This workshop was designed to develop a comprehensive perspective about available information on the evolution of lactation in most mammalian species, and to examine in some detail the origin, composition, and functional characteristics of different nutritional and immunologic components in mammalian milk and other lactation products. Specifically, the workshop focused on the following areas: (1) the evolution of mammalian lactation and breast feeding; (2) immunologic aspects of cellular and soluble products in milk; (3) the microbial composition of human milk and its effects on the development of the mucosal microbiome in the suckling infant, and (4) the role of oligosaccharides, antimicrobial peptides, and other important soluble compounds in the products of lactation, especially in the human milk. The evolution of mammary glands and lactation has been the subject of deep interest and considerable speculation since Darwin. More recently, it has been possible to compare the evolutionary development of mammary glands across diverse taxa. The workshop began with a comprehensive discussion of the genetic origins and functions of different components of lactation across the evolutionary tree. The keynote addressed by Olav T. Oftedal introduced the prevailing concepts about the evolution of mammary glands and different constituents in the products of lactation. It was suggested that appearance of a milk protein across monotremes, marsupials, and eutherians, the 3 major mammalian taxa, may indicate that the protein evolved before these groups diverged and is inherited from the ancestral taxa. Milk constituents of evolutionary interest are lactalbumin, caseins, milk fat globules, several proteins, including antimicrobial peptides, butrophylin 1A1, and xanthine, which will also be discussed in the context of their immunologic functions and host defense. The workshop program was designed to provide a historical perspective of the immunology of milk and its impact on the breastfeeding infant. During subsequent presentations in session I by Pearay L. Ogra, Jiri Mestecky, Helena Tlaskalová-Hogenová et al., and Valerie Verhasselt, we aimed to elucidate the recent observations on the common mucosal immune system and the maternal contribution to the development of immunity in the neonate in some detail. Subsequently, the immunomodulatory components of human milk, influence of breastfeeding on long-term health, and the role of breastfeeding in preventing allergy and infection will be explored in some detail. Session II of the workshop was designed to examine the importance of microbiota and their metabolic products in breast milk, and the breast milk microbiome with regard to neonatal colonization was discussed. Although there is no direct proof that the breast milk microbiome contributes significantly to coloni-

VIII Ogra/Walker/Lönnerdal zation, the combination of the breast milk microbiome and factors in breast milk which facilitate growth of health-producing bacteria was emphasized. We know that the microbiome of milk comes from various sources (skin, baby’s oral cavity, and the maternal gut). Since the newborn encounters breast milk (colostrum) at the beginning of life, breastfeeding is important as an initial colonizer. Maternal gut microbiota gain access to the breast milk through hormonal al- terations in the gut barrier and uptake by macrophages which travel to the breast. When we consider microorganisms in breast milk, we often overlook the virome. However, new evidence suggests that breast milk contains a considerable viral component. This viral component is being considered more relevant to infant health and to the microbiome composition. The dynamic composition of breast milk with regard to microbes and their metabolites and the molecules such as IgA which affect colonization collectively provides important defense to the newborn infant. A new observation regarding breast milk is that it supplies substrates for bacterial metabolism which allow anti-inflammatory metabolites to form. The topics of specific consideration include an overview of the milk, microbiome, and neonatal colonization; the origin of the milk microbiome and its potential utility; the current status of the milk virome; gut microbiota, host gene expression, and cell traffic via breastfeeding; and protection from necrotizing enterocolitis and breastfeeding and microbiota. These areas are discussed by Samuli Rautava, Leónides Fernández and Juan M. Rodríguez, Sindhu Mohandas and Pia S. Pannaraj, Josef Neu, and W. Allan Walker and Di Meng, respectively. Finally, we considered to examine in some detail other important protective factors in human milk. The objective of session III was to explore the structure and function of milk oligosaccharides, the role of select oligosaccharides in viral infection, milk fat globules, and their effects on the mucosal microbiome. Other presentations explore the effects of lactoferrin, osteopontin, and effects of other proteins and antibodies on the mucosal microflora in breastfed infants. These areas of milk protective factors will be discussed by Lars Bode, Franz-Georg Hanisch and Cem Aydogan, Olle Hernell et al., Nicholas D. Embleton and Jannet E. Berrington, Rulan Jiang and Bo Lönnerdal, and Vanessa P. Dunne-Castagna et al., respectively. The organizers of this workshop have made every attempt to provide a balanced state-of-the-art update on the current knowledge of milk, mucosal immunity, and the microbiome as well as their impact on breastfeeding in mammalian neonates. We hope the readership of the NNI workshop series will find this information helpful in their own endeavors in specific areas discussed during this workshop. Pearay L. Ogra W. Allan Walker Bo Lönnerdal

Preface IX Published online: April 8, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp X–XI (DOI:10.1159/000505245)

Foreword

There have been considerable advances in science to understand the varied mixture of bioactive components in human milk that influence the immune sta tus of infants not only by providing protection, but also by facilitating develop ment, tolerance, and an appropriate inflammatory response. Human milk is the communication vehicle between the maternal immune system and the infant, a system actively directing and educating the immune, metabolic, and microfloral systems within the infant. The physiological and pro tective functions of several immune components in human milk have been stud ied not only in infants, further evidence has also been obtained from what is known in other species and in vitro models. The 94th Nestlé Nutrition Institute (NNI) Workshop entitled Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate, which took place in Lau sanne in September 23–25, 2019, reviewed the latest data on immune develop ment in infants and the role of milk factors. The program was focusing on the current knowledge of how both the “classical” immune and nonimmune ingre dients found in human milk support maturation of the immune system, facili tate development of tolerance, and regulate inflammatory responses of infants. World experts in human milk research and nutrition, i.e., Pearay L. Ogra (Professor Emeritus Jacobs School of Medicine and Biomedical Sciences, Uni versity at Buffalo, NY, USA); W. Allan Walker (Conrad Taff Professor of Nutrition [Emeritus], Professor of Pedictrics, Harvard Medical School); and Bo Lönnerdal (Distinguished Professor Emeritus, Department of Nutrition and Internal Medicine, University of California, Davis, CA, USA), have contributed to the Workshop program. The 94th NNI Workshop was designed with the goal to provide a compre hensive overview on the latest human milk research and its potential in modu lating mucosal immunity, the microbiome, and its impact on the neonate. The program was comprised of three sessions. Session I, led by Prof. Pearay L. Ogra, reviewed data on the immunology of milk and lactation. The flow of the topics brought us from a historical perspective to the latest scientific findings in order to understand the complex immunobiology of mammalian milk. Session II, di rected by Prof. W. Allan Walker, discussed the microbiology of human milk and lactation in detail, with a focus on premature infants and necrotizing entero colitis. The objective of the third and last session, shepherded by Prof. Bo Lönnerdal, was to shed light on the protective factors in human milk, e.g., hu man milk oligosaccharides, bioactive milk fat components, and lactoferrin, and their role in influencing the neonate’s immune system. The program brought important new insights but also raised unanswered questions. We are only beginning to understand the complex milk composition, functions of different bioactive components, and the most important role of each in human development. The 94th NNI Workshop and its book are dedicated to an extraordinaire per sonality: Prof. Lars A. Hanson (MD, PhD), who is considered the founder of modern milk immunology by the human milk research community. We believe that this publication and online material will be a great scientific support to all people seeking a deeper understanding of human milk and its immunological properties and enlarge the knowledge of those who have already specialized in human milk research. We would like to thank the three chairpersons Pearay L. Ogra, W. Allan Walker, and Bo Lönnerdal who designed the scientific program. Our special thanks also go to all speakers and scientific experts in the audi ence who contributed to the content of the workshop and scientific discussions. Finally, we thank Maria Elena Munoz and the NNI team for making this workshop possible. Dr. Natalia Wagemans Global Head, Nestlé Nutrition Institute Vevey, Switzerland

Foreword XI Published online: April 16, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp XII–XIV (DOI:10.1159/000505986)

Contributors

Chairpersons & Speakers Dr. Cem Aydogan Dr. Nicholas D. Embleton PhytoNet AG Population Health Sciences Institute Untere Paulistrasse 5 Newcastle University CH–8834 Schindellegi-Feusisberg c/o ward 35 Royal Victoria Infirmary Switzerland Newcastle upon Tyne NE1 4LP [email protected] UK [email protected] Prof. Janet E. Berrington Population Health Sciences Institute Prof. Leónides Fernández Newcastle University Department of Galenic Pharmacy and c/o ward 35 Royal Victoria Infirmary Food Technology Newcastle upon Tyne NE1 4LP Complutense University of Madrid UK Avenida Puerta de Hierro, s/n [email protected] ES–28040 Madrid Spain Prof. Lars Bode [email protected] Division of Neonatology and Gastroenterology and Nutrition Prof. Franz-Georg Hanisch Department of Pediatrics Institute of Biochemistry II School of Medicine Medical Faculty University of California, San Diego University of Cologne 9500 Gilman Drive Joseph Stelzmann Street 52 La Jolla, CA 92093 DE–50931 Cologne USA Germany [email protected] [email protected]

Prof. Vanessa P. Dunne-Castagna Prof. Olle Hernell Department of Nutrition Department of Clinical Sciences/ University of California Pediatrics 1221 Robert Mondavi Institute Umeå University Davis, CA 95616 SE–90185 Umeå USA Sweden [email protected] [email protected]

XII List of Contributors Dr. Jiří Hrdý Prof. Jiri Mestecky Department of General Immunology Department of Microbiology and Institute of Immunology and Medicine Microbiology University of Alabama at Birmingham 1st Faculty of Medicine 845 19th Street South Charles University Birmingham, AL 35294 Kateřinská 1660/32 USA CZ–12108 Prague [email protected] Czech Republic [email protected] Dr. David A. Mills Department of Food Science and Prof. Rulan Jiang Technology Department of Nutrition University of California University of California 3142 RMI North Building 3217 Meyer Hall One Shields Avenue One Shields Avenue Davis, CA 95616-5270 Davis, CA 95616 USA USA [email protected] [email protected] Prof. Sindhu Mohandas Dr. Miloslav Kverka Division of Infectious Diseases Institute of Microbiology Department of Pediatrics Czech Academy of Sciences University of Southern California and Vídeňská 1083 Children’s Hospital Los Angeles CZ–14220 Prague 4650 Sunset Blvd., MS#51 Czech Republic Los Angeles, CA 90027 [email protected] USA [email protected] Prof. Bo Lönnerdal Department of Nutrition Dr. Josef Neu University of California Division of Neonatology 3217 Meyer Hall Department of Pediatrics One Shields Avenue University of Florida College of Medicine Davis, CA 95616 1600 S.W. Archer Road USA Gainesville, FL 32610 [email protected] USA [email protected] Prof. Di Meng Mucosal Immunology and Biology Prof. Olav T. Oftedal Research Center Smithsonian Environmental Massachusetts General Hospital for Research Center Children 647 Contees Wharf Road 114 16th Street Edgewater, MD 21037 Charlestown, MA 02129 USA USA [email protected] [email protected]

List of Contributors XIII Prof. Pearay L. Ogra Prof. Niklas Timby Department of Pediatrics Department of Clinical Sciences/ Jacobs School of Medicine and Pediatrics Biomedical Sciences Umeå University University at Buffalo Norrlands universitetssjukhus State University of New York 10, S-plan 875 Ellicott Street SE–90339 Umeå Buffalo, NY 14203 Sweden USA [email protected] [email protected] Prof. Helena Tlaskalová-Hogenová Prof. Pia S. Pannaraj Institute of Microbiology Division of Infectious Diseases Czech Academy of Sciences Department of Pediatrics Vídeňská 1083 University of Southern California and CZ–14220 Prague Children’s Hospital Los Angeles Czech Republic 4650 Sunset Blvd., MS#51 [email protected] Los Angeles, CA 90027 USA Prof. Valérie Verhasselt [email protected] Breastfeeding, Growth, and Immune Health Team Dr. Samuli Rautava School of Molecular Sciences Department of Pediatrics and University of Western Australia Adolescent Medicine M310 University of Turku and Perth, WA 6009 Turku University Hospital Australia Kiinamyllynkatu 4–8 [email protected] FI–20520 Turku Finland Prof. W. Allan Walker [email protected] Mucosal Immunology and Biology Research Center Prof. Juan M. Rodriguez Massachusetts General Hospital for Department of Nutrition and Children Food Sciences 114 16th Street Complutense University of Madrid Boston, MA 02129 Avenida Puerta de Hierro s/n USA ES–28040 Madrid [email protected] Spain [email protected]

XIV List of Contributors Published online: April 14, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp XV–XVI (DOI:10.1159/000507237)

Dedication

Lars A. Hanson

This workshop and the proceedings are dedicated to our friend and mentor Lars A. Hanson, the father of modern milk immunology and human breastfeeding. His original contributions in the field include: demonstration of the secretory IgA antibody in human milk, which is different from IgA in serum; existence of an enteromammary axis; novel insights regarding the effects of milk on host defense in the neonate; immunodeficiency of IgA antibody; and the impact of breastfeeding on child health globally. For his outstanding contributions, Dr. Hanson has received numerous awards. These include: the Medal of the University of Helsinki, Finland (1975); Oscar Medin Prize, Stockholm, Sweden (1979); Robert Koch Prize, Bonn, West Germany (1981); Anders Jahre Prize in Medicine, Oslo, Norway (1988); Honor- ary Consul, Costa Rica (1988); Lifetime Achievement Award, The Jeffrey Mod- ell Foundation, New York, NY, USA (1990); Elander Prize, Göteborg, Sweden (1996); Founder Fellow of the Royal College of Pediatrics and Child Health, UK (1996) (Hon. FRCPCH); The Rosen von Rosenstein Medal, Swedish Pediatric Association and Swedish Society of Medicine, Uppsala (1999); Member, Norwe- gian Academy of Science, Norway, 1999; Medal of Distinction, City of Göteborg (2002); Nutricia International Award (2004); Macy-György Award (2004), and Amningshjälpens Bröstpris (2005). Dr. Hanson obtained his MD at the School of Medicine, University of Göte- borg, Sweden, in 1961 and defended his thesis for a doctoral degree in medical science in 1961 (PhD). He has published over 750 scientific papers in bacteriol- ogy, pediatrics, and immunology and has served as an editor, contributor, and author of numerous books. He has mentored 79 PhD students from virtually every part of the world. Dr. Lars Hanson was born in Naverstad, Sweden. He is married to Monika Tunback, an outstanding scholar and journalist. Thank you Nenne!

XVI Dedication Immunology of Milk and Lactation

Published online: March 10, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 1–10 (DOI:10.1159/000505577)

The Evolution of Lactation in Mammalian Species

Olav T. Oftedal Smithsonian Environmental Research Center, Edgewater, MD, USA

Abstract Lactation is a defining characteristic of all mammals, and, indeed, mammals draw their name from mammae, or mammary glands. The evolution of mammary glands has been the subject of debate since Charles Darwin. The purpose of this brief review is not to examine all past theories of mammary evolution but to consider the evolution of the mammary gland in relation to (1) modern paleobiology, giving special attention to the mammaliaforms which had many mammalian features, including delayed tooth development suggestive of milk intake. (2) Comparative aspects of mammary development in monotremes, marsupials, and eutherians, which reveal the close developmental relation of mammary glands to other skin glands and hair follicles. (3) The evolution of caseins, which are now known to derive from secretory calcium-binding phosphoproteins, which have a long history in regulating biomineralization. (4) The evolution of lipid secretion, and especially the evolutionary incorporation of immune system components (such as xanthine oxidoreductase) into the fat globule membrane. (5) The evolution of lactose synthesis, and especially the synthesis of the wide array of oligosaccharides found in some milks, including monotremes, marsupials, caniform carnivores, and humans. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

Lactation is highly complex and of ancient evolutionary origin [1]; thousands of mammary genes are involved, and their patterns of expression during mammary development and milk secretion are under study [2]. Mammalian milk is distinctive in that it contains unique milk-specific proteins (αs1-, αs2-, β-, and κ-caseins, β-lactoglobulin, α-lactalbumin, and whey acidic protein), specialized membrane-enclosed lipid droplets, and saccharides (both lactose and oligosaccharides) not found elsewhere in nature. The similarities in mammary development, major milk constituents, and secretory pathways across the main taxonomic groups (monotremes, marsupials, and eutherians) indicate that mammary glands and lactation were inherited from a premammalian ancestor in the Jurassic and/or Cretaceous. The mammary gland and its secretion represent a major evolutionary novelty without any known intermediates. In the mid-19th century, Charles Darwin devoted a chapter of the 1872 edition of On the Origin of Species to a discussion of the problems of evolutionary novelty, including the mammary gland. Since then, a variety of theories have been put forth about the evolution of the mammary gland, some by analogy to avian brood pouches, some in reference to bird, lizard, and monotreme eggs, a few based on fossil evidence, and one citing similarities between α-lactalbumin and lysozyme [reviewed in 1]. More recently, studies of the structure and function of milk constituents, genetic pathways by which they have evolved, and developmental mechanisms and pathways in mammary development have provided new information about the evolution of mammary glands and milk secretion [3, 4]. The material herein has been ab- stracted from Oftedal [4], wherein the topics are treated more extensively.

The Paleobiology of Lactation

Ironically, the evolution of milk is not so much a story about mammalian evolution but rather a story that was largely complete before mammals appeared on earth (Fig. 1). The first vertebrates to set foot on land in the late Devonian (ca. 365 million years ago [mya]) were the tetrapods, a group ancestral to all subsequent terrestrial vertebrates [5]. Their progressive reduction in dermal protection was presumably accompanied by more rigid, mat-like webs of collagen in the skin and the development of more elaborate skin glands that secreted mu- cous and toxic compounds to facilitate gas exchange and to protect against des- iccation, infection, and predation, as do the multicellular glands of living frogs, salamanders, and caecilians. Granular glands of living amphibians secrete

2 Oftedal Amphibians Monotremes Marsupials Eutherians Birds Squamates

(Adult) 66 MPSU MG Multiple Feather Fcale glands (Pouch young)

145 APSU (pelage) Prototherians Therians Mesozoic Cenozoic Mammals 201 Mammaliaforms Archosaurs

Ancestral Lepidosaurs Cynodonts APSU-MPSU 252 Therapsids

α-Keratins α-Keratins + β-keratins Permian299 Triassic Jurassic Cretaceous Tertiary “Pelycosaurs” Synapsids Sauropsids

Paleozoic Basal amniotes iferous

Carbon- α-Keratins 359 Basal tetrapods nian Devo- mya

Fig. 1. Evolution of mammary glands. A schematic of the evolution of mammary glands in comparison to other integumentary structures. The left vertical axis shows time (in million years, mya) at the transitions between geological periods. The basal amniotes diverged into the synapsids (leading to mammals) and sauropsids (leading to dinosaurs, birds, and crocodilians, for example) in the late Carboniferous; the sketch indicates the belief that basal amniotes had glandular skin. Note sequential radiations of synapsids: “pelycosaurs,” therapsids, cynodonts, and finally mammaliaforms, which were directly ancestral to mammals, and believed to have well-developed milk production. The top central shaded box includes schematic views of developing mammary glands in monotremes, marsupials, and eutherians. Note that the evolution of β-keratins in sauropsids (red line to right) played a key role in both scale and feather evolution (reproduced from Oftedal and Dhouailly [16] with permission).

at least 500 antimicrobial peptides [6]. Some milk constituents, such as α-lactalbumin, β-lactoglobulin, whey acidic protein, and proteins in the mammary fat globule membrane, may have originated from antimicrobial components in carboniferous tetrapod skin secretions [4]. Another potentially significant tetrapod feature is the parental care of eggs, including the provision of secretions to keep them moist. Among all 3 living amphibian lineages (salamanders, frogs, and caecilians), a terrestrial system of egg development has evolved, with large-yolked eggs and direct development into adult-type hatchlings without a larval stage. In such species, parental care is nearly universal [7] and may include provision of supplemental moisture to the

Evolution of Lactation 3 eggs via transcutaneous water movement on direct contact or via skin secretions [8]. Among some caecilians, a lactation-like pattern of care occurs in that maternal epidermis swells with lipid-containing material after the young hatch from direct developing eggs, and the hatchlings ingest skin and/or secretions using specialized teeth [9]. The amniotes, ancestral to both synapsids (including mammals) and sauropsids (including dinosaurs/birds and crocodilians, for example), first appeared in the late Carboniferous (i.e., mid-Pennsylvanian, about 310 mya; Fig. 1). The amniotic egg was a major evolutionary novelty characterized by a fibrous eggshell and a set of specialized extra-embryonic membranes that partitioned and enhanced egg physiologic functions [10], permitting eggs to become larger and to contain more yolk to support development. The parchment-like eggshell of these early amniotes lacked a calcified layer and thus remained highly permeable to water [11]. It is possible that these amniote eggs were dependent on moisture provided by the parents, just as in some terrestrial amphibians. I have argued that lactation originated as a secretion that provided water, and other secretory constituents, to eggs [11]. The synapsids subsequently underwent a series of extensive radiations and massive extinction events, so that most lineages disappeared in the late Triassic, but the mammaliaforms persisted and radiated in the late Triassic and Jurassic, and were ancestral to mammals about 160 mya. The mammaliaforms became more mam- mal-like, progressively smaller in size, with increased metabolic expenditure, growth rates, and activity [12]. The associated miniaturization of eggs implies hatching of altricial young, possibly fed on milk. There is evidence in the regional specialization of the extra-embryonic membranes in monotreme eggs that nutrient absorption could occur at the abembryonic end [11]. The mammaliaforms also delayed tooth eruption, developing simple “milk teeth” first, followed by adult den- tition, suggesting the young were fed early in postnatal life on milk. Thus, a plausible scenario is that lactation initially provided a “proto-lacteal” secretion containing moisture, antimicrobial constituents, and perhaps a few nutrients (such as calcium) to eggs. At some point, hatchlings began to ingest milk secretions including constituents with nutritional value that supplanted egg nutrients. Vitellogenins began to disappear about 170 mya [13], indicating the replacement of egg nutrients by milk nutrients.

Gland Origin and Mammary Evolution

It was the similarity in development, structure, and secretory processes that led to the hypothesis that apocrine glands may resemble the ancestral condition of mammary glands [1]. Both apocrine and mammary glands secrete by exocytosis

4 Oftedal secretory vesicles and by budding out and pinching off cellular contents with loss of cytoplasm [1]. In mammary glands, budding out and pinching off occurs during secretion of the milk fat globule (MFG); cytoplasmic crescents may be present but are minimal [14]. In most mammals, an apocrine gland on the general skin surface is typically associated with both a hair follicle and a sebaceous gland in a triad termed an apopilosebaceous unit. The development of the apopilosebaceous unit occurs in coordinated fashion, typically leaving the apocrine gland duct opening into the infundibulum of the hair follicle, such that secretion contacts the hair shaft. A parallel is found in monotremes (Fig. 1), in which mammary glands, hair follicles, and sebaceous glands form what can be termed a mammopilosebaceous unit (MPSU) [1]. The lactiferous ducts also open up into the infundibulum of an enlarged, specialized mammary hair [15]. The mammary glands in monotremes are organized into bilateral oval mammary patches consisting of 100– 200 MPSUs [15, 16]; there is no nipple. In earliest lactation, when monotreme eggs are incubated and hatched, the secreting mammary gland is still relatively small and tubular, and thus it has a superficial resemblance to an apocrine gland. There is also a developmental association of mammary glands with hair follicles and sebaceous glands in marsupials, such as koalas and kangaroos (Fig. 1). An oval primary primordium separates into nipple primordia which deepen into knobs and generate hair follicles (primary sprout), mammary glands (secondary sprout), and sebaceous glands (tertiary sprout). In the opos- sum, the nipple primordium develops into 8 MPSUs. Developing hair follicles penetrate the nipple epithelium and are then shed, and the mammary gland ducts come to occupy these nipple-penetrating cavities. In the adult marsupial, the “mammary hairs” are no longer evident, but the galactophores reflect their prior existence. In eutherian mammals, apocrine glands retain an association with hair follicles (apopilosebaceous units), but the association of mammary glands with hair follicles, the presumed ancestral condition, is lost in many taxa. That this is a secondary condition is suggested by two phenomena. First in some taxa (e.g., the horse), the mammary gland develops as an MPSU association, and the mammary hairs and sebaceous glands remain in lactating mares [16]. Second, bone morphogenetic proteins inhibit hair follicle formation during nipple development in the mouse, and transgenic overexpression of a bone morphogenetic protein antagonist converts nipple epithelium into pilosebaceous units. This may suggest an ancestral state involving MPSUs that was altered by bone morphogenetic protein production.

Evolution of Lactation 5 Origin and Evolution of Caseins

Caseins are unique to milk, convey a large proportion of the amino acids required by offspring, and form large micelles containing calcium phosphate nanoclusters. Multiple caseins (characterized as αs1-, αs2-, β-, and κ-caseins) participate in these micelles, with κ-casein stabilizing the micelle in secreted milk. Caseins are a primary transport vehicle for amino acids, calcium, and phosphorus. After ingestion, κ-casein is vulnerable to proteases such as chymosin, causing the release of a macropeptide from κ-casein and precipitation of caseins into a gastric curd which entraps fat. Fat so entrapped is attacked by multiple lipases; the curd caseins themselves are hydrolyzed by proteases. A mechanism to retain milk as a semisolid (gastric curd) was no doubt an important evolutionary step. All mammalian milks that have been studied contain the 4 primary types of caseins: αs1-, αs2-, β-, and κ-caseins, indicating a premammalian origin, although casein genes have been variously duplicated in diverse mammalian groups. The caseins are members of a much larger family of proteins of unfolded nature that are secreted from cells, usually in association with tissue mineralization or regulation of calcium at target tissues. These proteins, termed secretory calcium-binding phosphoproteins (SCPP), are secreted by secretory epithelial cells or cells derived from underlying ectomesenchymal cells, and have an ancient history in the evolution of mineralized vertebrate tissue [17]. The SCPPs include extracellular matrix proteins secreted by ameloblasts, odontoblasts, and osteoblasts as well as salivary proteins that bind and transport calcium. As unfolded proteins, all SCPPs are low in cystine disulfide bridges, and a subclass of proteins (P/Q-rich SCPPs), which includes caseins, are particularly rich in proline and glutamine [17]. None of the milk casein genes have been found in sauropsids, but 2 members of this SCPP gene cluster, ODAM (odontogenic, ameloblast-associated protein) and FDCSP (follicular dendritic cell secreted peptide), have been found in the genomes of a frog (Xenopus) and a lizard (Anolis), respectively [17]. Based on the relative locations and structures of exons of P/Q-rich SCPPs, as well as their phylogenetic distribution, Kawasaki et al. [17] proposed that the αs- and β-caseins derive via gene duplication and exon changes from an ancestral gene (CSN1/2) that derives from another SCPP gene, either ODAM or SCPPPQ1 (which itself derived from ODAM), while κ-casein derives from the SCPP gene FDCSP (which also derived from ODAM). If this scenario is correct, the ODAM gene is ultimately the grandmother gene of all caseins and played a central role in the evolution of synapsid reproduction. The initial function of an ancestral SCPP in a protolacteal secretion may have been to regulate calcium delivery to the surface of an egg and to prevent pre-

6 Oftedal cipitation of calcium phosphate on the eggshell [17]. This hypothesized role of ancestral casein(s) in delivering calcium to eggs is consistent with the view that early amniotes produced eggs with a fibrous calcium-free eggshell, that such eggs take up environmental calcium especially at the abembryonal end, and that the limited calcium supply in yolk would have made this beneficial [1, 11]. Sub- sequently, there was an expansion in the number and types of casein genes and their interaction in producing the large and complex casein micelle, which was essential to neonatal nutrition. This may have been one of the most important evolutionary novelties of lactation.

Origin and Evolution of the Milk Fat Globule

Milk varies tremendously in lipid content among mammals (from < 1% in rhinos and some lemurs to > 60% in some seals [18]). A unique feature of milk lipids is that they are packaged into membrane-enclosed structures known as MFGs. The core triacylglycerols in MFGs are bounded sequentially by a phospholipid monolayer, an inner protein coat, a bilayered phospholipid membrane, and a glycosylated surface [19]. Transmembrane proteins such as mucins, butyro philin, and CD36 interact both with proteins of the inner coat, such as xanthine oxidoreductase (XOR), fatty acid-binding protein, and adipophilin, and with molecules at the milk-facing surface [19]. The glycosylated surface is primarily due to oligosaccharide chains attached to outwardly projecting domains of transmembrane proteins. Of particular interest is that cytoplasmic crescents become trapped in the MFG during the process of secretion, possibly akin to apocrine-type secretion involving terminal protrusion in a narrow (a bleb) or wide (an endpiece) extension, which detaches via pinching off (narrow blebs), or via merging of exocytotic vesicles to create a gap (wide endpieces). From an evolutionary perspective, mammary glands had to minimize cytoplasmic loss since the volume of secretion was increased from micro- to milliliters or liters. The secretion of the MFG and its associated MFG membrane (MFGM) en- velope is a highly regulated process. In particular 2 proteins, butyrophilin and XOR, play an obligatory structural role in MFGM synthesis, and if they are reduced or eliminated from mouse mammary cells via knockout of the coding genes, mice fail to produce normal milk [19]. These 2 proteins have apparently been coopted from other cellular functions during the evolution of the mammary gland. The butyrophilin in milk (butyrophilin 1A1) is one of the butyro philin family of proteins. In butyrophilin 1A1 in the MFGM, one domain projects inward into the underlying protein coat where it binds XOR with high affinity. XOR is an unusual partner for butyrophilin, as it is best known for its role

Evolution of Lactation 7 in catalysis of the last 2 steps in uric acid formation and other enzymatic functions. XOR has particularly high expression in epithelial surfaces of the gastrointestinal tract, liver, kidney, lung, skin, and mammary gland [20]. Yet, upregulation and apical membrane localization of XOR in mammary epithelial cells during mammary gland development indicates a novel function for XOR in the MFGM. Given the antiquity of XOR and its long conservative evolutionary history, evolution of this new mammary function must be considered a radical de- parture. MFG secretion presumably evolved from some prior less productive forms of fat secretion, perhaps by tetrapod or synapsid skin glands. If these were apocrine- like glands, they would have had the biochemical pathways for apocrine secretion. One can imagine a scenario in which an ancestral apocrine secretion entailed the pinching off of apical blebs containing cytoplasm, secretory vesicles, and perhaps cytoplasmic lipid droplets, similar to the process described for some specialized apocrine glands such as human axillary apocrine glands, glands of Moll, cerumi- nous glands in the outer ear canal, and rodent harderian glands. When the blebs disintegrate in the gland lumen, the various constituents are released.

Origin and Evolution of Milk Sugar Synthesis

All mammalian milks contain at least traces of sugar [18]; in many eutherians the predominant sugar is lactose (galactose [β1–4] glucose), while in monotremes, marsupials, and caniform carnivores oligosaccharides predominate, most of which contain a galactose (β1–4) glucose unit at the reducing end [21, 22]. Both lactose and oligosaccharides with lactose at the reducing end appear to be unique to mammary secretion and thus entail the evolution of a novel synthetic pathway. Some marine mammals produce milk devoid of lactose and lactose-based oligosaccharides, but this is a secondary loss. The evolution of lactose synthesis is a remarkable example of a protein (or in this case 2 proteins) adopting completely new functions with minimal changes in structure. During lactose synthesis, a transmembrane protein in the trans- Golgi, β-1,4-galactosyltransferase 1 (β4gal-T1) binds UDP-galactose, producing a conformational change in β4gal-T1 that allows α-lactalbumin to be bound [23]. α-Lactalbumin binding to β4gal-T1 alters its specificity, allowing glucose to become the acceptor sugar for galactose transfer, resulting in the synthesis of lactose. α-Lactalbumin does not have other known functions, but β4gal-T1 does facilitate transfer of galactose from UDP-galactose to N-acetyl glucosamine at the terminus of N-linked oligosaccharide chains, which was presumably its ancestral function.

8 Oftedal Based on amino acid sequence similarity, three-dimensional structure, and the structure of the exons that code for α-lactalbumin, it is apparent that α-lactalbumin is most closely related to c-type lysozyme and is derived from it via gene duplication and base pair substitution [24, 25]. Lysozymes are hydrolytic enzymes that cleave the glycosidic bonds in peptidoglycans, the major bacterial cell wall polymer, and thus play a key role in innate immune defense systems in both vertebrates and invertebrates. Amino acid substitutions have led to the loss of hydrolytic function in α-lactalbumin. The estimated date of origin of α-lactalbumin from c-lysozyme is ancient, prior to the time of the split of synapsids from sauropsids about 310 mya [24]. Given that c-lysozyme is widespread, being a normal antimicrobial constituent of egg white and epithelial secretions, including amphibian skin secretions, it seems likely that an ancestral c-lysozyme, already present in secretions provided to eggs, was coopted as a modifier of carbohydrate secretion, generating lactose. One uncertainty is whether the original milk carbohydrate was lactose or an oligosaccharide. A low rate of lactose synthesis, coupled with high activity of glycosyl transferases that could glycosylate lactose, may have produced diverse oligosaccharides rather than free lactose, similar to what is observed in many extant monotremes and marsupials. In a study of echidnas, Oftedal et al. [26] suggested O-acetylated sialyllactose may serve to provide protection for pouch young against microbial attack, and that this may have been an ancestral function of early oligosaccharides. However, the evolutionary factors that have produced such a diverse array of oligosaccharides in monotremes, marsupials, caniform carnivores, and primates (especially humans) remain uncertain.

Disclosure Statement

The author declares no conflicts of interest.

References

1 Oftedal OT: The mammary gland and its origin 4 Oftedal OT: Origin and evolution of the major during synapsid evolution. J Mammary Gland constituents of milk; in Fox PF, McSweeney PLH

Biol 2002; 7: 225–252. (eds): Advanced Dairy Chemistry-1A. Proteins: 2 Lemay DG, Lynn DJ, Martin WF, et al: The bo- Basic Aspects, ed 4. New York, Springer, 2013, pp vine lactation genome: insights into the evolution 1–42.

of mammalian milk. Genome Biol 2009; 10:R43– 5 Carroll R: The Rise of Amphibians. 365 Million R43. Years of Evolution. Baltimore, Johns Hopkins 3 Oftedal OT: The evolution of milk secretion and University Press,2009.

its ancient origins. Animal 2012; 6: 355–368. 6 Jenssen H, Hamill P, Hancock REW: Peptide

antimicrobial agents. Clin Microbiol Rev 2006; 19: 491–511.

Evolution of Lactation 9 7 Duellman W, Trueb L: Biology of Amphibians. 19 Mather IH: Milk fat globule membrane; in Baltimore, Johns Hopkins University Press, 1994. Fuquay J, Fox PF, McSweeney PL (eds): Encyclo- 8 Taigen TL, Pough FH, Stewart MM: Water bal- pedia of Dairy Sciences. San Diego, Academic ance of terrestrial anuran (Eleutherodactylus Press, 2011, vol 3, pp 680–690. coqui) eggs: importance of parental care. Ecology 20 Garattini E, Mendel R, Romão MJ, et al: Mamma-

1984; 65: 248–255. lian molybdo-flavoenzymes, an expanding family 9 Kupfer A, Müller H, Antoniazzi MM, et al: Paren- of proteins: structure, genetics, regulation, func-

tal investment by skin feeding in a caecilian am- tion and pathophysiology. Biochem J 2003; 372:

phibian. Nature 2006; 440: 926–929. 15–32. 10 Packard MJ, Seymour RS: Evolution of the amni- 21 Urashima T, Asakuma S, Leo F, et al: The pre- ote egg; in Sumida SS, Martin KLM (eds): Amni- dominance of type 1 oligosaccharides is a feature

ote Origins: Completing the Transition to Land. specific to human breast milk. Adv Nutr 2012; 3: San Diego, Academic Press, 1997, pp 265–290. 473S–482S. 11 Oftedal OT: The origin of lactation as a water 22 Urashima T, Messer M, Oftedal OT: Comparative source for parchment-shelled eggs. J Mammary biochemistry and evolution of milk oligosaccha-

Gland Biol 2002; 7: 253–266. rides of monotremes, marsupials and eutherians; 12 Kemp TS: The Origin and Evolution of Mam- in Pontarotti P (ed): Evolutionary Biology: Ge- mals. New York, Oxford University Press, 2005. nome Evolution, Speciation, Coevolution and 13 Brawand D, Wahli W, Kaessmann H: Loss of egg Origin of Life. Cham, Springer, 2014, pp 3–33. yolk genes in mammals and the origin of lacta- 23 Ramakrishnan B, Qasba PK: Crystal structure of

tion and placentation. PLoS Biol 2008; 6:e63. lactose synthase reveals a large conformational 14 Mather IH, Keenan TW: Origin and secretion of change in its catalytic component, the β-1,4-

milk lipids. J Mammary Gland Biol 1998; 3: 259– galactosyltransferase-I. J Mol Biol 2001; 310: 205– 273. 218. 15 Griffiths M: Biology of the Monotremes. New 24 Prager EM, Wilson AC: Ancient origin of lactal- York, Academic Press, 1978. bumin from lysozyme: analysis of DNA and ami-

16 Oftedal OT, Dhouailly D: Evo-devo of the mam- no acid sequences. J Mol Evol 1988; 27: 326–335.

mary gland. J Mammary Gland Biol 2013; 18: 25 Qasba PK, Kumar S: Molecular divergence of ly- 105–120. sozymes and α-lactalbumin. Crit Rev Biochem

17 Kawasaki K, Lafont A, Sire J: The evolution of Mol 1997; 32: 255–306. milk casein genes from tooth genes before the 26 Oftedal OT, Nicol SC, Davies NW, et al: Can an

origin of mammals. Mol Biol Evol 2011; 28: 2053– ancestral condition for milk oligosaccharides be 2061. determined? Evidence from the Tasmanian 18 Oftedal OT, Iverson SJ: Comparative analysis of echidna (Tachyglossus aculeatus setosus). Gly-

non-human milks. A. Phylogenetic variation in cobiology 2014; 24: 826–839. the gross composition of milks; in Jensen RG (ed): Handbook of Milk Composition. San Diego, Academic Press, 1995, pp 749–789.

10 Oftedal Immunology of Milk and Lactation

Immunology of Human Milk and Lactation: Historical Overview

Pearay L. Ogra Department of Pediatrics, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo, State University of New York, Buffalo, NY, USA

Abstract The development of the mammary glands and the process of lactation is an integral component of mammalian evolution, and suckling has been essential for the survival of the neonates of most mammalian species. The colostrum and milk, the major products of lactation, contain a wealth of biologically active products derived from the immunologic and microbiological experiences in the maternal circulation and in the maternal mucosal surfaces. These include major immunoglobulin isotypes in the maternal circulation, secretory IgA, a variety of soluble proteins, casein, nutritional components, hormones, a large number of cellular elements and their secreted functional products (cytokines and chemokines), several peptides, lipids, polysaccharides and oligosaccharides, and a diverse spectrum of microorganisms. During the past few decades, significant new information has become available about the evolutionary biology of mammalian lactation, the functional characterization of antibody and cellular immunologic products, the role of oligosaccharides and other proteins and peptides, and about the distribution and biologic functions of the microbiome observed in human products of lactation. This workshop explores this information in some detail in a series of presentations. A brief overview of the earlier observations on the immunologic aspects of lactation is presented here, and detailed reviews of more recent observations are reported in subsequent presentations in this workshop. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel

This review is dedicated to the late Dr. S.S. Ogra. She was instrumental as the principal investigator in the conduct of all milk-related investigations carried out in our program. She was a wonderful friend and my lifetime partner. She passed away in 1992 at 55 years of age. O, those with a beautiful face, may the child reared on your milk attain a long life, like the gods made immortal with drinks of nectar. Sushruta (600 BC)

Imagine that the world had created a new “dream product” to feed and immunize everyone born on earth. Imagine also that it was available everywhere, required no storage or delivery and helped mothers plan their families and reduce the risk of cancer. Then imagine that the world refused to use it. Frank Oski (1932–1996)

Introduction

These two quotes reflect the depth of human interest in breastfeeding for over 2,500 years [1, 2]. Milk and other lactational products of all mammals, including humans, have been associated with unique healing powers and beneficial effects. Mother’s milk has been considered a complete food for the infant in many ancient scriptures. With the evolution of agricultural civilization, long before the development of commercial milk formula foods, milks from buffalo, cow, sheep, camel, donkey, horse, elephant, and goat were highly recommended for the treatment of insomnia, loss of appetite, ascites, piles, infestations by worms, skin disorders, muscle weakness, dysfunctions of sexual activity, and a large variety of other human ailments [3]. In addition, breastfeeding also developed a religious and spiritual importance in the Middle Ages in Europe, as evidenced by the deep faith and respect for Nursing Madonna, the Virgin Mary breastfeeding the infant Jesus [4]. During the upsurge of Marian theology in Europe, milk was viewed as the processed blood, and the milk of the Virgin paralleled the role of the blood of Christ [5]. This is best exemplified by “the miracle of the lactation of St. Bernard,” based on a vision concerning St. Ber- nard of Clairvaux in France being hit with a squirt of milk traveling an impressive distance from the breast in the statue of the Virgin nursing the infant Jesus [6]. Such blessed milk is believed to have given him great wisdom and cured an infection in his eye. The modern history of immunology of mammalian lactation can be traced to as early as 1892 with observations by Paul Ehrlich that nonsuckling frequently resulted in death in the newborn foals, lambs, or piglets [7–9]. About the same time, observations by Escherich [10] provided for the first time evidence for ex- quisite sensitivity of intestinal microflora to human milk. Subsequently, the association of certain milk proteins such as immune lactoglobulin with specific

12 Ogra Table 1. Placentation of mammalian species and transport of maternal products to the fetus or neonate via the placenta or breast milk [13]

Mammalian Immunoglobulin transport to group the fetus via the neonate via the placenta the mammary gland

1 Man IgA (+++) Rabbit IgG (++) IgG (++) IgM (+) 2 Rodents IgA (+++) Carnivores IgG (+) IgG (++) IgM (+) 3 Horse IgA (+) Cattle IgG (+++) IgG (+++) Sheep IgM (+) Swine

Magnitude and transport from minimal (+) to maximal (+++). protective functions was identified by Smith [11]. This protein is now referred to as the immunoglobulin G in the bovine colostrum. Furthermore, studies by Dixon et al. [12] demonstrated that IgG in the colostrum is actively transported to the colostrum from maternal serum during lactation. It is now well known that newborns of many mammalian species will die from infections if they fail to receive lactational products adequately via breastfeeding or the neonate fails to suckle. Interestingly, however, human neonates can survive without any breast milk feeding, and they can be raised normally on nonhuman milk-based formula feeds. It has been shown that during placentation of mammals, different maternal immunologic components are transported selectively via the placenta or breastfeeding in different primates. For example, in rabbits, rodents, and some carnivores, maternal IgG is actively transported to the fetus in large amounts from the maternal serum across the placenta. On the other hand, such effective pla- cental transport does not occur in horses, cattle, swine, and other mammals, as reviewed in detail by Butler and Kehrli [13] and summarized briefly in Table 1. The immunologic composition of human milk and its biologic linkage to mucosal immunity was initially recognized by the identification of major classes of immunoglobulin in the milk by Gugler and von Muralt [14], and Hanson [15]. These elegant studies were followed by the identification of the unique immunoglobulin, the secretory IgA (SIgA), in human milk by Hanson and Johans- son [16]. Subsequent studies by Beer et al. [17], Ogra and Ogra [18], Ogra et al. [19], Mohr [20], and Okamoto et al. [21] identified several cellular and soluble

Immunology of Human Milk and Lactation 13 immunologic elements in the human milk and their possible transport to the suckling neonate via the process of breastfeeding. Finally, it is of interest that Ehrlich [7] demonstrated for the first time that maternal immunization and subsequent breastfeeding induced significant protection in suckling mice against the toxic effects of subsequently ingested ricin and abrin. His imaginative studies also raised the possibility of protection against infections such as syphilis, mumps, typhus, and measles via the process of breastfeeding [7–9]. During the past 3 decades, significant information has been obtained to suggest that the immunologic activity inherent in the products of lactation represents, to a major extent, the effector functional elements of the common mucosal immune system. Following the discovery of IgA in the serum by Heremans et al. [22], and of secretory IgA in the milk by Hanson and Johansson [16], the presence of SIgA was also demonstrated in other mucosal secretions by Chodirkar and Tomasi [23], Tomasi and Zigelbaum [24], and Bienenstock and Tomasi [25]. These observations were followed by the identification of antibacterial, antiviral, and antiparasitic activity in the milk associated with SIgA and other immunoglobulin classes, demonstration of a number of specific cellular elements and cell- mediated immune responses, and detection of cytokines and other immunoreg- ulatory factors in milk. The relationship of the immunologic activity in the milk and mammary glands to other mucosal surfaces was documented conclusively by several elegant studies which identified intestinal and respiratory tracts, and the sublingual tissues as the primary induction sites of specific IgA-committed B cells and their active migration to the mammary glands [26–29]. Since the discovery of Bifidobacterium bifidum subspecies in 1953, it is now clear that this organism predominates in the feces of the breastfed infants. Specific factors stimulating the growth of this organism are uniquely present in human but not in cow’s milk. A significant biologic database is now available to support the clinical observations dating back from antiquity to the last few centuries, which have suggested a strong association between breastfeeding and protection against a variety of infectious and noninfectious disease processes in humans. These include protection against infectious diarrheal diseases, fertility and childbearing, and immunomodulation of mucosal and systemic immune responses. This information has been extensively reviewed in many recent publications [30–34]. It is now clear that human milk contains a wealth of biologically active products. These include soluble proteins, hormones, a number of cellular elements and their functional products (cytokines, chemokines, and hormones), several peptides, proteins, lipids, oligosaccharides, and numerous microorganisms [32]. This information is briefly reviewed in Table 2. Significant new information has also become available about the following specific areas of milk: (a) evolution of lactation in mammalian species; (b) further characterization of immunologic

14 Ogra Table 2. Biologically active products in human milk [32]

Nutritional Immunologic products Microbiological products products soluble cellular bacteria viruses other

Bioactive proteins Immunoglobulins and peptides SIgA, Stable Vaccine Fungi IgA1, IgA2 Leukocytes microbiome associated Parasites Carrier proteins IgG Epithelial (physiological) IgM, SIgM cells Pathogenic Lipids, fat globules IgD Probiotics viruses Casein Free secretory Macrophages Glucosamine component Stem cells Pathogenic Retroviruses bacteria Electrolytes Lactoferrin B cells and Trace metals Lysozyme subsets Minerals α-Lactalbumin T cells and Carbohydrates Oligosaccharides subsets Oligosaccharides Glycoconjugates

Nucleotides Lipids and fat Vitamins globules

Albumin Cytokines Globulin Soluble CD14

Defensins

Hormones

Antisecretory factor

Nucleotides

components of human milk; (c) identification and functional characterization of the mucosal microbiome and its role in modulating the homeostasis of human biologic functions; (d) development and functional aspects of the microbiome and virome of human colostrum and milk, and (e) the role of milk oligosaccharides, other milk proteins, and peptides in the mechanism of protection induced by breastfeeding and ingestion of human milk. This workshop was designed to review this information in some detail. Re- cent information about the microbiological aspects of milk and lactation and its influences on the microbial colonization in the gut, and of other unique protective factors in the milk, is provided in the subsequent presentations in this workshop. A brief overview of the recent progress and highlights of earlier observations on the immunologic aspects of human milk are summarized below.

Immunology of Human Milk and Lactation 15 Table 3. Immunoglobulin levels in colostrum and milk at different periods and estimated delivery to the breastfed infant during lactation [37]

Postpartum Total protein by Total output of period, days immunoglobulin class, % immunoglobulin, mg/24 h IgG IgM IgA IgG IgM IgA

1 7 3 80 80 120 11,000 3 10 45 45 50 40 2,000 7 1–2 4 20 25 10 1,000 8–28 1–2 2 10 10 10 1,000 29–50 1–2 0.5–1.0 10 10 10 1,000

Immunology of Milk and Lactation

Lactational Performance: Secretion of Colostrum and Milk. The immunologically active products observed in the mammary glands and the products of lactation are (a) derived from local synthesis within the mammary glands, (b) secondary to the transport of components from the maternal circulation and blood stream, and (c) selective and specific transport from the induction sites in the mucosa-associated lymphoid tissue in the gastrointestinal tract, nasopharynx, sublingual lymphoid tissue, and lymphoid tissue in bronchoepi- thelial mucosal sites. Colostrum is the first postpartum product of lactation. It is dense in protein and fat, and it contains the highest amounts of soluble as well as cellular immunologic components compared to transitional or mature milk. Successful lactation with continued contribution of the transported or locally synthesized products in mature milk is also determined by continued contribution of neural, endocrine, and other maternal-infant interactions activated at the time of delivery. Early and frequent breast contact by the nursing infant is also important for continued stimulation of neural pathways to maintain prolactin and oxytocin release. Lactation often ceases when suckling stops.

Soluble Components Immunoglobulin. The major isotypes of immunoglobulin in the milk or colostrum is the 11SIgA. Other isotypes including 7SIgA, IgG, IgM, and IgD are also present in varying amounts, and IgE may be occasionally observed. The 11SIgA exists as a dimer of two 7SIgA molecules linked together by J-chain, a polypep- tide chain associated with a secretory component, and a polymeric immunoglobulin receptor [35]. Immunoglobulin Activity. The studies on the temporal distribution of class- specific immunoglobulins and different cellular components in the product of

16 Ogra lactation by S. Ogra and her colleagues [18, 19, 21, 30] in the late 1970s established a sequential database for their levels in the colostrum and mature milk. Highest levels of SIgA and IgM are observed during the first 3–5 days of lactation. The levels of IgA are usually 4–5 times higher than those of IgM and about 26–30 times higher than IgG levels [36]. As lactation progresses, the levels of IgA and IgM in the mature milk decline rapidly. However, this decline is compensated by the increase in the total volume of milk produced (Table 3). It is estimated that a fully breastfed neonate may consistently receive about 1 g of IgA each day and approximately 1% of this amount for IgM and IgG [36]. Comparative studies of immunoglobulin activity in the feces of breastfed infants have suggested that the fecal content of IgA may be 15–20 fold higher after human milk feeding compared to bovine IgG after feeding of bovine immunoglobulin products [37]. Immunoglobulin Reactivity. Studies carried out with many microorganisms, including respiratory syncytial virus, Escherichia coli 083 and Streptococcus pneumoniae, have demonstrated the appearance of antibodies in the milk IgA independent of, and even in the absence of, any detected antibody activity in the serum. These and other elegant studies have clearly demonstrated that SIgA and its functional activity is derived from the initial exposure to specific antigens in the neonatal respiratory and intestinal mucosal lymphoid tissue and the transport of specific antibody producing IgA B cells to the mammary glands. These observations have firmly established the concept of the enteromammary and the bronchomam- mary axis of immunologic reactivity in the mammary glands [28, 38, 39]. Antisecretory Factor. An important molecule which binds like a lectin to certain polysaccharides (and inhibits gut fluid secretion induced by cholera toxin) has been described by Lönnroth and Lange [40] in 1984. This hormone-like factor is produced in many tissues, including the mammary glands. It has been found in samples of human milk in poor socioeconomic settings. It is possibly induced by exposure to enterotoxin-producing bacteria. The antisecretory factor seems to have a protective anti-inflammatory effect by reducing fluid secretion, and it has been used to treat infectious diarrheas, anti- inflammatory bowel disease, and other inflammatory conditions [34]. Hu- man milk is also rich in other anti-inflammatory components. These include vitamins, especially A, C, E, as well as the enzymes catalase and glutathione peroxidase [32]. Soluble CD14 and Soluble Toll-Like Receptor. Colostrum and milk contain high concentrations of soluble CD14. This molecule helps lipopolysaccharides, a surface structure on gram-negative bacteria, to bind to TLR-4 receptor and activate phagocytes [41]. It appears that by means of milk CD14, phagocytes in the gut mucosa are activated by gram-negative as well as gram-positive organisms [34]. It has also been suggested that CD14 promotes differentiation and

Immunology of Human Milk and Lactation 17 expansion of B cells and the anti-inflammatory activity of lactoferrin, another important nonbinding protein found in human milk.

Cytokines. A large number (> 150) of cytokines, growth factors, and chemokines have been identified to date. These soluble complex molecules play a critical immunomodulating role in (1) stimulation of growth, differentiation, and immunoglobulin products of B cells, (2) enhancement of thymocyte proliferation, (3) inhibition of IL-2 production by T cells, and (4) suppression of IgE production. Cytokines such as IL-1β, TNF-α, IL-10, and TGF-β have been demonstrated in human milk and appear to be secreted by milk macrophages as well as by the epithelium of the mammary glands. In addition, IL-6 and interferon-α have also been described in human milk. Chemokines. They are small cytokines with selective and discrete target cell functions. They are classified based on spacing between the cysteine residues (CXC, CC). Some chemokines, including IL-8, have been detected in human milk. Most chemokines serve as signals to mobilize phagocytic cells to an area of inflammation, and such activated phagocytic cells produce more cytokines. Growth Factors. Finally, several growth factors, including IL-7, have been recovered from human milk [32]. It has been suggested that IL-7, which is a growth factor for T-cell progenitors for memory T cells, also supports thymic growth [42].

Cellular Components Human colostrum and milk are endowed with a variety of maternally derived cellular elements. These include epithelial cells, activated neutrophils, macrophages, stem cells, and B and T lymphocytes. In addition, recent observations have suggested that human milk is rich in bacteria and other cellular and subcel- lular living organisms [32]. 6 Leukocytes. There are > 1–3 × 10 leukocytes/mL in colostrum and early milk. 6 The number of such cells decreases gradually to < 1 × 10 /mL over the next 2–3 months of lactation [18]. Flow-cytometric analysis has indicated that lymphocytes constitute about 4% of the cell pool in the milk. It has been shown that by 4–6 months into lactation, epithelial cells constitute about 75–80% of milk cells [43]. Polymorphonuclear neutrophils represent about 40–50% of the total cell population in early colostrum [43]. In lactating mothers, the development of infection is often associated with an increase in milk leukocytes, and their numbers decline with resolution of infection. Experimental animal model data suggest that milk leukocytes cross the intestinal epithelial barrier in the infant and may en- graft in different organs, including mesenteric lymph nodes, liver, and spleen. The leukocytes may provide anti-infective benefits to the mammary gland of the lactating mother and possibly also to the breastfeeding infant [44–46].

18 Ogra Macrophages. Human milk macrophages exhibit several phenotypic and functional characteristics, and they are able to produce many cytokines, such as IL-1β, IL-6, TNF-β, and GM-CSF, spontaneously. Milk macrophages are capa- ble of antigen presentation and specific synthesis of prostaglandin E2, plasminogen activator, lysozyme, and C3 [47]. Milk macrophages have been shown to express the marker DC-SIGN which is a dendritic cell receptor for human immunodeficiency virus (HIV). As a result, it has been proposed that unstimulated milk macrophages may add to the risk of HIV transfer via the breast milk [48]. Activated milk macrophages exhibit enhanced phagocytosis and have been found to have a receptor for SIgA. Its role in the mechanism of protection in the breastfeeding neonate remains to be defined. Epithelial Cells. Human colostrum and milk contain both ductal and alveolar epithelial cells and myoepithelial cells. The majority of these cells are viable and can be propagated in vitro. It has been proposed that milk epithelial cells are de- tached from the stroma of mammary glands in an active process driven by differential gene expression. These cells appear in clusters and may serve a still to be defined useful biologic purpose [32]. Stem Cells. Some cells in human colostrum and milk have been found to express stem cell markers, CK5, nestin, P63, and CD498. Milk cells expressing such markers were able to self-replicate and differentiate into luminal or myoepithelial cells. It has also been observed that such milk stem cells express pluripo- tency markers and may differentiate into cells of other lineages, including adi- pocytes, hepatocytes, and pancreatic β-cells. The implication of these findings to the neonatal immune system and cellular homeostasis remain to be defined [32]. Lymphocytes and Cell-Mediated Immunity. The overall number of lymphocytes in the human colostrum and milk is relatively small. Of the 1–3 × 106 leukocytes/mL of colostrum during the first few days after birth, lymphocytes account for about 4–6% of total cells [20]. The B cells account for about 6% of the lymphocytes, and the bulk of the remaining cells are of T-lymphocyte phenotype (83%), with an additional small number of natural killer cells (8–10%). T Cells. Most T cells are CD8+ and CD4+ cells, with a relative predominance of CD8+ cells. Most cells are CD45RO+, a marker associated with activation and immunologic memory. The T cells are mainly αβ-receptor-bearing cells. Vδ1 and Vγ2 (but not Vδ2) marker-bearing cells are significantly overrepresented in milk T cells, possibly due to direct homing to the mammary glands. Milk lymphocytes share the CCR9 receptor from the thymus-expressed chemokine (TECK) in the epithelium, with all small intestinal CD4+ and CD8+ cells, and with other similar cells from tonsils, lungs, skin, synovium, and body fluids [49– 52]. Although the precise function of such cells in lactational products and

Immunology of Human Milk and Lactation 19 mammary glands remains to be defined, they may represent an important mechanism specifically targeting the immune system. Interaction of Cellular Elements with the Neonatal Immune System. In view of the large number and the diversity of cellular elements observed in the colostrum and milk and the large volume of milk consumed by the human infant during the neonatal period and early infancy, it is reasonable to postulate that these cells exert possible functions in the breastfeeding infant. During the course of normal breastfeeding, millions of viable cellular elements are ingested by the infant. Earlier studies have suggested that milk cells are taken up in the neonatal mucosal surface and may transfer to the neonate, with varying degrees of specific immunologic information [17, 19, 20, 53]. Studies carried out by Ogra et al. [19] in 1977 in a group of formula-fed infants after a single-feed administration of human colostrum have clearly demonstrated the uptake and transfer of IgA. In another group of infants of tuberculin-positive mothers, these investigators also demonstrated the transfer of in vitro correlates of T cells mediated against tuberculin after prolonged breastfeeding. The tuberculosis-specific T-cell response in such infants were short lived and undetectable after 10–12 weeks in spite of continued breastfeeding. Subse- quently, several experimental animal studies have demonstrated the engraftment of maternal DNA via milk leukocytes in infant tissue [44–46]. Although these observations have been linked to lymphocytes, such transfer may also occur with epithelial and stem cells. In other experimental investigations, the transfer of maternal T cells and HLA antigens appears to be associated with the development of immunologic tolerance to maternal HLA antigens. Breastfed infants also express a lower frequency of precursors of cytotoxic T lymphocytes reacting with maternal HLA than nonbreastfed infants [54, 55]. Recent observations have also shown that maternal cytotoxic T lymphocytes localize in the Peyer’s patches of breastfed infants [56]. Such localization may serve to compensate for the immature adaptive immune response functions in the neonate.

Risks and Benefits of Breastfeeding

In his keynote, Prof. Oftedal elegantly outlined the evolutionary biology of the mammary gland and lactation. This presentation also implied that the origin of the mammary gland is buried deep in time, and many of its evolutionary novelties and specific nutritional products, such as caseins and other milk-specific proteins, and the methods of sugar synthesis appear to have originated more than 300 million years ago [57]. It is clear that the evolution of the mammary gland and lactation has become an important characteristic of mammalian reproduction, and an essential mech-

20 Ogra Table 4. Benefits and possible risks associated with human breastfeeding [32]

Benefits Risks noninfectious infectious

Maternal benefits Reduced fertility Maternal sepsis Delayed conception Mastitis Improved contraception Breast abscess

Infant benefits Maturation of the mucosal Insufficient milk syndrome Active mycobacterial immune system and MALT Nutrient deficiency tuberculosis Gastrointestinal homeostasis Maternal metabolic disease Viral infections Gut flora colonization Galactosemia HBV, HCV Host gene expression Phenylketonuria HIV HTLV1, HTLV2 Reduction in Breast milk-associated HSV Infant mortality hyperbilirubinemia CMV Neonatal sepsis Sudden infant death syndrome Maternal medications Infections of the respiratory tract Chemicals Otitis media Radionucleotides Urinary tract infections Opioids Necrotizing enterocolitis Other Asthma, atopy, environmental allergies Metabolic syndrome, obesity, diabetes Malignancy Ischemic heart disease

Improved cognitive function

anism of defense and survival for the mammalian neonate. As a result, the relative benefits and risks of breastfeeding must be considered with regard to the immunologic, nutritional, and microbiological components of the products of lactation identified to date. This workshop was specifically designed to examine the recent progress made in these areas of lactation and maternal-neonatal interactions in some detail. The functional elements of the immunology of lactation and its benefits and risks are briefly summarized below. The contributions of the milk microbiome and virome and of nutritional and other protective factors in milk and their role in the mechanisms of host defense in the breastfed neonate are discussed in some detail in the presentations in sections II and III, respectively. The known benefits and potential risks of breastfeeding have been reviewed in extensive detail by Hanson [34] in a special publication in 2004 and more recently by Kim et al. [32] in 2015. An overview of these effects on the breastfeeding neonate as well as on the lactating mother is presented in Table 4.

Immunology of Human Milk and Lactation 21 Maternal Benefits Although lactation is designed to mobilize the best maternal attributes for the homeostasis and the survival of the infant, breastfeeding offers significant benefits to the mother as well. Feeding of about 6 sucklings per 24-h period has been found to provide significant contraceptive benefit for the mother. It has been proposed that more conceptions are prevented by breastfeeding than by all other contraceptive approaches and family planning programs in many parts of the world [58]. Such an important maternal influence has also been shown to contribute to the reduction in infant mortality correlated with reduced crowding in the family, with less risk of infection and improved availability of food and nutrition to the infant and the mother. An interpregnancy space of less than 2 years has been shown to increase the risk of infant mortality by over 50% before 5 years of age [59].

Neonatal Benefits Immunologic benefits for the neonate associated with breastfeeding include maturation of the mucosal immune system. The development of intestinal mucosal integrity is to a large extent determined by the maturation of mucosa- associated lymphoid tissue and other tissue sites, and the establishment of the mucosal microbiome. Recent investigations have demonstrated that SIgA antibody and other soluble immunologic products in the breast milk promote long- term gut homeostasis by regulating the acquisition of the mucosal microbiome and host gene expression [60, 61]. Exclusive breastfeeding in the first 6 months is clearly a major determinant of the prevention of diarrheal disease in infants, especially with Escherichia coli, Shigella, Vibrio cholera, Campylobacter, some parasitic infestations (Giardia lamblia), viruses (rotavirus), and possibly other mucosal infections. Case-control studies have suggested that breastfeeding and specific antibody activity in the colostrum and milk more often provide protection against severe disease and hospitalization rather than total prevention of colonization and infection [60, 61]. A number of studies have clearly demonstrated a beneficial role of human milk in preventing or modifying the severity of necrotizing enterocolitis in premature infants [32, 34]. Similarly, anti-infection benefits related to milk-associated antibodies, soluble cytokines, or other protective features have been observed in breastfed infants against several genitourinary respiratory infections, otitis media in childhood, neonatal sepsis, and possibly sudden infant death syndrome of unexplained origin [34]. Breastfeeding and the immunologic components of human milk have been shown to confer long-lived protection against reactive airway disease and bron- chial asthma, eczema, and other atopic and allergic states. The protective effects

22 Ogra may reflect multiple synergistic mechanisms, including maturation of gut and airway mucosa by growth factors in human milk, reduction in the absorption of allergens and other antigens by modulation of the mucosal microbiome, and induction of specific mucosal tolerance (oral tolerance), and immune exclu- sions. Combined with other protective factors in the gut, SIgA can impede allergen sensitization by blocking the transport of foreign macromolecules across the immature neonatal gut epithelia and modulating the development of specific antibodies or immune complexes. It should also be pointed out that cow’s milk protein and other food antigens ingested by the lactating mothers have been observed on colostrum and milk. Other studies have suggested that early breastfeeding may be associated with decreased serum antibody responses to cow milk proteins and other maternal dietary antigens in the breastfed infant. From an evolutionary biology perspective, it would seem that breastfeeding should provide only beneficial effects to the neonate. However, there is considerable debate regarding the protective “immune-mediated effects” of breastfeeding on the development of atopy and allergy. Some investigations have proposed that the act of breastfeeding itself, regardless of the constituents of the breast milk, may be more or equally important defense mechanisms for the infant. An interesting study has suggested no protective effect of indirect breastfeeding (breast milk fed by the bottle) compared to infants receiving direct breastfeeding [62].

Neonatal Risks Because of changing mammalian and (especially) human environments, there are now several documented risks in breastfeeding. However, these risks are not immunologic in nature; they are rather related to the maternal use of chemicals, radionucleotides, and other medications or drugs, maternal metabolic disorders, and possibly hyperbilirubinemia associated with breastfeeding. A number of infectious agents have been recovered from human colostrum and milk, and the transmission of infection to the neonate has been observed frequently. How- ever, the development of disease associated with such transmission is relatively infrequent. Infection in the breastfeeding infant has, however, been observed with active maternal Mycobacterium tuberculosis infection, active maternal sepsis or mastitis, and active infection with HIV, HBV, HSV, and CMV (Table 4). Existing information about the potential risks and benefits of the milk microbiome and other factors in human colostrum and milk will be explored in further detail in subsequent studies of this workshop. In conclusion, it may be worthwhile to recapitulate that from an evolutionary standpoint, human colostrum and milk continue to remain the single most im-

Immunology of Human Milk and Lactation 23 portant vehicle for the transport of all maternal immunologic experiences via breastfeeding to the neonate for its survival and well-being throughout its life span.

Acknowledgments

This historical overview of the immunology of milk and lactation and the select laboratory data summarized here are largely based on the investigations carried out by the following collaborators and co-authors from 1970 to 2000: Drs. Y. Chiba; J. Cumella; L. Duffy; J. Freihorst; M. Fishaut; R. Garofalo; G. Losonsky; S.S. Ogra; Y. Okamoto; and D. Wong The secretarial assistance of Mrs. Judith Maurino, my colleague and my secretarial support for almost 5 decades, in the preparation of this manuscript is gratefully acknowledged.

Disclosure Statement

The author declares to have no conflict of interest.

References

1 Sushruta: the father of Indian surgery and oph- 10 Escherich T: The intestinal bacteria of the neo- thalmology. https://www.ncbi,nlm.nih.gov/m/ nate and breast-fed infant. 1884. Rev Infect Dis

pubmed/9476614. 1988; 10: 1220–1225. 2 Oski F: Frank Oski quotes. https://www.azquotes. 11 Smith EL: The immune proteins of bovine colos-

com/author/56278-Frank_Oski. trum and plasma. J Biol Chem 1946; 164: 345–358. 3 Athavale B: Bala-Veda Pediatrics and Ayurveda 12 Dixon FJ, Weigle WO, Vazques JJ: Metabolism Proceedings of the XV International Congress of and mammary secretion of serum proteins in the

Pediatrics, New Delhi. Bombay, Shilp Associates, cow. Lab Invest 1961; 10: 216–237. 1977, pp 1–190. 13 Butler JE, Kehrli ME Jr: Immunoglobulins and 4 Tradigo A: Icons and Saints of the Eastern Ortho- immunocytes in the mammary gland and its se- dox Church. A Guide to Imagery. Los Angeles, cretions; in Mestecky J, Lamm ME, Ogra PL, et al Getty Publications, 2006. (eds): Mucosal Immunology, ed 3. San Diego, 5 Saxon E: The Eucharist in Romanesque France: Academic Press, 2005, pp 1763–1793. Iconography and Theology. Woodbridge, Boydell 14 Gugler E, von Muralt G: Immuno-electrophoretic Press, 2006, pp 205–207. studies on human milk proteins (in German).

6 Dewez L, van Iterson A: La lactation de Saint Ber- Schweiz Med Wochenschr 1959; 89: 925–929. nard: Légende et iconographie. Westmalle, Abdij 15 Hanson LA: Comparative immunological studies Westmalle, 1956. of the immune globulins of human milk and of 7 Ehrlich P: Über Immunität durch Vererbung und blood serum. Int Arch Allergy Appl Immunol

Saugung. Z Hyg Infect Krankh 1892. 1961; 18: 241–267. 8 Ehrlich P: Experimentelle Untersuchungen 16 Hanson LA, Johansson B: Immunological charac- über Immunität. I. Über Ricin. Dtsch Med terization of chromatographically separated pro-

Wochenschr 1891; 17: 976–979. tein fractions from human colostrum. Int Arch

9 Ehrlich P: Experimentelle Untersuchungen über Allergy Appl Immunol 1961; 20: 65–79. Immunität. II. Über Abrin. Deutsch Med 17 Beer AE, Billingham RE, Head JR: Natural trans-

Wochenschr 1891; 17: 1218–1219. plantation of leukocytes during suckling. Trans-

plant Proc 1975; 7: 399.

24 Ogra 18 Ogra S, Ogra PL: Immunologic aspects of human 34 Hanson LA: Immunobiology of Human Milk: colostrum and milk. II. Characteristics of lym- How Breastfeeding Protects Babies. Amarillo, phocyte reactivity and distribution of E-rosette Pharmasoft Publishing, 2004. forming cells at different times after the onset of 35 Johansen FE, Braathen R, Brandtzaeg P: The J

lactation. J Pediatr 1978; 92: 550–555. chain is essential for polymeric Ig receptor-medi-

19 Ogra S, Weintraub D, Ogra PL: Immunologic ated epithelial transport of IgA. J Immunol 2001;

aspects of human colostrum and milk. III. Fate 167: 5185–5192. and absorption of cellular and soluble compo- 36 Losonsky GA, Ogra PL: Mucosal Immune System nents in the gastrointestinal tract of the newborn. in Neonatal Infections: Nutritional and Immuno-

J Immunol 1977; 119: 245–248. logic Interactions. Orlando, Grune & Stratton, 20 Mohr JA: The possible induction and/or acquisi- 1984, pp 51–66. tion of cellular hypersensitivity associated with 37 Haneberg B: Immunoglobulins in feces from in-

ingestion of colostrum. J Pediatr 1973; 8: 1062– fants fed human or bovine milk. Scand J Immu-

1064. nol 1974; 3: 191–197. 21 Okamoto Y, Tsutsumi H, Kumar NS, Ogra PL: 38 Montgomery PC, Rosner BR, Cohn J, et al: The Effect of breast-feeding on the development of secretory antibody response: anti-DNP antibod- anti-idiotypic antibody response to F-glycopro- ies induced by dinitrophenylated type III pneu-

tein of respiratory syncytial virus in infant mice mococcus. Immunol Commun 1974; 3: 143–156. after post-partum maternal immunization. J Im- 39 McClelland DBL, Samson RR, Parkin DM, et al:

munol 1989; 142: 2507–2512. Bacterial agglutination studies with secretory IgA 22 Heremans JF, Heremans MT, Schultze HE: Isola- prepared from human gastrointestinal studies tion and description of a few properties of the with secretory IgA prepared from human gastro-

β2A-globulin of human serum. Clin Chim Acta intestinal secretions and colostrum. Gut 1972;

1959; 4: 96–102. 134: 50–458. 23 Chodirker WB, Tomasi TB Jr: Gamma-globulins: 40 Lönnroth I, Lange S: Purification and character- quantitative relationships in human serum and ization of hormone-like factor which inhibits

nonvascular fluids. Science 1963; 154: 533–536. cholera secretion. FEBS Lett 1997; 177: 104–108. 24 Tomasi TB Jr, Zigelbaum SO: The selective occur- 41 Vidal K, Labeta MO, Schiffrin EJ, et al: Soluble rence of 1A globulins in certain body fluids. J CD14 in human breast milk and its role in innate

Clin Invest 1963; 42: 1552–1560. immune responses. Acta Odontol Scand 2001; 59: 25 Bienenstock J, Tomasi TB Jr: Secretory γA in nor- 330–334.

mal urine. J Clin Invest 1968; 1162–1171. 42 Ngom PT, Collinson AC, Pido-Lopez J, et al: Im- 26 Lidin-Janson G, Sohl-Akerlund A: Antibody- proved thymic function in exclusively breastfed forming cells in human colostrum after oral im- infants is associated with higher interleukin 7

munization. Nature 1975; 257: 797–799. concentrations in their mothers’ breast milk. Am

27 Roux ME, McWilliams M, Phillips-Quagliata JM, J Clin Nutr 2004; 80: 722–728. et al: Origin of IgA-secreting plasma cells in the 43 Ho FC, Wong RL, Lawton JW, et al: Human co-

mammary gland. J Exp Med 1977; 146: 1311– lostral and breast milk cells. A light and electron

1322. microscopic study. Acta Paediatr Scand 1979; 68: 28 Fishaut M, Murphy D, Neifert M, et al: Broncho- 389–396. mammary axis in the immune response to respi- 44 Weller IJ, Hickler W, Sprenger R: Demonstration

ratory syncytial virus. J Pediatr 1981; 99: 186–191. that milk cells invade the suckling neonatal

29 Holmgren J, Czerkinsky C: Mucosal immunity mouse. Am J Reprod Immunol 1983; 4: 95–98.

and vaccines. Nat Med 2005; 11: 545–553. 45 Jain L, Vidyasagar D, Xanthou M, et al: In vivo 30 Ogra PL, Dayton DH: Immunology of Breast distribution of human milk leucocytes after in- Milk: A Monograph of the National Institute of gestion by newborn baboons. Arch Dis Child

Child Health and Human Development. New 1989; 64(7 Spec No): 930–933. York, Raven Press, 1979. 46 Schnorr KL, Pearson LD: Intestinal absorption of 31 Mestecky J, Blair C, Ogra PL (eds): Immunology maternal leucocytes by newborn lambs. J Reprod

of Milk and the Neonate. Adv Exp Med Biol. New Immunol 1984; 6: 329–337. York, Plenum, 1990, vol 310. 47 LeDeist F, de Saint-Basile G, Angeles-Cano E, et 32 Kim JH, Bode L, Ogra PL: Human milk; in Wil- al: Prostaglandin E2 and plasminogen activators son C, Nizet V, Maldonado Y, et al (eds): Rem- in human milk and their secretion by milk mac-

ington and Klein’s Infectious Diseases of the rophages. Am J Reprod Immunol Microbiol 1986;

Fetus and Newborn Infant, ed 8. Philadelphia, 11: 6–10. Saunders, 2015, chapt 5, pp 189–216. 48 Satomi M, Shimizu M, Shinya E, et al: Transmis- 33 Institute of Medicine: Infant Formula: Evaluating sion of macrophage-tropic HIV-1 by breast milk

the Safety of New Ingredients. Washington, Na- macrophages via DC-SIGN. J Infect Dis 2005; 191: tional Academies Press, 2001. 174–181.

Immunology of Human Milk and Lactation 25 49 Wirt DP, Adkins LT, Palkowetz KH, et al: Acti- 56 Cabinian A, Sinsimer D, Tang M, et al: Transfer vated and memory T lymphocytes in human of maternal immune cells by breastfeeding: ma-

milk. Cytometry 1992; 13: 282–290. ternal cytotoxic T lymphocytes present in breast 50 Bertotto A, Gerli R, Fabietti G, et al: Human milk localize in the Peyer’s patches of the nursed

breast milk T lymphocytes display the phenotype infant. PLoS One 2016; 11:e0156762. and functional characteristics of memory cells. 57 Oftedal OT, Dhouailly D: Evo-devo of the mam-

Eur J Immunol 1990; 20: 1877–1880. mary gland. J Mammary Gland Biol Neoplasia

51 Lindstrand A, Smedman L, Gunnlaugsson G, et 2013; 18: 105–120. al: Selective compartmentalization of γδ-T lym- 58 Rosa FW: Breast-feeding in family planning. PAG

phocytes in human breastmilk. Acta Paediatr Bull 1975; 5: 5–10.

1997; 86: 890–891. 59 Hobcraft JN, McDonald JW, Rutstein SO: Demo- 52 Kunkel EJ, Campbell JJ, Haraldsen G, et al: Lym- graphic determinants of infant and early child phocyte CC chemokine receptor 9 and epithelial mortality: a comparison analysis. Population

thymus-expressed chemokine (TECK) expression Studies 1985; 39: 363–385. distinguish the small intestinal immune compart- 60 Rogier EW, Frantz AL, Bruno ME, et al: Secretory ment: epithelial expression of tissue-specific che- antibodies in breast milk promote long-term in- mokines as an organizing principle in regional testinal homeostasis by regulating the gut micro-

immunity. J Exp Med 2000; 192: 761–768. biota and host gene expression. Proc Natl Acad

53 Schlesinger JJ, Covelli HD: Evidence for transmis- Sci USA 2014; 111: 3074–3079. sion of lymphocyte responses to tuberculin by 61 Donaldson GP, Ladinsky MS, Yu KB, et al: Gut

breast-feeding. Lancet 1977; 2(8037): 529–532. microbiota utilize immunoglobulin A for muco-

54 Zhang L, van Bree S, van Rood JJ, Claas FH: Influ- sal colonization. Science 2018; 360: 795–800. ence of breast feeding on the cytotoxic T cell al- 62 Soto-Ramirez N, Karmaus W, Zhang H, et al:

lorepertoire in man. Transplantation 1991; 52: Modes of infant feeding and the occurrence of 914–916. coughing/wheezing in the first year of life. J Hum

55 Moles J-P, Tuaillon E, Kankasa C, et al: Breast Lact 2013; 29: 71–80. milk cell trafficking induces microchimerism- mediated immune system maturation in the in-

fant. Pediatr Allergy Immunol 2018; 29: 133–143.

26 Ogra Immunology of Milk and Lactation

Published online: March 16, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 27–37 (DOI: 10.1159/000505336)

The Mammary Gland as an Integral Component of the Common Mucosal Immune System

a, b Jiri Mestecky a Department of Microbiology and Medicine, University of Alabama at Birmingham, b Birmingham, AL, USA; Laboratory of Cellular and Molecular Immunology, Institute of Microbiology, Czech Academy of Sciences, Prague, Czech Republic

Abstract The human mammary gland is an integral effector component of the common mucosal immune system. However, from physiological and immunological aspects, it displays several unique features not shared by other mucosal sites. The development, maturation, and activity of the mammary gland exhibits a strong hormonal dependence. Furthermore, in comparison to the intestinal and respiratory tracts, the mammary gland is not colonized by high numbers of bacteria of enormous diversity and does not contain mucosal inductive sites analogous to the intestinal Peyer’s patches. Consequently, when exposed to antigens, local or generalized immune responses are low or not present. Comparative evaluations of various immunization routes effective in the induction of antibodies in human milk are limited. Systemic immunization induces IgG antibodies in plasma, but due to the low levels of total IgG in human milk, their protective effect remains unknown. Oral or intranasal immunization or infection induces secretory IgA in milk, as demonstrated in several studies. Other routes of mucosal immunization, such as sublingual or rectal exposure effective in the induction of antibodies in various external secretions, have not been explored in the mammary gland. Because secretory IgA in milk displays protective functions, alternative immunization routes and antigen delivery systems should be explored. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

In nature, milk is the external secretion which is essential for the survival of the offspring due to its nutritional value and, from the immunological view, as the source of passive protection against infection [1, 2]. Humans are the only mammals in which maternal milk is frequently substituted by milk proteins derived from other mammalian species or plants. Although breastfeeding is preferred to alternative means of nutrition, the prenatal transplacental transport of plasma- derived immunoglobulins into the fetal circulation provides in many species effective protection in the systemic compartment [1]. In addition to the humoral factors of innate immunity, milk of various mammalian species contains high levels of immunoglobulin, which differ in their structural features and effector functions [1, 3]. In sharp contrast to the dominance of IgG in the milk of many mammals (e.g., pigs, cows, and horses), human milk contains more IgA in its secretory form as SIgA than immunoglobulins of other isotypes [1, 3–5]. Due to the essential role of antibodies in the prevention of mucosally acquired infections [6], extensive efforts have been devoted to the design of vaccines effective in the induction of antibodies of desired specificity in external secretions, including the evaluation of routes of immunizations, forms of antigen delivery of relevant antigens, and possible use of mucosal adjuvants [7, 8].

Properties, Origin, and Biological Activities of Antibodies in Milk

In contrast to IgG antibodies, which dominate in the milk of many species and are derived from the circulatory pool [1], human milk contains IgA as the main immunoglobulin isotype [5]. It is represented by SIgA in its dimeric (∼60%) and tetrameric (∼40%) forms, with small amounts of dimeric IgA lacking the secretory component (SC; see below) and trace amounts of monomeric IgA [9]. The presence of SIgA in its dimeric and tetrameric forms is of functional importance due to 4 and 8 antigen binding sites. Furthermore, the characteristic distribution of SIgA subclasses reflects, to a certain degree, the origin of cells producing SIgA1 or SIgA2 as well as the specificity of these antibodies [9, 10]. Contrary to the earlier proposal in which monomeric IgA was polymerized within epithelial cells during the transcytotic pathway through the acquisition of the epithelial IgA receptor called SC, immunochemical analyses of milk SIgA clearly revealed that polymeric IgA (pIgA) dimers and tetramers are produced in these forms by IgA plasma cells adjacent to mucosal and glandular epithelial cells expressing a receptor (polymeric immunoglobulin receptor), which selectively binds pIgA and IgM and after epithelial transcytosis remains associated with polymeric im-

28 Mestecky Table 1. Functional advantages of high SIgA and low IgG levels in human milk [6, 30]

SIgA IgG

Absorption from the neonatal gut – ± (species differences) Spectrum of antibodies reflects the maternal experience Resistance to proteolysis by endogenous enzymes + – Number of antigen-binding sites (bonus effect of multivalency)/virus neutralization ++++ + Effector functions Enhance interactions with factors of innate immunity (mucin, peroxidase, lactoferrin, and lysozyme) + – SIgA-associated glycans (interactions with bacteria) ++ – Inhibition of antigen absorption ++ + Inhibition of bacterial adherence ++ ±

munoglobulin molecules such as SC [11]. These structural studies convincingly excluded the origin of milk IgA in humans from the circulation and clearly demonstrated its local origin in the mammary gland. Although not analyzed in milk, other studies of salivary and intestinal IgA revealed the strict local origin: monoclonal pIgA present in abundance in the sera of patients with multiple myelo- matosis or intravenously administered radioactively labeled pIgA were detectable only in trace quantities in the saliva and intestinal fluid [11]. These findings are of considerable importance with respect to the previously considered possibility that the intravenously administered monoclonal pIgA of desired specificity would be selectively transported into external secretions and thus provide effective protection against mucosal pathogens. Large epidemiological studies indicate that breastfeeding provides nutritional, developmental, and anti-infectious advantages to the infant regarding diarrhea, neonatal septicemia, and respiratory and urinary tract infections, particularly in developing countries where it significantly reduced morbidity and mortality [1, 3]. SIgA antibodies from milk display their protective activity in several partially unrelated mechanisms (Table 1) [6]. Biologically active antigens such as viruses, toxins, and enzymes are effectively neutralized based on the specificity of vaccine-induced or naturally occurring antibodies. Importantly, the neutralizing activity may extend not only to the free SIgA in milk but also to cells containing passively acquired SIgA due to the expression of Fcα receptors [6, 11]. Mucosal bacteria, particularly in the intestinal tract and oral cavity, are in vivo coated with SIgA without harm [12]. In fact, antibody coating inhibits adherence of bacteria to the receptors expressed on the surface of epithelial cells and participates in the formation of a bacterial biofilm of the same species at

Common Mucosal Immune System 29 Table 2. Presence of natural or infection-induced antibodies to various antigens in human colostrum and milk [1–4, 13, 32]

Bacteria Viruses Other antigens

Escherichia coli (incl. LPS) Influenza Candida albicans Salmonella Rotavirus Shigella Rubella Bovine γ-globulin Vibrio cholerae Poliovirus Bovine β-lactalbumin Bacterioides fragilis Echoviruses Streptococcus pneumoniae Respiratory syncytial virus (incl. polysaccharide capsule) Cytomegalovirus Bordetalla pertussis Herpes simplex Clostridium tetani Human immunodeficiency virus Corynebacterium diphtheriae Arborviruses Streptococcus mutans (Semliki Forest virus, Dengue virus) (antigen I/II) Haemophilus influenzae Tetanus toxoid Neisseria meningitidis (polysaccharide)

relevant mucosal niches. It appears that in addition to specific antibody activity, abundant SIgA-associated glycan side chains on SC and heavy chains effectively bind to the corresponding bacterial glycan structures, thus preventing their adherence to epithelial receptors [12]. However, SIgA in concert with humoral factors of innate immunity enhances (or focuses) their antimicrobial activities [6]. The inhibition of absorption of soluble but biologically inert antigens from food by SIgA has been well documented as a means to prevent the overstimulation of the entire immune system by the increased absorption of such antigens in the absence of specific antibodies [6]. Furthermore, it should be stressed that due to the multivalency of milk SIgA (4 antigen binding sites in dimers and 8 in tetramers), the biological effectiveness for viral neutralization is enormously enhanced due to the bonus effect of multivalency [6, 12].

Specificity of Antibodies in Milk Reflects the Site of Antigenic Stimulation at Various Inductive Sites

The mammary gland as the effector site is populated by precursors of IgA-producing cells from remote inductive sites [7, 8, 10]. Consequently, the specificity of SIgA in milk depends on the encounter with antigens, dominantly in the gastrointestinal and respiratory tracts. As shown in Table 2, human colostrum and

30 Mestecky milk contain antibodies of the IgA isotype to a broad spectrum of environmental antigens of microbial and food origin. In response to the antigens encountered by the mother at the time of pregnancy and after the delivery, such antibodies provide the most relevant passive SIgA-mediated immunity [1–4, 13]. Although there is considerable individual variability in total and antigen-specific SIgA antibody levels among lactating mothers, antibodies to gram-positive and gram- negative bacteria, viruses, and food antigens are commonly found in all samples [4, 13]. Not surprisingly and in harmony with earlier results, antibodies to food antigens and antigens I/II of the oral bacterium Streptococcus mutans are associated with the IgA1 subclass, while those against lipopolysaccharides of gram- negative bacteria present in the large intestine; bacterial polysaccharides are mostly IgA2 [9, 13]. Interestingly, the distribution of total IgA1 and IgA2 was ∼53 and 47%, respectively, with marked individual variability [13]. In comparison to other external secretions, human milk is in this respect reminiscent of the IgA subclass distribution in the lower intestinal tract and markedly differs from the secretions of the upper respiratory and upper digestive tracts [5]. There are, however, several interesting observations which remain unexplained. External secretions, including milk and sera obtained from HIV-infected individuals, display extremely low levels of HIV-specific antibodies of the IgA isotypes irrespective of the route of HIV infection [14]. It appears that HIV, specifically its negative factor (nef), selectively suppresses IgA responses. Furthermore, milk collected from lactating mothers contains high titers of antisperm antibodies of the IgA isotype [15]. Because the female genital tract is a poor inductive site, and various antigens administered intravaginally stimulate only minimal or no local responses [16], it is possible to speculate that these antisperm antibodies are induced due to the exposure to sperm by alternative inductive sites, specifically through oral or anal receptive sexual encounter, both of which are effective in the induction of generalized mucosal immune responses [17].

Induction of Antibodies in Milk by Local or Systemic Immunization

The presence of antibody-secreting cells as well as T cells in secretory glands, including the mammary gland [18], was exploited in several attempts to induce local responses in animal models. Thus, the injection of microbial antigens or hapten carrier conjugates into the mammary glands of experimental animals has been used for the induction of antibody responses in milk [1, 19]. Importantly, such immunization induces the concomitant systemic responses dominated by IgG rather than IgA in sera of immunized animals. Furthermore, adjuvants causing local inflammatory reactions in the immunized gland were usually re-

Common Mucosal Immune System 31 quired for boosting of local responses. Although retrograde instillation of antigens into the duct of mammary glands has been explored with some success in animal experiments [1], the acceptance of such immunization attempts in humans is highly unlikely. The possibility that antibodies from the circulation could contribute significant quantities of IgG or IgA to external secretions, including the milk, has been explored in early studies [1]. In animals whose milk contains IgG as the dominant isotype, systemic immunization is effective, and IgG of plasma origin is present in their milk [1]. In contrast, in human milk, low levels of total IgG are present. Nevertheless, subcutaneous immunization with 3 selected rubella virus vaccines resulted in the induction of IgG-, IgM-, and IgA- specific antibodies in sera, but only minimal levels of IgG or IgM virus-specific antibodies were detectable in milk [20]. However, IgA antibodies were detectable in the milk of all systemically immunized women 2–4 weeks after immunization, with peak responses at the 4th week. Although primary immunization with previously unencountered antigen does not stimulate mucosal immune responses, in women who had been previously sensitized by the mucosal route, systemic immunization could evince an SIgA response in milk, as demonstrated with cholera vaccine [19]. In previously unexposed lactating Swedish women, systemically administered cholera vaccine did not induce SIgA antibodies in milk. In contrast, in lactating Pakistani women, presumably naturally exposed to Vibrio cholerae, such an immunization induced SIgA antibodies in milk [19]. Thus, the initial mucosal exposure profoundly influenced the outcome of subsequent systemic immunization and may favor the induction of milk SIgA antibodies to some antigens.

Induction of Generalized Mucosal Responses and the Common Mucosal Immune System

Pioneering studies performed in animals and later in humans helped to discov- er the origin of antibody-forming cells in anatomically remote mucosal tissues and associated secretory glands, including the mammary, salivary, and tear glands, and had an enormous impact on the feasible approaches and design of mucosally administered vaccines [19]. Oral administration of dinitrophenylated pneumococci to lactating rabbits led to the appearance of specific antibodies of the IgA isotype in milk [21]. Subsequent extensive experiments with oral administration of various particulate or soluble antigens confirmed earlier observations and demonstrated that, in addition to milk, SIgA antibodies were present in secretions of other mucosal tissues and glands [19]. The fundamental expla- nation for these observations of such importance was provided by earlier studies

32 Mestecky Table 3. Immunization routes: advantages and disadvantages [6, 19, 30]

Inductive sites Effector sites (advantages and disadvantages) oral respiratory1 intestinal mammary genital2 (saliva) (fluids) gland (milk)

Nasal route Low levels of Ag required + ++ + + + Ag presentation Oral route (enteric) Presence of other unrelated Ag, mucosal microbiota, enzymatic degradation, large amounts of Ag + + +++ ++ + Sublingual route Low Ag levels required + + ? + Rectal route Low acceptability + ++ ? + Intravaginal route Only female genital tract –– –? ±

Ag, antigen. 1 Nasal secretion. 2 Cervicovaginal secretions and semen.

concerning the tissue origin of IgA-producing cells found in mucosal tissues and glands [22]. Several investigators demonstrated that lymphocytes from the intestinal Peyer’s patches, bronchus-associated lymphoid tissue, and perhaps Waldeyer’s ring of the oropharynx are an enriched source of cells that express the potential to populate remote mucosal tissues and glands [19, 23]. Experi- ments performed primarily in mice led to the extended phenotyping of IgA precursor cells as surface IgA-positive cells that express surface receptors, later termed homing receptors, that are involved in interactions with specific ligands present on the surface of endothelial cells of postcapillary venules in mucosal tissues [24]. Importantly, for immunization studies and their physiologic inter- pretation, cells from such inductive sites, including Peyer’s patches, bronchus- associated lymphoid tissue, Waldeyer’s ring, rectal tonsils, and sublingual tissue, differ in their expression of homing receptors, which interact with corresponding ligands and, therefore, lead to the tissue-selective distribution of cells from various inductive site to tissue-elective effector sites (Table 3) [19, 23, 24]. It should be stressed, however, that information relevant to the human mammary gland with respect to the origin of IgA precursor cells from various inductive sites has not been, for obvious ethical reasons, explored and remains controversial. Based on immunohistochemical staining of 2 samples of lactating mammary glands and spectra of antibodies, some authors speculate that the IgA precursors originate in the upper respiratory tract and Waldeyer’s ring, while others

Common Mucosal Immune System 33 propose that the inductive sites in the intestinal tract are the most important site [13, 18, 19, 23]. This opinion is based on the distribution of IgA1 and IgA2 subclasses as well as specificity of SIgA antibodies in milk [13]. Subsequent immunization studies performed in animal models clearly demonstrated that immunization initiated at various inductive sites results in the appearance of specific antibodies not only at the site of the original antigen encounter but also at the remote mucosal effector sites, including the lactating mammary glands [19]. The existence of the common mucosal immune system was confirmed in humans by immunochemical analyses of spectra of naturally occurring or immunization-induced antibodies to a variety of antigens in tears, saliva, intestinal fluid, and genital secretions [19, 25, 26]. This concept was confirmed by the demonstration of specific antibody-secreting cells expressing mucosal homing receptors in peripheral blood and ultimately the presence of antibodies in external secretions [10, 19, 25, 26]. As indicated earlier, human milk contains SIgA antibodies to many antigens of microbial and food origin, but for ethical reasons, immunization studies have been performed in lactating women to a limited extent. Infection of the gastrointestinal tract with a serologically unique strain of Escherichia coli led to the induction of E. coli-specific antibodies in collected milk, and the ingestion of Streptococcus mutans in gelatin capsules induced specific antibodies in milk, tears, and saliva of a lactating mother [27, 28]; the levels of preexisting serum antibodies were not altered. In former studies [27], specific antibody-secretory cells were detectable by plaque assay. However, subsequent immunohistochemical and electron microscopy studies revealed that the plaque-forming cells were actually polymorphonuclear leukocytes, macrophages, and milk granules containing large amounts of internalized SIgA and not true antibody-secreting plasma cells, plasmablasts, or lymphoblasts [29]. In summary, the presence of common mucosal immune systems, including the mammary gland, provides the opportunity to further explore immunization efforts leading to the induction of humoral immune responses in milk to mucosal pathogens involved in the high level of morbidity and mortality in neonates [30].

Future Studies

As described earlier, the effectiveness of various immunization routes [19, 30] in the induction of antibody responses of desired specificity in human milk has not been extensively explored. However, because of new advances in the field of mucosal vaccination, including the identification of additional immunization routes, further experiments should be performed (Fig. 1).

34 Mestecky Effector sites Inductive sites

Antigens/vaccines NALT Retropharyngeal Upper respiratory LN tract Sublingual Submaxillary LN

Mammary glands

Cervical GALT LN GI tract Peyer’s Mesenteric patches LN

Genital tract Thoracic duct

Blood

Fig. 1. The common mucosal immune system in humans and the integration of the mammary gland. B cells from the bone marrow enter inductive sites such as the gut-associated lymphoid tissues (GALT; Peyer’s patches and rectal tonsil), upper respiratory tract sublingual sites, Waldeyer’s ring, and bronchus-associated lymphoid tissue. Under the influence of T cells, epithelial cells, and dendritic cells, they express SIgA. Microbial and environmental antigens enter the inductive sites through the pinocytotic and phagocytic M cells to interact with the antigen-processing and -presenting cells. IgA-committed and antigen-sensitized B cells leave the inductive sites to populate remote effector sites based on the interaction of their homing receptor with corresponding ligands on endothelial cells in postcapillary venules in the mucosal effector sites, including the mammary gland. In these locales, the terminal differentiation into IgA-secreting plasma cells occurs. The source of B cells from various inductive sites which populate the human mammary gland apparently includes both the upper respiratory as well as intestinal tracts. LN, lymph node. NALT, nasal-associated lymphoid tissue.

1. Which immunization route is effective in the induction of antibodies of desired specificity in human milk? The intranasal, sublingual, oral, and perhaps rectal routes of immunization should be compared. 2. Determine the phenotypes of specific antibody-secreting cells in the mammary gland, as compared to the peripheral blood, with respect to the immunoglobulin isotypes and expression of homing receptors. 3. Determine the levels and duration of humoral immune responses of specific antibodies with regard to the immunization route, immunoglobulin isotypes, and the antigen delivery systems in human milk.

Common Mucosal Immune System 35 4. Can such responses be boosted by repeated immunization at the original priming site, or will a combination of such routes be more effective? 5. Are there marked racially related differences (developed and developing countries) in the magnitude and specificity of immune responses to relevant antigens? 6. Determine the most effective timing of maternal mucosal immunization to provide optimal levels of protective antibodies in milk after birth. 7. Continue efforts in the exploration of effectiveness of currently evaluated mucosally administered adjuvants in the induction of specific antibodies in milk.

Acknowledgment

The research was supported by the Czech Science Foundation (grant No. 17-11275S).

Disclosure Statement

The author declares that he has no relevant or material financial interest that relates to the research described in this paper.

References

1 Butler JE, Rainard P, Lippolis J, et al: The mam- 6 Mestecky J: Protective activities in mucosal anti- mary gland in mucosal and regional immunity; in bodies, in Kiyono H, Pascual D (eds): Mucosal Mestecky J, Strober W, Russell MW, et al (eds): Vaccines: Innovation for Preventing Infectious Mucosal Immunology, ed 4. Amsterdam, Else- Disease, ed 2. Amsterdam, Elsevier/Academic vier/Academic Press, 2015, vol II, pp 2269–2306. Press, 2019, pp 71–84. 2 Sharma D, Hanson LA, Korotkova M, et al: Hu- 7 Russell MW, Mestecky J: Mucosal vaccines: an man milk: its components and their immunobio- overview: in Mestecky J, Strober W, Russell MW, logic functions; in Mestecky J, Strober W, Russell et al (eds): Mucosal Immunology, ed 4. Amster- MW, et al (eds): Mucosal Immunology, ed 4. Am- dam, Elsevier/Academic Press, 2015, vol II, pp sterdam, Elsevier/Academic Press, 2015, vol II, 1039–1046. pp 2307–2344. 8 Boyaka PN, McGhee JR, Czerkinsky C, Mestecky 3 Mestecky J, Blair C, Ogra PL: Immunology of J: Mucosal vaccines: an overview: in Mestecky J,

milk and the neonate. Adv Exp Med Biol 1990; Bienenstock J, Lamm ME, et al (eds): Mucosal

310: 1–488. Immunology, ed 3. Amsterdam, Elsevier/Aca- 4 Ogra SS, Weintraub DI, Ogra PL: Immunologic demic Press, 2005, vol I, pp 855–874. aspects of human colostrum and milk: interac- 9 Woof JM, Mestecky J: Mucosal immunoglobu- tion with the intestinal immunity of the neonate. lins; in Mestecky J, Strober W, Russell MW, et al

Adv Exp Med Biol 1978; 107: 95–112. (eds): Mucosal Immunology, ed 4. Amsterdam, 5 Jackson S, Moldoveanu Z, Mestecky J: Appendix Elsevier/Academic Press, 2015, pp 287–324. I: collection and processing of human mucosal 10 Pakkanen SH, Kantele JM, Moldoveanu Z, et al: secretions: in Mestecky J, Strober W, Russell Expression of homing receptors on IgA1 and MW, et al (eds): Mucosal Immunology, ed 4. Am- IgA2 plasmablasts in blood reflects differential sterdam, Elsevier/Academic Press, 2015, vol II, distribution of IgA1 and IgA2 in various body

pp 2345–2344. fluids. Clin Vaccine Immunol 2010; 17: 393–401.

36 Mestecky 11 Baker K, Blumberg RS, Kaetzel CS: Immunoglob- 21 Montgomery PC, Cohn J, Lally ET: The induction ulin transport and immunoglobulin receptors; in and characterization of secretory IgA antibodies.

Mestecky J, Strober W, Russell MW, et al (eds): Adv Exp Med Biol 1974; 45: 453–462. Mucosal Immunology, ed 4. Amsterdam, Else- 22 Craig SW, Cebra JJ: Peyer’s patches: an enriched vier/Academic Press, 2015, pp 349–409. source of precursors for IgA-producing immuno-

12 Mestecky J, Russell MW: Specific antibody activ- cytes in the rabbit. J Exp Med 1971; 134: 188–200. ity, glycan heterogeneity and polyreactivity con- 23 Brandtzaeg P: The mucosal B cell system. in Mes- tribute to the protective activity of S-IgA at mu- tecky J, Strober W, Russell MW, et al (eds): Mu-

cosal surfaces. Immunol Lett 2009; 124: 57–62. cosal Immunology, ed 4. Amsterdam, Elsevier/ 13 Ladjeva I, Peterman JH, Mestecky J: IgA subclass- Academic Press, 2015, pp 623–667. es of human colostral antibodies specific for mi- 24 Mikhak Z, Agace WW, Luster AD: Lymphocyte crobial and food antigens. Clin Exp Immunol trafficking to mucosal tissues; in Mestecky J,

1989; 78: 85–90. Strober W, Russell MW, et al (eds): Mucosal Im- 14 Mestecky J, Tamaras G: HIV-1/SIV humoral re- munology, ed 4. Amsterdam, Elsevier/Academic sponses in external secretions. Curr Immunol Press, 2015, pp 805–839.

Rev 2019; 15: 49–62. 25 Mestecky J, McGhee JR, Arnold RR, et al: Selec- 15 Ulcova-Gallova Z, Mestecky J, Fialova P: Anti- tive induction of an immune response in human spermatozoal antibodies in human colostrum external secretions by ingestion of bacterial anti-

and milk. Zentralbl Gynäkol 1994; 116: 636–638. gen. J Clin Invest 1978; 61: 731–737. 16 Russell MW, Mestecky J: Tolerance and protec- 26 Czerkinsky C, Prince SJ, Michalek SM, et al: IgA tion against infection in the genital tract. Immu- antibody-producing cells in peripheral blood af-

nol Invest 2010; 39: 500–525. ter antigen ingestion: evidence for a common 17 Kantele A, Hakkinen M, Moldoveanu Z, et al: mucosal immune system in humans. Proc Natl

Differences in immune responses induced by oral Acad Sci USA 1987; 84: 2449–2453. and rectal immunization with Salmonella typhi 27 Goldblum RM, Ahlstedt S, Carlsson B, et al: Anti- Ty21a: evidence for compartmentalization within body-forming cells in human colostrum after oral

the common mucosal immune system in hu- immunization. Nature 1975; 257: 797–799.

mans. Infect Immun 1998; 66: 5630–5635. 28 Gregory RL, Schöller M, Filler SJ, et al: IgA anti- 18 Brandtzaeg P: The secretory immune system of bodies to oral and ocular bacteria in human ex-

lactating human mammary glands compared ternal secretions. Protides Biol Fluids 1985; 32: with other exocrine organs. Ann NY Acad Sci 53–56.

1983; 409: 353–382. 29 Moro I, Crago SS, Mestecky J: Localization of IgA 19 Mestecky J: The common mucosal immune sys- and IgM in human colostral elements using im-

tem and current strategies for induction of im- munoelectron microscopy. J Clin Immunol 1983;

mune responses in external secretions. J Clin Im- 3: 382–391.

munol 1987; 7: 265–276. 30 Ogra PL, Losonsky GA, Fishaut M: Colostrum- 20 Losonsky GA, Fishaut JM, Strussenberg J, et al: derived immunity and maternal-neonatal inter-

Effect of immunization against rubella on lacta- action. Ann NY Acad Sci 1983; 409: 82–95. tion products. I. Development and characterization of specific immunologic reactivity in breast

milk. J Infect Dis 1982; 145: 654–660.

Common Mucosal Immune System 37 Immunology of Milk and Lactation

Immunomodulatory Components of Human Colostrum and Milk

a, b a, c b Helena Tlaskalová-Hogenová Miloslav Kverka Jiří Hrdý a b Institute of Microbiology, Czech Academy of Sciences, Prague, Czech Republic; Institute of Immunology and Microbiology, 1st Faculty of Medicine, Charles University, Prague, Czech Republic; c Institute of Experimental Medicine, Czech Academy of Sciences, Prague, Czech Republic

Abstract Human milk is a unique and complex secretion differing from lacteal secretions of other species. Besides nutrition, it provides protection during the newborn’s adaption to the extrauterine environment and reduces the morbidity and mortality caused by both infectious and noninfectious diseases. Its components act directly against infectious agents, but they also accelerate the newborn’s immune system development, increasing its capacity for defense and reducing the risk of allergy and other immune-related diseases. Cytokines show the most refined immunomodulatory effects, but oligosaccharides, hormones, and other components affect the newborn’s immunity as well. Furthermore, milk components substantially affect the microbial colonization of infant mucosa, which substantially influences the development of all parts of the immune system. All these components act primarily locally, on the mucosal membranes, preventing the penetration of microbes and other antigenic components into the circulation thus ensuring effective defense without the damaging inflammation. Human lacteal secretions contain a number of live cells. Although there are no major differences in the cytokine production between allergic and healthy mothers, they are able to respond to multiple stimuli. By increasing happiness, boosting protective immunity, and decreasing the risk of breast cancer, breastfeeding may have multiple benefits for the mother as well. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Breastfeeding: The Immunologic Connection between Mother and Infant

Breastfeeding represents the continuation of the tight relationship between mother and offspring after birth. In addition to the nutritional content, colostrum (early milk within 4–5 days after parturition) and mature milk contain many components that increase the resistance of the infant against infection. Due to its complex composition and multiple biologically active molecules, breastfeeding has numerous beneficial effects for both infant and mother (Fig. 1). Epidemiological and clinical studies conducted in different parts of the world concluded that breastfeeding reduces diarrheal diseases, e.g., shigellosis, cholera, and lambliasis, but it also protects against respiratory and urinary infections caused by various microbes including viruses [1]. This broad spectrum of diverse and beneficial effects is achieved by influencing multiple systems, most notably the immune system. The effect of breastfeeding on the infant’s immune system may be direct via immunomodulatory factors and cells contained in the milk, or it may be an indirect effect on the composition and function of commensal microbiota [2]. The effects of breastfeeding are not only beneficial for the infant, they also have positive effects on the mother. Breastfeeding improves maternal-infant bonding, it helps uterus involution, it delays the return of ovulation, and it may even decrease the risk of osteoporosis, cardiovascular diseases, and breast and ovarian cancer [3]. The molecular mechanisms responsible for the multiple beneficial effects of breastfeeding for the infant and mother are not yet completely understood; however, it is clear that they are realized via complex immune-endocrine-brain signaling networks. These signaling pathways are functioning also in very important host-microbiota interactions after birth. As this workshop deals with various immunologic and microbiological aspects of human milk, which are the objectives of other parts of the workshop, we will focus on selected factors exert- ing immunomodulatory effects.

Cytokines and Other Immunomodulatory Factors

Colostrum and milk are an important source of immunomodulatory components, such as cytokines, chemokines, and growth factors. These factors are important immunomodulators helping the infant’s mucosal immune system to establish the proper reactivity by regulating the reactivity of the infant’s immune cells and by accelerating the development of the mucosal barrier. This is a crucial step, because the infant is being colonized by a vast array of microorganisms, and the immature organism is suddenly exposed to the high load of antigenic

Immunomodulation by Human Milk 39 Mother – Immunomodulatory effect of lactation – Immunization against infant's oral cavity microbes – Prevention of breast cancer

Breastfeeding – Nutrients – Antigens and allergens – Immune cells – Antibodies – Cytokines/chemokines – Growth factors – Enzymes – Hormones – Commensal microbes

Infant – Effect on immune system development – Immune sytem priming – Protection from pathogens – Influence on the gut microbiota

Fig. 1. Bidirectional effect of breastfeeding on the infant’s and the mother’s immune system.

and mitogenic stimuli. In our laboratory, we used an antibody array to analyze the presence of the broadest spectrum of cytokines in colostrum and milk of healthy women and their changes early after delivery. We found a surprisingly broad spectrum of cytokines, and some of them were described in the colostrum for the first time [4]. Cytokines are secreted in the colostrum or milk by mammary gland epithelium and by resident leukocytes with a minor part being filtered from the mother’s blood [5]. Many of these factors perform a dual role, because they influence both the mammary gland and the suckling. TGF-β, one of the first cytokines reliably confirmed in human colostrum, is such an example. Most TGF-β in the lacteal secretion is in the form of TGF-β2, and its concentration is decreasing after birth, but its amount remains high due to the increase in the lacteal secretion volume. TGF-β influences both initial morphogenesis of the mammary gland and its involution during weaning by controlling epithelial cell death [6]. On the other hand, it supports the closure of the gut barrier and dampens the immune response to innocuous antigens thus supporting the development of oral tolerance to food and microbiota in the suckling [7]. It directly affects intestinal epithelial cells by inducing their differentiation and unresponsiveness to

40 Tlaskalová-Hogenová/Kverka/Hrdý antigenic stimuli and thus preventing their excessive reaction to gut microbiota, and accelerates the gut barrier closure [8]. While TGF-β-mediated gut barrier maturation is important for every neonate, it is crucial for immature preterm neonates, who may even benefit from oral TGF-β supplementation [8]. How- ever, TGF-β has a curious dual role, because it stimulates local B cells to switch toward secretory IgA production (SIgA; see later) [9], which is important for protecting infants from infections. However, the infant’s gut is also influenced by multiple other growth factors that influence epithelial and endothelial growth (e.g., epidermal, hepatocyte, and vascular endothelial growth factor, growth-regulated proteins, and angio- genin) and gut barrier closure, and prevent excessive stimulation of the underlying immune cells in the gut mucosa. But their effect may go beyond that and help to grow specialized tissue, such as the nervous system [10]. The joint activity of multiple cytokines (e.g., TNF-α, IL-17A, IL-1β, IL-6, IFN-γ, and CCL18) and chemokines (e.g., IL-8 and MCP-1) may help establish protective mucosal immune responses without the tissue damage associated with the unchecked inflammation. Balance between pro- and anti-inflammatory factors is crucial for dampening of inflammation and protecting the mammary gland from infection [11]. The balance in the lacteal secretion composition is significantly shifted during the infant’s infection, suggesting the existence of immunologic connection between lactating mothers and their infants [12]. These instructions help diseased children to cope with acute infection and subsequent recovery, but human milk significantly accelerates the development of gut-associated lymphoid tissue and other integral parts of mucosal and systemic immunity [13]. This effect increases the suckling’s humoral and cellular immune responses, which may extend the benefits of breastfeeding well beyond weaning.

Antibacterial Components, Enzymes, Hormones, and Oligosaccharides

Milk boosts the infant’s immature immune system with the factors that prevent infections. Colostrum and mature milk contain a broad range of anti-infectious agents, which belong to the innate arm of immunity. Apart from antibacterial activity, these substances often exert immunomodulatory effects as well. Lacto- ferrin, the most abundant protein in whey, is such example. It is not only bactericidal to several gram-negative and gram-positive pathogens and inhibits other microbes (viruses and fungi), it also modulates immune responses by binding to various membrane molecules (e.g., peptidoglycans and lectins) present on a wide variety of innate and adaptive immune cells. After its internalization, it in-

Immunomodulation by Human Milk 41 fluences multiple cytoplasmic and nuclear signaling pathways and functions of these cells [14]. Besides their primary enzymatic, hormonal, and nutritional functions, other milk components also influence the developing immune system of the infant either directly or indirectly by influencing their mucosal microbiota.

Secretory IgA and Autoantibodies

The large quantity of SIgA the mother delivers to the infant by colostrum and milk contains a broad spectrum of antibody specificities. The structure of SIgA determines its resistance to enzymatic digestion, thus guaranteeing its activity even in the distal parts of the infant’s intestine. These antibodies originate in the gut lymphoid tissue and are thus useful for gut mucosa protection of the immunologically immature infant, because they are directed primarily against gut microbes. However, the SIgA reactivity in milk has been studied mainly with regard to newborn’s defense, i.e., enhancing immunity against infection and preventing the development of inflammatory responses, and much less is known about other specificities. The common mucosal immune system of the mother ensures the transfer of antibodies to the mammary gland during lactation. SIgA antibodies are an important part of the intestinal barrier, because they prevent microbes from binding to the intestinal epithelium and penetration into the mucosa [15]. On the other side, together with other milk components, these secretory antibodies play an important role in the establishment and development of infant microbiota. During the epidemic of respiratory infections in the Prague maternity hospital, we had the opportunity to cooperate with neonatologists who preventively administered newborn infants with live probiotic strain Escherichia coli O83:K24:H31 (EcO83) and monitored the consequences of intestinal colonization in the following weeks. In particular, we were interested in the specific immune responses that we determined in stool filtrate samples and in infant sera but also in maternal milk. Repeated oral administration of this strain increased titers of specific antibodies directed against this microbe in stool and sera when compared with children who were not colonized [16]. To our surprise, we found a similar increase in EcO83-specific IgA in milk and sera of mothers of colonized children. This finding suggests that the mother’s immune system responds via breastfeeding to antigens that are present on the oral and intestinal mucosa of a breastfed baby even in the absence of these bacteria in the mother’s stool [17, 18]. Interestingly, this early postnatal colonization protected the children against the development of allergy even after 20 years [19].

42 Tlaskalová-Hogenová/Kverka/Hrdý Our laboratory has been studying the development of the antibody repertoire during the fetal and early postnatal periods in pigs. In this advantageous experimental model, 6-layered placenta does not allow the prenatal transfer of maternal antibodies to the newborn piglet, and, therefore, the colostrum is the only source of antibodies for piglets. The gnotobiological laboratory of the Institute of Microbiology of the Czechoslovak Academy of Sciences was established by Prof. Sterzl in the early 1960s and enabled to study the effect of microbiota on immune system development in germ-free piglets, rats, and mice [20]. Using newborn and germ-free, colostrum-deprived piglets, we have shown that the first antibodies produced by piglets were polyspecific in nature and reacted with autoantigens [21]. These natural autoantibodies play an important role in im- munoregulation during ontogeny but also in adults [22, 23]. We analyzed the human colostrum and milk SIgA for antigenic specificity of these autoantibodies. By using immunofluorescence, immunoblotting, and ELISA, we found that colostrum and milk of healthy mothers contain a broad array of autoantibodies. Most samples reacted with tissues of the ovary, pancreas, and adrenal glands. A minor portion of the colostrum samples reacted with liver, stomach, salivary glands, and kidneys. Colostral SIgA reactivity determined by ELISA was directed against extractable nuclear antigens, phospholipids, dsDNA, and cytoplasmic neutrophil antigen [24]. Furthermore, human milk contains antibodies to food antigens (including allergens), suggesting that SIgA antibodies prevent the penetration of these antigens into the circulation of the breastfed infant, which may be important for tolerance induction to food antigens [17, 18].

Cells

Maternal milk contains live cells of the maternal immune system. The highest concentration of live cells is early after delivery and then decreases in mature milk. The leukocyte concentration in colostrum varies broadly among women (approx. 106–107 cells/mL), and a typical colostral sample contains 40–50% macrophages, 40–50% neutrophils, and 5–10% lymphocytes [25]. There are significantly fewer cells in mature milk, and these are dominated by death cells, including cells of epithelial origin. Importantly, all the live cells are fully functional and activated as documented by the presence of CD45RO [26]. Colostral cells could enter the newborn via paracellular passage in the intestine because neonates do not have a fully developed gut barrier until approximately the 4th day after delivery. These colostral cells survive in neonates for a certain period of time and could influence the development of the immature neonatal immune system.

Immunomodulation by Human Milk 43 The neonatal immune system should recognize cells of the mother as foreign and develop effector immune responses against them, but this is rarely the case in real life. It seems that breastfeeding induces regulatory T cells that actively suppress neonatal immune responses against noninherited maternal antigens (NIMAs) [27]. In addition, the neonatal immune system is generally considered immature with Th2-biased and limited effector immune responses. Studies on experimental mouse models showed that for effective tolerance of NIMAs only prenatal exposure to NIMAs is not sufficient: postnatal exposure to NIMAs (via breastfeeding) is important [28]. This suggests that the cellular compounds of maternal milk are essential for priming the regulatory mechanisms of the neonate’s immune system. The close links between the mature maternal and the developing neonate immune system via breastfeeding suggest that the transmission of immune-mediated disorders from mother to newborn could be possible. The beneficial effects of the maternal milk are broadly acknowledged, but the abilities of colostral cells could be different in mothers with immune-mediated diseases. Therefore, we analyzed the differences in the proportion and functional capacity of colostral cells between healthy and allergic mothers. Surprisingly, there were no differences in Th1 or Th2 cytokine expression, and only increased expression of epidermal growth factor was detected in colostral cells of allergic mothers as compared to healthy ones [29]. To analyze the impact of these cells from the colostrum on the offspring’s cell-mediated immune responses, we cultivated them with the offspring’s cord blood cells using a Transwell system. There were no differences between colostral cells from healthy and allergic mothers in their capacity to induce cytokine expression typical for Th1, Th2, and regulatory T-cell immune responses in cord blood cells, which suggests that the quality and immunomodulatory properties of colostral cells of healthy and allergic mothers are comparable [30]. To test the capacity of colostral cells to produce these cytokines upon different types of stimulation, we stimulated cells isolated from the colostrum in vitro by pokeweed mitogen, EcO83, lipopolysaccharide, or phytohemagglutinin for 24 h and analyzed cytokine expression by RT-PCR (Fig. 2). Both cells from healthy and allergic mothers were able to induce significant expression of IFN-γ (Fig. 2a) and IL-10 (Fig. 2b), while they increased their expression of IL-13 only following phytohemagglutinin stimulation (Fig. 2c) and did not change IL-4 expression regardless of the stimulus (Fig. 2d). Surprisingly, there were no differences in gene expression between colostral cells of healthy and allergic mothers, which indicates that colostral cells from allergic mothers do not skew the development of the neonatal immune system into the proallergic phenotype. Therefore, it is unlikely that the transmission of “proallergic” cells of maternal origin is responsible for the induction of allergies in the offspring.

44 Tlaskalová-Hogenová/Kverka/Hrdý ns 25 15 ns * Healthy 20 * Allergic * 10 15 * ** * ** * * 10 5 5 Relative IL-10 expression Relative Relative IFN- γ expression Relative 0 0

LPS PHA LPS PHA LPS PHA LPS PHA PWM PWM PWM PWM a Control EcO83 Control EcO83 b Control EcO83 Control EcO83

100 * 6 ns 80 4 60 ** 40 2 20 Relative IL-4 expression Relative Relative IL-13 expression Relative 0 0

LPS PHA LPS PHA LPS PHA LPS PHA PWM PWM PWM PWM c Control EcO83 Control EcO83 d Control EcO83 Control EcO83

Fig. 2. Characterization of cytokine gene expression in colostral cells from healthy (n = 5) and allergic mothers (n = 4). Cells were stimulated by pokeweed mitogen (PWM), EcO83, lipopolysaccharide (LPS), or phytohemagglutinin (PHA) in vitro for 24 h, and expression of IFN-γ (a), IL-10 (b), IL-13 (c), and IL-4 (d) was analyzed by RT-PCR. There were no statistically significant differences in reactivity of similarly stimulated cells from healthy and allergic mothers, as analyzed by unpaired Student’s t test. ns, nonsignificant. * p < 0.05, ** p < 0.01, vs. nonstimulated cells (paired ANOVA with Dunnett’s multiple comparison test).

Conclusions

Human lacteal secretions are a source of numerous immunomodulatory components that influence immune system reactivity of the infant well beyond weaning. These factors include nutrients, immunoglobulins, cytokines, and other signaling molecules, immune cells, and microbes. All these factors contribute to the adaptation of the infant to the environment. Apart from colonization resistance and antigen exclusion, breastfeeding accelerates gut barrier closure, gut-associated lymphoid tissue development, and mucosal (oral) tolerance, and thus protects the infant from infectious and inflammatory diseases both imme- diately and later in life.

Immunomodulation by Human Milk 45 Acknowledgment

This work was supported by a grant from the Czech Science Foundation (No. 17-11275S).

Disclosure Statement

The authors declare that no financial or other conflicts exist in relation to the contents of this chapter.

References

1 Hanson LA: Breastfeeding provides passive and 12 Riskin A, Almog M, Peri R, et al: Changes in im- likely long-lasting active immunity. Ann Allergy munomodulatory constituents of human milk in

Asthma Immunol 1998; 81: 523–533; quiz 33–34, response to active infection in the nursing infant.

37. Pediatr Res 2012; 71: 220–225. 2 Ogra PL: Developmental aspects of the mucosal 13 Miklavcic JJ, Badger TM, Bowlin AK, et al: Hu- immune system: role of external environment, man breast-milk feeding enhances the humoral mucosal microflora and milk. Adv Exp Med Biol and cell-mediated immune response in neonatal

2009; 639: 41–56. piglets. J Nutr 2018; 148: 1860–1870. 3 Heinig MJ, Dewey KG: Health effects of breast 14 Telang S: Lactoferrin: a critical player in neonatal

feeding for mothers: a critical review. Nutr Res host defense. Nutrients 2018; 10:E1228.

Rev 1997; 10: 35–56. 15 Mestecky J, Russell MW: Specific antibody activ- 4 Kverka M, Burianova J, Lodinova-Zadnikova R, ity, glycan heterogeneity and polyreactivity con- et al: Cytokine profiling in human colostrum and tribute to the protective activity of S-IgA at mu-

milk by protein array. Clin Chem 2007; 53: 955– cosal surfaces. Immunol Lett 2009; 124: 57–62. 962. 16 Lodinova-Zadnikova R, Slavikova M, Tlaskalova- 5 Garofalo RP, Goldman AS: Cytokines, chemo- Hogenova H, et al: The antibody response in kines, and colony-stimulating factors in human breast-fed and non-breast-fed infants after artifi-

milk: the 1997 update. Biol Neonate 1998; 74: cial colonization of the intestine with Escherichia

134–142. coli O83. Pediatr Res 1991; 29: 396–399. 6 Brenmoehl J, Ohde D, Wirthgen E, Hoeflich A: 17 Lodinova-Zadnikova R, Tlaskalova-Hogenova H, Cytokines in milk and the role of TGF-beta. Best Kolesovova L: Effect of breastfeeding on immu-

Pract Res Clin Endocrinol Metab 2018; 32: 47–56. nologic development; in Ogra L (ed): Neonatal 7 Penttila IA: Milk-derived transforming growth Infections. Nutritional and Immunologic Inter factor-beta and the infant immune response. J actions. New York, Grune & Stratton, 1984.

Pediatr 2010; 156(suppl 2):S21–S25. 18 Casas R, Bottcher MF, Duchen K, Bjorksten B: 8 Donnet-Hughes A, Duc N, Serrant P, et al: Bioac- Detection of IgA antibodies to cat, beta-lactoglob- tive molecules in milk and their role in health ulin, and ovalbumin allergens in human milk. J

and disease: the role of transforming growth fac- Allergy Clin Immunol 2000; 105: 1236–1240.

tor-beta. Immunol Cell Biol 2000; 78: 74–79. 19 Lodinova-Zadnikova R, Cukrowska B, Tlaskalo- 9 Sonoda E, Hitoshi Y, Yamaguchi N, et al: Differ- va-Hogenova H: Oral administration of probiotic ential regulation of IgA production by TGF-beta Escherichia coli after birth reduces frequency of and IL-5: TGF-beta induces surface IgA-positive allergies and repeated infections later in life (after

cells bearing IL-5 receptor, whereas IL-5 pro- 10 and 20 years). Int Arch Allergy Immunol 2003;

motes their survival and maturation into IgA- 131: 209–211.

secreting cells. Cell Immunol 1992; 140: 158–172. 20 Sterzl J, Silverstein AM: Developmental aspects of

10 Bar S, Milanaik R, Adesman A: Long-term neuro- immunity. Adv Immunol 1967; 6: 337–459. developmental benefits of breastfeeding. Curr 21 Cukrowska B, Sinkora J, Rehakova Z, et al: Iso-

Opin Pediatr 2016; 28: 559–566. type and antibody specificity of spontaneously 11 Porcherie A, Gilbert FB, Germon P, et al: IL-17A formed immunoglobulins in pig fetuses and is an important effector of the immune response germ-free piglets: production by CD5– B cells.

of the mammary gland to Escherichia coli infec- Immunology 1996; 88: 611–617.

tion. J Immunol 2016; 196: 803–812.

46 Tlaskalová-Hogenová/Kverka/Hrdý 22 Avrameas S: Autopolyreactivity confers a holistic 27 Moles JP, Tuaillon E, Kankasa C, et al: Breastmilk role in the immune system. Scand J Immunol cell trafficking induces microchimerism-mediat-

2016; 83: 227–234. ed immune system maturation in the infant.

23 Tlaskalova-Hogenova H, Mandel L, Stepankova Pediatr Allergy Immunol 2018; 29: 133–143. R, et al: Autoimmunity: from physiology to pa- 28 Molitor-Dart ML, Andrassy J, Haynes LD, Bur thology. Natural antibodies, mucosal immunity lingham WJ: Tolerance induction or sensitization and development of B cell repertoire. Folia Biol in mice exposed to noninherited maternal anti-

(Praha) 1992; 38: 202–215. gens (NIMA). Am J Transplant 2008; 8: 2307– 24 Pribylova J, Krausova K, Kocourkova I, et al: Co- 2315. lostrum of healthy mothers contains broad spec- 29 Hrdy J, Novotna O, Kocourkova I, Prokesova L: trum of secretory IgA autoantibodies. J Clin Im- Cytokine expression in the colostral cells of

munol 2012; 32: 1372–1380. healthy and allergic mothers. Folia Microbiol

25 Wirt DP, Adkins LT, Palkowetz KH, et al: Acti- (Praha) 2012; 57: 215–219. vated and memory T lymphocytes in human 30 Hrdy J, Novotna O, Kocourkova I, Prokesova L:

milk. Cytometry 1992; 13: 282–290. The effect of the colostral cells on gene expres- 26 Bertotto A, Vagliasindi C, Gerli R, et al: Soluble sion of cytokines in cord blood cells. Folia Micro-

CD30 antigen in human colostrum. Biol Neonate biol (Praha) 2017; 62: 479–483.

1997; 71: 69–74.

Immunomodulation by Human Milk 47 Immunology of Milk and Lactation

Published online: March 27, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 48–58 (DOI: 10.1159/000505578)

Breastfeeding, a Personalized Medicine with Influence on Short- and Long-Term Immune Health

Valérie Verhasselt School of Molecular Sciences, University of Western Australia, Perth, WA, Australia

Abstract The neonatal immune system has its own reactivity, constraints, and challenges, which profoundly differ from the adult. Breast milk is most probably a key requirement both for optimal immune function in early life and for imprinting of the immune system for long-term immune health. Here, we will highlight how breast milk fills the needs and the gaps of the developing immune system and thereby represents the unbeatable way to prevent infectious disease. We will further focus on some factors in breast milk that we extensively studied and found to actively influence the immune trajectory and long-term immune health. More specifically, we will review how the presence of allergens in breast milk together with maternal milk cofactors such as TGF-β, vitamin A, and immunoglobulins influence mucosal immunity in early life with long-term effects on allergic disease susceptibility. We will see that, depending on the content and the nature of allergens in breast milk as well as the presence of immune modulators, very different outcomes are observed, ranging from protection to an increased allergy risk. We are starting to decipher the specific requirements for the neonatal immune system to function optimally. We are discovering how breast milk fulfills these requirements and guides immune trajectories from early life. Answering these questions will provide the infant with preventive and curative approaches that are tailored to this very specific period of life and will ensure long-term immune health. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Genetic Programming, Environment, and Growth Render Infancy a Period of High Susceptibility to Infectious and Allergic Diseases

As compared to adulthood, early postnatal life is a period that is characterized by rapid changes and multiple immune challenges. The neonates’ tissues are constantly changing due to the growth process. Therefore, the immune system has to respect growth constraints in this period of life, and strong inflammatory immune responses, such as Th1 immune responses, should be avoided as they may lead to scars. Furthermore, tissue development requires the secretion of cytokines such as IL-33, which are also involved in the differentiation of Th2 immunity and allergic immune responses [1]. Thus, in early life, tissue growth results in the predisposition to mount low inflammatory and pro-Th2 immune responses, which can explain the propensity of the neonate to be susceptible to infectious diseases (lack of strong inflammatory responses) and to allergic diseases (bias towards Th2 immune responses). Another characteristic of early life is that the immune system is developing, and gut microbiota, which are a major promotor of immune development [2, 3], are establishing. Genetically programmed immune system development and lack of microbiota in early life result in immune deficiency, where the levels of mucosal antibodies and immune cells in the tissues are much lower than in the adult [4, 5]. The neonate is also exposed to multiple new antigens, which are found in the environment, are present in the diet, or are associated with gut microbiota colonization. With all these antigens being new to the immune system, immune memory in early life is much lower than in the adult [6–8]. Naïve immune responses are known to be slower and less intense than memory ones, which also contribute to higher susceptibility to infection. Finally, yet importantly, requirements for immune activation in early life are different from adulthood. Elegant studies demonstrated that neonates are fully able to mount a cytotoxic immune response towards a viral infection when exposed to a dose that is 10,000 times lower than in the adult [9]. This has important implications for vaccination strategies and highlight the necessity of good hygiene especially in early life to prevent infection. Altogether, these observations indicate that for multiple reasons, early-life immune responses tend to be much less efficient in the fight against pathogens than in adulthood, and neonatal infection remains a common tragedy, with about 7 million cases and 700,000 deaths per year, accounting for 40% of mortality in those under 5 years of age [10]. Immune homeostasis also requires immune tolerance, an immune response that is characterized by the absence of inflammation and largely mediated by regulatory T cells. The induction of regulatory immune responses towards innocuous antigens, such as self-antigens and exogenous antigens derived from the diet, is critical to avoid autoimmune and allergic diseases, respectively. A pioneer study by

Breastfeeding: Personalized Medicine with Influence on Immune Health 49 Billingham et al. [11] demonstrated that immune tolerance is easily induced upon systemic antigen exposure in utero and early postnatal life, and 70 years later, studies are still ongoing to decipher the mechanisms underlying these characteristics [12]. However, the high prevalence of allergic diseases in infancy also indicates that mechanisms of tolerance towards exogenous innocuous antigens upon exposure through the skin and/or the mucosa are defective [13]. Instead, Th2 immune responses are preferentially induced [13]. As discussed here above, the liberation of mediators necessary for tissue growth contributes to this predisposition. The exposure to allergens through a skin barrier that is not fully developed can also favor allergic sensitization [14]. Furthermore, there is strong evidence from mouse models that the neonatal period is refractory to oral tolerance induction [15–19], the process by which immune tolerance to an antigen is acquired following its exposure through the digestive tract [20]. The spontaneous resolution of the majority of infant food allergies suggests a physiological maturation of the mechanisms of oral immune tolerance during the first years of life [13, 21]. Gut colonization with microbiota plays certainly a major role in the setup of immune tolerance and the inhibition of allergic responses. This is clear from studies in germ-free mice, which have elevated levels of serum IgE (a hall- mark of allergic immune responses) and are refractory to oral tolerance [22, 23]. Importantly, gut colonization with microbiota must occur in early postnatal life to efficiently regulate IgE immune responses [23], highlighting the concept of a window of opportunity to influence long-term immune responses [24]. There is accumulating evidence that the specific composition and the metabolites produced by the gut microbiota are critical for the generation of regulatory T cells and oral tolerance to be fully effective [2, 3, 20, 25]. In summary, the immune system in early life is not the one of a “small adult” and neither immature nor tolerance prone. The neonatal immune system is different from the adult one with specific requirements for activation and regulation. In the absence of specific interventions, the infant is highly susceptible to both infectious and allergic diseases. Infant vaccination to elicit memory immune responses and maternal vaccination during pregnancy to increase the levels of circulating antibodies during the first months are possible intervention targets, which are successful to a certain extent, to decrease morbidity and mortality due to infectious diseases in early childhood [10, 26]. Changes in allergy prevention guidelines and earlier introduction of potential allergens in the diet are promising strategies for food allergy prevention, with some limitations as we will discuss later on. A natural intervention does also exist, i.e., breastfeeding. We can distinguish the influence of breastfeeding on short-term immune outcomes, i.e. its effects while the infant is breastfed, and on long-term immune outcomes, i.e., months to years after breastfeeding has ceased (Fig. 1).

50 Verhasselt Immune system function in early life: sIgA, antimicrobial molecules, prebiotics, probiotics, leucocytes, growth factors -> Major decrease in susceptibility to infectious disease in early life Immune system trajectory Microbiota-shaping molecules, allergens, and microbial antigens, immune modulators -> Variable susceptibility to Disease risk allergy and infection

Time

Fig. 1. Short- and long-term effects of breast milk on immune outcomes. We can distinguish the influence of breastfeeding on short-term immune outcomes, i.e., its effects while the infant is breastfed, and on the long-term immune outcomes, i.e., months to years after breastfeeding has ceased. The major short-term effect of breast milk is infectious disease prevention, which relies on multiple factors with various mechanisms of actions that are briefly summarized here. Breast milk also contains various amounts of factors, which influence immune system development and reactivity in the long term and thereby immune- related susceptibility to diseases such as allergy and infection.

Breast Milk as a Physiological and Critical Strategy to Prevent Infectious Diseases in the Short Term

For years, breast milk was considered mainly as a source of nutrients for the developing child. The extensive observations that breastfeeding affords protection towards infectious diseases and could reduce the mortality rate of common infections by more than half have added another key role to breastfeeding [27]. A recent meta-analysis concludes that scaling up breastfeeding to a near universal level could prevent 823,000 annual deaths in children younger than 5 years, mostly due to infectious disease. This protection relies in great part on the transfer of mucosal immunity through breast milk, which importantly relies on multiple, various, and adapting mechanisms (Fig. 1). Breast milk contains potent antibiotics, such as lactoferrin, lysozyme, and antimicrobial peptides [27, 28]. The presence of prebiotics such as human milk oligosaccharides [29] and probiotics [30, 31] interferes with pathogenic bacterial expansion and invasion. Growth factors such as EGF and TGF-β contribute to maintain and repair potential pathogen-induced mucosal barrier breaks. Importantly, secretory IgA (SIgA) delivers personalized medicine. SIgA is present at high concentrations in colostrum (∼10 g/L) and at somewhat lower concentrations in mature breast milk (∼1 g/L). SIgA is specific for intestinal and respiratory pathogens in the maternal environment due to the selective migration of B cells originating from the maternal mu-

Breastfeeding: Personalized Medicine with Influence on Immune Health 51 cosa to the mammary gland [32]. They are thus providing mucosal immunity against pathogens, which are specifically found in the environment of the child. By their non-antigen-specific part, maternal SIgA will also contribute to the establishment of infant gut microbiota and protect the child form pathogens [33].

Breast Milk as a Key Player in the Education of the Immune System and Long-Term Susceptibility to Immune-Dependent Diseases

In addition to the provision of passive immunity and compensation for neonatal immune deficiencies, breast milk actively influences the development of the immune system (Fig. 1). Thereby, breast milk can guide immune trajectories and long-term susceptibility to diseases, which have an immune component in their physiopathology. There is evidence that breastfeeding decreases the risk of obesity and metabolic complications associated with obesity, such as type 2 diabetes [34]. However, there is lack of consistency regarding the possibility of long-term prevention of infectious diseases [35] and allergy [34, 36–40] by breastfeeding. Despite this, there are accumulating data indicating that breast milk has the potential to prevent these diseases by at least two major ways: by shaping the infant microbiota and by exposing the neonatal immune system to microbial antigens and allergens. The gut microbiota, and more specifically the microbiota-derived metabolites, are key for influencing the balance between health and disease [41]. In particular, gut microbiota influences susceptibility to infection and efficacy of vaccination [42], as well as susceptibility to allergy as we recently reviewed [2]. By the presence of prebiotics, probiotics, and antigen-specific and non-antigen- specific antimicrobial compounds, breastfeeding plays a major role in the initial seeding of the infant gut microbiota and in its constant evolution in postnatal life [2]. The major event being known to affect microbiota composition is cessation of breastfeeding [43]. Therefore, we can speculate that the various concentrations of microbiota-shaping compounds in each mother’s breast milk will contribute to the heterogeneity in gut microbiota found in children and thereby influence their susceptibility to disease [2]. Finding which microbiota is best adapted to each environment and how to influence microbiota-shaping molecules in breast milk will open up new avenues for long-term immune health. Breast milk also contains exogenous antigens, and our research in the last de- cade has been aimed at deciphering the long-term outcomes of the presence of allergens in breast milk on allergy susceptibility in the offspring. Recently, we expanded this research to infectious disease prevention. In the last part of this review, we will synthesize our main findings related to the presence of exogenous antigens in breast milk and discuss their implications for long-term immune health.

52 Verhasselt Allergy is a rising public health issue, with respiratory and food allergy affecting up to 20% of children (the Global Asthma Report 2018; http:// globalasthmareport.org/) [13]. After an era of avoidance, there has been a paradigm shift, and new strategies for allergy prevention aim now at inducing tolerance by antigen exposure. Following impressive results in large randomized clinical trials, the early introduction of eggs and peanuts in the diet is now recommended for food allergy prevention [13]. However, recent studies also demonstrated a significant proportion of infants already have egg sensitization and clinical reactivity (including anaphylaxis) prior to the first introduction of eggs into their solid food diet [13, 44–46]. This underscores the necessity to identify earlier and safer ways to promote oral tolerance development to food allergens in young infants, which may be particularly challenging knowing the refractori- ness of early life to oral tolerance. We proposed the hypothesis that early oral exposure to exogenous allergens in the presence of maternal milk immune modulators would alleviate oral tolerance induction in early life. Our experiments in mouse experimental settings demonstrated the key findings: 1. Mice exposed to only low amounts (ng/mL) of the egg-derived allergen ovalbumin (OVA) through breastfeeding are protected from allergic reactions to OVA when reexposed in adulthood via the oral (as a model of food allergy) [47] or respiratory route (as a model for respiratory allergy) [48–50]. This echoes to observations referenced here above indicating that the early-life immune system reacts to very low amounts of antigens. Importantly, other reports have shown the presence of dietary antigens in breast milk in the same range at very low concentrations (ng/mL), such as cow milk β-lactoglobulin [51–53], peanut allergens (Ara h 1 and Ara h 2) [54, 55], or wheat antigen (gliadin) [56]. By comparison, it is worth noting that allergen levels in cow’s milk are found in the range of mg/ mL. Thus, through breast milk, infants are exposed to a wide variety of dietary allergens at low concentrations while formula-fed infants are only exposed to cow milk-derived allergens and at much higher concentrations (unless hydrolyzed formulas are used). We can speculate this will lead to very different outcomes in terms of oral tolerance induction and allergy prevention in the offspring. Our recent observation in a human birth cohort demonstrated that the risk of egg allergy in children was reduced by a factor of 4 at 2.5 years when comparing children exposed to breast milk with or without egg antigen [57]. To fully confirm that OVA in breast milk is responsible for a decreased egg allergy risk in children, further randomized controlled trials will need to be conducted. The next steps will be to identify maternal interventions leading to the consistent presence of OVA in breast milk in order to move towards a successful prevention of egg allergy by breastfeeding.

Breastfeeding: Personalized Medicine with Influence on Immune Health 53 2. Successful oral tolerance and allergy prevention upon OVA exposure through breast milk required the concomitant presence of maternal immunomodulatory factors. We identified TGF-β, OVA-specific IgG, and vitamin A to stimulate oral tolerance induction in the offspring. Each cofactor was shown to act by different mechanisms of action. TGF-β promoted the induction of OVA- specific Th1 cells, which counteract allergic Th2 responses [48, 58]. Vitamin A accelerated gut epithelium maturation resulting in a stronger barrier in the first week of life [50]. It also promoted the expression of RALDH in small-intestine dendritic cells, which was associated with their increased efficiency at activating T lymphocytes [50]. Maternal vitamin A supplementation resulted in the possibility to induce oral tolerance to OVA in breast milk in the offspring from birth, which otherwise is only efficient from the third week of life [50]. OVA- specific IgG in breast milk was necessary for a protected transfer of OVA though the neonatal gut barrier and the induction of a prolonged and strong protection from allergy, which was found to be mediated by FoxP3 Tregs [59]. Recently, another group confirmed these findings and showed that OVA-specific IgG also promoted antigen presentation by neonatal dendritic cells [60]. In other words, early life is characterized by a relative lack of TGF-β in mucosal tissue, a physiological deficiency in vitamin A, and low mucosal and systemic immunoglobulin secretion, which contribute to the lack of oral tolerance induction in early life in the absence of breast milk. Breast milk is providing infants with these cofactors, which will affect gut epithelium barrier integrity, and antigen transfer and presentation for successful regulatory immune response induction. This will result in a long-term low risk for allergic disease as we showed for egg allergy both in an experimental mouse model and in a human birth cohort. Altogether, these data suggest that, according to the levels of immune modulators in breast milk, oral tolerance to dietary antigens in breast milk will be induced with more or less efficiency, which will condition long-term susceptibility to allergy. 3. Unexpectedly, we found allergens from respiratory sources such as from the house dust mites Dermatophagoides pteronyssinus (Der p) and Blomia tropi- calis in breast milk in similar amounts as dietary antigen [61, 62]. Since we detected Der p 1 in the digestive fluid of healthy adults [63], we propose that respiratory allergens are ingested by being trapped in the oropharynx or pulled back by the mucociliary epithelium, and follow the same route as dietary antigens to the mammary gland. In contrast to the observation with OVA, the transfer of Der p 1 through breast milk induced Th2 immune response priming and increased susceptibility to allergic disease in adult mice [64]. Importantly, in a human birth cohort, there was a positive association between allergic sensitization and respiratory allergies in children and the presence of Der p 1 levels in breast milk [61]. This

54 Verhasselt observation stresses that not all the antigens in breast milk induce oral tolerance, even though they reach the gut mucosa together with breast milk-tolerogenic factors. Our most recent findings further showed that Der p allergen in maternal milk abolished the capacity of neonatal mice to mount oral tolerance to bystander OVA egg antigen, which resulted in an increased risk of egg food allergy in the long term [65]. Importantly, human data showed heterogeneity in breast milk content in Der p and OVA. Based on the presence of Der p and/or OVA in breast milk, we could identify groups of lactating mothers which mirror the ones found in mice responsible for the different egg allergy risk [65]. These observations stress the need to identify how to counteract deleterious actions of some allergens in breast milk, such as those derived from house dust mite allergens, which prime for allergic sensitization to themselves and break oral tolerance induction to bystander ones. We have started to identify some potential targets as we showed that protease from Der p allergens are key players for the induction of gut Th2 mucosal immune im- balance in mice breastfed by mothers inhaling Der p [65]. Ongoing research is aimed at identifying how to modulate Der p levels and their enzymatic activity in breast milk and/or in the breastfed child. Ultimately, this should contribute to ensure food allergy prevention in children. 4. Based on the findings that house dust mite allergens in breast milk could prime immune reactivity in the long term, we proposed that novel strategies of early life prevention of infectious disease may take advantage of the possibility to stimulate antigen-specific immune responses through breast milk. Mi- crobial antigen transfer through breast milk would be a way to naturally vacci- nate the infant [26, 66]. Some evidence in the literature supports this hypothesis, such as the observation that maternal HIV infection of noninfected breastfed children is associated with infant stimulation of IgG and IgA secretion in their gut mucosa [67] and HIV-specific interferon-γ-secreting PBMC are found in 50% of cases [68]. Recently, we addressed the original hypothesis that the presence of malaria antigen in breast milk may stimulate antimalarial immune defenses and reduce the malaria risk in breastfed infants [69]. As a first critical step to address this hypothesis, we investigated whether Plasmodium falciparum his- tidine-rich protein 2 (pHRP-2) and lactate dehydrogenase (pLDH) are detectable in breast milk of mothers from Uganda, a country with endemic malaria [70]. We found that 15% of breast milk samples from mothers with asymptom- atic malaria do contain malaria antigens. Our preliminary data indicate that blood levels of malaria antigens determine their levels in breast milk. These landmark findings may have significant implications for susceptibility to malaria in children from endemic countries since malaria antigens in breast milk may strongly influence the immune responses to natural malaria infections and to malaria vaccines in breastfed children [26].

Breastfeeding: Personalized Medicine with Influence on Immune Health 55 Concluding Remarks and Perspectives

The way breast milk composition constantly adapts to the needs of each infant in each setting is fascinating. This results in a major success for the prevention of many infectious diseases in breastfed infants. There is also evidence that breastfeeding could have long-term protective effects on infections, even after breastfeeding has ceased, and this may open new avenues for infant vaccination. Unfortunately, and maybe unexpectedly, breastfeeding does not provide a consistent protection from allergy. A hypothesis behind this lack of protection may be that the lifestyle of modern societies has resulted in changes in breast milk composition, which is not suitable anymore for allergy prevention. Research conducted to identify which factors in breast milk condition protection versus susceptibility is providing new clues, which ultimately will result in the protection of breastfed children from allergy.

Disclosure Statement

Valérie Verhasselt has no conflict of interest with regard to the writing of this chapter.

References

1 de Kleer IM, Kool M, de Bruijn MJ, et al: Perina- 8 Thome JJ, Bickham KL, Ohmura Y, et al: Early- tal activation of the interleukin-33 pathway pro- life compartmentalization of human T cell differ- motes type 2 immunity in the developing lung. entiation and regulatory function in mucosal and

Immunity 2016; 45: 1285–1298. lymphoid tissues. Nat Med 2016; 22: 72–77. 2 van den Elsen LWJ, Garssen J, Burcelin R, Ver- 9 Sarzotti M, Robbins DS, Hoffman PM: Induction hasselt V: Shaping the gut microbiota by breast- of protective CTL responses in newborn mice by

feeding: the gateway to allergy prevention? Front a murine retrovirus. Science 1996; 271: 1726–

Pediatr 2019; 7: 47. 1728. 3 van den Elsen LWJ, Rekima A, Verhasselt V: Ear- 10 Kollmann TR, Kampmann B, Mazmanian SK, et ly-life nutrition and gut immune development. al: Protecting the newborn and young infant

Nestle Nutr Inst Workshop Ser 2019; 90: 137–149. from infectious diseases: lessons from immune

4 Torow N, Marsland BJ, Hornef MW, Gollwitzer ontogeny. Immunity 2017; 46: 350–363. ES: Neonatal mucosal immunology. Mucosal Im- 11 Billingham RE, Brent L, Medawar PB: Actively

munol 2017; 10: 5–17. acquired tolerance of foreign cells. Nature 1953;

5 Renz H, Brandtzaeg P, Hornef M: The impact of 172: 603–606. perinatal immune development on mucosal ho- 12 Ng MSF, Roth TL, Mendoza VF, et al: Helios en- meostasis and chronic inflammation. Nat Rev hances the preferential differentiation of human

Immunol 2011; 12: 9–23. fetal CD4(+) naive T cells into regulatory T cells.

6 Li N, van Unen V, Abdelaal T, et al: Memory Sci Immunol 2019; 4:eaav5947. CD4(+) T cells are generated in the human fetal 13 Sicherer SH, Sampson HA: Food allergy: a review

intestine. Nat Immunol 2019; 20: 301–312. and update on epidemiology, pathogenesis, diag- 7 Zhang X, Mozeleski B, Lemoine S, et al: CD4 T nosis, prevention, and management. J Allergy

cells with effector memory phenotype and func- Clin Immunol 2018; 141: 41–58. tion develop in the sterile environment of the

fetus. Sci Transl Med 2014; 6: 238ra72.

56 Verhasselt 14 Du Toit G, Sampson HA, Plaut M, et al: Food 30 Rautava S: Milk microbiome and neonatal coloni- allergy: update on prevention and tolerance. J zation – overview. Nestle Nutr Inst Workshop

Allergy Clin Immunol 2018; 141: 30–40. Ser 2020; 94. DOI: 10.1159/000505030. 15 Strobel S: Neonatal oral tolerance. Ann NY Acad 31 Fernández L, Rodriguez JM: Human milk micro-

Sci 1996; 778: 88–102. biota: origin and potential uses. Nestle Nutr Inst

16 Strobel S: Immunity induced after a feed of anti- Workshop Ser 2020; 94. DOI: 10.1159/000505031. gen during early life: oral tolerance v. sensitisa- 32 Brandtzaeg P: Mucosal immunity: integration

tion. Proc Nutr Soc 2001; 60: 437–442. between mother and the breast-fed infant.

17 Strobel S, Ferguson A: Immune responses to fed Vaccine 2003; 21: 3382–3388. protein antigens in mice. 3. Systemic tolerance or 33 Dunne-Castagna VP, Mills DA, Lönnerdal B: Ef- priming is related to age at which antigen is first fects of milk secretory immunoglobulin A on the

encountered. Pediatr Res 1984; 18: 588–594. commensal microbiota. Nestle Nutr Inst Work-

18 Hanson DG: Ontogeny of orally induced toler- shop Ser 2020; 94. DOI: 10.1159/000505335. ance to soluble proteins in mice. I. Priming and 34 Victora CG, Bahl R, Barros AJ, et al: Breastfeed-

tolerance in newborns. J Immunol 1981; 127: ing in the 21st century: epidemiology, mecha-

1518–1524. nisms, and lifelong effect. Lancet 2016; 387: 475– 19 Miller A, Lider O, Abramsky O, Weiner HL: Oral- 490. ly administered myelin basic protein in neonates 35 Hanson LA, Korotkova M, Lundin S, et al: The primes for immune responses and enhances ex- transfer of immunity from mother to child. Ann

perimental autoimmune encephalomyelitis in NY Acad Sci 2003; 987: 199–206.

adult animals. Eur J Immunol 1994; 24: 1026– 36 Dogaru CM, Nyffenegger D, Pescatore AM, et al: 1032. Breastfeeding and childhood asthma: systematic

20 Pabst O, Mowat AM: Oral tolerance to food pro- review and meta-analysis. Am J Epidemiol 2014;

tein. Mucosal Immunol 2012; 5: 232–239. 179: 1153–1167. 21 Wang J, Ji H: Influence of probiotics on dietary 37 Lodge CJ, Tan DJ, Lau MX, et al: Breastfeeding protein digestion and utilization in the gastroin- and asthma and allergies: a systematic review and

testinal tract. Curr Protein Pept Sci 2019; 20: 125– meta-analysis. Acta Paediatr 2015; 104: 38–53. 131. 38 Gungor D, Nadaud P, LaPergola CC, et al: Infant 22 Sudo N, Sawamura S, Tanaka K, et al: The re- milk-feeding practices and food allergies, allergic quirement of intestinal bacterial flora for the de- rhinitis, atopic dermatitis, and asthma through- velopment of an IgE production system fully sus- out the life span: a systematic review. Am J Clin

ceptible to oral tolerance induction. J Immunol Nutr 2019; 109(suppl 7): 772S–799S.

1997; 159: 1739–1745. 39 Greer FR, Sicherer SH, Burks AW: Committee on 23 Cahenzli J, Koller Y, Wyss M, et al: Intestinal mi- Nutrition, Section on Allergy and Immunology: crobial diversity during early-life colonization The effects of early nutritional interventions on shapes long-term IgE levels. Cell Host Microbe the development of atopic disease in infants and

2013; 14: 559–570. children: the role of maternal dietary restriction, 24 Renz H, Adkins BD, Bartfeld S, et al: The neona- breastfeeding, hydrolyzed formulas, and timing tal window of opportunity-early priming for life. of introduction of allergenic complementary

J Allergy Clin Immunol 2018; 141: 1212–1214. foods. Pediatrics 2019; 143:e20190281. 25 Rooks MG, Garrett WS: Gut microbiota, metabo- 40 Muraro A, Halken S, Arshad SH, et al: EAACI

lites and host immunity. Nat Rev Immunol 2016; food allergy and anaphylaxis guidelines. Primary

16: 341–352. prevention of food allergy. Allergy 2014; 69: 590– 26 Marchant A, Sadarangani M, Garand M, et al: 601. Maternal immunisation: collaborating with 41 Hand TW, Vujkovic-Cvijin I, Ridaura VK,

mother nature. Lancet Infect Dis 2017; 17:e197– Belkaid Y: Linking the microbiota, chronic dis- e208. ease, and the immune system. Trends Endocrinol

27 Labbok MH, Clark D, Goldman AS: Breastfeed- Metab 2016; 27: 831–843. ing: maintaining an irreplaceable immunological 42 Lynn DJ, Pulendran B: The potential of the mi-

resource. Nat Rev Immunol 2004; 4: 565–572. crobiota to influence vaccine responses. J Leuko-

28 Lawrence RM, Pane CA: Human breast milk: cur- cyte Biol 2018; 103: 225–231. rent concepts of immunology and infectious dis- 43 Backhed F, Roswall J, Peng Y, et al: Dynamics and eases. Curr Probl Pediatr Adolesc Health Care stabilization of the human gut microbiome dur-

2007; 37: 7–36. ing the first year of life. Cell Host Microbe 2015;

29 Bode L: Human milk oligosaccharides: structure 17: 852. and functions. Nestle Nutr Inst Workshop Ser 44 Bellach J, Schwarz V, Ahrens B, et al: Random-

2020; 94. DOI: 10.1159/000505339. ized placebo-controlled trial of hen’s egg consumption for primary prevention in infants. J

Allergy Clin Immunol 2017; 139: 1591–1599.e2.

Breastfeeding: Personalized Medicine with Influence on Immune Health 57 45 Palmer DJ, Metcalfe J, Makrides M, et al: Early 58 Turfkruyer M, Verhasselt V: Breast milk and its regular egg exposure in infants with eczema: a impact on maturation of the neonatal immune

randomized controlled trial. J Allergy Clin system. Curr Opin Infect Dis 2015; 28: 199–206.

Immunol 2013; 132: 387–392.e1. 59 Mosconi É: Mise en évidence des propriétés im- 46 Palmer DJ, Sullivan TR, Gold MS, et al: Random- muno-régulatrices des complexes immuns du lait ized controlled trial of early regular egg intake to maternel: implications dans la prévention des

prevent egg allergy. J Allergy Clin Immunol 2017; maladies allergiques; thesis. Nice, 2010.

139: 1600–1607.e2. 60 Yoshida M, Claypool SM, Wagner JS, et al: Hu- 47 Rekima A, Macchiaverni P, Turfkruyer M, et al: man neonatal Fc receptor mediates transport of Long-term reduction in food allergy susceptibil- IgG into luminal secretions for delivery of anti-

ity in mice by combining breastfeeding-induced gens to mucosal dendritic cells. Immunity 2004;

tolerance and TGF-beta-enriched formula after 20: 769–783.

weaning. Cli Exp Allergy 2017; 47: 565–576. 61 Baiz N, Macchiaverni P, Tulic MK, et al: Early 48 Verhasselt V, Milcent V, Cazareth J, et al: Breast oral exposure to house dust mite allergen through milk-mediated transfer of an antigen induces breast milk: a potential risk factor for allergic tolerance and protection from allergic asthma. sensitization and respiratory allergies in children.

Nat Med 2008; 14: 170–175. J Allergy Clin Immunol 2017; 139: 369–372.e10. 49 Mosconi E, Rekima A, Seitz-Polski B, et al: Breast 62 Macchiaverni P, Ynoue LH, Arslanian C, et al: milk immune complexes are potent inducers of Early exposure to respiratory allergens by placen-

oral tolerance in neonates and prevent asthma tal transfer and breastfeeding. PLoS One 2015;

development. Mucosal Immunol 2010; 3: 461– 10:e0139064. 474. 63 Tulic MK, Vivinus-Nebot M, Rekima A, et al: 50 Turfkruyer M, Rekima A, Macchiaverni P, et al: Presence of commensal house dust mite allergen Oral tolerance is inefficient in neonatal mice due in human gastrointestinal tract: a potential con- to a physiological vitamin A deficiency. Mucosal tributor to intestinal barrier dysfunction. Gut

Immunol 2016; 9: 479–491. 2016; 65: 757–766. 51 Fukushima Y, Kawata Y, Onda T, Kitagawa M: 64 Macchiaverni P, Rekima A, Turfkruyer M, et al: Consumption of cow milk and egg by lactating Respiratory allergen from house dust mite is women and the presence of beta-lactoglobulin present in human milk and primes for allergic and ovalbumin in breast milk. Am J Clin Nutr sensitization in a mouse model of asthma. Al-

1997; 65: 30–35. lergy 2014; 69: 395–398. 52 Host A, Husby S, Hansen LG, Osterballe O: Bo- 65 Rekima A, Bonnart C, Macchiaverni P, et al: A vine beta-lactoglobulin in human milk from role for early oral exposure to house dust mite atopic and non-atopic mothers. Relationship to allergens through breast milk in IgE-mediated maternal intake of homogenized and unhomog- food allergy susceptibility. J Allergy Clin Immu-

enized milk. Clin Exp Allergy 1990; 20: 383–387. nol 2020. DOI: 10.1016/j.jaci.2019.12.912. 53 Sorva R, Makinen-Kiljunen S, Juntunen-Back- 66 Verhasselt V: Is infant immunization by breast- man K: Beta-lactoglobulin secretion in human feeding possible? Philos Trans R Soc Lond B Biol

milk varies widely after cow’s milk ingestion in Sci 2015; 370: 20140139. mothers of infants with cow’s milk allergy. J Al- 67 Moussa S, Jenabian MA, Gody JC, et al: Adaptive

lergy Clin Immunol 1994; 93: 787–792. HIV-specific B cell-derived humoral immune 54 Vadas P, Wai Y, Burks W, Perelman B: Detection defenses of the intestinal mucosa in children ex-

of peanut allergens in breast milk of lactating posed to HIV via breast-feeding. PLoS One 2013;

women. JAMA 2001; 285: 1746–1748. 8:e63408. 55 Bernard H, Ah-Leung S, Drumare MF, et al: Pea- 68 John-Stewart GC, Mbori-Ngacha D, Payne BL, et nut allergens are rapidly transferred in human al: HV-1-specific cytotoxic T lymphocytes and breast milk and can prevent sensitization in mice. breast milk HIV-1 transmission. J Infect Dis

Allergy 2014; 69: 888–897. 2009; 199: 889–898. 56 Chirdo FG, Rumbo M, Anon MC, Fossati CA: 69 van den Elsen LWJ, Verhasselt V, Egwang T: Ma- Presence of high levels of non-degraded gliadin laria antigen shedding in the breast milk of in breast milk from healthy mothers. Scand J mothers from a region with endemic malaria.

Gastroenterol 1998; 33: 1186–1192. JAMA Pediatr 2020. DOI: 10.1001/jamapediat- 57 Verhasselt V, Genuneit J, Metcalfe JR, et al: Oval- rics.2019.5209. bumin in breast milk is associated with a de- 70 World Health Organization: World Malaria creased risk of IgE-mediated egg allergy in chil- Report 2018. https://www.who.int/malaria/ dren. Allergy 2019. DOI: 10.1111/all.14142. publications/world-malaria-report-2018/en/.

58 Verhasselt Immunology of Milk and Lactation

Published online: March 11, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 59–64 (DOI: 10.1159/000505425)

Summary on Immunology of Milk and Lactation

The information summarized in the various comprehensive presentations in this workshop represented a diverse spectrum of historical, evolutionary, and functional aspects of mammalian lactation and the process of breastfeeding. This workshop was dedicated to Prof. Lars A. Hanson (MD, PhD) for his outstanding contributions to the understanding of the biology of milk and the dissemination of knowledge on breastfeeding to advance current practices of breastfeeding in the contemporary human society worldwide. The dedication ceremony was followed by scientific presentations in session I of the workshop with the keynote addressed by Olav T. Oftedal. Oftedal provided an elegant perspective of the evolution of lactation in different mammalian species. Based on studies on synapsids (ancestral to mammals, which appear to have diverged from sauropsids [ancestral to crocodiles, lizards, and birds]), he proposed that lactation may have first evolved as a source of moisture and antimicrobial compounds for parchment-shelled eggs, followed by the evolution of some skin secretions, which eventually became milk. It was suggested that among basal animals (monotremes), each mammary gland develops as a triad in association with a hair follicle and sebaceous gland as a mammo- pilo-sebaceous unit (MPSU). In other mammalian species, such as marsupials, there is a similar triad, but the hair follicles are shed during development. In the diverse group of eutherian mammals, some show no association with the mammary hair, while others, such as the horse, develop as MPSU with mammary hair and sebaceous glands present in the mammary gland. The MPSU also bears significant resemblance to apocrine glands (APSU), suggesting that mammary glands may have also evolved from an APSU-type structure. Recent studies have suggested that most constituents of mammalian milk are unique and found only in mammary secretions. He proposed that if a milk protein occurs in the milk of monotremes, marsupials, and eutherians, the major mammalian taxa, then the protein must have evolved before the groups diverged and are inherited from the ancestral taxa. These observations have provided unique and new insights into the genetic origin and functions of specific mammary constituents in the products of lactation. Oftedal briefly alluded to the 4 primary types of caseins, members of the secretory calcium-binding phosphoproteins (SCPPs), as an evolutionary challenge because of their diversity and the large size of the micelles in milk. These proteins have an ancient history in the evolution of mineralized tissues. Based on related SCPP genes, caseins may have evolved as protolacteal secretion that delivered calcium to eggs. Finally, his presentation discussed briefly the evolution of the milk fat globule membrane, lactose, and other neutral and acidic oligosaccharides. The next presentation provided a brief historical overview of the immunology of milk and mammalian lactation. This presentation served as an introduction to the subsequent specific topics discussed in this workshop. Mother’s milk has been considered a complete food for the infant from times immemorial, and it has been associated with unique healing powers and beneficial effects. These include cure for insomnia, loss of appetite, ascites, piles, skin disorders, sexual dysfunction, muscle weakness, contraception, and prevention of cancer and infections. Breastfeeding developed a spiritual and religious importance in the Middle Ages in Europe, as evidenced by the deep faith and respect for Nursing Madonna, Virgin Mary, and the breastfed Jesus. The modern history of breast milk immunology can be traced to a publication by Paul Ehr lich as early as 1892 and subsequent demonstration of specific maternal antibody transport to the colostrum and milk. The immunologic composition of human milk and its biologic linkage to mucosa-associated lymphoid tissue was initially recognized by Gugler and Von Muralt, and by Lars Hanson. These elegant studies were followed by the identification of secretory IgA in human external secretions by Chodirkar and Tomasi, and Bienenstock and Tomasi, and in the human milk by Hanson and Johansson. Subsequent studies by Beer and Bellingham, Ogra et al., Mohr, and Okamoto and others identified several cellular and soluble immunologic factors in the human milk and their transport to the suckling neonate via the process of breastfeeding. It is now known that human colostrum and milk contains a wealth of immunologically active products derived from the innate and adoptive immunologic, microbiologic, dietary, and other maternal experiences in the maternal mucosal surfaces, es-

60 Ogra pecially the gut, and the maternal circulation. This historical review was dedicated to the memory of Dr. S.S. Ogra, the principal investigator of most milk- related research carried out in her laboratory in the early 1970s and 1980s in the School of Medicine at the University at Buffalo. This presentation briefly reviewed lactation performance and the presence and function of diverse soluble elements detected in mammalian colostrum and milk to date. These included: secretory IgA and other immunoglobulin isotypes, antisecretory factor, soluble CD14, and soluble Toll-like receptors, as well as several cytokines and lymphokines. It also introduced the role of colostrum- and milk-associated cellular components, such as leukocytes, macrophages, epithelial cells, stem cells, and T lymphocytes, and cell-mediated immune responses. This overview also summarized earlier studies on the transfer of tuberculin-specific maternal cellular immunity to the neonate via breastfeeding and more recent investigations on the transfer of maternal cellular immunity and engulfment of maternal DNA via the transfer of leukocytes and stem cells. Finally, the risks and benefits of the colostrum and milk to the neonate and the developing infant were briefly considered here. Detailed discussion of the issues identified here follows in subsequent presentations in this session and sessions II and III of this workshop. Jiri Mestecky reviewed in some detail the evidence for the existence of mucosa-associated lymphoid tissue and common mucosal immune sites for effective immunization in the mucosal system, and the importance of mammary glands as an integral component of the common mucosal immune system. He discussed recent studies on the structure, biologic activities, and the spectrum of antibodies of the IgA isotype specific for microbial, dietary, and other environmental antigens and macromolecules in the colostrum and milk. He concluded his presentation by identifying possible directions for future investigations in the immunobiology of the mammary gland and lactation. These include the routes for the most effective induction of IgA responses in milk, the identification of phenotypes of B lymphocytes that express homing receptors for the mammary gland, and the determination of effective timing for maternal immunization to provide optimal levels of protective immune reactivity in the colostrum and the milk for the neonate. Helena Tlaskalová-Hogenová, Miloslav Kverka, and Jiří Hrdý introduced the wide spectrum of immunomodulatory components present in human milk and colostrum, including those of innate and adaptive immunity, and factors influencing the composition and colonization of newborn gut microbiota. They discussed the nature of autoantibodies and the spectrum of newly detected cytokines and lymphokines in human milk. The presentation was completed with an overview of different cellular components of and cytokine gene expression on

Summary of Session I 61 colostral cells in healthy and allergic mothers. Studies carried out to date have identified over 35 cytokines in the colostrum and milk, and some of them have been identified for the first time in human milk. Their possible functions include the development of intestinal lymphoid tissue, functional development of the gut structure, angiogenesis, central and enteric nervous system development, and establishment of immunologic homeostasis in the mammary gland as well as in the breastfeeding neonate. Valerie Verhasselt discussed the influence of breastfeeding on the development of immunologic health in the breastfed neonate and infant. She began with the examination of the unique specificities of the neonatal immune system, including unique limitations to the development of immune responses after postnatal exposure to environmental antigens. She reviewed the role of TGF-β, vitamin A, several environmental allergens, and specific antibodies in the context of early life and long-term allergic disease susceptibility. Based on controlled epidemiological data and several experimental studies, she proposed that early-life oral exposure to allergens does not induce tolerance but may prime for allergic responses. It has been suggested that nonbreastfed infants are exposed to only few allergens, but in very high concentrations, such as β-lactoglobulins. On the other hand, breastfed infants are exposed to a wide variety of allergens in the maternal milk and colostrum but in extremely low concentrations. Additionally, breast milk provides the infant with significant amounts of TGF-β, vitamin A, and other cofactors which affect the integrity of the barrier of gut epithelium and regulate antigen transfer and presentation to the mucosa-associated lymphoid tissue. As a result, breastfeeding results in a low risk for allergic disorders in the long term. Such conclusions are supported by recent studies in human birth cohorts and by studies carried out in her lab- oratories with the induction of egg allergy in experimental (mouse) animal models. Carine Blanchard made the final presentation in session I. Due to certain un- avoidable circumstances, she could not provide a full-length manuscript of her presentation. Therefore, a more detailed summary of her talk is presented here. Her presentation focused on the immunologic evaluation of human milk oligosaccharides (HMO) with respect to disease expression in the neonate after exposure to allergens and infectious agents. She discussed in some detail the enzyme fucosyltransferase (FUT) and its genotypes FUT2 and FUT3. These enzymes are expressed on blood groups ABH and Lewis, intestinal mucosa, and other human body fluids. Recent studies have suggested that the early trajectory of neonatal microbial colonization is significantly influenced by the number of environmental factors. These include gestational age of the neonate, method of delivery, use of antibiotics, geographic location of birth, genetics, maternal stage

62 Ogra of lactation, maternal diet, and specific immunologic components delivered to the neonate via breastfeeding. These factors appear to determine the outcome of colonization and the composition of the neonatal microbiome as healthy or aberrant based on the metabolites generated by the microbiome. An aberrant microbiome has been associated with the development of sustained inflammation, induction of asthma, atopy, obesity, inflammatory bowel disease, and other disease states. Employing FUT2 and FUT3 genotypes as proxy for the HMOs, several ongoing investigations have provided important information on the role of HMOs in human milk and colostrum; 1. 2 -Fucosylated HMOs in human milk alleviate the negative effects of cesarean section on infant gut microbiota. ′ 2. HMOs in infant formula significantly improve the outcome of infections in infants. 3. Maternal FUT2- and FUT3-positive status is related to a lower risk of respiratory infections during the first 6 months of neonatal life. 4. Maternal HMOs are associated with the prevention of colonization and growth of the pathogenic microbiome. 5. Elevated serum IgE levels are associated with the absence of gut microbiota in experimental models of infection. She summarized the results of the LIFE child cohort studies from Leipzig (Germany) and Bangkok (Thailand) and other investigations involving cholera toxin/ovalbumin-induced food allergy in experimental animal models These studies have also demonstrated that the use of HMO and FUT2/FUT3 genotypes in infant feeding is associated with significantly decreased allergic sensitization. The mechanisms underlying such protection appear to be related to the modulation of regulatory T-cell function by HMOs and independent of the regular prebiotic effects associated with milk oligosaccharides and other soluble products in the milk. Based on the information summarized above from the presentations in session I of the workshop, it is apparent that we have come a long way in understanding the evolutionary biology of mammalian lactation, the presence and role of specific innate and adaptive immunologic mechanisms associated with human milk, and the impact of breastfeeding on the systemic and mucosal immunologic development in the neonate. Recent information about the microbiology of the milk and lactation and its influence on gut colonization, presented in session II, and studies about the role of HMOs and other soluble components of the colostrum and milk, presented in session III, are summarized next by W. Allan Walker and Bo Lönnerdal, respectively. It is gratifying to note the wealth of new information presented in this workshop, and it is hoped to generate further interest in exploring many unanswered

Summary of Session I 63 questions related to the mammary glands and lactation and its impact on the neonate. Finally, it would be appropriate to conclude this summary by recapitu- lating the statement made by Frank Oski (MD) several decades ago:

Imagine that the world had created a new dream product to feed and immunize everyone born on earth. Imagine also that it was available everywhere, required no storage or delivery and helped mothers plan their families and reduce the risk of cancer. Then Imagine that the world refused to use it. Frank Oski (1932–1996)

Pearay L. Ogra

64 Ogra Microbiology of Milk and Lactation: Influence on Gut Colonization

Published online: April 1, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 65–74 (DOI:10.1159/000505030)

Milk Microbiome and Neonatal Colonization: Overview

Samuli Rautava Department of Pediatrics and Adolescent Medicine, University of Turku and Turku University Hospital, Turku, Finland

Abstract Breastfeeding confers the infant short- and long-term health benefits and significantly modulates the developing infant gut microbiome. A specific human milk microbiome has relatively recently been discovered, but its origin remains poorly understood. Data from experimental and clinical studies suggest that the bacteria in milk may originate in the maternal gut and be transported via a specific enteromammary pathway, the details of which have not been elucidated yet. The milk microbiome is affected by the maternal metabolic state, antibiotic use, as well as the mode of delivery. We are only in the initial stages of understanding the biological function of the milk microbiome and its potential contribution to infant gut colonization. Several clinical studies indicate, however, that despite considerable differences in the overall composition of the milk and infant gut microbiomes, specific bacteria are detectable both in human milk and infant feces, and that the bacteria in milk are a source of microbes colonizing the neonatal gut. If the microbes in human milk are discovered to contribute to the beneficial effects of breastfeeding, modulating or mimicking the milk microbiome may provide a novel means of improving child health. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel

Introduction

Breastfeeding not only provides the neonate and infant with optimal nutrition, it also confers protection against acute and chronic diseases [reviewed in 1]. It has been estimated that, in the developing world, the risk of death of exclusively breastfed infants is only 12% of that of formula-fed infants [1]. Breastfeeding could potentially prevent half of all diarrheal diseases and one third of respiratory tract infections in infancy, which would correspond to an estimated 823,000 lives saved annually [1]. These protective effects are thought to be mediated by a variety of antimicrobial compounds present in human milk, albeit the fact that exclusively breastfed infants are less likely to be exposed to potentially contami- nated food and water probably also plays a role. Interestingly, in addition to protection against infectious disease in infancy, breastfeeding has been associated with long-term health benefits, including a reduced risk of developing chronic diseases in later life [1].

Breastfeeding and the Risk of Noncommunicable Diseases in Later Life

It is important to recognize the problems inherent in studies assessing the long- term health impact of breastfeeding. Performing randomized controlled studies comparing breastfeeding of different durations with formula feeding on the individual level is not possible for practical and ethical reasons. The results of epidemiological studies must be interpreted with caution because of potential confounding by factors including maternal obesity and mode of delivery, which are known or suspected to affect both breastfeeding rates [2, 3] and chronic disease risk in the child [4, 5]. Furthermore, the increased risk of infectious disease in infancy related to formula feeding may result in more frequent antibiotic use, which in turn has been suggested to increase the risk of childhood overweight [6] and asthma [7]. Many of the available epidemiological studies rely on parental reports of breastfeeding duration and childhood diseases, which together with potential recall bias in retrospective studies further decreases the reliability of the data. Despite the methodological difficulties discussed above, there are convincing data to suggest that breastfeeding confers health benefits, which extend beyond the period of breastfeeding and even into adulthood. It has been estimated that breastfeeding reduces the risk of overweight and obesity by 13–26% and that of type II diabetes mellitus by 35%, but no effects on the other components of the metabolic syndrome, including hypertension and hypercholesterolemia, have been reliably documented [1]. The data regarding the associations between breastfeeding and the risk of allergic diseases, asthma, and type I diabetes are somewhat inconsistent. The mechanisms between breast milk exposure and reduced risk of noncommunicable immune-mediated or inflammatory diseases are currently not well understood. Human milk contains a large variety of immunoactive molecules,

66 Rautava such as hormones, growth factors, cytokines, and chemokines, as well as maternal lymphocytes, which may and, to some extent, have been shown to modulate mucosal and systemic immune responses in the infant and induce immune maturation [reviewed in 8]. The individual variation in immune molecule concentrations in the milk associated with maternal health or the metabolic state may explain the somewhat discrepant data regarding breastfeeding and chronic disease risk. Future research will hopefully shed more light on these interesting associations and the underlying mechanisms.

Breastfeeding and the Infant Gut Microbiome

The developing intestinal microbiome is a potential mediator of the long-term health effects of breastfeeding. Early gut colonization and the establishment and development of the gut microbiome during the first year of life are a dynamic process which is profoundly affected by breastfeeding and exposure to formula and solid foods [reviewed in 9]. Disturbances in the intestinal microbiome in early life have been linked with later development of chronic noncommunicable diseases, including overweight and obesity [10] as well as allergic disease and asthma [11, 12]. Since breastfeeding has also been suggested to modulate the risk of these disorders, it is warranted to speculate whether causal connections exist between breast milk, the gut microbiome, and later disease. The gut microbiome of exclusively breastfed infants is drastically distinct from that of formula-fed infants of the same age. The initial neonatal gut microbiome in healthy term newborns is characterized by Escherichia coli, enterococci, streptococci, and clostridia, which are soon followed by anaerobes and particularly members of the genera Bifidobacterium and Bacteroides [13, 14]. Sev- eral studies indicate that bifidobacteria dominate the gut microbiome of breastfed infants [15–17], whereas that of formula-fed infants exhibits greater diversity [18]. Indeed, both the diversity and composition of the infant gut microbiome are affected by the intake of breast milk in a dose-dependent manner [19]. This has been observed also after the introduction of solid foods, and it appears that the cessation of breastfeeding modulates the infant gut microbiome more profoundly than complementary feeding [17]. Human milk contains a vast variety of molecules, which may influence the developing infant gut microbiome. The protection against infectious diseases alluded to above is mediated at least partially by IgA antibodies, Toll-like receptors, complement, and other antimicrobial proteins and peptides present in human milk [reviewed in 9], and it is likely that they elicit their effects on the potentially colonizing and indigenous as well as pathogenic bacteria. In addition

Milk Microbiome and Neonatal Colonization 67 to antibacterial components, human milk contains substances which selectively promote the growth of specific bacteria. A considerable body of scientific evidence suggests that human milk oligosaccharides (HMOs) are the most significant modulators of the developing infant gut microbiome present in human milk [reviewed in 20, 21]. HMOs are a large group of molecularly diverse non- digestable carbohydrates, the structure and function of which are discussed in detail elsewhere in this volume. One of the most comprehensively characterized biological effects of HMOs in breastfed infants is providing substrates and hence a survival benefit to specific members of the intestinal microbiome and particularly bifidobacteria. Whether individual differences in milk HMO composition resulting from maternal genetics, metabolic state, or disease are reflected in the effect of breastfeeding on gut microbiota or child health is the focus of future research.

The Human Milk Microbiome

Breastfeeding may influence the developing infant gut microbiome by directly inhibiting or promoting the growth of bacteria or by modulating intestinal immune function. In addition, breast milk harbors live bacteria, which are thought to provide a colonizing inoculum for the infant gut. We are currently only in the initial stages of understanding the composition, origin, and significance of the human milk microbiome, which are discussed in detail elsewhere in this volume. It is not always clear to what extent the microbes detected in human milk samples reflect the bacterial population in the mammary epithelium and whether the microbes detected in milk originate from the maternal skin or even the infant mouth. Nonetheless, bacterial DNA has been detected in surgically obtained samples from nonlactating mammary gland specimens [22], which has been interpreted to suggest that the human mammary epithelium harbors a distinct microbiome. As of present, more than 200 bacterial species have been detected in human milk samples [21]. Interestingly, the bacterial taxa in human milk are distinct from those encountered on the areolar skin, and gut-associated obligate anaerobes such as bifidobacteria are often detected [21, 23]. This may result from differences pertaining to the exposure to oxygen and other environmental factors, but also raises the question of the origin of the mammary gland and human milk microbiome. Relatively few studies have systematically approached this question. Several species of obligate anaerobic bacteria, including members of the genera Bifidobacterium, Bacteroides, Parabacteroides, and the clostridial family, usually thought to be characteristic of the human gut microbiome, were detected in both milk and fecal samples in a study of 7 lactating women [23].

68 Rautava Interestingly, in clinical trials, specific probiotic lactobacilli have been recovered in human milk after oral consumption by the mother [24, 25]. While these data do not provide insight into the mechanisms of microbial transfer from the gut to the breast milk, and contact contamination is possible, it is intriguing to speculate that a mechanism exists by which enteral bacteria are transported to the mammary gland and then milk. This notion is consistent with data from an elegant series of studies by Perez et al. [26], according to which increased bacterial translocation from the gut to mesenteric lymph nodes was detected in the perinatal period in mice. Furthermore, B. longum DNA was detected in the maternal gut microbiome, circulating immune cells, and milk, and finally also in infant feces [26]. These data indeed suggest that enteromammary transfer of microbes takes place at least during lactation. It may be hypothesized that this process is triggered by labor, since the human milk microbiome is reportedly different in mothers who have given birth by elective cesarean section delivery [27, 28].

The Human Milk Microbiome and Infant Gut Colonization

Dissecting the contribution of the human milk microbiome to infant gut colonization is challenging. Human gut colonization is a stepwise process, which is influenced by a series of exposures during early life [reviewed in 9]. In recent years, it has been suggested that microbes may be present already in the intrauterine environment, and that gut colonization may begin during fetal life [29, 30]. Whether an intrauterine microbiome exists during healthy pregnancy and whether fetal microbial colonization takes place remain open questions and active areas of debate and research. It is well established, however, that the neonate receives an important inoculum of maternal vaginal and intestinal bacteria during delivery. The importance of this early massive bacterial exposure is highlighted by studies showing differences in gut microbiome composition in children born by cesarean section and thus not exposed to the maternal microbiome days, weeks, months, and even years after the fact [reviewed in 9]. Whether the differences in the milk microbiome between mothers who have delivered by cesarean section or vaginal delivery play a role in the aberrant gut colonization in infants born by cesarean section is currently not known. As discussed above, breastfeeding is the single most important determinant of gut colonization patterns after delivery and during infancy. The role of breast milk bacteria as a source of colonizing bacteria is a new and largely uncharted area of research (Table 1). In a study of 15 infants born by elective cesarean section and their mothers [30], the colostrum microbiome was observed to be

Milk Microbiome and Neonatal Colonization 69 Table 1. Selected studies investigating the association between the human milk microbiome and infant gut colonization

Study n Method Results

Collado 15 16S The neonatal gut microbiome shifted during the first et al. sequencing week of life to resemble more closely that in colostrum [30] Grönlund 61 qPCR Maternal Bifidobacterium frequencies and counts in et al. feces but not in milk correlated with those detected in [32] the infant gut at 1 month of age Jost 7 Several, Members of Bifidobacterium, Bacteroides, et al. e.g., culture Parabacteroides, Blautia, Clostridium, Collinsella, and [23] and 16S Veillonella were detected in maternal feces, milk, and sequencing neonatal feces Viable Bifidobacterium breve was detected in maternal feces and milk as well as in infant feces Pannaraj 107 16S Infant fecal microbiome at the median age of 40 days et al. sequencing resembled more closely the microbiome in their own [19] mother’s milk as compared to nonrelated mothers As per source tracking analysis, 15% of the fecal microbiome was derived from the bacteria in milk Williams 21 16S The milk and infant fecal microbiomes have some et al. sequencing similarity in early life but become increasingly different [33] over time Source tracking analyses indicated that on day 2 of life, milk microbes contribute 4.9% to the infant gut microbiome, but this diminishes to 0.3% at 6 months of age

dominated by members of Enterobacteriaceae and streptococci. The dominant bacterial family detected in the meconium of these cesarean section-delivered neonates was staphylococci. However, in the newborn fecal samples obtained later in the first week of life, the relative abundance of staphylococci decreased and that of streptococci and Veillonellaceae increased. Moreover, using an un- weighted UniFrac distance matrix, the neonatal gut microbiome began to resemble the maternal milk microbiome during the first week of life. Even though not direct proof, these data may be interpreted to support the notion that microbes in milk might colonize the newborn gut. This notion is corroborated by a recent study according to which several bacterial taxa, particularly those belonging to the genera Streptococcus and Staphylococcus, are detected in preco- lostral samples collected before delivery and oral samples from the neonates. Importantly, the design of the study with milk collection before the child is born excludes the possibility of the milk bacteria originating from the infant mouth [31].

70 Rautava The contribution of human milk bacteria to infant gut colonization may indirectly be assessed based on reports on the similarities between the maternal milk and the infant gut microbiome (Table 1). In a study of 61 mother-infant pairs, bifidobacteria were analyzed by PCR from milk and infant fecal samples obtained at the age of 1 month [32]. In addition, the same analyses were performed on maternal fecal samples collected during the last trimester of pregnancy. Maternal fecal Bifidobacterium adolescentis and B. bifidum frequencies and counts correlated significantly with those detected in their infants’ gut. B. longum was the most frequently detected Bifidobacterium species in human milk, but no correlation was observed between human milk and infant gut bifidobacteria in total or at species level. According to these data, at least in the case of bifidobacteria, the mother is an important source of gut bacteria to the infant, but other routes of transfer may exceed human milk in importance. Further insight into the associations between the milk microbiome and infant gut colonization was provided by Jost et al. [23], who systematically studied the microbial composition of maternal feces and milk and serial infant fecal samples in 7 mother-infant pairs using both culture-independent and -dependent techniques. Bacteria belonging to the genera Bifidobacterium, Bacteroides, Parabacte- roides, Blautia, Clostridium, Collinsella, and Veillonella were observed to be detectable in all 3 sample types (maternal feces, milk, and neonatal feces). Further- more, viable Bifidobacterium breve was detected in maternal feces and milk as well as infant feces. These data lead the authors to conclude that their results support the hypothesis of enteromammary transfer of bacteria influencing infant gut colonization. Even more compelling yet still indirect evidence for the role of human milk microbes in infant gut colonization was provided by a recent study by Pan- naraj et al. [19], who collected and analyzed areolar, milk, and infant fecal samples from 107 mother-infant pairs. The infant gut microbiome assessed by fecal samples obtained at the median age of 40 days resembled more closely the microbial profiles detected in their own mother’s milk as compared to nonrelated mothers. It is important to note, however, that the milk and infant fecal microbiomes were clearly distinct from each other. Based on source-tracking analyses, approximately 15% of the fecal microbiome in predominantly breastfed infants originated from the bacteria in milk and 10% from the areolar skin during the first 30 days of life. Recently published data from serial fecal and milk sampling from 21 mother-infant pairs suggest that the human milk microbiome may contribute to infant oral colonization but less to gut colonization [33]. The milk and infant gut microbiomes displayed some similarities in the neonatal period with source-tracking analyses suggesting an approximately 5% contribution from milk bacteria to the neonatal gut microbiome during the first days of life. Later in infancy, the role of milk microbes in infant gut colonization patterns was found to be negligible.

Milk Microbiome and Neonatal Colonization 71 Table 2. Factors affecting the human milk microbiome

Exposure/factor Impact on the human milk microbiome

Excessive weight gain staphylococci ↑ during pregnancy [34] bifidobacteria ↓ Akkermansia muciniphila-type bacteria ↑ Cesarean section diversity ↓ delivery [27] richness ↓ bifidobacteria ↓ staphylococci ↑ Intrapartum antibiotics diversity ↑ [28] richness ↑ bifidobacteria ↓ Human milk oligosaccharide bifidobacteria with increased HMO total ↑ profile concentration and sialylated HMOs Akkermansia muciniphila with increased fucosylated ↑ human milk oligosaccharides

Clinical Significance of the Human Milk Microbiome

The clinical significance of human milk bacteria and their contribution to the developing infant gut microbiome is currently not well understood. Recent advances in research indicate, however, that maternal characteristics and exposures affect the bacterial population in human milk (Table 2). Maternal overweight or excessive weight gain have been reported to be associated with an altered milk microbiome [27, 34]. Interestingly, the HMO composition of human milk has been reported to be associated with the microbial profile in milk [35]. As alluded to above, the milk microbiome of mothers who have given birth by cesarean section is significantly different from those who deliver vaginally [27, 28]. Furthermore, maternal intrapartum antibiotic administration has recently been reported to affect the bacterial composition of human milk 1 month after delivery [28]. Interestingly, intrapartum antibiotics are also associated with increased abundance of antibiotic resistance genes in the neonatal gut microbiome, but this potentially detrimental consequence of antibiotic exposure appears to be ameliorated by breastfeeding, possibly at least in part via modulation of the taxonomic composition of the infant gut microbiome in favor of bifidobacteria [36]. Given the established significance of early gut colonization patterns to later health and the potential contribution of the bacteria in milk to the developing infant gut microbiome, future research should focus on establishing whether the milk microbiome is associated with health outcomes. Experimental and clinical

72 Rautava studies are needed to elucidate the contribution of the bacteria in milk to the beneficial effects of breastfeeding. In the future, modulating or mimicking the human milk microbiome may offer a means to influence early gut colonization and improve child health.

Disclosure Statement

The author declares no conflict of interest.

References

1 Victora CG, Bahl R, Barros AJ, et al; Lancet 10 Kalliomäki M, Collado MC, Salminen S, Isolauri Breastfeeding Series Group: Breastfeeding in the E: Early differences in fecal microbiota composi- 21st century: epidemiology, mechanisms, and tion in children may predict overweight. Am J

lifelong effect. Lancet 2016; 387: 475–490. Clin Nutr 2008; 87: 534–538. 2 Huang Y, Ouyang YQ, Redding SR: Maternal pre- 11 Abrahamsson TR, Jakobsson HE, Andersson AF, pregnancy body mass index, gestational weight et al: Low gut microbiota diversity in early infan- gain, and cessation of breastfeeding: a systematic cy precedes asthma at school age. Clin Exp Al-

review and meta-analysis. Breastfeed Med 2019; lergy 2014; 44: 559–566.

14: 366–374. 12 Fujimura KE, Sitarik AR, Havstad S, et al: Neo 3 Hobbs AJ, Mannion CA, McDonald SW, et al: The natal gut microbiota associates with childhood impact of caesarean section on breastfeeding ini- multisensitized atopy and T cell differentiation.

tiation, duration and difficulties in the first four Nat Med 2016; 22: 1187–1191. months postpartum. BMC Pregnancy Childbirth 13 Favier CF, de Vos WM, Akkermans AD: Develop-

2016; 16: 90. ment of bacterial and bifidobacterial communi-

4 Voerman E, Santos S, Patro Golab B, et al: Mater- ties in feces of newborn babies. Anaerobe 2003; 9: nal body mass index, gestational weight gain, and 219–229. the risk of overweight and obesity across child- 14 Jost T, Lacroix C, Braegger CP, Chassard C: New hood: an individual participant data meta-analy- insights in gut microbiota establishment in

sis. PLoS Med 2019; 16:e1002744. healthy breast fed neonates. PLoS One 2012; 5 Sevelsted A, Stokholm J, Bønnelykke K, Bisgaard 7:e44595. H: Cesarean section and chronic immune disor- 15 Harmsen HJ, Wildeboer-Veloo AC, Raangs GC, et

ders. Pediatrics 2015; 135:e92–e98. al: Analysis of intestinal flora development in 6 Rasmussen SH, Shrestha S, Bjerregaard LG, et al: breast-fed and formula-fed infants by using mo- Antibiotic exposure in early life and childhood lecular identification and detection methods. J

overweight and obesity: a systematic review and Pediatr Gastroenterol Nutr 2000; 30: 61–67.

meta-analysis. Diabetes Obes Metab 2018; 20: 16 Roger LC, Costabile A, Holland DT, et al: Exami- 1508–1514. nation of faecal Bifidobacterium populations in 7 Ahmadizar F, Vijverberg SJH, Arets HGM, et al: breast- and formula-fed infants during the first

Early life antibiotic use and the risk of asthma 18 months of life. Microbiology 2010; 156: 3329– and asthma exacerbations in children. Pediatr 3341.

Allergy Immunol 2017; 28: 430–437. 17 Stewart CJ, Ajami NJ, O’Brien JL, et al: Temporal 8 Rautava S, Walker WA: Academy of Breastfeed- development of the gut microbiome in early

ing Medicine founder’s lecture 2008: breastfeed- childhood from the TEDDY study. Nature 2018;

ing – an extrauterine link between mother and 562: 583–588.

child. Breastfeed Med 2009; 4: 3–10. 18 Azad MB, Konya T, Persaud RR, et al; CHILD 9 Rautava S: Microbial composition of the initial Study Investigators: Impact of maternal intrapar- colonization of newborns. Nestlé Nutr Inst Work- tum antibiotics, method of birth and breastfeed-

shop Ser 2017; 88: 11–21. ing on gut microbiota during the first year of life:

a prospective cohort study. BJOG 2016; 123: 983– 993.

Milk Microbiome and Neonatal Colonization 73 19 Pannaraj PS, Li F, Cerini C, et al: Association be- 28 Hermansson H, Kumar H, Collado MC, et al: tween breast milk bacterial communities and Breast milk microbiota is shaped by mode of de- establishment and development of the infant gut livery and intrapartum antibiotic exposure. Front

microbiome. JAMA Pediatr 2017; 171: 647–654. Nutr 2019; 6: 4. 20 Bode L: Human milk oligosaccharides: every 29 Aagaard K, Ma J, Antony KM, et al: The placenta

baby needs a sugar mama. Glycobiology 2012; 22: harbors a unique microbiome. Sci Transl Med

1147–1162. 2014; 6: 237ra65. 21 Jost T, Lacroix C, Braegger C, Chassard C: Impact 30 Collado MC, Rautava S, Aakko J, et al: Human of human milk bacteria and oligosaccharides on gut colonisation may be initiated in utero by dis- neonatal gut microbiota establishment and gut tinct microbial communities in the placenta and

health. Nutr Rev 2015; 73: 426–37. amniotic fluid. Sci Rep 2016; 6: 23129. 22 Hieken TJ, Chen J, Hoskin TL, et al: The microbi- 31 Ruiz L, Bacigalupe R, García-Carral C, et al: Mi- ome of aseptically collected human breast tissue crobiota of human precolostrum and its potential

in benign and malignant disease. Sci Rep 2016; 6: role as a source of bacteria to the infant mouth.

30751. Sci Rep 2019; 9: 8435. 23 Jost T, Lacroix C, Braegger CP, et al: Vertical 32 Grönlund MM, Gueimonde M, Laitinen K, et al: mother-neonate transfer of maternal gut bacteria Maternal breast-milk and intestinal bifidobacte-

via breastfeeding. Environ Microbiol 2014; 16: ria guide the compositional development of the 2891–2904. Bifidobacterium microbiota in infants at risk of

24 Jiménez E, Fernández L, Maldonado A, et al: Oral allergic disease. Clin Exp Allergy 2007; 37: 1764– administration of Lactobacillus strains isolated 1772. from breast milk as an alternative for the treat- 33 Williams JE, Carrothers JM, Lackey KA, et al: ment of infectious mastitis during lactation. Appl Strong multivariate relations exist among milk,

Environ Microbiol 2008; 74: 4650–4655. oral, and fecal microbiomes in mother-infant 25 Abrahamsson TR, Sinkiewicz G, Jakobsson T, et dyads during the first six months postpartum. J

al: Probiotic lactobacilli in breast milk and infant Nutr 2019; 149: 902–914. stool in relation to oral intake during the first 34 Collado MC, Laitinen K, Salminen S, Isolauri E:

year of life. J Pediatr Gastroenterol Nutr 2009; 49: Maternal weight and excessive weight gain dur- 349–354. ing pregnancy modify the immunomodulatory

26 Perez PF, Doré J, Leclerc M, et al: Bacterial im- potential of breast milk. Pediatr Res 2012: 72: 77– printing of the neonatal immune system: lessons 85.

from maternal cells? Pediatrics 2007; 119:e724– 35 Aakko J, Kumar H, Rautava S, et al: Human milk e732. oligosaccharide categories define the microbiota 27 Cabrera-Rubio R, Mira-Pascual L, Mira A, Col- composition in human colostrum. Benef Mi-

lado MC: Impact of mode of delivery on the milk crobes 2017; 8: 563–567. microbiota composition of healthy women. J Dev 36 Pärnänen K, Karkman A, Hultman J, et al: Mater-

Orig Health Dis 2016; 7: 54–60. nal gut and breast milk microbiota affect infant gut antibiotic resistome and mobile genetic ele-

ments. Nat Commun 2018; 9: 3891.

74 Rautava Microbiology of Milk and Lactation: Influence on Gut Colonization

Published online: March 13, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 75–85 (DOI: 10.1159/000505031)

Human Milk Microbiota: Origin and Potential Uses

a b Leónides Fernández Juan M. Rodríguez a Department of Galenic Pharmacy and Food Technology, Complutense University of Madrid, Madrid, b Spain; Department of Nutrition and Food Science, Complutense University of Madrid, Madrid, Spain

Abstract At the beginning of the 21st century, some pioneer studies provided evidence of the existence of a site-specific human milk microbiota. Hygienically collected milk samples from healthy women contain a relatively low bacterial load, which consist mostly of Staphylococ- cus, Streptococcus, lactic acid bacteria, and other gram-positive bacteria (Corynebacterium, Propionibacterium, and Bifidobacterium). DNA from strict anaerobic bacteria is also detected in human milk samples. The origin of human milk bacteria still remains largely unknown. Although the infant’s oral cavity and maternal skin may provide microbes to milk, selected bacteria of the maternal digestive microbiota may access the mammary glands through oral- and enteromammary pathways involving interactions with immune cells. In addition, when milk is collected using external devices, such as breast pumps, some microorganisms may arise from unhygienic handling as well as from the water used to clean and rinse the devices, for example. The human milk microbiota has a wide spectrum of potential uses. Most of them have been focused on the infant (including the preterm ones), but some bacterial strains present in human milk have also a big potential to be used to improve the mother’s health, mainly through the prevention or treatment of infectious mastitis during lactation. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Human Milk Is Not Sterile

Until recently, human milk and mammary glands were thought to be sterile under physiological conditions. Therefore, the presence of microbes in such locations was traditionally considered either as an infection or as a contamination. This negative view started to change in 2003 following the publication of 2 ar- ticles describing that human milk may be a source of lactic acid bacteria [1, 2]. Such bacteria were not only generally recognized as safe, they were also considered as beneficial and with key roles in infant gut colonization. This was a relevant finding in the context of increasing awareness of the importance ascribed to human microbiota for our health throughout life. The first article proposing the existence of a site-specific microbiome in human milk was published 1 year later [3]. Nowadays, it is generally accepted that human milk contains its own microbiota, with an increasing recognition of its role in early gut colonization in infants. In the last 15 years, the number of studies on the composition of the human milk microbiota has sharply increased [4, 5]. Cultivable bacteria in hygienically collected human milk samples are usually dominated by Staphylococcus (mainly S. epidermidis and other coagulase-negative species), Streptococcus (mainly S. mitis spp. and S. salivarius spp.), Corynebacteri- um, Propionibacterium, and other taxonomically related gram-positive bacteria. Lactic acid bacteria and bifidobacteria may also be isolated from human milk but to a lesser extent [4, 6]. Among them, species belonging to the genus Lactobacillus (e.g., L. salivarius, L. gasseri, L. fermentum, and L. reuteri) and Bifidobacterium (e.g., B. longum and B. breve) have attracted particular interest from a scientific, medical, and industrial point of view due to their potential application as probiotics. The bacterial concentration in human milk from healthy women usually 3 ranges from an undetectable level to ∼10 CFU/mL when the samples are collected hygienically by manual expression or by using single-use sterile pump devices. The concentration can be much higher (up to 106 CFU/mL) in mastitis- suffering women [7]. When the milk is collected by pumping, high concentrations of contaminating gram-negative bacteria (e.g., enterobacteria, Pseudomo- nas, and Stenotrophomonas) and yeasts may arise from the rinsing water and/or poor hygienic manipulation practices [8]. Traditional culture-dependent methods have the limitation of not being able to assess the presence of viable but noncultivable microbes and, particularly, that of the strict anaerobes, the DNA of which has also been detected in the human milk microbiome. Recent developments in culturomics have shown that if proper culture conditions are provided, most previously unculturable bacteria can be isolated from complex microbial ecosystems. Unfortunately, culturomic approaches have not been applied yet to milk and mammary microbiota.

76 Fernández/Rodríguez The application of the first available culture-independent molecular techniques, including different PCR approaches, provided a complementary assessment of the milk microbiome [4]. Soon they were replaced by high-throughput next-generation sequencing approaches, from metataxonomics (16S rRNA amplicon analysis) to metagenomics (total DNA sequencing). Sequencing of bacterial 16S rRNA genes does neither provide evidence for viability nor for function- ality of detected bacteria, but it is a first step in understanding the complexity of the milk microbial ecosystem and its role in infant gut colonization [5, 9, 10]. Metagenome studies involving shotgun sequencing of milk microbial DNA have been much scarcer [11, 12] and have provided data on the potential presence and roles of other components of the milk microbiome, including viruses, archaea, fungi, and protozoa. The microbiome of human milk seems to be a dynamic and complex ecosystem which is not randomly assembled but forms well-organized bacterial con- sortia and networks, which may be different among different populations [13]. At the genus level, culture-independent studies have confirmed the presence of DNA from bacterial genera previously identified with culture-dependent techniques, such as Staphylococcus, Streptococcus, Corynebacterium, Propionibacte- rium, Lactococcus, Leuconostoc, Bifidobacterium, and Lactobacillus in human milk. In addition, DNA from strict anaerobes, which are typically associated with the gut environment (e.g., Bacteroides, Eubacterium, Faecalibacterium, Roseburia, Ruminococcus, and Veillonella), has also been detected [5, 12]. Such bacteria are often noncultivable, and, in fact, they have not been isolated from human milk yet. Sequences from a third group of bacteria, typically associated with soil and water (Acinetobacter, Methylobacterium, Pseudomonas, Sphingobium, Sphin- gomonas, Stenotrophomonas, or Xanthomonas), are also frequently detected in human milk. However, they may be the result of technical artifacts derived from the presence of DNA from such genera in molecular-biological reagents. In fact, working with samples containing a low microbial biomass, such as human milk under physiological conditions, is a big challenge. Future studies will require suitable controls to determine which data are actually genuine in a given biological sample.

Where Does Human Milk Microbiota Come from?

After birth, the bacterial colonization process represents the first massive contact with microbes; the close link between early gut microbiota composition and the risk of disease later in life underlines the important role of the microbiota-

Origin and Uses of the Human Milk Microbiota 77 Maternal sources

Other sources Water, pumps, shields, partner, Oral other breastfed brother/sister, microbiota clothes, environment…

Translocation?

Breast skin, areola, nipple and Montgomery gland microbiota Milk Infant gut

Translocation?

Infant sources Mouth Nasopharynx Skin Intestinal microbiota

Maternal vaginal, Other sources: oral, intestinal and family, medical staff, skin microbiota water, foods, environment…

Fig. 1. Potential sources of the human milk microbiota. host interactions in early life. Colostrum and milk bacteria are, obviously, among the first colonizers of the infant gut and, therefore, may play a key role in driving the development of its microbiota. Several studies have reported a mother-to- infant transfer of human milk microorganisms (at the species and/or the strain level) using both culture-dependent and -independent techniques [4, 6, 14–16]. Although it has become evident that human milk is a source of infant gut bacteria, the origin of the bacteria present in human milk still remains largely unknown and is the subject of scientific controversy. Traditionally, any bacterial cell present in human milk was considered the result of contamination arising from the infant’s oral cavity or the mother’s skin. However, the detection of live bacterial cells and/or DNA from anaerobic species that are generally related to the gut environment and that do not survive in aerobic locations fueled a scientific debate on the origin of milk-associated bacteria (Fig. 1). Some bacteria from the infant’s oral cavity may contaminate milk during suckling due to milk flowing back into the mammary ducts; however, the presence of oral-related bacteria in milk may precede the first feeding since colos-

78 Fernández/Rodríguez trum collected within 24 h after birth has been found to contain typical oral bacteria like Veillonella, Leptotrichia, and Prevotella [10]. In addition, contamination from the breastfed baby’s oral cavity does not explain why precolostrum secreted by some women before delivery already contains some of the micro organisms that characterizes human milk [3]. A recent study has revealed the presence of typical oral bacteria (such as streptococci) in precolostrum collected during the first pregnancy and before contact with the newborn, a fact that indicates that they are not a contamination from the infant mouth [17]. The fact that some of these strains were shared by the mother’s milk and the infant mouth suggests that at least some oral bacteria reach the infant’s mouth through breastfeeding. Microbes inhabiting maternal skin (particularly the external surfaces closer to the ejected milk: nipples, mammary areolas, and Montgomery glands) may be transferred during ejection. Some bacteria commonly isolated in human milk, such as Propionibacterium acnes, Corynebacterium spp. or, particularly, S. epidermidis, are also common in some regions of the human skin. However, although staphylococci, corynebacteria, and propionibacteria have been traditionally associated with the skin, they are widespread in most, if not all, human mucosal surfaces; in fact, the populations of such bacterial groups reach their highest concentrations in the mucosal layers of the digestive and genitourinary tracts. A recent study comparing the bacterial composition of human milk, areolar skin, and infant stool samples of healthy mother-infant pairs found that the bacterial communities were distinct in these 3 ecosystems, differing in both composition and diversity [16]. Despite sharing some phylotypes, the comparison between the bacterial communities detected in milk and those found on breast skin or in the infant’s mouth reveals major differences between them. Both the skin and the infant’s mouth are highly improbable sources of strict anaerobic bacteria typically associated with the adult gut microbiome. Sharing of Bifidobacterium, Bacteroides, Para- bacteroides, and members of the clostridial class (Blautia, Clostridium, and Col- linsella) between maternal feces, human milk, and neonatal feces has already been reported [5, 18]. The assessment of bifidobacterial communities in several mother-infant pairs through bifidobacterial culturing and profiling analyses using an internal transcribed spacer revealed a large number of bifidobacterial strains that were commonly identified in maternal and infant fecal samples, as well as in the corresponding human milk sample [15]. All these studies reinforce the hypothesis that at least some bacteria, including obligate anaerobes, may be vertically transferred from mother to neonate via breastfeeding. Selected bacteria of the maternal digestive microbiota may access the mammary glands through oral and enteromammary pathways [3, 19, 20]. It has pre-

Origin and Uses of the Human Milk Microbiota 79 viously been observed that certain bacteria from the maternal digestive tract may spread to extradigestive locations in healthy hosts. In addition, some studies have offered a scientific plausible basis for such physiological translocation. The mechanism would involve mononuclear cells (dendritic cells and macrophages), which would be able to take up nonpathogenic bacteria from the gut lumen and, subsequently, carry them to other locations, including the lactating mammary gland [reviewed in 19, 20]. In vitro and in vivo data reinforcing this hypothesis have been obtained by different groups. It must be highlighted that there is an intense efflux of intestinal immune cells to the mammary gland during late pregnancy and lactation and that, in fact, the existence of an enteromammary circulation of IgA-producing cells is long known. An increased bacterial translocation from the gut to the mesenteric lymph nodes and the mammary gland in pregnant and lactating mice was observed by Perez et al. [18]. Bacteria could be observed histologically in the subepithelial dome and interfollicular regions of Peyer’s patches, and in the lamina propria of the small bowel, and associated with cells in the glandular tissue of the mammary gland. The Peyer’s patches of pregnant and lactating mice were macro- scopically larger than those of control animals and had a more prominent subepithelial dome and more dilated draining lymphatic vessels containing mononuclear cells. Similarly, oral administration of the strains Lactobacillus rhamnosus GG and L. gasseri K7 to mice during pregnancy and lactation led to changes in the mesenteric lymph nodes and mammary gland microbiota [21]. Live lactic acid bacteria were detected in blood, mesenteric lymph nodes, and mammary gland, while the Lactobacillus genus was detected exclusively in the mammary gland of the mice that ingested the strains. More recently, de Andrés et al. [22] also tried to elucidate if some lactic acid bacteria were able to translocate and colonize the mammary gland and milk in a murine model. For this purpose, L. lactis MG1614 and L. salivarius PS2 were transformed with a plas- mid containing the lux genes; subsequently, the transformed strains were orally administered to pregnant mice. The murine model allowed the visualization, isolation, and PCR detection of the transformed bacteria in different body locations, including mammary tissue and milk, reinforcing the hypothesis that physiological translocation of maternal bacteria during pregnancy and lactation may contribute to the composition of the mammary and milk microbiota. While some may argue that their presence in milk might be the result of superficial fecal contamination in mice, such a route can hardly explain their isolation and detection from mammary biopsies. In relation to human species, a study cited above [18] showed that human milk contains viable bacteria, including Streptococcus, Lactobacillus, and Bifido- bacterium, while acridine orange staining of milk and blood cytopreparations

80 Fernández/Rodríguez identified bacterial cells in association with maternal mononuclear cells. These results strongly suggest the involvement of mononuclear cells in the transport of intestinal bacteria to the mammary gland in late pregnancy. Other studies have reported that oral administration of L. reuteri, L. gasseri, L. fermentum, and L. salivarius strains isolated from human milk to pregnant or lactating women led to their presence in milk [reviewed in 23]. Gut bacterial translocation has usually been associated with pathogenic conditions, but a low rate of bacterial translocation (involving Bacteroides, lactobacilli, bifidobacteria, or enterococci) occurs in healthy hosts and may be associated with physiological immunomodulation of the infant [18]. Some bacterial strains seem to specifically mediate their own translocation without collateral translocation of other bacteria from the host digestive tract [24]. Many transient anatomical and physiological changes that occur during pregnancy and lactation may favor an increased bacterial translocation during such periods [reviewed in 19, 20]. Further studies are required to elucidate the mechanisms by which some bacterial strains may translocate physiologically in certain hosts or life stages. The existence of such bacterial oral- and enteromammary pathways would provide new opportunities for manipulating altered maternal-fetal microbiota, reducing the risk of preterm birth or infant diseases.

Potential Uses of the Human Milk Microbiota: The Mother’s Side

Up to the present, most of the research on potential applications of the human milk microbiota has been focused on infant health, including that of the preterm neonate. However, human milk bacteria may also be relevant for breast health. The lactating mammary gland ecosystem is hospitable to many microorganisms, including bacterial groups that have the potential to cause mastitis. Upon dis- turbance of this balanced state, infection can occur, and, in fact, some studies have reported that mastitis is a process characterized by a mammary bacterial dysbiosis, including a lower microbial diversity, increased abundance of opportunistic pathogens, and depletion of commensal obligate anaerobes (Fig. 2) [7, 25]. The etiopathogenesis of the different types of lactational mastitis (acute mastitis, subacute mastitis, and granulomatous lobular mastitis), as well as the factors that may predispose to or protect from mastitis, has recently been reviewed [23]. Since the resistance to antibiotics and the ability to drive the formation of biofilms are common properties among mastitis-causing bacteria, many cases are refractory to antibiotic therapy. In this context, the development of new

Origin and Uses of the Human Milk Microbiota 81 Healthy mammary gland

Host factors Infant factors Microbial factors Environmental factors Medical factors Other factors

Predisposing factors Protecting factors

Milk microbiota Low diversity High divesity High concentration Dysbiosis Low concentration

Healthy mammary gland Mastitis Correct treatment

Predisposing factors

Fig. 2. The human milk microbiota: from physiology to dysbiosis and mastitis. strategies for mastitis management based on human milk probiotics, as an alternative or complement to antibiotic therapy, is particularly appealing. To date, oral administration of a few Lactobacillus strains isolated from human milk has proven to be an excellent approach for the treatment and prevention of lactational mastitis [7]. The results of the first placebo-controlled clinical trial aimed to test the potential of the combination of L. salivarius CECT 5713 and L. gasseri CECT 5714 for the treatment of staphylococcal mastitis revealed that they were an efficient alternative for the treatment of lactational mastitis [23], leading to the disap- pearance of clinical symptoms and to significant reductions in milk staphylococcal concentrations. Subsequently, the efficacy of L. fermentum CECT 5716 or L. salivarius CECT 5713 for the same target was evaluated and compared to antibiotic therapy in a clinical trial involving 352 women with infectious mastitis [26]. The probiotic treatment led to significant reductions (1.7–2.1 log10 CFU/mL) in the staphylococcal and/or streptococcal counts in milk and to a rapid improvement in the condition. On the basis of the bacterial counts, pain

82 Fernández/Rodríguez scores and clinical evolution, women in any of the two probiotic groups had a significantly better outcome than those in the antibiotic group. The potential of oral administration of L. salivarius PS2 during pregnancy to prevent mastitis in women who had suffered infectious mastitis after, at least, one previous pregnancy was the subject of a subsequent trial [27]. For this purpose, 108 pregnant women were randomly divided in two groups: (a) a probiotic group, who ingested daily 9 log10 CFU of L. salivarius PS2 from 30 weeks of pregnancy until delivery, and (b) a control group, who received a placebo. The occurrence of mastitis was evaluated during the first 3 months after delivery. At the end of the study, the percentage of women suffering mastitis in the probiotic group (25%) was significantly lower than that in the control group (56%). Therefore, oral administration of such Lactobacillus strain during late pregnancy was an efficient method to prevent mastitis in a susceptible population. A parallel trial involving lactating women with mastitis was carried out in order to identify microbiological, biochemical, and/or immunological biomarkers of the probiotic effect of L. salivarius PS2 [28]. Samples of milk, blood, and urine were collected before and after the probiotic intervention and screened for a wide spectrum of microbiological, biochemical, and immunological parameters. In the mastitis group, L. salivarius PS2 intake led to a reduction in milk bacterial counts, milk and blood leukocyte counts, and IL-8 level in milk, increased levels of IgE, IgG3, EGF, and IL-7, a modification in the milk electrolyte profile, and a reduction in some oxidative stress biomarkers. In the same cohort, the characterization of the urine metabolic profiles at the beginning of the probiotic intervention showed increased energy metabolism (lactate, citrate, formate, acetate, and malonate) and decreased branched- chain amino acid catabolism (isocaproate and isovalerate) when compared to those obtained after the intervention [29]. In addition, probiotic supplementation led to a normalization of breast permeability. Changes in the levels of acetate and 2-phenylpropionate after probiotic intake suggested immunomodulation while an increased malonate level indicated an important antagonistic strategy of L. salivarius PS2 since this catabolite is a well-known repressor of the tricarboxylic acid cycle, which may alter staphylococcal and streptococcal metabolism and negatively affect their survival, virulence, and ability for biofilm formation. Transcriptomic profiling of human milk somatic cells and blood leukocytes was also applied to this cohort to explore potential targets responsive to probiotic intervention [30]. Despite the interindividual variability in the gene expression changes in both types of cells, their results showed the involvement of inflammatory and cell growth-related pathways and genes in the human milk somatic cells following the intake of L. salivarius PS2. Indi- vidual analyses of selected genes supported the upregulation of STC1 and IL19

Origin and Uses of the Human Milk Microbiota 83 and the downregulation of PLAUR and IFNGR1 in somatic cells of the patients as potential targets responsive to the probiotic. Other potential mechanisms by which some Lactobacillus strains may be able to control mastitis-causing agents in the breast after oral administration, including immunomodulation and local competitive exclusion and production of antimicrobials, have been reviewed by Fernández et al. [27]. Mastitis represents the first medical cause of undesired weaning. Therefore, this condition should be considered as a major public health issue since breastfeeding is associated with a reduced risk of many diseases in infants and mothers.

Disclosure Statement and Funding Sources

The authors are co-inventors in patents WO 2004/003235 A2 and PCT/NL2013/050924, involving the application of Lactobacillus strains for the prevention or treatment of mastitis. Work of the research group dealing with the potential origin of the human milk microbiota has been funded by grant AGL2016-75476-R (Spanish Ministry of Economy and Competitiveness).

References

1 Heikkilä MP, Saris PEJ: Inhibition of Staphylo- 9 Hunt KM, Foster JA, Forney LJ, et al: Character coccus aureus by the commensal bacteria of hu ization of the diversity and temporal stability of

man milk. J Appl Microbiol 2003; 95: 471–478. bacterial communities in human milk. PLoS One

2 Martín R, Langa S, Reviriego C, et al: Human 2011; 6:e21313. milk is a source of lactic acid bacteria for the in 10 Cabrera-Rubio R, Collado MC, Laitinen K, et al:

fant gut. J Pediatr 2003; 143: 754–758. The human milk microbiome changes over lacta 3 Martín R, Langa S, Reviriego C, et al: The com tion and is shaped by maternal weight and mode

mensal microflora of human milk: new perspec of delivery. Am J Clin Nutr 2012; 96: 544–551. tives for food bacteriotherapy and probiotics. 11 Ward TL, Hosid S, Ioshikhes I, et al: Human milk

Trends Food Sci Technol 2004; 15: 121–127. metagenome: a functional capacity analysis. BMC

4 Fernández L, Langa S, Martín V, et al: The hu Microbiol 2013; 13: 116. man milk microbiota: origin and potential roles 12 Jiménez E, de Andrés J, Manrique M, et al:

in health and disease. Pharmacol Res 2013; 69: Metagenomic analysis of milk of healthy and

1–10. mastitis-suffering women. J Hum Lact 2015; 31: 5 Jost T, Lacroix C, Braegger C, et al: Impact of hu 406–415. man milk bacteria and oligosaccharides on neo 13 Ma Z, Guan Q, Ye C, et al: Network analysis sug natal gut microbiota establishment and gut gests a potentially “evil” alliance of opportunistic

health. Nutr Rev 2015; 73: 426–437. pathogens inhibited by a cooperative network in

6 Murphy K, Curley D, O’Callaghan TF, et al: The human milk bacterial communities. Sci Rep 2015;

composition of human milk and infant fecal mi 5: 8275. crobiota over the first three months of life: a pilot 14 Asnicar F, Manara S, Zolfo M, et al: Studying ver

study. Sci Rep 2017; 7: 40597. tical microbiome transmission from mothers to 7 Fernández L, Arroyo R, Espinosa I, et al: Pro infants by strain-level metagenomics profiling.

biotics for human lactational mastitis. Benef mSystems 2017; 2:e00164.

Microbes 2014; 5: 169–183. 15 Duranti S, Lugli GA, Mancabelli L, et al: Maternal 8 Jiménez E, Arroyo R, Cárdenas N, et al: Mamma inheritance of bifidobacterial communities and ry candidiasis: a medical condition without scien bifidophages in infants through vertical trans

tific evidence? PLoS One 2017; 12:e0181071. mission. Microbiome 2017; 5: 66.

84 Fernández/Rodríguez 16 Pannaraj PS, Li F, Cerini C, et al: Association be 24 Cronin M, Morrissey D, Rajendran S, et al: Orally tween breast milk bacterial communities and administered bifidobacteria as vehicles for deliv

establishment and development of the infant gut ery of agents to systemic tumors. Mol Ther 2010;

microbiome. JAMA Pediatr 2017; 171: 647–654. 18: 1397–1407. 17 Ruiz L, Bacigalupe R, García-Carral C, et al: Mi 25 Patel SH, Vaidya YH, Patel RJ, et al: Culture inde crobiota of human precolostrum and its potential pendent assessment of human milk microbial

role as a source of bacteria to the infant mouth. community in lactational mastitis. Sci Rep 2017;

Sci Rep 2019; 9: 8435. 7: 7804. 18 Perez PF, Doré J, Leclerc M, et al: Bacterial imprint 26 Arroyo R, Martín V, Maldonado A, et al: Treat ing of the neonatal immune system: lessons from ment of infectious mastitis during lactation: anti

maternal cells? Pediatrics 2007; 119:e724–e732. biotics versus oral administration of lactobacilli

19 Rodríguez JM: The origin of human milk bacte isolated from breast milk. Clin Infect Dis 2010; 50: ria: is there a bacterial entero-mammary pathway 1551–1558. during late pregnancy and lactation. Adv Nutr 27 Fernández L, Cárdenas N, Arroyo R, et al: Pre

2014; 5: 779–784. vention of infectious mastitis by oral administra 20 Mira A, Rodríguez JM: The origin of human milk tion of Lactobacillus salivarius PS2 during late

bacteria; in McGuire M, McGuire MA, Bode L pregnancy. Clin Infect Dis 2016; 62: 568–573. (eds): Prebiotics and Probiotics in Human Milk: 28 Espinosa-Martos I, Jiménez E, de Andrés J, et al: Origins and Functions of Milk-Borne Oligosac Milk and blood biomarkers associated to the clin charides and Bacteria. London, Academic Press, ical efficacy of a probiotic for the treatment of

2016, pp 349–364. infectious mastitis. Benef Microbes 2016; 7: 305– 21 Treven P, Mrak V, Bogovič Matijašić B, et al: Ad 318. ministration of probiotics Lactobacillus rham- 29 Vázquez-Fresno R, Llorach R, Marinic J, et al: nosus GG and Lactobacillus gasseri K7 during Urinary metabolomic fingerprinting after con pregnancy and lactation changes mouse mesen sumption of a probiotic strain in women with

teric lymph nodes and mammary gland micro mastitis. Pharmacol Res 2014; 87: 160–165.

biota. J Dairy Sci 2015; 98: 2114–2128. 30 de Andrés J, Jiménez E, Espinosa-Martos I, et al: 22 de Andrés J, Jiménez E, Chico-Calero I, et al: An exploratory search for potential molecular Physiological translocation of lactic acid bacteria targets responsive to the probiotic Lactobacillus during pregnancy contributes to the composition salivarius PS2 in women with mastitis: gene ex

of the milk microbiota in mice. Nutrients 2017; pression profiling vs. interindividual variability.

10:E14. Front Microbiol 2018; 9: 2166. 23 Rodríguez JM, Fernández L: Infectious mastitis during lactation: a mammary dysbiosis model; in McGuire M, McGuire M, Bode L (eds): Prebiotics and Probiotics in Human Milk: Origins and Func tions of Milk-Borne Oligosaccharides and Bacte ria. London, Academic Press, 2016, pp 401–428.

Origin and Uses of the Human Milk Microbiota 85 Microbiology of Milk and Lactation: Influence on Gut Colonization

Published online: March 20, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 86–93 (DOI: 10.1159/000504997)

Beyond the Bacterial Microbiome: Virome of Human Milk and Effects on the Developing Infant

a a, b Sindhu Mohandas Pia S. Pannaraj a Division of Infectious Diseases, Department of Pediatrics, Children’s Hospital Los Angeles, b Los Angeles, CA, USA; Department of Molecular Immunology and Microbiology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA

Abstract Human milk microbes play an important role in infant health and disease. Emerging evidence shows that human milk viruses are also transmitted from the mother to the infant via breastfeeding. These viruses include eukaryotic viruses, bacterium-infecting viruses called bacteriophages, and other viral particles. Human milk viruses are instrumental in shaping the infant gut virome and microbiome. Eukaryotic DNA and RNA viruses contribute to pathogenic challenges and protection. Bacteriophages have the ability to kill bacteria or supply them with potentially beneficial gene functions, thereby shaping the microbiome. The early infant virome is dominated by bacteriophages that likely contribute to a highly dynamic microbiome in the early life. There is a critical window of early childhood growth with rapid maturation of metabolic, endocrine, neural, and immune pathways. The colonization of microbes in the infant body during this time plays an important role in the establishment and maturation of these pathways. The virome transmitted via breastfeeding may also be particularly important at these critical time points of immune development. More longitudinal studies of mother-infant pairs will help to better define the human milk virome and their functional impact on the development of the growing infant. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

Human milk is an important source of microbes that colonize the infant gut in early life and contribute to immune system maturation and protection against pathogen invasion [1]. While research on the human microbiome has primarily focused on the prokaryotic composition of the human microflora, viruses are also transmitted in human milk and play important roles in shaping innate and adaptive immune defenses starting from infancy [2, 3]. Highly diverse viral communities, including eukaryotic viruses and viruses-infecting bacteria (bacteriophages), archaea, and viral elements integrated in the host chromosomes, make up the virome. Viral identity, diversity, and life cycles in various human body habitats have been poorly studied until recently [4]. With increasing use of next-generation sequencing techniques and metagenomics sequencing, data can be analyzed not only for bacteria but also archaea and nonhuman eukaryotic and viral sequences. This has revolutionized the way we think about viruses and has been instrumental in contributing substantial amounts of data on the composition of the normal and pathogenic viromes across various human body habitats. This review summarizes the few existing studies that have analyzed viruses and phages in human milk and their functional impact on the infant and their microbiota.

Virome Diversity and Taxonomy

The human virome is composed of all the viruses found on the surface and inside the human body. It is comprised of viruses that infect eukaryotic and prokaryotic cells. These viral communities differ in terms of abundance and composition based on the anatomical sites they inhabit. Studies have evaluated the virome in blood [5], gut [6], respiratory tract [7], skin [8], and CSF [9]. Human stool has 109–1010 viral-like particles per gram of stool [6]. While some studies have studied specific viruses [10, 11] and bacteriophages [12] in milk, only one study has characterized the virome of human milk [2]. The milk virome was found to be distinct from adult stool, urine, saliva, and CSF viromes [2, 9]. Among the eukaryotic viruses identified in human milk, the most abundant viruses identified in human milk were of the Herpesviridae, Poxviridae, Mimi- viridae, and Iridoviridae families [2]. The majority (95%) of the viruses in human milk and infant stool were bacteriophages. The most abundant phages in human milk were from the Myoviridae, Siphoviridae, and Podoviridae families from the order Caudovirales (tailed viruses), which have predominantly lytic lifestyles [2]. Every individual harbored a morphologically unique phage popu-

Human Milk Virome 87 lation in milk. Infant stool contains abundant bacteriophages from the Myoviri- dae, Siphoviridae, Podoviridae, Inoviridae, and Microviridae families [2, 13, 14]. Bacteriophages are the most studied part of the human virome to date. As in human milk, phages represent a much larger proportion than eukaryotic viruses in most studied body habitats [15]. These phages can modulate and impact the bacterial ecology through their lytic and lysogenic cycles. Lytic phages infect and hijack their host cell replication and translation machinery to produce virions, subsequently leading to host cell lysis and virion release to infect new host cells. Lysogenic phages integrate into their host’s genome without interfering with its replication. The phage is incorporated into the host genome as a pro- phage that is transmitted to its progeny at each cell division. Gene transfers can lead to phenotypic or functional changes in the prokaryote host, which can sometimes provide a competitive advantage in their habitat. Virome dynamics in human milk over time is not known due to the lack of longitudinal studies. The infant virome shows great temporal diversity and is highly dynamic [6, 13]. A predator-prey relationship is believed to exist between the bacteriophages and bacteria in early life [13]. In contrast, in older children and adults, the viral communities are largely stable over time at body sites that have been longitudinally studied, including the mouth, gut, and skin [8, 16, 17]. Over 95% of virus types are retained in adult subjects who were followed up over 1 year [18].

Transmission

Factors including age, sex, genetics, environment, household contacts, anti biotic use, and diet are involved in the transmission of the human virome [5, 17, 19, 20]. Viruses are the most abundant biological entity, and humans are constantly exposed to viruses from the environment and through oral intake. The infant gut virome is seeded from the vaginal or cesarean section delivery [21], close contact with the mother and other household contacts, the environment, and oral intake. Human milk is one of the earliest and significant factors that is involved in direct transmission and establishment of the virome starting from the first few hours after birth. We previously examined the virome contents of milk and infant stool in a cohort of 10 mother-infant pairs in the first 10 days of life of the breastfed infant [2]. Viral communities were distinguishable between milk and infant stool, but the infant stool community was more similar to milk than to adult stool, urine, or saliva communities. This suggests that the transfer of viruses between mother and infant gut is greatest from milk in early life. It is likely that viruses cycle

88 Mohandas/Pannaraj between the mother’s milk and infant saliva in a constant exchange. However, we favor the idea that transfer of viruses primarily occurs from mother to infant in early life. Evidence for vertical transfer also exists for pathogenic viruses such as cytomegalovirus (CMV), Zika virus, and human immunodeficiency virus (HIV) [10, 11]. True mother-infant pairs harbored significantly more homolo- gous viral contigs than unrelated pairs [2]. Because of the significant proportion of bacteriophages shared between mother-infant pairs, we postulate that transmission of milk bacteriophages to the infant gut may help to shape the infant gut microbiome. A few studies have shown evidence for bacteriophages that are vertically transmitted from mother to infant if the corresponding bacterial hosts are transmitted. For example, bifidobacterial and bacteriophages specific for bifidobacteria (bifidophages) are transmitted together in human milk [12, 22]. Duranti et al. [12] analyzed the bifidobacterial population in milk and infant stool samples from 25 mother-infant pairs at 7 days and 1 month after birth. Matching bifidobacterial strains were found in the milk and infant stool of related pairs. They also found evidence of vertical transmission of B. longum phage 10029, identified in the fecal samples of a child as well as in the milk sample of the corresponding mother. The study authors suggested that the bifido(pro)phage was transmitted by the mother in breast milk as part of the bifidobacterial host and was then induced in the gut of the newborn.

Role of the Human Milk Virome in Infant Health

Once human milk viruses enter the infant gut, the eukaryotic viruses may impact infant health directly, and the bacteriophages impact bacterial ecology. The longitudinal role of human milk in the trajectory is currently unknown. Viruses may survive in different relative abundances while in milk. From birth to 2 years of age, the overall trajectory of the eukaryotic virome and the bacterial microbiome diversity expands as the bacteriophage virome richness contracts (Fig. 1) [13]. The start of infant gut viral colonization is unclear. It is well known that some eukaryotic viruses can be transmitted transplacentally and have been detected in amniotic fluid, e.g., CMV, Zika virus, herpes simplex virus, and others. However, Lim et al. [23] found that amniotic fluid from most term infants did not have unique viral sequences that were also not detected in negative controls. A direct epifluorescent microscope did not detect any viral particles in newborn meconium [14]. After delivery, human milk offers the first direct microbial seeding of the infant gut. Eukaryotic viruses may then have sporadic and transient increases in community members (e.g., Adenoviridae, Anelloviridae,

Human Milk Virome 89 Delivery Diet/ Milk Household Genetics Health Antibiotics Environments method nutrition stage contacts

Infant gut

Key Eukaryotic virus Bacteriophage Bacteria Species diversity Birth 1 2 3 Age, years

Fig. 1. Breastfeeding transfers milk microbes including bacterial and viral communities directly to the infant gut. Multiple factors impact the diversity and composition of the microbial communities in both the mother and infant. The eukaryotic virus, bacteriophages, and bacteria coexist in the infant gut in an interdependent and dynamic relationship. The trajectory of the microbial diversity over time is shown during the dynamic period in the first 2–3 years of life. Dashed lines represent hypothesized trajectories based on limited data.

Astroviridae, Caliciviridae, Picornaviridae, and Reoviridae), especially during infancy and toddler years. Other eukaryotic viruses may persist after the first infection (Herpesviridae). Bacteriophages and bacterial microbiome are highly dynamic in the infant gut and inter-dependent in a predator-prey relationship starting at birth [13]. Bacteriophage richness is high even during the first week of life and decreases thereafter [2, 13]. It is unclear when the bacteriophages initially colonize the gut, but 95% of the viruses in human milk are bacteriophages. Bacterial diversity is inversely correlated with bacteriophage diversity. An oscillatory predator-prey dynamic has been postulated but has not been clearly observed in the sparse longitudinal studies. A stable adult-like virome and microbiome is achieved at around 2–3 years of age. Continuous breastfeeding through at least 1 year of life, as recommended by the World Health Organization, may have a role in influencing the infant virome at various time points during infancy. There is a critical window of early childhood growth when rapid maturation of metabolic, endocrine, neural, and immune pathways occurs. Microbe colonization in the infant body during this time

90 Mohandas/Pannaraj plays an important role in the establishment and maturation of these pathways [1]. The virome may be particularly important at these critical time points due to their pathogenic or protective roles or their impact on bacterial communities. Several studies reported virome associations with disease although the causal direction of the associations needs to be determined. Concurrent presence of enteric nonpolio enteroviruses in the infant viral community is associated with a reduction in oral polio vaccine seroconversion [24]. This suggests that viruses can influence immune responses to pathogens, including the development of protective antibodies. The eukaryotic virome is altered in infants with diarrhea and adults with HIV and low peripheral CD4 T-cell counts [25, 26]. Increased prevalence of Anelloviridae and Circoviridae families and phage populations could distinguish malnourished from healthy twins [27]. Gut phageome composition has been linked to colorectal cancer and type I diabetes [28, 29]. In inflammatory bowel disease, phenotypes and response to fecal microbiota transplantation have been shown to be related to both the eukaryotic virome and bacteriophages [30]. After breastfeeding, infants sometimes regurgitate milk into the upper respiratory tract thereby seeding an additional site with microbes. A higher relative abundance of Propionibacterium phages in the respiratory virome was associated with elevated serum cytokines and recurrent respiratory tract infections compared to children who had a single respiratory infection. These findings indicate that respiratory microbe homeostasis has a role in the risk of recurrent respiratory infections in childhood [7]. Phage adherence to the mucus at mucosal surfaces in the gut or respiratory tract may benefit the human host by limiting mucosal bacteria [31]. The virome transmitted by human milk to the infant could thus have a significant impact on the long-term health of the infant.

Limitations of Virome Studies and Next Steps

Current studies have been limited to cross-sectional data, small numbers of subjects, or lack of including both the mother and infant as a pair. Though the availability of next-generation sequencing and metagenomics has remarkably accelerated the potential to characterize viruses, the technology continues to be ex- pensive and has significant ongoing challenges [15]. Current techniques for generating a viral metagenome include enriching and purifying steps that can lead to missing important constituents of the virome, such as latent viruses, integrated viruses, double-stranded DNA viruses, and large eukaryotic viruses. Removal of nonviral DNA is still inefficient. Optimized protocols have been published, but their efficiency has not yet been comparatively assessed. Other issues include lack of a comprehensive experimental pipeline for metagenomic

Human Milk Virome 91 analyses of viral populations. Viral databases are incomplete. Most bacteriophages revealed in the gut are novel and cannot be assigned a taxonomic position or linked to a bacterial host. These unknown sequences have been called the “viral dark matter” and may comprise 75–99% of the reads. Rapid advances in technology and bioinformatic pipelines are needed to overcome many of these challenges and to provide better access and a clearer understanding of the human milk virome and its impact on the developing infant. The relationships between mother and infant microbiomes and viromes are an important area for additional research. The infancy period represents a critical time of microbial interactions in host immunity and metabolism. It also confers a critical window of opportunity to provide the infant with healthy microbes, including viruses and bacteriophages that can be found in human milk. There is a definite need for more longitudinal paired mother-infant studies designed to capture the dynamic nature of the milk and infant virome and microbiome. Effects of breastfeeding duration, maternal health, age, geographic variation, and other factors on the milk and infant virome must be elucidated. More data are imperative to understand the causal relationship of the virome to disease so that we can optimize recommendations for a healthy maternal-infant virome.

Disclosure Statement

P.S.P. receives research funding from AstraZeneca, MedImmune, Sanofi Pasteur, Pfizer, and the US National Institutes of Health. She has received speaker honoraria from Sobi and the Nestle Nutrition Institute. S.M. declares no conflicts of interest.

References

1 Pannaraj PS, Li F, Cerini C, et al: Association be- 5 Moustafa A, Xie C, Kirkness E, et al: The blood tween breast milk bacterial communities and DNA virome in 8,000 humans. PLoS Pathog

establishment and development of the infant gut 2017; 13:e1006292.

microbiome. JAMA Pediatr 2017; 171: 647–654. 6 Beller L, Matthijnssens J: What is (not) known 2 Pannaraj PS, Ly M, Cerini C, Set al: Shared and about the dynamics of the human gut virome in

distinct features of human milk and infant stool health and disease. Curr Opin Virol 2019; 37: 52–

viromes. Front Microbiol 2018; 9: 1162. 57. 3 Freer G, Maggi F, Pifferi M, et al: The virome and 7 Li Y, Fu X, Ma J, et al: Altered respiratory virome its major component, anellovirus, a convoluted and serum cytokine profile associated with recur- system molding human immune defenses and rent respiratory tract infections in children. Nat

possibly affecting the development of asthma and Commun 2019; 10: 2288. respiratory diseases in childhood. Front Micro- 8 Oh J, Byrd AL, Park M, et al: Temporal stability

biol 2018; 9: 686. of the human skin microbiome. Cell 2016; 165: 4 Breitbart M, Hewson I, Felts B, et al: Metagenomic 854–866. analyses of an uncultured viral community from

human feces. J Bacteriol 2003; 185: 6220–6223.

92 Mohandas/Pannaraj 9 Ghose C, Ly M, Schwanemann LK, et al: The vi- 22 Lugli GA, Milani C, Turroni F, et al: Prophages of rome of cerebrospinal fluid: viruses where we the genus Bifidobacterium as modulating agents once thought there were none. Front Microbiol of the infant gut microbiota. Environ Microbiol

2019; 10: 2061. 2016; 18: 2196–2213. 10 Blohm GM, Lednicky JA, Marquez M, et al: Evi- 23 Lim ES, Rodriguez C, Holtz LR: Amniotic fluid dence for mother-to-child transmission of Zika from healthy term pregnancies does not harbor a

virus through breast milk. Clin Infect Dis 2018; detectable microbial community. Microbiome

66: 1120–1121. 2018; 6: 87. 11 Kuhn L, Kim HY, Walter J, et al: HIV-1 concen- 24 Praharaj I, Parker EPK, Giri S, et al: Influence of trations in human breast milk before and after nonpolio enteroviruses and the bacterial gut mi-

weaning. Sci Transl Med 2013; 5: 181ra151. crobiota on oral poliovirus vaccine response: a

12 Duranti S, Lugli GA, Mancabelli L, et al: Maternal study from South India. J Infect Dis 2019; 219: inheritance of bifidobacterial communities and 1178–1186. bifidophages in infants through vertical trans- 25 Phan TG, da Costa AC, Del Valle Mendoza J, et al:

mission. Microbiome 2017; 5: 66. The fecal virome of South and Central American 13 Lim ES, Zhou Y, Zhao G, et al: Early life dynamics children with diarrhea includes small circular of the human gut virome and bacterial microbi- DNA viral genomes of unknown origin. Arch

ome in infants. Nat Med 2015; 21: 1228–1234. Virol 2016; 161: 959–966. 14 Breitbart M, Haynes M, Kelley S, et al: Viral di- 26 Monaco CL, Gootenberg DB, Zhao G, et al: Al- versity and dynamics in an infant gut. Res Micro- tered virome and bacterial microbiome in human

biol 2008; 159: 367–373. immunodeficiency virus-associated acquired im- 15 Shkoporov AN, Hill C: Bacteriophages of the hu- munodeficiency syndrome. Cell Host Microbe

man gut: the “known unknown” of the microbi- 2016; 19: 311–322.

ome. Cell Host Microbe 2019; 25: 195–209. 27 Reyes A, Blanton LV, Cao S, et al: Gut DNA vi- 16 Altan E, Aiemjoy K, Phan TG, et al: Enteric viromes of Malawian twins discordant for severe

rome of Ethiopian children participating in a acute malnutrition. Proc Natl Acad Sci USA 2015;

clean water intervention trial. PLoS One 2018; 112: 11941–11946. 13:e0202054. 28 Zhao G, Vatanen T, Droit L, et al: Intestinal vi- 17 Ly M, Jones MB, Abeles SR, et al: Transmission of rome changes precede autoimmunity in type I

viruses via our microbiomes. Microbiome 2016; 4: diabetes-susceptible children. Proc Natl Acad Sci

64. USA 2017; 114:E6166–E6175. 18 Reyes A, Haynes M, Hanson N, et al: Viruses in 29 Hannigan GD, Duhaime MB, Ruffin MT 4th, et the faecal microbiota of monozygotic twins and al: Diagnostic potential and interactive dynamics

their mothers. Nature 2010; 466: 334–338. of the colorectal cancer virome. mBio 2018; 19 Minot S, Sinha R, Chen J, et al: The human gut 9:e02248-18. virome: inter-individual variation and dynamic 30 Conceicao-Neto N, Deboutte W, Dierckx T, et al:

response to diet. Genome Res 2011; 21: 1616– Low eukaryotic viral richness is associated with 1625. fecal microbiota transplantation success in pa-

20 Gorska A, Peter S, Willmann M, et al: Dynamics tients with UC. Gut 2018; 67: 1558–1559. of the human gut phageome during antibiotic 31 Barr JJ, Auro R, Furlan M, et al: Bacteriophage

treatment. Comput Biol Chem 2018; 74: 420–427. adhering to mucus provide a non-host-derived

21 McCann A, Ryan FJ, Stockdale SR, et al: Viromes immunity. Proc Natl Acad Sci USA 2013; 110: of one year old infants reveal the impact of birth 10771–10776.

mode on microbiome diversity. PeerJ 2018; 6:e4694.

Human Milk Virome 93 Microbiology of Milk and Lactation: Influence on Gut Colonization

Published online: March 23, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 94–102 (DOI: 10.1159/000504996)

Gut Microbiota, Host Gene Expression, and Cell Traffic via Milk

Josef Neu Department of Pediatrics, University of Florida College of Medicine, Gainesville, FL, USA

Abstract Contrary to common belief, the human neonate is often born with a nonsterile gastrointestinal tract, suggesting fetal colonization. This has been substantiated by numerous studies showing microbes in meconium. Shortly after birth, the infant is further colonized by microbes that reflect the diet, which in the newborn consists of milk. When fed milk from the mother’s breast, the infant derives a set of live microbes that have the capability of colonizing the gastrointestinal tract. This milk also provides a source of enzymes, such as lipase and alkaline phosphatase. Milk also provides a multitude of proteins, microRNAs, and other components that putatively interact with the host intestinal innate mucosal immune system to control infection, modulate intestinal inflammation, and provide signaling to distal sites for the development of adaptive immunity as well as growth and communication with the central nervous system. Colostrum differs from transitional and mature milk by being particularly rich in immunoglobulins as well as leukocytes. Live microbes found in fresh mother’s milk may be personalized for her infant and thus provide an impetus for either ensuring delivery of this personalized milk to the infant or, if that is not possible, to develop the means to personalize donor milk or formula. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Religion is a culture of faith; science is a culture of doubt. Richard P. Feynman

Introduction

Milk provided directly from the mother’s breast is the most appropriate food for most infants. This is a practice that has been supported by millions of years of history as well as scientific studies. Based on these studies, there has been a movement to provide milk to infants that simulates human milk as closely as possible whenever the mother cannot provide milk for her own baby. A baby born preterm presents a special circumstance. Many of these infants are too immature to feed directly from the breast and require tube feedings. Mothers of preterm infants are also often highly stressed and may not be able to provide adequate quantities of milk to meet their babies’ needs. Social circumstances such as illicit drug use or maternal infection may also preclude breastfeeding. Infants born preterm often have significant immaturities of their gastrointestinal tracts that preclude full feedings shortly after birth and require intravenous nutritional supplementation. When mothers’ milk is available, the composition of certain components such as protein, calcium, phosphorus, and other nutrients may not be completely adequate for a rapidly growing preterm infant. Be- yond macronutrients and major minerals, there are numerous other components of mother’s milk that appear to play very important bioactive roles. In comparison to macronutrient fortifications, attempts to mimic many of these bioactive factors from human milk have been targets of numerous studies in the past few decades. For such fortifications to be beneficial to infants, the underlying science needs to be sound. That said, much of the science pertaining to these bioactive components in human milk may be suboptimal since it is based on associative rather than mechanistic data. Said slightly differently, many potentially active components are found in human milk. Whether they actually have some important role to play in the developing infant who is the recipient of this milk has relied largely on conjecture and associations. When utilized as supplements in a milk background that is different from that of the mother, their function may be lost or significantly diminished. Some of the human milk components thought to play important roles include immunoglobulins, bioactive proteins such as lactoferrin and lysozyme, long-chain lipids, oligosaccharides, enzymes such as alkaline phosphatase and lipase, certain immune cells as well as stem cells, and more recently microbes and microbial components. In some situations, their mere presence in human

Milk Components and Host Interactions 95 milk has been translated to their need in formulas or other milk-type preparations that are fed to these infants. When considering these components of human milk, the following quote should be considered: “Correlation implies association, but not causation. Conversely, causation implies association but not correlation” [1]. This is important and should be kept in mind when discussing various potentially bioactive factors in human milk that are suggested to play important physiologic roles in the developing infant. Guidelines from the American Academy of Pediatrics published in 2012 [2] recommended that all preterm infants receive mother’s own milk or pasteurized donor milk if mother’s own milk is unavailable. In most cases, when donor milk is pasteurized, many bioactive components are lost [3–5]. Bile salt-stimulated lipase, an enzyme that hydrolyzes triglycerides, is completely inactivated with pasteurization. Alkaline phosphatase, an enzyme with the ability to dephosphor- ylate lipopolysaccharide, an important component of the cell wall gram-negative microorganisms known to be pathogenic, is also inactivated following the pasteurization process [6]. Assays for these enzymes have, in fact, been used to evaluate the adequacy of the pasteurization process for nearly 80 years [7]. The majority of resident live microbes in raw mother’s milk are inactivated by the pasteurization process, and donor milk provided by milk banks is usually sterile [8]. There are numerous cellular components present in fresh human milk as well as immunoglobulins that appear to respond to the status of the infant while the infant is being nursed by the mother, reminiscent of what has been termed the enteromammary system [9]. This dynamic system that involves the interplay between infant colonization, maternal exposure to the microbes the infant is colonized by, and subsequent dynamic production of specific antimicrobial factors relayed to the infant in the mother’s milk [10]. In this review, we will discuss several of the components of fresh mother’s milk, their potential roles in terms of interaction with the host, how this interaction may play a role in host health, and the state of the science as well as the gaps that remain for future investigation.

Human Milk Microbiota

The first area to be discussed is the microbiota found in human milk [11, 12]. It is clear from several studies that human milk feeding contributes to the microbial ecology of the developing gastrointestinal tract. Of note is the very fact that this may not provide the first colonizer to the infant gastrointestinal tract, since several studies have demonstrated that initial colonization of the infant gut may actually begin in the womb [13, 14].

96 Neu Studies suggesting a nonsterile in utero microbiome are controversial [15– 17]. Whether microbes are only present during pathological states or whether they are present as commensals is currently being debated and remains a topic for additional investigation. However, studies in human milk using both culture and nonculture techniques show that it contains a diverse array of microbes. From some studies, it appears that the microbial composition of milk from an individual mother over the first month after birth changes slightly, but it is very different from the microbial composition of milk from other mothers [18]. This suggests a personalization of the milk microbiota for each mother’s individual infant. The dose of microbes derived from mother’s milk appears to be substantial. Assuming an intake of 800 mL/day [19], it is estimated that the infant receives approximately 107 to 108 bacterial cells daily from this milk. The presence of these microbes in the milk have been suggested to play an important role in the development of the infant intestinal microbiome [20]. From where these microbes are derived, remains controversial. Several sources have been theorized, but the infant’s mouth, the mother’s skin, or the gastrointestinal tract are all likely sites of origin [21]. Exactly how these microbes transverse the mother’s body from her oral cavity or her gastrointestinal tract remains unclear [22]. At least one study in rodents has demonstrated that administering a certain strain of microbes to pregnant mice will result in that same strain being found in the milk of these lactating pregnant rodents [23]. Whether there are certain cell types such as antigen-presenting dendritic cells that carry these microbes from the maternal intestine to the breast is not clear but has been suggested. Despite these findings of microbes being present in human milk, their role in the infant remains poorly understood. It is thought that microbes and their components stimulate the innate immune system [24]. Low-grade stimulation by soluble Toll-like antigens, immunoglobulin interference with infection, and stabilization of barrier function, all appear to play a role, but the precise mechanisms remain to be determined. Whether and how adaptive immunity is affected remains to be elucidated. Introduction of these microbes into the gastrointestinal tract may also play a role in colonization resistance whereby they prevent the colonization by more pathogenic microorganisms.

Metabolomics

Another potential role of these microbes relates to their production of various highly active metabolites in conjunction with other nutritional substrates in the gastrointestinal tract. For example, the interaction of microbes with certain carbohydrates leads to the production of short-chain fatty acids such as acetate,

Milk Components and Host Interactions 97 propionic acid, and butyrate. These play important roles related to large intestinal energy metabolism (where butyrate is a major fuel source for colonic epithelium), proliferation and differentiation of epithelium, and maintenance of tight junctions [25]. Propionic and butyric acids also are able to inhibit IL-12- and IL-23-mediated stimulation of CD8 T cells via dendritic cells [26]. Intestinal microbes are also involved in tryptophan metabolism, which is the major pathway precursor for serotonin synthesis in the intestine [27]. This plays a major role in neuronal signaling pathways and has been related to various neurologic and psy- chiatric disease entities. Of interest is the fact that the metabolic profile of milk differs depending on the stage of lactation and gestational age [28]. Whether this has any consequence for the infant remains unknown. Milk metabolite composition also depends on the geographic region of origin [29]. It is highly likely that this effect is due to dietary differences in these geographic regions. Compared with donor milk or formula, mother’s own milk has a higher concentration of most microbes. Some mothers are not able to produce adequate quantities of milk shortly after birth to nourish her infant. Preterm birth, cesarean section, and stress are major contributors to the lack of ability to produce milk. Thus, if human milk is to be supplied to this mother’s infant, donor milk is needed, which is usually derived through donor milk banks. Since this milk has been pasteurized, live microbes have been killed, but microbial DNA may still be present. Donor milk is batched and difficult to personalize for any particular infant. Hence, it is possible that these infants do not derive the potential benefits of live and personalized microbes in terms of their metabolism and/or immunologic roles. Whether personalization of microbiota in human milk might have biological advantages is still to be shown. However, it is feasible to utilize small quantities of mother’s milk for personalization of banked donor milk. A recent study by our group demonstrated that incubation of donor milk with different concentrations of mother’s own milk for different periods of time can be utilized to faunate donor milk [30]. However, even if we can provide a personalized milk microbial ecology using refaunation techniques, it is not clear whether this will provide benefits for the infant.

Innate Immune Mechanisms

As mentioned previously, there are factors including microbes and their metabolites in human milk that have the ability to affect barrier function of the developing infant gastrointestinal tract, including mucus production and interepithelial tight junction integrity. This is critical from the perspective that regulation of interepithelial junctions is critical to health and disease [31, 32].

98 Neu There are also several Toll-like receptors present in human milk, causing it on the whole to be strongly anti-inflammatory. Components that contribute to this include sTLR2 and sCD14, which inhibit TLR2 signaling; sCD14, lact adherin, lactoferrin, and 2 -fucosyllactose, which attenuate TLR4 signaling [33]. ′ Another immune marker found in human mild milk is TGF-β. Previous investigations suggested an association between higher colostrum TGF-β levels and reduced risk of several immunologic outcomes in children [34]. Con- trary to the early studies, at least one study has shown a higher risk of eczema with higher levels of human milk TGF-β at 1 month of age [35]. The reason for these discrepancies is not fully understood but may be related to time of milk sampling where the colostrum composition is very different than that of a mature milk.

Immunoglobulins

There are various immunoglobulins present in human milk that have the capability for pathogen recognition [36]. Of the immunoglobulins, secretory IgA (SIgA) comprises over 90% of the immunoglobulin fraction. IgG is present in small quantities in colostrum and transitional milk but becomes a much larger proportion when the milk matures. SIgA is able to block pathogens without stimulating significant inflammatory responses. This is done by simply blocking the pathogen contact with the intestinal epithelial layer entrapping the pathogens within the mucin layers. There are several other effects of SIgA on the development of the gut microbiology and postintestinal immunity [36]. Studies in mice show that exposure of maternal SIgA prevents the translocation of aerobic bacteria from the neonatal gut into the draining lymph nodes. Maternal breast milk-derived SIgA can also promote intestinal epithelial barrier function and prevent systemic infection by blocking potential pathogens. Early exposure to SIgA in breast milk resulted in a pattern of intestinal epithelial cell expression that differed from mice that were not exposed to passive SIgA. This included intestinal inflammation-related genes that are associated with intestinal inflammation in humans. Colonic damage caused by epithelial disrupting agents, such as dextran sulfate sodium, could also be ameliorated by providing maternal SIgA.

Milk Components and Host Interactions 99 Growth Factors

Another set of bioactive compounds found in human milk include growth factors such as epidermal growth factor, vascular endothelial growth factor, and hepatocyte growth factor [37]. Some of these factors are thought to be very high in human milk, sometimes 20–30 fold that of maternal serum, which may point to important physiologic roles in the infant intestinal maturation. Despite this very strong association, the mechanistic framework for these factors actually inducing growth or having effects in the human infant gastrointestinal tract are lacking.

Exosomes

There is a growing interest in extracellular membrane vesicles, particularly the exosomes found in human milk [38]. These exosomes have been associated with a potential role in allergy prevention via the mechanism of T-cell proliferation and effects on cytokine production. The mechanism of how this alters allergic sensitization remains poorly understood. Exosomes purified from breast milk are able to promote intestinal epithelial cell growth in infants even when they receive formula feeding. Studies utilizing enteritis derived from neonatal mice or premature human small intestine have demonstrated the stimulating effect of breast milk on their growth-improved proliferation. However, whether this was due to any factor related to the exosomes remains unclear.

Stem Cells

Recent studies demonstrate that human milk contains a large number of cell types, including a certain population that are typical of stem cells [39]. Many of the stem cells are precursors to breast epithelial cells. These cells exhibit stem cell markers, and under certain conditions they undergo differentiation towards more mature epithelial lineages leading to either milk-producing cells or epithelial cells. Of interest is that the cells can also be transferred from the mother to the fetus. They can cross the wall of the gastrointestinal tract of the nursing pups and enter the circulatory system subsequently reaching different organisms and become fully functional after differentiation [40]. This microchimerism may lead to engraftment of the cells into the offspring organs, such as the liver carti- lage bone and duodenum. One possible consequence of such microchimerism is increased tolerance of a recipient to donor antigens by improved acceptance of maternal transplants by individuals who were breastfed as infants.

100 Neu There are other potential uses of breast milk-derived cells, including utilization for stroke therapy, especially when the subpopulations produce vascular endothelial growth factor and human growth factor. At this time, there are more questions than answers when it comes to breast milk stem cells, but their potential utilization for therapeutic purposes should stimulate some interesting research in the near future.

Conclusion and Future Directions

From this review, it should be obvious that we had just begun to scrape the surface of human milk-related factors that appear to play a role in the infant host. Many of the studies show correlations and associations but do not meet the cri- teria for clear cause and effect mechanisms. In the future, it is hoped that systems will be developed whereby mechanisms can reliably be tested utilizing in vitro models that are closer to the human situation. Furthermore, most in vivo studies have been done in rodents, but new technologies are being developed that provide better insights into the relationship of human milk components, the human host responses to these components, and what they mean for the subsequent development of that individual during his or her lifetime as well as potential effects that may traverse into additional generations.

Disclosure Statement

The author receives grant funding from Infant Bacterial Therapeutics and the National Institutes of Health.

References

1 Altman N, Krzywinski M: Association, correla- 5 Peila C, Moro GE, Bertino E, et al: The effect of

tion and causation. Nat Methods 2015; 12: 899– Holder pasteurization on nutrients and biologi- 900. cally-active components in donor human milk: a

2 Breastfeeding SO: Breastfeeding and the use of review. Nutrients 2016; 8:E477.

human milk. Pediatrics 2012; 129:e827–e841. 6 Yang Y, Rader E, Peters-Carr M, et al: Ontogeny 3 Rodriguez-Camejo C, Puyol A, Fazio L, et al: Im- of alkaline phosphatase activity in infant intes-

pact of Holder pasteurization on immunological tines and breast milk. BMC Pediatr 2019; 19: 2. properties of human breast milk over the first 7 Kay HD, Neave FK: Some results of the applica-

year of lactation. Pediatr Res 2019; 87: 32–41. tion of a simple test for efficiency of pasteurisa-

4 Guerra AF, Mellinger-Silva C, Rosenthal A, Lu- tion. Lancet 1935; 225: 1516–1518. chese RH: Hot topic: Holder pasteurization of 8 Vickers AM, Starks-Solis S, Hill DR, Newburg DS: human milk affects some bioactive proteins. J Pasteurized donor human milk maintains micro-

Dairy Sci 2018; 101: 2814–2818. biological purity for 4 days at 4 ° C. J Hum Lact

2015; 31: 401–405.

Milk Components and Host Interactions 101 9 Kleinman RE, Walker WA: The enteromammary 25 Tan J, McKenzie C, Potamitis M, et al: The role of immune system: an important new concept in short-chain fatty acids in health and disease. Adv

breast milk host defense. Dig Dis Sci 1979; 24: Immunol 2014; 121: 91–119. 876–882. 26 Nastasi C, Fredholm S, Willerslev-Olsen A, et al: 10 Neu J, Sullivan S: Baby and breast: a dynamic Butyrate and propionate inhibit antigen-specific

interaction. Pediatr Res 2012; 71: 135. CD8(+) T cell activation by suppressing IL-12 11 Fitzstevens JL, Smith KC, Hagadorn JI, et al: Sys- production by antigen-presenting cells. Sci Rep

tematic review of the human milk microbiota. 2017; 7: 14516.

Nutr Clin Pract 2017; 32: 354–364. 27 Waclawikova B, El Aidy S: Role of microbiota and 12 Gomez-Gallego C, Garcia-Mantrana I, Salminen tryptophan metabolites in the remote effect of S, Collado MC: The human milk microbiome and intestinal inflammation on brain and depression.

factors influencing its composition and activity. Pharmaceuticals (Basel) 2018; 11:E63.

Semin Fetal Neonatal Med 2016; 21: 400–405. 28 Sundekilde UK, Downey E, O’Mahony JA, et al: 13 Younge N, McCann JR, Ballard J, et al: Fetal ex- The effect of gestational and lactational age on

posure to the maternal microbiota in humans the human milk metabolome. Nutrients 2016;

and mice. JCI Insight 2019; 4: 127806. 8:E304. 14 Ardissone AN, Cruz DM, Davis-Richardson AG, 29 Gay MCL, Koleva PT, Slupsky CM, et al: World- et al: Meconium microbiome analysis identifies wide variation in human milk metabolome: indi- bacteria correlated with premature birth. PLoS cators of breast physiology and maternal life-

One 2014; 9:e90784. style? Nutrients 2018; 10:E1151. 15 de Goffau MC, Lager S, Sovio U, et al: Human 30 Cacho NT, Harrison NA, Parker LA, et al: Person- placenta has no microbiome but can contain po- alization of the microbiota of donor human milk

tential pathogens. Nature 2019; 572: 329–334. with mother’s own milk. Front Microbiol 2017; 8: 16 Aagaard K, Ma J, Antony KM, et al: The Placenta 1470. harbors a unique microbiome. Sci Transl Med 31 Camilleri M: Leaky gut: mechanisms, measure-

2014; 6: 237ra65. ment and clinical implications in humans. Gut

17 Parnell LA, Briggs CM, Cao B, et al: Microbial 2019; 68: 1516–1526. communities in placentas from term normal 32 Brandtzaeg P: Gate-keeper function of the intesti-

pregnancy exhibit spatially variable profiles. Sci nal epithelium. Benef Microbes 2013; 4: 67–82.

Rep 2017; 7: 11200. 33 He Y, Lawlor NT, Newburg DS: Human milk 18 Hunt KM, Foster JA, Forney LJ, et al: Character- components modulate Toll-like receptor-mediat-

ization of the diversity and temporal stability of ed inflammation. Adv Nutr 2016; 7: 102–111. bacterial communities in human milk. PLoS One 34 Kalliomaki M, Ouwehand A, Arvilommi H, et al:

2011; 6:e21313. Transforming growth factor-beta in breast milk: 19 Boix-Amoros A, Collado MC, Mira A: Relation- a potential regulator of atopic disease at an early

ship between milk microbiota, bacterial load, age. J Allergy Clin Immunol 1999; 104: 1251–1257. macronutrients, and human cells during lacta- 35 Khaleva E, Gridneva Z, Geddes DT, et al: Trans-

tion. Front Microbiol 2016; 7: 492. forming growth factor beta in human milk and 20 Pannaraj PS, Li F, Cerini C, et al: Association be- allergic outcomes in children: a systematic re-

tween breast milk bacterial communities and view. Clin Exp Allergy 2019; 49: 1201–1213. establishment and development of the infant gut 36 Rogier EW, Frantz AL, Bruno ME, et al: Lessons

microbiome. JAMA Pediatr 2017; 171: 647–654. from mother: long-term impact of antibodies in 21 McGuire MK, McGuire MA: Got bacteria? The breast milk on the gut microbiota and intestinal astounding, yet not-so-surprising, microbiome of immune system of breastfed offspring. Gut Mi-

human milk. Curr Opin Biotechnol 2017; 44: 63– crobes 2014; 5: 663–668. 68. 37 Kobata R, Tsukahara H, Ohshima Y, et al: High 22 Latuga MS, Stuebe A, Seed PC: A review of the levels of growth factors in human breast milk.

source and function of microbiota in breast milk. Early Hum Dev 2008; 84: 67–69.

Semin Reprod Med 2014; 32: 68–73. 38 Liao Y, Du X, Li J, Lönnerdal B: Human milk exo- 23 Martin R, Jimenez E, Heilig H, et al: Isolation of somes and their microRNAs survive digestion in bifidobacteria from breast milk and assessment vitro and are taken up by human intestinal cells.

of the bifidobacterial population by PCR-dena- Mol Nutr Food Res 2017; 61, DOI: 10.1002/ turing gradient gel electrophoresis and quantita- mnfr.201700082.

tive real-time PCR. Appl Environ Microbiol 2009; 39 Hassiotou F, Geddes DT, Hartmann PE: Cells in

75: 965–969. human milk: state of the science. J Hum Lact

24 Cacho NT, Lawrence RM: Innate immunity and 2013; 29: 171–182.

breast milk. Front Immunol 2017; 8: 584. 40 Ninkina N, Kukharsky MS, Hewitt MV, et al: Stem cells in human breast milk. Human Cell

2019; 32: 223–230.

102 Neu Microbiology of Milk and Lactation: Influence on Gut Colonization

Published online: March 31, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 103–112 (DOI: 10.1159/000505337)

Breast Milk and Microbiota in the Premature Gut: A Method of Preventing Necrotizing Enterocolitis

W. Allan Walker Di Meng Pediatric Gastroenterology Department, Mucosal Immunology and Biology Research Center, Massachusetts General Hospital for Children, Harvard Medical School, Boston, MA, USA

Abstract Necrotizing enterocolitis (NEC) is a devastating inflammatory condition of the intestine, which affects premature infants and causes untold damage. Its pathogenesis has to do with how colonizing bacteria interact with the immature newborn intestine. An immature innate immune response with increased TLR-4 on the cell surface and increased signaling molecules, such as NF-κB, can cause excessive inflammation. This is in conjunction with a decrease in the appearance of regulatory molecules which effect the control of innate responses. This condition is so devastating that it must be prevented and not treated. Fortu- nately, breast milk and probiotics can affect the condition leading to reduced inflammation. How does this effect work? We have shown that breast milk tryptophan and Bifidobacterium infantis result in a metabolite (indole-3-lactic acid) response, which is anti- inflammatory via inhibition of the aryl hydrocarbon receptor transcription factor which stimulates an IL-8 response. We have also shown that breast milk complex carbohydrates interacting with Bacteroides fragilis can cause short-chain fatty acids which exert anti- inflammatory effects on the newborn intestine. These breast milk metabolites could help prevent NEC if shown to be effective clinically. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

Necrotizing enterocolitis (NEC) is the most common intestinal emergency in premature infants [1]. It occurs in approximately 10% of infants under 1,500 g. It is in part a result of excessive inflammation occurring following contact with colonizing bacteria instead of the homeostasis that normally exists. As a result, the costs caused by morbidity, mortality, and direct hospital expenses are excessive [2]. Since NEC sequelae leave the infant at risk for short-bowel syndrome and significant lifelong cognitive problems, it is imperative that the disease be prevented rather than treated [3]. Several approaches have tried to prevent or lessen the expression of the condition. These include giving the premature infant its mother’s milk since breast milk has more protective factors [4] or giving the premature a variety of probiotics that may protect them from excessive inflammation with pathogens [5]. Recently, a study published in Pediatric Research suggested that the most effective way to prevent NEC is to give the premature infant a combination of expressed breast milk and a probiotic [6]. This was based on a study including a large number of premature infants treated for NEC presentation with probiotics and either breast milk and probiotics or formula and probiotics. The infants receiving mother’s milk and probiotics were protected from NEC, but those who received probiotics and formula were not. This suggested that probiotics interacting with breast milk to produce metabolites can reduce the excessive inflammation seen in the premature infant. We previously reported that Bifidobacterium infantis secretions had anti-inflammatory effects on the fetal intestine in a human fetal cell line [7]. We have also reported that Bacteroides fragilis is anti-inflammatory in the fetal human intestine [8]. Accordingly, in this study, we examined if incubating B. infantis or B. fragilis with breast milk can produce metabolites that are anti-inflammatory.

The Clinical Problem

When an immature intestine that is accustomed to residing in the interuterine environment where very few microorganisms are available to interact encounters the trillions of organisms which comprise the colonizing bacteria, excessive inflammation occurs instead of homeostasis that normally exists. This excessive inflammation could lead to bowel necrosis resulting in NEC [9]. We have studied this interaction and have shown that the reaction is in large part due to an inappropriate immature intestinal immune response [9]. For example, the immature immune system cannot distinguish between commensal versus pathogenic organisms, and responds to both with inflammation [10]. Furthermore, if

104 Walker/Meng SIGIRR TLR2, TLR4 MAL

MyD88 IRAK-M IRAKs Myd88s

TRAF6 Tollip A20 NF-κB I-κB

I-κB Inflammatory Ub gene expression Ub Ub

IL-8, TNF-α, IFN-γ

Mature Immature NEC ** 30 6 Mature 1.4 Mature Immature Immature 1.2 NEC 8 5 * * * 1.0 4 6 * * 0.8 3 * * 4 0.3 * 2 0.2 Relative mRNA level Relative Relative mRNA level Relative Relative mRNA level Relative 2 * 1 0.1 ** ** ** 0 0 0 a IL-8 b TLR2 TLR4 MyD88 TRAF NF-κB1 c SIGIRR Tollip A-20

Fig. 1. NF-κB/MyD88 acute innate immune-inflammatory and negative response genes measured by real-time PCR and expressed by relative fold mRNA levels of laser capture epithelial microdissection of 3 fetal/NEC/control intestines with control mRNA levels arbitrarily expressed as IL-8 response (a); receptors/signaling molecules/transcription factor (fetal/ control intestine only) (b); and negative regulators (c) (* p 0.01; ** p 0.001). Reproduction with permission from PLoS One [9]. we compare biopsies from premature infants, premature infants with NEC, and full-term infants, we note a marked difference in the expression of receptors, signaling molecules, and inhibitors [9, 11]. TLR-4 is expressed on the cell surface rather than inside the cell in larger quantities on immature cells; signaling molecules are increased in expression, and inhibiting molecules, which reduce the response, are decreased, resulting in an excessive, uninhibited inflammatory response that provides the basis for the prolonged necrosis that constitutes NEC (Fig. 1) [9, 12–14]. To overcome this excessive inflammation in the immature intestine, we found that we must find a clinically suitable way to minimize the inflammatory response to colonizing bacteria. One method that has been sug-

Effect of Breast Milk and Microbiota on the Immature Intestine 105 0.01 0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50 100 Study Year Pts, n Gross 1983 67 Tyson 1983 81 Lucas 1984 162 Schanler 1985 166

Overall 476 z = –2.062 p = 0.039 Favors breast milk Favors control

Fig. 2. Meta-analysis of the effect of mother’s breast milk on the incidence of NEC [18–22].

gested is corticosteroids, a trophic factor known to result in maturation of the gastrointestinal tract [15]. However, this approach has other unacceptable pul- monary side effects in the premature infant and cannot be used [16].

An Approach to a Solution of the Problem

Several studies have been done over the years to suggest that feeding the premature infant its own mother’s breast milk can reduce the incidence and severity of inflammation and prevent or minimize NEC [17]. These studies have been recorded in a meta-analysis (Fig. 2) [18–22] and appear to collectively reduce the incidence [23]. This is understandable since milk from mothers delivering prematurely have increased amounts of immunomediated factors (e.g., secretory IgA [SIgA] and cytokines) and anti-inflammatory factors (e.g., lactoferrin and ω-3 fatty acids), and the prebiotic effect of oligosaccharides, and it contains bacteria translocated from the mother’s intestine which are protective [24–27] and stimulate “pioneer” bacteria which preferentially help to yield SIgA and reduce IL-8 inflammation [28]. These factors undoubtedly reduce the excess inflammation seen after colonization. In addition, probiotics have been used in premature infants and have been shown to reduce NEC or to make its expression less problematic (Fig. 3) [29–40]. These studies suggest that certain bacterial species can also work on the immature intestine to reduce inflammation. We attempted to determine if secretions from Bifidobacterium infantis in the absence of the actual organism could affect the excessive inflammation. What we found out was components of the secretions in the absence of organisms were in fact anti-inflammatory [7]. We did this because of the concern of neonatologists that live organisms in an immature intestine could be dangerous to the immunocompromised gut of the premature infant.

106 Walker/Meng Review: Probiotics for NEC prevention Comparison: 01 NEC Outcome: 01 definite NEC Study Probiotic No probiotic RR (fixed) RR (fixed) or subcategory n/N n/N 95% CI Weight, % 95% CI Kitajima 1997 0/45 0/46 Not estimable Dani 2002 4/295 8/290 11.51 0.49 (0.15, 1.61) Costalos 2003 5/51 6/36 9.72 0.59 (0.19, 1.78) Bin-Nun 2005 1/72 10/73 13.73 0.10 (0.01, 0.77) Lin 2005 2/180 10/187 13.56 0.21 (0.05, 0.94) Manzoni 2006 1/39 3/41 4.04 0.35 (0.04, 3.23) Mohan 2006 2/21 1/17 1.53 1.62 (0.16, 16.37) Stratiki 2007 0/38 3/31 5.31 0.12 (0.01, 2.19) Lin 2008 4/217 14/217 19.35 0.29 (0.10, 0.85) Samanta 2008 5/91 15/95 20.29 0.35 (0.13, 0.92) Rouge 2009 2/45 1/49 1.32 2.18 (0.20, 23.21)

Total (95% CI) 1,094 1,082 100.00 0.35 (0.23, 0.55) Total events: 26 (probiotic), 71 (no probiotic) Test for heterogeneity: χ2 = 7.66, df = 9 (p = 0.57), I2 = 0% Test for overall effect: z = 4.64 (p < 0.00001)

0.01 0.1 1 10 100 Favors treatment Favors control

Fig. 3. A meta-analysis of the probiotic effect on NEC in premature infants [29–40]. Repro- duced with permission from Deshpande et al. [40].

Effect of B. infantis Interaction with Tryptophan in Breast Milk

Of interest was the observation made several years ago that probiotics were more effective when used with mother’s expressed breast milk rather than with infant formula [6]. In addition, we know that tryptophan exists in large quantities in mother’s milk [41], and that metabolites of tryptophan are immune mediators [42]. So, with the help of pharmacologists at the University of Salerno and the European Biomedical Research Institute in Salerno (EBRIS), we have been able to identify the anti-inflammatory molecule as indole-3-lactic acid (ILA), a known metabolite of tryptophan [43]. Since breast milk contains large amounts of tryptophan, we incubated the organism with breast milk and noted that the secretions contained large ILA quantities [Meng et al., unpubl. data]. Using 3- and 3- to 10 -kDa fractions of the culture media including breast milk, the same fraction that we previously had shown to be anti-inflammatory [7], we identified a specific peak using an ultrahigh-performance liquid chromatography tandem spectrometry and MoNA (MassBank of North America) that did not exist in breast milk or media alone. Using this database, we determined that this was a peak that was consistent with ILA, a known metabolite of tryptophan [Meng et al., unpubl. data]. We then tested various doses of ILA in H4 cells, NEC enterocytes, and orga noids from 2 additional aborted fetuses and found that on a dose-response basis

Effect of Breast Milk and Microbiota on the Immature Intestine 107 Lumen protein (1) Degradation Gut microbiota Tryptophan O – B. infantis – and other bacteria Serotonin pathway OH (under gut microbiota) NH NH2 Kynurenine pathway (4) (under gut microbiota) (2) 5-HT (3) Indole/AHR pathway (under B. infantis) Kynurenic acid

N O H OH OH N H ILA IL-1β Pathogen related (5) (6) inflammatory cytokine

AHR

5-HT: 5-hydroxytryptamine AHR: aryl hydrocarbon receptor IL-8 ILA: indole-3-lactic acid Fig. 4. Cartoon of the suggested mechanism of the anti-inflammatory effect of the tryptophan catabolite indole-3-lactic acid (ILA) in B. infantis secretions on an intestinal epithelial cell. (1) Degradation of lumen protein leads to the release of tryptophan (Trp). Under the influence of the gut microbiota, Trp is converted to (2) 5-hydroxytryptamine (5-HT) by the serotonin pathway, (3) ILA by the indole/aryl hydrocarbon receptor (AHR) pathways, and (4) kynurenic acid by the kynurenine pathway. ILA acts on AHR found in fetal enterocytes (5) thereby affecting the innate immune response in a ligand-specific fashion suppressing a pathogen-mediated inflammatory cytokine IL-1β-induced IL-8 secretion (6).

the compounds were anti-inflammatory, suggesting a generalized effect in immature intestine. The ILA was anti-inflammatory only in fetal mouse intestine but not in mature mouse intestine. When we tested it in Caco2 cells as a repre- sentative human mature enterocyte and in cortisone-treated H4 cells (a more mature version of H4 cells), we found that ILA was not anti-inflammatory [Meng et al., unpubl. data]. Since indole metabolites have a natural affinity for the receptor/transcription factor aryl hydrocarbon receptor (AHR), which functions to stimulate the IL-8 response as a transcription factor, we tested to see if an inhibition of that receptor affected the anti-inflammatory response. We noted that it did affect the anti-inflammatory response, suggesting it was a mecha-

108 Walker/Meng SCFA SCFA IL-1β SCFA receptor (109A) SCFA receptor HDACs (GPR 109A) inhibitors IL-1β (1) Increase (TSA) IL-1β receptor (2) HDAC Histon acylation (3) (2) activity (3) Increase the histon (1) Reduce histonacetylation acetylation of DNA of DNA Histon acylation

(2) (3)

Resulting in HDACs: Histone deacetylases (1) GPR 109A: G-protein-coupled receptor 109A TSA: Trichopstatin A The increase of proinflammatory cytokines (IL-8, IL-6 secretion)

Fig. 5. Short-chain fatty acids (SCFAs) decrease IL-1β-induced IL-8 secretion by inhibition of histone deacetylases (HDAC) activity in immature enterocytes. (1) IL-1β increases HDAC activity that will reduce the histone acetylation of DNA of the immature enterocytes resulting in the induction of proinflammatory cytokine secretion such as IL-8 and IL-6. (2) Via SCFA receptor, e.g., G-protein-coupled receptor 109A (GPR 109A), SCFAs inhibit HDAC activity leading to the increase in histone acetylation of DNA resulting in the inhibition of IL-1β- induced IL-8 and IL-6 induction. (3) HDAC inhibitors, e.g., trichostatin A (TSA), can inhibit HDAC activity leading to the increase in histone acetylation of DNA resulting in the inhibition of IL-1β-induced IL-8 and IL-6 induction. nism of the reaction. Figure 4 depicts our concept of the inflammatory response. Furthermore, when we examined the AHR receptor in Caco2 cells and cortisone-treated H4 cells, we found that they existed to a much lesser extent than fetal enterocytes. Accordingly, we suggest that breast milk also functions to protect premature infants by allowing B. infantis to metabolize tryptophan to produce an anti-inflammatory metabolite. This in part explains why probiotics work well in breastfed infants.

Additional Effects of Breast Milk Interaction with Bacterial Excessive Inflammation in Premature Infants

When the complex carbohydrates of breast milk interact with B. fragilis, it pro- duces short-chain fatty acids (SCFAs). We know from animal studies that SCFAs interacting with G-protein-coupled receptors can stimulate regulatory T

Effect of Breast Milk and Microbiota on the Immature Intestine 109 cells via the GPR 43 receptor to activate the original cells and to produce IL-10, which inhibits inflammation [44]. What happens to SCFAs in enterocytes of the premature intestine? Using SCFAs at various doses, we examined their effects on H4 cells, NEC cells, and organoids from other therapeutically aborted fetuses [45]. We found that the SCFAs were anti-inflammatory with these specimens, suggesting that they can affect the immature intestinal an anti-inflammatory response. When the SCFAs were used in fetal and adult mice, they were anti- inflammatory in both but worked by a different mechanism [Zheng et al., unpubl. data]. GPR 109A mediated the downregulated histone deacetylase 3 and 5 in the immature intestine but not in the mature intestine although SCFAs are anti-inflammatory in the mature intestine by a different mechanism. Figure 5 depicts that response with the immature intestine. In summary, breast milk can help to solve the premature infant’s problem with excessive inflammation by interacting with “pioneer” bacteria (B. infantis and B. fragilis) to produce metabolites that are effectively anti-inflammatory. These observations suggest another mechanism by which breast milk is anti-inflammatory in the premature infant, suggesting that breast milk works by multiple mechanisms and can protect the infant during the transition from the womb to the extrauterine environment.

Acknowledgments

This study was supported by the following grants to PI W. Allan Walker: − NIDDK (P01-DK033506) “Barrier function of the GI tract in health and disease” and − Beth Israel/Deaconess Medical Center (Award #01027741) “Impact of breastmilk in premature intestinal colonization.”

Disclosure Statement

No conflicts of interest, financial or otherwise, are declared by the authors.

References

1 Neu J, Walker WA: Necrotizing enterocolitis. N 4 Patel AL, Kim JH: Human milk and necrotizing

Engl J Med 2011; 364: 255–264. enterocolitis. Semin Pediatr Surg 2018; 27: 34–38. 2 Holman RC, Stoll BJ, Curns AT, et al: Necrotising 5 Ganguli K, Walker WA: Treatment of necrotizing enterocolitis hospitalisations among neonates in enterocolitis with probiotics. Gastroenterol Clin

the United States. Paediatr Perinat Epidemiol North Am 2012; 41: 733–746.

2006; 6: 498–507. 3 Henry MCW, Moss RL: Necrotizing enterocolitis.

Annu Rev Med 2009; 60: 111–124.

110 Walker/Meng 6 Repa A, Thanhaeuser M, Endress D, et al: Probi- 18 Henderson G, Anthony MY, McGuire W: Formu- otics (Lactobacillus acidophilus and Bifidobac- la milk versus maternal breast milk for feeding terium infantis) prevent NEC in VLBW infants preterm or low birth weight infants. Cochrane fed breast milk but not formula. Pediatr Res Database Syst Rev 2007;4:CD002972.

2015; 77: 381–885. 19 Gross SJ: Growth and biochemical response of 7 Ganguli K, Meng D, Rautava S, et al: Probiotics preterm infants fed human milk or modified in- prevent necrotizing enterocolitis by modulating fant formula. N Engl J Med 1983;308:237–241. enterocyte genes that regulate innate immune- 20 Tyson JE, Lasky RE, Mize CE, et al: Growth, met- mediated inflammation. Am J Physiol Gastroin- abolic response, and development in very-low-

test Liver Physiol 2013; 304:G132–G141. birth-weight infants fed banked human milk or 8 Jiang F, Meng D, Zhu W, et al: The symbiotic enriched formula. I. Neonatal findings. J Pediatr bacterial surface factor polysaccharide A on Bac- 1983; 103:95-104. teroides fragilis inhibits IL-1β-induced inflam- 21 Lucas A, Gore SM, Cole TJ, et al: Multicentre trial mation in human fetal enterocytes via toll recep- on feeding low birthweight infants: effects of diet

tors 2 and 4. PLoS One 2017; 12:e0172738. . on early growth. Arch Dis Child 1984; 59:722– 9 Nanthakumar N, Meng D, Goldstein AM, et al: 730. The mechanism of excessive intestinal inflamma- 22 Schanler RJ, Garza C, Nichol BL: Fortified moth- tion in necrotizing enterocolitis: an immature ers’ milk for very low birth weight infants: results

innate immune response. PLoS One 2011; 6: of growth and nutrient balance studies. J Pediatr e17776. 1985; 107:437–45. 10 Claud EC, Lu L, Anton PM, et al: Developmen- 23 Walker WA, Iyengar RS: Breast milk, microbiota, tally regulated IκB expression in intestinal epi- and intestinal immune homeostasis. Pediatr Res

thelium and susceptibility to flagellin-induced 2015; 77: 220–228.

inflammation. Proc Natl Acad Sci USA 2004; 101: 24 Palmer C, Bik EM, DiGiulio DB, et al: Develop- 7404–7408. ment of the human infant intestinal microbiota.

11 Nanthakumar N, Fusunyan RD, Sanderson IR, PLoS Biol 2007; 5:e177. Walker WA: Inflammation in the developing hu- 25 Weng M, Walker WA: The role of gut microbiota man intestine: a possible pathophysiologic basis in programming the immune phenotype. J Dev

for necrotizing enterocolitis. Proc Natl Acad Sci Orig Health Dis 2013; 4: 203–214.

USA 2000; 97: 6043–6048. 26 Jost T, Lacroix C, Braegger CP, et al: Vertical 12 Meng D, Zhu W, Ganguli K, et al: Anti-inflamma- mother-neonate transfer of maternal gut bacteria

tory effects of Bifidobacterium longum subsp via breastfeeding. Environ Microbiol 2014; 16: infantis secretion on fetal human enterocytes are 2891–2904. mediated by TLR4 receptors. Am J Physiol Gas- 27 Gareau MG, Sherman PM, Walker WA: Probiot-

trointest Liver Physiol 2016; 311:G744–G753. ics and the gut microbiota in intestinal health

13 Meng D, Zhu W, Shi HN, et al: Toll-like recep- and disease. Nat Rev Gastroenterol Hepatol 2010;

tor-4 in human and mouse colonic epithelium is 7: 503–514. developmentally regulated: a possible role in nec- 28 Donnet-Hughes A, Schriffin E, Walker WA: Pro-

rotizing enterocolitis. Pediatr Res 2015; 77: 416– tective properties of human milk and bacterial 424. colonization of the neonatal gut; in Duggan C, 14 Cahill C, Zhu W, Oziolor E, et al: Differential ex- Koletzko B, Watkins J, Walker WA (eds): Nutri- pression of the activator protein 1 transcription tion in Pediatrics – Basic Science and Clinical factor regulates interleukin-1β induction of inter- Aspects, ed 5. New Haven, Chinese Publications, leukin 6 in the developing enterocyte. PLoS One 2015, chapt 30, pp 165–250.

2016; 11:e0145184. 29 Kitajima H, Sumida Y, Tanaka R, et al: Early ad- 15 Lu L, Li T, William G, et al: Hydrocortisone in- ministration of Bifidobacterium breve to preterm duces changes in gene expression and differentia- neonates: randomised controlled trial. Arch Dis tion in immature human enterocytes. Am J Child Fetal Neonatal Ed 1997;76:F101–F107.

Physiol Gastrointest Liver Physiol 2011; 30 Dani C, Biadaioli R, Bertini G, et al: Probiotics 300:G425–G432. feeding in prevention of urinary tract infection, 16 Lu L, Bao Y, Khan A, et al: Hydrocortisone mod- bacterial sepsis and necrotizing enterocolitis in ulates cholera toxin endocytosis by regulating preterm infants. A prospective double-blind immature enterocyte plasma membrane phos- study. Biol Neonate 2002;82:103–108.

pholipids. Gastroenterology 2008; 135: 185–193. 31 Costalos C, Skouteri V, Gounaris A, et al: Enteral 17 Jost T, Lacroix C, Braegger CP, Chassard C: New feeding of premature neonates with Saccharomy- insights in gut microbiota establishment in ces boulardii. Early Hum Dev 2003;74:89–96.

healthy breast fed neonates. PLoS One 2012; 7:e44595.

Effect of Breast Milk and Microbiota on the Immature Intestine 111 32 Bin-Nun A, Bromiker R, Wilschanski M, et al: 39 Rouge´ C, Piloquet H, Butel MJ, et al: Oral sup- Oral probiotics prevent necrotizing enterocolitis plementation with probiotics in very-low-birth- in very low birth weight neonates. J Pediatr weight preterm infants: a randomized, double- 2005;147:192–196. blind, placebo-controlled trial. Am J Clin Nutr 33 Lin HC, Su BH, Chen AC, et al: Oral probiotics 2009;89:1828–1835. reduce the incidence and severity of necrotizing 40 Deshpande G, Rao S, Patole S, Bulsara M: Updat- enterocolitis in very low birth weight infants. ed meta-analysis of probiotics for preventing Pediatrics 2005;115:1–4. necrotizing enterocolitis in preterm neonates.

34 Manzoni P, Mostert M, Leonessa ML, et al: Oral Pediatrics. 2010; 125: 921–930. supplementation with Lacibacillus casei subspe- 41 Laursen MF, Bahl MI, Michaelsen KF, Licht TR: cies rhamnosus prevents enteric colonisation by First foods and gut microbes. Front Microbiol

Candida species in preterm neonates: a ran- 2017; 8: 356. domised study. Clin Infect Dis 2006;42:1735– 42 Roager HM, Licht TR: Microbial tryptophan ca- 1742. tabolites in health and disease. Nat Commun

35 Mohan R, Koebnick C, Schildt J, et al: Effects of 2018; 9: 3294. Bifidobacterium lactis Bb12 supplementation on 43 Aragozzini F, Ferrari A, Pacini N, Gualandris R: intestinal microbiota of preterm neonates: a dou- Indole-3-lactic acid as a tryptophan metabolite ble placebo controlled, randomized study. J Clin produced by Bifidobacterium spp. Appl Environ

Microbiol 2006;44:4025–4031. Microbiol 1979; 38: 544–546. 36 Stratiki Z, Costalos C, Sevastiadou S, et al: The 44 Smith PM, Howitt MR, Panikov N, et al: The mi- effect of a Bifidobacter supplemented bovine milk crobial metabolites, short-chain fatty acids, regu-

on intestinal permeability of preterm infants. late colonic Treg cell homeostasis. Science 2013;

Early Hum Dev 2007;83:575–579. 341: 569–573. 37 Lin HC, Hsu CH, Chen HL, et al: Oral probiotics 45 Senger S, Ingano L, Freire R, et al: Human fetal- prevent necrotizing enterocolitis in very low birth derived enterospheres provide insights on intesti- weight preterm infants: a multicenter, random- nal development and a novel model to study nec- ized, controlled trial. Pediatrics 2008;122:693– rotizing enterocolitis (NEC). Cell Mol

700. Gastroenterol Hepatol 2018; 5: 549–568. 38 Samanta M, Sarkar M, Ghosh P, et al: Prophylac- tic probiotics for prevention of necrotizing enterocolitis in very low birth weight newborns. J Trop Pediatr 2009; 55:128–131.

112 Walker/Meng Microbiology of Milk and Lactation: Influence on Gut Colonization

Summary on Microbiota of Milk and Lactation: Influence on Gut Colonization

The second session was devoted entirely to microbiota in breast milk. Further evidence was presented to support breastfeeding during the first 6 months of life by considering new aspects of the composition of breast milk which supports a healthy initial colonization in the newborn. Samuli Rautava stated that one of the major important functions of breastfeeding is the contribution it makes to establishing a healthy, normal initial colonization of bacteria. Breast milk provides an environment that facilitates the growth of health-promoting bacteria through the fermentation of milk constituents, such as oligosaccharides, which are metabolized and act as a substrate for bacteria, and the lower pH environment facilitates the growth of bifidobacteria and lactobacilli. In addition, breast milk has its own microbiome which is comprised of microbes from various sites, including the maternal gut. This enteromammary pathway is facilitated by hormonal influence on the intestinal microorganisms to be taken up through the intercellular pathway in the gut, engulfed by microphages, and transported to the breast. To date, it is not entirely clear to what extent the breast milk microbiome contributes to the ultimate colonization of the infant. This is an area of intense research. If the breast milk microbiome can be shown to influence intestinal colonization, this opens up a novel way for breastfeeding to be used in the prevention of disease. Leónides Fernández and Juan M. Rodríguez reported on human milk microbiota. Juan M. Rodríguez was the first to describe human milk microbiota, and his research has principally addressed this area. The human microbiota contains a low bacterial load consisting of Staphylococcus, Streptococcus, Corynebacteri- um, Propionibacterium, and gram-positive bacteria such as lactobacilli and bifidobacteria. Gram-negative anaerobic bacterial DNA can also be detected. This suggests that the source of these microbiota are several sites within the body. Evidence in animal models and humans suggests that the breast milk microbiota contributes to the infant’s gut microbiota, and there is some evidence for its function in the gut microbiota, although this is an area of ongoing research. For example, the milk microbiota in the infant’s gut may contribute to the development of innate and adaptive immunity. The enteromammary system may provide new approaches to prevent and treat disease. Sindhu Mohandas and Pia S. Pannaraj discussed a new topic with regard to the microbiome of the gut, namely the human milk virome. Evidence exists that human milk viruses are transmitted from mother to infant via breastfeeding. These viruses can help to shape the bacterial content of breast milk via bacteriophages transmitted via DNA and RNA to the colonizing bacteria. These bacteriophages can contribute to the dynamic effect of bacteria during initial colonization by contributing important genes to the organisms. In addition, the viral content of maternal milk may be important in the initial development of host defense. However, more studies are needed. Joseph Neu stated that the fetus first encounters microbes in the intrauterine environment based on cultured and molecular biologic studies in the placenta and amniotic fluid. An association exists between components of breast milk secretory IgA, lactoferrin, oligosaccharides, and microbiota and the protective health of the infant. Mechanisms whereby the actual effect is shown are necessary before we can unequivocally suggest a breast milk constituent is protective. Therefore, additional studies need to be done which address mechanisms of protection and not just associations. W. Allan Walker provides new evidence for the protective effect of breast milk. He showed that the composition of breast milk interacting with health- promoting bacteria, such as Bifidobacterium infantis and Bacteroides fragilis, can produce metabolites which are anti-inflammatory. For example, indole-lactic acid, a metabolite of breast milk tryptophan interacting with B. infantis, has been shown to help in the prevention of necrotizing enterocolitis caused by excessive intestinal inflammation. Similarly, short-chain fatty acids produced by B. fragilis interacting with breast milk complex carbohydrates can affect inflammation via the 109A G-protein receptor in immature enterocytes to produce short-chain fatty acids – another mechanism of protection against necrotizing enterocolitis. Session II provides additional data on the breast milk effect in protecting newborns in the extrauterine environment. It emphasizes the importance of colonization in this process. W. Allan Walker

114 Walker Protective Factors in Human Milk

Human Milk Oligosaccharides: Structure and Functions

Lars Bode Division of Neonatology and Division of Gastroenterology and Nutrition, Department of Pediatrics, Larsson-Rosenquist Foundation Mother-Milk-Infant Center of Research Excellence (MOMI CORE), University of California, La Jolla, CA, USA

Abstract Oligosaccharides are a group of complex glycans that are present in the milk of most mammals. However, human milk is unique as the concentrations of human milk oligosaccharides (HMOs) are much higher than those of other mammals, and their structural composition is more complex and varies between women. These observations prompt several questions: (i) Why are humans unique when it comes to milk oligosaccharides? (ii) Which maternal genetic and environmental factors drive the interindividual variation in HMO composition? (iii) What are the short- and long-term health benefits for the infant – and potentially also the mother? The combination of genome-wide association studies, milk transcriptomics, in vitro gene editing, and in silico pathway modeling allows us to reconstruct HMO biosynthetic pathways. Using new data mining approaches and leveraging samples and metadata from large mother-infant cohorts enable us to identify associations between HMO composition and infant and maternal health outcomes. Suitable preclinical models and clinical intervention studies allow us to corroborate the established associations for causal relationships and test for in vivo efficacy in humans. Knowledge generated from these different approaches will help us establish true structure-function relationships and provide the rigorous evidence required to improve infant health and development. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel What Are Human Milk Oligosaccharides?

Oligosaccharides (from the Greek ὀλίγος olígos, “a few,” and σάκχαρ sácchar, “sugar”) are saccharide (sugar) polymers containing a small number (typically 3–10 or more) of monosaccharides (simple sugars). Unlike the milk of most other mammals, human milk is unique as it contains a variety of more than 150 different and structurally distinct oligosaccharides at high concentrations. In fact, with 5–15 g/L, the total concentration of human milk oligosaccharides (HMOs) in mature milk often exceeds the total concentration of human milk proteins, making HMOs the third most abundant component after the simple milk sugar lactose and lipids, and not counting water [1]. HMOs contain up to 5 different building blocks (monosaccharides): glucose (Glc), galactose (Gal), N-acetylglucosamine (GlcNAc), fucose (Fuc), and sialic acid (Sia). Different HMOs are generated depending on which and how many of these building blocks are used, and how they are linked together [1]. Figure 1a shows the blueprint of the HMO structure assembly. All HMOs carry lactose (Galβ1–4Glc) at the reducing end. Lactose can be elongated by the addition of the disaccharides lacto-N-biose (Galβ1–3GlcNAc) or N-acetyllactosamine (Galβ1–4GlcNAc). Lactose or the elongated chains can be modified with sialic acid in α2–3- or α2–6-linkage and/or fucosylated in α1–2-, α1–3-, or α1–4- linkage, vastly expanding the diversity of the HMO structure portfolio. For example, each sialic acid monosaccharide contains a carboxyl group and intro- duces a negative charge to the HMO molecule altering its structural properties. The HMO structure often determines its functions [2]. Although the HMO composition follows a basic blueprint and more than 150 different HMOs have been identified so far, it is important to note that every woman synthesizes and secretes a distinct HMO composition profile that varies substantially between different women (Fig. 1b) but remains fairly constant over the course of lactation for the same woman [3]. So far, our lab has analyzed the HMO composition in more than 10,000 milk samples collected from women around the world as part of various collaborative projects. Figure 1c presents some of the data in a principal component (PC) plot, highlighting once again that HMO composition profiles vary between women, but also that there are distinct HMO profile clusters or HMO lactotypes.

116 Bode 25 Variable Lacto-N -biose Lactose 2’FL 3’FL LNnT 20 3’SL α1–2 DFLac 6’SL α1–3 LNT α1–4 β1–3 β1–3 β1–4 15 LNFP I LNFP II β1–4 β1–6 LNFP III α2–3 LSTb 10 LSTc α2–6 n = 0–15 DFLNT

HMO, µmol/mL LNH DSLNT 5 FLNH a N-acetyllactosamine DFLNH FDSLNH DSLNH 0 b Subjects (n = 1.206)

Genetics

Nutrition

PC2 (9.43%) Medications

Physical activity

Health status

Supplements (e.g., probiotics) PC1 (41.39%) PC3 (5.53%) d c

Fig. 1. Human milk oligosaccharide (HMO) composition varies between mothers, which is driven by genetics as well as environmental modifiers. a HMOs consist of the 5 monosaccharide building blocks: glucose (Glc, blue circle), galactose (Gal, yellow circle), N-acetylglucosamine (GlcNAc, blue square), fucose (Fuc, red triangle), and N-acetylneuraminic acid (NeuAc, purple diamond), and the HMO structural composition follows a basic blueprint. b HMO composition varies between mothers as exemplified by data from 1,206 mothers in the CHILD cohort [3]. Each bar on the x-axis represents a milk sample, each color is a specific HMO with concentrations indicated on the y-axis. c Principal component (PC) analysis of the HMO composition in over 10,000 milk samples collected from around the world shows how different HMO lactotypes cluster in different areas of the 3-dimensional space. Each dot in the space represents the HMO composition of a separate milk sample. The closer the dots are in the space, the more similar the HMO composition between the samples. The farther apart the dots, the more different the HMO composition in the samples. The left/right clustering is mostly driven by single nucleotide polymorphisms in the gene encoding for the enzyme fucosyltransferase 2 (FUT2) which catalyzes the addition of fucose in α1–2-linkage. d In addition to genetics, many other maternal factors drive the observed variation in HMO composition, many of them not well studied and the underlying mechanisms poorly understood.

Human Milk Oligosaccharides 117 What Drives the Variation in Human Milk Oligosaccharide Composition?

Figure 1d showcases just a few of the potential drivers of HMO composition. Genetics appears to be one of the strongest determinants of HMO composition. In fact, most of the drastic left/right clustering along the PC1 axis in the PC plot in Figure 1c can be explained by a single nucleotide polymorphism (SNP). In other words, the difference in 1 bp out of the approximately 3 billion bp of the human genome dramatically alters the overall oligosaccharide composition of human milk, establishing the so-called secretor and nonsecretor lactotypes. The affected gene encodes the enzyme fucosyltransferase 2 (FUT2), which catalyzes the addition of fucose to lactose or the elongated HMO chain in an α1–2-linkage [4]. Specific SNPs introduce a premature stop codon in the FUT2 reading frame, leading to incomplete FUT2 enzyme synthesis and lack of function. While the milk of women with active FUT2 (secretors) contains high amounts of α1–2- fucosylated HMOs like 2 -fucosyllactose (2 FL) or lacto-N-fucopentaose (LNFP) 1, these specific HMOs are almost completely absent in milk of women with in- ′ ′ active FUT2 (nonsecretors) [5]. Lack of FUT2 activity has ripple effects throughout the entire HMO biosynthetic pathway and impacts the concentration of almost all other HMOs, not just the ones that are α1–2-fucosylated. Similar, but slightly more subtle effects can be observed with SNPs in the gene encoding the enzyme fucosyltransferase 3 (FUT3) that is linked to the Lewis blood group antigen and catalyzes the addition of fucose to lactose or the elongated HMO chain in an α1–3- or α1–4-linkage [6]. Presence or lack of FUT3 activity establishes the Lewis-positive or -negative lactotypes. FUT2 and FUT3 are only 2 of the enzymes involved in HMO biosynthesis. Most of the other biosynthetic steps and catalyzing enzymes remain poorly characterized, leaving almost an entire biosynthetic pathway in human biology to be discovered. Remarkably, the complete pathway is only activated in humans, only in females, only in the mammary gland, and only during pregnancy and lactation. This specific restriction to species, gender, tissue, and time makes the pathway difficult to study, but the challenge may hold great opportunities to better understand the uniqueness of human biology. We are currently employing a combination of genome-wide association studies, human milk transcriptomics, HMO-omics as well as in silico pathway modeling and in vitro target validation to unravel HMO biosynthesis in the human mammary gland. In addition to genetics, environmental, modifiable factors like maternal diet and physical activity, use of supplements, maternal health status, and use of medications during pregnancy and lactation may also affect HMO composition [3]. For example, our first data from animal models show that high-fat diet low- ers the amount of mouse milk oligosaccharides while physical activity raises it

118 Bode [manuscript submitted]. Other data from our team show that women who use a combination probiotic during pregnancy have significantly different HMO composition than women who served as controls and did not receive the probiotic [7]. Whether maternal health conditions like obesity, gestational diabetes, or chronic inflammatory diseases affect HMO composition remains largely unknown and is currently an area of active investigation.

What Happens to Human Milk Oligosaccharides after Ingestion?

Once ingested, HMOs resist the low stomach pH as well as degradation through pancreatic and brush border enzymes in the small intestine [8, 9], with the potential exception of type 2 chains in which the terminal β1–4-linked Gal may be cleaved off by the enzyme lactase. Approximately 1% of the ingested HMOs are absorbed and can be measured in the systemic circulation as well as in the urine [10, 11], indicating that HMO effects extend to tissues and organs other than the intestine. Most HMOs reach the distal small intestine and colon intact, where they are either metabolized by microbes or excreted with the feces.

What Are Potential Human Milk Oligosaccharide Functions?

HMOs are often considered as human milk prebiotics, serving as metabolic substrates for potentially beneficial bacteria in the infant gut and as such shaping a healthy microbiome [12]. However, we strongly believe that HMOs are more than just food for bugs. In fact, it is believed that oligosaccharides originally evolved to serve the opposite purpose: not to feed bacteria but to keep them from growing [13]. Milk – or its evolution ancestor – is believed to have developed as a secretion to keep eggs moist. With the moisture came the risk of bacterial and fungal growth and contamination, which required the development of antimicrobial components, and oligosaccharides were likely part of that antimicrobial defense system. Although lactose is an integral structural part of all HMOs, lactose itself likely developed later during evolution, maybe as additional energy source. Thus, from the very beginning, HMOs developed as antimicrobials, and their additional prebiotic effects likely evolved much later. Research has shown that specific HMOs serve as bacteriostatic molecules that stop the growth of bacteria like group B streptococci [14]. HMOs also serve as antiadhesives, mimicking epithelial cell surface receptors used by many viruses, bacteria, and protozoan parasites to attach to host surfaces as a requirement for microbes to find their niche, proliferate, and in some cases invade and cause disease [1]. Thus, HMOs are soluble decoy receptors that prevent

Human Milk Oligosaccharides 119 microbes from binding to epithelial cells. As a consequence, potential pathogens are unable to attach, proliferate, and cause disease [1]. Antiadhesive effects of HMOs have been described for Campylobacter jejuni strains in vitro and in animal models [15], and the same HMOs are associated with lower Campylobacter diarrhea in a mother-infant cohort [16]. We have since been able to identify specific HMOs to also block enteropathogenic Escherichia coli (EPEC) adhesion and lesion formation in tissue culture as well as in mice [17; unpublished data]. Prebiotic or antimicrobial? Those effects do not have to be mutually exclusive. Commensals like certain Bifidobacteria seem to prefer simple, low-molecular- weight HMOs [18], whereas the antimicrobial properties seem to depend on higher-molecular structures. Smaller HMOs like fucosyllactoses, sialyllactoses, or lacto-N(neo)-tetraoses may serve as food for bugs, be preferentially utilized as metabolic substrates, and therefore protect the higher-molecular-weight HMOs from being degraded and available to exert their antimicrobial properties. However, evolution continues, and certain microbes may have started to ex- ploit human milk components to their advantage. We have recently shown that a specific rotavirus strain with a G10P [11] spike protein increases its infectivity in tissue culture assays in the presence of specific HMOs [19]. The same HMOs are at higher concentrations in the milk of women whose infants develop symptomatic rotavirus infections. Are pathogens getting ahead in the host-microbe arms race? Or can we leverage the gained knowledge to develop new vaccination strategies? Rotavac, a live attenuated vaccine against rotavirus, also increases its infectivity in tissue culture assays in the presence of the identified HMOs, pointing to new opportunities to include HMOs to boost vaccination success. In addition to modifying host-microbe interactions, HMOs are associated with infant growth and body composition. We have shown that specific HMOs associate with infant weight, lean mass, and fat mass in 25 mother-infant dyads in the US [20]. Employing the same HMO analytical platform, we have identified specific HMOs that are positively or negatively associated with excessive weight gain in exclusively breastfeed infants in 30 mother-infant dyads in Den- mark [21]. Remarkably, lacto-N-neo-tetraose (LNnT) was negatively associated with body fat and body weight in both studies, and 2 FL was positively associated with infant weight gain. While these are rather small cohorts, we have re- ′ cently completed the analysis of HMO composition in a larger cohort study including 802 mother-infant dyads in Finland [manuscript submitted], and the very same HMOs were significantly associated with infant weight gain all the way out to 5 years of age, long after breastfeeding occurred, suggesting long- term effects on infant growth and body composition. It is important to note that the data stem from association studies and do not allow us to draw conclusions on cause-and-effect relationships, but the observations from these 3 different

120 Bode Prevention Treatment First 1,000 days

Effects at every stage of the life cycle (“synthetic” HMOs)

Effects on pregnant and breastfeeding women

Immediate Immediate short-term short-term effects effects

Long-term effects (Developmental Origins of Health and Disease, DOHaD)

Fig. 2. The role of human milk oligosaccharides (HMOs) throughout life. HMOs have the potential to affect health and development at all stages of the life cycle. At one end of the spectrum, HMOs already appear in the amniotic fluid with potential immediate as well as long- term effects on the fetus. At the other end of the spectrum, HMOs reduce chronic inflammation and may serve as novel therapeutics to prevent or treat patients with heart attacks or stroke caused by atherosclerosis.

studies drive the need to understand causalities and underlying mechanisms, and may provide opportunities to use HMOs to support infant growth and weight gain when needed.

Can We Harness the Power of Human Milk Oligosaccharides to Develop New Therapeutics for Adults?

Rapidly accumulating data strongly suggests that HMOs have immediate benefits for infants with potential long-lasting effects throughout life, adding to the concept of Developmental Origins of Health and Disease (DOHaD) and the im-

Human Milk Oligosaccharides 121 portance of the first 1,000 days. Recently, we have shown that HMOs already appear in the amniotic fluid [22], suggesting they may also affect fetal health and development, again with potential long-lasting effects for life. However, HMOs may not only be good for infants, they may also affect maternal immediate and long-term health. HMOs appear in the maternal circulation as early as at the end of the first trimester of pregnancy and are excreted intact in maternal urine all throughout lactation [23]. Last, but certainly not least, HMOs now become available at large scale and fairly low cost, mostly because of recent advances in bioengineering microbes that utilize simple sugars to synthesize complex HMOs [24]. These synthetic, but structure-identical HMOs are now added to first infant formula products, but their application outside the maternal-infant space is also explored. First studies indicate the use of specific HMOs like 2 FL as novel therapeutics to improve gut health, and we have recently identified specific HMOs that accelerate macro- ′ phage resolution and reduce chronic inflammation in cell culture, and significantly reduce arthritis and atherosclerosis in an animal model [manuscripts submitted]. Most importantly, HMOs are natural components that are highly abundant in human milk and fed to infants every 2–3 h for several months. HMOs evolved to be safe and support infant immediate and long-term health and development, and they are likely safe for application in adults as well, opening new opportunities to develop HMOs as novel therapeutics for some of today’s most common, painful and often deadly diseases. In conclusion, we have now reached a point where HMOs are no longer an overlooked component of human milk, but considered to play a role in health promotion and disease prevention and treatment throughout the entire life span, from development of the fetus to frail- ty in the elderly (Fig. 2).

Disclosure Statement

The author declares no conflicts of interest related to this chapter.

122 Bode References

1 Bode L: Human milk oligosaccharides: every 14 Lin AE, Autran CA, Szyszka A, et al: Human milk

baby needs a sugar mama. Glycobiology 2012; 22: oligosaccharides inhibit growth of group B strep-

1147–1162. tococcus. J Biol Chem 2017; 292: 11243–11249. 2 Jantscher-Krenn E, Bode L: Structure-function 15 Ruiz-Palacios GM, Cervantes LE, Ramos P, et al: relationships of human milk oligosaccharides. Campylobacter jejuni binds intestinal H (O)

Adv Nutr 2012; 3: 383S–391S. antigen (Fucα1,2Galβ1, 4GlcNAc), and fucosyloli 3 Azad MB, Robertson B, Atakora F, et al: Human gosaccharides of human milk inhibit its binding

milk oligosaccharide concentrations are associ- and infection. J Biol Chem 2003; 278: 14112– ated with multiple fixed and modifiable maternal 14120. characteristics, environmental factors and feed- 16 Morrow AL, Ruiz-Palacios GM, Altaye M, et al:

ing practices. J Nutr 2018; 148: 1733–1742. Human milk oligosaccharides are associated with 4 Kumazaki T, Yoshida A: Biochemical evidence protection against diarrhea in breast-fed infants.

that secretor gene, Se, is a structural gene encod- J Pediatr 2004; 145: 297–303. ing a specific fucosyltransferase. Proc Natl Acad 17 Manthey CF, Autran CA, Eckmann L, Bode L:

Sci USA 1984; 81: 4193–4197. Human milk oligosaccharides protect against 5 Stahl B, Thurl S, Henker J, et al: Detection of four enteropathogenic E. coli attachment in vitro and human milk groups with respect to Lewis-blood- colonization in suckling mice. J Pediatr Gastroen-

group-dependent oligosaccharides by serologic terol Nutr 2014; 58: 167–170. and chromatographic analysis. Adv Exp Med Biol 18 LoCascio RG, Ninonuevo MR, Freeman SL, et al:

2001; 501: 299–306. Glycoprofiling of bifidobacterial consumption of 6 Johnson PH, Watkins WM: Purification of the human milk oligosaccharides demonstrates Lewis blood-group gene associated α-3/4- strain specific, preferential consumption of small fucosyltransferase from human milk: an enzyme chain glycans secreted in early human lactation. J

transferring fucose primarily to type 1 and lac- Agric Food Chem 2007; 55: 8914–8919. tose-based oligosaccharide chains. Glycoconj J 19 Ramani S, Stewart CJ, Laucirica DR, et al: Com-

1992; 9: 241–249. plex interplay between human milk oligosaccha- 7 Seppo AE, Kukkonen AK, Kuitunen M, et al: As- rides, milk microbiome and infant gut microbi- sociation of maternal probiotic supplementation ome modulates neonatal rotavirus infection. Nat

with human milk oligosaccharide composition. Commun 2018; 9: 5010.

JAMA Pediatr 2019; 173: 286–288. 20 Alderete TL, Autran CA, Brekke BE, et al: Asso- 8 Ruhaak LR, Stroble C, Underwood MA, et al: De- ciations between human milk oligosaccharides tection of milk oligosaccharides in plasma of in- and infant body composition in the first six

fants. Anal Bioanal Chem 2014; 406: 5775–5784. months of life. Am J Clin Nutr 2015; 102: 1381– 9 Goehring KC, Kennedy AD, Prieto PA, et al: 1388. Direct evidence for the presence of human milk 21 Larsson MW, Lind MV, Laursen RP, et al: Hu- oligosaccharides in the circulation of breastfed man milk oligosaccharide composition is associ-

infants. PLoS One 2014; 9:e101692. ated with excessive weight gain during exclusive 10 Rudloff S, Obermeier S, Borsch C, et al: Incorpo- breastfeeding – an explorative study. Front

ration of orally applied (13)C-galactose into milk Pediatr 2019; 7: 297.

lactose and oligosaccharides. Glycobiology 2006; 22 Wise A, Robertson B, Choudhury B, et al: Infants

16: 477–487. are exposed to human milk oligosaccharides al-

11 Dotz V, Rudloff S, Blank D, et al: 13C-labeled oli- ready in utero. Front Pediatr 2018; 6: 270. gosaccharides in breastfed infants’ urine: individ- 23 Jantscher-Krenn E, Aigner J, Reiter B, et al: Evi- ual-, structure- and time-dependent differences dence of human milk oligosaccharides in mater-

in the excretion. Glycobiology 2014; 24: 185–194. nal circulation already during pregnancy – a pilot

12 German JB, Freeman SL, Lebrilla CB, et al: Hu- study. Am J Physiol Endocrinol Metab 2019; man milk oligosaccharides: evolution, structures 316:E347–E357. and bioselectivity as substrates for intestinal bac- 24 Bode L, Contractor N, Barile D, et al: Overcoming teria. Nestlé Nutr Workshop Ser Pediatr Program the limited availability of human milk oligosac-

2008; 62: 205–218; discussion 218–222. charides: challenges and opportunities for re-

13 Oftedahl OT: The evolution of milk secretion and search and application. Nutr Rev 2016; 74: 635–

its ancient origins. Animal 2012; 6: 355–368. 644.

Human Milk Oligosaccharides 123 Protective Factors in Human Milk

Oligosaccharides and Viral Infection: Human Milk Oligosaccharides versus Algal Fucan-Type Polysaccharides

a b Franz-Georg Hanisch Cem Aydogan a Institute of Biochemistry II, Medical Faculty, University of Cologne, Köln, Germany; b PhytoNet AG, Schindellegi-Feusisberg, Switzerland

Abstract Norovirus infections belong to the most common causes of human gastroenteritis worldwide, and epidemic outbreaks are responsible for hundreds of thousands deaths annually. Strikingly, no antiviral treatment is available due to the difficulty in cultivating virions or in generating a vaccine, and due to the fact that their infection mechanisms are poorly understood. However, there is consent that noroviruses bind to histo-blood group antigens (HBGAs) on their way through the digestive tract. The HBGA profiles vary individually, making people more or less susceptible to different norovirus strains. In our current work, we tried to decipher the HBGA specificity of the most prevalent and clinically relevant norovirus GII.4 subfamily (Sydney 2012, JX459908) and its preferences for human milk oligosaccharides (HMOs) as potential anti-infectives. The structural evidence provided can explain at the molecular level why individuals with certain blood groups are at higher risk of infection, and how these infections may be prevented and treated by application of food additives. A central finding was that low-affinity binding of HMOs is surpassed by high-avidity binding of multivalent oligo- and polyfucoses as found in algal polysaccharides (fucoidans). Insight into structural details of fucoidans and their impact on noroviral-blocking efficiency is provided and discussed. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

Noroviruses belong to the family of Caliciviridae, which are small nonenveloped viruses that contain a single-stranded RNA genome surrounded by a capsid protein. The capsid is composed of 180 copies of a major capsid protein, VP1, and small numbers of a minor capsid protein, VP2 [1]. Based on variations within the VP1 gene, noroviruses are classified into 7 genogroups, termed GI–GVII [2]. Strains of the genogroup GII are responsible for most infections in humans [3]. To induce an infection, the norovirus needs to bind to the intestinal epithelia, a process that is at least partly mediated by specific interactions of a lectin-like activity in the norovirus capsid VP1 protein and blood group-active mucin-type O-glycans on membrane glycoproteins or mucins [4, 5]. Accumulating evidence suggests that fucose, as part of histo-blood group antigens (HBGAs) or Lewis- related antigens, plays an essential role in the lectin-like recognition by the VP1 capsid protein. This holds in particular true for the highly infectious GII.4 (Sydney 2012, JX459908) and the recently emerged GII.17 (Kawasaki, 2015, LC037415) noroviruses. Fucoses of ABH- and Lewis blood group antigens are introduced by the FUT2 and FUT3 fucosyltransferases, and homozygously re- cessive individuals lacking these enzyme activities were less susceptible or even resistant to an infection by certain strains in human challenge and outbreak studies [6, 7]. Certainly, in addition to the recently identified cellular protein receptor for murine noroviruses [8], there might be other host factors required for attachment and entry of human noroviruses. Besides vaccine-based approaches to combat the virus, one of the alternative strategies to prevent infections is based on competitors of epithelial receptors, which could be applied as food additives, like the human milk oligosaccharides (HMOs). HMOs represent an ideal source of potential competitors of viral glycan receptors, which mimic the structures of blood group-active O-glycans. For example, the trisaccharide 2 -fucosyllactose (2 FL) is able to block norovirus binding quite efficiently [9, 10], and it has reached market approval as a save ′ ′ food additive. Here we provide evidence for other milk oligosaccharides in the high-mass range to exert even stronger competitive effects on norovirus binding to gastric mucins [11, see below]. During our studies, we observed that oligo-valency of fucose in hepta- to decasaccharides promotes the competitive effects on norovirus binding. This became evident when L-fucose dendrimers with varying degrees of substitution were compared with respect to their competitive activity [11, see below]. High valency of α-L-fucose with no relationship to blood group structures is a feature of natural polysaccharides belonging to the group of polyfucoses or fucans.

Anti-Infectives Based on Fucoidan 125 Algal fucoidans are present in several orders, mainly Fucales and Laminari- ales, but also in Chordariales, Dictyotales, Dictyosiphonales, Ectocarpales, and Scytosiphonales. They are widely present among all the brown algae (Phaeophy- ceae) and exist either as a homopolymer of fucose or as a heteropolysaccharide. Fucoidans can be divided into two groups depending on their sources: fucoidans from Laminaria species have their central chains composed of (1→3)-linked α-L-fucose and those from Ascophyllum and Fucus species which are characterized by repeating (1→3) and (1→4) linked α-L-fucose [12]. Besides varying contents of other sugars (GlcA, Man, and Gal, for example), all fucoidans exhibit sulfation at high densities, as every second fucose can be substituted. Among fucoidans, those of the brown algae have previously attracted much attention, as they were claimed to exert a series of health beneficial effects [12]. Like other sulfated polysaccharides, fucoidans can inhibit virus infection of cells. This has been demonstrated for Herpes simplex, cytomegalovirus, and human immunodeficiency virus [13] as well as bovine viral diarrhea virus [14], probably by competing with cell surface heparan sulfate for binding to the virus; accordingly, the effect is strictly dependent on sulfation.

Human Milk Oligosaccharides: Applying Array Technology for the Identification of Potent Noroviral Binders

We screened fractions of high-molecular-mass HMOs with multiple fucose substitution for their binding capacities to the noroviral VLP capsid protein applying neoglycolipid array technology. We were able to prepare dipalmitoyl-phos- phatidylethanolamine-based neoglycolipids from over 40 complex HMO fractions containing preferentially fucosylated, high-mass glycans with different structures and with sizes in the range of 6- to 15-mers (Table 1, kindly provided by Prof. Clemens Kunz, University of Mainz). Immobilized on polystyrene surfaces, the amphipathic neoglycolipids present their carbohydrate portions in the aqueous phase forming clusters of ligands and enabling multivalent binding of the viral capsids similar to the situation on epithelial surfaces. Oligosaccharides with multiple fucosylation of the blood group H1 and Lewis-b type were found to exhibit highest binding capacities (Fig. 1) [11]. How- ever, besides valency, structural aspects were also found to play an important role. Among isomeric nonasaccharides containing 3 fucose residues, only those were active binders that carried the fucose at terminal galactoses (H1-type structures), whereas blood group Lewis-a-positive isomers were not active binders. Lewis-a antigen on HMOs or human gastric mucins was however recognized by other noroviral strains, like GII.10 (Vietnam 026, AF504671).

126 Hanisch/Aydogan Table 1. Human milk oligosaccharide (HMO) fractions and standard HMOs positive for GII.4 (Sydney, 2012, JX459908) VLPs

M+Na Composition HMO fraction/ IC50, mM standard HMO

731 H3N1 2, 107 876 F1H3N1 2 977 S1H3N1 94 1,022 F2H3N1 2 1,096 H4N2 2 1,241 F1H4N2 2a, 7, 46, 49, 54 1,270 S2H3N1 94 1,387 F2H4N2 7, 46, 107 1,388 S1H4N2 94 1,416 S2F1H3N1 94 1,420 H6N2 7, 107, 127 1,461 H5N3 54, 107 1,533 F3H4N2 2a, 7, 46 1,566 F1H6N2 107 1,634 S2H4N2 94 1,607 F1H5N3 7, 49, 51, 54, 107, 127 1,679 F4H4N2 7, 46 1,753 F2H5N3 47, 49, 51, 54, 127 1,780 S2F1H4N2 94 1,825 H6N4 7, 49 1,899 F3H5N3 49b, 51, 127 1,973 F1H6N4 127 2,483 F2H7N5 49 2,629 F3H7N5 49b 187 F1 Fucose >50 511 F1H2 2’FL 15–30 511 F1H2 3FL 15–30 876 F1H3N1 LNFP-I >50 876 F1H3N1 LNneoFP-I >50

MALDI mass spectrometry of native oligosaccharides in the positive-ion mode. S, NeuAc; F, fucose; H, hexose; N, N-acetylhexosamine. Molecular masses given in boldface represent compositional species with high fucosylation. HMO fractions in boldface contain mostly highly fucosylated species. IC50 values of VLP binding inhibition in the ELISA format are given for standard compounds of HMOs. a These components in fraction 2 were detectable in trace amounts. b HMO species in fraction 49 with >2 fucosyl residues were structurally characterized to expose terminal Lewis-a trisaccharides, but no H type 1 or Lewis-b.

Anti-Infectives Based on Fucoidan 127 1.0 46

0.9 2 0.8

0.7 Intensity, AU

0.6 7

0.5

0.4 m/z 0.3 127 0.2 47 51 54 Gll.4 VLP binding activity (OD 405 nm) 0.1 8 49 54 107 Solid-phase binding assay on immobilized neoglycolipids

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Neoglycolipid fraction

Fig. 1. Neoglycolipids of human milk oligosaccharides (HMOs) as probes for noroviral binding. VLPs bind preferentially to neoglycolipids containing oligovalent H1 and Lewis b-expressing HMOs when tested in solid-phase binding assays of GII.4 VLPs. The insets present MS and MS/MS data of methylated oligosaccharides in fraction 46. On the right hand side, structural models of the VP1 domains (S, blue; P1, red; P2, yellow) and of the VLP capsid are shown.

In-Solution Experiments with HMOs as Inhibitors of Noroviral Binding

In inhibition assays of VLP binding to human gastric mucins using free oligosaccharide competitors, neither the amounts needed to reach 50% inhibition nor the order of IC50 values was in accordance with expectations. First of all, the IC50 values were throughout in the millimolar range pointing to quite low HMO affinities. Whereas 2 FL and 3-fucosyllactose (3FL) exhibited IC50 values in the range of 15–30 mM, the blood group H1-active lacto-N-fucopentaose I (LNFP-I) ′ was far less active (IC50: > 50 mM) similar to free fucose (Table 1) [11]. There was also no discrimination in activity when comparing H1 and H2 blood group- active pentasaccharides LNFP-I and LNneoFP-I in binding inhibition assays, indicating that subterminal Gal or GlcNAc in the pentasaccharide cores are not significantly contributing to the binding. This is in contradiction to findings from other groups based on their work with P-domain dimers in STD (saturation-transfer-difference)-NMR and to binding studies with glycosphingolipids

128 Hanisch/Aydogan from small intestinal epithelium of humans (type 1 chains) and dogs (type 2 chains) [15, 16]. Neither in the ELISA format nor in surface plasmon resonance experiments with free oligosaccharides as inhibitors we could demonstrate a differential binding of type 1- and type 2-based fucopentaoses. The currently most prevalent strain GII.4 (JX459908, first described during a noroviral outbreak in 2012 in Sydney, Australia) was used in most of our experiments, and hence all the reported data refer generally to this strain. The newly emerged strain GII.17 (LC037415, first described in 2015 in Kawasaki, Japan) had been reported to exert higher virulence and to show better adaptation to the human blood group oligosaccharides. However, neither the relative affinities of P-domains to fucosylated receptors nor data on viral capsid binding to oyster tissue (see below) support the assumption of increased HBGA affinities of GII.17 strains. Oyster can function as vectors for the most common noroviruses, and although they become not infected, they could serve as a model for noroviral ac- cumulation in gastrointestinal tissues. A prerequisite for this is a moderate HBGA expression (A and H1 type) in digestive tissues and palps of the Pacific oyster Crassostrea gigas, whereas gills and mantle displayed only weak blood group expression [17]. Evidence from VLP binding studies on oyster tissues revealed however that the noroviral capsids from GII.4 (Sydney, 2012) accumu- lated on all investigated oyster tissues [17], whereas GII.17 (Kawasaki, 2015) did not bind to any of these.

Avidity Overrides Affinity: Dendrimers of α-L-Fucose Linked to α-Cyclodextrin

Cyclodextrins (CDs) are formed biotechnologically during enzymatic degradation of starch, and the safe α- and γ-CDs play an important role in food industry. α-CDs represent a cyclic glucose hexamer with unique chemical features. While the C3/C4-hydroxyl groups are chemically nearly inert, the C6-hydroxyl groups of the compound are easily accessible and highly reactive in substitution reactions. We used α-CD in an acid reversion chemistry approach to add α-L-fucose to the C6-hydroxyl groups and obtained (dependent on the molar excess of the fucose) mono- to hexa-substituted products [11]. Structural work on the products confirmed the linkage of primarily α-anomeric fucose to the C6 position of the hexa-glucose scaffold. Strikingly, the various fractions of differentially substituted Fuc-CDs revealed increasing inhibitory potentials that were correlated with the numbers of fucose in the dendrimers [11]. While fucose residues at C6 of the glucose scaffold

Anti-Infectives Based on Fucoidan 129 should exhibit a high degree of rotational freedom, the scaffold itself should be rather rigid and not contribute to the binding to the viral capsid. Hence, only the fucose in these dendrimers should be involved in the binding to VLPs. Accord- ingly, the available evidence (also from other groups [18]) suggests that despite a certain degree of blood group specificity of VLP binding, the valency-dependent avidity of fucose dramatically overrides the affinity aspect and should be considered in any design of prophylactic or therapeutic drugs or food additives.

Fucoidans and Processed Forms Are Highly Active Antinoroviral Compounds

VLPs of the norovirus strain GII.4 (Sydney, 2012, JX459908) bound in a concentration-dependent manner to immobilized native fucoidan from Fucus vesiculosus ranging from 0.2 to 100 µg/mL (not shown). For application of fucoidan as inhibitor, assays on immobilized human gastric mucins were established. GII.4 VLPs were binding actively to the gastric mucin, and this binding showed a characteristic pH dependency. Fucoidan was tested for its inhibitory capacities as native polysaccharide using different charges of commercial origin. VLP binding was effectively blocked by the native fucoidan (Fig. 2a) [11], and the competitive effect was shown to be concentration-dependent in the range from 1 to 20 mg/mL. An IC50 was determined at approximately 10 mg/mL corresponding to 260 µM for an average- sized fucoidan in crude F. vesiculosus preparations reported to be 38.2 kDa [19]. Treatment of the polyfucose with sodium periodate to oxidize peripheral fucoses (with expected involvement in VLP binding) totally abolished the inhibitory potential of fucoidan. Various routes of processing were followed to obtain fucoidan preparations of lower mass (e.g., partial acid hydrolysis, Fenton chemistry-based degradation, and hydrothermal degradation) or desulfated fucoidans (for mass spectra and linkage analysis by GC-MS, see Fig. 2b, c) [11]. With these preparations, we were able to demonstrate that the binding of fucoidan to VLPs and hence the antiviral effect was independent of sulfation (Fig. 2a). On the contrary, the observed inhibitory potential increased in accordance with the demasking of fucose residues. Moreover, the partial fragmentation to low-molecular-weight fucoidan and to oligofucoses (2- to 20-mers) gave rise to highly active preparations. X-ray cocrystallography studies with these oligofucoses revealed that only terminal single fucose residues were interacting with the P-domains of the capsid protein, indicating comparatively low affinity binding [Grant Hansman, pers. commun.].

130 Hanisch/Aydogan –7 p = 6.63 × 10 ×104 1.4 p = 1.08 × 105 1.2 5 DP10 DP15 DP20 DP25 1.0

4 11,931 13,553 16,488 15,016 10,628

0.8 17,941 19,404

3 20,866

0.6 22,339 23,501

0.4 25,264 Intensity, AU 2 26,736 28,199 29,661 on HGM (OD 405 nm) 0.2 31,134 Gll.4 VLP binding activity 1 32,586 34,049 35,509 36,969

0 38,430 100% Native Desulfated 0 a control fucoidan fucoidan 1,500 2,000 2,500 3,000 3,500 4,000

3, 4-Fuc 275 c m/z

2, 3-Fuc 262

2-Fuc 247 3-Fuc 4-Fuc 203 α α 2 α 190 3 4 α 3 Fuc 175 2 α 3 Ion current, % Ion current, α 3 α α 3 118 2 α TIC d

80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 b Time

Fig. 2. High avidity VLP blocking by fucoidan. Native fucoidan from Fucus vesiculosus and its desulfated/fragmented processing products (size range 2–20 dp) revealed strong inhibitory potentials in norovirus binding inhibition on human gastric mucins (a). Linkage analyses by GC-MS (b) support the proposed structural model that closely resembles previously published ones, but it is distinct by showing a higher degree of branching. c MALDI survey spectrum in the positive ion mode taken from underivatized oligofucoses formed during solvo- lytic desulfation of fucoidan. d Proposed model of average fucoidan from F. vesiculosus.

Our current efforts are focusing on species-specific and preparation-dependent variations of fucoidan structures and their impact on antiviral activity. An- imal model studies in oyster may provide an initial insight into possible applications of processed fucoidans as antinoroviral agents, adding to the many reported applications of this algal product [20].

Acknowledgments

The reviewed work was performed in a collaboration with groups in Mannheim (Dr. Vasily Morozov and Prof. Horst Schroten from the Children’s Hospital University Clin- ic, Germany) and in Heidelberg (Dr. Grant Hansman from the Schaller Research Group at the University of Heidelberg, Germany).

Anti-Infectives Based on Fucoidan 131 Disclosure Statement

The authors declare that they have no conflict of interests with the contents of this article.

References

1 Prasad BV, Hardy ME, Dokland T, et al: X-ray 12 Li B, Lu F, Wei X, Zhao R: Fucoidan: structure

crystallographic structure of the Norwalk virus and bioactivity. Molecules 2008; 13: 1671–1695.

capsid. Science 1999; 286: 287–290. 13 Hoshino T, Hayashi T, Hayashi K, et al: An anti- 2 Zheng DP, Ando T, Fankhauser RL, et al: Norovi- virally active sulfated polysaccharide from Sar- rus classification and proposed strain nomencla- gassum horneri (TURNER) C. AGARDH. Biol

ture. Virology 2006; 346: 312–323. Pharm Bull 1998; 21: 730–734. 3 Fankhauser RL, Monroe SS, Noel JS, et al: Epide- 14 Iqbal M, Flick-Smith H, McCauley JW: Interac- miologic and molecular trends of “Norwalk-like tions of bovine viral diarrhoea virus glycoprotein viruses” associated with outbreaks of gastroenteri- E (rns) with cell surface glycosaminoglycans. J

tis in the United States. J Infect Dis 2002; 186: 1–7. Gen Virol 2000; 81: 451–459. 4 Marionneau S, Ruvoën N, Le Moullac-Vaidye B, 15 Fiege B, Leuthold M, Parra F, et al: Epitope map- et al: Norwalk virus binds to histo-blood group ping of histo blood group antigens bound to nor- antigens present on gastroduodenal epithelial ovirus VLPs using STD NMR experiments reveals cells of secretor individuals. Gastroenterology fine details of molecular recognition. Glycoconj J

2002; 122: 1967–1977. 2017; 34: 679–689. 5 Schroten H, Hanisch FG, Hansman GS: Human 16 Nasir W, Frank M, Kunze A, et al: Histo-blood norovirus interactions with histo-blood group group antigen presentation is critical for norovi- antigens and human milk oligosaccharides. J rus VLP binding to glycosphingolipids in model

Virol 2016; 90: 5855–5859. membranes. Chem Biol 2017; 12: 1288–1296. 6 Lindesmith L, Moe C, Marionneau S, et al: Hu- 17 Morozov V, Hanisch FG, Wegner KM, Schroten man susceptibility and resistance to Norwalk vi- S: Pandemic GII.4 Sydney and epidemic GII.17

rus infection. Nat Med 2003; 9: 548–553. Kawasaki308 noroviruses display distinct speci- 7 Lopman BA, Trivedi T, Vicuna Y, et al: Norovirus ficities for histo-blood group antigens leading to infection and disease in an Ecuadorian birth co- different transmission vector dynamics in Pacific

hort: association of certain norovirus genotypes oysters. Front Microbiol 2018; 9: 2826.

with host FUT2 secretor status. J Infect Dis 2015; 18 Bücher KS, Yan H, Creutznacher R, Ruoff K, et al:

211: 1813–1821. Fucose-functionalized precision glycomacromol- 8 Orchard RC, Wilen CB, Doench JG, et al: Discov- ecules targeting human norovirus capsid protein.

ery of a proteinaceous cellular receptor for a nor- Biomolecules 2018; 19: 3714–3724.

ovirus. Science 2016; 353: 933–936. 19 Lahrsen E, Liewert I, Alban S: Gradual degrada- 9 Weichert S, Koromyslova A, Singh BK, et al: tion of fucoidan from Fucus vesiculosus and its Structural basis for norovirus inhibition by hu- effect on structure, antioxidant and antiprolifera-

man milk oligosaccharides. J Virol 2016; 90: 4843– tive activities. Carbohydr Polym 2018; 192: 208– 4848. 216. 10 Koromyslova A, Tripathi S, Morozov V, et al: 20 Fitton JH: Therapies from fucoidan: multifunc-

Human norovirus inhibition by a human milk tional marine polymers. Mar Drugs 2011; 9: 1731–

oligosaccharide. Virology 2017; 508: 81–89. 1760. 11 Hanisch FG, Hansman GS, Morozov V, et al: Avidity of α-fucose on human milk oligosaccharides and blood group-unrelated oligo/polyfucoses is essential for potent norovirus-binding

targets. J Biol Chem 2018; 293: 11955–11965.

132 Hanisch/Aydogan Protective Factors in Human Milk

Milk Fat Globule Membranes: Effects on Microbiome, Metabolome, and Infections in Infants and Children

a b a Olle Hernell Bo Lönnerdal Niklas Timby a Pediatrics, Department of Clinical Sciences, Umeå University, Umeå, Sweden; b Department of Nutrition, University of California, Davis, CA, USA

Abstract Dietary supplementation with bovine milk fat globule membrane (MFGM) concentrates has recently emerged as a possible means to improve the health of infants and young children, or defense against infections. We identified 5 double-blind, randomized, controlled trials (DBRCT) exploring the effects of supplementing the diet of infants and children with bovine MFGM concentrates on infections. We reviewed 3 studies which found a protective effect against infections at different ages during infancy and early childhood. Two of them have reported effects on the metabolome, and 1 study also on the microbiome and lipidome. MFGM supplementation had moderate, albeit interesting, effects on the oral and fecal microbiome, fecal and serum/plasma metabolome, and serum and erythrocyte membrane lipidome, which also are reviewed. We conclude that studies on MFGM supplementation during infancy and childhood indicate positive effects on the defense against infections and other outcomes, but more high-quality DBRCTs with well-defined MFGM fractions and outcome measures are needed before firm conclusions can be drawn. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel

Introduction

The composition of human milk has evolved to become uniquely optimized to meet the nutritional needs for infant growth and development during the first 4–6 months of life. One unique component of milk is the milk fat globule (MFG), which has a unique composition and complex structure with a core of mainly energy-rich triglycerides enveloped by a membrane structure, the MFG membrane (MFGM) [1]. An increasing number of studies have reported various health benefits from oral supplementation with bovine MFGM to humans in different age groups, including infants and children [2, 3]. MFGM is composed of a phospholipid and cholesterol triple layer with incorporated proteins and glycoproteins [1, 4]. Milk phospholipids, sphingomyelins, and gangliosides are largely located on the MFGM, although phospholipids are also secreted as smaller vesicles devoid of a triglyceride core, which typically separate with the whey fraction [4, 5]. The proteome of the human MFGM is very complex with several hundred identified proteins, including mucins, butyrophilin, lactoferrin, and lactadherin [1, 6, 7]. Bovine MFGM-rich fractions contain approximately the same number of proteins [8]. MFGM is also rich in sialic acid as part of gangliosides [5] and glycosylated proteins. The genes regulating MFGM synthesis are conserved across species suggesting a functional benefit of this fraction in milk [9], even if the detailed MFGM composition varies among species [7]. Sev- eral of the MFGM components have antimicrobial or immunological effects. Gangliosides play a role in the development of intestinal microbiota composition, gut immunity, and, consequently, in the defense against infections [10]. The glycoproteins butyrophilin, lactadherin, and mucin [11] have all antimicrobial effects, and the lipid fraction of bovine MFGM has an antiviral effect in vitro [12]. Breastfed infants have a higher intake of MFGM components than formula- fed infants because traditionally the MFGM fraction is discarded with the milk fat when this is replaced by blends of vegetable oils as the fat source in infant formulas. Resulting from advances in dairy technology, bovine MFGM concentrates are now commercially available and possible to use as food supplements, including infant formulas.

Clinical Studies on the Supplementation of Milk Fat Globule Membrane Concentrates to Infants and Children: Effects on Infections

In a literature search (August 10, 2019), we identified 5 double-blind, randomized, controlled trials (DBRCT) exploring the effects of supplementing the diet of infants or children with MFGM on infections (Table 1): In a Peruvian DBRCT, 550 healthy, primarily breastfed 6- to 11-month-old infants received 40 g/day of an instant complementary food fortified with 1 recommended dietary allowance (RDA) of multiple micronutrients as well as a protein source for 6 months. They were randomized to the protein source being

134 Hernell/Lönnerdal/Timby Table 1. Double-blind, randomized, controlled trials exploring the effects of MFGM supplementation to the diet of infants or children on infections compared with no supplementation

Study Age Supplementation Effects of MFGM on infections

Zavaleta et al. [13], 6–11 months MFGM (Lacprodan® Lower longitudinal prevalence of Lee et al. [25], MFGM-10; diarrhea Peru Arla Foods Ingredients) Lower incidence of bloody diarrhea

Veereman-Wauters 2.5–6 years, MFGM (INPULSE®; Fewer days with fever et al. [14], Belgium during 4 months Büllinger SA) Poppitt et al. [15], 8–24 months, Complex milk lipids No difference regarding diarrhea India during 12 weeks (Fonterra Co-operative Ltd)

Timby et al. [16, 17, 20] <2–6 months MFGM (Lacprodan® Lower incidence of otitis media He et al. [21, 24] MFGM-10; Grip et al. [22], Sweden Arla Foods Ingredients)

Li et al. [18], <1–4 months MFGM (Lacprodan® No differences regarding fever, China MFGM-10; diarrhea, or urinary tract infections Arla Foods Ingredients)

either an MFGM-enriched protein fraction (Lacprodan® MFGM-10; Arla Foods Ingredients, Viby, Denmark) or skim milk powder (control group) [13]. There was no difference between the groups in the incidence of diarrhea, but the longitudinal prevalence of diarrhea was significantly lower in the MFGM group than the control group (3.84 vs. 4.37%, p < 0.05). In a multivariate model adjusted for initial anemia and potable water facilities, the incidence of bloody diarrhea was lower in the MFGM group with an adjusted OR of 0.59 (95% CI 0.34–1.02, p = 0.025). In a Belgian DBRCT, 253 preschool children aged 2.5–6 years received 200 mL of a chocolate formula milk daily for 4 months [14]. They were randomized to a formula without supplementation (placebo group) or enriched with 500 mg of phospholipids by addition of 2.5% of a phospholipid-rich MFGM concentrate (INPULSE; Büllinger SA, Büllingen, Belgium) (intervention group). The intervention group had fewer days with fever (mean ± SD: 1.71 ± 2.47 vs. 2.60 ± 3.06, p = 0.028) and also a lower parental scoring of behavioral problems. In an Indian DBRCT, 450 infants aged 8–24 months were randomized to a daily dose of milk powder supplemented with 2 g of a spray-dried ganglioside concentrate (Fonterra Co-operative Group Ltd, Auckland, New Zeeland) or milk powder only (control group) for 12 weeks [15]. There was no difference between the groups, neither in the primary outcome rotavirus diarrhea nor in secondary outcomes including all-cause diarrhea. However, the authors noted that the incidence of rotavirus diarrhea during the study period was lower than expected, making the study underpowered compared to the intention of the design.

Immunologic and Metabolomic Effects of MFGM 135 In a Swedish DBRCT, 160 formula-fed healthy term infants were randomized to receive an experimental formula (EF) supplemented with a protein- rich MFGM fraction (Lacprodan® MFGM-10; Arla Foods Ingredients) or standard formula (SF) from < 2 to 6 months of age. The EF had lower energy density (60 vs. 66 kcal/100 mL) and protein concentration (1.20 vs. 1.27 g/100 mL). Proteins from the MFGM fraction made up 4% (wt/wt) of the total protein content in the formula. A breastfed reference group with 80 infants was also recruited. In the primary outcome (weight at 6 months of age), there was no difference between the group, but in cognitive testing at 12 months of age, the EF group achieved higher scores [16]. Among the secondary outcomes, the EF group had a lower incidence of acute otitis media than the SF group (1 vs. 9%, p = 0.034), a lower incidence and longitudinal prevalence of antipyretic use, and lower concentrations of secretory IgG against pneumococci after vaccination, all suggesting an infection-protective effect of the EF [17]. In a recent DBRCT in China (Nanjing, Shanghai, and Beijing), 600 healthy term infants were randomized to a formula supplemented with MFGM (Lacpro- dan® MFGM-10; Arla Foods Ingredients), a formula supplemented with the probiotic Lactobacillus paracasei ssp. paracasei strain F19, or to SF. A breastfed group (n = 200) was recruited as reference. The intervention lasted from age 21 ± 7 days until 4 months, and infants were followed until 1 year old. During the intervention, the MFGM and SF groups did not differ in days with fever or number of episodes of fever (> 38 ° C), upper respiratory tract infections, or diarrhea. The breastfed group had significantly fewer fever episodes (p = 0.021) and days with fever (p = 0.036) than the SF group but did not differ from the MFGM- supplemented group in any of the primary outcomes [18]. Of note is that otitis media was not an outcome in this study. In a recently published DBRCT conducted in Anhui, China, in which infection was not a primary outcome but included among adverse events, healthy 2-week-old infants were randomized to one of two staged study formulas [19]. The stage 1 formula was given to infants 180 days of age followed by a stage 2 formula through 365 days. The control group was fed SF (n = 228, 148 completed), and the intervention group received identical formulas supplemented with bovine lactoferrin (FrieslandCampina DMV, The Netherlands) (0.6 g/L) and bovine MFGM (Lacprodan® MFGM-10; Arla Foods Ingredients) (5 g/L) (n = 223, 143 completed). The primary outcome was cognitive development (BSID III) at 12 months, and secondary outcomes were medically confirmed adverse events, including respiratory and gastrointestinal infections, through 18 months. The incidence rates of upper respiratory tract infections and diarrhea were significantly lower for the MFGM + lactoferrin group than the control group

136 Hernell/Lönnerdal/Timby through day 545 (p = 0.02 and p = 0.003, respectively). How much of the difference between the groups was due to lactoferrin rather than MFGM is an open question.

Clinical Studies on Milk Fat Globule Membrane Concentrates to Infants: Effects on the Microbiome and Metabolome

The studies in Sweden and Peru mentioned above are the only identified DBRCT reporting effects of MFGM supplementation to infants or children on omics. In the Swedish study (Table 1), the oral microbiota was analyzed at 4 (n = 124) and 12 (n = 166) months of age using Illumina MiSeq multiplex sequencing and taxonomic resolution against the HOMD 16S rDNA database of oral bacteria. Secondary analyses of the fecal microbiome and metabolome were done in a randomized subsample of 30 infants in each group through secondary analyses of feces collected from these infants at baseline (∼2 months), and 4, 6, and 12 months of age via 16S rRNA amplicon sequencing and quantitative metabolomics profiling. There were moderate effects on the oral microbiome [20]. Species richness in the oral samples did not differ between the MFGM and SF groups, but a few taxa that were significantly associated with being in either group were identified, e.g., a lower level of Moraxella catarrhalis in the MFGM group. M. catarrhalis was highlighted, as it is one of the major otitis pathogens and may be associated with the decrease in otitis media seen in the same group [17]. The impact on the fecal microbiome was also moderate. During the intervention, some infants who consumed the MFGM formula had a higher percentage of Akkermansia, which was more evident before introduction of complementary food. At 12 months of age, the number of individuals with Haemophilus was lower in the MFGM group than the SF group, while the relative abundance of fecal Proteobacteria was higher in the MFGM infants to a level similar to breastfed infants. These findings support the hypothesis that MFGM may play a role in shaping gut microbial activity and function [21]. However, the microbiome differences decreased after introduction of complementary food and were virtually undetectable at 12 months of age. The difference in the erythrocyte membrane and plasma lipidomes between the groups was more distinct, particularly with higher levels of sphingomyelins and phosphatidylcholines in the MFGM group [22], which shows that the higher phospholipid intake among these infants had an effect which likely primarily contributed to the positive cognitive performance [16], but may also have affected innate immunity and reduced the risk of viral infections [23].

Immunologic and Metabolomic Effects of MFGM 137 The MFGM group also had higher concentrations of lyso-phosphatidylcho- line (PC) [24], an important carrier of docosahexaenoic acid across the blood- brain barrier. When comparing the metabolome of SF and MFGM infants, the latter had higher levels not only of phospholipids and lyso-PC but also tended to have higher levels of free fatty acids and their oxidation products, choline, betaine, myoinositol, and ketones, and lower circulating levels of amino acids. This suggests that these infants were more metabolically similar to breastfed infants (preference for fatty acid oxidation, including ketogenesis) [24]. Lyso-PC, choline, serine, and myoinositol as well as ketones in circulation are passing through the blood-brain barrier and serve as important substrates to support optimal phospholipid synthesis for the rapidly growing brain of the newborn infant. Ketone bodies are energy substrates but also building blocks and possibly a signal substance. Collectively, these data may also contribute to explain the better cognitive achievements in infants fed formula supplemented with MFGM [2, 3, 16, 19]. Analyses of the fecal microbiota and serum metabolome in the Peruvian study (Table 1) showed that the MFGM infants had improved micronutrient status, energy metabolism, and growth (irrespective of also being given 1 RDA of micronutrients daily), as reflected by increased levels of circulating amino acids and weight gain. A cytokine shift towards less T-helper type 1 responses was also observed, which was mainly attributed to lower IL-2 levels possibly causing enhanced immunity [25].

Conclusions

Studies on bovine MFGM supplementation concentrate on a safe diet for infants and children, and have shown promising results regarding the defense against infections. These findings are supported by known effects of individual MFGM components mostly based on in vitro and/or animal studies. Results from DBRCTs also support that MFGM supplementation has moderate effects on the oral and fecal microbiome, on the metabolome, on plasma and erythrocyte membrane lipidomes, and on the serum metabolome, which could be involved in the mechanisms explaining the positive clinical effects. However, the scientific base of knowledge for any effects of MFGM supplementation to infants and children on the metabolome, microbiome, or infections is very limited. More DBRCTs on MFGM supplementation to infants and children, designed to study the effects on infections as well as the microbiome and the metabolome, are needed.

138 Hernell/Lönnerdal/Timby Disclosure Statement

O.H. has participated as a clinical investigator and/or scientific advisory board member, speaker, consultant for Semper, Hero, Mead Johnson Nutrition, Arla Foods, Arla Ingre- dients, Cargill, Nestlé Nutrition Institute, and Hipp. B.L. has participated as a clinical investigator and/or scientific advisory board member, consultant, speaker for Semper, Hero, Mead Johnson Nutrition, Arla Foods, Arla Ingredients, Albion, Humana, Bios- time, Second Science, Triton, and Nestlé Nutrition Institute. N.T. has participated as a clinical investigator and/or speaker for Hero and Semper.

References

1 Lee H, Padhi E, Hasegawa Y, et al: Compositional 13 Zavaleta N, Kvistgaard AS, Graverholt G, et al: dynamics of the milk fat globule and its role in Efficacy of an MFGM-enriched complementary

infant development. Front Pediatr 2018; 6: 313. food in diarrhea, anemia, and micronutrient sta-

2 Hernell O, Timby N, Domellof M, Lönnerdal B: tus in infants. J Pediatr Gastroenterol Nutr 2011;

Clinical benefits of milk fat globule membranes 53: 561–568.

for infants and children. J Pediatr 2016; 173 14 Veereman-Wauters G, Staelens S, Rombaut R, et (suppl):S60–S65. al: Milk fat globule membrane (INPULSE) en- 3 Timby N, Domellof M, Lönnerdal B, Hernell O: riched formula milk decreases febrile episodes Supplementation of infant formula with bovine and may improve behavioral regulation in young

milk fat globule membranes. Adv Nutr 2017; 8: children. Nutrition 2012; 28: 749–752. 351–355. 15 Poppitt SD, McGregor RA, Wiessing KR, et al: 4 Bourlieu C, Michalski MC: Structure-function Bovine complex milk lipid containing ganglio- relationship of the milk fat globule. Curr Opin sides for prevention of rotavirus infection and

Clin Nutr Metab Care 2015; 18: 118–127. diarrhoea in northern Indian infants. J Pediatr

5 Rueda R, Maldonado J, Narbona E, Gil A: Neona- Gastroenterol Nutr 2014; 59: 167–171.

tal dietary gangliosides. Early Hum Dev 1998; 16 Timby N, Domellof E, Hernell O, et al: Neurode- 53(suppl):S135–S147. velopment, nutrition, and growth until 12 mo of 6 Liao Y, Alvarado R, Phinney B, Lönnerdal B: Pro- age in infants fed a low-energy, low-protein for- teomic characterization of human milk fat glob- mula supplemented with bovine milk fat globule ule membrane proteins during a 12 month lacta- membranes: a randomized controlled trial. Am J

tion period. J Proteome Res 2011; 10: 3530–3541. Clin Nutr 2014; 99: 860–868. 7 Lu J, Wang X, Zhang W, et al: Comparative pro- 17 Timby N, Hernell O, Vaarala O, et al: Infections teomics of milk fat globule membrane in differ- in infants fed formula supplemented with bovine ent species reveals variations in lactation and milk fat globule membranes. J Pediatr Gastroen-

nutrition. Food Chem 2016; 196: 665–672. terol Nutr 2015; 60: 384–389. 8 Affolter M, Grass L, Vanrobaeys F, et al: Qualita- 18 Li X, Peng Y, Li Z, et al: Feeding infants formula tive and quantitative profiling of the bovine milk with probiotics or milk fat globule membrane: a fat globule membrane proteome. J Proteomics double-blind, randomized controlled trial. Front

2010; 73: 1079–1088. Pediatr 2019; 7: 347. 9 German JB: Dietary lipids from an evolutionary 19 Li F, Wu SS, Berseth CL, et al: Improved neuro- perspective: sources, structures and functions. developmental outcomes associated with bovine

Matern Child Nutr 2011; 7(suppl 2): 2–16. milk fat globule membrane and lactoferrin in 10 Rueda R: The role of dietary gangliosides on im- infant formula: a randomized, controlled trial. J

munity and the prevention of infection. Br J Nutr Pediatr 2019; 215: 24–31.e8.

2007; 98(suppl 1):S68–S73. 20 Timby N, Domellof M, Holgerson PL, et al Oral 11 Liu B, Newburg DS: Human milk glycoproteins microbiota in infants fed a formula supplement- protect infants against human pathogens. Breast- ed with bovine milk fat globule membranes – a

feed Med 2013; 8: 354–362. randomized controlled trial. PLoS One 2017; 12 Fuller KL, Kuhlenschmidt TB, Kuhlenschmidt 12:e0169831. MS, et al: Milk fat globule membrane isolated from buttermilk or whey cream and their lipid components inhibit infectivity of rotavirus in

vitro. J Dairy Sci 2013; 96: 3488–3497.

Immunologic and Metabolomic Effects of MFGM 139 21 He X, Parenti M, Grip T, et al: Fecal microbiome 23 Voelker DR, Numata M: Phospholipid regulation and metabolome of infants fed bovine MFGM of innate immunity and respiratory viral infec-

supplemented formula or standard formula with tions. J Biol Chem 2019; 294: 428289. breast-fed infants as reference: a randomized 24 He X, Parenti M, Grip T, et al: Metabolic pheno-

controlled trial. Sci Rep 2019; 9: 11589. type of breast-fed infants, and infants fed stan- 22 Grip T, Dyrlund TS, Ahonen L, et al: Serum, plas- dard formula or bovine MFGM supplemented ma and erythrocyte membrane lipidomes in in- formula: a randomized controlled trial. Sci Rep

fants fed formula supplemented with bovine milk 2019; 9: 339.

fat globule membranes. Pediatr Res 2018; 84: 726– 25 Lee H, Zavaleta N, Chen SY, et al: Effect of bovine 732. milk fat globule membranes as a complementary food on the serum metabolome and immune markers of 6–11-month-old Peruvian infants.

NPJ Sci Food 2018; 2: 6.

140 Hernell/Lönnerdal/Timby Protective Factors in Human Milk

Clinical Trials of Lactoferrin in the Newborn: Effects on Infection and the Gut Microbiome

Nicholas D. Embleton Janet E. Berrington Newcastle Neonatal Service, Royal Victoria Infirmary, Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK

Abstract Newborn infants, especially those born preterm, are at risk of infections in early life. In preterm infants, necrotizing enterocolitis (NEC), a devastating inflammatory gut condition, and late-onset sepsis (LOS) are important causes of serious morbidity and are the commonest reasons for death after the first week of life. Fresh breast milk from the infant’s mother reduces the risks of these serious pathologies in a dose-dependent fashion. Considerable effort has been expended to better understand which specific components of human milk are likely to exert the greatest functional benefits, particularly those that have immune modulatory or anti-infectious properties. Lactoferrin is a whey glycoprotein present in especially high concentrations in colostrum and early milk. Studies show that lactoferrin impacts on immune function and, through a multitude of mechanisms, reduces the risk of viral, fungal, and bacterial infections. Supplemental enteral bovine lactoferrin has been tested in a series of randomized clinical trials, many of which suggested important reductions in LOS in preterm or low-birth-weight infants. However, the largest trial to date – the Enteral Lactoferrin in Neo- nates (ELFIN) trial – recruited 2,203 infants and failed to show any significant reductions in LOS or NEC. Challenges in conducting clinical research and the translational relevance of these studies for clinical practice will be considered. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Introduction

Every year, more than 15 million infants are born low birth weight (LBW, < 2.5 kg birth weight) or premature globally. Whilst survival rates are improving, infections in these infants remain prevalent and are the major causes of mortality and serious morbidity in most settings [1]. In low- and middle-income countries, the burden of parasitic and other infections is especially high throughout infancy. In comparison, death due to infection outside the perinatal period is relatively uncommon in high-income countries even in LBW infants. However, in infants born very preterm (< 32 weeks of gestation) in high-income countries, sepsis and necrotizing enterocolitis (NEC) are now the commonest reasons for death after the first week of life [2]. In part, this is due to improvements in neonatal care that have resulted in a dramatic reduction in early deaths due to respiratory disease. The etiology of NEC and sepsis is multifactorial, and various strategies have been employed in an attempt to reduce their occurrence in neonatal intensive care units (NICUs). Overall, the most successful strategy is promotion of mother’s own breast milk, which shows a dose-response reduction in just about every neonatal condition studied, but especially in reducing NEC and late-onset sepsis (LOS), which occurs after 72 h of life [3]. However, studies have also explored the role of using supplemental lactoferrin to reduce infectious morbidity [4].

Lactoferrin: A Key Immunonutrient in Mammalian Milk

It is well known that mammalian milk is incredibly complex and consists of a multitude of components with “functional” activity. Many of these components, including lactoferrin, a whey protein found in all mammalian milks, have ancient origins with evidence that they evolved before mammals and lactation even existed [5]. This knowledge has encouraged a paradigm shift in nutritional thinking for preterm infants, moving from a focus predominantly on the macronutrient content of milk to the recognition that functional components are key to health promotion and reduction of disease. Optimizing nutritional status in preterm infants requires knowledge of macro- and micronutrient needs, combined with an appreciation of functional components (for example, lactoferrin or oligosaccharides). Improving nutritional status also requires a healthy microbiome. In part, the gut microbiota is “acquired” from microbes present in the mammary gland (and therefore in breast milk), but gut microbiota is also “fed” by components in human milk. In addition to lactoferrin, other milk proteins may also have interactions with gut microbes or immunomodulatory properties such as IgA, osteopontin, lyso-

142 Embleton/Berrington zyme, and milk fat globule membrane proteins, and nonprotein components such as human milk oligosaccharides are also clearly important [6]. Identifica- tion of components in breast milk that may act as “immunonutrients” and thereby prevent NEC and LOS is seen by many as the holy grail of much current nutrition research in preterm infants. The challenge, however, is that neither NEC nor LOS are single pathology diseases: in reality, both diseases represent the final common pathway of a range of etiological factors, including the degree of pre- maturity-related illness (e.g., the need for respiratory or cardiovascular support), immunologic immaturity, birth mode, neonatal practices affecting skin and gut epithelial integrity, and the role of the gut microbiome. Studying these aspects in sick preterm infants is both practically and ethically challenging and requires considerable time and other resources. In recent years, our knowledge of gut microbiota in preterm infants has increased dramatically: this has paralleled improvements in next-generation sequencing methodologies that now enable the identification of thousands of microbes relatively easily and cheaply. Numerous studies show that the pattern of gut microbial communities differs between preterm and healthy term infants. Furthermore, these patterns appear to change prior to the onset of NEC, and perhaps LOS, offering hope that earlier disease identification may be possible in the future [7–10]. Whilst changes prior to disease have been termed dysbiotic, major challenges exist. Firstly, many completely healthy preterm infants exhibit gut community patterns that are not dissimilar to those who develop disease (i.e., many otherwise normal preterm infants appear to be dysbiotic), and, sec- ondly, many infants show rapid shifts and changes in microbial patterns without obvious precipitants giving the impression of gut microbial chaos. Furthermore, it is unclear whether shifts in microbial patterns are the cause of disease as op- posed to being a result of some other pathological process. Identification of an enteral immunonutrient with a range of enterocyte growth activity and immunomodulation that might stabilize bacterial communities or reduce microbial chaos and dysbiosis might then lead to a reduction in NEC and prevent translocation of gut organisms causing sepsis. Lactoferrin appeared to hold that promise for very preterm infants.

Lactoferrin: Interactions with Gut Microbes and Immunomodulatory Activity

Human breast milk has been known to have bacteriostatic activity against certain bacteria such as Escherichia coli for almost half a century [6]. It was postulated that much of this bacteriostatic activity derived from the presence of lac-

Clinical Trials of Lactoferrin in the Newborn 143 toferrin, an iron-binding glycoprotein present in high amounts especially in early colostrum [11]. Sequestration and binding of iron prevents many pathogens using their siderophores to acquire the necessary iron needed for growth and multiplication and appears to be a highly effective evolutionary strategy as it is widely present in animals even in nonmammalian species [5]. Subsequent studies showed that lactoferrin also has bactericidal activity against many pathogens by forming strong complexes with bacterial lipopolysaccharide creating a hole in the outer membrane of gram-negative bacteria. More recent studies, however, show that the interaction between lactoferrin, iron, and bacteria is likely much more complex than initially thought [12]. Enterally ingested lactoferrin survives gastric passage intact and can be detected in the feces in healthy breastfed infants, although some is hydrolyzed in the presence of gastric acid to a peptide, lactoferricin, which has also been shown to have anti-infectious activity. Lactoferrin stimulates enterocytes resulting in increases in small intestinal mass, length, and digestive enzyme expression [13]. To achieve this, lactoferrin binds to a receptor on enterocytes. This lactoferrin-receptor complex is then internalized where both iron-free (apo-) and iron-saturated (holo-)forms activate mTor signaling pathways involved in the cell cycle, as well as entering the nucle- us and activating gene transcription. Studies suggest that many of these biological activities can be effected by bovine as well as human lactoferrin [13]. Lactoferrin may also act as a “prebiotic” and promote the growth of certain “healthy” bacteria. The growth of Lactobacillus acidophilus was stimulated by the presence of bovine hololactoferrin but not apolactoferrin, whereas both apo- and holoforms stimulated the growth of bifidobacterial species B. breve, B. longum subsp. infantis and B. bifidum. In the original studies, the growth of B. longum subsp. longum did not appear to be affected, which was initially thought to be due to a lack of binding proteins in the cell membrane [14]. However, further studies suggest that the presence of a binding protein may in fact be universal in all bifidobacterial species, but the growth response to lactoferrin may differ at a strain level [15]. Data also show that high-dose lactoferrin may suppress the growth activity of certain species such as B. breve, suggesting there may be an optimal concentration of lactoferrin [16]. Higher levels of fecal lactoferrin appear in preterm compared to term infants and are directly associated with the amounts of lactobacilli and bifidobacteria [17]. Extensive reviews describe a range of potential roles of lactoferrin as an immunomodulator involved in innate and acquired pathways [18], all of which may be important in preterm infants. Studies show improved immune responses in newborn piglets fed high doses of lactoferrin with immune cells producing higher quantities of cytokines than those fed formula with lower levels of lactoferrin, and a trend to a higher serum

144 Embleton/Berrington IgG and lower mortality in those fed the higher intakes [19]. The same research group also showed increased blood NK cell populations (lymphocytes that kill target cells and secrete cytokines) and NK lactoferrin receptor activity in those fed higher lactoferrin intakes, although there was no increase in NK cell cytotox- icity [20]. These studies demonstrate that lactoferrin may support innate immune pathways that are critical for the development of pathogen-specific T- lymphocyte mechanisms. Further clinical data in humans are provided in a small randomized controlled trial (RCT) in Turkey, which recruited very LBW

(birth weight < 1.5 kg) infants to receive either supplemental oral lactoferrin (200 mg/day) or placebo [21]. The study used flow cytometry to show increases in Treg lymphocytes in the group receiving additional lactoferrin, which was accompanied by a significantly lower incidence of LOS.

Clinical Studies of Lactoferrin in Preterm Infants

As previously mentioned, NEC and sepsis remain important pathologies in neonatal practice with high mortality rates. Lactoferrin holds promise as a safe and effective approach to reduce the occurrence of these serious diseases with a strong biological basis and rationale to study in well-designed trials. Relatively small reductions in disease would also be important for such a potentially cheap intervention, and there may be other benefits such as reduced use and exposure to antibiotics, and, therefore, a lower risk of antimicrobial resistance in the NICU. Furthermore, lactoferrin may be as, or more, effective in other high-risk groups and settings, e.g., in infants in low- and middle-income countries. In a multicenter RCT in Peru, 190 LBW infants were randomized to receive lactoferrin 200 mg/kg/day versus placebo for 4 weeks [22]. The cumulative incidence of sepsis was lower in the lactoferrin group than in the placebo group (12.6 vs. 22.1%, respectively) with a difference that was more marked when the analysis was restricted to very LBW infants (20 vs. 37.5%, respectively). Preterm infants typically only receive small amounts of mother’s own colostrum in the first 1–2 days, but whilst the lactoferrin concentration in colostrum is very high compared to mature milk, the actual volume of milk received is small, and these infants usually take several days to achieve full enteral feeds. Whilst lactoferrin concentrations may be higher in the milk of mothers delivering premature than at-term infants [23], preterm infants may be more than 2 weeks old before their actual enteral intake of lactoferrin is similar to that in a term infant, a time period when the risk of LOS is especially high. Low intake of lactoferrin may be exacerbated by frequent stopping of milk feeds due to clinical concerns, and the use of pasteurized donor breast milk that has relatively lower

Clinical Trials of Lactoferrin in the Newborn 145 levels of lactoferrin compared to mother’s own breast milk [16]. Newer methods of pasteurization that better preserve lactoferrin levels are being explored [24], but a key question for many NICUs is whether routine enteral supplementation of lactoferrin improves outcome in preterm infants. Manzoni et al. [25] conducted a landmark, high-quality, multicenter RCT exploring whether a fixed dose of 100 mg/day supplemental lactoferrin reduced the occurrence of LOS in preterm infants (n = 472) randomized to 1 of 3 arms: (1) bovine lactoferrin (bLF); (2) bLF plus a probiotic Lactobacillus rhamnosus GG (LGG); or (3) placebo. Study treatment was given for up to 6 weeks and aimed to explore both the effect of bLF and any potential synergy with the co administration of LGG. Overall rates of LOS were significantly lower in the bLF (5.9%) and the bLF + LGG (4.6%) groups than the placebo group (17.3%), and this was accompanied by a reduction in sepsis-associated mortality (0 vs. 4.8%). The reduction in sepsis occurred for gram-negative, gram-positive, and fungal (candida) infections. A reduction in NEC was observed, but the authors noted that the study was not powered for a NEC outcome. An add-on study was planned and conducted using the same trial methods which concluded a reduction in NEC (2 vs. 5.4%) of borderline significance (p = 0.04) [26]. More recently, Sherman et al. [27] conducted an RCT enrolling 120 preterm infants weighing 750–1,500 g and administering 150 mg/kg of Tal-lactoferrin (TLf) or placebo every 12 h. TLf is a recombinant human lactoferrin produced commercially. The incidence of LOS was 50% lower in the TLf group compared to the placebo group (17 vs. 33%), and the numbers of infants never developing a gram-negative infection was much higher in the TLf group. As with previous studies, no specific harm or adverse effect was attributed to receiving supplemental lactoferrin. In addition, the investigators conducted stool microbiota profiling, which showed a reduction in Enterobacter and Klebsiella taxa, but increases in Citrobacter [28]. Interestingly, staphylococcus taxa were barely detectable in infants receiving TLf, although it should be noted that only 23/120 infants recruited to the study took part in this microbiota substudy. A Cochrane meta-analysis of lactoferrin supplementation added to enteral feeds in preterm infants included 6 RCTs and 886 participants, and showed a reduction in LOS with a relative risk (RR) of 0.59 (95% CI 0.40–0.87) and a lower incidence of NEC reported in trials with 750 participants with a RR 0.40 (95% CI 0.18–0.86) [4]. The authors concluded that lactoferrin supplementation with or without probiotics decreased bacterial and fungal LOS, but not chronic lung disease or length of hospital stay (low-quality evidence). A review suggested that lactoferrin should be started early in preterm infants, with a dose of > 100 mg/ day, and it should be combined with probiotics in order to reduce the incidence of NEC [29].

146 Embleton/Berrington Enteral Lactoferrin Supplementation in Preterm Infants: The ELFIN Trial

The ELFIN trial was planned in 2012–2013, commenced recruitment in 2014, and enrolled 2,203 infants from 37 neonatal units along with 97 continuing care hospitals in the UK until planned recruitment stopped in 2017 [30]. It represents the largest ever trial of lactoferrin supplementation in infants, and perhaps one of the largest interventional nutritional neonatal RCTs ever conducted. Very preterm infants (< 32 weeks gestation) were recruited before 72 h of age and al- located to lactoferrin supplementation (150 mg/kg/day) or placebo (using sucrose). Web-based randomization used a minimization algorithm to balance groups for site, gender, single versus multiple birth, and gestational age. Dry bLF has a pale pink-brown tinge, whereas sucrose was pale brown: considerable effort was taken to ensure blinding and masking, and the trial product was stored in opaque pots with a stopper that prevented clinical teams from seeing the product. The product was mixed with sterile water and milk (breast milk or formula depending on what the baby was receiving) and withdrawn into a purple enteral feeding syringe that continued to provide masking. The primary outcome was LOS either microbiologically confirmed (essentially a positive blood culture) or clinically suspected (essentially 5 or more days of treatment with antibiotics along with the presence of clinical or laboratory features of infection). Blinded endpoint review committees masked to participant allocation reviewed all case report forms of LOS or NEC and required 2 experienced neonatal clinicians to independently (i.e., blindly) agree on outcome. Any discrepancies involved further discussion and/or the involvement of a third independent clinician and, where necessary, further discussion with individual site investigators. No previous studies of lactoferrin appear to have applied this level of rigor to endpoint confirmation. The study was powered to detect a risk reduction of about 25% assuming an LOS event rate of approximately 20%. Primary outcome data were available for 2,182 infants (99%) and showed no difference in the rate of LOS between lactoferrin (28.9%) and placebo (30.7%), with an adjusted RR of 0.95 (95% CI 0.86–1.04). There were no significant differences in secondary outcomes or any preplanned subgroup analyses based on gestation. The incidence rates of microbiologically confirmed LOS (17 vs. 17%), NEC stage II/III (6 vs. 5%), and all-cause mortality (7 vs. 6%) did not differ. The ELFIN trial provides high-quality data that show that routinely providing supplemental bLF to preterm infants does not improve outcome, contradict- ing findings suggested by previous studies. Trial groups were well balanced for baseline characteristics, adherence to the protocol was high, event rates for outcomes were similar to predicted, and the trial enrolled more than twice as many

Clinical Trials of Lactoferrin in the Newborn 147 infants as all the previous studies combined. Although it is plausible that lactoferrin has the greatest benefits in the most immature infants, there was no sug- gestion of this in the subgroup analyses; similarly, there was no evidence of an interaction with the type of milk fed, i.e., human milk, formula, or both. There are, however, potentially several differences in the study cohorts, methodology, and NICU practices between the Italian trial of Manzoni et al. [25] and the ELFIN trial [30]. Invasive fungal infections occurred in 5.4% of the control group in the Italian trial, whereas only 5 babies in total (< 0.5%) in the ELFIN trial developed fungal sepsis. This may reflect differences in the use of antifungal prophylaxis or other infection control practices. Other factors are likely to be important when exploring the effects of immunonutrients on diseases such as NEC and LOS. In a large pragmatic trial, some babies recruited in the first 72 h may actually receive little or no lactoferrin if they are placed nil by mouth because of clinical concerns. In these cases, it is clearly not possible for lactoferrin to exert a beneficial effect. Although the timing of disease onset is not currently reported, it is likely that several cases of NEC or LOS occurred in the first 1–2 weeks when the cumulative amount of lactoferrin received will be quite low; indeed, in a recent RCT exploring the impact of donor human milk, almost 60–70% of all NEC or LOS cases occurred in the first 10 days of life [3]. It seems unlikely that administration of an enteral supplement such as lactoferrin for just a few days will have a major impact over just a few days. Several LOS cases are due to infections with coagulase-negative staphylococci (CoNS); whilst the origin of the CoNS may be in the gut, it is also likely that several cases of CoNS are due to skin colonization and infection of indwelling intravascular devices and cannulae. In this situation, lactoferrin would have to rapidly exert activity beyond the gut to reduce the incidence of LOS. Basic scientific studies, some of which were briefly reviewed earlier, suggest that lactoferrin works synergistically with other immunoreactive proteins such as osteopontin and lysozyme. Supplemental lactoferrin may require these or other “cofactors” to effectively reduce LOS – functional nutrients tend not to work in isolation. In addition, basic studies emphasize the impact of the degree of iron saturation of lactoferrin: the product used in the ELFIN trial underwent rigorous quality control to determine the degree of iron binding, but it is of course possible that the functional activity of commercially available bLF differs between suppliers, and other factors such as storage and transportation also have an impact on functional activity. Commercially available lactoferrin must also be pasteurized; differences in how this is done may impact on functional activity. Earlier studies and reviews suggest the possibility of a synergistic effect from probiotics. In the ELFIN trial, around one-third of infants were

148 Embleton/Berrington recorded as having received probiotics, but again there was no evidence of differential effects. Nevertheless, other study designs could explore that possibility more effectively because there are suggestions that certain probiotics may be synergistic with lactoferrin (i.e., lactoferrin is perhaps acting as prebiotic) whilst other data suggest the bLF may inhibit growth of certain probiotics at higher doses [16]. A further study (MAGPIE), nested within the ELFIN trial, recruited over 480 infants in whom daily stool and urine samples were collected for a range of microbiomic and metabolomic studies [31]. These data are not yet available but will provide additional evidence on basic mechanisms of lactoferrin, as well as differences occurring in the immediate time period before NEC or LOS onset.

Conclusion

NEC and LOS are devastating diseases in preterm infants. Considering the results of previous studies suggesting that lactoferrin might be beneficial, the results of the ELFIN trial are disappointing. Such is the nature of science. The ELFIN trial highlights the importance of adequately powered high-quality trials in providing the most robust evidence base for clinical practice. This requires collaboration at national and international level – it is simply not possible to adequately power a study to determine a realistic impact on a multifactorial disease such as LOS or NEC without recruiting at least 1,000 preterm infants. RCTs smaller than this should be interpreted with caution. Whilst large-scale pragmatic trials are essential, mechanistic studies (for example of gut microbiota, and immune and metabolic function) are also needed in order to better understand disease pathology and treatment effect, and to determine the optimal intervention to be tested in large RCTs. Necessarily, RCTs are practically and ethically challenging in preterm infants, and studies must be planned, developed, and conducted with parental involvement at every stage. The ELFIN trial does not negate the fact that lactoferrin is a key immunoprotein with a range of important health effects, it simply demonstrates that supplemental bLF in preterm infants does not significantly reduce the risk of LOS or NEC. It reminds us that nutrients do not function in isolation, and that the optimal source of immunonutrients remains fresh human milk. Whilst we continue to search for immunonutrients that improve health and reduce disease when provided as supplements, studies also need to be directed at elucidating how best to support mothers to meet their infants’ intake needs using their own milk.

Clinical Trials of Lactoferrin in the Newborn 149 Acknowledgments

The ELFIN and MAGPIE studies are the result of many years of collaboration with many investigators, especially Prof. W. McGuire and Prof. E. Juszczack, the ELFIN and MAGPIE study teams, and the clinical trial unit at the National Perinatal Epidemiology Unit, Oxford, UK. The studies would also not have been possible without local investigator teams of nurses and doctors. Most importantly, we wish to thank and acknowledge the parents who trusted us and could see the important benefits that clinical research will bring to future generations – our work is dedicated to them and all their babies.

Disclosure Statement

Dr Embleton and Dr Berrington report funding from the National Institutes for Health Research (NIHR), Prolacta Bioscience US, Danone Early Life Nutrition, Action Medical Research, and Tiny Lives. Dr Embleton declares speaker’s honoraria from the Nestlé Nu- trition Institute and Danone Early Life Nutrition. Neither author holds relevant stocks, shares options, or has any other relevant personal or family financial disclosures.

References

1 March of Dimes, PMNCH, Save the Children, 9 Stewart CJ, Embleton ND, Marrs ECL, et al: Lon- WHO: Born Too Soon: The Global Action Report gitudinal development of the gut microbiome on Preterm Birth. Geneva, WHO, 2012, vol 13, pp and metabolome in preterm neonates with late 1–126. onset sepsis and healthy controls. Microbiome

2 Berrington JE, Hearn RI, Bythell M, et al: Deaths 2017; 5: 75. in preterm infants: changing pathology over 2 10 Stewart CJ, Embleton ND, Marrs ECL, et al: Tem-

decades. J Pediatr 2012; 160: 49–53.e1. poral bacterial and metabolic development of the 3 Corpeleijn WE, de Waard M, Christmann V, et al: preterm gut reveals specific signatures in health

Effect of donor milk on severe infections and and disease. Microbiome 2016; 4: 67. mortality in very low-birth-weight infants: the 11 Embleton ND, Berrington JE, McGuire W, et al: Early Nutrition Study randomized clinical trial. Lactoferrin: antimicrobial activity and therapeu-

JAMA Pediatr 2016; 170: 654–661. tic potential. Semin Fetal Neonatal Med 2013; 18: 4 Pammi M, Suresh G: Enteral lactoferrin supple- 143–149. mentation for prevention of sepsis and necrotiz- 12 Jiang R, Lonnerdal B: Transcriptomic profiling of ing enterocolitis in preterm infants (review). intestinal epithelial cells in response to human,

Cochrane Database Syst Rev 2017; 6:CD007137. bovine and commercial bovine lactoferrins. Bio-

5 Oftedal OT: The evolution of milk secretion and metals 2014; 27: 831–841.

its ancient origins. Animal 2012; 6: 355–368. 13 Liao YL, Jiang RL, Lönnerdal B: Biochemical and 6 Lönnerdal B: Bioactive proteins in human milk: molecular impacts of lactoferrin on small intesti- health, nutrition, and implications for infant for- nal growth and development during early life.

mulas. J Pediatr 2016; 173:S4–S9. Biochem Cell Biol 2012; 90: 476–484. 7 Stewart CJ, Nelson A, Scribbins D, et al: Bacterial 14 Kim WS, Ohashi M, Tanaka T, et al: Growth-pro- and fungal viability in the preterm gut: NEC and moting effects of lactoferrin on L. acidophilus and

sepsis. Arch Dis Child Fetal Neonatal Ed 2013; 98: Bifidobacterium spp. BioMetals 2004; 17: 279–283. 298–303. 15 Rahman MM, Kim WS, Ito T, et al: Growth pro- 8 Berrington JE, Stewart CJ, Embleton ND, et al: motion and cell binding ability of bovine lactofer-

Gut microbiota in preterm infants: assessment rin to Bifidobacterium longum. Anaerobe 2009;

and relevance to health and disease. Arch Dis 15: 133–137.

Child Fetal Neonatal Ed 2013; 98:F286–F290. 16 Woodman T, Strunk T, Patole S, et al: Effects of lactoferrin on neonatal pathogens and Bifidobac- terium breve in human breast milk. PLoS One

2018; 13:e0201819.

150 Embleton/Berrington 17 Mastromarino P, Capobianco D, Campagna G, et 25 Manzoni P, Rinaldi M, Cattani S, et al: Bovine al: Correlation between lactoferrin and beneficial lactoferrin supplementation for prevention of microbiota in breast milk and infant’s feces. Bio- late-onset sepsis in very low-birth-weight neo-

metals 2014; 27: 1077–1086. nates: a randomized trial. JAMA 2009; 302: 1421– 18 Legrand D: Overview of lactoferrin as a natural 1428.

immune modulator. J Pediatr 2016; 173:S10–S15. 26 Manzoni P, Meyer M, Stolfi I, et al: Bovine lacto- 19 Comstock SS, Reznikov EA, Contractor N, Dono- ferrin supplementation for prevention of necro- van SM: Dietary bovine lactoferrin alters mucosal tizing enterocolitis in very-low-birth-weight neo- and systemic immune cell responses in neonatal nates: a randomized clinical trial. Early Hum Dev

piglets. J Nutr 2014; 144: 525–532. 2014; 90:S60–S65. 20 Liu KY, Comstock SS, Shunk JM, et al: Natural 27 Sherman MP, Adamkin DH, Niklas V, et al: Ran- killer cell populations and cytotoxic activity in domized controlled trial of talactoferrin oral solu-

pigs fed mother’s milk, formula, or formula sup- tion in preterm infants. J Pediatr 2016; 175: 68– plemented with bovine lactoferrin. Pediatr Res 73.e3.

2013; 74: 402–407. 28 Sherman MP, Sherman J, Arcinue R, Niklas V: 21 Akin IM, Atasay B, Dogu F, et al: Oral lactoferrin Randomized control trial of human recombinant to prevent nosocomial sepsis and necrotizing lactoferrin: a substudy reveals effects on the fecal enterocolitis of premature neonates and effect on microbiome of very low birth weight infants. J

T-regulatory cells. Am J Perinatol 2014; 31: 1111– Pediatr 2016; 173:S37–S42. 1120. 29 Manzoni P: Clinical benefits of lactoferrin for

22 Ochoa TJ, Zegarra J, Cam L, et al: Randomized infants and children. J Pediatr 2016; 173:S43–S52. controlled trial of lactoferrin for prevention of 30 ELFIN Trial Investigators Group: Enteral lacto- sepsis in Peruvian neonates less than 2,500 g. ferrin supplementation for very preterm infants:

Pediatr Infect Dis J 2015; 34: 571–576. a randomised placebo-controlled trial. Lancet

23 Turin CG, Rueda MS, Mercado E, et al: Lactofer- 2019; 393: 423–433. rin concentration in breast milk of mothers of 31 Embleton ND, Berrington JE, Dorling J, et al:

low-birth-weight newborns. 2017; 37: 507–512. Mechanisms affecting the gut of preterm infants

24 Moro GE, Billeaud C, Rachel B, et al: Processing in enteral feeding trials. Front Nutr 2017; 4: 14. of donor human milk: update and recommendations from the European Milk Bank Association

(EMBA). Front Pediatr 2019; 7: 1–10.

Clinical Trials of Lactoferrin in the Newborn 151 Protective Factors in Human Milk

Effects of Milk Osteopontin on Intestine, Neurodevelopment, and Immunity

Rulan Jiang Bo Lönnerdal Department of Nutrition, University of California, Davis, CA, USA

Abstract Osteopontin (OPN) is an acidic phosphorylated glycoprotein involved in a wide range of biological activities, such as cell proliferation and differentiation, as well as immunomodulatory functions. OPN contains integrin and CD44 binding sites, and it exerts its multiple functions by binding to its receptors on the cell membrane to trigger various cellular signaling pathways. It is generated by a variety of cell types, including epithelial cells and immune cells. OPN appears in most body fluids, such as milk and blood, and is present at a high concentration in human milk but not in bovine milk. Milk OPN is relatively resistant to digestion, and orally ingested OPN can enter the circulatory system. Milk OPN may, therefore, play essential roles in the development in early life. The impact of milk OPN on development has been investigated using cell models, animal models, and randomized clinical trials. Recent OPN studies strongly suggest that milk OPN plays important roles in intestinal proliferation and maturation, brain myelination, and neurodevelopment, as well as immune development. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel

Introduction

Osteopontin (OPN), consisting of 300 amino acids, is a phosphorylated glycoprotein with high levels of acidic amino acids. It is a multifunctional protein involved in a wide range of bioactivities, including cell proliferation and differentiation, immunomodulatory functions, biomineralization, as well as myelination [1, 2]. OPN contains Arg-Gly-Asp (RGD) and non-RGD integrin binding sites in the N-terminal and the central regions, as well as CD44 binding domains in the C-terminal region. The integrin RGD binding site is highly conserved among different species [3]. OPN exerts its multiple functions by binding to its receptors on cell membranes, activating various cell signaling pathways. OPN is expressed in a variety of cells, including osteoblasts, osteocytes, epithelial cells, endothelial cells, fibroblasts, and immune cells. There are two forms of OPN: secreted OPN present in body fluids and intracellular OPN found in immune cells [4]. Secreted OPN appears in most body fluids, such as milk, urine, blood, saliva, and bile. OPN is present at a high concentration in human milk, ∼178 mg/L in colostrum and ∼134 mg/L in transitional milk [5], whereas OPN concentrations in bovine milk and infant formula are markedly lower than in human milk, around 18 and 9 mg/L, respectively [6]. Consistent with these findings, microarray and immunoblotting analyses of cells in human milk have shown high expression of OPN during the whole lactation period [7]. Milk OPN is highly posttranslationally modified, with 36 phosphorylation sites and 5 O- glycosylation sites in human milk OPN [8] and 28 phosphorylated sites and 3 O-glycosylated sites in bovine milk OPN [9]. Milk OPN is relatively resistant to digestion by neonatal gastric juice [10] and intestinal digestive enzymes [11] in vitro. In addition, the presence of orally ingested milk OPN in the small intestine and plasma [12, 13] suggests that milk OPN may not only play a role in the intestine but also exert systemic functions. Methods for OPN isolation from milk have been established [14], and bovine milk OPN is commercially available [15], facilitating research on bioactivities and applications of bovine milk OPN. Functions of milk OPN have been investigated using cell models, animal models, and clinical trials. Based on the results from these OPN studies, it is evident that milk OPN plays important roles in intestinal, brain, and immune system development in infancy.

Milk Osteopontin Contributes to Intestinal Proliferation and Maturation

Since milk OPN is relatively resistant to digestion, intact or partly digested OPN may bind to receptors on the cell membrane of intestinal epithelial cells to exert pleiotropic functions in the intestine in infancy. Both human and bovine milk OPNs have been found to significantly stimulate proliferation of human crypt- like intestinal epithelial cells. To understand the underlying mechanisms by which milk OPN enhances intestinal cell proliferation, microarray assays (Illu- mina) were performed. Genes tightly related to cell proliferation and immune

Effects of Milk Osteopontin 153 function, such as MAPK13, CCNE1, CdGAP, CXCL10, IL6ST, and NF-κB, were significantly modified by milk OPN [16]. In addition, milk OPN has been found to enhance differentiation of intestinal epithelial cells (Caco-2) and to stimulate intestinal immunity by up-regulation of IL-18 secretion by intestinal epithelial (Caco-2) cells [11]. The jejunal transcriptomes of infant rhesus monkeys fed either regular formula or bovine milk OPN-supplemented formula for 3 months were compared; the transcriptome of the OPN-supplemented group was more similar to that of breastfed infant monkeys. OPN was shown to stimulate expression of genes involved in cell proliferation, migration, survival, and signaling pathways from integrin and CD44 receptors [17]. Survival of OPN in the gastrointestinal tract of breastfed infants may be facilitated by the formation of a stable complex between OPN and another major milk protein, lactoferrin (LF). Compared to OPN or LF alone, the LF-OPN complex has been shown to be more resistant to in vitro digestion and more effective in binding to and uptake by human crypt-like intestinal epithelial cells and in enhancing promotion of proliferation and differentiation of intestinal cells. Thus, by forming a complex in human milk, OPN and LF protect each other against proteolysis and enhance their individual bioactivities [11].

Milk Osteopontin Promotes Neurodevelopment in Early Infancy

Breastfeeding has been shown to promote cognitive development, and this is likely due to bioactive ingredients in human milk [18]. OPN is expressed abun- dantly in the brain in early life [19, 20]. OPN has also been shown to increase synthesis of myelin basic protein and formation of a myelin sheath in an in vitro primary culture model of myelination: mixed cortical cultures from embryonic mouse brain [2]. Myelination is one of the most critical cell-cell interactions for normal brain development, involving extensive information exchange between differentiating oligodendrocytes and axons [21]. Effects of milk OPN on neurodevelopment have been investigated in an OPN mouse model in which wild-type (WT) mouse pups nursed by WT or OPN-knockout dams received milk with abundant OPN or no OPN. The appearance of ingested milk OPN in the brain was shown by detection of orally gavaged iodine-125-labeled OPN and antibody- probed milk OPN in the brain of postnatal day 8 (P8) pups. The protein level of OPN was significantly higher in the WT pups nursed by WT dams than in those nursed by OPN-KO dams at P6 and P8. Cognitive tests performed at P30 showed that the pups nursed by WT dams had better memory and learning ability. Fur- ther experiments showed that the increased expression of OPN in the brain leads to increased expression of myelination-related proteins and elevated prolifera-

154 Jiang/Lönnerdal tion and differentiation of NG-2 glia into oligodendrocytes in the brain. These results indicate that milk OPN can play an important role in brain development and behavior during infancy by promoting brain myelination [22]. The high OPN concentration in human milk may thus be one of the factors resulting in better cognitive development in breastfed infants than in formula-fed infants.

Milk Osteopontin Stimulates Immune Development and Thus Provides Protection against Pathogen Insults

Animal studies and a randomized clinical trial have been carried out to investigate the impact of milk OPN on immune development in early life. Mouse pups nursed by WT or OPN-KO dams were utilized to investigate the effects of milk OPN on immune responses in early life [23]. To examine effects of milk OPN on immune responses, mouse pups (P20, P30, and P40) were treated with an intraperitoneal injection of Escherichia coli lipopolysaccharide (LPS), serotype O111:B4 (1.25 mg/kg; Sigma) in sterile saline. Mouse pups were then eutha- nized, and plasma samples were collected for cytokine measurements (ELISA; R&D Systems) at different time points (0.5, 1, 2, 4, 8, and 24 h). Diarrhea symptoms were seen starting at 4 h, and body weight was reduced by about 12.5% in all groups at 24 h. Plasma OPN and TNF-α were significantly up-regulated in both control and OPN-deficient groups after LPS injection, but a considerably higher increase in plasma OPN and TNF-α was observed in the pups nursed by OPN-KO dams. These observations indicate that milk OPN reduced inflammation resulting from LPS administration. Similarly, plasma IFN-γ was dramatically enhanced in response to the LPS challenge in pups nursed by both WT and OPN-KO dams, but the fact that the pups nursed by WT dams showed a significantly larger increase in IFN-γ suggests that milk OPN may contribute to resistance to LPS administration (bacterial infection) by altering immune responses [23]. Oral administration of bovine milk OPN (2 µg/mL) in drinking water to dextran sulfate sodium-treated WT adult mice reduced weight loss, spleen enlargement, and gut neutrophil activity, and it increased colon length and red blood cell counts [12]. Similarly, a lower inflammatory score and less neutrophil infiltration were observed in OPN-fed dextran sulfate sodium-treated mice [24], and a lower incidence of diarrhea was seen in LPS-injected piglets fed OPN-fortified (30 g/L) formula [25]. Moreover, OPN may improve immunity by altering the gut microbiota. Compared to ETEC (enterotoxigenic E. coli)-infected piglets fed regular algae, ETEC-infected piglets fed OPN-enriched algae showed decreased α-diversity, altered microbiota composition, and short- chain fatty acid profiles [26].

Effects of Milk Osteopontin 155 A randomized clinical trial was conducted on infants (1–6 months) to determine the effects of milk OPN during infancy. Infants were breastfed, fed regular formula, or OPN-supplemented formula containing 65 or 130 mg/L OPN (50 or 100% of the average OPN concentration in human milk, respectively). Infants fed formula supplemented with OPN had significantly lower serum TNF-α levels and fewer days of illness compared with infants fed regular formula [27]. Ad- ditionally, compared to infants fed regular formula, infants fed OPN-supplemented formula had an immune cell profile more similar to that of breastfed infants [28]. At 4 and 6 months, higher levels of OPN were found in plasma samples from the breastfed infants and infants fed OPN-supplemented formula than in samples from infants fed regular formula. These findings suggest that supplemental bovine milk OPN in infant formula may exert its beneficial effects by increasing endogenous OPN synthesis [5].

Conclusion

OPN appears at a high concentration in human milk. Since milk OPN is relatively resistant to digestion, and ingested milk OPN can be absorbed into the circulatory system, it is able to play essential roles in infancy both in the intestine and systemically. According to results from in vitro and in vivo studies, milk OPN contributes to intestinal proliferation and maturation, brain myelination and neurodevelopment, and immune development in early life.

Disclosure Statement

The authors declare no conflicts of interest.

References

1 Rittling SR, Singh R: Osteopontin in immune- 5 Jiang R, Lönnerdal B: Osteopontin in human milk

mediated diseases. J Dent Res 2015; 94: 1638– and infant formula affects infant plasma osteo-

1645. pontin concentrations. Pediatr Res 2019; 85: 502– 2 Selvaraju R, Bernasconi L, Losberger C, et al: Os- 505. teopontin is upregulated during in vivo demy- 6 Schack L, Lange A, Kelsen J, et al: Considerable elination and remyelination and enhances myelin variation in the concentration of osteopontin in

formation in vitro. Mol Cell Neurosci 2004; 25: human milk, bovine milk, and infant formulas. J

707–721. Dairy Sci 2009; 92: 5378–5385. 3 Wang KX, Denhardt DT: Osteopontin: role in 7 Nagatomo T, Ohga S, Takada H, et al: Microarray immune regulation and stress responses. Cyto- analysis of human milk cells: persistent high ex-

kine Growth Factor Rev 2008; 19: 333–345. pression of osteopontin during the lactation pe-

4 Rittling SR: Osteopontin in macrophage function. riod. Clin Exp Immunol 2004; 138: 47–53.

Expert Rev Mol Med 2011; 13:e15.

156 Jiang/Lönnerdal 8 Christensen B, Nielsen MS, Haselmann KF, et al: 18 Lönnerdal B: Human milk: bioactive proteins/ Post-translationally modified residues of native peptides and functional properties. Nestlé Nutr

human osteopontin are located in clusters: iden- Inst Workshop Ser 2016; 86: 97–107. tification of 36 phosphorylation and five O-glyco- 19 Choi JS, Cha JH, Park HJ, et al: Transient expres- sylation sites and their biological implications. sion of osteopontin mRNA and protein in amoe-

Biochem J 2005; 390(Pt 1): 285–292. boid microglia in developing rat brain. Exp Brain

9 Sørensen ES, Højrup P, Petersen TE: Posttransla- Res 2004; 154: 275–280. tional modifications of bovine osteopontin: iden- 20 Lee MY, Choi JS, Lim SW, et al: Expression of tification of twenty-eight phosphorylation and osteopontin mRNA in developing rat brainstem

three O-glycosylation sites. Protein Sci 1995; 4: and cerebellum. Cell Tissue Res 2001; 306: 179– 2040–2049. 185. 10 Chatterton DEW, Rasmussen JT, Heegaard CW, 21 Bercury KK, Macklin WB: Dynamics and mecha-

et al: In vitro digestion of novel milk protein in- nisms of CNS myelination. Dev Cell 2015; 32: gredients for use in infant formulas: research on 447–458. biological functions. Trends Food Sci Technol 22 Jiang R, Prell C, Lönnerdal B: Milk osteopontin

2004; 15: 373–383. promotes brain development by up-regulating 11 Liu L, Jiang R, Lönnerdal B: Assessment of bioac- osteopontin in the brain in early life. FASEB J

tivities of the human milk lactoferrin-osteopon- 2019; 33: 1681–1694.

tin complex in vitro. J Nutr Biochem 2019; 69: 23 Jiang R, Prell C, Lönnerdal B: Milk osteopontin 10–18. regulates immune responses in mice stimulated

12 da Silva AP, Ellen RP, Sørensen ES, et al: Osteo- by lipopolysaccharide (LPS). FASEB J 2016; 30: pontin attenuation of dextran sulfate sodium- 673.4.

induced colitis in mice. Lab Invest 2009; 89: 1169– 24 Kanwar JR, Kanwar RK, Stathopoulos S, et al: 1181. Comparative activities of milk components in

13 Ge X, Lu Y, Leung TM, et al: Milk osteopontin, a reversing chronic colitis. J Dairy Sci 2016; 99: nutritional approach to prevent alcohol-induced 2488–2501. liver injury. Am J Physiol Gastrointest Liver 25 Ren S, Hui Y, Goericke-Pesch S, et al: Gut and

Physiol 2013; 304:G929–G939. immune effects of bioactive milk factors in pre- 14 Sørensen ES, Petersen TE: Purification and char- term pigs exposed to prenatal inflammation. Am

acterization of three proteins isolated from the J Physiol Gastrointest Liver Physiol 2019; proteose peptone fraction of bovine milk. J Dairy 317:G67–G77.

Res 1993; 60: 189–197. 26 Wang M, Smith B, Adams B, et al: Osteopontin- 15 Kvistgaard AS, Matulka RA, Dolan LC, Ramanu- enriched algae modulates the gut microbiota jam KS: Pre-clinical in vitro and in vivo safety composition in weaning piglets infected with evaluation of bovine whey derived osteopontin, enterotoxigenic Escherichia coli (P06-069-19).

Lacprodan® OPN-10. Food Chem Toxicol 2014; Curr Dev Nutr 2019;3(suppl 1).

73: 59–70. 27 Lönnerdal B, Kvistgaard AS, Peerson JM, et al: 16 Jiang R, Lönnerdal B: Human and bovine osteo- Growth, nutrition, and cytokine response of pontin from milk and recombinant human osteo- breast-fed infants and infants fed formula with pontin may stimulate intestinal proliferation and added bovine osteopontin. J Pediatr Gastro

immune functions via various mechanisms re- enterol Nutr 2016; 62: 650–657.

vealed by microarray analysis. FASEB J 2013; 27: 41. 28 West CE, Kvistgaard AS, Peerson JM, et al: Effects 17 Donovan SM, Monaco MH, Drnevich J, et al: Bo- of osteopontin-enriched formula on lymphocyte vine osteopontin modifies the intestinal tran- subsets in the first 6 months of life: a randomized

scriptome of formula-fed infant rhesus monkeys controlled trial. Pediatr Res 2017; 82: 63–71. to be more similar to those that were breastfed. J

Nutr 2014; 144: 1910–1919.

Effects of Milk Osteopontin 157 Protective Factors in Human Milk

Published online: March 30, 2020 Ogra PL, Walker WA, Lönnerdal B (eds): Milk, Mucosal Immunity, and the Microbiome: Impact on the Neonate. Nestlé Nutr Inst Workshop Ser. Basel, Karger, 2020, vol 94, pp 158–168 (DOI:10.1159/000505335)

Effects of Milk Secretory Immunoglobulin A on the Commensal Microbiota

a, b a b Vanessa P. Dunne-Castagna David A. Mills Bo Lönnerdal a Department of Food Science and Technology, University of California, Davis, CA, USA; b Department of Nutrition, University of California, Davis, CA, USA

Abstract Secretory immunoglobulin A (SIgA) is intimately involved in the transfer of maternal immunity to the newborn breastfed infant. Recent research demonstrates the significance of SIgA in the initial development of the newborn’s microbiota and in the establishment of a tolerogenic immunologic disposition towards nonpathogenic organisms and environmental antigens. SIgA has long been known to prevent pathogen binding to the host epithelium through immune exclusion involving numerous mechanisms. This process primarily involves T-cell-dependent, somatically hypermutated monoclonal antibodies with high specificity towards pathogen surface antigens, and the success of the immune response is dependent upon the specific antigen recognition. Whereas this role is important, there is an alternate, dual role for SIgA in the health of the host – protection and promotion of commensal colonization and maintenance of homeostatic immunity. This latter role is primarily dependent upon N- and O-glycan moieties lining the secretory component and heavy chain of the SIgA dimer, with interactions independent of immunoglobulin specificity. These SIgA molecules are nonspecific polyclonal antibodies generated from plasma cells activated by dendritic cell sampling of luminal contents in the absence of inflammation. Breast milk is the primary supply of such polyclonal polyreactive SIgA in the initial stages of neonatal colonization, and it provides vital pathogen resistance while promoting colonization of commensal microbiota. © 2020 Nestlé Nutrition Institute, Switzerland/S. Karger AG, Basel Secretory IgA Production

In the mucosa, the majority of the immunoglobulin repertoire consists of somatically hypermutated IgA proteins, and over 25% are polyreactive to microbial-associated molecular patterns and other common antigens [1]. In addition, mucosal IgA production maintains a very high constant level in the absence of infection, up to 5 g/kg/day in an adult intestine [2], but it can be increased in an inflammatory environment. Recent studies using photoconversion to track Pey- er’s patch B cells showed marked small intestinal Peyer’s patch germinal center clonal exchange of memory B cells during a 3-day period [3], indicating considerable trafficking of memory B cells throughout mucosal surfaces in the gastrointestinal tract. In the mucosa, there seems to be a high level of low-affinity, di- versified germinal-center B-cell responses to microbial antigens that are required to maintain intestinal homeostasis. In addition to these T-cell-dependent responses which account for about 75% of IgA induction in the mucosa, B cells can be activated via T-independent processes. These may involve pattern recognition receptor signaling, as most IgA generated in mice deficient in T cells or with deficiencies in somatic hypermutation but without class switching produce primarily microbial-pattern (low- affinity) antibodies [4]. Interestingly, in these studies, the microbiota expanded into a dysbiotic state, with higher levels of Proteobacteria, suggesting the need for T-cell-dependent IgA responses to maintain intestinal homeostasis. Mature B cells are released into the circulation as plasmablasts and home back to the site of mucosal induction to secrete dimeric IgA. SIgA is derived upon transcytosis of dimeric IgA through mucosal epithelial cells bound to the polymeric immunoglobulin receptor (pIgR) on the basolateral membrane, and the subsequent apical cleavage of the immunoglobulin-bound secretory component (SC), which releases free SIgA into the lumen.

Immune Exclusion

Once in the lumen, there are a number of mechanisms by which SIgA functions and more are being discovered. Traditionally, it was thought that SIgA acts through agglutination and neutralization, excluding pathogens from binding to epithelial cells via Fab-dependent, high-affinity interactions to pathogen surface antigens. Although neutralization is supported by imaging of viral pathogens [5], agglutination has come into question recently as the concentration of an infectious agent is likely too low to provide enough localized pathogen to create immune complexes with SIgA. A recent study proposed a concept of “enchained

Milk SIgA and Commensals 159 growth” whereby bacterial pathogens coated with SIgA exhibit incomplete bi- nary fission and remain attached after replication [6]. In fact, this creates similar immune complexes as agglutination, preventing bacterial adhesion to epithelial cells and reducing pathology. In addition to these mechanisms, a novel function for Fab-dependent SIgA has been identified in work with the recombinant monoclonal SIgA, Sal4, directed at the acetylated O5 antigen of Salmonella enterica serovar Typhimurium (ST) [7]. In these studies, researchers found that Sal4 binding altered bacterial membrane integrity, compromising the cell energy gradient. As proton motif force is needed to power the bacterial flagella, motility was inhibited, as was ATP production and overall metabolic processes. While it is still uncertain how exactly the binding of SIgA to LPS alters the membrane potential, it likely involves cross-linking of several O-antigens, as the use of a recombinant immunoglobulin single chain did not create this effect.

Glycan-Mediated Binding

Although the Fab-mediated interactions of SIgA are vital for reducing pathogenic infections in the mucosa, as mice deficient in class-switch recombination and somatic hypermutation (AID–/–) [4] who are unable to create high-affinity immunoglobulins demonstrate heightened susceptibility to infection, antigen- independent, glycan-mediated interactions are also important in the function of this molecule. SIgA is highly decorated with N-linked glycans: 2 sites on each heavy chain, 8 in total, plus 7 on the SC, 1 on the J-chain, and an additional 2 on the hinge region on IgA2 (Fig. 1a) [8]. Many studies have now demonstrated that these glycans play a pivotal role in pathogen clearance, antigen presentation via M cells, and commensal homeostasis. For example, enteropathogenic Escherich- ia coli binds terminal sialic acid residues on the glycoprotein with intimin adhesion proteins [9], and Clostridium difficile toxin A neutralization is mediated by glycans [10]. Rochereau et al. [11] showed that sialic acid residues on IgA2 N- glycans (Fig. 1b) are required to bind on the apical surface of M cells via dectin-1 for reverse transcytosis and subsequent antigen presentation in the subepithelial dome. In addition, subepithelial dome dendritic cells require SIgA glycans for binding these immune complexes via DC-SIGN, a C-type lectin receptor [12]. Not only are these glycans on SIgA necessary for mucosal response to pathogens, more recent studies have also shown the utility of these glycans in binding to commensal bacteria, promoting immune tolerance, and maintaining homeostatic regulation. The first evidence of glycan-mediated commensal binding was shown utilizing a SIgA specific for Shigella flexneri LPS and 3 commensal bacte-

160 Dunne-Castagna/Mills/Lönnerdal N-linked glycans N-glycans on IgA2 Light chain Heavy chain SC a J chain

F(ab’)2

Sialic acid Mannose Fucose Galactose Glycan isoforms: Antigen binding cleft: N-acetylglucosamine Pathogen binding – High-affinity pathogen immune exclusion binding – immune b exclusion Commensal binding – protection, colonization Low-affinity commensal binding – colonization? Mucin localization Homeostatic regulation with self-antigen binding c

Fig. 1. Schematic representing the SIgA molecule (a), an example of the glycan isoforms (b), and some of the functions of the glycans and the antigen-binding region (c) on the molecule.

ria: Bifidobacterium lactis, Lactobacillus rhamnosus, and E. coli Nissle 1917 [13]. Using fluorescence microscopy, researchers visualized associations of this nonspecific SIgA to all bacteria, but when pretreated with glycosyl hydrolase (N- glycosidase F) to remove the N-linked glycans, SIgA was not able to bind to the 2 gram-positive bacteria, though it remained bound to the gram-negative one. Although further studies are needed to confirm the glycan dependency of commensals to SIgA, research from our lab and others are supporting the antigen- independent association of SIgA with commensal bacteria. Nakajima et al. [14] validated nonspecific SIgA binding using OTII mice. In their study, ovalbumin (OVA)-specific T cells were transferred into T-cell-deficient (CD3–/–) mice who were then provided with OVA antigen daily to induce an OVA-specific humoral response. Flow cytometry of fecal bacteria revealed complexes of commensal- SIgA-OVA, compared to commensal-SIgA complexes in wild-type counter- parts, which was recapitulated in vitro with hybridoma-derived OVA-specific IgA but not IgG. Our own studies utilizing flow cytometry and fluorescence microscopy have shown a concentration-dependent interaction of both hetero-

Milk SIgA and Commensals 161 0 µg SIgA 10 µg SIgA 100 µg SIgA 1,000 µg SIgA

1,000 1,000 1,000 1,000 0.017 0.016 0.12 10.7 0.12 35.1 0.30 91.7

800 800 800 800

600 600 600 600

400 400 400 400 FL2-W: Alexa 647 FL2-W: Alexa 647 FL2-W: Alexa 647 FL2-W: Alexa 647 200 200 200 200 SIgA (A650) SIgA 0.51 99.3 0.94 88.3 0.38 64.4 0.58 7.42 0 0 0 0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 FL1-H: Syto 9 FL1-H: Syto 9 FL1-H: Syto 9 FL1-H: Syto 9

Syto9 (cells)

Fig. 2. Flow-cytometric scatter plot showing increased pooled milk SIgA associated with a Bifidobacterium species with increased concentration of SIgA (percent circled). Syto9 is a DNA chelator indicating live bacteria, and A650 is anti-human-IgA with Alexa 647 fluoro- phore. Concentration is per 107 CFU bacteria.

geneously pooled milk SIgA (Fig. 2) and recombinant monoclonal SIgA with select commensal bacteria and enteropathogens [unpubl. data]. The utility of the nonspecific SIgA association with commensals is still unclear, though these studies and others have given some indication. Work using Caco-2 colonocytes exposed to the commensal bacteria Lactobacillus and Bifidobacterium demonstrated a 3.4-fold increase in the binding of bacteria to these cells, a reduction in NF-κB-induced proinflammatory cytokine production, and induction of the tight junction binding protein occludin when the bacteria were first associated with SIgA [15]. Our research group has recapitulated these studies and also found a significant reduction in the neutrophil recruiting IL-8 cytokine when SIgA is first associated to the commensals prior to binding in vitro on colonocytes. In addition, our studies have found that nonspecific SIgA association with commensal bacteria improves colonization in BALBc mice when the complex is provided orally [unpubl. data]. Gnotobiotic mice deficient in pIgR (pIgR–/–) are unable to produce SC-coated IgA, but serum and luminal IgA levels remain unchanged. These mice demonstrated a heightened translocation of commensal bacteria to the mesenteric lymph nodes, and an induction of a systemic immune response when introduced to commensal intestinal flora from specific pathogen-free mice. The systemic IgG response to colonization with specific pathogen-free microbio-

162 Dunne-Castagna/Mills/Lönnerdal ta was not seen in wild-type counterpart mice, demonstrating functional immune compartmentalization in the presence of SIgA [16]. Murine studies show that SIgA coating of commensal bacteria promotes colonization in the outer mucosal layer of the colon [17], preventing the stimulation of intraepithelial lymphocytes and epithelial pattern recognition receptors, and thereby reducing the induction of innate inflammatory responses to nonpathogenic bacteria. In the OTII-CD3–/– mouse work, researchers found that Fab-independent association of SIgA with Bacteroides thetaiotamicron altered the expression of several genes involved in carbohydrate utilization and in exopolysaccharide production [14]. This was also seen in a study by Donaldson et al. [18] investigating gene expression changes with Bacteroides fragilis upon nonspecific SIgA association, where they found a change in exopolysaccharide production that led to colonization of a different ecological niche in the small intestine. Our research group has shown a significant reduction in enteropathogenic invasion both in vitro in human colonocytes and in vivo in BALBc mice when SIgA-associated commensals are introduced into the cells or animal prior to pathogen challenge, but not when either the commensal or SIgA alone are provided [unpubl. data]. Corthesy and Phalipon [19] demonstrated that in the respiratory tract, tissue localization was dependent on the glycosylation of the SIgA molecule. Upon intranasal challenge with S. flexneri, the glycosylated SIgA molecule was viewed in close association with the mucosa, and dissemination of the pathogen was con- fined to the nasal cavity, but deglycosylation of SIgA led to pathogenic colonization in the deep lung alveoli. Gibbins et al. [20] demonstrated an association of SIgA with salivary mucin proteins that then bind with other epithelial cell-produced mucin proteins, indicating a mucin-mucin-driven SIgA interaction. While research supports the association of glycosylated SIgA with the outer mucosal layer in the large intestine, Rogier et al. [17] showed through fluorescent microscopy that the mucin MUC2 protein in the innermost mucosal lining of the gut epithelium is important not in binding to SIgA, but in excluding this complex from interacting with the epithelia. The exact mechanisms of SIgA interacting with the loose outer mucosa remain unclear, but examinations of other mucosal sites within mammals demonstrate the importance of the carbohydrate moieties in maintaining associations with both microorganisms and the mucosa.

Breast Milk Secretory IgA

The gut of the neonate, upon parturition, contains a number of innate immune barriers, an underdeveloped gut-associated lymphoid tissue, and naïve adaptive immunity, which can take up to 10 days to become stimulated [21]. These con-

Milk SIgA and Commensals 163 ditions can lead to an imbalanced inflammatory response without proper guid- ance from the mother’s passive immune defenses in breast milk that aid in establishing a regulated environment during the initial onslaught of microbial colonization. SIgA is the primary mucosal antibody found in the highest abundance over other immunoglobulins in milk, with concentrations up to 15 mg/ mL in colostrum and ∼1 mg/ml in mature milk [22], providing the breastfed infant 0.5–1 g/day. The origin of milk antibodies has been evaluated through many studies. Although antibody-secreting plasma cells can be detected in excreted milk [23], the production of SIgA and SIgM likely arises from secreting plasma cells in the basolateral region of the mammary gland to allow for transcytosis through the epithelium and release of SC-bound immunoglobulins. The induction site for these mammary-homing lymphocytes is predominantly the gut-associated lymphoid tissue, with migration through the established enteromammary pathway [24]; early work in this field showed that radioactively labeled IgA-secreting mesenteric lymphocytes were found in the mammary gland of lactating murine dams 3–10 times higher in number than lymphocytes of peripheral lymph node origin [25]. In addition, antigen-specific SIgA can be detected in milk following oral introduction of the antigen in lactating mothers [26]. Antibodies secreted in breast milk have been shown experimentally in rodents and epidemiologically in humans to prevent pathogenic infection in the infant gastrointestinal tract through targeted pathogen-specific SIgA both in infancy and extended through childhood [26]. Harris et al. [27] showed that C57BL/6 mouse pups were protected against the intestinal parasite Heligmoso- moides polygyrus when fostered on milk rich in pathogen-specific immunoglobulins (IgA and IgG), but not when the milk contained no H. polygyrus-specific immunoglobulins. Studies of breastfed human infants infected with the intestinal pathogen Campylobacter jejuni showed that milk containing antiflagellin or anti-outer membrane IgA antibodies [28] is protective against symptomatic C. jejuni infections. These studies and others have led to the exploration of a therapeutic SIgA approach: vaccinating mothers to promote protective SIgA antibodies in breast milk [29], a strategy warranting further investigation.

Milk Secretory IgA and Infant Commensal Colonization

Whereas the protective role of pathogen-targeted secretory milk antibodies is well established, recent studies have emerged focusing on the functional consequences of milk SIgA in association with commensals in the development of the newborn microbiota. Analyses of the milk antibody repertoire show significant

164 Dunne-Castagna/Mills/Lönnerdal amounts of polyclonal autoreactive antibodies with low levels of somatic hypermutation evident [30]. Multiple studies have observed cross-reactivity of both antigen-targeted monoclonal antibodies and polyreactive autoantibodies to commensal microbiota of the small intestine [1, 14, 18]. Thus, there is a high likelihood of the ability for the diverse polyclonal antibody repertoire of breast milk to bind with low affinity to bacteria found both within the mammary gland and in the gut of the breastfed infant. Whereas to date methodology has not been developed to differentiate the maternally versus endogenously derived SIgA in the SIgA-coated fraction of the infant gut microbiota, studies utilizing BugFACS reveal distinct microbiota coated with SIgA under homeostatic conditions [31] or in malnourished Malawian infants [32]. Consistent with IgA-seq studies in adults [33], malnourished infants had a higher relative abundance of Enterobac- teriaceae associated with intestinal inflammation, which represented a significant portion of the SIgA-associated taxa. Conversely, healthy infants showed consistently high SIgA association in members of the genera Akkermansia and Clostridium, regardless of dietary intake or age (1–24 months). Likely, these observed variations in IgA-coated fractions in the infant gut are a reflection of the differences in IgA repertoire development under varying disease conditions in both the infant mucosa and also in the breast milk environment. A recent study investigated the impact of breast milk SIgA on the development of necrotizing enterocolitis (NEC), a disease with complex etiology and high mortality afflicting preterm infants born under 33 weeks of gestation [34]. The study found that NEC development was correlated with a loss of IgA association to Enterobacteriaceae – through unclear mechanisms. The significance of IgA was confirmed in a NEC mouse model where suckling mouse pups were fostered on dams deficient in total immunoglobulins (Rag–/–) or IgA (Igha–/–). These mice developed NEC at the same rate as formula-fed pups, whereas control pups with milk sufficient in IgA had low rates of NEC development. Al- though milk has a plethora of immune components and oligosaccharides able to modify the microbiota and contribute to fortifying the immune system of the young infant, this study supports a crucial role for IgA in the preterm infant. Rogier et al. [35] sought to understand the difference in the function of maternally derived SIgA versus actively secreted endogenous IgA in the newborn gut. In their murine study, they found that the microbiome of adult mice, whose nursing mothers were deficient in pIgR and, therefore, provided no maternal SIgA during the suckling period (although milk IgA levels were unchanged), differed in key taxa from those who were exposed to SIgA during suckling. Specifically, they found a more pronounced presence of Proteobac- teria of the family Pasteurellaceae, and Firmicutes of the family Lachnospira- ceae. Analysis of the RNA component of intestinal epithelial cells of the same

Milk SIgA and Commensals 165 two mouse groups after challenge with epithelial-disrupting dextran sulfate sodium (DSS) showed two major groups of genes varying in expression level between the mice. Notably, many genes involved in cell repair were seen to be linked to DSS exposure and were up-regulated in adult mice with suckling exposure to SIgA, and a cluster of genes important in cell metabolism and growth that were not associated with DSS exposure were also up-regulated in the same group of mice. These reports, although based on a murine model, indicate the importance of early exposure to maternally derived SIgA to the long-term health of the child. Further research is needed to elucidate the mechanisms behind these results. In addition to these functions, SC may be important for protection against proteolytic degradation of the IgA antibody itself and perhaps the bacteria they coat, as evidenced by the presence of intact SIgA in the stool of exclusively breastfed infants [36] and in the 2011 study by Mathias and Corthésy [13], who showed that SC-bound, but not unbound, IgA resisted degradation in vitro. This protein-coating model may be one mechanism by which breast milk commensals survive intestinal digestion in order to colonize the gut of the infant. Indeed, our research group has found increased viability of commensal bacterial following in vitro intestinal digestion when the bacteria are first associated to pooled milk SIgA [unpubl. data].

Conclusion

SIgA plays important roles in the developing infant gut, both in pathogen defense and commensal homeostasis. Although breast milk provides a plethora of biologically active immune components and microbial-modifying factors, studies involving a reductionist approach to investigate individual bioactive molecules have helped elucidate the importance of the heavily glycosylated SIgA both in infancy and into adulthood. Studies are needed to determine the functional consequence of a SIgA-commensal complex both within the milk milieu and in the neonate gut with regard to commensal colonization and regulation of the infant mucosal immune response. Continued investigation into this relationship may translate into improvements in infant health and nutrition.

Acknowledgments

The authors wish to thank members of the Mills lab who contributed to the original research described in this study: Jassim Al-Oboudi, Eric Nonnecke, Hiu Ling Mak, Austin Ghera, Sarah Goldberg, and Meg Tseng.

166 Dunne-Castagna/Mills/Lönnerdal Disclosure Statement

Research was supported by a grant from the Bill and Melinda Gates Foundation and from the Peter J. Shields Endowment Fund for Dairy Food Science at the University of Cali- fornia Davis. V.P.D.-C. and B.L. received stipends from the Nestlé Nutrition Institute for participation in the 94th NNI workshop. D.A.M. has no financial relationship to disclose.

References

1 Bunker JJ, Erickson SA, Flynn TM, et al: Natural 13 Mathias A, Corthésy B: Recognition of gram-pos- polyreactive IgA antibodies coat the intestinal itive intestinal bacteria by hybridoma- and colos-

microbiota. Science 2017; 358:eaan6619. trum-derived secretory immunoglobulin A is

2 Mestecky J, Russell MW, Jackson S, Brown TA: mediated by carbohydrates. J Biol Chem 2011;

The human IgA system: a reassessment. Clin Im- 286: 17239–17247.

munol Immunopathol 1986; 40: 105–114. 14 Nakajima A, Vogelzang A, Maruya M, et al: IgA 3 Jarick KJ, Mokhtari Z, Scheller L, et al: Photocon- regulates the composition and metabolic function version of alloreactive T cells in murine Peyer’s of gut microbiota by promoting symbiosis be-

patches during acute graft-versus-host disease: tween bacteria. J Exp Med 2018; 215: 2019–2034. tracking the homing route of highly proliferative 15 Mathias A, Duc M, Favre L, et al: Potentiation of

cells in vivo. Front Immunol 2018; 9: 1468. polarized intestinal Caco-2 cell responsiveness to 4 Fagarasan S, Muramatsu M, Suzuki K, et al: Criti- probiotics complexed with secretory IgA. J Biol

cal roles of activation-induced cytidine deami- Chem 2010; 285: 33906-33913. nase in the homeostasis of gut flora. Science 16 Reikvam DH, Derrien M, Islam R, et al: Epitheli-

2002; 298: 1424–1427. al-microbial crosstalk in polymeric Ig receptor

5 Mazanec MB, Kaetzel CS, Lamm ME, et al: Intra- deficient mice. Eur J Immunol 2012; 42: 2959– cellular neutralization of virus by immunoglobu- 2970.

lin A antibodies. Proc Natl Acad Sci USA 1992; 89: 17 Rogier EW, Frantz AL, Bruno MEC, Kaetzel CS: 6901–6905. Secretory IgA is concentrated in the outer layer of 6 Moor K, Diard M, Sellin ME, et al: High-avidity colonic mucus along with gut bacteria. Pathogens

IgA protects the intestine by enchaining growing 2014; 3: 390–403.

bacteria. Nature 2017; 544: 498–502. 18 Donaldson GP, Ladinsky MS, Yu KB, et al: Gut 7 Forbes SJ, Martinelli D, Hsieh C, et al: Associa- microbiota utilize immunoglobulin A for muco-

tion of a protective monoclonal IgA with the O sal colonization. Science 2018; 360: 795–800. antigen of Salmonella enterica serovar 19 Corthesy B, Phalipon A: Molecular definition of Typhimurium impacts type 3 secretion and the role of secretory component in secretory IgA-

outer membrane integrity. Infect Immun 2012; mediated protection at mucosal surfaces. J Al-

80: 2454–2463. lergy Clin Immunol 2002; 109(1, suppl 1):S113. 8 Huang J, Guerrero A, Parker E, et al: Site-specific 20 Gibbins HL, Proctor GB, Yakubov GE, et al: SIgA glycosylation of secretory immunoglobulin A binding to mucosal surfaces is mediated by mu-

from human colostrum. J Proteome Res 2015; 14: cin-mucin interactions. PLoS One 2015; 1335–1349. 10:e0119677. 9 Perrier C, Sprenger N, Corthésy B: Glycans on 21 Torow N, Marsland BJ, Hornef MW, et al: Neona- secretory component participate in innate protec- tal mucosal immunology. Mucosal Immunol

tion against mucosal pathogens. J Biol Chem 2017; 10: 5–17.

2006; 281: 14280–14287. 22 Hanson LA: Breastfeeding provides passive and 10 Dallas SD, Rolfe RD: Binding of Clostridium dif- likely long-lasting active immunity. Ann Allergy

ficile toxin A to human milk secretory compo- Asthma Immunol 1998; 81: 523–533; quiz 533–

nent. J Med Microbiol 1998; 47: 879–888. 534, 537. 11 Rochereau N, Drocourt D, Perouzel E, et al: Dec- 23 Tuaillon E, Valea D, Becquart P, et al: Human tin-1 is essential for reverse transcytosis of glyco- milk-derived B cells: a highly activated switched sylated SIgA-antigen complexes by intestinal M memory cell population primed to secrete anti-

cells. PLoS Biol 2013; 11:e1001658. bodies. J Immunol 2009; 182: 7155–7162. 12 Baumann J, Park CG, Mantis NJ: Recognition of secretory IgA by DC-SIGN: implications for immune surveillance in the intestine. Immunol Lett

2010; 131: 59–66.

Milk SIgA and Commensals 167 24 Seppo AE, Savilahti EM, Berin MC, et al: Breast 30 Vassilev TL, Veleva KV: Natural polyreactive IgA milk IgA to foods has different epitope specificity and IgM autoantibodies in human colostrum.

than serum IgA – evidence for entero-mammary Scand J Immunol 1996; 44: 535–539.

link for food-specific IgA? Clin Exp Allergy 2017; 31 Planer JD, Peng Y, Kau AL, et al: Development of

47: 1275–1284. the gut microbiota and mucosal IgA responses in

25 Roux ME, McWilliams M, Phillips-Quagliata JM, twins and gnotobiotic mice. Nature 2016; 534: et al: Origin of IgA-secreting plasma cells in the 263–266.

mammary gland. J Exp Med 1977; 146: 1311– 32 Kau AL, Planer JD, Liu J, et al: Functional charac- 1322. terization of IgA-targeted bacterial taxa from un- 26 Nathavitharana KA, Catty D, McNeish AS: IgA dernourished Malawian children that produce antibodies in human milk: epidemiological mark- diet-dependent enteropathy. Sci Transl Med

ers of previous infections? Arch Dis Child Fetal 2015; 7: 276ra24.

Neonatal Ed 1994; 71:F192–F197. 33 Bunker JJ, Bendelac A: IgA responses to micro-

27 Harris NL, Spoerri I, Schopfer JF, et al: Mecha- biota. Immunity 2018; 49: 211–224. nisms of neonatal mucosal antibody protection. J 34 Gopalakrishna KP, Macadangdang BR, Rogers

Immunol 2006; 177: 6256–6262. MB, et al: Maternal IgA protects against the de- 28 Torres O, Cruz JR: Protection against Campylo- velopment of necrotizing enterocolitis in preterm

bacter diarrhea: role of milk IgA antibodies infants. Nat Med 2019; 25: 1110–1115. against bacterial surface antigens. Acta Paediatr 35 Rogier EW, Frantz AL, Bruno MEC, et al: Secre-

1993; 82: 835–838. tory antibodies in breast milk promote long-term 29 Dolby JM, Newell DG: The protection of infant intestinal homeostasis by regulating the gut mi- mice from colonization with Campylobacter crobiota and host gene expression. Proc Natl

jejuni by vaccination of the dams. J Hyg (Lond) Acad Sci USA 2014; 111: 3074–3079.

1986; 96: 143–151. 36 Davidson LA, Lönnerdal B: Persistence of human milk proteins in the breast-fed infant. Acta

Paediatr Scand 1987; 76: 733–740.

168 Dunne-Castagna/Mills/Lönnerdal Protective Factors in Human Milk

Summary on Protective Factors in Human Milk

It is well known that breastfed infants are protected against infections, have better cognitive development than infants fed formula, and that long-term metabolic diseases such as obesity and diabetes are less common. Many factors in breast milk contribute to these favorable outcomes, and this session intended to highlight recent research on some of the soluble components in human milk and the mechanisms behind their protective actions. Lars Bode started this session by presenting the biosynthesis and structure of human milk oligosaccharides (HMOs), and how they can vary among women with different genetic background (blood groups). Ingested HMOs largely resist digestion through the infant; a small percentage is absorbed and reaches the systemic circulation, and the rest of the HMOs arrives at the colon, where HMOs get metabolized by the infant gut microbiota or are excreted intact into the stool. In his studies in an animal model of necrotizing enterocolitis (NEC), a specific HMO (disialyllacto-N-tetraose) was able to improve survival and reduce pathology scores, which is in agreement with studies in preterm infants fed breast milk with low levels of this HMO having a higher risk of developing NEC. He also discussed recent studies correlating ratios of 2 HMOs with infant height and weight, and that a combination of HMOs is associated with food sensitization. Franz-Georg Hanisch and Cem Aydogan continued the discussion on HMOs. In their contribution, they presented research on how HMOs and algal fucan- type polysaccharides affect viral infections. Hanisch and his research group have found that the trisaccharide 2 -fucosyllactose (2 FL) can block norovirus bind-

′ ′ ing to gastric mucins and that higher-molecular-weight HMOs have even stronger competitive effects. They subsequently found that oligovalent forms (hepta- to decasaccharides) also have competitive effects on norovirus binding. They also described inhibitory effects of algal fucoidans both in gastric mucosa binding studies and in virion challenge experiments in the Pacific oyster. Thus, the antiviral effects displayed by HMOs may also be provided by the fucoidans, which could provide a step towards a prophylactic food additive. Olle Hernell, Bo Lönnerdal, and Niklas Timby described and discussed clinical trials on the milk fat globule membrane (MFGM), a dairy fraction that up to recently has not been part of infant formula. This membrane fraction in breast milk contains several components with anti-infectious activities which are also involved in brain development. Double-blind randomized controlled trials in infants and children have revealed protective effects against diarrhea and otitis media as well as fewer days of fever and less use of antipyretics. Serum lipidomic and metabolomic analyses demonstrate that addition of MFGM to formula diminishes the differences between breastfed and formula-fed infants, suggesting that MFGM may have a role in directing infant metabolism. Investigations into the oral and gut microflora, as well as fecal metabolomics, suggest that MFGM alters the microbiota and several metabolites in the feces, and thus MFGM supplementation renders infants fed formula more similar to breastfed infants. Clinical trials on lactoferrin in preterm infants were discussed in the presentation by Nicholas D. Embleton and Janet E. Berrington. Late-onset sepsis (LOS) and NEC are common causes of death in very preterm infants, and several studies have suggested that bovine lactoferrin or recombinant human lactoferrin supplementation can reduce their occurrence. Lactoferrin may have direct antibacterial activity and thus affect the microbiome, but it may also provide an effect on immune function in these infants. They presented results from a recently completed multicenter study on 2,203 very preterm infants, by far the largest trial to date. The primary outcome was LOS, and there was no difference between the intervention and the placebo group. NEC and all-cause mortality did not differ either. They discussed possible reasons for the lack of effects found, and why the results from their study differed from those obtained in other studies. Rulan Jiang and Bo Lönnerdal introduced the protein osteopontin (OPN), which is abundant in breast milk, but low in cow milk and infant formula. OPN is involved in cell proliferation and differentiation, biomineralization, immunomodulatory activities, and myelination. They described studies in mouse pups receiving milk from wild-type and OPN-knockout dams, respectively, and showed that milk OPN can in part survive digestion and affect intestinal maturation, immune function, brain protein expression, and cognitive development.

170 Lönnerdal In a clinical trial, infants fed regular formula or formula supplemented with bovine OPN at 2 different concentrations were compared to a breastfed reference group. Infants fed OPN-supplemented formula had significantly fewer days of illness than infants fed regular formula, similar to breastfed infants. Further, serum cytokine levels in infants fed formula with OPN were similar to those of breastfed infants. Thus, milk OPN may play essential roles in infant health and early infant development. In the final presentation of this session, Vanessa P. Dunne-Castagna, David A. Mills, and Bo Lönnerdal reported recent findings on the effect of breast milk secretory IgA (SIgA) on the commensal microbiota. They reviewed evidence for both naturally polyreactive and antigen-specific monoclonal immunoglobulins with capacity to bind to enteric pathogens, and then expanded on recent data showing antigen-independent binding of commensal bacteria. Such nonspecific binding of SIgA promotes colonization of commensal bacteria in a unique mucosal niche of the intestine, reducing competition within the complex micro biota and providing benefit to the host. Their own research has shown that polymeric human milk SIgA associated to select commensals protects them from proteolysis in vitro. Thus, milk SIgA may retain antipathogenic protective functions using high-affinity interactions while concurrently maintaining homeostatic commensal colonization in the infant intestinal mucosa. Bo Lönnerdal

Summary of Session III 171 Subject Index

Antisecretory factor, milk 17 Gut microbiome, see Microbiome

Bifidobacterium infantis, breast milk House dust mite, allergic sensitization tryptophan interactions 107–109 through breast milk 54, 55 Breastfeeding benefits Immunoglobulins, milk maternal 22 activity 16, 17, 42, 43 neonatal 22, 23 antigens 30 immunological connection between autoantibodies 43 mother and infant 39 induction by immunization 31, 32 neonatal risks 23, 24 milk microbiota effects 99 noncommunicable disease protection mucosal immunity 32–34, 61 in later life 66, 67 reactivity 17 secretory IgA Casein, evolution 6, 7 advantages in human milk 29 CD14, milk 17, 18, 99 glycan-mediated binding 160–163 Colostrum, see Immunology, milk, and immune exclusion 159, 160 lactation infant commensal colonization role 164–166, 171 Evolution, see Lactation evolution infection prevention 163, 164 Exosomes, milk 100 production 159 specificity and antigenic stimulation FDSCP, see Follicular dendritic cell 30, 31 secreted peptide types 16, 28, 29 Follicular dendritic cell secreted peptide Immunology, milk, and lactation (FDSCP) 6 chemokines 18 Fucosyltransferase (FUT) colostrum 16 functional overview 125 cytokines 18, 39–41, 62 genotypes and neonatal immunity 62, epithelial cells 19, 43 63 growth factors 18, 100 variants and milk oligosaccharide historical perspective 11–15, 60, 61 composition 118, 119 Immunoglobulins, see FUT, see Fucosyltransferase Immunoglobulins, milk infection prevention 51, 52 leukocytes 18 long-term susceptibility to immune- infants dependent diseases 52–55 gut colonization with breast macrophages 19 feeding 67–70 mucosal immunity 32–34 lactoferrin interactions 143–145 osteopontin in immune development milk fat globule membrane 155, 156 concentrate 137, 138 prospects for study 34–36 milk secretory IgA and stem cells 19, 100, 101 colonization 164–166, 171 T cells 19, 20, 44 mastitis prevention in nursing 81–84 Milk fat globule (MFG) Lactation evolution evolution 7, 8 casein evolution 6, 7 membrane mammary gland evolution 3–5 concentrate supplementation milk fat globule evolution 7, 8 effects in infants 137, 138 milk oligosaccharide evolution 8, 9 infections in infants and overview 2–4 children 134–137 Lactoferrin prospects 138, 170 gut microbe interactions 143–145 structure and synthesis 134 milk composition and function overview 133, 134 141–143 Milk microbiome, see Microbiome preterm infant studies Milk oligosaccharides, see ELFIN trial 147–149 Oligosaccharides, milk late-onset sepsis prevention 145, Milk virome, see Virome, milk 146–149 MPSU, see Mammopilosebaceous unit necrotizing enterocolitis prevention 145, 147–149 NEC, see Necrotizing enterocolitis Late-onset sepsis (LOS) Necrotizing enterocolitis (NEC) lactoferrin prevention 145, 146–149 immune responses 105 overview 142, 143 lactoferrin prevention 145, 147–149 LOS, see Late-onset sepsis milk microbiota in prevention 104, 106–110 Malaria, breast milk antigens 55 milk secretory IgA prevention 165 Mammary gland evolution 3–5 overview 104, 142 Mammopilosebaceous unit (MPSU), short-chain fatty acid studies 110 formation 5, 59, 60 Norovirus Metabolome oligosaccharide binding inhibition milk fat globule membrane dendrimer studies 129, 130 concentrates 137, 138 fucoidan binding 130, 131 MFG, see Milk fat globule in-solution binding experiments Microbiome 128, 129 human milk neoglycolipid array identification clinical significance 72, 73 of norovirus binders 126, 127 immunoglobulin effects 99 overview 125, 126 metabolomics 97, 98 necrotizing enterocolitis ODAM, see Odontogenic, ameloblast- prevention 104, 106–110 associated protein origins 77–81 Odontogenic, ameloblast-associated overview 68, 69, 76, 77, 96, 97 protein (ODAM) 6

Subject Index 173 Oligosaccharides, milk Preterm infants, see Late-onset sepsis; composition 116, 117 Necrotizing enterocolitis digestion 119 evolution 8, 9 Rotavirus, milk studies 22, 120, 135 functions 119–121 norovirus binding inhibition SCFAs, see Short-chain fatty acids dendrimer studies 129, 130 SCPPs, see Secretory calcium-binding fucoidan binding 130, 131 phosphoproteins in-solution binding experiments Secretory calcium-binding 128, 129 phosphoproteins (SCPPs) 6, 60 neoglycolipid array identification Short-chain fatty acids (SCFAs), of norovirus binders 126, 127 necrotizing enterocolitis studies 110 overview 125, 126 synthesis, see Fucosyltransferase T cells therapeutic applications 121, 122 milk 19, 20, 44 Oligosaccharides, see Milk Th2 responses and allergy 49, 50 oligosaccharides TLRs, see Toll-like receptors OPN, see Osteopontin Toll-like receptors (TLRs), milk 17, 18, 99 Osteopontin (OPN) milk Virome, milk functions 153, 170, 171 diversity and taxonomy 87, 88 immune development role 155 prospects for study 91, 92 intestinal proliferation and role in infant health 89-91 maturation role 153, 154 transmission 88, 89 neurodevelopment promotion 154, 155 Xanthine oxidoreductase (XOR) 7, 8 overview 152, 153 XOR, see Xanthine oxidoreductase OVA, see Ovalbumin Ovalbumin (OVA), allergy prevention in mice through breastfeeding 53, 54

174 Subject Index