Assessment of Human Milk in the Era of Precision Health

Abstract

Purpose of review:

Precision health provides an unprecedented opportunity to improve the assessment of infant nutrition and health outcomes. Breastfeeding is positively associated with infant health outcomes, yet only 58.3% of children born in 2017 were still breastfeeding at 6 months. There is an urgent need to examine the application of precision health tools that support the development of public health interventions focused on improving breastfeeding outcomes.

Recent findings:

In this review, we discussed the novel and highly-sensitive techniques that can provide a vast amount of omics data and clinical information just by evaluating small volumes of milk samples such as RNA sequencing, cytometry by time-of-flight, and human milk analyzer for clinical implementation. These advanced technics can run multiple samples in a short period of time making them ideal for the routine clinical evaluation of milk samples.

Summary:

Precision health tools are increasingly used in clinical research studies focused on infant nutrition. The integration of routinely collected multi-omics human milk data within the electronic health records has the potential to identify molecular biomarkers associated with infant health outcomes.

INTRODUCTION

Brief History of Breastfeeding:

Breastfeeding has been shown to be associated with both long-term and short-term health outcomes in mother and infants (1–3). Although little is known about breastfeeding habits in prehistoric humans, high-resolution trace-element geochemical analysis of prehistoric teeth collected from weaning infants found that breastfeeding as measured by barium concentration started to drop between 1–2 years after birth (4). These results are generally consistent with geochemical analysis of teeth collected from early humans nearly 2 million years ago that revealed children were routine breastfed up to 6 years after birth (5). In contrast, recent Centers for Disease Control and Prevention (CDC) data in the United States indicates that nearly 83.9% of infants born in 2018 started breastfeeding; however only 56.7% of these children were still breastfeeding at 6 months and 26.8% were exclusively through 6 months (6*). Currently, no racial or ethnic groups are meeting the Healthy People 2020 goals of 50% exclusive breastfeeding at 6 months (7). In response to these data, a systematic review conducted by U.S. Preventive Services Task Force (USPSTF) on breastfeeding promotion found interventions including formal breastfeeding education for mothers and their families, direct support, training of primary care staff, support techniques for breastfeeding, and peer support can promote breastfeeding from 1–3 months and 6–8 months (8). Despite evidence-based interventions, patient-level factors were ultimately the largest barriers to successful implementation (9,10*).

Exclusive Breastfeeding Challenges:

The health benefits of breastfeeding for mom and infant are well established; however, there are significant challenges to implementing health infant feeding practices at both the individual and societal-level. Challenges to the individual include learning from health guidelines (11), lack of knowledge of breastfeeding benefits (12*,13*), and health conditions of newborns (14). Societal-level challenges includes health care professional advice and social opportunity (15). To understand and prevent the social-level challenges of breastfeeding, a cross-sectional study in 653 women at 5–22 gestational weeks found that awareness of the WHO guidelines for breastfeeding had a significant effect on their intention to breastfeed or exclusively breastfeed (16). Moreover, knowledge of the impact of breastfeeding to infant health outcomes also plays an important role in the intention to breastfeed (17). Specifically, mothers of premature infants tend to overlook the benefits of breastfeeding (12) and insufficient physical suction along with respiratory and swallowing dysfunction due to preterm birth objectively reduce the chance of an infant being breastfed (14). One study in 100 infants admitted to the NICU at Nationwide Children’s hospital concluded the impact of personalized feeding programs in infants showed significant improvement, such as higher success in breastfeeding, earlier discharge, and reduction in economic burden using the personalized feeding method compared to traditional feeding method (18). Collectively, these studies demonstrate the need for personalized breastfeeding support that accommodate mothers and infants.

Precision Health in Infant Nutrition:

Precision health is an data-driven approach to developing proactive and personalized solutions to health problems that can be applied to the domains of precision public health and precision medicine (19). Precision public health is focused on community-level data outside the medical system, such as disease prevention and health improvement (20). Perinatal examples of precision public health have been applied to non-invasive prenatal screening (21). In contrast, precision medicine is an approach that is focused on individual differences in people’s genes, environment, and lifestyles into account for disease treatment and prevention (22–24**). An example of precision medicine to reduce preterm birth included transcriptomic analysis of RNA extracted from tissue to identify individuals with sign premature activation preterm labor (25). A review of precision medicine approach for detection and diagnosis in perinatal depression summarized that perinatal depression as one of the most common pregnancy complications is associated with alterations in the gut microbiome of mothers and infants (26). A morphine analgesia application of precision medicine developed a personalized neonatal pain treatment using genetic targets to predict the effects of opioid analgesia in newborns. Precision health studies requires data diversity, integrating information from various sources, such as patient-level clinical and multi-omics data collected from biospecimens (Figure 1).

Figure 1.

Personalized health data sources composition.

Human Milk Composition:

Human milk is an ideal food for infant nutrition (25) due to the bioactive components of human milk such as carbohydrates, proteins, and lipids (27). Nutritional and non-nutritional bioactive components of human milk differ based on length, time, and stage of lactation (28). The composition of bioactive components of human milk is complex and varies based on nutritional intake, diet, environmental influences, and many other factors that can impact human milk’s nutritional value and safety (28). Despites these observations, there is an increasing need for routinely collected information on human milk macronutrient and cellular composition as well as function information related to gene expression and metabolomics. Below, we review several innovative molecular technologies that can be used to profile human milk and inform precision health studies in the area of infant nutrition.

Mid-infrared Imaging System-human Milk Analyzer:

Evaluating the bioactive components of milk in clinical setting is critical to address the medical concerns in infant nutrition. But such techniques for being used in routine clinical practices need certain requirements such as clinical feasibility, accuracy, fastness and most importantly Food and Drug Administrator (FDA) approval. Conventional analyses of human milk composition and macronutrients have required individual and dedicated methodologies that require a relative large volume of milk (27). These methods are time-consuming and limited to research purposes rather than clinical application. As a consequence, there is currently no routinely collected data on human milk composition in the EHR. New methods are needed for clinical human milk analysis that require lower sample volume and shorter quantification time (27). More recently, the mid-infrared Imaging system-human milk analyzer (MIRIS-HMA) has been FDA approved to quantify the macronutrient including lactose, fat, and protein in human milk samples (29). Recent data has shown high correlation with conventional gold standards methods including Röse-Gottlieb extraction (30,31) for quantification of milk fat (24), lactose (32), and protein (32). Taken together, technological advances and FDA approved devices such as MIRIS-HMA are enabling routinely collected information on macro composition of human milk and infant nutrition that will be stored in EHR.

Mass Cytometry by Time-of-Flight:

Human milk has various cellular composition include lactocytes, myoepithelial cells, immune cells and a hierarchy of progenitor and stem cells, however the functional significance of these cellular components remain poorly characterized (33). Mass cytometry by time-of-flight (CyTOF) is an advanced cytometry technique that can be use to identify cellular characteristics of human milk using single cell analysis (34). As an example, high-dimensional mass cytometry was used to characterize a unique and previously uncharacterized sub-population of immune cells in fetal tissue (35) that indicates the high-throughput sensitivity and specificity of mass cytometry has potential to routinely analyze the large number of milk samples in a shorter time course. In addition, the integrated high-dimensional analysis algorithm makes CyTOF highly capable of identifying cellular heterogeneity (36). The integrated computational analysis algorithm consists of algorithms for cellular characteristic identifying, algorithms for comparing and identifying cellular population such as viSNE, SPADE, PhenoGraph, X-shift and Citrus (34). CyTOF also has a powerful application capacity to run a large number of samples in a short period of time (37*). Collectively, CyTOF provides an opportunity for researchers to understand and analyze the cellular composition in human milk and the impact of perinatal health outcomes.

RNA Sequencing:

Human milk feeding has been shown to provide healthy gastrointestinal mucosal stimuli, which can help improve infants’ immune systems (38) and also reduce the risk of infectious disease in infancy (39**). Mammary epithelial cells (MEC) derived from milk samples have recently emerged as a non-invasive model system to evaluate lactation physiology (40**). Previous work to characterize the MEC using transcriptome techniques in milk fat globules (MFGs) identified transcriptional signatures in women with normal lactation and diet-controlled gestational diabetes (40**). Single cell RNA sequencing (scRNA-seq) was used to profile transcriptome of the MEC in a randomized controlled trial with 59 diet-controlled women during early lactation who had gestational diabetes, but normal lactation. Epithelial and non-epithelial cell populations in human milk were identified with previously established immunostaining methods which labeled mammary epitlial cells with CD49f and EPCAM, and endothelial cells and hematopoietic lineages with CD31 and CD45 (40**). Through this developed scRNA-seq method the researchers identified immune cells, fatty acid synthesis pathways as well as identify the lower expression of the triglyceride synthesis. Collectively, these results show that single cell technologies such as RNA-seq have potential to identify cellular and molecular biomarkers in human milk samples that may inform infants immune system development.

Metabolomics in Human Milk:

In addition to nutritional components of human milk, there is a growing list of bioactive components including enzymes, cytokines, chemokines, hormones, growth factors, oligosaccharides, and antimicrobials (39**). Recent advances in metabolomics (MS) has increased the number of techniques available to characterize clinically relevant small molecules in biological tissues such as human milk. Currently, high-throughput MS-based analysis of human milk includes nuclear magnetic resonance (NMR) and mass spectrometry (MS)(41). The metabolomics signature and profile of early human milk (colostrum) has been identified via NMR (42**) to understand the impact of extreme fetal growth condition on the nutritional value of the milk (42**). Specifically, human milk samples were collected from 98 eligible participants across intrauterine-growth restriction that included appropriate-for-gestational-age (AGA), large-for-gestational-age (LGA), and intrauterine growth-restricted (IUGR) full-term neonates. A total of 35 metabolites were involved in transfer RNA biosynthesis and metabolism of branched chain amino acid (BCAA), glycerophospholipid and citrate cycle (42**). Moreover, human milk lactose, citric acid, choline, phosphocholine, and N-acetylglutamine were associated with LGA and IUGR while BCAA were associated with AGA (42**). These results showed alterations in different human milk metabolites may influence the nutrition value of early breast milk/colostrum in relation to fetal growth. In another study, gut microbiome and fecal were analyzed using MS-based techniques in breast fed infants during lactation (43**). The fecal microbiome was evaluated by 16S rRNA-seq and human milk oligosaccharides were measured by NMR (43**). These result showed that bifidobacterium species were correlated with changes in human milk oligosaccharides using time-series clustering analysis (43**). Collectively, these results demonstrate that MS-based analysis of human milk has the potential to identify novel human milk compounds that are associated with infant outcomes such as growth and development (44**).

CONCLUSION

Breastfeeding has a profound impact on early life health outcomes including infant growth as well as the functional development and immune system (45,46*). Integration of precision health tools and muli-omics human milk data may help to optimize the nutitional needs of infants at the individual level. Evaluating multi-omics human milk data can lead to identification and assessment of biomarkers and next-generation therapies focused on improving infant health outcomes. Barrier to routinely evaluating human milk composition in clinical settings include sample volumes, analysis time, and requirements for large laboratory equipment and technical staff, which prevents clinicians from implementing these conventional methods during routine perinatal care. In this review, we discussed precision health and novel high-throughput molecular (scRNA-seq, CyTOF, and MIRIS-HMA) techniques that can generate multi-omics human milk data with potential for routine clinical implementation. The integration of routinely collected multi-omics human milk data within the electronic health records has the potential to identify molecular biomarkers associated with infant health outcomes (Figure 2).

Figure 2.

Integration of personalized health data has potential advantage to switch from conventional infant feeding into personalized infant feeding that could lead to better early life health outcome.

KEY POINTS:

• There are challenges to implementing health infant feeding practices at both the individual and societal-level.

• Personalized health by utilizing integrated omics data could be the optimum approach to address infants’ nutritional needs at the individual level.

• Integration of multi-omics human milk data within electronic health records has signficant potential to identify human milk biomarkers that can inform public health interventions focused on improving exclusive breastfeeding.