Saturday, May 16, 2015

Milk responds: Changes to milk immune factors with infant (or maternal) infection

FOREWARD. As many of you know, my recent research has been looking at milk composition, infant growth, and maternal health in a population of ethnic Tibetans living in the Himalayas of Nepal. The communities I work with were within 50 miles of the epicenter of the earthquake on April 25.  While the loss of life in these communities was minimal (thankfully), there was considerable destruction of homes, clinics, schools, and infrastructure (water, latrines). With the upcoming monsoon season, there is considerable need for safe drinking water, food storage, medical care, and safe homes. Several NGOs with long standing relationships with the communities are currently fundraising for relief and rebuilding efforts. Please consider donating to these organizations if you can afford to do so (NepalSEEDS; Tsum-Nubri Relief Center). We are still committed to these communities, and will continue to support infrastructure and research to promote maternal and child health.

In the last blog post – January – I discussed the idea of immunological memory in milk, particularly the well described association between maternal exposure to pathogenic bacteria in early life and the immunological memory of those bacteria, by specific forms of secretory Immunoglobulin-A (sIgA), many years later. Milk is incredibly dynamic, and this is certainly true for the immune factors in milk. Three recent papers have investigated this responsiveness in several samples, using a variety of immune factors to measure immune activation in milk.

Breakey et al., (2015) have articulated this as a model of two systems within the mammary gland – a protective paradigm, where some immune factors in milk are always protecting against infection; and the responsive paradigm, where active infection will increase the concentrations of immune factors in milk. Of the hundreds of known (and many unknown!) immune factors in milk, some will be generally protective, and others will be responsive (Brandtzaeg 2010). A few, including secretory IgA, will be both.  

Breakey et al., (2015) investigated the responsiveness of immune factors in milk to current infection using two biomarkers – sIgA and lactoferrin – in a sample of 29 Toba mother-infant dyads followed longitudinally. Both of these biomarkers have come up before (for reviews: sIgA; lactoferrin).The Toba are indigenous population from Argentina (Figure 1); previous generations have subsisted as foragers, but more recently the population has become increasingly concentrated in peri-urban areas, often in informal settlements lacking access to sanitation and water facilities. 
Figure 1: Location of the Toba. Image from wikicommons, author Nazareno98; produced in 2008.

Milk samples and interviews were collected monthly, allowing for the researchers to investigate milk composition before, during, and after an infection in the infant. Infant infections during the preceding month were collected during monthly interviews; all infants in the study had at least one illness over the course of the longitudinal study. Mothers did not report frequent illnesses, although this may have been underreporting. 

In this sample, infants receiving milk with higher sIgA were less likely to be ill, while infants receiving milk with more lactoferrin were more likely to be ill. Although causation cannot be certain, the authors hypothesize that lactoferrin content of milk increases during an infection (responsive) while sIgA levels are more generally protective. 

The study with the Toba follows two earlier studies of immune responsiveness in milk, both done in WEIRD populations. The earliest, by Riskin et al., (2012) remains one of my favorite papers. In this study, Riskin et al., recruited 51 mother-infant dyads, younger than 3 months, from Haifa, Israel into the study. 31 mother-infant pairs were hospitalized for fever at the time of recruitment, with an additional 20 pairs serving as healthy controls. Milk samples were collected from the mothers while the infants were hospitalized, and then seven days later; samples from controls were collected at one week intervals. Milk samples were analyzed for immune cells (lymphocytes, neutrophils, macrophages, CD45+), sIgA, lactoferrin, TNF-alpha, and IL-10 (Figure 2). 

Figure 2: Important cells in the immune system. Not all are found in milk. Image credit:
For the purposes of analyses, the participants were grouped into 3 categories: controls (healthy mom, healthy baby; n=20), all sick (all infants sick, moms sick or not; n=31), and sick infant (only baby sick; n=20). For the control group, there were no changes in the immune factors measured in milk from time 1 to time 2. However, for the sick group, there were significant declines in CD45+ cells, lymphocytes, neutrophils, macrophages, IL-10, and TNF-alpha. Lactoferrin and sIgA also declined, but the differences were minor. It does not appear that the associations were simply responding to maternal infection either. In the 20 mothers of sick infants who were not ill themselves, milk cd14+ cells, neutrophils, and macrophages also showed a significant decline from the original to the after measure. All other immune factors also showed declines, but again these were relatively minor. 

In an additional study of 21 mother-infant pairs, Hassiotou et al., (2013) reported increased leukocytes, and sIgA in the milk of mothers with infections compared to earlier and later samples from the same mothers collected as part of a longitudinal study design. While both maternal and infant infection increased leukocytes and sIgA in milk, this was most pronounced for mothers with breast infections. 

One of the leading hypotheses for how maternal physiology may respond to infection in the infant is through oral contact. Saliva from the infant’s mouth may enter the breast, carrying the pathogens responsible for the infection. This would encourage a localized immune response to the pathogen in the mammary gland itself (Hassiotou et al., 2013), although Riskin et al., (2012) also propose a model of subclinical infection in the mothers.

The capacity for milk to balance between innate and adaptive immune responses is incredibly important, especially for infants living in highly pathogenic, low resource environments such as the Toba, or my own participants from Nubri (more on this to come). Certainly, having a milk to gut superhighway for immune factors should be incredibly important in promoting gut integrity, decreasing infant illness, protecting against growth faltering, and promoting infant survival. Infant – or maternal – illness becomes then not a reason to stop nursing, but a reason to nurse more.

Brandtzaeg P. (2010) The mucosal immune system and its integration with the mammary glands. J Pediatr. 156(2 Suppl):S8-15

Breakey AA, Hinde K, Valeggia CR, Sinofsky A, Ellison PT. (2015) Illness in breastfeeding infants relates to concentration of lactoferrin and secretory Immunoglobulin A in mother's milk. Evol Med Public Health. 2015(1):21-31. doi: 10.1093/emph/eov002.

Hassiotou F, Hepworth AR, Metzger P, Tat Lai C, Trengove N, Hartmann PE, Filgueira L. (2013) Maternal and infant infections stimulate a rapid leukocyte response in breastmilk. Clin Transl Immunology. 2(4):e3. doi: 10.1038/cti.2013.1
Riskin A, Almog M, Peri R, Halasz K, Srugo I, Kessel A. (2012) Changes in immunomodulatory constituents of human milk in response to active infection in the nursing infant. Pediatr Res. 71(2):220-5. doi: 10.1038/pr.2011.34.

Saturday, January 3, 2015

Milk remembers: Immune factors in milk “remember” childhood environments

It is well established that with very few exceptions, human milk is the preferred first food for infants. While the benefits of breastfeeding/receiving human milk are considerable and influence the development of multiple systems in the infant, perhaps the best known benefits of human milk are its immunoprotective properties. Worldwide, breastfeeding is associated with reduced risk of infectious diseases in infants, and these protections persist even in highly hygienic conditions such as the United States (Bartick & Reinhold 2010). Many immune factors are found in human milk, including immune cells, cytokines that regulate immune responses, and secretory Immunoglobulin-A (sIgA), perhaps the most common immunoprotein in human milk. It is well established that there is considerable variation in the immune factors in milk between individual mothers and between populations. It is also known that many of the immune factors in milk are highly responsive, changing in response to active infection of either the mother or infant (blog post on this topic coming next month). 

It has been traditionally held that the differences in immune factors in milk, especially sIgA, were reflecting the pathogenicity of the environment. The higher levels of sIgA found in the milk of women in developing countries was thought to be a proximate response to pathogen exposure in the immediate environment. However, some old – and some new – research suggests that the associations may be much more interesting. What if the past environment was just as important as the current environment in influencing sIgA and other immune factors in milk?

To study this, Nathavitharahna et al., (1994) decided to compare sIgA in the milk of women from three groups (sample size): women born in and currently living in Sri Lanka (n=64), women who had immigrated to England from Sri Lanka or other nearby countries in South Asia (n=20), and women born in and currently living in England (n=75).  Pathogen exposure for these groups broke down as follows: Sri Lankan women – high early life, high present; Immigrant women – high early life, low present; and British women – low early life, low present.
Surprisingly, as shown in Figure 1, there were no differences in the total amount of sIgA in the mean amounts of sIgA for each group – and within each group, milk sIgA ranged from 0.2 g/L to 19.1g/L! 

Figure 1: Comparison of total sIgA content in the milk of the three groups of mothers.
Mothers from England actually had more sIgA than mothers from Sri Lanka! However, total milk sIgA is only half the story. The researchers went on to look at specific sIgA antibodies to Escherichia coli (E. coli). They focused on 14 strands of E. coli commonly associated with moderate to severe diarrheal illness. Figure 2 is borrowed from the paper. Mothers living in Sri Lanka had the highest amount of E. coli specific sIgA in their milk. However, E. coli specific sIgA was much higher in the milk of immigrant mothers compared to their British neighbors. 
Figure 2: The amount of sIgA specific to each form of E. coli found in human milk for the three groups. Both horizontal and vertical axes are the same for each graph - white British women have much less E. coli specific sIgA in their milk than immigrant women or Sri Lankan women.

How can we interpret these findings? Nathavitharana et al., briefly considered that the pathogens may be maintained in the community, but further study demonstrated this was unlikely. Instead, the most logical explanation is that the sIgA in milk was the product of milk immunological “memory”. The sIgA in milk is produced by a type of immune cell called a B cell. These cells migrate to the mammary gland, often from the GI tract during last gestation/the onset of milk production. B cells include a special class of B cells, called memory cells. Memory B cells maintain “memories” of prior infections, allowing for a rapid antibody-mediated immune response should re-exposure occur. Memory B cells, it seems, were recording the mother’s own exposure history, migrating to the mammary gland, and providing infants with protection against the pathogenic E. coli experienced by their mother. Already having sIgA antibodies against common – and severe - pathogens may provide infants with increased capacity to resist or limit the severity of infection by these pathogens.

Six years later, another study, using a similar study design, provided further evidence for an immunological memory in milk. Holmlund et al., (2010) looked at three groups – women born and currently living in Mali (Africa), women who had migrated from Africa to Sweden (multiple countries represented), and women born and living in Sweden and analyzed their milk for several cytokines involved in immune function and sIgA. There were roughly 30 women in group. Holmlund and colleagues found few differences in the cytokines of the milks with two exceptions – Transforming Growth Factor beta ( forms 1 & 2). Women from Mali had the highest concentrations of each, with immigrant women having intermediate levels and Swedish women the lowest levels. Here’s the cool part – TGF-B2 is part of the signally cascade for sIgA, and sure enough – there was a significant association between TGF-B2 in the milk and sIgA. However, this association was only significant in the women living in Mali – but it was really, really close in the immigrant women.  

What does this all mean? It’s evidence for a complex immunological memory in milk. While this may not be important in highly hygienic environments such as the United States, it certainly suggests that there may be adaptive features in milk that record pathogen exposures during early life and provide a “dictionary” of potential infections to the infant. These highly specific forms of sIgA antibodies in human milk may allow for a more rapid immunological response by the infant. Milk “memories” therefore, would serve to protect the infant. How long these memories are retained is another question, and it seems likely that there will be rapid drift in the types of sIgA antibodies reflecting novel exposures by the infants. 

References - links included to open access papers

Bartick M, Reinhold A. 2010. The burden of suboptimalbreastfeeding in the United States: a pediatric cost analysis. Pediatrics 125(5):e1048-56. doi: 10.1542/peds.2009-1616.

Holmlund U, Amoudruz P, Johansson MA, Haileselassie Y, Ongoiba A, Kayentao K, Traoré B, Doumbo S, Schollin J, Doumbo O, Montgomery SM, Sverremark-Ekström E. 2010. Maternal country of origin, breast milkcharacteristics and potential influences on immunity in offspring. Clin Exp Immunol. 162(3):500-9. doi: 10.1111/j.1365-2249.2010.04275.x.

Nathavitharana KA, Catty D, and McNeish AS. 1994. IgA antibodies in human milk: epidemiological markers of previous infections? Archives of Disease in Childhood 71(F192-197).

Sunday, December 14, 2014

Human milk has a microbiome - and the bacteria are protecting mothers and infants!

The human microbiome project was a major undertaking by the National Institutes of Health, with a fairly simple mission: understand the bacterial communities living in and on the human body, and the potential impact these communities may have on health. Hundreds of individuals donated everything from feces to nasal secretions. However, one key system was ignored - human milk. That’s right – the microbiome of human milk was not studied.
Probably some of this had to do with a long standing myth that human milk was sterile. Why study something without bacteria, right? But, as we have quickly learned – human milk is far from sterile. The average baby consuming 800 mL/27 ounces of human milk will received between 100,000 and 10,000,000 million bacteria from human milk per day (Fernandez et al., 2013).

Figure 1: The Human Microbiome Project is not interested in milk. I fixed their image to better reflect this.
Fortunately, research into the human milk microbiome has continued despite this oversight by the Human Microbiome Project. It appears that nine “operational taxonomic units” (generally closely related species based on DNA analysis of the bacteria) are extremely common in most mothers studied to date: Streptococcus, Corynebacteria, Bradyrhizobiaceae, Staphylococcus, Serratia, Ralstonia, Propionibacterium, Pseudomonas, and Sphingomonas. These nine groups typically account for more than 50% of total bacteria. Bififobacterium and Lactobacillus are also common, but less universal (Fernandez et al., 2013).
The microbiota of milk appears to be quite stable (Fernandez et al., 2013), although a few factors appear to shape the composition. First, mothers with higher BMIs (in the obese range) produce colostrum with more Lactobacillus, and mature milk with more Staphylococcus and less Bifidobacterium (Cabera-Rubio et al., 2012). Cabera-Rubio and colleagues (2012) also found that greater pregnancy weight gain predicted more Staphylococcus in the milk in a small study of 18 mothers, half obese and half of normal weight. 

But here is the really neat part – guess what else altered the milk microbiota? Type of delivery. Mothers who had caesarian sections had a different milk microbiota than mothers who had a vaginal delivery. And the variation continued – mothers undergoing emergency caesarians after laboring had milk microbiotas closer to those of women who delivered vaginally than women with elective caesarians.

Where do the bacteria come from? Initially, it was thought that the milk microbiome was really just contamination from the skin microbiome. However, this is WRONG, WRONG, WRONG. While the milk microbiome does contain some of the same families of bacteria as skin, multi-site sampling of mammary skin and milk revealed that these are not the same species and/or genera. Instead, it appears that the microbiome of milk comes from several places, including the maternal gut microflora. Current evidence supports dendritic cells as the likely transfer mechanism. These cells, along with some macrophages, can open the tight junctions between cells forming the gut barrier and take in living bacteria. These cells can then maintain the live bacteria for several days in mesenteric lymph nodes scattered throughout the body (Fernandez et al., 2013). Dendritic cells are also pretty picky about what they take up – dead bacteria or latex beads will not activate immature dendritic cells for bacteria uptake, while commensal species, like Lactobacillus, show high levels of binding (Rescigno et al., 2001).
Figure 2: Dendritic cell (shown in blue). Image from

This allows for oral manipulation of the milk microbiome – mothers given supplemental Lactobacillus from three strands, L. gasseri, L. fermentum, L. salivarius, showed transfer of these strands to the milk (Jimenez et al., 2008). 

This lead to the logical question – could these strands be used to treat mastitis? Arroyo et al., (2010) randomized 352 women with mastitis to three groups – one dosed with L. fermentum, one dosed with L. salivarius, and one given standard antibiotic treatment (4 different drugs were used). Bacterial counts for milk were obtained for all mothers on Day 0 – that is before treatment started. All mothers had bacterial counts of 4.35-4.47 log10 CFU (colony forming units) – a little less than double the recommended bacterial counts for milk of 2.5 log10 CFUs. Mothers received 21 days of treatment, and milk bacterial counts were repeated on day 21. Women who received L. fermentum had mean bacterial counts of 2.61 log10 CFUs; L. salivarius had bacterial counts 2.33 log10 CFUs with clinical relief of mastitis, and all reported reductions in reported breast pain. Mothers who received antibiotics did not fare as well. Mean bacterial count for antibiotic receiving mothers was 3.28 log10 CFUs and pain scores were much higher. Three months later, only 8.8% of mothers receiving either L. fermentum or L. salivarius had experienced recurrent mastitis, while 30.1% of mothers receiving antibiotics had. All differences between antibiotic and probiotic groups were significantly different – the kind of significant difference that makes researchers do their happy dance.
Figure 3: This is how I picture the researchers after making this discovery - just substitute a computer for the piano. Gif by PEANUTS.

So the milk microbiome appears to be protecting mothers – but there is also good evidence that it is protecting infants. Little is known about the salivary microbiome of infants, but based on preliminary evidence, it appears to, not surprisingly, have some overlap with the milk microbiome (Nasidze et al., 2009). The milk microbiome also appears to contribute to the microbiome of the infant GI tract, as well as the development of immune function in the infant (Fernandez et al., 2013). Infants supplemented with Lactobacillus fermentum (yes, the same as used for the treatment of mastitis) showed significant reductions in diarrheal and respiratory infections in early infancy compared to control infants (Maldonado et al., 2012). Many of the bacteria in the milk microbiome are protecting both the mother and the infant from infection, and may even be involved in the development of immune tolerance.

Milk remains amazing – even the bacteria in milk!

Arroyo R, Martín V, Maldonado A, Jiménez E, Fernández L, Rodríguez JM. (2010) Treatment of infectious mastitis during lactation: antibiotics versus oral administration of lactobacilli isolated from breast milk. Clinical Infectious Diseases 50:1551–8.

Cabrera-Rubio R1, Collado MC, Laitinen K, Salminen S, Isolauri E, Mira A. (2012) The human milk microbiome changes over lactation and is shaped by maternal weight and mode of delivery. Am J Clin Nutr. 96(3):544-51.

Fernández L1, Langa S, Martín V, Maldonado A, Jiménez E, Martín R, Rodríguez JM. (2013) The human milk microbiota: origin and potential roles in health and disease. Pharmacol Research 69(1):1-10.

Jiménez E, Fernández L, Maldonado A, Martín R, Olivares M, Xaus J, et al. (2008) Oral administration of lactobacilli strains isolated from breast milk as an alternative for the treatment of infectious mastitis during lactation. Applied and Environment Microbiology 74:4650–5.

Maldonado J, Ca˜nabate F, Sempere L, Vela F, Sánchez AR, Narbona E, et al. (2012) Human milk probiotic Lactobacillus fermentum CECT5716 reduces the incidence of gastrointestinal and upper respiratory tract infections in infants. Journal of Pediatric Gastroenterology and Nutrition 54:55–61.

Martín R, Olivares M, Marín ML, Fernández L, Xaus J, Rodríguez JM. (2005) Probiotic potential of 3 lactobacilli strains isolated from breast milk. Journal of Human Lactation 21:8–17.

Nasidze I, Li J, Quinque D, Tang K, Stoneking M. (2009) Global diversity in the human salivary microbiome. Genome Research 19:636–43.
Rescigno M, Urbano M, Valzasina B, Francolín M, Rotta G, Bonasio R, et al. (2001) Dendritic cells express tight junction proteins and penetrate gut epithelial monolayers to sample bacteria. Nature Immunology 2:361–7.