In a comparison of before and after white blood cell, cytokine, immunoglobulin A (IgA), and lactoferrin responses in milk of mothers of babies with putative infectious illness vs. controls, Riskin et al. (1) have found that during active infection in the nursing infant, the total number of white blood cells—specifically, the number of macrophages—and levels of tumor necrosis factor-α increase. This study supports a very dynamic interaction between an infant' s health status and the mother's response that likely enhances the immunologic defenses of the mother's milk.

This interplay between the infant's health status and the response in the mother's milk is highly reminiscent of the concept that Kleinman and Walker (2) introduced more than 30 years ago, termed the “enteromammary system.” In their conceptual framework, antigen presented from the infant to the maternal gut via contact between the mother and the baby is brought into proximity to the mother's lymphoid follicles, which in turn commit lymphoblasts to specific IgA production, which then migrate to the breast and secrete sIgA, which is ingested by the infant. Furthermore, in this concept, T cells, B cells, and macrophages are also thought to extrude into the breast milk and are immunologically active. However, at the time the “enteromammary system” was conceptualized, the basis was primarily studies done in animals. The present study supports this concept nicely by showing responses in human infant mothers' milk (consisting of increased white blood cells, primarily macrophages, tumor necrosis factor-α to infectious illness thought to originate from the infant.

Secretory IgA and lactoferrin have been thought to be primary defenses in human milk, yet, curiously, neither showed a significant response in this study. While the total concentration of sIgA did not change, it is not known whether production shifted toward the specific, causative infectious agent. Although it is known that human milk changes in response to developmental changes in infant nutritional needs, the literature reports this more as a species phenomenon than as an individual mother-baby dyad interaction (3,4,5,6).

One caveat to these studies that Riskin et al. aptly raise is that it is possible that some of the findings in the first milk samples from the mothers of the sick infants (e.g., more CD45 cells and higher tumor necrosis factor-α levels, especially in comparison with those found in the controls) actually reflect a nonspecific stress response of the mothers whose infants were hospitalized. Although this is unlikely given the nature of the response, it cannot be ruled out.

Another interesting point is that approximately a third of the mothers reported being ill at the same time as the infant, although it was not specifically stated whether they had symptoms similar to those of their infant. The authors recognized this as a potential confounder and excluded these samples in additional analysis. It is reassuring that their findings of significant decrease of CD45 leukocytes and fraction of macrophages persisted and that trends for other markers persisted.

Overall, this study provides support of the dynamic nature of human milk and how the mother's immune system adjusts to her infant's infectious status. It is not known, though, how long this interaction continues. As the infant becomes more immunocompetent, does the maternal milk response wane? One also might wonder whether examination of maternal milk would be a useful adjunctive tool for diagnosis of infection in infants. Would skin-to-skin practices trigger antigen-specific antibodies and other defenses in milk that would aid preterm infants in neonatal intensive care units? Clearly, this paper accomplishes more than adding to the body of knowledge; it begs for more study on the mysteriously modulating mothers' milk.