Main

Infants who are exclusively fed human milk have a lower incidence of bacterial infections and necrotizing enterocolitis than formula-fed infants do (1–4). This benefit is thought to be the result of milk-borne hormones and peptides including secretory IgA, lysozyme, lactoferrin, and interferon. In many cases, these peptides are present in high concentrations in milk and are present in greater concentrations in preterm (PT) human milk compared with term (T) milk (5–7).

One factor present in human milk is the hematopoietic cytokine, G-CSF (8–10). Although the physiologic role of milk-borne G-CSF is not known, its recognized hematopoietic functions include support of neutrophil proliferation, differentiation, and survival (11). Recently, we identified the receptor for G-CSF on the villous border of developing human fetal intestine (12). Therefore, the G-CSF contained in human milk may function within the developing gastrointestinal tract or be absorbed and exert actions distantly, such as in the bone marrow. However, a prerequisite for either role is gastrointestinal luminal survival of the protein in either its intact form or as an active digestion product. In early postnatal life, the balance of factors affecting luminal proteolysis favors limited protein digestion (13, 14). Therefore, as a first step in evaluating the physiologic role of milk-borne G-CSF, we sought to determine whether endogenous G-CSF in human milk and rhG-CSF added to human milk and infant formulas survive in vitro conditions that simulate PT and T neonatal digestions.

METHODS

Human milk samples.

Human milk was obtained from breast-feeding mothers who delivered at term (n= 15, 39 ± 1.5 wk gestation, mean ± SD) or prematurely (n= 15, 30 ± 4.5 wk gestation). Samples were centrifuged at 14 000 ×g for 30 min, and the aqueous fractions were separated from the lipids and solid precipitate. Samples were frozen at −80°C until assayed. A subset of these samples was separated as described above, pooled, aliquoted, and used as a pooled human milk sample. All samples were thawed on ice at time of assay. All studies were approved by the University of Florida Institutional Review Board, and informed consent was obtained from all study participants.

Infant formulas and cow's milk.

Three commercially prepared infant formulas used in our neonatal intensive care unit were selected for study. These included Similac Special Care (Ross Laboratories, Columbus, OH), Similac 24 (Ross Laboratories), and Enfamil 24 (Mead Johnson Laboratories, Evansville, IN). Infant formulas were not centrifuged before incubation with PT gastric secretions. Similarly, commercially processed whole cow's milk (Publix, Lakeland, FL) was not centrifuged before incubation with T gastric secretions. The remainder of the experiments, in which either the infant formulas or cow's milk was used, were conducted using conditions identical to those for the analysis of human milk.

Gastric secretions.

Gastric juices were collected from premature and T infants who had an existing indwelling orogastric or nasogastric catheter, were <7 d of age, and were not receiving any oral feeds. Specifically, PT gastric secretions (n= 15, 27 ± 2.5 wk gestation, mean ± SD) were collected from infants <30 wk postconceptual age, and T gastric secretions (n= 20, 40 ± 2.5 wk gestation) were obtained from neonates >37 wk. The secretions were removed as part of routine care and were pooled and frozen at −80°C. At the time of assay, they were cleared of particulate material by centrifugation at 14 000 ×g for 10 min.

ELISA.

G-CSF was quantified by ELISA (Quantikine Human G-CSF Immunoassay, R & D Systems, Minneapolis, MN; lower limit of sensitivity, 7 pg/mL). The assay has no measurable cross-reactivity with other cytokines and recognizes both natural and recombinant G-CSF. In our laboratory, we validated the ELISA for human milk determinations of G-CSF, including linearity, intraassay precision, interassay precision, reproducibility, and percent recovery (10).

Simulations of digestion.

The susceptibility of milk-borne G-CSF to degradation was evaluated using neonatal gastric secretions as an in vitro simulation of neonatal digestion. For simulations involving PT human milk, PT gastric secretions were used. Similarly, for simulations using T human milk, T gastric secretions were used. Simulations were run for 0 (baseline degradation), 1, and 2 h. These intervals were selected because after human milk feedings, maximal gastric acid proteolytic activity occurs at 1 h (15), and, although continuous, stomach emptying generally occurs between 1 and 2 h (16). To approximate neonatal gastrointestinal luminal conditions without interfering with proteolytic activity, three incubation buffers were used in standard reaction mixtures (17). The buffers included 1) 0.1 M glycine at pH 3.2 to simulate preprandial gastric conditions;2) 0.1 M maleate at pH 5.8 to simulate postprandial gastric conditions; and 3) 0.01 M Tris 6.7 mM CaCl2 at pH 7.4 to simulate proximal small intestinal conditions. The standard reaction mixture included one of the three buffers, neonatal gastric juices, and the substrate (human milk, rhG-CSF in protein-free buffer, one of the commercially prepared premature infant formulas, or whole cow's milk).

To ensure that endogenous enzymes were active for each simulation, 50 μL of each of the three pH buffers was preincubated for 15 min at 37°C with 50-μL aliquots of pooled, thawed, and cleared neonatal gastric secretions. To start the reaction, a 50-μL aliquot of substrate was added and incubated at 37°C (150 μL total volume). The substrate was 1) the aqueous fraction of PT or T human milk, 2) the aqueous fraction of pooled human milk “spiked” with rhG-CSF, 3) commercially prepared premature infant formulas spiked with rhG-CSF, 4) commercially prepared whole cow's milk spiked with rhG-CSF, or 5) a known concentration of rhG-CSF in protein-free buffer. Reaction mixtures were evaluated at all three pH buffer conditions on all substrates for baseline, 1, and 2 h at 37°C. At the end of incubations, 5 μL of 1 M Tris buffer, pH 8, was added to stop the reaction. Samples were then frozen at −80°C until assay.

To determine whether a protective effect for G-CSF existed over the entire physiologic concentration range of endogenous G-CSF in milk (10), various concentrations of rhG-CSF ranging from 40 to 1250 pg/mL were added to pooled human milk samples. Reaction mixtures included both PT and T gastric secretions incubated with PT and T human milk, respectively. Samples were evaluated as previously described.

Data analysis.

Results are expressed as mean G-CSF concentration in pg/mL ± SEM. An symbol 97 level <0.05 was considered significant. G-CSF concentrations remaining in reaction mixtures at 1 and 2 h were compared with baseline by using multiple paired t tests.

RESULTS

We evaluated the capacity of human milk (n= 15 PT and 15 T) to protect against proteolytic degradation of endogenous G-CSF after exposure to neonatal gastric secretions. Baseline concentrations of endogenous G-CSF were lower in the PT milk samples (35 ± 5.5 pg/mL, mean ± SEM) than in the T samples (70 ± 10.7 pg/mL) (p< 0.05 for each pH level). After 1 and 2 h of simulated digestions, no significant degradation of G-CSF had occurred in either PT or T milk at any of the pH levels (Fig. 1).

Figure 1
figure 1

Concentration of endogenous G-CSF in human milk after in vitro simulations of digestion. G-CSF concentrations (pg/mL) in PT (A) and T (B) human milk are shown. PT milk was incubated with PT gastric secretions, and T milk was incubated with T gastric secretions. No significant degradation of endogenous G-CSF occurred at any time point for each of the pH levels (pH 3.2, white bars; pH 5.8, hatched bars; pH 7.4, black bars).

Full size image

When rhG-CSF was made to a final concentration of 120 pg/mL in protein-free buffer and combined with PT gastric secretions, >80% was measurable at pH 3.2 and 5.8 at 1 and 2 h (Fig. 2A). At pH 7.4, 60% of G-CSF was measurable at 1 h and only 40% at 2 h. The data seem to show decreased G-CSF concentrations at the highest pH, but the results were not statistically different (p= 0.5 and 0.4 compared with pH 3.2 at 1 and 2 h, respectively; and 0.2 and 0.2 compared with pH 5.8 at 1 and 2 h, respectively). When rhG-CSF was made to a final concentration of 60 pg/mL in a protein-free buffer and incubated with T gastric secretions, no significant differences were noted in the concentration of rhG-CSF at any of the times in any of the three pH levels examined (Fig. 2B).

Figure 2
figure 2

Concentrations of G-CSF after incubations of rhG-CSF in protein-free buffer mixed with gastric secretions from PT and T neonates. G-CSF concentrations (pg/mL) of rhG-CSF in protein-free buffer after incubations using PT (A) and T (B) gastric secretions are shown. No significant degradation of rhG-CSF occurred at any time point for each of the pH levels.

Full size image

Commercially prepared premature infant formulas to which a known quantity of rhG-CSF (final concentration, 312 pg/mL) was added were studied in a similar fashion (Fig. 3, A–C). Greater than 95% degradation was observed at 1 and 2 h when compared with baseline (p< 0.05) for each of the formulas. Whole cow's milk supplemented with rhG-CSF to achieve a final concentration of 50 pg/mL was evaluated at the three pH levels in a manner similar to human milk. Significant degradation of rhG-CSF occurred at 1 and 2 h at each of the pH levels (p< 0.05). To determine whether rhG-CSF was stable in infant formulas and cow's milk before the in vitro simulations, G-CSF concentrations were evaluated after incubations of 0, 10, and 30 min, and 1 and 2 h at room temperature. At each of the time points, >85% of supplemented rhG-CSF was measured.

Figure 3
figure 3

Concentrations of G-CSF in infant formulas and whole cow's milk supplemented with rhG-CSF after in vitro simulations of digestion. G-CSF concentrations (pg/mL) measured after incubations of supplemental rhG-CSF are shown for (A) Similac Special Care, (B) Similac 24, (C) Enfamil 24, and (D) whole cow's milk. Infant formulas were incubated with PT gastric secretions, and cow's milk was incubated with T gastric secretions. Significant degradation of rhG-CSF occurred at each pH level at 1 and 2 h for each of these formulas and cow's milk (★p< 0.05 compared with baseline, paired t tests).

Full size image

To determine whether G-CSF was protected from digestion over the entire range of rhG-CSF in human milk, various concentrations of rhG-CSF, from 39 to 1250 pg/mL, were added to PT and T milk and evaluated at each of the pH levels (Tables 1 and 2). No significant degradation of rhG-CSF occurred at 1 and 2 h at any pH level over the entire range.

Table 1 PT human milk with various concentrations of added rhG-CSF
Full size table
Table 2 T human milk with various concentrations of added rhG-CSF
Full size table

DISCUSSION

During in vitro conditions that simulate preprandial, postprandial, and small intestinal conditions of neonates, we found that endogenous G-CSF in PT and T human milk resists degradation. As previously reported by our group, baseline concentrations of G-CSF in this study were significantly greater in milk from women who delivered at term compared with those who delivered prematurely (10). Similarly, when rhG-CSF was added to human milk to correspond with the range of endogenous G-CSF previously reported in human milk (10), G-CSF resisted degradation over the entire range studied (40 to 1250 pg/mL). Although some degradation occurred when rhG-CSF was evaluated in a protein-free buffer, most was still measurable. However, when rhG-CSF was added to three commercially available premature infant formulas and to whole cow's milk, significant degradation occurred. However, rhG-CSF added to these infant formulas at room temperature and not subjected to digestive conditions was stable for the 2-h observation period, suggesting that rhG-CSF in infant formulas is not degraded by the infant formulas but, rather, was not protected by the formula during simulated digestive conditions.

Milk increases the pH of the stomach postprandially, and, thus, ingested proteins resist gastric digestion by proteolytic enzymes, which require low pH levels for activation (18). G-CSF may be protected from degradation by the presence of antiproteolytic agents known to be present in milk (18). Although small intestinal secretions were absent from our digestion mixture, secondary to the inherent difficulty in obtaining these, it is possible that a greater percentage of G-CSF would be degraded in their presence.

Although G-CSF was detected by ELISA, it is not known if this represents an intact and functional protein. The epitope of G-CSF to which the ELISA antibody was made has not been identified, and, therefore, if the ELISA antibody binds a digested G-CSF fragment, it is not necessarily a functional fragment. Further, it is not possible to discern between degradation of G-CSF and the presence of factors interfering with its measurement by the ELISA. Indeed, it is conceivable that proteins present in cow's milk or infant formulas, but absent in human milk, bind to G-CSF and interfere with epitope recognition.

Although it is unknown whether milk-borne G-CSF has any effect on the neonate, local and systemic effects after the oral administration of other growth factors have been reported. Epidermal growth factor, present in high concentrations in human and rodent milk (19), is a potent growth and maturation factor for the intestinal tract (19–21) and also has systemic effects (22).

Recently, Kling et al. (17) reported that milk-borne Epo resists degradation and is immunoreactive after exposure to neonatal gastric secretions. Other published data support the hypothesis that milk-borne Epo can affect erythropoiesis. In infant rats, oral administration of Epo results in elevated serum Epo concentrations and accelerated erythropoiesis (23–25).

Many of the biologically active peptides in human milk are absent in infant formulas. We speculate that the infant formulas used in this study were unable to protect rhG-CSF from degradation because they lack antiproteases that interfere with proteolysis or, alternatively, because glycosylation of the protein does not occur in infant formulas (13, 14). Alternatively, perhaps like IL-1 and IL-8, G-CSF is compartmentalized and, therefore, protected from digestion (26). Additional studies are necessary to define the specific physiologic mechanisms by which G-CSF is protected from proteolytic degradation.

Although the significance of G-CSF in human milk is not known, we speculate that its function may parallel that which is known for blood-borne G-CSF and that physiologic concentrations of milk-borne G-CSF are involved in immunologic protection of the gastrointestinal tract and, therefore, reduce the incidence of infection in the neonate. The finding that rhG-CSF combined with human milk is protected from degradation by neonatal gastric secretions suggests that incorporating pharmacologic doses of rhG-CSF in milk may be possible. Further studies are needed to determine whether orally administered rhG-CSF 1) has a local effect on the enterocytes that express G-CSF receptors, 2) is absorbed from the neonatal gastrointestinal tract in an active form, and 3) if absorbed, has an effect on granulocytopoiesis.