Human Milk Bactericidal Properties

Effect of Lyophilization and Relation to Maternal Factors and Milk Components

Abstract

Objective: 

Lyophilization appears to be a viable method for storing human milk, assuring no microbiological contamination and preserving its health benefits and antibacterial properties. The aim of the study is to evaluate and compare the effects of different storage methods (lyophilization and freezing at −20°C and −80°C) and maternal factors (gestational length or time postpartum) upon the microbiological contents and bactericidal activity of human milk. The possible relation between bactericidal activity and the content of certain nutrients and functional components is also investigated.

Methods: 

Microbiological content, bactericidal activity, sialic acid, and ganglioside contents, as well as protein, fat, and lactose concentrations were assessed in 125 human milk samples from 65 healthy donors in the Human Milk Bank of La Fe (Valencia, Spain).

Results: 

Lyophilization and storage at −80°C significantly reduced the content of mesophilic aerobic microorganisms and Staphylococcus epidermidis when compared with storage at −20°C. Bactericidal activity was not significantly modified by lyophilization when compared with freezing at either −20°C or −80°C. Bactericidal activity was not correlated with fat, protein, or lactose content, but was significantly correlated to ganglioside content. The bactericidal activity was significantly greater (P < 0.05) in mature milk and in milk from women with term delivery than in milk from early lactation (days 1–7 postpartum) and milk from women with preterm delivery, respectively.

Conclusions: 

Lyophilization and storage at −80°C of human milk yields similar results and are superior to storage at −20C with regard to microbial and bactericidal capacities, being a feasible alternative for human milk banks.

What Is Known/What Is New

What Is Known

• Present conditions for storing human milk include freezing at −20°C/−80°C, modifying its defensive properties and microbial content depending on time–storage conditions.
• Bactericidal capacity of human milk is considered as one of the most important defensive properties, remaining unknown factors controlling.
• Scarce studies have evaluated human milk lyophilization as storage alternative, founding none assessing its effects on nutritional or functional properties.

What Is New

• Human milk lyophilization does not modify its nutritional properties, bactericidal activity, or microbiological properties.
• Bactericidal capacity of human milk depends on the gestational age and time postpartum, and is related to total ganglioside content.

Breast-feeding is the best exclusive feeding option for infants during the first 6 months of life. Different authorities such as the World Health Organization, the American Academy of Pediatrics, and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition recommend pasteurized donor breast milk supplied by milk banks as the best alternative when direct breast-feeding is not possible ⁽¹⁾. In effect, donor milk affords additional benefits for premature or low-birth-weight infants in addition to its nutritional characteristics, such as protection against necrotizing enterocolitis (lessened incidence of necrotizing enterocolitis), infections, and improved food tolerance ^(2–4). The bactericidal capacity and microbiota of human milk are 2 properties that contribute to a specific intestinal colonization ⁽⁵⁾.

The retention of these properties in milk that has been pasteurized or stored at −20°C has been established ^(6–8), although the effects of other storage methods such as lyophilization (also known as freeze drying) or freezing at −80°C have not been explored to date.

Human milk banks (HMBs) are increasingly common throughout the world ⁽⁹⁾, and standards based on the guidelines published by different national HMBs have been adopted—although there may be some variations in the processing criteria applied in each case. From the time of expression of the donor milk to consumption, the storage conditions used aim to maximally preserve all of the product's properties and avoid the risk of contamination.

The most common storage method is freezing, and in this regard, some of the properties of milk are modified to 1 degree or other, depending on the temperature and duration of storage ^(10–14).

Lyophilization is highly valued as a storage technique for complex products that require quality guarantees such as drugs, microbial strains, biological tissues, and certain foods. In this context, lyophilization offers maximum preservation of the properties of the original product. The use of lyophilization for storing human milk may be a good alternative to freezing in HMBs, allowing prolongation of the shelf life of the milk and lowering the cost of storage.

Although many studies have evaluated the stability of cold-stored (refrigerated or frozen) milk, few have examined the effects of lyophilization upon human milk, and the existing publications are moreover old and largely limited to effects upon the fat globule ⁽¹⁵⁾ or certain protective actions ⁽¹⁶⁾. The effects of lyophilization upon the nutritional value, bioactive compound stability, sensory characteristics, and microbiological properties of human milk have not been established.

Our hypothesis is that lyophilization is a suitable alternative to freeze storage at −80°C and −20°C, and does not increase harmful microbial content or decrease its nutritional and bioactive functionality. Moreover, this study examines the possible effect of gestational length, time postpartum, and the contents of certain nutrients and bioactive compounds (sialic acid and gangliosides) upon the bactericidal capacity of milk.

METHODS

Subjects and Samples

A total of 125 human milk samples from 65 healthy donors in the HMB of La Fe Hospital (Valencia, Spain) were analyzed. All of the donors gave written informed consent to use the donated milk for research purposes. As part of the routine protocol of the HMB, a full clinical history was compiled in each case, including the characteristics of the donated samples. The study was approved by the hospital ethics committee. Milk was collected at home in sterile polypropylene containers using a breast pump (Lactina, Medela, Switzerland) and was immediately frozen at −20°C until transfer to the HMB, where the milk was stored under the same conditions until shipment to the laboratory of CEU-UCH University (Valencia, Spain). Frozen storage conditions were maintained at all times.

Of the 125 total samples collected, 50 were used to study the effect of lyophilization upon the microbiological and nutritional properties: 20 samples for microbiological counts (5 mL) and 30 for the study of bactericidal capacity (2 mL) and the analysis of proteins, lipids, and lactose (3 mL).

The effect of time postpartum and gestational length upon the bactericidal capacity of the milk was analyzed using 75 samples: 20 samples corresponded to the early lactation period (collected in the first 7 days after delivery) and 37 to mature milk (collected between 15 and 125 days after delivery). In turn, 27 of the samples came from women with gestations between 26 and 36 weeks (preterm), and 25 were donated by women with gestations between 38 and 41 weeks (term).

The relation between bactericidal capacity and ganglioside and sialic acid content was evaluated in 15 milk samples with sufficient volume (40 mL) to perform 3 determinations for each sample. Figure 1 summarizes the use of samples in the course of the study.

FIGURE 1: Flowchart summarizing the use of samples in the course of the study. HM = human milk.

At the time of processing, each sample was thawed and divided into 3 aliquots of 10 mL each: first aliquot (A) was destined for immediate analysis after thawing in the refrigerator at 6°C to −8°C; a second aliquot (B) was frozen at −80°C; and a third aliquot (C) was frozen at −80°C, followed by lyophilization. After a period of 10 to 14 days, aliquot B was thawed at room temperature, whereas lyophilized aliquot C was reconstituted to original weight with bidistilled water. Immediate analysis of the corresponding parameters was carried out in all cases.

Lyophilization

Sample lyophilization was carried out following the instructions of the equipment (Telstar Cryodos, Mannheim, Germany), under conditions of −40°C and 12 mbar. Briefly, 10 mL of sample was frozen at −80° C in sterile plastic tubes, and the equipment was turned on. Once the system reached working conditions (−80 °C, 0.01 mbar) and the samples were totally frozen, the samples were taken from the freezer, covered with loose parafilm, and introduced into the system. The samples were dry after 2 to 3 days.

Microbiological Count

For microbiological analysis, 1 mL of each sample was diluted with 9 mL of 0.1% wt/vol sterile peptone water (Oxoid; Unipath Ltd, Basingstoke, UK). To estimate microbial counts, appropriate dilutions of sample homogenates were prepared and inoculated in duplicate into growth media.

Mesophilic aerobic counts were determined by incubating plate count agar (Oxoid) at 31°C for 72 hours. Staphylococcus aureus was determined on Baird Parker Agar (Oxoid) with 10% (vol/vol) of Egg Yolk Tellurite Supplement by incubating the samples at 37°C for 24 hours. Enterococci were grown on Slanetz-Bartley (Oxoid) medium incubated at 37°C for 48 hours. Finally, Staphylococcus epidermidis was determined on Mannitol Salt Agar (Oxoid) incubated at 30°C for 48 hours. The results were expressed as colony-forming units per milliliter.

Determination of Bactericidal Capacity

The bactericidal effect against Escherichia coli NCTC 9111 serovar O111:K58 (B4):H− was assessed using the method described by Silvestre et al ⁽⁶⁾. The degree of bacteriolysis was calculated as the difference between E coli counts in the control and milk samples, expressed as a percentage of the control sample counts.

Autoanalyzer Nutrient Determination

Fat, protein, and lactose determinations were carried out using a human milk analyzer (Miris, Sweden). Briefly, samples were warmed at 37°C and homogenized with an ultrasonic homogenizer (VCX 130; Sonics & Material, Newtown, CT) for 30 seconds, and 2 mL were injected into the analyzer. Results were expressed as grams per 100 mL of milk. Before analysis, a daily calibration check was performed using the calibration solution (MIRIS check), which was provided by the supplier.

Determination of Gangliosides and Sialic Acid

Sialic acid content, expressed as Neu5Ac, was determined by spectrophotometry according to Salcedo et al ⁽¹⁷⁾. Briefly, samples were hydrolyzed with H₂SO₄ and purified with ion exchange chromatography. Neu5Ac content was determined by spectrophotometry following the resorcinol procedure of Svennerholm ⁽¹⁸⁾.

For the determination of gangliosides, lipid extraction was carried out following the methodology described by Puente et al ⁽¹⁸⁾. Once the lipid extract was obtained, the total ganglioside contents were determined as lipid-bound sialic acid using the resorcinol procedure of Svennerholm ⁽¹⁸⁾, whereas ganglioside identification was performed by high-performance thin-layer chromatography as previously described ⁽¹⁹⁾.

Statistical Analysis

The normal distribution of the data was evaluated using the Kolmogorov-Smirnov test. Parametric statistics were used in the event of P > 0.05, applying 1-factor analysis of variance (ANOVA) for the effect of storage methodology upon microbial content, or multifactorial ANOVA to evaluate the effects of storage method, gestational age, and time postpartum upon bactericidal capacity. In all cases, a later application of the Tukey test was used to check differences between groups. Moreover, multiple linear regression analysis was performed to explore the relation between different parameters (bactericidal capacity with nutritional composition (fat, protein, lactose), bactericidal capacity with ganglioside (total or individual) and sialic acid content). In the case of populations with a nonnormal distribution (P < 0.05), we used the nonparametric Kruskal-Wallis test to analyze differences between groups.

RESULTS

Effect of Storage Method Upon Microbiological Counts

Table 1 shows the results obtained in the microbiological analyses of mesophilic aerobic microorganisms, S aureus, S epidermidis, and Enterococcus spp, in the samples frozen at −20°C, samples frozen at −80°C, and samples frozen at −80°C with subsequent lyophilization.

TABLE 1: Effect of the storage method upon microbiological counts

The Kruskal-Wallis test used to assess the effect of the storage method upon the counts of each of the microbial species indicated that freezing at −80°C and lyophilization resulted in significant changes (P < 0.05) in the counts of mesophilic aerobic microorganisms and S epidermidis. Specifically, lyophilization significantly lowered the counts of both groups of organisms, whereas freezing at −80°C only affected S epidermidis, with a drop in count versus baseline. Neither S aureus nor Enterococcus spp showed significant differences in counts in either the lyophilized samples or the samples frozen at −80°C when compared with the corresponding counts in samples stored at −20°C.

Effect of Storage Method Upon the Bactericidal Capacity of Milk

All of the analyzed samples exhibited bactericidal capacity. Table 2 shows the bactericidal capacity of the samples frozen at −20°C and of those stored at −80°C, both with and without subsequent lyophilization. A normal distribution was confirmed in all of the cases.

TABLE 2: Effect of the storage method upon the bactericidal capacity of human milk

Multifactorial ANOVA used to investigate whether the nature of the milk samples or the method used to store the milk influenced the bactericidal capacity revealed no significant differences (P > 0.05) attributable to either the donors or the storage method used. Lyophilization of the milk produced no significant losses in bactericidal capacity.

Effect of Maternal Factors (Gestational Age and Time Postpartum) Upon the Bactericidal Capacity of Milk

An analysis was made of the possible relation between the properties of the milk (time postpartum and gestational age) and its bactericidal effect. Table 3 reports the bactericidal capacity corresponding to the early lactation period and mature milk, and for milk from women with preterm and term deliveries. A normal distribution was confirmed in all cases.

TABLE 3: Relation between maternal factors and bactericidal capacity of human milk

Significant differences (P < 0.05) were recorded for both time postpartum and gestational age. In effect, the samples of mature milk had greater bactericidal capacity than colostrum, and samples from women with term delivery exhibited greater bactericidal capacity than the samples from women with preterm delivery.

Relation Between Bactericidal Capacity and Nutritional Composition of Milk

The relation between the bactericidal capacity of milk and its lipid, protein, and lactose contents was analyzed in 30 samples. The average bactericidal capacity of these samples, expressed as the percentage inhibition of E coli growth, was 63.60% ± 13.75%. This value falls within the range obtained when bactericidal activity is evaluated depending maternal factors). The mean concentrations (in g/100 mL) of fat, protein, and lactose in the samples were 3.45 ± 1.80, 1.07 ± 0.54, and 6.90 ± 0.39, respectively.

Multiple regression analysis using the bactericidal capacity of the samples as dependent variable and the concentration of the 3 macronutrients mentioned (fat, protein, and lactose) as independent variables revealed no significant relation for any of them (P > 0.05), that is, the bactericidal capacity of the milk was not related to its fat, protein, or lactose contents.

Relation Between Bactericidal Capacity and Ganglioside and Sialic Acid Contents of Milk

The study of the relation between bactericidal capacity and functional components was carried out in 15 samples. Their mean bactericidal capacity (expressed as the percentage inhibition of E coli growth) was 70.83% ± 25.08%. The mean concentration of sialic acid and total gangliosides was 230.00 ± 22.10 mg/L and 550.80 ± 160.00 μg/L, respectively.

Multiple regression analysis of the relation between the bactericidal capacity of milk and its total ganglioside and sialic acid contents revealed a strongly significant correlation (P = 0.0000). Figure 2 shows the line adjusted to the model, with a correlation coefficient of r² = 0.8475.

FIGURE 2: Linear relation between total ganglioside content (μg/L) and bactericidal capacity (% inhibition).

DISCUSSION

Human milk is not sterile. The use of culture-independent methods has allowed to identify and classify specifically bacteria present in human milk, Staphylococcus, Streptococcus, Enterococcus, Lactococcus, Lactobacillus, Weissella, and Leuconostoc being the most common of the different microorganisms identified ^(20,21). A recent review has pointed out the beneficial roles of human milk microbiota in infant health: it favors maturation of the infant immune system, and contributes to reduction of incidence and severity of infections and possible promotion of a “healthy” microbiome in the infant gut ⁽²²⁾. Some microorganisms present in milk can, however, also cause illness under certain conditions, such as Clostridium or Staphylococcus, among others ^(23,24). The main concern of HMB is to ensure the safety of human milk, trying thus to eliminate HMB, not considering the possible lost of bacteria exerting positive effects for the body.

The effects of different freezing conditions upon the microbiological quality of human milk are known ^(25–28), although to our knowledge there have been no previous evaluations of the global effect of the successive freezing steps. Likewise, we know of no previous studies on the effects of lyophilization as milk storage method in HMBs. In the present study we compared the effect of lyophilization with that of present human milk storage methods (−20°C, −80°C) in terms of pathogenic bacterial abundance, the latter being a key concern in HMBs.

Mesophilic aerobic microorganisms and Staphylococcus (S epidermidis and S aureus) were the most frequently identified microorganisms, whereas the genus Enterococcus was isolated in only 20% of the samples analyzed. Staphylococcus and Streptococcus have been described as the predominant genera in fresh milk, and their counts do not vary when the product is frozen at −20°C for a period of 3 days to 6 months ⁽²⁹⁾. This suggests that the freezing of our reference samples at −20°C should not have altered the abundance of these genera.

Subsequent sample storage at −80°C or in lyophilized (freeze-dried) form resulted in a decrease in S epidermidis counts in both cases. The number of mesophilic aerobic microorganisms was not modified by freezing at −80°C, but was decreased significantly in the lyophilized samples when compared with both the reference values and the samples frozen at −80°C. Neither S aureus nor Enterococcus showed significant differences after lyophilization or after freezing at −80°C. From a microbiological perspective, the results obtained define lyophilization as being as good an alternative for storing milk in the HMBs as freeze storage at −80°C because it does not increase the microbial burden of the product.

Stability of Bactericidal Capacity

The results obtained corroborate the bactericidal activity of human milk against E coli, as already evidenced by previous studies ^(6,30). In our study, the mean bactericidal capacity of the nonstored samples (expressed as percentage inhibition of microbial growth) was lower than the capacity reported for fresh milk in previous studies ^(6,30,31).

The differences between our results and those found in the literature can be explained by possible milk bioactive compound loss during freezing ^(13,30,32).

Milk samples in our study were kept frozen at −20°C between 1 and 6 weeks before first analysis, consistent with routine practices in the HMB, and were seen to retain bactericidal capacity—although to a lesser degree than in freshly collected breast milk. These data confirm the existence of bactericidal properties in milk stored in the HMB, despite the decrease in relation to fresh milk attributed to storage of the product in the home of the donors.

We have found no studies on the additive effects of successive treatments such as those commonly used in milk banks, initial freezing in the home (−20°C), followed by storage at either −20°C or −80°C. Likewise, we have found no studies in the literature comparing the effects of lyophilization and freezing on milk in HMBs.

Our results confirm the existence of bactericidal capacity in all the milk samples, with values that differ widely among donors. The E coli growth-inhibiting capacity remained stable regardless of the storage method used. Freezing at −80°C during the period of time studied did not affect this capacity, and lyophilization of the milk likewise produced no changes when compared with either freezing at −80°C or the reference values.

Factors Influencing the Characteristics of Milk

Variability is characteristic of human milk because its components and properties are influenced by a broad range of factors. In-depth studies have been performed of the effect of the time postpartum upon its nutrient contents, although the variations in biocomponents and properties are not always known. In our study, we found both time postpartum and gestational age to exert a significant influence upon the bactericidal capacity of milk. Specifically, the bactericidal properties were less apparent in the first week of lactation and increased in mature milk. This finding is surprising because one of the most valued characteristics of colostrum is its contribution to the defenses of the newborn infant. Some authors have already analyzed this effect, and the results are not always consistent. Bertino et al ⁽¹²⁾ observed no differences in the bactericidal capacity of human milk against E coli during the first 15 days of nursing, whereas Ogundele ⁽³⁰⁾ reported a marked decrease in bactericidal activity during the period from colostrum to transitional milk. Other studies have analyzed the changes in certain components that contribute to the global effect with the observation of different trends ^(33–36).

The results confirm that the antibacterial capacity of milk against E coli is less pronounced in milk from women with preterm infants than in those with infants born at term. We have found no studies in the literature with which to establish comparisons in this respect, although some authors have examined the effect that pregnancy may have upon certain concrete components such as proteins, lipids, enzymes, or α-tocopherol ^(37,38). No studies have examined the overall effect, however.

Some of the bioactive components of human milk have been identified. Their chemical nature varies and includes lipids, proteins, and oligosaccharides, which are potentially active in protecting the newborn infant from infections ⁽³⁰⁾. The direct relation between each of these components and the global antibacterial protection afforded by breast milk, however, remains unclear.

To examine this relation, we analyzed the concentration of macronutrients in the samples. The results revealed no significant relation between the bactericidal capacity of milk and its protein, total fat, or lactose contents. The recorded values of these nutrients were consistent with the data found in the literature, and showed important variability among samples, possibly as a consequence of the heterogeneity of the latter.

Gangliosides and sialic acid derivatives are active compounds found in human milk that intervene in many biological processes and in the development and maturation of the immune system ^(39–42). The concentrations of the different ganglioside and sialic acid species in human milk vary greatly among different milk samples, and the factors underlying such variability have not been clearly established ^(43–46). The mechanisms of action of these compounds suggest that they may contribute to the antibacterial capacity of milk.

Regression analysis showed bactericidal effects to be directly correlated to the total amount of gangliosides present in human milk (P = 0.039, r = 0.896)—no individual ganglioside being shown to significantly contribute to this effect. To our knowledge, no studies have analyzed the relation between the bactericidal effect of human milk and the composition of the ganglioside fraction. As a result, it is not possible to compare our results with those of other authors. The bactericidal effect of bovine milk sphingolipids, which share part of the structure of gangliosides, has, however, been investigated. Sprong et al ⁽⁴⁷⁾ observed an inhibitory effect (on bacterial growth?) of sphingolipids present in bovine milk, suggesting that gangliosides are involved in the bactericidal action observed with this fraction. Fatty acids forming part of the ganglioside ceramide moiety may confer their bactericidal activity, as has been observed in vitro ⁽⁴⁸⁾, the sugar chain not being involved in this effect. The mentioned fatty acids may in fact be the active components of the lipid fraction described by Ogundele ⁽³⁰⁾.

CONCLUSIONS

The present work has demonstrated that lyophilization can be considered as a promising alternative for storing human milk in HMBs instead of freezing. Lyophilization does not increase harmful microbial content, decrease bactericidal capacity, or modify the nutritional value of human milk. Lyophilization, however, possibly may have negative aspects (eg, inducing a decrease in known beneficial bacteria content), hence emphasizing on the need of future studies using more comprehensive methods (likely culture-independent) to evaluate the effect of these preservation techniques on a wider range of bacterial taxa. This work also demonstrates that the bactericidal capacity of human milk depends on the gestational age and time postpartum, and is related to total ganglioside content.

Acknowledgments

The authors thank A. Ramon and R. Sirvent for their kind help in collecting the samples.