Human milk (HM) is considered the “gold standard” nutrition for feeding term and pre-term newborns [1]. When mother’s own milk is not available or not sufficient, despite significant lactation support, Donor Human Milk (DM) is an important and the first alternative, especially in high risk newborns admitted to Neonatal Intensive Care Units (NICU) [2, 3].

The main benefit deriving from the use of DM versus formula milks, in preterm infant feeding, is the reduction in the incidence of necrotizing enterocolitis (NEC) as shown by several meta-analysis [4–7]. Moreover, an enhanced feeding tolerance has also been reported [3, 8].

The DM must be collected, processed and stored ensuring its microbiological safety maintaining at the same time the nutritional quality (NQ) [9]. NQ control of HM storage is complex as well as mandatory and pasteurization procedure is the main processing step to inactivate pathogenic microorganisms [2, 10].

DM is typically pasteurized by Holder Pasteurization method (HoP) characterized by heating at 62.5 °C for 30 min. HoP is currently the requested procedure by the majority of Human Milk Banking Associations (HMBA), being a temperature below 62,5 °C considered not safe [2, 10].

The impact of HoP on the biological quality of HM has been investigated and results are still controversial and matter of debate. The main explanations reside in the different procedures regarding samples collection, storage and treatment [11–13]. In this regard, data on the total protein count and specific proteins have been provided by using immune-enzymatic techniques such as ELISA. Recently, GeLC-MS, a proteomic approach involving the separation of proteins in SDS-PAGE by one-dimensional electrophoresis (1DE) followed by identification by mass spectrometry (MS), has been suggested as a promising technique able to offer a qualitative HM protein profile [14]. Data on GeLC-MS pattern in HM previously treated by HoP and later on stored at Human Milk Bank (HMB) are still lacking.

Therefore, the purpose of the present study was to determine the effect of HoP on the protein profile of HM by using GeLC-MS analysis, under reducing and non-reducing conditions, in order to investigate the qualitative HM protein profile and the occurrence of protein aggregates, eventually formed after heat treatment.

When mother’s own milk is not sufficient, donor milk is the first best choice for feeding term and preterm newborns, due to its well-recognized nutritional advantages with respect to formula milk [4–8]. DM should be obtained from established HMBs that follow specific guidelines for screening, storage, and handling procedures to optimize its composition while ensuring its safety for the recipient. [2, 10] All the milk arriving to the Human Milk Bank must be pasteurized. The ideal pasteurization process should consist of a phase of rapid heating followed by a phase of constant maintenance of the temperature and a final phase of rapid cooling. Pasteurization of the milk minimizes the risk of disease transmission via HM, inactivating most of the viral and bacterial contaminants. In addition, donors are screened in a similar way as for blood donation. No report has been published showing transfer of diseases through pasteurized DM, although milk may contain microorganisms. [2, 10]. Currently Holder Pasteurization (62.5 °C for 30 min) is the gold standard for milk processing in Human Milk Banks and is recommended by several guidelines as the optimal compromise between quality and microbiological safety [2, 10]. Several studies have already been performed to evaluate the effects of pasteurization on mother’s milk macronutrients; the related results are often discordant, especially concerning proteins [8, 20–22].

Our study observed a non-significant reduction of the total protein count following HoP of milk samples. This result is in line with the available literature: several studies reached similar conclusions using different analytical methods [20, 21]. Only Vieira et al. in 2011 found that there was a significant reduction in the mean protein concentrations, between the raw and post-pasteurization samples (reduction of 3.9 %). However, these samples were only analyzed with the use of a FT-IR infrared analyzer [22].

The original contribution of our study consists of the use of a semi-quantitative analytical method, GeLC-MS analysis. This technique allows to evaluate the protein profile of human milk, which is constituted by a complex array of biologically active proteins. Each sample was tested and compared under reducing and nonreducing conditions, to highlight the possible presence of protein-complexes due to disulfide bond formation in the pasteurization step. However this method is not able to evaluate the protein changes associated whit interaction between proteins and sugars, or proteins and lipid due to thermic treatments. No differences were observed between the electrophoretic profile of the same sample under reducing and nonreducing conditions, except for the region of high molecular weight, probably corresponding to the formation of high-mass complexes which do not run in the gel. Over 75 kDa, we observed more numerous and well separated bands under reducing conditions. The peculiarity in our data consists of the observation that no variation is present in 14 out of 20 samples (previous studies have detected a variation in all samples). It is also noteworthy that amongst the 6 samples presenting a modification in the protein levels, 5 derived from colostrum milk and only 1 from transitional milk. This finding is relevant from a clinical practice standpoint, since donor milk usually consists of mature milk (which, according to our data, did not show any variation).

Each band showing a variation was then identified with the corresponding protein(s). The bands of greatest interest are the ones of medium molecular weight (25–30 kDa) since they contain beta casein, alpha-casein, lactoferrin fragments and immunoglobulin light chains (Igκ and Igλ); the other band of interest (250 kDa) was identified as tenascin.

Tenascin is a homohexameric disulfide-linked glycoprotein and its expression decreases with lactation. This protein was investigated in recent studies for its antimicrobial properties in milk and Fouda et al. have hypothesized its ability to neutralize the HIV-1 virus via binding to the chemokine receptor site [23]. Tenascin has never been evaluated previously in human milk after pasteurization and we observed a reduction only in reduced samples of colostrum.

Beta-casein and alpha-casein are two important proteins of human milk. Only one previous study has investigated the effect of Holder pasteurization on these proteins using an electrophoretic analysis method and has shown a slight decrease in a single pooled sample [14].

IgA antibodies in breast milk provide passive immunity and sufficient protection against infections to neonates and preterm infants [24]. Lactoferrin is a bioactive protein that serves as an immunomodulatory factor [24]. Several studies revealed a number of functions of these molecules: i) IgA has specific antigen-targeted anti-infective action; ii) lactoferrin plays important roles in immunomodulation, iron chelation and antimicrobial action, and exhibits the anti-adhesive and trophic properties necessary for intestinal growth [24, 25]. Overall, regarding lactoferrin and IgA, our study is partially in agreement with the results reported in the literature, in which a decrease was found in all samples tested, in contrast to our findings. There might be two reasons for these differences: i) we did not use pooled milk (as done in several previous studies) but we have analyzed all milk samples individually; furthermore the same samples have been analyzed before and after pasteurization; ii) we have utilized a GeLC-MS analysis (semi-quantitative technique), instead of an immunohistochemical analytical techniques like the ELISA test. Additionally, there is a significant variability in previous studies concerning the retention of these proteins after HoP, which can be ascribed not only to heterogeneity in methods of human milk collection but also to treatment and storage before and after HoP [11–13].

In conclusion, donor milk needs a heat treatment for a safety storage. Holder Pasteurization is the method recommended by all international guidelines for the constitutions of Human Milk Banks. Our data are in agreement with literature showing a small decrease in the total protein content of human milk after HoP. Nonetheless, samples obtained from different donors reacted differently to heat treatment, even if processed and stored using the same conditions: in particular, 30 % of the samples showed some differences in the protein profile after HoP, whereas 70 % of the samples showed no detectable differences. The detectable effects on the protein profile were mainly in colostrum samples; this variation is quite interesting from a clinical standpoint, since donor milk is mostly constituted by mature milk. The reasons for this apparent variability is not clear and would deserve further investigation. In this setting, future studies should be designed to investigate whether these differences are also confirmed by other techniques able to assess the protein changes due to thermic treatments including the interaction between proteins and sugars, or proteins and lipid with possible toxic derivatives.