Sample collection
We conducted a pretest–test design, where HM samples acted as their own controls. Breast milk was collected, at different stage of maturation (colostrum: n = 9; transitional milk: n = 5; mature milk: n = 6) according to Playford et al. [15], from 20 mothers delivered between 23 and 41 weeks of gestational age (GA). The study was approved by local ethic committee and mother gave informed and signed consent to the study.
Exclusion criteria were: maternal infections, tobacco smokers, drugs addiction and alcoholic; use of drugs or pharmacologically active substances; mothers who received blood transfusions or blood products, or organ transplants; fetal malformations, chromosomal abnormalities, perinatal asphyxia and dystocia.
HM samples were collected at two consecutive mornings, between 8 and 9 a.m., into disposable high density polyethylene sealed bottles (Flormed, Napoli, Italy) sterilized by using ethylene oxide. Milk expression was obtained by emptying one or two breasts with an electric breast pump (Medela Symphony). From each container, 10 mL of HM were taken, divided into two fractions: the first was immediately frozen at − 80 °C; the second was pasteurized in HMB and frozen at − 80 °C. HoP was performed with a Sterifeed Pasteuriser by Medicare Colgate Ltd (Cullompton, England), heating milk at 62.5 °C for 30 min. The last HoP phase requires a rapid and precise cooling of milk samples to 10 °C in approximately 20 min, by immersion into cold water.
Sample preparation and protein quantification
Skimmed HM samples were obtained by centrifugation at 2000 × g for 30 min at 10 °C, the pellet and the floating layer were discarded. Protein content was estimated according to Bradford [16].
GeLC-MS analysis and protein identification
Skimmed milk samples were mixed with Laemmli buffer (2 % w/v SDS, 10 % Glycerol, 5 % β-mercaptoethanol, 62 mM Tris-HCl pH 6.8), boiled for 5 min, and loaded on 10 × 8 cm vertical 12 % polyacrylamide gels. For non-reducing conditions the Laemmli buffer did not contain β-mercaptoethanol and samples were not boiled.
SDS-PAGE was performed at 10 mA per gel for 30 min and 30 mA per gel until the tracking dye front reached the bottom of the gel, at 10 °C with a Mini Protean II Xi System (Bio-Rad). The running buffer was 25 mM Tris-HCl, 200 mM Glycine, 0.1 % w/v SDS. The gels were stained overnight with Colloidal Coomassie brilliant blue G250 (Bio-Rad Laboratories) in accordance with Neuhoff et al. [17]. The Coomassie-stained gels were scanned using an Image Scanner III (GE Healthcare) at 300 dpi. The protein bands of interest were manually excised from 1DE gels and in-gel digested with trypsin as described by Spertino et al. [18]. The peptide mixtures were pooled and lyophilized in a SpeedVac for mass spectrometry analysis.
MS/MS analysis was performed using a QSTAR XL hybrid quadrupole-TOF instrument (Applied Biosystems, Foster City, CA, USA) coupled with a LC Packings Ultimate 3000 nano-flow LC system (Dionex, Amsterdam, The Netherlands), as described by Bona et al. [19]. Briefly, the QSTAR XL operated in positive mode and in information-dependent acquisition (IDA) mode, the dynamic exclusion feature of the Analyst QS 1.1 software (Applied Biosystems, Foster City, CA, USA) was enabled, with an exclusion mass width of +/−3 m/z for 60 s. LC/MS-MS files obtained from each protein sample were merged into a single MASCOT generic format (mgf) file and searched against the NCBI non-redundant database; tolerance for precursor and fragment masses was 0.25 Da. The proteins were identified in homology with significant ion scores (p < 0.05).
Statistical analysis
Clinical data are reported as the mean and SD. Protein content (mg/mL) are reported as median and interquartile ranges. Statistical analysis was performed using XLStat-Pro v.7.2.5 (Addinsoft, New York, USA). Results were compared between groups by Mann-Whitney U-two sided test when the data did not follow a Gaussian distribution. A value of P < 0.05 was considered significant.