Growth factor distribution between cream and skim in human milk: Relevance for premature baby nutrition

M. Saffon1, A. Laubscher2, and R. Jiménez-Flores2 1Université Laval, Quebec City, Canada

2Dairy Products Technology Center California Polytechnic State University, San Luis Obispo

Abstract Breast-feeding gives developmental advantages to infants, as numerous investigators have reported in these meetings of Milk Genomics. Some include reduced risks of necrotizing enterocolitis, gastroenteritis, respiratory infection, and immunologically based diseases, in addition to improved cognitive development later in life. Necrotic enterocolitis (NEC) is a grave illness among preterm babies all over the world. Efforts to correct this illness involve feeding human milk to these babies, and in many cases this is done via human milk obtained through banks. Future development of human milk distribution with active growth factors may depend on development of non-thermal treatments for microbial preservation. Many of these treatments need the separation of cream and skim for maximum efficiency. A central question for processing design in these cases is the characterization of growth factor partition between cream and skim milk. In this study we measured concentrations in cream and skim of three major growth factors in bank human milk using ELISA: vascular endothelial growth factor (VEGF; the 165-amino acid form is the most abundant), hepatic growth factor (HGF; a 674-amino acid polypeptide), and epidermal growth factor (EGF; a 53-amino acid polypeptide). We used pooled bank milk from 350 donors (100 L), and used pilot plant scale cold separation (5-7 oC) to emulate possible large-scale production of human milk for hospital use. The resulting skim fraction contained only 0.07-0.10% fat, and the cream fraction 35-45 % fat. In addition, we measured beta-casein, lactalbumin, lactoferrin and xantin oxidorreductase (by SDS-PAGE) IgA, IgG and IgM (ELISA) as controls for our experiments.

Acknowledgments This work was funded by: Dairy Management Inc. California Dairy Research Foundation California State University Agricultural Research Initiative We would also like to thank: Dairy Products Technology Center students and staff

Reprints can be obtained from rjimenez@calpoly.edu

Introduction Separation of skim and cream fractions of human milk is an indispensable and necessary step in the process of milk microfiltration for sterilization and concentration purposes. Considering human milk, several issues present a challenge for a successful process. The natural viscosity of human milk is much higher than bovine milk (presumably due to the high concentration of oligosaccharides in the serum of human milk). Normally the most efficient conditions for separation of cream is to process whole milk in a centrifugal separator at a temperature of 50 to 60oC (130 to 140oF). However, in human milk there are many temperature sensitive proteins and growth factors that need to be protected from thermal denaturation. Hence our need to make a detailed description of separation efficiency in human milk that has been frozen and thawed, at low temperatures and using a separator designed for human milk. One of the many advantages of donor milk programs for neonate and pre-term babies is the richness of biologically active molecules in human milk. The study of bioactive components in human milk has been difficult because of its biochemical complexity, the small concentration of certain bioactive components, the distribution of some of these agents between fractions of milk (i.e. skim vs cream fraction, or serum vs casein micelles), the dynamic quantitative and qualitative changes of milk during lactation, and the lack of specific reagents to quantify these agents. However, a host of bioactive substances, including hormones, growth factors, and immunological factors such as cytokines, have been identified in human milk. These molecules are susceptible, in some cases, to temperature-induced conformation and activity changes. Therefore optimal quality of any product for infants from human milk needs to be closely monitored. Since the separation process is a necessary step for processing human milk for our ultimate objective of pre-term baby nutrition, we need to closely characterize how this molecules partition between cream and skim fractions. These fractions are necessary for further low temperature membrane-based processing.

Table 1: Composition Results

Conclusions In conclusion, pooled human milk was successfully separated cold at a pilot plant scale. • Human milk was heat stable at a pseudo-UHT treatment. • The confocal imaging shows noticeable differences in the structure and

size of human milk fat globules in comparison to bovine. • There were high levels of growth factors (HGF, EGF, and VEGF) present

in the skim fraction. • TLR4 is also present in the skim fraction, and at potentially higher

concentration than in the cream fraction.

Composition Using the Foss Milkoscan FT2, we determined that 78.12% total protein was recovered by skimming. A new separator will be used for future runs, which hopefully will improve the composition of the skimmed fraction.

Figure 1: Human Milk

Confocal Microscopy Fresh cream samples were obtained (no freezing or refrigeration) and images were obtained immediately (picture on right) and after one-week storage at 4°C (picture on left). Phosopholipids were stained with phosphatidylethanolamine-lissamine rhodamine B and imaged at 60X under oil immersion.

Methods and Results

Samples were denatured and reduced prior to loading onto a 12% SDS-PAGE gel.

A cream sample after one-week storage at 4C and imaged at 100X under oil immersion (picture on left). A bovine cream sample for comparison (imaged at 60X under oil immersion, picture on right).

Precision Plus Protein Standard (Bio-Rad Laboratories)

250 kDa

75 kDa

50 kDa

25 kDa

10 kDa

Human Milk Skim Cream

Heat Stability The stability of human milk was examined after a pseudo UHT treatment (3 minutes at 121°C) at time points: zero, one day, one week, two weeks, and three weeks. Sedimentation was obtained by centrifugation at 4000 rpm for 60 minutes.

Figure 3: Western blot analysis of TLR4 in pooled bovine and human milk fractions. Lane 1: human milk; Lane 2: bovine colostrum; Lane 3: human cream; Lane 4 and 5: human skim milk; Lane 6: bovine cream; Lane 7: bovine skim milk; Lane 8: bovine milk.

The results give good indication that the pooled human milk is heat stable after a psuedo UHT treatment. There is no change in pH or amount of supernatant after treatment and over time. There is a slight yellowing of the milk over time, and this observation we are currently examining with a colorimeter.

Table 3: Calculated ELISA results from standard curves.

Toll-Like Receptors The presence of Toll-like receptor 4 (TLR4), a phylogenetically conserved mediator of innate immunity essential for microbial recognition, is examined in fractions of pooled bovine and human milk.

1 2 3 4

Figure 2: Samples collected from pooled human milk samples, separated into cream and skim fractions at 4°C, and run on a 12% SDS-Page gel with visualized through Coomassie staining.

6 7 8 5

Precision Plus Protein Standard (Bio-Rad Laboratories)

250 kD

75 kD

50 kD

25 kD

10 kD

88 kDa

TLR4 is present in the human skim fraction and this confirms the genomic observation (Medrano 2008). It was also interesting to see that the band intensity appears stronger than in the human milk and human cream fraction. This is also beneficial for future separation and membrane filtration processing.

Growth Factor Analysis The concentrations of hepatic growth factor (HGF), epidermal growth factor (EGF), and vascular endothelial growth factor (VEGF) were mesaured in pooled bovine and human milk fractions using enzyme-linked immunosorbent assay (ELISA) from Invitrogen (Carlsbad, CA).

Table 2: Results

In this trial, the three growth factors we measured (HGF, EGF, VEGF) we found that all of them had higher levels in the skim fraction. This is beneficial information for future separation and membrane filtration processing. We think that in bovine milk, excess casein interferes with our assay.