Donor human milk is the best alternative when mothers’ own milk is not available for the preterm infant because it contains nutritional, bioactive, protective and developmental components that cannot be found in infant formula (Wight, Reference Wight2001). However, to prevent the potential risk for the transmission of pathogens from donor mothers to preterm infants, the donor milk is pasteurized.
Holder pasteurization is a low-temperature long-time (LTLT) heat preservation method that is widely used in milk banks. Human milk is heated in a water bath and held at 62·5°C for 30 min (Arnold & Tully, Reference Arnold and Tully1994; Balmer & Williams, Reference Balmer and Williams1995; Hartmann et al. Reference Hartmann, Pang, Keil, Hartmann and Simmer2007). This treatment is capable of a log10 5 reduction of bacteria including Escherichia coli, Staphylococcus epidermidis, Enterobacter cloacae, Bacillus cereus and Staphylococcus aureus (Czank et al. Reference Czank, Prime, Hartmann, Simmer and Hartmann2009). However, bioactive proteins are only partially preserved during this process. For example, 72% of sIgA, 22% of lactoferrin and 39% of lysozyme were retained after Holder pasteurization. Furthermore, the retention of bile salt stimulated lipase (BSSL) was reported to be less than 1% (Henderson et al. Reference Henderson, Fay and Hamosh1998; Tully et al. Reference Tully, Jones and Tully2001; Czank et al. Reference Czank, Prime, Hartmann, Simmer and Hartmann2009). BSSL transforms triacylglycerol into monoglycerides and free fatty acids which aids in digestion of fat for the infant (Hernell & Bläckberg, Reference Hernell and Bläckberg1994). BSSL is a very heat labile enzyme and depending on exposure time the inactivation starts at 45°C (Wardell et al. Reference Wardell, Wright, Bardsley and D'Souza1984). It is likely that the inactivation of BSSL through the Holder pasteurization process is responsible for the lower growth rates of preterm infants fed pasteurized donor human milk compared with preterm infants fed mothers’ own milk (Williamson et al. Reference Williamson, Finucane and Ellis1978; Andersson et al. Reference Andersson, Sävman, Bläckberg and Hernell2007).
Treatments with power-ultrasound (20–100 kHz) are an emerging technology for the preservation of food alone (D'Amico et al. Reference D'Amico, Silk, Wu and Guo2006; Chouliara et al. Reference Chouliara, Georgogianni, Kanellopoulou and Kontominas2010; Wang et al. Reference Wang, Hu and Wang2010) and in combination with other novel processing technologies (Walkling-Ribeiro et al. Reference Walkling-Ribeiro, Noci, Cronin, Lyng and Morgan2009a, Reference Walkling-Ribeiro, Noci, Riener, Cronin, Lyng and Morganb; Arroyo et al. Reference Arroyo, Cebrián, Pagán and Condón2011). Power-ultrasound creates cavitation, that is, the formation, growth and implosive collapse of bubbles in liquids. During the collapse of these bubbles localised hot spots occurs with temperatures of roughly 5000°C, pressures of about 50 MPa and a lifetime of a few microseconds (Suslick, Reference Suslick1990). The pressure changes resulting from these implosions create shock waves that disrupt the cellular membranes of bacteria resulting in cell lysis (Allison et al. Reference Allison, D'Emanuele, Eginton and Williams1996; Cameron et al. Reference Cameron, McMaster and Britz2008). The combined effect of ultrasound and heat on human milk has been found to inactivate Esch. coli and Staph. epidermidis with a greater retention of sIgA, lysozyme, lactoferrin and BSSL than with Holder pasteurization (Czank et al. Reference Czank, Simmer and Hartmann2010).
The aim of this study was to investigate the effect of different ultrasound settings on the elimination of Esch. coli and preservation of BSSL activity in human milk. In a first experiment different ultrasound-energies were created by holding ultrasound-power constant and varying the exposure times. In a second experiment ultrasound-energy was held constant with varying ultrasound-power and exposure times.
Methods and Materials
Sample collection
Frozen human milk samples were donated by 3 mothers with excess milk supply. The collection of the milk was approved by the Human Research Ethics Committee of the University of Western Australia (RA/4/1/2369). All donors gave written consent for their donations to be used in research and all samples were de-identified.
Sample preparation
Escherichia coli K12 (ATCC1498, Southern Biological, Nunawading, Vic, Australia) was cultured in nutrient broth (Nutrient Broth No. 2, Oxoid Australia Pty Ltd, Adelaide, SA, Australia) over night at 37°C and enumerated using the optical density method with previously constructed standard curves (Monod, Reference Monod1949) measured with a spectrophotometer at 600 nm (PowerWave, Biotek, Winooski, VT, USA). The culture was then diluted and inoculated into previously thawed and pooled (3 donors) human milk to target a maximum accepted concentration in milk banks of 105 CFU/ml (colony forming units per millilitre) (Hartmann et al. Reference Hartmann, Pang, Keil, Hartmann and Simmer2007) and portioned into 12 ml samples and brought to 20°C using a water bath (CC-1, Huber GmbH, Offenburg, Germany).
Ultrasonication
A 20 kHz ultrasonic processor fitted with a 6 mm stepped microtip (VCX 130, Sonics & Materials Inc., Newtown, CT, USA) for a batch volume between 10 and 50 ml with a maximum amplitude of 123 μm was used for all experiments. Human milk (12 ml) was treated in batch mode (10 samples for each treatment condition) using 25 ml polystyrene test tubes (inner-diameter 25 mm) with 12 mm gap between the bottom of the tube and the microtip surface (approximately half the height of human milk). Temperature was recorded using a type T thermocouple (MT-29/5, Physitemp Instruments Inc., Clifton, NJ, USA) and USB data-logger (THERMES USB, Physitemp Instruments Inc., Clifton, NJ, USA) with data acquisition software (DASYlab10·0, Measurement Computing Corp., Norton, MA, USA) at a sampling rate of 16 Hz (Fig. 1). After treatment, the tube with human milk was transferred to an ice bath. Aliquots where taken and kept frozen at −80°C until the BSSL assay was performed. Cooled samples were transferred for immediate microbiological analyses.
Ultrasonication set-up.
Constant ultrasound-power and varying exposure times
The ultrasound processor was set on maximal amplitude and generated an ultrasound-power of 26 W. Human milk was treated for different times (19·2, 38·4, 57·6, 76·8, 96 and 115·2 s) to achieve an applied ultrasound-energy (500,1000, 1500, 2000, 2500 and 3000 J), respectively.
Constant ultrasound-energy and varying ultrasound-power and exposure time
The same ultrasound-energy (1000 J) was applied with different time-power ratios (276 s at 3·62 W; 114·5 s at 8·73 W; 70·5 s at 14·18 W; 50·5 s at 19·8 W and 39 s at 25·64 W).
Bacteria analysis
Cooled untreated and ultrasonicated samples (5 μl) were plated in duplicates onto nutrient agar (Plain agar, Southern Biological, Nunawading, Vic, Australia/Nutrient Broth No. 2, Oxoid Australia Pty Ltd, Adelaide, SA, Australia) and incubated at 37°C for 18 h. The numbers of colonies were then enumerated to determine the reduction in bacterial number following the ultrasound processing based on CFU/ml. Detection limit was at 200 CFU/ml.
BSSL assay
The BSSL activity in human milk was determined using a lipase assay kit (QuantiChrom Lipase Assay Kit, BioAssay Systems, Hayward, CA, USA). The frozen (−80°C) untreated and ultrasonicated samples were thawed and brought to 37 °C, diluted (1/100, v/v) with double deionized water and the assay was performed in duplicate according to kit instructions. The absorbance was determined with a spectrophotometer at 412 nm (PowerWave, Biotek, Winooski, VT, USA) and transformed into BSSL activity per litre (U/l) based on the formula in the kit instruction. The kit linear detection range is between 40 and 1600 U/l lipase activity in the 96-well plate assay. Unit definition: one unit of enzyme catalyses the cleavage of 1·0 μmole of substrate per min under the assay conditions (pH 8·5). The coefficient of variation (CV) within groups was <6% unless otherwise stated.
Statistical analysis
Results are presented as treatment means±sd values unless otherwise stated. All analyses were carried out using R 2.9.2 for Mac OS X (R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, 2009). ANOVA models with Tukey's HSD Post Hoc test were used to compare the values between groups (statistical significant when P<0·05, confidence level 95%). Graphical modelling was performed with OpenOffice 3.2.0 Spreadsheet (Sun Microsystems Inc. based on OpenOffice.org)
Results
Effect of holding ultrasound-power constant and varying the exposure times
A reduction of Esch. coli viability (CFU/ml) over time was observed when human milk was treated using ultrasound-power of 26 W. Untreated samples (0·0 s) inoculated with Esch. coli gave a total count of 221 600±48 157 CFU/ml (n=10). The temperature before treatment was 20·0±0·4°C. After 76·8, 96·0, and 115·2 s of treatment the Esch. coli viability was reduced to 42 160±13 598, 1611±1334 and undetectable CFU/ml (P<0·05) (n=10) with the temperature of the human milk increasing up to 51·4±0·3, 58·5±0·5, and 64·9±1·2°C, respectively (P<0·05) (n=10). There was a significant exponential regression (R 2 of 0·976) for bacterial count from untreated samples up to an exposure time of 76·8 s (Fig. 2).
Effect of holding ultrasound-power constant (26 W) and varying the exposure times on Esch. coli count (o, fine dashed line) and exponential regression for Esch. coli count between 0·0 and 76·8 s (solid line), and processing temperature (●) and its linear regression trend (long dashed line). All values are mean±sd (n=10).
Untreated human milk had a BSSL activity of 46 253±1969 U/l (n=6). The BSSL activity was significantly reduced to 93% between 0·0 and 19·2 s (P<0·05) but then remained relatively constant between 19·2 and 76·8 s (P>0·9) before decreasing significantly from 91 to 76% (P<0·05) between 76·8 and 96·0 s and to 8% at 115·2 s (P<0·05). Over this latter period the BSSL activity decreased from 42 052±1365 U/l at 76·8 s and 34 962±1268 U/l at 96·0 s (CV <6%) to only 3858±1921 U/l (CV <50%) at 115·2 s (n=6) (Fig. 3).
Effect of holding ultrasound-power constant (26 W) and varying the exposure times on BSSL activity (o, solid line) and processing temperature (●) and its linear regression trend (long dashed line). All values are mean±sd (n=6).
Effect of holding ultrasound-energy constant and varying the ultrasound-power and exposure time
Untreated human milk inoculated with Esch. coli gave a total bacteria concentration of 204 760±38 770 CFU/ml (n=10). Human milk treated with constant ultrasound-energy (1000 J) but varying ultrasound-power and exposure time showed a significant reduction of Esch. coli cell concentration compared with untreated human milk (P<0·05). Furthermore, a significant reduction in the Esch. coli load was found between the human milk treated with the longest exposure time and lowest power (276 s at 3·62 W; 7700±3731 CFU/ml) and all the other shorter but more powerful ultrasound settings (P<0·05) (Fig. 4). However, there was no significant difference in Esch. coli elimination between the treatments 114·5 s at 8·73 W, 70·5 s at 14·18 W, 50·5 s at 19·8 W and 39 s at 25·64 W (P>0·1).
Effect of holding ultrasound-energy constant (1000 J) and varying the ultrasound-power and exposure time. Esch. coli count (grey) (n=10), BSSL activity (white) (n=7) and final-temperature (black) (n=10). All values are mean±sd.
The temperature before ultrasonication was at 20·0±0·2 °C. The temperature after treatment was highest for the treatment with the longest time and lowest power. The temperature then progressively and significantly decreased at each time by power change as exposure time decreased from 276 to 39 s (P<0·05). Thus, the highest temperature of 53·7±0·5°C was observed at 276 s at 3·62 W and the lowest temperature of 37·6±0·3°C was observed at 39 s at 25·64 W (n=10) (Fig. 4).
Untreated human milk had a BSSL activity of 53 408±2590 U/l (n=7) and was only significantly reduced by the longest treatment time at the lowest power (276 s at 3·62 W) with the activity decreasing to 48502±2094 U/l (n=7), that is, to 91% of the untreated human milk BSSL activity (P<0·05) (Fig. 4).
Discussion
Effect of holding ultrasound-power constant and varying the exposure times
A previous study reported a synergistic effect when thermal and ultrasound treatments were used in combination (Czank et al. Reference Czank, Simmer and Hartmann2010). The time to reduce the bacterial content by log10 5 was significantly reduced when applied to human milk that had a starting temperature of either 45 or 50°C. However, ultrasound treatment of 60 ml human milk resulted in a consistent increase of 10·3°C from the starting temperature which stabilised within three minutes of the application of ultrasound. In our study the application of ultrasound to 12 ml human milk with a starting temperature of 20°C resulted in a linear increase in temperature over a period of 115·2 s (Fig. 2). We found an exponential decrease in the Esch. coli cell concentration up to 51·4°C (76·8 s exposure to ultrasound) and followed by a rapid decrease (log10 1·3) in viability between 51·4 and 58·5°C (96 s). It is of interest that the Esch. coli cells, when exposed to heat only, decreased when the temperature was increased to 56°C (Wills et al. Reference Wills, Han, Harris and Baum1982; Czank et al. Reference Czank, Prime, Hartmann, Simmer and Hartmann2009). A possible explanation for this trend could be that the rapid decrease in Esch. coli occurring as the temperature increased from 51·4 to 58·5°C was related to a synergistic effect of ultrasound and temperature (Fig. 2), because the application of temperature alone using Holder pasteurization at 57°C required a holding time of 16·4 min (Czank et al. Reference Czank, Prime, Hartmann, Simmer and Hartmann2009) to obtain a similar log10 decrease in Esch. coli to that observed as the temperature increase from 51·4 to 58·5°C (Fig. 2).
In contrast to the inactivation of the coliform bacterium, the loss in activity of BSSL occurred in two stages; from 100 to 93% BSSL activity between 20 and 28°C and then from 91 to 8% between 51·4 and 64·9°C (Fig. 3). Since it has been reported that BSSL is heat stable up to 45°C (Wardell et al. Reference Wardell, Wright, Bardsley and D'Souza1984), the initial significant but small decrease in BSSL activity may have been due to the exposure to ultrasound. However, the BSSL activity remained constant from 28 and 51·4°C, and this suggests that if the temperature of the human milk is maintained below 51·4°C the length of exposure to ultrasound would have a small effect on BSSL activity (Fig. 3). The high CV of 50% for BSSL activity at 115·2 s can be explained by the proximity to the linear detection range limit for this enzyme activity assay kit.
The large discrepancy between the exponential loss of Esch. coli and the relative stability of BSSL activity between 20 and 51·4°C suggests that initially ultrasound exposure caused cell death by disrupting the cell membranes of Esch. coli resulting in cell lysis (Allison et al. Reference Allison, D'Emanuele, Eginton and Williams1996; Cameron et al. Reference Cameron, McMaster and Britz2008). Subsequently, between 51·4 and 58·5°C there was a significant decrease in both BSSL activity and Esch. coli count suggesting that this resulted from the combined effect of ultrasound and temperature causing the denaturation of both BSSL and functional proteins present in Esch. coli.
Effect of holding ultrasound-energy constant and varying the ultrasound-power and exposure time
At constant ultrasound-energy (1000 J), the increase in temperature was highest when the human milk was treated for the longest time at the lowest ultrasound-power (276 s at 3·62 W) and lowest when treated for the shortest time at the highest power (39 s at 25·64 W) (Fig. 4). This finding demonstrates that the longer exposure time, rather than high ultrasound power, was responsible for the increase in temperature. Although, treatment at the longest time at the lowest power was the most effective in reducing the viability of Esch. coli, this longer exposure time resulted in a temperature that was approaching those associated with protein denaturation and compromised retention of the BSSL activity (Figs. 3 & 4). Furthermore, the BSSL activity was reduced at this setting (276 s at 3·62 W) but only to 91% of the untreated activity (Fig. 4). These findings again focus on a critical temperature being important in relation to decreasing both Esch. coli cell numbers and BSSL activity.
Extrapolation of the exponential regression for the reduction in Esch. coli from 0·0 s to 76·8 s, that is, during the time that ultrasound alone appeared to reduce bacteria (Fig. 2), predicts that there would be a log10 5 reduction in Esch. coli viability at 530 s. By 530 s with ultrasound power set at 26W, it can be calculated that the human milk would be subjected to 13·8 kJ of energy. With constant ultrasound-power it was found that the increase in temperature was linear up to 115·2 s (64·9°C) (Fig. 2). Furthermore when the ultrasound-energy was held constant the heating effect on human milk was the lowest when it was treated with high ultrasound-power and short exposure times (Fig. 4). Therefore, these findings predict that the loss of activity of BSSL could be minimised and the viability of Esch. coli could be reduced by log10 5 by applying 13·8 kJ of energy using high ultrasound power over a short exposure time to ensure that the temperature remains below the critical level for protein denaturation. Alternatively, the use of lower power settings such as the 26 W used in the present studies would require a cooling system to ensure the human milk BSSL activity was protected against temperature denaturation.
Although this study has only investigated the effects of ultrasonication on BSSL and Esch. coli, previous studies using either the LTLT Holder pasteurization or ultrasound processing methods have shown that other important proteins in human milk (sIgA, lactoferrin and lysozyme) were more resistant than BSSL to both forms of treatment types (Tully et al. Reference Tully, Jones and Tully2001; Czank et al. Reference Czank, Simmer and Hartmann2010).
On the other hand, rod-shaped bacteria such as Esch. coli were reported to have less resistance to ultrasound cavitation than coccoid bacteria (Zenker et al. Reference Zenker, Heinz and Knorr2003). Therefore, it could be expected that a longer exposure time will be required for coccoid bacteria, such as Staph. epidermidis and Staph. aureus, that are also commonly found in human milk. In this connection, Czank et al. (Reference Czank, Simmer and Hartmann2010) found that Esch. coli in human milk required 5·94 min exposure to ultrasound to achieve a log10 1 reduction in bacteria load, whereas, Staph. epidermidis required 16·01 min for a log10 1 reduction under similar conditions (Czank et al. Reference Czank, Simmer and Hartmann2010). Although, ultrasonication could be an alternative method to the thermal pasteurization of human milk, it would be necessary to ensure that the temperature of the human milk during ultrasonication remains below the level that causes a precipitous fall in the activity of BSSL (a marker enzyme for temperature denaturation of bioactive proteins). However a longer exposure time to ultrasound would be required to enhance possible synergistic treatment effects between temperature and ultrasound.
We would like to thank all the mothers who kindly donated human milk to make this research possible. Funding for this study was provided by an unrestricted research grant from Medela AG, Baar, Switzerland and a scholarship from Carag AG, Baar, Switzerland.
Ethic approval
Human Research Ethics Committee of the University of Western Australia (RA/4/1/2369).
Statement of financial support
Funding for this study was provided by an unrestricted research grant from Medela AG, Baar, Switzerland.
Lukas Christen receives an industry funded scholarship provided by Carag AG, Baar, Switzerland.
