The Journal of Nutrition

Nutrient Physiology, Metabolism, and Nutrient-Nutrient Interactions

Nutrient Fortification of Human Donor Milk

Affects Intestinal Function and Protein
Metabolism in Preterm Pigs
Jing Sun,1 Yanqi Li,1 Duc Ninh Nguyen,1 Martin S Mortensen,2 Chris HP van den Akker,5 Tom Skeath,6
Susanne E Pors,3 Stanislava Pankratova,1,4,7 Silvia Rudloff,8 Søren J Sørensen,2 Douglas G Burrin,9
Thomas Thymann,1 and Per T Sangild1,7

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

1
Section of Comparative Pediatrics and Nutrition, Faculty of Health and Medical Sciences; 2 Section of Microbiology, Faculty of Science;
3
Department of Veterinary and Animal Sciences; and 4 Laboratory of Neural Plasticity, Center for Neuroscience, University of Copenhagen,
Copenhagen, Denmark; 5 Department of Pediatrics, AMC, University of Amsterdam, Amsterdam, The Netherlands; 6 Newcastle Neonatal
Service, Royal Victoria Infirmary, Newcastle upon Tyne, United Kingdom; 7 Department of Pediatrics and Adolescent Medicine,
Rigshospitalet, Copenhagen, Denmark; 8 Institute of Nutritional Science, Justus-Liebig-University Giessen, Giessen, Germany; and
9
USDA/ARS Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX

Abstract
Background: Nutrient fortification of human milk is often required to secure adequate growth and organ development
for very preterm infants. There is concern that formula-based fortifiers (FFs) induce intestinal dysfunction, feeding intol-
erance, and necrotizing enterocolitis (NEC). Bovine colostrum (BC) may be an alternative nutrient fortifier, considering
its high content of protein and milk bioactive factors.
Objective: We investigated whether BC was superior to an FF product based on processed bovine milk and vegetable
oil to fortify donor human milk (DHM) for preterm pigs, used as a model for infants.
Methods: Sixty preterm pigs from 4 sows (Danish Landrace × Large White × Duroc, birth weight 944 ± 29 g) re-
ceived decreasing volumes of parenteral nutrition (96–72 mL  kg−1  d−1 ) and increasing volumes of enteral nutrition
(24–132 mL  kg−1  d−1 ) for 8 d. Pigs were fed donor porcine milk (DPM) and DHM with or without FF or BC fortification
(+4.6 g protein  kg−1  d−1 ).
Results: DPM-fed pigs showed higher growth (10-fold), protein synthesis (+15–30%), villus heights, lactase and pepti-
dase activities (+30%), and reduced intestinal cytokines (–50%) relative to DHM pigs (all P < 0.05). Fortification increased
protein synthesis (+20–30%), but with higher weight gain and lower urea and cortisol concentrations for DHM+BC
compared with DHM+FF pigs (2- to 3-fold differences, all P ≤ 0.06). DHM+FF pigs showed more diarrhea and re-
duced lactase and peptidase activities, hexose uptake, and villus heights relative to DHM+BC or DHM pigs (30–90%
differences, P < 0.05). Fortification did not affect NEC incidence but DHM+BC pigs had lower colonic interleukin (IL)-6
and IL-8 concentrations relative to the remaining pigs (–30%, P = 0.06). DHM+FF pigs had higher stomach bacterial
load than did DHM, and higher bacterial density along intestinal villi than did DHM and DHM+BC pigs (2- to 3-fold,
P < 0.05).
Conclusions: The FF product investigated in this study reduced growth, intestinal function, and protein utilization in
DHM-fed preterm pigs, relative to BC as fortifier. The relevance of BC as an alternative nutrient fortifier for preterm
infants should be tested. J Nutr 2018;148:336–347.

Keywords: formula, bovine colostrum, intestine, growth, preterm infant

enterocolitis (NEC), and sepsis (7, 8). To prevent growth fail-

Introduction
The survival rates of very premature infants are increasing but
many of these infants experience extrauterine growth restric-
tion. This is associated with long-term complications, such as
impaired neurologic development and later metabolic disease
(1–6). The factors contributing to poor growth include subopti-
mal nutrition, coupled with immature organ functions and
comorbidities such as respiratory distress, necrotizing
ure, it is important to provide optimal and adequate nutrients
and energy for preterm infants, often defined as an intake that
supports a growth rate similar to that in utero at the same post-
conceptual age (9). When the gut is immature, this growth rate
may be achieved by increasing nutrient intake via parenteral
nutrition (PN) (10), but PN is associated with higher risks of
catheter-related sepsis, liver disease, and impaired gut barrier
function (11, 12). Thus, early and gradual transition to enteral

© 2018 American Society for Nutrition. All rights reserved.

336 Manuscript received August 6, 2017. Initial review completed August 24, 2017. Revision accepted October 31, 2017.
First published online March 12, 2018; doi: https://doi.org/10.1093/jn/nxx033.
nutrition (EN) with mother’s own milk, or alternatively donor
human milk (DHM), is recommended (13).
Since human milk contains inadequate amounts of nutrients,
especially protein and minerals, to support optimal growth rates
of very premature infants, it is common to add nutrient for-
tifiers to human milk from 1–2 wk of age (9, 14). Early nu-
trient fortification increases protein intake for preterm infants
(15), but there is concern that early exposure to formula-based
products made from processed bovine milk and vegetable in-
gredients may increase the risks of gut inflammation, feeding
intolerance, and NEC (16–18). The evidence for risks associ-
ated with formula-based fortifier (FF) is weak (19), but it is
possible that bovine milk products and their processing proce-
dures (e.g., hydrolysis, filtration, heat treatment), and vegetable
ingredients (e.g., maltodextrin, oils) in FFs reduce the bioactiv-
ity of human milk after fortification (20–22). Milk fortifiers,
based exclusively on human milk, may be used in the future
(17), but their global availability is limited by high cost and
inadequate supply (23). It is therefore important to look for al-
ternative ways to provide very preterm infants with enough nu-
trients and protective milk factors for optimal growth, without
inducing additional risks.
Intact bovine colostrum (BC), the first milk after parturition
in cows, contains much higher levels of protein and bioactive
components than do processed FF and DHM (24). In preterm
pigs, feeding BC as the sole diet after preterm birth was equiv-
alent to DHM, and superior to infant formula, in preventing
NEC, intestinal dysfunction, and growth restriction (25–28).
Further, BC was superior to DHM in increasing weight gain
and intestinal pathogen resistance in preterm pigs (26). Given
these results, we have investigated the potential of BC as a sup-
plement to the first human milk feeding in preterm infants, and
these pilot studies show that BC supplementation was feasible
and not associated with adverse effects (29, 30).
With this in mind, we hypothesized that fortification of
DHM with FF would be inferior to BC in stimulating growth
and gut maturation. We tested this in preterm pigs because they
are very sensitive to feeding-induced gut and metabolic com-
plications and to inferior nutrient intakes and milk diets. We
first compared DHM-fed pigs with a group receiving donor
porcine milk (DPM), to establish the effect of feeding low-
protein human milk to preterm pigs. We then tested how nu-
trient fortification of DHM with BC or an FF product based on
Supported by the NEOMUNE (0603-00774B) and NEOCOL projects (6150-
00004B) from the Danish Strategic Research Council and Innovation Foundation
Denmark, and federal funds from the USDA, Agricultural Research Service un-
der Cooperative Agreement Number 58-6250-6-001 (DGB). JS was the recipient
of a PhD scholarship from the China Scholarship Council.
Author disclosures: The University of Copenhagen has filed a patent on the use
of bovine colostrum for human preterm infants. PTS is listed as sole inventor, but
has declined any share of potential revenue arising from commercial exploitation
of such a patent. JS, YL, DNN, MSM, CHPvdA, TS, SEP, SP, SR, SJS, DGB, and
TT, no conflicts of interest.
Supplemental Methods, Supplemental Figures 1–3, Supplemental Tables 1–3,
and Supplemental References are available from the ‘‘Supplementary data’’ link
in the online posting of the article and from the same link in the online table of
contents at https://academic.oup.com/jn/.
Address correspondence to PTS (e-mail: pts@sund.ku.dk).
Abbreviations used: AA, amino acid; ApA, aminopeptidase A; ApN, aminopep-
tidase N; BC, bovine colostrum; BUN, blood urea; CK, creatine kinase; CRP,
C-reactive protein; CSF, cerebrospinal fluid; DHM, donor human milk; DHM+BC,
donor human milk fortified with bovine colostrum; DHM+FF, donor human milk
fortified with a formula-based fortifier; Dist, distal small intestine; DPM, donor
porcine milk; DPPIV, dipeptidylpeptidase IV; FF, formula-based fortifier; FSR, frac-
tional synthesis rate; Mid, middle small intestine; NEC, necrotizing enterocolitis;
Prox, proximal small intestine.
commercial products affected growth, tissue protein synthesis,
clinical variables and intestinal structure, function, microbiol-
ogy, and immunology.
Methods
Donor milk and nutrient fortifier products. DPM was collected
from lactating sows in week 4 of lactation and pasteurized in a man-
ner similar to the DHM (62.5°C for 30 min, Holder pasteurization).
Holder-pasteurized DHM, collected 8–16 wks postpartum, was do-
nated by the Dutch Donor Milk Bank in Amsterdam. The macronu-
trient composition of DHM and DPM was determined by MilkoScan
(Foss, Hilleroed, Denmark). Bovine colostrum powder (ColoDan) was
donated by Damino Gesten, Biofiber, Denmark. Powdered Enfamil hu-
man milk fortifier (Mead Johnson, Evansville, IN) and Nutrilon Nena-
Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
tal protein fortifier (Danone, Utrecht, the Netherlands) were purchased
commercially. The nutrient composition of the donor milks and the for-
tifiers are listed in Supplemental Table 1.
Animals. All animal procedures were approved by the Danish Na-
tional Committee on Animal Experimentation. Sixty preterm pigs
were delivered from 4 sows by cesarean delivery at 106 d of
gestation (Danish Landrace × Large White × Duroc, term =
116 ± 2 d, birth weight 944 ± 29 g, male:female ratio 27:33, Askely-
gaard Farm), as previously described (27). Pigs were placed in individual
incubators with regulated temperature (37–38°C) and oxygen supply
(0.5–2.0 L/min, within the first 24 h). While pigs were sedated by the
maternal anesthetics, an umbilical catheter for PN and sow’s plasma
infusion and an orogastric feeding tube for EN were implanted (31).
Experimental design. Within 4 h after birth, all pigs began to re-
ceive daily decreasing volumes of PN [96–60 mL  kg−1  d−1 , com-
position as described previously (32)], and increasing volumes of EN
via orogastric feeding tube every 3 h (24–132 mL  kg−1  d−1 ) us-
ing their respective diets. Nine pigs were randomly allocated to receive
DPM until the end of study for comparison to the DHM group. The
remaining pigs were kept on DHM until day 3, when enteral feeding
reached 64 mL  kg−1  d−1 ; they were then stratified into 3 groups:
pigs continuing to receive unfortified DHM (n = 16), DHM fortified
with BC (DHM+BC, n = 19), and DHM fortified with the formula-
based fortifier (DHM+FF, n = 16). Gradual transition to fortification
was done by providing half-dose fortification for 24 h, then full-dose
until the end of the study. To reach a final protein concentration of
4.5 g/100 mL (close to that of DPM, 5–6 g/100 mL) for both fortifi-
cation groups, 7.0 g of BC powder was dissolved in 100 mL of DHM
for the DHM+BC group, and 5.7 g of Enfamil plus 1.5 g of Nu-
trilon powder were dissolved in 100 mL of DHM for the DHM+FF
group (for nutrient composition of products, see Supplemental
Table 1). The osmolality of the respective diets was measured with a
cryoscopic osmometer (Osmomat, Gonotec). Diets were prepared daily
and color-marked to blind the researchers to the interventions. The
macronutrient intake for each group via PN and EN are described in
Supplemental Table 2.
Clinical evaluation and tissue collection. Weight of pigs was
recorded daily and the dose of PN and EN adjusted to be comparable
to the feeding advancement rates used for moderately preterm infants
(25). Clinical symptoms and fecal consistency were evaluated 2 times/d
with the use of previously published scoring systems (33). The physical
activity of each pig was recorded when fortification was introduced,
with the use of infrared video surveillance cameras placed over each
incubator to analyze the proportion of active time (34).
Pigs were euthanized by intracardiac injection of sodium pentobar-
bital (60 mg/kg) if severe clinical symptoms appeared during the ex-
periment, or were killed at the end of day 8. In this study, 2 DPM, 5
DHM, 6 DHM+BC, and 4 DHM+FF pigs were euthanized 1–2 d be-
fore the end of the study because of dehydration, clinical signs of sepsis,
NEC, respiratory distress, or suspected cardiovascular problems. There
Human milk fortification in preterm neonates 337
were no statistical differences in total lifetime or mortality across the 4
groups and samples obtained from these pigs were included in the over-
all analyses. When the pigs were killed, blood samples were collected by
intracardiac puncture and organs were dissected, weighed, and scored
for NEC lesions, as described previously (35). Pigs with a score ≥3 in
any intestinal or colon region were diagnosed with NEC, and pigs with
a score of ≥4 as severe NEC. Blood samples and tissues from the prox-
imal, middle, and distal small intestine (Prox, Mid, and Dist, respec-
tively), and the colon region were snap frozen and stored at –80°C or
fixed in paraformaldehyde until further analysis.
In vivo intestinal function, brush border enzyme ac-
tivity, mucosal structure, and cytokine expression. In-
testinal hexose absorptive capacity was measured by detecting
plasma galactose concentration 20 min after an oral bolus of
15 mL/kg of 5% galactose solution on day 5 (35). Intestinal perme-
ability was tested before the pigs were killed by measuring urinary lac-
tulose:mannitol ratio at 3 h after an oral bolus (35). Gastric residual
was assessed by measuring postmortem stomach content weight after a
final feeding bolus (16.5 mL  kg−1  d−1 ) 60 min prior to killing. Gas-
tric pH was measured in 3-fold water-diluted stomach content with the
use of a pH meter [pH = log10 (10–pH × dilution factor)]. The activities
of brush border enzymes [lactase, maltase, sucrase, aminopeptidase A
(ApA), aminopeptidase N (ApN), dipeptidylpeptidase IV (DPPIV)] were
measured in tissue homogenates, as described earlier (35). Cytokines, in-
cluding IL-1β, IL-6, and IL-8, were measured in Dist and colon tissue
homogenates with the use of porcine DuoSet ELISA kits (R&D Sys-
tems), and adjusted for total protein concentration (Pierce BCA Protein
Assay Kit, Thermo Fisher Scientific). Mucosal weight proportion in the
small intestine was quantified by weighing the tissue before and after
scraping off the mucosa. Villus height and crypt depth were measured
on paraformaldehyde-fixed, hematoxylin and eosin-stained Prox, Mid,
and Dist samples with the use of Image J software (35).
Blood biochemistry, hematology, ex vivo phagocytotic ca-
pacity, bacterial culture, and cerebrospinal fluid cytokines.
Blood biochemistry and cell counting were analyzed in cardiac blood
collected when the pigs were killed with the use of a Siemens ADVIA
1800 Chemistry System and an ADVIA 2120i Hematology System (31),
and plasma cortisol and C-reactive protein (CRP) concentrations were
measured by ELISA (KGE008, DY2648, R&D Systems). In vitro phago-
cytic function of blood neutrophils was determined in whole blood
taken on day 5 and day 8 with the use of an Escherichia coli-conjugated
pHrodo Red BioParticles Phagocytosis kit (P35381, Thermo Fisher) on
a FACS Canto II flow cytometer (BD Biosciences) (36). Bacteria from the
aseptically sampled whole blood and homogenized bone marrow were
cultured on blood agar and colonies were enumerated (31). Cytokines
in cerebrospinal fluid were measured with the use of multiplex ELISA
kit (see Supplemental Methods).
In vivo protein synthesis and plasma free amino acid
(AA) profile. Tissue protein synthesis was measured in vivo with
the use of the modified flooding dose technique (37). Exactly 30
min before they were killed, pigs were infused with 10 mL/kg
body weight of a flooding dose of phenylalanine via an umbili-
cal catheter containing 135 mM of unlabeled l-phenylalanine and
15 mM of l-[ring-13 C6 ] phenylalanine (Cambridge Isotope Labo-
ratories). Blood, liver, muscle from the hind leg, and cerebellum
were taken 30 min after the infusion, snap frozen and stored at
–80°C. Protein fractional synthesis rates (FSRs) of various tissues, repre-
senting the amount of protein synthesized per day as a percentage of the
total protein pool, were calculated from the amount of free l-[ring-13 C6 ]
phenylalanine relative to protein bound l-[ring-13 C6 ] phenylalanine in
each tissue. Postprandial plasma-free AA profile after last feeding was
measured by reverse-phase HPLC of their phenylisothiocyanate deriva-
tives (38).
Microbiota and SCFAs. Genomic DNA was extracted from stom-
ach, cecum, and colon content with the use of the PowerMag Soil
DNA Isolation Kit optimized for epMotion robotic platform (MO-BIO
338 Sun et al.
Laboratories). 16S rRNA amplicon gene sequencing was performed
with the use of 2-step PCR amplification and the Illumina MiSeq plat-
form (see Supplemental Methods). Total bacterial load was quantified
by qPCR with the use of universal primer 341F and 534R, and stan-
dard DNA extracted from E. coli K12 strain with a known concen-
tration (31, 39). The density of bacteria adhering to the mucosa was
determined by fluorescence in situ hybridization of whole cross sections
of paraffin-embedded intestinal tissues and quantified by Visiopharm,
as described previously (31). The gut luminal concentrations of SCFAs,
including acetic acid, butyric acid, propionic acid, and pentanoic acid,
were measured in the colon content by GC-MS (26).
Statistical analyses. All analyses were performed with the use of
the statistical software environment R (version 3.3.2). First, separate
comparisons were made between DHM and DPM groups, and then
comparisons among the 3 DHM-based groups were made with the
Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
use of linear models. Continuous outcomes, such as organ weights
and hexose uptake, were analyzed with the use of the lm func-
tion, followed by Tukey’s post-hoc test for paired-comparison among
the 3 DHM-based groups. Repeated measurements, such as physi-
cal activity, were analyzed with the use of the lmer function, fol-
lowed by post-hoc paired-comparison with the use of the lsmeans
package. The models were checked for normality and variance ho-
mogeneity of the residuals and fitted values, and data transforma-
tion was performed if necessary. Binary data, such as NEC diagnoses,
were analyzed with the use of the glm function and the model was
checked for dispersion. All aforementioned statistical models were ad-
justed for potential covariates and confounders, such as litter, sex, and
birth weight. Nonparametric tests were performed for adrenal gland
weight, blood hematology, and neutrophil phagocytosis, and protein
FSR data, and post hoc tests were performed by Dunn’s test with
the use of the PMCMR package. For colonic IL-6 and IL-8 concen-
trations, the mean values for 1 of the fortified groups were tested
against the pooled mean value of pigs from the other 2 groups. Data
are presented as means ± SEMs, unless otherwise stated. P < 0.05
was considered significant and P ≤ 0.10 as a tendency to an effect.
Correlations between 2 continuous variables were tested by Pearson
correlation test, and differences of the slopes from 2 regressions were
tested by ANCOVA. Autoscaled plasma AA data were applied to princi-
pal component analysis (PCA) with the use of the pcaMethods package.
Microbiota analyses (α, β diversity, relative abundances) are described
in the Supplemental Methods.
Results
DHM compared with DPM to preterm pigs. DHM
pigs gained <10% of the weight gain of DPM pigs
(P < 0.05, Figure 1), although clinical scores and the physical ac-
tivity were similar (Supplemental Table 3). DHM pigs had lower
intestinal weight-to-length ratio, villus height, brush border
lactase, and ApA and DPPIV activities (all P < 0.01, Figure 1),
but no differences were found in crypt depth, hexose uptake,
gut permeability, and other brush border enzyme activities
(Figure 1, Supplemental Table 3). IL-6 and IL-8 concentrations
in Dist or colon were higher in DHM pigs (both P < 0.01,
Figure 1). DHM pigs had lower protein FSRs in muscle and
liver than DPM pigs (both P < 0.05, Figure 1). The weight of
other organs, gut functions and cytokines, blood biochemistry,
and hematology are summarized in Supplemental Table 3.
Collectively, the results show that DHM pigs had reduced
growth and intestinal maturation relative to DPM pigs.
Nutrient fortifier effects on body weight gain, clini-
cal conditions, organ dimensions, blood biochemistry.
DHM+BC pigs had higher body weight gain than DHM
and DHM+FF pigs (both P < 0.05, Figure 2A), and higher
 Weight gain, g/d * DPM
Diarrhea score
Intestinal weight-to-length ratio *
Villus height, μm *
Hexose uptake, μM *
Lactase activity, U/g *
ApA activity, U/g *
DPPIV activity, U/g *
Gut permeability, ratio
Dist IL6, pg/mg prot
Dist IL8, pg/mg prot
Colonic IL6, pg/mg prot
Protein FSR in liver, %/d *
Protein FSR in muscle, %/d *
Protein FSR in cerebellum, %/d
0 1
FIGURE 1 Body weight, gut function and inflammation, and tissue protein synthesis in DHM pigs relative to DPM pigs (set at 1.0). Values
are means ± SEMs, n = 5–9 (DPM) or 14–19 (DHM). *Different from DPM, P < 0.05. ApA, aminopeptidase A; DHM, pasteurized donor human
milk; Dist, distal small intestine; DPM, pasteurized donor porcine milk; DPPIV, dipeptidylpeptidase; FSR, fractional synthesis rate; prot, protein.
intestinal weight and weight-to-length ratio than DHM pigs (all
P < 0.05, Table 1). DHM+FF pigs had higher adrenal gland
weight than both DHM and DHM+BC pigs, and lower cerebel-
lum weight than DHM+BC pigs (all P < 0.05, Table 1). No dif-
ferences were found in clinical scores, fecal consistency, or physi-
cal activity before day 5 (Figure 2B–D), but thereafter DHM+FF
pigs had higher clinical scores and lower physical activity than
DHM and DHM+BC pigs (all P < 0.05, Figure 2B, D). On day
6, when enteral feeding volume plateaued (132 mL  kg−1  d−1 ),
DHM+FF pigs still tended to have lower physical activity than
DHM+BC pigs (P = 0.07, Figure 2D) and more diarrhea than
DHM and DHM+BC pigs (both P < 0.05, Figure 2C). No dif-
ferences were observed between DHM and DHM+BC pigs in
these parameters (Figure 2B–D).
Plasma cortisol concentrations were 2.5-fold higher in
DHM+FF pigs, relative to DHM+BC (P < 0.05, Table 2).
DHM+FF pigs tended to have higher albumin concentrations
than DHM pigs, and higher alanine aminotransferase values
than DHM+BC pigs (both P = 0.06, Table 2). Both BC and FF
fortifiers reduced total bilirubin concentrations (both P < 0.05,
Table 2). No electrolyte concentration differences were found
except elevated blood iron concentrations in DHM+FF pigs
relative to DHM+BC (P < 0.05, Table 2). DHM+FF pigs had
higher hemoglobin concentrations than DHM and DHM+BC
pigs, and higher hematocrit values than DHM+BC pigs (all P
< 0.05, Table 2). Collectively, the aforementioned data suggest
DHM
*
Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
*
*
*
2 3 4 5 6
Fold change
that DHM+FF pigs had more clinical problems than DHM and
DHM+BC pigs.
Nutrient fortifier effects on intestinal function and struc-
ture, and inflammation. Relative to DHM and DHM+BC
pigs, DHM+FF pigs had lower intestinal hexose absorptive ca-
pacity (P = 0.08, P < 0.05, respectively; Figure 3A) and maltase
activity (both P ≤ 0.01, Figure 3C). Sucrase and lactase activi-
ties were lower in DHM+FF pigs than in DHM pigs (P = 0.06,
P = 0.03, respectively; Figure 3C). ApN and ApA activities were
lower in DHM+FF pigs than in DHM+BC pigs (P = 0.05, P <
0.05, respectively; Figure 3C), whereas no significant differences
were found on DPPIV activity (Figure 3C). Gastric residual vol-
ume was higher in DHM+BC pigs than in both DHM+FF and
DHM pigs (both P < 0.05, Figure 3B). Gut permeability did not
differ among the groups (data not shown). Only DHM+BC pigs
tended to show increased proportion of mucosa (53.6% ± 2.3%
compared with 44.72% ± 2.1%, P = 0.07) and circumference
in the Prox region, relative to DHM (P < 0.05, Table 1). Intesti-
nal circumference in the Dist region was higher in both forti-
fier groups, relative to DHM pigs (P < 0.05, Table 1). Shorter
villi and longer crypts were observed in DHM+FF pigs across
the small intestine, relative to DHM+BC with values in DHM
pigs being intermediate (all P < 0.01, Figure 3D, E, G). Thus,
the villus height–to–crypt depth ratio was markedly reduced in
DHM+FF pigs, relative to the other groups (both P < 0.01,
Human milk fortification in preterm neonates 339
 TABLE 1 Organ dimensions, incidence of NEC and bone
A 20 b marrow infection, and bacterial load in bone marrow at autopsy in
Body weight gain, g/d
neonatal preterm pigs fed DHM, DHM+BC or DHM+FF for 8 d1

15 DHM DHM+BC DHM+FF

Birth weight, g 953 ± 68 957 ± 45 943 ± 54
a
Small intestinal weight, g/kg 30 ± 1a 40 ± 2b 36 ± 2ab
10
Small intestinal length, cm/kg 326 ± 20 356 ± 24 363 ± 23
a
Intestinal weight-to-length, mg/cm 97 ± 6a 115 ± 5b 101 ± 6ab
Colon weight, g/kg 12.1 ± 0.9 16.1 ± 2.0 13.3 ± 1.1
5
Stomach weight, g/kg 6.3 ± 0.3 8.2 ± 0.7 7.1 ± 0.3
Lung weight, g/kg 20.7 ± 0.7 20.9 ± 1.1 22.6 ± 0.9
0 Spleen weight, g/kg 2.69 ± 0.26 2.63 ± 0.27 2.48 ± 0.28
DHM DHM+BC DHM+FF Liver weight, g/kg 38 ±2 35 ±1 37 ±2
Kidney weight, g/kg 7.9 ± 0.3 8.0 ± 0.3 8.7 ± 0.3
B

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

3 DHM Adrenal gland weight, g/kg 0.20 ± 0.01a 0.21 ± 0.01a 0.27 ± 0.02b
DHM+BC Brain stem weight, g 2.70 ± 0.04 2.65 ± 0.04 2.60 ± 0.07
DHM+FF
Cerebrum weight, g 26 ±1 25 ±0 24 ±1
Cerebellum weight, g 2.82 ± 0.08ab 2.88 ± 0.04b 2.64 ± 0.08a
Clinical score

Hippocampal weight, g 0.44 ± 0.04 0.40 ± 0.03 0.41 ± 0.02

2 Circumference of Prox, cm 11.6 ± 0.4a 13.5 ± 0.6b 12.1 ± 0.5a
b
Circumference of Dist, cm 11.7 ± 0.5a 13.4 ± 0.6b 13.5 ± 0.6b
Intestinal NEC (>score 3), % 18 10 18
a Colon NEC (>score 3), % 18 21 44
Average intestinal NEC score 2.00 ± 0.37 1.79 ± 0.33 2.00 ± 0.37
a
Average colonic NEC score 2.69 ± 0.38 2.42 ± 0.31 3.06 ± 0.49
1 Bone marrow infection, % 60 44 46
Bacterial load in bone marrow, cfu/kg 6.0 ± 3.2 2.3 ± 1.5 331 ± 101
C 7 1
Organ weights were normalized to body weight at time of death (except for brain re-
gions). Values are means ± SEMs unless otherwise noted; n = 10–15. Labeled means
without a common superscript letter in a row differ, P < 0.05. DHM, pasteurized donor
6 b human milk; DHM+BC, donor human milk fortified with bovine colostrum; DHM+FF,
donor human milk fortified with a formula-based fortifier; Dist, distal small intestine;
Fecal score

5 a NEC, necrotizing enterocolitis; Prox, proximal small intestine.

a b

4 ab Figure 3F). Collectively, the data indicate that DHM+FF pigs

had impaired intestinal structure and digestive function.
3 a The incidences of NEC in the small intestine and colon
were similar among groups (Table 1), and there were no sig-
2 nificant differences in distal intestinal cytokine concentrations
(data not shown). The DHM+BC pigs tended to have lower
colonic IL-6 and IL-8 concentrations relative to the remaining
D 25 pigs (both P = 0.06, 18 pigs from DHM+BC, 28 pigs from
DHM and DHM+FF groups, Supplemental Figure 1A–C). In the
20 blood, the proportion of phagocytic neutrophils was reduced in
b the DHM+FF pigs, relative to DHM (P < 0.05, Supplemental
Active time, %

15 Figure 1D) and pigs diagnosed with NEC had lower phagocytic
b capacity compared with healthy pigs (P < 0.05, Supplemental
10 Figure 1E). Concentrations of plasma CRP and bone marrow
a bacterial load did not significantly differ among groups (Supple-
mental Figure 1F, Table 1), and no bacteria were cultured from
5
whole blood collected from any pigs. DHM+FF pigs showed
increases in the mean concentration of several proinflammatory
0 cytokines in the pooled cerebrospinal fluid sample relative to
3 4 5 6 7 8 DHM pigs (Supplemental Figure 2). Collectively, the data indi-
Day
cate that DHM+FF diet induced a slightly more inflammatory
FIGURE 2 Weight gain (A) and clinical (B) and fecal (C) scores mea- condition in the intestine and blood relative to the DHM+BC
sured twice daily from the beginning of fortification, and physical ac- diet.
tivity monitored by infrared camera (D) in neonatal preterm pigs fed
DHM, DHM+BC, or DHM+FF for 8 d. Values are means ± SEMs, n = Nutrient fortifier effects on protein utilization. Liver
16–19. Labeled means (A) or means at a time (B–D) without a common
protein FSRs were higher in both FF and BC fortification
letter differ, P < 0.05. DHM, pasteurized donor human milk; DHM+BC,
donor human milk fortified with bovine colostrum; DHM+FF, donor
groups relative to DHM pigs (both P < 0.05, Figure 4A),
human milk fortified with a formula-based fortifier. and muscle protein FSRs were higher in DHM+FF pigs
than in DHM, with intermediate values in DHM+BC pigs
340 Sun et al.
TABLE 2 Plasma biochemistry and blood cell counts at pattern differed between DHM+FF and DHM+BC pigs
autopsy in neonatal preterm pigs fed DHM, DHM+BC or (Supplemental Figure 3).
DHM+FF for 8 d1
Nutrient fortifier effects on microbiota in the stomach, ce-
DHM DHM+BC DHM+FF
cum, and colon. Fortification decreased the Shannon diver-
Biochemistry parameters sity in the stomach content (both P < 0.01, Figure 5A), but
Albumin, g/L 14.3 ± 0.9 15.0 ± 0.9 18.3 ± 1.2 not in the cecum or colon contents (data not shown). For-
Total protein, g/L 33 ±2 33 ± 1 39 ± 2 tification altered the cecum microbiota composition, and a
Alkaline phosphate, U/L 2687 ± 393 2046 ± 392 2141 ± 197 similar tendency was found in the colon content (P < 0.05,
Alanine aminotransferase, U/L 15.5 ± 1.0 12.9 ± 1.0 18.9 ± 3.0 P = 0.09, Figure 6A). β dispersion was increased by for-
Aspartate aminotransferase, U/L 66 ± 19 36 ± 11 95 ± 60 tification in the cecum contents (P < 0.05, Figure 6A). In
γ -Glutamyltransferase, U/L 27 ±2 23 ± 2 29 ± 4 the stomach, DHM+BC pigs had higher abundance of Bac-
Total bilirubin, µmol/L 4.52 ± 1.66b 1.88 ± 0.32a 1.61 ± 0.33a teroidales and Clostridiales than DHM pigs (both P < 0.05,
Cholesterol, mmol/L 2.20 ± 0.29 1.97 ± 0.19 2.19 ± 0.24 Figure 6B). The DHM+FF pigs had higher abundance of
Creatinine, µmol/L 56 ±5 48 ± 2 67 ± 10 Pseudomonas than DHM and DHM+BC pigs, and higher

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

BUN-to-creatinine ratio 49 ± 18a 36 ± 4a 86 ± 19b abundance of Lactococcus than DHM pigs (all P < 0.05,
Cortisol, ng/mL 152 ± 51ab 77 ± 12a 193 ± 46b Figure 6B). In the colon, Enterococcus (P = 0.07) and Bacil-
Calcium, mmol/L 2.58 ± 0.09 2.36 ± 0.07 2.49 ± 0.23 lales (P = 0.05) tended to be more abundant, and Escherichia
Magnesium, mmol/L 0.97 ± 0.10 0.80 ± 0.02 0.94 ± 0.08 (P = 0.05) and Staphylococcus (P = 0.07) tended to
Sodium, mmol/L 138 ±2 134 ± 2 139 ± 3 be less abundant in DHM+FF compared with DHM pigs
Potassium, mmol/L 4.93 ± 0.44 3.98 ± 0.19 5.43 ± 1.11 (Figure 6B). No group differences were found in the ce-
Iron, µmol/L 7.2 ± 1.2ab 6.6 ± 1.1a 12.1 ± 2.5b cum at genus concentration (Figure 6B). Analyzed across
Phosphate, mmol/L 1.74 ± 0.34 1.51 ± 0.07 1.79 ± 0.24 the 3 intestinal regions, the density of bacteria adher-
Hematology parameters ing to the mucosa did not differ among groups, although
Total erythrocytes,1012 /L 3.38 ± 0.13 3.33 ± 0.16 3.85 ± 0.16 in the middle intestine, the density was higher in the
Hemoglobin, mmol/L 3.85 ± 0.21a 3.89 ± 0.19a 4.60 ± 0.22b DHM+BC and DHM+FF groups than in the DHM group
Hematocrit, % 21 ± 1ab 20 ± 1a 25 ± 1b (P < 0.05, Figure 6C, E). When the density was expressed rel-
Total leukocytes, 109 /L 8.1 ± 1.0 8.1 ± 1.2 11.3 ± 2.0 ative to villus height (as a surrogate marker for surface area),
Neutrophils, 109 /L 4.69 ± 0.73 5.42 ± 0.93 7.66 ± 1.56 the values in DHM+FF pigs were higher than DHM pigs in the
Lymphocytes, 109 /L 2.86 ± 0.35 2.32 ± 0.26 2.82 ± 0.34 Mid, and higher than DHM+BC pigs in the Dist region (P <
Monocytes, 109 /L 0.21 ± 0.05 0.21 ± 0.06 0.40 ± 0.12 0.05, Figure 6D, E). The concentrations of total SCFAs in the
Thrombocyte, 109 /L 122 ± 25 263 ± 42 221 ± 50 colon content were lower in DHM+BC pigs than in DHM pigs
1
Values are means ± SEMs (n = 15–18 for biochemistry parameters; n = 9–13 for (77.4 ± 11.8 nM compared with 94.4 ± 15.3 mM, P < 0.05,
hematology parameters). Labeled means without a common superscript letter in a row DHM + FF: 89.3 ± 14.3 mM).
differ, P < 0.05. BUN, blood urea; DHM, pasteurized donor human milk; DHM+BC, Lower gastric acidity (higher pH) was associated with lower
donor human milk fortified with bovine colostrum; DHM+FF, donor human milk forti-
fied with a formula-based fortifier.
Shannon diversity (r = 0.60, P < 0.05, Figure 5B) and higher
total bacterial load (r = 0.63, P < 0.05, Figure 5C). Gastric pH
values were higher in DHM+FF than DHM and DHM+BC
(P < 0.05, Figure 4B). No significant differences were observed pigs (both P < 0.05, Figure 5D), and total bacteria count in
in the cerebellum protein FSR (Figure 4C). Despite similar the stomach content was increased in the DHM+FF compared
concentrations of protein fortification, DHM+FF pigs had 4- with DHM pigs (P < 0.05, Figure 5E). The association between
fold higher blood urea (BUN) concentrations than DHM+BC gastric pH and both relative and absolute abundance of Es-
pigs (P < 0.05, Figure 4D), leading to 2- to 3-fold higher cherichia/Shigella was fortification dependent, showing a pos-
BUN-to-creatinine ratios in these pigs relative to DHM and itive correlation only in pigs receiving fortification (P = 0.08,
DHM+BC pigs (both P < 0.05, Table 2). DHM+FF also P < 0.05, respectively, Figure 5F, G).
tended to have higher creatinine kinase (CK) concentrations
than DHM+BC pigs (P = 0.06, Figure 4E). BUN concentration
was positively correlated with cortisol concentration (r = 0.59,
Discussion
P < 0.01).
Following the last feeding, both BC and FF fortification in- Banked DHM is increasingly used to feed preterm infants, yet
creased the postprandial plasma concentrations of tyrosine, pro- human milk does not meet the nutritional needs of preterm in-
line, asparagine, and total amount of branched-chain AAs (all fants for growth and potential development. As a result, fortifi-
P < 0.05), whereas the valine values were only elevated cation of human milk is currently recommended for feeding very
relative to DHM by BC fortification (P < 0.05, Table 3). premature infants (14, 40), but without a clear consensus on
DHM+FF pigs had higher concentrations of ornithine, lysine, how and when this should be best made. Large variation exists
and isoleucine relative to the other groups (all P < 0.05, in the composition of commercially available, formula-based
Table 3). Concentrations of leucine, (all P < 0.05), sulfur AA nutrient fortifiers, and the optimal type of fortifier has yet to be
(P = 0.05), and essential AA–to–total AA ratio (P = determined. In this study, we used rapidly growing preterm pigs
0.06) were only increased in DHM+FF pigs, but not with an immature intestine, and a nutrient-restricted basal diet
DHM+BC, relative to DHM (Table 3). DHM+BC pigs of DHM, to compare how nutrient fortifiers based on bovine
had lower taurine concentrations than DHM and DHM+FF milk proteins (BC, FF), affected body growth, clinical variables,
pigs (both P < 0.05) and lower arginine than DHM+FF and intestinal health. First, we demonstrated that preterm pigs
pigs (P < 0.05, Table 3). The PCA score plot, com- fed DHM developed growth restriction and intestinal dysfunc-
bining all AA values, indicates that the postprandial AA tion compared with pigs fed DPM during the first week of life.
Human milk fortification in preterm neonates 341
 Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
FIGURE 3 Intestinal hexose absorptive capacity (A), gastric residual shown by the ratio of weight of stomach content to the weight of last
feeding bolus (B), brush border enzyme activities (C), villus height (D), crypt depth (E), and their ratio (F) across the small intestine in neonatal
preterm pigs fed DHM, DHM+BC, or DHM+FF for 8 d. Values are means ± SEMs, n = 12–14 (A, B), n = 15–18 (C–F). Labeled means without
a common letter differ, P < 0.05. (G) Representative images of histology in Prox small intestine from the 3 groups. Hematoxylin and eosin
staining, 100 × magnification. ApA, aminopeptidase A; ApN, aminopeptidase N; DHM, pasteurized donor human milk; DHM+BC, donor human
milk fortified with bovine colostrum; DHM+FF, donor human milk fortified with a formula-based fortifier; Dist, distal small intestine; DPPIV,
dipeptidylpeptidase; Mid, middle small intestine; Prox, proximal small intestine.
Protein FSR in cerebellum, %/d

A B
Protein FSR in muscle, %/d

C
Protein FSR in liver, %/d

100 b b 40 40
80 a b 30
30 ab
60 a
20 20
40
20 10 10

0 0 0

DHM +BC HM+FF DHM +BC HM+FF DHM HM+B

C +FF
DHM D DHM D D DHM
D E P=0.06
10 400
b
Plasma BUN, mM

Plasma CK, U/L

8 300
6 ab
200
4
a 100
2
0 0

DHM +BC HM+FF DHM HM+B

C +FF
DHM D D DHM

FIGURE 4 Protein FSR in the liver (A), muscle (B), and cerebellum (C), and plasma BUN (D) and CK (E) measured at time of
death in neonatal preterm pigs fed DHM, DHM+BC, or DHM+FF for 8 d. Values are means ± SEMs, n = 7–8 (A, B), n = 3–4 (C),
n = 15–18 (D, E). Labeled means without a common letter differ, P < 0.05. BUN, blood urea; CK, creatine kinase; DHM, pasteurized donor
human milk; DHM+BC, donor human milk fortified with bovine colostrum; DHM+FF, donor human milk fortified with a formula-based fortifier;
FSR, fractional synthesis rate.

342 Sun et al.

TABLE 3 Postprandial plasma AAs at autopsy in neonatal villus height, digestive enzyme activities, and hexose uptake
preterm pigs fed DHM, DHM+BC or DHM+FF for 8 d1 were also lower in DHM+FF compared with DHM and
DHM+BC pigs. It is unknown what factors in FF may lead
µmol/L DHM DHM+BC DHM+FF
to suboptimal intestinal health, but possible causes include ad-
Valine 266 ± 25a
382 ± 13b
350 ± 30ab verse processing procedures, such as hydrolysis and heat treat-
Tyrosine 26 ± 4a 95 ± 13b 101 ± 15b ments, that may destroy the protein structure and diminish the
Tryptophan 18 ± 1 19 ± 1 22 ± 2 bioactivity of the fortifier (45). Previous reports showed that
Threonine 149 ± 39 243 ± 33 240 ± 34 deeper crypts, shorter villi, and more diarrhea were observed in
Taurine 166 ± 26b 91 ± 13a 136 ± 13b early-weaned pigs after consuming hydrolyzed dietary protein
Serine 215 ± 18 308 ± 40 204 ± 21 (46), whereas bioactive compounds (e.g. IGF-1) present in larger
Proline 169 ± 18a 338 ± 38b 318 ± 38b amounts in BC, but not in FF, can stimulate intestinal cell prolif-
Ornithine 29 ± 6a 32 ± 2a 62 ± 15b eration and mucosal growth in newborn pigs (47, 48). Nonmilk
Methionine 59 ± 13 70 ± 5 77 ± 9 ingredients presented in FF, such as medium-chain triglycerides
Lysine 118 ± 24a 138 ± 11a 255 ± 30b extracted from vegetable oil, may also damage the mucosal bar-
Leucine 72 ± 10a 121 ± 14a 178 ± 34b rier function (20).

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

Isoleucine 89 ± 14a 100 ± 9a 162 ± 10b Maldigestion, bacterial colonization, barrier dysfunction,
Histidine 133 ± 18 128 ± 7 140 ± 22 and adverse immune responses make preterm neonates suscepti-
Glycine 1024 ± 223 1187 ± 154 701 ± 74 ble to gut inflammation, NEC, bacterial translocation and sepsis
Glutamate 246 ± 57 292 ± 42 353 ± 46 (49). We used a clinically relevant feeding regimen to simulate
Glutamine 192 ± 24a 374 ± 72ab 389 ± 62b that used for many preterm infants with a gradual increase in en-
Citrulline 62 ± 6 69 ± 5 87 ± 8 teral feeding volumes (reaching 132 mL  kg−1  d−1 by day 6).
Aspartate 20 ± 6 22 ± 3 30 ± 4 A more rapid transition to full enteral feeding generally in-
Asparagine 26 ± 3a 65 ± 14b 56 ± 11b creases NEC incidence in this model, as shown previously (25,
Arginine 34 ± 10ab 29 ± 2a 56 ± 10b 27). With The Use Of the current relatively gentle feeding ap-
Alanine 449 ± 76a 645 ± 49ab 733 ± 76b proach, we found that BC, but not FF fortification, tended to
Total AA2 3560 ± 493 4747 ± 415 4648 ± 404 alleviate gut inflammation via reducing colonic IL-6 and IL-8
Essential AA2 903 ± 126a 1200 ± 57ab 1423 ± 133b concentrations, 2 key cytokines that are often upregulated in re-
Sulfur AA 121 ± 17 139 ± 8 163 ± 15 sponse to bacterial stimulation and inflammation (50–52). Yet
Branched chain AA 426 ± 42a 603 ± 24b 690 ± 63b the FF pigs, specifically, showed a higher density of bacteria ad-
Essential AA:total AA 0.26 ± 0.02 0.27 ± 0.02 0.31 ± 0.01 hering to their stunted villi, potentially increasing the risk of
1
Values are means ± SEMs in µmol/L; n = 10–14. Labeled means without a common bacterial translocation in this group. The above differences may
superscript letter in a row differ, P < 0.05. AA, amino acid; DHM, pasteurized donor have been even more pronounced if the pigs had been followed
human milk; DHM+BC, donor human milk fortified with bovine colostrum; DHM+FF, for a longer growth period or transitioned faster to full enteral
donor human milk fortified with a formula-based fortifier.
2
feeding. We have observed that preterm pigs fed rapidly increas-
Phe was not included in the calculation of total AA, essential AA, and essential
AA:total AA ratio because an extra amount of unlabeled phenylalanine was infused ing volumes of DHM+BC during the first 4 d showed lower
30 min before blood collection for protein synthesis analyses. concentrations of NEC severity, gut permeability, and poten-
tial bacterial translocation compared with pigs fed DHM+FF
(unpublished data). Beyond the gut, the moderate gut inflam-
This can be explained by the lower concentrations of nutri- mation in the DHM+FF group also appeared to alter the sys-
ents and intestinal protective components in DHM compared temic immunity. Thus, pigs diagnosed with NEC, specifically the
with DPM (24, 41). When the low-protein, low-energy DHM DHM+FF pigs, were found to have a reduced phagocytic ca-
was fortified to match the protein and energy concentrations of pacity of blood neutrophils. This suggests an elevated risk of
DPM, we observed increased protein synthesis in response to systemic sepsis in these pigs, originating from infections via in-
both FF and BC fortification but reduced weight gain and in- dwelling catheters or from bacterial translocation via the im-
testinal function in pigs fortified with FF compared with BC. mature gut epithelium. While this was not clearly evident dur-
Further studies are required to clarify the mechanisms behind ing the first 8 d after birth, we did observe a nominally higher
these differential fortification responses, but BC might be rele- density of bacteria in the bone marrow (>100-fold) and higher
vant as an alternative fortifier for preterm infants. mean value of plasma CRP (>2-fold) in DHM+FF pigs relative
A main goal of human milk fortification for very premature to DHM+BC pigs.
infants is to prevent the extrauterine growth restriction that is The pattern of changes in protein metabolic endpoints sug-
associated with later neurodevelopmental delays, cerebral palsy, gests that BC provided a superior benefit in terms of pro-
and neurological abnormalities (42). Our results showed that tein digestion and metabolism via reducing protein catabolism.
fortification with FF compared with BC resulted in lower body Both fortifiers, especially FF, led to increased rates of protein
weight gain and intestinal dimensions, relative to DHM pigs. synthesis in the liver and muscle and higher plasma AA (e.g.
The FF used in this study appeared to trigger a degree of intesti- tyrosine) which is likely explained by higher protein intake
nal dysfunction shortly after preterm birth. Indeed, gut matura- (53–56). The PCA clustering of all the AAs indicates that
tion and function are very important factors for nutrient uptake, DHM+FF pigs had a different postprandial AA profile
growth, and organ development in preterm neonates, especially compared with DHM+BC pigs, mainly determined by higher
during the shift from PN to full enteral feeding. During this pe- concentrations of many AAs in the DHM+FF pigs. Corre-
riod, preterm infants frequently suffer from feeding intolerance, spondingly, the DHM+FF pigs had higher concentrations of
intestinal dysmotility, and maldigestion due to gut immaturity nitrogen excretion (high BUN concentrations) and more evi-
(43, 44). In preterm pigs, we observed more feeding intoler- dence of AA excretion and muscle protein breakdown (high
ance, diarrhea, and discomfort (lower physical activity) when CK concentrations) (57), relative to DHM+BC pigs. The
FF was used as a fortifier relative to BC. Consistently, the evidence of higher AA excretion may be explained by an
Human milk fortification in preterm neonates 343
 A 2.5 D 5 E 10 ab b
b c a

Shannon diversity
2.0 4 9

Log10 copy/g
a a b

Gastric pH
1.5 a 8
3
1.0 2 7
0.5 1 6
0.0 0 5

DHM +BC HM+FF DHM +BC HM+FF DHM HM+B

C +FF
DHM D DHM D D DHM

B 3.5 F 1 No fortification

Shigella, relative abundance

3 r=0.60 Fortification
P<0.05 0.8

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

Shannon diverstiy

2.5

2 0.6

1.5
0.4
1
0.2
0.5

0 0
2 3 4 5 6 2 3 4 5 6
Gastric pH Gastric pH
C G
11 11
Totoal bacteria, Log10 copy/g

Shigella, Log10 copy/g

10
10
9

9 8

7
8 r=0.63
P<0.05 6

7 5
2 3 4 5 6 2 3 4 5 6
Gastric pH Gastric pH

FIGURE 5 Shannon diversity index (A), correlations between gastric pH and Shannon diversity index (B) or total bacterial count (C) independent
of fortification groups, gastric acidity (D) and total bacteria count (E) measured in the stomach content collected from the neonatal preterm pigs
fed DHM, DHM+BC or DHM+FF for 8 d. Correlations between gastric pH and relative (F) or absolute (G) abundance of Escherichia/Shigella in
pigs treated with or without fortification. Values are means ± SEMs (A, D, E), n = 14–18, and labeled means without a common letter differ,
P < 0.05. DHM, pasteurized donor human milk; DHM+BC, donor human milk fortified with bovine colostrum; DHM+FF, donor human milk
fortified with a formula-based fortifier.

accelerated gastric emptying and a more rapid absorption of in the DHM+FF group may relate to increased physiologic
AAs and small peptides, following the addition of the pre- stress and digestive discomfort in response to suboptimal
hydrolyzed proteins in the DHM+FF group, consistent with diets (64).
earlier reports on the use of protein hydrolysates in infants The gut microbiota is not yet stable in early life, and thus it is
(58–61). In preterm infants, the net nitrogen use of hydrolyzed highly sensitive to external environment (i.e., enteral nutrition)
protein formula is ∼10% lower than that of intact protein and may actively interact with the host at many concentrations.
formula (45), and intake of hydrolyzed milk protein is also Fortification of DHM, especially by adding FF, resulted in a de-
associated with lower weight gain and higher urinary ex- creased gastric acidity (increased pH), which might be due to a
cretion of essential AAs (42, 62, 63). Lowered AA utiliza- relatively high buffer effect of the FF. This loss of gastric acidity
tion may also be linked to increased muscle tissue proteoly- not only delays the digestion of milk protein and lipid in the
sis, in part mediated by the elevated cortisol concentrations stomach (due to the low pH-optimum for gastric proteases and
(a highly catabolic hormone) in the DHM+FF group. Consis- lipases) but also may increase the risk of bacterial overgrowth,
tently, we found a positive correlation between cortisol and urea low bacterial diversity, and pathogen colonization, potentially
concentrations. In addition, the higher cortisol concentrations leading to bacterial translocation and NEC lesions further down

344 Sun et al.

 A Stomach Cecum Colon

0.2 0.50
0.25
0.0

PCoA2
0.25
0.00
−0.2
0.00
−0.4 DHM
−0.25
DHM+BC
DHM+FF
−0.6 −0.25
−0.25 0.00 0.25 0.50 −0.4 −0.2 0.0 0.2 0.4 0.6 −0.25 0.00 0.25 0.50
PCoA1

B 1.00

Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019

Adjusted P value

Enterococcus
0.10 Pseudomonas Staphylococcus

Order_Clostridiales Escherichia Order_Bacillales

Pseudomonas
Family_Clostridiaceae_1 Log10 total counts
Lactococcus 2
3 DHM vs DHM+BC
0.01
4 DHM vs DHM+FF
Order_Bacteroidales
Pseudomonas 5 DHM+BC vs DHM+FF

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
Fold change
number of particles/section

D
Adhering-bacteria density

C
Adhering-bacteria density

5000 15
per villus height, /μm

4000
b 10 b
3000

2000 b ab
5
b
1000 a ab
a
0 0 a
Prox Mid Dist Prox Mid Dist

E DHM DHM+BC DHM+FF

D
DHM DHM
HM
H M+BC
+B
BC
BC
DHM+BC
DHM DHM+BC DHM+FF

FIGURE 6 PCoA plot of β diversity (A) of the gut microbiota in neonatal preterm pigs fed DHM, DHM+BC or DHM+FF for 8 d, and bacteria
genera (B) plotted by fold change in relative abundance between groups (x axis) and significance of change (y axis, adjusted P < 0.1, n = 11–18).
(C) Density of adhering bacteria (number of particles with fluorescent red signal) on the mucosal surface in the 3 intestinal regions. (D) Density
expressed relative to villus height as a surrogate marker of surface area. Values are means ± SEMs, n = 11–15, and labeled means without a
common letter differ, P < 0.05. (E) Representative bacterial staining images from the 3 groups. DHM, pasteurized donor human milk; DHM+BC,
donor human milk fortified with bovine colostrum; DHM+FF, donor human milk fortified with a formula-based fortifier; Dist, distal small intestine;
Mid, middle small intestine; Prox, proximal small intestine.

the gastrointestinal (GI) tract (65, 66). In our study, we observed Early acquisition of bacteria in the upper GI tract may pro-
that Escherichia/Shigella, in particular, appeared to proliferate vide the seed for shaping the microbiota in the lower parts of
in the stomach where fortification caused a higher gastric pH. the GI tract, although there are few studies on the stomach
Escherichia/Shigella are 2 closely related Gram-negative genera microbiota in preterm neonates, and it is important to docu-
belonging to Enterobacteriaceae, and their abundances were in- ment in more detail if the combined stomach microbiota, or
creased prior to onset of NEC in an earlier infant study (67). specific groups of bacteria, are associated with gut inflammation

Human milk fortification in preterm neonates 345

during the first weeks of life especially in response to different
diets.
In conclusion, our study provides important preclinical evi-
dence for a diet-dependent response to nutrient fortification of
human donor milk to preterm neonates. Our results suggest that
BC is superior to the tested FF in promoting intestinal matura-
tion, nutrient metabolism, and body growth. Further studies are
required to investigate the long-term effects of nutrient fortifica-
tion on body growth, prevention of NEC, and sepsis, and neuro-
logical outcomes. It is critical to identify the specific fortification
products, product combinations, or industrial processing tech-
niques that appear to make commercial fortifiers less efficient in
promoting growth, nutrient utilization, and gut maturation in
preterm neonates than an intact, nutrient-dense, bioactive milk
diet, such as bovine colostrum.
Acknowledgments
We thank Jane Povlsen, Elin Skytte, Kristina Møller, Anders
Brunse, Päivi Worsøe, and Karina Ryom for technical sup-
port on animal procedures and laboratory analyses. We thank
Johannes B van Goudoever for kindly donating the leftover hu-
man donor milk. The authors’ responsibilities were as follows—
JS, YL, CHPvdA, TS, TT, DGB, and PTS: designed the research;
JS, DNN, MSM, CHPvdA, TS, SEP, SP, and SR: conducted the
research; JS, MSM, SJS, and DGB: analyzed the data; JS, DGB,
and PTS: wrote the paper; PTS: had primary responsibility for
the final content; and all authors: read and approved the final
manuscript.
References
1. Clark RH, Thomas P, Peabody J. Extrauterine growth restriction
remains a serious problem in prematurely born neonates. Pediatrics
2003;111:986–90.
2. Saigal S, Doyle LW. An overview of mortality and sequelae of preterm
birth from infancy to adulthood. Lancet 2008;371:261–9.
3. Wilson-Costello D, Friedman H, Minich N, Fanaroff AA, Hack M.
Improved survival rates with increased neurodevelopmental disability
for extremely low birth weight infants in the 1990s. Pediatrics
2005;115:997–1003.
4. Hofman PL, Regan F, Jackson WE, Jefferies C, Knight DB, Robinson
EM, Cutfield WS. Premature birth and later insulin resistance. N Engl J
Med 2004;351:2179–86.
5. Posod A, Odri Komazec I, Kager K, Pupp Peglow U, Griesmaier E,
Schermer E, Wurtinger P, Baumgartner D, Kiechl-Kohlendorfer U.
Former very preterm infants show an unfavorable cardiovascular risk
profile at a preschool age. PLoS One 2016;11:e0168162.
6. Guellec I, Lapillonne A, Marret S, Picaud JC, Mitanchez D, Charkaluk
ML, Fresson J, Arnaud C, Flamand C, Cambonie G, et al. Effect of
intra- and extrauterine growth on long-term neurologic outcomes of
very preterm infants. J Pediatr 2016;175:93–9 e1.
7. Marks K-A, Reichman B, Lusky A, Zmora E, in collaboration with the
Israel Neonatal Network. Fetal growth and postnatal growth failure in
very-low-birthweight infants. Acta Paediatr 2006;95:236–42.
8. Sakurai M, Itabashi K, Sato Y, Hibino S, Mizuno K. Extrauterine growth
restriction in preterm infants of gestational age ≤32 weeks. Pediatr Int
2008;50:70–5.
9. Di Natale C, Coclite E, Di Ventura L, Di Fabio S. Fortification of
maternal milk for preterm infants. J Matern Fetal Neonatal Med
2011;24(Suppl 1):S41–3.
10. Can E, Bulbul A, Uslu S, Comert S, Bolat F, Nuhoglu A. Effects
of aggressive parenteral nutrition on growth and clinical outcome in
preterm infants. Pediatr Int 2012;54:869–74.
11. Kansagra K, Stoll B, Rognerud C, Niinikoski H, Ou C-N, Harvey
R, Burrin D. Total parenteral nutrition adversely affects gut barrier
function in neonatal piglets. Am J Physiol Gastrointest Liver Physiol
2003;285:G1162–70.
346 Sun et al.
12. Kelly DA. Intestinal failure-associated liver disease: what do we know
today? Gastroenterology 2006;130:S70–7.
13. Arslanoglu S, Corpeleijn W, Moro G, Braegger C, Campoy C, Colomb
V, Decsi T, Domellof M, Fewtrell M, Hojsak I, et al. Donor human milk
for preterm infants: current evidence and research directions. J Pediatr
Gastr Nutr 2013;57:535–42.
14. Section on Breastfeeding. Breastfeeding and the use of human milk.
Pediatrics 2012;129:e827–41.
15. Shah SD, Dereddy N, Jones TL, Dhanireddy R, Talati AJ. Early versus
delayed human milk fortification in very low birth weight infants—a
randomized controlled trial. J Pediatr 2016;174:126–31 e1.
16. Abdelhamid AE, Chuang S-L, Hayes P, Fell JME. Evolution of in
vitro cow’s milk protein-specific inflammatory and regulatory cytokine
responses in preterm infants with necrotising enterocolitis. J Pediatr
Gastr Nutr 2013;56:5–11.
17. Sullivan S, Schanler RJ, Kim JH, Patel AL, Trawoger R, Kiechl-
Kohlendorfer U, Chan GM, Blanco CL, Abrams S, Cotten CM, et al.
Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
An exclusively human milk-based diet is associated with a lower rate
of necrotizing enterocolitis than a diet of human milk and bovine milk-
based products. J Pediatr 2010;156:U562–83.
18. Abrams SA, Schanler RJ, Lee ML, Rechtman DJ, Prolacta Study G.
Greater mortality and morbidity in extremely preterm infants fed a diet
containing cow milk protein products. Breastfeed Med 2014;9:281–5.
19. Brown JV, Embleton ND, Harding JE, McGuire W. Multi-nutrient
fortification of human milk for preterm infants. Cochrane Database Syst
Rev 2016;CD000343.
20. Lindmark T, Kimura Y, Artursson P. Absorption enhancement through
intracellular regulation of tight junction permeability by medium chain
fatty acids in Caco-2 cells. J Pharmacol Exp Ther 1998;284:362–9.
21. Tanaka S, Saitoh O, Tabata K, Matsuse R, Kojima K, Sugi K, Nakagawa
K, Kayazawa M, Teranishi T, Uchida K, et al. Medium-chain fatty
acids stimulate interleukin-8 production in Caco-2 cells with different
mechanisms from long-chain fatty acids. J Gastroenterol Hepatol
2001;16:748–54.
22. Li Y, Ostergaard MV, Jiang P, Chatterton DE, Thymann T, Kvistgaard
AS, Sangild PT. Whey protein processing influences formula-induced gut
maturation in preterm pigs. J Nutr 2013;143:1934–42.
23. Ganapathy V, Hay JW, Kim JH. Costs of necrotizing enterocolitis and
cost-effectiveness of exclusively human milk-based products in feeding
extremely premature infants. Breastfeed Med 2012;7:29–37.
24. Chatterton DE, Nguyen DN, Bering SB, Sangild PT. Anti-inflammatory
mechanisms of bioactive milk proteins in the intestine of newborns. Int
J Biochem Cell Biol 2013;45:1730–47.
25. Shen RL, Thymann T, Ostergaard MV, Stoy ACF, Krych L, Nielsen
DS, Lauridsen C, Hartmann B, Holst JJ, Burrin DG, et al. Early
gradual feeding with bovine colostrum improves gut function and NEC
resistance relative to infant formula in preterm pigs. Am J Physiol
Gastrointest Liver Physiol 2015;309:G310–23.
26. Rasmussen SO, Martin L, Ostergaard MV, Rudloff S, Li Y, Roggenbuck
M, Bering SB, Sangild PT. Bovine colostrum improves neonatal growth,
digestive function, and gut immunity relative to donor human milk and
infant formula in preterm pigs. Am J Physiol Gastrointest Liver Physiol
2016;311:G480–91.
27. Jensen ML, Sangild PT, Lykke M, Schmidt M, Boye M, Jensen BB,
Thymann T. Similar efficacy of human banked milk and bovine
colostrum to decrease incidence of necrotizing enterocolitis in preterm
piglets. Am J Physiol Regul Integr Comp Physiol 2013;305:R4–12.
28. Sty AC, Sangild PT, Skovgaard K, Thymann T, Bjerre M, Chatterton
DE, Purup S, Boye M, Heegaard PM. Spray dried, pasteurised bovine
colostrum protects against gut dysfunction and inflammation in preterm
pigs. J Pediatr Gastroenterol Nutr 2016;63:280–7.
29. Li Y, Juhl SM, Ye X, Shen RL, Iyore EO, Dai Y, Sangild PT, Greisen
GO. A stepwise, pilot study of bovine colostrum to supplement the first
enteral feeding in preterm infants (precolos): study protocol and initial
results. Front Pediatr 2017;5: 42.
30. Juhl SM, Ye X, Zhou P, Li Y, Iyore EO, Zhang L, Jiang P, van Goudoever
JB, Greisen G, Sangild PT. Bovine colostrum for preterm infants in
the first days of life: a randomized controlled pilot trial. J Pediatr
Gastroenterol Nutr 2017. DOI: 10.1097/MPG.0000000000001774,
PMID: 29019855.
31. Li Y, Nguyen DN, de Waard M, Christensen L, Zhou P, Jiang P,
Sun J, Bojesen AM, Lauridsen C, Lykkesfeldt J, et al. Pasteurization
procedures for donor human milk affect body growth, intestinal
 structure, and resistance against bacterial infections in preterm pigs. J
Nutr 2017;147:1121–30.
32. Ostergaard MV, Shen RL, Stoy AC, Skovgaard K, Krych L, Leth SS,
Nielsen DS, Hartmann B, Bering SB, Schmidt M, et al. Provision of
amniotic fluid during parenteral nutrition increases weight gain with
limited effects on gut structure, function, immunity, and microbiology
in newborn preterm pigs. J Parenter Enteral Nutr 2016;40:552–66.
33. Andersen AD, Sangild PT, Munch SL, van der Beek EM, Renes IB,
Ginneken C, Greisen GO, Thymann T. Delayed growth, motor function
and learning in preterm pigs during early postnatal life. Am J Physiol
Regul Integr Comp Physiol 2016;310:R481–92.
34. Cao M, Andersen AD, Van Ginneken C Shen RL, Petersen SO, Thymann
T, Jing J, Sangild PT. Physical activity level is impaired and diet
dependent in preterm newborn pigs. Pediatr Res 2015;78:137–44.
35. Cilieborg MS, Boye M, Thymann T, Jensen BB, Sangild PT. Diet-
dependent effects of minimal enteral nutrition on intestinal function
and necrotizing enterocolitis in preterm pigs. J Parenter Enter Nutr
2011;35:32–42.
36. Nguyen DN, Jiang P, Frokiaer H, Heegaard PM, Thymann T, Sangild
PT. Delayed development of systemic immunity in preterm pigs as a
model for preterm infants. Sci Rep 2016;6:36816.
37. Burrin DG, Stoll B, Guan X, Cui L, Chang X, Holst JJ. Glucagon-
like peptide 2 dose-dependently activates intestinal cell survival and
proliferation in neonatal piglets. Endocrinology 2005;146:22–32.
38. Stoll B, Chang X, Fan MZ, Reeds PJ, Burrin DG. Enteral nutrient
intake level determines intestinal protein synthesis and accretion rates in
neonatal pigs. Am J Physiol Gastrointest Liver Physiol 2000;279:G288–
94.
39. Muyzer G, de Waal EC, Uitterlinden AG. Profiling of complex microbial
populations by denaturing gradient gel electrophoresis analysis of
polymerase chain reaction-amplified genes coding for 16S rRNA. Appl
Environ Microbiol 1993;59:695–700.
40. Agostoni C, Buonocore G, Carnielli VP, De Curtis M, Darmaun D,
Decsi T, Domellof M, Embleton ND, Fusch C, Genzel-Boroviczeny
O, et al. Enteral nutrient supply for preterm infants: commentary
from the European Society of Paediatric Gastroenterology, Hepatology
and Nutrition Committee on Nutrition. J Pediatr Gastroenterol Nutr
2010;50:85–91.
41. Klobasa F, Werhahn E, Butler JE. Composition of sow milk during
lactation. J Anim Sci 1987;64:1458–66.
42. Simmer K. Human milk fortification. Nestlé Nutrition Institute
workshop series 2015;81:111–21.
43. Assad M, Elliott MJ, Abraham JH. Decreased cost and improved feeding
tolerance in VLBW infants fed an exclusive human milk diet. J Perinatol
2016;36:216–20.
44. Commare CE, Tappenden KA. Development of the infant intestine:
implications for nutrition support. Nutr Clin Pract 2007;22:159–73.
45. Vandenplas Y, Alarcon P, Fleischer D, Hernell O, Kolacek S, Laignelet
H, Lonnerdal B, Raman R, Rigo J, Salvatore S, et al. Should partial
hydrolysates be used as starter infant formula? A working group
consensus. J Pediatr Gastr Nutr 2016;62:22–35.
46. Vente-Spreeuwenberg MAM, Verdonk JMAJ, Koninkx JFJG, Beynen
AC, Verstegen MWA. Dietary protein hydrolysates vs. the intact proteins
do not enhance mucosal integrity and growth performance in weaned
piglets. Livest Prod Sci 2004;85:151–64.
47. Burrin DG, Wester TJ, Davis TA, Amick S, Heath JP. Orally
administered IGF-I increases intestinal mucosal growth in formula-
fed neonatal pigs. Am J Physiol Regul Integr Comp Physiol 1996;
270:R1085–91.
48. Xu R-J, Mellor D, Birtles M, Breier B, Gluckman P. Effects of oral IGF-I
or IGF-II on digestive organ growth in newborn piglets. Neonatology
1994;66:280–7.
49. Neu J. Necrotizing enterocolitis. World Rev Nutr Diet 2014;110:253–
63.
50. Mudter J, Neurath MF. Il-6 signaling in inflammatory bowel disease:
pathophysiological role and clinical relevance. Inflamm Bowel Dis
2007;13:1016–23.
51. Eckmann L, Jung HC, Schurer-Maly C, Panja A, Morzycka-
Wroblewska E, Kagnoff MF. Differential cytokine expression by human
intestinal epithelial cell lines: regulated expression of interleukin 8.
Gastroenterology 1993;105:1689–97.
52. Brannigan AE, O’Connell PR, Hurley H, O’Neill A, Brady HR,
Fitzpatrick JM, Watson RWG. Neutrophil apoptosis is delayed in
patients with inflammatory bowel disease. Shock 2000;13:361–6.
53. Burrin DG, Shulman RJ, Reeds PJ, Davis TA, Gravitt KR. Porcine
colostrum and milk stimulate visceral organ and skeletal muscle protein
synthesis in neonatal piglets. J Nutr 1992;122:1205–13.
54. Burrin DG, Davis TA, Ebner S, Schoknecht PA, Fiorotto ML, Reeds
PJ, McAvoy S. Nutrient-independent and nutrient-dependent factors
Downloaded from https://academic.oup.com/jn/article-abstract/148/3/336/4865974 by guest on 07 November 2019
stimulate protein synthesis in colostrum-fed newborn pigs. Pediatr Res
1995;37:593–9.
55. Sternowsky HJ, Heigl K. Tyrosine and its metabolites in urine and serum
of premature and mature newborns: increased values during formula
versus breast feeding. Eur J Pediatr 1979;132:179–87.
56. Rassin DK, Gaull GE, Raiha NC, Heinonen K. Milk protein quantity
and quality in low-birth-weight infants. IV. Effects on tyrosine and
phenylalanine in plasma and urine. J Pediatr 1977;90:356–60.
57. Paul GL, DeLany JP, Snook JT, Seifert JG, Kirby TE. Serum and urinary
markers of skeletal muscle tissue damage after weight lifting exercise.
Eur J Appl Physiol Occup Physiol 1989;58:786–90.
58. Zhao XT, McCamish MA, Miller RH, Wang L, Lin HC. Intestinal
transit and absorption of soy protein in dogs depend on load and degree
of protein hydrolysis. J Nutr 1997;127:2350–6.
59. Rigo J, Verloes A, Senterre J. Plasma amino acid concentrations in
term infants fed human milk, a whey-predominant formula, or a whey
hydrolysate formula. J Pediatr 1989;115:752–5.
60. McGraw NJ, Napawan N, Toland MR, Schulze J, Tulk BM, Krul ES.
Partially hydrolyzed soy protein shows enhanced transport of amino
acids compared to nonhydrolyzed protein across an intestinal epithelial
cell monolayer. J Food Sci 2014;79:H1832–40.
61. Mihatsch WA, Hogel J, Pohlandt F. Hydrolysed protein accelerates the
gastrointestinal transport of formula in preterm infants. Acta Paediatr
2001;90:196–8.
62. Picaud J-C, Rigo J, Normand S, Lapillonne A, Reygrobellet B, Claris
O, Salle BL. Nutritional efficacy of preterm formula with a partially
hydrolyzed protein source: a randomized pilot study. J Pediatr Gastr
Nutr 2001;32:555–61.
63. Maggio L, Zuppa A, Sawatzki G, Valsasina R, Schubert W, Tortorolo
G. Higher urinary excretion of essential amino acids in preterm infants
fed protein hydrolysates. Acta Paediatr 2005;94:75–84.
64. Thymann T, Burrin DG, Tappenden KA, Bjornvad CR, Jensen SK,
Sangild PT. Formula-feeding reduces lactose digestive capacity in
neonatal pigs. Br J Nutr 2007;95:1075.
65. Dinsmore JE, Jackson RJ, Smith SD. The protective role of gastric acidity
in neonatal bacterial translocation. J Pediatr Surg 1997;32:1014–
6.
66. More K, Athalye-Jape G, Rao S, Patole S. Association of inhibitors of
gastric acid secretion and higher incidence of necrotizing enterocolitis
in preterm very low-birth-weight infants. Am J Perinatol 2013;30:
849–56.
67. Zhou Y, Shan G, Sodergren E, Weinstock G, Walker WA, Gregory KE.
Longitudinal analysis of the premature infant intestinal microbiome
prior to necrotizing enterocolitis: a case-control study. PLoS One
2015;10:e0118632.
Human milk fortification in preterm neonates 347