ORIGINAL ARTICLE: GASTROENTEROLOGY

Oral Administration of Surfactant Protein-A Reduces

Pathology in an Experimental Model of
Necrotizing Enterocolitis


Hector D. Quintanilla, yYuying Liu, yNicole Y. Fatheree,
Constance L. Atkins,
z
Syed S. Hashmi, §Joanna Floros, ôFrancis X. McCormack, yJon Marc Rhoads,
and zJoseph L. Alcorn

ABSTRACT
Conclusions: In a rat model of experimental neonatal NEC, oral
Objectives: Necrotizing enterocolitis (NEC) frequently results in significant
administration of SP-A reduces intestinal levels of proinflammatory
morbidity and mortality in premature infants. Others reported that mice
cytokines and TLR4 protein and ameliorates adverse outcomes
deficient in pulmonary surfactant protein-A (SP-A) born and raised in a
associated with gastrointestinal pathologies.
nonhygienic environment succumb to significant gastrointestinal tract path-
ology, and enteral administration of purified SP-A significantly reduced Key Words: immunomodulation, NEC, proinflammatory cytokines, SP-A,
mortality. We hypothesized that oral administration of purified SP-A can TLR4
ameliorate pathology in an experimental model of neonatal NEC.
Methods: Experimental NEC was induced in newborn Sprague–Dawley rat (JPGN 2015;60: 613–620)
pups by daily formula gavage and intermittent exposure to hypoxia. Purified
human SP-A (5 mg/day) was administered by oral gavage. After 4 days,
surviving pups were sacrificed, and intestinal pathology was assessed by
histological examination of distal terminal ileal sections. Intestinal levels of
inflammatory cytokines (IL-1b, IFN-g, and TNF-a) were assessed by
N ecrotizing enterocolitis (NEC) continues to be one of the
most feared complications of prematurity, affecting up to
10% of extremely low birth weight infants (1). Despite improvement
enzyme-linked immunosorbent assay and levels of Toll-like receptor 4 in short-term survival with modern neonatal intensive care, long-term
(TLR4) by Western analysis. mortality for surgical NEC remains high (10%–50%) with increased
Results: Sixty-one percent of the gavaged rat pups that survived to day 4 met risk of neurodevelopmental delay for survivors (1,2). Because the
the criteria for experimental NEC after hypoxia, whereas treatment with SP- etiology of NEC remains incompletely understood, specific treatment
A significantly reduced mortality and assessment of NEC. Intestinal levels of strategies are lacking. Growing evidence points to stress or injury to
proinflammatory cytokines were significantly increased in pups exposed to the intestinal epithelium, combined with inadequate host defense and
hypoxia. Administration of SP-A to pups exposed to hypoxia significantly repair in the immature intestine, which results in intestinal injury
reduced IL-1b and TNF-a levels, but had little effect on elevated levels of mediated by multiple inflammatory mediators (3). It is thought that
IFN-g. SP-A treatment of hypoxia-exposed pups significantly reduced infectious agents are not directly responsible for NEC, but certain
expression of intestinal TLR4, key in NEC pathogenesis. bacterial phenotypes in combination with bowel ischemia, immatur-
ity of host mucosal defense, and enteral cow milk–based feeding may
initiate the development of NEC.
Bacterial pathogens interact with intestinal Toll-like recep-
Received March 21, 2014; accepted December 15, 2014. tors (TLRs), which recognize pathogen-associated molecules such
From the
Department of Pediatrics, Division of Neonatal-Perinatal as lipopolysaccharides (LPSs) (4). NEC in humans and animal
Medicine, the yDepartment of Pediatrics, Division of Pediatric Gastro- models is associated with increased expression and activation of
enterology, the zDepartment of Pediatrics, Pediatric Research Center, intestinal TLR4. Physiological stressors associated with the devel-
University of Texas Health Science Center at Houston, Houston, the opment of NEC (ie, intestinal immaturity and hypoxia) appear to
§Department of Pediatrics, Center for Host Defense, Inflammation and
Lung Disease, Pennsylvania State University College of Medicine,
sensitize the intestinal epithelium to LPS via the upregulation of
Hershey, and the ôDivision of Pulmonary, Critical Care Medicine TLR4 (5). In support of this concept, TLR4-null mice are protected
and Sleep, University of Cincinnati School of Medicine, Cincinnati, OH. from the development of NEC compared with wild-type mice in an
Address correspondence and reprint requests to Joseph L. Alcorn, PhD, experimental model of NEC (6). Increased intestinal expression of
Department of Pediatrics, University of Texas-Houston Medical TLRs and production of inflammatory cytokines appear to precede
School, 6431 Fannin, Suite 3.222, Houston, TX 77030 (e-mail: histological injury in experimental animal models of NEC (7).
[email protected]). Pulmonary surfactant protein-A (SP-A) is best known for its
This study was supported in part by the National Institutes of Health (NIH) role in the function of pulmonary surfactant to reduce alveolar
grant DK56338 (which supports the Texas Medical Center Digestive surface tension by maintaining surfactant homeostasis (8). SP-A,
Diseases Center) to J.L.A., NIH grant HL34788 to J.F.; and the Richard
however, is also a critical component of the innate immune system;
Warren Mithoff Professorship in Neonatal/Perinatal Medicine fellowship
grant to H.D.Q. SP-A assists in clearance of potential pathogens, and has an
The authors report no conflicts of interest. important role in the immunomodulatory properties of surfactant
Copyright # 2015 by European Society for Pediatric Gastroenterology, (9,10). SP-A is highly conserved among mammalian, avian, repti-
Hepatology, and Nutrition and North American Society for Pediatric lian and amphibian species, and lungfish, and polyclonal antibodies
Gastroenterology, Hepatology, and Nutrition raised against rat, mouse, or human SP-A typically cross-react with
DOI: 10.1097/MPG.0000000000000678 SP-A isolated from other species (11). Although expression of SP-A

JPGN  Volume 60, Number 5, May 2015 613

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 Alcorn et al JPGN  Volume 60, Number 5, May 2015

is the highest in respiratory tissues, expression does occur in a by oral gavage feeding of formula and exposure to hypoxia. Briefly,
variety of other tissues that interact with the surrounding environ- litters of newborn rats were separated from their dams after they
ment (ie, stomach, intestine, eye, kidney, uterus, skin), even though suckled for 12 hours. Rat pups were housed in an incubator and then
at much lower levels than found in lung (12,13). Mice deficient in starved for 12 hours before the initiation of formula feeding (day 1,
SP-A have no obvious phenotype, other than greatly enhanced 0.15 mL), 4 times daily (to deliver 200 kcal  kg1  day1) for
susceptibility to pulmonary infections (14,15). Nevertheless, a 3 days using sterile Solomon 22G 35 mm feeding needles (Instech
connection between SP-A and gastrointestinal pathology was Laboratories, Plymouth Meeting, PA). Rat pups were also subjected
reported previously by George et al (16), who found that SP-A to 10 minutes of hypoxia (5% oxygen, 95% nitrogen), 3 times daily
null juvenile mouse pups raised in a nonhygienic environment had for 3 days, in a hypoxic chamber (Billups-Rothenberg, Del Mar,
significant mortality. Rather than the expected lung pathology, CA). The formula consisted of 15 g Similac 60/40 (Ross Pediatrics,
however, mortality was associated with significant gastrointestinal Columbus, OH) in 75 mL of Esbilac canine milk replacement
tract pathology. They also found that oral administration of exogen- (Pet-Ag, Hampshire, IL) containing 1.86 kcal/mL. This technique
ous purified SP-A to newborn SP-A null pups improved survival, was developed by Caplan and Jilling (24) and Nadler et al (25).
suggesting that SP-A has a role in the immunoprotection of neonatal Modification of this well-established technique to induced intestinal
mice. injury in neonatal rats has been successfully used previously in our
The findings by George et al illustrate interesting associ- previous and present studies (7,26,27). Animals were monitored
ations between the known functions of SP-A and the factors that every 3 hours during the study. As in previously published studies,
contribute to the pathogenesis of NEC. SP-A has immunomodula- no analgesia was offered in this experimental NEC model (7,28).
tory properties, either increasing or decreasing proinflammatory or Pups were euthanized on day 4 after NEC protocol to collect tissues.
anti-inflammatory cytokines, depending on the context of the In some cases, pups were euthanized on day 3 if they were in pain,
interaction (10). SP-A is a pattern recognition molecule and binds demonstrating labored respirations, severe abdominal distension, or
to invasive bacteria and viruses, facilitating clearance by macro- gastrointestinal bleeding.
phages (17). SP-A has the capacity to bind LPS and modulate TLR4 To study the effect of SP-A in an experimental animal model
activity through LPS binding or binding directly to the receptor of NEC, pups were randomly divided into 5 groups: dam-fed rats
(18,19). Because many of these functions involve factors implicated (DF, n ¼ 6), formula-fed pups (FF, n ¼ 7 or 8), FF pups with formula
in the development of NEC, we hypothesized that orally adminis- supplemented with SP-A (FS, 5 mg/day, n ¼ 7 or 8), FF rat pups
tered SP-A could play a role in the amelioration of pathologies exposed to hypoxia to induce NEC (FH, n ¼ 10), and FF rat pups
associated with a rat pup model of experimental neonatal NEC. exposed to hypoxia with formula supplemented with SP-A (FHS,
n ¼ 10). To deliver SP-A to each pup, 5 mg of SP-A was diluted into
600 mL of the formula (the amount of formula required for feeding a
METHODS pup daily). Both preparations were prepared at a concentration of
Preparation of Purified SP-A 1 mg SP-A/mL.
The SP-A used in these studies was isolated from human
volunteers with pulmonary alveolar proteinosis, which results in Tissue Harvest
abnormal accumulation of surfactant in their lungs and must
undergo periodic pulmonary bronchoalveolar lavage (BAL) to Following incision of the abdomen, the gastrointestinal tract
remove the material (20). Two independent preparations of human was carefully removed. The small intestine was evaluated visually
SP-A were used in the present study. SP-A was purified from the for typical gross signs of NEC, such as intestinal distension, wall
BAL as described previously (21,22). Briefly, lavage material was hemorrhage, or necrosis. The terminal ileum was defined as the
pelleted and delipidated with isopropyl and 1-butanol. The aqueous distal 20% of the length of the small bowel. The terminal 5 cm of
phase underwent further precipitation, and SP-A was purified by small intestine (ileum) was excised. The distal 1 cm of each sample
column affinity. The purity of the SP-A (1 mg/mL) was assessed was formalin fixed and processed by the Cellular and Molecular
by SDS-PAGE and silver staining. Endotoxin content of the purified Morphology Core Lab (Texas Medical Center Digestive Diseases
SP-A was determined by the Limulus Ambocyte Lysate Assay Center, Houston) and stained with hematoxylin and eosin for
(GenScript USA; Piscataway, NJ): 260 fg endotoxin/mg SP-A histological evaluation. The remaining 4 cm of small intestines
(J.F.) and 180 pg endotoxin/mg SP-A (F.X.M.). It has been reported was immediately frozen in liquid nitrogen and transferred to 80oC
that latent TGF-b has been detected in purified SP-A preparations for further processing for protein and cytokine analysis.
as a result of noncovalent interactions, and this cocontainment has
profound unintended effects on the immunomodulatory properties
of SP-A (23). We assessed the level of TGF-b in the SP-A Histopathological Evaluation of NEC
preparations by immunoblot analysis. Although immunoblot
analysis was sensitive to detect 10 ng of purified, mature TGF- Pathological changes in intestinal architecture were quanti-
b, the SP-A used in these studies did not have detectable TGF-b. fied using a modified NEC scoring system developed and used by
The consistency of our results using these independently isolated the laboratories of Caplan and Jilling (24) and Ford and coworkers
SP-A preparations strongly suggests that differences between the (25). Briefly, histological scores in the ileum were scored by 3
preparations did not have a significant effect on the outcomes blinded evaluators on a scale of 0 (normal), 1 (mild NEC; separation
measured in this study. of the villous core, without other abnormalities), 2 (moderate NEC;
villous core separation, submucosal edema, and epithelial slough-
ing), and 3 (severe NEC; denudation of epithelium with loss of villi,
Experimental NEC Model full-thickness necrosis). The analysis was performed on 3 to 5
sections with 5-mm thick section for each sample similar to other
Use of newborn Sprague-Dawley rat pups (weighing 5–6 g, published studies (24). The final score was based on the worst-
Harlan Laboratories, Indianapolis, IN) was approved by the Animal appearing microscopic field. Animals with histological scores 2
Welfare Committee of the University of Texas Health Science were defined as having NEC. Histological scores were only
Center at Houston (No. HSC-AWC-10-147). NEC was induced obtained from the tissues of live animals.

614 www.jpgn.org

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 JPGN  Volume 60, Number 5, May 2015 SP-A Reduces Pathology in an NEC Model

Tissue Preparation for Western Blot and pups were divided into 5 treatment groups described in the Methods
Cytokine Analysis section. Pup survival was determined after 4 days of treatment
(Table 1). There was a significant difference in mortality between
Ileal tissues were homogenized in 0.4 mL of lysis buffer the different groups (P ¼ 0.019). The highest mortality was
(20 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 1 mmol/L observed in the FH group, in which nearly a third of the rat pups
EDTA, 1 mmol/L EGTA, 1% NP-40, and 1% sodium deoxycholate) died. A post hoc pairwise comparison between the FH group pups
containing protease inhibitors (2.5 mmol/L Na3VO4, 1 mg/mL and the other pups demonstrated that the mortality was significantly
leupeptin, 1 mg/mL aprotinin, and 1 mmol/L PMSF). The homo- higher in the FF pups exposed to hypoxia compared with the FF
genates were centrifuged at 14,000 g for 10 minutes at 4oC after pups exposed to hypoxia and SP-A (35% vs 5% mortality, respect-
incubation on ice for 30 minutes. The collected supernatants (tissue ively; P ¼ 0.043). Furthermore, mortality rates were similar
lysates) were used to measure the total protein using Bio-Rad Dc between the FHS and the FF groups (5% vs 7%, respectively;
Protein Assay (Bio-Rad Laboratories, Hercules, CA). To detect the P ¼ 0.999). Although there were more deaths in the FH group
expression levels of SP-A and TLR4 by Western immunoblot compared with the FF group, this difference failed to achieve
analysis, proteins (50 mL/lane) were treated with antibodies specific statistical significance (35% vs 7%, respectively; P ¼ 0.101).
for SP-A (No. sc-13977; Santa Cruz Biotechnology, Inc, Santa
Cruz, CA), TLR4 (IMG-5031A, Imgenex Corp, San Diego, CA),
and b-actin (No. ab6276; Abcam, Cambridge, MA), respectively, Orally Administered SP-A Reduced Intestinal
for 12 hours at 48C. After incubation with secondary IgG con- Pathology in a Rat Model of Experimental NEC
jugated to horse-radish peroxidase (Bio-Rad; 1:5000) for 1 hour at
room temperature, immunoreactive bands were visualized by We next assessed the effect of orally administered SP-A on
chemiluminescence (ECL Western Blotting System; GE Healthcare the incidence and severity of NEC in a rat model of experimental
Bio-Sciences, Pittsburgh, PA) and quantified on a Storm 840 NEC. The ileum was harvested from surviving animals in each
PhosphorImager (GE Healthcare, Piscataway, NJ). group and processed for histological examination. Shown in
Cytokines were assessed by the enzyme-linked immunosor- Figure 1A are representative histological examples of the sections
bent assay (ELISA) according to the manufacturer’s instructions obtained from the animals. As expected, sections from DF control
(RTA00, RIF00, and RLB00, respectively; R&D Systems, Minnea- pups showed normal, healthy villi, and submucosal structure. The
polis, MN). Each rat intestinal tissue lysate sample was tested in sections isolated from the FF pups had more vacuolated villi
duplicate on the 96-well ELISA plate; the samples with the indicating stress, but the overall structure of the ileum lacked the
percentages of confidence interval between the 2 readings <10 hallmarks of severe NEC. In FS, ileum structure was similar to that
were included in the data analysis. Calibration curves were prepared seen in pups fed formula alone. On the contrary, severe damage was
with a range of 12.5 to 800 pg/mL for TNF-a, 31.2 to 2000 pg/mL apparent in sections isolated from FF pups exposed to hypoxic
for IFN-g, and 31.2 to 1000 pg/mL for IL-1b. Serum cytokine conditions; the villi were short and sparse, and the lamina propria
detection by ELISA is normally expressed as pg/mL; however, was reduced in thickness. Oral administration of SP-A to FF pups
cytokine detection in tissue needs to be normalized based on either exposed to hypoxia resulted in ileum sections that appeared heal-
total protein concentrations in the tissue lysates or the weight of thier, with taller villi and more normal lamina propria.
tissues homogenized, otherwise errors occur because of the vari- The incidence and severity of NEC in the sections of all of
ations in volumes of lysis buffer added for tissue homogenization or the surviving pups after 4 days of formula gavage were quantified
in the size of tissues collected (different tissue weight). Therefore, it using a modified NEC scoring system developed and used by the
was extremely important to normalize the values directly obtained laboratories of Caplan and Jilling (24) and Ford and coworkers (25).
from ELISA (picogram/milliliter) to either picogram/milligram of In this scoring system, the assessment of NEC was graded by 3
total protein or picogram/milligram of tissue weight. Final cytokine separate masked evaluators, and animals with histological scores of
levels were normalized by the total protein concentration (milli- 2 were defined as having NEC; a score of 3 indicates severe NEC
gram/milliliter) and reported as picograms of cytokine per milli- (32). A summary of the analysis of 2 independent experiments is
gram of total protein (picogram/milligram protein). shown in Figure 1B. Not surprisingly, sections prepared from the
DF group showed no evidence of NEC (0 or 1). On the contrary, the
Statistical Analysis hallmarks of NEC were apparent in some of the sections prepared
from all of the other groups. Analysis of the results as a whole
Initial analysis of mortality and histological scores between
the groups was performed using a Fischer exact test. If significant
results were identified in these overall contingency tests, subsequent TABLE 1. Mortality of rat pups exposed to various conditions in an
post hoc pairwise comparisons were performed to identify differ- experimental model of NEC
ences between 2 groups of rat pups. Protein and cytokine levels

Number of rat pups P value
were compared across groups using a Kruskal-Wallis test or a
Mann-Whitney test. If a Kruskal-Wallis yielded significant results, Group Total, n Died, n (%) FF referent FH referent
subsequent post hoc pairwise comparisons were performed using a
Mann-Whitney test. Statistical significance was assumed at DF 12 0 (0) ns 0.029
P < 0.05. All of the analysis was performed using STATA (version FF 15 1 (7) Referent ns
10, StataCorp, College Station, TX). FS 15 4 (26) ns ns
FH 20 7 (35) ns referent
FHS 20 1 (5) ns 0.043
RESULTS
Total 82 13 (16)
Orally Administered SP-A Reduced Mortality in DF ¼ dam-fed; FF ¼ formula fed; FH ¼ formula-fed þ hypoxia;
a Rat Model of Experimental NEC FHS ¼ formula-fed þ hypoxia þ SP-A; FS ¼ formula-fed þ SP-A; NEC ¼
To address the hypothesis, an established rat model of necrotizing enterocolitis; ns ¼ not significant; SP-A ¼ surfactant protein-A.


experimental neonatal NEC was used (7,27,29–31). Newborn rat Post hoc pairwise comparison with specified referent group.

www.jpgn.org 615

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 Alcorn et al JPGN  Volume 60, Number 5, May 2015

A
DF FF FS FH FHS

B
4

P = 0.002
P = 0.047 P = 0.021

3
NEC histological scores

2
NEC

No
NEC
1

DF FF FS FH FHS

NEC (%) 0% 28% 18% 61% 16%

(0/12) (4/14) (2/11) (8/13) (3/19)

FIGURE 1. Effect of orally administered SP-A on ileal histology and assessment of NEC in an experimental rat model of neonatal NEC. A,
Representative hematoxylin and eosin sections of ileum isolated from the various groups. B, Assessment of NEC in sections of ileum of treated
animals. A score of 2 was defined as NEC. Each point represents assessment of a single animal that survived 4 days of treatment; the graph
represents animals from 2 independent experiments with animals treated with independent preparations of purified SP-A. P  0.05, determined
by Fisher exact test, was considered significant. DF ¼ dam-fed; FF ¼ formula-fed; FH ¼ formula-fed with hypoxia; FHS ¼ formula-fed with
hypoxia þ SP-A (original magnification 100); FS ¼ formula-fed þ purified SP-A (5 mg/day); NEC ¼ necrotizing enterocolitis; SP-A ¼ surfactant
protein-A.

indicated significant changes between some of the groups Orally Administered SP-A Reduced Intestinal
(P ¼ 0.019). Although the incidence and severity of NEC was more Levels of Proinflammatory Cytokines in a Rat
apparent in animals not exposed to hypoxia (FF and FS) as
compared with the DF group, these changes were not significantly Model of Experimental NEC
different. As predicted, the group exposed to hypoxia had signifi- Increases in inflammatory cytokines are associated with both
cantly higher numbers of pups with NEC (61%) compared with DF neonatal NEC in humans and experimental animal models of NEC
pups (0%, P ¼ 0.002). More important, oral administration of SP-A (7,33). We sought to determine whether oral administration of SP-A
to FF pups exposed to hypoxia significantly reduced the numbers of reduces the proinflammatory tone of the intestine. We measured 3
pups with NEC (16%) as compared with the FH group (P ¼ 0.021). Th-1 cytokines, IL-1b, IFN-g, and TNF-a, in the intestinal tissue
Administration of SP-A to pups exposed to hypoxia also reduced the lysates from rats in the 5 groups by ELISA. As seen in Figure 2A,
numbers of pups assessed to have NEC (18%) as compared with the intestinal levels of IL-1b were significantly increased in the FH
FF group (28%), but these changes were not significant. group compared with the DF, FF, and FS groups (P  0.004, post

616 www.jpgn.org

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 JPGN  Volume 60, Number 5, May 2015 SP-A Reduces Pathology in an NEC Model

A Mann-Whitney test). In contrast to the results shown with IL-1b and

(picogram/milligram total protein) (picogram/milligram total protein) (picogram/milligram total protein)

TNF-a, administration of SP-A did not significantly reduce these

IIeum IL1-β cytokine levels

40
P = 0.0015
P = 0.002 P = 0.009
elevated levels.
P = 0.004

30 Intestinal Levels of TLR4 Were Decreased Upon

Oral Administration of SP-A in a Rat Model of
20 Experimental NEC
TLR4 activity has a central role in the pathogenesis of NEC
10 by altering the inflammatory milieu and resulting in a proinflam-
matory environment (34,35). It has been reported that SP-A alters
activation and expression of TLR4 in the lung (32,36). We reported
0
B previously that levels of TLR4 protein were increased in FF pups
compared with DF pups in this rat pup model of neonatal NEC;
IIeum TNF-α cytokine levels

P = 0.01
4
however, the change was not significant (7). As seen in Figure 3A, a
representative blot of TLR4 protein in the intestine of DF and FF rat
pups showed similar results in these investigations. To determine
3 whether exogenous SP-A has an effect on expression of TLR4 in the
intestine, levels of TLR4 protein in the ileum were assessed by
2 Western analysis. In Figure 3B is shown a representative blot of the
resulting analysis. As can be seen, TLR4 expression was evident in
the ileal tissue of FF, FS, and FH animals, but expression was
1
decreased in the FHS pups. Densitometry was used to quantify
ND ND TLR4 levels relative to the loaded control, b-actin (Figure 3C). As
0 can be seen, exposure of the pups to hypoxia significantly increased
C P = NS TLR4 expression compared with the FS animals. Compared with
IIeum IFN-γ cytokine levels

the FF group, however, increased levels of TLR4 protein in the

16 P = 0.005 intestine of FH pups did not reach significance. On the contrary,
P = 0.003
P = 0.025
intestinal levels of TLR4 in animals exposed to hypoxia and
12 administered SP-A (FHS) were significantly reduced compared
with the FH group (P ¼ 0.014, Kruskal-Wallis test and post hoc
Mann-Whitney test).
8

4 Intestinal Levels of SP-A Were Not Significantly

Altered in a Rat Model of Experimental NEC
0 SP-A is recognized as one of the major pulmonary surfactant-
DF FF FS FH FHS associated proteins and the majority of the SP-A produced in
mammals is expressed in the lung; however, SP-A is expressed
FIGURE 2. Effect of orally administered SP-A on ileal IL-1b, TNF-a, and in nonpulmonary tissues, even though at lower levels (12,16). This
IFN-g cytokine levels in an experimental rat model of neonatal NEC. A, includes the intestine, in which differential allele-specific expres-
Levels of IL-1b in the ileum of rats treated for 4 days determined by sion of SP-A has been observed (12,13). We previously reported
ELISA. B, Levels of TNF-a in the ileum of rats treated for 4 days that SP-A expression is reduced in a human bronchiolar cell line
determined by ELISA. C, Levels of IFN-g in the ileum of rats treated (H441) upon transient exposure to infant formula (37). Based on
for 4 days determined by ELISA. Data represent box plots of cytokine those results and the fact that this model of neonatal NEC requires
measures in animals shown in B. P  0.05, determined by the Kruskal- formula feeding, we postulated that a reduction in intestinal levels
Wallis test and post hoc Mann-Whitney test, was considered signifi- of SP-A may play a role in the development of NEC in our
cant. DF ¼ dam-fed; ELISA ¼ enzyme-linked immunosorbent assay; experimental rat model. To investigate this possibility, we assessed
FF ¼ formula-fed; FH ¼ formula-fed with hypoxia; FHS ¼ formula-fed levels of SP-A in the ileum or treated animals by Western analysis.
with hypoxia þ SP-A; FS ¼ formula-fed þ SP-A; IFN ¼ interferon; We first sought to determine whether SP-A could be detected in
IL ¼ interleukin; ND ¼ not detectable; NEC ¼ necrotizing enterocolitis; intestinal tissue of rats. As seen in Figure 4A, Western analysis of
SP-A ¼ surfactant protein-A; TNF ¼ tumor necrosis factor. adult rat lung and intestinal tissue demonstrates that SP-A protein
can be detected in intestinal tissue. Western analysis for SP-A in the
treated pups is shown in Figure 4B. The results indicate that
hoc Mann-Whitney test). Oral administration of purified SP-A to although SP-A was detected in the intestine of the newly born
FHS resulted in a 2.5-fold decrease in IL-1b levels compared with rat pups, levels of SP-A did not appear to significantly change
the FH group (P  0.009, post hoc Mann-Whitney test). The effect between the various groups.
of SP-A on intestinal levels of TNF-a was then determined. This
proinflammatory cytokine was only detectable in the ilea of
formula-fed FH (Fig. 2B). The presence of exogenous SP-A sig- DISCUSSION
nificantly reduced intestinal levels of TNF-a by 90% (P  0.01, In 1823, Charles Billard published what may be the first case
post hoc Mann-Whitney test). Not all of the Th-1 cytokines, report of NEC in a ‘‘small, weak infant with infection, inflammation
however, were reduced by the presence of SP-A. As seen in and necrosis of the gastrointestinal tract.’’ NEC continues to be a
Figure 2C, levels of IFN-g were significantly increased in FH as major problem in pediatrics, and despite numerous advances in
compared with the DF, FF, and FS groups (P  0.025, post hoc medical treatment of premature infants, the pathogenesis and

www.jpgn.org 617

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 Alcorn et al JPGN  Volume 60, Number 5, May 2015

A DF FF pathophysiology of NEC remain incompletely understood, and

morbidity and mortality remain unacceptably high. One leading
TLR4
theory holds that NEC is initiated by factors that induce intestinal
β-actin
epithelial damage, which precipitate a dysregulated inflammatory
process driven by activation and overexpression of TLR4 (7,38).
Pulmonary SP-A is a key component of the innate immune system
B TLR4 FF FS FH FHS in the lung that participates in the clearance of pathogenic micro-
(+ control)
organisms and necrotic cells, downregulation of allergic reactions,
and resolution of inflammation (39). We sought to determine
TLR4 whether enteral administration of SP-A could reduce inflammation
in the intestine using a well-established rat model of experimental
β-actin neonatal NEC. Our results demonstrate that prophylactic adminis-
tration of purified human SP-A can reduce morbidity and mortality
in experimental NEC. Gross examination of the ileum sections from
C 4 neonatal rat pups suggests that SP-A can improve damage of the
P = 0.05 P = 0.014 ileum resulting from formula feeding and hypoxic stress, indicating
that oral administration of SP-A reduces the incidence and histo-
logical severity of NEC in a rat model of experimental neonatal
(SP-A/β-actin, arbitrary units)

3
NEC. SP-A also reduces intestinal levels of 2 proinflammatory
IIeum TLR4 protein

cytokines, IL-1b and TNF-a, consistent with its known immuno-

modulatory activity in the lungs. These results suggest that oral
2
administration of SP-A to neonatal rats exposed to hypoxia can
reduce intestinal levels of proinflammatory cytokines associated
with experimental NEC. SP-A is known to modulate expression and
1 activation of TLR4, a key mediator of inflammation during NEC,
and these investigations demonstrate that SP-A reduces intestinal
levels of TLR4 in newborn rats treated to induce experimental
neonatal NEC. This finding suggests that the ability of SP-A to
0
FF FS FH FHS reduce assessment of NEC and the levels of proinflammatory
cytokines in an experimental model of neonatal NEC may involve
FIGURE 3. Effect of orally administered SP-A on ileal levels of TLR4 in expression of TLR4. Our investigations suggest that the well-
an experimental rat model of neonatal NEC. A, Levels of ileal TLR4 described functions of SP-A to modulate inflammation potentially
protein in DF compared with FF rat pups. Levels of ileal TLR4 protein can be used as a means to ameliorate processes and adverse out-
were determined by Western analysis of 2 pups. Shown is a repre- comes associated with NEC.
sentative immunoblot. B, Levels of ileal TLR4 protein determined by The impetus for these investigations was provided by a report
Western analysis. Shown is a representative immunoblot. The TLR4- from George et al (16). In the relatively sterile environment of
positive control (þcontrol) was purchased (Santa Cruz Biotechnology; modern-day laboratories and animal care facilities, mice deficient
human erythroleukemia lysate, sc-2270). C, Densitometric quantifi- in the gene encoding SP-A (SP-A/) displayed few detrimental
cation of TLR4 in the ileum of rats treated for 4 days. Shown are the phenotypes other than enhanced susceptibility to pulmonary infec-
levels of TLR4 relative to b-actin levels (arbitrary units, means stan- tions (14,15). George et al wished to determine whether the effect of
standard error of the mean). Data represent n ¼ 4 to 6 animals from 2 a nonhygienic environment in juvenile mice would predispose the
independent experiments. P  0.05, determined by the Kruskal-Wallis SP-A/ mice to lung pathologies as a result of recurrent infections.
test and post hoc Mann-Whitney test, was considered significant. Indeed, they found a significant increase in mortality of the mice.
DF ¼ dam-fed; FF ¼ formula-fed; FH ¼ formula-fed with hypoxia; FHS ¼ Upon examination, however, the mice lacked the expected path-
formula-fed with hypoxia þ SP-A; FS ¼ formula-fed þ SP-A; NEC ¼ ologies associated with pulmonary infections. Instead, an intestinal
necrotizing enterocolitis; SP-A ¼ surfactant protein-A; TLR4 ¼ Toll-like pathology resembling NEC was observed in the afflicted mice.
receptor 4.

A Rat B
Mouse BAL

Intestine
Lung

DF FF FS FH FHS
kDa kDa
82 82
SP-A SP-A
64 64

37 37
SP-A

β-actin β-actin

FIGURE 4. Expression of SP-A in the lung and ileum of adult and newborn rats. A, Levels of SP-A in adult rat lung and ileum determined by Western
analysis. Shown are Western blots for SP-A and b-actin. Mouse BAL represents BAL material, used as a positive control for SP-A. Arrows indicate
monomer and dimerized forms of SP-A. B, Western analysis of ileal SP-A in rats treated for 4 days. Shown are Western blots for SP-A and b-actin.
BAL ¼ bronchoalveolar lavage; DF ¼ dam-fed; FF ¼ formula-fed; FH ¼ formula-fed with hypoxia; FHS ¼ formula-fed with hypoxia þ SP-A; FS ¼
formula-fed þ purified SP-A; SP-A ¼ surfactant protein-A.

618 www.jpgn.org

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 JPGN  Volume 60, Number 5, May 2015 SP-A Reduces Pathology in an NEC Model

They also found that although SP-A/ pups born from crosses of release of TNF-a in JAWS II dendritic cells (32). It is possible that
SP-A/ dams with SP-A/þ males displayed enhanced mortality, SP-A may interact directly with TLR4 and other receptors to
mortality of SP-A/ pups arising from SP-A/þ dams with prevent activation; alternatively, SP-A may interact directly with
SP-A/ males was not significantly different from SP-A/þ pups, LPS, preventing its binding to TLR4 (49). The fact that SP-A
suggesting transmission of a protective agent, presumably SP-A, participates in the clearance of pathogenic microorganisms and
from the dams. Because SP-A protein has been detected in human necrotic cells may indicate that SP-A reduces colonization by
mammary epithelium (40), they surmised that SP-A may be trans- pathogenic microorganisms (17). As each of these factors have
mitted to the pups via breast milk to provide protection to the gut. been implicated in the pathogenesis of NEC, our findings that a
They then demonstrated that oral administration of SP-A (5 mg/day) pulmonary protein can have a role in gastrointestinal pathology
to SP-A/ pups significantly decreased mortality compared with are not that surprising. Figure 5 is a schematic of a proposed
pups not treated with SP-A. Neither SP-A protein nor RNA, mechanism by which the known functions of SP-A can affect the
however, was found in lactating mammary tissues in this murine factors believed to result in NEC.
model (40a). In addition, it has been reported that milk from SP-A For years, clinicians and researchers have attempted to
null murine dams has significant cytokine alterations compared develop safe and effective strategies to prevent the development
with normal milk with an associated increase in pup mortality (41). of NEC by a variety of agents, such as by IgA supplementation,
Thus, although it seems that orally administered SP-A can transit intravenous dexamethasone, enteral antibiotics, polyunsaturated
through the stomach to provide protection to immature intestine, the fatty acid supplementation, or lactoferrin, and arginine supplement-
precise physiological mechanism responsible for this function of ation. Because of the inability to reproduce results and because of
SP-A remains unclear. potential toxicities, however, these approaches are not widely used.
This study demonstrates an interesting and unexpected inter- The potential beneficial effects of pulmonary surfactant on NEC
section of the mechanisms that are believed to initiate and propagate have been tested previously (50). An artificial surfactant consisting
NEC with the known functions of pulmonary SP-A. Factors that are of the hydrophobic SP-B and SP-C proteins and surfactant lipids
associated with the development of NEC, including hypoxia and the had no protective or therapeutic effect in an animal model of NEC;
exposure to high concentrations of LPS, are known to increase however, SP-A was not included in that preparation. The theoretical
TLR4 expression within the intestinal epithelium (42). Once acti- advantages of the therapeutic use of SP-A in patients with NEC or at
vated, TLR4 initiates a cascade of events, including synthesis of risk for NEC may be its favorable safety profile, its endogenous
proinflammatory cytokines, which in turn produces chemotaxis, expression in the gastrointestinal tract, its lack of systemic effects,
transmigration, and activation of leukocytes. Targeted treatments and its efficacy with an oral route of administration (enteral
(eg, probiotics, pentoxifylline, alkaline phosphatase, anti-TNF-a feeding). Although more research is necessary to define the mech-
antibodies) can reduce expression or activity of proinflammatory anism by which SP-A acts to alter the balance of factors that
cytokines and the incidence and severity of experimental NEC contribute to the development of NEC in the premature host, our
(43–46). These reports suggest that pathogenesis of NEC involves the results suggest that SP-A could be exploited in the prophylaxis of
intersection of LPS, TLR4, and proinflammatory cytokines. In the NEC of the newborn.
lung, SP-A is known to alter the cytokine environment (9,10,39).
SP-A has also been shown to regulate TLR4 expression and activity in REFERENCES
human macrophages and dendritic cells in the lung (32,36). SP-A 1. Jesse N, Neu J. Necrotizing enterocolitis: relationship to innate im-
can bind LPS and antagonizes LPS-enhanced TLR4 activity in munity, clinical features, and strategies for prevention. Neoreviews
alveolar macrophages (18,19,47). In the absence of SP-A, alveolar 2006;7:e143–9.
macrophages exhibit increased activation of NF-kB, because of 2. Vohr BR, Wright LL, Dusick AM, et al. Neurodevelopmental and
reduced levels of the inhibitory protein I-kB (48). SP-A also functional outcomes of extremely low birth weight infants in the
interacts with TLR4-MD2 protein and inhibits the LPS-stimulated National Institute of Child Health and Human Development Neonatal
Research Network, 1993–1994. Pediatrics 2000;105:1216–26.
3. Matin RJ, Fanaroff AA, Walsh MC. Fanaroff and Martin’s Neonatal-
Perinatal Medicine: Diseases of the Fetus Infant. St. Louis: Mosby/
Elsevier; 2011.
4. Caplan MS, Simon D, Jilling T. The role of PAF, TLR, and the
NEC
inflammatory response in neonatal necrotizing enterocolitis. Semin
Pediatr Surg 2005;14:145–51.
5. Hackam DJ, Good M, Sodhi CP. Mechanisms of gut barrier failure in the
pathogenesis of necrotizing enterocolitis: Toll-like receptors throw the
switch. Semin Pediatr Surg 2013;22:76–82.
6. Sodhi CP, Neal MD, Siggers R, et al. Intestinal epithelial Toll-like
receptor 4 regulates goblet cell development and is required for necro-
Proinflammatory tizing enterocolitis in mice. Gastroenterology 2012;143:708–18e1–e5.
LPS TLR4 cytokines
7. Liu Y, Zhu L, Fatheree NY, et al. Changes in intestinal Toll-like
receptors and cytokines precede histological injury in a rat model of
necrotizing enterocolitis. Am J Physiol Gastrointest Liver Physiol
2009;297:G442–50.
8. Kishore U, Greenhough TJ, Waters P, et al. Surfactant proteins SP-A and
SP-A SP-D: structure, function and receptors. Mol Immunol 2006;43:1293–
315.
FIGURE 5. Proposed model for the role of SP-A in ameliorating 9. Pastva AM, Wright JR, Williams KL. Immunomodulatory roles of
experimental NEC. Ovals indicate some of the factors believed to surfactant proteins A and D: implications in lung disease. Proc Am
initiate NEC in the intestine. SP-A is known to dampen the activities of Thorac Soc 2007;4:252–7.
these factors. LPS ¼ lipopolysaccharide; NEC ¼ necrotizing enteroco- 10. Wright JR. Immunoregulatory functions of surfactant proteins. Nat Rev
litis; SP-A ¼ surfactant protein-A; TLR4 ¼ Toll-like receptor 4. Immunol 2005;5:58–68.

www.jpgn.org 619

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.
 Alcorn et al JPGN  Volume 60, Number 5, May 2015

11. Sullivan LC, Daniels CB, Phillips ID, et al. Conservation of surfactant 31. Kelly N, Friend K, Boyle P, et al. The role of the glutathione antioxidant
protein A: evidence for a single origin for vertebrate pulmonary system in gut barrier failure in a rodent model of experimental necrotiz-
surfactant. J Mol Evol 1998;46:131–8. ing enterocolitis. Surgery 2004;136:557–66.
12. Akiyama J, Hoffman A, Brown C, et al. Tissue distribution of surfactant 32. Awasthi S, Madhusoodhanan R, Wolf R. Surfactant protein-A and Toll-
proteins A and D in the mouse. J Histochem Cytochem 2002;50:993–6. like receptor-4 modulate immune functions of preterm baboon lung
13. Lin Z, Wang Y, Zhu K, et al. Differential allele expression of host dendritic cell precursor cells. Cell Immunol 2011;268:87–96.
defense genes, pulmonary surfactant protein-A and osteopontin, in rat. 33. Caplan MS, Sun XM, Hseuh W, et al. Role of platelet activating factor
Mol Immunol 2004;41:1155–65. and tumor necrosis factor-alpha in neonatal necrotizing enterocolitis.
14. Korfhagen TR, Bruno MD, Ross GF, et al. Altered surfactant function J Pediatr 1990;116:960–4.
and structure in SP-A gene targeted mice. Proc Natl Acad Sci U S A 34. Gribar SC, Sodhi CP, Richardson WM, et al. Reciprocal expression and
1996;93:9594–9. signaling of TLR4 and TLR9 in the pathogenesis and treatment of
15. LeVine AM, Whitsett JA. Pulmonary collectins and innate host defense necrotizing enterocolitis. J Immunol 2009;182:636–46.
of the lung. Microbes Infect 2001;3:161–6. 35. Sodhi CP, Shi XH, Richardson WM, et al. Toll-like receptor-4 inhibits
16. George CL, Goss KL, Meyerholz DK, et al. Surfactant-associated enterocyte proliferation via impaired beta-catenin signaling in necrotiz-
protein A provides critical immunoprotection in neonatal mice. Infect ing enterocolitis. Gastroenterology 2010;138:185–96.
Immun 2008;76:380–90. 36. Henning LN, Azad AK, Parsa KV, et al. Pulmonary surfactant protein A
17. Haagsman HP. Interactions of surfactant protein A with pathogens. regulates TLR expression and activity in human macrophages. J Im-
Biochim Biophys Acta 1998;1408:264–77. munol 2008;180:7847–58.
18. Nguyen HA, Rajaram MV, Meyer DA, et al. Pulmonary surfactant 37. Chen MG, Atkins CL, Bruce SR, et al. Infant formula alters surfactant
protein A and surfactant lipids upregulate IRAK-M, a negative regulator protein A (SP-A) and SP-B expression in pulmonary epithelial cells.
of TLR-mediated inflammation in human macrophages. Am J Physiol Pediatr Pulmonol 2011;46:903–12.
Lung Cell Mol Physiol 2012;303:L608–16. 38. Frost BL, Jilling T, Caplan MS. The importance of pro-inflammatory
19. Stamme C, Muller M, Hamann L, et al. Surfactant protein a inhibits signaling in neonatal necrotizing enterocolitis. Semin Perinatol
lipopolysaccharide-induced immune cell activation by preventing the 2008;32:100–6.
interaction of lipopolysaccharide with lipopolysaccharide-binding pro- 39. Nayak A, Dodagatta-Marri E, Tsolaki AG, et al. An insight into the
tein. Am J Respir Cell Mol Biol 2002;27:353–60. diverse roles of surfactant proteins, SP-A and SP-D in innate and
20. Greenhill SR, Kotton DN. Pulmonary alveolar proteinosis: a bench-to- adaptive immunity. Front Immunol 2012;3:131.
bedside story of granulocyte-macrophage colony-stimulating factor
40. Braidotti P, Cigala C, Graziani D, et al. Surfactant protein A expression
dysfunction. Chest 2009;136:571–7.
in human normal and neoplastic breast epithelium. Am J Clin Pathol
21. Beharka AA, Crowther JE, McCormack FX, et al. Pulmonary surfactant
2001;116:721–8.
protein A activates a phosphatidylinositol 3-kinase/calcium signal
40a. George CLS, K Goss, PMcCray, et al. abstract 3756.10, Surfactant
transduction pathway in human macrophages: participation in the up-
protein-D in milk, abstract 3756.10. Paper presented at the Society of
regulation of mannose receptor activity. J Immunol 2005;175:2227–36.
Pediatric Research meeting, 2008.
22. Garcia-Verdugo I, Wang G, Floros J, et al. Structural analysis and lipid-
41. Gidvani MP, Theisen E, Leduc R, et al. Maternal surfactant protein A
binding properties of recombinant human surfactant protein a derived
influences the immunoprotective properties of milk in a murine model.
from one or both genes. Biochemistry 2002;41:14041–53.
Pediatr Res 2014;76:135–41.
23. Kunzmann S, Wright JR, Steinhilber W, et al. TGF-beta1 in SP-A
preparations influence immune suppressive properties of SP-A on hu- 42. Chan KL, Wong KF, Luk JM. Role of LPS/CD14/TLR4-mediated
man CD4þ T lymphocytes. Am J Physiol Lung Cell Mol Physiol inflammation in necrotizing enterocolitis: pathogenesis and therapeutic
2006;291:L747–56. implications. World J Gastroenterol 2009;15:4745–52.
24. Caplan MS, Jilling T. The role of polyunsaturated fatty acid supple- 43. Wu SF, Chiu HY, Chen AC, et al. Efficacy of different probiotic
mentation in intestinal inflammation and neonatal necrotizing entero- combinations on death and necrotizing enterocolitis in a premature
colitis. Lipids 2001;36:1053–7. rat model. J Pediatr Gastroenterol Nutr 2013;57:23–8.
25. Nadler EP, Dickinson E, Knisely A, et al. Expression of inducible nitric 44. Travadi J, Patole S, Charles A, et al. Pentoxifylline reduces the
oxide synthase and interleukin-12 in experimental necrotizing enter- incidence and severity of necrotizing enterocolitis in a neonatal rat
ocolitis. J Surg Res 2000;92:71–7. model. Pediatr Res 2006;60:185–9.
26. Liu Y, Fatheree NY, Dingle BM, et al. Lactobacillus reuteri DSM 17938 45. Rentea RM, Liedel JL, Fredrich K, et al. Intestinal alkaline phosphatase
changes the frequency of Foxp3þ regulatory T cells in the intestine and administration in newborns decreases systemic inflammatory cytokine
mesenteric lymph node in experimental necrotizing enterocolitis. PLoS expression in a neonatal necrotizing enterocolitis rat model. J Surg Res
One 2013;8:e56547. 2012;177:228–34.
27. Liu Y, Fatheree NY, Mangalat N, et al. Lactobacillus reuteri strains 46. Halpern MD, Clark JA, Saunders TA, et al. Reduction of experimental
reduce incidence and severity of experimental necrotizing enterocolitis necrotizing enterocolitis with anti-TNF-alpha. Am J Physiol Gastro-
via modulation of TLR4 and NF-kappaB signaling in the intestine. Am J intest Liver Physiol 2006;290:G757–64.
Physiol Gastrointest Liver Physiol 2012;302:G608–17. 47. Sender V, Lang L, Stamme C. Surfactant protein-A modulates LPS-
28. Dvorak B, Khailova L, Clark JA, et al. Comparison of epidermal growth induced TLR4 localization and signaling via beta-arrestin 2. PLoS One
factor and heparin-binding epidermal growth factor-like growth factor 2013;8:e59896.
for prevention of experimental necrotizing enterocolitis. J Pediatr 48. Wu Y, Adam S, Hamann L, et al. Accumulation of inhibitory kappaB-
Gastroenterol Nutr 2008;47:11–8. alpha as a mechanism contributing to the anti-inflammatory effects of
29. Caplan MS, Kelly A, Hsueh W. Endotoxin and hypoxia-induced surfactant protein-A. Am J Respir Cell Mol Biol 2004;31:587–94.
intestinal necrosis in rats: the role of platelet activating factor. Pediatr 49. Kawai T, Akira S. Pathogen recognition with Toll-like receptors. Curr
Res 1992;31:428–34. Opin Immunol 2005;17:338–44.
30. Dvorak B, Halpern MD, Holubec H, et al. Maternal milk reduces 50. Canpolat FE, Yurdakok M, Ersin SC. Effects of enterally administered
severity of necrotizing enterocolitis and increases intestinal IL-10 in surfactant in a rat model of necrotizing enterocolitis. Neonatology
a neonatal rat model. Pediatr Res 2003;53:426–33. 2012;102:53–8.

620 www.jpgn.org

Copyright 2015 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.