PLOS ONE

RESEARCH ARTICLE

Breadfruit flour is a healthy option for

modern foods and food security
Ying Liu1,2, Paula N. Brown2, Diane Ragone3, Deanna L. Gibson4, Susan J. Murch ID1*
1 Chemistry, University of British Columbia, Kelowna, British Columbia, Canada, 2 Natural Health and Food
Products Research Group, British Columbia Institute of Technology, Burnaby, British Columbia, Canada,
3 Breadfruit Institute, National Tropical Botanical Garden, Kauai, Hawaii, United States of America,
4 Biology, University of British Columbia, Kelowna, British Columbia, Canada

* susan.murch@ubc.ca

a1111111111
a1111111111
a1111111111 Abstract
a1111111111
Breadfruit is a traditional staple crop from Pacific islands with the potential to improve world-
a1111111111
wide food security and mitigate diabetes. Flour produced from breadfruit is a gluten-free,
low glycemic index, nutrient dense and complete protein option for modern foods but basic
scientific knowledge of health impacts of a breadfruit-based diet in animals and humans was
lacking. We designed a series of studies to provide basic and fundamental data on impacts
OPEN ACCESS
of a breadfruit-based diet through an in vitro and in vivo model. Cooked breadfruit flour was
Citation: Liu Y, Brown PN, Ragone D, Gibson DL,
digested through a multi-stage enzyme digestion model to estimate protein digestibility in
Murch SJ (2020) Breadfruit flour is a healthy
option for modern foods and food security. PLoS comparison to wheat flour. Breadfruit protein was found to be easier to digest than wheat
ONE 15(7): e0236300. https://doi.org/10.1371/ protein in the enzyme digestion model. The flour digestions were applied to Caco-2 cells to
journal.pone.0236300 test the cytotoxicity and to measure the immunogenicity through cytokine expression. No
Editor: Edy de Brito, Embrapa Agroindústria significant differences were observed for immune factors and cytokines (IL-4, IL-10, IL-8,
Tropical, BRAZIL TNF-α, IFN-γ) on Caco-2 cells between the breadfruit and wheat groups. A breadfruit-based
Received: April 21, 2020 rodent chow was formulated by substitution of all of the wheat in the standard formulation
Accepted: July 1, 2020 with breadfruit. The diets were isocaloric, nutrient equivalent and used to feed male and
female C57BL/6 mice for 21 days. No sign of malnutrition, discomfort, illness or death was
Published: July 23, 2020
observed among the mice because of the diet. The histology and the cytokine expression of
Copyright: © 2020 Liu et al. This is an open access
the mice ileum from both groups were analyzed and showed similar results. The expression
article distributed under the terms of the Creative
Commons Attribution License, which permits of major bacteria was measured in the colon and showed similar results. Mice fed the bread-
unrestricted use, distribution, and reproduction in fruit diet had a significantly higher growth rate and body weight than standard diet fed mice.
any medium, provided the original author and No negative health outcomes were observed in studies with in vitro or in vivo models and
source are credited.
breadfruit flour is a healthy alternative to other starches for modern foods.
Data Availability Statement: All relevant data are
within the manuscript and its Supporting
Information files.

Funding: The authors are grateful to the Natural

Sciences and Engineering Research Council of
Canada, the Canada Research Chairs program, the Introduction
Canadian Foundation for Innovation, the British
Breadfruit (Artocarpus altilis) is a traditional staple crop that most likely originated in Borneo
Columbia Knowledge Development Fund, the
Canadian Institutes of Health Research, the
[1] and was carried throughout the Pacific Islands with voyaging canoes [1, 2]. With high
University of British Columbia internal grants annual production (>400 kg per tree) [1, 3], and newly developed tissue culture propagation
program and the Breadfruit Institute at the National methods [4, 5], breadfruit is considered to be one of the best candidate plants to distribute to

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 1 / 19

PLOS ONE Breadfruit flour is a healthy option

Tropical Botanical Garden for funding for this undernourished populations in tropical areas for cultivation as a staple [6, 7]. It is often used
program. as a potato substitute in dishes as the fresh fruit can be baked, steamed, boiled, fried, micro-
Competing interests: The authors have declared waved, grilled, and barbecued [1, 2]. Our previous studies demonstrated that breadfruit pro-
that no competing interests exist. tein contains all of the essential amino acids and is especially rich in phenylalanine, leucine,
isoleucine and valine [7]. The most cultivar “Ma’afala” has been planted in nearly 50 countries
in the last decade and is an excellent food resource with higher total essential amino acid con-
tent than other staples including wheat, corn, rice, potato, soybean and yellow pea [1, 7].
Research on breadfruit starch has revealed its advantages over wheat flour on water and oil
holding capacity, swelling power, and viscosity [8, 9]. Moreover, cooking process was found to
cause little alteration in the bioactive compounds of breadfruit and water was the best extractor
compared to organic solvent, making it a promising functional ingredient and a great substi-
tute for wheat in the processed food products [10].
Although breadfruit has been consumed by humans for thousands of years, detailed and
systematic studies of the health impacts of a breadfruit diet have not previously been con-
ducted. Human studies related to breadfruit in the diet have mainly focused on the glycemic
index (GI) measurement. Several researchers [11–13], indicated that breadfruit had a low gly-
cemic index as compared to many common staples such as wheat, cassava, yam and potatoes.
However, the literature is somewhat complicated by a study in 1995 that reported the death of
four male Hooded Lister rats after consumption of breadfruit seed extracts [14] that led to the
classification of breadfruit as a “hazardous plant” by the United States Food and Drug Admin-
istration (FDA). The study indicated that the seeds were obtained from an Artocarpus altilis
tree at the Mayaguez Institute of Tropical Agriculture (Puerto Rico) but the records indicated
that the varieties of breadfruit at the institute were mostly seedless [15] suggesting that the spe-
cies was likely misidentified. Aka et al [16] reported that a raw and partially raw breadfruit diet
induced severe weight loss in male rats but cooked breadfruit resulted in weight gain. Unfortu-
nately, the reported methods were not clear with respect to the incorporation of breadfruit
either as a supplement along with the control diet or processed into the diet and the diet com-
position was not analyzed, making it difficult to understand exactly what the mice were con-
suming [16]. Adepeju et al., [17] also report weight loss in albino weanling rats fed diets with
various ratios of breadfruit, soy and groundnuts but the methods of preparing the diets were
not described and the nutrient composition was not reported.
The objective of the current study was to determine whether a diet containing breadfruit
flour poses any serious health concerns. The specific objectives were (a) to determine the
digestibility of protein in breadfruit flour, (b) to determine the impacts of digested breadfruit
flour on the health and viability of Caco-2 cells; which serves as a model of human intestinal
barrier and (c) to assess the overall growth and health of C57BL/6 mice fed a well-formulated
and -characterized standard diet that substituted breadfruit for wheat. These studies provide
the basic understanding of a breadfruit-based diet for human health.

Materials and methods

Sample preparation
Breadfruit (cultivar: Ma’afala, Artocarpus altilis) fruits were harvested from a single 4-year-
old-tree grown from micro-propagation at the National Tropical Botanical Garden’s McBryde
Garden on Kauai, Hawaii (21˚54’29.47˚N; 159˚30’41.45˚W). No permits were required for
harvest of the samples. Fruit were harvested when mature, peeled, cored and sliced with a
mandolin to 6.35 mm slices. Breadfruit slices were dried overnight at 115˚C in an Excalibur
Dehydrator to complete dryness and shipped to University of British Columbia Okanagan
campus. Flour was prepared by grinding dried breadfruit with a burr-style coffee grinder

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 2 / 19

PLOS ONE Breadfruit flour is a healthy option

(Bunn, Illinois, USA) controlling for excess heat. Wheat flour (all purpose, Robin Hood™) was
purchased from a local supermarket.

Multi-stage enzyme digestion model to mimic human digestion

The multi-stage enzyme digestion model for digesting breadfruit and wheat was modified
based on the previous methods [18–20]. The digestion process was divided by mouth diges-
tion, stomach digestion, and intestinal digestion (Fig 1; S1 Table). Mouth digestion was mim-
icked by mixing 6 g of cooked (water boiled) flour with 6 mL artificial saliva. The pH of the
solution was adjusted to 6–7 with HCl (1N), using a pH meter. The solution was incubated at
37˚C for 5 min in an incubator (Innova144, Eppendorf, Connecticut, USA). The solution was

Fig 1. A schematic representation of the in vitro digestion model.

https://doi.org/10.1371/journal.pone.0236300.g001

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 3 / 19

PLOS ONE Breadfruit flour is a healthy option

then mixed with 12 mL pepsin solution. The pH of the mixture was adjusted to 2–4 with HCl
(6N). The mixture was incubated at 37˚C and shaken consistently at 300 RPM for 2 hours. The
intestinal digestion was performed by adjusting the solution pH to 7.5 with NaOH (1N) and
then mixing the solution with 12 mL of pancreatic solution and 6 mL of bile solution. The
solution was incubated at 37˚C and shaken consistently at 300 RPM for 2 hours. The composi-
tions of saliva, pepsin solution, pancreatic solution and bile solution are listed in supplemen-
tary material. The digestion samples were adjusted to pH 7.5. The digestion extracts were
stored at -80˚C before analysis. A set of digestion reactions was conducted following the same
digestion process (Fig 1) in absence of flour samples to obtain a digestive enzyme solution that
contains all the enzymes and buffers used in the process.

Protein determination in the flours and digestion extracts

Protein was determined by two orthogonal methods to avoid incorrect determinations arising
from unidentified interferences as described by Jones et al. [2]. In brief, 10 mg flour was
extracted in 0.5 mL protein extraction buffer (20 mM HEPES, 150 mM NaCl, 0.3% Tween 20)
with 15 min sonication at room temperature. The tubes were centrifuged for 10 min at 13,000
RPM. The supernatant was saved for the protein determination of flours. The protein content
of flours and digestion extracts were first determined by BCA (bicinchoninic acid) protein
assay (Thermo Scientific, Manassas, USA) and then determined by a modified Lowry Protein
Assay kit (Thermo Scientific), following standard kit instructions. The protein content of
digestion extracts was measured using the same protocol with a 1:5 dilution of the digestion
extract in 0.9% NaCl prior to colorimetric determinations.

Effects of breadfruit digestion on Caco-2 cells: Cell viability

For the cell culture assay, Caco-2 cells were obtained from the American Type Culture
(ATCC1 HTB-37TM) and established according to previous protocol [21]. The digestion
extracts were centrifuged and filtered at 13,000 RPM for 10 min before applying to Caco-2
cells. Caco-2 cells were incubated for 4 hours with four different concentrations (1%, 5%, 10%,
and 50%) of three different digested extracts (digestive enzyme solution, wheat digestion solu-
tion, and breadfruit digestion solution) in the standard growth media. Cell viability was mea-
sured using a standard trypan blue method [22].

Effects of breadfruit digestion on Caco-2 cells: Cytokine expression

Digestion extracts were applied to Caco-2 cells under 4 different conditions; non-stimulated,
LPS (lipopolysaccharide) stimulated, IL-1β (Interleukin 1β) stimualted, and LPS plus IL-1β
stimulated for 24 hours (Table 1). LPS from Escherichia coli 0111:B4 (γ-irradiated, BioXtra,
suitable for cell culture; Sigma-Aldrich) was used for stimulating the cells and prepared as: 1
mg LPS was re-suspended with 1 mL nuclease-free water. 0.4 μL suspended LPS (1 mg/mL)
was mixed with 9.6 μL growth medium and used for the stimulation. IL-1β (animal-compo-
nent free, recombinant, expressed in E. coli, �98% (SDS-PAGE), �98% (HPLC); Sigma-
Aldrich) was used for stimulating the cells and prepared as: 10 μg IL-1β was prepared by re-
suspended 100 μL nuclease-free water. 0.4 μL suspended IL-1β was mixed with 9.6 μL PBS and
used for the stimulation. Non-stimulated Caco-2 cells received no additional stimulations
other than the digestion extracts.
RNA extraction was performed on the Caco-2 cells using RiboZol, following the manufac-
turer’s instruction. The concentration of the RNA samples was measured by a NanoDrop
2000c UV-Vis Spectrophotometer (Thermo Scientific). 1 to 2 μg of RNA was reversed tran-
scribed using an iScript cDNA synthesis kit based on manufacturer’s instructions An

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 4 / 19

PLOS ONE Breadfruit flour is a healthy option

Table 1. Experimental design for testing effects of digestion extracts combined with LPS and/or IL1β stimulation
on Caco-2 cells (n = 3) after 24 hours.
Block Stimulation Digestion extract
1 Non stimulated None
1% Digestive enzyme solution
1% Wheat digestion
1% Breadfruit digestion
2 1000 ng/mL LPS None
1% Digestive enzyme solution
1% Wheat digestion
1% Breadfruit digestion
3 100 ng/mL IL 1β None
1% Digestive enzyme solution
1% Wheat digestion
1% Breadfruit digestion
4 1000 ng/mL LPS+ 100 ng/mL IL 1β None
1% Digestive enzyme solution
1% Wheat digestion
1% Breadfruit digestion
https://doi.org/10.1371/journal.pone.0236300.t001

SsofastTM EvaGreen Supermix was used to prepare RT-qPCR (real time quantitative polymer-
ase chain reaction) mixture. The reference gene was 18S rRNA. Gene expression of TNF-α
(tumor necrosis factor-α), IFN-γ (interferon-γ), MCP-1 (monocyte chemoattractant protein-
1), IL-10, IL-6, iNOS (inducible nitric oxide synthase), IL-8 and IL-4 were examined. All the
primers were synthesized by Integrated DNA Technology (Coralville, USA) (S2 Table). Primer
efficiencies were verified according to the Minimum Information for Publication of Quantita-
tive Real-Time PCR Experiments guidelines [23]. Calculations of the expression value were
performed in the Bio-Rad CFX manager 3.1 (Bio-Rad Laboratories, Ontario, Canada) using
the ΔΔCt method.

Diet design for in vivo mice study

A breadfruit (BF) diet was formulated based on the standard 5LG4 diet (Purina, Gary Summit,
MO, USA) by replacing all the ground wheat and wheat components (~45.5% of 5LG4) with
breadfruit flour (Ma’afala). Analysis of the two finished diets was conducted by a commercial
lab to industry standard. The diets were isocaloric and nutritionally equivalent (S3 Table). To
determine whether the breadfruit flour included non-nutrient phytochemicals that may have
impacted the animal health, the antioxidant content of the diets was determined by standard-
ized DPPH protocols [24]. The BF diet and 5LG4 diet were not significantly different in their
scavenging activity at 15 minutes (S1 Fig), indicating that significant quantities of reactive phy-
tochemicals were not present.

Experimental animals
The animal experiments were performed by the staff at Jackson Laboratories, California, USA
according to standard operating protocols. The animal protocol was approved by the Univer-
sity of California San Diego Institutional Animal Care and Use Committee (IACUC) to ensure
the highest quality animal use and care. Mice (C57BL/6, Jackson Labs) were 7-weeks-old at the
beginning of the experiment and were housed for 3 weeks in positively ventilated polycarbon-
ate cages with HEPA-filtered air at a density of four mice per cage. The mice were ear-notched

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 5 / 19

PLOS ONE Breadfruit flour is a healthy option

for identification. The light source was provided by using artificial fluorescent lighting, creat-
ing a 12 hour light/dark cycle. The room temperature was controlled at 22±4˚C with 50±15%
humidity. A total of 32 mixed sex (half male and half female) mice were randomly assigned
into two groups. Each group had an equal number of female (8) and male (8) mice. Mice were
housed in individually and positively ventilated polycarbonate cages with HEPA filtered air at
a density of 4 mice per cage. The animal room was lighted entirely with artificial fluorescent
lighting, with a controlled 12 h light/dark cycle (6 am to 6 pm light). The normal temperature
and relative humidity ranges in the animal rooms were 22 ± 4˚C and 50 ± 15%, respectively.
The animal rooms were set to have 15 air exchanges per hour. Filtered tap water, acidified to a
pH of 2.5 to 3.0 and standard rodent chow (5LG4) or experimental BF diet were provided ad
libitum. Problems with the breadfruit diet crumbling in the cages meant that it was not possi-
ble to accurately determine food intake or absolute amount of food consumed.

Overall growth evaluation and hematology analysis

Body weight, and food and water intake were measured daily at the same time. Clinical obser-
vations were made daily, examining the general health condition and to see if any skin lesions
developed. At the terminal point, the mice were sacrificed by CO2 asphyxiation at the end of
the third week using approved humane practice. Whole-body composition measurements of
fat, lean, free water, and total water mass was conducted using EchoMRI-100 Analyzer
(EchoMRITM, USA) in the Jackson Laboratory.

Tissue collection
The organs were weighed individually and sent as snap-frozen tissue to the UBC Okanagan.
Feces were collected on days 3, 7, 14, and 21 from each cage, frozen in liquid nitrogen and
then sent to the UBC Okanagan campus. One-third of the ileum sections of the mice were
stored in RNAlater1, an RNA stabilization reagent (Qiagen1), and shipped frozen to the UBC
Okanagan campus. A 3- to 5-mm section of the ileum was fixed in 10% buffered neutral for-
malin with paraffin, sectioned following standard histological methods and then stained by
H&E method. The frozen tissues, feces and RNA later ileum section were stored in -80˚ before
analysis.

Ileum morphology examination

The morphological analysis of the H&E stained ileum slides was conducted, following the
method published by Generoso et al. [25]. Villi pictures (10 pictures × 4 mice) were taken
from each group using a camera (QImaging, Surrey, BC, Canada) attached to a microscope
(Olympus America Inc., PA, USA) and each parameter was measured blindly using ImageJ
1.47v software (Wayne Rasband, National Institutes of Health, Maryland, USA). The following
parameters were measured: distance between villi; villus height and thickness; the crypt depth;
the thickness of lamina propria and epithelium: the length of mucosa, submucosa, and muscu-
laris externa; and the number of goblet cells and red blood cells.

Major cytokine response on ileum

RNA extraction was performed on ileum tissue of the mice, using an RNeasy1 fibrous tissue
mini kit (Qiagen, Germany). The gene expression of TNF-α, IFN-γ, IL-10, IL-6, and iNOS was
examined using RT-PCR technology, using 18S rRNA as a reference gene. The design and test-
ing of the primers used in the study were conducted by Ghosh et al. [26].

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 6 / 19

PLOS ONE Breadfruit flour is a healthy option

Quantification of bacterial groups in the colon

DNA was extracted from the frozen colon tissue of the mice using a QIAamp DNA stool mini
kit (Qiagen). The production of Firmicutes members, Bacteroidetes members, Enterobacteria-
ceae family, Lactobacillus spp., and Bifidobacterium spp. in the colon was quantified by RT-
qPCR, using Eubacteria production as a reference in the analysis. The sequence specificity of
the primers was verified by Baker et al. [21].

Feces protein, total mineral analysis, and fecal occult blood detection
Feces from the same cage and the same day were homogenized and mixed using a CoorsTM
porcelain mortar and pestle (Sigma, USA). The protein was extracted in the protein extraction
buffer before BCA analysis. The mineral content of the feces was calculated by the difference
of weight of the feces before and after incineration at 600˚C for 2 h in an Isotemp programma-
ble muffle furnace (Fisher Scientific, Ottawa, ON), according to AACC 08-03.The fecal occult
blood was detected using a Hemoccult Sensa1 kit (Beckman Coulter, USA) following the
manufacturer’s protocol.

Statistical analysis
All statistical analysis was conducted using JMP1 10.0.0 (SAS Institute, Cary, NC) and Graph-
Pad Prism 5.0 software (GraphPad Software Inc., La Jolla, CA). A series of independent two
sample t-tests or two-way ANOVA analyses were conducted to determine if there were signifi-
cant differences between treatments with a type 1 error rate of 0.05. Graphs were created using
GraphPad Prism 5.0 software (Microsoft Corporation, Santa Rosa, CA).

Results
Protein digestibility
To eliminate the possibility of inaccuracies in the measurement of protein content in the
flour and digestions, two orthogonal protein measures were made at each step. The absolute
values of protein were different between the two methods, but the overall trends were con-
sistent. Overall, about 87% (modified Lowry assay) or 89% (BCA assay) of the breadfruit
protein was fully digested in the in vitro digestion model while 79% (modified Lowry assay)
or 71% (BCA assay) of the wheat protein was fully digested (Fig 2A). After the digestion
reactions, there was significantly more intact protein in the wheat digestion extract than the
breadfruit confirming that more of the breadfruit protein could be digested (Fig 2A). Both
breadfruit and wheat digestions had significantly higher protein than the digestive juice
based on BCA assay; however, this difference was not significant using modified Lowry
assay (Fig 2B).

In vitro evaluation of breadfruit cytotoxicity through cell viability

There was no significant difference in the viability of untreated cells and cells treated for 4
hours with flour digestion extracts (1% of either digestive enzyme solution, wheat digestion or
breadfruit digestion) (Fig 3A). When the digestion extracts were included at 5% in the cell cul-
tures, breadfruit digestion treated cells had significantly increased viability relative to both the
no digestion and wheat digestion treated cells (Fig 3B). When the digestion extract was
included at 10% in the cell culture medium, cell viability was significantly reduced in all treat-
ments (Fig 3C). At 50% of digestion extracts in the cell cultures, the breadfruit digestion extract
had a significantly higher cell viability (42% higher) than the digestive enzyme solution with-
out flours (Fig 3D). Overall, Caco-2 cells exposed to breadfruit digestion extracts had a

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 7 / 19

PLOS ONE Breadfruit flour is a healthy option

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 8 / 19

PLOS ONE Breadfruit flour is a healthy option

Fig 2. Protein content measured by BCA and modified Lowry assays. (A) Wheat and breadfruit flour before and
after digestion. (B) Digestion extracts. N = 3. � represents significant difference at α = 0.05, �� represents significant
difference at α = 0.01, ��� represents significant difference at α = 0.001, using two sample t-test or one-way ANOVA
with Tukey-Kramer Honest Significant Difference (HSD) test. Bar represents the standard error of 6 replicates within
each treatment.
https://doi.org/10.1371/journal.pone.0236300.g002

significantly higher cell viability than cells exposed to the digestive enzyme solution alone.
Increasing the amount of any of the dilution extracts significantly reduced the absolute num-
ber of viable cells and the 1% digestion extracts was most suitable for comparative studies.

Fig 3. Effects of digestion extracts of wheat and breadfruit on cell viability. (A) 1% digestion extracts. (B) 5% digestion extracts. (C) 10% digestion extract. (D) 50%
digestion extracts. Error bars represent the standard error of 3 replicates within each treatment. � represents significant difference at α = 0.05, �� represents significant
difference at α = 0.01, ��� represents significant difference at α = 0.001, using one-way ANOVA with Tukey-Kramer Honest Significant Difference (HSD) test.
https://doi.org/10.1371/journal.pone.0236300.g003

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 9 / 19

PLOS ONE Breadfruit flour is a healthy option

In vitro evaluation of breadfruit immunogenicity through cytokine

expression
Breadfruit flour digestion triggered a cytokine response on Caco-2 cells similar to wheat flour
digestion. All eight cytokines expression were examined, but no significant differences were
observed for IL-4, IL-10, IL-8, TNF-α or IFN-γ between wheat digestion treated group and
breadfruit digestion treated group (S2–S5 Figs). For digestion extract treated unstimulated
cells, 1% breadfruit treated cells had significantly higher MCP-1 expression than wheat-treated
cells and digestive enzyme solution treated cells (Fig 4A). When LPS stimulation was applied

Fig 4. Cytokine responses of Caco-2 cells towards digestion treatments under different stimulations. (A) non stimulated. (B) LPS stimulated (C) IL-1 beta
stimulated (D) LPS plus IL 1 beta. Error bars represent the standard error of 3 replicates within each treatment. � represents significant difference at α = 0.05, ��
represents significant difference at α = 0.01, ��� represents significant difference at α = 0.001, using one-way ANOVA with Tukey-Kramer Honest Significant Difference
(HSD) test.
https://doi.org/10.1371/journal.pone.0236300.g004

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 10 / 19

PLOS ONE Breadfruit flour is a healthy option

with 1% digestion extracts, cells exposed to the breadfruit digestive extract had a higher iNOS
expression than other groups (Fig 4B). When IL-1β stimulation was applied, the 1% wheat
digestion extract stimulated a significant higher IL-6 response than the breadfruit group (Fig
4C). When the LPS and IL-1β stimulations were combined, the expression of IL-6 was signifi-
cantly higher in wheat treated cells than breadfruit treated cells (Fig 4D).

Impact of breadfruit diet on mice overall health and growth

During the three-week-trial, there was no report of death or any abnormal clinical symptoms
for both male and female BF-fed (breadfruit-fed) mice. Both BF- and 5LG4-fed mice showed
similar growth patterns (Fig 5A). Three weeks of the breadfruit diet resulted in a slightly higher
growth rate (7.32% overall, 5.82% male, 8.82% female) in mice than the 5LG4 diet (Fig 5B),
with an equal effect on both male and female mice. Since the breadfruit diet was very crumbly,
the food loss during the trial was significant, and more breadfruit diet was used in the feeding,
therefore the usage cannot be interpreted as the food intake. There was a significant higher
daily water consumption (12.7% overall, 14.4% male, 10.6% female) in mice fed on the bread-
fruit diet compared to mice fed on the 5LG4 diet (Fig 5C & 5D), but there was no interaction

Fig 5. Comparison of growth parameters on mice fed on breadfruit (BF) and 5LG4 diets. (A) Daily body weight. (B) Growth rate.
(C) Daily water consumption. (D) Average daily water consumption. Bars represents the standard error of the measurement among
eight mice from the same group. � represents significant difference at α = 0.05, �� represents significant difference at α = 0.01, ���
represents significant difference at α = 0.001.
https://doi.org/10.1371/journal.pone.0236300.g005

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 11 / 19

PLOS ONE Breadfruit flour is a healthy option

Table 2. Differences in body composition and tissue weight between the breadfruit-fed (BF-fed) mice and 5LG4-fed mice.
Sex 5LG4 diet BF diet Male mice Female mice p
Diet 5LG4 diet BF diet 5LG4 Diet BF Diet Diet Gender Diet� Gender
a b
Body weight (g) 21.0±0.7 22.2±0.7 23.7±0.4 24.7±0.3 18.4±0.3 19.7±0.6 0.0042 <0.0001 0.629
Fat 17.1%±0.7% 15.7%±1.0% 15.0%±0.5%a 12.8%±0.6%b 19.1%±0.7% 18.5%±1.2% 0.0924 <0.0001 0.321
Lean 73.4%±0.9% 72.9%±0.9% 75.4%±1.1% 74.7%±0.7% 71.5%±0.9% 71.1%±1.5% 0.631 0.0016 0.908
Free water 0.5%±0.1% 0.5%±0.1% 0.5%±0.1% 0.4%±0.0% 0.4%±0.1% 0.6%±0.1% 0.741 0.664 0.0692
Total water 60.7%±0.7% 60.4%±0.8% 62.3%±0.9% 62.0%±0.5% 59.0%±0.8% 58.9%±1.3% 0.817 0.0013 0.926
Brain (g) 0.45±0.004 0.44±0.007 0.46±0.003 0.45±0.006 0.44±0.006 0.43±0.013 0.195 0.0461 0.661
Heart (g) 0.12±0.003 0.12±0.005 0.13±0.004 0.13±0.006 0.11±0.004 0.11±0.005 0.580 0.0002 0.662
Kidney (g) 0.25±0.009 0.26±0.008 0.28±0.009 0.29±0.006 0.22±0.008 0.24±0.008 0.232 <0.0001 0.736
Lungs (g) 0.23±0.010 0.22±0.010 0.24±0.011 0.22±0.010 0.21±0.014 0.21±0.018 0.397 0.003 0.0795
Liver (g) 0.94±0.052 0.88±0.066 1.12±0.033 0.93±0.126 0.76±0.033 0.83±0.047 0.505 0.194 0.589
Skin (g) 0.009±0.003 0.007±0.001 0.013±0.006 0.008±0.001 0.005±0.000 0.006±0.001 0.576 0.0134 0.371

Values highlighted in bold and with a superscript letter in each section are significant difference at alpha = 0.05, using 2 sample t test. p value was calculated using two-
way ANOVA analysis

https://doi.org/10.1371/journal.pone.0236300.t002

between diet and gender. Both the growth rate and daily water intake were very significantly
impacted by diets.
At the end of the three-week-trial, the body composition was similar between BF- and
5LG4-diet-fed mice (Table 2; S4 Table). Male mice consuming the BF diet had significantly
higher body weight (4% higher) and lower fat (2.2% less) than male mice consuming the 5LG4
diet (Table 2). There were few significant differences in the weights of various tissues and
organs in response to the diet. Female BF-fed mice had significantly larger duodena and male
BF-fed mice had significantly heavier femurs (S4 Table), while all the other organs, including
the brain, heart, liver, kidney, and skin were not significantly different in weight (Table 2).

Ileum morphology examination

Mice from both diet groups had no significant difference in the villus height and thickness; dis-
tance between the villi; crypt depth; thickness of lamina propria and epithelium; length of mucosa,
submucosa, and muscularis externa; and number of goblet cells and red blood cells (S5 Table).
There was no difference in the ileum morphology between 5LG4-fed and BF-fed mice (Fig 6).

Major cytokine responses on ileum

Cytokines expression, including TNF-α, IL-10, IL 6, and IFN-γ, were examined in the ileum
tissue of the mice and no significant difference was found as a result of consumption of the
5LG4 or breadfruit diet (S6 Fig). Diets had a significant impact on iNOS expression (Fig 7).
Female BF-fed mice had significantly higher iNOS expression than female 5LG4-fed mice (Fig
7); however, the difference was less than 0.5-fold. BF-fed mice had very similar cytokine
responses to 5LG4-fed mice in their ileums.
Gene expression of bacterial groups, Bifidobacterium spp., Lactobacillus spp., and Entero-
bacteriaceae in the colon were compared and no significant differences were found between
5LG4-fed and BF-fed mice (S7 Fig). Both male and female mice fed on the breadfruit diet
showed a higher gene expression of Bacteroidetes than mice fed on the 5LG4 diet (Fig 8A).
Female BF-fed mice showed a higher gene expression of Firmicutes (Fig 8B). Diets had a sig-
nificant impact on the gene expression of Bacteroidetes and Firmicutes. However, the ratio
between Firmicutes and Bacteroidetes was not significantly different between the two diets

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 12 / 19

PLOS ONE Breadfruit flour is a healthy option

Fig 6. Representative histopathology of the ileum tissues stained with haematoxylin and eosin (H&E). (A) 5LG4-fed male mice. (B) BF-fed male mice. (C) 5LG4-fed
female mice. (D) BF-fed female mice. Scale bar is 50 μm.
https://doi.org/10.1371/journal.pone.0236300.g006

(male 5LG4-fed mice: 0.61±0.17, male BF-fed mice: 0.20±0.05, female 5LG4-fed mice: 0.45
±0.12, female BF-fed mice: 0.30±0.13) (Fig 8). Other bacteria groups studied were not signifi-
cant between the diet groups.

Fecal protein, total mineral analysis, and fecal blood detection

Feces collected from the BF-fed mice tended to have a higher protein content than the
5LG4-fed mice (Fig 9A & 9B). Both male and female BF-fed mice showed a significantly higher
fecal protein than 5LG4-fed mice at Day 7 (Fig 9A & 9B). In terms of total mineral content,
however, the breadfruit diet fed mice showed a trend of lower percentage (1–6%) than the
5LG4 diet fed mice, especially at Day 7 and 14, where both male and female BF-fed mice had
significantly lower fecal mineral content than 5LG4 fed mice (Fig 9C & 9D). By the end of the
three-week trial (Day 21), there was no significant difference in the fecal protein or mineral
content between the two diet groups. Regardless of the collection date, gender, and diet, the
fecal blood detection showed that there was no blood in all of the feces collected.

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 13 / 19

PLOS ONE Breadfruit flour is a healthy option

Fig 7. Comparison of iNOS expression between ileums of breadfruit (BF)-and 5LG4-fed mice. � represents a significant difference between the two groups using 2
sample t-test at alpha level 0.05. Middle line denotes media, whiskers are determined by Tukey method, which represent largest value below 75th percentile plus 1.5 times
interquartile distance IQR or lowest value above 25th percentile minus 1.5 times IQR.
https://doi.org/10.1371/journal.pone.0236300.g007

Discussion
The motivation of this project was to contribute to the development of breadfruit as a sustain-
able, environmentally friendly, and high-production crop in developing countries and tropical
areas to feed and support peoples indigenous to these areas. Fundamental understanding of
the health impact of breadfruit digestion and breadfruit diets is necessary and imperative to
establishment of breadfruit as staple or as a functional food in the future.
As the first complete, fully-designed breadfruit diet study, our data showed that breadfruit
diet does not impose any toxic impact on Caco-2 cells or mice. We did not identify any factors
that might indicate potential for allergic or adverse reactions in either humans or animals.
There was no reported death, sickness, or malnutrition of the mice after consuming a bread-
fruit diet for 3 weeks. Similar to our findings, human studies related to breadfruit have not
reported any discomfort or death after consuming breadfruit [11–13, 27]. This is also

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 14 / 19

PLOS ONE Breadfruit flour is a healthy option

Fig 8. The comparison of bacteria expression between ileums of breadfruit diet and 5LG4 diet fed mice. (A) Bacterioidetes. (B) Firmicutes. � , �� , ��� represents
significant difference between the two groups using 2 sample t-test at alpha level 0.05, 0.01, 0.001, respectively. Middle line denotes media, whiskers are determined by
Tukey method, which represent largest value below 75th percentile plus 1.5 times interquartile distance (IQR) or lowest value above 25th percentile minus 1.5 times IQR.
https://doi.org/10.1371/journal.pone.0236300.g008

supported by the long history of breadfruit consumption by Pacific Island populations [1, 28].
Unfortunately, the breadfruit chow crumbled making it difficult to estimate the loss during
feeding and food intake. It is possible the mice consumed less/equal amount of breadfruit diet
but end up with better growth compared to wheat diet feed mice. Also possible those bread-
fruit diet mice grew better simply because they ate more. Reformulation of the chow to avoid
crumbling would require additional ingredients that may also have an effect. Therefore, a dif-
ferent approach is needed to determine whether breadfruit in the diet increases mouse growth.
Breadfruit digestion induced a higher (~1 fold) iNOS gene expression on Caco-2 cells than
the wheat digestion under LPS stimulation. This finding was observed in the in vivo study as
well. There was a slightly higher expression of iNOS (<0.5 fold) in the ileum of mice fed on the
breadfruit diet. LPS is an endotoxin found in Gram-negative bacteria, which are found to be
resident in the intestinal tracts of both humans and rodents [29]. iNOS is a pro-inflammatory
cytokine that can produce nitric oxide, which acts as a cytotoxic agent in the pathological pro-
cess, and can regulate mucosal barrier function [30, 31]. The production of iNOS in normal
mice ileum mucosa has been well demonstrated by Hoffman et al. [31], in which they
explained that normal mice ileum mucosa express iNOS due to the numbers of bacteria pres-
ent and this expression can vary depending on the bacterial state in the ileum. In this study,
only the gene expression of the cytokines was investigated using qPCR and the absolute
amounts of cytokines were not determined. In some cases, the RNA values do not accurately
reflect protein levels [32] but the method is generally considered effective for determination of
the relative differences between samples and treatments [32]
In a preliminary investigation of the microbiome, we examined the major bacterial groups
in the colon, which are closely correlated with the bacterial groups in the ileum. The composi-
tion of the gut microbiota has been previously found to be influenced by diets [33, 34]; there-
fore, it is not surprising that the gut microbe populations in the phyla Bacteroidetes and
Firmicutes were significantly higher in mice fed the breadfruit diet. This finding further sup-
ports the hypothesis that the iNOS expression difference in the breadfruit diet mice and con-
trol mice is a natural consequence of different bacterial populations in the ileum mucosa but

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 15 / 19

PLOS ONE Breadfruit flour is a healthy option

Fig 9. Comparison of fecal protein and total mineral content between breadfruit and 5LG4 diet fed mice. (A) Fecal protein content in male mice. (B) Fecal protein
content in female mice. (C) Fecal total mineral content in male mice. (D) Fecal total mineral content in female mice. Measurements were taken at Day 4, Day 7, Day 14,
and Day 21. � , �� , ��� represent a significant difference between the two groups using 2 sample t-test at alpha level 0.05, 0.01, 0.001, respectively. Bars represent the
standard error of the measurement among eight mice from the same group.
https://doi.org/10.1371/journal.pone.0236300.g009

further research is required to understand the full impacts of breadfruit consumption on the
microbiome.
Overall, these studies support the use of breadfruit as part of a healthy, nutritionally bal-
anced diet. The average consumption of grain in the US is 189 g (6.67 ounce) per day [35].
Consuming 189 g cooked breadfruit can meet up to near 57% of fiber, over 34% protein, vita-
min C and copper, about 28% potassium and manganese, and 5.75–11.5% of iron, calcium
and phosphorus of the daily recommended dietary allowances (RDA) [6, 7].

Supporting information
S1 Table. Composition of saliva, gastric solution, duodenal solution, and bile solution
used in the in vitro digestion model to mimic human digestion.
(DOCX)
S2 Table. RNA sequence, best annealing temperature and primer efficiency for the primers
used in the breadfruit in vitro cell model study.
(DOCX)

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 16 / 19

PLOS ONE Breadfruit flour is a healthy option

S3 Table. Comparison of nutritional value between the breadfruit (BF) diet and 5LG4 diet.
(DOCX)
S4 Table. Comparison of average body composition and tissue weight between the BF-fed
(breadfruit-fed) mice and 5LG4-fed mice.
(DOCX)
S5 Table. Histological examination of ileum morphology.
(DOCX)
S1 Fig. Comparison of scavenging activity percentage at 15 minutes between breadfruit
diet and 5LG4 diet using DPPH methods. Bars represents standard error calculated from the
three replicates.
(DOCX)
S2 Fig. Effect of digestive enzyme solution, wheat digestion and breadfruit digestion on
cytokine mRNA expression in Caco-2 cells without further stimulation for 24 hours. (A)
IL-10. (B). IL 8. (C) TNF-α. (D) IFN-γ. (E). IL-4. (F) iNOS. (G) IL-6.
(DOCX)
S3 Fig. Effect of digestive enzyme solution, wheat digestion and breadfruit digestion on
cytokine mRNA expression in Caco-2 cells with LPS stimulation for 24 hours. (A) IL-10.
(B). IL 8. (C) TNF-α. (D) IFN-γ. (E). IL-4. (F) MCP-1. (G) IL-6.
(DOCX)
S4 Fig. Effect of digestive enzyme solution, wheat digestion and breadfruit digestion on
cytokine mRNA expression in Caco-2 cells with IL-1β stimulation for 24 hours. (A) IL-10.
(B). IL 8. (C) TNF-α. (D) IFN-γ. (E). IL-4. (F) MCP-1. (G) iNOS.
(DOCX)
S5 Fig. Effect of digestive enzyme solution, wheat digestion and breadfruit digestion on
cytokine mRNA expression in Caco-2 cells with LPS and IL-1β stimulation for 24 hours.
(A) IL-10. (B). IL 8. (C) TNF-α. (D) IFN-γ. (E). IL-4. (F) iNOS. (G) MCP-1.
(DOCX)
S6 Fig. Comparison of cytokine expression between ileums of breadfruit (BF)-and
5LG4-fed mice. (A) TNF-α. (B) IL-6. (C) IFN-γ. (D) IL-10.
(DOCX)
S7 Fig. The comparison of bacteria expression between ileums of breadfruit diet and 5LG4
diet fed mice. (A) Enterobacteriacae. (B) Lactobacillus spp. (C) Bifidobacterium spp.
(DOCX)

Author Contributions
Conceptualization: Ying Liu, Diane Ragone, Susan J. Murch.
Data curation: Ying Liu, Susan J. Murch.
Funding acquisition: Diane Ragone, Susan J. Murch.
Investigation: Ying Liu, Deanna L. Gibson.
Methodology: Ying Liu, Paula N. Brown, Deanna L. Gibson, Susan J. Murch.
Project administration: Deanna L. Gibson, Susan J. Murch.

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 17 / 19

PLOS ONE Breadfruit flour is a healthy option

Resources: Paula N. Brown, Diane Ragone, Susan J. Murch.

Supervision: Deanna L. Gibson, Susan J. Murch.
Writing – original draft: Ying Liu, Susan J. Murch.
Writing – review & editing: Deanna L. Gibson, Susan J. Murch.

References
1. Lincoln NK, Ragone D, Zerega NJC, Roberts-Nkrumah LB, Merlin M, Jones AMP. Grow us our daily
bread: A review of breadfruit cultivation in traditional and contemporary systems. In: Warrington I, editor.
Horticultural review. John Wiley & Sons; 2019. pp. 380–384.
2. Jones AMP, Ragone D, Aiona K, Lane WA, Murch SJ. Nutritional and morphological diversity of bread-
fruit (Artocarpus, Moraceae): Identification of elite cultivars for food security. J Food Compos Anal.
2011; 24: 1091–1102.
3. Liu Y, Jones AMP, Murch SJ, Ragone D. Crop productivity, yield and seasonality of Breadfruit (Artocar-
pus spp., Moraceae). Fruits. 2014; 69: 345–361.
4. Murch SJ, Ragone D, Shi WL, Alan AR, Saxena PK. In vitro conservation and sustained production of
breadfruit (Artocarpus altilis, Moraceae): modern technologies for a traditional tropical crop. Naturwis-
senschaften. 2008; 95:99–107. https://doi.org/10.1007/s00114-007-0295-2 PMID: 17710379
5. Shi WL, Saxena PK, Ragone D, Murch SJ. Mass-propagation and bioreactor-based technologies for
germplasm conservation, evaluation and international distribution of breadfruit. Acta Horti. 2007;
757:169–176
6. Turi CE, Liu Y, Ragone D, Murch SJ. Breadfruit (Artocarpus spp.): A Traditional Crop with the Potential
to Prevent Hunger and Mitigate Diabetes in the Tropics. Trends Food Sci Technol. 2015; 45(2): 264–
272.
7. Liu Y, Ragone D, Murch S. Breadfruit (Artocarpus altilis): A source of high quality protein for food secu-
rity and novel food products. Amino Acids. 2015; 47: 847–856. https://doi.org/10.1007/s00726-015-
1914-4 PMID: 25588988
8. Huang S, Martinez M, Bohrer B. The compositional and functional attributes of commercial flours from
tropical fruits (breadfruit and banana). Foods; 2019; 8: 586.
9. Andrade IHP, Otoni CG, Amorim TS, Camilloto GP, Cruz RS. Ultrasound-assisted extraction of starch
nanoparticles from breadfruit (Artocarpus altilis (Parkinson) Fosberg). Colloids and Surfaces A. 2020;
586: 123277.
10. Do Nascimento T, Santos M, de Abreu J, de Almeida I, Da Silva Feijó M, Teodoro A, et al. Effects of
cooking on the phytochemical profile of breadfruit as revealed by higher solution UPLC–MS. J Sci Food
Agric. 2020; 100: 1962–1970. https://doi.org/10.1002/jsfa.10209 PMID: 31846074
11. Bahado-Singh PS, Wheatley AO, Ahmad MH, Morrisson EYStA, Asemota HN. Food processing meth-
ods influence the glycaemic indices of some commonly eaten West Indian carbohydrate-rich foods. Br J
Nutr. 2016; 96: 476–481.
12. Ramdath DD, Issacs CLR, Teelucksingh S, Wolever SMT. Glycaemic index of selected staples com-
monly eaten in the Caribbean and the effects of boiling v. crushing. Br J Nutr 2004; 91: 971–977. https://
doi.org/10.1079/bjn20041125 PMID: 15182400
13. Widanagamage RD, Ekanayake S, Welihinda J. Carbohydrate-rich foods: glycaemic indices and the
effect of constituent macronutrients. Int J Food Sci Nutr 2009; 60: 215–223. https://doi.org/10.1080/
09637480902849195 PMID: 19418327
14. Grant G, More LJ, Mckenzie NH, Dorward PM, Buchan WC, Telek L, Pusztai A. Nutritional and hae-
magglutination properties of several tropical seeds. J Agric Sci. 1995; 124: 437–445.
15. USDA, Germplasm Resources Information Network (GRIN); 2020 [cited 2020 Apr 13]. Database: GRIN
[Internet]. Available from https://www.arsgrin.gov/Pages/Collections#bkmk-ds
16. Aka LO, Obidike LO, Igokwe IR, Ezema WS. The effect of feeding differently prepared breadfruit (Arto-
carpus altilis) on the hematology, serum biochemistry, live and relative organ weight in albino rats. Nige-
rian Veterinary Journal. 2009; 30: 26–34.
17. Adepeju BA, Gbadamosi SO, Omobuwajo TO, Abiodum OA. Effect of breadfruit based complementary
food on weight and haematological parameters in albino rats. Food Science and Quality Management.
2014; 28:78–82.
18. Kiers LJ, Nout MJR, Rombouts MF. In vitro digestibility of processed and fermented soya bean, cowpea
and maize. J Sci Food Agri. 2000; 80: 1325–1331.

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 18 / 19

PLOS ONE Breadfruit flour is a healthy option

19. Tenore CG, Campiglia P, Ritieni A, Novellino E. In vitro bioaccessibility, bioavailability and plasma pro-
tein interaction of polyphenols from Annurca apple (M. pumila Miller cv Annurca). Food Chem. 2013;
141: 3519–3524. https://doi.org/10.1016/j.foodchem.2013.06.051 PMID: 23993515
20. Raiola A, Meca G, Mãnes J, Ritieni A. Bioaccessiblity of deoxynivalenol and its natural co-occurrence
with ochratoxin A and aflatoxin B in Italian commercial pasta. Food Chem Toxico. 2012; 50: 280–287.
21. Baker J, Brown K, Rajendiran E, Yip A, DeCoffe D, Dai C, et al. Medicinal lavender modulates the
enteric microbiota to protect against Citrobacter rodentium-induced colitis. Am J Physiol Gastrointest
Liver Physiol. 2012; 303: G825–G836. https://doi.org/10.1152/ajpgi.00327.2011 PMID: 22821949
22. Strober W. Trypan blue exclusion test of cell viability. Curr Protoc Immunol. 2015; A3B1.
23. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE Guidelines: Mini-
mum Information for Publication of Quantitative Real-Time PCR Experiments. Clin Chem. 2015; 55,
2009, 611–622.
24. Turi CE, Murch SJ. Targeted and untargeted phytochemistry of Ligusticum canbyi: indoleamines,
phthalides, antioxidant potential, and use of metabolomics as a hypothesis-generating technique for
compound discovery. Planta Medica. 2013; 79, 1370–1379. https://doi.org/10.1055/s-0033-1350618
PMID: 23877920
25. Generoso Sde V, Rodrigues NM, Trindade LM, Paiva NC, Cardoso VN, Carneiro CM, et al. Dietary sup-
plementation with omega-3 fatty acid attenuates 5-fluorouracil induced mucositis in mice. Lipids Health
Dis. 2015; 14: 54. https://doi.org/10.1186/s12944-015-0052-z PMID: 26063053
26. Ghosh S, Decoffe D, Brown K, Rajendiran E, Estaki M, Dai C, et al. Fish oil attenuates omega-6 polyun-
saturated Fatty Acid-induced dysbiosis and infectious colitis but impairs LPS-dephosphorylation activity
causing sepsis. PLoS One. 2013; 8: e55468. https://doi.org/10.1371/journal.pone.0055468 PMID:
23405155
27. Ragone D, Cavaletto CG. Sensory evaluation of fruit quality and nutritional composition of 20 breadfruit
(Artocarpus, Moraceae) cultivars. Economic Botany. 2006; 60: 335–346.
28. Ragone D. Breadfruit, Artocarpus altilis (Parkinson) Fosberg. Rome: International Plant Genetic
Resources Institute; 1997.
29. van der Waaij LA, Harmsen HJ, Madjipour M, Kroese FG, Zwiers M, van Dullemen HM, et al. Bacterial
population analysis of human colon and terminal ileum biopsies with 16S rRNA-based fluorescent
probes: commensal bacteria live in suspension and have no direct contact with epithelial cells. Inflamm
Bowel Dis. 2005; 11: 865–871. https://doi.org/10.1097/01.mib.0000179212.80778.d3 PMID: 16189415
30. Aktan F. iNOS-mediated nitric oxide production and its regulation. Life Sci. 2004; 75: 639–653. https://
doi.org/10.1016/j.lfs.2003.10.042 PMID: 15172174
31. Hoffman RA, Zhang G, Nüssler NC, Gleixner SL, Ford HR, Simmons RL, et al. Constitutive expression
of inducible nitric oxide synthase in the mouse ileal mucosa. Am J Physiol. 1997; 272: 383–392.
32. Amesn D, de Visser K, Town T. Approaches to determine expression of inflammatory cytokine. Meth-
ods Mol Biol. 2009; 511: 107–142. https://doi.org/10.1007/978-1-59745-447-6_5 PMID: 19347295
33. Brown K, DeCoffe D, Molcan E, Gibson DL. Diet-induced dysboisis of the intestinal microbiota and the
effects on immunity and disease. Nutrients 2012; 4(8):1095–1119. https://doi.org/10.3390/nu4081095
PMID: 23016134
34. Chan KY, Estaki M, Gibson DL. Clinical consequences of diet-induced dysbiosis. Ann Nutr Metab 2013:
63: 28–40. https://doi.org/10.1159/000354902 PMID: 24217034
35. USDA. What We Eat in America, NHANES 2015–2016, individuals 2 years and over (excluding breast-
fed children), day 1 dietary intake data, weighted; 2017. Database: Food Patterns Equivalents Data-
base (FPED) 2015–2016. [Internet]. Available from: www.ars.usda.gov/nea/bhnrc/fsrg

PLOS ONE | https://doi.org/10.1371/journal.pone.0236300 July 23, 2020 19 / 19