A Strong Impact of Genetic Background on Gut Microflora in Mice.

Abstract

Genetic background affects susceptibility to ileocolitis in mice deficient in two intracellular glutathione peroxidases, GPx1 and GPx2. The C57BL/6 (B6) GPx1/2 double-knockout (DKO) mice have mild ileocolitis, and 129S1/Sv (129) DKO mice have severe inflammation. We used diet to modulate ileocolitis; a casein-based defined diet with AIN76A micronutrients (AIN) attenuates inflammation compared to conventional LabDiets. Because luminal microbiota induce DKO ileocolitis, we assessed bacterial composition with automated ribosomal intergenic-spacer analysis (ARISA) on cecal DNA. We found that mouse strain had the strongest impact on the composition of microbiota than diet and GPx genotypes. In comparing AIN and LabDiet, DKO mice were more resistant to change than the non-DKO or WT mice. However, supplementing yeast and inulin to AIN diet greatly altered microflora profiles in the DKO mice. From 129 DKO strictly, we found overgrowth of Escherichia coli. We conclude that genetic background predisposes mice to colonization of potentially pathogenic E. coli.

Full text

Page 1

SAGE-Hindawi Access to Research
International Journal of Inflammation
Volume 2010, Article ID 986046, 12 pages
doi:10.4061/2010/986046
Research Article
AStrong Impactof GeneticBackgroundon
GutMicroflora inMice
R.Steven Esworthy,1DavidD.Smith,2and Fong-Fong Chu1
1Department of Cancer Biology, Beckman Research Institute of the City of Hope, 1500 Duarte Road, Duarte, CA 91010-3000, USA
2Division of Information Sciences, Beckman Research Institute of the City of Hope, 1500 Duarte Road, Duarte, CA 91010-3000, USA
Correspondence should be addressed to Fong-Fong Chu, fchu@coh.org
Received 29 March 2010; Accepted 9 June 2010
Academic Editor: Gerhard Rogler
Copyright © 2010 R. Steven Esworthy et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
Genetic background affects susceptibility to ileocolitis in mice deficient in two intracellular glutathione peroxidases, GPx1 and
GPx2. The C57BL/6 (B6) GPx1/2 double-knockout (DKO) mice have mild ileocolitis, and 129S1/Sv (129) DKO mice have severe
inflammation. We used diet to modulate ileocolitis; a casein-based defined diet with AIN76A micronutrients (AIN) attenuates
inflammation compared to conventional LabDiets. Because luminal microbiota induce DKO ileocolitis, we assessed bacterial
composition with automated ribosomal intergenic-spacer analysis (ARISA) on cecal DNA. We found that mouse strain had the
strongest impact on the composition of microbiota than diet and GPx genotypes. In comparing AIN and LabDiet, DKO mice were
more resistant to change than the non-DKO or WT mice. However, supplementing yeast and inulin to AIN diet greatly altered
microflora profiles in the DKO mice. From 129 DKO strictly, we found overgrowth of Escherichia coli. We conclude that genetic
background predisposes mice to colonization of potentially pathogenic E. coli.
1.Introduction
Gut microbiota play an important role in several diseases
including inflammatory bowel disease (IBD), type-1 dia-
betes, and obesity [1]. Recent metagenomic studies of the
gut microbiota have shown that bacteria dominate the gut
ecosystem [2]. Between 90 and 98% of bacteria, sampled
either from gut-surface-adherent population or feces, belong
to four bacterial phyla Firmicutes, Bacteroidetes, Proteobac-
teria, and Actinobacteria [3, 4]. Although there is a large
variation in bacterial population in different individuals,
the same bacterial phyla predominate in the stomach, small
intestine, colon, and feces from the same individual [2, 5].
However, some IBD patient guts have decreased bacterial
diversity with depletion of members of Firmicutes and
Bacteroidetes [3, 4]. Since understanding gut microbiota may
provide insight for IBD risk, pathogenesis, and treatment
strategies, there is surprisingly little information on the
microbiota information in mouse models of IBD.
While the metagenomic sequencing study on human
fecal microbial genes has expanded the database of bacterial
genomes deposited in the GenBank, the findings on the
gut microbiota composition also confirm the results from
methodsbasedonbacterial16SribosomalRNA(rRNA)gene
sequences [3–5]. Other noncultured PCR-based methods
have been used to appraise gut microbial composition;
theseincludeautomatedribosomalintergenicspaceranalysis
(ARISA) and terminal restriction fragment length polymor-
phisms [6, 7]. ARISA utilizes conserved 16S and 23S rRNA
gene sequences coupled with variability in the length of the
intergenic spacer to discriminate among bacterial species.
The PCR products are separated by an automated capillary
electrophoresis system with single-nucleotide resolution and
detected by a sensitive laser beam to produce an electro-
pherogram.
ARISA has been used as a crude microbe assay. Metage-
nomic study has estimated that each individual harbors
at least 160 bacterial species and entire cohort harbors
between 1,000 and 1,150 prevalent bacterial species [4]. A
single ARISA primer set on fecal samples only yields 20–
30 consensus products and 100 across all subjects [7–9].
Nevertheless, because ARISA generates a highly reproducible

Page 2

2International Journal of Inflammation
microbiota profile with conventional instrumentation, we
applied this method to assess cecal microbiota in a mouse
IBD model.
We have generated a mouse IBD model by disruption
of two genes encoding for two intracellular glutathione
peroxidases, GPx1 and GPx2 [10, 11]. These GPx1/2-double
knockout (DKO) mice (on a mixed C57BL/6 and 129S1/Sv
genetic background) have microflora-dependent ileocolitis,
since germ-free mice do not have inflammation [11]. Similar
to other mouse IBD models, genetic background has a
profound effect on disease severity in GPx1/2-DKO mice.
B6 DKO mice have mild ileocolitis, the mixed-strain B6;
129DKO mice have more severe disease [12], and 129DKO
mice have the most severe inflammation (from this study).
Since B6 and 129 strains may have different innate immune
responses, which can modulate microflora community [2,
13], we hypothesized that these two strains of mice also have
different gut microbiota.
In addition to genetics, diet also can modulate IBD.
Patients with Crohn’s disease (CD) can be managed by
prescribed diets, which are almost as efficacious as anti-
inflammatory corticosteroids [14–16]. For pediatric CD
patients, the enteral nutrition is preferred to corticosteroids
to avoid adverse effects in European countries [17]. The
major impact of enteral nutrition may rely on changes in gut
microbiota [16]. Thus, we also tested whether diet impacts
on the ileocolitis and microflora in the DKO mice.
In this manuscript, we analyzed the dietary effect on
mouse IBD on both B6 and 129 genetic backgrounds. Based
on the current knowledge on gut microbiota in IBD patients,
we tested whether mouse genetic background, inflammation
(DKO genotype), and diet affected gut microbiota.
2.MaterialsandMethods
2.1. Mice and Diets. Generation of GPx1/2-DKO mice
on the C57BL/6J (B6) × 129S1/SvimJ (129) background
(B6;129) was described previously [10]. B6 colony was
obtained after backcrossing B6;129 mice to B6 for 8
generations. N5 and N10 129 colonies were from B6;129
mice backcrossing to 129 strain for 5 and 10 generations,
respectively. Mice were fed either commercial chows (Lab-
Diet, Richmond, IN) or casein-based defined diets with
AIN76A micronutrients (AIN; Harland-Teklad, Madison,
WI) (Table 1). As specified in the experiments, some AIN
diets were supplemented with brewers’ yeast or inulin
(Oliggo-Fiber Inulin, a gift from Cargill Inc., Minneapolis,
MN).
When on commercial chows, breeders were maintained
on a high-fat LabDiet, and pups weaned to a low-fat LabDiet
at 22 days of age. When on AIN diets, breeders had 10% corn
oil (CO) and pups had 5% CO. Morbidity describes wasting
mice, which were likely to die in the next 24–48 hours, or
with poor health indicated by low body weight, no weight
gain and diarrhea, and unlikely to recover. When describing
diet effects on the pups before weaning, the diet refers to the
breeder diet. All experiments performed on these mice were
approved by City of Hope IUCUC.
2.2. Histology. Distal ileum and the entire colon were
processed for histopathology analysis. Tissues were scored
for inflammation and pathology in a blinded fashion using a
14-point system described previously [11]. Scoring includes
lymphocytes and neutrophils infiltration (0–3 points), ileal
Paneth cell or colonic goblet cell degranulation (0–2
points), epithelium reactivity, including crypt distortion
(0–3 points), inflammatory foci (0–3 points), and apoptotic
figures (0–3 points). The threshold for inflammation corre-
sponds to a score of 6–7 [18].
2.3. Microbiota Census with Noncultured ARISA and Culture
Methods. Cecal microflora were characterized in 22-day-
old pups or younger (16- to 21-day-old) sick mice when
morbidity criteria dictated.
For noncultured ARISA, DNA was isolated from mouse
cecal contents in 1mL TE buffer (10mM Tris-HCl, 1mM
EDTA, pH 8.0), 0.15mL phenol, and 0.2g of 1mm Zirco-
nia/Silica beads (BioSpec Products, Inc., Bartlesville, OK)
using a minibead-beater (BioSpec Products, Inc.) [19].
Approximately 150–300μg DNA was extracted from the
cecal contents of each mouse. The ribosomal integenic DNA
was amplified by PCR using a primer set of ITSF (5?-
GTCGTAACAAGGTAGCCGTA-3?) and ITSReub (Hex-5?-
GCCAAGGCATCCACC-3?) as described [6]. The Hex-
tagged fluorescent reverse primer is used to identify the
products on the DNA sequencing instrument. One μL
of 25μL reaction mixture was analyzed with a capil-
lary DNA analyzer (Hitachi AB model 3730) along with
Genescan1000-ROX standard (Applied Biosystems; City of
Hope Sequencing Core). To identify the DNA amplicons,
the rest of PCR products were separated in agarose gels,
and major DNA bands were excised and cloned into dT-
tailed pCR2.1 (Invitrogen) and sequenced. Cloning total
PCR products without gel separation only yielded one
new sequence. DNA sequence identified was determined by
BLAST (http://www.ncbi.nlm.nih.gov/).
The cecal content was also cultured under aerobic and
microaerobic conditions. The cecal contents were collected
individually; the volume was measured and diluted with
10 volumes of sterile phosphate-buffered saline (PBS). Each
sample was diluted 10,000 times the original sample volume
with PBS and then plated with the original volume on
two Luria-Bertani (LB) plates. One plate was incubated
for 24 hours in aerobic condition, the other for 6-7 days
under microaerobic conditions, both at 37◦C [20]. For
microaerobic culture, plates were placed in a GasPak (Becton
Dickerson, Cockeysville, MD) with an activated Anaerocult
A insert (EM SCIENCE, Gibbstown, NJ). The assembly
was purged with CO2. After counting the colonies, up to
10 colonies from each plate were collected and DNA was
amplified using ARISA primers. The PCR products were
resolved in 1.3% agarose gels and DNA fragments were
extractedwithQiagenGelExtractionKitandsequencedwith
the ARISA primers.
2.4. Statistics. For comparing time-to-event endpoints, such
as survival time on different diets, the results were plotted

Page 3

International Journal of Inflammation3
Table 1: Diet compositions.
Contents
LabDiet1
5061
Mixed
(23.4)
Brewer’s
(1)
(0.43)
Mixed
(5.5)
(3.7)
Mixed
(56)
(5.3)
3.3
LabDiet
5062
Mixed
(20.5)
Brewer’s
(1)
(0.48)
Mixed
(9.6)
(0.4)
Mixed
(53)
(2.7)
3.8
AIN1
5% CO
Casein
(23.4)
—4
AIN
10% CO
Casein
(17.7)
—
AIN + inulin
5–10% CO
Casein
(20–26.5)
—
Brewer’s
1% BY
Mixed
(17.8)
Brewer’s
(1)
(0.3)
Corn oil
(10.3)
(38)
Corn
(15)
(11.5)
3.8
Yeast (BY)2
10% BY
Mixed
(17.7)
Brewer’s
(10)
(0.3)
Corn oil
(10.7)
(35)
Corn
(15)
(11.5)
3.8
Protein
(%)3
Yeast
(% wt)
DL-Methionine, %
Fat
(%)
Sucrose, %
Other CHO5
(%)
Cellulose (fiber), %
Kcal/g
1LabDiets contained crude ingredients and micronutrients similar to AIN-76A (http://www.testdiet.com/). Other diets were defined diets containing AIN-
76A mineral and vitamin mix. All defined diets contained ∼0.002% ethoxyquin as an antioxidant, and had an adequate amount (0.2%) of choline bitartrate
or choline chloride.
2Protein, carbohydrate, and fat in Brewer’s yeast are not included in other contents.
3All % content is by weight.
4A dash means not present.
5CHO means carbohydrate, and Corn means corn starch.
(0.3)
Corn oil
(5.0)
(50)
Corn5
(5)
(18.4)
3.1
(0.3)
Corn oil
(10.0)
(39)
Corn
(15)
(11.5)
3.8
(0.3)
Corn oil
(5–10)
(5.0–38.4)
Corn
(27.5–15.5)
(13.4–6.5)
3.1–3.8
with Kaplan-Meier curves and analyzed with the log-rank
test. Analysis of variance (ANOVA) was used to compare
the means ± standard deviations (SDs). To compare pair-
wise diets versus a control, Dunnett’s correction was made
formultipletestingandtheDunnett-correctedP-valueswere
applied. Each ARISA data panel represents results pooled
from 6 to 21 mice analyzed individually. The electrophero-
grams were digitized, and the results for each group were
averaged using the statistical programming language R [21,
22]. The data were cleaned and passed to the Ribosort
package created for R by Scallan et al. [23]. Ribosort detects
and classifies peak-generating fragments in ARISA data with
a two-pass algorithm [24]. Output from Ribosort contains
information on the ribotype (represented by a specific
size of PCR product) abundances, ribotype proportions,
and sequencer detections. A Euclidean discriminant test
was applied to the final step in ARISA data set analysis.
Finally, a Czekanowski similarity index was run for pair-wise
comparisons from all panels in the ARISA Figures 4, 6, and
8. The quantitative version of the Czekanowski similarity
index is defined as 2W/(A + B), where A and B are the
abundance of species in two given sample conditions and
W is the number of species shared in the two samples. A
conventionforinterpretingtheCzekanowskisimilarityindex
is as follows: index between 0.5 to 0.75 indicates similar
abundances; index between 0.25 to 0.5 indicates different
abundances; index between 0 to 0.25 indicates very different
abundances.
3.Results
3.1. Genetic Background and Diet Had a Profound Effect
on GPx1/2-DKO Morbidity. Mouse strain background has
a big effect on morbidity outcome in GPx1/2-DKO mice
0.2
0.3
0.4
0.5
0.6
Surviving fraction
0.7
0.8
0.9
1
81216202428
Age (days)
32364044
B6 (52)
B6:129 (144)
129 N5 AIN (47)
129 N5 AIN + yeast (37)
129 N10 AIN (133)
129 N10 AIN + inulin (40)
48
Figure 1: Kaplan-Meier curves showing the effect of genetic
background and diet on DKO mouse survival. All mice are on
AIN base diets. Monitored from 8 days of age, the surviving
fraction excludes dead and culled, moribund mice. The numbers in
parenthesisarethenumberofmiceineachgroup.Thedataforyeast
are pools both 1 and 10% as no difference could be distinguished.
There are log rank differences for all interstrain comparisons on
AIN diet: P < .0001. 129 N5, yeast-supplemented AIN (AIN +
yeast) diet versus AIN; P = .0087. 129 N10, inulin-supplemented
AIN (Ain + inulin) diet versus AIN; P = .0003.
(Figure 1). As we noted before the strain difference in
survival on the conventional LabDiet, here we report the
same genetic effect on mouse survival on the AIN diet.
The AIN diet, formulated to mimic LabDiet for calories,

Page 4

4International Journal of Inflammation
0
1
2
3
4
5
#
∗
#
∗
6
7
8B6
2250
Age (days)
Inflammation/pathology score
(a)
0
2
4
6
8
10
12
Age (days)
B6; 129
2240
#
∗
#
∗
Inflammation/pathology score
(b)
0
1
2
3
4
5
#
∗
#
∗
6
7
8
9
10
129 N5
22
AIN lleum
LabDiet lleum
Age (days)
AIN colon
LabDiet colon
40
Inflammation/pathology score
(c)
∗
∗
0
2
4
6
8
10
12
129 N10
22
AIN lleum
LabDiet lleum
Age (days)
AIN colon
LabDiet colon
Inflammation/pathology score
(d)
Figure 2: Effect of diet on ileum and colon histology at weaning and peak pathology by strain in GPx1/2-DKO mice. Mice on LabDiet or
AIN diet were analyzed at 22 days of age or upon signs of morbidity, and at the peak of ileitis, which is 40 days for B6;129 or 129 N5 and
50 days for B6. 129 N10 mice were only analyzed at 22 days of age due to high morbidity. The AIN diet alleviated ileitis significantly in all
strains at both 22 days of age and at peak inflammation (the “∗” sign indicates P < .011). The ileal inflammation/pathology scores increased
significantly from 22 days to peak of inflammation in all strains (#P < .034). A diet-associated difference in colitis was only observed in
22-day-old 129 N10 pups (∗P < .001).
macro- and micronutrients, maintained better health than
LabDiets for the DKO mice.
B6 GPx1/2-DKO mice maintained fairly good health on
either a LabDiet or AIN diet. The surviving fraction of B6
DKO mice at 45 days on the AIN is 100%, which is virtually
the same as mice on LabDiet with 98% survival compared at
45 days of age (Figure 1 and data not shown). The B6;129
mixed-strain DKO mice had marginal health, with 97%
survival on AIN diet compared to 85% on LabDiet at 45
days of age. As expected, the 129DKO mice had poor health.
Eighty-five % of 129 N5 DKO on AIN diet survived 40 days
when only 65% survived on LabDiet. Only 47% of 129 N10
on AIN diet survived 40 days when merely 5% of 129 N10 on
LabDiet made it (P < .0001, Log rank; differences for strains
between AIN diet and LabDiet, except B6).
For 129DKO mice, the early morbidity is associated with
colitis. The postweaning morbidity in B6 and B6;129DKO
mice is correlated with later developing ileitis and rarely
involves diarrhea. Disease in the 129 N5 and N10 DKO starts
before weaning based on symptoms of wet tail and runting
as early as 11 days of age on AIN diet.
3.2. Morbidity of GPx1/2-DKO Mice Is Correlated with
Gut Inflammation. Morbidity at any time appears to be
reflection of acute inflammation, reflected in the inflamma-
tion/pathology scores of 6 or greater [18]. B6 DKO mice
had mean ileal pathology score around 5.5 at 50 days of
age, the peak of inflammation in this strain (Figure 2). 129
N10 DKO mice had mean ileal and colonic pathology scores
of 6.5 and 9, respectively, at 22 days of age. The pathology
scores correlate with morbidity. Typical histopathology in
129 at 22 days of age showed acute inflammation in the
cecum and distal colon with frequent skipping or less severe
inflammation in the proximal colon (Figure 3). In morbid
mice, disease often extended into the proximal colon. By
contrast, B6 mice had almost no pathology in the cecum,
proximal, and distal colon.
Diet significantly modified ileitis severity in all genetic
backgrounds, that is, B6, B6;129, as well as 129 N5 and
N10 DKO mice (Figure 2). Only in 129 strains was a dietary
effect on colitis observed. Non-DKO mice did not have gut
pathology.
3.3. Yeast and Inulin Supplementations to AIN Diet Increased
Morbidity on 129 N5 and N10 DKO Mice. Since LabDiet has
1% brewer’s yeast, and yeast antigens are associated with
some CD, we tested its effect in 129 N5 DKO mice [25].
Feeding 129 N5 DKO mice with AIN diet supplemented

Page 5

International Journal of Inflammation5
129 N10
Cecum
(a)
B6
(b)
Proximal colon
(c)(d)
Distal colon
(e)
1mm
(f)
Figure 3: Representative histopathology in the colon of 129 N10 and B6 mice on AIN at 22 days of age. Slides were stained with hematoxylin
and eosin. By 22 days of age, 90% of 129 N10 DKO mice have disease signs of diarrhea and many are lethargic. B6 DKO mice are without
these signs. This is correlated with histopathology in the cecum and colons of these mice. 129 mice show inflammation and distortion of
the cecum and distal colon with frequent crypt abscesses in the distal colon. The inflammation frequently skips the proximal colon (shown
here). In severe cases, where lethargy has progressed to wasting, the proximal colon can be involved. The cecum and colon of B6 mice show
mucin depletion and mild distortion due to apoptosis and proliferation.
with 1% and 10% brewer’s yeast increased their morbidity
(P = .0087)(Figure 1 shows pooled, no difference 1% versus
10%). The morbidity curve on yeast-containing AIN diet
resembles that obtained with LabDiet (data not shown).
Although the pups experienced more pronounced diarrhea,
the pathology score was not elevated in the N5 DKO mice
on the yeast-containing diet. The median colon pathology
score was 3 at 22 days of age on both diets (P = .58), and 4.5
versus 4 for mice on yeast-containing AIN versus AIN diets
at 40 days of age (P = .9; Mann-Whitney test). Thus, yeast-
supplementation had an adverse effect on the DKO mice by
increasing morbidity through intensifying diarrhea without
exacerbating colitis.
Inulin, a nondigestible fructooligosaccharide is a food
fiber with prebiotic properties, which may have prophylactic
or therapeutic potential for IBD [1, 26]. We supplemented
5% inulin in AIN diet to 129 N10 DKO mice on the premise
that it may retard the morbidity by fostering the colonization
of probiotic microflora. Disappointedly, supplementing 5%
inulin to AIN diet increased morbidity in 129DKO pups
(Figure 1; P = .0003). However, inulin did not affect colon
inflammation; both diets yielded a median disease score of
7, P = .28 (Mann-Whitney test). Thus, inulin also had an
adverse effect on DKO mice by increasing morbidity without
affecting colitis.
3.4. Diet Modified Cecal Microbiota Content in Wild-Type
(WT) B6 but Not DKO Mice. Since diet can alter gut
microflora[16],andmicrofloraareessentialforIBD,weused
ARISA to profile microbiota in these mice. We chose cecal
contenttoanalyzemicrobiotaforthefollowingreasons.First,