The synthetic food dye, Red 40, causes DNA damage, causes colonic inflammation, and impacts the microbiome in mice

Abstract

The incidence of colorectal cancer (CRC) among young people has been on the rise for the past four decades and its underlying causes are only just starting to be uncovered. Recent studies suggest that consuming ultra-processed foods and pro-inflammatory diets may be contributing factors. The increase in the use of synthetic food colors in such foods over the past 40 years, including the common synthetic food dye Allura Red AC (Red 40), coincides with the rise of early-onset colorectal cancer (EOCRC). As these ultra-processed foods are particularly appealing to children, there is a growing concern about the impact of synthetic food dyes on the development of CRC. Our study aimed to investigate the effects of Red 40 on DNA damage, the microbiome, and colonic inflammation. Despite a lack of prior research, high levels of human exposure to pro-inflammatory foods containing Red 40 highlight the urgency of exploring this issue. Our results show that Red 40 damages DNA both in vitro and in vivo and that consumption of Red 40 in the presence of a high-fat diet for 10 months leads to dysbiosis and low-grade colonic inflammation in mice. This evidence supports the hypothesis that Red 40 is a dangerous compound that dysregulates key players involved in the development of EOCRC.

Full text

Toxicology Reports 11 (2023) 221–232
Available online 6 September 2023
2214-7500/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
The synthetic food dye, Red 40, causes DNA damage, causes colonic
inammation, and impacts the microbiome in mice
Qi Zhang
a
, Alexander A. Chumanevich
a
,
*
, Ivy Nguyen
a
, Anastasiya A. Chumanevich
a
,
Nora Sartawi
a
, Jake Hogan
a
, Minou Khazan
a
, Quinn Harris
a
, Bryson Massey
a
,
Ioulia Chatzistamou
b
, Phillip J. Buckhaults
a
, Carolyn E. Banister
a
, Michael Wirth
c
,
James R. Hebert
d
, E. Angela Murphy
b
, Lorne J. Hofseth
a
a
Department of Drug Discovery and Biomedical Sciences, College of Pharmacy, University of South Carolina, Columbia, SC 29208, USA
b
Department of Pathology, Microbiology, and Immunology, School of Medicine, University of South Carolina, Columbia, SC, USA
c
Department of Biobehavioral Health & Nursing Science, College of Nursing, University of South Carolina, Columbia, SC 29208, USA
d
Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA
ARTICLE INFO
Handling Editor: Dr. L.H. Lash
Keywords:
Red 40
DNA damage
Colorectal cancer
Microbiome
P53
ABSTRACT
The incidence of colorectal cancer (CRC) among young people has been on the rise for the past four decades and
its underlying causes are only just starting to be uncovered. Recent studies suggest that consuming ultra-
processed foods and pro-inammatory diets may be contributing factors. The increase in the use of synthetic
food colors in such foods over the past 40 years, including the common synthetic food dye Allura Red AC (Red
40), coincides with the rise of early-onset colorectal cancer (EOCRC). As these ultra-processed foods are
particularly appealing to children, there is a growing concern about the impact of synthetic food dyes on the
development of CRC. Our study aimed to investigate the effects of Red 40 on DNA damage, the microbiome, and
colonic inammation. Despite a lack of prior research, high levels of human exposure to pro-inammatory foods
containing Red 40 highlight the urgency of exploring this issue. Our results show that Red 40 damages DNA both
in vitro and in vivo and that consumption of Red 40 in the presence of a high-fat diet for 10 months leads to
dysbiosis and low-grade colonic inammation in mice. This evidence supports the hypothesis that Red 40 is a
dangerous compound that dysregulates key players involved in the development of EOCRC.
1. Introduction
Colorectal cancer (CRC) in the young (<50 years old) - termed Early
Onset Colorectal Cancer (EOCRC) - has been on the rise over the past 40
years [1–6]. Although standard risk factors such as obesity, alcohol, and
smoking are linked to EOCRC in epidemiological studies [1–3,5,7–9];
there remains a gap in our understanding as to why we are seeing a rise
in CRC in otherwise healthy young people in their 20’s and 30’s.
Diet plays a pivotal role in inuencing the risk of colorectal cancer, a
signicant public health concern. Research suggests that a diet high in
processed meats, red meats, and saturated fats may elevate the risk of
developing colorectal cancer [10]. Conversely, diets rich in ber, whole
grains, fruits, and vegetables offer protective benets by promoting
regular bowel movements, maintaining gut health, and reducing
inammation. Furthermore, certain nutrients like calcium, vitamin D,
and antioxidants found in various foods have been associated with a
decreased risk of colorectal cancer. These ndings underscore the
crucial link between dietary choices and colorectal cancer development,
emphasizing the potential for preventive strategies and dietary modi-
cations to contribute signicantly to public health efforts aimed at
reducing the burden of this disease.
The prevalence of ultra-processed, “westernized” diets that are
typically high in fat and simple carbohydrates has increased signi-
cantly in the past 40 years [11–13], concomitant with the rise in EOCRC.
Abbreviations: ADI, Acceptable Daily Intake; CRC, colorectal cancer; EOCRC, early-onset colorectal cancer; HFD, high-fat diet; HFDR, High-Fat Diet +1x
Acceptable Daily Intake (ADI) of Red 40; HFD2R, HFD +2x Acceptable Daily Intake (ADI) of Red 40; IRS, Immunoreactivity Score; LFD, low-fat diet; LFDR, low-fat
diet +1x Acceptable Daily Intake (ADI) of Red 40; LFD2R, LFD +2x Acceptable Daily Intake (ADI) of Red 40.
* Corresponding author.
E-mail address: chumanev@sc.edu (A.A. Chumanevich).
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
https://doi.org/10.1016/j.toxrep.2023.08.006
Received 21 April 2023; Received in revised form 21 August 2023; Accepted 31 August 2023

Toxicology Reports 11 (2023) 221–232
222
This diet consists of foods that are heavily processed, include synthetic
chemicals; are low in ber, vitamins, minerals, and phytochemicals; and
are often high in added sugar [14,15]. Indeed, rigorous science has
consistently shown that a highly processed, high-fat, westernized diet
can drive inammation and gut dysbiosis [16–18], and may increase risk
of EOCRC [9,19–24]. There remains a gap, however, to the specic in-
gredients and their mechanisms toward EOCRC. Because of the preva-
lence of Red 40 in highly processed, westernized diets; it is reasonable to
investigate the impact of Red 40 - that often comes with the consump-
tion of a high-fat, westernized diet - on colon health in relation to
carcinogenesis.
Synthetic dyes are prevalent in the global food supply chain. Three
dyes [Allura Red (a.k.a. “Red 40”), Tartrazine (“Yellow-5”), and Sunset
Yellow (“Yellow-6”)] account for 90 % of all dyes used in food in the
USA; and Red 40 is by far the most common [25]. Alarmingly, 94 % of
people over 2 years old in the USA consume Red 40 [26]; and over 40 %
of foods marketed toward children in the USA contain such dyes [27]. It
is used extensively in processed foods as a coloring for beverages, frozen
treats, powder mixes, gelatin products, candies, icings, jellies, spices,
dressings, sauces, and baked goods. It is critical we better understand the
interaction of Red 40 with players involved in carcinogenesis.
The lack of rigorous prior research combined with high levels of
human exposure to westernized diets containing Red 40, underscores
the importance of researching this subject with current technologies,
platforms, and appropriate animal models. We posit that the synthetic
dye Red 40, acting as a foreign substance, induces a subtle and low-
grade inammatory response specic to the colon and rectum. This
chronic inammation may contribute to the development of CRC,
particularly in the distal colon and rectum. Supporting the premise of
this study, there evidence shows that Red 40 interacts with inamma-
tory components, impacts the microbiome, and can be perceived as
foreign by the body [28–33]. Indeed, it has recently been shown that
Red 40 drives colitis under experimental conditions by several groups
[34–36], and synthetic food dyes elevate inammatory cytokines and
modulate protein and gene expression related to inammation [28,37,
38]. Here, we present data consistent with the hypothesis that Red 40
damages DNA in vitro and in vivo; and that a westernized diet combined
with Red 40 causes dysbiosis, functional mutations, and low-grade
inammation in the distal colon and rectum.
2. Materials and methods
2.1. Cell line and reagents
Human colorectal carcinoma cell line HCT 116 were maintained in
media recommended by ATCC supplemented with 10 % New Born Calf
serum (NBCS) (Biouids, Rockville, MD), penicillin (10 U/ml, Biouids)
and streptomycin (10
μ
g/ml, Biouids) at 37 ◦C in a humidied chamber
with 5 % CO
2
atmosphere. Experiments with Red 40 were carried out by
treating the cells with indicated concentrations of Red 40 dissolved in
media. Red 40 (Allura Red AC) was purchased from Sigma-Aldrich.
2.2. Mouse experiments
We chose the A/J mice mouse model because these mice tend to
acquire more distal colon and rectal cancers with a carcinogen [39,40]
(as is the case for EOCRC) and are sensitive to chemical carcinogenesis
[39]. No carcinogen was used in our experiments because we wanted to
examine the impact of a HFD and/or Red 40 on the normal colon.
Three-week old, female mice of the A/J strain (Stock No: 000646,
Jackson Lab) were purchased from Jackson Laboratory (Bar Harbor,
ME). Animals were cared for in accordance with protocols approved by
the Institutional Animal Care and Use Committees of the University of
South Carolina. Care and use of animals were overseen by the Depart-
ment of Laboratory Animal Resources (DLAR) of the University of South
Carolina under the direction of a veterinarian. The DLAR is fully
accredited by the Association for Assessment and Accreditation of Lab-
oratory Animal Care International, is registered with the U.S. Depart-
ment of Agriculture (56-R-003) and has an active letter of Assurance of
Compliance on le at the NIH. Mice were maintained at 22
 C, with 12 h
dark, 12 h light cycle. Female mice were maintained on rodent chow for
one week and the experiment was started when the mice were four
weeks old. Mice were randomly assigned to experimental groups. For
the 10-month experiment, Red 40 was delivered in the drinking water ad
libitum; and volume consumed was measured every other day. The
mouse-equivalent of the human Acceptable Daily Intake (ADI, 7 mg/kg
daily) is 86 mg/kg daily. With the calculation that mice consume 2 ml
daily, mice were treated with the ADI and 2x ADI Red 40 and/or a LFD or
HFD for 10 months. For the in vivo experiment on DNA damages, Red 40
was administered orally (e.g., oral gavage) as one time dose (either ADI
or 2xADI).
2.3. Diets
We designed a diet that resembles a western-style diet with a high fat
content (‘high-fat diet, or HFD’). The HFD (Bio Serv® #F3282) con-
tained 20.5 % protein, 36.0 % fat, 35.8 % carbohydrate, 141 g/kg total
saturated fatty acids, 202.2 g/kg total unsaturated fatty acids, and 5.49
kcal/g total calories) (Supplemental Table 1). The corresponding low-
fat diet (LFD) was the control diet; has the same protein content; and is
matched in micronutrients (Bio Serv® #F4031, protein 20.5 %, fat 7.1
%, carbohydrate 61.5 %, total saturated fatty acids 27 g/kg, total un-
saturated fatty acids 40 g/kg, total caloric 3.39 kcal/g) (Supplemental
Table 1). Animal diets were kept at 4
 C. Food was re-lled every other
day for all the animals, which keep the food fresh and available for the
animals’ appetite. The drinking bottles were changed bi-weekly. Animal
diets and water were checked daily; and mice weight, diarrhea, and
hemoccult (fecal blood) were monitored twice a week. Food and drink
consumption were monitored twice a week; fecal samples were collected
once a month. Animal weight loss, diarrhea, hemoccult indicate the
disease condition in the intestinal tract. For example, hemoccult and
anemia suggesting pathological problems in the mucosa, which may
reect early age of onset, with carcinogenesis potentially starting in the
mucosa.
2.4. Macroscopic foci counting
Colon macroscopic foci were counted under a dissecting microscope.
The number and location of foci in the GI tract were recorded and
imaged. After all the necessary observations were made, the colon was
swiss-rolled and xed with 10 % neutral buffered formalin for 20 h and
embedded in parafn. Tissue sections were prepared and stained with
hematoxylin and eosin (H&E) for histological examination.
Table 1
Quantication of organoids derived in different culture environments. Data are
shown as mean ±SD.
Culture Medium Treatment Percentage (%)
WENRAS
(control media)
LFD 100
WENRAS +Nutlin
(only p53 mutant survive)
LFD
LFD +Red 40
20 ±2.3
52 ±5.1*
ENRAS
(only APC mutant survive)
LFD
LFD +Red 40
1 ±0.4
1 ±0.5
WENRAS
(control media)
HFD 100
WENRAS +Nutlin
(only p53 mutant survive)
HFD
HFD +Red 40
23 ±2.5
46 ±4.8**
ENRAS
(only APC mutant survive)
HFD
HFD +Red 40
16 ±1.8
20 ±2.2
* indicates a signicant difference from LFD/WENRAS+Nutlin
** indicates a signicant difference from HFD/WENRAS+Nutlin
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
223
2.5. Histology and quantifying inammation
H&E staining was performed on 5-µm frozen sections, followed by
bright eld microscopy image analysis and scored for inammation by
our pathologist in a blinded fashion. Scores were assigned to each slide
that reects three histological features by the percent area of involve-
ment. Inammation severity was scored as following: 0 for none, 1 for
minimal, 2 for moderate, and 3 for severe; inammation extent as
following: 0 for none, 1 for mucosa, 2 for mucosa and submucosa, and 3
for transmural; crypt damage as following: 0 for none, 1 for one-third of
crypt damaged, 2 for two-thirds of crypt damaged, 3 for crypt loss and
surface epithelium intact, 4 for crypt loss and surface epithelium loss.
Percent area involvement was scored as following: 0 for 0 %, 1 for 1–25
%, 2 for 26–50 %, 3 for 51–75 %, and 4 for 76–100 %. Therefore, the
minimal score is 0, and the maximal score is 40. We have used this
method of method extensively in previous publications [41–43]. For
immunohistochemical staining, formalin-xed, parafn-embedded se-
rial sections of mouse colon tissues were incubated overnight with an-
tibodies against iNOS (rabbit polyclonal, cat# 160862, diluted 1:3500;
Cayman Chemical, Ann Arbor, MI) by slow rocking using the Antibody
Amplier (ProHisto, Columbia, SC) to ensure even staining and repro-
ducible results. After incubation with primary antibodies, sections were
processed using EnVision+System-HRP kits (DakoCytomation, Car-
pinteria, CA) according to kit protocols. The chromogen was dia-
minobenzidine, and sections were counterstained with 1 % methyl
green. Intensity and degree of staining were evaluated independently by
three blinded investigators (QZ, AC and IC). For each tissue section, the
percentage of positive cells was scored on a scale of 0–5 for the per-
centage of tissue stained: 0 (0 % positive cells), 1 (<10 %), 2 (11–25 %),
3 (26–50 %), 4 (51–80 %), or 5 (>80 %). Staining intensity was scored
on a scale of 0–3: 0 (negative staining), 1 (weak staining), 2 (moderate
staining), or 3 (strong staining). The two scores were multiplied,
resulting in an immunoreactivity score (IRS) value ranging from 0 to 15
[43–46].
2.6. Comet assay (DNA damage) analysis
The single-cell gel electrophoresis Comet assay was performed on
HCT 116 cells, treated with 0–1500
μ
M, as well as on colon epithelial
cells of mice treated with human equivalent of the acceptable daily
intake (ADI) of Red 40 or double acceptable daily intake (2xADI) for 6 h,
24 h, or 1 week. The Comet Assay has been described by our group
previously [47]. Briey, following treatment of mice with Red 40, we
dissected out the colon, ushed it out with 1 ×PBS, opened it longitu-
dinally and cut the colon into two small pieces, which were incubated in
10 % fetal bovine serum/5 mM ethylenediaminetetraacetic acid in 1 ×
Ca
2+
/Mg
2+
-free PBS for 15 min at room temperature. Colon tissues were
then shaken to dislodge the epithelial layer into single-cells suspensions.
Cell viability was checked by trypan blue exclusion and >95 % cells
were viable. The single-cell suspension was centrifuged (200 g.m. for 5
min), and the pellet was brought up in freezing media and frozen at
–80 ◦C until Comet analysis. An alkali Comet assay was performed ac-
cording to instructions provided by the kit manufacturer (Comet-
Assay™, Trevigen, Gaithersburg, MD). Cells treated with hydrogen
peroxide (200
μ
M, 20 min) were used as positive controls. A minimum of
75 Comets per treatment were quantied after capturing with and
quantied by the Automated Comet Assay Analysis System Comet-
Assay™, Trevigen, Gaithersburg, MD). Olive tail moment was used to
evaluate DNA damage. The tail moment, expressed in arbitrary units, is
calculated by multiplying the percent of DNA (uorescence) in the tail
by the length of the tail in
μ
m. The tail length is measured between the
edge of Comet head and the end of the Comet tail. An advantage of using
the tail moment as an index of DNA damage is that both the amount of
DNA damage and the distance of migration of the genetic material in the
tail are represented by a single number. Three mice per treatment con-
dition were used and the average of three Comet assays was plotted. For
in vitro assay the average of the tree independent experiments were
plotted.
2.7. Colon organoids
Colonic tissue organoids developed from experimental mice crypts
were studied. Our understanding of the molecular and cellular mecha-
nisms that mediate the stem cell microenvironment have been leveraged
to assess mutations in colon cancer driver genes: APC and p53. To
quantify effects, the mouse colon was opened lengthwise and cut into
2–4 cm pieces, placed in a 50 ml conical tube lled with ice-cold 1X
RPMI 1640 (Gibco) buffer. The tissue was transferred into 20 ml ice-cold
DPBS (Gibco) for cleaning. Tubes were inverted gently four times, then
the tissue was transferred to sterile 50 ml tubes containing 30 mM EDTA,
1.5 mM DTT, diluted into 1X DPBS and incubated on ice for 20 min.
After dissociation, tissues were placed into a pre-warmed (37 ◦C) 30 mM
EDTA, diluted into 1X DPBS and incubated at 37 ◦C for 10 min; then
vigorously shook for 30 s to detach the epithelium from the basement
membrane. Suspended cells were transferred into a sterile 15 ml conical
tubes and pelleted by centrifugation at 500 g for 10 min at 4 ◦C. Su-
pernatant was decanted, and washed 2x with the same buffer, then
centrifuged at 500 g for 5 mins to pellet. Advanced DMEM/F12 was
supplemented with penicillin/streptomycin, 10 mM HEPES, 2 mM
GlutaMAX, 1 ×B27 (Life Technologies), 10 nM gastrin I (Sigma), and 1
mM N-acetylcysteine (Wako). The following niche factors were used: 50
ng/ml mouse recombinant EGF (Life Technologies), 100 ng/ml mouse
recombinant Noggin (Peprotech), 10 % R-spondin-1 conditioned me-
dium, 50 % Wnt-3A conditioned medium, 500 nM A83–01 (Tocris), and
10
μ
M SB202190 (Sigma). After matrigel polymerization, the cells were
overlaid with culture media representing different niche factor condi-
tions. Crypts were observed under the microscope and the nal pellet
was plated into matrigel (Corning Life Sciences, Durham, NC) which
supports 3D structural growth. The minimum amount of matrigel used
as 50/50 ratio of the pellet size, with 75 µL/well in a 12-well plate. 75 µL
of organoids/matrigel were placed in the center of each well forming a
dome. We maintained the plate in the incubator until the matrigel so-
lidied. Then, we added 1 ml of WENRAS media with ROCK and GSK
inhibitors in each well. After several days, organoids were placed in
standard media, or WENRAS+nutlin (surviving clones are p53 mutant),
or ENRAS (media minus Wnt-3a; surviving clones are APC-mutant).
Media was changed every other day; and cultured for 3 weeks. Orga-
noids were counted under a 40x microscope at the end of the
experiment.
2.8. Stool and microbiome analysis
Blood in stool was detected using Hemoccult (Beckman Coulter)
fecal immunochemical test. Immediately before sacrice, stool consis-
tency (0-fully formed stool; 2-loose stool; 3-diarrhea) and blood in the
stool (0-no blood; 2-detected using Hemoccult; 3-rectal bleeding) were
scored, and these measurements were used along with the weight dif-
ference in mice from the beginning to the end of the experiment (0 =no
weight loss; 1 =0–5 % weight loss; 2 =6–10 % weight loss; 3 =11–15 %
weight loss; 4 =16–20 % weight loss), to calculate a cumulative disease
index (CDI). Fecal samples were freshly collected from 1 month old, 7
months old, and 11 months old. During fecal sample collection, mice
were put in a large empty tip box, which have been punched with
multiple holes to maintain airow. After approximately 20 min, mice
produced enough fecal drops in the box, then collected fecal samples in a
clean 1.5 ml microcentrifuge tube. Fecal samples were freshly frozen in
– 80
 C for microbial analysis. DNA extraction was performed, and the
eluted DNA was divided into 2 plates; one of which was used to prepare
16 S libraries. After extraction, samples were quantied; then the V4
region of the 16 s rRNA gene was amplied from each sample using the
Dual indexing sequencing strategy. Sequencing was done on the Illu-
mina MiSeq platform. PCR was performed and products were visualized.
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
224
Libraries were normalized using SequalPrep Normalization Plate Kit
(Life technologies) following the manufacturer’s protocol for sequential
elution. The concentration of the pooled samples was determined and
the sizes of the amplicons in the library was determined. The nal li-
brary consisted of equal molar amounts from each of the plates,
normalized to the pooled plate at the lowest concentration. Library
Preparation for Sequencing and Sequencing Libraries were prepared
according to Illumina’s protocol for Preparing Libraries for Sequencing
on the MiSeq. PhiX and genomes were added in 16 s amplicon
sequencing to create diversity. A mock community was used for error
analysis.
2.9. Statistical analysis
The bioinformatics pipeline includes comparisons of each group’s
phylogenetic alpha diversity in Shannon vector and evenness vector,
which indicates diversity of microbial communities. To evaluate
phylogenetic and possible metabolomic alterations within samples,
QIIME reads processing with chimera removal, open reference, and
SILVA rRNA database (Silva_99) options were used. For PiCRUSt
application, a closed reference against the Greengenes database selec-
tion was used. Operational Taxonomic Unit (OTU) tables were gener-
ated from Nephele and were subjected to Linear Discrimination Analysis
of effect size (LEfSe). Data for each group presented as mean with
standard deviation. Multiple comparisons between the groups were
conducted by one-way analyses of variance (ANOVAs), followed by
Newman-Keuls’ post hoc tests. The p-value chosen for signicance was
0.05. All statistical analyses were performed in GraphPad PRISM 8.0
(GraphPad Software, San Diego, CA).
3. Results
3.1. Red 40 causes DNA damage in vitro
In vitro experiments were conducted to examine the impact of Red 40
on human HCT 116 colon cancer cells. To this, we carried out dose- and
time-course experiments with physiologically relevant doses of Red 40.
Fig. 1 shows that Red 40 causes DNA damage (determined by the comet
assay) in a dose (Fig. 1A) and time (Fig. 1B) dependent manner.
Consistent with these ndings, Red 40 also causes a dose-dependent
increase in the DNA damage markers, p53 and phosphorylated p53
(on serine 15) (Fig. 1C). It is important to mention that Red 40 does not
have signicant cytostatic or cytotoxic effect under these experimental
conditions as the cell grows rate /count and viability (94–96 %, n =3)
for Red 40 treated cells have not changed in comparison to non-treated
cells (viability of 96 %, n =3).
3.2. Red 40 causes DNA damage in vivo
Although previous studies from the early 2000’s have shown that
Red 40 causes DNA damage to the colon) [48–50]; others have shown a
lack of DNA damage [51]. Because this is an unresolved issue, and it
makes sense to conrm our in vitro results; here we evaluated the impact
of Red 40 on DNA damage in vivo. Fig. 2 shows that Red 40 – at doses
equivalent to the Accepted Daily Allowance (ADI) or twice the ADI,
causes DNA damage in the colon. Mice were given a bolus of Red 40 at
indicated doses, then colon cells collected at indicated times (0. 6, 24
hrs; and 1 week) by scraping and examined by the Comet Assay.
3.3. Impact of a HFD and Red 40 on health and the colon
We compared the food intake, body weight changes, and other
endpoints of health in the different treatment groups. Although the
average amount of food intake was higher in the LFD mice (Fig. 3A), the
weight gain was higher in mice consuming the HFD (Fig. 3C). The HFD
animals’ overall caloric intake was 10.29 Kcal/mouse/day; while the
LFD caloric intake was 8.88 Kcal/mouse/day (Fig. 3B). After con-
sumption of the HFD for 10 months, there was an increase in the size of
the spleen, liver, and kidney (Fig. 3D); and a decrease in hemoglobin
levels (Fig. 3E). For LFD group, liver, kidney, and spleen weights were
1.3 ±0.12 g, 0.34 ±0.03 g, and 0.08 ±0.06 g, respectively (Fig. 3F). A
HFD group, liver, kidney, and spleen weights were 1.25 ±0.35 g, 0.04
±0.09 g, and 0.09 ±0.04 g, respectively (Fig. 3F). After an addition of
the Red 40 to diets, a LFD group liver, kidney, and spleen weights were
1.46 ±0.13 g, 0.34 ±0.04 g, 0.09 ±0.02 g, respectively, and HFD
group liver, kidney, and spleen weights were 1.27 ±0.24 g, 0.34
±0.06 g, and 0.08 ±0.03 g, respectively. Colon length decreased
(although non-signicant) also with consumption of the HFD (9.26
±0.6 cm for LFD and 8.7 ±0.82 cm for HFD), and even further with an
addition of Red 40–8.75 ±1.38 cm for LFD and 7.79 ±1.27 cm. Daily
Fig. 1. Red 40 induces DNA damage in vitro. A. HCT116 cells were exposed to
Red 40 as indicated by dose then harvested after 24 h exposure. B. Time course
of exposure to 1 mM Red 40. C. Red 40 causes phosphorylation of P53 at serine
15 and stabilizes P53. For the comet assay (A, B), an alkaline Comet assay was
carried out, which detects both double and single stranded DNA breaks. Tail
moment is the product of the tail length and the fraction of total DNA in the tail;
and a longer tail moment indicates higher DNA damage. Error bars represent
the standard error (n =3).
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
225
liquid and food intake were monitored for all the groups; with no sig-
nicant difference in the amount consumed (Fig. 4A). In comparing the
impact of Red 40 on body weight, we saw that within the rst several
weeks after Red 40 consumption, the LFD with Red 40 group had sig-
nicant lower body weight than control group, (p <0.05) (Fig. 4C, D,
E), which indicate toxicity of Red 40 in mice fed a HFD. Indeed, survival
was lower in mice consuming Red 40 (Fig. 4F). Another observation was
that the HFD overlaid with Red 40, increased lymphocytes percentage in
peripheral blood (Fig. 4B).
Next, colon macroscopic foci were counted. All groups had foci by 11
months old, which indicates that aging is also a parameter for devel-
oping foci anomalies. Of note, mice consuming Red 40 had their intes-
tine content and tissue dyed red.
Consumption of a HFD with Red 40 increases foci number and size in
the colon compared to other groups. Flat structures, resembling aberrant
crypt foci, were counted under a dissecting microscope in the proximal
colon (Fig. 5A) and distal (Fig. 5B) colon/rectum. Red 40 increases the
number of foci in the distal colon in the presence or absence of the HFD.
A HFD appears to increase the number of foci in the proximal colon; and
Red 40 does not impact this endpoint in this part of the colon. We then
scored the inammatory “histological” changes in the distal colon and
rectum (Fig. 5C); which shows the degree of pathological inammation
increasing with the HFD ±Red 40; and Red 40 alone (+LFD) also in-
creases distal colon inammation. We also stained for inducible nitric
oxide synthase (iNOS) in the distal colon/rectum as a marker of
inammation (Fig. 5D). iNOS – a marker of inammation – is also
induced by Red 40 ±HFD. Finally, to see if Red 40 affects systemic in-
ammatory load, we measured serum IL-6; and saw that Red 40 indeed
elevates the levels of IL-6 in the presence of a HFD (Fig. 5E).
3.4. HFD and Red 40 impacts the microbiome in vivo
Because the microbiome is a key regulator of colonic health [52], we
wanted to examine how a Red 40 impacts the microbiome with and
without HFD. While alpha-diversity represents the diversity within each
sample, beta-diversity represents the difference between samples. In
other words, it describes how similar or different the two ecosystems
are. First, we asked how the fecal microbiome changes after 10 months
of consumption of the HFD vs. LFD. Fig. 6A shows a Bray-Curtis
dissimilarity between samples; and that diet signicantly changes the
beta-diversity of the samples. Also, after 10 months of consuming a HFD,
there is a decreased phylogenetic alpha diversity in Shannon vector
(Fig. 6B) and evenness vector (Fig. 6C), which indicate lower diversity of
microbial communities, and less healthy microbial environment. As
expected, 1-month old mice microbiota – still inuenced by mother - had
less diverse and richness (Fig. 6B, C). Quantication of bacteria based on
Phyla (Fig. 6D) shows that mice fed a HFD over 10 months had signif-
icantly increased Actinobacteria and Firmicutes; and signicantly lower
quantities of protective Bacteroidetes and Verrucomicrobia. Operational
taxonomic units at the family level were also compared (Fig. 6E). The
10-month consumption of the HFD results in fecal samples having
signicantly increased levels of Bidobacteriaceae and Lactobacillaceae;
and decreased levels of S24–7, Clostridiales, Lachnospiraceae, Rumino-
coccaceae and one undened family. Finally, we present microbial com-
munities through heatmap plots (Fig. 6F). Mice consuming a HFD had
elevated levels of Bidobacteriaceae;g__Bidobacterium, Lactobacillaceae;
g__Lactobacillus; and especially Erysipelotrichaceae; g_Allobaculum. There
also was a decreased S24–7, Verrucomicrobiaceae Akkermansia, and
Enterococcus.
We next asked how Red 40 affects the microbiome. Bray-Curtis
dissimilarity between samples shows that the beta-diversity is
different amongst the three groups; including the control LFD group
with and without Red 40 (Fig. 7A). However, Red 40 - alone - does not
appear to change the alpha diversity (Fig. 7B, C); a result consistent with
another recent study [35]. The introduction of the HFD with Red 40
however signicantly decreases benecial microbial communities and
increases harmful microbial communities in A/J mice (Figs. 7D, 7E).
Operational taxonomic units at the phyla level were quantied and
showed that the consumption of a HFD+Red 40 causes a signicant
increase in Actinobacteria, Firmicutes; and Proteobacteria; and a decrease
in Bacteroidetes (Fig. 7D). The benecial species, Verrucomicrobia, was
signicantly reduced by Red 40 alone (Fig. 7D). Finally, heatmap plots
show the gut microbiota composition between samples at the family
level. Consumption of Red 40 in the backdrop of a LFD causes an in-
crease in S24–7 and clostridiales; and a decrease in Lactobacillaceae
(Fig. 7F). Also, our HFD +Red 40 causes an increase in Erysipelo-
trichaceae compared with consumption of the LFD (Fig. 7F).
3.5. Impact of a HFD and Red 40 on functional p53 and APC mutations
Organoids were cultured according to well established mouse orga-
noid protocols [53]. Indeed, we have extensive experience establishing,
culturing, and testing colon organoids in our college core [54,55].
Organoids analyzed ex-vivo were functionally tested for the acquisition
of mutations to the gene by plating in the absence of ligands, as
described previously [56]. Briey, organoids are selected for TP53 and
APC mutants by adding Nutlin-3a (TP53) or removing Wnt-3A (APC) to
the media. Adding Nutlin-3a or Wnt-3A to the growth media will kill
cells with wildtype TP53 or APC, respectively, thus enriching for orga-
noids containing mutations. We observed that organoids derived from
HFD colonic culture are 2-fold larger than LFD colonic organoids. We
also observed that p53 is functionally mutated by Red 40 in the back-
drop of consuming either a LFD and a HFD. Although APC was func-
tionally mutated by consuming either the LFD or the HFD; Red 40 did
not induce further functional mutations (Table 1).
4. Discussion
Diseases such as Inammatory Bowel Disease (IBD) are characterized
by persistent, cyclic, and heightened inammation, often associated
with an elevated cancer risk within the affected tissue, such as the
increased risk of CRC due to IBD. While this demonstrates a robust
connection between intrinsic, recurrent, high-grade inammation and
cancer, emerging evidence highlights that even subtle, chronic inam-
mation can contribute to colorectal cancer development. In this study,
we postulated that Red 40, a synthetic food dye, might incite a con-
cealed, low-grade, tissue-specic inammation in the colon and rectum,
Fig. 2. Red 40 induces DNA damage in vivo. A. Red 40 given at a human
equivalent dose of 7 mg/kg (ADI) and 14 mg/kg (2x ADI) for 6 h, 24 h or 1
week causes DNA damage in cells collected from the colon after indicated time
points. Female A/J mice were given Red 40 orally (e.g., oral gavage) at indi-
cated doses (ADI or 2xADI), then colon cells collected at indicated times (0. 6,
or 24 h; and 1 week) by scraping and examined by the Comet Assay. Error bars
represent the standard deviation (n =3).
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
226
thereby laying the groundwork for colorectal carcinogenesis. With
ubiquitous exposure to synthetic chemicals through daily diets, partic-
ularly in highly processed, Western-style high-fat diets (HFDs), which
are often rich in synthetic food dyes, such as Red 40, the potential
impact is concerning. Disturbingly, the escalating consumption of Red
40 parallels the rising incidence of early-onset colorectal cancer
(EOCRC).
As our study investigates the effects of Red 40 on CRC-related end-
points—DNA damage, low-grade inammation, and the gut micro-
biome—we present foundational evidence substantiating its deleterious
inuence. Notably, our ndings align with recent research indicating
that Red 40 triggers colitis in experimental settings [34,35]. He and
Chen et al. [35,36] showed Red 40 (7 mg/kg/d) triggers experimental
colitis in mice expressing IL-23; and that this was dependent on
CD4 +T-cell and IFN-γ signaling [54]. Kwon et al. [34] showed that
chronic (100 ppm in diet ad libitum for 12 weeks) exposure to Red 40
prior to experimental induction of colitis exacerbates this condition
[34]; while intermittent Red 40 exposure (1 day per week for 12 week)
does not. Additionally, early life exposure to Red 40 primed mice to a
heightens susceptibility to DSS-induced colitis [34]. So, multiple studies
– including this one - demonstrate that chronic exposure to Red 40
provokes mild colitis, akin to our results following a ten-month chronic
exposure (Fig. 5). Importantly, Red 40 induces DNA damage and acti-
vates the p53 pathway, a signicant revelation considering p53’s pivotal
role in colorectal carcinogenesis (Figs. 1 and 2). Our nding that p53 is
functionally mutated by Red 40 in the backdrop of consuming either a
LFD and a HFD (Table 1) is novel and of signicance given the key role
that p53 plays in colorectal carcinogenesis [57]. Interestingly, Red 40
does not – in itself – functionally mutate APC. This impact of Red 40 on
p53 but not APC parallels the ndings that human EOCRC has a rela-
tively high p53 mutation load; but low APC mutation load [7,58,59].
Because Red 40 is metabolized by the gut microbiome to
Fig. 3. Impact of consuming a low-fat diet (LFD) vs. high-fat diet (HFD) on food intake (A), calorie intake (B), body weight changes (C), organ impact (D), systemic
hemoglobin (E) and organ weight (F). Female A/J mice were fed LFD or HFD starting at weaning. The average amount of food intake was higher in the LFD mice
(Fig. 3A); although the average daily calorie intake was lower in this group (Fig. 3B). Weight gain was higher in mice consuming the HFD (Fig. 3C). There was an
increase in the size of the spleen, liver, and kidney in the HFD group (Fig. 3D, F); and a decrease in hemoglobin levels in the HFD group (Fig. 3E).
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
227
cresidine-4-sulfonic acid (CSA-Na) and 1-amino-2-naphthol-6-sulfonic
acid (ANSA-Na), future studies will test whether these metabolites
also damage DNA. Indeed, previous studies have shown that ANSA-Na
can trigger colitis [35]; and that Red 40 and its’ metabolites can
impact the DNA and have pro-inammatory properties [28,29,60,61].
In studying the impact of Red 40 on the microbiome, our data show
that Red 40 does not appear to change the general composition of the
bacterial community - indicated by the lack of change in alpha diversity.
This result is consistent with other recent studies [35,36]. However, in
the presence of a HFD, the consumption of Red 40 contributes to a
change in the alpha and beta diversities of the fecal microbiome (Figs. 6
and 7). Although we have evidence of specic species of bacteria
impacted by Red 40, further studies are needed to delineate which
species mechanistically contribute to CRC because of Red 40 exposure.
Only two studies have previously explored which bacteria species
metabolize Red 40 and contribute to colon anomalies [35,62]. He et al.
showed that Red 40 is metabolized by B. ovatus and E. faecalis; but not
E. coli [35]. Chung et al. [62] showed high metabolism by Anaerobes,
Fusobacterium; modest metabolism by Bacteroides, Bidobacterium, and
Citrobacter; and that Red 40 is not metabolized by Acidamiococcus,
Peptostreptococcus [62]. Given the signicance of the gut microbiome to
health, and that Red 40 is consumed by a diverse and large number of
people, a further understanding of the interactions between Red 40 and
the gut microbiome will be of high importance to public health.
Because of the tight control of dietary exposures and careful mea-
surement of CRC-related outcomes, our study has added to the foun-
dation of growing scientic evidence that Red 40 and its metabolites
target endpoints that control the genesis of CRC: DNA damage, func-
tional mutation to p53, low-grade inammation in the distal colon, and
the microbiome. The use of in vitro and in vitro models has allowed us to
Fig. 4. Impact of consuming a LFD vs. HFD on liquid intake (A), systemic lymphocyte count (B), body weight changes (C, D, E), and percent survival (F). Female A/J
mice were exposed to Red 40 in the drinking water (mouse equivalent of 7 mg/kg/d (HFDR) or 14 mg/kg/d (HFD2R)) beginning at 3 months of age and continuing
until 11 months of age.
*, indicates signicant difference (p <0.05).
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
228
dive into mechanisms relevant to CRC. Despite its strengths, this study
has some weaknesses, including a lack of mechanistic insight into these
observational changes. Further studies will be needed to build on the
foundation that Red 40 causes low-grade inammation; and ask how it
does this, and which microbial species are most relevant to this process.
Until then, limitations remain to extrapolating results from cell culture
and animal studies to humans. Studies in humans will be needed to
deepen understanding of the role of Red 40 in the natural history
EOCRC. These studies will require careful planning and execution to
address practical and ethical concerns. Interventions will need to be
designed with temporal limitation in mind – e.g., using the same kinds of
assays employed in this study and taking advantage of methodologic
innovations that we have discussed earlier [63]. As with most epide-
miologic advancements, this also will require careful design, planning,
execution, and analyses of observational studies that do not share the
constraints of trials that are constrained to studying small study groups
over relatively short periods of time (i.e., the weeks or months needed to
see changes in intermediate biomarkers as opposed to the years or
Fig. 5. Red 40 causes inammation in the distal colon and rectum of the A/J mouse model. A/J mice were exposed to Red 40 in the drinking water (mouse
equivalent of 7 mg/kg/d or 14 mg/kg/d) beginning at 3 months of age and continuing until 11 months of age. Because of the use of Red 40 in a high-fat, westernized
diet; we also included groups consuming both a low-fat diet (LFD) and a high-fat diet (HFD). Flat structures, resembling aberrant crypt foci, were counted under a
dissecting microscope in the proximal colon (A) and distal (B) colon/rectum. We then scored the inammatory “histological” changes in the distal colon and rectum
as we have done many times previously (C). Foci count in the distal colon and rectum were also quantied (D). We also stained for inducible nitric oxide synthase
(iNOS) in the distal colon/rectum as a marker of inammation (E). Finally, to see if Red 40 affects systemic inammatory load, we measured serum IL-6; and saw that
Red 40 indeed elevates the levels of IL-6 in the presence of a HFD (F). HFDR =High-Fat Diet +1x Acceptable Daily Intake (ADI) Red 40; HFD2R =HFD +2x ADI Red
40.
*, indicates signicant difference (p <0.05).
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
229
decades needed to see CRC [64]. Future laboratory studies, conducted in
tandem with human studies, should elaborate relevant mechanisms of
action.
In conclusion, our study, distinguished by meticulous dietary control
and selected CRC-related outcome measurements, underscores Red 40’s
adverse effects. The combination of in vitro and in vivo models facilitates
nuanced exploration of CRC-relevant mechanisms. Nonetheless, our
study does possess limitations, including the need for more detailed
mechanistic insights into observed changes. As we continue to examine
how Red 40 triggers low-grade inammation, with emphasis on relevant
microbial species, we acknowledge the challenges in extrapolating
results from animal models to human contexts. To this, the current study
advances our understanding of Red 40’s detrimental effects on health,
highlighting the potential of chronic exposure to elevate CRC risk. We
demonstrate that Red 40 inicts DNA damage, particularly in the pres-
ence of a HFD, which leads to altered gut microbiota and subsequent
inammation in the distal colon. These ndings contribute to the
growing body of evidence illustrating Red 40’s adverse impact on
colorectal carcinogenesis. Future endeavors should incorporate human
studies to deepen insights into Red 40’s role in EOCRC’s natural history,
alongside further laboratory investigations to elucidate underlying
mechanisms.
Fig. 6. Microbial communities’ analysis: age and LFD vs HFD. A. Phylogenetic based beta diversity-principal component analysis (PCA) plot of fecal microbiota was
examined. Plots based on unweighted UniFrac distance matrices of microbial communities in fecal samples, three separated clusters were displayed. PC1 (x-axis)
explained 49.41 %, PC2 (y-axis) explained 11.78 % of variability, PC3 (z-axis) explained 8.73 % of the variability. B. Phylogenetic based alpha diversity-the Shannon
index. Lower Shannon vectors indicate the lower diversity of the microbial communities. LFD mice had the highest diversity than the other two groups. 1-month old
mice majority microbial communities came from the mother and less diverse and richness inuenced by diet. Microbiota in young mice tend to be less diverse than
that of older mice. C. Alpha diversity-the evenness was compared in three groups. The LFD group had the highest richness and evenness than the other two groups,
indicating LFD fed mice microbiome community has a small disparity between the number of individuals within each species. D. Quantitative phyla levels of
operational taxonomic unit (OTU) was compared. E. Family level of operational taxonomic unit (OTU) was compared. The HFD mice samples increased in Bi-
dobacteriaceae, Lactobacillaceae; decreased in S24–7, Clostridiales, Lachnospiraceae, Ruminococcaceae and one undened family. Signicance, *p <0.05, * *p <0.01,
* **p <0.001, * ** * p <0.0001. F. Microbial communities’ analysis. Heatmap plots of 11 months mice fecal samples showed the gut microbiota composition
between samples at the genus level. Relative abundances of individual taxa (rows) in each sample (columns) are indicated in the associated color scale.
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
230
Author contributions
QZ, AAC, and LJ conceived and designed the experiments; QZ, N,
AAC, NS, JH, MK, and QH performed the experiments; IC examined
pathology; PJB and CEB oversaw organoid experiments. All contributed
to writing the manuscript.
Funding
This work was supported by the National Institutes of Health, grants
1R01CA246809 and 1U01CA272977 issued to Lorne J. Hofseth.
Declaration of Competing Interest
The authors declare the following nancial interests/personal
Fig. 7. Microbial diversity of after 10 months exposure to Red 40. A. Phylogenetic based beta diversity, a qualitative principal component analysis (PCA) plot of fecal
microbiota. Compare LFD 11-month old mice fecal samples, LFD/HFD with Red 40 11 months old mice fecal sample. Unweighted UniFrac distance matrices of
microbial communities in fecal samples, PC1 (x-axis) explained 33.96 %, PC2 (y-axis) explained 16.31 % of variability, PC3 (z-axis) explained 9.96 % of the
variability. B. Phylogenetic based alpha diversity, the Shannon index. Lower Shannon vectors indicate the lower diversity of the microbial communities. Comparing
HFD with Red 40 11 months old mice to LFD with or without Allura Red AC 11 months old mice Shannon vector signicantly decreased. *p <0.05, * *p <0.01,
* **p <0.001. C. Comparing the alpha diversity evenness. Comparing the HFD with Red 40 11 months old mice to LFD with or without Red 40 11 months old mice
samples, diversity of evenness signicantly decreased. Signicant, D. Operational taxonomic unit (OTU) of the Phyla level were compared: LFD and HFD with and
without Red 40 11 months mice microbiota. The signicantly increased in Actinobacteria, Firmicutes; and Proteobacteria. Decreased in Bacteroidetes only in HFD with
Red 40 samples, and decreased Verrucomicrobia in both LFD and HFD with Red 40 samples. E. Comparing the OTU of the family level: Increased in Bidobacteriaceae
and Erysipelotrichaceae, especially in overlaid with HFD and Red 40, decreased in Bacteroidales family level in S24–7, especially in overlaid with HFD and Red 40;
Clostridiales family level in Ruminococcaceae and Lachnospiraceae, especially in overlaid with HFD and Red 40. F. Heatmap of family level. Heatmap plots show the gut
microbiota composition between samples at the family level. Relative abundances of individual taxa (rows) in each sample (columns) are indicated in the associated
color scale. *p <0.05, * *p <0.01, * **p <0.001, * ** * p <0.0001.
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
231
relationships which may be considered as potential competing interests:
Lorne Hofseth reports nancial support was provided by National In-
stitutes of Health.
Data Availability
No data was used for the research described in the article.
Acknowledgements
We thank Dr. Vitali Sikirzhytski, a Director of the Microscopy Core of
the University of South Carolina Center for Targeted Therapeutics for his
assistance with Comet assay imaging, and image analysis.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.toxrep.2023.08.006.
References
[1] R.L. Siegel, K.D. Miller, A. Jemal, Colorectal cancer mortality rates in adults aged
20 to 54 years in the United States, 1970-2014, Aug 8, JAMA 318 (6) (2017)
572–574, https://doi.org/10.1001/jama.2017.7630.
[2] R.L. Siegel, K.D. Miller, S.A. Fedewa, et al., Colorectal cancer statistics, 2017, May
6, CA: Cancer J. Clin. 67 (3) (2017) 177–193, https://doi.org/10.3322/
caac.21395.
[3] R.L. Siegel, S.A. Fedewa, W.F. Anderson, et al., Colorectal cancer incidence
patterns in the United States, 1974-2013, Aug 1, J. Natl. Cancer Inst. 109 (8)
(2017), https://doi.org/10.1093/jnci/djw322.
[4] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, 2016 (Jan), CA: Cancer J. Clin.
66 (1) (2016) 7–30, https://doi.org/10.3322/caac.21332.
[5] R.L. Siegel, A. Jemal, Percentage of colorectal cancer diagnosed in adults aged
younger than 50 years, May 1, Cancer 122 (9) (2016) 1462–1463, https://doi.org/
10.1002/cncr.29980.
[6] A. Venugopal, J.M. Carethers, Epidemiology and biology of early onset colorectal
cancer, Excli J. 21 (2022) 162–182, https://doi.org/10.17179/excli2021-4456.
[7] L.J. Hofseth, J.R. Hebert, A. Chanda, et al., Early-onset colorectal cancer: initial
clues and current views, Feb 21, Nat. Rev. Gastroenterol. Hepatol. (2020), https://
doi.org/10.1038/s41575-019-0253-4. Feb 21.
[8] L.H. Nguyen, P.H. Liu, X. Zheng, et al., Sedentary behaviors, TV viewing time, and
risk of young-onset colorectal cancer (Nov), JNCI Cancer Spectr. 2 (4) (2018)
pky073, https://doi.org/10.1093/jncics/pky073.
[9] X. Zheng, J. Hur, L.H. Nguyen, et al., Comprehensive assessment of diet quality and
risk of precursors of early-onset colorectal cancer, Nov 2, J. Natl. Cancer Inst.
(2020), https://doi.org/10.1093/jnci/djaa164. Nov 2.
[10] J.R. Hebert, L.J. Hofseth, Diet, Inammation, and Health, Elsevier Science,, 2022.
[11] B. Srour, M.C. Kordahi, E. Bonazzi, M. Deschasaux-Tanguy, M. Touvier,
B. Chassaing, Ultra-processed foods and human health: from epidemiological
evidence to mechanistic insights (Dec), Lancet Gastroenterol. Hepatol. 7 (12)
(2022) 1128–1140, https://doi.org/10.1016/s2468-1253(22)00169-8.
[12] N. Kliemann, A. Al Nahas, E.P. Vamos, et al., Ultra-processed foods and cancer risk:
from global food systems to individual exposures and mechanisms (Jul), Br. J.
Cancer 127 (1) (2022) 14–20, https://doi.org/10.1038/s41416-022-01749-y.
[13] F. Baygi, F. Mohammadi-Nasrabadi, B.C. Zyriax, et al., Global overview of dietary
outcomes and dietary intake assessment methods in maritime settings: a systematic
review, Aug 21, BMC Public Health 21 (1) (2021) 1579, https://doi.org/10.1186/
s12889-021-11593-z.
[14] B.M. Popkin, L.S. Adair, S.W. Ng, Global nutrition transition and the pandemic of
obesity in developing countries (Jan), Nutr. Rev. 70 (1) (2012) 3–21, https://doi.
org/10.1111/j.1753-4887.2011.00456.x.
[15] Y. Zhang, E.L. Giovannucci, Ultra-processed foods and health: a comprehensive
review, Jun 6, Crit. Rev. Food Sci. Nutr. (2022) 1–13, https://doi.org/10.1080/
10408398.2022.2084359.
[16] D. Statovci, M. Aguilera, J. MacSharry, S. Melgar, The impact of western diet and
nutrients on the microbiota and immune response at mucosal interfaces, Front.
Immunol. 8 (2017) 838, https://doi.org/10.3389/mmu.2017.00838.
[17] S.J. O’Keefe, Diet, microorganisms and their metabolites, and colon cancer (Dec),
Nat. Rev. Gastroenterol. Hepatol. 13 (12) (2016) 691–706, https://doi.org/
10.1038/nrgastro.2016.165.
[18] A.D. Benninghoff, K.J. Hintze, S.P. Monsanto, et al., Consumption of the total
western diet promotes colitis and inammation-associated colorectal cancer in
mice, Feb 20, Nutrients 12 (2) (2020), https://doi.org/10.3390/nu12020544.
[19] L.H. Nguyen, Y. Cao, J. Hur, et al., The sulfur microbial diet is associated with
increased risk of early-onset colorectal cancer precursors, Jul 14, Gastroenterology
(2021), https://doi.org/10.1053/j.gastro.2021.07.008. Jul 14.
[20] T. Ugai, L. Liu, F.K. Tabung, et al., Prognostic role of inammatory diets in
colorectal cancer overall and in strata of tumor-inltrating lymphocyte levels
(Nov), Clin. Transl. Med 12 (11) (2022), e1114, https://doi.org/10.1002/
ctm2.1114.
[21] D. Hang, L. Wang, Z. Fang, et al., Ultra-processed food consumption and risk of
colorectal cancer precursors: results from three prospective cohorts, Dec 7, J. Natl.
Cancer Inst. (2022), https://doi.org/10.1093/jnci/djac221. Dec 7.
[22] H.K. Joh, D.H. Lee, J. Hur, et al., Simple sugar and sugar-sweetened beverage
intake during adolescence and risk of colorectal cancer precursors: Adolescent
sugar intake and colorectal polyp, Mar 19, Gastroenterology (2021), https://doi.
org/10.1053/j.gastro.2021.03.028. Mar 19.
[23] J. Hur, E. Otegbeye, H.K. Joh, et al., Sugar-sweetened beverage intake in adulthood
and adolescence and risk of early-onset colorectal cancer among women, May 6,
Gut (2021), https://doi.org/10.1136/gutjnl-2020-323450. May 6.
[24] Y. Yue, J. Hur, Y. Cao, et al., Prospective evaluation of dietary and lifestyle pattern
indices with risk of colorectal cancer in a cohort of younger women, Mar 31, Ann.
Oncol.: Off. J. Eur. Soc. Med. Oncol. (2021), https://doi.org/10.1016/j.
annonc.2021.03.200. Mar 31.
[25] C. Potera, The articial food dye blues (Oct), Environ. Health Perspect. 118 (10)
(2010) A428, https://doi.org/10.1289/ehp.118-a428.
[26] D.L. Doell, D.E. Folmer, H.S. Lee, K.M. Butts, S.E. Carberry, Exposure estimate for
FD&C colour additives for the US population (May), Food Addit. Contam. Part A,
Chem., Anal., Control, Expo. risk Assess. 33 (5) (2016) 782–797, https://doi.org/
10.1080/19440049.2016.1179536.
[27] A. Batada, M.F. Jacobson, Prevalence of articial food colors in grocery store
products marketed to children (Oct), Clin. Pediatr. 55 (12) (2016) 1113–1119,
https://doi.org/10.1177/0009922816651621.
[28] L. Leo, C. Loong, X.L. Ho, M.F.B. Raman, M.Y.T. Suan, W.M. Loke, Occurrence of
azo food dyes and their effects on cellular inammatory responses (Feb), Nutr.
(Burbank, Los Angel Cty., Calif. ) 46 (2018) 36–40, https://doi.org/10.1016/j.
nut.2017.08.010.
[29] L.I. Khayyat, A.E. Essawy, J.M. Sorour, A. Soffar, Sunset yellow and allura red
modulate Bcl2 and COX2 expression levels and confer oxidative stress-mediated
renal and hepatic toxicity in male rats, PeerJ 6 (2018), e5689, https://doi.org/
10.7717/peerj.5689.
[30] (JECFA) JFWECoFA, Safety evaluation of certain food additives, World Health
Organization,, 2017.
[31] B. Raposa, R. Ponusz, G. Gerencser, et al., Food additives: sodium benzoate,
potassium sorbate, azorubine, and tartrazine modify the expression of NFkappaB,
GADD45alpha, and MAPK8 genes (Sep), Physiol. Int. 103 (3) (2016) 334–343,
https://doi.org/10.1556/2060.103.2016.3.6.
[32] S.K. Meyer, P.M.E. Probert, A.F. Lakey, et al., Hepatic effects of tartrazine (E 102)
after systemic exposure are independent of oestrogen receptor interactions in the
mouse, May 5, Toxicol. Lett. 273 (2017) 55–68, https://doi.org/10.1016/j.
toxlet.2017.03.024.
[33] J.P. Brown, Reduction of polymeric azo and nitro dyes by intestinal bacteria (May),
Appl. Environ. Microbiol. 41 (5) (1981) 1283–1286.
[34] Y.H. Kwon, S. Banskota, H. Wang, et al., Chronic exposure to synthetic food
colorant Allura Red AC promotes susceptibility to experimental colitis via
intestinal serotonin in mice, Dec 20, Nat. Commun. 13 (1) (2022) 7617, https://
doi.org/10.1038/s41467-022-35309-y.
[35] Z. He, L. Chen, J. Catalan-Dibene, et al., Food colorants metabolized by commensal
bacteria promote colitis in mice with dysregulated expression of interleukin-23,
May 11, Cell Metab. (2021), https://doi.org/10.1016/j.cmet.2021.04.015. May
11.
[36] L. Chen, Z. He, B.S. Reis, et al., IFN-γ(+) cytotoxic CD4(+) T lymphocytes are
involved in the pathogenesis of colitis induced by IL-23 and the food colorant Red
40 (Jul), Cell Mol. Immunol. 19 (7) (2022) 777–790, https://doi.org/10.1038/
s41423-022-00864-3.
[37] I.S. Khan, S. Ali, K.B. Dar, et al., Toxicological analysis of synthetic dye orange red
on expression of NFκB-mediated inammatory markers in Wistar rats, Sep 24, Drug
Chem. Toxicol. (2021) 1–11, https://doi.org/10.1080/01480545.2021.1979579.
[38] A. Vojdani, C. Vojdani, Immune reactivity to food coloring, Alter. Ther. Health Med
21 (Suppl 1) (2015) 52–62.
[39] K. Guda, J.N. Marino, Y. Jung, K. Crary, M. Dong, D.W. Rosenberg, Strain-specic
homeostatic responses during early stages of azoxymethane-induced colon
tumorigenesis in mice (Oct), Int. J. Oncol. 31 (4) (2007) 837–842.
[40] D.W. Rosenberg, C. Giardina, T. Tanaka, Mouse models for the study of colon
carcinogenesis (Feb), Carcinogenesis 30 (2) (2009) 183–196, https://doi.org/
10.1093/carcin/bgn267.
[41] A. Chaparala, H. Tashkandi, A.A. Chumanevich, et al., Molecules from American
ginseng suppress colitis through nuclear factor erythroid-2-related factor 2, Jun 21,
Nutrients 12 (6) (2020), https://doi.org/10.3390/nu12061850.
[42] A. Chaparala, D. Poudyal, H. Tashkandi, et al., Panaxynol, a bioactive component
of American ginseng, targets macrophages and suppresses colitis in mice, Jun 2,
Oncotarget 11 (22) (2020) 2026–2036, https://doi.org/10.18632/
oncotarget.27592.
[43] A.A. Chumanevich, A. Chaparala, E.E. Witalison, et al., Looking for the best anti-
colitis medicine: a comparative analysis of current and prospective compounds,
Jan 3, Oncotarget 8 (1) (2017) 228–237, https://doi.org/10.18632/
oncotarget.13894.
[44] A.A. Chumanevich, C.P. Causey, B.A. Knuckley, et al., Suppression of colitis in mice
by Cl-amidine: a novel peptidylarginine deiminase inhibitor (Jun), Am. J. Physiol.
Gastrointest. Liver Physiol. 300 (6) (2011) G929–G938, https://doi.org/10.1152/
ajpgi.00435.2010.
[45] A.A. Chumanevich, E.E. Witalison, A. Chaparala, et al., Repurposing the anti-
malarial drug, quinacrine: new anti-colitis properties, Aug 16, Oncotarget 7 (33)
(2016) 52928–52939, https://doi.org/10.18632/oncotarget.10608.
[46] H. Shaked, L.J. Hofseth, A. Chumanevich, et al., Chronic epithelial NF-κB
activation accelerates APC loss and intestinal tumor initiation through iNOS up-
Q. Zhang et al.

Toxicology Reports 11 (2023) 221–232
232
regulation, Aug 28, Proc. Natl. Acad. Sci. USA 109 (35) (2012) 14007–14012,
https://doi.org/10.1073/pnas.1211509109.
[47] Y. Jin, V.S. Kotakadi, L. Ying, et al., American ginseng suppresses inammation
and DNA damage associated with mouse colitis (Epub), Carcinogenesis 2008 29 (12)
(2008) 2351–2359.
[48] C. Shimada, K. Kano, Y.F. Sasaki, I. Sato, S. Tsudua, Differential colon DNA damage
induced by azo food additives between rats and mice (Aug), J. Toxicol. Sci. 35 (4)
(2010) 547–554.
[49] Y.F. Sasaki, S. Kawaguchi, A. Kamaya, et al., The comet assay with 8 mouse organs:
results with 39 currently used food additives, Aug 26, Mutat. Res. 519 (1–2) (2002)
103–119.
[50] S. Tsuda, M. Murakami, N. Matsusaka, K. Kano, K. Taniguchi, Y.F. Sasaki, DNA
damage induced by red food dyes orally administered to pregnant and male mice
(May), Toxicol. Sci.: Off. J. Soc. Toxicol. 61 (1) (2001) 92–99.
[51] M. Bastaki, T. Farrell, S. Bhusari, K. Pant, R. Kulkarni, Lack of genotoxicity in vivo
for food color additive Allura Red AC (Jul), Food Chem. Toxicol.: Int. J. Publ. Br.
Ind. Biol. Res. Assoc. 105 (2017) 308–314, https://doi.org/10.1016/j.
fct.2017.04.037.
[52] I. Yu, R. Wu, Y. Tokumaru, K.P. Terracina, K. Takabe, The role of the microbiome
on the pathogenesis and treatment of colorectal cancer, Nov 19, Cancers 14 (22)
(2022), https://doi.org/10.3390/cancers14225685.
[53] T. Sato, H. Clevers, Primary mouse small intestinal epithelial cell cultures, Methods
Mol. Biol. 945 (2013) 319–328, https://doi.org/10.1007/978-1-62703-125-7_19.
[54] C. Liu, C.E. Banister, C.C. Weige, et al., PRDM1 silences stem cell-related genes and
inhibits proliferation of human colon tumor organoids, May 29, Proc. Natl. Acad.
Sci. USA 115 (22) (2018) E5066–e5075, https://doi.org/10.1073/
pnas.1802902115.
[55] O. Saatci, A. Kaymak, U. Raza, et al., Targeting lysyl oxidase (LOX) overcomes
chemotherapy resistance in triple negative breast cancer, May 15, Nat. Commun.
11 (1) (2020) 2416, https://doi.org/10.1038/s41467-020-16199-4.
[56] O. Canli, A.M. Nicolas, J. Gupta, et al., Myeloid cell-derived reactive oxygen
species induce epithelial mutagenesis, Dec 11, Cancer Cell 32 (6) (2017) 869–883
e5, https://doi.org/10.1016/j.ccell.2017.11.004.
[57] M.C. Liebl, T.G. Hofmann, The role of p53 signaling in colorectal cancer, Apr 28,
Cancers 13 (9) (2021), https://doi.org/10.3390/cancers13092125.
[58] R.M. Xicola, Z. Manojlovic, G.J. Augustus, et al., Lack of APC somatic mutation is
associated with early-onset colorectal cancer in African Americans, Dec 13,
Carcinogenesis 39 (11) (2018) 1331–1341, https://doi.org/10.1093/carcin/
bgy122.
[59] A. Figer, L. Irmin, R. Geva, D. Flex, A. Sulkes, E. Friedman, Genetic analysis of the
APC gene regions involved in attenuated APC phenotype in Israeli patients with
early onset and familial colorectal cancer, Aug 17, Br. J. Cancer 85 (4) (2001)
523–526, https://doi.org/10.1054/bjoc.2001.1959.
[60] S. Kobylewski, M.F. Jacobson, Toxicology of food dyes (Jul-Sep), Int. J. Occup.
Environ. Health 18 (3) (2012) 220–246, https://doi.org/10.1179/
1077352512z.00000000034.
[61] H.S. Jabeen, S. ur Rahman, S. Mahmood, S. Anwer, Genotoxicity assessment of
amaranth and allura red using Saccharomyces cerevisiae (Jan), Bull. Environ.
Contam. Toxicol. 90 (1) (2013) 22–26, https://doi.org/10.1007/s00128-012-
0870-x.
[62] K.T. Chung, G.E. Fulk, M. Egan, Reduction of azo dyes by intestinal anaerobes
(Mar), Appl. Environ. Microbiol. 35 (3) (1978) 558–562.
[63] J.R. Hebert, E.A. Frongillo, S.A. Adams, et al., Perspective: randomized controlled
trials are not a panacea for diet-related research, Adv. Nutr. (Bethesda, Md) 7 (3)
(2016) 423–432, https://doi.org/10.3945/an.115.011023.
[64] J.R. Hebert, D.R. Miller, Methodologic considerations for investigating the diet-
cancer link, Am. J. Clin. Nutr. 47 (1988) 1068–1077.
Q. Zhang et al.