Short-term treatment with eicosapentaenoic acid improves inflammation and affects colonic differentiation markers and microbiota in patients with ulcerative colitis

Abstract

Patients with long-standing ulcerative colitis (UC) have an increased colorectal cancer (CRC) risk. In this pilot study we evaluated the effect of Eicosapentaenoic acid as free fatty acid (EPA-FFA) supplementation on mucosal disease activity, colonic differentiation markers and microbiota composition in UC patients. Twenty long-standing UC patients in stable clinical remission and with fecal calprotectin (FC) > 150 µg/g were enrolled (T0) and supplemented with EPA-FFA 2 g/daily for 90 days (T3). Endoscopic and histologic disease activities were measured by Mayo and Geboes scores, respectively. HES1, KLF4, STAT3, IL-10 and SOCS3 levels were determined using western blotting and qRT-PCR, while phospho-STAT3 levels were assessed by western blotting. Goblet cells were stained by Alcian blue. Microbiota analyses were performed on both fecal and colonic samples. Nineteen patients completed the study; seventeen (89.5%) were compliant. EPA-FFA treatment reduced FC levels at T3. Patients with FC > 150 µg/g at T3 (n = 2) were assumed as non-responders. EPA-FFA improved endoscopic and histological inflammation and induced IL-10, SOCS3, HES1 and KLF4 in compliant and responder patients. Importantly, long-term UC-driven microbiota composition was partially redressed by EPA-FFA. In conclusion, EPA-FFA supplementation reduced mucosal inflammation, promoted goblet cells differentiation and modulated intestinal microbiota composition in long-standing UC patients.

Full text

1
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
www.nature.com/scientificreports
Short-term treatment with
eicosapentaenoic acid improves
inammation and aects colonic
dierentiation markers and
microbiota in patients with
ulcerative colitis
Anna Prossomariti1,2, Eleonora Scaioli1, Giulia Piazzi2, Chiara Fazio1,2, Matteo Bellanova1,
Elena Biagi3, Marco Candela3, Patrizia Brigidi3, Clarissa Consolandi4, Tiziana Balbi5, Pasquale
Chieco2, Alessandra Munarini1,2, Milena Pariali2, Manuela Minguzzi1,2, Franco Bazzoli1,
Andrea Belluzzi6 & Luigi Ricciardiello
1
Patients with long-standing ulcerative colitis (UC) have an increased colorectal cancer (CRC) risk.
In this pilot study we evaluated the eect of Eicosapentaenoic acid as free fatty acid (EPA-FFA)
supplementation on mucosal disease activity, colonic dierentiation markers and microbiota
composition in UC patients. Twenty long-standing UC patients in stable clinical remission and with
fecal calprotectin (FC) > 150 µg/g were enrolled (T0) and supplemented with EPA-FFA 2 g/daily for 90
days (T3). Endoscopic and histologic disease activities were measured by Mayo and Geboes scores,
respectively. HES1, KLF4, STAT3, IL-10 and SOCS3 levels were determined using western blotting and
qRT-PCR, while phospho-STAT3 levels were assessed by western blotting. Goblet cells were stained by
Alcian blue. Microbiota analyses were performed on both fecal and colonic samples. Nineteen patients
completed the study; seventeen (89.5%) were compliant. EPA-FFA treatment reduced FC levels at
T3. Patients with FC > 150 µg/g at T3 (n = 2) were assumed as non-responders. EPA-FFA improved
endoscopic and histological inammation and induced IL-10, SOCS3, HES1 and KLF4 in compliant and
responder patients. Importantly, long-term UC-driven microbiota composition was partially redressed
by EPA-FFA. In conclusion, EPA-FFA supplementation reduced mucosal inammation, promoted goblet
cells dierentiation and modulated intestinal microbiota composition in long-standing UC patients.
Patients with ulcerative colitis (UC) have an increased risk to develop colitis-associated cancer (CAC) which is
proportionally related to the duration and the extent of the disease1. Current strategies to prevent CAC devel-
opment are mainly based on endoscopic surveillance in order to intercept and eradicate dysplasia which can
evolve to a malignant transformation2. However, persistent active intestinal inammation may hamper the iden-
tication of dysplastic areas during endoscopy. us, despite the reduction of advanced cancer incidence rates,
obtained through a regular endoscopic surveillance, critical goals for CAC prevention remain to preserve a con-
dition of histological remission3, 4, and to have predictive markers indicating those patients in whom endoscopic
1Department of Medical and Surgical Sciences, University of Bologna, Bologna, Italy. 2Center for Applied
Biomedical Research (CRBA), S.Orsola-Malpighi Hospital, University of Bologna, Bologna, Italy. 3Department
of Pharmacy and Biotechnology, University of Bologna, Bologna, Italy. 4Institute of Biomedical Technologies–
National Research Council (ITB-CNR), Segrate, Milan, Italy. 5Pathology Unit, S. Orsola-Malpighi Hospital, Bologna,
Italy. 6Gastroenterology Unit, S. Orsola-Malpighi Hospital, Bologna, Italy. Anna Prossomariti and Eleonora Scaioli
contributed equally to this work. Correspondence and requests for materials should be addressed to L.R. (email:
luigi.ricciardiello@unibo.it)
Received: 10 April 2017
Accepted: 4 July 2017
Published: xx xx xxxx
OPEN

www.nature.com/scientificreports/
2
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
surveillance would be more eective. Fecal calprotectin (FC) is a cytosolic protein belonging to the S100 protein
family, abundant in neutrophil granulocytes5, which represents a good predictor of endoscopic activity also in
asymptomatic UC patients6.
Several relevant molecular mechanisms contribute to the malignant epithelial transformation during chronic
intestinal inammation. Among these, aberrant activation of the signal transducer and activator of transcription
3 (STAT3), Interleukin (IL)-10 deciency or impaired function are critically involved in the onset of CAC7, 8.
Moreover, a thin and penetrable mucus layer, allowing a direct contact of bacteria with the epithelium, can
lead to persistent colonic inammation, thus promoting colon cancer development in UC patients9. Indeed, an
over-growth of mucosal and fecal bacteria in inamed colonic mucosa has been observed in UC patients, thus
supporting a critical role of the intestinal microbiota in the pathogenesis of UC and progression to CAC10, 11.
e canonical Notch signalling pathway, through the modulation of the transcriptional target hairy and
enhancer of split 1 (HES1), the antagonists atonal homolog 1 (HATH1) and kruppel-like factor 4 (KLF4) target,
is crucial to preserve a proper intestinal dierentiation12, 13. Our group recently proposed a tumor suppressor
function of HES1 during CAC progression14, while the role of KLF4 in CAC is still controversial15.
The abnormal regulation of these transcriptional factors result in a compromised epithelial differen-
tiation which can lead to an inefficient control of pathogenic microbes growth, favoring a tumor-prone
microenvironment16.
e use of anti-inammatory agents as tools for CAC prevention has been an intense focus of research17, 18.
To date, there are no uncontested chemopreventive agents for CAC. We have recently demonstrated that a
diet-containing highly-pure 1% eicosapentaenoic-acid as free fatty acid (EPA-FFA), an ω-3 polyunsaturated fatty
acid (ω-3 PUFA), was able to prevent colon cancer initiation and promotion in the azoxymethan/dextran sodium
sulfate (AOM-DSS) mouse model14. In the present exploratory study patients with long-standing UC in stable
clinical remission and active inammation identied by increased FC values, were supplemented with EPA-FFA
in order to test its eects on relevant mechanisms associated with UC disease and progression to CAC.
Results
EPA-FFA supplementation induces FC reduction and favors endoscopic and histological remis-
sion. In this study, twenty patients with long-standing UC were enrolled. Aer baseline colonoscopy, one
patient presented a clinical relapse before starting EPA-FFA supplementation, and was excluded from the trial.
Nineteen patients completed the study. e clinico-pathological features of all patients (n = 19) at T0 are shown
in Table1. Noteworthy, during EPA-FFA supplementation, no clinical relapse was observed.
Fatty acids composition was evaluated on RBC-puried membranes. Compared to T0, EPA-FFA supplemen-
tation led to a signicant increase of EPA (P < 0.0001; Fig.1a). e mean percentage values of EPA content
changed from 0.26 at T0 to 2.51 at T3. Capsules counting revealed that seventeen patients were adherent to
treatment with an overall compliance of 89.5%. Since EPA can be converted into the ω-3 PUFA docosahexaenoic
acid (DHA) in vivo through docosapentaenoic acid (DPA)19, we also measured the overall ω-3 PUFAs content
including EPA, DPA and DHA, in our patients. Interestingly, the combined percentage content of EPA, DPA
and DHA was signicantly increased at T3 compared to T0 (P < 0.0001; Supplementary FigureS1a), while the
percentage content of ω-6 PUFAs (arachidonic + linoleic acids) was unchanged upon EPA-FFA supplementation
(Supplementary FigureS1b).
Importantly, a signicant reduction of FC at T3 was observed (P < 0.0001; Fig.1b). e mean FC values
changed from 230 at T0 to 87.7 µg/g at T3. No side eects or serious adverse events were reported during the
trial. Two patients maintained FC levels >150 µg/g at T3 aer treatment and were considered non-responders.
Patients' Characteristics
Age, years median (range) 45 (23–80)
Male, n (%) 13 (68.4)
Current smokers, n (%) 1 (5.3)
BMI median (range) 24.16 (18.5–34)
Duration of UC, years median (range) 12 (8–27)
Time of remission, months median (range) 24 (4–60)
Fecal Calprotectin, (µg/g)median (range) 220 (150–300)
C-Reactive Protein, (mg/L)median (range) 0.3 (0.04–1.25)
SCCAI clinical score >3 n (%) 0 (0)
Mayo endoscopic sub-score ≥1 n (%) 13 (68.4)
Geboes histological score ≥3.1 n (%) 7 (36.8)
Concomitant medication, n (%)
Mesalamine 11 (57.9)
Azathioprine 2 (10.5)
Mesalamine + Azathioprine 4 (21.0)
Anti-TNFα1 (5.3)
None 1 (5.3)
Table 1. Clinico-pathological characteristics of patients at baseline (T0; n = 19).

www.nature.com/scientificreports/
3
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
EPA-FFA treatment signicantly promoted endoscopic and histological remission in compliant and responder
patients (n = 15).
Indeed, compared to baseline, endoscopic improvement was observed in 8 patients while no variations were
observed in 7 (P = 0.004; Fig.2a). Moreover, a resolution of histological inammation at T3 was observed in 5
patients, while the histological score remained unchanged in 10 (P = 0.03; Fig.2b). Endoscopic and histological
worsening was not observed.
EPA-FFA supplementation induces both IL-10 and SOCS3 expression reducing STAT3 acti-
vation. To elucidate the mechanisms responsible for the protective effect of EPA-FFA in patients with
long-standing UC, we investigated the modulation of the IL-10/STAT3/SOCS3 axis in compliant and responder
patients (n = 15). Compared to T0, we observed a concomitant signicant up-regulation of IL-10 (P = 0.03;
Fig.3a) and SOCS3 (P = 0.04; Fig.3b) mRNA levels at T3 associated with an increasing trend in IL-22 mRNA
(Supplementary FigureS2). Otherwise, no signicant dierences in IL-10 and SOCS3 protein expressions were
observed (Supplementary FigureS3a and b). Since STAT3 represents one of the major regulators of SOCS3, we
decided to characterize STAT3 activation in these patients. Treatment with EPA-FFA reduced STAT3 Tyr705
phosphorylation (p-STAT3) in 60% of patients (9/15) (Fig.3c and d), while not aecting STAT3 transcription
(Supplementary FigureS4). Noteworthy, 4/5 patients showing highest levels of p-STAT3 at T3 were poor compli-
ant patients with lower percentage of EPA in RBC aer supplementation. ese data suggest that over-expression
of SOCS3 following EPA-FFA supplementation, probably as a downstream eect of IL-10 induction, reduces
STAT3 activation, in particular in patients with highest percentage of EPA in RBC membranes. Correlation anal-
yses revealed a signicant positive correlation at T3 between transcriptional levels of SOCS3 and both IL-10
mRNA levels (P = 0.02; Supplementary FigureS5a) and p-STAT3 protein (P = 0.03; Supplementary FigureS5b),
thus supporting our hypothesis.
EPA-FFA supplementation modulates HES1 and KLF4 and stimulates goblet cell dierentia-
tion. Notch signaling, through the modulation of the transcriptional targets HES1 and KLF4, is crucial to pre-
serve a proper balance between the absorptive and the secretory cell lineages of the intestine13, 20. We observed a
signicant up-regulation of HES1 (P = 0.02; Fig.4a and c) and KLF4 proteins (P = 0.04; Fig.4b and c) in patients
with long-standing UC at T3 compared to T0, while no dierences were observed at mRNA level (Supplementary
FigureS6a and b). Correlation analysis indicated that these two transcription factors positively correlate with
each other (P = 0.0007; Supplementary FigureS7). Importantly, although no variations in the MUC2 mRNA
(Supplementary FigureS8a) and protein were found (Supplementary FigureS8b), compared to T0, in which
Figure 1. (a) Eicosapentaenoic acid (EPA; C20:5 n-3) percentage in RBCs and (b) FC levels (µg/g) (B) in all
patients (n = 19) at T0 and T3. Statistical signicance was calculated using the paired two-tailed t-test. Data are
shown as mean ± SEM.
Figure 2. (a) Mayo endoscopic score and (b) Geboes histological score in compliant and responder patients
(n = 15) at T0 and T3. Data are presented as percentage of patients according to Mayo and Geboes cut-os.

www.nature.com/scientificreports/
4
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
goblet cells depletion was found in 20% of patients, daily supplementation of EPA-FFA for 3 months was asso-
ciated with a signicant increased number of goblet cells in the colon (P = 0.04; Fig.4d and e). us, our results
unveil a role of EPA-FFA in improving secretory lineage dierentiation and intestinal epithelial cells turnover
through simultaneous induction of KLF4 and HES1. We found no dierences in terms of intestinal proliferation
measured by Ki-67 (Supplementary FigureS9a), c-MYC (Supplementary FigureS9b) and LGR5 upon EPA-FFA
supplementation (Supplementary FigureS9c).
EPA-FFA modulates the gut microbiota composition in UC patients. Given the critical role of
intestinal microbial imbalance in the pathogenesis of UC, the fecal and mucosal microbiota compositions were
also assessed in our patients.
Sequences are available at the following MG-Rast link:
http://metagenomics.anl.gov/mgmain.html?mgpage=project&project=mgp80642.
To identify the main microbiota dysbioses associated with the long-term UC disease, the fecal microbiota
composition of UC patients at T0 was compared to that of a group of Italian healthy adults (age 22–48 years,
enrolled in the same geographical area of the UC patients)21. An enrichment of the families Clostridiaceae (4.7
vs. 1%, P = 0.003; in particular genus SMB53, P = 0.001) and Ruminococcaceae (35.7 vs. 24.1%, P = 0.008), and
depletion of Verrucomicrobiaceae (0 vs. 0.4%, P = 0.002; in particular genus Akkermansia, 0 vs. 0.4%, P = 0.002),
Peptostreptococcaceae (0 vs. 0.3%, P = 0.0009) and Porphyromonadaceae (0 vs. 0.5%, P = 0.006; in particular
genus Parabacteroides, 0 vs. 0.5%, P = 0.006) families was found in UC patients at T0 (Fig.5a and b). Noteworthy
EPA-FFA supplementation increased Porphyromonadaceae (from 0 to 0.2%) and decreased Ruminococcaceae
(from 35.7 to 28%) (Fig.5b and c) in feces of UC patients. In addition, EPA-FFA had also eects on mucosal
microbiota of UC patients by decreasing the abundance of mucosal-adherent members of the Bacteroidaceae
family (in particular belonging to the genus Bacteroides, 27.4 vs. 14.7%) (Fig.5d and e).
Discussion
Dierent therapeutic approaches have been tested for CAC prevention in patients with Inammatory bowel dis-
ease (IBD) over the years. Importantly, an increasing number of data obtained from in vitro experiments, as well
as, animal and clinical studies support a protective role for ω-3 PUFAs (EPA and DHA) in gastrointestinal cancer
prevention including CRC (as reviewed by Eltweri et al.22).
However, data on pharmacological and natural compounds as anticancer agents in IBD patients are elusive and
still inconsistent23. It is well known that symptoms in IBD and serum biomarkers do not always properly mirror the
Figure 3. mRNA expression levels of (a) IL-10 and (b) SOCS3. Protein levels of (c) p-STAT3/STAT3 on
homogenized sigmoid colon tissues in compliant and responder patients (n = 15) at T0 and T3. Statistical
signicance was obtained using one-sample two-tailed t-test. Data are shown as mean of square root
transformed values ± SEM. (d) Western blot representative images of p-STAT3 (Y705) and STAT3 at T0 and T3
(n = 3).

www.nature.com/scientificreports/
5
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
inammatory degree of the mucosa24. FC is becoming the most useful non-invasive tool for monitoring the inamma-
tory status of the mucosa and the response to therapy, as well as for predicting clinical relapse in IBD patients25.
In this pilot study we tested, for the first time, the effects of EPA-FFA on asymptomatic patients with
long-standing UC in clinical remission who retained high FC levels (> 150 µg/g) despite stable maintenance
therapy. We found that short-term EPA-FFA supplementation at a dosage of 2 g/daily was associated with a signif-
icant increase of EPA and overall ω-3 PUFAs content (EPA, DPA and DHA) into RBCs, suggesting that EPA was
incorporated by most of patients (17/19) and eciently converted into DPA and DHA.
Although the primary end-point of this study was not to test the clinical benet of EPA-FFA but to explore its
eect on mucosal inammation and new potential chemopreventive mechanisms during long-standing UC, we
found that short-term EPA-FFA supplementation reduced mucosal inammation (with a signicant drop in FC)
favoring an improvement of both endoscopic and histological inammation in almost all patients. Since FC levels
have been demonstrated to correlate with the intensity of the neutrophilic inltrate12, our results, supported by
histological evaluation, indicate that EPA-FFA improved the inammatory state in patients with long-standing
UC. We believe that our results could be explained, at least in part, by the free-fatty acid-highly pure formulation
of EPA used in this study.
Previous evidence support a protective role for ω-3 PUFAs intake including both EPA and DHA in the pre-
vention of CRC in dierent settings26–28. However, data from ω-3 PUFAs supplementation in patients and murine
models of UC are still controversial29–31, and the impact of dietary ω-3 PUFAs supplementation for CAC preven-
tion is poorly dened.
Given the increased content of ω-3 PUFAs in our patients, it is reasonable to speculate that the observed
protective eects may be due to both EPA and DHA. Noteworthy, we found no relevant dierences in the ω-6
PUFAs content upon EPA-FFA supplementation. is result could be explained by the unchanged dietary habits
of enrolled patients during the study. Strikingly, the increased ω-3 PUFAs content was sucient to induce a rel-
evant protective response in UC patients, while possibly maintaining the same ω-6 PUFAs content as previously
suggested32.
In this study, in order to characterize the EPA-FFA short-term eects in long-standing UC patients, we rst
focused on the eects of EPA-FFA supplementation on IL-10/STAT3/SOCS3 signalling. e role of STAT3 during
UC is actually controversial. Indeed, studies on animal models of IBD suggested both a deleterious and protective
role of STAT3 hyperactivation during colitis33, 34. Importantly, increased levels of phospho-STAT3 were detected
in patients with active UC, as well as in dysplasia and cancer, while a progressive decreasing trend of SOCS3
levels was observed from low-grade dysplasia to UC-CRC35. However, more recent evidence obtained in UC
patients supported a role of SOCS3 over-expression in short-term disease relapse and mucosal inammation
impairing STAT3 activation36, 37. In this study, we found a concomitant signicant up-regulation of IL-10 and
SOCS3 mRNA upon EPA-FFA supplementation with a reduction of STAT3 activation in most of the patients
with highest EPA percentage levels at T3. However, no changes in IL-10 and SOCS3 proteins were appreciated
Figure 4. Protein expression levels of (a) HES1 and (b) KLF4 on homogenized sigmoid colon tissues in
compliant and responder patients (n = 15) at T0 and T3. Statistical signicance was measured using one-
sample two-tailed t-test. Data are shown as mean of square root transformed values ± SEM. (c) Western blot
representative images of HES1 and KLF4 at T0 and T3 (n = 3). (d) Alcian blue ranks and (e) representative
images of goblet cells staining at T0 (le panel) and T3 (right panel). Statistical signicance for Alcian blue
ranks was calculated using the paired two-tailed t-test. Data are shown as mean of ranks ± SEM.

www.nature.com/scientificreports/
6
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
upon EPA-FFA supplementation in our patients. As previously suggested by literature data, we hypothesized that
multiple post-transcriptional mechanisms may contribute to regulate SOCS3 and IL-10 proteins, thus explaining
the absence of a correlation between changes in their mRNA and protein levels38–42.
Importantly, considering responders, we did not observe endoscopic or histological worsening in any patient.
us, our data support a protective role of EPA-FFA during UC remission by turning o STAT3 activation
through SOCS3 transcriptional induction.
Notch signalling is also a key determinant for sustaining intestinal epithelial cells dierentiation and turno-
ver, for the integrity of the mucosal barrier, as well as for regulating malignant epithelial transformation in the
colon20. Evidence show possible oncogenic and tumor suppressor activities of HES1 and KLF4 in sporadic set-
tings, respectively43, 44. We previously showed in the AOM-DSS mouse model a loss of Notch1 signalling during
CAC development partially counteracted by EPA-FFA supporting a tumor-suppressor role of this pathway during
inammation-induced intestinal tumorigenesis14. Accordingly, Garg and colleagues previously demonstrated in
the same animal model, that Matrix metalloproteinase-9 (MMP-9), activating Notch1signalling and controlling
p53 cascade, exerts a strong protective eect toward CAC development45. Otherwise, in a recent in vitro work
conducted by our research group, we observed a MMP-9-dependent activation of Notch1 signalling in CRC cells
exposed to a conditioned medium (CM) containing multiple pro-inammatory cytokines secreted by activated
macrophages. e activation of MMP-9/Notch signalling was associated with increased CRC cells invasiveness,
suggesting a tumor-prone role of Notch1 signalling in sporadic CRC. Interestingly, EPA-FFA pre-treatment of
CM-exposed CRC cell lines led to reduced invasion through a Notch1 signalling switch o46. ese results, as
recently reviewed by our research group47, clearly indicate that the cell response to Notch signalling activation is
not univocal resulting in oncogenic or tumor-suppressive mechanisms depending on the specic pathological
context.
Figure 5. Median fecal microbiota composition at family level in (a) healthy adults, (b) UC patients at T0,
(c) UC patient at T3 and colon biopsies of (d) UC patients at T0 and (e) UC patients at T3, represented as pie
chart, in available samples from compliant and responder patients. Average relative abundance of families
representing at least 0.2% of the total microbiota in at least 10% of the sequenced samples are showed. Color
code for the most abundant bacterial families (present at an average abundance >1% in at least one group
of samples) is reported in approximate decreasing abundance order. Mann-Whitney U test was used to test
dierences among median groups.

www.nature.com/scientificreports/
7
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
In this study we demonstrated that EPA-FFA modulated intestinal dierentiation inducing both HES1 and
KLF4 proteins and increasing the number of goblet cells.
Patients with UC in remission are generally characterized by an intact mucus layer, although a defective and
penetrable intestinal barrier could be retained in some cases48. KLF4 has a crucial role on both maturation and
dierentiation of goblet cells in the colon49, and a critical role for IL-10 in the regulation of goblet cells activation
during inammation has been also previously described50. Moreover, microbiota analysis performed in our study
shows that the gut microbiota population constituents present in the UC group at T0 were partly modulated by
the EPA-FFA treatment. Indeed, the Porphyromonadaceae genus Parabacteroides, known to be decreased in UC51,
was signicantly increased in T3 samples compared to T0. Also EPA-FFA showed the capability to reduce the
fecal amount of Clostridium spp. compared to T0. Interestingly, these proteolytic microorganisms were known to
induce mucolytic metabolism in other species, i.e. Bacteroides52. Noteworthy, mucosal-adherent members of the
Bacteroides genus, known to include mucolytic species, were found to be decreased aer EPA-FFA treatment, pos-
sibly contributing to the protection of the epithelium. us, we hypothesize that the ability of EPA-FFA treatment
to promote goblet cells population could be a result of multiple mechanisms including the induction of KLF-4 and
IL-10, as well as the reduction of mucolytic bacteria. However, despite the impact of EPA-FFA supplementation
on important modulatory functions in UC patients, our study has some limitations and should be taken with
caution. Firstly, this explorative study involved a small number of subjects. Secondly, no patients received placebo
and we exclusively tested a single dosage of EPA-FFA. irdly, we cannot completely rule out that the clinical
outcome observed in our patients could be the result of a potential combined eect between EPA-FFA and stable
maintenance of UC therapies, such as 5-aminosalicylic acid which is taken by most of the recruited patients.
In conclusion, in this pilot study EPA-FFA improved endoscopic and histological inammation, aected the
IL-10/STAT3/SOCS3 cascade, stimulated goblet cells dierentiation and modulated the long-term UC-related
colonic alterations of intestinal microbiota. Future larger placebo controlled trials should be conducted in order
to conrm these results and to evaluate long-term eects of EPA-FFA supplementation on disease relapse and
CAC risk.
Methods
Study design. Eligible patients were asymptomatic subjects aged 18–70 years with long-standing (≥ 8 years)
UC, in stable clinical remission (simple clinical colitis activity index; SCCAI = 0), and FC levels higher than
150 µg/g53. Patients were included in the study aer signing the informed written consent. Concomitant stable
therapies for UC (mesalamine, immunomodulators and/or biological drugs) without modications in the previ-
ous 3 months were allowed. Exclusion criteria were: (1) recent use of steroids (< 2 months) or other experimental
drugs (< 3 months); (2) concomitant use of anticoagulants; (3) probiotic use; (4) pregnancy or breast-feeding;
(5) known or suspected hypersensitivity to eicosapentaenoic acid or ω-3 PUFAs; and (6) severe co-morbidities.
Subjects were given oral supplementation of 2 g/daily (two 500 mg capsules twice a day) of EPA-FFA (ALFA™,
SLA Pharma AG, Switzerland) for 90 days. During the study, subjects were asked to keep their dietary habits.
Patients underwent endoscopic examination at enrollment (T0) and aer 90 days of EPA-FFA supplementation
(T3). Six biopsies were taken from the sigmoid colon at each time point. Blood samples were obtained for isola-
tion of peripheral erythrocytes. Adherence to EPA-FFA supplementation was evaluated both by capsule counting
and assessing EPA incorporation into red blood cell (RBC) membranes. Compliant patients were considered
those who consumed at least 80% of the capsules, without interruption of the protocol for more than 14 con-
secutive days. e study was conducted in accordance to the Declaration of Helsinki and approved by the Ethic
Committee of the S.Orsola-Malpighi Hospital (Bologna, Italy).
e trial was registered on ClinicalTrials.gov with Identier: NCT02069561 on 19/02/2014. (https://clinical-
trials.gov/ct2/show/NCT02069561)].
Fecal calprotectin dosage. Fecal samples were collected within 24 hours before endoscopy and stored at
2–8 °C until assaying. Quantication of FC was carried out using CalFast (Eurospital, Trieste, Italy) according to
the manufacturer’s protocol. FC values >150 μg/g were considered predictive of mucosal endoscopic activity as
previously demonstrated53.
Endoscopic and histological evaluation. Two investigators (L.R., E.S.) performed all endoscopies.
According to the Mayo endoscopic sub-score, a cut-o ≥ 1 was used to discriminate the presence of endoscopic
inammation54. Histological activity was assessed by one expert blinded pathologist (T.B.) and scored according
to the Geboes grading system55. A Geboes cut-o score ≥ 3.1 was assumed to dene active histological inam-
mation56. When biopsies showed dierent degrees of activity, the highest degree of inammation was considered.
Acidic mucins quantication. Formalin-xed and paran-embedded (FFPE) biopsies were de-waxed in
toluene for 10 minutes, rehydrated, placed in the Alcian blue solution (Alcian blue 8GX in 3% acetic acid solu-
tion pH 2.5) for 30 minutes and counterstained with hematoxylin. For analysis, slides were placed in order of
increasing Alcian blue staining intensity using a rank order scoring system (1 = lower rank; 36 = higher rank).
Rank ordering method has been shown to be better than categorical scoring system to identify subtle dierences
between groups57.
Immunoistochemistry. Immunohistochemistry (IHC) was performed on FFPE colonic sections. Slides
were dewaxed, subjected to endogenous peroxidase inhibition, rehydrated and treated with citrate buer (pH 6.0)
at 120 °C for 15 minutes for antigen retrieval. en, slides were incubated overnight at +4 °C with the monoclonal
antibodies against Ki-67 and MUC2 (Supplementary Table1). Aer incubation with secondary antibody Rabbit/
Mouse (1:1000, DAKO EnVision™ System Peroxidase), the signal was detected with diaminobenzidine (DAB)

www.nature.com/scientificreports/
8
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
(Sigma-Aldrich, Saint Louis, Missouri, USA). Percentages of Ki67 positive nuclei and MUC2 positive DAB areas
were quantied using ImageJ soware (NIH, Bethesda, MD, USA).
Membrane fatty acid analysis. Membrane fatty acids content was measured in RBCs. Lipids extrac-
tion from RBC membranes, phospholipids separation and sample preparation were performed as previously
described58. Extracted fatty acid methyl-esters were then analyzed by gas-chromatography mass-spectrometry
(GC-MS). Fatty acid levels were expressed as relative percentages of total fatty acids.
Western Blotting. Total protein lysates were isolated from biopsies by sonication in RIPA buer. Forty µg of
proteins for each sample were separated on a 4–12% NuPAGE Novex Bis-Tris Gels (Invitrogen™, ermo Fisher
Scientic, Waltham, Massachusetts, USA) in MOPS buer (Novex™, ermo Fisher Scientic) and transferred
onto nitrocellulose membrane. Aer blocking, membranes were incubated overnight at +4 °C with primary anti-
bodies against HES1, KLF4, phosphorylated STAT3 (Y705), STAT3, IL-10, SOCS3 and GAPDH (Supplementary
Table1). Aer incubation with appropriate secondary Horse-Radish-Peroxidase (HRP) conjugated antibod-
ies (GE Healthcare Life Sciences, Little Chalfont, United Kingdom), the signal was detected with a luminol
enhancer solution (WESTAR EtaC, Cyanagen, Bologna, Italy) and images were acquired using the ChemidocTM
XRS + (Biorad, Hercules, CA, USA). Densitometric analysis performed using Image Lab™ soware.
Gene expression analysis. Total RNA was extracted from biopsies using Trizol® (Ambion, ermo Fisher
Scientic). One µg of total RNA was converted to cDNA using the High-Capacity RNA-to-cDNA™ Kit (Applied
Biosystems™, ermo Fisher Scientic) according to the manufacturer’s instructions. qRT-PCR reactions were
performed in duplicate on a MX3000p QPCR thermal cycler (Stratagene, San Diego, CA, USA) using the SYBR®
Select Master Mix for CFX (Applied Biosystems™, ermo Fisher Scientic) and the specic primers for IL-
10, IL-22, LGR5, C-MYC, MUC2, HES1 and KLF4. e primers sequences are listed on Supplementary Table2.
mRNA expressions of SOCS3 and STAT3 were analyzed using a 5′ nuclease probe (Assay ID: Hs.PT.58.4303529;
Integrated DNA Technologies, Coralville, Iowa, USA) and the Taqman® gene expression assay (Hs00374280_m1;
ermo Fisher Scientic), respectively. Fold induction levels were obtained using the 2−ΔΔCt method by normal-
izing against the reference gene RPS9.
Microbiota analysis. Fecal samples were collected prior to the endoscopic preparation while mucosal sam-
ples were taken during endoscopy.
Total bacterial DNA was extracted from feces using QIAamp DNA Stool Mini Kit (QIAGEN, Hilden,
Germany) and from biopsies using DNeasy Blood & Tissue Mini Kit (Qiagen). Due to a poor quality or quan-
tity of extracted DNA, data on fecal and mucosal microbiota were available from 14 and 16 of the 19 patients
included in the study, respectively. For all samples the V3–V4 region of the bacterial 16S rRNA gene was amplied
and sequenced using the Illumina platform (Illumina, San Diego, CA) using a 2 × 300 bp paired-end protocol.
Indexed libraries were pooled at equimolar concentrations, denatured and diluted to 6 pmol/L before loading
onto the MiSeq ow cell. Raw sequences were processed using a pipeline combining PANDAseq [S6] and QIIME
[S7]. High-quality reads were binned into operational taxonomic units (OTUs) at a 0.97 similarity threshold
using UCLUST [S8] and a “de novo” approach. Taxonomy was assigned using the RDP classier against the
Greengenes database (May 2013 release). All singleton OTUs were removed in an attempt to discard the majority
of chimera sequences. Relative abundance proles at family or genus level were obtained and plotted. For fecal
microbiota analysis, a comparison with a control population of Italian healthy adults enrolled in a previous study
was also performed21. Fecal samples from healthy subjects were collected and processed using the same proce-
dures applied for UC patients recruited in this study.
Statistical analysis. Data were analyzed with Graphpad 5.0 Soware (GraphPad Soware Inc., CA, USA)
and Statistix 9.0. e means of two matched groups (T0 vs. T3) were compared using the paired two-tailed
t-test. For statistical analysis (based on fold-changes) the mean of T0 samples was assumed as 1 and two-tailed
one-sample t-test was used to compare dierences between T0 and T3. Sign test, a test for analyzing simple +/−
dierences between paired comparisons59, was used to analyze dierences in the Mayo sub-score and Geboes
score. Correlation analyses were carried out using Spearman’s correlation coecient (rs). For qRT-PCR and
western blot analyses data were presented upon square-root transformation. For microbiota analysis, median
dierences among groups were tested using a non parametric approach (Mann-Whitney U test); P values were
corrected for multiple comparisons using the Benjamini-Hochberg method. P values < 0.05 were considered
statistically signicant.
All data generated or analysed during this study are included in this published article (and its Supplementary
Information les).
References
1. Beaugerie, L. & Itzowitz, S. H. Cancers Complicating Inammatory Bowel Disease. N. Engl. J. Med. 373, 195 (2015).
2. Cairns, S. . et al. Guidelines for colorectal cancer screening and surveillance in moderate and high ris groups (update from 2002).
Gut 59, 666–89 (2010).
3. Choi, C.-H. . et al. Forty-Year Analysis of Colonoscopic Surveillance Program for Neoplasia in Ulcerative Colitis: An Updated
Overview. Am. J. Gastroenterol. 110, 1022–1034 (2015).
4. Pineton de Chambrun, G., Peyrin-Biroulet, L., Lémann, M. & Colombel, J.-F. Clinical implications of mucosal healing for the
management of IBD. Nat. ev. Gastroenterol. Hepatol. 7, 15–29 (2010).
5. Poullis, A., Foster, ., Mendall, M. A. & Fagerhol, M. . Emerging role of calprotectin in gastroenterology. J. Gastroenterol. Hepatol.
18, 756–62 (2003).
6. Mooiweer, E., Fidder, H. H., Siersema, P. D., Laheij, . J. F. & Oldenburg, B. Fecal calprotectin testing can identify ineective
colorectal cancer surveillance procedures in patients with longstanding colitis. Inamm. Bowel Dis. 20, 1079–84 (2014).

www.nature.com/scientificreports/
9
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
7. Grivenniov, S. et al. IL-6 and Stat3 Are equired for Survival of Intestinal Epithelial Cells and Development of Colitis-Associated
Cancer. Cancer Cell 15, 103–113 (2009).
8. Sturlan, S. et al. Interleuin-10-decient mice and inammatory bowel disease associated cancer development. Carcinogenesis 22,
665–71 (2001).
9. Johansson, M. E. V. et al. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and
patients with ulcerative colitis. Gut 63, 281–291 (2014).
10. Chu, H. et al. Gene-microbiota interactions contribute to the pathogenesis of inammatory bowel disease. Science 352, 1116–20
(2016).
11. Tjalsma, H., Boleij, A., Marchesi, J. . & Dutilh, B. E. A bacterial driver-passenger model for colorectal cancer: beyond the usual
suspects. Nat. ev. Microbiol. 10, 575–82 (2012).
12. Neurath, M. F. New targets for mucosal healing and therapy in inammatory bowel diseases. Mucosal Immunol. 7, 6–19 (2014).
13. atz, J. P. et al. The zinc-finger transcription factor lf4 is required for terminal differentiation of goblet cells in the colon.
Development 129, 2619–28 (2002).
14. Piazzi, G. et al . Eicosapentaenoic acid free fatty acid prevents and suppresses colonic neoplasia in colitis-associated colorectal cancer
acting on Notch signaling and gut microbiota. Int. J. cancer 135, 2004–13 (2014).
15. Ghaleb, A. M., Laroui, H., Merlin, D. & Yang, V. W. Genetic deletion of lf4 in the mouse intestinal epithelium ameliorates dextran
sodium sulfate-induced colitis by modulating the NF-κB pathway inammatory response. Inamm. Bowel Dis. 20, 811–20 (2014).
16. Sartor, . B. & Mazmanian, S. . Intestinal Microbes in Inammatory Bowel Diseases. Am. J. Gastroenterol. 1(Suppl.), 15–21 (2012).
17. Terdiman, J. P., Steinbuch, M., Blumentals, W. A., Ullman, T. A. & ubin, D. T. 5-Aminosalicylic acid therapy and the ris of
colorectal cancer among patients with inammatory bowel disease. Inamm. Bowel Dis. 13, 367–71 (2007).
18. Lasry, A., Zinger, A. & Ben-Neriah, Y. Inammatory networs underlying colorectal cancer. Nat. Immunol. 17, 230–40 (2016).
19. Arterburn, L. M., Hall, E. B. & Oen, H. Distribution, interconversion, and dose response of n-3 fatty acids in humans. Am. J. Clin.
Nutr. 83, 1467S–1476S (2006).
20. Obata, Y. et al. Epithelial cell-intrinsic Notch signaling plays an essential role in the maintenance of gut immune homeostasis. J.
Immunol. 188, 2427–36 (2012).
21. Biagi, E. et al. Gut Microbiota and Extreme Longevity. Curr. Biol. 26, 1480–5 (2016).
22. Eltweri, A. M. et al. Potential applications of fish oils rich in omega-3 polyunsaturated fatty acids in the management of
gastrointestinal cancer. Clin. Nutr. 36, 65–78 (2017).
23. Subramanian, V. & Logan, . F. Chemoprevention of colorectal cancer in inflammatory bowel disease. Best Pract. es. Clin.
Gastroenterol. 25, 593–606 (2011).
24. Vermeire, S., Van Assche, G. & utgeerts, P. L aboratory marers in IBD: useful, magic, or unnecessary toys? Gut 55, 426–31 (2006).
25. Sandborn, W. J. et al. Correlation Between Concentrations of Fecal Calprotectin and Outcomes of Patients With Ulcerative Colitis
in a Phase 2 Trial. Gastroenterology 150, 96–102 (2016).
26. Song, M . et al. Marine ω-3 polyunsaturated fatty acids and ris of colorectal cancer according to microsatellite instability. J. Natl.
Cancer Inst. 107 (2015).
27. Fini, L. et al. Highly puried eicosapentaenoic acid as free fatty acids strongly suppresses polyps in Apc(Min/+) mice. Clin. Cancer
es. 16, 5703–11 (2010).
28. Cocbain, A. J., Toogood, G. J. & Hull, M. A. Omega-3 polyunsaturated fatty acids for the treatment and prevention of colorectal
cancer. Gut 61, 135–149 (2012).
29. Matsunaga, H. et al. Omega-3 fatty acids exacerbate DSS-induced colitis through decreased adiponectin in colonic subepithelial
myobroblasts. Inamm. Bowel Dis. 14, 1348–57 (2008).
30. Jia, Q. et al. educed colitis-associated colon cancer in Fat-1 (n-3 fatty acid desaturase) transgenic mice. Cancer es. 68, 3985–91
(2008).
31. Anantharishnan, A. N. et al. Long-term intae of dietary fat and ris of ulcerative colitis and Crohn’s disease. Gut 63, 776–84
(2014).
32. Davidson, L. A. et al. Chemopreventive n-3 polyunsaturated fatty acids reprogram genetic signatures during colon cancer initiation
and progression in the rat. Cancer es. 64, 6797–804 (2004).
33. Suzui, A. et al. CIS3/SO CS3/SSI3 plays a negative regulatory role in STAT3 activation and intestinal inammation. J. Exp. Med. 193,
471–81 (2001).
34. Picert, G. et al. STAT3 lins IL-22 signaling in intestinal epithelial cells to mucosal wound healing. J. Exp. Med. 206, 1465–72
(2009).
35. Li, Y. et al. Disease-related expression of the IL6/STAT3/SOCS3 signalling pathway in ulcerative colitis and ulcerative colitis-related
carcinogenesis. Gut 59, 227–35 (2010).
36. Xu, A. T. et al. High suppressor of cytoine signaling-3 expression impairs STAT3-dependent protective eects of interleuin-22 in
ulcerative colitis in remission. Inamm. Bowel Dis. 21, 241–50 (2015).
37. Li, Y. et al . Increased suppressor of cytoine signaling-3 expression predicts mucosal relapse in ulcerative colitis. Inamm. Bowel Dis.
19, 132–40 (2013).
38. Tone, M., Powell, M. J., Tone, Y., ompson, S. A. & Waldmann, H. IL-10 gene expression is controlled by the transcription factors
Sp1 and Sp3. J. Immunol. 165, 286–91 (2000).
39. Ma, W. et al. e p38 mitogen-activated inase pathway regulates the human interleuin-10 promoter via the activation of Sp1
transcription factor in lipopolysaccharide-stimulated human macrophages. J. Biol. Chem. 276, 13664–74 (2001).
40. Sharma, A. et al. Posttranscriptional regulation of interleuin-10 expression by hsa-mi-106a. Proc. Natl. Acad. Sci. USA 106,
5761–6 (2009).
41. Haan, S. et al. Tyrosine phosphorylation disrupts elongin interaction and accelerates SOCS3 degradation. J. Biol. Chem. 278,
31972–9 (2003).
42. Sasa i, A. et al. e N-terminal truncated isoform of SOCS3 translated from an alternative initiation AUG codon under stress
conditions is stable due to the lac of a major ubiquitination site, Lys-6. J. Biol. Chem. 278, 2432–6 (2003).
43. eedij, M. et al. Activation of Notch signaling in human colon adenocarcinoma. Int. J. Oncol. 33, 1223–9 (2008).
44. Zhao, W. et al. Identication of rüppel-lie factor 4 as a potential tumor suppressor gene in colorectal cancer. Oncogene 23,
395–402 (2004).
45. Garg, P. et al. Notch1 regulates the eects of matrix metalloproteinase-9 on colitis-associated cancer in mice. Gastroenterology 141,
1381–92 (2011).
46. Fazio, C. et al. Inammation increases NOTCH1 activity via MMP9 and is counteracted by Eicosapentaenoic Acid-free fatty acid in
colon cancer cells. Sci. ep. 6, 20670 (2016).
47. Fazio, C. & icciardiello, L. Inammation and Notch signaling: a crosstal with opposite eects on tumorigenesis. Cell Death Dis. 7,
e2515 (2016).
48. Johansson, M. E. V. et al. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and
patients with ulcerative colitis. Gut 63, 281–91 (2014).
49. Ghaleb, A. M., McConnell, B. B., aestner, . H. & Yang, V. W. Altered intestinal epithelial homeostasis in mice with intestine-
specic deletion of the rüppel-lie factor 4 gene. Dev. Biol. 349, 310–320 (2011).

www.nature.com/scientificreports/
10
SCIENTIfIC RePoRtS | 7: 7458 | DOI:10.1038/s41598-017-07992-1
50. Hasnain, S. Z. et al. IL-10 Promotes Production of Intestinal Mucus by Suppressing Protein Misfolding and Endoplasmic eticulum
Stress in Goblet Cells. Gastroenterology 144, 357–368.e9 (2013).
51. ajilić-Stojanović, M., Shanahan, F., Guarner, F. & de Vos, W. M. Phylogenetic analysis of dysbiosis in ulcerative colitis during
remission. Inamm. Bowel Dis. 19, 481–8 (2013).
52. Fischbach, M. A. & Sonnenburg, J. L. Eating for two: how metabolism establishes interspecies interactions in the gut. Cell Host
Microbe 10, 336–47 (2011).
53. Scaioli, E. et al. Clinical application of faecal calprotectin in ulcerative colitis patients. Eur. J. Gastroenterol. Hepatol. 27, 1418–24
(2015).
54. Schroeder, . W., Tremaine, W. J. & Ilstrup, D. M. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active
ulcerative colitis. A randomized study. N. Engl. J. Med. 317, 1625–9 (1987).
55. Geboes, . et al. A reproducible grading scale for histological assessment of inammation in ulcerative colitis. Gut 47, 404–9 (2000).
56. Mooiweer, E. et al. Low fecal calprotectin predicts sustained clinical remission in inammatory bowel disease patients: a plea for
deep remission. J. Crohns. Colitis 9, 50–5 (2015).
57. Holland, T. e comparative power of the discriminant methods used in toxicological pathology. Toxicol. Pathol. 33, 490–4 (2005).
58. Scaioli, E. et al. e Pharmacoinetic Prole of a New Gastroresistant Capsule Preparation of Eicosapentaenoic Acid as the Free
Fatty Acid. Biomed es. Int. 2015, 1–8 (2015).
59. Snedecor, G. & Cochran, W. Statistical Methods. (Iowa State University Press, 1989).
Acknowledgements
e authors thank Dr. Matteo Soverini for the bioinformatic analysis performed on microbiota data. is work
was supported by Italian Association for Cancer Research [Grant number: Investigator Grant IG14281 to L.R.,
Fellowship “David Raaelli” Number: 13837 to A.P.].
Author Contributions
A.P. performing experiments, draing of the manuscript, acquisition of data, analysis and interpretation of the
data, statistical analysis. E.S. performing endoscopies, collecting patients’ data, acquisition of data, analysis and
interpretation of the data. G.P. and C.F. technical support, acquisition of data, analysis and interpretation of
the data. M.B. technical support, acquisition of data. E.B. technical support, acquisition of data, analysis and
interpretation of data. M.C. acquisition of data, analysis and interpretation of data. P.B. analysis and interpretation
of data. C.C. technical support. T.B. analysis and interpretation of data. P.C. analysis and interpretation of data.
A.M. technical support, acquisition of data. M.P. technical support. M.M. technical support. F.B. critical revision
of the manuscript for important intellectual content. A.B. study concept and design; analysis and interpretation
of data; critical revision of the manuscript for important intellectual content. L.R. performing endoscopies, study
concept and design; study supervision; analysis and interpretation of data; obtained funding.
Additional Information
Supplementary information accompanies this paper at doi:10.1038/s41598-017-07992-1
Competing Interests: Luigi Ricciardiello received an unrestricted research grant by SLA Pharma, UK. Others
authors have nothing to disclose.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2017