© 2019. Published by The Company of Biologists Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License

(http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution and reproduction

in any medium provided that the original work is properly attributed.

Zebrafish modeling of intestinal injury, bacterial exposures, and medications

defines epithelial in vivo responses relevant to human inflammatory bowel

disease

Ling-shiang Chuang1,2, Joshua Morrison1,5, Nai-yun Hsu1,2, Philippe Ronel Labrias1,

Shikha Nayar1, Ernie Chen1,2, Nicole Villaverde1, Jody Ann Facey1, Gilles Boschetti3,

Mamta Giri1, Mireia Castillo-Martin4, Tin Htwe Thin4, Yashoda Sharma1, Jaime Chu5,

Judy H. Cho1,2*

1. Department of Genetics and Genomic Sciences, Icahn School of Medicine at

Mount Sinai, New York, United States

2. The Charles Bronfman Institute for Personalized Medicine, Icahn School of

Medicine at Mount Sinai, New York, United States

3. Department of Oncological Science, Icahn School of Medicine at Mount Sinai,

New York, United States

4. Department of Pathology, Icahn School of Medicine at Mount Sinai, New

York, United States

5. Department of Pediatrics, Icahn School of Medicine at Mount Sinai, New York,

United States

*To whom correspondence should be addressed.

The corresponding author’s contact information

Judy Cho

Hess CSM Building Floor 8th Room 118

1470 Madison Avenue

New York, NY 10029

TEL. (212) 824-8940

FAX. (646) 537-9452

E mail: Judy.cho@mssm.edu

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

http://dmm.biologists.org/lookup/doi/10.1242/dmm.037432Access the most recent version at First posted online on 23 July 2019 as 10.1242/dmm.037432

Abstract

Genome-wide association studies have identified over 200 genomic loci associated to

inflammatory bowel disease (IBD)1. High effect risk alleles define key roles for genes

involved in bacterial response and innate defense2. More high-throughput in vivo

systems are required to rapidly evaluate therapeutic agents. We visualize, in zebrafish,

the effects on epithelial barrier function and intestinal autophagy of one-course and

repetitive injury. Repetitive injury induces increased mortality, impaired recovery of

intestinal barrier function, failure to contain bacteria within the intestine, and impaired

autophagy. PGE2 administration protected against injury by enhancing epithelial

barrier function and limiting systemic infection. Effects of IBD therapeutic agents were

defined; mesalamine showed protective features during injury while 6-mercaptopurine

displayed marked induction of autophagy during recovery. Given the highly conserved

nature of innate defense in zebrafish, it represents an ideal model system with which

to test established and new IBD therapies targeted to the epithelial barrier.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Introduction

IBD is comprised of two subtypes, Crohn’s Disease (CD) and ulcerative colitis (UC).

While UC affects the superficial layers of the intestine, CD often affects deeper layers

of the bowel wall. The most significant loci identified in genome-wide association

studies (GWAS) for IBD include protein-altering alleles in IL23R (interleukin 23

receptor), NOD2 (nucleotide oligomerization domain 2), ATG16L1 (autophagy related

16 like 1)3, LRRK2 (Leucine-rich repeat kinase 2)4 and CSF2RB (colony stimulating

factor 2 receptor)5, as well as non-coding risk alleles near PTGER4 (prostaglandin E

receptor 4), which increase gene expression6. The NOD2 risk alleles show impaired

intracellular sensing of bacterial peptidoglycan7, resulting in impaired NF-B

activation8. The ATG16L1 alanine risk allele at codon 300 increases its proteolytic

cleavage by caspase 3 and 79, 10, implicating impaired autophagy in CD.

The striking overlap between genes implicated in CD and mycobacterial susceptibility

highlights a key role for innate defense and containment of microbes2. Recent literature

has defined the particular value of zebrafish in modeling innate defense and

immunity11-13. Because of their transparency over the initial weeks of life, zebrafish

serve as a powerful tool for performing live in vivo imaging of intestinal macrophages

and epithelial responses to injury and bacterial exposure.

Results

Repeated DSS treatments prevent recovery of acidified-lysosome and mucin

loss

To model the chronicity of IBD, we first compared the effects of single- and three-

course dextran sodium sulfate (DSS) injury (Fig. 1A, Fig. S1A) in zebrafish. Similar to

DSS injury in mice, the gut length is significantly shortened in DSS-injured zebrafish,

with no difference in the total length observed (Fig. S1B-1C). Using doses as

previously reported14, 15, we observed a dose-dependent induction of mortality with

repeated injury (Fig. 1A, 68% for 0.25% DSS, 29% for 0.1%). With both single and

repeated DSS injury, the goblet cells were observed by H&E histology (Fig. 1B).

We next sought to examine altered epithelial phenotypes contributing to the high

mortality with repetitive DSS injury. Neutral red is a dye that can diffuse across cellular

membranes at physiological pH (pH = 7), but is protonated under acidic conditions, at

which point the dye appears red and is unable to leave the acidic compartments it has

entered. Due to the acidic nature of lysosomes, neutral red will accumulate in properly

acidified-lysosomes (Fig. S2A-2C)14, 16, with a marked enrichment in lysosome-rich

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

enterocytes (LREs) in the posterior mid-intestine (Fig. S2D-2E)17-21. LREs are a

population of highly endocytic enterocytes in the zebrafish posterior midgut17. The

intensity of Neutral red staining indicates the amount of healthy lysosomes with acidic

pH16. Therefore, a decrease in Neutral red intensity indicates a depletion of healthy,

properly acidified-lysosomes. We observed a marked decrease in neutral red intensity

with single course DSS, which was significantly restored one day after DSS removal

and largely recovered by two days after DSS removal (Fig. 1C, Fig. S2F). In contrast,

with repetitive DSS injury, the reduction in neutral red staining observed after the third

course of DSS did not recover at one or two days following DSS removal (Fig. 1D).

Moreover, by qPCR, we show proinflammatory cytokines tnfa and Il1b were

predominately induced in the intestine with single DSS injury, but are also expressed

in non-intestinal tissues after repeated injury (Fig. S3A-3B). Similarly, Alcian blue

staining, which measures epithelial mucus expression (Fig. S4A-4B)14, demonstrated

a largely intact mucus with no changes in goblet cell number with single course DSS

injury, but did not recover with repeated injury when compared to controls (Fig. 1E,

Fig. S5A-5D).

DSS injury with bacterial exposure impacts LRE autophagy and bacterial protein

uptake

The uptake of bacterial proteins by gut-associated specialized enterocytes and

possibly mononuclear phagocytes serves to contain bacterial contaminants to the

intestine, a critical protective function. One of the main function for LREs is endocytosis

of proteins from the lumen of zebrafish intestine17. Using fluorescent and confocal

microscopy, we sought to visualize the local containment of bacteria in the absence

and presence of DSS. We used pHrodo-labeled, heat-killed E. coli, which fluoresces

at pH levels below 5.5. The amine of E. coli proteins was conjugated by activated

succinimidyl ester on pHrodo dye. After one hour of E. coli incubation, we observed

markedly increased uptake of E. coli proteins with DSS treatment (Fig. 2A). Cyto-ID

fluorescently stains autophagosomes in live cells, but does not stain lysosomes22. To

confirm the specificity of Cyto-ID staining for autophagy in the zebrafish intestine,

Bafilomycin A1 treatment was combined with both Cyto-ID and Neutral red staining.

Bafilomycin A1 is a late-stage autophagy inhibitor, which blocks both starvation-

induced and starvation-independent autophagy23. In the larval zebrafish, inhibiting

autophagy leads to loss of Cyto-ID staining but not neutral red staining LREs, indicating

depleted autophagy levels but normal levels of functional lysosomes (Fig. S6A-6B). Colocalization of labeled E. coli proteins and autophagy as measured by Cyto-ID was

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

observed in LREs (Fig. 2B, Movie S1-4). The fluorescence from E. coli proteins was

most intense in the posterior midgut (Fig. 2A), with pHrodo signals localizing to the

region enriched for LREs as shown by lysosensor-labeled lysosomes (Fig. S7A).

Similar to recovery trends seen with neutral red staining, we observed markedly

decreased Cyto-ID staining with either single or repeated DSS injury (Fig. 2C). The

single DSS injury finding was further validated with p62 immunoblot (Fig. S8)24, 25.

Outside of posterior midgut, Cyto-ID staining of autophagy is present and not limited

to the region with ingested E. coli proteins (Fig. S7B). With and without 0.25% DSS

treatment, we showed the decrease of Cyto-ID intensity in LRE region of posterior

midgut (Fig. S9). One course DSS injury demonstrated a marked recovery of

autophagy one day after removal of DSS (Fig. 2D); however, similar to trends observed

with neutral red staining, three course DSS injury did not demonstrate recovery of

Cyto-ID staining.

Endocytosis of proteins is an essential function of LREs in the zebrafish intestine17.

Pllp, a cellular marker for LREs, is also functionally necessary for proper apical

endocytosis17. Taken together, it follows that LREs may have the ability to endocytose

the proteins of immotile pathogens. Treatment with high-dose, but not low-dose, DSS

followed by E. coli exposure results in marked mortality (High-dose: Fig. 2E, left panel;

low-dose: Fig. S10). Using the zebrafish model, the role of bacterial invasion, an initial

step of sepsis and systemic infection, is dramatically demonstrated in vivo (Fig. 2E

right panel, Movie S6-7) and quantified (Fig. S11). The intensities of fluorescently-

labeled E. coli proteins are high in the intestine (S6-7 movie), which suggests that the

initiation of bacterial invasion originates from the intestine. However, further analysis

will be needed to exclude the possibility of DSS injury allowing E. coli invasion via

extra-intestinal surfaces of the zebrafish.

PGE2 and common IBD medications impact barrier function, bacterial

containment, and LRE function

We next sought to examine mechanisms modulating the mucus layer, with a particular

focus on the effects of treatment with prostaglandin E2 (PGE2). The CD predominant

association to the chromosome 5p13 region containing PTGER4 is the third most

significant association in Europeans, with the PTGER4 risk alleles correlating with

increased expression6, 26. Contrary to this, however, PGE2 is known to play a critical

role in intestinal mucosal cytoprotection mediated by PTGER427. Blockade of PGE2

through the use of COX2 (cyclo-oxygenase 2) inhibitors (e.g. non-steroidal anti-

inflammatory agents) impairs mucosal cytoprotection, frequently resulting in intestinal

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

ulceration28. Using our in vivo model, we observed a dose-dependent increase in

mucus expression by Alcian blue staining with PGE2 treatment for 24 hours (Fig. 3A-

3B). This increase was even more marked when confining the Alcian blue

quantification to the secreted, luminal mucus staining in the distal hindgut region (Fig.

3C14). With DSS injury, we observed the same level of increase of Alcian blue staining

with PGE2 treatment (Fig. S12A). Similar results were obtained with enteroids from

human intestinal epithelium, with a marked induction of Alcian blue staining (Fig. 3D;

images: Fig. S12B).

We next sought to define the time course dependencies of PGE2 treatment relative to

DSS injury and bacterial exposure. DSS injury followed by E. coli exposure results in

high levels of mortality (Fig. 2E-2F, Fig. 3F, green line). With one course of DSS injury,

followed by PGE2, and then E. coli treatment (Fig. 3E), we observed a significant

decrease in mortality (Fig. 3F, purple compared to green line); concomitant treatment

of PGE2 during the DSS injury period, followed by a 5 hour recovery period, then E.

coli treatment, resulted in an even further decrease in mortality (Fig. 3F, blue line).

With DSS administration (Fig. 3E-3G, green, purple and blue lines), the relative

mortality (Fig. 3F) parallels the fraction of ingested E. coli proteins at 1 and 2 hours

post-DSS treatment as measured by pHrodo intensities (Fig. 3G); after 3 hours of E.

coli exposure, among the surviving individuals, all three of the DSS-treated conditions

level off to equivalent amounts of ingested pHrodo-labeled E. coli proteins. In the

absence of DSS treatment (Fig. 3F-3G, black and red lines), E. coli treatment alone or

with antecedent PGE2 treatment before E. coli exposure (Fig. 3E, black and red bars),

no mortality is detected (Fig. 3F); successively increasing amounts of ingested pHrodo-

labeled E. coli proteins 1, 2 and 3 hours was observed after E. coli exposure (Fig. 3G,

black and red lines), in amounts that far surpass the DSS-treatment groups (Fig. 3G,

green, purple and blue lines). This indicates that in the absence of DSS-treatment, with

less abrupt bacterial exposure, robust uptake and containment of bacterial proteins by

LREs limits mortality.

Finally, we sought to model the effects of PGE2 as well as drugs commonly utilized in

IBD treatment on intestinal LREs phenotypes and function. Administration of either

PGE2 or 6-mercaptopurine during DSS injury did not result in enhancement of neutral

red or Cyto-ID staining compared to DSS treatment alone (Fig. 3H, light green points).

Interestingly, co-treatment of mesalamine during DSS injury resulted in increased

neutral red staining at both the 13 and 65 M doses, with an increase in Cyto-ID

observed at the 65 M dose (Fig. 3H, light green points). Neutral red and Cyto-ID

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

intensities were higher 5 hours after DSS removal (Fig. 3H, light blue points) compared

to measurements taken immediately after DSS removal (Fig. 3H, light green points).

Treatment with mesalamine administered for 5 hours immediately after removal of DSS

(Fig. 3H, light blue points, recovery phase), but did not result in enhancement of neutral

red or Cyto-ID intensities compared to untreated recovery controls. In contrast, 5 hours

of PGE2 treatment in the recovery phase resulted in significant induction of neutral red

and Cyto-ID intensities (Fig. 3H, light blue points), presumably reflecting the mucus

cytoprotective effects of PGE2 (Fig. 3A-3D). Importantly, 6-mercaptopurine treatment

resulted in increased neutral red and Cyto-ID intensities at both the 66 and 330 M

doses (Fig. 3H, light blue points), demonstrating a surprisingly rapid, in vivo induction

of autophagy in the recovery period following DSS injury.

Discussion

Traditionally, the most commonly utilized models in IBD involve murine models,

including IL10 pathway knockouts, the T cell transfer model, DSS treatment and, more

recently, the abcb1a (mdr1a)-/- knockout, which accurately predicted human

responses to blockade of IL12/23 and IL-17 as being beneficial and harmful,

respectively29-32. However, long-term remission rates as measured by mucosal healing

at week four, indicating that development of agents that address the epithelial

alterations in IBD are needed33. In particular, given the known role that decreased

autophagy in epithelial cells plays in driving Crohn’s disease34, 35, development of in

vivo models that can rapidly scale testing of new agents is needed.

Zebrafish models of IBD provide substantive advantages in modeling repetitive

intestinal injury and repair. In this regard, the most broadly utilized models of IBD have

utilized DSS which results in epithelial injury and inflammation, followed by prompt

recovery following DSS removal. It has been reported that DSS injury results from

complex formation with intraluminal fatty acids during DSS administration, resulting in

impaired barrier function. Chronic models of DSS in mice have been reported, but take

months to develop36. In the present models, we observe a striking failure to fully

recover from repetitive DSS injury as manifested by increased mortality, impaired

mucus production and impaired autophagy, all within two weeks; while the

inflammatory response is predominant in the intestine, given the mode of

administration, extraintestinal effects may occur. We observed the incomplete

recovery with repetitive injury by both measurements of neutral red, which measures

lysosomal pH, as well as by CytoID, which measures autophagosomes.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

The high mortality observed with repetitive DSS injury was associated with a striking,

systemic presence of microbial products (Fig. 2F). Both in vivo and in vitro (Fig. 3A-

3G), we observed that PGE2 potently induces mucus expression, which has previously

been reported to be mediated by PTGER4-mediated mucin exocytosis in a human

carcinoma cell line37. Furthermore, mice deficient in PTGER4 in the intestinal epithelial

layer did not develop wound-associated epithelial cells, resulting in impaired tissue

repair28. Our data highlight the critical role that PGE2 plays in protecting against

intestinal barrier associated bacterial invasion through increased mucin production,

blocking uptake of luminal E. coli proteins (Fig. 4). These findings are consistent with

a critical role of PGE2-PTGER4 in mediating intestinal barrier function, and

administration of PGE2 after removal of DSS resulted in enhanced recovery of Cyto-

ID staining (Fig. 3H). The complex actions of PGE2-PTGER4 in intestinal barrier

function, epithelial repair and innate immune functions have yet to be fully elucidated.

The presence of CD predominant PTGER4 associations that increase PTGER4 gene

expression6 would appear to contradict the overall protective actions of PGE2-

PTGER4 in intestinal barrier and epithelial barrier function. Future studies evaluating

pathogenic effects of increased PTGER4 expression, especially in CD, are of the

highest priority.

During DSS injury, only mesalamine resulted in enhanced Cyto-ID staining; this finding

is consistent with prior literature that mesalamine acts by blocking pro-inflammatory

cytokine-mediated induction of NF-kB thereby enhancing epithelial barrier function38,

39. In our study, mesalamine exerts its protective effects only during injury and not

during the recovery period following DSS removal. In contrast, 6-mercaptopurine

enhances autophagy solely after DSS removal, consistent with prior literature of direct

effects of thiopurines in enhancing autophagy40. While blockade of pro-inflammatory

cytokines such as anti-TNF or anti-IL12/2341 represents major mechanisms for treating

moderate to severe CD, it is still an open question as to whether management

combining pro-inflammatory cytokine blockade42 with 6-mercaptopurine provides

additional benefit43. As with many highly utilized, older medications, the precise

mechanism of action of 6-mercaptopurine is incompletely understood and likely

multifactorial in nature44, 45. Its use has been supplanted by more potent pro-

inflammatory blockade in moderate to severe disease. While the CD-predominant

associations to autophagy represent some of highest effect and previously

unappreciated pathways implicated by genome-wide association studies, given the

overall lower efficacy of 6-mercaptopurine in moderate to severe CD43, our studies

provide a cautionary note regarding the potency CD-based therapeutic development

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

focused solely on enhancing autophagy in the intestine. In the era of Precision

Medicine, scaling more rapidly a sophisticated understanding of time course factors,

disparate cellular effects, and mechanisms of action of established and new drugs will

be required. While mammalian systems will remain a linchpin, the superior speed,

sample size, visualization (e.g. microbes, autophagy) and key role of the epithelial

barrier makes high throughput studies in the zebrafish an important component of the

IBD therapeutic development pipeline.

Materials and Methods

Zebrafish maintenance

Adult zebrafish were maintained on a 14:10 hour light:dark cycle at 28°C. Wild-type

(WT; AB and Tab 14) and Tg(mfap4:turquoisext27) 46, 47. fish were used. Fertilized

embryos collected following natural spawning were cultured at 28.5°C in egg water

(0.6 g/l Instant Ocean; Blacksburg, VA) containing methylene blue (0.002 g/l) and 75

μM 1-Phenyl 2-thiourea. The Mount Sinai School of Medicine Institutional Animal

Care and Use Committee approved all protocols. For fluorescent imaging, the

methylene blue was removing within 24 hours of embryos collection. Larvae and

adults were fed once a day with Hikari First Bites (Hikari) after 6-day -post fertilization

and Zeigler Zebrafish Diet with Hatching Brine Shrimp Eggs (Pentair Aquatic Eco-

Systems, FL), respectively.

Enteroid culture establishment and differentiation

The human biopsies collection was approved by the Institutional Review Board of the

Mount Sinai School of Medicine. Human enteroid lines were generated according to

the method from VanDussen et. al.48 with modifications. Briefly, biopsies were obtained

from healthy and inflammatory bowel disease individuals during routine colonoscopy

visit. Biopsies were dissociated by collagenase I (Thermo Fisher Scientific) incubation

at 37C for 30 minutes and vigorously pipetting mixtures every 5 minutes. Biopsies

mixtures were passed through 70 m cell strainer (BD Biosciences) followed by two

washes with wash medium (DMEM/F12 medium with 15mM HEPES, 2mM L-

glutamine, 100U/ml of Penicillin and 0.1mg/ml of streptomycin). The crypt mixtures

were suspended in Matrigel (Corning) and cultured in 50% L-WRN (L cell-Wnt3a

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Rspondin3 and Noggin) conditioned medium with 10M of Y-27632 (ROCK inhibitor;

R&D Systems) and 10M SB-431542 (TGF-R inhibitor; R&D Systems). L-WRN

conditioned media was prepared as described49. Cells were passaged every 4 to 6

days by trypsinization, and fresh 50% L-WRN was replaced every 2 days. After 5

passages, cells were frozen in 8.3% DMSO in wash media and kept in liquid nitrogen.

Prior to differentiation, cells were thawed and expanded in 50% L-WRN with 10M of

Y-27632 and 10M SB-43154. To induce differentiation, 2.5 X 105 of trypsinized

spheroids were seeded in 6.5 mm PET transwell inserts (Corning) precoated with 1%

gelatin and cultured in 5% L-WRN with 10M of Y-27632 for 3 days. For mucus

quantification studies, 1 M of PGE2 or same volume of ethanol control was

supplement in 5% L-WRN during the last 24 hours of the differentiation.

Single and repeated DSS injury model

The single injury protocol was adapted and modified from Oehlers et. al.15. Batches

of 60 larvae were segregated into Petri dishes in 50 mL of egg water. Phenylthiourea

(75 μM; Acros Organics) was added 24 hours post-fertilization to prevent pigment

cell formation. To induce intestinal injury, 3 days post fertilization (dpf) larvae were

placed in freshly prepared 0.1 or 0.25% (w/v) colitis grade DSS (36,000 - 50,000

MW, MP Biomedical) for three days. In the repeated injury protocol, larvae were

treated with 0.05, 0.1 or 0.25% DSS at 5, 8, and 11 dpf for 24 hours, followed by

removal of DSS. Larvae were not fed for the duration of the single DSS treatment

experiments (6 dpf), the larvae were fed after 7 dpf with repeated DSS. Feeding was

skipped on the DSS treatment day.

Automated Hematoxylin and Eosin (H&E) Staining

An Autostainer XL (ST5010, Leica Biosystems, Nussloch, Germany) was used for

H&E staining. The automated staining protocol consisted of the following steps: three

changes of xylenes (2, 2 and 1 minutes), rehydration in an ethanol/water gradient

(100%, 100%, 95% ethanol for 2, 1 minutes and 20 seconds, respectively), followed

by washing with deionized (DI) water (2 minutes). The rehydrated slides were stained

with Gill 3 Hematoxylin (12 minutes; Sigma-Aldrich), washed with DI water (2

minutes), followed by 0.17% Acid Alcohol (1 minutes), washed with DI water (1

minutes), Ammonia water (20 seconds), washed with DI water (1 min), and 95%

ethanol (10 seconds). Staining with Eosin (45 seconds; Sigma-Aldrich) was

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

completed, followed by four washes with 100% ethanol (1 minute each), and four

washes with xylene (1 minute each). H&E stained slides were cover slipped with

Cytoseal XYL (Thermo Fisher Scientific) and dried prior to imaging.

Neutral red, Alcian blue staining and image quantification

The neutral red and Alcian blue staining protocols were adapted and optimized from

Oehlers et. al. 201315. Zebrafish larvae were staining live with 2.5 μg/ml neutral red

(ACROS, Bridgewater, NJ) in egg water for 5 hours. The intensity of neutral red

staining is non-saturated and consistent between 4 to 6 hours (Fig. S2A). After

moving stained larvae to fresh 50 ml egg water, larvae were anesthetized in tricaine-

S (Western Chemical), followed immediately by live imaging. For the zebrafish

intestinal injury model, mucin was stained as the quantifiable metric. Larvae were

fixed in 4% paraformaldehyde at room temperature for two hours, rinsed twice with

acid alcohol (70% ethanol with 1% hydrochloric acid) and stained in 0.01% Alcian

blue (w/v, in 80% ethanol with 20% glacial acetic acid) for three hours at room

temperature. The Alcian blue intensities were linear and non-saturated at two to four

hours of staining (Fig. S4A-4B). After rinsing with acid alcohol three times, fixed

larvae were imaged on a thin layer of 5% agarose gel.

For Alcian blue staining procedure of differentiated human enteroids, cells were fixed

in cold Methanol- Carnoy solution (60% Methanol, 30% Dry Chloroform, 10% Glacial

Acetic Acid; Sigma ) overnight and washed with 1X Phosphate-buffered saline (PBS,

10 mM PO43−, 137 mM NaCl, and 2.7 mM KCl) three times. Then, cells were kept in

1X PBS with 0.02% acetic acid solution for an additional 16 hours at 4°C. Prior to

Alcian blue staining, cells were kept in 3% acetic acid for 3 minutes at room

temperature. Cells were stained with 1% Alcian blue (in 3% acetic acid solution;

ACROS Organics) for 30 minutes at room temperature follow by three washes in of 1X

PBS solution. Transwell membranes were cut and mounted in fluoromount G (Electron

Microscopy Sciences) and visualized.

For both neutral red and Alcian blue staining, the images of full gut and/or lumen

were collected under 100x total magnification with the EVOS XL core microscope

(Thermo Fisher Scientific). For Alcian blue images, we extract intensity from blue

channel before we performed quantification. For both neutral red and Alcian blue

staining, we inverted the color of the image, traced the gut and/or lumen, or

monolayers cells from differentiated enteroids, and measured the mean intensity per

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

pixel using ImageJ software. To remove background signals, we subtracted the

mean intensity per pixel of images from the anal end of the gut. The relative

expressions (%) were calculated by normalizing to the untreated group. Statistical

tests were done with two-tailed Wilcoxon-test in RStudio.

Bacterial infection and autophagy

Heat-killed Escherichia coli (K-12 strain) were conjugated either with BioParticle-

Alexa 488 conjugates or pHrodo Red E. coli BioParticles Conjugate for uptake

studies (ThermoFisher, Waltham, MA). Bacterial concentrations were titrated and

optimized based on intestinal fluorescent intensity. Zebrafish larvae were incubated

in 2 μg/μL (6x107 CFU/ml) for one hour. Larvae were rinsed and moved to egg water,

with mortality and fluorescence measured at 0, 1, 2, and 3 hours.

We used Cyto-ID Autophagy detection kit 2.0 (Enzo Biochem) to analyze autophagy

with DSS and/or bacterial infection in live zebrafish. Zebrafish were incubated in 1 μL

CYTO-ID per 500 μL egg water for one hour with or without E.coli or DSS and wash

with 50 mL egg water before imaging. The fluorescent images for quantification were

collected after E. coli treatment with the EVOS FL Cell Imaging System. Using

ImageJ software, we measured the mean intensity per pixel for the gut and

subtracted this background for individual images. Auto-fluorescence was corrected

by comparing to the unstained fish gut. The relative expressions (%) were calculated

by normalizing to the untreated group. Statistical tests were done with Wilcoxon-test

in RStudio.

The confocal images of co-localization analysis and movies were generated with the

ZEISS LSM 880 with Airyscan. For co-localization, Tg(mfap4: turquoise) fish were

treated with pHrodo Red E. coli and stained for autophagy with CytoID. The confocal

z-stack movies (Movie S1-S4) and cell migration movies (Movie S5) were collected

after five hours of E. coli treatment. Systemic infection movies (Movie S6-S7) were

imaged with BioParticle-Alexa 488 conjugate at 90 minutes after the end of E. coli

treatment.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

PGE2 and drugs commonly utilized in IBD treatment

PGE2 and IBD drugs were added directly to zebrafish media either with or without DSS

at 5 dpf for cotreatment as a therapy. In the recovery groups, drugs were added after

removing DSS, and imaging was performed 5 hours later. The doses of drugs were

tested and the highest doses without fish death were chosen. Final concentrations of

PGE2 were 0.1 and 1 µM with equal amounts of ethanol used for controls. Mesalamine

(5-aminosalicyclic acid) was dissolved in egg water at a concentration of 10 mg/mL

and adjusted to pH 7 with dilute sodium hydroxide. 6-Mercaptopurine was at a

concentration of 35 mg/mL dissolved and diluted to the final concentrations with egg

water, with pH adjusted to 7.

Gene expression analysis of zebrafish intestines

From anesthetized, 6 days post-fertilization zebrafish larvae, whole intestines were

microdissected and separated from the remaining carcasses, with each collected in

30µL RLT Buffer (Qiagen). Total RNA was isolated from N = 15-25 intestines or

carcasses per sample by TRIzol extraction (Life Technologies). cDNA libraries were

generated by reverse transcription using the SuperScript cDNA synthesis kit

(Quanta). Quantatative Real Time PCR (qRT-PCR) was performed using PerfeCTa

SYBR Green Fast Mix (Quanta) with the LightCycler 480 (Roche). Gene expression

levels were normalized to ribosomal protein large P0 (rppo) using the comparative

threshold cycle (ΔΔCt) method. Primer sequences are listed in Table S1.

Single-cell suspension of zebrafish intestines for flow cytometry

6 dpf zebrafish were stained with LysoTracker Green DND-26 (Invitrogen) at a 1:100

dilution for one hour. Whole intestines were then dissected from anesthetized

zebrafish (N = 25 per group) and collected in 1mL MACS Rinsing Solution (Miltenyi

Biotech) with 0.5% BSA (Fisher Scientific), following by the single-cell suspension

protocol50. In short, samples were rocked at RT for 30 minutes before being

centrifuged at 700xg for 5 minutes. The cell pellet was suspended in PBS

with1mg/mL DNase I (Sigma Aldrich) and 1 mg/mL collagenase IV (Sigma Aldrich),

following by harsh pipetting and incubation at RT for 5 minutes. Samples were

centrifuged at 700xg for 5 minutes and the supernatant was removed. The cell pellet

was resuspended in DPBS 1X with 0.5% BSA (Gibco) before being strained through

70µM mesh. Single-cell suspensions were flow sorted in the Mount Sinai Flow

Cytometry CoRE using the BD influx cell sorter (BD Biosciences). RNA was

extracted using the RNAqueous kit (Invitrogen). cDNA synthesis and qPCR were

conducted as described in the gene expression section.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Immunoblot analyses

6 dpf whole zebrafish larvae were homogenized in lysis buffer (20 mM Tris pH 7.5,

150 mM NaCl, 1% NP-40, 2 mM EDTA, 10% glycerol, 1% BME, and protease

inhibitors (Roche Complete)). From each sample, 10 µg of protein was separated on

a Mini-Protean TGX 4-15% (Bio-Rad) gel and transferred to a PVDF membrane

using a Bio-Rad electrophoretic transfer apparatus (Trans-Blot Turbo, Bio-Rad).

After blocking with 5% BSA, the membrane was probed with the primary antibodies

anti-p62 (PM045, poly clonal, MBL; 1:1000) and anti-tubulin (12G10, DSHB; 1:2000).

A secondary incubation of either 1: 40000 dilution of horseradish peroxidase (HRP)

conjugated goat anti-mouse or 1:20000 dilution goat anti-rabbit secondary antibody

was applied to the membrane. The immunoblots were developed with SuperSignal

West Pico (Life Technologies)

Colocalization analysis of lysosensor and neutral red

The confocal images (Fig. S2B top panel) of zebrafish intestine stained with

lysosensor and neutral red were taken with a Zeiss LSM 880 confocal microscope

using 20X air lens (NA=0.8). The zoomed in images (Fig. S2B bottom panel) were

generated from the non-saturated area in intestine of the original images using Zen

2.3 Lite (Zeiss). Colocalization analysis of both lysosensor and neutral red positive

cells was performed using ImageJ Fiji software. The 2D intensity histogram and

Pearson correlation coefficient were generated by the Coloc2 plugin of Fiji.

Quantitation of E.coli intensity in cells of zebrafish dorsal aorta and cardinal

vein

Zebrafish were treated in the presence or absence of 0.25% DSS treatment for 24

hours before subject to fluoresce Alexa-488 labeled heat killed E.coli incubation (0 to

90 minutes ). The E.coli fluorescence intensity of 170 of zebrafish cells were

quantified using Zen 2.3 Lite software from Zeiss (38 cells for Control-Aorta: control

treatment in dorsal aorta; 53 cells for DSS-aorta: DSS treatment in dorsal aorta; 38

cells for Control-vein: control treatment in posterior cardinal vein; 41 cells for DSS-

vein: DSS treatment in posterior cardinal vein). P values were tested by the Mann

Whitney test.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Chemical inhibitor for autophagy

Bafilomycin A1 (Sigma-Aldrich) was dissolved in DMSO to create a 100 μM stock

solution and then diluted to a concentration of 10 and 20 nM in egg water. 5 dpf

Zebrafish larvae were treated for 24 hours, followed by Cyto-ID and Neutral red

staining.

Acknowledgements

Microscopy, intestinal sections and hematoxylin & eosin staining were performed at

the Microscopy, Biorepository and Pathology CoRE at the Icahn School of Medicine

at Mount Sinai with assistance from Nikos Tzavaras, Lisette Conde and Alan Soto,

respectively. Zebrafish were maintained and cultivated at Zebrafish Shared Research

Facility (Z-SRF) at the Icahn School of Medicine at Mount Sinai with assistance from

Charles Derossi and Kathryn Bambino. Mfap4 transgenic zebrafish line Tg(mfap4:

turquoise) was a generous gift from David Tobin at Duke University.

The authors disclose no conflicts of interest. This study was supported by a research

grants from Dr. Sanford J Grossman Trust for Integrative Studies in IBD, DK062429

and DK062422, K08 DK101340, Crohn’s and Colitis Foundation Visiting IBD

Research Fellowship, Gilead Liver Research Scholars Award, and The Mindich Child

Health and Development Institute at the Icahn School of Medicine at Mount Sinai.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

References

1. Liu, J.Z. et al. Association analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations. Nature genetics (2015).

2. Jostins, L. et al. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease. Nature 491, 119-124 (2012).

3. Huang, H. et al. Fine-mapping inflammatory bowel disease loci to single-variant resolution. Nature 547, 173-178 (2017).

4. Hui, K.Y. et al. Functional variants in the LRRK2 gene confer shared effects on risk for Crohn's disease and Parkinson's disease. Science translational medicine 10 (2018).

5. Chuang, L.S. et al. A Frameshift in CSF2RB Predominant Among Ashkenazi Jews Increases Risk for Crohn's Disease and Reduces Monocyte Signaling via GM-CSF. Gastroenterology 151, 710-723 e712 (2016).

6. Libioulle, C. et al. A novel susceptibility locus for Crohn's disease identified by whole genome association maps to a gene desert on chromosome 5p13.1 and modulates the level of expression of the prostaglandin receptor EP4. PLoS genetics (2007).

7. Bonen, D.K. et al. Crohn's disease-associated NOD2 variants share a signaling defect in response to lipopolysaccharide and peptidoglycan. Gastroenterology 124, 140-146 (2003).

8. Ogura, Y. et al. A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature 411, 603-606. (2001).

9. Lassen, K.G. et al. Atg16L1 T300A variant decreases selective autophagy resulting in altered cytokine signaling and decreased antibacterial defense. Proceedings of the National Academy of Sciences of the United States of America 111, 7741-7746 (2014).

10. Murthy, A. et al. A Crohn's disease variant in Atg16l1 enhances its degradation by caspase 3. Nature 506, 456-462 (2014).

11. Madigan, C.A. et al. A Macrophage Response to Mycobacterium leprae Phenolic Glycolipid Initiates Nerve Damage in Leprosy. Cell 170, 973-985 e910 (2017).

12. Cambier, C.J., O'Leary, S.M., O'Sullivan, M.P., Keane, J. & Ramakrishnan, L. Phenolic Glycolipid Facilitates Mycobacterial Escape from Microbicidal Tissue-Resident Macrophages. Immunity 47, 552-565 e554 (2017).

13. Cronan, M.R. et al. Macrophage Epithelial Reprogramming Underlies Mycobacterial Granuloma Formation and Promotes Infection. Immunity 45, 861-876 (2016).

14. Oehlers, S.H., Flores, M.V., Hall, C.J., Crosier, K.E. & Crosier, P.S. Retinoic acid suppresses intestinal mucus production and exacerbates experimental enterocolitis. Disease models & mechanisms 5, 457-467 (2012).

15. Oehlers, S.H. et al. Chemically induced intestinal damage models in zebrafish larvae. Zebrafish 10, 184-193 (2013).

16. Chazotte, B. Labeling lysosomes in live cells with neutral red. Cold Spring Harb Protoc 2011, pdb prot5570 (2011).

17. Rodriguez-Fraticelli, A.E. et al. Developmental regulation of apical endocytosis controls epithelial patterning in vertebrate tubular organs. Nature cell biology 17, 241-250 (2015).

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

18. Oehlers, S.H. et al. A chemical enterocolitis model in zebrafish larvae that is dependent on microbiota and responsive to pharmacological agents. Developmental dynamics : an official publication of the American Association of Anatomists 240, 288-298 (2011).

19. Farber, S.A. et al. Genetic analysis of digestive physiology using fluorescent phospholipid reporters. Science 292, 1385-1388 (2001).

20. Wallace, K.N., Akhter, S., Smith, E.M., Lorent, K. & Pack, M. Intestinal growth and differentiation in zebrafish. Mechanisms of development 122, 157-173 (2005).

21. Rombout, J.H., Lamers, C.H., Helfrich, M.H., Dekker, A. & Taverne-Thiele, J.J. Uptake and transport of intact macromolecules in the intestinal epithelium of carp (Cyprinus carpio L.) and the possible immunological implications. Cell and tissue research 239, 519-530 (1985).

22. Marx, V. Autophagy: eat thyself, sustain thyself. Nature methods 12, 1121-1125 (2015).

23. Hundeshagen, P., Hamacher-Brady, A., Eils, R. & Brady, N.R. Concurrent detection of autolysosome formation and lysosomal degradation by flow cytometry in a high-content screen for inducers of autophagy. BMC Biol 9, 38 (2011).

24. Schiebler, M. et al. Functional drug screening reveals anticonvulsants as enhancers of mTOR-independent autophagic killing of Mycobacterium tuberculosis through inositol depletion. EMBO molecular medicine 7, 127-139 (2015).

25. Wu, S.Y. et al. Salmonella spv locus suppresses host innate immune responses to bacterial infection. Fish & shellfish immunology 58, 387-396 (2016).

26. Gettler, K. et al. Prioritizing Crohn's disease genes by integrating association signals with gene expression implicates monocyte subsets. Genes and immunity (2019).

27. Takeuchi, K. Prostaglandin EP receptors and their roles in mucosal protection and ulcer healing in the gastrointestinal tract. Advances in clinical chemistry 51, 121-144 (2010).

28. Miyoshi, H. et al. Prostaglandin E2 promotes intestinal repair through an adaptive cellular response of the epithelium. The EMBO journal (2016).

29. Keubler, L.M., Buettner, M., Hager, C. & Bleich, A. A Multihit Model: Colitis Lessons from the Interleukin-10-deficient Mouse. Inflammatory bowel diseases 21, 1967-1975 (2015).

30. Ostanin, D.V. et al. T cell transfer model of chronic colitis: concepts, considerations, and tricks of the trade. American journal of physiology. Gastrointestinal and liver physiology 296, G135-146 (2009).

31. Chassaing, B., Aitken, J.D., Malleshappa, M. & Vijay-Kumar, M. Dextran sulfate sodium (DSS)-induced colitis in mice. Current protocols in immunology / edited by John E. Coligan ... [et al.] 104, Unit 15 25 (2014).

32. Wilk, J.N., Bilsborough, J. & Viney, J.L. The mdr1a-/- mouse model of spontaneous colitis: a relevant and appropriate animal model to study inflammatory bowel disease. Immunologic research 31, 151-159 (2005).

33. Holleran, G. et al. The Innate and Adaptive Immune System as Targets for Biologic Therapies in Inflammatory Bowel Disease. International journal of molecular sciences 18 (2017).

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

34. Ke, P., Shao, B.Z., Xu, Z.Q., Chen, X.W. & Liu, C. Intestinal Autophagy and Its Pharmacological Control in Inflammatory Bowel Disease. Front Immunol 7, 695 (2016).

35. Iida, T., Onodera, K. & Nakase, H. Role of autophagy in the pathogenesis of inflammatory bowel disease. World journal of gastroenterology 23, 1944-1953 (2017).

36. Wirtz, S. et al. Chemically induced mouse models of acute and chronic intestinal inflammation. Nature protocols 12, 1295-1309 (2017).

37. Belley, A. & Chadee, K. Prostaglandin E(2) stimulates rat and human colonic mucin exocytosis via the EP(4) receptor. Gastroenterology 117, 1352-1362 (1999).

38. Kaiser, G.C., Yan, F. & Polk, D.B. Mesalamine blocks tumor necrosis factor growth inhibition and nuclear factor kappaB activation in mouse colonocytes. Gastroenterology 116, 602-609 (1999).

39. Di Paolo, M.C., Merrett, M.N., Crotty, B. & Jewell, D.P. 5-Aminosalicylic acid inhibits the impaired epithelial barrier function induced by gamma interferon. Gut 38, 115-119 (1996).

40. Chaabane, W. & Appell, M.L. Interconnections between apoptotic and autophagic pathways during thiopurine-induced toxicity in cancer cells: the role of reactive oxygen species. Oncotarget 7, 75616-75634 (2016).

41. Sandborn, W.J. et al. Ustekinumab induction and maintenance therapy in refractory Crohn's disease. The New England journal of medicine 367, 1519-1528 (2012).

42. Abraham, C., Dulai, P.S., Vermeire, S. & Sandborn, W.J. Lessons Learned From Trials Targeting Cytokine Pathways in Patients With Inflammatory Bowel Diseases. Gastroenterology 152, 374-388 e374 (2017).

43. Colombel, J.F. et al. Infliximab, azathioprine, or combination therapy for Crohn's disease. The New England journal of medicine 362, 1383-1395 (2010).

44. Ordas, I., Mould, D.R., Feagan, B.G. & Sandborn, W.J. Anti-TNF monoclonal antibodies in inflammatory bowel disease: pharmacokinetics-based dosing paradigms. Clinical pharmacology and therapeutics 91, 635-646 (2012).

45. Poppe, D. et al. Azathioprine suppresses ezrin-radixin-moesin-dependent T cell-APC conjugation through inhibition of Vav guanosine exchange activity on Rac proteins. Journal of immunology 176, 640-651 (2006).

46. Oehlers, S.H. et al. Interception of host angiogenic signalling limits mycobacterial growth. Nature 517, 612-615 (2015).

47. Walton, E.M., Cronan, M.R., Beerman, R.W. & Tobin, D.M. The Macrophage-Specific Promoter mfap4 Allows Live, Long-Term Analysis of Macrophage Behavior during Mycobacterial Infection in Zebrafish. PloS one 10, e0138949 (2015).

48. VanDussen, K.L. et al. Development of an enhanced human gastrointestinal epithelial culture system to facilitate patient-based assays. Gut 64, 911-920 (2015).

49. Miyoshi, H. & Stappenbeck, T.S. In vitro expansion and genetic modification of gastrointestinal stem cells in spheroid culture. Nature protocols 8, 2471-2482 (2013).

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

50. Bresciani, E., Broadbridge, E. & Liu, P.P. An efficient dissociation protocol for generation of single cell suspension from zebrafish embryos and larvae. MethodsX 5, 1287-1290 (2018).

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Figures

Fig. 1. Impaired recovery of acidic-lysosome function and mucin production

after repeated DSS injury in zebrafish intestines. (A) Dose-dependent mortality

observed after repeated DSS injury (N=1739 from 11 clutches). Compared to the

single DSS Injury that has a 98-100% survival rate (day 6), repeated DSS injury

induces a high mortality rate of 63% in 0.25% DSS and 29% in 0.1% DSS. (B)

Hematoxylin and eosin (H&E) staining of a longitudinal section of zebrafish larvae

intestine. 20 larva per experimental conditions were embedded in paraffin and

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

sectioned with 10 μm per slides. Scale bar, 100 μm. (C, D) Quantification of neutral

red accumulation (top panel) and neutral red images (bottom panel) for single (C)

and repeated DSS injured zebrafish (D). Neutral red accumulation in the intestine

indicates normally functioning acidic lysosomes. The damage to lysosomal function is

fully recovered in two days with the single DSS injury in (C) (N=253 from 4 clutches)

but is impaired with the repeated DSS injury in (D) (N=137 from 3 clutches; day two

N=5 out of 13 from single clutch due to high mortality rate 61.5%). Neutral red

images for singe (C) and repeated injury (D) are shown as control, 0.1% and 0.25%

(w/v) DSS treatments at treatment time, one day and two days after. Scale bar, 100

µm. ***, P<0.001. **, P<0.01, *, P<0.05. Error bar, mean ± s.e.m.. (E) Images (left

panel) and quantification (right panel) of alcian blue staining intestine with single and

repeated injury with control and 24 hours after removing from DSS treatment. The

quantification of total alcian blue intensity from single (N=160 from 3 clutches) and

repeated injury (N=154 from 3 clutches). There was an increase in mucin production

after single DSS injury, but mucin production was impaired after repeated DSS injury.

Scale bar, 100 µm

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Figure 2. Lysosome-rich enterocyte-mediated bacterial protein uptake and

autophagy with injury and bacterial exposure. (A) Bright-field, fluorescent and

merged images of pHrodo-labeled heat-killed E. coli K12 MG1655 within a zebrafish

larval intestine with and without DSS treatment. 20 larva from each experimental

conditions were treated with the pHrodo-labeled E. coli proteins. The pHrodo dye is a

pH sensing dye indicating ingested E. coli proteins. Scale bar, 50 μm. The

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

quantification was shown at 0 hour time point of Fig. 3G. (B) Uptake of E. coli

proteins (red) co-localized with Cyto-ID positive autophagosomes (green) in intestine.

Scale bar, 20 μm. (C) Quantification of Cyto-ID intensities immediately following

single (top panel) and repeated DSS injury (bottom panel). 0.05% and 0.1% (w/v)

DSS were applied at treatment time. ***, P<0.001. Error bar, mean ± s.e.m.. (D)

Cyto-ID intensities measured one day after DSS treatment shows recovery and non-

recovery after single (top panel) and repeated DSS injury (bottom panel). Altogether

for c and d, N=544 from 6 clutches. ***, P<0.001. **, P<0.01. Error bar, mean ±

s.e.m.. (E) High levels of mortality with E .coli treatment following single DSS injury.

The left panel plots mortality rates after 1, 2 and 3 hours of heat-killed E. coli labeled

with Alexa-488 (green fluorescence) exposure following single DSS injury. *, P <

0.05, comparing mortality rates at each time point with and without DSS treatment.

N=341 from 3 clutches. Error bar, mean ± s.e.m.. The right panel shows systemic

penetration of E.coli in the dorsal aorta (upper arrow) and posterior cardinal vein

(lower arrow) with DSS treatment imaged 90 minutes after heat-killed E. coli

treatment. Scale bar 20 µm. Full bacterial invasion videos are shown in Movies S6

and S7.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Fig. 3. Effects of PGE2 and commonly utilized IBD medications on barrier

function, bacterial containment and LRE function. (A-C) PGE2 induces mucin

expression and release in a dose-dependent manner. (A) The images of Alcian blue

staining with treatments of 0.1, 1 and 10 μM PGE2. The quantification of whole gut

region in B, and the red box indicate quantification areas in C. Scale bar, 100 μm.

Alcian blue intensities of the full intestine (B) and the intestinal lumen (C) are shown

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

(N=100 from 3 clutches). **, P < 0.01. ***, P < 0.001. Error bar, mean ± s.e.m.. (D)

Quantification of relative Alcian blue intensity in a human enteroid differentiated

epithelial monolayer with and without 1 μM PGE2 treatment. ***, P < 0.001. N=3

patients with 3 biopsies each. (E) Experiment design of DSS and PGE2 treatment for

mortality assays (F) or ingested pHrodo-labeled E. coli proteins intensity (G). **, P <

0.01. *, P < 0.05. N=1145 from 9 clutches. Error bar, mean ± s.e.m.. (G) Fluorescent

intensity of ingested pHrodo-labeled E. coli proteins 1, 2 and 3 hours after removing

from DSS. N=240 from 3 clutches. Error bar, mean ± s.e.m.. (H) Applying PGE2 (left

panels), mesalamine (middle panels) and 6-mercaptopurine (right panels) to the DSS

injury model. The relative neutral red (top panels) and Cyto-ID (bottom panels)

intensities with treatment alone (dark blue), with treatment during DSS injury for 24

hours (light green) and with 5 hours of treatment after DSS removal (light blue) are

reported relative to untreated controls. ***, P < 0.001. **, P < 0.01. Neutral red-PGE2,

N=505 from 3 clutches. Neutral red-mesalamine, N=695 from 6 clutches. Neutral red-

6-mercaptopurine, N=705 from 3 clutches. Cyto-ID-PGE2, N=196 from 3 clutches.

Cyto-ID-mesalamine, N=189 from 3 clutches. Cyto-ID-6-mercaptopurine. N=213 from

3 clutches. Error bar, mean ± s.e.m..

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Fig. 4. Effects of PGE2 on mucus production and barrier protection.

Dis

ease

Mo

dels

& M

echa

nism

s •

DM

M •

Acc

epte

d m

anus

crip

t

Supplementary Figures

Fig. S1 Gut length is shortened with DSS injury. (A) Timeline of single DSS and repeated DSS

injuries. (B) Zebrafish images of untreated and single DSS injury (left panel). Red lines indicate

the measurement of gut length. Quantification of gut length (right panel). Comparing to

untreated controls, the gut length is significantly shortened after single DSS injury (P<0.001, N=

149 from 7 clutches). (C) The body axis length of untreated and 0.1% and 0.25% DSS treated

larvae.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S2. Optimization of neutral red staining. (A) Calibration curve of neutral red staining.

Neutral red staining equilibrium is between 4-6 hours. N=30 (B) Images of Neutral red co-

localizes with the lysosomal marker Lysosensor in zebrafish intestine (top panel). Total

magnification 200X. Bar, 50µm. Zoom in quantification images (bottom panel) from selected

area (white rectangle on top panel). Bar, 10µm. (C) Pearson correlation coefficient of Lyso-

sensor and Neutral Red intensities in (B) bottom panel. (D-E) Lysosome-rich enterocyte (LRE)

marker pllp enriched in Lysotracker-positive cells. 15 larvae intestines were dissected and cells

were collected after single cell isolation with both Lysotracker staining and no staining control.

The sorting result (D) showed 5.01% Lysotracker positive cells with minimum contamination

(less than 0.06%). (E) qPCR results of marker for macrophages (mpeg1), mfap4, epithelia

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

(epcam) and LREs (pllp) in Lysotracker positive, negative relative to unsorted control cells. (F)

The intestinal neutral red intensity recovers by 54 hours after removal from single DSS injury

(N=75 from 3 clutches).

Fig. S3. Inflammatory cytokines TNF and IL-1b mRNA are predominately induced in intestine

with single DSS injury but are also expressed in non-intestinal tissues after repeated injury. (A)

TNF expression after single (left) and repeated (right) injury (B) IL-1 expression after single

(left) and repeated (right) injury.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S4. Optimization of Alcian blue staining. (A) Calibration curve of Alcian blue staining. The

intensity of Alcian blue staining is linear but not saturated between 2 - 3.5 hours (N=40 from 3

clutches). (B) The R2 is 0.996 for the linear range.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S5. DSS reduces Alcian blue intensity after single DSS injury but does not affect goblet cell

numbers. (A) Images of Alcian blue staining with 0.1% and 0.25% DSS treatment. The data was

collected at the same time point as 0 hour in Figure 3e right panel. (B) Quantification of Alcian

blue intensity relative to control. (C) Quantification of goblet cells per mid-intestine. N=160 from

three clutches. (D) qPCR results of muc2.1 with single and repeated DSS injuries.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S6. Treatment of selective autophagy inhibitor Bafilomycin A1. (a) Dose-dependent

decrease of Cyto-ID staining with 10 and 20nM Bafilomycin A1 after 24 hours of treatment. ****,

P<0.0001 (b) No change in Neutral Red intensity.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S7. Confocal tiled image with the full zebrafish larvae intestine. (a) The images of bright-

field, lysosensor (green), pHrodo-E.coli (red) and merged images of red and green. pHrodo-

labelled E. coli were ingested in LRE-enriched region of the intestine. Bar, 200 µm. (b) The

images of bright-field, Cyto-ID (green), pHrodo-E.coli (red) and merged images of red and

green. Cyto-ID staining is present and not limited to the region with ingested E. coli. Bar, 200

µm.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S8. Immunoblot of the autophagy marker p62 with and without one course 0.25% DSS

treatment. p62 proteins accumulate in response of loss of autophagosomes 1, 2. In Agreement

with CytoID results (Fig. 2C), 0.25% DSS treatment induces accumulation of p62.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S9. Confocal tiled image with the full zebrafish larvae intestine with and without DSS

treatment. (a) The images of bright-field and Cyto-ID (green). Bar, 200 µm.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S10. Low levels of mortality with E .coli treatment following single low dose DSS injury. The

mortality rates after 1, 2 and 3 hours of heat-killed E. coli exposure following single DSS injury.

N=241.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S11. Quantification E. coli intensity in zebrafish aorta and vein with and without DSS

treatment. The E.coli fluorescence intensity of total 170 of zebrafish cells from two replicated

experiments were quantified (See Supplementary Methods). With DSS treatment, there are

significantly higher levels of E.coli uptake as measured by mean fluorescence in both dorsal

aorta and posterior cardinal vein. ****, P< 0.0001.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Fig. S12. (A) The images of Alcian blue staining with DSS alone or with treatments with 0.1, 1

and 10 μM PGE2 (left panel). The quantification of the whole gut region (red arrows) is shown in

the middle panel, and the blue arrows indicate quantification of lumen areas (right panel). Scale

bar, 100 μm. (N=100 from 3 clutches). **, P < 0.01. ***, P < 0.001. (B) Images of human

enteroid-differentiated epithelium monolayer with and without PGE2 treatment. Scale bar 100

µm.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Sequence (5' --> 3')

Gene Forward Reverse

epcam TTTGGATAAGAAACTTGTGTCTGAAG TGTAGTACGGTCGTCCTTATCTTTTT

il1b ATCAAACCCCAATCCACAGAGT GGCACTGAAGACACCACGTT

mfap4 ATGGCAATCGTGCTGTTCTT AACTTCTTGTGGCGTGTCA

mpeg1 TCACCTGCTGATGCTCTGCTG TCTGTGGAATGACAAAGACCTC

muc2.1 CAACATCGATGGCTGCTTCTG CTGACAGTAACATTCTTCCTCGC

p62 CGATGTTTTTGTCGGTCTCA CAAGAGCCAAACCCATCATT

pllp GGCCTTTGGTGCTGTTGGTAT GCTGGCTGTGTATAACAGGGT

rppo CTGAACATCTCGCCCTTCTC TAGCCGATCTGCAGACACAC

tnfa ACAACACTATTTACCTCGGC ACCAAACACCCCAAAGAAGG

Table S1. Primers information for qPCR.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Supplementary movies

Movie 1. The confocal z-stack movies of the posterior mid-intestine of Tg(mfap4: turquoise) fish.

Zebrafish treated with pH-rodo E. coli protein and co-staining with Cyto-ID for autophagy. The

mfap4 positive cells show as yellow puncta.

Movie 2. The confocal z-stack movies of the posterior mid-intestine of Tg(mfap4: turquoise) fish.

Zebrafish treated with pH-rodo E. coli protein and co-staining with Cyto-ID for autophagy. The E.

coli protein inside LRE cells shows in red puncta.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Movie 3. The confocal z-stack movies of the posterior mid-intestine of Tg(mfap4: turquoise) fish.

Zebrafish treated with pH-rodo E. coli protein and co-staining with Cyto-ID for autophagy. The

autophagy shows in green puncta.

Movie 4. The confocal z-stack movies of the posterior mid-intestine of Tg(mfap4: turquoise) fish.

Zebrafish treated with pH-rodo E. coli protein and co-staining with Cyto-ID for autophagy. A

merged movie of S. movies 1-3 was shown. The three-color colocalization suggests ingested E.

coli protein in the autophagy pathway.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Movie 5. The confocal time-lapse movies of posterior mid-intestine in wildtype zebrafish.

Zebrafish treated with pH-rodo E. coli protein in red color and co-staining with Cyto-ID for

autophagy in green color. The red and green color both localize cells in the posterior mid-

intestine over time.

Movie 6. Bacterial invasion movie without DSS treatment. Green color indicates fluoresce

Alexa-488 labeled heat-killed E. coli in the gut lumen but not dorsal aorta or cardinal vein.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n

Movie 7. Bacterial invasion movie with DSS treatment. Green color indicates fluoresce

Alexa-488 labeled heat-killed E. coli in the gut lumen, dorsal aorta, and cardinal vein.

Disease Models & Mechanisms: doi:10.1242/dmm.037432: Supplementary information

Dis

ease

Mo

dels

& M

echa

nism

s •

Sup

plem

enta

ry in

form

atio

n