Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice

Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice

Accepted Manuscript Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice Junichi Matsuo, Shunichi Kimura, Akihiro Y...

8MB Sizes 0 Downloads 61 Views

Accepted Manuscript Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice Junichi Matsuo, Shunichi Kimura, Akihiro Yamamura, Cai Ping Koh, Md Zakir Hossain, Dede Liana Heng, Kazuyoshi Kohu, Dominic Chih-Cheng Voon, Hiroshi Hiai, Michiaki Unno, Jimmy Bok Yan So, Feng Zhu, Supriya Srivastava, Teh Meng, Khay Guan Yeoh, Motomi Osato, Yoshiaki Ito PII: DOI: Reference:

S0016-5085(16)35113-7 10.1053/j.gastro.2016.09.018 YGAST 60710

To appear in: Gastroenterology Accepted Date: 19 September 2016 Please cite this article as: Matsuo J, Kimura S, Yamamura A, Koh CP, Hossain MZ, Heng DL, Kohu K, Chih-Cheng Voon D, Hiai H, Unno M, Yan So JB, Zhu F, Srivastava S, Meng T, Yeoh KG, Osato M, Ito Y, Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice, Gastroenterology (2016), doi: 10.1053/j.gastro.2016.09.018. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Identification of Stem Cells in the Epithelium of the Stomach Corpus and Antrum of Mice

RI PT

Short Title Identification of stomach stem cells by eR1

SC

Authors and Affiliations

Junichi Matsuo,1,11 Shunichi Kimura,1,2,11 Akihiro Yamamura,1,2,11 Cai Ping Koh,1 Md Zakir

M AN U

Hossain,1 Dede Liana Heng,1 Kazuyoshi Kohu,1 Dominic Chih-Cheng Voon1,13, Hiroshi Hiai,10 Michiaki Unno,2 Jimmy Bok Yan So,5 Feng Zhu,3 Supriya Srivastava,1 Teh Meng,4 Khay Guan Yeoh,3,6 Motomi Osato,1,7,8,9,12 and Yoshiaki Ito,1,3,12 1

Cancer Science Institute of Singapore, National University of Singapore, 14 Medical Drive,

2

TE D

Singapore 117599

Department of Surgery, Graduate School of Medicine, Tohoku University, 1-1 Seiryo-machi,

Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore,

AC C

3

EP

Aoba-ku, Miyagi, Japan 9808574

C/O NUHS Tower Block IE Kent Ridge Road, Singapore 119228 4

Department of Pathology, National University Health System, 1E Kent Ridge Road,

Singapore 119228 5

Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore

C/O NUHS Tower Block 1E Kent Ridge Road, Singapore 119228

1

ACCEPTED MANUSCRIPT 6

Department of Gastroenterology and Hepatology, National University Health System, 1E

Kent Ridge Road, Singapore 119228 7

Institute of Bioengineering and Nanotechnology, A*STAR, 31 Biopolis way, Singapore

8

RI PT

138669 Department of Pediatrics, Yong Loo Lin School of Medicine, National University of

Singapore, 14 Medical Drive, Singapore 119228

International Research Center for Medical Sciences, Kumamoto University, 2-2-1 Honjo,

SC

9

10

M AN U

Chuo-ku, Kumamoto 860-0811, Japan

Kyoto Disease Model Institute, Kyoto Science and Technology Center, Rm 22, 14

11

Co-first authors

12

Co-correspondance

TE D

Yoshidakawaramachi, Sakyo-ku, Kyoto 606-8305

JM, SK, and AK contributed equally to this work. YI and MO are co-corresponding authors.

Institute for Frontier Science Initiative and Division of Genetics, Cancer Research Institute,

AC C

13

EP

DCCV: Current address:

Kanazawa University, Kanazawa, Ishikawa, Japan Grant Support

This research is supported by the National Research Foundation Singapore and the Singapore Ministry of Education under its Research Centres of Excellence initiative. The study was also supported by the Singapore National Research Foundation under its Translational and Clinical Research Flagship Program and administered by the Singapore

2

ACCEPTED MANUSCRIPT Ministry of Health’s National Medical Research Council (Grant NMRC/TCR/001/2007 and NMRC/TCR/009-NUHS/2013). Acknowledgements

RI PT

We thank Dr. Nicholas Barker for providing us with Lgr5-GFP-ires-CreERT2 mouse and CreERT2 construct. This research is supported by the National Research Foundation Singapore and the Singapore Ministry of Education under its Research Centres of Excellence

SC

initiative. The study was also supported by the Singapore National Research Foundation under its Translational and Clinical Research Flagship Program and administered by the Ministry

of

Health’s

National

Medical

M AN U

Singapore

Research

Council (Grant

NMRC/TCR/001/2007 and NMRC/TCR/009-NUHS/2013). Abbreviations

TE D

b-HLH, Basic helix-loop-helix; Cldn18, Claudin18; EGFP, Enhanced green fluorescent protein; EYFP, Enhanced Yellow Fluorescent Protein; H. Pylori, Helicobacter Pylori; H&E, Hematoxylin and Eosin; IF, Immunofluorescence; LSL, Lox-Stop-Lox; MAPK, Mitogen-activated protein

AC C

metaplasia.

EP

kinases; TG, Transgenic; TFF2, Trefoil Factor2; SPEM, Spasmolytic polypeptide-expressing

Correspondence

Yoshiaki Ito, MD, PhD

Cancer Science Institute of Singapore, National University of Singapore, 14 Medical Drive, 117599, Singapore Tel: +65-6516-2242 / Fax: +65-6873-9664

3

ACCEPTED MANUSCRIPT Email: [email protected] Motomi Osato, MD, PhD Cancer Science Institute of Singapore, National University of Singapore, 14 Medical Drive,

RI PT

117599, Singapore Tel: +65-6601-1442 / Fax: +65-6873-9664

Disclosures

M AN U

No potential conflicts relevant to this manuscript.

SC

Email: [email protected]

Author Contributions

JM, SK and AY did most of the mouse experiments and writing the draft. JM worked more

TE D

on stem cell part, while SK and MU studied more on gastritis. Organoid studies were mainly done by AY. DLH, DCCV and KK participated in histological analysis. MZH made transgenic mice using previously reported eR1-GFP that is linked to CreERT2, and CPK crossed and

EP

maintained transgenic mice. KGY, JBYS and FZ provided human material and clinical data.

AC C

HH, SS and TM histologically analysed and interpreted the results. MO and YI conceived the aim of the study and YI is primarily responsible for writing of the manuscript.

4

ACCEPTED MANUSCRIPT ABSTRACT BACKGROUND & AIMS: Little is known about mechanisms of gastric carcinogenesis, partly because it has been a challenge to identify characterize gastric stem cells. Runx genes

RI PT

regulate development and their products are transcription factors associated with cancer development. A Runx1 enhancer element, eR1 is a marker of hematopoietic stem cells. We studied expression from eR1 in stomach and the roles of gastric stem cells in gastric

M AN U

SC

carcinogenesis in transgenic mice.

METHODS: We used in situ hybridization and immunofluorescence analyses to study expression of Runx1 in gastric tissues from C57BL/6 (control) mice. We then created mice that expressed enhanced green fluorescent protein (EGFP) or CreERT2 under the control of

TE D

eR1 (eR1-CreERT2;Rosa-LSL-tdTomato, eR1-CreERT2;Rosa-LSL-EYFP mice). Gastric tissues were collected and lineage-tracing experiments were performed. Gastric organoids were

EP

cultured from eR1-CreERT2(5-2);Rosa-LSL-tdTomato mice and immunofluorescence analyses were performed. We investigated the effects of expressing oncogenic mutations in stem

AC C

cells under control of eR1 using eR1-CreERT2;LSL-KrasG12D/+ mice; gastric tissues were collected and analyzed by histology and immunofluorescence.

RESULTS: Most proliferation occurred in the isthmus; 86% of proliferating cells were RUNX1 positive and 76% were MUC5AC positive. In eR1-EGFP mice, EGFP signals were mainly detected in the upper part of the gastric unit, and 83% of EGFP-positive cells were located in the

5

ACCEPTED MANUSCRIPT isthmus/pit region. We found that eR1 marked undifferentiated stem cells in the isthmus and a smaller number of terminally differentiated chief cells at the base. eR1 also marked cells in the pyloric gland in the antrum. Lineage tracing experiments demonstrated that stem cells in the isthmus and antrum continuously gave rise to mature cells to maintain the

RI PT

gastric unit. eR1-positive cells in the isthmus and pyloric gland generated organoid cultures in vitro. In eR1-CreERT2;LSL-Kras G12D/+ mice, MUC5AC-positive cells rapidly differentiated

SC

from stem cells in the isthmus, resulting in distinct metaplastic lesions similar to that

M AN U

observed in human gastric atrophy.

CONCLUSIONS: Using lineage tracing experiments in mice, we found that a Runx1 enhancer element, eR1, promotes its expression in the isthmus stem cells of stomach corpus as well

TE D

as pyloric gland in the antrum. We were able to use eR1 to express oncogenic mutations

EP

in gastric stem cells, proving a new model for studies of gastric carcinogenesis.

AC C

KEY WORDS: stomach cancer, mouse model, cancer stem cell, oncogene

6

ACCEPTED MANUSCRIPT Introduction Understanding the cells of origin in cancer is important for the development of effective strategies to detect, prevent and treat cancer. Tissue stem cells are generally

RI PT

studied as candidate cells from which cancer originates. Adult stem cells with a minimal differentiation phenotype1 and fully differentiated cells that have acquired the capacity to replicate and differentiate have been described.2-4 Furthermore, recent developments in

SC

stem cell research revealed multiple types of stem cells identified by different markers in the same tissues.5-7

M AN U

Gastric cancer is the second most frequent cause of cancer death in the world.8 However, until recently, molecular understanding of gastric carcinogenesis has been hampered by the paucity of studies concerning stem cells in the stomach epithelium. Anatomically, the stomach is roughly divided into the distal part (antrum) and the main

TE D

body (corpus).9 In the antrum, the upper three-quarters of the pyloric gland contain Muc5ac-producing cells, while the bottom portion contains rapidly growing stem/progenitor

EP

cells. By contrast, the gastric unit in the corpus contains Muc5ac-producing surface epithelial cells (pit cells). These cells are rapidly turned over and their half-life is about 3

AC C

days. Below this is a small region termed the isthmus, where corpus stem cells are presumed to be present based on radio-labeling and morphological studies using electron microscopy. Below the isthmus, there is a neck region where GS-II+ mucous neck cells and acid-producing parietal cells are located. Pepsinogen C-producing zymogenic cells, called chief cells, are present at the base.10 The half-life of parietal cells and chief cells is several months (see diagram in Supplementary Figure 1A).9, 11

7

ACCEPTED MANUSCRIPT Runx genes are developmental regulators and their products function as transcription factors and are often involved in cancer. There are three Runx genes in mammals, namely, Runx1, Runx2, and Runx3.12 Runx1 is a key regulator of hematopoiesis, is required for the generation of hematopoietic stem cells from endothelial cells,13 and is also

RI PT

critical for the maintenance of hematopoietic stem cells.14, 15 Anomaly of RUNX1 is widely involved in Leukemia.16 The non-coding region between the P1 and P2 promoters of Runx1

SC

is considered to be composed of multiple cis-regulatory elements in a mosaic fashion. One such element (~270 bp) was identified as an enhancer of Runx1 expression in hematopoietic

M AN U

stem cells and termed the +24 mouse conserved non-coding element (referred to as Runx1 enhancer element (eR1) in this study).17, 18 More recently, Runx1 was also demonstrated to have stem cell functions in skin.19 The many recent reports regarding Runx1 expression in epithelial cells of several organs20-23 prompted us to study Runx1 expression and eR1 activity

TE D

in various tissues, and we found that eR1 marked tissue stem cells of the stomach corpus

EP

and antrum.

Materials and Methods

AC C

Detailed Materials and Methods are described in Supplementary Materials. Mice and Treatment

Transgenic mice expressing eR1-EGFP were described previously.17 Transgenic mice expressing eR1 and tamoxifen-inducible CreERT2 were generated using the eR1 sequence, the mouse heat-shock protein 68 basal promoter sequence, and the CreERT2 transgene following a strategy similar to that previously described.17 The generation and details of

8

ACCEPTED MANUSCRIPT other mice are described in the Extended Experimental Procedures. All animals were handled in strict accordance with good animal practice as defined by the Institution of Animal Care and Use Committee (IACUC). All animal work was approved by the IACUC and

RI PT

the Office of Safety, Health, and Environment (OSHE) at the National University of Singapore. To induce gene expression, 3- and 6-week-old,eR1-CreERT2;Rosa-LSL-tdTomato (Rosa-tdTomato), eR1-CreERT2;Rosa-LSL-EYFP (Rosa-EYFP), eR1-CreERT2:LSL-K-rasG12D/+, and

SC

eR1-CreERT2:Rosa-tdTomato:LSL-K-rasG12D/+ mice were injected with 2, 4, or 6 mg of tamoxifen. Mice were analyzed 1 day, 1 week, 3 months, 6 months, and 1 year after

M AN U

treatment. LSL-K-rasG12D will be refer to as K-rasG12D in this paper.

Results

TE D

Runx1 Expression in Mouse Stomach Epithelial Cells

In situ hybridization with a Runx1 anti-sense probe detected Runx1 mRNA expression in the upper and lower parts of the corpus gastric unit (Figure 1A). Immunofluorescence (IF)

EP

staining using an anti-Runx1 antibody showed that Runx1 was expressed in E-cadherin+

AC C

epithelial cells in the upper part of the mouse stomach corpus (Figure 1B), and was more infrequently expressed at the base (Figure 1C, the histological locations of various parts of the stomach are illustrated in Supplementary Figure 1A). CD45+ immune cells expressed higher levels of Runx1 than epithelial cells (Figure 1B, arrows). The expression level of Runx1 protein relative to that of Runx1 mRNA is much lower in the lower part of the epithelium. This may be due to the difference in translation rate of Runx1 protein in the lower part of

9

ACCEPTED MANUSCRIPT the epithelium. The differential expression of mRNA and protein was also described for Trefoil Factor 2 (TFF2).24 The majority of Ki67 labeling was located in the isthmus, and 86% co-localized with

RI PT

Runx1+ cells (Supplementary Figure 1B and G). Subfractions of Muc5ac+ cells (Figure 1D) and GS-II+ neck cells (Figure 1E) were Runx1+. In total, 76% of Ki67+ cells expressed Muc5ac (Supplementary Figure 1C and G), 14% of Ki67+ cells were GS-II+ (Supplementary Figure 1D

SC

and G), and 3.5% of Ki67+ cells expressed HK-ATPase (Supplementary Figure 1E and G). Some Ki67+ cells at the base were also positive for pepsinogen C, but these constituted only

M AN U

0.58% of the total Ki67+ cell population (Supplementary Figure 1F and G). In addition to the corpus, Runx1 was also expressed in epithelial cells at and near the bottom of the pyloric gland (Supplementary Figure 1H). Well-characterized stem cell marker

TE D

Lgr5 was expressed at the bottom (Supplementary Figure 1I).23

EP

Identification of eR1-Enhanced Green Fluorescent Protein (EGFP)+ Cells in the Corpus A pilot investigation suggested that eR1 activity is also detected in stomach corpus

AC C

epithelial cells. We examined Enhanced Green Fluorescent Protein (EGFP)+ cells in the corpus epithelium of transgenic mice harboring eR1-EGFP (See experimental design in Figure 2A). EGFP signals were mainly detected in the upper part of the gastric unit (Figure 2B), and EGFP+ cells also expressed E-cadherin (Figure 2C and D). EGFP+ cells were found in 43% of gastric units and negative ones in 57% (Supplementary Figure 2A). In accordance with a previous report, eR1-EGFP signals were not detected in CD45+ differentiated blood cells (Supplementary Figure 2B), whereas some signals were observed in the stroma and

10

ACCEPTED MANUSCRIPT muscle layer of the corpus (Figure 2B).17 In total, 83%, 7%, and 10% of EGFP+ cells were located in the isthmus/pit region, neck, and base, respectively (Supplementary Figure 2C). Interestingly, 80% of EGFP+ cells co-localized with Ki67 at isthmus (Figure 2C, Supplementary Figure 2D-F). EGFP signals also co-localized with a portion of Runx1+ cells in the isthmus

RI PT

(Supplementary Figure 2G, arrows), and some EGFP+ cells expressed low levels of Runx1 (Supplementary Figure 2G, arrowhead).

SC

To assess whether eR1+ cells are involved in tissue regeneration, we treated mice with tamoxifen (6 mg/mouse), which induces damage in parietal cells, although the tissue

M AN U

rapidly recovered.25 One week after tamoxifen treatment, E-cadherin+ gastric units were mildly disorganized and the number of parietal cells was lower than in untreated tissue (Supplementary Figure 3A). The number of EGFP+ cells per gastric unit was higher in tamoxifen-treated tissue than in untreated tissue (Supplementary Figure 3B-E). In untreated

TE D

tissue, 93.3% and 6.7% of EGFP+ gastric units contained 1–2 and ≥3 EGFP+ cells, respectively. In tamoxifen-treated tissue at 3 days post-treatment, 56.6%, 32.1%, and 11.3% of EGFP+ gastric units contained 1–2, 3–4, and >4 EGFP+ cells, respectively. Moreover, in tamoxifen-

EP

treated tissue at 7 days post-treatment, 58.1%, 20.9%, and 20.9% of EGFP+ gastric units

AC C

contained 1–2, 3–4, and >4 EGFP+ cells, respectively (Supplementary Figure 3F). These results suggest that eR1+ cells play a significant role in tissue regeneration and possess stem cell properties.

Lineage Tracing in the Corpus Gastric Unit Using eR1 Activity We generated more than ten transgenic mouse lines harboring eR1-CreERT2 and studied two of them, TG(5-2) (higher copy number) and TG(3-17) (lower copy number).

11

ACCEPTED MANUSCRIPT Although the former had higher Cre recombinase activity than the latter, the results obtained using these lines were qualitatively indistinguishable. To assess the stem cell properties of eR1+ cells, we performed lineage tracing

RI PT

experiments. We activated CreERT2 using tamoxifen in transgenic mice carrying eR1CreERT2(5-2);Rosa-tdTomato to label eR1+ cells and their progeny with tdTomato and observed them at 1 day, 3 months, 6 months, and 1 year after tamoxifen treatment (see

SC

Supplementary Figure 4A for experimental design). One day after 6-week-old mice were injected with tamoxifen, tdTomato+ cells were detected in the isthmus, where eR1-EGFP

M AN U

was detected (Figure 2E and 3A, left panel). The number of tdTomato+ cells increased with time, as the period after tamoxifen injection increased (Figure 3A). One year after tamoxifen injection, more gastric units were entirely labeled by the fluorescent protein, and these signals co-localized with staining with anti-pepsinogen C, anti-HK-ATPase, and anti-Muc5ac

TE D

antibodies and GS-II lectin (Figure 3B–E). Fluorescent labeling of gastric units was maintained for at least 1 year (Figure 3A), which suggests that eR1+ cells kept differentiating into their progeny and regenerating corpus gastric units during this time. Therefore, we

EP

conclude that the activity of eR1 marks stem cells in the corpus and designated these cells

AC C

as stem cells at the isthmus (SC-ist). As described above, about 10% of eR1+ cells were detected at the base, which express pepsinogen C (Supplementary Figure 2H). Similarly, tdTomato could also be detected in chief cells in the lineage tracing experiment at 1 day after tamoxifen treatment (Figure 2F). eR1+ cells at the base possessed a tissue regeneration capacity. At 3 and 7 days after damage induction, increased number of EGFP+ cells were detected at the base of the gastric unit in eR1-EGFP mice (Supplementary Figure 3D and E). At 1 year after tamoxifen

12

ACCEPTED MANUSCRIPT injection to eR1-CreERT2;Rosa-tdTomato, ribbon-like signals were observed in pepsinogen C+ cells (Figure 3F, left). Similarly, we observed lineage tracings apparently originated from eR1+ cells in base which reached the GS-II+ cells, parietal cells and beyond (Figure 3F, middle and right). These results suggest that eR1+ cells at the base may have a capacity to

RI PT

proliferate. We have tentatively designated eR1+ pepsinogen C-producing cells as stem cells at the base (SC-bs), although we do not have definitive evidence that these lineage tracings

SC

originated from the base, not from isthmus.

To minimize toxicity of tamoxifen, we examined the effect of 2 mg, which is

M AN U

considered to be a non-toxic dose of tamoxifen,25 to 6 weeks old TG(5-2). As shown in Supplementary Figure 4B and C, the data obtained after 2 mg tamoxifen injection were qualitatively indistinguishable from the data obtained after 4 mg tamoxifen injection (Figure

TE D

3A).

Generation of Organoid Culture from eR1+ Cells.

EP

We then isolated gastric units from eR1-CreERT2(5-2);Rosa-tdTomato mice 1 day

AC C

after tamoxifen injection and digested them into single cells. Then, we incubated them in vitro for organoid culture. The eR1+/tdTomato+ cells generated tdTomato+ organoids in the culture medium with Wnt signaling pathway-dependent WENR (Wnt3a, EGF, Noggin, Rspondin) (Figure 4A).26 We also observed organoids generated from FACS-isolated eR1+/tdTomato+ (eR1high) cells, whereas organoids were rarely observed in eR1-/tdTomato(eR1neg) cells (Figure 4B, Supplementary Figure 5A). These results strongly suggest that only the stem cells have the capability to generate organoids in culture, as shown for intestinal stem cells.27 In addition to the WENR culture condition, eR1+/tdTomato+ cells formed

13

ACCEPTED MANUSCRIPT tdTomato+ organoids under Notch-signaling pathway-dependent ENJ (EGF, Noggin, Jagged1) culture conditions (Figure 4C).28 In total, 27 and 21 tdTomato+ organoids were generated from 500 cells under the WENR and ENJ culture conditions, respectively (Supplementary Figure 5B). The difference in the rate of successful organoid formation between the two

RI PT

different culture conditions was statistically insignificant (P = 0.34) (Supplementary Figure 5B).

SC

After incubation in vitro for 7 days, we examined whether the stem cells that initiated the organoid underwent differentiation. All cells in the organoid expressed E-

M AN U

cadherin (Figure 4E), indicating that they were epithelial cells. We observed that the stem cells differentiated into at least three lineages, namely, neck cells, pit cells and chief cells (Figure 4D). Many cells in the organoid were positive for Ki67 (Figure 4E). Runx1 expression was also detected (Figure 4F). Some Runx1 expressing cells could be stem cells while some

TE D

could be differentiated cells such as pit cells.

EP

Induction of Foveolar Hyperplasia and Antralization by eR1-CreERT2-activated K-rasG12D

AC C

We used the eR1-CreERT2 system to study step-wise carcinogenesis in the stomach (see Supplementary Figure 6A for the experimental approach). Hematoxylin and eosin (H&E) staining of tissue from eR1-CreERT2(3-17);K-rasG12D mice treated with tamoxifen, showed Ushape tube- or gland-like structures which seemed to be growing towards the base (Figure 5A). The results shown here were obtained using eR1-CreERT2 (3-17) strain to see morphological feature of well-separated U-shaped tube-like structure. E-cadherin staining clearly showed that the abnormal gland-like structure was composed of epithelial cells entirely (Figure 5B). P-Erk, a downstream target of K-rasG12D, was detected in such cells,

14

ACCEPTED MANUSCRIPT suggesting that they express K-rasG12D (Figure 5B). The eR1-CreERT2:K-rasG12D-induced glandlike structure was surrounded by CD45+ immune cells (Figure 5B, Supplementary Figure 6B and C).

RI PT

IF staining also showed that all the cells in the gland-like structure express Muc5ac and a cluster of cells near the bottom of this structure express Ki67 (Figure 5C). Ki67+ cells in normal epithelium clustered around the isthmus. This indicates that K-rasG12D induces robust

SC

differentiation of SC-ist mainly into Muc5ac+ cells such that Ki67+ cells in the structure moved towards the base and were positioned closer to the base than to the cluster of Ki67+

M AN U

cells in normal epithelium (Figure 5C). Foveolar hyperplasia is consistent with the reported findings of K19 promoter driven oncogenic K-ras expression in the stomach.29,30 The stem cells marked by eR1, therefore, must be present near the bottom of the gland-like structure. A notable feature of this gland-like structure is the presence of isthmus often observed near

TE D

the base (Figure 5B, arrows). The cells are densely packed in this area, suggesting rapid cell growth, and they show strong p-Erk staining (Figure 5B, right bottom). The cells in the isthmus area also overlaps with the staining of Ki67 (Figure 5C, Supplementary Figure 6D)

EP

and express high level of CD44 (Figure 5D). It appears that stem cells at the original isthmus

AC C

of the corpus identified by eR1 moved to the isthmus of newly formed gland-like structure induced by K-rasG12D.

The upper two-thirds or three-quarters of the fully elongated gland-like structure were Muc5ac+, the underlying region, including the base, was CD44+, and the base was also GS-II+ (Figure 5E, left of left panel). Ki67+ cells were located between Muc5ac+ cells and GSII+ cells (Figure 5E, left of middle panel). These labeling patterns were strikingly similar to those in the normal pyloric gland (Figure 5E, right of left and middle panel), i.e., the entire

15

ACCEPTED MANUSCRIPT pyloric gland-like structure was generated from SC-ist upon K-rasG12D activation. It is worth noting that eR1+ stem cells in the isthmus of corpus appear to be located in the isthmus of newly formed pyloric gland-like structure induced by K-rasG12D. It is known that isthmus of

RI PT

normal pyloric gland is the place where rapidly growing stem cells are located.31 An important function of the pyloric gland is to produce gastrin.32 We found that the normal pyloric gland expressed gastrin, whereas gastrin was undetectable in the K-rasG12D-

SC

induced pyloric gland-like structure (Figure 5E). Alcian blue, an intestinal mucin marker, is known to stain the base of the normal pyloric gland (Supplementary Figure 6F, left),

M AN U

whereas it weakly stained the pyloric gland-like structure (Supplementary Figure 6F, right). Muc2, an intestinal mucin, was also detected in the K-rasG12D-induced structure but the labeling was much weaker than in the intestine (Supplementary Figure 6G). However, we did not detect Cdx2, which is highly expressed in intestine and drives Muc2 expression

TE D

(Supplementary Figure 6I). Cdx1 was also not detectable (Supplementary Figure 6H). This indicates that K-rasG12D expression in SC-ist does not fully induce the intestinal phenotype in the stomach corpus within 3 months of observation period. These results suggest that K-

EP

rasG12D-induced metaplastic lesions in the corpus recapitulated antralization, which is often

AC C

observed in human stomach chronically infected with H. pylori. (see Discussion). When labeled with an anti-HK-ATPase antibody, a few parietal cells were detected within the gland-like structure (Figures 5D). Parietal cells were not differentiated from SC-ist after activation of K-rasG12D; therefore, parietal cells found within the gland-like structure must be present prior to K-rasG12D activation, presumably within the lower part of the same gastric gland. Chief cells also appeared to be incorporated into gland–like structure

16

ACCEPTED MANUSCRIPT (Supplementary Figure 6J). When this structure was fully elongated to the base, parietal cells and chief cells were no longer detectable within the gland-like structure (Figure 5F). Chronic inflammation after H pylori infection induces oxyntic atrophy. Gastric

RI PT

atrophy is considered to be precancerous.33, 34 However, the mechanism by which parietal cells are lost is not known. Since H. pylori activates Ras-Mapk pathway in infected cells,35 the phenomenon that we observe in K-ras expressed cells may represent the process that

SC

occurs in H. pylori infected stomach epithelium. We showed above that K-rasG12D expression in SC-ist attracts immune cells around the cells and immune system might be affecting the

M AN U

fate of parietal cells. We observed apoptosis of parietal cells that are present in the Ushaped gland (Supplementary Figure 7A and B) but not necrosis of the cells (Supplementary Figure 7C).36 Autophagy must be explored further (Supplementary Figure 7E).37,

38

Interestingly, we observed that parietal cells appear to be expelled from the gland–like

TE D

structure (Supplementary Figure 7D). These observations suggested that some parietal cells might be eliminated by cell competition.39 It appears that multiple mechanisms are involved

EP

in the process of elimination of parietal cells.

Runx1 functions in SC-ist. It is interesting to note that differentiated pit cells also

AC C

express Runx1, particularly strongly at the bottom of growing U-shaped gland where Ki67 and CD44 are also strongly expressed (Supplementary Figure 7F and G). We are studying the function of Runx1 within the stem cells and the cells after differentiation.

Induction of Metaplasia by eR1-CreERT2-Activated K-rasG12D in SC-bs

17

ACCEPTED MANUSCRIPT When K-rasG12D was expressed in SC-bs, abnormal lesions formed at the bottom of the corpus epithelium (referred to as bottom-type lesions) (Figure 6A). These lesions are distinct from the U-shaped gland-like structure formed when K-rasG12 is expressed in SC-ist. Three examples of bottom-type lesions are shown (Figure 6B–6D). These lesions were

RI PT

labeled with tdTomato, suggesting that each of them was derived from probably single eR1+ cells and underwent multiple cell divisions. They also expressed pepsinogen C, suggesting that they originated from chief cells. Most of these cells also reacted with GS-II, suggesting

SC

that chief cells are undergoing retro-differentiation into their precursor mucous neck cells.40

separated from chief cells (Figure 1E).

M AN U

In WT mice, GS-II+ cells are located in the central part of the gastric unit and are mostly

The phenomenon that chief cells express the mucous neck cell marker GS-II is similar to that described for spasmolytic polypeptide-expressing metaplasia (SPEM), which is

TE D

considered to be a precancerous state.33, 34, 41 Importantly, some cells in bottom-type lesions expressed Ki67, whereas normal chief cells rarely do, suggesting that chief cells acquired a proliferation capacity (Figure 6E). Bottom-type lesions also expressed CD44, which is often

EP

considered to be a stem cell marker (Figure 6F). The lower parts of these lesions were

AC C

surrounded by CD45+ immune cells (Supplementary Figure 6C), again suggesting that KrasG12D-expressing cells attract immune cells, similar to the observations made in previous studies.29, 41

RUNX1 Expression and Pseudopyloric Metaplasia in Human Stomach Consistent with mouse studies, RUNX1 was expressed in the isthmus, where stem cells reside, in human stomach (Supplementary Figure 8A). In addition, RUNX1 was strongly

18

ACCEPTED MANUSCRIPT expressed at the bottom of human pyloric glands which gradually decreased towards the surface (Supplementary Figure 8B). H. pylori infection activates the K-ras-MAPK pathway.35 To assess the clinical relevance of our mouse study to gastritis, we examined the presence of pseudopyloric metaplasia in serial tissue sections of the corpus of patients diagnosed

RI PT

with H. pylori-associated gastritis (Supplementary Figure 8C and D). Consistent with the healthy antrum, histological analysis of H&E staining did not detect parietal cells in the

SC

corpus of gastritis tissue (Supplementary Figure 8C and D). Immunohistochemical staining detected MUC6, a mucous neck cell marker, at the bottom of the gastritis corpus and in

M AN U

healthy antrum (Supplementary Figure 8D). Furthermore, MUC5AC was detected above the cluster of MUC6+ cells in healthy antrum and the gastritis corpus, and these MUC5AC+ cells occupied more than half of the glands (Supplementary Figure 8D). However, unlike in healthy antrum, gastrin was not detected in the gastritis corpus (Supplementary Figure 8D).

TE D

CDX2 was not detected in healthy antrum or MUC6+ lesions of gastritis tissue (Supplementary Figure 8D). Small CDX2+ lesions showed typical features of intestinal metaplasia (IM), as indicated by the presence of mucinous vacuoles (Supplementary Figure

EP

8D, arrow). These observations suggest that H. pylori infection induced the formation of pseudopyloric metaplasia in human corpus, and this process is similar to the induction of

AC C

pseudopyloric metaplasia when stem cells expressed K-rasG12D in mouse corpus.

eR1 Activities in the Pyloric Gland We also detected eR1 activity in the pyloric gland in the antrum (Figure 7A). eR1EGFP marked a small number of cells near the isthmus region where Ki67 is highly expressed (Figure 7A), and the position is significantly higher than the region where Lgr5+ and CD44+

19

ACCEPTED MANUSCRIPT cells are located.23,24 In one day post tamoxifen injection of eR1-CreERT2;Rosa-EYFP mice, eR1+ cells in the antrum did not seem to overlap with CD44 labeling (Figure 7B). Lineage tracing experiments showed that major cells in antrum were derived from eR1+ cells in 6–12

RI PT

months after tamoxifen injection (Figure 7C). When we isolated pyloric glands from eR1-CreERT2;Rosa-tdTomato mice 1 day post tamoxifen injection, we observed tdTomato near to but not at the bottom in most cases

SC

(Figure 7D). This is consistent with the results shown in Figure 7A–B. When we incubated these glands in vitro, tdTomato+ organoids developed (Figure 7E). tdTomoto+ organoids

M AN U

were also developed from single-eR1+ cells (Supplementary Figure 9A). These tdTomato+ organoids expressed Runx1 (Figure 7F). These results indicate that eR1+ cells in the antrum possess stem cell activity. There are multiple markers reported for stem cells in the pyloric gland, namely, Lgr5,26 CCK2R,42 and Sox2,43 in addition to eR1 described here. Elucidation of

TE D

the properties of the cells identified by these markers would be important to understand

AC C

DISCUSSION

EP

the basic properties of tissue stem cells.

In this study, we identified stem cells as eR1+ cells in the isthmus of corpus and pyloric gland in antrum. eR1 is a 270 bp-long enhancer element, within which multiple transcription factor-binding sites are present.17 These transcription factors would cooperatively regulate eR1 activity to precisely target expression in tissue stem cells. As the process of gastric carcinogenesis is still not fully understood, results reported in this paper provide a new experimental approach for better understanding of gastric cancer.

20

ACCEPTED MANUSCRIPT It is interesting to note that K-rasG12D expression in SC-ist of corpus induces apparent conversion of corpus gland to pyloric gland by eliminating parietal cells and chief cells. This conversion is similar to the process of gastric atrophy that is considered to be precancerous. Since H. pylori is known to activate MAPK pathway and activation of K-ras induces MAPK,35

RI PT

our study contributes to the understanding of the complex molecular mechanisms underlying early stage gastric carcinogenesis. Khurana et al. demonstrated that the parietal cell atrophy following H. pylori infection is characterized by induction of CD44 expression in

SC

the isthmus, which then expands towards the base of the gastric unit.44 Importantly, CD44

M AN U

was shown to promote proliferation of stomach epithelial cells upon H. pylori infection – in a p-Erk dependent manner.45 Our observation corroborates the findings of Khurana and others. Induction of metaplasia after activation of oncogenic K-ras in stomach have also been described.41, 46

TE D

Expression of K-rasG12D in SC-ist induces robust differentiation of pit (foveolar) cells but not differentiation of chief cells and parietal cells. Similar foveolar hyperplasia occurs in Menetrier disease, which is a hyperproliferative disorder of the stomach induced by over-

EP

expression of TGF-α, a ligand for EGF Receptor (EGFR).47 Furthermore, oxyntic atrophy with

AC C

antral phenotype in gastric epithelium is observed in Menetrier disease and its mouse model.48,47 We also observed the disappearance of parietal cells and chief cells and noted that the resultant structure resembles the pyloric gland. This process is often described as antralization and thus, the phenotype exhibited by our mouse model is likely to be foveolar hyperplasia accompanied by antralization. Owing to the activation of the Ras-MAPK pathway in the U-shaped structure in our study and presumably in the Menetrier disease model, there is a possibility that we may

21

ACCEPTED MANUSCRIPT be observing the same phenomenon in both models. However, there are distinct differences between the two biological systems. Unlike the Menetrier disease, our mouse model is driven by the oncogenic form of K-ras that elicit constitutively active growth promoting signals. The second difference is that in our model, K-rasG12D is expressed in gastric stem

RI PT

cells – the likely cell of origin of cancer. In contrast, stimulation by TGF-α in Menetrier disease is not confined to stem cells. Furthermore, the association of Menetrier disease with

SC

cancer is unclear,48 unlike the well-established tumorigenic effects of oncogenic Ras. Our model system and that of the Menetrier disease therefore appears to engage divergent

M AN U

molecular mechanisms. Observing the long-term effects of K-rasG12D expression in eR1+ stem cells is likely to resolve this issue and identify the earliest switch to oncogenic growth. While this study was in progress, Hayakawa et al. reported that rare cells derived from Mist1-Cre expressing cell could be found in the isthmus. They described such cells as

TE D

the origin of all epithelial lineages.28 Mist1 protein is known to be abundantly present in the chief cells and Hayakawa et al. also showed it. Although several reports indicated that a subfraction of chief cells has proliferating capacity,3, 4, 32 the results described in Hayakawa et al

EP

argued against proliferation of chief cells28. Conversely, eR1+ cells are present in a sub-

AC C

fraction of chief cells (designated as SC-bs in this study); the expression of K-rasG12D in pepsinogen-expressing SC-bs induced multiple rounds of cell division and retrodifferentiation of chief cells. Coupled with the expression of the proliferation marker Ki67 and CD44 in these eR1+ cells, this phenotype is reminiscent of SPEM. There is another significant difference in the property between eR1+ and Mist1+ cells in isthmus. Only 0.2% of Mist1+ cells in the isthmus were reported to express Ki67, indicating quiescence.33 On the other hand, 80% of eR1+ cells in the isthmus overlapped with Ki67

22

ACCEPTED MANUSCRIPT labeling, suggesting that eR1+ cells are rapidly dividing. In conclusion, our findings revealed another aspect of the complex nature of tumor formation in the stomach. The intriguing difference between our work and that of Hayakawa et al. can better be understood by

AC C

EP

TE D

M AN U

SC

RI PT

further in-depth studies of Stomach epithelial stem cells.

23

ACCEPTED MANUSCRIPT REFERENCES

7. 8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24.

RI PT

6.

SC

5.

M AN U

4.

TE D

3.

EP

2.

Barker N, Ridgway RA, van Es JH, et al. Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 2009;457:608-11. Friedmann-Morvinski D, Bushong EA, Ke E, et al. Dedifferentiation of neurons and astrocytes by oncogenes can induce gliomas in mice. Science 2012;338:1080-4. Nam KT, Lee HJ, Sousa JF, et al. Mature chief cells are cryptic progenitors for metaplasia in the stomach. Gastroenterology 2010;139:2028-2037 e9. Stange DE, Koo BK, Huch M, et al. Differentiated Troy+ chief cells act as reserve stem cells to generate all lineages of the stomach epithelium. Cell 2013;155:357-68. Asfaha S, Hayakawa Y, Muley A, et al. Krt19(+)/Lgr5(-) Cells Are Radioresistant CancerInitiating Stem Cells in the Colon and Intestine. Cell Stem Cell 2015;16:627-38. Barker N, van Es JH, Kuipers J, et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 2007;449:1003-7. Sangiorgi E, Capecchi MR. Bmi1 is expressed in vivo in intestinal stem cells. Nature Genetics 2008;40:915-920. Global Burden of Disease Cancer C, Fitzmaurice C, Dicker D, et al. The Global Burden of Cancer 2013. JAMA Oncol 2015;1:505-27. Mills JC, Shivdasani RA. Gastric epithelial stem cells. Gastroenterology 2011;140:412-24. Kim TH, Shivdasani RA. Notch signaling in stomach epithelial stem cell homeostasis. J Exp Med 2011;208:677-88. Karam SM, Leblond CP. Dynamics of epithelial cells in the corpus of the mouse stomach. The Anatomical Record 1993;236:259-273. Ito Y, Bae SC, Chuang LS. The RUNX family: developmental regulators in cancer. Nat Rev Cancer 2015;15:81-95. Chen MJ, Yokomizo T, Zeigler BM, et al. Runx1 is required for the endothelial to haematopoietic cell transition but not thereafter. Nature 2009;457:887-91. Jacob B, Osato M, Yamashita N, et al. Stem cell exhaustion due to Runx1 deficiency is prevented by Evi5 activation in leukemogenesis. Blood 2009;115:1610-1620. Okuda T, van Deursen J, Hiebert SW, et al. AML1, the target of multiple choromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 1996;84:321-330. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med 2016;374:2209-21. Ng CE, Yokomizo T, Yamashita N, et al. A Runx1 intronic enhancer marks hemogenic endothelial cells and hematopoietic stem cells. Stem Cells 2010;28:1869-81. Nottingham WT, Jarratt A, Burgess M, et al. Runx1-mediated hematopoietic stem-cell emergence is controlled by a Gata/Ets/SCL-regulated enhancer. Blood 2007;110:4188-4197. Scheitz CJ, Lee TS, McDermitt DJ, et al. Defining a tissue stem cell-driven Runx1/Stat3 signalling axis in epithelial cancer. EMBO J 2012;31:4124-39. Banerji S, Cibulskis K, Rangel-Escareno C, et al. Sequence analysis of mutations and translocations across breast cancer subtypes. Nature 2012;486:405-9. Ellis MJ, Ding L, Shen D, et al. Whole-genome analysis informs breast cancer response to aromatase inhibition. Nature 2012;486:353-60. Giannakis M, Stappenbeck TS, Mills JC, et al. Molecular properties of adult mouse gastric and intestinal epithelial progenitors in their niches. J Biol Chem 2006;281:11292-300. Osorio KM, Lilja KC, Tumbar T. Runx1 modulates adult hair follicle stem cell emergence and maintenance from distinct embryonic skin compartments. J Cell Biol 2011;193:235-50. Quante M, Marrache F, Goldenring JR, et al. TFF2 mRNA transcript expression marks a gland progenitor cell of the gastric oxyntic mucosa. Gastroenterology 2010;139:2018-2027 e2.

AC C

1.

24

ACCEPTED MANUSCRIPT

31.

32. 33.

34. 35. 36.

37. 38. 39.

40.

41.

42. 43. 44.

45.

RI PT

30.

SC

29.

M AN U

28.

TE D

27.

EP

26.

Huh WJ, Khurana SS, Geahlen JH, et al. Tamoxifen induces rapid, reversible atrophy, and metaplasia in mouse stomach. Gastroenterology 2012;142:21-24 e7. Barker N, Huch M, Kujala P, et al. Lgr5(+ve) stem cells drive self-renewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell 2010;6:25-36. Sato T, Vries RG, Snippert HJ, et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 2009;459:262-5. Hayakawa Y, Ariyama H, Stancikova J, et al. Mist1 Expressing Gastric Stem Cells Maintain the Normal and Neoplastic Gastric Epithelium and Are Supported by a Perivascular Stem Cell Niche. Cancer Cell 2015;28:800-14. Okumura T, Ericksen RE, Takaishi S, et al. K-ras mutation targeted to gastric tissue progenitor cells results in chronic inflammation, an altered microenvironment, and progression to intraepithelial neoplasia. Cancer Res 2010;70:8435-45. Ray KC, Bell KM, Yan J, et al. Epithelial Tissues Have Varying Degrees of Susceptibility to KrasG12D-Initiated Tumorigenesis in a Mouse Model. PLoS ONE 2011;6:e16786. Lee ER, Leblond CP. Dynamic histology of the antral epithelium in the mouse stomach: II. ultrastructure and renewal of isthmal cells. THE AMERICAN JOURNAL OF ANATOMY 1985;172:205-224. Watson SA, Grabowska AM, El-Zaatari M, et al. Gastrin - active participant or bystander in gastric carcinogenesis? Nat Rev Cancer 2006;6:936-46. Goldenring JR, Nam KT, Wang TC, et al. Spasmolytic polypeptide-expressing metaplasia and intestinal metaplasia: time for reevaluation of metaplasias and the origins of gastric cancer. Gastroenterology 2010;138:2207–2210. Nomura S, Baxter T, Yamaguchi H, et al. Spasmolytic polypeptide expressing metaplasia to preneoplasia in H. felis-infected mice. Gastroenterology 2004;127:582-594. Hatakeyama M. SagA of CagA in Helicobacter pylori pathogenesis. Curr Opin Microbiol 2008;11:30-7. Yang H, Rivera Z, Jube S, et al. Programmed necrosis induced by asbestos in human mesothelial cells causes high-mobility group box 1 protein release and resultant inflammation. Proc Natl Acad Sci U S A 2010;107:12611-6. Eskelinen EL. Doctor Jekyll and Mister Hyde: autophagy can promote both cell survival and cell death. Cell Death Differ 2005;12 Suppl 2:1468-72. Hwang JJ, Kim HN, Kim J, et al. Zinc(II) ion mediates tamoxifen-induced autophagy and cell death in MCF-7 breast cancer cell line. Biometals 2010;23:997-1013. Yamauchi H, Matsumaru T, Morita T, et al. The cell competition-based high-throughput screening identifies small compounds that promote the elimination of RasV12-transformed cells from epithelia. Sci Rep 2015;5:15336. Ramsey VG, Doherty JM, Chen CC, et al. The maturation of mucus-secreting gastric epithelial progenitors into digestive-enzyme secreting zymogenic cells requires Mist1. Development 2007;134:211-22. Choi E, Hendley AM, Bailey JM, et al. Expression of Activated Ras in Gastric Chief Cells of Mice Leads to the Full Spectrum of Metaplastic Lineage Transitions. Gastroenterology 2016;150:918-930 e13. Hayakawa Y, Jin G, Wang H, et al. CCK2R identifies and regulates gastric antral stem cell states and carcinogenesis. Gut 2015;64:544-53. Arnold K, Sarkar A, Yram MA, et al. Sox2(+) adult stem and progenitor cells are important for tissue regeneration and survival of mice. Cell Stem Cell 2011;9:317-29. Khurana SS, Riehl TE, Moore BD, et al. The hyaluronic acid receptor CD44 coordinates normal and metaplastic gastric epithelial progenitor cell proliferation. J Biol Chem 2013;288:16085-97. Bertaux-Skeirik N, Feng R, Schumacher MA, et al. CD44 plays a functional role in Helicobacter pylori-induced epithelial cell proliferation. PLoS Pathog 2015;11:e1004663.

AC C

25.

25

ACCEPTED MANUSCRIPT

RI PT SC M AN U TE D

48.

EP

47.

Matkar SS, Durham A, Brice A, et al. Systemic activation of K-ras rapidly induces gastric hyperplasia and metaplasia in mice. Am J Cancer Res 2011;1:432-445. Coffey RJ, Washington MK, Corless CL, et al. Menetrier disease and gastrointestinal stromal tumors: hyperproliferative disorders of the stomach. J Clin Invest 2007;117:70-80. Nomura S, Settle SH, Leys CM, et al. Evidence for Repatterning of the Gastric Fundic Epithelium Associated With Ménétrier’s Disease and TGFα Overexpression. Gastroenterology 2005;128:1292-1305.

AC C

46.

26

ACCEPTED MANUSCRIPT Figure Legends Figure 1. Runx1 expression in mouse corpus epithelial cells (A) In situ hybridization of the corpus of wild-type (WT) mice using an anti-sense probe

epithelium. A sense probe was used as the negative control (right).

RI PT

(blue) detected Runx1 in the upper (top) and lower (bottom) regions of the gastric

(B-C) Immunofluorescence (IF) staining of the stomach corpus of WT mouse. Runx1 is co-

SC

expressed with E-cadherin (B). Runx1 also co-expressed with CD45 (B, arrow). Runx1 and E-

M AN U

cadherin were co-expressed in the bottom part of gastric epithelium (C, arrow head). (D-E) IF staining of Runx1 and Muc5ac or Runx1 and GS-II in the corpus of WT mouse. Runx1+/Muc5ac+ and Runx1+/Muc5ac- cells reside in upper part of gastric units (D). A part of

TE D

Runx1 signals overlapped with GS-II (E). Scale bar = 100 µm.

Figure 2. Distribution of eR1-EGFP+ cells in the gastric corpus epithelium

EP

(A) Schematic model of the eR1-EGFP transgene. The Runx1-coding region is located on chromosome 21 and endogenous-eR1 is located between P1 and P2 promoters. The eR1-

AC C

mHSP68-EGFP transgene was inserted into a chromosome (termed Chr. A), and its activity (EGFP expression) was indicative of eR1 activity. (B) IF staining of EGFP (eR1-EGFP), E-cadherin and GS-II in the corpus of eR1-EGFP mouse. (C) IF staining of EGFP (eR1-EGFP), Ki67 and E-cadherin in the corpus of eR1-EGFP mouse. EGFP+ epithelial cells were observed in the Ki67+ region (isthmus). Panels-(a) show co-

27

ACCEPTED MANUSCRIPT localization of EGFP and Ki67 labeling (arrows), whereas panels-(b) show GFP labeling with no Ki67 labeling (arrow head). (D) IF staining of EGFP (eR1-EGFP) and E-cadherin in the corpus of eR1-EGFP mouse. EGFP+

RI PT

epithelial cells were also observed in the bottom of gastric units (arrow). (E-F) The tdTomato expression in corpus epithelium from eR1-CreERT2(5-2):Rosa-tdTomato mice with 1 day (1d) post 4 mg tamoxifen injection (p.i.). Expression of tdTomato was

M AN U

SC

observed in upper part (E) and lower part (F) of gastric epithelium. Scale bar = 100 µm.

Figure 3. Continuous differentiation of eR1+ cells in corpus epithelium. (A) tdTomato expression in corpus epithelium from eR1-CreERT2(5-2);Rosa-tdTomato mice with 1 day (1d), 3 month (3m), 6 month (6m) and 1 year (1y) post 4 mg tamoxifen injection.

TE D

Left panel in Figure 3B (1d p.i.) is adapted from Figure 2H. (B-E) IF staining of Pepsinogen C, GS-II, HK-ATPase and Muc5ac in the corpus of eR1-

EP

CreERT2(5-2);Rosa-tdTomato mice with 1 year post tamoxifen injection. The tdTomato signals were overlapped with Pepsinogen C+ (B), GS-II+ (C), HK-ATPase+ (D) and Muc5ac+ (E)

AC C

expressing regions.

(F) IF staining of Pepsinogen C in the corpus of eR1-CreERT2(5-2);Rosa-tdTomato mouse with 1 year post tamoxifen injection. The tdTomato signals were overlapped with Pepsinogen C (left panel), GS-II (middle panel) and HK-ATPase (right panel, arrow) expression. Scale bar = 100 µm.

28

ACCEPTED MANUSCRIPT Figure 4. Organoid formation from eR1+ cells. (A) Organoid generation from single eR1+ cells isolated from eR1-CreERT2(5-2);RosatdTomato mice at 1 day after injection of 2 mg of tamoxifen. Wnt3a, EGF, Noggin, and R-

RI PT

spondin with common growth factors (FGF and gastrin) were used for culture (WENR). In total, 1–3 tdTomato+ cells were observed in isolated gastric units (upper left panel). Single tdTomato+ cells at day 0 (bottom left panel) developed into tdTomato+ organoids at day 7

SC

(bottom right panel).

(B) Organoid generation rate of FACS-isolated tdTomato+ (eR1high) cells and tdTomato-

M AN U

(eR1neg) cells from eR1-CreERT2(5-2);Rosa-tdTomato mice at 1 day after tamoxifen injection. (C) Organoid formation from single eR1+ cells isolated from eR1-CreERT2(5-2);RosatdTomato mice at 1 day after injection of 2 mg of tamoxifen. EGF, Noggin, and Jagged1 with

TE D

common growth factors (FGF and gastrin) were used for culture (ENJ). (D) IF staining of paraffin section of organoids from eR1-CreERT2(5-2);Rosa-tdTomato mice.

EP

GS-II (upper panel), Muc5ac (middle panel), Pepsinogen C (lower panel). (E-F) IF staining of paraffin section of organoids from WT mouse. E-cadherin and Ki67 (E) and

AC C

E-cadherin and Runx1 (F). Scale bar = 100 µm.

Figure 5. Foveolar hyperplasia and antralization in oncogenic K-ras-expressing corpus tissue. (A) H&E staining of corpus tissue sections from eR1-CreERT2(3-17);K-rasG12D (eR1-K-rasG12D) and WT mice.

29

ACCEPTED MANUSCRIPT (B) IF staining of E-cadherin, p-Erk, and CD45 in the corpus of eR1-K-rasG12D mouse. The gland-like structures were surrounded by CD45+ cells. CD45+ cells are more easily recognizable in Supplementary Figure 6B. Enlarged image of bottom part of the lesions

RI PT

indicates isthmus shown by arrows (left panel). (C) IF staining of Ki67 and Muc5ac in the corpus of eR1-K-rasG12D mouse. The dashed-line shows a newly induced lesion.

SC

(D) CD44, HK-ATPase, and GS-II staining in the corpus of eR1-K-rasG12D mouse.

M AN U

(E) IF staining for CD44, Muc5ac, GS-II, Ki67 and Gastrin in the corpus of eR1-K-rasG12D (each panel of left) and WT mice (each panel of right).

(F) IF staining of HK-ATPase in the corpus of eR1-CreERT2(3-17);K-rasG12D;Rosa-tdTomato

TE D

mouse 2 months after tamoxifen injection. Scale bar = 100 µm.

Figure 6. Induction of metaplasia by expression of K-rasG12D in eR1+ cells in the base.

EP

(A) H&E staining of the corpus of tamoxifen treated eR1-CreERT2(3-17);K-rasG12D mouse.

AC C

(B-D) IF staining of Pepsinogen C and GS-II in the corpus of eR1-CreERT2(3-17);KrasG12D;Rosa-tdTomato mouse at 2 months after tamoxifen injection. (E-F) IF staining of Pepsinogen C with CD44 or Ki67 in the corpus of eR1-CreERT2(317);KrasG12D;Rosa-tdTomato mice. A part of tdTomato+ metaplastic lesions at the bottom of epithelium expressed Ki67 (E) and CD44 (F). Scale bar = 100 µm.

Figure 7. eR1+ cells regenerate pyloric glands in the antrum and generate organoids

30

ACCEPTED MANUSCRIPT (A) IF staining of EGFP (eR1-EGFP), Ki67 and E-cadherin in antrum of eR1-EGFP mouse. EGFP+ cells are detected in E-cadherin+ epithelial cells in pyloric glands. EGFP+ cells are mainly located in Ki67+ region (arrow). Ki67-/EGFP+ cells can also be detected in upper part of

RI PT

pyloric glands (pit). Scale bar = 100 µm. (B) The EYFP expression in pyloric gland from eR1-CreERT2(5-2);Rosa-EYFP mouse with 1 day post 2 mg tamoxifen injection. Scale bar = 100 µm.

SC

(C) tdTomato expression in pyloric gland from eR1-CreERT2(5-2):Rosa-tdTomato mice with 6 month (6m)(left panel) and 1 year (1y)(right panel) post 4 mg tamoxifen injection (p.i.). Scale

M AN U

bar = 100 µm.

(D) Isolated eR1+ pyloric glands from eR1-CreERT2(5-2);Rosa-tdTom mice with 1 day post 2 mg tamoxifen injection. Scale bar = 100 µm.

TE D

(E) Organoids generated from the pyloric glands cultured for 7 days in vitro. Scale bar = 500 µm.

(F) Wholemount IF staining of eR1+ cell-derived organoid. Runx1 and E-cadherin are

AC C

(6985 words)

EP

coexpressed in tdTomato+ organoid. Scale bar = 100 µm.

31

C

E

AC C

EP

D

TE D

M AN U

SC

B

RI PT

Figure 1. Runx1 expression in mouse corpus epithelial cells ACCEPTED MANUSCRIPT A

Figure 2. Distribution of eR1‐EGFP+ cells in the gastric corpus epithelium  ACCEPTED MANUSCRIPT A B

Chr. A

P1

P2 eR1

eR1

Runx1

mHsp 68  promoter

EGFP

D

M AN U

SC

C

RI PT

Chr. 21

E

AC C

EP

TE D

F

C

D

TE D

M AN U

SC

E

RI PT

Figure 3. Continuous differentiation of eR1+ cells in the corpus epithelium ACCEPTED MANUSCRIPT B A

AC C

EP

F

EP

TE D

F

AC C

E

D

M AN U

C

SC

RI PT

Figure 4. Organoid formation from eR1+ cells ACCEPTED MANUSCRIPT A B

RI PT

Figure 5. Foveolar hyperplasia and antralization in oncogenic K‐ras‐expressing corpus tissue. ACCEPTED MANUSCRIPT A C

D

M AN U

SC

B

F

AC C

EP

TE D

E

C

RI PT

Figure 6. Induction of metaplasia by expression of K‐rasG12D in eR1+ cells in the base ACCEPTED MANUSCRIPT A B

M AN U

SC

D

F

AC C

EP

TE D

E

RI PT

Figure 7. eR1+ cells regenerate pyloric glands in the antrum and generate organoids ACCEPTED MANUSCRIPT B A

D

TE D

M AN U

SC

C

F

AC C

EP

E

ACCEPTED MANUSCRIPT Supplemental Materials Materials and Methods Mice

RI PT

B6.129X1t(Rosa)26Sortm1(EYFP)Cos/J (Rosa-EYFP) mice, which harbor a loxP STOP sequence and enhanced yellow fluorescent protein (EYFP), B6.129S6-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze/J (Rosa-tdTomato) mice, which harbor a loxP STOP sequence and a CAG promoter-driven red

SC

fluorescent protein variant (tdTomato), and 129S/Sv-K-rastm4Tyj/J (K-rasG12D) mice, which

M AN U

harbor K-ras with the G12D point mutation, the expression of which is blocked by a loxp sequence, were purchased from The Jackson Laboratory. Lgr5-GFP-IresCreERT2 (Lgr5-GFPCreERT2) mice were described previously.1 All mice were backcrossed with C57BL/6 mice and maintained in this background. eR1-CreERT2 and Rosa-EYFP mice were crossed to

TE D

generate mice harboring both eR1-CreERT2 and Rosa-EYFP (eR1-CreERT2:Rosa-EYFP mice). eR1-CreERT2 and Rosa-tdTomato mice were crossed to generate eR1-CreERT2:RosatdTomato mice. eR1-CreERT2 and K-rasG12D mice were crossed to generate eR1-CreERT2:K-

EP

rasG12D mice. eR1-CreERT2:Rosa-tdTomato and K-rasG12D mice were crossed to generate eR1-

AC C

CreERT2:K-rasG12D:Rosa-tdTomato mice.

Induction of Damage to Corpus Tissue To induce damage in the stomach, 6 mg of tamoxifen (Sigma-Aldrich) dissolved in sunflower oil (Sigma-Aldrich) was injected into 3-4-week-old eR1-EGFP mice, which were analyzed 3 days and 1 week after injection.2

ACCEPTED MANUSCRIPT Human Clinical Tissues Primary healthy stomach tissues and clinical information were collected with patient consent in the “Gastric Cancer Biomarker Discovery II (GASCADII)” study, which was

RI PT

conducted at National University Hospital and was approved by the National Healthcare Group Domain Specific Review Board (ref. no. 2005/00440). Gastritis tissues were collected with patient consent from the National University Hospital tissue repositories or pathology

SC

archives and with the approval of the respective institutional research ethics review committees in accordance with local regulations and legislations.3 Clinical information was

Immunofluorescence (IF) Staining

M AN U

collected with the approval of the Institutional Review Board.

TE D

Cryosections (5 µm thick) were generated according to a general protocol.4 For paraffin sections, small intestine was washed with phosphate-buffered saline (PBS) containing 10% fetal calf serum and fixed in PBS containing 4% paraformaldehyde. Paraffin sections (5 µm

EP

thick) were generated from paraffin-embedded specimens according to a conventional

AC C

method. Cell morphology was determined by hematoxylin (Merck Millipore) and eosin (Merck Millipore) staining. Antibodies used for IF staining of sectioned tissues are described in the Antibodies for IF (see below). YFP expression was visualized using an anti-GFP antibody (MBL), and tdTomato expression was detected by endogenous fluorescence. Stained tissues were visualized using an Eclipse Ti (Nikon) or Axio Observer (Zeiss) microscope.

ACCEPTED MANUSCRIPT Antibodies for IF Staining Paraffin and frozen tissue sections were stained with the following primary antibodies: antiRunx1 (Abcam, 1:200), anti-E-cadherin (Abcam, 1:500), anti-E-cadherin Alexa488/555/647-

RI PT

conjugated (Invitrogen, 1:200), anti-CD45 (eBioscience, 1:500), GS-II lectin (Invitrogen, 1:2000), anti-Muc5ac (Santa Cruz, 1:200), anti-GFP (MBL, 1:1000), anti-GFP (Nacalai Tesque, 1:500), anti-Ki67 (Abcam, 1:200), anti-Ki67 (DaKo, 1:200), anti-Pepsinogen II (Abcam, 1:200),

SC

anti-HK-ATPase (MBL, 1:1000), anti-phospho-Erk (p-Erk) (Cell Signaling Technologies, 1:500), anti-CD44 (BioLegend, 1:500), anti-gastrin (DaKo, 1:200), anti-Muc2 (Santa Cruz, 1:500),

M AN U

anti-Cdx1 (Novus Biologicals, 1:200), anti-Cdx2 (Novus Biologicals, 1:200), anti-Cleavedcaspase3 (Cell Signaling Technologies, 1:200), anti-Hmgb1 (Abcam, 1:200) and anti-LC3 (Novus Biologicals, 1:200). Protein blocking and antibody dilution were used following of each reagents: 5 % skim milk (Fluka analytical), 2 % Donkey serum (Sigma-Aldrich), 3% BSA

EP

Immunohistochemistry

TE D

(Sigma-Aldrich) and Dako protein block (Dako).

AC C

Immunohistochemistry was performed as described previously.3 Anti-MUC5AC (Leica), antiMUC6 (Leica), anti-GASTRIN (DaKo), and anti-CDX2 (CELL MARQUE) antibodies were used. These antibodies were detected using a peroxidase-labeled streptavidin-biotin system (DaKo) with 3,3-diaminobenzidine (DaKo) according to the manufacturer’s instructions. Nuclei were stained with DAPI (Sigma). Stained tissues were counterstained with hematoxylin and visualized using an Olympus-IX71 microscope (Olympus).

ACCEPTED MANUSCRIPT In Situ Hybridization An anti-sense probe against 1523–2028 bp of the Runx1 sequence (NM_001111021.1) was developed following the guidance of Genostaff. Paraffin-embedded sections of mouse

RI PT

tissues were obtained from Genostaff. De-waxed tissue sections were treated with 8 µg/ml Protease K, 0.2 N HCl, 0.1 M tri-ethanolamine-HCl, 0.25% acetic anhydride, and ethanol. Hybridization was performed with 300 ng/ml probes prepared in Probe Diluent (Genostaff)

SC

at 60oC for 16 hr. Thereafter, sections were washed in HybriWash (Genostaff) and 50% formamide, treated with RNase, and washed with HybriWash and TBST (0.1% Tween 20

M AN U

prepared in Tris-buffered saline). After treatment with 1× G-Block (Genostaff), sections were incubated with an anti-digoxigenin antibody conjugated with alkaline phosphatase (Roche) diluted 1:2000 with 50× G-Block (Genostaff) prepared in TBST for 1 hr at room temperature. Sections were washed with TBST and incubated in 100 mM NaCl, 50 mM MgCl2, 0.1% Tween

TE D

20, 100 mM Tris-HCl (pH 9.5). Coloring reactions were performed overnight with NBT/BCIP solution (Sigma-Aldrich). Sections were counterstained with Kernechtrot stain solution

AC C

Alcian Blue Staining

EP

(Mutoh).

De-waxed tissue sections were soak in 3% acetic acid for 5 min and stained with 1% Alcian blue prepared in 3% acetic acid (pH 2.5) for 2 hr. Thereafter, tissue sections were washed with running water for 10 min and then with distilled water for 5 min. Nuclear Fast Red was used to stain nuclei.

ACCEPTED MANUSCRIPT Gastric gland isolation and organoid culture eR1-CreERT2:Rosa-tdTom mice were injected with 2 mg of tamoxifen. At 24 h after injection, mice were euthanized and the stomach was harvested. The stomach was opened up

RI PT

longitudinally along the greater curvature and washed with cold PBS. The opened stomach pinned on a silicon board was placed in a dish filled with ice-cold PBS. The fundic mucosa was scraped off using a surgical scalpel (blade no. 10). The tissue was chopped into

SC

approximately 5 mm2 pieces and incubated for 2 h on ice in Dulbecco’s PBS (Thermo Fisher Scientific) containing 5 mM EDTA. The gastric units were released by vigorous shaking in

M AN U

cold dissociation buffer, which consisted of 54.8 mM D-sorbitol and 44 mM sucrose prepared in Dulbecco’s PBS.5 Gland fractions were centrifuged at 150 g for 5 min at 4°C and washed with DPBS, and then the supernatant was removed. They were then washed again with 5 mL of advanced Dulbecco’s Modified Eagle’s Medium (DMEM)/F12 and centrifuged

TE D

at 150 g for 5 min at 4°C. The supernatant was then removed. For single-cell isolation, isolated glands were treated with TripLE Express (LIFE Technology) supplemented with 5 mg/mL DNAse I (Sigma-Aldrich) and 10 μM Y27632 (Sigma-Aldrich) for 10 min at 37°C. The

EP

cells were resuspended in Minigut medium (advanced DMEM/F12 (Thermo Fisher Scientific)

AC C

supplemented with 10 mM HEPES (Thermo Fisher Scientific), GlutaMAX (Thermo Fisher Scientific), penicillin/streptomycin (Thermo Fisher Scientific), N2 (Thermo Fisher Scientific), B27 (Thermo Fisher Scientific), 1 mM N-acetylcysteine (Sigma-Aldrich), and 1% bovine serum albumin), passed through a 40 μm cell strainer (BD Bioscience), and then washed with DMEM/F12.

Gastric gland and single-cell organoid culture

ACCEPTED MANUSCRIPT After counting the number of isolated glands or single cells, 200 glands/well or 500 cells/well were embedded in Matrigel (BD Bioscience) containing 100 ng/mL Noggin (R&D Systems), 50 ng/mL epidermal growth factor (EGF) (Sigma-Aldrich), 100 ng/mL fibroblast growth factor (FGF) 10 (Peprotech), 10 nM gastrin I (Sigma-Aldrich), 10 μM Y27632, and 1

RI PT

mM N-acetylcysteine with 100 ng/mL R-spondin 1 (R&D Systems) and 100 ng/mL Wnt3a (R&D Systems) (WENR) or with 1 μM Jagged 1 (R&D Systems) (ENJ). After polymerization of

SC

Matrigel, Minigut medium was overlaid. Medium containing growth factors was changed every 3 days. Images of gastric organoids were acquired using a fluorescence microscope

M AN U

(Nikon, ECLIPSE TS100, INTENSILIGHT C-HGFI).

FACS of eR1+ cells

TE D

Freshly isolated single gastric epithelial cells were sorted into eR1-negative (eR1neg) and eR1highly expressing (eR1high) populations. Viable epithelial cells were gated by forward scatter, side scatter, and SYTOX blue (Thermo Fisher Scientific) negative staining. Sorted cells were

EP

collected and counted. In total, 100–500 cells were embedded in Matrigel containing the WENR growth factors, followed by seeding in a 24-well plate. After polymerization of

AC C

Matrigel, Minigut medium was overlaid.

Processing for paraffin-embedded section of organoids Preparation of paraffin-embedded section was performed using previously descryibed procedure as a reference.6 Briefly, organoids were cultured in 8 well detachable chamber slides. Before harvesting the organoids, cryomold (Intermediate mold 15 mm x 15 mm x

ACCEPTED MANUSCRIPT 5mm, Tissue-Tek, Sakura Finet) were coated with 300 μl of warmed Hisotgel (Thermo Sientific). The medium was removed and the plastic chamber was detached from the glass slide. Matrigel block containing organoids were scraped off the slide with a clean razor blade and transferred onto the Histgel pre-coated cryomold. Another 300 μl of warm Histogel was

RI PT

added on top of the Matrigel. The mold was immediately transferred to ice and allowed to solidify for 10 minutes. Matrigel-Histgel complex were them out into processing pathology

SC

cassettes (Histosette I, Simport) and fixed in 4% paraformaldehyde. Following steps were

Whole mount staining of organoids

M AN U

the same as sample processing for IF staining of mice tissue.

Whole mount staining of organoids was performed as described previously.5 Briefly,

TE D

organoids were fixed with 4% paraformaldehyde for 30 min followed by treatment with 50 mM NH4Cl for 30 min. Permeabilization was performed with 0.5% Triton X-100 for 30 min, and then blocking was performed with 5% bovine serum albumin for 1 h. Thereafter,

EP

organoids were incubated with primary antibodies overnight at 4°C and then with the secondary antibody overnight at 4°C. The primary and secondary antibodies were used at

AC C

the same concentration as described in Antibodies for IF staining (see above). Images were acquired with a confocal microscope.

Supplementary Figure Legends Supplementary Figure 1. Ki67-expressing cells in corpus epithelium and Runx1 expression in the antrum

ACCEPTED MANUSCRIPT (A) Illustration for anatomical location of various parts of corpus gastric unit (GU). (B) IF staining of Runx1 and Ki67 in the corpus of WT mouse. Most of Ki67+ cells also expressed Runx1. Middle and right panels are higher magnification of the boxed areas in the

RI PT

left panel. (C-F) Corpus of gastric epithelium of WT mouse was stained with Ki67 together with Muc5ac, GS-II, HK-ATPase or Pepsinogen C. Muc5ac+ cells located at lower part (C) and those located

SC

at upper part of GS-II+ cells (D) overlap with Ki67 signals. Co-expression of Ki67 and HKATPase were rarely detected (E). A minor population of Pepsinogen C+ cells expresses Ki67 at

M AN U

the bottom of the gastric unit (F). Right panels are higher magnification of the boxed areas in the left panel.

(G) Fractions of Ki67+ cells that express Runx1, Muc5ac, GS-II, HK-ATPase, and Pepsinogen C

TE D

within the gastric unit are shown. Over 500 cells from four separate specimens were counted per marker.

(H) IF staining of Runx1 and E-cadherin of the antrum of WT mice detects co-expression of

EP

Runx1 and E-cadherin the pyloric gland. Stronger staining of Runx1 is observed at lower part

panel.

AC C

of the gland. Middle and right panels show higher magnification boxed regions in the left

(I) IF staining of Runx1 and GFP (Lgr5-GFP) in the antrum of mouse conditionally knocked out Runx1 (Runx1cKO) by Lgr5-GFP-CreERT2 (Ctr). Control (Ctr) mouse shows Runx1 expression in the Lgr5-GFP+ pyloric gland, which is enclosed by the dashed line (left), whereas Runx1cKO mouse treated with tamoxifen did not (right). Scale bar = 100 µm.

ACCEPTED MANUSCRIPT Supplementary Figure 2. eR1-GFP+ cells among corpus epithelial cells (A) Number of EGFP+ or EGFP- gastric units in corpus epithelium. More than 100 of gastric units were counted.

RI PT

(B) IF staining of EGFP (eR1-EGFP) and CD45 in the corpus of eR1-EGFP mouse. EGFP and CD45 staining do not overlap.

(C) Distribution of eR1-EGFP+ cells in the corpus gastric epithelium. The GS-II+ region was

SC

defined as the neck, and the upper and lower regions were defined as the isthmus/pit and

M AN U

base, respectively. More than 700 EGFP+/E-cadherin+ cells in four eR1-EGFP mice were counted.

(D-E) Two examples of IF staining of EGFP (eR1-EGFP), Ki67 and E-cadherin in the corpus of eR1-EGFP mouse. Panels show co-localization of EGFP and Ki67 labeling (arrows).

TE D

(F) Distribution of Ki67+/eR1-EGFP+ cells in the corpus gastric unit. The Ki67+ region was defined as the isthmus, and the above and below regions are defined as the pit and

EP

neck/base, respectively (See also Figure 2C). 400 EGFP+ and E-cadherin+ cells in three eR1EGFP mice were counted.

AC C

(G) IF staining of EGFP (eR1-EGFP) and Runx1 in the corpus of eR1-EGFP mouse. The eR1EGFP and Runx1 are co-expressed in the gastric epithelium (arrow). The eR1-EGFP is also observed in Runx1- cells (arrow head). The middle panels are higher magnification images of the boxed areas in the left panel. (H) IF staining of EGFP (eR1-EGFP) and Pepsinogen C in the corpus of eR1-EGFP mouse. EGFP+ epithelial cells were also observed in the bottom of gastric units, and they expressed pepsinogen C (arrow). Scale bar = 100 µm.

ACCEPTED MANUSCRIPT

Supplementary Figure 3. Tissue damage response of eR1-GFP+ cells (A) IF staining of HK-ATPase and GS-II to demonstrate tamoxifen-induced damage of the

RI PT

corpus in eR1-EGFP mice. The number of HK-ATPase+ parietal cells decreased and GS-II+ cells were detected near the bottom of gastric units at 3 days and 1 week after tamoxifen treatment (6 mg) (middle and right panels).

SC

(B-E) IF staining of EGFP (eR1-EGFP) and E-cadherin in the corpus of eR1-EGFP mice injected

M AN U

with 6 mg of tamoxifen. Compared with the number of EGFP+ cells 3 days after injection (3d) (B), the number of EGFP+ cells were markedly increased 7 days after injection (7d) (C) in the upper parts of the gastric units. Under the same conditions, similar increase of EGFP+ cells was observed at the bottom part of the gland unit (D and E). Right panels show higher

TE D

magnification images of the boxed areas in the left panels.

(F) Population of eR1-EGFP+ cells per gastric unit. The number of EGFP+ cells per gastric unit was counted. In total, more than 100 gastric units were examined in each 3 or 7 days

AC C

EP

tamoxifen-treated (TMX+) and non-treated (TMX-) mouse. Scale bar = 100 µm.

Supplementary Figure 4. Lineage tracing induced by low-toxic level of tamoxifen (A) Outline of lineage tracing experiments using Transgenic (Tg) mouse line 5-2, eR1CreERT2(5-2);Rosa-tdTomato. 2 or 4 mg of Tamoxifen was injected into 6-week-old mice. Mice were analyzed at indicated time post injection. (B-C) IF staining of E-cadherin in the corpus of eR1-CreERT2(5-2):Rosa-tdTomato mice with 1 day and 6 month post tamoxifen injection (2 mg). A few single tdTomato+ cells are observed

ACCEPTED MANUSCRIPT in upper part of corpus gastric epithelium (B, arrow), and tdTomato signals were observed many gastric unit (C). Scale bar = 100 µm.

RI PT

Supplementary Figure 5. Organoid formation from eR1+ cells (A) The eR1+/tdTomato+ cells from eR1-CreERT2(5-2);Rosa-LSL-tdTomato mice, 1 day post tamoxifen injection, were isolated by FACS for the analysis of organoid generation rate. Top

SC

1% of EGFP expressing cells were sorted.

M AN U

(B) Rate of organoid generation in WENR and ENJ culture conditions. A total of 500 cells from gastric units were plated, and tdTomato+ organoids were counted at day 7 of culture. The organoid generation rate did not statistically differ between the WENR and ENJ culture

TE D

conditions. *P = 0.34.

Supplementary Figure 6. K-rasG12D-induced foveolar hyperplasia and antralization in corpus

EP

epithelium

(A) Outline of lineage tracing experiments using eR1-CreERT2(3-17);K-rasG12D;Rosa-tdTomato

AC C

mouse. 4 or 6 mg of Tamoxifen was intraperitoneally injected into 3-week-old mice. Mice were analyzed at 2 months post injection. (B-C) IF staining of CD45 and E-cadherin of the stomach epithelium of eR1-CreERT2(3-17):KrasG12D (eR1-KrasG12D) mouse. CD45+ cells are attracted near the bottom of the metaplastic lesion (B, bracket).

ACCEPTED MANUSCRIPT (D-E) IF staining of phosphorylated Erk (p-Erk), Ki67, Muc5ac and GS-II of the corpus of eR1KrasG12D mice. The K-rasG12D-induced metaplasia are p-Erk+ and Ki67+ (D). The metaplasia express Muc5ac+ and the bottom is stained by GS-II (E).

RI PT

(F) Alcian blue staining of the antrum of WT mouse (left) and the corpus of eR1-KrasG12D mouse (right). Nuclei were stained with Nuclear Fast Red. Near the bottom of the metaplastic lesions are stained by Alcian blue.

SC

(G-I) IF staining of Muc2, Cdx1 and Cdx2 in the corpus of eR1-K-rasG12D mouse. Intestinal goblet cell marker Muc2 were detected in intestine of WT mice (G, left panel) and the

M AN U

corpus of eR1-KrasG12D mouse (G, right panel). Intestinal epithelial cell marker Cdx1 and Cdx2 (H and I, left panel) were not detected in the corpus of eR1-Kras mouse (H and I, right panel).

TE D

(J) IF staining of Pepsinogen C and GS-II in the corpus of eR1-CreERT2(3-17);K-rasG12D;RosatdTomato mouse 2 months after tamoxifen injection. Scale bar = 100 µm.

EP

Supplementary Figure 7. Further characterisation of K-rasG12D induced metaplasia

AC C

(A-B) IF staining of HK-ATPase with cleaved-caspase3 in eR1-K-rasG12D mouse. Cleavedcaspase3 was detected in the parietal cells present within the KrasG12D induced metaplastic lesion (A and B, arrow).

(C-D) IF staining of HK-ATPase together with Hmgb1 and CD44 in eR1-K-rasG12D mouse. Expression of Hmgb1 in parietal cells were localized in nucleus. Necrotic parietal cells which have cytoplasmic Hmgb1 expression are rarely detected (C). Some parietal cells which

ACCEPTED MANUSCRIPT appear to be expelled from CD44+ metaplastic lesion are observed (D, also in C). Right panels are higher magnification of the boxed areas in the left panel. (E) IF staining of LC3 and CD44 in eR1-K-rasG12D mouse. Right panels are higher magnification

RI PT

images of the boxed areas in the left panel. (F-G) IF staining of Runx1, CD44 and Ki67 in the corpus of eR1-KrasG12D mouse. Runx1 was expressed in K-rasG12D-induced lesion especially strongly near the bottom where Ki67 is also

SC

expressed (F). CD44 is expressed highly at the isthmus of metaplastic legion where Runx1 is

M AN U

also highly expressed (G).

Supplementary Figure 8. Characterisation of H. pylori-associated metaplasia in human stomach

TE D

(A-B) IF staining of RUNX1 and E-cadherin in healthy human corpus and antrum. RUNX1 was detected in E-cadherin+ epithelial cells in the corpus (A) and the antrum (B). Right panels

EP

show higher magnification the boxed areas in the left panel. (C) H&E staining of the healthy human antrum and the corpus of a patient with H. pylori-

AC C

associated gastritis (H. pylori gastritis). Higher magnification of the boxed areas are shown in the extreme left panels in (D). (D) H&E staining and immunohistochemical staining for MUC5AC, MUC6, GASTRIN, and CDX2 in the healthy human antrum and the corpus of a patient with H. pylori-associated gastritis. Similar to healthy antrum, MUC6 was detected at the bottom of gastritis lesions and MUC5aC was detected above MUC6. Unlike healthy antrum, GASTRIN was not observed in gastritis lesions. The arrows indicate the lesions exhibited CDX2+ intestinal metaplasia,

ACCEPTED MANUSCRIPT whereas CDX2 was not detected in healthy antrum. The dashed line shows the bottom of the epithelium. Scale bar = 100 µm.

RI PT

Supplementary Figure 9. Organoids generated from eR1+ cells in antrum (A) Tamoxifen (2 mg) was injected to eR1-CreERT2 (5-2) :Rosa-tdTomato mice for one day

M AN U

7 with the culture conditions of WENR are shown .

SC

and tdTomato+ were isolated. eR1+ cell derived organoids generated in vitro on day 1, 3 and

Supplementary References

1. Barker, N., van Es, J.H., Kuipers, J., Kujala, P., van den Born, M., Cozijnsen, M., Haegebarth, A., Korving, J., Begthel, H., Peters, P.J., et al. (2007). Identification of stem cells in small

TE D

intestine and colon by marker gene Lgr5. Nature 449, 1003-1007. 2. Huh, W.J., Khurana, S.S., Geahlen, J.H., Kohli, K., Waller, R.A., and Mills, J.C. (2012).

EP

Tamoxifen induces rapid, reversible atrophy, and metaplasia in mouse stomach.

AC C

Gastroenterology 142, 21-24 e27.

3. Srivastava, S., Liew, M.S., McKeon, F., Xian, W., Yeoh, K.G., Ho, K.Y., and Teh, M. (2014). Immunohistochemical analysis of metaplastic non-goblet columnar lined oesophagus shows phenotypic similarities to Barrett's oesophagus: a study in an Asian population. Dig. Liver Dis. 46, 170-175.

ACCEPTED MANUSCRIPT 4. Wakamatsu, Y., Watanabe, Y., Shimono, A., and Kondoh, H. (1993). Transition of Localization of the N-Myc Protein from Nucleus to Cytoplasm in Differentiating Neurons. Neuron 10, 1-9.

RI PT

5. Mahe, M.M., Aihara, E., Schumacher, M.A., Zavros, Y., Montrose, M.H., Helmrath, M.A., Sato, T., and Shroyer, N.F. (2013). Establishment of Gastrointestinal Epithelial Organoids. Curr. Protoc. Mouse Biol. 3, 217-240.

SC

6. Pinto, M.P., Jacobsen, B.M., and Horwitz, K.B. (2011). An immunohistochemical method to study breast cancer cell subpopulations and their growth regulation by hormones in

AC C

EP

TE D

M AN U

three-dimensional cultures. Front. Endocrinol. (Lausanne) 2, 15.

Supplementary Figure 1. Ki67‐expressing cells in corpus epithelium and Runx1 expression  ACCEPTED MANUSCRIPT in the antrum

E

RI PT

D

C

F

SC

B

G

TE D

M AN U

A

Percent of Ki67 + Cells Positive

100%

EP

80% 60%

20% 0% Runx1

AC C

40%

I

Negative

Muc5ac

GS‐II

HK‐ATPase Pepsinogen C

H

Supplementary Figure 2. eR1‐GFP+ cells among corpus epithelial cells ACCEPTED MANUSCRIPT A

C

B

60 50 40 30 20

RI PT

Percent of Gastric Units

70

10 0

E

F

TE D

M AN U

D

EGFP‐

SC

EGFP+

AC C

EP

G

H

C

D

TE D

M AN U

SC

E

RI PT

Supplementary Figure 3. Tissue damage response of eR1‐GFP+ cells ACCEPTED MANUSCRIPT A B

AC C

EP

F

Supplementary Figure 4. Lineage tracing induced by  low‐toxic level of tamoxifen ACCEPTED MANUSCRIPT C A B mHsp 68  promoter

Rosa 26

loxP

1 day, 3 month, 6 month  or 1 year p.i. (Analyze)

EP

TE D

M AN U

SC

6 weeks old (TMX 2 or 4mg) 

AC C

DOB

tdTomato

Stop loxP

CreERT2

RI PT

eR1

Supplementary Figure 5. Organoid formation from eR1+ cells ACCEPTED MANUSCRIPT B

AC C

EP

TE D

M AN U

SC

RI PT

A

Supplementary Figure 6. K‐rasG12D‐induced foveolar hyperplasia and antralization in corpus  ACCEPTED MANUSCRIPT epithelium C B A eR1‐CreERT2; LSL‐K‐rasG12D; Rosa‐LSL‐tdTomato mHsp 68  promoter

Stop Rosa 26

CreERT2

K‐rasG12D tdTomato

Stop loxP

loxP

E

G

H

F

J

AC C

EP

TE D

M AN U

SC

D

RI PT

eR1

I

D

F

G

E

AC C

EP

TE D

M AN U

SC

C

RI PT

Supplementary Figure 7. Further characterisation of K‐rasG12D induced metaplasia ACCEPTED MANUSCRIPT B A

Supplementary Figure 8. Characterisation of H. pylori‐associated metaplasia in human  ACCEPTED MANUSCRIPT stomach  A

C

SC

RI PT

B

AC C

EP

TE D

M AN U

D

Supplementary Figure 9. Organoids generated from eR1+ cells in antrum ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

A