Embed Size (px)
Citation preview
Accepted Manuscript
BMI1 and MEL18 Promote Colitis-Associated Cancer in Mice via REG3B andSTAT3
Xicheng Liu, Wendi Wei, Xiaowei Li, Pengcheng Shen, Dapeng Ju, Zhen Wang,Rukui Zhang, Fu Yang, Chunyan Chen, Kun Cao, Guoli Zhu, Hongyan Chen, LiangChen, Jianhua Sui, Erquan Zhang, Kaichun Wu, Fengchao Wang, Liping Zhao,Rongwen Xi
PII: S0016-5085(17)35976-0DOI: 10.1053/j.gastro.2017.07.044Reference: YGAST 61336
To appear in: GastroenterologyAccepted Date: 27 July 2017
Please cite this article as: Liu X, Wei W, Li X, Shen P, Ju D, Wang Z, Zhang R, Yang F, Chen C, CaoK, Zhu G, Chen H, Chen L, Sui J, Zhang E, Wu K, Wang F, Zhao L, Xi R, BMI1 and MEL18 PromoteColitis-Associated Cancer in Mice via REG3B and STAT3, Gastroenterology (2017), doi: 10.1053/j.gastro.2017.07.044.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
1
BMI1 and MEL18 Promote Colitis-Associated Cancer in Mice via
REG3B and STAT3
Short Title: PcG-REG3B-STAT3 signaling in CAC
Xicheng Liu1, Wendi Wei1, Xiaowei Li1,2, Pengcheng Shen1, Dapeng Ju1, Zhen Wang1,
Rukui Zhang1, Fu Yang1, Chunyan Chen1, Kun Cao1, Guoli Zhu1, Hongyan Chen1,
Liang Chen1, Jianhua Sui1, Erquan Zhang1, Kaichun Wu2, Fengchao Wang1, Liping
Zhao1, and Rongwen Xi1,3*
1 National Institute of Biological Sciences, No. 7 Science Park Road, Zhongguancun
Life Science Park, Beijing 102206, China
2 State Key Laboratory of Cancer Biology and Xijing Hospital of Digestive Diseases,
Fourth Military Medical University, Xi'an 710032, China
3Shanghai 10th People's Hospital, School of Life Science and Technology, Tongji
University, Shanghai 200072, China
Grant Support: This work was supported by the National Basic Research Program of
China (2014CB850002 to R.X.), National Key Research and Development Program
of China (2017YFA0103602 to R.X), National Basic Research Program of China
(2011CB812700 to R.X.), and a National Youth Fund (31601059 to F.Y.) from
National Natural Science Foundation of China.
Abbreviations: BMI1, BMI1 proto-oncogene, polycomb ring finger; MEL18,
polycomb group ring finger 2; AOM, azoxymethane; CAC, colitis-associated cancer;
ChIP, chromatin immunoprecipitation; Co-IP, co-immunoprecipitation; DAPI,
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
2
4',6-diamidino- Microarray Analysis 2-phenylindole; DSS, dextran sodium sulfate;
Gapdh, glyceraldehyde-3-phosphate dehydrogenase; H&E, hematoxylin and eosin;
IEC, intestinal Epithelial Cell; ISC, intestinal stem cell; JAK, Janus Kinase; PAP,
pancreatitis-associated proteins; PcG, polycomb group; Q-PCR, quantitative
polymerase chain reaction; SPF, specific pathogen-free; WT, wild-type.
*Corresponding Author: Rongwen Xi, National Institute of Biological Sciences,
Beijing, China. E-mail：[email protected]; Phone：86- 10-80723241.
Disclosures: The authors declare no potential conflicts of interest.
Transcript Profiling: The GEO accession numbers for the gene expression profiling
data reported in this paper are GSE57640, GSE57641 and GSE57642.
Access link:
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=krkdwkayrhidhep&acc=GSE57640
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=krkdwkayrhidhep&acc=GSE57641
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=krkdwkayrhidhep&acc=GSE57642
Author Contributions: X.L. and R.X. conceived of the experiments. X.L., W.W.,
X.L., P.S., Z.W., D.J., R.Z., F.Y., C.C., K.C., H.C., and G.Z. performed experiments.
J.S. produced recombinant human REG3β. L.C., E.Z., F.W., K.W., and L.Z. provided
mice and technical support. X.L. and R.X. wrote the manuscript. R.X. provided
funding.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
3
Short Summary: This study documents that BMI1 and MEL18 are required for
colitis-associated cancer (CAC) development via regulating a novel REG3B-STAT3
signaling pathway. The findings may have wide implications for the prevention and
treatment of CAC.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
4
Abstract:
Background & Aims: Polycomb group proteins are epigenetic factors that silence
gene expression; they are dysregulated in cancer cells and contribute to
carcinogenesis by unclear mechanisms. We investigated whether BMI1
proto-oncogene, polycomb ring finger (BMI1) and polycomb group ring finger 2
(PCGF2, also called MEL18) are involved in initiation and progression of
colitis-associated cancer (CAC) in mice.
Methods: We generated mice containing floxed alleles of Bmi1 and/or Mel18 and/or
Reg3b using the villin-Cre promoter (called Bmi1�IEC, Mel18�IEC, DKO, and TKO
mice). We also disrupted Bmi1 and/or Mel18 specifically in intestinal epithelial cells
(IECs) using the villin-CreERT2 inducible promoter. CAC was induced in cre-negative
littermate mice (control) and mice with conditional disruption of Bmi1 and/or Mel18
by intraperitoneal injection of azoxymethane followed by addition of dextran sulfate
sodium (DSS) to drinking water. Colon tissues were collected from mice and analyzed
by histology and immunoblots; IECs were isolated and used in cDNA microarray
analyses.
Results: Following administration of azoxymethane and DSS, DKO mice developed
significantly fewer polyps than control, Bmi1�IEC, Mel18�IEC, Reg3b�IEC, or TKO mice.
Adenomas in the colons of DKO mice were low-grade dysplasias whereas adenomas
in control, Bmi1�IEC, Mel18�IEC, Reg3b�IEC, or TKO mice were high-grade dysplasias
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
5
with aggressive invasion of the muscularis mucosa. Disruption of Bmi1 and Mel18
(DKO mice) during late stages of carcinogenesis significantly reduced the numbers of
large adenomas and the load of total adenomas, reduced proliferation, and increased
apoptosis in colon tissues. IECs isolated from DKO mice after azoxymethane and
DSS administration had increased expression of Reg3b, compared with control,
Bmi1�IEC, or Mel18�IEC mice. Expression of REG3B was sufficient to inhibit
cytokine-induced activation of STAT3 in IECs. The human REG3β protein, the
functional counterpart of mouse REG3B, inhibited STAT3 activity in human 293T
cells, and its expression level in colorectal tumors correlated inversely with pSTAT3
level and survival times of patients.
Conclusions: BMI1 and MEL18 contribute to development of CAC in mice by
promoting proliferation and reducing apoptosis, via suppressing expression of Reg3b.
REG3B negatively regulates cytokine-induced activation of STAT3 in colon epithelial
cells. This pathway might be targeted in patients with colitis to reduce carcinogenesis.
KEY WORDS: PcG, colon cancer, ulcerative colitis, PAP.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
6
Introduction
Colitis-associated cancer (CAC) is a subtype of colorectal cancer that is commonly
preceded by clinically detectable inflammatory bowel disease, such as Crohn's disease
or Ulcerative colitis. CAC is difficult to treat, and has high mortality.1-3 The mouse
model of CAC has provided important insights into inflammation-associated
tumorigenesis. Inflammation is associated with many types of cancers and may play a
causative role in tumor initiation and metastasis.3 Several major signaling pathways
through which inflammation promotes tumor incidence and size have been identified.
IKK β-dependent activation of NF-kappaB in premalignant epithelial cells prevents
apoptosis to facilitate tumor formation.4 Two key cytokines are IL-6 and IL-11, which
activate their respective receptors, followed by phosphorylation and activation of
STAT3 in malignant cells to promote their proliferation and survival.5-9 Activation of
the paracrine or autocrine STAT3 regulatory loop appears to be a major driver in
various types of cancers. In pancreatic ductal adenocarcinoma and lung
adenocarcinoma, cytokines activate STAT3 in an autocrine manner to promote
proliferation and survival of tumor cells.10-12 Therefore, STAT3 may serve as a
signaling node that connects autonomous, proto-oncogenic stimuli with
environmental inflammatory signals to initiate inflammation-associated
tumorigenesis.
Polycomb group (PcG) proteins are frequently unregulated in various cancers and are
considered proto-oncogenes.13-15 It is therefore important to understand the molecular
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
7
mechanisms underlying this oncogenic function. As a core component of the PRC1
complex, BMI1 is one of the best-studied mammalian PcG proteins. BMI1 was first
identified as a proto-oncogene in the initiation of lymphoma in collaboration with
Myc.16, 17 BMI1 exerts its oncogenic function by preventing Myc-induced apoptosis,
largely due to the repressive effect of BMI1 on the Cdkn2a locus, which contains two
tumor suppressor genes, p16Ink4a and p19ARF.18, 19 However, in some cases, BMI1
appears to have effects beyond silencing of the Cdkn2a locus. For example, deletion
of p16Ink4a and p19ARF only partially rescues neural stem cell maintenance in
Bmi1-deficient mice,20 BMI1 regulates the DNA damage response (DDR) pathway
independently of the Ink4a/Arf pathway.21 Moreover, the requirement of BMI1 in the
development of glioma is independent of Ink4a/Arf.22, 23 These observations suggest
that there are multiple pathways downstream of BMI1 that promote stem cell
self-renewal and tumorigenesis. MEL18, a homologue of BMI1, has been less well
studied. Knock-out studies indicate that Bmi1 and Mel18 have synergistic roles in Hox
gene regulation and skeletal patterning.24 Biochemically, both BMI1 and MEL18 are
able to effectively stimulate E3 ubiquitination ligase activity of recombinant Ring1B.
In addition, BMI1 and MEL18 constitute mutually exclusive PRC1 complexes, but
these complexes all accumulate at H3K27me3-rich regions.25-27 These observations
indicate that these two genes have a similar biological and biochemical function and
might be functionally redundant in tissues in which both genes are abundantly
expressed.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
8
In this study, we investigated the potential role of Bmi1 and Mel18 in the homeostasis
of colonic epithelium and colitis-associated CAC. BMI1 is expressed in intestinal
stem cells (ISCs) in the crypt of the small intestine,28 and is required for ISC
self-renewal and epithelial homeostasis in the small intestine.29, 30 However, the
contribution of BMI1 to CAC, particularly given its potentially redundant function
with MEL18, has not been defined. Here, we generated Bmi1 and Mel18 conditional
KO mice and specifically depleted their function in the intestinal epithelium.
Depletion of Bmi1 or Mel18 alone or simultaneously does not significantly affect
colonic homeostasis. However, deletion of both genes, but not alone, was able to
significantly inhibit tumor initiation and development in the experimental model of
CAC, indicating that BMI1 and MEL18 have redundant roles in colitis-associated
tumorigenesis. We determined that BMI1 or MEL18 is required for STAT3 activation
in premalignant cells in response to STAT3-activating cytokines released from
inflammatory cells and identified REG3B as a novel PcG target whose up-regulation
inhibits cytokine signaling and, consequently, tumorigenesis.
Materials and Methods
Experimental Animals
All animal experiments were approved by the Institutional Animal Care and Use
Committee at National Institute of Biological Sciences (NIBS) in accordance with the
China's Ministry of Health national guidelines for housing and care of laboratory
animals. No statistical method was used to predetermine sample size. Animal numbers
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
9
were determined to optimize minimum numbers necessary for statistical significance
according to the previous literature.31 All mice were bred and maintained under
specific pathogen-free (SPF) conditions at the animal facility of the National Institute
of Biological Sciences, and all experiments were performed using six- to
eight-week-old sex-matched mice. C57BL/6 mice were from Vital River Laboratory
Animal Technology Co. and all mice used in the experiments were backcrossed to
C57BL/6 mice for at least 6 generations. Cre-negative littermate mice were used as
wild-type (WT) controls. For generation and validation of conditional knockout Bmi1,
Mel18 and Reg3b alleles, see Supplementary Materials and Methods.
The following mouse alleles were used: villin-Cre 32 (Jackson Laboratory, Bar Harbor,
Maine), villin-CreERT2 33 (gift of Dr. Sylvie Robine, Paris, France) and Cdkn2aF/F 34
(Dr. Liang Chen, Beijing, China). For villin-CreERT2 induction, mice were injected
intraperitoneally with tamoxifen (Sigma, #T5648) in sunflower oil at a concentration
of 2 mg per 20 g body weight for 5 consecutive days.
Colitis-Associated Colon Cancer Induction
The CAC model was performed essentially as described previously.4 Briefly, six- to
eight-week-old sex-matched mice were intraperitoneally injected with 10 mg/kg of
azoxymethane (AOM; Sigma, #A5486). After 5 days, the mice were treated with
2.5% dextran sulfate sodium (DSS; MP Biomedicals, molecular weight 35–50 kDa,
#0216011080) in drinking water for 5 days, which was then followed by 16 days of
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
10
regular water. This cycle was repeated twice. On day 80 or 100, mice were
intraperitoneally injected with 100 mg/kg 5-bromo-2-deoxyuridine (BrdU; Sigma,
#B9285) and sacrificed 3 hr. later. In all experiments, littermate controls were used to
enable comparison with mice of the same genetic background. The animals that
exhibited health concerns not related to the study conditions were excluded from the
analysis. The animals were blindly monitored daily by weighing and clinical scoring
during AOM-DSS treatment. Polyp load was identified as a sum of the diameters of
all tumors in a given mouse.35
.
Isolation of Intestinal Epithelial Cells (IECs)
IECs were isolated from the freshly dissected colon as described previously.4, 36 After
removal of the Payer’s patches and the adventitial fat, the colons were cut open
longitudinally and washed with PBS. Colons were cut into 2-3 mm pieces and
incubated in HBSS containing 5mM EDTA at 37°C, shaking for 20 min. The
supernatant was collected and centrifuged. The pellet was washed in ice-cold PBS and
maintained in RNA stabilization solution or snap frozen in liquid nitrogen for western
blot analysis.
Statistical Analysis
Data were presented as mean ± SEM. Graphical analyses, statistical analysis, and
nonlinear regression analysis of the data were performed using GraphPad Prism
Software. Differences between groups were determined using one-way ANOVA,
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
11
two-way ANOVA, or unpaired Student’s two-tailed t-test, or Fisher’s exact test.
Kaplan–Meier method and log-rank test were used for survival analyses. A P value
that was less than 0.05 was indicated statistical significance for all data sets.
Results
BMI1 and MEL18 Are Required for the Initiation and Progression of CAC
Tumorigenesis
A germline knock-out of both the Bmi1 and Mel18 alleles was generated and
described previously. Analysis of these doubly deficient mice has revealed that BMI1
and MEL18 act in synergy and in a dose-dependent manner to repress Hox genes and
mediate survival of the developing embryos, suggesting that these two genes have
overlapping roles in these processes.24 Because the doubly deficient mice die at 9.5
dpc, to investigate the function of these genes in adult tissues, we generated floxed
alleles of Bmi1 (Supplementary Figure 1A-C) and Mel18 (Supplementary Figure
1D-F), respectively. Without excision, the homozygous mice with the floxed alleles
were fertile and healthy and had no apparent abnormalities.
To examine the function of these genes in CAC development, we generated mice
containing floxed alleles of Bmi1 and/or Mel18 and a transgene expressing Cre
recombinase under the control of the villin gene promoter (villin-Cre), an
intestine-specific promoter.32 These mice are hereafter referred to as Bmi1�IEC,
Mel18�IEC and DKO mice (Supplementary Figure 2A and B), respectively.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
12
Interestingly, intestinal epithelial cell (IEC)-specific depletion of Bmi1 and/or Mel18
did not have an overt effect on epithelial homeostasis in the colon (Supplementary
Figure 2C), and the baseline rates of cell proliferation and apoptosis were also
indistinguishable between WT and single or DKO mice (Supplementary Figure 2D
and E). Furthermore, Bmi1�IEC, Mel18�IEC and DKO mice were fertile and healthy and
did not exhibit any overt phenotype. These data suggest that Bmi1 and Mel18 are
largely dispensable in the epithelium for colonic development and homeostasis.
Next, we examined whether these genes are required for tumor development in the
CAC mouse model. In this model, mutation and chronic inflammation in mice of
appropriate genotypes was triggered by injection with the colonotropic mutagen
azoxymethane (AOM), followed by three cycles of treatment with the luminal toxin
dextran sodium sulfate (DSS) (Figure 1A), as previously described.4 After the
AOM-DSS treatment, upon gross inspection, the WT mice gradually developed
intestinal polyps. Similarly, Bmi1�IEC or Mel18�IEC mice gradually developed polyps.
However, DKO mice exhibited significantly fewer polyps compared to WT, Bmi1�IEC
or Mel18�IEC mice (Figure 1B). Histological analysis revealed that the DKO mice had
markedly decreased polyp multiplicity and polyp load (Figure 1C). In addition,
adenomas observed in the colons of DKO mice generally displayed only low-grade
dysplasias (Figure 1B), but adenomas with high-grade dysplasia characterized by
aggressive invasion of the muscularis mucosa were frequently observed in WT,
Bmi1�IEC or Mel18�IEC colons (Figure 1B). Immunohistochemistry with Ki-67 and
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
13
BrdU, both markers of cellular proliferation, revealed significant reduces in polyp
cells proliferation in the DKO mice compared to WT mice (Supplementary Figure 3A
and B). Active caspase-3 and TUNEL analysis revealed an increase in the apoptotic
response of the colonic mucosa in DKO mice (Supplementary Figure 3C). The
reciprocal decrease in proliferation and increase in cell death occurred specifically in
the polyp tissues, but not in adjacent normal epithelial cells in DKO mice
(Supplementary Figure 3A-C). The above data demonstrate that Bmi1 and Mel18 are
required for CAC tumorigenesis.
To define the stage at which the loss of Bmi1 and Mel18 impacts tumorigenesis, we
generated mice containing the floxed alleles of Bmi1 and Mel18 and the
villin-Cre-ERT2 driver,33 which allows controlled gene ablation in IECs upon
tamoxifen treatment. Similar to the above results with the non-inducible villin-Cre,
conditional deletion of both genes during the earliest stages of CAC resulted in a
significant decrease in polyp multiplicity and polyp load (Supplementary Figure 4A
and B). To determine whether Bmi1 and Mel18 are continuously required for tumor
progression after tumor initiation, we administered tamoxifen to the DKO mice with
villin-Cre-ERT2 after the last DSS treatment (Figure 1D). Remarkably, conditional
depletion of Bmi1 and Mel18 during late stages of CAC significantly reduced the
multiplicity of large adenomas and the load of total adenomas (Figure 1E and F,
Supplementary Figure 4C). Although a similar tendency was observed for the total
number of polyps, it did not reach statistical significance (Figure 1F). Consistent with
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
14
the profound effect on polyp size, there was a significant reduction in cell
proliferation in the DKO mice and a reciprocal increase in the epithelial apoptosis
(Supplementary Figure 4D and E). Taken together, these data suggest that BMI1 and
MEL18 are required for both initiation and progression of colorectal neoplasia in the
mouse model of CAC.
BMI1 and MEL18 Regulate the Proliferation and Survival of Premalignant Cells
Independently of Ink4a/Arf
Because inflammation plays a critical role in tumorigenesis,4, 7 we hypothesized that
the reduced adenoma multiplicity and load in DKO mice could be due to decreased
intestinal inflammation. To this end, we used the acute colitis model that employs a
single 5-day course of DSS.4 Unexpectedly, after AOM-DSS treatment, the DKO
mice exhibited greater body weight loss than WT or single KO mice (Figure 2A).
Histologically, these mice exhibited a significantly higher degree of mucosa damage
and an increased incidence of ulcerations (Figure 2A and B). These are typical
phenotypes of hyper-inflammatory response. In agreement with this notion,
proinflammatory genes, including IL-6, IL-1β, IL-11, TNFα and Cox-2, were
significantly elevated in the colonic mucosa of DKO mice (Figure 2C). Thus, the
reduced polyp multiplicity and load in DKO mice may not due to reduced
inflammation, or the DKO causes the disruption of downstream pathways that are
important for the inflammation-induced tumorigenic function.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
15
In contrast with the increased inflammatory phenotypes, immunohistochemical
analysis of Ki-67 revealed impaired proliferation of the colonic epithelium of DKO
mice compared to WT (Figure 2D). Staining for active caspase-3 revealed a reciprocal
increase in apoptosis of IECs in DKO mice (Figure 2D). In contrast, we did not
observe any differences in mass loss (Supplementary Figure 5A and B) or
histopathological morphology (Supplementary Figure 5C) between Bmi1�IEC or
Mel18�IEC and WT mice. Therefore, simultaneous depletion of Bmi1 and Mel18 but
not either alone reduces cell proliferation and increases cell death in IECs during
CAC tumorigenesis, resulting in tumor suppression. Collectively, these results
indicate that the reduced polyp multiplicity and load in DKO mice is due to impaired
epithelial cell proliferation and decreased epithelial cell survival, despite the
accompany of increased inflammatory response, a paradoxical disconnection that
have been observed previously.4, 6, 7
The Cdkn2a locus, which encodes the p16Ink4a and p19Arf tumor suppressors, is
targeted by BMI1.19 Significant up-regulation of p16 mRNA levels has also been
observed in Mel18-/- MEFs.19 Consistent with those observations, our qRT-PCR
analysis revealed that the transcription of p16Ink4a and p19Arf was already significantly
increased in Bmi1�IEC or Mel18�IEC single K.O mice (Supplementary Figure 6A).
Therefore, the increased cell death by apoptosis in Bmi1�IEC/ Mel18�IEC double K.O
cannot be simply explained by increased transcription of p16Ink4a and p19Arf. To
functionally test this hypothesis, we generated villin-Cre; Cdkn2aF/F (termed
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
16
Cdkn2a�IEC), Bmi1�IEC/Cdkn2a�IEC and Mel18�IEC/Cdkn2a�IEC double-deficient mice,
and Bmi1�IEC/Mel18�IEC/Cdkn2a�IEC triple-deficient (referred to as B/M/C�IEC
hereafter) mice. We confirmed the absence of the p16 and p19 proteins in the colon
epithelium of Cdkn2a�IEC mice by immunoblot analysis of isolated epithelial cells
(Supplementary Figure 5B). As expected, after AOM-DSS treatment, there was no
significant alteration in polyp multiplicity or polyp load between B/M/C�IEC and DKO
mice (Figure 2E), and between Bmi1�IEC and Bmi1�IEC/Cdkn2a�IEC double-deficient
mice or between Mel18�IEC and Mel18�IEC/Cdkn2a�IEC double-deficient mice
(Supplementary Figure 6C). In addition, when chronic DSS colitis was induced by
three cycles of DSS without AOM injection, no significant difference in the intensity
of epithelial damage was observed between B/M/C�IEC and DKO mice
(Supplementary Figure 6D). These data suggest that tumor suppression by Bmi1 and
Mel18 ablation is not caused by activation of the Ink4a/Arf locus.
BMI1 and MEL18 Are Required for STAT3 Activation in Premalignant Cells
The observed up-regulation of IL-6 and IL-11, which are major STAT3 activators in
the colon,6, 7 suggested that STAT3 might be highly activated in premalignant IECs in
the DKO mice. However, using immunoblot analysis with antibodies against STAT3
phosphorylated at tyrosine residue 705 (pSTAT3, the activated STAT3), we observed
instead dampened STAT3 activation during 8-16 days of induced colitis in IECs of
DKO mice, a time period that is considered critical for the initiation of tumorigenesis
(Figure 2F). By contrast, activation of several other signaling pathway effectors, such
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
17
as S6 kinase (S6K) and ERK remained largely unchanged, and the STAT3 negative
feedback regulator SOCS3 was not changed (Figure 2F). Reduced pSTAT3 staining in
IECs in sectioned tissues from DKO mice was also observed (Supplementary Figure
7A). Again, the reduction in the pSTAT3 level was observed only in DKO mice and
not Bmi1�IEC or Mel18�IEC mice (Supplementary Figure 7B and C), indicating that
BMI1 and MEL18 have redundant roles in facilitating STAT3 activation in
premalignant IECs.
Inflammation via IL-6/STAT3 signaling induces the expression of cell cycle-related
regulators and anti-apoptotic genes to promote the survival and proliferation of
premalignant IECs.6, 7 Consistent with a reduction in STAT3 activity in DKO mice,
RT-PCR analysis revealed that many STAT3-regulated genes, including cyclins
(cyclin B1, D1, D2 and E), c-Myc and cdc2, were moderately or significantly
down-regulated (Supplementary Figure 7D). In addition, the proapoptotic gene Bcl10
and p21 (cell-cycle inhibitor) were significantly up-regulated in DKO mice compared
to control mice (Supplementary Figure 7D). This alteration in STAT3 activation
provides a molecular explanation for the tumor suppression effect of Bmi1 and Mel18
ablation during CAC tumorigenesis.
Identification of Reg3b and Reg3g as Novel PcG Target Genes in IECs
To elucidate the molecular mechanisms of BMI1 and MEL18 in STAT3 regulation
and, consequently, CAC tumorigenesis, we performed cDNA microarray analysis of
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
18
premalignant IECs isolated from WT, Bmi1�IEC, Mel18�IEC, and DKO mice on day 12
after AOM-DSS treatment. As predicted, the Ink4a/Arf (Cdkn2a) locus was among the
top upregulated genes (Supplementary Table 1). Interestingly, two similar genes,
Reg3b and Reg3g, were also at the top in the list of genes that were specifically
upregulated in DKO IECs (Figure 3A and Supplementary Table 1). Reg3b and Reg3g
encode two small proteins that belong to a lectin family of secreted factors. Among
the 7 family members of the Reg genes in the mouse genome, Reg3b and Reg3g
exhibited pronounced up-regulation (>35 fold), Reg4 exhibited much less pronounced
up-regulation (~4 fold), and other Reg genes remained unaltered in DKO mice
(Figure 3A and B). The upregulation of Reg3b and Reg3g was further confirmed by
quantitative RT-PCR analysis (Figure 3B), immunoblot and immunohistochemical
staining for REG3B (Figure 3 C and D).
To determine whether BMI1 and MEL18 directly or indirectly regulate Reg3b and
Reg3g expression, we conducted chromatin immunoprecipitation (ChIP) assays with
BMI1 and MEL18 across the promoter regions of Reg3b and Reg3g loci in mouse
colonic epithelial tissues and CT26.WT cells, a mouse colon carcinoma cell line. The
results revealed that BMI1 and MEL18 were able to bind to promoter regions for
Reg3b and Reg3g and that this binding activity was accompanied by H2A
ubiquitination (Figure 3E and F, and Supplementary Figure 8A-D). In contrast, ChIP
assays did not detect any significant binding activity for BMI1 or MEL18 in the
promoter regions of other Reg family genes (Supplementary Figure 8E). These
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
19
observations are consistent with the results obtained by microarray analysis (Figure
3A) and indicate that the Reg3b and Reg3g loci are specifically targeted by BMI1 and
MEL18 in the colon epithelium. Interestingly, many PcG genes, especially M33 and
CBX8, were significantly upregulated during CAC tumor development
(Supplementary Figure 8F). Along with the observation that H2A ubiquitination was
enriched at Reg3b promoter, it is possible that the canonical PRC1 complexes are
involved in the regulation of REG3B expression.
REG3B Inhibits IL-6/ IL-11-Mediated STAT3 Activation
To determine whether REG3B/REG3G can negatively regulate STAT3 signaling, we
expressed full-length Reg3b and Reg3g in 293T cells and examined the effect on
STAT3 activation in response to IL-6 or IL-11. Normally, cellular STAT3 is strongly
phosphorylated upon IL-6 stimulation. Remarkably, in response to IL-6, STAT3
phosphorylation was considerably inhibited in Reg3b-transfected cells but not
Reg3g-transfected cells (Supplementary Figure 9A), and overexpression of REG3B
was able to inhibit STAT3 phosphorylation during the entire period of IL-6 treatment
in CT26.WT cells (Figure 4A). Similarly, IL-11 treatment induced moderate and
prolonged STAT3 activation in CT26.WT cells, and overexpression of REG3B also
significantly inhibited IL-11-induced STAT3 phosphorylation (Figure 4B). These data
demonstrate that expression of REG3B, but not REG3G, effectively interferes with
cytokine-mediated STAT3 activation in cultured cells.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
20
Mechanisms of IL-6/IL-11-mediated STAT3 activation are well understood. The
binding of extracellular IL-6/IL-11 to the cell surface receptor complex (IL-6R/GP130
or IL-11R/GP130) triggers activation of the receptor complex, which then activates
the associated Janus Kinase (JAK) via their mutual phosphorylation. The activated
JAK then phosphorylates cytoplasmic STAT3 and induces their dimerization and
activation. We observed that expression of REG3B reduced JAK phosphorylation
(Figure 4C), as well as GP130 phosphorylation (Figure 4D), indicating that REG3B
interferes with STAT3 signaling above JAK phosphorylation, possibly at the level of
the receptor complex. Interestingly, REG3B was able to interact with IL-6R and
IL-11R, respectively, in both CT26.WT cell extract (Supplementary Figure 9B and C)
and colonic epithelial tissue extract (Figure 4E). However, we failed to detect any
interactions of REG3B with IL-6 or GP130 by immunoprecipitation (Figure 4F and
Supplementary Figure 9B). Moreover, pre-incubation of cells with recombinant
REG3B protein inhibited IL-6-induced STAT3 activation in a dose-dependent manner
(Supplementary Figure 9D). In contrast, addition of REG3B protein after IL-6
treatment failed to inhibit STAT3 activation (Supplementary Figure 9E). Taken
together, these observations suggest that REG3B interferes with STAT3 signaling
extracellularly by binding to the receptor complex and inhibiting its activation by IL-6
or IL-11.
REG3B Treatment Suppresses CAC Tumorigenesis
To determine whether REG3B expression is sufficient to suppress CAC tumorigenesis,
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
21
we treated two groups of mice separately with recombinant REG3B protein. One
group was treated at “early stages” and another group was treated at “late stages” of
CAC, as illustrated in Figure 5A and Figure 5C. Strikingly, both early and late
REG3B treatment effectively reduced polyp multiplicity and polyp load compared to
PBS-treated mice (Figure 5B and D). Immunohistochemical staining revealed that
REG3B treatment significantly reduced cell proliferation and induced apoptosis in
epithelial cells (Supplementary Figure 10A-D). In addition, it markedly reduced
pSTAT3 levels in infiltrating inflammatory cells and IECs of CAC adenomas
(Supplementary Figure 10A and C). Meanwhile, we did not observe any altered
expression of BMI1, MEL18, p16, and p19 proteins (Supplementary Figure 11A)
between control and REG3B treatment mice. These data suggest that REG3B
treatment is sufficient to interfere with STAT3 activation in vivo and is effective in
preventing both tumor initiation and growth during CAC tumorigenesis.
The BMI1/MEL18-REG3B-STAT3 Signaling Axis in CAC Tumorigenesis
The above results imply that excessive production of REG3B following Bmi1 and
Mel18 ablation interferes with cytokine-mediated STAT3 activation in premalignant
cells and consequently suppresses tumorigenesis. Therefore, PcG proteins may
promote CAC tumorigenesis through a novel BMI1/MEL18-REG3B-STAT3
signaling axis. To further test the existence of this regulatory pathway, especially to
evaluate the contribution of REG3B in STAT3 inhibition followed by the loss of
BMI1 and MEL18, we generated Bmi1, Mel18 and Reg3b single, double or triple
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
22
mutant 3T3 cells, a mouse embryonic fibroblast cell line, using the CASPR/Cas9
system (Supplementary Figure 11B),37 and tested their responsiveness to IL-6- or
IL-11-induced STAT3 activation. Normally, treatment with IL-6 or IL-11 in 3T3 cells
was able to induce robust STAT3 activation (Figure 6A). Interestingly, the levels of
STAT3 activation were significantly reduced when both Bmi1 and Mel18, but not
either alone, were mutated (Figure 6A). Interestingly, upregulation of Reg3b was
observed in either Bmi1 or Mel18 K.O. cells, but a significantly higher magnitude of
upregulation was found in the Bmi1-/-/Mel18-/- (termed B-/-/M-/-) cells (Supplementary
Figure 11C), indicating that BMI1 and MEL18 have a partially redundant function in
regulating REG3B in these cells as well. Importantly, additional mutation of Reg3b
could rescue STAT3 activity in the Bmi1-/-/Mel18-/-/Reg3b-/- (termed B-/-/M-/-/R-/-) cells
back to control levels (Figure 6A), suggesting that the inhibitory effect on STAT3
following the loss of BMI1 and MEL18 is mainly due to the activation of REG3B.
To determine the contribution of this regulatory axis, especially the REG3B
component, to CAC tumorigenesis in vivo, we generated conditional Reg3b-deficient
mice by CRISPR-Cas9-mediated targeting (Supplementary Figure 12A), and studied
its role in the tumor suppressive effect following Bmi1 and Mel18 depletion. The
villin-Cre; Reg3bF/F deficient mice (termed Reg3b�IEC), which appeared
phenotypically normal, were almost completely devoid of REG3B protein in IECs
(Supplementary Figure 12B). We then engineered the Bmi1�IEC/Mel18�IEC/Reg3b�IEC
triple-deficient (referred to as TKO hereafter) mice, and studied the consequences in
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
23
the AOM-DSS model. Strikingly, TKO mice showed efficient development of
adenomas, with significantly increased adenomas multiplicity and load compared to
DKO mice (Figure 6B and C). Interestingly, polyp multiplicity, polyp load, and
adenomas proliferation were also significantly higher in TKO mice compared to that
in the WT control (Figure 6B and D, Supplementary Figure 12C). This is probably
caused by the contribution of a baseline tumor suppressive activity of REG3B in the
WT colonic epithelium, as Reg3b�IEC alone mice exhibited relatively small but
significantly increased adenomas proliferation and pSTAT3 level (Figure 6B-E,
Supplementary Figure 12C). Immunohistochemical staining also revealed the
upregulation of pSTAT3 expression in TKO adenoma cells (Figure 6E). Therefore, the
tumor inhibitory effect caused by Bmi1 and Mel18 depletion can be completely
eliminated by depleting Reg3b, demonstrating genetically that Reg3b is the major
downstream effector of BMI1 and MEL18 in regulating CAC tumor development.
Taken together, these observations strongly support a linear regulatory relationship
among BMI1/MEL18, REG3B and STAT3, which may constitute an important
regulatory pathway to control the initiation and development of CAC.
REG3β in Human Colorectal Cancer
We next assessed the potential role of the Reg family genes in human colorectal
cancers. There are 5 Reg family genes in the human genome, which are aligned
contiguously in a single cluster: REG1α, REG1β, REG3β, REG3γ and REG4.38 We
first expressed each of these genes in 293T cells using lentiviral-based vectors and
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
24
examined the effect on IL-6-mediated STAT3 signaling. Interestingly, among these
five genes, only REG3β exhibited robust STAT3 inhibition (Figure 7A). Coincidently,
REG3β is the most similar to mouse REG3B and is considered the human ortholog of
mouse REG3B.39, 40 To determine whether REG3β expression is sufficient to suppress
CAC tumorigenesis, we treated mice with recombinant REG3β protein at late stage of
CAC, as illustrated in Figure 7B, REG3β treatment significantly reduced polyp load
and marginally reduced polyp multiplicity (statistically not significant), compared to
PBS-treatment (Figure 7B).
To assess the clinical relevance of REG3β in colorectal cancer, we examined REG3β
expression in a tumour tissue microarray (TMA) consisting of 87 colorectal cancer
specimens, and found that patients with higher expression of REG3β in the tumor
tissue had a better 5-year survival rate than patients with moderate or lower
expression of REG3β (p < 0.01 and p< 0.0001, respectively) (Figure 7C). To evaluate
the relationships between REG3β expression and STAT3 activity, we examined
REG3β and pSTAT3 expression by immunostaining in TMA consisting of 92
colorectal cancer specimens. We found that REG3β was more frequently expressed in
tumor tissues with low pSTAT3 than with high pSTAT3 levels (Figure 7D and E).
Taken together, these data suggest that the expression level of REG3β inversely
correlates with pSTAT3 activity and disease prognosis, and indicate that REG3β could
be a tumor suppressor in CAC development by inhibiting STAT3 activity.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
25
Discussion
Here we have identified a novel and critical role for BMI1 and its homologous protein
MEL18 in the initiation and progression of CAC tumorigenesis by regulating cellular
response to cytokine signaling. We have identified Reg3b as a new tumor suppressor
gene targeted by BMI1/MEL18 in the experimental model of CAC. Gene expression
profiling identified significant up-regulation of Reg3b and Reg3g expression when
Bmi1 and Mel18 are depleted. The results of chromatin immunoprecipitation assays
suggest that BMI1 and MEL18 directly target the promoter regions of both Reg3b and
Reg3g. Interestingly, although most Reg genes are aligned in a single cluster in the
genome, BMI1 and MEL18 seem to specifically target Reg3b and Reg3g, at least in
colonic epithelial cells, indicating that this gene cluster is under sophisticated
regulatory control. Importantly, our triple conditional knock-out mice revealed that
Reg3b is responsible for the tumor suppressive effect following the ablation of Bmi1
and Mel18. Consistent with the genetic relationships, administration of recombinant
REG3B significantly suppresses tumor initiation and growth during CAC
tumorigenesis, and the effect is largely similar to that caused by ablation of Bmi1 and
Mel18. The importance of Reg3b repression by BMI1 and MEL18 for tumorigenesis
suggests that Reg3b represents another major tumor suppressor locus targeted by PcG
proteins.
Our in vitro and in vivo studies strongly suggest that the tumor suppressive effect of
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
26
REG3B in the CAC model is, at least in part, due to its anti-STAT3 activity. REG3B
belongs to a family of small, secreted proteins that contain a single C-type
(calcium-dependent) lectin domain. Reg proteins, also called pancreatitis-associated
proteins (PAP), were initially discovered as proteins that are strongly induced by
pancreatitis or during islet regeneration.38 Subsequent studies revealed that Reg
proteins are also expressed in a number of physiological or pathological processes.
For instance, it has been suggested that Reg3g functions as a carbohydrate-binding
bactericidal lectin to maintain microbial integrity in the gut.41 Prior to this report, a
few studies have implicated Reg proteins in regulating STAT3 activity. In Reg3b KO
mice, STAT3 activity is enhanced during liver regeneration.39 By contrast, STAT
activation is reduced during caerulein-induced pancreatitis.42 These seemingly
contradictory observations may indicate the complexity underlying the functions of
the Reg family proteins. Although Reg family proteins share a common lectin domain,
they may have distinct functions that are cell type-specific or context-dependent.
Indeed, although Reg3b and Reg3g are targeted by BMI1 and MEL18, only Reg3b
displays robust anti-STAT3 activity. We observed that the addition of recombinant
REG3B protein to the medium inhibited IL-6-mediated STAT3 activation, suggesting
that REG3B acts extracellularly to inhibit STAT3 signaling. Moreover, we found that
REG3B physically interacts with the receptor complex but not the ligand. These
observations indicate that REG3B interferes with STAT3 signaling by interfering with
the activation of the IL-6R/GP130 receptor complex (Figure 7F). Further molecular
and structural studies are needed to determine whether REG3B exerts its effect via
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
27
interference with the ligand-receptor interaction or through other mechanisms.
Because REG3B also pocesses bactericidal activity,43 in addition to the direct role in
interfering STAT3 signaling in tumor cells, it could also indirectly impact tumor
progression by modulating microbial composition in the gut, especially during late
stages of tumorigenesis, which warrantees further investigation in the future.
Consistent with a tumor-suppressive role for REG3B/REG3β in the CAC model, the
expression of REG3β is inversely correlated with the prognosis of colorectal cancers.
Interestingly, REG3B/REG3β is highly expressed in the small intestine but not the
colon.38 It is possible that the expression of these proteins and other potential tumor
suppressors in the small intestine provides a tumor-suppressive environment, in
agreement with the strikingly distinct incidences of carcinomas in the small intestine
and colon. Identification of REG3B as a polycomb target also indicates a potential
connection among polycomb function, gut microbiota and cancer. Further
investigation of the Reg family proteins as potential tumor suppressors and their
regulation under physiological and pathological conditions may facilitate the
development of effective therapeutics for colorectal cancers or other
inflammation-associated diseases.
Supplementary Material
Supplemental material includes three tables, twelve figures, supplementary data, and
supplementary materials and methods.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
28
Acknowledgments
We thank Dr. Sylvie Robine for providing the villin-Cre-ERT2 mice and Dr. Xiaodong
Wang and members of the Xi laboratory for critically reading the manuscript.
References
1. Terzic J, Grivennikov S, Karin E, et al. Inflammation and colon cancer. Gastroenterology
2010;138:2101-2114 e5.
2. Grivennikov SI. Inflammation and colorectal cancer: colitis-associated neoplasia. Semin
Immunopathol 2013;35:229-44.
3. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell
2010;140:883-99.
4. Greten FR, Eckmann L, Greten TF, et al. IKKbeta links inflammation and tumorigenesis in a
mouse model of colitis-associated cancer. Cell 2004;118:285-96.
5. Karin M, Clevers H. Reparative inflammation takes charge of tissue regeneration. Nature
2016;529:307-15.
6. Grivennikov S, Karin E, Terzic J, et al. IL-6 and Stat3 are required for survival of intestinal
epithelial cells and development of colitis-associated cancer. Cancer Cell 2009;15:103-13.
7. Bollrath J, Phesse TJ, von Burstin VA, et al. gp130-mediated Stat3 activation in enterocytes
regulates cell survival and cell-cycle progression during colitis-associated tumorigenesis.
Cancer Cell 2009;15:91-102.
8. Putoczki TL, Thiem S, Loving A, et al. Interleukin-11 is the dominant IL-6 family cytokine
during gastrointestinal tumorigenesis and can be targeted therapeutically. Cancer Cell
2013;24:257-71.
9. Matthews JR, Sansom OJ, Clarke AR. Absolute requirement for STAT3 function in
small-intestine crypt stem cell survival. Cell Death Differ 2011;18:1934-43.
10. Gao SP, Mark KG, Leslie K, et al. Mutations in the EGFR kinase domain mediate STAT3
activation via IL-6 production in human lung adenocarcinomas. J Clin Invest
2007;117:3846-56.
11. Fukuda A, Wang SC, Morris JPt, et al. Stat3 and MMP7 contribute to pancreatic ductal
adenocarcinoma initiation and progression. Cancer Cell 2011;19:441-55.
12. Lesina M, Kurkowski MU, Ludes K, et al. Stat3/Socs3 activation by IL-6 transsignaling promotes
progression of pancreatic intraepithelial neoplasia and development of pancreatic cancer.
Cancer Cell 2011;19:456-69.
13. Valk-Lingbeek ME, Bruggeman SW, van Lohuizen M. Stem cells and cancer; the polycomb
connection. Cell 2004;118:409-18.
14. Sparmann A, van Lohuizen M. Polycomb silencers control cell fate, development and cancer.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
29
Nat Rev Cancer 2006;6:846-56.
15. Wassef M, Rodilla V, Teissandier A, et al. Impaired PRC2 activity promotes transcriptional
instability and favors breast tumorigenesis. Genes Dev 2015;29:2547-62.
16. van Lohuizen M, Verbeek S, Scheijen B, et al. Identification of cooperating oncogenes in E
mu-myc transgenic mice by provirus tagging. Cell 1991;65:737-52.
17. Haupt Y, Alexander WS, Barri G, et al. Novel zinc finger gene implicated as myc collaborator by
retrovirally accelerated lymphomagenesis in E mu-myc transgenic mice. Cell 1991;65:753-63.
18. Jacobs JJ, Scheijen B, Voncken JW, et al. Bmi-1 collaborates with c-Myc in tumorigenesis by
inhibiting c-Myc-induced apoptosis via INK4a/ARF. Genes Dev 1999;13:2678-90.
19. Jacobs JJ, Kieboom K, Marino S, et al. The oncogene and Polycomb-group gene bmi-1
regulates cell proliferation and senescence through the ink4a locus. Nature 1999;397:164-8.
20. Molofsky AV, He S, Bydon M, et al. Bmi-1 promotes neural stem cell self-renewal and neural
development but not mouse growth and survival by repressing the p16Ink4a and p19Arf
senescence pathways. Genes Dev 2005;19:1432-7.
21. Liu J, Cao L, Chen J, et al. Bmi1 regulates mitochondrial function and the DNA damage
response pathway. Nature 2009;459:387-92.
22. Bruggeman SW, Hulsman D, Tanger E, et al. Bmi1 controls tumor development in an
Ink4a/Arf-independent manner in a mouse model for glioma. Cancer Cell 2007;12:328-41.
23. Gargiulo G, Cesaroni M, Serresi M, et al. In vivo RNAi screen for BMI1 targets identifies
TGF-beta/BMP-ER stress pathways as key regulators of neural- and malignant glioma-stem
cell homeostasis. Cancer Cell 2013;23:660-76.
24. Akasaka T, van Lohuizen M, van der Lugt N, et al. Mice doubly deficient for the Polycomb
Group genes Mel18 and Bmi1 reveal synergy and requirement for maintenance but not
initiation of Hox gene expression. Development 2001;128:1587-97.
25. Gao Z, Zhang J, Bonasio R, et al. PCGF homologs, CBX proteins, and RYBP define functionally
distinct PRC1 family complexes. Mol Cell 2012;45:344-56.
26. Elderkin S, Maertens GN, Endoh M, et al. A phosphorylated form of Mel-18 targets the
Ring1B histone H2A ubiquitin ligase to chromatin. Mol Cell 2007;28:107-20.
27. Taherbhoy AM, Huang OW, Cochran AG. BMI1-RING1B is an autoinhibited RING E3 ubiquitin
ligase. Nat Commun 2015;6:7621.
28. Sangiorgi E, Capecchi MR. Bmi1 is expressed in vivo in intestinal stem cells. Nat Genet
2008;40:915-20.
29. Lopez-Arribillaga E, Rodilla V, Pellegrinet L, et al. Bmi1 regulates murine intestinal stem cell
proliferation and self-renewal downstream of Notch. Development 2015;142:41-50.
30. Maynard MA, Ferretti R, Hilgendorf KI, et al. Bmi1 is required for tumorigenesis in a mouse
model of intestinal cancer. Oncogene 2014;33:3742-7.
31. Wilson JE, Petrucelli AS, Chen L, et al. Inflammasome-independent role of AIM2 in
suppressing colon tumorigenesis via DNA-PK and Akt. Nat Med 2015;21:906-13.
32. Madison BB, Dunbar L, Qiao XT, et al. Cis elements of the villin gene control expression in
restricted domains of the vertical (crypt) and horizontal (duodenum, cecum) axes of the
intestine. J Biol Chem 2002;277:33275-83.
33. el Marjou F, Janssen KP, Chang BH, et al. Tissue-specific and inducible Cre-mediated
recombination in the gut epithelium. Genesis 2004;39:186-93.
34. Aguirre AJ, Bardeesy N, Sinha M, et al. Activated Kras and Ink4a/Arf deficiency cooperate to
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
30
produce metastatic pancreatic ductal adenocarcinoma. Genes Dev 2003;17:3112-26.
35. Neufert C, Becker C, Neurath MF. An inducible mouse model of colon carcinogenesis for the
analysis of sporadic and inflammation-driven tumor progression. Nat Protoc
2007;2:1998-2004.
36. Wang F, Scoville D, He XC, et al. Isolation and characterization of intestinal stem cells based
on surface marker combinations and colony-formation assay. Gastroenterology
2013;145:383-95 e1-21.
37. Platt RJ, Chen S, Zhou Y, et al. CRISPR-Cas9 knockin mice for genome editing and cancer
modeling. Cell 2014;159:440-55.
38. Parikh A, Stephan AF, Tzanakakis ES. Regenerating proteins and their expression, regulation
and signaling. Biomol Concepts 2012;3:57-70.
39. Lieu HT, Simon MT, Nguyen-Khoa T, et al. Reg2 inactivation increases sensitivity to Fas
hepatotoxicity and delays liver regeneration post-hepatectomy in mice. Hepatology
2006;44:1452-64.
40. Liu JL, Cui W. Which gene, Reg2 or Reg3beta, was targeted that affected liver regeneration?
Hepatology 2007;45:1584-5; author reply 1585-6.
41. Cash HL, Whitham CV, Behrendt CL, et al. Symbiotic bacteria direct expression of an
intestinal bactericidal lectin. Science 2006;313:1126-30.
42. Gironella M, Folch-Puy E, LeGoffic A, et al. Experimental acute pancreatitis in PAP/HIP
knock-out mice. Gut 2007;56:1091-7.
43. Aden K, Rehman A, Falk-Paulsen M, et al. Epithelial IL-23R Signaling Licenses Protective IL-22
Responses in Intestinal Inflammation. Cell Rep 2016;16:2208-18.
Authors names in bold designate shared co-first authorship.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
31
Figure Legends
Figure 1. Double deficiency of Bmi1 and Mel18 decreases adenoma formation and
growth in the CAC model. (A) A schematic overview of the CAC model. (B)
Representative colon sections of WT, Bmi1�IEC, Mel18�IEC, and DKO mice at the end
of the CAC challenge stained with hematoxylin and eosin (H&E). (C) Polyp
multiplicity and average polyp load were analyzed in WT, Bmi1�IEC, Mel18�IEC, and
DKO mice. (D) Scheme of Bmi1 and Mel18 deletion during the late stage of CAC
growth. (E) H&E-stained of representative colon sections of WT and DKO mice at
the end of the CAC challenge showing the polyp reduction in DKO mice. (F) Polyp
number, number of polyps (polyp size > 2 mm), and polyp load were analyzed in WT
and DKO mice. Scale bars = 100 �m. The data are presented as the mean ± SEM (n =
6-30). ns, not significant; *p < 0.05; ***p < 0.001; by one-way ANOVA (C) or
two-tailed, unpaired t-test (F).
Figure 2. Colonic inflammation is increased in the DKO mice through dampened
STAT3 activation during DSS-induced colitis. (A) The body mass of DKO mice
decreased dramatically compared to WT littermates during acute DSS colitis (Upper
panel). Ulcer number and histological damage at day 15 in WT and DKO mice treated
with 3.5% DSS (Lower panel). (B) H&E-stained of colons from WT and DKO mice
at day 15. (C) Relative cytokine mRNA levels in whole colonic mucosa from WT and
DKO mice at day 15 of the CAC model. (D) Immunohistochemical analysis of Ki-67
and cleaved caspase-3 in colons from WT and DKO mice at day 15 of the CAC model.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
32
(E) Polyp numbers and polyp load in WT, Cdkn2a�IEC, DKO, and B/M/C�IEC mice. (F)
Colonic lysates were prepared at the indicated times, and the expression and
phosphorylation of the indicated proteins were analyze. β-actin was used as loading
controls. Scale bars = 100 �m. The data are presented as the mean ±SEM (n = 3-15).
ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001; by two-tailed, unpaired t-test
(A,D), two-way ANOVA (C), or one-way ANOVA (E). All data shown are
representative of at least three independent experiments.
Figure 3. Identification of reg3b as a novel PcG target. (A) Heat-map of the
microarray results shown significant up-regulation of the Reg3b and Reg3g genes in
IECs of the Bmi1 and Mel18 DKO compared to control mice. Red, up-regulated;
green, down-regulated; black, no change. Results are representative of two
independent experiments. (B) Quantitative PCR analysis of mouse Reg family genes
in IEC samples from DKO and WT mice. (C) Immunoblot analysis of the REG3B
protein in isolated colonic enterocytes from WT, Bmi1�IEC, Mel18�IEC and DKO mice.
β-actin was used as loading controls. (D) Immunohistochemical analysis of the
REG3B protein in colonic tissues from WT and DKO mice. (E and F) Schematic
representation of the Reg3b locus, with black squares marked 1, 2 indicating the
amplified regions in the ChIP studies (Upper panel). ChIP analysis of the binding of
the BMI1 (E) and MEL18 (F) antibodies at the Reg3b different loci in mouse colonic
epithelial tissues (Lower panel). IgG was used as a control. Scale bars= 100 �m. The
data are presented as the mean ± SEM (n = 3-6). *p < 0.05; **p < 0.01; by two-tailed,
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
33
unpaired t-test (E,F). All data shown are representative of at least three independent
experiments.
Figure 4. REG3B inhibits STAT3 signaling by interfering IL-6R/gp130 receptor
activation. (A) Immunoblot analysis of p-STAT3Y705, STAT3 and REG3B in lysates
from Reg3b-transfected or empty vector-transfected CT26.WT cells at different time
points after IL-6 (25 ng/ml) treatment. (B) Immunoblot analysis of p-STAT3Y705,
STAT3 and REG3B in lysates from Reg3b-transfected or empty vector-transfected
CT26.WT cells at different time points after IL-11 (50 ng/ml) treatment. (C)
Reg3b-transfected and empty vector-transfected CT26.WT cell lysates were first
immunoprecipitated with an anti-JAK2 antibody followed by the detection of tyrosine
phosphorylation with an anti-phosphotyrosine (PY20) antibody. Cell lysates were also
subjected to western blotting analysis using antibodies as indicated. (D)
Reg3b-transfected and empty vector-transfected CT26.WT cell lysates were
immunoprecipitated with an anti-GP130 antibody followed by the detection of
tyrosine phosphorylation with the PY20 antibody. (E) Co-immunoprecipitation (Co-IP)
analysis of REG3B/IL-6R association in mouse colonic epithelial tissues (n= 5 mice).
WCL: whole cell lysates, IP: Immunoprecipitation. (F) CO-IP analysis of the
interactions between REG3B and IL-6. β-actin was used as loading controls. All data
are representative of at least three independent experiments.
Figure 5. REG3B treatment suppresses CAC tumorigenesis. (A) Schematic overview
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
34
of REG3B treatment during CAC induction. Mice were injected intraperitoneally with
recombinant REG3B (5 µg/mouse) versus control PBS on days 1, 3, and 5 during
each DSS cycle. Colons were removed at day 100 after AOM injection. (B) The polyp
numbers and polyp load were determined microscopically. (C) Schematic overview of
REG3B administration during the late stage of the CAC model. Mice were treated
with recombinant REG3B (5 µg/mouse) versus control PBS every 3 days after the last
DSS cycle. Colons were removed at day 100 after AOM injection. (D) The polyp
numbers (polyp size > 2 mm) and polyp load were measured microscopically. The
data are presented as the mean ± SEM (n = 6-9). *p < 0.05; **p < 0.01; by two-tailed,
unpaired t-test.
Figure 6. The BMI1/MEL18-REG3B-STAT3 regulatory axis in CAC tumorigenesis.
(A) Immunoblot analysis of p-STAT3Y705 and STAT3 in lysates from WT, Reg3b-/-,
Bmi1-/-, Mel18-/-, B-/-/M-/-, and B-/-/M-/-/R-/- 3T3 cell after IL-6 treatment (Left panel)
and IL-11 treatment (Right panel). β-actin was used as loading controls. (B) Polyp
multiplicity and average polyp load were analyzed in WT, Reg3b�IEC, DKO, and TKO
mice. (C) H&E-stained of representative colon sections of WT, Reg3b�IEC, DKO
and TKO mice at the end of the CAC challenge shown the polyps are rescued in TKO
mice compared to DKO. (D) Percentage Ki-67 positive cells in WT, Reg3b�IEC, DKO
and TKO colonic polyps was calculated. (E) Immunohistochemical analysis of the
p-STAT3 protein in colonic polyp tissues of WT, Reg3b�IEC, DKO and TKO mice.
Scale bars=100 �m. The data are presented as the mean ± SEM (n = 5-15). ns, not
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
35
significant; **p < 0.01; ***p < 0.001; by one-way ANOVA (B,D).
Figure 7. REG3β in human colorectal cancer. (A) Immunoblot analysis of
p-STAT3Y705 in REG-overexpressing 293T cells incubated with IL-6 (25 ng/ml) or
untreated. β-actin was used as loading controls. Results are representative of one of
three experiments. (B) Upper panel: schematic overview of REG3β administration
during the late stage of the CAC model. Mice were treated with recombinant REG3β
(7 µg/mouse) versus control PBS every 3 days after the last DSS cycle. Colons were
removed at day 100 after AOM injection. Lower panel: the polyp numbers and polyp
load were measured microscopically. (n=14 for control and n=16 for REG3β). (C)
Upper panel: a Kaplan-Meier survival curve of 87 colorectal cancer patients grouped
based the levels of REG3β expression (high, moderate, low) in the tumor tissue. HR,
hazard ratio. Lower panel: representative micrographs of immunohistochemical
staining of REG3β in specimens of colon cancer tumor. (D) Representative
micrographs of immunohistochemical staining of phospho-STAT3 and REG3β in
matched specimens of human colorectal tumor. Insets show staining of specimens in
more detail. (E) Expression and clinical relevance of REG3β and phospho-STAT3
were assessed in human colorectal cancer TMAs. (F) Model of
BMI1/MEL18-REG3B-STAT3 regulatory axis in CAC Tumorigenesis. Scale bars =
100 �m. The data are presented as the mean ± SEM. ns, not significant; *p < 0.05;
**p < 0.01; ***p < 0.001; by two-tailed, unpaired t-test (B), log-rank test (C), or
Fisher’s exact test (E).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
1
Supplemental Material
BMI1 and MEL18 Promote Colitis-Associated Cancer in Mice via
REG3B and STAT3
Xicheng Liu, Wendi Wei, Xiaowei Li, Pengcheng Shen, Zhen Wang, Dapeng Ju,
Rukui Zhang, Fu Yang, Chunyan Chen, Kun Cao, Guoli Zhu, Hongyan Chen, Liang
Chen, Jianhua Sui, Erquan Zhang, Kaichun Wu, Fengchao Wang, Liping Zhao, and
Rongwen Xi
Inventory of Supplemental Material:
Supplementary Figures 1-12
Supplementary Tables 1-3
Supplementary Data
Supplementary Materials and Methods
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
2
Supplementary Figure 1
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
3
Supplementary Figure 1. Generation of floxed alleles of Bmi1 and Mel18. (A) Schematic diagram of the targeting strategy for the generation of a floxed allele of Bmi1. Exons are indicated by filled blocks with numbers. Neo, neomycin resistance gene; red triangle, loxP site; blue inverted triangle, frt site. (B and C) Southern blot analysis of genomic DNA from electroporated embryonic stem cell clones (129xC57BL/6 strain) after digestion with SacI (B). The 5' external probe detects a 5.7 kb wild type allele fragment and a 4.5 kb targeted fragment. PCR genotyping of targeted embryonic stem clones (C). PCR was performed on DNA prepared from tail biopsies, using a mixture of primers. To amplify the following two bands: 134 bp (floxed allele, lanes 1, 3 and 4), 70 bp (Wild type allele, lanes 1 and 2). (D) Schematic diagram of the targeting strategy for Mel18. Exons are indicated by filled blocks with numbers. Neo, neomycin resistance gene; red triangle, loxP site; blue inverted triangle, frt site. (E and F) Southern blot of DNA from targeted embryonic stem cell clones (129xC57BL/6 strain) after digestion with SmiI (E). The 5' external probe detects a 8.7 kb wild type fragment and a 5.3 kb targeted fragment. PCR genotyping of targeted embryonic stem clones (F). PCR was performed on DNA prepared from tail biopsies, using a mixture of primers. To amplify the following two bands: 113bp (floxed allele, lanes 3, 4, 5 and 6), 178 bp (Wild type allele, lanes 1, 2, 3 and 4).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
4
Supplementary Figure 2
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
5
Supplementary Figure 2. Villin-Cre-mediated deletion of Bmi1 and/or Mel18 does not significantly affect normal architecture or homeostasis of the colonic epithelium. (A) Quantitative PCR analysis of Bmi1 and Mel18 mRNA expression in IECs isolated from WT, Bmi1�IEC, Mel18�IEC and DKO mice. (B) Immunoblot analysis of the IECs confirmed the depletion of BMI1 and MEL18 proteins in WT, Bmi1�IEC, Mel18�IEC and DKO mice. (C) H&E-stained sections of control and experimental colon tissues revealed no gross changes in architecture or histology. Goblet cells were marked by PAS-haematoxylin staining and immunohistochemical analysis (IHC) against Chromogranin A was used to mark enteroendocrine cells. These cells appeared to be properly differentiated in the intestine of DKO mice. Scale bars = 100 µm. (D and E) Immunohistochemical analysis of Ki-67 (D) and cleaved caspase-3 (E) of colons from WT and DKO mice. Depletion of Bmi1 and Mel18 did not have a significant impact on the basal proliferation and differentiation of IECs. Scale bars = 100 µm. The data are presented as the mean ± SEM (n = 3-7). **p < 0.01; ***p < 0.001; by two-way ANOVA (A). All data shown are representative of at least three independent experiments.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
6
Supplementary Figure 3
Supplementary Figure 3. BMI1 and MEL18 regulate IEC proliferation and survival. (A) Ki-67 and BrdU staining revealed the decrease in proliferating cells in polyp of Bmi1 and Mel18 double deficient mice compared to WT mice. (B) Percentage Ki-67 positive cells in WT and DKO colonic crypts and colonic polyps was calculated. (C) Cleaved caspase-3 immunohistochemistry and TUNEL staining revealed the increase apoptosis in polyp of Bmi1 and Mel18 double deficient mice compared to WT mice. Scale bars = 100 �m. The data are presented as the mean ± SEM. ns, not significant; **p < 0.01; by two-way ANOVA (B).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
7
Supplementary Figure 4
Supplementary Figure 4. BMI1 and MEL18 are important for polyp growth and proliferation. (A) Scheme of Bmi1 and Mel18 deletion during the stage of CAC induction. (B) Polyp multiplicity and average polyp load were analyzed in WT and DKO mice. (C) Microscopic changes in the colons of WT and DKO mice on day 100 in the CAC model. (D) Immunohistochemical staining of Ki-67 and cleaved caspase-3 of polyps from WT and DKO mice. (E) Percentage Ki-67 positive cells in WT and TKO colonic polyps was calculated. Scale bars = 100 �m. The data are presented as the mean ± SEM (n = 4-9). ns, not significant; *p < 0.05; **p < 0.01; by two-tailed, unpaired t-test (B,E).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
8
Supplementary Figure 5
Supplementary Figure 5. Ablation of Bmi1 or Mel18 in IECs does not overtly affect the proliferation or apoptosis of colon epithelial cells during induced tumorigenesis. (A and B) The body mass of Bmi1�IEC (A) and Mel18�IEC (B) mice did not changed compared to WT littermates during acute DSS colitis. (C) H&E staining, immunohistochemical analysis of Ki-67, and cleaved caspase-3 immunohistochemistry revealed no change in proliferation and apoptosis in colons of WT, Bmi1�IEC, and Mel18�IEC mice at day 15. Scale bars= 100 �m. The data are presented as the mean ± SEM (n = 9). ns, not significant; by one-way ANOVA (C).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
9
Supplementary Figure 6
Supplementary Figure 6. Ink4a/Arf deficiency fails to rescue the DKO tumor phenotype. (A) Quantitative PCR analysis of p16 and p19 mRNA expression in IECs isolated from WT, Bmi1�IEC, Mel18�IEC and DKO mice at day 12 of the CAC model. (B) Immunoblot analysis of p16 and p19 in IEC lysates from WT and Cdkn2a�IEC mice. (C) Polyp numbers and polyp load were analyzed in WT, Cdkn2a�IEC, Bmi1�IEC and Bmi1�IEC/Cdkn2a�IEC mice (Left panel) or in WT, Cdkn2a�IEC, Mel18�IEC and Mel18�IEC/Cdkn2a�IEC mice (Right panel). (D) Representative H&E-stained histopathologic sections of colons from WT, Cdkn2a�IEC, DKO, and B/M/C�IEC mice with DSS-induced colitis. Scale bars = 100 �m. The data are presented as the mean ± SEM (n = 3-12). ns, not significant; by two-way ANOVA (A), or by two-tailed, unpaired t-test (C).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
10
Supplementary Figure 7
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
11
Supplementary Figure 7. Dampened STAT3 Activation in Bmi1- Deficient and Mel18-Deficient IECs. (A) Immunohistochemical analysis of phospho-STAT3 in colons from WT and DKO mice at day 15 of the CAC model. (B) Western blot analysis of p-STAT3Y705, STAT3, p-ERK and p-S6K expression in control and Bmi1 deficient mice at indicated time points (Upper panel). Examination of p-STAT3Y705, STAT3, p-ERK and p-S6K expression in control and Mel18 deficient mice at indicated time points(Lower panel). β-actin was used as loading controls. Results are representative of one of three experiments. (C) Immunohistochemical analysis of phospho-STAT3 in colons from WT, Bmi1�IEC, and Mel18�IEC mice at day 15 of the CAC model. (D) Relative expression levels of cell-cycle and apoptosis regulator mRNAs isolated from IECs of WT (white bars) and DKO (black bars) mice at day 12 of the CAC model. Scale bars= 100 �m. The data are presented as the mean ± SEM.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
12
Supplementary Figure 8
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
13
Supplementary Figure 8. Reg3b and Reg3g, but not other Reg family genes, are directly targeted by BMI1 and MEL18. (A and B) ChIP analysis of the binding of the indicated antibodies at the Reg3b (A) and Reg3g (B) loci in CT26-WT cells; IgG was used as a control. (C) ChIP analysis of the binding of the BMI1 antibody at the Reg3b promoter loci in colonic epithelial tissues from control and Bmi1 deficient mice; IgG was used as a control. (D) ChIP analysis of the binding of the MEL18 antibody at the Reg3b promoter loci in colonic epithelial tissues from control and Mel18 deficient mice; IgG was used as a control. (E) ChIP analysis for the indicated antibodies at the Reg1, Reg2, Reg3a, Reg3d, and Reg4 locus, with IgG as control. (F) qRT-PCR analysis of mRNA expression for multiple PcG genes (EZH2, EED, JARID2, CBX8, M33, MEIS1, MPH2, RAE28, RING1 and SUZ12) in IECs isolated from AOM-DSS-treated mice at indicated time points. The results showed many other PcG genes were significantly upregulated during CAC tumor development. The data are presented as the mean ± SEM. *p < 0.05; by two-tailed, unpaired t-test (C,D). All data shown are representative of at least three independent experiments.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
14
Supplementary Figure 9
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
15
Supplementary Figure 9. REG3B inhibits STAT3 signaling by interfering IL-6R/GP130 receptor activation. (A) Immunoblot analysis of p-STAT3Y705, STAT3 and REG3B in lysates from Reg3b- and Reg3g-transfected 293T cells. Cells were either remained untreated or treated with IL-6 (25ng/ml). (B and C) Co-IP analysis of REG3B/IL-6R (B) and REG3B/IL-11R (C) association in CT26.WT cells. WCL: whole cell lysates, IP: Immunoprecipitation. (D) Immunoblot analysis of p-STAT3Y705 and STAT3 in lysates from CT26.WT cells with or without REG3B treatment. The recombinant REG3B protein with the indicated dose was incubated with the cells first for 10 minutes before IL-6 treatment. (E) Immunoblot analysis of p-STAT3Y705 and STAT3 in lysates from CT26.WT cells with or without REG3B treatment. The recombinant REG3B with the indicated dose was added to the media 10 minutes after IL-6 treatment. β-actin was used as loading controls. All data shown are representative of at least three independent experiments.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
16
Supplementary Figure 10
Supplementary Figure 10. REG3B treatment alleviates STAT3 activation, inhibits proliferation and promotes apoptosis of premalignant IECs. (A) Immunohistochemical analysis of Ki-67, cleaved caspase 3, and phospho-STAT3 in colonic tumor tissues from CAC mice treated with PBS or REG3B during each DSS cycle. (B) Percentage Ki-67 positive cells in control and early REG3B treat mice colonic polyps was calculated. (C) Immunohistochemical staining of Ki-67, Cleaved Caspase 3, and phospho-STAT3 in colonic tumor tissues from CAC mice treated with PBS or REG3B after the last DSS cycle. (D) Percentage Ki-67 positive cells in control and late REG3B treat mice colonic polyps was calculated. Scale bars = 100 �m. The data are presented as the mean ± SEM. *p < 0.05; **p < 0.01; by two-tailed, unpaired t-test (B,D).
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
17
Supplementary Figure 11
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
18
Supplementary Figure 11. Generation of Bmi1, Mel18 and Reg3b-deficient cell lines. (A) Immunohistochemical analysis of BMI1, MEL18, p16 and p19 in colonic tumor tissues from CAC mice treated with PBS, REG3B during each DSS cycle, and REG3B after the last DSS cycle. (B) Generation of Bmi1-/-, Mel18-/-, Reg3b-/-, B-/-/M-/-, and B-/-/M-/-/R-/- 3T3 cells by CRISPR-Cas9-mediated targeting. TA clones from the PCR products were analyzed by DNA sequencing. # indicates allele’s number. The PAM sequences are highlighted in green; the targeting sequences in red; deletions are shown in dotted line. (C) qPCR analysis of Reg3b mRNA expression in 3T3 cell lysates from WT, Bmi1-/-, Mel18-/-, and B-/-/M-/- cells. Scale bars = 100 �m. The data are presented as the mean ± SEM. All data shown are representative of at least three independent experiments.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
19
Supplementary Figure 12
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
20
Supplementary Figure 12. Reg3b deletion rescues Bmi1 and Mel18 double-deficient mice. (A) Schematic diagram of the targeting strategy for the generation of a floxed allele of Reg3b using CRISPR/Cas9. Exons are indicated by filled blocks with numbers. Red triangle, loxP site. (B) Immunoblot analysis of REG3B protein in isolated colonic enterocytes from WT and Reg3b deficient mice. β-actin was used as loading controls. (C) Immunohistochemical analysis of Ki-67 in polyps from WT, Reg3b�IEC, DKO and TKO AOM-DSS-treated mice. (D) Immunoblot analysis of p-STAT3Y705 and REG3B at day 12 of the CAC challenge. β-actin was used as loading controls. (E and F) Luminal, mucosa-associated and epithelium-associated bacteria of WT mice and TKO mice with no (E) or DSS-treated (F) were assessed by qPCR determination of 16S rRNA gene copy number in the terminal colon. Data are from three groups of littermates. Scale bars= 100 �m. The data are presented as the mean ± SEM (n= 6-9). ns, not significant; by two-tailed, unpaired t-test (E,F). Results are representative of one of three experiments.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
21
Supplementary Table 1. Genes with an Absolute Fold Change of >|10| and a P-Value <0.05 on Either of the Two Gene Lists ProbeSetID DKO vs WT fold UniGene ID Gene Symbol 1459187_at 93.0040158 Mm.455233 --- 1448649_at 66.77703442 Mm.1193 Enpep 1424445_at 63.4912169 Mm.24400 Tm4sf5 1450140_a_at 62.49522949 Mm.4733 Cdkn2a 1425668_a_at 51.86565454 Mm.275973 St3gal4 1431213_a_at 47.83291482 --- LOC433762 1422610_s_at 33.50170088 Mm.281018 Igf2bp3 1448872_at 30.22333024 Mm.252385 Reg3g 1457068_at 28.50290367 Mm.359041 Naaladl1 1455869_at 25.9633332 Mm.458283 --- 1418486_at 25.24579449 Mm.27154 Vnn1 1416297_s_at 21.97107194 Mm.2553 Reg3b 1419400_at 20.05884001 Mm.2941 Mttp 1425470_at 19.34276189 --- --- 1424303_at 19.13095368 Mm.288805 Depdc7 1429994_s_at 18.9291097 Mm.271190 Cyp2c65 1446368_at 17.10073161 --- 9130221J18Rik 1451339_at 15.50876523 Mm.23352 Suox 1447845_s_at 13.91132386 Mm.27154 Vnn1 1418069_at 13.73191334 Mm.477720 Apoc2 1419393_at 13.0528788 Mm.289590 Abcg5 1417600_at 12.48559864 Mm.281804 Slc15a2 1448783_at 12.13590804 Mm.45874 Slc7a9 1417079_s_at 11.48989453 Mm.390793 Lgals2 1455223_at 11.4788323 Mm.476844 Igf2bp1 1455996_x_at 11.15406788 Mm.439707 Prap1 1417803_at 11.11540973 Mm.45481 1110032A04Rik 1418761_at 10.89019018 Mm.476844 Igf2bp1 1448290_at 10.58261418 Mm.2553 Reg3b 1427262_at 10.50277475 Mm.274770 Xist 1417889_at 10.41223436 Mm.281793 Apobec2
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
22
Supplementary Table 2. qRT-PCR Primers for mRNA Expression Name Sense Anti-sense Bmi1 5'-AGCAATGACTGTGATGCACTTGAG-3' 5'-CTCCAGCATTCGTCAGTCCATCCC-3'
Mel18 5'-GACACTTCCCAAATCTCCTCAG-3' 5'-ACAATGGTGGTGGCATCAAT-3'
bclx-F 5'-GGTCGCATCGTGGCCTTT-3' 5'-TCCGACTCACCAATACCTGCAT-3'
cdc2 5'-TCGCATCCCACGTCAAGA-3' 5'-GTTTGGCAGGATCATAGACTAGCA-3'
Cox-2 5'-CAGCCAGGCAGCAAATCCT-3' 5'-CTTATACTGGTCAAATCCTGTGCTCA-3'
cyclin B1 5'-ACTTCAGCCTGGGTCGCC-3' 5'-ACGTCAACCTCTCCGACTTTAGA-3'
cyclin D1 5'-CCCTGACACCAATCTCCTCAAC-3' 5'-GCATGGATGGCACAATCTCCT-3'
cyclin E 5'-ATGTGGCCGTGTTTTGCA-3' 5'-GGTCTGATTTTCCGAGGCTGA-3'
Cyclin D2 5'-GCGTGCAGAAGGACATCCA-3' 5'-CACTTTTGTTCCTCACAGACCTCTAG-3'
IL-6 5'-TTGGGACTGATGCTGGTGAC-3' 5'-TTGCCATTGCACAACTCTTTTC-3'
IL-11 5'-CTGCACAGATGAGAGACAAATTCC-3' 5'-GAAGCTGCAAAGATCCCAATG-3'
p21 5'-ATTCAGAGCCACAGGCACCAT-3' 5'-TCTCCGTGACGAAGTCAAAGTT-3'
IL-1β 5'-TGGGAAACAACAGTGGTCAGG-3' 5'-CCATCAGAGGCAAGGAGGAA-3'
survivin 5'-CCTGCACCCCAGAGCGAAT-3' 5'-AGAAAAAACACTGGGCCAAATCAG-3'
TNFα 5'-TCTGTGAAGGGAATGGGTGTT-3' 5'-CAGGTCACTGTCCCAGCATC-3'
PCNA GCACGTATATGCCGAGACCT-3' 5'-CCGCCTCCTCTTCTTTATCC-3'
bcl10 5'-CTTCTCTATGGCGTCGTCCC-3' 5'-CCCTCTTCCAACCGAAGGTC-3'
bcl2 5'-CCACCTGTGGTCCATCTGAC-3' 5'-CAATCCTCCCCCAGTTCACC-3'
c-Myc 5'-CGACTACGACTCCGTACAGC-3' 5'-GTAGCGACCGCAACATAGGA-3'
Reg1 5'-AAGAAGACCTGCCATCTGCC-3' 5'-GCCAGCGACGATTCCTTTTG-3'
Reg2 5'-AATCAACTGCCCAGAGGGTG-3' 5'-AGTGCCAACGACGGTTACTT-3'
Reg3a 5'-ACAAGGCTTATCGCTCCCAC-3' 5'-TGGGTTGTTGACCCATTGTTG-3'
Reg3b 5'-GAATATACCCTCCGCACGCA-3' 5'-GGTCATGGAGCCCAATCCAA-3'
Reg3d 5'-ACTGTGTTGCCTGATGTCCC-3' 5'-TGGCAGTGGATCTCTGCATT-3'
Reg3g 5'-TGCAAGGTGAAGTTGCCAAG-3' 5'-GGTTCATAGCCCAGTGTCGG-3'
Reg4 5'-TCATGCTGAGCTGGAGTGTC-3' 5'-GCGGTTGGCACATCCATTTT-3'
EZH2 5'-TTCGTGCCCTTGTGTGATAG-3' 5'-GGAAAGCGGTTTTGACACTC-3'
EED 5'-TGGCCATGGAAATGCTATC-3' 5'-CCTCCAAATATTGCCACCAG-3'
SUZ12 5'-TGCAGTTCACTCTTCGTTGG-3' 5'-GAACCAGGCTTGTTTTCCTG-3'
Jarid2 5'-ACAATGCTTCATCTTCGTGCC-3' 5'-GCTCTTTCTCCCGTGCTGAC-3'
Meis1 5'-CCTCTGCACTCGCATCAGTAC-3' 5'-GTTTGGCGAACACCGCTATATC-3'
Ring1 5'-CAGAATGCCAGCAAAACGTGGG-3' 5'-TTCTTCAGCATGTCCAGGCAGAT-3'
Cbx8 5'-GTTCGCGGCCGAAGCCCTCC-3' 5'-TGGGGCCATAGAGCTCCATCTC-3'
M33 5'-CAACAAAGGGGAAAAGCTGA-3' 5'-ACATCCGTGACAAAGACGTG-3'
Rae28 5'-GGCAGAAGCAGATGGGAGTG-3' 5'-GAGGGGCAGTGAGGGTTGTT-3'
Mph2 5'-CACTGGCATCTCCAGGTTTT-3' 5'-GAGGTATGGGGAAGGGGTTA-3'
GAPDH 5'-ATGACATCAAGAAGGTGGTGAAG-3' 5'-TCCTTGGAGGCCATGTAGG-3'
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
23
Supplementary Table 3. qPCR Primers for ChIP Used in the Study Name Sense Anti-sense Reg3b-CHIP1 5'-GTGTCTGGAAGAGGGTGTGGATGGT-3' 5'-TCCCGTTAGCCGACCTGCTTT-3'
Reg3b-CHIP2 5'-ATGCTGAAAGGTAGGCATGGGTTCT-3' 5'-AGACACCTGTCCTGTGTGTGAGA-3'
Reg3g-CHIP1 5'-TGGCAGGAAGAAACTCCCGCT-3' 5'-GGGAGGTAGGTGGTGCTTGGC-3'
Reg3g-CHIP2 5'-GGCTGCTAACAGGAAAGGGCCA-3' 5'-AGCGGGAGTTTCTTCCTGCCA-3'
Reg1-CHIP 5'-GGGCACTTGTTTGCTTCTGG-3' 5'-AACCTGGCAGTAGTCACAGC-3'
Reg2-CHIP 5'-AGTGGAGGCCCAAATTGCTT-3' 5'-AGGCTCCCTTGTATCCCCAT-3'
Reg3a-CHIP 5'-GGCAAAAGTGAAGGGAGGGA-3' 5'-ACAACAAACAGGGCACCAGA-3'
Reg3d-CHIP 5'-CAGGAGCTGGCACTTCTTCA-3' 5'-ATACCTGCCAGCCTTGCATT-3'
Reg4-CHIP 5'-GACAACTCGGGAACTCTGGG-3' 5'-GACAGGGGCAGAAGTGTTGA-3'
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
24
Supplementary Data
REG3B Is Required for Initiation of CAC Tumor
REG3B has antibacterial effects and recent studies suggest that it protects mice
against dissemination of Salmonella enteritidis and alcoholic steatohepatitis.1-3 To test
whether the effects of REG3B on STAT3 activity could be caused indirectly by its
anti-bacterial function, we analyzed the bacterial load in colonic epithelial lysates of
DSS-treated WT and TKO animals. The bacterial load was measured by total 16S
rRNAs using quantitative polymerase chain reaction (Q-PCR) in a
culture-independent manner, as described previously.1, 4 Upon DSS treatment, the
TKO mice indeed exhibited elevated phospho-STAT3 in the epithelial lysates
(Supplementary Figure 12D). However, consistent with findings by others,1 we
found that the bacteria load in lumen, mucus layer, or epithelial cells of the colon did
not significantly differ between TKO and cohoused WT littermates, under either
normal or DSS-treated conditions (Supplementary Figure 12E and F). These
observations suggest that the anti-STAT3 activity of REG3B, at least at the early
stages of tumor development, is likely not due to its impact on microbiota.
Supplementary Materials and Methods
Experimental Animals
The Bmi1 (Bmi1F/F) and Mel18 (Mel18F/F) conditional knockout mice were generated
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
25
by homologous recombination (depicted in Supplementary Figure 1A and D). The
8.9-kb genomic DNA fragment containing Bmi1 exons 2-10 and 8.6-kb genomic DNA
fragment containing Mel18 exons 5-13 were subcloned from C57BL/6 BAC library
using gap repair and cloned into PL253 vector.5 Final Bmi1 and Mel18 targeting
vector linearized by NotI were electroporated into 129xC57BL/6 ES cells. The
neomycin-resistant colonies were screened by southern blot according to the strategy
outlined in Supplementary Figure 1. Southern blot of Bmi1 was performed by using
the following probes: 3’ probe was amplified by forward primer:
GGGACCCAGTCAACA and reverse primer: CAAGAGGCTAAATGAGAT.
Southern blot of Mel18 was performed by using the following probes: 5’ probe was
amplified by the forward primer, AGCTGTCCCTGCGTCTT and reverse primer,
GGTGCTCTTACCCGTTG. Targeted ES clones were microinjected into C57BL/6
blastocysts to generate chimeric mice. Germline male chimeras were mated to
C57BL/6 females, and their agouti offspring were tested by PCR to confirm germ-line
transmission of the conditional allele. The Bmi1 conditional knockout mouse was
genotyped with the set of primers - Forward-WT: AAAATGGACATACCCAATAC,
Reverse-targeted: GTATAGCAT ACATTATACGAAG and Reverse -WT:
CATTTAAATTTGAAAATCTG. These primers yield a 134 bp band for floxed allele
and a 70 bp band for wild type locus. The Mel18 conditional knockout mouse was
genotyped with the set of primers - Forward-WT: CTGGGACAGCATCAGAAACCT,
Reverse-targeted: GTACCTGAC TGATGAAGTTCCTATA, and Reverse-WT:
CCCAACACAGAAGGTGCACCA. These primers yield a 113 bp band for floxed
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
26
allele and a 178 bp band for wild type locus.
The Reg3b conditional knockout mice were generated by co-microinjection of in
vitro-translated Cas9 mRNA, sgRNA and target vector into the C57BL/6 zygotes
(depicted in Supplementary Figure 9C). Founders were verified by sequencing of the
PCR fragments. The sgRNA sequence used to generate the Reg3b-CKO mice is
AAGCCTCTCTGCCCAGTGTT.
Reagents
Recombinant mouse REG3B (#51153-M08H) was purchased from Sino Biological.
Recombinant human IL-6 (#200-06) and IL-11 (#200-11) were purchased from
Peprotech. Recombinant human REG3β was expressed and purified from human
cells.
Histology and Immunohistochemistry
For histology, colons were excised and flushed gently with cold phosphate buffered
saline (PBS) and fixed with 10% neutral formalin overnight, embedded as Swiss rolls
in paraffin. Every 40 sections (5µm) of serial cuts were stained with hematoxylin and
eosin (H&E).6 Immunohistochemistry was performed according to the manufacturers’
protocols. The signals were developed using the ABC kit (Vector Laboratories)
according to manufacturers’ protocol. Counterstaining was done with hematoxylin.
All magnifications in legends represent objective lens. The following antibodies were
used: anti-BrdU (ab6326, 1:200, Abcam), anti-mouse Ki-67 (12202, 1:400, Cell
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
27
Signaling Technology), anti-mouse phospho-Stat3(Tyr705) (9145, 1:400, Cell
Signaling Technology), anti-human and mouse cleaved caspase-3 (9661, 1:300, Cell
Signaling Technology), anti-mouse Reg3b (AF5110, 1:500, R&D Systems),
anti-human and mouse Chromogranin A (1782-1, 1:200, Epitomics), and horseradish
peroxidase (HRP)–conjugated anti–rabbit IgG, anti–sheep IgG, and anti–mouse IgG
(ZSJQ-BIO). Polyps counts and immunohistochemistry analysis were performed in a
blinded fashion.
Generation of CRISPR-Cas9 Knockout Cell Lines
The CRISPRs designs were performed as described previously.7 SgRNAs were
designed and cloned into pX330 CRISPR/Cas9 vector (Addgene). The target
sequences used for Bmi1: ATGGCTCCAATGAAGACCGA; Mel18:
ACCTGCATCGTACGCTACTT, and Reg3b: TCTGTGCTCAATAGCGCTGA. To
construct the knockout cell lines, sgRNA-expressing plasmid with GFP, Cas9 plasmid
with RFP, and puromycin vectors were co-transfected into NIH3T3 cells. Selected
cells were sorted into single clones into the 96-well plate by flow cytometry using the
BD Biosciences FACSAria II. Single clones were screened by T7 endonuclease
I-cutting assay and the candidate knockout clones were verified by sequencing of the
PCR fragments.
Cell Culture, Lentivirus Production and Infection
CT26.WT, 3T3 and 293T cells were obtained from ATCC. All cells were cultured in
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
28
DMEM medium with 10% fetal bovine serum and 1% penicillin-streptomycin. All the
cell lines have been tested to be mycoplasma negative by the commonly used PCR
method and were authenticated by ATCC prior to purchase with their standard short
tandem repeat DNA typing methodology. All cells were grown at 37℃ in a 5% CO2
incubator. For over-expressing experiments, the full-length coding cDNAs of
candidate genes were cloned into the modified lentiviral vector pWPI. Constructed
pWPI plasmids were introduced to 293FT cells together with packaging plasmid
psPAX2 and envelope plasmid pMD2.G with a ratio of 5:3:2. After 48-50h
transfection, the medium containing lentiviruses was collected and centrifuged at
3000rpm for 5-10 min. Stable over-expressing cell lines were generated by infecting
HEK293T or CT26.WT cells with lentivirus in the presence of puromycin selection
(6µg/ml).
Western Blot Analysis
Protein was extracted from cell or tissue samples by standard methods using RIPA
lysis buffer (150 mM sodium chloride, 1.0% (vol/vol) Triton X-100, 0.5% sodium
deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0) containing protease inhibitors
(Complete Protease Inhibitor Cocktail Tablets, Roche) and phosphatase inhibitors
(PhosSTOP Phosphatase Inhibitor Cocktail Tablets, Roche). Solubilized proteins (30
µg) were separated by standard SDS-PAGE on a 12% polyacrylamide separating gel
with 5% stacking gel and then transferred to PVDF western blotting membrane
(Roche) by standard methods. The following primary antibodies were used:
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
29
anti-human and mouse IL6R (ab83053, 1:500, Abcam), anti-human and mouse
IL-11R (ab125015, 1:1,000, Abcam), anti-mouse p19 (ab80, 1:1,000, Abcam),
anti-human and mouse phospho-Stat3(Tyr705) (9145, 1:1,000, Cell Signaling
Technology), anti-mouse Stat3 (9132, 1:2,000, Cell Signaling Technology),
anti-human and mouse phospho-ERK (4370, 1:1,000, Cell Signaling Technology),
anti-human and mouse ERK (9102, 1:2,000, Cell Signaling Technology), anti-human
and mouse phospho-S6 (2211, 1:1,000, Cell Signaling Technology), anti-human and
mouse p16 (4824, 1:500, Cell Signaling Technology), anti-human and mouse Jak2
(3230, 1:1,000, Cell Signaling Technology), anti-human and mouse SOCS3 (2932,
1:1,000, Cell Signaling Technology), anti-mouse Reg3b (AF5110, 1:1,000, R&D
Systems), anti-human and mouse GP130 (SC-655, 1:250, Santa Cruz), anti-human
and mouse IL-6 (10395-MM02, 1:1,000, Sino Biological), anti-Flag M2 (F1804,
1:1,000, Sigma), anti-mouse Phosphotyrosine (05-947, 1:500, Millipore), and
anti-mouse β-actin (MAB1501, 1:20,000, Millipore). Secondary antibodies include
the following: HRP-conjugated goat anti-rabbit HRP (ZDR-5403, ZSJQ-BIO), goat
anti-mouse HRP (ZDR-5402, ZSJQ-BIO) and rabbit anti-goat HRP (ZDR-5105,
ZSJQ-BIO), all secondary antibodies were used at 1:5,000. Blots were visualized with
Immobilon Western Chemiluminescent HRP Substrate kit (Millipore) according to the
manufacturer’s instructions.
TUNEL Assay
TUNEL assay was performed using a fluorescein In Situ Cell Death Detection Kit
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
30
(Roche) following the manufacturer’s instructions. Sections were counterstained with
4',6-diamidino- Microarray Analysis 2-phenylindole (DAPI).
Isolation of RNA and Quantitative Real Time PCR (qRT-PCR)
For quantitative RT-PCR, tissue was homogenized using a rotor stator in TRIZOL
(Life Technologies), and total RNA was isolated according to the manufacturer’s
instructions. RNA was digested with DNaseI and purified using an RNeasy Mini kit
(Qiagen). The RNAs were reverse-transcribed to cDNA using the HiScript II 1st
Strand cDNA Synthesis Kit (Vazyme, R211-02), and the cDNA was mixed with
ChamQ SYBR qPCR master mix reagents (Vazyme, Q331-02). Real-time PCR was
performed using the ABI 7500 Fast Real-Time PCR System (Applied Biosystems).
The abundance of each cytokine mRNA was normalized to that of
glyceraldehyde-3-phosphate dehydrogenase (Gapdh) and compared to the levels in
WT intestines to calculate the fold induction. The qRT-PCR primers used in the study
are listed in Supplementary Table 2.
Isolation of Bacterial DNA and Quantitative PCR (Q-PCR) for Microbiota Analysis
Isolation of genomic DNA was essentially performed as described previously.4 Briefly,
the distal 5 cm of colon was isolated. The luminal contents were collected by flushing
with 2 ml sterile PBS. The remaining colons were cut longitudinally and washed
vigorously in 2 ml sterile PBS to collect the mucus and the bacteria that were adherent
to the mucosal surface. The remaining part of the colon was cut for isolating genomic
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
31
DNA. Genomic DNA was isolated from the pellets of lumen, mucus and epithelial
cell layer with the Qiagen Stool Kit.
Q-PCR analysis of bacterial genomic DNA was performed using the universal 16S
rRNA gene primers as described previously.1, 4 Bacterial 16S rRNA gene universal
primers-forward: ACTCCTACGGGAGGCAGCAGT, reverse: ATTACCGCGGCTGC
TGGC.
Co-Immunoprecipitation (Co-IP) Assay
Stable cell lines and mice IECs were lysed in NP-40 lysis buffer containing protease
inhibitors. The samples and controls lysates were incubated with anti-Flag resin
(A2220, Sigma) overnight with gentle shaking at 4℃. Protein complexes were then
washed three times in NP-40 lysis buffer and eluted with 3X FLAG peptide (F4799,
Sigma) for western detection. Anti-Flag M2 antibody (F1804, 1:1,000, Sigma) was
used to detect Flag-tagged REG3B.
Chromatin Immunoprecipitation (ChIP) Assays
The cells or tissue samples were fixed with 2% paraformaldehyde to crosslink the
DNA with bound proteins. CHIP assays were performed using a Chromatin
Immunoprecipitation (ChIP) Assay Kit (Millipore) following the manufacturer’s
instructions. The antibodies used for this study were anti-Bmi1 (ab14389, Abcam),
anti-Mel18 (ab5267,Abcam), anti-Ub-H2A (05-678, Millipore), and normal mouse or
rabbit IgG (Cell Signaling Technology) as a control. For CHIP primer design, we first
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
32
identify the gene promoter sequences using the NCBI database, and then search the
predicted binding site using the JASPAR database. We used PrimerPremier software
to design the random primers flanking the putative binding site sequences. The
resulting DNA was quantified and served as a template for the qRT-PCR reactions,
which were performed using the SYBR Premix Ex Taq reagents kit (Takara) and ABI
7500 Fast Real-Time PCR System (Applied Biosystems). The DNA enrichment after
ChIP was normalized using the fold enrichment method. All results were obtained
from three independent biological replicates. The sequence information for all the
primers are listed in Supplementary Table 3.
Microarray Analysis
Whole RNA was extracted using Trizol reagent (Invitrogen) from mouse IECs.
Microarray experiments were performed by the Beijing CapitalBio Corporation.
Material was processed for Affymetrix Mouse GeneChip 1.0 ST arrays (Affymetrix,
Inc.). For each experimental condition, three mice were used. The raw image files
were processed by Bioconductor “Affy” package, the intensity of probes in arrays
passed quality control were calculated and normalized for expression value. A fold
change of 2 and q value (FDR-corrected P value) of 0.05 were used as cutoffs.
Human Colon Adenocarcinoma Tissue Microarray Analysis
The study was ratified by the Institutional Review Board of the National Institute of
Biological Sciences (NIBS). Protein levels of REG3β expression in tissues were
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
33
determined using tissue microarray (TMA). Matched pairs of primary human colon
adenocarcinoma samples and adjacent normal tissues from 87 patients with survival
information and tumor grade information and primary human colon adenocarcinoma
samples from 92 patients were used for the construction of a tissue microarray
(Shanghai Outdo Biotech Co., Ltd.). Sections of tissue microarray were stained with
anti-human Reg3β antibody (ab134309, 1:500, Abcam) and anti-human and mouse
phospho-Stat3 (Tyr705) (9145, 1:400, Cell Signaling Technology). Positive staining
was scored as follows: high, large area staining; moderate, staining of multiple
smaller areas; low, staining of scattered few positive cells.
Supplementary References
1. Wang L, Fouts DE, Starkel P, et al. Intestinal REG3 Lectins Protect against Alcoholic
Steatohepatitis by Reducing Mucosa-Associated Microbiota and Preventing Bacterial
Translocation. Cell Host Microbe 2016;19:227-39.
2. van Ampting MT, Loonen LM, Schonewille AJ, et al. Intestinally secreted C-type lectin Reg3b
attenuates salmonellosis but not listeriosis in mice. Infect Immun 2012;80:1115-20.
3. Miki T, Holst O, Hardt WD. The bactericidal activity of the C-type lectin RegIIIbeta against
Gram-negative bacteria involves binding to lipid A. J Biol Chem 2012;287:34844-55.
4. Vaishnava S, Yamamoto M, Severson KM, et al. The antibacterial lectin RegIIIgamma
promotes the spatial segregation of microbiota and host in the intestine. Science
2011;334:255-8.
5. Liu P, Jenkins NA, Copeland NG. A highly efficient recombineering-based method for
generating conditional knockout mutations. Genome Res 2003;13:476-84.
6. Greten FR, Eckmann L, Greten TF, et al. IKKbeta links inflammation and tumorigenesis in a
mouse model of colitis-associated cancer. Cell 2004;118:285-96.
7. Platt RJ, Chen S, Zhou Y, et al. CRISPR-Cas9 knockin mice for genome editing and cancer
modeling. Cell 2014;159:440-55.
Authors names in bold designate shared co-first authorship.
