A RIPK3 PGE2 circuit mediates myeloid derived suppressor...4 1. the correlation of . RIPK3. and indicated genes and survival, respectively. 2. 3. Animal experiments. 4. C57BL/6 mice

1

A RIPK3-PGE2 circuit mediates myeloid-derived suppressor 1

cell-potentiated colorectal carcinogenesis 2

3

Guifang Yan1,2,#

, Huakan Zhao1,2,#

, Qi Zhang1,2,#

, Yu Zhou1,2

, Lei Wu1,2

, Juan Lei1,2

, Xiang Wang1,2

, 4

Jiangang Zhang1,2

, Xiao Zhang1,2

, Lu Zheng3, Guangsheng Du

4, Weidong Xiao

4, Bo Tang

5, 5

Hongming Miao6,*

, Yongsheng Li1,2,*

6

7 1Institute of Cancer,

2Clinical Medicine Research Center,

3Department of Hepatobiliary Surgery, 8

4Department of General Surgery,

5Department of Gastroenterology, Xinqiao Hospital, Third 9

Military Medical University, Chongqing 400037, China. 10 6Department of Biochemistry and Molecular Biology, Third Military Medical University, 11

Chongqing 400038, China 12

13 #

These authors contributed equally to this work. 14 *Corresponding Authors: Hongming Miao ([email protected]) and Yongsheng Li 15

([email protected]). 16

17

Running title: RIPK3-PGE2 circuit regulates MDSC. 18

19

Research. on September 18, 2020. © 2018 American Association for Cancercancerres.aacrjournals.org Downloaded from

Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on July 16, 2018; DOI: 10.1158/0008-5472.CAN-17-3962

http://cancerres.aacrjournals.org/

2

ABSTRACT 1

Receptor-interacting protein kinase 3 (RIPK3) is essential for mucosal repair in inflammatory 2

bowel diseases (IBD) and colorectal cancer (CRC). However, its role in tumor immunity is 3

unknown. Here we report that decreased RIPK3 in CRC correlates with the accumulation of 4

myeloid-derived suppressor cells (MDSC). Deficiency of RIPK3 boosted tumorigenesis via 5

accumulation and immunosuppressive activity of MDSC. Reduction of RIPK3 in MDSC and CRC 6

cells elicited nuclear factor kappa B (NF-κB)-transcribed cyclooxygenase-2 (COX-2), which 7

catalyzed the synthesis of prostaglandin E2 (PGE2). PGE2 exacerbated the immunosuppressive 8

activity of MDSC and accelerated tumor growth. Moreover, PGE2 suppressed RIPK3 expression 9

while enhancing expression of NF-κB and COX-2 in MDSC and CRC cells. Inhibition of COX-2 10

or PGE2 receptors reversed the immunosuppressive activity of MDSC and dampened 11

tumorigenesis. Patient databases also delineated the correlation of RIPK3 and COX-2 expression 12

with CRC survival. Our findings demonstrate a novel signaling circuit by which RIPK3 and PGE2 13

regulate tumor immunity, providing potential ideas for immunotherapy against CRC. 14

15

Keywords: RIPK3; necroptosis; colorectal cancer; MDSC; COX-2; PGE2 16

17

Significance: 18

A novel signaling circuit involving RIPK3 and PGE2 enhances accumulation and 19

immunosuppressive activity of MDSC, implicating its potential as a therapeutic target in 20

anticancer immunotherapy. 21

22




3

Introduction 1

Colorectal cancer (CRC) is one of the most common malignant tumors and the third leading 2

cause of cancer-associated mortality in the world (1). The 5-year survival rate is only ~10% for the 3

patients with advanced CRC (2). Exploring the pathogenesis and effective therapeutic targets of 4

CRC is of great clinical significance. Inflammatory bowel diseases (IBD) are recognized as 5

precancerous diseases of CRC. The CRC-infiltrating immune cells, including myeloid-derived 6

suppressor cells (MDSC) promote the carcinogenesis, which is one of the most important causes 7

for tumor progression and therapeutic failure (3-5). MDSC can induce immune tolerance by 8

overexpressing arginase 1 (Arg-1), inducible nitric oxide synthase (iNOS or NOS2), and reactive 9

oxygen species (ROS) to suppress the activation of cytotoxic T lymphocytes (CTL). MDSC can 10

also differentiate toward tumor-associated macrophages (TAMs) and promote the 11

immunosuppressive function of regulatory T cells (Tregs). In addition, MDSC secrete 12

prostaglandin E2 (PGE2), calcium-binding protein S100A8/A9, fibroblast growth factors (FGFs), 13

matrix metalloproteinases (MMPs), transforming growth factor β (TGF-β), vascular endothelial 14

growth factor (VEGF) and other cytokines to promote tumor proliferation, angiogenesis and 15

metastasis (6). Therefore, addressing mechanisms regulating MDSC will provide new ideas for the 16

immunotherapy of CRC. 17

Inflammation initiates necroptosis which parallels with caspases-mediated apoptosis and 18

nuclear factor kappa B (NF-κB)-mediated proliferation and plays an essential role in 19

carcinogenesis (7). Receptor-interacting protein kinase 3 (RIPK3) is a central regulatory molecule 20

for necroptosis (8), whereas its role in tumor immunity remains unknown. It has been reported that 21

RIPK3 promotes the mucosal repair in IBD (9). More importantly, RIPK3 also inhibits the 22

tumorigenesis of CRC and the expression of proinflammatory factors including S100A8, 23

chemokine (C-X-C motif) ligand 1 (CXCL1), interleukin (IL)-1β, IL-6, and tumor necrosis factor 24

alpha (TNFα) (10). Since these proinflammatory factors correlate with the accumulation and 25

maintenance of MDSC (11), the above studies suggest that RIPK3 may regulate the 26

tumor-infiltrating MDSC. 27

Here, we demonstrate that the downregulation of RIPK3 in tumor-infiltrating MDSC 28

potentiates NF-κB activation and cyclooxygenase-2 (COX-2)-derived PGE2 production. PGE2 in 29

turn further reduces RIPK3 and promotes the immunosuppressive activity of MDSC and 30

carcinogenesis. Therapy targeting this signaling circuit involving RIPK3 and PGE2 potently blunts 31

the accumulation and activity of MDSC and protects colorectal against malignancy. Our data 32

provide molecular basis for RIPK3 regulating MDSC and tumor immunity, and suggest potential 33

immunotherapeutic idea for CRC. 34

35

Methods 36

Human databases 37

The correlations between RIPK3 gene expression and CRC were determined through analysis 38

of Kaiser colon and Skrzypczak colorectal cancer datasets, respectively, which are available at 39

Oncomine (http://www.oncomine.org/). 40

The National Center for Biotechnology Information Gene Expression Omnibus databases 41

GSE21510 (12) and GSE17536 (13) containing 148 and 177 patients with CRC were evaluated for 42




4

the correlation of RIPK3 and indicated genes and survival, respectively. 1

2

Animal experiments 3

C57BL/6 mice were purchased from the Chinese Academy of Medical Sciences (Beijing, 4

China). RIPK3 knockout (KO) mice with C57BL/6 background were kindly provided by 5

Xiaodong Wang and Zhirong Shen (National Institute of Biological Sciences, Beijing, China). All 6

wildtype (WT) and KO mice were age and sex matched, and cages were randomly assigned to the 7

treatment groups. All animal procedures were conducted in accordance with the national and 8

international Guidelines for the Care and Use of Laboratory, approved by the Animal Care and 9

Use Committee of Third Military Medical University (Chongqing, China) and complied with the 10

Declaration of Helsinski. 11

The model of acute IBD was established by feeding C57BL/6 mice with 2% DSS dissolved 12

in sterile pure water for 6 days. For CRC induction, 6 week C57BL/6 mice were injected 13

intraperitoneally with 10mg/kg azoxymethane (AOM, Cat No. 25843-45-2, Sigma-Aldrich) and 14

after 7 days, they were fed with sterile pure water containing 2% dextran sodium sulfate (DSS, Cat 15

No. 0216011080, MW 40,000-50,000, MPbio) for 3 cycles. Colons and spleens were removed 16

upon sacrifice at indicated interval. Macroscopic tumors were measured with calipers. Portions of 17

the distal colons were either frozen in -80℃ or fixed with formaldehyde and paraffin embedded 18

for histological analysis. For some indicated experiments, GSK872 (0.75 mg/kg, Cat No. 2673, 19

Biovision), AH-6809 (5 mg/kg, Cat No. HY-10418, Med Chem Express), or ONO-AE3-208 (5 20

mg/kg, Cat No. 402473-54-5, Med Chem Express) was injected i.p. every 2 days until the mice 21

were sacrificed. Anti-Gr-1 (12.5mg/kg, Cat No. BE0075, Bioxcell) or CXCR2 antagonist 22

(CXCR2-a) SB225002 (4 mg/kg, Cat No. 182498-32-4, Med Chem Express) was injected i.p. 23

every 2 days from the third cycle until the mice were sacrificed. In the in vivo aspirin (ASA) 24

treatment experiments, mice were subjected 0.02% ASA (Cat No. 50-78-2, Med Chem Express) 25

containing water during the CRC induction. 26

Body weight, stool consistency and rectal bleeding were monitored daily. The values before 27

DSS exposure were recorded as baseline. The diarrhea scores were calculated as follows: 0-stool 28

formed pellets; 1-diarrhea; 2-hematochezia; 3-serious hematochezia or archoptosis; 4-die. The 29

average total scores were calculated in each cycle for 21 days. 30

For the chimerism experiments, C57BL/6 WT or RIPK3-KO mice were irradiated (850 cGy) 31

and injected i.v. with 10×106 bone marrow (BM) cells from congenic WT or RIPK3-KO mice, 32

respectively. For two weeks after engraftment, mice were given antibiotic water (containing 33

Trimethoprim and Sulfamethoxazole). After 7 weeks, the peripheral MDSC were analyzed to 34

confirm chimerism. 35

36

Flow cytometry (FCM) 37

The single cell suspension was prepared by mechanic dispersion and enzymatic digestion of 38

indicated tissues. For extracellular staining of target proteins, cells (1×106/ml) were preincubated 39

in a mixture of PBS, 1% fetal bovine serum (FBS), and 0.1% (w/v) sodium azide with 40

FcgIII/IIR-specific antibody to block nonspecific binding and stained with different combinations 41

of fluorochrome-coupled antibodies including CD45 (Cat No. 103108), CD11b (Cat No. 42

101208/101224), Gr-1 (Cat No. 108426), Ly6G (Cat No. 127618), Ly6C (Cat No. 128007), F4/80 43

(Cat No. 123110), CD11c (Cat No. 117308), CD206 (Cat No. 141704), CD3 (Cat No. 44




5

100206),CD8α (Cat No. 301030), CD4 (Cat No. 100422), CD33 (Cat No. 303304), HLA-DR (Cat 1

No. 307606), and APC Annexin V Apoptosis Detection Kit with 7-AAD (Cat No. 640930) from 2

Biolegend, Fixable Viability Dye EfluorTM

780 (Cat No. 65-0865-14) from eBioscience, ROS (Cat 3

No. s0033-1) from Beyotime. For intracellular staining of RIPK3 (Cat No. 95702, Cell Signaling 4

Technology; ab152130, Abcam), COX2 (Cat No. 12282, Cell Signaling Technology), Arg-1 (Cat 5

No. 42284, GeneTex), NOS2 (MA5-17139, Thermo), interferon-γ (IFN-γ, Cat No. 505810, 6

Biolegend) and granzyme B (GzmB, Cat No. 515403, Biolegend), we followed the manufacturers’ 7

protocols after cells were treated with PGE2 (Cat No. 363-24-6, Cayman Chemical), GSK872 (Cat 8

No. 2673, Biovision), AH-6809 (Cat No. HY-10418, Med Chem Express), Caffeic Acid Phenethyl 9

Ester (Cat No. S7414, Selleck), ASA (Cat No. 50-78-2, Med Chem Express), 10

N-Hydroxy-nor-L-arginine (NHNL, Cat No. 399275, Calbiochem), Bevacizumab (Bev, Roche), 11

Cetuximab (Cet, Merck), Nimotuzumab (Nim, Biotech Pharma), irinotecan (CPT-11, Pfizer), 12

Oxaliplatin (OXA, Cat No. HY-17371, Med Chem Express), 5-Fluorouracil (5-FU, Cat No. 13

HY-9006, Med Chem Express) or gemcitabine (GEM, Eli Lilly and Company), respectively. The 14

fluorescence were determined on a FACS Canto II system (BD Biosciences) and analyzed with 15

FlowJo software (Tree Star). 16

17

MDSC induction in vitro 18

MDSC were isolated and induced as indicated previously (14,15). Briefly, BM cells from WT 19

or RIPK3-/-

mice were stained by anti-mouse Gr-1 particles (Cat No.558111, BD) which were 20

optimized for positive selection and collected together using the BD IMagTM

Cell Separation 21

Magnet. Fresh harvested MDSC were incubated in RPMI1640 medium containing 5% FBS with 22

GM-CSF (20 ng/ml, 315-03, Peprotech) for 48h. 23

24

Cell culture 25

The mouse colorectal cancer CT26 cell line was purchased from ATCC (Manassas, VA) and 26

authenticated via STR profiling. Cells were cultured in RPMI-1640 (Gibco) supplemented with 10% 27

FBS (Gibco) and 1% penicillin-streptomycin (Gibco). The CT26 cells were routinely verified 28

mycoplasma-free using MycAwayTM-Color One-Step Mycoplasma Detection Kit (Yeasen 29

Bio-technol) and the most recent date of testing was April 5, 2018. Cells were used within 12 30

passages following thawing in all experiments. 31

32

CD8+ T cell isolation, purification and proliferation assay 33

CD8+ T cells were isolated from the spleen of C57BL/6 mice by CD8

+ T cell Isolation Kit 34

(Cat No. 480007, Biolegend). BM-derived MDSC were co-cultured with CFDA-SE 35

(56-carboxyfluorescein diacetatesuccinimidyl ester, CFSE; Cat No. 2011-11-2, Dojindo, 36

Kumamoro, Japan) labeled CD8+ T cells (1×10

6) at 10:1 in the medium containing anti-CD3 (1 37

μg/ml) and anti-CD28 (1 μg/ml). At day 3 post co-cultivation, cells were harvested and 38

CFSE+CD8

+ T cells were detected by FCM. 39

For the analysis of CD8+ T cell function, CD8

+ T cells were co-cultured with MDSC (5:1, 40

10:1 or 16:1) and were harvested to stimulate using Cell Stimulated Cocktail (Cat No. 4303372, 41

eBioscience) for 4 hrs and then were collected for the determination of GzmB and IFN-γ by FCM. 42

43

Western blot 44




6

Samples were lysised in RIPA buffer containing PMSF. The protein quantification was 1

determined by BCA protein assay (Cat No. P0068, Beyotime, China), and equal amounts of 2

proteins (40 μg) were subjected to SDS/PAGE (12% gels). After electrophoresis, proteins were 3

transferred onto PVDF membranes (0.45mm) in running buffer with 20% methanol. Non-specific 4

sites were blocked with 5% (w/v) non-fat dried skimmed milk powder in TBST (2M Tris-HCl 5

buffer, pH 7.6; 0.05 M NaCl; and 0.05% Tween-20) for 60 min at 37℃. The membranes were then 6

incubated overnight at 4°C with the following antibodies, which were diluted in TBST: 7

anti-RIPK3 (1:1000; Cat No. 2283, ProSci), anti-COX2 (1:500; Cat No. 12282, Cell Signaling 8

Technology), anti-p65 (1:500; Cat No. 41556, Gene Tax), anti-PKA (1:1000; Cat No. ab76238, 9

Abcam), anti-Actin (1:1000; Cat No. A1978, Sigma-Aldrich), Cell Signaling Technology 10

antibodies including anti-CREB (1:500; Cat No. 9197), anti-p-CREB (1:500; Cat No. 9198), 11

anti-p-stat3 (1:1000; Cat No. 9145), anti-stat3 (1:1000; Cat No. 4904), anti-p-stat6 (1:1000; Cat 12

No. 56554s), and anti-stat6 (1:1000; Cat No. 5397). After four washes in TBST, the membranes 13

were incubated with horseradish-peroxidase conjugated secondary antibodies (Cat No. A0562, 14

Beyotime) for 1 h in TBST (dilution of 1:5000). Protein bands were visualized by using Enhanced 15

Chemiluminescence (ECL) Plus Western blotting detection kit (Cat No. P0018-2, Beyotime). 16

17

Confocal microscopy. 18

Mice or human tissues were fixed and permeabilized with Fixation & Permeabilization 19

Buffers (BD Biosciences) for 15min and then incubated with FC-block (BD Biosciences) for 30 20

min at room temperature. Subsequently, cells were stained with Gr-1 (1:50, Cat No. MAB1037, 21

R&D Systems), RIPK3 (1:100, Cat No. 95702, Cell Signaling Technology), or CD33 (1:100, Cat 22

No. ab213050, Abcam), RIPK3 (1:50, Cat No. ab152130, Abcam) for overnight at 4 °C, washed 23

thrice with PBS before incubation with fluorochrome associated secondary antibodies for 30min 24

with Alexa-488, and 647 (Bioss). Afterwards, sections were washed thrice with wash buffer (BD 25

Biosciences), and then were incubated with DAPI and mounted on slides using Prolong Gold 26

antifade reagent (Beyotime). The sections were imaged with a Leica TCS SP5 laser scanning 27

confocal microscope (Leica Microsystems). The co-localization and average intensity were 28

assessed by using Leica LASX (Microsystems software). 29

30

Immunohistochemistry analysis 31

Colon and tumor tissues were fixed with formaldehyde. Paraffin sections were stained with 32

hematoxylin and eosin or subjected to immunohistochemistry for Gr-1 (Cat No. MAB1037, R&D 33

Systems), RIPK3 (Cat No. 2283, ProSci) and COX-2 (Cat No. 12282, Cell Signaling Technology). 34

35

Real-time PCR 36

Real-time PCR was performed as previously (16). Total RNA was extracted from cells with 37

RNA queousTM

Mico kit (Cat No. 00490515, Invitrogen). Real-time quantitative PCR was 38

performed on a CFX384TM

system (BIO-RAD). Primers used in this study are below: RIPK3-F: 39

CAG TGG GAC TTC GTG TCC G, RIPK3-R: CAA GCT GTG TAG GTA GCA CAT C; EP1-F: 40

CTT AAC CTG AGC CTA GCG GAT, EP1-R: ATG TGC CAT TAT CGC CTG TTG; EP2-F: 41

GGA GGA CTG CAA GAG TCG TC, EP2-R: GCG ATG AGA TTC CCC AGA ACC; EP3-F: 42

GCT CAT GGG GAT CAT GTG TGT, EP3-R: CAC CAC CCC GAA GAT GAA CAT; EP4-F: 43

ACC ATT CCT AGA TCG AAC CGT, EP4-R: CAC CAC CCC GAA GAT GAA CAT; ACTIN-F: 44




7

TGA CAG GAT GCA GAA GGA GA; ACTIN-R: GTA CTT GCG CTC AGG AGG AG. 1

2

PGE2 determination 3

Cell supernatants were collected for evaluating PGE2 concentration with UPLC-MS/MS as 4

previously (16). Prior to sample extraction, d4-PGE2 (500 pg) was added to permit quantification. 5

Extracted samples were separated by an Acquity UPLC I-Class system (Waters, MA, USA) and 6

mass spectrometry was performed on an AB Sciex 6500 QTRAP. PGE2 was analyzed using 7

scheduled multiple reaction monitoring (MRM). Data acquisitions were performed using Analyst 8

1.6.2 software (Applied Biosystems). 9

10

Cell proliferation assay 11

Cell counting kit-8 (CCK8) assay was used to assess the proliferation of MDSC and CT-26 12

cells. For indicated experiments 5 × 105 BM-derived MDSC or 5 × 10

3 CT-26 cells were seeded in 13

96-well plates. After 48 hrs, a batch of cells in 100 μl medium was stained with 10 μl of CCK8 14

reagent (Dojindo, Kumamoto, Japan) at 37℃ for 2 hrs. The data was quantified with an automatic 15

plate reader (Thermo) at 450 nm. 16

17

Statistics 18

The number of animals used in the experiments was estimated to give sufficient power 19

(>90%) on the basis of the effect sizes observed in our preliminary data. The statistical analysis 20

was performed using Excel (Microsoft), Origin 9.1 (OriginLab) or GraphPad Prism 7 (GraphPad 21

Software). Statistical significance for binary comparisons was assessed by 2-tailed Student’s t test. 22

For comparison of more than 2 groups, ANOVA with Sidak’s multiple comparisons test was used. 23

For correlation analysis, Pearson’s correlation coefficient was applied. Overall survival was 24

calculated using the Kaplan-Meier method, and the differences in survival curves were analyzed 25

using the log-rank test. All data are reported as mean ± SEM. The P value of 0.05 or less was 26

considered significant. 27

28

Results 29

RIPK3 is downregulated in CRC-infiltrating MDSC 30

The RIPK3 expression was first evaluated in CRC patient databases with Oncomine which 31

showed a consistent decrease of RIPK3 in CRC tissue (Supplementary Fig. S1A). We next 32

employed AOM plus DSS-induced mouse CRC model (Fig. 1A, left panel) (17). The body weight 33

reduced during the DSS treatment and rebound subsequently after DSS withdrawn at each cycle 34

(Supplementary Fig. S1B). Upon sacrifice after Day 90, the colorectal was collected and tumors 35

were separated (Fig. 1A, right panel). We compared the percentage of immune cells in tumor and 36

colorectal tissue and found that both leukocytes (CD45+) and MDSC (CD11b

+Gr-1

+) were 37

significantly higher in tumor than in colorectal tissues (Fig. 1B and 1C). The tumor-infiltrating 38

MDSC showed lower RIPK3, compared with MDSC in the colorectal of tumor-bearing mice (Fig. 39

1D and 1E). 40

We collected clinical colorectal cancer and adjacent normal tissues and found that the 41

accumulation of MDSC was much higher. Consistently, RIPK3 expression in MDSC was 42

significantly suppressed in the tumor microenvironment (TME) than in adjacent tissue (Fig.1F). 43




8

Of interest, we found that RIPK3 expression was highest in the colorectal of IBD mice whereas 1

was lower in the colorectal of CRC mice and lowest in tumor tissues (Supplementary Fig. 2

S1C-S1E), suggesting a differential RIPK3 expression pattern during the development of CRC. In 3

addition, we evaluated other immune cells in colorectal and tumor tissue of CRC mice. CD8+ T 4

cells and dendritic cells (DCs, CD11c+) were less while macrophages (MΦs, F4/80

+) were more 5

abundant in tumor than in colorectal tissues (Supplementary Fig. S1F). We also determined RIPK3 6

expression but did not found significant changes in DCs, MΦs, or T cells when compared to that 7

in colorectal tissues (Supplementary Fig. S1G). We hence wondered whether the down-regulation 8

of RIPK3 in tumor-infiltrating MDSC was caused by factors from TME. We stimulated mouse 9

BM cells with supernatants from CT26 colorectal cancer cells in vitro and found that the 10

percentage of MDSC increased significantly whereas the expression of RIPK3 was 11

down-regulated significantly (Fig.1G). Together these results indicated that RIPK3 was 12

downregulated in tumor tissues and CRC-infiltrating MDSC. 13

14

Enhanced MDSC accumulation and tumorigenesis in RIPK3 deficient mice 15

To investigate the role of RIPK3 in the tumorigenesis of CRC, we employed RIPK3 16

knockout mice (KO). These mice showed decreased body weight, higher diarrhea score, shortened 17

colorectal length, increased tumor number in colorectal, heavier spleen, and significantly reduced 18

survival, compared with wildtype (WT) mice (Fig. 2A-2E; Supplementary Fig. S2A). The 19

accumulation of MDSC in tumor, colorectal and spleen also increased in KO mice (Fig. 2F-2I). Of 20

note, only the granulocytic MDSC (g-MDSC, CD11b+Ly6G

+) increased in the tumor compared 21

with that in the colorectal tissues, which was not shared by the monocytic MDSC (m-MDSC, 22

CD11b+Ly6C

+) (18), MΦs, DCs, or T cells (Fig. 2J and 2K). 23

We also used GSK872, a specific RIPK3 inhibitor (19) to treat the WT mice. GSK872 24

significantly aggregated AOM plus DSS induced weight loss, colorectal shortening, tumor mass, 25

splenomegaly and MDSC accumulation, while it did not alter the infiltration of MΦs and DCs 26

(Supplementary Fig. S2B-S2H). These results demonstrated that RIPK3 deficiency promoted 27

colorectal carcinogenesis and MDSC infiltration. 28

29

Deficiency of RIPK3 promotes the proliferation and immunosuppressive activity of MDSC 30

in vitro 31

The role of RIPK3 on MDSC was next sought in vitro. We found that the percentage of 32

MDSC was much higher in RIPK3-KO group than in WT after mouse BM cells were stimulated 33

by GM-CSF (Fig. 3A), although there was no difference between WT and RIPK3-KO groups 34

without stimulation (Supplementary Fig. S3A). RIPK3 absence in MDSC also resulted in a 35

modest higher proliferation (Fig. 3B) but did not show significant change in cell death (Fig. 3C), 36

which were consistent with the results from GSK872-treated MDSC (Supplementary Fig. S3B). 37

The differentiation of MDSC was also assessed in vitro. After induction with GM-CSF (20 ng/ml) 38

for 48 hrs, the percentage of MΦs, especially M2 type MΦs (F4/80+CD206

+) were significantly 39

higher in KO group, while DCs were lower as compared with WT (Fig. 3D), suggesting that 40

RIPK3 absence in MDSC promoted the M2-like differentiation upon GM-CSF induction. 41

In addition, Arg-1 but not NOS2 or ROS increased in the RIPK3-KO and GSK872-treated 42

MDSC (Fig. 3E and 3F; Supplementary Fig. S3C). The co-cultivation of RIPK3-KO and 43

GSK872-treated MDSC significantly dampened anti-CD3 and anti-CD28 induced proliferation of 44




9

CD8+ T cells (Fig. 3G), as well as the expression of GzmB and IFN-γ (Fig. 3H; Supplementary 1

Fig. S3D), compared with that co-cultured with WT MDSC. Administration of NHNL, an Arg-1 2

inhibitor, significantly rescued the activity of CD8+ T cells (Fig. 3I). Of note, GSK872-treated T 3

cells also showed moderately impaired expression of GzmB and IFN-γ (Supplementary Fig. S3E), 4

suggesting that RIPK3 deficiency in CTL may also contribute to colorectal carcinogenesis. 5

Furthermore, the supernatant from RIPK3-KO MDSC enhanced the proliferation of CT26 cells 6

(Fig. 3J). These results indicated that RIPK3 deficiency enhanced proliferation and 7

immunosuppressive function of MDSC. 8

9

Carcinogenesis is accelerated after RIPK3-deficient BM chimerism 10

To further test the role of RIPK3 on MDSC function in vivo, we generated chimeras by 11

infusing WT or KO BM cells into WT or KO recipient mice after irradiation. The presence of 12

chimerism after 7 weeks was confirmed using FCM. These animals were subsequently induced 13

CRC using model 2 (Supplementary Fig. S3F). We found that the WT recipients exhibited more 14

severe weight loss, higher mortality and tumor formation ratio, shorter colorectal length, 15

splenomegaly and more MDSC in colorectal and spleen when engrafted with cells from 16

RIPK3-KO donors, compared with that from WT donors (Fig. 4A-4G). The percentage of 17

g-MDSC were higher, while MΦs and DCs showed no significant difference in the colorectal and 18

spleen of mice received KO BM, compared with that of mice received WT BM. The leukocytes in 19

colorectal tissues were higher in mice received KO BM, while their percentages did not show 20

significant change in spleen (Fig. 4H and 4I; Supplementary Fig. S3G and S3H). Moreover, 21

although KO recipients showed a moderate weight loss, increased mortality and modest higher 22

tumorigenecity compared with WT recipients after engrafted with WT BM cells, the colorectal 23

length, spleen weight, MDSC infiltration, MΦs and DCs in colorectal and spleen did not show 24

significant change (Fig. 4A-4I). 25

To further validate the essential role of MDSC in the tumorigenesis of CRC, we 26

administrated anti-Gr-1 (to deplete MDSC) and CXCR2-a SB225002 (to inhibit MDSC 27

chemotaxis) every 2 days from the third cycle of DSS until the mice were sacrificed. Both 28

anti-Gr-1 and CXCR2-a reversed the weight loss, mortality, tumor formation, colorectal length 29

and the MDSC infiltration in colorectal and spleen compared with KO control (Fig. 4J-4O). 30

Together these findings supported the conclusion that RIPK3 deficiency in MDSC promoted 31

tumorigenesis. 32

33

NF-κB/COX-2/PGE2 axis is upregulated in RIPK3-deficient MDSC 34

We next explored the underlying mechanism by which RIPK3 regulated MDSC. 35

Aforementioned, PGE2 is a pro-inflammatory and immunosuppressive lipid mediator that 36

potentiates MDSC activity and tumor growth (20). COX-2 is an essential enzyme for the 37

production of PGE2. We found that COX-2 expression was significantly upregulated in the 38

tumor-infiltrating MDSC than in colorectal MDSC (Fig. 5A), but no significant difference of 39

COX-2 expression was observed in CD45- cells of tumor and colorectal tissues (Supplementary 40

Fig. S4A). Compared with WT mice, COX-2 expression were upregulated in tumor-infiltrating 41

MDSC (Fig. 5B) and in CD45- cells of both tumor and colorectal tissues of RIPK3-KO mice 42

(Supplementary Fig. S4B and S4C). Using UPLC-MS/MS, we found that RIPK3-KO MDSC 43




10

produced more PGE2 than that in WT MDSC (Fig. 5C), which was consistent with the results of 1

COX-2 expression. 2

Given that NF-κB is a well-known transcription factor of COX-2 and an essential controller 3

for the immunosuppressive activity of MDSC (21,22), we examined the expression of NF-κB in 4

MDSC. Significant upregulated NF-κB p65 and COX-2 were observed in RIPK3-KO MDSC, 5

compared with WT (Fig. 5D). We administrated aspirin (ASA, COX inhibitor) and caffeic acid 6

phenethyl ester (CAPE, NF-κB inhibitor) (23) and found that they both inhibited PGE2 production 7

from RIPK3-KO MDSC but only showed a trend in decreasing PGE2 production from 8

GSK-872-treated CT26 cells (Supplementary Fig. S4D and S4E). We also assessed other key 9

signaling molecules that drive the accumulation and function of MDSC including stat3 and stat6. 10

However, they showed no difference between WT and KO MDSC (Supplementary Fig. S4F). 11

These findings demonstrated that RIPK3 reduction in MDSC promoted the activation of 12

NF-κB/COX-2/PGE2 axis. 13

14

Inhibitors targeting COX-2 and EP2 blunt the immunosuppressive activity of MDSC and 15

carcinogenesis 16

PGE2 exerts its function by binding to its receptors including EP1-4. We found that the 17

RIPK3-KO MDSC showed higher EP2 and EP4, compared with WT MDSC (Supplementary Fig. 18

S5A). Therefore, the CRC mice model was treated with ASA or EP inhibitors (EP1 and EP2 19

inhibitor AH6809 and EP4 inhibitor ONO-AE3-208) (24). We found that ASA significantly 20

protected the mice against tumorigenesis and reduced the accumulation and COX-2/Arg-1 21

expression of MDSC (Fig. 6, A-E; Supplementary Fig. S5B). AH6809 but not ONO-AE3-208 22

attenuated AOM plus DSS induced tumorigenesis and MDSC accumulation (Fig. 6F-6L; 23

Supplementary Fig. S5C). 24

In vitro, PGE2 significantly enhanced Arg-1 expression in MDSC and the differentiation 25

toward M2 macrophages, which were reversed by ASA and AH6809 (Fig. 6M and 6N). 26

Antagonists of NF-κB/COX-2/PGE2/EPs signaling pathway consistently rescued the CD8+ T cell 27

activation dampened by RIPK3-KO MDSC co-cultivation (Fig. 6O and 6P). 28

In addition, PGE2 promoted the proliferation of CT26 cells, which was blunted by AH6809 29

(Supplementary Fig. S5D, left panel). The co-cultivation with supernatant from RIPK3-KO 30

MDSC also enhanced the proliferation of CT26 cells which was reversed by ASA or AH6809 31

pretreatment in these MDSC (Supplementary Fig. S5D, right panel). Together these results 32

indicated that antagonists targeting NF-κB/COX-2/PGE2 signaling improved the prognosis of 33

CRC. 34

35

RIPK3-PGE2 circuit in tumor microenvironment potentiates malignancy 36

We next questioned whether PGE2 in turn regulated RIPK3 and the downstream 37

NF-κB/COX-2 signaling. BM-derived MDSC were treated with or without PGE2, and we indeed 38

found that PGE2 significantly suppressed RIPK3 while enhanced the expression of p65 and 39

COX-2 in MDSC (Fig. 7A). 40

It is well-known that EP receptors are G-protein coupled receptors that activate cAMP 41

dependent protein kinase A (PKA) and promote the subsequent translocation of the transcription 42

factor cAMP responsive element binding protein (CREB) (25). Of note, a recent report indicated 43

that CREB reduced the promoter activity of RIPK3 (26). Here we found that the 44




11

PGE2-downregulated RIPK3 could be rescued by both H89 (PKA inhibitor) and AH6809 in 1

MDSC (Fig. 7B and 7C). ASA also could upregulate RIPK3 expression in MDSC (Fig. 7D), 2

indicating that PGE2 suppressed RIPK3 via PKA-CREB signaling. Consistently, PGE2 decreased 3

RIPK3 via PKA-CREB pathway while both PGE2 and GSK872 promoted the expression of p65 4

and COX-2 in CT26 cells (Supplementary Fig. S6A-S6D). These results identified a novel circuit 5

involving RIPK3, NF-κB, COX-2, and PGE2 in the tumor microenvironment. 6

To explore the correlation of clinical chemotherapy and targeted therapy with RIPK3 7

expression, we treated MDSC with some common drugs used in cancer. The expression of 8

RIPK3 in MDSC were upregulated by bevacizumab or cetuximab (Fig. 7E) but was 9

downregulated by CPT-11, OXA, 5-FU and GEM (Fig. 7F). 10

We also evaluated the clinical relevance of the RIPK3-PGE2 circuit in CRC cancer patients. 11

We found that with the development of CRC clinical stage, the expression of RIPK3 in 12

tumor-infiltrating MDSC decreased (Fig. 7G; Supplementary Table 1). The relationship between 13

RIPK3 and indicated gene transcripts were examined in 148 patients with CRC from National 14

Center for Biotechnology Information Gene Expression Omnibus database (GSE21510) (13). Our 15

results demonstrated that RIPK3 expression negatively correlated with CD33 and S100A8 which 16

are MDSC markers (27) (Fig. 7H). Importantly, RIPK3 also negatively correlated with PTGS2 17

(COX-2) (Fig. 7I). Since the database GSE21510 lacked the survival results, we analyzed the 18

correlation of RIPK3 and survival of CRC patients with another one (GSE17536) which involved 19

177 patients. We divided patients into “low” and “high” groups based on the median values of 20

RIPK3 and PTGS2. We found that RIPK3high

PTGS2low

patients showed longest survival while low 21

RIPK3 and high PTGS2

was associated with poor survival (Fig. 7J; Supplementary Table 2). 22

Therefore RIPK3 downregulation and COX-2/PGE2 upregulation in the tumor microenvironment 23

formed a circuit that promoted the accumulation of immunosuppressive MDSC and colorectal 24

carcinogenesis. 25

26

DISCUSSION 27

The infiltration of MDSC in tumor microenvironment is closely related to poor prognosis 28

(3,11). Here, we found that that the down-regulation of RIPK3 promoted the infiltration and 29

immunosuppressive activity of MDSC in tumor microenvironment. The chimeric mice experiment 30

also indicated the pivotal role of RIPK3 on CRC-infiltrating MDSC. Therefore, we identified that 31

RIPK3 regulated tumor immunity by modulating MDSC. 32

MDSC play an immunosuppressive function mainly via multiple signal pathways. First, the 33

lipid metabolite PGE2 derived from arachidonic acid via COX-2 catalysis in tumor 34

microenvironment stimulates the expression of Arg-1, IL-6, VEGF and other cancer-promoting 35

molecules in MDSC (11,28). Of note, MDSC also express COX-2 which promotes their own 36

immunosuppressive activity (29). Second, NF-κB activation in MDSC can promote the 37

proliferation and inhibit the differentiation of MDSC (22). Third, the secretion of VEGF from 38

MDSC is also promoted by stat signaling pathway, which enhances angiogenesis (11). Our present 39

study showed that the expression of COX-2 in RIPK3-KO MDSC of CRC tissues was 40

significantly enhanced, compared with that in WT MDSC. In vitro experiments also demonstrated 41

that RIPK3 deficient MDSC exhibited increased COX-2 expression and PGE2 secretion. However, 42

we did not observe significant changes in ROS, NOS2, stat3 and stat6 between WT and 43




12

RIPK3-KO MDSC. These data suggested that the loss of RIPK3 in tumor-infiltrating MDSC 1

promoted the immunosuppressive function by activating COX-2/PGE2. 2

A previous study reported a necrosis-independent pathway of IBD by regulating DCs (9). 3

They showed that in DSS-induced IBD, RIPK3 deficiency impaired NF-κB activation and caspase 4

1-mediated processing of IL-1β in DCs, thereby dampened the tissue repair. Actually, we also 5

observed that RIPK3 was upregulated in the colorectal tissue during the IBD acute induction of 6

DSS, while it significantly reduced at the stage of CRC. Furthermore, we showed that the 7

RIPK3-KO MDSC did not tend to differentiate into DCs. These cells possessed higher 8

immunosuppressive function and were prone to differentiate toward M2 macrophages. Another 9

recent report showed that RIPK3 deficiency enhanced lipopolysaccharide (LPS) induced IL-1β 10

and TNFα expression in macrophages (30). Since LPS is an endotoxin that was found in the outer 11

membrane of Gram-negative bacteria such as E.coli (31) and NF-κB also transcripts IL-1β and 12

TNFα (32), this study was consistent with our results that RIPK3 deficiency in MDSC enhanced 13

NF-κB activation that in turn upregulated COX-2 expression and PGE2 production during the 14

carcinogenesis of CRC. 15

The mechanisms of RIPK3 upregulating NF-κB are complex, which we did not investigate in 16

detail in this study but discuss below. Aforementioned NF-κB is a parallel proliferation pathway of 17

RIPK3-mediated necroptosis and caspases-associated apoptosis (7,33). Knockout of RIPK3 is 18

supposed to lead to a compensatory promotion of NF-κB pathway. A recent study indicated that 19

RIPK3 absent activated NF-κB via a MLKL-independent pathway (34). Moreover, ubiquitination 20

degradation is an important mechanism for the down-regulation of multiple transcription factors in 21

cells. It has been shown that Cullin-RING E3 ligases (CRLs) can mediate NF-κB ubiquitination 22

degradation and reduce its entry into nuclei (35). Hence RIPK3 may also phosphorylate CRLs by 23

mimicking MLKL activation, thereby promoting ubiquitination degradation of NF-κB. 24

Of note, the downregulation of RIPK3 and the subsequent NF-κB/COX-2/PGE2 signaling 25

was also identical in CRC cells in vivo and in vitro, which demonstrated the negative correlation 26

between RIPK3 expression and tumorigenesis. Administration of PGE2 inhibited RIPK3 27

expression but enhanced NF-κB/COX-2 signaling in both MDSC and CT26 colorectal cancer cells 28

via activating PKA-CREB signaling, indicating an unappreciated negative signaling circuit that 29

aggregate the malignancy. Moreover, PGE2 was reported to directly and indirectly blunt the 30

activation of CD8+ T cells (20,36). Our results showed that inhibition of COX-2 or the PGE2 31

receptors significantly reversed the downregulated RIPK3 and attenuated the immunosuppressive 32

activity of MDSC thereby dampened the tumorigenesis of CRC. 33

In summation, our present study identified a novel RIPK3-PGE2 circuit that regulated the 34

infiltration and function of MDSC and the tumorigenesis of CRC. RIPK3 reduction led to NF-κB 35

activation and upregulation of downstream COX-2 which catalyzed the synthesis of PGE2. PGE2 36

in turn further inhibited RIPK3 and promoted NF-κB/COX-2 and Arg-1 expression in MDSC. 37

This signaling circuit also existed in CRC cells and accelerated tumor growth. Importantly, using 38

compounds or drugs targeting this signaling circuit clarified the immunoregulatory role of RIPK3 39

and attenuated the carcinogenesis of CRC. Daily consumption of low-dose ASA has now been 40

applied to efficiently prevent and cure CRC (37,38), delineating the significance of PGE2 41

blockade in the tumor microenvironment. PGE2 was reported to directly stimulate CRC cells to 42

secrete CXCL1 which bound to CXCR2 to recruit MDSC into TME (39,40). Of note, CRC cells 43

also express CXCR2 which predicts poor prognosis. The CXCR2 antagonist inhibits the 44




13

proliferation and metastasis of CRC cells (41,42). Since the RIPK3-PGE2 circuit exists in both 1

MDSC and CRC cells, our results demonstrated a mutual role of PGE2 blockade and CXCR2 2

antagonist in inhibiting CRC tumorigenesis in RIPK3-KO mice. Moreover, our data demonstrated 3

that cetuximab and bevacizumab upregulated RIPK3, while CPT-11, OXA, 5-FU and GEM 4

suppressed RIPK3 in MDSC. Of interest, cetuximab or bevacizumab are both the first-line 5

targeted drugs for CRC patients and they were reported to inhibit COX2 (43,44). Therefore, 6

targeting RIPK3 in MDSC might be considered for rational use of chemotherapeutic and targeted 7

drugs, which is essential for re-educating the immunosuppressive TME and enhance the 8

anti-tumor immunity. These findings provided molecular basis and potential ideas for the 9

immunotherapy of CRC. 10

11

Disclosure of Potential Conflicts of Interest 12

The authors have declared that no conflict of interest exists. 13

14

Grant Support 15

The work was supported by Youth 1000 Talent Plan (to Y. Li) and the National Natural 16

Science Foundation of China (81472435 and 81671573 to Y. Li) and cstc2017jcyjBX0071 (to H. 17

Miao) from the Foundation and Frontier Research Project of Chongqing. 18

19

Acknowledgments 20

We sincerely thank Xiaodong Wang and Zhirong Shen (National Institute of Biological 21

Sciences, Beijing, China) for providing RIPK3 knockout (KO) mice. We extremely appreciate 22

Chunyan Hu for her support on FCM. We are also very grateful to Rong Xin and Lu Jiang for their 23

valuable assistance in IHC and IF procedure, and to Jin Peng and Qian Chen for confocal 24

microscopy experiments. 25

26

References 27

1. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017. CA Cancer J Clin 2017;67:7-30. 28

2. Brenner H, Kloor M, Pox CP. Colorectal cancer. Lancet 2014;383:1490-502. 29

3. Medina-Echeverz J, Aranda F, Berraondo P. Myeloid-derived cells are key targets of tumor 30

immunotherapy. Oncoimmunology 2014;3:e28398. 31

4. Katoh H, Wang D, Daikoku T, Sun H, Dey SK, Dubois RN. CXCR2-expressing 32

myeloid-derived suppressor cells are essential to promote colitis-associated tumorigenesis. 33

Cancer Cell 2013;24:631-44. 34

5. Tauriello DVF, Batlle E. Targeting the Microenvironment in Advanced Colorectal Cancer. 35

Trends Cancer 2016;2:495-504. 36

6. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the immune 37

system. Nat Rev Immunol 2009;9:162-74. 38

7. Han J, Zhong CQ, Zhang DW. Programmed necrosis: backup to and competitor with apoptosis 39

in the immune system. Nat Immunol 2011;12:1143-9. 40

8. Weinlich R, Oberst A, Beere HM, Green DR. Necroptosis in development, inflammation and 41

disease. Nat Rev Mol Cell Biol 2017;18:127-36. 42

9. Moriwaki K, Balaji S, McQuade T, Malhotra N, Kang J, Chan FK. The necroptosis adaptor 43




14

RIPK3 promotes injury-induced cytokine expression and tissue repair. Immunity 1

2014;41:567-78. 2

10. Bozec D, Iuga AC, Roda G, Dahan S, Yeretssian G. Critical function of the necroptosis 3

adaptor RIPK3 in protecting from intestinal tumorigenesis. Oncotarget 2016;7:46384-400. 4

11. Ostrand-Rosenberg S, Sinha P. Myeloid-derived suppressor cells: linking inflammation and 5

cancer. J Immunol 2009;182:4499-506. 6

12. Tsukamoto S, Ishikawa T, Iida S, Ishiguro M, Mogushi K, Mizushima H, et al. Clinical 7

significance of osteoprotegerin expression in human colorectal cancer. Clin Cancer Res 8

2011;17:2444-50. 9

13. Smith JJ, Deane NG, Wu F, Merchant NB, Zhang B, Jiang A, et al. Experimentally derived 10

metastasis gene expression profile predicts recurrence and death in patients with colon cancer. 11

Gastroenterology 2010;138:958-68. 12

14. Lechner MG, Liebertz DJ, Epstein AL. Characterization of cytokine-induced myeloid-derived 13

suppressor cells from normal human peripheral blood mononuclear cells. J Immunol 14

2010;185:2273-84. 15

15. Dolcetti L, Peranzoni E, Ugel S, Marigo I, Fernandez Gomez A, Mesa C, et al. Hierarchy of 16

immunosuppressive strength among myeloid-derived suppressor cell subsets is determined by 17

GM-CSF. Eur J Immunol 2010;40:22-35. 18

16. Li Y, Dalli J, Chiang N, Baron RM, Quintana C, Serhan CN. Plasticity of leukocytic exudates 19

in resolving acute inflammation is regulated by MicroRNA and proresolving mediators. 20

Immunity 2013;39:885-98. 21

17. Neufert C, Becker C, Neurath MF. An inducible mouse model of colon carcinogenesis for the 22

analysis of sporadic and inflammation-driven tumor progression. Nat Protoc 23

2007;2:1998-2004. 24

18. Ouzounova M, Lee E, Piranlioglu R, El Andaloussi A, Kolhe R, Demirci MF, et al. Monocytic 25

and granulocytic myeloid derived suppressor cells differentially regulate spatiotemporal 26

tumour plasticity during metastatic cascade. Nat Commun 2017;8:14979. 27

19. Qiu X, Klausen C, Cheng JC, Leung PC. CD40 ligand induces RIP1-dependent, 28

necroptosis-like cell death in low-grade serous but not serous borderline ovarian tumor cells. 29

Cell Death Dis 2015;6:e1864. 30

20. Zelenay S, van der Veen AG, Bottcher JP, Snelgrove KJ, Rogers N, Acton SE, et al. 31

Cyclooxygenase-Dependent Tumor Growth through Evasion of Immunity. Cell 32

2015;162:1257-70. 33

21. Nguyen LK, Cavadas MA, Kholodenko BN, Frank TD, Cheong A. Species differential 34

regulation of COX2 can be described by an NFkappaB-dependent logic AND gate. Cell Mol 35

Life Sci 2015;72:2431-43. 36

22. Yu J, Wang Y, Yan F, Zhang P, Li H, Zhao H, et al. Noncanonical NF-kappaB activation 37

mediates STAT3-stimulated IDO upregulation in myeloid-derived suppressor cells in breast 38

cancer. J Immunol 2014;193:2574-86. 39

23. Natarajan K, Singh S, Burke TR, Jr., Grunberger D, Aggarwal BB. Caffeic acid phenethyl 40

ester is a potent and specific inhibitor of activation of nuclear transcription factor NF-kappa B. 41

Proc Natl Acad Sci U S A 1996;93:9090-5. 42

24. Rosch S, Ramer R, Brune K, Hinz B. Prostaglandin E2 induces cyclooxygenase-2 expression 43

in human non-pigmented ciliary epithelial cells through activation of p38 and p42/44 44




15

mitogen-activated protein kinases. Biochem Biophys Res Commun 2005;338:1171-8. 1

25. Alvarez Y, Municio C, Alonso S, Sanchez Crespo M, Fernandez N. The induction of IL-10 by 2

zymosan in dendritic cells depends on CREB activation by the coactivators CREB-binding 3

protein and TORC2 and autocrine PGE2. J Immunol 2009;183:1471-9. 4

26. Guida N, Laudati G, Serani A, Mascolo L, Molinaro P, Montuori P, et al. The neurotoxicant 5

PCB-95 by increasing the neuronal transcriptional repressor REST down-regulates caspase-8 6

and increases Ripk1, Ripk3 and MLKL expression determining necroptotic neuronal death. 7

Biochem Pharmacol 2017;142:229-41. 8

27. Heinemann AS, Pirr S, Fehlhaber B, Mellinger L, Burgmann J, Busse M, et al. In neonates 9

S100A8/S100A9 alarmins prevent the expansion of a specific inflammatory monocyte 10

population promoting septic shock. FASEB J 2017;31:1153-64. 11

28. Condamine T, Gabrilovich DI. Molecular mechanisms regulating myeloid-derived suppressor 12

cell differentiation and function. Trends Immunol 2011;32:19-25. 13

29. Rodriguez PC, Hernandez CP, Quiceno D, Dubinett SM, Zabaleta J, Ochoa JB, et al. Arginase 14

I in myeloid suppressor cells is induced by COX-2 in lung carcinoma. J Exp Med 15

2005;202:931-9. 16

30. Moriwaki K, Bertin J, Gough PJ, Chan FK. A RIPK3-caspase 8 complex mediates atypical 17

pro-IL-1beta processing. J Immunol 2015;194:1938-44. 18

31. Wang JH, Manning BJ, Wu QD, Blankson S, Bouchier-Hayes D, Redmond HP. 19

Endotoxin/lipopolysaccharide activates NF-kappa B and enhances tumor cell adhesion and 20

invasion through a beta 1 integrin-dependent mechanism. J Immunol 2003;170:795-804. 21

32. Banerjee D, Liou HC, Sen R. c-Rel-dependent priming of naive T cells by inflammatory 22

cytokines. Immunity 2005;23:445-58. 23

33. Vlantis K, Wullaert A, Polykratis A, Kondylis V, Dannappel M, Schwarzer R, et al. NEMO 24

Prevents RIP Kinase 1-Mediated Epithelial Cell Death and Chronic Intestinal Inflammation by 25

NF-kappaB-Dependent and -Independent Functions. Immunity 2016;44:553-67. 26

34. Najjar M, Saleh D, Zelic M, Nogusa S, Shah S, Tai A, et al. RIPK1 and RIPK3 Kinases 27

Promote Cell-Death-Independent Inflammation by Toll-like Receptor 4. Immunity 28

2016;45:46-59. 29

35. Schweitzer K, Bozko PM, Dubiel W, Naumann M. CSN controls NF-kappaB by 30

deubiquitinylation of IkappaBalpha. EMBO J 2007;26:1532-41. 31

36. Weinlich R, Bortoluci KR, Chehab CF, Serezani CH, Ulbrich AG, Peters-Golden M, et al. 32

TLR4/MYD88-dependent, LPS-induced synthesis of PGE2 by macrophages or dendritic cells 33

prevents anti-CD3-mediated CD95L upregulation in T cells. Cell Death Differ 34

2008;15:1901-9. 35

37. Liao X, Lochhead P, Nishihara R, Morikawa T, Kuchiba A, Yamauchi M, et al. Aspirin use, 36

tumor PIK3CA mutation, and colorectal-cancer survival. N Engl J Med 2012;367:1596-606. 37

38. Algra AM, Rothwell PM. Effects of regular aspirin on long-term cancer incidence and 38

metastasis: a systematic comparison of evidence from observational studies versus 39

randomised trials. Lancet Oncol 2012;13:518-27. 40

39. Kumar V, Donthireddy L, Marvel D, Condamine T, Wang F, Lavilla-Alonso S, et al. 41

Cancer-Associated Fibroblasts Neutralize the Anti-tumor Effect of CSF1 Receptor Blockade 42

by Inducing PMN-MDSC Infiltration of Tumors. Cancer Cell 2017;32:654-68 e5. 43

40. Wang D, Wang H, Brown J, Daikoku T, Ning W, Shi Q, et al. CXCL1 induced by 44




16

prostaglandin E2 promotes angiogenesis in colorectal cancer. J Exp Med 2006;203:941-51. 1

41. Jamieson T, Clarke M, Steele CW, Samuel MS, Neumann J, Jung A, et al. Inhibition of 2

CXCR2 profoundly suppresses inflammation-driven and spontaneous tumorigenesis. J Clin 3

Invest 2012;122:3127-44. 4

42. Wu W, Sun C, Xu D, Zhang X, Shen W, Lv Y, et al. Expression of CXCR2 and its clinical 5

significance in human colorectal cancer. Int J Clin Exp Med 2015;8:5883-9. 6

43. Kaliberova LN, Kusmartsev SA, Krendelchtchikova V, Stockard CR, Grizzle WE, Buchsbaum 7

DJ, et al. Experimental cancer therapy using restoration of NAD+ -linked 8

15-hydroxyprostaglandin dehydrogenase expression. Mol Cancer Ther 2009;8:3130-9. 9

44. Half E, Sun Y, Sinicrope FA. Anti-EGFR and ErbB-2 antibodies attenuate cyclooxygenase-2 10

expression and cooperatively inhibit survival of human colon cancer cells. Cancer Lett 11

2007;251:237-46. 12

13

14

15




17

FIGURE LEGENDS 1

Figure 1. RIPK3 is down-regulated in MDSC of CRC tissues. A, Schematic of mice treated 2

with AOM and DSS (Model 1, 2% DSS drinking for 5 days each cycle). The entity image of the 3

colorectal after mice sacrifice is shown on the right panel. B, The percentage of leukocytes 4

(CD45+) in the tumor and colorectal tissue. C, The percentage of MDSC (CD11b

+Gr-1

+) in CD45

+ 5

cells in the tumor and colorectal tissue. D,E, RIPK3 expression in MDSC in the tumor and 6

colorectal tissue by FCM (D) and confocal microscopy (E) (RIPK3 red,Gr-1 green). F, RIPK3 7

expression in MDSC in human colorectal cancer and adjacent tissue. G, After mouse BM cells 8

were treated with PBS or tumor supernatants (from CT26 cells) for 48 hrs, the percentage of 9

CD11b+Gr-1

+ cells in BMCs and RIPK3

+ cells in CD11b

+Gr-1

+ cells were determined. Data of 10

B-G were expressed as mean±SEM. *P<0.05,

**P<0.01, and

****P<0.0001, by Student’s t test. 11

12

Figure 2. RIPK3 deficiency promotes MDSC infiltration and CRC tumorigenesis. A-K, WT 13

and RIPK3-KO (KO) mice model were established with the protocol as Model 1 in Fig. 1A. The 14

body weight were monitored (A), the tumor number (B), colorectal length (C), spleen weight (D), 15

survival (E), MDSC infiltration in tumor by immunofluorescence staining (F), MDSC percentage 16

in tumor (G), colorectal (H) and spleen (I); the percentages of immune cells (J), CD11b+Ly6G

+ 17

and CD11b+Ly6C

+ cells (K) in the tumor tissue upon sacrifice were assessed. Data were expressed 18

as mean±SEM. *P<0.05,

**P<0.01 and ***P<0.001, WT vs. KO, by Student’s t test. 19

20

Figure 3. RIPK3 absence in MDSC enhances the immunosuppressive activity in vitro. A-C, 21

After bone marrow cells from WT and RIPK3 knockout (KO) mice were treated with GM-CSF 22

(20 ng/ml) for 48 hrs, the proportion (A), proliferation (B) and death (C) of MDSC were examined. 23

After treatment with GM-CSF (20 ng/ml) for 48 hrs, the differentiation of WT and KO MDSC 24

into MΦs and DCs (D), as well as the expression of Arg-1, NOS2 and ROS (E,F) were examined. 25

G, CD8+ T cells were co-cultured with WT/RIPK3-KO MDSC (10:1) for 3 days, the proliferation 26

of CD8+ T cells were determined by CFSE. H, After CD8

+ T cells were co-cultured with 27

WT/RIPK3-KO MDSC (10:1) for 48 hrs, the expression of GzmB and IFN-γ were assessed with 28

FCM. I, After MDSC were treated with vehicle (PBS) or NHNL (30 μM), CD8+ T cells were 29

co-cultured with MDSC (5:1) for 48 hrs, and the expression of GzmB and IFN-γ were assessed. J, 30

After CT26 colorectal cancer cells were cultured with conditioned medium (supernatant of 31

WT/RIPK3-KO MDSC : culture medium=1:1) for 48 hrs, the proliferation was assessed. Data 32

were expressed as mean±SEM. *P<0.05,

**P<0.01, and

***P<0.001, by Student’s t test. 33

34

Figure 4. RIPK3 in myeloid derived cells is essential for inhibiting colorectal tumorigenesis. 35

A-I, The chimeric mice and CRC model were established as indicated in Methods. The body 36

weight (A), mortality (B), tumor formation ratio (C), colorectal length (D), spleen weight (E), 37

MDSC percentage in colorectal (F), and MDSC percentage in spleen (G) , g-MDSC percentage in 38

colorectal (H), and g-MDSC percentage in spleen (I) upon sacrifice were assessed. J-O, CRC 39

model of MDSC depletion and CXCL1 receptor CXCR2 antagonist treatment were established as 40

indicated in Methods. The body weight (J), mortality (K), tumor number (L), colorectal length 41

(M), MDSC percentage in colorectal (N) and in spleen (O). Data were expressed as mean±SEM. 42 *P<0.05 and

***P<0.001, by Student’s t test. 43

44




18

Figure 5. NF-κB/COX-2/PGE2 signaling is enhanced in RIPK3 deficient MDSC. A, The 1

expression of Gr-1 and COX-2 in tumor and colorectal tissue. B, COX-2 expression in 2

tumor-bearing WT/RIPK3-KO MDSC was evaluated by FCM. C, MRM chromatograms, MS/MS 3

spectrum and production of PGE2 from BM-derived WT/RIPK3-KO MDSC identified with 4

UPLC-MS/MS. D, The expression of COX-2 and p65 in BM-derived WT/RIPK3-KO MDSC. 5

Data were expressed as mean±SEM. *P<0.05,

**P<0.01 and

***P<0.001, by 1-way ANOVA with 6

Sidak’s multiple comparisons test (A) or Student’s t test (B,C). 7

8

Figure 6. Blockade of COX-2 or EP2 attenuates tumorigenesis. A-E, WT and RIPK3-KO mice 9

CRC model 2 (Supplementary Fig. S3F) treated with or without 0.02% ASA containing water 10

drinking as indicated in Methods, the body weight, tumor number, colorectal length, and spleen 11

weight upon sacrifice were determined (A-D). The accumulation of MDSC and COX-2 expression 12

in MDSC of tumor, colorectal and spleen were analyzed (E). F-L, WT and RIPK3-KO Mice CRC 13

model 2 treated with or without AH-6809 (5mg/kg) as indicated in Methods, the body weight (F), 14

tumor number, colorectal length, spleen weight (G-I), and the percentage of MDSC in tumor, 15

colorectal and spleen upon sacrifice were determined (J-L). M,N, After BM-derived MDSC from 16

WT or RIPK3-KO mice were treated with vehicle (PBS), PGE2 (10 μM), ASA (10 μM), or 17

AH6809 (10 μM) for 48 hrs, Arg-1 expression (M) and percentage of differentiation toward M2 18

macrophages (N) were assessed with FCM. O,P, After BM-derived MDSC from WT or 19

RIPK3-KO mice were treated with/without ASA (10 μM), CAPE (2 μM) or AH-6809 (10 μM) for 20

48 hrs, CD8+ T cells were co-cultured with MDSC (16:1) for 24 hrs. The expression of IFN-γ (O) 21

and GzmB (P) in CD8+ T cells were assessed with FCM. Data were expressed as mean±SEM. 22

*P<0.05,

**P<0.01,

***P<0.001 and

****P<0.0001, by Student’s t test (A-L) or 1-way ANOVA with 23

Sidak’s multiple comparisons test (M-P). 24

25

Figure 7. PGE2 negatively regulates RIPK3 and upregulates NF-κB and COX-2 in MDSC 26

and CRC cells. A, BM-derived MDSC were treated with vehicle (PBS) or PGE2 (10 μM) for 48 27

hrs, the expression of RIPK3, p65 and COX-2 were assessed with western blot. B,C, BM-derived 28

MDSC were treated with vehicle (PBS), PGE2 (10 μM), H89 (20 μM) or AH6809 (10 μM) for 48 29

hrs, the expression of PKA, CREB, p-CREB and RIPK3 were determined with western blot (B), 30

the RIPK3 mRNA level was determined with qPCR (C). D, BM-derived MDSC were treated with 31

vehicle (PBS) or ASA (10 μM) for 48 hrs, the protein expression of RIPK3 was assessed. E,F, 32

RIPK3 expression of MDSC from bone marrow cells of WT mice were treated with vehicle (PBS), 33

Bevacizumab (2.5mg/ml), Cetuximab (0.5mg/ml), Nimotuzumab (0.5mg/ml) (E), CPT-11 (300nM), 34

OXA (30nM), 5-FU (250nM) or GEM (100nM) (F). G, Confocal microscopy determination of 35

RIPK3 expression in MDSC in human polypus and cancer tissues at different stages. Data were 36

expressed as mean±SEM. *P<0.05, by 1-way ANOVA with Sidak’s multiple comparisons test. 37

(H,I) Pearson’s correlation coefficient was used to determine the correlation between RIPK3 and 38

indicated genes including CD33, S100A8 (H) and PTGS2 (COX-2) (I) in 148 patients with CRC 39

was examined (GSE21510). J, Patient survival data were obtained from GEO database 40

(GSE17536), overall survival probability were then calculated using the Kaplan-Meier method, 41

and the differences in survival curves were analyzed using the log-rank test. 42

























Published OnlineFirst July 16, 2018.Cancer Res Guifang Yan, Huakan Zhao, Qi Zhang, et al. cell-potentiated colorectal carcinogenesisA RIPK3-PGE2 circuit mediates myeloid-derived suppressor

Updated version

10.1158/0008-5472.CAN-17-3962doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2018/07/13/0008-5472.CAN-17-3962.DC1

Access the most recent supplemental material at:

Manuscript

Authorbeen edited. Author manuscripts have been peer reviewed and accepted for publication but have not yet

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/early/2018/07/14/0008-5472.CAN-17-3962To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-17-3962

http://cancerres.aacrjournals.org/content/suppl/2018/07/13/0008-5472.CAN-17-3962.DC1

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/early/2018/07/14/0008-5472.CAN-17-3962


Documents

A RIPK3 PGE2 circuit mediates myeloid derived suppressor...4 1. the correlation of . RIPK3. and indicated genes and survival, respectively. 2. 3. Animal experiments. 4. C57BL/6 mice