Therapeutic Co-Targeting of WEE1 and ATM Downregulates PD-L1 Expression in Pancreatic Cancer

pISSN 1598-2998, eISSN 2005-9256

Cancer Res Treat. 2020;52(1):149-166 https://doi.org/10.4143/crt.2019.183

Original Article Open Access

Therapeutic Co-targeting of WEE1 and ATM Downregulates PD-L1 Expression in Pancreatic Cancer

Mei Hua Jin, MD, MS1 Purpose Ah-Rong Nam, MS1 Pancreatic cancer (PC) is one of the most lethal cancers worldwide, but there are currently no effective treatments. The DNA damage response (DDR) is under investigation for the Ji Eun Park, MS1 development of novel anti-cancer drugs. Since DNA repair pathway alterations have been MS1 Ju-Hee Bang, found frequently in PC, the purpose of this study was to test the DDR-targeting strategy in 1,2 Yung-Jue Bang, MD, PhD PC using WEE1 and ATM inhibitors. Do-Youn Oh, MD, PhD1,2 Materials and Methods We performed in vitro experiments using a total of ten human PC cell lines to evaluate anti- tumor effect of AZD1775 (WEE1 inhibitor) alone or combination with AZD0156 (ATM inhibitor). We established Capan-1–mouse model for in vivo experiments to confirm our findings. Results In our research, we found that WEE1 inhibitor (AZD1775) as single agent showed anti-tumor 1Cancer Research Institute, Seoul National effects in PC cells, however, targeting WEE1 upregulated p-ATM level. Here, we observed University College of Medicine, Seoul, that co-targeting of WEE1 and ATM acted synergistically to reduce cell proliferation and 2Department of Internal Medicine, Seoul migration, and to induce DNA damage in vitro. Notably, inhibition of WEE1 or WEE1/ATM National University Hospital, Seoul, Korea downregulated programmed cell death ligand 1 expression by blocking glycogen synthase kinase-3! serine 9 phosphorylation and decrease of CMTM6 expression. In Capan-1 mouse xenograft model, AZD1775 plus AZD0156 (ATM inhibitor) treatment reduced tumor growth and downregulated tumor expression of programmed cell death ligand 1, CMTM6, CD163, and CXCR2, all of which contribute to tumor immune evasion.

+ + + + + + + + + + + + + + + + + + + + Conclusion Correspondence: Do-Youn Oh, MD, PhD Dual blockade of WEE1 and ATM might be a potential therapeutic strategy for PC. Taken + D +e p+a r +tm +e n +t o +f I n+t e +rn +al M+ e+d i +ci n +e , + + + + + + + + + + + + + + + + + + + + + + + + + + together, our results support further clinical development of DDR-targeting strategies for + S +e o +u l +N a+ti o+n a+l U +n i+v e +rs i+ty +C o +ll e +g e + o f + M +ed +ic i n+e ,+ + 1 +0 1 + D a+e h+a k +-r o+, J+o n +g n +o - g+u ,+ S e+o u +l 0 +3 0 +80 ,+ K +or e+a + PC. + T +e l :+ 8 2+-2 -+2 0 +7 2 +-0 7+0 1 + + + + + + + + + + + + + F +a x +: 8 2+- 2 +-7 6+2 - +96 6+2 + + + + + + + + + + + + + E +-m +a i +l: o +h d +o y +o u +n @ +s n+u .+ac .+k r + + + + + + + + + + + + + + + + + + + + + + + + + + + + Key words + R +e c +ei v +e d + A +p r+il 5+, 2 +0 1 +9 + + + + + + + + + + + A +c c +e p +te d + J +un +e 2 +1, 2+0 1+9 + + + + + + + + + + Pancreatic neoplasms, DNA damage response, WEE1, ATM, PD-L1 + P +u b +li s +h e d+ O +n l+in e+ J+u n +e 2 +5 , +20 1+9 + + + + + + + + + + + + + + + + + + + + + + + + + + +

Introduction telangiectasia mutated (ATM) and partner and localizer of BRCA2 (PALB2) are present in PC patients with frequencies of 5% (germline mutations) and 12% (somatic mutations) [1]. Pancreatic cancer (PC) is one of the most lethal diseases These genes play critical roles in the DNA damage response worldwide, and there is an urgent need to develop effective (DDR), which signals the presence of strand breaks and other therapies for this disease [1,2]. Recently, genomic analyses forms of DNA damage and coordinates their repair. Para- have revealed that many cancer susceptibility genes are fre- doxically, mutations in genes that compromise the DDR can quently mutated in PC, including KRAS (92% of cases) and both cause and protect against cancer. On the one hand, TP53 (78%). In addition, germline and somatic mutations in defects in the DDR can lead to genomic instability and the genes encoding breast cancer 1 and 2 (BRCA1, BRCA2), ataxia accumulation of mutations that increase the probability of

│ https://www.e-crt.org │ Copyright ⓒ 2020 by the Korean Cancer Association 149 This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Cancer Res Treat. 2020;52(1):149-166

cancer. On the other hand, DDR pathway dysfunction can Increasing evidence suggests the existence of crosstalk render tumor cells susceptible to chemotherapeutic agents between the DDR signaling network and immune pathways that damage DNA and/or impair alternative DDR pathways [13,14]. For example, recent studies have demonstrated that [3]. Thus, targeting of specific molecules in the DDR is a the DDR regulates PD-L1 expression in cancer cells via a potential strategy for the development of new drugs for can- pathway involving activation of signal transducer and acti- cers with urgent unmet needs, including PC. vator of transcription (STAT) signaling and inactivation of The cellular response to DNA damage, including single- glycogen synthase kinase-3! (GSK-3!) [15,16]. However, stranded or double-stranded DNA breaks, is controlled by a such interactions between the DDR and immune signaling network of proteins that include damage-sensing proteins have not yet been studied in PC. such as poly(ADP-ribose) polymerase (PARP); transducers Here, we evaluated the anti-tumor effects of targeting the such as the kinases ATM and ataxia telangiectasia and Rad3 DDR using a WEE1 kinase inhibitor (AZD1775) and an ATM related (ATR); and effectors such as the nuclear kinase WEE1, kinase inhibitor (AZD0156) in PC cells in vitro and in a mouse which is a key protein in cell cycle checkpoint control and xenograft model. We also examined the expression and acti- inhibits entry into mitosis [3]. Ultimately, the DDR results in vation of a number of DDR-related and immune signaling- one of three outcomes for the cell: successful DNA repair, cell related molecules to identify potential crosstalk between the cycle arrest, or apoptosis. DDR and immune system in PC. Therapeutic targeting of the DDR pathway has been examined in diverse tumor types [3]. Notably, inhibition of PARP, ATM, ATR, or WEE1 has been shown to abrogate DNA repair via homologous recombination (HR) in many cancers Materials and Methods with a genetically defective DDR, thus leading to synthetic lethality [4-7]. WEE1 acts as a gatekeeper of the G2/M cell cycle checkpoint, and its activity increases during the S and 1. Human cell lines and reagents G2 phases; thus, WEE1 inhibition can induce growth arrest in S phase [8]. In contrast, cancer cells expressing mutant Ten human PC cell lines were employed in this study: TP53 lack a functional G1 checkpoint, and DNA damage AsPC-1, Capan-1, Capan-2, MIA PaCa-2, PANC-1, SNU213, must be repaired during the G2/M transition. Given that SNU324, and SNU410 were purchased from the Korean Cell TP53 is often mutated in PC [8], PC is the good candidate for Line Bank (Seoul, Korea), and SNU2913 and SNU2918, pati- the development of the DDR-acting agents. ent-derived cell lines, were successfully established from The tumor microenvironment plays a critical role in cancer patient. Cells were cultured in medium (MIA PaCa-2 and progression [9]. PC is unique compared with other tumor PANC-1 cells in Dulbecco's modified Eagle's medium, all types in being surrounded by strong stroma. Abundant other cell lines in RPMI-1640, both from Welgen Inc., Gyeong- immunosuppressive cells reside in the tumor microenviron- san, Korea) supplemented with 10% fetal bovine serum and ment, including regulatory T cells, myeloid-derived suppres- 10 µg/mL gentamicin and were maintained at 37°C in a 5% sor cells (MDSCs), M2-type macrophages, and cancer-asso- CO2 atmosphere. The WEE1 inhibitor AZD1775 and ATM ciated fibroblasts (CAFs) [9,10]. Recruitment of these cells inhibitor AZD0156 were kindly provided by AstraZeneca establishes a barrier to the anti-tumor immune response (Macclesfield, UK). [9,10]. In addition, signaling via the chemokine receptor CXCR2 can drive PC growth by recruiting MDSCs and 2. PD-L1 expression analysis by flow cytometry tumor-associated neutrophils and by enhancing the metasta- tic process [10]. Programmed cell death ligand 1 (PD-L1) is a Cells (2!105) were seeded in 60-mm dishes and incubated negative regulator of the immune response and acts by bind- with AZD1775 and/or AZD0156 for 72 hours. The adherent ing to its receptor programmed cell death 1 (PD-1) on T cells, cells were harvested, resuspended in cell staining buffer which inactivates the cells and thus allows the tumor to (#420201, BioLegend, San Diego, CA), and incubated with escape immune surveillance [11]. Data from genomic analy- anti–PD-L1 antibody (#329708, BioLegend) for 30 minutes at ses indicate that immunogenic subtype of PC, which exhibits room temperature. Cells were then washed once with the high levels of PD-L1, cytotoxic T-lymphocyte-associated pro- same buffer and analyzed on a FACSCalibur. The results are tein 4 and CXCR2 among several subtypes [1]. Notably, the presented as the means of three independent experiments. chemokine-like factor-like MARVEL transmembrane domain containing family member 6 (CMTM6) has been suggested as one of the mechanisms of regulation of PD-L1 through preventing PD-L1 degradation by lysosome [12].

150 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

3. Human cytokine array were < 0.05. Half-maximal inhibitory concentrations (IC50) were calculated using SigmaPlot software. Combined drug Cells (1!106) were seeded in 60-mm dishes and exposed to effects were analyzed by calculating the combination index AZD1775 and/or AZD0156 for 24 hours. The cell super- (CI) with CalcuSyn software (Biosoft, Cambridge, UK). CI natant was then collected and 500 µL/sample was analyzed values of < 1, 1, and > 1 indicate synergistic, additive, and using the Proteome Profiler Human Cytokine Array Kit antagonistic effects, respectively. (#ARY-005B, R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. Spot intensities were meas- 8. Others ured using ImageJ software (National Institutes of Health, Betheda, MD). Further information on cell viability assay, colony-forming assay, western blot analysis, immunoprecipitation, cell cycle 4. Human phospho-kinase array analysis, Annexin V/propidium iodide (PI) apoptosis assay, comet assay, transwell migration assay, and immunohisto- Cells (1!106) were seeded in 100-mm dishes and exposed chemistry can be found in supplementary methods. to AZD1775 and/or AZD0156 for 72 hours. The cells were harvested, and lysate samples containing 300 µg of proteins 9. Ethical statement were analyzed using the Proteome Profiler Human Phospho- Kinase Array Kit (#ARY003B, R&D Systems) according to Animal experiments were performed at the Biomedical the manufacturer’s instructions. Spot intensities measured Center for Animal Resource Development of Seoul National using ImageJ software. University (Seoul, Korea) according to institutional guide- lines, and prior approval of the study protocol was obtained 5. Tumor xenograft experiments from the Institutional Animal Care and Use Committee.

Four-week-old female athymic nude mice were purchased from Orient Bio Inc. (Seongnam, Korea). Capan-1 cells were resuspended at 3!107 cells in 100 µL of phosphate-buffered Results saline and injected subcutaneously. The tumor volume was calculated using the formula: volume=[(width)2!height]/2. When the tumor volume reached 200 mm3, the mice were 1. WEE1 inhibition inhibits the proliferation of PC cells randomly assigned to four groups of five mice to receive (1) vehicle (2-hydroxypropyl-!-cyclodextrin solution), (2) AZD- To evaluate the anti-tumor effects of WEE1 inhibition in 1775 once daily at 30 mg/kg for 4 weeks (5 days on/2 days PC, we used the MTT assay to assess the proliferation of 10 off), (3) AZD0156, as described for (2), or (4) AZD1775 plus human PC cell lines in the presence of AZD1775 for 72 hours. AZD0156, as described for (2). All treatments were adminis- As shown in Fig. 1A, AZD1775 inhibited the proliferation of tered by oral gavage. Body weights and tumor sizes were all PC cell lines in a dose-dependent manner. The concentra- measured every other day. tions causing 50% inhibition (IC50) ranged from 0.5 to 2.1 µM (S1 Table). To verify these data, we examined the ability of 6. Mouse cytokine array the PC cell lines to form colonies after 10 days of incubation with AZD1775. This analysis also indicated a profound sup- Immediately before sacrifice, the mice were bled and serum pression of colony formation in all PC cell lines (Fig. 1B), and samples were prepared. Aliquots of 500 µL were analyzed the low IC50 values for colony formation (0.03-0.36 µM) con- using the Proteome Profiler Mouse Cytokine Array Kit, Panel firmed the sensitivity of human PC cells to WEE1 inhibition A (#ARY006, R&D Systems) according to the manufacturer’s (S1 Table). Because all 10 of the PC cell lines showed compa- instructions. Spot intensities were measured using ImageJ rable inhibition by AZD1775, we randomly selected four cell software. lines (Capan-1, SNU213, SNU410, and SNU2913) for the following experiments. 7. Statistical analysis To determine whether AZD1775 blocked signaling in the DDR pathway, we performed western blot analysis of the Analyses were conducted using SigmaPlot version 10.0 expression and activation (phosphorylation) of a number of (Systat Software Inc., San Jose, CA). Data are presented as molecules involved in DDR signaling downstream of WEE1. the mean±standard errors. All statistical tests were two- For these experiments, PC cells were incubated with or with- sided. Differences were considered significant if the p-values out AZD1775 for 24 hours before analysis by western blot-

VOLUME 52 NUMBER 1 JANUARY 2020 151 Cancer Res Treat. 2020;52(1):149-166

MTT assay A Colony forming assay B 140 AsPC-1 140 AsPC-1 120 Capan-1 120 Capan-1 Capan-2 ) Capan-2 ) % (

% 100 MIA PaCa-2 100 MIA PaCa-2 ( s

y PANC-1 PANC-1 i t i n l 80 80 i SNU213 SNU213 o l b

SNU324 o SNU324 a i c

60 60 v

SNU410 f SNU410 l l o

e SNU2913 SNU2913 40 . 40 o C SNU2918 SNU2918 20 N 20 0 0 0 0.001 0.01 0.1 1 10 0 0.001 0.01 0.1 1 10 AZD1775 (µM) AZD1775 (µM)

C Capan-1 SNU213 SNU410 SNU2913

AZD1775 (µM) - 0.1 1 - 0.1 1 - 0.1 1 - 0.1 1 (kDa) p-WEE1 100

WEE1 100

p-CDC2 35

CDC2 35

γ-H2AX 17

c-Myc 55

p–NF-kB 55

NF-kB 55

GAPDH 40

Fig. 1. Growth inhibitory effect of AZD1775 in pancreatic cancer (PC) cell lines. (A) Ten PC cell lines were exposed to control (0) or 0.001, 0.01, 0.1, 1, and 10 µM AZD1775 for 72 hours, and cell viability was measured using the MTT assay. (B) Cell lines were treated as described in A and colony formation was analyzed after 10 days. (C) Western blot analysis of total or phosphorylated signaling molecules in Capan-1, SNU213, SNU410, and SNU2913 cells treated with 0, 0.1, or 1 µM AZD1775 for 24 hours. Experiments were repeated three times. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; p–NF-kB, phosphorylated nuclear factor kappa-light-chain-enhancer of activated B cells. (Continued to the next page)

ting. As shown in Fig. 1C, we confirmed that AZD1775 dose- 2. WEE1 inhibition induces S-phase arrest and apoptosis in dependently reduced the expression of total and phospho- PC cells rylated WEE1 (p-WEE1) in all PC cells tested. In addition, AZD1775 also decreased the expression of phosphorylated Since WEE1 inhibition reduced PC cell proliferation, we cell division cycle protein 2, c-Myc, and phosphorylated next determined whether AZD1775 induced apoptosis. For nuclear factor kappa-light-chain-enhancer of activated B cells this, the cells were incubated with or without AZD1775 for (p–NF-kB), and upregulated the expression of !-H2AX com- 48 hours and apoptosis was examined by flow cytometry of pared with control (Fig. 1C). These results suggest that AZD- Annexin V/PI–stained cells or by western blot analysis of an 1775 monotherapy has anti-proliferative activity in PC cells. apoptosis regulator, MCL-1, and an effector, cleaved caspase 7. We found that AZD1775 treatment significantly increased the proportion of apoptotic cells compared with control cells

152 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

D Capan-1 SNU213 SNU410 SNU2913 105 4.8 105 4.8 105 4.2 105 2.0

l 4 4 4 4 o 10 10 10 10 r Q1 Q2 t Q1 Q2 Q1 Q2

n Q1 Q2

o 3 3 3 3

C 10 3.3 10 10 10 2.8 3.8 Q4 2.9 Q4 Q4 102 102 102 Q4 102

102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105 PI Capan-1 SNU213 SNU410 SNU2913 105 14.8 105 19.8 105 13 105 6.9

5 4 4 4 4 7 10 10 10 10 7 Q1 Q2 Q1 Q2 1 Q1 Q2 Q1 Q2 D

Z 3 3 3 3 10 20.3 10 10 10 16.8 A Q4 17.7 16.2 Q4 Q4 102 102 102 Q4 102

102 103 104 105 102 103 104 105 102 103 104 105 102 103 104 105

Annexin V

EF 50 Control Capan-1 SNU213 SNU410 SNU2913 AZD1775 1 µM

) 40 AZD1775 (µM) - 0.1 1 - 0.1 1 - 0.1 1 - 0.1 1 %

( (kDa)

l l

e 30 t-Caspase-7 35 c

c i t o t 20 c-Caspase-7 17 p o p

A 10 MCL-1 40

0 β-Actin 40 Capan-1 SNU213 SNU410 SNU2913

G H 70 Control Capan-1 SNU213 SNU410 SNU2913 60 AZD1775 1 µM AZD1775 (µM) - 0.1 1 - 0.1 1 - 0.1 1 - 0.1 1

) 50 (kDa) % (

p-RPA32 (S4/S8) 35

e 40 s a

h 30 p

S RPA32 35 20 10 GAPDH 40 0 Capan-1 SNU213 SNU410 SNU2913

Fig. 1. (Continued from the previous page) (D) Annexin V/propidium iodide (PI) apoptosis assay of Capan-1, SNU213, SNU410, and SNU2913 cells treated with 0 or 1 µM AZD1775 for 48 hours. (E) Quantification of three independent Annexin V/PI assays. (F) Western blot analysis of apoptosis-related proteins in cells treated with 0, 0.1, or 1 µM AZD1775 for 48 hours. (G) Cell cycle analysis of PI-stained PC cells treated with 0 or 1 µM AZD1775 for 24 hours. (H) Western blot analysis of cell cycle arrest-related signaling molecules in cells treated with 0, 0.1, or 1 µM AZD1775 for 24 hours. *p < 0.05, **p < 0.01, ***p < 0.001. Experiments were repeated three times.

VOLUME 52 NUMBER 1 JANUARY 2020 153 Cancer Res Treat. 2020;52(1):149-166 ; - r c e i s n o h t i t o t s e

c i h t a n g p o i f

r s e r d p e a n

n e u y g a l l

s l c n

a i a B u t v d t

c 3 n n o x

a 1 t 2 a

e r 4 9

d f , 8 e 2 d e 3 o e t t

0 U n . v s i n a 0

i i N l i t - s S n y i n r y o d i l i o 0 t e a d 2 1 t h a 9 a n 4 8

o p n a , 5 y r U

i s 2 c t . a i p b N o 0 s t

o S l h s s g m i b a p

o n

n i , 3 T n o C 7 1 r T d B

1 g ) 2 e 1 k n t 0 M a i - U B t 0 s . ( b F N

0 e n

. l S a s N C

a ) r W –

e a n ) u t 1 p i 0 D - 1

o a 5 5 0 0 0 5 5 0 . A 1 k n c m ( 3 ( 5 5 4 1 4 7 7 4 h s i a

8 r r . p 0 4 d . e u h a 0 t 2 t n 1 1

- o C i r

h w C

o 1

o f , 2 1

0 5 0 5 0 5 0 - 1 r < P 9 7 3 2 2 1 1 0 5

. g ......

I 2

7 t e r 0 0 0 0 0 0

U l

7 s

) D E ( x e d n i n o i t a n i b m o C o 0 5 - 1 C , e f 1 a N

; 1 c n

D 6 1 e

n > 5 - - Z

g o n 1

i x o A 0 s

e e r r t D d d o o M - t 1 1 n Z r i µ i y

p b A h 1

0 e r g - r h 1 o 1 i d o n 4 o n

t A )

i i / ,

e U t a a ) t 1 d - 1 0 0 0 0 n n N . D a i 0 5 5 5 5 n e 0 k S M h

4 2 2 2 ( 2 b a , T

m p 0 - - m 5

s e A l 7 o 1 ( o h

t 7 p h i C d 1 3

p . m e 1 w - 1 1 s t D

9 o 1 r 3 . a 2

c d Z t 0 u - e e U 3 u s A t - o 1 1 d

N s a 2 h y m

S o e

U M - r h r 2 a - t 1 c µ i N e 7

s S s d r r 1 l a l i

l t o a r a - - f e c r

o p c e 1 e

6 i e c 0 5 n r g

0 y 1

l n h 1 ) n 1 1 0 t a 1 4 g i . l

o R , D i U 0 e T 1 s w t Z N

- i

- H 1 A c n a S d A ( i D a x

e - x t P e p d d -

1 a a a , e t n A e t C 1 a a r

- a

G t l

- -

d 5 ; C e s 7 1 n n r l C

l 7 a o

3 i e 3 1 R ) ) 1 1 1 1 t 1 P c 1 B B H d - E I 5

T T D 2 k k p E t 1 M M

a D f C . i - - d ; A A µ µ Z C U E r P o 0 F F s R a C ( ( T T

c N A A R l R N N s 5 6 S S s

d W i S e - E

– G 7 5 n s h n c p p 7 1 h t -

a y a t i 1 0

l r i B t D D a a

w i w

f Z Z n

d s t t o A A a e a

t t n r t 1 e c a e o o e t v l 1 m i i m - a t b t t g

n a v c a 1 i n a . n e a e t a r 0 r

p r c l t f e t a

a e t o

t C n s

d r e o a - o i i e i n t t c x a W a

a n ) r n t ) n a i R R e H i a C M M

h c T T b M ( b D T T

d n µ u A A . P ( A A m - m e

n d - y A - p o l 5 o s a p

c G 7 e n c

i 7 v f r a a M i 1 e o r t

h t t T D c

f s c Z l t e - A a

a t

c A p

e n y d h s f a e e g g f t i i s r l E

s a s

, -

, a . s a

y t T 2 r

p c T . o e A p T g f h a i T f F p M e S k

154 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

A B Control AZD1775 AZD0156 Combination Control AZD0156 1 µM AZD1775 1 µM Combination 100

Capan-1 80 y t i s

n 60 e t n i

i 40 a T SNU213 20

0 Capan-1 SNU213 SNU410 SNU2913

Control AZD0156 1 µM SNU410 AZD1775 1 µM Combination 100

80 t n e 60 SNU2913 m o m

l i 40 a T 20

0 Capan-1 SNU213 SNU410 SNU2913

C Capan-1 SNU213 SNU410 SNU2913 AZD1775 (µM) - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 AZD0156 (µM) - - 1 1 - - 1 1 - - 1 1 - - 1 1 (kDa) p-ATM 250

ATM 250

γ-H2AX 17

GAPDH 40

Fig. 3. Effect of WEE1 and ataxia telangiectasia mutated (ATM) co-inhibition on DNA damage. (A, B) Displayed Comet Assay results after treated with 0, AZD1775 (1 µmol/L) alone, AZD0156 (1 µmol/L) alone, or both for 24 hours. Tail intensity and moment were analyzed using the Comet Assay IV program. (C) The related DNA damage signals were detected after exposed at combination treatment setting for 24 hours. **p < 0.01, ***p < 0.001. Experiments were repeated three times. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

(Fig. 1D and E) and concomitantly diminished the expression S-phase. To verify this, we examined expression of phospho- of MCL-1 and elevated the expression of cleaved caspase-7 rylated replication protein A 32 (p-RPA32), an S-phase mar- (Fig. 1F). Since WEE1 functions as a regulator of cell cycle ker that binds to single-stranded DNA. Indeed, AZD1775 progression, we also assessed the proportion of cells in the treatment for 24 hours resulted in increased p-RPA32 at ser- cell cycle phases by flow cytometric analysis of PI-stained ine 4 and 8, indicating that WEE1 inhibition leads to replica- cells. As shown in Fig. 1G, a significantly greater proportion tion stress (Fig. 1H). of cells treated with AZD1775 than control was arrested in

VOLUME 52 NUMBER 1 JANUARY 2020 155 Cancer Res Treat. 2020;52(1):149-166

3. Co-inhibition of WEE1 and ATM profoundly impairs gression. Importantly, very few studies have evaluated the activation of the HR pathway of DNA repair effects of DDR-targeted agents on tumor cell migration. In this study, Transwell migration assay was employed. Inter- Having evaluated the effects of AZD1775 as a single agent, estingly, we found that cell migration and invasion of all four we asked whether WEE1 inhibition influences the activity of PC cell lines were significantly suppressed by treatment with other core members of the DDR, such as ATM or ATR. 1 µM AZD1775 or AZD0156 alone. However, the combina- Indeed, we found that phosphorylation of ATM and ATR tion of both drugs resulted in efficiently inhibition of cell was increased in cells treated for 24 hours with AZD1775 migration (Fig. 4A and B). Thus, WEE1 and ATM appear to (Fig. 2A). Since AZD1775 had a greater effect on promoting have previously unrecognized functions in promoting PC phosphorylation of ATM than of ATR, we next examined the cell migration. anti-proliferative effects of AZD1775 in combination with the The matrix metallopeptidases 9 and 2 (MMP-9 and MMP- ATM inhibitor AZD0156 using the MTT assay. As shown in 2) are known to be crucial for cancer metastasis and invasion Fig. 2B, the CI values of AZD0156 plus AZD1775 were less [17,18]. Therefore, we asked whether the effects of WEE1 than 1 for all four PC cell lines, indicating that the drugs had and/or ATM inhibition on PC cell migration were mediated a synergistic effect on proliferation. via these enzymes. Western blot analysis revealed that treat- To better understand the effects of combination AZD1775 ment with AZD1775 or AZD0156 for 24 hours slightly and AZD0156 treatment, we examined the expression levels decreased the expression of the smaller, active form of MMP- of several molecules required for DNA repair via HR, after 9 compared with control cells, and this effect was augmented treatment of cells for 72 hours with AZD0156 and/or AZD- by co-treatment with both agents (Fig. 4C). In contrast, a 1775. We found that expression of p–NF-!B, Rad51, excision reduction in active MMP-2 expression was only observed in repair cross-complementing protein 1, C-terminal binding combination-treated Capan-1 and SNU410 cells (Fig. 4C). protein-interacting protein, and p-STAT1 were all downreg- Next, we investigated the effects of DDR targeting on ulated by AZD1775 or AZD0156 when added alone, but a chemokine expression in Capan-1 cells using a human cyto- greater effect was observed in cells co-treated with both kine/chemokine array. We found that the chemokines inter- agents (Fig. 2C). These data suggest that co-inhibition of leukin 8 (IL-8), CXCL1, CCL5, and CCL2 were significantly WEE1 and ATM strongly blocks the HR pathway of DNA downregulated by AZD1775 treatment, combination therapy repair. again showed an enhanced inhibitory effect on all four chemokines (Fig. 4D and E). Interestingly, previous work showed 4. Co-inhibition of WEE1 and ATM synergistically induces that signaling via the receptors for these chemokines (CXC- DNA damage R2, CCR5, and CCR2), facilitate the release of MMP-9, which contributes to enhanced angiogenesis and tumor metastasis Having demonstrated that co-inhibition of WEE1 and [19-21]. Collectively, our results suggest that AZD1775 and ATM blocks DNA repair, we next asked whether these AZD0156 act synergistically to suppress PC cell migration agents could induce DNA damage. We exposed the cells to by reducing chemokine expression and MMP-9 release. AZD1775 and/or AZD0156 for 24 hours, and then monitored DNA fragmentation at the single-cell level using a comet 6. Co-inhibition of WEE1 and ATM downregulates PD-L1 assay. We found that AZD1775 or AZD0156 alone promoted expression in PC cells DNA damage compared with control, but both agents in combination caused markedly increased fragmentation, as Next, given that the pivotal status of PD-L1 expression in indicated by comet tail intensity and moment (Fig. 3A and cancer cell or cancer microenvironment [22], we explored the B). Consistent with this, western blot analysis revealed potential interaction between the DDR and the anti-tumor upregulated expression of the DNA damage marker "-H2AX, immune response by examining the effects of DDR-targeted with a greater effect observed in cells subjected to combina- agents on total and cell surface expression of PD-L1 in PC tion AZD1775 and AZD0156 treatment compared with cells. Western blot analysis revealed that total cellular PD-L1 monotherapy (Fig. 3C). expression was decreased by AZD1775 or AZD0156 treatment alone, but combination treatment was even more effec- 5. Co-inhibition of WEE1 and ATM efficiently suppresses tive (Fig. 5A). This pattern was also observed when cell PC cell migration surface PD-L1 expression was examined by flow cytometry, particularly in SNU2913 cells, which express high PD-L1 lev- To determine the consequences of WEE1 and ATM inhibi- els (Fig. 5B). tion on PC cell function, we examined cell migration and To understand in more detail, the effects of WEE1 and ATM invasion, which play well-characterized roles in cancer pro- inhibition on PD-L1 expression, we investigated potential

156 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

A 24 hr Control AZD1775 AZD0156 Combination

Capan-1

SNU213

SNU410

SNU2913

B Capan-1 SNU213

120 120 ) )

% 100 % 100 ( (

n n o o i 80 i 80 s s a a v v n n i 60 i 60 - - n n o o i i t 40 t 40 a a r r g g i 20 i 20 M M 0 0 l l ro M M on ro M M on t µ µ ti t µ µ ti on 1 1 a on 1 1 a C 5 6 in C 5 6 in 77 15 b 77 15 b 1 0 om 1 0 om ZD ZD C ZD ZD C A A A A

SNU410 SNU2913

120 120 ) ) % % 100 100 ( (

n n o o i i 80 80 s s a a v v n n i i 60 60 - - n n o o i i t t 40 40 a a r r g g i i 20 20 M M 0 0 l n l n ro M M io ro M M io nt µ µ t nt µ µ t o 1 1 na o 1 1 na C 5 6 bi C 5 6 bi 77 15 77 15 1 0 om 1 0 om ZD ZD C ZD ZD C A A A A Fig. 4. Effect of WEE1 and ataxia telangiectasia mutated co-inhibition on migration of pancreatic cancer cells. (A) The anti- migration effect was evaluated using cells with 0, AZD1775 1 µmol/L, or AZD0156 1 µmol/L or both. The image was captured at 24 hours after treatment. (B) The percentage of migratory cells was analyzed after image was captured. The data represents three independent times. (Continued to the next page)

VOLUME 52 NUMBER 1 JANUARY 2020 157 Cancer Res Treat. 2020;52(1):149-166

C Capan-1 SNU213 SNU410 SNU2913 AZD1775 (µM) - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 AZD0156 (µM) - - 1 1 - - 1 1 - - 1 1 - - 1 1 (kDa) MMP-9 75

MMP-2 75

GAPDH 40

DE Control AZD0156 1 µM AZD1775 1 µM Combination Control 12,000 10,000 y t i 8,000 s n e t

n 6,000 i

t o

AZD1775 p 4,000 S 2,000 0 IL-8 CXCL1

AZD0156 Control AZD0156 1 µM AZD1775 1 µM Combination 12,000 10,000 y t i 8,000 s n e

Combination t

n 6,000 i

t o

p 4,000 S 2,000 0 CCL5 CCL2

Fig. 4. (Continued from the previous page) (C) Western blot analysis of migration-related molecules in cells treated as described in panel A for 24 hours. MMP, matrix metallopeptidases; GAPDH, glyceraldehyde 3-phosphate dehydrogenase. (D, E) Human cytokine array analysis of Capan-1 cells treated as described for panel A for 24 hours. Spot intensities were quantified using ImageJ software. 1, interleukin 8 (IL-8); 2, CXCL1; 3, CCL5; 4, CCL2. *p < 0.05, **p < 0.01, ***p < 0.001.

mechanisms that regulate PD-L1: CMTM6, which is thought ing with anti–PD-L1 and probing for the presence of CMTM6 to prevent PD-L1 degradation by lysosomes [12], and GSK- in the immunoprecipitates. We found that CMTM6 was pres- 3!, which has recently been identified as a novel regulator of ent in western blots of anti–PD-L1 immunoprecipitates, but PD-L1 expression [23]. Western blot analysis of CMTM6 not control IgG immunoprecipitates, of all four PC cell lines expression showed a marked decrease in the protein levels (Fig. 5C). Notably, immunoprecipitation (IP) with anti– after treatment for 72 hours with AZD1775 and AZD0156, PD-L1 antibody showed the CMTM6–PD-L1 binding was either alone or in combination (Fig. 5A). To probe CMTM6- decreased upon WEE1 or ATM inhibition, with the greatest mediated regulation of PD-L1 further, we asked whether this effects observed upon dual inhibition. However, IP of CMT- molecule is associated intracellularly by immunoprecipitat- M6 followed by western blotting of PD-L1 did not show a

158 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer ) e 1 d P d e L I - ( b y

i D h s r e c e P

t s d f l a e t o a

i d r

s p i s e i s B c a c

y y e l l d r a g e

p t n , o a a

3 H e n M t 1 r n µ u t

o D 9 o

i l 1 t

2 P C s

b a l U

5 n l A i n o i t a n i b m o 1 C m N n e b 0 i r G c S

D m

e .

Z o

t 3

A C 6 s 6 5 1 0 D Z A 1 r M e 9 u M T 2 T o

h U M M 5 7 7 1 D Z A )

M µ

2 C C N 1 A

: 7 (

S i

t P

l 7 . r 1 I l o r t n o 7 C o d - n o n r 1 1 t f n a n -

o D n

a i

Z a

n o 6 r

C A a 5 n o i t a n i b m o C o a s 1

p 1 - C e , a 0 r

n 1

C a

D L 6 5 1 0 D Z A - x p Z e a D

A )

C P

1 I r 5 7 7 1 D Z A –

0 0 0 0 0 0 : n i L o

5 4 3 2 1 t -

P o /

I ) % ( 1 L - D P e c a f r u S

d n a l o r t n o C P n o , ( i

s G 1 s

5 g e 7 I d

7 l

p 1 n o i t a n i b m o C o a x r g D t e i

l n

o 6 5 1 0 D Z A h A 1 L c

t ) - A L

a - f a

M D e

o D D

5 7 7 1 D Z A µ 0 P d 7 0 k

P s

( 4 1 4 : i l 1 e

l s

c r e I

y a 1 1 o l o r t n o C c l

1 a r 0 - d

3 n u

n e 1 h -

a 1 s t a

9 n o i t a n i b m o C i l t m p 2 l o a w U e l

- 1 C

N b a

d 6 5 1 0 D Z A r f S e n t o g

r - - G

e g

r e i I t 5 7 7 1 D Z A r s : s p t

y e P l s I

l 1 1

l o W l o r t n o C

n e ) n c a 0

- o 1 C ) 1 ) i g ( 4 t

e C i . n U g i

s P b t

a - 1 ( i N e

3 1 9 2 U N S r p r S h

o m t e s n i

x i c - - t

l n 0 1 4 U N S e o 1 n e a

c L e c c

e -

d c

D h d t 1 1 i 3 1 2 U N S t e

P e t

o t : a t a d a

e t

e - I

1 d r v t u 1 - n a p a C i e 2 c t a u U n c m e

n -

a 1 a N p i

- t p i 3 1 9 2 U N S S e

e s r n c

n a o - - i e t n

r C

c s e ( 0 1 4 U N S e e e c

i . l s s w g G u e

r 1 1

r c g

n s I 3 1 2 U N S u t e

o a : l l o n 1 u P o e - l - h 1 e I

F m

2 1 - n a p a a C m

a ) i i 7 p g

r - 1 B x a r ( n e a

i C o . t l p f e

a a x - -

) s 6 1 n E d a

L A g . - ( M n n i

s ) ) T 1 s e a D r 6 r

H L g P o - u 1 y M M M D M f o r

µ µ o D C E T P r ( ( o

d P h E t A d

5 6 M e a G 2 y 7 5 C l b W 7 1 7 i

h u

f 1 0 r e r g c o D D d

o s e

Z Z t f r e

e c A A – t

d e A

1 a f

s f l L h a - e E

n s D d . a e o 5 P t

p . h

a d g r p e i - n o r 3 F a f t

VOLUME 52 NUMBER 1 JANUARY 2020 159 Cancer Res Treat. 2020;52(1):149-166 , s 5 e

i ; t i 6 s 2 4 n r e t y n T i

t s E o e y p - M M S n µ µ

p o . i 1 1

s , 5 6 a r 4 l 7 5 n i

7 1 o u ; r b 1 0 t o 9 D D n m 1 h Z Z o o

A C A C 4 4 r 2

y r T o

f c

3 r 5 A S P

- - l p p e

, n 3 a

. ; p

K 1 3 r A 0 3 o F 0 1 - f .

r p 0 d e

e S <

b i B s p r * e E c * Y s * - R

e p , C d 1 -

0 s p .

a , 0

r 2 d

< S ;

e - t 9 p p r a * e e *

r S , t

5 B ) s E 0 ! l . l R 3 0 e - C

- c

< K p

1 S - p β * G n

3 . – a - 5 K p p 1 ( S a r G e ! C -

3 S p f

- ) o e

0 0 0 0 3 s s 0 0 0 0 i 5 a 0 0 0 0 s , , , , p n

- y i 2 0 8 6

1 1 p k y t i s n e t n i t o p S a (

n e 3 s a

5 a y p h

a t r d n r e y t a

s a

e l s n y a r e D n o g i o h k c - p y s o l o h g

h p s p d

o e , t 6 h

a 6 6 6 6 ; l p

7 y n 9 r a 3 o r h m y p u T s

) o H

) h K E p

A 1

1 1 1 , , F 1 - D

( . p ( ) e

e r e g a s a 5 5 5 5 a w p

t n f s i o u k

s o

i n J v o 2 2 2 2 e e i r g s p a e

e h m h 3 4 3 4 3 4 3 4 I d t

a l

g 5 6 n l m o 7 5 n o r a o i i t 7 1 t r c s 1 0 f n a

o o u D D f n d

i C Z Z e b d d A A u e e m i n t i f o a t i l C t n y n o r a C o ( u

q .

p 5 s e

. r o g e h i w F p

160 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

l s e h d i a t d s n n y y a i w l

h o a 5 m r e r 7 n g d e 7

a l

t r 1 a t

o r o d D D l e ) n m c Z b

a y

u 0 0

A D l n

, o 5 5 5 5 0 0 7 7 0 0

T i k r

) g t

2 ( 2 5 5 4 4 1 1 4 4

e a )

, 0

t t n

i 0 B

b H

2 n , e

m b

i D o !

t A

r C (

a W P (

n o

i ) .

7 b A

l ,

) D m e

G (

o g

d C ; .

6 K k

5 s

B 1 l

, m /

0 l )

m D µ e

g 2 0

Z c

0 e

0 A

6 m

5 s 2 0

7 B 1

0 1 2 u 0 !

d D ( 3

o =

Z ( e

s 4

A t 5 3

r m 2

7 6 a

6 7 a

t t

5 1 v 1

n b i 0 D 1

t a

e Z

D 2

0 e

r c A

5 l

m 7 C

a g

t D a

f 1 , c o a

1 ) g D g

o e S n

Z 0 r k k

. A e / / r t

4 s g l g , e

r o 4

)

c e r m m

o 1

t g n

! n f n

- 0 0 (

o 0 o k i

a 3 3

a m n t

1 C

)

/ l h 6 5 a

a s u

l o 5 7 g t

n y r n

p 1 t 7 o

L 8

a e r b n 0 1

a - d t m

D o

D D m n e C C n D Z Z 0 o o s

i i 6 3 P C A C A e a c ( ( h

x t c 5 1

e 3 -

) ) 7 t 6 2 n s t X B B H d i g 7 M M t r h R

k k A f D n M 1 T T n o - - g C 2 1 n a o T P F F i h a A A l D X r l o H ( - s A 5 0 5 0 5 0 g -

N N i - ( M C

i Z p

t 2 2 1 1 1 a – G γ

l i C o 1 L ) g ( t h g i e w y d o B p A p L - b n

i h - , p D e t h e D a x P a l

n P k e c i

1 i - r - d h

o o n l e t c l a

c v e ) p

a c f a

M t r d C i T

a e n e w A l i A

( m c C

d g u d e m n t e n i a

n t a r n o a d e i i g t t r e a t o t a t u

r 9 a s n e

i l 2 , p m c b

)

y i f r a 0 7 m i o o 0 m 2 o s

2 h s C a d i t ! p 4 s e ( c s 2

t y e ) t f o i l n a L h 0 a g r e 2 E p n n

g m , 6 a a t N o

l 5 7 B l a 1 n e 1 U k ) g g e t a 0 e -

r k k e T c / / t x D ( i

F a 4 g

g g i r – Z 1 a g x e N m m m 1 p A

t n -

. n a f e 0 0 o t 0 i t s i n 3 3 a l h t

x r a a e 5 6 c a e s o l / 7 5 n p y b o n i 7 1 o 1

t 8 a a a r b m 1 0 e t l s E

D i D D C n m u h

t E t Z Z d o o h

5 6 n C A A C n o o 7 d i W t

e 7 e

) t 1 d k d u 3 B a e D c l ( n i

Z u m o a s n

n A s

t 1 i i m 1

t P h I 0 0 0 0

E n g T ) m 0 0 0 i o E 0 0 0 o e U C , , , C r

( 3 2 1 ( f W d

. ) m m ( e m u l o v r o m u

T .

f e s 3 d e y s o l e s k

t a d o a e s r s r o t e t n e e c b f n e v

e w s o g r f

e f n C a s o 4 e

a r h u

o r d t o s t

o l y n m i y p h

m 3 1 7 L e e d u t 6 6 L d E i u

- - d 1 t t o i r

n N - r D o D i K e a o U t t

P e f p C

) l

T a n s c 6 h A u A u 5 (

o 1

n s i s . 0 y r o e 6

a x D . v h o v r g

p Z e i f - u 3 F o c d A

VOLUME 52 NUMBER 1 JANUARY 2020 161 Cancer Res Treat. 2020;52(1):149-166

Control

AZD1775

AZD0156

Combination

F 12,000 C5/C5a 12,000 TIMP-1 12,000 M-CSF 10,000 10,000 10,000 y y y t t t i 8,000 i 8,000 i 8,000 s s s n n n e e e t t t

n 6,000 n 6,000 n 6,000 i i i

t t t o o o

p 4,000 p 4,000 p 4,000 S S S 2,000 2,000 2,000 0 0 0 l n l n l n ro M M io ro M M io ro M M io nt µ µ t nt µ µ t nt µ µ t o 1 1 na o 1 1 na o 1 1 na C 5 6 bi C 5 6 bi C 5 6 bi 77 15 77 15 77 15 1 0 om 1 0 om 1 0 om ZD ZD C ZD ZD C ZD ZD C A A A A A A

CD54 IL-16 CXCL12 12,000 12,000 12,000 10,000 10,000 10,000 y y y t t t i 8,000 i 8,000 i 8,000 s s s n n n e e e t t t

n 6,000 n 6,000 n 6,000 i i i

t t t o o o

p 4,000 p 4,000 p 4,000 S S S 2,000 2,000 2,000 0 0 0 l n l n l n ro M M io ro M M io ro M M io nt µ µ t nt µ µ t nt µ µ t o 1 1 na o 1 1 na o 1 1 na C 5 6 bi C 5 6 bi C 5 6 i 77 15 77 15 77 15 b 1 0 om 1 0 om 1 0 om ZD ZD C ZD ZD C ZD ZD C A A A A A A

Fig. 6. (Continued from the previous page) (E, F) Mouse cytokine array analysis of serum collected from xenografted mice. Spot intensities were measured using ImageJ software. 1, C5/C5a; 2, TIMP metallopeptidase inhibitor (TIMP-1); 3, macrophage colony-stimulating factor (M-CSF); 4, CD54; 5, interleukin 16 (IL-16); 6, CXCL12. *p < 0.05, **p < 0.01, ***p < 0.001.

162 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

comparable decrease in CMTM6–PD-L1 binding, even in vitro findings in the isolated tumors. In line with the in vitro cells co-treated with both AZD0156 and AZD1775 (Fig. 5C). results, we confirmed that AZD1775 treatment upregulated It is possible that the reduction of CMTM6 by AZD1775 and p-ATM while combination treatment with AZD-1775 and the AZD0156 led to decrease of CMTM6-PD-L1 bound form, ATM inhibitor reversed this. Moreover, the expression of which ultimately increase the PD-L1 degradation by lyso- p–NF-"B and CXCR2 were profoundly blocked by AZD1775 some supported by previous report [12]. and/or AZD0156 treatment, and AZD1775 plus AZD0156 To determine whether PD-L1 expression might be affected enhanced #-H2AX accumulation. Furthermore, we validated by the regulator GSK-3! in PC cells [23], we examined changes the in vitro findings with PD-L1 and CMTM6 by confirming in phosphorylated kinase expression in AZD1775 and/or that both proteins were downregulated by single- or dual- AZD0156-treated cells using a human phospho-kinase array. agent treatment (Fig. 6D). We observed that while AZD1775 and AZD0156 both down- Finally, we assessed the effects of the DDR-targeting regulated p–GSK-3! expression, the combination treatment agents on immune cell activity by analyzing cytokine levels was more effective than either agent alone (Fig. 5D and E). in the sera of tumor-bearing mice. Using a mouse cytokine This indicated that PD-L1 expression not only controlled by array, we found that release of C5a, TIMP metallopeptidase CMTM6 but also influenced by GSK-3! activity to a certain inhibitor 1 (TIMP-1), macrophage colony-stimulating factor, extent. Moreover, we found the same inhibitory effects on CD54, IL-16, and CXCL12 were markedly reduced by AZD- several other phospho-kinases, including phosphorylated 1775 and/or AZD0156 treatment, with larger effects obser- cAMP response element-binding protein (p-CREB), p-Src, ved in drug combination-treated mice (Fig. 6E and F). Collec- phosphorylated focal adhesion kinase (p-FAK), p-Yes, and tively, these in vivo data confirm that dual targeting of the p-p53 (Fig. 5D and E). Taken together, these results indicate DDR pathway components WEE1 and ATM profoundly sup- that dual blockade of WEE1 and ATM may reduce PD-L1 press tumor growth in vivo compared with blockade of either expression by downregulating the expression of CMTM6, molecule alone. and inactive GSK-3!.

7. Co-inhibition of WEE1 and ATM augmented anti-tumor growth in Capan-1-xenograft model Discussion

To confirm the anti-tumor effect of AZD1775 and AZD- 0156 in vivo, Capan-1 xenograft model was established. Although ATM is a tumor suppressor gene, both of WEE1 Despite high PD-L1 expression was detected in vitro in SNU- and ATM promote DNA repair to maintain cell survival dur- 2913 cells, since Capan-1 was more tumorigenic than SNU- ing the progression of the cancers [24]. Inhibition of WEE1 2913, we selected Capan-1 xenograft models to conduct the and ATM activity are our opportunity to kill the cancers by following experiments. Consistent with the in vitro observa- facilitating DNA damage. tions, we found that single-agent treatment with AZD1775 In this study, we demonstrate for the first time that target- or AZD0156 significantly decreased tumor growth, but the ing the DDR by inhibition of WEE1 has anti-tumor effects in combination treatment was markedly more effective than PC and that co-inhibition of WEE1 and ATM amplifies these either agent alone (p < 0.05) (Fig. 6A). We observed no overt effects. We also identified crosstalk between the DDR path- evidence of drug toxicity, as reflected by a lack of significant way and immune system by demonstrating DDR-dependent change in body weight (Fig. 6B). regulation of PD-L1 expression in PC cells. Next, we excised the tumors and performed immunohis- One of the main findings of this study is that dual inhibi- tochemical (IHC) staining of various markers. As we had tion of WEE1 and ATM has profound effects on PC cell observed in vitro, PD-L1 expression was decreased by AZD- migration. Only a few studies have been performed on the 1775 or AZD0156 monotherapy and decreased by combina- anti-migratory effects of DDR-acting agents. One report tion treatment (Fig. 6C). Moreover, CD163, an M2-type showed that AZD1775 inhibits the migration of gastric cancer macrophage marker, was expressed at high levels in tumors cells, although the mechanism of action was not elucidated from control mice but at lower levels in tumors from [25]. Another study suggested that ATM might promote cell AZD1775- and/or AZD0156-treated mice (Fig. 6C). Combi- migration by regulating IL-8 expression independently of its nation treatment with both drugs also potently inhibited role in DNA double-strand break repair [26]. We found that staining of the proliferation marker Ki-67 and increased ter- WEE1 and ATM co-inhibition reduced the expression of minal deoxynucleotidyl transferase dUTP nick end labeling MMP-9, IL-8, CXCL1, CCL2, and CCL5, which is consistent staining, which detects fragmented DNA (Fig. 6C). In paral- with an anti-migratory effect. This is the first demonstration lel, we performed western blot analysis to confirm our in of the involvement of WEE1 or ATM in MMP-9 or IL-8 reg-

VOLUME 52 NUMBER 1 JANUARY 2020 163 Cancer Res Treat. 2020;52(1):149-166

ulation. Moreover, downregulation of CXCR2, which is the increase the HIP1R expression and compete for CMTM6 to IL-8 and CXCL1 receptor, in PC tumors was confirmed in our bind with PD-L1. Secondly, CMTM6 might be transcription- in vivo experiments. Although CXCR2 inhibition was shown ally downregulated by WEE1 or ATM, and these events fur- to profoundly suppress metastasis and augments anti–PD-1 ther disrupt the interaction between PD-L1 and CMTM6. The therapy in PC [10], the current study is the first to examine mechanism by which DDR targeting blocks CMTM6 expres- CXCR2 expression in PC. Since CXCR2 and CCR2 are also sion remains to be clarified. expressed by M2 macrophages, we speculate that combina- As mentioned earlier, KRAS and TP53 are frequently tion AZD1775 and AZD0156 treatment might impair M2 mutated in PC. Recent work suggests that KRAS mutant can- type polarization [27]. In addition, the involvement of Src cer cells can enhance the recruitment of MDSCs and M2 and FAK in promoting cell migration and invasion is well macrophages to the tumor, thereby contributing to evasion known [28], and a recent report implicates a similar role for of the immune response. M2 macrophages can also be recrui- p-CREB [29]. Thus, our finding that expression of p-CREB, ted by cancer cells carrying mutant TP53 [9]. These macro- p-Src, p-Yes, and p-FAK are profoundly inhibited by WEE1 phages promote tumor evasion through multiple mecha- and ATM co-inhibition is consistent with a role for these mol- nisms, including enhanced angiogenesis and tissue remod- ecules in PC cell migration. Prior to this study, there had eling, recruitment of MDSCs and CAFs, and inhibition of T been no reports of a relationship between Src/FAK/CREB cell function and proliferation, thus enhancing metastasis and the DDR pathway in PC. and invasion [33]. Here, we evaluated M2 macrophage An interesting recent study reported that PARP inhibition recruitment to the tumor in the mouse xenograft model by upregulates PD-L1 expression in breast cancer via inactiva- IHC staining of CD163, and we found that the high basal lev- tion of GSK-3! [16]. We also found a similar increase in els of tumor-associated CD163+ cells were markedly decrea- PD-L1 expression in PARP inhibitor (Olaparib)-treated PC sed by AZD1775 and/or AZD0156 treatment. This finding cells (data not shown). In contrast, the present study showed supports the possibility that combination therapy suppresses here that targeting of the DDR by WEE1 and/or ATM inhi- recruitment of M2 macrophages to the tumor, and it will be bition reduced PD-L1 expression concomitantly with down- of interest to understand how other cells in the tumor micro- regulation of p–GSK-3! Ser9 level, the inactive form of GSK- environment are affected by DDR-targeting agents. 3!. This mechanism might at least partly explain the reduc- We used high concentrations of WEE1 inhibitor (AZD1775- tion in PD-L1 induced by WEE1 and/or ATM inhibition. It 1 µM) and ATM inhibitor (AZD0156 1 µM) in current in vitro is interesting to note that PARP, WEE1, and ATM have dis- experiments considering the active pharmacology of com- tinct effects on PD-L1 expression, despite the fact that they pounds. Particularly, AZD0156 1 µM might have off-target are all core members of the DDR signaling network. Thus, effects. while PARP inhibition increases PD-L1 expression by enhan- Taken together, our results demonstrate that DDR-target- cing GSK-3! inactivation, WEE1 and/or ATM inhibition ing agents such as WEE1 and ATM inhibitors have potent, does the opposite by decreasing the expression of inactive synergistic anti-tumor effects in PC. We also identified cro- GSK-3!. sstalk between the DDR and immune response in PC, since Among the mechanisms known to regulate PD-L1 expres- WEE1/ATM inhibition downregulates CXCR2, CD163, and sion, many act at the transcriptional level, including the JAK- PD-L1 expression, the latter by blocking expression of inac- STAT pathway, c-Myc, and NF-"B [15,30,31]. Consistent with tive p–GSK-3! and CMTM6. The findings in this study sup- this literature, we also observed a reduction in p-STAT-1 and port further clinical development of DDR-targeting strategies p-NF-"B concomitant with PD-L1 downregulation in WEE1 for the treatment of PC. and/or ATM inhibitor-treated cells. More recently, Burr and colleagues used CRISPR/Cas9 technology to screen approx- Electronic Supplementary Material imately 20,000 genes in the human PC cell line BxPC-3, and they identified a novel protein, CMTM6, as a major regulator Supplementary materials are available at Cancer Research and of PD-L1 expression [12]. Importantly, we observed that Treatment website (https://www.e-crt.org). CMTM6 binds to PD-L1 in PC cells and that CMTM6 expression was reduced upon inhibition of WEE1 and/or ATM. As Conflicts of Interest mentioned earlier, CMTM6 could rescue and recycle the PD-L1 from the lysosome to the cell surface. More recently, D-Y Oh received AstraZeneca-KHIDI (Korea Health Industry an interesting study revealed that HIP1R facilitates the deli- Development Institute) Oncology Research Program and Y-J Bang very of PD-L1 to the lysosome [32]. Here, we speculated two is a consultant/advisory board member of AstraZeneca. No con- potential mechanisms of the CMTM6 reduction by WEE1 or flicts of interest were disclosed by the other authors. ATM. First of all, WEE1 or ATM might post-transcriptionally

164 CANCER RESEARCH AND TREATMENT Mei Hua Jin, Co-targeting of WEE1 and ATM in Pancreatic Cancer

Acknowledgments (grant No. 06-2016-2920), Research Resettlement Fund for the new faculty of Seoul National University (grant No. 800-2016-0312) and This research was supported by 3rd AstraZeneca-KHIDI (Korea Research fund from SNUH (grant No. 03-2017-0100). Health Industry Development Institute) oncology research program

References

1. Bailey P, Chang DK, Nones K, Johns AL, Patch AM, Gingras ogy. 2013;139:428-37. MC, et al. Genomic analyses identify molecular subtypes of 15. Sato H, Niimi A, Yasuhara T, Permata TB, Hagiwara Y, Isono pancreatic cancer. Nature. 2016;531:47-52. M, et al. DNA double-strand break repair pathway regulates 2. Choi Y, Oh DY, Kim K, Chie EK, Kim TY, Lee KH, et al. Con- PD-L1 expression in cancer cells. Nat Commun. 2017;8:1751. current chemoradiotherapy versus chemotherapy alone for 16. Jiao S, Xia W, Yamaguchi H, Wei Y, Chen MK, Hsu JM, et al. unresectable locally advanced pancreatic cancer: a retrospec- PARP inhibitor upregulates PD-L1 expression and enhances tive cohort study. Cancer Res Treat. 2016;48:1045-55. cancer-associated immunosuppression. Clin Cancer Res. 2017; 3. O’Connor MJ. Targeting the DNA damage response in cancer. 23:3711-20. Mol Cell. 2015;60:547-60. 17. Farina AR, Mackay AR. Gelatinase B/MMP-9 in tumour 4. Richer AL, Cala JM, O’Brien K, Carson VM, Inge LJ, Whitsett pathogenesis and progression. Cancers (Basel). 2014;6:240-96. TG. WEE1 kinase inhibitor AZD1775 has preclinical efficacy 18. Gialeli C, Theocharis AD, Karamanos NK. Roles of matrix in LKB1-deficient non-small cell lung cancer. Cancer Res. metalloproteinases in cancer progression and their pharmaco- 2017;77:4663-72. logical targeting. FEBS J. 2011;278:16-27. 5. Kaufman B, Shapira-Frommer R, Schmutzler RK, Audeh MW, 19. Zarogoulidis P, Katsikogianni F, Tsiouda T, Sakkas A, Katsiko- Friedlander M, Balmana J, et al. Olaparib monotherapy in giannis N, Zarogoulidis K. Interleukin-8 and interleukin-17 for patients with advanced cancer and a germline BRCA1/2 cancer. Cancer Invest. 2014;32:197-205. mutation. J Clin Oncol. 2015;33:244-50. 20. Chuang JY, Yang WH, Chen HT, Huang CY, Tan TW, Lin YT, 6. Chen CC, Kass EM, Yen WF, Ludwig T, Moynahan ME, et al. CCL5/CCR5 axis promotes the motility of human oral Chaudhuri J, et al. ATM loss leads to synthetic lethality in cancer cells. J Cell Physiol. 2009;220:418-26. BRCA1 BRCT mutant mice associated with exacerbated 21. Tang CH, Tsai CC. CCL2 increases MMP-9 expression and cell defects in homology-directed repair. Proc Natl Acad Sci U S motility in human chondrosarcoma cells via the Ras/Raf/ A. 2017;114:7665-70. MEK/ERK/NF-!B signaling pathway. Biochem Pharmacol. 7. Mohni KN, Kavanaugh GM, Cortez D. ATR pathway inhibi- 2012;83:335-44. tion is synthetically lethal in cancer cells with ERCC1 defi- 22. Topalian SL, Drake CG, Pardoll DM. Immune checkpoint ciency. Cancer Res. 2014;74:2835-45. blockade: a common denominator approach to cancer therapy. 8. Matheson CJ, Backos DS, Reigan P. Targeting WEE1 kinase in Cancer Cell. 2015;27:450-61. cancer. Trends Pharmacol Sci. 2016;37:872-81. 23. Li CW, Lim SO, Xia W, Lee HH, Chan LC, Kuo CW, et al. Gly- 9. Binnewies M, Roberts EW, Kersten K, Chan V, Fearon DF, cosylation and stabilization of programmed death ligand-1 Merad M, et al. Understanding the tumor immune microen- suppresses T-cell activity. Nat Commun. 2016;7:12632. vironment (TIME) for effective therapy. Nat Med. 2018;24:541- 24. Pilie PG, Tang C, Mills GB, Yap TA. State-of-the-art strategies 50. for targeting the DNA damage response in cancer. Nat Rev 10. Steele CW, Karim SA, Leach JD, Bailey P, Upstill-Goddard R, Clin Oncol. 2019;16:81-104. Rishi L, et al. CXCR2 inhibition profoundly suppresses metas- 25. Chen D, Lin X, Gao J, Shen L, Li Z, Dong B, et al. Wee1 inhi- tases and augments immunotherapy in pancreatic ductal ade- bitor AZD1775 combined with cisplatin potentiates anticancer nocarcinoma. Cancer Cell. 2016;29:832-45. activity against gastric cancer by increasing DNA damage and 11. Chen DS, Mellman I. Oncology meets immunology: the can- cell apoptosis. Biomed Res Int. 2018;2018:5813292. cer-immunity cycle. Immunity. 2013;39:1-10. 26. Chen WT, Ebelt ND, Stracker TH, Xhemalce B, Van Den Berg 12. Burr ML, Sparbier CE, Chan YC, Williamson JC, Woods K, CL, Miller KM. ATM regulation of IL-8 links oxidative stress Beavis PA, et al. CMTM6 maintains the expression of PD-L1 to cancer cell migration and invasion. Elife. 2015;4:e07270. and regulates anti-tumour immunity. Nature. 2017;549:101-5. 27. Dalton HJ, Armaiz-Pena GN, Gonzalez-Villasana V, Lopez- 13. Chatzinikolaou G, Karakasilioti I, Garinis GA. DNA damage Berestein G, Bar-Eli M, Sood AK. Monocyte subpopulations in and innate immunity: links and trade-offs. Trends Immunol. angiogenesis. Cancer Res. 2014;74:1287-93. 2014;35:429-35. 28. Kolli-Bouhafs K, Sick E, Noulet F, Gies JP, De Mey J, Ronde P. 14. Rosado MM, Bennici E, Novelli F, Pioli C. Beyond DNA repair, FAK competes for Src to promote migration against invasion the immunological role of PARP-1 and its siblings. Immunol- in melanoma cells. Cell Death Dis. 2014;5:e1379.

VOLUME 52 NUMBER 1 JANUARY 2020 165 Cancer Res Treat. 2020;52(1):149-166

29. Zhang ZX, Zhang WN, Sun YY, Li YH, Xu ZM, Fu WN. CREB Boyer-Guittaut M, et al. PD-L1 expression is regulated by both promotes laryngeal cancer cell migration via MYCT1/NAT10 DNA methylation and NF-kB during EMT signaling in non- axis. Onco Targets Ther. 2018;11:1323-31. small cell lung carcinoma. Oncoimmunology. 2018;7:e1423170. 30. Garcia-Diaz A, Shin DS, Moreno BH, Saco J, Escuin-Ordinas 32. Wang H, Yao H, Li C, Shi H, Lan J, Li Z, et al. HIP1R targets H, Rodriguez GA, et al. Interferon receptor signaling path- PD-L1 to lysosomal degradation to alter T cell-mediated cyto- ways regulating PD-L1 and PD-L2 expression. Cell Rep. 2017; toxicity. Nat Chem Biol. 2019;15:42-50. 19:1189-201. 33. Aras S, Zaidi MR. TAMeless traitors: macrophages in cancer 31. Asgarova A, Asgarov K, Godet Y, Peixoto P, Nadaradjane A, progression and metastasis. Br J Cancer. 2017;117:1583-91.

166 CANCER RESEARCH AND TREATMENT