dkk2 high dkk2 low - nature...supplementary figure 2. effects of dkk2 blockade on the polyps of the...
TRANSCRIPT
Supplementary Figure 1. Upregulation of DKK2 expression by APC-loss. (a) Upregulation of DKK2 expression in human CRC samples over normal colorectal samples and in MSS CRCs over MSI CRCs. The numbers in the chart denote the sample sizes. (b-d) Upregulation of DKK2 expression in mouse intestinal polyps. DKK2 mRNA levels isolated from polyps and normal intestines from 4 Apcmin/+ and normal mice, respectively, were determined by quantitative RT-PCR (data are presented as means±sem; Two-sided Student t-test, n=4) (b), whereas the DKK2 mRNA (c) and protein (d) in intestinal sections was detected by in situ hybridization and Immunostaining, respectively. The sections were processed and exposed simultaneously, and the immunostaining intensity was quantified by ImageJ software. Five independent sections per mouse with 5 mice per group were quantified. Data are presented as means±sem (two-sided Student t-test). Red signals are Dkk2 mRNA, whereas blue signals are nuclear counter stain in c. (e,f) Upregulation of DKK2 expression in APC-loss MC38 cells. DKK2 expression was determined by quantitative RT-PCR (data are presented as means±sem; two-sided Student t-test, n=4) using RNAs isolated from MC38 cells with or without the APC mutation (e) or from APC
mutant MC38 cells transfected with different β-catenin siRNAs (f). Western analysis of β-catenin levels is also shown. (g) Upregulation of DKK2 expression in APC-loss HCT116 human colon cancer cells. DKK2 expression was determined by quantitative RT-PCR. Data are presented as means±sem (Two-sided Student t-test, n=4). (h) Correlation between DKK2 expression and CRC patient survival rates. The overall and relapse-free survival rates of the high (top 15 percentile) and low (bottom 15 percentile) DKK2 expressers are compared using the TCGA provisional datasets of colorectal adenocarcinoma (P=0.031, n=56 for overall survival; P=0.036, n=50 for relapse-free survival) by the two-sided Mantel-Cox Log-Rank test. (i) Verification of the lack of DKK2 protein in DKK2-null mice. Polyps from Apcmin/+ (DKK2 WT) and Apcmin/+Dkk2-/- mice were analyzed by Western blotting using the anti-DKK2 antibody. The experiments were repeated twice.
0
5
1 0
1 5
2 0
0
1
2
3
4
5
0 .0
0 .5
1 .0
1 .5
0
5
1 0
1 5
0
2
4
6
1.0
0.0
-1.0
-2.0
Normal Tumor
DK
K2 e
xpre
ssio
n
(Log2)
0 50 100 150
100
80
60
40
20
0Ove
rall
surv
ival
(%)
Months
0 50 100
100
80
60
40
20
0
Rela
pse-fre
e
surv
ival
(%)
Months
c
h
DK
K2 m
RN
A (
fold
)e f
DK
K2 m
RN
A (
fold
) HCT116MC38 g
DK
K2 m
RN
A (
fold
)
SiCtr SiCtnnb1
1 2 3
Tubulin
-catenin
MC38Apc
-catenin
Tubulin
65
65
Tubulin
-catenin
APC: WT Mut
DKK2 highDKK2 low
DKK2 highDKK2 low
30
165
P<0.05
MSI MSS
0
-0.2
-0.4
0.4
0.2
P<0.01
DK
K2 e
xpre
ssio
n
(Z s
core
)
P=0.004
ApcMin/+ Dkk2-/-ApcMin/+WT
d
DK
K2 s
tain
(A
U)
DK
K2 m
RN
A (
fold
)
Normal Polyp
bP<0.001
P=0.011 P<0.001P<0.001
DKK2
Actin
-100
-75-50
-37
-100
-75
-50
-37
-100
-75
-25
-50 -50
i
ApcMin/+ Dkk2-/-ApcMin/+WT
a
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 2. Effects of DKK2 blockade on the polyps of the ApcMin/+ mice. (a,b) The anti-DKK2 antibody reduces tumor burdens in ApcMin/+ mice. Mice (16 weeks old male) were treated with 5F8 and an isotype antibody (IgG3) (10 mg/kg, once a week, i.p.) for 4 weeks. Polyps numbers in small intestines were counted. Data are presented as means±sem (IgG vs 5F8; n=5; two-way Anova). (c) The effect of 5F8 on tumor suppression depends on DKK2. ApcMin/+ and ApcMin/+DKK2-/- mice (10 weeks old female) were treated with 5F8 and an isotype antibody (IgG3) (8 mg/kg, twice a week, i.p.) for 6 weeks. Polyps numbers in small intestines were counted under a stereomicroscope after staining with methylene-blue. Data are presented as means±sem (5F8 vs IgG treated; n=5; two-way Anova). (d-f) DKK2-deficiency increases apoptosis and granzyme B-positive cells without affecting proliferation in the polyps of the ApcMin/+ mice. Histological sections of polyps collected from the ApcMin/+ and ApcMin/+Dkk2-/- mice (20 weeks old) were stained with anti-activated caspase 3, granzyme B or Ki67 antibody together with DAPI. Scale bars are 150 µm. Five independent sections per mouse were quantified from five mice per genotype. Data are presented as means±sem (Two-sided Student t-test). (g-h) Quantification of the numbers of infiltrated CD8+ and NK1.1+ cells in intestine tumors from ApcMin/+ and ApcMin/+DKK2-
/- mice based on flow cytometry analysis of dissected polyps. Five large polyps per mouse (20 weeks old) were collected and analyzed by flow cytometry after staining with CD45, CD8+, and NK1.1+. Data are presented as means±sem (ns, not significant; n=5; Two-sided Student t-test). (i) DKK2 blockade does not affect differentiation in the polyps. Relative expression of a number of intestinal epithelial cell stem and differentiation markers in the polyps dissected from IgG or 5F8 treated Apcmin/+ mice were examined by quantitative RT-PCR and normalized against the mRNA levels of GAPDH. Data are presented as means±sem (n=5; Two-sided Student t-test).
0
1 0
2 0
3 0
4 0
5 0
IgG 5F8
Ki67
0
1 0
2 0
3 0
ba
0
1 0
2 0
3 0
Poly
ps
Poly
ps
Activated caspase 3
GZMB
Dkk2-/-ApcMin/+
ApcMin/+ Dkk2-/-ApcMin/+
ApcMin/+
e
d
0
1 0
2 0
3 0
4 0
5 0
0
5
1 0
1 5
2 0
2 5
c
>5 1-5 <1
Size (mm)
Proximal Middle Distal
Poly
ps
WT Dkk2-/-
IgG
5F8
0
2 0 0 0 0
4 0 0 0 0
6 0 0 0 0
8 0 0 0 0
CD
8a
+;C
D4
5+
Ce
lls
/
1x
10
e6
Tu
mo
r C
ell
s
g
h
0
2 0 0 0
4 0 0 0
6 0 0 0
NK
1.1
+;C
D4
5+
Ce
lls
/
1x
10
e6
Tu
mo
r C
ell
s
ns
ns
Cells
/fie
ldC
ells
/fie
ld
0
2 0
4 0
6 0
8 0
1 0 0f
CK
-18
CK
-19
Gale
ct in-3
Lam
nin
a1
Lam
nin
a2
Odc1
AR
P-P
0
CE
A
MU
C1
EpC
AM
0 .0
0 .5
1 .0
1 .5Ig G
5 F 8
Fo
ld C
ha
ng
e
i
Cells
/fie
ld
IgG
5F8
IgG
5F8
P=0.001 P=0.019 P<0.001P=0.012 P<0.001P<0.001
P<0.001
P=0.005
P=0.031
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 3. Flow cytometry analysis of leukocytes in tumors or lymph nodes. (a-g) Flow cytometry analysis of tumor infiltrated leukocytes. C57BL/6 mice were inoculated s.c. with the MC38 cells. Treatment of 5F8 (10 mg/kg, every three days, i.p.) commenced at Days 9 and 12. Tumors were collected at Day 14 and analyzed by flow cytometry. Panels d-f are derived from Panel c, whereas Panel g is derived from e. Data for tumor growth (P=0.034 and 0.024 for Days 12 and 14, respectively; two-way Anova) and weight (two-sided Student’s t-test) are presented as means±sem (n=5). (h-k) Flow cytometry analysis of tumor draining lymph nodes. Inguinal lymph nodes were collected from mice described above and analyzed by flow cytometry. Data are presented as means±sem (n=5; Two-sided Student’s t-test).
0 .0
0 .5
1 .0
1 .5
2 .0
IgG 5F8
1.08 0.1 0.84 0.2
CD4-CD8-
NK
1.1
FCS
NK1.1+
IgG
5F8
MFI:
138 5
189 13
P=0.035
Counts
IgG
5F8
MFI:
79 4
95 8
CD8+
GZMB
Counts
0
200
400
600
800
1000
6 8 10 12 14
CD4
CD
8
IgG 5F8
NK
1.1
FSC
29.1 4.8 20.7 3.1
11.2 0.8 13.3 3.0
10.8 2.2 12.1 2.1
IgG 5F8
c
e
Gr-
1
23.8 5.1
35.0 8.7 45.2 9.5
20.8 3.9
IgG 5F8
CD11b
b
d
CD45 +
CD45 +
CD4-CD8-
a
Tum
or
volu
me (m
m3)
Days
IgG
5F8
Tum
or
weig
ht
(g)
IgG 5F8
GZMB
CD8+ NK1.1+
IgG
5F8
MFI:
284 5
431 59
P=0.012
f g
Counts
Counts
IgG
5F8
GZMB
MFI:
237 6
303 22
P=0.035
CD4 +
Foxp3
IgG 5F8
CD25
23 3.8 19 4.8
-103 0 103 104 105
23 3.8
105
104
103
0
-103
105
104
103
0
-103
-103 0 103 104 105
CD4
CD
8
IgG 5F8
31.4 1.9
17.4 0.5
31.8 3.0
18.4 1.5
CD45+h
j
GZMB
k
i
P=0.012
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 4. Effects of DKK2 blockade on immune effector cell activation. (a,b,e,f) Flow cytometry analysis of leukocytes in Peyer’s patches (PPs) of ApcMin/+ mice. Leukocytes from PPs of ApcMin/+ or ApcMin/+Dkk2-/- mice (20 weeks old) (a,b) or from ApcMin/+ mice injected with one dose of 5F8 (10 mg/Kg, i.p.) for 24 hours (e,f) were prepared and analyzed by flow cytometry. The populations shown were pre-gated for CD45. Data are presented as means±sem (Two sided Student’s t-test; n=5). (c,d) Flow cytometry analysis of additional markers of tumor infiltrated CD8+ and NK1.1+ cells described in Fig. 3g-h (n=10).
CD4
CD
8
GZMB
3.1±0.8
8.2±0.8 8.0±0.6
3.8±0.5
3.3±1.6 12.8±5.5
5F8IgG 5F8IgG
CD
8
*
3.7±1.9
14.2±2.9 15.4±2.1
3.1±1.1
CD4
CD
8
4.2±0.9 7.4±2.1
CD
8
a b
GZMB
*
f
Dkk2-/-ApcMin/+ApcMin/+ Dkk2-/-ApcMin/+ ApcMin/+
e
Ki6
7
T-b
et
Eom
es
p-m
TO
R
p-S
6
p-A
kt
CD
122
CD
127
KLG
R1
0 .0
0 .5
1 .0
1 .5
2 .0Ig G
5 F 8
Ki6
7
T-b
et
Eom
es
p-m
TO
R
p-S
6
p-A
k t
KLR
G1
CD
122
CD
127
0 .6
0 .8
1 .0
1 .2
1 .4 Ig G
5 F 8
Rela
tive
M
FI
CD8+ NK1.1+dc
P=0.032
P=0.027
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 5. Cytotoxic immune cell depletion. (a-c) Confirmation of cell depletion efficiencies by flow analysis for Fig. 3i and j. (d) Effects of CD8+ or NK1.1+ depletion on polyp formation in the Apcmin/+ mice. Data are presented as means±sem (ns, not significant; n=6, two-way Anova for left panel; n=5, two-sided Student t-Test for right panel).
Iso αNK1.1
CD8
DX5
2.7
3.7
0.0
4.7
CD
8
CD4
Iso αCD820.7
13.3
0.2
15.3
NK
1.1
FSC
8.3 8.6
CD45+CD4-CD8-
CD45+
Iso αCD8
b
c
a CD45+
d
Poly
ps
IgG
5F8
ns80
60
40
20
0CD8
depletion
NK1.1
depletion
No
depletion
80
60
40
20
0
P=0.049
P<0.001P<0.001 P<0.001
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 6. DKK2 directly suppresses NK cell activation. (a) Reversal of DKK2-mediated inhibition of mouse NK cells by 5F8. NK cells were treated and analyzed as in Fig. 4d in the presence or absence of 36 µg 5F8 or IgG3. Data are presented as means±sem (**, p<0.001vs IgG or DKK2+5F8; n=5 except for IFNγ n=3; two-way Anova). (b) DKK2 inhibits mouse NK cells stimulated with IL12 and IL-18. Mouse primary NK cells were cultured with IL-12 and IL-18 (50 ng/ml) for 24 hrs. DKK2 protein (400 ng/ml) were then added for another 24 hours, followed by flow cytometry analyses. Data are presented as means±sem (**p<0.001 vs Mock+IgG or DKK2+5F8; n=3; Two-way Anova). The experiments were repeated twice. (c) DKK2 inhibits human NK cells. Human NK cells were isolated from peripheral bloods pooled from multiple normal individuals and incubated with human IL-15 (50 ng/ml) with or without 400 ng/ml human DKK2 protein for 24 hours before analysis by flow cytometry. Data are presented as means±sem (n=3; two-way Anova). The experiments were repeated three times. (d) DKK2 inhibits IL-15-mediated activation of mouse and human primary CD8+ T cells. Primary mouse CD8+ cells were isolated from spleens, whereas human CD8+ cells were isolated from peripheral blood using negative selection. These cells were cultured in IL-15 + IL15Rα-Fc (50 ng/ml) for 4 days. DKK2 (400 ng/ml) was then added for 24 hours before flow analysis. Data are presented as means ± sem (n=3; two-way Anova). The experiments were repeated two times. (e) DKK2 inhibits mouse IECs. Mouse IECs were isolated from normal mouse intestines and incubated with IL-15 (200 ng/ml) with or without 400 ng/ml DKK2 protein for 24 hours before analysis by flow cytometry. Data are presented as means±sem (n=3; Two-sided Student’s t-test). The experiments were repeated twice.
0 .0
0 .5
1 .0
1 .5
GZ
MB
CD
69
IFN
KLR
G1
CD
122
CD
107a
CD
314
NK
p46
CD
127
Eom
es
T-b
et
0 .0
0 .5
1 .0
1 .5
Re
lati
ve
MF
I
Ig G
5 F 8
D K K 2 + Ig G
D K K 2 + 5 F 8
**
********
****
0 .0
0 .5
1 .0
1 .5
Rela
tive
M
FI
IgG 5F8 IgG 5F8
Mock
DKK2
c
a
bR
ela
tive
M
FI
Mock
DKK2
CD107CD69GZMB
α47+CD8α+
Gzm
b(M
FI)
Mock DKK2
e
d
G B C D 6 9 N K G 2 D IF N
0 .0
0 .5
1 .0
1 .5
G B C D 6 9 N K G 2 D IF N
0 .0
0 .5
1 .0
1 .5
Rela
tive
M
FI
Rela
tive
M
FI
Human CD8Mouse CD8
Mock
DKK2
P=0.013
P=0.008P=0.003
P=0.014 P=0.004
0 .6
0 .7
0 .8
0 .9
1 .0
1 .1
1 .2
P=0.021
P=0.001
**
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 7. DKK2 acts independently of Wnt-β-catenin signaling. (a) Wnt reporter gene assay. DKK2 (400 ng/ml), and Wnt3a (100 ng/ml), and GSK3 inhibitor CHIR (1 µM) were added for 6 hrs to cells transfected with TOPFLASH the day before as indicated. Data are presented as means±sem (Two-way Anova; n=3). The experiments were repeated twice. (b) LRP5-deficiency does not affect WNT3A-indcued accumulation of β-catenin in primary mouse NK cells. Isolated primary mouse NK cells were expanded with IL-15 (50 ng/ml) for 24 hrs and then incubated with WNT3A (100 ng/ml) for 24 hours and analyzed by Western. The β-catenin blots were quantified and are shown as means±sem (two-way Anova, n=3). The experiments were repeated twice.
0
1
2
3
0
2
4
6
Mock DKK2CHIR
Rela
tive
W
nt re
port
er
gene a
ctivi
ty (
fold
)
Mock
Wnt3a
a
β-cat
Actin
Mock:
WNT3A:
WT Lrp5-/-
+
-
+
-
-
+
-
+
b
-100
-75
-50
-37
P=0.004 P=0.009
β-c
at
(fold
)
P=0.007
P<0.001
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 8. DKK2 impedes phospho-STAT5 nuclear localization. (a-c) DKK2 does not affect the MTORC1 pathway, phosphorylation of STAT1 or STAT4, or cell size. NK cells were prepared and analyzed as in Fig. 5a. The experiments were repeated twice. (d-e) Analysis of RNA sequencing results reveals relationship of DKK2-treatment to STAT signaling in mouse NK cells. Mouse NK cells were prepared and treated as Fig. 4d, and mRNAs isolated from these NK cells were subjected to NGS. Panel d shows pathway enrichment, whereas Panel e shows alterations in expression of STAT5 motif genes. Gene names are listed in Supplementary Table I. (f,h) Individual channels for Figure 5b and c. (g) DKK2 alters distribution of p-STAT5. NK cells were treated as in Fig. 5a and subjected to subcellular fractionation. The Nuclear (Nu) and non-nuclear (Non-Nu; cytosolic and membrane fraction pool) fractions were analyzed by Western blotting. The p-STAT5 blots were quantified and are shown as means±sem (Two-sided Student’s t-test, n=3).
DKK2 (10 nM)Mock
0 .9 0
0 .9 5
1 .0 0
1 .0 5
1 .1 0
Ge
om
etr
ic M
ea
n
(%
co
ntr
ol)
M o c k
D K K 2
FSC
SS
C
50K 100K 150K 200K
200K
Mock
DKK2
150K
100K
50K
0
c
-75
-50
-250
-75
-50
-50
-75
-50
-37
pS6K
pMTOR
Mock DKK2
IL-15
AKT
pAKT
Actin
-100
-50
-75
-100
-75
-37
pSTAT4
pSTAT1
Mock DKK2
IL-15
Actin
a
DAPIRAB8 pStat5
IgG
3D
KK
2M
ock
f
h
dTN
FA
_S
IGN
ALIN
G_V
IA_
NF
KB
JAK
_S
TA
T3_S
IGN
AL
ING
AP
ICA
L_JU
NC
TIO
N
STA
T5_S
IGN
ALIN
G
XE
NO
BIO
TIC
_M
ETA
BO
LIS
M
CO
AG
ULA
TIO
N
3.0
2.0
1.0
0
FD
R q
-valu
e (L
og
10)
b
e
gNuNon-Nu
5F
8
DKK2: - + - +
-100
-75
-250
-150
-25
p-STAT5
Histone
LRP5
P-S
TA
T5 (
fold
)
0
2
4
6
P=0.023
P=0.025
SSC FSC
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 9. DKK2 suppresses NK cells via LRP5, but not LRP6. (a) LRP6 is not required for DKK2-mediated inhibition of NK activation. Primary mouse NK cells were prepared from WT and Lrp6-/- mice and treated as Fig. 6a, followed by flow cytometry and Western analysis. Data are presented as means±sem (n=3; two-way Anova). The experiments were repeated twice. (b) LRP6 is required for WNT3A-induced stabilization of β-catenin in NK cells. Mouse NK cells were treated as in Supplementary Fig. 7b. The β-catenin blots were quantified and are shown as means±sd (Two-way Anova, n=3). (c,d) Flow cytometry analyses of infiltrated leukocytes in tumors described in Fig. 6c. Data are presented as means±sem (**, p<0.001; n=5; two-way Anova). (e-g) LRP5 in NK cells has a significant role in DKK2 blockade-mediated tumor suppression and NK cell activation. NK-specific knockout mice were generated by transferring bone marrows of WT or NCR1-CreLrp5f/f mice into WT recipient mice, followed with inoculation of MC38 cells. Treatment of 5F8 (10 mg/kg, i.p.) was given at Day 14, 19, and 22, followed by flow cytometry. Data are presented as means±sem (P=0.002 IgG+WT vs 5F8+WT in e; **, p<0.001; n=5 in f; two-way Anova). (h) A model for DKK2 to impart tumor immune evasion. DKK2 produced by tumor cells and possibly tumor infiltrated stromal cells binds to LRP5 on NK cells, leading to sequestration of phospho-STAT5 to endosomes and reduction in its nuclear localization. This in turn leads to an impediment in NK cell activation including a reduction in granzyme B production and attenuated NK-mediated tumor cell killing.
0
1
2
3
0
2
4
6
G B C D 6 9 I F N
0
1
2
3
4
G B N K p 4 6 I F N
0
1
2
3
0
1
2
3
200
300
400
500
600
700
800
14 16 18 20 22
-100
-75
-50
-37
WT Lrp6-/-
WNT3A: +- +-
β-cat
Actin
NK
CD107 CD69 CD314 GZMB
Rela
tive
M
FI
5F8+Lrp5-/-
IgG+WT
IgG+Lrp5-/-
5F8+WT
****CD8
CD107 CD69 CD314 GZMB
Rela
tive
M
FI
**
5F8+Lrp5-/-
IgG+WT
IgG+Lrp5-/-
5F8+WT
ba
d
Rela
tive
G
ZM
B M
FI Mock
DKK2WT Lrp6-/-
LRP6
LRP5-250
-150
-250
-150
c
WNT3A: +- +-
WT Lrp6-/-
β-c
at
(fold
)
P<0.001
Rela
tive
M
FI
Rela
tive
M
FI
e f
g
h
**
**
Tum
or
volu
me (m
m3)
Day
5F8+NK-Lrp5-/-
IgG+WT
IgG+NK-Lrp5-/-
5F8+WT5F8+NK-Lrp5-/-
IgG+WT
IgG+NK-Lrp5-/-
5F8+WT
5F8+NK-Lrp5-/-
IgG+WT
IgG+NK-Lrp5-/-
5F8+WT
****
NK
CD8
0 .5
1 .0
1 .5
P=0.004P=0.011
**
**
**
P=0.045P=0.002
P=0.016
WT Lrp6-/-
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 10. DKK2 suppresses responses elicited by PD-1 blockade. (a) Augmented anti-tumor effects of DKK2 and PD-1 blockade combination in the MC38 tumor model. C57BL/6 mice were inoculated s.c. with MC38 cells. Treatment of IgG (20 mg/kg, n=8), 5F8+IgG (n=7), anti-PD-1+IgG (n=7), and 5F8+anti-PD-1 (n=8)(10 mg/kg each) in 100 µl was done at every 5 days starting Day 18. Survival was evaluated by the two-sided Log-rank (Mantel-Cox) multiple comparison test with Bonferroni correction (P<0.001, IgG vs combo; P=0.003, IgG vs 5F8, P=0.003 IgG vs αPD-1; P=0.082, combo vs anti-PD-1; P<0.012, combo vs 5F8). (b-d) Effects of the antibody treatments on cytotoxic immune cells. 57BL/6 mice were inoculated s.c. with the MC38 cells. Treatments of 5F8 and/or anti-PD-1 (10 mg/kg, i.p) were done at Days 13 and 18. Tumors were collected for flow cytometry analysis on Day 20. Tumor growth data are presented as means±sem (P=0.025, IgG vs anti-PD-1; P=0.008, IgG vs 5F8; P<0.001 IgG vs combo for Day 20). Flow data are presented as means±sem (** p<0.01; two-way Anova). (e) Effect of recombinant DKK2 protein on cytotoxic immune cell responses to PD-1 blockade. C57BL/6 mice were inoculated s.c. with the MC38 cell. When tumors grew to 500 mm3, they were injected with DKK2 protein (600 ng/25 µl/tumor; multiple injection sites per tumor) for three times every 8 hours. One hour after the last inject, tumors were collected, and infiltrated leukocytes were analyzed by flow cytometry. Data are presented as means±sem (n=7; one-way Anova). (f) DKK2 blockade impedes progression of APC-loss MC38 tumors. C57BL/6 mice were inoculated s.c. with the APC mutant MC38 cells. Treatment of control IgG (20 mg/Kg), 5F8+IgG, anti-PD-1+IgG or 5F8+anti-PD-1 (10 mg/kg each) was done at Days 11, 16 and Day 20. Tumor growth data are presented as means±sem (P=0.016, IgG vs 5F8, two-way Anova for Day 20; P<0.001, 5F8 vs combo, Student’s t-test for Day 23; n=5). Survival was evaluated by the two-sided Log-rank (Mantel-Cox) multiple comparison test with Bonferroni correction (p=0.002, IgG vs 5F8 or combo; p=0.003, anti-PD-1 vs 5F8 or combo; p=0.008, 5F8 vs combo). (g) Effects of DKK2 blockade on immune effector cell activation in APC mutant MC38 tumors. The tumor implantation and antibody treatments were done as in f, but at Day 11 and 16. Tumors were collected at Day 18 and analyzed by flow cytometry. Data are presented as means±sem (**, p<0.001; Two-way Anova; n=5).
0
500
1000
1500
13 15 17 19 21
a
bDays
Surv
ival
(%)
500
1000
1500
GZM
B (
MF
I)
IgG
αP
D1
5F
8
Com
b
CD8
60
40
20Perc
ent
CD45
IgG
αP
D1
5F
8
Com
b
d
1000
2000
IgG
αP
D1
5F
8
Com
b
GZM
B (
MF
I)
NK
Tum
or
volu
me (m
m3)
Days
12
8
4
Perc
ent
CD8
IgG
αP
D1
5F
8
Com
b
16
20
IgG
αP
D1
5F
8
Com
b
10
Perc
ent
NK
6
4
2
8
e
1500 4000
1000
500
3000
2000
1000
+
-
-
-
+
-
-
+
+
+
-
-
-
+
-
-
+
+
GZM
B (
MF
I)
c
CD8 NK
IgG:
αPD1:
DKK2:
Perc
ent 40
30
20
50
+
-
-
-
+
-
-
+
+
CD45
IgG:
αPD1:
DKK2:
10
20
10
15
5
+
-
-
-
+
-
-
+
+
CD8
**
**
**
αPD-1
IgG
5F8+αPD-1
5F8
αPD-1
IgG
5F8+αPD-1
5F8
2 0 3 0 4 0
0
5 0
1 0 0
P<0.001
P<0.001
P=0.042P=0.005
P=0.015
P=0.040
P=0.016
P<0.001
P=0.016
1500
1000
500
1500
1000
500
1500
1000
500
1500
1000
500
Tum
or
volu
me (m
m3)
Days20 30 40 50 20 30 40 50 20 30 40 50 20 30 40 50
IgG 5F8 αPD1 Comb
n=8 n=7 n=7 n=8
2 0 2 2 2 4 2 6 2 8
0
5 0
1 0 0
C o n tro l
a n ti-P D 1
5 F 8
a n ti-P D 1 + 5 F 8
0
200
400
600
800
1000
1200
11 13 15 17 19 21 23
Days
Tum
or
volu
me (m
m3)
αPD-1
IgG
5F8+αPD-1
5F8
f
Days
Surv
ival
(%)
G B C D 6 9 C D 1 0 7 a C D 3 1 4 I F N
0
1
2
3
Re
lati
ve
MF
I
G B C D 6 9 N K p 4 6 C D 1 0 7 a C D 3 1 4 I F N
0
1
2
3
Re
lati
ve
MF
I
NK
CD8g
**
P=0.024**
P=0.044**
**
P=0.005
**
**
**
P=0.012 **
**
**
**
**
**
**
**
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 11. Effects of DKK2 and PD-1 combination blockade on a mouse melanoma tumor model. (a) Upregulation of DKK2 in human melanomas containing PTEN-loss and/or PI3K activated mutations. The numbers in the chart denote the sample size. (b) A trend of correlation between DKK2 expression upregulation and resistance to anti-PD-1 treatment in human melanomas. (c) PI3K inhibitor Wortmannin reduces DKK2 expression in YUMM1.7 cells. The cells were treated with Wortmannin (5 µM) for 24 hours and DKK2 mRNA levels were determined by qRT-PCR. Data are presented as means±sem (Two-sided Student’s t-test; n=4). The experiment was repeated three times. (d,e) Anti-tumor effects of DKK2 and PD-1 blockades in the YUMM1.7 tumor model. C57BL/6 mice were inoculated s.c. with the YUMM1.7 cells. Treatment of 5F8 and/or anti-PD-1 (10 mg/kg, i.p) was done at every 5 days starting Day 12. Survival was evaluated by the two-sided Log-rank (Mantel-Cox) multiple comparison test with Bonferroni correction (p<0.001, IgG vs combo; p=0.004, IgG vs 5F8, p=0.003 IgG vs αPD-1, p=0.018 combo vs 5F8, p=0.007 Combo vs αPD-1). Mean tumor growth data are presented as means±sem (P values for Day 30 are: <0.001, IgG vs combo; 0.045, IgG vs 5F8 or αPD-1; 0.023, combo vs 5F8; 0.025, combo vs αPD-1; two-way Anova). (f) Effects of antibody treatments on cytotoxic immune cells on YUMM tumors in the 57BL/6 mice. Treatments of 5F8 and/or anti-PD-1 were done at Days 16 and 20. Tumors were collected for flow cytometry analysis on Day 21. Data are presented as means±sem (**, p<0.01; Two-way Anova).
0 .0
0 .5
1 .0
1 .5
1000
500
20 25 30 40
Tum
or
volu
me (m
m3)
CD8
1000
2000
0
CD
69 (
MF
I)
NK
10
0
5
15
Perc
ent
50
40
30
20
10
0
CD8
IgG
αP
D1
5F
8
Com
b
Days
DKK2 expression
FP
KM
4
3
2
1
a
P=0.09
b
1.0
0.5
0
PTEN-/-
PIK3CAAC
-
-
+
-
-
+
+
+
P=0.006
P=0.019
P=0.045
Log2 m
edia
n-c
ente
red ra
tio
259
19
10
29
**
****
**
IgG
αP
D1
5F
8
Com
b
IgG
αP
D1
5F
8
Com
b
NK
G2D
(M
FI)
NK3000
2000
1000
**
IgG
αP
D1
5F
8
Com
b
or
αPD-1IgG 5F8+αPD-15F8
c
35 20 25 30 4035 20 25 30 4035 20 25 30 4035
n=8 n=10 n=9 n=10
P<0.001
Dkk2 m
RN
A (
fold
)
P=0.004
P=0.011
3 0 3 5 4 0
0
5 0
1 0 0
e
dS
urv
ival
(%)
Days
αPD-1
IgG
5F8+αPD-1
5F8
0
500
1000
1500
20 25 30
Tum
or
volu
me (m
m3)
Days
αPD-1
IgG
5F8+αPD-1
5F8
1000
2000
3000
0
4000NKCD8
GZM
B (
MF
I)
1000
2000
3000
0
****
IgG
αP
D1
5F
8
Com
b
IgG
αP
D1
5F
8
Com
b
f
P=0.034
P=0.039
P=0.054
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 12. Anti-DKK2 antibody suppresses tumor progress and activate immune effector cells in lungs of Apcfl/flKrasact/+ mice. The mice (9 weeks old) were instilled intra-nasally with 65 ul of MEM containing adenovirus expressing Cre (2x107 plague-forming unites) and CaCl2. After 6 weeks, the mice were treated weekly with IgG3 or 5F8 (10 mg/Kg). After another 6 weeks, the lungs (n=10 IgG and n=7 5F8) were perfused and the number of tumors visible at the surfaces are shown in a (Data are presented as means±sem; two-sided Student’s t-test). Some of the lungs are sectioned, and five histological sections from a lung of a representative IgG or 5F8-treated lung are also shown in a. The rest of the lungs were analyzed by flow cytometry (b-e). Data are presented as means±sem (**, p<0.001 vs IgG; n=5; Two-sided Student’s t-test).
CD
69
CD
107
CD
314
KLR
G1
0
1
2
3
Re
lati
ve
MF
I
Ig G
5 F 8CD8
CD
69
CD
107
CD
314
KLR
G1
0
1
2
3
4
5
Re
lati
ve
MF
I
Ig G
5 F 8
NK
P=0.0
04
P=0.0
16
P<0.0
01
P=0.0
02
P=0.0
36
P<0.0
01
P<0.001 P=0.006
a
b c
d e
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 13. Augmented anti-tumor effects of DKK2 and PD-1 combination blockade. (a) Correlations between DKK2 expression and cancer patient survival rates and granzyme B expression. The survival rates of the high (top 15 percentile) and low (bottom 15 percentile) DKK2 expressers are compared using the TCGA provisional datasets of kidney renal papillary carcinoma (P=0.006; n=43) and bladder urothelial carcinoma (P=0.012; n=61) by the two-sided Mantel-Cox Log-Rank test. Trends of correlation of DKK2 expression with granzyme B are also noted. (b,d) DKK2- or LRP5-deficiency does not alter the NK1.1+, CD8+, CD4+, or CD11b+Gr+ population in spleens. Mice of 8 weeks old female were used. Data are presented as mean±sem (n=5). (c,e) DKK2- or LRP5-deficiency does not alter NK cell developments. Mice of 8 weeks old female were used. Data are presented as mean±sem (n=5). Cells are pre-gated for the CD3e- CD19- NK1.1+ population. (f) Upregulation of LRP5 mRNA in tumor infiltrated NK cells. RNA was isolated from NK cells sorted from MC38 grafted tumors and primary NK cells from spleen and analyzed by qRT-PCR. Data are presented as mean±sem (n=4; two-sided Student t-test).
0
1 0
2 0
3 0
CD4CD8NK1.1 CD11b
Gr1
Cell
(%)
0
2 0
4 0
6 0
8 0Bone marrow
0
2 0
4 0
6 0
8 0
Cell
(%)
Cell
(%)
SpleenSpleenb c
d
WT
Dkk2-/- WT
Dkk2-/-
WT
Dkk2-/-
a
0
2
4
6
8
1 0
LR
P5 m
RN
A (
fold
)
P<0.001
0 50 100
100
80
60
40
20
0
Months
Ove
rall
surv
ival
(%)
f
0 50 100
100
80
60
40
20
0
Months
Ove
rall
surv
ival
(%)
200
Kidney renal papillary carcinoma Bladder urothelial carcinoma
DKK2 highDKK2 low
e
0
5
1 0
1 5
2 0
2 5
0
2 0
4 0
6 0
8 0
0
2 0
4 0
6 0
8 0Bone marrow SpleenSpleen
Cell
(%)
Cell
(%)
Cell
(%)
WT
Lrp5-/-
WT
Lrp5-/-
WT
Lrp5-/-
CD4CD8NK1.1 CD11b
Gr1
P=0.07
P=0.10
Log2 m
edia
n-c
ente
red ra
tio
DKK2 Low High
Kidney
Bladder
DKK2 highDKK2 low
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 14. Uncropped Western blots
Figure 6a
Figure 6d
Figure 6g
Figure 6f
Figure 6i
Figure 5a
P-STAT5
STAT5
P-ERK
P-AKT
AKT
GZMB
PFN
ACTIN
ERK
P-STAT5 STAT5
IL2/15Rβ
LRP5C
Rγc
LRP6
LRP5
LRP5
P-STAT5
STAT5
JAK1CA
LRP5C
LRP6
LRP5
LRP6
β-tubulin
STAT5
LRP5C
LRP5C
STAT5
Nature Medicine: doi:10.1038/nm.4496
Supplementary Figure 1ISupplementary Figure 1e
Supplementary Figure 1f
Supplementary Figure 1g
Supplementary Figure 9a
Supplementary Figure 9b
DKK2
Actin
β-catenin
β-tubulin
β-catenin
β-tubulin
β-catenin
β-tubulin
Supplementary Figure 7b
β-catenin
Actin
Supplementary Figure 8a
p-MTOR
p-S6K
p-AKT
AKT
Actin
Supplementary Figure 8b
p-STAT1
p-STAT4
Actin
Supplementary Figure 8g
p-STAT5
Histone
LRP5
LRP5
LRP6
β-catenin
Actin
Nature Medicine: doi:10.1038/nm.4496
Gene symbol 1.Control 2.Control 1.DKK2 2.DKK2
DACH2 1.134291418 0.837028498 -0.7989581 -1.17236182
ATP2A2 1.125646644 0.843963719 -1.18600457 -0.7836058
TRAF6 0.707956444 1.245124141 -1.12007019 -0.83301039
GABPA 0.795629313 1.145082364 -0.67660469 -1.26410698
CPSF7 0.61789126 1.317507259 -1.03662422 -0.8987743
LIF 1.267547989 0.666072892 -1.16709042 -0.76653046
KBTBD8 1.311749802 0.611822649 -0.79607561 -1.12749684
CXADR 0.961553377 0.960529814 -0.57014795 -1.35193524
RYBP 1.110673586 0.811190912 -1.32258949 -0.59927501
GZMB 1.327468792 0.586402749 -0.78015019 -1.13372135
MTF2 1.356238161 0.555187192 -0.84254954 -1.06887582
MBNL3 0.819714839 1.074392976 -1.38290085 -0.51120697
MBNL1 1.406649866 0.478676452 -0.85612604 -1.02920028
CNOT2 0.728268669 1.147095489 -1.38226127 -0.49310288
PHF17 0.494819716 1.377503379 -0.70720631 -1.16511679
ATP11C 0.635351836 1.231913801 -0.52413559 -1.34313004
SMC1A 0.900380767 0.962219247 -1.44550288 -0.41709713
FZD6 1.156898099 0.703866253 -1.39671741 -0.46404694
RUNX1 1.34830992 0.474703147 -1.29558267 -0.52743039
ELMO1 1.045394847 0.772946974 -1.48211031 -0.33623152
TMEM163 1.447204212 0.368095023 -1.15508685 -0.66021239
SPRED1 0.316037626 1.479245294 -0.67355086 -1.12173206
PURA 0.559594139 1.235114653 -0.37254187 -1.42216692
SPRY4 1.496913016 0.293910016 -0.70920631 -1.08161672
CMTM6 1.44587472 0.34237932 -1.20500999 -0.58324405
ZBTB20 1.175749257 0.599404997 -1.47184634 -0.30330792
AKAP1 0.268512591 1.502859469 -1.10977856 -0.6615935
PTCH1 1.480121654 0.219832553 -1.24746382 -0.45249038
ANKRD28 0.252777962 1.445609037 -0.3997168 -1.2986702
HMGN5 0.293850769 1.389849125 -0.31055508 -1.37314481
WIPF1 1.251465807 0.429709726 -1.48711978 -0.19405575
PPP2R3A 0.253338693 1.427711392 -0.3483386 -1.33271149
Prf1 1.450652674 0.195478548 -1.3243245 -0.32180672
OPN3 0.148401627 1.426873165 -0.19549527 -1.37977952
LCOR 0.115537556 1.410049502 -1.40882121 -0.11676585
Supplementary Table I_STAT5 regulated genes
Nature Medicine: doi:10.1038/nm.4496
qRT Primers
hDKK2 CGACACACCATGCAGGCCGA
hDKK2 CCTGGTCAGGCCGCCAATCG
mDKK2 TCAACTCCATCAAGTCCTCTC
mDKK2 TCACATTCCTTATCACTGCTG
SiRNA sequences
siCTNNB1-1
Mixed (GUGAAAUUCUUGGCUAUUA; GCGCUUGGCUGAACCAUCA; AAGCUGACCUGAUGGAGUU;
CAGCAAAUCAUGCGCCUUU)
siCTNNB1-2 GUGAAAUUCUUGGCUAUUA
siCTNNB1-3 GCGCUUGGCUGAACCAUCA
SgRNA guiding sequences
Cas9 mAPC-1F CACCGTTGGAGAGAGAGCGAGGTAT
Cas9 mAPC-1R AAACatacctcgctctctctccaaC
Cas9 mAPC-2F CACCGTGCCACACAATGGAACTCGG
Cas9 mAPC-2R AAACccgagttccattgtgtggcaC
hAPC 855aa Cas9-1F1 CACCGTTGGAGAGAGAACGCGGAAT
hAPC 855aa Cas9-1R1 AAACattccgcgttctctctccaaC
hAPC 900aa cas9-1F2 CACCGGTCTTCCTGAGAGGTATGAA
hAPC 900aa cas9-1R2 AAACttcatacctctcaggaagacC
hAPC 1345aa cas9-2F1 CACCGGTTTATCTTCAGAATCAGCC
hAPC 1345aa cas9-2R1 AAACggctgattctgaagataaacC
hAPC 1385AA cas9-2F2 CACCGTACATCTGCTAAACATGAGT
hAPC 1385AA cas9-2R2 AAACactcatgtttagcagatgtaC
Supplementary Table II_Primer sequences
Nature Medicine: doi:10.1038/nm.4496