Effective combinatorial immunotherapy for castration-resistant prostate cancer

Article abstract of DOI:10.1038/nature21676

A significant fraction of patients with advanced prostate cancer treated with androgen deprivation therapy experience relapse with relentless progression to lethal metastatic castration-resistant prostate cancer (mCRPC). Immune checkpoint blockade using a

Full text of DOI:10.1038/nature21676

LEttER
doi:10.1038/nature21676
Effective combinatorial immunotherapy for
castration-resistant prostate cancer
Xin Lu1†, James W. horner2, Erin Paul2, Xiaoying Shang1, Patricia troncoso3, Pingna Deng1, Shan Jiang2, Qing Chang2,
Denise J. Spring1, Padmanee Sharma4, John A. Zebala5, Dean Y. Maeda5, Y. Alan Wang1 & Ronald A. DePinho1
A significant fraction of patients with advanced prostate cancer model of mCRPC to efficiently test combination therapies in an
treated with androgen deprivation therapy experience relapse autochthonous setting. Combination of anti-CTLA4 and anti-PD1
with relentless progression to lethal metastatic castration-resistant engendered only modest efficacy. Targeted therapy against
prostate cancer (mCRPC)1. Immune checkpoint blockade using mCRPC-infiltrating MDSCs, using multikinase inhibitors such
antibodies against cytotoxic-T-lymphocyte-associated protein 4 as cabozantinib and BEZ235, also showed minimal anti-tumour
(CTLA4) or programmed cell death 1/programmed cell death 1 activities. Strikingly, primary and metastatic CRPC showed
ligand 1 (PD1/PD-L1) generates durable therapeutic responses robust synergistic responses when immune checkpoint blockade
in a significant subset of patients across a variety of cancer types2. was combined with MDSC-targeted therapy. Mechanistically,
However, mCRPC showed overwhelming de novo resistance combination therapy efficacy stemmed from the upregulation
to immune checkpoint blockade3–5, motivating a search for of interleukin-1 receptor antagonist and suppression of MDSC-
targeted therapies that overcome this resistance. Myeloid-derived promoting cytokines secreted by prostate cancer cells. These
suppressor cells (MDSCs) are known to play important roles in observations illuminate a clinical path hypothesis for combining
tumour immune evasion6. The abundance of circulating MDSCs immune checkpoint blockade with MDSC-targeted therapies in the
correlates with prostate-specific antigen levels and metastasis in treatment of mCRPC.
patients with prostate cancer7–9. Mouse models of prostate cancer
Mouse models of prostate cancer (PCa) engineered with signature
show that MDSCs (CD11b+Gr1+) promote tumour initiation10 mutations of human PCa exhibit autochthonous tumour evolution in
and progression11. These observations prompted us to hypothesize an intact immune system12–14. However, traditional germline genetic
that robust immunotherapy responses in mCRPC may be elicited modelling has limited capacity in generating the cohort sizes needed to
by the combined actions of immune checkpoint blockade agents conduct multi-arm drug testing. This issue is particularly pressing for
together with targeted agents that neutralize MDSCs yet preserve PCa models based on PB-Cre12, with optimal intercrosses producing
T-cell function. Here we develop a novel chimaeric mouse 12.5% PCa-prone males (Extended Data Fig. 1a). Here, we employed a
PB-Cre+ PtenL/L p53L/L Smad4L/L mTmGL/+ LSL-LUCL/+ (CPPSML mouse)
a
c
3.1 months 3.4 months
4.7 months 5.2 months
MRI
3 months
5 months
b
d
Germline
PB-Cre+ PtenL/L
Germline
Chimaera
MRI, T0
Prostate > 150 mm3
Enzalutamide
diet, 3 weeks
Chimaera
Castrate
MRI, T3 weeks
CPPSML
Prostate tumour
Metastasis to draining lymph node
#
Micrometastasis to lung
#
e
f
#
****
***
****
**
**
**
****
**
***
**
**
****
***
*
***
****
5
4
3
2
1
0
5
4
3
2
1
0
25
20
15
10
5
**
0
PB-Cre+ PtenL/L mice and CPPSML mice (n =5, biological replicates).
d–f, Preclinical trial results of prostate tumour mass, lymph node
metastasis score, and lung micrometastasis number (n =11, 6, 7, 4, 4,
9, 6, and 7, respectively, biological replicates). Red bar, mean. *P <0.05,
**P<0.01, ***P<0.001, ****P<0.0001, #P>0.05, Mann–Whitney U-test.
Figure 1 | Strong combination synergy by ICB with cabozantinib or
BEZ235 in mCRPC. a, Spontaneous prostate tumour development in the
CPPSML model with tumours detected by fluorescence, bioluminescence,
and MRI. Scale bar, 5 mm. b, c, Procedures and representative MRI images
for testing prostate tumour response to castration and enzalutamide diet in
1Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA. 2Institute for Applied Cancer Science, The University of Texas MD Anderson
Cancer Center, Houston, Texas 77030, USA. 3Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA. 4Department of Genitourinary Medical
Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA. 5Syntrix Biosystems, Inc., Auburn, Washington 98001, USA. †Present address: Department of Biological
Sciences, University of Notre Dame, Notre Dame, Indiana 46556, USA.
0 0 M o n t h 2 0 1 7
| V o L 0 0 0 | n A t U R E | 1
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
novel non-germline mCRPC model in a C57BL/6 background through phase trials but failure to improve overall survival in phase III trials
first establishing JH61 and JH58 mouse embryonic stem cell (mES cell) of mCRPC, and/or (3) purported immunomodulatory activities and
clones (Extended Data Fig. 1b–d) derived from the following genotypes: thus the potential to enhance or negate immune checkpoint blockade
PB-Cre+ PtenL/L p53L/L Smad4L/L mTmGL/+ LSL-LUCL/+ (CPPSML), (ICB). The agents selected were the tyrosine kinase inhibitors dasatinib
which exhibited age-dependent green fluorescent protein (GFP+)LUC+ (Dasa)15 and cabozantinib (Cabo)16, and the phosphoinositide 3-kinase
PCa development (Fig. 1a). In high-percentage chimaeras derived from (PI3K)/mTOR dual inhibitor BEZ235 (BEZ)17,18. Of relevance to this
JH61 or JH58 mES cells (Extended Data Table 1a), 50% of mice (4 out study, previous evidence suggests that PI3K pathway activation in both
of 8 necropsied) developed GFP+ cancer cells at 3 months of age and cancer cells and cancer-associated myeloid cells can mediate immuno-
showed dissemination of cancer cells to draining lymph nodes and lung suppression19–23 and that BEZ exhibits minimal inhibitory activity
(Extended Data Fig. 1e, f). In prostate, GFP+ areas corresponded to on mouse T cells24. For ICB, we used a cocktail of anti-CTLA4 and
CK8+/CK5+ adenocarcinoma (Extended Data Fig. 1g, h). Importantly, anti-PD1 antibodies to maximize the blockade of checkpoint path-
we observed a fourfold increase in the rate by which PCa-bearing mice ways, a regimen in line with a clinical treatment protocol currently
can be generated using chimaeric modelling (Extended Data Fig. 1a). being tested in a phase II trial for mCRPC (NCT02601014). CPPSML
To study combination therapy targeting mCRPC, we first employed the chimaeras (generated from JH61) with induced and MRI-documented
PB-Cre+ PtenL/L p53L/L Smad4L/L germline model and demonstrated mCRPC were randomized to receive single or combination treatments
that an androgen deprivation therapy protocol (castration followed by for 4 weeks before endpoint analysis (Extended Data Fig. 2d). While
enzalutamide-admixed diet) generated a significant, albeit transient, all targeted agent monotherapies or dual ICB cocktail had minimal
survival benefit (Extended Data Fig. 2a). Next, CPPSML chimaeras impact on prostate tumour mass, the combination of Cabo+ICB or
were subjected to the same androgen deprivation therapy to induce BEZ+ICB showed potent synergistic efficacy in targeting primary
CRPC. To ensure consistency, MRI was used to assign chimaeras with and metastatic PCa growth (Fig. 1d–f and Extended Data Fig. 2e). In
prostate tumour volumes over 150mm3 before 18 weeks of age: 86 out contrast, ICB alone or Dasa+ICB showed minimal impact on primary
of 107 (80.4%) chimaeras met this criterion (Fig. 1b and Extended Data or metastatic disease burden, although ICB alone resulted in signifi-
Fig. 2b). We validated emergence of CRPC in CPPSML chimaeras by cant reduction of lymph node metastasis and lung micrometastasis
comparing the response of size-matched primary prostate tumours (Fig. 1d–f and Extended Data Fig. 2e). In a corroborating study with
with androgen deprivation therapy in three cohorts: the chimaeras, chimaeras derived from JH58, Cabo+ICB also generated significant
CPPSML mice through breeding, and castration-sensitive PB-Cre+ efficacy in the mCRPC setting (Extended Data Fig. 3a–c). At necropsy,
PtenL/L mice (Fig. 1c). All treated chimaeras succumbed to primary Cabo+ICB- and BEZ+ICB-treated CRPC mice showed minimal
CRPC, with metastases in lymph nodes and micrometastases in lungs residual tumour cells in the prostate (Extended Data Fig. 2e), reduced
(Extended Data Fig. 2c). Thus, the CPPSML chimaera models provide proliferation, and pronounced apoptosis (Extended Data Fig. 3d–g).
a speedy platform to test multiple therapies on mCRPC.
We catalogued the constellation of intratumoural immunocytes
Next, mCRPC-bearing chimaera mice were assigned to therapeutic by time of flight mass cytometry (CyTOF)11 in the various treatment
trials. The targeted agents were selected on the basis of (1) strong arms as part of end-point analyses. Continued Dasa, but not Cabo
activity in preclinical PCa models, (2) initial activity and safety in early or BEZ, treatment was associated with a significant reduction of
a
b
c
Gr-MDSC (IC50 = 1.6 μM)
d
Gr-MDSC (IC50 = 0.17 μM)
CD8+ T cell (IC50 = 11.7 μM)
GFP+ PCa cell (IC50 = 3.2 μM)
CD8+ T cell (IC50 = 13.9 μM)
GFP+ PCa cell (IC50 = 95.6 μM)
***
***
***
***
*
**
*
80
60
40
20
0
15
10
5
100
50
0
100
50
0
0
–2 –1
0
1
2
–2 –1
0
1
2
log(Cabo concentration)
log(BEZ concentration)
Moderate prolif.
High prolif.
No prolif.
CD8+ T cells
e
f
g
h
i
CD4+ T cells
CD8+ T cells
CD4+ T cells
CD8+, high prolif.
CD8+, low prolif.
CD4+, high prolif.
CD4+, low prolif.
100
80
100
10
1
80
60
40
20
0
1.2
0.9
0.6
0.3
0.0
100
50
0
60
40
0.1
20
0.01
Stimulation
Cabo (μM)
BEZ (μM)
0
Stimulation
Cabo (μM)
BEZ (μM)
Stimulation
Cabo (μM)
BEZ (μM)
Dasa (μM)
+
+
1
+
+
1
+
+
+
+
+
1
+
10
+
1
+
+
+
+
+
1
+
10
+
1
+
+
+
Stimulation
Cabo (μM)
BEZ (μM)
Dasa (μM)
+
+
1
+
10
+
1
+
+
+
10
10
10
10
10
Dasa (μM)
0.5 5
0.5 5
Dasa (μM)
0.5 5
0.5 5
Figure 2 | Cabozantinib and BEZ235 attenuate MDSC frequency and
immunosuppressive activity in the tumour microenvironment.
a, b, CyTOF quantification of intratumoural Gr-MDSC frequency and
CD8+ T/Treg ratio by indicated treatments (n =17, 8, 4, 6, 8, and 4,
respectively). Red bar, mean. *P <0.05, **P <0.01, ***P <0.001,
Mann–Whitney U-test. c, d, In vitro sensitivity to Cabo (c) and BEZ (d)
by MDSCs, CD8+ T cells, and GFP+ PCa cells isolated from CRPC in
CPPSML mice with IC50 indicated (n =3, biological replicates). e, T-cell
proliferation (prolif.) assay when co-cultured 1:1 with MDSCs isolated
from control or drug-treated CRPC. High and low proliferation was
defined as T-cell division ≥3 and ≤2, respectively (n ≥4, biological
replicates). f, g, Drug effect on CD8+ (f) and CD4+ (g) T-cell proliferation
assay. High, moderate, and no proliferation was defined as T-cell division
≥3, ≤2, and 0, respectively (n =3, biological replicates). h, i, Drug effect
on IFNγ secretion (by CD8+ T cells) and IL-2 secretion (by CD4+
T cells) in the assay in f and g measured by enzyme-linked immunosorbent
assay (ELISA) (n =3, biological replicates). In c, d, h, and i, data represent
mean s.d.
2
| n A t U R E | V o L 0 0 0 | 0 0 M o n t h 2 0 1 7
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
a
b
c
Control
BEZ
ICB
Cabo
Control
Cabo BEZ
Recombinant active
p70S6K + ERK2 (μg)
200
150
100
50
pMET
pVEGFR2
pErk
0
0.25 1.0
pErk
pAkt
pmTOR
pS6K
pS6
Vinculin
0
pS6
Vinculin
d
e
f
g
***
***
***
**
***
***
75
50
25
0
Control
Cabo + ICB
BEZ + ICB
**
*** ******
1.0
0.8
0.6
0.4
0.2
1.0
200
150
100
50
0.8
0.6
0.4
0.2
IL-1α (10 ng ml–1
)
)
)
)
0.0
Cabo
0.0
BEZ
+
+
+
+
0
IL-1β (10 ng ml–1
IL-1ra (50 ng ml–1
+
+
+
+
+
+
0
+
+
+
0
+
+
p70S6K;
ERK2 ( μg)
p70S6K;
ERK2 ( μg)
+
+
+
+
+
+
0
0.25 1.0
0
0.25 1.0
Anti-IL-1ra (2 μg ml–1
h
i
j
CM treated with vehicle
CM treated with Cabo
CM treated with BEZ
CM treated with vehicle
CM treated with Cabo
CM treated with BEZ
Unconditioned
Conditioned medium
300
200
100
0
1.5
1.0
0.5
0.0
100
10
1
***
***
***
***
0.1
Arg1 Ncf1 Ncf4 Cybb
Arg1 Ncf1 Ncf4 Cybb
Figure 3 | Cabozantinib and BEZ235 inhibit PI3K signalling and
modulate MDSC-regulating cytokines. a, Relative phospho-RTK
intensity measured with phospho-RTK array for treated CRPC (n =2,
replicates). h, RNA expression levels of indicated genes by intratumoural
MDSCs when cultured in conditioned medium by CPPSML PCa cell
lines, measured by quantitative polymerase chain reaction with reverse
biological replicates). b, Level of phospho-proteins in RTK and PI3K–Akt– transcription (qRT–PCR) (n =3, biological replicates). i, Effect on gene
mTOR pathways in treated tumours, detected with western blot.
c, p-Erk and p-S6 intensity in MDSCs when co-transfected with ERK2
and p70S6K, detected with western blot. d, e, Effect of ERK2 and p70S6K
on the resistance of MDSCs to Cabo (1.5 μM) or BEZ (0.15 μM), measured
with WST-1 assay (n =3, biological replicates). f, Quantification of tumour BEZ (1 μM), measured with cytokine array (n =2, biological replicates). In
cytokine levels in treated CRPC using cytokine array (n =2, biological
replicates). g, Trans-well MDSC migration assay (n =3, biological
expression by pre-treating CPPSML PCa cells with Cabo and BEZ before
conditioned medium was collected to culture MDSCs (n =3, biological
replicates). j, Quantification of cytokine levels in conditioned medium
from CPPSML PCa cell lines pre-treated with vehicle, Cabo (1 μM), or
a and d–j, data represent mean s.d. **P <0.01, ***P <0.001, Student’s
t-test. For gel source data, see Supplementary Fig. 1.
tumour-infiltrating T cells (Extended Data Fig. 4a), suggesting that GM-CSF and pre-conditioned for 12h by PCa cell lines established from
the small impact of Dasa+ICB may reflect depletion of T cells in the the CPPSML model (Extended Data Fig. 5c). Moreover, Cabo or BEZ
tumour microenvironment. This finding is consistent with reported treatment alleviated the suppressive activity of intratumoural MDSCs
Dasa inhibition of TCR-mediated signal transduction and prolifera- on CD4+ and CD8+ T-cell proliferation (Fig. 2e and Extended Data
tion25. CPPSML PCa tumours predominantly contain granulocytic Fig. 5d). On the other hand, CD8+ and CD4+ T-cell in vitro proliferation
MDSCs (Gr-MDSCs, CD11b+Gr1+Ly6G+Ly6Clow) (Extended Data was only moderately suppressed by Cabo or BEZ, yet completely
Fig. 4b). Either Cabo or BEZ treatment resulted in a significant reduc- blocked by Dasa (Fig. 2f, g and Extended Data Fig. 5e). Equivalent drug
tion of Gr-MDSCs, while ICB alone had no impact (Fig. 2a). However, effects on interferon (IFN)-γ and interleukin-2 (IL-2) production by
ICB significantly increased the ratio of CD8+ to regulatory T cells T cells were observed (Fig. 2h, i). In summary, Cabo and BEZ elicited a
(Treg) (Fig. 2b), a defining feature for ICB-based therapies. Cabo+ICB pronounced effect on the infiltration and activity of MDSCs.
and BEZ+ICB further elevated CD8+/Treg ratios (Fig. 2b). MDSC
Next, phospho-receptor tyrosine kinase (phospho-RTK) anti-
depletion with anti-Gr1 neutralizing antibody11 sensitized CRPC in body arrays were used to assess the effect of Cabo and BEZ on the
CPPSML to ICB (Extended Data Fig. 4c), arguing that MDSCs mediate phospho-RTK signalling in treated CPPSML tumours (Extended
de novo resistance to ICB and suggesting that Cabo or BEZ enhances Data Fig. 6a), revealing Cabo-induced downregulation of pEGFR,
ICB through diminishing MDSCs.
pErbB2, pErbB3, pAxl, and pPDGFRα, and partial downregulation
To explore the impact of these agents on cells in the tumour micro- of pEGFR, pErbB3, and pAxl through indirect effect by BEZ (Fig. 3a).
environment, in vitro viability assays were performed to audit Cabo, We further observed that Cabo or BEZ also reduced phosphorylated
BEZ, and Dasa activity on MDSCs, CD8+ T cells, and GFP+ PCa MET (pMET) and phosphorylated vascular endothelial growth factor
cells collected from CRPC in CPPSML. Relative to CD8+ T cells and receptor 2 (pVEGFR2) levels, and significantly suppressed PI3K–Akt–
GFP+ PCa cells, MDSCs displayed a significantly higher sensitivity mTOR signalling in CPPSML tumours (Fig. 3b). Correspondingly,
to Cabo and BEZ, but not to Dasa (Fig. 2c, d and Extended Data Cabo or BEZ decreased pS6 signalling in intratumoural Gr1+ MDSCs
Fig. 5a). Similar half-maximum inhibitory concentration (IC50) results in the CPPSML tumour microenvironment (Extended Data Fig. 6b, c).
were obtained when MDSCs were assayed in medium supplemented These findings raised the possibility that Cabo and BEZ compromise
with 10ng ml−1 granulocyte–macrophage colony-stimulating factor MDSCs through inhibition of PI3K signalling. To test this hypothesis,
(GM-CSF) (Extended Data Fig. 5b) or in medium supplemented with we sought to rescue the viability of Cabo- or BEZ-treated MDSCs with
0 0 M o n t h 2 0 1 7
| V o L 0 0 0 | n A t U R E | 3
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
****
**
****
****
a
***
***
****
****
***
****
#
*
#
#
#
#
#
**
#
*
*
*
*
8
6
4
2
0
25
20
15
10
5
5
4
3
2
1
0
#
**
***
**
0
b
GSK2636771 +
ICB
Control
ICB
PI-3065
PI-3065 + ICB GSK2636771
SX-682
SX-682 + ICB
T cells > 30% T cells < 30%
c
d
e
40
30
20
10
0
50
40
30
20
10
0
Dendritic cell
T cell
Macrophage
B cell
*P = 0.024
Gr-MDSC
NK
0
10
20
30
CD8+ T cells (%)
Figure 4 | Gr-MDSCs inversely correlate with CD8+ infiltrating
T cells in human PCa. a, b, Efficacy of isoform-selective p110 inhibitors
or Cxcr1/2 inhibitor as single agents or in combination with ICB to treat
mCRPC in CPPSML model (n =15, 8, 3, 6, 4, 5, 5, and 10 respectively,
biological replicates). Red bar, mean. *P <0.05, **P <0.01, ***P <0.001,
****P <0.0001, #P >0.05, Mann–Whitney U-test. Scale bars: 3 mm for
organ images, 200 μm for haematoxylin and eosin (H&E) images.
c, SPADE tree derived from CyTOF analysis of human samples. All live
infiltrating immune cells (CD45+) were used to construct the tree. NK,
natural killer cells. d, Frequency of infiltrating immune cell populations
from patient samples. e, Linear regression analysis of Gr-MDSC frequency
with CD8+ T-cell frequency. In c–e, n =12 (biological replicates).
enforced downstream activation of signalling surrogates. Specifically, medium may drive the gene upregulation. Through cytokine array
MDSCs were isolated from the induced CRPC in CPPSML and (Extended Data Fig. 7d), we identified ten cytokines significantly
co-transfected with recombinant active ERK2 and p70S6K proteins downregulated in the conditioned medium of PCa cells pre-treated
(Fig. 3c). Transfected MDSCs showed improved survival under Cabo with Cabo or BEZ (Fig. 3j), among which CCL5 was also identified as a
or BEZ treatment (Fig. 3d, e). Similar results were obtained when the downregulated cytokine in Cabo+ICB- and BEZ+ICB-treated CRPC
assay was performed in enhanced medium (Extended Data Fig. 6d, e). (Fig. 3f). When MDSCs isolated from CRPC tumours were treated with
These results reinforce the view that Cabo and BEZ exert an impact each of the ten cytokines in the presence of Cabo or BEZ, significant
on the PCa tumour microenvironment in part via selective depletion upregulation of Arg1, Cybb, Ncf1, and Ncf4 was observed with most
of MDSCs.
of the cytokines tested (Extended Data Fig. 7e). Therefore, PCa cells
As cytokine signalling plays a pivotal role in the recruitment and are capable of driving immunosuppression-related gene expression in
activation of MDSCs6, we further explored the impact of combination MDSCs through secretion of multiple cytokines, and this paracrine
treatment on cytokine production in primary CRPC. Cytokine arrays signalling is impaired by Cabo or BEZ treatment.
revealed that several key cytokines involved in regulating recruitment
BEZ targets multiple p110 isoforms. To test whether isoform-specific
and activity of immunosuppressive myeloid cells, including CCL5, PI3K inhibitors would generate a comparable level of synergistic
CCL12, CD40, and HGF, were reduced by Cabo+ICB or BEZ+ICB advantage when combined with ICB, we tested PI-3065 (p110δ-
treatment. These treatments were also associated with increased selective inhibitor27) and GSK2636771 (p110β-selective inhibitor28).
IL-1ra, CD142, and VEGF (Fig. 3f and Extended Data Fig. 7a). IL-1ra Moreover, given the critical role of Cxcr2 in MDSC recruitment10,11
may contribute to reduced MDSC infiltration26. We confirmed that and the downregulation of Cxcr2 ligands Cxcl1 and Cxcl2 in the con-
recombinant IL-1ra inhibited IL-1 induced chemoattraction of MDSCs, ditioned medium of PCa cells treated with Cabo and BEZ (Fig. 3j),
and this effect was blocked by IL-1ra neutralizing antibody (Fig. 3g). we also tested a novel clinical-stage Cxcr1/2 inhibitor SX-682 (Extended
Notably, the cytokine changes were significantly less pronounced in Data Fig. 7f, g) as monotherapy or in combination with ICB in the
Dasa+ICB treatment (Extended Data Fig. 7b). Noting that cytokine CPPSML model. Mild to moderate effects of PI-3065, GSK2636771,
production by cancer cells may influence the functional status of or SX-682 as single agents on CRPC progression were observed,
myeloid cells, we cultured MDSCs isolated from CRPC with con- yet combination with ICB produced strong efficacy (Fig. 4a, b).
ditioned medium from the CPPSML PCa cell lines (Extended Data To give further credential to our model and assess the pattern of
Fig. 7c) and showed upregulated expression of genes responsible for MDSCs in human PCa, a 32-antibody CyTOF panel was developed
MDSC-induced immune suppression, including Arg1, Cybb, Ncf1, and (Extended Data Table 1b) and used to analyse 12 fresh fine-needle
Ncf4 (refs 6, 11) (Fig. 3h). Critically, the expression induction was largely biopsy samples from ten treatment-naive PCa tumours (Extended Data
abolished if the PCa cells were pre-treated with Cabo or BEZ before Table 1c). Analysis with SPADE software displayed the heterogeneous
conditioned medium was collected (Fig. 3i), whereas direct treatment immune cell populations (Extended Data Fig. 8). Results showed the
of MDSCs with Cabo or BEZ caused insignificant expression changes prominence of Gr-MDSCs relative to Mo-MDSCs, with the latter
(data not shown), suggesting that certain cytokines in the conditioned occupying no nodes in the SPADE tree (Fig. 4c). Total T-cell load across
4
| n A t U R E | V o L 0 0 0 | 0 0 M o n t h 2 0 1 7
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
the samples varied considerably (Fig. 4d), yet the frequency of CD8+
T cells correlated inversely with the frequency of Gr-MDSCs (Fig. 4e),
a pattern consistent with the conservation of antagonistic activity of
Gr-MDSCs on CD8+ T cells in human PCa.
17. Carver, B. S. et al. Reciprocal feedback regulation of PI3K and androgen
receptor signaling in PTEN-deficient prostate cancer. Cancer Cell 19,
575–586 (2011).
18. Jia, S. et al. Opposing effects of androgen deprivation and targeted therapy on
prostate cancer prevention. Cancer Discov. 3, 44–51 (2013).
19. Peng, W. et al. Loss of PTEN promotes resistance to T cell-mediated
immunotherapy. Cancer Discov. 6, 202–216 (2016).
20. Kim, K. et al. Eradication of metastatic mouse cancers resistant to immune
checkpoint blockade by suppression of myeloid-derived cells. Proc. Natl Acad.
Sci. USA 111, 11774–11779 (2014).
21. Rivera, L. B. et al. Intratumoral myeloid cells regulate responsiveness
and resistance to antiangiogenic therapy. Cell Reports 11, 577–591
(2015).
22. De Henau, O. et al. Overcoming resistance to checkpoint blockade therapy by
targeting PI3Kγ in myeloid cells. Nature 539, 443–447 (2016).
23. Kaneda, M. M. et al. PI3Kγ is a molecular switch that controls immune
suppression. Nature 539, 437–442 (2016).
24. Frick, A. et al. Immune cell-based screening assay for response to anticancer
agents: applications in pharmacogenomics. Pharm. Genomics Pers. Med. 8,
81–98 (2015).
25. Schade, A. E. et al. Dasatinib, a small-molecule protein tyrosine kinase
inhibitor, inhibits T-cell activation and proliferation. Blood 111, 1366–1377
(2008).
26. Tu, S. et al. Overexpression of interleukin-1β induces gastric inflammation and
cancer and mobilizes myeloid-derived suppressor cells in mice. Cancer Cell
14, 408–419 (2008).
Here, by using a chimaeric model of mCRPC, we have shown that
ICB alone is insufficient to generate an effective response, but a combi-
nation of ICB with drugs that inactivate MDSCs demonstrates superior
synergistic efficacy against de novo resistance to ICB (Extended Data
Fig. 9). The differential sensitivity of MDSCs and CD8+ T cells to Cabo,
BEZ, and possibly other phosphokinase inhibitors should provide an
avenue for optimizing the dose and schedule for effective silencing of
MDSCs while simultaneously sparing cytotoxic T lymphocytes to attack
cancer cells. As with all preclinical model systems and human clinical
pathological correlations, prospective clinical trials will be needed to
substantiate the hypotheses of our work. Future studies should explore
the combination therapy in the context of both established mCRPC and
newly diagnosed PCa together with selective anti-androgens to achieve
durable clinical response in this major cancer of men.
Online Content Methods, along with any additional Extended Data display items and
Source Data, are available in the online version of the paper; references unique to
these sections appear only in the online paper.
27. Ali, K. et al. Inactivation of PI(3)K p110δ breaks regulatory T-cell-mediated
immune tolerance to cancer. Nature 510, 407–411 (2014).
28. Costa, C. et al. Measurement of PIP3 levels reveals an unexpected role for
p110β in early adaptive responses to p110α-specific inhibitors in luminal
breast cancer. Cancer Cell 27, 97–108 (2015).
received 15 November 2016; accepted 31 January 2017.
Published online 20 March 2017.
1. Watson, P. A., Arora, V. K. & Sawyers, C. L. Emerging mechanisms of resistance
to androgen receptor inhibitors in prostate cancer. Nature Rev. Cancer 15,
701–711 (2015).
2. Sharma, P. & Allison, J. P. The future of immune checkpoint therapy. Science
348, 56–61 (2015).
3. Kwon, E. D. et al. Ipilimumab versus placebo after radiotherapy in patients with
metastatic castration-resistant prostate cancer that had progressed after
docetaxel chemotherapy (CA184-043): a multicentre, randomised, double-
blind, phase 3 trial. Lancet Oncol. 15, 700–712 (2014).
4. Beer, T. M. et al. Randomized, double-blind, phase III trial of ipilimumab versus
placebo in asymptomatic or minimally symptomatic patients with metastatic
chemotherapy-naive castration-resistant prostate cancer. J. Clin. Oncol. 35,
40–47 (2017).
5. Topalian, S. L. et al. Safety, activity, and immune correlates of anti-PD-1
antibody in cancer. N. Engl. J. Med. 366, 2443–2454 (2012).
6. Gabrilovich, D. I. & Nagaraj, S. Myeloid-derived suppressor cells as regulators
of the immune system. Nature Rev. Immunol. 9, 162–174 (2009).
7. Vuk-Pavlovic´, S. et al. Immunosuppressive CD14+HLA-DRlow/− monocytes in
prostate cancer. Prostate 70, 443–455 (2010).
Supplementary Information is available in the online version of the paper.
Acknowledgements We thank N. Feng for advice on preclinical drug dosing;
X. Xu and K. O’Connor for technical support on mES cells and chimaera
generation; P. Jones and C. Carroll for providing 19F NMR spectroscopy data;
K. Zhao, Z. Xu, and Z. Fang for animal husbandry; F. Giancotti for suggestions
on manuscript writing; A. Varma and J. W. Han for technical assistance;
G. Wang, E.-J. Jin, P. Dey, and all members of the DePinho laboratory for a
variety of technical support and suggestions. The project was supported
by U01CA141508 (R.A.D.), P01CA117969 (R.A.D.), Department of Defense
Prostate Cancer Research Program (PCRP) Idea Development Award–New
Investigator Option W81XWH-14-1-0576 (X.L.), National Institutes of Health
grant R44HL072614 (D.Y.M.), and the Clayton & Modesta Williams Cancer
Research Fund (R.A.D.). Facility-based experimental support was provided
by the Small Animal Imaging Facility (C. Kingsley, V. Tran, K. Maldonado,
C. Kaffes), Flow Cytometry and Cellular Imaging Core Facility (J. Burks, D. Mak,
and K. Dwyer), and Laboratory Animal Genetic Services (F. Benavides) at The
University of Texas MD Anderson Cancer Center (Cancer Center Support Grant
P30CA016672).
8. Brusa, D. et al. Circulating immunosuppressive cells of prostate cancer
patients before and after radical prostatectomy: profile comparison. Int. J. Urol.
20, 971–978 (2013).
Author Contributions X.L., R.A.D., Y.A.W., and J.W.H. conceived the project and
discussed experiments; J.W.H. designed the methodology for, and oversaw,
the chimaeric modelling; X.L., E.P., and X.S. acquired and analysed the data;
P.D. performed mouse genotyping; X.S., P.D., and Q.C. performed histology
staining; P.T. collaborated on providing fresh human samples; S.J. oversaw and
performed animal colony management; D.Y.M. provided study drug SX-682
and technical assistance for its use; J.A.Z. suggested MDSC experiments with
SX-682 in the model; P.S. provided key suggestions on experiments; Y.A.W.
and R.A.D. supervised the research; X.L., R.A.D., Y.A.W., and D.J.S. wrote the
manuscript.
9. Hossain, D. M. et al. TLR9-targeted STAT3 silencing abrogates
immunosuppressive activity of myeloid-derived suppressor cells from prostate
cancer patients. Clin. Cancer Res. 21, 3771–3782 (2015).
10. Di Mitri, D. et al. Tumour-infiltrating Gr-1+ myeloid cells antagonize senescence
in cancer. Nature 515, 134–137 (2014).
11. Wang, G. et al. Targeting YAP-dependent MDSC infiltration impairs tumor
progression. Cancer Discov. 6, 80–95 (2016).
12. Shen, M. M. & Abate-Shen, C. Molecular genetics of prostate cancer: new
prospects for old challenges. Genes Dev. 24, 1967–2000 (2010).
13. Ding, Z. et al. Telomerase reactivation following telomere dysfunction yields
murine prostate tumors with bone metastases. Cell 148, 896–907 (2012).
14. Ding, Z. et al. SMAD4-dependent barrier constrains prostate cancer growth and
metastatic progression. Nature 470, 269–273 (2011).
15. Araujo, J. C. et al. Docetaxel and dasatinib or placebo in men with metastatic
castration-resistant prostate cancer (READY): a randomised, double-blind
phase 3 trial. Lancet Oncol. 14, 1307–1316 (2013).
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Readers are welcome to comment on the online version of the paper. Publisher’s
note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations. Correspondence and requests
for materials should be addressed to R.A.D. (rdepinho@mdanderson.org)
or Y.A.W. (yalanwang@mdanderson.org).
16. Smith, M. et al. Phase III study of cabozantinib in previously treated metastatic
castration-resistant prostate cancer: COMET-1. J. Clin. Oncol. 34, 3005–3013
(2016).
reviewer Information Nature thanks D. Gabrilovich, M. Galsky, R. Madan and the
other anonymous reviewer(s) for their contribution to the peer review of this work.
0 0 M o n t h 2 0 1 7
| V o L 0 0 0 | n A t U R E | 5
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
MethOdS
Specifically, cell lines were cultured to roughly 80% confluence on inactivated
Ethics statement and transgenic mice. All animal work performed in this mouse embryonic fibroblasts in wells of a 24-well plate in mES cell culture media
study was approved by The University of Texas MD Anderson Cancer Center described above. Cells cultured for injection were split 11–16h before preparation
Institutional Animal Care and Use Committee. All animals were maintained for microinjection. Cells were removed from incubation, washed with sterile PBS,
in pathogen-free conditions and cared for in accordance with the International and trypsinized with 0.25% trypsin-EDTA. Trypsin was quenched with mES cell
Association for Assessment and Accreditation of Laboratory Animal Care policies culture media containing FBS and the cell culture plate was returned to incubation
and certification. All surgeries were performed with isoflurane anaesthesia. for 45min to allow the larger mouse embryonic fibroblasts to re-attach to the
Analgesic was administered after surgery along with temperature-controlled bottom of the cell culture well. After that incubation, the desired mES cells were
post-surgical monitoring to minimize suffering. PB-Cre+ PtenL/L p53L/L Smad4L/L collected and transferred to a 15ml conical tube. The suspension was brought
mice were described previously13. Mice were crossed to Rosa26-Lox-tdTomato-Lox- up to 5ml with mES cell culture media, gently pipetted up and down to mix, and
EGFP (mTmG) allele29 (The Jackson Laboratory, 007676) and Rosa26-pCAGGs-LSL- then spun down at 4°C and 800r.p.m. for 4min. All except 100–200μl of media
luciferase (LSL-LUC) allele (MMRRC, 01XAC), both of which were already congenic was aspirated from the tube. The bottom of the conical tube was tapped externally
to C57BL/6, and further backcrossed to C57BL/6 background for four generations. to break up the cell pellet and the tube was placed on ice for microinjection use.
Mice were then intercrossed to obtain ‘CPPSML’ males with homozygous
C57BL/6NTac-Tyrtm1Arte female mice that had achieved successful strain-
status of Pten, p53, and Smad4, heterozygous status of mTmG and LSL-LUC, matched mating after superovulation via timed gonadotropin administration were
and hemizygous status of PB-Cre. To calculate the expected frequency of pups used as donors for 3.5 day blastocysts. Blastocysts were collected by uterine flush
to develop prostate tumours, several factors were included: only males were and cultured in M16 medium (Millipore, MR-016-D) overlaid with embryo-tested
considered; PB-Cre had to be transmitted from males to the next generation owing oil in a 35mm culture dish at 37°C with 5% CO2. Blastocysts were each micro-
to the non-specific expression in oocytes30, thus only half of the litter was Cre+; manipulated to insert roughly 12 individual mES cells into its blastocoel. Injected
moreover, to circumvent paternal infertility caused by the diseased prostate, the blastocysts were then implanted into the uteri of pseudo-pregnant females. Each
key loxP-flanked tumour suppressor alleles (for example, Pten/p53 (ref. 31), or pseudo-pregnant female received up to 14 micromanipulated blastocysts (up to
Pten/Smad4 (ref. 14)) that were needed for aggressive PCa were maintained as seven blastocysts per uterine horn). Chimaeras were verified with prostate tumour
heterozygous state for the Pten loxP allele in the males for continued fertility (p53 formation by MRI and assigned into preclinical studies through randomization
or Smad4 could be homozygous)13,14, thus only half of the litter was homozygous which maximized the chance that male mice housed in the same cage received
for all tumour suppressor alleles. Mice with spontaneous prostate tumours were different drugs. Blinding was not applied, because allocation required investigators
euthanized at designated time points for tumour collection. Owing to the internal to assess mouse prostate tumour size with MRI before randomization. Pups were
status of the tumours, we used signs of lethargy, reduced mobility, and morbidity, excluded from tumour analysis if they died without significant sign of tumour
rather than maximal tumour size, as a protocol-enforced end point.
formation.
mES cell derivation. PtenL/L p53L/L Smad4L/L mTmGL/L or PtenL/L p53L/L Smad4L/L Single-nucleotide polymorphism analysis. Single-nucleotide polymorphism
LSL-LUCL/L female mice 3–4 weeks old were administered 5 international units analysis was performed in the Genetic Services core at MD Anderson Cancer
(IU) of pregnant mare serum gonadotropin (NHPP: National Hormone & Center. For background characterization of the JH61 ES cell line, 80 microsatellite
Peptide Program) by intraperitoneal injection at 15:00–16:00. Females were then (simple sequence length polymorphism, SSLP) markers, polymorphic between
administered 5 IU of human chorionic gonadotropin (Sigma-Aldrich, C1063) FVB/N and 129S6 inbred strains, and evenly distributed along the genome (that
by intraperitoneal injection 46h after pregnant mare serum gonadotropin and is, genome scan), were used as described32. To analyse the MHC region on mouse
mated 1:1 with PB-Cre+ PtenL/+ p53L/L Smad4L/L LSL-LUCL/L or PB-Cre+ PtenL/L chromosome 17, an extra 15 polymorphic SSLPs flanking the H2 complex were
p53L/L Smad4L/L mTmGL/L male mice, respectively. Successful mating was eval- added. The results (allelotype) obtained for the SSLP in the H2 region were con-
uated and noted by presence of copulus plug 14–16h after mating. Blastocyst- firmed by means of H2 polymerase chain reaction–restriction fragment length
stage embryos were isolated from those plugged females at embryonic day 3.5 polymorphism (PCR–RFLP) genotype as described33
.
by uterine isolation and flush. Blastocysts were collected and cultured on 16h Non-invasive mouse and ex vivo imaging. For MRI imaging with Bruker ICON,
pre-plated inactivated mouse embryonic fibroblasts in pre-equilibrated mES cell the animals were anaesthetized with 1–3% isoflurane and placed on the ICON
culture media. Culture media included knockout serum replacement (Invitrogen, animal bed. The MRI coil was secured into position over the animal and the entire
10828010), knockout DMEM (Invitrogen, 10829-018), GlutaMAX (Invitrogen, bed assembly was placed into the Bruker ICON 1T MRI bore. Rapid acquisition
35050079), sodium pyruvate (Invitrogen, 11360-070), non-essential amino with relaxation enhancement (RARE) T2-weighted images were acquired in both
acids (Invitrogen, 11140050), 2-mercaptoethanol (Invitrogen, 21985-023), the coronal and axial planes. After the imaging was completed, the animals were
penicillin–streptomyocin (Invitrogen, 15140-122), insulin (Sigma-Aldrich, allowed to recover under a heating lamp until fully conscious. MRI images were
12643), LIF (Millipore, ESG1107), mitogen-activated inhibitor (Stemgent, loaded into ImageJ to manually demarcate the contour of the prostate and calculate
PD0325901), and glycogen synthase kinase inhibitor (Stemgent, CHIR99021). the total volume. Bioluminescence imaging with IVIS Spectrum was performed
All components were prepared and filter sterilized before use. The culture dish as previously described34. Fluorescence imaging of dissected prostate tumours,
containing the isolated blastocysts was left undisturbed in a water-jacketed copper- the pair of draining lymph nodes, and lung was performed with a Leica M165FC
shelved incubator at 37 °C and 5% CO2. Six days after isolation and culture, fluorescence stereomicroscope. Lymph node metastasis score was assigned as 0 (no
mES cell outgrowths were visible and individually picked using an Oxford P20 GFP+ tumour cells), 1 (sparse GFP+ tumour cells on one lymph node), 2 (sparse
pipette with sterile 100μl round gel filtered tips (USA Scientific, 1022-0810). Each GFP+ tumour cells on both lymph nodes, or strong GFP+ tumour cell patches
outgrowth was transferred to a well of a sterile 96-well U-bottomed dish containing on only one lymph node), 3 (sparse GFP+ tumour cells on one lymph node and
15 μl sterile PBS. Forty-five microlitres of 0.25% trypsin-EDTA (Invitrogen, strong GFP+ tumour cell patches on the other lymph node), or 4 (strong GFP+
25200-056) was added to each well. After incubation for 4 min at 37 °C with tumour cell patches on the both lymph nodes). Spontaneous lung micrometastasis
5% CO2, each well was quenched with 300 μl of mES cell media containing was quantified by counting tumour cell clusters with more than ten GFP+ tumour
fetal bovine serum (FBS) and pipetted up and down six to eight times using a cells in each cluster.
multichannel pipette with sterile elongated tips, to generate cell clumps. Individual Immunohistochemistry, immunofluorescence, and western blot. Tissues were
cell clump populations were transferred to corresponding wells of a 48-well plate fixed in 10% formalin overnight and embedded in paraffin. Immunohistochemical
containing 14–16h pre-plated mouse embryonic fibroblasts in pre-equilibrated and immunofluorescence staining was performed as previously described11,14
.
mES cell culture medium. Once the wells reached ~80% confluency they were split Immunohistochemical slides were scanned with a Pannoramic Digital Slide
to triplicate plates for expansion in culture, freezing, and DNA isolation. Generated Scanner (3DHISTECH) and images were cropped from virtual slides in
mES cell lines were genotyped and clones with the genotype PB-Cre+ PtenL/L p53L/L Pannoramic Viewer. Immunofluorescence slides were imaged with a Nikon
Smad4L/L mTmGL/+ LSL-LUCL/+ (CPPSML) were selected. They were confirmed A1R Confocal Laser Microscope and quantified with ImageJ. Primary anti-
to be mycoplasma negative, and screened for the presence of the Y chromosome. bodies used included CK5 (Covance, PRB-160P), CK8 (Covance, MMS-162P),
Chromosome spreads were prepared and observed for abnormalities. Background Ki67 (Fisher, RM-9106-S1), cleaved caspase 3 (Cell Signaling Technology, 9661),
characterization was performed by single-nucleotide polymorphism analysis.
Gr-1 (BioLegend, 108401), phospho-S6 (Cell Signaling Technology, 4858). For
Chimaera cohort development. Derived mES cell lines JH58 and JH61 were western blot analysis, cells or fresh tissues were lysed on ice using RIPA buffer
genotyped as PB-Cre+ PtenL/L p53L/L Smad4L/L mTmGL/+ LSL-LUCL/+ (CPPSML) (Boston BioProducts) supplemented with protease and phosphatase inhibitors
and confirmed to contain the Y chromosome. Chimaera cohorts were produced (Roche). Western blot procedure was performed as previously described11,14
.
by blastocyst microinjection of the mES cells into C57BL/6NTac-Tyrtm1Arte Primary antibodies used included phospho-Met (Cell Signaling Technology, 3077),
(Taconic, 11971) then followed by uterine implantation into pseudo-pregnant phospho-VEGFR2 (Cell Signaling Technology, 3770), phospho-Erk1/2 (Cell
CD-1 (Charles River, 022) or Swiss Webster (Taconic, SW-F) female mice. Signaling Technology, 4370), phospho-Akt (Cell Signaling Technology, 4060),
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
phospho-mTOR (Cell Signaling Technology, 5536), phospho-p70 S6K (Cell using MACS technology were added to the upper chamber of a 24-well trans-well
Signaling Technology, 9234), phospho-S6 (Cell Signaling Technology, 4856), and system (BD Falcon). MDSC short-term culture medium (RPMI1640 supplemented
vinculin (Millipore, 05-386).
with 10% FBS) with the indicated concentrations of recombinant mouse IL-1α
Targeted pathway inhibitors. Enzalutamide (MedKoo Biosciences, 201821) was (BioLegend, 575002), IL-1β (BioLegend, 575102), IL-1ra (VWR, 10006-448), or
admixed with Purina 5053 Chow at 50mg of drug per kilogram of diet (Research mouse IL-1ra neutralizing antibody (R&D Systems, AF-480-NA) was added to
Diets), which, when fed to mice, could reach approximately 10mg per kg per day the bottom chamber. Cells were allowed to migrate to the bottom well for 6h at
dose35 (calculation based on average body mass of 25g and daily food intake of 37°C with 5% CO2. Migrated cells were fixed and stained with crystal violet for
5g per day, with online calculator at www.researchdiets.com/resource-center-page/ quantification.
diet-dose-calculator). To induce castration-resistant PCa, mice with prostate T-cell suppression assay and quantification of cytokine production. A T-cell
tumour confirmed by MRI were surgically castrated and started on an Enza- suppression assay was performed as we previously described11 using equal numbers
mixed diet for 3 weeks before being moved back to regular chow and treated of MACS-sorted intratumoural Gr-MDSCs and CFSE- (Invitrogen) labelled
with therapeutic drugs. For in vivo pharmacological inhibition, BEZ235 (Selleck MACS-sorted CD8+ or CD4+ T cells from spleen of wild-type C57BL/6 mice
Chemicals, S1009), cabozantinib (Selleck Chemicals, S1119), dasatinib (Selleck (The Jackson Laboratory, 000664). To assess the effect of drugs (BEZ, Cabo, Dasa)
Chemicals, S1021), PI-3065 (MedKoo, 407192), and GSK2636771 (MedKoo, on the proliferation of T cells, drugs at different concentrations were added to
205844) were orally administered at daily doses of 45mg per kg, 30mg per kg, anti-CD3/anti-CD28-stimulated CFSE-labelled and MACS-sorted CD8+ or CD4+
30mg per kg, 50mg per kg, and 30mg per kg, respectively, daily on a Monday– T cells from spleen of wild-type C57BL/6 mice at time zero. CFSE intensity was
Friday schedule. SX-682 (Syntrix Biosystems) was orally administered twice a day quantified 72h later with peaks identified by a BD LSRFortessa Cell Analyzer. CFSE
at 50mg per kg actual dose on a Monday–Friday schedule. Similar drug dosing peaks indicated the division times. Division times 0, 1–2, and ≥3 were defined
methods were described previously27,28,36–38. The doses we used for cabozantinib as no proliferation, moderate proliferation, and high proliferation, respectively.
and BEZ235 are clinically relevant: for cabozantinib, to convert the mouse dose Viable CD4+ or CD8+ T cells (viability defined as negative DAPI (4′,6-diamidino-
into human dose, we calculated 30mg per kg×(1/12.3)=2.4mg per kg daily in 2-phenylindole) stain) falling in each category were quantified as the percentage of
human (the conversion factor 12.3 can be found in FDA guidance at http://www. total live CD4+ or CD8+ T cells, respectively. The supernatant medium was used to
fda.gov/downloads/Drugs/.../Guidances/UCM078932 and in ref. 39). Given a typ- quantify IFNγ and IL-2 production by ELISA following the manufacturer’s manual
ical human weight of 60kg, the clinical dose range of 60mg daily (PCa, renal cell (BioLegend, 430804 and 431004).
carcinoma) to 140mg daily (medullary thyroid cancer) converts to a human dose Antibody array. Mouse prostate tumours treated with defined agents were
range of 1–2.3mg per kg daily. Therefore, our cabozantinib dose in mice approx- processed as instructed by the manufacturer’s protocol, and an equal amount
imates the approved dose for treating medullary thyroid cancer, and is 1.4 times (micrograms) of lysate was used to quantify phospho-RTK proteins with a
higher than used in COMET-1 and COMET-2 phase III trials of advanced PCa. For Proteome Profiler Mouse Phospho-RTK Array Kit (R&D Systems, ARY014) or
BEZ235, previous studies showed that while 300 or 400mg twice daily40 showed a Proteome Profiler Mouse XL Cytokine Array Kit (R&D Systems, ARY028).
significant toxicity in patients, 200mg twice daily displayed higher tolerability41
.
Medium conditioned by CPPSML PCa cell lines was analysed with a Proteome
The dose we used in the mice study, 45mg per kg daily, is approximately equivalent Profiler Mouse Cytokine Array Kit - Panel A (R&D Systems, ARY006).
to 225mg daily in humans, and thus is within range of tolerable clinical doses. Quantification of the spot intensity in the arrays was conducted with background
For ICB and Gr1-neutralizing antibody treatment, anti-PD1 (clone RMP1-14, subtraction in ImageJ.
BioXcell, BE0146), anti-CTLA4 (clone 9H10, BioXcell, BE0131), or anti-Gr1 (clone Transfection of recombinant proteins. MACS-sorted MDSCs isolated from
RB6-8C5, BioXcell, BE0075) antibodies (or their respective isotype IgG controls) mouse prostate tumours were co-transfected with recombinant active ERK2 (R&D
were intraperitoneally administered at 200μg per injection three times per week. Systems, 1230-KS-010) and p70S6K (R&D Systems, 896-KS-010) following the
The duration of drug treatment was typically 4 weeks before endpoint analysis.
manufacturer’s protocol of Chariot Protein Delivery Reagent (Active Motif, 30025).
CyTOF of human specimens and mouse tumours. CyTOF analysis of mouse The transfection efficiency was verified to be over 80% in separate wells of MDSCs
prostate tumours with data analysed in FlowJo (Tree Star) followed our previous transfected with β-galactosidase and stained with a Beta-Galactosidase Staining
methods11. Human prostate fine-needle aspirate specimens were acquired under Kit (Clontech, 631780). The effect of co-transfection on the resistance of MDSCs
approved institutional review board Protocol PA14-0420 at MD Anderson Cancer to drug treatment was evaluated 12h after transfection with Cell Proliferation
Center. Informed consent was obtained from all subjects. Prostate tumour single Reagent WST-1 (Sigma-Aldrich, 11644807001).
cells were isolated using a Human Tumour Dissociation kit (Miltenyl Biotec, 130- Treatment of MDSCs with PCa-cell conditioned medium. CPPSML PCa cell
095-929). All isolated cells were depleted of erythrocytes by hypotonic lysis. Cells lines established and grown in complete DMEM supplemented with 10% FBS were
were blocked with FcR Blocking Reagent (Miltenyl Biotec, 130-059-901) for 10min cultured to 90% confluence and renewed with fresh medium followed by medium
and incubated with CyTOF antibody mix for 30min at room temperature. Cells collection (0.45μm filtered) 12h later. The homozygous deletion of Pten, p53, and
were washed once and incubated with Cell-ID Cisplatin (Fluidigm, 201064) at Smad4 was confirmed with genotyping, and all cell lines were confirmed free of
2.5μM for 2.5min for viability staining. Cells were fixed with MaxPar Fix and mycoplasma with a MycoAlert Mycoplasma Detection Kit (Lonza). When the
Perm Buffer containing Cell-ID Intercalator-Ir (Fluidigm, 201192A) at 0.125μM cells were required to be pre-treated with Cabo or BEZ, the renewed medium was
and 4°C overnight to stain the nuclei. The samples were analysed with a CyTOF supplemented with the drugs at indicated concentration and added to cells. Twelve
instrument (Fluidigm) in the Flow Cytometry and Cellular Imaging Core Facility hours later, conditioned medium was collected with any floating cells removed
at MD Anderson Cancer Center. Data were analysed with FlowJo (Tree Star) and by 0.45μm filter, and subsequently concentrated 50-fold with an Amicon Ultra-
SPADE software42
.
15 Centrifugal Filter Unit with Ultracel-3 membrane (Millipore, UFC900324) to
Cell isolation and in vitro survival and migration assay. Over 95% of all deplete small compounds but retain proteins larger than 3kDa (including most
CD11b+Gr1+ myeloid cells in established naive or castration-resistant prostate cytokines). Such concentrated medium was diluted back to onefold with complete
tumours in CPPSML mice are granulocytic (Ly-6G+ Ly-6CLow), similar to the level DMEM. The conditioned medium was used within 12h of collection to culture
in the PB-Cre+ PtenL/L Smad4L/L model that we recently reported11. Gr-MDSCs MACS-sorted MDSCs isolated from mouse prostate tumours. Unconditioned
were isolated from prostate tumours by first enriching for lymphocytes using complete DMEM supplemented with 10% FBS was used as negative control.
Lympholyte-M Cell Separation Media (Cederlane, CL5031) followed by MACS- To test whether supplementing drug pre-treated conditioned medium with
based isolation using a Mouse MDSC Isolation Kit (Miltenyi Biotec, 130-094-538) specific cytokines could reverse the effect on MDSCs, the ten cytokines (all from
and plated in RPMI1640 supplemented with 10% FBS and antibiotics. From the BioLegend) were added individually at 100ng ml−1
.
same mice, CD8+ T cells were isolated from the spleen using a Mouse CD8a+ Quantitative RT–PCR. RNA was isolated by RNeasy Kit (Qiagen) and reversed tran-
T Cell Isolation Kit (Miltenyi Biotec, 130-104-075). MDSCs and CD8+ T cells were scribed using a Superscript III cDNA synthesis Kit (Life Technology). Quantitative
cultured as described20. GFP+ PCa cells were isolated from the prostate tumours by PCR was performed using a SYBR-GreenER Kit (Life Technology). The following
fluorescence-activated cell sorting with a BD FACSAria III Cell Sorter, and cultured primers were used: Arg1_F, CTCCAAGCCAAAGTCCTTAGAG; Arg1_R,
in complete DMEM supplemented with 10% FBS and antibiotics. The purity of AGGAGCTGTCATTAGGGACATC; Ncf1_F, ACACCTTCATTCGCCATATTGC;
Gr-MDSCs (CD11b+Gr1+Ly6G+Ly6ClowF4/80−), CD8+ T-cell (CD3+CD8+), Ncf1_R, TCGGTGAATTTTCTGTAGACCAC; Ncf4_F, GTGAACTCGGCCT
and PCa-cell (GFP+Tomato−CD45−) populations was greater than 90%, as deter- GGATCTG; Ncf4_R, AAGCTGCTCAAAGTCGCTCT; Cybb_F, CCTCTAC
mined by flow cytometry. A survival assay was performed on 96-well plates with CAAAACCATTCGGAG; Cybb_R, CTGTCCACGTACAATTCGTTCA;
Cell Proliferation Reagent WST-1 (Sigma-Aldrich, 11644807001) as described20
.
internal control Gapdh_F, AGGTCGGTGTGAACGGATTTG; Gapdh_R,
Serially diluted drugs (BEZ, Cabo, Dasa) were added to the culture to construct TGTAGACCATGTAGTTGAGGTCA.
dose–response curves and the IC50 value was calculated with GraphPad Prism Preparation of SX-682. SX-682 synthesis was performed at Syntrix Biosystems
6 software. For the migration assay, Gr-MDSCs isolated from prostate tumours and is described in the patent US 8969365 B2 (publication date 3 March 2015).
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
General chemicals, reagents, and precursors for synthesis were purchased from suspension was gravity filtered to remove the insoluble salts, and the resulting
Sigma-Aldrich (Milwaukee, Wisconsin, USA), Boron Molecular (Research Triangle clear yellow solution was added via pipette to a flask of H2O (2 l) and placed in
Park, North Carolina, USA), and Frontier Scientific (Logan, Utah, USA). Solvents the refrigerator. After cooling for about 1.5h, the resulting off-white precipitate
were purchased from either VWR International (West Chester, Pennsylvania, was collected by vacuum filtration, rinsing with water. The funnel was dried
USA) or Sigma-Aldrich (Milwaukee, Wisconsin, USA) and used without further overnight in a vacuum desiccator to afford 3.87g (80%) of 2-(2-boronic acid-5-
purification. The synthesis steps are described below as illustrated in Extended trifluoromethoxy-benzylsulfanyl)-pyrimidine-5-carboxylic acid (4-fluoro-phenyl)-
Data Fig. 7g.
amide (4) (SX-682). Melting point=211–214°C; 1H NMR (300MHz, DMSO-d6)
2-Chloro-pyrimidine-5-carboxylic acid (1) (3.16g, 20mmol, Frontier Scientific, δ 10.49 (s, 1H), 9.09 (s, 2H), 8.33 (bs, 2H), 7.78–7.73 (m, 2H), 7.66 (d, J=8.5Hz,
Logan, Utah, USA) was suspended in dichloromethane (40ml), and oxalyl chloride 1H), 7.46 (s, 1H), 7.25–7.19 (m, 3H), 4.70 (s, 2H); 13C NMR (100.6 MHz,
(3.30g, 26 mmol) was added, followed by DMF (three drops) as catalyst. The DMSO-d6) d 173.7, 161.9, 159.8, 157.4, 156.8, 148.8, 144.7, 136.0, 135.7, 135.0,
reaction started to vigorously evolve gas. The reaction was heated to reflux for 1h, 123.8, 122.3, 122.2, 121.7, 121.4, 118.8, 118.3, 115.5, 115.3, 34.3; 19F NMR
and then allowed to cool to room temperature. 4-Fluoroaniline (2.44g, 2.2mmol) (300 MHz, DMSO-d6) δ −56.5 (3F), −118.2 (1F); 11B NMR (128.4 MHz,
was added, vigorous bubbling was seen again, and the reaction mixture warmed DMSO-d6) δ 21.7; HRMS (m/z): [M]+ calculated for C19H15BF4N3O4S, 468.0807;
up considerably. Triethylamine (4.05g, 40mmol) was added, and a flocculent found, 468.0803; analysis (calculated, found for C19H14BF4N3O4S): C (48.84, 48.91),
precipitate immediately formed. The reaction mixture was heated to reflux once H (3.02, 3.20), N (8.99, 8.99), S (6.86, 6.73).
again for another hour, removed from heat, and stirred at room temperature for Statistical analysis. Data are presented as mean s.d. unless indicated otherwise.
18h under nitrogen. The reaction was diluted with EtOAc (100ml), and the organic Sample size was chosen to ensure 80% power to detect significant effect size on
layer washed with H2O, saturated NaHCO3, H2O, 1 N HCl, H2O, brine, then dried the basis of our recent publications using the transgenic PCa mouse models and
over Na2SO4. The liquid was filtered, and evaporated to yield 3.44g (68%) of the thereof derived primary tumour and myeloid cells11,13,14. A Student’s t-test assum-
2-chloro-pyrimidine-5-carboxylic acid (4-fluoro-phenyl)-amide as a light yellow ing two-tailed distributions or a non-parametric Mann–Whitney U-test was used
solid. Electrospray ionization–mass spectrometry (ESI–MS) (m/z): [M]+ =252.0. to calculate statistical significance between groups (no assumption was made that
This intermediate was carried forward without further purification.
variance was similar between the groups being statistically compared). P<0.05
In a round-bottomed flask, the intermediate 2-chloro-pyrimidine-5-carboxylic was considered statistically significant.
acid (4-fluoro-phenyl)-amide (2.52g, 10.0mmol) and anhydrous sodium hydrogen Data availability statement. Source Data for the main and Extended Data Figures
sulfide (1.22 g, 21.8 mmol) were suspended in anhydrous DMF (20 ml). The are provided in the online version of this paper. All other data are available from
suspension was stirred at room temperature, and the reaction mixture turned a the corresponding authors upon reasonable request.
deep green colour. After 1h, the reaction mixture was partitioned between EtOAc
and H2O, and transferred to a separatory funnel. After the layers were separated,
the organic layer was washed twice with a 2:1 mixture of H2O and 5% aqueous
NaHCO3. The combined aqueous layers were acidified with 1 N HCl precipitating
a yellow solid. The suspension was left to stand at room temperature for 2h, and
then the precipitate was collected by vacuum filtration, and rinsed with water. The
yellow solid was dried overnight in a vacuum desiccator to yield 2.3g (92%) of
the thiopyrimidinamide intermediate 2. 1H NMR (300MHz, DMSO-d6) δ 10.29
(s, 1H), 8.77 (bs, 2H), 7.77–7.70 (m, 2H), 7.24 (t, J=8.9Hz, 2H); ESI–MS (m/z):
[M]+ =250.0.
2-Mercapto-pyrimidine-5-carboxylic acid (4-fluoro-phenyl)-amide intermediate
2 (2.32g, 9.3mmol) and 2-bromomethyl-4-trifluoromethoxy-phenylboronic acid,
pinacol ester (3.85g, 10.1mmol, Boron Molecular, Raleigh, North Carolina, USA)
were suspended in anhydrous DMF (20ml). Sonication was used to dissolve the
compounds. To the reaction flask triethylamine (2.8ml, 20.1mmol, Sigma Aldrich,
Milwaukee, Wisconsin, USA) was added and a precipitate (triethylamine-HBr)
immediately formed. The reaction was layered with nitrogen gas and left to stand
at room temperature for 3.75h. The reaction was poured into H2O (500ml) and
layered with EtOAc. The biphasic solution was transferred to a separatory funnel
and diluted further with EtOAc and brine. The layers were separated, and the
aqueous layer was extracted twice more with EtOAc. The combined organic
layers were dried over Na2SO4, gravity filtered, and dried by rotary evaporation to
yield 5.7g (98%) of an oil, 2-[2-(4,4,5,5-tetramethyl-[1,3,2]dioxaborolan-2-yl)-5-
trifluoromethoxy-benzylsulfanyl]-pyrimidine-5-carboxylic acid (4-fluoro-phenyl)-
amide (3). 1H NMR (500 MHz, DMSO-d6) δ 10.52 (s, 1H), 9.11 (s, 2H), 7.81
(d, J=8.2Hz, 1H), 7.78–7.75 (m, 2H), 7.55 (s, 1H), 7.28–7.22 (m, 3H), 4.72 (s, 2H),
1.32 (s, 12H); ESI–MS (m/z): [M]+ =550.1. The NMR spectrum also contained
peaks consistent with the presence of residual DMF. The intermediate was carried
forward without further purification.
29. Muzumdar, M. D., Tasic, B., Miyamichi, K., Li, L. & Luo, L. A global double-
fluorescent Cre reporter mouse. Genesis 45, 593–605 (2007).
30. Birbach, A. Use of PB-Cre4 mice for mosaic gene deletion. PLoS ONE 8,
e53501 (2013).
31. Chen, Z. et al. Crucial role of p53-dependent cellular senescence in
suppression of Pten-deficient tumorigenesis. Nature 436, 725–730 (2005).
32. Benavides, F. et al. Microsatellite DNA variants between the inbred SENCAR
mouse strains. Mol. Carcinog. 28, 191–195 (2000).
33. Saha, B. K. Typing of murine major histocompatibility complex with a
microsatellite in the class II Eb gene. J. Immunol. Methods 194, 77–83 (1996).
34. Lu, X. et al. VCAM-1 promotes osteolytic expansion of indolent bone
micrometastasis of breast cancer by engaging α4β1-positive osteoclast
progenitors. Cancer Cell 20, 701–714 (2011).
35. Asangani, I. A. et al. Therapeutic targeting of BET bromodomain proteins in
castration-resistant prostate cancer. Nature 510, 278–282 (2014).
36. Maira, S.-M. et al. Identification and characterization of NVP-BEZ235, a new
orally available dual phosphatidylinositol 3-kinase/mammalian target of
rapamycin inhibitor with potent in vivo antitumor activity. Mol. Cancer Ther. 7,
1851–1863 (2008).
37. Nguyen, H. M. et al. Cabozantinib inhibits growth of androgen-sensitive and
castration-resistant prostate cancer and affects bone remodeling. PLoS ONE 8,
e78881 (2013).
38. Vitali, R. et al. Activity of tyrosine kinase inhibitor dasatinib in neuroblastoma
cells in vitro and in orthotopic mouse model. Int. J. Cancer 125, 2547–2555
(2009).
39. Nair, A. B. & Jacob, S. A simple practice guide for dose conversion between
animals and human. J. Basic Clin. Pharm. 7, 27–31 (2016).
40. Fazio, N. et al. A phase II study of BEZ235 in patients with everolimus-resistant,
advanced pancreatic neuroendocrine tumours. Anticancer Res. 36, 713–719
(2016).
41. Carlo, M. I. et al. A phase Ib study of BEZ235, a dual inhibitor of
phosphatidylinositol 3-kinase (PI3K) and mammalian target of rapamycin
(mTOR), in patients with advanced renal cell carcinoma. Oncologist 21,
787–788 (2016).
The pinacolyl boronate ester was deprotected via hydrolysis of the interme-
diate trifluoroborate43. Compound 3 (5.66g, 10.3mmol, 1 eq.) was dissolved in
methanol (100ml). The reaction vessel was charged with 4.5M aqueous potassium
hydrogen fluoride (11.5ml, 5eq.) and the resulting solution was stirred for 1h.
The methanol was removed by rotary evaporation at room temperature and the
resulting mixture of yellow and off-white solids was suspended in acetone. The
42. Bjornson, Z. B., Nolan, G. P. & Fantl, W. J. Single-cell mass cytometry for
analysis of immune system functional states. Curr. Opin. Immunol. 25,
484–494 (2013).
43. Yuen, A. K. L. & Hutton, C. A. Deprotection of pinacolyl boronate esters via
hydrolysis of intermediate potassium trifluoroborates. Tetrahedr. Lett. 46,
7899–7903 (2005).
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
Extended Data Figure 1 | Chimaeric modelling as an efficient approach
to generating spontaneous metastatic PCa. a, Comparison of probability
of obtaining PCa-bearing males with CPPSML genotype in a litter,
through breeding or chimaeric modelling. In chimaeric modelling,
>75% coat colour contributed by injected mES cells (mESC in the figure)
is defined as positivity for chimaera. b, c, Predicted and experimental
results for PCR–RFLP genotyping of the H2 locus from several mouse
strains and two mES cell lines (JM8 is a standard mES cell line derived
from C57BL/6 strain as control; Ep61 is also known as JH61). Red asterisk
in b and yellow highlight wells in c indicate that the H2 haplotyte for JH61
is H2b, the same as the C57BL/6 strain. d, SSLP marker analysis of the
region on chromosome 17 flanking the H2 complex locus (34–46 Mb),
showing that JH61 has 100% C57BL/6 background in the H2 locus,
identical to the standard C57BL/6 mES cell line JM8. e, Experimental steps
for generating the CPPSML chimaeras. f, Fluorescence images of prostate,
draining lymph node (LN), and lung from a representative chimaera at
3 months old. GFP+ signals indicate the presence of metastasis to lymph
nodes and disseminated tumour cells and micrometastasis in lung.
Scale bars: prostate, 5 mm; lymph node and lung, 1 mm. g, Fluorescence
microscopy and H&E image of snap-frozen prostate tumour from
chimaera showing that the GFP+ area corresponds to adenocarcinoma
and the GFP− area corresponds to normal host cells. Scale bar, 500 μm.
h, Immunohistochemical staining showing the expansion of both CK8+
luminal lineage and CD5+ basal lineage in the prostate tumour formed in
CPPSML chimaera. Scale bar, 50 μm.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
Extended Data Figure 2 | Experimental design for preclinical therapy
of mCRPC in CPPSML chimaeras. a, Significant yet transient survival
benefit by castration followed by diet admixed with enzalutamide
(50 mg per kg diet) in PB-Cre+ PtenL/L p53L/L Smad4L/L mice (n =40 and
18, respectively). ****P <0.0001, log-rank test. b, Record of assignment
for drug trials showing the time range of prostate tumour formation
in the CPPSML chimaera. c, Representative CPPSML chimaera with
primary CRPC, lymph node metastasis, and micrometastasis in lung.
Scale bars: prostate, 5 mm; lymph node and lung, 1 mm. AP, DLP, and
VP denote anterior, dorsolateral, and ventral prostate lobes, respectively.
d, Experimental flow for creating mCRPC cohorts and preclinical
testing of monotherapy and combination therapy, followed by tumour
characterization. e, Representative images of prostate tumours with H&E
staining, and GFP+ tumour cells in the lymph node and lung. Scale bars:
prostate, 5 mm; lymph node and lung, 1 mm; H&E, 200 μm.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
Extended Data Figure 3 | Significant combination efficacy by
cabozantinib and ICB observed in chimaeras generated from JH58.
a, Experimental design for JH58 chimaeras, similar to the JH61 chimaera
experiments. b, Longitudinal MRI images from representative chimaeras
in control or combination cohorts. Red contour denotes area of prostate
tumour. c, Strong anti-tumour effect by combination therapy in JH58
chimaeras shown by prostate tumour mass, lymph node metastasis
scores and lung micrometastasis number (n =3, biological replicates).
**P <0.01, Student’s t-test. d, e, Quantification of tumour cell
proliferation by Ki67 immunohistochemistry (n =4, biological replicates)
with representative images. Anterior prostate (AP) and dorsolateral
prostate (DLP) were quantified separately. f, g, Quantification of tumour
cell apoptosis by cleaved caspase 3 immunohistochemistry (n =5,
biological replicates) with representative images. Scale bar, 100 μm.
In c, d, and f, data represent mean s.d. *P <0.05, **P <0.01,
***P <0.001, compared with control using Student’s t-test.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
Extended Data Figure 4 | Combination efficacy by Gr1 neutralizing
antibody with ICB. a, Dasatinib, but not cabozantinib or BEZ235,
significantly reduced the frequency of infiltrating T cells in CRPC of
CPPSML mice (n =4, biological replicates). Data represent mean s.e.m.
*P <0.05, Mann–Whitney U-test. b, Frequency of Gr-MDSCs and Mo-
MDSCs in CPPSML prostate tumours (n =13, biological replicates).
c, Mass and representative whole-organ and H&E images of prostate
tumours from CPPSML chimaeras induced to develop CRPC and treated
with 1 month of control IgG, ICB (anti-CTLA4 plus anti-PD1 antibodies),
anti-Gr1 neutralizing antibody, or combination of ICB and anti-Gr1
(n =4, biological replicates). Scale bars: 3 mm for organ images, 200 μm for
H&E images. In b and c, data represent mean s.d. *P <0.05, **P <0.01,
****P <0.0001, #P >0.05, Mann–Whitney U-test.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
Extended Data Figure 5 | Characterization of the effect of drugs on
MDSCs. a, Comparison of in vitro sensitivity to Dasa by MDSCs, CD8+
T cells, and GFP+ cancer cells isolated from CRPC of CPPSML mice.
Cell viability was measured 24 h after the start of drug treatment using
the WST-1 assay. IC50 values are indicated. b, Comparison of in vitro
sensitivity to BEZ, Cabo, and Dasa by MDSCs isolated from CRPC of
CPPSML mice. The assay was performed in RPMI1640 supplemented
with 10% FBS and 10 ng ml−1 GM-CSF (n =2, biological replicates).
c, Comparison of in vitro sensitivity to BEZ, Cabo, and Dasa by MDSCs
isolated from CRPC of CPPSML mice. The assay was performed in
RPMI1640 supplemented with 10% FBS, 10 ng ml−1 GM-CSF, and pre-
conditioned for 12 h by PCa cell lines established from the CPPSML model
(n =2, biological replicates). d, Representative CFSE flow cytometry
histograms showing the effect on in vitro T-cell proliferation by MDSCs
isolated from CRPC of CPPSML mice treated with the indicated drugs.
Position of CFSE peaks can be used to denote the T-cell division times.
e, Representative CFSE flow cytometry histograms showing the effect of
Cabo, BEZ, and Dasa on in vitro T-cell proliferation.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
Extended Data Figure 6 | Cabozantinib and BEZ235 inhibit PI3K
signalling in prostate tumour and intratumoural MDSCs. a, Mouse
phospho-RTK array measuring phospho-RTK activity in prostate tumours
with indicated treatments. Numerals 1–5 represent pEGFR, pErbB2,
pErbB3, pAxl, and pPDGFRα, respectively (n =2, biological replicates).
b, c, Reduced pS6 signal in intratumoural MDSCs by Cabo and BEZ
treatment, revealed by immunofluorescent co-staining of pS6 and Gr-1
(n =3, biological replicates). Scale bar, 100 μm. d, WST-1 assay showing
that co-transfection of active ERK2 and p70S6K proteins mediated the
resistance of MDSCs isolated from CPPSML tumours to the cytotoxicity
by Cabo (1.5 μM) or BEZ (0.15 μM). The assay was performed in
RPMI1640 supplemented with 10% FBS and 10 ng ml−1 GM-CSF
(n =3, biological replicates). e, WST-1 assay similar to d, but performed
in RPMI1640 supplemented with 10% FBS, 10 ng ml−1 GM-CSF, and
pre-conditioned for 12 h by PCa cell lines established from the CPPSML
model (n =3, biological replicates). In c–e, data represent mean s.d.
***P <0.001, Student’s t-test.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
Extended Data Figure 7 | Cabozantinib or BEZ235 suppress secretion
by PCa cells of several cytokines that promote MDSC activity.
a, Quantification of intratumoural cytokine levels in CRPC chimaera
tumours with indicated treatment using cytokine array (n =2, biological
replicates). Numerals 1–9 represent CCL5, CCL12, CCL21, CD40,
CD142, HGF, IGFBP-6, IL-1ra, and VEGF, respectively. b, Quantification
of intratumoural cytokine levels in Dasa +ICB combination-treated
CPPSML chimaera CRPC with mouse cytokine assay, with image and
relative intensity of the numbered cytokines shown (n =2, biological
replicates). c, Experimental design for MDSC culture in the presence
of PCa conditioned medium. d, Cytokine array results for conditioned
medium from CPPSML PCa cell lines treated with vehicle, Cabo
(1 μM), or BEZ (1 μM) for 12 h (n =2, biological replicates). Boxed
cytokine is CCL5. e, Effect of supplementation of individual cytokines to
the conditioned medium from PCa cell lines treated with Cabo (1 μM) or
BEZ (1 μM) on Arg1, Cybb, Ncf1, and Ncf4 from cultured MDSCs (n =3,
biological replicates). f, g, Chemical structure and synthesis of allosteric
CXCR1/2 antagonist SX-682 (Syntrix Biosystems). For details, please refer
to the corresponding section in Methods. In b, e, and f, data represent
mean s.d.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
Extended Data Figure 8 | Detailed cell population annotation in SPADE tree. a, SPADE tree coloured by the median intensity of individual markers
(indicated above colour bar) to facilitate the assignment of tree branches to individual cell populations (shown on the top of each plot) (n =12, biological
replicates). b, Surface markers of different immune subpopulations representing small branches of the SPADE tree.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

Letter reSeArCH
Extended Data Figure 9 | Model depicting the combination therapy strategy in treating mCRPC. As demonstrated in the CPPSML chimaera model,
targeted therapy with agents that inhibit MDSC infiltration frequency and immunosuppressive activity can synergize with ICB to invigorate T-cell
immunity in the prostate tumour microenvironment and thus impair CRPC progression.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

reSeArCH Letter
extended data table 1 | Additional information on mouse model, clinical samples, and methodology
a, Coat colour chimaerism for chimaeras derived from JH61. b, CyTOF antibody panel for human prostate tumour samples. c, Clinical information of the fresh fine-needle biopsy specimens.
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.