Nature | Vol 579 | 12 March 2020 | 285
study found that combined epigenetic modifiers blunt metastases in a
mouse model of aggressive NSCLC with an immune-competent micro-
environment^14. Here we reveal a key mechanism of action of adjuvant
epigenetic therapy (AET) on premetastatic niches that could underlie
these previous findings. Low-dose AET modulates innate immune fac-
tors in the lung microenvironment to inhibit tumour recurrence after
resection in three syngeneic models of aggressive pulmonary metas-
tasis: Lewis lung carcinoma (LLC), HNM007 oesophageal squamous
cell carcinoma (mice of the LLC and HNM007 models have no extrapul-
monary metastases) and 4T1 mammary cancer (Fig. 1a, Extended
Data Fig. 2a–e).
It is known that the microenvironments of distant organs—the
targets of future metastases—are not passive receivers of
circulating tumour cells, but instead are selectively modified by the
primary tumour before metastasis^6 ,^7 ,^15 ,^16. Settlement of tumour cells at
distant sites is dependent on tumour-secreted factors and tumour-
shed extracellular vesicles that enable the premetastatic microenviron-
ment to support their colonization^6 ,^7 ,^15 ,^16. To study whether low-dose
AET affects the dynamics of the premetastatic microenvironment, we
used mouse models of highly aggressive, pulmonary metastasis that
exhibit pulmonary metastases in 90–100% of mice after resection,
with median disease-free survival times of fewer than 14 days (Fig. 1a,
b, Extended Data Fig. 2a–d).
In mice of the LLC model (hereafter LLC mice), no tumour-cell infiltra-
tion of the lung was detectable by haematoxylin and eosin (H&E) and
no green fluorescent protein (GFP)+ tumour cells were observed by
b
Daysafterresection
Mean
area
of
GF
+P
nodul
es
(m
(^2) m
)
Day 0 Day 3 Day 6Day 9 Day 12 Day 15
c
Pe
rcentag
eo
f
CD45
+cells
Pe
rcentag
eo
f
CD45
+cells
Pe
rcentag
eo
f
CD45
+cells
Pe
rcentag
eo
f
CD45
+cells
Pe
rcentag
eo
f
CD45
+cells
Pe
rcentag
eo
f
CD
- 3
cells
Pe
rcentag
eo
f
CD - 3
cells
Pe
rcentag
eo
f
CD - 4
cells
**
**
CD11
bC
D11b
LLC
HNM00
7
Mock LD-AET Mock
LD-AET
Daysafterresection
20
40
60
(^036)
Mock
LD-AET
Daysafterresection
Pe
rcentag
eo
f
MDSCs in CD45
+cells
Pe
rcentage
of
MDSCs in CD45
+cells
f
Num
ber
of
GF
+P
nodul
es
d
P = 0.01 6
0
20
40
60
Mock Pep-H (^6) Alveolar
macrophages
Interstitial
macrophages
MoPep-Hck 6
0
2
4
6
8
10
NS NS
Su
rvival (%
)
Lung-m
etas
tasis-fr
ee
survival
(%
)
Lung-m
etas
tasis-fr
ee
survival
(%
)
Lung-m
etas
tasis-fr
ee
survival
(%
)
0510 15 010203040
- Mock(n= 9)
Daysafterresection
0
50
100
0
50
100
Daysafterresection
Pep-irrel (n= 7)
Pep-H6(n= 8)
Pep-irrel (n= 7)
Pep-H6(n= 8)
Mock(n= 9)
GR 1
CD11
b
Mock Pep-H6 MDSCs
MDSCs
6.26
MDSCs
28.4
Macrophages
a
e
0
50
100
0 5110 5
0
50
100
0 264 8
M-transMDSCfer(n= 7)
PMN-MDSC
transfer (n= 7)
Mock(n= 7)
M-transfMDSCer (n= 7)
PMtransN-MDSCfer(n= 7)
Mock(n= 7)
Daysafterresection
Daysafterresection
0 Survival (%) Survival (%)
50
100
04268 0
50
100
MDSCtransfer(n= 10)
Mock(n= 10)
MDSCtransfer(n= 10)
Mock(n= 10)
05110 52025
Daysafterresection Daysafterresection
20
Tumour
implantation
Tumour resection and
treatmentstart
Vehicleor
adjuvanttreatment
Treatment
stop
Micediedand
survival timerecorded
CTscanevery 2–3daysCTscanevery2–3days
Day –9 Day 0Day 14
20
40
60
0
80
093 6 12
0
2
4
4080
120
03691215
Daysafterresection
(^003691215)
20
40
60
CD11b+cells CD11b+GR 1 +cells CD 3 +cells CD 8 +cells CD 4 +cellsTreg cells
0
20
NCD0D2D4D7D9D1^2 NCD0D2D4D7D9D12 NCD0D2D4D7D9D1^2 NCD0D2D4D7D9D12 NCD0D2D4D7D9D1^2 NCD0D2D4D7D9D1^2
40
60
80
0
20
40
60
80
0
20
40
60
0
20
40
60
0
20
40
60
0
20
40
60
105
104
103
102
101
100
100101102103104105
105
104
103
102
101
100
100101102103104105
105
104
103
102
101
(^100100101102103104105)
105
104
103
102
101
(^100100101102103104105)
105
104
103
(^101021)
101102103104105
105
104
103
(^101021)
101102103104105
MDSCs30.7 MDSCs4.66
MDSCs36.1 MDSCs8.25
GR1
GR1
Fig. 1 | Low-dose AET disrupts the lung premetastatic microenvironment by
affecting MDSCs. a, Timeline of mouse models of metastasis with treatment
schedules. CT, computed tomography. b, Longitudinal H&E and
immunof luorescence staining of lung tissue showed the presence of tumour
cells from day 6 in the LLC model. Immunof luorescence staining was
performed using GFP (green) antibodies. Merged images contain
4′,6-diamidino-2-phenylindole (DAPI) DNA staining, which demarcates cell
nuclei (blue). Scale bars, 2 mm. Graph shows the area and numbers of
metastatic nodules (n = 3 mice at each time point). Two-sample, two-sided t-
test. c, Immune-cell profiles of lungs in the LLC model. Single-cell suspensions
from both lungs were analysed by f luorescence-activated cell sorting (FACS)
(n = 3 mice at each time point). NC, negative control (normal lungs from
C57BL/6 mice). D, day. Two-sample, two-sided t-test was used in comparison
with the negative control. Treg cells, regulatory T cells. d, FACS showing lung
MDSCs in the LLC model at day 3 were depleted using pep-H6 (leftmost two
panels). Column diagram showing the effect of pep-H6 on the percentages of
lung MDSCs and macrophages (middle two panels) at day 3 (n = 3 mice in each
group). Two-sample, two-sided t-test. Kaplan–Meier curves showing disease-
free survival and overall survival of LLC mice after MDSC depletion (rightmost
two panels). Two-sided log-rank test. Pep-irrel, irrelevant peptibody, used as
control. e, Kaplan–Meier curves showing disease-free and overall survival of
LLC mice after transfusion of lung MDSCs (5 × 10^6 ) (leftmost two panels) and of
monocytic (M-)MDSCs (5 × 10^6 ) or polymorphonuclear (PMN-)MDSCs (5 × 10^6 )
from bone marrow (rightmost two panels). All transfusions were conducted on
day 1 and day 4. Two-sided log-rank test. f, FACS showing representative effects
of low-dose (LD-)AET on lung MDSCs in LLC and HNM007 mice at day 3. Column
diagrams showing the effect of low-dose AET on lung MDSCs in LLC and
HNM007 mice (n = 3 mice at each time point). Two-sample, two-sided t-test.
Bars show mean ± s.e.m. NS, not significant, P < 0.05, P < 0.01, P < 0.001.