Supplementary Materials for
Targeting autocrine CCL5-CCR5 axis reprograms immunosuppressive myeloid cells
and reinvigorates antitumor immunity
Method and Materials
Measurement of CCL5 in TME:
CCL5 in TME was measured by ELISA. 1mg tumor (2 week) was lysed in 1mL ELISA lysate
buffer [10mM Tris.HCL (pH.7.5), 100mM NaCl, 0.5mM EDTA, 0.5% NP40, 1mM PMSF, 1µg/mL
Leupeptin and 1µg/mL DNAase I] and briefly sonicated. Supernatants were collected and
subject to ELISA.
Primer information:
β-actin (Forward: 5’-TGAGCTGCGTTTTACACCCT-3’, Reverse: 5’TTTGGGGGATGTTTGCTCCA-3’) Rb1 (Forward:5’-ACAACCCAGCAGTGCGTTA-3’, Reverse:
5’-TGGGGATTGCAGCTGTTTTG-3’); NOS2 (Forward: 5’-AGGAGGAGAGAGATCCGATTTAG3’, Rerverse: 5’-TCCAAGGTGCTTGCCTTATAC-3’); CCL5 (Forward: 5’AGGACTCTGAGACAGCACAT-3’, Reverse:5’-AGCAATGACAGGGAAGCTATAC-3’); CCR5
(Forward: 5’-TGGTTTATGACCTGGTTGACTT-3’, Reverse: 5’TCTTGGGCCCTCTATGATACT-3’).
In vitro MDSC generation system:
To culture MDSCs, fresh erythrocyte-depleted BM aspirates were plated in a petri dish in
complete media supplemented with GM-CSF (20ng/mL) and IL-6 (20ng/mL), or in complete
media supplemented with 30% 4T1supernatant, or in combination of GM-CSF, IL-6 and 4T1
supernatant. To collect PGE2-depleted 4T1 supernatant, 4T1 cells were initially treated with
Celecoxib (10µM) for 6 hours, followed by thorough washes with PBS and a continued culture in
complete media containing Celecoxib (10µM) for another 24 hours.
Description of MDSC subset-switching:
The role of PGE2 in MDSC subset-switching-- The basic MDSC media (GM-CSF+IL-6supplemented complete media) was only able to stimulate a small percentage of Ly6C+ MMDSCs from ccl5-/- BM progenitors, in contrast to a full range of MDSCs with WT BM
progenitors (Supplementary Fig. S4A). ELISA analysis on secreted cytokines showed 2.02 ±
1.06 ng/mL of prostaglandin E2 (PGE2) in the supernatant collected from WT MDSC culture,
but no detectable PGE2 in ccl5-/- MDSC supernatant. The addition of exogenous PGE2 or 4T1
supernatant which contains a certain amount of PGE2 to the culture of ccl5-/- BM progenitors
strongly promoted the generation of the “atypical” Ly6Chigh/Ly6Ghigh cells. Depletion of PGE2
alone from 4T1 supernatant via pretreatment with Celecoxib, a specific COX-2 inhibitor, caused
a dramatic accumulation of Ly6C+ M-MDSCs at the expense of the Ly6Chigh/Ly6Ghigh cells
(Supplementary Fig. S4B), suggesting that PGE2’s main role is to promote the transition of
Ly6C+ cells to the Ly6Chigh/Ly6Ghigh cells. Further investigations on cell morphologies as well as
the function of each population are needed to confirm the role of PGE2 in MDSC subsetswitching. Interestingly, LPS was exchangeable with PGE2 in terms of promoting the transition
from M-MDSCs to Ly6Ghigh/Ly6Chigh MDSCs (data not shown), suggesting that the role of PGE2
is redundant.
The role of CCL5 in MDSC subset-switching-- none of the proinflammatory agents we tested,
including LPS, PGE2, IL-6, CCL5 and etc., could rescue the suspended transition from
Ly6Ghigh/Ly6Chigh to G-MDSCs, indicating a non-redundant role of autocrine CCL5-CCR5
signaling axis in complete G-MDSC development.
Supplementary Figures
Supplementary Figure S1. Comparable CCL5 levels in serum and in tumor between tumorbearing WT and ccl5-/- hosts were observed. Serum and 4T1 tumors were collected at 4 weeks
after inoculation. Data are representative of 4 tumors/ group.
Supplementary Figure S2. No up-regulation of CCL5 expression was observed in myeloid
CD34+ and CD90+ cells in response to tumor growth. 4T1 tumor cells and PyMT tumor
suspension were transplanted into mammary pads of Balb/c and C57/B6 mice respectively. 2-4
weeks after tumor inoculation, CD34+, CD90+ and CD11b+/Gr-1+ BM cells were screened for
intracellular expression of CCL5. Representative data of 2 independent experiments showed no
CCL5 expression in CD34+ and CD90+ myeloid cells in response to tumor growth. BM cells were
pooled from 3 mice / group.
Supplementary Figure S3. CCL5 upregulation in MDSCs responding to PyMT tumor growth.
PyMT breast tumor cells were transplanted into mammary pads of C57/B6 WT mice. 2-4 weeks
after tumor inoculation, the upregulation of CCL5 expression in BM MDSCs (CD11b+/Gr-1+) was
observed. Data are representative of 2 independent experiments with pooled BM cells from 3
mice per group.
Supplementary Figure S4. Transition from Ly6C+ M-MDSCs to Ly6Chigh/Ly6Ghigh MDSCs
mediated by PGE2. (A) Flow cytometric analysis on Ly6C/Ly6G expression of GM-CSF / IL-6induced MDSCs derived from Balb/c WT or ccl5-/- BM progenitors. (B) Flow cytometric analysis
on Ly6C/Ly6G expression of MDSCs derived from WT or ccl5-/- BM progenitors under various
culture conditions as indicated. Data are representative of more than 3 independent
experiments. BM progenitors were cultured for 6 days.
A
B
Supplementary Figure S5. NOS2 mRNA level in WT vs ccl5
(Gr-1+/CD11b+) from Balb/c WT or ccl5
-/-
-/-
MDSCs. Sorted BM MDSCs
with or without 4T1 tumor were subject to quantitative
PCR for relative expression of NOS2 (mRNA expression of NOS2 in WT naïve MDSCs was set
to 1). End products of qPCR were further subject to agarose gel electrophoresis. Data are
representative of 2 independent experiments with pooled BM cells from 3 mice per group (N:
naïve, Tu:tumor).
Supplementary Figure S6. Efficacy of nanoparticle-delivered CCL5-silencing in BM.
(A, left) PyMT tumor burden (weight: g) of control nanoparticle-treated and CCL5-targeting
nanoparticle-treated mice (C57/B6). Data are representative of 2 independent experiments (n=4
to 5 per group), mean ± SEM.*P˂0.05. (B, right) Sorted BM MDSCs (Gr-1+/CD11b+) from PBS
(mock), control nanoparticle (control) and CCL5-targeting nanoparticle (CCL5-targeting) treated
mice were subject to quantitative PCR of relative expression of CCL5, CCR5 and NOS2. mRNA
expressions in sorted MDSCs from mock treatment group was set to 1. mean ± SEM.**P˂0.01.
Supplementary Figure S7. Model of CCL5/CCR5 axis-regulated MDSC differentiation and
function.
Supplementary Figure S8. Different gene expression patterns between tumor-infiltrating Mand G-MDSCs. The expression profiling of indicated genes of M-MDSCs (CD11b+/Ly6C+) and
G-MDSCs (CD11b+/Ly6G+) sorted from 4T1 tumors carried by WT or ccl5-/-mice. Results
represent pooled data from 4 tumors per group.