Available via license: CC BY 4.0
Content may be subject to copyright.
Supplementary figure 1. Tumour cell and fibroblast phenotypes in vitro. (A) Representative
phase contrast images of tumour cells B16.F10, B16.OVA, Lung tumour and Lung
tumour.OVA and fibroblasts CAF and NORM in in vitro culture. Scale bar, 140m. (B) Flow
cytometry characterization of fibroblasts used (NORM, lung CAF, skin CAF and FRC
fibroblasts) to exclude tumour cell and immune cells. Representative profiles for typical CAF
markers Podoplanin, PDGFR and PDGFRepithelial marker EPCAM; and immune
markers CD45and MHC II. (C) Fibroblast collagen gel contractility assay displayed as gel
diameter as a percentage of its starting diameter.
Supplementary figure 2. Ovalbumin processing by fibroblasts and tumour cells in vitro. (A)
Time-lapse phase images showing debris (red arrow) breaking away from a donor cell
(green arrow, upper panel 0-4hrs) and its subsequent uptake by an independent acceptor
cell (blue arrow, 21-24hrs). Scale bar. 60m. (B) Representative FACS dot plots showing
ovalbumin (DQ-OVA) proteolytic degradation over time for FRC, NORM, and CAF fibroblasts
and B16.F10 and lung tumour cells.
1 hr 2 hr 4 hr
21 hr 22 hr 23 hr
24 hr 60mm
0 hr
FS Lin
BODIPY DQ-OVA (MFI)
FRC
CAF
NORM
B16.F10
Lung Tumour
0 min 15 min 45 min 90 min 120 min
0 min 15 min 45 min 90 min 120 min
0 min 15 min 45 min 90 min 120 min
0 min 15 min 45 min 90 min 120 min
15 min 45 min 90 min 120 min
10
m
m
A
B
Supplementary figure 3. CAFs exhibit slower processing kinetics than normal fibroblasts.
(A) Representative micrographs of endosomal compartments (red) showing early
endosomes (Rab5), and late endosomes (Rab7) at 15 and 120 minutes post ovalbumin
pulse (green). Areas of co-localization (white arrows) and ovalbumin not within labelled
compartments (blue arrows) are indicated in each image. Scale bar: 10μm. (B) Flow
cytometry plots showing MHC I SIINFEKL staining in FRCs, CAFs and normal fibroblasts 18
hours after pulsing with ovalbumin (unstained control; stained but no OVA pulse control;
OVA pulsed). (C) Representative micrograph of LysoTracker (red) labelled CAFs containing
engulfed B16.OVA tumour debris (green). Nuclei counterstained with DAPI (blue). Scale bar,
5μm. Assays performed in triplicate from 2 experiments (B).
Supplementary figure 4. CAFs reduce CD8+ T cell viability in an antigen-specific, death
ligand dependent manner. (A) T cell capacity to kill tumour cells following fibroblast co-
culture (normal fibroblasts, yellow bars; CAFs, red bars) and subsequent transfer to parental
(B16.F10) or antigen bearing (B16.OVA) cells. (B) T cell capacity to kill tumour cells when
control parental (B16.F10) or antigen bearing (B16.OVA) tumour cells were used as source
of antigen rather than addition of OVA. Tumour cells were cultured with fibroblasts prior to
addition of activated OT-I T cells. Flow cytometry quantification of tumour cell viability in
cultured alone or in presence of normal fibroblasts (yellow bars), CAFs (red bars). (C) T cell
capacity to kill tumour cells when control parental (LC.WT) or antigen bearing (LC.OVA)
tumour cells were used as source of antigen rather than addition of OVA. (D) Flow cytometry
quantification of Lag-3 expression by T cells cultured in presence or absence of fibroblasts,
antigen and target cells as indicated. Data shown as mean ± SEM. * P < 0.05, ** P < 0.01,
*** P <0.001 (one-way ANOVA with Tukey’s or Dunnett’s post hoc analysis). ns, not
significant. Comparisons indicated by horizontal lines. Assays performed in duplicate from
three independent experiments.
Supplementary figure 5. Antigen-specific deletion of T cells occurs in vivo. (A-B) Gating
strategy for recovered T cells after adoptive transfer to tumour bearing mice. Representative
plots identifying differentially labelled OT1 T cells (CFSE green) and WT T cells (CFSE APC)
from lymph nodes (A) and f spleen (B) of tumour bearing mice. Main panels depict live-
labelled T cells with minor panels illustrating gating ancestry. (C) Adoptive transfer
experimental design. (D) Intratumoural T cell counts/million events at 18hrs and (E) 72hrs
post transfer. (F) Frequency of dead labelled T cells per tumour at 18hrs and (G) 72hrs post
transfer. (H-I) FAS and PD-1 expression by live T cells within tumours at (H) 18hrs and (I)
72hrs post transfer. (J-K) T cells adoptively transferred show equal viability in the lymph and
spleen of recipient mice. CD8+ T cell viability of cells recovered from adoptive transfer of
either OT1 or WT T cells in to recipient mice bearing ovalbumin (OVA) or WT B16 tumours
from the (J) lymph nodes or (K) spleen. Data shown as mean ± SEM. * P < 0.05, ** P <
0.01, *** P < 0.001 (one- way ANOVA with Tukey’s post hoc analysis). ns, not significant.
Comparisons indicated by horizontal lines. Assays performed in 10 mice (D-G), 5 mice (H, I)
or 4 mice (J, K) from two independent experiments. Each point is a individual tumour (D-I),
lymph node (J) or spleen (K).
Supplementary figure 6. Blocking antibodies against PD-L2 and FASL have no overall
effect on total immune cell number or stromal cell populations in tumours and lymph nodes.
(A) Total CD45+ immune cell frequency in tumours and (B) lymph nodes from mice receiving
control antibody or blocking antibodies. (C) Tetramer positive CD8 T cells in lymph nodes
from mice receiving control antibody or blocking antibodies. (D-F) Stromal cell populations
within tumours from animals receiving either control or blocking antibodies as a percentage
of CD45- cells stained for (D) Podoplanin, (E) CD31, (F) both. (G-I) Stromal cell populations
within lymph nodes from animals receiving either control or blocking antibodies as a
percentage of CD45- cells stained for (G) Podoplanin (FRC), (H) CD31 (blood vessels) (I)
both (LECs). (J) (K) Data shown as mean ± SEM. Significance determined by two-tailed
unpaired Student’s t-test. Symbols show individual animals.
Supplementary figure 7. Impact of deletion of host FASL on tumour volume and antigen
specific T cells. (A) Representative confocal image confirming presence of injected (GFP-,
yellow arrowheads) Thy1+Podoplanin+ CAFs within day 9 tumours at the periphery (left) and
centre (right) of the lesion. Host-derived cells were detected as GFP+ (white arrowheads)
Scale bar: 20μm. (B) Flow cytometry quantification of CFSE-Labelled injected CAFs
expressed as relative percentage within the non-immune stromal compartment (CD45-GFP-
CFSE+PDGFR+Thy1+ cells). ** P < 0.01 (one-way ANOVA with Tukey’s post hoc analysis).
n = 3 individual tumours. (C) Growth curves of tumours in FASL-deficient GLD mice.
B16.OVA tumour cells were implanted alone (no FASL present in host) or at a 1:1 ratio with
CAFs (FASL on CAFs only), and compared with tumours implanted in WT C57 (D) Volumes
of day 9 tumours. (E) Antigen specific T cells in tumours expressed as fold change of
Tetramer positive CD8 T cells within the CD45+ compartment. Data shown as mean ± SEM.
* P < 0.05 (one-way ANOVA with Tukey’s post hoc analysis). n = 6 tumours from 2
independent experiments.
0 2 4 6 8 10
0
100
200
300
Days
GLD:
B16
GLD:
B16+CAF
WT:
B16
GLD:
B16
GLD:
B16+CAF
WT:
B16
0
100
200
300
400
500
volume mm^3
Volume mm3
GLD:
B16
GLD:
B16+CAF
WT:
B16
0.0
0.5
1.0
1.5
2.0
Fold change in tet+ CD8
C D
* *
B16 B16:CAF CAF
0
1
2
3
4
Relative %
GFP-Thy1+PDGFR+ CAFs
**
A B
Host GFP Podoplanin Thy1 DAPI
Tumour periphery Tumour centre
E
Volume mm3
*
Supplementary figure 8. PD-L1 and FASL expression in human CAFs and normal
fibroblasts. (A) PD-L1 expression from tumours where data was available (colon, only
GSE1257). B) FASL expression. Data shown as mean ± SEM., *** P < 0.001. (two-tailed
unpaired Student’s t-test). Symbols represent individual samples.