Letter
https://doi.org/10.1038/s41586-018-0130-2
Bystander CD8+ T cells are abundant and phenotypically distinct in human tumour infiltrates Yannick Simoni1*, Etienne Becht1, Michael Fehlings1,2, Chiew Yee Loh1, Si-Lin Koo3, Karen Wei Weng Teng1, Joe Poh Sheng Yeong1,4, Rahul Nahar5, Tong Zhang5, Hassen Kared1, Kaibo Duan1, Nicholas Ang1, Michael Poidinger1, Yin Yeng Lee5, Anis Larbi1, Alexis J. Khng5, Emile Tan6, Cherylin Fu6, Ronnie Mathew6, Melissa Teo7, Wan Teck Lim3, Chee Keong Toh3, Boon-Hean Ong8, Tina Koh7, Axel M. Hillmer5, Angela Takano4, Tony Kiat Hon Lim4,5,9, Eng Huat Tan3, Weiwei Zhai5, Daniel S. W. Tan3,5, Iain Beehuat Tan3,5,9 & Evan W. Newell1*
Various forms of immunotherapy, such as checkpoint blockade immunotherapy, are proving to be effective at restoring T cellmediated immune responses that can lead to marked and sustained clinical responses, but only in some patients and cancer types1–4. Patients and tumours may respond unpredictably to immunotherapy partly owing to heterogeneity of the immune composition and phenotypic profiles of tumour-infiltrating lymphocytes (TILs) within individual tumours and between patients5,6. Although there is evidence that tumour-mutation-derived neoantigen-specific T cells play a role in tumour control2,4,7–10, in most cases the antigen specificities of phenotypically diverse tumour-infiltrating T cells are largely unknown. Here we show that human lung and colorectal cancer CD8+ TILs can not only be specific for tumour antigens (for example, neoantigens), but also recognize a wide range of epitopes unrelated to cancer (such as those from Epstein–Barr virus, human cytomegalovirus or influenza virus). We found that these bystander CD8+ TILs have diverse phenotypes that overlap with tumourspecific cells, but lack CD39 expression. In colorectal and lung tumours, the absence of CD39 in CD8+ TILs defines populations that lack hallmarks of chronic antigen stimulation at the tumour site, supporting their classification as bystanders. Expression of CD39 varied markedly between patients, with some patients having predominantly CD39− CD8+ TILs. Furthermore, frequencies of CD39 expression among CD8+ TILs correlated with several important clinical parameters, such as the mutation status of lung tumour epidermal growth factor receptors. Our results demonstrate that not all tumour-infiltrating T cells are specific for tumour antigens, and suggest that measuring CD39 expression could be a straightforward way to quantify or isolate bystander T cells. Using mass cytometry and a panel dedicated to the detailed profiling of tumour infiltrating T cells we observed that, consistent with previous reports5,6,11, CD8+ TILs constitute a highly heterogeneous cell population both within individual tumours (Fig. 1a) and among patients with lung and colorectal tumours (n = 144 patient tumours analysed by mass cytometry in this study) (Fig. 1b and Extended Data Fig. 1). We therefore decided to investigate the antigen specificity of CD8+ TILs to better understand the basis for this heterogeneity. In total, we screened for 1091 putative neoantigens, 123 tumour-associated antigens (TAA) and 46 cancer-unrelated epitopes (mostly virus-derived) using mass cytometry coupled to multiplex major histocompatibility complex (MHC)-tetramer staining, as reported previously12 (Fig. 2, Extended Data Fig. 2 and Supplementary Tables 1–3). Two positive hits were detected for neoantigen epitopes from a total of 24 patients tested (Fig. 2b, c and Supplementary Table 4). As 0.18% of the 1,091 computationally-predicted putative neoantigens could be confirmed
experimentally, these data are in line with other publications reporting identification rates for neoantigen-specific CD8+ T cells from predicted neoantigens of between 0% and 0.5%13–15. The small number of neoantigen-specific T cell populations detected may also be related to the relatively low mutational burden of these tumours (Supplementary Table 4), even though neoantigen-specific T cell responses have previously been reported in the context of other tumours with low mutational burden13,16. Nevertheless, these results highlight the challenge of accurately predicting and validating neoantigens for therapeutic purposes9. We also detected two tumour-specific CD8+ TIL populations in an unusual case of lung cancer associated with Epstein–Barr virus (EBV) infection (lymphoepithelioma-like carcinoma, LELC) (Extended Data Fig. 3). Despite testing 40 patient tumours with large panels of TAA-derived epitopes, we failed to identify TAA-specific CD8+ TILs. Data from in vitro expanded CD8+ T cells have shown that these cells can be detected by MHC-tetramer staining17–19. MART-1 and NY-ESO-1 epitope-specific T cells have also been detected in unexpanded TILs20,21. It is possible that TAA-specific CD8+ TIL cells were absent or present at undetectably low frequencies in all of the samples we tested. Unexpectedly, we detected cancer-unrelated MHC-tetramer+ cells (n = 46 CD8+ T cell populations) in cohorts of patients with lung cancer or colorectal cancer (in 9 of 24 lung cancer patients, 37.5%; and in 21 of 42 colorectal cancer patients, 50%) (Fig. 2b). In these cases, MHCtetramer+ CD8+ TILs were specific for various Epstein Barr virus (EBV), human cytomegalovirus (HCMV) or influenza virus epitopes that were presented by three different HLA alleles (Fig. 2d). Frequencies for individual epitopes varied between 0.07% and 3.3% of total CD8+ TILs, and 11 examples of these were validated using fluorescence flow cytometry (Fig. 2d, e, Extended Data Fig. 4). The expression of CD69 and/or CD103 in many of these cancer-unrelated CD8+ TILs suggests that they are not derived from blood contamination (Extended Data Fig. 4). These data therefore show that CD8+ TILs are not all specific for tumour antigens, but can include bystander CD8+ TILs that are specific for cancer-unrelated epitopes. Having identified cancer-unrelated bystander and tumour-specific CD8+ TILs, we next compared the phenotypes of these two populations with those of remaining CD8+ TILs of unknown specificity. All the tumour-specific CD8+ TILs that we identified displayed resident memory T cell-like phenotypes and expressed various co-stimulatory and inhibitory receptors, such as PD-1 (Fig. 3a and Extended Data Fig. 3, 5). Surprisingly, we observed overlapping but diverse phenotypic profiles for cancer-unrelated CD8+ TILs with respect to these markers. Many of the bystander CD8+ TILs also expressed resident memory T cell-like phenotypes as well as various co-stimulatory
1
Agency for Science, Technology and Research (A*STAR), Singapore Immunology Network (SIgN), Singapore, Singapore. 2immunoSCAPE, Singapore, Singapore. 3Division of Medical Oncology, National Cancer Centre Singapore (NCCS), Singapore, Singapore. 4Department of Anatomical Pathology, Singapore General Hospital, Singapore, Singapore. 5Agency for Science, Technology and Research (A*STAR), Genome Institute of Singapore (GIS), Singapore, Singapore. 6Department of Colorectal Surgery, Singapore General Hospital, Singapore, Singapore. 7Division of Surgical Oncology, National Cancer Centre Singapore (NCCS), Singapore, Singapore. 8Department of Cardiothoracic Surgery, National Heart Centre Singapore (NHCS), Singapore, Singapore. 9 Duke–National University of Singapore Medical School, Singapore, Singapore. *e-mail:
[email protected];
[email protected] N A t U r e | www.nature.com/nature
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
RESEARCH Letter a
Lung tumour (CD8+ TILs – patient A167) CCR7
CD45RO
CD69
CD103
CD49a
CD27
CD28
CD127
ICOS
OX-40
TIGIT
2B4
CD57
PD-1
CD39
t-SNE1
Normalized intensity
t-SNE2
b
All patients
Low Patient A139
Patient A167
Patient A170
Patient A248
Patient A250
Fig. 1 | Tumour-infiltrating CD8+ TILs are phenotypically heterogeneous within a tumour and across patients. a, t-distributed stochastic neighbour embedding (t-SNE) map of CD8+ TILs isolated from a colorectal tumour. t-SNE was performed on one patient to explore the heterogeneity of CD8+ TILs within an individual (see also Extended Data Fig. 1 and Methods). Representative data from one patient (see Methods for source data availability). b, t-SNE map of CD8+ TILs isolated from lung tumours or colorectal tumours. Mass cytometry and t-SNE were performed simultaneously on six different patients from each cohort to explore the heterogeneity of CD8+ TILs across patients. Patient identifiers refer to individual patients. Representative data from n = 6 patients for each cancer type.
High
Patient A255
103 101 100 0 101 102 103
0 101 102 103
0 101 102 103
0 101 102 103
0 101 102 103
All patients
Patient 980
Patient 1011
Patient 1053
Patient 1054
0 101 102 103
0 101 102 103
0 101 102 103
0 101 102 103
0 101 102 103
0 101 102 103
Patient 1201
0 101 102 103
Patient 1227
103 102
t-SNE1
Colorectal tumour (CD8+ TILS)
Lung tumour (CD8+ TILs)
102
101 100 0 101 102 103
0 101 102 103
t-SNE2
and activation marker molecules. The inhibitory receptors TIGIT and PD-1, two markers that were previously shown to be expressed by tumour-antigen-specific CD8+ T cells22, were also expressed by many of these cells (Fig. 3a, c). Although PD-1 has been proposed as a marker of tumour-specific CD8+ T cells23, our results are consistent with previous reports of virus-specific CD8+ T cells infiltrating tumours that express PD-1 in mice24. However, we observed a striking lack of CD39 expression in bystander CD8+ TILs (5.2 ± 8.4% (s.d. is used throughout), n = 46). By contrast, CD39 was highly expressed by tumour-specific CD8+ TILs and variably expressed by cells of unknown specificity (40.4 ± 27.2%, P < 0.0001) (Fig. 3b, c). CD39 is a transmembrane extracellular ATPase that is widely expressed by regulatory T cells, B cells and some tumour cells. In conjunction with the enzymatic activity of CD73, CD39 can catalyse the conversion of ATP to adenosine, which has been shown to have immunosuppressive activity25,26. Based on these data, we hypothesize that the lack of CD39 could be used to enrich for cancer-unrelated bystander CD8+ TILs. Conversely, though we think that these results suggest that CD39 could also be a useful marker of tumour-specific CD8+ TILs, this link could be observed only in two neoantigen responses from two patients and two tumour-specific CD8+ TIL populations in an unusual LELC tumour. To better compare the characteristics of CD39– CD8+ and CD39+ CD8+ TILs, we performed transcriptomic profiling. Using principal component analysis (PCA) and gene set enrichment analysis (GSEA), we found that CD39+ CD8+ TILs were enriched in expression of genes related to cell proliferation and exhaustion, which are characteristics of chronically stimulated T cells27–29 (Fig. 4a, b and Extended Data Fig. 6), consistent with previous reports in both cancer25 and infectious disease26. In line with this, T cell receptor (TCR) sequencing indicated a skewed and reduced diversity of TCR sequence diversity in CD39+ CD8+ TILs (Extended Data Fig. 7), supporting the notion of enrichment for cells that have undergone tumour-antigen-driven
clonal expansion29,30. At the protein level, compared to their CD39– counterparts, CD39+ CD8+ TILs from colon and lung tumours showed hallmarks of exhausted cells in terms of both phenotypic and functional markers (Fig. 4c, d and Extended Data Fig. 8), consistent with the transcriptomic profiling data. Thus, expression of CD39 defines a population of highly exhausted cells; whereas the absence of CD39 in CD8+ TILs defines a population whose phenotype is inconsistent with chronic antigen stimulation at the tumour site, consistent with a bystander role. Next, we investigated whether expression of CD39 by CD8+ TILs was linked to clinical parameters measured in either of the studied patient cohorts. In colorectal tumours, we detected highly heterogeneous frequencies of CD39 expression among CD8+ TILs (n = 94; mean, 44.5 ± 23.7%; minimum, 0.2%; maximum, 85.8%) (Fig. 4e and Extended Data Fig. 9). No significant correlations with any clinical parameters (tumour mutational burden, driver mutational status or consensus molecular subtype (CMS)) were obtained from this cohort (Extended Data Fig. 9). However, based on the transcriptomic profiles of adjacent frozen tumour sections, we found that tumours with higher percentages of CD39+ CD8+ TILs had gene expression profiles indicative of T cell inflammation and expression of pathways associated with antigen processing and presentation (Extended Data Fig. 10). In lung cancer patients, the expression of CD39 in CD8+ TIL cells was also highly heterogeneous across patients (n = 50; mean, 21.9 ± 23.23%; minimum, 0%; maximum, 88.5%) (Fig. 4e and Extended Data Fig. 9). For this type of cancer, the frequency of CD39+ CD8+ TIL cells clearly correlated with the epidermal growth factor receptor (EGFR)-mutation status, an oncogenic driver mutation that is especially common in East Asian patients with lung cancer31. Preliminary reports suggest that patients with EGFR-mutated tumours are relatively poor responders to anti-PD-1 treatment and have low CD8 + T cell density compared to patients with EGFR-wild-type tumours32.
N A t U r e | www.nature.com/nature
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
101
102
103
0 101
0
102
103
Streptavidin (Tb159)
Streptavidin (Sm154)
101
0 101
0
102
103
103
101
0 101
0
Streptavidin (Ho165)
102
103
EBV epitope (154+ 174+ 175+)
103
c
102 101
0 101
0
Streptavidin (Yb171)
102
103
Streptavidin (Yb174)
0.1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
0.01
0
103
104
105
Patient 974 1.11%
0
CD8
103
104
105
0.1 0.01
Patient 1128 1.3%
0
CD8
104
105
103
104
105
103
0
0 0
103
104
105
CD8
9- EBV (LMP2) 10- HCMV (IE1) 11- HCMV (pp65 #1) 12- HCMV (pp65 #2) 13- HCMV (pp65) 14- HCMV (pp65) 15- Influenza (NP #1) 16- Influenza (NP #2)
104
105
CD8
Patient 1278 0.73%
0
103
104
105
CD8
DST epitope (mutated residue is underlined), ISDEMFKTFK; MHCtetramer mutAHR epitope, GISQELPYK. Percentages represent the MHC tetramer+ cells among CD8+ TILs for each patient. Data from two independent experiments. d, Frequencies of cancer-unrelated CD8+ TILs identified by mass cytometry and multiplex MHC-tetramer staining in lung tumours (right, n = 11 MHC tetramer+ populations) or colorectal tumours (left, n = 35 MHC tetramer+ populations). Peptide sequences are listed in Supplementary Table 3. Inf., influenza virus. e, Representative flow cytometry dot plot showing cancer-unrelated CD8+ TILs specific for different epitopes, identified from colorectal tumour CD3+ TILs using mass cytometry screening. Percentages are of MHC tetramer+ cells among CD8+ TILs for each patient. See also Extended Data Fig. 4. Data are from two independent experiments.
c
Patient 1128
Patient 1414
Patient 1854
9
-1
D3
PD
C
7
G IT
D5
TI
C
RG 1
5
KL
O
C
D2
40 X-
S
R
O IC
27
LA H
Patient 1539
Patient 1053
104
PD-1
*
-D
8
D1
D2
C
C
C
7 D2
9a
03
D4 C
D1 C
D6
5R
C C
C
D4
9
Tumour-specific tetramer+ CD8+ TILs
O
100 90 80 70 60 50 40 30 20 10 0
103
103
All CD8+ TILs
C
CD39+ CD8 TILs (%)
b
0
104
Cancer-unrelated tetramer+ CD8+ TILs
R7
Positive cells (%)
100 90 80 70 60 50 40 30 20 10 0
0
Patient 1053 4.38%
Tetramer A*02:01 Tetramer A*11:01 Tetramer A*24:02
Patient 1171 0.31%
CD8
Fig. 2 | Tumour-specific and cancer-unrelated CD8+ T cells infiltrate tumour tissues. a, Schematic for screening of neoantigens (NeoAg), tumour-associated antigens (TAA) and cancer-unrelated epitopes by mass cytometry coupled to multiplex MHC-tetramer staining. See also Extended Data Fig. 2 and Supplementary Tables 1–3 for examples and list of peptides. b, Total number of different MHC class I tetramers screened for neoantigens, TAA and cancer-unrelated epitopes by mass cytometry (left). Total number of different MHC class I tetramers identified for neoantigens, TAA and cancer-unrelated epitopes by mass cytometry (right). Neoantigens are colour-coded by patient. See also Supplementary Table 4 for patient information. c, Flow cytometry dot plots representing MHC tetramer+ CD8+ TILs identified from colorectal tumour CD3+ TILs using mass cytometry screening. MHC-tetramer mutated (mut) a
103
1- EBV (BMFL1) 2- EBV (BMFL1) 3- EBV (BRFL1) 4- EBV (BRFL1) 5- EBV (EBNA3A) 6- EBV (EBNA3B) 7- EBV (EBNA3B) 8- EBV (EBNA4)
1
Tetramer A*24:02 EBV (EBNA3A)
Patient 1539 3.25%
104
105
Cancer-unrelated epitopes
HCMV Inf.
EBV 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1
0.11%
CD8
Tetramer A*24:02 Influenza (NP)
HCMV Inf.
Tetramer+ CD8+ TILs in lung tumours (%)
EBV 10
Tetramer A*11:01 EBV (LMP2)
e
Tetramer A*02:01 HCMV (pp65)
d
Tetramer+ CD8+ TILs in colorectal tumours (%)
Ex vivo antigen specific CD8+ TILs phenotype characterization
Patient 1218 105
TA A un Ca re nc la er te d
neoantigen epitope (167+ 171+ 174+)
102
Tetramer A*24:02 HCMV (pp65)
0
102
Tetramer A*11:01 neoAg mutDST
0
101
103
Streptavidin (Lu175)
101
103 102
Streptavidin (Er167)
102
Streptavidin (Dy164)
Streptavidin (Dy161)
Streptavidin (Gd158)
Multiplex tetramer staining by mass cytometry (CD8+ TIL cells – patient 1053) 103
N eo an tig en
Colorectal tumour
16 14 12 10 8 6 4 2 0
Tetramer A*11:01 neoAg mutAHR
Lung tumour
1,200 1,000 800 600 400 200 0 N eo an tig en
Multiplex tetramer screening 1- Tissue collection 2- Tumour/normal exome sequencing Tumour-specific epitopes (Neoantigens, tumour-associated antigens) 3- Mutation identification Cancer-unrelated epitopes 4- Neoantigen MHC class I (HCMV, HIV, Influenza, ...) binding prediction
TA A un Ca re nc la er te d
b
Number of epitopes screened
a
Number of epitopes identified
Letter RESEARCH
103 102 0
4%24.7% 0 102
103
CD39
104
2% 0 102
103
CD39
34% 104
1.1% 7.1% 0 102
103
CD39
Fig. 3 | Cancer-unrelated CD8+ TILs do not express CD39. a, Expression of markers by cancer-unrelated (blue, n = 46 biologically independent MHC tetramer+ cells) and tumour-specific CD8+ TILs (red, n = 4 biologically independent MHC tetramer+ cells) in human tumours. Triangles represent neoantigen-specific CD8+ TILs and squares represent tumour-specific CD8+ TILs derived from an LELC (See Extended Data Figs. 3, 5). Data are mean ± s.d. from at least ten independent mass cytometry experiments. Each data point represents an antigen-specific population. b, Expression of CD39 by cancer-unrelated (blue, n = 46 biologically independent MHC tetramer+ cells) or tumour-specific CD8+ T cells (red, n = 4 biologically
104
2.2% 24.9% 0 102
103
CD39
104
96.9% 80% 0 102
103
104
CD39
independent MHC tetramer+ cells) with paired total CD8+ TILs. Triangles represent neoantigen-specific CD8+ TILs and squares represent tumourspecific CD8+ TILs in an LELC (See Extended Data Fig. 3). Data are mean ± s.d.; paired two-tailed t-test; lung tumour only, P = 0.0064; colorectal tumour only, *P < 0.0001. c, Representative flow cytometry dot plot representing the expression of PD-1 and CD39 by cancer-unrelated (blue) or tumour-specific CD8+ TILs (red) identified from colorectal tumours by mass cytometry. Frequencies of CD39+ cells among cancer-unrelated (blue), tumour-specific (red) or all CD8+ TILs (grey) for each patient. Data are from two independent experiments. N A t U r e | www.nature.com/nature
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
RESEARCH LETTER a
b
Exhaustion gene set enrichment
100
PC1
CD39
f
60 40 20
PB M Tu C m ou r
0
M Tu C m ou r
Lung cancer Colorectal cancer
100
Pembrolizumab (anti-PD-1) Blood collection Day 24 Day –1 Day 9
Colorectal tumour
40
0
7
0
80
20
KI 6
20
0.0234
60
C D2 5
40
g
Lung tumour
CD39+ CD8+ T cells (%)
*
80
PB
103
E EG GF FR R W m T ut an t
0.0003
100
102
0.0003
*
4
0
101
*
A-
103
(172Yb)
e CD39+ CD8+ T cells (%)
102
*
3
0
101
*
60
7
103
*
80
TL
0
102
*
100
C
101
6.3% 74.7%
0
D5
0
Positive cells (%)
2.5% 24.6%
48% 90.6%
All CD8 (day 9) Cluster (day 9)
Day –1
Day 9
Day 24
103
t-SNE1
102 101
CTLA-4 (166Er)
103
PD-1 (160Gd)
TIM-3 (167Er)
CD39 (172Yb)
20
M
103
C
102
TI
101
0.0021
40
102 101 0
2.7% 0
101 102 103
25.9% 0
101 102 103
10% 0
101 102 103
t-SNE2
Ki67 (152Sm)
0
IC O S O X40
103
*
IT
102
*
G
101
*
60
-1
0
*
TI
103
*
80
7 H LA -D R
102
0.75
PD
101
0
4.8% 45.9%
0.0288 100
C D1 2
0
67.9%
Colorectal tumour Positive cells (%)
102 101
3%
Ki67 (162Dy)
56.4%
5%
103
10 5 0 –5
d
Colorectal tumour (CD8+ TILs) ICOS (152Sm)
OX40 (149Sm)
c
CD39– CD8+ TILs CD39+ CD8+ TILs
5,000 10,000 15,000 Position in the ranked list of genes
4
Naive CD8+ Effector CD8+
0
2B
0
0.0
Phenotypes
−200 –100
0.1
–200
0.2
MKI67 PDCD1 CD244 CTLA4
−100
P < 0.001
0.3
1
ENTPD1
C D2 8
RG
PC2
KL
0
0.4
C D2 7
Enrichment score
100
103 102 101 0 0
101 102 103
CD39 (160Gd)
Fig. 4 | Comparative analysis of CD8+ TILs stratified by CD39 expression. a, Projection of the whole transcriptome of sorted blood naive (CCR7+ CD45RO−, white; n = 4 patients), blood effector (CCR7− CD45RO+, grey; n = 5 patients), tumour CD39− (blue, n = 7 patients) and CD39+ CD8+ TILs (red, n = 8 patients) using PCA. Ellipses represent the standard deviation around the centroid of a phenotype. See also Supplementary Table 5. b, Enrichment of the gene set for exhausted T cells28,29 in CD39+ CD8+ TILs. Genes towards the left are enriched in CD39+ CD8+ TILs (n = 8 patients), genes towards the right are enriched in CD39− CD8+ TILs (n = 7 patients). Two-sided GSEA empirical test. See also Extended Data Figs. 6, 7. c, Mass cytometry dot plots representing expression of OX-40, ICOS, Ki67, TIM-3, PD-1 and CTLA-4 according to CD39 status by CD8+ TILs. Representative data from one patient with a colorectal tumour. Data are from at least ten independent mass cytometry
experiments. d, Frequencies of the expression of activation markers (top) and inhibitory markers (bottom) by CD39− (blue) and CD39+ (red) CD8+ TILs in colorectal tumours (n = 9 to 45 patients). Data are mean ± s.d. from at least ten independent mass cytometry experiments. Two-tailed paired t-test. *P < 0.0001. See also Extended Data Fig. 8. e, Frequencies of CD39+ CD8+ T cells in lung tumours (n = 50 patients), colorectal tumours (n = 94 patients) and the matched PBMCs. Data are mean ± s.d. f, Frequencies of CD39+ CD8+ T cells in lung tumours stratified by mutation status of EGFR. Wild-type (WT) EGFR, n = 17 patients; mutant EGFR, n = 25 patients. Data are mean ± s.d., two-tailed unpaired t-test. g, Schematic for blood collection time point in colorectal cancer patients treated with pembrolizumab (anti-PD-1 treatment) (top). t-SNE map of CD8+ T cells isolated from PBMC at −1, 9 and 24 days relative to pembrolizumab treatment (bottom).
In our study cohort, the frequency of CD39+ CD8+ TIL cells was significantly higher in patients with EGFR-wild-type tumours (32.3% ± 20.35%) in comparison to those with EGFR-mutant tumours (16.24% ± 23.51%), where these cells were virtually absent (
