Background
Immune checkpoint blockades (ICBs) therapy opens a new era of cancer immunotherapy, which creates a holistic version and radically changes for cancer treatment [
1,
2]. Unfortunately, its clinical application has been limited by a low response rate. Especially, certain molecular subgroups of cancer, like EGFR-mutated lung cancer, were reported to obtain low efficacy of ICB therapy in clinical [
3,
4]. EGFR mutation leads to tumor immune escape and compromises infiltration of tumoricidal effector of T cells [
5]. Although tumor microenvironment (TME) restriction on immune cells has been well studied [
6‐
8] and significant efforts have been taken to explore the potential way to enhance T cell infiltration into the tumor bed [
9,
10], the mechanism of EGFR-mutation inducing un-inflamed TME remains unknown. With the aim of mapping proteins related to the non-inflamed TME of EGFR-mutated lung cancer, global proteomics and phosphoproteomics data from cancer tissues were analyzed. Based on the non-inflamed phenotype of EGFR-driven non-small cell lung cancer (NSCLC), we found that targeting PKCδ is a promising strategy to induce tumor infiltrating lymphocytes (TIL). These turn “cold” tumors “hot” and make them more susceptible to ICB therapy. Our findings provide a novel avenue for enhancing the efficacy of tumor immunotherapy.
Methods
Materials
Rottlerin and PEP-005 were purchased from Sigma-Aldrich. CD3/28 antibody, IL-2, and fluorescence conjugated PD-L1/CD3/CD4/CD8 was obtained from Biolegend. Anti-PD-1 mAb (αPD-1) was purchased from BioXcell. CRISP-Cas9 plasmid of PKCδ was purchased from GenScript. RIPA buffer and primary antibodies against GAPDH, PD-L1, p-PKCδ, NF-κB, and p-P65 were purchased from Cell Signaling Technology. Total-PKCδ was obtained from Abcam. Anti-rabbit and anti-mouse secondary antibodies were purchased from Odyssey.
Cell lines and culture
All cell lines were purchased from ATCC. Lung cancer cell lines H1975 (EGFRL858R+T790M mutation), A549 (KRAS mutation), H460 (KRAS mutation), H1650 (EGFRExon19 del), H820 (EGFRL858R+T790M mutation), and H1819 (EGFR overexpression) were cultivated in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin. Mouse cell lines LLC1 (mouse lung cancer cell) and NIH/3T3 (mouse embryonic fibroblast) cells were cultured in DMEM medium, while BEAS-2B (normal lung epithelial cell) was cultured in BEBM medium. PKCδ knock out cell line was generated in H1975 cells. EGFRmut cells were generated by overexpressing EGFRL858R+T790M in BEAS-2B cells. All the cells were cultivated at 37 °C in a 5% CO2 incubator. Agents (rottlerin and PEP-005) were dissolved in DMSO to generate the stock solutions (20 mM), and the stock solutions were diluted with full culture medium to their target concentrations.
CRISPR-Cas9
CRISPR-Cas9 was used to knockout the target gene in NSCLC cells. Briefly, PRKCD CRISPR Guide RNA (Sequence: CTCCGCGGCGGTTCATCGTT) was constructed into lentivirus vector pLentiCRISPR v2 which was used as vehicle control. After transfected to target cells (H1975), puromycin was applied for screening the stable clone of PRKCD knockout.
Real-time PCR
The expression level of mRNA was quantified by Real-time PCR by using FastStart Universal SYBR Green Master (Roche), following this protocol: 94 °C for 10 min, followed by 40 cycles at 94 °C for 10 s and 60 °C for 30 s. Actin was considered as internal standard. The primers sequences are as following:
ICAM1 Forward primer: TCTTCCTCGGCCTTCCCATA
ICAM1 Reverse primer: AGGTACCATGGCCCCAAATG
Actin Forward primer: GATATTGGCAACGACCCCCA
Actin Reverse primer: CCCAGCCAGGATCTTGAAGG
Flow cytometry analysis of PD-L1 expression
After cells were trypsinized, 1 × 105 cells were re-suspended in 100 mL of staining buffer containing 1μl APC-conjugated anti-human PD-L1 antibody (BioLegend) and incubated at room temperature for 15 min. After washing with PBS for 3 times, cells were analyzed by flow cytometry.
T-cell killing assay
PKCδ−/− H1975-GFP cells and the control were seeded in a 96-well plate overnight for adhesion. Human PBMCs were isolated from healthy donors by Ficoll centrifugation and immediately frozen in − 80°. PBMCs were activated with 2 μg/mL CD3/28 antibodies and 10 ng/mL IL-2, and then co-cultured with H1975 luciferase cells at 10:1 ratio. The cell number was calculated using an INCell Analyzer 6000 imaging system.
Database acquisition
A large-scale and publicly available collection of multi-omic datasets from 103 lung adenocarcinoma (LUAD) cases in Chinese patients was released by the Tan Minjia group [
11]. Integrative analysis of proteomic and phosphoproteomic data from this collection revealed cancer-associated characteristics in patients with EGFR mutations. The normalized iBAQ intensities released by Tan’s group were used in quantitative analysis of proteomic data. Downstream statistical analysis by Perseus software (
https://www.perseus-framework.org/, version 1.5.5.3) was done following the standard protocol [
12]. Samples were then grouped into NT and tumor group which was further divided into groups with or without EGFR mutation. Only proteins with 3 valid values in at least one group were kept. Student’s
T-test was performed between experimental and control groups with or without EGFR mutation, with false discovery rate (FDR) < 0.05 and S0 = 0.1.
Next, the intensities of the phosphopeptide signals released by Tan’s group were used in quantitative analysis of phosphoproteomic data. Student’s T-test was operated first between tumor samples with and without EGFR mutation and, second, between experimental and control groups in EGFR mutation samples. Results with p < 0.05 were considered statistically significant. The missed proteins (zero or one value in the NT group, but five values in the tumor group) were rescued and combined with the phosphosites found to be significant.
Proteomic analysis
One hundred μg of proteins were reduced and alkylated with 1mM dithiothreitol and 0.5mM iodoacetamide, respectively, followed by digestion with trypsin in 1:100 (w/w) ratio overnight. Then, peptides were desalted as described previously [
13]. Finally, the desalted peptides were dried in vacuum and dissolved for LC-MS/MS analysis. LC-MS/MS analysis was performed on an Easy nLC system (Thermo Scientific, USA) coupled to a Q Exactive mass spectrometer (Thermo Scientific, USA) for 60 min. The mass spectrometer was operated in positive ion mode by full-scan MS scan (m/z 300-1550, resolution 70000) followed by data-dependent MS/MS scan (top 10 modes, resolution 17500). The MS data were analyzed using MaxQuant software (version 1.5.5.1) [
14]. Proteins were identified against the human proteome sequences from UniProtKB (state July 2017, 70698 entries). 0.02 ppm and 7 Da were set for fragment ion and precursor ion tolerances, respectively. “2 missed cleavages” was enabled for tryptic peptide. Carbamidomethylation was chosen as static modification and oxidation and deamidation were selected as dynamic modifications. FDR of 0.01 was used in peptide identification. The label-free protein quantitation (LFQ) was performed using the LFQ algorithm [
15]. Bioinformatics and statistical analyses were performed in Perseus software. Gene ontology enrichments were computed using the ‘enrichGO’ function from R package ‘clusterProfiler’ and top significantly enriched terms were selected [
16].
Protein−protein interaction analysis
The protein-protein relationship data was obtained from the STRING database based on experimental sources (version 11.0;
https://string-db.org/). To find more novel EGFR interacting proteins, we also selected text mining resource in STRING analysis. Low-confidence edges (edges with a confidence score < 0.4) were removed from the network.
Immunohistochemical staining (IHC) of human lung cancer tissue samples
One hundred patients’ samples were collected from West China Hospital following the hospital guidelines and patients signed informed consent in all cases. The slides with specimens were incubated with primary antibodies for PD-L1 and phosphor-PKC-δ (1:100 dilutions) overnight at 4 °C, which was detected by a biotin labeled anti-IgG secondary antibody and streptavidin-Horseradish peroxidase (HRP). Staining on the slides was quantified by colorimetric detection. According to the intensity of staining, specimens were classified into three levels: high (+++), medium (++), and low/negative (+/−).
NIH/3T3 cells and H1975 were trypsinized and collected. 2.5 × 104 NIH/3T3 cells and 0.5 × 104 H1975 cells were mixed and suspended in 200 μL culture medium in cell-repellent surface 96 well plates and centrifuged at 180g × 3 min. Plate was cultured overnight, and spheroid formation was observed.
Cell surface marker and intracellular cytokines staining
Blood cells were collected by centrifuging at 350 g for 5 min. Cell pellets were re-suspended in 3 ml 1X RBC Lysis Buffer and incubated on ice for 5 min to lyse red blood cells. The lysate was then centrifuged for 5 min at 350 g, and supernatant was discarded. To prepare cells for cell surface marker staining, pre-incubation with Fc Receptor Blocking Solution was required, and then cells were added with conjugated fluorescent antibodies (e.g., anti-CD8-APC) on ice for 15–20 min in the dark. For some samples, further staining of intracellular cell components was required. These cells were fixed and permeabilized in Perm/Wash Buffer overnight. The fixed/permeabilized cells were then suspended in intracellular staining perm/wash Buffer with a conjugated antibody for 20 minu in the dark at room temperature. In the end, stained cells were loaded into a flow cytometer for analysis.
Tumor dissociation, CD8+ and NK1.1+ isolation
The tumor tissues were dissociated, using a kit (Miltenyi Biotec), first by enzymatic digestion then gentleMACS™ Dissociators are used for mechanical dissociation steps. After dissociation, mixture was filtered using a 70-mm filter to obtain single-cell suspension. Next, CD8+ and NK1.1+ cells were isolated by manual magnetic labeling kit. Briefly, each 107 total cells were incubated with 10 μL of Biotin-Antibody Cocktail and incubated for 5 min in 4 °C and then 20 μL of CD8+/NK1.1+ T Cell Micro-Bead Cocktail was mixed for 10 min. Finally, labeled cells were added to the column for isolation by magnetic MACS separator.
Immune cell quantification in tumor spheroids
Three days after the tumor spheroids were seeded, PBMCs were added. PBMCs were isolated from healthy donors using the density gradient technique with the Ficoll PLUS from GE Healthcare. 5 × 104 immune cells/well were added and co-cultured overnight. Before harvest, tumor spheroids were washed with PBS three times (1 minutes each time) to remove surface attached cells. After dissociation with Accumax (eBiosciences), the number of immune cells within the spheroids were quantified by flow cytometry.
Xenograft mouse model
Animal studies were approved by the Ethical Committee of Macau University of Science and Technology. The mouse tumor model was established as previously described [
17]. 1 × 10
6 LLC1 mouse lung cancer cells/100 μl were subcutaneously injected into the right forelimb of C57BL/6 mouse. After 5 days, the mice with tumor volume reached 5 mm
3 were divided into different treatment groups (
n = 6): Control (treated with 200ul PBS /day by I.P), rottlerin (5mg/kg), anti-PD-1(200ug/time), and a combination treatment of rottlerin and anti-PD-1. Rottlerin was administered once a day and anti-PD-1 once a week by I.P administration. The tumor dimensions (length and width) were measured every 3 days, and the tumor volume was calculated by following equation: volume = (width
2 × length)/2. Mice of each group were sacrificed at day 21 for taking images of tumor. For survival analysis, another six mice of each group were used for long-term study and calculated overall survival data.
Statistical analysis
All data for the three experiments was analyzed using GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA). One way analysis of variance (ANOVA) was used to analyze the differences between three or more groups, and Student’s t test was used for comparisons of only two groups. Results were represented as means ± SEM. Any results with p < 0.05 were considered statistically significant.
Discussion
Infiltration of functional cytotoxic T lymphocytes into the TME is essential for inducing durable clinical responses to ICB therapy, and the presence of sufficient TILs is a critical indicator of good prognosis for patients [
27]. In other words, ICB works best against so-called “hot” tumors (T cell inflamed tumors). These resulted in which only a minority of cancer patients can recruit sufficient TILs in established tumors and merely 20–30% clinical response rate of ICB. Patients with EGFR mutations especially exhibited uninflamed phenotypes and weak immunogenicity and consequently showed an unfavorable response to PD-1 blockade immunotherapy [
18,
28]. Development of an effective way to facilitate T cell infiltration into the TME is urgently required for current clinical therapy.
In this study, we found that PKCδ is responsible for the non-inflamed phenotype of EGFR-mutated tumors. PKCδ activates NF-κB and mediates the upregulation of cell-cell adhesion genes, which, in turn, results in the formation of a physical barrier, decreasing T cell infiltration into the TME and ultimately failure of ICB therapy. The finding was further verified by combining treatment of PKCδ inhibition and αPD-1, resulting in significantly enhanced antitumor efficacy of αPD-1. Therefore, our study provides insight into overcoming the lack of effective strategies to enhance the clinical efficacy of ICB therapy.
Recently, PKC isozymes are demonstrated to be closely involved in multiple signal transduction systems that respond to a variety of external signals, including hormones, growth factors, and other membrane receptor ligands [
29,
30]. Meanwhile, PKC isozymes are also widely involved in tumorigeneses. For example, PKCι/λ and PKCζ, are now considered fundamental regulators of tumorigenesis [
31]; PKCε acts as a key regulator of mitochondrial redox homeostasis in acute myeloid leukemia [
32]. Among different isoforms, PKCδ was reported as a critical regulator of immune homeostasis and closely involved in autoimmune disease and cancer progression [
21,
33]. The oncogenic role of PKCδ has been demonstrated in preclinical and clinical data, including promotion of lung KRAS-dependent tumorigenesis [
34] and negative correlation with the prognosis of ErbB2-driven tumorigenesis [
35]. Interestingly, ectopic expression of PKCδ in NSCLC was shown to lead to TKI-resistance in EGFR-mutant lung cancer patients [
36]. This resistance is typical of “cold” tumors. This previous research is consistent with our findings and further supports the important role of PKCδ in inducing an immunosuppressive TME.
As a signaling molecule downstream of PKCδ, NF-κB family members and their regulated genes have been linked to malignant transformation, tumor cell proliferation, survival, angiogenesis, invasion/metastasis, and therapeutic resistance. NF-κB is reported to be closely involved in cancer initiation and progression [
37,
38]. The concept of “NF-κB addiction” was widely accepted in cancer [
39]. It was reported to constitutively activate in different types of human cancers and regulate various oncogenic genes in cancer development and progression. Moreover, the association of the immunosuppressive TME with NF-κB has been proven in many cancers. For example, human ovary cancer constitutively activates NF-κB signaling and produces cytokines which impair T cell activity and promote expansion of immunosuppressive MDSCs [
40]. Although the pro-inflammatory effect of NF-κB has been well established, based on this research, for certain types of tumors, temporary blocking of NF-κB signaling will contribute to enhancing clinical efficacy of cancer immunotherapy.
Induction of PD-L1 is another important mechanism of PKCδ promoting cancer progress and escaping immune surveillance, which also uncovers a novel pathway between EGFR and PD-L1. Especially for TKI resistant NSCLC, it provides a potential explanation why such tumors are prone to the uninflamed status in TME.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.