Background
The application of immune checkpoint blockade (ICB), such as programmed death-1 receptor (PD-1) and programmed death ligand 1 (PD-L1) antibodies, was shown to reinvigorate T cell function and prolong survival in various cancer types [
1‐
5]. However, due to intrinsic or acquired drug resistance, only a minority of patients experience long-term benefits from ICB. Pre-existing cytotoxic T cells in the tumor microenvironment (TME) are a prerequisite for the reinvigoration of T cells and the resulting inflammation, indicating that an immuno-hot TME induces a beneficial response to ICB administration [
6]. However, the intrinsic tumor-driven force that leads to an immuno-hot TME remains unclear.
It is known that the form of tumor cell death instructs an immuno-hot or -cold TME. Immunogenic cell death (ICD) supplies an immuno-hot TME by promoting antigen release, antigen presentation, and cytotoxic T cell activation, inducing successful antitumor immunity [
7]. Strategies aimed at eliciting ICD have been used to overcome resistance to ICB treatments [
8]. Necroptosis, pyroptosis, and ferroptosis are the predominant immunogenic forms of cell death, and apoptosis is usually regarded as an immune-tolerogenic process [
9‐
11]. Furthermore, cells that undergo necroptosis activate the immune system, particularly through antigen presentation and cross-priming of CD8
+ T cells [
12]. In pyroptosis, gasdermin proteins are cleaved by inflammatory caspases, leading to inflammatory cytokine release and cell death [
13]. In cancer cells, gasdermin E cleaved by caspase-3 is an essential mediator of pyroptosis, which converts non-inflammatory apoptotic signals into pyroptotic cell death and suppresses tumor growth [
14].
Caspase-8 (Casp8) is a switch for immune-tolerogenic apoptosis, immunogenic necroptosis, and pyroptosis [
15]. It has been reported that in the presence of Casp8 malfunctions, the form of cell death could switch to necroptosis [
16]. It is likely that Casp8 malfunction leads to ICD in cancer cells, which may provoke an adaptive immune response, facilitating CD8
+ T cell infiltration and inducing an inflamed (hot) TME, which in turn improves the efficacy of ICB immunotherapies. Consistent with this theory, in a TRAF
−/− melanoma mouse model, tumor cells redirected the TNF signaling pathway to favor RIPK1-dependent necroptosis, enhance tumor eradication, and show a better response to anti-PD-1 therapy than the control group [
17].
However, evidence also suggests that necrosis-induced inflammation only facilitates tissue repair responses and is not sufficiently effective to induce anticancer immunity [
18,
19]. Moreover, the function of Casp8 may be catalytically activity-dependent or -independent. Apart from the aforementioned role of cleavage-dependent Casp8 function, the expression of catalytically inactive Casp8 is both necessary and sufficient to induce inflammasome formation [
15]. This implies a complicated role of Casp8 in cell death and adaptive and innate immune responses.
In this study, we explored The Cancer Genome Atlas (TCGA) database and ICB-treated cohorts to determine the role of Casp8 in the TME and ICB responsiveness. We further established a Casp8 knockout cell line and animal models to understand the underlying mechanisms.
Methods
Cell lines
B16F10 cells were purchased from the American Type Culture Collection and cultured in Dulbecco’s modified Eagle’s medium supplemented with fetal bovine serum (10%), penicillin (100 U/mL), and streptomycin (100 mg/mL) at 37 °C in a humidified atmosphere containing 5% CO
2. The caspase-8-knockout B16F10 cell line (B16-C8KO) was generated using CRISPR/Cas9 technology. The gRNAs encoding caspase-8 are shown in Additional file
1: Fig. S1.
Animals and animal models
Female C57BL/6 mice aged 6–8 weeks were purchased from the Center of Experimental Animals of the Third Military Medical Univercity (TMMU). Nude mice were purchased from VitalStar Biotechnology Co. Ltd. (Beijing, China). The mouse handling protocols were approved by the Institutional Animal Care and Use Committee of TMMU. To establish tumor models, B16F10 cells (2 × 105 in 100 µL of PBS) were subcutaneously inoculated into the right flank of 6–8-week-old female C57BL/6 mice or nude mice. When the tumors became palpable, tumor volume was monitored twice per week. In immunotherapeutic models, 2 × 105 B16F10 cells were subcutaneously inoculated into the right flank of 6–8-week-old female C57BL/6 mice. Mice received 200 µg of intraperitoneal anti-PD-L1 monoclonal antibody (10F.9G2, Be0101, BioXcell) or the equivalent isotype control antibody (BioXcell, BE0090) on days 4, 7, and 10.
For radiation-combined immunotherapeutic models, 2 × 105 B16F10 cells were subcutaneously inoculated into the right legs of C57BL/6 mice. When the tumors reached approximately 50 mm3, the mice were locally irradiated using the Varian Trilogy Stereotactic System at a single dose of 20 Gy. On the same day, 200 µg of anti-PD-1 monoclonal antibody (10F.9G2, Be0101, BioXcell) or equivalent isotype control antibody (BioXcell, BE0090) was injected intraperitoneally every three days three times.
Detected cell surface ecto-calreticulin (ecto-CTR) expression and total calreticulin (total CRT) expression
The B16-C8KO cells or control cells (B16F10 cells treated with Z-IETD-FMK or dimethyl sulfoxide (DMSO) for 8 h) were collected and washed with PBS with 0.3% goat serum, fixed with 4% formaldehyde with 10% goat serum solution, and incubated with calreticulin antibody at 4 °C for 1 h. After washing three times with FACS buffer, ecto-CRT was detected using flow cytometry. To detect total CRT, the cells were fixed using a Cytofix/Cytoperm Kit (554714, BD). After washing twice with wash buffer, the samples were enclosed in 10% goat serum, incubated with the calreticulin antibody at 4 °C for 30 min, and detected by flow cytometry.
In vivo phagocytosis assay
In vivo phagocytosis was performed in accordance with a previously established protocol [
20]. Briefly, B16F10 cells were stained with 1 µM Cell Tracker Deep Red dye (Invitrogen), following the manufacturer’s protocol, and then treated with 50 µM Z-IETD-FMK or DMSO for 30 min, follow by 25 µM doxorubicin for 24 h. Cells were harvested and adjusted to 5 × 10
7 cells/mL in PBS. Labeled tumor cells (5 × 10
6 in 100 µL PBS) were injected into the spleen. After 2 h, the mice were sacrificed, the spleens were harvested and stained with anti-mouse CD11c antibody, and phagocytosis was assessed by flow cytometric analysis.
Flow cytometry
In subcutaneous animal models, tumors were harvested on days 18–20. After euthanasia, the tumors were collected and filtered through a 70-μm cell strainer to obtain single-cell suspensions. For the analysis of tumor-infiltrating immune cells, the samples were stained with anti-CD45 (30-F11), anti-CD11b (M1/70), anti-CD3 (17A2), anti-CD8 (53-6.7), anti-CD4 (GK1.5), anti-F4/80 (BM8), anti-Gr-1 (RB6-8C5), and Fixable Viability Dye eFluor 780 (65-0865) (eBioscience). For T cell function analysis, samples were cultured with a cell stimulation cocktail (00-4975-03; eBioscience) for 6 h and subsequently stained with anti-CD45 (30-F11), anti-CD3 (17A2), anti-CD8 (53-6.7), anti-IFN-γ (XMG1.2), and anti-GZMB(GB11) using the Cytofix/Cytoperm™ Kit (554714, BD).
In the subcutaneous mouse model, draining lymph nodes were harvested for dendritic cell analysis. After filtering through a 70-μm cell strainer, the single-cell suspensions were stained with anti-CD11c (N418), anti-CD103 (2E7), anti-CD45 (30-F11), anti-MHC-II (M5/114.15.2), and Fixable Viability Dye eFluor 780 (65-0865) (eBioscience). Data were collected using a Gallios flow cytometer (Beckman Coulter) and analyzed using FlowJo software. To detect the cell surface and total calreticulin, cells were treated with the caspase-8 inhibitor Z-IETD-FMK or DMSO and then irradiated using the Varian Trilogy Stereotactic System with a single dose of 20 Gy. After 16 h, the cells were collected for analysis. Staining protocols were performed according to the manufacturer’s instructions, using anti-calreticulin (EPR3924, Abcam).
Data collection
Bulk RNA sequencing data from 31 tumor types and four ICB-treated datasets with clinical information are available in the database of TCGA and under Accession codes: PRJEB23709 [
21], GSE78220 [
22], and GSE91061 [
23]; see Table S8 of the original paper [
24].
RNA sequencing
Five pairs of control or B16-C8KO subcutaneous tumors from the right flank of 6–8-week-old female C57BL/6 mice were collected for RNA sequencing.
To investigate the contribution of caspase-8 to the inflamed TME, the cases were split into CASP8-high or CASP8-low groups based on the expression level of caspase-8 in each tumor type. The 1st and 4th quartiles were defined as CASP8-high and CASP8-low, respectively. Gene rank was generated based on the log2 fold change between the two groups, which was calculated using the R package
DESeq2 [
25]. Then, the normalized enrichment scores and p-values of the inflamed TME gene set [
26] were computed by
fgsea for each tumor type. To confirm the significance, enrichment analysis was applied to the cases with group information using Gene Set Enrichment Analysis (GSEA). A heatmap was plotted based on the group information for each gene in the gene set.
To validate the benefit of caspase-8 for ICB treatment, cases in each ICB-treated cohort were grouped as CASP8-high and CASP8-low based on the median expression level of CASP8. The hazard ratio was calculated by survival and survminer, and forest plots were plotted using forest plots. A heatmap was plotted for each gene of the inflamed TME gene set in ICB-treated datasets. Columns of the heatmap were arranged according to the expression level of CASP8.
To characterize the transcriptome profile of Casp8 knockout and control cells in B16-bearing mice principal component analysis was performed using FactoMineR, and the top two principal components were used for plotting. Enrichment signaling pathways were analyzed using Gene Ontology based on the differentially expressed genes between caspase-8-Casp8 knockout and control mice.
Statistics
Comparisons between two groups of continuous variables were performed using an unpaired t-test or Mann–Whitney U test. Comparisons of continuous variables from three or more groups were performed using one-way analysis of variance (ANOVA). The association between responders and the categorical variables Casp8-high and Casp8-low was compared using the χ2 test or Fisher’s exact test. Tumor growth was compared using one-way ANOVA. Survival was estimated using Kaplan–Meier curves, and the p-value and hazard ratio were determined using a log-rank test. Statistical analyses were performed using Prism 6 software (GraphPad, Prism Software Inc., CA, USA) and R version 4.0.0. Statistical significance was set at p < 0.05.
Discussion
ICD can be induced by different stressors such as chemotherapy, irradiation, and targeted anticancer agents. Anthracycline chemotherapy drugs, such as doxorubicin, induce caspase-dependent ICD by emitting damage-associated molecular patterns [
27]. Many chemotherapeutic drugs, such as cisplatin, can induce casp8 expression and lead to apoptosis, but these drugs are non-specific inducers [
28]. Casp8 is a key regulator of cell death [
15]. As both necrosis and pyroptosis are immunogenic, we could infer that a loss of Casp8 function leads to ICD, which triggers a stronger antitumor immune response and benefits ICB therapy. However, our findings suggest that the role of Casp8 is more complex. The clinical data from TCGA and ICB-treated datasets revealed that in certain cancer types, especially melanoma, Casp8 plays a pro-inflammatory role.
Previous studies have revealed that Casp8 may be related to ICB responsiveness, and Casp8 mutant cells accumulated in tumors with a highly cytotoxic TME [
29]. One explanation is that a loss of Casp8 function leads to resistance to immune cell death. To test this theory, tumor cells were treated with CRISPR and co-cultured with natural killer cells and T cells [
30,
31]. However, caspase-8 knockout was enriched in MC38/MC38-OVA tumors, but not in B16F10 cells, implying that the role of Casp8 may vary among cancer types. In contrast, in MC38-OVA cells co-cultured with OT-I T cells in vitro, treatment with anti-PD-1 failed to enhance tumor eradication [
31].
These results imply that the loss of Casp8 function may not be immunogenic, as expected. In clinical samples, high Casp8 expression was related to better overall survival and cytotoxicity of T cells in cancer patients [
32]. Furthermore, based on the function of this protein, it is reasonable to infer that this phenomenon occurs because of resistance to inducers of cell death in tumor cells. However, in another study, when the CRISPR-treated tumor cells were treated with cytotoxic T cells, Casp8 knockout tumors were not identified. This indicates that resistance might occur ahead of T cell death [
15].
In addition to its role in cell death, our findings showed that Casp8 plays a crucial role in antigen presentation. CRT is a fundamental molecule involved in ICD. When CRT is blocked or knocked down, the immune response is impaired. More precisely, cell-surface-bonded CRT is predominant in ICD. We found that Casp8 knockout downregulated ecto-CRT in our B16F10 model, which is consistent with previous studies [
33]; Casp8 knockout mice also showed a weak antitumor response in the CT26 mouse model. These reports support our finding that Casp8 is a key regulator of ecto-CRT, which affects the response to ICB. However, in the endoplasmic reticulum stress model, ecto-CRT was not Casp8-dependent, and a similar finding was reported in the photodynamic therapy model [
34]. In clinical practice, the detection of Casp8 mutations is practical, implying that such mutations might be independent predictive markers of the response to ICB.
Furthermore, we tried to rescue the resistance to ICB caused by the Casp8 mutation. Irradiation is an important therapeutic method used to overcome low responsiveness to ICB in clinical practice. As far as immunity was concerned, irradiation was thought to have a dual effect, both inhibiting and promoting immunity [
35,
36]. Lamerton verified that if the whole body of the animals was exposed to radiation of 1.76 Gy/day or 0.84 Gy/day, their immune system first responded positively and peripheral blood count increased; however, within 20 days, their bone marrow failed to produce platelets and leukocytes, and their immune system was destroyed [
37], ultimately resulting in death. However, an increasing number of studies have shown that local irradiation might enhance antitumor immunity; for example, 8.5 Gy × 5 irradiation of tumors was reported to enhance MHC class I expression and dendritic cell function and improve the efficacy of tumor immunotherapy [
38,
39]. High LET/RBE irradiation, such as particle and heavy-ion radiation, could induce single- and double-strand DNA breaks [
40]. Meanwhile, in living tissues generating ROS/RNS and H
2O
2, irradiation damages DNA, proteins, and membranes, resulting in new antigen production and strengthening of the immune response. A previous study by our group confirmed that a single local irradiation dose of 20 Gy for tumors could enhance the number of tumor-infiltrating CD8
+ CTLs in the TME of B16F10 tumors, and the depletion of CD8
+ T cells significantly weakened the therapeutic effect of irradiation [
41]. In this study, irradiation combined with ICB improved the ORR and prolonged progression-free and overall survival. In our model, we found that in Casp8-deficient patients, radiation might be an effective approach for overcoming ICB resistance. We explained a new mechanism where stereotactic body radiation therapy enhances ICB therapy by inducing calreticulin expression through caspase-8 inhibition, thus enriching the immunological theory of stereotactic body radiation therapy-enhanced immunotherapy.
Conclusions
Casp8 deficiency, knockout, and inhibitor treatment led to an impaired ecto-calreticulin transition, which in turn resulted in hampered antigen presentation and a cold TME, which are traits associated with ICB resistance. Irradiation could rescue the ecto-calreticulin expression of tumor cells and improve phagocytosis to overcome ICB resistance. Consistent with TCGA and ICB-treated cohorts, Casp8 expression was correlated with an inflamed TME and a better response to ICB. These results imply that patients with Casp8 loss-of-function mutations may not benefit from ICB alone; however, a radiation combination strategy might sensitize non-responders and improve clinical outcomes.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.