CX3CL1 release during immunogenic apoptosis is associated with enhanced anti-tumour immunity

Abstract

Introduction: Immunogenic cell death (ICD) has emerged as a novel option for cancer immunotherapy. The key determinants of ICD encompass antigenicity (the presence of antigens) and adjuvanticity, which involves the release of damage-associated molecular patterns (DAMPs) and various cytokines and chemokines. CX3CL1, also known as neurotactin or fractalkine, is a chemokine involved in cellular signalling and immune cell interactions. CX3CL1 has been denoted as a "find me" signal that stimulates chemotaxis of immune cells towards dying cells, facilitating efferocytosis and antigen presentation. However, in the context of ICD, it is uncertain whether CX3CL1 is an important mediator of the effects of ICD.

Methods: In this study, we investigated the intricate role of CX3CL1 in immunogenic apoptosis induced by mitoxantrone (MTX) in cancer cells. The Luminex xMAP technology was used to quantify murine cytokines, chemokines and growth factors to identify pivotal regulatory cytokines released by murine fibrosarcoma MCA205 and melanoma B16-F10 cells undergoing ICD. Moreover, a murine tumour prophylactic vaccination model was employed to analyse the effect of CX3CL1 on the activation of an adaptive immune response against MCA205 cells undergoing ICD. Furthermore, thorough analysis of the TCGA-SKCM public dataset from 98 melanoma patients revealed the role of CX3CL1 and its receptor CX3CR1 in melanoma patients.

Results: Our findings demonstrate enhanced CX3CL1 release from apoptotic MCA205 and B16-F10 cells (regardless of the cell type) but not if they are undergoing ferroptosis or accidental necrosis. Moreover, the addition of recombinant CX3CL1 to non-immunogenic doses of MTX-treated, apoptotically dying cancer cells in the murine prophylactic tumour vaccination model induced a robust immunogenic response, effectively increasing the survival of the mice. Furthermore, analysis of melanoma patient data revealed enhanced survival rates in individuals exhibiting elevated levels of CD8+ T cells expressing CX3CR1.

Conclusion: These data collectively underscore the importance of the release of CX3CL1 in eliciting an immunogenic response against dying cancer cells and suggest that CX3CL1 may serve as a key switch in conferring immunogenicity to apoptosis.

Keywords: CX3CL1; chemokines; cytokines; fractalkine; immunogenic apoptosis; immunogenic cell death.

Figure 1

CX3CL1 release is associated with immunogenic apoptosis. (A, D) Cell death measured by flow cytometry of MCA205 cells (A) and B16-F10 cells (D) treated with 2 μM or 8 μM MTX (MCA205 and B16-F10 respectively), 2.5 μM RSL3, or three cycles of F/T. Quantification was done by AnV and Sytox Blue staining. The values are the means ± SEM and represent three independent experiments. (B, E) Representative dot plots of the cell death measurement shown in (A, D). (C, F) The concentration (pg/mL) of CX3CL1 measured in the supernatants of dying cells using Luminex xMAP technology. The values are the means ± SEM and represent three independent experiments. Statistical significance was calculated by one-way ANOVA followed by Tukey’s multiple comparisons test: **p < 0.01, ****p < 0.0001. MTX, mitoxantrone; RSL3, RAS-lethal selective 3; F/T, freeze/thaw; AnV, Annexin-V; Sytox, Sytox Blue.

Figure 2

CX3CL1 reverts non-immunogenic apoptosis to ICD. (A) Schematic representation of the tumour prophylactic vaccination mouse model. On day 0, mice were vaccinated in the left flank with either 5 x 10⁵ or 2.5 x 10⁵ MTX-treated MCA205 cells. On day 7, the mice were challenged in the opposite flank with 10⁵ viable cancer cells of the same type and tumour growth was monitored with a digital calliper. (B) Kaplan-Meier curve of the progression of tumour development over time. The reduction of dose of MTX-treated MCA205 cells from 5 x 10⁵ cells to 2.5 x 10⁵ cells significantly decreased tumour-free survival from 70% to 20%. The statistical differences were calculated by a log-rank (Mantel-Cox) test. Survival curves comparison: **p < 0.01. (C) Schematic representation of the tumour prophylactic vaccination mouse model. On day 0, mice were vaccinated in the left flank with either PBS, 2.5 x 10⁵ MTX-treated MCA205 cells alone or 2.5 x 10⁵ MTX-treated cells in combination with different doses of recombinant CX3CL1 (1 ng, 10 ng or 100 ng). On day 7, mice were challenged in the opposite flank with 10⁵ viable cancer cells of the same type and afterwards tumour growth was followed with a digital calliper. (D) Kaplan-Meier curve of the progression of tumour development over time. The addition of 1 ng or 10 ng of rCX3CL1 to MTX-treated MCA205 cells significantly increased tumour-free survival from 10% (MCA205 MTX alone) to 50%. Interestingly, 100 ng of rCX3CL1 had no significant effect on tumour-free survival. The statistical differences were calculated by a log-rank (Mantel-Cox) test. Survival curves comparison: *p < 0.05, **p < 0.01, ***p < 0.001. ICD, immunogenic cell death; MTX, mitoxantrone; PBS, phosphate-buffered saline; rCX3CL1, recombinant CX3CL1.

Figure 3

CX3CR1 is associated with increased CD8⁺ T cells and increased patient survival probability. (A) The association between CX3CR1 expression and overall survival (OS) in the TCGA-SKCM dataset. OS curves were generated by setting median expression as cut-off. (B) Statistical analysis of cell infiltration stratified based on CX3CR1 expression in the TCGA-SKCM dataset. Cellular deconvolution was performed by five algorithms (EPIC, TIMER, quanTIseq, MCP-counter and xCell). The orange square means that the abundance of cells is significantly greater in patients with a high expression level of the receptor than in those with a low expression level. A grey square means that cell proportion does not differ statistically between groups, ns – not significant. A dash means that the method does not determine the proportions of the corresponding cells. Mann-Whitney U test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (C) Boxplots of CD8+ T-cell signature expression stratified according to CX3CR1 expression in the TCGA-SKCM dataset. High (coral) and low (cyan) expression level groups were generated by setting median expression as cut-off. Mann-Whitney U test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns, not significant.