IL-18 derived from microwave-ablated osteosarcoma facilitates antitumor response

Abstract

Background

Osteosarcoma (OSA) is the most common primary malignant bone tumor in children and adolescents, while no further improvement in survival has been achieved in the past few decades. Microwave ablation (MWA) is an emerging thermal therapy to ablate malignant bone tumors that induces antitumor immunity, while intrinsic mechanisms remain elusive. This study uncovered that MWA treatment promoted interleukin (IL)-18 expression and enhanced the dendritic cell (DC) functions, as well as CD8⁺ T cells infiltration. Combining IL-18 and anti-programmed cell death protein 1 (PD-1) displayed the additive tumor suppression effects.

Methods

MWA treated OSA was co-cultured with bone marrow dendritic cells (BMDC) in vitro. RNA-sequencing was performed to analyze the intrinsic mechanism of BMDC after stimulated with MWA treated OSA. Mouse model of primary osteosarcoma of the femur was established for in vivo immune response evaluation, and this model was used to measure the additive tumor eradication of IL-18 and anti-PD-1 combination therapy.

Results

MWA promoted the activation, antigen uptake and antigen cross-presentation of BMDCs. Mouse models demonstrated that MWA treatment could elevate the proportion of CD8⁺ T cells and CD11c⁺DCs both in tumors and draining lymph nodes. The tumor suppression was mediated through the release of IL-18 by MWA-treated OSA and interaction with the IL-18 receptor on DCs, which was reversed after anti-IL-18 antibodies application. Besides, IL-18 cytokine exerted the tumor eradication effect and showed the additive role against OSA when combining anti-PD-1 therapy.

Conclusions

Our study demonstrated a critical role of IL-18-mediated signaling in the induction of protective immune responses against OSA, which provided a novel mechanism to the explanation of MWA-induced immune activation, and IL-18 cytokine treatment could be considered a feasible, effective therapeutic approach for OSA in the clinic.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Microwave ablation (MWA) has been demonstrated to promote immune activation, while the intrinsic mechanisms remain to be fully elucidated. MWA is established as an efficient modality for malignant bone tumor ablation and postoperative functional recovery, its immunomodulatory effects have been less defined. The potential of MWA to stimulate antitumor immunity is often hindered by a lack of understanding regarding its intrinsic molecular mechanisms, thereby restricting its combinatorial clinical application.

WHAT THIS STUDY ADDS

• Our work elucidates the mechanism that MWA enhanced the interleukin (IL)-18 expression of osteosarcoma, which elicits immune activation and assists in anti-programmed cell death protein 1 (PD-1) therapy. Herein, we uncover that MWA induces immune response in osteosarcoma via an IL-18-dependent axis, driving dendritic cell maturation and CD8⁺ T-cell priming. Notably, integrating MWA with anti-PD-1 blockade resulted in additive therapeutic benefits.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The finding of this paper addresses the fundamental gap of mechanism by which MWA induces immune activation, providing a rational basis for treating osteosarcoma either by applying cytokine or combining immune therapy. Our work not only deciphers the immune microenvironment remodeling following thermal ablation but also proposes actionable strategies to augment systemic antitumor immunity, applicable to both standalone and combined therapeutic regimens.

Introduction

Osteosarcoma (OSA) ranks as the most prevalent malignant primary bone tumor mostly affecting children, adolescents, and patients over 60 years of age.1,4 The aggressive nature of OSA is evidenced by a higher incidence of lung metastasis, which is observed in an average of 74% of patients. Additionally, 80% of all confirmed cases of OSA have undetected microscopic metastases, and 20% of patients already present with lung metastases at the time of initial diagnosis.5,7 Neoadjuvant chemotherapy, surgery plus chemotherapy postoperation, has been recommended as standard therapy in OSA according to the NCCN (National Comprehensive Cancer Network) guideline in the 1970s, while there was unfortunately not much improvement in therapeutic outcomes since then.8,10 In addition, classical in situ removal of tumors by surgery and subsequent mechanical reconstruction has been associated with severe complications, such as prosthesis infection and loosening, as well as tumor recurrence.11,14 Thus, in order to augment the surgical outcome and reduce postoperative complications, a series of novel approaches have emerged and were employed, among which thermal therapy has been shown to suppress tumor progression with high efficiency.¹⁵

As one kind of thermal therapy, microwave ablation (MWA) acts by resonating water molecules within cells and thereby inducing heat-mediated cell death based on electromagnetic waves, which comes with several treatment advantages. First, MWA is not hampered by impedance, such as caused by the high tissue density, water and adipose tissues. Second, MWA can overcome the so-called “heat sink effect”, which refers to the phenomenon that tissues with elevated temperatures can transfer their heat towards surrounding tissues of lower temperature, thereby comprising the therapeutic effects of thermal treatment.^{16 17} These advantages make MWA an ideal application to reach target tissues in inner organs, such as the liver, bone and lung. This way, it has previously been demonstrated that MWA treatment followed by surgical resection could achieve tumor ablation, while bones of affected limbs showed improved functional recovery in patients with OSA.1115,19 In general, MWA treatment can improve the clinical outcomes of surgical operations.

In recent years, significant interest has been triggered by observations linking the instantaneous heat-mediated effects of tumor ablation induced by MWA with concomitant immune stimulation that was shown to contribute to the overall clinical effect.^{20 21} During MWA treatments, the core temperature can exceed 50°C, which induces coagulative necrosis with protein denaturation. However, the vicinity of these necrotic areas retains temperatures between 41°C and 46°C, which allows for the active synthesis of immunogenic effector molecules and the initiation of antitumor immune responses.¹⁵ During this process, the exposure of lymphocyte-priming dendritic cells (DCs) to damage or danger-associated molecular patterns, like the high mobility group protein box 1 (HMGB-1) or calreticulin (CRT), and inflammatory cytokines, like interleukin (IL)-1β and IL-6, can induce DC maturation, which in the presence of tumor (neo)antigens that are simultaneously released can prime T-cell responses.22,29 By vaccinating mice with MWA-treated OSA cells in vivo, tumor antigen-specific CD8⁺ T cells were primed and tumors were suppressed through an Fas/FasL-mediated apoptosis pathway.^{22 28} In the context of hepatocellular carcinoma, local MWA treatment was observed to raise systemic interferon-γ (IFN-γ) and IL-5 levels in peripheral blood, which was accompanied by augmented T-cell infiltration into ablated tumors and an overall improved prognosis for the patients.28,31 Nevertheless, the intrinsic mechanisms of how MWA-treated tumors induce immune activation remain elusive. Uncovering these mechanisms, however, could help to further augment immune-mediated effects after MWA treatment and other tumor therapies that could either be implemented alone or in combination with MWA to improve tumor elimination in patients. In this study, we sought to define the immune responses induced by MWA in OSA, and to identify key molecular mediators responsible for its immunomodulatory effects. Using plasma samples from patients with OSA, we first identified a distinct inflammatory cytokine signature after MWA and surgical resection treatment compared with surgical resection alone. Through integrated in vitro and in vivo approaches, we demonstrated that MWA-treated OSA cells enhance DC activation and promote antigen cross-presentation to CD8⁺ T cells. In syngeneic murine models, MWA increased intratumoral and lymph node infiltration of DCs and CD8⁺ T cells. Transcriptomic analysis revealed that MWA-treated K7M2 augmented IL-18R expression on DCs, which ligand IL-18 expression of K7M2 was confirmed to be promoted after MWA treatment. Functional studies using IL-18 blocking antibodies abrogated MWA-induced T cells activation and antitumor efficacy, confirming the pivotal role of IL-18/IL-18R signaling in mediating the ablative immune response. In addition, IL-18 cytokine exhibited an additive effect in OSA treatment when combined with anti-programmed cell death protein 1 (PD-1). Our findings thus elucidate an IL-18-dependent mechanism whereby MWA stimulates antitumor immunity in OSA, providing a rationale for combining MWA with IL-18-based immunotherapies.

Results

MWA treatment induces an enhanced inflammatory cytokine signature and promotes the activation of CD8⁺ T cells in patients

Based on the efficient tumor eradication effect induced by MWA treatment and surgical resection as reported,30,32 it was of great significance to further unveil subsequent immune responses involved in tumor regression. As the induction of immune responses can likely reduce the probability of detecting residual or metastatic tumor cells after treatment, we performed multicytokine detection assays with plasma samples of patients at day 7 after MWA to test whether MWA treatment can enhance systemic inflammation compared with patients experiencing surgical resection alone. Plasmas from patients before any surgical performance were set as baseline analysis. The results of this analysis revealed enhanced inflammation in the MWA group (figure 1A). Notably, excluding IL-6, IL-10 and tumor necrosis factor (TNF)-β which are deemed to promote an immunosuppressive tumor microenvironment through recruiting immune inhibitory cells, such as regulatory T cells (Treg),33,38 the remaining cytokine expressions were elevated in the MWA group, suggesting the potential for enhanced inflammation and responses by cytotoxic CD8⁺ T cells.^{39 40} Conversely, only surgical resection (No MWA) treatment displayed the enhanced systemic expression of IL-6 and IL-10, implying the immune suppressive effects of surgical trauma,39,41 and this immune inhibitory response was reversed when MWA treatment was added (figure 1B). Together, MWA treatment followed by surgical resection still promoted a peripheral inflammation enhancement with immune inhibitory cytokines suppression, and the induction of a systemic inflammatory signature could be an important component for the improved elimination of tumors after MWA treatment.

Figure 1. MWA treatment induces an inflammatory cytokine signature, and promotes the activation of CD8⁺ T cells in patients. Heatmap depicting cytokines expression (Z-score standardization, consisting of IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17A, IL-17F, IL-22, TNF-α, TNF-β and IFN-γ) changes of plasma from MWA and surgical resection treated OS patients (MWA, n=23), only surgical resections treated OS patients (No MWA, n=14) compared with patients before surgical operation (preoperation, n=15). (B) Cytokines concentration measurement of IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17A, IL-17F, IL-22, TNF-α, TNF-β and IFN-γ of plasma from patients. (C) Flow cytometry detection of CD8⁺CD25⁺ T cells within total CD8⁺ T cells (left), CD8⁺CD44⁺ T cells within total CD8⁺ T cells (second from left), ratio of CD8⁺CD25⁺ T cells/CD4⁺CD25⁺ T cells (second from right), and ratio of CD8⁺CD44⁺ T cells/CD4⁺CD25⁺Foxp3⁺ T cells (right) of total PBMC from MWA and surgical resection treated OS patients (n=23) compared with surgical resections (n=14). Each circle represents one value of cell percentages/ratio. Lines depict mean±SD. Unpaired t-test was performed for statistical tests. *p<0.05 and **p<0.01. IFN, interferon; IL, interleukin; MWA, microwave ablation; OS, osteosarcoma; PBMC, peripheral blood mononuclear cell; TNF, tumor necrosis factor.

Next, we were interested to test whether MWA treatment induced systemic changes in the composition of major immune cell subsets in the T-cell compartment. Therefore, we analyzed relative changes in the presence of activated (CD8⁺CD25⁺, CD8⁺CD44⁺) and regulatory (CD4⁺CD25⁺Foxp3⁺) T cells within peripheral blood mononuclear cells 7 days post-surgery by flow cytometry (figure 1C, online supplemental figure S1). These experiments revealed that there was an increase in the proportion of activated CD8⁺ T cells in patients who had undergone MWA treatment and resection, when compared with patients in the control group that received surgical resection alone. In contrast, no significant changes were obtained in activated CD4⁺ T cells (CD4⁺CD25⁺) and Tregs (CD4⁺CD25⁺Foxp3⁺) (figure 1C, online supplemental figure S2). The higher ratio of activated CD8⁺ T cells to Tregs observed after MWA treatment might indicate enhanced tumor antigen-specific CD8⁺ T-cell formation that could support the response against residential tumor cells under this condition (figure 1C). These alterations in inflammatory cytokine production and immune cell composition observed in the MWA and surgical resection treated group pointed towards a beneficial phenotype promoting tumor eradication.

MWA-treated OS cells display characteristics of apoptosis

As one type of thermal therapy, MWA treatment has been demonstrated to induce cellular stress in tumor cells, such as heat shock and reactive oxide species to induce cell death.^{42 43} In order to monitor temperature changes elicited by MWA treatment in cell culture media, an infrared thermometer was employed. The power of the MWA device was set to 5 W, and temperature was increasing with time as expected. Specifically, MWA was found to reach maximal temperatures of approximately 39.2°C after 1 min, 46.6°C after 2 min, 57.6°C after 3 min, 68.7°C after 4 min and 79.6°C after 5 min of treatment, as indicated by infrared imaging (figure 2A, online supplemental figure S3A). It has previously been shown that tumor cell-secreted cytokines and tumor-associated antigens can be damaged at very high temperatures.⁴⁴ In contrast, more moderate temperature induction in tumor cells by MWA treatment promoted intratumoral CD8⁺ T-cell infiltration and systemic immune activation.^{43 45 46} To preserve immune-active molecules and to thereby maximize the immune-activating effects of MWA treatment, 2 min with 5 W seemed to be the optimal MWA treatment condition in vitro, according to our temperature monitoring.

Figure 2. MWA treatment of K7M2 cells induces apoptosis and necrosis in a time-dependent manner. (A) Infrared thermometer imaging of MWA treatment-induced temperature changes in culture media after various treatment durations. (B) Cell viability of K7M2 cells at days 1, 3 and 7 after MWA treatment as measured by cell counting kit 8 assay (n=3 per group from three independent experiments). (C) Annexin V and PI staining were detected by flow cytometry to analyze the apoptotic and necrotic status of K7M2 cells at 24 hours after MWA treatment or indicated durations (n=3 per group from four independent experiments). (D) Transmission electron microscopy was applied to depict subcellular changes of K7M2 cells 24 hours after MWA treatment for indicated durations. A Turkey’s multiple comparisons test was performed for statistical tests (B). Bar graphs depict mean±SD. ****p<0.0001 (B). OS, osteosarcoma; MWA, microwave ablation; PI, propidium iodide.

Next, we further investigated the relationship between cell viability after MWA treatment and the corresponding immune-stimulatory capacity of treated tumor cells. Compared with the untreated murine OS (Osteosarcoma) cell line K7M2, MWA treatment for 1 min did not significantly alter cell viability as measured 24 hours after MWA exposure, while cell viability reduced to almost 50% after 2 min of treatment, as determined by cell counting kit 8 (CCK-8) assay that is based on the detection of formazan formation (figure 2B). Cell viability was further reduced by approximately 75% after 3 min of MWA treatment, which further reduced after 4 and 5 min. Additionally, detection of apoptosis by flow cytometry demonstrated that MWA treatment primarily induced apoptosis in cells after 2 min of exposure, as evident by the majority of cells exhibiting an Annexin V-positive and propidium iodide (PI)-negative profile 24 hours post MWA treatment (figure 2C). With the extension of MWA treatment time, a transition to predominantly necrotic cell death was observed, as characterized by an increase in the proportion of Annexin V-positive and PI-positive cells (figure 2C, online supplemental figure S3B). Transmission electron microscopy (TEM) was performed to reveal subcellular changes of organelles in MWA-treated cells. Cells receiving 1 min of treatment did not display specific differences compared with untreated cells. However, this observation changed after 2 min of treatment, when highly pyknotic and fragmented nuclei, aggregated mitochondria (light blue arrows) and budding apoptotic bodies (light yellow arrows) became visible, indicating the induction of apoptosis. Consistent with our flow cytometry staining (figure 2C), longer MWA treatment periods of 3, 4, and 5 min led to cellular images displaying necrotic characteristics, including condensed and fragmented nuclei, as well as coagulated and lytic cytoplasm and organelles (figure 2D). Collectively, these observations suggested that MWA treatment predominantly induced apoptosis in tumor cells within the initial 2 min of exposure under the conditions studied in these in vitro experiments.

MWA-treated OS cells promoted BMDC activation, antigen uptake, and cross-presentation

After observing that MWA treatment for 2 min could induce highest frequencies of apoptosis in K7M2 cells, we aimed to further explore the immunogenicity of this form of cell death. Therefore, a co-cultured system was established in which we incubated murine bone marrow dendritic cell (BMDC) with either untreated or MWA-exposed K7M2 cells, with the aim to study differential effects on their activation and capacity to endocytose and cross-present antigens (figure 3A). We defined differentiated BMDC without any treatment as “blank” group, and those cultured with untreated K7M2 cells as “control” group. First, we used expression of the co-stimulatory molecules CD80 and CD86 as a measure of BMDC activation (measured as per cent double-positive cells) and observed elevated levels in response to K7M2 cells that were exposed to MWA (2 min), compared with BMDC incubated with untreated control cells (figure 3B,C, online supplemental figure S4). In line with this, quantification of mean fluorescence intensities detected for CD80 and CD86 on BMDC also demonstrated a significant induction of co-stimulatory molecule expression after incubation with MWA-treated cancer cells (figure 3D,E). For both experiments, treatment with the toll-like receptor 4 agonist lipopolysaccharide was used as a positive control for BMDC activation. From cell culture supernatants of the same experiments, we performed measurements of IL-6, IL-10 and IL-12p40 that are secreted on BMDC activation. These measurements indicated enhanced levels of these molecules in samples containing MWA-treated K7M2 cells, further indicating enhanced BMDC activation under this condition (online supplemental figure S5). Moreover, to elucidate whether MWA treatment promoted BMDC activation was limited to K7M2 cells, we incorporated another osteosarcoma cell LM8, melanoma cell B16h10 and colorectal cell MC38 (B16h10 and MC38 are treated as tumor cells with high immunogenicity) to evaluate the activation of BMDC in the same co-culture system. Conformably, there is no significant difference between ‘Blank’ and ‘Control-LM8’ groups in activated BMDCs proportions, while remarkable enhancement was achieved in ‘MWA-LM8’ group with CD80 and CD86 double positive BMDCs proportion up to 73.2% (online supplemental figure S6A,B). Synchronously, MFI analysis of CD80 and CD86 also demonstrate the same trends with K7M2 (online supplemental figure S6C,D). As we have studied the ‘cold’ OSA tumor cell lines in two models, the effects of BMDCs activated by MWA treated ‘hot’ tumor cell lines are also measured through B16 melanoma or MC38 colorectal cell lines. Being different from the results of OSA models, both B16 and MC38 without MWA treatment display a certain degree of enhancing BMDCs activation, which may be due to the internal relatively higher immunogenicity of B16 and MC38 (online supplemental figure S7A). After MWA treatment, both B16 and MC38 cell lines exhibit promoted BMDC activation effects, which are consistent with the result of osteosarcoma models (online supplemental figure S7B). Furthermore, MFI analysis of CD80 and CD86 also clarify the same trends with osteosarcoma models (online supplemental figure S7C,D). Generally, MWA treatment can enhance the immunogenicity of tumor cells whether it is a ‘cold’ tumor or a ‘hot’ tumor, and promote the activation of DC.

Figure 3. MWA treatment of K7M2 cells promotes BMDC activation, antigen uptake and cross-presentation after co-culture in vitro. (A) Diagram of co-culture model in which MWA-treated or control K7M2 cells were incubated with BMDC to study effects on their activated and function. (B) The activated BMDC proportions after co-cultured with K7M2 or MWA treated K7M2 for 12 hours were detected by flow cytometry staining CD80 and CD86 (n=5). (C, D) Histogram graphs of CD80 and CD86 expression measurement in BMDC after co-cultured with K7M2 or MWA treated K7M2 for 12 hours. MFI measurements of CD80 and CD86 expressed on BMDC after co-culture (n=5). (E) Representative FACS plots of the frequency of FITC labeled OVA proteins uptake by BMDC. BMDC were treated with K7M2 or MWA treated K7M2 for 12 hours, and treated BMDC were incubated with FITC labeled OVA proteins for 30 min before detection was performed by FACS. (F) Proportions of BMDC uptaking FITC labeled OVA proteins after different treatments (G). MFI quantification of FITC-OVA uptake by BMDC after incubated under indicated conditions (n=5). (H) The histogram graphs of H-2K^b SIINFEKL positive BMDC after cross-presentation. BMDC were treated with K7M2 or MWA treated K7M2 for 12 hours, and 100 µg/mL soluble protein OVA were added for another 20 hours. FACS was performed to analyze the percentage and MFI of BMDC expressing H-2K^b SIINFEKL (n=5). One-way ANOVA multiple comparison tests were performed for statistical analysis of (C, E, F, H, I and K). Lines in bar graphs depict mean±SE with *p<0.05, **p<0.01 and ****p<0.0001. ANOVA, analysis of variance; BMDC, bone marrow dendritic cell; FACS, Fluorescence-Activated Cell Sorter; FITC, Fluorescein 5(6)-isothiocyanate; LPS, lipopolysaccharide; MFI, mean fluorescence intensity; MWA, microwave ablation; OVA, ovalbumin.

As BMDC are known as professional antigen presenting cells that can perform antigen cross-presentation and priming of T cells,^{47 48} we were interested to investigate whether MWA treatment of antigen donor cells could influence the efficiency of related pathways. First, FITC (Fluorescein 5(6)-isothiocyanate)-labeled chicken ovalbumin (OVA) was therefore added to the previously described co-cultured system, which allowed the quantification of antigen uptake in the presence or absence of MWA-treated cancer cells. We defined differentiated BMDC treated with OVA only as the “OVA” group. For these experiments, BMDC were co-cultured with either untreated or MWA-exposed (2 min) K7M2 cells for 12 hours, before 10 µg/mL FITC-OVA was added into the culture media at 37°C for 30 min. Next, the presence of FITC-OVA positive cells within the CD11c and major histocompatibility complex (MHC)-II double-positive BMDC population was quantified as a measure of antigen uptake efficiency (figure 3G). While untreated control cells rather dampened the endocytic ability of BMDC, compare to BMDC that were cultured in the absence of cancer cells, co-culture with MWA-exposed K7M2 cells seemed to rescue the endocytic ability of BMDC and remain at a similar level, as evident by a comparable percentage of FITC-OVA-positive BMDC (figure 3H). In addition, the amount of OVA endocytosis per cell was highest in BMDC incubated with MWA-treated K7M2 cells, as measured by FITC mean fluorescence intensity (MFI), while incubation with untreated K7M2 cells seemed to again suppress antigen uptake (figure 3I, online supplemental figure S8). Although MWA-exposed K7M2 cells could not promote more BMDC to uptake antigen in our co-culture system, the ability of antigen uptake of each BMDC appeared elevated, which might contribute to the enhancement of antitumor immune effects.

In addition to antigen uptake, efficient antigen processing and peptide loading onto MHC class I by DCs for cross-presenting on the cell surface is an essential procedure to prime CD8-positive T cell responses against cancer.⁴⁹ The OSA cell line K7M2 used in our study was originally derived from BALB/c mice, which express the MHC-I haplotypes H-2K^d, H-2D^d, H-2L^d,⁵⁰ and no cancer-specific peptides restricted to these haplotypes have been identified at present to study antigen cross-presentation. Consequently, we transduced K7M2 cells with OVA (abbreviated as K7M2_OVA) to allow the antigen cross-presentation measurement accessible. Through the co-culture system used above containing gene-edited K7M2 cells and BMDC differentiated from C57BL/6 mice bone marrow stem cells (abbreviated as C57-BMDC) whose MHC-I haplotype is primarily H-2K^b, we could detect the SIINFEKL H-2K^b expression of C57-BMDC to analyze the antigen cross-presentation regulation. The group “BMDC” was referred to C57-BMDC without any treatment. The “Blank” group was set as C57-BMDC incubated with K7M2 without gene edition or MWA treatment. K7M2_OVA without MWA treatment displayed moderate antigen cross-presentation efficacy, which might be due to the immunosuppressive effects of K7M2. The improved function was obtained when C57-BMDC was treated with MWA-exposed K7M2, according to the MFI measurement of SIINFEKL H-2K^b (figure 3H,I). Obviously, MWA-exposed K7M2 cells could improve the antigen processing or cross-presentation abilities of BMDC, and this advantage might assist in overcoming the intratumor BMDC suppressive functions.

Taken together, we characterized MWA-exposed K7M2 cells could induce the maturation of BMDC and enhanced their capacity to perform antigen uptake by endocytosis as well as antigen cross-presentation ability.

RNA sequencing revealed the induction of diverse transcriptional modules in BMDC stimulated with MWA-treated K7M2 cells

After our finding that MWA-treated OS cells could enhance BMDC activation and function, we further performed RNA sequencing (RNA-seq) to uncover specific gene regulation patterns in BMDC after co-incubation with MWA-treated K7M2 cells. For these experiments, we isolated BMDC from untreated cultures (DC), or co-cultures with either control (DC-K) or MWA-treated K7M2 cells (DC-MK) using a magnetic cell sorting approach with CD11c-specific beads to investigate BMDC responses at the transcriptional level (figure 4A). To confirm the purity of isolated BMDC, we tested for OS-related gene expression (Pten, Egfr, Rb1, Runx2, Myc and Trp53)51,53 and tumor cell malignancy-associated genes (Ccne1, Cdkn2a, Vegfa and Sp7)^{54 55} in the DC-K and DC-MK groups compared with the DC group in our RNA-seq data to exclude potential contamination with K7M2 cells. There was no significant upregulation for the majority of these genes when compared with the DC group. There seemed to be an elevated expression of Sox9 in DC-K and DC-MK groups, but the actual Fpkm values were extremely low and could be treated as non-biological significance (figure 4B). Thus, we concluded that there was no tumor cells contamination in the isolated BMDC samples. Principal component analysis demonstrated that BMDC stimulated by MWA-treated K7M2 cells (DC-MK) were clearly separated from both K7M2 cells treated (DC-K) and control groups (DC) (figure 4C). The distinctly separated cluster of DC-MK indicated that MWA-treated K7M2 cells induced significant gene expression changes in BMDC in co-culture models.

Figure 4. BMDC co-cultured with MWA-treated K7M2 cells showed distinct transcriptional changes linked to an inflammatory gene signature. (A) Diagram depicting the co-culture of BMDC with either MWA-exposed or naive K7M2 cells, isolation of CD11c⁺ dendritic cells, and the subsequent RNA sequencing to identify transcriptional patterns induced by MWA treatment. (B) Heat map analysis of genes expression relative to tumor cell proliferation and mutation. (C) PCA plot showing the variable genes on rlog transformed counts. (D) Volcano plot showing upregulated and downregulated genes when comparing DC-K and DC-MK. (E) Top 10 significant pathways enriched according to a KEGG analysis when comparing transcriptional changes in DC-MK with DC-K. (F, G) Heat map analyses of cytokines and cytokines receptors with significant expression levels in MWA-exposed K7M2 cell-treated BMDC compared to BMDC incubated with untreated control cells. (H) Transcriptional quantification of the co-stimulatory molecules Cd80, Cd86 and Cd40 in BMDC. *p<0.05, **p<0.01 and ****p<0.0001. BMDC, bone marrow dendritic cell; DC, dendritic cell; KEGG, Kyoto Encyclopedia of Genes and Genomes; MWA, microwave ablation; PCA, principal component analysis; RNA-seq, RNA sequencing.

Next, we aimed to investigate whether MWA treatment of M7M2 cells led to the induction of specific genes associated with beneficial immune responses in cancer. First, we generated a volcano plot indicating the number of genes that are differently expressed between DC-MK and DC-K (figure 4D). Next, we used this dataset to perform a Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis, which revealed enrichment of the gene set terms “cell adhesion molecules” and “axon guidance”, which related to migration and antigen capture, as well as “cytokines – cytokine receptor interactions” in DC-MK (figure 4E). We speculated that the upregulated cell adhesion molecules and axon guidance pathway might indicate the enhanced residence of DCs locally that is needed for prolonged retention through adhesion molecules, and accelerating antigen capture abilities through “tendrils”. Besides, the enrichment of the gene set term “cytokine-cytokine receptor interaction” might infer that altered functional regulation of BMDC incubated with MWA-treated K7M2 cells might be a consequence of downstream signaling events after receptor binding with ligands originating from treated tumor cells.

As it has previously been shown that patients with MWA treatment and surgical resection displayed an enhanced inflammatory cytokine signature in plasma samples compared with those without MWA treatment,⁵⁶ we next filtered our dataset for the top 15 differentially expressed genes encoding cytokines and cytokine receptors in BMDC in order to identify major mechanisms of immune regulation induced by MWA-treated tumor cells (figure 4F,G). Gene expressions of the co-stimulatory molecules CD80, CD86 and CD40 were highest in DC-MK samples, compared with the DC-K and DC condition (figure 4H), which was consistent with the results of our earlier flow cytometric analysis of BMDC activation (figure 3B–D).

With the aim to identify critical mediators of BMDC activation that were secreted by MWA-treated K7M2 cells, we studied the known function of our top 15 cytokines and cytokine receptors to identify important ligands potentially produced by MWA-exposed K7M2 cells to elicit antitumor immunity. Seven cytokine receptors including IL-27Ra,57,59 IL-17Rb,^{60 61} IL-12Rb2,⁶² IL-2Ra,⁶³ IL-2Rb,⁶³ IL-21R,⁶⁴ CCR-7^{65 66} and CSF2Rb⁶⁷ were listed in online supplemental table S1 as well as their corresponding cytokines, which were not produced mainly by tumor cells and excluded for further analysis. In addition, another four cytokine receptors including IL-1R2,^{68 69} CXCR-2,⁷⁰ CXCR-4⁷¹ and CXCR-5⁷² were also excluded due to the suppressive tumor microenvironment induction (online supplemental table S2).

The ligand of IL-18R1 is IL-18, which is produced by T cells, natural killer (NK) cells, DCs, Kupffer cells, articular chondrocytes, osteoblasts and synovial fibroblasts.^{46 73 74} In addition, we had measured the exact expressions of CXCL-9, CXCL-10 and CXCL-11, whose co-receptor is CXCR-3, from the supernatant of untreated and MWA-treated K7M2 cells. In addition, we also measured the expression of CXCL-16 (corresponding receptor is CXCR-6). All these four chemokine expression measurements demonstrated that there was no obvious significance or just a modest alteration in MWA-treated K7M2 cells compared with untreated controls, which indicated that CXCR-3 and CXCR-6 might not be appropriate candidates for further studies (online supplemental figure S9). In addition, we excluded granulocyte-macrophage colony-stimulating factor (GM-CSF) as an important player in our system, as there was no significant difference in actual expressions in either MWA-exposed or control K7M2 cells (online supplemental figure S9). In all, we decided to select IL-18 as the promising target cytokine and further test if IL-18 was actually secreted by MWA-treated K7M2 cells to explore its function in regulating antitumor immune response in our model.

MWA treatment enhanced IL-18 expression by K7M2 cells both in vitro and in vivo

Our RNA-seq analysis indicated that the transcriptional level of IL-18R1 in BMDC was elevated after incubation with MWA-treated K7M2 cells. This prompted further investigations into whether expression of the corresponding ligand IL-18 could be enhanced in MWA-exposed K7M2 cells compared with untreated controls. It has previously been established that IL-18 is produced by several types of cells, can promote effector functions of NK and T cells, and enhance the cytotoxic functions of CAR-T (Chimeric Antigen Receptor-T) cells.4673,75 Based hereon, we hypothesized that IL-18 could function as a catalyst for functional BMDC activation, and we therefore measured the expression of IL-18 on both the transcriptional and translational level in either MWA-exposed or untreated K7M2 cells to confirm its biological significance (figure 5A).

Figure 5. MWA-treatment promoted the IL-18 expressions. (A) Schematic diagram illustrating how IL-18 released from MWA-treated K7M2 cells might interact with the IL-18R on DCs derived from either tumor tissue or draining lymph nodes. (B) Analysis of IL-18 transcription in either control K7M2 or MWA-treated cells 24 hours after exposure as measured by real-time PCR (n=9). (C, D) Evaluation of IL-18 protein expressions in either untreated K7M2 controls or MWA-exposed cells 24 hours after treatment by western blot (n=4). (E) Detection of IL-18 release from K7M2 control or MWA-treated cells 6 and 24 hours after MWA exposure through an ELISA assay (n=9). (F) Representative image of H&E, p53 (green) and IL-18 (red) expression detection through immunofluorescent staining in control and MWA treatment groups. (G) Statistical analysis of p53 and IL-18 expression by MFI measurement (n=5). (H) Co-localization analysis between intratumor p53 and IL-18 expressions. (I) Measurement of IL-18 concentration from patients’ plasma at day 7 after operations between different groups. (J) Paired comparison of IL-18 concentration from “No MWA” (only resection) treated patients’ plasma before operations and at day 7 after operations (n=13). (K) Paired comparison of IL-18 concentration from “MWA” treated patients’ plasma at day 7 before and after operations (n=12). An unpaired t-test was performed to analyze for significant difference in (B+D+I). Two-way ANOVA multiple comparison tests were performed for statistical analysis in (E+G). Simple linear regression was used for correlation analysis. A paired t-test was performed to analyze for significant difference in (J+K). Lines in bar graphs depict mean±SEM with *p<0.05, **p<0.01, ***p<0.001 and ****p<0.0001. ANOVA, analysis of variance; DC, dendritic cell; IL, interleukin; MFI, mean fluorescence intensity; MWA, microwave ablation; OS, osteosarcoma.

Therefore, we measured RNA expression of IL-18 by real-time PCR (figure 5B) and total intracellular protein levels by western blotting (figure 5C,D). These experiments revealed enhanced IL-18 expression by MWA-treated K7M2 cells on both the transcriptional and translational level 24 hours after MWA exposure. In line with this, MWA-treated K7M2 cells released more IL-18 into the surrounding medium compared with untreated controls, as measured by an IL-18-specific ELISA (figure 5E). In addition to our in vitro findings, when investigating IL-18 expression in MWA-treated OSAS in vivo by confocal microscopy staining, we observed increased IL-18 signals in MWA-exposed tumor tissue. To identify whether IL-18 originated dominantly from tumor cells, we performed the co-staining immunofluorescence assay including IL-18 staining and OSA tumor cells identified through p53 labeling (the mostly mutative gene across tumors^{76 77}). As it is shown in the figure below, cells with p53 positive staining mostly point to the highly viable OSA tumor cells. Low expression of IL-18 was detected in the control group accompanied by high p53 positive staining. However, MWA treatment promoted the intratumor IL-18 expression while potentiated remarkable p53 suppression (figure 5F,G). Within the OSA tumor tissues, the distribution of IL-18 exhibited a positive relation with p53, demonstrating the high degree of co-localization between IL-18 and OSA tumor cells after MWA treatment (figure 5H). This result concluded that MWA induced IL-18 enhanced expression appeared to originate from OSA tumor cells. This indicated MWA-treatment of osteosarcoma in our mouse model could induce IL-18 expressions. We collected plasma samples before operations and at day 7 after operations from 25 patients divided into “No MWA” (only surgical resection) and “MWA” (MWA and surgical resection) treated groups, and detected the IL-18 concentrations. Two groups were comparable after our analysis of patients baseline characteristics, including age, gender, malignancy, tumor location and IL-18 concentration before operations (online supplemental figure S10). Strikingly, “MWA” treated patients achieved elevated IL-18 levels when compared with the “No MWA” treated patients (figure 5I). Paired samples analysis from the same patients before operations and at day 7 after operations also demonstrated the IL-18 boosting effects of MWA (figure 5J,K), which confirmed the previous results in vitro and mouse model.

Collectively, these findings indicated that MWA-treatment stimulated IL-18 production both in vitro and in vivo, and this was also confirmed through clinical samples measurement, suggesting that IL-18 might contribute to the immune-regulatory properties after MWA treatment of malignant bone tumors.

Neutralization of IL-18 dampened the tumor-suppressing effects of MWA treatment

Our discovery that MWA-exposed K7M2 cells released increasing amounts of IL-18 compared with untreated controls triggered us to study whether IL-18 release from MWA-treated cancer cells contributed to the beneficial therapeutic effects that we observed in vivo. We therefore performed a series of OSA experiments in mice, in which we combined MWA treatment with depletion of IL-18 using neutralizing antibodies (figure 6A). Therefore, the femur in situ OSA mouse model was established. Briefly, the lateral condyle of femurs were drilled to have round defects/lesions of 1.0 mm diameter and 2.0 mm depth without penetrating the cortical bone on the opposite side. K7M2 cells suspended in 10 µL matrigel were orthotopically injected into the bone lesions on day 14 before MWA performance. One group received MWA treatment on day 0, while the other was left untreated as control. On the same day, both groups were again subdivided into one part receiving an IL-18 neutralizing antibody and one that was treated with an isotype control every 3 days. The results of this experiment indicated that anti-IL-18 treatment could completely revert the MWA treatment-induced suppression of tumor growth (figure 6B, online supplemental figure S11).

Figure 6. Blocking the IL-18/IL-18R axis reversed MWA treatment-induced tumor suppression and promoted antitumor immunity in vivo. (A) Schematic diagram of MWA exposure in combination with neutralizing anti-IL-18 treatment in a mouse model of osteosarcoma in vivo. (B) Tumor volume development after MWA treatment in the presence or absence of an IL-18 neutralizing antibody, using injection of an isotype as control (n=5). (C) Kaplan-Meier survival curve of mice after indicated treatments (n=8). (D, E) MWA-treated K7M2 cells enhanced intratumoral CD8⁺ T cells infiltration, as detected by flow cytometry after gating on all CD45-positive events (n=5). (F, G) Expression measurements of the co-stimulatory molecules CD80, CD86 and IL-18R on DCs isolated from tumor-draining lymph nodes after indicated treatments throughflow cytometry (n=5). (H, I) Detection of intratumoral CD11c⁺ cells infiltration and IL-18 expression after indicated treatments (n=5). For statistical analysis, a Kaplan-Meier analysis (C) was performed to compare survival difference among different treatment conditions. A one-way ANOVA test was performed to analyze significant difference in (E, G, I). Lines in bar graphs depict mean±SEM with *p<0.05, **p<0.01, ***p<0.001 and ****p<0.0001. ANOVA, analysis of variance; DC, dendritic cell; IL, interleukin; i.p., intraperitoneal; MFI, mean fluorescence identity; MHC, major histocompatibility complex; MWA, microwave ablation; mAb, monoclonal antibody; OS, osteosarcoma.

In line with this, IL-18 neutralization prevented the beneficial effects of MWA treatment on overall survival, indicating its crucial role as a mediator of anticancer immune responses in this model (figure 6C). Additionally, we aimed to study the effect of IL-18 depletion on immune cell recruitment and composition in MWA-treated tumors. By performing flow cytometry staining of ex vivo tumor material, we observed that MWA treatment of tumors led to a significant increase in the number of tumor-infiltrating CD8⁺ T cells, while CD4⁺ T cells infiltration remained unaltered (figure 6D,E). Importantly, while IL-18 neutralizing antibodies did not suppress intratumoral infiltration of CD8⁺ T cells in the absence of MWA treatment, CD8⁺ T-cell levels that were elevated after MWA treatment were reduced back to control levels after depletion of IL-18 with neutralizing antibodies. Additionally, the expression of co-stimulatory molecules CD80 and CD86 by CD11c⁺ DCs isolated from tumor-draining lymph nodes was highest after MWA treatment, but strongly dependent on the presence of functional IL-18 (figure 6F,G, online supplemental figure S12). At the same time, MWA treatment induced the expression of IL-18R on DCs, which however did not depend on the presence of IL-18, as treatment with neutralizing antibodies did not change this phenotype. This indicated that IL-18R expression by DCs was not the result of a positive feedback loop induced by IL-18, implying that other pathways triggered by MWA treatment might also induce IL-18R expression in DCs (figure 6G). Furthermore, our experiments indicated that untreated OS tumors displayed a “cold” immune status with a low number of CD8⁺ T cells and CD11c⁺ DCs, which could be converted into a “hot” immune phenotype after MWA treatment, as indicated by the detection of enhanced numbers of these cells by confocal imaging of tumor tissues (online supplemental figure S13). Furthermore, we performed immune fluorescence imaging of tumor tissue sections to analyze the expression of IL-18 and the DC marker CD11c. Here, we found that IL-18 was elevated in MWA-treated cancer samples, concomitant with increased CD11c⁺ DCs infiltration, suggesting an advantageous role of IL-18 in recruiting DC (figure 6H). While administration of IL-18-neutralizing antibodies did not prevent the detection of IL-18 by immunofluorescence microscopy, concomitant CD11c⁺ signals were reduced, indicating decreased DC infiltration into the tumors (figure 6H,I).

In addition, MWA-treatment could inhibit the infiltration of immuno-suppressive Tregs (CD4⁺ Foxp3⁺) and accelerate numbers of early activated CD8⁺ T cells (CD8⁺CD69⁺), a phenotype that was reversed through application of anti-IL-18 neutralized antibody treatment (online supplemental figures S14 and S15). Together with the enhanced co-stimulation of DC (CD80⁺CD86⁺) as described in our in vitro studies, these findings indicate that MWA treatment might have the potential to support in situ priming of DC and T-cell responses to boost systematic immune responses against OSA.

IL-18 cytokine therapy exhibited remarkable tumor suppression and additive therapeutic effects when combined with anti-PD-1 in mouse model

Given that MWA promoted the IL-18 production in experimental models and clinical patient plasma, we speculated that whether IL-18 could exhibit tumor suppression effect in OSA treatment, and employed an OSA mouse model for further analysis. Since OSA has been reported to respond poorly to anti-PD-1 monoclonal antibodies therapy,78,81 we also explored the potential role of IL-18 in reversing anti-PD-1 therapy resistance. Only anti-PD-1 therapy showed weak tumor suppression, which was assistant with the reported patients’ studies. Strikingly, only IL-18 cytokine treatment could achieve satisfactory tumor eradication compared with control and anti-PD-1 therapy. When combining IL-18 cytokine and anti-PD-1 treatment, we observed the additive tumor suppressive effects, demonstrating that additional IL-18 cytokine could assist in overcoming tumor resistance to anti-PD-1 therapy (figure 7A–E). More intratumor CD8⁺ T cells infiltration was detected using immunofluorescence both in IL-18 cytokine and IL-18 cytokine and anti-PD-1

Figure 7. IL-18 could be a reliable prognosis factor in patients with osteosarcoma and suppressed tumor growth in mouse models. (A, B, C, D) Tumor volume development after application of anti-PD-1, IL-18 cytokine and combining IL-18 cytokine and anti-PD-1 (n=4). (E) Tumor growth curve of mice after indicated treatments (n=4). (F) Immunofluorescence staining of CD8 in tumor tissues of different treatments. (G) Quantification of CD8 MFI in different groups (n=4). For statistical analysis a one-way ANOVA test was performed to compare survival difference among different treatment conditions (E). A one-way ANOVA test was performed to analyze significant difference in (G). Lines in bar graphs depict mean±SEM with ****p<0.0001. ANOVA, analysis of variance; IL, interleukin; MFI, mean fluorescence identity; PD-1, programmed cell death protein 1.

combination therapy, while both control and anti-PD-1 groups displayed the CD8⁺ T cells “desert” signature (figure 7F,G). It was suggested that enhanced CD8⁺ T cells infiltration might be the intrinsic mechanism to assist with anti-PD-1 therapy.

Discussion

In this study, we highlighted the beneficial role of MWA treatment in OSA tumor suppression and related immune responses. MWA-exposed K7M2 OSA cells showed enhanced apoptosis and improved immunogenicity. In addition, we uncovered that MWA treatment enhanced the IL-18 expression of K7M2 cells and this enhanced release promoted the infiltration of CD8⁺ T cells and CD11c⁺ DCs into both the tumors and lymph nodes. This immune stimulating effect was reversed by the addition of IL-18 neutralizing antibodies, which reverted the tumor suppression effect of MWA treatment. Together, we uncovered a novel mechanism of enhanced IL-18/IL-18R signaling induced by MWA treatment of OSA that promotes antitumor immune responses.

OSA is a particularly aggressive form of primary bone cancer threatening the survival and limb function of children and adolescents, without any further therapeutic improvement in overall survival under the current standard therapy, including neoadjuvant chemotherapy, the administration of chemotherapy postoperationally. More recently, MWA treatment has been extensively applied in treating OSA, and achieved efficient tumor control and preservation of limb function.1882,84 During and after the clinical application of MWA, the induction of concomitant immune responses had been observed and efforts have been made to uncover the underlying mechanisms, with limited success until now.⁸⁵ New insights into the molecular regulation of immune responses after MWA treatment, however, could be employed to further improve therapeutic applications and to augment beneficial immune responses in patients with cancer.

Our study demonstrated that MWA-treated K7M2 OSA cells displayed an apoptosis state with the condition of 5 W for 2 min, rather than coagulation necrosis, which promoted an immune-stimulatory state of the treated cells. Specifically, elevated expression of CRT and HMGB-1 24 hours after treatment indicated enhanced immunogenicity of MWA-exposed K7M2 cells, as CRT had been shown to promote phagocytosis by DCs through an enhanced “eat me” signature,⁸⁶ and HMGB-1 was found to bind specific receptors on myeloid cells and primed immune responses.⁸⁷ Meanwhile, our co-culture models in vitro demonstrated that co-stimulation, antigen uptake and cross-presentation of BMDC were enhanced after co-cultured with MWA-exposed K7M2 cells. These results indicated that MWA-treated K7M2 cells might promote the initiation of antitumor immune responses after DC activation. To test this hypothesis in vivo, we established the OSA femur in situ model and performed tumor regression studies in the presence or absence of MWA treatment. Of note, untreated tumor tissue was almost void of CD8⁺ T cells and CD11c⁺ DCs in tumor tissue and tumor draining lymph nodes, a phenotype previously defined as a “cold” tumor microenvironment.⁸⁵ In contrast, MWA treatment provoked CD8⁺ T cells and CD11c⁺ DCs infiltration into both tumor tissues and draining lymph nodes, indicating that the immune-deserted OSA was successfully converted into a tumor with a “hot” microenvironment. This was evident by the observation of enhanced immune cells recruitment into the tumor and the promotion of stronger antitumor effects.

Aiming to uncover the underlying mechanism of how MWA-exposed K7M2 cells could induce antitumor immunity, we explored a co-culture model in which either control or MWA-treated K7M2 cells were co-incubated with BMDC for 24 hours. Next, CD11c⁺ BMDC were isolated using CD11c positive magnetic selection beads for further RNA-seq analysis. KEGG pathway enrichment analysis of the acquired datasets indicated the importance of specific cytokine and cytokine receptor pathways, which implied that MWA-treated K7M2 cells might release cytokines that interact with receptors on BMDC to regulate immune responses. Thus, we analyzed the top 15 cytokine and chemokine receptor genes that were differentially expressed when comparing MWA-treated versus untreated K7M2 cells as stimulation of BMDC. The results highlighted IL-18R as an interesting candidate to further study in this system. IL-18, the ligand for IL-18R, was expressed by MWA-treated tumor cells and has a functional role in immune activation. Specifically, we detected IL-18 expression on the transcriptional and translational level in MWA-exposed K7M2 cells by real-time PCR and western blotting, and we measured the exact secretion by ELISA measurements of cell culture supernatants. Microscopic analysis of tumor tissue sections also confirmed elevated expression levels of intratumoral IL-18 after MWA treatment. Before performing further experiments, we additionally performed an ELISA assay for the IL-18 binding protein (IL-18BP) using the same supernatants from K7M2 cells 24 hours after MWA treatment, as IL-18BP has previously been reported to competitively bind IL-18R.⁸⁸ However, the concentrations of IL-18BP measured were extremely low and below the detection limit of the assay, and we therefore concluded that this inhibitory factor could not suppress possible immune-activation induced by IL-18 as a consequence of MWA treatment.

IL-18 is a member of the IL-1 family of cytokines. In resting cells, it is constitutively expressed as an inactive pro-form in the cytosol, which binds to caspase-1 or 4 to form a complex, and it is cleaved into its active form under conditions of inflammation.^{89 90} Active IL-18 is released into the extracellular space and functions as an inflammation mediator. In our study, we investigated IL-18 as a mediator of anticancer immune responses following MWA treatment of OSA cells, and we applied IL-18 neutralizing antibodies to confirm its critical role in vivo. Enhanced infiltration of CD8⁺ T cells and CD11c⁺ DCs was detected by flow cytometry after MWA-treatment of tumors both within tumor tissues and draining lymph nodes, which could be reversed by administration of IL-18 neutralizing antibodies. Additionally, blocking the IL-18/IL-18R axis reversed the therapeutic benefits of MWA treatment on tumor outgrowth. We could therefore demonstrate a correlation between IL-18 release by MWA-treated K7M2 cells and the infiltration of CD8⁺ T cells and CD11c⁺ DCs, which resulted in tumor eradication. Besides, the IL-18R expression on the cell surface of DCs did not decline after IL-18 neutralizing antibodies performance, which illustrated that there might be additional regulators expressed by MWA-treated K7M2 cells that promote IL-18R expression, such as IL-15,⁹¹ although we did not detect them.

As already indicated above, OSA is considered a “cold” tumor with low CD8⁺ T-cell infiltration and limited expression of immune checkpoint molecules, resulting in resistance to immune checkpoint inhibitor therapy and poor prognostics.⁸⁵ In addition, thermal therapies inducing a “hot” tumor phenotype have previously shown to boost antitumor immune responses, and show synergies with immune checkpoint inhibitor therapies through enhancing tumor antigens exposure and releasing inflammatory cytokines, thereby promoting an “in situ vaccination” effects.92,94 Meanwhile, patterns of MWA-treated tumors and following resection also promote immune activation.^{20 56} A different study demonstrated that by designing IL-18-secreting DLL3-targeting CAR T cells, tumor cytotoxicity could be augmented in several xenograft and syngeneic small cell lung carcinoma models. In addition, IL-18 stimulated persistence and activation of both CAR T cells and tumor-reactive endogenous T cells, as well as myeloid cells in the tumor microenvironment.⁷⁵ All these reported cases of IL-18-mediated T-cell invigorating and tumor eliminating effects support our findings that MWA treatment enhanced therapeutic efficiency by boosting antitumor immune responses through promoting IL-18 expression and enhancing the functions of DC and CD8⁺ T cells.

At the same time, there are several additional mechanisms and advantages of MWA treatment, which were not specifically explored in our study. Extratumor matrix has been considered a barrier for chemotherapy or immunotherapy due to the gathering of cancer-associated fibroblasts and abnormal vessels.⁹⁵ In this context, MWA treatment has been shown to break down these barriers that surround tumor tissues, thereby facilitating increased chemotaxis and infiltration of immune cells, as well as drug infusion.^{96 97} In addition, thermal therapies regulated abnormal vascular structures to enhance drug delivery to tumors and thereby enhance therapeutic effects.^{96 98} Another important mechanism by which MWA-treatment promoted antitumor immune responses is by boosting the release of tumor-associated antigens (TAAs) important for T-cell priming.^{25 99} Although advanced strategies of next-generation sequencing technologies nowadays allow the identification of tumor-specific mutations as a basis for the development of targeted cancer immunotherapies, similar approaches in OSA face the challenge that the tumor mutation rate is low and shared mutations are almost absent among different patients.^{99 100} Therefore, novel strategies are emerging to use TAAs to induce antigen-specific T cells in these tumor types that lack detectable mutations and tumor-specific antigens (TSAs).^{101 102} In this setting, CD8⁺ T cells are primed by APCs (Antigen-presenting cells) against TAAs with the help of adjuvant cytokines to generate effector TAA-specific T cells that can target tumor cells with TAAs presented on the surface. Afterwards, tumors would be eradicated, while the specificity of killing might be lower compared with CD8⁺ T cells cytotoxicity induced by recognition of TSAs.103,105 From this perspective, MWA treatment might be a suitable method to induce TAA-specific antitumor immune responses. Based on the immune promoting effects of MWA treatment, combination with additional therapies, such as immune checkpoint blockade, tumor vaccination, or cytokine treatments, might show synergies for improved OSA tumor control and enhanced antitumor immune responses compared with traditional therapies, such as chemotherapy and immunotherapy.

Given that MWA promoted IL-18 expression and suppressed OSA progression, which was reversed on IL-18 blockade, it is essential to investigate whether IL-18 cytokine could be developed as a therapeutic agent for OSA. Notably, IL-18 monotherapy exerted a protective effect against OSA in mouse models. Moreover, IL-18 demonstrated an additive antitumor effect when combined with anti-PD-1 monoclonal antibodies, accompanied by enhanced intratumoral infiltration of CD8⁺ T cells. These findings highlight a potential strategy to overcome immunotherapy resistance in OSA.

In its current stage, our study has still some limitations. First, MWA is a thermal therapy combined with magnetic wave ablation, and it would be best to distinguish the specific effects of MWA treatment from a simple rise in temperature. However, it remains challenging to differentiate between the effects of specific physical MWA parameters, such as radiation pressure,¹⁰⁶ thermal mechanical stress¹⁰⁷ and acoustic wave strength,¹⁰⁸ which ideally would all be studied individually in the context of cancer treatment to gain mechanistic insights into the therapeutic mechanism of MWA treatment action. Meanwhile, it has been suggested that fever (~39°C) can enhance the metabolic fitness and antitumor efficacy of CD8⁺ effector T cells, indicating the protective role of hyperthermia in cancer therapy.^{109 110} It would be more convincing to distinguish between direct thermal effects and MWA-promoted IL-18 expression on T cells activation in future studies. Second, clinical OSA samples would be most optimal to verify our conclusions, but ethical restrictions make these tissues mostly inaccessible. The rapidly advancing organoid technology might provide more data that is in line with clinical samples. Third, it would be interesting to directly compare MWA-induced immune effects with other thermal therapies, including radioablation, photothermal ablation, and cryoablation.⁷² By thoroughly exploring the related mechanisms of action, further progress could be made to improve the therapeutic efficacy of thermal therapies against cancer.

In summary, our study unveiled a critical mechanism underlying MWA treatment-induced immune activation in OSA, which was mediated by IL-18-induced DC activation and CD8⁺ T cells priming. Blockage of IL-18 with neutralizing antibodies reverted MWA treatment-induced immune stimulation and tumor control. Thus, our findings conceptually advanced our understanding of the mechanism underlying MWA treatment in cancer immunotherapy and might provide an attractive approach to augment antitumor immunity in patients with tumor.

Materials and methods

Cell lines

The murine OSA cell line K7M2 (SCSP-5217, Cell bank of Chinese Academy of Sciences, Shanghai, China) was cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, C11995500BT) supplemented with 10% fetal bovine serum (FBS, VivaCell, C04001-500) and 1% streptomycin and penicillin at 37°C in 5% CO₂. Cells were harvested using 0.25% Trypsin-EDTA (Gibco, 25200–072).

Primary cells

BMDC were differentiated from progenitor cells from bone marrow. After extraction from the femur cavity of 4–6 weeks old male BALB/c or C57BL/6 mice, cells were cultured in RPMI (Roswell Park Memorial Institute) media supplemented with 10% (FBS) and 1% streptomycin and penicillin, along with 20 ng/mL murine GM-CSF and 20 ng/mL IL-4 at 37°C in 5% CO₂ for further differentiation on the 6-well plates. Culture media was refreshed at day 4, and differentiated BMDC were obtained on day 7. Cell scrapers were used to detach the BMDC for further experiments.

Surgical procedure of patients

The surgical procedures involved in situ MWA with intralesional resection and subsequent mechanical reinforcement. After identifying the extent of the tumor and dissecting the tumor-bearing bone from surrounding normal tissues with a safe margin, an ablation area was constructed to completely inactivate the tumor in situ. After the tumor was confirmed to be inactivated by radiologists and surgeons, affected bone tissues were resected. Besides, tumor edges were ablated by MWA to avoid potential residual tumor cells.

MWA procedure of experiments

A microwave therapy apparatus (WB-3100AI, Baoxing, Xuzhou, China) was used for MWA treatments. Cells or tumor tissues were treated with the microwave antenna at 5 W to induce treatment effects. For in vitro experiments, the microwave antenna was fixed at the center of a 1.5 mL tube containing 1.0 mL cell suspension. Regarding the femur OSA in situ mouse models, the microwave antenna was placed at the center of the tumor for effective treatment. Cooled sterile saline was dropped on the surface of the skin to avoid burns.

To measure the MWA-induced heating effect, 1×10⁶ K7M2 cells resuspended with 1 mL complete medium into 1.5 mL tapered tubes were prepared. At the time MWA was performed, temperature changes were monitored dynamically by using an infrared thermal imager (FLIR, Germany).

Detection of cell viability using the cell counting kit 8

To detect cellular viability after MWA treatment, K7M2 cells in logarithmic growth phase were harvested with 0.25% trypsin containing EDTA, followed by cell counting and MWA treatment, as previously described. Next, K7M2 cells were seeded into 96-well plates at a density of 1×10⁴ cells per well and cultured at 37°C with 5% CO₂. After being cultured for 1, 3, or 7 days, respectively, media was discarded and cells were washed two times with phosphate-buffered saline (PBS). Subsequently, 100 µL CCK-8 working solution (Dojindo, CK04, Japan) per well was added and incubated at 37°C with 5% CO₂ for 2 hours. For quantification of cell viability, OD (Optical Density) values were measured on a microplate reader (Multiskan GO, Thermo, USA) at 450 nm.

Detection of MWA-induced apoptosis and necrosis

MWA-treated K7M2 cells were cultured for 24 hours and subsequently trypsinized and washed with PBS two times. Annexin V, FITC Apoptosis Detection Kit (Dojindo, AD10, Japan) was used to measure MWA treatment-induced apoptosis and necrosis according to the manufacturer’s instructions. Specifically, K7M2 cells were stained with Annexin V working solution in the dark for 15 min, followed by PI working solution staining in the dark for 5 min. All the working solutions were prepared in buffers from the kit. Stained cells were analyzed by flow cytometry (Beckman CytoFLEX, USA). All the flow data were analyzed by FlowJo software (FlowJo, USA).

Transmission electron microscope

After MWA treatment, K7M2 cells were cultured for 24 hours before cells were collected and fixed with 2.5% glutaraldehyde for 2 hours in the dark at room temperature. 0.1 M PBS (pH=7.4) was used to wash cell pellets three times for 10 min each. Next, 1% OSO4 (Roles-Bio, RBGP1060) in 0.1 M PBS (pH=7.4) was applied for fixation at room temperature for 1–2 hours. After that, cell pellets were washed in 0.1 M PBS (pH=7.4) three times for 10 min each, and then dehydrated with a concentration gradient of ethanol (30%, 50%, 70%, 80%, 90%, 95% and 100%) for 15 min each. Finally, samples were dried, embedded, cured and cut into sheets of 70 µm. Microstructure of cells was observed by TEM (Hitachi H-750, Japan).

Enzyme linked immunosorbent assay

Cultured media of MWA-treated K7M2 cells was centrifuged at 2,000 g for 20 min at 4°C to separate supernatant and cell pellets for ELISA detection. The supernatant was removed to analyze HMGB-1 (Invitrogen, EEL102, USA), IL-18 (Abcam, ab216165, UK) and IL-18BP (Abcam, ab254509, UK) concentrations. Cell pellets were lysed with RIPA (Radio-Immunoprecipitation Assay) lysate buffer for CRT detection (Abcam, ab284624, UK). Expression of HMGB-1, CRT, IL-18 and IL-18BP were calculated from OD values at 450 nm combined with a standard curve (Multiskan GO, Thermo, USA).

For detection of cytokines derived from BMDC, cultured media was centrifuged at 2,000 g for 20 min at 4°C to separate supernatant and cell pellets. Supernatants were obtained for analyses of the levels of IL-6 (BioLegend, 575709, USA), IL-10 (BioLegend, 431411, USA) and IL-12 (Invitrogen, BMS616, USA). All the concentrations were calculated from OD value at 450 nm combined with standard curve (Multiskan GO, Thermo, USA).

Mult cytokine assay

To measure the inflammatory signature changes after MWA treatment, we collected peripheral blood samples from patients with OSA, and centrifuged at 500 g for 8 min at 4°C to obtain plasma for cytokine detection using a panel of 14 human cytokines (QuantoBio, C60011, CN) in accordance with the recommendations of the manufacturer. The selected cytokines included IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17A, IL-17F, IL-22, TNF-α, TNF-β and IFN-γ. Cytokine levels were evaluated with a Luminex 200 system (Luminex).

Isolation of intratumoral CD45⁺ cells

All tumor tissues were processed within 1 hour after resection. Tumor tissues were minced in sterile FBS free DMEM and digested with Collagenase II (0.5 mg/mL) and Collagenase IV (1 mg/mL) at 37°C for 0.5 hours. Single-cell suspensions were filtered (70 µm) and counted. A mouse CD45⁺ positive selection kit (EasySep, #18945) was used to obtain CD45⁺ immune cells according to the manufacturer’s instruction, and cells were further stained with fluorochrome-conjugated monoclonal antibodies for flow cytometric detection.

Flow cytometry

The following antibodies were applied for flow cytometry: anti-human CD3 (clone SP34-2, BD Pharmingen, 5 µL/test), anti-human CD45 (clone HI30, BD Pharmingen, 5 µL/test), anti-human CD4 (clone RPA-T4, BD Pharmingen, 5 µL/test), anti-human CD8 (clone RPA-T8, BD Pharmingen, 5 µL/test), anti-human CD25 (clone 2A3, BD Pharmingen, 5 µL/test), anti-human CD44 (clone G44-26, BD Pharmingen, 10 µL/test), anti-human Foxp3 (clone 259D/C7, BD Pharmingen, 5 µL/test), anti-mouse CD11c (clone N418, BioLegend, 1:200 diluted), anti-mouse MHC-II (clone M5/114.15.2, BioLegend, 1:200 diluted), anti-mouse CD80 (clone 16–10 A1, BioLegend, 1:200 diluted), anti-mouse CD86 (clone GL1, BioLegend, 1:200 diluted), anti-mouse CD45 (clone 30-F11, BD Biosciences, 1:200 diluted), anti-mouse CD3 (clone 17A2, BioLegend, 1:400 diluted), anti-mouse CD8 (clone 53–6.7, BioLegend, 1:200 diluted), anti-mouse IFN-γ (clone XMG1.2, BioLegend, 1:250 diluted), anti-mouse IL-18R (clone P3TUNYA, Thermo Fisher, 1:80 diluted). Cell surface markers were stained with fluorochrome-labeled antibodies in flow cytometry staining buffer (PBS containing 3% fetal bovine serum). Surface staining was performed in the dark for 20–30 min at 4°C before further intracellular staining or acquisition by flow cytometry. For intracellular staining, after the surface staining was performed, Cyto-Fast Fix/Perm Buffer was added to the samples in the dark for 20 min at room temperature, followed by washed with staining buffer two times. Then fluorochrome-labeled antibodies were added in the dark for 20–30 min at 4°C before further acquisition by flow cytometry. Next, cells were washed two times with staining buffer and resuspended for flow cytometric detection. Stained cells were analyzed by flow cytometry (MBP FACSymphony S6, BD Biosciences). FVS 780 (Fixable Viability Dye eFlour 780, eBioscience, 65–0865-14, USA, 1:100 diluted) was used for live or dead cells selection. All the flow data were analyzed by FlowJo software (FlowJo, USA).

RNA sequencing

Total RNA was extracted using TRIzol reagent, and messenger RNA (mRNA) was enriched with Oligo (dT) beads. The enriched mRNA was fragmented, reverse-transcribed into complementary DNA (cDNA) using the NEBNext Ultra RNA Library Prep Kit (New England Biolabs), and purified. Double-stranded cDNA was end-repaired, ligated to Illumina adaptors and purified with AMPure XP Beads. The ligated fragments were size-selected by agarose gel electrophoresis, PCR amplified and sequenced on the Illumina NovaSeq 6000 by Gene Denovo Biotechnology. For RNA-seq data analysis, high-quality reads were mapped to the reference genome using the alignment tool HISAT2 (V.2.4). For each transcription region, an FPKM (fragment per kilobase of transcript per million mapped reads) value was calculated to quantify its expression abundance using StringTie software (V.10). Differential gene expression between groups was analyzed using the edgeR software with a false discovery rate <0.05 and |logFC|≥1 as the threshold. Gene Ontology and KEGG analyses were conducted to functionally annotate these differentially expressed genes.

Quantitative real-time PCR

To analyze IL-18 gene expression of K7M2 cells in the presence or absence of MWA treatment at the transcriptional level, quantitative real-time PCR was employed. In brief, a total RNA extraction kit (Omega BIO-TEK, R6834-02) was used to extract total RNA 24 hours after treatment, which was then quantified using a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, Massachusetts, USA). RNA was reverse transcribed with PrimeScript RT Master Mix (Takara RR036Q). Quantitative PCR was performed with ChamQ SYBR qPCR Master Mix (Vazyme, Q311-02). (Primers used for PCR are listed in online supplemental table S3).

Western blot analysis

MWA-treated K7M2 cells or non-exposed controls were lysed in RIPA buffer (Beyotime Biotechnology, catalog no. P0013B) containing a protease inhibitor cocktail (TargetMol, catalog no. C0001) for 15 min on ice. The protein concentration was detected using a BCA Assay Kit (Thermo Fisher, catalog no. A53225), and 25 µg of total protein were separated by an SDS-PAGE (Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis) and subsequently blotted onto a polyvinylidene fluoride (PVDF) membrane (Merck Millipore, catalog no. IPVH00010). After blocked with 5% bovine serum albumin solution, the PVDF membranes were incubated with primary antibody anti-IL-18 (1:1000, Abcam, catalog no. ab191860) in TBS-T (Tris-Buffered Saline with Tween 20) at 4°C overnight. Next day, the PVDF membrane was washed with TBS-T three times and incubated with the second antibody (1:5,000, Abcam, catalog no. ab7090) dissolved in TBS-T for 1 hour at room temperature, before it was washed again three times with TBS-T. The bands were then imaged using a Bio-Rad ChemiDoc system following the manufacturer’s instructions.

Immunofluorescence

Multiplex immunofluorescence experiments were performed using consecutive staining cycles in the following ways: Tumors were isolated from the mice, fixed with 10% neutral formalin and embedded with paraffin, followed by wax box section. The sectioned slices were processed with deparaffinized in xylene and rehydrated in graded alcohol. Sections were then performed with antigen retrieval, endogenous peroxidase inactivation and non-specific blocking. They were subjected to serial staining with antibodies specific for CD11c (1:100, Abcam, catalog no. ab219799) and IL-18 (1:50, Abcam, catalog no. ab191860). Afterward, each sample was stained with corresponding secondary antibodies. Serial staining for CD8 (1:1,000, Abcam, catalog no. ab217344) and CD11c (1:100, Abcam, catalog no. ab219799) was carried out as described above. All slides were imaged using Nikon AX-NIS-Elements confocal microscopy.

In vivo experiments

To establish the mouse model of orthotopic OSA in the femur with intact immunity, 1×10⁶ K7M2 cells were injected into the femurs of 4–6 weeks old male Balb/c mice. Tumor volume was calculated as formula: 1/2×L (L: longest diameter of the tumor)×S² (S: the maximum width in the vertical direction of “L”). When the volume of tumors reached 300 mm³, MWA treatment was carried out using an intensity of 5 W for 2 min. Experimental groups were defined as control (sham-operation), MWA (with MWA treatment), anti-IL-18 (100 µg per dose, intraperitoneal injection, every 3 days) and MWA+anti-IL-18 (MWA treatment combined with anti-IL-18 administration). Tumor volume was measured and recorded every 3 days. Mice were sacrificed at day 18 post-MWA treatment, and tumors as well as draining lymph nodes were harvested for FCM (Flow Cytometry) analysis or immunofluorescence staining.

For anti PD-1 monoclonal antibodies and IL-18 cytokine applications, performances were carried out when the tumor volume reached 100 mm³. Both anti PD-1 monoclonal antibodies (12.5 mg/kg body weight, Bio X Cell, BE0146) and IL-18 cytokine (1 µg per dose, Sino Biological, 50073-MNCE) were intraperitoneally injected every 3 days. Tumor volume was measured and recorded every 3 days. Mice were sacrificed at day 18 post-different treatments, and tumors were harvested for immunofluorescence staining.

Statistical analysis

Data are presented as the mean±SD. Differences between two groups were analyzed with Student’s two-tailed t-test. One-way analysis of variance was used to determine statistical differences among three or more groups. Statistical analyses were conducted using Origin Pro software (Origin Lab Corporation, Massachusetts, USA). Graph generation and statistical analyses were performed using GraphPad Prism, V.10.1.2 (GraphPad Software). Wilcoxon rank-sum/Mann-Whitney U test was performed followed by univariate and multivariate analyses to compare cytokines in this study. Correlation coefficients between the cytokines were calculated using a Pearson correlation coefficient and presented in correlograms. P value<0.05 was considered statistically significant for the comparisons.

Data availability statement

Data are available upon reasonable request.