Hepatocellular carcinoma (HCC) is the fifth most common malignancy and the third leading cause of cancer-related deaths worldwide, occurring most frequently in the setting of chronic liver injury and cirrhosis. Despite improvements in the diagnostic and treatment modalities, the long-term prognosis has been far from satisfactory [1, 2].

Radiofrequency ablation (RFA) has been established as the primary ablative modality for HCC. However, similar to other treatments, RFA for HCC is frequently followed by tumour recurrence [3]. There is an urgent need to develop novel therapies with systemic activity to avoid spontaneous progression or recurrence of HCC. Immunotherapy represents a potential therapeutic option.

It has been shown that RFA can not only effectively kill HCC tumour cells but also provoke the immune response to remove the debris [4, 5]. At temperatures above 45 °C to 50 °C, cell membranes are destroyed, proteins are denatured, and a region of necrosis surrounding the electrode is generated [6]. Hyperthermia-related tumour antigenicity suggests that thermal ablation-inactivated tumour tissues are endowed with immunogenicity, which might be closely related to the necrotic tumour cell-released molecules with high antigenic characteristics. In addition, the inflammatory response resulting from tumour necrosis following thermal ablation may lead to the accumulation of a large number of antigen-presenting cells (APCs) and lymphocytes at the treatment sites. These necrotic tumour cell-activated APCs, which are more dynamic than those activated by apoptotic cells, can more effectively promote the effect of antigen presentation to effector cells [7, 8].

Therefore, RFA favours immune activation and the presentation of otherwise cryptic antigens, thus inducing a specific anti-tumour immune response. However, unfortunately, there is no consensus regarding which antigens could constitutively be generated after RFA and could potentiate the immune response.

To identify the immune-potentiating antigens, proteomic analysis with two-dimensional electrophoresis (2-DE) and mass spectrometric analysis (MS) was used in this study. We found that Ficolin-3 was overexpressed in the serum of most HCC patients after RFA. In addition, elevated Ficolin-3 predicts better prognosis, which might be a biomarker for RFA treatment efficacy and a potential therapeutic target for HCC immunotherapy.

RFA is a minimally invasive technique used as standard local therapy for HCC. Its immune-potentiating effect has been frequently reported with few studies concentrating on the exact mechanisms. In the present study, we performed a gel-based serum proteomic analysis before and after RFA for HCC patients and identified 18 proteins to be differentially expressed by mass spectrometry. Through literature review, we found three proteins, CLU, Ficolin-3, and RBP4, that might be related to an immune response. Further ELISA assay showed that only Ficolin-3 significantly changed after RFA. Survival analysis showed that patients with higher Ficolin-3 after RFA tended to have better DFS, indicating that Ficolin-3 might be one of the immune-potentiating tumour antigens.

The exact explanation for the association between the enhanced immune response and RFA is not fully clarified. There are several possible explanations. First, RFA destroyed the tumour and relieved the body of the tumour burden, possibly leading to the reversal of immune suppression and unmasking a population of primed tumour-specific T cells that can mediate protective immunity [11]. Second, RFA leads to enhanced release and exposure of immune-potentiating tumour antigens, such as nuclear protein high mobility group B1 and heat shock proteins (HSPs), which might induce the antitumour immune response through the activation of dendritic cells (DCs) [12]. At 24 h after percutaneous treatment of HCC by thermal ablation, the expression of HSP70 in the cytoplasm and on the cell membrane of tumour cells was augmented by 8 times and that of HSP90 was augmented by 1.2 times. The two HSPs are evidently antigenic stimulatory to the immune system [13]. The anti-tumour effect of in vitro-inactivated vaccine is comparatively weak, which might be due to the rinsing of many immunogens (such as HSP) released by necrotic tumour cells during multiple centrifugal washing processes, resulting in immunogenicity decline of the tumour cells finally collected. This indirectly proves the effectiveness of the anti-tumour immune response stimulated by tumour-associated antigens released in situ from tumour tissues following thermal ablation. Ficolin-3 might be one of these in situ-released antigens. Third, hyperthermia induced by RFA causes both immunologic and biologic effects, such as the accelerated migration of peripheral blood mononuclear cells, activation of effector cells, and induction and secretion of cytokines, all of which might enhance the antitumour immune response [14]. Hyperthermia-killed HCC cells express a large number of HSPs. Compared with tumour cells untreated with heat shock, hyperthermia-treated ones have sensitized dendritic cell vaccines more effective in inducing CD4⁺ and CD8⁺ T cells to participate in the anti-tumour immune response [15]. In addition, HCC ablation induces a functional transient activation of myeloid dendritic cells (MDC) associated with increased serum levels of TNF-α and IL-1β [16].

We found that serum Ficolin-3 was significantly overexpressed after RFA, which was verified by ELISA in 52 specimens. Furthermore, we found that patients with higher Ficolin-3 after RFA had better 1-, 3-, and 5- year DFS. However, no significant difference was found for CLU and RBP4. Ficolin-3 (H-ficolin; Hakata antigen) is a thermolabile beta-2-macroglycoprotein found in all human sera and is a member of the ficolin/opsonin p35 lectin family. The protein can activate the complement pathway in association with MASPs and sMAP, thereby aiding in the host defence through the activation of the lectin pathway [17]. It was found that a low concentration of Ficolin-3 was associated with an increased risk of fever and neutropenia (FN), particularly FN with bacteraemia, in children treated with chemotherapy for cancer [18]. In addition, Ficolin-3 was found to bind to IgMs, and the IgM-Ficolin-3 complex deposited on cancer cells could induce complement attack. Therefore, the IgM-Ficolin-3 -mediated complement activation pathway might be a new defensive strategy for human cancer immunosurveillance. These two studies showed that high Ficolin-3 represents strong immunity against infection or cancer. We proposed this might be one of the reasons for better survival for HCC patients with higher serum Ficolin-3 after RFA.

Because the immune-potentiating effect of thermal ablation is not sufficiently strong to prevent tumour recurrence, there is an urgent need to develop novel therapies with systemic activity to avoid spontaneous progression or recurrence after RFA treatment [19]. In a rabbit VX2 hepatoma model, it was found that combined RFA and Toll-like receptor 9 (TLR9) agonist stimulation not only induced increased antitumour T-cell stimulation/cytotoxicity and a longer mean survival of animals but also significantly inhibited tumour spread to the lungs and peritoneum and prohibited new tumour growth in animals receiving a secondary systemic tumour cell injection [20]. In HCC patients, Cui et al. found that cellular immunotherapy (CIT) with autologous mononuclear cell-derived natural killer (NK) cells, γδT cells and cytokine-induced killer (CIK) cells could enhance progression-free survival for HCC patients after RFA with no adverse response [21]. den Brok et al. found that RFA induced a weak immune response, but it was not sufficient to protect against tumour recurrence. Therefore, they combined immune-potentiating strategies via the administration of CTLA-4-blocking antibodies to lower the threshold for T-cell activation and obtained a better immune response, finally leading to increased tumour protection [22]. Actually, we conceived the same idea as den Brok that tumour antigens released by tumours in situ or in the serum might induce an antitumour immune response, leading to a better prognosis, and Ficolin-3 might be one of those antigens.

Our study possesses limitations. First, in this study, we tried to identify immune-potentiating tumour antigens released after RFA, but we could only find antigens soluble in serum instead of other tumour antigens in situ. It would have been more ideal to have used ablated tumour tissues in this study. However, in clinical practice, it is very difficult to obtain sufficient tumour antigens. This idea is undoubtedly possible in animal models, although it would not reflect the true circumstance. Second, most of the HCC patients were HBV associated in the present study. Thus, it remains to be verified in what type of model would Ficolin-3 change after RFA using other hepatitis-associated HCC samples. Additionally, although we found that Ficolin-3 might be used as a tumour vaccine for the prevention and therapy of residual tumour through stimulation of the immune system, we have not proven this hypothesis, which we hope to confirm in a future animal study or even with preclinical tests.

We found that Ficolin-3 was overexpressed in the serum of most HCC patients after RFA and might be a potential biomarker for RFA treatment efficacy and tumour vaccine development for HCC immunotherapy. The combination of RFA and active immunotherapy with Ficolin-3 may have a synergistic effect in HCC treatment.