EBV-induced upregulation of CD55 reduces the efficacy of cetuximab treatment in nasopharyngeal carcinoma

Abstract

Cetuximab, an anti-epidermal growth factor receptor (EGFR) antibody, has been shown to improve survival in nasopharyngeal carcinoma (NPC) patients. However, a correlation between the expression of EGFR and the response to cetuximab has not been observed, indicating that the mechanism underlying the effects of cetuximab needs to be further elucidated. The antitumour response involves immunotherapeutic mechanisms that target tumour-associated antigens, including complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC), act either alone or, more often, in combination. However, EBV infected NPC cells often develop resistance mechanisms that allow them to evade immune surveillance. Here, we found that overexpression of the complement-regulated protein CD55 in EBV-associated NPC cells mainly suppresses ADCC activity thus reduces the efficacy of cetuximab. Mechanistically, EBV latent membrane protein 1 (LMP1) mediated upregulation of CD55 through the NF-κB signalling pathway. The present study provides a rationale for the development of CD55 inhibitors to improve the clinical efficacy of cetuximab in NPC.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-024-05822-3.

Introduction

Nasopharyngeal carcinoma (NPC) is a malignancy arising from the nasopharynx and has high rates of incidence throughout southern China and Southeast Asia. Approximately 98% of undifferentiated NPC patients have latent infection with EBV [1]. EBV‑infected nasopharyngeal epithelial cells usually express EBV antigens, including EBV-associated nuclear antigen-1 (EBNA1), latent membrane protein 1 and 2 (LMP1 and LMP2), and BamHI A rightward transcript-microRNAs (BARTs) [2–4]. The complement system is the first line of innate immunity, which can rapidly recognize and kill virus-infected cells [5]. However, EBV-infected NPC cells can express these gene products to evade complement cytotoxicity and resist antitumour therapy [6, 7]. Therefore, further understanding the role of EBV in the tumour microenvironment may provide effective therapeutic targets for EBV-associated NPC.

Cetuximab, a monoclonal antibody that binds to epidermal growth factor receptor (EGFR), has been used for the treatment of EBV-infected NPC with some success [8, 9]. However, the expression of EGFR does not ensure effective therapy [10], indicating that other mechanisms contribute to anti-EGFR treatment in EBV-infected NPC. Increasing evidence suggests that multiple interacting mechanisms, including complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC), act either alone or, more often, in combination, in the antitumour response of monoclonal antibodies(mAbs), including cetuximab [11, 12]. CDC activation can lead to the formation of a membrane attack complex, inducing lysis of target cells. Additionally, complement may enhance natural killer (NK) cell cytotoxicity, thus increasing ADCC-induced cytotoxicity [13]. Xenograft models of non-small cell lung cancer have demonstrated that the antitumour activity of cetuximab is associated with the complement-mediated immune response [14]. Moreover, cetuximab can elicit a specific NK cell-mediated killing response ranging from 20 to 90% in vitro [15], suggesting that the mediators of NK cell cytotoxicity constitute an important mechanism of cetuximab treatment. However, the major limitation of complement is the complement-regulatory proteins [16], which provide a means of evasion from complement attack. Therefore, understanding the interactions between complement activation and cetuximab will increase the clinical efficacy and improve the long-term survival of NPC patients.

In this study, we observed that CD55 was overexpressed in EBV-infected NPC cells. CD55, or decay accelerating factor (DAF), was first discovered on the surface of erythrocytes. Structurally, CD55 has three major domains: four short consensus repeat (SCR) domains called SCR-1, SCR-2, SCR-3, and SCR-4, a serine/Threonine (S/T)-rich domain; and a GPI anchor domain [17]. Functionally, CD55 is a complement-regulatory membrane protein that participates in the regulation of complement activation and protects cells from complement-mediated attack by limiting the formation of C3/C5 convertase enzymes [18]. Therefore, multiple studies have focused on the role of CD55 in immune surveillance. More recently, aberrant CD55 expression has also been described in many solid tumours, including NPC, where it is higher than that in normal cells and promotes malignant progression [19–21]. However, whether the regulation of CD55 expression by EBV infection can induce complement–mediated immune evasion in NPC patients treated with cetuximab has yet to be studied. In this study, we found that CD55 overexpression in EBV-associated NPC cells mainly suppressed NK cell–mediated antitumour activity during treatment with cetuximab. Furthermore, CD55 knockdown markedly enhanced antitumour responses to cetuximab therapy in vivo, indicating the potential application of CD55 as a therapeutic target via enhanced NK cell-mediated cytolytic activity to increase the efficacy of cetuximab in NPC.

Results

EBV infection promotes complement-regulatory protein CD55 expression in NPC

To investigate the role of EBV infection in NPC, we first compared the transcriptomes of previously established EBV-infected (CNE2-EBV and TW03-EBV), and EBV-negative (CNE2 and TW03) cell pairs and EBV infected cells with EBV nuclear antigen 1 (EBNA1) deletion by two gRNAs (#1 and #2) [22]. The construction of stable cell lines has been demonstrated by Western blotting in our previous studies [22–24]. In this study, in situ hybridization was further used to detect the presence of EBER in the majority of EBV-infected cultures (Fig. 1b). We found that the transcription of CD55 was significantly upregulated in EBV-infected NPC cells. Moreover, a reduction in CD55 expression was detected in EBV-infected cells following EBNA1-deletion (Fig. 1a). Western blotting and real-time PCR confirmed that CD55 protein and mRNA expression were increased in these EBV-infected NPC cells (Fig. 1c and d). As CD55 is a membrane-expressed protein, we conducted flow cytometric analysis to detect CD55 expression and detected the same positive association between CD55 expression and EBV infection in NPC cell lines (Fig. 1e). Consistently, the upregulation of CD55 was further confirmed by IHC in EBV-infected NPC biopsies, but was barely detectable in normal epithelial cells (Fig. 1f and g). CD59 is also known to be a complement inhibitory protein. Numerous studies have shown that CD59 overexpression is related to poor prognosis [25, 26]. In this study, a minor difference in CD59 expression was detected via whole-transcriptome sequencing (Fig. 1a). Further analysis revealed that NPC cells expressed CD59 at markedly higher levels than did normal nasopharyngeal epithelial cells. However, the association between EBV infection and CD59 expression was not detected in other NPC cell lines (Supplementary Fig. 2). Therefore, in this study, we conducted an analysis to evaluate the function of CD55.

Fig. 1.

CD55 is overexpressed in EBV-infected nasopharyngeal carcinoma (NPC). (a), Cellular RNA was sequenced by RNA-Seq in EBV- and EBV + NPC cell pairs and in EBV-infected cells with EBNA1 deletion by gRNAs. (b), EBV-encoded RNA (EBER) was detected in NPC cells at 100x. (c, d), The expression of CD55 was detected by Western blotting and RT‒PCR. (e), Flow cytometric analysis of the cell surface expression of CD55 in EBV-negative (HK1 and CNE2), EBV-positive (HK1-EBV and CNE2-EBV) and normal nasopharyngeal epithelial (N2 and NP69) cells. (f), Representative NPC tissue sections stained with H&E, EBER and CD55. Black scale bars, 10 μm. (g), IHC scores of the indicated genes in NPC biopsies. Means ± SDs, N = 81, ns: not significant, ***p < 0.001

EBV-induced upregulation of CD55 reduces complement-mediated cytotoxicity of cetuximab in NPC cell lines

Considering that the EGFR expression level is critically important for the efficacy of cetuximab in NPC, we first detected the expression of EGFR via Western blotting and flow cytometry. The HK1 cell lines, both in the wildtype and EBV⁺ NPC cell lines, can express EGFR protein (Supplementary Fig. 3a). Though the level of surface EGFR expression detected by flow cytometry was slightly increased after EBV infection (Supplementary Fig. 3b), the elevated efficacy of cetuximab (anti-EGFR antibody) was not detected, suggesting the EBV-induced upregulation of CD55 may play an important role in the resistance of cetuximab treatment. We next incubated NPC cells/EBV infected NPC cells with human AB serum containing complement and cetuximab, the complement-mediated death of the cells was then evaluated via laser scanning confocal microscopy (Fig. 2a). After coincubation, cell membrane perforation was observed in the EBV -negative NPC cells, but not observed in the EBV-infected NPC cells (Fig. 2b). Flow cytometry further revealed that the apoptosis rate of the EBV-infected NPC cells was significantly lower than that of the EBV-negative cells. However, after inactivating the serum, the difference between the two groups was eliminated (Fig. 2c and d). The Western blotting results confirmed that the expression of CD55 in NPC cells infected with EBV was successfully knocked out (Fig. 2e). Flow cytometry results revealed that after stable knockout of CD55, complement inhibition in EBV-infected cultures was reversed (Fig. 2f). These data suggest that the overexpression of CD55 induced by EBV infection is associated with complement dysfunction in NPC cell lines.

Fig. 2.

Complement-mediated cytotoxicity is inhibited by CD55 in EBV infected NPC cells. (a), Representative images of apoptosis detection by immunofluorescence. Black scale bars: 100 μm. (b), Apoptosis in NPC cells with or without EBV infection was assessed by immunofluorescence. (c), Flow cytometry analysis was used to detect cell apoptosis in the above NPC cells. (d), Statistical analysis of the flow cytometry analysis results. Means ± SDs, N = 3. (e), CD55 expression in empty vector- and CD55-knockout EBV + NPC cells was detected via Western blotting. (f), Apoptosis in empty vector- and CD55-knockout EBV-infected NPC cells. Black scale bars, 10 μm. ns: not significant, *p < 0.05; ns: not significant

Cetuximab-induced NK cell cytotoxicity is inhibited by CD55 in EBV-infected NPC cells

We subsequently investigated whether increased levels of CD55 resulted in the inhibition of cetuximab-induced NK-cell cytotoxicity. After coculture with cetuximab, human AB serum and NK cells, cytotoxicity was evaluated via flow cytometry by assessing the degranulation of NK cells and direct lysis of NPC cells (Fig. 3a). The phenotype of the NK cells was identified before they were added to the coculture (Fig. 3b). Depletion of CD55 in EBV positive NPC cells increased Granb, Perforin and IFN-γ expression in NK cells (Fig. 3c, d and Supplementary Fig. 1). In line with these findings, CD55 depletion improved NPC cell apoptosis (Fig. 3e). We next investigated whether the NK cell cytotoxicity was complement dependence. After incubated with inactive serum (without complement), NK cell and cetuximab, we conducted flow cytometry to evaluate the NK cell cytotoxicity. The results revealed that though the cetuximab-mediated cell death was reduced in the absence of complement, the significant differences were still detected among the groups, suggesting the NK cell cytotoxicity was noncomplement-dependent in EBV-infected NPC (Supplementary Fig. 4).

Fig. 3.

NK cell activation is inhibited by CD55 in EBV-infected NPC cells. (a), Experimental model diagram of how CD55 inhibits NK-cell functions as observed in the in vitro studies. (b), Flow cytometry analysis for NK cell phenotypic identification. (c), Representative flow cytometry results of the expression of cytokines secreted by NK cells. (d), The amount of IFN-γ in the supernatant was determined by ELISA. Means ± SDs, N = 3. (e), Flow cytometry analysis of tumor cell apoptosis after coincubation of NK cells with cetuximab. *,p < 0.05; **,p < 0.01; ***p < 0.001

The EBV/LMP1/NF-κB signalling pathway enhances CD55 expression

Previously, we demonstrated that the inhibition of NF-κB activity enhances the sensitivity of tumours to either chemotherapy or radiotherapy in NPC [27]. To assess whether EBV infection influences NF-κB activity, we first performed immunofluorescence and found that EBV infection improved the nuclear translocation of p65 (Fig. 4a). In addition, we detected a positive correlation between the number of p65-positive cells and the CD55 staining score of tumour cells in NPC tissues (Fig. 4b and c). We further detected the phosphorylation status of NF-κB downstream genes (P65, P50 and p-IKK) via western blotting, and found that EBV/NF-κB activation promoted CD55 expression in NPC. Moreover, an NF-κB inhibitor (ammonium pyrrolidine dithiocarbamate, PDTC), which can effectively inhibit NF-κB and block NF-κB translocation to the nucleus, reduced CD55 expression in EBV-infected cell lines (CNE2-EBV and HK-EBV) (Fig. 4d). We next attempted to identify the viral genes that contribute to the activation of NF-κB after EBV infection. Previous studies have shown that the latent viral oncogenes LMP1 and LMP2a are expressed in EBV-infected NPC cell lines [28]. We have previously established NPC cell lines in which LMP1 or LMP2a was overexpressed or LMP1 was knocked out and verified its stable expression through Western blotting [22, 23]. Here, we found that the overexpression of LMP1 in NPC cells (CNE2 and HK1) led to increases in CD55 expression, and the same association between LMP1 and CD55 was detected in the C666-1 cell line (Fig. 4e). EBV-encoded LMP1 is believed to play an important role in NPC pathogenesis and immune evasion and contributes to the development of the NPC tumour microenvironment [29, 30]. A previous study demonstrated that very low levels of LMP1, even those which are virtually undetectable, are sufficient to activate NF-κB signalling, a key pathway contributing to LMP1-mediated cell growth and survival [31]. In the present study, the positive association between CD55 and LMP1 was further detected at the RNA level via RT-PCR (Fig. 4f). Therefore, we blocked the NF-κB pathway with gradient concentrations of the kinase inhibitor PDTC in LMP1-overexpressing NPC cells, and the results revealed a tendency toward a decrease in the expression of CD55 (Fig. 4g). We further detected the surface expression of CD55 via flow cytometry. The results revealed the same decrease of CD55 expression after blocking the NF-κB pathway in LMP1-overexpressing NPC cells (Fig. 4h). These data indicate that LMP1 participates in the regulation of CD55 expression through the NF-κB signalling pathway after EBV infection.

Fig. 4.

The EBV/LMP1/NF-κB signalling pathway enhances CD55 expression. (a), Immunofluorescence detection of the nuclear translocation of p65 in EBV⁻ and EBV⁺ NPC cell lines. (b), NPC tissues were subjected to IHC with antibodies targeting human CD55 and nuclear p56. (c), Statistical analysis of nuclear p56 staining and CD55 expression in tissues from NPC patients treated with cetuximab. N = 168. (d), Western blotting was performed with the indicated antibodies in cytoplasmic extracts and nuclear extracts after incubation with gradient concentrations of the NF-κB inhibitor PDTC (0 mmol/L, 25 mmol/L and 50 mmol/L). (e), Western blotting and (f) RT‒PCR comparing the level of CD55 in empty control-, LMP2a/LMP1-overexpressing NPC (CNE2 and HK1), empty control- and sgLMP2a/LMP1-C666-1 cells. (g) Western blotting of the cytoplasmic levels of CD55 and the nuclear levels of p65 and p50 in empty control- and LMP1-overexpressing NPC cells after treatment with various concentrations of PDTC (0 mmol/L, 25 mmol/L and 50 mmol/L). (h), Flow cytometry analysis of surface expression of CD55 in the above cells. *, p < 0.05; **, p < 0.01; ****p < 0.0001; ns: not significant

Inhibition of CD55 restores cetuximab-induced ADCC activity and shows therapeutic potential in vivo

We inoculated NCG mice (NK-cell-deficient) with EBV-positive HK1 cells with or without CD55 knockdown to explore the potential effects of inhibiting CD55 on the sensitivity of EBV-positive tumours to cetuximab. Primary human NK cells were injected intravenously. Ten days after tumour cell inoculation, the mice were injected with IL-15 every day for 1 week and with IL-2 every other day for 21 days to support the survival of the human NK cells. Moreover, cetuximab (0.1 mg) was injected intraperitoneally weekly to induce tumour apoptosis (Fig. 5a). Although treatment with NK-cell reinfusion alone exerted a nonsignificant antitumour effect, the combination of cetuximab and NK-cell reinfusion significantly reduced tumour growth in HK1-EBV tumour-bearing mice (Fig. 5b-d). Moreover, compared with that in the control group of HK1-EBV-Vector tumour-bearing mice, tumour growth was significantly limited following CD55 knockout in HK1-EBV tumour-bearing mice treated with the combination treatment. In addition, NK-cell expression of Granb and Perforin was significantly increased in CD55-knockout xenograft tumours (Fig. 5e and f). Notably, no significant difference of CDC-mediated cytotoxicity was observed in cetuximab treated B-NDG Hc mice models (Supplementary Fig. 5a). However, the combined activity of complement and ADCC was significantly stronger than the activity of ADCC alone (Supplementary Fig. 5b-d), suggesting that complement activity may attribute to promote NK-cell function, thus sensitizing tumours to cetuximab.

Fig. 5.

Inhibition of CD55 restores sensitivity to cetuximab in an NCG mouse model. (a), Schematic of in vivo studies using HK1-EBV-Vector or HK1-EBV-sgCD55 in an NCG xenograft mouse model treated with NK cells reinfused with or without cetuximab treatment or a negative control, along with cytokine administration. Growth curves (b), weight (c) and images (d) of NPC xenograft tumours from NCG mice. (e, f) Flow cytometric analysis of the expression of cytokines secreted by NK cells. (g), Mechanistic diagram of how EBV-enhanced CD55 regulates complement-mediated cytotoxicity and NK cell cytotoxicity to inhibit the efficacy of cetuximab. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns: not significant

CD55 overexpression is associated with cetuximab resistance in NPC patients

A previous study revealed that increased expression of CD55 was correlated with tumour progression and poor prognosis in NPC [32]. However, only 80 NPC patients were included in that study. In the present study, 215 NPC patients treated with cetuximab at Sun Yat-Sen University Cancer Center between October 2008 and October 2015 were retrospectively reviewed. We then conducted propensity score matching (PSM) to exclude the impact of preexisting confounding factors, such as sex, age, clinical stage and treatment. After PSM, we selected a total of 168 patients and conducted survival analysis in this well-balanced large cohort. The baseline demographic and clinical characteristics were well balanced between the two groups (Fig. 6a) and were not significantly different between the CD55 low- and high-expression groups (Supplementary Table 1). Moreover, CD55 overexpression was associated with poor overall survival and progression-free survival in NPC patients treated with cetuximab (Fig. 6b, c). The results of the quantitative analysis via multi-immunofluorescence staining also revealed that the number of cytotoxic NK cells (CD56⁺CD3⁻GranzymeB⁺) was significantly decreased in the biopsies from patients with high CD55 expression in NPC patients (Fig. 6d, e). These data further support that NK-cell cytolytic activity is directly inhibited in CD55 overexpressing NPC patients treated with cetuximab.

Fig. 6.

NK-cell cytolytic activity is inhibited in CD55 overexpressing NPC patients treated with cetuximab. (a), Propensity score matching (PSM) was used to exclude the impact of preexisting confounding factors, such as sex, age, clinical stage and combination treatment strategies, in the NPC cohort treated with cetuximab. (b), Kaplan–Meier survival curves of OS and (c) PFS in the matched NPC cohort (n = 168). (d), Images of cytotoxic NK cells (PanCK^– CD3– CD56⁺ GranB⁺ cells) in CD55 high and CD55 low expression biopsies assessed by multi-immunofluorescence staining. Scale bars: 100 μm. (e), Statistical analysis of cytotoxic NK cells in NPC tissues with high CD55 expression and low CD55 expression. ***, p < 0.001

Methods

Cell culture and cell culture conditions

The NPC cell lines TW03, CNE2, HK1, and C666-1 were a kind gift from Prof. Musheng Zeng (Sun Yat-sen University Cancer Center) and were cultured in RPMI 1640 with 10% FBS (Gibco). The EBV⁺ cell lines were maintained in RPMI 1640 medium supplemented with 700–900 µg/ml G418. Latent EBV infection in the cell lines was confirmed via in situ hybridization of EBERs with a kit (Zhongshan Jinqiao). All the cell lines were routinely tested by PCR at the beginning of the present study [22]. All the cells were tested to ensure that they were free from Mycoplasma infection.

Whole-transcriptome sequencing

Total RNA from the two EBV⁻ NPC cell lines (CNE2 and TW03), the EBV⁺ NPC cell lines (CNE2-EBV and TWO3-EBV) and the EBV⁺ NPC cell lines after EBV genome destruction (#1 and #2) was isolated with TRIzol and purified using the RNeasy Plus Micro Kit (Qiagen) according to the manufacturer’s protocol. Libraries preparation for RNA sequencing was performed using the Illumina HiSeq4000 platform (Novogene Genomic Sequencing Center, Beijing, China). The RNA-seq sequence reads were aligned to the reference genome using STAR (v2.5.1b). Differential expression analysis was performed using the DESeq2 R package (1.10.1). After alignment, the RNA-seq data were converted to gene expression data as fragments per kilobase of exon per million fragments mapped (FPKM) values via HTSeq v0.6.0. To control the false discovery rate, the P-values of the results were adjusted using the Benjamini–Hochberg method. Genes with an absolute fold change of two and an adjusted p-value (Padj) < 0.05 were set as the threshold for significantly differential expression.

Clinical specimens

This study was approved by the Institutional Ethical Review Board of Sun Yat-sen University Cancer Center (Approval No.GZR2020-093). All patients provided oral informed consent. In group one, 81 samples were obtained from Sun Yat-sen University Cancer Center, Guangzhou, China, from 2010 to 2015, including 65 EBV-positive samples, 6 moderate EBV-positive samples and 10 EBV-negative NPC samples. This group was used to analyse the association between EBV and CD55 expression. In group two, 215 NPC patients treated with cetuximab at Sun Yat-Sen University Cancer Center between October 2008 and October 2015 were retrospectively reviewed. All the patients had primary NPC tumours. Propensity score matching (PSM) was conducted to exclude the impact of preexisting confounding factors, such as sex, age, clinical stage and treatment. After PSM, 168 patients were selected for further analysis. We subsequently conducted a survival analysis and evaluated the association between the number of cytotoxic NK cells and CD55 expression in this well-balanced cohort.

Real-time PCR, Western blotting and immunohistochemical analyses were performed as previously described [33].

Generation of EBV-infected cells

After being cocultured with GFP-expressing Akata EBV for 24 h, the infected cells (HK1-EBV, CNE2-EBV and TW03-EBV) were sorted by flow cytometry (Beckman) and maintained in selection medium containing G418. Latent EBV infection in the cell lines was detected with an EBER ISH kit (Zhongshan Jinqiao) according to the manufacturer’s instructions.

Complement-mediated cytotoxicity

Normal human serum (NHS) from human male AB plasma was purchased from Sigma-Aldrich (H4522) and used as a source of complement. As a negative control, we inactivated complement by heat treatment with NHS (30 min at 56 ℃). Subsequently, 2 × 10⁶ NPC cells were mixed with cetuximab (40 µg/mL) to a final volume of 50 µL for 30 min at 4 ℃. Fifty microlitres of human serum, inactivated serum or PBS was added, and the mixture was incubated for 1 h at 37 ℃ in a 5% CO₂ incubator. After washing, immunofluorescence and flow cytometry analyses were conducted to determine the percentage of dead cells. All experiments were performed in triplicate for each condition. The results were combined and statistically analysed via Student’s t- test.

Complement-mediated NK-cell cytotoxicity assays

Umbilical cord blood-derived NK cells were kindly provided by the China Cord Blood Bank of Shandong Province. Enriched NK cells (purity > 95%) were isolated by negative selection using an NK cell isolation kit (cat# 130-092-657, Miltenyi Biotec). The target NPC cells were labelled with CSFE (FITC channel; 65085084, eBioscience) and cultured with human serum (2 × 10⁶/100µl) for 1 h at 37 ℃ prior to the addition of NK cells (NPC: NK cells = 1:50 cell ratio) and cetuximab (40 µg /mL). After incubation for 5 h at 37 °C in a 5% CO₂ incubator, CD3⁻CD56⁺ NK cells were gated and analysed for the expression of surface perforin (BD Biosciences) and Granb (BD Biosciences) using fluorescent tagged antibodies. NK-cell cytotoxicity was further determined by the percentage of dead NPC cells as described in our previous study [23]. Briefly, labelled NPC cells were stained with the viability probe 7-AAD (PC5.5 channel; 00699350, Invitrogen) after incubation with NK cells and cetuximab. The gating strategy is shown in Fig. 3e-h; (e and f) NPC cells were gated on forwards (FSC) and side scatter plots (SSCs), and (g) NPC cells were identified as CFSE⁺ cells. (h) The percentage of 7-AAD -positive NPC cells was calculated as the percentage of dead NPC cells. All samples were measured in triplicate per experiment.

Flow cytometry

Single-cell suspensions prepared from cocultures in vitro were immunostained in cell staining buffer for 30 min at 4 ℃ in the dark. The data were analysed using FlowJo v10 (Treestar, Ashland, OR, USA).

Nuclear and cytoplasmic extraction

The adherent cells were harvested with trypsin-EDTA and then centrifuged at 1000 × rpm for 5 min. Two hundred microlitres of ice-cold CER I and 11 µl of ice-cold CER II were gradually added to the cell pellet. After the tube was centrifuged at maximum speed in a microcentrifuge (approximately 16,000 × g), the supernatant (cytoplasmic extract) was immediately transferred to a clean prechilled tube. Then, the remaining insoluble (pellet) fraction, which contained nuclei, was suspended in ice-cold NER. The samples were placed on ice and vortexed for 15 s every 10 min for a total of 40 min. After the tube was centrifuged at maximum speed (approximately 16,000 × g) in a microcentrifuge for 10 min, the supernatant (nuclear extract) fraction was collected for subsequent analyses.

IFN-γ ELISA

ELIspot plates (human IFN-γ ELISPOT Set, BD™ ELISPOT) precoated with an anti-IFN-γ antibody were used to evaluate the expression of IFN-γ as described previously [24]. The spots were then enumerated on the plates using Immunospot (Cellular Technology Ltd.; Shaker Heights, OH, USA).

Immunofluorescence

Immunofluorescence and image analyses were performed as described previously [23]. In brief, NPC tissue sections were subjected to microwave treatment in antigen retrieval solution and then incubated with anti-CD3 (1:200, Abcam), anti-Granzyme B (Gra-B) (1:100, CST), anti-CD55 (1:500, Abcam), anti-CD56 (1:100, CST) or panCK (1:4,000, Sigma) antibodies overnight at 4 °C. The sections were then incubated with an HRP-conjugated secondary antibody (Panovue, 10013001010). All the slides were subsequently counterstained with 4′,6′-diamidino-2-phenylindole (DAPI, Panovue, 10012100500) for 10 min and mounted with ProLong Diamond Antifade Mountant (Panovue, 10022001010).

Xenograft model

Female NOD/ShiLtJGpt-Prkdc^em26Cd52Il2rg^em26Cd22/Gpt (NCG) mice (6–8 weeks old) purchased from GemPharmatech Co., Ltd. (Jiangsu, China) were inoculated s.c. with 5 × 10⁶ HK1-EBV or HK1-EBV-derived cells. Several days later, each mouse was intravenously injected with or without 1.5 × 10⁷ NK cells from umbilical cord blood. NK-cell survival was promoted by the injection of hIL-2(10,000 units/mouse every 2 days for 21 days) and hIL-15 (10 ng/mouse for 7 days). Moreover, intraperitoneal injection of cetuximab (0.1 mg/mouse) or an isotype control antibody was performed once a week for three weeks after NK-cell reinfusion in the combination treatment studies. Tumour size and body weight were measured every 4 days. NOD.CB17-Prkdc^scidIl2rg^tm1Bcgen Hc^{tm1(c645insTA)Bcgen}/Bcgen(B-NDG Hc) mice purchased from Biomice (Jiangsu, China) were used to detect the complement mediated cytotoxicity in vivo. This study protocol was approved by the Experimental Animal Ethics Committee of Sun Yat-sen University Cancer Center (Approval No. L102012020060G).

Isolation of immune cells from xenograft tumours

Tumour tissues from the mice were collected on a 6-well plate that contained 3 ml of cold serum-free RPMI 1640. The tissues were cut into small pieces of 2–4 mm with scissors in 50 ml conical tubes and then digested in 5 ml of enzyme mixture. The tubes were transferred to a shaker for 40 min (220 rpm at 37 °C). Subsequently, 10 ml of ice-cold RPMI 1640 medium supplemented with 10% FBS was added to terminate the enzymatic reaction. The cell pellets were filtered through a 70 μm MACS SmartStrainer. After centrifugation at 500 × g for 5 min, the collected cell pellets were resuspended in 6 mL of 40% Percoll and then carefully layered onto 80% Percoll with a Pasteur pipette. The samples were centrifuged at 1,260 × g for 20 min at room temperature, and then the middle interface layer was collected in a new 15-mL tube containing 10 mL of FACS buffer. The mixture was centrifuged at 500 × g for 5 min, the supernatant was discarded, and the cell pellet was collected for flow cytometry analysis.

Statistical analysis

Statistical analyses were performed using GraphPad Prism v9.0. Normalized data were compared by a two-tailed unpaired t test. Data with a nonnormal distribution were compared with a two-tailed Mann–Whitney test. Survival curves were generated via the Kaplan–Meier method and compared using the log-rank test. Two-tailed P < 0.05 was considered statistically significant.

Discussion

EBV-associated NPC is a common head and neck cancer that is highly prevalent in southern China and Southeast Asia [34, 35]. Although NPC is highly sensitive to radiotherapy and chemotherapy, up to 30% of patients with locoregionally advanced NPC still develop metastasis and recurrence [36]. Patients with metastasis and recurrence often succumb to this cancer [37]. Considering that EGFR overexpression is very common in NPC with a detection rate of > 80% [38], cetuximab, an anti-EGFR antibody, has been shown to improve survival in NPC patients [8, 9]. However, a correlation between the expression of EGFR and the response to cetuximab has not been observed [8]. More importantly, only a small percentage of NPC patients benefit from this targeted therapy [39]. Therefore, the mechanism of EGFR-mAbs in the treatment of NPC still needs to be further elucidated, and an in-depth understanding of this mechanism may further increase the survival of NPC patients.

Growing evidence suggests that multiple interacting mechanisms, including complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC), act either alone or, more often, in combination, in the antitumour response of monoclonal antibodies, including cetuximab [11, 12]. CDC activation can lead to the formation of a membrane attack complex inducing lysis of target cells. ADCC activation is mediated primarily by NK cells through binding of the FcγRIIIa (CD16a) receptor with the Fc portion of the antibodies bound to the specific antigen expressed on the target cells. This binding initiates the secretion of granzyme and perforin by NK cells, resulting in lysis of the target cell [40]. Given the complexity of immune activation induced by mAbs, which of these mechanisms is most important for the efficacy of cetuximab-treated EBV-associated NPC remains incompletely understood. In the present study, we found that the EBV-induced upregulation of the complement regulatory protein CD55 reduced cetuximab-induced CDC and ADCC cytotoxicity in vitro. However, a significant difference in cetuximab-induced CDC cytotoxicity was not detected in the in vivo models, indicating that CDC was not the key mechanism that reduces the efficacy of cetuximab. Importantly, the inhibition of CD55 efficiently activated ADCC-mediated cell lysis in the absence of complement. Although surprising, these results obtained in the absence of complement are in agreement with of previous studies [41–43], which suggesting additional novel function for CD55 in tumour immune surveillance that may be unrelated to complement inhibition. Although not specifically tested in our experiments, it is possible, as previously documented by others, that in the absence of serum, CD55 may inhibit the soluble complement fragments and stimulate complement receptors expressed by NK cells [44, 45]. Moreover, the combined activity of complement and ADCC was significantly stronger than the activity of ADCC or CDC alone. In a series of papers, Weiner et al. reported that the deposited complement fragments sterically hinder the interaction of NK cell FcγRIII (CD16) with the Fc region of mAbs [46]. Similarly, Derer S et al. reported that cross-talk between complement and immune cells optimized ADCC cytotoxicity, thus enhancing the efficacy of anti-EGFR mAbs [47]. Such interactions between complement and immune cells may explain the stronger efficacy detected in both in vivo and in vitro models. However, the specific mechanism by which CD55 regulates cetuximab-mediated cell death still needs to be further studied.

As a complement regulatory protein, CD55 has been detected in many solid tumours, including NPC, and is expressed at higher levels than in normal cells and promotes malignant progression [19–21]. The complement regulatory protein can protect cells against complement activation and NK-cell -induced cell death [48]. Previous studies have demonstrated the downregulation of CD55 by siRNA sensitizes breast cancer [49] and uterine serous carcinoma [50] to antibody-based immunotherapy. Further study revealed that the combination treatment involving cetuximab and NK-cell reinfusion in HK1-EBV NPC cells after CD55 knockdown markedly enhanced the efficacy of cetuximab, suggesting that CD55 may function as a novel immune checkpoint molecule in EBV-associated NPC. Whether this treatment benefit holds true in EBV-associated NPC patients still need further investigation.

EBV oncogenesis is principally associated with the latent phase in infected cells [51]. Generally, latent membrane protein 1/2 (LMP1 and LMP2) are the main latent proteins and have been shown to have oncogenic properties [6, 52]. LMP1 is a member of the tumor necrosis factor receptor (TNFR) superfamily, which is known as an inflammation-promoting factor inducing a panel of pro-inflammatory cytokines via NFκB, AP1, and STAT3 [53–55]. Recently, LMP1-initiated NF-κB activation has been shown to upregulate the expression of cell surface antigens (CD95, CD54 and CD40) and proinflammatory cytokines [56]. The constitutive activation of NF-κB plays a pivotal role in NK cell function, innate immunity and malignancy-associated genes [57, 58]. Therefore, we first investigated the nuclear expression of NF-κB p65 by immunofluorescence and found that the translocation of p65 to the cell nucleus in EBV⁺ NPC cells was significantly greater than that in EBV^– NPC cells. The same result was detected in the nuclear extracts of EBV⁺ NPC cells via IHC and Western blotting, further supporting the hypothesis that LMP1-NF-κB upregulates another member of the cell surface antigen, CD55. To further verify the role of LMP1-NF-κB-CD55 signalling, we used the NF-κB inhibitor PDTC to block its activation in LMP1-overexpressing NPC cells. As expected, we found that the increase in CD55 expression could be abolished by PDTC treatment, highlighting the importance of the LMP1-NF- κB pathway in regulating CD55 expression.

In conclusion, the results of our study demonstrate a novel molecular mechanism in which EBV latent membrane protein 1 (LMP1) mediates the upregulation of CD55 through the NF-κB signalling pathway. The upregulation of CD55 directly inhibited NK cell–mediated antitumour activity. Moreover, overcoming the immunosuppressive role of CD55 may be an efficient strategy for improving the efficacy of cetuximab in EBV-associated NPC.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

NPC

Nasopharyngeal carcinoma

EGFR

Epidermal growth factor receptor

ADCC

Antibody-dependent cellular cytotoxicity

CDC

Complement-dependent cytotoxicity

EBNA1

EBV nuclear antigen 1

LMP1

Latent membrane protein1

LMP2

Latent membrane protein2

Author contributions

CMY and TX designed the project and revised the manuscript; QZ, XBD and HH performed the experiments and analyzed the data. QZ drafted the manuscript. RY and TY contributed a lot in the revision and provided scientific advice. TLX offered part of the cohorts’ database and clinical samples. All authors have read and approved the final version of the manuscript.

Funding

This work was supported by National Natural Science Foundation of China (No.82002872, 82172795 and 82230034) and Guangdong Medical Research Foundation, China (A2022142).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Internal Review and the Ethics Boards of the Sun Yat-Sen University Cancer Center. Written informed consent was obtained from all subjects. All animal experiments were conducted in agreement with the Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of the Sun Yat-Sen University Cancer Center, Sun Yat-Sen University.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.