The TLR3 Agonist Inhibit Drug Efflux and Sequentially Consolidates Low-Dose Cisplatin-Based Chemoimmunotherapy while Reducing Side Effects

In clinical, it is gradually recognized and accepted that the strategies targeting the tumor microenvironment (TME) for the activation of immunocytes or the abrogation of the interaction between tumor and stroma could complement traditional chemotherapy or radiotherapy and improve clinical benefit further (1). Pathogen-associated molecular patterns with a high capacity to boost the immune response of innate immune system are investigated as adjuvants in cancer vaccines for immunotherapy, especially the toll-like receptors (TLR; ref. 2). Polyriboinosinic–polyribocytidylic acid (poly I:C, PIC) and its derivative poly-ICLC (Hiltonol) are the agonist of the endosomal TLR3. Notwithstanding, in published clinical studies, the TLR3 agonist as cancer vaccine adjuvant is safe and well tolerated to activate the immune system (3), but poly-ICLC as stand-alone treatment have no or limited significant activity of antitumor (4, 5) and produce significant toxicities (2). In addition, accumulated studies have demonstrated that the different ways of administration, including intravenous, intra- or peritumoral, intramuscular injection or transfected with poly(I:C) (PIC) contribute to the disparity of therapy efficacy (6–10). Previous clinical trials have indicated that glioblastoma patients pretreated with immunotherapy and followed by chemotherapy show better clinical outcomes than those received chemotherapy alone (11). Notably, once initial tumor control is achieved by regular dosage of chemotherapy, the evolution-based therapeutic strategy with low-dose chemotherapeutic drug could inhibit the tumor progression and prolong the survival of mice with breast cancer (12). Therefore, the optimal dosage and combination strategies of TLR3 agonist and chemotherapeutic drugs should be further investigated (13).

Chemoimmunotherapy is developed by the combination of immunotherapy and anticancer agents. In hepatocellular carcinoma, administration of TLR3 agonist together with sorafenib activates NK cells, T cells, macrophages, DCs but impairs the infiltration of myeloid-derived suppressor cells (MDSC), as well as directly increases tumor apoptosis independently of immune system to restrict tumor growth in immunodeficient NOD/SCID gamma (NSG) mice (14), suggesting that immune system might be a partial attributor to the TLR3 agonist–induced chemoimmunotherapy. The ATP-binding cassette (ABC) transporter family including P-gp (ABCB1) and MRP-1(ABCC1) etc., have been found to be involved in the development of multidrug resistance (MDR) in cancer treatment though the transportation of many anticancer drugs (e.g., cisplatin) through cellular membranes, limiting the cytoplasmic residence (15–19). However, the correlation between ABC transporter family and poly (I:C) in cisplatin-based chemotherapy is totally unknown.

Long noncoding RNAs (lncRNA) have been discovered to participate in the tumor initiation, development and metastasis, but its role in chemotherapy resistance is unclear (20). In this study, we, for the first time, uncovered that preculture tumor with poly (I:C) sensitized tumors to the low-dose cisplatin-based chemotherapy via downregulation of ABC transporter family and targeting the TME in vitro and in vivo. Moreover, as a humanized monoclonal antibody targeting the EGFR, cetuximab has been applied in the chemotherapy of head and neck squamous cell carcinoma (HNSCC), including OSCC. We here identified that the addition of cetuximab in this new strategy efficiently enhanced the complete response (CR) ratio with no significant adverse side effects in OSCC. Furthermore, Lnc-IL7R was determined as a factor of chemotherapy resistance in this strategy.

We here constructed a cohort including 30 normal oral tissues, 30 leukoplakia tissues, 166 primary OSCC samples that were retrospectively collected to determine the protein expression of TLR3 in OSCC. Forty matched normal-malignant fresh tissue samples collected from surgery were used for TLR3 mRNA analysis and cancer-associated fibroblasts (CAF) isolation. All of the patients diagnosed with primary OSCC were confirmed by hematoxylin and eosin staining by experienced pathologists from the Department of Pathology at Nanjing Stomatology Hospital. The ethical approval for this study was obtained from the Research Ethics Committee of Nanjing Stomatology Hospital. Patients who were diagnosed with autoimmune or other malignant diseases and pregnant or lactating individuals were excluded from in this study. No patients underwent preoperative chemotherapy and/or radiotherapy. All of the TSCC tissues were evaluated according to WHO classifications and International Cancer Control (UICC) tumor–node–metastasis (TNM) staging system.

Immunohistochemistry and immunofluorescence assays are performed according to our previous study described (21). The IHC scores of TLR3 in tumor nest were indicated by the scores and in fibroblast was indicated by the percentage of positive cells. The positive staining was defined per specimen by a 25% of median expression threshold. Thus, the IHC sores of tumor >4 and the IHC sores of fibroblast >10% were regarded as positive expression.

The human OSCC cell lines HSC-3, OSCC3, and SCC-14a/b were obtained from Professor Yvonne L. Kapila (Michigan University, USA) in 2011. The human OSCC cell line SCC-4, SCC1, CAL27, HIOEC-B (HB) cells and the normal human keratinocyte line HaCaT were kindly provided by Shanghai Ninth Hospital (Shanghai, China) in 2011. The cells were characterized by mycoplasma detection, DNA fingerprinting, isozyme detection, and cell vitality detection by the provider. No further authentication of cell lines was conducted. All cells were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% FBS (Life Technologies) at 37°C, 5% CO₂ condition. Six- to 8-week-old female C57BL/6 mice and Balb/c mice were purchased from Model Animal Research Center of Nanjing University. Five- to 6-week-old male BALB/c-nu/nu T cells-deficient mice were purchased from Cavens (Changzhou, china). Detailed information on the antibodies and reagents used in this study is provided in Supplementary Table S1.

Therapeutic protocols of tumor models and all other methods were described in Supplementary Materials and Methods.

Statistical analyses were performed on Statistical package for social sciences version 16.0 (SPSS 16.0, SPSS Inc.) and Prism statistical software package (GraphPad Software Inc.). The relationships between the presence of TLR3 and clinicopathologic characteristics were determined by Spearman χ² tests. Unpaired t test or Mann–Whitney U test was used to compare the two groups and the differences between more than two groups were analyzed by the Kruskal–Wallis test etc. Differences were considered statistically significant with P < 0.05.

Recently, the programmed death ligand-1 (PD-L1) checkpoint blockade has been demonstrated to be produced objective responses in patients with non–small cell lung cancer, melanoma and renal cell cancer (22). As the most critical facts, the expressions of PD-L1 in TME including tumor cells and tumor-infiltrating immune cells (TIL) manipulate the objective response rate.

To investigate the feasibility of the application of TLR3 agonist, we here collected a larger cohort including 30 normal oral tissues, 30 leukoplakia tissues, 166 retrospective primary OSCC samples and 40 matched normal-malignant fresh tissue samples for estimating the expression of TLR3 in OSCC. Immunohistochemistry (IHC) analysis showed that the TLR3 expression was absent in the normal epithelial part but relatively increased in the paracancerous leukoplakia tissue while the OSCC tissue harbored a strong positive stain in tumor cells, TILs and CAFs (Fig. 1A). TLR3 positivity was defined per specimen by a 25% expression threshold of IHC scores and we found that positive TLR3 expression (the IHC score > 4) was widely present in 115 of 166 OSCC patients (69%) and 123 of 166 (74%) patients showed positive expression of TLR3 in CAF (Fig. 1B). Moreover, the mRNA level of TLR3 was also identified to be upregulated during the carcinogenesis in the 40 matched normal-malignant fresh tissue samples (Supplementary Fig. S1).

Furthermore, we analyzed the correlations between aberrant TLR3 expression and the clinicopathological characteristics of OSCC patients. Significant higher expression of TLR3 was found in patients with higher tumor–node–metastasis (TNM) stage, lymph node metastasis and poor differentiation, but high TLR3 expression in CAF predicted a good clinical outcome, which was conversed in tumor cells (Supplementary Table S2). These data underlined the function heterogeneity of TLR3 in TME, which was similar to TLR7 in OSCC as our previous findings (23). Taken together, the overexpressed TLR3 expression in OSCC provided a foundation of TLR3 agonist-based therapy.

Clinical data have proved that increasing dose intensity of cisplatin (CDDP or DDP) does not improve the overall survival (OS) compared with standard chemotherapy but aggravate the adverse events (AE) for advanced epithelial ovarian cancer (24).

To overcome these obstacles, we proposed the possibility that TLR3 agonist-based therapy is a promising strategy for the complements of cisplatin. First, we evaluated the sensitivity of four OSCC cell lines SCC1, SCC4, OSCC3, and CAL27. Series concentration of cisplatin (μmol/L) were set for estimating the corresponding inhibition rate of cell vitality (Fig. 1C) and the results showed that the SCC4 cell line was less sensitive to cisplatin with the higher IC₅₀ value of cisplatin [8.532, 95% confidence interval (CI), 7.327–9.935] than OSCC3 (5.138, 95% CI, 4.109–6.426). We further explored the TLR3 expression in four cell lines and found that the sensitivity to cisplatin might negatively correlate to the TLR3 level in tumor cells, because the TLR3 expression in SCC4 is much higher than that in OSCC (Fig. 1D). Second, TLR3 agonist poly (I:C) (20 μg/mL), which has no significant cytotoxicity (Supplementary Fig. S2), was used in combination with different doses of cisplatin, including IC₂₅, IC₅₀, and IC₇₅ for the CCK-8 assay, namely the “DDP+PIC” group (Fig. 1E). We found that the combined efficiency reached maximization and statistical significance in the IC25 cisplatin in SCC1, SCC4, and CAL27, but not the high-dose IC₅₀ value of cisplatin, whereas the IC₇₅ value of cisplatin had no combined affects with poly (I:C), which could be explained by the high cell vitality inhibition of the IC₇₅ value of cisplatin only (Fig. 1F–H). However, this combined effect was not observed in OSCC3 cell line with the relatively low TLR3 expression (Fig. 1I).

Unexpectedly, other than the optimized dose of cisplatin, the sequence of administration was found to be an interesting factor in combined chemotherapy (11). When the tumor cells were precultured with 20 μg/mL poly (I:C) for 12 h followed by co-stimulation with poly (I:C) and the IC25 cisplatin for 48 h, namely the “PIC+DDP” group (Fig. 1E), the combined effects were remarkably elevated when compared to the tumor cells without the pretreatment of poly (I:C) in “DDP+PIC” group (Fig. 1F–H) and this inhibition rate was equal to the IC₅₀ value of cisplatin-induced inhibition rate. These findings strongly implicate that the orderly utilization of poly (I:C) consolidate the low-dose cisplatin-induced chemotherapy.

Of note, this orderly administration of poly (I:C) and the low-dose cisplatin could also show the combined effect in OSCC3 cell lines which failed to response to the combination without the pretreatment of poly (I:C) (Fig. 1I), suggesting that pretreatment of poly (I:C) could make the tumor cells with low expression of TLR3 sensitive to the combination of poly (I:C) and low-dose cisplatin.

To confirm the role of poly (I:C) in the sensitivity of tumor cell to cisplatin, we next precultured the SCC4 cells with poly (I:C) for indicated time and then the combined chemotherapy was performed (Fig. 2A). The results indicated that the inhibition rate was positively related to the preculture time, which demonstrated that poly (I:C) could induce tumor cell more sensitive to cisplatin.

Multiple drug resistance (MDR) is a pivotal characteristic of chemotherapy resistance. ATP-binding cassette (ABC) transporters (e.g., P-gp and MRP-1) could participate in the MDR by transporting the endogenous metabolites and xenobiotics including chemotherapy drugs out of cells (10). In this study, we found that the treatment of poly (I:C) in different concentration and time could significantly repress the majority of ABC transporters expression, including P-gp, ABCB4, MRP-1, ABCC2, CFTR, ABCC8, ABCD2, ABCE1, ABCG1, and ABCG2 (Fig. 2B; Supplementary Fig. S3A and S3B). Besides, the activity of P-gp in the drug efflux was determined. Rho-123 as a substrate for P-gp was used to indicate the activity of P-gp by flow cytometry. Poly (I:C)-treated SCC4 cells showed a decreased P-gp expression and an increased fluorescent intensity of Rho-123 which might be explained by the increased intracellular Rh-123 accumulation. Verapamil was used as a positive control to inhibit the activity of P-gp (Fig. 2C; Supplementary Fig. S3C). Moreover, the P-gp and MRP-1 expression were also found to be downregulated in response to the combined chemotherapy (i.e., “PIC+IC25” group; Fig. 2D).

To provide direct evidence that poly (I:C) make tumor sensitive to cisplatin-based chemotherapy, we precultured SCC4 cells with poly (I:C) for the subsequent combined chemotherapy for one or 2 days. The culture medium containing free cisplatin was totally removed and the atomic absorption spectrographic assay was performed to measure the average Pt concentration (μg/L) in tumor cells (Fig. 2E). We found that the average Pt concentration in “PIC+IC25” group was higher than that in the group treated with the IC₂₅ value of cisplatin only. Furthermore, the tumor cells treated with orderly chemotherapy were replaced with fresh culture medium and were allowed to cisplatin efflux for additional 2 days and then Pt concentration was measured again. The results showed that the cells in the “PIC+IC25” group still harbored detectable Pt concentration when compared to the “IC₂₅” group (Fig. 2E). These results indicate that poly (I:C) could impair the activity of drug efflux by inhibition of ABC transporters in tumor cells for the low-dose cisplatin-based chemotherapy.

We next aimed to elaborate the characteristics of this chemotherapy strategy. Three important features were found in vitro. First, the orderly combination Poly(I:C) and DDP induced a significant G₂ cell-cycle arrest in day 2 but a delayed remarkable apoptosis in day 4 in SCC4 cells, which suggested that this strategy was suitable for long-time low-dose chemotherapy for tumor patients with poor physical conditions (Supplementary Fig. S4A and S4B). Moreover, we found that knockdown of TLR3/TRIF signals, but not the MYD88 or RIG-1 signals, could restore the Poly(I:C)/DDP-induced inhibition of cell vitality, which indicated that our combined chemotherapy is TLR3-dependent (Supplementary Fig. S4C). The mechanism of apoptosis was investigated, and the results indicated that Poly(I:C)/DDP induced significant activation of PARP and caspase-3. The inhibitor of caspase-3 Ac-DEVD-CHO could inhibit its activation and impaired the Poly(I:C)/DDP-induced apoptosis (Fig. 2F).

Second, we focused on the infiltration of lymphocytes in the tumor chemotherapy. Spleen lymphocytes were isolated and cultured and the activation of lymphocytes were assessed. The CD3⁺ CD8⁺ T cells and the activated CD11c⁺ MHCII⁺ dendritic cells (DC) were significantly increased in the “PIC+IC25” group although there exited the low-dose cisplatin (Fig. 2G). Moreover, because the tumor cells in vivo dynamically interact with TME including immunocytes, the splenocytes were directly co-cultured with tumor cells. We found a direct inhibition of tumor cells with no treatment in the co-culture system, and the orderly combination Poly(I:C) and DDP could further impair the proliferation and induce more apoptosis (Supplementary Fig. S4D and S4F).

Third, as the most important infiltrated non-lymphocytes in stroma, CAFs use paracrine and reciprocal cancer-stromal communication network to stimulate tumor proliferation and metastasis (25). We hypothesized that the chemotherapy strategy could also effectively impair the interaction between CAF and tumors. Then two CAFs were isolated from two OSCC patients. The phenotypic and functional characterizations were performed using α-SMA/FSP-1 as the CAFs markers (Fig. 2H) and the pan-cytokeratin as the marker of epithelial cells (Supplementary Fig. S4G). Although CAFs promoted the tumor cell proliferation, the orderly chemotherapy strategy effectively dampened the CAFs-supported proliferation of tumor cells (Fig. 2I).

To provide the evidence of the efficiency of this strategy in vivo, xenograft model of human OSCC was established. We evaluated the responses of SCC4 tumors to two different concentration of Poly(I:C), which were reported to activate the immune system. No significant tumor regression was observed in different groups and the administration of increased dose of Poly(I:C) was associated with significant aggravating weight loss in mice treated with 50 μg/dose Poly(I:C) (Fig. 3A and B), although it could significantly activate the spleen immunocytes including CD86⁺ CD11c⁺ DCs and CD69⁺ CD49⁺ NKs (Fig. 3C). These findings implicate that the activation of immune system alone is not sufficient to eliminate tumors.

We next evaluated the efficiency of the combined chemotherapy. As the low-dose cisplatin in clinical trials is 10 to 20 mg/m², the corresponding dose used in mice was 5 mg/kg which was used as high dose of administration (i.e., the “DDP⁵” group) and the 2.5 mg/kg was used as low dose (i.e., the “DDP^2.5” group). The combination group consisted of 25 μg/dose Poly(I:C) and 2.5 mg/kg low-dose cisplatin and the mice in this group were treated with 25 μg/dose Poly(I:C) one day in advance (Fig. 3D). We observed a significant tumor remission in the combination group than that in the “DDP^2.5” group. Importantly, this efficiency of the “Poly(I:C)²⁵+DDP^2.5” group was equal to which in the high-dose “DDP⁵” group (Fig. 3E). However, the mice treated with low-dose cisplatin in the “DDP^2.5” group only showed no significant reduction of tumor volume. Cisplatin has a toxicity profile characterized by nausea and vomiting, renal toxicity and neurotoxicity. The “DDP⁵” group had a remarkable weight loss and enhanced level of Creatinine (Cr) but the “Poly(I:C)²⁵+DDP^2.5” group showed alleviated adverse side effects. Hepatotoxicity was measured by the liver enzyme levels (ALT, AST) that showed no significant changes in this study (Fig. 3F).

Furthermore, to confirm the regulation of Poly(I:C) on the drug efflux, we determined the expression of ABC transporters in tumor samples. The combination chemotherapy inhibited the expression of P-gp, ABCB4, MRP-1, ABCC2, CFTR, ABCC8, ABCD2, ABCE1, and ABCG1 (Fig. 3G). To be noted that the metabolism of cisplatin in body is very fast, using atomic absorption spectrometry, we measured the free platinum concentrations in the tumor samples which were collected 5 days after the last time administration at day 34. A detectable Pt was still found in the tumor samples of the “Poly(I:C)²⁵+DDP^2.5” group but not the other groups after 5 day's metabolism (Fig. 3H). Altogether, 25 μg/dose Poly(I:C) administration could inhibited the cisplatin efflux for the low-dose cisplatin-based chemotherapy in non-direct cytotoxicity manner.

We next identified the impacts of this orderly combined chemotherapy on TME. The poly(I:C)-based cancer vaccines implement antitumor activity have been identified to mainly depend on activating DC, stimulating T cells, enhancing NKs cytotoxicity, and the DC–NK cell interactions (2). Whereas the myeloid suppressive cells (MDSC) and the tumor-associated macrophages (TAMs) are reported to correlate with poor clinical outcomes and contribute to the carcinogenesis of OSCC (26, 27).

In spite of the equal efficiency of different treatments in the “Poly(I:C)²⁵/DDP^2.5” group and the “DDP⁵” group, we observed a distinct activation pattern of immunocytes. The combination strategy could remarkably alleviate the downregulated percentage of activated CD11c⁺ CD86⁺ DCs in spleen (Fig. 4A) and upregulate the percentage of activated CD69⁺ CD49⁺ NKs in tumor (Fig. 4B). In addition, the mice treated with Poly(I:C)²⁵/DDP^2.5 had significantly impaired CD11b⁺ F4/80⁺ Gr1⁺ TAMs (Fig. 4C) and CD45⁺ CD11b⁺ Gr1⁺ MDSCs (Fig. 4D). A decreased of α-SMA⁺ FSP-1⁺ CAFs were also observed in Poly(I:C)²⁵/DDP^2.5-treated mice and the mRNA expression changes of α-SMA and FSP-1 showed the same results (Fig. 4E and F). These data emphasized on a role of the “Poly(I:C)²⁵/DDP^2.5” strategy in overcoming the immunosuppressive TME and these cells could be the biomarkers of response to the combination chemotherapy.

Currently, cetuximab is the only targeted monoclonal antibody approved for recurrent or metastatic HNSCC, which targets the EGFR. Therefore, we further combined the 200 μg/dose cetuximab and the Poly(I:C)²⁵/DDP^2.5 and investigated the optimal administration sequence (Fig. 5A and F). Based on the strategy mentioned in Fig. 3D, we set three groups, namely the “Poly(I:C)²⁵/DDP^2.5”group as control group, and the mice in group 2 were co-treated with Poly(I:C)²⁵/DDP^2.5/cetuximab²⁰⁰; In the group 3, cetuximab²⁰⁰ was administrated at third day (Fig. 5B). The results showed that the combination of Poly(I:C)²⁵, DDP^2.5 and cetuximab²⁰⁰ contributed to a complete tumor regression and the complete response rate was up to 87.5% (7 of 8) in group 2 when the mice were simultaneously treated with Poly(I:C)²⁵/DDP^2.5/cetuximab²⁰⁰. Interestingly, the strategy of group 2 had minimal effect on body weight when compared to group 3. Although the mice in group 3 also showed a diminished tumor burden, the adverse side effects on body weight were noticed (Fig. 5B). Altogether, the simultaneous administration of Poly(I:C)²⁵/DDP^2.5/cetuximab²⁰⁰ could effectively eliminate the tumors with high security.

To extend the application of orderly combination of Poly(I:C)²⁵/DDP^2.5 to other type of tumor. The B16 metastatic melanoma and A20 non-Hodgkin's lymphoma models were established (Fig. 5A). The results showed that the combined chemotherapy could also limit tumor foci conformation in lung (Fig. 5C) and intestinal system (Fig. 5D) and restrict the tumor progression, respectively. Altogether, these results underscore the importance of this orderly combined chemotherapy in different applications.

LncRNAs have been discovered to contribute to hallmarks of cancer, including proliferation, growth inhibition, immortality, angiogenesis, and metastasis (20). However, its role in chemotherapy resistance is poorly known. TLR-related human lncRNAs were limited, including lnc-IL7R, THRIL, and NEAT1. Interestingly, we here found that the expression of Lnc-IL7R was upregulated in response to the treatment of combined chemotherapy in vivo (Supplementary Fig. S5A). Recently, lnc-IL7R was found to impair the inflammatory response by downregulation of E-selectin, VCAM-1, IL6, and IL8 induced by TLR2 and TLR4, but not by TLR3 in THP-1 cells (28). However, the TLR3 expression showed the decreased trends (Fig. 6A). In addition, the manipulation of TLR3 and Lnc-IL7R expression by Poly(I:C)and DDP were confirmed in vitro in tumor cells. The Lnc-IL7R was positively regulated by Poly(I:C) and DDP, which was accompanied by downregulation of TLR3 in time and concentration-dependent manner (Fig. 6B and C). Moreover, high expression of TLR3 was found to be related to low expression of Lnc-IL7R in OSCC cell lines (Fig. 6D) and 40 matched clinical OSCC tissue samples (Fig. 6E). Knockdown of TLR3 in SCC4 could directly enhance the Lnc-IL7R expression (Fig. 6F). Together, the administration of Poly(I:C)²⁵/DDP^2.5 could promote the tumor to elevate the inflammation-related long noncoding RNA Lnc-IL7R via TLR3.

We then transfected the SCC4 cell with the siRNA targeting Lnc-IL7R, the expression of TNF-α and IL8 was increased by the inhibition of Lnc-IL7R (Fig. 6G; Supplementary Fig. S5B). Besides, the knockdown of Lnc-IL7R showed no effects on the apoptosis but a litter inhibition of cell cycle of tumor cells (Supplementary Fig. S6A and S6B). Therefore, the Lnc-IL7R had no or slight direct impacts on the proliferation and vitality.

We next determined the role of Lnc-IL7R in chemotherapy. The results showed that the cells with Lnc-IL7R knockdown was more sensitive in response to the treatment of Poly(I:C)²⁵/DDP^2.5 with more apoptosis when compared to the cells transfected with siRNA-negative control for 48 hours (Fig. 6H). Furthermore, the SCC4 cells were treated with lentiviral vectors of Lnc-IL7R siRNA or negative control, the function of lentiviral vectors RNAi-Lnc-IL7R was also confirmed (Supplementary Fig. S7A and S7B) and then we investigated its role in the sensitivity of chemotherapy in vivo as described in Fig. 3D. The mice with Lnc-IL7R knockdown in tumor were observed with a higher sensitivity to the treatment of Poly(I:C)²⁵/DDP^2.5 than the control mice without Lnc-IL7R knockdown (Fig. 6I and J). Altogether, upregulation of Lnc-IL7R could be a factor of chemotherapy resistance and its inhibition would be conducive to the chemotherapy efficiency of Poly(I:C)²⁵/DDP^2.5.

Although the TLR3 agonist has been found to be cancer vaccine adjuvant and favor the activation of immune system, its treatment alone had unsatisfactory efficiency on tumor elimination (2, 3). Chemoimmunotherapy could integrate the advantages of immunotherapy and chemotherapy, but the side effects are still responsible for the grim survival prospects. Besides, the patients bearing the tumor could not benefit from increasing dose intensity of cisplatin but decrease their quality of life (QOL; ref. 29). This study emphasized on the strategy that orderly combination of low-dose cisplatin-based chemotherapy and TLR3-related immunotherapy with high efficiency but low adverse side effects.

The TLR3 agonist, including poly (A:U), poly (I:C) or poly-ICLC. Previous studies mainly focused on its role in immune system. Nagato and colleagues (30) reported that the combined chemotherapy of poly-IC and anti–PD-L1 mAb induced a potent immune responses leading to the complete eradication or reduction of tumor growth in melanoma, lung, and colon cancer models, which was involved with CD8⁺ T cells but not CD4⁺ T cells or NK cells in this combinatorial therapy. The poly-IC in that study was administered intravenously at 50 μg/dose 5 days apart for three times and the treatment of poly-IC alone could also, to some extent, restrict the tumor growth. The anticancer properties of poly(A:U) was displayed in TLR3-expressing melanoma models. Conforti and colleagues (31) had combined CpG ODN-based vaccine, oxaliplatin-based chemotherapy and poly(A:U) and administrated in a sequence that vaccine was treated first and 100 μg/dose poly(A:U) was treated intraperitoneally 3 days apart for four times but the oxaliplatin was treated for 1 time during the whole therapy. However, poly(A:U) in this study had no significant effects on the tumor regression and the combined chemotherapy failed to protect mice from tumor progression in the immunodeficient nu/nu mice, indicated that the vaccine/ poly(A:U)/oxaliplatin might function well relied on the intact immune system. Recently, poly-ICLC (40 μg/dose, i.m.) was demonstrated to impair hepatocellular carcinoma progression in the combination of sorafenib though the activation of NKs and CD8⁺ T cells and the inhibition of monocytic MDSCs, and the antitumor role of poly-ICLC/sorafenib was also observed in T cells, B cells, and NK cells-deficient NOD/SCID gamma (NSG) mice (13). We observed no significant antitumor effects of poly (I:C) (25 or 50 μg/dose, i.p.) in our study, whereas the orderly combination of 25 μg/dose poly (I:C) and low-dose cisplatin showed a remarkably effects on tumor eradication in T cells–deficient nude mice. The activation of immunity was assessed and we found the activated DCs and NKs, the inhibition of MDSCs and TAMs were positively related with the response to treatment in our study. Altogether, these data implicate that the discrepancy of administration or dosage contribute to a different efficiency of TLR3 agonist for cancer chemotherapy, which was partly dependent on immunity.

To be noted that we here found a new role of poly (I:C), which was independent of the immune system in cisplatin-based chemotherapy. Drug resistance could be mediated by decreased uptake of water-soluble drugs that enter cells via ATP-binding cassette transporters. The relationship between poly (I:C) and these transporters was poorly known. Micheline and colleagues (32) reported that the viral-induced inflammation stimuli affected the maternal disposition and fetal exposure to clinically important endogenous and exogenous compounds during pregnancy. Because they found that poly(I:C)-treated pregnant rats had significant inhibition of transporters expression in placental including P-gp, MRP-1, and ABCG2 etc. Besides, placental explants from the third trimesters were treated with poly(I:C), which induced a dose-dependent decrease in ABCB1 mRNA levels (33). However, two studies had no direct evidences on the correlation between drug efflux and poly(I:C). In our study, we found poly(I:C) paly the similar role in cisplatin-related tumor chemotherapy. poly(I:C) treatment was able to inhibit the expression of the majority of ATP-binding cassette transporters, including P-gp, ABCB4, MRP-1, ABCC2, CFTR, ABCC8, ABCD2, ABCE1, ABCG1, and ABCG2, leading to increased cisplatin retained in tumor cells and downregulated Pt efflux in vivo. These findings underlined the unexpected role of poly(I:C) in cisplatin-based chemotherapy resistance.

Currently, an emerging role of long noncoding RNA in chemotherapy drug resistance had been discovered, such as maternally expressed gene 3 (MEG3; ref. 34), long noncoding RNA activated by TGF-β (lnc-ATB; ref. 35), in response to cisplatin, trastuzumab etc. The inflammatory long noncoding RNA lnc-IL7R was identified in LPS-stimulated THP-1 and induced by TLR2 and TLR4, but not by TLR3, which was confirmed in human peripheral blood mononuclear cells (PBMCs; ref. 28). However, we uncovered the TLR3 negatively regulated the expression of Lnc-IL7R in carcinogenesis of OSCC and enhanced Lnc-IL7R functioned as a chemotherapy resistance factor in response to our orderly combination of poly(I:C)/cisplatin. Abrogation of Lnc-IL7R could increase the sensibility of the combined chemotherapy of OSCC but not affect the apoptosis of tumor cells.

In summary, our results illustrated a novel strategy of orderly chemoimmunotherapy involved with poly(I:C) and low-dose cisplatin. The orderly utilization of poly (I:C) could downregulate the ATP-binding cassette transporters to retain cisplatin in tumor cells and abrogate the immunosuppressive TME for consolidating the low-dose cisplatin-induced chemotherapy with alleviated adverse side effects. Importantly, the simultaneous utilization of poly(I:C)/cisplatin/cetuximab could be safely applied in the chemotherapy of OSCC.

No potential conflicts of interest were disclosed.

Conception and design: L. Ding, Y. Li, Q. Hu, Y. Ni, Y. Hou

Development of methodology: L. Ding, Y. Li, Y. Hou

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Ding, J. Ren, D. Zhang, Y. Li, Y. Hou

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Ding, J. Ren, Y. Li, J. Ji, Q. Hu, Y. Hou

Writing, review, and/or revision of the manuscript: L. Ding, Y. Li, H. Wang, Y. Ni, Y. Hou

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Ding, Y. Li, X. Huang, Y. Hou

Study supervision: L. Ding, Y. Li, Q. Hu, Y. Hou

We thank the School of Stomatology of Nanjing Medical University for helpful and essentially critical samples.

This work was supported by the National Natural Science Foundation of China (No. 81402238, 81072213, 81271698; to Y. Ni and G Qin), the Nanjing Medical Science & Research Project (No. YKK13145; to Y. Ni).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.