Introduction
CpG-oligodeoxynucleotides (CpG-ODNs) are synthetic activators of toll-like receptor 9 (TLR9) and TLR21 in different species. Mammals express TLR9 but lack of TLR21. In contrast, avian species only express TLR21, while fish species contain both TLRs [
1,
2]. The activation of mammalian TLR9 by CpG-ODNs induces immune responses including an innate immune response elicited within hours after CpG-ODN stimulation, followed by the second phase of an adaptive immune response that occurs several days later. During this process, the CpG-ODN activated antigen-presenting cells become competent for their antigen presentation and production of Th1 response-promoting cytokines. Increased expression of co-stimulatory molecules enhances the antigen-presenting activity of the cells to naïve
T cells. The produced cytokines promote a
T helper (Th) 1 polarized immune response and CD8 positive
T cells responses with an effective killing activity [
3,
4]. Due to the fact that the activated immune response facilitates the eradication of cancer cells from body, the antitumor effect of CpG-ODN was investigated and has been demonstrated in various cancer animal models [
5,
6]. In addition, CpG-ODNs are being investigated in clinical trials as a therapeutic agent for cancer treatments, but thus far, no CpG-ODN has yet been approved for cancer therapy [
7‐
9].
The immunostimulatory activity of a CpG-ODN is determined by its nucleotide sequence and structure, which include the content of its CpG-dideoxynucleotides containing hexamer motifs (CpG-hexamer motifs), and the number, position, spacing, and surrounding bases of these CpG-motifs [
10,
11]. Based on their structures, CpG-ODNs can be divided mainly into three different immune stimulatory types. Type A CpG-ODNs induce the production of IFN-α and activate the maturation of plasmacytoid dendritic cells (pDCs), but have little effect on
B-cell activation. Type B CpG-ODNs strongly induce
B-cell proliferation, cytokine production and have some effect on pDC and monocyte maturation, and NK cell activation. The immune stimulatory property of type C CpG-ODNs is between that of the type A and type B CpG-ODNs [
12,
13]. Type B CpG-ODNs are the most commonly used CpG-ODNs. In addition, a CpG-ODN usually has different strengths of activity in different species. This species-specific activity of a CpG-ODN is mainly determined by the nucleotide context of its CpG-hexamer motifs [
14‐
16].
CTLA-4 and PD-1 are the two best investigated immune checkpoint regulators that play important roles in maintaining the homeostasis of the immune system in preventing disorders caused by the over-activation of immune responses. CTLA-4 is essential for immune tolerance and plays a central in the regulation of
T-cell activation. PD-1 controls the late immune response of
T cells in peripheral tissues, as its ligands are mainly expressed in non-lymphoid tissues [
17‐
19]. Various CTLA-4 and PD-1/PD-L1 monoclonal antibodies have been developed for anti-tumors by immune checkpoint blockade. An anti-CTLA-4 antibody was approved by the US FDA in 2011; since then, six additional PD-1 or PD-L1 antibodies have been approved for immunotherapy of different cancer types [
20,
21]. Cancer therapy with these immune checkpoint inhibitors was demonstrated to have a notable efficacy; nevertheless, the response rate of patients with solid tumors is generally less than 30%. Thus, there is an immense need to improve the efficacy of the therapy with immune checkpoint inhibitors [
22,
23].
Head and neck cancers are one of the most common cancers worldwide and head and neck squamous cell carcinoma (HNSCC) accounted for more than 90% of the cancer phenotype. The cancer patients suffer from a poor quality of life. Currently, there is lack of highly effective and satisfied therapeutic strategy for the treatment of this type of cancers. Only less than 20% of the patients respond to therapy with immune checkpoint blockade. This is consistent with the fact that this type of cancer is often characterized by an immunosuppressive microenvironment [
24,
25]. Recently, we developed an orthotopic syngeneic mouse model with an immortal cell line derived from mouse HNSCC for studying immunotherapy of the cancers [
26]. Further, in previous studies, we developed a CpG-ODN called CpG-2722, which activates human and mouse TLR9s, and fish TLR21 [
27]. In this study, we further characterized the immunostimulatory properties of this CpG-2722 and investigated its tumor-suppressive activities alone and in combination with anti-PD-1 using this developed orthotopic syngeneic HNSCC animal model. The CpG-2722 was potent in inducing the expression of IL-12 and IFN-
γ as type B CpG-ODN, but also induced type I IFNs like type A CpG-ODN. CpG-2722 and anti-PD-1 alone suppressed tumor growth. In addition, a combination of CpG-2722 and anti-PD-1 showed a cooperative effect on the regression of HNSCCs.
Materials and methods
Reagents, antibodies, and human peripheral blood mononuclear cells (PBMCs)
All CpG-ODNs were purchased from Integrated DNA Technologies, Inc. CpG-ODNs dissolved in DNase/RNase free water, and aliquots of CpG-ODNs were stored at – 20 ℃. Anti-PD-1 antibody used for in vivo treatment was purchased from InvivoGen. (Cat. No. mpd1-mab15-10). Rat anti-mouse CD8 antibody used for immunohistochemistry was purchased from Invitrogen (clone: 4SM15, Cat. No. 14-0808-82). Trizol reagent and SuperScript™ IV kit were purchased from Invitrogen. SYBR® Green PCR kit was purchased from Qiagen. Human PBMCs were purchased from ZenBio, Inc.
Mouse splenocytes and bone marrow-derived macrophages (BMDMs) preparation
Mouse splenocytes and BMDMs were isolated from 6- to 8-week-old C57BL/6 J mouse (National Laboratory Animal Center, Taiwan). To prepare splenocytes, mouse spleen was collected and pounded by using the plunger of a syringe. Single cells were squeezed out of the spleen fragments, passed through the 40-μm nylon cell strainer (BD FalconTM), and centrifuged at 1500 rpm for 5 min. Cell pellet was resuspended with RBC lysis buffer for 2 min, and the lysis reaction was terminated by adding 30 ml PBS. Splenocytes were spin down and cultured in RPMI 1640 completed medium at 37 °C in a 5% CO2 incubator. To prepare BMDMs, bone marrow cells were washed out of tibias and femurs, passed through a 40-μm nylon cell strainer, and centrifuged at 1500 rpm for 5 min. Cell pellet was resuspended with RBC lysis buffer for 2 min, and lysis reaction was terminated by adding 30 ml PBS. Bone marrow cells were spin down and cultured in 70% DMEM completed medium containing 10% FBS, L-glutamine, antibiotics, 10 mM HEPES buffer, and 30% L929 conditional medium at 37 °C in a 5% CO2 incubator for 7 days.
RNA isolation
Total RNA from mouse splenocytes, BMDMs, and human PBMCs was isolated using the illustra™ RNAspin Mini Kit (GE Healthcare) following the manufacturer's protocol. RNA samples from NHRI-HN1-derived tumors were isolated using the TRIzol reagent.
Reverse transcription-quantitative PCR (RT-qPCR) analysis
Cells were treated with different CpG-ODNs at 0.5 μM for 4 h. RNA samples were then isolated, and reverse transcription was performed using the SuperScript™ IV First-Strand Synthesis System (Invitrogen). We performed quantitative PCR by using QuantiNova™ SYBR® Green PCR Kit (Qiagen) and Applied Biosystems ViiA™ 7 Real-Time PCR System with gene-specific primers (Supplementary Table1 and Supplementary Table 2) for gene expression analysis. The expression level of β-actin was used as the loading control.
Enzyme-linked immunosorbent assay for cytokine production
Human PBMCs were treated with or without different CpG-ODNs as indicated for 24 h, and cell culture media were collected. Cytokines production was measured using enzyme-linked immunosorbent assay (ELISA) kits from eBioscience (San Diego, CA, the USA) following the manufacturer’s protocol.
Syngeneic orthotopic head and neck cancer animal model
Indicated amount of NHRI-HN1 cells were mixed with matrigel (BD Biosciences) at 1:1 ratio to a total volume of 100 μl. The cells were intramucosally injected into the 6–8-week-old C57BL/6 J mice through a side of buccal region to grow the tumor [
28]. When tumors reached the indicated size, the mice were intratumorally injected with the indicated amount of CpG-2722 twice/week, in combination with or without 10 μg of anti-PD-1 antibody once/week. All groups contained five mice and five tumors. Tumor volume of the mice bearing NHRI-HN1-derived tumor was measured using the formula = length × (width)
2 × 0.5.
Immunohistochemistry
Paraffin-embedded NHRI-HN1-derived tumors were sectioned into 5-μm tissue slides. These tissue slides were rehydrated with graded concentrations of ethanol to PBS and blocked endogenous peroxidase with 3% hydrogen peroxide for 5 min. For CD8 staining, a rat monoclonal antibody against mouse CD8 (clone: 4SM15, Invitrogen) was used at a dilution of 1:50 and incubated at room temperature for 1 h. The tissue sections were incubated with HRP-conjugated secondary antibody at room temperature for 30 min following washing with PBST. The detection was processed in the Discovery XT automated IHC/ISH slide staining system (Ventana Medical System, Inc. Tucson), using the ultraView Universal DAB Detection Kit (Ventana Medical System, Inc. Tucson), according to the manufacturer’s instructions. Immunostaining was visualized after counterstaining with hematoxylin. CD8-positive cells and leukocyte infiltration were counted using ImageJ software.
Discussion
Head and neck cancers comprise a group of malignancies arising from the oral cavity, oropharynx, hypopharynx, larynx and lips, paranasal sinuses, nasopharynx, and nasal cavity. Squamous cell carcinoma (SCC) constitutes the majority (> 90%) of histopathological types of head and neck cancers [
34‐
36]. Immune checkpoint blockade with anti-PD-1 antibodies has been approved by US FDA for the treatment of recurrent and metastatic HNSCC. Nevertheless, the majority of patients do not respond to the therapy, underscoring the need for a strategy to alleviate the resistance of HNSCC to immunotherapy [
24,
25]. Previously, only few syngeneic animal models such as the MOC1/2 and TC-1 were available for the study of head and neck cancers. The MOC1/2 cell lines were derived from gene-deficient mice, and the TC-1 cell line was derived from primary lung cells by immortalization and retroviral transduction with HPV16 E6/E740 [
37,
38]. Recently, a 4NQO-induced murine oral squamous cell (4MOSC) line was developed, and we developed a stemness-enriched murine HNSCC cell line, NHRI-HN1 for generating syngeneic orthotopic head and neck cancer animal models with C57BL/6 J mice [
26,
39]. The NHRI-HN1 cells were demonstrated to have similar gene expression and signaling pathway modulation as human oral squamous cell carcinoma tissues; therefore, the established cancer animal model is reliable for the study of human HNSCC [
26]. In this study, we used this newly developed model to study the antitumor effect of combining CpG-2722 and anti-PD-1 on suppressing HNSCC growth.
The CpG-2722 used in this study is a B type CpG-ODN containing 19 nucleotide bases, two copies of GTCGTT-hexamer motifs, and four thymidines in between these two hexamer motifs (Table
1) [
27]. CpG-ODN with GTCGTT-hexamer motif often contains species-specific activity to human cells [
14‐
16]. However, in addition to the grouper and human cells, the CpG-2722 also contains immunostimulatory activity to mouse cells. Thus, the CpG-2722 is universal for different species, and research result of this CpG-ODN obtained from cancer animal model is more likely to be replicated in humans. In this study, we further characterized this CpG-ODN and compared its immunostimulatory activity with the CpG-ODNs of different types. CpG-2722 has a good activity on the induction of inflammatory cytokines, particularly the IL-12 genes and IFN-
γ gene in immune cells, as a type B CpG-ODN. In addition, it also induces the expression of type I IFNs like a type A CpG-ODN. This cytokine-inducing profile is not only seen in immune cells, but it also observed in the tumors of the HNSCC animal models.
The structural basis for the cytokines-inducing profile of CpG-2722 is not clear. Nevertheless, it is known that CpG-ODNs with different structures have different cytokine-inducing profiles. The distinct abilities of type A and type B CpG-ODNs in the induction of type I IFNs are resulted from their higher-order structures. Type A CpG-ODNs are capable of forming multimeric aggregates, whereas type B CpG-ODNs are monomeric and do not have such a feature [
40]. A model of spatiotemporal TLR9 activation has been suggested to explain the differential immunostimulatory activities of different CpG-ODNs in dendritic cells. Type A CpG-ODNs activate TLR9 in early endosomes to trigger IRF7 activation for the induction of large amounts of type I IFNs. Type B CpG-ODN is quickly transported to late endosomal/lysosomal compartments for TLR9 activation in order to activate NF-κB and the production of inflammatory cytokines. In contrast, class C CpG-ODNs have the capability to be retained in these endosomal compartments to activate the production of IFNs and inflammatory cytokines [
41,
42]. In line with these, encapsulation of class B CpG-ODNs into particles allows their retention in early endosomes for the induction of higher levels of type I IFNs [
43]. The CpG-2722 contains activities to induce the production of inflammatory cytokines and IFNs; whether it contains some structural characteristic of the type A CpG-ODNs are still to be investigated.
CpG-2722 had a good activity in inducing IL-12 and IFN-
γ in different cell types. A heterodimeric form of IL-12 referred to as p70 is composed by the IL-12A (p35) and IL-12B (p40). The promoter of IL-12A contains the binding sites for transcription factors such as NF-κB, c-Rel, and IRF-1. The promoter of IL-12B contains NF-κB, PU.1, IRF-1, IRF-8, NFAT, and AP-1 binding sites. Thus, the production of IL-12 is regulated by multiple signal pathways, leading to the activation of different transcription factors for these binding sites [
44‐
46]. The characteristics of CpG-2722 to regulate multiple pathways including NF-κB and IRF for inductions of inflammatory cytokines and type I IFNs could play some role for its superior activity in regulating the expression of the IL-12 genes. Most of the IL-12-induced antitumor effects include generation of T helper type 1 (Th1) and cytotoxic
T-cell responses, which are mediated by NK- and
T-cell-generated inflammatory responses [
29,
30]. IL-12 activates the production of IFN-
γ from these cell types. IFN-
γ in turn regulates the differentiation and activation of NK cells and
T cells. In addition, similar to TLR agonists such as CpG-ODN, the IFN-
γ is effective in inducing the polarization of macrophages into M1 phenotypes. The M1 macrophages are inflammatory, while the M2 macrophages are immunosuppressive cells. These two types of macrophages are convertible. Thus, the increased M1 macrophages in the tumor microenvironment are favorable for generating antitumor responses [
31].
The functional mechanisms for the antitumor activities of anti-PD-1 and CpG-ODN have been well investigated. Interaction between PD-1 and its ligands initiates a signal transduction to inhibit TCR activation. In this pathway, PD-1 activates SHP2 tyrosine phosphatase, leading to the dephosphorylation of signaling molecules downstream to TCR and inhibition of TCR activation-mediated cell survival, cell proliferation, and cytokines’ production. Anti-PD-1 blocks this inhibitory signaling, leading to an increased killing activity of CD8
T cells in the tumors [
17,
18]. In contrast, the antitumor activity of CpG-ODN is mainly mediated by inducing the production of inflammatory cytokines from immune cells, which prime the antitumor responses resulting from the activation of immune cells. Of these induced cytokines, IL-12 and IFNs have been shown to contribute to the CpG-ODN-induced antitumor effect [
5‐
7,
47,
48]. CpG-ODN monotherapy often showed good activities in inducing tumor regression in a cancer animal model. In addition, the injection of CpG-ODN into the tumor exerted better antitumor activity than the administration of the CpG-ODN at distant sites such as via intraperitoneal injection or intravenous injection [
49,
50]. Based on the positive results of preclinical studies, CpG-ODNs have been investigated in clinical trials as therapeutic antitumor agents [
7‐
9]. However, no CpG-ODN has been approved for cancer treatment so far, suggesting that CpG-ODN alone may not be sufficient for boosting an efficient antitumor immune response in humans. In contrast, although cancer therapy with anti-PD-1 has been approved, only a small portion of patients benefited from this therapy.
The immune system employs coordinated innate immunity and adaptive immunity to elicit an antitumor immune response. The resistance of patients to immunotherapy may result from deficiencies in various aspects of the antitumor response, resulting from an immune-suppressive tumor microenvironment. In this study, we showed the cooperative effect of CpG-2722 and anti-PD-1 on the suppression of HNSCC growth. HNSCC is characterized by an immune-suppressive tumor microenvironment [
24,
25]. The function of CpG-2722 to activate various cytokines including IL-12, IFN-
γ, and type I IFNs leading to the accumulation and activation of immune cells including pDCs, M1 macrophages, and CD8 positive
T cells in the tumor microenvironment, thus sharpening up the microenvironment into a “hot” one favorable for
T-cell-mediated tumor killing. Thus, CpG-2722 could be a candidate of combinational therapy with immune checkpoint inhibitors for tumors with an immune suppressive microenvironment.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.