Integration of Oncogenes via Sleeping Beauty as a Mouse Model of HPV16⁺ Oral Tumors and Immunologic Control

Human papillomaviruses (HPV) are small DNA viruses responsible for a multitude of diseases ranging from benign warts to anogenital and oropharyngeal malignancies (1–3). In most cases, HPV infections spontaneously regress or remain subclinical. However, persistent infections with high-risk HPV types (i.e., HPV 16 and 18) may progress into precancerous lesions, and eventually cancer (4, 5). In the case of HPV⁺ head and neck squamous cell carcinoma (HNSCC), which is mainly driven by HPV16, patients do not usually display symptoms until an advanced stage, during which tumor metastasis, particularly to tumor-draining lymph nodes, is common (3, 6). Typically, surgery, radiation therapy, and chemotherapy are used for the treatment of HNSCC, but these therapeutic methods often result in irreversible damage to anatomical structures and reduced quality of life (7–9). It has been suggested that antitumor immunity and immune microenvironments are important for tumor prognosis. As such, many attempts have been made to generate HPV tumor models useful for the evaluation of candidate immunotherapies against HPV-associated HNSCC that may be better tolerated and more effective than conventional therapies (10).

Traditionally, preclinical HPV⁺ oral tumor models are generated by implanting tumor cell lines that express the HPV oncoproteins into the oral cavity of mice (11–13). One major drawback of these tumor challenge models is their inability to capture major aspects of tumorigenesis caused by HPV infection, including the processes of cellular transformation after viral entry and the disease progression from precancerous lesion to invasive cancer (14). An alternative method to create a preclinical HPV tumor model involves the generation of immunocompetent transgenic mice that expresses the HPV oncogenes E6 and E7 (15–20). Nonetheless, due to continuous HPV16 E6 and E7 expression, the transgenic mice often become immunotolerant to these oncogenes, rendering the model less effective at evaluating immunotherapy (10). Zhong and colleagues have published an improved inducible transgenic HPV⁺ oral tumor model that may be able to circumvent this issue of immune tolerance (21). However, because HPV16-E6/E7 becomes universally expressed in the epithelium after induction, the antitumor immune response is expected to be different from the clinically localized tumor. Therefore, more research is required to improve preclinical HPV oral tumor models. Ideally, a preclinical HPV oral tumor model should possess the following characteristics: (i) forms spontaneous, localized, HPV⁺ oral carcinomas; (ii) displays an immune microenvironment that closely resembles those of HPV⁺ oral tumors in patients; (iii) the tumor-bearing mice should respond appropriately to immune therapies and generate antitumor immunity; and (iv) tumor formation should follow clinical progression starting from precancerous to invasive and metastatic states.

HPV infects basal epithelial cells through wounding. After entering the host cell and uncoating its capsid, viral DNA is transferred into the cell nucleus and maintained in low copy numbers, ranging from 50 to 100 copies per basal cell (22). Under normal circumstances, these basal cells divide, migrate off the basement membrane, change shape, and terminally differentiate into cornified cells at the epithelial surface. Once the basal cells detach from the basement membrane, they silence the cell cycle and stop dividing. Occasionally, HPV DNA integrates into the genomic DNA of the infected cell, during which several HPV DNA encoded genes, including L1, L2, E1, E2, E4, and E5, may be lost. Importantly, E2 proteins serve as inhibitors for E6 and E7 oncogenes. Oncoproteins E6 and E7 can degrade p53, disrupt the interaction between retinoblastoma protein (pRb) and E2 transcription factors (E2F), and reinitiate the cell cycle, resulting in rapid cell proliferation (23, 24). Thus, integration of the HPV viral genome into the host genome and the continual expression of E6/E7 are important for HPV oncogenesis.

In order to simulate HPV oncogenesis, we hypothesized that local introduction of high-risk HPV16-E6/E7 DNA can mimic viral DNA entering the cells after HPV infection. The introduction of plasmids encoding HPV16 oncogenes into epithelial cells can be achieved with the use of electroporation, a physical transfection method that uses an electrical pulse to create temporary pores in cell membranes through which substances like nucleic acids can pass into cells. It has been shown that electroporation is effective in introducing foreign nucleic acids into many cell types in mice (25). However, although electroporation achieves very stable transfections, the gene expression is typically transient in mice. Therefore, an additional method is required to reproduce the persistent expression of E6/E7 oncogenes resulting from HPV DNA integration.

Sleeping beauty (SB) is a synthetic transposable element composed of a transposon DNA substrate and a transposase enzyme. SB transposase regulates excision and insertion of transposon DNA into a TA-dinucleotide of the host genomic DNA in a cut-and-paste manner (26, 27). The SB transposable element has been previously used to introduce antiangiogenic genes for long-term therapeutic effects and to generate spontaneous tumor models with persistent oncogene expression (28–30). We hypothesized that the utilization of the SB transposase system could reproduce the essentially random integration process of HPV oncogenes into the host genome and result in continual expression of E6/E7 in the transfected cells.

Previously, it has been demonstrated that HPV16-E6/E7 and mutant Ras oncogenes cooperate to form tumors in transgenic mice (31). This synergistic, carcinogenic effect is restricted to anatomic sites with epithelial transition zones classically affected by HPV16-E6/E7 oncogene expression, which, in the buccal area, is located between the outer skin of the lip and the inner buccal mucosa. Therefore, introduction of viral oncogenes HPV16-E6/E7 with mutant Ras oncogene locally in the buccal area may also result in efficient tumor formation in mice.

In this study, we developed a spontaneous, HPV⁺ buccal tumor model by injecting mice with plasmids encoding oncogenes HPV16-E6/E7, mutant Ras, Luciferase (as a reporter gene), and SB transposase into the buccal area followed by electroporation. Using a candidate therapeutic HPV DNA vaccine, pNGVL4a-CRT/E7(detox), we evaluated responses to immunotherapy and examined the ability of the transformed tumor cells to migrate to distant locations.

Six- to 8-week old female C57BL/6NCr (strain #556), FVB/NCr (strain #559), BALB/cAnNCr (strain #555), and athymic nude (Athymic NCr-nu/nu, strain #553) mice were purchased from Charles River Laboratories. All mice were maintained at the Johns Hopkins University School of Medicine Animal Facility (Baltimore, MD) under specific pathogen-free conditions. All procedures were performed according to protocols approved by the Johns Hopkins Institutional Animal Care and Use Committee and in accordance with recommendations for the proper use and care of laboratory animals.

To generate pcDNA3-Luc, luciferase was cloned into the XbaI/XhoI sites of pcDNA3 by using the following primers: 5′-aaatctagaatggaagacgccaaaaacat-3′ (forward) and 5′-aaactcgagcacggcgatctttccgccct-3′ (reverse).

To generate pcDNA3-Luc-T2a-E7, T2a-E7 was cloned into the Xho I/EcoRI sites of pcDNA3-Luc by using the following primers: 5′- aaactcgaggagggcagaggaagtcttctaacatgcggtgacgtggaggagaatcccggccctatgcatggagatacacct-3′ (forward) and 5′-tttgaattctggtttctgagaacagatgg-3′ (reverse).

To generate the luciferase E7 and E6 oncogene fusion construct, pcDNA3-Luc-T2a-E7-T2a-E6), T2a-E6 was cloned into the EcoRI/BamHI sites of pcDNA3-Luc-T2a-E7 by using the following primers: 5′- aaagaattcgagggcagaggaagtcttctaacatgcggtgacgtggaggagaatcccggccctatgcaccaaaagagaact-3′ (forward) and 5′-tttggatcccagctgggtttctctacgtg-3′ (reverse).

To generate Pkt2-Luc-T2a-E7-T2a-E6, Luc-T2a-E7-T2a-E6 was removed from pcDNA3-Luc-T2a-E7-T2a-E6 using XbaI/BamHI sites, and the sequence was cloned into the XbaI/Bgl II sites of the Pkt2/clp-akt vector, as previously performed (30).

To generate Pkt2-Luc-T2a-E7, T2a-E7 was cloned into the XhoI and BstXI sites of Pkt2-Luc-T2a-E7-T2a-E6 by using the following primers: 5′-aaactcgaggagggcagaggaagtcttct-3′ (forward) and 5′-tttccagctagctggttatggtttctgagaacaga-3′ (reverse).

To generate Pkt2-Luc-T2a-E6, T2a-E6 was cloned into the XhoI and BstXI sites of Pkt2-Luc-T2a-E7-T2a-E6 using the primers 5′-aaactcgaggagggcagaggaagtcttct-3′ (forward) and 5′-tttccagctagctggttacagctgggtttctctacg-3′ (reverse).

In the DNA constructs, T2a (EGRGSLLTCGDVEENPGP), a self-cleavage peptide from the T. asigna virus, was inserted between each gene, allowing for the production of each protein individually from a single fusion protein (32). Cleavage of the T2a sequence occurs due to a ribosomal skip mechanism at the point between the final glycine and proline residues (32).

The construction of the pT/Caggs-NRasV12 plasmid, pT2/C-Luc//PGK-SB13 plasmid, and the pKT2/CLP-AKT plasmid has been described previously (30). The plasmids were purchased from Addgene (plasmid #20205, plasmid #20207, and plasmid #20281, respectively). The construction of the pCMV(CAT)T7-SB100 plasmid has also been described previously (33) and was purchased from Addgene (plasmid #34879).

The DNA vaccine, pNGVL4a-CRT/E7(detox), has been previously described (34) and expresses a fusion of human calreticulin (CRT) and detoxified HPV16-E7 antigen. CRT is a 46-kDa protein located in the lumen of the cell's endoplasmic reticulum and has been shown to strengthen MHC class I presentation for the activation of CD8⁺ T cells. E7(detox) represents a HPV16-E7 gene with mutations at positions 24 and 26, which disrupt the Rb-binding site of E7 and, thus, abolishes the capacity of E7 to transform cells (35). The plasmid backbone, pNGVL4a, was obtained from the NIH National Gene Vector Laboratory and has been used in several clinical studies (36).

For the generation of tumors in immunodeficient mice, athymic nude mice (5 mice/group) were submucosally injected with (i) plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100; (ii) plasmids encoding luciferase and HPV16-E6, mutant Ras, and SB100; (iii) plasmids encoding luciferase and HPV16-E7, mutant Ras, and SB100; (iv) plasmids encoding luciferase, HPV16- E6/E7, and SB100; or (v) plasmids encoding mutant Ras, SB13, and luciferase into the buccal area (10 μg/plasmid, 30 μL/injection), followed by electroporation (8 pulses, 72v for 20 ms/pulse, 20 ms intervals between each pulse).

For the generation of tumors in immunodeficient mice with Akt expression, athymic nude mice (5 mice/group) received submucosal injection of DNA plasmids encoding luciferase, HPV16-E6/E7, constitutively active AKT, and SB100 into the buccal area, followed by electroporation as described (10 μg/plasmid, 30 μL/injection).

For the generation of tumors in immunocompetent mice, C57BL/6NCr, Balb/c, or FVB/NCr mice (5 mice/group) were submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation (10 μg/plasmid, 30 μL/injection).

For the generation of tumors in immune-suppressed mice, C57BL/6NCr mice (5 mice/group) received one of the following treatments:

(i) intraperitoneal (i.p.) injection of 100 μg monoclonal anti-CD3 (clone: 17A2; catalog #BE0002, Bio X Cell) per day for 3 consecutive days, and on the day after the final anti-CD3 injection, submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation; (ii) i.p. injection of 100 μg monoclonal anti-CD4 (clone: GK1.5; catalog #BE0003-1, Bio X Cell) per day for 3 consecutive days, and on the day after the third anti-CD4 injection, submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation. Anti-CD4 injections were continued at weekly intervals; (iii) i.p. injection of 200 μg monoclonal anti-CD8 (clone: 2.43; catalog #BP0061, Bio X Cell) per day for 3 consecutive days, and on the day after the third anti-CD8 injection, submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation. Anti-CD8 injections were continued at 4-day intervals; (iv) submucosal injection of plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation with no antibody treatment; or (v) i.p. injection of both 100 μg monoclonal anti-CD4 and 200 μg monoclonal anti-CD8 per day for 3 consecutive days, and on the day after the final anti-CD4 and anti-CD8 injection, submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant Ras, and SB100 into the buccal area, followed by electroporation. Ten micrograms/plasmid and 30 μL/injection were used for all treatment arms.

Following oncogene plasmid transfection, buccal tumor growth was monitored 1 to 2 times per week via luminescence imaging (37). In addition, tumor size was measured with digital calipers. To reduce mouse suffering due to the negative impacts of the presence of buccal tumors, mouse death due to tumor burden was defined as (i) display of a bioluminescence signal of 10⁹ p/sec/cm²/sr; (ii) a tumor volume of 150 mm³; or (iii) a 10% reduction of weight compare with the average weight of healthy mice of the same age. Mice were subsequently euthanized via CO₂ asphyxiation when any of these three characteristics were observed.

In vivo bioluminescence imaging was performed using the IVIS Spectrum in vivo imaging system series 2000 (PerkinElmer). In summary, mice were anesthetized by intramuscular (i.m.) injection of a ketamine/xylazine solution (5:1 ratio). After i.p. injection of substrate d-luciferin (GoldBio), bioluminescence imaging for luciferase expression was conducted on a cryogenically cooled IVIS system using Living Image acquisition and analysis software (Xenogen). Images were acquired 10 minutes after d-luciferin administration. Bioluminescence was detected by the IVIS imager, integrated, and digitized. The region of interest from displayed images was quantified as total photon counts using Living Image 2.50 software (Xenogen).

C57BL/6NCr mice (5 mice/group) received i.p. injections of 100 μg monoclonal anti-CD3 per day for 3 consecutive days, and on the day after the last anti-CD3 injection, submucosally injected with plasmids encoding luciferase and HPV16-E6/E7, mutant NRas, and SB100 into the buccal area, followed by electroporation as described. Ten days after plasmid injection, mice were vaccinated with pNGVL4a-CRT/E7(detox) or empty pNGVL4a vector only (20 μg/dose/mouse) via i.m. injection in the hind leg, followed by electroporation. The mice continued to receive the same vaccination at 4-day intervals 3 additional times. Following vaccination, tumor growth was monitored twice a week by in vivo bioluminescence imaging, and the survival and weight of mice were monitored for up to 50 days after oncogenic plasmids transfection. Mice were euthanized via CO₂ asphyxiation on day 50 or when any of the following conditions were observed: (i) display of a bioluminescence signal of 10⁹ p/sec/cm²/sr; (ii) a tumor volume of 150 mm³; or (iii) a 10% reduction of weight compare with the average weight of healthy mice of the same age.

After euthanizing the tumor-bearing mice, buccal tumors were surgically isolated and placed into 10% neutral buffered formalin solution for adequate fixation with a minimum of 48 hours at room temperature. The tumor samples were then formalin fixed and paraffin embedded. Hematoxylin and eosin (H&E) staining of tissue sections was performed by the Johns Hopkins University Oncology Tissue Services. Immunohistochemical (IHC) staining was performed for the following markers: proliferation marker Ki-67 (clone: D3B5; Cell Signaling Technology; catalog #12202), oncoprotein NRas (Santa Cruz Biotechnology, catalog #sc-519), and HPV marker p16 (Abcam; catalog #ab51243), secondary antibody (poly-HRP anti-rabbit IgG; Leica, catalog #PV6119) was used for all IHC staining. The histology slides were reviewed by a board-certified gynecologic pathologist (Dr. T.-C. Wu) of the Pathology Department in the Johns Hopkins University School of Medicine.

Seven days after the last vaccination, mice were euthanized, and the tumor tissues were surgically removed, minced into 1- to 2-mm pieces, digested with serum-free RPMI-1640 medium containing collagenase I (0.05 mg/mL; Sigma-Aldrich, catalog #C9891), collagenase IV (0.05 mg/mL; Sigma-Aldrich, catalog #C5138), hyaluronidase IV (0.025 mg/mL; Sigma-Aldrich, catalog #H3506), DNase I (0.25 mg/mL; Sigma-Aldrich, catalog #DN25-g), and penicillin (100 U/mL) and streptomycin (100 μg/mL; Thermo Fisher, catalog #15140122), and incubated at 37°C with periodic agitation. The tumor digest was then filtered through a 70-μm nylon filter mesh to remove undigested tissue fragments. The resultant single tumor cell suspensions and tumor-infiltrating lymphocytes were washed with PBS containing 0.5% BSA, and viable cells were determined using trypan blue dye exclusion.

In situ hybridization was performed using the RNAscope 2.0 HD Brown Chromogenic Reagent Kit (Advanced Cell Diagnostics) using the supplied protocol and a target probe against HPV16-E6/E7 (Advanced Cell Diagnostics #311521). Briefly, freshly cut, formalin-fixed, paraffin-embedded slides were baked at 60°C overnight. After deparaffinization, slides were air dried, circled with a hydrophobic barrier pen (Vector Labs, ImmEdge pen catalog # H-4000), and then exposed to pretreatment solutions 1, 2, and 3. Target probes were hybridized for 2 hours at 40°C in hybridization chambers, followed by a series of signal amplification and washing steps. The signals were detected by chromogenic reactions using DAB chromogen followed by undiluted Gill's hematoxylin (Sigma-Aldrich) counterstaining.

To evaluate the systemic immune response, peripheral blood was collected into 1.5 mL EDTA-coated Eppendorf tubes from the ventral arteries of mice tails. After red blood cell lysis with ACK lysing buffer (Quality Biological), peripheral blood mononuclear cells (PBMC) were then stained with fluorescein isothiocyanate (FITC)-conjugated monoclonal rat anti-mouse CD3 (BD Pharmingen, catalog # 555274), allophycocyanin (APC)-conjugated monoclonal rat anti-mouse CD8a (BioLegend, catalog #100712), and phycoerythrin (PE)-conjugated H-2D^b tetramer loaded with HPV16-E7aa49–57 peptide (provided by NIH Tetramer Core Facility) at 4°C for at least 30 minutes. After washing with FACS washing buffer (0.5% BSA in PBS), the cells were stained with 7-AAD before flow cytometry analysis to exclude dead cells.

To evaluate the presence of E7-specific CD8⁺ T cells in the tumor microenvironment (TME) following vaccination, tumor-infiltrating lymphocytes (TIL) obtained from the tumor tissue of mice were stained with FITC-conjugated anti-mouse CD8a (BD Pharmingen, catalog # 553031) and PE-conjugated HPV16-E7aa49-57 peptide-loaded H-2D^b tetramer. To evaluate the presence of regulatory T cells in the TME, TILs were stained with FITC-conjugated rat anti-mouse CD4 (eBioscience, clone RM4-5) and APC-conjugated anti-mouse CD25 (eBioscience, catalog #17-0251-82), permeabilized and fixed (Fixation/Permeabilization Solution Kit, BD Biosciences, catalog #555028), and stained with PE-conjugated anti-mouse Foxp3 (eBioscience, clone FJK-16s). To evaluate the presence of myeloid derived suppressor cells in TME, TILs were stained with FITC-conjugated anti-mouse Gr-1 (eBioscience, catalog #11-5931-82) and PE-conjugated anti-mouse CD11b (eBioscience, catalog #12-0112-82). After staining and washing, the cells were acquired with a FACS Calibur flow cytometer (BD Biosciences) and analyzed with FlowJo V.10 software.

When the tumor-bearing C57BL/6 mice experienced one of the following conditions: (i) display of a bioluminescence signal of 10⁹ p/sec/cm²/sr; (ii) a tumor volume of 150 mm³; or (iii) a 10% reduction of weight compare with the average weight of healthy mice of the same age, mice were statistically considered to have died from tumor burden, and euthanized according to the protocol. Enlarged tumor-draining lymph nodes (TDLN) and non–tumor-draining lymph nodes (NTDLN) were surgically remove for ex vivo bioluminescence assays. The TDLNs were minced, treated with tissue digestion buffer [collagenase (0.05 mg/mL), collagenase IV (0.05 mg/mL), hyaluronidase IV (0.025 mg/mL), and DNase I (0.25 mg/mL) in deionized water], filtered, and then cultured in 6-well plates with RPMI 1640 medium (Thermo Fisher, catalog #11875093) supplemented with penicillin (100 U/mL) and streptomycin (100 μg/mL), 1× MEM nonessential amino acids (Thermo Fisher, catalog #11140050), 1 mmol/L sodium pyruvate (Thermo Fisher, catalog #11360070), 2 mmol/L l-glutamine (Thermo Fisher, catalog #25030081), 55 nmol/L 2-mercaptoethanol (Thermo Fisher, catalog #21985023), and 10% fetal bovine serum (Thermo Fisher, catalog #16000036). After 1 week of culturing, whole cell mixtures were transferred to T75 tissue culture flasks. Once the cells expanded to cover more than 70% of the flask surface, the cells were trypsinized, PBS-washed, and 5 × 10⁵ cells were resuspended in 40 μL PBS solution, followed by submucosal buccal injection into immunodeficient athymic mice. After injection, the presence and growth of tumor cells in mice were monitored via in vivo bioluminescence imaging.

Athymic nude mice received submucosal injection of plasmid DNA in the buccal area, followed by electroporation. The plasmids included luciferase, E6/E7, NRas^G12V, and SB100. When the tumor size exceeded 7 mm in diameter, mice were considered to have died from tumor burden, and euthanized according to protocol. Tumor tissue, TDLNs, and spleens were harvested for luciferase assays. Tissues from naïve mice were used as negative controls. The tissue was homogenized by a Mini-Beadbeater-1 with 1 mm and 0.1 mm glass beads (Bio Spec Products Inc.) in 50,000. Bioluminescence assays of the tissue samples were conducted using the superficGloMax-Multi Detection System, according to the manufacturer's instructions (Promega Corporation). Relative light units (RLU) were adjusted to each gram of tissue.

The statistical analyses were performed with GraphPad Prism V.6 software. Data were expressed as means ± standard deviation (SD). Kaplan–Meier survival plots were constructed to estimate either tumor-free rate or survival percentage. The log-rank test was used to compare survival times between treatment groups. Comparisons between individual data points were analyzed by two-tailed Student t tests. A P < 0.05 was considered statistically significant.

To generate spontaneous oral tumors that continuously express HPV16-E6 and E7, different combinations of plasmids (Fig. 1A) were submucosally injected into the buccal area of athymic nude mice, followed by electroporation. The luciferase enzymes encoded in the plasmids resulted in the generation of bioluminescence from viable, transfected cells that could be measured quantitatively using an IVIS imager. After plasmid injection, tumor growth was monitored through in vivo bioluminescence imaging. We observed that a combination of plasmids encoding sleeping beauty transposase (SB100), luciferase, HPV16-E6/E7, and NRas^G12V oncogenes resulted in the fastest tumor formation and growth rate as measured by the luminescence intensity (Fig. 1B and C). Based on our results, the combination of E6, E7, and mutant NRas oncogenes is the most efficient at causing oral tumor formation in mice.

After demonstrating the ability to generate spontaneous oral tumors in immunodeficient mice by oncogene-encoded plasmid transfection with SB transposase, we sought to apply the same strategy for the generation of spontaneous oral HPV tumors in immunocompetent mice. Plasmids encoding luciferase, HPV16-E6/E7, NRas^G12V, and SB100 were submucosally injected into the buccal area of C57BL/6NCr, Balb/c, and FVB/NCr mice. However, in contrast to our observation in immunodeficient mice, the luciferase gene expression decreased steadily after transfection, as indicated by bioluminescence intensity, indicating the clearance of transfected cells by immunocompetent mice, regardless of genetic strain (Fig. 2A). Thus, we hypothesized that an immunosuppressed environment is necessary for the spontaneous formation of tumors following plasmid transfection. To test this hypothesis, we treated C57BL/6NCr mice with or without anti-CD3, anti-CD4, or anti-CD8 antibodies to deplete various T-cell populations prior to plasmid transfections (illustrated in Fig. 2B). Mice that received anti-CD3 (total T-cell depletion) experienced a rapid increase in luminescence signal after plasmid transfection, suggesting the expansion of plasmid-transfected cells. In contrast, mice that received anti-CD4 or anti-CD8 antibodies (i.e., CD4⁺ or CD8⁺ T-cell depletion, respectively) generated similar luminescence signals as the control group (no antibody depletion; Fig. 2C). When cotreated with anti-CD4 and anti-CD8, plasmid-transfected mice showed an initial increase in luminescence signal, which decreased back to baseline by day 28 (Supplementary Fig. S1).

In addition to the observed increase in luminescence intensity, visible tumors were observed in CD3-depleted mice around 17 days after oncogene transfection. To examine the relationship between T-cell depletion and the formation of tumors, total T-cell populations in CD3-depleted, plasmid-transfected mice was measured. As shown in Fig. 3A, the amount of systemic CD3⁺ T cells in mice declined sharply following anti-CD3 injection. On day 9, the luminescence intensity in mice began to increase and continued to increase even as the CD3⁺ T-cell population began to recover. The volume of the spontaneous tumors formed after plasmid transfection was strongly correlated with luminescence intensity (Fig. 3B). The presence of growing tumor in plasmid-transfected mice did not influence the rate of T-cell recovery when compared with that of CD3-depleted mice without oncogenic plasmid transfection (Supplementary Fig. S2).

To assess the influence of growing oral tumors on the health status of mice, the formation and growth of tumors, as well as the weight of oncogenic plasmids-transfected mice with or without CD3 depletion, were monitored and evaluated. Similar to those observed in Fig. 2C, the luminescence signal in control mice (without CD3 depletion) decreased rapidly following oncogenic plasmid transfection, resulting in no tumor formation, as well as a steady weight gain (Supplementary Fig. S3A). In comparison, a steady increase in luminescence signal was observed with oncogenic plasmid-transfected, CD3-depleted mice, which corresponded to formation of tumors and reduction in mouse weight (Supplementary Fig S3B). Four weeks after oncogenic plasmid transfection, significant differences in mouse weight (∼10%) were observed between oncogenic plasmid-transfected mice with or without initial CD3 depletion (Supplementary Fig. S3C).

Together, these results suggest that initial total T-cell depletion, as compared with partial depletion, is required for the formation of spontaneous oral HPV tumors in immunocompetent mice and that luminescence intensity is an effective and reliable measurement for monitoring tumor growth. These data demonstrate that the size of spontaneous HPV buccal tumors negatively impacts the health status of mice. Thus, to minimize mouse suffering due to the presence of growing buccal tumor, we subsequently defined mouse death due to tumor burden as discussed in the Materials and Methods section, and mice were euthanized when any of the listed characteristics were observed.

To characterize the spontaneous buccal tumors formed after the transfection of oncogene-encoding plasmids, we treated C57BL/6NCr mice with anti-CD3 daily for 3 days, followed by transfection of plasmids encoding E6/E7, NRas^G12V, SB100, and luciferase at the buccal area. When the resultant spontaneous tumors reached 7 mm in diameter, the mice were sacrificed, and the buccal tumors were excised, fixed with formalin, and embedded in paraffin. The tissue sections of the tumors were prepared and stained for various tumor markers. The spontaneous tumors originated from the buccal mucosa area appeared as solid papillary masses with bleeding ulcers (Fig. 4A). H&E staining showed that the tumors were composed of spindle-shaped tumor cells with high mitotic activity (Fig. 4B). IHC staining revealed high expressions of Ki-67, a cell proliferation biomarker, as well as positive NRas and HPV biomarker p16 expression (Fig. 4B and C). RNA in situ hybridization with HPV16-E6/E7 probes also showed strong staining in the tumor (Fig. 4C).

After we successfully demonstrated the ability to repeatedly generate the HPV16⁺ spontaneous buccal tumors in C57BL/6NCr mice, we sought to verify whether this model could be used to assess the efficacy of immune therapy. Initial CD3-depleted, oncogene plasmid-transfected C57BL/6NCr mice were vaccinated with either a therapeutic HPV DNA vaccine, pNGVL4a-CRT/E7(detox), or empty pNGVL4a vector only 10 days after the plasmid transfection. The mice received the same vaccination a total of 4 times at 4-day intervals (Fig. 5A). As shown in Fig. 5B, mice vaccinated with pNGVL4a-CRT/E7(detox) had significantly better survival percentages compared with mice vaccinated with empty pNGVL4a control. Eighty percent of the mice that received pNGVL4a-CRT/E7(detox) DNA vaccination survived more than 8 weeks after plasmid transfection, whereas all mice that received empty pNGVL4a vector vaccination died in less than 4 weeks after plasmid transfection. When the experiment was repeated over an extended time frame, we observed that all tumor-bearing mice treated with pNGVL4a-CRT/E7(detox) survived more than 20 weeks after oncogenic plasmid transfection, whereas all mice treated with empty pNGVL4a vector died in less than 6 weeks (Supplementary Fig. S4). The control of spontaneous HPV buccal tumors in vaccinated mice corresponded with the generation of a potent systemic E7-specific CD8⁺ T-cell response (Fig. 5C and D). Mice vaccinated with pNGVL4a-CRT/E7(detox) showed lower luminescence signals compared with the mice treated with empty pNGVL4a vector, demonstrating the clearance of plasmid-transformed tumor cells (Fig. 5E and F).

In addition to evaluating the systemic E7-specific CD8⁺ T-cell response elicited in tumor-bearing mice following vaccination, the influence of vaccination on the immune cell populations in the TME was assessed using the tumor tissues harvested from mice 1 week after the last vaccination. In comparison with tumor-bearing mice treated with empty pNGVL4a plasmid vector, a significantly higher population of E7-specific CD8⁺ T cells was observed in the tumors of mice treated with pNGVL4a-CRT/E7(detox; Fig. 6A). In contrast, no differences in the abundance of regulatory T cells or myeloid-derived suppressor cells in the tumor were observed among mice treated with various regimens (Fig. 6B and C). These results indicate that therapeutic vaccination can lead to the generation of a potent antigen-specific immune response against the spontaneous HPV buccal tumors, resulting in effective tumor control and prolonged the survival of tumor-bearing mice.

In our initial experiments involving the use of immunodeficient mice, a substantial amount of luciferase activity was observed in the tumor, tumor-draining lymph nodes (TDLNs), and spleens of the plasmid-transfected, athymic nude mice as compared with naïve mice (Fig. 7A). Likewise, in C57BL/6NCR mice bearing spontaneous HPV buccal tumors generated from plasmid transfection, enlarged and superficial TDLNs were observed with detectable luciferase activity as compared with non- tumor-draining lymph nodes (NTDLNs; Fig. 7B and C). We hypothesized that these TDLNs contained transformed tumor cells that metastasized from the buccal area. To test our hypothesis, when the mice were euthanized due to tumor burden, luciferase-positive TDLNs were surgically removed and subsequently minced, digested, and cultured. After cell proliferation was observed, the cell mixture was collected and implanted into the buccal area of immunodeficient nude mice. The reimplanted LN cells showed positive luciferase activity, and the formation of buccal tumors in implanted nude mice was observed (Fig. 7D). IHC staining revealed that the tumors generated by LN cell implantation were positive for HPV marker p16, as well as NRas (Fig. 7E). These histologic analyses of the tumors formed from the LN cell implantation revealed characteristics similar to that of HPV16 spontaneous buccal tumors originating from plasmid transfection (Fig. 4), suggesting that the HPV16 tumor cells can migrate to different locations inside tumor-bearing mice, replicating the metastatic process of late-stage HPV-associated cancer.

Although we demonstrated that transfection of plasmids encoding HPV16-E6/E7, NRas^G12V, luciferase, and SB100 leads to the spontaneous formation of tumors in mice, Ras mutation is not typically observed in the HPV⁺ oral cancers presented by patients in the clinic (38). In comparison, many clinical HPV⁺ oral tumors are associated with mutations in the PI3K signaling pathway (38). Thus, we explored the ability to generate spontaneous HPV⁺ buccal tumors in immunocompromised mice by changing the plasmid encoding NRas^G12V to plasmid encoding constitutively active AKT, a downstream molecule of the PI3K pathway. We observed that the introduction of AKT with HPV16 E6 and E7 oncogenes could also lead to the spontaneous formation of oral tumors in athymic nude mice (Supplementary Fig. S5).

In this study, we developed a preclinical, spontaneous HPV buccal tumor model through transfection of plasmids encoding oncogenes HPV16-E6/E7, NRas^G12V, as well as SB transposase and luciferase reporter genes. Plasmid transfection resulted in the formation of spontaneous tumors in immunodeficient, athymic nude mice and in immunocompetent C57BL/6NCr mice subjected to early immune suppression. We demonstrated that these spontaneous tumors expressed relevant HPV16 biomarker p16 and detectable HPV16 RNA, could metastasize and migrate to distant locations, such as TDLNs, and could be controlled through immunotherapeutic treatments.

The spontaneous HPV buccal tumor model developed in the current study complements existing models. As mentioned in the Introduction, although prior HPV⁺ HNSCC tumor models can result in consistent tumor formation and have the ability to utilize human-derived HPV⁺ tumor cells, they fail to replicate several crucial characteristics of HPV cancers (11–13). Likewise, conventional HPV transgenic mice require multiple rounds of selection to ensure successful genetic modification, which can take months to a year or more for tumors to form and often result in immune tolerance (10, 15–19, 39–41). The improved, Cre-LoxP–inducible, transgenic HPV oral model addressed some issues of earlier models, making it more clinically relevant compared with conventional transgenic models (21). However, the modified transgenic model requires the integration of three transgenes, which further complicates the selection process. Mammalian genomes have been shown to contain many “pseudo” LoxP sites that can potentially be recombined by Cre-recombinase with unknown effects (42). The spontaneous HPV tumor model developed in the current study does not require xenograft implantations of preexisting tumor cells but instead utilizes the SB transposon system to induce oncogene integration and long-term expression (26, 28–30). The method used in our current study resulted in the formation of spontaneous tumors in both athymic nude mice and wild-type mice within a few weeks following plasmid transfection, with a tumor formation rate of greater than 80%, allowing for time-efficient and time-flexible experimental designs compared with the transgenic models mentioned previously. The high tumor formation rate upon plasmid transfection can potentially reduce the mouse turnover while ensuring the generation of a sufficient amount of tumor-bearing mice to be used for subsequent evaluations. Therefore, our model provides a cost-, time-, and labor-efficient method to generate spontaneous HPV oral tumors that mimic localized viral gene integrations and cellular transformations that occur during natural HPV disease progressions.

In our spontaneous tumor model, we demonstrated that initial CD3 T-cell depletion in immunocompetent C57BL/6NCr mice is necessary for tumor formation to occur. This likely reflects the natural immunogenicity of HPV oncoproteins E6 and E7, as the generation of antigen-specific immune responses following exposure to HPV antigens, either through infection or through introduction of DNA encoding the HPV antigenic proteins, has been well documented in both human and murine settings (43–48). It is likely that the initial anti-CD3 injection depleted the CD3⁺ T cells in mice at the time of plasmid transfection and prevented the generation of rapid antitumor immune responses, including the priming of antigen-specific T cells. As a result, no generation of systemic or local E7-specific CD8⁺ T-cell responses was observed in initial CD3-depleted, plasmid-transfected control mice, which did not receive subsequent antigen-specific immunization. In clinical settings, it is also observed that patients experiencing immune suppression, such as those infected with human immunodeficiency virus (HIV), have elevated risks for persistent HPV infections and cervical cancers, further emphasizing the importance of the immune system in mediating HPV-associated disease progression (49–55). Nonetheless, we demonstrated that despite the decrease in the total T-cell population caused by early CD3 depletion, vaccination with a therapeutic HPV DNA vaccine, pNGVL4a-CRT/E7(detox), still effectively enhanced HPV16/E7-specific CD8⁺ T-cell responses in tumor-bearing mice, leading to potent tumor control and prolonged mouse survival. These findings suggest that a potent therapeutic HPV vaccine can potentially trigger effective HPV-specific immune responses, even in immune-suppressed patients. This is in-line with our previous preclinical study demonstrating that HPV DNA vaccination is capable of controlling HPV-associated tumors in the absence of CD4⁺ T cells (56), and results in the clearance of HPV infection. Although the current study utilized immune therapy as the modeled therapeutic method, the spontaneous HPV⁺ buccal tumor model may also be used to evaluate other forms of treatment, including chemotherapy and radiotherapy.

In addition to CD4, CD8, or CD3 depletion, we also tested the coadministration of anti-CD4 and anti-CD8 for the induction of tumor formation. However, although the CD4/CD8-depleted mice experienced an initial increase of luciferase signal, the signal eventually decreased to baseline levels, indicating clearance of transfected cells in CD4/CD8 depleted mice as compared with that in CD3-depleted mice. The discrepancy between the experimental outcomes in oncogenic plasmid-transfected mice treated with anti-CD3 versus anti-CD4 anti-CD8 suggests the potential involvement of CD3⁺ CD4⁻ CD8⁻ immune cells, such as natural killer T (NKT) cells, γδ T cells, and double-negative αβ T (DNT) cells, in preventing the formation of tumors following oncogenic plasmid transfection, in addition to the antigen-specific CD4⁺ and CD8⁺ T cells. NKT cells have been reported to recognize α-galactosylceramide in conjunction with the MHC class I–like CD1d molecule and produce a large amount of IFNγ to activate NK cells and CD8⁺ T cells to eliminate MHC-negative and MHC-positive tumors, respectively (57). NKT cells have been suggested to contribute to the immune surveillance of chronic lymphocytic leukemia (58), prevent metastasis of preclinical breast cancer model (59), and it has been demonstrated to be well tolerated and potentially efficacious as an antitumor therapy in patients with non–small cell lung cancers, head and neck tumors, and advanced melanoma (for review see refs. 60 and 61). γδ T cells have been suggested as an early source of IFNγ in tumor immune surveillance, and they regulate the function of tumor-specific CD8⁺ T cells (62), as well as to exert direct cytotoxicity against various tumor cells (for review see refs. 63, 64). γδ T cell-based immunotherapy has been suggested to have better efficacy than current second-line therapies for a few types of cancers (for review see ref. 65). Lastly, DNT cells have been shown to have the capacity to restrict the growth of a lymphoma tumor cell line in mice (66), to have potent cytotoxic effects against allogeneic and autologous primary leukemic blasts isolated from AML patients (67), and to inhibit the growth of pancreatic carcinoma in vivo via MHC class I chain-related molecules A and natural killer group 2 member D (NKG2D) mediated manner (68). It would be of particular interest to examine the potential roles of these immune cell populations in regulating the formation and growth of spontaneous HPV⁺ oral tumors following oncogenic plasmid transfection in future studies. Further studies are also warranted to explore additional methods to induce an immunosuppressive environment in mice for the spontaneous formation of tumors.

At around 4 weeks after oncogene injection, when the mice were sacrificed, enlarged TDLNs were observed. These LN cells had positive luciferase activity and, when implanted into the buccal area of immunodeficient, athymic nude mice, resulted in the formation of gross tumors. Based on the difference in the luminescence intensity observed in the buccal tumor tissue and TDLNs, our data suggest that the transformed tumor cells detected in the LNs originated in the buccal region and subsequently metastasized to the TDLNs. Previous studies also support the notion that buccal tumor formation precedes metastasis in the oral TDLNs, thereby resulting in tumor-positive TDLNs (69–72). We hope to potentially utilize our model in the future to study and analyze in vivo tumor migration and metastasis, and future studies are warranted to evaluate how the metastasis of these tumor cells influences the immune microenvironment of TDLNs and to correlate the results with those in clinical settings.

Despite all of the potential benefits that our preclinical tumor model may introduce, we recognize that our model still has room for further improvements. In particular, while our spontaneous tumor model expresses several relevant markers that are present in clinical HPV oral tumors, the morphology of the tumors derived through our approach is closer to that of a sarcoma rather than carcinoma. This may be due to the fact that submucosal injection delivers the plasmid below the epithelial layer, causing transfection and transformation of nonepithelial cells. This issue could potentially be addressed through the use of the K14 promoter. Although the conventional promoters currently used allow for more efficient expression of the oncogenes, K14 promoters can restrict the oncogene expression within transfected epithelial cells, increasing the likelihood for the generation of spontaneous HPV16⁺ buccal carcinomas. Our current spontaneous HPV tumor model involved the transfection and integration of HPV16-E6/E7 and NRas^G12V oncogenes into the mice. Although it has been well documented that this combination is effective in triggering carcinogenesis in vivo (31), clinically reported HPV⁺ oral cancers are not typically associated with Ras mutation (38). In comparison, mutations in the PI3K–AKT signaling pathway are observed in over 50% of HPV⁺ head and neck squamous cell carcinomas (38). We have demonstrated that transfection of HPV16-E6/E7, AKT, luciferase, and SB100 could also lead to the spontaneous formation of oral tumors in athymic nude mice, although the tumor growth rate of AKT tumors was slower than that formed from transfection of plasmid encoding NRas^G12V. Further investigation is warranted for the generation and optimization of the HPV AKT tumor model, which may further enhance the clinical relevance of the spontaneous HPV⁺ oral tumor model.

Although the incidence of HPV-associated oropharyngeal cancer has been on the rise in developed countries (73), a significant portion of oropharyngeal cancer cases worldwide are still attributed to causes independent of HPV infections such as tobacco smoking (74). Unlike HPV⁺ oral cancers that are associated with HPV-E6/E7 and mutations in the PI3K signaling pathway, HPV⁻ oral cancers are mainly characterized by mutations in TP53 (38). Previous studies utilized the administration of carcinogen 4-nitroquinoline 1-oxide (4-NQO) in mice to generate HPV⁻ oral cancer (75–78). However, the induction of carcinogenesis by administration of 4-NQO alone is very slow, requiring up to 36 to 40 weeks before the tumor becomes visible. We believe our method to generate spontaneous HPV⁺ oral tumors may also be applied for the generation of spontaneous HPV⁻ oral tumors. By integrating mutant TP53 genes into cells located at the buccal area of mice via the SB transposase system and exposing the mice to carcinogen 4-NQO, we believe we can generate clinically relevant spontaneous HPV⁻ tumors at a much faster rate, allowing for rapid evaluation of therapeutic options against HPV⁻ oral tumors.

In conclusion, our model provides a fast, efficient, and simple option to induce spontaneous formation of HPV⁺ tumors, which can be used to assess therapeutic strategies. This model can potentially be used to identify effective therapeutic intervention, analyze tumor migration, and conduct tumor biology and TME research.

No potential conflicts of interest were disclosed.

Conception and design: Y.-H. Lin, A. Yang, C.-F. Hung, T.-C. Wu

Development of methodology: Y.-H. Lin, M.-C. Yang, C.-F. Hung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.-H. Lin, M.-C. Yang, R. Jiang, A. Yang, S. Peng, T. Henkle

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y.-H. Lin, M.-C. Yang, A. Yang, S. Peng, C.-F. Hung, T.-C. Wu

Writing, review, and/or revision of the manuscript: Y.-H. Lin, R. Jiang, A. Yang, E. Farmer, S. Peng, C.-F. Hung, T.-C. Wu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.-H. Tseng, A. Yang, E. Farmer, Y.-N. Chang, C.-F. Hung

Study supervision: C.-F. Hung

This work was supported by the NIH Cervical Cancer Specialized Program of Research Excellence (SPORE; P50CA098252), R21 grant (R21AI109259), R21 grant (R21CA194896), and R01 (R01CA114425) and Tri-Service General Hospital (TSGH-C107-086).

The authors thank Dr. Richard B.S. Roden for helpful discussions and critical review of this article.

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.