Results of in vitro clonogenic survival assays with increasing radiation doses. A, Human papillomavirus–positive (HPV+) human tumor cells (UMSCC-47 and UPCI-SCC90) are more resistant to radiation than are HPV-negative (HPV−) cells (UMSCC-1, -19, and -84). Cells were irradiated with the indicated dose, and percentage survival was calculated by comparing the number of surviving colonies with the number of cells plated (3 plates per group; the results were averaged across 2 experiments). B, The HPV+ mouse tonsil epithelial cells are more resistant than the HPV− cells. Colony-forming assay of cells was performed as described in Figure 1A. Error bars represent SE.
Radiation-related clearance of human papillomavirus–positive (HPV+) tumors in mice is enhanced by an intact immune response. Tumor growth curves and survival curves are shown for wild-type C57BL/6 mice (A and B) and C57BL/6 B6.129S7-Rag1tm1Mom/J (RAG-1) mice (C and D) implanted with HPV+ or HPV negative (HPV−) mouse tonsil epithelial cells. Six mice were used in each group. Tumors were locally irradiated with the indicated amount of radiation 7 days after implantation (XRT). Results of the log-rank test, with α = .01, show that survival of wild-type mice vs RAG-1 mice implanted with HPV+ tumors is significant at 0 Gy (P = .002), 8 Gy (P = .007), 16 Gy (P = .006), and 24 Gy (P = .003) but did not reach significance at 32 Gy (P = .03) due to 1 RAG-1 mouse with tumor clearance. The difference in survival between wild-type mice implanted with HPV+ vs HPV− tumors was significant at 32 Gy (P = .002).
Growth inhibition of bulky human papillomavirus–positive (HPV+) tumors is enhanced by vaccination with an adenovirus expressing HPV-16 E6 and E7 oncogenes (Ad-E6/E7). The C57BL/6 mice were injected with HPV+ mouse tonsil epithelial cells, and tumors were allowed to grow to 1 cm. Mice were vaccinated with either control (AdEmpty) or Ad-E6/E7 on day 14 after tumor implantation. To determine whether timing of vaccination alters response, a separate group of mice was also vaccinated with Ad-E6/E7 21 days after implantation. Six mice were used in each treatment group. Mice flanks were locally irradiated with 20 Gy of radiation (XRT) on day 28 after tumor implantation. Growth curves (A) and survival curves (B) for the various groups show a slight improvement in survival (P = .003, log-rank test, with α = .01) for the mice receiving vaccination with Ad-E6/E7 14 days before radiation therapy.
Human papillomavirus–positive (HPV+) mouse tonsil epithelial cells (MTECs) are more resistant to cisplatin than are HPV-negative (HPV−) HPV− MTECs. The MTECs were plated, were allowed to attach, and were treated with cisplatin for 24 hours; colony formation was then assessed. The HPV+ and HPV− MTECs were incubated with escalating doses of cisplatin and were allowed to grow until a 15-cell colony size was achieved. Three plates were used per condition, and the results were averaged across 2 experiments. The percentage of surviving cells that formed colonies were quantified. The HPV+ MTECs are more resistant (approximately 63%) to cisplatin than are the HPV− cells (P < .02, Mann-Whitney test). Error bars represent SE.
Human papillomavirus–positive (HPV+) tumors are more sensitive to treatment with cisplatin than are HPV-negative (HPV−) tumors in immune-competent mice. Tumor growth rates and survival rates were quantified after the indicated doses of cisplatin in C57BL/6 mice (A and B) C57BL/6 and B6.129S7-Rag1tm1Mom/J (RAG-1) mice (C). Six mice were used in each group. Mice received 3 weekly intraperitoneal injections of cisplatin, with the first dose starting 1 week after tumor implantation with 1 × 106 HPV+ or HPV− mouse tonsil epithelial cells. Difference in survival for HPV+ tumors comparing wild-type vs RAG-1 mice and the difference in survival for HPV+ vs HPV− tumors in wild-type mice was different at 10- and 20-mg/m2 doses (P < .01, log-rank test, with α = .01). Error bars represent SE.
Cisplatin response can be partially restored by adoptive transfer of splenocytes into B6.129S7-Rag1tm1Mom/J (RAG-1) mice. Immune cells from wild-type C57BL/6 mice were adoptively transferred into C57BL/6 RAG-1 mice, human papillomavirus–positive mouse tonsil epithelial cells were implanted, and response to weekly cisplatin was compared with that of naive RAG-1 mice (6 mice per group). Mice receiving adoptive transfer had slower tumor growth and greater partial response than did control RAG-1 mice. Treatment with cisplatin, 20 mg/m2 every week for 3 weeks, was initiated 1 week after tumor cell implantation (cisplatin treatment). Error bars represent SE.
Cisplatin response is insufficient to induce clearance of bulky tumors. To compare the response for bulky vs small tumor burdens, C57BL/6 mice were implanted with human papillomavirus–positive mouse tonsil epithelial cells and were treated with cisplatin, 20 mg/m2 (3 weekly doses), either 7 or 21 days (approximately 1 cm in greatest dimension) after implantation (6 mice per group). Cisplatin treatment of bulky tumors slowed tumor growth but was not sufficient to result in complete remission when tumors were larger. The difference in survival between treatment at day 7 and day 21 was significant (P = .007, log-rank test, with α = .01).
Clonogenic survival of human papillomavirus–positive (HPV+) and HPV-negative (HPV−) mouse tonsil epithelial cells treated with concurrent radiation and cisplatin in vitro. The cells were split 24 hours after treatment with the indicated doses of radiation and 0.25 μg/mL of cisplatin, and colonies greater than 15 cells were counted at 14 days. Three plates were used per condition, with results averaged from 2 experiments. Surviving fractions of cells were compared with initial numbers of plated cells. No difference was seen in sensitivity to combined radiation and cisplatin use between HPV+ and HPV− mouse tonsil epithelial cells. Error bars represent SE.
Cisplatin and radiation responses in vivo with and without the adenovirus-expressing human papillomavirus (HPV) 16 E6 and E7 oncogene (Ad-E6/E7) vaccination. A, In vivo response of bulky tumors either untreated (RAG-1 and wild type, no treatment) or treated with 3 doses of concurrent radiation (8 Gy weekly) and cisplatin (3 doses of 20 mg/m2 weekly) implanted in either wild-type C57BL/6 or C57BL/6 RAG-1 mice (the other 4 groups). The HPV-positive (HPV+) cells were injected in C57BL/6 or C57BL/6 RAG-1 mice and were allowed to grow for 14 days (6 mice per group). In the treatment groups, mice were divided again to be vaccinated with either control AdEmpty or Ad-E6/E7 14 days after tumor implantation (adenovirus treatment). Cisplatin and radiation administration in the treatment groups began on day 21 after implantation of tumor cells. Growth curves (A) and survival curves (B) were calculated for each mouse. Wild-type mice had a significantly better response compared with C57BL/6 RAG-1 mice (P < .01, log-rank test, with α = .01). Wild-type mice vaccinated with Ad-E6/E7 had improved survival and slower growth compared with mice receiving control AdEmpty. C, Luminescent images of HPV+ tumors that co-express luciferase. The groups of the indicated mice are shown after treatment on day 59. Wild-type (C57BL/6) mice treated with Ad-E6/E7 and cisplatin and radiation cleared tumors compared with persistent tumor in RAG-1 mice treated with cisplatin and radiation or C57BL/6 mice treated with AdEmpty and cisplatin and radiation. CRT indicates chemoradiation therapy; XRT, radiation therapy.
Spanos WC, Nowicki P, Lee DW, Hoover A, Hostager B, Gupta A, Anderson ME, Lee JH. Immune Response During Therapy With Cisplatin or Radiation for Human Papillomavirus–Related Head and Neck Cancer. Arch Otolaryngol Head Neck Surg. 2009;135(11):1137-1146. doi:10.1001/archoto.2009.159
Human papillomavirus (HPV) is the most identifiable cause of head and neck squamous cell cancer (HNSCC). Compared with HPV-negative HNSCC, HPV-positive HNSCC presents at an advanced stage but with significantly better survival. We created a syngeneic mouse model of HPV-positive and HPV-negative HNSCC by transforming mouse primary tonsil epithelial cells with either HPV oncogenes or a nonantigenic RNA interference strategy that affects similar oncogenic pathways.
To examine the effect of radiation therapy on HPV-positive and HPV-negative tumors in immune-competent and immune-incompetent mice and to examine responses in human cancer cell lines.
Prospective in vivo murine model.
Main Outcome Measures
Survival and tumor growth.
For human and murine transformed cell lines, HPV-positive cells were more resistant to radiation and cisplatin therapy compared with HPV-negative cells. In vivo, HPV-positive tumors were more sensitive to radiation, with complete clearance at 20 Gy, compared with their HPV-negative counterparts, which showed persistent growth. Cisplatin in vivo cleared HPV-positive tumors but not HPV-negative tumors. However, neither radiation or cisplatin therapy cured immune-incompetent mice. Adoptive transfer of wild-type immune cells into immune-incompetent mice restored HPV-positive tumor clearance with cisplatin therapy.
The HPV-positive tumors are not more curable based on increased epithelial sensitivity to cisplatin or radiation therapy. Instead, radiation and cisplatin induce an immune response to this antigenic cancer. The implications of these results may lead to novel therapies that enhance tumor eradication for HPV-positive cancers.
Squamous cell cancers (SCCs) of the tonsillar region can be separated on a molecular level into cancers that are and are not associated with human papillomavirus (HPV). Previous findings1- 3 have confirmed that HPV-associated cancers defy the previously held paradigm in which the stage at diagnosis correlates with prognosis. Advanced-stage HPV-associated cancers have a higher cure rate than do HPV-negative cancers.3,4 The presence or absence of viral oncogenes is a distinguishing feature that allows for molecular examination of this difference in therapeutic response.
The HPV status of tonsillar cancers is not used as a routine indicator to guide the choice among the various therapeutic options. Typical treatment involves surgical excision of early cancers and/or concurrent platinum-based chemotherapy (cisplatin or carboplatin) and fractionated radiation therapy. Radiation therapy has developed as a successful treatment option for head and neck SCC (HNSCC). Current therapies provide fractionated doses of radiation, with a significant proportion of individuals reaching complete remission. Radiation therapy for stages III and IV oropharyngeal cancer offers reported 5-year survival of 10% to 20%.5 Cisplatin therapy concurrent with radiation therapy improves clinical response rates and survival rates4,6 for advanced-stage cancers. Cisplatin acts by reacting with guanine and forming intrachromosomal and interchromosomal bonds.7 Formation of these bonds overwhelms DNA repair, thus inducing irreversible damage that leads to cell death. Rapidly dividing tumor cells are, thus, more sensitive to this DNA damage. The combination of radiation and cisplatin synergize to produce irreversible damage.
The improved survival of HPV-positive patients could be owing to multiple factors. For example, it is possible that HPV-positive cells are intrinsically more sensitive to these standard therapies and, thus, respond better to treatment. Recent study findings3,8 suggest that HPV-related cancers actually display enhanced sensitivity to concurrent chemoradiation therapy. Another possibility is that HPV-positive tumors uniquely express foreign (viral) proteins and that an immune response is induced during therapy that helps clear tumors and prevent recurrence.
To better understand why HPV-associated cancers more readily respond to therapy, we developed in vitro and in vivo methods using human cancer cell lines and primary tonsil keratinocytes (mouse and human) transformed with HPV oncogenes.9,10 These methods allow us to better understand responses to common therapies used for this cancer type. In this study, we examine the response of HPV-positive and HPV-negative cancer cells to these therapies in vitro and in a mouse model that replicates many aspects of human disease. The findings demonstrate that an immune response augments the known antitumor action of combined cisplatin and radiation therapy. This immune response is required to clear established cancer. These findings have implications for (1) other antigenic tumors (such as Epstein-Barr virus–related nasopharyngeal carcinoma) that are treated with platinum-based chemotherapy and radiation, (2) the development of further adjuvant therapies that could potentially enhance this immune response, and (3) understanding the poor response to therapy in individuals with suppressed immunity.
Normal mouse tonsil epithelial cells (MTECs) were isolated from the epithelium overlying C57BL/6 mouse and human tonsil tissue as previously described.9,11 The E6E7/Ras and shPTPN13/Ras MTEC lines, generated from normal MTECs as previously described, were maintained in E-media.10,11 The HNSCC tumor lines UMSCC-1, -19, and -84 (HPV negative) and UMSCC-47 and UPCI-SCC90 (HPV positive) were maintained in a combination of Dulbecco modified Eagle medium, 10% fetal calf serum, and penicillin-streptomycin, 1%.
Cell lines were seeded on 60-mm plates (500 cells per plate) in triplicate at each concentration. Cisplatin (0.025, 0.05, 0.1, 0.25, 0.5, and 1.0 μg/mL in dimethyl sulfoxide) was added 5 hours later. Cisplatin was removed along with a medium change 24 hours later. In experiments in which cells were treated with combined cisplatin and radiation, plates were irradiated (0, 2.5, 5, 8, and 16 Gy) 2 hours after the addition of cisplatin. Cisplatin doses for combined cisplatin and radiation therapy were 0 and 0.25 μg/mL. Head and neck cancer cells were treated only with radiation (0, 2, 4, 6, and 8 Gy). Cells were allowed to grow until colonies of untreated controls reached 50 or more cells (10-14 days). Plates were fixed with 70% ethanol and were stained with Coomassie blue; colonies with more than 15 cells were counted. Each experimental condition was repeated in triplicate (3 plates per condition), and the surviving fraction was calculated using the following formula: (colonies counted) / (cells plated × plating efficiency / 100). Plating efficiency was calculated as the number of cells present 24 hours after plating divided by the number of cells plated at time zero. The experiment was repeated, and the average of the experiments is plotted with standard errors. Statistical analysis of the cisplatin in vitro assay was performed using the Mann-Whitney test.
In vivo growth was assayed using previously described techniques.11 All experiments were performed in accord with institutional and national guidelines and regulations; the protocol was approved by the animal care committees at the University of Iowa and Sanford Research. Briefly, using an 18-gauge needle, C57BL/6 mice (immune competent) and B6.129S7-Rag1tm1Mom/J (RAG-1) mice (immune incompetent, lacking B and T cells) were injected with 1 × 106 cells in the subcutaneous tissue of the flank (6 mice per treatment condition; the experiment was repeated, and the results were averaged). The mice were treated with cisplatin dissolved in an isotonic sodium chloride solution (5, 10, or 20 mg/m2) and administered intraperitoneally weekly for 3 weeks. Treatment with cisplatin began 1 week after tumor cell injection, unless otherwise specified. For animals treated with radiation alone, 0, 8, 16, 24, or 32 Gy of radiation was administered to the tumor site, with lead shielding to the rest of the body, at the University of Iowa animal radiation facility (cesium source). Animals were euthanized when the tumor size was greater than 20 mm in its greatest dimension or when the animal was substantially emaciated. Mice were considered tumor free when they showed no evidence of tumor after 3 months. Survival graphs were calculated by standardizing the date of death for a 2-cm tumor. Statistical analysis for the survival graphs was performed using the log-rank test, with α = .01.
The development of adenovirus with no reporter gene (AdEmpty) and adenovirus expressing HPV-16 E6 and E7 oncogenes (Ad-E6/E7) was previously described.12 Adenovirus (107 particles) suspended in 50 μL of phosphate-buffered saline solution was delivered to the nasal passages of anesthetized C57BL/6 mice either 7 or 14 days before undergoing radiation therapy alone (6 mice per treatment group; the results were averaged across 2 experiments). The mice were treated with 20 Gy. For combined cisplatin and radiation therapy plus vaccine experiments, the same dose of virus was delivered 7 days before treatment, with cisplatin at 20 mg/m2 and radiation at 8 Gy given once a week for 3 weeks, to a total dose of 24 Gy.
The mice were euthanized, and their spleens were harvested through an abdominal incision. The spleens were homogenized in a phosphate-buffered saline solution. The resulting phosphate-buffered saline–splenocyte mixture was spun at 0.2 relative centrifugal force for 5 minutes, supernatant was aspirated and discarded, and the pellet was resuspended in 10 mL of ACK-lysing buffer (ammonium chloride, 8.29 g/L; potassium bicarbonate, 1.00 g/L; and disodium EDTA–2H2O, 0.0372 g/L). Four milliliters of a phosphate-buffered saline solution (pH, mean [SD], 7.4 [0.2]) was used to neutralize the ACK-lysing buffer, and the resulting solution was flash spun to isolate larger debris. The supernatant was poured into a new tube and was spun at 0.2 relative centrifugal force for 5 minutes. The resulting supernatant was discarded, and the pellet composed of splenocytes was resuspended in 10 mL of splenocyte medium (5% fetal calf serum, penicillin [1 U/mL]–streptomycin [1 μg/mL], 1%, in RPMI media). Splenocyte concentration was determined using a hemocytometer.
C57BL/6 RAG-1 mice were anesthetized using ketamine, 0.1 mL, intraperitoneally, and then 1.6 × 107 splenocytes were injected retro-orbitally into each mouse. Seven days after splenocyte transfer, the C57BL/6 RAG-1 mice were injected with 1 × 106 E6E7/Ras MTECs and then were treated with cisplatin as described in the “In Vivo Assay” subsection of the “Methods” section (6 mice per treatment group; the results were averaged across 2 experiments). Seven weeks after initiation of the experiment, the C57BL/6 RAG-1 mice were euthanized, their splenocytes were isolated using the aforementioned protocol, and splenocyte populations of CD4, CD8, CD40, and CD25 were examined using flow cytometry.
It is possible that the survival advantage of patients with HPV-positive HNSCC treated with chemoradiotherapy is due to an inherent radiosensitivity of HPV-positive tumor cells compared with HPV-negative tumor cells. To test this hypothesis, we completed clonogenic survival assays with increasing radiation doses in vitro for several HPV-positive and HPV-negative HNSCC lines. The HPV-positive cell lines (UMSCC-47 and UPCI-SCC90) were slightly more resistant to increasing doses of radiation compared with the HPV-negative cell lines (UMSCC-1, UMSCC-19, and UMSCC-84) (Figure 1A). In preparation for anticipated in vivo studies with mice, clonogenic survival assays were performed with HPV-positive and HPV-negative transformed MTECs treated with graded radiation doses. The transformation and characterization of these MTECs have been previously described.10,11 Similar to the human cell lines, the HPV-positive MTECs exhibited greater survival in vitro than did the HPV-negative MTECs (Figure 1B). These findings are contrary to the hypothesis of an inherent sensitivity of HPV cells to radiation therapy.
The in vitro results showed relative resistance to radiation therapy for HPV-positive cells. Next, we used the immune-competent mouse model and examined response to the HPV-positive and HPV-negative cell lines in vivo (Figure 2). In matched mice, HPV-positive and HPV-negative cells demonstrated a dose response to radiation; however, the HPV-positive tumors had sustained complete responses in 80% of mice receiving more than 20 Gy of radiation. On the other hand, the HPV-negative cell lines showed only a partial response to therapy, and only 1 mouse that received the highest dose of radiation (32 Gy) remained disease free. Therefore, the HPV-positive tumors in mice are more easily cleared in vivo.
Although many factors could result in a survival difference in vitro vs in vivo, we hypothesized that an immune response may be enhancing clearance of the HPV-positive cells in vivo. To understand whether an immune response played a role in clearance of the HPV-positive tumors, we repeated the same assay in C57BL/6 RAG-1 mice. The RAG-1 mice are inbred mice that are genetically nearly identical to C57BL/6 mice except that they have a mutated RAG-1 gene, which is essential for the development of functional T and B cells.13 A dose-dependent radiation response was again seen, but complete remission was attainable in only 1 mouse at the highest dose of radiation (Figure 2C). Unlike immune-competent mice (C57BL/6 mice), RAG-1 mice rarely cleared the HPV-positive tumors after undergoing radiation. The survival difference (80%) between wild-type and RAG-1 mice with HPV-positive implanted and treated with equivalent doses of radiation (0, 8, 16, and 24 Gy) was significant (P = .002, .007, .006, and .003, respectively). Thus, the immune system is important in clearing HPV-positive cells treated with radiation, and the presence of the HPV-positive oncogenes E6 and E7 is necessary for the observed survival advantage.
Unlike the mice we tested in the experiments described herein, most patients present with tumors well past the approximate 10-million-cell stage, and their tumors have been present for weeks to months. To test whether a single dose of radiation was sufficient to induce clearance in the treatment of established HPV-positive tumor volume of approximately 1000 mm3, mice were irradiated with 20 Gy. Compared with no treatment (control), radiation therapy limited tumor growth and resulted in prolonged survival time (Figure 3). However, all of the mice had tumors at the end of the experiment, and all of the mice died as a result of persistent disease. Given the role of the immune system noted previously herein, we postulated that enhancing the HPV-specific immune response may help improve response during radiation therapy. We previously developed an adenoviral vaccine that targets the HPV oncogenes E6 and E7 and is effective at inducing clearance of HPV-positive tumor cells expressing E6 and E7 if given before tumor implantation.12 We used the Ad-E6/E7 vaccine or an adenovirus control (AdEmpty) to test whether an immune response to the E6 and E7 HPV oncogenes would improve radiation response to HPV-positive tumors. Compared with vector controls, priming the mice with Ad-E6/E7 7 or 14 days before undergoing radiation therapy resulted in slower tumor growth and prolonged survival but did not result in tumor eradication (0% survival in all groups) (Figure 3). In this HPV-16 murine tumor system, a single treatment with 20 Gy of radiation alone cannot cure established tumors, and enhancing the immune response by vaccination against viral oncoproteins enhances response but is insufficient to induce cure.
The most common chemotherapy agent for treating HPV-related oropharyngeal cancers is cisplatin.14 Examination of HPV-positive and HPV-negative head and neck cancer cell lines for in vitro sensitivity to cisplatin showed no significant difference based on HPV status (data not shown). To determine the cellular toxic effects of cisplatin therapy on HPV-positive and HPV-negative tumorigenic MTECs, we performed clonogenic assays using these 2 cell lines. A dose response to increasing concentrations of cisplatin for each cell line is shown in Figure 4. The HPV-positive cells are more resistant (63%, P < .02) to cisplatin compared with the HPV-negative cell lines. Based on these data, we predicted that the HPV-positive tumors would be more resistant to cisplatin therapy in vivo.
To test cisplatin sensitivity in vivo, immune-competent mice were injected with either HPV-positive or HPV-negative tumorigenic MTECs. We began treatment with cisplatin after 7 days, when tumors were barely palpable. Figure 5A and B show the weekly average tumor volume (6 mice per treatment dose) for mice with HPV-negative and HPV-positive tumors. As in the in vitro experiment, cisplatin again elicited a predictable dose response, but there was a difference at the maximum dose used. Cisplatin treatment at 20 mg/m2 slowed HPV-negative tumor growth and prolonged survival but did not cure a single mouse. On the other hand, the same dose of cisplatin completely cleared all 6 HPV-positive tumors. One mouse developed a recurrence 5 weeks after the last cisplatin treatment, which brought overall survival to 83%. Thus, in vivo, HPV-positive tumors are more sensitive to treatment with cisplatin.
These findings led us to ask whether an immune response was partially responsible for total clearance of the HPV-positive tumors. To test this, we repeated the in vivo experiment using the same doses of cisplatin in RAG-1 mice. The tumor volumes of RAG-1 mice treated with cisplatin at 20 mg/m2 were averaged and plotted weekly (Figure 5C). Cisplatin slowed tumor growth and prolonged survival, but, ultimately, all RAG-1 mice died of their HPV-positive tumors. The difference in survival for wild-type vs RAG-1 mice with HPV-positive tumors was significant with 10- and 20-mg/m2 doses of cisplatin (P < .01). These data prove that cisplatin therapy induces an immune response that enhances clearance of HPV-positive tumors.
The in vivo experiments in immune-incompetent mice showed that T and B lymphocytes were necessary for clearance of HPV-positive tumors during chemotherapy. Because the mice are so genetically similar, it is possible to restore components of the immune response in the RAG-1 mice by transferring wild-type mice immune cells into RAG-1 mice. To reconstitute the immune response in RAG-1 mice, we harvested splenocytes from immune-competent wild-type C57BL/6 mice and transferred them into RAG-1 mice before tumor implantation and cisplatin therapy. Mice with adoptively transferred immune cells had significantly slower tumor growth and enhanced initial regression of tumors after treatment with cisplatin compared with mice that did not undergo adoptive transfer (Figure 6). However, the improved antitumor response after adoptive transfer was transient, and overall survival (0%) was not different from that of RAG-1 mice treated with cisplatin alone (data not shown).
To ascertain that an adoptive transfer occurred in these mice, we removed their spleens after the experiment reached its end point (tumor size >20 mm). The RAG-1 splenocyte population was examined using flow cytometry for levels of CD8, CD4, CD40, and CD25 cells, which were then compared with splenocyte populations from naive RAG-1 and C57BL/6 mice. All adoptively transferred RAG-1 mice contained donor CD8-, CD4-, CD40-, and CD25-expressing cells, but the levels were lower than those for wild-type C57Bl/6 mice (data not shown). These data further support the finding that an immune response aids tumor clearance during cisplatin therapy, but they also suggest that the transfer of cells did not completely reconstitute a fully functional immune response, which is required for total tumor clearance.
The HPV-positive tumors cured with cisplatin therapy in the in vivo experiment represented relatively small-volume early disease. To better understand whether cisplatin therapy and the immune response are sufficient to also clear large-volume established disease, C57BL/6 mice were injected with HPV-positive cells. The tumors were allowed to grow for 1 or 3 weeks (average tumor volume, 1000 mm3) before beginning the cisplatin regimen (20 mg/m2 weekly for 3 weeks). Cisplatin treatment slowed progression of tumor growth. However, if tumors were larger at the initiation of therapy, cisplatin was insufficient to induce a complete response or clearance in wild-type mice (Figure 7).
Combining cisplatin chemotherapy with radiation therapy has been shown in multiple studies to increase complete response and survival.4,6 In vitro studies have shown that this enhanced response is at least in part due to a synergistic action of tumor toxicity for the combined therapies.15 Because our data show that an immune response is as important in clearing HPV-positive tumors in mice treated with radiation therapy or chemotherapy alone, we examined the effect that the immune response played in clearing HPV-positive cancers during concurrent radiation therapy and chemotherapy in vitro. Figure 8 shows that cisplatin enhanced cell toxic effects for HPV-positive and HPV-negative cells. Differences in responses between the cell lines to concurrent therapy are not statistically significant.
We next tested whether an immune response was required to clear well-established HPV-positive disease. In these experiments, we let the tumors grow to 1 cm in greatest dimension before administering treatment. Previous experiments treating larger tumors with either cisplatin (20 mg/m2 every week for 3 weeks) or radiation alone (a single dose of 24 Gy) produced a partial response but did not lead to a complete response or survival for any mice. Therefore, to test response to concurrent therapy, cisplatin was delivered in weekly doses (20 mg/m2 every week for 3 weeks) and radiation (8 Gy every week for 3 weeks) was given on the infusion days. To compare responses in immune-competent and immune-incompetent animals, we delivered the same therapies to wild-type and RAG-1 mice. Comparison of RAG-1 mice receiving no treatment vs chemoradiation in Figure 9 shows that concurrent chemoradiation produces a partial response in the absence of an immune response (RAG-1 AdEmpty vs RAG-1 no treatment). A complete response was seen in 50% of immune-competent mice treated with chemoradiation therapy (wild-type AdEmpty). The addition of Ad-E6/E7 14 days after tumor injection combined with chemoradiation cleared tumors in 90% compared with 50% clearance with AdEmpty. We concluded that the vaccine plus concurrent cisplatin and radiation therapy improves tumor killing and prolongs survival.
Tonsillar cancer, although histologically similar, is caused by random mutations (HPV negative) or by viral oncogenes plus limited random mutations (HPV positive). This dichotomy allows exploration of targeted therapy to viral oncogenes in HPV-positive head and neck cancer with treatment based on a molecular test for HPV. Typing of HPV can already provide prognostic information for patients. Data from this study examining response to current standard therapies allow us to draw several conclusions: (1) HPV-positive tumors are not more sensitive to radiation therapy in vitro; (2) in addition to direct cell toxic effects, an immune response helps clear HPV-positive cells that are directly irradiated; (3) HPV-positive cells are not more sensitive to cisplatin therapy; (4) similar to radiation therapy, cisplatin has a direct toxic effect and the immune response required for clearing HPV-positive cells; and (5) therapies that improve an anti-HPV immune response will likely provide improved treatment outcomes.
Culture of tumor cells in vitro is, by definition, an artificial environment. Previous studies16,17 have shown a correlation between the radiosensitivity of cells in vitro and in vivo tumor growth. However, the human cell lines were placed into a xenograft mouse model with inherent immune reactivity limitations. More recently, the radiosensitivity of human tumor cell lines was correlated with the genotype—that is, cells with more mutations, such as p53 mutations, were more sensitive to radiation therapy.18 Carcinoma cells (RKO) transformed with E6 and E7 exhibited alterations in cell cycle progression but not in radiosensitivity compared with baseline RKO cells.19 This is similar to the findings that HPV-positive HNSCC and MTECs are more resistant to radiation therapy than are HPV-negative cells. The previously mentioned studies do not point to a clear association with oncogene transformation and radiation sensitivity. Instead, cells with fewer mutations of key intracellular tumor suppressors or cell cycle checkpoint controls may be more resistant to the chromosomal DNA damage induced by radiation. Although we noticed a difference in sensitivity to radiation between HPV-positive and HPV-negative cell lines, all the cell lines we tested are relatively sensitive to radiation-induced cell death compared with known radiation-resistant cells. Therefore, the relatively minor differences in radiation sensitivity we detected may not be significant in vivo because, during human therapy, the cells receive 10 times the cumulative dose of radiation.
It has become increasingly clear for many cancers that the immune system has a critical role in response during therapy.20,21 Although cisplatin is a known immunosuppressive agent, it does not suppress the function of the vaccine. This agrees with the current literature that cisplatin does not impair the effectiveness of CRT/E7 DNA vaccine against E7-expressing tumors.22 Recent evidence suggested that certain therapies not only have the direct effect of killing tumor cells but that the manner of cell death induced by the cytotoxic drug is important for inducing an immune response. Radiation and oxaliplatin (a drug similar to cisplatin) have previously been shown to induce an immune response during therapy in experimentally tested models.23 The HPV-positive tumors contain foreign viral proteins. The presence of viral oncoproteins in the cell could trigger an HPV-specific immune response. For example, in an antigenic model of fibrosarcoma and colon carcinoma, radiation has been shown to cause calreticulin to transit to the cell surface.24 The radiation-induced enhancement of calreticulin has been shown to be necessary to develop an antitumor response. Similarly, the mechanism for oxaliplatin has been shown to have direct cellular toxic effects and enhancement of tumor-mediated cell killing.23 Therefore, it is possible that antigen presentation, immune response, and E6 and E7 clearance are enhanced through radiation-mediated tumor damage. This enhanced immune clearance of cells with E6 and E7 antigens could explain the improved survival and decreased recurrence seen in HPV-positive HNSCC treated with cisplatin and radiation. The importance of the immune response in HPV-positive cells is shown in the durable clearance of tumor in immune-competent mice but not in RAG-1 mice. We cannot re-establish HPV-positive tumor cells in mice that clear their tumors during chemoradiation (data not shown). Such a finding would coincide with the finding that people with HPV-positive HNSCC have fewer recurrences after treatment.3 Although the mouse provides a useful model to complete the basic scientific work, future work needs to confirm these findings in humans and begin to determine mechanisms for the HPV-induced immune clearance.
Current treatment of HNSCC of the tonsillar region involves combined treatment with surgery, chemotherapy, and radiation therapy. Treatment is not stratified based on HPV status, and many tumors go without HPV tissue typing. The finding that mouse HPV-positive tumors are more sensitive to platinum-based chemotherapy (cisplatin) and radiation than are HPV-negative tumors correlates with the improved survival noted in previous clinical trials.25 The present mouse data and the review of previous human studies suggest that it would be wise to further investigate whether primary treatment with concurrent cisplatin and radiation should be a mainstay in the treatment of all HPV-positive cancers. The fact that neither therapy alone cleared the larger bulky tumors supports the finding that survival is improved in advanced-stage oropharyngeal cancers with concurrent delivery of cisplatin and radiation. It is possible that surgical removal of the tumor before administering postoperative chemoradiation therapy may further increase survival. Such a study will be examined in future experiments using the present mouse model and also could be applied to a correctly designed human study.
The present data also suggest that individuals who have an intact immune response will do better during therapy. Patients undergoing renal transplantation are at increased risk for viral-related, including HPV, cancers.26 In future work, it is important to also examine whether patients who are immunosupressed from human immunodeficiency virus, organ transplantation, or systemic disease (such as severe cachexia) may have a more limited response to treatment of HPV-positive HNSCC with combined cisplatin and radiation. It would also be important when evaluating any adjuvant chemotherapy. If this therapy disrupts immune function, it actually may make treatment response worse.
Patients do not typically present with microscopic or small tumor burden. Although we achieved clearance of larger tumors with combined cisplatin and radiation therapy, the best survival and complete response of HPV-positive tumors was achieved with the addition of Ad-E6/E7. Recently, results of a vaccination trial27 with viral oncogenes also indicated that vaccination enhances clearance of HPV-transformed cells during chemoradiation therapy. Thus, enhancing the immune response with an anti-HPV oncogene vaccine will likely augment the immunologic clearance induced by combined cisplatin and radiation therapy. Because it may be difficult to obtain approval to inoculate with wild-type oncogenes expressed in adenovirus in humans, an alternative delivery of E6 and E7 antigens could be devised to enhance tumor immune clearance. Such a future approach will likely require a better understanding of how tolerance is initially induced to allow the expression of viral oncoproteins and how to break this tolerance to initiate immune-mediated clearance of the cells containing HPV.
Mouse and human tonsil cells transformed with HPV-16 E6 and E7 oncogenes are less sensitive to radiation and cisplatin in vitro than are HPV-negative tonsil cells. The HPV-positive tumors in mice were more sensitive to combined cisplatin and radiation therapy than were their HPV-negative counterparts and required an intact immune response for tumor clearance. Adoptive transfer of wild-type immune cells into the immune-incompetent mice restored HPV-positive tumor clearance with cisplatin. Large HPV-positive tumors were cleared in approximately half of the mice treated with concurrent cisplatin and radiation, but the large tumors required the addition of Ad-E6/E7 to achieve clearance in 90% of mice.
Correspondence: John H. Lee, MD, Department of Otolaryngology–Head and Neck Surgery, Sanford Health, 1310 W 22nd St, Talley Bldg, Sioux Falls, SD 57105 (LeeJ@sanfordhealth.org).
Submitted for Publication: May 11, 2009; final revision received June 12, 2009; accepted June 16, 2009.
Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Spanos, Nowicki, D. W. Lee, Hostager, and J. H. Lee. Acquisition of data: Spanos, Nowicki, D. W. Lee, Anderson, and J. H. Lee. Analysis and interpretation of data: Spanos, Nowicki, D. W. Lee, Hoover, Gupta, and J. H. Lee. Drafting of the manuscript: Spanos, Nowicki, and J. H. Lee. Critical revision of the manuscript for important intellectual content: Spanos, Nowicki, Hoover, Hostager, Gupta, Anderson, and J. H. Lee. Statistical analysis: Nowicki. Obtained funding: J. H. Lee. Administrative, technical, and material support: Anderson and J. H. Lee. Study supervision: Spanos and J. H. Lee.
Financial Disclosure: None reported.
Funding/Support: This study was funded by the American Cancer Society Pilot Grant and by Sanford Research/University of South Dakota institutional internal funding.
Previous Presentation: This study was presented at the American Head and Neck Society 2009 Annual Meeting; May 30, 2009; Phoenix, Arizona.
Additional Contributions: The UPCI-SCC90 cells were provided by Susanne Gollin, PhD (University of Pittsburgh, Pennsylvania).