Systemic and local human papillomavirus 16-specific T-cell immunity in patients with head and neck cancer

Authors


Abstract

Squamous cell carcinomas of the head and neck (HNSCC), in particular those of the oropharynx, can be caused by human papilloma virus Type 16 (HPV16). Whereas these HPV-induced oropharyngeal carcinomas may express the HPV16 E6 and E7 oncoproteins and are associated with better survival, the nonvirally induced HNSCC are associated with overexpression of p53. In this study we assessed the presence of systemic and local T cells reactive against these oncoproteins in HNSCC. An exploratory study on the presence, type and function of HPV16- and/or p53-specific T cells in the blood, tumor and/or metastatic lymph node as measured by several immune assays was performed in an unselected group of 50 patients with HNSCC. Tumor tissue was tested for HPV DNA and the overexpression of p53 protein. Almost all HPV16+ tumors were located in the oropharynx. Circulating HPV16- and p53-specific T cells were found in 17/47 and 7/45 tested patients. T cells were isolated from tumor cultures and/or lymph nodes of 20 patients. HPV16-specific T cells were detected in six of eight HPV+ tumors, but in none of the 12 HPV-tumors. Tumor-infiltrating p53-specific T cells were not detected. In depth analysis of the HPV16-specific T-cell response revealed that this response comprised a broad repertoire of CD4+ T-helper Type 1 and 2 cells, CD4+ regulatory T cells and CD8+ T cells reactive to HPV16. The local presence of HPV16-specific T-cell immunity in HPV16-induced HNSCC implicates a role in the antitumor response and support the development of immunotherapy for HNSCC.

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer1 and comprises cancers from the oral cavity, hypopharynx and oropharynx. Human papilloma virus Type 16 (HPV16) induces HNSCC squamous cell carcinoma almost exclusively at the oropharynx, the incidence of which is rising worldwide.2 Patients with HPV-induced HNCSCC often present with advanced disease but their prognosis is better than HNCSCC associated with smoking and drinking.3–5 Notably, the majority but not all patients with an HPV+ tumor display a strong gene signature for adaptive immune response in their tumor6 as well as strong tumor infiltration by T cells4 suggesting that the presence of HPV may be related to an enhanced local tumor-specific T-cell response.

HPV16-induced cancer cells express two virally derived oncoproteins E6 and E7 and because these oncoproteins are foreign to the body, they are expected to evoke an immune response. So far, there are only few reports on HPV16-specific T-cell immunity in HNCSCC. They describe elevated levels of circulating HPV16 E7-specific CD8+ T cells7 as well as the presence of HPV16-specific IFNγ-producing T cells in in vitro cultures of PBMC from patients with HPV16+ HNSCC.8 In addition, circulating antibodies to HPV16 have been detected in HNCSCC patients with high viral load and it is suggested that the HPV-specific antibody status is related to clinical outcome.9, 10

In nonviral-induced, smoking and drinking-associated HNCSCC, p53 overexpression is one of the most common abnormalities identified. Mutation of p53 is correlated with poor prognosis. Mutations in p53 protein or in the p53 regulating pathway can cause overexpression of the protein. The expression of the HPV16 E6 oncoprotein, however, induces increased degradation of p53. Both mechanisms may result in an enhanced presentation of p53-derived peptides to T cells, the latter mechanism predominantly to CD8 T cells. P53-specific IgG antibodies have been found in patients with HNSCC and this was suggested to be linked with bad prognosis.11, 12 These p53-specific IgG responses indicate that p53- specific CD4+ T cells are also present in these patients. Indeed, the presence of both circulating and tumor-infiltrating p53-specific T cells has been reported.13, 14

In HNSCC a number of cytokine-, antibody- and vaccine-based immunotherapeutic approaches have been tested or are currently underway.15–17 Recently, clinical success was achieved in the field of the immunotherapy of HPV16-induced (pre-)malignancies of the anogenital region, in particular with therapeutic vaccines.18–20 Similar vaccine approaches are tested for the treatment of p53-overexpressing cancers.21–23 There is a strong notion that the presence and type of preexisting circulating and local tumor-specific immunity influences the clinical outcome as well as that of immunotherapeutic approaches.6, 18 Thus, to optimize immune resistance to HNSCC, a better understanding of the character and specificity of tumor-infiltrating-lymphocytes (TIL) in HNSCC is needed. Therefore, an explorative study was performed in which the blood of 50 patients and 20 successfully obtained T-cell cultures from their tumors and/or lymph nodes were studied for the presence, type and function of HPV- or p53-specific T cells. In addition the tumors were studied for the presence of HPV and p53-overexpression.

Material and Methods

Patients and material

Patients presenting with an epithelial lesion of the head and neck region were included after informed consent. After inclusion 50 ml of blood and two (research and diagnostic) biopsies were taken in parallel with physical examination under sedation however for several patients a diagnostic biopsy had been taken 2–3 weeks earlier before the patients were referred to our hospital. None of the patients received treatment prior to study inclusion. Diagnosis was based on histopathological analysis of biopsies. PBMC were isolated by Ficoll-density centrifugation and obtained cells were directly tested in a lymphocyte stimulation test (LST). Remaining PBMC was preserved in liquid nitrogen until further use. In a number of cases, PBMC were transformed with EBV to obtain B-LCL lines.

Immunohistochemistry

Immunohistochemical analysis was performed on 3-μm paraffin sections, mounted on aminopropylethoxysilane-coated slides. Sections were deparaffinized, rehydrated and treated with 0.3% H2O2 in methanol for 20 min to block endogenous peroxidase activity. Antigen retrieval was performed (0.01 M citrate, pH 6.0) and sections were blocked with 1% BSA. Subsequently, sections were stained for p53 (overnight, room temperature) using a 1:2,000 dilution of anti-human p53 (Clone DO-7, Neomarkers, Fremont, CA) in PBS containing 1% BSA. Next the slides were incubated with Powervision-Poly/HRP (Immunogenic, Duiven, The Netherlands) and immune complexes were visualized with diaminobenzidine. Slides were evaluated by light microscopy and scored for the presence, intensity and percentage of tumor cells expressing p53.

HPV typing

DNA was isolated from formalin-fixed, paraffin-embedded biopsy samples as previously described.24 The aqueous solution obtained from the DNA analysis was diluted 1:10 and 1:50. The presence of HPV in the diluted samples was determined by genotyping the samples that demonstrated 65 basepair PCR amplimers on a 3% agarose gel using an INNO-LiPA Genotyping Extra test (Innogenetics, Ghent, Belgium), according to the manufacturer's instruction. This assay allows the detection of the following HPV types: HPV 6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 43, 44, 45, 51, 52, 53, 54, 56, 58, 59, 66, 68, 69/71, 73 and 74. Hybridization patterns were visually inspected and interpreted using a grid.

Antigens and lymphocyte stimulation test

Pools of overlapping 22-mer peptides spanning the entire HPV16 E6 and E7 proteins were used and pools of overlapping 30-mer peptides spanning the whole p53 protein as described previously.25, 26 The first and last amino acid within the sequence of the indicated antigen that is represented by the pooled peptides is indicated (e.g., E6 1-92). Memory response mix (MRM) was used as positive control.26, 27 The capacity of T cells to proliferate on recognition of the antigen was determined by LST as described earlier.27 The average of eight tested wells divided by the average of the medium control plus 3 SD was calculated and is referred to as the stimulation index (SI).

Day 6 supernatant was analyzed for the presence of IFNγ and IL-10 by ELISA (Sanquin, The Netherlands). Cut-off value (IFNγ 50 pg ml−1; IL-10 20 pg ml−1) and >2× the concentration of the medium control.

Isolation and culture of TIL

All T-cell cultures and tests were performed in IMDM supplemented with 10% human AB serum (Greiner Bio-one, Germany). Biopsies were incubated in IMDM+10% AB serum supplemented with Fungizone 25 μg ml−1, 1,000 U pen/strep and 20 μg ml−1 gentamycin (all Invitrogen) for 1 hr to decrease the bacterial load in the tissue. Subsequently, the biopsies were minced in to pieces of ∼ 1 mm3 and cultured in medium with 10% T-cell growth factor (Zeptometrix, Buffalo, NY) and 5 ng ml−1 IL-15 (Peprotech, Rocky Hill, NJ). During the first day 5 ng ml−1 IL-7 (Peprotech) was added. TIL were expanded for 2–3 weeks with only addition of medium and cytokines and analyzed for phenotype and function.

In two patients with large lymph node metastases, a needle aspirate was taken and frozen in liquid nitrogen until further use.

Analysis of TIL specificity

Autologous PBMC were cultured for 2–3 days with GM-CSF and the obtained differentiated monocytes were loaded with peptide pools (5 μg ml−1). When more than 1.5 × 106 T cells grew out, peptide pools of p53 were also tested as well as separate smaller pools of HPV-peptides were used. Monocytes were washed and TIL were added (20,000–50,000) in fresh IMDM+10% AB serum. All tests were performed in triplicate and proliferation was determined by incorporation of 0.5 Ci/well [3H]thymidine for the last 16 hr. Phytohaemagglutinin (PHA) (Remel, Germany) was taken along as a positive control for T-cell activation. A 48 hr-supernatant was pooled and analyzed by Th1/Th2 cytometric bead array (CBA, BDbiosciences). Responses were considered positive when three times medium control and above cut-off value of 20 pg ml−1. One TIL and LN culture were stimulated with E6 peptide pool to obtain enough HPV-specific T cells to measure the breadth of the response with respect to single peptide-antigens as described earlier.28 Data was recorded on a FacsCalibur or LSR-II (BD biosciences) and analyzed using Flowjo software (Treestar, USA).

Isolation of T-cell clones

T-cell clones from the TIL of Patient 27 were isolated using limited dilution as described earlier.29 T-cell clones were tested for phenotype and TCRVβ usage by flow cytometry and specificity was tested by 48-hr proliferation assay on peptide or protein loaded B-LCL. Cytokine production was analyzed by ELISA and CBA.

Suppression assays

The capacity to suppress the proliferation of stimulated naive CD4+CD25+ cells by the T-cell clones of Patient 27 was tested as previously described.30

Results

Circulating HPV16- and p53-specific T cells are present in patients with head and neck cancer

A total of 50 patients with HNSCC in a variety of regions were included in this study (Table 1). In 41 and 43 cases we were able to test the tumor tissue of the biopsy for the presence of HPV and the over-expression of p53 protein, respectively. The patients presenting with an HPV16+ tumor were slightly younger (average 60.4 years) compared to HPV16- and unknown oropharyngeal tumors (average 63.3 years) or all HPV-patients (average 65.8 years).

Table 1. Patients with squamous cell carcinoma of the head and neck region
inline image

Of the 41 tested tumors, 12 contained DNA of HPV 16. It is known that HPV infection is often present in the oropharynx. Indeed, when we divided our cohort according to region 11 of the 12 HPV16+ tumors were located in the oropharynx region (52% of all oropharyngeal tumors) and 1 was present in the oral cavity (8% of oral cavity tumors). No HPV was detected in 10 tested tumors of the hypopharynx (Table 1). In 12 out of 43 tumors p53 was over-expressed, more often (6/10) in the tumors of the hypopharynx (Table 1).

PBMC were isolated and directly tested for the presence of specific T cells to peptides of the HPV 16 oncoproteins or p53 by LST (four examples are shown in Fig. 1). In 17 of 47 tested patients a proliferative response to HPV16 was detected. Interestingly most responses to HPV were found in patients with an HPV+ tumor. There is a trend that the detection rate is higher in this group (7 out of 11 patients displayed a detectable response to HPV16) than in patients with a HPV- tumor (7 out of 28 patients; p = 0.061 Fisher's exact). Only in a minority of patients (5/17) this proliferative response was accompanied by detectable IFNγ production (Patient 8, 9, 22, 28 and 52; range 77–440 pg ml−1). The response of Patient 8 was also accompanied by detectable IL-10 (28 pg ml−1). Notably, in four of these five cases the patient's tumor was HPV16 negative.

Figure 1.

HPV- and p53-specific T cells circulate in the blood. Lymphocyte stimulation test of freshly isolated PBMC. Proliferation of T cells stimulated with pools of peptides indicated at the x-axis after 7 days of culture. The positive control is the memory response mix (MRM). Shown are the counts of eight separate wells and the mean. Top: Patient 9; positive response to peptide pools covering E6 aa81-158 and p53 aa241-393. Patient 35 responding to E6 aa81-158 and E7. Bottom: Patient 26; weak response to E6 aa81-158. Lower right: Patient 25 no response.

Proliferative T cell responses to p53 were found in the blood of 7 out of 45 tested patients. In only one patient this response was associated with IFNγ-production (Patient 7, 61 pg ml−1). No correlation existed between p53 over-expression or the presence of HPV in the tumor and the detected p53-specific immune response in the blood. Only one of the patients with a p53 over-expressing tumor displayed p53-specific immunity (Table 1).

Thus, within an unselected group of patients treated with head and neck cancer 52% of patients with an oropharyngeal tumor presented with tumor-integrated HPV16 DNA. Furthermore, circulating HPV16-specific T cells are more often detected in this group.

The percentage of circulating and tumor-associated CD4+CD25+Foxp3+ T cells

For 12 patients of whom enough PBMC were available the number of regulatory T cells was determined. The mean percentage of CD25+Foxp3+ cells within the CD4+ T cell population in the blood was 2.7% (range 0.68–6.29, n = 12) and no difference was found between patient groups or HPV status (data not shown). Interestingly a substantial number of CD4+ T cells present in the tumor cultures displayed Foxp3+, the transcription factor associated with regulatory function (average 9%; range 2–25 %, Fig. 2a and not shown).

Figure 2.

HPV-specific T cells are present in HPV+ tumors. (a) Phenotypic analysis of freshly isolated T cells derived from tumor tissue of Patient 53 (left panel) or after 3 weeks of culture (right panel) by flow cytometry. An increase in the percentage of CD4+ and CD8+ T cells (while their ratio remains relatively unchanged) during the 3-week culture period is observed. Within the CD4+ T cell population CD25+Foxp3+ regulatory cells are found. (b) Proliferation assay of TIL of five different patients. Monocytes were loaded overnight with medium only (med), peptide pool of overlapping 22 mer peptides of the HPV16 E6 and E7 oncoproteins. As positive control for T-cell reactivity PHA stimulation was used (PHA). The number of conditions that could be tested depended on the number of TIL obtained. T cells were added and proliferation was measured at 48 hr by the incorporation of [3H]-thymidine (top row). Mean of triplicate + SEM is shown. Supernatants isolated at 48 hr of culture were analyzed by CBA (bottom row). All tumors were HPV16+.

HPV-specific T cells can be isolated from oropharyngeal tumors and draining lymph nodes

We were able to culture T cells from the tumor biopsies of 19 patients. In all cultures, both CD8+ and CD4+ T cells grew out (CD4/CD8 ratio median 2.45, average 5.1 and range 0.1–19). The composition of T cells was not clearly influenced by the culture conditions (Fig. 2a). After a 3-week homeostatic expansion period (no exogenous antigen was added) the T cells were tested for their reactivity against autologous monocytes loaded with indicated pools of HPV16 E6 or E7 peptides (n = 19) and when enough cells were available to monocytes pulsed with pools of p53 peptides (n = 10). Nonspecific stimulation with PHA revealed that the TIL from most tumors comprised a heterogeneous population of T cells that displayed a high capacity to proliferate and to produce the Type 1 cytokines IFNγ and TNFα as well as the Type 2 cytokines IL-5 and IL-4 (Fig. 2b and not shown).

While none of the TIL isolated from HPV negative tumors displayed a proliferative response to HPV (data not shown), we were able to detect HPV16-specific T-cell proliferation in five out of seven patients with an HPV16+ tumor (Fig. 2b). TIL from Patient 27 reacted against HPV16E6 by both proliferation and the production of IFNγ. TIL from Patient 28 reacted with strikingly high proliferation and high production of IFNγ, TNFα and IL-5 against both HPV16 E6 and E7. The TIL from a third HPV16+ patient displayed a modest proliferative response to E6 but without the concomitant production of IFNγ (patient 35; Fig. 2b). In all these patients we had also detected an HPV16-specific response in the blood. The TIL of patient 46 displayed an IFNγ-associated HPV16 E7-specific proliferative response whereas the TIL of Patient 53 displayed a potent IFNγ and TNFα-associated HPV16 E6-specific T-cell response. In the latter two cases no response was detected in the blood suggesting that measurements of HPV16-specific immunity in PBMC may underestimate the total number of patients that have mounted an immune response. Of the ten different TIL that could be tested for the presence of p53-specific T cells, none displayed p53-specific reactivity despite that two patients had tumors that overexpressed p53 (Patient 22 and 31; not shown).

In two patients large lymph node metastasis were palpable in the neck and a needle aspirate was taken (Patient 6 and 35, both HPV16+). For Patient 6, the cells were stimulated ones with pools of E6 or E7 peptides to obtain enough cells to test. These bulk cultures were tested by overnight activation and intracellular cytokine staining.28 As indicated by the up regulation of the activation markers CD137 and CD154, the lymph node cells comprised HPV16-specific T cells reactive to peptide pools of E6 and E7 and which were able to produce both IFNγ and IL-2 (Fig. 3b). The T cells were able to recognize their cognate antigen when naturally processed and presented as E7 protein-pulsed APC were clearly recognized but not APC pulsed with control protein. Unfortunately in the LN of Patient 35 we did not detect HPV-specific T cells although both the TIL and blood of this patient displayed HPV16-specific T-cell reactivity.

Figure 3.

Specificity as well as IFNγ and IL-2 production of TIL by multiparameter flow cytometry. (a) Analyses of the specificity and cytokine production of TIL of Patient 46. T cells were stimulated with single peptides and analyzed for expression of activation markers CD137 and CD154 and the production of the cytokines IL-2 and IFNγ. Cells were gated on CD3+CD4+ expressing cells, the numbers represent the percentage of gated cells. Bottom row; IL-2 and IFNγ expression of CD154+, CD137+ cells gated in the plots of the top row. Numbers indicate percentage of cells in that quadrant. * peptide 21–32 is representative for all other peptides covering E7 since no response was found in this group. (b) Analysis of HPV-specific reactivity in LN cells of Patient 6. T cells were cultured with B-LCL loaded with medium only (med) or peptide pools of E6 or E7, HPV16 E6 or E7 protein and HIV-RT protein (ctrl prot) measured by multi parameter flowcytometry as described in (a). Shown is the percentage of CD4+ T cells expressing activation marker CD137 and CD154 in response to the indicated stimulus as well as the percentage of cells that produced the depicted cytokines within the activated cell population.

Thus in six out of eight patients with HPV-induced tumors of the oropharynx, we detected functionally active HPV-specific T cells in the tumor or lymph nodes.

The breadth of the HPV16-specific CD4+ and CD8+ T–cell response in TIL

To assess the type and specificity of the HPV16-specific T cells in the TIL, the tumor-derived T cells of Patient 46 were stimulated overnight with autologous B-LCL pulsed either with a pool or each individual single 22-mer peptide of HPV16 E7 and analyzed by multiparameter flow cytometry. The CD4+ T cells within this TIL population responded to a sequence located in the first part of the HPV16 E7 protein by the expression of both CD154 and CD137 and the production of IL-2, IFNγ and IL-2, or IFNγ only (Fig. 3a). Not only B-LCL pulsed with the E7 22-mer peptides were recognized but also B-LCL loaded with whole protein confirming that these T cells recognized the naturally processed antigen. The population of CD4+ T cells predominantly produced IL-2 and IFNγ although some T cells only produced IL-2, IFNγ or neither cytokine upon antigen-specific stimulation.

As the number of TIL isolated from Patient 27 was to low to do a direct analysis we decided to isolate T-cell clones from this TIL culture. This resulted in the isolation and expansion of 18 HPV16-E6 specific T-cell clones, 14 of these where CD4+ T cells and 4 CD8+ T-cell clones. Upon antigenic stimulation all T-cell clones proliferated and most T-cell clones produced high amounts of IFNγ. A limited number of clones produced IL-10 (Fig. 4a). Interestingly, in two clones IFNγ production was accompanied by high levels of IL-5 upon peptide or processed protein recognition (Fig. 4c and not shown). All CD4+ T cell clones tested responded to epitopes within the sequence covered by amino acids 41–72 of HPV16E6 and recognized naturally processed E6 protein. At least 3 distinct epitopes were recognized (examples see Figs. 4b and 4c). Unfortunately, the exact CD8+ T-cell epitope recognized could not be determined since the in vitro life span of these cells was limited. Phenotypical analysis of surface markers and intra-cellular FoxP3 by flow cytometry on 2–3 weeks rested T-cell clones revealed that some CD4+ T cells expressed low levels of the transcription factor Foxp3 (Fig. 5a, Clone 4 and 245). Interestingly, two of the three HPV16-specific CD8 T-cell clones expressed high levels of CD25 and Foxp3 as well (Fig. 5a). The analysis of the T-cell receptorVβ (TCRVβ) usage of the T-cell clones revealed the presence of at least five different TCRVβ families. Because the T-cell clones expressing TCRVβ17 recognized different epitopes (Clones 4 and 142; E6 aa51-72 and Clone 245; E6 aa41-62) (Figs. 4 and 5) and also the 2 T-cell clones expressing TCRVβ2 recognized different epitopes (Clone 49; E6 aa51-72 and Clone 251 E6 aa41-62 and 51-72), we conclude that at least 6 distinct HPV16E6-specific CD4+ T cell populations were present in the tumor. It was reported for CD4+ T cells isolated from oropharyngeal tumors that they exert suppressive function via ectonucleotidases CD39 and CD73.31 Interestingly, all isolated CD4+ T cells tested expressed the ectonucleotidases CD39 and CD73 on the surface whereas this is not the case for peripheral CD4+ T cells isolated from a healthy control (Fig. 5b).

Figure 4.

TIL isolated from Patient 27 contains functional CD4+ and CD8+ HPV-specific T cells. Clones were isolated by limited dilution and tested for the recognition of HPV16E6 peptides loaded on autologous B-LCL (black) or B-LCL only (white). (a) T cell function was tested by proliferation (top row), IFNγ production (middle) and IL-10 production (bottom) after 48 h. Clone number is depicted at the X-axis, 14 CD4+ and 4 CD8+ clones were isolated. (b) All clones recognized a sequence between aa41 and aa82 of the E6 protein as shown by the IFNγ production upon stimulation with peptide loaded B-LCL. Shown are 4 typical examples of the recognition of aa41-62, aa51-72 or the overlapping peptides aa41-62 and aa51-72. (c) Not only stimulation with peptide but also with naturally processed protein results in T-cell proliferation (left) and production of cytokines (right). Clone 25 produces IFNγ, TNFα as wells as high amounts of IL-5 whereas Clone 251 mostly produced IFNγ. Two representative clones out of six are shown.

Figure 5.

HPV specific helper and regulatory cells are present within the tumor. (a) Flowcytometric analysis of resting clones by surface markers CD4 and CD8 (top row) and intracellular staining for the Treg markers CD25 and FoxP3 (second row). Gates are adjusted to unstained control. Bottom row shows the TCRVβ usage of the clones as determined with sets of antibodies. Shown is the antibody set that stained positive as depicted in the graphs and the concomitant TCRVβ is indicated. (b) Ectonucleotidase expression was analyzed by flowcytometry for five clones and PBMC. All clones express high CD39 and CD73 receptor (grey fill is background fluorescence). C, The capacity of the T-cell clones to suppress the response of naive T cells was tested by the activation of CD4+CD25- T cells with a pool of mixed B-LCL anti-CD3 antibodies. Proliferation and IFNγ production is measured at 48 hr. The T-cell clones were added in increasing numbers (100/10, 100/50, 100/100). Example of suppressive Clone 4 (left) and helper Clone 251 (right). Proliferation test was performed in triplicate, mean with SEM.

On the basis of the expression of Foxp3 by the isolated CD4+ T-cell clones, as well as on our experience with HPV16-specific T cells in cervical tumor, we tested the HPV16-specific CD4+ T cells for their capacity to suppress the proliferation and cytokine production of allogeneic naïve CD4+ T cells (Fig. 5c). This revealed that in this TIL population suppressive (Clone 4) but also helper T cells (Clone 251) were present. Taken together we conclude that the local HPV-specific T-cell response comprises a mixed population of Th1, Th2 and regulatory T cells, as well as IFNγ producing CD8+ T cells.

Discussion

In this explorative study of the tumor-specific immune response in 50 patients with HNSCC we found HPV16 DNA to be present in almost half of the patients presenting with a malignancy in the oropharynx but not in other regions. Circulating T cells reacting against HPV16 E6 or E7 peptides were detected in the blood, often in patients with an HPV16+ tumor, largely extending previous data on HLA-A2 restricted T cells in a small group of patients7 Similar to what was observed in patients with cervical carcinoma the HPV16-specific T-cell proliferative responses in HNSCC were not/hardly ever accompanied by the production of IFNγ.26, 32 Although these responses in the blood clearly show that adaptive immune responses are induced upon exposure to the HPV viral oncoproteins, this reaction does not necessarily reflect an ongoing immune response against the tumor. This is exemplified by the detection of circulating HPV16-specific T cells in the blood of some patients with HPV-negative tumors, which in a number of cases displayed a similar cytokine profile as described for HPV16-specific reactivity in healthy females.27

Here, we showed that HPV16-specific T cells are present in five out of seven TIL populations and in one out of two tested tumor-draining lymph nodes from patients with HPV16+ tumors. Notably two out of these six patients displayed a local HPV16-specific T cell immunity without a concomitant detectable response in the blood. This indicates that the peripheral T-cell compartment not always reflects the local anti-tumor response. Importantly our data indicate that HPV16-specific T cells are locally present in the majority of patients with HPV16-positive oropharyngeal tumors, this in sharp contrast to the only 32% of the patients with HPV16+ cervical cancers.29 We speculate that the location of the disease is likely to contribute to this as the oropharynx consists of lymphoid tissue and our studies on cervical tumor-draining lymph nodes suggest that they generally contain HPV-specific T cells.28, 29 In addition, the high load of bacterial and fungal coinfections present in the oropharynx of these patients may also act as stimulus for the local immune system.

So far, the data on TIL in HNSCC was limited to the quantification and phenotyping of tumor infiltrating immune cells by immunohistochemical studies and microarray.6, 33, 34 We now show that such infiltrating immune cells comprise functionally different T-cell subsets and importantly that they are reactive to the HPV16 oncoproteins. In TIL we find different subsets of HPV16-specific T cells (Th1, Th2, CTL and Tregs) with different cytokine profiles (IL-2, IFNγ, IL-5) comparable to what is found in anogenital HPV16 induced lesions.29, 35, 36 Together this advocates that local adaptive immune cells might play a role in the response to therapy and the altered survival of this patient group.

P53 specific CD8+ T cells have also been detected in HNSCC by means of tetramer staining, albeit that these cells were present at low frequencies.14 In this study 28% of the tumors overexpressed p53 protein but we did not find evidence for p53-specific T cells in the 10 TIL tested. We have previously shown that it is possible to isolate p53-specific CD4+ T cells from the TIL of patients with ovarian cancer.23 Furthermore both p53-specific CD4+ and CD8+ T cells can be isolated from the vaccine site in our studies on p53 vaccination.22 Although the existence of HPV16-specific regulatory T cells is reported, there is no evidence for the existence of p53-specific regulatory T cells.22, 37 However other types of regulatory T cells may have suppressed the outgrowth of p53-specific T cells. We can not exclude that the high numbers of CD4+Foxp3+ T cells detected in the tumors may have played a role. In some patients circulating p53-specific T cells were detected but there was no obvious relation with either p53-overexpression or the potential HPV-mediated increased degradation of p53 as some tumors did not overexpress p53 and were HPV negative. Previously it has been argued that such p53-negative tumors may represent immune escape variants,13, 14 alternatively the tumors may harbour p53-mutations that do not result in overexpression.

We report that HNSCC are infiltrated by high numbers of CD4+CD25+Foxp3+ T cells and that the frequency of these cells exceeded the frequency in the blood. Furthermore, we isolated both HPV16-specific CD4+ and CD8+ T-cell clones that expressed high levels of Foxp3 under resting conditions. In HNSCC the ratio between CD8+ and regulatory T cells is associated with disease.5, 33 Similarly, in HPV16 induced genital malignancies the presence of FoxP3 positive as well as FoxP3 negative HPV16 specific regulatory T cells in tumor and LN is described30, 36, 38 and a low CD8/Treg T-cell ratio is associated with worse outcome.39 While the regulatory function of the CD4+CD25+Foxp3+ T cells present in TILs needs to be confirmed, at least one of our isolated clones clearly exerted regulatory function. The expression of CD39 and CD73 has been found on CD4+ T cells as well as on the great majority of regulatory T cells in HNSCC, and has been implicated as a mechanism for T cell suppression.31 Analyses of these markers on the isolated T cell clones revealed that all CD4+ T-cell clones highly expressed CD39 and CD73 at the cell surface in resting state, but these included CD4 helper T cells that were not suppressive in a classical suppression assay. Notably, also influenza virus-specific CD4+ helper T cells and regulatory T cells that were isolated from PBMC cultures40 express similar high levels of CD39 and CD73 (unpublished observations). Although the transcription factor Foxp3 is mainly reported in CD4+ regulatory T cells we found CD8+CD25+FoxP3high T cells in TIL and in CD8+ T-cell clones derived from the tonsil region. Siegmund et al. describe this phenotype to be uniquely present in human tonsil tissue,41 however their presence was also demonstrated in human prostate cancer.42 Whether these T cells with tolerogenic phenotype are a reflection of suppressed local immunity or part of an ongoing adaptive immune response can not be concluded from this data.

Stimulation of local immunity might change the balance from a protumor microenvironment to a hostile tumoricidal environment. Recently, different forms of immunotherapy—vaccination to enhance the HPV16-specific T cells response, imiquimod to enhance the local innate immune response, and combination hereof—have been successful in the treatment of HPV16-induced lesions of the vulva.19, 20, 43 Similar to HPV-induced lesions of the vulva, the HPV-induced HNSCC are relatively accessible and one could envisage that these strategies might be applicable to tumors of the oropharynx.

Acknowledgements

The authors thank all patients that participated in this study and also acknowledge S. Uljee for the HPV typing.

Ancillary