Int. J. Cancer: 122, 2791–2799 (2008) ' 2008 Wiley-Liss, Inc.
HPV16 E629-38-specific T cells kill cervical carcinoma cells despite partial evasion of T-cell effector function Karen J. Thomas1, Kelly L. Smith1, Sarah J. Youde2, Mererid Evans3, Alison N. Fiander4, Leszek K. Borysiewicz5 and Stephen Man1* 1 Department of Medical Biochemistry and Immunology, School of Medicine, Cardiff University, Henry Wellcome Building, Heath Park, Cardiff, United Kingdom 2 Department of Oral Surgery, Medicine and Pathology, School of Dentistry, Cardiff University, Cardiff, United Kingdom 3 Clinical Oncology, Velindre NHS Trust, Whitchurch, Cardiff, United Kingdom 4 Department of Obstetrics and Gynaecology, School of Medicine, Cardiff University, Heath Park, Cardiff, United Kingdom 5 Medical Research Council, London, United Kingdom Persistent human papillomavirus type 16 (HPV16) infection is associated with the development of more than 50% of cervical cancers. The HPV16 E6 and E7 oncoproteins are constitutively expressed in cervical carcinomas and are attractive targets for cytotoxic T lymphocyte (CTL)-based immunotherapy. However, cervical carcinomas may possess multiple evasion mechanisms for HPV16 E6/E7-specific CTL. In this study, we investigated whether HPV161 cervical carcinoma cell lines (CaCxCL) could evade all effector functions of HPV16 E629-38-specific T cells. Such CD81 T cells were detected in the blood (4/10) or invaded lymph node (1/1) of cervical cancer patients using HLA-A*0201/HPV16 E629-38 tetramers after in vitro stimulation. T cells cultured from 3 different donors killed HPV16 E629-38 peptide-pulsed target cells but not HPV161 CaCxCL in 51Cr release assays. The absence of killing correlated with limited T-cell degranulation against CaCxCL, but this was not due to antigen processing defects per se; CaCxCL could induce specific T-cell release of IFN-c and TNF-a, and CaCxCL could be killed in longer cytotoxicity assays (>20 hr). Interestingly, the ‘slowÕ killing of CaCxCL could be partially inhibited by concanamycin A, a known perforin inhibitor. The results suggest that CaCxCL was only partially activating T cells, but this was still sufficient for slow killing. Overall, our results highlight the need to examine multiple T-cell effector functions in the context of endogenous antigen presentation by tumour cells. In this study, testing for cytotoxicity using short-term assays only would have ruled out a candidate epitope for immunotherapy. ' 2008 Wiley-Liss, Inc. Key words: human papillomavirus; T lymphocytes; cervical cancer
Worldwide, cervical cancer (CaCx) is the second most common cancer in women. Cervical neoplasia, both invasive cervical carcinoma and premalignant cervical intraepithelial neoplasias, are associated with persistent human papillomavirus (HPV) infection. HPV 16 is the most prevalent HPV type globally, being found in more than 50% of cervical cancers.1 Despite the recent introduction of prophylactic vaccines for HPV, which could prevent cervical cancer in future generations, there is still a need for research into immunotherapeutic approaches against HPV-induced disease. Conventional treatment modalities for recurrent/advanced cervical cancer have low success rates.2 Furthermore, there exists a global reservoir of women with persistent HPV infection, for whom prophylactic vaccines will not be effective. These women (particularly those in developing countries) are likely to develop premalignant cervical intraepithelial neoplasia (CIN3) and cervical cancer without clinical intervention. The HPV E6 and E7 gene products are attractive candidate target antigens for immunotherapeutic approaches. E6 and E7 are nuclear proteins that are constitutively retained and expressed in cervical tumour cells, and have immortalizing and transforming properties.3,4 The therapeutic potential of CD81 cytotoxic T lymphocyte (CTL) directed against HPV16 E6 and E7 oncoproteins has been shown in murine studies.5–7 However, studies investigating the function of human HPV-specific CTL have been limited, mainly because it has been difficult to isolate and propagate such CTLs.8,9 Consequently, there have been only a few studies characPublication of the International Union Against Cancer
terizing human CTL epitopes derived from HPV16 E6/E7.10,11 Previously, we demonstrated that human CTL, specific for HPV16 E629-38, were unable to kill HPV161 cervical carcinoma cell lines in 51Cr release assays. We proposed that this was due to defects in antigen processing, because we showed that these cell lines had low or undetectable levels of several intracellular proteins involved in the MHC class I antigen processing pathway.12 However, we subsequently demonstrated that CTL against a different peptide epitope, HPV16 E711-20 was capable of killing the same carcinoma cell lines.13 This suggested that the defects in antigen processing in the carcinoma cell lines were not global, but perhaps might be specific for the E629-38 epitope. In this study, we further studied CD81 T-cell responses against this epitope, in particular, investigating whether failure of HPV16 E629-38-specific CTL to kill cervical carcinoma cells in 51Cr release assays correlates with a failure of other CD81 T-cell effector functions. We showed that this was not the case, and that cervical carcinoma cells can still be killed specifically but by slower pathways of cytotoxicity. This highlights the importance of monitoring multiple effector functions of CD81 T cells for cancer immunotherapy. Material and methods Patient samples The study was approved by the South Glamorgan local ethics committee (LREC) and informed signed consent was obtained for all samples. Ten cervical cancer (CaCx) and 8 CIN3 patients were recruited from patients presenting for treatment at the University Hospital of Wales, Cardiff. Healthy donors were laboratory staff; 5 male and 5 female (no known history of cervical dysplasia or neoplasia). Peripheral blood mononuclear cells (PBMC) from patients and controls were screened for HLA-A*0201 expression using the monoclonal antibodies MA2.1 (HLA-A2 and HLAB1714) and CR351-11 (recognizing HLA-A*0201 and HLA-A28, One Lambda, Gateshead, UK). For patients with cervical cancer undergoing radical hysterectomy, lymph node and/or tumour biopsies were obtained at the time of surgery or shortly afterwards on discussion with a pathologist. T-cell lines and clones The derivation of HLA-A*0201-restricted, HPV16 E629-38-specific T-cell clones 3C11 and 7E7 has been described previously.12 The first two authors contributed equally to this work. Grant sponsors: Medical Research Council (UK), Cancer Research UK, Cancer Research Wales, The Royal Society. *Correspondence to: Department of Medical Biochemistry and Immunology, School of Medicine, Cardiff University, Henry Wellcome Building, Heath Park, Cardiff CF14 4XX, UK. Fax: 144-29-20687303. E-mail:
[email protected] Received 3 October 2007; Accepted after revision 3 January 2008 DOI 10.1002/ijc.23475 Published online 25 March 2008 in Wiley InterScience (www.interscience. wiley.com).
2792
THOMAS ET AL.
Briefly, 3C11 was cloned from the PBMC of a healthy donor, after stimulation with dendritic cells transfected with HPV16 E6 and E7 antigens. 7E7 was cloned from the PBMC of a patient with invasive carcinoma following in vitro restimulation with HPV16 E629-38 peptide. Both clones were derived by limiting dilution cloning in 96-well plates using irradiated allogeneic PBMC feeders, phytohaemagglutinin (PHA, Bio Stat, Stockport, UK) and lymphocult T (Biotest Ag, Dreieich, Germany). The clones were subsequently expanded in T75 flasks using irradiated allogeneic PBMC, PHA and IL-2 (Chiron UK, Harefield, UK). Both clones have similar TCR avidity for HPV16 E629-38 based on peptide dose:response experiments (Ref. 12, unpublished observations); however, 7E7 has been used predominantly for experiments because it has superior growth. Cell lines The CIR.A2 cell line is a B-LCL expressing a transfected genomic clone of HLA-A*0201.15 It was grown in RPMI/FCS (RPMI 1640 with 10% FCS, 2 mM glutamine, 100 lg/ml streptomycin, 100 U/ml penicillin and 25 mM Hepes) and 400 lg/ml G418 (Gibco, Paisley, UK). Both CaSki and SiHa cell lines were obtained from the American Tissue Culture Collection. CaSki (ATCC CRL-1550) is a HLA-A2, HPV16-transformed cervical carcinoma cell line derived from a small bowel metastasis,16 and expresses HPV16 E6 and E7. SiHa (ATCC HTB-35) is a HLA-A2 negative, HPV16-transformed cervical carcinoma cell line derived from a primary squamous carcinoma of the cervix.17 Generation of the SiHa-A2 transfectant expressing HLA-A*0201 has been previously described.13 CaSki was grown in RPMI/FCS, SiHa cells were grown in DMEM/FCS (DMEM with 10% FCS, 2 mM glutamine, 100 lg/ml streptomycin, 100 U/ml penicillin, 25 mM Hepes and 1% nonessential amino acids). SiHa-A2 cells were grown in DMEM/FCS with the addition of 400 lg/ml G418. Degranulation assays CD81 T-cell degranulation in response to antigen stimulation was measured by flow cytometry and adapted from Rubio et al.18 In brief, 5 3 105 CD8 T cells were added to 2.5 3 105 target cells (6peptide pulsing) in a 96-well tissue culture plate (Nunc, Loughborough, UK), together with Monensin (1 ll of 1:5 dilution added to well; GolgiStop, BD Pharmingen, Oxford, UK) and 1 ll of CD107a-FITC antibody (BD Pharmingen, Oxford, UK). The plate was briefly centrifuged before incubation at 37C for 4 hr. After incubation, the plate wells were centrifuged again to pellet cells, and the supernatant was removed. The cells were resuspended in cold PBS/0.1% FCS and further stained with anti-CD8-PE antibody for 20 min at 4C. After 2 more washing steps with cold PBS/0.1% FCS, the cells were resuspended in PBS/0.1% FCS/2% paraformaldehyde and analysed by flow cytometry. 51
Chromium release assays Short-term cytotoxicity was measured in a standard 4-hr 51Cr release assay as described previously.8 For the peptide titration experiments, C1R-A2 target cells were pulsed with varying doses of the peptides for 1 hr after labelling for 2 hr with 51Cr (Na251CrO4, Amersham Biosciences, Little Chalfont, UK). Target cells were washed twice with RPMI media before coincubation with T cells. Infection of target cells with recombinant vaccinia virus (TA-HPV) and treatment with recombinant IFN-g has been previously described.12 After 4 hr of incubation, radioactive counts were obtained by b plate liquid scintillation counting on a 1,450 microbeta Trilux liquid scintillation and luminescence counter (Wallac, Milton-Keynes, UK). Overnight cytotoxicity assay For measurement of ‘slowÕ killing, target cells were first labelled with CFSE (Molecular Probes, Invitrogen, UK). Approximately 5 3 106 adherent cell lines were removed from flasks by Trypsin-EDTA (Invitrogen, Paisley, UK) treatment and resus-
pended in 1 ml of PBS. CFSE was diluted to 0.5 mM, and 1 ll of this is added to the cell suspension for 8 min at 37C. The cells were then washed with RPMI/FCS to stop the reaction, and the cells resuspended at 5 3 104 cells per 500 ll in either RPMI/FCS or DMEM/FCS according to the cell line used. The CFSE labelled target cells were then added at 5 3 104 cells per well in 48-well plate overnight. T cells were added the following day at 2.5 3 105 per well in 500 ll 10% AB RPMI, to provide an approximate E:T ratio of 5:1. The plates were incubated overnight before assessment of killing first by fluorescence microscopy and then by propidium iodide (PI) staining. For microscopy, the wells were washed twice with PBS to remove RPMI and nonadherent T cells, before digital photographs (Leica DM IRBE microscope with a Hamamatsu ORCA-ER camera, and Improvision Openlab software) were taken of individual wells. Next, the remaining target cells were removed by trypsin-EDTA (Invitrogen, Paisley, UK) treatment and resuspended in 200 ll of PBS (containing 1% FCS). PI (Sigma, Gillingham, UK) was added to the resuspended cells at a final concentration of 2.5 lg/ml before immediate flow cytometric analysis. Concanamycin A (CMA, Sigma, UK) was dissolved in DMSO, and was used in some overnight experiments by treating CTL with 100–1 nm of CMA for 2 hr19 before washing the CTL and adding them to target cells. The % inhibition of killing in the presence of CTL was calculated from (% PI-positive CaSki cells 2 % PI-positive CaSki cells 1 CTL)/(% PI-positive CaSki cells), and multiplying by 100. Cytokine assays Cytokines secreted by 7E7 T cells after coculture with target cells were measured initially using a cytometric bead array kit (Th1/Th2, Becton Dickinson (BD, Oxford, UK)). The target cells were dispensed into 24-well plates at 1 3 105 cells/well in 2 ml of RPMI/FCS media. Sixty microlitres of supernatant was removed from the wells for the zero time-point sample, then 5 3 104 7E7 T cells were added, and further 60 ll samples removed at 4 and 8 hr. Supernatants were analysed using the CBA kit according to manufacturer instructions, and concentrations of individual cytokines determined using BD CBA software. The secretion of IFN-g by T cells was also measured using an IFN-g ELISA kit (Mabtech, Nacka Strand, Sweden). T cells at 105/well were coincubated with 5 3 104 adherent target cells (prepared as earlier) in 48-well plates for 18–20 hr. Supernatants (250 ll) were removed, and 100 ll of this was analysed for IFN-g content compared to standards in an IFN-g ELISA. Detection of HPV1629-38-specific T cells using tetramers This was carried out as previously described in Youde et al.8 PElabelled tetrameric HLA-A*0201/HPV16 E629-38 complexes were generously provided by Dr R. Dunbar and Prof. V. Cerundolo (Oxford University) and synthesized as previously described.20 Short-term T-cell lines were generated for tetramer staining as previously described.12 In brief, 2 3 106 PBMC in 1 ml of RPMI1640 (10% human AB serum) were stimulated with 20 lg/ml of peptide and supplemented with 20 ng/ml rIL7. On day 3, rIL2 was added at a final concentration of 10 U/ml. The T cells were restimulated every 7 days (up to day 21) by coculturing with irradiated peptide pulsed (10 lg/ml) autologous PBMC at a 2:1 T cell:PBMC ratio. Recombinant IL2 at 10 U/ml was added every 3 days. Results CD81 T cells recognizing HPV16 E629-38 can be detected in the blood of patients with cervical cancer We have previously reported the derivation of 2 HLA-A*201 restricted human cytotoxic T-cell clones specific for HPV16 E62938 that were obtained from a healthy donor and a patient with cervical cancer respectively.12 To examine whether T cells recognizing HPV16 E629-38 occurred frequently in patients with HPVassociated cervical neoplasia, we used tetramers of HLA-A*0201/
PARTIAL EVASION OF HPV-SPECIFIC T CELLS
2793
FIGURE 1 – Flow cytometric detection of CD81 T cells recognizing HPV16 E629-38 in cervical cancer patients using tetramers. The following T-cell lines were tested with HLA-A*0201/HPV16 E629-38-PE tetramers: (a) HPV16 E629-38-specific T-cell clone 7E7; (b) HPV16 E711-20 Tcell line D4; additionally, PBMC from patients were stimulated with the HPV16 E629-38 peptide for between 7 and 21 days before testing with tetramer: (c) Patient CaCx005 day 21 (Stage 1B squamous cell carcinoma, HPV161 tumour and lymph node). (d) Patient CaCx001 day 21 (Stage 2B squamous cell carcinoma, HPV161 tumour); (e) Patient CaCx006 day 7 (Stage 1B adenocarcinoma, HPV type not available); (f) Lymph node lymphocytes from patient CaCx005 day 21; (g) Patient CaCx008 day 14 (Stage 2B squamous cell carcinoma, HPV type not available); (h) Patient CaCx010 day 14 (Stage 4B metastatic, HPV161 lung metastases); (i) Patient CIN006 day 14 (CIN3, no HPV type available). No significant staining (>0.01%) with tetramer was ever obtained with ex vivo PBMC samples (data not shown).
HPV16E629-38 to quantitate specific T cells in blood. The specificity of this reagent was verified by positive staining of an HPV16 E629-38 T-cell clone 7E7 (Fig. 1a) but not an HPV16 E711-20 specific T-cell clone D4 (Fig. 1b). An additional T-cell clone 3C11 (generated from a healthy donor) against HPV16 E629-38 was also stained with similar results to 7E7 (data not shown). When we tested PBMC from cancer patients, there was no significant detection of tetramer staining cells directly ex vivo from any patient tested (data not shown); however, we were able to detect significant populations of tetramer-positive cells if the cells were cultured in vitro with HPV16 E629-38 peptide for 7–21 days prior to flow cytometric analysis (Figs. 1d and 1g–1i). Such CD81 tetramer1 T cells could be detected in the blood of 4 out of 10 patients with invasive cervical cancer (Figs. 1d and 1g–1i, data not shown), and occasionally patients with preinvasive CIN3 (1/8 patients, data not shown) or healthy donors (1/8 donors, data not shown). Interestingly, patient CaCx005 for whom we were unable to detect tetramer staining cells in blood after 21 days of culture (Fig. 1c) did display obvious tetramer positive cells after culture of lymphocytes isolated from an HPV161 tumour-invaded lymph node (Fig. 1f). In general, there were low cell recoveries from HPV16 E629-38 stimulated PBMC cultures, precluding further functional analyses (data not shown). However, sufficient T cells were recovered from the lymph node of CaCx005 to allow comparison with the previously described T-cell clones 7E7 and 3C11.12 The lymph node derived T cells recognized HLA-A*02011 B-LCL (C1R-A2) cells pulsed with exogenous HPV16 E6 29-38 peptide (Fig. 2a). Further-
more, they could recognize endogenously processed E6 antigen, as they could kill C1R-A2 cells transfected with HPV antigen using recombinant vaccinia virus (TA-HPV) (Fig. 2a). However when tested with cervical carcinoma cells (Caski, HPV161, HLAA*02011) cells, there was no cytotoxicity (Fig. 2b). This could be induced if the Caski cells were treated with IFN-g prior to transfection with HPV16 E6 genes. This was also the case for HPV16 E629-38-specific T-cell clones such as 7E7 (Fig. 2c) and 3C11 (Fig. 2d), confirming our previous findings.12 Both 7E7 and 3C11 also failed to kill another HPV161 HLA-A*02011 cervical carcinoma cell line SiHa-A2 (Ref. 13, and data not shown). Therefore, HPV16 E629-38-specific T cells can be detected and expanded from patients with cervical cancer. T-cell lines/clones generated from 3 different individuals (patient blood, patient lymph node and healthy donor blood) had similar cytotoxic specificity; all can kill B-LCL transfected with E6 antigen, but fail to kill a cervical carcinoma cell line (CaSki) that expresses HPV16 E6 as a result of natural transformation. Cervical carcinoma cell lines fail to induce significant degranulation in HPV16 E629-38 CTL The results with the patient samples demonstrated that it was possible to generate short-term T-cell lines against HPV16 E62938, however, to study more precisely the effector function of HPV16 E629-38 specific CTL, T-cell clones (7E7 and 3C11) were used. Previous studies on T-cell recognition of tumour antigens have demonstrated that it is difficult to interpret results when polyclonal T-cell populations are used.21,22
2794
THOMAS ET AL.
FIGURE 2 – HPV16 E629-38-specific T cells fail to kill HPV161 HLA*02011 cervical carcinoma cells. HPV16 E629-38 specific T cells were generated from the lymph node lymphocytes of patient CaCx005 and tested in 51Cr release assays against: (a) C1R-A2 B-LCL targets that were untreated (u), pulsed with HPV16 E711-20 peptide (D), infected with Wyeth vaccinia (s), infected with TA-HPV vaccinia ( ), or pulsed with HPV16 E629-38 peptide (m). (b) Caski cells that were untreated (u), infected with TA-HPV vaccinia ( ), or treated with IFN-g prior to TAHPV infection (¤). (c) T-cell clone 7E7 was tested against 51Cr release assays against Caski cells that were untreated (u), infected with TAHPV vaccinia ( ), or treated with IFN-g prior to TA-HPV infection (¤). (d) T-cell clone 3C11 was tested in against the same targets as 7E7.
Short-term 51Cr release assays primarily measure CTL killing mediated via the granule exocytosis pathway.23 To investigate whether the failure of HPV16 E629-38 specific T cells (7E7 and 3C11) to kill cervical carcinoma cell lines reflected a failure to release lytic granules, we tested surface CD107a expression of T cells.24 Coincubation of T cells with either Caski or SiHa-A2 cervical carcinoma cell lines resulted in low levels of degranulation as measured by CD107a expression (Figs. 3a and 3b). By contrast when exogenous HPV16 E629-38 peptide was presented by the same cell lines, there was appreciable degranulation (Figs. 3a and 3b). Cervical carcinoma cell lines can induce Th1 cytokine production by HPV16 E629-38-specific T cells The results obtained earlier were consistent with a failure of HPV16 E629-38-specific T cells to recognize cervical carcinoma cell lines in the absence of exogenous peptide. This might result from insufficient TCR avidity and/or constraints on epitope generation imposed by inadequate endogenous antigen processing and presentation.12 However, previous attempts to remedy such defects by treating cell lines with IFN-g12 or triggering via CD4025 were unsuccessful, unless accompanied by transfection with HPV16 E6. If epitope density was the only limiting factor, then it would be predicted that these carcinoma cells would not stimulate CD81 T cells to exhibit other effector functions such as cytokine release. The CTL clone 7E7 had been shown to have moderate/low peptide affinity (50% of maximal lysis at 1 nM12), and could be activated to secrete Th1 cytokines by presentation of
exogenous peptide (peptide-pulsed Caski and SiHa-A2 carcinoma cells, Fig. 4b). Interestingly, and in contrast to the 51Cr release assay results (Fig. 2), 7E7 cells could also secrete detectable IFNg and TNF-a in direct response to cervical carcinoma cells (Caski and SiHa-A2, Fig. 4a). There was no cytokine secretion in response to HLA-A*02012, HPV161 cells (SiHa, Fig. 4a) or HLA-A*02011, HPV162 cell lines (C33A) (Fig. 4a), suggesting that the cytokine response was specific. There was no detectable release of IL10, IL4, IL5 or IL2 by 7E7 using either peptidepulsed or unpulsed carcinoma cells (data not shown). The positive cytokine secretion results appear to contradict the negative results obtained with the cytotoxicity assays. This might simply reflect an increased sensitivity for assays measuring cytokine release versus cytotoxicity (51Cr-release).26 To test whether this was the case for 7E7 T cells, cytotoxicity and cytokine release were assayed against varying doses of the cognate peptide HPV pulsed onto HPV negative (B-LCL) target cells. It was found that cytokine release typically required 2–3 orders of magnitude higher doses of peptide than cytotoxicity (Fig. 5), in line with previous studies of antiviral human CD81 T cells.27,28 Similar titration results were obtained using peptidepulsed HLA-A*02011 cervical carcinoma cell targets (data not shown). These results demonstrate that for 7E7 T cells, cytokine release is a more stringent measure of T-cell activation than cytotoxicity. Importantly, they suggest that the cervical carcinoma cells are capable of processing and presenting the HPV16 E629-38 epitope sufficiently to activate an effector function of CD81 T cells.
PARTIAL EVASION OF HPV-SPECIFIC T CELLS
2795
FIGURE 3 – HPV16 E629-38-specific T-cell response to cervical carcinoma cells measured by degranulation assay. (a) T-cell clone 7E7 was tested against cervical carcinoma cell lines (Caski or SiHa-A2) in the absence or presence of HPV16 E629-38 peptide. 7E7 was incubated with cervical carcinoma cell lines at a 2:1 ratio for 4 hr before analysis of CD107a expression by flow cytometry. (b) T-cell clone 3C11 was tested against cervical carcinoma cell lines under the same conditions as 7E7. The results are representative of at least 3 experiments for each T-cell clone.
FIGURE 4 – HPV16 E629-38-specific T-cell response to cervical carcinoma cells measured by cytokine cytometric bead array. 7E7 was cultured with cervical carcinoma cell lines (in absence or presence of HPV16 E629-38 peptide) at a 1:2 ratio for 8 hr, before supernatant was removed for cytokine testing by cytometric bead array.
HPV16 E629-38 CTL kill cervical carcinoma cell lines via slow mechanisms In the absence of the perforin/granzyme pathway, target cells can also be lysed by CTL via slower mechanisms through death
receptors such as FAS or TNF-R.23 To test this with HPV16 E62938-specific T cells, we incubated 7E7 T cells with CFSE-labelled Caski cells over a longer time period (>20 hr). Cytotoxicity was assessed visually by light microscopy and quantitatively by flow
2796
THOMAS ET AL.
cytometry (PI staining). Evidence of cell death was observed microscopically through disappearance of CFSE-labelled Caski cells in wells containing HPV16 E629-38-specific CTL (7E7), but not in those wells containing influenza M158-66-specific CTL (C25, Fig. 6). These results correlated with an increased number of dead Caski cells detected by PI staining (Fig. 6) for the cultures containing both specific T cells and CaSki cells. The specificity of this slow killing was investigated by testing 7E7 T cells on HLA-A*2012 SiHa cervical carcinoma cells. These were not killed significantly when coincubated with 7E7 T cells (Figs. 7a and 7b). By contrast, SiHa cells that had been transfected with genes encoding HLA-A*0201 (SiHa-A2, Figs. 7a and 7b), were killed when cocultured with 7E7 T cells. Similar results were obtained with 3C11 T cells (obtained from a different donor); there was increased cytotoxicity against SiHa-A2 cells versus SiHa cells (Fig. 7c).
FIGURE 5 – Different peptide dose requirements for cytotoxicity versus cytokine secretion for HPV16 E629-38-specific T cells. HLAA21 B-LCL (C1R-A2) targets were pulsed with varying doses of peptide before being tested with 7E7 (HPV16 E629-38 specific) T cells, either for cytotoxicity in 51Cr-release assays, or for IFN-g secretion in ELISA assays. The titration curves are representative of at least 3 experiments, where the maximum response (% specific lysis or pg/ml IFN-g) for each assay was normalized to 100% before plotting results. 7E7 T cells were tested at E:T ratios of between 6.25:1 and 12.5:1 on C1R-A2 targets (% specific lysis at maximum peptide dose is 71%). For the IFN-g ELISA assays, 7E7 T cells were incubated with C1RA2 targets at a 2:1 E:T ratio (IFN-g secretion at maximum dose 2,366 pg/ml). Curves were fitted using Graphpad Prism (IFN-g secretion r2 5 0995, n 5 5, 50% maximum response at 939.6 ng/ml; 51Cr release r2 5 0999, n 5 3, 50% maximum response at 4.7 ng/ml).
The kinetics of killing suggested the possibility that the cytotoxicity of the Caski and SiHa-A2 cells was mediated through FASL (CD178). However, we were unable to inhibit this killing using a blocking antibody directed against FAS (SM1/23; KLS, data not shown). Interestingly, killing but not cytokine secretion by 7E7 T cells could be partially inhibited by CMA, an inhibitor of perforinmediated cytotoxicity (Fig. 8). Therefore, despite the absence of significant killing in short-term 51Cr assays and low levels of degranulation, HPV16 E629-38 T cells could mediate specific slow killing of cervical carcinoma cells.
Discussion We had previously suggested that T cells specific for HPV16 E629-38 might be ineffective for immunotherapy, because cervical carcinoma cells were defective at presenting this epitope.12 This was based in part on the inability of HPV16 E629-38-specific T cells to mediate cytotoxicity against untreated cervical carcinoma cells in 51Cr-release assays. In this report, we studied additional effector functions of HPV16 E629-38-specific T cells after recognition of cervical carcinoma cells. We show that cervical carcinoma cells can induce cytokine release and activate slow cytotoxicity. There have been several studies on HPV16 epitope-specific CD81 T cells in patients with cervical neoplasia,8,29–31 but few have precisely quantitated the frequency of HPV-specific T cells. In this study, we detected T cells specific for HPV16 E629-38 in the blood of patients with cervical cancer (0.02–0.7%, Fig. 1), but only after in vitro culture. The low frequency detection of systemic HPV-specific T cells may reflect the localized nature of HPV infection and disease. In support of this, we were able to detect HPV16 E629-38-specific T cells in a tumour-invaded lymph node of a patient with cervical cancer, even though no CTL could be detected in PBMC. These studies on patients need to be extended but do raise questions about monitoring systemic T-cell responses for HPV-associated diseases. There are several potential reasons why HPV16 E629-38-specific T cells fail to kill cervical carcinoma cells in 51Cr release assays. These could involve either defects at the level of the effector or the target cells. However, the results in this study appear to exclude the most obvious explanations. The CaCxCL used (CaSki and SiHa) could have a global resistance to CTL killing.32–34 However, SiHa-A29,13 and CaSki cell lines9,10,30,35 can be killed by HPV16 E6- or E7-specific CTL. Alternatively, the CTL may have a low avidity for the cognate
FIGURE 6 – HPV16 E629-38-specific T cells can mediate cytotoxicity of cervical carcinoma cells via slow mechanisms. 7E7 (HPV16 E629-38 specific) T cells were cocultured with CFSE-labelled CaSki cells in 48-well plates for 18–20 hr before analysis by fluorescent microscopy. Digital images were captured from individual wells. Following this, remaining CaSki cells were stained with PI, and immediately analysed by flow cytometry. The results are representative of 3–7 experiments for each experimental combination. In the absence of T cells, the mean % (6SD) of PI-positive Caski cells was 13.8 (67.1). When 7E7 T cells were present, the mean % (6SD) was 43.9 (611.6), which was significantly different (p < 0.001, unpaired t-test). A CD81 T-cell clone specific for influenza A M158-66 (C25) was also tested under the same conditions.
PARTIAL EVASION OF HPV-SPECIFIC T CELLS
2797
FIGURE 7 – Specificity of slow cytotoxicity mediated by HPV16 E629-38-specific T cells. This used the same methodology as for Figure 6. (a) 7E7 T cells were tested on HPV161 SiHa cervical carcinoma cells and SiHa-A2 (transfectants) before analysis by fluorescent microscopy. (b) 7E7 T cells were tested as in (a), but target cells were stained with PI before quantitation of dead cells by flow cytometry. (c) 3C11 T cells were tested against CaSki, SiHa and SiHa-A2 cells, and dead cells quantitated by PI staining as in (b). The results are representative of duplicate experiments.
FIGURE 8 – Effect of concanamycin A (CMA) on slow cytotoxicity mediated by HPV16 E629-38-specific T cells. 7E7 T cells were treated with varying doses of CMA prior to set up of overnight cytotoxicity experiments. Cytotoxicity was assessed as for Figure 6. % PI-positive CaSki cells in absence of T cells was 5.72% and in presence of T cells was 42.2%. Supernatants were also harvested to simultaneously assess IFN-g secretion using ELISA as for Figure 5. The results shown are representative of 3 experiments.
peptide-MHC ligand, and fail to recognize the carcinoma cells because of low epitope density. The main argument against this explanation is that the T cells tested were clearly capable of specific functional recognition (cytokine release or slow killing) of CaCxCL. Activation of T cells for cytokine secretion requires a greater epitope density than for activation of cytotoxicity,27,28 and this was also shown for HPV16 E629-38-specific T cells (Fig. 5). Therefore, low epitope density per se cannot explain the lack of killing in 51Cr release assays. How can the apparently contradictory results with the cytokine release (Fig. 4) and 51Cr release/CD107 assays (Figs. 2 and 3) be reconciled? The pattern of effector function (cytokine secretion, no granule mediated cytotoxicity) observed in this study suggests that the T cells were only partially activated. This phenomenon has been demonstrated for murine CD81 T cells using altered peptide ligands.36,37 One hypotheses to explain the results in this
study would be that cervical carcinoma cells preferentially present an altered version of the HPV16 E629-38 peptide that only partially activates CD81 T cells. Increasing the antigen processing capacity of cervical carcinoma cells (IFN-g and transfection12) may increase the ligand density of the ‘wild typeÕ peptide sufficiently to activate the full gamut of T-cell effector functions. Proving this hypothesis will require direct sequencing of peptides eluted from cervical carcinoma cells.38 The perforin/granzyme B mechanism is the dominant mechanism for rapid T-cell killing of virus-infected cells and tumour cells.23 The slower killing mediated by the CD95 (FAS) pathway is only evident when this pathway is compromised.23 We were unable to demonstrate FAS-mediated killing, despite the high level expression of FAS on both CaSki and SiHA cells.34 We found that HPV16 E629-38 T cells were negative for FASL expression before and after incubation with CaCxCL, and no blocking was observed with an anti-FAS antibody (KLS and SM unpublished observations). Moreover, SiHa cells have been shown to be inherently resistant to death receptor-mediated apoptosis.34,39,40 Unexpectedly, we found that the slow killing was partially inhibited by CMA, a known inhibitor of perforin and granule mediated cytotoxicity. CMA has profound inhibitory effects on cytotoxicity in short-term 51Cr assays,19 but to our knowledge this is the first use of CMA in longer term cytotoxicity assays. The partial inhibition by CMA might reflect the existence of other cytotoxicity mechanisms or the biological stability of this compound in tissue culture. The low level of cytotoxic granule release by HPV16 E629-38 specific T cells is clearly ineffective in short-term assays; however, we speculate that tumour cells will succumb slowly to persistent and sustained granule release. The upregulation of antigen processing components on tumour cells by T-cell-derived IFN-g could also contribute to the effect. Recent studies with human T cells that similarly demonstrate the slow elimination of tumour cells in the apparent absence of killing in 51Cr assays26,41,42 suggest that this phenomenon is worthy of further investigation. What are the general implications of this study for T-cell mediated immunotherapy of cervical carcinoma? We demonstrated that HPV16 E629-38 CTL can kill HLA-A*02011, HPV161 cervical carcinoma cell lines that originated from either a metastases
2798
THOMAS ET AL.
(Caski) or primary tumour (SiHa-A2). Only HPV16 E629-38-specific T cells were tested in this article, but our results are consistent with studies of 2 other HPV16 E6 epitopes.43,44 The novel aspect of the current study is that we have shown that lack of killing is not due to lack of T-cell recognition; CaCxCL can induce cytokine secretion and be destroyed by slow killing mechanisms. Slow killing mechanisms (perforin-independent) can mediate tumour cell regression45,46 or tissue damage47 in murine models. Despite many immunotherapy trials, the contribution of different T-cell cytotoxicity mechanisms in mediating tumour regression in patients is still unknown. Our results highlight the importance of investigating multiple T-cell effector functions in the context of antigen presentation by tumour cells. For instance, only assaying in vitro for cytotoxic activity-mediated via granule exocytosis (51Cr release or CD107
assays) could rule out potentially useful T-cell epitopes. The demonstration that HPV16 E629-38 T cells can be grown from patients with cervical cancer, detected at sites of disease, activated to produce inflammatory cytokines and to kill cervical carcinoma cells, suggests that this epitope may be useful for immunotherapy of cervical cancer. Acknowledgements The authors thank the following for their valuable advice: Prof. D. Price (CD107 assays), Dr. N. Blake (antigen processing/cytokine assays), Dr. P. Brennan (apoptosis), Dr. R. Stanton (microscopy) and Dr. E. Wang (CBA assays). They also thank Mrs. Sian Llewellyn-Lacey (MRC Tissue Culture Facility) for mycoplasma testing and Prof. Enzo Cerundolo for helpful advice and kind provision of tetramers.
References 1. 2. 3.
4. 5.
6.
7.
8.
9.
10.
11. 12.
13. 14. 15. 16.
17.
Clifford GM, Smith JS, Plummer M, Munoz N, Franceschi S. Human papillomavirus types in invasive cervical cancer worldwide: a metaanalysis. Br J Cancer 2003;88:63–73. Cannistra SA, Niloff JM. Cancer of the uterine cervix. N Engl J Med 1996;334:1030–38. Crook T, Morgenstern JP, Crawford L, Banks L. Continued expression of HPV-16 protein is required for maintenance of the transformed phenotype of cells co-transformed by HPV-16 plus EJ-ras. EMBO J 1989;8:513–19. Vousden K. Interactions between papillomavirus proteins and tumor suppressor gene products. Adv Cancer Res 1994;64:1–24. Lin K-Y, Guarnieri FG, Stavely-O’Carroll K, Levitsky HI, August JT, Pardoll DM, Wu T-C. Treatment of established tumours with a novel vaccine that enhances major histocompatibility class II presentation of tumour antigen. Cancer Res 1996;56:21–26. Zwaveling S, Ferreira Mota SC, Nouta J, Johnson M, Lipford GB, Offringa R, van der Burg SH, Melief CJ. Established human papillomavirus type 16-expressing tumors are effectively eradicated following vaccination with long peptides. J Immunol 2002;169:350–8. Velders MP, McElhiney S, Cassetti MC, Eiben GL, Higgins T, Kovacs GR, Elmishad AG, Kast WM, Smith LR. Eradication of established tumors by vaccination with Venezuelan equine encephalitis virus replicon particles delivering human papillomavirus 16 E7 RNA. Cancer Res 2001;61:7861–7. Youde SJ, Dunbar PR, Evans EM, Fiander AN, Borysiewicz LK, Cerundolo V, Man S. Use of fluorogenic histocompatibility leukocyte antigen-A*0201/HPV 16 E7 peptide complexes to isolate rare human cytotoxic T-lymphocyte-recognizing endogenous human papillomavirus antigens. Cancer Res 2000;60:365–71. Schreurs MW, Scholten KB, Kueter EW, Ruizendaal JJ, Meijer CJ, Hooijberg E. In vitro generation and life span extension of human papillomavirus type 16-specific, healthy donor-derived CTL clones. J Immunol 2003;171:2912–21. Ressing ME, Sette A, Brandt RMP, Ruppert J, Wentworth PA, Hartman M, Oseroff C, Grey HM, Melief CJM, Kast WM. Human ctl epitopes encoded by human papillomavirus type 16 e6 and e7 identified through in vivo and in vitro immunogenicity studies of HLA-A *0201 binding peptides. J Immunol 1995;154:5934–43. Nakagawa M, Kim KH, Moscicki AB. Different methods of identifying new antigenic epitopes of human papillomavirus type 16 E6 and E7 proteins. Clin Diagn Lab Immunol 2004;11:889–96. Evans M, Borysiewicz LK, Evans AS, Rowe M, Jones M, Gileadi U, Cerundolo V, Man S. Antigen processing defects in cervical carcinomas limit the presentation of a CTL epitope from human papillomavirus 16 E6. J Immunol 2001;167:5420–8. Youde SJ, McCarthy CM, Thomas KJ, Smith KL, Man S. Cross-typic specificity and immunotherapeutic potential of a human HPV16 E7specific CTL line. Int J Cancer 2005;114:606–12. McMichael AJ, Parham P, Rust N, Brodsky FM. A monoclonal antibody that recognizes an antigenic determinant shared by HLA A2 and B17. Hum Immunol 1980;1:121–29. Storkus WJ, Howell DN, Salter RD, Dawson JR, Cresswell P. NK susceptibility varies inversely with target cell class I HLA antigen expression. J Immunol 1987;138:1657–59. Pattillo RA, Hussa RO, Story MT, Ruckert AC, Shalaby MR, Mattingly RF. Tumor antigen and human chorionic gonadotropin in CaSki cells: a new epidermoid cervical cancer cell line. Science 1977; 196: 1456–8. Baker CC, Phelps WC, Lindgren V, Braun MJ, Gonda MA, Howley PM. Structural and transcriptional analysis of human papillomavirus
18. 19.
20.
21.
22. 23. 24.
25.
26.
27. 28.
29.
30. 31.
32.
type 16 sequences in cervical carcinoma cell lines. J Virol 1987;61: 962–71. Rubio V, Stuge TB, Singh N, Betts MR, Weber JS, Roederer M, Lee PP. Ex vivo identification, isolation and analysis of tumor-cytolytic T cells. Nat Med 2003;9:1377–82. Kataoka T, Shinohara N, Takayama H, Takaku K, Kondo S, Yonehara S, Nagai K. Concanamycin A, a powerful tool for characterization and estimation of contribution of perforin- and Fas-based lytic pathways in cell-mediated cytotoxicity. J Immunol 1996;156:3678–86. Dunbar PR, Smith CL, Chao D, Salio M, Shepherd D, Mirza F, Lipp M, Lanzavecchia A, Sallusto F, Evans A, Russell-Jones R, Harris AL, et al. A shift in the phenotype of melan-A-specific CTL identifies melanoma patients with an active tumor-specific immune response. J Immunol 2000;165:6644–52. Jochmus I, Osen W, Altmann A, Buck G, Hofmann B, Schneider A, Gissmann L, Rammensee HG. Specificity of human cytotoxic T lymphocytes induced by a human papillomavirus type 16 E7 derived peptide. J Gen Virol 1997;78:1689–95. Speiser DE, Cerottini JC, Romero P. Can hTERT peptide (540–548) specific CD8 T cells recognize and kill tumor cells? Cancer Immun 2002;2:14–20. Shresta S, Pham CT, Thomas DA, Graubert TA, Ley TJ. How do cytotoxic lymphocytes kill their targets? Curr Opin Immunol 1998; 10:581–7. Betts MR, Brenchley JM, Price DA, De Rosa SC, Douek DC, Roederer M, Koup RA. Sensitive and viable identification of antigenspecific CD81 T cells by a flow cytometric assay for degranulation. J Immunol Methods 2003;281:65–78. Hill SC, Youde SJ, Man S, Teale GR, Baxendale AJ, Hislop A, Davies CC, Luesley DM, Blom AM, Rickinson AB, Young LS, Eliopoulos AG. Activation of CD40 in cervical carcinoma cells facilitates CTL responses and augments chemotherapy-induced apoptosis. J Immunol 2005;174:41–50. Lee SP, Brooks JM, Al-Jarrah H, Thomas WA, Haigh TA, Taylor GS, Humme S, Schepers A, Hammerschmidt W, Yates JL, Rickinson AB, Blake NW. CD8 T-cell recognition of endogenously expressed epstein-barr virus nuclear antigen 1. J Exp Med 2004;199:1409–20. Valitutti S, Muller S, Dessing M, Lanzavecchia A. Different responses are elicited in cytotoxic T lymphocytes by different levels of T-cell receptor occupancy. J Exp Med 1996;183:1917–21. Betts MR, Price DA, Brenchley JM, Lore K, Guenaga FJ, Smed-Sorensen A, Ambrozak DR, Migueles SA, Connors M, Roederer M, Douek DC, Koup RA. The functional profile of primary human antiviral CD81 T-cell effector activity is dictated by cognate peptide concentration. J Immunol 2004;172:6407–17. Ressing ME, van Driel WJ, Celis E, Sette A, Brandt MP, Hartman M, Anholts JD, Schreuder GM, ter Harmsel WB, Fleuren GJ, Trimbos BJ, Kast WM, et al. Occasional memory cytotoxic T-cell responses of patients with human papillomavirus type 16-positive cervical lesions against a human leukocyte antigen-A *0201-restricted E7-encoded epitope. Cancer Res 1996;56:582–88. Evans EM, Man S, Evans AS, Borysiewicz LK. Infiltration of cervical cancer tissue with human papillomavirus-specific cytotoxic T-lymphocytes. Cancer Res 1997;57:2943–50. Nakagawa M, Kim KH, Moscicki AB. Patterns of CD8 T-cell epitopes within the human papillomavirus type 16 (HPV 16) E6 protein among young women whose HPV 16 infection has become undetectable. Clin Diagn Lab Immunol 2005;12:1003–5. Medema JP, de Jong J, Peltenburg LT, Verdegaal EM, Gorter A, Bres SA, Franken KL, Hahne M, Albar JP, Melief CJ, Offringa R. Block-
PARTIAL EVASION OF HPV-SPECIFIC T CELLS
33.
34.
35.
36. 37. 38.
39.
40.
ade of the granzyme B/perforin pathway through overexpression of the serine protease inhibitor PI-9/SPI-6 constitutes a mechanism for immune escape by tumors. Proc Natl Acad Sci USA 2001;98:11515–20. Aguilar-Lemarroy A, Gariglio P, Whitaker NJ, Eichhorst ST, zur Hausen H, Krammer PH, Rosl F. Restoration of p53 expression sensitizes human papillomavirus type 16 immortalized human keratinocytes to CD95-mediated apoptosis. Oncogene 2002;21:165–75. Hougardy BM, van der Zee AG, van den Heuvel FA, Timmer T, de Vries EG, de Jong S. Sensitivity to Fas-mediated apoptosis in highrisk HPV-positive human cervical cancer cells: relationship with Fas, caspase-8, and Bid. Gynecol Oncol 2005;97:353–64. Santin AD, Hermonat PL, Ravaggi A, Chiriva-Internati M, Zhan D, Pecorelli S, Parham GP, Cannon MJ. Induction of human papillomavirus-specific CD4(1) and CD8(1) lymphocytes by E7-pulsed autologous dendritic cells in patients with human papillomavirus type 16and 18-positive cervical cancer. J Virol 1999;73:5402–10. Brossart P, Bevan MJ. Selective activation of Fas/Fas ligand-mediated cytotoxicity by a self peptide. J Exp Med 1996;183:2449–58. Cao W, Tykodi SS, Esser MT, Braciale VL, Braciale TJ. Partial activation of CD81 T cells by a self-derived peptide. Nature 1995; 378:295–8. Skipper JCA, Hendrickson RC, Gulden PH, Brichard V, Vanpel A, Chen Y, Shabanowitz J, Wolfel T, Slingluff CL, Boon T, Hunt DF, Engelhard VH. An hla a2 restricted tyrosinase antigen on melanoma cells results from posttranslational modification and suggests a novel pathway for processing of membrane proteins. J Exp Med 1996; 183:527–34. Aguilar-Lemarroy A, Kirchhoff S, Whitaker N, Gariglio P, zur Hausen H, Krammer PH, Rosl F. Differential sensitivity of human papillomavirus type 16(1) and type 18(1) cervical carcinoma cells to CD95-mediated apoptosis. Int J Cancer 2001;93:823–31. Hougardy BM, Maduro JH, van der Zee AG, de Groot DJ, van den Heuvel FA, de Vries EG, de Jong S. Proteasome inhibitor MG132 sensitizes HPV-positive human cervical cancer cells to rhTRAILinduced apoptosis. Int J Cancer 2006;118:1892–900.
2799
41. Kuball J, Schmitz FW, Voss RH, Ferreira EA, Engel R, Guillaume P, Strand S, Romero P, Huber C, Sherman LA, Theobald M. Cooperation of human tumor-reactive CD41 and CD81 T cells after redirection of their specificity by a high-affinity p53A2.1-specific TCR. Immunity 2005;22:117–29. 42. Griffioen M, Kessler JH, Borghi M, van Soest RA, van der Minne CE, Nouta J, van der Burg SH, Medema JP, Schrier PI, Falkenburg JH, Osanto S, Melief CJ. Detection and functional analysis of CD81 T cells specific for PRAME: a target for T-cell therapy. Clin Cancer Res 2006;12:3130–6. 43. Morishima S, Akatsuka Y, Nawa A, Kondo E, Kiyono T, Torikai H, Nakanishi T, Ito Y, Tsujimura K, Iwata K, Ito K, Kodera Y, et al. Identification of an HLA-A24-restricted cytotoxic T-lymphocyte epitope from human papillomavirus type-16 E6: the combined effects of bortezomib and interferon-g on the presentation of a cryptic epitope. Int J Cancer 2007;120:594–604. 44. Kim KH, Dishongh R, Santin AD, Cannon MJ, Bellone S, Nakagawa M. Recognition of a cervical cancer derived tumor cell line by a human papillomavirus type 16 E6 52-61-specific CD8 T cell clone. Cancer Immun 2006;6:9–16. 45. Seki N, Brooks AD, Carter CR, Back TC, Parsoneault EM, Smyth MJ, Wiltrout RH, Sayers TJ. Tumor-specific CTL kill murine renal cancer cells using both perforin and Fas ligand-mediated lysis in vitro, but cause tumor regression in vivo in the absence of perforin. J Immunol 2002;168:3484–92. 46. Caldwell SA, Ryan MH, McDuffie E, Abrams SI. The Fas/Fas ligand pathway is important for optimal tumor regression in a mouse model of CTL adoptive immunotherapy of experimental CMS4 lung metastases. J Immunol 2003;171:2402–12. 47. Liu AN, Mohammed AZ, Rice WR, Fiedeldey DT, Liebermann JS, Whitsett JA, Braciale TJ, Enelow RI. Perforin-independent CD8(1) T-cell-mediated cytotoxicity of alveolar epithelial cells is preferentially mediated by tumor necrosis factor-a: relative insensitivity to Fas ligand. Am J Respir Cell Mol Biol 1999;20:849–58.
