Immunological and Clinical Responses in Women with Vulval Intraepithelial Neoplasia Vaccinated with a Vaccinia Virus Encoding Human Papillomavirus 16/18 Oncoproteins¹

VIN³ is a high-risk HPV-associated condition (1, 2, 3) which has increased in incidence over the last 20 years, particularly in younger women (4, 5). The multifocal presentation, high recurrence rate, and uncertain risk of progression to malignant disease makes VIN a difficult disease to treat (6, 7, 8). Intractable symptoms including itch, pain and psychosexual dysfunction are common (9), and in the past many women have been subjected to repeated and disfiguring surgery in an attempt to control their disease (10). The close association with a high-risk HPV infection provides the opportunity for the use of vaccines to treat VIN. HPV oncogenes E6 and E7 are expressed throughout the spectrum of HPV-associated disease and therefore are potential targets for immunotherapy (11). One approach is to generate specific CTL that are capable of recognizing and killing HPV-infected cells (12, 13). TA-HPV (Xenova Research, Ltd., Cambridge, United Kingdom) is a live recombinant vaccinia virus encoding modified versions of HPV 16 and 18 E6 and E7 (14). Borysiewicz et al. (15) administered TA-HPV to eight patients with advanced cervical carcinoma. Three patients developed an HPV-specific antibody response, and one of three evaluable patients developed transient HPV-specific CTL 9 weeks after vaccination. In a multicentric European Organization for Research and Treatment of Cancer Phase II trial, early stage cervical cancer patients were vaccinated, and immune responses were measured before and after immunization. TA-HPV was well tolerated in all patients with only mild to moderate local toxicity and no serious adverse events attributable to the vaccine. After a single vaccination, HPV-specific CTL were found in four patients and eight patients developed HPV-specific serological responses (16). The aim of the present study was to investigate clinical and immunological responses in women with high-grade VIN vaccinated with TA-HPV.

Eighteen women with HPV 16-positive high-grade VIN were recruited from the Vulva Clinic at St. Mary’s Hospital, Manchester, United Kingdom. All were immunocompetent as assessed by a full blood count, complement and immunoglobulin levels, the absolute CD4⁺ count, and a Mantoux test. Vaccination with TA-HPV (∼2.5 × 10⁵ plaque-forming units) was by a single dermal scarification to the deltoid region of the upper arm. The dose used was based on previous studies with TA-HPV and use of vaccinia virus during the smallpox eradication campaign (15, 16). The vaccination site was covered with an occlusive dressing which was changed twice weekly until spontaneous scab dehiscence (usually 6 weeks postvaccination). Any local and systemic adverse events as well as tolerability were assessed twice weekly during vaccination site dressing visits and then at monthly intervals until the end of follow-up. Vulval lesions and symptoms were assessed, and blood for immunological assays was taken at entry, 4, 8, 12, and 24 weeks after vaccination. Vulval lesions were measured after the application of acetic acid and recorded on vulval diagrams. Photographs of the entire vulva and of a designated marker lesion were taken at each visit; the baseline measurements/photographs were taken immediately before vaccination, 1–2 weeks after the screening vulval biopsy. Further biopsies were taken from the marker site at 12 and 24 weeks for pathology and HPV analysis; postvaccination results are based on the 24-week biopsy except where this was not available. For patients 3 and 13, the 12-week biopsy was used for postvaccination analysis. Histological grading of all tissue specimens was performed by two independent consultant histopathologists with a special interest in gynecological oncology, according to strict criteria (17). The United Kingdom Medicines Control Agency, the Gene Therapy Advisory Committee, and the Central Manchester Local Research Ethics Committee approved the study, and all patients gave written informed consent.

HLA-A, -B, -DRB1, and -DQB1 DNA-based typing was by PCR-sequence-specific amplification. DNA was extracted and purified from peripheral blood using the Dynabeads DNA Direct System 2 kit (Dynal AS, Oslo, Norway) and processed according to the manufacturer’s instructions. HLA-A, -B, -DRB1, and -DQB1 molecular typing was performed using the Dynal Allset⁺ sequence-specific amplification “low resolution” kit (Dynal Biotech, Ltd., Bromborough, United Kingdom) according to the manufacturer’s instructions.

DNA was extracted from the vulval biopsies by incubating the samples overnight at 37°C on a rotary mixer in 500 μl of guanidinium isothiocyanate lysis buffer (12 g of guanidinium isothiocyanate, 635 μl of 1 m sodium citrate, 425 μl of 30% Sarkosyl, 50 μl of glycogen, 250 μl of 0.1 m DTT, up to 25.5 ml with double distilled H₂O). Samples were spun for 1 min, and DNA was precipitated from the supernatant using isopropanol. The mixture was centrifuged at 13,000 rpm for 15 min, and the pellet was washed with 70% ethanol for 5 min and then dried at 56°C for 15 min before being dissolved in 400 μl of double distilled H₂O. All specimens were positive for β-globin (18). HPV DNA was detected using GP5+/GP6+ consensus primers (19) and genotyped using type-specific primers for HPV 6/11, 16, 18, 31, and 33 (20, 21, 22). Amplification products were resolved on 2% agarose gels and visualized by ethidium bromide staining.

The amount of HPV 16 DNA in vulval biopsies was determined by quantitative PCR analysis, which measures the real time accumulation of fluorescence (Taqman) (23). Fluorescence was detected by an ABI Prism 7700 Sequence Detection System (PerkinElmer, Wellesley, MA). The nucleotide sequences of the HPV 16 oligonucleotides were: forward primer HPV 16 L1-278F, 5′-CAGATACACAGCGGCTGGTTT-3′ (nt position 278–299 of HPV 16 L1 ORF), reverse primer HPV 16 L1–417R 5′-TGCATTTGCTGCATAAGCACTA-3′ (nt position 396–417 of HPV 16 L1 ORF) and probe HPV 16 L1-302pr 5′-FAM-TGACCACGACCTACCTCAACACCTACACAGG-3′ (nt position 302–333 of the HPV 16 L1 ORF) (24). A commercially available set of primers and VIC-labeled probe for human β-actin were used as an endogenous control (part 4310881E, PE Biosystems, proprietary sequence, Warrington, United Kingdom). All PCR reactions were carried out in Applied Biosystems (Foster City, CA) Taqman universal PCR master mix (containing Amplitaq Gold, DNA polymerase, AmpErase UNG, dNTPs with dUTP, and optimized buffer). PCR was initiated by a preincubation step at 50°C for 2 min followed by an incubation step at 95°C for 10 min. Subsequently, 40 two-step PCR cycles with a denaturation step at 95°C for 15 s and an annealing step at 60°C for 1 min were completed. Samples were simultaneously analyzed for HPV and β-actin (multiplexed real time PCR). All samples were tested in triplicate on two separate occasions with comparable results. SDs for replicate samples were always <10%. The number of copies of HPV 16 per cellular genome was estimated by subtracting the threshold cycle number (C_t) at which HPV 16 was detected from that at which β-actin was detected, to give the δC_t for the sample. The δC_t derived for the HPV 16-positive cervical carcinoma cell line SiHa, which has a known HPV 16 load of 1–2 copies/cell, was then subtracted from the sample δC_t to give the sample δδC_t. The formula 2^δδCt gives the approximate number of copies of HPV 16 per cellular genome relative to SiHa.

Vaccinia-specific IgG was measured in patient sera by ELISA using Wyeth strain vaccinia-infected Vero cell lysates as antigen and mock-infected Vero cell lysates as the control (15). A positive vaccinia-specific IgG response was defined as a titer of >2 at an absorbance of 0.5 absorbance units. A response postvaccination was defined as ≥10-fold increase in the prevaccination titer. HPV 16 and 18 E6 and E7 IgG levels were measured by ELISA using recombinant proteins fused to glutathione S-transferase; glutathione S-transferase without HPV protein sequences was used as specificity control, and any background response to the latter was subtracted to give the HPV-specific OD as previously described (25). IgG levels were measured in sera from a group of women matched by age to the patient group, and with a lifetime history of normal cervical cytology, and no previous history of cervical or vulval pathology (n = 17); all but one were seronegative by HPV 16 L1 VLP serology (26). The mean specific absorbance ± 3 SD of this group of normals for each assay was used as the cutoff value to define seropositivity in the VIN patients. The normal and baseline VIN patient serological responses were compared using Mann-Whitney U (nonparametric) tests. Pre- and day 28 postvaccination VIN patient serological responses were compared by paired Wilcoxon signed ranks tests with P ≤ 0.05 considered statistically significant.

IFN-γ producing vaccinia- and HPV-specific T cells were quantified using ELISPOT. Fifteen of the patients were HLA-A2 positive. Thus HPV-specific T cell responses were assessed using HLA-A2-restricted CD8⁺ nonamer peptides. Cryopreserved patient PBMC were stimulated overnight with Wyeth strain vaccinia virus-infected Vero cell lysates or mock-infected Vero cell lysates (1:2500 in PBS) or with HLA-A2-binding nonamer peptides (Immune Systems, Ltd.), influenza M1 (58–66), EBV BMLF1 (280–288; positive controls), HPV 16 E7 (11–20), E7 (82–90), E7 (86–93) and E6 (29–38; 25 μg/ml), PHA (Murex Diagnostics; 1 mg/ml) or without antigen (medium control). Three replicate wells at 5 × 10⁵ cells/well of a multiscreen 96-well Immobilon-P filtration plate (Millipore, Bedford, MA) coated with anti-IFN-γ antibody (1-D1K) were prepared and processed as per kit instructions (ELISPOT assay kit; Mabtech, Nacka, Sweden). Results were plotted as the number of IFN-γ-producing cells per 10⁶ PBMC after subtracting the background (medium control) responses. A positive ELISPOT was defined as ≥20 spots/10⁶ PBMC. In addition, the ELISPOT responses from all replicate experiments from all HLA-A2 individuals for each E7 peptide (11–20, 82–90, 86–93) were compared from pretreatment levels to those at days 28, 56, and 84 using a paired Wilcoxon signed ranks test with P ≤ 0.05 considered statistically significant.

PBMC were seeded in triplicate wells of a 96-well round bottomed microtiter plate (Alpha Laboratories, Ltd., Eastleigh, United Kingdom) at 2 × 10⁵ cells/well in RPMI 1640 supplemented with 10% human AB serum (Quest Biomedical, Norwich, NY), 100 μg/ml streptomycin and 100 IU/ml penicillin (Life Technologies, Inc., Gaithersburg, MD). PBMC alone (medium control) or PBMC with 50 μg/ml recombinant HPV 16 L2E6E7 protein (TA-CIN; Xenova Research, Ltd.), 50 μg/ml recombinant HPV 6 L2E7 protein (TA-GW; Xenova Research, Ltd.), 35,000 units/ml tuberculin purified protein derivative (Evans Vaccines, Ltd., Liverpool, United Kingdom), or 1 μg/ml PHA were incubated for 5 days at 37°C. During the final 18 h of culture, 1 μCi/well [³H]thymidine (NEN Life Science Products, Boston, MA) was added. The cells were harvested using a Packard 96-well vacuum cell harvester onto Unifilter plates (Packard BioSciences) and left to dry overnight, and 30 μl/well Microscint 20 scintillation fluid (Packard) was added. [³H]Thymidine incorporation was measured using a Topcount scintillation counter (Packard). Results were plotted as the mean number of counts for the antigen-stimulated PBMC divided by the mean number of counts for the medium only (negative control) PBMC, to give the stimulation index. A preexisting proliferative T cell response to HPV 16 L2E6E7 was defined as a stimulation index of ≥2. A postvaccination proliferative T cell response to HPV 16 L2E6E7 was defined as a 2-fold increase in the stimulation index postvaccination compared with the prevaccination value. In addition, a Wilcoxon signed ranks test was used to test for population differences in the responsiveness to HPV 16 L2E6E7 or HPV 6 L2E7 before and after vaccination, with P ≤ 0.05 considered statistically significant.

Vulval biopsy sections (7 μm) were allowed to come to room temperature, fixed in acetone for 10 min, air-dried, and rehydrated in PBS. Endogenous peroxidase activity was quenched by incubation with DAKO peroxidase block (DAKO, Ely, United Kingdom) for 5 min. Sections were washed in PBS and blocked with 10% casein (DAKO) for 10 min at room temperature. Primary mouse monoclonal antibodies to CD1a (Langerhans cell), CD4 (T helper cell), CD8 (cytotoxic T cell), and CD68 (macrophage cell, all DAKO) were prepared in PBS at a concentration of 1:50 (CD1a, CD4, and CD8) or 1:100 (CD68). Antibody (50 μl) was added to each section, and the slides were incubated at room temperature for 1 h in a humidified chamber. Sections were washed in PBS, and a horseradish peroxidase-conjugated goat antimouse secondary antibody (DAKO EnVision+ System) was applied for 30 min at room temperature. Diaminobenzidine chromagen substrate was used to develop the stain, and the reaction was stopped after 10 min with running water. The slides were counterstained with Gills X1 hematoxylin, dehydrated with ethanol, and cleared in xylene. Coverslips were applied with PERTEX mounting medium (Cell Path plc, Powys, United Kingdom). Quantification of infiltrating immune cells was performed using a 10 × 10 eyepiece graticule and a 40× objective lens. Three representative high-powered fields were counted per section in a blinded manner by two independent observers, E. J. D. and P. L. S. CD1a-positive cells were counted in the epidermis, whereas CD4-, CD8-, and CD68-positive cells were counted in the upper dermis. Significance was calculated from clinical responder and nonresponder group median scores using the Mann-Whitney U (nonparametric) test for between group differences and Wilcoxon signed ranks test for within group differences pre- and postvaccination. A value of P < 0.05 was considered statistically significant.

TA-HPV was found to be safe and well tolerated, with no unexpected vaccine-related adverse events reported. Two patients complained of a sore throat (patient 1 on days 10–17 and patient 2 on days 0–7) and one patient developed swelling of the vaccinated arm in the first week postvaccination (patient 5). Patient 9 reported flu-like symptoms after vaccination that coincided with an increased T cell response to influenza virus by IFN-γ ELISPOT.

Our 18 patients had been diagnosed with VIN 3 an average of 6.4 years (range, 6 months–17 years) before recruitment into this study. All patients were immunocompetent as determined by standard laboratory tests. Fourteen women had undergone at least one previous treatment and nine women had undergone more than three previous treatments for VIN 3 (surgical excision or laser therapy) before trial entry.

Clinical improvement as defined by at least a 50% reduction in lesion diameter was evident in 8 of 18 (44%) patients (Table 1; Fig. 1). In most responder cases, there was evidence of a global improvement of other nonmarker lesions. One patient (9) showed complete regression of all her vulval lesions that was confirmed histologically, with PCR also showing viral clearance. Of the seven partial responder patients, three showed improvement in histological grade (patients 5, 11, and 16) but HPV was still detected by PCR analysis. A significant relief from pruritis vulvae and vulval discomfort was reported by nine patients, including five with lesion shrinkage and four others; one of the latter showed histological improvement to VIN 2, and in another no HPV was detected by conventional PCR analysis postvaccination. Four patients were asymptomatic pre- and postvaccination; three of these also showed lesion resolution or reduction. Five patients with persistent vulval symptoms showed no evidence of improvement in the extent of their disease over the course of the study.

VIN can certainly recur after apparently successful eradication by various means so continued follow-up of the patient group was important to determine whether the responses observed here were stable with time. Clinical data are available for a subset of 12 of the trial patients for an extended period of up to 15 months postvaccination (average, 12 months; Table 2). Of these women, eight have been stable, two have shown continued improvement in the extent of their vulval lesions (patient 3, complete response) or symptoms (patient 2), and two have shown disease relapse, with symptoms (patient 4) or lesion measurements (patient 12) approximating those documented before vaccination. All of the patients are being held under every-6-months review at the Vulva Clinic at St. Mary’s Hospital in Manchester, United Kingdom.

Baseline HPV 16 real time PCR analysis revealed an apparently wide variation in viral load between patients, ranging between approximately 1 and 30,000 copies of HPV 16 per cellular genome (Table 1). There was no obvious association between baseline viral load and disease severity (as assessed by extent of vulval lesions, self-reported severity of symptoms, or number of failed previous treatments for VIN 3 at screening). Two patients showed viral clearance by conventional PCR analysis 24 weeks after vaccination. This was confirmed by real time PCR for patient 9, but patient 4 showed viral persistence albeit at a very low level (∼0.3 copy/cellular genome). Overall, 12 patients showed a reduction in viral load postvaccination by real time PCR analysis. This included six of eight of the patients showing an objective clinical response, three of four of the nonresponders with significant symptomatic improvement, and three of six of those who failed to show a clinical response to the vaccine.

Vaccine immunogenicity (Table 1; Fig. 2) was confirmed by evidence of new or increased serological and/or cell-mediated immunity to vaccinia virus in all patients. Six women had evidence of preexisting vaccinia-specific IgG (patients 1, 2, 4, 8, 12, and 15). After vaccination, all 18 women had evidence of vaccinia-specific IgG and in 17 of 18 there was a ≥10-fold increase from the prevaccination titer.

All patients showed a IFN-γ ELISPOT response to PHA (data not shown). Six patients had evidence of preexisting IFN-γ-releasing vaccinia-specific T cells (20–100/10⁶ PBMC) before vaccination (patients 2, 4, 6, 7, 15, and 17). Three of these patients also had preexisting vaccinia-specific IgG (patients 2, 4, and 15). Seventeen women had evidence of IFN-γ-releasing vaccinia-specific T cells after vaccination. In 14 of these women, an increase from the prevaccination level of T cells was documented (range, 2- to 60-fold). Patient 10 showed no T cell response to vaccinia, patients 6 and 7 had no boosted response, and for patient 13 there were no prevaccination data. Overall, these data show clear evidence of responsiveness to the vaccine vector in all patients by serology and/or cell-mediated immunity. Furthermore all of the patients exhibited a classical “clinical take” after scarification with TA-HPV.

To investigate HPV-specific immunity, ELISPOT using HLA-A2 binding peptides was performed. Of 18 patients, 15 were HLA-A2 positive. Memory responses to HLA-A2 influenza and EBV peptides were used as positive controls. Patients 4, 6 and 17, patients 3, 9, 12, 14, 16, and 18, and patient 13 gave specific responses to both influenza and EBV, influenza alone, or EBV alone, respectively (range, 20–300/10⁶ PBMC). All these patients typed for HLA-A2. Patients 1, 2, 5, 7, and 15 were HLA-A2 but did not show an influenza or EBV peptide response (consistently negative or marginally positive on one occasion). As expected, non-HLA-A2 patients 8, 10, and 11 gave no responses to these HLA-A2-restricted peptides. There was therefore evidence of HLA-A2-restricted memory responses to influenza and/or EBV in 10 of 15 evaluable patients. The fact that not all of the patients responded to both positive control antigens presumably reflected the fact that the majority, but not 100%, of individuals are infected with influenza virus and/or EBV, and therefore memory T cell responses cannot be expected from all patients in the sample.

ELISPOT assays were performed with PBMC taken from patients on days 0, 28, 56, and 84 on up to four separate occasions, depending on the availability of lymphocytes. Positive influenza responses were reproducible in 75% of assays across all four time points (range, 50–100%, SE 6%), and marginal or possibly false positive influenza peptide responses were seen in only 9 ± 2% of assays.

We next examined responses to four HLA-A2-restricted E6 or E7 peptides in all patients. No responses were seen in any patient to E6 (29, 30, 31, 32, 33, 34, 35, 36, 37, 38). Patient E7 ELISPOT responses at day 0 were compared with those at day 28, 56, and 84 for peptides E7 (11, 12, 13, 14, 15, 16, 17, 18, 19, 20), E7 (82–90), and E7 (86–93) by Wilcoxon signed ranks tests. The Ps for days 28, 56, and 84 were, respectively, 0.22, 0.02*, and 0.2 for peptide E7 (11, 12, 13, 14, 15, 16, 17, 18, 19, 20); 0.2, 0.15, and 0.002* for E7 (82–90); and 0.3, 0.05*, and 0.04* for E7 (86–93). Thus, there were significant increases in the responses of the patients measured at day 56 and/or 84 postvaccination for each of the E7 peptides tested.

Because HPV-specific T cell responses were generally of a lower frequency than influenza responses and may be only transient in the peripheral blood, a positive HPV-specific T cell response postvaccination was further defined as one in which patient lymphocytes reproducibly reached the threshold of positivity on at least one of the three time points postvaccination (i.e., the threshold was reached in 66–100% of all of the assays performed on that patient’s samples). A borderline response was defined as one in which patient lymphocytes reached the threshold of positivity on two or more of the three time points postvaccination in 50% of all assays performed on that patient’s samples. Three patients (3, 4, and 6) had evidence of an increased response to E7 (11, 12, 13, 14, 15, 16, 17, 18, 19, 20), four patients (1, 3, 6, and 16) had evidence of an increased response to E7 (82–90) and five patients (1, 4, 5, 6and 15) had evidence of an increased response to E7 (86–93) after vaccination. As expected, non-HLA-A2 patients 8, 10, and 11 did not respond to any of the HLA-A2-restricted HPV 16 E7 peptides. In summary, one patient showed increased but low level T cell responses to three different HPV 16 E7 peptides, three patients showed increased but low level T cell responses to two different HPV 16 E7 peptides, and three patients showed increased but low level T cell responses to one of the HPV 16 E7 peptides after vaccination. Five of eight patients showing HPV-specific HLA-A2-restricted peptide responses also had influenza or EBV memory responses. Five of 10 patients negative for HPV responses were either not HLA-A2 (patients 8, 10, and 11) or did not give an influenza/EBV memory response (patients 2 and 7). Four of seven HLA-A2 clinical responders showed increased HPV-specific T cell responses postvaccination; two of these made responses to two different E7 peptides. Four other patients had evidence of increased HPV-specific T cells postvaccination; two of these had preexisting T cells to one or more E7 peptide and showed increased T cell responses to a different E7 peptide postvaccination. These two patients showed a reduction in viral load and significant symptom relief after vaccination; one additionally showed histological improvement. Patient 13 had evidence of an HPV-specific T cell response postvaccination, but there was no prevaccination ELISPOT data. Thus, of the seven patients with increased HPV-specific ELISPOT responses after vaccination, four were clinical responders, two were symptom improvers, and one was a nonresponder patient. Thus, HPV-specific ELISPOT responses are increased over the course of the study, but it is difficult to identify a direct correlation with clinical improvement.

Proliferative T cell responses to HPV 16 L2E6E7 protein were measured by [³H]thymidine incorporation after a 5-day incubation of patient PBMC with antigen. All patients showed a proliferative response to PHA and tuberculin purified protein derivative. The statistical analysis of the proliferative responses to HPV 16 L2E6E7 showed that there was a significant increase in proliferation at day 28 (P = 0.05), day 56 (P = 0.006), and day 84 (P = 0.002) compared with baseline in the group as a whole. By contrast, responses to the control protein HPV 6 L2E7 were not significantly different to baseline at any time point (respectively, P = 0.5, P = 0.6, and P = 0.9). As defined by a stimulation index of >2, patient 15 had evidence of a preexisting proliferative T cell response to HPV 16 L2E6E7 fusion protein and eight others (3, 7, 10, 11, 12, 16, 17 & 18) showed at least a 2-fold increased proliferative T cell responses to HPV 16 L2E6E7 postvaccination (Table 1). Patient 13 had a stimulation index of >2 for HPV 16 L2E6E7 postvaccination, but there was no prevaccination lymphocyte sample. There is therefore clear evidence of vaccine-delivered HPV-specific immunogenicity, as measured in the proliferation assay in 8 of 17 patients.

Serological responses were measured to HPV 16 and 18 E6 and E7 by ELISA. Eight patients had positive IgG responses to HPV 16 and/or 18 E6 or E7 before vaccination (Table 1). Overall, the baseline responses were stable with time and did not differ significantly with sera sampled 84 days after vaccination. Two patients showed a marginal positive response before and a marginal negative response after vaccination. Mann-Whitney U comparison of normal versus VIN patient antibody responses at baseline or at day 28 was not significantly different for any of the antigens tested, except HPV 16 E6. The significance depended on the data from the seropositive individuals. Overall, there is very little difference in serological response to any of the HPV antigens pre- and postvaccination, but there were two patients who showed clearly increased HPV-specific serological responses detected at 4 weeks after vaccination; patient 5 to HPV 18 E7 (4.5-fold) and patient 16 to HPV 16 E6 (2-fold; Table 1). In both cases, IgG levels returned to the prevaccination level by 12 weeks postvaccination. Both patients with HPV-specific serological responses also showed increased HPV-specific T cell responses and demonstrated clinical improvement over the course of the study.

After ELISPOT, proliferation, and serology assays, three evaluable patients showed no HPV-specific immune responses after vaccination; this included the patient exhibiting a complete clinical response. Only patient 16 showed increased HPV-specific immunity by all three assays. Overall, HPV-specific immunogenicity defined by these various criteria was established for the vaccine for 13 patients, 7 by ELISPOT and a further 6 by proliferation.

The mean number of lesion-associated CD1a (Langerhans cell)-, CD68 (macrophage)-, CD4 (T helper cell)-, and CD8 (cytotoxic T cell)-positive immune cells was determined by immunohistochemistry pre- and postvaccination for each patient. Clinical responder patients had significantly more lesion-associated CD1a, CD4, and CD8-positive immune cells before vaccination than nonresponders (P = 0.02, P = 0.0004, and P = 0.001, respectively; Table 3; Fig. 3). There was no difference between responder and nonresponder patients in lesion-associated CD68-positive immune cells before vaccination. The number of CD1a- and CD68-positive immune cells did not change after vaccination for responder or nonresponder groups of patients. By contrast, there were significantly more CD4- and CD8-positive immune cells in postvaccination nonresponder biopsies than prevaccination (P = 0.05 and P = 0.001, respectively). A similar trend was seen for responder patients, with increased numbers of lesion-associated CD8-positive immune cells postvaccination compared with prevaccination, although this did not reach statistical significance.

In this study, we have investigated the use of a recombinant vaccinia virus encoding modified HPV 16 and 18 oncoproteins as immunotherapy for high-grade VIN. VIN is a chronic, relapsing condition in which conventional surgical treatment is disfiguring and often of limited therapeutic benefit (10). After vaccination, 12 of 18 patients showed either an objective clinical response or symptom relief. This compares favorably with the success rates of other therapies for VIN 3 (27, 28, 29). This Phase II trial was not placebo controlled, but given the history of chronic disease in our patient group, in whom multiple previous treatments had been unsuccessful, spontaneous clinical improvement over a 6-month period seems unlikely. Spontaneous regression of high-grade VIN has been reported in the literature in occasional case reports and in a single retrospective case note review (30, 31, 32). There is a striking similarity in the patient profiles and clinical appearance of the lesions where spontaneous regression has been reported. Women are typically nonwhite and much younger (median age, 19.5 years) than the case mix of patients reported here. A significant proportion of patients who show spontaneous regression present during pregnancy or in the immediate postnatal period with asymptomatic, multifocal, pigmented lesions in the perineal region. This clinical description differs substantially from that observed in the present study, in which all patients were Caucasian and nonpregnant with a mean age of 41.1 years. All but four of the patients were symptomatic, and only one patient had pigmented lesions (patient 2). This patient failed to respond to vaccination. Most of the lesions were found over the labia minora and posterior fourchette, with perineal involvement only present in those women with a chronic history of relapsing high-grade VIN (patients 2, 5, 7, 8, and 18). Thus, spontaneous regression is observed infrequently and in a very distinct population of patients, not represented in this study. Nevertheless, without a placebo group, it cannot be concluded with certainty that the clinical responses observed here were directly attributable to the effects of vaccination. A future option may be to compare vaccination with surgical excision or laser therapy for their relative rates of disease resolution, recurrence and symptom relief. Conventional treatment modalities for high-grade VIN can produce disease reduction and symptomatic improvement in the short term but can be associated with significant morbidity, disfigurement, and high rates of recurrence in the long term (33, 34). Vaccination resulted in minimal morbidity and no disfigurement, but continued clinical follow-up of the vaccinated patient cohort is essential to determine the longevity of the responses observed.

All of the patients demonstrated a vaccinia-specific humoral and/or cell-mediated immune response after vaccination, confirming the immunogenicity of TA-HPV. A number of patients had evidence of preexisting vaccinia-specific immunity, probably because of previous smallpox vaccination (not specifically recorded). All six women with preexisting vaccinia-specific T cells failed to show an objective clinical response over the course of the study. An explanation for this observation may be that vaccinia-specific T cells rapidly eliminate vaccinia-infected target cells and thus reduce the immunogenicity of the HPV vaccine, although previous clinical studies with this vaccine have not observed this effect (15).

Antigen-specific ELISA, proliferation assay, and ELISPOT methods were used to assess HPV-specific immunity after vaccination. The measurement of appropriate HPV-specific immune responses pre- and postvaccination requires an accurate prediction of the nature and kinetics of the mechanistically important immune responses in the regression of HPV-associated lesions. HPV-specific CTL responses are assumed to play an important role, but there is no knowledge of the precise specificity(ies) of the principal T cells, their longevity in the peripheral circulation, or the best assays for detection. ELISPOT offers advantages over CTL assays, which are logistically demanding on time and lymphocytes because they frequently require multiple T cell restimulations and the generation of appropriate antigen-expressing target cells (35, 36). ELISPOT may be more sensitive and quantitative than CTL assays, but the limited patient lymphocytes for each time point made exhaustive repetition of assays impossible. Two observations facilitated the use of HLA-A2-restricted peptides in ELISPOT. First, all trial patients had HPV 16-associated disease, and indeed there is a significantly higher frequency of HPV 16 infection (91%) in high-grade VIN than in local cervical cancer patients (61%). Secondly, there is a significant over representation of HLA-A2 in these patients compared with local controls (37). In the vaccinated patients, influenza- and EBV-positive control assays consistently demonstrated the competence of the lymphocytes in recall responses measured by ELISPOT at each time point. HPV-specific ELISPOT responses detected after vaccination were at a lower frequency and present transiently. The transient nature of HPV-specific CTL has been documented before in follow-up studies of cervical intraepithelial neoplasia patients who naturally clear their HPV infections (38). The HLA-A2-restricted ELISPOT assay using four HPV peptides is obviously limited in assessing full potential CD8⁺ HPV immunity. Nevertheless, there were significant increases in the ELISPOT responses of patients measured at day 56 and/or 84 compared with baseline for each of these peptides. Individually, 5 of 15 patients showed the strongest vaccine-related responses to HPV peptide E7 (86–93), 4 of 15 to HPV 16 E7 (82–90), and three to HPV 16 E7 (11, 12, 13, 14, 15, 16, 17, 18, 19, 20) postvaccination. These responses were also obviously preexisting in two patients, and the changes seen after vaccination may represent a boost to existing immunity rather than a de novo response in the others. Significantly, all but one of the VIN patients showing a “new” HPV-specific ELISPOT response also showed clinical responses either in terms of lesion shrinkage or by histological improvement and/or symptom relief. Thus, with one exception, these data support the hypothesis that immunological factors contribute to disease regression (39, 40, 41, 42, 43). Such IFN-γ-releasing T cells have been shown to transfer tumor immunity in other model systems (44). It is therefore possible that the measured immunity is at least a correlate of the clinically relevant immunological responses that are important in promoting HPV-associated disease resolution.

The presence of peripheral blood-derived HPV-specific T cells measured by different methodologies in patients with HPV-associated lesions is not necessarily associated with viral clearance or disease resolution (45, 46). The proliferative responses of the vaccinated patients were significantly increased at all postvaccination time points compared with baseline for HPV 16 L2E6E7 but not for HPV 6 L2E7. The strongest increase in proliferative HPV-specific T cell responses seen in nine of the patients did not correlate precisely with clinical outcome. Four of these patients were nonresponders (patients 7, 10, 17, and 18) and did not demonstrate increased HPV-specific CD8⁺ T cell responses by ELISPOT postvaccination. The remaining four patients with clearly increased HPV-specific proliferative responses were clinical responders (patients 3, 11, 12, and 16). Two of these patients additionally showed increased HPV-specific CD8⁺ T cell responses by ELISPOT postvaccination (patients 3 and 16); patient 11 was not evaluable by HLA-A2-restricted ELISPOT. Thus, individually, increased HPV-specific proliferative responses did not appear to correlate with clinical responsiveness to the vaccine unless they were also associated with increased HPV-specific CD8⁺ responses. Of the 18 patients, 12 demonstrated a reduction in viral load postvaccination; all 6 women who failed to show a reduction in viral load demonstrated strongly increased proliferative T cell responses to HPV 16 L2E6E7 after vaccination. If such proliferative responses reflect T helper cell activity, this may not be sufficient to deliver the clinical response but probably is necessary for the development of a CD8⁺ response (47, 48).

Clinical responsiveness was associated with increased numbers of lesion-associated CD1a (Langerhans), CD4 (T helper), and CD8 (cytotoxic T cell)-positive cells in situ before vaccination. This suggests that pretreatment local immune factors are critical in promoting the capacity for responsiveness to immunotherapy. Furthermore, pretreatment biopsies may provide a means for prognostic assessment for patients undergoing vaccination therapy. Previous work has shown an association between reduced numbers of Langerhans cells and increasing levels of dysplasia in the cervix (49), indicating a role for local professional antigen-presenting cells in viral clearance. Langerhans cells have been shown to push T cell responses toward a predominantly T helper 1 phenotype (50), and the natural regression of warts and papillomas is associated with an extensive infiltration of lymphocytes (39, 40). In this study, nonresponders showed significantly higher numbers of CD4- and CD8-positive immune cells in situ postvaccination, but at lower levels overall compared with responding patients. The local expression of cytokines is likely to direct the lymphocyte response toward a predominantly T helper 1 or T helper 2 phenotype, and this in turn will determine the efficacy of the resulting immunity (51). It is also possible that other key effectors, including natural killer and lymphokine-activated killer cells that we have not measured, play an important role in producing disease regression (52). The future challenge is how to measure HPV-specific T cell responses in the lesions themselves. Double staining of lesion-associated immune cells with antibodies directed against CD4 or CD8 molecules and markers of activated T cells, including granzyme B, T cell intracellular antigen-1, and OX40, may provide useful information (53, 54). Alternative assays, including intracellular FACS, which measures the production of cytokines such as IFN-γ by T cell subsets (55) and quantitative PCR analysis of cytokine transcripts in the lesions themselves (56, 57) are being developed and may provide useful alternatives to ELISPOT and proliferation assays.

We examined HLA class I expression in successive patient biopsies by methods described previously (3) and there was consistent down-regulation of one or more HLA alleles in 4 of 18 patients (data not shown). Three of these patients showed an objective clinical response, and the other had symptom relief. These observations may reflect the immune selection of the lesion, which will further confound the efficacy of any vaccination that generates HPV-specific T cells. However, there is also evidence of heterogeneity of HLA expression in the lesions and together with non-HLA restricted effector mechanisms playing a role, it is too complex to look for simplistic links to outcome.

There was a large variation in the HPV 16 viral load between patients (<1–35,000 copies of HPV per cellular genome). Patients stratified by <10, >10, and >100 HPV copies/cellular genome did not differ in disease severity or clinical responsiveness to the vaccine. A high viral load may indicate a productive HPV infection with multiple episomal copies of HPV in each cell. In contrast, a low viral load may indicate either viral integration into the host cell genome or a low level of productive HPV infection by a subset of cells (58). Twelve patients demonstrated a reduction in viral load postvaccination, and this varied from a 2-fold decrease at one extreme to a 70-fold decrease at the other. The relative change in viral load was not obviously correlated with any clinical response in the patients. We used DNA derived from whole vulval biopsies to measure viral load in this study. Because these biopsies contained not only cells from the lesion but also stromal cells, the measured viral load is not an accurate estimate of the viral load in the lesions. The use of laser capture to identify and isolate dysplastic cells from lesions before extracting DNA would allow a more accurate assessment of viral load for these studies (59, 60). It is clear that estimates of viral load will be improved by taking into account analyses on multiple biopsies from the same patients over a longer period of follow-up (24). Another major confounding factor is the possibility that high-grade VIN lesions result from HPV 16 field effects, yielding heterogeneous polyclonal disease as has been shown in cervical neoplasia (61).

There are a number of therapeutic and prophylactic immunotherapies being investigated for HPV-associated disease (62). The results presented here provide a promising platform on which to refine the vaccination strategy and highlight some of the difficulties that need to be addressed before any definite conclusions can be drawn from clinical trials such as this. This is the first demonstration of clinical and immunological responses after immunotherapy in women with HPV 16-associated high-grade VIN. Vaccination with TA-HPV was associated over the course of the study with objective clinical responses in 44% of patients and symptom relief in a further 22%. To confirm the role of vaccination in these responses, however, subsequent trials should use random allocation of patients to both an immunotherapy group and a placebo group, possibly comparing vaccination with one or other method of conventional treatment.

Whereas 12 women demonstrated a reduction in viral load after vaccination, only one patient showed complete viral clearance. Delivery of cure in this and other HPV-associated conditions, however, may depend on viral clearance or at least sustained anti-HPV immunity, and it is possible that a single vaccination is just not enough. This is now being addressed in a similar patient group by a first immunization with the HPV16 L2E6E7 fusion protein (TA-CIN) followed by a boost with TA-HPV. In preclinical studies, this prime-boost immunization strategy has shown enhanced immunogenicity compared with the use of either agent alone (63).

We thank Dr. Peter Snijders and Dr. Chris Meijer for providing primer sequences for the real time PCR analysis. Thanks also to Godfrey Wilson, Gerald Corbett, Andrew Bailey, Garry Ashton, Stuart Pepper, Yvonne Connolly, Lindsay Jack, and the TA-HPV project team at Xenova Research, Ltd.