Base line sera from 35 women with HPV 16 mono-infected high grade cervical intraepithelial neoplasia (CIN 2/3) participating in a vaccination trial with chimeric HPV 16 L1-E7 virus-like particles [17] were used as examples of naturally acquired HPV 16 antibodies. For vaccine-induced HPV 16 antibodies, 72 post-vaccination sera of this vaccination trial were used. Ethics statement [17]: From December 2000 until May 2002, 39 patients were enrolled into the study at 10 German study sites (Berlin, Duesseldorf, Hamburg, Hanover, Jena, Mainz, Munich (2), Muenster, Rostock). Positive votes were obtained from each internal review board of the participating centers. Ethics regulations were coordinated by the ‘Ethik-Kommission der Friedrich-Schiller-Universität Jena an der Medizinischen Fakultät’ (http://www.ethikkommission.uniklinikum-jena.de/Ethik_KommissionderFSUJena.html). All investigations involving humans have been performed in accordance with the principles embodied in the Declaration of Helsinki. The study was performed in compliance with the guidelines of Good Clinical Practice as described in the International Conference on Harmonization of Good Clinical Practices guidelines step 4 (1996). All patients signed an informed consent form. Study participation took 24 weeks with 2 optional follow-up visits after 36 and 48 weeks. Blood samples were collected at baseline, 2 weeks after each vaccination and after 24 weeks. Peripheral blood mononuclear cells (PBL) and plasma were stored frozen until analysis.

A post-vaccination serum collected after obtaining written informed consent from a 21 years old female volunteer 3 years post completion of a 3-dose Gardasil vaccination course was used as positive standard on all assay plates. Approval for obtaining and testing this serum derives from ‘Ethik-Kommission der Bayerischen Landesärztekammer’, Munich, Germany.

Pseudovirions (PSV) were prepared as previously described [18], [19] with some modifications. Pseudovirions were prepared by co-transfecting 293TT cells with plasmids encoding humanized HPV L1 and L2 genes. For HPV 16 we used a bi-cistronic L1 and L2 expression construct, exceeding packagable size. For Production of HPV 18 and BPV 1 PSV we used two plasmids encoding L1 and L2 separately. As a reporter we used a plasmid encoding the enzyme Gaussia luciferase or a plasmid encoding secreted human placental alkaline phosphatase (SEAP). Co-transfection was performed on 4×10⁶ 293TT cells using TurboFect (Fermentas) according to the manufacturer’s instructions with minor modifications. Briefly, 10–15 µg DNA were mixed with 2 volumes of TurboFect and 1 ml of unsupplemented DMEM and added to 293TT cells which were subsequently incubated at 37°C, 5%CO₂ for 72 hours. Absence of cell culture contaminations and authentication of the cell line were determined by the Multiplex cell Contamination Test [20] and a SNP-based human cell line authentication test as described [21].

For pseudovirion extraction, 293TT cells were harvested and the cell pellet was resuspended in an equal volume of lysis buffer. Lysis buffer was prepared by adding 58,3 µl of 10% Brij58 (w/v) (Sigma) and 6,6 µl RNAse A/T (Fermentas) to every ml of DBPS (Invitrogen). Cells were rotated overnight at 37°C to induce pseudovirion maturation. On the next day, salt extraction was performed by adding 0,17 volumes of 5 M NaCl, followed by the addition of 250 units of benzonase for 1 h at 37°C. The pseudovirions were purified by Optiprep gradient and fractions were stored at −80°C.

The SEAP manPBNA was essentially performed as described [16]. As target cells 293TT cells were seeded at a concentration of 15,000 cells per well on 96-well plates in DMEM (Sigma), supplemented with 10% FCS (Gibco, BRL), 1% penicillin/streptomycin (Life Technologies) and 125 µg/ml hygromycin (Roche). The following day, the pseudovirions were diluted in DMEM (1∶5000) and mixed with the sera at different dilutions. After 15 min incubation at room temperature, the medium of the 293TT cells was replaced by 200 µl of the pseudovirion solution. The following controls were included: untreated cells, cells treated with pseudovirions alone, and cells treated with pseudovirions in the presence of neutralizing monoclonal antibodies. Detection of SEAP in the cell culture supernatant was performed five days later with the chemiluminescent SEAP Reporter Gene Assay (Roche) according to the manufacturer’s instructions. All sera were tested in duplicates. The neutralization activity was calculated using the following formula: % neutralization = (Psv_alone−Psv_serum)/(Psv_alone−Psv_Ab)*100%. Sera with at least 70% neutralizing activity were regarded as neutralizing.

Antibodies binding to HPV L1 proteins were measured in a multiplexed glutathione S-transferase capture immunosorbent assay with fluorescent bead-based Luminex technology as previously described in detail [22], [23] [24], [25]. The cut off for the Luminex assay was calculated from the reactivity of pre-vaccine sera that were negative at 1∶100 dilution (<109 MFI). At the 1∶2700 dilution that was used to measure MFI-values in a linear range for the vaccine sera, the maximum reactivity of the above mentioned negative pre-vaccine sera was 10 MFI and therefore used as cut off value for the sera at 1∶2700 dilution.

The 293TT cells originally described as target cell line for the PBNA were not suited for the add-on assay layout for the HT-PBNA since attachment of the 293TT cells to the assay plates was very variable and resulted in high assay variability. Therefore, we generated a HeLa cell line stably transfected with a linearized expression construct for the SV 40 large T-antigen. HeLa cells (tested for absence of contaminations and for authenticity as described above [20]) were transfected and then cultivated under selection using hygromycin B. Several single clones were isolated and tested for transduction by HPV 16 pseudovirions. One clone (HeLaT clone-4) was subsequently used in all assays.

In preceding experiments we tested different cell lines in the HT-PBNA. The 293TT cell line [13] appeared ill-suited for the add-on assay layout of the HT-PBNA, where trypsin-treated cells are added to premixed PSV and sera. 293TT cell attachment to the assay plates was highly variable and relatively slow (>4 hours for >90% cell attachment) causing a significant delay in reporter gene expression. In contrast, more than 90% of cells of a newly established HeLa subline (HeLaT clone-4) stably expressing the SV40 large T-antigen were found to attach within 1 hour of seeding (data not shown). With the HT-PBNA protocol, a range of 500–3000 HeLaT clone-4 cells per well in the 384-well assay plates showed little influence on the intensity of the Gaussia luciferase signal (Figure 2). The luminescence signal in an ATPlite1step™ assay (Perkin Elmer) is directly correlated with cell number/metabolic activity and was proportional to the number of seeded cells, up to 2500 cells/well when confluency was reached after 2 days of incubation. In all subsequent assays, 1500 cells/well were used to ensure that cell density is not exceeding sub-confluency. The robustness of the PBNA towards variations in cell number is crucial for the application in a 384 well format where “edge effects” of the outer most wells often influence cell growth.

Typically, for the luminescence signal of the Gaussia luciferase reporter assay in 384 well plates we observed a coefficient of variation (CV) of less than 3% within rows and less than 6% for the whole plate (data not shown). The higher CV for the whole plate is mainly a consequence of the decay of the Gaussia luciferase luminescence during the time that is necessary to read the 384 well plate (75 sec), even when employing a glow substrate with reduced turnover rates. The luminescence reader measures the plate row-wise so that the decrease in signal is negligible within a row. Sera are diluted within a row and therefore the effect of a decreasing luminescence signal on the calculation of the ED₅₀ is minimal and no correction for position effects is necessary.

PSV preparations of the same or different papillomavirus types can vary considerably in regard to particle concentration, L1 capsid protein to reporter plasmid ratio and transduction activity. A standardization of PSV preparations by L1 or plasmid concentration is therefore not the best indicator of assay performance. Titration of PSV preparations in the HT-PBNA demonstrated significant variation (Figure 3). At high PSV concentrations signals were frequently reduced. All PSV preparations reached similar maximal intensities but the dilutions giving maximal luminescence signal spanned a wide range even for PSV preparations of the same HPV type (e.g. HPV 16 PSV in Figure 3). To investigate the effect of PSV dilution on sensitivity and variability of the PBNA we determined the effective PSV dilution yielding 50% neutralization (ED₅₀) of a serum derived from a Gardasil-immunized individual for different dilutions of HPV 16 and HPV 18 PSV (Figure 4). As expected, lowering the PSV concentrations raised the ED₅₀ values indicating an increasing analytical sensitivity. At low PSV concentrations the ED₅₀ reached a plateau. The maximal gain in sensitivity was typically 3-fold relative to the highest PSV concentrations. However, the variability of the ED₅₀ determination markedly increased with lower PSV concentrations, indicating reduced assay robustness. Therefore, PSV concentrations slightly lower than those generating maximal luminescence intensity were considered to represent the best compromise between low variability and high analytical sensitivity and were used in all future HT-PBNA.

In the manPBNA described by Pastrana et al., sera and PSV are pre-incubated in a separate plate for 1 hour to ensure complete antibody binding before adding the mix to pre-plated, attached reporter cells. In contrast, for the add-on HT-PBNA trypsin-treated reporter cells are added as a suspension to a mixture of PSV and serum in the cell culture assay plates. To investigate the effect of incubation time of the PSV/serum-premix on the ED₅₀-value in the HT-PBNA, our standard serum from a Gardasil® immunized individual was pre-incubated for 2–120 minutes with HPV 16 or HPV 18 pseudovirions before the reporter cells were added. Notably, we found little effect of the PSV-serum pre-incubation time on the ED₅₀-value (Figure 5). For all incubation times investigated ED₅₀-values did not differ significantly for both HPV 16 and HPV 18 PSVs. This indicates that uptake of the PSV by the added reporter cells is sufficiently slow to allow for antibody binding and neutralization irrespective of incubation time. In the automated HT-PBNA we used a standard pre-incubation time of 1 hour which allows the handling of larger batches of assay plates under identical conditions.

At a dilution of 1∶20 we observed that most human sera yielded 60–90% inhibition of bovine papillomavirus 1 (BPV-1) PSV activity in the HT-PBNA, irrespective of whether the sera originated from HPV-vaccinated or non-vaccinated persons. In addition, at a 1∶20 dilution, these sera also neutralized PSV of all other tested HPV types. This effect was not due to reduced viability of the reporter cell line as high serum concentrations rather stimulated cell growth (data not shown). Pastrana and colleagues observed a similar “cross-species inhibition” at 1∶20 dilution for some sera in their SEAP PBNA [16] and consistent with their report, we were also not able to eliminate the inhibition of BPV PSV by heat inactivation, freeze-thaw cycles or centrifugation of the sera. At a dilution of ≥1∶40 none of the sera inhibited BPV PSV activity and sera from vaccinated persons specifically neutralized PSV from vaccine-HPV types or closely related HPV-types only. We therefore conclude that this low titer inhibition is non-specific and not due to cross-neutralizing antibodies.

The reproducibility of titers obtained with the HT-PBNA for HPV 16 and HPV 18 was determined with our serum standard tested a total of 58 times on individual plates in seven independent runs over a period of two months in the course of a larger serological study (Figure 6). A mean coefficient of variation (CV) of 13% was obtained for the ED₅₀-values of 58 repeats with both, the HPV 16 and HPV 18 HT-PBNA. The CV for the intra-run variability with 8–10 repeats per run ranged from 6 to 12%. For a cell-based assay, these CV indicate a low variability. As the underlying assay is newly developed, no data on reproducibility in other laboratories are yet available.

Previously, we analyzed a total of 1271 sera from a different vaccine trial (25% pre-vaccination sera; data not shown) for neutralization of HPV 16, HPV 18 and BPV 1 PSV in the HT-PBNA and found only 8 sera (0.6%) with a BPV 1-specific ED₅₀>80 (maximal BPV-specific ED₅₀ was 120). Considering the neutralization of BPV 1 in the PBNA as unspecific we defined an ED₅₀ of 80 as the cut off value to classify sera positive in the neutralization assay for all HPV types.

First, to determine the PV type-specific detection of antibodies in our PBNA we tested the WHO international standards for antibodies to HPV 16 [26] and HPV 18 [27] using PSV of HPV types 16, 18, 31, 45, 52, 58 and BPV 1. The HPV 16 standard neutralized specifically HPV 16 in the HT-PBNA with an ED₅₀ of 463 (Figure 7), which is at least five times higher than the titers obtained with the manPBNA in five laboratories participating in the collaborative study to assess the suitability of the serum as an International Standard for antibodies to HPV 16 in immunoassays and pseudovirion neutralization assays [26]. With an assigned potency of 10 mIU/µl for the WHO standard after reconstitution as directed in 0.5 ml distilled water, the neutralization titer of 463 in the HT-PBNA would correspond to an analytical sensitivity of 0.864 mIU. The HPV 18 standard neutralized specifically HPV 18 in the HT-PBNA with an ED50 of 579 (Figure 7) corresponding to an analytical sensitivity of 1.105 mIU.

To compare manPBNA with the HT-PBNA, we tested sera from a German HPV 16 L1-E7 chimeric virus-like particles vaccination trial in CIN patients [17] for neutralization of HPV 16 PSV. In addition, data previously obtained with a Luminex-based total IgG-assay using GST-L1 as an antigen were available for these sera. Thirty-six patients with HPV 16 DNA-positive, high-grade cervical intraepithelial neoplasia (CIN) in groups of 12 participants had received a high- or low-dose vaccine or placebo at weeks 0, 2 and 4. A total of 107 analyzed sera included 35 base line (pre-vaccination) sera and 24 post-vaccination sera each from week 2, 4 and 14 of the study.

First, we investigated the performance of the HT-PBNA in detecting neutralizing antibodies as a result of natural infection with 35 available pre-vaccine serum samples using PSVs from HPV types 16, 18, and 31 (Figure 8). All patients from the trial had CIN2+ lesions and were HPV 16 DNA positive at the time of enrollment.

Using an ED₅₀ of 80 as cut-off, among the 35 sera tested, 29 (83%), 6 (17%) and 8 (23%) were found positive for neutralization of HPV 16, HPV 18 and HPV 31 PSV, respectively (Figure 8). The geometric mean titer (GMT) of all sera was 449 (GMT of positive sera: 723), 59 (123) and 69 (299) for HPV 16, HPV 18 and HPV 31, respectively. All HPV 18 and HPV 31 neutralizing sera were also positive in the HPV 16 HT-PBNA, with the exception of one serum that was only positive for HPV 18 (ED₅₀ = 111). The HPV 31 DNA status of the patients has not been determined, but the well-documented cross-reactivity of HPV 16 antibodies against HPV 31 might explain neutralization of HPV 31. These HPV types are closely related and induction of cross-protection against HPV 31 has been demonstrated by the two commercial HPV-vaccines containing HPV 16 virus-like particles [28]. Interestingly, two sera neutralized HPV 31 PSVs with a 3-fold higher titer compared to HPV 16 PSVs, which highly indicates that the antibodies might be derived from an infection with HPV 31.

In general there is a direct, linear correlation between the titers determined in HT-PBNA and manPBNA based on using SEAP, especially for the vaccine-induced high titer sera (Figure 9B). However, the HT-PBNA yields titers about 10-fold higher than the manPBNA (Figures 9A and 9B). In the subgroup of the 35 pre-vaccination sera that contain only antibodies from natural infection, 20 sera were non-neutralizing in the manPBNA. These 20 sera comprise all 6 sera that were negative in the HT-PBNA but also 14 HT-PBNA positive sera with an ED₅₀ GMT of 258 and a maximal ED₅₀ of 1703. All 15 manPBNA positive sera (GMT (ED₅₀) = 250) were also positive in the HT-PBNA (GMT (ED₅₀) = 1895). This indicates a higher sensitivity of the HT-PBNA in the detection of sera with low neutralization titers.

The HT-PBNA also showed a good agreement with a Luminex-based total IgG-assay using GST-HPV 16 L1 as antigen (Figure 9C). An R²-value of 0.7 (0.67 for the 29 HT-PBNA positive sera only) was determined for the linear regression analysis between GST-HPV 16 L1 antibody reactivity (expressed as median fluorescence intensity, MFI) and the HT-PBNA ED₅₀ values. Of the 35 sera, 28 (80%) were considered HPV 16 seropositive by the GST-HPV 16 L1 antibody binding assay (reactivity >109 MFI) and 29 (83%) sera neutralization antibody positive by the HT-PBNA (ED₅₀>80). A Cohen’s kappa coefficient of 0.72 indicated good agreement between these two assays.

For the 72 analyzed post-vaccination sera of the trial, covering a very wide range of titers, a direct correlation of the manPBNA and the HT-PBNA was observed, but again the HT-PBNA yielded approximately 10 fold higher titers (figure 9B). All except two of the post-vaccination sera were positive in both assays. These two sera were negative in the manPBNA and had low titers in the HT-PBNA with ED₅₀-values of 135 and 204, respectively.

As for the low titer pre-vaccine sera, the HT-PBNA and the Luminex-based total IgG-assay showed a good correlation with a R² of 0.61 for the linear regression (Figure 9D). All 72 sera were positive in both assays applying an ED₅₀ of 80 as cut off for the HT-PBNA and 10 MFI for the HPV 16 L1 antibody binding assay. Some sera with relative low reactivity in the Luminex assay compared to the HT-PBNA were most probably caused by handling issues in the Luminex experiment as they had high reactivity in both PBNAs.

An important application of PBNA, ‘the gold standard in vitro assay’, for assessing the protective potential of vaccine-induced antibodies, is the detection of cross-neutralization activity against non-vaccine HPV types. The high sensitivity of our HT-PBNA allows the detection of cross-protective antibodies induced by commercial L1-based vaccines as shown in Figure 10A for our in-house standard serum, a collected from a 21 year old female volunteer 3 years after completion of a 3-dose Gardasil® vaccination. The ED₅₀ of this serum for HPV 16, 18, 31 and 58 was 19308, 1952, 111 and 205, respectively. No neutralization of HPV types 45 and 52 or BPV-1 PSV was observed.

In Figure 10B the reactivity of a high titer Cervarix® serum in the HT-PBNA is shown. In addition to the vaccine-types 16 and 18 with titers >30,000 also cross-neutralization activity against the non-vaccine HPV types 31, 45, 52 and 58 with ED₅₀-values of 2842, 230, 86 and 95, respectively, was detected.

As the PSV are composed of L1 and L2 it is also possible to measure neutralizing antibodies directed against the minor capsid protein L2. In Figure 10C the titration of an experimental serum from a mouse immunized with HPV 16 and HPV 31 L2 (aa 17–36) inserted in the capsid of adeno-associated virus 2 (AAV2) [29] in the HT-PBNA is shown. In addition to the vaccine type HPV 16 PSVs of HPV 18, 45, 58 and BPV are also neutralized by the serum with ED₅₀- values of 1796, 179, 289, 149 and 109, respectively. HPV 31 and HPV 58 however, escape this neutralization.