HPV16-L1-specific Antibody Response Is Associated with Clinical Remission of High-risk HPV-positive Early Dysplastic Lesions

Discussion

In the present study, we showed that an HPV16-L1-specific antibody response is strongly associated with a clinical remission of high-risk HPV-positive early dysplastic lesions and, therefore, could be a useful tool to improve the clinical management of these women.

Only one (5.9%) HPV16-L1 antibody-positive woman out of 17 with a productive high-risk HPV-positive early dysplastic lesion progressed to histologically-confirmed CIN3 during the follow-up period of 66 months. In addition to that, only 4 (8.7%) out of 46 women with a CIN lesion during follow-up were L1 antigen and L1 antibody double-positive.

It is generally accepted that regression of HPV-associated lesions is due to a cell-mediated immune response to early, but not late HPV proteins, such as L1. Nonetheless, our data are not in contradiction to the well-established model that the cell-mediated immune response is the key driver destroying virus-infected cells.

Measurements of this cell-mediated immune response in clinical practice is not easily performed, but the detection of antibodies in patient serum is routine and the standard-of-care for most infectious agents. As shown here, this approach is indicative of an HPV-specific immune competence in patients and positive test results are associated with clinical remission.

The study of Nicholls et al., who provided a complete chronological picture of wart regression, supports our serological data (13). They showed for canine oral papillomavirus that a clinical remission of the lesions were due to a cell-mediated immune response that was closely-followed by seroconversion and production of antibodies to the major coat protein, L1 (14).

On the other hand, our findings regarding the clinical outcome of the L1 antibody-positive women with abnormal smears (LSIL/HSIL) are in clear contrast to the results reported by deGruijl et al. and Ochi et al. Even with a similar setting as ‘prospective, non-intervention, cohort study’ deGruijl et al. reported that seropositivity for HPV16-L1/L2 capsids was strongly-associated with an increased cervical cancer risk, but not clinical remission of early dysplastic lesions (15). Using a HPV16 neutralisation assay Ochi et al. reported similar findings (16). The reasons for these completely different outcomes are difficult to explain.

Our and de Gruijl et al.'s study protocol were nearly identical, looking at the numbers of enrolled patients (133 vs. 112), the cytology-based classification as mild-to-moderate dysplasia, and the follow-up times (56 months vs. 66 months). The study protocol of Ochi et al. differed slightly, recruiting 242 patients with punch biopsy-confirmed CIN1 and CIN2. This invasive procedure was avoided in both cytology-based studies because of the potential negative impact on the natural outcome of HPV infection due to wound-healing processes and ‘non natural’ antigen release. Follow-up time of this study was a little bit longer, with up to 84 months. In all three studies, the age of the patients enrolled ranged from 18-56 years (18-50, 19-56, 19-52 years, respectively). The mean age of our patients was 29.4 years, whereas the women were a little bit older in the other two studies, at 35.1 and 35.7 years (15, 16).

In contrast to de Gruijl et al. and Ochi et al., we used Hybrid-Capture II test for high-risk determination but not a HPV16 type-specific Polymerase Chain Reaction. This means that we are not able to rule out that non-HPV16 infections were also followed-up.

However, assuming a similar HPV-type distribution of the 87 high-risk HPV-positive dysplasias in our study as described by Delere et al. for a similar German collective, we would expect that about 50% (42/85) of our cases would most likely have had HPV16 infection (17). Overall, 22 cases in our study showed an HPV16-VLP antibody response, which would be 52.4% of the expected HPV16-positive cases. This positivity rate is similar to that reported by deGruijl et al. (55.1%), Ho et al. (56.7%), Carter et al. (59.5%) as well Ochi et al. (61.8%) (15-19).

Interestingly Choi et al. reported that 50% of HPV16-L1 capsid protein-positive dysplasias regressed within one year (10).

In addition to that, and even more conclusive due to the use of the known type-restricted HPV16-VLP serological assay, we can be quite sure that HPV16 infection was present to induce the antibodies measured in our study. This means that the different HPV-DNA detection systems used in the different studies most likely do not explain the discrepant clinical outcomes observed.

One additional difference in the studies is the characterisation of the lesions enrolled. Ochi et al. used biopsies confirmed of CIN1 and CIN2 lesions, whereas de Gruijl et al. used only cytological criteria to classify the lesions as mild to moderate dysplasia. Both were unable to differentiate the early dysplastic lesions in productive and non-productive infections. We, for the first time, added HPV-L1 capsid protein detection with the L1 biomarker, Cytoactiv. As reported previously, this allows for identification of lesions with different risk profiles for remission and progression within the heterogenic group of early dysplastic lesions.

Therefore, in contrast to de Grujil et al. and Ochi et al., we are able to link the L1 antibody response to the expression of the L1 capsid protein, as Nicholls et al. also did (13, 15, 16). Twenty-two (26%) patients of the study group but only two (4%) women of the control group showed an HPV16-L1-specific antibody response. This observed 6.5-fold increase of the antibody response for women with high-risk HPV-positive early dysplastic lesion is indicative of an L1-specific activation of the immune response during a productive HPV infection.

One could argue that the antibody response measured by de Gruijl et al. and Ochi et al. has to be due to the expression of the L1 capsid-protein, even the expression was not confirmed (15, 16).

Nevertheless, the detection of the serological antibody response is not indicative of when the antibodies were generated, i.e. during a past infection or the present ongoing infection. Ochi et al. for example reported that 12.5% of HPV16-DNA-negative early dysplastic lesions were positive in the HPV16 neutralisation assay, which could be indicative of HPV16 antibodies produced in the past. However, even this would only be able to explain some, but not all, of the discrepant clinical outcomes (16).

Interestingly in our study, 4 out of 5 L1 antibody-positive women within the group of L1 capsid protein-negative women showed clinical remission during the follow-up period, which might mean that even antibodies produced in the past could be indicative of clinical outcome for early dysplastic lesions.

Most likely, the different findings in the clinical outcome can be ascribed to the different serological assays used in the studies. de Gruijl et al. and Ochi et al. measured L1/L2 VLP antibody responses, whereas we used an epitope-specific, competitive HPV16-L1 VLP assay (15, 16). This means that deGruijl et al. and Ochi et al. measured antibodies against L1 and L2, whereas we in contrast measured only the L1-specific antibody response of a specific epitope.

Nevertheless, it is still not clear to us why the HPV-specific immune response measured by deGruijl et al. and Ochi et al. was not supportive in controlling the infection, but was accompanied by disease progression in almost all cases. Unfortunately they did not discuss this issue (15, 16).

We are convinced that our data is well in accord with Stanley's model (5). Papilloma viruses have developed several mechanisms to escape immune recognition. There is no viremia in natural infections, and free virus particles are shed from the surfaces of squamous epithelia, with poor access to vascular and lymphatic channels and thus to lymph nodes, where immune responses would be initiated. Nevertheless, these free viral particles, consisting of 360 L1 capsid proteins, are the only fully-accessible antigen source in the early stage of virally induced SIL. These particles are released from the apical layers of keratinocytes during the productive phase at the end of the natural viral lifecycle of the HPV infection.

To generate an effective virus-specific immune response the virus particles have to be detected by antigen-presenting cells of squamous epithelium. We imagine that this L1 stimulus might function like an adjuvant, attracting additional immune cells to this area, which could lead to a T-cell-specific immune response against the early HPV proteins, subsequently or in parallel. Microlesions in the epithelial transformation zone could promote this process, also leading to an activation of B-lymphocytes as antibody-producing cells. A block at any stage of the activation cascade, such as no or diminished antigen release, reduced numbers of antigen-presenting cells or Major.

Histocompatibility Complex (MHC) incompatibility or deficiency, may result in an ineffective immune response, manifest as the lack of L1 antibody response, and failure to clear HPV, propagating progression of the intraepithelial lesion. In addition to the established model of a cell-mediated immune response to early proteins, such as E1 and E6, it remains to be elucidated if an L1-specific cell-mediated immune response is also of importance in the clinical remission of early dysplastic lesions. Additional studies are necessary to answer this question.

Taken together, our data emphasize the importance of the immune response in the control of HPV infections and HPV-related pre-malignant lesions. HPV-specific immune competence as evidenced by the detection of HPV16-L1-specific antibodies strongly correlates with clinical remission of early dysplastic lesions. The risk of progression to CIN3 for L1 antigen and HPV16-L1 antibody double-positive women is extremely low (about 6%) and very rarely (in 8.7%) associated with CIN during follow-up. Therefore, the rapid HPV16-L1 antibody test could be a promising tool to improve the risk assessment and appropriate clinical management of high-risk HPV-positive early dysplastic lesions when combined with the Cytoactiv® test.