There are several lines of evidence indicating that there is a causal relationship between infections by certain (high-risk) types of human papillomaviruses (e.g., HPV 16, 18, 31, 33, or 45) and the development of malignant tumors of the uterine cervix. For example, in more than 95% of tumor biopsies, HPV genomes were detected and transforming activity of the viral E6 and E7 genes was demonstrated in cells in culture and in experimental animals. Therefore, it is generally believed that prevention of HPV infection will reduce the incidence of cervical cancer (and other tumors associated with high-risk HPV infection, such as cancer of the penis and cancer of the vulva). As cancer usually arises many years after the primary infection, it will likely take a long time for the decrease in tumor incidence to become apparent. However, as one can expect a reduction of the short-term clinical consequences of infection (cervical intraepithelial neoplasias), one can also expect that a successful vaccination will result in the modification of the screening program and thus in the reduction of health care costs within a few years.

Epidemiological studies strongly suggest that papillomavirus infections are controlled by the immune system and as a result, therefore, vaccination against viral infection should be successful. Papillomavirus infections are more frequent in immunocompromised individuals such as organ-transplant recipients or HIV-infected patients. Studies on the age-dependent prevalence of HPV infections revealed a transient appearance of HPV DNA and of the morphological signs of the infection (i.e., abnormal Pap smear), whereas virus capsid-specific antibodies persist for several years. In regressing warts, enhanced cytokine activity and infiltration of lymphocytes are found, whereas persisting warts do not show signs of an immune response when analyzed using immunohistochemistry.

Direct proof that immunization protects against papillomavirus infection was obtained by experiments in cattle, dogs, or rabbits. Animals were successfully vaccinated against challenge with the respective papillomaviruses, i.e., bovine papillomavirus (BPV) 1 or 4, canine oral papillomavirus (COPV), or cottontail rabbit papillomavirus (CRPV). Protection is based on the induction of neutralizing antibodies following vaccination with complete inactivated virions or with the virus-like particles (VLP). These VLPs assemble after expression of the structural proteins L1 and L2 (or L1 alone) in eukaryotic systems such as yeast or insect cells infected by recombinant baculovirus. A suitable model for genital papillomavirus infections in man (e.g., HPV 16) is the oral papillomatosis in beagles induced by the COPV. The neutralizing serum antibodies (IgG) raised by the vaccination are likely to access the infectious virus through the small wounds that are induced during the procedure of experimental challenge. Although serum IgG antibodies may to a certain degree be involved in providing protection during naturally occurring mucosal infections, and also by leaking into mucosal secretions, local immunity is principally conferred by secretory IgA (sIgA) antibodies. Such a response can be induced either by direct application of the immunogen to the mucosal epithelia (e.g., within the nose or the gut) or via transgenic organisms (bacteria or plants).

Experiments with animal papillomaviruses (BPV 4, CRPV) demonstrate that immunization with viral proteins prevents not only the virus infection (as mentioned previously) but also the progression of already existing lesions. For example, a recent publication by Jensen et al. describes the regression of papillomas after vaccination of rabbits with CRPV E1 recombinant Listeria monocytogenes as a consequence of the induction of an E1-specific T cell proliferative response.

In the case of high-risk human papillomaviruses HPV 16 and 18, it was shown that malignant progression of infected epithelial cells is associated with the expression of viral proteins E6 and E7 that are required for the maintenance of the transformed state. Therefore, these proteins are considered tumor markers, and it has been hypothesized that they represent ideal target molecules for immune therapy strategies. In fact, protection from tumor growth by HPV-transformed syngeneic tumor cells was obtained in laboratory rodents that were immunized with either E6 or E7 proteins, with peptides representing T cell epitopes or with specific cytoxic T cells (CTLs) in adoptive transfer experiments.

The first clinical trials are presently being conducted in late-stage cervical cancer patients. Either recombinant vaccinia virus expressing the E6 and E7 protein of HPV 16 and 18 or synthetic peptides restricted for a certain HLA haplotype (A2) are used for immunization. Initial results suggest that an E7-specific CTL response can be induced in some patients; the issue of a clinical response after vaccination will be addressed in subsequent studies. It is unclear, however, at present whether HPV-specific immune therapy is efficient only in patients with precursor lesions of cervical cancer (and not in cancer patients themselves): in malignant cells, there is often a partial loss of the MHC class I molecules that are required for presentation of virus-specific peptides.

In an attempt to develop a vaccine that combines prophylactic and therapeutic properties, we have generated chimeric virus-like particles (CVLP) consisting of a C-terminally truncated L1 protein fused to sequences of the HPV 16 E7 oncoprotein. We were able to demonstrate that CVLPs induce neutralizing antibodies and CTLs, and that they also confer protection against HPV 16 E7 positive tumors. Thus, we speculate that they are suitable not only to prevent infection by HPV 16 but also to cure infected diseased individuals.

We have found that on expression by recombinant baculovirus, a 34-amino acid carboxy-terminal truncated L1 protein of HPV 16 (designated HPV 16L1?C) was able to assemble into VLPs at an even higher efficiency than the wild type L1. Based on this observation, we fused several different sequences of HPV 16 of heterologous origin onto HPV 16L1?C and analyzed them after expression in insect cells. Insertion of a maximum of 60 amino acids was shown by electron microscopy to be compatible with the formation of virus capsids (albeit more heterogeneous in size than HPV 16L1?C VLPs). One of these chimeras containing the nucleotide coding for the first 60 amino acids of the HPV 16 E7 protein was characterized in more detail. After banding in CsCl gradients, the yield of HPV 16L1?CE7 CVLPs was estimated by electron microscopy and found to be about five-fold lower when compared to HPV 16L1?C VLPs. The presence of HPV 16 E7 sequences within the VLPs was proven by Western blot analysis of the CsCl purified material and also by immunofluorescence using E7-specific antibodies. Therefore, HPV 16L1?C VLPs containing the first 60 aa of the HPV 16 E7 protein were named HPV 16L1?CE7_1-60 chimeric virus-like particles (CVLPs).

The HPV 16 L1 open reading frame (ORF) was excised from plasmid 16-114/k-L1/L2 using Bgl II and subcloned into pUC19 (New England Biolabs) (Schwalbach, Germany) using the single BamHI site of this vector. By PCR we introduced a 34 C-terminal amino acid deletion and an EcoRV restriction site (followed by a stop codon: TAA) that was used to link the nucleotides coding for the first 60 amino acids of the HPV 16 E7 protein generated by PCR from the complete HPV 16 clone (upper part). Subsequently, the EcoRV site was removed and replaced by two additional L1-specific amino acids (lower part). The resulting construct was cloned into pVL1393 (Invitrogen, Groningen, The Netherlands) to generate recombinant baculoviruses: details of the cloning procedure were described elsewhere. Generation of recombinant baculovirus and purification of virus-like particles was carried out as published earlier.