Expert Reviews in Molecular Medicine http://www.expertreviews.org
Request PDF version (367K Acrobat file) How to cite this article
Corrigendum to Figure 1

Accession number: txt001smc; Accession date: 3 July 1998;
Updated: Yes, this version; Last update: 19 October 1998; Archived previous version: No; Replaced by revised version: No
[19 October 1998: Slight modification to schematic of Fig. 1 (fig004smc) to show four domains of monomeric CD4]

Human cellular immune responses against human papillomaviruses in cervical neoplasia

Abbreviations: HPV, CaCx

Stephen Man

Human papillomaviruses (HPVs) are ubiquitous DNA viruses that infect cutaneous and mucosal epithelia. A subset of HPVs infects the female genital tract, to induce cervical lesions that can progress to malignancy in some women. DNA from HPVs can be found in >94% of cervical carcinomas (CaCx) worldwide; this strong association suggests that it might be possible to develop either prophylaxis or therapies for cervical neoplasia, based on the manipulation of human immune responses against HPVs. This review examines the current research into human immune responses against HPVs in CaCx and the potential impact of this research on human health.

Author contact details

Cervical cancer (CaCx) is the second most common cause of cancer-related deaths in women worldwide, with 500 000 new cases reported annually (Ref. 1). In the UK alone, CaCx accounts for ~1400 deaths per year, with five-year survival rates for patients with advanced cancer being ~45% (Ref. 2). Cervical neoplasia encompasses a spectrum of cervical epithelial cell abnormalities, ranging from pre-invasive cervical intraepithelial neoplasia (CIN) to CaCx (Fig. 1, fig001smc). It is thought that CaCx develops from pre-invasive CIN lesions rather than arising spontaneously. Cervical smear testing and treatment for pre-invasive CIN have reduced the overall incidence of CaCx in the UK and other developed countries (Ref. 3). Nevertheless, women treated for CIN still have a 20-fold higher risk of developing CaCx than the general population (Ref. 4); furthermore, adenocarcinoma of the cervix is increasing in incidence, in the UK at least (Ref. 5). Current therapies for cancer (surgery, radio- or chemotherapy) are of limited effectiveness for curing recurrent disease, and their use in developing countries is limited by economic constraints. The global burden of HPV-associated cervical neoplasia requires the development of novel approaches for prevention and treatment.

Association between papillomaviruses and cervical neoplasia
HPVs comprise a large family of double-stranded DNA (ds DNA) viruses that infect epithelial cells of cutaneous (skin) and mucosal surfaces. To date, at least 75 distinct types have been isolated from various human tissue biopsies. A subset (~25) of these HPVs is sexually transmitted and causes a spectrum of diseases in the male and female genital tract. These HPVs can be categorised according to their potential to cause malignancy as low risk, intermediate risk or high risk. HPV types 6 (HPV-6) and 11 (HPV-11), for example, are associated with low-risk disease, such as genital warts. By contrast, high-risk HPVs, such as HPV-16, HPV-18, HPV-31, HPV-33 and HPV-45, are associated with high-grade intraepithelial lesions and invasive cancer. HPV DNA has been found in >94% of CaCx cases studied, with HPV-16 and HPV-18 being the types most commonly detected. In this review, the emphasis is on the current knowledge of human immune responses against HPVs that are associated with the development of cervical neoplasia, including both premalignant disease (CIN) and invasive CaCx.

The close association of both CIN3 (the most advanced stage of CIN) and CaCx with HPV infection provides an opportunity for immunologists to develop strategies based on harnessing the human immune response to prevent or treat cervical neoplasia. This review focuses on cell-mediated immune (CMI) responses, but it is worth noting that the first evidence that HPV proteins were immunogenic in humans came from the detection of HPV-specific antibody responses. However, natural infection of humans by HPVs does not appear to provoke strong antibody responses; antibody responses against E6 and E7 proteins from HPV-16 are most prevalent in patients with advanced CaCx and appear to have little prognostic value (reviewed in Ref. 6). Prophylaxis against HPV infection using a vaccine will require the induction of a strong neutralising antibody response against HPV structural proteins (see Fig. 2, fig002smc). This will need to take place at the point of infection, namely at the mucosal surfaces of the female genital tract. Recently, the development of synthetic virus-like particles (VLPs) based on HPV structural proteins L1 and L2 has greatly advanced knowledge of immune responses to HPV infection (reviewed in Ref. 6). VLPs have been used as vaccines in animal papillomavirus models with encouraging results; for example, dogs were protected against papillomavirus infection by vaccination with VLPs (Ref. 7). To date, however, VLPs based on HPV L1 and L2 proteins have not been tested on humans and evaluation of clinical effectiveness in preventing the development of CaCx will require studies that are both large scale and long term.

The association between HPV infection and development of CaCx is not absolute, because the majority of women infected with HPVs do not develop CaCx. A recent study of college-age women in the USA demonstrated that ~40% of this population had evidence of HPV infection but the majority were able to eliminate HPV infection within 6 months (Ref. 8). The role of the immune system in this viral clearance is not clear, but there is circumstantial evidence that CMI is important in the control of persistent HPV infection. First, HPV-associated cervical lesions are more prevalent in immunosuppressed individuals, such as the recipients of kidney transplants who are undergoing immunosuppressive therapy (Ref. 9) or women who are infected with human immunodeficiency virus (HIV; Ref. 10). Second, in humans, so-called ‘spontaneous’ regression of HPV-induced genital warts is associated with an infiltration of lymphocytes into the lesions (Ref. 11); similar observations have also been made in animal models (Refs 12, 13, 14). Third, the presence of lymphocytes in the lesion correlates with improved prognosis in squamous CaCx (Ref. 15). These observations have fuelled an intense worldwide study of CMI responses against HPVs, to understand the roles of these responses in disease, ultimately with the aim of developing immunotherapies against CaCx.

HPV infection and gene expression
The study of human CMI against HPVs is limited by the technical inability to propagate virion particles on a large enough scale in vitro to generate HPV antigens and HPV-infected target cells. Furthermore, the HPVs that are associated with the development of cervical neoplasia have both species- and tissue-specific tropism, infecting only human keratinocytes (or cells with the potential for squamous maturation). Fortunately, advances in the molecular genetics of HPVs have allowed immunologists to express HPV-gene products in non-HPV infected cells of many species, using DNA transfection or recombinant viral vectors.

The organisation of the HPV-16 genome, consisting of 8000 DNA base pairs (bps), has been extensively studied and is shown in a simplified form in Figure 2 (fig002smc; reviewed in detail in Ref. 16). Two types of genes are expressed: the early (E) genes are involved in DNA replication, protein transcription and cellular transformation; and the late (L) genes code for viral capsid proteins. The target cell for HPV infection is the keratinocyte (epithelial cell) and expression of HPV genes in vivo is exquisitely dependent on the differentiation programme of keratinocytes. During a productive viral infection there is a cascade of HPV-gene expression that follows the differentiation programme of keratinocytes; it starts with the suprabasal epithelial stem cells and culminates in release of HPV virions from terminally differentiated keratinocytes that are being exfoliated at the skin or mucosal surface (Fig. 1, fig001smc). The genome of HPVs is maintained extrachromosomally as an episome (a self-replicating genetic element that can move into or out of chromosomes) in keratinocytes, but is frequently found to be integrated into a human chromosome in HPV-associated tumours. The molecular and cellular events leading to the change from a benign, persistently infected basal epithelial cell to a malignant transformed keratinocyte that is no longer capable of supporting HPV replication are still poorly understood. It is important to stress that this transformation process, which can take two decades, is dependent not only on HPV infection (most of the individuals who are infected with HPV do not develop cancer), but also on host genes and other co-factors. However, cellular immunologists have (unashamedly) focused their attention on the HPV-gene products that are expressed in transformed cervical keratinocytes. The HPV proteins that are invariably expressed in HPV-associated tumours are E6 and E7. These bind to the tumour-suppressor proteins, Rb and p53, respectively, to induce the transformed cell phenotype (Refs 17, 18). The absolute requirement for the continued expression of E6 and E7 to maintain the transformed phenotype makes these proteins extremely attractive tumour-specific targets for the immunotherapy of cervical neoplasia.

Processing of HPV antigens for recognition by immune cells
Specific CMI responses have been defined in humans into two broad categories: those mediated by cytotoxic T lymphocytes (CTLs) and those mediated by T-helper lymphocytes (THLs).

Cytotoxic T-lymphocyte responses
CTLs are usually CD8⁺ (express the CD8 receptor on their cell surface) and play a vital role in the clearance of virally infected cells. Foreign viral antigens are recognised on the surface of virally infected cells in the form of short peptides bound to major histocompatability complex (MHC) class I molecules [in humans these are called human leucocyte antigens (HLAs)]. These foreign peptides are 8–10 amino acids long and are generated by the proteolysis of viral proteins in the cytoplasm of an infected cell. These peptides are then transported from the cytosol into the endoplasmic reticulum where they bind to newly synthesised MHC class I molecules. A complex of MHC class I and antigenic peptide plus b₂ microglobulin (b₂m) then migrates to the cell surface where recognition by a CTL bearing a T-cell receptor (TCR) of the appropriate specificity can take place (Fig. 3, fig003smc). MHC class I molecules are expressed on the surface of all nucleated cells, therefore anti-viral CTLs are a powerful effector mechanism for the clearance of virally infected cells. Activation of CTLs requires recognition of foreign antigens on so-called ‘professional’ antigen-presenting cells (APCs), such as macrophages, activated B lymphocytes (B cells) and dendritic cells. All of these cells (1) express high levels of both MHC class I and MHC class II molecules (see below), (2) are able to process exogenous antigens, and (3) are rich in the co-stimulation molecules (such as B7-1) that are important for T-cell activation. For CTL recognition of HPVs, it is envisaged that expression of the early HPV genes (E1–E7) within the cytoplasm of infected cells will generate HPV-derived peptides that are able to bind to HLA class I molecules (Fig. 3, fig003smc). In HPV-transformed cervical keratinocytes, it is assumed that the constitutive expression of E6 and E7 proteins should result in E6- or E7-derived peptides presented by HLA class I molecules at the keratinocyte surface. It is not clear, however, whether HPV-infected or HPV-transformed keratinocytes are able to activate HPV-specific CTLs directly, because keratinocytes have been shown to be poor APCs for CD4⁺ T cells (Ref. 19).

T-helper lymphocyte responses
THLs are usually CD4⁺ and can have both ‘positive’ and ‘negative’ roles in immunity. They can provide cytokines to activate B cells for the production of antibodies, and cytokines to activate and sustain CD8⁺ CTL responses. They can also promote inflammatory responses, which in the context of a response to pathogens can be beneficial but their role is detrimental in other circumstances such as in asthma and autoimmune diseases. In contrast to CD8⁺ T cells, CD4⁺ T cells recognise so-called ‘exogenous’ antigens, which are sampled from the surroundings of the APCs by endocytosis or by phagosomes. Once inside the cell, the antigens are encapsulated into vacuoles called endosomes, which degrade the proteins into peptide fragments (generally 11–13 amino acids in length). These low-pH endosomes fuse with other vesicles that contain MHC class II molecules, allowing binding of the peptide to MHC class II molecules (Fig. 4, fig004smc). Unlike MHC class I molecules, which are expressed on all nucleated cells, MHC class II expression is largely restricted to professional APCs such as macrophages, activated B cells and dendritic cells. All of these cells have high levels of MHC class II expression, are able to process exogenous antigens efficiently, and are rich in the co-stimulation molecules (such as B7-1) that are important for T-cell activation.

The majority of peptide epitopes that have been defined for CD4⁺ T cells have been derived from membrane-bound or exogenous proteins, presumably because these peptides have access to the MHC class II pathway (Ref. 20). Processing and presentation of HPV antigens via HLA class II molecules could occur in a number of ways. There could theoretically be ‘direct’ presentation of HPV antigens (i.e. peptides) because most cervical carcinomas express HLA class II molecules (Ref. 21). This is unlikely, however, because HPV early gene proteins are mostly expressed in the cytoplasm and nucleus, suggesting that these proteins would be preferentially degraded in the cytoplasm and presented by HLA class I molecules. There is also the possibility that HPV proteins might be released into the extracellular space as a result of keratinocyte death (or cell and/or tissue damage arising through trauma or surgery); these antigens could then be taken up by resident APCs and presented to local CD4⁺ T cells. This is possible in cervical carcinoma and high-grade CIN, where infiltrates of lymphocytes (including potential APCs) have been observed (Refs 15, 22, 23). However, low-grade CIN (and HPV infection in general) are characterised by a lack of local inflammation and an absence of cell death (Ref. 24). These observations suggest that the MHC class II pathway would be less effective at presenting antigens from keratinocytes that had been freshly infected with HPV than from keratinocytes that had been transformed by HPV (Fig. 1, fig001smc).

Individual viral antigens can be exogenously or endogenously processed in human APCs, and this can predict whether they are presented by HLA class I or HLA class II molecules. However, there are some notable exceptions; for example, CD4⁺ T cells can recognise endogenous viral antigens via HLA class II molecules (Refs 25, 26). In a reciprocal manner, CD8⁺ T cells have been shown to recognise exogenous antigens presented by HLA class I molecules (Refs 27, 28). Thus, it is conceivable that HPV antigens could be presented by these so-called ‘alternative’ pathways. For example, (1) exogenous HPV antigens could be presented exogenously to HPV-specific CD8⁺ CTLs by local APCs, or (2) endogenous HPV antigens could be presented by keratinocytes to CD4⁺ T cells. Further studies of the antigen-processing function of both HPV-infected and HPV-transformed keratinocytes are required to understand which HPV antigens might be recognised by T cells in vivo.

Methods for measuring human cell-mediated immunity
In common with human immune responses against the majority of viruses (with the exception of HIV), CMI responses to HPVs cannot be demonstrated directly from peripheral blood; activation of lymphocytes in vitro is required. The main methods for activating CMI are summarised in Table 1 (tab001smc). In general, CD8⁺ T-cell responses are studied by the measurement of cytotoxicity against 51Cr-labelled target cells, whereas CD4⁺ T-cell responses rely on the measurement of specific T-cell proliferation and/or cytokine release from T cells. The advantages and disadvantages of the different methodologies are discussed below.

CD4⁺ T-cell-mediated responses: proliferative responses
The most commonly used method to assess specific CMI in vitro has been to test the ability of peripheral blood mononuclear cells (PBMCs) derived from patients to proliferate in response to HPV antigens. It has been assumed that these assays largely measure CD4⁺ T-cell responses; however, CD8⁺ T cells or natural killer (NK) cells can also contribute to the response. Early studies using relatively crude antigen preparations from warts that contained HPVs demonstrated that PBMCs from patients with a history of skin warts were able to proliferate in vitro (Ref. 29). Interestingly, PBMCs from some individuals who had no clinical history of warts also responded specifically to wart-antigen preparations. This high frequency of responders in the control group (of asymptomatic individuals) has been seen in many studies, and some of the possible reasons for this are discussed below.

As the molecular knowledge of HPVs and the mechanisms of antigen processing have advanced, it has become possible to measure T-cell responses against individual HPV proteins, and even to map T-cell epitopes within these proteins (for example, using synthetic peptides). However, interpretation of these assays has certain provisos; One problem associated with the use of soluble proteins to stimulate T cells in vitro is the provision of the most appropriate controls. Positive controls are needed to check that the PBMCs that are isolated are functional and can react to other antigens, whereas negative controls are needed to show that the in vitro conditions that are being used are supporting only HPV-specific T-cell proliferation. It is necessary to be wary if recombinant proteins have been produced in bacterial vectors because it is possible that insufficient purification will result in the carry over of bacterial products (such as endotoxins), which might affect immune responses in vitro. A further consideration is the use of recombinant proteins that have been ‘tagged’ with glutathionine S-transferase (GST) to facilitate their purification on chromatography columns. GST itself can be immunogenic and responses against GST can mask weak, HPV-specific T-cell responses. Most of the studies that have used soluble HPV proteins to assay for immune reactivity have not taken into account (and added controls) the possibility of a response against the protein tag itself; for example, by using GST tags on non-HPV-derived proteins as a control (Table 2, tab002smc).

One of the first studies to address whether HPV-specific memory CD4⁺ T-cell responses existed in patients with CIN used soluble E6 proteins from HPV-16 and HPV-18. Weak proliferative responses were observed both in CIN patients and in normal controls, but there was no statistically significant difference between the two groups (Ref. 30). The most clear-cut results so far have been obtained using panels of synthetic peptides that overlap in sequence to stimulate T cells in vitro. Use of E7 peptides from HPV-16 (Ref. 31) and E6 and L1 peptides from HPV-16 (Ref. 32) defined eight different epitopes that were recognised by CD4⁺ T cells from normal donors. However, the prior exposure status of the donors to HPVs was not known, so it was not clear whether these T-cell proliferative responses represented ‘fortuitous’ primary responses or indicated the presence of memory T cells that were specific for HPVs. Later studies addressed the clinical relevance of CD4⁺ T cells that are specific for HPV-derived peptides, by comparing responses between patients with well-characterised HPV-associated disease and normal controls. Responses against peptides derived from HPV-16 E7 (Ref. 33) and HPV-16 L1 (Refs 34, 35) were not significantly associated with HPV disease because T-cell responses were also seen in normal controls. However, decreased numbers of patients with CaCx were able to make T-cell responses against HPV-16 E7 and HPV-16 L1 peptides, compared with either patients with CIN3 or normal controls. However, other studies using HPV-16 E7 peptides have shown more clear-cut clinical correlations. In these studies, the T-cell responses correlated with both the stage of disease and the presence of HPV-16 DNA (Refs 36, 37, 38). A limitation of all of these studies, however, is that there is no indication that the peptide epitopes used are actually produced as the result of processing of the soluble proteins by APCs, or that they are representative of those found on either an HPV-infected keratinocyte or a transformed keratinocyte.

T-cell responses from control groups
A common feature of CD4⁺ T-cell responses against HPVs is the high frequency of proliferative T-cell responses from normal (apparently asymptomatic) control groups of volunteers or patients. There are two possible reasons for this. First, the T-cell responses to HPV antigens in the patients might not bear any relationship to HPV-associated disease (and, as such, are an artefact of the in vitro system used); owing to the lack of appropriate controls used, this cannot be excluded in many of the studies reported so far (summarised in Tables 2 and 3, tab002smc, tab003smc). Second, some researchers might argue that HPVs are ubiquitous viruses in humans and that most individuals will be able to make a T-cell memory response as a result of their prior exposure to HPV. Thus, any relevance to disease must be inferred either from a statistically significant increase in the number of people who respond (qualitatively) in a patient group (Ref. 36), or conversely from a significant decrease in the number of patients who respond (Ref. 33). This argument suggests that the T-cell proliferative responses that are often reported in asymptomatic individuals must result from long-lasting T-cell memory against HPV infection; however, there is little evidence to support this. Studying human volunteers after a natural primary infection with HPVs could address this important question: to establish both the longevity of HPV-specific immunity (measuring antibodies, CD4⁺ and CD8⁺ T-cell responses) and its role in the development of cervical neoplasia. However, setting up such studies will be ethically and technically difficult and would require a long-term follow-up period.

Cytokine production
CMI is regulated by cytokines that are produced by CD4⁺ T-helper cells. The major cytokines associated with the development of antiviral CMI are interleukin 2 (IL-2) and interferon g (IFN-g), whereas cytokines such as interleukin 4 (IL-4), interleukin 5 (IL-5), interleukin 10 (IL-10) and interleukin 13 (IL-13) are thought to inhibit the development of CMI. However, few studies have addressed the production of specific cytokines by HPV-specific CD4⁺ T cells. Clerici and colleagues (Ref. 39) measured cytokine release by PBMCs after stimulation with soluble influenza antigens, allogeneic APCs or mitogens. They studied patients who had either extensive HPV infection (HPV DNA was found at multiple sites of the lower genital tract) or localised HPV infection (HPV DNA was found only in the cervix). There was a significant difference between the levels of cytokines produced by patients who had extensive HPV disease and by those who had localised HPV disease. Decreased levels of IL-2 production were found in patients who had extensive HPV disease, compared with those who had localised HPV disease or those in control groups. By contrast, elevated levels of IL-4 and IL-10 were found in the patient group that had extensive disease, compared with those with only localised HPV disease or no HPV disease (controls). This study, which was performed on a small group (30) of patients using non-HPV antigens, suggests that there is a shift towards increased T-cell production of cytokines (IL-10, IL-4) that are associated with the suppression of CMI. A more recent study (examining IL-2 production by patient-derived T cells in response to HPV-16 E7-derived peptides) suggests that the generation of T-cell responses is dependent on additional factors (Ref. 40). No significant responses were observed from T cells derived from: (1) umbilical cord blood, (2) patients with evidence of HPV-16 infection but without CIN, or (3) patients with CIN but no evidence of HPV infection. Interestingly, significant responses were observed in patients with CIN who had persistent HPV infection, and also to a significantly lesser extent in patients with CaCx. On follow-up, patients who cleared HPV-16 [measured using the polymerase chain reaction (PCR)] were still able to make a positive IL-2 response; however, this decreased with time after clearance of HPV infection. This might suggest a role for THLs in the clearance of HPVs, but this interpretation must be reconciled with the increased T-helper activity found in patients with persistent HPV infections. This study illustrates the difficulties in determining whether HPV-specific T cells play any role in disease, and suggests that additional (immune and non-immune) parameters will need to be studied to gain a clearer picture.

Because immune responses are regulated by HLA molecules, it is possible that possession of certain HLA class II alleles might contribute to disease susceptibility. This has been extensively studied in patients with CIN and CaCx; however, few studies have reported significantly increased or decreased frequencies of HLA alleles in patient groups compared with controls (Ref. 41).

CD8⁺ T-cell responses
HPV-specific CTLs can be readily induced in rodent models after injection of the animals with recombinant viruses containing HPV genes (Ref. 42) or rodent cells transfected with HPV genes (Ref. 43). Furthermore, such HPV-specific CTLs can protect against subsequent challenge with tumour cells transfected with HPV genes (Refs 43, 44). This has driven the search in humans for so-called ‘naturally occurring’ CTL responses. Until recently the results have been disappointing, with a consensus that HPVs are able to evade the human immune system and establish a state of immune tolerance (Refs 45, 46). This view did not, however, hinder the development of vaccines designed to elicit HPV-specific immunity. It is noteworthy that early clinical trials of such vaccines were set up in the absence of any data that CMI responses occurred naturally in human HPV infection or, indeed, were relevant to human cervical disease. Three clinical trials involving patients with advanced CaCx have been set up using: (1) HPV-derived peptides (Ref. 47), (2) HPV-derived soluble proteins (Ref. 48), or (3) recombinant vaccinia viruses expressing HPV proteins (Ref. 49). Although the clinical results or immune responses from these trials have not been extensively reported in the literature, it is clear from at least two of the trials that HPV antigens can be immunogenic in humans, inducing a T-cell response (Refs 48, 49). All of the trials were performed on groups of patients with advanced CaCx, making assessment of ‘natural’ T-cell immunity in the absence of the vaccine difficult. More recently, there has been significant progress with the first reports that HPV-specific human CTLs can occur naturally (Table 3, tab003smc). Ressing and co-workers (Ref. 50) demonstrated that peptides derived from HPV-16 E7, which were able to bind to HLA-A*0201 (the commonest HLA class I allele among Caucasians), were capable of inducing peptide-specific CTLs in vitro from normal donors. Some of these CTLs were able to recognise a cell line derived from human cervical carcinoma that had been transformed by HPV-16 and was expressing HLA-A*0201; this suggests that at least some of the CTLs that were stimulated by peptides in vitro also recognised naturally processed HPV epitopes. This approach was extended to the study of patients with CIN3 and CaCx, and showed that the same peptides were recognised by T cells present in PBMCs (Refs 50, 51) and also by T cells isolated from cervical biopsy specimens (Ref. 51). By contrast, another study (Ref. 52), which used the same peptides, was unable to detect CTLs among patients unless populations of APCs that were enriched for dendritic cells were used for T-cell stimulation in vitro. This study raises the possibility that the peptides that are selected for re-stimulation in vitro might not be truly representative of the HPV epitopes that are actually expressed in vivo as a consequence of HPV-associated disease. Another disadvantage of the peptide re-stimulation approach is that for technical reasons the HPV responses examined will be restricted to a single (or only a few) HLA allele(s). Evans and co-workers (Ref. 53) have used HPV-16-transformed CaSki cells (an HPV-transformed cervical keratinocyte line, which expresses HLA-A*0201) for re-stimulation of PBMCs in vitro, rather than peptides. This approach should re-stimulate CTLs with naturally processed HPV epitopes; however, the use of a tumour cell line that has not been derived from the same person as the responder cells were derived from can result in responses preferentially directed against allogeneic MHC molecules, rather than HPV antigens. This might, in part, explain the low numbers of patients (9%) who were able to mount an HPV-specific CTL response in this study. An alternative approach to re-stimulation with HPV-derived peptides (or tumour cells) is to use full-length HPV proteins, either in a soluble form (Ref. 54) or expressed by recombinant viral vectors (Ref. 55). Using the latter approach (recombinant adenoviruses expressing full-length HPV-16 or HPV-18 E6/E7), naturally occurring HPV-specific CTLs were found in the PBMCs of most (6/10) of the patients with CIN3 tested (Ref. 55). Some of the CTLs generated were also able to lyse the HPV-transformed CaSki target cells, and CTLs could be detected at multiple time points over an eight-month period. The CTL responses also appeared to be disease associated, because no responses were detected among normal controls. Further testing of the MHC restriction of the CTL lines from one patient revealed that HPV antigens could be recognised in the context of at least three different HLA class I alleles. These data suggest that the natural CTL response against HPVs encompasses a broad range of MHC- and HPV-antigen specificities. A high CTL response rate (6/8) was also observed using soluble HPV-16 E6 and E7 proteins (Ref. 54); however, it was not clear whether these responses represented CD8⁺ or CD4⁺ CTLs. The CTL responses were most frequent among individuals who were HPV-16⁺ (by PCR typing) with no CIN; controls who were HPV-16^- with no CIN were not tested.

Overall, the studies of CTL function in humans with HPV-associated cervical disease suggest that naturally occurring HPV-specific CTLs do occur, although with variable detection rates (resulting from the differing methodologies employed). Encouragingly (and unlike the studies of CD4⁺ T-cell function in human HPV-associated cervical neoplasia), responses in ‘normal’ asymptomatic subjects appear to be rare, suggesting the CTL responses are associated with HPV disease. However, these preliminary studies have not addressed whether these CTLs have any role in the prevention, control or treatment of disease.

Clinical implications/applications
The in vitro assays described above suggest that HPV proteins are immunogenic in humans both to CD4⁺ T cells and to CD8⁺ T cells. From epidemiological studies, it is clear that most individuals who encounter ‘genital’ HPV types are able to clear them without further complications (Ref. 8). It is not known whether CMI plays any part in this clearance, largely because immune parameters after primary HPV infection have not been studied. Interestingly, it is known that among women with pre-invasive cervical neoplasia (specifically, CIN3) spontaneous regression of lesions occurs in ~30% of cases (Ref. 56). However, we do not know whether HPV-specific CMI plays any role in this regression, or whether the absence of an HPV-specific CMI response predisposes towards the development of cancer. Carefully controlled studies of large cohorts of women who have been exposed to HPV and exhibit stages of cervical disease that are well characterised are needed; furthermore, these studies will require long-term follow-up. It is also possible that CMI is important in both the control of disease and susceptibility towards cancer without being directed against HPV itself; alternatively, other types of CMI (such as that mediated by NK cells or gd T cells, which are a subset of T cells) might play an equally or more important role. Paradoxically, HPV-specific CMI (both CD4⁺ and CD8⁺ T-cell responses) has been most consistently found in patients with persistent CIN3 or cancer (Refs 36, 37, 38, 40, 50, 51, 55). It could be argued that this represents a failure of CMI to halt the progression of disease, and that these CMI responses have arisen as an indirect consequence of disease. Further studies will be required to resolve this; however, these are still likely to be limited by the need to use conventional assays of T-cell function, which require extensive in vitro culture. Recent advances in technology have increased considerably the sensitivity of detection of effector T cells in blood samples, allowing direct quantification and functional assessment with minimal in vitro handling (Ref. 57).

Another consideration of the available data on CMI against HPV is that they are largely sampled from PBMCs; T-cell responses in this cellular compartment might be different from those occurring at sites of disease. Recently, it has been shown that HPV-specific CTLs can be found among lymphocytes that are infiltrating cervical tumours (Ref. 51) and that the numbers of these CTLs are higher than those detected in PBMCs in the same individual. Histological analysis of tumour tissue has demonstrated an enrichment of activated CD8⁺ CTLs, but it is not known whether these cells are HPV specific (Ref. 58).

The measurement of HPV-specific CMI in situ in the cervix is important, because there are several methods by which HPV-transformed cervical keratinocytes could evade immune recognition. First, it has been shown that reduced or absent expression of MHC class I molecules occurs frequently in cervical tumour cells (Refs 59, 60). This is a possible mechanism for HPV-transformed keratinocytes to evade CTL responses; however, it is not known whether MHC class I downregulation is a direct consequence of immune selection (as seen for melanoma; Ref. 61), or simply the result of the genetic instability of tumour cells. What is clear is that downregulation of the MHC pathway does correlate with poorer prognosis for patients with CIN3 (Ref. 62) or CaCx (Ref. 63). Second, HPV-transformed keratinocytes might produce cytokines that suppress immune responses (Ref. 39) or proteins that can inactivate stimulatory cytokines (Ref. 64). Third, HPV-transformed keratinocytes might resist CTL killing, either by producing proteins that interfere with CTL-lytic mechanisms or by inducing apoptotic cell death in the CTLs themselves. For these reasons, it will be important in the future to measure CMI in tissue biopsies, although this will be technically more challenging. Such analyses will be vital to understanding the role of CMI in the natural history of HPV-associated cervical disease and for assessing the therapeutic potential of vaccines designed to induce HPV-specific CTLs.

Human antiviral CTLs have been shown to have powerful therapeutic effects both against infectious virus (Ref. 65) and against virally induced tumours (Refs 66, 67). Regardless of the natural role of HPV-specific CMI against HPVs, there remains considerable optimism that human immune responses against HPVs can also be harnessed for the therapy of CaCx. Continuing studies of the interplay between HPV-transformed keratinocytes and HPV-specific CTLs (both in vitro and at sites of disease) will allow the rational design of immunotherapies. These immunotherapeutic approaches, in combination with conventional treatments and prophylaxis, should aim to diminish the global impact of CaCx.

Research in progress and outstanding research questions
HPV-specific killing of cervical tumour cells by CTLs
For successful immunotherapy of CaCx, HPV-specific CTLs must be able to destroy cervical tumours. The studies described above have used either synthetic peptides or recombinant viral vectors to re-stimulate HPV-specific CTLs in vitro. It is not known whether these methods will mimic adequately the HPV antigens displayed on cervical tumour cells, such that the CTLs generated can kill tumour cells.

Interactions at the sites of disease between HPV-transformed cervical keratinocytes and T cells
It is known that downregulation of MHC class I molecules frequently occurs in CaCx, but it is not known whether this downregulation has a functional impact on CD8⁺ CTL recognition of tumour cells. For example, does downregulation of HLA-A2 in cervical tumour cells effect the recognition of tumour cells by CTLs that are restricted by HLA-A2? If this is so, then CTLs that are not restricted by HLA-A2 should be selected for use in immunotherapy.

Cytokine regulation of the immune response
Activated CTLs can be found among lymphocytes infiltrating cervical tumours. Do cytokines, or other mechanisms, suppress the action of these CTLs? If the interactions between T cells and cervical keratinocytes are understood, it might be possible to circumvent the effects of MHC downregulation (or immunosuppression) in the cervix, for example by local delivery of cytokines.

HPV-specific CTLs and immunotherapy of cervical cancer
Promising results in other systems suggest that HPV-specific CTLs might be effective in the immunotherapy of CaCx. Ongoing vaccine trials (designed to elicit HPV-specific CTLs in vivo) are not yet measuring clinical responses (tumour regression). Results from immunological assays will not be known for 2–3 years. Subject to answering the questions posed above, the most direct method of testing the concept that CD8⁺ CTLs are important in the immunotherapy of CaCx is to infuse patients who have CaCx with large numbers of (autologous) HPV-specific CTLs that have been generated in vitro. As fundamental research continues, it is increasingly likely that this approach will be attempted within the next five years.

Prophylactic vaccination, design and testing
Considerable effort (by academic groups and industry) is now being spent on the development of vaccines against HPVs, some of which are being tested in clinical trials. In brief, two main approaches are being taken: the vaccination of individuals therapeutically to control or eliminate an existing infection (and/or cancer), or the development of prophylactic vaccines to protect an individual from acquiring HPV disease (and HPV-related cancer).

The therapeutic vaccine approaches are based on the fact that E6 and E7 proteins of the high-risk HPV types are continually expressed in the transformed cells; as such, they are potential tumour-specific target antigens for an immune response, particularly CMI responses, such as CTLs. Several strategies are being used to try to induce these responses including: (1) recombinant vaccinia vectors that express E6 and E7, (2) soluble recombinant proteins with adjuvant, or (3) synthetic peptides that represent predicted CTL epitopes (for example, delivered parenterally or coated onto dendritic cells).

Attempts to generate effective prophylactic vaccines to prevent HPV infection are focusing on mechanisms to induce long-lasting mucosal immunity directed against L1 and L2, the major structural proteins of HPVs. The strategies being tried include immunisation with DNA (DNA vaccines) or the highly immunogenic VLPs. In VLPs, L1 is expressed with or without L2 (in eukaryotic or, in some cases, prokaryotic systems); these proteins self-assemble into ‘empty’ virus particles (that do not contain any viral DNA) consisting of proteins in their correct conformation (as found in the virus).

Acknowledgements and funding
Funding for Stephen Man’s research has been provided, in part, by the Royal Society (RS University Research Fellowship) and the Higher Education Funding Council of Wales (HEFCW). The author is grateful to Hetty Bontkes (Department of Pathology, Free University Hospital, Amsterdam, The Netherlands) and Dr Alison Fiander (Department of Obstetrics and Gynaecology, Llandough Hospital, Penarth, Wales) for critically reviewing this manuscript.

References

1 Boyle, P. (1997) Global burden of cancer. Lancet 349 (suppl. II), 23–26 [not in PubMed]

2 Cannistra, S.A. and Niloff, J.M. (1996) Cancer of the uterine cervix. N. Engl. J. Med. 334, 1030–1038 [PubMed]

3 Forsmo, S. et al. (1997) Treatment of pre-invasive conditions during opportunistic screening and its effectiveness on cervical cancer incidence in one Norwegian county. Int. J. Cancer 71, 4–8 [PubMed]

4 Soutter, W.P. et al. (1997) Invasive cervical cancer after conservative therapy for cervical intraepithelial neoplasia [see comments]. Lancet 349, 978–980 [PubMed]

5 Stockton, D., Cooper, P. and Lonsdale, R. (1997) Changing incidence of invasive adenocarcinoma of the uterine cervix in East Anglia. J. Med. Screen. 4, 40–43 [PubMed]

6 Frazer, I.H. (1996) Immunology of papillomavirus infection. Curr. Opin. Immunol. 8, 484–491 [PubMed]

7 Suzich, J.A. et al. (1995) Systemic immunization with papillomavirus L1 protein completely prevents the development of viral mucosal papillomas. Proc. Natl. Acad. Sci. USA 92, 11553–11557 [PubMed]

8 Ho, G.Y. et al. (1998) Natural history of cervicovaginal papillomavirus infection in young women. N. Engl. J. Med. 338, 423–428 [PubMed]

9 Halpert, R. et al. (1986) Human papillomavirus and lower genital neoplasia in renal transplant patients. Obstet. Gynecol. 68, 251–258 [PubMed]

10 Petry, K.U. et al. (1994) Cellular immunodeficiency enhances the progression of human papillomavirus-associated cervical lesions. Int. J. Cancer 57, 836–840 [PubMed]

11 Coleman, N. et al. (1994) Immunological events in regressing genital warts. Am. J. Clin. Pathol. 102, 768–774 [PubMed]

12 Jarrett, W.F. et al. (1991) Studies on vaccination against papillomaviruses: prophylactic and therapeutic vaccination with recombinant structural proteins. Virology 184, 33–42 [PubMed]

13 Knowles, G., O'Neil, B.W. and Campo, M.S. (1996) Phenotypical characterization of lymphocytes infiltrating regressing papillomas. J. Virol. 70, 8451–8458 [PubMed]

14 Selvakumar, R. et al. (1997) Regression of papillomas induced by cottontail rabbit papillomavirus is associated with infiltration of CD8⁺ cells and persistence of viral DNA after regression. J. Virol. 71, 5540–5548 [PubMed]

15 Tosi, P. et al. (1992) Prognostic factors in invasive cervical carcinomas associated with human papillomavirus (HPV). Path. Res. Pract. 188, 866–873 [PubMed]

16 zur Hausen, H. and de Villiers, E.M. (1994) Human papillomaviruses. Annu. Rev. Microbiol. 48, 427–447 [PubMed]

17 Crook, T. et al. (1989) Continued expression of HPV-16 E7 protein is required for maintenance of the transformed phenotype of cells co-transformed by HPV-16 plus EJ-ras. EMBO J. 8, 513–519 [PubMed]

18 Vousden, K.H. (1994) Interactions between papillomavirus proteins and tumor suppressor gene products. Adv. Cancer Res. 64, 1–24 [PubMed]

19 Bal, V. et al. (1990) Antigen presentation by keratinocytes induces tolerance in human T cells. Eur. J. Immunol. 20, 1893–1897 [PubMed]

20 Rammensee, H. (1995) Chemistry of peptides associated with MHC class I and class II molecules. Curr. Opin. Immunol. 7, 85–96 [PubMed]

21 Glew, S.S. et al. (1992) HLA class II antigen expression in human papillomavirus-associated cervical cancer. Cancer Res. 52, 4009–4016 [PubMed]

22 Ferguson, A. et al. (1985) Expression of MHC products and leucocyte differentiation antigens in gynaecological neoplasms: an immunohistological analysis of the tumour cells and infiltrating leucocytes. Br. J. Cancer 52, 551–563 [PubMed]

23 Coleman, N. and Stanley, M.A. (1994) Analysis of HLA-DR expression on keratinocytes in cervical neoplasia. Int. J. Cancer 56, 314–319 [PubMed]

24 Tay, S.K. et al. (1987) Subpopulations of Langerhans cells in cervical neoplasia. Br. J. Obstet. Gynaecol. 94, 10–15 [PubMed]

25 Jaraquemada, D., Marti, M. and Long, E.O. (1990) An endogenous processing pathway in vaccinia virus-infected cells for presentation of cytoplasmic antigens to class II- restricted T cells. J. Exp. Med. 172, 947–954 [PubMed]

26 Nuchtern, J.G., Biddison, W.E. and Klausner, R.D. (1990) Class II MHC molecules can use the endogenous pathway of antigen presentation. Nature 343, 74–76 [PubMed]

27 Carbone, F.R. and Bevan, M.J. (1990) Class I-restricted processing and presentation of exogenous cell-associated antigen in vivo. J. Exp. Med. 171, 377–387 [PubMed]

28 Speidel, K. et al. (1997) Priming of cytotoxic T lymphocytes by five heat-aggregated antigens in vivo: conditions, efficiency, and relation to antibody responses. Eur. J. Immunol. 27, 2391–2399 [PubMed]

29 Lee, A.K. and Eisinger, M. (1976) Cell-mediated immunity (CMI) to human wart virus and wart-associated tissue antigens. Clin. Exp. Immunol. 26, 419–424 [PubMed]

30 Cubie, H.A. et al. (1989) Lymphoproliferative response to fusion proteins of human papillomaviruses in patients with cervical intraepithelial neoplasia. Epidemiol. Infect. 103, 625–632 [PubMed]

31 Altmann, A. et al. (1992) Definition of immunogenic determinants of the human papillomavirus type 16 nucleoprotein E7. Eur. J. Cancer 28, 326–333 [PubMed]

32 Strang, G. et al. (1990) Human T cell responses to human papillomavirus type 16 L1 and E6 synthetic peptides: identification of T cell determinants, HLA-DR restriction and virus type specificity. J. Gen. Virol. 71, 423–431 [PubMed]

33 Luxton, J.C. et al. (1996) Proliferative T cell responses to the human papillomavirus type 16 E7 protein in women with cervical dysplasia and cervical carcinoma and in healthy individuals. J. Gen. Virol. 77, 1585–1593 [PubMed]

34 Shepherd, P.S. et al. (1996) Proliferative T cell responses to human papillomavirus type 16 L1 peptides in patients with cervical dysplasia. J. Gen. Virol. 77, 593–602 [PubMed]

35 Luxton, J.C. et al. (1997) Serological and T helper cell responses to human papillomavirus type 16 L1 in women with cervical dysplasia or cervical carcinoma and in healthy controls. J. Gen. Virol. 78, 917–923 [PubMed]

36 Kadish, A.S. et al. (1994) Cell-mediated immune responses to E7 peptides of human papillomavirus (HPV) type 16 are dependent on the HPV type infecting the cervix whereas serological reactivity is not type-specific. J. Gen. Virol. 75, 2277–2284 [PubMed]

37 Kadish, A.S. et al. (1997) Lymphoproliferative responses to human papillomavirus (HPV) type 16 proteins E6 and E7: outcome of HPV infection and associated neoplasia. J. Natl. Cancer Inst. 89, 1285–1293 [PubMed]

38 de Gruijl, T.D. et al. (1996) T-cell proliferative responses against human papillomavirus type 16 E7 oncoprotein are most prominent in cervical intraepithelial neoplasia patients with a persistent viral infection. J. Gen. Virol. 77, 2183–2191 [PubMed]

39 Clerici, M. et al. (1997) Cytokine production patterns in cervical intraepithelial neoplasia: association with human papillomavirus infection. J. Natl. Cancer Inst. 89, 245–250 [PubMed]

40 de Gruijl, T.D. (1998) Differential T helper cell responses to human papillomavirus type 16 E7 related to viral clearance or persistence in patients with cervical neoplasia: a longitudinal study. Cancer Res. 58, 1700–1706 [PubMed]

41 Stern, P.L. (1996) Immunity to human papillomavirus-associated cervical neoplasia. Adv. Cancer Res. 69, 175–211 [PubMed]

42 Meneguzzi, G. et al. (1991) Immunization against human papillomavirus type 16 tumor cells with recombinant vaccinia viruses expressing E6 and E7. Virology 181, 62–69 [PubMed]

43 Chen, L.P. et al. (1991) Human papillomavirus type 16 nucleoprotein E7 is a tumor rejection antigen. Proc. Natl. Acad. Sci. USA 88, 110–114 [PubMed]

44 Chen, L. et al. (1992) Induction of cytotoxic T lymphocytes specific for a syngeneic tumor expressing the E6 oncoprotein of human papillomavirus type 16. J. Immunol. 148, 2617–2621 [PubMed]

45 Beverley, P.C. et al. (1994) Strategies for studying mouse and human immune responses to human papillomavirus type 16. In Ciba Found. Symp. 187, pp. 78–86, Wiley, Chichester, UK [PubMed]

46 Tindle, R.W. and Frazer, I.H. (1994) Immune response to human papillomaviruses and the prospects for human papillomavirus-specific immunisation. Curr. Top. Microbiol. Immunol. 186, 217–253 [PubMed]

47 Melief, C.J.M. et al. (1996) Peptide based cancer vaccines. Curr. Opin. Immunol. 8, 651–657 [PubMed]

48 Tindle, R.W. (1996) Human papillomavirus vaccines for cervical cancer. Curr. Opin. Immunol. 8, 643–650 [PubMed]

49 Borysiewicz, L.K. et al. (1996) A recombinant vaccinia virus encoding human papillomavirus type 16 and type 18, E6 and E7 proteins as immunotherapy for cervical cancer. Lancet 347, 1523–1527 [PubMed]

50 Ressing, M.E. et al. (1996) 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. 56, 582–588 [PubMed]

51 Evans, E.M. et al. (1997) Infiltration of cervical cancer tissue with human papillomavirus-specific cytotoxic T-lymphocytes. Cancer Res. 57, 2943–2950 [PubMed]

52 Jochmus, I. et al. (1997) Specificity of human cytotoxic T lymphocytes induced by a human papillomavirus type 16 E7 derived peptide. J. Gen. Virol. 78, 1689–1695 [PubMed]

53 Evans, C. et al. (1996) HLA-A2-restricted peripheral blood cytolytic T lymphocyte response to HPV type 16 proteins E6 and E7 from patients with neoplastic cervical lesions. Cancer Immunol. Immunother. 42, 151–160 [PubMed]

54 Nakagawa, M. et al. (1997) Cytotoxic T lymphocyte responses to E6 and E7 proteins of human papillomavirus type 16: relationship to cervical intraepithelial neoplasia. J. Infect. Dis. 175, 927–931 [PubMed]

55 Nimako, M. et al. (1997) Human papillomavirus-specific cytotoxic T lymphocytes in patients with cervical intraepithelial neoplasia grade III. Cancer Res. 57, 4855–4861 [PubMed]

56 Ostor, A.G. (1993) Natural history of cervical intraepithelial neoplasia: a critical review. Int. J. Gynecol. Pathol. 12, 186–192 [PubMed]

57 McMichael, A. and O’Callaghan, C. (1998) A new look at T cells. J. Exp. Med. 187, 1367–1371 [PubMed]

58 Bontkes, H.J. et al. (1997) Assessment of cytotoxic T-lymphocyte phenotype using the specific markers granzyme B and TIA-1 in cervical neoplastic lesions. Br. J. Cancer 76, 1353–1360 [PubMed]

59 Connor, M.E. and Stern, P.L. (1990) Loss of MHC class-I expression in cervical carcinomas. Int. J. Cancer 46, 1029–1034 [PubMed]

60 Cromme, F.V. et al. (1994) Loss of transporter protein, encoded by the TAP-1 gene, is highly correlated with loss of HLA expression in cervical carcinomas. J. Exp. Med. 179, 335–340 [PubMed]

61 Ikeda, H. et al. (1997) Characterization of an antigen that is recognized on a melanoma showing partial HLA loss by CTL expressing an NK inhibitory receptor. Immunity 6, 199–208 [PubMed]

62 Bontkes, H. et al. (1998) Specific HLA class I downregulation is an early event in cervical dysplasia associated with clinical progression. Lancet 351, 187–188 [PubMed]

63 Ellis, J.R. et al. (1995) The association of an HPV16 oncogene variant with HLA-B7 has implications for vaccine design in cervical cancer. Nat. Med. 1, 464–470 [PubMed]

64 Smith, G. (1996) Virus proteins that bind cytokines, chemokines or interferons. Curr. Opin. Immunol. 8, 467–471 [PubMed]

65 Riddell, S.R. et al. (1992) Restoration of viral immunity in immunodeficient humans by the adoptive transfer of T cell clones. Science 257, 238–241 [PubMed]

66 Papadopoulos, E.B. et al. (1994) Infusions of donor leukocytes to treat Epstein Barr virus associated lymphoproliferative disorders after allogeneic bone marrow transplantation. N. Engl. J. Med. 330, 1185–1191 [PubMed]

67 Rooney, C.M. et al. (1995) Use of gene-modified virus-specific T lymphocytes to control Epstein–Barr-virus-related lymphoproliferation. Lancet 345, 9–13 [PubMed]

68 Ressing, M.E. et al. (1995) 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. 154, 5934–5943 [PubMed]

69 Konya, J. et al. (1997) Identification of a cytotoxic T lymphocyte epitope in the human papillomavirus type 16 E2 protein. J. Gen. Virol. 78, 2615–2620 [PubMed]

| home | search | glossary | links | sitemap | contact |

Expert Reviews in Molecular Medicine © Cambridge University Press ISSN 1462-3994 (Disclaimer and copyright)
Editorial Office: Centre for Applied Research in Educational Technologies (CARET), 1st Floor, 16 Mill Lane, Cambridge, CB2 1SB, UK. Tel: +44 (0)1223 765 375; Fax: +44(0)1223 765 505; E-mail: ermm@caret.cam.ac.uk