Mechanisms of cell entry by human papillomaviruses: an overview

Host cell entry of HPV is initiated by binding of the virus particle to cell surface receptors (Figure 1). It has been suggested that virions bind initially to the basement membrane prior to transfer to the basal keratinocyte cell surface [18]. It is important to note that the entry of HPV in vitro is initiated by binding to a cell surface receptor in contrast to the in vivo situation where the basement membrane has recently been identified as the primary site of virus binding [18, 21].

Early work investigating the cell surface receptors found that HPVs bind to a widely expressed and evolutionary conserved cell surface receptor and that the interaction depends primarily on L1 [24‐27]. Glycosaminoglycans (GAGs), especially heparan sulfate, were suggested as initial attachment receptors for HPV VLPs [28‐31]. Heparan sulfate proteoglycans (HSPG) are frequently found in the extracellular matrix (ECM) and on the surface of most cells. They are involved in several biological functions and because of their location they are appropriate molecules for viral infection [32, 33]. Heparan sulfate is often found on two membrane-bound proteoglycans, syndecans and glypicans [34]. Glypicans are predominantly expressed in the central nervous system, whereas syndecans are the predominant HSPG in epithelial cells, the target cells of HPV. Especially syndecan-1 may serve as the primary attachment receptor in vivo due to its high expression level in the appropriate target cells and upregulation during wound healing [27, 35]. Furthermore, other candidate receptors for HPV have been suggested, such as laminin-5. Several in vitro studies have shown that HPV can also bind to a receptor in the ECM, identified as laminin-5 which is able to mediate binding to the ECM [36‐38]. However, laminin-5 interaction seems to be of lesser importance for a productive infection and even though the affinity to laminin-5 is higher than to heparan sulfate, infectious transfer from the ECM seems to require heparan sulfate binding [27, 37, 38].

The classical notion of a virus binding to a single receptor to enter cells through a single defined uptake mechanism is quickly being overtaken by a more complex picture. New findings, such as a specific co-receptor and virus attachment to multiple receptors, have raised the question that viruses known to bind to a non-specific receptor may turn out to also have a more specific co-receptor [39].

Like HPVs, mammalian herpesviruses adsorb strongly to proteoglycans, especially HSPGs. For the herpes simplex virus (HSV) this high affinity attachment step enhances infectivity, although it does not appear to be an absolute requirement for the virus to infect the cell. HSPG is preferred and is considered to be a binding receptor, as opposed to an entry receptor. It is obvious that for cell penetration, HSV usually interacts with co-receptors that are distinct from the proteoglycans attachment receptor [7, 40].

Accumulating evidence suggests that a secondary receptor or co-receptor is also involved in the infectious internalization of HPV subsequent to interaction with HSPG [38, 41]. It appears that HSPG functions as more than a simple attachment factor in HPV infection in that this interaction promotes essential conformational changes in the viral capsid, but HSPGs are clearly not the cell surface receptors that mediate virion internalization or later events in infection [41].

The cell adhesion receptor α6-integrin, which is involved in cell to cell interactions, has been suggested as secondary receptor even though its involvement in HPV infection is rather controversial [29, 35, 37, 42‐44]. Given the close association of proteoglycans and integrins as matrix components, it is possible that the experimental association of α6-integrin with HPV binding and entry is a secondary effect due to its interaction with HSPGs [7].

Several studies suggest a role for L2 in facilitating infection via interaction with a secondary receptor(s) [45‐48]. Although cell surface interactions predominantly depend on the major capsid protein L1, it seems likely that the secondary cell surface receptor is L1-specific, although, it is possible that L2 may contribute to surface interactions [21].

These observations could indicate that the cell surface binding is indeed mediated by more than one receptor. A reasonable hypothesis is that a productive infection would require an initial low specificity binding mediated by L1, followed by the interaction of a more specific protein component with L2 [7]. A specific region in the L2 protein was proposed to interact with a cell surface molecule after attachment of the virus to a primary receptor. This interpretation suggests a post-attachment conformational change at the cell surface to unmask this specific domain in L2, a process that many other viruses use to trigger downstream events such as secondary receptor interactions [27, 48].

Initial attachment to HSPG moieties functions primarily to facilitate the critical step of L2 proteolytic cleavage which is essential for successful infection [41]. The minor capsid protein L2 is cleaved by furin on the cell surface at a consensus cleavage site that is conserved among all papillomaviruses [17]. These sequences are inaccessible at the surface of mature virions in solution in order to prevent host antibody response to the conserved epitopes [27]. As mentioned above, capsid interaction with HSPG results in a conformational change which results in the exposure of the furin cleavage region. After cleavage, an additional conformational change may expose the binding site for the secondary cell receptor, or it lowers the affinity for the primary receptor, which results in the hand-off to a secondary receptor [27, 41, 49].

Taken together, capsid interaction with HSPG induces conformational changes that result in the exposure of the L2 amino terminus. Exposure of this L2 N-terminus allows access to highly conserved consensus furin convertase recognition site and subsequent furin cleavage which is essential for successful infection. Moreover, the L2 N-terminus is essential for the L2 protein to adopt a correct conformation within the assembled capsid. Correct folding may also require the formation of a disulfide bond between HPV16 L2 cysteine residues Cys22 and Cys28, which was recently identified. Mutation of the contributing cysteine residues rendered mutant virions non-infectious [15, 21, 50, 51].

Even if keratinocytes are the main targets of HPV and only entry in these cells has been shown to result in a productive infection, HPV-VLP are also able to enter other cellular types such as dendritic cells (DC) or Langerhans cells (LC). Interactions between these antigen presenting cells (APCs) and HPV are likely to be important for the establishment of the immune response after a prophylactic vaccination or a natural infection. Bousarghin et al. showed that these APCs differentially interact with HPV16 VLPs. Although DC and LC are able to bind and internalize HPV16 VLPs, there are differences in VLP binding to DC and LC. DC use heparan sulfates to bind HPV16 VLPs in contrast to LC on which heparin does not have any inhibitory effects [52]. Various studies showed that VLPs co-localize with langerin in LC [52, 53]. Although still controversial, the investigation on the immunogenicity of VLPs supports a key contribution for the low-affinity Fcγ receptors, expressed on DC, as an important molecule in a HPV-VLP receptor complex [54, 55].

After binding to cell surface receptors HPV must be internalized into the cell to establish an infection. To date, the dynamics of HPV interaction with the cell surface during the initial stages of infection are not completely understood and the entry mechanisms and the molecules involved are contradictory and still a subject of scientific debate (table 1).

The conflicting data could be due to the "maturity" state of the VLPs and PsVs used. HPV capsids extracted from replicating cultured cells can exist in two forms. "Immature" capsids are larger, less regular and less protease resistant than "mature" capsids indicating a substantial change in conformation during the maturation process [56]. Therefore, it is likely that the omission of a maturation step could result in assay variability due to particle heterogeneity [7]. Moreover, HPVs exhibit promiscuous cell association while only completing their life cycle in differentiating squamous epithelium [57]. Therefore, while the early events of infection may be similar in permissive and non-permissive cell types, there is a restriction of viral replicative functions and virion production that is determined by factors tied to the keratinocyte differentiation program [7].

Productive entry of HPV involves internalization by endocytosis, a process that for HPV occurs slowly and asynchronously over a period of several hours, except for some non-epithelial cells [8, 52, 58]. Multiple studies have shown an unusually extended residence on the cell surface for HPVs [7, 29, 59, 60]. Most ligands, including the majority of viruses, are internalized rapidly, within minutes after the initial receptor encounter and engagement. The reason for the delayed kinetics for HPVs is unknown, although it is noteworthy that syndecans have been reported to have a slow rate of internalization after ligand binding [61]. Alternatively, the conformational changes or the transfer to a secondary receptor that is sparsely arrayed or exhibits particular requirements for endocytosis are a possible explanation for the slow kinetics [8, 27, 58]. Moreover, in vitro experiments showed that cell surface dynamics of HPV indicated a transport mechanism along actin rich cell protrusions to access the endocytic machinery and thus enhance infectious entry. This transport was facilitated by binding to receptors that were likely to interact with actin filaments to mediate the transport towards the cell body by retrograde flow. This requirement may contribute to the prolonged residence on the cell surface and the impeded kinetics [8].

Several endocytic pathways have been described and clathrin- and caveolae-mediated are two main pathways used by non-enveloped viruses to infect cells [5, 62]. A possible approach to distinguish between the clathrin-dependent and caveolar pathways is the analysis of biochemical inhibition of ligand uptake, although non-specific effects must be considered. The development of molecular inhibitors in the form of dominant-negative molecules has surpassed the use of biochemical inhibitors in terms of decreasing these non-specific effects. Selinka et al. examined a set of biochemical inhibitors for effects on HPV33 PsV infection and found a dependence upon an intact actin cytoskeleton and microtubules. Day et al. investigated the uptake of bovine papillomaviruses (BPVs) through biochemical inhibitor analysis and co-localization studies with established markers of endocytic compartments. Both studies could not demonstrate the involvement of caveolar endocytosis and concluded that uptake of these viruses occurs via a clathrin-dependent pathway [59, 63]. However, a study utilizing PsVs, generated by mixing VLPs with naked DNA, unexpectedly found that HPV31 was sensitive to caveolar inhibition. In contrast, the entry of HPV16, which phylogenetically, is closely related to HPV31, and HPV58 was found to be blocked by inhibitors of clathrin-mediated uptake [64]. The data on the entry of HPV31 was confirmed by Smith et al. who described a caveolar uptake of HPV31 virions in keratinocytes [65]. However, another study found that biochemical inhibition of clathrin-dependent uptake did prevent HPV31 infection [66]. HPV31 appears to interact with HSPG similarly to HPV16 for in vivo infection. Possibly HPV31 interacts differently with or has a unique co-receptor that shunts it into a different internalization pathway [67].

Most studies investigating the uptake of HPV16 concluded the involvement of clathrin-dependent endocytosis [63‐66]. In contrast to these studies, Spoden et al. observed clathrin- and caveolae-independent internalization of HPV16 PsVs. Entry occurred by a mechanism that was resistant to combined siRNA-mediated down regulation of caveolin-1 and clathrin heavy chain as well as being resistant to over-expression of dominant negative mutants of caveolin-1 and eps-15, which plays a role in clathrin coated vesicle formation [58]. The authors suggested the involvement of tetraspanin-enriched microdomains that serve as a platform for uptake by an uncharacterized internalization mechanism. None of the conducted studies demonstrated an effect of caveolar disruption on HPV16 infection.

Initiation and progression of HPV-associated cervical cancer have been shown to be related to functional alterations of LC within the cervical epithelium. Because of their role in initiating an antiviral immune response, DC and LC represent an ideal target for immune evasion by viruses. The study of the interactions between HPV16 VLPs and DC or LC showed that the entry of virus particles is different as suggested by Fausch et al. and Yan et al. Fausch et al. showed that DC use a clathrin-mediated endocytosis whereas LC use a different pathway which is not associated with clathrin or caveolae [68]. Yan et al. show that LC uptake of HPV6 L1 was blocked by filipin pretreatment confirming a role for caveolin-mediated uptake of VLPs by LC [53]. Another study, however, showed that virus particles use the same clathrin-dependent endocytic pathway to enter DC and LC [52].