Abstract
Human papillomavirus (HPV) infection is clinically very common. It is usually a major risk factor in the development of cutaneous benign lesions, cervical cancer and a variety of other malignancies. The biological function of ubiquitination as an intracellular proteasomal-mediated form of protein degradation and an important modulator in the regulation of many fundamental cellular processes has been increasingly recognized over the last decade. HPV proteins have been demonstrated to evolve different strategies to utilize the ubiquitin system for their own purposes. The putative roles of E3 ubiquitin ligases in HPV-induced carcinogenesis have become increasingly apparent, although the mechanisms remain unclear. In this review we provide an update on the mechanisms of the involvement of E3 ubiquitin ligases in HPV-induced carcinogenesis, focusing on their interaction with HPV proteins and their roles in several signalling pathways. Targeting the E3 ubiquitin ligases might offer potential therapeutic strategies for HPV-related diseases in future.
Introduction
Human papillomaviruses (HPV) are small double-stranded DNA viruses. More than 130 HPV genotypes have been cloned from a wide variety of clinical lesions. 1 HPV have a predilection for either cutaneous or mucosal epithelial surfaces and can be divided into low- and high-risk HPV. Low-risk HPV predominantly generate benign squamous epithelial lesions, commonly known as warts, whereas the high-risk types are associated with malignant diseases such as cervical cancer, skin tumours in patients with epidermodysplasia verruciformis, anal and rectal cancers, and head and neck squamous cell carcinoma (HNSCC), including oral cavity cancer.2–4 The two most common low-risk viruses causing warts on the anogenital epithelium are HPV6 and HPV11; the two major high-risk or oncogenic viruses are HPV16 and HPV18. Although the role of HPV in HNSCC is highly controversial, recent data from meta-analyses and case–control studies indicate that HPV may be an independent risk factor for the development of oropharyngeal and oral carcinoma.5–7
The HPV genome can be divided into three regions: an early region encoding six nonstructural proteins (E1, E2, E4, E5, E6 and E7); a late region encoding two structural proteins (L1 and L2); a noncoding long control region. The E1, E2, E4 and E5 proteins are required for viral DNA replication, the E6 and E7 oncoproteins co-operate to transform and immortalize infected cells, and the L1 and L2 proteins are needed for the production of viral particles. 8 Because they infect the basal layer of the squamous epithelium and have no viraemic phase, HPV are isolated from circulating immune cells. The limited innate immune response, the low levels of viral gene expression in the basal epithelium and the lack of cytopathic effects generally lead to a delayed adaptive immune response to initial HPV infection, thus favouring the establishment of persistent viral infection. 3
Because of the rather limited coding capacity of viral genomes, many aspects of the life cycle of HPV rely on specific interactions between viral proteins and intracellular regulatory proteins, so as to redirect cellular signalling to the benefit of the virus. Modification of cellular proteins by the reversible covalent attachment of ubiquitin, much like phosphorylation, is involved in the control of most fundamental cellular processes. As one of the most pivotal regulatory mechanisms in biology, the ubiquitin system is extensively used to promote HPV replication, persistent viral infection and HPV-induced carcinogenesis. Good examples are the interactions of HPV E6 and E7 oncoproteins with tumour suppressors p53 and pRb, respectively, leading to their degradation and inactivation, and contributing to persistent viral infection, cellular transformation, immortalization and carcinogenesis. 9 In this review we will provide an update on the contributions of E3 ubiquitin ligases to HPV-induced carcinogenesis and the associated mechanisms, focusing on their interaction with HPV proteins and their roles in several signalling pathways.
The ubiquitin system
As a highly conserved 76-amino-acid polypeptide, ubiquitin serves as a signal for the target protein to be recognized and degraded in the 26 S proteasome. The ubiquitin–proteasome pathway plays a critical role in various cellular functions including antigen processing, cell-cycle regulation, apoptosis, signal transduction and transcriptional regulation. 10 Ubiquitin is covalently attached to one or more lysine residues of cellular proteins via an enzymatic cascade involving three classes of enzyme, termed E1 (ubiquitin-activating enzymes), E2 (ubiquitin-conjugating enzymes) and E3 (ubiquitin–protein ligases). The process of ubiquitination and the ubiquitin–proteasome pathway have been discussed in detail.10,11 E3 ubiquitin ligases can be classified into two groups according to their substrates. One group is characterized by homology to the E6-associated protein carboxyl terminus (HECT) domain. 12 The archetype is the E6-associated protein (E6-AP) recruited by HPV E6 protein. 13 The second group, the so-called really interesting new gene (RING) finger E3 ubiquitin ligases, can be categorized further into unimolecular and multimolecular E3s. Examples of unimolecular RING E3s are murine double minute 2 (Mdm2) 14 and Casitas B-lineage lymphoma (Cbl). 15 Prototypes of the multimolecular RING E3s are the Skp1/Cullin1/F-box (SCF) protein family. 16
It is now clear that polyubiquitin chains linked through lysine at position 48 of ubiquitin (Lys48) target substrate proteins for proteasome-mediated degradation, whereas polyubiquitin chains of alternative linkages (such as Lys63) carry out signalling functions independently of proteolysis. 17 Mono-ubiquitination of some proteins (such as histones, calmodulin and actin) serves as a signal for endocytosis and histone-mediated transcriptional regulation without targeting for degradation. 18 The removal of ubiquitin residues from substrates is catalysed by deubiquitinating enzymes. The deubiquitinating enzymes are of great importance in numerous critical cellular functions such as cell growth, gene silencing, oncogenesis, endocytosis and DNA replication. 19
HPV oncoproteins that assemble E3 ubiquitin ligases
Examples of interactions between human papilloma virus (HPV) proteins and E3 ubiquitin ligases.
See text for details.
APC, anaphase-promoting complex; Bak, Bcl-2 antagonist killer; c-Cbl, cellular Casitas B-lineage lymphoma; cIAP, cellular inhibitor of apoptosis: c-myc, cellular myelocytomatosis oncogene; E6-AP, E6-associated protein; E6TP1, E6 targeting protein 1; EGFR, epidermal growth factor receptor; hADA3, human alteration/deficiency in activation protein 3; hDlg, human homologue of Drosophila discs large tumour suppressor protein; hScrib, human homologue of the Drosophila Scribble tumour suppressor protein; IGFBP3, insulin-like growth factor binding protein 3; MCM7, minichromosome maintenance protein 7: Mdm2 murine double minute 2; MGMT, O6-methylguanine DNA methyl transferase; NFX1, nuclear transcription factor, X-box binding 1; PTPN3, protein tyrosine phosphatase, nonreceptor tyrosinase 3; RIP1, receptor-interacting protein 1; SCF, Skp1/Cullin1/F-box; SOCS1, suppressor of cytokine signalling 1; TIP-2/GIPC, Tax interacting protein, clone 2/GAIP interacting protein, C terminus; TRAF, tumour necrosis factor receptor-associated factor; EDD, E3 identified by differential display.
Annexin A1 and human alteration/deficiency in activation protein 3 (hADA3) have been shown to be novel substrates for E6-AP-mediated degradation (Table 1). Annexin A1 is thought to be a novel substrate for E6-AP-mediated ubiquitination and subsequent degradation, 32 indicating that E6-AP plays a role in controlling the diverse functions of annexin A1, such as inhibition of cell proliferation, regulation of cell differentiation, apoptosis and processes involved in carcinogenesis. These results may explain research showing that, in comparison with normal cervical epithelium, the expression of annexin A1 is significantly reduced during the progression of cervical neoplasia. 33 hADA3 appears to be required for p53-mediated transcriptional activation of various target promoters and p53-mediated G1 cell-cycle arrest. 34 A study reported significantly increased hADA3 expression, decreased cell proliferation and increased apoptotic rate in HPV-16-positive SiHa cells treated with either E6 siRNA or E6-AP siRNA. 35 These data suggest that HPV16 E6 might initiate cellular transformation as a consequence of hADA3 inhibition, induced by E6-AP-dependent ubiquitination and proteasomal degradation.
The ability of high-risk HPV to target cellular proteins for proteasome-mediated degradation is not restricted to E6. It has been reported that HPV E7 proteins bind pRb through the conserved LXCXE motif, resulting in disruption of the pRb/E2F repressor complex and uncontrolled G1 exit and S-phase entry.36,37 High-risk HPV E7 proteins destabilize pRb via a ubiquitin/proteasome-dependent mechanism.38,39 Huh et al. 40 proposed that the HPV16 E7-associated cullin 2/ubiquitin ligase complex contributes to aberrant degradation of pRb in HPV16 E7-expressing cells (Table 1). Low-risk HPV E7 proteins bind pRb with lower affinity than the high-risk HPV E7 proteins. 37 Recent reports show that expression of the low-risk HPV6 E7 protein causes destabilization of the pRb family member p130 by ubiquitination of phospho-p130 through the cullin 1/ubiquitin ligase complex.41,42
HPV proteins that interfere with E3 ubiquitin ligases
The third oncoprotein in HPV, HPV E5, has been shown to induce transformation and anchorage-independent growth, 43 to co-operate with E7 to stimulate proliferation of primary cells in vivo 44 and to enhance immortalization by E6/E7. 45 HPV E5 delays degradation of activated epidermal growth factor receptor (EGFR) and upregulates EGFR-mediated signal transduction. This may be a consequence of interaction of E5 with the EGFR, which disrupts binding of EGFR to cellular Cbl (c-Cbl) (Table 1), 46 an E3 ubiquitin ligase that ubiquitinates phosphorylated tyrosine kinase receptors. 15 Expression of the HPV E2 protein leads to G2/M arrest. 47 Retardation of the G2/M transition is believed to be the result of interference with the activity of anaphase-promoting complex (APC), a large multiprotein E3 ubiquitin ligase that is crucial for progression into and out of M-phase. Furthermore, high-risk, but not low-risk, HPV E2 proteins induce genomic instability as a consequence of perturbation of APC-dependent proteolysis 48 (Table 1).
HPV proteins degraded by E3 ubiquitin ligases
Human papillomavirus E7 is a short-lived protein ubiquitinated on the N-terminal residue and degraded by the ubiquitin–proteasome pathway. 49 Two independent pathways for E7 ubiquitination have been clarified (Table 1): one involves the SCF ubiquitin ligase complex; 50 the second involves the interferon-γ-inducible suppressor of cytokine signalling-1 (SOCS1), a cytokine signalling suppressor functioning as an antioncogene against various haematopoietic oncogenic proteins. SOCS1 induces ubiquitination and degradation of E7 in a SOCS-box-dependent manner, causing a concomitant increase in pRb levels. 51 These two E7 degradation pathways seem to be complementary, as the first appears to be predominant in the nucleus 50 whereas the second appears to be predominant in the cytoplasm. 51 Interestingly, the stability of E7 is also regulated by its interaction with the deubiquitinating enzyme ubiquitin-specific peptidase 11 (USP11). E7 binds to USP11, which reduces the ubiquitination and therefore enhances the stability of E7. 52 Thus, the stability of E7 depends on the net result of two opposing activities controlling the rate of ubiquitination. In addition to E7, HPV E2 and E6 have both been reported to be ubiquitinated and degraded by the proteasome.53–55 In the case of E2, the amino-terminal transactivation domain is required, 53 whereas with E6 its degradation is independent of E6/E6-AP binding capability. 55 The major ubiquitin ligase involved in E2 degradation was reported to be SCF. 56
E3 ubiquitin ligases in signalling pathways associated with HPV-induced carcinogenesis
E3 ubiquitin ligases in NF-κB signalling
Activation of nuclear factor-κB (NF-κB) has been implicated in a variety of cellular processes related to transformation and oncogenesis. NF-κB-dependent proliferation and protection from apoptosis are likely to be important in HPV-induced carcinogenesis. HPV E6- and E7-positive cells have been shown to be sensitized to interleukin-1β-mediated NF-κB activation and exhibit elevated NF-κB components. 57 In addition, microarray data suggest that NF-κB-responsive genes are upregulated by E6 protein in cervical keratinocytes. 58 Data on NF-κB expression and activation in relation to HPV in oral cancer have also been reported. 59 Although the exact mechanisms are still being debated, further studies have revealed that HPV16 E6 and E7 proteins inhibit tumour necrosis factor-mediated apoptosis by upregulating the expression of cellular inhibitor of apoptosis-2 (cIAP-2), a potent and critical antiapoptotic protein, leading to activation of the NF-κB signalling pathway.60,61 The RING E3 ubiquitin ligases tumour necrosis factor receptor-associated factor (TRAF) and cIAP are involved in NF-κB activation by HPV oncoproteins (Table 1).
E3 ubiquitin ligases in TGF-β/Smad signalling
One of the most widely studied roles of transforming growth factor-β (TGF-β) is its function as a tumour suppressor during the early stage of tumourigenesis. TGF-β binds to its receptors and transduces the antiproliferative signal by the intracellular proteins termed signalling mother against decapentaplegic peptides (Smads). 62 Alterations of the TGF-β/Smad signalling pathway, which is tightly regulated by E3 ubiquitin ligases, 63 are related to a range of human cancers, including cervical carcinoma. 64 Favre-Bonvin et al. 65 reported that HPV18 E6 protein interacts with a protein containing the PDZ domain, TIP-2/GIPC (Tax interacting protein, clone 2/GAIP interacting protein, C terminus) (Table 1), triggering its degradation through the ubiquitin–proteasome pathway. As this PDZ protein has been found to be involved in TGF-β signalling by favouring expression of the TGF-β type III receptor at the cell membrane, the depletion of TIP-2/GIPC in HeLa cells hampers TGF-β signalling and also diminishes the antiproliferative effect of TGF-β. Like the degradation of other PDZ proteins by E6-AP, the interaction between TIP-2/GIPC and HPV18 E6 is possibly mediated through E6-AP, though this point remains to be further defined. Moreover, HPV16 E7 sequesters Smads in the nucleus and inhibits the ability of Smads to bind to their target DNA sequence, 66 indicating that suppression of Smad-mediated signalling by HPV E7 oncoprotein may contribute to HPV-induced carcinogenesis. However, whether ubiquitination plays a role in the interaction of HPV E7 with Smad proteins needs further investigation.
E3 ubiquitin ligases in IGF signalling
Insulin-like growth factor (IGF)-I is a broad-spectrum growth factor that has mitogenic and antiapoptotic activities: high levels of IGF-I in the circulation are reported to be positively associated with heightened risk of many epithelial and glandular tumours.67,68 Data suggest a possible role of IGF-I receptor in primary and metastatic undifferentiated carcinoma of the head and neck. 69 IGF binding protein 3 (IGFBP3) is a major component of IGF binding proteins in the circulation and is an important regulator of biologically active IGFs. Functioning independently of IGF-I, IGFBP3 contributes to both p53-dependent and -independent apoptosis. 70 It has been shown that the serum concentration of IGFBP3 is inversely associated with the incident detection of oncogenic HPV and the incidence of oncogenic HPV-positive cervical neoplasia. 71 HPV16 E7 protein binds to and inactivates IGFBP3, 72 which may augment the serum level of active IGF-I, contributing to the progression of cervical disease. Further studies revealed that HPV16 E7 oncoprotein induces polyubiquitination and proteasome-dependent proteolysis of nuclear IGFBP3 in cervical cancer cells, 73 suggesting that the inactivation of IGFBP3 and the inhibition of apoptosis are important for the oncogenic activity of E7. The E3 ubiquitin ligase mediating the degradation of IGFBP3 has not yet been identified.
Other E3 ubiquitin ligases
Mdm2
Research data suggest that Mdm2 might be related to HPV-induced diseases. Overexpression of Mdm2 protein was found to contribute to the pathogenesis of oral carcinoma and oral verrucous hyperplasia. 74 In addition, the risk of oral squamous cell carcinoma associated with HPV16 L1 seropositivity was modified by Mdm2 promoter polymorphisms. 75 Identified as a p53 E3 ubiquitin ligase, Mdm2 was considered to be the regulator of normal physiological p53 degradation. 14 Mdm2 may be an alternative mechanism causing p53 protein dysfunction in oral squamous cell carcinoma. 76 Furthermore, p53 was shown to drive Mdm2 expression in a negative-feedback loop. 77 The role of Mdm2 in degrading p53 was best demonstrated by murine studies in which inactivation of p53 was able to entirely rescue the embryonic lethality of loss of Mdm2 function. 78 The effects of Mdm2 on p53 are concentration-dependent: low Mdm2 induces p53 monoubiquitination and nuclear export; high Mdm2 promotes polyubiquitination and nuclear degradation. 79 Interestingly, research showed that Mdm2 enhances transcriptional activity of HPV E2 protein, which is critical for the regulation of the viral life cycle. 80 (Table 1).
Nedd4
Gap junction biogenesis requires the oligomerization of six connexin proteins into a hexameric connexon and their subsequent signalling to the plasma membrane. 81 Among the gap junction proteins encoded by the 21 connexin genes identified in humans, connexin 43 (Cx43) is the most widely expressed and best studied. 82 Certain HPV-associated cancers, especially those of the cervix, have previously demonstrated downregulation of Cx43. In addition, loss of Cx43 expression was believed to be an early event in cervical carcinogenesis, and re-expression of this gene in cervical carcinoma cells was associated with decreased neoplastic potential. 83 Tomakidi et al. 84 found an almost complete disappearance of Cx43 when HaCaT cells were grown in an organotypic raft culture system to induce differentiation. However, the mechanisms behind these data remain unexplained. Leykauf et al. 85 revealed that Cx43 binds, both in vitro and in vivo, the HECT E3 ubiquitin ligase Nedd4 (neuronal precursor cell-expressed developmentally down-regulated 4). All three WW domains of rat Nedd4 interact with rat Cx43, and this interaction is modulated by the phosphorylation state of Cx43. The authors further showed that the Nedd4 domain WW2 binds to the canonical PY motif at the C-terminus of Cx43, triggering the internalization of Cx43 and posterior degradation of the gap junction proteins. Thus, the disruption of Cx43 by the Nedd4-mediated pathway may contribute to the development of HPV-associated cervical cancer.
EDD
E3 identified by differential display (EDD), which was originally isolated as a progestin-induced gene, encodes a 350-kDa E3 ubiquitin ligase containing a HECT domain. 86 Although alterations in EDD function have been linked to cancer development, 87 little is known about the biochemical activities of this protein. EDD was identified as a new HPV18 E6 binding partner. 88 Moreover, the authors claimed that changes in the levels of EDD expression may have a significant impact on the ability of HPV E6/E6-AP ubiquitin ligase complex to direct the degradation of several cellular substrates, p53 in particular, and thereby directly influence the development of HPV-induced malignancy. 88
Therapeutic intervention with E3 ubiquitin ligases for HPV-related disease
Only a small fraction of the genes involved in the ubiquitin system have been identified, and no comprehensive evaluation has been made of the genetic alterations of these components in human diseases. Defining the genetic and genomic targets of new drugs will therefore be important in finding therapeutic strategies. Selective inhibition of a deubiquitinating enzyme may allow more rapid degradation of given substrate(s). Bortezomib, a highly selective proteasomal inhibitor that has been approved by the US Food and Drug Administration, is the product of rational drug design directed towards selective inhibition of the proteolysis of a group of proteasomal substrates. 89 Similarly, specific targeting of an E3 ligase would diminish the ubiquitination of its substrate(s). The E3 ligases will probably provide exceptional specificity in the therapeutic modulation of the ubiquitin system because the system is hierarchical, with the greatest diversity at the E3 ligase level. 90
As described above, Mdm2 is an oncogenic RING finger E3 for p53. Theoretically, inhibition of the Mdm2–p53 interaction or inhibition of the conjugation of ubiquitin to p53 might lead to stimulation of the tumour-suppressor activity of p53. Furthermore, inactivating Mdm2 might prove beneficial in the treatment of tumours carrying wild-type p53, in that Mdm2 might work as a ubiquitin ligase for other antioncogenic proteins. 91 Chemical library screens for Mdm2 inhibitors have identified Nutlins (cis-imidazoline derivatives) that occupy the p53 binding pocket and thus prolong p53 action, including p53-dependent cell-cycle arrest and apoptosis. 92 Importantly, Nutlins are effective in vivo upon oral administration to halt the growth of nude mouse tumour xenografts without noticeable toxicity to healthy tissues. 93 Recently Nutlin-3 has been shown to be effective against many cancer cell lines, including HNSCC. 93 Moreover, another p53-stabilizing small molecule, 2,5-bis(5-hydroxymethyl-2-thienyl)furan (RITA), isolated from a screen at the National Cancer Institute, 94 binds the N-terminus of p53 and seems to prevent the recognition of p53 by Mdm2 or to stabilize the N-terminal α-helix domain of p53 in the Mdm2 groove, or both. 95 The RITA-stabilized p53 is transcriptionally active as it can facilitate expression of endogenous p53-target genes. 95 Most excitingly, RITA slows the growth of mice tumour xenografts in a concentration-dependent manner, 96 though its oral bioavailability was not tested in this study. One question that has not been answered is whether RITA and Nutlins might show synergistic antitumour activity in the mouse tumour xenograft model, as they attack different aspects of the Mdm2–p53 recognition process.92,95 It is also necessary to understand why RITA causes cell-cycle perturbations and a slight increase in apoptosis in p53-negative cells. 95 Does this mean that RITA targets other cellular proteins, such as other p53-related proteins p63 or p73? Finally, possible side-effects of long-term Nutlin/RITA administration should be further addressed in animal models. So although it is clear that certain small molecules have the desired effect of activating p53 by suppressing its degradation, further studies are required to determine whether these molecules will be useful in the treatment of human cancers, including cervical cancer and oral cavity carcinoma.
Although the exact nature of the contribution of E6-AP to HPV-induced disease is not fully understood,, E6-AP and HPV itself are considered to be pharmacological targets for cervical cancer. It has been shown that knocking down E6-AP expression by in vitro-selected ribozymes in HeLa cells induces apoptosis on exposure to genotoxic compounds. 96 Additionally, pharmacological targeting of HPV E6 with zinc-ejecting compounds inactivates E6 by disrupting its functionally crucial zinc fingers and thus silences aberrant E6-AP activity. 97 Such small-molecule zinc ejectors could potentially be applicable for clinical use in the topical treatment of cervical papillomas and cervical cancer, if their in vivo safety and bioavailability can be more thoroughly demonstrated. In principle, inhibitors of the HECT domain of E6-AP might be useful in the therapy of cervical cancer. Such HECT domain inhibitors might take advantage of the extra step involved in HECT-dependent protein ubiquitination that is not required for other classes of E3s. Because this step requires a large conformational change in the HECT domain, 98 it might be possible to find molecules that block this transition or stabilize the HECT domain in a nonfunctional conformation.
Conclusions and future prospects
Ubiquitin modification is a mechanism for controlling the function and availability of the regulatory proteins in cells; it may also provide a platform for HPV to achieve successful infection and carcinogenesis. In some cases, HPV proteins have evolved so that they structurally resemble the cellular E3 ubiquitin ligases, to permit degradation of target proteins and thereby enhance the efficiency of replication. In other cases, HPV proteins are capable of inhibiting targets of these E3 ubiquitin ligases that might retard viral replication. In addition, some HPV proteins are themselves regulated by degradation. Other E3 ubiquitin ligases also participate in cytokine signalling during HPV infection. It is likely that further studies of HPV will reveal additional examples of such processes, and, as in the past, these examples may shed light on important cellular pathways.
In order to further understand the significance of E3 ubiquitin ligases in the immunopathogenesis of HPV infection and HPV-induced carcinogenesis, future studies should focus more on the identification of new substrates of ubiquitination and the identification of more ubiquitin-binding domains. Fortunately, technological advances (such as the use of mass spectrometry) should assist the discovery of new substrates of ubiquitination. The discovery of nearly 20 types of ubiquitin-binding domains in recent years will provide a boost to the identification of ubiquitin receptors. If we know both the ubiquitination targets and the ubiquitin receptors, it may be possible to reconstruct ubiquitin signalling events biochemically as they happen during HPV infection in vivo. Advanced mechanistic understanding of ubiquitin signalling would be extremely valuable in the future development of possible treatments for HPV-associated diseases, including cervical cancer and HNSCC.
Footnotes
Declaration of conflicting interest
The authors declare that there are no conflicts of interest.
Funding
This work was supported by grants from the Education Bureau of Zhejiang Province (Y201121544) and the Health Bureau of Zhejiang Province (2012RCB025).
Acknowledgements
The authors thank Dr Qiang Zhou for his critical review of this article.