hrHPV E5 oncoprotein: immune evasion and related immunotherapies

hrHPV is the main cause of cervical cancer and is also related to other tumors, such as head-and-neck [11] and anogenital [12]. It’s genome is composed of a Long Control Region (LCR) of nearly 1 kb and a region that encodes the early (E1, E2 and E4 to E8) and late proteins (L1-major capsid and L2-minor capsid proteins) [13].

Virus infection mainly occurs through sexual intercourse, generally by means of micro-injuries on cervical tissue. This event enables the virus to bind laminin-5 or -332 (a non-related heparan sulfate molecule) or heparan sulfate moieties, which may act as transient receptors in the extracellular matrix to attract the virus to the keratinocytes surface. Once the virus reaches the basement membrane of the epithelium, it interacts with several receptors for docking: first L1 and L2 bind the heparan sulfate proteoglycan (HSPG), which leads to conformational changes of these proteins; later, the attachment is to a non-HSPG molecule, whose identity has not been entirely elucidated yet. Then, the viral genome goes through the cell membrane by either calveolae or clathrin-dependent or -independent endocytic pathways, depending on the specific HPV genotype, and reaches the endosomal compartment [14].

Once internalized, HPV is uncoated and L1 dissociates from L2. L2 is suggested to be critical for HPV infection since it forms a complex with the viral genome and the nuclear domain 10, which are led together to the nucleus [14]. It is in this region that Papillomavirus initiates its replication and expression programs as an episome, and generates RNAs that undergo a large alternative splicing process. At this moment, the virus is into the basal layer, the expression of E1 and E2 genes is increased and a pool of infected cells with 10 to 50 copies of viral genome is formed. At this phase, there is no elimination of immunogenic particles that can be recognized by the immune system [13].

Once infected cells attain the parabasal and superficial layers, the expression of the other early proteins takes place. They keep keratinocytes in the proliferative stage, which prevents their normal differentiation process. In a final stage of virus cycle, novel virions are synthesized with the support of L1, L2, E2 and chaperone proteins [15]. Only now these particles can be released spontaneously along with the viral proteins, following the host cell natural apoptosis. This time is the suitable moment to ensure the nescience of the immune system, since the immune cells access is limited in the upper layers [5].

The oncogenic viral proteins E5, E6 and E7, are able to modulate the expression of pivotal proteins that control mitogen activity, differentiation program and immune evasion mechanisms. As a result of E2 gene loss during viral genome integration, both E6 and E7 reach elevated expression levels, which are essential for cell transformation. They disrupt the normal cell capacity for apoptosis and replication and interfere with the S-phase entry. E6 is able to interact with several key proteins which are directly or indirectly involved in cell cycle regulation, gene expression, apoptosis and differentiation/mitogen balance, such as p53, E6AP ligase, p300/CBP, histone acetyltransferases (ADA3), AP-1, Bak, Bax, FADD, procaspase 8, ERC-55 and paxillin [16]. E7, in turn, is capable of interacting with pRb, cyclin A/CDK2 (cyclin-dependent kinase 2), cyclin E/CDK2 (indirectly), p27 and p21 [17].

The transforming activities of E6 and E7 are supported by E5, which also has its own oncogenic properties. E5 induces mitogenesis in several ways and regulates important growth factor receptors or molecules, such as the epithelial growth factor receptor (EGF-R), Bcl-2, Bax, Fas and calnexin, which are involved in the control of cell differentiation, survival and growth [3, 5].

However, different HPV types express distinct forms of E5 and this causes differences in the carcinogenic competence of the Papillomavirus. Indeed, low-risk HPVs (lrHPVs) lack E5 or encode different polymorphic types with less transforming ability (E5β, -γ, -δ), resulting in non-tumoral clinical outcomes, whereas hrHPVs encode another E5 form (E5α) that may lead to lesion progression and tumor development [18]. It is known that different forms of E5 have different abilities of binding the EVER and ZnT-1 proteins (these proteins are important in viral genome replication and immune response). HPV16 E5 (E5α) is able to bind and prevent the activities of EVER1, EVER2 and ZnT-1, which results in the inhibition of MTF-1 transcriptional effect. Hence, the different cell transformation ability of E5 from high- and low-risk HPVs could be directly related to this feature, insofar as the other types of E5 (β, γ and δ) are not able to bind these proteins [19]. Thus, this finding indicates that the difference between hrHPVs and lrHPVs could be determined by the presence of a particular E5 ORF in ELR region (region between the early and late genes of HPV genome). Therefore, the oncogenic competence of HPV could be extensively related to E5 oncoprotein and this subject is worthy of further experimental studies [18, 20].

It is well established that HPV16 E5 oncoprotein acts in the early stages of infection/transformation [5] and, in particular, in DNA synthesis in suprabasal epithelium and in viral genome amplification along with the E7 oncoprotein [21]. After the viral DNA is integrated into the host’s genome, E5 gene is usually lost, suggesting that its activity is not required for the late transformation stage. However, some studies demonstrated that HPV16 E5 remains expressed even in high grade lesions and invasive cancer. This might be due to the presence of an episomal form or head-to-tail integration, which is responsible for E5 expression [22]. In unpublished data, our group confirmed the HPV16 E5 mRNA expression in biopsies from patients with high grade and cancerous lesions. Thus, it is likely that this oncoprotein may interfere with all stages of tumorigenesis.

Finally, since E5 is mainly active in the early phase of infection, it has become a key target of therapeutic studies which seek to prevent the progression of the lesion to pre-malignant stages and carcinoma [23].

The host’s immune system harbors both innate and adaptive immune responses, which are responsible for a successful protection against the papillomavirus and tumor development [24]. However, HPV can sometimes escape from all immunologic efforts, due to its oncoproteins activities, including E5. In the following paragraphs, it is presented how some key agents from the host’s immune system act in a non-infection condition and, in the following topic the E5 activity on these agents is discussed.

The initial barrier to genital infection by HPV is the cervical epithelium that contains keratinocytes (KC) and dendritic cells (DC), both of which are responsible for antigen presentation to T cells through MHC I and II. Keratinocytes are able to induce DC and NK cell maturation and to promote CD4⁺ and CD8⁺ T cells activity. Similar to KC, DC recognizes antigens through the pattern recognition receptors (PRRs) and synthesizes a wide range of signaling molecules, including interferons and cytokines. This cell is also capable of direct interaction with the NK cell, promoting its activation. These initial innate immune response efforts are essential to the development of an effective clearance of HPV, but in a minority of occasions, E5 and the other oncogenes are able to successfully escape from host immune surveillance [25, 26].

In the next sequence of events, the activation of T cells comprises the major effector cells for the regression and eradication of HPV infection. These cells have cytolytic capacity and synthesize antigen-specific antibodies and a wide range of signalling mediators in order to maintain antigen-specific memory B cells and lyse infected/tumour cells. They establish a specific cytokine profile concerning the antitumor response, in a complex balance between Th1, Th2, Th17 and Treg cytokines which won’t always result in cancer clearance. The absence of T cells were associated with persistent infection and neoplastic progression [5, 25] and, because of that, the priming of these cells is continuously used as immunotherapeutic targets [27, 28].

In this scenario, the presence of checkpoint molecules/receptors and their inhibitors also play an essential role supporting cancer development. As an example, the receptors KIR, CTLA-4 (Cytotoxic T-lymphocyte antigen-4) and PD-1, and its ligand, PD-L1, are immunosuppressive molecules, currently associated with a poor prognosis [29, 30].

The identification of CTLA-4, a protein receptor on the surface of T cells, dates back to the late 1980s, when it was shown that it puts brakes on T cells, preventing them from launching full-out immune attacks and reducing them to a small pool of memory cells [31]. As well as CTLA-4, PD-1 is also present in T cells and activates an inhibitory cascade, hampering T cell responsiveness and its cytokines secretion. Its ligand, PDL-1, which is expressed in antigen presenting cells (APCs), was also found substantially expressed in tumour cells of HPV-related cancers [32‐34]. However, a complete description of this important immunosuppressive milieu goes beyond the scope of this review and many reviews dealing with this issue have been recently published [29, 30, 35].

The major immune mechanism disrupted by E5 oncoprotein is antigen presentation, accomplished by MHC antigen processing. Following the guiding principles of the “missing-self” hypothesis, other viruses, as HPV, disrupt MHC expression in a variant-selective manner to protect infected cells against the NK and CTL cytotoxicity. This occurs by a selective downregulation: while HLA-C and the non-classical MHC class I molecule HLA-E, which binds to the inhibitors receptors of NK cells, maintain their constant levels, HLA-A and -B are downregulated and are not able to induce CTL activation and accomplish infected cell lysis [36, 37].

Classical MHC is classified in class I (HLA-A, -B and -C genes) and class II (HLA-DR, -DQ, -DP genes). MHC I is a complex formed by a polymorphic heavy chain codified by the loci HLA-A, -B and -C in chromosome 6, and by a light chain, denominated β2-microglobulin, codified in chromosome 15 [38]. These HLA types are responsible for coupling antigen peptides in ER, which are transported to GA and shown at the cell surface to T cytotoxic lymphocytes [39]. In infected cells, however, HPV16 E5 blocks this mechanism. It is of common knowledge that MHC I is downregulated at the membrane surface of keratinocytes and, thus, the viral antigen recognition and maturation of NK and T cells are impaired. This effect is linked to the interaction between the hydrophobic transmembrane domain of HPV16 E5 and the heavy chain of MHC I in Golgi/ER membrane [40] and not by the blockage of the expression of the heavy chain or the transporter protein TAP1 [36].

It was also shown that HPV16 E5 can downregulate the MHC I membrane expression, by causing alkalization inside the membrane compartments, which keeps MHC I inside the GA [6], and by interaction with the chaperone calnexin, which causes retention of MHC I inside the ER [41]. A similar mechanism to this occurs for CD1d, a MHC I-like glycoprotein that presents self or microbial lipid antigen to NKT (natural killer T) cell. The downregulation of this molecule, which will eventually be degraded at the cytosol, is an immune evasion strategy performed by several other viruses besides HPV. CD1d is important for antigen presentation to CD1d-restricted invariant NKT cell (iNKT) that, once activated, plays a key role in anti-viral response since it causes the lysis of infected cells and modulates Th1/Th2 polarization [42].

In vitro studies showed that the HPV16 E5 interaction with the 16 kDa subunit of ATPase resulted in decreased MHC I levels and a reduction of CLT cells recognition and activity [5]. The same results were found in a bovine model. This mechanism of altered transport was further evidenced by the lack of increased surface levels of MHC I even when total levels of MHC I were increased by interferon (beta and gamma) treatment [43]. Similarly, HPV16 E5 was able to interfere with MHC II surface expression [44].

Besides interference with antigen presentation, HPV16 E5 also modulates the immune and inflammatory pathways through other manners by interfering in EGF-R activation and transduction signaling such as the mitogen-activated protein kinases (Ras/Raf/MAP kinase) and the phophoinositide 3-kinase (PI3K/Akt) [6]. The first transduction pathway is involved in cytokines synthesis [45] while the second regulates chemokine production, DC differentiation from monocytes, NK maturation and leukocytes migration [46]. EGF-R activation also induces an increase in the expression of ganglioside-1 (GM-1) and caveolin-1 on the cell surface, affecting cell signaling and vesicular trafficking. GM-1 inhibits cytotoxic T lymphocytes and immune synapse formation, and increases the EGF-R proliferative response [47]. Figure 1 summarizes several other transduction molecules which are involved in EGF-R signaling, including the ones exemplified above, besides the E5 interferences in TGF-β signaling mediators discussed in "Transforming growth factor β signaling" topic.

EGF-R activation by HPV16 E5 also interferes in the inflammatory pathway. After EGF-R phosphorylation, this receptor is able to induce the transcription of COX-2, whose upregulation was usually associated with malignant processes and with a decreased overall survival rate and increased metastasis [48, 49]. The kappa B nuclear factor (NF-κB), as well as the prostaglandin E2 (PGE₂), both related to COX-2 pathway, were also proposed as bad prognostic factors. NF-κB is a key factor of immune response [50, 51] and PGE₂ seems to be involved in anti-inflammatory response, since it impairs T cell arrest and interacts with APCs, which compromises T cell maturation [52]. In addition, this prostaglandin stimulates cellular proliferation, angiogenesis, cell migration, invasion and survival [53] (COX-2 pathway can be seen in details in Fig. 2). Therefore, both molecules may be an alternative target for therapeutic interventions, since recent trials have shown that the inhibition of COX-2 led to adverse side effects [49].

In addition, E5 is responsible for the deregulation of other important key participants in the immune response, such as the natural killer cells, the transforming growth factor-β and interferons, as highlighted in the following sections. Table 1 summarizes the E5 influence on host immune activities.

The study of the interaction between HPV and the host’s immune system has opened up a number of therapeutic opportunities [102, 103]. These therapeutic strategies aim to modulate immune responses for treating or preventing infection. Since there are no antiviral drugs for HPV infection, cervical lesions are treated by chemotherapy, radiotherapy, surgery or hysterectomy as the most radical of all. However, these procedures are not effective in all cases, which results in a high recurrence and significant global mortality rate, around 52% or more than 270.000 deaths per year [104]. When HPV-related cancers are included, the number of deaths raises to 295.000 per year or one every two minutes [105]. About 95% of anal, 70% of oropharyngeal and 65%, 50% and 35% of vaginal, vulvar and penile cancers, respectively, are caused by hrHPV [106].

Two prophylactic vaccines (bivalent and quadrivalent) were produced to prevent HPV infection [107], and recently a 9-valent vaccine was approved, which protects from nine HPV genotypes (6, 11, 16, 18, 31, 33, 45, 52, and 58) [108]. However, there is still a lack of therapeutic vaccines. These vaccines usually aim to activate immune cells, either from innate or adaptive response [109], in order to simulate the natural immune response and reverse the viral immunomodulatory mechanisms. It was reported, for example, that T CD4⁺ and CD8⁺ cells levels were reduced in high-grade cervical lesions [110], unlike what happens in spontaneous regression lesions [111]. Beyond that, a specific cytokine expression profile was also described in these opposite situations (lesion progression and lesion remission) [112, 113] and all these mechanisms allowed the adoption of a number of immunotherapeutic approaches of great potential, including those related to E5. Nowadays, the use of new immunotherapeutic approaches have emphasized the strengthening of non-specific immune response (i.e. cell-mediated responses, cytokines production and secretion) and stimulated researches about the synthesis of monoclonal antibodies against checkpoint inhibitors such as PD-1/PDL-1 and CTLA-4 [114].

HPV oncoproteins have been widely employed in cervical cancer immunotherapy and E6 and E7 were extensively tested. However, these two oncoproteins were not able to fully eradicate pre-cancerous lesions and for this reason, E5 has emerged as a third alternative [9]. The use of HPV E5 protein in therapeutic vaccines does not aim to eradicate E5 function, but to stimulate the host’s immune system to combat viral infection and lesion progression. It is well established, for example, that dendritic cells play a key role in the development of innate and adaptive responses due to its antigen presentation activity [115] and E5 can impair this activity. In addition, circulating DCs are reduced during viral-induced cell transformation and cervical cancer progression [116]. Therefore, an autologous DC administration, previously pulsed with E5 antigen in culture, could be performed to bypass this problem and stimulate properly the host’s immune system against virally infected cells.

So far, several types of therapeutic vaccines against HPV16 were analyzed: peptide-based, protein-based, DNA/RNA, viral/bacterial vector, plant-derived and DC-based vaccines. All of those methodologies were applied for E6 or E7 vaccines but only three of them (peptide-based, DNA, and viral vector preparations) were utilized for E5 vaccines. Among the methodology tested for E5, viral vector vaccines are a promising therapeutic strategy for the presentation of pathogen antigens to sensitize the host’s immune system, inducing a strong cytotoxic response. These vaccines are highly immunogenic since they are able to go inside the host cells where they synthesize the antigens of interest and cause the lysis of infected cells [109].

One of the first studies to evaluate HPV16 E5 therapeutic vaccine utilized recombinant adenovirus bioengineering to express E5 in a murine model. This vaccine produced a reduction of tumor growth in vitro and the immunological response was T CD8⁺ dependent and T CD4⁺-independent, suggesting that HPV16 E5 is an antigen able to induce tumor rejection [117]. In another study, a recombinant vaccinia virus for Bovine papillomavirus 1 (BPV-1) [118] E5 was effective in reducing experimental mouse tumors, but the same technology applied to HPV16 E5 elicited no positive immune effect, unlike HPV16 E6 and E7 [119]. In a DNA vaccine encoding HPV16 oncoproteins, E5 was associated with E6 and E7 and genetically fused to herpes virus glycoprotein D. This vaccine showed a specific and substantial T CD8⁺ response for each oncoprotein individually in a TC-1 tumor mouse model. This vaccine exhibited a higher anti-cancer (to 100%) effect when vectors encoding GM-CSF or IL-12 were co-administrated [120]. These cytokines are co-stimulators of immune response and activate/recruit central cells of innate system such as dendritic and NK cells [121], which results in maturation of APC and CTL response. Both co-stimulators were successfully utilized in pre-clinical and clinical trials of cervical cancer therapy, in particular IL-12 [103, 121, 122]. This strategy was utilized with some success by Borysiewicz et al. in the first human trial of immunotherapy with E6 and E7 recombinant vaccinia virus more than 20 years ago [123].

When comparing the results, a bias could be the different viral vector chosen. Unlike adenovirus, which was capable of inducing a CTL response for different recombinant virus, the vaccinia virus was unable to stimulate an appropriate immune response against HPV16 E5. It is known that adenovirus is capable of inducing a CTL response for different recombinant virus. Moreover, different tumor cells were used to induce tumors in mice model, which makes it difficult to compare the results. However, these studies suggested that HPV16 E5 is able to stimulate a host immune response with the induction of E5-specific cytotoxic T lymphocytes. These cells were activated by dendritic cells, thus, receptor agonists can further strengthen their stimulation to achieve a successful therapeutic outcome.

Additionally, the T cell-mediated immune response induced by HPV16 E5-peptide (delivered through recombinant adenoviruses) may be influenced by specific HLA-A, in terms of intensity-response. The HPV16 E5 epitope reduced the oncogenic potential and generated specific T memory cells that are restricted to HLA-A*0201, observed in patients samples, unlike the whole HPV16 E5 oncoprotein, which did not show signs of this ability [124].

Potent peptides can be used to induce a host immune response against viruses [93, 109] and HPV16 E5 showed in silico evidence of being an useful therapeutic strategy [125]. Peptide-based vaccine is a safe and easy methodology to execute despite having a poorer immunogenic capability than protein vaccines [9]. The in silico study evaluated the most potent epitopes from the HPV16 E5 oncoprotein, which were capable of stimulating T and B cell activities and were also entirely immunogenic through MHC class I and II [125]. This recent work identified the epitopes that are most likely to be effective as a vaccine, but in vitro and in vivo studies need to be carried out to substantiate these findings.

The use of adjuvants and its features can also have influence on the success of the vaccine. In a work using HPV16 E5 peptide vaccine in mice, activated T CD8⁺ cells and synthesis of IFN-γ were observed when administered with CpG oligodeoxynucleotides as an adjuvant [126]. CpG is a common adjuvant used to induce dendritic cell responses through TLR9 activation and the cell-mediated response in a final stage, improving the vaccine efficacy. The strong immune response caused by CpG/E5-peptide vaccine was correlated with the reduction of tumor growth [8].

Long peptides present positive features of both peptide and protein vaccines: safety, easy production and good immunogenicity. Moreover, long peptides require professional APCs, which reduce immune tolerance induction, and provide more epitopes for MHC presentation, thus increasing the response levels. HPV16 E5 long peptide together with an E6 and E7 construct was able to reduce tumor burden in mice as well as to induce a strong and prolonged immune response, with the induction of specific T CD4⁺ and CD8⁺ cells. Along with these results, the E5 + E6 + E7 vaccine had a synergistic effect and was more effective than the E5 or E6 + E7 vaccines alone. Recombinant adeno-associated virus was used as delivery system to ensure a strong immunogenicity [9].

A possible good alternative to activate key cells or to create a favorable tumor milieu to achieve an effective immune response is the cytokine modulation. Several studies have given proof of the benefits of cytokine administration or their expression modulation for induction of the host’s immune system by itself (mainly through the activation of cytotoxic cells) or by supporting other immunotherapeutic approaches [121, 122, 127].

Among the vaccine methodologies already mentioned, DNA vaccines are a promising therapy for tumors and viral infections. Key features of this methodology are: easy production with high purity and ability to induce stable expression of antigens [128]. However, this type of vaccine has a weak intrinsic potency when compared with peptide and protein vaccines [9]. Thus, the use of synergistic approaches is preferable. In a work published recently, the host immunologic activation by E5 DNA vaccine was observed in mice using the whole HPV16 E5 ORF gene and the HPV16 E5multi, a harmless version of the oncoprotein designed to express multi epitope sequences. The HPV16 E5 vaccine caused a great reduction of the tumor volume, similar to that induced by HPV16 E6 and HPV16 E7 DNA vaccines, although E6 and E7 vaccines were more effective in delaying tumor growth. E5 vaccine, unlike E6 and E7, was also able to induce T CD8⁺ cell activity causing tumor shrinkage in absence of any adjuvant [10]. In another study, the same HPV16 E5 and E5multi sequences were fused to a capsid protein sequence of the plant virus PVX that is known to induce strong immunological responses [128]. These new E5 DNA vaccines were able to induce T CD8+ response and improved antitumor activity in new murine models of ano-genital or oropharynx tumors [129].

However, other vaccine strategies for E5 remain untested. One example of these unexplored strategies is the autologous DC vaccine. Autologous monocytes were differentiated in DC in vitro and these cells were administered back in the patient after being loaded with HPV 16/18 antigens in cultured system such as E7 oncoprotein, which led to a rise in CD8⁺ T cell activity and secretion of IFN-γ. DC is known as the most potent cell for CTL induction, and the use of HPV16/18 oncoproteins as antigens for development of DC vaccines showed good results [130‐132], but no trials were carried out for the E5 oncoprotein. In some cases, an additional synergistic approach can be used to improve vaccine efficacy, such as the blockade of PDL-1 [132]. A DNA vaccine comprising a gene fusion of HPV16 E6/E7 with CTLA-4 induced a greater CTL response in a mouse tumor model.

Another important example of untested strategy is the use of recombinant vaccines with bacteria as vectors applied for E5. This strategy was tested with the E7 oncogene, using Listeria monocytogenes (Lm) and proved to be safe in a phase-1 safety protocol. In this vaccine, E7 was fused with the non-hemolytic protein fragment listeriolysin O (LLO). Lm itself was found to induce strong CD8⁺ and CD4⁺ responses in patients with cervical cancer, and the recombinant E7/LLO antigen showed a significant degree of specific immunogenicity as well [133]. However, systemic listeriosis following vaccination was reported in a patient. Following this serious adverse event, the trial was immediately put on hold and thereafter, following consultation with the vaccine manufacturer (Advaxis), the trial was terminated early [134].

Together, these works indicate the benefit of using different immunotherapeutic approaches at the same time to act synergistically, potentializing therapeutic effects. The use of E5 by means of these strategies should provide more knowledge about E5 activity as an antigen and help in the search of an appropriate HPV-related cancer therapy. In general, E5 vaccines may represent a good choice for immunotherapeutic strategies. Although more studies are needed, E5 seems to be a valid tool to achieve a more effective therapeutic HPV vaccine, possibly integrated by check point inhibitor treatments or miRNA regulation [135]. Finally, other E5-related activities could be used in HPV-related therapy and some of them are summarized in Table 2.