Among serological biomarkers, LDH remains to date the strongest prognostic markers in melanoma [26]. LDH is less sensitive in the early stage. However, it is indicative of the tumor growth and malignancy progression [27, 28]. It is still noteworthy that LDH, which occurs also in infection, inflammation and hemolysis, might lead to false positive readouts. S100 serves as diagnostic and prognostic marker for melanoma, as its expression is much higher in the malignant melanoma than in the normal tissues [29, 30]. It has been found that elevated level of S100, in particular S100B, is related to the poor prognosis and low survival rate [31, 32]. Vascular endothelial growth factor (VEGF) and melanoma-inhibitory activity (MIA) protein are involved in tumor-associated angiogenesis and melanoma cell interaction with the ECM respectively. Studies have shown that elevated levels of VEGF and MIA are associated with more advanced disease and poor survival rates [33, 34]. Other serum biomarkers for melanoma include tyrosinase [35], osteopontin [36], IL-8 [37] etc.

In addition to the serum biomarkers, molecules over-expressed in the tumor tissues also function as an effective indicator of the melanoma progression and metastasis. For example, matrix metalloproteinases (MMP), which proteolytically degrade ECM proteins, have been shown to promote melanoma cell migration and tumor progression. The MMP-1/3 positive melanoma was shown to correlate with higher malignancy and decreased overall survival [38]. Cell adhesion molecules (CAMs) such as the carcinoembryonic antigen-related cell adhesion molecule-1 and melanoma-associated antigen (MAA), which play an essential role in cell adhesion, show abnormal expression in melanoma and participate in melanoma migration and metastasis [39, 40]. Other tissue-specific biomarkers such as cyclooxygenase-2 (Cox-2) and galectin-3 have also been shown to associate with melanoma tumor size and malignancy and overall patient survival.

About half of the melanomas harbor the mutation of BRAF gene (BRAF^V600E, and possibly BRAF^V600K), which results in constant activation of BRAF kinase [6]. This likely induces the promotion of Ras-Raf-MEK-ERK signaling pathway and therefore the proliferation of melanoma. Like BRAF mutation, N-Ras gene mutation is also considered as the “driver mutation” and found in approximately 15-20% metastatic melanoma patients [4]. BRAF and N-Ras mutations have been shown to link to the metastasis and poor outcome [41]. With lower frequencies, KIT mutations have been observed in mucosal and acral melanomas and melanomas with sun-damaged skin [42]. Cutaneous melanomas are characterized with mutations of BRAF, N-Ras, MEK, NF1 and KIT, while most ocular melanomas harbor the mutations of GNAQ and GNA11 [43]. In order to discriminate Spitz nevi and Spitzoid melanomas, biomarkers such as CDKN2A, RREB1, MYC and CCND1 have been utilized [44].

Cytotoxic T-lymphocyte antigen-4 (CTLA-4), a protein receptor found on the surface of some T cells, functions as an immune checkpoint. By binding to the B7 on the APCs, it transmits an inhibitory signal to T cell activation, which confers vulnerability to the tumor invasion [45]. Ipilimumab (FDA approved for melanoma therapy in 2011), the human monoclonal immunoglobulin G1 against CTLA-4, has been shown to expand anti-tumor T cell activity and inhibit immune tolerance, therefore significantly enhancing the melanoma patients overall survival rate [46, 47]. Another anti CTLA-4 monoclonal antibody is tremelimumab, which is still undergoing human trials [48].

Programmed death receptor-1 (PD-1) is a protein expressed on the cell surface of T cells. By binding to the ligands PD-L1 and PD-L2 expressed on tumor and stromal cells, it transmits the signal in down-regulating T cell activation and promoting self-tolerance [49]. As PD-1 is also expressed on B cells and macrophages, it is likely more potent compared to CTLA-4 in negatively inhibiting immune responses [50]. PD-1 inhibitors, which block PD-1 and activate immune responses, have achieved varying success in different cancers including melanoma. Nivolumab, lambrolizumab and pembrolizumab, which are anti-PD-1 antibodies, have been approved by FDA for melanoma therapy [51, 52].

One of the major challenges of engineering OVs is that the natural tropism of viruses does not match the therapeutic need. For example, the administration of adenovirus systemically, among which the majority is sequestered by hepatic macrophages and hepatocyte transduction, results in adenovirus-mediated liver toxicity [68]. Therefore, approaches that retarget the viruses are needed to avoid the toxic side effects and facilitate the efficient infection to the targeted tumor cells. By modifying the viral coat proteins, which is referred to as transductional targeting, the viruses are engineered to infect the tumor cells that they do not infect naturally [69]. For example, the modification of hypervariable loop 5 in the capsid protein hexon in adenovirus prevents infection of hepatocytes and expand the anticancer therapeutic window [70].

Pseudotyping, which is the first approach of transductional targeting, is to replace viral attachment proteins of one virus with the tropism determinant ligands of other viruses whose tropism is against the targeted tumor cells. Exciting results have been observed by using oncoretroviral vector pseudotyped with a modified chimeric Sindbis virus envelop glycoproteins and conjugated with antibody specific for P-glycoprotein, for targeting metastatic melanoma [71].

While pseudotyping is limited by the viral attachment proteins and the receptors exclusively present on the target cells, the other approach of transductional targeting by using adaptors, the molecules with one end binding the viral attachment protein and the other binding the target cells’ receptors, possesses more flexibility without changing the structure of viral vectors [72]. For example, the fusing of ectodomain of coxsackie virus and adenovirus receptor (CAR) and a single chain antibody against carcinoembryonic antigen allowed the adenovirus to target on tumors and ablated the liver tropism [73]. Avidin, Biotin and monoclonal antibodies are also used in adaptor systems. For example, a biotin-acceptor peptide has been incorporated into the adenovirus fiber capsid protein and biotinylated and coupled to an avidin-binding ligand [74]. MMP and MAA, as the biomarkers of melanoma, play a critical role in melanoma progression and metastasis. Efficient targeting of retrovirus to melanoma cells has been done by fusing a single chain antibody recognizing the high molecular weight MAA (HMWMAA) followed by a proline-rich linker and a MMP-2 cleavage site, to the amino terminus of murine leukemia virus (MLV-A) envelop glycoprotein [75, 76]. Following the attachment with HMWMAA, MMP cleaves the MMP-2 cleavage site and allows the infection of MLV. Another example combining the avidin/biotin and monoclonal antibody is by first biotinylating VV to which the avidin is subsequently added and then coating the avidin bridge with biotinylated antibodies. The antibodies are against the cell surface makers of target cells [77].

The transcriptional regions consisting of promoters and enhancers regulate the kinetics and levels of mRNA production, and activators recruit RNA polymerase II for initiating the transcription of mRNA. Cancer-specific promoters can be activated in the presence of a subset of activators existing in cancerous cells but not in normal cells [78]. Therefore, transcriptional targeting in which part of the viral genome is placed under the control of cancer-specific promoter serves as a feasible way for generating tumor selectivity.

Tissue targeted virotherapy has been investigated in a variety of cancers including melanoma. Promoters of genes over-expressed in melanoma have been chosen and utilized, among which most are the above mentioned melanoma biomarkers’ promoters such as Cox-2, C-X-C chemokine receptor type 4 (CXCR-4), tyrosinase and survivin promoters [10, 78-80]. Cox-2 is closely associated with the progression and metastasis of melanoma, whereas benign nevi and tissues are Cox-2 negative [81]. CXCR-4 over-expressed on human melanoma cells has also been shown to play a role in melanoma progression [82]. Incorporation of Cox-2 and CXCR-4 promoters in adenovirus genome has been shown to significantly enhance the virus-mediated gene expression in the malignant melanoma instead of non-malignant primary melanocytes [79, 80]. Tyrosinase, which plays an essential role in melanin synthesis and melanocyte differentiation, has been used as a melanoma biomarker for disease detection purposes [83]. It has been demonstrated that incorporation of tyrosinase promoter to drive the expression of the viral genes of adenovirus leads to an obvious melanoma-selective viral cytotoxicity [84]. Survivin is expressed on a variety of tumors but undetectable in terminally differentiated tissues. It is associated with tumor progression and resistance to chemo- and radio-therapies, which is likely due to its ability of modulating apoptosis pathways and inhibiting apoptotic enzymes such as caspase-3 and caspase-7 [11, 85]. Lu et al. have reported that compared with Cox-2 and CXCR-4 promoters, the activity of survivin promoter is higher in melanoma cells and lower in normal tissues such as livers and epithelial melanocytes [79]. A continued search for a better cancer-specificity is warranted, as higher expression in tumor cells and lower expression in normal tissues for transgenes in viral vectors is always desirable.

Mutations in ras gene, which lead to the constant activation of Ras and therefore cell over-proliferation and tumorigenesis, have been reported in 30% of all cancers [86]. The incidence of ras mutations in melanomas is around 20% and the majority is N-ras mutations [87]. It has been demonstrated that reovirus and VSV, in unaltered form, replicate preferentially and effectively in the Ras-mutated cancer cells and therefore serve as natural OVs [88, 89]. It has been shown that reovirus efficiently infects and kills melanoma cells and regresses the xenografted melanoma tumors. In addition, Ras signaling pathway has been shown to be involved in reovirus-induced apoptosis in human melanoma cells [9]. Similarly, Noser et al. [89] have reported that Ras-overexpressing cancer cells are effectively killed by VSV and the oncolytic properties of VSV are dependent on the activation of Ras/Raf/MEK/ERK signaling pathway. It is noteworthy that the activation of Ras pathway is usually coupled with the down-regulation of IFN-induced antiviral responses such as the induction of protein kinase PKR, which further favors the virus replication in cancer cells.

Dysfunction of IFN signaling pathway has been reported in different cancers such as melanoma and breast cancer. Impaired IFN pathway aids tumor progression and immune escape, as IFN signaling is critical for the proper activation of key immune cells such as T, B and natural killer (NK) cells [90]. It has been reported that several melanomas are deficient in the key components of IFN signaling pathway such as STAT1 and STAT2 [91]. One example of enhancing the viral oncoselectivity via targeting the impaired IFN pathway in tumors involves in the nonstructural protein NS1 of influenza virus which has been demonstrated to block the function of PKR induced by IFN [92]. Influenza virus with NS1 deleted has been shown to selectively infect and lyse the melanoma cells that are PKR-deficient and regress the melanoma tumors in vivo, while failing to replicate in normal cells with intact PKR-mediated antiviral responses [93]. It is worthwhile to further investigate the application of viruses with the IFN or PKR viral inhibitor deletion for melanoma virotherapies, as IFN/PKR inhibitors are expressed by a variety of viruses such as VV and retrovirus [93].

Deficiency in apoptosis is a critical contributor to the therapeutic resistance of melanoma. Two pathways, the intrinsic and extrinsic pathways lead to the apoptosis. While Bcl-2 family proteins such as Bax and Bak and the induction of apoptosome and caspases are involved in the intrinsic pathway, the extrinsic pathway features the death ligands of tumor necrosis factor (TNF)-α protein family including TNF-α, Fas ligand (CD95L) and TNF-related apoptosis-inducing ligand (TRAIL) [94]. The application of tyrosinase promoter controlled CD95L has been shown to selectively induce apoptosis in melanoma cells while sparing the non-melanoma cells [95]. Viruses such as Newcastle disease virus (NDV) have been demonstrated to possess the ability of naturally inducing apoptosis in various cancer cells [96]. A wide array of viral proteins which have been shown to induce apoptosis include: NS3 protein of bovine viral diarrhea virus, viral protein R of human immunodeficiency virus, hemagglutinin-neuraminidase of NDV and EIA of adenovirus [96-99]. It is appropriate to incorporate these apoptosis-inducing proteins into the OVs for expression. In addition, the over-expression of anti-apoptotic proteins is a critical factor for melanoma’s resistance to the chemotherapy. Therefore, it is feasible to increase the viral oncolysis by co-administrating with inhibitors of the anti-apoptotic proteins, such as BCL-2 inhibitors. It has been reported that the co-treatment of BCL-2 inhibitor improves the viral oncolysis of VSV in the tumors which are resistant to the treatment of virus alone [100].

The components in the environment where tumors exist, including the blood vessels, fibroblasts, immune cells and elements of ECM, are collectively termed as “tumor microenvironment”. The interaction between tumor cells and tumor microenvironment is critical in tumor initiation, progression and invasiveness. Also, tumor microenvironment affects the delivery and spread of the therapies such as OVs in tumors. Therefore, in this section, we will review the strategies used for increasing the efficacy of OVs and combating the melanoma progression via targeting the melanoma tumor microenvironment.

Oncolytic virotherapy may represent a unique combination of viral tumor lysis and the activation of innate and adaptive immune responses. Targeted infection of tumor cells causes the release of TAAs and localized inflammation, which subsequently induces the tumor recognition of host immune system [113]. Therefore, in this section, we will review the strategies for enhancing immune recognition of cancer specific neo-antigens and antitumor immune responses.

In the past decade, there has been a rapid development of viruses with oncolytic activity as the clinical or preclinical anticancer therapeutics. More and more studies have shown that rather than using OVs as tumor therapy alone, it would be likely to realize OVs’ full therapeutic potential by combining with other treatment strategies such as radiotherapy, chemotherapy or immunotherapy.

A major part of the current standard therapeutic strategy for melanoma includes a range of targeted treatment entities based on specific gene mutations (BRAF inhibitors), pathways or enzymes (MEK inhibitors, tyrosine kinase inhibitors) and growth factors (VEGF inhibitors) [6, 19, 53-60]. Melanoma associated biomarkers like LDH, MIA, VEGF and MMP have shown a notable prognostic value where the level of these biomarkers has closely correlated with the advancement of the disease [33, 34, 38]. In the context of FDA approved therapies targeting melanoma biomarkers such as BRAF, MEK and VEGF, it is safe to say that biomarkers can not only serve as prognostic tools but are also useful in developing new therapies targeting melanoma. Targeted therapies fail to provide expected outcomes in many patients due to the genetic variability of melanoma, as melanoma tumors can be genetically heterogeneous and may or may not harbor mutations in target genes of therapeutic interest. Efficacy of treatment strategy for melanoma increases significantly when multiple targeted therapies are used in combination versus use of a single therapeutic agent, indicating the high level of dependence of targeted melanoma therapies on the genetic profile of the cancer. While nearly 50% of melanoma patients usually respond to BRAF inhibitors, the number of non-responding patients decreased by 40% when BRAF inhibitors were used along with MEK inhibitors as a combination therapy [6, 19, 55, 160]. In regard of this, immunotherapy (IFN-α and IL-2) and oncolytic virotherapy that are not heavily dependent on genetic heterogeneity of melanoma would circumvent limitations associated with targeted therapies that target specific gene mutations.

The OVs’ ability of causing tumor lysis, stimulating local inflammatory responses and serving as gene delivery vehicles, supports them as promising therapeutic agents for melanoma. While OVs are able to induce cell lysis and local inflammation that are not relying on genetic heterogeneity of melanoma, some viruses, in native forms, preferentially infect melanoma cells with aberrant genetic properties. Viruses such as reovirus, NDV and VSV have been shown to effectively kill the Ras-mutated tumor cells or lead to apoptosis, which make them as natural OVs for melanoma [88, 89, 96]. In addition, for those viruses without natural oncoselectivity, modifications targeting them to the genetic abnormalities of melanoma, often represented as biomarkers, have shown significant effectiveness in enhancing their oncolytic activity. For example, promoters of melanoma biomarkers such as Cox-2, CXCR-4, tyrosinase and survivin have been incorporated into the viral genomes, which result in significantly enhanced oncospecificity [10, 78-80]. A single chain antibody recognizing the HMWMAA, which is another melanoma biomarker, has been engineered into the virus for transductionally targeting melanoma cells [77]. Beyond the genetic and epigenetic alterations within the melanoma cells, the interaction between the melanoma cells and the microenvironment critically determines the transformation process of melanoma. Microenvironmental factors such as growth factors (VEGF, FGF, PDGF and TGF) and proteolytic enzymes (MMPs) involve in angiogenesis and tissue remodeling, highly promoting melanoma progression and metastasis. Modifications of OVs to target these microenvironmental factors have resulted in enhanced viral penetration and spread and melanoma regression [103, 110, 111].

OVs represent a class of novel agents as they can be engineered to harness the benefits of targeted therapy and immunotherapy, by inducing cancer cell death along with induction of a strong innate and adaptive immune response against cancer cells. The greatest success has been achieved with HSV (T-vec) and VV (JX-594) when developing OVs for cancer treatment. Some other viruses that are being studied in pre-clinical and clinical trials include HF10 (HSV), reovirus, NDV, VSV and coxsackie virus [115, 127, 152, 161, 162]. TAAs and immunostimulatory cytokines have been engineered into OVs, intended to attain enhanced immune recognition of tumor antigens and antitumor immune responses [115]. Using the viruses as the platforms for expressing immunostimulatory cytokines such as GM-CSF is appealing as it is likely to increase the local concentration of the gene products and minimize the side effects due to the systemic exposure [163]. While in some instances the expression of cytokines showed no inhibitive effect in virus replication, in some other cases the incorporation of cytokines such as IL-2, IFN and GM-CSF resulted in significantly decreased virus titer [126, 128, 164-166]. To cope with this issue, strategies have been designed to fuse the cytokine transgene to an exogenous cytokine-expressing regulator, which regulates the expression of the cytokine at specific time [167].

While some OVs are being studied for their efficacy as monotherapy, some are under investigation in combination with other therapies such as chemotherapy, radiotherapy, targeted therapy and more recently adoptive T cell therapy. T-vec is under investigation in several clinical studies where its efficacy is being assessed when combined with ipilimumab (CTLA4 inhibitor), and the results of a phase Ib/2 study of T-vec+ipilimumab suggest the enhanced synergistic efficacy with higher complete response and objective response rate [168]. Similar synergistic effects have also been observed when OVs were combined with anti-PD-1 and IL-2 [157]. Adoptive T cell therapy, with priming of T cells against tumor antigens followed by transfusion of the primed T cells in cancer patients, has shown consistent efficacy in clinical trials for melanoma [169]. Adoptive T cell therapy is also being investigated as combination therapy along with OVs. It has been shown that oncolytic adenovirus expressing IFN-γ combined with adoptive T cell therapy resulted in overall objective response of 38.5% in melanoma patients [170].

Although combination therapies have shown significant improvement and encouraging results, careful attention is still paid to the clinical trial designs, timing and dosing regimens, biosafety and biodistribution. For example, when combining the OVs and immunotherapy such as anti-PD-1, the timing of immunotherapy administration would be critical to avoid augmenting the anti-viral immune response or any earlier clearance of the virus. Rajani et al. [157] have demonstrated that administration of anti-PD-1 7 days after the intratumoral injection of virus is optimal to maximize the virus spread within the tumors. When combining OVs with chemotherapies such as CPA, CPA is able to eliminate the neutralizing antibodies against the virus and allow the virus to cause toxicities in systemic organs [153]. Therefore, it is essential to critically supervise the toxicity and the real-time levels of neutralizing antibodies against the virus, and carefully make dose-escalation decisions. For the combinational therapy of OVs and immunotherapy, it has been proposed to additionally combine with the systemic treatment such as T-reg depletion, to exert even higher synergistic efficacy [125]. However, one should be cautious to avoid autoimmune reactions when combining T cell enhancement with T-reg depletion, as immunotherapy such as anti-CTLA-4 treatment is already closely related to decreased number of myeloid derived suppressor cells [171].

Although it is reasonable to expect sufficient efficacy exerted by a one-shot curative agent, the immune “barriers” such as virus neutralization and sequestration make it more realistic to attain the optimal therapeutic effect with serial OVs for priming the antitumor immune responses. Strategies have been developed to prime the antitumor responses with one viral vector and boost with another, in order to minimize the anti-viral immune responses [172, 173].

In summary, this review describes OVs and their mechanisms in melanoma tumor regression. We also discuss different combinational therapies, which overcome the limitations associated with each therapeutic agent such as chemo-, radio- and targeted- therapies. Nonetheless, the quest for identifying new melanoma biomarkers that can channelize efforts for developing newer targeted therapies for melanoma should be continued.