Retroviruses (RVs) infect all vertebrates and those that infect humans are associated with various types of cancer, inflammatory diseases and human immune deficiency syndrome [12, 13]. This large family is classified into six genera that include alpha-, beta-, delta- and gammaretroviruses, lentiviruses and spumaviruses [14], which are based on genomic structure and sequence relationships. RVs are enveloped RNA viruses that consist of two identical copies of highly condensed positive-sense, single-stranded (ss) RNA enclosed by a capsid. The genome contains three essential genes, namely gag, pol and env that encode for the viral core proteins (matrix, capsid and nucleocapsid), the viral enzymes (protease, reverse transcriptase and integrase) and the viral envelope glycoproteins (surface and transmembrane proteins), respectively [9, 15–17]. RVVs efficiently transfer genes in vitro to a broad range of targeted cells and, since they have the capacity to integrate into the host genome, achieving long expression [2, 13, 18, 19]. For these reasons, RVs were among the first viruses engineered for gene therapy and have become the most commonly used RNA virus vectors [2]. Although an advantage, the integration of the vector into the host genome is also of major concern as random insertion into a pre-oncogenic site or a site responsible for the inactivation of tumor suppressor genes increases the risk of subsequent tumor development [10, 13, 20].

The most extensively studied viral vectors are from the poxvirus (PV) family, a family of viruses that can infect both vertebrates and invertebrates [44, 45]. This family includes variants of vaccinia virus (from the Orthopoxvirus genus) as well as fowlpox and canarypox viruses (from the Avipoxvirus genus) [45]. PVs contain large linear dsDNA genomes. In contrast to other DNA viruses, poxviruses encode their own transcription machinery, a viral DNA-dependent RNA polymerase and post-transcriptional modifying enzymes, allowing self-sufficient virus replication in the host cell [44, 46]. Since poxvirus DNA is replicated in the cytoplasm, random insertion of the viral genome into the host chromosome is not a concern [18, 45, 47]. Additionally, PVVs have a wide host tropism, form stable recombinants, have accurate replication and efficient post-translational modification of transgenes and can accommodate large inserts of foreign DNA [44, 45]. Interestingly, some strains have a natural tropism for tumor tissue, with a 10³–10⁴-fold higher expression than in other organs [6]. PVVs have long been used in safe and successful vaccination programs, with the vaccinia vector used to vaccinate more than one billion people during the eradication of smallpox [44, 48–50]. To overcome safety concerns, attenuated vaccinia viruses (such as modified vaccinia virus Ankara or MVA) [51, 52], which can infect mammalian cells and express transgenes, but cannot produce infective virus particles, were developed [45, 53]. Unfortunately, vaccinia and MVA vectors can only be administered once or twice to vaccinia-immune or vaccinia-naïve patients due to the development of neutralizing antibodies to the vector [54].

Adenoviruses (AdVs) are non-enveloped DNA viruses than can infect and replicate in a wide range of organs, including the respiratory tract, the eye, bladder, gastrointestinal tract and liver [9]. Generally, these viruses are known to cause benign upper respiratory tract illness and epidemic gastroenteritis and conjunctivitis in humans. Human AdVs are classified into six species (A–F), which are further subdivided into more than 50 serotypes [1–10, 12–52] based on their ability to agglutinate red blood cells, oncogenic potential, genomic organization and DNA homology [2, 3, 55–57]. These viruses are produced and purified to high titres (up to 10¹³ virus particles/ml) and can infect both dividing and post-mitotic cells [6, 55], replicating very efficiently in permissive cells [5]. The AdV genome is easily manipulated and inserted transgenes are maintained throughout successive rounds of replication [5, 6]. To date, the vast majority of recombinant AdV vectors are derived from the human AdV serotypes 2 and 5 of species C [55, 58]. AdV vectors can be genetically engineered to remain replication-sufficient, conditionally replication-sufficient or replication-deficient in the host [58]. First generation vectors are generated by substituting the early gene 1 (E1), or the E1 and early gene 3 (E3), with an expression cassette [8, 58–61], rendering the AdV replication-deficient and capable of replication only in specifically designed complementing cell lines [55]. Second generation vectors have had the E1, E3, as well as early gene 2 (E2), deleted to minimize the host inflammatory response and potential toxicity [62], while third generation vectors (also known as gutted, gutless, high capacity or helper-dependent vectors) are generated by deleting the entire AdV genome with the exception of the ITRs (inverted terminal repeats) and Ψ (packaging signal) regions required in cis for viral DNA replication and packaging (see Fig. 7.1) [63, 64]. For this reason, a replication incompetent helper virus is required to provide all the necessary AdV functions in trans, which has to be later separated from the AdVVs [65]. Deletion of a large amount of genome increases the cloning capacity of the vector and the adaptive immune response of the host is also decreased [2].

Adeno-associated viruses (AAVs) belong to the Parvoviridae family containing viruses that infect numerous species of mammals, including humans [2, 7]. Even though more than 100 AAV serotypes have been identified, only 12 of these have been shown to infect humans, with AVV-2 the prototype [10]. AVVs are small, non-enveloped ssDNA viruses and the genome could consist of either the sense or anti-sense strand. The genome contains two genes (rep and cap which encode for seven proteins Rep40, Rep52, Rep68, Rep78, VP1, VP2, VP3), flanked by 5′ and 3′ palindromic sequences (ITRs). The Rep regulatory proteins are required for replication and packaging, whereas VIP1-3 are structural proteins that form the capsid; the ITRs, on the other hand, are indispensable for viral replication, packaging and integration. AAV requires co-infection with another helper virus (either HSV or an adenovirus) or a stressed-cell environment (e.g. when cells are irradiated or treated with genotoxic compounds) to mediate its replication; in the absence of these it establishes latency by integrating site-specifically into host chromosome 19 [2, 66–69]. The different AAV serotypes each use unique mechanisms for cell entry, which results in different host tropisms. A safe, non-rescueable helper plasmid is used to complement the AAV coding sequences for cap and rep in trans. Following infection of a permissible cell line by either a wild-type AdV or Herpes simplex virus (HSV), the two plasmids are co-transfected into the cells, allowing for the formation of recombinant AAV [9, 10, 70]. Alternatively, a helper virus-free procedure has been developed, in which a mini-adenovirus helper plasmid is co-transfected with the vector and a packaging plasmid into an AdV E1-expressing cell line [2, 9, 10], resulting in the absence of production of infectious AdV particles.

Wild-type herpes simplex virus-1 (HSV-1) is an enveloped, dsDNA virus that is spread by direct contact and infects and replicates in the skin and mucosal membranes, before infecting cells of the central nervous system [2, 10]. The large complex viral genome encodes for more than 80 proteins that can be classified as either essential or non-essential for virus replication [71, 72]. HSVVs mimic the latent state of HSV-1, producing a highly infectious, efficient vehicle for delivery of foreign genetic material to both neural and non-neural tissue cells [2]. Vectors engineered from HSV-1 have the capacity to deliver large pieces of foreign DNA (up to 150 kbs) to the nucleus of most dividing and non-dividing cells. HSV has several other advantages, including the fact that it can infect many different host cell types (including cells of the nervous system), the viral DNA will not integrate into the host genetic material, and the complex nature of the virus genome, which contains about 40 genes that are not essential for virus replication and can be deleted without interfering with virus production in vitro, while the latent behavior of HSV can be used for stable, long-term expression of therapeutic transgenes [11, 71, 73]. High titers of pure, non-pathogenic HSV-1 vectors can be produced by introducing null mutations into viral immediate early genes. This disrupts the capacity for viral replication, but allows production of the vectors by in vitro complementation of these genes in trans [2]. HSV-1 is currently genetically engineered to generate three different types of vectors i.e. (i) recombinant attenuated virus vector, (ii) defective, replication-incompetent non-pathogenic recombinant vector, and (iii) amplicon vectors. Of the HSVVs, amplicons have been used in most anti-cancer applications [11, 74, 75].

Since the first clinical gene therapy trial in 1990s, great attention has been paid to cancer as a potential candidate for gene therapy applications. Today, among a total of 1,902 gene therapy clinical trials worldwide, those for cancer gene therapy constitute about 64 %. This number not only represents the high enthusiasm in this field but also indicates the multiple pathways that can be targeted to stop cancer growth. However, the ongoing failure of any of these clinical trials to yield an efficacious gene therapy product, with the exception of the Chinese approved Gendicine, denotes the multiple challenges that are yet to be overcome. Several strategies have been attempted to stop cancer growth, for example through re- or over-expression of tumor suppressor genes [95], introduction of a suicide gene followed by pro-drug administration [96], knocking down oncogenes or enhancing radio-and chemo-sensitivity of the cancer cells [97]. Another strategy is aimed at targeting the host non-cancerous tissue, primarily to enhance the anti-tumor immune response. This can be achieved by expression of transgenes possessing direct anti-tumor activity (e.g. cytokines) or by indirectly activating the host immune response (e.g., GM-CSF) [98]. This last strategy can also convey chemo-protection to highly vulnerable cells in the bone marrow to protect them against frequent high doses of chemotherapy [99]. The inability to target all cancer cells, especially after tumor dissemination means that systemic cancer gene therapy aimed at developing anti-cancer immunity appear most promising. Specific targeting of cancer cells is also crucial as non-selective expression of toxic transgenes in normal cells can lead to severe toxicity. Fortunately, using promoters specifically upregulated in cancer cells to drive expression of suicide genes can be used, such as epidermal growth factor receptor promoter, transferrin promoter, telomerase promoter or Prostate Specific Antigen promoters [100, 101].

Some of the therapeutic targets used in cancer gene therapy are summarized below;

As tumors grow they require a supply of nutrients and oxygen designated as angiogenesis. No matter how unregulated, it is crucial for neoplastic growth, expansion and metastasis [132]. Several growth factors such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), platelet-derived growth factor (PDGF), epidermal growth factor (EGF), interleukins (ILs) and matrix proteins are mandatory for angiogenesis [133]. Moreover, proteolytic enzymes such as cathepsin, urokinase-type plasminogen activator, gelatinases A/B, and stromelysin are critical to the neovascularization process [134]. Anti-angiogenic modulators include natural inhibitors such as endostatin, thrombospondin and angiostatin or inhibitors of the pro-angiogenic factors (antibodies, antisense RNA and soluble receptors for FGF, VEGF). Therefore, anti-angiogenesis-based therapies targeting ECs instead of tumor cells are a potentially powerful therapeutic approach to fight cancer. They possesses several advantages including; reduced toxicity and less capacity to develop resistance. In addition, the therapy is independent on tumor type [135]. In addition, ECs in the tumor are more sensitive to anti-angiogenic therapy because they proliferate more rapidly and express several unique surface markers [136].

Nevertheless, most of the antitumor benefits reported in the clinic after targeting of angiogenesis were subtle and only achievable after intratumoral administration with very high doses of the agents. This could be explained by the presence of large numbers of factors that control tumor angiogenesis, as well as the malleability of cancer cells to compensate for any one missing factor. Therefore, as in most of the gene therapy approaches, combination of antitumor therapies is the most likely way to accomplish effective anticancer results with minimal side effects.

In recent years, substantial progress has been made in vaccine development for malignant diseases. New technologies have been developed for the identification of a large number of tumor antigens that can be utilized to stimulate the patient’s immune system in order to specifically recognize and destroy tumor cells. Cancer vaccination encompasses therapies that involve the administration of some form of antigen to induce a specific antitumor immune response. The vast majority of vaccine studies today employ measures designed to activate tumor-specific T cells. In addition to tumor antigens, cytokine genes (such as GM-CSF) or co-stimulatory genes (such as B7) are often included to improve the effectiveness of the vaccine. While the success of vaccine strategies are dependent on the route of antigen delivery and adjuvant, it is important to keep in mind that their success often depends as much on the particular antigen being used. This can be illustrated with the human mucin tumor antigen, MUC-1, which in early vaccine studies caused cellular immune responses in mice, but humoral responses in humans.