Retroviruses are enveloped RNA viruses of 80–130 nm in diameter with a genome size of 8–11 kilobases (kb) [11]. Retroviral genomes are made up of gag, pol, env, and 5′ and 3′ long terminal repeats (LTR). Retroviral vectors are rendered replication defective by deleting the structural, envelope, and enzymatic genes, and instead, the desired transgenes are inserted [12]. These replication-defective vectors, or proviruses, are incapable of continuing to spread after the initial infection of the target cells. The initial round of vector replication is dependent on several cis acting elements from the viral genome, including a promoter and polyadenylation signal, reverse transcription signal, transfer RNA (tRNA), primer binding site (PBS), and a polyurine tract, to initiate complementary DNA (cDNA) synthesis. Cellular tRNA binds to the PBS to initiate reverse transcription. Vectors containing the transgene are then packaged into viral particles as directed by a packaging signal (Ψ), PBS, and LTRs. Helper viruses, or plasmids, that carry the viral genes gag/pol and env, but lacking the packaging signal ψ [13], are required to express the viral proteins needed to produce vector-containing infectious viral particles. The helper function can also be provided by transfecting helper plasmids into cell cultures to produce helper cells [14]. The process of expressing the transgene in the retroviral vectors and delivering them to the target cells is called transduction [13].

Entry of the viral vectors into targeted cells is an essential first step in viral-vector-based gene delivery. Gaining entry into cells involves the interaction of the viral envelope with the host cell receptor. For lentiviral vectors, which are derived from human immunodeficiency virus type 1 (HIV1) [15], this involves a two-stage process [16]. Initially, the glycoprotein 120 (gp120) subunit of the envelope protein interacts with the target cell receptor CD4. The resulting conformational changes in gp120 allow binding with the co-receptors CXCR4 or CCR5, which belong to the chemokine family [17, 18]. Subsequently, another envelope protein subunit, gp41, signals fusion that is then followed by the release of viral capsid into the cytoplasm of the target cells. In addition to viral entry, the envelope plays an important role in determining cellular tropism by a technique called pseudotyping, where a viral vector is combined with foreign envelope proteins. Cellular tropism enhances the range of susceptible cells for the viral vectors. The lentiviral vector envelope can be substituted by the vesicular stomatitis virus G glycoprotein (VSV-G). This pseudotyping strategy has been shown not only to increase the range of possible target cell types, but also to improve stability and increase viral titers [13, 19]. Gammaretroviral vectors, e.g., murine leukemia viral vectors, transduce only dividing cells [20] and have a narrow range of cells that can have genetic material transferred to them. However, these vectors have ready access to the host genome because the nuclear membrane is removed during cell division. In contrast, lentiviral vectors transduce a wide variety of cell types, both dividing and nondividing cells such as neurons and B and T cells, albeit with varying degree of transduction efficiency [19, 21, 22].

A distinguishing feature of retroviruses is their ability to integrate into the host genome, resulting in sustained expression of the transgene. However, integration presents challenges in vector design if the vector is inserted at a site where the regulatory elements in the vector result in transcriptional activation of oncogenes [23]. Such insertional mutagenesis was the underlying cause for the adverse events in an early human gene transfer trial involving patients with severe combined immunodeficiency defect (SCID) [24]. Some patients involved in this trial subsequently developed T-cell leukemia. For lentiviral vectors, the risk of insertional mutagenesis can be reduced by designing self-inactivation (SIN) vectors that lack the viral transcriptional control elements, promoter/enhancer, and, thus, reduce the possibility of activating an oncogene located adjacent to the vector integration site [25]. Strategies have also been developed to target specific chromosomal insertion sites for the transgenes [26]. Research has also led to approaches to avoid integration. Because integration requires integrase attachment on the LTRs, mutating the integrase gene or modifying the attachment sequences of the LTRs may eliminate integration [27]. The transgene can also be expressed as episomal DNA without integration into the host genome. Such expression can occur for a relatively long duration in nondividing cells, such as retina cells [28, 29], or transiently in dividing cells.

Gammaretroviral vectors can accommodate up to 7 kb sequences [30, 31], while lentiviral vectors can accommodate larger transgenes, up to 10 kb [25]. In addition to the length of the inserted sequences, vector design must also consider stability issues. Retroviral vectors have half-lives of a few hours at 37 °C and up to a few months at −80 °C. A contributing factor to the instability is the loss of function of the envelope protein. Another factor to consider is viral titer. Lentiviral vectors can be produced at 10⁷ transducing units/ml. The titer can be increased to 10⁹–10¹⁰ transducing units/ml by ultracentrifugation [25]. However, titers may decrease when larger transgenes are inserted.

The first human gene transfer was conducted in 1989, as an immunotherapy to treat patients with advanced melanoma, using retroviral vectors and tumor-infiltrating lymphocytes [32, 33]. Targeted gene therapy began in 1990 with the adenosine deaminase gene for treatment of SCID [34]. Since then, numerous disorders have been treated by gene therapy [35]. They include inherited genetic disorders that involve autosomal X-linked recessive single genes or some autosomal dominant genes, and some acquired diseases, such as cancer, vascular diseases, neurodegenerative disorders, and inflammatory diseases. The first clinical trial using lentiviral vectors was approved in 2002 [36]. Currently several clinical trials are underway in which lentiviral vectors have been used to: (1) treat HIV infection, (2) transduce neuronal cells of the central nervous system for the treatment of Parkinson’s disease [37], and (3) deliver beta-globin gene for beta-thalassemia treatment [38].

There are safety concerns with the use of retroviral vectors for gene delivery. Both gammaretroviral and lentiviral vectors have the potential to illicit an immune response from the host. Immune reactions toward the viral vector result in rejection of all expressing cells. However, it should be noted that immune response is a beneficial outcome if lentiviral vectors are used to deliver genetic vaccines to treat HIV infection. Another serious safety issue is homologous recombination, which occurs when the packaging virus recombines with the vector to produce replication competent viruses. These viruses can produce harmful infections, and additionally gammaretroviruses can also cause cancer if an oncogene is activated by insertional mutagenesis [39]. In contrast, there is no evidence that lentiviral vectors result in oncogene activation through insertional mutagenesis. Another concern is the spread of the vector beyond the intended target tissue, which may cause persistent unwanted biological activity or unpredictable responses. All of these concerns have increased caution in the use of retroviral vectors although their advantages can be overwhelmingly attractive.

Adeno-associated virus (AAV) is a small, single-stranded DNA virus of 4,681 nucleotides (nt). The wild-type (wt) genome is made up of two genes, rep and cap, that encode four replication proteins and three capsid proteins, respectively. The three capsid proteins, Vp1, Vp2, and Vp3, are produced from the same open reading frame, but from differential splicing (Vp1) and alternative translational start sites (Vp2 and Vp3, respectively) [40]. Vp3 is the most abundant subunit in the virion and interacts with the host cell receptor. Recognition of Vp3 by the receptor determines cellular tropism of the virus. A phospholipase domain, essential for viral infectivity, has been identified in the unique N-terminus of Vp1 [41, 42]. The functional significance of Vp2 remains to be resolved. The viral genome is flanked on either side by 145-bp inverted terminal repeats (ITR) [43]. With the deletion of most of the viral genes, transgenes can be inserted into the cis 145-bp ITRs, to create a recombinant AAV (rAAV). The transgene can be expressed in the transduced cells without integration into the host cell genome and persists as episomal DNA [44]. The small genome size poses a limitation for AAV vector in delivering large transgenes. However, recent studies have shown delivery of genomes up to 6 kb, although delivery of these larger payloads was less efficient [43, 45]. Typically, transgenes of up to 5 kb are delivered. Strategies have been developed to improve efficiency, for example by splitting the vectors, with each vector containing approximately half of the transgene within the same cell. This approach, however, is still limited by viral packaging capacity. Additionally, cells have to be infected with different viral particles to achieve full transgene expression.

Transcapsidation is an approach to improve the packaging capacity, increase tissue tropisms, and transduction efficiency, where more than 100 different capsids from different serotypes can be exchanged to produce dozens of rAAV containing the same genome [45]. AAV serotype 2 is the best-studied AAV and was the first one used for gene transfer. However, vectors derived from alternative serotypes, e.g., 1, 4, 5, and 6 have been packaged with the same vector genome, but different viral capsids to improve efficiency and tropisms [46, 47].

rAAVs have been used in gene therapies in human muscle, liver, lung, central nervous system [48], and recently in the retina [49]. In a gene therapy trial of Parkinson’s patients, there was an improvement in the Unified Parkinson’s Disease Rating Scale (UPDRS) after 6 months for patients who received the glutamic acid decarboxylase (GAD) gene carried by an AAV2 vector that was delivered to the subthalamic nucleus [50]. rAAV2 vector carrying the gene encoding retinal pigment epithelium-specific 65-kilodalton protein (RPE65) has been used in a gene therapy trial to treat severe retinal dystrophy [51]. These results show promise, but further clinical trials are needed to demonstrate clinical significance.

There has been a four-decade long history of nonviral delivery system development, prompted by on-going safety concerns associated with viral vectors. DNA transfection protocols emerged in the late 1970s, and liposome-based gene delivery strategies were first reported in the 1980s, though the field really took off in the late 1990s with the discovery of siRNA which required the use of a delivery system in vivo. Nonviral delivery systems are attractive because they typically have lower immunogenicity, lower toxicity, and their production can be easily scaled for widespread clinical use; but they also have disadvantages. These disadvantages lead to lower efficiency at each stage of the delivery process, with an overall efficiency <0.1 %.

There are a number of basic engineering considerations in the design of all synthetic vectors. They should be nontoxic, nonimmunogenic, and biodegradable, protect their nucleic acid cargo against degradation, and efficiently deliver their cargo to both the cells of interest and the desired intracellular target. Optimizing the efficiency of nonviral delivery has been the focus of much research, which includes understanding the best targeting approach for selecting the cells of interest, how the stability of formulations change after administration, and how to get efficient endosomal release. Some of the barriers to delivery are illustrated in Fig. 18.2 and include: (1) formation of nucleic acid complex, (2) entry of the complex into the cells of interest, (3) endosomal escape of the nucleic acids, (4) dissociation of the complex, and (5) transport of the nucleic acid to site of action [52]. In contrast to viral vectors, synthetic vectors are poorly optimized to take advantage of existing cellular architecture. In particular, they cannot control endosomal release, have low diffusion rate, and are unable to take advantage of active transport mechanisms. Ideally for DNA, delivery is perinuclear, and for siRNA, it is targeting of the RNA-induced silencing complex (RISC). However, the rate-limiting steps of delivery are not fully understood, and there is not a good mechanistic understanding of how to rationally optimize loading rate.

A further challenge has been that the optimization of synthetic vectors for efficient in vitro cell culture delivery rarely translates to similar in vivo results. Monolayers of selected cells, in a carefully controlled environment, are not representative of the complex in vivo environment. The in vivo environment is a three-dimensional heterogeneous structure in an extracellular matrix, complete with enzymes, different cell morphologies, and a circulatory and immune system that cannot be easily replicated in vitro. In the worst case, some parameters can be optimized in vitro based on misleading effects which do not occur in vivo. In cell culture, size can be an advantage due to sedimentation efficiency, but in vivo smaller (less than 40 nm) particles are favored because of faster diffusion rates [53]. Another example is that positively charged vectors are beneficial in vitro for binding nucleic acids and enhancing uptake by interacting with the negatively charged cell membrane. However in vivo, negatively charged serum will bind to the vector, significantly reducing effectiveness. Some of the common designs for synthetic nanoscale delivery systems are illustrated in Fig. 18.3 and described in more detail in the following sections.

Complementary to the carrier-mediated viral and nonviral systems described previously, delivery of nucleic acids can be enhanced locally using a range of forces, or delivered locally using a number of noninvasive and minimally invasive techniques. Local delivery or the use of fields to target regions is attractive to avoid initial systemic clearance processes, enhance cargo concentration in the region of the cells of interest, and avoid side effects associated with large doses. However, disruption of the normal physiology of the local region because of these delivery methods may have unintended consequences, including generating an inflammatory response, altering lymphatic flow and perturbation of physiological functions. Ideally, the goal for directed delivery is to combine both regional and cell-specific delivery, to minimize disturbance of neighboring cells.