Heart and Metabolism

Future developments in gene therapy

Andrew H. Baker Division of Cardiovascular and Medical Sciences, University of Glasgow, UK
Correspondence: Dr Andrew H. Baker, Division of Cardiovascular and Medical Sciences,
University of Glasgow, Glasgow G11 6NT, UK.
Tel: +44 141 2112100/2116, fax: +44 141 2111763, e-mail: ab11f@clinmed.gla.ac.uk

Abstract

Gene therapy for diseases of the cardiovascular system has received an enormous amount of attention over the last decade, and numerous preclinical studies provide testimony to its potential. Importantly, clinical trials are underway, for which further results are eagerly awaited.
However, major concerns have arisen pertaining to the side effects of unregulated therapeutic gene overexpression as well as to the toxicity and immunogenicity of vectors used to deliver these genes. Hence, the development and refinement of gene-based therapeutics are crucial for the future long-term success of this technology in the clinic. In particular, stringent control of gene expression in defined target tissues is required. Furthermore, recent advances in the understanding of virus biology and pathology now provide the opportunity to create “designer” viruses tailored for individual cardiovascular disease applications. n Heart Metab. 2002;18:28–35.

Keywords: Gene therapy, cardiovascular disease, viruses

Introduction
Within the broad discipline of cardiovascular and related diseases, gene therapy holds promise for effective and sustained correction of human diseases, both inherited and acquired. In 1990 the direct transfer of genes to the vessel wall was first described [1], which unleashed an astounding sense of anticipation for the creation of novel therapies to alleviate cardiovascular disease. Feasibility studies for high-level gene delivery to diseased blood vessels, endothelium, and myocardium using “marker” or “reporter” genes were quickly followed by evaluation of candidate therapeutic genes in preclinical in vivo models. The unparalleled use of viruses, particularly recombinant adenoviruses, resulted in high-level therapeutic protein production, and strategies were adopted to deliver such viruses to target sites in vivo, including blood vessels and myocardial tissue [2–4]. The wave of enthusiasm has led to clinical trials, particularly of gene therapies to alleviate myocardial and peripheral ischemia, postangioplasty restenosis, vein graft failure, and familial hypercholesterolemia [5]. However, in the late 1990s scepticism arose, particularly relating to toxicity and immunogenicity of viruses and the impact of sustained, high-level overexpression of therapeutic genes on biological systems. Following the tragic death of Jesse Gelsinger, attributed directly to viral gene transfer [6], the need to design, refine, and fully evaluate the safety of gene-based medicines became overwhelmingly apparent. Where does this leave gene therapy for cardiovascular disease?

Local gene delivery
The vast majority of applications being considered for cardiovascular gene therapy require the genetic manipulation of distinct cell populations in a localized and selective manner, eg, diseased myocardial tissue, damaged blood vessels, saphenous vein conduits for bypass grafts, etc. Catheters and ex vivo genetic manipulation provide localized access to the tissue through which high concentrations of the vector can be exposed to the target cell population. Although there has been success with nonviral gene delivery to the vasculature in human subjects [7], short exposure times and lack of efficiency restrict gene delivery. Hence, the efficient adenovirus system has been used with great success. Adenovirus-mediated transduction of target cells is depicted in Figure 1.

Figure 1. Cell transduction by adenovirus vectors. (A) Adenoviruses bind to the cell surface by interaction with the coxsackie and adenovirus receptor (CAR) [8, 9], and internalize into endosomes using integrins ·v‚3 or ·v‚5. Following endosomal escape, adenoviruses target to the nucleus where they are maintained episomally and express the therapeutic protein encoded by the foreign gene inserted into the recombinant genome. (B, C) Adenovirus binding and internalization can be monitored by fluorescent microscopy of labeled particles, indicated in red (C). (B) Uninfected control cells.

However, for efficient and high-level gene transfer, high titers are required, leading to acute toxicity and immunogenicity [10, 11], as well as virus dissemination to distant sites [12]. Strategies to improve the efficiency of gene transfer in this context would, in theory, result in the requirement for reduced viral loads to achieve equivalent levels of gene transfer. Additionally, the reduced virus titer combined with the use of cell-specific promoters would circumvent the potentially deleterious effects of virus dissemination. This is an important objective for the development of local gene therapy for cardiovascular disease.

Systemic gene delivery
Thus far, systemic gene therapy for cardiovascular disease has been limited to the expression of therapeutic transgenes from cells in the liver, as this is the principle site of virus sequestration following systemic administration [8]. Logically, overexpression of genes that regulate lipid metabolism has met with success [13, 14], as well as the use of the liver as a factory to flood the bloodstream with therapeutic proteins that are soluble in nature, such as metal-loproteinase inhibitors [15]. The use of first-generation adenoviruses in this setting is very limited due to the transient nature of gene expression, attributed to the high immunogenicity of the vector system [16]. For this reason, strategies to prolong adenovirus-mediated gene transfer as well as the use of other vector systems are being developed, with success.

Figure 2. New vectors for cardiovascular gene therapy. (A) First-generation adenoviruses are deleted in the E1 region with insertion of the expression cassette at that site (yellow). The remainder of the genome (grey) is from the adenovirus and is flanked by the inverted terminal repeats (red). Second-generation vectors contain additional modifications in the E2 (green) or E4 (blue) region to prolong transgene expression in vivo. Third-generation vectors are deleted of the entire adenovirus genome except the terminal repeats and essential sequences and contain “stuffer” DNA (pink) to ensure correct virus packaging. (B) Adeno-associated viruses have a 4.7-kB genome which, in recombinant form, contains the expression cassette and inverted terminal repeats (TR) only. (C) Lentiviruses have a complex genome, although the majority of genes can be deleted and provided as helper functions in trans, increasing their safety profile [19]. Recombinant lentiviruses are a promising gene therapy vector and require testing in applications pertinent to cardiovascular gene therapy.

Viral vectors with reduced immunogenicity
Modification of the adenovirus genome (Figure 2A) to reduce “leaky” expression of viral genes prolongs the expression profile of trans genes following systemic administration [17], but not following local application to the vessel wall [11]. Deletion of the entire adenoviral genome (Figure 2A), with the exception of essential packaging sequences, creates “helper-dependent” or “gutless” viruses that provide long-term gene expression in vivo [18]. Such systems have recently been used for long-term cardiovascular gene therapy studies in mice for reduction of atherosclerosis [14]. Aside from these encouraging recent developments, other viral vector systems may come to the fore in the future. Adeno-associated viruses and lentiviruses (Figure 2B, C) are less immunogenic than adenoviruses and offer the possibility of integrated, long-term gene expression, principally determined by elegant studies in other gene therapy disciplines [20–22]. Again, in similarity to adenoviruses, both adeno-associated viruses and lentiviruses transduce the liver upon systemic administration [20, 23]. These vectors, once developed appropriately for efficient and selective gene delivery to the cardiovascular system, may offer the ability to provide sustained correction of disease phenotypes.

The creation of “designer” viruses
Detailed knowledge of virus biology and infection pathways has led to the development of novel, disease-specific gene delivery vectors. This offers the opportunity to develop virus vectors that selectively transduce vascular cells, such as vascular smooth muscle, myocardium, or the vascular endothelium.
For the latter, studies in mice and humans demonstrate the distinct heterogeneity of the vascular endothelium whereby vascular addresses (ie, cell surface receptors specific for the endothelium of individual organs) can be exploited for targeted therapeutics [24, 25].
This has important implications for the development of “organ homing” gene therapy vectors. Such designer viruses can be developed by two approaches — transductional and transcriptional targeting.

Transductional targeting
As critical amino acids involved in the adenovirus-adenovirus receptor interaction have been identified [26], it has been possible to mutate the virus fiber protein to abolish this interaction (virus “detargeting”). This alone would merely reduce the infectivity of all cell types. However, within the terminal knob domain of the fiber protein, the HI loop, which is exposed at the fiber surface, can be modified to accommodate the insertion of targeting ligands. Simply, upon the identification of a small peptide sequence that can mediate binding to a cell surface receptor expressed highly and selectively on the target cell (eg, vascular endothelial cell or cardiac myocyte), it can be inserted into the virus fiber to “retarget” virus binding through that receptor (Figure 3). Adenoviruses can also be retargeted to alternative receptors by “pseudotyping” and by the use of “bispecific” antibodies (Figure 3).

Figure 3. Construction of designer viruses. Adenoviruses bind to coxsackie and adenovirus receptor (brown receptor), expressed highly on nontarget tissue, such as hepatocytes in the liver, but at a lower level on cardiovascular tissue including endothelial cells, smooth muscle cells, and myocytes. Adenovirus tropism can be modified by three main methods: (A) new targeting ligands (eg, small peptides, green) can be genetically inserted into the adenovirus knob to target it to a new receptor expressed selectively on target cells (green); (B) pseudotyping involves the replacement of the entire adenovirus knob region (brown) with the knob domain of another adenovirus serotype (yellow) that has a receptor with a more favorable distribution for gene transfer to cardiovascular cells; (C) the knob domain can be incubated with a bispecific antibody to block adenovirus binding to its receptor and redirect it to an alternate receptor targeted by the second component of the bispecific antibody (blue).

As the human genome project filters through to proteomics, the expression profile of cell surface receptors on normal and diseased tissue will be documented in detail and, when combined with vascular mapping by phage display [25], will define a repertoire of targeting opportunities. Viruses targeted through these receptors can then be assembled. Thus far, all three methods have proven successful for retargeting adenoviruses to the vascular endothelium [27–29]. There is no reason why selective adenovirus vectors for efficient gene delivery to smooth muscle cells or the myocardium cannot be constructed in the same ways and utilized for local gene therapy.

Transcriptional targeting
The vast majority of preclinical cardiovascular gene therapy studies have expressed transgenes from strong viral promoters such as the cytomegalovirus promoter. Such promoters, although very active, are nonselective and sensitive to silencing in vivo [30]. Therefore, in addition to transductional targeting, transcriptional targeting using cell-specific promoters can add selectivity to the target tissue without compromising efficiency. Efficient promoters that maintain activity and selectivity when engineered into gene therapy vectors exist for vascular smooth muscle cells [31], endothelium [32, 33], and cardiac tissue [34]. It is important to note that the efficiency and selectivity of candidate promoters can be altered unpredictably when engineered into the complex genome of viral vectors and must be rigorously evaluated [35]. Alternatively, it is possible to switch transgene expression on or off using regulated promoters such as the tetracycline system [36], a highly desirable safety feature for future in vivo gene therapy protocols.

A combined targeting approach
In theory, a combined transductional and transcriptional approach should produce highly effective and selective vascular gene therapy vectors, highlighted by a recent landmark study [37]. Systemic administration of an adenovirus targeted through the endothelial cell-specific angiotensin-converting enzyme, combined with transgene expression from the FLT-1 promoter, produced a remarkable 300 000-fold shift in gene expression from the liver to the lung (Figure 4) [37].

Figure 4. Combined vascular cell targeting. (A) A bispecific antibody (pink) that binds to the angiotensin-converting enzyme (ACE), as well as to the adenovirus fiber knob (to neutralize binding to its receptor), was incubated with an adenovirus vector expressing a reporter gene from the constitutive cytomegalovirus (CMV) or endothelial cell-selective FLT-1 promoter (B). (C) Analysis of gene transfer following systemic administration to rats demonstrated highly efficient and selective gene transfer to the pulmonary endothelium using the anti-ACE antibody and FLT-1-mediated transcription demonstrated by endothelial staining (green) with nuclei counterstained with DAPI (blue). As shown, the CMV promoter mediates gene transcription in the liver and spleen (green), a phenotype that is not present using the FLT-1 promoter. Adapted from [37].

This now enables the use of gene therapy as an approach to pulmonary vascular disease such as primary pulmonary hypertension. Similar strategies to target the vasculature of other organs pertinent to cardiovascular disease including the heart, kidneys, and brain may now be possible provided appropriate targeting systems are developed.

Concluding remarks
Although in the early 1990s gene therapists promised to rapidly deliver effective therapies for a host of diseases, that goal has yet to be realized. In other disciplines, children have had genetic defects corrected by gene therapy [38]. Recent results in vascular gene therapy are encouraging. However, at this time, cardiovascular gene therapy needs to be further developed and this can only be achieved through refinement, optimization, and stringent safety testing of therapeutic gene delivery vectors.

Acknowledgments
I would like to thank Campbell Nicol and Stuart Nicklin (University of Glasgow) for the adenovirus fluorescent microscopy and critical reading of this review, respectively.

REFERENCES

1: Science. 1990 Sep 14;249(4974):1285-8.