Background

According to the “Consensus document on European brain research” (Olesen et al, J Neurol Neurosurg Psychiatry 77, 2006) the socio-economic burden of diseases affecting the human brain is estimated to constitute 35% of all EU disease burden. Demographic changes in ageing societies of the EU will increase this rate considerably and this will represent a crucial challenge to forthcoming generations. Fortunately, enormous progress has been made during the last decade in understanding mechanisms and principles of both normal brain function and CNS disorders. This increase in knowledge makes new treatments for nervous system diseases realistic. Unfortunately, current pharmacological treatment strategies for neurodegenerative diseases are at best partially effective. Typically, they offer only some symptomatic relief. Generally, they are not able to affect disease progression. Furthermore, pharmacotherapies are limited by restrictions set by the blood-brain-barrier and by their relatively untargeted mode of action. As a consequence, they often provoke side effects, especially following long-term treatment. Thus, the targeted delivery of protective, or even curative genes to specific populations of disease-affected brain cells is an exciting future alternative. This approach may offer significant relief or even cure to presently untreatable or only symptomatically treatable brain diseases. Several proteins and regulatory RNAs involved in normal brain function and disease progression have been identified or will be identified in the near future. Targeting these molecules directly rather than developing drugs that modulate them may drastically reduce cost and time efforts in developing medications as compared to classical pharmacology. Ultimately, targeted and tightly controlled delivery of genetic medicines may open up venues for treatment strategies taking into account specific personal requirements of individual patients.

The objectives of the NEUGENE consortium are thus to develop and to validate the tools necessary for safe, efficient, durable, targeted and regulated expression of therapeutic molecules within the central nervous system (CNS). We will make use of the most advanced viral vectors for CNS applications and integrate new features for cell-type specific, regulated and sustained transgene expression to allow the use of such vectors for gene therapy of diverse brain disorders in human patients. These developments will focus on the special features and requirements of the CNS as a target tissue:
   
i) The CNS is composed of several types of cells with different functions in a complex three-dimensional structure (astroglia, oligodendroglia, microglia and many different types of neurons). Delivery of curative genes or regulatory RNAs critically depends on appropriate targeting of gene transfer vectors to specific CNS cell types. For example, astrocytes may be important targets for expression of secreted neurotrophic factors, but microglial cells should not be transduced as they may eventually present antigenic peptides and/or initiate immune reactions. Neurotransmitters, receptors or signal transducers used for therapeutic purpose should only be expressed in neurons affected by a given disease, as their expression in other types of neurons may interfere with normal functions and induce side effects.
   ii) Most neurons of the CNS are terminally differentiated cells that normally are not replaced. Thus, cytotoxicity of gene transfer vectors can not be tolerated. Furthermore, as a consequence of lack of division there is little turnover of neuronal cells in the brain and therefore transduced cells may retain the transferred genetic information for the life-span of the patient. Therefore, in case of non-tolerable side effects a regulated system in which transgene expression could be tightly regulated would be highly beneficial. Equally important, demands for expression levels of e.g. neurotrophic factors or neurotransmitters may vary drastically in different stages of a brain disease, depending on disease progression or progress of curative intervention. In terms of both safety and therapeutic efficacy, it is therefore mandatory to have tight control over transgene expression levels by making use of regulated transgene expression systems.
   iii) Gene transfer vectors based on recombinant viruses are the most effective vehicles for gene therapy to date, but as they are partially composed of proteins not recognized as “self” by the mammalian organism, there is the potential risk to provoke immunological responses. Immune reactions in the brain even more than those in peripheral organs, are sometimes difficult to control and thus require special attention. On the other hand, the relative immune privilege of the brain may allow for modes of vector application not possible in peripheral organs. As an important safety issue CNS tissue reactions like activation of microglia and production of (pro)-inflammatory cytokines and astrogliosis must be critically evaluated especially for novel vectors with enhanced transduction properties.

   A large number of viral and non-viral gene delivery tools have been developed for CNS applications during the past decade (for review e.g. de Lima et al, Curr Drug Targets CNS Neurol Disord, 2005). Non-viral systems are generally hampered by relatively inefficient transfection and short-term expression of transgenes. Viral vectors ensure generally more efficient transduction but their transgene expression may also be limited. For example, early generation adenoviral vectors provoked strong immune responses after CNS transduction, due to expression of residual adenoviral genes. This problem was diminished with the development of “gutless” adenoviral vectors, which do not contain any residual Ad genes, and use of non-human Ad vectors (Perreau and Kremer, Mol Biotechnol, 2006), but the complex capsid of these vectors and the possible prevalence of pre-existing immunity remain as potential risks for recognition by the immune system. Naturally neurotrophic viruses like HSV have been engineered into efficient CNS gene transfer tools, but they are still associated with some cytotoxicity, are relatively complex to produce and may interact significantly with the host immune system (Bowers et al, Gene Ther. 2003). A number of vectors based on other viruses (Rabies virus, SV40) are currently under development but it is not clear whether they will match the efficiency of the current vectors of choice for the CNS: adeno-associated (AAV) and lentiviral (LV) vectors.
   Viral vectors based on AAV are currently considered as the safest viral gene delivery systems, since wild-type AAVs are not associated with any known human or animal disease, and recombinant vectors persist predominantly as episomes without known cytotoxicity. AAV-2 based vectors have been used in clinical trials including 6 trials in the CNS to demonstrate safety in patients suffering from Parkinson's, Alzheimer’s, Batten or Canavan┤s disease. AAV vectors do not provoke significant immune reactions upon administration into experimental animals immunologically na´ve to the vector. However, in a clinical trial aimed at correcting factor IX deficiency (haemophilia), pre-existing immunity to the wild-type virus, which is acquired during childhood in the majority of humans, has led to treatment failure due to destruction of transduced hepatocytes (Manno et al, 2006). Limited data are available from clinical trials regarding immunological response following AAV administration in the CNS. Open-label trials in the CNS have so far demonstrated no significant side effects and have led to some clinical improvement in patients, arguing for a potentially advantageous immunological milieu in the CNS as compared to peripheral vector application, reminiscent of what has been observed in animal models (McPhee et al, 2006; Kaplitt et al, 2007). However, essential safety issues like intracerebral immune responses, regulation of transgene expression and other relevant aspects have not been addressed so far and are especially lacking for the novel serotypes of AAV, which might provide significantly enhanced transduction characteristics.

   Lentiviral vectors have recently challenged the dominant position of AAV vectors for CNS application. LV vectors are extremely efficient at transducing dividing and quiescent cells, leading to a high level and stable transgene expression, and lack cytotoxicity. While their titres are still 3-4 logs below those achievable with AAV vectors, they lead to comparable levels of gene expression upon injection of similar volumes in the CNS. This is a significant advantage considering the potential for side effects associated with high vector doses. These characteristics, together with their relatively simple production, have made them highly attractive and thus have led to widespread use for applications in various animal models and even clinical trials. However, LV vectors share a major drawback with other retroviral vectors: the potential for insertional mutagenesis associated with their natural integration in the host cell genome. While this may be irrelevant when transducing terminally differentiated cells in the CNS (e.g. neurons), the potential risk for glioma formation or tumours of other bystander cells cannot be altogether disregarded. Fortunately, one of the consortium members has recently demonstrated that non-integrating LV vectors are as efficient as their integrating counterparts for CNS transduction. These integration-deficient LV vectors, simply produced through the use of integrase mutations, become non-replicating episomal plasmids in transduced cells and lead to high level and stable transgene expression. Non-integrating LV are considered highly safe systems for CNS therapy for the following reasons: (i) similarly to standard LV, they are non-replicating and self-inactivating; (ii) they have a highly reduced risk of causing insertional mutagenesis; (iii) being non-replicating they do not support multiplication of a potential HIV-based recombinant virus; (iv) being episomal they would be diluted out if undesired dividing cell populations were transduced.