Gene Therapy (2003) 10, 946–954 & 2003 Nature Publishing Group All rights reserved 0969-7128/03 $25.00 www.nature.com/gt REVIEW Inflammation and adaptive immune responses to adenoviral vectors injected into the brain: peculiarities, mechanisms, and consequences PR Lowenstein1,2 and MG Castro1,2 1Gene Therapeutics Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, USA; and 2Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA, USA Gene Therapy (2003) 10, 946–954. doi:10.1038/sj.gt.3302048 It has been repeatedly shown – first by Shirai (1921) and stereotyped responses, such as cytokine release, and Murphy (1926), and most recently by Tansley (1946) – that influx of inflammatory nonantigen-specific cells. Once foreign homologous tissues grafted to the brain either do not stimulated, innate immunity’s effector arm is repre- provoke an immunity reaction or, if they do, do not respond to sented by the activation of local astrocytes and microglial it. It will be shown here that they do respond to an immune cells, and the influx of macrophages, NK-cells and state of proved effectiveness called forth by a preliminary dendritic cells. In systemic organs, such as the skin or grafting of foreign tissue elsewhere. The two questions, relating liver, professional antigen-presenting cells (ie dendritic as they do to quite distinct immunological properties, may thus cells) will carry antigens to draining lymph nodes, where have contrary answers. (taken from the introduction) priming of naı¨ve T cells occurs. Toll receptors, located on Skin homografts transplanted to the brains of specifically cells interacting with invading viruses or bacteria, can immunized rabbits respond by total breakdown to an immune recognize pathogenic epitopes embedded within the state already in being. On the testimony of other workers, they viral capsids, or bacterial membranes themselves. Thus, enjoy prolonged or indefinite survival when grafted to the viral vector capsids can per se elicit innate immune brains of non-immunized animals. responses, and thus cause inflammation, stimulating an It is concluded that skin homografts transplanted to the immune response.3 brain submit to but cannot elicit an immune state. (taken from Primed T cells divide and exit the lymph nodes as the final summary) activated or armed T cells and search systemic organs for the presence of the specific antigenic epitopes recognized Peter Medawar, 194816 by their unique T-cell receptor. Activated T cells, CD4+ and CD8+, will recognize viral antigenic epitopes on MHC molecules expressed by target tissues, and, upon antigen recognition, release proinflammatory cytokines, Introduction such as IL-1, IL-6, and TNFa. CD8+ cells can have a Most gene therapy vectors used for gene transfer into the cytotoxic phenotype (CTL); CTLs can kill target cells CNS are derived from pathogenic viruses.1 Therefore, infected with viruses, and express antigenic epitopes inflammation and immune responses are one of the through a number of cytotoxic effectors, for example, challenges facing successful and long-term gene therapy. perforin or Fas-L. Alternatively, activated T cells can eliminate virus replication using noncytolytic mechan- Upon injection into the brain, viral vectors will encounter 4 various cell types, for example, specific tissue cells (for isms, for example, through the secretion of IFNg. Virus example, neurons and glial cells), vascular endothelial interactions with B cells lead to virus-specific and high cells, connective tissue fibroblasts, and local (for exam- titer anti-viral antibodies that recognize non-linear ple, perivascular macrophages and microglial cells) and/ antigenic epitopes, and are thus able to detect intact or systemic circulating immune cells. These early viral capsids. Binding of antibodies to viral capsids interactions lead to a complex sequence of molecular activates a series of effector mechanisms that sequester and cellular responses that trigger innate and eventually virus from the circulation. adaptive immune responses.2 Through cellular and humoral mechanisms, the Innate immune responses are elicited almost immedi- immune system constrains virus infection and replica- ately following vector–cell interactions, and consist of tion. Similar mechanisms will constrain viral vector- mediated transduction of target cells. An important Correspondence: Dr PR Lowenstein, Gene Therapeutics Research Institute, difference between viral vectors and pathogenic viruses Cedars-Sinai Medical Center, Davis Building, Research Pavilion, Room is in the number of initial virus particles entering the R5090, 8700 Beverly Blvd, Los Angeles, CA 90048, USA organism. Low amounts of pathogenic viruses cause Inflammation and adaptive immune responses to adenoviral vectors PR Lowenstein and MG Castro 947 disease, since they rely on intracellular multiplication. inoculations were rapidly rejected. This suggested that Viral vectors that do not replicate are usually adminis- immune responses in the brain were blunted. This issue tered at high doses. This leads locally to high multiplicity was addressed again in 1948, when Peter Medawar16 of infection, causing potentially different inflammatory enquired if extended transplant survival in the brain was and immune responses. because of failure to activate the immune system, or That, in spite of the immune system, viruses can cause whether the effector arm of the immune system was life-threatening diseases indicate that viruses and the impaired in the brain. immune system must establish a delicate equilibrium, From a series of complex, yet elegant, technically where neither virus infection nor immune responses are challenging experiments, Peter Medawar concluded that completely efficient. Viruses may have defective infec- brain immune privilege is because of immune ignorance, tious pathways or the immune system may be unable to or failure to activate the immune system. His experi- eliminate rapidly and completely virus infections, or ments demonstrated that preceding peripheral immuni- viruses may have evolved strategies to overcome zation against skin transplants eliminated these from the immune detection, and the immune system’s sophisti- brain. This demonstrated that systemic immune re- cated mechanisms to detect virus infection.5–8 Examples sponses were fully active against antigen placed into of these strategies are downregulation of MHC mole- the brain. Of relevance to the present discussion, cules, inhibition of antigen-presenting pathways, and Medawar presented evidence that activated cells from secretion of binding proteins for cytokines and chemo- the effector arm of the immune system could detect and kines, among others. To these challenges, the immune eliminate foreign cells within the CNS. The lack of brain system has developed mechanisms to block virus entry lymphatics was suggested to underlie the inability of to the body, a set of innate immune responses that have transplants into the brain to elicit immune responses. evolved throughout time and are fixed in the genome Peter Medawar’s16 experiments ushered the modern that can recognize specific or shared pathogen features, exploration of the peculiar brain immune responses, and an adaptive immune system that can rapidly evolve referred to as the brain’s immune privilege. This review in the face of pathogen infection using mutation and of the immune responses against viral vectors will be adaptive molecular mechanisms not engraved in genome made in the context of what is commonly refered to as structure.9–11 The removal of most if not all virally the brain’s immune privilege.12 encoded proteins from the vector genomes also elim- inates those viral functions, which, in the context of viral infection, allow viruses to escape deleterious immune Brain inflammation upon injection of viral responses. On the other side, once the target cell is vectors infected, viral vectors will express few, if any, viral antigenic proteins. Although several virus-derived vectors have been shown Viruses used in gene therapy have been adapted for to cause brain inflammation, most information is avail- safe, nonpathogenic gene transfer. To convert a patho- able on adenoviral vector-induced inflammation. 17–19 genic virus into a vector, two minimum objectives need Cytokine20 and cellular21 inflammatory reactions are to be achieved. One, virus pathogenicity must be elicited upon injection of first-generation adenoviral eliminated (eg viral functions encoding for replication, vectors. This is reflected in a release of the cytokines inhibition of cellular funcitons, etc, must be removed), IL-1, IL-6, TNFa, and the stimulation of a fever and the potentially therapeutic transgene and transcrip- response,20 when adenovirus is injected into the third tional control elements cloned into the vectors. Details of ventricle. Cytokine levels increase very rapidly, with a systemic immune responses to different viral vectors are peak observed at 1.5 h after virus injection, and a time discussed further in other reviews in this issue. course similar to the fever response, returning to normal within 24 h. Injection of adenovirus into the striatum elicits an increase of IL-1 and IL-6, but not TNFa or fever. How do immune responses in the brain differ Moreover, the fever response upon injection of adenovir- from those in other organs? al vectors into the lateral ventricle is blocked by antagonists of IL-1, demonstrating that the fever Immune responses to viral vectors injected directly into response