Directed evolution of biochemical pathways in vivo: hepatitis C virus (HCV) replication in the murine liver

A recurring theme in modern chemistry is the use of synthesis as a tool for understanding how natural systems work.  Bio-organic chemists have long employed de novo design and directed evolution of biological molecules to understand mechanism and to engineer new function, and we would like to use this approach to understand biochemical pathways in vivo. 

Optimally, the pathway targeted for directed evolution would be comprised by non-endogenous biochemical reactions that can be integrated into the normal biochemistry of the host cell.  In addition, the coordinated reactions that comprise this pathway should be encoded by an episome that can replicate in quiescent as well as dividing cells since most highly differentiated cells are non-dividing in vivo.  Viral replication in vivo is a process that meets these criteria and provides an elegant system in which to develop selection markers and other tools for directed evolution in vivo.  We are using hepatitis C virus (HCV) replication as a model system to develop methodologies that permit the directed evolution of non-endogenous biochemical reaction in a mammalian in vivo setting. The development of a murine model of HCV replication is an ambitious project that has not been readily solved by classical approaches in virology but is ideally suited to the techniques of in vivo evolution.  Likecess of HCV replication in many ways represents an ideal model system for the directed evolution of multi-reaction processes in vivo. 

This strategy is based on our hypothesis that HCV clones derived from human infections fail to replicate in the mouse liver because they are unable to interact productively with the host cellular machinery required for replication and/or because they are unable to evade an antiviral response in the murine host.  If this is true, it should be possible to discover specific mutations that relieve these constraints when introduced into an HCV genome. 

In order to perform this directed evolution experiment truly in vivo, we are taking advantage of the phenomenon of therapeutic liver repopulation.  In murine models of liver repopulation, resident hepatocytes are engineered to be unhealthy due to genetic deletion of an essential metabolic enzyme or to overexpression of a toxic transgene.  Wild-type transplanted hepatocytes or resident hepatocytes in which the genetic abnormality has been corrected are “more fit” and grow in a nodular pattern until they have repopulated the organ. The original resident hepatocytes are “less fit” and are selected against during this process because they are diseased.   To exploit this phenomenon, we have chosen the fumarylacetoacetate hydrolase (FAH) gene as our in vivo selection marker.  FAH catalyzes the second to last step in tyrosine catabolism.  Mice homozygous for a targeted disruption of the Fah gene (FAH^-/- mice) accumulate cytotoxic levels of intracellular fumarylacetoacetate (FAA) and succinyl acetone (SA). 

If untreated, this leads to acute liver failure and death. FAH^-/- mice can be rescued by oral administration of Orfadin™, an inhibitor of homogentisate oxidase, which is upstream of FAH, or by repopulation of the liver by a starting population of as few as 1000 hepatocytes expressing functional FAH. We have generated HCV replicons bearing FAH expression cassettes (A). Delivery of this in vitro-transcribed RNA to the livers of FAH^-/- mice creates an in vivo system in which survival and proliferation of an individual hepatocyte is coupled to its propagation of the HCV replicon RNA driving expression of FAH (B).

We transfected FAH^-/- RAG1^-/- mice with in vitro-transcribed FAH-HCV replicon RNA and negative control neoR-HCV replicon RNA, and then applied selective pressure by withholding Orfadin™.  Encouragingly, a measurable survival advantage has been observed for animals transfected with FAH-HCV RNA versus those transfected with neoR-HCV RNA.  Moreover, FAH and NS5B are detected by RT-PCR in livers of animals transfected with FAH-HCV RNA but not neoR-HCV RNA. 

Given the half-lives of HCV virions in vivo and neoR-HCV replicon RNA in vitro, it is highly unlikely that the FAH RNA signal we detect is due to the input RNA; rather, it more likely reflects newly synthesized HCV RNA. 

We are now in the process of performing additional RT-PCR and immunohistochemical analysis of liver tissues taken from FAH-HCV RNA-transfected animals(“experimental”) and negative control animals (“control”) for the presence of FAH and HCV species.