At A Glance
Name: John Pezacki
Position: Research officer, chemical biology, Steacie Institute for Molecular Sciences, National Research Council Canada
Background: Visiting scientist, Scripps Research Institute, 2001-2002; Postdoc, chemistry, University of Toronto, 1998-2002; PhD, chemistry, McMaster University, 1998
At A Glance
Name: Angela Tonary
Position: Research associate, Biomolecular Sensing and Imaging Program, Steacie Institute for Molecular Sciences, NRC
Background: Research scientist, project leader, and manager of in vitro biology (successive), Adherex Technologies, 1999-2002; Assistant coordinator, Let's Talk Science Program, University of Ottawa, 1997-1999; PhD, physiology, University of Ottawa, 2001
Hepatitis C remains one of the biggest scourges in public health, yet there is a relative dearth of reliable screening methods for anti-HCV compounds. This is primarily because HCV is highly sensitive to the host cell state, and therefore difficult to model in vitro.
Recent methods based on sub-genomic HCV replicons have helped researchers face this challenge, but they are still striving to create a more relevant disease model for HCV infection without having to resort to small-animal research.
Angela Tonary and John Pezacki of the Steacie Institute for Molecular Sciences of the National Research Council Canada are two such researchers. The pair recently developed a screening assay based on normalizing luciferase reporter activity based on cell number in 2-D and 3-D HCV replicon cell cultures, work that is detailed in the March 15 issue of Analytical Biochemistry. Last week Tonary and Pezacki took a few moments to discuss their research with Cell-Based Assay News.
What was the reason you developed this assay?
JP: Our lab is interested in developing new tools for understanding host-virus interactions in hepatitis C virus. HCV is a global health problem there are no vaccines, and only partially successful therapeutic approaches to the disease. HCV is a liver-borne pathogen, and is very sensitive to local environment. In fact, an in vitro infectious model has been developed only recently, because HCV requires an in vivo environment in order to replicate and become infectious.
It is very difficult to culture HCV in [existing hepatoma] cell lines, so we wanted to try and develop this system that modeled the in vivo conditions better, so we moved to three dimensions.
How do subgenomic HCV replicons work in this type of research?
JP: The HCV replicon was developed in the lat 1990s in Germany by Ralf Bartenschlager, and was improved by Charles Rice, who's at Rockefeller University now. They were developed as models for viral replication in cells. They're subgenomic pieces of RNA that can be propagated in cell culture, and the portions of the genome contain all the functional proteins that are required for replication. In fact, the RNA, once it gets into a suitable host, can self-replicate, like a little molecular machine, if you will. That produces the proteins that are required for it to replicate, and then it copies itself using the materials of the host cell.
The subgenomic replicons that are the most popular and widely used for studying host-virus interactions in HCV have the NS3 to NS5B portion of the HCV genome. NS3 is the major pharmaceutical target it's the viral protease; and the NS5B is the other major pharmaceutical target, which is the RNA-dependent RNA polymerase. These models contain those functional proteins.
What are the typical current methods for HCV drug screening?
JP: We're not a pharmaceutical company, we're a government lab, but certainly from my limited knowledge, approaches vary from in vitro assays large-scale screens against a protein target in vitro, and then validate and further study those targets in replicons, small-animal, and finally large-animal models in disease.
What is the importance of conducting this assay in 3-D cell cultures if you end up conducting end-point assays on live cells?
AT: From the most general perspective, the importance of extending it to 3-D is that the 3-D cell culture is much more representative of the in vivo state of that tissue. The 3-D cell culture more closely mimics physiologically what would be happening in those cells that make up the tissue. So regardless of the fact that at our end point we are actually lysing the cells, prior to lysing them we have more accurately modeled the tissue in that 3-D culture.
Are 3-D cell cultures as disease models relatively new?
JP: It's still a frontier, I would say, in the field of cell biology.
Did your group develop the 3-D cell culture system for this assay?
AT: This particular 3-D cell culture system, which used a roller bottle apparatus, is not something that we developed. The roller bottle apparatus is commercially available, and 3-D spheroids generated by this apparatus are not new in and of themselves. It's simply an existing model that we used as part of the assay development.
So your innovation was a way to measure cellular DNA content and luciferase-HCV replicon expression in the same assay?
AT: Correct.
What is new about that compared to what's been done in the past?
AT: To my knowledge, what was quite unique about this assay was that very fact that from the same single lysate, you can determine two very important endpoints, which are luciferase activity and DNA content, without having to do two separate assays. That way we were able to calculate back to determine actual cell number based on DNA content, so that we could normalize luciferase expression in those model systems based on cell number.
JP: Culturing cells in 3-D is not a trivial matter compared to 2-D cultures in monolayers. We can get heterogeneous mixtures of 3-D spheroids that might have widely varied cell number, and if we want to make some sort of judgment about the levels of HCV replicon RNA that are present in one 3-D culture versus another, we need to be able to figure out how many cells we're dealing with and have some way to normalize the data, without costs going through the roof. Part of this is also driving toward something that we will ultimately be able to format in medium to high throughput.
What are the challenges of achieving a higher throughput format for this assay?
JP: Part of the challenge is technological, and we currently have projects to try and develop new formats whereby we can reproducibly culture cells in three dimensions in a high-throughput format. That roller bottle is not inherently a high-throughput technology, but if you were to imagine a miniaturized version of that sort of setup in 96- or 384-'roller bottle' format, then you would have the ability to start thinking about real drug screening.
Have you thought about using high-content or automated imaging for this assay? Would this type of approach be possible with 3-D cell culture?
JP: It is possible, and of course we have some collaboration with prominent people involved in molecular imaging. It would certainly be useful to be able to look at phenotypic changes in 3-D; morphological changes are something that is an easy endpoint in image-based assays. Again, however, there is a technological challenge there to be able to link the two formats. There are some basic requirements for growing cells in 3-D that are incompatible with current high-throughput microscopes. There are still technological challenges, I think, with being able to do that meaningfully, but certainly it's something that we're interested in.
How about live-cell assays as opposed to fixed end-point assays? Is this something that would be advantageous in any way? What are the challenges?
JP: The problem with live-cell assays is understanding how to normalize that data. There are secretion-based assays that have come out of Stan Lemon's group [at the University of] Texas that are very effective high-throughput replicon-based live-cell screening tools. However, in the context of a 3-D model of a tissue, you need to be able to figure out how many cells had replicon RNA to begin with, and how many cells you had per spheroid, and that's not a trivial question to answer. So in principle, yes you can do that, but you still need to be able to normalize that data. You could do imaging and try to guess how many cells there are based on the size of the spheroid, but that will only give you semi-quantitative information.
What are the next steps in the near future, and several months down the road, for developing this assay?
AT: For me, the development of this assay has been extremely useful because I can use it for other experiments, as well. In terms of further development, I am involved with the ongoing effort of miniaturizing these assays, and being able to, in general, develop the models into a high-throughput format. Moving this forward depends very much on being able to accurately develop these miniaturized liver models.
JP: This is a very difficult problem. We're taking small steps toward better models, but of course, hepatocytes and primary liver cells in general are the most difficult to work with of all cell types.
What are the major difficulties?
AT: The major difficulty is that they have very short-lived viability and functionality once they are removed from the body. It takes a lot of very careful tweaking of your cell culture conditions to not only keep the cells alive, but to keep them functioning as primary hepatocytes. Those are ongoing challenges for all researchers working with these cells, but especially for us trying to develop very liver-specific models.
Does this assay have commercial potential?
JP: This is a step toward something that could be quite commercially relevant. The closer we get to being able to model a real liver and the in vivo conditions, the more relevant this assay will become.
Can this method be extended to other disease models?
JP: 3-D cultures are most convenient for studying cancer. In principle, the approach that we presented here could be used for screening anti-mitotics and other anti-cancer small molecules, using an appropriate luciferase reporter system, or based on morphology.