At A Glance
Name: Raffi Manoukian
Position: Research associate, flow cytometry core facility, Amgen, since 1999
Background: Clinical flow cytometry facility/clinical immunology labs, Children's Hospital Boston; MSc, anatomy and cellular biology, McGill University, 1993
GENEVA — Straight-up flow cytometry has been around for several decades, and has always provided scientists with a "high-content" method for analyzing cells — in other words, multiple parameters are measured — but it has always lacked the imaging aspect that would allow researchers to better understand just what exactly is happening inside of cells. Actual high-content screening — automated, multiplexed imaging of cells on well plates or microscope slide — can provide some of these answers, but also suffers from limitations, mostly the speed with which the images can be analyzed and the resulting data used.
So it was only a matter of time before vendors seized the opportunity to combine the strengths of these two technologies into one platform. Companies such as CompuCyte and TTP Labtech have attempted to tackle the problem by designing so-called laser scanning cytometers, which provide the speed of a flow cytometer, along with rudimentary imaging prowess, on a static plate or slide. Others, such as Amnis, have developed actual flow cytometers that can take images.
Biotech giant Amgen has invested in both a CompuCyte iCyte imaging cytometer and an Amnis ImageStream 100 imaging flow cytometer, and Amgen scientist Raffi Manoukian works in the core cytometry lab responsible for both instruments. At the Society for Biomolecular Screening conference held in Geneva, Switzerland, last week, Manoukian gave a presentation about how he and his colleagues at Amgen are using the technologies separately and in conjunction to conduct high-content secondary-type screening of drug candidates. After the presentation, he sat down with CBA News to further discuss his work.
How did you develop your interest in flow cytometry and imaging?
I started doing a lot of cell-based assays while earning my master's at McGill, and I did a lot of electron microscopy, so I guess that was a high-content approach without me even knowing it. So right there I had an interest in imaging. And then, going from one lab to another early in my career, I became interested in flow cytometry, the power of flow, and how it can be applied to all kinds of research applications. That eventually led me to pharma, where I was exposed to a wide variety of technologies, and my love for imaging was renewed when we had the opportunity to do laser-scanning confocal microscopy.
How do you incorporate imaging cytometry in your research at Amgen?
We're a core service lab, so we serve all aspects of research and development. A lot of our assays come from a therapeutic area, so we have oncology, inflammation, neurology, metabolic disorders, and other therapeutic areas; also, we do lab work for toxicology and pathologies. But most of it is for specific therapeutic areas, and specific assays that they need done.
What types of cellular assays do you most frequently run using this approach?
We do a lot of the nuclear translocation assays that you've seen at this show. Also, we do a lot of cell-cycle assays, because there are a lot of compounds that affect cell cycle. Tumors are basically uncontrollably dividing cells, so cell cycle is a really important feature, and this is a strong platform on which to study it. We do a lot of phenotyping, which is important for inflammation, and [it is] combined with intracellular labeling of cytokines, and other intracellular proteins — so again, we do this multi-parametric approach and get a lot of information from flow and imaging cytometry.
If you didn't use this approach, what are some of the ways you might go about obtaining similar data?
We'd be doing manual scanning. We have other microscopes that are very basic, but are high quality — epifluorescence microscopes with CCD cameras on them, and we'd basically be taking snapshots of the images, bringing it on to some kind of image-analysis software, and this would have to be done well-by-well, field-by-field, so it would take an incredible amount of time and resources. We don't have that luxury.
What is the importance of image analysis and informatics in the flow-imaging approach?
Informatics is very important, probably more in the sense of data storage. I'm not involved too much in how the algorithms are written or how the software is designed. We'll definitely give our suggestions to the vendors for that. But the main concern is data storage, and our in-house IT people usually have solutions for that in terms of saving data on terabyte servers and what not.
Amnis, whose flow imaging cytometry platform you use in your work, didn't actually come to this show … they had recently said that pure screening applications may not be the strongest application area for the instrument (see CBA News, 9/5/2005). What do you think about that?
It depends on what you mean by screening, I think. You have your high-throughput screening, but the type of screening we do is more like secondary bioassays. So after the large hundreds of thousands of compound libraries are screened using more binding-type assays, then you take those lead compounds and put them into specific assays, and screen them based on functionality. So I think they still have a place in that area, and if you design your experiment carefully, you can get a lot of information, especially with cytometric imaging.
Are there any other advances on the horizon, or that you would like to see in the area of combining flow and imaging, or either flow or imaging alone?
Not that I know of. I think that quantitative imaging is the newest and hottest thing. A lot of attention is being given to magnetic resonance imaging, especially with the recent popularity of biomarkers. MRI has been done to follow biomarkers in clinical trials and patients, and maybe we can bring that technology in, and get even more information. There might be a day where they have miniaturized MRI machines that can scan an entire plate in something like 10 seconds, and then get a three-dimensional and colorimetric analysis. So who knows, anything's possible.