At A Glance
Name: Philippe Marchand
Position: Director of engineering, Genoptix
Background: Project director, OMM, 1999-2001; researcher, University of California, San Diego, 1991-1999; MS and PhD, electrical engineering, Universite de Haute, Alsace, France.
Name: William Butler
Position: Director of advanced research and development, Genoptix
Background: Scientific investigator, Nanogen; Various positions, Hybritech; BA, physics, Rice University; PhD, physics, University of California, San Diego.
Along with several of their colleagues at San Diego-based biotech Genoptix, Philippe Marchand and William Butler recently co-authored a paper in Nature Biotechnology entitled “Microfluidic sorting of mammalian cells by optical switching” [Nature Biotechnology. 2005 Jan; 23(1): 83-7]. In it, the researchers describe the technology underlying the Microsorter, a new microfluidic-based flow cytometry platform being developed at the company. The technology, which uses an infrared laser to “guide” cells down microfluidic channels, has a much lower throughput than most existing cell sorters, and is meant for smaller amounts of cells than traditional flow cytometers. Why would such an instrument be attractive? Marchand and Butler took a few moments last week to discuss the answer with Inside Bioassays.
Marchand: It is related to the patents we won. We have a few of the fundamental, basic technology patents that have been awarded. In one form or another, the basic technology patents do cover some of what’s in that instrument embodiment. We have another instrument platform developed here that also uses some of the same basic technology for a completely different application, a completely different embodiment. We have basic technology patents, and I don’t believe at this point that any of the more specific instrument patents have come out yet.
Compared with other microfluidic flow instruments on the market — [ones from] Beckman Coulter, Becton Dickinson, and Guava, for example — how is this new technology different?
Marchand: There is one fundamental thing. It is flow cytometry, and in flow cytometery, about two-thirds or so of the instruments out there are what’s called analytical instruments. You use a flow cytometer to do analysis of cells through fluorescence detection and things like that. But you don’t necessarily do sorting. A sorter such as this uses some of the same basic technology as in the BD or Beckman instruments, or [DakoCytomation’s] MoFlo, or a Partek instrument. Those are geared towards using the same analytical technique, and doing the analysis, but at the same time, those cells that actually match the proper criteria — get them out. Separate them and get that population out of the system, so you can do further testing. You can take those cells and do PCR, or gene analysis, or whatever it is you can do with a cell that’s still intact. The Guava instrument, as far as I know, are analytical instruments, and do not perform the sorting operation.
One of the points of the paper was that this technology maintained cell viability. Is this an issue with flow cytometry?
Marchand: It can be. It’s just a matter of the performance those systems offer. They go so fast, the cells are submitted to mechanical stress. They’re being pushed down at about 10 meters per second in those systems — that’s very fast — and there are therefore very large pressures being exerted on the cells. It’s not uncommon to see cells going through a system and being really physically modified in terms of their shape because of the stress that they’re under.
Butler: And there’s a lot of stress when they’re put in individual drops, and when you actually do the sorting, those drops stop very quickly at the end of their paths.
Marchand: They’re just being thrown down very quickly into a cup, where they land very hard. Because of the very nature of a typical sorting instrument, we expect that we would be much gentler. The thing that could have been a worry is because we’re subjecting the cells to a lot more laser exposure than other cytometers, because we have the laser switching mechanism in there, that there might have been heat-induced stress. But after doing the work, we realized there is essentially no heat-induced stress being generated.
Throughput-wise, how does this compare to cell sorting instruments that are currently available?
Marchand: The first answer is: It’s much slower. But the real answer is: It’s almost on purpose. The traditional sorters, the big sorting machines, can sort up to 10,000 cells per second, no problem, and in some cases, you can even go faster than that. However, you also need an awful lot of cells to be able to sort. You won’t be able to sort 10,000 cells on a BD FACS Array, for example. That just won’t work, because by time you’ve set up the instrument and start doing your work, the cells have already disappeared. So that’s the first thing. This instrument is geared towards small populations, and it’s a combination of design and engineering limits at the same time. It can’t go all that fast, but that’s OK, because it can deal with small populations.
What would be some examples of applications where sorting such small cell populations would be useful?
Marchand: Anything that’s in many cases primary tissues. Not blood, obviously, because you can get, without exaggeration, tens or hundreds of millions of cells from a blood sample. However, think about fine-needle aspirates from a tumor. Precious tissues that have been preserved, like paraffin-embedded tissues, is another. There are banks or hospitals that have large numbers of those things, but they’re very precious, because once you consume them, you don’t have them. Some of those are from patients that go back 15, 20, or 25 years. And if you could go back to those because you have the knowledge of the clinical outcome of those patients — unfortunately, in many cases, it’s when the patient died and what he died from and how he responded to therapy — but you have all that knowledge that’s been accumulated after you preserved that tissue. So you can go back to the original tissue, which is preserved in paraffin, get the cells out of it, and then do some sorting to be able to do further analysis to help you either develop new drugs, or new diagnosis techniques, or new markers. But those are very precious and come in small amounts, and are often difficult to deal with — they come dirty, which means there is a lot of extraneous material in there. Traditional flow cytometry systems do not like to have those in the systems, because there’s potential for clogging and contamination. With our system, you have a self-contained microfluidics cartridge, and once you put the sample on it, it doesn’t go anywhere else, and it’s a little easier to deal with. We’ve had pretty good success with primary samples like pancreatic cells, which we worked on with the University of Chicago. Nobody really wanted to sort those cells for them in any of the facilities, and we took that and managed to sort it with very good results.
Butler: Other possibilities — we mentioned that it’s not ideal for most blood analysis. However, if researchers are looking for rare cells in blood, or other things, there are some techniques that are used right now to isolate some populations of rare cells, but they’re usually not very pure at all. Our device would be very good as follow-on to that to help purify the first cancer population of rare cells, and further enrichment of those rare cells, so you can do analysis on them. Another application would be in early research laboratories — not at all in clinical applications — where they’re working with small populations of cells, perhaps many small populations, and it’s necessary to purify them. For instance, binding transfected cells in a population where transfection is never at all near 100 percent — in fact, it’s usually a very low percentage of cells that’s transfected — if you want to purify those out, this is an approach to doing that.
Has Genoptix been primarily focused on tools for cancer diagnostics and biomarker detection?
Marchand: Not exactly. The main revenue stream for the company is: We have a clinical lab, so we’re certified in California, and we’re also CAP [College of American Pathologists] certified, now. And what we offer there are standard flow services and marker services for diagnosis and/or detection. But we also offer tests where patient’s cells are tested against a panel of drugs — right now, we’re working on all kinds of leukemia — and we return the tests to the physicians and say: This patient seems to show more resistance to this drug than that drug. It’s part of the whole personalized medicine development.
How about drug discovery applications?
Marchand: That’s actually what we used as our benchmark testing biological model. We used GFP in HeLa cells, and we measured expression at that level. The direct corollary of that would be in bioprocessing applications, where you want to find and sort and select the highest expressers in a population, where you have vats and vats of those that you intend to do, but you must first select the best.
Butler: Just to clarify something regarding the company’s revenue stream and focus to this point — the Microsorter is not involved in those applications at all.
Is this product on the market yet?
Marchand: We have our alpha prototype, which has been integrated in a system, and it’s been beaten up from an engineering and design point of view, so right now we’re focusing on application development. Before we can go out and offer that and do the investment — which is quite costly for developing and manufacturing such an instrument — we have to really put the word out, and the best way is to demonstrate viable applications.
What are the major applications so far that you feel might really drive sales of this product?
Marchand: One is what we were talking about — transfected cells that you want to sort for highest expressers. Anything that has to do with samples that are not and can not be run on traditional cell sorters, for multiple reasons; that’s another one. And a third one, which is a corollary to the last one, is anything that has to do with bio-safe, or bio-unsafe particles, I should say. Because of the nature of the instrument, particles are confined on the cartridge. There is no exchange of fluids between that cartridge and the outside world. So you could load up the cartridges in your bio-safe environment, bring it to the instrument, run your sort, go back to your biosafe environment, and get the assorted populations out. This is much more difficult to do with a traditional flow sorter, which essentially has aerosols and droplets propagating in air. With traditional flow sorters, there are issues with biosafety. The way around is usually that the entire system is confined to a bio-safe environment. In this case, you don’t need that. It’s another small advantage, but when combined with the overall instrument, can lead to valuable applications.
Butler: In terms of other applications, we anticipate that the research community will identify further applications once their interest is aroused in this capability — one that hasn’t been available before. So defining the full set of applications is probably beyond our current capabilities, because we don’t recognize all of them. So we need some viable subsets of applications to start with.
I see you used a 488-nanometer laser for excitation, which makes sense for some of the more popular fluorescent markers that are used …
Marchand: That’s actually why we went with that. When we started the project, we actually used a green laser for internal testing because it was much easier and much cheaper. But a blue laser is really what you need if you’re going to target the more popular fluorescent tags.
Do you anticipate the need for more or different laser lines in the future?
Marchand: Well, if you have blue, you cover a much wider area of what’s out there. There are probably people that will say, ‘Well, I really need a red laser, because I’m doing this certain thing.’ But those people early on should be such a small subset of the commercial targets and customers that we’re going to have to forego it. It’s a matter of trading off pricing of the instrument with capabilities. Right now, the instrument has four-color detection — four fluorescent detection channels — with one blue excitation laser. For sorting, that seems to be what you need. When people do very extensive analysis in flow, they use all the capabilities — the 12 or 16 parameters that some of the more expensive instruments offer — but when you do sorting, you don’t really need so many parameters. You usually do two or three parameters.
Butler: There’s nothing that would prevent us from putting other lasers in, except for manufacturing difficulties, complexity, and the associated cost of that. So in the future, we might be able to offer multiple lines, but that would be after you’ve reached the point where you’re successfully marketing the product in the first place.
What is the anticipated price range of this instrument?
Marchand: We’ve been working on two versions of the instrument. One would be a lower-cost, lower-performance instrument — lower-performance in terms of throughput. It wouldn’t go much faster than 20 cells per second. That would be the target price in the range of above $50,000 and below $75,000. The advanced version — which probably people in R&D and life sciences would hopefully be more interested in — would probably be around $100,000. That would have throughput up to the numbers that we quoted in the paper — the 100 cells per second, the full four-color detection, et cetera.