Skip to main content
Premium Trial:

Request an Annual Quote

Yale s Rimm on Improving Tissue Microarray Technology and Analysis

David Rimm
Yale Cancer Center Tissue Microarray Facility

Name: David Rimm

Title: Director, Yale Cancer Center Tissue Microarray Facility

Education: 1989 — MD, PhD, cell biology, Johns Hopkins School of Medicine; 1981 — BS, molecular biology, University of Wisconsin, Madison.

David Rimm is an associate professor in Yale University School of Medicine's pathology department and has been the director of the Yale Cancer Center's Tissue Microarray Facility since he founded it in 2000. He is also a founder and consultant for HistoRx, a Yale-spin off that has pioneered a protein expression analysis technique called Automated Quantitative Analysis, or AQUA.

Rimm and co-authors last month published a paper in the Journal of the National Cancer Institute that details how the concentration of antibodies in a protein expression experiment can affect the assay results, which in turn can skew the results of array-based diagnostic tests.

Entitled "Automated Quantitative Analysis of In Situ Protein Expression, Antibody Concentration, and Prognosis" and published on Dec. 21, 2005, the paper prompted a lead editorial in the journal that said the study would send researchers developing predictive diagnostics based on immunohistochemistry "back to the drawing board." To discuss the study at length and learn more about YCC's tissue array facility, BioArray News spoke with Rimm last month.

What kind of tools are being used on a regular basis at the Yale Cancer Center's Tissue Microarray Facility?

We've got a room full of equipment, but our tissue microarray facility does not provide any analysis capacity. What our facility does is we take [orders], make tissue microarrays, [and] give the arrays back to the client. We don't actually read them.

In our facility with have the standard histology tools, and then we have arraying devices. We have two conventional Beecher instrument manual arrayers. We have tested the Chemicon arrayer and we had that in our lab for awhile but we didn't like it; we gave it back to them. We also tested this NACC arrayer from the Nippon Automated Control Company. We also had the Automated Tissue Arrayer-27 from Beecher Instruments which is an automated tissue microarraying device, [but] … we just couldn't get it to work. So ultimately we convinced Beecher to trade us two manual ones and a semi-manual one. And they gave us a letter that said they would get us the other materials shortly and we are still waiting on that.

The bottom line is that, unless you are making arrays of liver from the same animal, the automated device may be good in toxicology, but we found that it was worthless for the kinds of tests we needed. So we just used the [manual] one. And so we make the arrays in that facility and then we cut the slides and give them to whoever is going to read them.

Who are your typical clients, then?

Eighty percent of our clients are Yale faculty. The other 20 percent are faculty from other academic centers and from the National Cancer Institute and the National Institutes of Health and then a smattering of private companies that want our services.

So we've licensed slides to a number of organizations, both public and private.

How did you get involved with HistoRx?

HistoRx is completely unrelated to the YCC Tissue Microarray Facility. HistoRx is the result of [work] that was going on in my research lab. Back in 2000, my research lab wanted to analyze tissue microarrays. So I went to my chairman and said, 'I want to make tissue microarrays for my research lab because that's the format we want to use for analysis.' He said, 'OK, if you are going to make tissue microarrays for your lab, then you are going to have to make them for everybody.' So I talked to the Yale Cancer Center and they seeded the facility, and so then I was running the facility as well as running my research lab, but I was instructed to make sure I kept the budget separate and so they are to this day two separate entities. In my research lab we focus on analysis of tissue microarrays that we buy just like anybody else from the tissue microarray facility.

The analysis is what my lab has been working on since about 1999 with the concept that finally we have a small enough piece of tissue that we can be totally quantitative about. A whole section was just too daunting to be quantitative about, but we can get our arms around a little 0.6 millimeter spot. And so we, in my lab, started trying five or six pieces of commercially available software. We couldn't get any of them to do it the way we thought it should be done.

And then one of the senior fellows in my lab, Robert Camp, said 'How about if I write the software?' So he started writing and we went back and forth and so the first version of AQUA was actually written in Visual Basic. I started talking about it with various people and they said they wanted it. I went to our office of cooperative research and they patented that, but instead of licensing it they decided to found a company based on it.

So the company was founded in September 2004 and it has done very well. I don't work for the company but I have decided to serve as a consultant. So I am a consultant for HistoRx and a shareholder and those are my disclosures.

Why did you decide to use AQUA on this project, and were you looking to solve this problem specifically or did you accidentally discover it?

Well, that's the cool thing. We weren't looking at it, we were just characterizing quantitative expression analysis and we were using AQUA to characterize it. One time we were doing the titering and we thought, 'Let's try an actual experiment with too high a titer,' because normally you use 1: 8,000 for Her2. And we used 1: 500, instead of 1: 8,000, because we thought, 'What happens if we use the wrong titer?' And we got the completely opposite result. And we thought, 'Wow, that's really striking that we can completely flip the survival curves by just changing the titer.' And that was back in 2002. This was a long time coming. It took a long time to figure out how to do the cell block standard curves and getting that standardization and being able to build those standard curves at the two different titrations took us years. It was much trickier than anyone thought it would be to get the cell line's protocols just right and to do all the analyses just right.

But that was the proof of our hypothesis that was generated by that experiment of using the wrong titer. We started with Her2, we used the wrong titer; we got a different result. We wanted to know how general this was. We had estrogen receptor in the lab and we tried it on that and it didn't happen. Then we tried it on p53 and it happened again, just like we saw for Her2. So then we realized that there was a general principle here that has to do with the dynamic range of the protein being expressed. And if you have a broad enough dynamic range — and it's not a linear relationship, it's a U-shaped relationship with outcome — then you are going to be able to see that with our technology. You probably wouldn't be able to see it with conventional technologies.

How does AQUA work?

AQUA is sort of an unusual set of analysis tools that are based on molecular interactions, instead of being based on shape and feature extraction. So far, all the conventional software for quantitative pathology has always been based on feature extraction. But we don't want to do that. And every compartment or target we look at is based on a molecular interaction, not a contrast-generating edge. We use fluorescence instead of a traditional pathologist's brown stain because fluorescence has a much broader dynamic range and it also can be used for multiplexing.

In light of the results of your study, what does it mean for companies that are developing diagnostics based on biomarkers?

A lot of those companies are now working with HistoRx because they realize the importance of being quantitative. And what this illustrates is that with the traditional ways pathologists do things, everything is sort of relative. There's no handle to hold onto to be truly quantitative and to ensure reproducibility. What this shows is that it is not about how good your pathologist is — it's about how good your assay is. And this sort of provides an anchor for the assay, something you can go back to. A standard curve just like we would use in chemistry and hematology, the things where there really is an anchor in laboratory medicine that's reproducible year in and year out and is really robust.

To me this is in some sense obvious. I mean, why wouldn't we require the same kind of quantitative rigor from anatomic pathology that we require from clinical pathology? In clinical pathology if we said we did an assay by grading it on a scale of 1, 2, and 3, that wouldn't be acceptable. But in anatomic pathology that was acceptable historically because it was so complex to do the quantitative analysis. Well now we've sort of broken through that complexity and we can provide quantitative analysis to the anatomical pathology world. I think, and clearly I'm biased, that in the future we are going to see all of anatomic pathology moving to a platform that allows the kind of reproducibility and stability and objectivity that you get in a laboratory medicine test.

Are there any competitive technologies?

Well, what we've seen is that there are a few companies that are starting to delve into fluorescence on the basis of the data that they've seen from us over the past couple of years. Depending upon how broad our patent coverage is …

Have you filed?

We've filed and they are going back and forth and they allowed some claims and not others and the attorneys were fighting for more claims and … patent law is something I find [I am] completely unable to interpret. Some things get patented even when there's clear evidence of those things in the literature and then it happens that patents are turned down when it seems like a very novel technology. So I can't pretend to understand it. The AQUA has a very broad patent filed for using this technology for co-localization. If we get that then there won't be many people playing in the fluorescence space. If we don't get that then maybe other people will play in the space as well.

Where do the arrays fit into the system?

The arrays are a convenient platform for doing AQUA. The arrays are not cDNA arrays, they are tissue arrays — there are just 0.6 mm spots of tissue.

So what's next for this technology?

What's in the pipeline is that we've used this technology to look at response to therapy to drugs. And that's why a lot of pharmaceutical companies are interested because it looks like we'll be able to use this technology to make companion diagnostics that wouldn't work in any other way. In fact, even for some of the current drugs that are out there, there may be a desire to use this technology instead of some of the current technologies. So one could envision the brown stain in pathology labs for any companion diagnostic where a drug is going to be prescribed on the basis of the presence or the amount of protein expression, like Herceptin or tamoxifen and estrogen receptor.

With the current technology for the HercepTest, you may be able to predict 40 percent responders. Well if this technology was applied you may be able to predict maybe 60 or 70 percent responders. And so the oncologists who are prescribing these drugs may be the driving forces behind the introduction of this test to the market. What is on the horizon is that HistoRx is in the process of making a second-generation box that would be placed into the pathologists' labs so that they would do the analysis themselves. They would just need the reagents. They would be able to put the reagents on their slide and then put it into the HistoRx box and then they could read some of these current tests or future tests.

File Attachments
The Scan

Wolf Howl Responses Offer Look at Vocal Behavior-Related Selection in Dogs

In dozens of domestic dogs listening to wolf vocalizations, researchers in Communication Biology see responses varying with age, sex, reproductive status, and a breed's evolutionary distance from wolves.

Facial Imaging-Based Genetic Diagnoses Appears to Get Boost With Three-Dimensional Approach

With data for more than 1,900 individuals affected by a range of genetic conditions, researchers compared facial phenotype-based diagnoses informed by 2D or 3D images in the European Journal of Human Genetics.

Survey Suggests Multigene Cancer Panel VUS Reporting May Vary Across Genetic Counselors

Investigators surveyed dozens of genetic counselors working in clinical or laboratory settings, uncovering attitudes around VUS reporting after multigene cancer panel testing in the Journal of Genetic Counseling.

Study Points to Tuberculosis Protection by Gaucher Disease Mutation

A mutation linked to Gaucher disease in the Ashkenazi Jewish population appears to boost Mycobacterium tuberculosis resistance in a zebrafish model of the lysosomal storage condition, a new PNAS study finds.