The goal of the EuroBiochips meeting, which opened Tuesday at the Berlin Hilton, appeared similar to that of the German World Cup spectators who huddled around televisions in the hotel lobby: to advance to the next level.
Instead of the usual array of pitches for basic DNA chips, the first day of the meeting focused on how to improve sample preparation, and novel approaches to DNA and protein array chemistry and biochemistry.
Speakers in the sample preparation talks Tuesday morning emphasized how careful sample handling is a key prerequisite to good results both with DNA and protein arrays, and plugged the products and processes their companies are marketing in this direction. During the afternoon, another group of speakers — mostly from German companies and institutions — demonstrated their commitment to genomics over soccer (football) by staying in the conference room and delivering presentations on microarraying machines, peptide chips, PCR vs. oligo arrays, and the latest synthetic nucleotides. Following is a summary of key talks during the day.
Qiagen’s Remedy for Spoiled Samples
“Preparation, fractionation, stabilization, etcetera, is becoming a very important issue” in microarray experiments, said Qiagen researcher Uwe Oelm ller, in introducing his company’s RNA stabilization chemistry for sample preparation.
Oelm ller presented a future scenario in which a doctor’s office sends off a room-temperature blood sample to a diagnostic lab, and the sample degrades over the hours, or days, when it is in transit. “There is a clear risk that the sample will [undergo] big changes in gene expression,” he said. Qiagen and PreAnalytics have tested mouse kidney samples, he said, and have found that these changes in gene expression can happen in ten minutes after the sample is drawn.
The solution? Not surprisingly, Oelm ller recommended stabilizing and preparing the RNA immediately with a reagent such as the company’s RNALater.
Christian Korlhage, also of Qiagen, followed Oelm ller’s talk with a discussion of cleanup methods. The company, he announced, plans to introduce a new product, LabelStar, next week, which will remove non-incorporated oligonucleotides as well as other contaminating factors. The LabelStar array kit will also include a denaturing solution, and LabelStar RT, which is designed to destabilize the hydrogen bonds between secondary RNA structures and to reduce non-specific binding to RNA, he said.
Qiagen has found in its internal research that this kit reduces the level of false positives while retaining sensitivity, but these results have not been independently confirmed.
Ethical and Accurate Clinical Samples
Marc Reymond, a co-founder and the chief medical officer of Europroteome, followed up this sample prep plug with a warning about the underlying issues involved in collecting samples to begin with. “I strongly believe that it is of fundamental importance when doing business with human samples data ... to obtain the patients’ written informed consent,” he stated, noting that there has been an “explosion” lately in committees and other discussion bodies around Europe to set the rules for patient sample collection in post-genomic research. “You can bet that within three to five years we will have guidelines on what we will be allowed to do and not do with human samples and data.” Reymond emphasized that patients should be informed additionally that their samples could be used for business, not just pure academic research.
Reymond went on to describe his company’s sample collection procedures, which involve paying an employee to be in charge of obtaining and transporting a surgical sample from the operating room to the lab. Not only does this dedicated person deal with the informed consent end of things, the person also assures that a pathologist has properly taken a sample from the proper area of the organ being studied. For example, in colorectal cancer, the first of seven epithelial cancers that Europroteome is studying, “you have to be aware of a lot of contamination and necrosis.” Also, since recent studies have shown that there is genetic variation within tumors, the sample must be taken from the place on the tumor that is under study and must be labeled as such. While an internal control sample can be taken from, say, a non-tumor area of the patient’s colon, this cannot be regarded as a “normal,” sample, said Reymond. However, the problem is not solved by comparing the patient’s tumor sample to a “normal” sample from another patient, as gene expression and protein patterns can vary from patient to patient. The sample, additionally, must be at least 5 x 5 x 5 millimeters — a problem with small tumors such as breast cancer.
Europroteome is currently gathering such samples and doing both gene expression profiling with Agilent cDNA arrays as well as proteomic profiling with 2D gels. The company, which has its headquarters outside of Berlin, has grown out around its magnetic bead-based invention for sorting out epithelial cells from samples. In this patented process, antibodies for a ubiquitous epithelial antigen are attached to the beads, then the sample is run through them in a test tube. The beads “catch” the epithelial cells, then are separated from the rest of the sample.
With protein profiling, Europroteome also has tried the Ciphergen SELDI-based system, but finds that 2D gels are still the most informative technology. As far as protein microarrays go, Reymond thinks they will only be used for confirmation of results, with 100 or fewer protein capture agents on an array. “I don’t believe in high-throughput proteomics. I believe in targeted proteomics in combination with gene expression” studies, he said.
The session’s leader, Joel Rossier, chief technology officer of DiagnoSwiss, expressed even more skepticism about protein microarrays during a discussion on the subject at the end of the morning’s session. “The biggest problem is the dynamic range on the chip. You can have a lot of features on cDNA arrays, but if you have different concentrations — which can be as many as nine orders of magnitude — it is hard to match low-abundance and high-abundance proteins.”
In the first afternoon talk, Jörg Funk of febit in Mannheim, Germany, illuminated the audience with an explanation of the company’s reflective micromirror microarray machine, called geniom one. While admitting that other companies, including NimbleGen and Xeotron, are using the Texas Instruments micromirrors for light-based in situ synthesis of oligonucleotides on microarrays, he noted that febit is the only company “that offers an integrated solution.” Indeed geniom one seems like a classic marvel of German engineering ingenuity, in that it combines sample preparation, hybridization, and reading in one groovy futuristic-looking box. The user need only label the samples and inject them into a cartridge that is then inserted into the machine. The particular probes on an array are designed on the computer as digital images, then reproduced on the chip through a series of bouncing angled mirrors. Funk demonstrated the accuracy of the system by showing how he fed a picture of the development team at febit into the machine, then the machine printed out an array that made a facsimile of the picture, albeit in black and white.
The geniom one, however, is not scheduled to go onto the market until next year, and is only available for early access customers at a fee of $300,000 per machine, as well as $1,000 per chip cartridge.
Even when it does go on the market, there are inherent limitations. Only one chip can be inserted at a time. Funk said he did not know about the coupling (hybridization) efficiency of the system, but session chair Jörg Hoheisel of the German Cancer Research Center in Heidelberg, told BioArray News that the chips produced by febit’s machine are as reproducible and work as well as other microarrays he has tried.
GeneScan of Freiburg, Germany, described another innovation: on-chip PCR. The process involves using v-shaped oligos anchored to the surface of the chip. The company has a CYP2D5 chip, as well, which it expects to be approved as a diagnostic screening device in Europe by the end of the first quarter, 2003, and by the end of 2003 in the US.
Jerini of Berlin is also aiming for the commercial microarray market, but is pursuing a different angle. The company is making peptide arrays with an inkjet-style non-contact printer for kinase and protease profiling, as well as antibody profiling, according to Jerini representative Ulf Reimer. Almost all of Jerini’s customers are big pharmas, and the company prefers to do fee-for-service work than to sell its chips, said Reimer.
PCR vs. Oligos
Flying in the face of the prevailing wisdom that oligonucleotide arrays are preferable to PCR arrays, Daniel Marechal of Eurogentec in Liege, Belgium said in his talk that the company has found that the PCR chips offer a more cost-effective and wider selection of probes. Additionally, he said, the probe sequences, which now are as short as 200 to 500 base pairs, can be randomly primed rather than primed solely at the 3’ end, leading to less bias in probe selection.
“PCR microarrays have more advantages over oligo arrays,” he said. Marechal showed a graph of an internal study the company has performed indicating that PCR product arrays were significantly more sensitive than those containing oligos of any length between 30 and 70 bases. Eurogentec sells PCR-based arrays for seven different microbial pathogens as well as different species of yeast, Drosophila, breast cancer, and leukemia.
Although Eurogentec developed two of these arrays in collaboration with the German Cancer Research Center, Marechal’s statement about PCR’s advantages elicited polite protests from Hoheisel. “An issue in our lab was purification of PCR products. We don’t do it any more,” he said. In fact, Hoheisel has found that when unpurified PCR products are spotted on glass slides, the resulting signal intensities are 94 percent as good as those obtained with purified PCR products.
In addition to saving time and expense, the elimination of the PCR purification process has also meant keeping the 50 percent of PCR product that can be lost during this process. “Not having to do purification, we can do twice as many arrays as we could before,” he said.