Abstract
The completion of the Human Genome Project has been hailed as the most important scientific advance of recent times. Some commentators have even predicted that this definitive knowledge of our genetic makeup will eventually transform medicine by leading to the development of revolutionary new therapies and drugs. Genes will be identified that are responsible for particular diseases, and methods will be developed that allow specific genetic defects to be corrected or circumvented. However, biologists have been quick to point out that these kinds of developments will require much more than just the sequence of the human genome. Even identifying the 50,000 or so genes that this sequence contains will not be enough. The challenge now is to understand in detail what the function of all these thousands of genes is and how they interact. This vast new field has been called “Functional Genomics”, and it may prove to be even more of a challenge than posed by the Human Genome Project. The success of that project depended upon the automation of nucleotide sequencing, and it is clear that significant progress in Functional Genomics will also require the development of automated methods for many of the analytical techniques that are currently carried out manually.
A group at the Max Planck Institute of Experimental Endocrinology in Hannover has collaborated with the automated solutions company, Tecan to develop GenePaint, an automated system for carrying out in situ hybridisation. This technique is used to reveal the way in which particular genes are expressed in different tissues, information which is an important first step in understanding gene function.
The Director of the Experimental Endocrinology Institute, Professor Gregor Eichele, explained the background to their work: “Gene function is critically illuminated by knowing where and when a gene is expressed. We need to know where genes are expressed in pathological samples, but also of course in normal tissue and during development. When you look at any tissue, human or animal, or even plants, genes are often only expressed in certain cell types.
“If you have a piece of tumour tissue, a cancer gene may only be expressed in certain cells, and this is very often used as a diagnostic tool. The sequence of a gene itself is not very useful functionally and there are many tens of thousands of genes whose expression pattern has to be determined.”
One way of investigating which genes are expressed in a particular tissue is to use messenger RNA extracted from that tissue to probe a micro-array containing thousands of dots, each containing copies of a different gene. Although this technique is rapid, it is also relatively crude since it is not suitable for providing any information about whether a gene is expressed in particular cells or regions within a tissue. Much more information can be obtained from the technique of in situ hybridisation. This method involves the hybridisation of a probe derived from a single gene to thin tissue sections mounted on microscope slides. Bound probe is then detected using radioactive or histological techniques. Microscopic examination of these slides allows the precise pattern of gene expression to be determined at the level of an individual cell. However, a major drawback of this technique is that it is very labour intensive. One person might only be able to process 20 slides in two days. Until now, time and labour restraints have limited the extent to which in situ hybridisation can be used to study gene expression on a genome-wide scale.
Professor Eichele commented “In situ hybridisation is of course well known as a manual technology, but it is very labour intensive. We were already using it a lot in our lab and it became imperative that we automate the process.”
This motivation led Professor Eichele's group to develop GenePaint, an automated system for carrying out in situ hybridisation. This system uses the Tecan GENESIS Robotic Sample Processor to control the addition of solutions, incubation of hybridisation reactions, subsequent washing of preparations, and the colour reactions for gene expression detection. The GENESIS platform is combined with a special instrument described by Professor Eichele: “The key part is a small hybridisation chamber which is composed of the microscope slide which carries the tissue, and then a glass plate which is pressed against this slide with a couple of spacers between so that you get a very thin flow-through chamber. This is mounted on a frame that is placed into a temperature control rack. A probe complementary to the endogenous cellular mRNA is hybridised at 55°C, and then you need to wash off the excess at about 62°C. You require a very careful temperature control just like you do for polymerase chain reactions. We use a non-radioactive method to detect the bound probe which requires a number of washing and reaction steps that are delicate and time consuming so automation really helps enormously.”
The Tecan GENESIS Robotic Sample Processor plays a central role in the GenePaint system by controlling the accurate delivery of a sequence of almost 40 different incubation and wash solutions to individual flow-through chambers. The pattern of gene expression within each tissue section is detected using epitope-tagged riboprobes. These probes are stable and can be visualised using commercially available histological reagents. The development of an automated system for documenting results has just been completed in a collaborative effort between Leica Microsystems and Professor Eichele's group. This equipment consists of an auto-focusing Leica microscope equipped with a computer-controlled scanning stage and a digital camera. The images are stored on a computer and can be processed by automated image analysis. GenePaint can be used with either paraffin-embedded or frozen sections and is capable of processing up to 200 standard microscope slides within 20 hours. Gene expression can be detected at the single cell level with an exceptionally low background. The whole procedure is a general one that is applicable to a broad variety of tissue types and genes.
Professor Eichele identified several advantages of having an automated system for carrying out in situ hybridisation in this way. “If you were doing it manually, you would need one technician to do twenty slides for two continuous days. Every few minutes they would have to do something such as immersing slides in a series of solutions, pipette reagents onto the sections and place slides into incubators. It would take more than a day and a half for twenty slides. Now a single person will do at least ten times as much. The rate of data generation is restricted by how many GenePaint machines you have, although obviously you have to keep an eye on them. Another advantage is the reproducibility of the process because the robotic system always does things in the same way. Finally, we found that unless individuals are very skilled, the manual procedure resulted in damage to the tissue. For example, you had to incubate the specimens underneath a coverslip that you then had to remove, and in doing so, you would often damage the delicate tissue section which are only a few thousandths of a millimetre thick.”
The development of the GenePaint automated system means that it is now possible to process hundreds or even thousands of genes within a few months. Consequently, it is now possible to envisage the technique being used as a generalised screening method for identifying genes with particular patterns of expression. For example, it will be possible to track down genes that are only expressed in a specific region of the brain. Such genes might encode potential drug targets or have a key role in basic brain physiology and pathophysiology. These genes might be involved in the development or function of the region in which they are expressed, and might therefore play a role in the aetiology of those diseases that are characterised by abnormal brain function. Another way in which the system could be used is to build up a detailed three-dimensional picture of gene expression within a tissue. The expression of a single gene could be studied in serial sections and then a combined image can be created by computer analysis. This would provide a more realistic and accurate picture of gene expression than is provided by examining individual sections.
Development of the prototype for the GenePaint system began in 1998. After a detailed evaluation of available sample processing systems, the Tecan GENESIS Robotic Sample Processor was chosen as the optimal platform. Then, in collaboration with Reiner Pisalla, Tecan's Global Solutions Manager for Genomics, and his team, the prototype design was further refined and tested, and modifications made to the existing GENESIS software to control additional functions. Beta tests for GenePaint are currently being carried out at Harvard University in Cambridge, MA, USA, and the system will soon be available commercially from Tecan. The team has also been approached by many diagnostic groups interested in applying the technology to FISH (fluorescence in situ hybridisation) techniques.
Looking to the future, Professor Eichele expects the GenePaint system to be useful in investigations of the fundamental mechanisms of disease and non-disease processes. “This sort of application opens new avenues. Accordingly, there is a very large interest in this technology from the academic community and also from pharmaceutical companies.” One can safely predict that the automation of analytical techniques using liquid handling machines such as the Tecan GENESIS Robotic Sample Processor will play an increasingly important role in making Functional Genomics a reality. As Professor Eichele noted, “Any kind of repetitive work that you do, machines can do also. If any part of a protocol becomes a real bottleneck then we must automate it.”