WO2001061052A1 - Dna arrays for determining drug selectivity - Google Patents

Abstract

Generally the present invention is directed to a method of screening drug candidates as well as to compositions identified thereby. Hybridization experiments utilize the immobilized sequences as 'bait' which are used to analyze a mix of targets, which are the cDNAs derived be reverse transcription (RT) of cellular mRNA. Fluorescently-tagged nucleotides are included in the RT reactions, so that the RT-PCR generated cDNA will be fluorescently labeled. The binding of the labeled cDNA to the template DNA can be evaluated by confocal microscopy of the slide under laser illumination. Statistical and comparative analysis of the resulting data shows which genes are significantly expressed.

Description

DNA ARRAYS FOR DETERMINING DRUG SELECTIVITY FIELD OF THE INVENTION

The present invention combines functional genomics, vis-a-vis cellular gene expression and drug discovery techniques to achieve rational development of potential drug candidates. Pattern(s) of gene expression are utilized to direct and facilitate drug research as well as identify drug candidates.

BACKGROUND OF THE INVENTION

Gene expression assays are becoming practical tools for quantitating the effect of a drug or drugs on a large fraction of genes and/or proteins in a cell culture. Raw data from gene expression assays, however, are often difficult to coherently interpret, as measurement technologies typically return numerous genes with altered expression in response to a drug (anywhere from as few as 10 to up to 1,000 or more). The fact that one or a few genes among many has an altered expression yields little insight and fails to provide meaningful information from a drug design perspective.

Without effective methods of analysis, one is left to ad hoc further experimentation to interpret gene expression results in terms of biological pathways and mechanisms. Procedures for guiding the interpretation of such data, at least in the case of biological pathways is disclosed, for example, in U.S. Patent No. 5,965,352, which is hereby incorporated in its entirety herein by reference thereto.

The focus of current models of drug discovery and design are generally on screening compounds for activity against defined molecular targets thought to be important for a particular disease process. During the process of lead optimization, attempts are made to improve potency and specificity of a drug through analog development and testing. However, potency without selectivity is of questionable value since it is the therapeutic index (the ratio of the toxic to therapeutic dose) that is often a drug's critical feature. Often, the search for potency during lead optimization is a surrogate for expected cellular and in vivo specificity of the compound based on the belief that lowering the concentration of a drug that is required to produce a desired effect will result in fewer nonspecific effects. Using these techniques it has only been possible to assess selectivity of a compound for a target by comparison to a finite number of related targets. Additionally, it is only possible to assess selectivity of a compound for a target by comparison to a very finite number of related targets. Often, side effects due to interactions of a drug with related or even non-related targets may not be discovered until late in the drug development process when the drug enters animal testing. Unexpected cross reactions and systemic toxicity can be reasons for a drug to fail in development and is often the most costly part of pre-clinical drug development.

Another approach to drug screening involves testing numerous compounds for a specific effect on a known molecular target, typically a cloned gene sequence or an isolated enzyme or protein. High- throughput assays can be developed in which numerous compounds can be tested for the ability to alter enzyme activity (e.g., by changing the level of transcription from a specific promoter or by altering the binding of identified proteins). Although the use of high-throughput screens is a powerful methodology for identifying drug candidates, it has limitations. A major drawback is that the assay provides little or no information about the effects of a compound on related or non-related targets. It also provides no information at the cellular or organismal level. These effects must be tested separately by using the drug in a series of additional assays such as cell biologic and whole animal studies to determine toxicity or side effects in vivo.

Currently many pharmaceutical companies have developed secondary screens that use thirty or more related targets. This large scale secondary screening is extremely costly and may have limited value since there may be hundreds or even thousands of other targets on which the drug could act, as well as unrelated targets whose inhibition could lead to unexpected toxicity.

There is a recognized need in the early stage of drug development to significantly improve the identification of compounds which impact single and/or multiple molecular targets. Early and accurate identification of such compounds dramatically reduces the time and cost in the later stages of drug development.

SUMMARY OF THE INVENTION

The present invention is generally directed to a method of identifying a drug candidate which comprises isolating and labeling both target and test genetic material and hybridizing the target and test genetic material to an uncharacterized or substantially uncharacterized template of genetic material and comparing the expression of the test genetic material to the target genetic material. A potential drug candidate is selected based on the similarity of expression of the test and target genetic material. It is preferable that the target and test genetic material are hybridized to individual templates along with control genetic material so that differences in the pattern of expression may be obtained.

More explicitly, a method of identifying a drug candidate is disclosed. An initial step is the construction of a template by depositing reference genetic material (e.g., cDNA) from a cell onto a substrate. A novel aspect of the present invention is that the genetic material deposited on the template to form the array does not need to be characterized. In other words, only the pattern of genetic material has to be reproducible. Accordingly, advances in drug design can be made without knowing the precise sequence of DNA being analyzed. The method includes the step of isolating a control genetic material (e.g., cDNA) from the cell fine (generally the same cell line used to create the template), and labeling the control genetic material. The cell line is altered to form target genetic material and the target genetic material (e.g., cDNA) is isolated and labeled. The target genetic material and control genetic material are preferably labeled with different fluorophores so that once they are hybridized to the template, a visible differential pattern (e.g., fluorometrical or spectrophotometrical) is exposed. The control genetic material and the target genetic material are hybridized to the reference genetic material. Hybridizing occurs preferably (e.g., through application of a mixture of target and control genetic material) in a simultaneous manner but could occur sequentially without mixing of the target or control genetic material. A cell from the cell line is exposed to a potential drug candidate and genetic material from the exposed cell is isolated and labeled to form test genetic material (e.g., cDNA). The control genetic material (labeled) and the test genetic material (labeled) are hybridized with the template genetic material to determine a second pattern of expression. The first and second patterns are compared and a drug candidate is selected based on similarity between the first and second patterns of expression. The reference genetic material forming the template may be randomly deposited in a fixed pattern, and may be substantially uncharacterized. The cell may be selected from the group consisting of a mammalian cell, a bacterial cell, and a yeast cell. The target genetic material preferably may include an alteration selected from the group consisting of a homozygous knockout, a heterozygous knockout, a dominant negative alteration, an antisense alteration, and a drug induced alteration. The potential drug candidate may be an anticancer drug.

In addition, the target genetic material and the test genetic material are usually labeled similarly and the control genetic material is labeled differently. The target genetic material and test genetic material, respectively, along with the control genetic material, may be hybridized to the reference genetic material in a sequential manner or simultaneously as a mixture. An alternate embodiment of the present invention is a method of identifying a drug candidate and a drug candidate identified thereby which comprises depositing reference genetic material from a cell line on a substrate to form a template; altering the cell line to form target genetic material which is isolated and labeled; exposing a cell from the cell fine to a potential drug candidate; labeling genetic material isolated from the cell exposed to the potential drug candidate to form test genetic material; and hybridizing the target genetic material and the test genetic material to the reference genetic material to determine a pattern of expression for both. In this case, a drug candidate is preferably selected based upon a null pattern.

Yet a further embodiment of the invention resides in a kit comprised of a microarray of substantially uncharacterized genetic material (e.g., cDNA) in a fixed pattern, the microarray capable of being hybridized with genetic material (e.g., cDNA) of a test cell; and an array pattern, the array pattern corresponding to a differential expression pattern of genetic material (e.g., cDNA) isolated from a control cell and a cell altered in a desired manner which has been hybridized with genetic material (e.g., cDNA) isolated from a reference cell. The kit may further include a test cell and means for isolating and labeling genetic material. The kit may also include means for hybridizing cDNA of the test cell to the microarray. The array pattern may a digital image (e.g., a picture of colored spots).

Another feature of the invention resides in a method of screening drug candidates comprised of: providing a hybridizable microarray of genetic material (e.g., cDNA) whose template DNA pattern is known to a first party; providing a known expression pattern of genetic material (e.g., cDNA) (or for that matter a library of expression patterns) which was previously isolated, labeled and hybridized to the microarray, preferably along with a control. The identity of a candidate drug or drugs that are used to treat the cells to obtain a test is preferably known only to a second party. The expression patterns or library of expression patterns generally correspond to known alterations. The known alterations may include alterations selected from the group consisting of a homozygous knockout, a heterozygous knockout, a dominant negative alteration, an antisense alteration, and a drug induced alteration. The identity of the cDNA library may be known only to the first party.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a schematic representation of a preferred embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

Genome sequencing is currently underway for several eukaryotic organisms, including humans, nematodes, Arabidopsis, and flies as well as many strains of yeast. Extensive gene expression profiling work has been done with the budding yeast Saccharomyces cereυisiae using arrays of more than 260,000 synthetic oligonucleotides covering the entire S. cereυisiae genome. Messenger RNAs present at a frequency of 1:150,000 to 1:30,000 can be unambiguously detected by these arrays.

DNA microarray technology is a result of advancements in several fields including fluidics, robotics, chemistry, optics, and computing. It is now possible to immobilize tens of thousands of defined DNA sequences in very regular patterns of very small spots (for example 100 micron spots on 200 micron centers) on a chemically-activated glass substrate (microscope slide) of the type available from Sigma. These sequences can either be known genes, or Expressed Sequence Tags (ESTs- cDNA sequences derived from the ends of cellular mRNA which code for a gene of unknown function). For use in microarrays, ESTs are typically chosen to represent the 3' or 5' end of the mRNA. These sequences can then be used as binding templates to perform an inverse-Northern hybridization. These hybridization experiments utilize the immobilized sequences as "bait" which are used to analyze a mix of targets, which are the cDNAs derived by reverse transcription (RT) of cellular mRNA. Fluorescently-tagged nucleotides are included in the RT reactions, so that the RT-PCR generated cDNA will be fluorescently labeled. Thus, the binding of the labeled cDNA to the template DNA can be evaluated by confocal microscopy of the slide under laser illumination. Binding is indicated by fluorescent spots on the slide. When different fluorophores are used during the labeling of mRNA populations from different sources, and the microarrays are scanned at the appropriate wavelengths, the relative amounts of mRNA can be simultaneously determined for different samples. Statistical analysis of the resulting data shows which genes are significantly expressed. Comparisons that have been made in this manner include normal and tumor tissue, transfected cell lines and their parental fines, and cells before and after treatment with drugs.

Researchers have used EST arrays to study S. cereυisiae as it changed from anaerobic to aerobic metabolism, a diauxic shift, which is the switching from one metabolic state to another. These studies revealed that only 19 mRNAs showed more than a 2-fold change under control conditions. However, as the glucose in the medium was depleted, more than 700 genes showed at least a 2-fold increase in expression and more than 1000 genes showed a 50% decrease. About half of the differentially expressed genes were represented by ESTs, but others were of known function. A detailed map of the metabolic reprogramming that occurs during the diauxic shift was constructed. These results demonstrate that a change in cell function, particularly changes in such a basic process as cell growth will produce measurable changes in gene expression. The present invention will allow detection of these changes in gene expression which will form the "expression fingerprint."

S. cereυisiae oligonucleotide arrays have also been used to compare the effects of a potent purvalanol analogue flavopiridol, a CDK-2 inhibitor currently in clinical trial. A series of purine analogues related to purvalanol as selective inhibitors have been synthesized using combinatorial chemistry of human cyclin-dependent kinase 2 (CDK-2). The lead compound was also a selective inhibitor of the S. cereυisiae CDK homologue Cdc28p. Signaling genes that were clearly decreased by both purvalanol and flavopiridol included G2 cyclin, which is important in the transition to mitosis; EGT2, which controls the timing of cell separation after cytokinesis; and the histone genes HTA2 and HTB2. Transcription of all these genes is known to be regulated by CDK-2. Surprisingly, it was also found that genes involved in cellular phosphate metabolism were affected. This was explained by the finding that Pho58p kinase, which monitors intracellular phosphate levels in yeast and which is closely related to Cdc28ρ, was also inhibited by the purvalanol analogue. Genes that were upregulated by both purvalanol and flavopiridol included heat shock proteins, a number of genes encoding members of the ATP -binding cassette superfamily and other drug transport proteins, as well as cell wall proteins implicated in increased drug resistance in yeast. Thus, these agents were not specific for inhibiting CDK-2 and instead cross- reacted with many other important cellular processes. Unfortunately, these side effects were only identified after the design and synthesis of the analog. It is likely that an even more complex pattern of gene changes will be seen in mammalian cells exposed to growth inhibitory drugs. Thus, using these techniques alone, it is difficult and expensive to predict and design effective analogues by traditional methods in mammalian cells.

The present invention uses DNA microarrays to assist with lead drug selection during drug development. An aspect of the present invention is the testing of compounds for their effects on cellular gene expression, and using this pattern of expression as a measure of selectivity. These patterns of altered gene expression are used to identify and to direct the rational development of more selective drugs. The present invention identifies drugs that most closely reproduce the pattern of gene expression changes (the differential "expression fingerprint").

Thus, the present invention increases the rate at which effective drugs are discovered as well as decreases the cost of developing the drugs. The desirable target expression fingerprint is generated by any number of methods including selective genetic deletion ("knockout"), molecular inhibition ("dominant negative" or "antisense"), or less preferentially pharmacologically induced modification of the cell of interest.

In accordance with the present invention, drugs are any compounds of any degree of complexity that impact a biological system, whether by known or unknown mechanisms and whether or not they are used therapeutically. Drugs thus include: typical small molecules of research or therapeutic interest; naturally-occurring factors, such as endocrine, paracrine, or autocrine factors or factors interacting with cell receptors of all types; intracellular factors, such as elements of intracellular signaling pathways; factors isolated from other natural sources; and so forth. The biological effect of a drug may be a consequence of, inter aha, drug-mediated changes in the rate of transcription or degradation of one or more species of RNA, the rate or extent of translation or post-translational processing of one or more polypeptides, the rate or extent of the degradation of one or more proteins, the inhibition or stimulation of the action or activity of one or more proteins, and so forth. By identifying the expression fingerprint of a drug with a desirable therapeutic activity, it is possible to identify other compounds having a similar therapeutic activity, as well as to identify compounds with greater specificity. In such an application, the significance of the similarity between a genetic deletion or molecular inhibition is determined to see if it meets a certain threshold of significance.

The methods of the present invention can be used to identify new drugs that mediate the therapeutic actions or that mediate the side- effects of a drug of interest by comparison of the drug of interest with other drugs having similar therapeutic effects. Two drugs are considered to have similar therapeutic effects if they both exhibit similar therapeutic efficacy for the same disease or disorder in a patient or in an animal disease model. The methods of this invention can be applied to determine the effect of the drug of interest and also of a second drug with similar therapeutic effects. By comparing expression fingerprints for additional drugs with similar therapeutic effects, the therapeutic effects of the drug of interest can be further narrowed, identified or modified.

The present invention allows determination of the drug's downstream effects at the transcriptional level. This determination is based upon the assumption that, for example, a genetic deletion or the molecular inhibition of a target in a cell will produce a desirable biological effect and induce a measurable change in gene expression. Similarly, exposure to a drug that inhibits the same target would be predicted to produce similar changes in gene expression. However, since a drug is never completely specific, there will almost certainly be additional changes in gene expression seen with the drug. For example, as a result of regulatory, homeostatic, and compensatory networks and systems known to be present in cells, even an "ideal drug," i.e., a drug that directly affects only the desired target, without effects on other constituents, will have complicated and often unpredictable indirect effects. Accordingly, a drug that is not ideal, e.g., one that directly affects more than one molecular target, may have still more complicated downstream effects. Thus, the present invention will identify the compounds that show the same temporal, qualitative and quantitative changes in gene expression as seen with the genetic or molecular target deletion. In instances where the compound fails to produce the same changes as those seen with the genetic or molecular target deletion, it would be assumed that the compound acts by a different mechanism and this compound would be eliminated from further testing. It is expected that the compound may induce many changes in gene expression that include early response, late response, stress and death. The present invention will identify these changes by detecting specific changes due to exposure to an experimental agent against a background of other, subsequent changes resulting from cellular debilitation caused by those specific effects. The pattern of genes which are specifically affected will produce a characteristic fingerprint for the target, a fingerprint that will serve as a blueprint for drug development for that particular target. A unique aspect of the present invention is its utility regardless of whether the affected genes encode proteins of known function or whether the genes used as the reference material are of unknown function. Also the affected genes do not even have to be known but can be anonymous or substantially anonymous. All that is required is a reproducible pattern of reference genetic material. The present invention relies on the expression fingerprint, preferably resulting from the deletion or inhibition of a particular molecular target in a cell. The specific functions of the affected genes do not have to be known for interpretation of these expression fingerprints. It is anticipated that initially many of the genes exhibiting changes in expression will be of unknown function since ESTs predominate in the currently available libraries.

In the preferred embodiment of the present invention schematically illustrated in Fig. 1, an anonymous gene array 21 - a genetic material library (e.g. cDNA) is prepared from an appropriate drug- sensitive reference cell 23 in which the individual genes are not identified is utilized. This reference cell 23, or more precisely reference cell line 23, preferably expresses the target of interest and shows a desirable biological response to a lead compound (e.g., for an anticancer drug this could be growth inhibition). The reference cell fine 23 may be human, non-human, prokaryotic or eukaryotic. The reference cell line 23 is preferably the source of reference genetic material 24 that is used to form the template or base gene array 21. Next, either a genetically modified form of the cell is created (e.g., knockout, either homozygous or heterozygous), or a molecularly inhibited form of the cell is constructed (e.g., a dominant negative or an antisense). This is target cell 25. This modified cell 25 could be a knockout or a stable or a transient transfection. The modified cell line 25 is the source of the target genetic material 26 (see Scheme B below). Importantly, the reference cell line 23 is then treated with a drug of interest at an appropriate concentration for an appropriate time to form drug treated or test cell 27. This is described below in Scheme C. After sufficient amounts of the cell lines under conditions or Schemes A, B, and C are grown, and total mRNA is isolated from the cell lines. The poly- adenylated mRNA may be purified from the total mRNA. This mRNA or the purified polyA mRNA 24a, 24b, 26 and/or 28 is then labeled by a reverse transcriptase reaction with suitable different fluorescent or other probes. Examples of suitable fluorescent probes include dCTP labeled Cy3 and Cy5 available from Amersham Pharmacia. Reference cell line 23 preferably runs as a control cell 23a along side the experimental manipulation of both the target cell 25 and test cell 27. Control cell 23a is cell fine A described below. Preferably, control genetic material 24a is labeled with one probe and target genetic material 26 and test genetic material 28, respectively, are labeled with a different probe. These reverse transcribed, labeled templates serve as the probe material in the microarrays of the present invention.

The cDNA is prepared by reverse transcriptase polymerase chain reaction (RT-PCR) from three sources including:

Scheme A. Reference cell line 23 (human or non human - the template or wildtype cDNA). Source of reference genetic material 24 and/or control genetic material 24a.

Scheme B. Target cell fine 25. This is the reference cell line described in (A) in which the drug target molecule has been modified (e.g., deleted either by genetic knockout or inhibited by molecular means, such as antisense or dominant negative). This is the source of target genetic material 26.

Scheme C. Test cell line 27. This is the same cell line described in (A) after drug treatment. In other words, the reference cell line exposed to a potential drug candidate. This is source of test genetic material 28.

The pattern of gene expression is determined by hybridizing the probe cDNA from A, B, or C described above to arrays of the template DNA. Detection can be by any means having sufficient sensitivity (radioactivity, colorimetric or fluorescence). Differential hybridization can be employed using two or more probe DNAs (e.g., A-C) with different molecular tags being hybridized at the same time. The cDNA arrays are prepared on a suitable support such as a membrane, or a microscope slide (chip). The cDNAs can be individual genes or 3' or 5' ESTs of the partial or entire genome of the species from which the cell line under investigation originated. Alternatively, it could be a cDNA library prepared from the cell line under investigation. Importantly, the identity of the individual cDNAs, ESTs, or cDNAs from the library may be known substantially uncharacterized or entirely uncharacterized. The patterns of gene expression obtained by the present invention are:

A compared to A = The "normalization" expression fingerprint. That is preferably cDNA 24 to cDNA 24a or cDNA 24b. This should be a null pattern.

B compared to A = The target expression fingerprint 30. This is preferably target cDNA 26 and control cDNA 24a hybridized to template 21.

C compared to A = The test expression fingerprint 32. This is preferably test cDNA 28 and control cDNA 24b hybridized to template 21.

Analysis of the data obtained by the present invention is performed by computer using appropriate pattern recognition algorithms. The objective is to identify a compound, or compounds with an expression fingerprint (C compared to A) that most closely matches or preferably is identical to the target expression fingerprint (B compared to A). This analysis could also be obtained by a direct comparison of C to B, looking for a pattern most similar to a null change. Nucleic acid hybridization and wash conditions are chosen so that the probe "specifically binds" or "specifically hybridizes" to a specific array site, i.e., the probe hybridizes, duplexes or binds to a sequence array site with a complementary nucleic acid sequence but does not hybridize to a site with a non-complementary nucleic acid sequence. It can easily be demonstrated that specific hybridization conditions result in specific hybridization by carrying out a hybridization assay including negative controls (see, e.g., Shalon et al., supra, and Chee et al., supra). General parameters for specific (i.e., stringent) hybridization conditions for nucleic acids are described in Sambrook et al., supra, and in Ausubel et al., 1987, Current Protocols in Molecular Biology, Greene Publishing and Wiley- Interscience, New York, which is incorporated in its entirety for all purposes. When the cDNA microarrays of Schena et al. are used, typical hybridization conditions are hybridization in 5X SSC plus 0.2% SDS at 65°C for 4 hours followed by washes at 25°C in low stringency wash buffer (IX SSC plus 0.2% SDS) foUowed by 10 minutes at 25°C in high stringency wash buffer (0.1X SSC plus 0.2% SDS) (Shena et al, 1996, Proc. Natl. Acad. Sci. USA, 93:10614).

When fluorescently labeled probes are used, the fluorescence emissions at each site of a transcript array can be, preferably, detected by scanning confocal laser microscopy. In one embodiment, a separate scan, using the appropriate excitation line, is carried out for each of the two fluorophores used. Alternatively, a laser can be used that allows simultaneous specimen illumination at wavelengths specific to the two fluorophores and emissions from the two fluorophores can be analyzed simultaneously (see Shalon et al., 1996, A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization, Genome Research 6:639-645, which is incorporated by reference in its entirety for all purposes). In one embodiment, the arrays are scanned with a laser fluorescent scanner with a computer controlled X-Y stage and a microscope objective. Sequential excitation of the two fluorophores is achieved with a multi-line, mixed gas laser and the emitted fight is split by wavelength and detected with two photomultiplier tubes.

Signals are recorded and, in a preferred embodiment, analyzed by computer, e.g., using a 12 bit analog to digital board. In one embodiment the scanned image is despeckled using a graphics program (e.g., Hijaak Graphics Suite) and then analyzed using an image gridding program that creates a spreadsheet of the average hybridization at each wavelength at each site. If necessary, an experimentally determined correction for "cross talk" (or overlap) between the channels for the two fluors may be made. For any particular hybridization site on the transcript array, a ratio of the emission of the two fluorophores can be calculated. The ratio is independent of the absolute expression level of the cognate gene, but is useful for genes whose expression is significantly modulated by drug administration, gene deletion, or any other tested event.

To measure drug response data, the cells are preferably exposed to graded levels of the drug or drug candidate of interest. When the cells are grown in vitro, the compound is usually added to their nutrient medium. The drug added is a graded amount that depends on the particular characteristics of the drug, but usually will be between about 1 ng/ml and 100 mg/ml. In some cases a drug will be solubilized in a solvent such as dimethyl sulfoxide ("DMSO").

There are several methods of creating the modified cell lines whose gene expression would serve as a template for comparison of the treated cells. Creation of a homozygous knockout (-/-) will ensure complete deletion of the gene target. However, this type of knockout may be lethal if the gene is essential for a cellular process. Even if this homozygous knockout is not lethal, it may cause drastic changes in expression of other genes such as stress genes. This additional response is problematic when it is not observed with treatment with the inhibitory agent (e.g., drug). Since drugs rarely produce complete inhibition of a target, it is likely that engineered cell lines with stable transfections of a dominant negative or antisense construct (which typically exhibit between 50 and 70% reductions in gene expression) will serve as an appropriate comparison. Alternatively, a heterozygous knockout cell fine could be used if it showed similar but less drastic phenotypic changes to the homozygous (-/-) knockout. Preferred methods of creating the modified cell fines of the present invention include dominant negative and antisense inhibition which typically produce less than complete inhibition of a target. It is, however, desirable that the degree of inhibition is sufficient to produce the desired biological change in the cell phenotype.

Referring back to Figure 1, the expression fingerprint of the present invention is generated by hybridizing the labeled cDNA (A (24a) and B (26), and A (24b) and C (28)) to the cDNA array template 21. The pattern of differential gene expression is read based upon the intensity of binding between the labeled cDNAs. A pattern of changes caused by the genetic or molecular inhibition of the target and a pattern of changes caused by the drug is obtained and compared. The patterns of gene expression caused by the genetic or molecular target inhibition can be compared with the patterns produced by exposure to the compounds being tested using, for example, the publicly available COMPARE analysis program. This program was developed to analyze and compare drug cytotoxicity profiles of compounds in the National Cancer Institute human tumor cell fine panel. Test compounds will be assigned a correlation coefficient based upon their expression pattern similarity to the target inhibition pattern. The higher the correlation coefficient, the more selective the compound. It may be preferable to develop customized software to more specifically and effectively analyze this data. Interpretation of this data will employ cluster analysis to identify changes in linked sets of genes that may indicate mechanistic side effects or therapeutic similarities or differences among compounds. Although generally the probe colors are red, green and yellow for neutral, Fig. 1 illustrates the differential expression pattern is illustrated with solid, hatched, and dotted lines respectively. Information regarding the quality control during the production of microarrays and from the results of individual microarray experiments is stored, managed, and analyzed in a system built around a current generation relational database and a user interface preferably accessible via the internet. Basic analysis, performed within the database identifies genes which are significantly up-or down-regulated in single experiments and will identify genes with similar changes in multiple repetitions of the same experiment. Genes which are coordinately regulated across time courses of drug exposure or within families of related cell types can be identified using custom software to identify hierarchical clusters. This same software can be used for identifying gene expression patterns (fingerprints) against the variable background that will be obtained in screens conducted in different cell lines. Additionally, analysis of related subfamilies of genes (e.g., in the Kyoto Enzyme Database) can identify agents which coordinately affect members of specific metabolic and signaling pathways.

The appropriate time points for analysis of expression changes induced by an experimental drug, as well as the appropriate expression pattern to use as a reference will be determined on a case by case basis. The changes in gene expression caused by a drug will be dynamic depending on whether early or late response genes are affected; the initial changes could be lost or masked in a cascade of subsequent effects. There is a wide range of turnover times for cellular mRNA from minutes to days. This mRNA stability could affect the time that any given effect may become manifest. The changes will be compared with the engineered cell fines where the cell has had multiple generations to adapt to the changes and where lethal effects have been selected against. Preliminary experiments have demonstrated significant changes in gene expression at time points of several hours using standard chemotherapeutic agents and established model cell lines.

The specific protocols of the present invention will depend on the system being tested and the following protocols are meant for illustrative purposes only. The protocols are a synthesis of methods used by groups at the NHGRI, Stanford University, and Telechem. All reactions are run using RNAse-free reagents and labware where appropriate, following good lab practice for the handling of samples sensitive to RNAse degradation and fluors sensitive to photo bleaching. All solutions are filtered to 0.2μm to eliminate particulate contaminants.

A MCR-7 human breast cancer cell 3' EST cell library is to be prepared and arrayed in 96 well plates. Clones are to be PCR amplified directly from bacterial culture using a single set of primers. Overnight cultures in 200 μl of LB broth with Ampicillin at 100 μg/mL are prepared in 96-well plates using a 1 μl aliquot from glycerol stocks of the clones. PCR reactions are run in 96-well format using pre-aliquoted plates from Marsh Biomedical (#ZAM-600-415) which contain all needed reagents except primers and template. The total reaction volume is 50 μL; 5 μL of this is available for the addition of 1-2 μL of the bacterial culture, depending on the attained ODεoo after overnight culture, and 3-4 μL for addition of primers. The PCR is run in an MFJResearch 96-well thermocycler:30 sec 96°C for 1 cycle then (45 sec 94°C, 45 sec 55°C, 2.5 min 72°C) for 40 cycles then 10 min 72°C.

Unincorporated nucleotides and unreacted primers are removed from the PCR reactions using QIAquick 96-well PCR cleanup plates (Qiagen) following the manufacturer's protocols. The 50μL eluates are dried in a SpeedVac and are resuspended in 14 μL of 2X SSC. One microliter of the solution is analyzed by electrophoresis and ethidium bromide staining on a 1.2% agarose gel and 1 μL is quantitated using the PicoGreen reagent from Molecular Probes. Quantization results are crosschecked against the agarose gel. As needed, volumes are adjusted to produce a concentration of at least 300 ng/μl for printing.

The microarrays should be printed on nuclease-free, pretreated, silylated shdes (CEL Industries, Houston TX or Silane slides from Sigma) using an OmniGrid printing robot (Beckman Instruments). After printing, the slides are allowed to air dry. The printed nucleic acids are then immobilized onto the slide surface by rehydration in a 100% humidity chamber, immersion in 0.2% SDS for 1 min, immersion in dH₂O for 1 min, immersion in a solution of 1 g sodium borohydride in 300 mL PBS and 100 mL 95% EtOH for 5 min, and four sequential immersions in dH₂O for 2 min. Slides are briefly centrifuged to dry and can be stored for at least one year.

Total RNA will be extracted from cells (or e.g., mouse xenograft tumors) delete using the TRIzol reagent from GIBCO BRL/Life Technologies following the manufacturer's protocols. Qiagen Oligotex beads can be used to select polyA+ RNA from the total RNA following the manufacturer's protocols. The final mRNA will be adjusted to a concentration of 1 μg/mL by dilution with 10 mM TRIS pH 8.5 or concentration using a microcon 30 column (millipore) as needed.

This protocol is based largely on the recommended protocol for the Superscript II reverse transcriptase (GIBCO BRL/Life Technologies). In the reaction the concentration of unlabeled dCTP is reduced relative to dCTP, the balance is made up using Cy3- or Cy5-labeled dCTP (Amersham Pharmacia). RNAse inhibitor (GIBCO BRL/Life Technologies, lOU/μl) is included in the reaction. Following incubations the products are denatured using NaOH and HC1 neutralization.

Unincorporated nucleotides are separated using four sequential 0.5 mL washes with 10 mM TRIS pH 8.5 in a microcon 30 spin column. The final preparation is lyophilized without heat in a SpeedNac centrifuge. Pellets can be stored in dark at -20°C until hybridization.

Hybridizations are performed using individual shde chambers (TeleChem International). Slides, coverslips (VWR) and chambers are cleaned before use with EtOH and dried using compressed air and kimwipes. Once cleaned, components are handled with tweezers. The hybridization solution and microarray slides are prepared in parallel. The hybridization solution has a final volume of 10 μl and contains yeast tRNA, poly-dA and Cotl DNA (GIBCO BRL/Life Technologies) to act as blocking agents in 0.1% SDA in 3X SSC. Immediately before application to the microarray shde, this solution is incubated in a boiling water bath for 2.5 min and an ice water bath for 10 sec. Microarray shdes are prepared by immersion in ddH2θ at 93-97°C for 2 minutes followed by immersion in 100% EtOH for 15 sec and centrifugation at 500g for 1 minute to dry the slides. The hybridization reaction is assembled in a hybridization chamber to which 20μl 4X SSC has been added to each reservoir in the chamber. The 10 μl hybridization solution is applied as a single bead to the area of the shde where the DNA targets are located. To avoid trapping bubbles it is best to set the coverslip down on-edge on the slide not making contact with the probe solution and then to slowly tilt until its surface contacts the probe solution. Surface tension and gravity will then pull the other edge of the coverslip to the surface of the slide while spreading the probe solution in the process. The cover is attached to the hybridization chamber and the hybridization is allowed to proceed overnight with the chamber submerged in a 62°C water bath. The hybridized slides are washed before scanning by immersions in 0.5X SSC, 0.01% SDS for 5 min, following by immersion in 0.06X SSC, 0.01% SDS for 5 min following by immersion in 0.06X SSC, for 2 min and a brief centrifugation to dry the shde.

The results presented herein, specifically those for PX-12 were obtained using an Axon GenePix 4000A scanner. The determination of spot intensities in the Cy3 and Cy5 channels, and statistical analysis of the intensity distributions within the spots, is handled by dedicated control software for the instrument.

Our previous understanding of the interaction of an asymmetric disulfide compound on the thioredoxin reductase/(TR) thioredoxin (thioredoxin) redox system was utilized to test the present invention. PX-12 (1-methylpropyl 2-imidazolyl disulfide ) is a competitive inhibitor of the reduction of thioredoxin by thioredoxin reductase with a Ki value of 31 μM. It is not a substantial substrate for reduction by human glutathione reductase. This disulfide caused reversible thioalkylation of thioredoxin at the redox catalytic site as evidenced by the fact that there was no initial measurable reaction of the mutant thioredoxin where both the catalytic site Cys³² and Cys³⁵ residues were replaced by Serine. In addition, the disulfide caused a slower irreversible inactivation of thioredoxin as a substrate for reduction by thioredoxin reductase, with a half life of 4 hrs. This irreversible inactivation of thioredoxin occurred at concentrations of the disulfides an order of magnitude below those that inhibited thioredoxin reductase, and involved the Cys⁷³ of thioredoxin which is outside the conserved redox catalytic site, as shown by the resistance to inactivation of a mutant thioredoxin where Cys⁷³ was replaced by Ser. Electrophoretic analysis (not shown) and mass spectral analyses (also not shown) of the products of the reaction between the disulfide and thioredoxin, show modification of 1 to 3 Cys residues of the protein occurs in a concentration dependent fashion. The disulfides inhibited the thioredoxin dependent proliferation of MCF-7 breast cancer cells with IC50 values of 1.2 μM, respectively Based on previous studies PX-12 has IC50 value of about 1.2μM. Knowing this information allowed us to analyze PX-12 as a drug of choice and determine its "expression fingerprint", and as an extension of the present application utilize this information as a first pattern of expression to determine the level of similarity of a potential drug candidate. Table I below shows the results of exposing MCF7 cells at 70% confluency to PX- 12 at 50μM for 24 and 48 hours.

TABLE I

T l yyppee N iNaammee 24 hrs 48 hrs

Apoptosis "Homo sapiens TRAIL receptor 2 mRNA, complete cds" 1.59 2.29

Apoptosis Programmed cell death 2 -2.70 -2.45

Apoptosis apoptosis-related protein TFAR15 -1.76 -0.86

Cell Cycle H.sapiens mRNA for cyclin Gl - 1.68 -0.48

Cell Cycle CDC28 protein kinase 1 - 1.37 -2.97 Type Name 24 hrs 48 hrs

Cell Cycle Cyclin A 024 -214

Cell Cycle Cyclin D2 394 25

Cell Cycle "Human checkpoint suppressor 1 mRNA, complete cds" 159 173

Growth Signal transducer and activator of transcription 5A 165 129

Growth Proliferating cell nuclear antigen -245 -288

Growth V-myb avian myeloblastosis viral oncogene homolog-like 2 173 009

Growth P55-C-FOS PROTO-ONCOGENE PROTEIN 137 249

Growth "Retinoic acid receptor, gamma 1" 197 153

Growth Phosphatidyhnositol 3-kιnase pi 10 beta isoform 165 013

Growth Apo l_Human (MER5 (Aopl-Mouse)- ke protein) -237 -117

Growth ETS-RELATED TRANSCRIPTION FACTOR ELF- 1 202 162

Growth "Human K-ras oncogene protein mRNA, complete cds" 196 127

Growth "Human c-jun proto oncogene (JUN), complete cds, clone hCJ- 1" 070 169

Other CD9 antigen -190 -180

Other "Human chromatin assembly factor-I p60 subunit mRNA, 034 -171 complete cds"

Other Ribosomal protein S5 -063 -167

Other DNA polymerase alpha subunit -019 -192

Other Topoisomerase (DNA) II alpha (170kD) -059 -210

Other GATA-binding protein 3 -124 -206

Other Proteasome component C2 -135 -229

Other Ribosomal protein L32 -209 -097

Other Zinc finger protein 3 (A8-51) 215 026

Other PRE-B-CELL LEUKEMIA TRANSCRIPTION FACTOR-2 153 161

Redox Matrix metalloproteinase 2 113 165

Redox NAD(P)H menadione oxidoreductase -174 -184

Redox Glutathione S-transferase Ml -177 -139

Redox Quinone oxidoreductase (NQ02) 029 -162

Redox Glutathione reductase 206 056

Redox Human metallothionein-Ie gene (hMT-Ie) -130 -210

Redox H sapiens mRNA for metallothionein -189 -159

Redox Glutathione S-transferase theta 1 -201 026

Redox IMP (inosine monophosphate) dehydrogenase 1 175 012

Redox Glutathione S-transferase M3 (brain) -179 -109

Redox "Cytochiome P450, subfamily IIC (mephenytoin 4-hydroxylase) -216 -092

Redox Cytochiome P450 reductase - human, placenta 220 057

Reported values are the change in expression ratio as the number of standard deviations relative to the mean (no change in expression). A positive value indicates an increase in expression in the cells treated with PX-12. The fisted genes were selected based on a criterion for identifying outliers in the distribution by requiring that the expression ratio for at least one of the time points was more than 1.6 standard deviations from the mean. For microarrays of 502 genes, there was significant signal in both channels for about 100 genes. In this case, for a normal sampling distribution with a mean ratio of 1, by random chance 10 genes would be predicted to fall outside this range. However, for the two time points, 24 and 48 hours, the prediction is that much less than one gene (no genes) would show concordance by chance at both time points. Results of the experiment show that within a range of half of one standard deviation, 16 of these genes are in agreement at the two time points. These data show that following exposure to PX-12, the genes most affected could be categorized by their relation to apoptosis, cell cycle, growth, or redox control. These effects are logical given the role of PX-12 as a disulfide inhibitor of the thioredoxin signaling pathway.

A cDNA microarray was produced by printing duplicates of a selected set of 502 cancer-related genes and ESTs. Following RT-labeling and hybridization of samples discussed in Table I, the fluorescent intensities of the bound Cy3 and Cy5 label were determined independently based on gray-scale images. Shown were 1152 features representing 502 gene/ESTs printed in duplicate along with "landing lights" and other features to demarcate the limit of each field. Each field is made up of 24 columns of 12 rows of spots and is printed by a single printing of the array printer. Spots were approximately 160 microns in diameter and spaced on 225 micron centers. The two channel intensities at each spot were translated to intensity levels in the red and green channels, respectively, to produce a pseudo-color composite image. Spots with approximately equal binding of Cy3- and Cy5-label appear yellow, a greater bound proportion of the Cy3-label will appear red in this image, and if Cy5 predominates, the spot will appear green. The limits of the area of the array printed by each of four printing pins are marked by duplicate spotting oligonucleotides pre-labeled to a high level with Cy3. Table II below shows decreasing expression during the time course for MCF7 cells exposed to Doxorubicin. MCF7 cells at 70% confluency were incubated with Doxorubicin at 2μM for 1, 3 or 12 hours.

TABLE II

Type Name lhr 3hr 12 hr

Apoptosis ADP-ribosyltransferase (NAD+; poly (ADP-ribose) polymerase) 091 012 -063

Apoptosis Human cysteine protease Mch2 isoform alpha (Mch2) mRNA, 045 027 -073 complete cds

Cell Cycle Human cell cycle checkpoint control protein mRNA, complete cds 194 116 -016

Cell Cycle Cyclin Dl (PRAD1, parathyroid adenomatosis 1) 164 063 052

Cell Cycle Cyclin A 083 -033 -125

Cell Cycle Human cychn-dependent protein kinase mRNA, complete cds 076 046 -133

Damage Damage-specific DNA binding protein 1 (127 kD) 284 167 042

Damage DNA repair hebxase ERCC3 216 123 -033

Damage DNA repair protein MSH2 106 -043 -097

Damage Superoxide dismutase 1 (Cu/Zn) 103 012 -019

Damage H sapiens thiol-specific antioxidant protein mRNA 028 -111 -122

Damage ESTs, Weakly similar to HEAT SHOCK 27 KD PROTEIN 016 -054 -131 [H sapiens]

Damage Homo sapiens 9G8 splicing factor mRNA, complete cds -043 -071 -135

Growth Transferrin receptor (p90, CD71) 268 058 -018

Giowth IMP (inosine monophosphate) dehydrogenase 1 231 183 -174

Growth DUAL SPECIFICITY MITOGEN-ACTIVATED PROTEIN 212 066 -019 KINASE KINASE 1

Growth Eukaryotic translation initiation factor 4A (eIF-4A) isoform 1 165 088 -106

Growth Thrombopoietin 084 065 -012

Giowth Thioredoxin reductase 075 045 005

Giowth thioredoxin mRNA, nuclear gene encoding mitochondrial protein, 073 050 -054 complete cds

Growth S-ADENOSYLMETHIONINE SYNTHETASE GAMMA FORM 041 -048 -091

Growth Metallothionein IL -012 -018 -021

Growth RAS-RELATED PROTEIN RAB-5A -021 -051 -094

Growth Human mRNA for suppiessor for yeast mutant, complete cds 250 095 -091

Other GAPDH 210 067 -057

Other H sapiens mRNA for 55 11 binding protein 175 132 -008

Other Human inducible poly(A)-bιndιng protein mRNA, complete cds 158 129 002

Other TRANSFORMING PROTEIN RHOB 148 004 001

Other Small nuclear ribonucleoprotein polypeptides B and Bl 139 118 -116

These genes/ESTs exhibited a decrease in expression ratios across the time course. Reported values are the change in expression ratio shown as the number of standard deviations relative to the mean (no change in expression) at each time point.

Table III shows increasing expression during a time course for MCF7 cells exposed to Doxorubicin. MCF7 cells at 70% confluency were incubated with Doxorubicin at 2μM for 1, 3 or 12 hours.

TABLE III

Type Name lhr 3hr 12 hr

Apoptosis apoptosis-related protein TFAR15 -157 -111 065

Cell Cycle Cyclin H -169 -119 -083

Damage RNaseP protein p38 (RPP38) mRNA, complete cds -111 -055 -013

Growth Insulin-like growth factor 1 receptor -279 068 171

Growth Small inducible cytokine A2 -181 -002 072

Growth Human Ets transcription factor (NERF-2) mRNA, complete cds -130 -115 -028

Growth RAS-RELATED PROTEIN RAL-A -117 -097 -092

Growth Erythropoietin receptoi -113 -001 251

Growth STRESS-ACTIVATED PROTEIN KINASE JNK1 -072 031 042

Growth SWI/SNF complex 60 kda subunit -072 026 131

Growth V-ski avian sarcoma viral oncogene homolog -041 016 193

Growth Fibroblast growth factor receptor 3 (achondroplasia, -007 080 294 thanatrophoric dwarfism)

Growth V-myb avian myeloblastosis viral oncogene homolog-hke 2 049 097 257

Other Ribosomal protein L21 -212 -111 -071

Other INTERFERON GAMMA UP-REGULATED 1-5111 PROTEIN -136 -068 174 PRECURSOR

Other HISTONE HID -119 -113 -067

Other Human histone H2B 1 mRNA, 3' end -107 -086 -046

Other Cytochrome P450 IB1 (dioxin-inducible) -100 -071 -061

Other ATP synthase, H+ transporting, mitochondrial FO-complex -100 -055 -023

Other IMMEDIATE-EARLY RESPONSE PROTEIN NOT -096 079 197

Other Spermidine/spermine Nl-acetyltransferase -095 -063 137

Other Geneial transcription factor IIB -089 -080 -065

Other GATA-binding protein 1 (globin transcription factor 1) -016 113 163

Other TUMOR-ASSOCIATED ANTIGEN L6 007 034 056

These genes/ESTs exhibited a constant increase in expression ratio across the time course. Reported values are the change in expression ratio expressed as the number of standard deviations relative to the mean (no change in expression).

Tables II and III list genes where expression exhibited a unidirectional trend across a time course of 1 to 12 hours upon exposure to the anticancer drug Doxorubicin. For each gene, the reported value is the number of standard deviations above or below the mean that the expression ratio for that gene fell within the distribution of expression ratios for all genes at that time point. (The mean of this distribution was a ratio of 1 for each of the time points in Tables I, II, and III, e.g., on average the expression ratio of most genes does not change.).

The genes hsted in Tables II and III were identified based on the constraint that the changes from time point to time point must be in the same direction. Using the reported standard deviation at each time point to estimate the chance occurrence of the inclusion of the gene in the unidirectional trend, the random inclusion of a gene would be predicted by chance to occur less than once for a gene set of this size. Following exposure to Doxorubicin the genes most affected could be categorized by their relation to apoptosis, cell cycle, growth, or DNA damage with the greatest numbers of genes related to damage and growth. Doxorubicin is an anthracycline antibiotic which binds tightly to DNA and can lead to inhibition of DNA- directed synthesis of DNA and RNA and, subsequently, of protein. It also inhibits the repair of double stranded breaks mediated by topoisomerase-II. Doxorubicin acts nonspecifically within the genome and thus, it is not surprising that multiple changes are observed.

The data shown in Tables I-III illustrate the utility of the present invention on a small scale. The microarrays tested in these experiments consisted of several hundred genes. Future experiments will utilize microarrays comprised of 1000's of genes. Importantly, the illustrative results shown in Tables I-III provide useful characterizations of the effects of the drugs or compounds tested. In all three of these experiments, a number of genes showed measurable, statistically relevant responses. Furthermore, these gene expression responses could be categorized into groups related to apoptosis, cell cycle control, growth, or redox control. These gene expression responses could then serve as the expression patterns against which to compare additional experimental agents/compounds. Additionally, a test compound which produces expression pattern changes across numerous and diverse categories of molecular processes may be eliminated from further study, since broad effects indicate the likelihood of side-effects. Thus, these examples, while not meant to be limiting, illustrate the power and selectivity of the present invention in determining drug responses in complex mammalian systems.

Microarray technology allows for the determination of which genes are expressed in a particular cell type. Knowing when and which genes are expressed in a particular cell type is important in understanding key molecular events such as cell growth, cell cycle changes, and responses to environmental change. This understanding is important because even though most cells in our body contain the same genes, not all of these genes are utilized and thus "turned on" in each cell at any particular time. For example, some genes are turned on or expressed only when needed. Some genes are only turned on in specific cell types (e.g., pancreas cells are the only ones which express genes for producing insulin). In order to understand these expression patterns, it is necessary to identify which genes are turned on in different cell types.

The power of this technique can be more fully realized by considering that disease is often the result of inappropriately transcribed genes- either too much or too little transcription. Often, disease is a manifestation of regulatory genes that are deleted, inactivated, or become constitutively activated. For example, cancer often presents a particularly complicated case when similar clinical symptoms are often the result of heterogeneous changes. For instance, prostate cancer may be caused by several different, independently regulated genetic defects even in a single patient. In a group of prostate cancer patients, it is likely that every one of them may have a slightly different set of defective genes, each with differing implications for prognosis and treatment of the disease. It is the goal of the present invention to be able to determine the expression pattern for a particular cell type that is inhibited in some determined manner, and then to duplicate this expression pattern by treatment of the wild-type cells with a compound. Thus, a compound that produces a particular, desired expression pattern (for example, an analog of PX-12) can be identified and utilized as a therapeutic agent. This therapeutic agent would likely be void of side effects due to the careful characterization and analysis of its expression profile.

While the foregoing has been set forth in considerable detail, the sequences are presented for elucidation, and not limitation. Modifications and improvements, including equivalents, of the technology disclosed above which are within the purview and abilities of those in the art are included within the scope of the claims appended hereto. A further embodiment of the invention resides in a kit comprised of a microarray of substantially uncharacterized genetic material (e.g., cDNA) in a fixed pattern, the microarray capable of being hybridized with genetic material (e.g., cDNA) of a test cell; and an array pattern, the array pattern corresponding to a differential expression pattern of genetic material (e.g., cDNA) isolated from a control cell and a cell altered in a desired manner which has been hybridized with genetic material (e.g., cDNA) isolated from a reference cell. The kit may further include a test cell and means for isolating and labeling genetic material. The kit may also include means for hybridizing cDNA of the test cell to the microarray. The array pattern may a digital image (e.g., a picture of colored spots). As a matter of convenience, reagents employed in the present invention can be provided in the kit packaged in combination with predetermined amounts of reagents for use in determining and/or quantitating expression patterns. For example, a kit can comprise in packaged combination with other reagents any or all of the components described above depending on need. The test kits of the invention can further comprise any of the following components for additional convenience a positive control and/or a negative control. Additionally included in the kit may be a means for isolating and/or labeling of the genetic material, reagents associated with these techniques, as well as tubes or receptacles for preparation of the sample. Preferably included in the kits are means for lysing cells and extracting nucleic acid, etc.

It will be readily apparent to those skilled in the art that numerous modifications, alterations and changes can be made with respect to the specifics of the above description without departing from the inventive concept described herein. The embodiments described above are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is defined by the following claims rather than the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.

Claims

What is claimed is:

1. A method of identifying a drug candidate comprised of:

depositing reference cDNA from a cell line on a substrate to thereby form a template;

isolating and labeling cDNA from said reference cell to thereby form labeled control cDNA;

altering genetic material from said cell line to thereby form altered cDNA;

labeling said altered cDNA to thereby form labeled altered cDNA;

hybridizing said labeled control cDNA and said labeled altered cDNA to said reference cDNA to form a first hybridized pattern;

exposing a cell from said reference cell line to a potential drug candidate;

isolating and labeling cDNA from said cell exposed to said potential drug candidate to form a labeled test cDNA;

hybridizing said labeled control cDNA and said labeled test cDNA to said reference cDNA to form a second hybridized pattern; and

comparing said second hybridized pattern to said first hybridized pattern and selecting a drug candidate based on differences between said first and second hybridized patterns.

2. The method of claim 1, wherein said altered cDNA and control cDNA are hybridized to said reference cDNA in a sequential manner.

3. The method of claim 1, wherein said test cDNA is mixed with said control prior to contacting said reference cDNA.

4. The method of claim 1, wherein said reference cDNA is deposited in a fixed pattern.

5. The method of claim 1, wherein said cell is selected from the group consisting of a mammalian cell, a bacterial cell, and a yeast cell.

6. The method of claim 1, wherein said altered cDNA includes an alteration selected from the group consisting of a homozygous knockout, a heterozygous knockout, a dominant negative alteration, an antisense alteration, and a drug induced alteration.

7. The method of claim 1, wherein said potential drug candidate is an anticancer drug.

8. The method of claim 1, wherein a differential pattern of expression results from said step of comparing said first and second pattern of expression.

9. The method of claim 1, wherein said reference cDNA is substantially uncharacterized.

10. The method of claim 1, wherein the reference cDNA is completely uncharacterized.

11. A method of identifying a drug candidate, comprising: depositing reference genetic material from a cell line on a substrate to thereby form a template;

altering the cell line to form target genetic material;

labeling said target genetic material;

exposing a cell from said cell line to a potential drug candidate;

labeling genetic material isolated from the cell exposed to said potential drug candidate to form test genetic material;

hybridizing said target genetic material and said test genetic material to said reference genetic material to determine a pattern of expression; and selecting the drug candidate based on the similarity of the pattern of expression to a null pattern.

12. The method of claim 11, wherein said reference genetic material is randomly deposited in a fixed pattern.

13. The method of claim 11, wherein said reference genetic material is substantially uncharacterized.

14. The method of claim 11, wherein said target genetic material includes an alteration selected from the group consisting of a homozygous knockout, a heterozygous knockout, a dominant negative alteration, an antisense alteration, and a drug induced alteration.

15. The method of claim 11, wherein said target genetic material and said test genetic material are hybridized to the reference genetic material in a sequential manner.

16. A kit comprised of: a microarray of uncharacterized cDNA in a fixed pattern, said microarray capable of being hybridized; and

an array pattern, said array pattern corresponding to a differential expression pattern of cDNA isolated from a control cell and a cell altered in a desired manner which has been hybridized with cDNA isolated from a reference cell.

17. The kit of claim 16, further including a test cell.

18 The kit of claim 16, further including means for isolating and labeling genetic material.

19. The kit of claim 16 further including means for hybridizing cDNA of said test cell to the microarray.

20. A method of screening drug candidates comprised of:

providing a microarray of cDNA whose pattern is known to a first party, said microarray being capable of being hybridized with cDNA isolated from a cell treated with a drug candidate known to a second party; and

comparing an expression pattern of cDNA hybridized to said microarray with a known expression pattern.

21. The method of claim 20, wherein the identity of said candidate drug is known only to the second party.

22 The method of claim 20, wherein the known expression pattern corresponds to a known alteration in a cell line.

23. The method of claim 21, wherein the known expression pattern is known only to the first party.

24. The method of claim 22, wherein said alteration includes an alteration selected from the group consisting of a homozygous knockout, a heterozygous knockout, a dominant negative alteration, an antisense alteration, and a drug induced alteration.

25. The method of claim 20, wherein said cell is selected from the group consisting of a mammalian cell, a bacterial cell, and a yeast cell.

26. The method of claim 20, wherein the known expression pattern is selected from a library of expression patterns.