DDT Vol. 7, No. 22 November 2002 updateconference

1359-6446/02/$ – see front matter ©2002 Elsevier Science Ltd. All rights reserved. PII: S1359-6446(02)02506-0 1121www.drugdiscoverytoday.com

nanoparticles is related to toxicity. Pharm.Res. 16, 1836–1842

William M. PardridgeDepartment of Medicine

UCLA School of MedicineWarren Hall 13-164900 Veteran Avenue

Los Angeles, CA 90024, USAe-mail: wpardridge@mednet.ucla.edu

Genetic approach tochemical genetics ▼▼

Many drugs on the market today wereinitially identified because theydemonstrated activity in patients, animalmodels or cellular assays that werebelieved to be physiologically relevant.The advent of molecular biology allowedmore detailed dissection of biologicalprocesses into their componentbiochemical processes. Thus, clear linksbetween the physiological effects ofcompounds and their molecularmechanisms are relatively recent in thehistory of the pharmaceutical industry.Some of the clearest associations betweengenes and disease processes come fromanalyses of naturally arising mutations inthe human population. Only in very fewcases, however, is there a tight linkbetween a single gene and a particularhuman disease that identifies a cleartarget for pharmaceutical intervention.

There are a number of molecularbiology approaches that can help to

delineate the functions of individualgenes at the gene, transcript or proteinlevel. These include genetic tools such astransgenic animals where only one (or agroup) of genes are altered. Currenttechniques enable the expression of thegene to be restricted to specific tissues,and there is some degree of regulationpossible by exogenous intervention, butin the vast majority of transgenic animalsavailable today, this type of regulationwas not engineered in. Biochemicalapproaches such as RNA interference(RNAi) can be use to reduce specificprotein levels by interfering with thetranscript, but it is technically difficult toapply these approaches to a broad rangeof cells or in whole animals.Alternatively, dominant–negativemutants, either virally encoded or stablytransfected, can be used to probe genefunction by reducing protein activity.

In their recent paper, Skokat andVelleca [1] describe a hybrid of theseapproaches where they introduce amutant copy of a gene of interest thatpossesses wild-type activity but isengineered to bind to specific uniqueligands. Replacement of the wild-typecopy of this gene by homologousrecombination with this engineeredversion results in mice where theactivities of the encoded mutantproteins can be examined specifically bydosing with exogenous low molecularweight ligands. Because the engineered

proteins are wild type in every otherrespect, these recombinant mice provideexcellent models of what might beexpected if unique selective inhibitorswere developed as therapeutic agents.Not only can the overall physiologicalresponse to inhibition be measured, butmore detailed analyses can also beperformed; for example, assessingchanges in transcriptional and/orproteomic profiles or direct quantitationof downstream products.

As with any method designed to be asurrogate for the effects of a novel drug,there are drawbacks. Creating themutants, although no longerscientifically challenging, is technicallydemanding. In addition, the high degreeof selectivity that can be achieved withcompounds interacting with mutants isin many cases extremely difficult if notimpossible to achieve using drug-likemolecules designed to interact with theendogenous target. Despite theselimitations, Shokat and Velleca havedemonstrated the potential power thatcan result when genetics and chemistryare ingeniously combined.

Reference1 Shokat, K. and Velleca, M. (2002) Chemical

genetics: novel approaches to the discoveryof signal transduction inhibitors. DrugDiscovery Today 7, 872–879

Paul R. CaronVertex Pharmaceuticals

130 Waverly StCambridge, MA 02139, USA

Mining the human ‘kinome’David A. Dunn, Research Manager, Pharmacopeia, PO Box 5350, Princeton, NJ 08543, USA; tel: +1 609 452 3731; fax: +1 609 655 4187, e-mail: ddunn@pharmacop.com

Whenever there is a major advance inscience, new tools and paradigms changeand accelerate the pace of discovery. Thesequencing of the human genome hashad a major effect on the way we pursue

the discovery and development of newdrugs.

The now commonly used ‘omics’ ter-minology has come to refer to the para-digm shift that genomics has had on the

way we see biology. At the 2nd Inter-national IBC Conference on Protein Kinasesheld in Boston (MA, USA; 9–10 September2002), use of the terms ‘kinomics’ and‘phosphonomics’ reflect the impact that

genomics has had on the developmentof new drugs in this important target area.

The conference covered on-going research in the discovery of therapeuti-cally relevant pathways, technologies forscreening kinase targets in HTS, the de-velopment of leads and the testing ofthose leads in the clinic. This provided anexciting glimpse down a path that willhopefully lead to new drugs for treatingdiseases, such as cancer and inflammation.

Control of complex intra-cellularprocessesProtein kinases (PKs) are an importantclass of intracellular enzymes involved in the regulation of a large variety of biochemical pathways. Because of theirregulatory control, PKs have a major rolein human disease. More than 500 genesin the human genome code for PKs and∼ 400 diseases are associated with PK-mediated pathways. A large fraction ofthese are involved in cancer and inflam-mation.

In drug discovery, complex and inter-dependent regulatory pathways oftenserve as therapeutic targets. In his keynoteaddress, Lewis Cantly of Harvard MedicalSchool (http://www.hms.harvard.edu)described the signaling pathways con-trolled by phosphoinositide 3-kinase(PI3K). He showed how gene knockoutsand bioinformatics could be used tomap these pathways and their relation to disease. Stressing the importance ofbioinformatics, he described how ScanSite(http://scansite.mit.edu), a data miningtool co-developed by researchers at theMassachusetts Institute of Technology(http://med.mit.edu/) and HarvardMedical School, can help identify andrank structural motifs that confer selec-tivity for many PKs. Using ScanSite, tu-berin was identified as an AKT substratethat biochemically linked PI3K to tuber-ous sclerosis.

The need for new technologyThe demand to increase the magnitudeand comprehensiveness of ‘kinomic’ drug

discovery requires new technology. Topicscovered in the conference ranged fromcellular and biochemical tools to compu-tational approaches.

The architecture of the kinome is keyto understanding the cellular pathwaysinvolved in disease. We need new cell-based technology to map these path-ways. The Food and Drug Administartion(http://www.fda.gov) and the NationalCancer Institute (http://www.nci.nih.gov)have formed a new jont research programcalled the Clinical Proteomics Program,directed by Emmanuel Petricoin andLance Liotta. Petricoin and co-workershave taken an ambitious approach toprofiling pathways at the bedside. Theyhave developed laser capture micro-dis-section to excise healthy and diseasedcells from patient biopsies. Protein ex-pression can be probed directly and dis-ease profiles can be derived. More than130 proteins that are associated with numerous cancers, including PKs, havebeen identified. Changes in protein ex-pression can be followed over the courseof treatment. This work will impact thedevelopment of individualized therapiesfor the treatment of cancer.

Kinexis (http://www.kinexis.ca) usesimmunoblotting to enable comprehen-sive analysis of the expression patterns of >250 known signaling proteins. Thecompany has evaluated >1400 antibodiesin >4000 immunoblots. Steven Pelech,President and CEO of Kinexis, describedhow this information is used to developanalytical tools to provide a ‘bar-codereadout’ to quantify phosphorylationand decipher kinase pathways.

A dominant theme of the conferencewas the importance of structural knowl-edge in the development of potent andspecific inhibitors. BioFocus (http://www.biofocus.co.uk) combines ligand SARand structural knowledge of the enzymeto develop ‘family–activity relationships’,where SAR is understood in the context ofspecific families of targets. Brad Sherborne,Head of Computational Chemistry atBioFocus, reviewed the development of

several focused combinatorial librariesbased on amidine functionalized hetero-cyclic core-structures. Using CDK2 as anexample, he showed how the approachcould be used to identify potent and selective chemotypes in support of ‘hit-to-lead’ development programs.

Cellular and computational approachesare not the only ways to move kinase tar-gets through the discovery and develop-ment pipeline. New biochemical ap-proaches to screening kinase targets arebeing developed. Miniaturization in theHTS laboratory is becoming crucial to efficiently screen the growing number oftargets. Development of miniaturizedHTS assays must be simplified. Homo-geneous assays like fluorescence polar-ization (FP) and time resolved fluorescenceresonance energy transfer (TR-FRET) arepopular. Companies such as Promega(http://www.promega.com), MolecularDevices (http://www.moleculardevices.com) and Panvera (http://www.panvera.com) offer kits that are suitable forscreening kinases using FP.

New technology is being developed to address specific problem areas. BillRicketts from Ribapharm (http://www.ribapharm.com) described his evaluationof a new approach to screening serineand threonine kinases using fluorogenicsubstrates developed by Promega thatdoes not require substrate-specific anti-bodies. Luis Stancato of Eli Lilly (http://www.lilly.co.uk) described the Calipermobility-shift assay technology.

Steven Anderson of Abbott Labo-ratories (http://www.abbott.com) re-ported on the development of µARCS(Microarrayed Compound Screening) forscreening kinases. The technology usescompounds arrayed at high density onplastic sheets that are transferred to gelscontaining assay reagents. Anderson alsodescribed ASMS (Affinity Selection MassSpectrometry). This approach is used toidentify compounds that bind directly totyrosine, serine and/or threonine kinases.

Detailed knowledge of enzyme structurecan be indispensable to a development

update conference DDT Vol. 7, No. 22 November 2002

1359-6446/02/$ – see front matter ©2002 Elsevier Science Ltd. All rights reserved. PII: S1359-6446(02)02509-61122 www.drugdiscoverytoday.com

DDT Vol. 7, No. 22 November 2002 updateconference

1123www.drugdiscoverytoday.com

program. Syrrx (http://www.syrrx.com)uses nano-volume crystallization to sim-plify and speed the production of pro-tein crystals that can be used to solve the3D structures of proteins in importanttarget classes such as kinases. Twelvenovel structures have been solved thisyear and the company is targeting an-other 13 by the end of 2002. DominiquePerrin of Serono (http://www.serono.com)described the use of discrete sub-structuralanalysis (DSA) to identify chemical deter-minants related to compound activity.DSA helped Serono complete a hit–leadprogram for JNK within 90 days.

Into the clinicNearly every major pharmaceutical andbiotechnology company has compoundsin their development pipelines that arefocused on kinase-related diseases. Ac-cording to Jeffrey Hanke of AstraZeneca(http://www.astrazeneca.com), there are∼ 50 programs in clinical testing. Many ofthese programs are targeted toward ty-rosine kinases, which are often implicatedin cancer. Major tyrosine kinase targetsinclude EGFR and VEGFR (cancer) and P38MAP kinase (inflammation) and PDGFR.

For PKs, selectivity is an important de-terminant for a viable drug candidate.There can be as many as 200–300 pro-tein kinases mediating a large number ofpathways, for example in a single cell.Action at a specific pathway requires tar-get selectively; this can depend on themechanism-of-action of the drug. Drugsthat inhibit the enzyme by competingwith substrate binding are often selec-tive, although selective ATP-inhibitorshave also been identified. Chemistry basedon substituted purines and pyrimidinesare common, as would be expected forinhibitors of ATP binding. Compoundsdescribed by Pfizer (http://www.pfizer.com), Scios (http://www.sciosinc.com),Vertex (http://www.vpharm.com), Boehr-inger Ingelheim (http://www.boehringer-ingelheim.com), and Johnson & Johnson(http://www.jnjpharmarnd.com), use dif-ferent classes of core structures.

P38 MAP kinase is associated with sev-eral inflammatory diseases, includingrheumatoid arthritis. Jeffery Madwed of Boehringer Ingleheim showed thatBIRB796 inhibits the enzyme with a po-tency of 24 nM and is 10,000-fold selec-tive over a panel of kinases, includingERK, another member of the MAP kinasefamily. BIRB796 inhibits P38 via a novelmechanism involving two distinct sitesthat interferes with the binding of ATP.Preclinical results show that the com-pound inhibits production of tumournecrosis factor-α (TNFα), and preventedthe progression of collagen induced arthri-tis in a mouse model. Phase II studies arecurrently under way.

Scott Wadsworth of Johnson & Johnsonshowed that small-molecule inhibitors of P38 MAP could be used for noveltherapeutic uses, including allergy and HIV infection. Robert Kaufman of Vertex described how genetic and structural information has been used to take a‘chemogenomics’ approach to the dis-covery of potent and highly selective inhibitors, such as VX745. David Liu of Scios described how they have usedhuman and rat cDNA microarrays toidentify novel pathways associated withdisease. This approach was used to iden-tify ‘cross-talk’ between P38 and TGF-βsignaling pathways.

MEK is a member of the MAP kinasefamily and is involved in tumour prolifer-ation. Haile Tecle (Pfizer) showed resultsof the development of a series of com-pounds with activity against MEK. CI1040is extremely selective (>10,000-fold)against a large panel of kinases includingP38. Inhibition is not dependent on

ATP concentration, implying a substratecompetitive binding mechanism. CI1040is currently in Phase II clinical trials.

EGFR is a receptor tyrosine kinase (TK)that mediates several downstream signal-ing pathways that control cellular prolif-eration and survival. EGFR-TK is associatedwith a large number of cancers, includingprostate, lung colon and pancreatic can-cer. There are several programs activelypursuing the development of EGFR-TKinhibitors. Recently, the FDA has putIressa (AstraZeneca) on an acceleratedapproval track for the treatment of non-small cell lung carcinoma. Jeffery Hankedescribed the use of a structure-basedapproach to find selective inhibitors ofATP binding. Tarceva™ is another EGFRinhibitor being developed by Hoffman-La Roche (http://www.roche/com) incollaboration with OSI Pharmaceuticalsand Genentech (http://www.gene.com).Tarceva™, like Iressa™, acts to block ATPbinding to EGFR. Novartis has an orallybioavailable drug PKI166 in Phase IB tri-als; this is a dual inhibitor of EGFR –Her2tyrosine kinases.

ConclusionThis is the second conference on proteinkinases sponsored by the IBC this year.Both conferences were extremely wellattended, each with >250 delegates. Theconference presented a broad and com-prehensive view of the current status ofdrug discovery in the area of protein ki-nases. This is, and will continue to be, anactive area in pharmaceutical research.Based on the broad role that PKs have indisease, we should expect more activityin this area in the near future.

View the editors’ choice of BMN news and reviews at our Gateway sites:

http://bmn.com/microbiology http://bmn.com/immunologyhttp://bmn.com/cancerbiology http://bmn.com/neurosciencehttp://bmn.com/genomics http://bmn.com/biochemistryhttp://bmn.com/cellbiology http://bmn.com/geneticshttp://bmn.com/pharmacology