The IBC EuroBiochips 2001 conference(5–8 June 2001, Munich, Germany) reviewed the current state-of-the-artminiaturization and high-throughputtechnologies and comprised ~40 presen-tations, the majority of which was by European speakers. The conference focused on three main technologies: nucleic acid arrays, protein arrays andmicrofluidic technologies. The term‘biochip’, which was defined neither before nor during the conference, couldbe, and was, applied equally well to each of the technologies presented; and, despite the sometimes ambiguousdefinitions, all presenters agreed on the following: miniaturization is the key.

Why miniaturize?There are several obvious reasons: minia-turization allows the packing of morereagents into the same volume. For example: hundreds of different protein-affinity reagents on a chip instead of old-fashioned western blots and Sepharoseaffinity-columns; hundreds of thousandsof DNA hybridization reactions on a sin-gle glass-slide instead of old-fashionedSouthern and northern blots; and an integrated laboratory inside an ordinaryCD-ROM-like disk.

Other, not as obvious but equally im-portant, advantages of miniaturizationinclude the potential to increase reactionkinetics as a result of smaller reaction-volumes, and significantly increased sur-face:volume ratios1. In addition to theseadvantages, most of the microfluidic applications allow capillary forces to ma-nipulate the reagent solution, and wherecapillary force is insufficient for pumpingand mixing the reagents, spinning of thewhole CD-like laboratory will finish thejob! In terms of costs, miniaturization is

also attractive because less reagents areneeded for smaller size reactions. Capillarychannels the size of a single DNA mol-ecule can now be produced on a largescale. This could enable single-moleculeapplications to be carried out routinely.However, there is little data available yeton how DNA, and other large biologicalmolecules, would behave under suchconditions.

Nucleic-acid arraysOnly a few years ago DNA arrays seemedto be technology of the future; now re-searchers are frequently spotting DNAon slides and membranes. Although thepatent situation is complex, as was out-lined by Dominique Trosch and BernardGanahl (Reinhardt Sollner Ganahl,Kirchheim bei Munchen, Germany), themarket does appear open and usable.After an initial period, characterized byintensive speculations (illustrated by theabnormally high proportion of publishedreviews over research papers; Fig. 1),DNA microarrays almost became a commodity, but uptake and use of thistechnology in drug discovery has notbeen as fast as some DNA arraying companies would have liked.

There are two major types of arrays –oligonucleotide-based (short nucleic-acidfragments) and cDNA based (longerDNA-fragments). Both are successfullyused in industry (e.g. Affymetrix, SantaClara, CA, USA; and Incyte, Palo Alto,CA, USA) and in more academic do-it-yourself settings (e.g. catalogue and cus-tom arrays from MWG-Biotech, Edersberg,Germany; and Agilent Technologies).

Although the initial costs for apreprinted DNA chip might seem pro-hibitive, in comparison to the staff, equip-ment, information-technology supportand quality control required to manufac-ture them in-house, it is cost-effective tofirst know how many experiments are tobe performed. Two cheaper approachesdescribed at this conference are RNAChipTM and AtlasTM technologies fromClonTech Laboratories (Palo Alto, CA,USA). These relatively inexpensive ap-proaches provide a convenient disease-profiling technology-platform and couldbe used without large initial investmentsin expensive equipment.

Now that DNA chips have becomemore reproducible, efforts are being fo-cused on improving speed and sensitiv-ity, as highlighted by the advent of new

DDT Vol. 6, No. 15 August 2001 updateconferences

1359-6446/01/$ – see front matter ©2001 Elsevier Science Ltd. All rights reserved. PII: S1359-6446(01)01919-5 775

EuroBiochips: spot the difference!Mikhail Soloviev, Oxford GlycoSciences, The Forum, 86 Milton Park, Abingdon, Oxon, UK OX14 4RY, tel: +44 (0)1235 207673, fax: +44 (0)1235 207670, mikhail.soloviev@ogs.co.uk

Figure 1. DNAmicroarray publications(red). Value for 2001(blue) is an estimate,based on the first twoquarters of 2001.Proportion of reviews(green) among totalpublications on thesubject.

0

20

40

60

80

1994 1995 1996 1997Year

1998 1999 2000 2001

Drug Discovery Today

0

200

400

600

800

Num

ber

of p

ublic

atio

ns

Proportion of review

s (%)

assay techniques and bioinformatics.Zeptosens (Witterswil, Switzerland) havecollaborated with Qiagen (Chatsworth,CA, USA) to apply planar wave-guidetechnology to DNA chips. EMBL–EBI(European Molecular Biology Institute–European Bioinformatics Institute;Cambridge, UK) and IBM (White Plains,NY, USA) have embarked on data analy-sis systems to extract meaningful datafrom massive and complex data sets.With the current increase in use, DNA-array technology is expected to becomea routine application (such as DNA elec-trophoresis) in 2–3 years for HTS ofSNPs, mutational analysis and as a firststep in the screening of differentially expressed targets.

Protein microarrays – evolution orrevolution?The review by the chair of this session(Thomas Joos, Natural and MedicalSciences Institute (NMI); University ofTuebingen, Reutlingen, Germany) andthe first presentation by Hanno Langen(Roche, Basel, Switzerland) introducedthis session that discussed recently de-veloped technologies and those of thefuture, rather than protein-chip deriveddata.

To make an array of biological mol-ecules, the molecules either have to beavailable and purified (before spotting)or be synthesizable in situ. This task is rel-atively easy with nucleic acids, for whichmost of the sequencing information is available, and there are only four different, but chemically similar, bases.Furthermore, nucleic acids are not onlyinformation carriers, but also perfect af-finity reagents. For these reasons, nucleicacids have attracted a lot of attentionand are widely used in a variety of appli-cations, including microarray-basedtechniques. Millions of DNA fragmentscan be made cheaply and immobilizedto result in high-throughput, massivelyparallel diagnostic applications.

So, why bother with proteins? AlthoughDNA is an essential information carrier

and mRNA is an equally vital messengermolecule, the majority of biological in-teractions in our cells are performed byproteins. Genomic DNA can be analyzedfor single nucleotide polymorphisms(SNPs) and mutations and mRNA ex-pression levels can be analyzed in aquantitative manner by a variety of tech-niques (including arrays). However, acorrelation between the level of mRNAexpression and the expression and func-tion of respective proteins is limited.Protein turnover rate in the cell dependson the rate of protein synthesis (which is the only cellular process dependent on mRNA concentration) and proteindegradation. The concentration andfunction of proteins also depends on the immediate cellular environment and on post-translational modifications.Therefore, it is the level of protein expression and post-translational modi-fication that is a better marker of biological processes in a cell. This is why proteomics is a superior tool in disease diagnostics and drug-discoveryapplications.

To date, the use of two-dimensional(2D) gels coupled to MS has providedmost quantitative proteomics data (aspresented by Hanno Langen of Rocheand Jon Terrett of Oxford Glycosciences,Oxford, UK). This approach has resultedin massive databases of protein sequencesand their expression levels, and enablesthe selection of proteins or protein isoforms, which can be used in diseasediagnostics or other related applications.But why have protein arrays not progressed along with DNA arrays? The answer is because there are no completesets of affinity reagents with the requiredspecificities and affinities [althoughCiphergen (Fremont, CA, USA) showedsome progress with its non-selective protein-binding matrices using variousbinding conditions]. This is a result oftwo factors. First, the absence of infor-mation on exactly which protein andwhich post-translational modificationpattern is expressed differentially underparticular conditions. (Having access tosuch data enables the choice of relevantprotein targets.) Second, a protein array

update conferences DDT Vol. 6, No. 15 August 2001

776

Figure 2. Spot the difference: Typical output of a differential protein expression analysis.Two different cell lysate samples (a) and (b) were separately labelled and hybridized toidentical arrays of 48 different different affinity capture reagents (antibodies). This imageis reproduced, with permission, from the presentation by Oxford Glycosciences (UK) Ltd(Oxford, UK).

(a)

Drug Discovery Today

(b)

100%

0

100%

Sig

nal i

nten

sity

Sig

nal i

nten

sity

0

for quantitative proteomics is effectivelyan array of high-affinity reagents, for example, antibodies. The developmentof each antibody requires significantlymore time and resources than, for exam-ple, the synthesis of an oligonucleotideor purification of a PCR product.Therefore, it is unlikely that a generic setof affinity reagents against all proteins(even from one species) will be availablein the near future, thus, to date, publi-cations on protein-chip technology comprise more reviews than primary research papers.

However, this is not the only obstaclein making useful protein chips. DNA canbe dried, rehydrated, frozen and boiled,and even partially degraded without the loss of its information content (itsunique sequence identifier) or its high-affinity binding to its complementarystrand. Proteins, except in rare cases, willnot withstand such harsh treatment.Further, unlike nucleic acids, which canattach to the surface of a solid support in

any direction, protein-affinity reagentsmust be correctly oriented and activesites must be easily accessible. In addi-tion, the protein ‘hybridization’ stepscannot be carried out under denaturingconditions, unlike DNA hybridization,and this creates additional difficultieswith background false-positive staining.

Other important factors include thechoice of affinity templates [e.g. anti-bodies versus protein–RNA fusions (presented by Phylos; Lexington, MA,USA)] and the variety of assay optionssuch as quantitative proteomics (OxfordGlycosciences), protein–protein inter-actions and functional genomics and proteomics (Biacore, Uppsala, Sweden).

It was concluded, therefore, that thedevelopment of protein arrays requiresadditional efforts as well as the ‘know-how’ element. There is no doubt thatprotein-array based proteomics is advan-tageous over the traditional approachesbased on 2D gels and/or chromatogra-phy coupled with MS. Protein arrays can

enable the scale-up of analysis, both byrunning multiple affinity-recognitionsteps in a single chip and by reducingthe time of analysis (Fig. 2). Protein-affinity arrays should also improve reproducibility and enable fully quantita-tive protein-expression analysis. It is notknown, however, whether the develop-ment of protein arrays will follow theway paved by DNA-array developers, bycontinued improvements in the surface-immobilization techniques, further in-crease in the affinity and density of theimmobilized antibody and widespreadacceptance of these new technologies.Early evidence from this conference sug-gests much greater variability in the styleof protein arrays in all areas includingsurfaces, arrayers, affinity reagents, assays, analyses and array-readers.

Reference1 Ekins, R.P. and Chu, F.W. (1991) Multianalyte

microspot immunoassay – microanalytical‘compact disk’ of the future. Clin. Chem. 11,1955–1967

DDT Vol. 6, No. 15 August 2001 updateconferences

777

Drug Discovery Today will protect your identity!

Here is an unrivalled opportunity to get off your chest those things that really irritate you!

…and to be able to tell the relevant people what really irritates you withoutthem knowing it’s you!

Send in your letters and get some real debate going!

Please send all contributions to Dr Rebecca Lawrencee-mail: rebecca.lawrence@drugdiscoverytoday.com

Publication of letters is subject to editorial discretion