REVIEW

Microfluidics in macro-biomolecules analysis: macro insidein a nano world

Iuliana Oita & Hadewych Halewyck & Bert Thys &Bart Rombaut & Yvan Vander Heyden &Debby Mangelings

Received: 27 February 2010 /Revised: 13 May 2010 /Accepted: 18 May 2010 /Published online: 13 June 2010# Springer-Verlag 2010

Abstract Use of microfluidic devices in the life sciencesand medicine has created the possibility of performinginvestigations at the molecular level. Moreover, micro-fluidic devices are also part of the technological frameworkthat has enabled a new type of scientific information to berevealed, i.e. that based on intensive screening of completesets of gene and protein sequences. A deeper bioanalyticalperspective may provide quantitative and qualitative tools,enabling study of various diseases and, eventually, mayoffer support for the development of accurate and reliablemethods for clinical assessment. This would open the wayto molecule-based diagnostics, i.e. establish accuratediagnosis and disease prognosis based on identificationand/or quantification of biomacromolecules, for exampleproteins or nucleic acids. Finally, the development ofdisposable and portable devices for molecule-based diag-nosis would provide the perfect translation of the sciencebehind life-science research into practical applicationsdedicated to patients and health practitioners. This reviewprovides an analytical perspective of the impact of micro-fluidics on the detection and characterization of bio-macromolecules involved in pathological processes. The

main features of molecule-based diagnostics and thespecific requirements for the diagnostic devices are dis-cussed. Further, the techniques currently used for testingbio-macromolecules for potential diagnostic purposes areidentified, emphasizing the newest developments. Subse-quently, the challenges of this type of application and thestatus of commercially available devices are highlighted,and future trends are noted.

Keywords Microfluidics . Diagnostics . Proteins . Nucleicacids . PCR . Virus

Introduction

The technological advances of the last century enabledscientific clarification of several major mysteries of life, i.e.all organisms are made of cells, which are chemical systemscomposed mainly of complex carbon chains and sharing thesame information system [1]. A deeper insight into thecells, on the nanometre scale, highlights the main compo-nents: nucleic acids—the source of genetic information—and proteins—the main executive molecules. Elucidation ofthe way in which sub-cellular components work together toform functional cells and organisms will enable completeunderstanding of cellular processes and the way the cellresponds to the environment [1]. Eventually, the decipher-ing of cellular processes will help us to understand diseasemechanisms, and to find the proper way to diagnose andcure them.

Fast assessment of bio-macromolecules such as proteins,peptides, over/under expression of gene markers, and genemutations can be extremely valuable information fordiagnosis and prognosis of several pathologies [2, 3]. Forexample, the presence of various cancers and diseases is

I. Oita :Y. Vander Heyden :D. Mangelings (*)Department of Analytical Chemistry and PharmaceuticalTechnology, Center for Pharmaceutical Research (CePhaR),Vrije Universiteit Brussel-VUB,Laarbeeklaan 103,Brussels 1090, Belgiume-mail: debby.mangelings@vub.ac.be

H. Halewyck :B. Thys :B. RombautDepartment of Pharmaceutical Biotechnology & MolecularBiology, Center for Pharmaceutical Research (CePhaR),Vrije Universiteit Brussel-VUB,Laarbeeklaan 103,Brussels 1090, Belgium

Anal Bioanal Chem (2010) 398:239–264DOI 10.1007/s00216-010-3857-7

sometimes linked to abnormal concentrations of specificproteins [4]. Also, investigation of nucleic acid sequences,especially for identification of a genetic mutation, is apractical approach used to identify or confirm differentpathologies [4, 5]. In infectious diseases, the main cause ofmortality in developing countries and one of the majorcauses in developed countries [6], the pathogenic sourcebecomes even more traceable and is perfect candidate formolecule-based diagnostics. In all these situations, complexsamples, available in small amounts, have to be processedrapidly, preferably near the patient’s bedside, and aclinically significant response has to be obtained. In thiscase, the medical response can be adjusted more rapidly tothe patient’s reaction, enabling a personalized medicalapproach.

For years, consistent efforts have been made to developanalytical applications enabling fast, accurate, precise, andreproducible insights into the world of macro-biomolecules.Pandora’s box in the biosciences was opened around the1990s, when advances in incremental technology, theminiaturization boom, and progress in engineering emergedin microfluidic devices [7]. Microfluidics can be regardedas a framework, an enabling technology [8] related to fluidflowing in channels of micro or nano-size, offering thepossibility of developing products with better performanceand additional features (Fig. 1). The dawn of expectationsstarted with the concept of the micro-total analytical system(μTAS) [9]. μTAS was supposed to perform automaticsampling, sample transport, any necessary chemical reac-tions, and detection on a single, miniaturized, platform [9].The concept was declared the state-of-art strategy, becauseof its amazing advantages, i.e. faster separations, shortertransport time, lower sample and reagent consumption andthe possibility of multi-component testing under the sameconditions [9].

Since the concept of μTAS was announced, there was anexplosive interest on the topic, as shown by the huge amountof papers published so far (7745 accordingly to Scopusdatabase, December 2009), of which almost a fifth is related tothe life sciences and medical applications (Fig. 2). Over 70%of all papers on the use of microfluidic devices in the lifesciences and medicine are targeted on the characterization ofbiological systems at the molecular level (Fig. 2). The use ofmicrofluidics was perceived as a possible new dimension inbioanalysis. Microfluidic device-based applications shiftedthe perspective from the general to the cellular and sub-cellular level and enabled the measurement of molecularvariation, the dynamics of seconds-long processes, and bio-macromolecular motion [10]. Valuable bio-molecules wereisolated, characterized, quantified, and used to exploreinteractions with other molecules.

Once the euphoria generated by this promising newsolution for old unsolved problems became transformed

into a chase for application development, a long list ofpractical problems was revealed. Because a completely newplatform was to be developed, choice of the material andthe design of the device were the first problems to occur.Further, applications development revealed that the tech-nology required for production of fully integrated deviceswas not yet mastered. Also, solutions were needed for theextremely sensitive and miniaturized detection systems,while the complexity of samples emphasized the need forsample clean-up or enrichment of analytes of interest. Sometechnical problems have already been answered, or at leastbetter understood, while several are still open challenges forthe technology available.

This review tries to give an overview, from an analyticalperspective, of the impact of microfluidics on the detectionand characterization of the bio-macromolecules involved inpathological processes, focussing especially on those with ahigh potential to be developed as diagnostic devices forinfectious diseases and cancer. Given the importance of thefield, in which volume of literature doubles every fouryears, manuscripts mainly published during the last threeyears are discussed.

We will first review the main requirements for develop-ing diagnostically relevant applications. Further, on-chipsample treatment, on-chip PCR, separation, and immunoaf-finity techniques, and detection schemes will be identified.Subsequently, the challenges of this type of application andthe status of commercially available devices will bediscussed. We will conclude with future opportunities ofthe research. Throughout the review, examples of thenewest research, promising approaches, and opportunitieswill be emphasized.

Manuscripts reporting the use of microfluidics for micro-vascular network chips, the isolation of biomacromoleculeson chips, and applications related to cells other than pathogensare beyond the scope of this review. High-densitymicroarrays, although a source of clinically valuableinformation, are also not included because of disadvan-tages such as complexity, high cost, lack of robustness,and difficulty of interpretation.

Main requirements for developing diagnostic relevantapplications

An impressive amount of research has focused on developingapplications that can help medical practitioners achieve fasterand more accurate diagnosis, reliably assess disease progno-sis, or monitor treatment, on a solid quantitative basis [11].The final objective would be the development of portableautomatic devices able to provide fast laboratory graderesults without the need for special reagents. The devicesshould be able to provide results not only for patient bedside

240 I. Oita et al.

use but also in major public health threats such as pandemicsrisk and suspicions of biowarfare agents use [12]. Theapplication of μTAS concept would fit perfectly to such anapplication—a sample of bodily fluid, for example blood,urine, or nasal secretion, is collected and is introduced intothe workstation where minimal treatment is applied (e.g. ablood sample is diluted with EDTA to prevent clotting) andseparation is performed if necessary (e.g. plasma is separatedfrom whole blood or cells are lysed). Furthermore, multipleanalytes are then captured at the receptor site. After washingand introduction of secondary reagents, analyte levels areread using the workstation read out [13].

During the development of molecular miniaturizeddiagnostics, bioanalysis should provide two solutions:

1. extraction of the analyte of interest from the sample,and

2. conversion of analyte properties into a readable signal.

Finding these solutions would be equivalent to translationof the science into practical applications dedicated to massconsumers and health practitioners.

The technology available today has enabled the develop-ment of a number of biosensors for cancer biomarkersanalysis, whereas most of the multi-array sensor chips fortesting at or near the patient bedside are still in development orresearch stage [3]. Unfortunately, the use of proteins for

diagnostic tests is limited by current detection methods,which are only sensitive enough when the disease issignificantly advanced and protein concentrations havealready reached critical thresholds [4]. The development ofmicrofluidic devices enabling differential proteomic profilingand detection of a panel of several proteins has the potentialto revolutionize the biomedical research and to increase thesensitivity of tests [14]. Viral detection is another field wherethe use of microfluidic devices has the potential to improvedetection limits, simplify procedures, and reduce the timeneeded for confirmation of a viral infection [15].

Current techniques used to study and/or identify macro-biomolecules for diagnostic relevant applications

Several macro-biomolecules are routinely tested for diag-nostic purposes using a number of classical molecularbiology tools, for example slab gel electrophoresis orenzyme-linked immunosorbent assay (ELISA). Usually,the methods involve multiple manual steps, with severalcritical steps and long incubation times, and use largevolumes of buffers and expensive reagents available inminute amounts (antibodies). For a more precise assess-ment, blotting can be performed by transferring the bandsfrom the slab gel electropherogram to a nitrocellulose or

a b c

d e

f

Fig. 1 Several commercially available microfluidics-based devicesused for bio-analytical purposes a. Dynaflow System for ion-channeldrug discovery (Cellectricon, Mölndal, Sweden); b. LC–MSmicrofluidics-based chip (Agilent, Santa Clara, CA, USA); c. Nano-titer plates (microfluidic ChipShop, Jena, Germany); d. 15-cyclescontinuous-flow polymerase chain reaction (PCR) chip (microfluidic

ChipShop); e. 96-sample Sentrix Array Matrix (top) and the multi-sample Sentix Bead Chips (Illumina, San Diego, CA, USA); f.Disposable chip for generation of picoliter-volume droplets. Eachdroplet is further used as a PCR reactor (RainDance Technologies,Lexington, MA, USA)

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 241

Nylon membrane using a high electric field. The membranecontaining the transferred bands is then incubated withfunctionalized antibodies for specific proteins. Detection isgenerally achieved using functionalized antibodies, eitherradioactively or fluorescently labelled or covalently boundto an enzyme the activity of which can be easily measured.

Nucleic acids have a huge advantage over proteins,because use of polymerase chain reaction (PCR) enablessignal increases through target-based amplification andreliable detection of just a few copies of nucleotidesequences [2, 16]. For proteins, enzyme-linked immuno-sorbent assay (ELISA) is the most used technique. Limitsof detection (LOD) for ELISA are in the picomolar range,but large volumes are needed and the technique is rathercomplicated involving target capturing by an antibody andsandwiching with a second antibody, which is alsoresponsible for signal generation [2].

Within this context, applications developed using themicrofluidic devices would bring portability, higher sensi-tivity, cost reduction, shorter analysis time, and lesslaboratory space consumption, overcoming most of theinconveniences of the classical tools of molecular biology[17].

In microfluidics, the reduction in size results in a highsurface area-to-volume ratio and surface effects becomeextremely prominent. The physics underlying microfluidicdevices has been excellently described by Squires andQuake [18]. The paper emphasizes the variety of phenom-ena and the manner in which they had been exploited todevelop microfluidic devices.

The initial promise of microfluidics—the developmentof a μTAS that would integrate all analytical operations ona single platform [9]—is only partially fulfilled at themoment. The number of manuscripts reporting the devel-opment of a complete application, integrating sampletreatment is equal to or below the number of manuscriptsreporting partial applications, for example the analysis of analready processed sample or the development of a methodof sample preparation. In this context the integration of thepre-separation sample processing is currently the weakestlink [16, 19]. Expectations have become more mature andrealistic and the solution has been found by changing theperception of μTAS: the omnipotent chip is now seen as amicrofluidics platform, composed of a set of combinablebuilding blocks, where each block performs a singleoperation [17]. Almost all devices reported in the literature

Chemical sciences Environmental sciences

Others

Health sciences

Device manufacturin and

theoretical aspects

Biochemistry Geneticsand Molecular BiologyMedicine

Neuroscience

Multidisciplinary

Immunology andMicrobiologyPharmacology Toxicologyand PharmaceuticsHealth Professions

0

50

100

150

200

250

300

350

400

1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

Nu

mb

er o

f p

aper

s

a

b

Fig. 2 Characterization of thescientific literature available onmicrofluidic devices. a. Yearlydynamics of the number ofpublished papers b. Domains ofinterest for microfluidic devicesresearch

242 I. Oita et al.

can be further integrated in such a microfluidics platform.This approach enables the implementation of bio-analyticalassay in a better, foreseeable and less risky manner [17].The development of single-unit blocks avoids conflictsbetween the technical requirements for various operations;for example, in sample treatment high throughput is themain demand whereas for analytical separations highresolution is most important [16].

The transport of samples and reagents in microfluidicchannels is performed by routine approaches for separationscience, for example high pressure, vacuum, or electricalfield. The last is preferred for most applications, eitheruniform, as in electrophoretic separations, or non-uniform,as in dielectrophoresis in which a force is exerted on adielectric particle subjected to a non-uniform electric field.Dielectrophoresis is mostly used for analytes such asdielectric particles, without the need for them to be charged.The selectivity of dielectrophoresis can be easily tuned byaltering the field frequency. Other researchers are trying touse the magnetic field for controlled transport of paramag-netic particles [20]. An alternative lab-on-a-chip technologyis droplet technology, also called digital microfluidics. Inthis case, the flow is not continuous but as discrete droplets.The devices operate similarly to bench-top equipment, onlywith a significant volume reduction to the nL range andmore automation. Generation and manipulation of dropletsare performed in accordance with three main principles—electrowetting, dielectrophoresis, and immiscible-fluidflows. Detailed characterization of different microfluidicplatforms and recommendations for selection of the mostappropriate approach based on application can be foundelsewhere [17].

For analytical purposes, two main approaches are usedfor separation of the analyte of interest, from the samples—affinity-based separation and capture, based for exampleantigen–antibody reactions, and physicochemical-basedseparations, for example capillary electrophoresis (CE) orliquid chromatography (LC).

Sampling and sample treatment

Sampling and even minimal sample preparation, forexample separating plasma from whole blood, are per-formed off-microfluidic device. The main interests in usingmicrofluidic devices for sample treatment are related to theisolation and/or concentration of analytes or the removal ofinterfering components. Therefore methods involving con-tinuous phase separations should be favoured, because theyhave no need for careful sample loading [16]. However, theprocessing of larger sample volumes (sometimes evenhundreds of microliters) is still a problem.

Antigen–antibody and streptavidin–biotin affinity are thefavourite approaches for isolating target analytes. The

isolations are performed within a magnetic field usingmagnetic beads or in electrical fields using polystyrenebeads (Table 1). Sometimes, the successful isolationdescribed in the literature is more a proof of concept shownfor standard solutions and not yet tested on real clinicalsamples (Table 1). A trial for alternative ways to avoid theuse of antibodies [31] has been reported, but in some casesthe use of immobilized antibodies can result in extremelysensitive devices [32–35]. A very interesting example, withan intended use for single nucleotide polymorphismdiagnosis, is presented in Fig. 3. Dielectrophoresis [36],solid-phase extraction (SPE) [21, 37–41] or size-exclusionbased separations with nanopores [42] and microfabricatedplastic membranes [43] are other possible methods reportedfor target isolation.

Isolation and concentration is particularly important forproteins, usually present in complex mixtures (for instancein serum there are over 10000 different types) and in highdynamic ranges (the protein of interest is present atpicomolar levels, while less interesting proteins can bepresent at 30–50 gL−1 concentrations). For the particularcase of nucleic acid isolation, commonly used procedureswith commercially available kits are cumbersome, andinclude a substantial number of manual steps. For example,some purification kits require approximately ten pipettingsteps, three mechanical mixing steps, six centrifugationsteps, and ten tube-transfer steps [44]. The commonautomation approaches for these tests are benchtop-dependent, and therefore not suited to emerging bedsideapplications that require compact, automated, and robustoperational settings [19]. Still, several successful attemptsto automate nucleic acids extraction using microfluidicdevices have been reported. A plastic chip has been usedfor viral RNA extraction from a lysate of mammalian cellsinfected with influenza A (H1N1) virus [40]. The RNAisolation was achieved by μSPE, by reversible binding ofthe nucleic acids to silica particles trapped in a porouspolymer monolith [40]. The procedure is extremely simpleand only requires 10 min. A solid-phase method to isolatePCR-amplifiable genomic DNA from chemically lysedblood cells has also been used [37]. In this case, a poroussilicon matrix integrated in the biochip was used to performthe DNA extraction in 20 min [37].

Blood was initially regarded as the main target for thedevelopment of diagnostic tests. More and more applicationsfocus on development of tests for less invasive and complexbody fluids, for example urine, saliva, or nasal secretions.However, whole blood and plasma samples remain a constanttarget for sample-preparation devices using DNA extraction[37, 45, 46], pathogen capturing [47], protein depletion [38,48], or plasma separation from whole blood [49].

Microfluidics may also offer solutions for manipulationof samples ranked as highly bio-hazardous. An interesting

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 243

Table 1 Selected on-chip sample treatment using conjugated beads for affinity-based isolation of the target analyte

No. Targetanalyte

Testedsample

Capturingtechnique

Microfluidicdevice

Processedsamplevolume

Total time(isolation+separation)

Ref.

1 Dengue viruses Clinical serum Antibody-conjugatedsuperparamagneticbead

Microfluidic systemwith three integratedfunctional devicesfor pumping,mixing, andseparation

25 μL 10 min [21]

2 West Nile virus Sera frominfected chicken

Microarray of probemolecules immobilizedon a semipermeable membranefollowed by extraction fromthe membrane usingfunctionalized magnetic beads

Electrophoretic flowcell

20 μL min−1 2–3 min [22]

3 Dengue virusserotype 2 andenterovirus (EV)71

Clinical serum Antibody-conjugatedsuperparamagnetic bead

PDMS chip integratingthe samplepurification/enrichment and RT-PCR diagnosis

20–100 μL min−1

60 min [23]

4 Cholera toxin subunitB (CTB)

Solution Antibody-conjugatedsuperparamagnetic bead

PDMS chip integratingthe samplepurification/enrichment anddetection

100 μL 60 min [24]

5 Peptides displayedon E. coli cells

A library of E.coli cells

Streptavidin-functionalizedpolystyrene microspheres

Continuous-flowmicrofluidic sortingdevice

>108 cells h−1 [25]

6 gDNA fromleukocytes

Human wholeblood sample

Antibody-conjugatedsuperparamagnetic bead

PDMS chip integratingthe leukocytespurification, DNAextraction and fastanalysis of geneticgene

200 μL 20 min [26]

7 Alpha-fetoprotein(AFP)

Spiked serum(0.1 μg mL−1)

Streptavidin-functionalizedpolystyrene microspheres

PDMS-glass hybridimmunoassaymicrochip

100 μL min−1 55 min [27]

8 AFP, CEA, and PSAantigen

Solution,10 ng mL−1

Streptavidin-functionalizedpolystyrene microspheres

Multiplex electro-immunosensingsystem PDMS-glasshybrid

90 μL 55 min [28]

9 Alpha-fetoprotein(AFP)

Solution 1–1,000 ng mL−1

Antibody-conjugatedsuperparamagnetic bead

PMMA chips 5 μL 20 min [29]

10 Antibodies associatedwith an infectionby the dengue virus(immunoglobulin G(IgG) andimmunoglobulin M(IgM)

Serum Virus-conjugatedsuperparamagnetic bead

Integrated chip iscomposed of threepolydimethylsiloxanelayers and one glasslayer

100 μL 30 min [30]

11 Pathogen-specificDNA

Whole bloodspiked withHepatitis Bvirus (HBV)and E. coli

Antibody-conjugatedsuperparamagnetic bead

Centrifugalmicrofluidics on apolymer based CDplatform

100 μL 12 min [46]

244 I. Oita et al.

application consists in the manipulation and disruption ofbacterial cells and viruses in a closed system and withminimum intervention from the operator [37, 39, 47, 50–52]. Dielectrophoresis (DEP) has been used for manipu-lation and disruption of cells of Bordella pertussiss, abacterial respiratory pathogen [36], and for capture andlysis of Vaccinia virus particles [51]. The disintegration ofthe virus outer layer was proved by revealing the damagedand exposed tubules networks by scanning electronmicroscopy (SEM) [51].

Rapid isolation and counting of human immunodeficien-cy virus (HIV) from only 10 μL unprocessed whole bloodhas been performed on a microfluidic platform [47]. Thechip surface was coated with anti-gp120 antibodies tocapture HIV by binding to the gp120-glycoprotein on thesurface of HIV envelope. The research group that devel-oped the on-chip isolation of HIV intends to develop arapid (

system, which is more sensitive, “user friendly”, and fasterthan PCR [55].

As proved also by clinical standards, PCR is an invaluabletool for nucleic acids analysis, especially in viral diseases andcancer diagnosis [56]. Analysis of heterogeneous nucleic acidmixtures is a source of important clinical information,especially in assessment of antiretroviral resistant mutationsor early cancer detection [57]. Individually, genotypicinvestigation of minority viral populations might be a sourceof valuable information such as the tendency to develop drugresistance.

Integration of PCR on microfluidic platforms reducescost by reducing reagent volumes to tens of nanolitres. Italso reduces complexity by integration of several assaysteps into a single device, and reduces the times requiredfor thermocycling [58]. In the current mode of operation ofPCR, sample processing, amplification, and analysis of thePCR mixtures are stand-alone operations.

The miniaturization of PCR has the potential to identifythe missing link that integrates sample processing withdownstream PCR and analysis of PCR mixtures [56]. Inthis mode, sample manipulation, with potential effects onthe measurement, can be tightly regulated and accountedfor. Moreover, avoiding conventional extraction and manip-ulation limits sample loss. The integration of theseoperations would be beneficial for characterization ofcancers at the molecular level, enabling meaningfulquantitative assessment of cancer pathogenesis and devel-opment of more effective therapy [56]. Miniaturization alsobrings other potential benefits, for example multi-paralleltreatment of a defined numbers of cells, single cell-basedanalysis, or even single gene analysis [56].

The feasibility of rapid, single-molecule amplification ofnucleic acids was established using a microfluidic systemdeveloped to enable rapid PCR analysis of individual DNAmolecules with precise temperature control [57]. PCR wasperformed on heterogeneous samples containing syntheticCYP2D6.6 wild-type and mutant templates, on a quartzchip designed to enable adequate mixing by Browniandiffusion after each stage of reagent dispensing. Sampleswere loaded from a microtitre plate on to the microchipthrough an integrated capillary, which minimizes thecontamination risk. The chip design allowed eight parallelPCR reactions. Thermocycling was performed by heatingwith nine embedded resistive heaters built from platinumtracers and cooling by recirculating water underneath thechip. The system had an impressive detection set-upcomprising two different lasers for excitation, i.e. a488 nm optically pumped solid-state laser and the 633-nmline of an HeNe laser, a series of dichroic mirrors and bandpass filters, and three charged-coupled device (CCD)cameras, each collecting light of a different wavelength,i.e. 515, 550, and 685 nm. Thermocouples and optical

DNA melt analysis were used to demonstrate the chip’sability to rapidly thermocycle. When the desired tempera-ture was 68.5 °C, seven out of the eight channels managedaccurate control within a 1° range. The efficiency ofamplification was proved by measuring the fluorescenceemission resulting from amplification of a 1:1 mixture ofwild and mutant templates of CYP2D6.6 using two Taqmanprobes. The procedure is fast, reproducible and able toamplify a heterogeneous sample containing two templateswithout mixing the amplification products between tem-plates. Use of the proposed microfluidic setup overcomesthe high reagent cost and cumbersome reaction assemblyrequirements of plate-based limiting dilution PCR methodsand the low throughput and manual handling of currentmicrochip methods. Almost 96 templates were analyzed in25 min by use of ∼1 μL PCR reaction volume. Comparedwith digital PCR this operation was performed almost fivetimes faster and using a PCR reaction volume a factor of 1400 lower [57]. Small channel-to-channel differences inamplification were noticed, but these are potentiallyreducible by refinement of the heating procedure.

When transfer-messenger (tm) RNA purification, nucleicacid sequence-based amplification (NASBA), and real-timedetection were integrated on a microfluidic device for thefirst time (Fig. 4), the chip was able to identify crude E. colibacterial lysates in less than 30 min [59]. Cell lysis wasperformed off-chip, using a commercially available device.tmRNA, from 100 lysed E. coli cells, was purified by SPEusing silica beads immobilized on the chip surface. Deviceclogging by debris, or bubble formation because of passageof air were avoided by immobilization of silica beads in athin layer that leaves enough free space within the channel.RNA was eluted in 5-μL fractions by flow of deionizedwater through the silica bed chamber. The amplificationused custom-designed high-selectivity primers and real-time detection was performed at 530 nm using molecularbeacon probes (oligonucleotides with an FAM fluorophoreat the 5′ end and a BHQ1 quencher at the 3′ end) [59].

An infrared temperature control system has been describedfor completely contactless temperature control PCR in micro-fluidic chips in a fluidic channel too small to enableconventional temperature control using thermocouple-basedsensing [60]. The system comprised an IR pyrometer sensingthe surface temperature above a PCR chamber. The design ofthe system ensured rapid equilibration between the PCRsolution and the chamber surface [60]. For non-contacttemperature control, the surface temperature relative to thatof the PCR solution temperature was calibrated using theboiling point of water and an azeotrope within the chip.Successful PCR of a fragment of a Bacillus anthracis genewas performed by use of the described system [60].

For the first time, reverse transcription PCR has beenperformed on single-copy viral RNA in monodisperse

246 I. Oita et al.

isolated pico-droplet reactors using a fused-silica chip withhydrophobic coating. The device was coupled with an offchip valving system for generation of mono-dispersedroplets with ∼70 nL volume [61]. RNA was isolated inmonodisperse picoliter droplets emulsified in oil. In eachdroplet, real-time reverse transcription PCR with fluores-cence detection of amplification was performed. Afterapproximately 23 amplification cycles, RNA from 0.05–47 plaque forming units (pfu)/droplet was detected by real-time fluorescence [61]. The use of microdroplet technologylimits the interaction between the microfluidic surface andPCR sample/reagents; this is responsible for PCR inhibitionand carry-over contamination.

Several other manuscripts (Table 2) have reported fullyintegrated devices with impressive lowest amplified con-centration, but a real μTAS is still in its research phase.

Separation techniques

On chip capillary electrophoresis (CE) is the most usedseparation technique. Detection is frequently achieved byfluorescence, although electrochemical methods involvingamperometric detection [73] and contactless conductivitydetection [74] have also been reported. Separations are per-formed on home-made, hybrid glass-polydimethylsiloxane(PDMS) [65, 73] chips or poly(methyl methacrylate)(PMMA) chips [74–77]. A few applications use commer-cially available devices [75, 78, 79].

CE on chip has been used to quantify PCR products [65,80], oxidative stress biomarkers in urine [73], hepatic cancerbiomarkers in spiked serum [77, 81, 82], thrombin—amarker for various haemostasis-related diseases and con-ditions—in diluted plasma [76], viruses, for example humanrhinovirus (HRV) or swine influenza virus [75, 78, 83],inflammatory biomarkers [84], and K-ras, the oncogene formutations closely associated with colorectal cancer [79](Table 3). The separation modes involved free-solutionelectrophoresis [73, 75, 77–79, 84], capillary gel electropho-

resis (CGE) [65, 73–75, 80, 82, 86, 87], or affinity capillaryelectrophoresis [42, 77, 78, 81, 88].

CGE has been performed to resolve and investigate theabundance of proteins in complex samples in order toidentify viruses and bacteriophages [89], to detect thepresence of food-borne pathogenic bacteria in decayed foodsamples [80], or to analyse of RNA–RNA interactions [90],as alternatives to cell culture or PCR.

Several manuscripts report the use of on-chip transientisotachophoresis (ITP) to achieve preconcentration ofanalytes [82, 85, 88]. Using ITP, human serum albumin(HSA) and its immunocomplex with a monoclonalantibody were preconcentrated 800-fold [85] and 2000-fold [88] online on standard cross-channel PMMA micro-chips. Another interesting application uses an integratednanoporous membrane to perform simultaneous concen-tration and detection of the inactivated swine influenzavirus. Detection was achieved by coupling with afluorescent labelled antibody. The fluorescent antibodycomplex was electrophoretically separated from the un-bound antibody in 6 min by use of less than 50 μL clinicalsample [83].

Intact protein separations from E. coli cell lysate havebeen performed on a two-dimensional microfluidic systemwith ten channels combining isoelectric focusing (IEF) andsodium dodecyl sulfate (SDS) polyacrylamide gel electro-phoresis (PAGE) [86]. Protein profiling was used forbacterial identification and characterization using a micro-chip separation platform [87]. The method is potentiallyuniversally applicable, especially for bio threat agents, andovercomes the disadvantages of PCR, i.e. high cost and theneed for special working procedures to avoid contamina-tion. The method includes four steps—the bacteria orspores are harvested and lysed, and the constituent proteinsare solubilized, labelled with a fluorescent tag, andanalyzed using chip gel electrophoresis (CGE). The proteinfingerprints from the model organisms were used forbacterial identification or characterization.

Chamber for future

application

RPC

RNA Purification Silica Bead Chamber

Input

NASBA Chamber

(NC)

NASBA Port

Waste Output

a bFig. 4 Integrated microfluidicdevice for RNA purificationand real-time NASBA. a Photo-graph of the device. Each chipcan perform two separate reac-tions with the same reagents, butdifferent samples, to incorporatecontrols. b Single-device archi-tecture showing the distinctfunctional microfluidic modules:RNA purification chamber(RPC) and real-time NASBAchamber. (Reproduced, withpermission, from Ref. [59])

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 247

Table 2 Selected on-chip PCR applications

Objective Chip architecture Amplified sequence Lowestconcentrationsuccessfullyamplified.

Detection of theamplified sequence

Ref.

Micro circulating PCRchip

Three bio-reactors withsuction-type membrane andthree microvalves operatingat three different temperatures

150 base pairs associatedwith the hepatitis C virus

102copies μL−1 Off line afterextraction fromopen reactionchambers afterfinishing the PCRprocedure

[62]

Concurrentelectrochemicaldetection

Eleven parallel channels 489-bp gene fragment On-line,electrochemical,square-wave vol-tammetry

[63]

Extraction of genomicDNA and detectionof single nucleotidepolymorphism

Three major modules for rapidpurification, DNA extractionand fast analysis of geneticgene

Genomic DNA from leukocytes 33.26±2.5 ngDNA μL−1

Off-line, opticalevaluationin UV

[64]

Integration of PCRand CE on a singleplatform

Tri-layered glass-PDMS withintegrated pneumatically-actuatedvalves and pumps for fluid han-dling, a thin-film resistive elementthat acts simultaneously as a heaterand a temperature sensor, andchannels for capillary electropho-resis (CE)

On-line, a laser diodeand a chargedcoupled device(CCD) camera

[65]

Detection of α-thalassemia-1 dele-tion using salivasamples

DNA extraction chamber, sampleloading chamber, waste collectionchamber, PCR reaction chambers

gDNA extracted from saliva 12.00 pg μL−1 Off-line, fluorescence,by an externaloptical detectionmodule

[66]

Device architecture forelectrochemicalpatterning anddetection of multipleDNA sequences

Amplicons diagnostic of human(H1N1) and avian (H5N1)influenza

400 nmol L−1 On-line integratedelectrochemicalarray

[67]

Unsealed reactors forreal-time isothermalhelicase-dependentamplification

An array of 4 unsealed reactorsfor real-time helicase-dependentamplification

BNI-1 fragment of SARScDNA

On-line fluorescence,CCD camera andimage analysis

[68]

Quantitative PCRsystem for DNAamplification anddetection

Two micro modules for thermaland microfluidic control withthree serpentine-shape micro-pumps

350 and 150-bp detection genesassociated with two viruses,specifically hepatitis B virus(HBV) and hepatitis C virus(HCV),

10 copies μL−1 On-line fluorescence [69]

Fungal pathogenicnucleic acid detection

Microfluidic microarray assemblydevice on a CD-like glass chip

Botrytis cinerea and Didymellabryoniae

0.5 nmol L−1 On-line, confocallaser fluorescentscanner followed byimage analysis

[70]

Single nucleotidepolymorphismgenotyping of PCRamplicons fromwhole blood

Thin film transistor photosensorintegrating a microfluidic channel,a DNA chip platform, and aphotodetector

Biotinylated target DNA 0.5 nmol L−1 On-linechemiluminescencephotodetector

[71]

Poly(methylmethacrylate)continuous-flow PCRmicrofluidic chip

Chip on the PMMA substrate with20 parallel channels

DNA template with a 990-basepair fragment of Pseudomonas

Off-line [72]

248 I. Oita et al.

Tab

le3

SelectedCEapplications

onachip

Analyte

Detectio

nPerform

ance

Ref.

RT-PCRmixture

obtained

afteram

plificationof

a23

4-base

pairRNA

isolated

from

amultip

lemielomacancer

line

Laser-ind

uced-fluorescence

Sufficientsensitivity

even

with

dram

atic

redu

ctionin

instrumentcostandcomplexity

;separatio

nresolutio

ncomparablewith

abenchtop

,commercially

availablesystem

;sign

al-to-no

iseratio

32.3;LOD

0.1

ngμL−1

[65]

8-Hyd

roxy

deox

yguano

sine

(8-O

H-dG)DNA

addu

ctin

urine(oxidativ

estress

biom

arker)

Electrochem

ically

(amperometricdetection)

andviascanning

electron

microscop

e(SEM)im

aging

LOD20

attomoles;rang

efrom

100nm

olL−1–150

μmol

L−1);separatio

nefficiencies

ofapprox

.12

0,00

0–17

0,00

0plates

m−1

[73]

DNA

Con

tactless

cond

uctiv

itymeasurement

Highsign

al-to-no

iseratio

;labelfree

detection

[74]

Hum

anrhinov

irus

serotype

2(H

RV2)

Fluorescence

Analysistim

e10

s;pu

rity

assessmentof

fractio

nscollected

from

size-exclusion

chromatog

raph

ypu

rificatio

nof

thelabelling

mixture

andmon

itoring

affinity

complex

form

ation

[75]

Throm

binlevelsin

plasmadilutedto

10%

(v/v)

Fluorescence

Lessthan

1min;run-to-run

andchip-to-chip

reprod

ucibility

(RSD)of

migratio

ntim

es<10

%;LOD

540

nmol

L−1

[76]

AFPfluo

rescently

labelledAFPin

spiked

serum

samples

Fluorescence

Quantitativ

eassay;either

themetho

dof

standard

additio

nor

acalib

ratio

ncurve;

AFP

atng

mL−1

levelsin

10μLhu

man

serum

inafew

tens

ofminutes

[77]

K-ras

oncogene

formutations

high

lyassociated

with

colorectal

cancer

Fluorescence

Analysistim

e11

min

usingtheCAEsystem

and85

sforPMMA

microchips

[79]

α-Fetop

rotein

(AFP)from

spiked

serum

samples

Laser-ind

uced-fluorescence(LIF)

Total

assaytim

e<10

min;LOD

0.1ng

mL−1;CV

<2%

;qu

antitationrang

efrom

24to

922ng

mL−1;go

odcorrelationof

testresults

for68

patient

serum

samples

with

acommercially

availablereferencemetho

d

[81]

Inactiv

ated

swineinfluenza

Fluorescence

Microchip-based

concentration;

separate

thevirus/fluo

rescentantib

odycomplex

from

theun

boun

dantib

odyelectrop

horetically;totalassaytim

e6min;<50

μLsample

[83]

Hum

anserum

albu

min

(HSA)andits

immun

ocom

plex

with

amon

oclonalantib

ody

Fluorescence

800-fold

sign

alenhancem

ent;LOD

7.5

pmol

L−1;analysistim

e25

s[85]

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 249

Microchip LC coupled with MS has been reported foridentification of autoantigens [91], defining a proteomesignature for invasive ductal breast carcinoma [92], andbiomarker screening applications in MCF-7 breast cancercellular extracts [14]. A method very applicable in a clinicalenvironment was also developed [92].

Immunoaffinity techniques

Affinity-based methods exploit the specific binding ofbiomolecules in order to isolate and characterize them inthe presence of thousands of other compounds [93]. Themain actors are antibodies, but aptamers, for example cell-surface receptors and oligonucleotides, are also valuablealternatives [94]. Low-molecular-mass aptamers have sev-eral advantages over antibodies—faster tissue penetration,longer shelf-life, sustaining reversible denaturation, lowertoxicity, the possibility of being produced against targetssuch as membrane proteins, and use of highly automatedtechnology [95].

A wide array of clinical diagnostic tests employsimmunoassays and immunoblotting approaches [96].Performing immunoassays on microfluidic devices has theadvantages of high throughput, short analysis time, smallvolume and high sensitivity, and fulfils most of theimportant criteria for clinical diagnoses [97]. The mainadvantages of microfluidics as enabling technology hasbeen clearly proved by several applications [77, 98, 99]with similar or improved performance compared withELISA but much simpler and faster and with lowervolumes. Future developments will increase even morethe advantages over ELISA, creating fully integrateddevices [100].

Analysis of inflammatory biomarkers, C-reactive pro-tein (CRP), prostate-specific antigen (PSA), argininevasopressin (AVP), alpha-fetoprotein (AFP) cancer cells,and infection markers are just some of the clinicallyimportant applications currently implemented on micro-fluidic immunoassay chips. A chip has been used forrapid isolation and quantification of inflammatory bio-markers in microdissected areas of a skin biopsy. Thebiomarkers were isolated by immunoaffinity capturewithin the extraction port of the chip by use of a panelof 12 antibodies immobilized on a disposable glass fibredisk [84].

A specific aptamer, i.e. an oligonucleotide, has beenused for highly selective capture and enrichment of argininevasopressin (AVP) and possible diagnosis of immunologi-cal shock or congestive heart failure based on AVPquantification [101]. Trace amounts of AVP were enriched,eluted isocratically using a microfluidic platform, anddetected label-free by coupled matrix-assisted laser desorp-tion/ionization mass spectrometry (MALDI-MS). The

aptamer–analyte binding was thermally disruptable en-abling easy device regeneration [101].

Affinity-based isolation has enabled the first steps fromthe bench top to the clinic in the development of aptamericbiosensors for cancer cells. Prostate tumour cells weresuccessfully isolated and identified on a PMMA microchip[32]. The procedure managed to discriminate rare circulat-ing prostate tumour cells resident in a peripheral bloodmatrix without staining, using antibodies and aptamers forprostate-specific membrane antigen (PSMA) immobilizedon the surface of a capture bed fixed within the chip [32].Similarly, the efficacy of on-chip recognition and capture ofbreast cancer cells using an antibody-based microfluidicdevice has been reported [34]. The procedure used abiochip etched on to PDMS with the inner surface of themicrochannels coated with epithelial membrane antigen(EMA) and epithelial growth factor receptor (EGFR) [34].

Applying the same principle but from a reversedperspective, virus-bound magnetic bead complexes havebeen used for rapid serological analysis of antibodiesassociated with an infection by the Dengue virus [30].Dengue virus infection was confirmed by use of a micro-fluidic system that integrated one-way micropumps, a four-membrane-type micromixer, two-way micropumps, and anon-chip microcoil array [30]. Detection is achieved usingfluorescence-labelled secondary antibodies. The procedurewas performed automatically on a single chip within30 min, which is a factor of eight faster than the traditionalmethod. Also, the LOD (21 pg) was reduced by a factor ofapproximately 38 compared with the traditional method[30].

C-reactive protein (CRP), a general inflammation andcardiovascular disease risk assessment marker, has beendetected in a one-step sandwich immunoassay using afluorescence microscope [35]. Sample collection andimmunoreaction were integrated on a microfluidic chip.CRP was detected from 5 μL human serum at concen-trations of 10 ng mL−1 in less than 3 min, and after 13 minfor concentrations below 1 ng mL−1 [35]. Anothermanuscript reported CRP assay in a simulated serum matrixby on-chip immunoaffinity chromatography [102]. CRPwas fluorescently labelled in a one-step reaction andinjected directly into the immunoaffinity capillary contain-ing monoclonal anti-CRP attached to a 5.0-μm streptavidin-coated silica bead. The limit of detection was 57.2 ng mL−1

and chromatographic run times were less than 10 min[102].

An important step in the miniaturization and integrationof laboratory operations in self-contained devices was madeby designing a microfluidic chip for combinatorial libraryscreening (Fig. 5) [25]. The chip was used to map acombinatorial peptide library of possible epitopes for anti-T7•tag antibody and anti-FLAG •tag antibody. The peptides

250 I. Oita et al.

were displayed on E. coli cells as insertions within anexternal loop of the outer membrane protein OmpX. Thebinding peptides were selected in several steps (Fig. 5). Inthe first step, the library was incubated off-chip with thetarget biotinylated antibodies. Cells with the binding peptideswere captured on streptavidin-functionalized 5.6-μm polysty-rene microspheres. In the second step, a disposable chip wasused for serial dieletrophoresis activated cell sorting toseparate cells with binding peptides from cells with non-binding peptides. The sorting was performed on the basisof size, in continuous-flow, with a sorting speed of morethan 108 cells h−1. The antibody-binding target cellscaptured on microspheres were funnelled dielectrophoreti-cally because of their different polarization [25]. In thethird step, the collected beads with attached cells weregrown overnight and the sorting in the second step wasrepeated. The DNA of the sorted cells was sequencedautomatically to determine the sequence of the bindingpeptides. Cell library screening using microfluidic sortingwas comparable with a combination of conventional cellsorting-methods, i.e. one round of sequential magneticselection and two rounds of fluorescence activated cellsorting. The microfluidic sorting chip [25] had increasedtolerance of flow disturbances and microbubbles and more

robust purity performance using high cell concentrationsat the inlet. The authors also mention a possibility ofincreased throughput by optimization of chip architecture,using parallel channels fabricated on a single chip. Thetechnique could be further developed to monitor poly-clonal signatures of serum antibodies for disease profiling,for other affinity-based screening using substrate-functionalized beads, and for automated reagent genera-tion, wherein ligands for a given protein or cell type couldbe discovered using self-regenerating libraries.

Prostate-specific antigen (PSA) has been rapidly andsensitively quantified in human serum samples using animmunosensor coupled to a glassy carbon electrode (GCE)modified with multiwall carbon nanotubes (MWCNT)(CNT–GCE) integrated with microfluidic systems [103].PSA was captured immunologically with the immobilizedanti-tPSA and horseradish peroxidase (HRP) enzyme-labelled second antibodies specific to PSA. The detectionrelies on back electrochemical reduction of 4-tert-butylca-techol catalyzed by HRP in the presence of hydrogenperoxide. The procedure had an LOD of 0.08 μg L−1 whenelectrochemical detection was used [103].

A new approach to immunoblotting combines on-chipintegration of polyacrylamide gel electrophoresis (PAGE)

Fig. 5 Schematic depiction ofantibody fingerprinting using amicrofluidic sorting device, cas-sette B (not to scale). Left:Micrograph of the sorting devicein operation showing the firstsorting stage. Right: Micrographof the second stage at the col-lection point. (Reproduced, withpermission, from Ref. [25])

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 251

with subsequent in-situ immunoblotting [98]. The manu-script reports “hands free” electrophoretic transfer ofresolved species to a blotting membrane as a directed,efficient method for protein identification without a needfor pressure-driven flow and valving [98]. alpha-Actininand PSA were identified and quantified from multi-proteinsamples in the 101–105nmol L−1 range with LODs of0.05 pg and 1.8 pg, respectively. Moreover, detectionsensitivity was enhanced approximately fivefold by targetprotein enrichment on the blotting membrane [98].

An interesting application of affinity-based methods isthe development of encoded microbeads. These are smartmicrostructures with both molecular recognition ability andbuild-in codes for rapid microbeads identification [104].The codes can be optical, electronic, graphical, or physicalsignatures creating combinations similar to the black andwhite stripes of traditional bar codes [105, 106]. The mostpopular approach uses combinations of quantum dots(QDs) to generate fluorescent codes. QDs are inorganicnano-crystals with fluorescent properties which depend onboth their composition and size [106]. They have severalamazing properties, for example high fluorescence yield,remarkable stability, and extremely narrow emission,enabling many non-overlapping colours to be used simul-taneously [106]. Moreover, the use of the encoded microbe-ads can add multiplexing capability to the assay byenabling multi-analyte analysis [97].

Detection schemes

Optical detection schemes are still the favourite choice formeasurements in microfluidic systems [107–109]. There aretwo major approaches—coupling of the macro-scale opticalinfrastructure as “off-chip approach” or integration of micro-optical functions on to microfluidic devices as “on-chipapproach” [110]. Several solutions for integration, forexample planar waveguides, coupling schemes to the outsideworld, evanescent-wave based detectors, and optical fluidicsintegration problems, and perspectives and limitations ofthese different solutions have been discussed in detail [107].Irrespective of the detection approach, the LODs achievedare profoundly affected by sensor size and shape, because ofanalyte transport limitations rather than signal transductionlimitations [110]. Production of inexpensive, sensitive, andportable optical detection systems is, therefore, currently ofmajor importance in the manufacture of commercial devicesfor portable diagnostic devices [108].

Laser-induced fluorescence (LIF) is the most populardetection technique in microchip-based applications. Be-cause the number of fluorescent analytes is rather limited, aderivatization step with a fluorescent dye is usuallyinvolved [42, 51, 65, 89, 111]. Unlike UV detection, an

alternative to “diode array detectors” enabling simultaneousdata acquisition at multiple wavelengths is not thataccessible for fluorescent detection. Fluorescent detectorsare bulky and expensive. In some cases, the ability tocombine the available detection technology with the mostappropriate label is a real challenge, but also the labellingitself is a major analytical burden. In many cases thedetection relies on measurement of fluorescence intensity,performed by capturing the signal with a CCD cameracoupled to a fluorescence microscope and followed bygraphical analysis of the captured images. For instance,CCD cameras have been used to monitor, capture, andspecifically isolate viruses and bacteria [112, 113], tocontrol DNA hybridization [114], to type Staphylococcusaureus strains [115], to quantify proteins after separation[98], or to screen and identify new serological biomarkersfor inflammatory bowel disease [62].

As fluorescent probes, fluorescein’s derivatives are stillextremely popular [89, 116–119], but the cyanine family(Cy)dyes [75, 120, 121] and Alexa Fluor type dyes [31, 33,120, 122, 123] are also used. In most immunoassays, thedetection problem is solved by using fluorescent labelledantibodies. The use of fluorophores with higher intensityyield, for example QDs, has enabled visualisation of HIVafter capture on a microfluidic device using only a standard10× fluorescence microscope [47]. Similarly, the use of QDprobes resulted in 30-fold signal amplification, whichimplied a reduction in observed limits of detection bynearly two orders of magnitude [124].

A combination of two dyes with different colours hasbeen used to study Vaccinia virus infection [51]. A bluefluorescent cell-permeable DNA counterstain dye and agreen-fluorescent cell-permeable lipophilic dye were used[51]. Evolving to more complex hardware, a two-colourdetection system was incorporated for the first time into aCE–LIF chip to measure simultaneously fluorescence fromreference standards (650 nm) and the analyte (450 nm) inthe sample [89]. This approach enabled location of standardpeaks without interference from sample or backgroundpeaks [89].

Fluorescence detection has the advantage of sensitivity,but the derivatization steps required are a major drawback.Alternatives enabling label-free detection are, therefore,always of interest. When an immediate answer from anunprocessed sample is needed, fluorescence detection is oflimited value. To combine the advantages of fluorescencewith label-free detection, the gene for green fluorescentprotein has been incorporated into the Vesicular stomatitisvirus (VSV) genome to monitor the infection on a chip[125].

Another approach enabled rapid detection of bacteria bymonitoring off-chip bioluminescence [126]. Adenosinetriphosphate (ATP) extracted from bacterial cells was

252 I. Oita et al.

treated with luciferin to induce the bioluminescencereaction of firefly, luciferin-ATP [126].

Surface plasmon resonance (SPR) tends to be themethod of choice for label-free optical detection as analternative to fluorescence. The technique detects variationsin refractive index by observing changes in an optimumplasmon coupling angle or wavelength when bindingoccurs at metal–dielectric interfaces. The main disadvan-tage of this technique is the non-specific measure for massaccumulation, thus any change due to non-specificallyloaded molecules cannot be differentiated from the target[127]. In case of surface enhanced Raman spectroscopy(SERS), the absorption and scattering of properties ofmetallic nanoparticles enable their use as enhanced chro-mophores in molecular labelling [128]. In this way, thedisadvantages of SPR are overcome. SERS applications formicrofluidic devices are described in detail elsewhere[127]. The technique has been used to identify Denguevirus serotype 2 on chip at levels higher than 30 pmol L−1

and with excellent specificity against other serotypes [129].The combination of liquids and optics in the same

physical volume is one way to enrich the functionality ofthe sensors [123]. Qβ phages have been detected andmonitored on an optofluidic chip using anti-resonantreflecting optical waveguides [123]. Time-dependent fluo-rescence correlation spectroscopy data were used tocalculate diffusion coefficients, flow velocities, and con-centrations of viruses. The device can also be used as aninexpensive and portable sensor capable of discriminatingbetween viruses of different sizes. The technique issensitive to picomolar concentrations and can be used todetect and distinguish fluorescent objects in the size rangeof viruses, e.g. phages of 26 nm [123].

Electrochemical detection is an alternative with growingpopularity. It is sensitive, compatible with a wide array ofbiochemical reactions, and easily miniaturized. There arethree possibilities: voltammetric, conductometric, and poten-tiometric detection. This detection approach is not affectedby scale reduction and has an excellent performance evenwhen micrometer-size electrodes are employed. An overviewof the available approaches is given elsewhere [130]. Severalinteresting applications using electrochemical detection,SERS, and SPR are listed in Table 4.

Mass spectrometry (MS) enables detailed qualitative andquantitative assessment of cellular biomarkers with highsensitivity and reliability. The fabrication of mass spectrom-eters and interfaces between microfluidic platforms and MSdetectors on micro and nano scales is currently one of themost investigated topics [140–142]. The reduction in scalebrings advantages such as improved process control andautomation, shorter analysis times (minutes or seconds),reduced sample consumption, amenability to multiplexingand high-throughput processing, and lower analysis costs [9,

140–142]. Unfortunately, MS requires multi-step sample pre-treatment procedures, sometimes including liquid chromato-graphic (LC) separations [14]. Two main MS strategies arecurrently successfully implemented for quantitative proteo-mics, namely label-free and stable isotope labelling. Thesestrategies are described in detail elsewhere [140–142].

Analysis of a complex cellular extract on a fullyintegrated microfluidic system using MS detection hasbeen reported [14]. Proteins from breast cancer cellularextracts were tryptically digested, cleaned from salts andlabelled with an isobaric tag for relative and absolutequantitation (iTRAQ) reagents. Bovine proteins were alsoadded to the sample as standards. The separation wasperformed using a glass microchip LC–MS, designed in-house and enclosing four distinctive functional elementsincluding a pump, a sampling valve, a separation channel,and an electrospray ionization (ESI) interface. Mobile phasepropulsion through the LC channel and ESI interface wasachieved by EOF pumping. The pumping unit consisted oftwo arrays of 200 microchannels, each of 2 cm length and1.5–1.8 μm depth, connected in parallel. The large numberof channels ensured sufficient flow rate and the small sizeresulted in sufficient hydraulic resistance to pressurize aback-flow leakage. The EOF generated in the multichannelspump ensured mobile phase propulsion. The valving had asimilar design. Separation was performed in a channelpacked with 5 μm Zorbax C18 particles. Sample injectionwas performed electrokinetically and a gradient wasgenerated to perform the separation. Quantitative analysisof an entire protein extract has been performed without anysample pre-fractionation and differential protein expressionanalysis in MCF-7 cells cultured in the presence of β-oestradiol and tamoxifen [14]. The chip enabled reliableidentification of 40–50 proteins and, in another experiment,was able to identify five proteins of several previouslyreported human putative cancer biomarkers that were up ordownregulated [14].

Challenges and trends

Researchers developing applications or techniques usefulfor molecular diagnosis employing microfluidics are facingchallenges at three levels—the device level, the samplelevel, and the application level.

Challenges at the device level

The device itself is a source of challenges resulting fromdesign, fabrication and operation. Several commercialsolutions are available, but home-made adapted devicesare very popular and it is impressive to see that thecreativity of researchers has no boundaries. In the academic

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 253

world, lack of funds sometimes results in good brainactivity. However, in several cases, the enthusiasm ofresearchers leads to complicated solutions, which are nottested for reproducibility and enlarge the gap betweenacademia and industry.

The devices are fabricated in various ways, andfabrication is a significant part of microfluidics-relatedmanuscripts. Artificial polymers, PDMS [33, 36, 51, 113,114, 125, 143–145], PMMA [47, 77, 80, 132] or cyclicpolyolefin [40, 99] are the materials of choice for massproduction and most home-made applications. Polymershave a “Jekyll and Hyde” character in comparison withother material classes [146]. They have several advantages,i.e. are biocompatible and UV–visible transparent, and theiruse enables relatively inexpensive fabrication of complexmicro and nanostructures, with high reproducibility. More-over, huge numbers of materials and methods are availablefor microstructure fabrication. Selection of the appropriatematerial and microfabrication method for a given applica-tion is, therefore, extremely difficult, especially for scien-tists unfamiliar with polymer chemistry. This explains whythe scientific literature is dominated by a few favouritematerials only [146].

Polymeric materials can also be a source of challenges,especially those containing polyimide, which is incompat-ible with aggressive steps that might occur during manu-facture [147]. A modular approach has been used formanufacturing an electrical biosensor composed of single-stranded modified DNA probe used to perform simulta-

neous monitoring and differentiation of DNA sequencerepresentatives of PCR amplicons derived from human(H1N1) and avian (H5N1) influenza. Initially, the electrodeand chamber substrates were individually processed; thetwo substrates were then bonded to complete the processingof the device. The modular approach was a solution to therelatively harsh conditions occurring as a result of the useof sulfuric acid for in situ cleaning and preparation of theelectrodes [147].

The design of the device should be kept as simple aspossible to enable mass production, but sometimes it isproblematic to control the accurate delivery of manyreagents simultaneously, to ensure adequate mixing, andto avoid contamination. Some researchers have managed tofind simple solutions, for example integration of separatereservoirs for sample and reference solutions [77], hencereducing the number of washing steps and the risks ofcontamination. A simple solution has been described inwhich magnetic beads used throughout the application wereeffectively retained by placing a magnet on a side of thechip [126].

Although technological advances in recent decades haveenabled ready access to micro-components, fabrication on amicrometer scale (typical surface areas for micro-electromechanical systems are 100–10.000 μm2 [120]) withor without the use of biological reagents is a challenge thatcannot be neglected. When surfaces are functionalized withbiological macromolecules, supplementary aspects such asbiomolecule stability and compatibility also become criti-

Table 4 Applications employing a detection method other than fluorescence

Analyte LOD Detection Ref.

α-Fetoprotein (AFP), hepato-cellular carcinomabiomarker

0.1 ng Electrochemical [52]

Thyroglobulin, cancer biomarker 1 pg mL−1 Surface plasmon resonance [31]

8-Hydroxydeoxyguanosine (8-OH- dG) DNAadduct, biomarker for oxidative stress

20 attomoles Electrochemical [73]

Breast carcinoma markers

cal. A wide range of biomolecules have been immobilizedon the inner surfaces of devices, e.g. anti-Micoplasmapneomoniae antibodies [148], anti-gp120 antibodies [47],anti-AFP antibodies [77], bacterial cells [149], DNA [114,147], H1N1 probe [147], and Hsp60 [113]. A microspottingtool has been reported for sequential deposition ofbiomolecules at the same location on an active surface[120]. The technique enables proper coating of the sensingsurface with bioactive layers and parallel deposition ofthree different biomolecules in a single run [120]. Anothertechnique, the “layer-by-layer” (LbL) technique [150],enables polyelectrolyte coatings to be applied to the surfaceof digitally encoded microcarriers. The coating of micro-carriers with antibodies, and the use of the coated micro-carriers as capturing agents, have been reported [150].

Microfluidic devices are characterized by large surfacearea-to-volume ratios. Large capillary forces are hencegenerated, which may drive fluids into unwanted areas ofthe device, risking contamination of other fluid streams.Within the channels, the mixing of liquids can only beperformed by diffusion, because of the laminar flows. Toperform adequate mixing, long channels are needed. Anew membrane-type micromixer has been described [112].Mixing was achieved by injecting compressed air con-trolled by an electromagnetic valve (EMV). The micro-mixer was used in the process of incubation of the viralsamples and the magnetic beads [112]. Other approachesfor mixing liquids include the use of centrifugal force andcapillary action [114], and magnetic force [151].

Real sample analysis involves the complete integrationof sample preparation, and analyte separation anddetection [19]. Several manuscripts have reported inte-grated detection devices, for example a miniature laser-induced fluorescence detection module [89], a prototypeof an integrated fluorescence detection system, and anoptical fibre light guide on a laminate-based multichannelchip [117], and even a new design of a controllable micro-lens structure capable of enhancing an LIF detectionsystem [116].

Challenges at the sample level

Bioanalytical samples are well known for their complexity,i.e. often a large number of different molecules is presentover wide ranges of concentration in different matrices.Moreover, available sample volumes are usually low, i.e. inthe microlitre range. Also, biological colloids, for examplemacromolecular solutions and viral or bacterial suspen-sions, are known for wide distributions of charges, sizes,and shapes, which can be affected also by the experimentalconditions. Compared with these, small-molecule speciesare a homogenous population composed of virtuallyidentical entities. The distribution of analyte properties will

generate a distribution of electrophoretic mobilities, forexample, when capillary electrophoresis is used as separa-tion technique. This is one of the causes of the broad andoften irregular peak shapes sometimes obtained for samplesof biological origin [152].

Occasionally, the non-uniformity of analyte propertiesrequires alternative solutions, for example use of anasymmetric electrical field for electrophoretic separations.Dielectrophoresis was the solution found for separation ofcells or particles, when the large size variations turned intoa separation asset [36, 51, 113, 144].

Wall interactions are another common problem forsamples of biological origin, especially for protein-containing samples. BSA-FITC has been used as modelprotein to demonstrate how protein molecules are adsorbedand distributed on the inner wall surface of the PMMAmicrochannel [132]. The surface of the channels wasinitially coated with polyethyleneimine (PEI) containingabundant amino groups to covalently immobilize AFPmonoclonal antibody. BSA–FITC binding within thechannels was then studied to enable optimization of thesystem to reduce non-specific binding, and quantification ofAFP was achieved with LOD down to 1 pg mL−1. Theutility of the chip to detect AFP from healthy human serumwas demonstrated [132].

Gonzales et al. [153] studied the adsorption of majorpolymerase chain reaction (PCR) mixture components onthe capillary channel wall. None of the polymeric materialsor flow velocities tested was found to affect subsequentPCR amplification. PCR inhibition occurred only afterexposure of the mixture to tubing lengths of 3 m or whenthe sample volume was reduced. Individual testing of PCRproducts revealed significant DNA adsorption and an evengreater adsorption of the fluorescent dye used. Thesefindings imply that adsorption of reaction components bywall surfaces is responsible for inhibition of PCR inpolymeric tubing. These phenomena increase substantiallywith increasing tubing lengths or with sample volumereduction, but not with contact times or typical flowvelocities for dynamic PCR amplification [153]. Theseresults indicate the need for careful consideration ofchemical compatibility between polymeric capillaries andDNA dyes when quantitative microfluidic devices aredeveloped [153].

Challenges at the application level

Development of applications is usually focused on main-taining the stability of samples and reagents, preventingfalse-positive results because of non-specific binding,increasing precision, and reducing the LOD.

To avoid non-specific binding, more sensitive andselective approaches are used for the design of devices.

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 255

Molecular-imprinted polymers (MIP), for instance, areartificial recognition materials designed to interact non-covalently with the analyte. A cross-linked polymercontaining highly selective recognition sites created bymeans of soft-lithography has been used for specificrecognition of viruses [154, 155]. The shape and surfacechemistry of the MIP facilitated highly specific interactionwith the virus, and non-specific interactions were thuseliminated [154, 155].

In some cases, the applications are more than routineanalyses such as separation, identification, or assay—forinstance, a device used for cell growth and on-chipinfectivity assay of swine influenza virus [125].

Trends

The huge number of research projects on the developmentof portable diagnostics is an objective means of quantifyingthe great expectations of microfluidics in the moleculardiagnosis field.

Paper is increasingly regarded as a promising supportfor inexpensive, portable, fully disposable, and easy touse devices for complicated molecular diagnostics—aseasy to interpret as the home-used pregnancy test.Photolithography can be used to build selectivelyhydrophobic barriers in the filter paper, enabling thehydrophilic paper channels to control the transport ofaqueous solutions by capillary forces without the needfor external pumping [156]. Several other research groupsfocused on the calibration of paper-based microfluidicdevices [157], on the techniques used to generatehydrophilic channels in filter paper [158], on the applica-tion of electrochemical detection [159], or on thedevelopment of a hand-held optical colorimeter [160].

Use of microfluidic devices gives deeper insight intothe life sciences and medicine because it enables, forinstance, study of the proteins in a single cell [161],quantification of multiple proteins in a single sample [33,162], identification of a single nucleotide polymorphism[135], or detection below the limits of classical methods,for example the “bio-bar code” assay [163–165]. The“bio-bar code” assay (Fig. 6) uses two types of function-alized particles:

1. magnetic particles functionalized with recognitionelements, i.e. monoclonal antibodies or a hapten-modified oligonucleotide; and

2. gold nanoparticles functionalized with a second recog-nition element and a bifunctional oligonucleotide bar-code DNA [2].

The analyte of interest, the target, usually a protein, isinitially captured and enriched by the magnetic micropar-

ticle (Fig. 6). Further, the target is sandwiched betweenmagnetic microparticle and the gold nanoparticles. Thesandwiches are easily separated from the sample in amagnetic field. In a further step, the DNA of the bar code isreleased by heating. Signal amplification is achievedbecause for each target recognition thousands of bar-codesare released. Half of the released DNA is detected by use ofa scanometric assay based on the affinity of the releasedDNA for a complementary “universal” scanometric goldnanoparticle DNA probe. The other part is complementaryto the chip immobilized DNA, responsible for sorting andbinding barcodes complementary to the target sequence.The concept has unparalleled sensitivity for target detec-tion, i.e. it can be between one and six orders of magnitudemore sensitive than conventional ELISA. By use of thismethod, PSA was measured in the serum of patients afterradical prostatectomy, even in cases when the availableimmunoassays were not able to detect it [166].

At the crossroad of four major scientific fields, i.e.biology, physics, chemistry, and medical science, biosen-sors are one of the most studied subjects of recent years.Although significant progress has been achieved, andanalysis at the single-molecule level [5] or based onsingle-cell composition [161, 167] is possible, viablesolutions for real-time, bedside diagnostic devices are stillawaited.

Recent progress in technology opens the way towardmass production of biosensors and bedside devices. The useof polymeric materials for fabrication of microfluidicsystems will simplify manufacturing processes [168]. Newtransducing and biocompatible interfaces are expected to bedeveloped based on composites which integrate nano-particles, carbon nanotubes (CNTs), and nanoengineered“smart” polymers [168]. The difficulty in producing low-cost, sensitive, and portable optical detection systems limitsthe number of the commercial devices for bedside diag-nostics, but several technological solutions for biomolecule-compatible optofluidic integration have already beendescribed [169].

Linder [170] has described two types of technologyexpected to extend the use of microfluidic devices outsideclinical laboratories, i.e. amplification chemistry that resultsin the accumulation of an opaque material at the surface ofreaction site, and a solution for long-term storage ofmultiple reagents and for the sequential delivery of thereagents to the reaction site inside a microfluidic device.Innovative techniques involving nanoengineered materialshave been used to develop potentially less expensivedetection systems without alignment requirements, withoutminiaturization disadvantages, and which are more readilyadapted for bedside use [108]. The combination of progressin technology and the life sciences brings us one step closerto microfluidic devices with higher sensitivity, and im-

256 I. Oita et al.

proved biocompatibility and stability of the immobilizedmolecules, which makes them feasible biosensors andbedside devices.

The use of protein biomarkers for development of micro-electrical sensors, and the underlying technical concepts,have been described [171]. Similarly, emerging optical andmicrofluidic technology suitable for bedside genetic anal-ysis systems [169] and DNA biosensors [172] have beenreviewed. Some early stage commercial products based onelectrochemical DNA biosensors integrated in analyticalmicrofluidic devices have also been reported [172].

There is currently a discrepancy between the molec-ular based diagnostic tests approved by regulatoryagencies such as the FDA (Table 5) within the last threeyears, and the literature and the available technology. Theapproved tests are rather traditional, and mostly based onimmunoassays. Several tests use PCR for confirmation ofviral infection, cancer diagnosis, and for prognosis ordiagnosis of genetic diseases, for example cystic fibrosisin new-borns (Table 5). Still, compared with the numberof published papers (hundreds or even thousands) FDA-cleared diagnostic devices involving microfluidics arerestricted mostly to immunochromatographic assays toconfirm several viruses and bacteria, and several otherdevices intended for measurement of cholesterol level inblood control or for glycaemic control in people withdiabetes. All the other tests make use of highly specialized

reagents and equipment. The whole situation is even morecontradictory if we consider that more than hundredcompanies are producing and commercializing miniatur-ized analytical devices [173].

According to a study published by Analytical Chemistry[8], microfluidics is now at the “slope of enlightenment”stage of the Gartner hype cycle model of the life cycle oftechnology (Fig. 7a). If the yearly change in paperspublished on microfluidics (Fig. 7b) is considered, it isclear that major events in science and technology also actedas triggers for microfluidics. The technology was redefined,a new cycle started, more powerful applications weredefined, and deeper insight was obtained.

Concluding remarks

Microfluidics are currently among the most fashionableresearch topics. Much of the research exploits the advan-tages of microfluidics to solve current problems inbioanalysis, e.g. small volumes, limited stability, and highcost. An overwhelming number of manuscripts is devotedto trying to extend research findings into routine clinicaluse.

However, only a few diagnostic devices are availablecommercially, and the increase in the number of registereddevices is not following the number of published papers or

Fig. 6 Biobar code assay developed for quantification of PSA in patient serum. Schematic representation of the PSA Au–NP probes (A) and thePSA bio-barcode assay (B). For details, see text (reproduced with permission from Ref. [166])

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 257

Table 5 Molecular diagnostics FDA approved within the last three years (source: https://www.accessdata.fda.gov/scripts/cdrh/devicesatfda/index.cfm, accessed 02.2010)

ProductName

Marketingauthorisationholder

Approvaldate

Applicationrange

Working principle Instrumentation

1 VIDASfPSA rtAssay

Biomerieux 10/8/09 Assessment of the chance that theman has prostate cancer basedon measurement of free prostate-specific antigen (fPSA)

Two step enzyme immunoassaysandwich method with a finalfluorescence detection

Ready-to-use reagentstrips (10 wells/strip)to be used withbenchtop chemistryanalyzers

2 ArchitectCore

AbbottLaboratories

04/10/09 Detects antibodies (anti HBc)associated with the hepatitisB virus (HBV) core antigenin serum or plasma

Two-step immunoassay withchemiluminescence detection

Benchtop analyzer

3 CervistaHPV

Third WaveTechnologies

03/12/09 Identify DNA from humanpapilloma virus (HPV) types 16and 18 in cervical samples

DNA extracted from cervicalsamples is amplified in athermal cycler. A secondisothermal reaction will generatea fluorescent signal

Benchtop PCR

4 CobasTaqMan

RocheMolecularSystems

10/03/10 Quantify the amount of hepatitisC viral RNA in a patient’s bloodto help the physicians to determinean individual’s response totreatment

Nucleic acid (RNA) is separatedfrom the cells in the bloodsample. Separated RNA isamplified and the amount ofHCV RNA in the patient’s bloodis measured on the basis of theamount of light produced

Benchtop qPCR

5 Tspot. TB OxfordImmunotec

07/30/08 Detect the immune response ofthymus cells (T cells) found inan individual’s white bloodcells that are stimulated byproteins produced by thebacteria that causes tuberculosis

Combination of a two step enzymeimmunoassay sandwich methodwith a final visual detection

Benchtop-wells

6 Spot-LightHER2CISH

Invitrogen 07/01/08 Measure the number of copies ofHer-2 gene on chromosome 17in breast cancer cells to assess ifthe patient is eligible for treatmentwith the cancer drug Herceptin(Trastuzumab).

On-slide binding of Her-2 genewith a matching digoxigenin-tagged DNA probe. A fluorescent(FITC) tagged antibody to digox-igenin followed by a horseradishperoxidase conjugated antibodyto FITC and DAB further revealsthe probes.

Fluorescencemicroscopy

7 xTAGrespiratoryviral panel(RVP)

LuminexMolecularDiagnostics

01/03/08 Identifies nucleic acids of multiplerespiratory viruses innasopharyngeal swab specimensfrom individuals suspected ofrespiratory tract infections

Multiplex detection of viral nucleicacids based on beads; selectiveisolation of viral RNA followedby PCR

Multiple steps,benchtop PCRequipment

8 DakoTOP2AFISHpharmDx

DakoDenmark

01/11/08 Assessment of the risk of post-surgical recurrence of breastcancer and long-term survivalbased on the measurement ofthe number of copies of theTOP2A (Topoisomerase 2 alpha)gene on chromosome 17 inbreast cancer cells

On slide binding of Her-2 genewith a matching a red fluorescent-tagged DNA probe binds tomatching DNA of the TOP2Agene and a green fluorescent-tagged DNA probe binds to thematching central portion of chro-mosome 17 in cells on the slide

Fluorescencemicroscopy

9 GeneSearchBLN Test

Veridex LLC 07/16/07 Rapid detection of metastaseslarger than 0.2 mm in lymphnodes tissue removed frombiopsies of breast cancerpatients

Chemical amplification of twogene products abundant in breasttissue and scarce in lymph nodecells. Fluorescence detection

Benchtop PCR

10 Mesomark FujirebioDiagnostics

01/24/07 Assessment of metastases fora rare cancer of the internalbody lining (mesothelioma)

Colorimetric reaction for specificprotein fragments released bycells of malignant mesotheliomainto a patient’s blood

Benchtop analyzer

258 I. Oita et al.

https://www.accessdata.fda.gov/scripts/cdrh/devicesatfda/index.cfmhttps://www.accessdata.fda.gov/scripts/cdrh/devicesatfda/index.cfm

funding of biomarker analysis on microfluidic platforms.There are two different views on the evolution of thesedevices: the optimistic view is that microfluidics willbecome a integral part of the future of bioanalysis; thepessimistic view is that microfluidics will remain a niche

preoccupation for research without any practical consequences[8].

Although a “killer application” has not yet been de-veloped [8], small steps towards it are taken every day. Thebeginning is complete, and applications, for example

Table 5 (continued)

ProductName

Marketingauthorisationholder

Approvaldate

Applicationrange

Working principle Instrumentation

11 MammaPrint AgendiaLaboratory

2009 Assess a patient’s risk of distantbreast cancer metastasis

Microarray-based gene expressionanalysis of RNA extracted frombreast tumour tissue

Benchtop–bioanalyzer

12 AVantage A/H5N1 flutest

Arbor Vita 04/01/09 Rapid diagnostic test for the H5N1influenza A viral subtype (avian orbird flu)

Rapid proteomics test for thespecific detection of H5N1

Portable

b 1996 – cloning of Dolly the sheep

2004 – first carbon nanotube transistor

% in

crea

se o

f th

e n

um

ber

of

pap

ers

2002

2000

140

20092007

2005

2004

2003

2001

1999

1998

1997

1996

1995

1994

0

100

120

160

1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Year

60

80

20

40

20082006

2001 – decoding of human genome

aFig. 7 a Gartner hype cyclemodel (reproduced with permis-sion). b Evolution of percentagechange in the number ofpapers published yearly onmicrofluidics

Microfluidics in macro-biomolecules analysis: macro inside in a nano world 259

MammaPrint (Table 5), that uses a microfluidic platform togive a prognosis of the future evolution of breast cancer,will increasingly become part of our life. Technically,applications within the life sciences are increasingly tryingto resemble the world depicted in ’70s–’80s science fictionmovies: devices were always small (handheld), incrediblysleek, and provided all the information necessary at amoment’s notice [173]. Use of microfluidics devices in lifescience and medicine have revealed new perspectives andenabled deeper investigation, but they have still to find theirway to more generalized and commercialized applications.

Acknowledgements This work was supported by a HorizontaleOnderzoeksactie (HOA) of the Vrije Universiteit Brussel and a researchgrant (G.0051.08) of the FWO-Vlaanderen. Debby Mangelings is apostdoctoral fellow of the Research Foundation Flanders (FWO).

References

1. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P(2008) Molecular biology of the cell, 5th edn. Garland Science,New York

2. Giljohann DA, Mirkin CA (2009) Drivers of biodiagnosticdevelopment. Nature 462:461–464

3. Tothill IE (2009) Biosensors for cancer markers diagnosis.Semin Cell Dev Biol 20:55–62

4. Rosi NL, Mirkin CA (2005) Nanostructures in Biodiagnostics.Chem Rev 105:1547–1562

5. Shen YP, Wui BL (2009) Microarray-based genomic DNAprofiling technologies in clinical molecular diagnostics. ClinChem 55:659–669

6. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ(2006) Global and regional burden of disease and risk factors,2001: systematic analysis of population health data. Lancet367:1747–1757

7. Fields S (2001) The interplay of biology and technology. PNAS98:10051–10054

8. Mukhopadhyay R (2009) Microfluidics: on the slope ofenlightenment. Anal Chem 81:4169–4173

9. Manz A, Graber N, Widmer HM (1990) Miniaturized totalchemical analysis systems: a novel concept for chemical sensing.Sensors and Actuators: B Chemical 1:244–248

10. Arata HF, Kumemura M, Sakaki N, Fujita H (2008) Towardssingle biomolecule handling and characterization by MEMS.Anal Bioanal Chem 391:2385–2393

11. Amur S, Frueh FW, Lesko LJ, Huang SM (2008) Integration anduse of biomarkers in drug development, regulation and clinicalpractice: a US regulatory perspective. Biomark Med 2