Molecular Diagnostics on Microfabricated Electrophoretic ... · slab gel electrophoresis to identify products and provide semiquantitative data. CE provides a more rapid, auto-mated

Molecular Diagnostics on MicrofabricatedElectrophoretic Devices: From Slab Gel- to

Capillary- to Microchip-based Assays for T- andB-Cell Lymphoproliferative Disorders

Nicole J. Munro,1 Karen Snow,2 Jeffrey A. Kant,3 and James P. Landers1,4*

Background: Current methods for molecular-based di-agnosis of disease rely heavily on modern molecularbiology techniques for interrogating the genome foraberrant DNA sequences. These techniques typicallyinclude amplification of the target DNA sequencesfollowed by separation of the amplified fragments byslab gel electrophoresis. As a result of the labor-inten-sive, time-consuming nature of slab gel electrophoresis,alternative electrophoretic formats have been developedin the form of capillary electrophoresis and, more re-cently, multichannel microchip electrophoresis.Methods: Capillary electrophoresis was explored as analternative to slab gel electrophoresis for the analysis ofPCR-amplified products indicative of T- and B-cellmalignancies as a means of defining the elements forsilica microchip-based diagnosis. Capillary-based sepa-rations were replicated on electrophoretic microchips.Results: The microchip-based electrophoretic separa-tion effectively resolved PCR-amplified fragments fromthe variable region of the T-cell receptor-g gene (150–250 bp range) and the immunoglobulin heavy chaingene (80–140 bp range), yielding diagnostically relevantinformation regarding the presence of clonal DNA pop-ulations. Although hydroxyethylcellulose provided ad-equate separation power, the need for a coated micro-channel for effective resolution necessitated additionalpreparative steps. In addition, preliminary data are

shown indicating that polyvinylpyrrolidone may pro-vide an adequate matrix without the need for micro-channel coating.Conclusions: Separation of B- and T-cell gene rear-rangement PCR products on microchips provides diag-nostic information in dramatically reduced time (160 svs 2.5 h) with no loss of diagnostic capacity whencompared with current methodologies. As illustrated,this technology and methodology holds great potentialfor extrapolation to the abundance of similar molecularbiology-based techniques.© 1999 American Association for Clinical Chemistry

DNA fragment analysis has become an integral compo-nent of clinical diagnostics. Detection of PCR-amplifiedfragments currently is used for a wide variety of clinicalapplications such as disease diagnosis and minimal resid-ual disease detection (1, 2). Separation of DNA fragmentshas traditionally been performed by electrophoresis onslab gels utilizing densitometry or Southern blot detec-tion. The potential for translating these separations tocapillary electrophoresis (CE)5 has been illustrated (3 ). Inaddition to the minute quantities of sample and reagentsrequired for CE, there is a dramatic reduction in analysistime. This time reduction results from the translation ofthe electrophoretic format from the slab gel to the capil-lary, where higher applied fields can be tolerated as aresult of the rapid heat transfer through the micron-scaleinternal diameter capillary. Unfortunately, incorporatingparallel processing, as is performed on slab gels, has notbeen as prolific, although several examples have beenshown [for example, see Ref. (4 )] and commercial instru-

1 Department of Chemistry, University of Pittsburgh, Pittsburgh, PA15260.

2 Division of Laboratory Genetics, Mayo Clinic, Rochester, MN 55905.3 Department of Pathology, University of Pittsburgh Medical Center,

Pittsburgh, PA 15213.4 University of Pittsburgh Cancer Institute, Pittsburgh, PA 15260.*Address correspondence to this author at: Department of Chemistry,

University of Virginia, Charlottesville, VA 22901. Fax 804-243-8852; [email protected].

Received May 27, 1999; accepted August 10, 1999.

5 Nonstandard abbreviations: CE, capillary electrophoresis; TCR, T-cellreceptor; IgH, immunoglobulin heavy chain; LIF, laser-induced fluorescence;TBE, Tris-borate-EDTA; HEC, hydroxyethylcellulose; PVP, polyvinylpyrroli-done; FC, fluorocarbon; and PA, polyacrylamide.

Clinical Chemistry 45:111906–1917 (1999) Molecular Diagnostics

and Genetics

1906

ments are available. For high-throughput analysis appli-cations such as genomic sequencing, clinical diagnostics,or drug screening, parallel processing will be essential.

A miniaturized variant of CE exploits extant microfab-ricated chip technology to create an electrophoretic chipdevice that provides even greater advantages for rapidand cost-effective clinical analysis. Perhaps the greatestadvantage of the electrophoretic chip platform over theslab gel and even the capillary formats is the potential forintegrating sample processing steps directly (5–8). Incombination with parallel processing, this has the poten-tial to create the rapid, high-throughput platform de-manded by current clinical diagnostic applications. Thepotential impact on the clinical sector for electrophoresisusing microfabricated devices has been discussed (9 ) withsupporting preliminary data from applications such asserum protein separations (10 ) and immunoassays (11–13). Clinically relevant DNA separations on microfabri-cated devices have also been shown, including the sepa-ration of PCR products of the dystrophin gene (14 ) andDNA-sequencing products (15 ).

The development of electrophoretic microchip technol-ogy is in its infancy, although an increased number ofresearch groups are reporting their production and use(5, 6, 8, 11–19). Despite the fact that few reports havedemonstrated the application of electrophoretic microchipsto real-world analysis, it is clear that this technology isideally suited to the rapid analysis of PCR-amplified DNA.

A multitude of diagnostic assays currently in useincorporate PCR-based amplification for detection of ab-errant DNA sequences (mutations, rearrangements, dele-tions, and translocations) correlative with the onset ofdisease (1 ). Such is the case for the diagnosis of T- or B-celllymphoma, where PCR results can be available within 1day whereas Southern blot studies require 1 week (20 ).Clinical diagnosis of T- or B-cell lymphoproliferativedisorders takes advantage of the maturation process ofthe T and B cells by focusing on the analysis of the genesencoding for the T-cell receptor (TCR) and immunoglob-ulin heavy chain (IgH). The majority of oncogenic eventsbegin after rearrangement (21 ); therefore, the offspring ofa transformed cell will possess the same specific genesequence produced when the parent cell underwent rear-rangement. A predominance of a single sequence will,therefore, be a mark of clonality, an indication of malig-nancy. However, as a result of the combinatorial nature ofthe rearrangement process, a normal cell population willpossess a polyclonal population, a diverse variety of lowabundance sequences. PCR-based assays have been de-veloped to analyze the lengths of the DNA fragmentsproduced when a variety of primers for the variable andjoining regions are used (multiplexed PCR) (22 ). Whenmultiplexed PCR and subsequent fragment separation areused, a normal cell population displays a large variety ofDNA fragment sizes, whereas a malignant populationdisplays a predominance of one DNA fragment size.

As with other PCR assays, established protocols for

analyzing B- and T-cell amplified products have relied onslab gel electrophoresis to identify products and providesemiquantitative data. CE provides a more rapid, auto-mated platform for high-resolution PCR fragment analy-sis with excellent detection sensitivity provided by laser-induced fluorescence (LIF) detection (23 ). In an earlierreport, Oda et al. (24 ) took the first steps toward utilizingCE for T-cell malignancy diagnosis by defining initialconditions for separation of the PCR-amplified fragments.In this report, this work is extended to include theanalysis of B-cell malignancy assay samples and themethodology translated to the microfabricated chip,where separations can be achieved over an order ofmagnitude faster. Comparisons are made between slabgel, capillary, and microchip electrophoretic results withrespect to analysis time, resolution, and diagnostic capa-bilities. In addition, with a vision toward a simple androbust microchip platform for diagnostic DNA fragmentdetection, the results of separations that were performedin a microchip containing uncoated channels are pre-sented; these results, although preliminary in nature,illustrate the feasibility for this methodology.

Materials and Methodssieving matrix preparationA 13 Tris-borate-EDTA [TBE; 89 mmol/L Tris (SigmaChemicals), 89 mmol/L boric acid (EM Chemicals), and 2mmol/L EDTA (Sigma)] solution was adjusted to pH 8.6with 2 mol/L sodium hydroxide. Hydroxyethylcellulose(HEC, Mr 250 000; Aldrich) and polyvinylpyrrolidone(PVP, Mr 360 000; Acros) sieving matrixes were preparedby heating a 13 TBE solution to the point were conden-sation was seen on the inside glass surface (;55 °C) andthen adding the polymer (10 or 20 g/L) slowly to therapidly stirring solution. Heated mixing at ;55 °C contin-ued until the solution cleared (;15 min); it was thenallowed to stir for 1 h with no heating. The solution wasfiltered through an 0.8 mm filter (Millipore) and stored at4 °C. Before the sieving matrix was used, 1-(4-[3-methyl-2,3-dihydro-(benzo-1,3-oxazole)-2-methylidene]-quino-linium)-3-trimethyl-ammonium propane diiodide (YO-PRO-1; Molecular Probes) was added at 1 mL of YO-PROper milliliter of sieving matrix (a 1:1000 dilution of the 1mmol/L solution obtained from the manufacturer).

instrumentationCE instrumentation. A Beckman P/ACE System 5510(Beckman Instruments) was used for CE analysis. Fluo-rescence detection used a P/ACE System Laser Module488 with a P/ACE LIF detector, which excites at 488 nmand collects emission at 520 nm 6 10 nm. A laser power of0.25 mW was presented at the capillary. Instrumentcontrol and data collection were performed with an IBM486 ValuePoint computer utilizing System Gold software(Ver. 8.1).

Clinical Chemistry 45, No. 11, 1999 1907

Electrophoretic microchip instrumentation. The separationvoltages were supplied by an in-house manufacturedhigh-voltage power supply controlled by a program writ-ten in Labview. For fluorescence detection, the 488-nmline of an argon ion laser (532R-BS-AO4; Melles Griot)was expanded to fill the back aperture of an objective. Theexpanded beam was reflected off a beam splitter(505DRLP02; Omega Optical) set at 45° to the incidentbeam and into the channel of an electrophoretic microchipby an objective (203/numerical aperture, 0.5). Fluores-cence emitted by the sample was collected by the objectiveand focused by a 200-mm lens onto a 400-mm pinhole(25 ). The fluorescence wavelength was spectrally filteredby a 530 nm bandpass filter (Omega Optical 530DF30,FWHM 30 nm), collected by a photomultiplier tube (PMT,Hamamatsu R38960), and processed via a program writ-ten in Labview.

Electrophoretic microchip microfabrication. Electrophoreticchips were microfabricated by the Alberta Microelectron-ics Center. Channels were etched in one glass plate towhich a glass top plate was bonded. Electrophoretic chipsconsisted of an eight-channel arrangement, with eachchannel composed of an injection cross with a sample-to-sample waste distance of 1.45 cm, inlet-to-outlet distanceof 6.65 cm, and the junction 0.5 cm from the inlet (Fig. 3).Detection occurred 4.2 cm from the injection cross. Thesample channel was 100 mm wide and 10 mm deep, andthe separation channel was 50 mm wide and 10 mm deep.One-millimeter holes were drilled in the top glass plate toallow access to the channels. Cut pipette tips were ep-oxied onto the top plate to form larger wells for solutionsand placement of gold-coated electrodes.

pcr amplification of tcrg rearrangementsRearranged T-cell receptor-g (TCRG) gene sequences wereamplified using multiplexed PCR as described previously(20 ). For the TCRG variable segments, five primers (TCRGV2, 3, 4, 8, and 9) were used along with three primers forthe TCRG joining segments (JGT12 with consensus se-quences for J1.3 and J2.3, JGT3 with consensus sequencesfor J1.1 and J1.2, and JGT4 with consensus sequences forJ1.2). The PCR reaction mixture included 0.75 U of Taqpolymerase (Taq Gold-P; Perkin-Elmer), 200 mmol/L eachof dATP, dCTP, dGTP, and dTTP, 1.5 mmol/L MgCl2,and 1.0 mmol/L of each primer in standard PCR buffer(Perkin-Elmer) for a total volume of 25 mL. PCR cyclingconditions included a 10-min hold at 95 °C, followed by40 cycles of PCR (94 °C for 30 s, 55 °C for 1 min, 72 °C for1 min) and a 10-min extension at 72 °C.

pcr amplification of igh rearrangementsIGH PCR was performed in 25-mL reactions using 1 U ofTaq polymerase (PE Biosystems) and a consensus 59framework III variable region primer (AGG TGC AGCTGG TGC AGT CTG G) with a mixture of three consensus

primers directed at heavy chain joining regions 1, 2, 4, and5 (ACC TGA GGA GAC GGT GAC CAG GGT), 3 (TACCTG AAG AGA CGG TGA CCA TTG T), and 6 (ACCTGA GGA GAC GGT GAC CGT GGT). PCR thermocy-cling conditions included an initial denaturation for 5 min at95 °C, followed by 35 cycles of PCR (95 °C for 15 s, 52 °C for20 s, 72 °C for 30 s) and a 5-min extension at 72 °C.

analysis via slab gel electrophoresis with sybrgreen i stainingFor the TCRG gene rearrangement assay, 3.5 mL of theamplified PCR mixtures was loaded onto an 8% poly-acrylamide (19:1, by weight) minigel. The DNA markerwas pGEM (Promega). Electrophoresis was performedusing a 13 TBE buffer for 75 min at 200 V. Gels werestained with SYBR Green 1 (1 mg/L; Molecular Probes)with gentle agitation for 30 min at room temperature.Stained gels were illuminated with a 300 nm ultraviolettransilluminator and photographed with Polaroid 667black and white print film through a SYBR Green I gelstain photographic filter (1- to 2-s exposure; f-stop, 4.5).

For IGH gene rearrangement PCR assays, 30 mL of PCRreaction product mixed with 3 mL of blue sucrose wasloaded into lanes of an 8% polyacrylamide gel (16 cm 3 20cm 3 1.5 mm) and electrophoresed in 13 TBE buffer for2.5 h at 180 V. The gel was stained with SYBR Green I andphotographed with type 57 Kodak film. MspI digests ofpBR322 (New England Biolabs) were used as molecularweight markers.

pcr product analysis via lif-ceFor CE separations using HEC (described above), a 37cm 3 50 mm (effective length, 30 cm) fluorocarbon (FC)-coated m-SIL capillary (J & W Scientific) was used. YO-PRO-1 (1 mmol/L) was used as a fluorescent intercalator.The DNA marker, HaeIII digest of pBR322 (BoehringerMannheim Biochemicals), was diluted 1:100 with 10mmol/L Tris, 1 mmol/L EDTA for a final concentration of4.20 mg/L. The PCR-amplified samples were diluted 1:10in 10 mmol/L Tris, 1 mmol/L EDTA. FC-coated capillar-ies were first rinsed with 20 column volumes of water.HEC separations used the following method: 3-s electro-kinetic injection of water at 3.5 kV, 10-s injection of sampleat 8.5 kV (230 V/cm), 3-s injection of water at 3.5 kV.Separation followed at 8 kV (216 V/cm), using reversedpolarity (inlet cathode, outlet anode) and maintaining thecapillary at 20 °C. The capillary was rinsed with 5 columnvolumes of water between runs and prerinsed with 10column volumes of fresh (unelectrophoresed) buffer be-fore each run.

For CE separations using PVP (described above), 37cm 3 50 mm bare silica capillaries (Polymicro Technology)were first conditioned by rinsing with 20 column volumesof water. Injection and run settings were as stated for theHEC/FC-capillary system.

1908 Munro et al.: T- and B-Cell Lymphoma Assay by Microchip Electrophoresis

pcr product analysis via microchipelectrophoresisMicrochip channels were coated with linear polyacryl-amide (PA) by a modified Hjerten method (26 ). Sampleswere desalted by placing 10 mL of undiluted sample in aMicrocon® YM-10 Centrifugal Filter Device (Amicon)with 90 mL of water and then centrifuged at 7g for 10 min.One hundred microliters of water was then added, vor-tex-mixed, and centrifuged for 10 min at 7g; this proce-dure was repeated five times, and the eluent was dis-carded. To reconstitute the sample, 10 mL of water wasadded, vortex-mixed, centrifuged for 30 s at 2g; thecartridge was then inverted and centrifuged at 7g for 2min. The same HEC and PVP sieving matrices were usedas described for the CE analyses. Channel preparationand between-run rinses were performed as described forthe capillary system. Microchip sample injection wasperformed by applying a 400 V (275 V/cm) potentialacross the sample and sample waste reservoirs, with thesample at ground. For separation, the sample and samplewaste were grounded and 2200 V was applied to the inlet

and 900 V to the outlet (165 V/cm). Fluorescence wascollected at 10 Hz.

Resultsslab gel and ce analysis of tcrg and igh generearrangementsConventional methodology for detecting gene rearrange-ments consistent with T- and B-cell lymphomas involveselectrophoresis of T-cell receptor-g (TCRG) and immuno-globulin heavy chain (IGH) gene rearrangement PCRproducts on an acrylamide gel. Fig. 1 shows SYBR Green-stained polyacrylamide gels for these assays. In theseanalyses, a normal polyclonal population of lymphoidcells will yield a series of DNA fragments that appear asa “smear” or a series of discrete bands, depending on theresolution. Conversely, a clonal lymphocyte population,indicative of a lymphoid neoplasm, will yield one or moredominant bands. The PCR primers used for multiplexedamplification of the TCRG variable region(s) yield frag-ments in the 150- to 250-bp range (Fig. 1A). The polyclonalfragments typical of a normal sample (negative for a T-cell

Fig. 1. Slab gel separations of TCRG and IGH gene rearrangement products.(A), slab gel of nine TCRG gene rearrangement products and a DNA ladder, pGEM. Lanes 1 and 5 were diagnosed as negative equivocal; lanes 2–4, 7, and 8 as positive;and lanes 6 and 9 as negative. Lane 7 shows the results for a positive control that contains a 5% dilution of DNA from a clonal population into genomic DNA from whiteblood cells. Lane 8 represents a positive control containing a 10% dilution of DNA from the same clonal population into DNA from thymus. Lane 9 is a negative control.(B), slab gel of IGH gene rearrangement products and a DNA ladder, pBR322/Msp1. Lanes 1, 5, 7, and 8 were diagnosed as positive for monoclonality; lanes 2–4and 6 as negative for monoclonality. Lanes 1 and 2 are controls.


lymphoproliferative disorder) are shown in lane 9 (nega-tive control) and in lane 6 (patient sample). Both casesdisplay the smear of DNA fragments typical of polyclonalpopulations. In contrast, the presence of a dominatingDNA fragment indicative of clonality is illustrated by thepatient samples in lanes 2, 3, and 4, which exhibit one ormore densely stained bands in the 150- to 250-bp regionand compare favorably with the positive controls in lanes7 and 8. Samples in lanes 1 and 5 were negative/equivocalfor clonality.

Interrogation of B-cell populations via the same molec-ular approach is shown in Fig. 1B. Here, multiplex PCRamplification of the variable region(s) of the IGH geneyields fragments in the 80- to 140-bp region (Fig. 1B). Aswith the T-cells, normal samples produce a polyclonalpopulation of DNA fragments that are observed as low-abundance DNA bands spanning the 80- to 140-bp range(lanes 2–4 and 6). The lower molecular weight fragments

yielded by the IGH gene amplification are resolved betterthan the corresponding T-cell amplification products and,therefore, are not observed as a smear but rather as aseries of discrete bands. However, with B-cell samplesthat display clonality, dominant band(s), which signal thepresence of the monoclonal population, are clearly ob-served and indicative of a B-cell lymphoproliferativedisorder (lanes 1, 5, 7, and 8).

The results from CE analysis of selected samples ana-lyzed via slab gel electrophoresis in Fig. 1 are shown inFig. 2. CE conditions included a 37 cm 3 50 mm FC-coatedcapillary containing 10 g/L HEC solubilized in 13 TBEbuffer at pH 8.6. A fluorescent intercalator, YO-PRO-1,was added to the buffer for LIF detection. Fig. 2A showsCE separations representative of TCRG gene rearrange-ment products and compares these electropherogramswith a separation of the lower molecular weight frag-ments in a pBR322 HaeIII digest (bottom panel). The top

Fig. 2. CE separation of gene rearrangement products.(A), negative/equivocal sample (T1) and positive (T2) TCRG gene rearrangement samples analyzed via CE with corresponding slab gel separations (insets) and DNAmarker, HaeIII digest of pBR322. (B), CE separation of a positive (B1) and negative (B2) IGH gene rearrangement sample with analogous slab gel separations (inset)and DNA marker. Electropherogram labels correspond to those in Fig. 1. Conditions: 37 cm 3 50 mm FC capillary, effective length, 30 cm; 10 g/L HEC in 13 TBE; 10 selectrokinetic injection at 8.5 kV; separation at 8 kV.


panel (T1) shows a DNA fragment pattern indicative of apolyclonal population, with the stained acrylamide gel(inset) displaying an equivalent banding pattern. Becauseof the band (gel) and peak (capillary) indicated by thearrows, this patient was diagnosed as negative/equivocal.A monoclonal fragment pattern is illustrated in the mid-dle panel (T2, with acrylamide gel inset). For this sample,the diagnosis is obvious, based on the presence of adominant band on the gel and the corroborative peak inthe electropherogram. The corresponding PCR-amplifiedIGH rearrangement products were also analyzed by CE,and the results are shown in Fig. 2B along with thepBR322 HaeIII digest marker (bottom panel). The toppanel (B1) shows a positive sample, with a broad band onthe gel being partially resolved into two peaks by CE; thisis unequivocal identification of a monoclonal population.Conversely, the middle panel (B2) illustrates a character-istic profile for a sample negative for a B-cell malignancy,with the widespread polyclonal population of fragmentsdisplayed as a sawtooth pattern produced by the high-resolution separation.

microchip-based electrophoretic analysis oftcrg and igh gene rearrangementsEfficient microchip-based electrophoretic analysis re-quires a multichannel electrophoretic microchip and anoptical apparatus capable of sensitive fluorescence detec-tion (Fig. 3). The microchip illustrated in Fig. 3A containseight independent microchannels with the correspondingreservoirs for each. Although parallel analysis can beperformed, only single-channel analysis was required in

this initial study. Injection of DNA was accomplishedelectrokinetically at the intersection between the separa-tion channel (50 mm wide and 10 mm deep) connecting theinlet and outlet reservoirs and the cross channel (100 mmwide and 10 mm deep) that connects the sample andsample waste reservoirs. The sample reservoir-to-samplewaste distance was minimized to negate any electro-phoretic artifacts that may result from separation in thecross channel itself during injection. With a separationdistance (cross channel intersection-to-detection point) of4.2 cm, the effective length of the separation channel wassevenfold shorter than the capillary used to generate theseparations shown in Fig. 2. Detection was accomplishedby focusing the laser beam on the center of the microchan-nel at the detection window of the microchip (shown inFig. 3A). Fig. 3B shows the configuration of the home-built confocal epifluorescence microscopic detection sys-tem. The system uses an argon ion laser that is filteredspectrally and expanded to fill the back aperture of amicroscope objective lens to focus the beam tightly insidethe channel. Fluorescence is collected via the same objec-tive lens and, after passage through the optical compo-nents displayed in Fig. 3B, is detected by a photomulti-plier tube. A single software program controls both thedetection system and the electrophoresis high-voltagesystem, synchronizing the separation with the detection.This configuration is ideal for microchip-based electro-phoresis because it allows for sensitive detection whileleaving the microchip surface completely accessible formanipulation of sample, solutions, and electrophoresishardware.

Fig. 3. Configuration of the electro-phoretic microchip (A) and layout ofthe confocal fluorescence system foron-chip detection (B).PMT, photomultiplier tube.


Microchip-based electrophoretic separation of DNA markers.After the microchannels were coated with PA, the micro-channels were filled and equilibrated with 10 g/L HEC in13 TBE containing micromolar concentrations of YO-PRO-1 as an intercalating dye for double-stranded DNAdetection. A typical separation of DNA markers, HaeIIIdigest of pBR322, produced the microchip electrophero-gram shown in Fig. 4A. The high resolution observed iscomparable to separation of the same DNA markers in the37 cm 3 50 mm capillary (Fig. 4B).

Analysis of TCRG and IGH gene rearrangements. The sam-ples evaluated by gel electrophoresis (results shown inFig. 1) were also analyzed by microchip electrophoresiswith optimized conditions for sample preparation (sam-ple desalting and a 1:10 dilution in 10 mmol/L Tris, 1mmol/L EDTA). Fig. 5 shows a comparison of the capil-lary and electrophoretic microchip formats for separationof the TCRG gene rearrangement products for four sam-ples. The profile for sample T1 (Fig. 5B) displays the sameseries of low-abundance fragments (fragment size range,

150–250 bp) as observed for gel and capillary separationsand also allows for detection of the suspicious single band(fragment size, ;140 bp on the gel; indicated by the arrowin Fig. 5) that led this sample to be classified as negative/equivocal. A typical negative sample profile is displayedfor sample T4, which contrasts the profiles for positivesamples T2 and T3 that clearly demonstrate clonality.

Analysis of the IGH gene rearrangement products viamicrochip electrophoresis also yielded profiles similar tothose seen for the CE analysis (Fig. 6). ElectropherogramsB2 and B3 are representative of negative samples, whereaselectropherograms B1 and B4 are characteristic of positivesamples. A slight reduction in resolution can be seen inthe electropherograms for sample B1 with the microchip(Fig. 6B) compared with the capillary (Fig. 6A) separa-tions. However, it is improbable that this slight reductionwill influence the diagnostic capacity.

Evaluation of PVP as a sieving matrix for DNA separations inbare silica channels. PVP, a polymer that has been shown tofunction as a sieving matrix for DNA and also a dynamiccoating for bare silica capillaries (27 ), was evaluated forDNA separations in bare silica channels. Dynamic coat-ings can be added to the separation buffer to temporarilyadhere to, and therefore deactivate, the silica surface. Thisis in contrast to traditional capillary coatings, such as theFC and PA coatings, that are covalently bound to thecapillary surface and require time-consuming coatingprocesses. Initially, this dynamic coating/sieving matrixpolymer was tested in a 37 cm 3 50 mm bare silicacapillary for comparison with the standard HEC/FC-coated capillary approach. Fig. 7A shows that the PVP/bare silica approach for separating the DNA fragments(HaeIII digest of pBR322) provides effective resolution ofDNA in less than 15 min (Fig. 7A) in addition to achievingsingle-base resolution as evidenced by the separation ofthe 123- and 124-bp fragments.

Because of the potential advantages of a dynamiccoating sieving matrix and the positive results obtainedwith the capillary format, PVP was tested as a DNAsieving polymer for microchip electrophoresis. Fig. 7Bshows the separation obtained in a 4.2-cm channel. Theresolution obtained with the microchip was clearly infe-rior to that obtained with HEC on the microchip with thesame channel length (Fig. 4A). Despite the lower resolu-tion, the IGH gene rearrangement products were analyzedwith the PVP system to test applicability to a clinicallyrelevant analysis (Fig. 8). As shown, those samples thatwere positive (B1 and B4) and negative (B2 and B3) forB-cell clonality are clearly distinguishable. Similar profilesare observed for the negative samples (B2 and B3) whencompared with the capillary and microchip separations inHEC (Fig. 6, B2 and B3). However, the positive sample (B1and B4) profiles display multiple peaks that were notobserved in the HEC systems (Fig. 6, B1 and B4).

Fig. 4. Separation of the HaeIII digest of pBR322 using microchip (A)and CE (B) with 10 g/L HEC in 13 TBE and an FC-coated capillary andPA-coated channel.Other conditions as stated in text.


DiscussionMany clinical diagnostic assays rely heavily on molecularbiology techniques for identification of nucleic acid se-quences indicative of disease. The cornerstone of manysuch analyses is the amplification of target sequences viaPCR; such is the case for the diagnosis of lymphoprolif-erative disorders. Although dependent on the type ofT-cell lymphoproliferative disorder, PCR/slab gel electro-phoresis methods on average detect 75–90% of the clonalpopulations that are detected by Southern blot(20, 22, 28). Validation studies in the Department of Pa-thology at the University of Pittsburgh Medical Centerhave demonstrated that the IGH PCR method used inthese studies shows a clonal band in 70% of the casesdemonstrating rearrangement by the gold standard,Southern blot analysis, with a probe specific for the heavychain joining region. Despite the lower detection rates ofthe PCR/slab gel assay, PCR methods offer the advan-tages of decreased turnaround time, the ability to analyze

a wider range of specimen types and amounts, and insome cases, greater sensitivity for detection of smallertumor burden. As a result of these advantages, the PCR/slab gel assays are commonly used, as evidenced by theirroutine use in many clinical laboratories including theMolecular Diagnostics Laboratories at both the MayoClinic and the University of Pittsburgh Medical Center.However, substantial room for improvement in the meth-odology clearly still exists, especially with respect toexpediting the separation portion of the assay.

Although slab gel electrophoresis is clearly the estab-lished method for interrogation of PCR-amplified DNAfragments, capillary-based electrophoresis has been gain-ing momentum since its introduction at the beginning ofthis decade as an alternative format (3 ). Although notperformed in this work, sensitivity comparisons havebeen performed and have shown that sensitivity is equiv-alent between the CE and slab gel techniques (24 ). Bothelectrophoretic techniques (capillary and slab gel) can

Fig. 5. Capillary and microchip electrophoresis of TCRG gene rearrangement products.Capillary (A) and microchip (B) separations of samples labeled in Fig. 1A. Sample T1 was diagnosed as negative/equivocal because of the band indicated by the arrow.Sample T4 is a characteristic separation of a negative sample, whereas T2 and T3 represent positive samples. Separations performed in FC-coated capillaries andlinear PA-coated channels using 10 g/L HEC in 13 TBE. Other conditions as stated in text.


detect a positive control diluted 1:100 in thymus-derivedDNA but failed to detect a 1:1000 dilution. Microchipelectrophoresis, a miniaturized format of CE, has begun toattract attention as a higher efficiency embodiment of CE(14, 15) where the potential for parallel processing andintegration of other chemistries can be realized (9, 23).

Oda et al. (24 ) achieved TCRG gene rearrangementproduct separation by CE in 17 min using 10 g/L HEC ina 47 cm 3 50 mm DB-17-coated capillary at 260 V/cm. Theresults given in Fig. 2A parallel these results and showthat the amplified products of TCRG genes can be sepa-rated in less than 15 min via CE in FC-coated m-SIL-FCcapillaries, which were found to provide better reproduc-ibility and longer lifetimes for DNA analysis than DB-17-coated capillaries. The resolution obtainable with theconditions described in the original study (24 ), althoughadequate for the T-cell analysis, was not readily applica-ble to B-cell analysis. The smaller fragments (80–140 bp)amplified with B-cell gene rearrangements (see Fig. 1)

require higher resolution conditions in this region thanconditions that provide adequate separation of the prim-ers and product.

Perhaps most important is the demonstration that theoptimized conditions developed with CE could be trans-lated directly to the microchip for DNA analysis. In thecapillary system, FC-coated m-SIL-FC capillaries wereused because of their adequate stability and reproducibil-ity; however, microchips with FC-coated channels are notcommercially available. The method of Hjerten (26 ),which provides a relatively stable PA coating of the silicasurface, is one of the most common methods utilized forsilica deactivation. Using this methodology, comparisonof the microchip-based electrophoretic separation withseparation in the capillary clearly demonstrates a substan-tial decrease in analysis time with a negligible loss ofresolving power. In fact, a standard approach for calcu-lating resolution (29 ) showed that resolution on themicrochip was higher for the larger fragments (increased

Fig. 6. IGH gene rearrangement product separations in a capillary and electophoretic microchip.Separation of samples labeled in Fig. 1B via capillary (A) and microchip (B) electrophoresis. Samples B1 and B4 are positive, and samples B2 and B3 are negative.Separations performed using 10 g/L HEC in 13 TBE in FC-coated capillaries and PA-coated channels. Other conditions as stated in text.


;28% for 267- to 587-bp fragments) but lower for theshorter fragments (decreased ;24% for 184- to 234-bpfragments; decreased ;55% for 51- to 104-bp fragments).Despite these minor changes in resolving power (whichcan be accounted for by differences in the silica surfaces,the surface coatings, and/or discrepancies in the appliedvoltages), it is clear that the same diagnostic informationcan be extracted from both systems. This is exemplified bythe ability to detect the suspicious band/peak (fragmentsize ;140 bp on the gel; indicated by the arrow in Fig. 5)in sample T1, which led this sample to be classified asnegative/equivocal. This is of seminal importance whenenvisioning how multiplex microchip electrophoresis(parallel analysis of many samples) could impact through-put and turnaround time in molecular diagnostics. Thebenefits (e.g., in this particular case, of reducing analysistime from 2.5 h with slab gel electrophoresis to 160 s withmicrochip electrophoresis with no apparent compromisein the quality of the information provided) are obvious.

Although the performance (resolution) of the micro-chip and capillary systems were comparable in terms ofextracting diagnostic information, the robustness of thetwo platforms differ at this point. Effective analysis of thePCR samples via CE required only a simple 1:10 dilutionof the PCR product with Tris/EDTA buffer. This is

consistent with the previous observations that the saltconcentrations typically associated with standard PCRmixtures do not to cause extensive problems with capil-lary-based separations (30–32). In fact, studies haveshown that PCR product injected directly into the capil-lary for CE analysis does not affect the quality of theseparation (32 ). This contrasts results with the PA-coatedmicrochip, which appeared to be extremely sensitive tothe high salt concentrations of the PCR mixture. This is inagreement with previous studies in which either desalting(33, 34) or extensive dilutions (14 ) were necessary samplepretreatment steps. The loss of electrophoretic functional-ity is probably the result of deterioration of the channelcoating because exposure to salt led to loss of efficiency(broader peaks) and increased migration times. Althoughdesalting the samples before microchip analysis circum-vented this problem, this presents a disadvantage becauseit burdens the test protocol with additional steps that

Fig. 7. Separation of a DNA marker, HaeIII digest of pBR322, using PVPin uncoated capillaries and channels.(A), capillary separation of the DNA marker by 20 g/L PVP in 13 TBE in a baresilica 37 cm 3 50 mm capillary. (B), separation on a bare silica microchip using20 g/L PVP in 13 TBE.

Fig. 8. Separation of IGH gene rearrangement products using 20 g/LPVP in 13 TBE on an uncoated microchip.Samples correspond to those in Fig. 1B and Fig. 6. Samples B1 and B4 arepositive, and samples B2 and B3 are negative.


eventually impact the cost of the test. However, it isnoteworthy that this problem may be resolved by usingdifferent surface passivation approaches. Hofgartner et al.(35 ) have recently shown that use of chemistries thatprovide a more stable surface coating can obliterate thesalt-sensitive nature of microchip electrophoresis. In thatstudy, the successive injection of several hundred PCRsamples that had not been desalted or extensively dilutedwas achieved without impacting the quality of the sepa-ration.

Under the conditions described in this report, replicateanalysis of the same sample or of many different samplesshowed reproducible profiles, provided that the PCRsamples were desalted before analysis. However, an ex-tensive statistical analysis of reproducibility needs to beperformed to evaluate the robustness of this electro-phoretic platform. It is clear that the reproducibility ofDNA electrophoresis and robustness of the microchip willbe directly related to how well the microchannel surface ispassivated. As far as covalent modification of the surfaceis concerned, the PA coating appears adequate but saltsensitive, and improved deactivation chemistries (35 ) arelikely to improve on this. Perhaps the most convenientapproach will be the use of a polymer that not onlyprovides sieving of the DNA but also deactivates thesurface in a dynamic fashion. This is attractive becausemost covalent coatings that deactivate silica surfaces aredifficult to produce uniformly and reliably and havelimited lifetimes. The virtues of PVP (27, 36) and poly-(ethylene oxide) (37 ) as polymers that can dynamicallydeactivate silica surfaces and provide a sieving matrix forcapillary DNA separations have been proclaimed. Weshow, for the first time, that the microchannel can bedynamically deactivated with PVP (Mr 360 000) and PCR-amplified DNA resolved with the same polymer. Theresolution was poorer than that observed with HEC,possibly because the surface of the microchannel wasrougher than in a capillary (38 ) and, hence, not aseffectively deactivated. The multiple peaks observed inthe IGH gene rearrangement products (Fig. 8) were notseen in the HEC/FC capillary system (Fig. 6A), butsimilar multiple peak profiles were obtained for the samesamples separated in bare silica capillaries using PVP.Despite the multiple peak profile, which is currentlyunder investigation, the use of PVP as a dynamic coating/sieving matrix for DNA separations in uncoated electro-phoretic microchips under unoptimized conditions allowsdiagnostic information about IGH gene rearrangement tobe obtained. Although there are idiosyncratic issues to beresolved, these results illustrate the potential of dynamiccoatings that also function as sieving matrices for simpli-fying the use of uncoated electrophoretic microchips formolecular diagnostics.

In conclusion, slab gel, capillary, and microfabricated chipelectrophoresis have been used to analyze TCRG and IGHgene rearrangements, molecular diagnostic assays used to

detect T- and B-cell lymphoproliferative disorders. Thesame information obtained from the slab gel and CE wasextracted by the electrophoretic microchip, except with ananalysis time that was dramatically reduced. Decreasingthe time needed for electrophoresis from as long as 2.5 hon the slab gel to 15 min on the capillary and finally to160 s on the electrophoretic microchip has obvious impli-cations for diagnostic testing. Although the microchipseparations primarily used a PA channel coating and HECfor separation of the DNA fragments, the use of PVP as adynamic coating/sieving matrix for DNA separationswas shown to be feasible. This approach may eliminatethe time-consuming channel coating process and irrepro-ducibility problems accepted with silica coatings, allow-ing for minimal preparation and rapid analysis. Inherentin considering electrophoretic microchip technology as ananalytical tool in clinical diagnostics will be its compari-son with conventional methods. In this report, microchipsare shown to reduce electrophoretic analysis time by60-fold while maintaining full diagnostic capacity. This isthe first step toward accepting this technology as a newparadigm in diagnostics.

This work was funded in part by the National CancerInstitute (Grant 1R21CA78865-01 to J.P.L.), the LubrizolCorporation (to N.M.), and ACS Analytical Division Sum-mer Fellowship sponsored by the R.W. Johnson Pharma-ceutical Institute (to N.M.). We would like to thank Dr.Zhili Huang, University of Pittsburgh, for designing,building, and writing the Labview program for the mi-crochip high-voltage power supply, and Beckman Coulterfor instrumentation.

References1. Rapley R, Theophilus BD, Bevan IS, Walker MR. Fundamentals of

the polymerase chain reaction—a future in clinical diagnostics.Med Lab Sci 1992;49:119–28.

2. Reiss J. The polymerase chain reaction and its potential role inclinical diagnostics and research. J Intern Med 1991;230:391–5.

3. Oda RP, Bush VJ, Landers JP. Clinical applications of capillaryelectrophoresis. In: Landers JP, ed. Handbook of capillary electro-phoresis, 2nd ed. Boca Raton: CRC Press, 1997:639–73.

4. Huang XC, Quesada MA, Mathies RA. Capillary array electrophore-sis using laser-excited confocal fluorescence detection. AnalChem 1992;64:967–72.

5. Woolley AT, Hadley D, Landre P, deMello AJ, Mathies RA, NorthrupMA. Functional integration of PCR amplification and capillaryelectrophoresis in a microfabricated DNA analysis device. AnalChem 1996;68:4081–6.

6. Kopp MU, de Mello AJ, Manz A. Chemical amplification: continu-ous-flow PCR on a chip. Science 1998;280:1046–8.

7. Shoffner MA, Cheng J, Hvichia GE, Kricka LJ, Wilding P. ChipPCR. I. Surface passivation of microfabricated silicon-glass chipsfor PCR. Nucleic Acids Res 1996;24:375–9.

8. Burns MA, Johnson BN, Brahmasandra SN, Handique K, WebsterJR, Krishnan M, et al. An integrated nanoliter DNA analysis device.Science 1998;282:484–7.

9. Colyer CL, Tang T, Chiem N, Harrison DJ. Clinical potential of


microchip capillary electrophoresis systems. Electrophoresis1997;18:1733–41.

10. Colyer CL, Mangru SD, Harrison DJ. Microchip-based capillaryelectrophoresis of human serum. J Chromatogr A 1997;781:271–6.

11. Chiem N, Harrison DJ. Microchip-based capillary electrophoresisfor immunoassays: analysis of monoclonal antibodies and the-ophylline. Anal Chem 1997;69:373–8.

12. Koutny LB, Schmalzing D, Taylor TA, Fuchs M. Microchip electro-phoretic immunoassay for serum cortisol. Anal Chem 1996;68:18–22.

13. von Heeren F, Verpoorte E, Manz A, Thormann W. Micellarelectrokinetic chromatography separations and analyses of bio-logical samples on a cyclic planar microstructure. Anal Chem1996;68:2044–53.

14. Cheng J, Waters LC, Fortina P, Hvichia G, Jacobson SC, RamseyJM, et al. Degenerate oligonucleotide primed-polymerase chainreaction and capillary electrophoresis analysis of human DNA onmicrochip-based devices. Anal Biochem 1998;257:101–6.

15. Schmalzing D, Adourian A, Koutny L, Ziaugra L, Matsudaira P,Ehrlich D. DNA sequencing on microfabricated electrophoreticdevices. Anal Chem 1998;70:2303–10.

16. Hofmann O, Che D, Cruickshank KA, Muller UR. Adaptation ofcapillary isoelectric focusing to microchannels on a glass chip.Anal Chem 1999;71:678–86.

17. Rodriguez I, Zhang Y, Lee HK, Li SFY. Conventional capillaryelectrophoresis in comparison with short-capillary capillary elec-trophoresis and microfabricated glass chip capillary electrophore-sis for the analysis of fluorescein isothiocyanate anti-humanimmunoglobulin G. J Chromatogr A 1997;781:287–93.

18. Walker PA, Morris MD, Burns MA, Johnson BN. Isotachophoreticseparations on a microchip-normal Raman spectroscopy detec-tion. Anal Chem 1998;70:3766–9.

19. McCormick RM, Nelson RJ, Alonso-Amigo MG, Benvegnu DJ,Hooper HH. Microchannel electrophoretic separations of DNA ininjection-molded plastic substrates. Anal Chem 1997;69:2626–30.

20. Trainor KJ, Brisco MJ, Wan JH, Neoh S, Grist S, Morley AA. Generearrangement in B- and T-lymphoproliferative disease detected bythe polymerase chain reaction. Blood 1991;78:192–6.

21. Coad JE, Olson DJ, Lander TA, McGlennen RC. Molecular assess-ment of clonality in lymphoproliferative disorders. II. T-cell recep-tor gene rearrangements. Mol Diagn 1997;2:69–81.

22. Rezuke WN, Abernathy EC, Tsongalis GJ. Molecular diagnosis of B-and T-cell lymphomas: fundamental principles and clinical appli-cations. Clin Chem 1997;43:1814–23.

23. Tian H, Huhmer AFR, Landers JP. New vistas in clinical diagnos-tics: from capillary electrophoresis to rapid diagnostics on glasschips. Curr Top Anal Chem 1998;1:171–89.

24. Oda RP, Wick MJ, Rueckert LM, Lust JA, Landers JP. Evaluation ofcapillary electrophoresis in polymer solutions with laser-inducedfluorescence detection for the automated detection of T-cell generearrangements in lymphoproliferative disorders. Electrophoresis1996;17:1491–8.

25. Ocvirk G, Tang T, Harrison DJ. Optimization of confocal epifluores-cence microscopy for microchip-based miniaturized total analysissystem. Analyst 1998;123:1429–34.

26. Hjerten S. High-performance electrophoresis: elimination of elec-troendoosmosis and solute adsorption. J Chromatogr 1985;347:191–8.

27. Gao Q, Yeung ES. A matrix for DNA separation: genotyping andsequencing using poly (vinylpyrrolidone) solution in uncoatedcapillaries. Anal Chem 1998;70:1382–8.

28. Lust JA. Molecular genetics and lymphoproliferative disorders.J Clin Lab Anal 1996;10:359–67.

29. Snyder LR, Kirkland JJ. Introduction to modern liquid chromatog-raphy, 2nd ed. New York: Wiley Interscience, 1979:863pp.

30. Ulfelder KJ, Schwartz HE, Hall JM, Sunzeri FJ. Restriction fragmentlength polymorphism analysis of ERBB2 oncogene by capillaryelectrophoresis. Anal Biochem 1992;200:260–7.

31. Guttman A, Schwartz JE. Artifacts related to sample introduction inCE affecting separation performance and quantitation. Anal Chem1995;67:2279–83.

32. Pancholi P, Oda RP, Mitchell PS, Landers JP, Persing DH. Diag-nostic detection of herpes simplex and hepatitis C viral ampliconsby capillary electrophoresis: comparison with Southern blot detec-tion. Mol Diagn 1997;2:27–37.

33. Wooley AT, Sensabaugh GF, Mathies RA. High-speed DNA geno-typing using microfabricated capillary array electrophoresis chips.Anal Chem 1997;69:2181–6.

34. Wooley AT, Mathies RA. Ultra-high-speed DNA fragment separa-tions using microfabricated capillary array electrophoresis chips.Proc Natl Acad Sci U S A 1994;91:11348–52.

35. Höfgartner WT, Hühmer AFR, Landers JP, Kant JA. Rapid diagnosisof herpes simplex encephalitis using microchip Electrophoresis ofPCR products. Clin Chem 1999;45:in press.

36. Braun B, Blanch HW, Prausnitz JM. Capillary electrophoresis ofDNA restriction fragments: effect of polymer properties. Electro-phoresis 1997;18:1994–7.

37. Fung EN, Yeung ES. High-speed DNA sequencing by using mixedpoly (ethylene oxide) solutions in uncoated capillary columns. AnalChem 1995;67:1913–9.

38. Simpson PC, Woolley AT, Mathies RA. Microfabrication technologyfor the production of capillary array electrophoresis chips.J Biomed Microdevices 1998;1:7–26.


Documents

Molecular Diagnostics on Microfabricated Electrophoretic ... · slab gel electrophoresis to identify products and provide semiquantitative data. CE provides a more rapid, auto-mated