| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Department of Biotechnology, University of Turku, Tykistökatu 6A, FIN-20520 Turku, Finland.
a Author for correspondence. fax 358-2-3338050; e-mail harri.harma{at}utu.fi
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Genetically engineered thiol-Fab or thiolated monoclonal antibodies (mAbs) were covalently attached onto uniformly sized 60-µm maleimide-activated microparticles. Free and total PSA were detected with europium- or terbium-labeled Fab fragments on a single microparticle using a microfluorometer in a time-resolved mode.
Results: The detection limit of the free- and total-PSA assays (mean + 3 SD of zero calibrator) was 0.35 µg/L, with a total volume of 330 nL per particle. An excellent correlation was found in microparticle and microtiter-well assays for 21 serum samples: slopes for free and total PSA were 1.06 ± 0.03 and 1.03 ± 0.02, respectively (Sy|x = 0.084 and 0.057 µg/L), with intercepts of 0.013 ± 0.018 and 0.013 ± 0.017 µg/L (R >0.99). Furthermore, the particle-immobilized Fab fragment had a PSA binding capacity 1.5-fold higher than the intact mAb capacity on a single microparticle. Capacity, kinetics, and sensitivity of the Fab fragment and intact mAb assays in the microparticle and microtiter well formats are discussed.
Conclusions: With site-specific (cysteine tail) covalent attachment of Fab fragments on a microparticle, subattomole amounts of PSA can be detected quantitatively.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Down-scaling the physical dimensions of a reaction cell is beneficial, but reducing the size of the biomolecules can also be advantageous in various assay formats. For example, fragments of an antibody, such as Fab or scFv fragments, can improve surface capacity and, hence, the kinetics of assays. An improved capacity is extremely important in miniature assay formats because, generally, less surface area is available for solid-phase reactions. Correct orientation of a surface capture antibody also increases the potential use of antibody fragments, which can be tailored for solid-phase coupling by the introduction of a site-specific group, such as a thiol group, through genetic engineering (Meretoja et al., submitted for publication). Therefore, a combination of microparticles with large surface areas and antibody fragments is an attractive approach. The use of recombinant antibody fragments can reduce false-positive and -negative results by eliminating the effects of heterophilic antibodies, and complement and rheumatoid factors (6)(7). Less expensive and time-consuming production of recombinant antibody fragments in g/L titers further favors the use of fragmented antibody species (8).
The RecAll prostate-specific antigen
(PSA)1
assay, based entirely on site-specific recombinant Fab
fragments for capturing and detecting the analyte, has been
demonstrated previously (9). The Fab fragments were
site-specifically labeled with nonfluorescent lanthanide chelates,
whose signals were enhanced with fluorescence enhancement solution. The
use of intrinsically fluorescent lanthanide chelates allows more
compact and rapid assays as well as multiplex assays to be performed
because no additional enhancement step is required. We describe here a
PSA assay that relies on the use of recombinant Fab fragments,
intrinsically fluorescent lanthanide chelates, and porous
microparticles as solid support (Fig. 1
). To detect low PSA concentrations, we used an advanced
microscope-based detection system to measure analyte quantitatively. We
previously have shown that subattomole quantities of PSA can be
detected by this time-resolved fluorescence detection technology
(10). For the first time, clinical functionality of this
assay concept is demonstrated.
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
microparticle coating
The microparticles were coated with specific thiolated monoclonal
antibodies (mAbs) or recombinant Fab fragments against PSA (H117-mAb-SH
or H117-Fab-SH). The H117-Fab-SH was produced in our laboratories by
genetic engineering for insertion of a free cysteine residue at the
COOH terminus of the Fd chain (Meretoja et al. and Lamminmäki et
al., submitted for publication).
The H117-mAb was thiolated with N-succinimidyl
S-acetylthioacetate (SATA; Pierce). SATA was dissolved in
dimethyl sulfoxide to a concentration of 4.7 mol/L, and 4 µL
of this solution was added to a H117-mAb solution in a total volume of
400 µL. The H117-mAb solution contained 850 µg of intact antibody
in a 50 mmol/L phosphate buffer, pH 7.8, containing 1 mmol/L EDTA. This
reaction was carried out at room temperature for 30 min. Thiolated mAb
was purified from unreacted SATA by NAP-5 and NAP-10 columns (Amersham
Pharmacia Biotech). The thiolated mAbs were eluted with 50 mmol/L
phosphate buffer, pH 7.8, containing 1 mmol/L EDTA, and concentrated
with a Centricon 50 concentrator (Millipore) to 
100 µL. The latent
thiol group was released by the addition of 0.5 mol/L hydroxylamine in
50 mmol/L phosphate buffer, pH 7.5, containing 25 mmol/L EDTA. The
number of thiol groups per mAb were quantified with Ellman reagents
(Pierce).
The recombinant H117-Fab-SH fragment was stored at 4 °C in a 50 mmol/L glycine-Tris buffer, pH 6.0, containing 1 mmol/L dithiothreitol. Before coating, dithiothreitol was removed by NAP-5 and NAP-10 columns with 50 mmol/L phosphate buffer, pH 7.8, containing 25 mmol/L EDTA (buffer A) or 50 mmol/L borate buffer, pH 8.3, containing 5 mmol/L EDTA (buffer B). Immediately after the purification, Tris[2-carboxyethyl]phosphine hydrochloride (Pierce) was added to a concentration of 20 µmol/L. The eluted H117-Fab-SH was concentrated with a Centricon 30 concentrator.
Approximately 15 000 amino-grafted particles were activated with
0.5–50 mmol/L N-[
-maleimido caproyloxy]
sulfosuccinimide ester (EMCS; Pierce) or 0.1–5 mmol/L
sulfosuccinimidyl[4-iodoacetyl] aminobenzoate (sulfo-SIAB; Pierce) in
500 µL of buffers A and B, respectively. The activation was carried
out by end-over-end rotation (Rotamix; Heto Lab Equipment) for 30–80
min at room temperature. The sulfo-SIAB activation was performed in the
dark. The activation solution was removed, and 10–60 ng of H117-Fab-SH
or 115–170 ng of H117-mAb-SH was added to the microparticles. The
reaction was mixed in a Rotamix at room temperature overnight
(sulfo-SIAB in dark) and stopped by incubation of the particles with
400 µL of 0.01 mol/L cysteine in 50 mmol/L phosphate buffer, pH 7.8,
in an end-over-end rotation for 2 h at room temperature.
Finally the particles were washed four times with 2 mmol/L Tris buffer,
pH 7.5, containing 0.1 g/L Tween 20 and stored in the Tris-HCl
solution at 4 °C.
microparticle binding capacity
The PSA-binding capacities of antibody-coated beads were
determined using Eu(III)-labeled free PSA to saturate the
surfaces. The assays were performed as described elsewhere
(11).
labeling of Fab FRAGMENTS
The PSA-specific Fab fragments were produced in our laboratories
(Meretoja et al. and Lamminmäki et al., submitted for
publication). Briefly, the genes coding for Fabs were cloned from
hybridoma cells producing corresponding mAbs, and the clones were
transferred to pKKtac vector. The Fabs were produced
to Escherichia coli periplasmic space and purified using a
cation-exchange chromatography column (SP Streamline; Amersham
Pharmacia Biotech) and a protein G affinity column (Amersham Pharmacia
Biotech). The 5A10-Fab recognizing the free form of PSA was labeled
with the europium(III) chelate of
2,2',2'',2'''-{{6,6'-{4''-[2-(4-isothiocyanatophenyl)ethyl]1H-pyrazol-1,3-diyl}bis(pyridine)-2,2'-diyl}
bis(methylenenitrilo)}tetrakis(acetic acid) (chelate I)
(12). The H50-mAbs recognizing free and complexed forms of
PSA were labeled with the same ligand, using terbium(III) ion. The
labeling procedure has been described previously (11). The
labeling of the Fab fragments was 4.5 europium or terbium molecules per
Fab when determined against Eu(III) and Tb(III) calibrators. The
affinity constant of the labeled 5A10-Fab fragment was determined as
described previously (9).
assay kinetics
The kinetics of the two-step PSA immunoassays were determined in a
total volume of 330 nL/microparticle (20 microparticles/reaction tube)
using 67 nL of sample per microparticle at room temperature. The assays
were carried out at a PSA concentration of 50 µg/L in the assay
buffer (Wallac). After incubation with PSA in an end-over-end rotation
or a vortex-type mixer (Vortex; Scientific Industries), the
microparticles were washed three times with 50 mmol/L Tris-HCl, pH 7.8,
containing 9 g/L NaCl, 1 g/L Germall II, 5 g/L Tween 40, and 5 g/L
Tween 20 (buffer C). In the tracer incubation, 1.5 ng of
5A10-Fab-Eu(III) or 4.5 ng of 5A10-mAb-Eu(III) was incubated in
a total volume of 330 nL/particle and washed with buffer C. For
detection, the particle suspension was centrifuged, and the particles
were pipetted on a Cyclopore track-etched membrane (5 µm; Whatman)
and dried out through the membrane with a tissue. The quantity of the
analyte was measured directly on the surface of the microparticle with
a microfluorometer (Wallac) in a time-resolved mode (10).
One-step PSA assays were performed as above, but with the analyte and
tracers incubated in the same reaction.
The microtiter plate assays were carried out as recommended by the manufacturer (Wallac). Briefly, in the two-step assay, 25 µL of PSA calibrator (Dual PSA reagent set; Wallac) was incubated in the assay buffer for 1 h and the tracers for 2 h in total volumes of 125 and 200 µL, respectively. Finally, europium ions were dissolved from the chelate into 200 µL of the DELFIA enhancement solution (Wallac). The signal was measured with the Victor 1420 multilabel counter (Wallac).
calibration curves
The assay conditions for the calibration curves were as described
under "Assay Kinetics". The PSA incubation time was 180 min, and
the tracer was incubated with the PSA-bound microparticles for 10 min.
Ten replicates were measured for the zero concentration, and 8
replicates were measured for the calibration samples containing
analyte. The detection limit of the assays was calculated from
the zero calibrator plus 3 SD. Serum samples were a kind gift from Dr.
Maciej Kwiatkowski (Kantonspital, Aarau, Switzerland).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Originally, the microparticles carried epoxy groups on the surface as a result of the glycidylmethacrylate copolymer used in the polymerization process. Additional surface modification was made by Sintef to obtain primary amino groups. Two thiol-active linkers, sulfo-SIAB and EMCS, were investigated for optimum loading capacity and assay performance of H117-Fab. Equal Fab fragment-loading capacities were found with both linkers in the Eu(III)-labeled PSA assay when 0.5 mg of sulfo-SIAB and 10 mg of EMCS were used. Because nonspecific binding was increased in the PSA assay of sulfo-SIAB-activated microparticles, EMCS activation was used in the final mAb and Fab fragment immobilizations.
The PSA-binding capacities were 1.0 and 1.5 ng on single H117-mAb and H117-Fab microparticles, respectively. Maximum loading of the surface capturing antibodies was achieved when 170 ng of H117-mAb or 50 ng of H117-Fab was used per microparticle. Higher amounts of antibody molecules did not increase the binding of europium-labeled PSA on the particles.
labeling of Fab FRAGMENTS
Although site-specific bioconjugation was made possible through
recombinant Fab fragment production, the Fab fragments were not labeled
through the thiol groups in the current study because one chelate per
Fab fragment did not yield satisfactory signals. Fab fragments were
labeled through the amino groups, yielding a labeling of 4.5 chelate
molecules/Fab molecule for both europium and terbium chelates
I. The affinity constant for the labeled 5A10-Fab fragment
was calculated to determine whether relatively high labeling had an
impact on the PSA-binding efficiency. The assay revealed that the
affinity constant of the tracer 5A10-Fab, 2 x
109 L/mol, was not significantly
different from that of the nonlabeled Fab fragment, 3 x
109 L/mol.
kinetics
To increase the speed of the microparticle assay, two mixing
systems were investigated, end-over-end mixing and vortex mixing. In
end-over-end mixing, the assay solution and the microparticles did not
move actively, which was observed visually using a stroboscope. More
vigorous vortex mixing enhanced the movement of the microparticles,
substantially improving assay speed. In comparison with the
end-over-end mixing (two-step assay), kinetics in the one-step
PSA assay were approximately fourfold faster (Fig. 2
). Hence, vortex mixing was used in the subsequent PSA assays.
|
The kinetic curve of the PSA incubation with the microparticles in the
two-step assay is shown in Fig. 2
. Equilibrium was reached in 160
min. Reaction volume, amount of the tracer molecules, and mixing
efficiency were varied to test whether the smaller size of the Fab
fragment compared with the mAb had an effect on assay kinetics. No
significant kinetic difference was found in the Fab and mAb assays
(data not shown). Free PSA was used as a model, but in the assays,
complexed PSA should give identical results because the rate was
limited by the association rate to the solid-phase antibody as well as
the mixing efficiency. The assay kinetics were further studied using
different numbers of particles in the same assay volume. Increasing the
amount of particles by a factor of 10 decreased the total PSA assay
time by the same factor (data not shown).
calibration curves
The sensitivity of an immunoassay is strongly dependent on the
ratio of sample to total assay volume. The sample volume was one-fifth
of the total volume in all assays performed in this study. Volumes of
200, 330, and 1000 nL/microparticle were investigated to test the
effect of total volume on the efficiency in binding PSA. In
these assays, >90% of the analyte was captured. Because no
significant difference was found among various assay volumes, the
330-nL volume was chosen for further assays to achieve reasonable assay
speed and sensitivity.
The calibration curves and precision profiles at different free- and
total-PSA concentrations when the total assay volume was 330
nL/microparticle are shown in Fig. 3
. Eight individual microparticles were detected at each analyte
concentration. The assay had excellent linear response, and the
detection range was three orders of magnitude. The lower detection
limit was 0.35 µg/L for both free and total PSA when calculated from
the response for the zero calibrator plus 3 SD. The within-run
CV was well below 10% for all measured concentrations.
|
Correlation data for free PSA by the microparticle and the commercial
DELFIA microtiter plate assays are illustrated in Fig. 4
. Excellent correlation was found for serum samples. The slopes
for the microparticle and microtiter plate assays were 1.06 ±
0.03 and 1.03 ± 0.02 for free and total PSA, respectively
(Sy|x = 0.084 and 0.057 µg/L) with intercepts of
0.013 ± 0.018 and 0.013 ± 0.017 µg/L, and the regression
coefficient was >0.99.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A prerequisite for detecting any multiple analytes, such as free and total PSA, in submicroliter assay volumes is an advanced labeling technology. Fluorescence-based assays are suitable for simultaneous detection of such multiple targets. The sensitivity of assays performed with prompt fluorescent labels is limited by nonspecific fluorescence and autofluorescence. The practical sensitivity of biological assays can be improved by the use of luminophores with high specific activity and low nonspecific fluorescence, such as luminescent inorganic phosphors (14), cofluorescent chelates (15), lanthanide cryptates (16), and lanthanide chelates (10)(17)(18)(19). We previously have shown that with lanthanide chelates, 1 x 104 europium molecules can be detected on a single microparticle (10). To perform a functional assay on a microparticle, we labeled PSA-specific Fab fragments with europium and terbium chelate I. Relatively high labeling of the tracer Fab fragments, 4.5 chelate molecules/Fab molecule, was achieved. One could anticipate that such high labeling would impair the binding efficiencies of the antibody fragments, but the effect was found to be insignificant.
Intact H117-mAbs and H117-Fab fragments were immobilized onto the microparticle by the same thiol-coupling chemistry, enabling a comparison of the PSA-binding capacity of the two antibody surfaces. More available binding sites were found when Fab fragments were immobilized onto a microparticle than when intact mAbs were conjugated because of the smaller size and improved orientation (site-specific conjugation through cysteine residue) of the Fab fragment. We have obtained similar results for microtiter plate PSA assays in which biotinylated capture Fabs or mAbs were immobilized on a streptavidin surface (unpublished data).
The calculated PSA-binding capacity on a microparticle per surface area
(1 ng/5 x 10-6 m2 =
2 x 10-4 g/m2) was
approximately fivefold lower than on a commercial polystyrene
microtiter well (130 ng/1.4 x 10-4
m2 = 9 x 10-4
g/m2; Wallac) when H117-mAb was immobilized on
both surfaces. The microparticle surface area had been estimated with a
Brunauer-Emmet-Teller isoterm study by the manufacturer. This method
fails to determine whether all of the surface area is available for
protein binding. One can readily think that a porous medium containing
access-restricted channels decreases the available protein-binding
surface. Hence, the calculated surface area is likely to be
overestimated, yielding an actual PSA-binding capacity per surface area
close to that of a microtiter well. Thus, the idea of using single
porous microparticles in clinical immunoassays has a particular
interest because the assay volume can be reduced substantially from
that of the microtiter plate format because of the mobile and densely
packed solid phase. In the commercial DELFIA anti-PSA microtiter plate
assay, a total assay volume of 125 µL is recommended by the
manufacturer. If we consider that both microparticles and microtiter
wells hold the same PSA-binding capacity per surface area, the PSA
assay could be performed in 1 µL without sacrificing any of the
assay performance. The volume of a single microparticle is in the order
of 100 fL, which can be neglected when calculating a total assay volume
because the effect in any practical assay volume is negligible.
The kinetic comparison of labeled Fab fragments and intact antibodies was performed in the single microparticle PSA assay. We varied the amount of tracer molecules, assay volume, and mixing efficiencies in the tracer incubation. Because in the noncompetitive PSA immunoassay the tracer molecule was introduced in an excess into the reaction and the mixing of the reaction was apparently efficient, very similar kinetic behavior was found for both the Fab and mAb (this applies to mAbs recognizing both the free and complexed forms of PSA) assays, indicating that diffusion was not limiting in the current assays. Although no size-dependent kinetic improvement was found with recombinant Fab fragments, the genetic engineering technique is a promising tool to produce antibody fragments with improved on-rate values to increase the speed of an assay. The smaller size, however, becomes extremely important when high surface loading is of interest. With a simple kinetic simulation, one can show that the reaction rate is improved in a linear manner when the concentration of the surface capturing antibody is increased. In a microtiter plate format, we found improved PSA assay kinetics with increasing loading capacity when we used biotinylated Fab fragments (unpublished data). Several studies on porous microparticles have indicated that high loading capacity might not be an ideal approach kinetically. In these experiments, slow assay kinetics were likely to originate from restricted access of molecules into the pores and slow diffusion of molecules inside the pores (20)(21)(22). The use of small recombinant Fab fragments should, however, reduce these effects.
The surface-to-volume ratio of a microparticle assay plays a
predominant role in determining the limits of the assay speed, as
reported previously (11). A 10-fold increase in the particle
amount (and hence, a 10-fold decrease in the surface-to-volume ratio)
reduced the time for the PSA assay from 160 to 20 min (total
assay volume decreased from 330 to 35 nL). To perform a single-particle
PSA assay in 12 min (the saturation time in the commercial DELFIA
microtiter plate format), a total assay volume of 20 nL/particle should
be used. This volume is 50-fold lower than expected from the
surface-to-volume calculations of a microparticle and a microtiter
plate. This gives an idea of the magnitude of the kinetic limitation in
a densely packed porous solid phase, although the microparticle and
microtiter well assay formats cannot be compared fully, mainly because
of differences in liquid-transfer efficiencies. We believe that the
vortex mixing was very efficient because the motion of the solution and
microparticles was rapid when observed with a stroboscope. Hence, the
50-fold difference in the estimated value and experimental data likely
originated from the densely packed porous structure and not from a
reduced external mass transport. It is difficult to confirm
experimentally whether this effect is attributable to the porosity or
mass transport because the limited loading capacity of a 60-µm
compact microparticle does not allow an equal comparison. Attempts have
made to overcome the problem of mass transport in porous particles in
chromatographic media by using highly porous beads (pore size, 800 nm
in diameter) (23). Although some convection occurs in these
porous beads, flow around the particles is still preferred
(21). Thus, the passage of molecules within the pores is
considered to be controlled by diffusion, and no significant
improvement in kinetics would be expected if a larger pore size was
used.
In terms of kinetics, the single-microparticle approach with volumes <50 nL is feasible for rapid assays. Without system automation, such a low volume assay is not, however, practical. Today, numerous examples can be found in the literature where microtechnology has been used to make reaction cells, channels, pumps, and detection systems suitable for extremely low-volume assays (24). These technologies should enable the application of instrumentation to the single-microparticle concept.
We obtained excellent linear response and a detection range covering more than three orders of magnitude in the miniature Fab fragment immunoassay of free and total PSA when the analyte was detected on a single microparticle. The microparticle assay and the corresponding DELFIA immunoassay in microtiter plates correlated well when 21 serum samples were studied. In these microparticle assays, subattomole quantities were monitored. Although the clinical utility of the present assay has been verified, further improvements in sensitivity are needed to perform rapid PSA assays using low nanoliter volumes. However, fast analysis could be performed with analytes such as C-reactive protein in the range of clinical interest with the current method. During the past year, we have studied alternative time-resolved fluorescence labels to detect analyte concentrations that could not be monitored with the present labeling technology. Using these labels with higher specific activity, we could achieve an improved detection limit for single microparticles.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
-maleimido caproyloxy] sulfosuccinimide ester; and sulfo-SIAB, sulfosuccinimidyl[4-iodoacetyl] aminobenzoate. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-hydroxyprogesterone, immunoreactive trypsin, and creatine kinase MM isoenzyme in dried blood spots. Clin Chem 1992;38:2038-2043.[Abstract]
-chymotrypsin biocatalysts prepared by deposition for use in organic medium. Biotechnol Bioeng 1998;59:366-373.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  P. Tarkkinen, T. Palenius, and T. Lovgren Ultrarapid, Ultrasensitive One-Step Kinetic Immunoassay for C-Reactive Protein (CRP) in Whole Blood Samples: Measurement of the Entire CRP Concentration Range with a Single Sample Dilution Clin. Chem., February 1, 2002; 48(2): 269 - 277. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) using vortex
mixing, and for end-over-end (
) and vortex (•) mixing in the
one-step PSA assay.
