The immuassay handbook parte19

581 views
474 views

Published on

Published in: Health & Medicine
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
581
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
27
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

The immuassay handbook parte19

  1. 1. 67© 2013 David G. Wild. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/B978-0-08-097037-0.00006-3 Homogeneous immunoassays require only mixing of a sample and immunochemical reagents followed by detection. Immunochemical binding produces a physi- cally detectable signal that obviates the need to separate bound from free label. Because the rate of the binding reaction is not limited by slow diffusion to a surface, incubation times are fast, usually only a few seconds to a few minutes; and the non-separation assay protocols minimize the requirements for automation. At least the- oretically homogeneous methods are more sensitive than heterogeneous immunoassays. This is because the separation and washing steps of heterogeneous methods are inherently more error prone and tend to reverse weak binding reactions with an attendant decrease in sensitivity. However, because sample constituents are not removed by a wash step, variations in signal caused by nonspecific effects of the sample matrix often prevent realization of this potential advantage. Recent research has focused on methods that will circumvent the sample interference problem and offer sensitivities equal to or better than the best heterogeneous methods. Homogeneous methods have been developed for large and small analytes using both competitive and noncom- petitive protocols. Immunochemical binding can be fol- lowed either kinetically or after attainment of binding equilibrium. The unique and common characteristic of each of these methods is that they provide a mechanism for modifying the signal produced by a label as a function of an immunochemical binding event. This contrasts with heterogeneous methods, which depend on identifying the location of an inert label after a separation step. The majority of labels are therefore environmentally sensitive sensors that are responsive to local variations in pH, solute concentration, electric field, steric constraints, radiation intensity, solvation, etc. The first reported immunoassays were homogeneous but did not employ a label. They are attributed to Kraus, who in 1897 coined the term precipitin for the precipitate formed upon mixing an antigen and an antibody (Kraus, 1897). Likewise, the earliest immunoassays that employed a label were homogeneous (Meyer, 1922). Sheep erythro- cytes, serving as a label, were coated with human immuno- globulin, and anti-immunoglobulin antibodies appearing in rheumatoid arthritis patients were shown to cause read- ily visible clumping of the cells. This method became known as hemagglutination. Subsequently, with the advent of spectrophotometry, it was shown that the visibly scat- tered light in Kraus’ precipitin reaction could be measured by turbidimetry to quantitate precipitin formation (Boyden et al., 1947). However, the method was unsuitable for routine use. Not only was it very sensitive to other components in the sample but the signal was biphasic, increasing and then decreasing upon adding antigen to a fixed amount of antibody (frequently referred to as the prozone phenomenon or hook effect). This occurs because only when antibody and antigen are near equimolar concentrations can antibody bridge between antigens to produce a polymeric complex. At low antigen concentra- tions, the excess antibody coats every antigen molecule, and at high antigen concentrations, the excess antigen occupies every antibody-binding site. Particle Agglutination ERYTHROCYTES AND LATEX An important improvement in the evolution of homoge- neous immunoassays was the replacement of Meyer’s erythrocytes with more readily controllable latex particles (Singer & Plotz, 1956). Typically antibodies are bound to the surface of the latex particles or erythrocytes, which form aggregates when a polyvalent antigen is present. Alternatively, the particles can be coated with an antigen to permit detection of antibodies. Both competitive and sandwich assay architectures can be used. A noncompeti- tive“sandwich”assaycanbesetupwithtwoantibody-coated particles that bind a multivalent antigen. The more antigen that is present, the more agglutination occurs. Although not as troublesome as the precipitin reaction, a sufficient excess of antigen will produce a biphasic response. Alternatively, added antigen can inhibit antibody-induced aggregationofantigen-coatedparticles.This“agglutination inhibition” method avoids the biphasic response but is generally less sensitive. In each case, it is only necessary to combine the sample and reagents and optically measure the agglutination. One particularly innovative variant of this method was introduced by Agen Corp., which uses the patient’s own erythrocytes as the particles in an assay for HIV antibodies. An HIV antigen conjugated to antibodies that bind an erythrocyte surface antigen is added to a whole blood sample. The antigen immediately becomes bound to the erythrocytes, which subsequently aggluti- nate if HIV antibodies are present (see Fig. 1; Wilson et al., 1991). Visual observation of agglutination is moderately sensi- tive but of course is not quantitative. Most instrumental measurements rely on turbidimetry or nephelometry. Turbidimetry measures the intensity of a beam of light transmitted through the sample, and nephelometry mea- sures the light that is scattered at an angle away from the beam. Nephelometry is more sensitive but is more subject to interference from particulate matter in the sample. Sim- ple subtraction of sample background is not possible because light scattering is not linearly dependent on the concentration of the particles. Various strategies have been developed to avoid this problem. By monitoring the rate of change in a nephelo- metric signal, it is possible to reduce the problem suffi- ciently that standard urine assays (Abuscreen®, Roche Diagnostics) for drugs of abuse, and common serum pro- teins (ICS-II™, Beckman) can be quantitatively measured. Homogeneous Immunoassays Edwin F. Ullman (tullman@earthlink.net) C H A P T E R 2.3
  2. 2. 68 The Immunoassay Handbook Another approach is to detect individual particles in a flowing stream. These “particle counting immunoas- says” (PACIAs) estimate the change in the number of unaggregated particles during an immunochemical reac- tion (Masson et al., 1981). This method is found to be most reliable for lower concentration analytes. High sensitivity can be achieved by measuring the angular anisotropy of laser light scattered from a suspension of parti- cles (laser nephelometry). The light scattering is a sensitive function of the particle size, particularly when the wavelength of light is on the order of the size of the particles. The method provides greater sensitivity than simple nephelometry, but interference from particulate material in the sample and reagents is also increased (Von Schulthess et al., 1976). GOLD NANOPARTICLES A related method for detecting agglutination uses parti- cles that have greatly enhanced light scattering relative to latex or erythrocytes. Colloidal metals have high indices of refraction and produce particularly strong light scatter- ing. The scattering intensity is strongly dependent on particle size. For gold sols, relatively sharp apparent absorption and scattered light maxima are seen that increase from about 520 to 620nm as the size of the par- ticles increases from about 50 to 120nm. Gold particle concentrations in the low femtomolar range can be detected. The theoretical basis for this phenomenon has been described (Yguerabide & Yguerabide, 1998). Agglutination of gold particles changes the effective par- ticle size, which is reflected in a change in color of the sus- pension. Although exploration of the use of gold nanoparticles (AuNPs) is currently enjoying a renaissance, a number of “sol particle immunoassays” (SPIAs) based on this phenomenon were developed over 30 years ago. For example, a sandwich SPIA capable of detecting 5.4pM human placental lactogen (HPL) was carried out by mixing the sample with gold particles conjugated to anti-HPL antibodies and measuring the change in color with a color- imeter. Other assays, including competitive immunoassays for hormones, were also demonstrated (Leuvering et al., 1980). The current interest in nanotechnology has led to more detailed investigation of this method but with little improvement in sensitivity. Liu et al. (2008) described the use of dynamic light scattering (DNS) in place of light absorption to monitor the extent of agglutination of anti- body labeled with spherical AuNPs and another antibody labeled with gold nanorods. Despite improved control of particle size and shape and a more sophisticated detection method, the detection limit of 0.1ng/mL (4pM) of free- prostate-specific antigen (PSA) was essentially unchanged. However, by attaching gold particles to magnetic particles, magnetic field enhancement of clustering enabled DNS detection of as little as 140fM alpha-fetoprotein (Chun et al., 2011). Another means of detecting gold particle aggregation takes advantage of their ability to absorb light without photobleaching. This property was used in an assay for alpha-fetoprotein, based on photothermal beam deflec- tion (PBD). Antibodies were bound to both 20nm gold particles and to large 50µm glass particles. In the presence of a multivalent antigen, many gold particles aggregate around each glass particle. Irradiation of the suspension with an intense light beam produces highly localized heat- ing near each aggregate. The resulting change in the index of refraction of the solution near the aggregates causes a measurable angular deflection of the beam. This assay pro- vided an order of magnitude increase in sensitivity relative to agglutination of latex particles although rapid settling of the large glass particles detracts from the convenience of the method (Sakashita et al., 1995). MAGNETIC NANOPARTICLES Even without attachment of gold particles, detection of immunochemically induced aggregation of magnetic nanoparticles is a promising approach for detection of viral andcellularanalytes.Nanoparticleswithsuperparamagnetic iron oxide cores have single magnetic domains that become aligned when subjected to a magnetic field. When they are FIGURE 1 Hemagglutination. Erythrocytes in anti-HIV positive blood are agglutinated upon addition of HIV antigen conjugated to antibodies to the cell surface antigen, glycophorin. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  3. 3. 69CHAPTER 2.3 Homogeneous Immunoassays magnetized in this way, they form clusters that can influ- ence the nuclear spin relaxation times of water, the same parameters that are measured in conventional magnetic resonance imaging (MRI). In one study, using particles with a hydrophilic surface coating and conjugated to anti-HSV-1 antibodies, as few as 50 HSV-1 viral particles in 100 µL of serum (0.8aM) could be detected by monitor- ing the relaxation times with a bench top MRI instrument (Perez et al., 2003). Small molecule antigens have been similarly assayed by taking advantage of their ability to interfere with the association of antibody- and antigen-coated magnetic nanoparticles. Clinical applica- tions of this method appear likely as progress is made in reducing the size and cost of MRI instrumentation. The method has recently been reviewed by Lowery (2009). Ranzoni et al. (2012) described a particularly innovative approach to monitoring immunochemically induced mag- netic particle aggregation that avoids the use of expensive instrumentation. Application of a magnetic field to 300 or 500nm antibody-coated magnetic particles induces clus- tering and acceleration of the formation of antigen sand- wich particle pairs. Replacing the static field with a rotating magnetic field disperses the unattached particles and induces controlled rotation of the particle pairs. By simul- taneously illuminating the sample with a laser, the light scattering by the rotating particle pairs is modulated as a function of the changing particle pair orientation. The sig- nals can be analyzed as a function of the analyte concentra- tion. BSA in buffer could be detected at 0.4pM and in plasma at 5pM in less than 3min. Lysis Immunoassays Cell lysis initiated by binding of immunoglobulins to cellu- lar antigens is a fundamental part of the body’s defense mechanism. The process depends on the action of comple- ment, a complex mixture of proteins that undergoes a cas- cade of events triggered by binding to antibodies on the cell surface and culminating in lysis of the cell. This process is the basis for a much explored but little utilized homoge- neous immunoassay, hemolysis immunoassay. Antigens covalently conjugated to erythrocytes are allowed to bind to an antibody in the presence of a sample to be analyzed. The amount of antibody available for binding to the cells is affected by competitive binding to free antigen present in the sample. Following binding, serum that contains com- plement is added. After further incubation to complete lysis, the remaining intact cells are removed, and the concentra- tion of hemoglobin released into the solution is measured by its light absorption (see Fig. 2). Because many molecules of hemoglobin are released as a result of relatively few bind- ing events, the assay sensitivity approaches that of radioim- munoassay (Arquilla & Stavitsky, 1956). Despite this early success, hemolysis immunoassay is now of little more than historical interest. Erythrocytes are difficult to store; the protocol is complex; complement is unstable; and precautions must be taken to deactivate complement that may be present in the sample. Numerous attempts to overcome these problems and improve assay sensitivity have met with some degree of success. A simpler protocol was devised that is based on the measurement of the peroxidase activity of the released hemoglobin rather than its optical absorption. Because fluorescent or chemi- luminescent substrates can be used that do not enter the cells, removal of the cells is avoided. However, reagent sta- bility and interference from sample components remain problematic (Tatsu & Yoshikawa, 1990; Tatsu et al., 1992). A major advance in lysis immunoassays came when it was shown that liposomes, which are more stable than erythrocytes, can also be lysed by complement when anti- bodies are bound to a surface antigen (Kataoka et al., 1971; FIGURE 2 Complement-mediated lysis immunoassay. Erythrocyte lysis is promoted by binding of antibodies to erythrocytes labeled with an antigen. Lysis is inhibited by the presence of free antigen. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  4. 4. 70 The Immunoassay Handbook Kinsky, 1972). The substance that is released need no lon- ger be limited to hemoglobin. Nearly any compound that is in solution during formation of the liposomes can in principle be encapsulated. Proteins and other large mole- cules can be retained nearly indefinitely, and sufficiently hydrophilic smaller molecules leak out only slowly. Many encapsulated labels have been investigated. Encapsulated fluorescent compounds at sufficiently high concentration are usually quenched and become fluorescent only when released (Yasuda et al., 1988), and dyes showing concentra- tion dependence perform similarly (Frost et al., 1994). Solutions of entrapped stable nitroxide radicals show a change in their electron paramagnetic resonance (EPR) spectra upon dilution (Chan et al., 1978). Encapsulated chelators can bind to rare earth ions upon their release to form fluorescent complexes (Ius et al., 1996). Encapsulated enzyme substrates when released can be converted to a detectable product by an enzyme present in the bulk solu- tion (Thompson & Gaber, 1985), and coenzymes can sim- ilarly be released to react with the corresponding apoenzyme (Haga et al., 1990). Much attention has been given to the development of commercial liposome lysis immunoassays, based on the use of encapsulated enzymes (Canova-Davis et al., 1986; Yu et al., 1987). Not only are enzymes resistant to leakage, but they also offer the opportunity for enormous amplification because many enzyme molecules can be released from each liposome and each molecule can then generate many molecules of a detectable product. Unfortunately, practi- cal applications of this potentially powerful method have been difficult to develop. Methods for stabilizing standard- ized solutions of complement remain elusive, and the pres- ence of complement activity and inhibitors in serum samples has not been successfully overcome. Spin Immunoassays The use of molecular reporter groups other than particles in immunoassays was first described by Coons who used antibodies labeled with fluorescein to visualize binding to tissue slices (Coons et al., 1942). Although fluorescent labels eventually became very useful, stable nitroxide radicals were the first reporter groups commonly used in homoge- neous immunoassays. Nitroxides have one unpaired elec- tron, which is detectable by EPR spectroscopy. Electrons have a spin of 1/2 which permits two (2S+1) orientations in a magnetic field. Transitions between these states are stim- ulated at a precise microwave frequency depending upon the field strength. The closely associated nitrogen nucleus has S=1, which imparts three possible local fields, and thus, there are three separate resonance frequencies. This hyper- fine coupling is anisotropic with respect to the orientation of the nitroxide group in the magnetic field. The many pos- sible orientations in a solution of a nitroxide produce a broad spectrum when the molecules are relatively immo- bile. However, sharp three-line spectra are produced if the molecules are small enough to cause averaging of the hyperfine coupling due to rapid tumbling. Binding of a small nitroxide-labeled drug such as mor- phine to anti-morphine antibodies sharply reduces its rate of tumbling leading to line broadening (Fig. 3). In the presence of free morphine, binding of the nitroxide- labeled drug is inhibited, and the spectral lines become narrow. This was the basis of “spin immunoassay” which was the first widely used immunoassay, homogeneous and heterogeneous, to be employed commercially (Leute et al., 1972). Sold under the trade name FRAT, it was adopted by the US Army for screening its personnel for abused drugs during the Vietnam War. The method was completely manual but very simple and fast. A capillary was used to draw in a defined volume of urine. The contents of the capillary were then injected into a reagent comprising a mixture of the labeled drug and the antibody. The solution was drawn back into the capillary, which was then plugged at one end and dropped into an EPR cavity. The whole process could be done manually in about 30s with little training. Following the war, the method fell into disuse because of its poor sensitivity (~10−7 M) and the costly instrumentation. FIGURE 3 Spin immunoassay. The line-broadened EPR spectrum of a slowly tumbling spin-labeled drug bound to an antibody is replaced by the sharp line spectrum of the free label when drug from a sample competes for antibody-binding sites. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  5. 5. 71CHAPTER 2.3 Homogeneous Immunoassays Fluorescence Immunoassays FLUORESCENCE POLARIZATION An alternative method of monitoring changes in tumbling rates is fluorescence polarization immunoassay (FPIA). Light is differentially absorbed by molecules depending on how they are oriented relative to the direction and polariza- tion of the exciting light. The light subsequently emitted as fluorescence by each of the resulting electronically excited molecules will usually be polarized. However, freely tum- bling molecules rotate during the period that they remain excited thereby randomizing their orientation with a net reduction in fluorescence polarization. The more rapid the tumbling, the less polarization is observed (see Fig. 4). This phenomenon was first applied in a homogeneous immuno- assay by Dandliker (Dandliker & Feigen, 1961; Dandliker et al., 1973). Like spin immunoassay, binding of antibody to a labeled antigen, for example fluorescein bound to penicil- lin, affected the tumbling rate of the label. The presence of free antigen inhibited binding leading to an increase in freely tumbling labels and, thus, a decrease in the polariza- tion of the emitted light. Initially, the method was little more than a laboratory curiosity because of the primitive state of development of commercial spectrofluorometers and the requirement for two separate measurements differ- ing by a 90° rotation of a polarizing lens. FPIA, like spin immunoassay, has since been widely used though it has mostly been applied to small molecule ana- lytes. The sample and antibody are combined, and the antigen in the sample competes with fluorescer-labeled antigen for binding to the antibody. Increasing concentra- tions of the antigen produce decreased polarization. Abbott Laboratories uses FPIA primarily for therapeutic drug monitoring and drug abuse testing on their immunochem- istry systems. The success of the method after years of dis- use stemmed in larger measure from the development of improved solid-state methods for analyzing polarized light. High molecular weight analytes are difficult to analyze by FPIA for two reasons. First, binding of a large antigen to an antibody that may be of similar mass produces a smaller relative change in the tumbling rate and thus a smaller change in polarization. Secondly, most fluores- cent labels have excited state lifetimes in the range of 10−9–10−7 s, which is too short to permit rotational reori- entation of larger biopolymers. A rhenium-based fluoro- phore was identified that shows polarized fluorescence with a lifetime of 3ms. This extends the theoretically accessible molecular weight range for special applications (Guo et al., 1998). However, interference from quenching impurities becomes increasingly problematic with long- lived excited states because quenching affects the mea- sured polarization. Interference from adventitious fluorophores and non- specific binding of the label to proteins in biological sam- ples has historically restricted the detection limit of FPIA to concentrations of about 100pM. By using a highly hydrophilic long wavelength dye and time-delayed mea- surements that permit discrimination between the emis- sion from the background and the label, detection of an oligonucleotide down to 10pM in buffer has been claimed (Devlin et al., 1993). A more detailed account of fluores- cence polarization as applied to immunoassay has recently been published (Jameson and Ross, 2010). FLUORESCENCE RESONANCE ENERGY TRANSFER Fluorescence (or Förster) resonance energy transfer (FRET) refers to the transfer of energy from an electroni- cally excited donor molecule to a nearby acceptor mole- cule that has an energetically accessible excited state. The excited states of one or both of the donor and acceptor can decay with fluorescence emission. When both are fluores- cent, energy transfer is observed by a reduction in the emission intensity of the donor with a concomitant increase in the intensity of the longer wavelength emission from the acceptor. The energy transfer mechanism of most rel- evance in immunoassays depends on through-space cou- pling of the electronic transition dipoles (Förster, 1948). The rate of energy transfer is inversely dependent on the sixth power of the distance between the donor and Bound label Slow rotation Polarized emission Free label Rapid rotation Unpolarized emission Excitation Emission Emission Excitation FIGURE 4 Fluorescence polarization. Excitation of fluorescent labels leads to selective absorption of polarized light by appropriately oriented molecules. Polarized emission of these oriented molecules occurs when their rate of rotation is low relative to the rate of fluorescence emission. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  6. 6. 72 The Immunoassay Handbook acceptor and is directly related to the spectral overlap of the donor emission and acceptor absorption spectra. Dis- tances at which the donor fluorescence is reduced by 50%, the FRET distance Ro, can approach 10 nm depending on the dyes that are used. FRET immunoassays were originally developed to avoid the need for measuring fluorescence polarization by the primitive methods available at that time (Ullman et al., 1976; Ullman & Khanna, 1981). Only a simple flu- orometer is needed, and the method is applicable to both small and large molecules. Competitive immunoassays can be easily constructed by conjugating a fluorescent donor to the antigen and an acceptor molecule to the antibody. Usually, multiple acceptor molecules are attached to the antibody to assure that there will be at least one acceptor at an efficient energy transfer distance within the immune complex. In immunometric (sand- wich) assays, multiple donors and acceptors on each of the antibodies may be needed (see Fig. 5). Usually, the assays are carried out by following the rate of change in fluorescence during the initial phase of immunochemical binding. This reduces interference from fluorescence by the sample, which normally does not change over the course of the assay. In competitive immunoassays, the antigen causes a reduction in the rate of decrease in donor fluorescence by competing for antibody-binding sites. In sandwich immunoassays, increasing concentrations of antigen accelerate quenching of the donor emission due to accelerated binding of the two antibodies to the anti- gen. As in the precipitin and latex agglutination methods, a sufficient antibody concentration must be used to avoid the prozone phenomenon that occurs at antigen excess where only single-antibody molecules become available for binding to an antigen. The first commercial use of FRET immunoassays was in Syva’s Advance®immunochemistry system. In this system, the reduction in fluorescence of the donor was monitored rather than the theoretically more sensitive appearance of fluorescence of the acceptor. This was necessary because of the difficulty in finding a fluorescent acceptor that is not directly excited by the light used to excite the donor. Nev- ertheless, by using phycoerythrin as the donor, a highly fluorescent protein with an extinction coefficient near 2×106/M/cm, sufficient sensitivity could be obtained for a homogeneous serum digoxin assay that could quantitate 500pM of the drug in serum (25pM in the assay medium). Subsequent improvements in FRET immunoassays have circumvented this problem through the development of specialized acceptors and alternative energy donors. For example, gold nanoparticles (AuNPs) have been found to be highly efficient FRET acceptors that permit direct monitoring of donor fluorescence without interfering acceptor auto-fluorescence. The advantage of this type of quencher was illustrated by a homogeneous troponin sand- wich immunoassay using an anti-troponin antibody bound to a AuNP and a second anti-troponin antibody labeled with the highly fluorescent Cy3 label. Association of the two antibodies upon binding to troponin led to up to 95% fluorescence quenching, and troponin concentrations as low as 20pM could be detected (Mayilo et al., 2009). Simi- larly, graphene has been shown to be a highly efficient acceptor with fluorescent quantum dots as the donor. The Förster energy transfer distance exceeded 10 nm, and a sensitive FRET immunoassay capable of detecting alpha-fetoprotein could be constructed (Liu et al., 2010). The use of quantum dots (QDs) as an energy acceptor has recently received considerable attention. The com- monly used CdSe/ZnS core/shell QDs can be prepared in various sizes in the range of 2–10nm. They all have similar but nonidentical absorption spectra and fluoresce effi- ciently with narrow band emission at different wavelengths depending on their size. QDs are therefore ideally suited to serve as energy acceptors using a common fluorescent donor in multiplexed immunoassay in which each analyte can be independently analyzed by deconvoluting the com- bined emission spectra. In a model system, Geissler et al. (2010) used a combination of time, excitation wavelength, and emission wavelength resolution to carry out a five-plex assay using different terbium donors and five different quantum dots as fluorescent acceptors. The use of quan- tum dots in FRET assays has recently been reviewed (Algar and Krull, 2010). The development of time-resolved FRET immunoas- says, TR-FRET, provided a major improvement in fluo- rescence immunoassays. Rare earth chelates that serve as the donor are particularly useful for this purpose (see SIGNAL GENERATION AND DETECTION SYSTEMS). These labels have long fluorescent lifetimes of the order of milliseconds FIGURE 5 Fluorescence resonance energy transfer. FRET produces a reduction in donor fluorescence and an increase in acceptor fluorescence upon binding of the two labels within an immune complex. Competitive and sandwich immunoassay protocols are illustrated. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  7. 7. 73CHAPTER 2.3 Homogeneous Immunoassays whereas the fluorescent emission of most fluorescent sub- stances decays in less than a microsecond. Following a brief pulse of light, emission from sample impurities and accep- tor molecules that have been directly excited decays very rapidly. The delayed fluorescence that is subsequently mea- sured is associated only with emission from the donor and from those acceptor molecules that are excited by reso- nance energy transfer from the donor. The reduced back- ground emission provides greater sensitivity than when using steady-state fluorescence measurements, and quench- ing effects of the sample matrix are reduced by measuring the ratio of light emitted by the acceptor and donor. The method is used by Cisbio under the trade name TRACE™ (time-resolved amplified cryptate emission) and employs an europium chelate donor together with a deriva- tive of allophycocyanin, an energy-accepting fluorophore obtained from red algae (Mathis, 1993). Homogeneous binding assays for drug discovery are carried out with Cisbio’s Kryptor™ time-resolved fluorescence instrument. There is currently much interest in various additional com- binations of time resolution, quantum dots, metallic parti- cles, and graphene acceptors that may lead to additional improvements in sensitivity and multiplexing of FRET immunoassays. Up-converting phosphors represent an alternative type of fluorescent donor that can avoid unwanted direct excitation of the acceptor and fluorescent sample compo- nents. Under usual irradiation conditions, fluorescent molecules emit light at a longer wavelength than that used for excitation. This occurs because the initially excited state of the fluorophore loses vibrational energy as heat, prior to reemission of light. With sufficiently intense longer wavelength excitation, the same excited state can be reached either by simultaneous or stepwise absorption of two or more lower energy photons. The wavelength of the ensuing fluorescence emission from the excited state remains substantially unchanged but is now shorter than that of the exciting light. Because the excitation light can be in the near infrared, background emission from the acceptor and from fluorophores in the sample is minimal. Fluorophores such as 1,4-bis-(4- diaminostyryl) benzene derivatives that have a large two- photon absorption cross section at 800 nm have been used together with a nonfluorescent acceptor to detect as low as 50 pM anti-BSA antibodies despite strong back- ground fluorescence (Liu et al., 2008). A more robust approach to upconversion is based on the use of upconverting particles (UCPs). UCPs are ceramic materials in which rare earth atoms, which typically absorb 970–1000 nm, are embedded in a crystalline matrix. Sub- micrometer to less than 10nm sized particles can be pre- pared. Because upconversion into the visible spectral region occurs by sequential excitation of relatively long- lived excitation states, intense exciting light intensity is not needed (Austin and Lim, 2008). Quantitation of estradiol in 20% whole blood, at levels as low as 0.5nM, has been achieved using UCP conjugates of the antibody and Alexa Fluor 680-labeled 17β-estradiol (Kuningas et al., 2007). Near infrared excitation at 980nm resulted in upconver- sion to 660nm by the phosphor, which in turn transferred energy to the Alexa dye with emission at 740nm with no sample background fluorescence. FLUORESCENCE PROTECTION ASSAYS An alternative method for modulating fluorescence in an immunoassay is to arrange the system so that those fluo- rescent labels that are not bound in an immune complex are quenched. This is a direct way to obviate the need to separate free from bound label and can be used to convert any heterogeneous fluorescence immunoassay into a homogeneous format. A competitive assay, for example, can then be run by first incubating the analyte, antibody, and a fluorescent-labeled antigen. The antibody can be in solution or bound to a surface. A quenching reagent that reacts selectively only with unbound label is then added, and the residual fluorescence associated with the bound label is measured. The most general fluorescence protection assay strat- egy is to use an anti-fluorescer antibody as the quench- ing reagent. Some fluorophores such as fluorescein are quenched simply on binding to an anti-fluorescer anti- body. When binding provides insufficient quenching, a nonfluorescent energy acceptor can be attached to the antibody. Provided the fluorescent label is conjugated to a low-molecular-weight hapten by a short chain, simul- taneous binding of antibodies to the hapten and the label is sterically hindered. Fluorescent labels linked to hap- tens are therefore quenched by anti-fluorescer antibod- ies unless they are protected by being bound to an anti-hapten antibody (see Fig. 6). Fluorescent labels on protein antigens are less efficiently protected by anti- body binding. This can be overcome by providing a means to sterically hinder the binding of anti-fluorescer antibodies to the antigen immune complex. An effective way to do this is by increasing the size of the anti-fluo- rescer antibody by using the anti-fluorescer antibody as a preformed complex with anti-immunoglobulin anti- body or a conjugation with dextran (Zuk et al., 1979). Fluorescence protection can similarly be applied to assays that use fluorescer-labeled antibodies provided the binding components are rendered sufficiently bulky. When fluorescer-labeled antibodies to a protein antigen bind to antigen that is immobilized on suspended aga- rose particles, the fluorescent label becomes sequestered and unavailable for binding to subsequently added FIGURE 6 Fluorescence protection immunoassay. Binding of antibody to an antigen–fluorophore conjugate prevents anti-fluorescer antibody from binding and quenching the fluorescer. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  8. 8. 74 The Immunoassay Handbook anti-fluorescer antibodies. The more free antigen that is present, the less labeled antibody becomes bound to the agarose, and the greater the quenching. Discrimination between bound and free antibodies can be enhanced by addition of a suspension of carbon particles coated with anti-fluorescer antibody. The carbon particles not only provide increased quenching but may also inhibit residual slow binding of anti-fluorescer antibodies to fluorescent labels on the agarose (see Fig. 7; Ullman, 1981). Substan- tial improvement in fluorescence protection immunoas- says is likely by replacing quenching by carbon particle by the recently discovered highly efficient quenching of AuNPs and graphene. Particularly for competitive immunoassays of small ana- lytes, fluorescence protection offers an advantage over FRET because only one label is required. It is therefore possible to multiplex several assays by using different fluo- rescent labels. Additionally, up to 98% of the fluorescent signal can be modulated as compared with 30–80% modu- lation by FRET. The sensitivity of both methods is limited by the fluorescent background from the sample except where time-resolved fluorescence is employed. As in FRET, this problem can be minimized by measuring the quenching rate rather than the end point after addition of the final reagent. This effectively subtracts out the static sample background, although with lipemic serum samples, changes in light scattering due to dissolution of lipids fol- lowing dilution can be a problem. Surprisingly, fluorescence protection immunoassays have received little attention despite their advantages over FRET. The main application has been in a non-immunochemical assay for the binding of ligands to cellular receptors. A fluorescein-labeled ligand and an unlabeled test compound are allowed to compete for binding to receptors on a cell surface. Addition of anti-fluorescein antibodies quenches only the unbound conjugate so that higher fluorescence intensity relates to less effective competition by the test compound (Sklar et al., 1982). FLUORESCENCE FLUCTUATION ANALYSIS A number of different approaches have been developed for homogenous assays that are based on detecting immuno- chemical binding by monitoring temporal variations in flu- orescence intensity. Binding of an antigen–fluorescer conjugate to latex particles coated with an antibody to the antigen will of course cause the particles to become fluorescent. Provided the number and affinity of the anti- bodies on a particle is sufficiently high, the concentration of the conjugate within the volume defined by the particle will be higher than in the bulk solution. Binding can be detected by interrogating many small volumes of a suspension of these particles using a laser. Using a fluorescence fluctua- tion-correlationmethod,increasedfluctuationsareobserved upon binding due to the presence of fluorescent particles in some volumes and not in others (Elings et al., 1983). The method has been applied to a homogeneous blood typing system. Cells in whole blood are stained with a membrane-soluble fluorescent dye, and aggregation of the fluorescent erythrocytes by blood type-specific anti- bodies is detected by fluorescence fluctuation analysis. For this purpose, a probe is immersed in the solution that has two optical fibers, one that delivers a narrow beam of light to the solution and the other that collects fluorescent emission from a short segment of the light path and delivers it to a photomultiplier tube (PMT). By using an interrogation volume of about 1 nL, fluctua- tions produced by single cells can be distinguished from dimers without waiting for larger aggregates to form (see Fig. 8; Ghazarossian et al., 1988). A related approach is used for homogeneous detection of binding of ligands to cells or latex particles for use in high-throughput screening. In Applied Biosystems’ FMAT® system cells or particles having a receptor or antibody on their surface are mixed with a fluorescer-labeled ligand and the test compounds in microtiter wells. The bottom FIGURE 7 Fluorescence protection immunoassay (as in Fig. 6) employing carbon particles to enhance quenching. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  9. 9. 75CHAPTER 2.3 Homogeneous Immunoassays surface of well is scanned with a laser to permit activation of only a small volume at a time. The magnitude of the fluorescence fluctuations permits estimation of the degree of binding of the label, which is inversely related to the ability of the test compound to bind. The method allows multiplexing by labeling with two dyes that emit at differ- ent wavelengths. The ability to measure fluorescence from sufficiently small volumes has enabled the use of fluorescence correlation spectroscopy (FCS) in place of simple detection of aggregates of fluorescent labels. At suffi- ciently low fluorophore concentrations and a sufficiently small irradiation volume (typically about 1 fL), the fluo- rescence of single molecules can be observed. Molecules to be detected are distinguished from other fluorescent species by their characteristic rates of diffusion through the detection volume, which are affected by their masses and shapes. The method is used for high-throughput screening for drugs. Sensitive homogeneous immunoas- says have been demonstrated by selecting conditions that assure that the free and bound fluorescent binding part- ners have very different diffusion rates. For example, Tang et al. (2010) carried out a sandwich immunoassay by using one antibody labeled with a fluorescence mole- cule and a second antibody bound to much larger 14 nm silver nanoparticles. Serum alpha-fetoprotein concentra- tions as low as 1.4 pM could be detected. Even greater sensitivity could be obtained using different antibodies to human IgG bound separately to 40 nm gold particles (Chen et al., 2009). The particles are fluorescent and photostable (He et al., 2008), and their diffusion charac- teristics are greatly altered by immunochemically induced aggregation. As low as 5 pg/mL (30 fM) concen- trations of human IgG could be detected in 20% dog serum. The need for discrimination between different rates of diffusion has been circumvented by causing two different fluorescent molecules to become associated immunochemically (Winkler et al., 1999). Coincident emission from the paired fluorescers can then be easily differentiated from that of unpaired molecules. The method, called confocal fluorescence coincidence analysis, since renamed fluorescence cross-correlation spectros- copy (FCCS), has been further enhanced by including FRET interactions in the analysis of the spectral fluctua- tions. Various additional combinations of analysis are described by Eggeling et al., 2005. Chemi- and Bioluminescent Immunoassays Efforts to replace FRET with a potentially more sensi- tive method using chemiluminescence resonance energy transfer (CRET) have not been encouraging. The process differs from FRET only by using chemical excitation instead of light to activate a FRET donor. In early attempts to achieve higher sensitivity, several hap- tens and protein antigens labeled with isoluminol were shown to sensitize fluorescein emission when bound to fluorescein-labeled antibodies (Patel & Campbell, 1983). Although the assays were equivalent to radioim- munoassay in sensitivity, the high susceptibility of the chemiluminescent reaction to differences in sample composition led to unacceptable performance, at least for clinical applications. More recently, highly efficient macromolecular energy acceptors such as graphene (Lee et al., 2012) have been used to quench the chemilumi- nescence of peroxidase-catalyzed oxidation of luminol. A homogeneous C-reactive protein (CRP) sandwich immunoassay using two different antibodies conjugated respectively to graphene and peroxidase permitted detections of serum CRP at clinically relevant (nM) lev- els. Quenching using AuNPs in place of graphene has provided sensitivities in the low pM range in a similar sandwich immunoassay for alpha-fetoprotein (Huang & Ren, 2011). The use of natural bioluminescent enzymes offers a potential advantage in immunoassays, because the chro- mophores are usually sequestered and minimally affected by interfering substances that may be in the sample. Bio- luminescence resonance energy transfer (BRET) has been extensively studied as a means of monitoring protein- binding interactions, but BRET immunoassays have received surprisingly little attention. In a notable example, bioluminescent Renilla luciferase was used as a donor with enhanced yellow fluorescence protein as the acceptor. The VH and VL fragments of an anti-lysozyme antibody were expressed respectively as chimeric proteins with each of these labels. The presence of lysozyme caused binding and reassembly of the two antibody fragments, resulting in an increase in emission from the acceptor protein in the pres- ence of the substrate coelenterazine (Arai et al., 2001). The same group demonstrated a competitive BRET immuno- assay for a decapeptide using a fusion protein of firefly luciferase and the decapeptide antigen together with an antibody labeled with a fluorescent (Cy3.5) acceptor (Yamakawa et al., 2002). Neither method afforded excep- tional sensitivity. FIGURE 8 Fluorescence fluctuation assay. Bent optical fibers immersed in a suspension of particles and capable of interrogating nanoliter volumes are used in fluorescence fluctuation analysis to distinguish between free fluorescent label and fluorescer-labeled particles. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  10. 10. 76 The Immunoassay Handbook Surface-Enhanced Raman Scattering Raman spectroscopy measures the very small fraction of light that is scattered with shifted frequency during irra- diation of a chemical substance. The frequency shift is caused by coupling with vibronic oscillations and is inde- pendent of the excitation wavelength. Molecules in close proximity to a metal surface display surface-enhanced Raman scattering (SERS) that can result in up to 1011-fold Raman signal enhancement. AuNPs and silver nanoparti- cles can produce large enhancements, depending on their size and shape and the proximity of other enhancers. Very sensitive heterogeneous immunoassays have been con- structed by immunochemically-induced particle binding to roughened metal surfaces. There has been considerable interest in developing SERS homogeneous immunoassays, but the potentially high sensitivity of the method has not yet been realized. Chen et al. (2008) reported a homoge- neous immunoassay for human IgG, using SERS. The AuNPs were labeled with anti-IgG and an additional enhancer. Aggregation of the particles upon binding to IgG enhanced Raman scattering, but the sensitivity was limited to about 0.7nM IgG. Li et al. (2008) reduced the detection limit by an order of magnitude in a model assay system for biotinylated IgG, using gold particles labeled with protein A and fluorescein (enhancer)-labeled avidin. By contrast, a heterogeneous sandwich immunoassay for IgG in which antibody labeled with uniquely fabricated cubic gold particles became bound to a gold surface pro- vided a sensitivity of 1pM (Narayanan et al., 2008). Enzyme Immunoassays ENZYME-MULTIPLIED IMMUNOASSAY TECHNIQUE Homogeneous enzyme immunoassays were first described in 1972, and assays for drugs of abuse in urine were introduced commercially the following year under the trade name EMIT® (Rubenstein et al., 1972). The assays used a simple mix-and-read protocol and had the advan- tage of using a general-purpose standard laboratory spectrophotometer. The method was based on the ability of antidrug antibodies to inhibit the enzyme activity of a lysozyme–drug conjugate. The enzyme catalyzes bacterial cell wall hydrolysis, which is monitored by the clearing of a turbid bacterial suspension. In the presence of free drug, less antibody is available to bind to the conjugate, which results in an increased rate of clearing. Lysozyme was cho- sen because reaction with its large substrate seemed likely to be susceptible to steric inhibition by the antibody (see Fig. 9, left arrow). However, the light scattering assay was not very sensitive, and the assays could not be used with serum because of serum-induced agglutination of the bacteria. An ideal enzyme should not be present in the test sam- ple; it should be detectable at low concentrations by a gen- eral-purpose spectrometer, and the sample matrix should not affect its activity. Additionally, it is necessary for the enzyme to retain activity and stability following conjuga- tion, and, most importantly, the activity of the resulting conjugate must be modulated by antibody. Glucose-6- phosphate dehydrogenase (G6PDH) from the bacterium, Leuconostoc mesenteroides, was found to meet most of these requirements. This 109kDa enzyme has two identical sub- units and converts NAD to NADH, which is measured by its absorbance at 340nm. Endogenous G6PDH in serum requires NADP and thus does not interfere with the mea- surement. Hapten conjugates of G6PDH are inhibited by anti-hapten antibodies when the hapten is bonded to lysine amino groups. Although the antibodies are unlikely to effect steric exclusion of the small substrates, molecular contact between the bound antibody and the enzyme pro- duces an inhibitory effect by effecting a change in the enzyme conformation (see Fig. 9, right arrow). Up to 80% inhibition by antibodies is occasionally observed, but the percent inhibitability depends strongly on the hapten structure and identity of the antibody. Increasing the num- ber of haptens per enzyme molecule increases inhibitabil- ity but with concomitant loss in enzyme activity. Targeting of the hapten just to those lysines that effect inhibition is difficult because different constellations of sites are active with different haptens. As a result, extensive screening to identify suitable means of linking the conjugates and iden- tifying effective antibodies is often needed. Despite these limitations, serum EMIT G6PDH assays for a large number of therapeutic drugs and urine assays FIGURE 9 EMIT. Antibodies can inhibit the activity of antigen-labeled enzymes either by sterically hindering approach of the enzyme substrate or by inducing a conformational change in the enzyme. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  11. 11. 77CHAPTER 2.3 Homogeneous Immunoassays for abused drugs have been successfully developed. Several alternative enzymes utilizing macromolecular substrates have been described including amylase, phospholipase C, mitochondrial malate dehydrogenase, and dextranase, but the enzymes either are too sensitive to serum components or are difficult to measure at low concentrations. Current EMIT assays use genetically modified G6PDH to achieve greater modulation with fewer haptens. The conjugates are more stable and provide greater assay response. Haptens are attached to a single cysteine per subunit (there are no natural cysteines) that is incorpo- rated at a site that maximizes antibody-induced modula- tion. Suitable sites were identifying by mapping an enzyme epitope that is recognized by an inhibitory anti-G6PDH monoclonal antibody which induces a conformational change of the enzyme. The amino acids within the epitope were then systematically replaced one at a time with cyste- ine (Ullman, 1999). The activity of conjugates prepared by attachment of drugs to the cysteine sulfhydryl group is not impaired. EMIT assays using recombinant G6PDH are currently sold by Siemens Healthcare and used for assays such as digoxin that must provide accurate quantitation down to 500pM. The primary limitation of immunoassays based on simple modulation of the activity of enzyme conjugates is the difficulty in designing assays for larger molecules. The problem stems from the need to cause an antibody to bind close enough to the enzyme to have a strong effect on the enzyme activity. One method that has met with some success depends on the ability of antibodies to form aggregates with protein antigens. It has been found that horseradish peroxidase (HRP) is inhibited by high concentrations of hydrogen peroxide but is protected when it is polymerized as part of a large aggregate (Hoshino et al., 1987). Homogeneous enzyme immuno- assays for protein C, alpha-fetoprotein, and polymor- phonuclear leukocyte elastase have been reported, using this principle. In the assay, an antibody–HRP conjugate binds to the antigen to form an aggregate (precipitin) that retains its enzyme activity when high peroxide con- centrations are used. Sufficient antibody must be used to avoid the prozone phenomenon, and the sensitivity of the assays is limited by the susceptibility of the system to serum interference. Another approach to protein assays is based on the previously discussed potential for large enzyme substrates to be sterically excluded from antibody:enzyme conjugate complexes. This was nicely demonstrated with β-galactosidase–antigen conjugates, which normally are not modulated by antibodies, even when the antigen is small. However, antibody-induced modulation of the enzyme conjugates of both small and large antigens was observed when the substrate, o-nitrophenyl-β-galactoside, was tethered to a large dextran polymer (Gibbons et al., 1980). EMIT assays for human IgG and CRP were dem- onstrated using this substrate. By using a dextran-linked fluorogenic substrate, a competitive ferritin assay was set up with a detection limit of about 4pM in serum (Armenta et al., 1985). Unfortunately, occasional low levels of anti- bodies to β-galactosidase in serum samples prevent appli- cation of this method for very high sensitivity serum assays. CHARGE-INDUCED ENZYME ACTIVATION An alternative approach to immunochemical modulation of enzyme activity is to engineer a change in the microen- vironment of an enzyme when it becomes bound in an immune complex. An assay for CRP in which the immune complex is highly charged is illustrative. Binding of CRP to a β-galactosidase conjugate of anti-CRP antibody does not significantly affect enzyme activity. Subsequent addi- tion of excess unconjugated anti-CRP antibody increases congestion within the immune complex and reduces the enzyme activity when a macromolecular substrate is used. However, if the unconjugated antibody is succinylated and the macromolecular substrate carries a positive charge, the enzyme activity is increased due to coulombic attraction of the substrate to the immune complex (see Fig. 10). Enzyme activation immunoassays for both protein antigens and antibodies have been demonstrated based on this principle (Gibbons et al., 1981). The prozone phenomenon is avoided simply by adding a large excess of the succinylated antibody. Despite the apparent attractiveness of this approach, the method has several pitfalls. Antigen concentrations as low as 60fM in the assay mixtures can be detected, but the use of greater than 0.1% serum affects quantitation. Hence, the practical detection limit is closer to 60pM. Addition- ally, while the assays do not show a prozone effect, data handling is difficult, because the response curves of high sensitivity assays do not increase monotonically. FIGURE 10 Enzyme activation immunoassay. Binding of enzyme-labeled antibody and succinylated antibody to a multi-epitopic antigen produces increased catalytic activity when using a polycationic substrate because of increased charge attraction (lower KM) between the enzyme and substrate. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  12. 12. 78 The Immunoassay Handbook ENZYME CHANNELING Enzyme channeling provides a method of detecting the proximity of two enzymes in an immune complex. The first enzyme catalyzes the formation of an intermediate substrate that is converted by the second enzyme into a detectable product. When both enzymes are independently dispersed in the same solution, the rate of product formation is slow at first but accelerates as the concentration of the intermediate substrate builds up. This kinetic behavior changes when both enzymes are closely associated at a surface (Mosbach & Mattiasson, 1970). The local concentration of the interme- diate in the vicinity of molecules of the first enzyme is deter- mined by the rate of formation of the intermediate and its rate of diffusion away from the enzyme. A local steady-state concentration is rapidly reached that is higher than the con- centration in the bulk solution. Localization of several mol- ecules of the first enzyme at a surface increases the rate of formation of the intermediate and thus increases its local concentration. When the second enzyme becomes bound to this surface, it experiences a relatively constant elevated concentration of the intermediate leading to an increased linear rate of formation of the final product. Homogeneous enzyme channeling immunoassays take advantage of this phenomenon (Litman et al., 1980). Vari- ous surfaces have been employed including agarose parti- cles, latex beads, and the polystyrene surface of a microtiter well. One enzyme serves to label an antibody or antigen, and an excess of the other enzyme is bound to the surface. Usually, the first enzyme is attached to the surface because more linear kinetics are obtained although channeling also occurs when the roles of the enzymes are reversed. A vari- ety of enzyme pairs have been used including alkaline phosphatase/β-galactosidase, hexokinase/G6PDH, and glucose oxidase (GO)/HRP. When a natural substrate is not available, as in the case of alkaline phosphatase/β- galactosidase, synthetic constructs can be prepared that permit the sequential reaction to occur. A competitive assay for HIgG can be carried out with agarose particles labeled with HIgG and GO (see Fig. 11). Upon reaction with glucose, these particles become sur- rounded by a halo of hydrogen peroxide. As the peroxide diffuses into the bulk solution, it is diluted, and the con- centration is further reduced by catalase that is present in the reaction mixture. When anti-HIgG antibodies labeled with HRP bind to the particles in the presence of ABTS, a chromogenic HRP substrate, there is a nearly constant rate of color formation that depends inversely on the con- centration of the HIgG. The most sensitive applications of enzyme channeling avoid the use of a preformed surface in favor of the in situ formation of a colloidal precipitate. An assay for polyribose phosphate (PRP), a component of the cell wall of Haemophilis influenzae, was demonstrated using a reagent containing anti-PRP antibody labeled with GO (Ab-GO), anti-PRP antibody labeled with HRP (Ab-HRP), and free GO (Ullman et al., 1984). Combination of this reagent with a clinical sample to which anti-GO antibody had been added produced an Ab-GO:PRP:Ab-HRP sandwich com- plex that was incorporated into a colloidal GO:anti-GO immune complex (precipitin) (see Fig. 12). Addition of glu- cose, ABTS, and catalase initiated the enzyme-channeling reaction. The assay response was nearly linear with a detec- tion limit of about 10fM PRP in the assay mixture, suffi- cientforacerebralspinalfluidassayforbacterialmeningitis. FIGURE 11 Enzyme-channeling immunoassay. Hydrogen peroxide is more concentrated at the surface of agarose particles labeled with GO than in bulk solution. (a) HRP-labeled antibody bound to an antigen at the agarose surface catalyzes the oxidation of a leuco dye. (b) With no antigen present, color formation requires diffusion of hydrogen peroxide into bulk solution. This reaction is suppressed by the presence of catalase. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  13. 13. 79CHAPTER 2.3 Homogeneous Immunoassays Remarkably, despite high sensitivity and simple assay pro- tocols, homogeneous enzyme-channeling immunoassays have received scant attention following these initial studies. Enzyme Effector Immunoassays Instead of using an intact enzyme, homogeneous immuno- assays can be constructed with catalytically inert labels that can combine with an enzyme or enzyme fragment and effect a change in enzyme activity. The labels that have been studied include enzyme substrates, coenzymes and inhibitors, and enzyme fragments. SUBSTRATE-LINKED FLUORESCENCE IMMUNOASSAY Substrate labels have been used for monitoring therapeutic drug levels where there is a relatively high concentration of the drug in serum (Li et al., 1981). Conjugation of β-galactosidylumbelliferone to theophylline does not affect its ability to act as a fluorogenic substrate for β-galactosidase. When anti-theophylline antibodies are present, the reaction is inhibited because the substrate, when bound to the antibodies, is less sterically accessible to the enzyme. If theophylline is present, it competes for antibody binding sites, and there is an increase in the rate of appearance of the fluorescent umbelliferone product. A major problem with this method is the necessarily low concentration of the substrate that must be used for the analyte to enter into effective competitive binding. At a low concentration, the enzyme turnover is slow. Signifi- cant increases in rate can be achieved at very high enzyme concentrations, but then the cost becomes prohibitive. Moreover, the method does not have the advantage of signal amplification that is provided by other enzyme immunoassay methods. That is, the number of molecules that provide a signal can never be greater than the number of molecules of the analyte. One approach to overcome these limitations is to use a substrate that gives a very easily detectable signal. ATP for example can be used as the substrate. Reaction of the ATP–antigen conjugate with firefly luciferase produces a chemiluminescent signal that is suppressed when anti- body is present. The method has been applied to com- petitive assays for HIgG and 2,4-dinitrophenylalanine (Carrico et al., 1976). While this overcomes some of the sensitivity limitations, sample-to-sample variations due to the susceptibility of chemiluminescent reactions to matrix effects more than offset the theoretical gain in sensitivity. ENZYME COFACTOR IMMUNOASSAY Cofactors or other prosthetic groups are more attractive labels for stimulation of enzyme activity. Unlike substrate labels, these groups provide the opportunity for signal amplification. The most successful example is the use of FAD–analyte conjugates that can complex with inactive apo-glucose oxidase to form active GO. When these con- jugates are bound to an anti-analyte antibody, complex- ation is sterically inhibited, and the enzyme is not activated. Competition by analyte from a sample increases the avail- able conjugate concentration and thus increases the enzyme activity (see Fig. 13). The method was commer- cialized by Miles Laboratories for protein analytes, such as thyroid-binding globulin. The reagents have also been dried onto strips for single step measurements of drugs in serum. Called ARIS™ (apoenzyme reactivation immu- noassay system), strips are impregnated with the dry FAD conjugate, antibody, apoenzyme, glucose, and reagents for detecting hydrogen peroxide that is a product of the FIGURE 12 Immune complex-enhanced enzyme-channeling immunoassay employing GO and antibodies to GO to create particles in situ. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  14. 14. 80 The Immunoassay Handbook enzyme reaction. When contacted with a sample, assembly of the apoenzyme and FAD conjugate is enhanced when antibody becomes bound to the analyte instead of the con- jugate. The intensity of color that is formed is used to determine the concentration of drugs such as phenytoin and theophylline. This is an ideal format for this type of assay because it permits the otherwise unstable apoenzyme to be stored in a dry form. ENZYME INHIBITOR IMMUNOASSAY Enzyme inhibitors can be used as labels in much the same manner as coenzymes (see Fig. 13). S-substituted ethoxy- methylphosphonothioates are irreversible inhibitors of acetylcholinesterase. Conjugation of haptens such as thy- roxine and theophylline to the sulfur does not affect the inhibition. However, the rate of enzyme inhibition is reduced upon binding of the conjugate to an anti-hapten antibody. Free hapten competes for antibody-binding sites and increases the concentration of available inhibitor with a corresponding acceleration of inhibition and decrease in enzyme activity. The reaction can be followed by the rate of color formation produced by reaction of 5,5′-dithiobis- (2-nitrobenzoic acid) with thiocholine that is formed upon enzymatic hydrolysis of acetylthiocholine (Blecka et al., 1983). Dihydrofolate reductase has been used for a theophylline assay in a similar manner by using a conjugate of theophylline and methotrexate, a reversible inhibitor of the enzyme (Place et al., 1983). The cholinesterase inhibition assay was sold briefly by Abbott using their bichromatic analyzers. The scope and limitations of the method have not been discussed. It is clear that adequate sensitivity requires very strong, prefer- ably irreversible, binding of the inhibitor to the enzyme and a means for sensitive detection of enzyme activity. Further investigation of this method has probably been impeded by the difficulty of identifying a better inhibitor/ enzyme pair. ENZYME COMPLEMENTATION IMMUNOASSAY Some enzymes can be disassembled into enzymatically inert fragments that become active only when they associ- ate into the original holoenzyme. Complementation of the enzymatically inactive S-protein from ribonuclease A with its 20 amino acid N-terminal S-peptide leads to recovery of enzyme activity. Conjugation of thyroxine to the S-pep- tide can be achieved without blocking complementation with the larger S-protein fragment. However, the activity of the reassembled enzyme is reduced. Binding of anti- thyroxine antibodies to the reassembled enzyme produces a recovery in enzyme activity, which is the basis for a homogeneous immunoassay (Gonnelli et al., 1981). Con- versely, when anti-thyroxine antibodies are added to the free S-peptide conjugate, complementation is blocked, and enzyme activity is inhibited. This process can similarly be used in a homogeneous assay for thyroxine (Farina & Gohlke, 1983). The former is simply an EMIT-type assay in which the antibodies provide enhancement of activity instead of inhibition. The latter is based on a related but different phenomenon that is conceptually similar to an enzyme cofactor immunoassay. In both cases, competition by free thyroxine for the antibody leads to a change in enzyme activity. Although ribonuclease has not proven to be attractive, possibly because there are no highly sensitive methods for the detection of ribonuclease activity, complementation immunoassaysusinganalternativeenzyme,β-galactosidase, have been commercialized under the trade name CEDIA® (cloned enzyme donor immunoassay). The native enzyme has four identical subunits, but there is not yet a consensus as to how many must be assembled to impart activity. However, it has long been recognized that two peptides that together comprise a subunit can be combined with recovery of nearly full activity, a process known as α-complementation (Ullmann, et al., 1967). These frag- ments include a smaller “donor” peptide and a larger “acceptor” protein. There are several donor–acceptor pairs that have this property. The members of each pair together contain the entire amino acid sequence of the native enzyme. In CEDIA, donor peptides are usually derived from the N-terminus, and the acceptor has an amino acid deletion including some portion of the donor. The donor is modified to permit conjugation to an antigen at a specific site and genetically engineered to provide good recovery of enzyme activity (Engel & Khanna, 1992; Henderson et al., 1986). In CEDIA assays, the analyte and the donor conjugate first compete for antibody-binding sites. The substrate and excess acceptor are then added. The higher the analyte concentration, the less antibody is available to bind the conjugate, and the faster the rate of assembly and final rate of catalyzed formation of product (see Fig. 14). CEDIA has been applied to a large variety of small molecules as well as to a few proteins. Like EMIT assays, a considerable amount of selection is required to identify antibodies that not only have the desired specificity and affinity but are also able to inhibit complementation. Donor conjugates of macromolecular antigens are difficult to design because bulky substituents can interfere with complementation, but the ability to genetically engineer the donor provides more flexibility than in EMIT. These assays offer some- what higher sensitivity than EMIT because of the high turnover and good detectability of β-galactosidase. Their primary disadvantage arises from the need to reconstitute the relatively unstable acceptor enzyme fragments, which must be stored dry, and the additional time required for the FIGURE 13 Enzyme cofactor immunoassay. Free FAD-labeled antigen assembles with inactive apo-GO to form the active holoenzyme. The reaction is inhibited when FAD-labeled antigen is bound to an antibody. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  15. 15. 81CHAPTER 2.3 Homogeneous Immunoassays complementation reaction. Like EMIT, convenient ways to construct practical noncompetitive sandwich immuno- assays have not been reported. A related process based on β–galactosidase complemen- tation is currently sold by Discoverx to quantify intracel- lular protein–protein binding (Wehrman et al., 2005). The process differs from CEDIA in that the protein bind- ing brings the two enzyme fragments together and initi- ates complementation with the appearance of enzyme activity. One of the fragments is expressed as a fusion pro- tein that becomes localized at a particular subcellular region, such as the cell membrane or nucleus, and the other fragment is fused to the protein undergoing trans- location and binding. Success of the method depends critically on selecting enzyme fragments that do not affect the translocation and binding processes. The method is commonly used for high-throughput screening by mea- suring drug-induced binding of β-arrestin to G-protein coupled receptors (GPCRs). Application of the technol- ogy to in vitro protein immunoassays has not yet been explored. Proximity-Induced Hybridization A hallmark of many cellular processes is the enablement of chemical reactions by localization of two or more reactants at binding sites within cellular compartments. Localization leads to an increase in the effective concen- tration of the reaction components and thereby pro- duces increased rates of reaction and enhances binding events. Proximity-induced hybridization, like channel- ing and charge-induced enzyme immunoassays discussed above, is based on this principle. Immune complexes typically occupy spherical volumes on the order of 0.01 zeptoliter. A single freely diffusing molecule in this vol- ume has an effective concentration of about 200 mM. Rates and equilibria within the complex between oligo- nucleotides attached to different members of the com- plex can therefore be enhanced by many orders of magnitude. Two applications of this principal have received attention. PROXIMITY LIGATION IMMUNOASSAY Proximity ligation assays (PLAs) employ two or three “proximity probes,” each capable of being incorporated into a single complex containing the analyte. Each of the probes comprises a different oligonucleotide. One of the simplest PLA configurations is an immunoassay that uses two such labeled polyclonal or monoclonal antibody prep- arations (Gullberg et al., 2004). Initially, a mixture of the labeled antibodies is combined with a sample whereupon immune complexes are formed that include both of the oligonucleotides. In the second step, a combination of reagents is added that includes a ligase and a ligation tem- plate oligonucleotide. The template binds to both nearby oligonucleotides in the immune complexes and thereby enables them to be ligated to each other. The lower con- centration of probe sequences external to the immune complexes greatly reduces the rate of nonspecific ligation. All elements required for PCR are also included in the combined reagent. The third and final step of the assay only requires PCR amplification of the ligated probes by thermal cycling. No additional handling of the reaction mixture is needed. Using this method, detection of low fM concentrations of IL2, IL4, and VEGF in as little as 1µL of sample were demonstrated (Gullberg et al., 2004). A modification of the method, called 3PLA, utilizes three proximity probes (Schallmeiner et al., 2007). The probes can be any binding substance such as antibodies, aptamers, oligonucleotides, etc. The binding substances can be directed against different sites on the same mole- cule, or molecular complexes can be detected by using binding agents directed against different molecules in the complex. In 3PLA, two of the proximity probes are labeled as above with different oligonucleotides, and the third proximity probe is labeled with the above template oligo- nucleotide. Because all three oligonucleotides must be in close proximity for ligation to occur, the method is expected to have increased specificity and should tolerate higher concentrations of unbound antibodies without undue non- specific ligation. Additionally, a means of increasing sensi- tivity is provided by protecting the oligonucleotides from nonspecific ligation by including complementary oligonu- cleotides that can only be displaced when in close proxim- ity to the template (see Fig. 15). The method has been shown to detect as little as 100 molecules of VEGF, PSA, and troponin in 1µL (~200aM). Although more time-con- suming than other very high-sensitivity homogeneous immunoassays such as luminescent oxygen-channeling immunoassay (LOCI) (see below), PLA is likely become a powerful tool for studying molecular interactions at or near the single-cell level. PROXIMITY HYBRIDIZATION IMMUNOASSAY A related approach to proximity-induced hybridization in homogeneous assays uses FRET detection (Heyduk et al., 2008). Similar to PLA, two antibody preparations that are labeled with different oligonucleotides are used in a sandwich immunoassay. In this approach, the oligonucleotides are labeled respectively with an FRET acceptor such as Cy5 and an FRET donor such as fluorescein or a europium chelate. FIGURE 14 Enzyme complementation immunoassay. Free peptide- labeled antigen (Ag-donor) binds to a larger inactive acceptor fragment of β-galactosidase to form the active holoenzyme. The reaction is inhibited by antibody binding to the Ag-donor. It has not been established whether the product is a fully assembled octamer, as shown. (The color version of this figure may be viewed at www.immunoassay- handbook.com).
  16. 16. 82 The Immunoassay Handbook The oligonucleotides have short (6–7bp) complementary portions designed to minimize hybridization of probes bound to free antibodies without interfering with hybridization within the immune complex. The approach is an attractive alternative to FRET immunoassays that use antibodies labeled with multiple dyes, some of which will be too remote to permit efficient energy transfer. Intra-complex oligonu- cleotide binding also is expected to have the effect of increas- ing sensitivity by stabilizing the termolecular sandwich complexes. Model assays for troponin and CRP were stud- ied. A detection limit of about 30pM troponin was obtained. Isotopic Labels SCINTILLATION PROXIMITY ASSAY One advantage of radioactive labels is that labeled ligands can be constructed that are chemically identical to the free ligand. This is of particular value in assays where chemical labeling of a ligand may interfere with binding to a natural receptor. However, this advantage must be weighed against the biohazard, stability, and disposal problems associated with radioactive isotopes, and the inconvenience of the separation and washing steps of standard radiobind- ing assays. Scintillation proximity assays (SPAs) are radioisotopic assays that avoid separation and washing. SPA requires a radiolabel that emits alpha or weak beta particles that only traverse short distances in a condensed medium. A receptor or antibody is bound to the surface of a polymer in which a fluorophore is dissolved such as a latex bead. If the radiola- bel becomes bound to the surface, radiation from the label passes through the polymer and produces pulses of light that are detected by scintillation counting. Most of the unbound labels will be too distant for radiation to reach the active surface (see Fig. 16). Instead of using fluorescent latex beads, fluorescent wells can also be used where the antibody or receptor is bound to the well surface. In either format, it is only necessary to incubate the labeled antigen and sample together with the antibody-coated surface and then measure the light emission (Hart & Greenwald, 1979). Reagents for SPA are sold for high-throughput drug screening by Perkin Elmer. In this application, natural receptors are used as well as antibodies, and both competi- tive and sandwich immunoassays can be carried out. An antibody, antigen, or receptor can carry the radiolabel. An important consideration in SPA is the choice of the label. The most light per decomposition event is produced by alpha particles, which have very high energy and short ionization tracks that have a high density of electrons. Additionally, the short path lengths provide optimal dis- crimination between bound and unbound label. However, the high energies associated with alpha emitters render them particularly hazardous if ingested and unacceptable for routine use as labels. Weak beta particles also have short path lengths but produce a weaker emission per decay FIGURE 16 Scintillation proximity assay. Beta particles emitted by tritium-labeled antigen are intercepted by a scintillator-loaded particle more efficiently when labeled antigen is bound to the particle. Free antigen inhibits binding. (The color version of this figure may be viewed at www.immunoassayhandbook.com). i ii iii Proximity probe 1 Proximity probe 2 Proximity probe 3 FIGURE 15 Antibody-based 3PLA proximity ligation assay. Three oligonucleotide-labeled antibodies bind to a protein analyte, thereby enabling ligation of two of the oligonucleotides. The ligated strand is amplified by PCR. Random ligation of oligonucleotides that are not in the immune complex is prevented by added blocking sequences. (The color version of this figure may be viewed at www.immunoassayhandbook.com). © 2009 BioTechniques. Reproduced with permission from Darmanis et al., Biotechniques 43, 443–450 (2007).
  17. 17. 83CHAPTER 2.3 Homogeneous Immunoassays event. β-Decay isotopes such as 3H, 33P, 35S, and 125I, each of which emits beta particles with energies below 300keV, are the most practical (Udenfriend et al., 1987). Electroactive Labels AMPEROMETRIC DETECTION Ferrocene can be readily oxidized at a voltage that does not affect most serum components (340mV vs saturated calomel electrode (SCE)). Because electron transfer to an electrode requires a very close approach, the current falls off as the ferrocene is made more bulky. Homogeneous electrochem- ical immunoassays have been designed using this principle. Ferrocene is conjugated to a hapten such as thyroxine. Elec- trolytic oxidation of the conjugate is measured in the pres- ence of an electrochemically inert agent designed to rapidly reduce the ferrocenium ions back to the neutral label. Glu- cose and GO are used for this purpose. When an antibody to the thyroxine is present, the current is reduced because of decreased accessibility of the conjugate to the electrode (see Fig. 17). Free thyroxine competes for the antibody and causes an increased current (Robinson et al., 1986). ELECTROCHEMILUMINESCENCE Measurement of electrochemiluminescence (ECL) has enabled the development of highly sensitive immunoas- says. The method was originally demonstrated by using a pyrene label that can be oxidized at an electrode and then rapidly reduced by cyclic voltammetry. At the appropriate voltages, the latter process is sufficiently exothermic that the pyrene ends up in an excited state that subsequently emits a photon. As in amperometric detection, the process is hindered when the label is associated with a bulky com- plex. Thus binding of anti-albumin antibodies reduces the emission produced by pyrene-labeled albumin, and com- petitive binding by free albumin causes recovery of the sig- nal (Ikariyama et al., 1985). Much improved sensitivity can be obtained by using a strongly fluorescent ruthenium II chelate in place of the pyrene (Blackburn et al., 1991). Like pyrene, the ruthe- nium is oxidized at an electrode, but it is then reduced without reversing the electrode polarity by including tri- propylamine in the reaction mixture. Electrochemical oxi- dation of the amine occurs along with ruthenium oxidation, and the resulting radical cation undergoes deprotonation to form an amine free radical. This species is itself a strong reducing agent and transfers an electron back to ruthe- nium. The process is sufficiently exothermic to leave the ruthenium in an electronically excited state that subse- quently emits a photon (see Fig. 18). Immunoassays are carried out by causing the ruthenium label to bind to antibodies on suspendable beads. Depend- ing on whether the assay is competitive or a sandwich (immunometric) immunoassay, an analog of the analyte or a second antibody is labeled. The beads serve to reduce the accessibility of the ruthenium to the electrode. In a com- petitive assay, labeled antigen is inhibited from binding to the beads so that the signal increases with increasing anti- gen concentration. In sandwich assays, labeled antibody binds to the beads in response to increasing antigen con- centration, causing the signal to decrease. Methods have also been worked out that reverse the pro- tective effect of the beads so that the signal increases when the label is bound to the beads. This is achieved by using mag- netic beads that are concentrated on the electrode and washed prior to reading. This heterogeneous protocol provides FIGURE 17 Homogeneous electrochemical immunoassay. Ferrocene (Fc)-labeled antigen produces a current by shuttling electrons between an anode and reduced glucose oxidase [GO(H)]. The ferrocene-labeled antigen is inactive when bound to antibody. (The color version of this figure may be viewed at www.immunoassayhandbook.com). FIGURE 18 Homogeneous electrochemiluminescence immunoassay. Ru (II)-labeled antigen and tripropylamine are simultaneously oxidized at an anode. Reaction between the oxidation products, aminopropyl radical and Ru (III), populates an excited state of Ru (II), which subsequently emits light. Oxidation of Ru (II)-labeled antigen is inhibited when bound to antibody. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  18. 18. 84 The Immunoassay Handbook higher sensitivity and extends the method to very low con- centration analytes such as serum TSH. ECL assays of this type have been designed for many clinically important ana- lytes and are used in Roche’s Elecsys® immunoassay system. Oxygen-Channeling Immunoassays LUMINESCENT OXYGEN-CHANNELING IMMUNOASSAY In order to achieve ultrahigh sensitivity in a homogenous immunoassay without resorting to time-consuming amplifi- cation steps, a label must be identified that not only is detectable at very low concentrations but also is able to pro- duce a signal that can be modulated. With the exception of isotopic labels, which provide only intermediate sensitivity, all modulatable labels are affected to some degree by the sample matrix. High quantum yield fluorescent compounds with high extinction coefficients, and light-scattering metal nanoparticles offer excellent sensitivity with single-molecule detection techniques, but weak signals can be obscured by background emission from bulk samples. Chemilumines- cence detection is inherently much less sensitive for single- molecule detection because, at most, only one photon can be produced by each molecule. However, the chemilumi- nescent background is immeasurable in most types of sam- ples, and chemiluminescent labels are therefore the best choice for measuring very low concentrations as distinct from very low absolute amounts. Unfortunately, the high susceptibility of chemiluminescent reactions to medium effects has thwarted many attempts to design practical homogeneous chemiluminescent immunoassays. Luminescence oxygen channeling immunoassay (LOCI) is the only homogeneous immunoassay method developed, to date, that can take full advantage of the exquisite sen- sitivity of chemiluminescence (Ullman et al., 1996). This is accomplished by inducing a chemiluminescent reaction within the interior of latex beads, where it is completely isolated from the sample matrix. The method uses two types of ligand- or receptor-coated latex beads that are suf- ficiently small (~250nm) that they do not settle out from a water suspension. One of the beads is a chemiluminescer bead in which is dissolved an olefin that can react rapidly with singlet oxygen (1∆gO2) to produce an electronically excited product. The reaction yields a dioxetane that spon- taneously decomposes in about 1s with efficient emission of light. The other bead contains a dissolved photosensitizer such as a phthalocyanine that is capable of exciting oxygen molecules to their singlet state upon exposure to light. The sensitizer is also physically separated from contact with the sample and thus insulated from nonspecific medium effects. The LOCI assay principle has similarities to enzyme- channeling immunoassays (see Fig. 19). The sensitizer beads are excited by wavelengths of light that are suffi- ciently long that there is little excitation of the chemilumi- nescer. The singlet oxygen that is formed within the beads diffuses into the aqueous solution. Because it has only a 4µs lifetime in water, it exists only in the immediate vicinity of each sensitizer bead and is virtually undetectable beyond about 300nm from the bead surface. When a chemilumi- nescer bead is bound to a sensitizer bead, it is exposed to the singlet oxygen, which diffuses into the bead and initiates the chemiluminescent reaction. Unbound beads are too remote to be affected. Although the change in the number of pho- tons produced per binding event is less than FRET immu- noassays, the method produces over 104 more photons per binding event than standard chemiluminescent assays that are limited to less than one photon per binding event. Any of the binding reagents required for a particular assay format can be bound to the bead surfaces, provided the reaction leads to analyte-dependent binding between the two types of beads. Following the binding reaction, the bead suspension is irradiated using a 680nm laser for several 0.1– 1s periods, and the shorter wavelength chemiluminescent FIGURE 19 LOCI. Singlet oxygen produced upon photoexcitation of a sensitizer particle (S) is intercepted by a bound chemiluminescent particle (CL) and induces a delayed chemiluminescent emission. In the absence of an antigen, binding does not occur, and the singlet oxygen decays prior to encountering a CL. (The color version of this figure may be viewed at www.immunoassayhandbook.com).
  19. 19. 85CHAPTER 2.3 Homogeneous Immunoassays emission at 550–650nm that is produced is integrated for similar time periods. Only a luminometer equipped with a laser and a mechanical shutter is required. Because the sig- nal intensity is seldom limiting, the measurements are made well before the binding equilibria are established. The sig- nal therefore is a measure of the number of bead pairs formed in a specific time period. Assay sensitivity is at least equal and frequently superior to the most sensitive heterogeneous immunoassays. In a TSH sandwich immunoassay using different anti-TSH antibodies on each of the beads, as few as 120,000 mole- cules (230aM in the final 900µL assay solution or 4.1fM in the 50µL serum sample) can be detected with a total incu- bation period of 14min. Competitive assays suitable for detection of pM concentrations of drugs such as digoxin can be run in under a minute. Protocols for sandwich immunoassays have been devel- oped that minimize incubation times and limit the very weak background chemiluminescence produced by long wave- length excitation of the chemiluminescent olefin. In a typical sandwich assay, the sample is first incubated with a mixture of antibody-coated chemiluminescer beads and an excess of biotinylated antibody against a different antigen epitope. The concentration of the beads is kept low to reduce back- ground, but moderate antibody concentrations can be obtained because of the 1000 or more antibody-binding sites per bead. This allows efficient formation of sandwich structures without dependence on slow bead–bead binding. Following a short incubation period, streptavidin-coated sensitizer beads are added and become bound in proportion to the number of biotins bound to the chemiluminescer beads. A large excess of the sensitizer beads can be used to accelerate bead–bead binding because the sensitizer beads do not contribute to background. The chemiluminescence emission is measured after a second short incubation period. Using these conditions, assay ranges in excess of five orders of magnitude in analyte concentration are obtained. Assays for drugs, proteins, antibodies, cell surface antigens, recep- tors, and nucleic acids have all been demonstrated. The sample matrix has little effect on the LOCI signal. The singlet oxygen can in principle be intercepted by sam- ple components during transit between beads, but its life- time is so short that it can only be captured by compounds that are both very reactive and in high concentration. Pro- teins are the primary quenchers of the singlet oxygen in serum (Wagner et al., 1993). This has little effect on sam- ple to sample variation, and the problem is usually avoided by using a final concentration of no more than 10% serum or plasma. High ascorbate concentrations in urine follow- ing megadose vitamin C therapy represent a more serious problem that necessitates a relatively high final dilution step (<1% urine) or inclusion of an oxidant. The method is used commercially on the Dimension Vista® analyzers from Siemens and in Perkin Elmer’s high-throughput AlphaScreen® and Alphalisa® for drug discovery. Conclusion Despite limited sensitivity relative to some heterogeneous assays, older homogeneous immunoassay methods such as EMIT, CEDIA, FRET, and latex agglutination have found widespread use due to ease of automation and the opportunity to carry out the assays manually with com- mon laboratory instrumentation. Over the past decade, new methods have been developed that achieved greater sensitivity and reduced susceptibility to matrix effects. The more recently developed ECL, magnetic aggrega- tion, LOCI, and PLA immunoassays retain to varying degrees the simplicity of earlier methods. These methods are minimally affected by the sample matrix and can be carried out with very small sample volumes. The recent commercial availability of generic homogeneous immuno- assay reagents for LOCI is likely to accelerate replacement of conventional heterogeneous methods such as RIA and ELISA used in both commercial applications and basic research. One important attribute of homogeneous methods that does not compare favorably with heterogeneous methods is their limited multiplexing capability. Considering the continuing strong interest in developing new homoge- neous methods, it is likely that this limitation will soon be overcome. A useful method for multiplexing is based on measuring binding of a fluorescent ligand to beads that are coated with different capture antigens or antibodies. Each capture ligand is coded by a unique fluorescent signature associated with its bead. Binding of a fluorescent label to a specific ligand can in this way be determined by using flow cytometry (Lima & Zhang, 2007). Arguably, this is a homogeneous method although a separation step is often used and the instrumentation is complex. By substituting flow cytometry with homogeneous fluctuation analysis, individual coded beads could in principle be analyzed in bulk suspension with a much simpler protocol. Addition- ally, the use of quantum dots and metallic nanoparticles as identity codes offers the opportunity for increased sensi- tivity. By further using light scattering and diffusion rates as additional coding parameters, it is not unreasonable to anticipate future development of highly sensitive homoge- neous methods that can rival the level of multiplexing of large scale heterogeneous methods. References and Further Reading Algar, W.R. and Krull, U.J. New opportunities in multiplexed optical bioanalyses using quantum dots and donor-acceptor interactions. Anal. Bioanal. Chem. 398, 2439–2449 (2010). Arai, R., Nakagawa, H., Tsumoto, K., Mahoney, W., Kumagai, I., Ueda, H. and Nagamune, T. Demonstration of a homogeneous noncompetitive immunoas- say based on bioluminescence resonance energy transfer. Anal. Biochem. 289, 77–81 (2001). Armenta, R., Tarnowski, T., Gibbons, I. and Ullman, E.F. Improved sensitivity in homogeneous enzyme immunoassays using a fluorogenic macromolecular substrate: an assay for serum ferritin. Anal. Biochem. 146, 211–219 (1985). Arquilla, E.R. and Stavitsky, A.B. The production and identification of antibodies to insulin and their use in assaying insulin. J. Clin. Invest. 35, 458 (1956). Austin, R.H. and Lim, S. Proc. Natl Acad. Sci. U. S. A. 105, 17217–17221 (2008). Baudry, J., Rouzeau, C., Goubault, C., Robic, C., Cohen-Tannoudji, L., Koenig, A., Bertrand, E. and Bibette, J. Acceleration of the recognition rate between grafted ligands and receptors with magnetic forces. Proc. Natl Acad. Sci. U. S. A. 103, 16076–16078 (2006). Blackburn, G.F., Shah, H.P., Kenten, H., Leland, J., Kamin, R.A., Link, J., Peterman, J., Powell, M.J., Shah, A., et al. Electrochemiluminescence detection for development of immunoassays and DNA probe assays for clinical diagnos- tics. Clin. Chem. 37, 1534–1539 (1991). Blecka, L.J., Shaffer, M. & Dworschack, R. Inhibitor enzyme immunoassays for the quantification of various haptens: a review. In: Immunoenzymatic Techniques (eds Avrameas, S., Druet, P., Masseyeff, R. and Feldmann, G.), 207–214 (Elsevier, New York, 1983). Boyden, A., Bolton, E. and Gemeroy, D. Precipitin testing with special reference to the measurement of turbidity. J. Immunol. 57, 211 (1947).

×