CA1214991A - Immunoassay - Google Patents

Abstract

IMMUNOASSAY
Abstract of the Disclosure Apparatus and method for providing an immunoassay of a binding reaction between a ligand and an antiligand which are typically an antigen and an antibody, including a spatial pattern formed by a spatial array of separate regions of antiligand material, and ligand material dispersed to interact with the spatial array of separate regions of antiligand material for producing a binding reaction between the ligand and the antiligand in the spatial patterns and with the bound complexes labeled with a particular physical characteristic. A
source of input energy and with the input energy at a particular spectrum for interacting with particular physical characteristic of the labeled binding reaction. Scanning the spatial pattern with the input energy at the particular spectrum for producing output energy having amplitude levels formed by a substantially random background component and a non-random component representing the labeled bound complexes, and the non random component representing the labeled bound complexes detected to produce an output signal in accordance with the labeled binding reaction.
************

Description

2 ~

4 The present invention is directed to a new type of immunoassay, which includes a sensitive technique for the 6 quantitative detection of low concentrations of molecules of a 7 particular type such as molecules in solution.

9 Description of the Prior Art , 11 It is desirable in certain circumstances to measure 12 very low concentrations of certain organic compounds~ In 13 medicine, or example, it is very useful to determine the 14 concentration of a given kind of molecule, usually in solution, which either exists naturally in physiological fluids (eOg.
16 blood or urine~ or which has been introduced into the living 17 system (e.g. drugs or contaminants). Because of the rapidly 18 advancing state of understanding of the molecular basis of both 19 the normal and diseased states of living systems, there is an increasing need for methods of detection which are quantitative, 21 specific to the molecule of interest, highly sensitive and 22 relatively simple to implement. Examples of molecules of 23 interest in a medical and/or biological context include, but are 24 not limited to, drugs, sex and adrenal hormonesl biologically active peptides, circulating hormones and excreted antigens 26 associated with tumors. In the case of drugs, for example, it 27 is often the case that the safe and efficacious use of a 28 particular drug requires that its concentration in the 29 circulatory system be held to ~ithin relatively narrow bounds, referred to as the therapeutic range~
' l ~,

3~

1 One broad approach used to detect the presence of a Z particular compound, referred to as the analyte, is the 3 immunoassay, in which detection of a given molecular species,

4 referred to generally as the ligand, is accomplished through the use of a second molecular species~ often called the antiligand, 6 or the receptor, which specifically binds to the first compound 7 of interest. The presence of the ligand of interest is detected 8 by measuring, or inferring, either directly or indirectly, the 9 extent of binding of ligand to antiligand~ The ligand may be either monopitopic or polyepitopic and is generally defined to 11 be any organic molecule for which there exists another molec~le 12 (i.e. the antiligand) which specifically binds to said ligand, 13 owing to the recognition of some portion of said ligand.
14 Examples of ligands include macromolecular antigens and haptens (e.g. drugs)~ The antiligand, or receptor, is usually an 16 antibody, which either exists naturally or can be prepared 17 artificially. The ligand and antiligand together form a 18 homologous pair. Throughout the text the terms antigen and 19 antibody, which represent typical examples, are used Z interchangeably with the terms ligand and antiligand, 21 respectively, but sueh usage does not signify any loss of 22 generality. In some eases, the antibody would be the ligand and 23 the antigen the antiligand, if it was the presence of the 24 antibody that was to be detected.

26 Implementation of a successful immunoassay requires a 27 detectable signal which is related to the extent of 28 antigen-antibody binding which occurs upon the reaetion of the 29 analyte with various assay reagents. ~sually that signal is provided for by a label which is conjugated to either the ligand or the antiligand, depending on the mode of operation of the ~2 immunoassay. Any label which provides a stable, conveniently 2 ~

1 detectable signal is an acceptaole candidate~ Physical or 2 chemical effects which produce detectable signals, and for which 3 suitable labels exist, include radioactivityt fluorescence, 4 chemiluminescence, phosphorescence and enz~natic activity~ to name a few.
7 Broadly speaking, immunoassays fall into two general 8 categories -- heterogeneous and homogeneous~ In heterogeneous 9 assays, the purpose of the label is simply to establish the location of the molecule to which it is conjugated -- iae. to 11 establish whether the labeled molecule is free in solution or is 12 part of a bound complex. ~eterogeneous assays generally 13 function by explicitly separating bound antigen-antibody 14 complexes from the remaining free antigen and~or antibody. A
me~hod which is frequently employed consists of attaching one of 16 the members of the homologous pair to a solid surface by 17 covalent binding, physical absorption~ or some other means.
18 When antigen-antibody binding occurs, the resulting bound 19 complexes remain attached to this solid surface (composed of any suitably inert material such as plastic, paper, glass, metal, 21 polymer gel, etc.), allowing for separation of free antigen ~2 and/or antibody in the surrounding solution by a wash step. A
23 variation on this method consists of using small (typically 0.05 24 to 20 microns) suspendable particles to provide the solid surface onto which either antigen or antibody i5 immobilized.
26 Separation is effected by centrifugation of the solution of 27 sample, reagents and suspendable beads at an appropriate speed, 28 resulting in selective sedimentation of the support particles 29 together with the bound complexes.

vl ~2 1 Nothwithstanding the successful application of 2 heterogeneous assay procedures, it is generally desirable to 3 eliminate separation steps, since the latter are time-consuming, 4 labor-intensive and sometimes the source o~ errors in the signal measurement~ Furthermore, the more complicated protocols 6 associated with heterogeneous assays make them less suitable for 7 automated instrumentation of the kind needed for large-scale 8 clinical applications. Consequently, homogeneous assays are 9 more desirable. In the homogeneous format, the signal obtained from the labeled ligand or antiligand is modified, or modulated, 11 in some systematic, recognizable way when ligand-antiligand 12 bin~ing occurs. Consequentlyt separation of the labeled bound 13 complexes from the free labeled molecules is no longer 14 required.

16 There exist a number of ways in which immunoassays can 17 be carried out. For clar-ity a heterogeneous format is assumed, 18 although each aproach can be utilized (with varying degrees of 19 success) in a homogeneous format, given a suitable label which is modulated by the binding reaction.

22 In the competitive mode, the analyte, assumed to he 23 antigen, is allowed to compete with a known concentration of 24 labeled antigen ~provided in re~agent form in the assay kit) for binding to a limited number of antibody molecule5 which are 26 attached to a solid matrix. Following an appropriate incubation 27 period, the reacting solution i5 washed away, ideally leaving 28 just labeled antigen-antibody complexes attached to the binding 29 surface, thereby permitting the signal from the labels to be quantitated.
~1 ~2 3~

1 In another method, called the sandwich mode, the 2 analyte, again assumed to be antigen, reacts with an excess of 3 surface-i~mobilized antibody molecules~ After a suitable 4 incubation period, an excess of label-conjugated antibody is added to the system. After this reaction has gone to essential 6 completion, a wash step removes unbound labeled antibody and 7 other sources of contamination, permitting measurement of the 8 signal produced by labels which are attached to 9 antibody-antigen-antibody complexes~

11 In yet another approach, called the indirect mode, the 12 analyte, this time assumed to consist of specific antibody, is 13 allowed to bind to surface-immobilized antigen which is in 14 excess. The binding surface is then washed and allowed to react with label-conjugated antibody. After a suitable incubation 16 period the surface is washed again, removing free labeled 17 antibody and permitting measurement of the signal due to labeled 18 antibody. The resulting signal strength varies inversely with ~9 the concentration of the starting (unknown) antibody, since labeled antibody can bind only to those immobilized antigen 21 molecules which have not already complexed to the analyte.

23 One of the most sensitive immunoassays developed 24 thusfar is the radioimmunoassay (RIA), in which the label is a Z5 radionuclideO such as I125, conjugated to either member of 26 the homologous (binding) pair. This assay, which is necessarily 27 heteroqeneous, has achieved extremely high sensitivities, 28 extending down to the vicinity of 10-17 molar for certain 29 analytes. The obvious advantage of radioactive labeling, and the reason for the extremely high sensitivity of RIA-type ~,`1 9~

1 assays, is that there exists negligible natural background 2 radioactivity in the samples to be analyzed. Also, RIA is 3 relatively insensitive to variations in the overall chernical 4 composition of the unknown sample solution. However~ the radioactive reagents are expensive, possess relatively short 6 shelf lives and require the use of sophisticated, expensive 7 instrumentation as well as elaborate safe~y measures or both 8 their use and disposal. Hence, there is an increasing ~ motivation to develop non-isotopic assays.

11 Fluorescence provides a potentially attractive 12 alternative ~o radioactivity as a suitable label for 13 immunoassays. For examplet fluorescein (usually in the form of 14 fluorescein isothiocyanate, or "FITC"~ and a variety of other fluorescent dye molecules can be attached to most ligands and 16 receptors without significantly impairing their binding 17 properties. Fluorescent molecules have the property that they 18 absorb light over a certain range of wavelengths and (after a lg delay ran~ing from 10-9 to 10-4 secondsl emit light over a range of longer wavelengths. Hence, through the use of a 21 sui~able light source, detector and optics, including excitation 22 and emission filters, the fluorescence intensity originating 23 from labeled molecules can be determined.
2~
Several heterogeneous fluorescence-based immunoassays 26 (FIA) have been developed, including the FIAX/StiQTM method 27 (IDT Corp., Santa Clara, CAo ) and the FluoromaticTM method 28 (Bio-Rad Corp., Richmond, CA.). In the former case/ antigen is 29 immobilized on an absorbant surface consisting of a cellulose-likè polymer mounted on the end of a portable 1 "dipstick", which is manually inserted into sample, reagent and 2 wash solutions and ultimately into the fluorescence measuring 3 instrument. A competitive reaction utilizing FITC-labeled 4 monospecific antibody is typically em~loyed. In the Bio-Rad assay kit, the solid surface is replaced by suspendable 6 polyacrylamide gel microbeads which carry covalently-bound ~ specific antibody. A sandwich mode is typically employed, with 8 centrifugal sedimentation, followecl by resuspension, of the 9 beads for separation and measurement. Photon-counting techniques can be used to extend the sensitivity of the 11 fluorescence intensity measurementO

13 Use of an enzyme as a label has produced a variety of 14 useful enzyme immunoassays (EIA), the most popular of which is known as ELISA. In the typical heterogeneous format a 16 sandwich~type reaction is employed, in which the ligand of 17 interest, assumed here to be antigen, binds to 18 surface-immobilized specific antibody and then to an 19 enzyme~antibody conjugate. After suitable incubation, any remaining free enzyme conjugate is eliminated by a wash or 21 centrifugation step. A suitable substrate for the enzyme is 22 then brought into contact with the surface containing the bound 23 complexes. The enzyme-substrate pair is chosen to provide a 24 reaction product which yields a readily detectable signal~ such as a color change or a fluorescence emission. The use of an 26 enzyme as a label services to effectively amplify the 27 contribution of a single labeled bound complex to the measured 28 signal, because many substrate molecules can be converted by a 29 single enzyme molecule~

~.)1 1 As discussed previously, it is generally desirable to 2 eliminate the separation steps associated with typical 3 heterogeneous assays and, instead, use homogeneous techniques.
4 One of the first homogeneous assays to be developed was the fluorescence polar;zation immunoassay. Here, the polarization 6 of the emission of the fluorescent dye label is modulated to an 7 extent which depends on the rate of rotational diffusion, or 8 tumblin~, of the label in solution~ Free labeled molecules 9 which rotate rapidly relative to the life-time of their excited states emit light of relatively random polarization (assuming a 11 linearly polarized exciting beam, for example). However, when 12 the label becomes attached to a relatively large bound complex, 13 the rate of tumbling becomes relatively slow/ resulting in 14 fluorescence emission of substantially linear polarization (i.e.
essentially unchanged). Unfortunately, this technique is 16 limited in practice to the detection of low molecular weight 17 ligands, e.gO drugs, whose rate of tumbling is sufficiently 18 rapid to produce a measurable change in fluorescence 19 polarization upon binding to the antiligand. The extent of modulation of the signal, in any caset is quite small.

22 Another useful fluorescence-based homogeneous 23 technique is the fluorescence excitation transfer immunoassay 24 (FETI), also Xnown simply as fluorescence quenching. Here, two different dye labels, termed the donor and the acceptor, or 26 quencher are used. The pair has the property that when the 27 labels are brought close together, i.e. to within distances 28 characteristic of the dimensions of antigen-antibody complexesr 29 there is non-radiative energy transfer between the ~0 1 electronically excited donor molecule and the acceptorO That 2 is~ the acceptor quenches the fl~orescence emission of the 3 donor, resulting in a decreased intensity of the latter. In a 4 typical competitive mode, the donor label is attached to the ligand o interest and the acceptor label fixed to the specific 6 antibody. When ligand is present in the unknown sample7 some fraction of the acceptor-labeled antibody binds to the free 8 ligand, leaving a fraction of the labeled ligand unquenched and 9 therefore able to emit fluorescence radiationO The intensity of the latter increases with increasing analyte concentration.

12 The principal drawback of the FETI technique is the 13 requirement that the donor-labeled ligand be relatively pure.
14 Substan~ial concentrations of labeled impurities produce a large background signal, making detection of a small change due to ~.S
16 complexing all the more difficult. Along these lines,~Patent 17 #4,261,968 describes an assay in which the quantum efficiency of 18 a fluorescent label is decreased when the labeled antigen 19 bec~mes bound to the antibody, resulting in a decrease in the total fluorescence emission of the sample solution.

22 One of the main factors which limits the sensitivity 23 and reproducibility of all non-isotopic assays to varying 24 degrees is the presence of background false signals. For example, in fluorescence-based assays the use of untreated blood 26 serum may yield relatively high and variable background 27 fluorescence levels due to the presence of proteins, bilirubin 28 and drugs. In addition, there may exist variations in the 29 absolute fluorescence intensity from one sample to the next due to fluorescence from sample cell surfaces, light scattering from ~,1 1 impurities in solution, aberrations on optical surfaces, 2 temperature dependent effects, etc~ Problems related to 3 impurities are particularly troublesome in homogeneous assays.
4 However, the background false signal contrib~tions are often relat~vely constant in time for any given sample measurementO
6 Hence, a very useful technique for reducing the background ~ contribution without the necessity of making additional control 8 measurements i5 to determine the time rate of change of the 9 signal. Such a rate determination in the early stages of the antigen-antibody binding reaction (i~e. when the rate is 11 largest) should, in principle, be independent of the ~constant) 12 background level.

14 In principle, thenr the rate determining procedure can be applied to any homogeneous assay technique, with the added 16 advantage that the binding reaction need not be taken to 17 essential completion, thereby resulting in a faster assay 18 measurement. However, this approach becomes less feasible or 19 advantageous the smaller the total signal change due to binding, relative to the background level. Hence, there are invariably 21 practical limitations to the sensitivity which can be achieved 22 using any of the existing homogeneous non-isotopic immunoassays, 23 given the typical courses of background false signalsr 24 interferences and nonspecific effects.

~.,1 1 SUMMARY OF THE IN~ENTIO~
3 The present invention is directed to an 4 immunoassay technique in which the contrihutions to the measured signal due to free labeled ligand or antiligand as well as 6 contaminants and other sources of "background" signal are 7 effectively suppressed, thereby permitting measurement of the 8 desired signal due to labeled bound molecules at very low 9 concentrations. The resulting assay, therefore, possesses a higher sensitivity than those currently in use. This 11 requirement of insensitivity to free labeled molecules plus 12 background contamination holds regardless of the type of 13 labeling and/or detection scheme employed.

The underlying principle of the present invention 16 is that labeled bound ligand-antiligand complexes are caused to 17 reside preferentially in a predetermined spatial pattern. rrhe 1~ useful signal due to these labels therefore is forced to exist 19 at or be associated with only certain predetermined locations within the sample solution volume or on a surface. This 21 behavior is in sharp contrast to the origin of the signal 22 due to unbound labeled ligand or labeled antiligand molecules as 23 well as background contamination sources of signal which can be 24 expected to be spatially random. The spatial pattern is scanned spatially while the antigen-antibody binding reaction is 2~ proceeding or, if desired, after the reaction has run to 27 completion. Using signal enhancement techniques such as 28 filtering, phase-sensitive detection or autocorrelation, the 29 desired signal level can be ~reatly enhanced with respect to the ~ contribution from free labeled molecules and background ~ 1 _ 1 contamination sources~ thereby yielding a more sensitive assay.
2 It is to be appreciated that the present invention is not 3 limited to homogeneous assays although such homogeneous assays 4 are desirable. Alsor the general principle of having the bound complexes reside in a known spatial pattern is not limited to 6 the use of fluorescent labels~ The detected si~nal can be 7 related, for example, to changes in optical density teither at a 8 specific wavelength or over a broad range of wavelengths), light 9 scattering, color, reflectance, birefringence, magnetism or any other physical variable which can be detected with suitable 11 sensitivity and spatial resolutiong 13 The present invention may be accomplished with a 14 number of different embodiments, but the common principle in each case is to force the bound complexes produced by the 16 ligand-antiligand reaction to occur predominantly in a 17 predetermined spatial (or geometric~ pattern and then to scan 18 the region in space with a suitable detector where this pattern 19 is expected to be located and, using noise reduction techniques, to measure the signal due to the binding reaction and 21 effectively suppress the contribution due to free labeled 22 molecules plus background contaminants, which display no such 23 spatial pattern. For example, if the analyte consists of 24 antigen, and labeled antigen is employed in a competitive-type assay, specific antibody can be attached to a solid surface in 26 the form of a spatially periodic array of stripes of 27 predetermined width and spacing. After completion of a 28 competitive reaction as described above, labeled bound antigen 29 will be found only on the antibody-coated stripes, while free 1 labeled antigen and other signal-pro~ucing interfering 2 contaminants will in general exist throughout the sample space 3 with more-or-less equal density when there is free diffusion and 4 adequate mixing of all unbound species prior to completion of the hinding reaction. Therefore~ the signal due to the 6 particular label used will be enhanced at the locations of the 7 antibody-coated stripes.

9 ~ the Drawings 11 A clearer understanding of the invention will be 12 had with reference to the following description and drawings 13 wherein;

14 Figure 1 is a first embodiment of the invention;

16 Figure 2 illustrates the output waveform from the 17 embodiment of Figure 1;

1~

19 Figure 3 is a second embodiment of the invention;

2Z Figure 4 is a third embodiment of the invention;

~3 24 Figure 5 illustrates the output waveform from the embodiment of Figure 4;

27 Figures 6(a), 6(b) and 6(c) illustrate three 28 alternate methods of providing an output signal representative 29 Of the periodic signal;

~,1 ~2 _ 13 -1 Figure 7 illustrates a waveform representative of 2 a correlation function provided by the method shown in Figure 3 6(c); and Figure 8(a) and 8(b~ illustrates alternate forms 6 of a fourth embodiment of the invention~

8 Descriptlon of the Preferred Embodiments A sample solution 10 is contained in a ll transparent cylindrical glass or plastic tube l2. For the 12 purpose of describing this particular embodiment, it is assumed 13 that the analyte to be measured oonsists of antigen and that à
14 competitive mode assay is employed. The inside surfa~e of the tubc 12 is coated with specific antibody in the form of a ribbed 16 pattern of stripes 14. Typically~ the stripes 14 of antibody 17 coating are of equal width and with the spacing between adjacent 18 stripes e`qual to the width of the stripes, so that, half the 19 inner tube surface is coated with antibody. The tube 12 is rotated by a motor 16 at a constant angular speed ~O~
21 equivalent to fO revolutions/sec, where fO = wo/2~ The 22 tube 12 is illuminated by an exciting light source such as, blue 23 light of wavelength ~ = 485 nm in the case of fluorescein 24 laheled molecules. The exciting light consists of a narrow, 1at ribbon of light and may be produced using a broad spectrum 26 light source 18 and with the light energy from the light source 27 18 directed through a collimating lens 20 and a cylindrical lens 28 22 to have the plane of the focused beam aligned parallel to the 29 stripes on the rotating tube 12 and focused on the surface of the tube. A filter 24 is used to remove all wavelengths but ~,1 l those in a range suitable or excitation of the fluorescent 2 labels. The focused beam excites any Eluorescently laheled 3 molecules or impurities which lie in the path of the beam within the sample tube 12. The exciting beam should ideally possess a width which is matched to the width of the antibody-coated 6 stripes 14 but the width of the beam should not exceed the width 7 of the stripes.

9 An imaging lens 26 is placed close to the rotating tube to collect a fraction of the resulting ll fluorescence emission and to image the emission onto a detector 12 such as a photo-multiplier tube 28 so as to produce a detected 13 fluorescent signal I. Although other types of detectors such as 14 solid-state photovoltic detectors may be employed~ a photo-multiplier tube would typically be used for the most 16 sensitive assay. A filter 30 is used to eliminate light at the 17 exciting wavelength which reaches the detector 28 due to stray 18 reflections from the sample tube 12 and/or scattering from 19 surface imperfections and particles in the solution 10. The ~esired fluorescent signal is at a longer wavelength than the 21 exciting light.
~2 23 There are two kinds of contributions to the 24 resulting detected fluorescent signal I. First, there is a signal of magnitude Io~ consisting of background fluorescence 26 from the glass tube 12 plus impurities in the sample solution as 27 well as unbound labeled antigen ~assuming, for example, a 28 competitive-type assay). In general these background 29 fluorescent sources are expected to be randomly distributed throughout the tube volume, so that the resulting signal bears l no systematic relationship to the angular orientation of the 2 tube 1~ and there is no characteristic frequency component or 3 phase in this background signal contribution. Second, there is 4 a desired signal of magnitude AI due to labeled antigen which has become bound to the fixed antibody pattern. This~
6 contribution of magnitude ~I consists of a periodically varying 7 signal of known frequency f, given by f = nfO, where n is the 8 number of antibody-coated stripes 14 located on the inner tube 9 surface. The phase of this bound-label oscillating signal of magnitude ~I may be determined with respect to an oscillating ll reference signal of the same frequency, which can be derived 12 from the rotation of the tube 12. For example, the reference 13 signal may be derived from a pattern of spots or holes 32 of the 14 same periodici~y as the antibody stripes 14 located on a disc 34 attached to the rotor shaft of the motor 16~ The reference l6 signal may be obtained by an interrupted light beam from a light 17 source 36 and detected by a photo-detector 38. Other simple 18 techniques, such as a magnetic-field-actuated switch may also be l9 used to produce a reference signal.

21 The detected signal, I, from the photo-multiplier ~2 detector 28 is ideally as shown in Figure 2. A5 can be seen in 23 Figure 2 the total detectable fluorescence signal I has been 24 converted into the sum of two components: a periodic waveform of frequency f and magnitude ~ I due to the bound complexed labeled 26 antigen~ superimposed on a baclcground level of intensity Io 27 which has been represented as a constant~ Regardless of the 28 detailed shape of the periodic component, which depends on the 29 precise relationship between the exciting beam size and the stripe width, plus other factors, the desired signal of 1 magnitude ~I contains a stron~ Fourier component at the 2 fundamental frequency f. Of course, in reality there is a 3 certain amount of random and/or systematic "noise" superimposed 4 on the ideal signal shown in Figure 2 as will be discussed below.

7 A simple method of extracting the desired signal 8 of magnitude AI consists of passing the output signal from the 9 photo-multiplier 2B through a narrow-frequency bandpass filter 40 as shown in Figure 6(a). The filter 40 is centered at 11 frequency f, and the peak-to-peak amplitude ~I or the 12 root-mean-s~uare (R.M.S.) amplitude may then be measured by an 13 amplitude detector 42 to provide an output signal representative 14 of the oscillating component of the total detected signal. This 1~ method effectively blocks the large D~C. component of the 16 signal, Io~ due to unbound labeled antigen plus the remainin~
1~ background fluorescence. In situations where the signal ~rom 18 the pattern is weak due to a low density of labeled bound 19 complexes, the oscillating component may be diffi~ult to measure using a bandpass filter due to the presence of one or more of 21 the following sources of fluetuations (iOe., "noise") in the 22 system which may get through the filter:

24 (t) Fluctuations in the exciting ,beam intensity7 27 (2) Shot noise in the photo-multiplier detector.
28 The fluorescent light consists of discrete 29 photons and therefore has random noise , , associated with it.
~,1 1 (3) Imperfections in the glass tube surface 2 a~d/or non-uniformities in antibody 3 coating (4) Fluorescent impurity "spots" fixed to the S surface of the ~ube, whose brightness may 7 be comparable to or exceed that of the 8 labeled antigen attached to the 9 antibody-coated stripes.

11 A more powerful method of extracting the 12 periodic oscillating signal component from the total 13 fluorescence signal is the technique of phase-sensitive 14 detection, using a "lock~in" or phase-lock amplifier 44 shown in Figure 6(b). Lock-in amplifiers are frequently used for 16 extracting periodic signals, of fixed frequency and phase, from 17 large backgrounds which are random in time. Such amplifiers are 18 available in a sophisticated form from Princeton Applied 19 Research, Princeton, N.J. Simple lock-in amplifiers can be made with a few integrated circuits, for instance, ~lsing the GAP-01 21 Analog Processing Subsystem made by Precision honolithics, Inc., 2~ Santa Clara, CA. The basic idea of a lock-in amplifier is to 23 alternate the polarity of the input signal with a reference 24 signal and then average the resultant signal.

26 For signals which are not in phase with the 27 reference signal, such as background fluorescence, the average 28 of the output will be zero since half the time the output signal 29 will be positive and half the time negative. For a periodic signal such as a sine wave in phase with the reference signal 3~

- lB -the average of the output will not be zero. For instance if 2 when the input periodic signal is positive the polarity is not 3 reversed but when the input is negative the input polarity is 4 reversed by the reference signall the output will always be positive, and the average of this periodic signal will be 6 non-zero.

8 The phase-lock amplifier 44 therefore can 9 selectively amplify that portion of the input signal which possesses a Fourier component at the fundamental modulation 11 frequency (equal to f in the present case~ and whose phase is 12 fixed with respect to a reference signal of the same frequency 13 such as the reference signal produced from the detector 38. In 14 this way, a relatively weak signal at the fundamental frequency can be detected by an amplitude detector 46 from masking noise 16 sources whose r.m.s. amplitudes exceed the desired signal 17 amplitude by as much as several orders of magnitude The 18 important distinction is that the phases of the Fourier 19 components of the noisy background contribution are random with respect to the periodic reference signal from the detector 38J

22 Fluctuations (1~ and (2) defined above, occur 23 randomly in time, or with respect to the pattern period, and 24 therefore can be largely suppressed using a phase-lock amplifier ~5 44O Fluctuations (3) and (4) defined above, can in principle be 26 reduced by making the system (i.e., glass tube 12, antibody 27 coatings 14, etc.) as uniform as possible and by designing a 2a pattern which is less sensitive to single or only occasional 29 imperfections. Computer techniques also may be used in conjunction with pha5e-sensitive detection to recognize and 19 _ l reject a given stripe 14 whenever the stripe produces an 2 anomalously large signal contribution which the computer 3 determines is due to a major imperfection. The influence of a 4 particular stripe on the resulting output signal may thereby be ~ minimized.

~ As described above~ the assay can be performed in 8 a non-competitive mode, as in the "sandwich" assay. The 9 sandwich assay can employ the same pattern consisting of surface-immobilized antibody shown in the embodiment of Figure ll 1. The unknown sample 10 containing the antigen of interest is 12 put into the tube 12 and allowed to react with the 13 antibody-coated stripes 14. After a suitable incubation period 14 to allow for essential completion of the binding reaction an excess of fluorescently-labeled antibody is then added to the 16 sample solution 10O Assuming that the antigen possesses at 17 least two active binding sites, the antigen can indirectly bind lB labeled antibody to the surface-immobilized antibody, resulting l9 in antibody-antigen-antibody complexes. As the binding reaction proceeds a fluorescent pattern begins to develop on the pattern 21 of antibody stripes 14O With the sandwich assay, the intensity 22 of fluorescence emission associa~ed with the pattern is ideally 23 directly proportional to the number of antigen molecules in the 24 sampleO It is presumed that the number of antibody molecules attached to the tube surface in the form of a pattern 26 substantially exceeds the maximum number of antigen molecules in 27 the sample, so that essentially all of the latter become bound 28 to the surfaceO The narrow filter 40 or a phase-lock amplifier 29 44 can be used to extract the amplitude ~I of the periodic signal corresponding to the pattern.

1 In both the competitive and sandwich assays, the 2 la~eled ligand usually carries only a single fluorescent label 3 molecule. For this case the binding of one labeled antigen or 4 antibody molecule adds only one fluorescent molecule to the S pattern, which increments the measured siynal by an exceedingly 6 small amount. The tagging could be performed by coating a 7 tagged "carrier" particle with the antigen or antibody to be 8 tagged. For instance in the "sandwich" mode assay the labeled 9 antibody may be added to the system in the form of fluorescent "carrier" particles which are coated with antibody. These 11 particles rnay be composed of a suitable inert material such as 12 polystyrene~ polyacrylamide gel, etc., which has been 13 impregnated with fluorescent dye molecules such as by covalent 14 binding, hydrophobic association, etc. Antibody may be attached to the outer surface of the particles by covalent binding, 16 physical absorption, etc. The size of these particles would 17 typically be submicron~ The fluorescent carrier particles offer 1~ the advantage that they produce a fluorescent signal having a 19 much greater amplitude than that obtained from a single fluorescent label molecule~ The signal produced by the 21 fluorescent carrier particles may be thousands of times larger 22 depending on the number of fluorescent molecules which are 23 attached to each carrier particle~ Therefore~ a single binding 24 event increases the fluorescent intensity of one segment of the pattern by a much larger amount than would be the case using 26 ordinary labeled antibody. In effect, the use of carrier 27 particles provides additional amplification for the pattern 28 technique~ This arnplification becomes important if the assay is 29 to be highly sensitive so as to detect extremely low levels of antigen.

~,1 ~2 - 21 ~

1 If carrier particles are used in this way to 2 amplify the effect of labeling, the pattern technique can also 3 be used as an alternative to the well-known slide test, based on 4 particle agglutinationO A popular qualitative method for

5 performing an assay uses carrier particles (typically latex,

6 submicron) coated with specific antibody. The antigen in the

7 unknown sample binds to the antibody~ thereby crosslinking

8 various numbers o carrier particles into aggregates of various

9 sizes. If the agglutination proceeds far enough, representing that the antigen is present in sufficiently large concentration, 11 the aggregates scatter enough light, given a carrier particle 12 diameter of a fraction of a micron or larger, to be detectable 13 with the unaided eye. The resulting degree of flocculation of 14 the typically concentrated carrier particles provides a qualitative indiation of the presence of the antigen above a 16 certain concentrationQ For example, the pregnancy assay (hCG) 1~ is usually performed in this way.
lE~
13 An improved visual agglutination assay can be designed by attaching spe~ific antibody to a solid surface in 21 the form of a macroscopic spatial pattern which is easily 22 recognizable by the naked eye (i.e. any simple geometric design, 23 including an alphabet letter, number, etc.~. As in the simple 24 slide testt a certain amount of antib~ody-coated carrier particles in solution is used in contact with the 26 antibody-coated surface. Upon addition of antigen (unknown~ and 27 agitation of the solution in contact with the surface, some of 28 the carrier particles begin to bind to the surface-immobilized 29 antibody pattern. To the extent that this preferential positioning of some of the particles (and particle aggregates) 1 on the pattern causes a visually discernable change in color, 2 optical densityl reflectance/ light scattering or any other 3 optical property, the pattern begins to "stand out" amidst the 4 more-or-less random positioning of all the remaining particles.
For example, strongly fluorescing carrier particles can be used.
~ Of course, the antigen also induces some aggregation of free 7 particles in solution which remain unbound to the pattern. The 8 resulting contrast of the pattern with respect to the 9 surrounding solution then depends on the extent of binding to the pattern, the strength of the "signal" caused by the 11 par~icles, the ratio of the amount of part'icle-coated antibody 12 to pattern antibody~ and other factors.

l* Although the present invention has been described above with reference to the pattern technique using fluorescence 16 as the signal-producing label to be detected, it is clear that a 17 variety of other physical quantities can be employed as the 18 detectable label variable. One example is optical density~
19 For example, one can employ carrier particles which are relatively opaque (or semi-opaque) to white light or to light of 21 a given wavelength. Again~ taking the example in which the 22 unknown analyte is antigen and the carrier particles are coatea 23 with antibody, a~ the antigen-antibody binding reaction 24 progresses the antibody-coated stripes in the pattern will become steadily more optically dense, transmitting less light 26 over the course of the reaction. The signal can be obtained by ~7 a simple measurement of transmitted intensity through the 2~ pattern segments. In this case, however, another embodiment of 29 the present invention might be preferred, which embodiment is shown in Figure 3.
~1 1 The principal difference between the embodiments 2 of Fi~ures 1 and 3 is that the pattern now consists of radial 3 stripes 5n, or segments, fixed on a circular disc 52. The 4 sample solution exists as a layer 54 of liquid on the disc surface~ The d;sc 52 may be rotated by the motor 16 and with 6 the light source 18~ Collimating lens 20 and filter 24 7 provides for light energy directed to th~ stripes 50 on the disc 8 52. If the measurement is of optical density, the filter 24 is 9 optional~ The output signal is detected by the imaging lens 26, the filter 30 and the photo-multiplier tube ~8~ Again, if 11 optical density is being measured, the filter 30 is optionalO A
1~ phase-lock detection scheme, as shown in Figure 6(b) r may be 13 used to obtain the effective modulation amplitude of the 14 periodic signal, proportional to the concentration of optically dense particles attached to the pattern. Alternatively, the 16 filter system of Figure 6(a) may be used to provide a 17 measurement of the periodic signal.
1~ .
19 Another suitable label, related to optical density, is color. Colored dye molecules can be attached to the 21 antigen, for example, in the case of a competitive assay, or to ~2 the excess specific antibody, in the case of a sandwich assay 23 and the dye molecules can be detected by a color change.
24 Colored carrier particles may be used to enhance the change in coior which occurs due to binding to the pattern as described 26 above for optical density. As another example, red blood cells 27 can be used to hind to the pattern via the antigenic sites on 28 their membrane outer surfaces. A measurement of the optical 29 density at an appropriate wavelength would provide a sensitive test of the degree of binding of the red cells to the pattern.
~1 - 2~ -1 Of course~ other kinds of cells can be used for the same 2 purpose, to detect the presence of certain molecules on their 3 surfaces, provided that there results some detectable change in 4 colorr turbidity, etc.

6 Another physical quantity wh;ch may be measured 7 is reflected light~ Here, for example, the carrier particles 8 can be metallic particles (e.g. colloidal gold) coated with 9 antibody. As the binding progresses, the refl2ction of light from the rotating pattern design increases, giving a measure of 11 the amount of antigen-antibody binding and hence the amount of 12 antigen in the sample. In all of these above described 13 measurements the rate at which the pattern appears may be 14 measured. It is important to appreciate that in all of the 1~ pattern-related measurements described earlier, as well as those 16 to be described below, the magnitude of the fluctuating 17 component of the signal produced by pattern-bound labels, ~I, is 1~ obtained continuously in time. It obviously increases as more 19 labeled molecules become fixed to the pattern Hence~ in every case one has a homogeneous technique in which the time rate of 21 change of pattern-assooiated label binding can be obtained. As 22 pointed out in the Background of the Invention, such a rate 23 measurement is highly desirable, in that it provides additional 24 discrimination over back~round sources of false signal, which are usually approximately constant in timeO As well, rate 26 determination generally permits a shorter overall measurement 27 time, in that it does not require that the binding reaction go 28 to essential completion in order to measure the total signal 29 change associated with complete binding.

- ~5 -1 In all embodiments of the invention, the pattern 2 caused by binding can be formed either by a spatial 3 rearrangement of the labeled molecules or by a modulation of the 4 strength of the label at the binding sites on the pattern. In the flrst case it is important that all molecules comprising the 6 analyte, as well as all other compounds provided for as 7 reagents, labeled or otherwise, be able to come into efficient 8 contact with each of the antibody or antigen-coated segments 9 which comprise the spatial pattern~ In the second case this "mixing" is not needed but one could use a label whose strength 11 is modulated when the binding occurs. A fluorescent label which 12 is quenched upon binding is an example of such a label. Here 13 the fluorescent pattern can form not because of a spatial 14 rearrangement of the labels but because the strengths of the labels on the unbound molecules will be dif~erent than the 16 strengths of the labels on the molecules bound to the pattern.
17 Thls se~ond case could be used where the pattern is on a porous 18 medium and the labeled molecules do not have an opportunity to 19 move about over dimensions comparable to the characteristic size o~ the pattern and, therefore their concentration remains 21 essentially uniform spatially.

23 A third embodiment of the pattern technique of 24 the present invention is shown in Figure 4. In the embodiment of Fiyure 4, the antibody coating takes the form of a set of n 26 parallel stripes 60 on a surface 62. For convenience it is 27 assumed that the stripes 60 are of equal width w, with a 28 constant spacing d between centers, where d is greater than w 29 but not necessarily equal to 2w~ An output signal due to labeled molecules which are attached to the pattern may be f~

1 detected by scanning a light beam across the surface while 2 keeping the sample surface stationary or vice versa. As in the 3 first embodiment, a flat ribbon-like beam of light can be 4 created by using the cylindrical focusing lens 22 or by using an opaque mask with a slit aperture. If fluorescence labeling is 6 used to tag either the antigen or antibody, depending on the 7 mode of the assay, then the filter 24 is used to yield a beam 8 whose wavelength range is appropriate to excite the fluorescent 9 tags. The detection is accomplished using the imaging lens 26 and the filter 30 to ~ass light at the appropriate wavelength to 11 the detector Bifurcated fiber optics may also be used to 12 advantage to conduct the exciting and fluroscent light to and 13 from the pattern respectively.

The surface 62 could be made of a porous material 16 such as paper. If a porous material were used, the sample 17 solution would absorb into the material and the diffusion of 18 molecular species in the solution would be limitedO In this 19 case the tagging of the reaction could be done, for example, with fluorescence quenching so that binding of the tagged 21 molecule would give less fluorescence at the stripes than 22 between stripes even though diffusion were limited.

24 One method of examining the pattern consists of scanning the light beam back and forth across the array of 26 stripes 60 and specifically along a direction perpendicular to 27 the stripe length and parallel to the pattern surface, with a 28 constant linear speed s in each direction~ The scanning 29 amplitude is assumed to be larger than the overall width of the 1 array of stripes so that the detected fluorescence intensity I
2 resembles the waveform of Figure 5.

4 The phase~lock amplifier 44 of Figure 6(b) may be used to extract the pattern signal, from the embodiment of 6 Figure 4 r but this would require a second reference pattern 7 having the same periodicity as that of the antibody pattern.
8 Furthermore, the beam sweep amplitude and alignment with respect 9 to the pattern would in general have to be adjusted carefully to avoid noise artifacts due to loss of phase-lock at the 11 turn-around points of the sweep. The filter method of Figure lZ 6(a) would also present difficulties to the extraction of the ~3 pattern signal due to these turn-around artifacts. As a third 14 alternative, an autocorrelation technique may be used as shown in Figure 6(c). The technique of autocorrelation is a powerul 16 method for extracting a periodic signal superimposed on a 17 non-periodic background without the need for a reference 18 signal.
1~
The autocorrelation function, A(t), is given by 22 A (tj = (I(t'~ I(t' + t)>

24 where I (t') represents the signal strength (i.e. intensity3 ~5 detected at time t' and the brackets ~ > t denote an 26 integration, or running sum in the case of discrete sampling of 27 the intensity, over times t'~ In the example chosen of 2~ fluorescense labeling, I(t') represents the fluorescense 29 intensity at time t'. In the case of a reciprocating scan of constant speed s and a uniform spacing d between ad~acent ~1 - ~8 -1 stripes 60 (n in number), A (t) resembles the plot of Figure 7.

3 The correlation function A(t) produced by an 4 autocorrelator 70 basically possesses a series of peaks of spacing d/s in time, plus a peak at t=0 and one at the scan 6 period T, plus multiples of T~ superimposed on a baseline which 7 is closely related to the square of the average fluorescent 8 intensity obtained from the sample~ The actual A(t) obtained 9 differs somewhat because of differing relative peak heights, existence of extra peaks etc, from the ideal plot of Figure 7, 11 depending on the relationship between the beam sweep amplitude 12 and the pa~tern array size as well as the pattern alignment.
13 Nevertheless~ from one or more of the peak heights at t = d/s, 14 2d/s~ ....~ nd/s, and the size of the baseline the pattern intensity, independent of the background intensity, can be 16 determined by a detector 72. The peaks in A(t) in Figure 7 are 17 due to corre1ations of higher signal intensity corresponding to 18 spacings d, 2d; ...., nd. It is to be appreciated that the 19 autocorrelation technique of Figure 6(c~ also may be used with the first and second embodiments of Figures 1 and 3. Whether 21 the phase-lock amplifier or autocorrelator proves to be the 22 superior method of signal detection will depend on the 23 particular design of the apparatus, the detailed nature and 24 stren~th of the background false signals and the strength of the signal due to pattern-associated labels.

27 This invention based on binding in a spatial 28 pattern in principle permits multiple homogeneous assays to be 29 performed on the same sampl~. As a first example several ~0 ~:1 1 patterns may be located in close proximity on a given surface.
2 E~ch of the patterns would carry a different antibody, specific 3 to each antigen of interest. The patterns would be 4 distinguishable by having, for example~ different periodicities such as, different stripe separations d for the embodiment of 6 Figure 4 The different patterns are discriminated between by searching for different fundamental frequencies using the phase-8 lock detector or different correlation peak times using the 9 autocorrelator.

11 As a second example, the same pattern may be used 12 for each of several assays, but with a different identifying 13 label for each test. This could consist of a different 14 fluorescent wavelength, a different color carrier particle, etc.
By using different optical filters a measurement can be made of 16 each pattern intensity either simultaneously, using multiple 17 sets of detecting apparatus, or sequentially using a single 18 basic apparatus.

It shoul~ also be appreciated that a variety of 21 scanning schemes may be employed other than the ones already 22 discussed. These include the use of computer-controlled 23 stepping motors to accurately position a scanning beam or sample 24 location representing either rotation or linear displacement, as well as different ilumination/collection schemes, including the 26 use of fiber optics. For example, an optic fiber or fiber 27 bundle can be used to both illuminate a portion of the pattern 28 and detect the fluorescence. Reticles or masks, either 29 stationary or moving~ may be used to advantage in conjunction ~- with other optical elements to effectively provide for the 1 scanning of the spatial pattern in a manner which minimizes the 2 need for exacting alignment procedures. One co~ld focus the 3 pattern on a TV camera and scan the pattern electronically.

The existence of substantial numbers of unbound 6 labeled molecules in the assay solution increases the background 7 signal level (e.g. Io in Figure 2)~ regardless of the nature 8 of the labell In principle, however, the existence of a large 9 background, or baseline, level does not influence the measurement of the magnitude of the oscillating signal component 11 associated with the spatial pattern (e.g. ~I in Figure 2), since 12 the background signal due to freely-diffusing labeled molecules 13 is unrelated in requency or phase to the pattern. On the other 14 hand, fluctuations in the measured signal associated with imperfections in the antibody coating of the pattern or in 16 interposed optical surfaces necessarily remain fixed in phase 17 and frequency and become 'llocked in" to the spatial pattern, 18 thereby contributln~ to the apparent signal due to 19 pattern-associated labels. In this case, the use of labeled carrier particles to increase the signal strength associated 21 with the pattern, as discussed above, serves to proportionately 22 reduce the effect of these imperfections. The existence of 23 these non-random fluctuations which are physically fixed ~ith 24 respect to the pattern establishes a practical limit to the ultimate sensitivity which can be achieved by the spatial 26 pattern technique of the present invention~ since it is 27 dificult in practice to separate their contribution from that 28 of the desired pattern signal.
2~
~0 ~1 1 False signals associated with imperfections in 2 the surface on which the pattern is located, or of the pattern 3 coating itself, will not change during the assay measurement, so 4 a determination of the rate of change of the signal due to antigen-antibody binding to the pattern allows for the effective 6 removal of these unwanted signals. Also, as indicated above, 7 the phase-lock detector or autocorrelator may be designed 8 (perhaps in conjunction with a computer) to measure the signal 9 due to each segment of the pattern and effectively to ignore those segments which yield an abnormally large or small signal~
11 since such a large deviation in siqnal is probably due to an 12 imperfection in the pattern surface, optical surface or pattern 13 coating.

All of the embodiments discussed thusfar provide 16 for different means for determining the signal due to labeled 1~ molecules which become attached to segments of a spatial pattern 1~ as a consequence of antigen-antibody binding reactions~ All of 19 these embodiments share one common feature -- each requires the exis~ence of a surface onto which a particular spatial pattern 21 of antibody or antigen has been fixed. However, the design of a 22 homogeneous immunoassay based on the detection of binding to a 23 spatial pattern need not be confined to embodiments in which the 24 binding of labeled molecules must occur on a surface. Instead, the pattern may consist of a predetermined, preferential spatial 26 localization of labeled bound complexes which occurs within the 27 bulk volume of the assay solution, requiring no active surface~
2~ All that is required for this modification is a means for 29 causing the.labeled complexes to seek a set of locations or regions within the solution due to the application of a 1 particular external force field, with the remaining unbound 2 labeled molecules unafEected by that field and hence randomly 3 located throughout the solution. Once this localization of 4 bound labels has been achieved, the signal from the resulting three-dimensional pattern of labels can be readily detected 6 using one or more Oe the methods descrihed above in connection 7 with the first three embodiments (e.y. filtering, phase-lock 8 detection and autocorrelation).
One such external force field which can be 11 utilized is a magnetic field such as shown in Figure 8(a~. The 12 magnetic susceptibilities of typical macro-molecules are 13 prohibitively small. To magnify the forces produced by applying 14 an external magnetic field to the system~ one may use small magnetic carrier particles 80 which are designed to form stable 16 suspensions in a sample solution 82 contained in a sample cell 1~ 84 and which particles can be coated with antibody or antigen.
18 The-size of the particles 80 would typically lie in the range 19 0.01 to 50 micronsO For the purpose of describing the magnetic-field-induced pattern method, a competitive-type assay 21 with fluorescence labeling is assumed, where the analyte 22 consists of antigen. The magnetic carrier particles 80 are 23 coated with specific antibody (by covalent binding~ physical 24 absorption, etc.)~ After the analyte, labeled antigen (of known concentration) and antibody-coated magnetic particles have been 26 introduced into the assay cell volume, a particular magnetic 27 field is applied to the volume -- for example, a spatially 28 periodic field resembling a set of "fingers" of high field 29 strength, characterized by an approximate width w and center-to-center spacing d, as illustrated in Figure 8~a~ Such 1 a magnetic field pattern can be formed by using a set of magnets 2 86 which may be permanent magnets or a series of electromagnets~
3 appropriately spaced. In addition, a material of hiyh 4 permeability may be used to help shape the magnetic flux into the desired pattern within the solution volume.

When the magnetic field pattern is applied to the 8 sample solution 82, the magnetic carrier particles 80 experience 9 forces which cause them to move toward the finger-like zones of high magnetic flux in the solution. The speed and extent to 11 which the concentration of particles rises w;thin these zones 12 depends on the magnetic field strength and gradient acting on 13 the particles 80, their size, the solution viscosity and the 14 length o~ time during which the field is applied. This preferential localization of antibody coated carrier particles 16 due to magnetic fields forms the essentlal feature of the first 17 form of the fourth embodiment shown in Figure 8~a). The 18 resulting localization of the particles ~0 is shown in Figure 19 8(a~

21 As a result of the sequestering of the magnetic 22 particles 80 in these predetermined regions in the solution 82, 23 clearly the majority of labeled antigen mole~ules which become 24 bound to antibody are found within these high flux "channels" in solution. The remaining unbound labeled molecules plus any 26 background impurities which fluoresce are unaffected by the 27 applied field and therefore are located randomly throughout the 28 sample volume~ The resulting pattern of labeled complexes can 29 now be detected using any of the schemes discussed earlier. It should be appreciated that a complete sequestering or _ 34 1 localization of the magnetic carrier particles 80 within the 2 regions of high magnetic flux is not required for s~ccessful 3 measurement of the bound labels. The detection schemes previously discussed are in general capable of extracting a ~ particular frequency component or temporal correlation in the 6 overall signal whose amplitude is very small relative to the ~ size of the overall signal, most of which may consist of 8 baseline due to labeled molecules distributed uniformly 9 throughout the sample volume. Depending on the detailed nature of the magnetic flux pattern, the magnetic carrier particles 11 tend either to rather uniformly fill the high flux zones or 12 to be distributed inhomogeneously within those regions.
13 However, the assay technique can be made to function regardless 14 of the form of particle localizationO The carrier particles 80, themselves may give rise ~o a periodic signal during the 16 scanning bu~ this signal could be determined at the beginning of 1~ the reaction and substracted from subsequent signals.

19 Because the spatial pattern of labeled bound complexes as shown in Figure 8(a) is solely the result of the 21 application of an appropriate external magnetic field, clearly 22 the pattern can be made to appear or disappear or to generally 23 move throughout the assay solution 82 by judicious manipulation 24 of the applied fieldO For example, simple translation of the ~5 zones of high magnetic flux shown in Figure 8(a) can be easily 26 accomplished, for example, by mechanical translation of an 27 external array of magnets 86 when formed by permanen~ magnets.
28 Alternatively, the flux pattern can be moved when the magnets 29 are formed as a set of ele~tromagnets, whose windings are energized by a temporal sequence of currents designed to produce ~1 1 a "phased array" of fields~ with the net result that the zones 2 of high magnetic flux can be maAe to move spatially (i.e.
3 translate or rotate) in discrete increments of distance, without 4 the necessity of physical movement.

6 The fact that the pattern of localization of the 7 magnetic carrier particles 80 can be caused to move within the 8 solution 82 conveys at least two advantages to this assay 9 method. First, there is a greater flexi~ility and ease in designing the pattern scanning system needed to extract the 11 bound-label signal. In the case of fluorescent labeling, for 12 example, the pattern can be reciprocated back and forth 13 electrically, requiring no mechanical motion~ with the result 14 that no moving parts are required and minimal requirements ~5 for alignment of optical components. Second, the fact that the 16 magnetic carrier particles 80 can be moved back and forth 17 through the assay solution effectively eliminates the need for 18 gross stirring or agitation of the solution, normally required 19 ~o insure the adequate exposure of all free molecules to the surace~bound antibody. In the present embodiment, translation 21 of the magnetic carrier particles 80 serves to bring the coated ~2 antibody into efficient contact with the antigen and labeled 23 antigen throughout the assay solution 82~ This feature should ~4 ultimately speed up the binding reaction and reduce the overall time needed to perform a homogeDeous rate determination. Given 26 the fact that active coated surfaces are not required, the 27 fourth embodiment offers the potential advantages of ease of 2a manufacturing of the apparatus and elimination of potentially 29 troublesome periodic false signals due to surface-related imperfections-1 It should be understood that the geometry shown 2 in Figure 8(a) can be modified considerably. For example, the 3 magnetic flux pattern may be radially configured, so as to 4 resemble the fixed pattern of the second embodiment, in which case the pattern would be rotated in time rather than translated 6 in reciprocating fashionJ The pattern could consist of just two 7 stripes0 so that on alternate cycles the magnetic particles 8 preferentially reside either in one half of the sample solution g or the adjoining half. Rather than using ferromagnetic particles one may instead choose to use parama~netic particles.

12 Other kinds of externally applied force fields 13 can as well be utilized to set up a spatial pattern 3f labeled 14 bound complexes within the bulk assay solution 82 without the requirement of an active surface pattern~ For example, an 16 arrangement of electrodes 88 in contact with the assay solution 17 82 can be used to spatially translate charged molecules and/or 18 char~ed carrier particles ~0 in the solution as shown in Figure 19 8~b). The direction of translation depends on the sign of the charge and the speed of motion depends on the amount of charge 21 and the friction factor, in turn related to the size of the 22 particle 90. If a charged carrier particle 90 is used whose 23 charge is of opposite sign to that of the unbound labeled 24 molecules, then the bound complexes can be spatially separated from the free labels and a spatial pattern set up in solution.
26 By comparing, for example, the signal obtained from the pattern 27 to that obtained in the absence of the pattern (i.e. with the 2~ applied electric field first turned on, and then off), the 29 amount of bound label can be inferred. If the charge states and resulting mobilîties of the free labeled antigen and the labeled 7`1 1 antigen-antibody complexes are sufficiently different, two 2 species may be displaced with an applied electric field, thereby 3 setting up a spatial patternJ without the use of carrier 4 particles, which would be advantageous~

6 One could also use carrier particles 90 which are 7 electrically polarizable instead of charged. The applied 8 electric field would induce a dipole moment in the carrier 9 particle 90 and the gradient of the electric field would then exert a force on the particle so that the carrier particle could 11 be made to form a predetermined pattern in the solution 82.

13 It is also important to appreciate that the 14 pattern employed need not be spatially regular. Through the use of a computer and signal averaging/enhancing techniques, some 16 specific, non-periodic pattern may be detected. In facty the 17 pattern method may be used to replace the standard latex lB agglutination slide test now used for pregnancy testing, as 19 discussed above~ The binding of carrier particles onto a specific pattern may be detected with the eye since the 21 human visual system is excellent in being able to detect 22 non-random patterns or shapes in relatively noisy background 23 environments. The sensitivity associated with the apparatus 24 and method of the present invention is better than that which can be achieved when only random agglutination of the carrier 26 particles is the event which is to be recognized. Multiple 27 assays could then be performed by using a set of patterns which 28 are deposited on a surface in the form of different alphabet 29 letters, numbers or other simple geometric shapes whose identification would be r~latively unambiguousn 1 It is also important to appreciate that all of 2 the techniques associated with the formation of spatial patterns 3 described herein can be used to assay whole cells. Heret for 4 example, one may wish to detect the concentration of certain antigens located on the cell surface or to determine the 6 fraction of ce71s which contain a yiven molecule on their cell ~ surface. In this case, the antigen-antibody complexes consist 8 of cells bound to the antibody-coated pat~ern segments.

Although the present invention h~s been described 11 with reference to particular embodiments, it is to be 12 appreciated that various adaptations and modifications may be 13 made and the invention is only to be limited to the appended 14 claimsO

.9 2~

~2

Claims (23)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR
PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. Apparatus for providing an immunoassay of a binding reaction between a ligand and an antiligand, including a spatial pattern formed by a spatially periodic array of areas of antiligand material, ligand material dispersed to interact with the spatially periodic array of areas of antiligand material for producing a binding reaction between the ligand and the antiligand in the spatial pattern of areas, comprising means for labeling the binding reaction with a particular physical characteristic, a source of input energy and with the input energy at a particular spectrum for interacting with the particular physical characteristic of the labeled binding reaction, means for scanning the spatial pattern with the input energy at the particular spectrum for producing output energy having amplitude levels formed by a substantially random background component and a periodic component representing the labeled binding reaction, and means responsive to the output energy for detecting the periodic component representing the labeled binding reaction and for producing an output signal in accordance with the labeled binding reaction.

2. The apparatus for providing an immunoassay of Claim 1, additionally including a means for forming said spatial pattern selected from the group consisting of a cylindrical member and with the spatial pattern formed as a plurality of vertical stripes arranged circumferentially around the inside surface of the cylindrical member, a disc member and with the spatial pattern formed as a plurality of radial stripes arranged around a center point of the disc, a substantially flat member and with the spatial pattern formed as a plurality of horizontal stripes arranged across the flat member, and a spatially periodic force field to which carrier particles to which the antiligand is attached are responsive.

3. The apparatus for providing an immunoassay of Claim 2 wherein the means for forming the spatial pattern is a spatially periodic force field, the carrier particle is magnetic, and the force field is a magnetic field.

4. The apparatus for providing an immunoassay of Claim 2 wherein the means for forming the spatial pattern is a spatially periodic force field, the carrier particle is electrically charged, and the force field is an electric field.

5. The apparatus for providing an immunoassay of Claim 2 wherein the means for forming the spatial pattern is a spatially periodic force field, the carrier particle is electrically polarizable, and the force field is an electric field.

6. The apparatus for providing an immunoassay of Claim 2 wherein the means for forming said spatial pattern is one of said cylindrical member or said disc member, and the means for scanning includes means for rotating the means for forming said spatial pattern while the input energy is directed toward the spatial pattern.

7. The apparatus for providing an immunoassay of Claim 2 wherein the means for forming said spatial pattern is said flat member, and the means for scanning includes means for scanning the input energy back and forth across the flat member in a direction perpendicular to the horizontal stripes.

8. The apparatus for providing an immunoassay of Claim 1 wherein the means for scanning includes means for moving the spatial pattern while the input energy is directed toward the moving spatial pattern.

9. The apparatus for providing an immunoassay of claim 1 wherein the means for scanning includes means for maintaining the spatial pattern stationary while scanning the input energy across the stationary spatial pattern.

10. The apparatus for providing an immunoassay of claim 1 or 2 wherein the ligand material is dispersed in a liquid within said means for forming said spatial pattern.

11. A method for providing an immunoassay of a binding reaction between a ligand and an antiligand including the following steps, providing a spatial pattern formed by a spatial array of separate areas of antiligand material, providing a ligand material dispersed to interact with the spatial array of separate areas of antiligand material for producing a binding reaction between the ligand and the antiligand in the spatial pattern, labeling the binding reaction with a particular physical characteristic, providing input energy at a particular spectrum for interact-ing with the particular physical characteristic of the labeled binding reaction, scanning the spatial pattern with the input energy at the particular spectrum for producing output energy having amplitude levels formed by a substantially random background component and a non-random component representing the labeled binding reaction, and detecting the non-random component representing the labeled binding reaction and producing an output signal in accordance with the labeled binding reaction.

12. The method of claim 11 wherein the spatial pattern is formed as a plurality of vertical areas arranged circumferentially around a cylindrical member, and wherein the scanning includes rotating the cylindrical member while the input energy is directed toward the spatial pattern.

13. The method of claim 11 wherein the spatial pattern is formed as a plurality of radial areas arranged around a center point of the disc, and wherein the scanning includes rotating the disc member while the input energy is directed toward the spatial pattern .

14. The method of claim 11 wherein the spatial pattern is formed as a plurality of horizontal areas arranged across the flat member, and wherein the scanning includes scanning the input energy back and forth across the flat member in a direction perpendicular to the horizontal stripes.

15. The method of claim 11 wherein the labeled binding reaction between the ligand and antiligand is performed as a competitive assay including a known labeled ligand and an unknown unlabeled ligand both interacting with the antiligand to produce the binding reaction.

16. The method of claim 15 wherein the ligand to be labeled is attached to labeled carrier particles.

17. The method of claim 11 wherein the labeled binding reaction between the ligand and antiligand is performed as a sandwich assay including a known labeled antiligand and a known unlabeled antiligand and with an initial binding reaction between the ligand and the known unlabeled antiligand of a quantity sufficient to bind all of the ligand and with a subsequent binding reaction between the known labeled antiligand and the product of the initial binding reaction.

18. The method of claim 17 wherein the antiligand to be labeled is attached to labeled carrier particles.

19. The method of claim 11 wherein the scanning includes moving the spatial pattern while the input energy is directed toward the moving spatial pattern.

20. The method of claim 11 wherein the scanning includes maintaining the spatial pattern stationary while scanning the input energy across the stationary spatial pattern.

21. The method of claim 11 or 17 wherein the detecting includes filtering with a frequency range for passing the non-random component while discriminat-ing against the random background component.

22. The method of claim 11 or 17 wherein the detecting includes phase-lock detecting by amplifying the non-random component while discriminating against the random background component.

23. The method of claim 11 or 17 wherein the labeling of the binding reaction is a fluorescent label.